Poster Abstracts

1 downloads 0 Views 2MB Size Report
Oct 5, 2015 - importance in patients with chronic renal failure (CRF), and has a ..... strategies such as amorphous dispersion could be evaluated to increase oral exposure. ... Faculty of Pharmaceutical Sciences, Hiroshima International University, Kure, Japan ..... In this study, consecutive linear equations successfully.
Drug Metabolism Reviews

ISSN: 0360-2532 (Print) 1097-9883 (Online) Journal homepage: http://www.tandfonline.com/loi/idmr20

Poster Abstracts To cite this article: (2015) Poster Abstracts, Drug Metabolism Reviews, 47:sup1, 41-297, DOI: 10.3109/03602532.2015.1071933 To link to this article: http://dx.doi.org/10.3109/03602532.2015.1071933

Published online: 05 Oct 2015.

Submit your article to this journal

Article views: 1092

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=idmr20 Download by: [112.134.238.102]

Date: 10 February 2016, At: 07:30

http://informahealthcare.com/dmr ISSN: 0360-2532 (print), 1097-9883 (electronic) Drug Metab Rev, 2015; 47(S1): 41–297 ! 2015 Taylor & Francis. DOI: 10.3109/03602532.2015.1071933

ABSTRACTS

Poster Abstracts P1. LEVERAGING ISOTOPES FOR THE QUANTIFICATION OF METABOLITES FROM NONRADIOLABELED MASS BALANCE STUDY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Su-Er W. Huskey, Chun-qi Zhu, Grazyna Ciszewska, Roger Kramp, James Mangold, and Tapan Ray Department of Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, East Hanover, NJ, USA A radiolabeled tracer is typically used in human ADME (mass balance) studies to facilitate the understanding of routes of excretion and elimination pathways of new drug candidates following oral administration during phase II of drug development. However, this approach may not be suitable under certain circumstances where there is the potential for high sustained radioactivity exposure. For example, the rationale to conduct a human ADME with radiolabel might be questioned, (i) when the drug and/or metabolites have unusually long half-lives, (ii) when the drug and/or metabolites remain localized in a specific tissue for an unusually long time, or (iii) if the target population is pediatric. Thus, our objective was to provide an alternative approach to obtain relevant human ADME information when a human ADME using radio-tracer is not suitable. Compound Z was synthesized with dual labels (stable isotope labeling and 14C labeling) to explore the applicability of this approach using rat as an animal model. A mass balance study was conducted in rats in two subgroups: (Group 1) [13C614C]compound Z and (Group 2) [12C]compound Z. Following oral dosing of compound Z to rats, excreta were collected daily for two days. A metabolite mixture of [13C614C]compound Z was generated in vitro using hepatocytes and/or liver microsomes. The metabolite mixture of [13C614C]compound Z was spiked into fecal homogenates from rats dosed with [12C]compound Z. The fecal homogenates from both the subgroups were extracted and analyzed under identical conditions. Recoveries of compound Z and metabolites in excreta were expressed in terms of % of dose. Findings from both subgroups will be presented and the implication of this approach will be discussed in this presentation.

P2. USING TARGET ENHANCEMENT AND TOF-MRM TO ACHIEVE FEMTOGRAM ON COLUMN SENSITIVITY FOR QUANTITATION OF DRUGS IN HUMAN PLASMA Yun W. Alelyunas1, Mark Wrona2, and Paul Rainville2 1 Department of Pharmaceutical Life Sciences, Waters Corporation, Milford, MA, USA and 2Pharmaceutical & Life Sciences Group, Waters Corporation, Milford, MA, USA Purpose: To determine and compare quantification attributes including detection limit, quantification limit, and linear dynamic range of full scan MS, MSE, and Tof MRM (target enhancement) mode of data acquisition using Synapt G2-S QTof high resolution mass spectrometer (HRMS). Methods: Serial diluted solutions, containing alprazolam, verapamil, buspirone, and clopidogrel, were prepared using quenched and diluted human plasma. The concentration of the solution ranges from 100 pg/mL down to 1 pg/mL in seventeen 1:1 serial dilution. The quenched and diluted human plasma was prepared by adding 3 volumes of acetonitrile to human plasma. After centrifugation to remove precipitated protein, the supernatent was transferred to a new vial and diluted with 1 volume of H2O. The LC system used was an ACQUITY I-Class UPLC and BEH UPLC C18 1.7 mm 2.1  50 mm column. A generic linear gradient from 2 to 60 B% in 4 min was used (mobile phase A was Water with 0.1% formic acid, and mobile phase B was acetonitrile containing 10% MeOH and 0.1% formic acid (v/v)). The flow rate was 0.6 ml/min and column temperature was 60 C. Data were collected using several resolution modes of acquisition including full scan MS, MSE, and Tof MRM (target enhanced high duty cycle) modes of operation. Data acquisition was performed using Masslynx. Data processing was completed using Targetlynx. Results: With widespread adoption of HRMS in drug discovery and development, there are increasing interest in using HRMS for quantitation in addition to traditional qualitative studies. In this way, one can take advantage of high mass resolution capability of HRMS to achieve selective/sensitivity, which may otherwise be challenging using tandem quadruple MS due to complex matrix interferences. Recently, Waters introduced a new target enhancement mode of data acquisition (Tof MRM) on Synapt instrument for routine analysis. In this mode, a precursor ion is selected and the Tof pusher is synchronized with the precursor or a product ion to maximize duty cycle for a target m/z range. This 100% duty cycle should result in an increase in response, sensitivity and selectivity. The present study compares linearity and sensitivity for four model compounds determined using one of three modes of available data acquisition on Synapt G2S: Tof MRM, full scan MS, and MSE. Results indicate that using Tof MRM there is 4 to 410 fold of LOQ enhancement with excellent linear dynamic range

42

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

from 3.6 to 3.9. The focused range of data acquisition also resulted in 90% filesize reduction compared to either full scan MS or MSE. Conclusion: Tof MRM mode of operation on QTofs provides added benefits of enhanced sensitivity and large linear dynamic range often needed for quantitative studies. The observed excellent fentogram on column sensitivity and upto to 3.9 log linear dynamic range suggest that Synapt G2S is capable of performing both quantitative and qualitative studies. In this way one will be able to obtain kinetics as well as parent or metabolite information. The smaller file size of TofMRM also makes the data more manageable for high throughput studies.

P3. SIMPLE DERIVATIZATION FOR THE SIMULTANEOUS LC/MS QUANTIFICATION OF AMINO ACID NEUROTRANSMITTERS IN EXTRACELLULAR FLUID FROM RAT BRAIN MICRODIALYSIS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Minli Zhang1, Chengwei Fang2, and Gennady Smagin3 1 Department of Drug Metabolism and Pharmacokinetics, AstraZeneca, Waltham, MA, USA, 2Department of Drug Metabolism and Pharmacokinetics, AstraZeneca, Wilmington, DE, USA, and 3Department of Neuroscience, AstraZeneca, Wilmington, DE, USA Quantification of amino acid-based neurotransmitters in extracellular fluids, such as those in the neuron synapse, presents a challenge to the analytical chemistry because of their absence of UV- or fluorescence-detectable functional groups, and low sensitivities in mass spectrometric analysis. This report describes a novel use of thesuccinimide reagent, N-a-BocL-tryptophan hydroxysuccinimide ester (Boc-TRP), for the pre-column derivatization to quantify amino acid neurotransmitters in the rat brain microdialysis samples. The Boc-TRP derivatization was rapid and quantitative in phosphate buffer (pH 7.4) at room temperature. The derivatizedneurotransmitters were suitable for rapid LC/MS quantification with less than 3-min chromatographic separation. The Boc-group in the derivatized product generated unique fragmentation patterns in the triple quadruple mass spectrometric analysis under Multiple Reaction Monitoring mode, therefby significantly increasing the specificity and sensitivity. The derivatization and rapid LC/MS quantification method developed in this study showed a linear dynamic range of 1–4 nM to 2000 nM with coefficient greater than 0.992. Since there was no concentration and reconstitution in the sample workup process, this derivatization approach simplified the neurotransmitters quantification of the brain microdialysis samples.

P4. EVALUATING IN VIVO RECEPTOR OCCUPANCY OF DOPAMINE D2 RECEPTOR IN RAT BRAIN USING LIQUID EXTRACTION SURFACED ANALYSIS MASS SPECTROMETRY Jun Tadano, Toichiro Yamada, Kenichi Watanabe, Tetsuya Nakagawa, and Masashi Yabuki Preclinical Research Laboratories, Sumitomo Dainippon Pharma Co., Ltd., Osaka, Japan Introduction: In vivo receptor occupancy is a key factor needed to understand the pharmacological effects of the central nervous system (CNS) drugs. Although studies of receptor occupancy are primarily conducted using isotope-labeled tracers (3H, 14C), labeled tracers are rarely available at the early discovery stage. As LC-MS/MS analysis can be used to estimate the specific distribution in a target area even without labeled tracers, it is often used to evaluate receptor occupancy in the brain. However, despite the sensitivity of this methodology, one drawback is that different regions of the brain need to be collected and quantified separately. Recently, liquid extraction surface analysis mass spectrometry (LESA-MS) has been reported as a useful tool to estimate the specific distribution of drugs in tissue sections (Eikel et al., 2011). Here, we investigated the specific distribution of raclopride (a dopamine D2 receptor probe) in the rat brain and in vivo receptor occupancy using LESA-MS. Method: In order to evaluate its specific distribution, raclopride (0.003 mg/kg) was administered intravenously to rats, whose brains were then collected after 10 min and sliced into 10 mm thick sections. Olanzapine was orally coadministered at doses between 0.1 and 100 mg/kg to evaluate receptor occupancy. The concentration of raclopride in each slice was analyzed by LESA-MS. Additionally, we quantified each site (frontal cortex, striatum, hippocampus, and cerebellum) by the traditional LC-MS/MS method so that the quality of results obtained through LESA could be effectively determined. Result: Region-specific distribution of raclopride was observed and the ratio between peak areas in striatum to those in the cerebellum was determined by LESA-MS to be approximately 3.5 at low doses. Co-administration of olanzapine dose-dependently decreased this ratio. The LESA-MS method yielded results that were highly consistent with those by traditional LC/MS/MS. In conclusion, our study suggests that the LESA-MS technique is a useful tool for evaluating the receptor occupancy of drugs using non-labeled compounds.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

43

Reference Eikel D, Vavrek M, Smith S, et al. (2011). Liquid extraction surface analysis mass spectrometry (LESA-MS) as a novel profiling tool for drug distribution and metabolism analysis: the terfenadine example. Rapid Commun Mass Spectrom 25:3587–3596.

P5. ANTIOXIDANT AND RENOPROTECTIVE ACTIVITY OF 2-HYDROXYPROPYL-BETA-CYCLODEXTRIN IN NEPHRECTOMIZED RATS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Makoto Anraku1, Daisuke Iohara1, Kaneto Uekama2, Fumitoshi Hirayama1, and Masaki Otagiri1 1 Faculty of Pharmaceutical Sciences, Sojo University, Kumamoto, Japan and 2DDS Research Institute, Sojo University, Kumamoto, Japan Oxidative stress, which involves the production of excessive levels of reactive oxygen spices, is a pathogenic condition of great importance in patients with chronic renal failure (CRF), and has a great impact on their survival. Thus, the development of effective anti-oxidant therapy for treating CRF would be highly desirable. One of the proposed mechanisms of oxidative stress in CRF involves the accelerated production of oxidants, such as uremic toxins and their reduced renal clearance. Therefore, the removal of such substances from the systemic circulation would be expected to lead to a reduction in oxidative stress in CRF. Cyclodextrins are a class of carbohydrates that are finding increased medicinal use, due to their ability to complex with hydrophobic molecules. One cyclodextrin in particular, 2-hydroxypropyl-b-cyclodextrin (HP-b-CyD), is used as a carrier to solubilize lipophilic drugs and is itself being considered as a therapeutic agent for the treatment of Niemann–Pick Type C disease, due to its ability to mobilize cholesterol. Therefore, HP-b-CD might also be useful as a therapeutic agent for the treatment of CRF, due to its ability to mobilize uremic toxins. In this study, the effect of HP-b-CyD and other CyDs on oxidative stress and CRF was investigated using 5/6 nephrectomized rats. The ingestion of HP-b-CyD over a 4-week period resulted in a significant decrease in serum indoxyl sulfate and creatinin levels, compared with the ingestion of a-CyD, b-CyD or g-CyD. The ingestion of HP-b-CyD also resulted in an increase in antioxidant potential, compared with the ingestion of a-CyD, b-CyD or g-CyD. These results suggest that HP-b-CyD has a high antioxidant activity as well as renoprotective effects in CRF model rats. In in vitro studies, the affinity of HB-b-CyD for a uremic toxin such as indole was much higher than that of other CyDs, including a-CyD, b-CyD or g-CyD. These results suggest that the ingestion of HP-b-CyD can result in a significant reduction in the levels of pro-oxidants, such as uremic toxins, in the gastrointestinal tract, thereby inhibiting the subsequent development of oxidative stress in the systemic circulation.

P6. FINDING OUT WHAT HAPPENED TO YOUR BIOMOLECULE WITH THE HELP OF LC-MS/MS: A CASE FOR AN INTACT PROTEIN PK ANALYSIS Carrie Xu, Baomin Xin, and Timothy Olah Pco/bar, Bristol-Myers Squibb, Princeton, NJ, USA LC-MS/MS is rapidly becoming an invaluable tool and a complementary technique to ligand-binding assays for the assessment of biomolecules in early discovery since it has the capability to provide both quantitative and qualitative information on biologic drugs or biomarkers in biological matrices. In addition, observed differences in results obtained by ELISA and LC-MS/MS methods can provide a distinctive insight into the fate of therapeutic proteins following administration in early PK/PD studies. Recently, after overcoming many technical challenges, a sensitive LC/MS/MS-based method (LLOQ ¼ 1.0 nM) was developed to measure a small intact protein in PK studies following the administration of a therapeutic protein. Although the ELISA and LCMS/MS determinations matched well for early time points, there were significant differences at later time points. The differences in values sparked a systematic investigation into the potential metabolic instability of the bimolecular drug candidate. Carefully designed experiments combined with LC-high resolution accurate mass (LC-HRMS) analysis provided us with in vivo oxidation on one of the methionines to explain this discrepancy and helped to direct the project team with an appropriate path forward. The LC/MS/MS method development, PK results and metabolic instability findings will be presented in this poster.

P7. MOUSE LIVER TRANSPORTER QUANTITATION WORKFLOW USING AB SCIEX TRIPLETOF 6600 System Xi Chen1, Lei Xiong1, Suma Ramagiri1, Mollah Sahana1, Liling Liu2, Yuzhong Deng2, Brian Dean2, Yuan Chen2, and Xiaorong Liang2 1 AB Sciex, Redwood City, CA, USA and 2Genentech, South San Francisco, CA, USA Purpose: Transport proteins play an important role in the absorption, distribution, and elimination of a variety of drugs. In recent years, a large number of transporters, both efflux (ATP-binding cassette (ABC) family) and influx (solute carrier (SLC) family members), have been identified and well-characterized. However, the abundance of these transporters in intestine, liver, and kidney has not been accurately quantitated due to technique challenges. The expression level of transporter is one of the key determinants in translating in vitro data to in vivo for the understanding of drug disposition, drug–drug interactions, and variability in drug response and toxicity. This work aims to build a robust liquid chromatography-mass spectrometry (LC-MS) workflow on the AB Sciex TripleTOFÕ 6600 platform for the quantitation of a variety of SLC and ABC drug transporters

44

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

expressed in on the mouse liver plasma membrane. Method: Mouse liver was harvested and perfused with saline to remove the blood. Liver samples were homogenized in pH 7.4 phosphate buffer containing protease inhibitors using an Omni bead ruptor. The plasma membrane fraction was prepared using a rate-dependent centrifugation through a sucrose density gradient. The final plasma membrane fraction was suspended in a 20 mM Tris-HCl buffer. These transporter proteins at the plasma membrane were denatured, reduced, alkylated, and digested using trypsin. Protein specific peptides generated by trypsin digestion were used as targeted surrogates for quantitation of each transporter on TripleTOF 6600 instrument. This targeted list of peptides was generated by a combination of protein identification experiment followed by mouse protein database UniprotkB database search and Skyline software. For the absolute quantitation of transporters, two stable isotope-labeled specific peptides were employed. One is used as the surrogate analyte for the specific peptide, while the other as internal standard. Results and conclusions: The use of TripleTOF 6600 system offers several advantages over the triple quadruple instrument for transporter quantitation, which includes higher specificity due to higher resolution and accuracy for MS and MS/MS fragmentation data, simplified and faster method development, and the ability to quantify larger number of transporters in a single run. Several transporter proteins of known concentrations in mouse liver plasma membrane were used as models to validate this workflow. Our preliminary data show that measured concentrations of these transporters using the proposed workflow are comparable to those reported in literatures. Currently, we are exploring a wide variety of these transporters aiming to simultaneously quantify them in a single MS analysis.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P8. A SIMPLE HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY FOR DETERMINING LAPATINIB AND ERLOTINIB IN HUMAN PLASMA Masahiro Ohgami1, Masato Homma2, Yoshiharu Suzuki3, Shoichi Mitsuhashi4, Fumie Fujisawa4, Takayuki Kaburagi5, Keiko Uchiumi5, Yutaka Yamada5, Hiroko Bando6, and Keiji Takei1 1 Department of Pharmacy, Ibaraki Prefectural Central Hospital, Kasama, Japan, 2Department of Pharmaceutical Sciences, Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Japan, 3Department of Pharmacy, Tsukuba University Hospital, Tsukuba, Japan, 4Department of Medical Oncology, Ibaraki Prefectural Central Hospital, Kasama, Japan, 5Department of Respiratory Medicine, Ibaraki Prefectural Central Hospital, Kasama, Japan, and 6Department of Breast and Endocrine Surgery, University of Tsukuba, Tsukuba, Japan Background: Lapatinib and erlotinib, tyrosine kinase inhibitors (TKIs) which suppress the growth of cancer cells via inhibiting signal transduction pathways, are used for the treatment of breast cancer and non-small cell lung cancer, respectively. It has been still unclear whether or not therapeutic drug monitoring (TDM) for lapatinib and erlotinib is useful for assessing the clinical outcomes and adverse events because of limited TDM data. One of the reasons for the limitation of lapatinib and erlotinib TDM is the assay method, in which high-performance liquid chromatography with tandem mass-spectrometric detection (LC-MS/MS) has been commonly used for previous studies. Since LC-MS/MS is not generally available in the clinical institutions such as hospitals, alternative way to determine for both drugs is required. Objective: We developed a simple high-performance liquid chromatography (HPLC) ultraviolet method for the determination of lapatinib and erlotinib in human plasma. Methods: Optimized conditions can permit detection of lapatinib and erlotinib in the same chromatogram, so either compound can be used as an internal standard (IS) for another determinant. An aliquot of plasma sample (200 ml) spiked with 50 ml of IS solutions was treated with 700 ml of acetonitrile to precipitate proteins. After vortex mixing for 20 s, the sample was centrifuged at 5000 g for 10 min, and the supernatant was evaporated to dryness under the nitrogen gas stream. The residue was reconstituted with 100 ml of mobile phase and 20 ml was injected into the HPLC system. Lapatinib and erlotinib were separated on octadecylsilyl silica-gel column using a mobile phase consisting of a mixture of acetonitrile:methanol:water:trifluoroacetic acid (26:26:48:0.1) pumped at a flow rate of 1.0 mL/min. The detection wavelength was set at 316 nm. Results: The calibration curves of lapatinib and erlotinib were linear (r ¼ 0.9999) in the range of 0.625–10 mg/mL and 0.125–4 mg/mL, respectively. The recoveries for both lapatinib and erlotinib at the plasma concentration of 0.125–5.0 mg/mL were higher than 92.3% with the coefficients of variation (CVs) less than 1.8%. The CVs for intra- and inter-day assay of lapatinib and erlotinib were less than 3.6% and 4.3%, respectively. The present method can be applied for determining lapatinib or erlotinib in plasma samples obtained from patients receiving of either TKIs. Conclusion: Present HPLC can be used for practical TDM for lapatinib or erlotinib in completely same conditions.

P9. SIMULTANEOUS DETERMINATION OF 3-CAFFEOYLQUINIC ACID, 3,5-DICAFFEOYLQUINIC ACID, ACACETIN, APIGENIN, DIOSMETIN, LINARIN AND LUTEOLIN IN RAT PLASMA BY LC-MS/MS Ju Hyun Kim, Soon Sang Kwon, Eun Nam Kim, and Hye Suk Lee Drug Metabolism and Bioanalysis Laboratory, College of Pharmacy, The Catholic University of Korea, Bucheon, South Korea A sensitive and rapid liquid chromatography-tandem mass spectrometry (LC-MS/MS) method was developed to simultaneously evaluate the pharmacokinetic properties of seven active constituents of Chrysanthemum zawadskii extract including 3-caffeoylquinic acid (3-CQA), 3,5-dicaffeoylquinic acid (3,5-diCQA), acacetin, apigenin,diosmetin, linarin and luteolin in

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

45

Downloaded by [112.134.238.102] at 07:30 10 February 2016

rat plasma. Following a simple protein precipitation by methanol, the analytes were separated on a C8 reverse phase column with a gradient mobile phase system of methanol and water containing 0.1% of formic acid. Selected reaction monitoring (SRM) in positive and negative electrospray-ionization modes was performed and fully validated. The analytical responses were linear (r240.99) over the concentration ranges from 5 to 1000 ng/mL for 3-CQA and 3,5-diCQA and from 1 to 200 ng/mL for acacetin, apigenin, diosmetin, linarin and luteolin. The inter- and intra-day precisions of analytes were 515% and accuracies ranged from 85 to 115%. Each analyte was stable in the stability assessments, including a short-term (on ice for 2 h), post-preparation (4  C for 24 h in autosampler) and three freeze-thaw cycle (80  C to room temperature). The coefficients of variation of matrix effect were within 15%, which demonstrated that the matrix effects of the analytes were negligible. The developed method was applied to pharmacokinetic study after oral administration of Chrysanthemum zawadskii extract to rats at the dose of 500 mg/kg. The plasma concentrations of 3-CQA, 3,5-diCQA and linarin were well-determined while those of other analytes were not measurable. Based on the evidences that flavonoids are extensively metabolized into glucuronide metabolite in vivo, rat plasma samples treated with b-glucuronidase were analyzed by the same procedure. After the treatment of b-glucuronidase, the plasma exposure levels of all analytes were measurable and higher than intact samples indicating glucuronide metabolites are the majority plasma circulating forms of analytes. In the case of linarin, however, the plasma concentrations were not significantly different in the presence and absence of b-glucuronidase, which means the glucuronidation of linarin is negligible. The glucuronide conjugates of analytes except linarin were confirmed by in vitro using rat liver microsomes incubated with uridine diphosphoglucuronic acid. Finally, the developed method is reliable for the pharmacokinetic study of Chrysanthemum zawadskii extract in rats.

P10. A NOVEL TOOL FOR OPTIMIZING PEPTIDE DRUGS’ HALF-LIFE: MEMBRANE-FREE PLASMA PROTEINBINDING ASSAYS Hinnerk Boriss Sovicell, GmbH, Leipzig, Germany Measuring plasma protein binding of peptide drugs is important for dosage scaling and helps optimizing the peptides’ half-life. Typically, peptides exhibit difficulties to penetrate dialysis or ultrafiltration membrane pores, which makes it difficult to use these techniques to assess plasma protein binding of peptide drugs. Another challenge is to maintain peptides in solution in pure buffer. The poster presents a new method that assesses binding to an immobilized phase and plasma in the liquid phase. Thus, the assay eliminates the need for membranes and assay compartments without plasma, thereby solving the two biggest challenges for measuring protein binding of peptides.

P11. SIGNIFICANCE OF PH ON THE CYTOTOXIC POTENTIAL OF THE WATER DISINFECTION BY-PRODUCT IODOACETIC ACID Vicki M. Richardson1, Susan D. Richardson2, Mary P. Moyer3, Jane Ellen Simmons1, and Anthony B. DeAngelo1 1 ORD/Nheerl/Istd, US EPA, Research Triangle Park, NC, USA, 2Department of Chemistry and Biochemistry, University of South Carolina, Columbia, SC, USA, and 3INCELL Corporation, San Antonio, TX, USA Disinfection of water provides significant health benefits through reduction of morbidity and mortality from water-borne disease. However, disinfection by oxidizing chemicals results in the formation of disinfection byproducts (DBPs). To date, more than 600 DBPs have been identified; the US EPA regulates eleven. Iodoacids have been identified in drinking water disinfected by either chlorination or chloramination. Based on vitro studies in Chinese hamster ovary cells, iodoacetic acid (IAA) is the most toxic DBP assessed to date. As the iodine group in the chemical structure of IAA is possibly labile, it has been suggested that IAA is unstable under the low-pH conditions in the stomach. Due to the possibility of IAA being acid labile, its bioavailability, as well as its toxic potential may be reduced. To determine whether IAA was stable at the low pH conditions of the stomach, biological and chemical endpoints were examined at pH 2 and at body temperature (37  C). To assess the biological stability of IAA, we used epithelial cells derived from normal human colon (NCM460) whose cytochrome P450 and glutathione S-transferase theta 1 (GSTT1) activities are comparable to those in the intact large intestine. Human colon cells were selected for this study as some epidemiology studies have reported associations, albeit weak, between consumption of disinfected water and colon cancer. NCM460 cells were treated (96 h) with IAA previously incubated in acidic water (pH 2) or control water (pH 7) for 1 h at 37  C and tested for cytotoxicity using a crystal violet growth inhibition assay (IC50). The IC50 for IAA previously incubated in acidic water and control water was 1.0  105 M and 1.3  105 M, respectively. IAA and any potential degradation products formed during the low pH, body temperature incubation were also analyzed using gas chromatography/mass spectrometry (GC/MS). The results showed no apparent degradation of IAA following exposure to pH 2 as compared to the neutral pH controls. This study, which combined analytical chemistry and toxicology, demonstrates that IAA remains intact and has similar toxic potency under conditions comparable to those of the human stomach. We have also demonstrated the utility of NCM460 cells in testing chemicals for cytotoxicity.

46

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Declaration of interest This abstract of a proposed presentation does not necessarily reflect the policies of the U.S. EPA.

P12. EVALUATING FACTORS AFFECTING LOW-BIOAVAILABILITY IN VIVO FOR A NOVEL, LOW CLEARANCE SMALL MOLECULE INHIBITOR FOR AN ONCOLOGY TARGET

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Gauri Deshmukh, Bianca M. Liederer, and Lesley Murray Department of Drug Metabolism and Pharmacokinetics, Genentech Inc., South San Francisco, CA, USA Compound A was identified during discovery phase for a selective small molecule inhibitor against a novel oncology target. The objective of this study was to elucidate the potential cause of low oral (PO) exposure (F ¼ 11.6%) in female CD1 mice following 5 mg/kg PO administration (MCT suspension) of a low-clearance compound (Cl ¼ 25 mL/min/kg) following 1 mg/ kg intravenous bolus (IV) administration (30% PEG400/70% water solution). Metabolism, permeability, stability, and solubility were evaluated to determine if these factors were contributing to the low oral exposure. To determine whether the first pass effect was occurring, compound A was dosed at 5 mg/kg via intraperitoneal (IP) (30% PEG400/70% water solution) and intramuscular (IM) (10% HPbCD solution) administration in female CD-1 mice. This yielded F values of 22.5% and 10.3%, respectively. In addition, mouse liver perfusion was conducted and FH (fraction of the amount entering portal blood that escapes hepatic metabolism and biliary secretion) was 64%, suggesting that the liver had a low to moderate contribution to low bioavailability post oral dosing. To further determine the relative contribution of liver and intestine to first-pass clearance, compound A was incubated in CD-1 mouse liver and intestinal S9 fractions. The mouse liver S9 + NADPH incubations resulted in disappearance of 60% of parent, while the intestinal S9 incubations showed very little disappearance of parent. Compound A showed moderate metabolic stability with Cl equal to 52 mL/min/kg and 58 mL/min/kg in liver microsomes and hepatocytes, respectively. In silico prediction for apparent permeability (Papp A to B) was high with 12.5  106 cm/s and no efflux, with a ratio of 0.95. Apparent permeability in MDCK cells was moderate with 3.7  106 cm/s and an efflux ratio of 3.9. To determine stability, compound A was incubated at 37  C for 24 h in FeSSIF (simulated intestinal fluid in fed state, pH 5) or FaSSIF (simulated intestinal fluid in fasted state, pH 5) or at 4  C in acetonitrile (LC-MS/MS solvent) at either 0.5 mM or 306 mM (the dose concentration administered in vivo). The compound was stable under all conditions. Solubility was tested in FeSSIF and FaSSIF and compound A had a low solubility in both FeSSIF (90.4 mM) and FaSSIF (53.6 mM). The available data suggest that hepatic and intestinal metabolism, permeability, solvent and intestinal fluid stability were not the key factors affecting poor bioavailability. The low solubility of the compound appears to be the main contributor to the observed poor bioavailability, suggesting that various formulation strategies such as amorphous dispersion could be evaluated to increase oral exposure.

P13. ACAMPROSATE INTESTINAL ABSORPTION IS PARTLY PARACELLULAR AND PARTLY CARRIER-MEDIATED Frans Franek1, Sibylle Neuhoff2, and Bente Steffansen1 1 Department of Pharmacy, University of Copenhagen, Copenhagen, Denmark and 2Department of Translational Science in DMPK, Simcyp Ltd (a Certara Company), Sheffield, UK Acamprosate is a small (MW ¼ 181.2), hydrophilic (log Po:w ¼ 3.57) sulfonic acid compound fully ionized at intestinal fluid pH (pKa ¼ 1.8), has negligible metabolism and is mainly eliminated renally (BDDCS Class III). The aim of this study is to investigate the different mechanisms of intestinal absorption of acamprosate. In vitro and in vivo information relevant for acamprosate absorption, disposition and elimination was determined or collated from the literature and applied in combination with physicochemical data on acamprosate in a two-step modeling approach for membrane permeability and absorption. Firstly, passive diffusional transcellular, paracellular and effective-intestinal segmental permeabilities (Ptrans, Ppara and Peff, man, passive, respectively) were estimated using a newly-developed, mechanistic, physiologically-based absorption model in MatLab. Secondly, physiologically-based pharmacokinetic (PBPK) modeling of acamprosate, using the estimated permeabilities, was performed using the ‘‘Advanced Dissolution, Absorption and Metabolism’’ (ADAM) model within the Simcyp Population-based Simulator (V12-R2). As performance verification of the full PBPK-model, simulated area under the plasma concentration-time profile (AUC) and oral fraction absorbed (fa), following single intravenous or oral acamprosate doses over a range of 666–2310 mg were compared to reported data for healthy volunteers. In vitro determined IC50values for acamprosate inhibition of transporter-mediated influx or efflux of substrates to selected transporters were studied in Caco-2 cells. Acamprosate disposition and elimination was adequately predicted, as the predicted and observed AUC, after 666 mg infusion of acamprosate, were 43.2 ± 8 (Mean ± SD) and 39.7 ± 8.5 mg ml1 h1, respectively. Results regarding intestinal absorption of acamprosate indicate that the Ppara of 9.8  108 cm/s in the duodenum and jejunum together with yet not identified membrane transporter mediated permeability (Pcarr) account for acamprosate intestinal absorption, whereas Ppara in ileum and colon (1.3  1010 cm/s) as well Ptrans throughout the intestine (2.8  1010 cm/s) is negligible. Neither the investigated apical influx, i.e. OATP2B1, ASBT, PEPT1, EAAT1/3; TAUT, PAT1, B0+ATrBAT, nor the basolateral bi-directional transporter(s), OSTa/b, seem to be important for acamprosate

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

47

carrier mediated permeability. In conclusion, acamprosate intestinal absorption in human seems to be partly paracellular and partly (currently unidentified) transporter mediated.

P14. PREDICTING EFFECT OF SOLUBILIZERS ON THE ORAL ABSORPTION OF A LOW-SOLUBLE DRUG BY USING RAT SIMULATED GASTROINTESTINAL FLUIDS: THE IMPORTANCE OF REGIONAL DIFFERENCES IN MEMBRANE PERMEABILITY AND SOLUBILITY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yusuke Tanaka, Toshiyuki Baba, and Shunji Nagata Faculty of Pharmaceutical Sciences, Hiroshima International University, Kure, Japan Background: Currently, simulated intestinal fluid such as FaSSIF and FeSSIF has been developed to assess drug dissolution in the gastrointestinal (GI) tract. However, these simulated intestinal fluids are developed based on total bile acid and phospholipid concentrations only in upper small intestinal fluid of dogs and humans. Our previous report revealed large differences in these components between the upper and lower small intestine in rats (Tanak et al., 2012). Considering the different environment in each GI fluid, evaluating the solubility and permeability of drugs in the lower region of the GI tract, as well as the upper small intestine, is considered to be very important for accurate prediction of oral absorption. Purpose. This study aimed to develop a novel approach for predicting the oral absorption of low-solubility drugs and the effect of solubilizers by considering regional differences in solubility and permeability within the GI tract using newly developed simulated upper and lower intestinal fluids. Methods. Simulated intestinal fluids were prepared based on in vivo bile acid and phospholipid concentrations in the upper and lower small intestine of fasted rat. The saturated solubility of albendazole (AZ), a low soluble drug, was measured using these simulated fluids with or without solubilizers. To estimate apparent permeability by in situ measurement, AZ was dissolved in each simulated GI fluid, and each drug solution was injected into the corresponding GI segment. Fraction absorbed (Fa) at time t after oral administration was calculated from equation for estimation of the maximum absorbable dose (MAD) using these solubility and permeability data. Results. The saturated solubility and apparent permeability of AZ estimated using the upper and lower simulated GI fluids were about 15.8 and 34.3 mg/mL and 3.9 and 3.3  105 cm/sec, respectively. When adding dimethyl sulfoxide and b- cyclodextrin as solubilizers into the two simulated GI fluids, the solubility and permeability of AZ were almost unchanged in the both fluids. On the other hand, the AZ solubility was significantly increased by adding Vitamin E-TPGS while the apparent permeability was to some extent decreased. The predicted Fa values using these solubility and permeability data were found to correspond closely to the in vivo data, and its absorption enhancement by solubilizers was also predicted accurately. Conclusion. These results indicated the importance of evaluating regional differences in drug solubility and permeability in order to predict oral absorption of low-solubility drugs accurately. The new methodology developed in the present study could be useful for new oral drug development.

Reference Tanaka Y, Hara T, Waki R, Nagata S. (2012). Regional differences in the components of luminal water from rat gastrointestinal tract and comparison with other species. J Pharm Pharm Sci 15:510–518.

P15. ORAL DELIVERY OF PENTAMIDINE FOR HUMAN AFRICAN TRYPANOSOMIASIS Sihyung Yang, and Michael Zhuo Wang Department of Pharmaceutical Chemistry, The University of Kansas, Lawrence, KS, USA Human African trypanosomiasis (HAT), a neglected tropical disease, is caused by the protozoan parasite Trypanosoma brucei and affects millions of people in Sub-Saharan Africa. HAT appears in two stages: first stage, when the parasites are confined in the blood and lymph, and second stage, when the parasites infect the central nervous system. Although pentamidine is used to treat490% of all first-stage HAT, it requires parenteral administration via 7–10 daily intramuscular injections due to its poor oral bioavailability. This creates significant logistical and societal problems in the remote rural areas of Africa, where HAT is endemic. Previous efforts using a prodrug approach resulted in the development of pafuramidine (the bis-methamidoxime prodrug of furamidine, which is a structural analog of pentamidine) that demonstrated good oral activity against first-stage HAT in clinical trials. Unfortunately, pafuramidine development program was suspended after reports of unexpected renal toxicity. Here, we aimed to directly enhance the oral bioavailability of pentamidine by increasing its intestinal absorption using intercellular junction-modulating E-cadherin peptides (ECPs). E-cadherin proteins belong to the cadherins superfamily of cell– cell adhesion molecules and forms adherens junctions. The model ECP, HAV6, substantially increased pentamidine permeability across the MDCK-MDR1 cell monolayers (7- to 9.5-fold) compared to vehicle (water) and a scrambled peptide HAV6scr. Pharmacokinetic study in fasted rats showed HAV6 (50 mg/kg) substantially increased the systemic exposure (plasma AUC0–24 h and Cmax) of pentamidine (30 mg/kg) after oral administration compared to vehicle (water; 4.0- and 4.6-fold). HAV6 was also stable (490% remaining in 4 h) in simulated gastric fluid with pepsin, rat small-intestinal mucosal scrapings and human plasma. These results strongly suggest that HAV6 has a remarkable effect on the gut absorption of pentamidine and that HAV6-based oral formulations of pentamidine may have desirable efficacy against first stage HAT.

48

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P16. CHARACTERIZATION OF THE SECRETORY TRANSPORT OF DRUGS IN THE MOUSE SMALL INTESTINE USING USSING CHAMBER SYSTEM

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Takeshi Nakayama1, Kazuya Maeda1, Shinya Fukizawa1, Yukio Kato2, Paul A. Dawson3, and Hiroyuki Kusuhara1 1 Laboratory of Molecular Pharmacokinetics, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan, 2Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan, and 3Division of Pediatric Gastroenterology, Hepatology and Nutrition, Emory University School of Medicine, Atlanta, GA, USA Background: Intestinal absorption is a critical process for orally administered drugs to determine their systemic exposure. The molecular mechanisms of the membrane transport of drugs have been studied extensively, and the involvement of P-glycoprotein (P-gp) and the breast cancer resistance protein (BCRP) in the suppression of the intestinal absorption of several drugs has been well-established. It has been reported recently that the secretory transport from the blood into the intestinal lumen might also contribute to the elimination of certain drugs from the systemic circulation. However, the molecular mechanisms of intestinal drug secretion have not been well characterized so far. Objective: The objective of our study was to clarify the contribution of transporters to the secretory transport of drugs in the mouse small intestine using the Ussing chamber system. We focused on the involvement of P-gp and BCRP on the apical side and organic solute transporter a/b (OSTa/b) on the basal side in the intestinal secretion of drugs (rosuvastatin, pitavastatin, apixaban, rivaroxaban, fexofenadine, and talinolol). Methods: The secretory transport of tested drugs was quantified across the intestinal tissue sections prepared from the duodenum, jejunum, and ileum in mice with the Ussing chamber. The inhibitory effects of several compounds on the basal-to-apical transport of the drugs were also investigated. The contribution of P-gp and BCRP to the drug efflux from the ileum section to the apical side was investigated by comparing the drug transport across the ileum section obtained from Mdr1a/1 b(–/–)/Bcrp(–/–) and wild-type mice. The role of Osta/b in the drug influx at the basal side of the ileum section was investigated using ileum sections prepared from Osta(–/–) and wild-type mice. Results: The basal-to-apical transport of drugs was 6- to 10-fold higher in the ileum sections compared with the duodenum and jejunum sections. The secretory transport of all the drugs across the ileum section was significantly lower in Mdr1a/1 b(–/–)/Bcrp(–/–) mice than in wild-type mice, and the absorptive transport was significantly higher in Mdr1a/1 b(–/–)/Bcrp (–/–) mice. Fluvastatin (200 mM) and apixaban (100 mM) inhibited the secretory transport and tissue accumulation of the tested drugs. On the other hand, the secretory transport of drugs across the ileum section was not decreased by the knockout of Osta, denying the involvement of OSTa/b in the basal uptake of drugs in the overall secretion. Conclusions: The secretory transport of the tested drugs was higher in the ileum of mice compared with the duodenum and jejunum, implying the selective expression of active transport systems contributes to their intestinal secretion. In secretary transport, P-gp and/or BCRP are responsible for the drug efflux from the intestinal epithelial cells to the lumen, whereas OSTa/ b, which is predominantly expressed in the rodent ileum, does not contribute to the drug uptake from the basal side; therefore, clarification of its molecular mechanisms is needed.

P17. ESTIMATION OF IN VIVO DISSOLUTION PROFILE OF ORALLY-DISINTEGRATING TABLET DOSED WITHOUT WATER BY DECONVOLUTION METHOD Masayo Oishi, Yuko Mori, and Yoshiro Tomono Department of Clinical pharmacology, Pfizer Japan Inc., Tokyo, Japan Introduction: Orally-disintegrating tablet (ODT) is defined as ‘‘a solid dosage form containing medicinal substances which disintegrates rapidly, usually within a matter of seconds, when placed upon the tongue’’ (CDER, 2008). Main purpose to develop ODT is to provide a formulation which rapidly disintegrates in saliva without the need for chewing or liquids, i.e., dosing without water. Therefore, it is essential to confirm bioequivalence (BE) of ODT dosed without water to a reference formulation which is usually immediate release (IR) oral solid formulation dosed with water. For general IR formulation, in vitro in vivo correlation (IVIVC) is well investigated; however, there are few investigations on IVIVC for ODT dosed without water. Conventional dissolution tests are conducted with plenty of liquid under the assumption that a tablet disintegrates and dissolves in gastrointestinal tract fluid. However, it is not the case of ODT dosed without water. It disintegrates in mouth and dissolves both in saliva and in gastric fluid. It is considered that this complicated situation is one of the causes why the estimation of in vivo dissolution profile of ODT without water from in vitro data is difficult. Objective: The purpose of our investigation is to estimate in vivo dissolution profile of ODT without water by deconvolution method using data of human BE studies and assess the impact of change of in vivo dissolution profile on concentration profiles by convolution method. Methods: Plasma concentration profiles dosed with and without water in BE studies of commercially available ODTs in Japan market (pitavastatin, naftopidil and sildenafil) were deconvoluted by profiles of reference IR data to estimate in vivo dissolution profiles. In addition, simulations by convolution method were conducted using estimated in vivo dissolution profile and reference IR profile with some different conditions e.g., different absorption/elimination profiles, blood sampling points etc. Results: Discontinuous in vivo dissolution profiles were estimated for ODT without water whereas the profile for ODT with water is similar to a reference IR formulation. The simulation by convolution showed discontinuous in vivo dissolution profile could impact on plasma concentration profile of ODT without water especially for compounds with rapid absorption, i.e. compounds with short Tmax. Conclusion: In vivo dissolution profile of ODT without water which is hardly estimated from in vitro dissolution data was estimated by

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

49

deconvolution method using human plasma concentration profiles and it was considered that absorption rate is one of the important factors which can affect plasma concentration profile of ODT without water when discontinuous in vivo dissolution occurs.

Reference CDER. (2008). Guidance for industry orally disintegrating tablets food and drug administration. Silver Spring, (MD): Center for Drug Evaluation and Research (CDER).

P18. PREDICTING HUMAN ORAL BIOAVAILABILITY USING EX VIVO PORCINE INTESTINAL TISSUE AND PBPK MODELING

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Joost Westerhout, Sieto Bosgra, Evita van de Steeg, Maria LH Vlaming, and Heleen M. Wortelboer TNO, Zeist, the Netherlands A reliable prediction of the oral bioavailability in humans is crucial for pharmaceutical and food industry. Here, we present a new approach to evaluate the predictability of human oral bioavailability of compounds with different chemical characteristics from permeability studies with healthy porcine intestinal tissue combined with physiologically-based pharmacokinetic (PBPK) modeling. To this end, a representative set of compounds was chosen and their apparent permeability (Papp) data in ex vivo porcine intestinal tissue were determined and compared to absolute human oral bioavailability data reported in literature as previously described (Westerhout et al., 2014). In the present study, the Papp data of in total of 21 compounds across porcine intestinal tissue were determined and PBPK modeling was used to predict the oral bioavailability of these compounds. For comparison, the same approach was taken to predict human oral bioavailability from Papp data of the same compounds determined with Caco-2 cells and human intestinal tissue mounted in the Ussing chamber system. Of the current set of 21 compounds, 11 were known to be substrates for one or more active transporters present at the intestinal epithelium, whereas 13 were known to be substrates for one or more of the different metabolizing enzymes expressed in the intestinal epithelium. Rather than applying a generic modeling approach for the prediction of human oral bioavailability based on in vitro/ex vivo permeability data, we applied a compound-specific approach which incorporated the differences in abundance levels of transporters and metabolizing enzymes between porcine intestine, Caco-2 cells and human intestine. To account for differences between the in vitro/ex vivo models, the abundance levels of transporters and metabolizing enzymes in porcine intestine, Caco-2 cells and human intestine were also incorporated in the PBPK model. Results show that human oral bioavailability can be predicted reasonably well for the passively transported compounds using a generic modeling approach, independent of which in vitro model is used (R2 ¼ 0.64 for porcine intestine, 0.77 for Caco-2 and 0.80 for human intestine). However, for the compounds that are substrates for one of the active transport systems and/or metabolizing enzymes, the predictability of human oral bioavailability could be improved by applying a more specific PBPK approach.

Reference Westerhout J, Steeg EV, Grossouw D, et al. (2014). A new approach to predict human intestinal absorption using porcine intestinal tissue and biorelevant matrices. Eur J Pharm Sci 63:167–177.

P19. CHANGES IN DRUG ABSORPTION BY PHARMACEUTICAL EXCIPIENTS IN RAT SMALL INTESTINE Yusuke Takizawa, Hisanao Kishimoto, Takahito Furuya, and Katsuhisa Inoue Department of Biopharmaceutics, Tokyo University of Pharmacy and Life Sciences, Tokyo, Japan Purpose: Pharmaceutical excipients are inactive ingredients commonly used in the formulation of oral dosage forms, and considered to have little effect on the pharmacokinetic and pharmacological properties of an active pharmaceutical ingredient. However, it has recently been reported that the several types of detergents that have been classified as pharmaceutical excipients can inhibit the function of membrane transporters such as P-gp and BCRP, suggesting possible interactions of other pharmaceutical excipients in drug absorption. We, therefore, examined the effects of 21 pharmaceutical excipients on the drug absorption in rat small intestine. Methods: The effects of pharmaceutical excipients on the membrane permeability were examined by in vitro non-everted gut sac method using 5-carboxyfluorescein (5-CF) and b-naphthol as absorption markers of paracellular and transcellular routes, respectively. The change of the membrane permeability was evaluated by the change of the area under the drug concentration-time curve (AUC) calculated from the concentration of each substrate in the extra tissue chamber. The concentrations of pharmaceutical excipients were in the range of clinical dosage. Results and discussion: The membrane permeability of 5-CF was significantly increased in the presence of methyl-b-cyclodextrin in jejunum, but such effect was not observed in ileum. Similarly, several excipients such as sodium carboxymethyl starch, low-substituted hydroxypropyl cellulose and cross carmellose sodium, were found to increase the membrane permeability of 5-CF in jejunum, but not in ileum. The effect of hydroxypropyl cellulose was dependent on the concentration used in the experiments, as shown that the 5-CF absorption was decreased at 0.02%, but significantly increased at 0.20% in both jejunum and ileum. On the other hand, the membrane permeability of b-naphthol was significantly decreased by 0.08% corn starch in ileum, but not in jejunum. Similarly, in the case of microcrystalline cellulose, the b-naphthol absorption was significantly changed in jejunum, but not in ileum.

50

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Other excipients used as binders (hydroxypropyl cellulose and hydroxypropyl methylcellulose) did not alter b-naphthol absorption in ileum, except ethyl cellulose, which increased those in both jejunum and ileum. Conclusions: We demonstrated that pharmaceutical excipients have abilities to alter drug absorption by modulating paracellular and transcellular routes, and their effects are different between jejunum and ileum in rat small intestine. Therefore, it would be necessary to further examine the effects of all types of pharmaceutical excipients and to classify them based on the influence on intestinal drug permeability.

P20. ESTIMATION OF PEDIATRIC DOSAGES CONSIDERING RAPID DEVELOPMENT OF PHYSIOLOGICAL FUNCTIONS AFTER BIRTH BY APPROXIMATED SIMPLE FORMULAE: IMPLEMENTATION AND COMPARISON WITH DOSAGES DESCRIBED IN THE LABELING

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Akihiro Hisaka1, Tomoko Mayuzumi2, Hiroki Koshimichi2, and Hiroshi Suzuki2 1 Department of Geriatric Pharmacology and Therapeutics, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan and 2The Department of Pharmacy, The University of Tokyo Hospital, Tokyo, Japan Background and objective: Previously, we reported an estimation method for pediatric dosages which is applicable to newborn mature infants and also premature infants considering rapid development of hepatic and renal functions after birth (Fujino et al., 2012). In this study, simple formulae were newly implemented by approximating the previous method, and they were applied to 45 oral and 28 intravenous drugs of which pediatric dosages have been described in the Japanese product labeling for comparison. To estimate pediatric dosages with these formulae, age of the child and two basic parameters of adult, i.e. oral clearance and urinary excretion ratio of unchanged form, are required. Method: Pediatric dosages were calculated to achieve the same serum concentration of drugs with the therapeutic dose in adults. The pediatric hepatic clearance was estimated considering the expression change of CYP enzymes and the development of hepatic weight. The pediatric renal clearance was estimated from the development of GFR. The age of children was divided into three parts, and changes of the hepatic and the renal clearances were respectively approximated with linear equations for each part. The equations were implemented in Microsoft Excel, thus the pediatric dosage for arbitrary age can be calculated as a ratio versus adult (per body weight) from oral clearance and urinary excretion ratio of unchanged form in adults. The equations were applied to drugs of which pediatric dosages have been described in the Japanese product labeling for comparison. Result: In the original estimation method for pediatric dosages based on physiologically based pharmacokinetics, various factors including development of the hepatic and renal functions and changes in the protein binding were considered. In this study, consecutive linear equations successfully approximated the previous calculation within 20% error considering only development of the hepatic and renal functions. We compared the calculated dosages with those in the Japanese product labeling for ages from 0.25 to 6 years old. For 45 oral drugs, the median of the ratio (estimated versus described) was 1.15 and the interquartile range was 0.93–1.36. For 28 intravenous drugs, the median was 0.93, and the interquartile range was 0.64–1.48. Overall, the estimated dosages were agreed excellently with those described in the product labeling. However, estimated dosages of four drugs were less than half of the described ones among oral drugs. Discussion: The present study suggested that the proposed equations of pediatric dosage estimation were practical and reliable for various ages and situations. It would be expected that pediatric dosages are calculated reasonably for many drugs with this method even though the information is unavailable from the product labeling. On the other hand, for drugs whose discrepancies between the estimated and described dosages are significant, it seems necessary to check their efficacy, safety and pharmacokinetics in pediatric population as possible.

Reference Fujino et al. (2012). World conference on pharmacometrics, Seoul, Sep.

P21. PREDICTING HEPATIC CLEARANCE OF SLOW METABOLIZED COMPOUNDS USING HEPATOCYTE SANDWICH PERFUSION SYSTEM Chan Way Ng, Yu Yu, Lei Xia, and Hanry Yu Yong Loo Lin School of Medicine, Department of Physiology, National University of Singapore, Singapore, Singapore Predicting hepatic clearance of a new chemical entity using in vitro model is always of great interest to the pharmaceutical industry. With the knowledge of hepatic clearance and targeted therapeutic concentration, therapeutic dose can be estimated for first-in-man study. However, extrapolating in vitro data to in vivo results often leads to undeprediction of hepatic clearance, especially for the compounds with low turnovers. This is mainly due to the lack of biotransformative enzyme activities and the limitation of the incubation period in the in vitro models. We have previously developed an immediate-overlay sandwich culture that can restore hepatocyte polarity and improve liver specific functions (Ng et al., 2006). We have further developed a laminarflow perfusion bioreactor that can maintain liver-specific functions in the sandwich culture for two weeks (Xia et al., 2009). This study evaluated the ability of our hepatocyte sandwich perfusion system to predict hepatic clearance of three commercially available compounds with slow clearance (56 mL/min/kg). The amount of parent compounds remained in the culture was quantified using LC/MS analysis, while hepatic clearance was predicted using well-stirred model. The predictions of the hepatic clearance fall within 2-fold of observed human in vivo clearance value. In addition, Monte Carlo Simulation showed that the hepatocyte sandwich perfusion culture has higher sensitivity in predicting hepatic clearance compared to conventional

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

51

hepatocyte suspension. Taken together, these results demonstrated the capability of laminar-flow immediate-overlay hepatocyte sandwich perfusion system to predict hepatic clearance by providing a larger clearance prediction window compared to hepatocyte suspension.

References Ng S, Han R, Chang S, Ni J, et al. (2006). Improved hepatocyte excretory function by immediate presentation of polarity cues. Tissue Eng 12:2181–2191. Xia L, Ng S, Han R, et al. (2009). Laminar-flow immediate-overlay hepatocyte sandwich perfusion system for drug hepatotoxicity testing. Biomaterials 30:5927–5936.

P22. CLEARANCE PREDICTION FOR METABOLICALLY ELIMINATED COMPOUNDS: IN VITRO TO IN VIVO EXTRAPOLATION (IVIVE) OR ALLOMETRY?

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Hong Mei1, Sujal Deshmukh2, and Christopher Gibson3 1 Department of Pharmacokinetics Pharmocodynamics and Drug Metabolism, Merck Research Laboratories, Kenilworth, NJ, USA, 2Department of Pharmacokinetics Pharmocodynamics and Drug Metabolism, Merck Research Laboratories, Boston, MA, USA, and 3Department of Pharmacokinetics Pharmocodynamics and Drug Metabolism, Merck Research Laboratories, West Point, PA, USA Accurate prediction of hepatic clearance (CLh) using in vitro systems is critical for estimating human dose and drug exposure at the drug discovery stage. A general trend of under-prediction for drugs metabolized by CYPs and UGTs has been reported using in vitro systems (Ito & Houston, 2005; Kilford et al., 2009). We have evaluated several in vitro models of human clearance prediction using state-of-the-art methodologies for determination of intrinsic clearance (CLint) and binding (plasma and nonspecific). First, drugs were selected where metabolic elimination was dominant and clearance values by intravenous dosing were available in human and at least 2 preclinical species. ‘‘Difficult to predict’’ compounds, i.e. low CLh, high-plasma protein binding and large species differences in liver extraction ratio (Eh) were intentionally included. Twenty-eight compounds (16 Merck proprietary and 12 commercial) were selected with a representative distribution of physiochemical properties. These compounds were tested in the following in vitro metabolizing systems: liver microsomes, hepatocytes and recombinant human CYPs (rCYPs) to determine their CLint at a concentration of 0.1 uM (10  lower than published methods). This low-substrate concentration should prevent underestimation of CLint for compounds with low Km or poor solubility. Microsomal binding, plasma protein binding, and blood to plasma ratios were all re-measured under the same conditions. Binding in hepatocytes was predicted from measured binding in microsomes using an in silico model. Various CLh prediction methods (e.g. well-stirred model with and without an animal correction factor, Lave’s hybrid allometric method (Lave, 1997)) were evaluated. For CYP substrates, rCYP and human hepatocytes provided the best clearance prediction without the need of any correction factor. Human liver microsomes also provided adequate CLh prediction for CYP substrates. Lave’s hybrid allometric method generated a consistent underestimation of Clh for all the compounds. For UGT substrates, human hepatocytes tended to provide an underestimation of CLh that required an animal correction factor to improve the prediction. Allometry, on the other hand, poorly predicted CLh for CYP substrates with low CLh (Eh50.3), but provided an adequate prediction for CYP substrates with Eh40.3. Allometry also showed an adequate prediction for UGT substrates with Eh from 0.03 to 0.6.

References Ito K, Houston JB. (2005). Prediction of human drug clearance from in vitro and preclinical data using physiologically based and empirical approaches. Pharm Res 22:103–112. Kilford PJ, Stringer R, Sohal B, et al. (2009). Prediction of drug clearance by glucuronidation from in vitro data: Use of combined cytochrome P450 and UDP-glucuronosyltransferase cofactors in alamethicin-activated human liver microsomes. Drug Metab Dispos 37:82–89. Lave T, Dupin S, Schmitt C, et al. (1997). Integration of in vitro data into allometric scaling to predict hepatic metabolic clearance in man: Application to 10 extensively metabolized drugs. J Pharm Sci 86:584–590.

P23. PREDICTION OF DRUG CLEARANCE IN GENENTECH: A RETROSPECTIVE ANALYSIS OF IN VITRO IN VIVO EXTRAPOLATION METHODS AND ALLOMETRY METHODS Jialin Mao, Yuan Chen, Peter Fan, Jane Kenny, Bianca M. Liederer, Xingrong Liu, Matthew Wright, and Cornelis E. C. A Hop Department of Drug Metabolism and Pharmacokinetics, Genentech, South San Francisco, CA, USA Prediction of human drug clearance is a critical aspect of drug discovery in order to select good drug candidates and, in preclinical development, to plan for clinical trials. The objective of this study was to retrospectively assess the reliability of 12 in vitro and in vivo methods in prediction of human drug clearance for 11 proprietary compounds for which in vitro, preclinical pharmacokinetic data and single oral dose human pharmacokinetic data were available. The 12 methods include five in vitro in vivo extrapolation (IVIVE) and seven allometry methods. The five IVIVE methods were human hepatocytes without the binding correction, human liver microsomes without the binding correction, human hepatocytes with the binding correction,

52

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

human liver microsomes with the binding correction and Patrick Poulin method. The seven allometry methods were simple allometry, free allometry, allometry with maximum life-span potential, fu corrected intercept method (FCIM), CL-QSAR, rat with rat plasma protein-binding correction and monkey method. Eleven compounds evaluated are anti-cancer protein kinase inhibitors and their oral clearance was collected from oncology patients. Human hepatic clearance was estimated from the observed oral clearance using the well-stirred model assuming complete oral absorption. In this limited data set, there are seven low clearance drugs (ER50.3), three moderate clearance drugs (0.3  ER  0.7) and one high-clearance drug (ER40.7), and 2-, 1.5- and 1.25-fold were used as criteria of successful prediction for the three categories, respectively. Human hepatocytes without the binding correction (with 80% accuracy in 10 compounds) and FCIM (with 73% accuracy in 11 compounds) provided somewhat better prediction of apparent human clearance than the other methods. It was intriguing to notice that hepatocytes without the binding correction provided reasonable predictions in mouse, dog and cynomolgus monkey, but did not perform well in rat for these compounds. This observation suggested that the elimination pathways might be different between species, which highlighted the importance of the mechanistic understanding of the clearance pathways across species. In conclusion, convergence of different prediction methods combined with more mechanistic understanding may increase the confidence in the human clearance prediction prospectively.

P24. RETROSPECTIVE ANALYSIS FOR PREDICTION OF HUMAN PHARMACOKINETICS USING CYNOMOLGUS MONKEY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yusuke Aratsu, Toshio Taniguchi, Yukihiro Nomura, and Motohiro Kogayu Drug Metabolism & Pharmacokinetics Research Laboratories, Japan Tobacco Inc, Takatuki, Osaka, Japan Purpose: In recent years, we have a greater opportunity to use cynomolgus monkeys for pharmacokinetics studies in drug discovery stage. Since there is high amino-acid sequence homology between human and cynomolgus monkey for not only drugmetabolizing enzymes such as CYPs but also many transporters, cynomolgus monkeys are expected to be a useful animal species for the prediction of human pharmacokinetics profile. However, there are not many research papers reporting cynomolgus monkey’s pharmacokinetics in detail, and there are many issues to be solved, such as individual differences in pharmacokinetics for cynomolgus monkey. In this study, pharmacokinetics profiles in cynomolgus monkey will be investigated using several drugs which can be eliminated from body through various pathways in order to establish an easy-to-use approach for the prediction of human pharmacokinetics based on monkey IVIVE (In vitro–In Vivo extrapolating), which is applicable in drug discovery stage. Methods: We administered intravenously Midazolam (CYP3A substrate), Antipyrine (CYPs substrate), Raloxifene (UGT substrate), Fexofenadine (a drug excreted in bile) and Famotidine (a drug excreted in urine) to six cynomolgus monkeys, calculated pharmacokinetics parameters such as total clearances (CLtot) and volume of distribution (Vdss), and evaluated the individual difference in pharmacokinetics. Moreover, in vitro clearances (CLint) of above and several other drugs were obtained using the monkey microsomes and hepatocytes. Then, human pharmacokinetics parameters were predicted using the monkey IVIVE results obtained by the in vitro clearances (CLint) and in vivo clearances (CLtot), and the validity of the IVIVE was evaluated by comparing human predicted parameters with actual values. Results and discussion: There was little or no individual difference in CLtot and Vdss of Midazolam, Raloxifene, Fexofenadine and Famotidine when intravenously administered to the cynomolgus monkeys. We are planning to evaluate several more drugs and will present an easy-to-use approach for the prediction of human pharmacokinetics based on monkey IVIVE which is applicable in drug discovery stage.

P25. TRANSLATIONAL ADME IN DRUG DISCOVERY – PREDICTIONS OF PHARMACOKINETICS, TOXICOKINETICS AND DRUG INTERACTION POTENTIAL USING PBPK MODELING TOOLS Minxia Michelle He1, Jennifer Ott1, Shobha Bhattachar2, Daniel Mudra1, Stephen Hall1, and Nita J. Patel3 1 Department of Drug Disposition, Eli Lilly & Co Rsch Labs, Indianapolis, IN, USA, 2Department of Product Design & Developability, Eli Lilly & Co Rsch Labs, Indianapolis, IN, USA, and 3Eli Lilly & Co Rsch Labs, Indianapolis, IN, USA The applications of PBPK modeling tools (Simcyp and Gastro+) have demonstrated significant impact on drug discovery and development. The predictive tools were effectively utilized in three main areas as indicated by the case examples. PBPK modeling/prediction of human PK using Gastro+ or Simcyp (the models were first validated with rodent and large animal data) has helped to identify gaps for early discovery lead molecules. As an example, the human clearance of a lead compound A was predicted to be rapid using the in vitro and in vivo ADME/PK data. Subsequent simulation based on the predicted human PK parameters and animal efficacious exposure suggested that a high dose (3000 mg QD or 400 mg BID) would be required to achieve a clinical PK profile with sufficient target coverage. However, modified input parameters suggested that increasing potency 3-fold and reducing clearance by 50% would significantly reduce the human dose to 200 mg QD. Further sensitivity analysis provided information about whether changes of phys–chem properties such as solubility, clogP or pKa would impact the PK prediction. Those results provided clear SAR directions to optimize the lead molecules to have desired human ADME properties. Secondly, PBPK modeling with Gastro+ or Simcyp can be used for the prediction of toxicokinetics at the high dose range, especially for the compound with absorption limitation that required enabling formulation. Incorporating multiple parameters including solubility/permeability/precipitation time, the predicted AUC exposure for Compound B matched well with the actual data. Further effort to refine the model is ongoing and the prediction would help dose selection for toxicology studies

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

53

to achieve targeted exposure. Finally, PBPK models are powerful tool in predicting risk for clinical drug-drug interactions. For CYP3A4 time-dependent inhibition (TDI), the liver effect can be determined by in vitro liver microsomal TDI assay or hepatocytes assay in the presence of human plasma protein. However, the gut effect could not be readily addressed experimentally. Compound C was evaluated for CYP3A4 TDI with mixed results from the microsomal and hepatocytes assays. Simcyp modeling was performed to predict both hepatic and gut 3A4 interaction based on the liver microsomal data. Although there was a less than 2-fold change in AUC, the gut intrinsic clearance of midazolam in the presence of Compound C was reduced from 11 to 2 L/h (5-fold). The significant gut effect based on the prediction has allowed the risk assessment of CYP3A4 related drug-drug interaction.

P26. PREDICTION OF MAJOR CLEARANCE PATHWAYS OF DRUGS FROM THEIR PHYSICOCHEMICAL PROPERTIES AND SCREENING DATA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kota Asahina, Yusuke Aratsu, Yuki Soejima, Mitsuru Takahashi and Motohiro Kogayu Drug Metabolism & Pharmacokinetics Research Laboratories, Japan Tobacco Inc, Takatsuki, Osaka, Japan Purpose: The major clearance pathway(s) of drugs is very important pharmacokinetic information in the process of drug discovery and development. In vivo total clearance is the sum of all clearance mechanisms, such as metabolism and biliary/renal excretion, and may be dominated by one or more major pathways. Predicting major clearance pathways of a drug is useful not only for understanding its pharmacokinetic properties, such as the potential of drug-drug interactions in humans, but also for optimizing pharmacokinetic properties and selecting appropriate screening tools for drug candidates. In this study, we established a prediction method of the major clearance pathways of drugs based on their physicochemical properties and in vitro screening data obtained in drug discovery. Method: A dataset of 75 compounds (marketed drugs and under development compounds) was used. Human major clearance pathways of marketed drugs were collected from the literature. Rat clearance pathways were determined based on the concentrations of parent drug and its metabolites in bile and urine from bile cannulated rats. Physicochemical parameters, Caco-2 permeability (apparent permeability: Papp) and intrinsic clearance (CLint) in rat and human liver microsomes/hepatocytes were obtained by conventional methods. Result: Sum of the percentage of dose excreted in urine and bile showed increasing trend with increase in polar surface area (PSA) and with decrease in Papp across Caco-2. Compounds with high PSA (470) and low Caco-2 permeability (Papp51.0  106 cm/s) were excreted in bile and/or urine as parent drug (more than 50% of dose). Of the compounds excreted in urine and bile, those with low molecular weight (MW) tended to be excreted in urine, and those with high MW tended to be excreted in bile: the MW weight threshold was estimated to be &400Da. On the other hand, the major clearance pathway of the compounds with low PSA (PSA570) was metabolism. Furthermore, these compounds were classified into CYP-metabolized type and Non-CYP-metabolized type based on the CLint ratio (CLint in liver microsomes/ CLint in hepatocytes). Result and discussion: PSA, Caco-2 permeability, MW, and CLint ratios were found to be the determinants of the major clearance pathways. We established a diagnostic tree to predict the major clearance pathways of drugs based on the physicochemical parameters and in vitro screening data. This diagnostic tree is useful in selecting an appropriate prediction method of clearance and drug-drug interactions in humans and can aid chemists in optimizing the pharmacokinetic properties of drug candidates in early discovery.

P27. ABSTRACT WITHDRAWN

P28. PLATED HEPATOCYTE RELAY ASSAY (PHRA) FOR THE ESTIMATION OF INTRINSIC HEPATIC CLEARANCE AND METABOLITE PROFILE OF SLOWLY METABOLIZED COMPOUNDS Albert P. Li1, Chi-Chi Peng2, Utkarsh Doshi1 and Chandra Prakash2 1 In vitro ADMET Laboratories LLC, Columbia, MD, USA and 2Department of Drug Metabolism and Pharmacokinetics, BiogenIdec, Cambridge, MA, USA The current practice of metabolic stability screening of new chemical entities (NCE) has led to the collection of chemicals that are resistant to metabolism by in vitro metabolic experimental system such as human liver microsomes and human hepatocytes. The evaluation of metabolic fate and estimation of in vivo intrinsic clearance of these ‘‘slowly metabolized’’ compounds is a technical challenge, as in general, little metabolism can be observed using the routine short-term incubations with human liver microsomes, S9, or human hepatocyte suspensions. To overcome this challenge, a relay long-term metabolism assay has been developed with primary cultured hepatocytes. The relay assay consists of firstly culturing plateable cryopreserved human hepatocytes (pooled from three donors) on day 1 for approximately 4 h to allow attachment, followed by incubation with the chemicals to be evaluated for a 24 h period. On day 2, the incubation medium from day 1 is transferred to a new culture of human hepatocytes prepared from the same lots used in day 1 followed by another 24 h of incubation (1st relay). The process is repeated on days 3 (2nd relay), 4 (3rd relay), and 5 (4th relay). With this novel relay method, a chemical can be evaluated for metabolism by human hepatocytes for up to 5 days. This Plated Hepatocyte Relay Assay (PHRA) was validated with 10 model drugs with in vivo hepatic intrinsic clearance ranged from 0.081 mL/min/kg (warfarin; slowly metabolized) to 7.67 mL/min/kg

54

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

(diclofenac; rapidly metabolized). Linear time-dependent parent disappearance was observed, with the calculated hepatic clearance values falling within 2-3 fold of the known in vivo values. The results suggest that PHRA can be used for the estimation of in vivo hepatic clearance of slowly metabolized compounds that cannot be readily evaluated using routine methods. The assay can also be used for the metabolite profiling and phenotyping of slowly metabolized compounds.

P29. DEVELOPMENT OF A RELAY ASSAY WITH PLATED HUMAN HEPATOCYTES FOR THE EVALUATION OF SLOWLY METABOLIZED COMPOUNDS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Utkarsh Doshi and Albert P. Li In vitro ADMET Laboratories LLC, Columbia, MD, USA Routine screening of new chemical entities (NCE) for metabolic stability using either human liver microsomes or human hepatocytes in drug development has led to the collection of compounds that are resistant to metabolism by conventional in vitro procedures. These compounds nevertheless are metabolized in vivo, albeit slowly (slowly-metabolized compounds). These slowly metabolized compounds in general cannot be readily evaluated using the routine methods such as incubations with liver microsomes and hepatocytes. To overcome this challenge, a relay assay has recently been established by Di et al. (2012) with which an NCE is subjected to metabolism by up to 5 incubations (relays) using human hepatocyte suspensions, each for 4 h using freshly thawed human hepatocyte, thereby leading to a total incubation duration of 20 h (5 relays of 4 h). We report here a relay assay based on the same concept, but using plated human hepatocytes, with incubation durations of 24 h for each relay, thereby extending the time duration of each relay 4-fold of the assay of Di et al. Proof of concept studies were performed with a known slowly metabolized drug, tolbutamide. Three human hepatocyte lots were used in the assay as individual lots as well as a mixture of the three lots (pooled hepatocytes). Three relays were performed: Plateable cryopreserved human hepatocytes were thawed and cultured on a collagen-coated 24 well plate (350 K cells/well; 0.5 mL; in serum free Hepatocyte Incubation Medium (HQM)) for 4 h, followed by addition of an equal volume of HQM containing 2 mM of tolbutamide (final concentration of 1 mM). The plate was incubated at 37  C for 24 h after which an aliquot was taken for analysis. The remaining incubated medium was transferred to a newly prepared 24-well plates in which freshly thawed hepatocytes were cultured for 4 h (1st relay). The process was repeated one more time (2nd relay), resulting in incubation durations of 24, 48 and 72 h. Linear (R2 values ranged from 0.9837 to 0.9991) time-dependent decrease of tolbutamide was observed. Individual differences were observed, with the pool hepatocytes yielding results equivalent to the mathematical mean of the results of the three individual donors. Our results suggest that this modified relay method with plated hepatocytes can be used for the evaluation of the metabolic fate of slowly metabolized chemicals. Additional studies are being performed with the assay which includes the evaluation of additional model slowly metabolized chemicals, the use of pooled plateable human hepatocytes to minimize individual differences, and phenotyping studies using P450 inhibitors.

P30. MICROSOMAL STABILITY SCREENING USING SINGLE-TIME-POINT METHOD. A RETROSPECTIVE DATA ANALYSIS Pauline C.K. Chai1, Ali Tabatabaei2, Kathe Stauber2, and Anne Danks2 1 Department of Drug Metabolism and Pharmacokinetics, Preclinical Development, Dart NeuroScience LLC, San Diego, CA, USA and 2Dart NeuroScience LLC, San Diego, CA, USA To evaluate metabolic stability and potential first-pass extent, in vitro liver microsomal stability determination has become one of the first-tier screening assays routinely run in DMPK group during lead selection and lead optimization process. Intrinsic clearance generated from the study is used to rank order the compounds and predict the in vivo clearance and bioavailability. Traditionally, it is calculated using the slope determined from % remaining at multiple-time points vs. incubation time curve (multiple-time-point approach). However, with limited resources and expectation for faster turnaround time, more labs seek better experimental design using single-time-point approach instead (Li & Deng, 2014). For this approach, the slope is determined using the natural log value of the % remaining at one specific time point divided by the time (single-time-point method). To examine the accuracy and feasibility using single-time-point approach to determine the in vitro intrinsic clearance in early drug discovery, we conducted a retrospective data analysis from in-house data for more than 7000 compounds synthesized in DNS with different chemical structures run through the HLM and RLM stability assays. The assays were designed and run by multiple-time -point methods (0 plus 15, 30 and 60 min for HLM stability assay; or 0 plus 5, 15 and 30 min for RLM stability assay). The intrinsic clearance values calculated by two methods were compared. While CLint ¼ (slope)*(mL/mg protein); for multiple-time-point method, slope was obtained directly from the % reaming vs. time curve; for single-time-point method, slope was calculated from ln (%remaining at 60 min)/60 for HLM or ln (%remaining at 30 min)/30 for RLM. The analysis showed the intrinsic clearance calculated by the single-time-point method is in great agreement to that from multiple-time-point method; the correlation coefficient is greater than 0.95 for both HLM and RLM stability assays. The results suggested that the intrinsic clearance calculated from single-time-point method is comparable to that calculated by multiple-time-point method, and therefore can rank order the stability of the compounds in a right direction.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

55

Reference Li C, Deng Y. (2004). A novel method for the preparation of liposomes: Freeze drying of monophase solutions. J Pharm Sci 93:403–1414.

P31. THE ROLE OF EXTRA-HEPATIC METABOLISM IN THE PHARMACOKINETICS OF TARGETED COVALENT INHIBITORS AFATINIB, IBRUTINIB, NERATINIB

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yoshihiro Shibata and Masato Chiba Pharmacokinetics Research Laboratories, Taiho Pharmaceutical Co. Ltd., Tsukuba, Japan Despite the fact that much progress has been recently made in the development of targeted covalent inhibitor (TCI) candidates, the pharmacokinetics (PK) has not been well-characterized with an emphasis on the extra-hepatic clearance (CLextH) by GSH/GST-dependent conjugations attributable to the unique electrophilic structure (e.g. acrylamide moiety) incorporated into the TCI compounds. In the present study, the CLextH values were examined in rat, dog and monkey, in order to predict the contribution of CLextH to the PK in human for the model TCI compounds afatinib, ibrutinib and neratinib. Afatinib and neratinib both underwent extensive conjugations with GSH in the buffer and cytosol fractions of liver and kidney, while ibrutinib showed much lower reactivity/susceptibility toward GSH/GST-dependent conjugations. The CLextH in each species was calculated from the difference between observed total body clearance and predicted hepatic clearance (CLH) from metabolic stability in the cryopreserved hepatocytes suspended in 100% serum of the corresponding species (Shibata et al., 2000, 2002, 2014). The power-based simple allometry relating CLextH for the unbound compound to animal body weight was applicable across species for afatinib and neratinib (R240.9), while it was not for ibrutinib (R2 ¼ 0.04). The predicted AUC after oral administrations reasonably well agreed with reported AUC in the Phase I dose-escalation studies for afatinib and neratinib. Comparisons of CLextH and CLH predicted that the CLextH largely determines the PK of afatinib (490% as a % of total body clearance) and neratinib (34%) in human. The method can provide preliminary information on the potential contribution of CLextH by GSH/GST-dependent conjugations to total body clearance in the PK of TCI candidates in human, which also facilitates the optimization of PK at the drug discovery stage for the development of TCI candidates.

References Shibata Y, Takahashi H, Ishii Y. (2000). A convenient in vitro screening method for predicting in vivo drug metabolic clearance using isolated hepatocytes suspended in serum. Drug Metab Dispos 28:1518–1523. Shibata Y, Takahashi H, Chiba M, Ishii Y. (2002). Prediction of hepatic clearance and availability by cryopreserved human hepatocytes: an application of serum incubation method. Drug Metab Dispos 30:892–896. Shibata Y, Kuze J, Chiba M. (2014). Utility of cryopreserved hepatocytes suspended in serum to predict hepatic clearance in dog and monkey. Drug Metab Pharmacokinet 29:168–176.

P32. GETTING THE PASSIVE TUBULAR REABSORPTION RIGHT. PREREQUISITE FOR MECHANISTIC BOTTOM-UP MODELING OF RENAL CLEARANCE Daniel Scotcher1, Christopher Jones2, Amin Rostami1, and Aleksandra Galetin1 1 Centre for Applied Pharmacokinetic Research, Manchester Pharmacy School, University of Manchester, Manchester, UK and 2 Department of Oncology iMED, AstraZeneca, Macclesfield, UK Despite increasing application of in vitro–in vivo extrapolation (IVIVE) methods, mechanistic translation of in vitro data to predict human renal excretion clearance (CLR) in vivo is still lacking. In the current study a minimal physiologically-based mechanistic model was applied to predict fraction of tubular reabsorption (Freab) and non-secretion CLR (CLR,Non-sec) from in vitro permeability data generated in Caco-2 and MDCK cell lines. A database of CLR values for 63 drugs was collated from clinical studies. Passive permeability and transporter information were collated from literature and AstraZeneca in house databases. The mechanistic model was used to predict Freab and CLR, non-sec following changes in tubule surface area and filtrate flow rate in a stepwise manner. The drugs in the database represented a range of therapeutic classes and physical chemical properties; CLR values ranged from 0.02 to 119 mL/min for isoxicam and varenicline, respectively. Assuming that glomerular filtration (CLR,filt) was the only contributing mechanism to CLR resulted in an over-prediction of CLR for 21 drugs (43-fold error). In contrast, application of the minimal model of CLR, Non-Sec reduced the CLR over-prediction to 14 drugs. Accounting for water reabsorption and microvilli in the nephron, either alone or in combination, improved prediction accuracy of CLR, Non-sec such that CLR was over-predicted for 11, 5 and 2 drugs, respectively. Although successful for extensively reabsorbed drugs, use of CLR,Non-Sec approach under-predicted CLR for 32 drugs, of which 40% were reported as substrates for renal transporters. Of the remaining drugs, 13 were weak bases. Accounting for ionization of weak bases as a result of pH differences between urine and in vitro assays reduced the extent of CLR under-prediction. In conclusion, the minimal physiologically based model represents a promising approach to predict Freab and improve IVIVE of renal clearance. Further work is required to refine the model and incorporate contribution of active secretion.

56

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P33. STRATEGY FOR ACCURATE PREDICTION OF HUMAN CLEARANCE AND VOLUME OF DISTRIBUTION BY THE CONSENSUS LEVEL THEORY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Norimasa Jinno1, Michiko Niwa-Sakai1, Saori Koh1, Masashi Ishiguro1, Tomohito Kozaki2, Atsushi Ose3, Kiyohiko Sugano3, Mizuka Tagashira1, and Kazuyuki Tsurui1 1 Laboratory for Safety Assessment & ADME, Asahi Kasei Pharma Corporation, Izunokuni-shi, Shizuoka, Japan, 2Statistics Analysis Group, Data Science Department, Clinical Development Center, Asahi Kasei Pharma Corporation, Chiyoda-ku, Tokyo, Japan, 3Department of Clinical Pharmacology, Development Planning, Clinical Development Center, Asahi Kasei Pharma Corporation, Chiyoda-ku, Tokyo, Japan, and 4Laboratory for Formulation Research, Pharmaceuticals Research Center, Asahi Kasei Pharma Corporation, Izunokuni-shi, Shizuoka, Japan Purpose: Clearance (CL) is one of the most important parameters in human pharmacokinetics (PK) that relates to total drug exposure (area under the plasma concentration curve, AUC). Volume of distribution (Vd) is also an important parameter that relates to maximal concentration (Cmax) and half-life (T1/2). Although there are many methods to predict such PK parameters, it is difficult to decide which method is the best. The purpose of this study was to improve the predictions of CL and Vd at steady state (Vss) from preclinical data based on the consensus levels of several methods. Methods: Three human CL prediction methods were selected, hereinafter referred to as the methods providing the best predictions from the 29 methods previously reported (Ring et al., 2011; Jones et al., 2011; Lombardo et al., 2013a): (i) fu Correction Intercept Method (FCIM): CL ¼ 33.35  (a/Rfup)0.77 (ii) Two-species scaling approach (TS)rat–dog: CL (human) ¼ a (rat–dog)  (BW (human))0.628 (iii) Single-species scaling approach with monkey (SS)monkey: CL (human) ¼ 0.4  CL (monkey) In the same way, three human Vss prediction methods were selected from the 33 methods previously reported (Jones et al., 2011; Lombardo et al., 2013b): (i) Two-species-based allometry (rat, dog): log Vss (human) ¼ (0.07714 log Vss (rat)  log Vss (dog)) + 0.5147 log Vss (dog) + 0.586 (ii) Oie–Tozer_in vivo: Vss (human) ¼ Vp + (fup  Ve) + [(1 . fup)  RE/I  Vp] + Vr  fup/fut (human), fut (average preclinical species) ¼ fut (human), fut (animal) ¼ (Vr  fup)/Vss (animal). Vp  (fup  Ve). [(1  fup)  RE/I  Vp] (iii) SSS monkey: Vss (human) ¼ 0.79  Vss (monkey) Retrospective analyses were performed using 68 compounds, for which rat, dog, monkey, and human PK data were available (Lombardo et al., 2013a, b)). In addition, prediction accuracy was evaluated based on the closeness of each predicted value to actual values. Results: With regard to human CL prediction, when the maximum difference of two or three CL prediction methods was less than 2-fold, the discrepancy between the predicted and observed CL values was within 2-fold in about 60% of cases. Similarly, for human Vss prediction, when the maximum difference of three Vss prediction methods was less than 2-fold, the discrepancy between predicted and observed Vss values was within 2-fold in about 80% of cases. Based on these findings, the CONSENSUS LEVEL was defined based on the closeness of the predicted values. When the predicted values were within 2-fold, the predictabilities were considered ‘‘High’’ and the CONSENSUS LEVEL was set as ‘‘A’’; for predicted values not within 2-fold, the CONSENSUS level was set as ‘‘B’’. Conclusions: The confidence of predicting human PK was improved by considering the CONSENSUS LEVEL.

References Jones RD, Jones HM, Rowland M, et al. (2011). PhRMA CPCDC initiative on predictive models of human pharmacokinetics, part 2: Comparative assessment of prediction methods of human volume of distribution. J Pharm Sci 100:4074–4089. Lombardo F, Waters NJ, Argikar UA, et al. (2013a). Comprehensive assessment of human pharmacokinetic prediction based on in vivo animal pharmacokinetic data, part 2: Clearance. J Clin Pharmacol 53:178–191. Lombardo F, Waters NJ, Argikar UA, et al. (2013b). Comprehensive assessment of human pharmacokinetic prediction based on in vivo animal pharmacokinetic data, part 1: Volume of distribution at steady state. J Clin Pharmacol 53:167–177. Ring BJ, Chien JY, Adkison KK, et al. (2011). PhRMA CPCDC initiative on predictive models of human pharmacokinetics, part 3: Comparative assessment of prediction methods of human clearance. J Pharm Sci 100:4090–4110.

P34. IDENTIFICATION OF A RESIDUE POSSIBLY LOCATED IN THE SUBSTRATE BINDING SITE OF BOVINE GLYCINE N-ACYLTRANSFERASE Christoffel P. S. Badenhorst and Alberdina A. van Dijk Department of Biochemistry, North-West University, Potchefstroom, South Africa Glycine N-acyltransferase (GLYAT, E.C. 2.3.1.13) is a mitochondrial enzyme that plays a vital role in the metabolism of coenzyme A in the mammalian liver and kidney. GLYAT catalyses the glycine conjugation of benzoyl-CoA, an intermediate in the metabolism of phenylpropionate, a major product of microbial fermentation in the gut. In the absence of glycine conjugation by GLYAT, benzoyl-CoA would accumulate, leading to coenzyme A (CoASH) sequestration. CoASH is perhaps one of the most important molecules in biochemistry, and can be seen as a central hub connecting the metabolism of fatty acids, amino acids, and carbohydrates. Therefore, GLYAT is an essential enzyme that plays a very important role in metabolism by maintaining adequate levels of free CoASH in hepatic mitochondria. In modern times humans are constantly exposed to ever increasing numbers of

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

57

Downloaded by [112.134.238.102] at 07:30 10 February 2016

xenobiotics, such as industrial solvents and pharmaceutical drugs, that are metabolised to acyl-CoA intermediates. This presents a problem, since biotransformation systems do not always exist to facilitate the degradation of these unusual acyl-CoAs, which can lead to CoASH sequestration. Additionally, accumulating acyl-CoA thioesters can have several toxic effects, such as inhibiting important enzymes, serving as alternative substrates for normal metabolic enzymes, and forming adducts with cellular nucleophiles such as glutathione and certain amino acid residues of proteins. Accumulation of acyl-CoA metabolites, that cannot be further metabolised, is one of the major mechanisms of toxicity of several important pharmaceuticals, such as nonsteroidal antiinflammatory drugs and valproic acid, an anticonvulsant. It is therefore of great value to understand the factors which determine whether an acyl-CoA metabolite will accumulate or not. Since glycine conjugation is the major system preventing the accumulation of acyl-CoA metabolites in the liver, the aim of this study was to investigate the residues of GLYAT that form part of the substrate binding site, and therefore determine substrate selectivity. Because the structure of GLYAT has not been solved, a molecular model of bovine GLYAT was generated and used to identify Asn131 as a residue that appears to be situated in the enzyme’s binding site. A paralogue of bovine GLYAT, the bovine phenylacetyltransferase, has a cysteine residue, corresponding to Asn131 of bovine GLYAT, situated in its substrate-binding site. This active site cysteine residue makes the phenylacetyltransferase sensitive to inhibition by sulfhydryl reagents, such as DTNB. Therefore, site-directed mutagenesis was used to introduce a cysteine residue to this position of a recombinant bovine GLYAT, thus forming an N131C mutant. The N131C variant had enzyme activity comparable to that of the wild-type recombinant bovine GLYAT. However, treatment with DTNB significantly inhibited the N131C variant, similar to the inhibition of the phenylacetyltransferase by DTNB. These results suggest that the Asn131 residue of bovine GLYAT is situated in the enzyme’s substrate-binding site. Future work will investigate the effects of amino acid substitutions at this position on the substrate binding and substrate selectivity of recombinant human and bovine GLYAT enzymes.

P35. THE UNDERESTIMATED ROLE OF GLYCINE CONJUGATION IN METABOLIC HOMEOSTASIS Christoffel P. S. Badenhorst, Elardus Erasmus, Rencia van der Sluis, Carla Nortje, and Alberdina A. van Dijk Department of Biochemistry, North-West University, Potchefstroom, South Africa In 1841, Alexander Ure discovered the first biotransformation reaction, glycine conjugation, after finding that ingested benzoic acid is excreted as hippuric acid in the urine. And yet today, more than 170 years later, glycine conjugation is still a very poorly understood metabolic process that has been largely neglected by the scientific community. The lack of interest in glycine conjugation may be explained by the scarcity of pharmaceutically active xenobiotics that are metabolised to glycine conjugates. However, the fact that hippurate is one of the major components of animal urine suggests that glycine conjugation is a very important biotransformation process. This view is supported by the observation that no genetic defect of glycine conjugation has been reported in the literature. Because of the emerging clinical relevance of glycine conjugation, interest in this pathway has increased in recent years. Therefore, it is very important to understand the role of glycine N-acyltransferase (E.C. 2.3.1.13) and glycine conjugation under normal metabolic conditions. Two alternative hypotheses about the role of glycine conjugation have been proposed. The first is the traditional view that glycine conjugation is important for the detoxification and elimination of toxic organic acids such as benzoic acid. The more recent ‘‘glycine deportation’’ hypothesis claims that the detoxification value of glycine conjugation is low, and that glycine conjugation should instead be viewed as a neuroregulatory process, preventing accumulation of the neurotransmitter glycine in the central nervous system. This apparent conflict in the literature was addressed by re-evaluating the classic assumption that glycine conjugation is a detoxification mechanism, and by carefully analysing the glycine deportation hypothesis. One of the major observations underlying the glycine deportation hypothesis is that, unlike glucuronidation, glycine conjugation does not significantly improve the water solubility of organic acids such as benzoic acid. However, we argue that the detoxification value of glycine conjugation does not depend on improved water solubility of the glycine conjugates. The toxicity of benzoic acid results from its lipophilicity, and the detoxification value of glycine conjugation lies in the formation of the less lipophilic hippuric acid, which can be more readily excreted in the urine. We conclude that benzoic acid should not be seen as the typical substrate for glycine conjugation, which is more appropriately viewed as the final reaction of the phenylpropionate catabolism pathway. Phenylpropionate, a product of microbial fermentation in the gut, is metabolised to benzoylCoA in the liver, followed by glycine conjugation, which regenerates coenzyme A and forms the less toxic hippurate, which is excreted. Finally, we suggest that the depletion of plasma and CSF glycine levels, resulting from the ingestion of large amounts of phenylpropionate or benzoate, serves as the neurological signal for animals to cease feeding on diets rich in these toxins.

P36. WHOLE CELL-DEPENDENT PRODUCTION OF PHASE II METABOLITES USING GENETICALLY ENGINEERED BUDDING YEAST Shinichi Ikushiro, Miyu Nishikawa, and Toshiyuki Sakaki Toyama Prefectural University, Toyama, Japan Some drugs with carboxylic acid moieties are metabolized to acyl glucuronide and it has been thought to be associated with the toxicity in some case. Therefore, in vitro synthesis of phase II metabolites often becomes critical during toxicity studies of new pharmaceutical products. In order to synthesize the conjugates as phase II metabolites, we have developed several mammalian UDP-glucuronosyltransferases (UGT)sulfotransferase (SULT) ,or catechol-O-methyltransferase (COMT)expression systems in budding yeast. For the glucuronide production, mammalian UGT including human and model animals, and UDP-glucose

58

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

dehydrogenase (UGDH) were co-expressed in budding yeast, Saccharomyces cerevisiae AH22, using a genome integrated vector. For sulfate production, five human SULT isoforms (1A1, 1A3,1B1,1E1,2A1) or human COMT were expressed in yeast using a multicopy plasmid vector. Using genetically engineered yeast containing human UGT1A6 and rat UGDH, glucuronide formation of 7-hydroxycoumarine was examined. Most glucuronides were found in the reaction medium with time-dependent production, suggesting the functional expression of both enzymes and the presence of endogenous transport system for glucuronide in yeast. Mycophenolic acid used in the immunosuppretant drug has a phenolic and carboxylic hydroxyl groups. Human UGT1A9 in yeast cells can mainly catalyze the penolic glucuronide formation. In contrast, rat UGT2B1 can specifically catalyze the acyl glucuronide formation. Selecting the suitable mammalian UGT isoforms in yeast, we can get the desirable glucuronides of various compounds with multiple conjugation sites. In the presence of glucose and ammonium sulfate, formation of sulfated drug and dietary compounds were observed in whole-cell production system with SULT isoforms. In the presence of glucose and L-methionine, formation of methylated O-catechol compounds were observed in whole-cell production system with COMT. These expression system of mammalian UGT, SULT or COMT in budding yeast would be a powerful and convenient tool for enzyme-assisted synthesis of various phase II metabolites including glucuronides, sulfates and methylated conjugates.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P37. DRUG-DRUG INTERACTION POTENTIAL AND EFFECTS ON INTRINSIC CLEARANCE RESULTING FROM CONJUGATION BY BOTH N-METHYL AND SULFATRANSFERASES: A CASE STUDY Carlo Sensenhauser1, Ian E. Templeton1, Shannon Dallas1, Peter Ward2, and Jose Silva1 1 The Pharmacokinetics, Dynamics and Metabolism (PDM) Division, Janssen Research & Development, LLC, Spring House, PA, USA and 2GPCD Unit, Janssen Research & Development, LLC, La Jolla, CA, USA Metabolic clearance of xenobiotic compounds can be affected by the enzymatic conversion of the substrate to a more hydrophilic conjugated metabolite (e.g. glucuronides or sulfonates), resulting from the transfer of a polar group from a specific cofactor, such as uridine 5-diphospho glucuronic acid (UDPGA) or, in the case of sulfatransferase (SULT)-catalyzed reactions, 30 phosphoadenosine-50 -phosphosulfate (PAPS). On the other hand, methyltransferases are a large family of enzymes that catalyze the transfer of a methyl group from S-adenosylmethionine (SAM) to a carbon or heteroatom (O, N or S) on the substrate molecule, which can result in a more hydrophobic conjugate. The primary metabolic routes of a compound, JNJ-Y, currently in development, are sulfonation of a hydroxyl group and methylation of a pyridinyl nitrogen. Both monosubstituted products (either N-methyl or sulfate) and disubstituted product (N-methyl sulfate) are detected as circulating metabolites. Determination of the kinetic parameters, Km and Vmax, indicated an intrinsic clearance (CLint) value for N-methylation about one order of magnitude greater than sulfonation. In addition, CLint values for the conversion of the parent compound to either a monosubstituted metabolite were also approximately an order of magnitude higher than the corresponding pathway to convert a mono- to the disubstituted metabolite. The effects of selective inhibition of either or both pathways on CLint were studied in vitro in both human liver and intestinal cytosol. Cytosolic fractions (0.4 mg/mL) were incubated with JNJ-Y in the presence of PAPS, SAM, dithiothreitol and potential inhibitors of either pathway (e.g., nicotinamide, homocysteine, 17-alpha ethynylestradiol). JNJ-Y was also screened with a panel of recombinant human SULTs to identify the isoforms involved in the formation of the sulfonated metabolites. Overall, metabolite formation either as a function of time, substrate or inhibitor concentration, was analyzed with an LC/MS/MS method on a Sciex API4000 triple quadrupole mass spectrometer. In addition, the effects of JNJ-Y as a perpetrator on the metabolism of nicotinamide and prototypical SULT substrates (e.g. acetaminophen, dehydroepiandrosterone) were investigated. Nicotinamide the N-methylation of JNJ-Y and 17-alpha ethynylestradiol inhibited the sulfonation pathway. IC50 values were 40 and 1 microM, respectively. Multiple SULT isoforms appear to be involved in the formation of the sulfate metabolites, with SULT2A1 and possibly SULT1E1 providing the major contributions. Further, pharmacokinetic modeling was utilized to estimate the in vivo impact of the collected in vitro data.

P38. IN VITRO EFFECT OF ANTHOCYANINS AND CATECHINS ON ACTIVITY OF SELECTED BIOTRANSFORMATION ENZYMES IN INTESTINAL CELL LINES Katerina Lnenickova, Petra Pokorna, and Barbora Szotakova Department of Biochemical Sciences, Faculty of Pharmacy, Charles University in Prague, Hradec Kralove, Czech Republic Anthocyanins and catechins are two important groups of natural polyphenols. Their ability to scavenge reactive oxygen species makes them valued compounds with a wide range of use in the health care. Polyphenols protects cells from damage caused by free radicals, and thus help to reduce the risks of cardiovascular disease and cancer. The consumption of dietary supplements with high concentrations of flavonoids is still increasing, but in addition to positive effects on human health it can also have side effects, especially influence drug metabolizing enzymes and thus may affect pharmacokinetic as well as pharmacodynamic profiles of co-administered drugs. The potential cytotoxicity of selected anthocyanidins (cyanidin, delphinidin, malvidin, pelargonidin) and green tea catechins (epigallocatechin-3-gallate (EGCG) and polyphenon 60) towards Caco-2 cells were evaluated by MTT assay. The aim of this study was to evaluate the effect of anthocyanins and catechins on the activity of selected biotransformation enzymes: glutathione S-transferase (GST), carbonyl reductase (CBR), UDP-glucuronosyl transferase (UGT) and cytochrome P450 (CYP). Results showed that tested anthocyanidins except malvidin have no effect on the proliferation of intestinal cell line Caco-2. Malvidin showed a slight cytotoxic effect in a concentration-dependent manner.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

59

Furthermore, anthocyanidins inhibited the activity of GST. Pelargonidin was the most potent inhibitor of GST. In contrast, catechins showed no effect on the activity of GST. Delphinidin and pelargonidin inhibited the activity of CBR. Conversely, malvidin, cyanidin and both catechins increased the activity of CBR. The activity of CYP (CYP1A1, CYP1A2 and CYP3A4) and UGT in control and treated microsomes of Caco-2 cells has not been detected. Observed changes in the activity of biotransformation enzymes induced by selected flavonoids were mild, therefore significant interactions with concomitantly administered drugs are not expected. This project was supported by Czech Science Foundation, Grant No. P303/12/G163 and by the Grant Agency of Charles University, Grant No. 2014/1874214.

P39. UDP-GLUCURONOSYLTRANSFERASE1A3, 1A9 AND 2B7 ARE MAJOR ENZYMES RESPONSIBLE FOR 20-HETE GLUCURONIDATION IN HUMAN LIVER MICROSOMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yazun B. Jarrar1, Young-Cha Eun1, Ah-Seo Kyung1, Hyun-Kim Dong1, Su-Jun Lee1, and Jae-Gook Shin2 1 Department of Pharmacology and Pharmacogenomics Research Center, Inje University College of Medicine, Busan, South Korea and 2Department of Clinical Pharmacology, Inje University College of Medicine, Busan, South Korea The 20-hydroxyeicosatetraenoic acid (20-HETE) is involved in manyphysiological functions, including blood pressure and platelet aggregation. The 20-HETE is glucuronidation by UDP-glucuronosyltransferases (UGTs) and it is eliminated through human urine mainly in the glucuronidated form.The present study identified major UGTs responsible for 20-HETE glucuronidation and investigated their genetic influenceon the glucuronidation reaction using human livers (n ¼ 44). Twelve recombinant UGTs were screened to identify themajor contributors to 20-HETE glucuronidation. The results showed thatUGT2B7, UGT1A9, and UGT1A3 were themajor contributors to 20-HETE glucuronidation. Protein expression levels and genetic variants of UGT1A3, 1A9 and 2B7 were analyzed in human livers using Western blotting and genotyping, respectively. Glucuronidationof 20-HETE was significantly correlated with the protein levels of UGT2B7 (r2 ¼ 0.33, p50.001) and UGT1A9 (r2 ¼ 0.31, p50.001), but not UGT1A3 (r2 ¼ 0.02, p40.05). The genetic variations UGT2B7802C4T, UGT1A9118T94T10,and 1399T4C significantly altered 20-HETE-glucuronide formation (p50.05–0.001). As altered levels of 20HETE correlated with cardiovascular homeostasis and diseases, the present data may increase our understanding of 20-HETE metabolism and cardiovascular complications.

P40. ASSESSMENT OF THE IMPACT OF GLUTATHIONE-S-TRANSFERASE (GST) M1 GENETIC POLYMORPHISMS ON THE CLEARANCE AND FRACTIONAL METABOLISM OF COVALENT MODIFIERS Timothy J. Strelevitz1, Xin Yang1, Linda Wood1, Jean-Claude Marshall1, Mike Zientek2, Larry Tremaine1, and Louis Leung1 1 Department of Pharmacokinetics, Pharmacodynamics and Metabolism, Pfizer, Groton, CT, USA and 2Department of Pharmacokinetics, Pharmacodynamics and Metabolism, Pfizer, San Diego, CA, USA Assessment of the Impact of Glutathione-S-Transferase (GST) M1 Genetic Polymorphisms on the Clearance and Fractional Metabolism of Covalent Modifiers Tim Strelevitz, Xin Yang, Linda Wood, Jean-Claude Marshall, Larry Tremaine, Mike Zientek, Louis Leung. The design of targeted covalent enzyme modifiers may seem attractive considering the potential for enhanced biochemical efficiency associated with an irreversible or a slow-reversible mechanism, which can lead to an increase in selectivity and therapeutic margins. Due to the presence of the inherently reactive chemical moiety, clearance of covalent modifiers may involve direct reaction with glutathione (GSH) or via a glutathione S-transferase (GST)-mediated process. To assess the role of GST enzymes in the hepatic metabolic clearance of covalent modifiers in human, we utilized in vitro approaches including recombinant human GST enzymes and liver S9 fraction genotyped for GSTM1, the major GST enzyme in the liver known to exhibit genetic polymorphisms. A representative acrylamide covalent modifier exhibited GSTM1 genotypedependent GSH-conjugation activity and fractional metabolism among the different liver S9 fractions, with corresponding changes in the fractional metabolism by cytochrome P450 3A and no noticeable changes in fractional metabolism by UDPglucuronosyltransferase 1A9, enzymes that are also involved in its metabolism. The in vitro results suggested that GSTM1 genotypes may need to be taken into consideration during the clinical drug development of certain covalent modifiers.

P41. FORMATION OF NOVEL GLUTATHIONE ADDUCTS OF BENZBROMARONE METABOLITES BY HUMAN LIVER MICROSOMES AND CYTOSOLS Naoki Cho, Kaoru Kobayashi, Mina Yoshida, Minaka Shibuya, and Kan Chiba Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba City, Japan Benzbromarone (BBR) is a uricosuric agent that occasionally causes fatal hepatic failure in human beings. Recently, we found that 10 -OH BBR, which is a BBR metabolite, is primarily metabolized to 10 ,6-(OH)2 BBR by cytochrome P450 2C9 (CYP2C9), which is mainly expressed in the human liver. We also found that 10 -OH BBR shows cytotoxic effects on human hepatoma FLC7 cells stably expressing CYP2C9 when in the presence of buthionine-sulfoximine, which is an inhibitor of glutathione (GSH) synthesis. These findings suggest that 10 ,6-(OH)2 BBR, and/or other metabolites generated from 10 ,6-(OH)2 BBR by CYP2C9,

60

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

are involved in 10 -OH BBR induced cytotoxicity. Generally speaking, the drug metabolites that cause hepatotoxicity are detoxified by the formation of GSH adducts. Therefore, the formation of GSH adducts from 10 ,6-(OH)2 BBR and/or other metabolites appears to provide an important pathway to detoxification. However, there is currently no information available on GSH adducts formed from 10 ,6-(OH)2 BBR or other metabolites generated from 10 ,6-(OH)2 BBR in the human liver. Accordingly, this study aims to clarify whether GSH adducts are formed from 10 ,6-(OH)2 BBR in human liver microsomes or cytosols. To obtain 10 ,6-(OH)2 BBR, large-scale conversion of 10 -OH BBR to 10 ,6-(OH)2 BBR was conducted by using rat liver microsomes. Formed 10 ,6-(OH)2 BBR was isolated by preparative high performance liquid chromatography and metabolites were analyzed by amaZon SL ion trap mass spectrometer operating in the electrospray negative ionization mode. In the presence of GSH, incubation of 10 ,6-(OH)2 BBR with human liver microsomes or cytosols resulted in the formation of three metabolites (M1, M2, and M3) having clustered (dibromo-containing) peaks centered at m/z 716, m/z 744, and m/z 778, respectively. M1 and M2 expelled fragment ions at m/z 443 and m/z 471, respectively, which were derived from cleavage of the bond linking the sulfur to the carbon atom of the the tripeptide chain. The M1, M2, and M3 mass numbers were identical to those of a GSH adduct generated from 10 ,6-(OH)2 BBR lacking C2H4O, that was derived from the BBR monohydroxide, and that from the further hydroxide of 10 ,6-(OH)2 BBR, respectively. We found that M1 and M3 were formed in human liver microsomes, while M2 was formed in human liver cytosols. These results indicate that GSH adducts are formed from 10 ,6-(OH)2 BBR by microsomal and cytosolic enzymes, which then provides a detoxification process for BBR-induced hepatotoxicity.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P42. ENZYMATIC ACTIVITIES OF NOVEL NAT1 AND NAT2 POLYMORPHIC VARIANTS FOUND IN A POPULATION OF MODEL PRIMATE MACACA MULATTA (RHESUS MACAQUE) Sotiria Boukouvala1, Theodora Tsirka1, Nafsika Drakomathioulaki1, Despina Giannouri1, Nikolaos Marinakis1, Sofia Rizou1, Sofia Zaliou1, Audrey Sabbagh2, Brigitte Crouau-Roy3, and Giannoulis Fakis1 1 Department of Molecular Biology and Genetics, Democritus University of Thrace, Alexandroupolis, Greece, 2Institut de Recherche pour le De´veloppement (IRD), Universite´ Paris Descartes, Paris, France, and 3ENFA, UMR5174EDB (Laboratoire E´volution & Diversite´ Biologique), CNRS, Universite´ Paul Sabatier, Toulouse, France Arylamine N-acetyltransferases (NATs) are xenobiotic metabolizing enzymes studied for their roles in drug metabolism and chemical carcinogenesis. NAT homologues are found in human and other primates, including rhesus macaque (Macaca mulatta). taxon mnemonic MACMU), a well-established model in medical and pharmacological research. Most primates have two NAT isoenzymes (NAT1 and NAT2), both of which are genetically and functionally polymorphic in human populations. In rhesus, one polymorphism (p.Val231Ile) has been described for NAT2 and is unique in that it alters function by decreasing enzyme affinity towards NAT2 substrates and increasing affinity towards NAT1 substrates. In the present study, we sequenced NAT1 and NAT2 from 25 rhesus individuals, identifying 12 haplotypes for each of the two genes. The (MACMU)NAT1 haplotypes bear 5 synonymous and 7 non-synonymous SNPs, while the (MACMU)NAT2 haplotypes contain 6 synonymous and 7 non-synonymous SNPs, including known SNP c.G691A (p.Val231Ile). One (MACMU)NAT1 non-synonymous SNP, c.G560A (p.Arg187Gln), is also found in human populations and is responsible for the slow acetylator phenotype conferred by human NAT1*14 alleles. Sitedirected mutagenesis was performed to introduce the non-synonymous NAT1 and NAT2 SNPs of our rhesus population into the reference sequence of each corresponding gene, followed by expression and purification of recombinant variants for functional studies. The (MACMU)NAT1 variants p.Gly51Ala, p.Met82Val, p.Leu89Phe, p.Asp115Tyr, p.Glu155Gln, p.Phe175Leu and p.Arg187Gln were assayed with substrates p-aminobenzoate (PABA), 5-aminosalicylate (5AS), sulphamethazine (SMZ) and p-anisidine (PANS). The (MACMU)NAT2 variants p.Met42Val, p.Glu49Lys, p.Asp142Tyr, p.Gln155Glu, p.His186Asn, p.Ser221Stop and p.Val231Ile were investigated with substrates procainamide (PA), SMZ, PABA and PANS. Of the substrates used, PABA and 5AS are NAT1-selective, PA and SMZ are NAT2 selective and PANS generates activity with both isoenzymes. For (MACMU)NAT1 variants, our results indicate that p.Glu155Gln, p.Gly51Ala and p.Met82Val cause marginal to moderate decrease in enzymatic activity, while p.Leu89Phe, p.Asp115Tyr, p.Phe175Leu and p.Arg187Gln reduce acetylation activity. Immunoblot analysis with antibody against the N-terminal hexa-histidine tag of recombinant proteins indicated that most (MACMU)NAT1 variants demonstrate compromised integrity. In contrast, the same antibody detected intact recombinant (MACMU)NAT2 variants, although effects on protein thermostability were evident by differential scanning fluorimetry and matched the diverse enzymatic activity pattern observed with different substrates. Variants p.Glu49Lys and p.Gln155Glu caused marginal to moderate decrease in (MACMU)NAT2 enzymatic activity, while p.Met42Val was a slow acetylator variant. In contrast, p.His186Asn caused marginal to moderate increase in enzymatic activity, while p.Asp142Tyr had an unusual effect in that it increased (MACMU)NAT2 enzymatic activity substantially. The effect of p.Val231Ile polymorphism was confirmed to be as previously described, while p.Ser221Stop was a truncated inactive variant. Our experimental observations for rhesus NAT1 and NAT2 variants were supported by structural modelling, allowing insight into novel mechanisms by which polymorphisms may affect NAT enzymatic function. Rhesus NAT proteins are expressed more efficiently than their human homologues, providing a practicable model for chemical screens.

P43. EFFECT OF CARBAMAZEPINE ON EXPRESSION OF UDP-GLUCURONOSYLTRANSFERASE 1A6 IN RAT BRAIN Yuki Asai, Miki Katoh, Yukiko Sakakibara, and Masayuki Nadai Faculty of Pharmacy, Meijo University, Nagoya, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

61

UDP-glucuronosyltransferase (UGT) is one of the major phase II drug-metabolizing enzymes, which catalyzes glucuronidation of many drugs and endogenous compounds. While UGT is mainly expressed in the liver, it is also found in the brain. Therefore, UGT may be responsible for metabolizing drugs, such as anti-depressants and anti-epileptics in the brain. There are several drugs and chemicals capable of inducing UGTs, including UGT1A6. Carbamazepine (CBZ) is an anti-epileptic drug and a typical inducer of many drug-metabolizing enzymes. There is, however, no evidence that UGT is induced by CBZ in the brain. In the present study, we determined whether Ugt1a6 is induced by treatment with CBZ in rat brains. Eight-week-old male Sprague-Dawly rats were treated intraperitoneally for 7 days with 100 mg/kg CBZ. The rats were decapitated one day after the last treatment and the brains were divided into nine regions (cerebellum, frontal cortex, parietal cortex, piriform cortex, hippocampus, medulla oblongata, olfactory bulb, striatum and thalamus). Expression of Ugt1a6 mRNA was quantified by real-time polymerase chain reaction. Serotonin glucuronidation, which is catalyzed mainly by Ugt1a6 in the brain, was measured by liquid chromatography-tandem mass spectrometry. The Ugt1a6 mRNA was increased by 3.4- and 1.9-fold in the cerebellum and hippocampus, respectively. For serotonin glucuronidation, there was a 1.8-fold increase in the hippocampus. It is suggested that the increase in serotonin glucuronidation resulted from the induction of Ugt1a6 mRNA in the hippocampus. Next, a further experiment was conducted to clarify the mechanism of Ugt1a6 induction in rat brain. NF-E2 related factor 2 (Nrf2), which is one of the nuclear receptors, is the key factor of Ugt1a6 induction and is activated by reactive oxygen species (ROS). It has been reported that ROS is generated by CBZ b (Pavone & Cardile, 2003). In the hippocampus, the Nrf2 mRNA was increased by 2.3-fold and nuclear translocation of Nrf2 protein was elevated by 1.3-fold after the treatment with CBZ. Thus, there is a possibility that Nrf2 can be activated by ROS that is formed by CBZ, leading to an increase in Ugt1a6 mRNA expression in the hippocampus. On the other hand, it is predicted that another mechanism may be responsible for Ugt1a6 induction in the cerebellum since Nrf2 mRNA and nuclear translocation of Nrf2 were not changed. In conclusion, Ugt1a6 mRNA and serotonin glucuronidation were induced by the treatment with CBZ, suggesting that the drug concentration is reduced in the brain by CBZ.

Reference Pavone A, Cardile V. (2003). An in vitro study of new antiepileptic drugs and astrocytes. Epilepsia 10:34–39.

P44. IDENTIFICATION OF RESPONSIBLE ORGAN AND UDP-GLUCURONOSYLTRANSFERASE (UGT) ISOFORMS FOR THE GLUCURONIDATION OF IPRAGLIFLOZIN Yasuyuki Kasai1, Yasuharu Nakanishi2, Fumihiko Ushigome1, Hiroshi Arai1, and Aiji Miyashita1 1 Analysis & Pharmacokinetics Research Laboratories, Astellas Pharma Inc., Osaka, Japan and 2Pharmacokinetics and Bioanalysis Center, Shin Nippon Biomedical Laboratories, Ltd., Wakayama, Japan Background: Ipragliflozin, a selective sodium glucose co-transporter 2 inhibitor, is a therapeutic agent for type 2 diabetes mellitus (T2DM). In T2DM patients with moderate or severe renal impairment, the ipragliflozin exposure was 1.4–1.5-fold greater than those in T2DM patients with normal renal function (Veltkamp et al., 2011). Therefore, it is meaningful to investigate the responsible organ of ipragliflozin metabolism for clarifying the mechanism of increase in ipragliflozin exposure in these patients. Metabolic profiling data indicated that glucuronidation is the primary elimination pathway in human, yielding the three regioisomeric metabolites, M2, M3 and M4. It has been reported that UDP-glucuronosyltransferase (UGT) 2B7, UGT2B4 and UGT1A9 are involved in M2 formation in human liver (Ushigome et al., 2011). Objective: The purpose of this study was to identify the responsible organ for the ipragliflozin glucuronidation, and the responsible UGT isoforms for the metabolites formation in human liver and kidney. Method: The metabolites formation activities were compared using human liver microsomes (HLM), human intestine microsomes (HIM), and human kidney microsomes (HKM). Furthermore, the UGT isoforms responsible for the metabolites formation in human liver and kidney were identified using chemical UGT inhibitors and recombinant human UGTs. Results: HKM showed the highest M2 and M4 formation followed by HLM and HIM, while HLM showed the highest M3 formation followed by HKM. These results indicate that HKM have comparable or higher ipragliflozin catalytic activity than HLM, while HIM has markedly low activity. Among 12 recombinant human UGTs tested, high-catalytic activity was exhibited by UGT2B7 for M2 formation, UGT1A4, UGT1A9 and UGT2B4 for M3 formation, and UGT1A9 for M4 formation. Prior to identification of UGT isoforms using HLM and HKM, the selectivity of inhibitory effect of hecogenin, niflumic acid, gemfibrozil and diclofenac was evaluated using recombinant human UGTs. Hecogenin, niflumic acid, and gemfibrozil selectively inhibited UGT1A4, UGT1A9, and UGT2B7, respectively, while diclofenac inhibited multiple UGTs including UGT2B4. In HLM, gemfibrozil, diclofenac and niflumic acid strongly inhibited M2, M3, and M4 formation, respectively, whereas hecogenin and niflumic acid did not inhibit M3 formation. In HKM, gemfibrozil strongly inhibited M2 formation, and niflumic acid strongly inhibited M3 and M4 formation. Conclusion: HKM and HLM showed high ipragliflozin catalytic activity. Taking tissue weight and microsomal protein content into consideration, it was suggested that in vivoipragliflozin catalytic activity would be the highest in the liver, followed by kidney. Furthermore, it was suggested that M2 formation is catalyzed by UGT2B7 in human liver and kidney, M3 formation is catalyzed by UGT2B4 in human liver and UGT1A9 in human kidney, and M4 formation is catalyzed by UGT1A9 in human liver and kidney.

62

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

References Ushigome F, Kasai Y, Uehara S, et al. (2011). Identification of UDP-glucuronosyltransferase (UGT) isozymes involved in ipragliflozin metabolism in human liver. Poster W4421 presented at 25th American Association of Pharmaceutical Scientists Annual Meeting and Exposition. Veltkamp SA, et al. (2011). The effect of renal impairment on the pharmacokinetics and urinary glucose excretion of the SGLT2 inhibitor ipragliflozin (ASP1941) in type 2 diabetes mellitus patients. Poster 1127-P presented at the 71st Scientific Sessions of the American Diabetes Association.

P45. CORRELATION BETWEEN PROTEIN LEVELS AND GLUCURONIDATION ACTIVITIES OF UGT MARKER SUBSTRATES AND CORRELATION BETWEEN UGT PROTEIN LEVELS IN HUMAN LIVER MICSOROMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yuichiro Sato and Masanori Nagata Analysis & Pharmacokinetics Research Laboratories, Astellas Pharma Inc., Osaka-shi, Japan Uridine 50 -diphosphate-glucuronosyltransferases (UGTs) are microsomal enzymes that catalyze the glucuronidation of various xenobiotics as well as endogenous compounds. In humans, at least 20 UGTs have been identified and those isoforms are known to possess broad and often overlapping substrate specificities, which complicates the analysis of individual isoform function. Although several marker substrates for this purpose have been identified, their utility and selectivity have not been fully investigated. One way to determine their validity as marker substrate is to investigate the proportionality between glucuronidation activities and protein levels of the corresponding UGT isoform. We previously reported optimized targeted peptide-based methods for quantification of 13 UGT isoforms (UGT1A1, 1A3, 1A4, 1A6, 1A7, 1A8, 1A9, 1A10, 2B4, 2B7, 2B10, 2B15, and 2B17) using liquid chromatography-tandem mass spectrometry and expression of 10 isoforms (UGT1A1, 1A3, 1A4, 1A6, 1A9, 2B4, 2B7, 2B10, 2B15, and 2B17) in human liver microsomes (Sato et al., Drug Metabolism and Disposition 42:885–889, 2014). Here, we investigated the correlation between protein levels and glucuronidation activities of potential marker substrates (estradiol: 1A1, trifluoperazine: 1A4, naphthol: 1A6, propofol: 1A9, and morphine: 2B7) using 16 individual human liver microsomes. Good correlation between protein level and glucuronidation activity was observed for UGT1A1 (r2 ¼ 0.9256 and p50.0001), 1A4 (r2 ¼ 0.6519 and p ¼ 0.0002) and 2B7 (r2 ¼ 0.6408 and p ¼ 0.0002), indicating the utility of estradiol, trifluoperazine, and morphine as respective marker substrates. Correlation between UGT1A6 protein level and naphthol glucuronidation was poor (r2 ¼ 0.1500 and p ¼ 0.1384), which might be in line with the previous studies indicating remarkable contribution of other isoforms to the glucuronidation of naphthol. Protein level of UGT1A9 was significantly correlated with propofol glucuronidation after removing one apparent outlier whose glucuronidation was very low compared with the other samples (r2 ¼ 0.6531 and p ¼ 0.0004). Correlation between UGT protein levels was also investigated. While no significant correlation was observed between protein levels of UGT1A isoforms, significant correlation was observed between UGT2B4, 2B7, 2B10 and 2B15. These results suggest that transcription of UGT2B4, 2B7, 2B10 and 2B15 might be regulated similarly. Protein level of UGT2B17 was undetectable or minimally detectable in five individuals, and no significant correlation with any UGT2B isoforms was observed.

P46. MECHANISM FOR THE RETENTION OF UDP-GLUCURONOSYLTRANSFERASE 2B7 IN THE ENDOPLASMIC RETICULUM: THE ROLE OF THE LUMINAL DOMAIN Sora Kimura1, Yuu Miyauchi1, Yuko Hirota2, Yoshitaka Tanaka2, Peter I. Mackenzie3, Hideyuki Yamada1, and Yuji Ishii1 1 Laboratory of Molecular Life Sciences, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan, 2 Department of Pharmaceutical Cell Biology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan, and 3Department of Clinical Pharmacology, Flinders Medical Center and Flinders University, Adelaide, Australia Objectives: UDP-Glucuronosyltransferase (UGT) is a type-I membrane protein, which is bound to the endoplasmic reticulum (ER) membrane via the C-terminal region; the bulk of the protein is located in the luminal space of the ER. The double lysine motif (DLM: KKXX or KXKXX) of UGT has been considered to serve as an ER-retention signal. However, this motif seems not to be the sole determinant for correct targeting, because its deletion fails to affect the localization of human UGT1A6 to the ER. To confirm if this is also true for UGT2B7, we examined the role of the cytosolic tail containing DLM and the luminal domain in the localization of this UGT to the ER. Methods: A baculovirus-Sf-9 cell system for expressing UGT2B7 and its mutants lacking the C-terminus was established using a Bac-to-Bac system. The cells were subfractionated by differential centrifugation, and then the subcellular localization of recombinant UGT2B7 and its deletion mutants was analyzed by immunoblotting. Further, the same constructs in pcDNA3.1 were also expressed in COS cells, and their subcellular localization was analyzed by immunofluorescent method using a confocal microscopy. Results and discussion: Like wild-type UGT2B7, mutants lacking DLM and the whole cytosolic tail were detected in the microsomal fraction but not in the nucleus, mitochondria and cytosol. Such subcellular localization was similar to calnexin, an ER marker protein. Furthermore, truncated UGT2B7 mutants lacking some or all of residues 134-253 were expressed in Sf9 microsomes and were co-localized with calnexin in COS cells. Conclusions: The DLM present in the C-terminus of UGT2B7 is not essential for localizing this enzyme to the ER membrane. It is suggested that the luminal domain of UGT2B7 plays a role in retention in the ER. Although the present study could not identify the specific domain causing the targeting of UGT2B7 to the ER, one key region seems to be located prior to residue 133.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

63

P47. PROTEIN EXPRESSION LEVELS OF DRUG-METABOLIZING ENZYMES AND TRANSPORTERS IN HUMAN SMALL INTESTINE, LIVER AND KIDNEY MICROSOMES BY SWATH-MS BASED COMPREHENSIVE PROTEIN QUANTIFICATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kenji Nakamura1, Mio Hirayama2, Shingo Ito2, and Sumio Ohtsuki2 1 Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan and 2Department of Pharmaceutical Microbiology, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan Protein expression levels of metabolizing enzymes and transporters in liver, intestine and kidney are important for understanding ADME of drugs. Protein quantification by multiple reaction monitoring (MRM) with LC-MS/MS is becoming a useful tool for ADME study; however, it has the limitation in number of measurable proteins in a single injection. SWATH (sequential window acquisition of all theoretical fragment ion spectra) is a new acquisition method of LC-MS/MS and solves this limitation in principle. The purpose of the present study was to validate simultaneous protein quantification by SWATH and determine the protein expression levels of CYPs, UGTs and transporters in human small intestine, liver and kidney microsomes. The absolute protein expression levels of 20 CYPs and 9 UGTs, 31 ABCs and 73 SLCs were measured in commercially available human liver microsomes (HLM, pool of 50), human intestinal microsomes (HIM, pool of 13) and human kidney microsomes (HKM, pool of 4) by LC-MS/MS with spiking stable isotope labeled internal standard peptides. Quantified values of 65 peptides were compared between SWATH and MRM, and 85% of the values were less than 1.5 fold difference. There was no significance in the variance of quantified values between the two methods (p ¼ 0.55, Student’s t-test). The absolute protein expression levels of 20 CYPs and 9 UGTs, 31 ABCs and 73 SLCs were measured in the microsomes by SWATH. The number of quantified proteins were 15, 8, and 3 of the CYPs, 9, 4, and 3 of the UGTs, 7, 9 and 7 of the ABCs, and 12, 25 and 38 of the SLCs in microsomes of the liver, small intestine or kidney, respectively. Expression of CYPs in liver was greater than those in the intestine and kidney, e.g. the expression level of CYP3A4 in the liver was 4.1 times greater than the small intestine. In contrast, some UGT isoforms exhibited greater expression in the intestine or kidney than in the liver, e.g. UGT1A9 in the kidney was 3.1 times greater than in the liver. These results indicate that SWATH provides comprehensive protein quantification, which is as accurate as that by MRM. In addition, absolute protein expression levels of CYPs, UGTs, ABCs and SLCs reported in the present study are important information for understanding contribution of each protein to drug ADME.

P48. DIFFERENTIAL EXPRESSION OF SMALL HETERODIMER PARTNER POTENTIALLY CONTRIBUTES TO INTERINDIVIDUAL VARIABILITY IN CYP2D6-MEDIATED DRUG METABOLISM Xian Pan and Hyunyoung Jeong University of Illinois at Chicago, Chicago, IL, USA Cytochrome P450 2D6 (CYP2D6)-mediated drug metabolism displays a large interindividual variability. Previous studies suggest that differential transcriptional regulation of CYP2D6 may play a role in determining CYP2D6 activity (Carcillo et al., 2003; Temesvari et al., 2012); however, factors governing CYP2D6 transcription are largely unknown. We recently demonstrated small heterodimer partner (SHP) as a novel transcriptional regulator of CYP2D6 expression; SHP represses transactivation of CYP2D6 promoter by hepatocyte nuclear receptor 4a (HNF4a) (Koh et al., 2014). The objective of this study is to investigate whether differential expression of SHP leads to altered CYP2D6 expression and activity. To this end, we examined (I) the effects of a SHP inducer (an agonist of farnesoid X receptor, GW4064) on CYP2D6 expression and activity in primary human hepatocytes and CYP2D6-humanized transgenic (Tg-CYP2D6) mice and (II) the correlation between SHP and CYP2D6 expression in healthy human liver tissues. Firstly, in primary human hepatocytes, treatment with GW4064 (10 mM) for 48 hours led to a 3-fold increase in SHP and a 2-fold decrease in CYP2D6 mRNA expression. Results from S9 phenotyping (by using debrisoquine as a probe drug for CYP2D6) showed 20% decrease in CYP2D6 activity upon GW4064 treatment. Secondly, in Tg-CYP2D6 mice, the administration of GW4064 (10 mg/kg, ip for 5 days) increased SHP mRNA expression by 2-fold, and decreased CYP2D6 expression and activity by 2-fold. Results from chromatin immunoprecipitation assay revealed the increased recruitment of SHP and decreased recruitment of both HNF4a and RNA polymerase II on CYP2D6 promoter in Tg-CYP2D6 mice treated with GW4064, indicating that the enhanced SHP expression by GW4064 potentially represses HNF4a transactivation of CYP2D6 promoter. Thirdly, in healthy human liver tissues (n ¼ 8 donors), SHP protein and CYP2D6 mRNA levels were negatively correlated with each other (r ¼ 0.78, p ¼ 0.02). Together, these results indicate that enhanced SHP expression decreases CYP2D6 expression and activity in livers and suggest that differential expression level of SHP is a potential contributor to interindividual variability in CYP2D6-mediated drug metabolism.

References Carcillo JA, Adedoyin A, Burckart GJ, et al. (2003). Coordinated intrahepatic and extrahepatic regulation of cytochrome p4502D6 in healthy subjects and in patients after liver transplantation. Clin Pharmacol Therap 73:456–467. Koh KH, Pan X, Shen HW, et al. (2014). Altered expression of small heterodimer partner governs cytochrome P450 (CYP) 2D6 induction during pregnancy in CYP2D6-humanized mice. J Biol Chem 289:3105–3113.

64

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Temesvari M, Kobori L, Paulik J, et al. (2012). Estimation of drug-metabolizing capacity by cytochrome P450 genotyping and expression. J Pharmacol Exp Therap 341:294–305.

P49. EVALUATION OF LUCIFERIN-4A O-DEMETHYLASE ACTIVITY AS A CYP4A11 MARKER IN HUMAN LIVER AND KIDNEY MICROSOMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Satoshi Yamaori1, Mio Shionoiri2, Shigeru Ohmori1, and Kazuhito Watanabe2 1 Shinshu University Hospital, Matsumoto, Japan and 2Faculty of Pharmaceutical Sciences, Hokuriku University, Kanazawa, Japan Cytochrome P450 4A11 (CYP4A11) is mainly expressed in human liver and kidney. This enzyme catalyzes !-hydroxylation of saturated and unsaturated fatty acids, especially lauric acid and arachidonic acid although little is known about a role in drug metabolism. In addition, the CYP4A11 genetic polymorphisms are suggested to be associated with hypertension and coronary artery diseases. Thus, the functional analysis of CYP4A11 is very important to clarify its physiological role. Lauric acid has been widely used as a probe substrate of CYP4A11 in vitro. However, the measurement of lauric acid !-hydroxylation requires a radiolabeled substrate and complicated procedures. On the other hand, a luminogenic assay using 2-(6-methoxyquinolin-2-yl)4,5-dihydrothiazole-4-carboxylic acid (Luciferin-4A) is a simple and rapid method. However, CYP-isoform specificity of Luciferin-4A O-demethylase activity in human liver and kidney microsomes (HLMs and HKMs) remain to be unclear. In this study, we examined the specificity of luciferin-4A O-demethylation using HLMs and HKMs. Kinetic analysis indicated that the Vmax and Km values were 44.2 pmol/min/mg protein and 54.9 mM, respectively, for pooled HLMs and 42.5 pmol/min/mg protein and 39.2 mM, respectively, for pooled HKMs. On the other hand, the Vmax and Km values for recombinant CYP4A11 (rCYP4A11) were 705 pmol/min/nmol P450 and 33.6 mM, respectively. Luciferin-4A O-demethylase activities in pooled HLMs and HKMs as well as rCYP4A11 were efficiently inhibited by HET0016, a selective inhibitor of arachidonic acid !-hydroxylase including CYP4A11 (Miyata et al., 2001). The IC50 values for HLMs, HKMs, and rCYP4A11 were 33.3, 21.3, and 17.2 nM, respectively. In contrast, the other CYP isoform-selective inhibitors tested did not show any significant inhibition. Luciferin-4A O-demethylase activities in 16 individual HLMs significantly correlated with lauric acid !-hydroxylase activities measured by XenoTech (r ¼ 0.904, p40.0001). These results indicate that luciferin-4A O-demethylase activity is a good marker of CYP4A11 activity in HLMs and HKMs.

Reference Miyata N, Taniguchi K, Seki T, et al. (2001). HET0016, a potent ant selective inhibitor of 20-HETE synthesizing enzyme. Br J Pharmacol 133:325–329.

P50. RAPID IDENTIFICATION AND PRODUCTION OF METABOLITES USING CYPEXPRESS-2D6, 3A4 AND 2C9 Mani Subramanian1, Shuvendu Das1, Enrique Martinez2, and Denis M. Callewaert2 1 Center for Biocatalysis and Bioprocessing, The University of Iowa, Coralville, IA, USA and 2Oxford Biomedical Research, Inc., Rochester Hills, MI, USA Every drug in development needs to be fully characterized with respect to all the metabolites produced in humans. A wide range of methods has been developed to rapidly screen drug candidates for metabolism by cytochromes P450. However, production and purification of metabolites in quantities sufficient for Metabolite in Safety Test (MIST) studies is presently a challenging, time-consuming and expensive process – whether the preparative-scale synthesis of metabolites is performed using classical chemical synthesis, or liver microsomes. These technologies often have low yields, multiple byproducts, and microsome preparations suffer from lot-to-lot variability – and the requirement for large numbers of animals. Biocatalytic tools involving recombinant human P450s expressed in various hosts have been available for several years and can be very useful for metabolite identification. However, these systems are typically only stable for short periods, and require addition of NADPH or a system to generate NADPH, so that their use to generate milligram quantities of metabolites can prove extremely expensive. We have developed a novel system, CypExpressÔ; in which human liver CYPs are co expressed with human NADPH-P450 oxidoreductase in Pichia cells. Using proprietary processes, these cells are grown in large-scale fermenters, harvested and dried to a stable semi-permeable powder that retains very high levels of P450, NADPH-P450 oxidoreductase and NADPHregenerating activity. CypExpressÔ should prove invaluable for large-scale preparation of drug metabolites, production of reference standards, drug–drug interaction analysis, and rapid identification of human drug metabolites. CypExpressÔ CYP2D6, CYP3A4 and CYP2C9 products are already available. All three of these products mimic the human liver reactions of corresponding CYPs when tested using recommended standard substrates for each CYP. The present studies are focused on (a) the utility of CypExpressÔ to scale up production of specific metabolite(s) – including the ability to recycle CypExpressÔ for higher efficiency production of metabolites, (b) optimization of catalytic parameters for this system vs. traditional biocatalytic systems, and (c) direct comparison of the yields obtained using CypExpressÔ vs. commercial liver microsomes. CypExpressÔ2C9 catalyzed conversion of diclofenac (DN) to 4-hydroxydiclofenac (HDN) in a 400 mL reaction generated 53.6 mg of HDN from 59.2 mg of DN in only 1.5 h. In two-cycles of a 200 mL reaction containing 14.4 mg of testosterone (TE),

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

65

CypExpressÔ3A4 catalyzed the production of 7.01 mg of 6b-hydroxytestosterone (HTE) in 3.0 h. In two cycles of a 200 mL reaction containing 27 mg of dextromethorphan, CypExpressÔ2D6 catalyzed production of 7.5 mg dextrorphan. Reaction conversion rates and product formation in reactions catalyzed by CypExpressÔ 3A4, 2C9, and 2D6 were directly compared to reactions catalyzed by rat liver microsomes. CypExpressÔ is a very efficient, cost-effective, and less complex alternative for scale-up production of specific metabolites.

P51. INHIBITORY POTENTIAL OF THELEPHORIC ACID AND TSAHC ON CYP2J2 ACTIVITIES IN HUMAN LIVER MICROSOMES Zhexue Wu, Boram Lee, Se Yong Kim, Jong-Cheol Shon, Kwang-Hyeon Liu, and Kyung-Sik Song College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, South Korea

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Cytochrome P450 2J2 (CYP2J2) is an enzyme mainly found in human extrahepatic tissues, with predominant expression in the cardiovascular system. CYP2J2 plays important roles in the metabolism of endogenous metabolites and therapeutic drugs, such as arachidonic acid, astemizole, ebastine, and terfenadine. CYP2J2 is also over-expressed in human cancer tissues and cancer cell lines and may represent a potential target for therapy of human cancers. In this study, 50 compounds were screened as potential CYP2J2 inhibitors. Among them, thelephoric acid and TSAHC showed strong inhibition of astemizole-demethylation activity (IC50 ¼ 3.23 mM and 2.60 mM, respectively) in a dose-dependent manner. The present data suggest that these compounds might be potential candidates for further evaluation for anticancer activity.

P52. POTENTIAL OF DECURSIN TO INHIBIT THE HUMAN CYTOCHROME P450 2J2 ISOFORM Zhexue Wu, Boram Lee, Taeho Lee, Min Young Lee, Kyung-Sik Song, and Kwang Hyeon Liu College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, South Korea CYP2J2 enzyme is highly expressed in human tumors and carcinoma cell lines, and epoxyeicosatrienoic acids, CYP2J2mediated metabolites, have been implicated in the pathologic development of human cancers. To identify a CYP2J2 inhibitor, 50 natural products obtained from plants were screened using astemizole as a CYP2J2 probe substrate in human liver microsomes. Of these, decursin noncompetitively inhibited CYP2J2-mediated astemizole O-demethylation and terfenadine hydroxylation activities with Ki values of 8.34 mM and 15.8 mM, respectively. It also showed cytotoxic effects against human hepatoma HepG2 cells in a dose-dependent manner while it did not show cytotoxicity against mouse hepatocytes. The present data suggest that decursin is a potential candidate for further evaluation for its CYP2J2 targeting anti-cancer activities. Studies are currently underway to test decursin as a potential therapeutic agent for cancer.

P53. EVALUATIONS OF CYTOCHROME P450 (CYPS) 2D6 ACTIVITY IN THE SMALL INTESTINE OF JAPANESE PATIENTS. IN COMPARISON WITH THE CYP3A4 ACTIVITY AND THE RELATION WITH GENETIC POLYMORPHISM, MRNA, AND PROTEIN EXPRESSION Momoko Kawakami1, Sachiko Takenoshita-Nakaya2, Yuko Takeba2, Yuki Nishimura3, Minoru Watanabe2, Yuki Ohta2, Shinichi Kobayashi1, Naoki Matsumoto2, and Naoki Uchida1 1 Department of Clinical Pharmacology, Showa University School of Medicine, Tokyo, Japan, 2Department of Pharmacology, St. Marianna University School of Medicine, Kanagawa, Japan, and 3Department of Pharmacology, Showa University School of Medicine, Tokyo, Japan Introduction: Cytochrome P450 (CYPs) are expressed not only in the liver but also in the small intestine. CYP2D6 is known to be involved in the metabolism of various psychotropic agents. We previously reported about expressions of CYP2D6 mRNA, protein and genetic polymorphism in the small intestine of Japanese patients. However, CYP2D6 enzyme activity is still unclear as well as the relation between the CYP2D6 enzyme activity and genetic polymorphism, mRNA and protein expression in the small intestine of Japanese patients. Methods: Small intestine tissues were obtained from 50 Japanese patients who had pancreatouduodenectomy at St. Marianna University Hospital. Genotyping for CYP2D6 was performed using a direct sequence method. CYP2D6 mRNA level was measured by real time RT-PCR. CYP2D6 protein level was analyzed by Western Blot analysis.The activity of CYP2D6 from the intestinal microsome was measured by Luminescent method. Results: CYP2D6 genotype showed following three different patterns (*1/*1, 50%), (*1/*10, 50%), (*10/*10, 50%). The mRNA levels of CYP2D6 *10/*10 were higher than that of other CYP2D6 genotypes. There was no correlation between CYP2D6 mRNA levels and CYP2D6 protein levels. There was no relation between CYP2D6 genetic polymorphism and CYP2D6 protein levels. CYP2D6 activities were below of detection even using 80 mg microsomal protein content in samples, although CYP3A4 activities were detected in 1.6 mg protein content. Conclusions: As the detection of CYP2D6 activity was difficult, we could not clarify the relationship between the CYP2D6 activity, genetic polymorphism, mRNA and protein expression, respectively. These results suggest the CYP2D6 function for the first-pass metabolism may be so little as to be neglected in Japanese small intestine.

66

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P54. ROLE OF MICRORNA IN SPECIES-SPECIFIC INDUCTION OF CYP2B

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Lyudmila F. Gulyaeva1, Semen K. Kolmykov2, Mikhail D. Chanyshev1, Irina N. Krylova2, and Sergey S. Nechkin1 1 Department of Molecular carcinogenesis, Institute of MolBiol and Bioph SB RAMS, Novosibirsk, Russia and 2Department of Natural Sciences, Novosibirsk State University, Novosibirsk, Russia Cytochromes P450 of the 2B subfamily are xenobiotic-inducible enzymes responsible for the metabolism of a variety of drugs, but the mechanisms of this induction remain an enigma. Such, while a nuclear receptor CAR mediates CYP2B inducibility, no direct binding to CAR ligand binding domain is demonstrated for some of the most potent inducers, implying the involvement of the alternative mechanisms in the induction. Furthermore, the CAR-mediated CYP2B inducibility differs even between common rodent animal models: TPD (2,4,6-tryphenildioxane-1,3) is a potent inducer in rats but has no effect in mice, while TCPOBOP (1,4-bis[2-(3,5-dichloro)pyridyloxy]benzol) is a potent inducer in mice but not in rats. Previously, we showed that the speciesdependent differences of CYP2B induction occur on the level of the transcriptional activation. Here, we investigated the roles of microRNA in the regulation of the disparate CYP2B inducibility in mouse and rat species. To that end, we performed an in silico analysis of the CYP2B1 and cyp2b10 genes and their 30 -UTR sequences, followed by the experimental confirmation of the selected miRs expression in the livers of phenobarbital-, TPD- and TCPOBOP-treated animals. In silico analysis revealed the presence of the binding sites for a variety of miRNA species in the 30 -UTR sequences of the CYP2B1 and cyp2b10 genes. Using multiple search algorithms (Targetscan, miRDB, Diana-microT) and miR databases (miRBase, UCSC, NCBI, Ensembl) we selected the following microRNAs: -214, -143, -23, -27, -673 as potential regulators of rat and mouse CYP2B genes. Importantly, through the in silico analysis of 3’UTRs we found the presence of the binding sites for the discrete microRNAs in rat (miR-3557, -146b) and mouse (miR-3057,-421) CYP2B1 and cyp2b10 genes. RT-PCR analysis (with small nuclear U6 RNA employed as a reference gene) has confirmed the expression for the selected miRs species in rat and mouse livers upon PB, TPD, or TCPOBOP exposures. The induction of CYP2B enzymes was measured by the (1) RT-PCR the of CYP2B mRNA, (2) pentoxyresorufin O-dealkylation, and (3) the Western-blot analysis. We found that all candidate miRs were constitutively expressed in the liver of male Wistar rats and male C57BL mice. Expression of microRNAs -214, -143, -23, -27, -673 under the exposure with CYP inducers was associated with the changes in levels of transcripts and the levels of activities of CYP2B. Strikingly, the changes in expression of the candidate species-specific miRs were consistent with the type of the inducer and the animal species used. Thus, we showed that one mechanism of the species-specific inducibility of the CYP2B enzymes is via the species-specific miR regulation. This observation has implication for the improvement in the drug discovery success rate, in the light of understanding of the specifics of drug metabolism in different species and the possible variations in the drug metabolism between different human subjects. This study was supported by RFBR (grant No. 12-04-00642-a).

P55. PHENOTYPE-GENOTYPE ANALYSIS OF CYP2C19 IN SICKLE CELL DISEASE PATIENTS AND NON-SICKLE CELL VOLUNTEERS IN NIGERIA Olufunmilayo Adejumo1, Chinedum Peace Babalola2, Adeyinka Gladys Falusi3, Swati V. Nagar4, and Taiwo Kotila5 1 Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Olabisi Onabanjo University, Ago-Iwoye, Nigeria, 2 Department of Pharmaceutical Chemistry, University of Ibadan, Ibadan, Nigeria, 3Genetic and Bioethics Unit, IMRAT, University of Ibadan, Ibadan, Nigeria, 4Department of Pharmaceutical Sciences, Temple University School of Pharmacy, Philadelphia, PA, USA, and 5Department of Haematology, College of Medicine, University of Ibadan, Ibadan, Nigeria Background: Cytochrome P4502C19 is a polymorphic enzyme responsible for the metabolism of proguanil (PG), a prophylactic antimalarial to its active metabolite, cycloguanil (CG). Nigeria carries the highest sickle-cell disease (SCD) burden globally (2%) with a high prevalence of malaria, hence PG is used daily by SCD patients. Information on the genetic polymorphism of PG in the Nigerian population is scanty and no report exists for SCD patients. The aim of this study was to investigate and compare CYP2C19 phenotype and genotype prevalence in SCD patients and non-SCD controls in Nigeria. Methods: One hundred and fifty-eight unrelated Nigerians comprising 115 SCD and 43 non-SCD volunteers sampled in South-west Nigeria were genotyped for CYP2C19*1, CYP2C19*2 and CYP2C19*3 from DNA prepared from blood samples collected pre-dose. Polymerase chain reaction (PCR)-restriction fragment length polymorphism analysis was employed for CYP2C19 genotyping. Phenotyping for proguanil metabolism to CG was carried out by collecting pooled urine samples for 8 hours after PG administration. Proguanil and cycloguanil concentrations were measured by HPLC and metabolic ratio (MR) calculated as PG/CG concentrations. Results: Genotyping for non-SCD volunteers showed 95.3% extensive metabolisers (EMs) and 4.7% poor metabolisers (PMs) while 99.1% EM and 0.9% PM were observed for SCD volunteers. Genotypic prevalence frequency of homozygous CYP2C19*2/*2 for PMs was significantly lower in SCD (0.9%) compared with nonSCD (4.7%) (p ¼ 0.04). General frequencies of CYP2C19*1 (84.3%; 84.9%) and CYP2C19*2 alleles (15.7%; 15.1%) in SCD and non-SCD were not significantly different (p40.05). Sickle cell disease volunteers exhibited a lower frequency of homozygous CYP2C19*1/*1 genotype for EMs (69.6%) compared with non-SCDs (74.4%) (p ¼ 0.04). Frequencies of heterozygous CYP2C19*1/*2 were not significantly different in SCD and non-SCD volunteers (p40.05). No CYP2C19*3 allele was detected in the two groups. Phenotyping showed that 99.2% were EMs and 0.8% PM of PG among SCD while 95.5% and 4.4% non-SCD volunteers were EMs and PMs, respectively. Relating genotype with phenotype, MRs for the SCD EMs with CYP2C19*1/*1 allele ranged from 0.02 to 8.70, while those with CYP2C19*1/*2 allele ranged from 0.06 to 7.60.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

67

The MRs for the non-SCD EMs with CYP2C19*1/*1 allele ranged from 0.22 to 8.33 while those with CYP2C19*1/*2 allele ranged from 1.0. 8.33. Two volunteers with MRs of 18.18 and 25.76 among the non-SCD and one SCD patient with MR of 16.77 were genotyped as CYP2C19*2/*2. They were also phenotyped as PMs showing complete concordance between genotyping and phenotyping of CYP2C19. Conclusion: Prevalence of CYP2C19 allele polymorphisms was defined in Nigerian sickle-cell disease and non-sickle-cell disease populations. Poor metabolisers may be predicted from phenotyping or genotyping of CYP2C19. A clear difference in metabolic disposition of sickle-cell disease and healthy persons was established for proguanil and CYP2C19.

P56. Characterization of the cytochrome P450 enzymes involved in the metabolism of TSAHC

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Boram Lee1, Ki Hun Park2, and Kwang-Hyeon Liu1 1 College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, South Korea and 2Division of Applied Life Science, Gyeongsang National University, JinJu, South Korea 40 -(p-Toluenesulfonylamide)-4-hydroxychalcone (TSAHC) is a synthetic sulfonylamino chalcone compound which has inhibitory effects on anti-cancer activity. This study was performed to elucidate the metabolism of TSAHC in human liver microsomes (HLMs) and to characterize the cytochrome P450 (P450) enzymes that are involved in the metabolism of TSAHC. Human liver microsomal incubation of TSAHC in the presence of NADPH resulted in the formation of two metabolites, hydroxy-TSAHC M1 and hydroxy TSAHC M2. Correlation analysis between the known P450 enzyme activities and the rate of the formation of TSAHC metabolites in the 15 HLMs showed that TSAHC metabolism is correlated with CYP2C and CYP3A activity. The P450 isoformselective inhibition study in liver microsomes and the incubation study of cDNA-expressed enzymes also showed that the hydroxy TSAHC M1 and M2 are mainly mediated by CYP2C and CYP3A, respectively. These findings suggest that CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4 and CYP3A5 isoforms are major enzymes contributing to the disposition of TSAHC.

P57. TIME-DEPENDENT INACTIVATION AND PRODUCTION OF A REACTIVE QUINONE METHIDE METABOLITE BY CYP3A7 METABOLISM OF NEVIRAPINE: IMPLICATION FOR HEPATOTOXICITY IN HIV INFECTED NEONATES Jed N. Lampe Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, KS, USA In 2014, the second baby born to an HIV infected mother was declared ‘‘cured’’ of HIV after more than three years with no detectable blood borne virus (Wilson & Young, 2014). The treatment, consisting of high doses of the drug cocktail AZT, 3TC, and nevirapine, was given to the infant from 4 h. after birth until 12 months of age (Wilson & Young, 2014). In adults, one of the primary routes of nevirapine metabolism and elimination is through oxidation mediated by hepatic CYP3A4 (Erickson et al., 1999; Wen et al., 2009). In addition to several other metabolites, CYP3A4 is known to produce a reactive quinone methide metabolite at the C-12 position of nevirapine that is thought to be responsible for the idiosyncratic liver toxicity seen in approximately 9% of adult patients taking this drug (Sharma et al., 2012). However, despite the fact that this drug is used clinically to treat HIV in neonates, its metabolism has never been fully characterized in this population subgroup. Our hypothesis was that CYP3A7, the predominant neonatal hepatic CYP isoform, would produce a qualitatively similar metabolic profile as seen with adult CYP3A4, albeit with important quantitative differences. To this end, the metabolism of nevirapine was analyzed using both recombinant CYP3A7 and pediatric human liver microsomes. Metabolites were quantified using LCMS and authentic metabolite standards. The reactive quinone methide metabolite was quantified by GSH trapping experiments. Here, we demonstrate that CYP3A7 produces a metabolic profile similar to CYP3A4, although with increased production of the reactive quinone methide metabolite. Additionally, we demonstrate that CYP3A7 is subject to the same time-dependent inactivation kinetics as CYP3A4, indicating that neonates prescribed with nevirapine may be at increased risk for both drug-drug interactions and hepatic idiosyncratic toxicity, as seen in the adult population.

References Erickson DA, Mather G, Trager WF, et al. (1999). Characterization of the in vitro biotransformation of the HIV-1 reverse transcriptase inhibitor nevirapine by human hepatic cytochromes P-450. Drug Metab Dispos 27:1488–1495. Sharma AM, Li Y, Novalen M, et al. (2012). Bioactivation of nevirapine to a reactive quinone methide: Implications for liver injury. Chem Res Toxicol 25:1708–1719. Wen B, Chen Y, Fitch WL. (2009). Metabolic activation of nevirapine in human liver microsomes: Dehydrogenation and inactivation of cytochrome P450 3A4. Drug Metab Dispos 37:1557–1562. Wilson J, Young S. (2014). Second baby possibly ‘‘cured’’ of HIV. Available from: http://www.cnn.com [last accessed 27 Jun 14].

P58. DEVELOPMENT OF AN IN VITRO SCREENING ASSAY FOR CYP1B1-TARGETED ANTICANCER PRODRUGS USING THE KLE HUMAN ENDOMETRIAL CARCINOMA CELL LINE Zhiying Wang, Yao Chen, and Michael Zhuo Wang Department of Pharmaceutical Chemistry, University of Kansas, Lawrence, KS, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

68

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Overexpression of cytochrome P450 1B1 (CYP1B1) has been associated with a number of human steroid hormones-related tumors (e.g. endometrial cancer) and CYP1B1 has been proposed as a novel target for the development of anticancer therapeutics. One strategy is to develop prodrugs specifically activated by the tumor CYP1B1, resulting in a targeted tumor suppression. To this goal, an in vitro screening assay needs to be developed to select promising CYP1B1-targeted anticancer prodrugs. KLE human endometrial carcinoma cell line was previously shown to express CYP1B1 and its up-regulation played a crucial role in endometrial carcinogenesis. Here, we aimed to employ KLE cells to develop an in vitro screening assay for CYP1B1-targeted anticancer prodrugs. First, mRNA levels of CYP1 enzymes (CYP1A1, CYP1A2 and CYP1B1) and ethoxyresorufin-O-deethylase (EROD) activity were determined in KLE cells and effect of growth conditions (e.g. postseeding growth time and presence of phenol red in culture media) on CYP1B1 expression was also characterized. Second, using two proposed CYP1B1-activated prodrugs, DB289 and DB844, cell viability after 72 h drug treatment was determined using AlamarBlue assay and compared to drug treatment in the presence of a CYP1B1 inhibitor, a-naphthoflavone (a-NF). In comparison to human liver and differentiated HepaRG cells, KLE cells showed substantially higher CYP1B1, but markedly lower CYP1A1 and CYP1A2 expression. In KLE cells, a significant growth days-dependent CYP1B1 expression can be observed in both mRNA levels and functional activities. CYP1B1 mRNA expression reached the highest level on day 9 post-seeding and the EROD activity increased from day 5 to day 11 post-seeding before decreasing afterwards. In contrast, CYP1B1 mRNA levels reduced quickly in HEK293T cells transiently transfected with CYP1B1 (approximately 62% and 8% remaining 2 days and 8 days post-transfection relative to 24 h post-transfection, respectively). DB289 and DB844 at 50 mM exhibited moderate toxicity towards KLE cells with 19% and 18% of cells remaining, respectively. Importantly, inhibition of CYP1B1 activity by a-NF (1 mM) significantly ameliorated the cytotoxicity of DB289 and DB844 in KLE cells (55% and 47% of cells remaining). In summary, a KLE cell-based screening assay has been developed to identify CYP1B1targeted anticancer prodrugs and this assay can be further optimized as a high-throughput assay to screen large compound libraries.

P59. A NOVEL EVALUATION SYSTEM FOR INHIBITION OF CYP3A4 ACTIVITY BY DIETARY SUPPLEMENTS Yu Sato, Shogo Takahashi, Takamitsu Sasaki, Takeshi Kumagai, and Kiyoshi Nagata Department of Environmental and Health Sciences, Tohoku Pharmaceutical University, Miyagi, Japan Dietary supplements, being believed by many people to be safe, have used for self-medication and treat at a variety of health conditions. In this case, drug–herb interactions could represent one of the most important problems in therapeutics; however, the information on this interaction is very limited. Drug interactions that inhibition of cytochrome P450 (P450) was most accrued. Generally, for in vitro P450 inhibition experiments, human liver microsomes have been used; however, these experiments do not consider the absorption of compounds into the cells. Hydrophobic compounds that included in dietary supplements such as glycosides and carboxylated compounds are likely to be poorly transferred into the intestinal epithelium through the membrane. In this study, we developed a novel evaluation system to evaluate drug metabolism, in which we modeled the absorption of compounds into cells using frozen P450-expressing cells. Using this system, we evaluated the extent of CYP3A4 inhibition by dietary supplements. Human hepatocarcinoma HepG2 cells were infected with an adenovirus vector expressing human CYP3A4 (Ad-CYP3A4 cells). The Ad-CYP3A4 cells were cryopreserved (fAdCYP3A4 cells), and cytolysis was performed at the time of evaluation. CYP3A4 mRNA and enzyme activity of AdCYP3A4 cells, fAd-CYP3A4 cells and CYP3A4 microsomes were measured using real-time PCR and P450-GloTM assay, respectively. Similarly, inhibition of CYP3A4 activity by ketoconazole (a typical CYP3A4 inhibitor), or the dietary supplements was measured. fAd-CYP3A4 cells showed CYP3A4 mRNA levels and enzyme activity levels that were similar to those of non-cryopreserved Ad-CYP3A4 cells. The IC50 values for inhibition of CYP3A4 activity by the test compounds in fAd-CYP3A4 cells did not differ from those in Ad-CYP3A4 cells and CYP3A4 microsomes. However, the inhibitory effects of dietary supplements were stronger in CYP3A4 microsomes than in fAd-CYP3A4 cells. This difference may be attributable to the differential penetration abilities of the compounds present in the dietary supplements into fAd-CYP3A4 cells. These results suggest that our evaluation system mimicking human intestinal drug metabolism might be useful for studying the inhibition of CYP3A4 activity by dietary supplements.

P60. ROLES OF AUTO-INDUCTION OF HUMAN P450 3A ENZYMES INVOLVED IN THE TERATOGENICITY OF THALIDOMIDE Norie Murayama1, Christophe Chesne´2, Norio Shibata3, F. Peter Guengerich4, and Hiroshi Yamazaki1 1 Showa Pharmaceutical University, Machida, Japan, 2Biopredic International, Rennes, France, 3Nagoya Institute of Technology, Nagoya, Japan, and 4Vanderbilt University School of Medicine, Nashville, TN, USA Teratogen thalidomide has been approved for treatments of myeloma. The teratogenicity of thalidomide is species-specific, but its mode of action still remains unclear. We recently demonstrated heterotropic cooperativity of human cytochrome P450 (P450) 3A4/5 by thalidomide using the model substrate midazolam in various in vitro and in vivo models. In the present study, we further investigated the regulation of drug metabolizing enzymes by thalidomide in human livers.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

69

Chimeric mice with humanized liver, but not control mice, displayed enhanced midazolam clearance upon pre-treatment with orally administered thalidomide, presumably because of human P450 3A induction. Thalidomide enhanced levels of P450 3A4 mRNA, protein expression, and/or oxidation activity in human HepaRG cells indirectly, suggesting activation of gene transcription factors. Interestingly, the resulting ratios of P450 3A induction were apparently high when hydrocortisone was omitted from basal culture medium and pregnancy-related hormones were fortified. Thalidomide and its human proportionate metabolite 5-hydroxythalidomide displayed significant modulation of real-time coregulator-nuclear receptor interaction with pregnane X receptor (PXR) ligand-binding domains, similar to established agonists for this receptor in a microarray assay. These results collectively suggest that thalidomide acts as a ligand for PXR and cause enzyme induction and auto-induced P450 3A4 might involve in the metabolic activation of thalidomide and causing teratogenicity.

P61. ABSTRACT WITHDRAWN

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P62. THE ROLE OF HEPATOCYTE NUCLEAR FACTOR 3b IN CYP2B6 GENE EXPRESSION Linhao Li1, Scott Heyward2, and Hongbing Wang3 1 Department of Pharmaceutical Sciences, University of Maryland at Baltimore, Baltimore, MD, USA, 2Celsis In vitro Technologies, Baltimore, MD, USA, and 3Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore, MD, USA Cytochrome p450 2B6 (CYP2B6) plays an increasingly important role in xenobiotics metabolism and detoxification. Nuclear receptors constitutive androstane receptor (CAR) and pregnane X receptor (PXR) have been clearly established as the predominant regulators for the inductive expression of CYP2B6 gene in human liver. However, there is dramatic interindividual variability in CYP2B6 gene expression that cannot be fully explained by a simple CAR/PXR-based modulation model. Here we show that induction of CYP2B6 by phenobarbital (PB), a typical CYP2B6 inducer, was correlated with the expression of hepatocyte nuclear receptor 3b (HNF3b) (R = 0.7549, **p50.01) in human primary hepatocyte (HPH) cultures prepared from 29 liver donors. The mRNA level of HNF3b was also moderately up-regulated by typical CAR activators, PB and 6-(4-chlorophenyl)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde-O-(3,4-dichlorobenzyl) oxime (CITCO), as well as PXR activator rifampicin. Computer-based analysis and cell-based serial deletion assays revealed that presence of HNF3b significantly increased the basal activity of the CYP2B6 luciferase reporter. Two novel enhancer modules located upstream of the CYP2B6 gene transcription start site were identified and postulated to be key elements governing the HNF3b-mediated CYP2B6 expression. Nevertheless, overexpression of HNF3b in HepG2 cells was not synergistically associated with the activation of CYP2B6 reporter-mediated by CAR or PXR. Overall, whereas additional experiments aiming to further confirm the recruitment of HNF3b to these enhancers in the promoter region of CYP2B6 gene are underway, our results thus far indicate that HNF3b may represent another liver enriched transcription factor that contributes to the transcription of CYP2B6 gene in human liver.

P63. IN VITRO MODULATORY EFFECT OF MACHILIN A ISOLATED FROM MACHILUS THUNBERGII ON HUMAN CYTOCHROME 1A2 AND 2B6 Sunju Kim1, JuHee Sim1, Oh Kwang Kwon1, Jong-Sup Bae1, Jeong Ah Kim1, Seung Ho Lee2, Kwang Hyeon Liu1, and Sangkyu Lee1 1 College of Pharmacy, Kyungpook National University, Deagu, South Korea and 2College of Pharmacy, Yeungnam University, Gyeongsan, South Korea Machilin A (MA), one of the lignans isolated from Machilus thunbergii, showed several biological activities as a folk medicine for the treatment of leg edema, abdominal pain and abdominal distension in Korea. Although there have been studies of the biological effects of MA, the potential drug-interaction effect of MA following evaluation of inhibitory effects of cytochrome P450 (CYP) activities was not investigated. Here, we investigated the inhibitory effects of MA on the activities of CYP isoforms using a cocktail of probe substrates in pooled human liver microsome (HLMs) and human recombinant cDNA-expressed CYP. MA strongly inhibited CYP1A2-mediated phenacetin O-deethylation and CYP2B6mediated bupropion O-deethylation with an IC50 value of 3.0 mM and 3.9 mM, respectively, while other CYPs were not significantly inhibited. A Dixon plot resulted in the inhibition mechanism of MA being competitive inhibition. In addition, in human recombinant cDNA-expressed 1A2 and 1A1, MA decreased CYP 1 A-catalyzed phenacetin O-deethylase activity with IC50 values of 1.72 mM and 1.53 mM, respectively. Also, CYP 2B6-catalyzed bupropion hydroxylation activity with IC50 values of 2.59 mM in human recombinant cDNA-expressed 2B6. Overall, this is the first investigation of potential drug–drug interactions associated with MA conducted by identifying the competitive inhibitory effects of MA on CYP1A2 and CYP2B6 in HLMs.

70

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P64. NEW TOOLS TO DIFFERENTIATE CYP3A4 AND CYP3A5 ACTIVITY Xiaohai Li1, Theodore M Kamenecka1, Valer Jeso2, Glen Micalizio2, Scott Heyward3, Gregory S. Walker4, and Michael D. Cameron2 1 Department of Molecular Therapeutics, The Scripps Research Institute, Jupiter, FL, USA, 2 Department of Chemistry, Dartmouth College, Hanover, NH, USA, and 3Bioreclamation IVT, Baltimore, MD, USA, and 4Department of Pharmacokinetics, Dynamics & Metabolism, Pfizer, Groton, CT, USA Almost half of the prescription medications are metabolized by cytochrome P450 3A4 and 3A5. Substrate overlap between CYP3A4 and 3A5 have led to the erroneous habit of attributing the cumulative activity to CYP3A4. While CYP3A4 expression is ubiquitous, CYP3A5 expression is polymorphic, with large individual differences in CYP3A5 expression level. Altered pharmacokinetics and variable drug-interactions have been observed for several compounds based on CYP3A5 genotype. We detail new tool compounds recently discovered within our lab to elucidate the contribution of CYP3A5 in biological samples containing both enzymes. These include a CYP3A5 selective substrate that is over 100-fold selective over CYP3A4 and other common hepatic P450 enzymes. Additionally, a series of highly selective CYP3A4 inhibitors have been discovered capable of isolating CYP3A5 activity.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P65. DIFFERENTIAL INTERACTIONS OF STEROIDOGENIC CYTOCHROME P450 17A1 WITH ITS SUBSTRATES Elyse M. Petrunak1, Natasha DeVore2, and Emily E. Scott1 1 Department of Medicinal Chemistry, University of Kansas, Lawrence, KS, USA and 2Department of Molecular Biosciences, University of Kansas, Lawrence, KS, USA Cytochrome P450 17A1 (CYP17A1) differs from xenobiotic-metabolizing P450 enzymes in both its key role in endogenous steroid biosynthesis and in performing a two-reaction sequence performed to generate testosterone precursors. This reaction sequence involves a typical hydroxylation reaction followed by an unusual carbon-carbon bond cleavage (lyase) reaction thought to be catalyzed by a peroxo intermediate. Although CYP17A1 catalyzes the hydroxylation with little disparity between its two physiological substrates (D4-progesterone and D5-pregnenolone), the lyase reaction shows preference for one substrate (D5-17ahydroxypregnenolone over its counterpart D4-17a-hydroxyprogesterone), suggesting that differential interactions of the CYP17A1 enzyme with its substrates might regulate lyase reactivity. No structures are currently available for wild type CYP17A1 with any substrate to compare such interactions. However, mutation of a single active site residue (Alanine 105 a` Leucine) facilitated this goal. Previous examination of this mutation, prompted by species comparisons, reported increases in turnover for both progesterone hydroxylation (2-fold) and lyase activity (5-fold). Current data show that this mutation increases the binding affinity of CYP17A1 for progesterone and 17a-hydroxypregnenolone (both 2-fold) and facilitates cocrystallization with all four physiologically relevant substrates. The crystal structures of CYP17A1 A105L with both hydroxylase ˚ . The general orientation substrates and both 17,20-lyase substrates were determined at resolutions ranging between 2.5 and 3 A the steroidal nucleus is conserved among substrates. Both hydroxylase substrates and the inefficient lyase substrate exhibit a single binding mode (‘‘Position 1’’) in which the substrate forms a hydrogen bond with a key active site residue (Asn 202) and the site of metabolism is projected toward the heme iron. In contrast, the preferred lyase substrate can assume an additional binding mode (‘‘Position 2’’) in which the substrate does not form the aforementioned hydrogen bond to Asn202 and the site of ˚ closer to the heme iron than for Position 1. Observation of only the preferred lyase substrate in a position metabolism is 0.4 A closer to the heme iron is consistent with spectroscopic observations suggesting that the 17-hydroxyl group of this substrate interacts with the proximal oxygen of the peroxo species proposed for the lyase reaction. In comparison, the poor lyase substrate is not observed closer to the heme, consistent with spectroscopic studies suggesting that its 17-hydroxyl group interacts with the distal oxygen of the peroxo catalytic species. Such differential interactions with the catalytic intermediate could facilitate efficient catalysis for D5-17a-hydroxypregnenolone, but not its counterpart, D4-17a-hydroxyprogesterone. Such structural studies provide useful insights into how this enzyme functions, furthering our understanding of human steroidogenesis.

P66. CYCLIN-DEPENDENT KINASE 1 INVOLVED IN REGULATION OF CYP2D6 EXPRESSION Keita Inami1, Takamitsu Sasaki1, Syu-ichi Kanno2, Takeshi Kumagai1, and Kiyoshi Nagata1 1 Department of Environmental and Health Science, Tohoku Pharmaceutical University, Sendai, Japan and 2Department of Clinical Pharmacotherapeutics, Tohoku Pharmaceutical University, Sendai, Japan Induction of cytochrome P450 (CYP) is known to be under the control of ligand-binding transcription factors such as the aryl hydrocarbon receptor (AhR) and nuclear receptors. Polycyclic aromatic hydrocarbons (PAHs) are well-known AhR ligands that induce CYP1A1 and CYP1A2 expression via activation of AhR (Nebert et al., 2004). Furthermore, several studies have shown that PAHs can also induce CYP3A4 via the pregnane X receptor-mediated pathway (Luckert et al., 2013, Kumagai et al., 2012). In addition to these CYPs, we have observed CYP2D6 to be induced by PAHs; however, the mechanism of this CYP2D6 induction is yet to be elucidated. Recently, studies have reported that expression of CYP genes is regulated by events associated with the cell cycle (Sugatani et al., 2010). Therefore, we hypothesized that the induction of CYP2D6 by PAHs is mediated

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

71

through a cell cycle-associated pathway. We treated HepG2 cells with benzo[a]pyrene (B[a]P), dibenz[a,h]anthracene (DB[a,h]A), or b-naphthoflavone (b-NF) at a final concentration of 1 mM. After 24–72 h of exposure to these PAHs, the CYP mRNA expression and cell cycle distribution were analyzed by real-time PCR and flow cytometry, respectively. Exposure to B[a]P and DB[a,h]A increased in CYP2D6 expression, as well as CYP1A1 and CYP3A4 expression. The maximum increase in CYP2D6 expression (23.9-fold) was observed when the cells were treated with DB[a,h]A for 72 h. In addition, B[a]P and DB[a,h]A markedly increased the number of cells in the G2-M DNA damage checkpoint. b-NF did not increase CYP2D6 expression or alter the cell cycle distribution, even when the cells were treated with a high concentration of b-NF (50 mM) for 72 h. We further investigated whether various compounds that cause accumulation of cells at the G2-M checkpoint could increase CYP2D6 expression. Treatment with microtubule inhibitors nocodazole and vincristine led to G2-M arrest; however, increased CYP2D6 expression was not observed. In contrast, cyclin-dependent kinase 1 (Cdk1)-selective inhibitor RO-3306 increased CYP2D6 expression. Cdk1 is known as a key regulator of G2-M transition in eukaryotic cells. When the cells were treated with 10 mM RO-3306, CYP2D6 expression increased by 5.9-fold in comparison to that in the control cells. In conclusion, our data suggest that Cdk1 is involved in the regulation of CYP2D6 expression.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

References Kumagai T, Suzuki H, Sasaki T, et al. (2012). Polycyclic aromatic hydrocarbons activate CYP3A4 gene transcription through human pregnane X receptor. Drug Metab Pharmacokinet 27:200–206. Luckert C, Ehlers A, Buhrke T, et al. (2013). Polycyclic aromatic hydrocarbons stimulate human CYP3A4 promoter activity via PXR. Toxicol Lett 222:180–188. Nebert DW, Dalton TP, Okey AB, Gonzalez FJ. (2004). Role of aryl hydrocarbon receptor-mediated induction of the CYP1 enzymes in environmental toxicity and cancer. J Biol Chem 279:23847–23850. Sugatani J, Osabe M, Kurosawa M, et al. (2010). Induction of UGT1A1 and CYP2B6 by an antimitogenic factor in HepG2 cells is mediated through suppression of cyclin-dependent kinase 2 activity: Cell cycle-dependent expression. Drug Metab Dispos 38:177–186.

P67. ABSTRACT WITHDRAWN

P68. MECHANISM OF OXIDATION BY MITOCHONDRIAL CYP2E1 DIFFERS FROM MICROSOMAL CYP2E1 Jessica H. Hartman, H. Cass Martin, and Grover P. Miller Department of Biochemistry & Molecular Biology, University of Arkansas for Medical Sciences, Little Rock, AR, USA Cytochrome P450 2E1 (CYP2E1) is responsible for the metabolism of many low molecular-weight compounds, resulting in detoxification or bioactivation to toxic or carcinogenic metabolites. CYP2E1 is primarily localized in the endoplasmic reticulum (microsomal CYP2E1); however, the same active full-length enzyme is found in significant amounts in the mitochondria, where it forms an alternate catalytic complex with the mitochondrial adrenodoxin (Adx/Adr) system. This complex could potentially exhibit different reaction kinetics for substrates, involving changes in the efficiency and/or mechanism of CYP2E1 oxidation. Therefore, in this study, we analyzed in vitro steady-state kinetics of metabolism by mitochondrial CYP2E1 for probe substrates aniline and 4-nitrophenol, which have been reported to undergo cooperative (two-site) metabolism by microsomal CYP2E1. Mitochondria from livers of six Sprague-Dawley rats were purified by gradient centrifugation, homogenized, pooled, and used for analysis of mitochondrial CYP2E1 activity. The kinetic profile of microsomal CYP2E1 4-nitrophenol oxidation displays significant substrate inhibition. However, CYP2E1 in mitochondria undergoes simple Michaelis–Menten kinetics for 4nitrophenol metabolism with a Km of 35 mM and a Vmax of 28 pmol/min/mg total protein. Similarly, for aniline 4-hydroxylation to 4-aminophenol, negative cooperativity is observed for microsomal CYP2E1; however, aniline hydroxylation by mitochondrial CYP2E1 follows Michaelis–Menten kinetics (Km = 13 mM, Vmax = 14 pmol/min/mg protein). The lack of cooperativity observed in the kinetic profiles suggests that mitochondrial CYP2E1 catalysis utilizing Adx/Adr may change the cooperative binding site and/or rate limiting steps in the catalytic cycle. These distinct kinetic mechanisms could impact the relative importance of CYP2E1 localized in different cellular compartments in a substrate concentration-dependent manner and ultimately impact cellular toxicity arising from metabolism of CYP2E1 substrates. Ongoing experiments will expand the study to include other substrates of CYP2E1 which are metabolized to toxic or carcinogenic metabolites, including styrene and acetaminophen. Ultimately, the results of this study will provide the first detailed mechanistic perspective of metabolism by mitochondrial CYP2E1.

P69. ENZYMATIC PROPERTIES OF NADPH-CYTOCHROME P450 REDUCTASE FROM HOT PEPPER Young Hee Joung, Ga-Young Lee, Hyun Min Kim, Sang Hoon Ma, Se Hee Park, and Chul-Ho Yun School of Biological Sciences and Technology, Chonnam National University, Gwang-Ju, South Korea Genome sequencing of the hot pepper revealed that it has 447 genes encoding P450s and two genes encoding NADPHcytochrome P450 reductase (CPR) genes (Kim et al., 2014). When the expression patterns of two CPR genes (CaCPR1 and CaCPR2) from hot pepper (Capsicum annuum L. cv. Bukang) were examined, the expression level of CaCPR1 was higher than

72

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

that of CaCPR2 in all tested tissues (leaves, flowers, and fruits) and gradually increased during fruit ripening (Lee et al., 2014). The CaCPR2 gene seemed to be expressed constitutively in all of the tested tissues. The CaCPR1 gene was found to be inducible and considered as the major CPR in most hot pepper tissues. To investigate the enzymatic properties of CaCPR1, the cDNA of CaCPR1 was heterologously expressed in Escherichia coli without any modification of amino acid sequences, and CaCPR1 was purified. The enzymatic properties of CaCPR1 were confirmed using cytochrome c and cytochrome b5 as protein substrates. CaCPR1 also catalyzed the reduction of rabbit cytochrome b5 and supported the P450-catalysed reaction. It also reduced tetrazolium salts such as 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and 5-cyano-2,3-ditolyl tetrazolium chloride (CTC), and ferricyanide. Cytochrome c, cytochrome b5, MTT, CTC, and ferricyanide can be used in continuous assays. In particular, the fluorogenic substrate CTC can be used to detect CaCPR1 activity in vivo. CaCPR1 showed about 4-fold higher reductase activity toward doxorubicin and much higher thermal stability when compared to rat CPR. Taken together, it is suggested that the enzymatic properties of plant CaCPR1 are similar to those of CPRs from other plant (Jensen & Møller, 2010) and mammalian sources (Pandey & Flu¨ck, 2013). Although the CaCPR1 showed a high preference for NADPH over NADH as reducing coenzyme, the CaCPR1 was found to be an efficient NADH-utilizing reductase for P450-catalysed oxidation and cytochrome b5 reduction. It is suggested that the CaCPR1 can be used a prototype for studying biological functions and enzymatic properties of plant CPRs.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

References Jensen K, Møller BL. (2010). Plant NADPH-cytochrome P450 oxidoreductases. Phytochem 71:132–141. Kim S, Park M, Yeom SI, et al. (2014). Genome sequence of the hot pepper provides insights into the evolution of pungency in Capsicum species. Nat Genet 46:270–278. Lee GY, Kim HM, Ma SH, et al. (2014). Heterologous expression and functional characterization of the NADPH-cytochrome P450 reductase from Capsicum annuum. Plant Physiol Biochem 82C:116–122. Pandey AV1, Flu¨ck CE. (2013). NADPH P450 oxidoreductase: Structure, function, and pathology of diseases. Pharmacol Ther 138:229–254.

P70. REGIOSELECTIVE HYDROXYLATION OF OMEPRAZOLE ENANTIOMERS BY BACTERIAL CYP102A1 MUTANTS Chul-Ho Yun, Sang-Hoon Ryu, Bo-Yeon Park, Hyung-Sik Kang, and Young Hee Joung School of Biological Sciences and Technology, Chonnam National University, Gwang-Ju, South Korea A large set of Bacillus megaterium CYP102A1 mutants are known to metabolize various drugs to form human metabolites (Kang et al., 2014; Yun et al., 2007). Omeprazole (OMP), a proton pump inhibitor, has been widely used as an acid inhibitory agent for the treatment of gastric acid hypersecretion disorders including dyspepsia, peptic ulcer disease, gastroesophageal reflux disease, laryngopharyngeal reflux, and Zollinger–Ellison syndrome. It is primarily metabolized by human CYP2C19 and CYP3A4 to 50 -OH OMP and a sulfone product, respectively (Yamazaki et al., 1997). Recently, it was reported that several CYP102A1 mutants can oxidize racemic (Butler et al., 2013) and S-OMP (Butler et al., 2014) to 50 -OH OMP and that these mutants can further oxidize 50 -OH racemic OMP to 50 -COOH OMP (Butler et al., 2013). Here, we report that the S-and R-enantiomers of OMP are hydroxylated by 26 mutants of CYP102A1 to produce one major metabolite (50 -OH OMP) regardless of the chirality of the parent substrates. Chemical structure of the major metabolite (50 -OH OMP) was determined by NMR and LC-MS. The aim of this study was to investigate the enzymatic properties of CYP102A1 mutants to generate the 50 -OH product from S- and R-OMP enantiomers. Generally, the 50 -OH product formation rates of mutants for R-OMP were much higher than those for S-OMP. A triple R47L/F87V/L188Q mutant showed the highest kcat values toward S-OMP (97 min1) and R-OMP (103 min1) among the tested mutants. Although the binding of R-OMP to the CYP102A1 active site caused a more apparent change of heme environment compared to binding of S-OMP, there was no correlation between the spectral change upon substrate binding and catalytic activity of either enantiomer. The 5’-OH OMP produced from racemic, S-, and R-OMP could be obtained with a high conversion rate and high selectivity when the triple R47L/F87V/L188Q mutant was used. These results suggest that bacterial CYP102A1 mutants can be used to produce the human metabolite 5’-OH OMP from both the S- and R-enantiomers of OMP.

References Butler CF, Peet C, Mason AE, et al. (2013). Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias. J Biol Chem 288:25387–25399. Butler CF, Peet C, McLean KJ, et al. (2014). Human P450-like oxidation of diverse proton pump inhibitor drugs by ‘‘gatekeeper’’ mutants of flavocytochrome P450 BM3. Biochem J 460:247–259. Kang JY, Ryu SH, Park SH, et al. (2014). Chimeric cytochromes P450 engineered by domain swapping and random mutagenesis for producing human metabolites of drugs. Biotechnol Bioeng 111:1313–1322. Yamazaki H, Inoue K, Shaw PM, et al. (1997). Different contributions of cytochrome P450 2C19 and 3A4 in the oxidation of omeprazole by human liver microsomes: Effects of contents of these two forms in individual human samples. J Pharmacol Exp Ther 283:434–442. Yun CH, Kim KH, Kim DH, et al. (2007). The bacterial P450 BM3: a prototype for a biocatalyst with human P450 activities. Trends Biotechnol 25:289–298.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

73

P71. WARFARIN METABOLISM MEDIATED BY POLYMORPHIC P450 2C19 IN CYNOMOLGUS MONKEYS Hiroshi Yamazaki1, Yasuhiro Uno2, Mitsunori Shukuya1, Takahiro Yoshikawa2, Akinori Matsushita2, Norie Murayama1, Koichiro Fukuzaki3, and Masahiro Utoh2 1 Showa Pharmaceutical University, Machida, Japan, 2SNBL Ltd, Kainan, Japan and 3SNBL USA Ltd, Everett, WA, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Cynomolgus monkeys are widely used as primate models in preclinical studies, due to their close evolutionary relationship to humans. Monkey cytochrome P450 2C19 (formerly known as P450 2C75), highly homologous to human P450 2C19, has been identified in cynomolgus and rhesus monkeys, non-human primate species widely used in drug metabolism studies. In the present study, sequencing of P450 2C19 in 78 cynomolgus and 36 rhesus monkeys identified 34 non-synonymous variants. Among these, 6 were located in substrate recognition sites important for protein function. Three variants (p.Phe100Asn, p.Ala103Val, and p.Ile112Leu) showed substantially reduced activity as compared with wild type in R- and S-warfarin 7hydroxylation. Kinetic analysis for R-warfarin 7-hydroxylation and P450 docking simulation indicated that P450 2C19 Ala103Val would change the function and conformation of this enzyme. Ala103Val variation diminished homotropic cooperativity of P450 2C19 with R-warfarin yielding low metabolic capacity. When warfarin was orally administered to fasted male cynomolgus monkeys (n = 11, from Indochina, 4–8 years of age, 3.5–7.4 kg of body weight) at 1 mg/kg, the P450 2C19 genotypes were associated with the in-vivo pharmacokinetics of warfarin and its metabolites in these monkeys. These results indicated that the interindividual variability of P450 2C-dependent warfarin metabolism is at least partly accounted for by P450 2C19 variants in cynomolgus monkeys.

P72. IDENTIFICATION OF HUMAN CYTOCHROME P450 MOLECULAR SPECIES INVOLVED IN THE METABOLISM OF GEISSOSCHIZINE METHYL ETHER Takashi Matsumoto1, Hirotaka Kushida1, Junko Watanabe1, Hiroaki Nishimura2, Takeshi Maruyama2, Kazuya Maemura1, and Yoshio Kase1 1 Tsumura Research Laboratories, Kampo Scientific Strategies Division, Tsumura & Co., Inashiki-gun, Japan and 2Kampo Scientific Strategies Division, Kampo Formulations Development CenterTsumura & Co., Inashiki-gun, Japan Purpose: Geissoschizine methyl ether (GM) is an active component of the traditional Japanese medicine Yokukansan, which is sometimes prescribed for the neurosis, insomnia, and behavioral and psychological symptoms of dementia. GM activity, such as a partial agonist action on 5-HT1A, reportedly contributes to the mechanism underlying the action of Yokukansan. We previously examined the metabolic profile of GM in human liver microsomes and confirmed the formation of hydroxylated, de-methylated, dehydrogenated, methylated, and water-adduct metabolites. Given that hirsuteine, which is also present in Yokukansan and structurally similar to GM, can be metabolized by cytochrome P450 (CYP), GM may also be metabolized by CYP. If indeed this is the case, it is critical to identify CYP molecular species to avoid their interaction with CYP-inducing or CYP-inhibiting drugs. Therefore, we attempted to determine whether GM was metabolized by CYP in humans and the type of CYP molecular species that might be involved in the metabolism. Methods: GM and human liver S9 or CYP-expressing microsomes (CYP1A1, 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, and 3A4) were allowed to react at 37  C for 60 or 20 min, respectively. The reaction mixture was extracted with ethyl acetate and then measured by LC-MS/MS. For the CYP-specific antibody test, we added CYP1A1, 2C9, 2C19, 2D6, and 3A4 antibodies to the reaction system of human liver microsomes and GM. Results and discussion: GM-metabolizing reactions with human liver S9 were markedly inhibited with the addition of 1-aminobenzotriazole, a non-specific inhibitor of CYP. Experiments in a CYP-expressing microsome system indicated that CYP molecular species involved in the production of hydroxylated metabolites were 1A1, 2C19, 2D6, and 3A4, de-methylated metabolites were 2C9, 2C19, and 2D6, dehydrogenated metabolites were 1A1, 2C9, 2C19, and 3A4, methylated metabolites were 1A1 and 2D6, and water-adduct metabolites were 2C9, 2C19, 2D6, and 3A4. Of the CYP-specific antibodies tested, antibodies for CYP2C9, 2C19, 2D6, and 3A4 significantly inhibited the GM-metabolizing reactions in human liver microsomes. Conclusions: The present study demonstrated that GM is metabolized primarily by CYP. Furthermore, the responsible molecular species were suggested to be CYP2C9, 2C19, 2D6, and 3A4. Our findings will help to avoid interactions between drugs that induce or inhibit these enzymes when prescribing Yokukansan containing GM as an active component.

P73. OXIDATIVE METABOLISM OF 1-METHYL-4-PHENYL-1,2,3,6-TETRAHYDROPYRIDINE (MPTP) BY LIVER AND BRAIN MICROSOMES FROM COMMON MARMOSETS Shotaro Uehara1, Makiko Shimizu1, Norie Murayama1, Takashi Inoue2, Erika Sasaki2, and Hiroshi Yamazaki1 1 Showa Pharmaceutical University, Machida, Japan and 2Central Institute for Experimental Animals, Kawasaki, Japan The common marmoset (Callithrix jacchus) is a non-human primate that could prove useful as human pharmacokinetic and biomedical research models. 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), a potential neurotoxin, induces the degeneration of dopaminergic neurons and cause parkinsonism-like syndrome in common marmosets, other primates, and humans. MPTP-treated common marmoset was established as a Parkinson’s disease model to evaluated in preclinical studies. However, MPTP metabolism in common marmosets has not been fully investigated. The aim of this investigation was to

74

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

elucidate the oxidative metabolism of MPTP in common marmoset liver and brain. MPTP was metabolized to toxic metabolite, 1-methyl-4-phenylpyridinium ion (MPP+) via an intermediate 1-methyl-4-phenyl- 2,3-dihydropyridinium ion (MPDP+) in brain microsomes from common marmosets. MPTP was also metabolized to 4-phenyl-1,2,3,6-tetrahydropyridine (PTP) and MPTP N-oxide by liver from common marmosets. Recombinant marmoset cytochrome P450 enzymes and flavin containing monooxygenases supported these reactions. To examine the individual variability of MPTP metabolism, formation activities of MPTP to MPP+, MPDP+, and PTP was measured in liver and brain microsomes of 11 common marmosets in the presence of an NADHP-generating system. The individual variations for MPP+, MPDP+, and PTP formation activities in liver and brain microsomes were 3- to 5- and 2- to 5-fold, respectively. In marmoset liver microsomes, PTP formation activities significantly correlated with bufuralol 1’-hydroxylation activities (a P450 2D probe reaction) and MPP+ formation activities. These results suggested that marmoset P450 2D enzyme(s) contributed to oxidative metabolism of MPTP to PTP in common marmosets. The present finding should be provide a foundation for mechanistic studies of MPTP in common marmosets and successful production of Parkinson’s disease model in common marmosets.

P74. GENERATION AND EVALUATION OF A RELATIVE ACTIVITY FACTOR FOR CYP3A5 THROUGH MECHANISMBASED INACTIVATION OF CYP3A4 BY CYP3CIDE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Jeffrey M. Weber, Barbara J. Ring, and J. Matthew Hutzler Department of In vitro Metabolism, Quintiles, Indianapolis, IN, USA Relative activity factor (RAF) approaches are commonly used for extrapolating in vitro results from recombinantly expressed cytochrome P450 (rCYP) enzymes to activity observed in human liver microsomes (HLM). The fraction metabolized by each P450 (fm, CYP) may subsequently be estimated. The ability to estimate the relative contributions of CYP3A4 and CYP3A5 to total CYP3A metabolism was extremely challenging, thus historical RAF values represented total CYP3A activity. However, recent literature shows that CYP3A4 can be selectively inhibited using CYP3cide (Walsky et al., 2012), thereby allowing CYP3A5 activity to be characterized. To generate distinct RAFs for CYP3A4 and CYP3A5, determination of linear rate kinetic parameters for enzymatic activity (midazolam 10 -hydroxylation) in rCYP3A4 and rCYP3A5, as well as a 150-donor pooled lot of HLM (BD UltraPool 150) with/without CYP3cide treatment was determined. Based on CLint values determined in HLMs, CYP3A5 only contributes 7% of total midazolam 10 -hydroxylation, agreeing with vendor data demonstrating that CYP3A5 represents 12% of total CYP3A (Western Blot). Using CLint data obtained from HLMs and rCYP3A4 and 3A5, a RAF for CYP3A4 was estimated to be 33.5, whereas the RAF for CYP3A5 was 1.18. Matrix

Km (mM)

Vmax (pmol/min/mg or pmol P450)

CLint (mL/min/mg or pmol P450)

RAF (pmol/mg)

HLM (3A4 only) HLM + CYP3cide (3A5 only) rCYP3A4 SupersomeTM rCYP3A5 SupersomeTM

1.90 4.50 1.41 1.10

1177 191 24.3 39.6

577* 42.4 17.2 36.0

33.5 1.18 N/A N/A

*HLM CLint adjusted by CYP3A5 HLM CLint (HLM CLint – (HLM + CYP3cide CLint)) = CYP3A4 HLM CLint. RAF = HLM CLint/rCYP CLint. To confirm the inability of the commonly-used UltraPool 150 HLM matrix for differentiating CYP3A4 from CYP3A5 contribution to total CYP3A metabolism, three probes with varying contributions to metabolism by CYP3A4 and CYP3A5 were incubated in the various in vitro systems. The fraction of CYP3A4 and CYP3A5 attributed to CYP3A metabolism was determined.

Test Article

Literature: Fraction CYP3A5 of total CYP3A in CYP3A5*1/*1 donor (Tseng et alDMD 42: 2014)

rCYP3A4CLint (mL/min/pmol)

rCYP3A5CLint (mL/min/pmol)

Atazanavir Tacrolimus Diltiazem

40.89 0.45 0.18

7.46 12.3 3.99

5.45 21.6 1.25

rCYP3A4 Scaled CLint(mL/min/mg)

rCYP3A5 Scaled CLint(mL/min/mg)

Fraction CYP3A4: CYP3A using RAF

Fraction CYP3A5: CYP3A using RAF

250 411 134

6.44 25.5 1.47

0.97 0.94 0.99

0.03 0.06 0.01

Scaled CLint = CLint  RAF. Despite a known CYP3A5 genotype affect on the pharmacokinetic profile of atazanavir and tacrolimus (40–70% increase in apparent oral clearance, CL/F, for CYP3A5 expressers), this RAF method failed to predict a high contribution of CYP3A5. This finding calls into question the utility of large-donor HLM pools in predicting CYP3A5 liabilities. Additional studies using CYP3A5-genotyped microsomes will be conducted in order to more thoroughly assess the utility of use of this population-based RAF approach estimated from large-donor HLM pools, to predict the contributions of CYP3A4 and 3A5 to total CYP3A metabolism.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

75

P75. TIZANIDINE IMIDAZOLE METABOLITE AS A MARKER OF CANINE P450 1A2 ACTIVITY Chuck Locuson and J. Scott Daniels Vanderbilt University Medical Center, Nashville, TN, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

The canine P450 1A2 enzyme metabolizes veterinary drugs and preclinical candidate drugs being developed as human therapeutics. Hepatic canine 1A2 expression is equal to or greater than the human 1A2 isoform, plus the enzyme is inducible and polymorphic, creating the potential for large inter-subject differences in pharmacokinetic properties (up to 17-fold). Hence, the ability to probe canine 1A2 activity and inhibition is highly relevant towards pharmacokinetic scaling, toxicokinetics and toxicology findings, and other studies used to support drug development. As previous studies have relied on human 1A2 probe compounds with questionable specificity for dog 1A2, it was hypothesized that recombinant canine 1A2 could be used to find a specific 1A2 substrate that could be used as an in vitro and in vivo probe. Intrinsic clearance experiments demonstrated that phenacetin was a better substrate for P450 1A1 than for 1A2, partially explaining why phenacetin does not serve as an indicator of 1A2 activity (Whiterock et al., 2012). Conversely, tizanidine was a better substrate for canine 1A2 than for 1A1. Profiling the metabolites of 1A2, 1A1, and liver microsomes demonstrated that the tizanidine imidazole metabolite was a significant primary metabolite of all three enzyme systems. Dog plasma concentrations of the imidazole metabolite were significantly decreased (45% mean reduction; n = 3) when tizanidine was coadministered with the P450 1 A inhibitor enrofloxacin versus administration of tizanidine alone. Considering that enrofloxacin is a moderate inhibitor (Ki = 4 uM) and changes in tizanidine metabolism were detected at a low [inhibitor]/Ki ratio suggests that tizanidine might be a useful canine 1A2 probe.

Reference Whiterock VJ, Morgan DG, Lentz KA, et al. (2012). Phenacetin pharmacokinetics in CYP1A2-deficient beagle dogs. Drug Metab Dispos 40:228–231.

P76. EFFECT OF BAN-LAN-GEN ON TRANSACTIVATION OF THE CYP1A1 AND CYP1A2 GENES Takeshi Kumagai, Takayuki Kiyomatsu, Takamitsu Sasaki, and Kiyoshi Nagata Department of Environmental Health Science, Tohoku Pharmaceutical University, Sendai, Japan Purpose: Ban-Lan-Gen is a common name for the dried roots of indigo plants, which include Polygonum tinctorium, Isatis indigotica, Isatis tinctoria, and Strobilanthescusia. Ban-Lan-Gen is used as an anti-inflammatory and anti-viral medicine for hepatitis, influenza, and various kinds of inflammation. Major compounds in Ban-Lan-Gen include indigo, indirubin (IND) and trypthanthrin (TPT), and Ban-Lan-Gen is known to cause transactivation of cytochrome P450 (CYP) genes CYP1A1 andCYP1A2 via the aryl hydrocarbon receptor (AhR). In this study, we investigated the effects of Ban-Lan-Gen on transcriptional activation of the CYP1A1 and CYP1A2genes. Methods: Reporter assays were performed with HepG2-derived cells that stably expressed the 23-kb region involved in the transcriptional activation of theCYP1A1 and CYP1A2 genes (5-1 dual cells). CYP1A1 and CYP1A2 mRNA levels were determined by real-time RT-PCR. Results: Ban-Lan-Gen increased the transactivation of the CYP1A1 and CYP1A2 genes in a dose-dependent manner. IND and TPT, which are the primary constituents of Ban-Lan-Gen, increased CYP1A1and CYP1A2 gene reporter activities. In addition, overexpression of the human pregnane X receptor (PXR) increased the CYP1A1 gene reporter activity, and introduction of the human PXR-short hairpin RNA (hPXR-shRNA) inhibited the CYP1A1 gene reporter activity induced by treatment with IND or TPT. Furthermore, activation of the CYP1A1 reporter gene by TPT significantly enhanced upon treatment with rifampicin (RIF), a typical PXR ligand, whereas RIF had no effect on activation by IND. RIF treatment also significantly enhanced the activation of the CYP1A1 reporter gene by 2,3,7,8-tetrachlorodibenzo-pdioxin and benzo[a]pyrene, typical ligands of AhR. Conclusion: These results suggest that induction of CYP1A1 and CYP1A2 expression by Ban-Lan-Gen is primarily dependent on AhR activation, and enhanced through PXR activation by IND and TPT.

P77. EFFECTS OF CYP2D6 AND CYP3A5 GENOTYPES AND CO-ADMINISTRATED DRUGS ON THE METABOLIC CLEARANCE OF ANTIDEPRESSANT MIRTAZAPINE Maho Okubo1, Norie Murayama1, Jun Miura2, Yasuji Chiba3, and Hiroshi Yamazaki1 1 Showa Pharmaceutical University, Machida, Japan, 2Muroran Institute of Technology, Muroran, Japan, and 3Sanai Hospital, Noboribetsu, Japan Mirtazapine is a noradrenergic and specific serotonergic antidepressant used in major depression and other psychiatric disorders. It is important to develop a treatment method for each patient because mirtazapine has high frequencies of side effects including characteristic drowsiness. Mirtazapine is mainly metabolized by cytochrome P450 (CYP) 1A2, 2D6, and 3A4 to 8-hydroxymirtazapine and N-demethylmirtazapine. Therefore, plasma concentrations of mirtazapine could be influenced by its oxidation metabolism. The aim of this study was to investigate the factors for individual variations of metabolic clearance of mirtazapine in humans. Mirtazapine oxidation activities of pooled and individual human liver microsomes and recombinantly expressed P450 enzymes were examined in vitro analysis. Mirtazapine 8-hydroxylation activities of human liver microsomes were suppressed by 5 mM quinidine (a CYP2D6 inhibitor), and N-demethylation activities were by 5 mM ketoconazole (a CYP3A

76

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

inhibitor). Mirtazapine oxidation activities by CYP1A2, 2D6, 3A4, and 3A5 were confirmed. Kinetic parameters for mirtazapine N-demethylation by recombinant CYP3A5 were close to those by CYP3A4, and their Vmax values were higher than the other P450 enzymes. On the other hand, the Kmvalue for CYP2D6-mediated for mirtazapine 8-hydroxylation was low. Significantly correlations were confirmed between 8-hydroxylation and bufralol 10 -hydroxylation (CYP2D6-dependent) activities, and Ndemethylation and testosterone 6b-hydroxylation (CYP3A-dependent) activities. Metabolic clearance of mirtazapine and mirtazapine 8-hydroxylation activities in individual liver microsomes were significantly lower in the CYP2D6 intermediate metabolizers (IM) and poor metabolizers (PM) groups than extensive metabolizers (EM) group (*p50.05). Trough plasma concentration/dose (C/D) ratios of mirtazapine from 14 patients were significantly higher in the CYP2D6 IM group than EM group, and were also higher in the CYP3A5 poor-expressors group than the expressors group (*p50.05). These results provide that mirtazapine metabolic clearance could be influenced by the CYP2D6 and CYP3A5 genotypes and co-administrated drugs in clinical patients.

P78. ETHAMBUTOL INDUCES CYP2E1: COMPUTATIONAL MODELING AND IN VIVO INVESTIGATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Ganna M. Shayakhmetova1, Alla K. Voronina1, Svitlana I. Anisimova1, Valentina M. Kovalenko1, and Volodymir Kitam2 1 Department of General Toxicology, Institute of Pharmacology and Toxicology, National Academy of Medical Sciences of Ukraine, Kyiv, Ukraine and 2Department of Molecular Oncology, Institute of Molecular Biology and Genetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine The necessity to minimize adverse effects of tuberculosis chemotherapy requires a comprehensive evaluation of the effects of anti-tuberculosis drugs (ATD) on the organism. There are a few reports relative to hepatotoxicity and male gonadotoxicity of ethambutol ((2 S,20 S)-2,20 -(ethane-1,2-diyldiimino)dibutan-1-ol). Ethambutol (EMB) is used mainly in combination with other anti-TB drugs, which can induce or inhibit CYP2E1. It was reasonable to investigate the effect of E on CYP2E1. Aim of study: computer simulation of E and CYP2E1 interaction, and investigation of CYP2E1 induction/inhibition indices in rats’ liver and testes following E administration. Optimized by us, computer model of human CYP2E1 spatial structure was used. For modeling of EMB and CYP2E1 interaction docking of this ligand to enzyme active center was carried out. Spatial structure of E was taken from PubChem data base (CID 14052) and optimized using ArgusLab Software. The same software using Lamarkian Genetic Algorithm was applied for docking. E was administered to Wistar male rats in therapeutic dose (2 month). CYP2E1 mRNA expression in liver and testes was evaluated by RT-PCR. p-Nitrophenol hydroxylase activity. a selective enzyme marker for CYP2E1 was determined in microsomal fraction of liver and testes according to the method of Koop et al. Computer simulation showed that EMB can directly interact with CYP2E1 active site and heme. Such location is typical for CYP2E1 substrates, but ligand interacts with heme by his hydroxyl group, i.e. it could not be metabolized. Enzyme-EMB binding energy approaches to protein-substrate interaction level. Aforementioned allow supposing that EMB could be a weak CYP2E1 competitive inhibitor. But, following EMB administration to male rats we have demonstrated that CYP2E1 mRNA contents in their liver and testes increased by 4.4 and 8.7 times respectively. Also p–nitrophenol hydroxylase activity in liver and testes of EMB-treated animals rose by 4 times and 3 times as compared with control. Thus, results of in vivo experiments evidenced CYP2E1 induction in these organs. Data suggest that EMB can interact with CYP2E1 active site. EMB ability to cause increase of hepatic and testicular ˚ 1 mRNA level, as well as p-nitrophenol hydroxylase activity intensification, possibly could be realized by different CYP2A mechanisms. This phenomenon needs to be profoundly investigated. At least we could assume a mechanism for increasing of CYP2E1 activity to be similar to such for ethanol (enzyme molecule post-transcriptional stabilization). Results suggest the capability of EMB to participate in drug–drug interaction and metabolic idiosyncratic reaction during pharmacotherapy of tuberculosis.

P79. EFFECT OF SINGLE-WALLED CARBON NANOTUBES ON EXPRESSION OF PHASE I DRUG-METABOLIZING ENZYMES IN HUMANS Miki Katoh1, Yukinori Kuwabara1, Chiaki Kato1, Kotaro Hitoshi1, Yoshinori Ando2, and Masayuki Nadai1 1 Faculty of Pharmacy, Meijo University, Nagoya, Japan and 2Faculty of Science and Technology, Meijo University, Nagoya, Japan Carbon nanotube is a cylindrical molecule consisting of hexagonally arranged carbon atoms (graphene sheet) and is currently a promising key nanotechnology material. Among the several types of carbon nanotubes available, the single-walled carbon nanotube (SWCNT) in particular may be used in a wide variety of applications. Recently, various applications of SWCNTs in the medical field as drug and gene delivery agents have been proposed. Drug metabolism is one of the key factors that influence the pharmaceutical and toxicological effects of a drug. The purpose of this study was to clarify the effect of SWCNTs on phase I drug-metabolizing enzymes in humans. Primary cultures of human hepatocytes obtained from BD Gentest were exposed to 100 mg/mL SWCNTs for 24 h, and then a polymerase chain reaction array was conducted to examine 84 phase I drugmetabolizing enzyme genes. After exposure to SWCNTs, the expression levels of 14 genes were found to be significantly downregulated, by more than 50%. Cytochrome P450 (CYP) 1A1 and aldehyde dehydrogenase 1A1 mRNA expression decreased by 87% and 80%, respectively, compared to the corresponding levels in the control. Expression levels of two genes were significantly upregulated, by more than two-fold. Of the major CYPs, CYP1A2, CYP2C9, CYP2C19, CYP2D6, and

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

77

CYP3A4 mRNA expression decreased to 55%, 35%, 54%, 82%, and 83%, respectively. Therefore, the effects of SWCNTs differed between drug-metabolizing enzymes. In addition, changes in CYP1A2 mRNA expression were investigated after HepG2 cells were exposed to SWCNTs for 24 h. CYP1A2 mRNA expression decreased in response to exposure to 100 and 500 mg/mL SWCNTs. To clarify the mechanism of CYP1A2 suppression by SWCNTs, mRNA expression of the arylhydrocarbon receptor (which is one of the nuclear receptors for regulation of CYP1A2 expression), CYP1A1/1A2 reporter activity, and methylation status of the GC box in CYP1A2 were determined after SWCNT exposure in HepG2 cells. The mRNA expression of the arylhydrocarbon receptor was unchanged. CYP1A1/1A2 reporter activity was found to decrease in response to exposure to 100 and 500 mg/mL SWCNTs upon using a CYP1A1/2 Induction STEP Reporter Assay Kit. Bisulfite sequencing showed the methylation status of the GC box in CYP1A2 to be changed after exposure to SWCNTs. Therefore, it is suggested that reduction of CYP1A2 mRNA expression may be caused by a decrease in its reporter activity and a change in its DNA methylation. In conclusion, SWCNTs altered the expression of drug-metabolizing enzymes and may affect drug metabolism in human liver.

P80. INTEGRATING HEPATOCYTE AND IMMUNE SYSTEM MODELS TO EXPLORE MECHANISTIC EXPLANATIONSOF IMMUNE-MEDIATED P450 DOWN-REGULATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Brenden K. Petersen1 and C. Anthony Hunt2 1 UCSF/UCB Joint Graduate Group in Bioengineering, University of California, Berkeley, Berkeley, CA, USA and 2Department of Bioengineering and Therapeutic Sciences, University of California, San Francisco, CA, USA In silico modeling and simulation (M&S) methods are viewed as a promising countermeasure to the recent decline in pharmaceutical R&D productivity. Improved insight into drug clearance mechanisms in various contexts can improve decisions as the R&D process progresses. We previously developed and validated an in silico hepatocyte culture (ISHC) model to test mechanistic hypotheses about drug-induced liver toxicity. The ISHC is object-oriented, multi-scale, multi-attribute, discreteevent, and agent-based. The model structure maps to a monolayer culture of isolated rat hepatocytes. The model responds to an initial dose of chemical entities (e.g. xenobiotic) using various concrete mechanisms. These mechanisms serve as mechanistic hypotheses that map to cellular processes, including xenobiotic metabolism, inflammation, and enzyme regulation. In silico outputs mimic wet-lab experimental data, including dose-response curves and measures of in vitro intrinsic clearance. Integrating the ISHC with an immune system model allows expanding the set of targeted attributes to include immune cell interaction with liver parenchyma and the downstream effects on intrinsic clearance and drug hepatotoxicity. Specifically, the targeted attribute is the ability of Kupffer cells to down-regulate P450 drug-metabolizing enzymes via cytokine mediators in response to an inflammatory stimulus. In this case, the inflammatory stimulus is lipopolysaccharide (LPS); in the future, it may be a toxic drug metabolite. The validation data include: (1) a dose-response curve between inflammatory stimulus dose and cytokine response after a fixed period of Kupffer cell incubation, (2) time-course reduction in hepatic P450 levels after an inflammatory stimulus, and (3) fold-changes in in vitro intrinsic clearance before/after an inflammatory stimulus in a hepatocyteKupffer cell co-culture. Validation data #1 and #2 were obtained from wet-lab experiments using Kupffer cell cultures and hepatocyte cultures, respectively. We facilitate simulating these experiments by using a modular modeling framework capable of flexible changes in model structure. Validation data #3 was obtained using a hepatocyte-Kupffer cell co-culture; thus, simulating this experiment requires integration of hepatocyte and Kupffer cell model components and structure. Validation is achieved when in silico results fall within an acceptable, pre-specified measure of quantitative similarity. Often, a posited mechanism fails to achieve validation and is thus falsified; it is then revised via an iterative refinement protocol. The multiple, falsified mechanisms encountered during iterative refinement cycles provide insight by reducing the set of plausible mechanistic explanations for immune-mediated P450 down-regulation. An ISHC that achieves validation targets will eventually be falsified with newly added validation targets. In parallel, the same modularized mechanisms can be used within a whole-liver model – the in silico liver (ISL) – to explore the mechanisms space of P450 down-regulation in vivo.

P81. EFFECT OF CYP3A DEFICIENCY ON ANDROGEN ACTION IN THE MOUSE PROSTATE Mana Yamazaki1, Kaoru Kobayashi1, Mari Hashimoto1, Yasuhiro Kazuki2, Shoko Takehara2, Mitsuo Oshimura2, and Kan Chiba1 1 Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan and 2 Chromosome Engineering Research Center, Tottori University, Yonago, Japan Testosterone is converted into dihydrotestosterone (DHT), an active metabolite of testosterone, by steroid 5a-reductase (SRD5A) in the prostate. Testosterone and DHT activate androgen receptor (AR) as ligands and stimulate the development of the normal prostate tissue and the growth of prostate cancer cells. On the other hand, testosterone is inactivated to 6b-hydroxytestosterone by cytochrome P450 3A (CYP3A), suggesting that CYP3A may play some role in the regulation of the circulating testosterone levels and androgen actions in the prostate. Moreover, AR transcriptionally activates sterol regulatory element binding protein (SREBP) cleavage activating protein (SCAP), which stimulates the processing of SREBP-2. Since SREBP-2 increases the expression of Srd5a2 in the prostate, AR may be activated by increased DHT levels through the promotion of SRD5A2 activity by SREBP-2. In this study, we examined the effect of Cyp3a deficiency on prostatic androgen actions and activation of AR via

78

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

increased SRD5A2 activity by using male Cyp3a-knockout (Cyp3a/) mice aged between 10 and 11 weeks. 6b-Hydroxylation activities of testosterone in the liver microsomes of Cyp3a/ mice were lower than those of WT mice. Hepatic and plasma testosterone levels of Cyp3a/ mice were higher than those of WT mice. The mRNA expression levels of AR target genes, including Scap gene, in the prostates of Cyp3a/ mice were higher than those of WT mice. In addition, binding of AR to androgen response element located in a promoter region of spermine-binding protein gene, an AR target gene, was more abundant in the prostates of Cyp3a/ mice. Furthermore, protein levels of prostate stem cell antigen, a maker of the androgen actions, in the prostates of Cyp3a/ mice were higher than those of WT mice. These results suggest that suppressed 6b-hydroxylation activities of testosterone in Cyp3a/ mice consequently activated androgen actions in the prostates via increased plasma testosterone levels. In addition, the mRNA expression levels of Srd5a2 and some SREBP-2 target genes in the prostates of Cyp3a/ mice were higher than those of WT mice. Thus, increased mRNA expression levels of Scap gene in the prostates of Cyp3a/ mice might promote SRD5A2 expression through enhancing SREBP-2 activity. In conclusion, it was suggested that deficiency of Cyp3a increases the plasma testosterone levels by repressing the inactivation of testosterone to 6b-hydroxytestosterone, and then activates androgen actions in the prostate via AR activation. Furthermore, activated AR may up-regulate the expression of SCAP which activates SREBP-2, resulting in the further AR activation through the promotion of SRD5A2.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P82. ENDOGENOUS AND DRUG SUBSTRATE BINDING TO CYP2J2 IN LIPOSOMES AND LIPID BILAYER NANODISCS Eric A. Evangelista, Wynton D. McClary, William M. Atkins, and Rheem A. Totah Department of Medicinal Chemistry, University of Washington, Seattle, WA, USA Cytochrome P450 2J2 (CYP2J2) is the only isoform in the human 2J subfamily. It is expressed mostly in extrahepatic tissues including the intestine, brain and especially the heart where it is the most abundant isozyme. CYP2J2 can metabolize exogenous compounds and is proposed to be responsible for 70% of ebastine first pass metabolism. In cardiac tissue, CYP2J2 is known to oxidize endogenous arachidonic acid (AA) into cis-epoxyeicosatrienoic acids (EETs), which have established cardioprotective properties and any reduction in EETs formation can lead to cardiotoxicity. CYP2J2 expression is carefully regulated in the heart and expression is not affected by typical CYP inducing agents such as rifampicin and phenobarbital. Furthermore, in vitro CYP2J2 displays atypical kinetics and substrate dependent inhibition. For example, grepafloxacin efficiently inhibits AA but has no effect on the oxidation of probe substrate terfenadine, danazol inhibits metabolism of both substrates, while lansoprazole inhibits terfenadine metabolism but does not affect AA metabolism. To understand the interaction of CYP2J2 with its exogenous and endogenous substrates, binding to full length and truncated recombinantly expressed CYP2J2 was investigated in liposomes and lipid bilayer nanodiscs. Data indicate that the presence of a lipid bilayer is necessary to observe ferric spin state changes. Further, binding affinity of terfenadine and astemizole to truncated CYP2J2 increase by as much as fifteen-fold when the enzyme is in lipid bilayer nanodiscs compared to liposomes. These results indicate a significant role for the enzyme’s micro environment for binding and, as a result, catalysis. Elucidating the different factors that affect binding and catalysis of CYP2J2’s substrates will help determine exogenous-endogenous substrate interaction and are a first step to provide better ways to predict druginduced cardiotoxicity.

P83. SIX YEARS OF EXPERIENCE WITH THE MIST GUIDANCE Frank Runge, Bojan Bister, Stefan Blech, and Thomas Ebner Drug Metabolism & Pharmacokinetics, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach an der Riss, Germany Statement of hypothesis: Discussions about approaches to address metabolites in safety testing (MIST) is available in the literature since 2002 and resulted in the release of formal guidance by FDA in 2008 and ICH in 2009. The emphasis of these guidances is placed upon obtaining comparative exposure data of circulating metabolites in human compared to animal studies as early as possible in the drug development process. However, there is no single comprehensive clinical trial that unequivocally allows the identification and determination of steady-state exposure of human metabolites. Therefore, a stepwise approach is chosen using in vitro and in vivo data at different stages of the drug development process with the aim to mitigate the risk of identifying disproportionate human metabolites (DHM) at late stages of drug development. It is triggered by the timing of the Boehringer Ingelheim development process and provides data for the risk assessment of DHMs. The first work package compares human in vitro data and rat ADME data. Work package two and three provide an in vivo assessment of relevant human metabolites and potential DHMs from single rising dose (SRD) and multiple rising dose (MRD) studies compared to animal toxicity studies. The fourth and last work package assembles all animal and human metabolism data including results of the human ADME study. Method: Metabolite structures were elucidated on either a LTQ Orbitrap XL mass spectrometer equipped with a Triversa NanomateTM nanospray ion source or a Synapt G2s mass spectrometer. Plasma, urine, feces and bile samples from rats, mice, dogs, minipigs and human plasma samples were prepared and analyzed during these studies. Typically urine and bile samples were centrifuged. Feces homogenates were extracted several times with optimized solvents. The extracts were combined and concentrated by lyophilization. Plasma samples were either processed by solid-phase extraction (SPE) or protein

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

79

precipitation. Supporting data: Data from 10 Boehringer Ingelheim drug candidates in development from the last 6 years were analyzed from in vitroexperiments (human hepatocytes and human liver microsomes), from animal ADME and toxicity studies, as well as from human single rising dose, multiple rising dose and human ADME studies. Data will be presented in detail regarding coverage of human metabolites in the respective animal samples and in vitro samples. Results and conclusion: A stepwise approach using data from different in vitro and in vivo experiments gives a good understanding of drug metabolism and mitigates the risk of identifying a DHM at late stages of drug development. Nevertheless, identification of all circulating human metabolites based only on LC-MS analytical approaches cannot be guaranteed prior the human ADME study.

P84. CHARACTERIZATION OF IN VITRO AND IN VIVO METABOLIC PATHWAYS IN THE COTTON RAT

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Christine L. Talarico, Feng Wang, Angela S. Mote, and Sonia R. Miranda Infectious Diseases TAU, GlaxoSmithKline, Research Triangle Park, NC, USA The cotton rat (S.hispidus) is susceptible to numerous human infectious pathogens, such as respiratory syncytial virus, influenza A and B, adenoviruses and parainfluenza virus. Therefore, it is an excellent model to use for discovering compounds effective against respiratory diseases. There is very limited information on the metabolic capabilities in the cotton rat; hence the main goal was to determine the hepatic cytochrome P450 metabolism. A set of CYP P450-substrates representing the major CYP subtypes were tested in cotton rat hepatic microsomes and hepatocytes. Metabolic rates were compared to those found in Hans Wistar and Sprague Dawley rats and CD1 mice microsomes and hepatocytes. In microsomes, differences in T1/2 values were observed across species for two substrates, diclofenac and atorvastatin, suggesting decreased metabolism. The microsomal metabolism of the remaining compounds was similar across all species. In hepatocytes, s-mephenytoin showed moderate metabolism in cotton rats, but was stable (T1/2490 min) in the other species. To determine if an in vitro/in vivo correlation exists, a known CYP3A4 substrate, midazolam, was orally administrated to cotton rat and HW rats. As a specific substrate for CYP3A4, midazolam is known to suffer extensive first pass metabolism in the rat, especially in liver. Our preliminary results indicated that midazolam exhibited similar oral exposure in cotton rats and HW rats. These results suggest some similarities in the P450 metabolic capabilities between the cotton rat and the Han Wistar rat compared to the other rodent species. More work is ongoing to characterize the Phase II metabolic capabilites in the cotton rat.

P85. COMPARISON OF PHARMACOKINETIC PARAMETERS OF CYTOCHROME P450 PROBE SUBSTRATES IN HUMANS, MICE, RATS, DOGS, MONKEYS, AND MICROMINIPIGS Chizuka Sakai1, Shunsuke Iwano1, Yuri Yamazaki1, Akihiro Ando1, Hiroshi Yamazaki2, and Yohei Miyamoto1 1 Toxicology and Pharmacokinetics Laboratories, Pharmaceutical Research Laboratories, Toray Industries, Inc., Kamakura, Japan and 2Laboratory of Drug Metabolism & Pharmacokinetics, Showa Pharmaceutical University, Machida, Japan Experimental animals such as mice, rats, dogs, monkeys, and minipigs are commonly used in the nonclinical studies of new drug candidates to predict human pharmacokinetics. Therefore, it is important to clarify species differences in cytochrome P450 (P450 or CYP) activities between humans and such experimental animals. In this study, we assessed multiple in vivo P450 activities in mice, rats, dogs, monkeys, and Microminipigs, registered as a novel variety of pigs, by simultaneous dosing with typical human P450 substrates such as caffeine (human CYP1A2 substrate), losartan (human CYP2C9 substrate), omeprazole (human CYP2C19 substrate), dextromethorphan (human CYP2D6 substrate), and midazolam (human CYP3A substrate) to these animals. In addition, the in vitro intrinsic clearance (CLint) of the 5 P450 substrates using liver microsomes from humans and the experimental animals were investigated. To determine in vivo and in vitro PK parameters, the concentrations of 5 P450 substrates were measured by LC/MS/MS. The high BAs and low CLint values of caffeine in the 5 experimental animals were similar to those in humans. Mice and monkeys showed lower BAs and higher CLint values of losartan than those in humans, rats, dogs, and Microminipigs. Mice, rats, and monkeys showed lower BAs and higher CLint values of omeprazole than humans, dogs, and Microminipigs. The lowest BAs of dextromethorphan in monkeys and Microminipigs were observed, and the CLint values of dextromethorphan in mice, rats, monkeys, and Microminipigs were higher than those in humans and dogs. The BA of midazolam in Microminipigs was similar to that in humans as compared with the other animals, but the CLint values in the 5 experimental animals were the relatively similar to that in humans. These results indicated that little first pass metabolism may occurs after oral administration of caffeine in all animals, and the lower BAs of losartan, omeprazole, and dextromethorphan in the experimental animals compared to humans were due to species differences in hepatic availability (Fh). On the other hands, the lower BAs of midazolam in mice, rats, dogs, and monkeys compared to humans and Microminipigs may depend on the fraction absorbed multiplied by intestinal availability (FaFg) other than the Fh. These results are useful in predicting pharmacokinetics in humans from in vitro and in vivo experimental data using these animals.

P86. ABSTRACT WITHDRAWN

80

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P87. IN VITRO TO IN VIVO PREDICTION OF CYP3A INDUCTION IN THE FIRST AND SECOND TRIMESTERS OF PREGNANCY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Zufei Zhang, Muhammad Farooq, Bhagwat Prasad, and Jashvant D. Unadkat Department of Pharmaceutics, University of Washington Seattle, Seattle, WA, USA Background: Cytochrome P450 3A (CYP3A) activity, as measured by midazolam oral clearance, is increased in pregnant women by 2 fold during the third trimester (T3) (Hebert MF, et al., 2008). This can result in subtherapeutic plasma concentrations of CYP3A cleared drugs (e.g. indinavir) requiring a change in dosing regimens of these drugs. For ethical and logistical reasons, the magnitude of change in CYP3A activity during the first trimester (T1) and second trimester (T2) has not been characterized. Hypothesis: The magnitude of induction in CYP3A activity during T1 and T2 of human pregnancy can be predicted using HepaRG cells. Methods: Differentiated HepaRG cells were plated in 96-well plates and maintained in HepaRG general purpose media for 72 h. Then, cells were incubated for 72 h with the unbound, total, 10X total T1, T2, or T3 plasma concentrations of cortisol (C), estradiol (E2), progesterone (P), growth hormone (GH), and placental growth hormone (PGH), either individually or in combination. Depletion of the hormones was accounted for by renewing the media every 12 h to maintain the average hormone concentrations in the media equivalent to that observed during T1, T2 or T3. At the end of the incubations, CYP3A activity was assessed by measuring 10 OH midazolam formation. CYP3A activity was expressed as fold-induction relative to that produced by HepaRG cells incubated with hormone concentrations observed in non-pregnant women. Results: Of all the hormones incubated individually, only C significantly induced (48-fold) CYP3A activity at 1X and 10X T1-T3 plasma concentrations. At T3 and T2 plasma concentrations, the combination, E2 & C or P & C, induced CYP3A activity greater than each hormone alone, indicating possible synergism between these hormones. The combination of all five hormones at 1X T1, T2 or T3 plasma concentrations resulted in fold-induction of CYP3A activity that was not significantly different for the three treatments: T1 16.2 ± 12.8, T2 24.5 ± 12.3, T3 16.7 ± 4.4. Conclusions: Consistent with our previous findings, cortisol was the major inducer of CYP3A activity in HepaRG cells (Papageorgiou et al., 2013). Based on these HepaRG cells data, we predict that the in vivo induction of CYP3A activity will be comparable across the three trimesters. Indeed, the in vivo clearance of indinavir, a drug cleared by CYP3A metabolism, is similar during the second and third trimester of pregnancy (Cressey et al., 2013). To determine if these results are replicated in human hepatocytes, corresponding experiments using sandwich cultured human hepatocytes are ongoing in our laboratory. Supported by P01 DA032507.

References Cressey TR, Best BM, Achalapong J, et al., (2013). Reduced indinavir exposure during pregnancy. BJCP 76:475–483. Hebert MF, Easterling TR, Kirby B, et al., (2008). Effects of pregnancy on CYP3A and P-glycoprotein activities as measured by disposition of midazolam and digoxin. CPT 84:248–253. Papageorgiou I, Grepper S, Unadkat JD. (2013). Induction of hepatic CYP3A enzymes by pregnancy-related hormones: Studies in human hepatocytes and hepatic cell lines. DMD 41:281–290.

P88. SPECIES AND TISSUE DIFFERENCES IN SEROTONIN GLUCURONIDATION Yukiko Sakakibara, Miki Katoh, Taisho Kawayanagi, and Masayuki Nadai Faculty of Pharmacy, Meijo University, Nagoya, Japan UDP-glucuronosyltransferase catalyzes major phase II reaction, which is involved in the metabolism of drugs as well as endogenous compounds. Glucuronidation is one of the metabolic pathways of an essential neurotransmitter, serotonin, and serotonin glucuronidation is catalyzed selectively by UGT1A6 in humans (Krishnaswamy et al., 2003). Serotonin is mainly found in the extrahepatic tissues, such as intestine, kidney and brain, where some UGT1As are expressed at a higher level than in the liver. It is important to clarify the serotonin kinetics in each tissue in order to understand its various pharmacological effects in the body. It is, therefore, of value to clarify the characterization of serotonin glucuronidation in the extrahepatic tissues. The aim of the present study was to characterize the serotonin glucuronidation in extrahepatic tissues of humans and rats and to compare the glucuronidation between species or tissues. Pooled microsomes from human intestine, liver or kidney were purchased from BD Gentest or KAC (Kyoto, Japan), respectively. Pooled microsomes from rat intestine, kidney, brain and liver were prepared from 8-week-old male Sprague-Dawley rats. Serotonin glucuronide was quantified by liquid chromatographytandem mass spectrometry. The kinetics of serotonin glucuronidation in human intestine was fitted to the Hill equation. Serotonin glucuronidation in human kidney and liver were fitted to the Michaelis–Menten model, which was consistent with the result of the previous report (Krishnaswamy et al., 2003). Serotonin glucuronidation exhibited different kinetics in the intestine and kidney or liver, indicating that it would be catalyzed not only by UGT1A6 but also by another UGT1A isoform, which has a different affinity to serotonin. Serotonin glucuronidation in the rat intestine, kidney and liver were well fitted to the Michaelis– Menten model and was exhibited monophasic kinetics in the kidney, whereas biphasic kinetics in the rat intestine and liver. It is suggested that serotonin glucuronidation was catalyzed by more than two common UGT isoforms in the rat intestine and liver. On the other hand, serotonin glucuronidation in the rat brain was fitted to the Hill equation, suggesting that the affinity of

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

81

UGT1A isoform which was responsible for serotonin glucuronidation in the rat brain was different from that in other tissues. In conclusion, we show that tissue and species differences exist in serotonin glucuronidation. It is possible that the kinetics of serotonin glucuronidation in the human brain may differ from that in the intestine, kidney and liver. It is necessary to take account of these potential differences when considering serotonin pharmacodynamics and pharmacokinetics.

Reference Krishnaswamy S, Duan SX, von Moltke, et al., (2003). Drug Metab Dispos 31:133–139.

P89. ASSESSMENT OF POTENTIAL FACTORS RELATED TO LOW CYP2C19 ACTIVITY IN LIVER CANCER PATIENTS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yuki Nishimura1, Norimitsu Kurata1, Mariko Iwase1, Rika Yasuhara2, Hajime Yasuhara1, Yasuna Kobayashi3, Toshinori Yamamoto3, Takeshi Aoki4, Masahiko Murakami4, Koji Saito5, Shinichi Kobayashi1, and Katsuji Oguchi1 1 Department of Pharmacology, School of Medicine, Showa University, Tokyo, Japan, 2Department of Oral Diagnostic Sciences, School of Dentistry, Showa University, Tokyo, Japan, 3Department of Pharmacotherapeutics of Clinical Pharmacy, School of Pharmacy, Showa University, Tokyo, Japan, 4Department of Surgery, School of Medicine, Showa University, Tokyo, Japan, and 5 Department of Pathology, School of Medicine, Showa University, Tokyo, Japan In our previous study, the remarkable low activity of cytochrome P450 2C19 (CYP2C19) was observed in liver cancer patients. Although the incidence of the CYP2C19 poor metabolizer is known to be approximately 20% in the Japanese population, the higher frequency of reduced activity was shown in the metastatic liver cancer status. Therefore, the current study was conducted to clarify the isozyme selectivity and the related factors to lower CYP activity in liver cancer morbidity. The study protocol was approved by Ethics Committee on Human Genome Analysis of Showa University. Liver and blood samples were donated from 21 Japanese liver cancer patients who received partial hepatectomy. The normal part which does not contain pathological cancer focus was used as liver samples. CYP2C19 genotype was determined using both of liver and blood samples by PCR-RFLP or the Taqman assay. The protein contents and the mRNA levels of this enzyme were analyzed by western blot analysis, and the Taqman real-time PCR method, respectively. CYP2C9, CYP2C19, CYP2D6 and CYP3A4 activities were assayed by HPLC method using the typical substrate for each isoform. The level of 27 kind of cytokines were determined by Bio-Plex suspension array system. The result showed that the CYP2C19 genotype analyzed by liver samples were identical with that in blood samples in all patients. No patients had CYP2C19*17 allele. Most of patients showed lower CYP2C19 activity compared to control that was estimated by pooled liver microsomes. Even in six patients with CYP2C19*1/*1, the activity was lower than 15% of the control value, suggested no association with the genotype. Furthermore, the reduced activity did not correlate CYP2C19 protein and mRNA levels. The hierarchal cluster analysis on 21 patients classified in 5 groups by the similarity of 27 cytokines. Between two groups of them, a significant difference of CYP2C19 activity was observed (p50.05, 12.09 ± 4.66 vs. 62.00 ± 38.7 pmol/mg protein/min). In addition, the levels of several cytokines such as IL-ra, FGF basic, VEGF, MIP-1a and G-CSF were different significantly between these groups. CYP2C9, CYP2D6 and CYP3A4 activities did not show any remarkable difference compared to the control activities. In conclusion, we found that the decline of metabolic activity in liver cancer status arises selectivity to CYP2C19 isozyme. The lower activity was not associated with its genotype, mRNA and protein level, but associated with certain cytokine level. Further studies are needed to elucidate the mechanisms and significance of the results obtained in this study, and to clarify the role of CYP2C19 in cancer patients.

P90. EVALUATION OF A PXR-CAR-CYP3A4/3A7-CYP2D6-CYP2C9 MULTIPLE HUMANIZED MOUSE MODEL TO PREDICT DRUG–DRUG INTERACTIONS IN MAN Nico Scheer1, Yury Kapelyukh2, Lesley A. McLaughlin3, Mike McMahon2, Anja Rode1, Colin J. Henderson4, and C. Roland Wolf4 1 ADMET, TaconicArtemis, Cologne, Germany, 2University of Dundee, Dundee, UK, 3Cancer Research UK Molecular Pharmacology Unit, Dundee, UK, and 4Cancer Research UK Molecular Pharmacology Unit, Biomedical Research Institute, Dundee, UK The use of genetically humanized mouse models to predict drug–drug interactions in man is often limited by the restricted number of proteins involved in drug metabolism and disposition which are humanized in the models described to date. This is particularly important in light of the complex interactions of many compounds with various drug metabolizing enzymes and the proteins regulating their expression. In order to overcome this limitation, we have generated a multiple humanized mouse model in which the murine pregnane X receptor (Pxr), constitutive androstane receptor (Car) and the cytochrome P450 (Cyp) 3a, Cyp2d and Cyp2c gene clusters have been replaced with human PXR, CAR, CYP3A4/3A7, CYP2D6 and CYP2C9. Here we provide a first characterization of the PXR-CAR-CYP3A4/3A7-CYP2D6-CYP2C9 humanized mouse model. We show that the basal hepatic expression of CYP2D6 and CYP2C9 in this model is close to the average expression in human liver. While the basal CYP3A4 expression in the liver is at the lower end of that in humans, it is strongly inducible with the human specific PXRactivator rifampicin. In vitro inhibition studies with the CYP3A4 inhibitor ketoconazole and the CYP3A4 substrate midazolam using liver microsomes from humanized PXR-CAR-CYP3A4/3A7 and PXR-CAR-CYP3A4/3A7-CYP2D6-CYP2C9 mice show significantly higher IC50 values in the former model compared to the latter. These findings suggest an involvement of other

82

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

enzymes, presumably from the mouse Cyp2c family, in the metabolism of midazolam in the humanized PXR-CAR-CYP3A4/ 3A7 model. Furthermore, liver microsomes from humanized PXR-CAR-CYP3A4/3A7-CYP2D6-CYP2C9 mice and humans showed markedly higher affinity and lower IC50 to the CYP3A4 mechanism-based inhibitor erythromycin compared to samples from WT mice using midazolam as a probe substrate. In summary, our data demonstrate the utility of the humanized PXR-CARCYP3A4/3A7-CYP2D6-CYP2C9 mice in studying CYP3A4-mediated drug–drug interactions. Particularly, the confounding contribution of mouse Cyp2c enzymes to midazolam metabolism can be overcome with these animals. In vivo DDI studies using these mice are currently on going and the results will be presented. We anticipate that the PXR-CAR-CYP3A4/3A7-CYP2D6CYP2C9 humanized mouse model will have significant value in predicting cytochrome P450-mediated drug pharmacokinetics in man.

P91. MECHANISMS UNDERLYING DRUG-INDUCED CHOLESTASIS IN PRIMARY HUMAN HEPATOCYTE CULTURES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Ce´line Parmentier1, Marlies Oorts2, Davide Germano3, Philippe Couttet3, Arno Lukas4, Philipp Hewitt5, Armin Wolf3, Pieter Annaert6, Marianne Uteng3, and Lysiane Richert1 1 KaLy-Cell, Strasbourg, France, 2Department of Drug Delivery & Disposition, KU Leuven, Leuven, Belgium, 3Novartis Institutes for Biomedical Research, Basel, Switzerland, 4Emergentec Biodevelopment GmbH, Vienna, Austria, 5Merck Serono, Darmastadt, Germany, and 6Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium An incidence of 1–2% of cholestasis has been reported in patients treated with therapeutic doses of Chlorpromazine (CPZ). Similarly, several cases of Ibuprofen (IBU) hepatotoxicity have been reported, including cholestasis. The most common abnormalities related to Cyclosporine A (CsA) hepatotoxicity are increases in bile salt levels, hyperbilirubinemia and cholestasis. In the present study, we assessed the mechanisms underlying the cholestatic effects of these three drugs in primary human hepatocyte (PHH) cultures with known and different CYPs expressions. Between-donor differences in CPZ exposure and biokinetics could not account for inter-donor variability in the transcriptomic cholestatic signature (Parmentier et al., 2013) of CPZ observed in PHH cultures. The inter-donor variability in the strength of this transcriptomic cholestatic signature in response to CPZ treatment at 0.1 and/or 1 mM (in the range of human cmax) was associated with variability in the strength of the transcriptomic inflammation signature. This, further confirms our recent observation that inflammation could be related to the idiosyncratic hepatotoxic effects of CPZ in humans (Parmentier et al., 2013). PHH cultures with a low CYP2C9 activity were highly sensitive to IBU cytotoxicity, with a high number of deregulated genes and a cholestastic transcriptomic signature at 100 mM, whereas PHH cultures with a high CYP2C9 activity were less sensitive to IBU cytotoxicity and exhibited a cholestatic signature only at 1000 mM. PHH cultures treated with the highest CYP3A4/5-dependent CsA metabolizing capacity showed the lowest number of deregulated genes and the signature of cholestatis was observed at 7 mM CsA only while in PHH cultures with lower CYP3A4/5 activities the cholestatic signature was evident at both 0.7 mM and 7 mM CsA. Both CsA and IBU induced an inhibition of the MRP2 transporter treatment in PHH cultures while CPZ did not. The large inter-donor variability in the response to CPZ and CsA was confirmed using the recently developed assay for Drug Induced Cholestasis Index (DICI)(Chatterjee et al. 2014) with cryopreserved plateable PHH cultures. Taken together, our data show that the use of PHHs can be considered a valuable. a valuable tool to study potential drug-induced cholestasis across individuals.

References Chatterjee S, Richert L, Augustijns P, Annaert P. (2014). Hepatocyte-based in vitro model for assessment of drug-induced cholestasis. Toxicol Appl Pharmacol 274:124–136. Parmentier C, Truisi GL, Moenks K, et al. (2013). Transcriptomic hepatotoxicity signature of chlorpromazine after short- and longterm exposure in primary human sandwich cultures. Drug Metab Dispos 41:1835–1842.

This work was supported by the Seventh Framework Programme – PredictIV, grant #202222 ‘‘Profiling the toxicity of new drugs: A non-animal-based approach integrating toxicodynamics and biokinetics’’, and by the EU-EFPIA Innovative Medicine Initiative (IMI) project #115336-2 ‘‘Mechanism-Based Integrated Systems for the Prediction of Drug-Induced Liver Injury MIP-DILI’’.

P92. THE VALUE OF SELECTED CRYOPRESERVED HUMAN HEPATOCYTES FOR THE STUDY OF THE ROLE OF BIOKINETIC IN LONG-TERM REPEATED TOXICITY STUDIES Emma Di Consiglio1, Giuliana Pomponio1, Ce´line Parmentier2, Jessica Broeders3, Bas J. Blaauboer3, Frederic Y. Bois4, Emanuela Testai1, and Lysiane Richert2 1 Istituto Superiore di Sanita`, Roma, Italy, 2KaLy-Cell, Strasbourg, France, 3Utrecht University Institute for Risk Assessment Science, Utrecht, the Netherlands, and 4DRC/Viva/Meto, INERIS, Verneuil en Halatte, France The role of biokinetics in long-term repeated toxicity assays with primary human hepatocytes (PHH) was studied using Ibuprofen (IBU), Amiodarone (AMI) and Chlorpromazine (CPZ) with distinct outcomes of the biotransformation processes: (i) detoxifying by CYP2C9; (ii) bioactivation by CYP3A4 and 2C8; (iii) mixed pathways by CYP1A2 and 2D6, respectively. Selected plateable cryopreserved PHH with various and known CYPs activities were submitted to daily dosing for 14 days: in the case of IBU no great differences with respect to a single exposure is evidenced. A dynamic equilibrium was reached practically

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

83

in one day, as revealed by the PBBK model applied, according to what is reported in human healthy volunteers after oral administration of standard IBU formulations. The opposite occurs with AMI. This drug enters hepatocytes very quickly, where it is biotransformed to its active metabolite mono-N-desethylamiodarone (MDEA). Over time an intracellular accumulation of the parent compound as well as MDEA, trapped inside the cells producing it, was noticed. In the case of repeated exposure a dynamic equilibrium was reached in about a week, reflecting the in vivo situation when a second dosage of the drug is administered before the previous one is completely eliminated. Higher cytotoxicity was exerted by IBU in PHH cultures with the lowest CYP2C9 activity. After repeated exposure a twenty fold different intracellular IBU was measured among PHHs, very well related with the high variable CYP2C9 activity. An inter-individual variability in AMI biokinetic and toxicity was also observed, most probably related to differences in drug metabolizing Phase 1 and 2 enzymes as well as transporter activities. In contrast, and as expected, no correlation between accumulation process and cytotoxicity was observed for CPZ. In conclusion, the use of various and well characterized PHHs can be considered a valuable tool to evidence possible differences in biokinetics that could explain inter-individual variability in drug bioavailability and consequently toxicity. This work was supported by the Seventh Framework Programme – PredictIV, grant # 202222 ‘‘Profiling the toxicity of new drugs: A non-animal-based approach integrating toxicodynamics and biokinetics’’, and by the EU-EFPIA Innovative Medicine Initiative (IMI) project #115336-2 ‘‘Mechanism-Based Integrated Systems for the Prediction of Drug-Induced Liver Injury MIP-DILI’’.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P93. EVIDENCE FOR A NEW TESTOSTERONE METABOLITE UNIQUELY FORMED BY CYP3A7: A POTENTIAL PROBE FOR CYP 3A7-MEDIATED METABOLISM IN THE NEWBORN Nicole R. Zane1, Yao Chen2, Michael Zhuo Wang2, and Dhiren R. Thakker1 1 Division of Pharmacotherapy and Experimental Therapeutics, UNC Eshelman School of Pharmacy, Chapel Hill, NC, USA and 2 Department of Pharmaceutical Chemistry, University of Kansas, Lawrence, KS, USA Background: CYP3A4 is the most abundant cytochrome P450 in the liver and is the primary enzyme that mediates oxidative metabolism. Despite the significant role it plays toward hepatic drug metabolism in adults, it is virtually absent in fetal tissues where the dominant isoform is CYP3A7. Downregulation of CYP3A7 expression parallels the upregulation of CYP3A4 until CYP3A7 expression is undetectable. Several reviews have reported that CYP3A7 activity as measured by 6b-hydroxylation of testosterone is lower compared to CYP3A4. The aim of this study was to define ontogeny of the CYP3A family by correlating protein expression to the functional activity of the main hepatic CYP3A isoforms. Methods: Fetal (14–39 weeks gestation), pediatric (2–10 years) and adult (418 years) liver tissue was obtained from the NICHD Brain and Tissue Bank. CYP3A4, 3A5, and 3A7 protein expression was determined by quantitative mass spectrometry and then correlated with functional data. Testosterone, a CYP3A-specific probe, was utilized to measure activity in recombinant CYP3A4, 3A5, 3A7 and human liver microsomes prepared from the liver tissues (n = 6/each). 6b-OH testosterone formation was measured by LC-MS/MS; retention time of 6b-OH testosterone was 2.51 min and linearity ranged from 0.005 mM to 10 mM (r = 0.9943). A Kruskal–Wallis test was utilized to test significance for all assays. Results: CYP3A4 expression was similar between adult and pediatric tissues (0.069 versus 0.078 nmol/ mg, respectively), but was substantially higher compared to fetal tissues (0.0004 nmol/mg, p50.002). CYP3A5 protein expression was comparable among the three populations. Conversely, CYP3A7 expression decreased with increasing age across the three populations (0.54 to 0.051 to 0.02 nmol/mg). Fetal CYP3A7 expression was higher than the pediatric and adult population by 10fold and 30-fold, respectively (p50.006). Adult and pediatric microsomes, as well as recombinant CYP3A4 and 3A5, all metabolized testosterone to 6b-OH testosterone (peak at 2.51 min) and to a minor M + 16 product that eluted with at 2.61 min. Interestingly, analysis of the metabolites formed by fetal tissues and recombinant CYP3A7 revealed that the metabolite eluting at 2.61 min was the major product, with 6b-OH testosterone (2.51 min) as a minor product. Assuming the same response factor for both metabolites, the metabolic rate, normalized to the total protein concentration, was higher in fetal tissues (666 nmol/min/mg protein) compared to adult and pediatric tissues (278 and 257 nmol/min/mg protein, respectively; p50.004). When normalized to the total CYP3A expression, the metabolic rate was 1.2 and 1.6 nmol/min/pmol CYP3A for fetal and pediatric tissues, respectively. The metabolic rate was 2.8 nmol/min/pmol CYP3A in adult tissues, higher than both fetal and pediatric tissues (p50.001). Conclusion: This study, for the first time, provides evidence that fetal tissue can metabolize testosterone at a higher rate than pediatric or adult liver tissue, and that expression of CYP3A7 in fetal tissue accounts for this high rate of testosterone metabolism. Furthermore, we report that testosterone is differentially metabolized by the CYP3A7 isoform to a novel metabolite with a retention time distinct from 6b-OH testosterone, the major metabolite formed by CYP3A4 and CYP3A5.

P94. ASSESSMENT OF THE IMPACT OF RENAL OR HEPATIC IMPAIRMENT VS. PHARMACOLOGIC INHIBTION ON SYSTEMIC EXPOSURE OF DRUGS Catherine K. Yeung1, Makiko Kusama2, Huixia Zhang3, Isabelle Ragueneau-Majlessi4, Sophie Argon4, Li Li3, Peter Chang3, Ping Zhao3, Lei Zhang3, Kenta Yoshida3, Issam Zineh3, Yuichi Sugiyama2, and Shiew Mei Huang3 1 Department of Pharmacy, University of Washington, Seattle, WA, USA, 2Laboratory of Pharmaceutical Regulatory Science, Graduate School of Pharmaceutical Sciences,The University of Tokyo, Tokyo, Japan, 3Office of Clinical Pharmacology, CDER, U.S. Food and Drug Administration, Silver Spring, MD, USA, and 4Department of Pharmaceutics, Drug Interaction Database Program, University of Washington, Seattle, WA, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

84

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Various intrinsic factors (e.g. organ impairment) and extrinsic factors (e.g. drug interactions) can affect a drug’s exposure and thus its response. Whether organ impairment and drug–drug interactions (DDIs) result in concordant effects on drug pharmacokinetics is likely to be a function of a drug’s elimination pathways; this inter-relationship has not been comprehensively assessed. We therefore conducted a systematic review of both published literature (using the University of Washington Drug Interaction DatabaseÕ organ impairment module and the FDA’s drug reviews ([email protected] FDA)) in order to: (1) compare the maximum level of increase in drug exposure observed with hepatic impairment (HI) versus pharmacologic inhibition using CYP3A4 as a model system, and (2) evaluate the association between renal impairment (RI) and pharmacologic inhibition of renal transporters on drug exposure for drugs with a high (430%) fe (where fe is defined as the fraction excreted unchanged in the urine). The analysis demonstrated that: (1) For most CYP3A4 substrates (16/19; 84%), the percent change in the AUC values obtained in subjects with severe HI was lower than those in subjects taking a strong pharmacologic CYP3A4 inhibitor. This effect was greatest for drugs with a fraction metabolized (fm) CYP3A4450%, wherein the AUC of HI subjects was 530% of the AUC observed in pharmacologically inhibited (ketoconazole) healthy subjects. (2) For drugs (n = 30) with fe  50%, the mean AUC ratio of renally impaired/healthy control subjects was 291% greater than the AUC ratio of pharmacologically inhibited (by inhibitors of renal transporters, such as cimetidine, cyclosporine, probenecid)/uninhibited controls. (3) For drugs (n = 22) with 50%4fe  30%, the mean AUC ratio of renally impaired/healthy control subjects was 29% less than the AUC ratio of pharmacologically inhibited/uninhibited controls. However, interpretation of this data is difficult as many of the pharmacologic inhibitors used in this group were not primarily inhibitors of renal transport (e.g. ketoconazole, cyclosporine, fluconazole) Our analyses show that, for drugs that are highly cleared via CYP3A4 (i.e. fm450%), a well-conducted pharmacologic inhibition with a potent and specific inhibitor reflects the worst case scenario for changes in AUC that would be observed in patients with severe HI. For drugs that are cleared mainly by renal excretion (i.e. fe  50%), the AUC changes observed in a pharmacologic inhibition study are less than the AUC changes observed in patients with severe renal impairment, likely due to the unavailability of potent and specific renal transport inhibitors. Further evaluation is needed to determine the implications of these findings for drug development and dose optimization in clinical practice.

P95. ACUTE RENAL FAILURE ALTERS THE INDUCED mRNA EXPRESSION OF HEPATIC CYP3A SUBFAMILY IN RATS Masami Doi1, Kazuki Fujimoto1, Noriko Kajikawa1, and Tetsuya Aiba2 1 School of Pharmaceutical Sciences, Okayama University, Okayama, Japan, and 2The Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, Okayama, Japan Background: As demonstrated previously, the hepatic metabolism of the CYP3A substrates alters in rats with glycerol-induced acute renal failure (ARF), due to an altered expression of the hepatic drug metabolizing enzyme. Although the mechanism for the alteration is not fully understood at present, it is speculated that the protein expression mechanism in the liver is modulated when renal function is seriously impaired. Assuming that the altered protein expression with the kidney impairment reflects a perturbation of the regulation mechanism of the hepatic mRNA expression, we probed the regulation mechanism by examining whether the dexamethasone (DEX)-induced hepatic mRNA expression occurs in a similar manner between the rats with or without the kidney impairment. Methods: Male Wistar rats of 7 weeks old were used in this study. They were intramuscularly treated with 50% glycerol to induce ARF. Rats assigned to the sham group were treated with saline instead. After 24 hours from the glycerol treatment, the induction procedure was performed, in which DEX was orally administered to each rat in the sham and ARF groups at a dose of 80 mg/kg once a day for three successive days. After another 24 hours from the last DEX administration, each rat was sacrificed and the liver was quickly removed. The liver was subsequently processed to extract total RNA, followed by a reverse transcription to have cDNA. The hepatic mRNA expressions of CYP3A1 and 3A2 were semiquantitatively evaluated with real-time PCR. Results and conclusions: Prior to the DEX treatment, there was no detectable difference in the hepatic mRNA expressions of CYP3A1 and 3A2 between the sham and ARF rats. It was shown that the DEXinduced mRNA expression of CYP3A1 significantly increases in ARF rats as compared to that in the sham rats, suggesting that the hepatic mRNA expression of CYP3A1 is facilitated in ARF rats, although its protein expression was previously shown not to increase in ARF rats. As for the hepatic CYP3A2 mRNA expression, it varies to a large extent experiment by experiment with the DEX treatment. It seems that the mRNA expression does not alter in ARF rats, being inconsistent with the previous finding that the hepatic CYP3A2 protein expression decreases in ARF rats. These observations imply that the altered hepatic expressions of CYP3A1 and 3A2 proteins in ARF rats are primarily regulated in a subsequence process following their mRNA expressions.

P96. EFFECT OF AGE AND GENOTYPE ON CYP2D6 ACTIVITY IN CHILDREN AND ADOLESCENTS Andrea Gaedigk1, Robin E. Pearce1, Jessica Tay-Sontheimer2, Kenneth E. Thummel2, Yvonne S. Lin2, and J. Steven Leeder1 1 Division of Clinical Pharmacology, Toxicology & Therapeutic Innovation, Children’s Mercy Hospital, Kansas City, MO, USA and 2Department of Pharmaceutics, University of Washington, Seattle, WA, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

85

Background: The majority of drugs prescribed to adolescents are not well- studied, even though this special patient population is undergoing considerable physiological changes with age. There is a need to better understand the impact of physiological maturation and genetic variation on drug metabolism and disposition in general and CYP2D6 activity in particular, as many antipsychotics and antidepressants are increasingly used in pediatrics and are metabolized by this enzyme. The goal of this study was to investigate the impact of developmental changes and genotype on CYP2D6 activity in children and adolescents over 7 visits covering a 3-year period. Here we report on the first 3 study visits. Methods: Healthy subjects 6–15 yrs of age at enrollment participated in a longitudinal phenotyping study involving administration of 0.5 mg/kg dextromethorphan (DM) and 4-hour urine collection every 6 months for 3 years (7 total visits); the urinary molar ratio of DM over its metabolite dextrorphan (DX) served as phenotype measure. Ethnicity was self-defined, and 189 subjects (Caucasians, n = 80; African Americans (AA), n = 79, mixed race/other n = 20) completed visit 1, with 160 subjects completing all 3 study visits. At each visit, height, weight, BMI and Tanner stage were recorded. A subset of subjects was diagnosed with Attention Deficit Hyperactivity Disorder (ADHD). CYP2D6 genotype was determined using XL-PCR and TaqMan and a quantitative copy number assay. Subjects were classified as PM, IM or EM/UM. Analyses of age and gender effects excluded PM subjects. Results: A total of 160 children and adolescents completed the first 3 study visits. Log(DM/DX) was highly correlated with CYP2D6 genotype (PM4IM4EM = UM, p50.0001 for visits 1–3 (r2adj = 0.586, 0.592 and 0.544 for visits 1, 2 and 3, respectively). A similar relationship was apparent when the analysis was repeated for AA (p50.0001) and Caucasians (p50.0001) separately, although only one AA participant was a PM. Log(DM/DX) was associated with urinary pH (r2adj = 0.06–0.19; p50.002), but was not affected by height, weight, BMI, gender or Tanner stage, nor was there any significant difference between the control and ADHD groups. No consistent statistically significant effect of age on log(DM/DX) was noted; an exception was a marginally significant finding at visit 2 for the entire nonPM group (r2adj = 0.019; p = 0.042). However, this association disappeared when the IM and EM/UM groups were analyzed separately. Subgroup analyses revealed a consistent, statistically significant age-related decrease in log(DM/DX) with increasing age in AA (p = 0.029, 0.023 and 0.018 for visits 1, 2, and 3, respectively) that was not observed in Caucasians. Overall, CYP2D6 genotype accounted for 60% and age accounted for 6% of the variance in log(DM/DX) (r2adj = 0.048, 0.057 and 0.068 for visits 1–3, respectively). There was no effect of age on log(DM/DX) when ADHD and non-ADHD subjects were compared. Discussion: CYP2D6 genotype was the dominant factor contributing to variability in CYP2D6 activity, as determined by log(DM/ DX) ratio in urine. In our analysis, a significant effect of age on log(DM/DX) was restricted to African Americans. This research was supported by NIH grants R01 HD058556 and U54 RR031295.

P97. PHARMACOKINETICS OF METFORMIN AND R,S-VERAPAMIL IN FOUR LINEAGES OF MINI-SWINE Nita Patel1, Alain Stricker-Conrad2, Sarah Renna2, Helen Huang1, Nathan Yumibe1, Kenneth Ruterbories1, Jeffrey Tan3, David White2, Jason Liu2, Larry Brown2, and Guy Bouchard2 1 Department of Drug Disposition, Eli Lilly and Company, Indianapolis, IN, USA, 2Department of Preclinical Services, Sinclair Research, Columbia, MO, USA, and 3Department of Product Design and Developability, Eli Lilly and Company, Indianapolis, IN, USA Minipigs are useful models in drug development due to their anatomical and functional similarities to humans with respect to their skin, GI, and immune system and considered as relevant disease models for arteriosclerosis, metabolic syndrome, gastric ulcer, and wound healing (Bode et al, 2010). While pharmacology evaluations may utilize one strain or another, selection based on strain differences in pharmacokinetic (PK) properties of compounds is unknown. Oral and intravenous PK studies were conducted in four different strains: Sinclair, Yucatan, Hanford, and Go¨ttingen after administration of Metformin (MET, renally cleared) or R,S-Verapamil (R,S-VER, metabolically cleared). Age and weight-matched castrated juvenile males (4.2–5.1 mos, 11–14 kg) were dosed with MET or R,S-VER (N = 3 or 4). Animals were serially bled via standard vascular access port blood sampling, plasma samples were analyzed by LC/MS/MS, and PK parameters were determined using non-compartmental analysis. The results indicated that the PK of MET was similar between all mini-swine strains, except for the Go¨ttingen which had higher plasma clearance (19.6 mL/min/kg, compared to the other three strains ranging from 8.7 to 10.1 mL/min/kg, p50.0001 Dunnett’s test). The plasma clearance of both R- and S-VER was significantly lower in Sinclair (18 and 20 mL/min/ kg, p50.05, Student’s t-test) compared to the clearance in other strains (ranging from 32 to 42 mL/min/kg). N-demethylation of R,S-VER to form R- and S-Norverapamil (NOR) was negligible after intravenous administration. Following oral dosing, however, the AUC0–48h of S-NOR was negligible compared to R-NOR in all strains. While the R-NOR to R,S-VER ratio was not different for Handford (6 ± 7) or Yucatan (9 ± 2) versus Go¨ttingen (4 ± 3), it was significantly higher in the Sinclair (15 ± 6, p50.05, Dunnett’s test) compared to the Go¨ttingen strain. Unlike in humans (Robinson & Mehvar, 1996). S-VER plasma protein binding was higher than R-VER in all mini-swine strains, while on average the plasma protein binding of R,S-VER in human (87.2%) was most similar to Sinclair mini-swine (84.0%). The preferential conversion of R,S-VER to R-NOR over S-NOR in mini-swine may be due to stereo-selectivity of protein binding. This work highlights the importance of considering the impact of metabolic and dispositional differences in mini-swine strains when selecting the relevant strain for PK studies.

References Bode J, Clausing P, Gervais F, et al. (2010). The utility of the minipig as an animal model in regulatory toxicology. J Pharmacol Toxicol Methods 62:196–220.

86

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Robinson M, Mehvar R. (1996). Enantioselective distribution of verapamil and norverapamil into human and rat erythrocytes: The role of plasma protein binding. Biopharm Drug Dispos 17:577–87.

P98. SEX-, SPECIES-, AND TISSUE-SPECIFIC METABOLISM OF EMPAGLIFLOZIN IN MALE MOUSE KIDNEY FORMS AN UNSTABLE HEMIACETAL METABOLITE (M466/2) THAT DEGRADES TO 4-HYDROXYCROTONALDEHYDE, A REACTIVE AND CYTOTOXIC SPECIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Mitchell E. Taub1, Eva Ludwig-Schwellinger2, Naoki Ishiguro3, Wataru Kishimoto3, Hongbin Yu1, Klaus Wagner4, and Donald Tweedie1 1 Department of Drug Metabolism & Pharmacokinetics, Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, CT, USA 2 Department of Drug Metabolism & Pharmacokinetics, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany, 3 Department of Pharmacokinetics and Non-Clinical Safety, Nippon Boehringer Ingelheim Co., Ltd., Kobe, Japan, and 4 Department of Development, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany Following daily oral administration of empagliflozin ((2S,3R,4R,5S,6R)-2-[4-chloro-3-[[4-[(3S)-oxolan-3-yl]oxyphenyl]methyl]phenyl]-6-(hydroxymethyl)oxane-3,4,5-triol) at 1000 mg/kg/day to male and female CD-1 mice for 2 years, renal tubular injury above background cystogenic changes were identified in male mice. Similar renal injury was not detected in male mice administered empagliflozin 300 mg/kg/day, in female mice at any dose level (up to 1000 mg/kg/day), or in male or female Han Wistar rats at doses up to 700 mg/kg/day. Studies were conducted to explore this sex and species dependent observation by evaluating possible transporter and/or metabolism differences in the disposition and metabolism of empagliflozin. Using oocytes transfected with selected mouse, rat, and human solute carrier transporters, it was determined that empagliflozin is a substrate of mouse oat3, rat Oat3, mouse oatp1a1, rat Oatp1a1 and human SGLT2. However, using isolated kidney slices obtained from male and female CD-1 mice and Han Wistar rats, there was no sex-based difference in the overall extent of uptake of empagliflozin. In vitro metabolism studies using hepatic and renal microsomes from male and female mouse, rat, and human revealed a hemiacetal metabolite of empagliflozin (M466/2), predominantly formed in male rat kidney microsomes, which had not previously been detected in plasma, urine, or feces in these species following oral dosing. The formation of M466/2 in male mouse kidney microsomes was 31-fold higher compared to female mouse kidney microsomes and was 29- and 20-fold higher compared to male and female mouse liver microsomes, respectively. Formation rates of M466/2 in the male mouse kidney microsomes were 73- and 15-fold higher than that observed with rat kidney and rat liver microsomes, respectively. M466/2 is an unstable species, which degrades to form a phenol metabolite (M380/1) and 4-hydroxycrotonaldehyde (4-OH CTA). Reduced, radiolabeled glutathione (GSH) was used to trap formed 4-OH CTA, and the structure of the GSH adduct was confirmed by mass spectrometry. Stoichiometric formation of M380/1 from M466/2 was observed (93-96% at 24 hr). However, formation of 4-OH CTA was considerably lower (17.5% at 40 hr), which is consistent with 4-OH CTA being a highly reactive species. These data provide an example of a highly selective tissue-, species-, and sex-specific lesion in male CD-1 mice arising from a cytotoxic metabolite product, 4-OH CTA. Empagliflozin metabolism in human differs significantly from the mouse, where conjugation of empagliflozin to 3 glucuronide metabolites is the most prevalent pathway and oxidation is a very minor pathway. As such, renal toxicity caused by the formation of 4-OH CTA from empagliflozin is not expected in humans.

P99. IS THE CHIMERIC HUMANIZED LIVER MOUSE MODEL READY TO PREDICT HUMAN DRUG METABOLISM IN VIVO? Edwin C.Y. Chow1, Jason Z. Y. Wang1, Hui Tang1, Ricky Ng1, Jose Silva2, David C. Evans2, and K. Sandy Pang1 1 Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, ON, Canada and 2 Department of Pharmacokinetics Dynamics & Metabolism, Johnson & Johnson Pharmaceutical Research and Development, L.L.C., Spring House, PA, USA Humanized liver mouse model is being appraised for the prediction of human metabolism and toxicity of lead, novel drug candidates in vivo. However, a thorough examination is needed to assess liver integrity (zonation, microcirculation, and remnant mouse liver tissue) and inter-organ communication between humanized liver and other mouse tissues. Thus, FRGN [Fah(/), Rag2(/), and IL-2rg(/) on NOD background] mice and FRGN mice containing livers repopulated with mouse (mFRGN; 129S7 strain) or human (hFRGN; 3 different donors) hepatocytes were investigated in single pass prograde (P) and retrograde (R) perfusion studies to study enzyme zonation and microcirculation, and liver histology, bile acid formation, and gene expression changes not only in liver, but in small intestine, brain, and kidney were examined. In FRGN liver perfusion studies with harmol, the ratio of steady-state harmol sulfation/glucurondation rates (HS/HG = 0.078 ± 0.018 to 0.045 ± 0.010, p50.05) decreased by 73% and extraction ratio (E = 0.633 ± 0.077 to 0.732 ± 0.052, p50.05) increased by 16% when flow was switched from P to R flow, events reflective of intact enzyme zonation and an anterior distribution of sulfotransferase vs. UDPglucuronosyltransferase activities. By contrast, there was no difference in the HS/HG ratio (0.133 ± 0.050 vs. 0.142 ± 0.109) and E (0.697 ± 0.057 vs. 0.714 ± 0.070) for hFRGN livers in P and R directions, suggestive of loss in liver enzyme zonation. Results from mFRGN livers were similar to those for hFRGN livers in P and R perfusions and showed loss of zonation. Histology revealed that mFRGN and hFRGN livers were more tightly packed with hepatocytes within sinusoids. Furthermore, mRNA analysis of human genes from the same donor (n  3) showed highly variable gene expression in different recipient hFRGN

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

87

livers. Moreover, mouse genes such as mPXR, mMDR1a, mBCRP, mSULT2A1, mGSTa4-4, and mVLDLR persisted in hFRGN livers, with levels present at equal or higher levels compared to FRGN livers. Disruption of bile acid homeostasis was evident in hFRGN mice, displayed as higher frequency of enlarged gall bladders and necrotic livers. The plausible explanation was the miscommunication between the mouse intestine and human liver: hFRGN mice produced higher amounts of total bile acids (in bile, plasma, liver, and feces), especially cholic, chenodeoxycholic, and deoxycholic acids, high affinity FXR ligands, whereas FRGN and mFRGN mice formed mostly muricholic acids (a-MCA, b-MCA, and !-MCA), lesser affinity FXR ligands. Upon reaching the hFRGN mouse intestine, human bile acids elicited greater changes in ileal mFXR genes, namely the small heterodimer partner (mSHP), fibroblast growth factor 15 (mFGF15), and ileal bile acid binding protein (mIBABP). Furthermore, other subtle changes in mRNA expression of nuclear receptors, transporters and enzymes in intestine, brain and kidney was found in hFRGN mice as compared to FRGN mice: brain and intestinal mMDR1a, renal and intestinal mPXR, mCAR, mOSTab, mMRP2, mMRP4, and mGSTa4-4, renal mMRP3, and intestinal mSULT1A1 and mUGT1A1 were increased, and brain mCYP1A2, mCYP2C29, mCYP2E1, and mCYP3A11 and intestinal mMRP3 decreased. The composite data suggest that careful consideration must be given to transporter and enzyme changes in alternate mouse target/elimination tissues, in addition to those in the humanized liver, when using chimeric mouse model to study human ADME.

P100. SPECIES- AND GENDER-SPECIFIC CHIRAL INVERSION OF A SECONDARY ALCOHOL BY ADH

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Karsten Menzel, Sheri Smith, Bo Liu, and Christine Fandozzi Department of Drug Metabolism, Merck & Co., West Point, PA, USA Chiral inversion of a drug is a known phenomenon and has been reported most frequently for NSAIDs. The conversion of one enantiomer to its antipode happens predominantly in the presence of enzymes. In such cases, the reactions contributing to the chiral inversion can depend on the species or in some rare instances can be dependent on the gender. This presentation will summarize the in vitro and in vivo data supporting the species and gender specific chiral inversion of a secondary alcohol. Metabolite profiling studies of in vitro and in vivo samples suggested that an investigational compound underwent chiral inversion of a secondary alcohol with a species and gender dependence. The chiral inversion was favored primarily in rats and to a lesser extent in dogs and rhesus monkeys. Additional studies showed, that the conversion of configuration of the secondary alcohol was observed primarily in plasma and hepatocytes from male rats and to a significantly smaller extent in female rats. The change in configuration of a secondary alcohol is hypothesized to follow a reaction sequence involving an oxidation followed by a reduction. While the oxidation from the secondary alcohol to the ketone forms an achiral center, the reduction back to the alcohol can either result in retention or inversion of stereochemistry at the carbon atom bearing the secondary alcohol. It was hypothesized that alcohol dehydrogenase (ADH) might be involved in both reaction steps. Mechanistic studies of the metabolic chiral inversions for the investigational compound were carried out and data will be presented to support the hypothesis that ADH can catalyze not only the oxidation of the secondary alcohol to the ketone but also catalyze the reduction of the ketone back to the secondary alcohol. In vitro studies with rat specific ADH isoforms have been reported to be stereospecific, however, at the moment it is unknown if the ketone intermediate of the investigational compound can be reduced stereoselectively by rat ADH in a gender-dependent manner. This presentation will summarize the in vitro and in vivo data supporting the gender specificity for the chiral inversion of the secondary alcohol and the possible involvement of ADH in the reaction sequence.

P101. ABSOLUTE QUANTIFICATION OF WITHOUT INTERNAL STANDARDS: ACCELERATOR MASS SPECTROMETRY AND MICROTRACERS IN PHARMACEUTICAL RESEARCH TO ADDRESS MIST LIABILITIES Stephen Dueker1 and Jason Giacomo2 1 Eckert & Ziegler Vitalea Science, Davis, CA, USA and 2AMS Laboratory, Eckert & Ziegler Vitalea Science, Davis, CA, USA Metabolites can be both mediators of drug toxicity and efficacy and recent FDA Guidance for Industry strongly recommends that studies leading to in vivo drug metabolism in humans be performed as early as feasible. In this light, a complete accounting of significant metabolite(s) in both safety species and humans during clinical testing is a high priority. The quantitative sensitivity of 14C-Accelerator Mass Spectrometry (AMS) empowers early human metabolism studies to be conducted within Phase I trials with microtrace (nCi) quantities of administered 14C-labeled drug. Direct atom counting with AMS reveals the full complement of metabolites at absolute abundance levels with no requirement for internal standards. These measurements facilitate informed decision-making in and around Phase 1 studies with incremental added cost to the program using trace radiolabel. The results of Plasma extract UPLC radioprofiles are presented from human and two safety species (rat and dog) after microtracer oral administration of an experimental drug. The radiochromatograms are arranged in a stacked array and retention times are locked relative to co-chromotography standards. Radioactivity recovery was monitored through the full extraction and separation procedure to provide assurance that the full complement of metabolites were assessed. Lower Limits of Quantification were 8 amol of 14C per collected LC fraction and in all cases, less than 1 dpm of radioactivity was loaded onto the UPLC system. Multiple metabolic differences are noted not only between the human and the animals but also between the two species. Specifically, the metabolite included as a cold spike is disproportionately found in humans and not in either species. Total radioactivity of 14C in plasma was measured across the time course and the residual

88

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

14C not extracted using protein precipitation is plotted. The recovery of total 14C from the plasma went from 90% at early time points to less than 40% at later time points. These data suggest strong and possible covalent binding of reactive metabolites. These data show AMS to be a suitable tool for assessing metabolite liabilities without prior knowledge of metabolite structure or internal standards.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P102. INVESTIGATION OF THE FUNCTIONAL SNPs IN THE BCRP TRANSPORTER AND SUSCEPTIBILITY TO COLORECTAL CANCER ¨ zhan1, Fatih Sari2, and Hakan Yanar3 Gu¨l O 1 Department of Pharmaceutical Toxicology, Istanbul University, Istanbul, Turkey, 2Department of Pharmaceutical Toxicology, Istanbul University, Istanbul, Turkey and 3Department of General Surgery, Istanbul University, Istanbul, Turkey Breast cancer resistance protein (BCRP) protects tissues by actively transporting xenobiotics and their metabolites out of the cells. BCRP is expressed abundantly in the apical membrane of normal intestinal and colonic epithelium. Its substrates include a number of structurally unrelated compounds including drugs, pesticides, carcinogens and endogenous compounds. Although the functional and common BCRP alleles, 34G4A and 421C4A, are shown to vary by ethnicity, their potential mechanism is not satisfactorily described in affecting susceptibility to colorectal cancer yet. We aimed both to evaluate the effects of BCRP variants on the susceptibility to colorectal cancer and to predict the individual response of Turkish people to xenobiotics transferred by BCRP. BCRP 421C4A was significantly associated with colorectal cancer risk (OR = 16.12; p = 0.204). The findings are the first results of BCRPs allele distributions in the Turkish population and provide an understanding of correlation between therapeutic approaches and aetiology of colorectal cancer.

P103. AGE-RELATED CHANGES IN EXPRESSION AND ACTIVITY OF HUMAN MITOCHONDRIAL GLUTATHIONE TRANSFERASE ZETA1 Guo Zhong1, Marci Smeltz1, Stephan C. Jahn1, Taimour Langaee2, Peter W. Stacpoole3, and Margaret O. James1 1 Department of Medicinal Chemistry, University of Florida, Gainesville, FL, USA, 2Center for Pharmacogenomics, University of Florida, Gainesville, FL, USA, and 3Department of Medicine and Biochemistry and Molecular Biology, University of Florida, Gainesville, FL, USA Glutathione transferase zeta1 (GSTZ1) catalyzes glutathione (GSH)-dependent dechlorination of the investigational drug, dichloroacetate (DCA), to an inactive metabolite, glyoxylate. DCA is taken up by mitochondria where it acts to inhibit the pyruvate dehydrogenase complex and has therapeutic potential for cancer, cardiovascular and metabolic disorders. An understanding of its metabolism is important for appropriate dose selection. GSTZ1 was first identified in hepatic cytosol and later in the mitochondria (Li et al., 2011). In people, age has been shown to influence DCA pharmacokinetics and side effects (Shroads et al. 2008). Although the ontogeny of human hepatic cytosolic GSTZ1 has been investigated (Li et al., 2011), the developmental pattern and properties of human hepatic mitochondrial GSTZ1 are not known. Our hypothesis was that mitochondrial GSTZ1 would exhibit different enzyme kinetic properties and ontogeny compared with cytosolic GSTZ1. Human liver samples from 101 donors aged 1 day to 84 years were processed into subcellular fractions and the GSTZ1 haplotype determined. GSTZ1 expression and activity with DCA were determined in mitochondrial fractions. Expression of mitochondrial GSTZ1, assessed by Western blot, increased in an age-dependent manner to a plateau after age 7 years to 0.05 ± 0.02 ng GSTZ1/mg mitochondrial protein. Mitochondrial GSTZ1 activity was 3 to 4-fold higher in samples from GSTZ1A carriers than in samples from carriers of other haplotypes. A correlation (r2) of 0.96 was obtained between activity and expression in mitochondrial samples from GSTZ1A non-carriers (n = 86) and 0.83 for GSTZ1A carriers (n = 15). The absolute values of both activity and expression were lower in mitochondria than cytosol. The apparent enzyme kinetics for GSH in cytosol best fit the Michaelis–Menten model, and the Km,app (1.98 ± 0.59 mM; mean ± S.D., n = 3) was lower than the reported physiological GSH concentration of 5 mM. In mitochondria the data fit the Hill equation (h51; Kprime app = 6.30 ± 4.35 mM; mean ± S.D., n = 3), and maximum activity was measured at higher concentrations of GSH than are reported physiologically. In summary, we found an age-related increase in the expression of human hepatic mitochondrial GSTZ1 similar to that reported for cytosolic GSTZ1, however the GSH requirements of cytosolic GSTZ1 and mitochondrial GSTZ1 differed. Both mitochondrial and cytosolic GSTZ1 are expected to contribute to overall DCA metabolism.

References Li W, Gu Y, James MO, et al. (2012). Prenatal and postnatal expression of glutathione transferase z 1 in human liver and the roles of haplotype and subject age in determining activity with dichloroacetate. Drug Metab Dispos 40:232–239. Li W, James MO, McKenzie SC, et al. (2011). Mitochondrion as a novel site of dichloroacetate biotransformation by glutathione transferase zeta 1. J Pharmacol Exp Ther 336:87–94. Shroads AL, Guo X, Dixit V, et al. (2008). Age-dependent kinetics and metabolism of dichloroacetate: Possible relevance to toxicity. J Pharmacol Exp Ther 324:1163–1171.

This work was supported in part by the US Public Health Service, grant RO1 GM 099871.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

89

P104. ALTERED HEPATIC DRUG METABOLIZING ACTIVITY IN RATS WITH ACUTE LUNG IMPAIRMENT

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yuki Hori1, Yasumasa Simizu2, and Tetsuya Aiba2 1 School of Pharmaceutical Sciences, Okayama University, Kita-ku Okayama-shi, Japan and 2Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama University, Kita-ku Okayama-shi, Japan Background: As oxidative reactions play an important role in the hepatic drug metabolism, a lung impairment presumably influences the hepatic drug metabolizing activity when it causes a considerable decrease in the oxygen supply to circulation, and yet, such an impact of the lung impairment is not fully understood. We, therefore, examined whether and to what extent the lung impairment affects the hepatic drug metabolizing activity in rats, aiming to conduct an efficient pharmacotherapy for patients suffering from serious lung diseases. Methods: Male Wistar rats (260–330 g) were used in the study. After anesthetized, rat was administered diluted hydrochloride solution to the right bronchus via nasal cavity to induce acute lung impairments (ALI). Following a 24-hour recovery period, the level of the blood oxygen saturation was checked with a veterinary pulse oximeter, and blood was collected to evaluate the plasma concentration of a1-acid glycoprotein (AGP), an acute phase inflammation–related protein. Rat was then sacrificed for hepatic excision. The liver microsomes were prepared from pooled livers procured from 4 rats with an ultracentrifugation method. The drug metabolizing activity of hepatic cytochrome P450 (CYP) was examined in the incubation study with the microsomes, in which midazolam, warfarin, and propranolol were used as representative substrates of CYP3A, 2C, and 2D, respectively. Their concentrations in the incubation mixture were determined with an HPLC method. The hepatic expressions of the CYP3A subfamily and CYP oxidoreductase (POR), a protein participating in electron transfer to CYP, were evaluated with western blotting analysis. Results and conclusions: In ALI rats, the blood oxygen saturation significantly decreases from the normal level of 94% to 86%, which corresponds to the decrease in the blood oxygen tension from the normal level of 72 mmHg to 52 mmHg. The plasma AGP concentration increases approximately 3.5 times in ALI rats as compared to that in normal rats. The rats are therefore considered to be suffering from a serious lung impairment. As for the hepatic drug metabolizing activity, an altered metabolic profile was observed in the midazolam metabolism in ALI rat, in which the Michaelis constant Km decreases and Vmax remains unchanged. Altered metabolic profiles were also observed for the warfarin and propranolol metabolism. On the other hand, little change was noted in the hepatic expressions of CYP3A1 and CYP3A2, while that of POR seems to increase in ALI rats. With these findings, it is probable that when the pulmonary function is impaired, the hepatic expression of POR increases to compensate an inadequate blood oxygen saturation. This facilitates the oxidation process in the hepatic drug metabolizing reaction, altering the metabolic profiles of various compounds in a non-specific manner.

P105. LABEL-FREE MASS SPECTROMETRY IMAGING OF DRUG AND METABOLITES IN TISSUE Angela Wehr and Chandra Prakash Dmpk, Biogen Idec, Cambridge, MA, USA Tissue distribution analysis of drugs is essential in drug development. Localization of the drug at the target site enhances confidence in the mechanism and duration of action. Traditionally, distribution analysis has been achieved through the use of autoradiography or LC-MS. While autoradiography studies track the distribution of all drug related compounds, they require synthesis of a radiolabeled drug and lack the ability to distinguish the parent drug from metabolites. Alternatively, while LC-MS may differentiate drugs and metabolites, this process requires the tissue be homogenized, destroying the spatial orientation of the analytes within the tissue. High-resolution imaging mass spectrometry permits label-free imaging of drugs in tissue with the added ability to determine the mass of the compound, thus distinguishing drugs from metabolites. Here, we report the distribution of drug within rat tissue following oral dosing. Rats were euthanized at multiple time points post-dose. Following harvest, tissues were flash frozen and sectioned. DESI, MALDI, and LESA-MS were performed in the positive ion mode. The distribution of olanzapine in whole body sections was determined by these techniques. The advantages of these techniques over autoradiography and LC-MS were successfully demonstrated in the analysis of the distribution of this drug in rat tissue. The studies were conducted in house, without radiolabeled compound, and the analysis time was reduced to 1–2 months, 75% faster than radiolabeled studies. The distribution of the drug to the brain, spinal cord, liver, kidney, and testes, was easily obtained. Additionally, preliminary work with internal and external standards suggests that normalization of the response may permit quantitative analysis with these tissues as well. In summary, we used MS imaging to determine the distribution of drug in rat tissue. HRMS enabled the specific detection and MS/MS analysis confirmed the identity of the drug. This label-free imaging technology enables distribution analysis earlier in the drug discovery process and at a lower cost than traditional techniques.

P106. THE DISPOSITION OF ELND005 (SCYLLO-INOSITOL) IS CONSERVED ACROSS SPECIES AND PUTATIVELY MEDIATED BY INOSITOL SPECIFIC SOLUTE CARRIER TRANSPORTERS Kevin P. Quinn1, Earvin Liang2, Holly Lin3, George R. Tonn1, John Michael Sauer1, Matthias Kurth2, Michael Skov4, Bruce Connop5, Aleksandra Pastrak6, and Susan Abushakra2 1 Department of Drug Metabolism and Pharmacokinetics, Former Elan Pharmaceuticals Inc, South San Francisco, CA, USA, 2 Clinical Development, Transition Therapeutics Inc, San Mateo, CA, USA, 3Department of Bioanalytical Development, Former Elan Pharmaceuticals Inc, South San Francisco, CA, USA, 4Department of Safety Evaluation, Former Elan Pharmaceuticals Inc,

90

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Downloaded by [112.134.238.102] at 07:30 10 February 2016

South San Francisco, CA, USA, 5Department of Preclinical Development, Transition Therapeutics, Inc, Toronto, ON, Canada, and 6Department of Clinical Development, Transition Therapeutics, Inc, Toronto, ON, Canada ELND005 (Scyllo-Inositol), one of the nine naturally occurring stereoisomers of myo- inositol (1,3,5/2,4,6-hexahydroxycyclohexane), is currently being evaluated as a potential treatment for Agitation and Aggression in Alzheimer’s Disease (AD) and as a potential cognitive enhancing agent in young adults with Down Syndrome without dementia. Myo-inositol is important in second-messenger signaling, but there is a paucity of information on its disposition and even less on the other stereoisomers. Inositols are very hydrophilic, and their absorption, distribution and excretion putatively require carrier-mediated active transport (SLC 2A13, SLC5A3 and SLC5A11). Inositols are also known to be catabolized to CO2 and H2O by the D-gluconate-xylulosepentose-phosphate pathway. The objective of this work was to elucidate the importance of these pathways to the disposition of ELND005 in humans, dogs, and rats. The definitive disposition of scyllo-inositol was investigated following a single oral administration of 14C-ELND005 to rats and dogs (30 mg/kg and 100 mCi/kg) or to healthy male human subjects (1000 mg and 100 mCi). Total radioactivity and ELND005 concentrations were measured in each specimen where possible. The results demonstrated similar plasma concentration-time profiles in all three species, suggesting that the disposition of scyllo-inositol is conserved. Absorption was rapid to moderate and complete, followed by an initial rapid distribution phase and then a shallow protracted terminal phase with a half-life estimated in weeks. This latter phase is thought to be mediated by a slow redistribution of compound from tissues. In all species, total body clearance was low while the volume of distribution was substantially greater than total body water. Binding of 14C-ELND005 to plasma proteins was negligible in all species. In a rat Quantitative-WholeBody-Autoradiography study, 14C-ELND005-derived radioactivity was widely distributed by 24-hours and was preferentially distributed into the CNS and glandular tissues that highly express inositol transporters. As part of a mass-balance study in the dog, tissues were harvested two-months after dosing and showed a similar distribution pattern of 14C-ELND005-derived radioactivity as seen in the rat. Evidence for biotransformation in all three species was negligible in the plasma, urine and feces, which were predominately composed of intact 14C-ELND005. 14CO2 was detected in expired air and accounted for approximately 3% and 18% of the total dose in the rats and humans, respectively. In all species the majority of the dose was recovered in the urine as intact 14C-ELND005, suggesting that renal excretion was the major route of scyllo-inositol elimination. Total recoveries from the rats and humans were 90% or greater, while the recovery from the dogs was 80%, most likely due to the technical limitation to capture expired air. In conclusion, the overall similarities in the dispositional properties for scyllo-inositol across species support a conserved mechanism which may be typical for inositols in general.

P107. ABSTRACT WITHDRAWN

P108. INTRACEREBROVENTRICULAR (ICV) AND/OR INTRATHECAL (IT) SINGLE DOSE ADMINISTRATION OF AN 35S-LABELED SECOND GENERATION ANTISENSE OLIGONUCLEOTIDE (ASO) TO GENERATE PHARMACOKINETIC (PK) DATA IN MICE AND RATS C. Webber1, RA Shabir1, H. Murphy1, K. Read1, DA Norris2, C. Mazur2, CF Bennett2, J.S. Grundy2, M. Beconi3, and Sarah Pass1 1 Huntingdon Life Sciences Ltd, Alconbury, Cambridgeshire, UK, 2Isis Pharmaceuticals, Carlsbad, CA, USA, and 3CHDI Foundation, Los Angeles, CA, USA CD-1 mice and Sprague-Dawley rats received a single dose of an 18-mer 35S-labeled second generation ASO [20 -Omethyoxyethyl (MOE) partially modified (chimeric gapmer) phosphorothioate oligonucleotide] in a study designed to obtain information on (i) the time-course of whole blood and plasma radioactivity concentrations; (ii) the distribution of radioactivity in body tissues, including cellular distribution within the brain; (iii) the rates and routes of excretion of radioactivity. Within each of 3 phases (PK/Mass Balance, QWBA and Microautoradiography), a group of mice received a single ICV dose (0.1 mg) and separate groups of rats received either single ICV or IT doses (0.5 mg). ICV doses were administered by slow bolus injection at a maximum dose volume of 30 mL (rat) or 10 mL (mouse) using stereotaxic apparatus to set appropriate dose site co-ordinates. IT doses were administered by bolus injection into the intrathecal space at a maximum dose volume of up to 30 mL via a catheter sutured in place. Following dosing in the PK/Mass Balance phase, whole blood, plasma, and CSF were sampled at intervals until 168 h (ICV) or 672 h (IT) and excreta collected until 168 h. Each group showed rapid absorption (Cmax  1 h) followed by multiphasic decline and terminal half-lives of 468 h. Excretion of radioactivity was protracted, with approximately 20–25% dose in urine, 5–10% dose in faeces and the balance retained in the carcass at 168 h. Following dosing in the QWBA phase, one animal was prepared for imaging from each group (ICV and IT) at 5 different time points from 4 to 336 h. The only appreciable differences in the pattern of distribution of radioactivity between species, routes of administration and dose levels was a slower absorption seen in tissues of the CNS at the first sacrifice time (4 h) in rats administered via the IT route. Clearance of drugrelated material from tissues was slow. Following dosing in the microautoradiography phase, one animal from each group (ICV and IT) was prepared for imaging at 4 h and 168 h post dose. Intense silver granules representing radioactivity were present in the peripheral areas of the cerebrum and cerebellum following ICV injection and, to a lesser extent, IT injection. After ICV or IT administration in rodents, significant drug uptake and distribution throughout the CNS was observed, as well as to systemic

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

91

tissues. Delivery of ASOs directly into the CSF via IT delivery represents a promising approach to treat various CNS disorders that are amenable to antisense pharmacology.

P109. PREDICTING MAJOR ELIMINATION ROUTE OF ORALLY ADMINISTERED DRUGS UTILIZING IN VITRO AND IN SILICO MODELS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Chelsea M. Hosey and Leslie Z. Benet Department of Bioengineering and Therapeutic Sciences, University of California – San Francisco, San Francisco, CA, USA In vitro permeability measurements and in silico permeability predictions can predict the extent of metabolism of drugs in humans. Drugs that are poorly metabolized inhumans will have a relatively low permeability rate, and their major route of elimination (renal or biliary elimination of unchanged drug) can be predicted by an in silicomodel utilizing calculated polarizability and metabolic stability. In vitro permeability rates were measured in Caco-2, MDCK, and PAMPA datasets and in silicopermeability rate predictions were calculated in VolSurf+ and ADMET Predictor. Area under the receiver operating characteristic curves determined if in vitro or insilico permeability rates were capable of differentiating extensively metabolized compounds from those eliminated as parent drug. For each cell line, drugs wereconsidered as a standard to deLine high/low permeability if they were present in 4 or more datasets. The drug that gave the highest sensitivity-specificity average was selected as a standard. Performance measures (sensitivity, specificity, positive predictive value, and negative predictive value) of permeability rate of drugs compared toeach standard to predict the extent of metabolism were evaluated for external datasets. Highly and poorly permeable compounds were segregated to predict the extent ofmetabolism; the elimination route (biliary or renal) of poorly permeablecompounds was predicted by a logistic regression model that we created included calculatedpolarizability and metabolic stability. In vitro permeability rate measurements and in silico predictions are differentiable between extensively and poorly metabolizedcompounds with area under the receiver operating curves of 0.93 ± 0.08, 0.91 ± 0.03, 0.93 ± 0.05 for Caco-2, MDCK, and PAMPA in vitro models, respectively, and0.82 ± 0.03 for the VolSurf + Caco-2 model. Sensitivity, specificity, positive predictive value when predicting extent of metabolism by comparing drug permeabilities to the standards salicylic acid or labetalol were 40.8 in training and testing sets, while negative predictive values were 40.7. The twotier elimination route prediction approach correctly predicted 61.3 ± 7.6%, 84.0 ± 0.9%, and 80.1 ± 9.3% of metabolized, biliarily eliminated, and renal eliminated drugs respectively. The optimalpermeability standards used to predict the extent of metabolism are salicylic acid and labetalol for Caco-2 and MDCK cell lines, respectively. The major eliminationroute of a compound can be predicted in a two-tier approach, predicting the extent of metabolism from high or low permeability rates as compared to labetalol orsalicylic acid, and by utilizing metabolic stability and polarizability in a logistic regression model to predict the major elimination route of poorly metabolized drugs.

P110. EVALUATION FOR CHANGE IN THE DISPOSITION OF MIDAZOLAM UNDER HYPOTHERMIA IN RATS BY IN VITRO METABOLISM AND PROTEIN BINDING STUDY Hirotaka Miyamoto, Mariko Taniguchi, Haruna Hirata, Shintaro Fumoto, and Koyo Nishida Graduate School of Biomedical Sciences, Nagasaki University, Nagasaki, Japan Introduction: We have already reported that the hepatic disposition of model compounds with different elimination processes was altered under hypothermic condition in rats (Miyamoto et al., 2012). Midazolam (MDZ) is one of the sedating drugs used during therapeutic hypothermia and metabolized by CYP3A4 in human. In this study, we examined the effect of temperature on the distribution and metabolism of MDZ, as a probe drug of CYP3A activities, to clarify the mechanism of an alternation in the pharmacokinetics of drugs metabolized by CYP3A under hypothermia. Methods: In the in vitro study, the hepatocytes were obtained from male Wistar rats and incubated at 37, 32 or 28  C for 10 min after the addition of MDZ (0.25–10 mg/mL). Male Wistar rats were anesthetized with pentobarbital, and their body temperatures were maintained at 37  C by heat lump and 32 or 28  C by ice pack. MDZ (5 mg/kg) was injected into jugular vein, followed by collection of blood. The brain and liver were removed at appropriate time. The unbound fraction ratio of MDZ in rat serum at 37, 32, or 28  C was determined by the equilibrium dialysis method. The MDZ concentration was determined by HPLC-UV (wavelength: 220 nm). Results and discussion: The Michaelis constant (Km) and maximum eliminate velocity (Vmax) of MDZ for elimination from hepatocyte suspension were obtained by the Eadie-Hofstee plot. The Vmax of MDZ was decreased about 30% at 32  C and 50% at 28  C compared to 37  C respectively, although the Km was not altered. The plasma concentration of MDZ was significantly increased at 28 oC compared to 37  C and total body clearance was decreased according to body temperatures. The distribution volume at steady state of MDZ obtained by AUC/MRT was significantly decreased at 28oC compared to 37oC. The MDZ concentration in the brain and liver was not altered at 32 or 28  C compared to 37oC and the tissue-to-plasma ratio of MDZ was decreased under hypothermia. This result suggests that the distribution of MDZ into the brain and liver was inhibited under hypothermia. The unbound fraction of MDZ in rat serum was about 2.4% at 37  C and was significantly decreased at 32oC (1.2%) and 28oC (1.1%). This result suggests that the reduction in unbound fraction of MDZ under hypothermia might cause the inhibition of tissue distribution of MDZ. In addition, we estimated the plasma concentration of MDZ at 37 and 32  C considering the CYP3A activity and unbound fraction ratio of MDZ based on physiologically based pharmacokinetic model analysis. Conclusion: The

92

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

change in pharmacokinetics of MDZ under hypothermia in rats could caused by the reduction in CYP3A activity and unbound fraction ratio of MDZ in rat serum under low temperature.

Reference Miyamoto H, Miyake H, Yoshikawa N, et al. (2012). Evaluation of changes in hepatic disposition of phenolsulfonphthalein, indocyanine green and FITC-dextran at low temperatures by rat liver perfusion system. J Pharm Pharmacol 64:848–854.

P111. EFFECT OF VEHICLE ON THE DISPOSITION OF SUNSCREEN INGREDIENT 2-HYDROXY-4METHOXYBENZOPHENONE IN RODENTS FOLLOWING DERMAL APPLICATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Suramya Waidyanatha1, Jacob A. McDonald2, Melanie Doyle-Eisele2, Barry S. McIntyre1, and J. Michael Sanders1 1 Division of the National Toxicology Program, National Institute of Environmental Health Sciences, Research Triangle Park, NC, USA and 2Lovelace Respiratory Research Institute, Albuquerque, Albuquerque, NM, USA 2-Hydroxy-4-methoxybenzophenone (HMB) is used as a UV stabilizer in a variety of products including sunscreens, cosmetics and pharmaceuticals and hence there is potential for widespread human exposure. The National Toxicology Program is investigating the toxicity and carcinogenicity of HMB in rodents. To aid in the vehicle selection for dermal toxicity studies, the effect of vehicle on the disposition of HMB was investigated in male and female Harlan Sprague Dawley Rats and B6C3F1/N Mice following a single dermal application of 0.1 or 10 mg/kg [14C]HMB. The vehicles under consideration were a lotion (olive oil:emulsifying wax:water,15:15:70), coconut oil, light paraffin oil, ethanol, and ethanol:coconut oil (1:1). Following application of 10 mg/kg in male rats, the highest absorption was observed in paraffin oil (79.8 ± 6.9%) with a trend paraffin oil4ethanol:coconut oil4ethanol4coconut oil4lotion. The absorption was not dose-dependent in male rats (paraffin oil vehicle; 79.8 ± 6.9 and 72.5 ± 8.5%, respectively for 10 and 0.1 mg/kg) and male mice (ethanol vehicle; 43.7 ± 3.7 and 41.4 ± 3.7%, respectively for 10 and 1 mg/kg). There was no species difference in absorption following application of 10 mg/kg [14C]HMB formulated in lotion vehicle (male rat, 34.6 ± 8.5%; male mouse, 36.8 ± 8.2%), however, the absorption was higher in rats (67.8 ± 3.2%) than in mice (43.7 ± 3.7%) in ethanol vehicle. There was no apparent sex difference in the absorbed dose in both species. The total administered dose recovered in tissues for male rats were higher (2.8–8.3%) than for male mice (0.6– 1.2%). Excretion of HMB-derived radioactivity was mainly via urine and feces in both rats and mice. In conclusion, these studies indicate that the vehicle used has an impact on the systemic absorption of HMB following dermal application. This work was conducted under NIEHS Contract N01-ES-75562.

P112. EVALUATING BRAIN PENETRATION OF G01 IN P-GLYCOPROTEIN KNOCKOUT RATS AND WILD-TYPE RATS PRE-TREATED WITH ELACRIDAR Justin Ly1, Linda Bao2, Kang-Jye Chou2, Jonathan Cheong1, Alan Deng1, Anthony Estrada3, Liling Liu1, Emile Plise1, and Lesley Murray1 1 Drug Metabolism and Pharmacokinetics, Genentech Inc., South San Francisco, CA, USA, 2Pharmaceutics, Genentech Inc., South San Francisco, CA, USA, and 3Discovery Chemistry, Genentech Inc., South San Francisco, CA, USA P-glycoprotein (P-gp) is part of a larger superfamily of efflux transporters found in the gut, kidneys, liver, brain, and other organs. Particular in the CNS, it plays a major role in limiting the brain penetration of compounds that are found to be P-gp substrates. Recently, a P-gp knock-out (KO) rat model lacking the expression of mdr1a/b has become available. However, the use of the P-gp KO model may not be practical when screening multiple compounds due to the high cost of this model. The objective of this work was to assess the brain penetration of a P-gp substrate, G01, in P-gp KO rats and to evaluate a potential simple and cost effective method using wild-type (WT) rats pretreated with elacridar. Male Spraque-Dawley P-gp KO and male SpraqueDawley WT rats were used. WT rats received either no pre-treatment, vehicle, or elacridar formulated in DMSO/MCT or in 50% Cremophor EL/35% 1-Methyl-2-Pyrrolidone/15% Myglyol 810 (SEDDS). After 2 h, WT and P-gp KO groups received 10 mg/kg of G01 orally. At 1 and 3 hr post-dose of G01, brains and plasma from the two groups were collected (N = 3). LC/MS/MS was used to quantitate levels of G01 in brain homogenates and plasma samples. Following an oral dose of 10 mg/kg, there was no difference in the levels of G01 in the brains between no pre-treatment and vehicle treated groups. Levels of G01 in the brains of KO rats and WT rats pre-treated with elacidar formulated in SEDDS at 1 hr post-dose were approximately 10- and 7-fold higher than no pre-treatment, respectively. At 3 hr post-dose, G01 levels were approximately 16- and 18-fold higher than no pretreatment, respectively. Formulating elacridar in SEDDS yielded approximately 16- and 3-fold higher brain levels at 3 h postdose when compared to DMSO/MCT, respectively. In addition, double doses of elacridar appeared to inhibit P-gp better than a single dose. G01 levels in the brain from single and double doses of elacridar at 3 hr post-dose were approximately 6- and 16-fold higher than no pre-treatment, respectively. Although P-gp KO rat model is available and was shown to be effective in increasing brain concentration of G01, alternative model using WT rats pre-treated with a chemical inhibitor, such as elacridar, can also be employed to achieve similar results as the KO model. Data showed that elacridar concentration correlated with the extent of change in the levels of G01 in the brain. These data suggest that elacridar pre-treated WT rats may be used to assess the brain penetration of G01. However, sufficient exposure of elacridar may be required to elicit significant effect. Optimizing formulation or increasing the frequency of dosing to enhance exposure may enable inhibition of P-gp efflux at the blood-brain barrier.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

93

P113. TISSUE DISTRIBUTION OF DABRAFENIB AND ITS METABOLITES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Lauren E. Richards-Peterson1, Mike Morris1, Reid Groseclose2, Kristen Jurusik1, Elisabeth Minthorn3, Tom Wilde1, Donna Mamaril-Fishman1, Molly Karlinsey1, Katie Olson2, David Bershas1, Chet Bowen1, Peter Gorycki1, and Stephen Castellino2 1 PTS Dmpk, GlaxoSmithKline, King of Prussia, PA, USA, 2PTS Dmpk, GlaxoSmithKline, Research Triangle Park, NC, USA, and 3Oncology R&D, GlaxoSmithKline, Collegeville, PA, USA Dabrafenib (TAFINLAR, GSK2118436, N-{3-[5-(2-Amino-4-pyrimidinyl)-2-(1,1-dimethylethyl)-1,3-thiazol-4-yl]-2-fluorophenyl}-2,6-difluorobenzene sulfonamide, methanesulfonate salt) has been approved for treatment of patients with unresectable or metastatic Braf V600E-positive melanoma, and has been shown to be effective in shrinking brain metastases in patients. We have conducted in vivo studies in mice and rats to assess the tissue distribution of dabrafenib and its main metabolites and have investigated their human Pgp transporter substrate status in vitro. Metabolites hydroxy- and carboxy-dabrafenib are formed from dabrafenib via sequential CYP oxidations, while desmethyl- dabrafenib is generated by a pH-driven decarboxylation of carboxydabrafenib. The relative potency of dabrafenib and its metabolites ranked in the following order: dabrafenib4hydroxydabrafenib  desmethyl-dabrafenib 44 carboxy-dabrafenib. Daily oral administration of dabrafenib free base (30 mg/kg) for 5 days to mice bearing subcutaneous A375pF11s xenograft tumors resulted in relative plasma exposures (Cmax and AUC0-t) of carboxy-dabrafenib 44 parent 4 hydroxy-dabrafenib desmethyl-dabrafenib. Homogenization and extraction of liver, kidney, brain and tumor tissue showed the presence of dabrafenib and all three metabolites in liver, kidney and tumor. In brain, relative exposures were carboxy-dabrafenib 44 desmethyl-dabrafenib 4 parent (hydroxy-dabrafenib was not detected) as quantified by UHPLC/MS/MS, however mouse brain tissue was BQL for all analytes by MALDI-IMS. Excised xenograft tumor tissue contained 5-fold lower exposures of carboxy- dabrafenib versus plasma, but contained 5-fold higher exposures of desmethyldabrafenib versus plasma, while the dabrafenib exposures in tumor tissue were similar to plasma. These data suggest that the subcutaneous xenograft tumor and mouse brain tissues are depots for the pharmacologically active desmethyl-dabrafenib, whether by selective uptake or formation in situ via decarboxylation of carboxy-dabrafenib. Hydroxy-dabrafenib exposures were lower than dabrafenib in all mouse tissues tested except liver, however carboxy- and desmethyl-dabrafenib were the predominant mouse liver components. In mouse kidney, relative exposures were carboxy-dabrafenib desmethyl-dabrafenib44dabrafenib4 hydroxy-drafenib. MALDI-IMS analysis of the mouse kidney tissue illustrated that carboxy-dabrafenib was localized to the medulla, while desmethyl-dabrafenib was predominantly in the cortex. In rat, following a [14C]dabrafenib free base dose, QWBA showed dabrafenib drug-related material well distributed to tissues, with the notable exception of brain, while radioprofiling showed homogenized rat liver contained predominantly hydroxy-dabrafenib. Using in vitro methodologies, dabrafenib, hydroxy-dabrafenib and desmethyl-dabrafenib were demonstrated to be substrates for human Pgp, an efflux transporter at the blood brain barrier; therefore, penetration into human brain may be restricted. Carboxy-dabrafenib was not a substrate for Pgp; therefore, may have the highest potential for brain penetration, despite its hydrophilic nature. Together these studies illustrate the distribution of dabrafenib and its main metabolites across tissue type and species.

P114. ABSTRACT WITHDRAWN

P115. PHARMACOKINETICS, METABOLISM, AND BILIARY EXCRETION OF THE INOSINE 50 -MONOPHOSPHATE DEHYDROGENASE INHIBITOR AS2643361 Kazuhiro Tetsuka1, Nicolas Gerst2, Tomonori Nakanishi2, Toshiko Yamada-Kubo2, Kouichi Tamura2, and Jeffrey N. Masters2 1 Analysis & Pharmacokinetics Research Labs., Astellas Pharma Inc., Tsukuba-shi, Japan and 2Astellas Research Institute of America LLC, Skokie, IL, USA Purpose: We previously reported that the novel inosine 500 -monophosphate dehydrogenase (IMPDH) inhibitor AS2643361 exhibits more potent immunosuppressive activity and less gastrointestinal (GI) toxicity in rats than mycophenolic acid mofetil (MMF), a pro-drug of another IMPDH inhibitor of mycophenolic acid (MPA) (Nakanishi et al., 2012). Here, we further characterized the pharmacokinetics, metabolism, and biliary excretion of AS2643361. Methods: Total plasma clearance (CLtot) in rats was calculated from the plasma concentration of AS2643361 following intravenous (i.v.) administration, and bioavailability (BA) was calculated from the plasma concentration following i.v. and oral administration. Biliary excretion (%) in rats was estimated by measuring the biliary concentration of AS2643361 following intravenous administration. In vitro cytochrome P450 (CYP) metabolism and glucuronidation was assessed in small intestinal and liver microsomes of rats and humans, and in vitro biliary excretion and canalicular efflux were evaluated in rat and human sandwich-cultured hepatocytes (SCHs). Results: Pharmacokinetic findings of AS2643361 were as follows: CLtot, 17.7 mL/min/kg; BA, 28%; and biliary excretion, 0.13%. CYP metabolism of AS2643361 was observed in small intestinal and liver microsomes of rats and humans, but glucuronidation was not. This finding contrasted with that of MPA, which exhibited glucuronidation to 7-O-glucuronide (MPAG). Biliary excretion of AS2643361 in rat SCHs was not observed in vitro, which was consistent with its markedly low biliary excretion in vivo. In contrast, a significant biliary excretion of MPAG after MPA metabolism in rat SCHs was detected, consistent with previous in vitro and in vivo findings. Similarly, canalicular efflux of AS2643361 in human SCHs was not

94

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

observed while that of MPAG following glucuronidation from MPA was confirmed, consistent with our previous report.2) Conclusions: The present study demonstrates that CYP metabolism, not glucuronidation or biliary excretion, is the major disposition pathway of AS2643361 in rats. This profile is in contrast to that of MMF, whose active form MPA undergoes enterohepatic circulation via cycles of glucuronidation to MPAG, biliary excretion, deconjugation to MPA in the GI tract, and reabsorption. In addition, this profile might partially account for the GI toxicity of AS2643361 being lower than that of MMF in rats, which is supported by its reduced contact with the GI tract. Given that human liver microsomes exhibit CYP metabolism but not glucuronidation or canalicular efflux in vitro, these findings might be applicable to a clinical setting.

References Nakanishi T, Kozuki Y, Eikyu Y, et al. (2012). In vitro and in vivo characterization of AS2643361, a novel and highly potent inosine 50 monophosphate dehydrogenase inhibitor. Eur J Pharmacol 674:58–63. Tetsuka K, Gerst N, Tamura K, Masters JN. (2014). Species differences in sinusoidal and canalicular efflux transport of mycophenolic acid 7-O-glucuronide in sandwich-cultured hepatocytes. Pharmacology Research and Perspectives 2:e00035. doi:10.1002/prp2.35.

P116. INVESTIGATION OF THE BIODISTRIBUTION AND METABOLISM OF A TRITIUM-LABELED ANTISENSE OLIGONUCLEOTIDE IN THE MOUSE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Andrew McEwen, Stuart G. Wood, Claire Henson, Stephen Harris, and Kathryn Sweeney Quotient Bioresearch Limited, Rushden, UK The potential use of oligonucleotides as therapeutic agents continues to elicit a great deal of interest. A key activity in the development of these molecules is a basic understanding of their pharmacokinetic properties and biodistribution. The gold standard for studies of adsorption, distribution, metabolism and excretion (ADME) remains the use of radiolabelled drugs. Small molecule ADME studies have traditionally utilised carbon-14 labelled materials because of the perceived stability of the label, whereas, the tritium isotope has often been dismissed as inappropriate. Oligonucleotides have been labeled with a variety of non standard isotopes: 18F, 32P and 111In amongst others. The aim of the current work was to revisit the use of tritium as a convenient radiolabel for oligonucleotides. For this study a [3H]-labelled oligonucleotide was prepared by Quotient Radiochemicals. The radiolabelled test compound (50 -ACAGCUCC[3H](U)GACACCUCUCdTdTdT-30 ) was supplied at a specific activity of 5.9 Ci/mmol and a purity of 92%. The procedure involved synthesis of a brominated version of the oligonucleotide followed by tritium–bromine exchange. The [3H]-labelled oligonucleotide was supplied as a solution in ammonium acetate:acetonitrile (9:1 v/v) and was prepared for dosing by dialysis in phosphate buffered saline (PBS) for ca 18 hours and mixed with non-radiolabelled material and PBS in order to achieve the desired concentration and specific activity for dosing. Female CD-1 mice were dosed intravenously with the radiolabelled oligonucleotide at 5 mg/kg. Blood spot samples were taken at 0.5, 1, 2, 4, 6 and 24 hours post dose to determine the pharmacokinetic properties of the test material. Blood spot samples were combusted using a Packard Oxidizer model 307 and the 3H2O produced mixed with scintillant, MonoPhase S prior to liquid scintillation counting (Packard 2300TR Scintillation Counter). Blood versus time data was analysed using PC ModFit (Version 4.0) and the kinetic data characterised by a non-compartmental analysis (NCA). Animals for determination of tissue distribution were killed at 1, 3, 6 and 24 hours post dose. Carcasses were subjected to whole-body autoradiography using procedures based on the work of Ullberg (Acta. Radiol. Suppl. 118, 22–31, 1954) and sections obtained using a Leica CM3600 microtome (Leica Microsystems). Distribution of radioactivity was determined and quantified using a Fuji FLA-5100 image analyser and associated Tina and SeeScan Software. Tissue concentrations of radioactivity were determined using a standard curve produced using calibrated whole blood standards. The data obtained in this study demonstrates that good quality pharmacokinetic and tissue distribution data can be obtained following radiolabelling of oligonucleotides with tritium. The properties of the tritium label (high activity, short path length) provide quantification of tissue concentrations with excellent spatial resolution. The tritium label also enables cellular distribution to be determined using microautoradiography.

P117. ABSTRACT WITHDRAWN

P118. ASSESSMENT OF BLOOD-BRAIN BARRIER CROSSING USING MSI: NEW PREDICTIVE TOOLS FOR CNS TARGETED DRUG EFFICACY STUDY Gregory Hamm1, Fabien Pamelard1, David Bonnel1, Raphael Legouffe1, Guillaume Hochart2, and Jonathan Stauber1 1 Imabiotech, Loos, France and 2MS Imaging Department, Imabiotech, Loos, France Introduction: The development of drug targeting CNS is one of the most challenging for pharmaceutical industries. In fact, the main limitation concerns the drug ability to cross the BBB. In this context, Mass Spectrometry Imaging is used to localize the drug near blood vessels in the brain and evaluate the permeability of BBB for the drug by following CNS positive and negative markers. We propose here a new tool to accurately assess the potential of BBB crossing by a drug which combines spatial and spectral information of markers and drug. Methods: Mice (C57 Black 6 strain) were dosed with two drugs, the Clozapine (Sigma) at 50 mg/kg (P.O.) and the Verapamil (Sigma) at 5 mg/kg and sacrificed 30 min post injection. A pg-p substrate, the cyclosporine

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

95

A (CsA) is used to modulate the properties of verapamil to cross the BBB and administrated 15 min before the others. Snap frozen brain cryosections were carried out with a Microm cryostat HM560 (Thermo Scientific) at 10 mm thick and mounted on several conductive ITO glass slides before matrix automated deposition. Mass spectra were acquired in positive detection mode using a SolariX FTICR 7.0T Mass Spectrometer (Bruker Daltonics). Preliminary results/abstract: The fine evaluation of bloodbrain barrier crossing by therapeutics using MSI needs reliable markers of tissue targeting. In one hand, heme b (m/z 616.1) is used on imaging data to follow blood vessels and vasculature networks in the brain, as a CNS negative marker. In the other hand, the clozapine (m/z 327), an anti-psychotic agent well known to cross the BBB and targeted some specific neurotransmitters acts as CNS positive markers. In the present study, we evaluate the crossing of the blood-brain barrier by the verapamil with or without pg-p substrate action. MALDI-MS images were obtained at low (100 mm) and high (20 mm) spatial resolution to show the distribution of the molecules of interest in fine histological region of the brain. As expected, Clozapine was observed homogenously in the brain and heme accumulates specifically in discrete voxels corresponding to blood vessels region. Concerning the verapamil, its distribution is less specific than for heme and less homogenous than for clozapine. Finally, we can easily generate a scale of blood-brain barrier crossing potential where the verapamil takes place between our two markers. Novel Aspect: A new way to consider MSI data is presented to go beyond the basic visual interpretation for CNS study.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P119. COMPUTER ASSISTED METABOLITE IDENTIFICATION, APPLICATION in GSH TRAPPING Esra Nurten Cece Jr1, Fabien Fontaine Sr2, Guillem Plasencia Sr2, Marieke Teppner Jr3, Andreas Brink Sr3, Axel Paehler Sr3, and Ismael Zamora Sr4 1 Department of Experimental and Health Sciences, University Pompeu Fabra, Barcelona, Spain, 2Department of Molecular Discovery, London, UK, 3F. Hoffmann-La Roche Ltd., Basel, Switzerland, and 4Lead Molecular Design, S.L., San Cugat del Valle´s, Spain Chemically reactive metabolites have been implicated in the potential toxic response for new chemical entities. One way to address this issue is to perform in vitro metabolism studies with glutathione (GSH) as nucleophile in the media, in order to trap any thiol-reactive electrophilic metabolite. Such screening assays have been implemented by most pharmaceutical companies as part of the toxicity screening cascade during the discovery. High Resolution Mass Spectrometry (HRMS) can effectively analyze the samples typically used in GSH trapping studies. The methodology used in this work utilizes Human Liver Microsomes (HLM) incubations where GSH is added in excess to the drug at a substrate concentration of 20 mM. Samples were analyzed by LC-HRMS on a Waters UPLC system coupled to an AB Sciex TripleTOFÔ 5600 mass spectrometer. A generic, non-targeted LC-HRMS method was applied to a set of 95 drugs, 57 of which have been reported to form GSH adducts and 38 without reports of GSH adduct. Post-acquisition data analysis was performed with a new algorithm to detect chromatographic peaks as first step. As second step in the analysis, potential drug-related GSH adducts were identified based on GSH related fragment ions and the loss of constant neutral fragments (NL) using the computer-assisted MassMetaSite procedure. Lastly, structural elucidation for each of the GSH adducts was performed. All data (chromatograms, spectra, GSH adduct structures; definitive and markush) are captured in a searchable database and mining tools within WebMetabase could be used, e.g. SAR analysis (structure–activity relationships), identification of problematic chemical motifs or to compare GSH related metabolic pathways for a large compound set. The GSH adducts identified by MassMetaSite were compared with a manual inspection of the HRMS data and extraction of the data by a DMPK expert. All peaks that were reported manually were also shown in the computational driven approach. According to the results, 41 compounds were manually identified as forming reactive metabolites while the computer-assisted MassMetaSite automated approach identified GSH adducts for 59 compounds. Based on the results of this study only about 43.5% of all GSH adducts identified by MMS showed the NL of 129 Da (pyroglutamate) commonly applied for the detection of GSH adducts. Most of the GSH adducts that were not detected manually (28.2%) revealed the GSH fragment ion at m/z 308 and 28.2% the showed additional less common GSH related neutral losses or fragment ions. In conclusion, MassMetaSite in combination with WebMetabase is a powerful tool to process large datasets of analytical data to identify drug-related GSH adducts. The acquired structural information and biotransformation data enables the DMPK expert to translate this into knowledge for GSH adduct formation in a highly automated and accelerated fashion.

P120. DEVELOPMENT OF HAMINTM TESTOSTERONE TRANSDERMAL DELIVERY SYSTEM AND ITS IN VITRO AND IN VIVO EVALUATIONS Z. Chik1, Didi Emh1, Noordin MI2, and Mohamed Z1 1 Department of Pharmacology, University of Malaya, Kuala Lumpur, Malaysia and 2Department of Pharmacy, University of Malaya, Kuala Lumpur, Malaysia Background: Transdermal preparations becoming popular recently, because of their unique advantages such as convenience, avoidance of first-pass effect, and reduction of systemic side effects. Testosterone taken orally is rapidly inactivated by first-pass hepatic metabolism, making oral therapy an ineffective means of delivering testosterone especially for testosterone replacement therapy in the treatment of male hypogonadism. The objective of this study was to develop a new testosterone transdermal delivery system by using palm oil base called HaminTM and to evaluate its characteristics and effectiveness by using in vitro and

96

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

in vivoapproaches. Methods: 0.5% HaminTM Testosterone Transdermal Delivery System (TTDS) was formulated from palm oil base called HaminTM. Liquid chromatography mass spectrometry (LCMSMS) method was developed and validated to analyse testosterone in pharmaceutical formulation and blood. This method was later used in the evaluation of testosterone. The formulation was tested for its physical characteristics such as organoleptic, viscosity, uniformity content, and particle size. On the other hand, in vitro skin permeation studies were performed by using Franz diffusion cells. In vivo study was carried out in rats to further confirm the effectiveness of this testosterone transdermal delivery system. Results: The validated LCMSMS method was linear over the concentration range of 1–100 ng/mL. Organoleptic test of the cream showed the homogenous appearance with soft texture and viscosity measurement showed the pseudoplastic behaviour. The uniformity content was in the range of 90.90–115.00%. The particle size of the cream developed as observed by dynamic light scattering method was in the range of 97–7740 nm. The permeation study showed that the TTDS was able to deliver testosterone across the membrane with the cumulative amount diffused was up to 500 ng/cm2 after 5 hours application of TTDS which contained 5 mg of testosterone. In the in vivo study using rats, application of 0.5% TTDS resulted in an increased serum testosterone level and reached the mean Cmax and Tmax of 349.5 ng/mL and 0.25 hours, respectively. Conclusions: We have successfully developed and tested a novel TTDS which has all characteristics for transdermal delivery and successful in delivering testosterone across the membrane in in vitro and in vivo studies.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P121. A HIGHLY SENSITIVE AND SELECTIVE NEW TRAPPING REAGENT, STABLE ISOTOPE LABELED GLUTATHIONE ETHYL ESTER, FOR DETECTION OF REACTIVE INTERMEDIATE(S) Toshikazu Yamaoka1 and Yoshiaki Kitamura2 1 Development Research Laboratories, Kyorin Pharmaceutical Co., Ltd., Tochigi, Japan and 2Discovery Research Laboratories, Kyorin Pharmaceutical Co., Ltd., Tochigi, Japan Background: Despite extensive safety testing of drug candidates, extremely rare adverse events still occur after approval. To predict the post-marketing risk for idiosyncratic drug reactions (IDRs), glutathione (GSH)-trapping assays are widely used in the pharmaceutical industry. Although the mechanisms of IDRs are poorly understood, reactive metabolic intermediates may be involved in the onset of some IDRs. GSH derivatives e.g. glutathione ethyl ester (GSHEE) (Soglia et al., 2004) and stable isotope labeled GSH ([13C2,15N]GSH) (Ma et al., 2008), have been introduced as trapping reagents for reactive intermediates. However, more sensitive and selective reagents are being sought to help reduce the risk of IDRs. In this study, we designed a stable isotopelabeled GSHEE (GSHEE-d5) and examined its detection capabilities. Methods: GSHEE-d5 was synthesized as a new trapping reagent and tested with eight drugs reported to cause IDRs (positives: acetaminophen, diclofenac, clozapine, tienilic acid, imipramine, omeprazole, trovafloxacin, and mefenamic acid) and three control drugs (negatives: testosterone, midazolam, and terfenadine). The trapping reagent and [13C2,15N]GSH as a reference were incubated with each positive and negative drug in human liver microsomes. The resulting adducts between the reagents and reactive metabolites were concentrated by solid-phase extraction and analyzed with an ultra-performance liquid chromatograph coupled with a linear ion trap mass spectrometer (UPLC/MS). The labeled reagents allowed detection of the adducts with isotope pattern-dependent scanning and neutral loss filtering. When each of the labeled reagents was analyzed in a 1:1 mixture with non-labeled GSHEE or GSH, they gave a characteristic isotopic doublet with a 5- or 3-Da difference, respectively. Results: A one-step reaction of GSH with ethanol-d6 produced GSHEE-d5 (yield, 85%). GSHEE-d5 formed adducts with all eight positive drugs and no adducts with any of the three negative drugs, whereas [13C2,15N]GSH failed to form adducts with three of the positive drugs. In the case of diclofenac, GSHEE-d5 showed about 10 times greater sensitivity (signal/noise ratio) than [13C2,15N]GSH. Conclusions: A new trapping reagent, GSHEE-d5, was easily synthesized from GSH in a one-step reaction. A combination of this reagent and UPLC/MS will allow rapid analysis of new chemical entities (11 min/sample) with greater sensitivity and selectivity than current methods. In conclusion, GSHEE-d5 may be a useful tool to help evaluate the risk of IDRs from drug candidates in early stages of drug development.

References Ma L, Wen B, Ruan Q, Zhu M. (2008). Rapid screening of glutathione-trapped reactive metabolites by linear ion trap mass spectrometry with isotope pattern-dependent scanning and postacquisition data mining. Chem Res Toxicol 21:1477–1483. Soglia JR, Harriman SP, Zhao S, et al. (2004). The development of a higher throughput reactive intermediate screening assay incorporating micro-bore liquid chromatography-micro-electrospray ionization-tandem mass spectrometry and glutathione ethyl ester as an in vitro conjugating reagent. J Pharm Biomed Anal 36:105–116.

P122. DEVELOPABILITY ASSESSMENT FOR A NOVEL AMIDE-CONTAINING NICOTINAMIDE PHOSPHORIBOSYLTRANSFERASE (NAMPT) INHIBITOR Bianca M. Liederer1, Leslie Wang1, Justin Q. Ly1, Kang-Jye Chou2, Peter S. Dragovich3, Xiaorong Liang1, and Lulu X. Yang1 1 Department of Drug Metabolism and Pharmacokinetics, Genentech Inc., South San Francisco, CA, USA, 2Department of Pharmaceutics, Genentech Inc., South San Francisco, CA, USA, and 3Department of Discovery Chemistry, Genentech Inc., South San Francisco, CA, USA

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

97

GNE-618 is a potent, selective NAMPT (nicotinamide phosphoribosyltransferase) inhibitor, intended as an oral therapy for human cancers. In this study, its developability is assessed from both formulation and DMPK perspectives. GNE-618 is a crystalline solid, for which a low melting polymorph (LMP) and a high melting polymorph (HMP) were identified. The HMP showed good solid stability but lower solubility than the LMP, except when PEG was the solvent. No viable salts could be identified because of its low pKa values. A co-solvent formulation, 60% PEG 400/10% ethanol/30% D5W, was used for initial preclinical PK studies, and dose-limiting exposure was observed, likely due to poor solubility. A comparison of the pharmacokinetics of GNE-618 using various formulations was conducted in rats. Capryol/PEG400, a type IV lipid formulation, showed improved oral exposure over crystalline suspension and co-solvent formulations. Amorphous solid dispersion did not provide much solubility advantage because of quick recrystallization of the compound. Oral bioavailabilities in preclinical species were acceptable (64.3, 38.4, 74.5, 124% for mice, rats, monkeys and dogs, respectively). Plasma clearance values were low in mice, rats, monkeys and dogs (1.70, 5.51, 0.49, and 2.37 mL/min/kg, respectively), and the clearance category was well-predicted by hepatocytes. Transporter studies in MDCKI-MDR1 and MDCKII-hBCRP cells suggested that GNE-618 is likely a substrate for both MDR1 and BCRP. The volume of distribution ranged from 0.31 in monkeys to 1.38 L/kg in rats. Plasma protein binding was high (fu = 0.013 to 0.027). Allometric scaling predicted a low clearance and a volume of distribution of 0.7 L/kg in human. Overall, GNE-618 exhibited acceptable preclinical properties and projected human PK estimates. Oral exposure using type IV formulations were acceptable, suggesting potential for development as an oral anticancer drug.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P123. ABSTRACT WITHDRAWN

P124. STRATEGIC COMPOUND SELECTION AND HUMAN PHARMACOKINETIC PREDICTION OF CALCIUM SENSING RECEPTOR ANTAGONIST FOR TREATMENT OF OSTEOPOROSIS Ryo Matsuyama1, Etsuko Fujita1, Yuri Sakai1, Yohei Matsueda1, Masaru Ebisawa2, and Kazuoki Tsuruda1 1 Teijin Institute for Bio-medical Research, Teijin Pharma Limited, Hino Tokyo, Japan and 2Pharmaceutical Research Planning Department, Teijin Pharma Limited, Hino Tokyo, Japan It is known that the intermittent administration of parathyroid hormone (PTH) increases bone mass, but the continuous infusion of PTH increases serum calcium and induces bone resorption. Teriparatide (Exogenous recombinant PTH) achieves intermittent PTH profile and is approved for the treatment of osteoporosis. The endogenous PTH secretion is regulated by calcium sensing receptor (CaSR) expressed in the parathyroid gland, and CaSR antagonists which stimulate transient PTH secretion are hoped that the drug for the treatment of osteoporosis. The objective of this study was to find promising CaSR antagonist compounds for the treatment of osteoporosis. As a strategy for achieving CaSR-mediated transient PTH secretion, we defined the target pharmacokinetic (PK) profile of CaSR antagonist as rapid absorption and rapid elimination with orallyavailable potential. The rapid absorption property was examined solubility and Caco-2 cell permeability assay in vitro. The rapid elimination property was assessed in rat PK study following intravenous administration. The pharmacodynamics studies were also assessed in rat following oral administration. As a result of early ADME screening assay, two compounds indicated good in vitro and in vivo PK profiles which showed good solubility and permeability, no CYP inhibition, and sharp PK profile with orally-available potential in rat. In rat PD study, these compounds indicated more transient PTH profiles and less influence on serum Ca level than reference compound. These PTH profiles correlated with sharp PK profiles, which were earlier time of maximum concentration and more rapid elimination from plasma than those of reference compound. Human volume of distribution and clearance were predicted from animal data using allometric methods recommended by the PhRMA CPCDC initiative on predictive models of human pharmacokinetics (Jones et al., 2011; Ring et al., 2011). Predicted human half-life of these two compounds ranged from 0.37 to 1.39 h and 0.35 to 1.91 h. These preclinical data were expected that the PK of the CaSR antagonist compounds would be characterized by rapid absorption (good solubility and permeability) and rapid elimination (half-life of approximately 1 h), resulting in transient PTH secretion in human.

References Jones RD, Jones HM, Rowland M, et al. (2011). PhRMA CPCDC initiative on predictive models of human pharmacokinetics. Part 2: Comparative assessment of prediction methods of human volume of distribution. J Pharm Sci 100:4074–4089. Ring BJ, Chien JY, Adkison KK, et al. (2011). PhRMA CPCDC initiative on predictive models of human pharmacokinetics. Part 3: Comparative assessment of prediction methods of human clearance. J Pharm Sci 100:4090–4110.

P125. ABSTRACT WITHDRAWN

98

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P126. DEVELOPMENT OF A PHOTOMETRIC KINETIC GLUTATHIONE REACTIVITY ASSAY FOR SCREENING COVALENT WARHEADS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Nickolas A. Neitzel1, Brian R. Baer1, James F. Blake2, Mark L. Boys3, Mark J. Chicarelli3, Jay B. Fell3, Erik J. Hicken3, Ellen R. Laird2, Michael A. Lyon3, Francis X. Sullivan4, and Tony P. Tang3 1 Department of Drug Metabolism and Pharmacokinetics, Array BioPharma Inc., Boulder, CO, USA, 2Department of Computational Research, Array BioPharma Inc., Boulder, CO, USA, 3Department of Medicinal Chemistry, Array BioPharma Inc., Boulder, CO, USA, and 4Department of Biology, Array BioPharma Inc., Boulder, CO, USA Off-target reactivity of a covalent inhibitor is traditionally assayed by incubating an electrophilic compound with glutathione (GSH) and, using various techniques, monitoring the loss of compound or GSH. The goal is to rapidly determine the extent to which a compound might react with endogenous thiol-containing biological nucleophiles, thereby providing drug discovery scientists with a high-throughput screen to mitigate potential toxicological consequences associated with a therapeutic electrophile. It is hypothesized that attributes common to most GSH reactivity assays, such as reagent selection, assay duration, and temperature, inadvertently preclude the differentiation of more desirable low-to-medium reactivity compounds. In response, an assay was designed to quantify and rank-order constituents of Array’s electrophilic ‘‘warhead’’ library according to their pseudo-first-order rate of reactivity with GSH (kGSH). In an optimized method, 500 mM warhead and 140 mM GSH were incubated together in buffer for up to 24 hours at 37  C. Reactions were quenched at several time-points with 4,40 dithiodipyridine (DTDP), and remaining GSH was quantified using spectrophotometric detection of 4-thiopyridine (4-TP). During the development of the assay, 5,50 -dithio-bis(2-nitrobenzoic acid) (DTNB; Ellman’s Reagent) was replaced with DTDP for quenching reactions and DMSO was substituted with acetonitrile for solvating warheads due to cross-reactivity with integral assay components. Also, temperature was increased from room temperature to 37  C and assay duration was increased from 2 to 24 hours to promote further delineation of low-to-medium reactivity warheads. Of Array’s 711 available warheads, 61 were considered highly reactive (4 1000 hr1), 76 were considered to exhibit medium reactivity (100: 1000 hr1), and 224 were deemed to have low reactivity (5: 100 hr1). Of the remaining 350 warheads, 306 exhibited no measurable activity (55 hr1) and 44 were dropped from the analysis due to issues with solubility or absorbance. Values of kGSH were compared to values of electrophilicity calculated in silico and the correlation was found to be poor (r250.010), indicating that either the chosen model was inadequate for determining warhead reactivity across a broad structural spectrum or warhead reactivity was best ascertained empirically. The kGSH data were also compared with data collected in a protein modification LC-MS assay, in which the extent of modification of specific cysteine residues on a target protein was monitored. This comparison identified warheads that exhibited low reactivity towards endogenous thiols and were inherently selective for target cysteines. Combining these selective covalent warheads with non-covalent, target-specific inhibitors, identified via traditional SAR and protein crystallography, offers a new approach to inventing highly selective covalent inhibitors.

P127. LOW LEVEL RADIOACTIVE COUNTING: MAKING THE MOST OF YOUR VALUABLE SAMPLES Andrew McEwen, Gill Ford, Sebastian Lopez, and Stuart G. Wood Quotient Bioresearch Limited, Rushden, UK Determination of the comparative metabolism in pre-clinical species and man remains an essential component of the drug development process. Whilst early metabolic information can be obtained using ‘‘label free’’ techniques such as mass spectrometry, data obtained in this manner is at best semi quantitative in nature and is generally based on comparative ‘‘response factors’’. For chromatographic data generated during early and late stage development to be useful, accurate methods of metabolite quantification are required. The separated chromatographic peaks can be subjected to structural analysis using techniques such as mass spectrometry or NMR. By linking the structural data with the chhromatography data, a metabolism pathway can be elucidated on a quantitative level. The gold standard approach remains the use of radiolabelled compounds. The current FDA and ICH guidelines designed to address the safety of drug metabolites (commonly referred to as Metabolites in Safety Testing (MIST)) have placed the emphasis upon obtaining comparative data on circulating metabolites in both human and animal studies as early as possible. In pre-clinical studies the amount of radioactivity that can be administered is not limited, whereas in the clinical setting the dose is generally limited, on safety grounds, to around 100 mCi. In a standard human mass balance study the concentrations of radioactivity in the urine and faeces are generally sufficient to obtain useful information on the nature of the excreted components. By contrast, concentrations of circulating components can be extremely low in plasma and the analysis is further complicated by the low sample volumes usually obtained. High-performance liquid chromatography (HPLC) coupled with online radiodetection is a regularly used method for quantification and identification of metabolites. The method has a relatively high limit of detection due to the low residence time in the detector cell. It is important, therefore, to conssider alternative methods with a lower limit of detection and higher sensitivity. In light of this a comparative investigation was carried out into the limits of detection of offline radiodetection methods coupled with HPLC, such as, LSC, MSC and phosphorimaging. Phosphorimaging shows promise as a technique for reducing the sample quantity required and improving sensitivity.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

99

P128. ABSTRACT WITHDRAWN

P129. ALBUMIN FUSION RENDERS THIOREDOXIN AN EFFECTIVE ANTI-OXIDATIVE AND ANTI-INFLAMMATORY AGENT FOR PREVENTING ACUTE KIDNEY INJURY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Azusa Kodama1, Hiroshi Watanabe1, Ryota Tanaka1, Yu Ishima1, Masafumi Fukagawa2, Masaki Otagiri3, and Toru Maruyama1 1 Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan, 2Division of Nephrology, Endocrinology and Metabolism, Tokai University School of Medicine, Kanagawa, Japan, and 3Faculty of Pharmaceutical Sciences Sojo University, Kumamoto, Japan Purpose: Acute kidney injury (AKI) is closely associated with increased morbidity and mortality. Drug-induced AKI accounts for about 20% of hospital- and community-acquired AKI. Therefore, a strategy for preventing drug-induced AKI due to enhanced oxidative stress and inflammatory response is highly desirable. Thioredoxin-1 (Trx), an endogenous redox-active protein, has a short retention time in the blood. A long acting form of Trx, human serum albumin-Trx (HSA-Trx), was produced by recombinant HSA fusion and its effectiveness in preventing drug-induced AKI (contrast-induced nephropathy (CIN) and cisplatin-induced nephropathy) was examined. Methods: The genetic fusion of Trx and HSA was performed and the fusion protein was prepared in Pichia expression system. A CIN model was established by administering indomethacin (10 mg/kg), followed at 15 and 30 min, respectively, by N-nitro-L-arginine methyl ester (10 mg/kg) and ioversol (contrast media: 8.3 mL ioversol/kg) to rats. Cisplatin-induced nephropathy model mouse was established by a single administration of cisplatin (15 mg/ kg). Results: Compared with saline, a mixture of HSA and Trx, or Trx alone, intravenous HSA-Trx pretreatment significantly attenuated the ioversol-induced elevation in renal damages, such as serum creatinine, blood urea nitrogen and urinary N-acetylb-D-glucosaminidase along with a decrease in creatinine clearance. In addition, HSA-Trx caused a substantial reduction in the histological features of renal tubular injuries and in the number of apoptosis-positive tubular cells. HSA-Trx treatment also attenuated the cisplatin-induced renal dysfunction and histologic alterations such as renal tubular injuries and apoptosis-positive tubular cells. Changes in oxidative stress markers (superoxide, 8-OHdG, malondialdehyde, glutathione and nitrothyrosine levels) clearly indicated that HSA-Trx significantly suppressed renal oxidative stress in both model mice. Moreover, HSA-Trx showed anti-inflammatory action as evaluated by the marked suppression of the elevation of TNF-a, IL-1b and IL-6 levels in cisplatininduced nephropathy. Administered fluorescein isothiocyanate-labeled HSA-Trx was found partially localized in the proximal tubular cells whereas majority remained in the blood circulation. Specific cellular uptake and the scavenging of intracellular reactive oxygen species, which was induced by hydrogen peroxide or cisplatin, by HSA-Trx were observed in human proximal tubular cells (HK-2 cells). Conclusion: HSA-Trx could be a novel and effective approach for preventing drug-induced acute kidney injury due to its prolonged anti-oxidative and anti-inflammatory action not only in extracellular compartment but also inside the proximal tubular cell.

P130. ABSTRACT WITHDRAWN

P131. GENETICALLY ENGINEERED ALBUMIN-THIOREDOXIN FUSION PROTEIN, LONG-ACTING ANTI-OXIDANT, AMELIORATES ACUTE LUNG INJURY ASSOCIATED WITH INFLUENZA VIRUS INFECTION Toru Maruyama1, Ryota Tanaka1, Hiroshi Watanabe1, Azusa Kodama1, Yu Ishima1, and Masaki Otagiri2 1 Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan and 2Faculty of Pharmaceutical Sciences Sojo University, Kumamoto, Japan Purpose: Infection of Influenza virus is a leading cause of human illness and death throughout the world. Especially, acute lung injury is an important cause of death in humans infected influenza virus, including H1N1, H5N1. It has been found that an excessive immune response induced by reactive oxygen species (ROS) and reactive nitrogen species (RNS) play a crucial role for the onset and development of influenza-mediated acute lung injury. The purpose of this study is to evaluate the therapeutic impact of HSA-Trx, a long-acting anti-oxidant, against acute lung injury induced by influenza A H1N1 virus infection. Methods: Genetically engineerd HSA-Trx was produced using Pichia pastoris systems as same as commercially available recombinant HSA preparation. Mice were infected with influenza A (H1N1) virus (1.5  LD50) via the transtracheal route. Results: First, to evaluate the validity of a mouse model used in this study, oseltamivir, a standard therapy of influenza virus infection, was used as a positive control. It was administrated intraperitoneally 2 days after the influenza virus infections, and was continued to five consecutive days as same as a therapy in human. As expectedly, total protein level in bronchoalveolar lavage fluid (BALF) of model mice was significantly decreased by the oseltamivir treatment. Similary, oseltamivir attenuated the lung damages as evaluated by HE staining. Moreover, the total number of neutrophils and the total protein levels in BALF were significantly elevated and the total amounts of virus in lung were maximized at day 4 after infection. These data confirmed that a mouse model used in this study is able to use for the evaluation of therapeutic efficacy of HSA-Trx against influenza viral pneumonitis.

100

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

HSA-Trx significantly improved the survival rate (470%) of mice 10 days after virus infection, whereas more than 80% of mice were died in saline treated group. Similarly, HSA-Trx significantly reduced the lung damages in influenza virus infected mice as judged by the decrease in the total protein levels in BALF and HE staining of lung tissues. Under such a circumstances, HSA-Trx also significantly inhibited the total number of neutrophils in BALF, the accumulation of 8-OHdG and peroxynitrite in lung tissue of model mice. In contrast, HSA-Trx did not affect the total number of macrophage cells and the levels of IFN-gamma in BALF, expression of iNOS in lung. Moreover, in contrast to oseltamivir, HSA-Trx did not affect the amount of influenza virus in lung. These data indicate that HSA-Trx ameliorates the lung damages associated with influenza virus infection due to the effective inhibition of the ROS production. Conclusion: Present study demonstrated that HSA-Trx has a potential as a resue therapy against acute lung injury induced by exaggerated host immunological responsiveness during the infection of influenza virus through its superior anti-oxidative action.

P132. DISEASE IN A DISH: PATIENT-SPECIFIC hiPSC-DERIVED CARDIOMYOCYTES TO STUDY LONG QT INTERVAL AND hERG CHANNEL TOXICITY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Daniel A. Ryan1, Karl J. Okolotowicz1, Wesley L. McKeithan2, Alex Savtchenko2, Robert S. Kass3, Mark Mercola2, and John R. Cashman1 1 The Human BioMolecular Research Institute, San Diego, CA, USA, 2Muscle Development and Regeneration Program, Sanford-Burnham Medical Research Institute, La Jolla, CA, USA, and 3Department of Pharmacology, College of Physicians and Surgeons, Columbia University, New York, NY, USA In the United States an estimated 850 000 adults are hospitalized for cardiac arrhythmias each year, making arrhythmias one of the top five causes of healthcare expenditures with a direct cost of more than $40 billion annually for diagnosis, treatment and rehabilitation. Atrial fibrillation (a common type of arrhythmia) affects millions of people and the incidence is rising. However, some medicines used to treat arrhythmias cause arrhythmias as a side effect. Deleterious interactions of drugs and drug candidates with sodium and potassium channels in cardiomyocytes are primarily responsible for the adverse effects and toxicity. The Hypothesis of our work is that Disease in a Dish approaches can be used to develop more effective and safer drugs useful for heart disorders. The Methods use patient-derived hiPSC-cardiomyocytes to distinguish on-target and off-target effects of drugs or drug-candidates on cardiac ion channels. This research and product development provides a novel method for developing drugs for treating human cardiac arrhythmias by targeting the late Na+ channel current that is the on-target effect of the drug candidate. Certain patients are genetically predisposed to a potentially fatal arrhythmogenic response to existing drugs used to treat LQT3 because the drugs also have off-target effects on other important ion channels in cardiomyocytes. We will use patient-derived hiPSC-cardiomyocytes to develop a safer drug development candidate that will retain efficacy against the "leaky" Na+-channel yet minimizes off-target effects, including the K+ hERG channel that can be responsible for a proarrhythmic effect. Because this problem is thought to be exacerbated in patients with the common KCHN2 variant, K897T (with a frequency of 33% of the white population), removing the off-target liability addresses a serious unmet clinical need. The Results show that using a Kinetic Imaging Cytometer (Vala Sciences, Inc.) for evaluation of the effect of drug candidates on Na+- and K+-channels leads to efficient screening of analogs of existing drugs (i.e. drug rescue). The proposed idea is that the path from patient-specific hiPSCs to clinic might be easier than for a completely new chemical entity. In addition, an appealing aspect is that the hiPSCs can be derived from a patient to test individualized therapy. In conclusion, we have developed a Disease in a Dish approach to drug development for cardiac arrhythmias and aim to produce a better drug for this important health problem. The work was financially supported by Grant Number TR4-06857 from the California Institute of Regenerative Medicine.

P133. ATP METABOLISM IN RBC AS SYSTEMIC BIOMARKER FOR CARDIOVASCULAR TOXICITY Pollen K. F. Yeung1,2, Chad Purcell2, Patrick Tilman2, Shyam Sundar2, and Sarah Gallant2 1 Department of Medicine, Dalhousie University, Halifax, NS, Canada and 2College of Pharmacy, Dalhousie University, Halifax, NS, Canada Although clinical drug development has made significant stride along with pharmaceutical sciences over the last 3 decades, from the application of pharmacokinetics in the 1970s, controlled clinical studies for efficacy in 1980s, pharmacodynamics and pharmacogenetics in the 1990s, to a focus on drug safety in the past decade, the success rate to introduce new effective and safe therapeutic agents has not kept up with expectations from the financial investment and those of patients. In another word, there is inadequate improvement in drug therapy or financial reward. Identification and application of biomarkers for lead selection and optimization has been heralded as one of the most likely scientific approach to increase the success of drug development. It is widely conceived that biomarker is a scientific basis and an effective tool for disease management and personalized medicine. We have shown in a rat model that cardiovascular injury induced by an isoprotereneol injection (30 mg/kg) caused significant mortality and increased break down of ATP in red blood cell (RBC) in vivo. Those did not survive from the injury had significantly higher RBC concentrations of AMP. We have also shown that exercise increased RBC concentrations of ATP which may be a key factor for post-exercise hypotension. Exercise also reduced mortality and RBC concentrations of AMP in the acute myocardial infraction model induced by isoproterenol. Further, the calcium channel antagonist diltiazem which lowers blood

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

101

pressure and heart rate was shown to have a dose dependant attenuating effect on the cardiovascular injury induced by isoproterenol. The presentation will focus on the potential of using ATP metabolism as biomarker target for cardiovascular toxicities and regulation of cardiovascular homeostasis. It will also discuss the opportunities, challenges and obstacles of exploiting ATP metabolism in the RBC as targets for drug development and personalized medicine. The research was supported in part by Canadian Institute of Health Research (CIHR), Nova Scotia Health Research Foundation (NSHRF) and Dalhousie Pharmacy Foundation.

P134. POLYTHIOLATED- AND RECOMBINANT MANNOSYLATED-ALBUMIN AS A NOVEL CD68+/CD206+ KUPFFER CELL-TARGETED NANO-ANTIOXIDANT FOR THE TREATMENT OF ACUTE AND CHRONIC HEPATITIS MODELS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Hitoshi Maeda1, Shota Ichimizu1, Hiroshi Watanabe1, Yu Ishima1, Ayaka Suenaga1, Masaki Otagiri2, and Toru Maruyama1 1 Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan and 2Faculty of Pharmaceutical Sciences, Sojo University, Kumamoto, Japan Purpose: Reactive oxygen species (ROS) derived from Kupffer cells (KC), especially CD68+, play a key role in the development of hepatic oxidative stress and injuries. To examine the therapeutic impact of polythiolated- and recombinant mannosylatedhuman serum albumin (SH-Man-HSA), a novel nano-antioxidant delivering thiols (SH) to the CD68+ KC that also expressed mannose recepror (CD206), against concanavalin-A (Con-A)-induced acute and chronic hepatitis models because it was widely used as a model of various immune related hepatopathy, such as autoimmune hepatitis and viral hepatitis C. Methods: Man-HSA was prepared in yeast expression system. SH-Man-HSA was further synthesized using 2-iminothiolane and its modification efficacy was approximately 7.5 mol SH/mol. Mice received an intravenously injection with Con-A at a dose of 0.25 mg/mouse (single) or 0.3 mg/mouse (once a week) for 6 weeks, respectively. Results: In vitro EPR measurement and pharmacokinetic analysis revealed that SH-Man-HSA possessed powerful radical scavenging activity and was rapidly and selectively delivered SH to the liver via CD206 presented on the CD68+ KC surface. SH-Man-HSA (20 mmol SH/kg) significantly improved the survival rate of both leathal Con-A (1.0 mg/mouse) and repeated Con-A (0.3 mg/mouse) treated mice. Moreover, hepato-protective effect of SH-Man-HSA against Con-A (0.25 mg/mouse) treated models was superior to N-acetyl-cysteine (NAC), Man-HSA, SH-HSA, due to the remarkable inhibition of hepatic oxidative stress. These data suggest that both the SH and mannose groups are essential for the therapeutic effect of SH-Man-HSA. Flow cytometric analysis further confirmed that Con-A induced in an increase of F4/80+/ROS+ cells was dramatically decreased by SH-Man-HSA. The administration of SH-Man-HSA at 4 hours following a Con-A injection exhibited profound hepato-protective action against experimental hepatitis model, while this was not observed in the case of NAC. SH-Man-HSA also suppressed the hepatic injury and fibrosis in a chronic hepatopathy induced by the repeated Con-A (0.3 mg/mouse) injections without the changes in the mRNA levels of factors that associated with the chronic hepatic injury and fibrogenesis, such as TGF-b, TNF-a and HMGB-1. Conclusion: The results obtained in this study reveals that even though immune related hepatitis model, ROS derived from activated CD68+/CD206+ KC could be a new target for the treatment of acute and chronic hepatic injury, and for this purpose, SH-Man-HSA has a great potential as a therapeutics because of its efficient and rapid delivery of SH to CD68+/CD206+ KC.

P135. SINGLE SOFTWARE SOLUTION FOR RAPID AND CONFIDENT METABOLITE PROFILING AND IDENTIFICATION Caroline Ding Thermo Fisher Scientific, San Jose, CA, USA In vitro drug metabolite identification by LC coupled with HRAM MS is an essential component of drug discovery to select and prioritize compounds with the best chance of success. Data processing and data collation are the current rate limiting step of MetID process. To improve the throughput of MetID workflow, this study demonstrates the use of new node based processing software Compound Discoverer from processing to reporting for rapid screening and confident identification of metabolites using bench-top Orbitrap mass spectrometer Q ExactiveÔ. Commercial compound Darunavir were incubated in rat liver microsomes (1 mg/mL) fortified with NADPH (1 mM) for 0 and 2 hour. Samples were chromatographically separated using Ultimate3000 UHPLC Thermo Fisher ScientificTM on an Accucore aQ C18 column (100  2.1 mm 3mm) with mobile phase consisting of (A) water/0.1% FA and (B) acetonitrile/0.1% FA. The HRAM analysis was conducted on Q Exactive equipped with a HESI II ion source. Full scan MS and HCD MS/MS data were collected in a data-dependent fashion at resolution 70 000 and 15 000, respectively. High-quality HRAM data acquired on Q Exactive provide a wealthy of information for metabolite identification, especially in the presence of complicated biologic matrix. The accurate mass full scan data with high fidelity isotopic fine structure allow confident metabolite identification and elemental composition assignment. The information rich HCD ms/ms fragments provide substructure confirmation and localization of site of transformation. The incubated samples of Darunavir were analyzed using processing workflow from Compound Discoverer for both metabolite identification and structure elucidation. The software uses combinatorial approach with dealkylation prediction to enable comprehensive search for expected metabolites. Adducts were grouped together which greatly reduces false positives. The unexpected component detection node Unknown Detector ensured no major unexpected metabolites were missed. Compare with control effectively removed false

102

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

positives. Elemental compositions of putative metabolites were automatically confirmed by fine isotopic pattern matching. Fragment ion matching (FIShTM) and annotations automatically provided substructure information and localization of site of transformation for the putative metabolites. The preliminary results in this study demonstrate HRAM data from Q Exactive and the unique features introduced by Compound Discoverer enable fast, efficient, and confident metabolite profiling in an all-in-one UHPLC/HR-MS/MS platform, which can significantly increase the throughput of metabolite identification in drug discovery.

P136. CYCLOLINOPEPTIDES AND ENTEROLACOTNE IN COMBINATION IMPROVE IN-VITRO INTESTINAL BARRIER FUNCTION FOLLOWING AN INFLAMMATORY STIMULUS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Ahmed Almousa1, Ed S. Krol1, Francois Meurens2, Martin Reany3, Jennifer Jones4, and Jane Alcorn1 1 College of Pharmacy and Nutrition, Drug Discovery and Development Research Group, University of Saskatchewan, Saskatoon, SK, Canada, 2Vido, University of Saskatchewan, Saskatoon, SK, Canada, 3College of Agriculture and Bioresources, University of Saskatchewan, Saskatoon, SK, Canada, and 4College of Medicine, University of Saskatchewan, Saskatoon, SK, Canada Inflammatory Bowel Disease (IBD) is a chronic idiopathic inflammation of intestinal epithelium and is sub-classified into Crohn’s Disease (CD) and Ulcerative Colitis (UC). IBD has a complex pathogenesis, which renders most pharmacologically active compounds incapable of fully resolving the inflammation. The maintenance of intestinal barrier has been widely accepted as a target to improve inflammation in IBD. Cyclolinopeptides (CLPs) and Enterolactone (ENL) are essential components derived from flaxseed, and experimental evidence suggests these bioactives mediate the anti-inflammatory and/or immunosuppressive properties associated with flaxseed supplementation. Here we hypothesize that CLPs and/or ENL are capable of reducing inflammation in an in-vitro model of IBD utilizing different protective pathways. HCT-8 intestinal epithelial cell line was grown on polystyrene transwells to allow development of a polarized epithelium. On day six Trans-Epithelial Electrical Resistance (TEER) values were measured, media was replaced with phosphate buffer saline (PBS) containing TNF-a (10 ng/mL) and INF-g (50 ng/mL), to induce inflammation, in addition to CLPs and ENL either alone or in combination. TEER reduction was measured at 24 hours and 48 hours for the same wells, followed by addition of Lucifer Yellow (LY) to assess maintenance of tight junctions. Subsequently, cells were collected to extract total mRNA and perform qRT-PCR analysis for target genes (IL-6, IL-8, IL-1b, TNF-a, PPAR-g, CCL-20 and MUC-2). TEER values indicate a protective effect of CLP-A at 24 hours, and a concentration-response relationship for CLP-J and ENL at both 24 and 48 hours. The combination of ENL and either of the CLPs resulted in significantly improved TEER values in comparison to each bioactive when solely administered. Genetic expression results are supportive of TNF-a down-regulation for CLPs combined to up-regulation of PPAR-g that positively affect the production of tight junction proteins. ENL was able to suppress the production of pro-inflammatory cytokines. Immunocytochemical analysis of tight junction proteins (occludins and claudins) is underway in order to support the maintenance of barrier integrity upon exposure to target bio-actives. In conclusion, bio-actives isolated from flaxseed are able to provide greater anti-inflammatory effects and improved maintenance of barrier integrityin-vitro when combined together, possibly due to the multimodal activity of these bio-actives.

P137. DEVELOPMENT OF A NANOTECHNOLOGY-BASED CARBON MONOXIDE DONOR AND ITS THERAPEUTIC IMPACT ON IDIOPATHIC PULMONARY FIBROSIS Saori Nagao1, Kazuaki Taguchi2, Ryota Tanaka1, Hiromi Sakai3, Hiroshi Watanabe1, Masaki Otagiri2, and Toru Maruyama1 1 Department of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan, 2 Faculty of Pharmaceutical Sciences, Sojo University, Kumamoto, Japan, and 3Department of Biopharmaceutics, Chemistry, school of Medicine, Nara Medical University, Kashihara, Japan Purpose: Carbon monoxide (CO) has potent anti-inflammatory and anti-oxidative effects, and has attracted interest as a possible clinically viable medicinal agent. However, it is difficult to regulate the body disposition of CO gas as therapeutics. To solve this problem, we focused on hemoglobin-vesicles (HbV), which is encapsulated human hemoglobin in phospholipid vesicles and has comparable gas binding ability to red blood cells, as CO delivery carrier. In this study, we evaluated the safety of CO-bound HbV (CO-HbV) in normal mice, and subsequently investigated the therapeutic impact of CO-HbV against idiopathic pulmonary fibrosis (IPF), which is thought to involve inflammation and the production of reactive oxygen species (ROS), and its mechanism using bleomycin-induced IPF model mice. Methods: CO-HbV was prepared by bubbling CO gas into deoxyHbV under sterile conditions. As one of the safety evaluations, we analyzed the influence of CO-HbV on the organ injury based on the results of laboratory test of plasma and histopathological examination in vivo, and on the human hematopoietic stem/progenitor cells by colony assay in vitro. IPF model mice were induced by intratracheal bleomycin administration (5 mg/kg) on day 0. Saline, HbV or CO-HbV (1000 mg Hb/kg) was intravenously administered on day 0 and 1. Pulmonary fibril formation and respiratory function were quantitatively evaluated by measuring hydroxyproline levels and forced vital capacity, respectively. Results and discussion: There were no observation of organ injury by tissue examination and serum laboratory parameters reflecting hepatic, renal and pancreatic function for the experimental period after CO-HbV administration, compared to normal mice. In colony assay, no effect was observed on the formation of colonies of burst-forming units of erythrocyte (BFU-E). These data demonstrated the lack of any irreversible sign for a deteriorative damage to the organs and suppressive effect on hematopoiesis.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

103

In IPF model mice, CO-HbV suppressed the progression of pulmonary fibril formation and improved respiratory function compared to saline and HbV. In addition, the histopathological evaluation using Masson’s trichrome stain also showed CO-HbV suppressed the pulmonary fibrosis. The mechanism underlying the suppressive effect of CO-HbV on pulmonary fibrosis can be attributed to a decrease in ROS generation and the production of inflammatory cells, cytokines and TGF-b in the lung. Furthermore, ROS generation was decreased as the result of the inhibition of the activity of the NADPH oxidase family (Nox4), which is an important role in the pathogenesis of IPF, with no detectable changes in its protein expression. Conclusions: CO-HbV is safe and effective agent for IPF and exerts both anti-inflammatory and anti-oxidative effects of CO without any severe side effects. CO-HbV could be new type of pharmaceutical therapeutic agent for using CO as a medical gas that would arrest ROS and inflammation-related disorders.

P138. A PREDICTION TOOL FOR THE DOSE-PROPORTIONALITY OF PHARMACOKINETICS TO GUIDE DECISION MAKING ON MICRODOSING STUDIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sieto Bosgra1, Joost Westerhout1, Maria LH Vlaming2, and Wouter H. Vaes1 1 TNO, Zeist, Netherlands and 2Department of Microbiology and Systems Biology, TNO, Zeist, Netherlands Microdosing studies allow investigation of clinical pharmacokinetics of drug candidates (first-in-man) earlier in development, at a stage where not all high dose safety concerns have yet been sorted out. Furthermore, microdosing allows inclusion of patients or target groups that cannot be admitted in high dose phase 0/1 trials. However, a potential cause for concern when considering a microdosing study is that a particular drug candidate may display nonlinear pharmacokinetics (PK). Saturation of processes such as membrane transport or metabolism at exposure levels between microdose and (intended) therapeutic dose may limit the predictivity of high dose PK from microdose observations. Guidance on the assessment of the likelihood of appreciable nonlinear PK based on preclinical information can be helpful in staging the clinical phase and the place of microdosing in it. We present a prediction tool in the form of a decision tree that evaluates concerns about nonlinearities raised in the preclinical phase and their potential impact on the proportionality between microdose and intended therapeutic dose as predicted from preclinical information. The tool considers several processes that may show saturation at different sites: gastrointestinal dissolution; intestinal, hepatic and renal uptake and efflux transport; cytochrome P450 and UGT metabolism; and protein binding. The expected maximum concentrations at the relevant sites are estimated by non-compartmental methods. These are compared with solubility, Michaelis constants for active or enzymatic processes and bound concentrations, or worst-case estimates thereof, to assess the potential saturation of the processes below therapeutic doses. We applied the decision tree to cases comparing microdose and therapeutic dose PK described in public literature (Ieiri et al., 2011; Lappin et al., 2006, 2010, 2011; Maeda et al., 2011). Except for nonlinearities caused by processes not considered in the tool, we were able to discriminate cases showing substantial nonlinearities from cases displaying dose-proportional PK based on preclinical input. The decision tree may provide valuable guidance to decide whether a microdosing study is a sensible option to gain early insight in clinical PK of drug candidates.

References Ieiri I, Nishimura C, Maeda K, et al. (2011). Pharmacokinetic and pharmacogenomic profiles of telmisartan after the oral microdose and therapeutic dose. Pharmacogenet Genomics 21:495–505. Lappin G, Kuhnz W, Jochemsen R, et al. (2006). Use of microdosing to predict pharmacokinetics at the therapeutic dose: Experience with 5 drugs. Clin Pharmacol Ther 80:203–215. Lappin G, Shishikura Y, Jochemsen R, et al. (2010). Pharmacokinetics of fexofenadine: Evaluation of a microdose and assessment of absolute oral bioavailability. Eur J Pharm Sci 40:125–131. Lappin G, Shishikura Y, Jochemsen R, et al. (2011). Comparative pharmacokinetics between a microdose and therapeutic dose for clarithromycin, sumatriptan, propafenone, paracetamol (acetaminophen), and phenobarbital in human volunteers. Eur J Pharm Sci 43:141–150. Maeda K, Takano J, Ikeda Y, et al. (2011). Nonlinear pharmacokinetics of oral quinidine and verapamil in healthy subjects: a clinical microdosing study. Clin Pharmacol Ther 90:263–270.

P139. WHERE MICROSAMPLING MEETS THE 3RS: IMPLEMENTATION AND BEST PRACTICES FOR SMALL MOLECULE TOXICOKINETIC ASSESSMENTS IN RATS Bianca M. Liederer1, Shelly Zhong2, Jessica Couch3, Xiao Ding1, Eric Harstad3, Xiaorong Liang1, Kirsten Messick4, Trung Nguyen4, Michelle Schweiger4, Jacqueline Tarrant5, and Brian Dean1 1 Department of Drug Metabolism and Pharmacokinetics, Genentech Inc., South San Francisco, CA, USA, 2Department of Nonclinical Operations, Genentech Inc., South San Francisco, CA, USA, 3Department of Toxicology, Genentech Inc., South San Francisco, CA, USA, 4Department of In Vivo Studies, Genentech Inc., South San Francisco, CA, USA and 5Department of Pathology, Genentech Inc., South San Francisco, CA, USA Blood microsampling is a technique that enables the measurement of circulating drug concentrations by use of reduced blood volume. The smaller sample volume decreases animal numbers in rodent toxicology studies needed for toxicokinetic (TK) evaluations by collecting small-volume, sparse samples from rats traditionally designated only for toxicity assessment, thereby

104

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

eliminating the need for TK satellite groups. The objectives of this study were to select and optimize blood collection methods, address technical/logistical challenges, and support a path forward for implementation of microsampling in both GLP and nonGLP toxicology studies with TK sampling. This initiative was driven by a commitment to the 3Rs (replacement, reduction, refinement) while maintaining high-quality TK evaluations. Both hemolyzed blood sampling (HBS) and microplasma sampling (MPS) were evaluated in pilot rat toxicity studies for accuracy of TK measurements and the impact on routine clinical pathology. A feasibility assessment on the impact of decreased volume collections (1.5 mL reduced to 150–200 mL) was conducted in rodents. Here, we propose new ‘‘base case’’ study designs for TK studies that eliminate or reduce TK satellite animals. In addition, we describe microsampling at a higher volume than commonly accepted in order to balance the needs of sample size (i.e. ISR, repeated analysis) with blood volume limits and clinical pathology impact. On the basis of data and gap analyses, the elimination or reduction of TK satellite groups in standard rodent GLP and non-GLP studies is currently being implemented.

P140. DEVELOPMENT OF HEMIACETAL ESTERFIED NEW QUINOLONE TO PREVENT CHELATION WITH METAL CONTAINING DRUGS AND TO PREVENT PSEUDOMEMBRANOUS COLITIS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kenji Matsuyama1, Toru Otori1, Atsusi Kawase1, Masahiro Iwaki1, Tetsutaro Kimachi2, and Yoshikazu Ishii3 1 Faculty of Pharmacy, Kinki University, Higashi-Osaka, Japan, 2School of Pharmacy & Pharmaceutical Sciences, Mukogawa Women’s University, Nishinomiya, Japan, and 3College of Medicine, Toho University, Funahashi, Japan Objective: The absorbtion of new quinolone antibiotics is reduced through a reaction with metal-containing drugs such as lanthanum carbonate and aluminum hydroxide. In the present study, we esterfied the center of chelation of the carboxylic group of the parent drug levoflaxin (LVFX) with the ethoxycarbonyl 1-ethylhemiacetal ester resulting in the synthesis of the pro-drug levofloxacin ethoxycarbonyl 1-ethyl hemiacetal ester (LVFX-EHE). We hypothesize that the LVFX-EHE is hydrolyzed into LVFX at the intestinal epithelium. Therefore, we performed a comparison of blood concentrations after oral administration of equimolar concentrations of LVFX and LVFX-EHE in rats. Also, a comparative examination of antibacterial activity for both LVFX and LVFX-EHE using the E. coli strains ATCC 25922 and NIH JC2, and Enterobacter,TCC23355. Experiment: Equimolar concentrations of LVFX (22.5 mg/kg) and LVFX-EHE (29.8 mg/kg) with or without aluminum hydroxide (30 mg/kg) were administered orally to rats. Blood sampling was carried out at 15, 30, 45, 90, 120, 180 and 240 min, followed by measurement of blood concentrations of LVFX by HPLC. Antimicrobial susceptibility testing was performed using the broth microdilution method described by the ‘‘Clinical and Laboratory Standards Institute’’ M07-A09 (2012) document. Results: The same blood concentration of LVFX was obtained after the oral administration of equimolar concentrations of LVFX or LVFXEHE without aluminum hydroxide. Furthermore, co-administration of LVFX and aluminum hydroxide reduced the blood concentration of LVFX by about 50%, while co-administered with LVFX-EHE and aluminum hydroxide observed same blood concentration of LVFX and LVFX-EHE without aluminum hydroxide. Regarding the antibacterial action of LVFX-EHE in the digestive tract before intestinal absorption, a significant decreased sensitivity of LVFX-EHE was observed when compared with that of the parent drug LVFX, suggesting a soft quinolone that is less susceptible to pseudomembranous colitis. This study suggests the benefit ofesterification of the carboxyl group of new quinolone.

P141. ASSESSMENT OF SYNERGISTIC EFFICACY AND SAFETY OF IBD98-M CONTROLLED-RELEASE CAPSULES (MESALAMINE AND SODIUM HYALURONATE) FOR TREATMENT OF MILD TO MODERATE ULCERATIVE COLITIS: COMPARISON TO MESALAMINE Eskouhie Tchaparian Preclinical Research, HolyStone Healthcare, Taipei, Taiwan Ulcerative colitis (UC) is one of the main types of inflammatory bowel disease (IBD). It is a long-term (chronic) condition caused by ulcers and inflammation of the inner lining of the large intestine (colon).The most common symptoms of UC include intermittent rectal bleeding, crampy abdominal pain and diarrhea. Patients also may experience fatigue, weight loss and loss of appetite. Treatments of UC may involve both medications such corticosteroids, aminosalicylates (mesalamine/5ASA) and thiopurines or surgery in more severe cases. In addition, applications of hyaluronic acid, a natural biopolymer has been shown to be effective in improving and treating symptoms of IBD including UC. These studies were undertaken to evaluate the pharmacological relevance of IBD98-M, a delayed release capsule, as a combination therapy of both mesalamine and sodium hyaluronate for the treatment of mild to moderate ulcerative colitis (UC). The synergistic effect of mesalamine and hyaluronate suspensions was assessed in TNBS-induced colitis rat models in vivo. Further, to confirm targeted and controlled drug delivery of mesalamine (active ingredient), the systemic exposure of oral administration of 1000 mg of IBD98-M was compared to that of Pentasa (mesalamine/5-ASA) in pigs. Macroscopic scoring results of colitis implied an overall improvement in the colon tissue pathology in IBD98-M treated rats relative to HA, mesalamine or PBS treatment groups. Reduction in colon tissue thickness and significant decrease in MPO activity was also observed in response to IBD98-M. Body weight gain and stool consistency were also observed characteristics of IBD98-M administration. In contrast to mesalamine group, the PK studies showed significant decrease in serum mesalamine levels, in addition to prolonged drug exposure profile in IBD98_M treatment. These results provide supporting evidence for the potential clinical application of IBD98-M for treatment of mild to moderate ulcerative colitis.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

105

P142. ALBUMIN AND ALBUMIN-DRUG CONJUGATES: COMPARATIVE PHARMACOKINETICS AND TISSUE DISTRIBUTION STUDIES USING LYSINE AND CYSTEINE SPECIFIC LABELLING AGENTS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Andrew McEwen, Stuart G. Wood, Claire Henson, and Stephen Harris Quotient Bioresearch Limited, Rushden, UK The therapeutic use of proteins, peptides, and small molecule drug candidates is hampered by short in vivo serum half-life. One approach to alter their biodistribution and pharmacokinetics is to take advantage of the exceptionally long circulation half-life of serum albumin or IgG, which is attributed to interaction with the neonatal Fc receptor (FcRn). Low-molecular weight drugs and bioactive proteins have been conjugated to albumin, or encapsulated into albumin nanoparticles. The observed accumulation of albumin in solid tumors has resulted in the development of albumin-based drug delivery systems for tumor targeting. Modified albumins have also been developed in an attempt to further extend the half-life of the carrier molecule. Whilst the pharmacokinetics of albumin is generally well described within the native species (ie human albumin in man), there is a notable lack of data regarding the pharmacokinetics of albumins across species (i.e. human albumin in the rat). This study was performed to provide pharmacokinetic information for albumin in pre-clinical species. Albumin (human, mouse and bovine) was labeled using both [3H] and [14C} N-Succinimidyl Propionate, a lysine specific reagent and [14C] N-Ethyl Maleimide, a cysteine specific reagent. Proteins were purified using a size exclusion column and checked by HPLC using BioSep SEC-2000 column (300  7.8 mm) with mobile phases 100 mM Sodium phosphate pH6.8 and 300 mM Sodium Chloride. Sprague-Dawley rats were dosed intravenously with radiolabelled human albumins (6.25 mg/kg) and were killed 0.25, 1, 3, 8, 24, 72, 120 and 240 hours post dose. Terminal blood samples were taken for preparation of plasma and carcasses were retained for whole body autoradiography. For each radiolabel one animal was placed in a glass metabowl and urine and faeces were collected up to 240 hours post dose. Male CD-1 mice were dosed intravenously with radiolabelled mouse or bovine serum albumin (6.25 mg/kg), as above. Serial samples were taken 0.25, 1, 3, 8, 24, 72, 120 and 240 hours post dose for preparation of plasma. The plasma halflife for radiolabeled human albumin in the rat was found to be about 25 hours for all three labeled forms. Radioactivity was excreted slowly in the urine (90%) during 240 hours consistent with the plasma half-life with about 10% excreted in the faeces. Autoradiograms showed radioactivity distributed throughout the animal with maximal concentrations in tissues generally observed 3 hours after dose administration. The plasma half-life for radiolabeled mouse albumin in the mouse was found to be about 38 hours, whilst the plasma half-life for bovine serum albumin in the mouse was 72 hours. The data obtained in this study provides basic pharmacokinetic information that can be used in the development of albumin drug conjugates. It has been shown that the half-life of albumin can vary dependent on the source and species used for testing. Similar pharmacokinetics and tissue distribution of radioactivity was observed for all three radiolabelled forms indicating that the radiolabel was placed in a stable position in the protein.

P143. PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELLING IN DRUG DISCOVERY AND DEVELOPMENT: A PHARMACEUTICAL INDUSTRY PERSPECTIVE Yuan Chen1, Hannah Jones2, Christopher R. Gibson3, Tycho Heimbach4, Neil Parrott5, Sheila Peters6, Jan Snoeys7, Vijay Upreti8, Ming Zheng9, and Stephen David Hall10 1 Department of Drug Metabolism and Pharmacokinetics, Genentech, Inc., South San Francisco, CA, USA, 2Pfizer Worldwide R&D, Pfizer, Cambridge, MA, USA, 3Department of Drug Metabolism, Merck Research Laboratories, West Point, PA, USA, 4 Novartis Institutes for Biomedical Research, East Hanover, NJ, USA, 5Department of Non-Clinical Safety, Hoffmann-La Roche AG, pRED, Pharma Research & Early Development, Basel, Switzerland, 6Innovative Medicines and Early Clinical Development, AstraZeneca, Mo¨lndal, Sweden, 7Janssen Research & Development, Beerse, Belgium, 8Department of Quantitative Pharmacology, PKDM, Amgen, Seattle, WA, USA, 9Exploratory Clinical Development, Bristol-Myers Squibb Company, Wallingford, CT, USA, and 10Department of Drug Disposition, Eli Lilly & Company, Indianapolis, IN, USA Physiologically based pharmacokinetic (PBPK) models are employed broadly throughout the pharmaceutical industry to provide quantitative, integrated analyses of drug absorption and disposition in pre-clinical species and humans. A group comprising of industrial scientists with expertise in PBPK modelling was assembled under the umbrella of the Innovation and Quality (IQ) Drug Metabolism and Clinical Pharmacology Leadership Groups. The focus of this group has been to develop a united view on PBPK best practice and application in the pharmaceutical industry. The first position paper on PBPK modelling that represents the collective view of PBPK experts from 10 leading global pharmaceutical companies will be published. In this unique industry wide collaboration, we describe the principles that underlie the utility and value of PBPK modelling in drug discovery and development. Through detailed case studies and an extensive compilation of  80 examples we exemplify the diverse opportunities for efficiency gains provided by PBPK modelling. The case studies demonstrating PBPK impacts from early discovery through late development will be shared: (a) Pediatric PBPK modelling for formulation selection and informing PIP, (b) organ impairment modelling (c) PBPK modelling for DDI predictions using reversible and time dependent inhibition, (d) ACAT modelling for absorption of preclinical species, (e) Recent PBPK model-based FDA approvals with ‘‘lessons learned’’. The perspective on the confidence and challenges with these different applications are also described and presented in a table. While the confidence level is relatively high for PK and CYP-based DDI prediction, major challenges remain in the prediction of transporter and non-CYP enzyme based DDI, ontogeny of enzymes and transporters proteins and changes in enzyme/transporter

106

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

abundance in disease populations of interest. We highlight the new opportunities arising as global regulatory agencies recognize that well-planned and well-designed PBPK modeling may not only optimize preclinical or clinical studies throughout the course of the development of the molecule but also support new product labels. Industry best practices for implementation and validation of PBPK models are also presented.

P144. A FIT-FOR-PURPOSE APPROACH TO NONCLINICAL ASSESSMENTS FOR ADALIMUMAB AND INFLIXIMAB BIOSIMILAR DEVELOPMENT

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sharon L. Ripp1, Michael Bolt2, Mazin Derzi2, Gregory L. Finch2, Susan Hurst1, Theodore J. Johnson3, Michael W. Leach2, and Teresa A. Smolarek1 1 Department of Pharmacokinetics, Dynamics & Metabolism, Pfizer, Groton, CT, USA, 2Department of Drug Safety Research & Development, Pfizer, Andover, MA, USA, and 3Department of Pharmacokinetics, Dynamics & Metabolism, Pfizer, La Jolla, CA, USA Biosimilars are biologic products that are developed to be highly similar to an existing licensed biologic product such that there are no clinically meaningful differences between the two. Biosimilars have the potential to increase patient access to key medicines. Given that there are multiple and variable parameters considered when developing a biosimilar, the nonclinical development strategy is typically fit-for-purpose and based on regulatory guidance and data available for the innovator. The nonclinical assessments for two potential biosimilars, adalimumab and infliximab, are described herein. Both agents are monoclonal antibodies (mAbs) targeting tumor necrosis factor (TNF) and are used in the treatment of numerous inflammatory and autoimmune diseases. Infliximab, also known as RemicadeÕ , is a chimeric mAb that does not bind to TNF from standard nonclinical species; therefore, the nonclinical assessment was limited to a single dose toxicokinetic (TK)/tolerability study in rats comparing the tolerability, exposure, and anti-drug antibody (ADA) response of the proposed biosimilar (PF-06438179) to approved infliximab sourced from Europe (infliximab-EU). Following administration of a single intravenous dose to rats, PF06438179 and infliximab-EU were well-tolerated. Systemic exposures [Cmax and area under the concentration time curve (AUC)] in animals dosed with PF-06438179 or infliximab-EU appeared similar, with mean exposure ratios of PF-06438179 relative to infliximab-EU ranging from 0.88 to 1.16. None of the animals dosed with PF-06438179 or infliximab-EU developed ADA. Adalimumab, also known as HumiraÕ , is a fully human mAb and binds TNF from cynomolgus monkeys, making this species pharmacologically-relevant. Therefore, a repeat-dose toxicity study was conducted in cynomolgus monkeys comparing the toxicity, TK and ADA response of the proposed biosimilar (PF-06410293) with approved adalimumab sourced from Europe (adalimumab-EU). PF-06410293 and adalimumab-EU were administered subcutaneously once per week for four weeks at a dose of 157 mg/kg; both agents were well tolerated. Systemic exposures (Cmax and AUC) appeared similar, with mean exposure ratios of PF-06410293 relative to adalimumab-EU ranging from 1.0 to 1.2. PF-06410293- and adalimumab-EU-related effects were limited to microscopic observations of decreased cellularity of lymphoid follicles and germinal centers in the spleen, with similar incidence and severity across groups. These results were expected based on the mechanism of action and were consistent with data reported by the innovator for adalimumab. The data from these nonclinical studies support further clinical investigation of both of the proposed biosimilars.

P145. IN VITRO AND IN VIVO EVALUATION OF THE ANTI-TUMOR ACTIVITY OF CA102N, A NOVEL HYALURONIC ACID CONJUGATED DRUG (HACD), AGAINST COLORECTAL CANCER Eskouhie Tchaparian Preclinical Research, HolyStone Healthcare, Taipei, Taiwan Hyaluronic Acid conjugated drugs (HACD), are novel carrier-mediated drug delivery systems (Targeting CarrierÔ) for selective and targeted transport of therapeutics. The technology is based on CD44-mediated endocytosis of Hyaluronic Acid (HA) and intracellular drug release. In the present study, the antitumor activity of CA102N, an HA-Nimesulide (COX2 inhibitor) conjugate, was investigated in vitro and in vivo. CD44 expression and CD44-mediated HA binding was evaluated in vitro in HCT15 (CD44-low) and HT29 (CD44-overexpressed) colorectal cancer cell lines by Flow Cytometry (FACS analysis) and confocal microscopy. A high CD44-expression level (62.5%) was detected in HT29 compared to that of HCT15 cells (18.6%). The fluorescence microscopy analysis confirmed HA endocytosis into the cytosol of HT29 cells, but not sufficiently in HCT15 due to low CD44 expression levels. Further, cytotoxicity assessment assay (MTT) of HT29 colorectal cancer cells demonstrated CD44-mediated anticancer activity of the conjugate relative to free drugs (XNO2, XNH2) or HA. To determine the in vivo therapeutic effects of CA102N, a dose-dependent intravenously (i.v.) administration of CA102N or free drug (XH2) (1.5 or 4.5 mg/kg) was investigated in female athymic mice bearing HT29 colorectal cancer cells. Substantial tumour growth inhibition was recorded in CA102N treatments relative to the control or free drug (XNH2). Additionally, tumour shrinkage was seen within 3 days of the high-dose CA102N administration that almost completely disappeared by day 7 in 1/3 of the treated mice; therefore, indicating potential CA102N mediated induction of apoptosis in human colon tumour (HT29 cells). Interestingly, there were no apparent signs of treatment-related toxicity to the mice. The results of these studies suggest that CA102N exhibits antitumor efficacy on human colorectal cancer cells (HT29) xenografts and may represent a highly promising agent with significant therapeutic benefits and toxicity was not observed in mice. Hyaluronic

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

107

acid-conjugated drug (HACD) delivery system is an innovative approach that may have great potential for development of new and effective therapeutics for cancer therapy.

P146. TRANSPORT CHARACTERISTICS OF EXOSOMES IN HUMAN BRAIN CAPILLARY ENDOTHELIAL CELLS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Hiroki Kuroda1, Masanori Tachikawa2, Yasuo Uchida2, and Tetsuya Terasaki2 1 Faculty of Pharmaceutical Sciences, Tohoku University, Sendai, Japan and 2Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan Introduction: Drug delivery to the brain is still a big challenge in the therapy of central nervous system (CNS) disorder because the blood-brain barrier (BBB) hinders the distribution of drugs into brain. There has been a growing interest in the applications of exosomes, cell-secreting microvesicles, to the CNS drug delivery as an endogenous carrier of low molecular weight drugs and nucleic acid medicine. Therefore, clarifying the mechanism(s) of exosomes transport at the BBB will provide a new insight into an alternative drug delivery route at the BBB. The purpose of this study was to investigate the characteristics of exosomes internalization into human brain capillary endothelial cells as the first step of the BBB transcellular transport. Methods: Exosomes fraction was extracted from the exosomes-free culture medium of colon26 cells, a mouse colon cancer cell line by using an Exoquick-TC kit. Immuno blot analyses of TSG101, an exosomes marker protein, and GRP78, a Golgi marker, were performed to confirm the enrichment of exosomes used in the transport assay. The isolated exosomes were labeled with a PKH67 fluorescent cell linker kit. A human brain capillary endothelial cell line (hCMEC/D3 cells) was used as an in vitro human BBB model. The internalization activity of exosomes into hCMEC/D3 cells was estimated by the detection and/or quantification of fluorescence with the confocal laser scanning microscopy and microplate reader. Results and discussion: Immuno blot analyses revealed that the band of TSG101 was detected in the exosomes-enriched fraction whereas the band of GRP78 was not detected. This result confirmed the enrichment of exosomes derived from colon26 cells. The amount of PKH67-labeled exosomes taken up by hCMEC/D3 cells was increased in a dose-dependent manner. The uptake of labeled exosomes by hCMEC/D3 cells was diminished under 4 degree Celsius condition and significantly decreased in the presence of the excess unlabeled exosomes. Methyl-b-cyclodextrin and chlorpromazine, an inhibitor of caveolae-lipid raft endocytosis and clathrin-mediated endocytosis, respectively, had no effect on the uptake. Furthermore, the green fluorescence derived from labeled-exosomes was detected as dots shape in the cells. These lines of evidence suggest that the exosomes are internalized into hCMEC/D3 cells in an energydependent manner presumably via caveolae- and clathrin-independent pathway. Conclusion: Human brain capillary endothelial cells have the ability to take up the exosomes via a vesicular internalization pathway.

P147. A NEW HEPATOCYTE BASED ASSAY THAT DELIVERS AN ESTIMATION OF FRACTION METABOLIZED VIA A SINGLE ENZYMATIC PATHWAY SUPPORTED WITH STRUCTURAL INFORMATION OF METABOLITES Carina Leandersson1, Kajsa Kanebratt1, Tommy B. Andersson1, Emre M. Isin1, and Richard A. Thompson2 1 Cardiovascular & Metabolic Diseases iMed DMPK, AstraZeneca R&D Molndal, Mo¨lndal, Sweden and 2Respiratory, Inflammation & Autoimmunity iMed DMPK, AstraZeneca R&D Molndal, Mo¨lndal, Sweden Metabolism of a candidate drug via a single enzymatic pathway may implicate a victim drug–drug interaction (DDI) risk which could delay further development of the compound. Therefore, experimental methods addressing enzyme phenotyping and DDI liabilities are important not only in late stage drug discovery, but also in early stages where there still is a possibility to influence the design of new molecules. To assess the victim DDI risk in early drug discovery projects, two assays are used as part of our workflow today. One is a cytochrome P450 phenotyping (CYPID) assay using intrinsic clearance (CLint) from ten different recombinant cytochrome P450 (CYP450) enzymes together with intersystem extrapolation factors (ISEFs) to estimate fraction metabolism via a single CYP450 (fm,CYP). The other assay estimates fm,CYP by comparing the CLint of a compound incubated in pooled human hepatocytes in the absence and presence of a specific enzyme inhibitor (FrCLint). The differences between the two CLint-values is then used for the estimation of fm,CYP (Desbans et al., 2014). Both these assays have drawbacks with respect to design of new compounds. The CYPID assay only takes into account ten CYP450s and thus excludes information on other hepatic enzymes. The FrCLint assay on the other hand includes all hepatic enzymes, but is often unsuccessful in delivering data on low CLint compounds. Furthermore, neither of these assays deliver metabolite structural information which is crucial in building SARs and understanding how chemical modifications affect interactions with drug metabolizing enzymes. Both assays are labor intensive and thereby costly, making them unsuitable for the design-make-test phase. To overcome these drawbacks we have developed a new assay. The assay is a refinement of the FrCLintassay using human hepatocytes and ketoconazole to investigate the CYP3A4 contribution. However, instead of following depletion of parent at different time points to calculate a CLint value, the extent of metabolite formation from a compound incubated in the absence and presence of ketoconazole at one time point is used to estimate fm,CYP3A4. The evaluation of the fm,CYP3A4 for this assay shows good correlation with FrCLint. It has been used successfully in delivering results for low CLintcompounds as well as structural information of the metabolites.

References Desbans C, Hilgendorf C, Lutz M, et al. (2014). Prediction of fraction metabolized via CYP3A in humans utilizing cryopreserved human hepatocytes from a set of 12 single donors. Xenobiotica 44:17–27.

108

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P148. EVIDENCE OF IN VITRO AND IN VIVO INTERACTION BETWEEN a-HYDROXY ALDEHYDE METABOLITE OF ATAZANAVIR AND ISONIAZID

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Shalu Jhajra and Saranjit Singh Department of Pharmaceutical Analysis, National Institute of Pharmaceutical Education and Research, Mohali, India The treatment of HIV requires use of multiple antiretroviral drugs. The standard therapy to cure tuberculosis (TB) in patients with HIV infection is a 6-month regimen comprising again of multiple drugs (Blumberg et al., 2003). Thus, the treatment in patients having both HIV and TB is doubly complicated. There are a few reports in the literature that have focused upon drug–drug interaction among HIV and anti-TB drugs. Mainly, these studies have been carried out on HIV protease inhibitors and rifampicin, with and without ritonavir (Acosta et al., 2007, Burger et al. 2006, Haas et al. 2009, Mallolas et al. 2007). However, the literature is silent on interaction between atazanavir and other HIV protease inhibitors with isoniazid (and remaining anti-TB drugs). The purpose of this study was to explore interaction between isoniazid and atazanavir at the metabolism level. The interaction product was targeted both in vitro (rCYP3A4, human and rat liver S9 fraction and microsomes) and in vivo (plasma, bile, urine and feces) matrices of rat. The oral doses of atazanavir and isoniazid were 6 and 5 mg/kg, respectively. The presence of interaction product was confirmed by HRMS data, and through its fragmentation pattern, which was established using Tandem mass (MSn). The latter was also used to confirm the presence of interaction product through MRM mode. During the data analysis, interaction was found between a-hydroxy aldehyde metabolite of atazanavir and isoniazid, which is proposed to happen via Schiff base mechanism. The presence of interaction product was found in all in vitro samples. In case of in vivosamples, the presence of interaction product was found in bile and feces of rat, while the same was not present in plasma. The urine samples are being analyzed. This interaction may result in decrease in the toxicity of atazanavir, by neutralizing its reactive aldehyde metabolite. The interaction product did not appear in plasma and was excreted out by biliary excretion system, which also reflects that this interaction product was not exposed to other tissues, so there may be decreases in extra-hepatic toxicity along with hepatic toxicity. Perhaps this is the first report on in vivointeraction between reactive metabolite of one drug with a co-administered drug.

References Acosta EP, Kendall MA, Gerber JG, et al. (2007). Effect of concomitantly administered rifampin on the pharmacokinetics and safety of atazanavir administered twice daily. Agents Chemother 51:3104–3110. Blumberg H, Burman W, Chaisson RE, et al. (2003). American Thoracic Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America: treatment of tuberculosis. Am J Respir Crit Care Med 167:603. Burger D, Agarwala S, Child M, et al. (2006). Effect of rifampin on steady-state pharmacokinetics of atazanavir with ritonavir in healthy volunteers. Antimicrob Agents Chemother 50:3336–3342. Haas DW, Koletar SL, Laughlin L, et al. (2009). Hepatotoxicity and gastrointestinal intolerance when healthy volunteers taking rifampin add twice-daily atazanavir and ritonavir. J Acquired Immune Deficiency Syndromes 50:290–293. Mallolas J, Sarasa M, Nomdedeu M, et al. (2007). Pharmacokinetic interaction between rifampicin and ritonavir boosted atazanavir in HIV infected patients. Medicine 8:131–134.

P149. VALIDATED UGT INHIBITION ASSAYS TO ASSESS POTENTIAL DRUG–DRUG INTERACTION LIABILITY OF NEW CHEMICAL ENTITIES Kwang-Jong Chen, Jon Gilbert, and Mohammed Taimi Cyprotex, Watertown, MA, USA UDP-glucuronosyltransferases (UGTs) are a family of conjugating enzymes that participate in the metabolism of many drugs. These enzymes catalyze the conjugation of glucuronic acid with endogenous (e.g. steroids and bilirubin) and xenobiotic (e.g. drugs and carcinogens) chemical compounds containing hydroxyl, amine, sulfhydryl, and carboxylic acid functional groups. While the cytochrome P450 enzymes are the greatest contributor to drug metabolism for most medications, the UGTs are known to be the second greatest contributor to drug metabolism. Glucuronidation pathways are a listed clearance mechanism for 1 in 10 of the top 200 prescribed drugs. There are currently eleven different UGT enzymes that are known to contribute to the phase II metabolism of many of the most prescribed medications used in clinical practice, with UGT2B7 having the greatest number of substrates followed by UGT1A1, UGT1A9 and UGT1A4. Inhibition of UGT enzymes might lead to clinically relevant drug– drug interactions and have been associated with significant increases in bilirubin concentrations that have been connected with jaundice. In this study, in vitro UGT inhibition assays using human liver microsomes and human recombinant UGT enzymes were developed as screening approaches to assess the impact of new chemical entities on altering the exposure of drugs metabolized by UGT enzymes. These assays are based on analysis and quantification by high-performance liquid chromatography–tandem mass spectrometry of glucuronides formed from selective probe substrates, namely, naloxone (UGT2B7), b-estradiol (UGT1A1), 1-naphthol (UGT1A6), propofol (UGT1A9), trifluoperazine (UGT1A4) and sulindac sulfone (UGT1A3). The enzyme kinetics have been characterized, through the evaluation of protein and incubation time, linearity, and enzyme kinetics (Km and Vmax). Using these optimized conditions, we then established the profile of the enzyme inhibition using chemical inhibitors. In conclusion, understanding of a compound’s potential for UGT interference may be a useful tool for limiting or avoiding drug–drug interaction liabilities in humans.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

109

P150. MECHANISM-BASED EVALUATION OF INTERACTION POTENTIALS BETWEEN STEVIOL AND SELECTED DRUGS Meiyu Wang, Jiajun Li, Huixin Qi, Jia Lu, and Yunting Xu Center for Metabolism and Pharmacokinetics, College of Pharmaceutical Sciences, Sochoow Universiy, Suzhou, China

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Steviol is the aglycone of steviosides and rebaudioside A found in stevia extracts. Pharmacokinetic investigations in rats and humans indicate that steviosides and rebaudioside A are extensively hydrolyzed in the gastrointestinal tract after oral ingestion, leading to the formation of steviol. Steviol is then absorbed into the circulation and cleared from the body via glucuronidation and subsequent excretion of steviol glucuronide (SV-G) into bile and urine. In the present study, UGT enzymes involved in the glucuronidation of steviol were first characterized and their potential to interact with selected drugs was examined. Using stably transfected cell lines overexpressing human transporters, carrier-mediated transport of SV-G was then investigated. Results suggest that steviol glucuronidation was primarily mediated by UGT2B7 at low concentration (2 mM) and UGT2B7 and UGT1A3 at high concentration (20 mM). Inhibition studies indicate that diclofenac displayed a relatively strong inhibition against steviol glucuronidation in human liver microsomes with a Ki of 4.2 mM, while haloperidol, lamotrigine and zidovudine were weak inhibitors of the reaction. Efflux transport study showed that SV-G was not a substrate of breast cancer resistance protein and P-glycoprotein. Among uptake transporters tested,OAT3 played a predominant role in the uptake of SV-G at low concentration (5 mM), while OAT1-, OATP1B3-, OATP1B1- and OATP2B1-mediated uptakes of SV-G were evident at high concentration (20 mM).The above results suggest that drugs or natural occurring compounds may interfere with steviol clearance via UGT enzyme inhibition or transporter activity modulation.

P151. THE ‘MORITELA-TSHWENE TEA EXTRACT’ INHIBITED CYP450 ENZYMES IN VITRO BUT NOT IN VIVO A. Walubo1, Z. Bekker1, M. R. Lekhooa1, J. B. du Plessis1, and M. G. Matsabisa2 1 Department of Pharmacology, University of the Free state, Bloemfontein, South Africa and 2Indigenous-Knowledge-System, Medical Research council, Cape town, South Africa The increasing use of traditional medicines by patients together with conventional medicines has led to a rise in the incidence, and consequently, recognition of the dangers of herbal–drug interactions. It is, therefore, appropriate for anyone contemplating developing traditional remedies for human use to explore their potential interactions with the conventional medicines. The aim of this study was to screen Moritela-Tshwene Tea (MTT) extracts, a product in development by the Medical Research CouncilIndigenous knowledge Systems program (MRC-IKS) for wide community medicinal use, for possible interaction with the major drug metabolising human cytochrome-P450 (CYP450) isoforms; CYP1A2, CYP2B6, CYP2C9, CYP2C19, CYP2D6 and CYP3A4. First, the effect of the MTT Tea extracts on the activity of each of the six CYP450-isoforms (Baculosomes) in vitro was tested by observing for changes in the rate of metabolism after addition to the reaction mixture of increasing amount of the MTT extract solutions at 0, 31.25, 62.5, 125, 250, 500 and 1000 ug/ml; n = 6. Thereafter, the effect of MTT extracts on the activity and quantity of CYP1A2, CYP2B1/2, CYP2D1 and CYP3A2 in vivo was determined using a rat model. The animals were administered with the MTT extracts (15 mg/kg) for 3 and 7 days, after which liver-microsomes were analysed for enzyme activity, and then for enzyme quantity by SDS-PAGE. The study was approved by the Animal Ethics committee of the University of the Free State. Whereas the MTT extracts inhibited the activities of all the six CYP450 enzymes in vitro, the MTT extracts had no effect on the activity and quantity of rat CYP3A2, CYP2D6, CYP1A2 and CYP2B6 in vivo. This implies that the MTT extracts would not interfere with drugs that are significantly metabolized by CYP450 enzymes. In conclusion, although the in vitro study demonstrated that the MTT contains potent CYP450 enzyme inhibitors, the in vivo experiment showed that these enzyme inhibitors had no effect on the enzymes.

P152. INHIBITION AND INDUCTION POTENCIAL OF NEW ANTI-TUBERCULOSIS DRUG, DELAMANID, ON CYTOCHROME P450 ENZYMES Katsunori Sasahara, Yoshihiko Shimokawa, Noriaki Yoda, Katsuhiko Mizuno, and Ken Umehara Tokushima Research Institute, Otsuka Pharmaceutical Co., Ltd., Tokushima, Japan Introduction: Delamanid (OPC-67683) inhibits mycolic acid synthesis in Mycobacterium tuberculosis and is a new medicine for the treatment of multidrug-resistant tuberculosis (MDR-TB) in combination with a background drug regimen. In addition, TB patients are often infected with other diseases such as human immunodeficiency virus and are treated with concomitant drugs for these other indications. Purpose: The current study was investigated the effects of delamanid on cytochrome P450 (CYP) enzymes in in vitro to reveal the drug–drug interaction (DDI) potential. Methods: The inhibitory effects of delamanid at up to 100 feˆm on eight CYP isoforms (CYP1A2, CYP2A6, CYP2B6, CYP2C8/9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4) were investigated using probe substrates and human liver microsomes by high performance liquid chromatography coupled with ultraviolet and fluorescence and liquid chromatographic-tandem mass spectrometry (LC-MS). The inductive effects of delamanid at up to 10 feˆm on CYP1A2, CYP2B6, CYP2C9, and CYP3A4 were investigated using human hepatocytes; each CYP mRNA level in the hepatocytes was analyzed by quantitative real-time PCR. In addition, the enzyme activities except CYP2B6 were measured by the LC-MS method. Results: Delamanid showed no inhibitory effects on any CYP isoform, and there is little possibility for mechanism-based inactivation. In addition, there were small or no increases in the mRNA levels and marker enzyme activities in human hepatocytes. Conclusion: Taken together, these data suggest that delamanid does not affect the metabolizing capacity of other drugs which are

110

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

metabolized via CYPs. The highest concentration of delamanid in this study was higher than the maximal plasma concentration observed in the clinical trial. Therefore, delamanid is unlikely to cause clinically relevant DDI via changes in exposure of other drugs.

P153 . A METABOLISM-DRIVEN COMBINATORIAL APPROACH TO OVERCOME ER/HER2 CROSSTALK IN BREAST CANCER

Downloaded by [112.134.238.102] at 07:30 10 February 2016

James Chun Yip Chan, Peirong Lim, Pei Shi Ong, and Eric Chun Yong Chan Department of Pharmacy, National University of Singapore, Singapore, Singapore For more than three decades, tamoxifen, an estrogen receptor (ER) antagonist is the cornerstone treatment for ER + breast cancer. However, clinical use of tamoxifen is complicated by significant inter-individual variability attributed to polymorphic expression of CYP2D6 and CYP3A which sequentially metabolize the pro-drugtamoxifen to its primary active metabolite, endoxifen (Figure 1). Additionally, acquired resistance to tamoxifen occurs in 30% of patients, and is attributed to crosstalk between pro-proliferative ER and human epidermal growth receptor 2 (HER2) pathways, resulting in incomplete inhibition of proliferative signals. Preclinical studies demonstrated synergism between lapatinib, a HER2 inhibitor, and 4-hydroxytamoxifen, another active metabolite of tamoxifen in resistant breast cancer cells, prompting initiation of 3 clinical trials to investigate in vivo evidence of synergism between lapatinib and tamoxifen. However, a major but neglected confounding factor is the potent, irreversible mechanism-based inactivation of CYP3A by lapatinib, essentially inhibiting bioactivation of tamoxifen to endoxifen, resulting in pharmacokinetic antagonism which may abrogate the predicted pharmacodynamic synergism. In this work, we aimed to investigate the effect of lapatinib on tamoxifen bioactivation and strategically exploit the synergistic effects of ER and HER2 inhibition. Human liver microsomes were incubated with or without 10 mM lapatinib for 30 minutes, followed by a 10  dilution step into secondary incubations containing 0–100 mM tamoxifen or 4-hydroxytamoxifen, and formation kinetics of N-desmethyltamoxifenand endoxifen respectively mediated by CYP3A were monitored. Subsequently, the cytotoxic effects of lapatinib in combination with tamoxifen or its active metabolites 4-hydroxytamoxifen and endoxifen were determined in MCF-7 (ER + /HER2-), BT474 (ER + / HER2 + ) and MCF-7/HER2 (ER + /HER2 + ) breast cancer cells using the combination index (CI) design by Chou and Talalay. Intrinsic clearance values in the absence and presence of lapatinib were 6.63 vs. 2.99 pmol/min/mg for tamoxifen, and 2.39 vs 0.58 pmol/mg/min for 4-hydroxytamoxifen, indicating that lapatinib reduced CYP3A-mediated bioactivation by more than 2- and 4-fold respectively. Cytotoxicity studies revealed for the first time potent synergism between lapatinib and endoxifen in all cell lines, where significantly lower lapatinib (0.9–134 nM) andendoxifen (78–160 nM) concentrations were required to achieve synergism than with tamoxifen (41000 nM) or 4-hydroxytamoxifen (4600 nM). Notably, antiproliferative activity was achieved using 30- to 4000-fold lower concentrations than plasma concentrations achievable with current therapeutic regimens, even in BT474 which is reported to be poorly responsive to tamoxifen. By directly administering endoxifen, inter-individual variability in tamoxifen response due to CYP450 polymorphism, and bioactivation inhibition due to CYP3A inactivation by lapatinib is circumvented, while achieving potent synergism due to simultaneous inhibition of crosstalk between ER and HER2 at lower doses, potentially improving safety profiles due to reduced drug exposure. This work showcases how development of a novel and effective rational drug combination is driven by judicious application of metabolism investigations (Figure 1).

Figure 1. Bioactivation of tamoxifen to its primary active metabolite, endoxifen. Thickness of arrows indicates predominant metabolic pathways.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

111

P154. ABSTRACT WITHDRAWN

P155. PREDICTION OF DRUG–DRUG INTERACTIONS BASED ON CONTRIBUTION OF CYTOCHROME P450 ENZYME TO DRUG METABOLISM USING THE RAF APPROACH IN DRUG DISCOVERY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Koji Yokoo1, Toshifumi Tohda2, Yoichi Naritomi1, Naoki Ueda2, Keitaro Kadono1, and Kenji Tabata1 1 Astellas Pharma Inc., Ibaraki, Japan and 2Astellas Research Technology Co., Ltd., Ibaraki, Japan During drug discovery, evaluation of the relative contributions of cytochrome P450 (CYP) enzymes to the metabolism of new drug candidates is required to assess issues such as potential drug–drug interactions (DDIs), genetic polymorphisms, and interindividual variations. Previous proposals for the evaluation of the relative contribution of each CYP isoform have included the relative activity factor (RAF), the inter-system extrapolation factor, and the inhibitor approaches. Here, we assessed the relationship between the relative contributions of CYP – calculated using the RAF approach – and DDI risk. The RAF approach was used to predict the relative contribution of each CYP isoform (CYP1A2, 2B6, 2C8, 2C9, 2C19, 2D6, and 3A4) to the net hepatic intrinsic clearance of drugs. RAF was estimated from the in vitro intrinsic clearance (CLint) of marker substrate for each CYP isoform (CYP1A2, phenacetin; 2B6, bupropion; 2C8, repaglinide; 2C9, diclofenac; 2C19, S-mephenytoin; 2D6, dextromethorphan; and 3A4, midazolam and nifedipine). The CLint of marker substrates and drugs were determined by measuring substrate depletion. The magnitude of in vivo DDIs induced by strong inhibitors of CYP3A4 was evaluated using the Metabolism and Transport Drug Interaction DatabaseTM. The RAF of each CYP and CLint using microsomes expressing recombinant enzymes and human liver microsomes (CLint,HLM) was obtained for 24 drugs. For CYP3A4, the RAF of nifedipine (16.7) was different from that of midazolam (56.2). Differences between CLint,HLM and predicted CLint were within 2-fold for 14 of the 24 drugs when using RAF calculated by nifedipine and for 16 of the 24 drugs when using RAF calculated by midazolam. In contrast, by using RAF calculated by nifedipine or midazolam, which can be successfully predicted CLint,HLM, the number of compounds with good prediction (within 2-fold) increased to 20 of the 24 drugs. A correlation between fraction metabolized of CYP3A4 (fm,CYP3A4) and area under the curve (AUC) ratio with strong inhibitors of CYP3A4 was observed. For drugs exhibiting fm,CYP3A4 less than 0.6, a modest DDI risk was observed (AUC ratio 52-fold). In contrast, drugs with an AUC ratio greater than 5 had a high fm,CYP3A4 (40.9). The RAF approach enables the efficient estimation of the relative contributions of CYP. Results from microsomal substrate depletion experiments can be used to assess issues with DDIs.

P156. USE OF NOMINAL OR TIME-AVERAGED MEDIA CONCENTRATIONS OF TEST COMPOUNDS IN PLATED HEPATOCYTE INDUCTION ASSAYS TO DETERMINE EC50 VALUES AND RELATIVE INDUCTION SCORES: IMPACT ON PREDICTION OF CYP3A4 INDUCTION POTENTIAL IN VIVO George Zhang, Thuy Ho, Robert J. Clark, Lisa G. Fox, and David M. Stresser Corning Gentest Contract Research Services, Corning Life Sciences, Woburn, MA, USA Cytochrome P450 (CYP) induction is often considered a liability in drug development and this has prompted the development of in vitro models to predict CYP induction potential in vivo. As we previously reported, a strong and predictive relationship between the extent of midazolam AUC change in human and the relative induction score (RIS) calculated from either mRNA or enzyme activity as endpoint was found when EC50 and RIS were calculated from nominal concentrations in assay media for 9 compounds. In this study, we measured the time-weighted average media concentration of test compounds during the final 24 hours of a 48-hour treatment period in a CYP3A4 hepatocyte induction assay. With endpoints of CYP3A4 mRNA or testosterone 6b-hydroxylase activity, EC50 values and RIS were determined and compared using either the time-averaged or nominal concentrations. For compounds exhibiting no significant loss during the last day of treatment, EC50 values were found to be similar between the two methods. However, as expected, much lower EC50 values were found for test compounds such as nifedipine and clotrimazole that had undergone substantial metabolic depletion. Somewhat surprisingly, construction of RIS calibration curve based on ‘‘time-averaged’’ EC50 values showed no significant improvement to the calibration curve fits or subsequent predictions, based on R2, root mean square and geometric mean fold error analysis. Results from this limited data set of one donor support the use of calibration curves generated from in vitro mRNA induction response curves incorporating use of either nominal or measured concentrations to predict CYP3A4 induction potential in human.

P157. MONOCLONAL ANTIBODY AGAINST CD30 (IRATUMUMAB) INDUCES HEPATIC CYP ENZYMES IN VITRO: AN ATTRACTIVE MODEL TO STUDY DRUG–DRUG INTERACTIONS Maciej Czerwinski1, Catherine Wiegand1, Rachel Williams2, Masood Khan2, and David B. Buckley1 1 XenoTech, LLC, Lenexa, KS, USA and 2KCAS, Bioanalytical Services, Shawnee, KS, USA Some therapeutic proteins (TP) cause a cytokine-mediated suppression of hepatic drug-metabolizing enzymes which results in pharmacokinetic drug interactions of TP with co-administered small molecule drugs. Here we evaluated the blood-mediated effects of immuno-modulatory TP, namely, Muromonab (Orthoclone, a murine mAb against CD3) and Iratumumab (MDX-060,

Downloaded by [112.134.238.102] at 07:30 10 February 2016

112

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

a human mAb against CD30) on drug-metabolizing enzymes in primary human hepatocytes co-cultured with Kupffer cells. Human whole blood was treated with Muromonab (1 mg/mL) or Iratumumab (100 mg/mL) at 37  C for 24 h to model the in vivo release of cytokines from peripheral blood mononuclear cells (PBMC). Plasma levels of cytokines were measured with electrochemiluminescence detection-based sandwich immunoassay. Subsequently, plasma from the TP-treated blood and controls was added to the hepatocyte co-cultures (50% v/v in cell culture medium) for 72 h to evaluate effects of the plurality of released cytokines on the mRNA expression of constitutive androstane receptor (CAR), pregnane X receptor (PXR), CYP enzymes (1A2, 2B6, 2C9, 3A4, 3A5), alpha-1-acid glycoprotein 1 (ORM1) and C-reactive protein (CRP) by RT-PCR. Muromonab, caused a significant (p50.05, t-test: paired two samples for means) increase in concentration of IFN-g (interferong, in 5 out of 7 donors), IL-1b (interleukin-1b, 1/7), IL-2 (2/7), IL-4 (2/7), IL-5(1/7), IL-6 (1/7), IL-8 (3/7), IL-13 (1/7) and TNFa (tumor necrosis factor-a, 3/7) as compared to vehicle control. Muromonab caused a significant decrease in IL-1b (3/7), IL-6 (1/7), IL-8 (1/7), IL-13 (1/7) and TNF-a (1/7). Correspondingly, Muromonab increased mean plasma concentration (n = 7) of all cytokine except IL-1b and IL-6. On the other hand, Iratumumab caused a significant decrease in IFN-g (in 2 out of 8 donors), IL-1b (8/8), IL-6 (8/8), IL-8 (8/8), IL-10 (6/8), IL-12p70 (1/8), IL-13 (2/8) and TNF-a (7/8) and an increase in IL-5 (2/8) and TNF-a (1/8). Similarly, Iratumumab decreased mean plasma concentration (n = 8) of all cytokines. Treatment of the co-cultures with Interleukin-6 (10 ng/mL), a positive control, resulted in marked suppression of PXR and CYP mRNA levels and induction of ORM1 and CRP mRNAs. Treatment of the co-cultures with Muromonab alone or with plasma from Muromonab-treated blood had little or no effect on the mRNA expression of CAR, PXR, CYP enzymes, ORM1 or CRP. Similarly, treatment of the co-cultures with Iratumumab alone had little or no effect on mRNA expression. In contrast to Muromonab-treated plasma, plasma from Iratumumab-treated blood caused time-dependent increase of CYP1A2, CYP2B6, CYP2C9, CYP3A4 and CYP3A5 mRNA levels and suppression of CRP mRNA levels. These results suggest that applying plasma from human blood treated with a therapeutic protein to hepatocytes co-cultured with Kupffer cells can help identify those therapeutic proteins with the potential to induce CYP expression by an indirect mechanism, namely the release of cytokines from PBMC. Furthermore, these results indicate for the first time that a CYP-enzyme induction, not only suppression, could facilitate an interaction between TP and small molecule drugs. This study demonstrates that hepatocytes co-cultured with Kupffer cells can be used as an effective test system for evaluation of direct or cytokine-mediated drug interactions between immuno-modulatory TP and small molecule drugs.

P158. EFFECTS OF THE RUTA GRAVEOLENS EXTRACT AND ITS INGREDIENTS ON CYTOCHROME P450-DEPENDENT MONOOXYGENASE IN MICE Yune-Fang Ueng1, Chien-Chih Chen2, I-Jung Lee2, Yu-Ling Huang2, Chiung-Chiao Huang2, and Chul-Ho Yun3 1 Division of Basic Chinese Medicine, National Research Institute of Chinese Medicine, Taipei, Taiwan, 2National Research Institute of Chinese Medicine, Taipei, Taiwan, and 3School of Biological Sciences and Technology, Chonnam National University, Gwangju, South Korea Ruta graveolens (rue) has been used for the treatment of heart disease, wounds and rheumatism. To elucidate the effects of rue on cytochrome P450 (CYP, P450)-dependent monooxygenase, C57BL/6 J mice were treated with the aqueous extract of rue. Oral administration of rue extract to males increased hepatic Cyp1a and Cyp2b activities in a dose-dependent manner. Under a 7-day treatment regimen, rue (0.5 g/kg) induced Cyp1a and Cyp2b activities and protein levels in the livers of males and females, but not in the kidneys. Cytochrome c reduction activity of the electron-transfer partner, NADPH-P450 reductase was not affected by this treatment. Compared with the P450 induction by prototype inducers, 3-methylcholanthrene (CYP1A) and phenobarbital (CYP2B), the induction by rue was relatively weak. Based on the contents of rutin and furanocoumains in the rue extract, mice were treated with a mixture of rutin and furanocoumains for 7 days. The mixture of rue-ingredients also increased hepatic Cyp1a and Cyp2b activities in males, suggesting their contribution at least in part to the P450 induction by rue extract.

P159. IDENTIFICATION OF TRUE INHIBITOR CONCENTRATION THAT DETERMINES CYTOCHROME P450-MEDIATED DRUG–DRUG INTERACTION Shinji Iwasaki, Hideki Hirabayashi, Miyuki Funami, and Nobuyuki Amano Drug Metabolism & Pharmacokinetics Research Laboratories, Takeda Pharmaceutical Company Limited, Fujisawa, Japan It is important to predict the possibility of hepatic cytochrome P450 (CYP)-mediated drug–drug interaction (DDI) for the development of new drugs, and the methodology of DDI prediction has been discussed for a long time. Although the classical theory indicates that the unbound drug concentration in the liver should be used in the prediction of DDIs, it has not been experimentally demonstrated yet. In this research, we attempted to develop an accurate DDI prediction method from in vitro. To identify which inhibitor concentration should be employed for the most accurate prediction of the magnitude of DDI, we utilized an in situ rat liver perfusion technique and investigate DDI effects in terms of the intrinsic clearance (CLint) of nifedipine, the CYP substrate probe. The in situ CLint for nifedipine was decreased by co-infusion of typical CYP inhibitors, fluconazole, ketoconazole, ritonavir or indinavir. The decrease in in situ CLint was the most comparable to that in in vitro CLint against the unbound liver concentration ([I]liver,u), compared with the other in situ concentrations investigated: total and unbound concentrations in the inlet (portal vein), outlet (hepatic vein) and liver tissue. These results suggested that [I]liver,u is the most

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

113

reliable determinant for CYP-mediated DDI. To test the above findings in vivo, we also investigated DDI effects of orally administered nifedipine in rats with constant infusion of fluconzole, ketoconazole or ritonavir. The accurate prediction of magnitude of DDI was achieved when [I]liver,u was applied as the inhibitor concentration. In conclusion, we demonstrated that the [I]liver,u is the most reasonable determinant in the prediction of CYP-mediated DDIs. The technique and methodology taken in this study would provide valuable insights into prediction of DDI.

P160. Quantitative prediction of drug–drug interactions by PBPK model using the in vivo Ki values

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Akiko Kobashi1, Jose Martin Ciloy1, Mayumi Matsushita1, Kazuya Maeda2, and Kiyomi Ito3 1 Fujitsu Kyushu Systems Limited, Fukuoka, Japan, 2Laboratory of Molecular Pharmacokinetics, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan, and 3Research Institute of Pharmaceutical Sciences, Musashino University, Tokyo, Japan Objective: Quantitative predictions of drug–drug interaction (DDI) based on metabolic inhibition rely on the values of the inhibition constant (Ki) of the perpetrator drug. Based on an analysis of DDI studies involving 11 metabolic inhibitors, Kato et al. have reported a discrepancy between the Ki values obtained from in vitro studies (in vitro Ki) and those obtained from fitting the results of clinical DDI studies to a physiologically-based pharmacokinetic (PBPK) model (in vivo Ki). They have also reported that predictability of the PBPK model is best achieved using in vivo Ki values and that the accuracy of DDI predictions is improved by correcting the in vitro Ki values based on the correlation between Ki ratio(vivo/vitro) and the clogP value (predicted in vivo Ki). In this study, we compared the prediction accuracy of a PBPK model using three different Ki values(in vitro Ki, predicted in vivo Ki, and in vivo Ki) for a larger set of data. Methodology: Based on a PBPK model incorporating the reversible inhibition of hepatic metabolism, in vivo Ki values were calculated for 85 DDI studies involving 23 reversible inhibitors of cytochrome P450 enzymes, including those reported by Kato et al. In vitro Ki values for these inhibitors were collected from literature and the correlation between their Ki ratio (vivo/vitro) and the clogP value was investigated to derive a formula for calculating the predicted in vivo Ki. 109 DDI studies reported in the literature were quantitatively predicted by the PBPK model using the software DDI Simulator V2.3 (Fujitsu Kyushu Systems Limited) and the predictive accuracies were compared between the cases using in vitro Ki, predicted in vivo Ki, and in vivo Ki. Results and discussions: The Ki ratio (vivo/vitro) correlated well with the clogP value, which was used to calculate the predicted in vivo Ki values. DDI prediction using in vivo Ki gave the highest accuracy followed by predicted in vivo Ki and in vitro Ki in that order. This finding shows that for compounds whose in vivo Ki is unknown, the use of predicted in vivo Ki allows for more accurate quantitative predictions of DDI than using in vitro Ki without correction. In particular, predictions using predicted in vivo Ki is more effective for lipophilic inhibitors.The use of PBPK model has been proposed in the FDA draft guidance and also in the new Japanese draft guideline to assess the need for further DDI clinical studies of a drug that is under development. The use of predicted in vivo Ki and in vivo Ki is expected to facilitate the DDI predictions with high accuracy.

References Kato M, Shitara Y, Sato H, et al. (2008). The quantitative prediction of CYP-mediated drug interaction by physiologically based pharmacokinetic modeling. Pharm Res 25:1891–1901.

P161. THE INHIBITORY EFFECTS OF GASTROINTESTINAL AND RESPIRATORY DRUGS ON HUMAN HEPATIC MICROSOMAL CYTOCHROME P450 ACTIVITIES Mariko Iwase, Yuki Nishimura, Norimitsu Kurata, Atsushi Tsukurimichi, and Katsuji Oguchi Department of Pharmacology, School of Medicine, Showa University, Tokyo, Japan Over the counter (OTC), drugs have an important role for self-medication. However, pharmacokinetic properties of OTC drugs are not fully investigated and the reports concerning to their drug interaction are insufficient. Our previous data showed that OTC drugs such as antihistamines inhibited cytochrome P450(CYP) activities, suggesting the risk of drug interaction. Several gastrointestinal and respiratory drugs are available as OTC drugs. Because of the pharmacological properties, these drugs are often to use concomitantly with other drugs.Therefore, it is important to predict the possible drug interactions of these drugs by using the in vitro assay method. In the current study, we investigated the inhibitory effects of gastrointestinal and respiratory drugs on CYP activities in human liver microsomes. Each CYP activity was assayed using conventional substrates as follows: ethoxyresorfin O-deethylation (CYP1A2), S-warfarin 7-hydroxylation (CYP2C9), S-mephenytoin 4-hydroxylation (CYP2C19), bufuralol 10 -hydroxylation (CYP2D6) and midazolam 10 -hydroxylation (CYP3A). Gastrointestinal drugs (loperamide hydrochloride, oxethazaine, papaverine hydrochloride, pirenzepine hydrochloride hydrate, trimebutine maleate) and respiratory drugs (cloperastine hydrochloride, L-carbocisteine and potassium guaiacolsulfonate, diprophylline, methoxyphenamine hydrochloride) were estimated for its inhibitory effect on CYP activities. The apparent Ki values and the inhibition manner of these drugs on CYP activities were evaluated by nonlinear regression analysis. Preincubation study was also performed to clarify their possibility of mechanism-based inhibition. Loperamide, oxethazaine and cloperastine showed competitive inhibitory effect on CYP2D6 activities with low Ki values. In all of the drugs described above, cloperastine extremely inhibited CYP2D6 activity with apparent Ki value of 0.15 mM. CYP3A4 activity was inhibited by trimebutine, papaverine and oxethazaine. Their

114

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

inhibition manners were competitive and the Ki values were lower than 10 mM. Any other notable inhibitory effects on CYP activities were not observed by other drugs employed in this study. Preincubation analysis did not show any remarkable effects. This study showed that trimebutine, papaverine and oxethazaine inhibited CYP3A4 activity. The Ki values of trimebutine and papaverine were closed to the reported Cmax values, administered orally in human. It is noteworthy that cloperastine strongly inhibited CYP2D6 activity with extermely low Ki value. These results suggested that the drug–drug interactions mediated by these CYPs might occur during the treatment using these OTC drugs. To avoid the risk of adverse reaction, further appropriate human study will be needed.

P162. A "HUMAN FIRST" APPROACH TO ASSESS FRACTION METABOLIZED (FM) AND INFORM VICTIM DRUG INTERACTIONS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Jackie Bloomer, and Aarti Patel Department of Drug Metabolism and Pharmacokinetics, GlaxoSmithKline, Hertfordshire, UK In vitro enzymology methods to determine the major enzymes responsible for metabolism are well established and regularly applied early in drug discovery and development. However, without information on the routes of drug elimination in human, the clinical implications of this data are unclear. This can have a significant impact on clinical development plans including the conduct of unnecessary drug interaction studies and exclusion of co-medications in clinical trials, thereby impacting recruitment. We have investigated the outcomes of CYP3A4 victim drug interaction studies in GSK to illustrate the value of a ‘‘human first’’ approach to assessing risk and defining drug interaction strategies. The drug interaction results from over 10 studies were collated along with any relevant in vitro enzymology, preclinical and human metabolism data for each drug. The fraction metabolised (fm) by CYP3A4 was then back calculated from the drug interaction data and retrospectively predicted from the in vitro enzymology data and/or preclinical metabolism data using recognized extrapolation approaches. This analysis indicated that the fm by CYP3A4 and the subsequent drug interaction was over-estimated in over half of the cases investigated, which resulted in some unnecessary drug interaction studies. The limited value of predicting human fm from preclinical metabolism data was evident and the caution required when using in vitro metabolism data to predict drug interactions in the absence of clinical drug elimination data was highlighted. These observations have led to a revised strategy of understanding the routes of elimination in human before interpreting, or even generating, in vitro metabolism data in early discovery and development. In addition metabolism in preclinical species is not required prior to FTIH. This ‘‘human first’’ strategy is achievable by advances in metabolite identification and non-invasive collection of bile that can be applied in FTIH. Observations of the proportion of metabolites versus parent drug in human urine and bile can help define more relevant and bespoke in vitro substrate experiments (enzymes or transporters) to understand the mechanisms of drug elimination. The outcome of these experiments, when considered with the human data, will help optimize the conduct, design and timing of clinical drug interaction studies.

P163. ABSTRACT WITHDRAWN

P164. TESTING FOR POTENTIAL INTERACTION BETWEEN PHELA, A HERBAL EXTRACT AND CYCLOSPORINE, CYCLOPHOSPHAMIDE AND DEXAMETHASONE Makhotso Rose Lekhooa1, Andrew Walubo1, J. B. du Plessis1, and M. G. Matsabisa2 1 Pharmacology Department, University of the Free State, Bloemfontein, South Africa and 2Indigenous-Knowledge-System, Medical Research Council, Cape Town, South Africa Introduction: Phela is a herbal extract, derived from four South African traditional medicinal plants. It has been used for decades for wasting conditions. Previous anecdotal reports of patients and healers indicated that Phela stimulates the immune system. Furthermore, preliminary results showed Phela’simmunomodulatory effects in healthy rats. Currently, Phela is under development for use as an immune booster in immune-compromised individuals. However its drug–drug interactions profile is unknown. Cyclosporine, cyclophosphamide and dexamethasone are known immunesuppresants with varying mechanism of action. Since the subsequent animal experiment in our laboratory would involve co-administration of the drugs [cyclosporine (CsA), cyclophosphamide (CP) and dexamethasone (Dex)] with Phela, there was a need to test for potential interactions with Phela. Thus, the aim of the experiment was to screen for direct interactions with any of the three drugs using an equilibrium dialysis by slide-a-lyzer technique. Methods: Plasma (3 ml) spiked with the drugs[cyclosporine, cyclophosphamide and dexamethasone] separately was dialyzed against 0.1 M of sodium phosphate buffer (pH 7.4) until equilibrium protein binding for each drug was reached, after which Phela was added and dialysis was continued for six more hours. Aliquots (200 ml) were taken every 2 hours from plasma in the chamber and buffer, and were analysed for the respective drug concentrations. Thereafter, the free and bound concentrations were determined. Results: Cyclophosphamide and dexamethasone equilibrated at 4 hours and cyclosporine at 22 hours. Free fraction (%) was (mean ± SD): [CP-only, 75.83 ± 22.26% vs CP + Phela, 37.90 ± 9.57%, p = 0.054; Dex-only, 91.21 ± 5.14% vs Dex + Phela, 87.75 ± 1.38%, p = 0.085; and CsA-only, 55.95 ± 12.92% vs CsA + Phela, 79.34 ± 12.18%, p = 0.323]. Despite wide variations, there was no significant difference between the free fractions of drug-only

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

115

group when compared with drug + Phela group. Conclusion: In conclusion, there was no interaction between Phela and cyclosporine, cyclophosphamide and dexamethasone. This implies that the drugs can be co-administered with Phela without interference.

P165. SUCCESSFUL APPLICATION OF PBPK MODELING TO SIMULATE THE LACK OF BRINCIDOFOVIR INHIBITION OF CYP3A4-MEDIATED CLEARANCE OF MIDAZOLAM

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Alison Margolskee1, Mark A. Bush2, and Tim Tippin1 1 Chimerix, Inc., Durham, NC, USA and 2Nuventra, Inc., Durham, NC, USA Background: Brincidofovir (BCV) is an orally administered nucleotide lipid conjugate in Phase 3 clinical development for the prevention of cytomegalovirus infection in hematopoietic stem cell transplant recipients, and is also being evaluated for treatment of adenovirus infection. BCV exhibits reversible in vitro inhibition of CYP3A4 with an IC50 of approximately 12 mM. The purpose of this analysis was to characterize the BCV drug–drug interaction with midazolam (MDZ) by developing physiologically-based pharmacokinetic (PBPK) models for each drug. Methods: A ‘‘lumped’’ tissue PBPK scheme was developed for BCV, which included slowly perfused tissue, highly perfused tissue, kidney, fat, liver, and an intestinal compartmental absorption and transit model. Tissue:blood partition coefficients were estimated for BCV using calculated pKa (2.4) and log P (5.9) and an experimentally determined unbound fraction in plasma of 0.1%. Passive and hypothetical carriermediated intestinal transport of BCV were incorporated to achieve an adequate model to describe the plasma concentration versus time data. A PBPK model for MDZ, adapted from Fenneteau-2010, was used to simulate MDZ plasma concentrations following IV or oral administration. Results: The proposed PBPK model adequately described the BCV plasma concentrations obtained after oral administration of BCV to 30 healthy subjects, based on the chi square statistic and comparison of observed and simulated PK parameters (values within 2-fold of actual). Based on similar criteria, the proposed PBPK model adequately described MDZ plasma concentrations. Simulations from the linked BCV/MDZ models predicted a slight increase in MDZ plasma exposures. Simulated area under the plasma concentration-time curve ratios (AUCR, defined as AUCMDZ + BCV/AUCMDZ alone) were 1.06 and 1.02 following oral and IV administration of MDZ, respectively, suggesting a minimal effect of BCV on MDZ plasma concentrations. Clinical results were consistent with PBPK simulations. In the clinical study, single oral (2.5 mg) and IV (1 mg) doses of MDZ were administered to healthy subjects either alone or combination with single oral doses of BCV (200 mg). MDZ AUCRestimates were 1.12 and 1.05 following oral and IV administration of MDZ, respectively, confirming the linked PBPK model predictions. Conclusions: PBPK models were developed for BCV and MDZ that adequately simulated plasma concentrations for each drug when administered alone and in combination. Based on simulations and clinical results, no meaningful interaction between BCV and MDZ was observed. Therefore, no DDIs are expected with other CYP3A substrates in the BCV patient population. Overall, this work demonstrates the general utility of PBPK modeling for the prediction/ confirmation of drug–drug interaction potential, and in particular to assess the low potential for BCV to affect the PK of MDZ via inhibition of CYP3A4.

P166. ANALYSIS OF DRUG–DRUG INTERACTIONS BETWEEN CERIVASTATIN AND INTERACTING DRUGS, CYCLOSPORINE A AND GEMFIBROZIL, USING PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODEL Soo-Jin Kim1, Yoshiaki Yao2, Takashi Yoshikado1, Kota Toshimoto1, and Yuichi Sugiyama1 1 Sugiyama Laboratory, RIKEN Innovation Center, RIKEN, Yokohama, Japan and 2Analysis & Pharmacokinetics Research Labs, Drug Discovery Research, Astellas Pharma Inc., Ibaraki, Japan Cerivastatin (CER) was withdrawn from the world market due to fatal side effect, rhabdomyolysis. CER is taken up into the liver by organic anion transporting polypeptide (OATP)1Bs and then metabolized by CYP3A4 and CYP2C8. Coadministrations of CER and cyclosporine A (CsA) or gemfibrozil (GEM) were shown to increase the plasma concentration of CER by 2.5–5 fold. CsA is an inhibitor of OATP1B1 and CYP3A4, whereas GEM and its metabolite, GEM-glucuronide, are inhibitors of OATP1B1 and CYP2C8. The increase of systemic exposure of CER by DDI may result in myopathy and/or fatal rhabdomyolysis. The purpose of this study was to analyze these DDIs based on unified PBPK model including hepatic uptake and metabolism of CER. The plasma concentration profiles of CER, CsA and GEM/GEM-glucuronide were reproduced by a PBPK model. Tissue-to-blood concentration ratios (Kp) in the peripheral compartments were predicted by the in silico calculation. Relevant in vitro and in vivo parameters for PBPK model were obtained from literatures, our previous studies and in vitro experiments. The concentration-time profiles for CER with and without coadministration of CsA or GEM were simultaneously fitted to obtain the optimized parameters for inhibitor constant (Ki) against OATP1B1 and CYPs. The analyses indicated that the major mechanism of the in vivo DDI by CsA or GEM was the inhibition of OATPs mediated uptake or the inhibition of CYP2C8 mediated metabolism, respectively. These results were consistent with the previously reported consequences. In addition, the in vivo Ki value for OATP1B1 inhibition by CsA obtained was comparable with an in vitro Ki value obtained in the presence of CsA preincubation using the OATP1B1-expressing HEK293 cells. The present PBPK modeling approach is expected to be applicable to quantitative predictions of other DDIs involving both transporters and metabolic enzymes, providing valuable information for increasing the efficacy of drug development and the avoidance of toxic interactions in clinical studies.

116

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P167. INDIRECT INHIBITION OF ORGANIC ANION TRANSPORTING POLYPEPTIDE (OATP) 1B3-MEDIATED TRANSPORT BY PROTEASOME INHIBITOR BORTEZOMIB

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Khondoker Alam1, John Powell1, Kai Ding2, and Wei Yue1 1 Department of Pharmaceutical Sciences, University of Oklahoma Health Sciences Center, Oklahoma city, OK, USA and 2 Department of Biostatistics and Epidemiology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA Purpose: OATP1B3 is a liver-specific membrane transport protein which mediates hepatic uptake of many clinically important drugs (e.g. statins and paclitaxel). To date, OATP-mediated drug–drug interactions have focused primarily on direct mechanisms of inhibition by drugs/metabolites. Although it is likely that indirect mechanism(s) may be involved in OATP-mediated drug– drug interactions (DDIs) through modulation of OATP1B3 function, currently, little is known about how OATP1B3 function is regulated. Proteasome has emerged as a new therapeutic target. Proteasome inhibition is associated with aging and disease states. We hypothesized that OATP1B3 is post-translationally modified by ubiquitination and that proteasome inhibition affects OATP1B3 transport function. The aim of this study was to determine the ubiquitin-modification of OATP1B3 and the impact of bortezomib, the first-in-class proteasome inhibitor, on OATP1B3 transport function in a HEK293-OATP1B3 stable cell line and human sandwich-cultured hepatocytes (SCH). Methods: Ubiquitination of OATP1B3 was determined by co-immunoprecipitation of HA-Ub and FLAG-OATP1B3 in HEK293 cells transiently transfected with plasmid vectors expressing HA-Ub, FLAGOATP1B3 or both. Direct inhibition of bortezomib (0.1. 5 mM) to OATP1B3-mediated [3H]CCK-8 transport (1 mM, 3 min) was determined in a HEK293-OATP1B3 stable cell line. To determine the potential indirect effect of bortezomib on OATP1B3 transport function, HEK293-OATP1B3 and human SCH were pre-treated with bortezomib, 100 nM and 250 nM, respectively, or vehicle control for up to 7 h. After rinsing with HBSS buffer, [3H]CCK-8 accumulation (1 mM, 3 min) was compared between bortezomib treatment and control. LDH toxicity assay was performed to determine cytotoxicity following bortezomib treatment. Results: HA-Ub and FLAG-OATP1B3 were co-immunoprecipitated in HA-Ub/FLAG-OATP1B3-cotransfected HEK293, indicating that OATP1B3 undergoes ubiquitination. Bortezomib did not directly affect [3H]CCK-8 uptake in HEK293OATP1B3. However, bortezomib pre-treatment for as short as 2 h significantly decreased [3H]CCK-8 uptake both in HEK293OATP1B3 and human SCH. In HEK293-OATP1B3 cells, 2 h and 7 h bortezomib pre-treatment significantly decreased [3H]CCK-8 accumulation to 75.6 ± 7.7% and 71.9 ± 2.5% of control, respectively (n = 3). In human SCH, 7 h bortezomib pretreatment significantly decreased [3H]CCK-8 accumulation to 61.6 ± 9.9% of control (n = 4 livers). LDH assay indicated negligible toxicity after bortezomib treatment. Conclusions: This is the first indication that OATP1B3 is post-translationally modified by ubiquitin and that proteasome inhibitor inhibits OATP1B3 transport function. The inhibition of OATP1B3-mediated transport by bortezomib occurred in an indirect manner, presumably by affecting ubiquitin modification of OATP1B3. Inhibition of OATP1B3 transport function by bortezomib highlights the clinical significance of proteasome inhibition as a potential determinant for OATP1B3-mediated DDIs affecting the pharmacokinetics and efficacy of OATP1B3 substrates (e.g. statins, paclitaxel). This work wqas supported by NIH R01 GM094268.

P168. PRECISE ESTIMATIONS OF THE INHIBITION CONSTANTS BY PBPK ANALYSES OF METABOLITES’ PHARMACOKINETIC ALTERATIONS USING CLUSTER NEWTON METHOD Kazuya Maeda1, Kenta Yoshida1, Yuichi Sugiyama2, Akihiko Konagaya3, and Hiroyuki Kusuhara1 1 Department of Molecular Pharmacokinetics, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan, 2Sugiyama Laboratory, RIKEN Innovation Center, RIKEN, Research Cluster for Innovation, Yokohama, Japan, and 3 Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, Yokohama, Japan Quantitative predictions of the degrees of pharmacokinetic drug–drug interactions (DDIs) are important for the developments of new drugs or for the proper use of existing drugs. There are two key factors determining the degrees of DDIs caused by the inhibition of metabolic enzymes: the inhibition constants of coadministered drugs (Ki) and the contribution of the metabolic enzyme(s) to the overall eliminations of substrate drugs (fm). While precise estimation of these two parameters are essential in the accurate predictions of DDIs, high variability of estimated values in vitro makes ‘‘bottom-up’’ predictions of DDIs unreliable. It is also difficult to estimate these values from clinical observations (‘‘top-down’’ approach) to validate in vitro values or to predict the effect of new combinations, owing to the complementary nature of these two factors in determining DDIs. We hypothesized that physiologically-based pharmacokinetic (PBPK) analyses of the pharmacokinetic alterations in clinical DDI cases for not only substrate drugs, but also metabolites of substrate drugs, makes the simultaneous determination of these two parameters possible. In this study, to overcome the difficulty of estimating pharmacokinetic parameters for both parent drugs and metabolites, we introduced a new method called Cluster Newton Method (CNM) (Aoki et al., 2011; Yoshida et al., 2013), which allows us to estimate large number of unknown parameters from limited information available. We first collected DDI cases with information of metabolites’ pharmacokinetic alterations using the University of Washington Drug Interaction Database (http://www.druginteractioninfo.org). Pharmacokinetic alterations in the collected cases were then analyzed with PBPK models and CNM under MATLAB software environment, with or without including the information of metabolites in the analyses. As a result, multiple set of pharmacokinetic parameters reproducing observed plasma concentration-time profiles

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

117

and/or urinary accumulations were successfully obtained for respective cases. In most cases, inclusion of metabolites pharmacokinetic information in analyses markedly improved the precisions of estimated Ki and fm. Furthermore, estimated Ki from different DDI cases were similar to each inhibitor, indicating that the obtained values are reliable and can be used to predict new DDIs. These results suggest the importance of metabolites information in precise determination of parameters in PBPK analyses of DDIs.

References Aoki Y, Hayami K, De Sterck H, Konagaya A. (2011). Cluster Newton method for sampling multiple solutions of an underdetermined inverse problem: Parameter identification for pharmacokinetics. NII Technical Report, NII-2011-002 E. Yoshida K, Maeda K, Kusuhara H, Konagaya A. (2013). Estimation of feasible solution space using Cluster Newton Method: Application to pharmacokinetic analysis of irinotecan with physiologically-based pharmacokinetic models. BMC Syst Biol 7:S3.

P169. INHIBITION OF OATP1B1 BY GLUCURONIDES: DOES IN VITRO TRANSLATE TO IN VIVO RELEVANCE?

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Rebecca A. Sulivan1, Aleksandra Galetin1, Pradeep Sharma2, Helen Rollison2, and Katherine Fenner2 1 Centre for Applied Pharmacokinetic Research, Manchester Pharmacy School, University of Manchester, Manchester, UK and 2 Department of Drug Metabolism and Pharmacokinetics, Drug Safety and Metabolism, AstraZeneca, Cheshire, UK Transporter mediated drug–drug interactions (DDI) are increasingly recognised to be of clinical relevance and the requirement for their investigation is highlighted in the latest EMA and FDA guidelines. A number of drugs have been reported to inhibit hepatic uptake transporter OATP1B1, including metabolites like cyclosporine AM1. In addition, recent studies have reported more pronounced inhibition of OATP1B1 in vitro, following a pre-incubation step, as illustrated in the case of cyclosporine and its AM1 metabolite1. The aim of the current work was to investigate the inhibitory potential of a range of compounds (including 10 glucuronides) against OATP1B1 and to assess the effect of pre-incubation using estradiol 17B glucuronide (E17BG) and pitavastatin as probes. Initially, OATP1B1 inhibition was characterised for 18 compounds in total with E17BG as a probe at 0.02 mM and over a two minute incubation time. The inhibitory potential was assessed in stably transfected HEK-293 cells on three separate occasions with and without a 30-minute pre-incubation step; in all cases rifamycin (0–100 mM) was used as a positive control. Subsequent experiments were performed for a subset of 13 most potent inhibitors using pitavastatin (1 mM) as a clinically relevant OATP1B1 probe. Potential OATP1B1 DDI risk in vivo was assessed for a subset of inhibitors using a static model and the in vitro data generated. Of all glucuronides, telmisartan and repaglinide glucuronides were identified as the most potent OATP1B1 inhibitors with IC50 values of 1.4 and 5.4 mM, respectively, following pre-incubation using E17BG as probe; comparable values were obtained with pitavastatin. For the remaining glucuronides IC50 values were410 mM. Parent compound OATP1B1 IC50 ranged from 0.6–92 mM for telmisartan and ezetimibe, respectively and inhibitory potency was similar to or less than that of glucuronide metabolites. E17BG and pitavastatin IC50 values showed comparable OATP1B1 inhibition, with 7/13 IC50 within 2-fold; excellent agreement was seen in the case of cyclosporine, rifampicin and telmisartan glucuronide. A timedependent increase in inhibitory potency was not observed in all cases and trends were inconsistent between probes/inhibitors. Of all glucuronides investigated, risk of clinically relevant OATP1B1 inhibition was only seen in the cases of telmisartan and gemfibrozil glucuronides (R41.25), highlighting the possible contribution of glucuronides to in vivo transporter mediated DDIs (Gertz et al., 2013).

Reference Gertz M, Cartwright CM, Hobbs MJ, et al. (2013). Cyclosporine inhibition of hepatic CYP3A4, uptake and efflux transporters: Application of PBPK modelling in the assessment of drug–drug interaction potential. Pharm Res 30:761–780.

P170. APPLICATION OF A PHYSIOLOGICALLY BASED POPULATION PHARMACOKINETIC MODEL FOR SIMVASTATIN AND ITS ACTIVE METABOLITE SIMVASTATIN ACID TO PREDICT CLINICALLY RELEVANT DRUG–DRUG INTERACTIONS Nikolaos Tsamandouras, Amin Rostami-Hodjegan, Aleksandra Galetin, and Leon Aarons Centre for Applied Pharmacokinetic Research, Manchester Pharmacy School, University of Manchester, Manchester, UK Simvastatin (SV), a commonly used HMG-CoA reductase inhibitor, is a prodrug with complex pharmacokinetics due to the inter-conversion between the parent drug and its main active metabolite, simvastatin acid (SVA). Both SV and SVA undergo oxidative metabolism mainly by CYP3A4/5 and, therefore, are subject to clinically important drug–drug interactions (DDIs) and associated risk of myopathy. Therefore, in the current study, a model with physiologically realistic compartmental structure was developed allowing inter-conversion between the lactone (SV) and acid (SVA) form of the drug in different tissues. The model was developed with nonlinear mixed effects software (NONMEM 7.2) using SV and SVA plasma concentrations from 34 healthy volunteers. The ability of the developed model to predict DDI effects was assessed by evaluating the model’s predictive performance using a range of CYP3A inhibitors, namely clarithromycin, itraconazole, erythromycin and diltiazem. The concentration–time profiles of inhibitors in the sites of interaction were generated in Simcyp v13 and were used as forcing functions in the small intestinal wall and liver tissue compartments of the developed SV/SVA

118

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

mechanistic model. For each of the DDI investigated, the model predicted changes in SV and SVA Cmax and AUC were compared to the observed values from reported clinical DDI studies. Overall, the developed model predicted the increase in both SV and SVA AUC and Cmax within 2-fold of the observed values for the evaluated DDIs. The best model performance was observed in the case of clarithromycin DDI. Specifically, the model predicted an increase in plasma SV AUC and Cmax by 10.02 - and 6.49-fold respectively and 11.46- and 8.97-fold increase in SVA AUC and Cmax respectively, following a week of co-administration of 40 mg simvastatin with twice-a-day 500 mg clarithromycin. The respective observed values were 9.95, 7.14, 12.17 and 10, respectively (Jacobson, 2004). A significant advantage of the developed mechanistic model is that it can predict the DDI effect on the SV/SVA concentrations in the clinically relevant tissues such as the liver (site of action) and muscle (site of toxicity). Our model predictions indicate that the evaluated DDIs cause a significant increase in the muscle exposures of both SV and SVA, which is concordant with the clinically observed cases of rhabdomyolysis. In conclusion, the presented model-based approach can be of significant use during the drug development for assessing DDI risk of compounds likely to be co-administered with simvastatin.

Reference

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Jacobson TA. (2004). Comparative pharmacokinetic interaction profiles of pravastatin, simvastatin, and atorvastatin when coadministered with cytochrome P450 inhibitors. Am J Cardiol 94:1140–1146.

P171. ANTIBODY-DRUG CONJUGATES (ADCS) DRUG INTERACTIONS CAN BE PREDICTED USING A PHYSIOLOGICALLY-BASED PHARMACOKINETIC (PBPK) MODELING APPROACH Sophie Mukadam1, Divya Samineni2, Harvey Wong1, Ben-Quan Shen3, Dan Lu2, Sandhya Girish2, Cornelis Hop1, Yan Jin Jin2, Chunze Li2, and Yuan Chen1 1 Department of Drug Metabolism and Pharmacokinetics, Genentech, Inc., South San Francisco, CA, USA, 2Department of Clinical Pharmacology, Genentech, Inc., South San Francisco, CA, USA, and 3Department of Preclinical and Translational PKPD, Genentech, Inc., South San Francisco, CA, USA Background and objectives: While the therapeutic strategy for the antibody–drug conjugates (ADCs) entails a significant expansion of the therapeutic index of the chemotherapeutic, the development and optimization of their highly complex components still remain challenging. The unconjugated cytotoxic agent i.e. MMAE (monomethyl auristatin E) is expected to undergo clearance mechanisms consistent with small molecules. Hence, evaluation of the risk of drug–drug interaction associated with unconjugated cytotoxic agent is important in support of clinical development of ADCs. Methods: In this study, a PBPK model linking acMMAE (antibody conjugated MMAE) as a parent drug to the catabolite, MMAE, was developed and implemented in the SimcypÕ simulator using a mixed bottom-up and top-down approach. The initial built model was used to simulate PK of acMMAE and MMAE in patients dosed with the anti-CD22-vc-MMAE ADC. Subsequently, the model performance was verified by comparing predicted MMAE PK profiles with observed data obtained from dosing another vcMMAE ADC, i.e. Brentuximab vedotin. Finally, the validated model was used to simulate the DDI between brentuximab vedotin and midazolam, ketoconazole, and rifimpcin. The simulated DDI (AUC and Cmax ratio) were compared with the published data. Results: The developed acMMAE-MMAE linked PBPK model described the shape of acMMAE and MMAE PK profile reasonably well. The simulated Cmax and AUC for acMMAE and MMAE were in good agreement with observed mean values. The validation of the MMAE PBPK model was performed using brentuximab vedotin clinical PK data at the three clinical dose levels. The PK parameters predicted were comparable to the observed data. The model simulated DDI, expressed as AUC and Cmax ratios, were well within the 2 fold of the observed data from brentuximab clinical DDI study. Conclusion: This work demonstrates the utility of the first PBPK model developed to predict the MMAE PK following administration of the vc-MMAE ADC and provides a valuable tool for the prediction of MMAE drug interaction potential. It is conceivable that this model can be extended to other vc-MMAE ADCs to simulate the MMAE drug interaction potential and inform its DDI strategy during the clinical development.

P172. DRUG INTERACTION GUIDELINE FOR DRUG DEVELOPMENT AND LABELING RECOMMENDATIONS: FINAL DRAFT OF JAPANESE NEW GUIDELINE Yoshiro Saito1, Akihiro Hisaka2, Toshiyuki Kume2, Kazuya Maeda2, Hiroshi Suzuki2, Kiyomi Ito2, Ken-ichi Inui2, Yukio Kato2, Shogo Ozawa2, Hiroshi Watanabe3, Shin-ichi Miura4, Toshinari Mitsuoka5, Keiko Maekawa1, Masanobu Sato6, Akihiro Ishiguro6, Reiko Sato6, Naomi Nagai6, and Yasuo Ohno1 1 The MHLW Drug Interaction Working Group and National Institute of Health Sciences (NIHS), Tokyo, Japan, 2The MHLW Drug Interaction Working Group and the Japanese Society for the Study of Xenobiotics (JSSX), Tokyo, Japan, 3The MHLW Drug Interaction Working Group and the Japanese Society of Clinical Pharmacology and Therapeutics (JSCPT), Tokyo, Japan, 4 The MHLW Drug Interaction Working Group and Japan Pharmaceutical Manufacturers Association (JPMA), Tokyo, Japan, 5 The MHLW Drug Interaction Working Group, the Ministry of Health, Labour and Welfare, Tokyo, Japan, and 6The MHLW Drug Interaction Working Group and Pharmaceuticals and Medical Devices Agency (PMDA), Tokyo, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

119

Background: Evaluation of drug interactions (DIs) in the new drug development is important to secure efficacy and to reduce incidence of adverse reactions in clinical trial and proper use of drugs after marketing. The current Japanese regulatory document ‘‘Methods of Drug Interaction Studies (The MHLW, 2001)’’ from Ministry of Health, Labour and Welfare (MHLW) was issued in 2001. Since then, a lot of progress has been made in the research related to DI and the clinical information. New guideline (The EMA, 2013) and draft guidance (The US FDA, 2012) were also publicized recently from EMA and FDA, respectively. Based on these situations, we started investigation to draft a new Japanese guideline. Methods: Since December 2012, the MHLW DI-working group (WG) and three Sub-WGs (Transporters, Drug Metabolizing Enzymes and Modeling & Labeling) organized by experts from academia, industry and regulatory bodies, have intensively discussed and collaboratively worked with the other MHLW Group (Labeling Guideline Working Group) to complete the final draft version of new Japanese DI guideline. Results: The draft guideline was presented in the 10th ISSX Meeting in Oct. 2013, and notified in Dec. 2013 for public comments. After examining the comments from Japan and overseas obtained, the draft guideline was finalized in May 2014 and submitted to the MHLW. In addition to the overall update of existing regulatory document, the final draft supplemented detailed explanation in every area. Followings are major parts of the revision: (1) transporter studies and their general methods on absorption in small intestine, urinally excretion from kidney, and biliary excretion from liver; (2) lists of typical substrates and inhibitors of transporters in vivo and in vitro, (3) stepwise examination for DIs on metabolizing enzymes and transporters using decision trees; (4) evaluation by modeling and simulation using in vitro and in vivo data; (5) lists of known inhibitors and inducers on major cytochrome P450 isoforms classified in strongly-, moderately- and weakly-affecting groups based on the changes of AUC or CL/F, and a table of ‘‘sensitive and moderately sensitive drugs susceptible to pharmacokinetic DIs’’; (6) Interactions with therapeutic proteins; (7) labeling recommendations, and (8) detailed examples and rich references for more science-based approaches. Conclusion: The new Japanese guideline will be issued in the near future to effectively design necessary drug interaction studies during drug development. It is also expected that the guideline will be appropriately utilized for the purpose of safe and effective use of DI-related information after marketing.

References The MHLW. (2001). The method of drug interaction studies. Notification No.813, June 2001. Japan: Ministry of Health, Labor and Welfare. The EMA. (2013). Guideline on the investigation of drug interactions. January 2013. London, UK: European Medicines Agency. The US FDA. (2012). Drug interaction studies-study design, data analysis, implications for dosing, and labeling recommendations (draft guidance). February 2012. Maryland, USA: US Food and Drug Association.

P173. DRUG DISPOSITION AND DRUG–DRUG INTERACTION DATA IN 2013 FDA NEW DRUG APPLICATIONS: A SYSTEMATIC REVIEW Jingjing Yu, and Isabelle Ragueneau-Majlessi Drug Interaction Database Program, Department of Pharmaceutics, University of Washington, Seattle, WA, USA Background: The aim of the work was to perform a systematic analysis of DMPK, transport and drug interaction data available in New Drug Applications (NDAs) approved in 2013, and highlight significant findings. Methods: The University of Washington Drug Interaction Databaseß was used to extract in vitro metabolism and transport data, as well as results of clinical drug interaction studies. Results: Among 27 New Molecular Entities (NMEs) approved in 2013, 22 (81%) had drug metabolism or transport data available and were fully analyzed (the other compounds were radioactive diagnostic or therapeutic agents, Gadolinium-based contrast agent, and cytolytic antibody). Eight drugs evaluated (30%) were cancer therapies, including 4 kinase inhibitors, making oncology the most represented therapeutic area. Seventeen of the 22 NMEs (77%) were metabolized by CYP3A in vitro, and 10 of them were confirmed to be in vivo CYP3A substrates. As perpetrators, 17 NMEs inhibited at least one enzyme in vitro, with only 2 being established in vivo CYP inhibitors (simeprevir moderately inhibited CYP3A and alogliptin weakly inhibited CYP2D6). Of the 21 NMEs evaluated for in vitro enzyme induction, five (23%) were found to be CYP inducers in vitro. In vivo, dabrafinib and eslicarbazebine acetate induced CYP3A moderately and weakly, respectively. Twenty-one of the 22 NMEs had in vitro transport data available, either as substrates or as perpetrators, with P-glycoprotein (P-gp) being evaluated in all cases. Twelve NMEs were substrates of P-gp while 8 NMEs inhibited P-gp in vitro. Four NMEs were tested in vivo with the P-gp probe substrate digoxin and only simeprevir significantly increased digoxin exposure (AUC ratio = 1.4). For other transporters recommended in the 2012 draft FDA DDI guidance, BCRP, OATP1B1, and OATP1B3 were the most studied (n = 9, 9, and 7, respectively), while information on OAT1, OAT3, and OCT2 was found in 5, 4, and 4 NMEs, respectively. Amongst them, 5 NMEs showed in vitro inhibition of OATP1B1/ 1B3, while in vivo, only simeprevir moderately increased the AUC of the OATP substrate rosuvastatin by 180%. Four NMEs inhibited BCRP in vitro, but these results were not expected to have any clinical relevance. Overall, 10 of the 22 drugs (45%) had at least one in vivo DDI study with a change in exposure of clinical significance (AUC increase 2-fold or AUC decrease  50% for the affected drugs), with NMEs being mainly victim drugs. Two compounds, ibrutinib and simeprevir, were found to be sensitive substrates of CYP3A with an AUC ratio 45-fold when co-administered with potent CYP3A inhibitors. Conclusion: New drugs approved in 2013 were thoroughly evaluated for transporter-based DDIs, reflecting recent

120

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

regulatory recommendations. Clinically significant DDIs (2-fold change in exposure) were observed for 45% of the compounds. The largest interaction was observed with the kinase inhibitor ibrutinib, a substrate of CYP3A, when co-administered with the CYP3A inhibitor ketoconazole (over 20-fold increase in ibrutinib AUC).

P174. EXPLORING THE DRUG–DRUG INTERACTION BETWEEN GEMFIBROZIL AND REPAGLINIDE IN RATS: METABOLISM AND TRANSPORT

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Joanna E. Barbara, Seema Muranjan, Forrest Stanley, Chandra Kollu, Sylvie Kandel, Clayton Otwell, and David B. Buckley XenoTech, LLC, Lenexa, KS, USA A clinically-relevant drug–drug interaction (DDI) between the dyslipidemia drug gemfibrozil and the antidiabetic repaglinide is well-documented throughout the literature. In humans, repaglinide is predominantly cleared by hepatic metabolism involving cytochrome P450 (P450) 3A4 and 2C8 and UGT1A1 and 1A3. Gemfibrozil and its glucuronide metabolite inhibit CYP2C8 (irreversibly) and UGT1A1, which has been proposed as a key cause of the clinical interaction. Additionally, gemfibrozil and gemfibrozil glucuronide are established inhibitors of the hepatic uptake transporter OATP1B1, proposed as a confounding factor. A mechanistic assessment of the gemfibrozil/repaglinide DDI was undertaken in male Sprague-Dawley rats. Subjects (n = 3/group) received 200 mg/kg/day gemfibrozil or vehicle control by oral gavage for 2 days. On day 3, rats were fasted for 12 h and dosed once orally with 1 mg/kg repaglinide coadministered with gemfibrozil or vehicle. Blood, bile and urine were collected for 12 h following repaglinide administration. Livers were extracted for microsome preparation. Inlife work was performed by Xenometrics LLC (Stilwell, KS). Rat plasma time-points, Hamilton-pooled plasma, and 0–12 h urine and bile pools were analyzed by LC-MS/MS for PK and metabolite profiling. P450 and UGT enzyme activities in liver microsomes were analyzed by LC-MS/MS. Gemfibrozil treatment greatly reduced repaglinide clearance in rats. Gemfibroziltreated rats exhibited a 4-fold higher repaglinide Cmax and a 3.5-fold greater AUC0–12 than control rats, but the tmax (1.2 and 1.7 h) and t1/2 (2.6 h) values were similar. The Vd,obs and plasma CLobs were approximately two-thirds lower in gemfibroziltreated rats than control rats. Repaglinide, repaglinide glucuronide and three oxidative metabolites were observed in control rat plasma. In gemfibrozil-treated rat plasma, which contained seven additional metabolites, repaglinide glucuronide was present at 3-fold higher abundance and the oxidative metabolites at 414-fold abundance compared to control plasma. In control rats, the vast majority of the repaglinide-related material was excreted in bile. Treatment with gemfibrozil resulted in a vectoral shift in elimination pathways leading to a marked increase in urinary excretion of repaglinide and its metabolites. Most repaglinide metabolites were observed in bile, where repaglinide was also detected. Control and gemfibrozil-treated rat bile contained 42 and 38 metabolites, respectively. In gemfibrozil-treated rat bile, repaglinide and its metabolites were present at lower abundance (30. 80%) than control rat bile. Repaglinide and 6 metabolites were detected in control rat urine but were present at up to 900-fold higher abundance in gemfibrozil-treated rat urine. Analysis of microsomal P450 and UGT activity towards a panel of marker substrates revealed little difference between the gemfibrozil-treated and control rats, with the exception of a 2.5-fold increase in CYP4A1 activity in gemfibrozil-treated rats possibly caused by PPARa activation. Clearance experiments with the gemfibrozil-treated, control, and naı¨ve rat liver microsomes did not show a marked decrease in repaglinide clearance following gemfibrozil treatment, even when a pre-incubation period with gemfibrozil or gemfibrozil glucuronide was incorporated. The increased repaglinide exposure associated with gemfibrozil treatment cannot be correlated with P450 or UGT inhibition in rats. The PK data and altered excretion profile are consistent with a hepatic transporter-based DDI.

P175. IN VITRO HERB-DRUG INTERACTION POTENTIAL OF HS-23, LONICERA FLOWER EXTRACT FOR THE TREATMENT OF SEPSIS Tae Yeon Kong1, Hyeon-Uk Jeong1, Yeon Kim2, Sung-Woon Hong2, Sung Hum Yeon2, Yong Yeon Cho1, and Hye Suk Lee1 1 College of Pharmacy, The Catholic University of Korea, Bucheon, South Korea and 2Huons Co., Ltd., Ansan, South Korea HS-23, an extract of the dried flower buds of Lonicera japonica, is a new botanical drug currently being evaluated in a phase I clinical study in Korea for the treatment of sepsis. Herb-drug interactions resulting from concurrent use of botanical drugs with prescription and over-the-counter drugs may cause the adverse reactions such as toxicity and treatment failure. The underlying mechanisms of herb–drug interactions involve inhibition or induction of cytochrome P450 (CYP) enzymes, UDPglucuronosyltransferase (UGT) enzymes, and drug transporters. The in vitro induction and inhibition potentials of HS-23 on the drug-metabolizing enzymes using human hepatocytes and liver microsomes were assessed to evaluate metabolism-based herb–drug interaction. We studied transporter-mediated herb–drug interaction potential of HS-23 in HEK293 cells overexpressing organic cation transporter (OCT)1, OCT2, organic anion transporter (OAT)1, OAT3, organic anion transporting polypeptide (OATP)1B1, OATP1B3 and LLC-PK1 cells overexpressing P-glycoprotein (P-gp), and breast cancer resistance protein (BCRP). HS-23 slightly inhibited CYP2A6, CYP2B6, CYP2C9, CYP2C19, and CYP3A4 enzyme activities in human liver microsomes with IC50 values of 80.6, 160.7, 169.5, 85.4, and 76.6 mg/mL, respectively. HS-23 showed negligible inhibition of CYP1A2, CYP2C8, CYP2D6, UGT1A1, UGT1A4, UGT1A9, and UGT2B7 activities in human liver microsomes. Based on these results, HS-23 may not cause the inhibition of the metabolism of CYP2A6, CYP2B6, CYP2C9, CYP2C19, and CYP3A4-

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

121

catalyzed drugs in humans. HS-23 did not affect the mRNA expression of CYP1A2, CYP2B6, CYP3A4, CYP2C8, CYP2C9, UGT1A1, and UGT1A4 after 48 h treatment at three concentrations (0.5, 5, and 50 mg/mL) in three independent human hepatocytes, indicating that HS-23 is not a perpetrator of herb–drug interactions due to the up- or down-regulation of CYP1A2, CYP2B6, CYP3A4, CYP2C8, CYP2C9, UGT1A1 and UGT1A4 enzymes. HS-23 (up to 100 mg/mL) did not significantly inhibit transport activities of OCT1, OCT2, OAT1, OAT3, OATP1B1, OATP1B3, P-gp and BCRP. In conclusion, these results indicate that HS-23 may not cause clinically relevant inhibition and induction of these CYP and UGT enzymes, and transporter-mediated drug interaction, although their contribution to in vivo pharmacokinetics need to be further investigated. This study was supported by a grant (A121826) from health technology development program through the Ministry of health and welfare of Korea.

P176. APPLICATION OF A PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODEL FOR PREDICTION OF THE DRUG–DRUG INTERACTION (DDI) BETWEEN LORAZEPAM AND PROBENECID

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sibylle Neuhoff, Helen Musther, and Karen Rowland-Yeo Simcyp Ltd (a Certara Company), Sheffield, UK Purpose: Application of PBPK models to assess the UGT2B7-mediated DDI between lorazepam and probenecid in healthy volunteers. Methods: Prior in vitro and in vivo information on the metabolism of lorazepam was combined with physicochemical data within the Simcyp Population-based Simulator (V13 release 2) to generate concentration-time profiles of lorazepam following single (SD) and multiple, intravenous (iv) or oral doses over a range of doses (1 mg to 4.5 mg (0.057 mg/kg)). Similarly, a full PBPK model was also developed for the UGT2B7 inhibitor probenecid. Where possible, simulations were compared with corresponding observed data for each compound. In addition, in vitro Ki data relating to inhibition of UGT2B7 glucuronide formation by probenecid were used to investigate the effects of this compound on the systemic exposure of lorazepam. Results: The simulated concentration-time profiles of lorazepam were consistent with observed data across 4 independent studies and for probenecid for 2 independent studies at the inhibitor dose of 500 mg. The predicted increase in exposure of lorazepam (2 mg SD iv) following administration of probenecid was similar to observed. The predicted and observed ratios of the area under the plasma concentration-time profile were 1.37 (range: 1.14–1.79) and 1.825 (range: 0.97–2.82), respectively. Conclusions: PBPK modelling in conjunction with reliable inhibition data can be used to assess the importance of interactions affecting the glucuronidation pathways. Indeed, PBPK models can be used to evaluate other UGT2B7-mediated DDIs using the selective victim drug, lorazepam, or the inhibitor, probenecid.

P177. CO-ADMINISTRATION OF INTRAVENOUS PROTON PUMP INHIBITORS IS A RISK FACTOR FOR DELAYED ELIMINATION OF PLASMA METHOTREXATE IN HIGH-DOSE METHOTREXATE THERAPY Masato Homma1, Kunihiro Suzuki1, Kosuke Doki1, and Shigeru Chiba2 1 Department of Pharmaceutical Sciences, University of Tsukuba, Tsukuba, Japan, and 2Department of Hematology, University of Tsukuba, Tsukuba, Japan Purpose: Our previous study described that co-administration of proton pump inhibitors (PPIs) potentially delayed elimination of plasma methotrexate (MTX) in limited number of patients treated with high-dose MTX (HDMTX) therapy for malignant diseases (Suzuki et al., 2009). To confirm clinical impact of this drug interaction, we further examined effects of administration route of PPIs on delayed elimination of MTX in large number of patient population and population pharmacokinetic (PPK) analysis. Methods: Plasma MTX concentration data on 383 cycles of HDMTX therapy performed in 164 patients who received intravenous or oral PPIs (omeprazole, lansoprazole and rabeprazole) concomitantly were examined. Patients with plasma MTX concentration 10 mmol/L, 1 mmol/L or 0.1 mmol/L at 24 h, 48 h or 72 h after the administration of HDMTX therapy, respectively, were classified into the delayed elimination group (95 cycles). Multiple logistic regression analysis was performed to evaluate the PPIs co-administration as a risk factor for delayed elimination of MTX. Population pharmacokinetic analysis was conducted to assess the impact of intravenous or oral PPIs on the systemic clearance (CL) of plasma MTX. Results: Multiple logistic regression analysis identified co-administration of PPIs as a risk factor for delayed elimination of plasma MTX (odds ratio 1.73, 95% confidence interval 1.06–2.85) as well as presence of third space effusions, pleural effusions and ascites. Population pharmacokinetic analysis for plasma MTX revealed that creatinine clearance and intravenous omeprazole were significant factors for estimating MTX CL. Conclusions: Present results supported a previous finding in which PPIs co-administration was associated with delayed elimination of plasma MTX in patients with HDMTX. Among the co-administration of several PPIs with intravenous or oral administration, it was suggested that intravenous omeprazole co-administration was the most important for delayed elimination of plasma MTX. The mechanism underlying this drug interaction remained to be clarified in future work.

Reference Suzuki K, Doki K, Homma M, et al. (2009). Co-administration of proton pump inhibitors delays elimination of plasma methotrexate in high-dose methotrexate therapy. Br J Clin Pharmacol 67:44–49.

122

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P178. DRUG INTERACTION STUDIES DURING DRUG DEVELOPMENT AND NEW DRUG APPLICATIONS (1): CURRENT STATUS AND PMDA PERSPECTIVES ON DRUG INTERACTIONS INVOLVING METABOLIZING ENZYMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Akihiro Ishiguro1, Mai Hirano2, Naoko Honma3, Motohiro Hoshino4, Daisuke Iwata1, Shinichi Kijima3, Shinichi Okudaira3, Masanobu Sato5, Reiko Sato2, Kumiko Takeuchi3, Shingo Watanabe3, and Naomi Nagai1 1 Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of New Drug, Tokyo, Japan, 2Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Safety, Tokyo, Japan, 3Pharmaceuticals and Medical Devices Agency (PMDA), Office of New Drug, Tokyo, Japan, 4 Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Conformity Audit, Tokyo, Japan, and 5Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Std and Guidelines Dev, Tokyo, Japan Multiple-drug therapies are often prescribed in current medical practice, making it more difficult for healthcare professionals to adequately manage drug–drug and drug-food interactions. In Japan, intensive works toward the revision of the regulatory document, ‘‘Methods of Drug Interaction Studies’’ issued in 2001 by the Ministry of Health, Labour and Welfare (MHLW) (Sato et al., 2015) have been conducted by the MHLW Drug Interaction Working Group and three Sub-Working Groups (Transporters, Drug Metabolizing Enzymes, and Modeling & Labeling) since Dec. 2012. While the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) published the new guideline in 2013 and the draft guidance in 2012, respectively (The EMA, 2013; The MHLW, 2001), the new MHLW guideline drafted based on the recent progress of clinical pharmacology as well as consideration of global harmonization, was presented in the 10th ISSX Meeting in October 2013 (The US FDA, 2012) and notified in December 2013 for public comments. Based on the public comments, for example, (1) approaches for transportermediated drug interaction studies and induction studies, (2) application of modeling and simulation to develop both in vitro to in vivo studies and in vivo studies to labeling implications, and (3) detailed clinical drug interaction study design had been discussed, and the final draft of new Japanese guideline, ‘‘Drug interaction guideline for drug development and labeling recommendations’’ was prepared. In order to finalize the new guideline and respond to the public comments, it is useful to summarize recent cases regarding drug interaction studies during new drug development. In this presentation, we show the latest information on the drug interaction data related to drug metabolizing enzymes in Japanese NDAs and discussion points during PMDA review process. Based on our study, it is concluded that the new guideline will contribute to planning, conducting and evaluating drug interaction studies appropriately during new drug development. We expect continuous discussion and further international harmonization of methods of drug interaction studies among US and EU regulatory agencies.

References Sato M, et al. (2015). Drug interaction studies during drug development: Current status and regulatory perspectives in Japan. The 10th International ISSX Meeting, P166. The EMA. (2013). Guideline on the Investigation of Drug Interactions. London, UK: European Medicines Agency. The MHLW. (2001). The method of drug interaction studies. Notification No.813, Jun. 2001. Japan: Ministry of Health, Labor and Welfare. The US FDA. (2012). Drug interaction studies – Study design, data analysis, implications for dosing, and labeling recommendations (draft guidance). Maryland, USA: US Food and Drug Association.

P179. DRUG INTERACTION STUDIES DURING DRUG DEVELOPMENT AND NEW DRUG APPLICATIONS (2): CURRENT STATUS AND PMDA PERSPECTIVES ON DRUG INTERACTIONS INVOLVING TRANSPORTERS Daisuke Iwata1, Mai Hirano2, Naoko Honma3, Motohiro Hoshino4, Akihiro Ishiguro1, Shinichi Kijima3, Shinichi Okudaira3, Masanobu Sato5, Reiko Sato2, Kumiko Takeuchi3, Shingo Watanabe3, and Naomi Nagai1 1 Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of New Drug, Tokyo, Japan, 2Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Safety, Tokyo, Japan, 3Pharmaceuticals and Medical Devices Agency (PMDA), Office of New Drug, Tokyo, Japan, 4 Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Conformity Audit, Tokyo, Japan, and 5Pharmaceuticals and Medical Devices Agency (PMDA), The PMDA Drug Interaction Working Group and Office of Std and Guidelines Dev, Tokyo, Japan Multiple-drug therapies are often prescribed in current medical practice, making it more difficult for healthcare professionals to adequately manage drug–drug and drug-food interactions. In Japan, intensive works toward the revision of the regulatory document, ‘‘Methods of Drug Interaction Studies’’ issued in 2001 by the Ministry of Health, Labour and Welfare (MHLW) (Sato et al., 2015) have been conducted by the MHLW Drug Interaction Working Group and three Sub-Working Groups (Transporters, Drug Metabolizing Enzymes, and Modeling & Labeling) since December 2012. While the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) published the new guideline in 2013 and the draft guidance in 2012, respectively (The EMA, 2013; The MHLW, 2001), the new MHLW guideline drafted based on the recent progress of clinical pharmacology as well as consideration of global harmonization, was presented in the 10th ISSX Meeting in in October 2013 (The US FDA, 2012) and notified in December 2013 for public comments. Based on the public comments, for example, (1)

Poster Abstracts

DOI: 10.3109/03602532.2015.1071933

123

approaches for transporter-mediated drug interaction studies and induction studies, (2) application of modeling and simulation to develop both in vitro to in vivo studies and in vivo studies to labeling implications, and (3) detailed clinical drug interaction study design had been discussed, and the final draft of new Japanese guideline, ‘‘Drug interaction guideline for drug development and labeling recommendations’’ was prepared. In order to finalize the new guideline and respond to the public comments, it is useful to summarize recent cases regarding drug interaction studies during new drug development. In this presentation, we show the latest information on the drug interaction data related to transporters in Japanese NDAs and discussion points during PMDA review process. Based on our study, it is concluded that the new guideline will contribute to planning, conducting and evaluating drug interaction studies appropriately during new drug development. We expect continuous discussion and further international harmonization of methods of drug interaction studies among US and EU regulatory agencies.

References

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sato M, et al. (2015). Drug interaction studies during drug development: Current status and regulatory perspectives in Japan. The 10th International ISSX Meeting, P166. The EMA. (2013). Guideline on the Investigation of Drug Interactions. London, UK: European Medicines Agency. The MHLW. (2001). The method of drug interaction studies. Notification No.813, Jun. 2001. Japan: Ministry of Health, Labor and Welfare. The US FDA. (2012). Drug interaction studies – Study design, data analysis, implications for dosing, and labeling recommendations (draft guidance). Maryland, USA: US Food and Drug Association.

P180. EFFECTS OF TRICYCLIC COMPOUNDS ON ALMOTRIPTAN TRANSPORT THROUGH OATP1A2 Jennifer Lu1, Alexia Grangeon1, Fleur Gaudette2, Markus Keiser3, Ve´ronique Michaud2, and Jacques Turgeon2 1 Faculty of Pharmacy, Montreal University, Montreal, QC, Canada, 2CRCHUM, University of Montreal, Montreal, QC, Canada, and 3Department of Pharmacology and Clinical Pharmacology, Center of Drug Absorption and Transport (C_DAT), University Medicine of Greifswald, Greifswald, Germany Background: OATP1A2 is a membrane drug-transporter highly expressed at the human blood-brain barrier that may potentially mediate penetration of drugs across the brain. Recently, Cheng et al. demonstrated that hydrophilic antimigraine triptans are substrates for OATP1A2. Triptans are selective agonists of the serotonin receptors 5-HT1B and 5-HT1D located on the smooth muscle cells of intracranial blood vessels and neurons, respectively. Thus, triptans need to cross the blood-brain barrier to reach their site of action. Using an in vitro model overexpressing OATP1A2, our results confirmed that several triptan drugs were substrates for OATP1A2 transport. We have previously demonstrated that OATP1A2-mediated transport of rosuvastatin was inhibited by compounds composed of a tricyclic ring with a short aliphatic amine chain. The goal of this study was to determine whether almotriptan transport via OATP1A2 can be affected by tricyclic drugs. Methods: A HEK293 cell line overexpressing OATP1A2 was used as in vitro model. OATP1A2 cells and control cells (7.5  105 cells/well in 12well plates) were incubated at 37 C for 2 minutes with almotriptan (15 mM) and different concentrations of inhibitors (0.15, 0.5, 1.5, 5.0, 15.0, 50.0, 150.0 mM). The tricyclic compounds evaluated were amitriptyline, carazolol, carbamazepine, carbazole, carvedilol, chlorpromazine, clomipramine, desipramine, doxepin, imipramine, nortriptyline, phenothiazine, and trimipramine. The quantity of intracellular almotriptan was quantified by HPLC-UV. Results: The transport of almotriptan through OATP1A2 was inhibited by most of the tricyclic compounds evaluated. The IC50values are shown in the table. Carvedilol had the greatest inhibition potency on almotriptan transport through OATP1A2. Conclusion: Our data suggest that compounds composed of a tricyclic ring with a short aliphatic amine chain such as carvedilol and tricyclic antidepressants may limit the penetration of almotriptan in the brain by modulating OATP1A2 transport. Consequently, the impact of concomitant administration of triptans with potent inhibitors of OATP1A2 on antimigraine efficiency of almotriptan needs to be confirmed. Compound

IC50 (mM)

Amitriptyline Carazolol Carbamazepine Carbazole Carvedilol Chlorpromazine Clomipramine Desipramine Doxepin Imipramine Nortriptyline Phenothiazine Trimipramine

4.6 1.6 No effect No effect 0.7 8.7 6.1 16.2 2.5 4.3 4.5 No effect 7.6

124

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P181. Quantitative analysis for induction of CYP3A in a new double humanized mice of PXR/SXR (knock-in) and CYP3A (transchromosomic)

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Jiro Kuze1, Kaoru Kobayashi2, Shoko Takehara3, Daisuke Satoh3, Toshiharu Mutoh4, Masato Chiba4, Kanako Takano2, Katsuhide Igarashi5, Satoshi Kitajima6, Jun Kanno6, Kan Chiba2, Mitsuo Oshimura3, and Yasuhiro Kazuki3 1 Pharmacokinetics Research Lab, Taiho Pharmaceutical Co., Ltd, Tsukuba, Japan, 2Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan, 3Chromosome Engineering Research Center, Tottori University, Yonago, Japan, 4Discovery Drug Metabolism & Pharmacokinetics, Taiho Pharmaceutical Co., Ltd, Tsukuba, Japan, 5Life Science Tokyo Advanced Research center, Hoshi University School of Pharmacy and Pharmaceutical Science, Tokyo, Japan, and 6Division of Cellular and Molecular Toxicology, Biological Safety Research Center, National Institute of Health Sciences, Tokyo, Japan Pregnane X receptor (PXR), also known as steroid and xenobiotic receptor (SXR), is a major activator of CYP3A transcription. Induction of CYP3A in response to PXR/SXR ligands shows species-specific differences. Thus, we generated a double humanized mouse model of PXR/SXR and CYP3A to study the species-specific induction of human CYP3A. The CYP3Ahumanized mice generated by using mouse artificial chromosome (MAC) vector containing the entire genomic human CYP3A locus (CYP3A-MAC mouse line) (Kazuki et al., 2003) were bred with the PXR/SXR-humanized mice in which the ligand binding domain of human PXR/SXR was homologously knocked-in to the murine gene replacing the endogenous ligand binding domain (hSXRki mouse line) (Igarashi et al., 2012) to generate the double humanized mice (CYP3A-MAC/hSXRki mouse line). Previously, we presented that treatment with rifampicin, a typical human PXR/SXR ligand, increased the expression of hepatic CYP3A4 mRNA in CYP3A-MAC/hSXRki mice but not in CYP3A-MAC mice. In this study, we quantitatively assessed an in vivo and in vitro induction response to rifampicin by using CYP3A-MAC/hSXRki mice. CYP3A-MAC/hSXRki or CYP3AMAC mice were treated with rifampicin (10 mg/kg/day, 3 days, p.o.) or vehicle. At 1 day after final treatments, mice were intravenously or orally administered with triazolam (1 mg/kg). Plasma concentrations of triazolam and its metabolites (10 - and 4-hydroxyltriazolam) were quantitated by LC/MS/MS analysis. In addition, we assessed an in vitro induction response to rifampicin by using liver and small intestine microsomes of CYP3A-MAC/hSXRki and CYP3A-MAC mice. Formation rates of 10 - and 4-hydroxyltriazolam in liver and small intestine microsomes were determined by HPLC analysis. In CYP3A-MAC/ hSXRki mice, the plasma concentration of triazolam was significantly altered by rifampicin treatment. Cmax and AUC of triazolam after oral dosing were decreased by 51% and 32%, respectively. AUC after intravenous dosing was also decreased by 67%. 10 -Hydroxytriazolam was a major metabolite of triazolam in plasma of CYP3A-MAC/hSXRki mice. AUC ratios of 10 -hydroxytriazolam to triazolam were also increased by rifampicin treatment. In contrast, treatment with rifampicin to CYP3AMAC mice had an insignificant effect on plasma concentrations of triazolam and its metabolites. In addition, rifampicin treatment increased formation rates of 10 - and 4-hydroxyltriazolam in liver and intestine microsomes of CYP3A-MAC/ hSXRki mice, but not in those of CYP3A-MAC mice. These data suggest that the CYP3A-MAC/hSXRki mouse line may be a useful model to predict human CYP3A induction mediated via human PXR/SXR in vivo and in vitro.

References Igarashi K, Kitajima S, Aisaki K, et al. (2012). Development of humanized steroid and xenobiotic receptor mouse by homologous knock-in of the human steroid and xenobiotic receptor ligand binding domain sequence. J Toxicol Sci 37:373–380. Kazuki Y, Hoshiya H, Kai Y, et al. (2003). Trans-chromosomic mice containing a human CYP3A cluster for prediction of xenobiotic metabolism in humans. Hum Mol Genet 22:578–592.

P182. ESTABLISHMENT OF AN ASSAY METHOD FOR CYP3A INDUCTION POTENCY WITH LS180 CELL LINE EXPRESSING HUMAN PREGNANE X RECEPTOR Miwa Watanabe, Maki Hasegawa, Shinobu Shioya, and Harunobu Tahara R&D Division, Transrational Reserch Unit, KYOWA-KIRIN, Shizuoka, Japan Intestinal human colon adenocarcinoma LS180 cell line expresses endogenous pregnane X receptor (PXR) and has CYP3A activity (Anshul Gupta et al., DMD 36:1172–1180, 2008). LS180 cell line, however, has low ability of CYP3A induction. We hypothesized that LS180 cell line highly expressing PXR could have higher CYP3A induction potency, and could use as a screening tool to evaluate CYP3A induction potency in preclinical stage. We obtained LS180 cell line stably expressing human PXR (hPXR-LS180). CYP3A induction potency to rifampicin was compared with hPXR-LS180, LS180 cell line and cryopreserved human hepatocytes. Selected 14 known CYP3A inducers were examined their effects on CYP3A activity in hPXR-LS180. Their induction potency were normalized by that of rifampicin (as the reference; 3 micromol/L). R20 value was determined as the concentration of 20% of induction compared with the reference. From the reported value of clinical area under the plasma concentration-time curve (AUC) and the plasma unbound fraction (fu), the unbound plasma exposures (fu  AUC) to inducers were calculated. The fu  AUC/R20 value was determined as the in vivo induction index. The relationship between in vitro and in vivo induction potency of CYP3A was examined by using the in vivoinduction index. CYP3A induction potency to rifampicin (3 micromol/L) in hPXR-LS180 cell line was 5-times higher than that in LS180 cell line and was comparable to that in cryopreserved human hepatocytes. When the fu  AUC/R20 obtained in hPXR-LS180

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

125

versus the percentage decrease in AUC of coadministered CYP3A substrate was plotted, the positive correlation was observed. From the result of in vitro–in vivo correlation, it is suggested that AUC decrease in coadministered CYP3A substrate is predictable from the fu  AUC / R20 value. These results were comparable with the data obtained in HepaRG cells. In conclusion, these results indicated that hPXR-LS180 cell line can use as simple screening system to evaluate CYP3A induction potency, and the method reported in the present study will be applicable to predict clinically DDIs caused by CYP3A induction in a semiquantitative manner.

P183. PREDICTION OF CYP2B6 AND 3A4 INDUCTION POTENTIAL WITH RIS IN PRIMARY HUMAN HEPATOCYTES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Jun Sun1, Bruno Carvalho1, Mohamad Shebley2, Yau Lau1, Sonie de Morais1, and Ganesh Rajaraman1 1 Department of Drug Metabolism, Abbvie Inc, North Chicago, IL, USA and 2Translational PKPD, Abbvie Inc, North Chicago, IL, USA Drug–drug interactions (DDI) may be caused by variety of reasons including CYP inhibition, CYP induction, and drug transporter interactions. CYP induction results in increases of drug-metabolizing enzyme activities potentially leading to decreased drug efficacy and/or increased drug toxicity (e.g. Toxic metabolites, Pro-drugs). CYP3A4 is a major drug metabolizing enzyme in humans and its induction leading to DDI has been extensively studied in the clinic. Literature has shown that there is a good in vitro in vivo extrapolation (IVIVE) using relative induction score (RIS) approach to assess the potential DDI risk mediated by CYP3A4 induction. In comparison, CYP2B6 induction potential and its impact in clinic are much less known. In most of the case, if a compound induces CYP3A4, it also induces CYP2B6 due to the cross link between PXR and CAR. In our study, we used primary human hepatocytes, the gold standard to study CYP2B6 and 3A4 induction. We determined the maximal fold induction (Emax) and the concentration causing half maximal induction (EC50) by Rifampicin, Carbamazepine, Phenytoin, Efavirenz, Pleconaril, and Rosiglitazone at mRNA level (determined by quantitative polymer chain reaction, qtPCR, TaqManÔ) for CYP3A4 induction and Carbarmazepine, Rifampicin, Efavirenz, Metamizole, Nelfinavir and Ferulic Acid for CYP2B6 induction. The RIS values determined by our in vitro studies showed good correlation to in vivo observed percent AUC decreases of Midazolam (CYP3A4) and Bupropion (CYP2B6). The results indicate that CYP2B6 and 3A4 induction potential can be assessed with RIS in primary human hepatocytes. The 3 known CYP2B6 and 3A4 in vivo inducers (Efavirenz, Carbamazepine, Rifampicin) induced both CYP2B6 and 3A4 in vitro and the CYP2B6 in vivo induction risk was able to be predicted using CYP3A4 RIS static modeling. CYP3A4 induction RIS could be used as a surrogate of CYP2B6 induction potential for these compounds. Further studies are needed to confirm the IVIVE (RIS) for CYP2B6 induction due to the limited availability of clinical DDI studies for CYP2B6.

P184. BRINCIDOFOVIR AND IN SITU-GENERATED METABOLITES DO NOT INDUCE CYTOCHROME P450 ENZYMES, AS DETERMINED CONCURRENTLY FOR BOTH BRINCIDOFOVIR AND METABOLITES Kathy Van Sickle1, David Buckley2, and Tim Tippin1 1 Chimerix, Inc., Durham, NC, USA and 2XenoTech, LLC, Lenexa, KS, USA Background: Brincidofovir is an orally administered lipid conjugate nucleotide in Phase 3 clinical development for the prevention of cytomegalovirus (CMV) infection in hematopoietic stem cell transplant (HCT) recipients and for the treatment of adenovirus (AdV) infection. In humans, brincidofovir is metabolized primarily in the liver to three metabolites, CMX103, CMX064, and cidofovir (CDV) [Figure 1]. The purpose of this study was to assess the ability of brincidofovir and in situgenerated metabolites to induce cytochrome P450 (CYP) enzymes CYP1A2, CYP2B6 and CYP3A4/5. Methods: Cryopreserved cultured human hepatocytes from three individual donors were treated once daily for three consecutive days with six concentrations of brincidofovir ranging from 0.1 to 10 mM or control (vehicle, negative control, or known inducers). Following the third day of treatment, aliquots of media were collected at 0, 1, 3, 6, and 24 hours post treatment. The hepatocytes were subsequently incubated with fresh media containing substrates of CYP1A2, CYP2B6, and CYP3A4/5 to measure changes in CYP activity. Cells were then harvested and analyzed by quantitative reverse transcription-polymerase chain reaction (qRTPCR) to assess changes in CYP mRNA levels. Media concentrations of in situ generated CMX064, CMX103 and CDV were measured using LC-MS/MS. To assess the exposure of hepatocytes to brincidofovir metabolites, an area under the media concentration versus time curve (AUC0–24) was determined for each metabolite using non-compartmental analysis. Results: Compared to the prototypical inducers omeprazole, phenobarbital, and rifampin which caused anticipated increases in CYP1A2, CYP2B6, and CYP3A4/5 activity and mRNA levels, brincidofovir did not cause increases in CYP activity (2 fold change over control) or mRNA levels (510% of the increase caused by positive controls). Following incubation with 10 mM, exposure (AUC0–24 and/or Cmax) to each metabolite on Day 3 was greater than or equal to systemic concentrations measured in humans following a single 200 mg oral dose of brincidofovir. Conclusions: No induction of CYP1A2, CYP2B6 or CYP3A4/5 by brincidofovir was observed in hepatocytes at concentrations approximately 10-fold higher than peak plasma concentrations observed in humans after brincidofovir administration. The exposure to CDV, CMX064 and CMX103 measured in media were equal to or exceeded those in humans; therefore, a low CYP induction potential of brincidofovir metabolites can be inferred. Thus, brincidofovir and its major metabolites are predicted to have a low risk for perpetrating a drug–drug interaction via

126

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

induction of CYP enzymes. This study demonstrates the utility of measuring in situ-generated metabolites to assess CYP enzyme induction potential due to parent drug and/or metabolites.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Figure 1. Brincidofovir and metabolites.

P185. EVALUATION OF VARIOUS APROACHES TO PREDICT CLINICAL CYP1A2 INDUCTION FROM IN VITRO HUMAN HEPATOCYTE MRNA DATA Lai Wang and Heidi J. Einolf Drug Metabolism and Pharmacokinetics, Novartis Institute for Biomedical Research, East Hanover, NJ, USA Prediction of clinical drug–drug interaction (DDI) due to cytochrome P450 (CYP)1A2 induction can be an important aspect of drug development, particularly for informing the need for a designated clinical DDI trial. It was the objective of this study to evaluate several approaches to predict the clinical drug interaction of known in vivo inducers/non-inducers of CYP1A2 using in vitro mRNA induction data. Clinical drug-interaction trials with the sensitive CYP1A2 substrate, theophylline, were found from literature sources (8 trials were used in this analysis). The compounds identified in the literature included albuterol, omeprazole, rifampin, moricizine, phenobarbital, ritonavir, and phenytoin. The in vitro induction parameters (EC50 and Emax values) of the clinical inducers/non-inducers were determined using cryopreserved human hepatocytes and evaluation of CYP1A2 mRNA induction by real-time PCR. Various approaches were evaluated to correlate thein vitro induction parameters of CYP1A2 with the reported changes in clinical theophylline exposure. Those methods included the correlation (or calibration curve) approaches: Cmax /EC50, AUC/F2 (where F2 is concentration of inducer that elicits a 2-fold in vitro induction effect), and the relative induction score (RIS), as well as calculation of the R3 value defined by the FDA. These approaches were then used to predict the magnitude of clinical DDI for a test in vitro inducer of CYP1A2, but unknown clinical inducer, in addition to the aforementioned known clinical inducers/non-inducers. A good correlation of the in vitro CYP1A2 induction data with the in vivo DDI data were found with the Cmax /EC50 approach, particularly using Cmax,total (R2 value of 0.89). Use of Cmax,unbound or hepatic inlet Cmax,unbound also yielded a decent correlation (R2 value of 40.7). The other correlation approaches did not yield as good as a fit of the data (R2 values of 0.4). A similar trend was found with the correlation of CYP1A1 mRNA induction data with the theophylline clinical DDI results. The R3 approach predicted the DDI within 2-fold of the actual value for the majority of the trials of this limited dataset, particularly using Cmax,unbound. Examination of the unknown clinical inducer by the Cmax /EC50 approach (using Cmax,total) predicted a 10% decrease in theophylline AUC. This was in-line with the R3 method where the value was estimated to be 0.9 using Cmax,total. Using other surrogates for the inducer concentration such as hepatic inlet Cmax,unbound resulted in predictions of up to 30% decrease in AUC. Based upon the analysis of this dataset, the use of in vitro induction data to predict the clinical outcome of CYP1A2 induction has shown great promise. Increased confidence in the modeling of CYP1A2 induction will come with continued evaluation of more clinical trials.

P186. THE IMPACT OF INCOMPLETE DOSE-RESPONSE CURVES ON EC50 AND Emax DETERMINATIONS IN ENZYME INDUCTION ASSESSMENT Immaculate Amunom, Elizabeth Lafreniere, Rebecca R. Campbell, Paul Bolliger, Bradley A. Klaus, Janet Sawi, Sara McKinney, Racquel Mueller, Catherine M. Wiegand, and David B. Buckley XenoTech, LLC, Lenexa, KS, USA In evaluating the enzyme induction potential of a new drug candidate, it is common practice to calculate EC50 and Emax values to aid the prediction of drug–drug interactions. Typically, EC50 and Emax values are calculated when an increase of two-fold or higher is observed and these values are used to assess clinical induction potential. In literature, there are two common approaches, but not limited to, the sigmoid 3-parameter and Hill 3-parameter equations. In general, the sigmoid 3-parameter

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

127

equation forces the calculated Emax towards the maximal experimentally observed value regardless of whether a ‘‘true’’ Emax has been reached. Conversely, the Hill 3-parameter equation extrapolates the data set to calculate Emax and EC50 values that are closer to those obtained with a comprehensive data set. Therefore, these equations can generate disparate Emax, and, therefore, EC50values with data sets where maximal induction is not reached. In the present study, these two statistical approaches were evaluated for their fidelity to estimate EC50 and Emax when experimental Emax is not reached. CYP3A4 mRNA induction doseresponse curves were obtained from multiple preparations of cryopreserved human hepatocytes treated once per day for three consecutive days with a vehicle control (0.1% DMSO) or one of three CYP3A4 enzyme inducers, namely rifampin (0.05. 20 mM), ritonavir (0.05. 20 mM), and pioglitazone (0.5. 50 mM). mRNA expression was assessed by qRT-PCR. Fold increase (fold change – 1) was used to calculate the EC50 and Emax for CYP3A4 with both the sigmoid 3-parameter fit and the sigmoidal Hill 3-parmater fit equations with all concentrations included (n = 5). Fidelity of each equation was evaluated by incrementally decreasing the number of concentrations (starting from highest to lowest) prior to calculation of Emax and EC50 values. This method mimics data sets obtained when Emax cannot be reached experimentally. Results confirm that when experimental induction curve reaches a plateau (i.e. rifampin), the calculated EC50 is consistent across cultures; whereas the Emax exhibits intra-culture variability commonly associated with donor variability. In these cases, use of the Hill or sigmoid 3-parameter equations yields similar results. However, when toxicity or solubility precludes the use of complete dose-response curves (e.g. ritonavir, pioglitazone), the Hill equation exhibits a tendency to over-predict Emax whereas the sigmoid equation exhibits a tendency to under-predict Emax. In these cases, wide variations in the potency (EC50) were observed across multiple hepatocyte donors. In general, the Hill equation resulted in a better prediction of the ‘true’ Emax and EC50 values compared to the sigmoid equation when incomplete curves were used. These data sets produced lower apparent EC50 values (i.e. more potent) when processed with the sigmoid equation. When the experimental Emax is not reached (i.e. the resulting curve is bell-shaped or remains exponential through the highest concentration), the resulting Emax and EC50values may be subject to variability based on data processing methods. In these cases, it is important to understand and consider the pitfalls of these methods when evaluating the potential for clinical enzyme induction from in vitro data sets.

P187. ASSESSMENT OF CYP450 ENZYME INDUCTION IN FRESH HUMAN HEPATOCYTES: COMPARING FDA AND EMA DDI GUIDELINES Ilona Schreck1, Satoru Yasuda2, Stefan Beck1, Christine Baumhauer1, Dominik Heer1, Yumi Nishiya2, Noriko Okudaira2, and Veronika Rozehnal1 1 Tissue and Cell Research Center Munich, Daiichi Sankyo Europe GmbH, Munich, Germany and 2Drug Metabolism and Pharmacokinetics Research Laboratories, Daiichi Sankyo Co., Ltd, Tokyo, Japan To date, primary cultures of human hepatocytes are supposed to be the best in vitro test system to predict cytochrome P450 (CYP) induction in the human liver in vivo (European Medicines Agency, 2012; FDA, 2000). In the present study differences between FDA and EMA drug–drug interaction (DDI) guidelines are addressed. Particularly, both authorities differ in their recommendation for CYP2B6 inducers. FDA suggest to use the indirect constitutive androstane receptor (CAR) activator phenobarbital (PB) in a concentration range of 0.5–1 mM whereas EMA stated to use CITCO (100 nM), a direct activator of CAR. Beside that only FDA recommends to include a non-inducer control in the study design, however without any concrete recommendation. Moreover, EMA requests to use hepatocytes with an initial viability 480%, whereas FDA gives no recommendation on viability. In order to meet both FDA and EMA guideline criteria, sandwich cultured freshly isolated human hepatocytes were incubated for 72 hours with the known inducers rifampicin at 0.1–30 mM (for CYP3A4), PB at 0.01–3 mM or CITCO at 0.01–10 mM (both for CYP2B6). As non-inducers gatifloxacin (1–10 mM) and quercetin (10–100 mM) were tested (Grasela et al., 2000; Li et al., 2009). After the incubation CYP3A4, CYP1A2 and CYP2B6 mRNA levels were determined using qPCR. Formation of hydroxybupropion from bupropion was used as specific reaction to determine CYP2B6 enzyme activity. In addition, 45 batches of fresh human hepatocytes with an initial viability between 70. 90%, determined by trypan blue exclusion method, were analyzed for correlation with their CYP3A4 induction potential. At the authorities recommended concentrations, PB (1 mM) showed higher induction potential than CITCO at 100 nM (14.8 ± 3.4 vs. 3.1 ± 1.6-fold induction on enzyme activity level and 8.0 ± 2.2 vs. 2.6 ± 2.4 -fold induction on mRNA expression level). Moreover, interindividual differences in CYP2B6 induction were more pronounced upon CITCO treatment. Compared to PB CYP2B6 induction was not seen in all donors after 100 nM CITCO. No induction potential for all investigated CYP enzymes was observed after treatment with gatifloxacin and quercetin, however quercetin showed cytotoxicity at 100 mM (63.8 ± 24.1% of control mitochondrial activity). The correlation of initial viability and induction potential revealed that cells with an initial viability between 70 and 80% showed similar induction potential compared to hepatocytes with a viability higher than 80%. In conclusion, PB induced CYP2B6 not only markedly higher compared to CITCO but also in all donors without exception. Therefore, PB seems to be more suitable as positive control for CYP2B6 induction in fresh human hepatocytes. The preferred non-inducer control in our study design is gatifloxacin due to the quercetin cytotoxicity. The correlation analysis displayed that hepatocytes with a viability between 70 and 80% are acceptable for CYP induction studies.

References European Medicines Agency. (2012). Guideline on the investigation of drug interactions. London, UK: European Medicines Agency.

128

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

FDA. (2000). FDA guidance for industry: Drug interaction studies – Study design, data analysis, and implications for dosing and labeling, draft guidance. Grasela DM, LaCreta FP, Kollia GD, et al. (2000). Open-Label, nonrandomized study of the effects of gatifloxacin on the pharmacokinetics of midazolam in healthy male volunteers. Pharmacotherapy 20:330–335. Li L, Stanton JD, Tolson AH, et al. (2009). Bioactive terpenoids and flavonoids from Ginkgo biloba extract induce the expression of hepatic drug-metabolizing enzymes through pregnane X receptor, constitutive androstane receptor, and aryl hydrocarbon receptormediated pathways. Pharm Res 26:872–882.

P188. CYTOCHROME B5 AND NADH INCREASE THE POTENCY OF CYP1A1 TO DETOXIFY AND ACTIVATE BENZO[A]PYRENE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Petr Hodek1, Radek Indra1, Michaela Moserova1, Eva Anzebacherova2, Jiri Hudecek1, Eva Frei3, Volker M. Arlt4, and Marie Stiborova1 1 Department of Biochemistry, Charles University in Prague, Prague, Czech Republic, 2Department of Medical Chemistry and Biochemistry, Palacky University, Olomouc, Czech Republic, 3German Cancer Research Center (DKFZ), Heidelberg, Germany, and 4Analytical and Environmental Sciences Division, King’s College London, London, UK Benzo[a]pyrene (BaP) is an environmental carcinogen, which needs a metabolic activation to form covalent adducts with DNA. Microsomal cytochrome P450 (CYP) 1A1 and epoxide hydrolase (mEH) are the most important enzymes involved in the BaP activation. Rat hepatic microsomes oxidized BaP to BaP-9,10-dihydrodiol, an unknown metabolite Mx, BaP-7,8-dihydrodiol, BaP-4,5-dihydrodiol, BaP-1,6-dione, BaP-3,6-dione, BaP-9-ol and BaP-3-ol. Even though the majority of BaP metabolites are detoxification products, BaP-9-ol is a precursor of 9-hydroxy-BaP-4,5-epoxide, which can, similarly to BaP-7,8-dihydrodiol9,10-epoxide, react with deoxyguanosine in DNA. Thus, factors effecting the balance between detoxification BaP metabolites and products forming DNA adducts are of a great importance. While the role of CYP1A1 and mEH in the metabolism of BaP seems to be clear, the involvement of cytochrome b5, NADH:cytochrome b5 reductase, and NADH as a donor of electrons for the CYP-mediated BaP metabolism is still rather enigmatic. That is why, we investigated the impact of cytochrome b5 and NADH on the CYP1A1-mediated BaP metabolism, namely on the metabolite profile and the BaP-DNA adduct formation in vitro. Since human and rat CYP1A1 are quite similar in BaP oxidation, the rats were selected as suitable animal model mimicking BaP oxidation in human. An up to more than 3.5-fold increase in total amount of BaP metabolites was found in hepatic microsomes, in which CYP1A1 was induced. Interestingly, comparable amounts of same BaP metabolites were detected when NADPH was substituted with NADH. In an artificial system of SupersomesTM containing expressed CYP1A1 and CYPOR, BaP was oxidized to the analogous spectrum of metabolites as in microsomes in the presence of either NADH or NADPH. The supplementation of this system with cytochrome b5 increased total BaP oxidation more than 2-fold, while the levels of BaP-7,8-dihydrodiol, BaP9,10-dihydrodiol and BaP-3-ol were elevated more intensely. The formation of two BaP-derived DNA adducts was investigated by 32P-postlabeling. Induction of CYP1A1 in microsomes resulted in a more than 25-fold increase in total level of DNA adducts. The BaP-DNA adduct formation was not dependent on the type of reduced cofactors (NADH or NADPH). In the reconstituted system containing CYP1A1 and CYPOR, the only one BaP-DNA adduct was formed. Hence, this DNA adduct is attributed solely to BaP activation via CYP1A1. When mEH was added to the CYP1A1-CYPOR reconstituted system the second adduct was also detected. The addition of cytochrome b5 to this reconstituted system stimulated the formation of both BaP-DNA adducts, but their levels were dependent on ratios of CYP1A1:CYPOR:mEH:cytochrome b5. These results demonstrate that the CYP1A1-mediated BaP metabolism is stimulated by cytochrome b5 and suggest that NADH can, to some extent, substitute NADPH as an electron donor. This work has been supported by GACR (P303/10/G163).

P189. EPIDERMAL GROWTH FACTOR PREFERENTIALLY DOWN-REGULATES BASAL, BUT NOT INDUCED CYP3A4 AT SUPRA-PHYSIOLOGICAL CONCENTRATIONS IN PLATED HUMAN HEPATOCYTES George Zhang, Thuy Ho, Robert J. Clark, and David M. Stresser Corning Gentest Contract Research Services, Corning Life Sciences, Woburn, MA, USA Epidermal growth factor (EGF) is a well-known mitogen that has importance in cell proliferation and differentiation. EGF is common additive to certain cell culture media, including hepatocyte culture media. Like other growth factors, EGF has been previously shown to modulate constitutive cytochrome P450 (CYP) expression in vitro. Here we assessed the influence of EGF on the basal expression and inducibility of CYP3A4, CYP1A2 and CYP2B6 in plated human hepatocytes. Cryopreserved human hepatocytes from two donors were treated with EGF over a range of concentrations (0.001–500 ng/mL) in the presence or absence of a single concentration of a known inducer: 10 mM rifampicin (RIF) for CYP3A4, 50 mM omeprazole for CYP1A2 and 1 mM phenobarbital for CYP2B6. After the three day treatment, enzyme activity was determined with probe substrates and relative CYP mRNA levels measured by qPCR analysis. Treatment of hepatocytes with EGF resulted in a concentration-dependent decrease in basal CYP3A4 activity up to 92% and 89% for donors 295 and 311, respectively. IC50 values were 0.81 and 0.76 ng/mL for the two donors. In contrast, the RIF-induced CYP3A4 activity was reduced only 18% to 23% by the same treatment (therefore, no IC50 values were calculated). These results demonstrated that EGF caused a preferential decrease in basal CYP3A4 activity. IC50 values

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

129

ranging from 0.54 to 0.83 ng/mL were obtained for down-regulation in basal but also RIF-induced CYP3A4 mRNA in the two donors. While maximum decrease of mRNA was substantial in both basal and RIF-induced (477%), basal mRNA levels were decreased to a much greater extent than RIF-induced mRNA levels. Thus, the fold induction exhibited by RIF for both activity and mRNA can be significantly enhanced with addition of EGF to the media. With the supra-physiological (10) concentrations of EGF (4–10 ng/mL), fold induction of mRNA and activity ranged from 25 to 26-fold and 22 to 28-fold for the two donors, respectively, compared to 4.3- to 5.2-fold and 2.0- to 2.4-fold induction for comparable treatment without EGF supplementation. In contrast to CYP3A4, CYP1A2 and CYP2B6 activity and mRNA were not as highly suppressed by EGF nor was the effect preferential to basal or induced. These data demonstrate that EGF as a media additive can enable a considerably higher dynamic range in a CYP3A4 induction assay. This finding has implications for increasing assay dynamic range and decreasing the treatment exposure time needed to achieve an acceptable induction response in a CYP3A4 screening assay.

P190. THE HETEROCYCLIC AROMATIC AMINE, 2-AMINO-3-METHYLIMIDAZO [4,5-F] QUINOLINE (IQ) INDUCES HUMAN CYTOCHROME P450 1A2 THROUGH THE ARYL HYDROCARBON RECEPTOR (AHR) AND XENOBIOTIC RESPONSIVE ELEMENT (XRE) IN HUMAN HEPATOCYTES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Medjda Bellamri1, Julie Rageul1, Georges Baffet1, Robert Turesky2, and Sophie Langouet1 1 Inserm U1085 IRSET, University of Rennes 1, Rennes, France and 2Department of Medicinal Chemistry, Univeristy of Minnesota, Minneapolis, MN, USA Heterocyclic aromatic amines (HAA) are produced in cooked beef, poultry, and fish, as well as in tobacco smoke condensate and exhaust gases. Activation of HAAs, mainly occurring in the liver, is known to involve N-oxidation by cytochrome P450 (CYP) 1A2, followed by O-acetylation with subsequent formation of a reactive intermediate, nitrenium ion, able to form DNA adducts. Important interspecies differences exist in term of metabolism activation underlying the importance of studingHAA bioactivation in human liver cells. In order to establish if HAAs were able to modulate metabolic activation, we studied the regulation of metabolism enzymes in human primary hepatocytes. Among the 20 HAAs actually identified, 2-amino-3methylimidazo [4,5-f] quinoline (IQ) is a potent carcinogen inducing tumors in multiple tissues and in different species. By treating human hepatocytes with IQ, we observed a marked increase of CYP1A1/2 expression at mRNA, protein levels as well as at activity levels using several substrates. The CYP1A induction is specific with IQ because no effect was observed with the other HAAs studied (PhIP, MeIQx, AaC). The extent of induction was similar to that found with 3-methylcholanthrene, a wellrecognized CYP1A inducer and was not effective in rat hepatocytes in primary culture. This effect required active gene transcription and was mediated by IQ ability to transactivate xenobiotic responsive element (XRE) sequence of the CYP1A gene promoter; it was abolished by a co-treatment with resveratrol, a well-known AhR antagonist. Direct in vitro ligand binding competition assays confirmed the AhR activation by IQ. These results represent the first demonstration that IQ can specifically induce CYP expression in human hepatocytes and that IQ is a potent AhR agonist. These effects need to be considered in human risk assessment of HAAs as their bioactivation is CYP1A2 dependent in human liver.

P191. EVALUATION OF A SINGLE-WELL CYTOCHROME P450 INDUCTION ASSAY FOR CYP3A4/5, 1A2 AND 2B6 USING HEPATOCYTES DERIVED FROM HEPARG CELL LINE Rachel Mandell Basile1, Gary Park2, David Lau2, David Sweeny2, Maurice Standlee2, and Susan Grepper3 1 Department of Drug Metabolism and Pharmacokinetics, Rigel Pharmaceuticals, South San Francisco, CA, USA, 2Department of Drug Metabolism and Pharmacokinetics, Rigel Pharm, South San Francisco, CA, USA, and 3CellzDirect/Life Technologies, Durham, NC, USA Induction of drug-metabolizing enzymes can represent a major hurdle in drug development. Many drugs are enzyme inducers, thereby resulting in autoinduction (time-dependent change in pharmacokinetics due to increase in clearance) or causing pharmacokinetic interaction with other drugs metabolized by the induced pathways. Since enzyme induction could be speciesspecific, fresh primary human hepatocytes had been used as a standard non clinical assay to evaluate potential of enzyme induction. Recently, a new cell line (HepaRG) was shown to express intact PXR,CAR and AhR nuclear receptors, and was shown to be inducible with known CYP inducers. In this study, CYP induction assays using HepaRG cells were constructed and evaluated, in place of fresh human hepatocytes. The extent of induction, based on CYP activity and mRNA quantitation, were compared. For each run of the assay, one vial of HepaRG cells (10  106 cells per vial) was plated onto a full 96-well plate for assay with three positive controls (rifampin, phenobarbital and omeprazole), two vehicle controls (0.1%DMSO and 0.5% EtOH), and triplicate aliquots of 9 unknown compounds at three concentrations. The mRNA assay included three housekeeping genes, allowing correction by normalizing mRNA expression to one of these genes for identifying cytotoxicity as opposed to a lack of induction. Cells were incubated with compound beginning on the 4th day after plating for 72 hours with media changes each day. Probe substrate metabolism (testosterone, phenacetin and bupropion) was measured at 90 minutes by relative production of a known metabolite (6 beta hydroxy testosterone, acetaminophen, and 3 hydroxy bupropion) as measured by LC/MS/MS. mRNA measurement was by a specialized luminex assay read on a Flexmap 3D. Although inter-plate variability exists based on the fold of CYP induction (generally within 3-fold) by the positive controls, the enzyme induction potential of unknown compounds can be consistently evaluated based on comparison with the positive control data within the same plate. The microsomal CYP3A4

130

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

mechanism based inhibition assay was run for the 9 unknown compounds and the Ritonavir positive control. In the future this may allow us to predict this circumstance without confirming by mRNA quantitation. There was a good correlation between the fold induction in activity and mRNA expression using the positive controls for both the 3A4 (rifampicin fold induction of expression/ fold induction of activity = 1.48 and phenobarbital fold expression/fold induction = 1.25) and the 2B6 isoforms(phenobarbital fold expression/fold induction = 2.80). In the mRNA assay, we also tested those wells for induction of other proteins and found In a single well, 3A4/5, 1A2, and 2B6 activity by probe substrate can be simultaneously tested. The mRNA expression levels of the cells in that same well can be measured and compared to the activity information along with housekeeping gene expression. A simple, data-rich assay for predicting potential CYP450 induction was evaluated and implemented. In the future, induction of other enzymes or transporters could be potentially tested using these cells or co cultured cells.

P192. ENHANCEMENT OF OUR UNDERSTANDING OF CYP3A4 INDUCTION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Xin Yang1, Cheng Chang2, Keith Riccardi2, Li Di2, Ronald S Obach2, and Odette Fahmi2 1 Department of Pharmacokinetics, Dynamics & Metabolism, Pfizer, Groton, CT, USA and 2PDM, Pfizer, Groton, CT, USA Induction is an important underlying mechanism of clinical drug–drug interactions (DDI). Accurate prediction of human clinical DDI from in vitro induction data is important to guide timing of DDI studies and DDI trial designs. Induction EC50 and Emax values are ordinarily estimated using the nominal concentration of test compound added to in vitro assays. However, it is possible that the nominal concentration may not be truly reflective of the actual concentration to which the induction receptor (e.g. PXR, AhR, etc) is exposed because the test compound may reversibly bind assay components and it may also undergo metabolism and/or degradation. In fact, such possibilities have been highlighted in recent documents from government regulatory authorities. To further explore this, we measured the free drug concentration over time in a standard human hepatocyte induction assay and compared with nominal concentration. We also evaluated the relationship between duration of exposure and induction to better understand if induction is more related to inducer Cmax or AUC. The results showed that rifampin induction response based on enzymatic induction was driven by unbound AUC. Furthermore, the actual measured free concentration was 11-fold lower than the nominal concentration; it is likely due to non-specific binding to plastics and hepatocytes, as well as metabolic clearance. This study provides a framework on more accurate characterization of in vitro induction parameters to further enhance clinical DDI prediction.

P193. MODULATION OF SELECTED DRUG-METABOLIZING ENZYMES BY MONEPANTEL IN SHEEP IN VIVO Lucie Stuchlikova1, Lenka Lecova1, Tereza Stolcova1, Hana Petnikova1, Jiri Lamka2, Barbora Szotakova1, Vladimir Kubicek3, and Lenka Skalova1 1 Department of Biochemical Sciences, Faculty of Pharmacy, Charles University in Prague, Hradec Kralove, Czech Republic, 2 Department of Pharmacology and Toxicology, Faculty of Pharmacy, Charles University in Prague, Hradec Kralove, Czech Republic, and 3Department of Biophysics and Physical chemistry, Faculty of Pharmacy, Charles University in Prague, Hradec Kralove, Czech Republic Monepantel (MOP), anthelmintic drug intended for treatment and control of gastrointestinal roundworms (nematodes) infection and associated diseases in small ruminants, was introduced in the veterinary practice in 2009 in New Zealand as ZolvixÕ . Contact of an organism with certain drug or other xenobiotics may cause induction of drug-metabolizing enzymes. Induction of these enzymes may affect efficacy of simultaneously or subsequently administered drugs and may be manifested as drug–drug interactions. Although MOP have been used in veterinary practice for several years, no information about its effect on cytochromes P450 (CYPs) and other drug-metabolizing enzymes in target species have been available. The present study was focused on the induction effect of MOP on selected drug-metabolizing enzymes, as these information could be important of prediction of possible drug–drug interactions as well as for effective therapy and limitation of drug-resistance development. For this purpose, sheep were divided into two groups: the first group was treated with MOP (in one recommended dose), the second group represented the untreated controls. After 24 hours, liver and small intestines were removed and subcellular fractions were prepared. Specific activities of selected drug-metabolizing enzymes (CYPs, UDP-glucuronosyltransferases; UGT, glutathione S-transferases; GST and sulfotransferases; SULT) were assayed using spectrophotometric orspectrofluorimetric methods. CYPs and UGT activities were found only in liver, while GST and SULT were detected also in small intestine. MOP treatment led to significant increase of CYPs and GST-related activities in ovine liver over to controls. MOP-mediated induction of CYPs and GST could accelerate deactivation of drugs which are substrates of these drug-metabolizing enzymes. Therefore, interaction of MOP with several other drugs is highly probable in concurrent or consecutive administrations. This project was supported by Czech Science Foundation, grant No P502/10/0217 and by Grant Agency of Charles University, grant No 673612/B-CH/2012.

P194. OIL FROM SEA BUCKTHORN BERRIES AS A POTENTIAL SOURCE OF BIOACTIVE COMPOUNDS Jiri Hudecek, Adela Michalikova, Jan Blecha, and Petr Hodek Department of Biochemistry, Charles University in Prague, Prague, Czech Republic

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

131

Sea buckthorn (Hippophae rhamnoides L.) is a popular raw material for numerous nutritional products. This usage relies mainly on the known high content of various kinds of biochemically active compounds in the sea buckthorn berries. These are, in the first place, a very rich source various of antioxidants (vitamin C, tocopherol, flavonoids). In addition to the water soluble components, the sea buckthorn berries are rich also in lipids, which may be extracted or pressed out to form the sea buckthorn oil. In difference to most other plant species, the lipids are present not only in the seeds, but also in the flesh of fruits. In some areas, e.g. in Mongolia, the sea buckthorn oil is used in supportive ointments for treatment of injuries, and burns, or is applied orally to aid the reconvalescence. Chemically, it is known to contain a large proportion of unsaturated fatty acids. A distinctive chemical feature of the seabuckthorn oil is its relatively high content of palmitooleic acid. We studied several samples of commercial sea buckthorn oil (from Mongolia and the Czech republic), comparing their basic physical and chemical characteristics. The differences found suggest that the commercial ‘‘sea buckthorn oil’’ is often diluted with some other edible oil and colored with plant extracts (containing carotene). To some extent, though, the differences may also reflect the considerable variability of the oil, reported previously for the products from different geographic regions. We examined also the potential of the orally applied seabuckthorn oil to induce cytochrome P450 in rat liver. This work has been supported by GACR (P303/10/G163).

P195. COMPARISON OF ENDOGENOUS 4BETA-HYDROXYCHOLESTEROL AND 6BETA-HYDROXYCORTISOL WITH CYP3A PROBE DRUGS AS MARKERS FOR CYP3A INDUCTION IN CYNOMOLGUS MONKEY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Harunobu Tahara, Miwa Watanabe, and Maki Hasegawa Translational Research Unit, R&D Division, Kyowa Hakko Kirin Co. Ltd., Shizuoka, Japan The induction of cytochrome P450 enzymes is one of the risk factors for drug–drug interactions (DDI) leading to pharmacokinetic and/or pharmacodynamic changes. Several DDI studies including CYP3A induction in cynomolgus monkey to predict the human DDI were evaluated in preclinical stage of drug development. To date, 4beta-hydroxycholesterol ratio in plasma and 6beta-hydroxycortisol ratio in urine have been reported to be endogenous CYP3A makers in human and cynomolgus monkey (Bjo¨rkhem-Bergman et al., 2013; Li et al., 2014). The purpose of this study was to compare the endogenous 4betahydroxycholesterol level in plasma and 6beta-hydroxycortisol level in urine with plasma level of CYP3A probe drugs as markers for CYP3A induction by rifampicin in cynomolgus monkey. Before and after oral administration of rifampicin to male cynomolgus monkey at doses of 0.2, 2 and 20 mg/kg for 4 days (n = 4), the plasma concentration of midazolam was examined after oral administration of midazolam at dose of 2 mg/kg. In addition, the plasma concentration of 4beta-hydroxycholesterol and the urinary concentration of 6beta-hydroxycortisol were also examined to compare the detection ability of CYP3A induction between endogenous and exogenous CYP3A makers. The CYP3A induction could be detected by CYP3A probe drugs, even at the lowest dose of rifampicin (0.2 mg/kg), but not by 4beta-hydroxycholesterol ratio and 6beta-hydroxycortisol ratio at the lowest dose. The mean fold-induction from baseline at rifampicin doses of 0.2, 2 and 20 mg/kg was 1.8, 5.4 and 6.8, respectively, for midazolam clearance, 1.0, 1.3 and 2.1, respectively, for 4beta-hydroxycholesterol ratio and 1.2, 2.5 and 2.6, respectively, for 6beta-hydroxycortisol ratio. The fold-induction estimated by midazolam clearance was higher than those by the endogenous markers. The intestinal metabolism of midazolam by CYP3A might be contributed to the fold-induction. The endogenous CYP3A makers are simple and easy to measure and useful for monitoring the CYP3A induction in multiple ascending dose study such as toxicity test, but not enough for detection of intestinal CYP3A induction. In conclusion, CYP3A probe drugs were more sensitive to detect the CYP3A induction rather than endogenous CYP3A makers in cynomolgus monkey.

References Bjo¨rkhem-Bergman L, Ba¨ckstro¨m T, Nyle´n H, et al. (2013). Comparison of endogenous 4b-hydroxycholesterol with midazolam as markers for CYP3A4 induction by rifampicin. Drug Metab Dispos 41:1488–1493. Li K, Zhao S, Zhang L, et al. (2014). 4b-Hydroxycholesterol as an endogenous biomarker of CYP3A activity in cynomolgus monkeys. Drug Metab Dispos 42:839–843.

P196. CORNINGÕ HEPATOCELLS: AN IN VITRO SCREENING TOOL FOR PREDICTING CLINICAL CYP3A4 INDUCTION Rongjun Zuo1, Feng Li1, Sweta Parikh1, Mercyanne Andes1, Kirsten Cooper1, Enne Akor2, George Zhang2, Jin Liu3, Ronald A. Faris3, and Christopher J. Patten1 1 Corning Life Sciences, Bedford, MA, USA, 2Corning Life Sciences, Woborn, MA, USA, and 3Department of Science and Technology, Corning Incorporated, Corning, NY, USA Induction-mediated drug–drug interactions need to be carefully characterized in vitro for drug candidates in order to determine and/or predict their induction potential in patients. Currently, both the FDA and EMA recommend using primary human hepatocytes as an in vitro test system for this purpose. Several models ranging from simple to complex, such as basic models (Cmax /EC50, relative induction score, R3) and mechanistic models (net effect model and PBPK model), are being proposed to predict clinical CYP3A4 induction using in vitro data from primary human hepatocytes. In light of the large lot-to-lot variation inherent in primary human hepatocytes, CorningÕ HepatoCells (derived from primary human hepatocytes) was evaluated for its

132

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

applicability as an in vitro screening tool for predicting clinical CYP3A4 inducers. Three lots of cells were treated with a list of clinical strong, moderate/weak inducers, and non-inducers at more than 8 concentrations each. Both enzymatic activity (testosterone 6b hydroxylase activity) and mRNA expression were measured as endpoints using LC-MS/MS and RT-PCR, respectively. EC50 and Emax were estimated from concentration-dependent induction response curves. Three basic models (Cmax /EC50, RIS, R3) were employed using EC50 and Emax data for the model drugs, and were evaluated for their ability to predict potential clinical CYP3A4 inducers/non-inducers. It was found that all 3 models using both enzymatic activity and mRNA data correlated very well (R240.9) with clinical clearance data of CYP3A4 substrates. The results also showed that all 3 lots of CorningÕ HepatoCells behaved similarly to primary human hepatocytes in terms of prediction, with only minor lot-to-lot variations for the 3 lots of cells. In conclusion, CorningÕ HepatoCells can be used as a potential in vitro screening tool for prediction of clinical CYP3A4 induction.

P197. ABSTRACT WITHDRAWN

P198. THE ROLE OF INHIBITORY METABOLITES IN THE P450-BASED DRUG–DRUG INTERACTIONS OF AMIODARONE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Matthew G. McDonald and Allan E. Rettie Department of Medicinal Chemistry, University of Washington, Seattle, WA, USA Amiodarone, a class III anti-arrhythmic agent, is the most commonly used drug for treatment of patients with atrial fibrillation. Amiodarone is known to interact with a wide variety of therapeutic agents, with many of these drug–drug interactions (DDIs) resulting from inhibition of cytochrome P450. Four specific P450 enzymes appear to be implicated in the majority of these metabolism-dependent in vivo interactions: CYP1A2, CYP2C9, CYP2D6 and CYP3A4. Amiodarone, itself, appears to be a fairly weak in vitro inhibitor of these enzymes, raising the possibility that inhibitory metabolites play a more direct role than the parent drug. Our goals were to investigate the role that inhibitory metabolites play in P450-based amiodarone DDIs, determine the mechanism(s) by which these interactions might occur and, finally, to use in vitro inhibitory data to predict clinical DDIs. To this end, IC50 shift and time dependent inhibition experiments were carried out to measure the ability ofamiodarone and its circulating human metabolites to reversibly inhibit/inactivate CYP1A2, CYP2C9, CYP2D6 and CYP3A4 activities in human liver microsomes. [I]u/Ki,u values were calculated and used to predict in vivo amiodarone DDIs for pharmaceuticals metabolized by these four enzymes. Based on these values, the minor metabolite di-desethylamiodarone (DDEA) was predicted to be the major cause of DDIs with xenobiotics that are primarily metabolized by CYP1A2, CYP2C9 or CYP3A4, while amiodarone and its mono-desethyl derivative (MDEA) were the most likely cause of interactions involving inhibition of CYP2D6 metabolism. Amiodarone drug interactions predicted from the reversible inhibition of CYP2C9, CYP2D6 and CYP3A4 activities were found to be in excellent agreement with the magnitude of reported clinical DDIs with warfarin, metoprolol and simvastatin. Time dependent inhibition experiments showed DDEA to be a potent inactivator of CYP1A2 (Ki = 0.46 mM, kinact = 0.030 min  1), while MDEA was a moderate inactivator of both CYP2D6 (KI = 2.7 uM, kinact = 0.018 min1) and CYP3A4 (KI = 2.6uM, kinact = 0.016 min1). For both DDEA and MDEA, mechanism-based inactivation appears to occur through formation of a metabolic intermediate (MI) complex. Additional metabolic studies strongly suggest that CYP3A4 is the primary enzyme involved in the metabolism of amiodarone to both MDEA and DDEA. These studies demonstrate both the diversity of likely inhibitory mechanisms with amiodarone and the need to consider metabolites as the ‘culprit’ in inhibitory P450-based DDIs.

P199. IN VITRO SYSTEM-DEPENDENT INHIBITION OF CYTOCHROME P450 ENZYMES (CYP), UDP-GLUCURONOSYLTRANSFERASES (UGT) AND TRANSPORTERS BY OLIGONUCLEOTIDES Faraz Kazmi, Phyllis Yerino, Ellis Bixler, Chase McCoy, and David B. Buckley XenoTech, LLC, Lenexa, KS, USA We previously examined the cytochrome P450 (CYP) inhibition potential of oligonucleotides and reported that the phosphorothioate, but not phosphodiester, backboned molecules caused potent inhibition of CYP1A2 and CYP2C8 in human liver microsomes (HLM) but not in cryopreserved human hepatocytes (CHH) (Buckley et al., 2009). In this study we expanded upon our previous work to include direct, time- (TDI) and metabolism-dependent inhibition (MDI) of 7 CYP and 8 UGT enzymes in pooled HLM and pooled CHH. Additionally, inhibition of OATP1B1, OATP1B3, OAT1, OAT3, OCT1, OCT2, Pgp and BCRP was also examined. Briefly, two oligonucleotides with either phosphodiester (oligo # 1 and #2) or phosphorothioate (oligo #3 and #4) backbones were evaluated in HLM to determine IC50values without a pre-incubation (direct inhibition) or with a 30 min pre-incubation +/ NADPH (MDI and TDI evaluation respectively). For UGTs, UDPGA was added to HLM and inhibition potential was assessed with a single concentration of oligos (30 mM). For CHH assays, MDI was assessed with 30 and 90 min pre-incubations, and 4-methylumbelliferone (4-MU) was used as a marker of overall UGT activity. Marker substrate incubations with CYP and UGT enzyme probes were performed for 5–10 min and residual activity was assessed by LC/MS/MS analysis. For uptake transporter assays, oligos were first pre-incubated for 15 min with transporter transfected HEK cells

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

133

followed by incubation of substrate-oligo mix for 2 min. Uptake of radiolabeled substrates was measured by scintillation counting. Inhibition of efflux transporters was evaluated with either Caco-2 or MDCK cells in bi-directional assays. The results indicated that oligos #1 and #2 with the phosphodiester backbone had little to no inhibitory effect on all CYP and UGT enzymes in HLM and CHH with the exception of oligo #2 in HLM for CYP2C19. Conversely, the oligos with the phosphorothioate backbone caused direct inhibition of CYP and UGT enzymes. The rank order of direct inhibition in HLM for oligo #3 was CYP1A24CYP2C84UGT1A14UGT2B174CYP2B64UGT1A94CYP2C194UGT2B104UGT2B154UGT1A34 UGT1A64CYP2D64CYP2C9. Similar results were observed for oligo #4: CYP1A24UGT1A14CYP2C84 UGT2B174CYP2B64UGT2B104UGT1A94UGT2B154UGT1A34UGT1A64CYP2C19. There was evidence of NADPH-independent inhibition of CYP enzymes with oligo #3 and #4 for CYP2B6, CYP2C8, CYP2C19, CYP2C9, CYP2D6 and CYP3A4/5. In contrast to HLM, we observed little to no direct inhibition by any oligo in CHH (with the exception of oligo #2 with CYP2C19 and TDI by oligo #3 with CYP2C8), demonstrating system-dependent outcomes. Inhibition was observed for all uptake but not efflux transporters by the phosphorothioate oligos. Overall this study has implications on the design of in vitro DDI studies for oligonucleotide based therapeutics based upon test system-dependent differences.

References Buckley DB, Kazmi F, Yerino P, et al. (2009). Inhibition of cytochrome P450 (CYP) enzymes, CYP1A2 and CYP2C8, by oligonucleotides in human liver microsomes (HLM): A system-dependent outcome. Drug Metab Rev 41:94–95.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P200. MULTI-STEP INACTIVATION OF CYP1A2 BY ENOXACIN AND SUSCEPTIBILITY TO FUTILE CYCLING Ryan P. Seguin1, Dustin M. Smith2, Kantipudi N. Babu1, Wendel L. Nelson1, and Kent L. Kunze1 1 Department of Medicinal Chemistry, University of Washington, Seattle, WA, USA and 2Department of Drug Discovery Support, Boehringer-Ingelheim Pharmaceuticals Inc, Ridgefield, CT, USA Enoxacin causes clinically-significant DDI with cytochrome P450 1A2 (CYP1A2) substrates which cannot be explained by reversible inhibition by parent drug. We found enoxacin causes time- and NADPH-dependent inhibition of CYP1A2 activity. This observation prompted investigation into the role of metabolites in CYP1A2 inhibition. The N-hydroxylated metabolite, N-OH-enoxacin, was shown to inactivate CYP1A2 faster than enoxacin in both human liver microsomes (HLM) and CYP1A2 bactosomes. We sought to test the hypothesis that enoxacin is sequentially metabolized by CYP1A2 to form a metabolicintermediate complex (MIC). MIC formation was detectable by UV-Vis spectroscopy and enzyme activity was restorable by treatment with ferricyanide. Our proposed mechanism for enzyme inactivation involves at least three oxidative steps. N-hydroxylation appears to be the first on-path step towards enzyme inactivation. We suspect further oxidation occurs to yield a cyclic nitrone followed by ring-opening and oxidation to a nitroso/oxime. Cyclic nitrone and nitroso/oxime metabolites have been tentatively identified by mass spectrometry. The cyclic nitrone is trappable with cyanide and in rapid equilibrium with the ring-opened primary hydroxylamine as indicated by co-elution of a + 18 Da peak under UPLC-MS/MS. Each intermediate species leading up to enzyme inactivation may be released from the enzyme active site or further metabolized without release to generate the irreversible enzyme-metabolite complex. Our data suggest that a significant portion of enzyme inactivation events occur via a non-dissociative process whereby intermediates are not released. The concentration of N-OH-enoxacin observed in a typical incubation would not be expected to drive an appreciable rate of enzyme inactivation. Importantly, NADPH-dependent reduction of N-OH-enoxacin to parent drug was observed in HLM. The reductive process is expected to disallow rapid accumulation of hydroxylamine metabolites and, therefore, suggests that enzyme inactivation will not be driven primarily by hydroxylamine free in solution. Non-dissociative enzyme inactivation proceeds without release of intermediates and is, therefore, not susceptible to reductase activity. This may explain why enzyme inactivation occurs in systems where intermediates accumulate slowly and reductase activity dominates.

P201. ASSESSMENT OF INHIBITORY POTENTIAL OF DARINAPARSIN ON THE MAJOR HUMAN CYP ENZYMES Gang Luo1, Hongliang Cai2, Ryan Keehn1, Mikaela Mullady1, Michael Schicker1, Richard Voorman1, and John Barrett2 1 Department of Drug Metabolism and Disposition, Covance Laboratories Inc., Madison, WI, USA and 2ZioPharm Oncology, Inc., Boston, MA, USA Darinaparsin (S-dimethylarsino-glutathione), a metabolic intermediate of inorganic arsenicals that occurs in vivo, is believed to generate volatile cytotoxic arsenic compounds when glutathione concentrations are low. Darinaparsin has demonstrated inhibition activity against cultured myeloma cells and is currently in Phase II testing against T-cell lymphoma. Since Darinaparsin will likely be coadministered with drugs that are substrates of CYP enzymes, the present study was undertaken to assess the in vitro inhibitory potential of Darinaparsin on the major human CYP enzymes. Darinaparsin did not show direct inhibition of CYP2B6 (bupropion hydroxylase), CYP2C9 (diclofenac 40 -hydroxylase), CYP2C19 (S-mephenytoin 40 hydroxylase), or CYP2D6 (bufuralol 10 -hydroxylase) with concentrations up to 100 mM. However, Darinaparsin demonstrated direct inhibition of CYP1A2 (phenacetin O-deethylase), CYP2C8 (amodiaquine N-deethylase), and CYP3A4/5 (testosterone 6bhydroxylase and midazolam 10 -hydroxylase). The IC50 value of Darinaparsin was 13.5 mM for CYP2C8 and 25 mM for CYP3A4/ 5 (midazolam 10 -hydroxylase). The IC50 value was estimated to be 4100 mM for CYP1A2 and 4400 mM for CYP3A4/5 (testosterone 6b-hydroxylase) because Darinaparsin showed less than 50% inhibition on these two CYP enzymes. Additional

134

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

characterization experiments demonstrated competitive inhibition of CYP1A2 (Ki 96.7 mM), mixed inhibition of CYP2C8 (Ki 4.69 mM) and CYP3A4/5 (midazolam 10 -hydroxylase, Ki 29.2 mM). Darinaparsin showed metabolism–dependent inhibition of CYP2B6 but not CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, and CYP3A4/5. Metabolism–dependent inhibition of CYP2B6 had KI 46.9 mM and kinact 0.0232 minute1. Given the plasma Cmax of Darinaparsin (738 ± 134 ng/mL, equivalent to 1.79 ± 0.326 mM) in patients following intravenous administration of a therapeutic dose (300 mg/m2) is 0.382 of the lowest Ki value, the weak inhibitory effect of Darinaparsin on the major CYP enzymes should have minimal impact on the CYP-mediated metabolism of concomitant drugs.

P202. MECLIZINE, A PREGNANE X RECEPTOR AGONIST, IS AN INHIBITOR OF CYTOCHROME P450 3A ENZYME ACTIVITY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Winnie Yin Bing Foo1, Eric Chun Yong Chan1, and Aik Jiang Lau2 1 Department of Pharmacy, National University of Singapore, Singapore, Singapore and 2Department of Pharmacy and Pharmacology, National University of Singapore, Singapore, Singapore Pregnane X receptor (PXR), which belongs to the superfamily of nuclear receptors, regulates the expression of numerous genes, including those encoding CYP3A4 and CYP3A5. Previously, meclizine, which is a H1 antagonist, was reported to be an agonist of human PXR. It increased CYP3A4 mRNA expression, but decreased CYP3A-mediated testosterone 6b-hydroxylation in primary cultures of human hepatocytes. Given that the decrease in CYP3A activity was still evident at 24 h after the last dose of meclizine and that the hepatocytes were washed at the end of the treatment period prior to the addition of testosterone substrate, we hypothesized that meclizine inhibits CYP3A activity in an irreversible or quasi-irreversible manner. CYP3A catalytic activity was assessed by a testosterone 6b-hydroxylation assay conducted in pooled human liver microsomes and recombinant CYP3A4 and CYP3A5 enzymes. The amount of 6b-hydroxytestosteronne was quantified by a validated ultra-high performance liquid chromatography tandem mass spectrometry method. Meclizine inhibited CYP3A activity in human liver microsomes, CYP3A4 recombinant enzyme, and CYP3A5 recombinant enzyme with an IC50 of 9 mM, 3 mM, and 12 mM, respectively. Enzyme kinetic analysis indicated that meclizine inhibited hepatic microsomal CYP3A activity by a mixed-mode mechanism and with an apparent Ki of 30 ± 6 mM. To assess whether meclizine is a mechanism-based inactivator of CYP3A, enzymes were preincubated with meclizine prior to initiating testosterone 6b-hydroxylation assay with an aliquot of the preincubation mixture. Preincubation of meclizine with microsomal enzymes for 30 min resulted in a decrease in CYP3A activity, compared to the meclizine group without preincubation and the vehicle control group at 30 min preincubation. Positive (erythromycin) and negative (ketoconazole) controls for mechanism-based inactivation of CYP3A gave the expected results. Detailed kinetic characterization indicated that meclizine decreased human liver microsomal CYP3A activity in a time- and concentration-dependent manner. The kinactivation was 0.15 min1, half-life was 4.7 min, KI was 83 mM, and ratio of kinactivation to KI was 0.002 min1mM1. The inactivation of microsomal CYP3A was NADPH-independent at preincubation time of 0–15 min but NADPH-dependent at preincubation time 30 min. Dialysis of microsomal enzymes preincubated with meclizine did not restore the CYP3A activity to that of the control level, indicating irreversible inactivation. The inactivation of recombinant CYP3A4 occurred in the presence and absence of NADPH, suggesting that both the parent drug and/or its metabolite(s) may be an inactivator. In contrast, meclizine did not show time-dependent inactivation of recombinant CYP3A5. In conclusion, meclizine inhibited CYP3A by both direct mixed-mode inhibition and mechanism-based inactivation. These findings illustrate that a PXR activator could also be an inhibitor of a PXR-regulated enzyme, thereby giving rise to opposing effects on the functional activity of the enzyme.

P203. INHIBITIORY EFFECTS OF DRONEDARONE AND ITS N-DESBUTYL METABOLITE ON HUMAN CYP3A4 AND CYP3A5 ACTIVITIES Yanjun Hong, Ray Hng Yeo, and Eric Chun Yong Chan Department of Pharmacy, National University of Singapore, Singapore, Singapore Mechanism-based inactivation (MBI) of CYP450 has the potential to cause clinically relevant drug–drug interactions (DDIs). Dronedarone is an antiarrhythmic agent approved in 2009 for the treatment of atrial fibrillation. However, hepatotoxicity and DDIs limit its clinical usage. Dronedarone is mainly metabolized by CYP3A4 and CYP3A5 via N-desbutylation and oxidation. In-house preliminary findings demonstrate that dronedarone inhibits CYP3A4 and CYP3A5 in a time-dependent manner. Based on the terminal amino functional group that may be metabolized to a nitroso intermediate, we further hypothesize that dronedarone and N-desbutyl dronedarone (NDD) have the propensity to cause MBI of CYP450 via the formation of a metaboliteintermediate (MI) complex. Therefore, this study aims to investigate the MBI of CYP3A4 and CYP3A5 by dronedarone and NDD. In vitro dilution experiments were performed using recombinant CYP3A4 and CYP3A5 to establish their time-, concentration-, and NADPH-dependent inactivation by dronedarone and NDD. Inactivation kinetic and partition ratio parameters were further derived. Excess CYP3A4/5 probe substrates were added in the dilution experiments to determine whether the MBI could be protected. Spectral difference scanning was performed to assess the presence of MI complex. Our findings demonstrated for the first time that both dronedarone and NDD cause inactivation of CYP3A4 and CYP3A5, exhibiting time-, concentration-, and NADPH-dependence and saturation kinetics. For CYP3A4 inactivation, KI and kinact were 0.873 mM and 0.0389 min1 respectively for dronedarone, and 6.242 mM and 0.0988 min1 respectively for NDD. For CYP3A5

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

135

inactivation, KI and kinact were 2.186 mM and 0.0056 min1 respectively for dronedarone, and 5.445 mM and 0.0563 min1 respectively for NDD. The partition ratios for the inactivation of CYP3A4 and CYP3A5 were approximately 61.1 and 32.2 for dronedarone, and 35.3 and 36.6 for NDD, respectively. Testosterone protected both CYP3A4 and CYP3A5 from inactivation by dronedarone and NDD. The formation of the Soret peak that was reversible by addition of potassium ferricyanide confirmed the presence of MI complex between both dronedarone and NDD and the CYP enzymes. In conclusion, dronedarone and NDD cause MBI of CYP3A4 and CYP3A5 via the formation of MI complex with ferrous heme iron. Our results highlight the unique inhibitory actions of dronedarone against CYP3A enzymes and shed new lights on the possible mechanisms of its clinical DDIs.

P204. ABSTRACT WITHDRAWN

P205. CHEMICAL INHIBITION OR GENETIC DISRUPTION OF SOLUBLE EPOXIDE HYDROLASE DOES NOT PROTECT FROM ACUTE KIDNEY INJURY IN A MOUSE RENAL ISCHEMIA REPERFUSION MODEL

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Svetlana Markova1, Yingmei Liu2, Jean Olson3 and DL Kroetz4 1 Department of Drug Metabolism and Pharmacokinetics, MyoKardia, South San Francisco, CA, USA, 2Department of Bioengineering and Pharmaceutic Science, University of California San Francisco, San Francisco, CA, USA, 3Department of Anatomic Pathology, University of California San Francisco, San Francisco, CA, USA, and 4Department of Biopharmaceutical Sciences, University of California, San Francisco, San Francisco, CA, USA Ischemia-reperfusion (IR) acute kidney injury (AKI) is a significant clinical problem. Endogenous molecules recruited in IR pathways have been targeted over the past decade to ameliorate renal IR injury. Epoxyeicosatrienoic acids (EETs) elicit antiinflammatory, anti-apoptotic and vasoprotective effects in various tissues (Bellien & Joannides, 2013). Unfortunately, the biological effects of EETs are limited by their rapid metabolism by soluble epoxide hydrolase (sEH), encoded by EPHX2. Given anecdotal evidence of EETs being renoprotective in renal IR AKI (Lee et al., 2012) and our previous results demonstrating that EETs are protective against cisplatin-induced AKI (Liu et al., 2012), we designed this study to investigate renoprotective properties of EETs in a well-established mouse model of IR AKI using Ephx2 (/) mice and a novel selective inhibitor of sEH t-TUCB. A 20 min renal ischemia period followed by 24 hr reperfusion resulted in significant kidney damage measured by increases in plasma creatinine (10-fold) and urea nitrogen (6-fold), kidney Kim-1 mRNA (88-fold) and urine NGAL (30-fold) (P50.05). Treatment of the mice with the sEH inhibitor t-TUCB resulted in an increase in plasma 9,10-EpOME/ DiHOME ratios (8- and13-fold) and 12,13-EpOME/ DiHOME ratios (57- and 86-fold) in sham and IR mice, respectively (P50.001). However, inhibition of sEH had no protective effect on renal injury markers. Similarly, sEH inhibition with t-TUCB had no effect on IR-induced increases in renal markers of inflammation (TNFa, TNFR1/R2, iNOS, Cox-2, IL-1b, IL-6, and MCP-1), hypoxic damage (HIF-1a) and neovascularization (VEGF, VEGFR). There was also no protection of sEH inhibition against apoptotic changes induced by IR as confirmed by TUNEL staining of kidney sections. Similar results were obtained in IR studies performed in Ephx2/ mice. Despite previous evidence that sEH inhibition has anti-inflammatory and anti-apoptotic effects in the mouse kidney, we found that genetic or chemical disruption of sEH function does not attenuate IR AKI in a mouse model. These data suggest that the involvement of EETs is not critical for the manifestation of AKI under conditions of IR. The possibility of transient renoprotective effects of EETs during early times following the renal insult requires further investigation.

References Bellien J, Joannides R. (2013). Epoxyeicosatrienoic acid pathway in human health and diseases. J Cardiovasc Pharmacol 61:188–196. Lee JP, Yang SH, Lee H-Y, et al. (2012). Soluble epoxide hydrolase activity determines the severity of ischemia-reperfusion injury in kidney. PLoS One 7:e37075. Liu Y, Webb HK, Fukushima H, et al. (2012). Attenuation of cisplatin-induced renal injury by inhibition of soluble epoxide hydrolase involves nuclear factor kB signaling. J Pharmacol Exp Ther 341:725–734.

P206. SYNTHESIS OF FURANOCOUMARIN, BENZOFURAN AND COUMARIN DERIVATIVES POSSESSING AN INHIBITORY EFFECT ON HUMAN CYP2A6 Yuki Yamaguchi1, Naozumi Nishizono1, Teruki Yoshimura2, Keiji Wada2, and Kazuaki Oda1 1 Department of Pharmaceutical Chemistry, School of Pharmaceutical Sciences, Health Sciences University of Hokkaido, Hokkaido, Japan and 2Department of Food and Chemical Toxicology, School of Pharmaceutical Sciences, Health Sciences University of Hokkaido, Ishikari-Tobetsu, Hokkaido, Japan Human microsomal cytochrome P450 (CYP) enzymes are responsible for the great portion of the metabolism of known drugs in humans. However, CYP2A6 which metabolizes less than 3% of therapeutic drugs, including nicotine and tegafur, has been proposed as a novel target for smoking cessation. CYP2A6 also has been shown to be involved in the mutagenic activation of promutagens such as the tobacco-specific nitrosoamines, for example pharmacogenomic studies have suggested that male smokers completely lacking CYP2A6 were more resistant to lung cancer. Methoxsalen (1, CYP2A6 Ki: 0.80 mM; CYP3A4 Ki: 25.0 mM) decreases both tobacco-related cancer risk and smoking frequency by inhibiting CYP2A6 activity. CYP2A6 inhibitor

136

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Downloaded by [112.134.238.102] at 07:30 10 February 2016

(R)-(+)-menthofuran (2, CYP2A6 Ki: 0.84 mM; CYP3A4 IC50: 4100 mM) has a more selective inhibitory effect than 1, however it induces hepatotoxicity and hepatic carcinogenic activity. The primary aim of this work was to synthesize a series of furanocoumarin, benzofuran, and coumarin derivatives with a view to determining the relationship between the structure of the inhibitors and their ability to inhibit CYP2A6 activity. Another aim was to develop CYP2A6 inhibitors that do not inhibit CYP3A4 activity, because CYP3A is responsible for more than 50% of CYP-dependent human xenobiotic metabolism. Furanocoumarin, benzofuran and coumarin derivatives having structures similar to those of 1 or 2 were synthesized, and then their ability to inhibit CYP2A6 was examined. Furanocoumarin derivative 3 showed weak CYP2A6 inhibitory effect (CYP2A6 IC50: 4.42 mM). Benzofuran derivative 4 (CYP2A6 Ki: 0.45 mM; CYP3A4 Ki: 12.5 mM) displayed CYP3A4 inhibitory activity comparable with that of 1, but decreased selectivity against CYP2A6. Aromatization of 2 leads to an increase in CYP3A4 inhibitory activity, because 4 may be able to undergo -stacking with the phenylalanine cluster that is involved in the initial recognition of CYP3A4 substrates. Coumarin derivatives 5 (CYP2A6 Ki: 0.08 mM; CYP3A4 Ki: 424.0 mM) and 6 (CYP2A6 Ki: 0.01 mM; CYP3A4 Ki: 46.60 mM) showed a stronger and more selective inhibitory effect on CYP2A6 than 1. Coumarin derivatives 5 and 6 may serve as leads structure to develop even more potent and selective inhibitors of CYP2A6 that may prove to be valuable for the development of novel non-nicotine smoking cessation agents.

P207. BIOACTIVATION OF FURANS BY HUMAN LUNG AND LIVER MICROSOMES AND S9 Arno Knorr, Quentin Dutertre, Martin Almstetter, Elyette Martin, Mark Bentley, and Patrick Vanscheeuwijck Philip Morris International R&D, Philip Morris Products S.A., Neuchatel, Switzerland Direct exposure to inhaled mainstream cigarette smoke causes smoking-related damange in the human lung. Although the lung has a significant capability for biotransforming such compounds many of the harmful and potentially harmful compounds in cigarettes smoke (‘‘HPHCs’’) thereby reducing their acute toxicity, in some instances biotransformation can result in the generation of more reactive, and thus more toxic, metabolites. This is the case for furan and simple alkylfurans1. As part of our research on potentially harmful smoke constituents, we are assessing whether the generation of such reactive metabolites can enable the establishment of a link between exposure to such compounds and disease relevant mechanisms or pathways (e.g. oxidative stress, inflammation, apoptosis). Several representative furans present in cigarette smoke were evaluated for their potential to generate reactive electrophilic intermediates or unstable toxic phase II conjugates from subcellular biotransformation (model system: liver microsomes and liver S9; target system: smoker and non-smoker lung microsomes, and lung S9). Metabolic activation of the selected compounds (furan, 2-methylfuran, 2,4- and 2,5-dimethylfuran, furfural, 5-methylfurfural, 5-hydroxymethylfurfural (5– HMF), furfuryl alcohol and 5-methylfurfuryl alcohol) was assessed as a simple mixture and as components within a complex aqueous cigarette smoke fraction. The aqueous smoke fraction, commonly referred to as whole smoke bubbled phosphate buffered saline (wsPBS), was generated by passing the mainstream whole smoke of 6 conventional reference cigarettes through 36 mL of phosphate buffered saline at pH 7.4 using a sintered glass impinger. The concentrations for each furan present within the mixtures applied to subcellular fractions were 2 nmol/mg and 40 nmol/mg protein for liver and 5 nmol/mg protein for lung subcellular fractions, representing concentrations relevant to those present within the applied wsPBS fraction (between 4 and 40 nmol/mg protein for liver and between 1 and 10 nmol/mg protein for lung subcellular fractions). Due to the highly labile nature of reactive electrophilic intermediates, a trapping approach using glutathione as a nucleophile to generate more stable adducts was adopted. Analysis of these adducts by liquid chromatography coupled to mass spectrometry (LC-MS) was used to screen for reactive intermediates generated by CYP2E1 phase I metabolism of the furans. A mixture of unlabeled and stable isotope labeled glutathione (glycine-13C2, 15N) provided a specific isotopic signature for the conjugates, enabling a generic screening approach for glutathione adducts and the identification of downstream modifications of the initial adducts. Metabolites of the selected furans, representing the known major metabolic pathways, were identified and the rates of metabolite formation were assessed for the different human subcellular systems employed. A broad range of glutathione-adducts of electrophilic reactive intermediates for furans could be found in both liver and lung subcellular systems, formation rates for toxified species being 100 times lower in the lung compared to the liver. Reactive sulfate conjugates of 5–HMF and 5-methylfurfuryl alcohol were also formed by both liver and lung S9, with similar differences in formation rate for lung, being 50 times lower compared to liver.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

137

Reference Peterson LA. (2013). Reactive metabolites in the biotransformation of molecules containing a furan ring. Chem Res Toxicol 26:6.

P208. Comparison of intestinal metabolism for CYP3A substrates between rats and humans

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Keitaro Kadono, Yoichi Naritomi, and Kenji Tabata Astellas Pharma Inc., Ibaraki, Japan Understanding the species differences between rats and humans is important, as the former are frequently used in pharmacokinetic studies during drug discovery. Although net fraction of dose absorbed (Fa) has been reported to be similar between rats and humans (Chiou & Barve, 1998; Zhao et al., 2003), little information is available regarding species differences in intestinal availability (Fg) between the two. Here, we used the portal-systemic concentration difference method to directly estimate the fraction of dose absorbed in the portal vein (FaFg) of male Sprague-Dawley rats for nine CYP3A substrates for comparison with human FaFg. No distinct differences in FaFg between the two species were observed, and values for seven of the nine compounds were within a two-fold difference. Given that the Fa values of both species are expected to reach complete absorption (Fa = 1) from parallel artificial membrane assays, this result indicates a moderate correlation in Fg between the two species. Rats might therefore be appropriate for evaluating the intestinal absorption and metabolism of CYP3A substrates. However, our colleagues have also reported poor correlation of FaFg between rats and humans for eight UGT substrates that were expected to have a high Fa. (Furukawa, 2014) The correlation between Fg between rats and humans might therefore differ depending on the enzymes contributing to intestinal metabolism.

References Chiou WL, Barve A. (1998). Linear correlation of the fraction of oral dose absorbed of 64 drugs between humans and rats. Pharm Res 15:1792–1795. Furukawa T, Naritomi Y, Tetsuka K, et al. (2014). Species differences in intestinal glucuronidation activities between humans, rats, dogs and monkeys. Xenobiotica 44:205–216. Zhao YH, Abraham MH, Le J, et al. (2003). Evaluation of rat intestinal absorption data and correlation with human intestinal absorption. Eur J Med Chem 38:233–243.

P209. IN VITRO-IN VIVO EXTRAPOLATION OF UGT1A9-MEDIATED CLEARANCE AND APPLICATION IN A PBPK MODEL FOR DAPAGLIFLOZIN Jian Lin, Alyssa L. Dantonio, Kimberly Lapham, Mark Niosi, Christine C. Orozco, Ernesto Callegari, Louis Y. Leung, and Theunis C. Goosen Pharmacokinetics, Dynamics, and Metabolism, Pfizer Global Research and Development, Groton, CT, USA Quantitative prediction of human UDP-glucuronosyltransferase (UGT)-isoform specific clearance (CL) could be achieved if selective substrate relative activity factors (RAFs) are established with recombinantly expressed UGTs (rUGTs) and human microsomes from drug clearing organs. The anti-diabetic agent dapagliflozin is primarily cleared through glucuronidation by UGT1A9 to a major inactive circulating metabolite, dapagliflozin 3-O-glucuronide. In this study, we evaluated the experimental variables in the determination of RAFs utilizing propofol (PRO) and mycophenolic acid (MPA) as UGT1A9-selective probe substrates. These data were used in a physiologically-based pharmacokinetic (PBPK) modeling approach to evaluate the prediction accuracy of hepatic and extrahepatic CL for dapagliflozin obtained from an intravenous 14C-microdose study (Boulton et al., 2013). Unbound intrinsic clearance (CLint,u) values for PRO and MPA were obtained from enzyme kinetic studies in the presence and absence of 2% BSA with rUGTs and microsomes from human liver (HLM), kidney (HKM) and intestine (HIM). These data were subsequently used to establish RAFs for PRO and MPA. The UGT1A9 RAF values for HLM, HKM, and HIM determined with propofol and mycophenolic acid were 0.40, 0.81; 0.82, 1.3; and 0.008 (PRO), respectively. Dapagliflozin CLint,u in rUGT1A9 was 62 mL/min/mg and following application of the two RAF values, the scaled in vivo CLint was 22 and 45 mL/min/kg for liver and 2.9 and 4.5 mL/min/kg for kidney. Applying the well-stirred model for metabolic CL, the predicted systemic blood CL for dapagliflozin was 10.5 and 18.7 L/h using the two different RAFs. These values are in excellent agreement with the observed systemic blood CL of 14.1 L/h1. Since the RAF scalar for intestinal metabolism is small, the impact of intestinal first-pass metabolism during dapagliflozin absorption is predicted to be negligible. This case study exemplifies the value of applying PBPK models to describe the pharmacokinetics of UGT substrates and should be useful in the prediction of human pharmacokinetics of drugs primarily cleared by UGT1A9.

Reference Boulton DW, Kasichayanula S, Keung CF, et al. (2013). Simultaneous oral therapeutic and intravenous 14C-microdoses to determine the absolute oral bioavailability of saxagliptin and dapagliflozin. Br J Clin Pharmacol 75:763–768.

138

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P210. IDENTIFICATION OF DIET-DERIVED CONSTITUENTS THAT INHIBIT INTESTINAL GLUCURONIDATION AT CLINICALLY ACHIEVABLE CONCENTRATIONS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Brandon T. Gufford1, Gang Chen2, Philip Lazarus2, Tyler N. Graf3, Nicholas H. Oberlies3, and Mary F. Paine1 1 Department of Experimental and Systems Pharmacology, College of Pharmacy, Washington State University, Spokane, WA, USA, 2Department of Pharmaceutical Sciences, College of Pharmacy, Washington State University, Spokane, WA, USA, and 3 Department of Chemistry and Biochemistry, The University of North Carolina at Greensboro, Greensboro, NC, USA Drug metabolizing enzymes within enterocytes constitute a key barrier to xenobiotic entry into the systemic circulation. Furanocoumarins in grapefruit juice are cornerstone examples of diet-derived constituents that perpetrate interactions with drugs via mechanism-based inhibition of intestinal cytochrome P450 (CYP) 3A4. Relative to intestinal CYP3A4-mediated inhibition, dietary substance-drug interactions mediated by other enzymes remain understudied. Many drugs and diet-derived constituents undergo extensive glucuronidation in the gut, raising concern for clinically relevant dietary substance-drug interactions mediated via inhibition of intestinal UDP-glucuronosyl transferase (UGT) activity. This hypothesis was tested by applying a working systematic approach to a panel of structurally diverse diet-derived constituents/extracts (n = 15) to identify and characterize potential perpetrators of intestinal UGT-mediated dietary substance-drug interactions. A plate reader-based assay involving the fluorescent UGT probe substrate 4-methylumbelliferone (100 mM), human intestinal microsomes (0.2 mg/mL), and human embryonic kidney cell lysates overexpressing gut-relevant UGT1A isoforms (UGT1A1, UGT1A8, UGT1A10) (0.05, 0.025, or 0.02 mg/mL, respectively) was used to screen milk thistle flavonolignans (silybin A, silybin B, isosilybin A, isosilybin B, silychristin, isosilychristin, silydianin) and associated extracts (silibinin, silymarin), naringin, naringenin, apigenin, kaempferol, quercetin, and epigallocatechin gallate (10, 100 mM). Constituents/extracts that inhibited UGT activity by 450% in at least one enzyme source (n = 14) were selected for IC50 determination. The IC50 of six constituents (silybin A, silychristin, isosilychristin, silydianin, naringenin, kaempferol) with at least one enzyme source was 3 mM, which was more than 30 times lower than putative intestinal (Hoh et al., 2006; Mohamed & Frye, 2010) or relevant citrus juice concentrations (Erlund et al., 2001; Vandermolen et al., 2013), suggesting that these constituents can inhibit intestinal UGT activity at clinically achievable concentrations. These high potencies imply that drug substrates for intestinal isoforms, specifically UGT1A8 and UGT1A10, carry increased risk of interactions mediated by inhibition of intestinal glucuronidation. Such intestinal UGT1A substrates include the cholesterol lowering agent ezetimibe, the immunosuppressant mycophenolic acid, and the anti-cancer agent raloxifene. These results identified multiple diet-derived constituents as potent inhibitors of intestinal UGT activity, warranting advanced modeling and simulation to inform clinical assessment. Ultimately, these efforts will help to provide evidence-based recommendations to both clinicians and consumers about the risk or safety of certain dietary substance-drug combinations.

References Erlund I, Meririnne E, Alfthan G, Aro A. (2001). Plasma kinetics and urinary excretion of the flavanones naringenin and hesperetin in humans after ingestion of orange juice and grapefruit juice. J Nutr 131:235–241. Hoh C, Boocock D, Marczylo T, et al. (2006). Pilot study of oral silibinin, a putative chemopreventive agent, in colorectal cancer patients: silibinin levels in plasma, colorectum, and liver and their pharmacodynamic consequences. Clin Cancer Res 12:2944–2950. Mohamed MF, Frye RF. (2010). Inhibition of intestinal and hepatic glucuronidation of mycophenolic acid by Ginkgo biloba extract and flavonoids. Drug Metab Dispos 38:270–275. Vandermolen KM, Cech NB, Paine MF, Oberlies NH. (2013). Rapid quantitation of furanocoumarins and flavonoids in grapefruit juice using ultra-performance liquid chromatography. Phytochem Anal 24:654–660.

P211. IMPACT OF CYP2J2*7 ON DRUG METABOLISM IN THE HUMAN HEART Jade Huguet, Fleur Gaudette, Francois Be´langer, Ve´ronique Michaud, and Jacques Turgeon CRCHUM, University of Montreal, Montreal, QC, Canada Background & objective: CYP450 epoxygenase 2J2 (CYP2J2) metabolizes arachidonic acid to epoxyeicosatrienoic acids (EETs) which have cardiovascular protective properties. CYP2J2 is the most highly expressed CYP450 in the human heart. CYP2J2*7 (G50T) polymorphism, the most frequent and the most studied functional variant for CYP2J2, results in a reduced transcription of the gene and lower EETs plasma concentrations. The objectives of our study were 1) to evaluate the allelic frequency of CYP2J2*7 in explanted hearts of patients with ischemic vs. non-ischemic heart disease vs. healthy subject; and 2) to determine the impact of CYP2J2*7 polymorphism on heart metabolic activity using ebastine as a CYP2J2 probe-drug. Methods: Human hearts from patients undergoing cardiac transplant (n = 40) and from normal subjects (n = 9) were obtained from the RETEB tissue bank. Microsomes were prepared from ventricular specimen using the differential centrifugation technique. Incubations were performed with heart microsomes and ebastine (0.075–8 mM). Ebastine and hydroxyebastine (OH-EBA) were analyzed by LC-MSMS and enzyme kinetics parameters were calculated. CYP2J2*7 variant was determined by TaqManÕ qRT-PCR. Results: In our cohort of human hearts, CYP2J2*7 allele frequency was 15% (n = 6/40) and 11% (n = 2/18) in non-ischemic and normal heart groups, respectively. In contrast, noCYP2J2*7 variant allele (0/40) was found among the heart group with ischemic disease. No significant difference was observed for ebastine kinetic parameters betweenCYP2J2 genotype groups (CYP2J2*1*1 vs.*1*7) in both groups (non-ischemic and normal heart groups, p40.3). A comparison between CYP2J2*1*7 genotype groups showed higher OH-EBA

Poster Abstracts

DOI: 10.3109/03602532.2015.1071933

139

formation in patients undergoing cardiac transplant vs. normal. Ebastine metabolism in heart microsomes for the CYP2J2*1*1 genotype was lower in normal group compared to both cardiac transplant groups (0.1950.2850.44 mL/h, respectively). Non-ischemic hearts Kinetic parameters

Ischemic hearts

Healthy hearts

CYP2J2*1/*1 n = 14 CYP2J2*1/*7 n = 6 CYP2J2*1/*1 n = 20 CYP2J2*1/*1 n = 6 CYP2J2*1/*7 n = 2

Km (uM) Vmax (pmol OH-EBA/ min/mg prot) Clint (mL/hr)

0.13 ± 0.06 3.02 ± 2.81 0.44 ± 0.27

0.12 ± 0.04 3.78 ± 2.98 0.52 ± 0.29

0.18 ± 0.12 2.67 ± 3.18 0.28 ± 0.18

0.15 ± 0.10 2.02 ± 2.64 0.19 ± 0.22

0.09 ± 0.03 1.72 ± 1.57 0.36 ± 0.28

Conclusions: Our results demonstrated that CYP2J2*7 heterozygous (-50GT) polymorphism did not significantly affect ebastine metabolism in human heart when compared to the respective wild-type group. This finding is in opposite with previous reports which indicate a decrease of function associated with the allelic variant of CYP2J2. We did not find any variant of CYP2J2 in the ischemic group. Some epigenetic factors associated with this pathologic condition might modulate CYP2J2 activities and could explain this discrepancy observed in our study.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P212. CYP450 MEDIATED DRUG METABOLISM IN HUMAN LUNG TISSUE Sophie Gravel1, Jade Huguet1, Eric Rousseau2, Fleur Gaudette1, Jacques Turgeon1, and Ve´ronique Michaud1 1 CRCHUM, University of Montreal, Montreal, QC, Canada and 2CHUS, Centre Hospitalier de l’Universite´ deSherbrooke, Sherbrooke, QC, Canada Background: Lung tissue may play an important role in the metabolism and clearance of drugs particularly following their intravenous administration or by inhalation. Local CYP expression in the lungs and surrounding tissues may affect the homeostasis of the lung as well as its susceptibility to disease or response to therapy.Our objective was to characterize CYP450 activity in human lung tissue using cocktails of CYP450 probe-markers. Methods: Lung tissue specimens were obtained from 6 patients (1.7–13 g). Microsomes were prepared using the differential centrifugation technique. Incubations were performed with validated cocktails of CYP450 substrate-markers containing; (1) bupropion (CYP2B6) (465 uM), ebastine (CYP2J2) (1.5 uM), midazolam (CYP3A4/5) (3 uM) and chlorzoxazone (CYP2E1) (600 uM); (2) bufuralol (CYP2D6) (30 uM) and repaglinide (CYP2C8) (10.2 uM); (3) tolbutamide (CYP2C9) (300 uM); and 4) lauric acid (CYP4A11) (15 uM). Substrate markers and their metabolites were quantified by LC-MS/MS. mRNA levels of CYP450s were quantified by RT-PCR. Results: Our results showed that CYP450 expressed in the lung exhibit drug metabolism capacity. CYP450 activity in human lung was detected particularly for CYP2J2, CYP3A4/5 and CYP2C9. The greatest activity was observed for the formation of hydroxyebastine (0.375 ± 0.350 pmol/min/mg prot) followed by hydroxytolbutamide and 1-hydroxymidazolam (0.173 ± 0.068 and 0.078 ± 0.024 pmol/min/mg prot, respectively). No significant levels of metabolites were detected for the other CYP450 substrates tested. Conclusion: Our data indicate that local drug metabolism through CYP2J2, CYP2C9 and CYP3A4/5 enzymes can be detected in human lung tissue. CYP2J2 showed the higher CYP450 activity in the lung. In contrast with previous reports suggesting low intersubject variability in expression of CYP2J2, we observed a high level of interindividual variability in CYP2J2 activity (19-fold of variability) in human lungs.

P213. ABSTRACT WITHDRAWN

P214. ARE DRUGS METABOLIZED IN HUMAN SKIN? IS THERE A LINK TO SKIN TOXICITY? Nenad Manevski1, Kamal Kumar Balavenkatraman2, Armin Wolf2, Gian Camenisch1, Olivier Kretz1, and Karine Litherland1 1 Department of Drug Metabolism and Pharmacokinetics, Novartis, Basel, Switzerland and 2Department of Pre-clinical safety, Novartis, Basel, Switzerland Human skin, the largest organ of the body, has a potential to metabolize therapeutic drugs, cosmetics, and environmental xenobiotics (Gotz et al., 2012; Hewitt et al., 2013; van Eijl et al., 2012). Despite the clear relevance for the pharmacotherapy, drug metabolism in skin remains poorly understood. Existing experimental models often show lack of robustness and consistency, leading to difficulties in prediction of skin drug clearance. Moreover, high incidence of drug-induced skin toxicity, for example skin rash, raises a question whether reactive drug metabolites are involved in their onset. Taken together, drug development requires a robust model to assess drug metabolism and drug-induced skin toxicity. To address these challenges and closely mimic the physiological conditions, we developed a model based on full-thickness human skin explants. Fresh human skin was procured from the University Hospital in Basel, expediently transported to our Novartis laboratories in organ preservation solution, and used in a number of drug metabolism and toxicity assays. Skin samples were incubated with suitable enzyme probes, and corresponding extracts were analyzed by LC-MS for the presence of specific metabolites. To evaluate drugand metabolite-induced toxicity in human skin, we successfully developed a long-term skin explant culture, up to 14 days. In parallel to activity assays, we initiated experiments to quantify the gene expression of most relevant drug-metabolizing enzymes.

140

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Present results demonstrate that human skin has a significant capacity to metabolize drugs via both phase I and phase II reactions. Several new metabolic reactions were identified for the first time. Activity assays combined with long-term toxicity readouts, such as histology slides and cytokine release profiles, offer a unique model to assess drug metabolism and metaboliteinduced toxicity in the human skin.

References Gotz C, Pfeiffer R, Tigges J, et al. (2012). Xenobiotic metabolism capacities of human skin in comparison with a 3D-epidermis model and keratinocyte-based cell culture as in vitro alternatives for chemical testing: phase II enzymes. Exp Dermatol 21:364–369. Hewitt NJ, Edwards R, Fritsche E, et al. (2013). Use of human in vitro skin models for accurate and ethical risk assessment: metabolic considerations. Toxicol Sci 133:209–217. van Eijl S, Zhu Z, Cupitt J, et al. (2012). Elucidation of xenobiotic metabolism pathways in human skin and human skin models by proteomic profiling. PLoS One 7:e41721.

P215. STRUCTURAL CHARACTERIZATION OF A NOVEL CYCLIZED PRODUCT OF A HYDROXAMIC ACID ANALOGUE AND ITS CONTRIBUTION TO THE HIGH CLEARANCE (CL) AND IN VITRO-IN VIVO CORRELATION (IVIVC) DISCONNECT OF COMPOUND 1

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Amin Kamel and Mithat Gunduz Department of Metabolism and Pharmacokinetics, Novartis Institutes for Biomedical Research Inc., Cambridge, MA, USA The hydroxamic acid moiety plays an important role as a pharmacophore in a variety of biological active agents. Such agents include enzyme inhibitors, anticancer agents, antimicrobial compounds and agents used in the treatment of cardiovascular diseases. In general, hydroxamic acids undergo enzymatic hydrolysis and reduction to the corresponding acids and amides, respectively. Reduction of hydroxamic acids may be mediated via several enzymatic and non-enzymatic systems including blood and Aldehyde Oxidase (AO). The use of 14C-labeled compound1 earlier in discovery was essential to understand the IVIVC disconnect of compound1. Sample analysis was carried out on a Thermo-Finnigan LTQ-OrbitrapÕ with accurate mass measurement and MSn capabilities with radiometric detection using b-RAM. Compound1 is rapidly cleared in rats and dogs following intravenous administration and its CLint values in rat and dog hepatocytes were 8 and 2-fold lower than in vivo clearance values suggesting possible extrahepatic metabolism and /or aldehyde oxidase mediated metabolism. In vitro studies in microsomes and hepatocytes suggested that oxidative metabolism, hydrolysis and reduction to the corresponding acids and amides, respectively, and direct glucuronidation of compound 1 were the major biotransformation pathways. Quantitative and qualitative similarities were observed between rat hepatocytes and rat plasma metabolic profiles and reduction of compound 1 to the corresponding amide metabolite represented the major metabolic pathway. Additional in vitro studies in plasma, blood and cytosol suggested that compound 1 was stable in plasma and AO did not contribute to the reduction of compound 1 to the corresponding amide metabolite. In blood, however, compound 1 seems to have undergone cylization following 1 hr incubation. Data suggested that the formation of the cyclized product proceeds non-enzymatically and catalyzed by the heme moiety of the hemoglobin in blood in the presence of both a reduced pyridine nucleotide and a flavin and may contribute to high CL of compound 1. The blood cyclized product was identified using accurate mass measurements, multiple stage MS and 1H NMR its structure was confirmed unequivocally with synthetic std. A mechanism for the formation of the cyclized product is proposed and its metabolic fate was studied. The reducing system in blood is widely distributed in the animal body, and is likely to be of significance in the in vivo metabolism of hydroxamic acids.

P216. ABSTRACT WITHDRAWN

P217. A NOVEL CAR-MEDIATED MECHANISM FOR REGULATION OF RAT CYP3A1 EXPRESSION Toshie T. Gamou, Wataru Habano, Jun Terashima, and Shogo Ozawa Department of Pharmacodynamics and Molecular Genetics, School pf Pharmacy, Iwate Medical University, Shiwa-Gun, Iwate, Japan Purpose: The cytochrome P450s (CYPs) 3A subfamily members are as important as being involved in the oxidative metabolism of endogenous and exogenous compounds. Constitutive androstane receptor (CAR) is one of the principal regulators of hepatic CYP3A as well as CYP2B. Three most important human CYP3A4 regulatory elements (REs) have been well characterized, namely, a proximal ER6 (proER6), a xenobiotic responsive enhancer module (XREM) and a constitutive liver enhancer module (CLEM). However, as there are no significant REs in rat CYP3A except for the proER6, we set out to investigate a possible involvement of CAR in the rat hepatic CYP3A expression. Methods: Expression plasmids of rCAR were transfected to a rat FAA-HTC1 hepatoma cells by lipofectamine reagents. Gene expressions of rCYP3As and rCYP2Bs in transfectant were analyzed by qRT-PCR. Transactivations of rCYP3a1 promoter regulatory elements by rCAR were assayed for firefly luciferase activity.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

141

Results: By over-expression of mature rCAR in rat FAA-HTC1 hepatoma cells, mRNA expression of rCyp3a1 was dramatically increased. In reporter assays, over-expression of rCAR induced reporter activity through the region containing from 33.2 kb to 30.5 kb of rCyp3a1 promoter. As a result of further analysis, we narrowed down the core enhancer regions located from 31739 to 31585 of rCyp3a1. This region contained two putative binding motifs of nuclear receptors (DR4 and DR2), and putative transcription factor recognition sites; a hepatocyte nuclear factor-1 binding motif (HNF1), a nuclear factor-kappa B binding motif (NFkB) and a activator protein 1 binding motif (AP-1). This region seemed to be direction-dependent enhancer activity. The reporter plasmid containing multiple DR4 and DR2 motifs showed extremely strong transactivation in a rCARdependent manner. Meanwhile, the reporter plasmid containing multiple transcription factor recognition sites indicated weaker transactivation activity. Conclusions: rCYP3a1 was remarkably induced in rCAR-over-expressed FAA-HTC1 cells. We found a novel rCAR-dependent enhancer in a region from 31739 to 31585 of rCYP3a1. On analysis of this region, core enhancer region contained two putative binding motifs of nuclear receptors (DR4 and DR2), and putative transcription factor recognition sites (HNF1), which was involved in the transcriptional activation. We have reached following conclusions: First, DR4 and/or DR2 motifs were primarily responsible for the rCAR-mediated transcription; second, the region of HNF1 worked in CARmediated transcriptional activation in a synergistic manner.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P218. TRANSCRIPTOMIC EVALUATION OF THE RESPONSE OF A DUAL HUMAN HEPATOCYTE-KUPFFER CELL CULTURE SYSTEM TO PROINFLAMMATORY STIMULI Amy C. Ditewig1, Eric A.G. Blomme1, Rita Ciurlionis1, Yi Yang2, and Michael J. Liguori1 1 Department of Cell, Molecular, & Exploratory Toxicology, Abbvie, Inc., North Chicago, IL, USA and 2Department of Investigative Toxicology and Pathology, Abbvie, Inc., North Chicago, IL, USA Primary hepatocyte cultures have been the standard model for in vitro toxicological evaluation of experimental compounds in the pharmaceutical industry, but have limitations for the prediction of drug-induced liver injury (DILI) in humans. A dual culture (DC) system combining human hepatocytes (HHs) and Kupffer cells (KCs) may have more relevance to predict DILI. The purpose of this study was to evaluate a DC system compared to HHs or KCs alone. Hepatocytes alone, KCs alone, or a 1:4 ratio of KCs:HHs were cultured on collagen-I coated plates for 48 hours before dosing for 48 hours with naı¨ve media, 10 mg/mL lipopolysaccharide (LPS), 50 ng/mL interleukin-6 (IL-6), or 50 ng/mL TNFa. RNA was analyzed using Human Genome Affymetrix arrays. Overall, 32.7% of transcripts on the array were expressed by all three naı¨ve culture types. The presence of KCs resulted in the detection of a large number of unique transcripts (5.8% of probe sets were specific to the KC only culture compared to 1.8% which were unique to the HH cultures). When compared to HHs, naı¨ve DCs showed higher expression levels for genes in cell motility, macrophage activation, and immune regulation pathways, likely reflecting cell composition of the cultures. A strong decrease in cytochrome P450 (CYP) and PXR mRNA levels was observed in the DC system, although mRNA levels of CYP1B1 were consistently higher in the DC system. When exposed to LPS, DCs showed enhanced immune mediator stimulation (e.g. IL-1, GM-CSF, MIP-1, IL-17A, and NF-kB signaling), down-regulation of several cytochrome P450s (CYP2B6 and CYP3A7), Phase II enzymes (UGTs, SULTs, ALDHs and GSTs), and transporters (P-GP, MRP-2, OATP2B1, OATP1B1). LPS downregulated mitochondrial respiratory complexes in DCs only and stimulated the hepatic stellate cell activation pathway more robustly in DCs. In the DC system, TNFa-induced changes were more robust when compared to IL-6-induced changes, however both treatments impacted the expected pathways (acute phase response initiation, stellate cell activation, IFN signaling, and antigen presentation). Altogether, these data demonstrated the functionality of this DC in vitro model and suggest that this model may be useful to investigate the role of the innate immune system in the development of DILI. Disclosures: AD, EB, RC, YY and ML are employees of AbbVie. The design, study conduct, and financial support for this research was provided by AbbVie. AbbVie participated in the interpretation of data, review, and approval of the publication.

P219. GENE EXPRESSION OF CYTOCHROME P450S AND MDR1A IN THE SMALL INTESTINE AND LIVER OF RATS WITH INDOMETHACIN-INDUCED SMALL INTESTINAL ULCERS Shoji Kawauchi1, Tsutomu Nakamura2, Ikuya Miki1, Jun Inoue1, Sayo Horibe1, Tsuneo Hamaguchi1, Toshihito Tanahashi1, and Shigeto Mizuno1 1 Kobe Pharmaceutical University, Kobe, Japan and 2Faculty of Pharmaceutical Sciences, Himeji Dokkyo University, Himeji, Japan Background: The small intestine is the primary site of absorption for orally administered drugs, and intestinal metabolism and active extrusion of absorbed drug have been recognized as barriers to drug absorption for over a decade. Non-steroidal antiinflammatory drugs induce the serious side effect of small intestinal ulcerations (SIUs), but little information is available regarding the consequences to drug metabolism. Meanwhile, we previously found the down-regulation of cytochrome P450 (CYP) 3A and P-glycoprotein (mdr1a) in the liver and upper small intestine of the mice with experimental colitis (Kawauchi et al., 2014). Aim: In the present study, we examined the existence of secondary hepatic inflammation in rats with indomethacin (INM)-induced SIUs and assessed its relationship to the CYPs and mdr1a in the small intestine and the liver. Methods: Animals were administered INM (10 mg/5 ml/ kg) by subcutaneous injection. Gene expression of the CYP family of enzymes and mdr1a

142

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

was measured with quantitative real-time polymerase chain reaction. The effects of INM on the expression of human CYP3A4 (isoform of rat CYP3A1) induced by rifampicin were examined in human hepatocellular carcinoma HepG2 cells. Results: INM induced SIUs predominantly in the lower region of the small intestine with high expression of inflammatory markers. Liver dysfunction was also observed, which suggested a secondary inflammatory response in rats with SIUs. In rats with normal and SIU tissues, a significant difference was observed only in CYP2D2 mRNA expression in the upper region of the small intestine. In the liver of rats with SIUs, the levels of CYP2C11, CYP2E1, and CYP3A1 mRNA expression were significantly decreased by more than 85%, and loss of CYP3A protein was observed. In in vitro study, the CYP3A4 expression was not changed by the concomitant treatment of INM with rifampicin, compared to that with rifampicin treatment alone, and also no significant inductive effect of INM was observed. Conclusion: INM-induced SIUs had a subtle effect on intestinal CYP expression, but had an apparent action on hepatic CYP, which was influenced, at least in part, by the enterohepatic circulation.

Reference Kawauchi S, Nakamura T, Miki I, et al. (2014). Down-regulation of CYP3A and P-glycoprotein in the secondary inflammatory response of mice with dextran sulfate sodium-induced colitis and its contribution to cyclosporine A blood concentrations. J Pharmacol Sci 124:180–191.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P220. IN 3D SPHEROID, THE REGULATION PATHWAY OF DRUG-METABOLIZING ENZYME EXPRESSION IS DIFFERENT FROM THE PATHWAY IN 2D MONOLAYER CULTURED CELL Jun Terashima, Shinpei Goto, Ayako Nakamura, Yoko Jinma, Sawaka Hoshi, Wataru Habano, and Shogo Ozawa Pharmacodynamics and Molecular Genetics Lab, School of Pharmacy, Iwate-Medical University, Morioka, Japan We have analyzed the expression pattern and regulation pathway of drug-metabolizing enzyme in 3D tumor spheroid. The drugmetabolizing enzyme expression in 3D tumor spheroid shows distinctive pattern. We report that CYP expression pattern and the regulation mechanism of the expression in 3D tumor spheroids of human liver cancer cells. CYP1A1 and CYP3A4 expression showed locoregional expression, which were overlapping to ATF4 and HIF1A expression region that are marker of glucosedeprived and hypoxia conditions respectively. Previously, we reported that glucose deprived stress induced increasing of CYP expression, and aryl hydrocarbon receptor (AhR) related to the various stress responses in liver tumor cells (Terashima et al., 2011, 2013). CYP1A2 expression showed distinctive pattern, like mosaic pattern in 3D tumor spheroid. Moreover, we defined that CYP expression in 3D tumor spheroids were induced the different pathway from 2D monolayer cultured tumor cells. These results will be helpful in a better understanding of anticancer agent metabolism in clump of cancer cells in vivo.

References Terashima J, Habano W, Gamou T, Ozawa S. (2011). Induction of CYP1 family members under low-glucose conditions requires AhR expression and occurs through the nuclear translocation of AhR. Drug Metab Pharmacokinet 26:577–583. Terashima J, Tachikawa C, Kudo K, Habano W, Ozawa S. (2013). An aryl hydrocarbon receptor induces VEGF expression through ATF4 under glucose deprivation in HepG2. BMC Mol Biol 14:27.

P221. UDP-GLUCOSYLTRANSFERASES (UGTs) GENE EXPRESSION CHANGES IN HAEMONCHUS CONTORTUS AFTER ALBENDAZOLE TREATMENT Lenka Lecova, Lucie Stuchlikova, Petra Matouskova, Barbora Szotakova, and Lenka Skalova Department of Biochemical Sciences, Faculty of Pharmacy, Charles University in Prague, Hradec Kralove, Czech Republic The prevalence of anthelmintic resistance has dramatically increased and has evolved into a serious problem facing small ruminant production in many countries. Benzimidazole anthelmintics, including albendazole, have been used for decades to control gastrointestinal parasites. Resistant parasites have several mechanisms that protect them from the effects of these drugs. Well known b-tubulin gene polymorphism is one of the factors correlated with the resistance. Moreover, highly effectivemechanisms of resistance are differences in anthelmintic metabolism and detoxification. Biotransformation enzymes represent an excellent defence mechanism against the negative action of xenobiotics. An important role in the inactivation and excretion of a great variety of these compounds plays glucosidation. This task in parasites is performed by enzymes from the class of UDP-glucosyltransferases (UGTs). Taking advantage of the newly assembled genome sequences of Haemonchus contortus, we performed a pilot analysis of some gene candidates from UGTs family in this worm. We identified several differences in UGTs expression between susceptible (ISE) and multi-resistant (WR) strain of both genders. Different UGTs expression may play a role in albendazole resistance of WR strain. In this study we tested the ability ofalbendazole to induce UGTs expression in both strains. We analysed expression of UGTs in adult worms exposed to sublethal concentration of albendazole for 24 hours in liquid culture compared to non-exposed adult worms. We analysed both genders of ISE and WR strain of H. contortus. The relative quantity of each gene of interest was calculated using the DDCT method.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

143

This study was supported by the operational programme ECOP, registration number CZ.1.07/2.3.00/30.0061, Increasing of the R&D capacity at Charles University through new positions for graduates of doctoral studies, Czech Republic.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P222. FREQUENCY DETERMINATION OF ALELLIC VARIANTS, OF XENOBIOTIC AND HORMONAL RECEPTORS AHR AND ERa, IN A MEXICAN PRE- AND POSTMENOPAUSAL BREAST CANCER GROUP WOMEN ´ lvarez-Librado, Enoc M. Corte´s-Malago´n, Jose´ Bonilla-Delgado, Mo´nica Sierra-Martı´nez, Carlos A. A Juan C. Bravata-Alcantara, Sonia Cha´vez-Ocan˜a, Claudia V. Arellano-Gutierrez, Diana C. Rodrı´guez-Ballesteros, and Octavio D. Reyes-Herna´ndez Hospital Jua´rez de Me´xico, Unidad de Investigacio´n, Mexico D. F., Mexico Breast cancer (BC) involves hormonal and genetic risk factors. Estrogens, and its metabolites play important roles in BC, and there are several different mechanisms by which estrogens induce cellular and genomic alterations. One of the most important estrogens mechanisms is to bind estrogen receptor alpha (ERa). On the other hand, the arylhydrocarbon receptor (AhR) mediates the adverse effects of dioxins, including modulation of sex steroid hormone signaling. The ligand-activated AhR directly associates with ERa and modulates their function both, positively and negatively. This may explain, in part, the sex steroid hormone-related adverse effects of dioxins. According to a previous study in Thai women, the rs2066853 single nucleotide polymorphism (SNP) in the AhR gene may play an important role in estrogen metabolism and modify individual susceptibility to breast cancer development. Also, several SNPs in the ESR1 (ERa gene) have been identified and found to be associated with either, an increased, or a decreased risk for various diseases such as breast cancer. The best characterized SNPs of ESR1 are the PvuII and XbaI restriction site polymorphisms, both located in the first intron. The development of genetic markers will allow us to identify susceptible individuals to BC. In this study, 96 patients with histopathologically confirmed breast cancer and 153 controls were included. The rs2066853 AhR and PvuII and XbaI ESR1 variants were determinated by qPCR. No differences were observed for the allelic frequency for these 3 variants between both groups. Comparing genotype frequencies, among case and control individuals, only XbaI homozygote carriers shown a statistical difference.

P223. ABSTRACT WITHDRAWN

P224. DYNAMIC PLASTICITY OF ‘‘CLUSTER OF DIFFERENTIATION ANTIGEN’’ EXPRESSIONS AT THE INFLAMMATORY BLOOD–BRAIN BARRIER Kazuki Sato1, Masanori Tachikawa2, Yasuo Uchida2, Sumio Ohtsuki3, and Tetsuya Terasaki2 1 Faculty of Pharmaceutical Sciences, Tohoku University, Sendai, Japan, 2Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan, and 3Department of Pharmaceutical Microbiology, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan Introduction: Neuro-immune diseases such as sepsis and multiple sclerosis recruit leukocytes migration to the central nervous system across the blood–brain barrier (BBB). This prompted us to hypothesize that blocking the leukocytes migration machinery at the BBB is an alternative therapeutic strategy in the diseases. By considering that a variety of cluster of differentiation (CD) antigens play an important role in the migration process, the ones which exhibit the significant changes in the protein expression levels under the disease state could be the most potential drug targets. Thus, the purpose of this study was to establish the protein quantification method of the CD antigens and clarify those protein expression levels at the BBB under inflammatory conditions. Methods: LC-MS/MS-based quantitative targeted proteomics for the CD antigens was established by using a series of internal standards of stable isotope labeled peptides. For quantification, specific peptide sequences of 70 kinds of mouse CD antigens including CD11a, CD18, CD29, CD49d, CD54, CD62E, CD62P, CD102, CD106, and CD162 were selected by means of our in silico peptide selection criteria. To produce a sepsis inflammatory model, ddY mice were subjected to lipopolysaccharide intraperitoneally. Brain capillary (Bcap) were prepared from the mice by the mesh method and digested by trypsin. Results and discussion: This is the first study to determine 17 kinds of protein expression amounts of the CD antigens including CD18, CD29, CD49d, CD54, CD102, CD106, which are involved in the leukocytes migration process, in isolated brain capillaries of the inflammatory model. Remarkably, the protein expression levels of CD54 and CD106 were up-regulated under inflammatory conditions; CD54: 1.55 vs 7.27 fmol/mg protein, and CD106: 0.503 vs 1.78 fmol/mg protein (control vs sepsis model). In contrast, the protein expression levels of CD9 and CD97 were significantly down-regulated under inflammatory conditions; CD9: 10.9 vs 5.79 fmol/mg protein, and CD97: 0.436 vs 0.145 fmol/mg protein (control vs sepsis model). These results suggest that the leukocytes migration machinery at the BBB under inflammatory conditions involves the dramatic changes in the expression levels of CD54, CD106, CD9, and CD97. These CD antigens could be the most potential targetfor blocking the leukocytes migration. Conclusion: The BBB exhibits a dynamic plasticity of the leucocyte migration by changing the expression levels of the CD antigens under inflammatory conditions.

144

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P225. DEVELOPMENT OF SIMULTANEOUS ANALYSIS METHOD FOR IMMUNOGLOBULINS BY LC-MS/MS: INTRALABORATORY COMPARISON WITH ELISA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Katsuaki Ito, Kumiko Wakamiya, Tomomi Hioki, Masanobu Mukoyoshi, Tsutomu Mochizuki, and Yoshinori Kasahara Pharmaceutical Development Research Laboratories, Teijin Pharma Limited, Hino Tokyo, Japan Serum levels of immunoglobulin (Ig) including IgG, IgA, and IgM are useful biomarkers for several diseases such as multiple myeloma and Waldenstrom’s Macroglobulinemia. Although ELISA is a common approach for the bioanalysis of protein biomarkers, it needs independent assays for each Ig. LC-MS/MS analysis has an advantage that multiple analytes can be quantified in one assay, leading to save analytical runs and biological samples. Therefore, in this study, we developed the simultaneous analysis method for IgG, IgA, and, IgM by LC-MS/MS. As mentioned in FDA draft guidance, the compatibility between different analysis methods is a discussion point. Therefore, we compared the quantified values between LC-MS/MS and ELISA. First, the purified Ig sample was prepared from human serum by the sodium sulfate precipitation. After that, the human serum and the purified Ig sample were digested by trypsin, and the specific peptides for each Ig were quantified by scheduled MRM (QTRAP4500). By the developed LC-MS/MS method, the accuracy of calibration curve and QC samples (LQC, MQC, HQC) was within 15% for the three Igs, satisfying the criteria for the bioanalysis. In addition, the LC-MS/MS analysis showed wider dynamic range than that of ELISA (LC-MS/MS: 1000-fold, ELISA: 64-fold). On the analysis of the human serum and the purified Ig sample, the LC-MS/MS analysis showed high precision that the coefficient of variation was within 5% for the three Igs. In comparison with ELISA, the quantified values of IgG, IgA, and IgM were 1.26, 1.59, and 1.07fold higher by LC-MS/MS in the human serum, respectively. In the purified Ig sample, the LC-MS/MS analysis showed 1.25, 1.58, and 1.11-fold higher values for IgG, IgA, and IgM, respectively. Thus, the LC-MS/MS analysis showed higher quantified values than ELISA. Similar difference was also described in previous literature about monoclonal antibody and explained by the hypothesis that ELISA only measures free protein, while LC-MS/MS assay can measure total protein (Swann & Shapiro, 2011). Next, we calculated the purification recovery of each Ig from the human serum. As the result, the calculated recoveries were similar between the LC-MS/MS analysis and ELISA. The recovery of IgG, IgA, and, IgM was 89.7, 84.7, and 88.4% by LC-MS/MS, while 89.7, 85.4, and 84.9% by ELISA. Therefore, it is conceivable that the relative changes of the Ig levels such as fold-change from baseline are comparable between the LC-MS/MS analysis and ELISA, although the absolute quantified values were different. In conclusion, we successfully developed the LC-MS/MS analysis method for IgG, IgA, and IgM. The developed method can quantify the three Igs in one assay with high accuracy, high precision, and wider dynamic range. In addition, it is conceivable that the relative changes of the Ig levels are comparable between the LC-MS/MS analysis and ELISA.

Reference Swann PG, Shapiro MA. (2011). Regulatory considerations for development of bioanalytical assays for biotechnology products. Bioanalysis 3:597–603.

P226. LC-MS/MS-BASED TARGETED PROTEOMICS ASSAY TO DETERMINE THE ABSOLUTE PROTEIN EXPRESSION OF CLINICALLY RELEVANT PHASE I & II ENZYMES Diana Busch1, Christian Gro¨er1, Marek Drozdzik2, Werner Siegmund1, and Stefan Oswald1 1 Department of Clinical Pharmacology, University of Greifswald, Greifswald, Germany and Pharmacology, Pomeranian Medical University, Szczecin, Poland

2

Department of Clinical

Background: The pharmacokinetics of many drugs is markedly influenced by intestinal and hepatic biotransformation enzymes such as cytochrome P450 (CYP450) enzymes and UDP-glucuronosyltransferases (UGT). In order to predict their impact on drug disposition as well as on drug–drug interactions, data on their absolute intestinal and hepatic abundance are required. Therefore, it was the aim of this study to develop and comprehensively validate LC-MS/MS methods for the absolute quantification of clinically relevant CYP and UGT enzymes. Methods: LC-MS/MS methods were developed for CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, CYP3A5, UGT1A1, UGT1A3, UGT2B7 and UGT2B15. Proteins were quantified by measuring proteospecific tryptic peptides using stable isotope-labelled internal standard peptides. The assays were validated with respect to specificity, linearity, within-day and between-day accuracy and precision, stability as well as digestion efficiency and matrix effects according to current bioanalytical guidelines. Results: For the aforementioned 13 proteins, two LC-MS/MS assays were developed. All methods were shown to be selective for the respective enzyme and the analytical range was in each case 0.25–50 nmol/l. Within-day (intra-day) as well as between-day (inter-day) accuracy was between 13.1 and 12.5% (relative error) and precision 1.1–14.8% (coefficient of variation). All peptides were shown to be stable during preparation, storage in the autosampler (24 h at 4  C), during overnight digestion (16 h at 37  C) and during three freeze-thaw cycles. Significant matrix effects could not be observed. The method was successfully applied to measure CYP and UGT expression in human intestinal and human liver microsomes. Conclusion: The developed methods were shown to possess sufficient specificity, sensitivity, accuracy, precision and stability to measure clinically relevant metabolizing enzymes in human tissues. These absolute expression data may allow more precise prediction of drug disposition using PBPK modelling-based approaches.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

145

P227. COMPARISON OF UGT PROTEIN EXPRESSION IN HUMAN COLON AFTER MULTIDOSE, ORAL CURCUMINOID ADMINISTRATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Gary N. Asher1, John K. Fallon2, and Philip C. Smith2 1 Department of Family Medicine, University of North Carolina School of Medicine, Chapel Hill, NC, USA and 2Division of Molecular Pharmaceutics, University of North Carolina Eshelman School of Pharmacy, Chapel Hill, NC, USA Introduction: In vitro studies have demonstrated that curcumin is a substrate for UGT with a putative ability to both induce expression and inhibit function of UGTs, highlighting the potential for interaction with some drugs. Therefore, we sought to evaluate the effect of 4-week exposure to oral curcumin in healthy volunteers on intestinal UGT protein expression. Methods: Healthy volunteers who had received a recent screening colonoscopy at UNC Hospitals were recruited. Participants were ages 40–80 years and did not have any gastrointestinal or bleeding disorders, lab abnormalities, or recent antibiotic use. All participants received daily curcuminoid extract, 4 g, (C3, Sabinsa Corp.) for 28 consecutive days. Unprepped, rectal mucosal pinch biopsies were obtained at baseline and on the final day of curcumin administration. 20–30 mg tissue samples were obtained through a rigid, disposable sigmoidoscope using a disposable flexible biopsy forceps, and immediately frozen in liquid nitrogen and stored at 80  C until analyzed. Microsomes were prepared from biopsy samples using sequential centrifugation. Microsomal protein concentrations were determined using Pierce BCA protein assay. The equivalent of 20 mg protein per sample was used for analysis of protein expression. Quantification of 14 UGT 2As and 2Bs was performed by LC-MS/MS using previously published methods for quantitative targeted absolute proteomics (QTAP) (Fallon et al., 2013). LOD was 0.1 pmol/ mg protein. Comparison of UGT concentrations at baseline and after curcumin exposure was performed using a paired t-test. Results: Paired baseline and 4 week biopsy samples were available for 38 participants. Quantifiable concentrations of UGT 1A10 and 2B17 were detected in 35 and 33 paired samples respectively, while all other UGTs were BLOQ. UGT1A1 was detected at very low concentration in one participant. Mean (SD) concentration for UGT 1A10 at baseline was 0.77 pmol/mg (±0.71) and 0.87 pmol/mg (±0.72) after four weeks of oral curcumin intake (p = 0.48). For UGT 2B17, mean concentration at baseline was 1.46 pmol/mg (±1.73) and 1.60 pmol/mg (±1.58) after four weeks (p = 0.59). Conclusions: We found no differences in rectal mucosal UGT concentrations before and after 28 days of oral curcumin administration, indicating that daily curcumin use is unlikely to alter intestinal UGT expression in the colon. Distal gut biopsies may not accurately reflect the proximal gut environment where UGT expression is higher and curcumin concentrations may also be higher.

Reference Fallon JK, Neubert H, Hyland R, et al. (2013). Targeted quantitative proteomics for the analysis of 14 UGT1As and -2Bs in human liver using nanoUPLC-MS/MS with selected reaction monitoring. J Proteome Res 12:4402–4413.

P228. TARGETED QUANTITATIVE PROTEOMIC ANALYSIS OF UGTS AND CYP450S IN RECTAL, CERVICAL AND VAGINAL TISSUES AND IMPLICATIONS FOR XENOBIOTIC DISPOSITION AND TREATMENT OF DISEASE John K. Fallon1, Corbin G. Thompson2, Angela D. M. Kashuba2, and Philip C. Smith1 1 Eshelman School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA and 2Center for AIDS Research, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Introduction: Identification and quantification of drug metabolizing enzymes in the female genital tract (FGT) and surrounding tissues could enhance the understanding of xenobiotic disposition and treatment of disease. The FGT may be a site of steroid metabolism and has been identified as a potential HIV reservoir. Here, we show application of a targeted proteomics method for the quantification of 14 UGTs and 8 CYP450s, and the qualitative identification of a further 3 UGTs and 2 CYP450s, in FGT and rectal tissue. We examine implications with respect to disposition and treatment. Methods: Normal rectal, cervical and vaginal (n = 10 for each) tissues were collected from gynecologic surgeries and snap frozen. Subsequent homogenization was followed by fractionation using sequential centrifugation. Samples were analyzed by modification of previously reported methods (Fallon et al., 2013). Membrane microsomal protein (10 mg) was denatured with heat, reduced and alkylated. Heavy labeled JPT and Thermo peptides (1 pmol) served as internal standards for quantification. Samples were digested with trypsin (1:20 w/w) at 37  C for 4 h. The reaction was stopped with cold acetonitrile. Samples were treated with solid phase extraction. A Waters nanoACQUITY LC coupled to an AB SCIEX QTRAP 5500 (MRM mode) was used for analysis. Concentrations were calculated relative to heavy labeled peptide standards. For those proteins without heavy labeled peptides qualitative identification criteria included detecting at least two peptides and a minimum of two MRMs for at least one of these. A pooled liver sample was analyzed in duplicate for comparison. Results: UGT2B17 could be quantified in 9 of 10 rectal samples (mean ± SD, 0.99 ± 1.2 pmol/mg protein) above an LOD of 0.1 pmol/mg protein. No other UGTs or CYP450s were quantifiable above similar LODs in any of the three tissues examined. UGT2A1 (2 MRMs for one peptide and at least 1 for a second peptide) was qualitatively identified in all 10 cervical samples, 7 of 10 vaginal samples and 5 of 10 rectal samples. UGT2A1 total peak area response for cervical samples was roughly four times higher than for vaginal samples and five times higher than for rectal samples (n = 10 for each). Conclusions: This is the first report of a UGT/CYP450 quantitative proteomic analysis of cervical, vaginal and rectal tissue. UGTs 2B17 and 2A1 are known to glucuronidate steroids and thus their presence in these tissues is

146

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

perhaps unsurprising. If either isoform has a role in metabolizing, for example, antiretroviral drugs used to treat HIV then this information could be of value for drug development and treatment of disease.

Reference Fallon JK, Neubert H, Hyland R, et al. (2013). Targeted quantitative proteomics for the analysis of 14 UGT1As and -2Bs in human liver using nanoUPLC-MS/MS with selected reaction monitoring. J Proteome Res 12:4402–4413.

P229. THE ROLE OF CYTOCHROMES P450 AND ALDO-KETO REDUCTASES IN PROGNOSIS OF BREAST CARCINOMA PATIENTS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Pavel Soucek1, Viktor Hlava´cˇ1, Veronika Brynychova´2, Radka Va´clavı´kova´2, Marie Ehrlichova´2, David Vra´na3, Va´clav Pecha4, Marke´ta Trnkova´5, Marcela Mrhalova´6, Roman Kodet6, Katerˇina Kuba´cˇkova´3, and Jirˇ´ı Gateˇk7 1 Toxicogenomics Unit, National Institute of Public Health, Prague, Czech Republic, 2National Institute of Public Health, Prague, Czech Republic, 3Palacky University Medical School and Teaching Hospital, Olomouc, Czech Republic, 4Institute for the Care for Mother and Child, Prague, Czech Republic, 5Biolab Praha, Ltd., Prague, Czech Republic, 6University Hospital Motol, Prague, Czech Republic, and 7Tomas Bata University, Zlin, Czech Republic Metabolism of anticancer drugs affects their antitumor effects. This study has investigated the associations of gene expression of enzymes metabolizing anticancer drugs with therapy response and survival of breast carcinoma patients. Gene expression of thirteen aldo-keto reductases (AKRs), carbonyl reductase 1, and ten cytochromes P450 (CYPs) was assessed using quantitative real-time PCR in tumors and paired adjacent non-neoplastic tissues from 68 post-treatment breast carcinoma patients. Eleven candidate genes were then evaluated in an independent series of 50 pre-treatment patients. Protein expression of the most significant genes was confirmed by immunoblotting. AKR1A1 was significantly overexpressed and AKR1C1–4, KCNAB1, CYP2C19, CYP3A4, and CYP3A5 downregulated in tumors compared with control non-neoplastic tissues after correction for multiple testing. Significant association of CYP2B6 transcript levels in tumors with expression of hormonal receptors was found in the post-treatment set and replicated in the pre-treatment set of patients. Significantly higher intratumoral levels of AKR1C1, AKR1C2, or CYP2W1 were found in responders to neoadjuvant chemotherapy compared with non-responders. Patients with high AKR7A3 or CYP2B6 levels in the pre-treatment set had significantly longer disease-free survival than patients with low levels. Protein products of AKR1C1, AKR1C2, AKR7A3, and CYP3A4 were found in tumors and those of AKR1C1 and AKR7A3 correlated with their transcript levels. Prognostic and predictive roles of drug metabolizing enzymes strikingly differ between post- and pre-treatment breast carcinoma patients. Particularly AKR1C2, AKR7A3, CYP2B6, and CYP3A4 shall continue to be further followed as putative prognostic or predictive biomarkers in breast carcinoma patients. This work was supported by the Czech Science Foundation (No. 13-25222J) and the Internal Grant Agency of the Czech Ministry of Health (No. NT/14055-3).

P230. COMPARISON OF PLASMA LIPIDOMIC PROFILE OF HUMANS WITH PRECLINICAL ANIMALS Kosuke Saito1, Masayo Urata1, Katsuko Toyoshima1, Masaki Ishikawa1, Mayumi Murayama1, Yoko Tajima1, Yuya Senoo1, Kazuhisa Takemoto1, Yuji Kumagai2, Keiko Maekawa1, and Yoshiro Saito1 1 Medicinal Safety Science, National Institute of Health Sciences, Tokyo, Japan and 2Clinical Research Center, Kitasato University School of Medicine, Kanagawa, Japan Endobiotic metabolites, including lipids, are associated with the biological processes in the body and thus suitable candidates for biomarkers to evaluate disease state, therapeutic efficacy, and toxicity of drugs. However, information regarding how differences between humans and preclinical animals remain limited. In the present study, we employed metabolomic approach to obtain profiles of plasma lipid molecules in humans (Caucasian, n = 15) and three preclinical animals (mice [BALB/c, n = 5], rats [Sprague-Dawley, n = 5] and rabbits [New Zealand White, n = 5]). All subjects were young male and overnight fasted to minimize the effect of gender, age, and daily nutrients on lipid profiles. In total, the levels of 208 lipids, including 123 phospholipids (11 lysophosphatidylcholine [LPC], 5 lysophosphatidylethanolamine [LPE], 45 PC, 19 ether-type PC [ePC], 13 PE, 18 ePE, 11 phosphatidylinositol [PI] and 1 phosphatidylglycerol [PG]), 39 sphingolipids (26 sphingomyelin [SM], 5 ceramide [Cer], 5 hexosylceramide [HexCer], 3 sulfatide [Su]) and 45 neutral lipids (cholesterol [Ch], 30 Ch ester [ChE], 12 diacylglycerol [DG] and 2 coenzyme Q [CoQ]), were determined by LC/MS. Principal component analysis showed clear separation of four groups. Loading scatter plot have demonstrated that polyunsaturated fatty acid containing ChEs (23:6 ChE and 24:6 ChE), sulfatides (40:1 Su and 42:2 Su) and CoQ10 specified rodents (mice and rats), rabbits and humans, respectively. When compared with humans, levels of LPCs were at higher levels in mice and rats but not in rabbits. PEs were at higher levels in rabbits and mice but at lower levels in rats. Majority of ePCs and SMs were at lower levels in all preclinical animals. In addition, mice showed similar levels of ChEs and DGs with humans, while rats and rabbits did about 2-times lower levels. On the other hand, the levels of each PC varied among species. Thus, we now perform structural analysis to understand the differences in composition of fatty acid chains in PCs among species. Overall, our results clearly demonstrated that plasma lipidomic profiles varied from humans to preclinical animals. This observation would provide useful fundamental information for

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

147

exploring and validating biomarkers in future preclinical studies and may also help to establish the regulatory standards for these studies.

P231. STABLE ISOTOPE RESOLVED METABOLOMICS OF ERLOTINIB AND GEFITINIB RESISTANT NON-SMALL-CELL LUNG CANCER (NSCLC) CELL LINES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Robert M. Mader1, Markus Joerger2, Anton Vychalkovskiy3, and Walter Jaeger4 1 Department of Medicine I, Medical University of Vienna, Vienna, Austria, 2Department of Medical Oncology & Hematology, Cantonal Hospital, St. Gallen, Switzerland, 3Department of Biology, ETH Zurich, Zurich, Switzerland, and 4Department of Clinical Pharmacy and Diagnostics, University of Vienna, Vienna, Austria Non-small-cell lung cancer (NSCLC) is still one of the leading causes of death by cancer worldwide. Initially this aggressive tumor is responsive to erlotinib and gefitinib, but almost always reoccurs in a resistant form resulting in a low 5 year survival rate. To better understand the mechanisms of resistance, endogenous small molecules were profiled by LC/MS-MS in cytoplasm, cellular supernatant and cell-derived exosomes in wild type, and erolotinib- and gefitinib-resistant cell lines after 24 h of drug administration. Indeed, stable isotope resolved analysis revealed pronounced changes in the metabolic pathways between wildtype and erlotinib- and gefitinib-resistant HCC4006 cells. The level of most amino acids (n = 22), biogenic amines (n = 17, sphingomyelins (n = 15), glycerophospholipids (n = 77), acylcarnitines (n = 41) and lysophosphatidylcholines (n = 14) was strongly increased in the cytoplasm and cellular supernatant of gefitinib- but decreased in erlotinib-resistant HCC4006 cells indicating differences in energy metabolism, regulation of cell growth, phospolipase activity and fatty acid beta-oxidation. Interestingly, metabolic pattern of cell-derived exosomes transmitted at least partly the drug resistance phenotype. Especially the biogenic amines creatinine, kynurenine, spermidine, spermine, taurine, the spingomyelins SM (OH) C22:1, SM C16:0, SM C18:1, SMC24:0, SM C24:1, C12-DC, the lysophosphatidylcholines lyso PC a C14:0, lyso PC a C16:0, lyso PC a C18:0 and lyso PC a C28:0 and the acylcarnitines C8 and C12-DC, which showed increased concentrations in the cytoplasm and in the cellular supernatant of gefitinib-resistant cells, were also elevated in exosomes. Metabolomic data from exosomes derived from erlotinibresistant cells, however, did not match with cytoplasm and cellular supernatant and showed profiles similar to those of gefitinib. In conclusion, our results successfully differentiated resistant erlotinib and gefitinib HCC4006 cells from wild type controls. Targeted metabolomics may therefore be a useful tool to identify novel biomarkers for the early diagnosis of drug resistance in lung cancer patients.

P232. CORNINGÕ HEPATOCELLS: AN ALTERNATIVE HEPATIC MODEL FOR IN VITRO ADME/TOX APPLICATIONS Feng Li1, Sweta Parikh1, Mercyanne Andes1, Kirsten Cooper1, Jin Liu2, Enne Akor3, Ronald A. Faris2, Christopher J. Patten1, and Rongjun Zuo1 1 Corning Life Sciences, Bedford, MA, USA, 2Department of Science and Technology, Corning Incorporated, Corning, NY, USA, and 3Corning Life Sciences, Woburn, MA, USA Primary human hepatocytes are considered the ‘‘Gold Standard’’ for studying the metabolic fate of drugs, as they contain the full complement of drug metabolism enzymes, cofactors, transporter proteins, and nuclear receptor pathways for regulation of enzyme expression. However, due to large lot-to-lot variations and limited lot size of primary hepatocytes, significant time and resources are spent on screening a suitable lot (or lots) for desired studies. An alternative model with consistent performance and constant supply will be highly beneficial for in vitro screening and characterization of drug candidates. Here we report characterization of cryopreserved CorningÕ HepatoCells (derived from primary human hepatocytes) for applications such as CYP induction, metabolism, and cytotoxicity studies. When cultured on CorningÕ BioCoatÔ Collagen I coated tissue culture plates with a MatrigelÔ overlay, CorningÕ HepatoCells maintained mature hepatocyte morphology very similar to primary hepatocytes. Upon exposure to prototypical inducers (10 mM Rifampicin, 50 mM Omeprazole, 1 mM Phenobarbital) for 3 consecutive days, the cells demonstrated induction responses in CYP3A4, 1A2, and 2B6 comparable to primary human hepatocytes (measured by both enzymatic activity and mRNA level). Inter-lot and intra-lot variations were assessed, and showed significantly lower variations than primary human hepatocytes. CorningÕ HepatoCells demonstrated concentration-dependent response to aflatoxin B1, a metabolism-based hepatotoxic compound; the toxic response was reverted by incubation with CYP3A4 inhibitor ketoconazole. In conclusion, this model represents a renewable source of human hepatic cells with consistent performance for ADME applications such as CYP induction, metabolism and cytotoxicity.

P233. ORNITHINE CARBAMOYLTRANSEFERASE IS A PROMISING BLOOD BIOMARKER FOR DRUG-INDUCED LIVER INJURY Akira Koibuchi1, Tomomi Furihata1, Takayuki Aizawa1, Minaka Shibuya1, Makoto Murata2, Hiroshi Murayama2, Kaoru Kobayashi1, and Kan Chiba1 1 Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Science, Chiba University, Chiba, Japan and 2 Diagnostic Departments, YAMASA Corporation, Choshi, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

148

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Ornithine carbamoyltransferase (OCT) is one of the urea cycle enzymes that are specifically localized in liver mitochondria, and elevation of serum OCT level has been observed in several types of rat and human liver injuries. Because the extent of OCT elevation is often higher than that of alanine aminotransferase (ALT), OCT has the potential to become a promising biomarker that enables sensitive detection of hepatocellular damages. However, it is currently unclear whether OCT is released from hepatocytes in drug-induced liver injuries (DILI). Accordingly, in this study, we aimed to clarify in vitro profiles of OCT release from hepatocytes in various hepatotoxic drug treatments. To accomplish this, primary hepatocytes from Sprague-Dawley male rats (5–6 weeks) were prepared by a collagenase-perfusion method. The cells were treated with celecoxib, fluoxetine, imipramine, amiodarone or their vehicle (0.5% dimethylsulfoxide, DMSO) at 0–250 mM. The OCT and ALT levels in culture medium were assessed at each concentration point by enzyme-linked immunosorbent assay and enzyme activity determination, respectively. In each treatment, the relative difference in the OCT or ALT release level of drug-treated cells compared to those of DMSO-treated (control) cells was calculated, and maximum relative release level of OCT or ALT (hereafter referred to as Rmax,OCT or Rmax,ALT, respectively) was estimated using a computer program (Deltagraph, JPD, Tokyo). Rmax value was expressed as mean ± S.D. When rat primary hepatocytes were treated with celecoxib, the results showed that Rmax,OCT value (78 ± 25-fold) was clearly higher than Rmax,ALT value (12 ± 4-fold), so that the ratio of Rmax,OCT to Rmax,ALT (Rmax ratio) was 7 ± 2. Similarly, the result of fluoxetine treatment showed that Rmax,OCT value (54 ± 16-fold) was much higher than Rmax,ALT value (8 ± 4-fold), so that the Rmax ratio of which was 8 ± 3. However, when compared to these Rmax,OCT and Rmax,ALT values, it was found that, in imipramine and amiodarone treatments, the Rmax,OCT values were apparently lower (23 ± 9- and 26 ± 9-fold, respectively), while similar Rmax,ALT values were obtained (8 ± 3- and 7 ± 2-fold, respectively). Accordingly, Rmax ratios of imipramine and amiodarone treatments were 3 ± 1 and 4 ± 1, respectively, which were still above 1.0, but lower than those of celecoxib and fluoxetine treatments. To summarize, our results show that OCT is released from rat hepatocytes during the treatments of all four drugs tested in this study. The results also show that Rmax,OCT values are higher than Rmax,ALT values, where it should be noted that the Rmax ratios are clearly different depending on the drugs utilized. Therefore, our present results indicate that OCT is a promising candidate for use as a sensitive and unique blood DILI biomarker.

P234. THE USE OF POOLED PLATED CRYOPRESERVED HUMAN HEPATOCYTES FOR THE DETERMINATION OF METABOLIC CLEARANCE, CYTOCHROME P450 ENZYME INDUCTION AND UPTAKE TRANSPORTER STUDIES Faraz Kazmi, Phyllis Yerino, Donald Miller, Kevin C. Lyon, Catherine Wiegand, Elizabeth Lafreniere, Racquel Mueller, Sara McKinney, Eliza Hodes, Rebecca R. Campbell, Maciej Czerwinski, Seema Muranjan, Joanna E. Barbara, and David B. Buckley XenoTech, LLC, Lenexa, KS, USA During early drug discovery and development, conventional test systems such as subcellular hepatic fractions or pooled, suspended cryopreserved human hepatocytes are used to determine metabolic clearance (intrinsic clearance; CLint) or in the case of cytochrome P450 (CYP) induction, individual cryopreserved or fresh plated primary human hepatocytes. Conventional test systems often fail to predict CLint and half-life (t1/2) of low turnover drugs due to limitations in test system stability (4 hours), leading to inaccurate scaling of in vitro to in vivo CLint. In the case of CYP induction screening, the use of individual plated primary human hepatocytes can lead to inter-individual variability in the induction response or susceptibility to cytotoxicity. Lastly, hepatic uptake experiments are frequently conducted in pooled suspended hepatocytes but typically only with individual plated hepatocytes. In the present study we evaluated the utility of pooled plated cryopreserved human hepatocytes to determine metabolic clearance, CYP induction, and uptake transporter functionality. Briefly, individual plateable primary human hepatocytes were pooled (n = 5) using a proprietary single-freeze pooling process. In vitro metabolic clearance of a range of low to high clearance drugs (typically 1 mM) was evaluated with pooled plated hepatocytes cultured in 48-well plates for 4 hours followed by incubations lasting up to 48 hours. Loss of parent drug was determined by LC/MS/MS analysis. For CYP induction, pooled human hepatocytes were cultured in 48-well plates and treated once daily for 3 days with MCM + medium containing either 0.1% DMSO, rifampin (0.1–20 mM), omeprazole (50 mM), or phenobarbital (750 mM). Following treatment, enzyme activities of CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19 and CYP3A4/5 were assessed by a 2-step cocktail in situ incubation. Additionally, CYP1A2, CYP2B6, CYP2C8, CYP2C19, CYP2C9, and CYP3A4 mRNA expression was assessed by RT-qPCR. Transporter functionality was assessed with 1 min incubations (at 0.1 mM) of estrone sulfate (OATP1B1), CCK-8 (OATP1B3), MPP + (OCT1) or TCA (NTCP), followed by scintillation counting. Characterization of the pooled plated hepatocytes indicated that enzymatic activities were comparable to the theoretical average activity calculated based on individual donor data. The rank order of in vitro CLint was determined as follows: diclofenac4naloxone4midazolam4 verapamil & phenacetin & repaglinide4dextromethorphan4diazepam4tolbutamide & timolol4S-warfarin4disopyramide = zolmitriptan4theophylline; and was consistent with their literature classification as high, medium or low clearance drugs. CYP activity and mRNA analysis indicated that all CYPs tested were inducible with the greatest induction seen in CYP3A4 (rifampin 6.9-fold activity; 6.6-fold mRNA), CYP1A2 (omeprazole; 3.3-fold activity; 20-fold mRNA), and CYP2B6 (phenobarbital; 9.4fold activity; 16-fold mRNA). Uptake transporters were functional with OATP1B1, OATP1B3, OCT1 and NTCP uptake ratios of 42. Overall, these results indicate that pooled plated cryopreserved hepatocytes are a simple and cost effective test system appropriate for assessing metabolic clearance of low turnover compounds, early CYP induction screening and uptake transporter studies.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

149

P235. IN VITRO-IN VIVO EXTRAPOLATION OF HUMAN BILIARY CLEARANCE IN SANDWICH CULTURED HUMAN HEPATOCYTES Emi Kimoto, Yi-an Bi, Renato Scialis, Sumathy Mathialagan, and Manthena Varma Department of Pharmacokinetics, Dynamics & Metabolism, Pfizer Inc., Groton, CT, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Biliary excretion is an important route for the elimination of drugs and/or their metabolites. It may impact drug exposure and efficacy when predicting human dose in clinic, therefore, there is a need to establish the reliable estimation of the biliary clearance in the early drug discovery/development process. The aim of this study was to evaluate the prediction of human biliary clearance from in vitro hepatocyte tool, sandwich-cultured human hepatocytes (SCHH). Human biliary clearance data were collected from the literature for 12 compounds. The hepatic uptake and disposition of selected drugs were assessed using SCHH and the transporter affinity was conducted using transfected-Human Embryonic Kidney 293 cells. No relationship was observed between % dose excretion in human and % biliary excretion index in SCHH. However, a good in vitro in vivo correlation (IVIVC) was observed between human biliary clearance and predicted biliary clearance from SCHH, when corrected for free concentration in cells. Relationship was relatively good for non-OATP substrates. On the other hand, a few OATP substrates (methotrexate, pravastatin and rosuvastatin) were under predicted, although other OATP substrates (valsartan and mezlocillin) were well predicted. In conclusion, the results of the present study suggest that SCHH model would provide good estimate of human biliary clearance and can be used for pharmacokinetic predictions.

P236. EVALUATION OF METABOLITE-INDUCED HEPATOTOXICITY BY A NEWLY ESTABLISHED PRIMARY HUMAN HEPATOCYTE CULTURE SYSTEM WITH GAS-PERMEABLE LUMOX PLATES Kazunobu Aoyama1, Akihiko Sumita2, Naomi Kamiguchi3, Teruaki Okuda3, Hideki Hirabayashi1, Toshiya Moriwaki1, and Junzo Takahashi1 1 Drug Metabolism and Pharmacokinetics Res. Labs., PRD, Takeda Pharmaceutical Company Limited, Fujisawa, Japan, 2 Analytical Development Labs., CMC Center, Takeda Pharmaceutical Company Limited, Fujisawa, Japan, and 3Drug Safety Res. Labs., PRD, Takeda Pharmaceutical Company Limited, Fujisawa, Japan Generally, the activities of drug metabolizing enzymes (DMEs) in human hepatocytes rapidly decrease after seeding on culture plates, which often makes it difficult to evaluate the long-term metabolic profiles of the drugs with in vitro systems. One major reason for this might be an insufficient supply of oxygen to the hepatocytes in vitro. Oxygen is adequately supplied to the liver in vivo by mixture of arterial and venous blood, while an oxygen supply to the in vitro cultured hepatocytes is limited by a low concentration of oxygen in a culture medium (Nahmias et al., 2006). We have established a new primary human hepatocyte culture system with sustainably high DME activities using lumoxÕ (lumox) plates, which can increase oxygen supply to the hepatocytes through a gas-permeable fluorocarbon film bottom. We investigated the effects of lumox plates on the enzyme activities in cryopreserved human hepatocytes. These activities after three-day culture on lumox plates were 2.4- to 41.8-times higher than those of the hepatocytes on conventionally used polystyrene plates. In the current study, we tested the feasibility of the established hepatocyte culture system for evaluation of metabolite-induced cytotoxicity. Furosemide, reported to induce cytotoxicity probably by its reactive metabolites, was used as a model compound for metabolite-induced cytotoxicity. As a result, in the presence of L-buthionine-sulfoximine, a glutathione depletor, furosemide exhibited cytotoxicity to the hepatocytes on the lumox plates, but not on polystyrene plates. In addition, the cytotoxicity by furosemide was prevented with 1aminobenzotriazole, a multi-CYP inhibitor, added to the hepatocytes on the lumox plates. These results suggested that cytotoxicity by furosemide was induced viaformation of oxidative metabolites, not acyl glucuronide. In conclusion, we have successfully demonstrated the applicability of the established primary human hepatocyte culture system with lumox plates to evaluate metabolite-induced cytotoxicity and identify cytotoxic metabolites of new chemical entities in the drug discovery and development stage.

Reference Nahmias Y, Kramvis Y, Barbe L, et al. (2006). A novel formulation of oxygen-carrying matrix enhances liver-specific function of cultured hepatocytes. FASEB J 20:2531–2533.

P237. A NOVEL GENERATION OF HUMAN PLURIPOTENT STEM CELL-DERIVED HEPATOCYTES WITH SUBSTANTIALLY IMPROVED FUNCTIONALITY AND SEVERAL ADULT CHARACTERISTICS Barbara Ku¨ppers-Munther1, Annika Asplund2, Mariska van Giezen1, Nidal Ghosheh2, and Peter Sartipy1 1 Cellartis AB / Cellectis AB, Go¨teborg, Sweden and 2Systems Biology Research Center, School of Bioscience, University of Sko¨vde, Sko¨vde, Sweden Hepatocytes derived from human pluripotent stem cells (hPSC) have the potential to serve as predictive human in vitro model systems for drug discovery and toxicity testing provided that they display biologically relevant hepatic functions. However, up to

150

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

now, the functionality of hPSC-derived hepatocytes has been insufficient for applications requiring high expression of multiple drug metabolizing enzymes. We have developed a novel, highly robust 2D differentiation protocol in order to obtain more functional hepatocytes from human induced pluripotent stem cells (hiPSC). The resulting hiPSC-derived hepatocytes, Enhanced hiPS-HEP, show substantial CYP1A, 2C9, 2C19, 2D6, 2E1, and 3A4 enzyme activities which are required for toxicity testing and drug metabolism studies. Importantly, the Enhanced hiPS-HEP also have several adult hepatic characteristics, such as low expression of fetal genes (e.g. CYP3A7 and alpha-Fetoprotein), and high expression of adult genes (e.g. CYP2C9, 2C19, 2E1, and 3A4). Importantly, we can generate hepatocytes from multiple hPSC lines using the same differentiation protocol. Interestingly, Enhanced hiPS-HEP generated from different hPSC lines show different CYP activity profiles reflecting interindividual variations present in the population, which is typically observed when using human primary hepatocytes from different donors. The key achievements with the present study are (1) a substantially improved metabolic activity of hPSCderived hepatocytes and (2) the robustness of the differentiation protocol across many hPSC lines. These developments now provide a platform for an inexhaustible source of functional human hepatocytes suitable for toxicity testing and drug metabolism studies with an opportunity to also select the genetic background of the cells.

P238. EFFECTS OF COLD STORAGE ON THE RESPONSE OF PRIMARY CULTURED HUMAN HEPATOCYTES TO MODEL P450 INDUCERS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kirsten Amaral, Qian Yang, Utkarsh Doshi, Nick Ring, and Albert P. Li In vitro ADMET Laboratories LLC, Columbia, MD, USA The development of cryopreserved human hepatocytes that can be cultured (plateable) allows P450 studies to be performed, using pre-characterized hepatocytes–cryopreserved hepatocyte lots that have been thoroughly tested to be responsive to model inducers. Conditions of the cultured hepatocytes, especially cell density and confluency, are critical to their responsiveness towards P450 induction. One challenge in the performance of P450 induction studies is the lack of expertise of the investigators in the culturing of the human hepatocytes. To overcome this challenge, we investigated procedures for the shipment of human hepatocyte cultures from one laboratory to another laboratory for P450 induction studies. Our investigation involved the following: 1. Development of transportation conditions to maintain the health of the hepatocytes during transit. 2. Evaluation of the morphology and cellular ATP content of the transported hepatocytes. 3. Evaluation of the responsiveness of the recovered hepatocytes to model P450 inducers. Cryopreserved human hepatocytes from 5 donors were cultured in 96-well collagen-coated plates for 24 hrs. Replicate plates of hepatocytes from each donor were prepared. Set 1: Control cultures: These hepatocytes were kept in a cell culture incubator for the duration of the experiment with daily medium change. Set 2: Shipment cultures: These cultures were designated for shipment. On the day after plating, medium of the shipment cultures was changed to Transport Medium and maintained at 4 deg. C, with replicate plates recovered daily (medium changed to culture medium and returned to the cell culture incubator). Control and shipment cultures were monitored for morphology and cellular ATP contents for up to 5 days during which the control cultures were cultured for 5 days, and the shipment cultures were cultured for 1 day, followed by storage in the cold Transport Medium for 1, 2, or 3 days. The hepatocytes were found to remain confluent with normal cell morphology in both sets of plates during the 5 day period. ATP contents of the hepatocytes were similar between the control and shipment sets, with no apparent decrease with time in culture. The results suggest that cultured cryopreserved human hepatocytes could be maintained for 3 days in the Transport Medium at 4 deg. C. Human hepatocytes cultured from 5 donors were then shipped from one of our laboratories (in Malden, MA) to another (in Columbia, MD) with a shipment duration of approximately 24 hrs. Upon receipt, medium was changed and the plates were returned to a cell culture incubator. On the next day, the hepatocytes were treated for 3 days with model isoform-selective P450 inducers omeprazole (CYP1A), phenobarbital (CYP2B6), and rifampin (CYP3A4). Induction was evaluated based on gene expression using RT-PCR. Dose dependent and significant induction of the isoforms was observed for the shipped hepatocytes from all 5 donors. Our results show that cryopreserved human hepatocytes can be cultured in one laboratory with expertise in cell culture, and be transported to another laboratory where the cells can be subjected to experimentation such as P450 induction studies.

P239. CORNINGÕ HEPATOCELL-CELLS CLOSELY MODEL THE BEHAVIOR OF PARENTAL CELLS FOR PREDICTING HEPATOTOXICITY Ronald A. Faris1, Yulong Hong1, Jin Liu1, Ronjun Zuo2, Feng Li2, Mercyanne Andes2, and Christopher J. Patten2 1 Department of Science and Technology, Corning Incorporated, Corning, NY, USA and 2Corning Life Sciences, Corning Incorporated, Bedford, MA, USA Unforeseen hepatotoxicity is one of the primary reasons for aftermarket black box warnings and removal of drugs. As a result the FDA recently recommended the development of novel tools and cell-based assays to more accurately identify drug induced liver injury. Primary human hepatocyte (PHH) culture is presently the gold standard model for studying the ADMET properties of drug candidates. However, due to large lot-to-lot variations and limited lot size of primary hepatocytes, significant time and resources are spent on prequalifying primary hepatocytes. Donor cells should ideally express the complement of Phase I and II drug metabolism enzymes and drug transporters which impact hepatotoxicity. The availability of renewable sources of human

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

151

hepatocytes that closely model the behavior of parental hepatocytes in primary culture will enable the long-term use of these cell models for comparative pharmacogenomics toxicity studies. In the present hepatotoxicity studies, we compare the performance of CorningÕ HepatoCells, a cell-line derived from primary human hepatocytes, to the parental hepatocyte cells and HepaRGTM. Cells were plated on collagen I coated dishes and cultured in recommended media for 24 hours and then pretreated with a panel of 80 compounds that contained known non-toxic and toxic compounds for 24 hours. Cell viability was measured using the CellTiterÔ GLO assay (Promega). Correlative analysis revealed that the toxicity profile of renewable hepatocytes more closely modeled the behavior of parental primary hepatocytes (R2 = 0.7) than HepaRG (R2 = 0.5). Of note, 7/80 compounds were toxic to both renewable and parental hepatocytes but not to HepaRGTM. In contrast 3/80 compounds were only cytotoxic to HepaRGTM. These results suggest that CorningÕ HepatoCells closely model the functional behavior of parental cells and therefore may be used as a primary hepatocyte surrogate to study pharmacogenomic-related cytotoxicity.

P240. MULTIPLE DONOR POOLED PLATEABLE CRYOPRESERVED HUMAN HEPATOCYTES FOR P450 INDUCTION STUDIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Qian Yang, Utkarsh Doshi, and Albert P. Li In vitro ADMET Laboratories LLC, Columbia, MD, USA We are the first to report here that re-cryopreservation of plateable cryopreserved human hepatocytes would retain their ability to be cultured, thereby allowing the preparation of pooled plateable cryopreserved human hepatocytes. A recent development in the application of human hepatocytes is to minimize individual differences in drug metabolism via the application of cryopreserved hepatocytes pooled from multiple donors (pooled human hepatocytes). The preparation of the pooled human hepatocytes involves the thawing of hepatocytes prepared and cryopreserved from individual donors, mixing (pooling) of the thawed hepatocytes, and re-cryopreservation of the pooled hepatocytes. Pooled human hepatocytes are generally used as suspension cultures for relative short-term studies including metabolic stability, metabolite profiling, and P450 inhibition. We recently developed a novel technique (QuickRefreezeÔ, patent pending) for re-cryopreservation of human hepatocytes. With this technique, cryopreserved human hepatocytes can be thawed and cryopreserved a second time (re-cryopreserved) with only minimum damage to the cells. We report here our results with re-cryopreservation of plateable cryopreserved human hepatocytes. Five individual lots of plateable cryopreserved human hepatocytes were thawed and cryopreserved either individually or as a pool of the five donors. The re-cryopreserved hepatocytes were found to remain plateable, albeit with slightly reduced plating efficiency and confluency of the cultures when compared to the cells before the second cryopreservation. P450 induction studies were performed with isoform-selective model inducers omeprazole (CYP1A2), phenobarbital (CYP2B6), and rifampin (CYP3A4) with both individually and pooled re-cryopreserved hepatocytes. The hepatocytes were thawed and recovered using Universal Cryopreservation Recovery Medium (UCRM), with viability (based on trypan blue exclusion) found to be approximately 90%. The hepatocytes were cultured for 2 days on collagen-coated 96-well plates, followed by treatment with the model inducers for 3 days. P450 induction was evaluated using mRNA as endpoint, with gene expression quantified by RT-PCR. Dose dependent induction of all three isoforms by the model inducers was observed for the re-cryopreserved hepatocytes from the five individual donors and the pooled re-cryopreserved hepatocytes. The observed fold induction values (OI) in the pooled hepatocytes were similar to the mathematic average (MI) of that for the individual donors: The OI/MI values are 27.7/26.1 for CYP1A2, 7.6/7.7 for CYP2B6, and 10.2/13.8 for CYP3A4. The pooled plateable human hepatocytes therefore represent a useful experimental system for P450 induction evaluation. Additional applications of this novel experimental system are being evaluated in our laboratory including in vitro hepatotoxicity, long-term incubation for the evaluation of slowlymetabolized compounds, time-dependent P450 inhibition, and uptake and efflux transporter studies.

P241. IN VITRO EVALUATION OF HUMAN HEPATOCYTES ISOLATED FROM CHIMERIC MICE WITH HUMANIZED LIVERS (PXB-MICEÕ ) TRANSPLANTED USING CELLS FROM THREE DIFFERENT DONORS Chihiro Yamasaki, Ami Yanagi, Yasumi Yoshizane, Yutaka Kageyama, Yumiko Iwasaki, Yuji Ishida, and Chise Tateno PhoenixBio Co. Ltd., Hiroshima, Japan Background: Fresh human (h)-hepatocytes are regarded as the best in vitro model for xenobiotic metabolism and cytotoxicity studies. However, an ‘‘on demand’’ supply of fresh h-hepatocytes from the same donor is not possible. Chimeric mice with humanized livers (PXB-miceÕ ) are considered a useful animal model for predicting h-type drug metabolism and toxicity, and a good tool for supplying fresh h-hepatocytes on demand. In a previous study, we showed that during culture the expression pattern of several h-CYP and UGT mRNAs, the level of h-albumin (h-alb) secretion, and the activity of CYP3A4 in h-hepatocytes isolated from the PXB mice (PXB-cells), were similar to those found in commercially available cryopreserved h-hepatocytes. Purpose: Here, we evaluated the differences in individual hepatic functions such as h-alb secretion, hCYP and transporter mRNA expression, CYP3A4 activity levels, and CYP induction abilities, between PXB-cells isolated from Chimeric mice transplanted with three different donor hepatocytes. Methods and results: PXB mice were produced by transplanting commercially available cryopreserved h-hepatocytes (2-year-old Hispanic girl, 5-year-old African American boy and 2-year-old Caucasian girl) into cDNA-urokinase-type plasminogen activator-transgenic/SCID mice. The PXB-cells were isolated from three PXB mice per donor by using the collagenase perfusion method. Isolated PXB-cells were cultured on type I

152

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

collagen-coated dish for three weeks, and their hepatic functions were analyzed. Irrespective of the plateability of the original donor cells, all the PXB-cells were plateable and cultured at least for 3 weeks. H-albumin levels in the culture medium were maintained throughout the culture period. Expression levels of several hCYPs and transporters in the cultured PXB-cells were comparable to those in freshly isolated PXB-cells. There were no significant differences in the CYP3A4 activities between the freshly isolated PXB-cells and those cultured for three weeks. CYP1A1, 1A2 and 3A4 induction abilities were also maintained for three weeks in vitro. The PXB-cells derived from different donors retained the individuality of hepatic functions between the donors and the inter-donor deviations between the PXB-cells were small. Conclusion: The chimeric mouse with humanized liver may be a useful tool for supplying fresh h-hepatocytes from the same donor on demand. The isolated ‘‘PXB-cells’’ can maintain several hepatic functions on the culture plate, and should be a valuable tool for in vitro metabolic and pharmacotoxicological studies.

P242. VALIDATION OF A 3-DIMENSIONAL METABOLICALLY COMPETENT LIVER MICROTISSUE MODEL FOR LONG-TERM CYTOTOXICITY STUDIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Timmothy Moeller1,2 1 BioreclamationIVT, Halethorpe, MD, USA and 2Brad Larson, Applications, BioTek Instruments, Winooski, VT, USA Non-steroidal anti-inflammatory drugs (NSAIDs) are a class of drug commonly used as analgesics and antipyretics, as well as for management of rheumatological disorders. They are one of the most highly prescribed drug families around the world, and consequently, along with antimicrobial agents, are the most frequent causes of drug-induced liver injury (DILI) (Bjornsson et al., 2010). Diclofenac, in particular, frequently used to treat chronic pain and inflammatory disorders, elicited an FDA warning in 2009 concerning potential hepatic effects. Multiple in vitro animal models have been used to determine the varied mechanisms of action (MOA) of NSAID-related hepatotoxicity. Studies using rat liver mitochondria and freshly isolated rat hepatocytes demonstrated that diphenylamine, a common NSAID structure, uncoupled oxidative phosphorylation, and decreased hepatic ATP content (Masubuchi et al., 2000). Mitochondrial permeability transition (MPT) has also been shown to be important in diclofenac-induced liver injury, as well as the role that oxidative stress plays in MPT induction (Gomez-Lechon et al., 2003). Finally, according to work done by Schmitz et al. (1992), CYP-related metabolic activation of the drug, and the formation of reactive metabolites is also related to diclofenac hepatotoxicity. The combined hepatotoxic effects of these mechanisms usually occurs within weeks of therapy commencement. This emphasizes the fact that in vitro safety testing should incorporate dosing experiments that analyze the effects of a potential drug for periods reaching out to 7–14 days or longer. However, primary hepatocytes cultured in a two-dimensional (2D) manner on the bottom of a microplate, standard practice for in vitro liver toxicity testing, have been shown to undergo rapid loss of differentiated function and metabolic capacity (Cheng et al., 2008), and have less complex inter-cellular and cell-matrix interactions compared to in vivo. What is required is a cell culture model where cells retain viability and function long-term, by creating a favorable environment for these interactions to reform. Here we demonstrate the incorporation of a cell culture system to create human liver microtissues that retain hepatocyte function such as metabolism that is suitable for use in long-term toxicity studies. Hepatocytes are dispensed into a hanging drop and re-aggregate into a functional microtissue, demonstrating in vivo-like cell-cell and cell-matrix interactions and retained viability over weeks. A panel of assays was run to assess cell health and the different MOA exhibited by diclofenac using the 3D liver microtissues, in addition to primary hepatocytes cultured in 2D, and extended dosing periods of one to two weeks. Quantification of the various luminescent and fluorescent emissions via microplate reading and imaging was carried out by a novel cell imaging multi-mode reader. Results confirm the validity of the in vitromicrotissue model to retain metabolic capacity that may be required to predict in vivo hepatotoxic effects of lead molecules.

P243. VALIDATION OF A NOVEL 3D CELL CULTURE SYSTEM TO PERFORM IN VITRO FUNCTIONAL AND TOXICITY ANALYSES USING PRIMARY HEPATOCYTES Timmothy Moeller1 and Brad Larson2 1 Bioreclamation IVT, Halethorpe, MD, Grant Cameron, TAP Biosystems, Royston, UK and 2Applications, BioTek Instruments, Winooski, VT, USA Hepatocytes are the primary cell type of the liver providing the majority of the detoxification which increases the potential for cellular dysfunction and death. Though the source of the insult may be caused by several factors, exposure to drugs represents a significant concern warranting FDA guidance on drug–drug interactions (DDI) and drug-induced liver injury (DILI). In vivo studies are still the gold standard; however, in vitro screening has gained importance for reducing animal exposure, amendable to high-throughput platforms and better equipped to study cellular mechanisms of action. Typically in vitro screening has incorporated primary hepatocytes cultured in a two dimensional (2D) format where the cells form a monolayer across the bottom of a well. However, when cultured and studied in this fashion they rapidly lose their key functions and de-differentiate over the course of only a few days. The ability to culture, characterize and challenge primary cells in a biomimetic 3D environment enables longer term studies. Here we present data demonstrating the differences in response between human primary cells cultured in 2D and in the RAFT 3D cell culture system which has the benefits of a collagen hydrogel with tissue-like properties. General function was assessed measuring ATP and CYP3A4 activity using solution-based assays, and cytotoxicity markers using

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

153

imaging system. Camptothecin and pyocyanin were tested for their ability to cause short-term oxidative stress, as well as longterm induction of apoptosis and necrosis. Kinetic live cell imaging was performed using multiple fluorescent non-perturbing probes to monitor the various effects in real time with incubations up to 48 hours which enabled a thorough assessment of the toxin profile. Image overlay allowed for discrete cellular analysis. Variations in cytotoxicity levels were observed after a 24-hour 800 nM camptothecin treatment (75%:2D; 28%:3D). ROS induction was less in 3D than 2D system. Overall, cells exhibited greater viability using the RAFT system, and were less sensitive to toxins than observed in traditional 2D culture for the endpoints which may indicate a more robust cell culture system.

P244. FUNCTIONAL ASSESSMENT OF LIVER MITOCHONDRIA FROM LPS-PRETREATED EXPERIMENTAL DILI MODEL

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Koichi Arakawa, Shuichi Sekine and Kousei Ito Laboratory of Biopharmaceutics, Graduate School of Pharmazeutical Sciences, Chiba University, Chiba, Japan Purpose: Drug induced liver injury (DILI) is hard to be reproduced by drug administration alone in experimental animals, while pretreatment with non-toxic small amount of lipopolysaccharide (LPS) dramatically sensitizes them to DILI. LPS is known to exert variety of biological effects such as induction of inflammatory cytokines, blood coagulation, oxidative stress, depending on the dosage. Among them, blood coagulation is reported to involve in rats DILI model (Zou et al., 2009). However, it has not been demonstrated what is happen inside hepatocytes particularly in mitochondria, because it is one of the key organelles regulating cell death. In this study, we examined the effect of non-toxic small amount of LPS on the hepatic mitochondrial function to discuss about how it sensitizes the animals to DILI. Method: Rats were fasted for 24 hours, then administered with LPS (1 mg/kg, i.p.). Two hours later, diclofenac (DCF, 100 mg/kg, i.p.) was administered, and serum ALT was measured at 12 hours after DCF administration. In a separate experiment, liver mitochondria were isolated at 2 hours after LPS administration with or without heparin (3000 units/kg, s.c.) to measure ATP, GSH, respiration rate, and ATP production ability. Result and discussion: ALT was significantly increased when LPS was pretreated before DCF. Respiration rate and ATP production ability were significantly increased in isolated mitochondria from LPS treated rat liver, but they returned to control levels by co-administration with heparin. In spite of the increased respiration rate in isolated mitochondria, amount of ATP in both mitochondria and in liver homogenate were significantly decreased after LPS treatment. Mitochondrial GSH was also significantly decreased after LPS treatment. Collectively, it was implicated that non-toxic amount of LPS might induce blood coagulation inside liver circulation as reported previously and resulted in hypoxia. Under such hypoxic condition, reactive oxygen species expended hepatic GSH, and ATP was decreased in spite of feedback stimulation of the respiratory function. Drugs with potential mitochondrial toxicity, such as DCF, will give final blow to induce cell death. Conclusion: Coagulation-dependent oxidative stress and compromised ATP production in mitochondria may be the underlying mechanisms which sensitize hepatocytes to DILI after non-toxic small amount of LPS treatment. Reference Zou et al. (2009). Toxicol Sci 108:184193.

P245. IN VITRO ASSESSMENT OF DRUG-INDUCED CHOLESTASIS BY ALTERATION OF BILE ACID HOMEOSTASIS IN SANDWICH-CULTURED HUMAN HEPATOCYTES: MIP-DILIs SET OF TRAINING COMPOUNDS Marlies Oorts1, Lysiane Richert2, and Pieter Annaert3 1 Drug Delivery & Disposition, KU Leuven, Leuven, Belgium, 2KaLy-Cell, Strasbourg, France, and Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium

3

Department of

Drug-induced liver injury (DILI) is known to be a leading cause of drug candidate failures during clinical trials or even postmarketing withdrawal. Symptoms of cholestasis occur in about 50% of the DILI cases reported in literature (Oorts et al., 2014). Recently, our research group successfully established an in vitro model, based on sandwich-cultured human hepatocytes (SCHH), for identification of compounds with a liability to cause cholestasis by altering bile acid (BA) homeostasis in human2. In the present study, this in vitro model was applied to investigate the potential of 14 training compounds to exert cholestasis. These training compounds were chosen, based on clinical data, by the MIP-DILI consortium. Day-5 SCHH (n = 1 donor) were exposed to 3 concentrations of each test compound (TC), in presence or absence of a concentrated mixture of BAs for 24/48 h. Urea assay was used to assess hepatocyte functionality. The cholestatic potential of the TCs was expressed by the drug-induced cholestasis index (DICI) (Chatterjee et al., 2014). It was observed that troglitazone and nefazodone, which exert cholestasis in the clinic, showed enhanced toxicity in presence with BAs, yielding DICI values 0.8. However, bosentan, which is known to cause cholestasis in vivo, did not decrease the urea formation in presence of BAs. After 24 of incubation, buspirone (50 mM) showed a significant increase in toxicity in presence with BAs, while no toxicity was observed when the TC was incubated alone. This was a remarkable observation, since buspirone does not inhibit the bile salt export pump, which represents the most wellknown mechanism for cholestasis (Kostrubsky et al., 2006). Tolcapone and entacapone yielded safety margins (SM) of 1 and 30, respectively, meaning that both compounds are flagged for in vivo cholestasis risk. Other hepatotoxic compounds, i.e.

154

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

paracetamol, pioglitazone, diclofenac and amiodarone did not show any decrease in urea formation at their highest concentrations (1000 mM, 300 mM, 500 mM and 25 mM, respectively). On the other hand, fialuridine (50 mM), metformin (150 mM) and perhexiline (65 mM) did show significant signs of toxicity. Interestingly, DICI values  0.8 were observed for ximelagatran at low concentrations, while this was not the case with higher concentrations. In conclusion, our results confirm the utility of this in vitro model to detect the cholestatic potential of compounds. Ongoing mechanistic research addresses the findings for bosentan, buspirone, tolcapone, entacapone and ximelagatran.

References Chatterjee S, Richert L, Augustijns P, Annaert P. (2014). Hepatocyte-based in vitro model for assessment of drug-induced cholestasis. Toxicol Appl Pharmacol 274:124–136. Kostrubsky SE, Strom SC, Kalgutkar AS, et al. (2006). Inhibition of hepatobiliary transport as a predictive method for clinical hepatotoxicity of nefazodone. Toxicol Sci 90:451–459. Oorts M, Richert L, Annaert P. (2014). In vitro drug-induced cholestasis detection with freshly-isolated and cryopreserved rat hepatocytes. Manuscript in preparation.

This work was supported by the EU-EFPIA Innovative Medicine Initiative (IMI) project #115336-2 ‘‘Mechanism-Based Integrated Systems for the Prediction of Drug-Induced Liver Injury MIP-DILI’’

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P246. TROGLITAZONE AND ITS SULFATE CONJUGATE CAUSE DYSFUNCTION OF THE BILE CANALICULAR NETWORK FORMATION Akinori Takemura, Aya Izaki, Shuichi Sekine, and Kousei Ito The Laboratory of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan Background and aims: Idiosyncratic drug induced liver injury (DILI) is one of the major reasons for withdrawal from the market. Troglitazone (Tro), an antidiabetic drug and classified with thiazolidinedione, has been withdrawn from Japanese market in 2000 because of idiosyncratic DILI. Although several mechanisms (eg. inhibition of BSEP, mitochondrial toxicity and production of reactive metabolites) have been suggested for Tro induced idiosyncratic DILI, they could not be reproduced experimentally at therapeutic range (Cmax: 6.6 mM). We are particularly interested in the formation of bile canalicular network during the recovery process. Because the structure is essential to exert liver detoxification function, inhibition of its formation process possibly leads to delay of recovery or even deterioration of liver injury. However, little is known about the effect of Tro on canalicular network formation. Here we examined the effect of Tro and other cholestatic drugs on this process in sandwich cultured rat hepatocytes (SCRHs). Method: SCRHs were treated with the drugs for 24 hours, then Mrp2 and F-actin were stained to visualize bile canalicular network. Fluorescence images were taken using confocal laser scanning microscopy and bile canalicular network formation was quantified as the length of positively stained networks/cell count in five fields. Medium and cell were collected and amounts of Tro and its sulfated conjugate (Tro-sul) were determined by LC-MS/MS. Result and conclusion: The formation of canalicular network was inhibited by the treatment with Tro, haloperidol, ibuprofen, chlorpromazine, imipramine and lithocholic acid (positive control). Inhibition was the most potent with Tro (IC50 value of 0.6 mM), while other thiazolidinedione including pioglitazone and rosiglitazone inhibited only weakly. During the incubation, approximately 60% of Tro was metabolized to Tro-sul until 24 hours. Moreover Tro-sul itself inhibited bile canalicular network formation in a concentration dependent manner (IC50 value for 0.27 mM). Considering the close proximity of IC50 values of Tro and Tro-sul to the therapeutic range, inhibition of bile canalicular network formation is possible under clinical situation.

P247. INVESTIGATING DRUG–INFLAMMATION INTERACTION OF ACETAMINOPHEN IN HEPATOCYTES AND KUPFFER CELLS CO-CULTURE SYSTEM FOR IN VITRO DRUG SCREENING APPLICATION Yu Yu1, Nisha Hari Singh2, Rashidah Binte Sakban1, Lei Xia1, and Hanry Yu1 1 Department of Physiology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore and 2 InvitroCue, Singapore, Singapore Background: Idiosyncratic drug induced liver injury (IDILI) is a leading cause for drug withdrawal and post-marketing drug restrictions by FDA. Limited knowledge has been obtained for the screening of IDILI due to its undefined mechanisms. Currently there is lack of sufficient in vitro biomarkers to predict IDILI. Primary hepatocyte and Kupffer cell co-culture shows its potential to study the IDILI based on the inflammatory stress hypothesis in the cellular and molecular levels. Aim and hypothesis: In this project, we propose to apply rat hepatocytes and Kupffer cells co-culture to investigate and establish key biomarkers with in vivo correlations for idiosyncratic hepatotoxins screening, using acetaminophen (APAP) as a model drug. We hypothesize that the co-culture in the inflammatory condition would provide a more sensitive platform to detect the cell injury induced by APAP via regulation of Kupffer cells and alteration of hepatocyte intracellular events. Methods: Lipopolysaccharides (LPS), a bacterial endotoxin was applied to induce the inflammatory stress at an optimized concentration (100 ng/ml). Proinflammatory cytokines production was assessed by ELISA kits. APAP induced cytotoxicity was evaluated by MTS assay. And LC/MS/MS was used to study the activity of cytochrome P450 (CYPs) which help metabolize APAP and produce reactive metabolites. Function of glutathione S-transferases (GSTs), critical enzymes in the detoxification process, was investigated via a

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

155

respective kit from Caymen Chemical. Results: APAP induced a minor and severe cell injury at the concentrations of 1 mM and 15 mM, respectively, in the co-culture with LPS treatment, while no cytotoxicity was observed in the hepatocyte pure culture or co-culture without LPS treatment. Production of pro-inflammatory cytokines such as IL-1b, IL-6 and TNF-a was found in the presence of Kupffer cells upon LPS treatment, indicating the inflammatory condition. Results of CYP activities revealed that with LPS treatment, the enzyme activities were significantly repressed, similar as what has been observed from in vivo situations. Treatment of APAP helped improve the CYP activities level, but not as high as that without LPS treatment. Treatment of LPS in the co-culture also resulted in the GST activities decrease. However, no functional improvement of GSTs was observed with APAP treatment, indicating that the cell injury may be resulted from the impaired detoxification ability. Conclusions: This study illustrates that the presence of Kupffer cells in the inflammatory condition would change the enzymatic levels in hepatocytes and may lead to the final cell injury induced by APAP at otherwise safe doses, suggesting the potential usage of rat hepatocytes and Kupffer cells co-culture as a sensitive in vitro screening model for drug-inflammation interaction study.

P248. HEPARG CLONE 5 F – A HIGHLY ACTIVE HUMAN LIVER CELL LINE FOR DRUG METABOLISM, TRANSPORTER AND TOXICOLOGY STUDIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Tim Brayman, Yongling Xiao, Maureen Bourner, and David Thompson Sigma-Aldrich, Saint Louis, MO HepaRG cells have been increasingly used over the past several years for a variety of human liver assay applications (e.g. drug metabolism and clearance, cytochrome p450 [CYP] induction, bioreactors and toxicology studies) due primarily to their enhanced drug metabolism activities compared with other hepatoma-derived cell lines. Compared with primary human hepatocytes, HepaRG offer a more convenient source of hepatocytes, lack of donor variability and comparable CYP3A4 activity. In the course of developing a series of drug metabolism and transporter knockout cell lines in HepaRG using zinc finger nuclease (ZFN) technology, we isolated a clone (5F) for use as a control cell line. 5F was transfected, single cell sorted and clonally expanded following the identical procedure used for all knockout cell lines. Compared with low passage wildtype HepaRG cells, 5F demonstrated several unexpected properties, including exceptional growth rate (42-fold faster) and differentiation, more consistent and robust response to induction of CYP2B6 and CYP3A4 using CITCO and rifampin, respectively, and higher levels of basal activity of CYP2D6. Enhanced transporter function was observed as the rate of the OATP1B1/1B3 substrate estradiol 17-b glucuronide uptake was increased up to 2-fold compared to wild type cells. In addition, 5F maintained these properties long past the passage number when wildtype HepaRG cells begin to lose some of their key functional properties (p20). Active cultures of 5F were observed well past passage 35. In summary, we have developed a proprietary ZFN genetically engineered HepaRG cell line, clone 5F, with improved growth characteristics, differentiation, and enhanced functionality. Clone 5F should be useful as a control with HepaRG knockout cell lines as well as on its own as an alternative, highly active human hepatocyte cell line.

P249. COMPARATIVE ANALYSIS OF HUMAN FETAL AND ADULT HEPATOCYTES BY METABOLOMICS AND GENOMICS Seiichi Ishida1, Su-Ryang Kim1, Takashi Kubo1, Yukie Kuroda1, Takaaki Ishii2, Maki Hojyo1, Atsuko Miyajima3, Taku Matsushita2, and Yuko Sekino1 1 Divison of Pharmacology, National Institute of Health Sciences, Tokyo, Japan, 2Depatment of Applied Life Science, Sojo University, Kumamoto, Japan, and 3Divison of Medical Devices, National Institute of Health Sciences, Tokyo, Japan Purpose: Liver is the central organ of the metabolism in human body. Fetal liver is also the important organ, though its function is distinct from that of adult liver. In this study, the differences of the basal metabolic functions between fetal and adult hepatocytes are analyzed by metabolome analysis. In parallel, genome-wide gene expressions are measured in both cells. The results of metabolome analysis and gene expression information are compared to clarify their functional differences and susceptibility to chemical compounds. Methods: Human fetal hepatocytes (Hc cells, CS-ABI-3716) were prepared from six normal fetal livers (gestation average 16 weeks) by Applied Cell Biology Research Institute (Kirkland, WA, USA). The primary human hepatocytes (HEP220) were obtained from three adult patients undergoing resection for primary or secondary tumors, and isolated by collagenase perfusion of histologically normal liver fragment by Biopredic International (Rennes, France). Metabolites in fetal and adult hepatocytes were prepared by methanol-chloroform-water extraction and comprehensively analyzed by capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS). Total RNAs in fetal and adult hepatocytes were prepared and analyzed by Affymetrix Human U133A gene chip. Fetal and adult hepatocytes were exposed to chemical compounds, including tributyltin, acetaminophen, sodium valproate and perfluorooctanesulfonic acid with several concentrations. The IC50 value of each chemical compound was measured by WST-8 assay. Results and Discussion: We identified 211 metabolites (116 anions and 95 cations) in the hepatocytes by CE-TOFMS. Principal component analysis and hierarchical clustering analysis of the relative amounts of metabolites clearly classified hepatocytes into 2 groups that were consistent with their origin, i.e. the fetus and adult. The amounts of most metabolites in the glycolysis/glyconeogenesis pathway, tricarboxylic acid cycle and urea cycle were lower in fetal hepatocytes than in adult hepatocytes. These results suggest different susceptibility of the fetal and adult liver to toxic insults affecting energy metabolism. Actually, it was showed that three chemical compounds,

156

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

tributyltin, acetaminophen and sodium valproate, were more toxic to the fetal hepatocytes than to the adult hepatocytes. On the contrary, perfluorooctanesulfonic acid was more toxic to the adult hepatocytes. Analyses of genome-wide gene expressions are going to clarify the molecular mechanisms of these differences. The relationships of the basal functions of fetal and adult hepatocytes and their sensitivities to the chemical compounds will be discussed.

P250. TOXICOLOGICAL STUDIES OF HEPATOTOXIC COMPOUNDS IN HUMAN HEPATOMA HEPARG CELLS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yuichi Yokoyama1,2,3, Yoshifumi Sasaki2, Natsuko Terasaki2, Taku Kawataki2, Koji Takekawa2, Yumiko Iwase2, Seigo Sanoh3, and Shigeru Ohta3 1 Safety Research Laboratories, Chiba, Japan, 2Mitsubishi Tanabe Pharma Corporation, Chiba, Japan, and 3Graduate School of Biomedical and Health Sciences, Hiroshima University, Chiba, Japan HepaRG is a newly established human hepatoma cell line, which were obtained from a human hepatocellular carcinoma (Gripon et al., 2002). Differentiated HepaRG cells have been shown to maintain liver-specific functions such as CYPs, Phase II drugmetabolizing enzymes and transporters (Aninat et al., 2006; Guillouzo et al., 2007). In this study, the feasibility of the HepaRG cells as a human hepatocyte model, which express the drug metabolizing enzyme and have the activities, for toxicological studies were examined using some hepatotoxic compounds. Firstly, basal drug metabolizing enzyme activities were measured in HepaRG cells, human hepatocytes and HepG2 cells. Six phase I enzyme activities (CYP1A2, CYP2B6, CYP2C9, CYP2C19, CYP2D6 and CYP3A4) and two phase II enzyme activities (UGT and SULT) were determined in each cell. Determined enzyme activities in differentiated HepaRG cells showed the almost comparable to those in human hepatocytes and much higher than those in HepG2 cells except for SULT activity. Secondly, we have examined the cytotoxicity of some hepatotoxic compounds, such as aflatoxin B1 (AFB1) and cyclophosphamide (CPA) in HepaRG cells and human hepatocytes. In these examinations, AFB1 and CPA-induced cytotoxicities in HepaRG cells were comparable to those observed in human hepatocytes. Furthermore, as the cytotoxicities of these compounds were inhibited by 1-aminobenzotriazole, which is a non-specific CYP inhibitor, in both cells, the cytotoxicity of which is mediated via metabolic activation by CYP was detectable in HepaRG cells as well as human hepatocytes. Lastly, toxicogenomics analysis was performed in HepG2 and HepaRG cells after exposure of AFB1 and CPA. It is well known that cytotoxicity of AFB1 and CPA is related to the cellular DNA damage which is induced by reactive metabolites of these compounds. Therefore, we focused on gene expression profiles of p53 signaling pathway, which is related to the DNA damage response. In HepaRG cells, much number of p53-related genes was up-regulated than those in HepG2 cells. According to these results, gene expression profiles in HepaRG cells were much more reflected to the mechanism of AFB1 and CPA toxicity than those in HepG2 cells. In the light of our investigation, HepaRG cells would useful human hepatic cellular models for toxicity studies.

References Aninat C, Piton A, Glaise D, et al. (2006). Guillouzo, Expression of cytochromes P450, conjugating enzymes and nuclear receptors in human hepatoma HepaRG cells. Drug Metab Dispos 34:75–83. Gripon P, Rumin S, Urban S, et al. (2002). Guguen-Guillouzo, infection of a human hepatoma cell line by hepatitis B virus. PNAS 99:15655–15660. Guillouzo A, Corlu A, Aninat C, et al. (2007). Guguen–Guillouzo, The human hepatoma HepaRG cells: A highly differentiated model for studies of liver metabolism and toxicity of xenobiotics. Chem–Biol Interact 168:66–73.

P251. THE USEFULNESS OF IN VITRO SANDWICH CULTURED HEPATOCYTES ASSAY FOR PREDICTING THE CLINICAL RISK OF DRUG-INDUCED CHOLESTATIC LIVER INJURY Takeshi Susukida1, Shuichi Sekine1, Mayuka Nozaki1, Hitoshi Matsui2, Tomoharu Osada2, Toshiharu Horie3, and Kousei Ito1 1 Laboratory of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan, 2Drug Development Service Segment, LSI Medience Corporation, Ibaraki, Japan, and 3Faculty of Pharmaceutical Sciences, Teikyo Heisei University, Tokyo, Japan Purpose: Drug induced liver injury (DILI) is serious adverse event leading to the dropout of candidate compounds from drug developing and the withdrawal of pharmaceuticals from clinical use. Accumulation of bile acids (BAs) in hepatocyte was suggested as an underlying mechanism of cholestatic DILI. It has been already proved that many drugs which cause cholestatic DILI produce BAs retention in hepatocytes. We have constructed an in vitro BAs-dependent hepatotoxicity assay system using rat sandwich cultured hepatocytes (SCHs) to mimic cholestatic DILI in vivo (Ogimura et al., 2011). However, it has not yet been confirmed if this assay system could predict the cholestatic DILI in clinical situation. In this study, we assayed BAs-dependent hepatotoxicity in several drugs using rat and human SCHs and compared with frequency rate of cholestatic DILI (in vivo data obtained in the clinical studies). [Methods] We selected 30 drugs from PMDA adverse drug reaction report database considering the relative frequency of cholestatic DILI. Absolute frequencies of increase in cholestatic DILI markers (serum alkaline phosphatase (ALP) and g-glutamyltransferase (gGTP)) were calculated based on the clinical data in the Interview Form and determined as the index of in vivo data. Rat and human SCHs were exposed to test drugs with or without serum components of BAs. Drug concentration was set to 50 mM except for Cyclosporine A and Everolimus (10 mM). After 24 hours incubation, we measured LDH release into medium and determined as the index of BAs-dependent

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

157

hepatotoxicity. [Results and Discussion] Almost similar tendency of BAs-dependent hepatotoxicity was observed between in human and rat SCHs, and its correlation was statistically significant (r2 = 0.65). When obtained data of BAs-dependent hepatotoxicity in rat SCHs was applied in Receiver Operating Characteristic (ROC) analysis (medians of absolute frequencies of increase in ALP and gGTP extracted from each compounds (0.22%, 0.45%) were determined as the standard), threshold values of in vitro toxicity were set to 12.2% (ALP) and 8.6% (gGTP). In this condition, the sensitivity and specificity was 75% (9/12) and 78% (11/14) for ALP, and 92% (12/13) and 92% (11/12) for gGTP, respectively. Likewise in human SCHs, threshold values of in vitrotoxicity were set to 7.2% (both ALP and gGTP) and the sensitivity and specificity was 77% (10/13) and 69% (9/13) for ALP, and 85% (11/13) and 73% (8/11) for gGTP, respectively. Therefore, these results suggest that BAsdependent hepatotoxicity observed in both rat and human SCHs correlates well with the frequency of ALP/gGTP increase in patients. [Conclusion] BAs-dependent hepatotoxicity assay using SCHs might be useful preclinical screening tool to predict the clinical risk of cholestatic DILI.

Reference Ogimura E, Sekine S, Horie T. (2011). Bile salt export pump inhibitors are associated with bile acid-dependent drug-induced toxicity in sandwich-cultured hepatocytes. Biochem Biophys Res Commun 416:313–317.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P252. CHARACTERIZATION OF CORNINGÕ HEPATOCELLS FOR DRUG UPTAKE TRANSPORT ASSAY Rongjun Zuo1, Kirsten Cooper1, Mercyanne Andes1, Feng Li1, Sweta Parikh1, Na Li1, Jin Liu2, Ronald A. Faris2, and Christopher J. Patten1 1 Corning Life Sciences, Bedford, MA, USA and 2Departmant of Science and Technology, Corning Incorporated, Corning, NY, USA Increasing evidence shows that drug transporters play an important role in drug-induced hepatotoxicity and adverse drug–drug interactions. Regulatory agencies (USFDA and EMA) recommend in their drug interaction guidelines to investigate clinical relevant transporters for new molecular entities. Current in vitro models including recombinant vesicles, transfected cell lines, suspension or plated cultures (or sandwich culture) of primary hepatocytes, are widely used for efflux or uptake transporter study with various pros and cons. Here we report characterization of CorningÕ HepatoCells (derived from primary human hepatocytes) for uptake transporter study, specifically OCT1, OATP1B1/1B3, and NTCP. When cultured on CorningÕ BioCoatÔ Collagen I coated tissue culture plates for 3–4 days, CorningÕ HepatoCells demonstrated time-dependent and concentration-dependent uptake of prototypical substrates of OCT1, OATP1B1/1B3, and NTCP. Multiple lots of CorningÕ HepatoCells generated similar kinetic profiles which are aligned with published data generated using existing SLC transporter models. Inhibition profiles were also evaluated with several inhibitors and multiple substrates. Gene expression profile of these important uptake transporters supported functional data. In conclusion, the study demonstrated that CorningÕ HepatoCells actively expresses functional uptake transporters, and this model represents a renewable source of human hepatic cells with consistent performance for uptake transporter study.

P253. SIMULTANEOUS EVALUATION OF SIX UDP-GLUCURONOSYLTRANSFERASES ENZYME ACTIVITIES IN HUMAN LIVER MICROSOMES USING AN LC-MS/MS Boram Lee, Jeongmin Joo, Doohyun Lee, Daehun Kim, Kwang-Hyeon Liu, and Taeho Lee College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, South Korea UGT enzymes, one of phase II drug metabolizing enzyme, metabolize the formation of glucuronide conjugates of xenobiotics such as drug and pesticide. The function of these enzymes is essential for the clearance of many xenobiotics, however, alteration of UGT activity is a potential cause of adverse drug–drug interaction. The early detection of potential drug–drug interactions is an important issue of drug discovery that has led to the development of high-throughput screening (HTS) methods for drug interactions. We developed a HTS method for the evaluation of six human liver microsomal UGT enzymes activities using cocktail incubation and tandem mass spectrometry in vitro. The two cocktail doses in vitro were developed to minimize mutual drug interactions among substrates: cocktail A was composed of three substrates inclusing SN-38 for UGT1A1, chenodeoxycholic acid for UGT1A3, and trifluoperazine for UGT1A4; and cocktail B was composed of three substrates including N-acetylserotonin for UGT1A6, mycophenolic acid for UGT1A9, and naloxone for UGT2B7. Present cocktail combination was validated by the comparison of enzyme activities and IC50 values on known UGT inhibitors between cocktail and individual incubation. This HTS method will be useful for the evaluation of drug interaction potential of new chemical entities and mechanistic understanding of UGT-mediated drug interactions.

P254. AN IN VITRO HIGH-THROUGHPUT CYP COCKTAIL INHIBITION ASSAY USING UHPLC-MS-MS Guannan Li, Ke Huang, Dejan Nikolic´, and Richard B. van Breemen Department of Medicinal Chemistry and Pharmacognosy, UIC/NIH Center for Botanical Dietary Supplements Research, Chicago, IL, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

158

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Early detection of possible drug–drug interactions (DDI) is essential during drug discovery and development, and the discovery of possible drug-botanical interactions is important for the safe use of botanical dietary supplements. Since inhibition of cytochrome P450 (CYP) enzymes is the most common cause underlying drug–drug or drug-botanical interactions, in vitro assessment of CYP inhibition potential is important. Here, an efficient and cost-effective mass spectrometry-based in vitro high-throughput CYP cocktail inhibition assay was developed that uses 10 substrates simultaneously against 9 CYP isoforms. Potassium phosphate buffer (100 mL 0.1 M, pH 7.4) containing 13 mM NADPH, 0.2 mg/mL human liver microsomes, and a cocktail of 10 CYP substrates (1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1 and 3A4) was incubated at 37  C for 10 min. The reactions were terminated adding 20 mL of a stop solution (water/acetonitrile/ formic acid, 92:5:3 v/v/v) containing stable isotope-labeled surrogate standards. All metabolites and internal standards were analyzed in a single run using UHPLC-MS-MS with collision-induced dissociation and selected reaction monitoring on a Shimadzu (Kyoto, Japan) Nexera UHPLC and LCMS-8050 triple quadrupole mass spectrometer. The concentrations of CYP substrates were selected based on the Km values of each that were determined in-house. The interaction between 10 probe substrates was evaluated. Concentrations of 2B6 substrate (bupropion) and 2E1 substrate (chlorzoxazone) were lowered to minimize the interference. Incubation time were carefully evaluated and controlled. Minimal amount of organic solvent was introduced. All 10 metabolites of the probe substrates and their corresponding isotope-labeled internal standards were separated on a ACQUITY UPLC BEH C18column (2.1  100 mm, 1.7 mm) using a rapid gradient of only 2.5-min from 10% to 80% acetonitrile in water containing 0.1% formic acid. The flow rate was 0.6 mL/min, and the column oven temperature was 40  C. Unlike some previous cocktail approaches that omitted several possible substrates due to ionization limitations or lack of sensitivity of the analytical method, both positive and negative ion electrospray were carried out with rapid polarity switching (5 milliseconds) on an ultrafast, high-sensitivity triple quadrupole mass spectrometer. The method was validated using known inhibitors of each CYP enzyme and then applied to the analysis of potential drug-botanical interactions involving licorice.

P255. AN EX VIVO EQUILIBRIUM DIALYSIS (ED) ASSAY FOR MEASURING IN VIVO BRAIN AND PLASMA DRUGFREE FRACTION Adarsh S. Gandhi1, Manuel Cajina1, Megan Nattini1, Megan McAleavy2, and Gamini Chandrasena1 1 Department of Bioanalysis and Physiology, Lundbeck Research Inc. USA, Paramus, NJ, USA and 2Neuroinflammation Disease Biology Unit, Lundbeck Research Inc. USA, Paramus, NJ, USA The extent of non-specific brain tissue binding of CNS drug candidates is essential for evaluating their pharmacokinetic and pharmacodynamic profiles. While total drug concentration can be directly measured, it is not a true estimate of the amount available at target site. Studies have shown unbound drug concentration in the brain (Cu,brain) or plasma (Cu,plasma) correlated with pharmacodynamic readouts (target occupancy) in rodent models (Watson et al., 2009). Although total brain drug concentrations can be easily measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS), estimating relevant fraction unbound (fu,brain) is more challenging. Nevertheless, in vitro and in situ methodologies are available to measure fu,plasma, fu,brain or Cu,brain. Intracerebral microdialysis is one of the methods utilized for measuring Cu,brain, but it’s a complex and an invasive method, and not very amenable to higher-throughput. On the other hand, in vitro ED using either plasma or tissue homogenates is a widely used method, and offers a static system for higher-throughput analysis. However, in vitro ED can impart potential errors at execution either in manual or automation mode, including analytical phase. We have developed an ex vivo assay for measuring in vivo fu,plasma and fu,brain and evaluated against in vitro ED assay using three clinically relevant CNS compounds. For in vitro ED analysis, untreated rat brain homogenate solution (50:30:20 Water:Isopropanol:DMSO, pH 7.4) and plasma were incubated with Verapamil (0.44 mM), Buspirone (10 mM) or Nortriptyline (10 mM) for 2.5 h at 37  C in 5% CO2for maintaining pH 7.4 during dialysis against Phosphate Buffered Saline (PBS). For in vivo ED analysis, rats were orally dosed (n = 3) with Buspirone (10 mg/kg), Nortriptyline (10 mg/kg) or Verapamil (5 mg/kg). Blood and brains were collected after 1 h. Plasma and brain homogenates were dialyzed under same conditions described above. The total and free drug concentrations in plasma plus brain homogenates and PBS, respectively were measured by LC-MS/MS. The preliminary data (mean ± SD) showed ex vivo fu,plasma (1 h) for Verapamil, Buspirone and Nortriptyline were (0.013 ± 0.005), (0.27 ± 0.07), and (0.025 ± 0.008), respectively. The fu,brain (1 h) estimates for Verapamil, Buspirone and Nortriptyline were (0.014 ± 0.004), (0.01 ± 0.001), and (0.002 ± 0.0003), respectively. The observed ex vivo estimates were similar to in vitro estimates generated for rat, and were also well correlated with published in vitro data (Maurer et al., 2005). This ex vivo method provides an advantage, enabling simultaneous measurement of all tissue partitioning parameters generated from same animal, allowing better prediction of disposition mechanisms (influx and efflux) at earlier drug design phase. Additional in vivo validation studies (in mice and rats) are under way, including co-administration of Verapamil as a normalization control to account for potential assay to assay variability.

References Maurer TS, Debartolo DB, Tess DA, Scott DO. (2005). Relationship between exposure and nonspecific binding of thirty-three central nervous system drugs in mice. Drug Metab Dispos 33:175–181. Watson J, Wright S, Lucas A, et al. (2009). Receptor occupancy and brain free fraction. Drug Metab Dispos 37:753–760.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

159

P256. MALDI-MS/MS-BASED HIGH-THROUGHPUT BIOMARKER PROTEIN QUANTIFICATION WITH A NEW INTERNAL STANDARD METHOD

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Toshihiro Yoneyama1, Sumio Ohtsuki2, Yasuo Uchida3, Masanori Tachikawa1, and Tetsuya Terasaki1 1 Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan, 2Department of Pharmaceutical Microbiology, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan, and 3Tohoku University, Sendai, Japan Purpose: Quantitative targeted absolute proteomics with liquid chromatography-mass spectrometry (LC-MS/MS) has enabled us to estimate the abundance of biomarker proteins in biological fluids such as plasma, serum and cerebrospinal fluid. However, the long LC separation time and the small numbers of samples to be analyzed within a day sometimes limits highthroughput analysis. Matrix-assisted laser desorption ionization (MALDI)-MS/MS may also be a powerful tool to overcome this limitation except for low signal reproducibility which is caused by uneven ionization efficiency. The purpose of this study was thus to establish MALDI-MS/MS based high-throughput biomarker protein quantification using a newly developed internal standard method. Method: Target peptide sequences to be quantified were selected from fibrinogen protein and synthesized. The sequences of internal standard (IS) peptides were also synthesized as a scrambled sequences of the target peptide. A standard curve was made by MALDI-LTQ-based MS/MS analysis using a mixture of the IS peptide with a fixed amount, a dilution series of target peptide and a-cyano-4-hydroxycinnamic acid as a matrix for peptide ionization. The standard human plasma were digested with trypsin and analyzed for fibrinogen quantification. Result and discussion: The standard curve with the IS peptide produced much better linearity (R240.99) within the range of 50–1000 fmol/microL, as compared to that without IS peptide (R250.7). Coefficient of variation (CV) of over 300 cps signal intensities produced from 5 different serum concentrations was within approximately 20%.The peptide selection criteria to produce high quantification sensitivity was established: especially the peptide sensitivity significantly went down by replacing aromatic amino acid with another non-aromatic amino acid. The relative error of quantitative values analyzed by MALDI-MS/MS and LC-MS/MS was within 20% from 0.5 to 5 microL plasma. These results suggest that (i) the MALDI-MS/MS problem associated with low reproducibility was overcome by using the IS peptide methods, (ii) containing an aromatic amino acid residue is a key criteria to select the peptide sequence to be quantified, and (iii) the quantitative values analyzed by MALDI-MS/MS are as valid as those analyzed by LC-MS/MS. Conclusion: A MALDI-MS/MS-based biomarker protein quantification was established using a newly developed internal standard method.

P257. DEVELOPMENT OF PHARMACOKINETICS EVALUATION SYSTEMS FOR EFFECTIVE FORMULATION DESIGN (II) APPLICATION OF PBPK MODELING FOR NON-LINEAR PHARMACOKINETICS USING GASTROPLUSTM Yuka Haruna, Motoki Onishi, Maki Matsuda, Hidetoshi Hashizume, Kazuhide Imagaki, Takashi Matsui, and Kozo Tagawa Medicinal Analysis Research Department, Research & Development Division, Towa Pharmaceutical Co., Ltd., Kyoto, Japan Background: In vitro traditional dissolution testing does not necessarily reflect drug’s in vivo kinetics due to the complexity of drug disposition. This fact is a major obstacle to formulation design for pharmaceutical bioequivalence. To resolve this problem, we have been studying pharmacokinetics evaluation systems by using pharmacokinetics simulation software and by using an in vitro dissolution system which mimics in vivo conditions, by retrospective or prospective analysis technique. This time, we studied the usability of GastroPlusTM at the early stage of drug formulation design. In addition, we studied the application of PBPK modeling for drugs which exhibit non-linear pharmacokinetics. Methods: We first classified the drugs based on BCS, and then set the basic models of GastroPlusTM for each drug. Next, on the basis of those basic models, we conducted the parameter sensitivity analysis for gastric residence time, gastric pH, degree of solubility, rate of dissolution, membrane permeability and particle size, and further analyzed the influence of food on drug absorption. Then, we constructed PBPK models from in vitro metabolic testing, and simulated PK profiles and conducted the parameter sensitivity analysis for the above factors. Conclusion: The results lead us to a better understanding the impact of drug concentration in gut on first-pass effect. Furthermore, we are studying an in vitro dissolution system to deepen understanding about thein vivo actual concentration of drugs.

P258. MODELING THE REACTIVITY OF DRUG METABOLITES S. Joshua Swamidass1 and Tyler Hughes2 1 Department of Pathology and Immunology, Washington University in Saint Louis, Saint Louis, MO, USA and 2Washington University in Saint Louis, Saint Louis, MO, USA Idiosyncratic drug reactions (IADRs) have profound financial and health implications, and are believed to be caused by reactive metabolites. Using a large dataset of 734 diverse molecules, we built mathematical models of reactivity. In the database, reactive molecules were annotated with the specific atom known to be reactive, which enables our approach to identify which specific substructure in a molecule are reactive. Cross-validation accuracy on this dataset was greater than 88%. This model can score arbitrary molecules for reactivity, based on their structure alone. For example, the reactivity for a set of quinones is strongly dependent on substituents. In contrast, structural alerts (another way of identifying IADR risk) would treat all these molecules as equally problematic even though some are safe. The reactivity scores our model assigns these 10 quinones correlates closely with

160

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

glutathione reactivity (R2 = 0.622, p = 0.007) and rat hepatocyte toxicity (R2 = 0.732, p = 0.002). In this case, metabolism is ignorable here because quinones are reactive without bioactivation. Even though this model works for a wide range of molecule, these correlations are just as good as those obtained in the literature by a quinone-specific QSAR model. Our approach of using large amounts of reactivity data builds higher quality models than a focused QSAR study, and more accurately identifies reactivity than does structural alerts. In the future, we aim to integrate these models with our group’s metabolism prediction model, Xenosite.

P259. PREDICTION OF THREE-DIMENSIONAL STRUCTURES AND STRUCTURAL FLEXIBILITIES OF WILD-TYPE AND MUTANT CYTOCHROME P450 1A2 BY MOLECULAR DYNAMICS SIMULATIONS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yurie Watanabe1, Shuichi Fukuyoshi1, Masahiro Hiratsuka2, Noriyuki Yamaotsu3, and Akifumi Oda1 1 Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan, 2Graduate School of Pharmaceutical Sciences, Tohoku University, Aoba-ku, Sendai, Miyagi, Japan, and 3School of Pharmacy, Kitasato University, Minato-ku, Tokyo, Japan Cytochrome P450 1A2 (CYP1A2) which is the major drug-metabolizing enzyme among CYPs, is known to have many variant alleles. The genetic polymorphism of CYP1A2 may cause individual differences in the pharmacokinetics of medicines. In the present study, we investigated the effect of genetic polymorphism on the three-dimensional (3D) conformation of CYP1A2 using minimizations and molecular dynamics (MD) simulations of wild-type and mutant CYP1A2. MD simulation is useful for investigating the stability and dynamics of biopolymers such as proteins in a simulated biological environment using the numerical solution of Newton’s equation of motion. The initial 3D structures of the wild-type and mutant CYP1A2 were constructed from experimentally determined structure registered in Protein Data Bank (PDB ID: 2HI4). Using the 3D structure of CYP1A2.1 refined by the MD simulation, the structures of the mutants were constructed. The explicit water solvent was used to simulate the behavior of the enzymes in a water system. The MD simulations each longer than 100 nanoseconds were performed at 300 K under constant pressure. We employed our previously determined force field parameters around the heme iron (Oda et al., 2005), and the AMBER ff12SB force field was adopted for the amino acids. The minimizations and MD simulations were carried out by using the Amber 12 program, and the tleap module of AmberTools was used for the construction of the mutants. The root mean square deviation for the main-chain atoms was calculated to evaluate the MD simulations. The root mean square fluctuation (RMSF) was calculated to determine the flexibilities of the amino acid residues in wild-type and mutant CYP1A2. By comparing the RMSFs of the wild-type and mutants, we investigated the influence of amino acid mutation on the dynamic properties of the proteins. HBOP and HBSITE (Oda et al., 2009) were used to search for hydrophobic sites, which are candidates for ligand-binding pockets or protein-protein interface regions. The results showed that the static structures were changed by the mutations of amino acid residues, not only near the mutated residues but also in distant portions of the proteins. Moreover, the mutation of only one amino acid was shown to change the structural flexibility of the protein, which may influence the substrate recognition mechanism and enzymatic activity.

References Oda A, Yamaotsu N, Hirono S. (2005). New AMBER force field parameters of heme iron for cytochrome P450s determined by quantum chemical calculations of simplified models. J Comput Chem 26:818–826. Oda A, Yamaotsu N, Hirono S. (2009). Evaluation of the searching abilities of HBOP and HBSITE for binding pocket detection. J Comput Chem 30:2728–2737.

P260. PREDICTING REGIOSELECTIVITY AND LABILITY OF CYTOCHROME P450 METABOLISM USING QUANTUM MECHANICAL SIMULATIONS Jon D. Tyzack, Peter Hunt, Matthew D. Segall, and Nicholas W. Foster Optibrium Ltd., Cambridge, UK Many computational methods have been developed that predict the regioselectivity of metabolism by drug metabolising isoforms of the Cytochrome P450 class of enzymes (P450) (Cruciani et al., 2005; Hennemann et al., 2009; Jones et al., 2002; Rydberg et al., 2010; Zaretzki et al., 2011). Here, we describe recent developments to a method for predicting P450 metabolism that combines quantum mechanical (QM) simulations to estimate the reactivity of potential sites of metabolism on a compound with a ligandbased approach to account for the effects of orientation and steric constraints due to the binding pockets of different P450 isoforms. These new developments include modelling reaction pathways for epoxidation and developing models for an extended range of P450 isoforms. The resulting models achieve accuracies of 85-90% on independent test sets. While valuable, predicting the relative proportion of metabolite formation at different sites on a compound is only a partial solution to designing more stable compounds. The advantage of a quantum mechanical approach is that it provides a quantitative estimate of the reactivity of each site, from which additional information can be derived regarding the vulnerability of each site to metabolism in absolute terms. One such measurement is the site lability, as calculated by StarDropÔ (StarDrop), which is a measure of the efficiency of the product formation step. This is an important factor influencing the rate of metabolism and we will illustrate how this provides valuable guidance regarding the potential to redesign compounds to overcome issues due to rapid P450 metabolism.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

161

References Cruciani G, Carosati E, De Boeck B, et al. (2005). MetaSite: Understanding metabolism in human cytochromes from the perspective of the chemist. J Med Chem 48:6970–6979. Hennemann M, Friedl A, Lobell M, et al. (2009). CypScore: Quantitative prediction of reactivity toward cytochromes P450 based on semiempirical molecular orbital theory. ChemMedChem 4:657–669. Jones JP, Mysinger M, Korzekwa KR (2002). Computational models for cytochrome P450: A predictive electronic model for aromatic oxidation and hydrogen abstraction. Drug Metab Dispos 30:7–12. Rydberg P, Gloriam DE, Jed Zaretzki J, et al. (2010). SMARTCyp: A 2D method for prediction of cytochrome P450-mediated drug metabolism. ACS Med Chem Lett 1:96–100. StarDrop, Optibrium Ltd, Cambridge, UK. Available from: www.optibrium.com/stardrop Zaretzki J, Bergeron C, Rydberg P, et al. (2011). RS-predictor: A new tool for predicting sites of cytochrome P450-mediated metabolism applied to CYP 3A4. J Chem Inf Model 51:1667–1689.

P261. IN SILICO PREDICTION OF IN VIVO EFFICACY OF SMALL MOLECULES ON MYCOBACTERIUM TUBERCULOSIS USING MACHINE LEARNING APPROACHES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sitarama Gunturi, Narayanan Ramamurthi, and Pratyusha Vadagam Life Sciences R & D Division, Tata Consultancy Services Ltd., Hyderabad, India Infections by Mycobacterium tuberculosis (Mtb), the causative bacterial agent of tuberculosis (TB), represents an escalating threat for global health with the increased prevalence of multi- and extensively drug-resistant (MDR and XDR, respectively) TB. In some countries, MDR/XDR strains can account for up to 22% of infections (Wright et al., 2009) and it remains a threat to control tuberculosis. It is observed that there is handful of published reports addressing this objective and hence, we have initiated a study to mine, build predictive models and create insights based on the in vivo profiles of 934 candidates whose in vivo anti-TB activity is available and further, assess the stability of the models based on test compounds. The current study is more comprehensive for the following reasons: (a) OECD guidelines have been followed in developing the QSAR model; (b) the test set contains compounds belonging to both the classes, active and inactive; and (c) feature selection algorithm is employed to select features that are significant to predict the mTB of small molecules. In the present study, we have used a dataset of 934 in vivo anti-TB compounds collected from literature (Ballell et al., 2013; Ekins et al., 2014) and 469 two-dimensional calculated physico-chemical and structural properties. Further, we have applied (i) Kennard-Stone algorithm for data division (ii) Genetic Algorithm (GA) for feature selection and (iii) three machine learning approaches, namely, Support Vector Machines (SVM), Classification Trees (CT) and Logistic Regression (LR) to build robust classification models. The performances of the models generated, are compared based on a) sensitivity b) specificity and c) classification accuracy and it is inferred a) Models generated by SVM outperform those generated by CT and LR and b) a six descriptor SVM model is shown to be the best and it has the following statistical criteria, sensitivity (0.7359), specificity (0.6089) and the classification accuracy (0.6766) on the test set.

References Ballell L, Bates RH, Young RJ, et al. (2013). Fueling open-source drug discovery: 177 Small-molecule leads against tuberculosis. ChemMedChem 8:313–321. Ekins S, Pottorf R, Reynolds RC, et al. (2014). Looking back to the future: Predicting in vivo efficacy of small molecules versus Mycobacterium tuberculosis. J Chem Inform Model 54:1070–1082. Wright A, Zignol M, Van Deun A, et al. (2009). Raviglione, epidemiology of anti-tuberculosis drug resistance 2002–07: An updated analysis of the Global Project on Anti-Tuberculosis Drug Resistance Surveillance. Lancet 373:1861–1873.

P262. HUMAN SMALL INTESTINAL EPITHELIAL CELLS DIFFERENTIATED FROM ADULT INTESTINAL STEM CELLS AS A NOVEL SYSTEM FOR PREDICTING ORAL DRUG ABSORPTION IN HUMANS Toru Takenaka1, Harada Naomoto2, Jiro Kuze1, Masato Chiba1, Takahiro Iwao, and Tamihide Matsunaga3 1 Pharmacokinetics Research Lab, Taiho Pharmaceutical Co., Ltd., Tsukuba, Japan, 2Evaluation Research Lab., Taiho Pharmaceutical Co., Ltd., Tsukuba, Japan, and 3Department of Clinical Pharmacy, Graduate School of Pharmaceutical Sciences, Nagoya City University, Nagoya, Japan Widely used cell-based system, Caco-2 cells have a limited capability for the prediction of oral absorption: its tight junctions are markedly tighter than those in the small intestine, which may reduce the permeability of drugs with significant paracellular absorption. Recently, we demonstrated that the commercially available primary human small intestinal epithelial cells (HIEC), composed of adult intestinal stem cells, can continuously proliferate and spontaneously differentiate into enterocytes over 25 passages. Furthermore the mRNA expression levels of uptake/efflux transporters and metabolizing enzymes were comparable between HIEC and Caco-2 cells. In the present study, we investigated the applicability of HIEC to the permeability assay. Under optimized culture condition, HIEC formed morphologically-matured monolayer consisted of polarized columnar epithelia with dense microvilli, tight junctions, and desmosomes on the cell culture insert. Transepithelial electrical resistance across the monolayer was 9-fold lower in HIEC (98.9  cm2) than in Caco-2 cells (900  cm2), which indicated that the looseness of the tight junctions in the HIEC monolayer was similar to that in the human small intestine (40  cm2) (Sjo¨berg et al., 2013),

162

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

whereas that of Caco-2 cells was not. Furthermore, in order to estimate the values of paracellular pore radius and porosity (the fraction of the paracellular space of the membrane surface) of the HIEC and Caco-2 cells monolayers, the effusion-based method was applied to the obtained apparent permeability (Papp) values of polyethylene glycol oligomers (Linnankoski et al., 2010). We observed that there were no significant differences in the pore radius among HIEC, Caco-2 cells, and human small intestine, but the porosity of the HIEC monolayer was closer to human small intestine than that of Caco-2 cells monolayer. Transport studies across HIEC and Caco-2 cells monolayers were conducted using 24 drugs which are absorbed via transcellular diffusion, paracellular pathway, or carrier-mediated transport. The relationship between Papp values and fractions absorbed in humans (Fa) was stronger in HIEC monolayer than that in Caco-2 cells, especially for the drugs which show incomplete absorption (Fa580%) or undergo paracellular transport. In conclusion, the HIEC monolayer can serve as a novel and superior alternative to the conventional Caco-2 cells monolayer for predicting oral absorption of the drugs including paracellularly absorbed drugs.

References Linnankoski J, Ma¨kela¨ J, Palmgren J,et al. (2010). Paracellular porosity and pore size of the human intestinal epithelium in tissue and cell culture models. J Pharm Sci 99:2166–2175. ˚ , Lutz M, Tannergren C, et al. (2013). Comprehensive study on regional human intestinal permeability and prediction Sjo¨berg A of fraction absorbed of drugs using the Ussing chamber technique. Eur J Pharm Sci 48:166–180.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P263. LYSOSOMAL TRAPPING OF CLARITHROMYCIN IN HUMAN ALVEOLAR MACROPHAGES – IN VITRO AND IN SILICO ASSESSMENT Ayse Ufuk1, Michael Gertz2, Jonathan Plumb3, J.Brian Houston1, and Aleksandra Galetin1 1 Centre for Applied Pharmacokinetic Research, Manchester Pharmacy School, University of Manchester, Manchester, UK, 2 F. Hoffmann-La Roche, Basel, Switzerland, and 3University Hospital Of South Manchester NHS Foundation Trust, University of Manchester, Manchester, UK Alveolar macrophages (AMs) may play an important role in respiratory drug accumulation, affecting both lung and blood systemic concentrations. These cells are rich in lysosomes, associated with ‘lysosomal trapping’ of cationic amphiphilic drugs due to their increased ionisation resulting from the pH differences between lysosomes and cytosol. The present study investigated accumulation of clarithromycin in primary human AMs isolated from lung cancer patients, predominantly smokers (n = 9). The extent of lysosomal trapping of clarithromycin was assessed at 5 mM over 10 minutes ± 20 mM ammonium chloride (NH4Cl). Imipramine, a prototypical lysosomotropic compound, was included in the studies as a positive control. Cell-to-medium concentration ratios (Kp,tot) in the absence and presence of NH4Cl were determined. In addition, lysosomal trapping of both drugs (Kp,lysosome) was predicted using a mechanistic in silico cell model developed in Matlab 2012a. Drug physicochemical properties (logP, pKa) and AM specific parameters (e.g. lysosomal volume, pH of different organelles) were used to predict the extent of lysosomal trapping based on pH differences between external medium, cytosol and lysosome; lipid partitioning and electrostatic interactions were also accounted for (Trapp et al., 2008). The mean clarithromycin Kp,tot in human AMs was 60 under control conditions, but showed large variability (CV: 51%). The extent of accumulation was comparable to clarithromycin Kp,tot obtained in AM cell line NR8383. On average, 61% reduction in Kp,tot was observed in the presence of NH4Cl (ranged between 57–67%) indicating an important, but not exclusive, contribution of lysosomal sequestration to clarithromycin accumulation in human AMs. This reduction in accumulation was not as pronounced as in the case of NR8383 (84%), suggesting higher lipid partitioning of clarithromycin in human AMs. Similar trends were observed for imipramine which showed much higher cellular accumulation than clarithromycin (Kp,tot = 609) and 65% reduction in Kp,tot in the presence of NH4Cl. Thein silico cell model correctly predicted high lysosomal accumulation of both drugs in AMs with predicted Kp,lysosome of 68 and 51 for clarithromycin and imipramine, respectively. Furthermore, the predicted reduction in Kp,lysosome of both compounds was comparable with the observed changes in Kp,tot in the presence of NH4Cl. These results highlight important contribution of lysosomal sequestration for accumulation of basic drugs in AMs. Moreover, they indicate substantial variability in drug accumulation between patients due to possible differences in lysosomal abundance, volume and phospholipid content of AMs, in particular in smokers, all of which may have important clinical implications. The developed mechanistic cell model represents a promising in silico tool for predicting the potential of drugs for lysosomal sequestration based on their physicochemical properties and specific cell characteristics.

Reference Trapp S, Rosania GR, Horobin RW. (2008). Kornhuber, quantitative modeling of selective lysosomal targeting for drug design. Eur Biophys J 37:1317–1328.

P264. DETECTION OF PHASE II METABOLITES IN NANOPILLAR-CULTURED HUMAN HEPATOCYTE SPHEROIDS Ryosuke Takahashi1, Katsuhiro Kanda2, Shinji Nagai2, and Masaki Yoshie2 1 Central Research Laboratory, Hitachi, Ltd, Saitama, Japan and 2Hitachi High-technologies Corp., Ibaraki, Japan An in vitro analysis of metabolites using human hepatocytes is expected in early stages of new drug development process, because some pharmacologically active metabolites may contribute to therapeutic and adverse effects to the human body. This

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

163

Downloaded by [112.134.238.102] at 07:30 10 February 2016

enables the prediction of an accurate view of potential in vivo metabolites and metabolic pathways. The purpose of this study is to analyze phase II metabolites of lamotrigine and salbutamol under feeder-free hepatocyte spheroid culture system using Nanopillar Plate.Human cryopreserved hepatocytes were seeded on the Nanopillar Plate and cultured for five days by exchanging the medium daily until complete spheroid formation. Lamotrigine (anticonvulsant drug) or salbutamol (bronchodilator drug) were added to the spheroids respectively on day 5, 8, 12, 19 and 26, and were incubated in hepatocyte spheroids for three and seven days. The targeted phase II metabolites of lamotrigine-N-glucuronide and salbutamol-4-o-sulfate in the respective culture medium were analyzed by LC/MS/MS.An extracted ion chromatogram of lamotrigine-N-glucuronide parent ion (m/z 432) showed two peaks of RT 8.8 and 9.5. Additionally, MS analysis of these peaks also revealed an isotopic ion (m/z 434) specific for 37Cl of lamotrigine along with the ion of m/z 432 specific for 35Cl. Moreover, lamotrigine-N-glucuronide fragment ion (m/z 256), corresponding to lamotrigine molecule, was detected by MS/MS analysis as well as isotopic fragment ion of m/z 258. These two peaks of different RT were speculated as lamotrigine-2N-glucronide, a major metabolite by UGT1A3 and 1A4, and a minor metabolite of lamotrigine-5N-glucronide. In terms of an extracted ion chromatogram of salbutamol-4-o-sulfate parent ion (m/z 318) showed a single peak of RT 8.0. MS/MS analysis of this ion showed two fragment ions consisted of salbutamol molecule (m/z 238) and sulfate molecule (m/z 97). Both targeted conjugates were detected in all samples collected in time course at least up to one month of incubation. Furthermore, Semi-quantitative analysis showed the amounts of these two metabolites tend to be higher at seven days of exposure rather than three days.In conclusion, feeder-free hepatocyte spheroid culture system using the Nanopillar Plate maintained sufficient phase II metabolism activities, which is suitable for practical long-term in vitro assay to predict human metabolites.

P265. DEVELOPMENT OF TRAPPING ASSAY FOR UGT-MEDIATED METABOLIC ACTIVATION USING CYSTEINE IN HUMAN LIVER MICROSOMES Yasuyuki Toyoda1, Hiroshi Harada1, Noboru Kamada1, Masaru Tsutsui1, Takuro Endo1, Makoto Murakami1, and Mamoru Kobayashi2 1 Pharmacokinetics Research, Kissei Pharmaceutaical Co., Ltd., Matsumoto-City, Japan and 2Development Research, Kissei Pharmaceutical Co., Ltd., Azumino-City, Japan Purpose: Drugs with carboxylic acid can be metabolized to acylglucuronide (AG) by UDP-glucuronosyltransferases (UGT). AGs have been thought to be intrinsically reactive metabolite and their reactivities have been thought to be associated with covalent binding to protein and occurrence of adverse drug reactions. Therefore, assessment for the reactivity of AGs is important for evaluating the potential risk of toxicity of new chemical entities in early drug discovery phase. Some trapping assays have been widely used for the detection and characterization of reactive metabolites, especially for those of cytochrome P450-mediated metabolic activation. However, only a few trapping assays have been performed for AGs as yet. In this study, we developed a new in vitro approach for detecting reactive AGs by using cysteine (Cys) as a trapping agent and evaluated the reactivity of AGs qualitatively and quantitatively. Methods: Some drugs with carboxylic acid were incubated with human liver microsomes (HLMs) in the presence of UDP-glucuronic acid and Cys. After the reaction was quenched by adding acetonitrile, the samples were analyzed using LC-TOFMS to identify the structure of Cys adduct. Next, we examined the trapping ability for reactive AGs using various Cys derivatives, such as glutathione, N-acetylcysteine and so on. Finally, to evaluate the reactivity of AGs quantitatively, we used [35S]Cys as a trapping agent, and [35S]Cys adducts were analyzed by radio-HPLC. Results and conclusions: N-Acyl-Cys amide adducts formed via their reactive AGs showed a mass shift of 103 amu, which was consistent with the mass shift of Cysteine conjugation and dehydration, compared with the unchanged form. In addition, it was suggested that the formation of Cys adduct needs anN-terminal Cys moiety with not only a free amine group but also a free thiol group, from the results using various Cys derivatives as the trapping agents. It was also proposed that the S-acyl-Cys thioester adduct can be formed as the intermediate from reactive AGs, and it then transforms to the N-acyl-Cys amide adduct by intramolecular Sto N-acyl rearrangement. Furthermore, it was indicated that the peak area resulted from the formation of [35S]Cys adduct corresponds to the reactivity of AGs. A structure effect on the reactivity of AGs obtained from this assay demonstrated the rank order: Acetic acid4Isopropionic acid4Benzoic acid derivatives. In conclusion, we detected the reactive AGs formed in HLMs as the stable Cys amide adducts and quantitatively evaluated the degree of AG reactivity. Therefore, this approach has a possibility to detect the reactivity of AGs easily and to screen the potential risk of new chemical entities usefully in early drug discovery phase.

P266. LONG-TERM HEPATOCYTE CULTURE FOR DRUG SAFETY ASSESSMENT SCREENING Jessica A. Bonzo, Lauren E. Sangenario, Shyam Kumar, and Rafal P. Witek Department of Cell Biology, Thermo Fisher Scientific, Frederick, MD, USA Primary human hepatocyte cultures are the gold standard for preclinical drug safety assessment as well as required by the FDA for evaluation of drug induction potential. Cryopreserved human hepatocytes have been validated as a comparable model to fresh human hepatocytes. However, these cultures are limited by their relatively short life-span of 5–7 days accompanied by a rapid drop in metabolic function which precludes, among other experiments, evaluation of low metabolic turnover compounds, multi-

164

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Downloaded by [112.134.238.102] at 07:30 10 February 2016

day toxicity studies, and long term viral infection studies. The need for longer term hepatocyte cultures has increased with recent initiatives by European regulatory bodies to limit animal testing. To address this need, we are evaluating a new cell culture media that maintains the function and viability of cryopreserved human hepatocytes for at least 10 days. The new culture media has been tested for the ability to prolong hepatocyte life as determined by morphological assessment for polarity and bile-canaliculi formation using carboxy dichlorofluorescein diacetate (CDFDA) staining. Cell health was evaluated by performing CyQuant, ATP, and LDH assays. Albumin production was assessed using a human-specific albumin ELISA. In addition, to understand the effect of long term culture on metabolic function, Phase I and II enzymatic activities have been assessed by luminescent assay and confirmed by LC/MS. Hepatocytes cultured in the new long term culture optimized media displayed CYP1A2, CYP2B6, and CYP3A4 activity levels significantly greater than those achieved using standard culture media after 5 days. Cultures maintained in the new media survived for at least 10 days with metabolic activities comparable to day 5; this was not achievable with standard culture media. In addition supporting data shows hepatic cultures in the new media after 10 days maintained polarity, bile canaliculi formation, ATP, and albumin production. Finally, minimal LDH levels were observed at day 10 suggesting that the new media supports healthy hepatic cultures for minimum of 10 days. As our data suggests, hepatocyte long term culture media supports hepatocyte health and metabolic function for an extended period of time. This enables researchers to perform metabolic and toxicity experiments not achievable using currently available standard culture conditions. This further allows for multi-day toxicity studies and more predictive data interpretation. Finally, extended hepatic culture allows pharmaceutical researchers to comply with new regulatory requirements.

P267. A NOVEL 3D-CULTURE SYSTEM ACTIVATING HEPATIC FUNCTION OF HEPG2 CELLS UTILIZING A COLLAGEN VITRIGEL MEMBRANE CHAMBER AND ITS APPLICATION TO ADME/TOX ANALYSES Ayumi Oshikata-Miyazaki1, Seiichi Ishida2, and Toshiaki Takezawa1 1 Division of Animal Sciences, National Institute of Agrobiological Sciences, Tsukuba, Japan and 2Divison of Pharmacology, National Institute of Health Sciences, Tokyo, Japan Purpose: A collagen vitrigel membrane (CVM) is composed of high-density collagen fibrils equivalent to connective tissue in vivo and possesses excellent strength, transparency, and permeability of protein with high molecular weight, so that it functions as a scaffold for anchorage-dependent cells and a carrier of drug delivery (Takezawa et al., 2004, 2007). Recently, we fabricate a human corneal epithelium model in a CVM chamber and developed a novel eye irritancy test (EIT) method ‘‘VitrigelEIT method’’ (Yamaguchi et al., 2013). Meanwhile, it was reported that the liver-specific functions were activated by incorporating an oxygen permeable scaffold and/or a three-dimensional co-culture system with a different type of cells into the culture system of hepatic cells. In this study, we aimed to develop a hepatic model useful for ADME/Tox assays as a tool for drug development by utilizing a CVM chamber with advantages of both oxygen permeability and co-culture system. Methods: HepG2 cells were seeded and cultured in CVM chambers fixed in a normal culture dish for 48 hours, and subsequently the cells were cultured on the liquid-solid, liquid-liquid and liquid-gas interface. Also, the co-culture system was fabricated by overlaying collagen sol on the HepG2 cells and seeding human umbilical vein endothelial cells (HUVECs) on the overlaid collagen gel. Cells in culture were periodically observed with a phase-contrast microscope and their albumin secretion level, urea synthesis level and CYP3A4 activity were analyzed. The ADME behavior of a model drug was observed with a fluorescence microscope after adding fluorescein diacetate (FD) into culture media. Also, the toxicity of acetaminophen (APAP) was evaluated by fluorescent assays for mitochondrial membrane potential and cell membrane permeability. Results and discussion: Liver-specific functions such as albumin secretion level, urea synthesis level and CYP3A4 activity were significantly activated in cells cultured on the liquid-gas interface for only 1 day in comparison to the cells in other conditions. Also, the effect of concentration-dependent APAP was confirmed by culturing HepG2 cells on the liquid-gas interface. Moreover, FD was incorporated into the cells and metabolized into a fluorescent molecule in every condition. Interestingly, the cells in the sandwich-co-culture system most rapidly excreted the intracellular fluorescence into a bile canaliculus-like structure. Especially, bile-canaliculus-like structures formed by HepG2 cells in monolayer were most clearly observed in the sandwich-co-culture system in comparison to other culture systems. Therefore, we hope the co-culture system of HepG2 cells and HUVECs fabricated in a CVM chamber on the liquid-gas interface would provide a new aspect toward the studies on hepatic ADME/Tox essentially required for new drug development.

References Takezawa T, Ozaki K, Nitani A, et al. (2004). Collagen Vitrigel: A novel scaffold that can facilitate a three-dimensional culture for reconstructingorganoids. Cell Transplant 13:463–473. Takezawa T, Takeuchi T, Nitani A, et al. (2007). Collagen vitrigel membrane useful for paracrine assays in vitro and drug delivery systems in vivo. J Biotechnol 131:76–83. Yamaguchi H, Kojima H, Takezawa T. (2013). Vitrigel-eye irritancy test method using HCE-T cells. Toxicol Sci 135: 347–355.

P268. HUMAN PCIS AS EX VIVO MODEL TO STUDY THE INTERPLAY BETWEEN P-GP AND CYP3A4 IN THE INTESTINE Ming Li, Inge A.M. de Graaf, Marina H. de Jager, and Geny M. M. Groothuis Department of Pharmacy, University of Groningen, Groningen, the Netherlands

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

165

Introduction: Both efflux transporters and metabolic enzymes are expressed in the intestinal epithelium, and they work coordinately to reduce the intracellular concentration of xenobiotics and the absorption of orally taken drugs. The drug–drug interactions related to this are of clinical importance and require investigation in the pre-clinical phase and preferably with human tissue. However, proper and efficient in vitro or ex vivo models of the human intestine that express drug transporters and metabolizing enzymes at physiological levels in each region are lacking.The Precision-Cut Intestinal Slice (PCIS) model has been established to investigate drug metabolism, toxicity, and more recently transport in both human and animal tissue. The aim of this study is to show its application toinvestigate the interplay between P-gp and CYP3A4 in human intestine by studying the transport and metabolism of model compound simultaneously. Methods: Human PCIS were prepared from the explant of human intestine. Quinidine was used as duo-substrate of P-gp and CYP3A4. CP100356 and PSC 833 were employed as selective P-gp inhibitors,while verapamil and ketoconazole were used as inhibitors of both transporters and metabolic enzymes. The metabolite, 3OH-quinidine, was measured by LC-MS in both medium and tissue in the presence or absence of the inhibitors. Furthermore, the tissue content of quinidine was also quantified by LC-MS to investigate the influence of transporter inhibition. Results: In human jejunum,the intracellular concentration of quinidine was limited by P-gp effluxat low concentrations (below 5 mM), whereas at concentrations between 5 to 50 mM CYP3A metabolism became limiting. The selective P-gp inhibitors increased the intracellular concentrations of both quinidine (2.3–2.7fold) and 3-OH-quinidine (9.4–11.4 fold), and the total metabolite production (1.9–2.3-fold).Verapamil and ketoconazole increased the intracellular quinidine concentrations but decreasedthe metabolite production significantly. As expected,the effect of this interplay was different in PCIS of the different regions of the intestine. Conclusion:This study showed that human PCIS is a reliable, simple and fast ex vivomodel to study the intestinal transport and metabolism together. The results indicate that the interplay between P-gp and CYP3A4is dependent on the concentration of the drug and the selectivity of the inhibitors. Furthermore, it exhibits regional differences, probably due to the different expression profiles of P-gp and CYP3A4.In future studies, human PCIS can be used to predict the interplay between transport and metabolism to decrease the risk of drug–drug interactions and inadequate pharmacotherapeutic effects of newly developed compounds.

P269. COMPARISON OF THE CORRELATION BETWEEN CYP3A ACTIVITY AND PROTEIN EXPRESSION IN SPHEROID AND MONOLAYER CULTURE Masataka Santoh1, Seigo Sanoh1, Masashi Takagi1, Yoko Ejiri2, and Shigeru Ohta1 1 Graduate School of Biomedical & Health Sciences, Hiroshima University, Hiroshima, Japan and 2Microdevices Development Team, New Business Development Division, Kuraray Co., Ltd, Tsukuba, Japan Introduction: Cytochrome P450 (CYP) families, drug metabolic enzymes, are well known to reduce their activities quickly depending on periods in hepatocyte culture. Recently, various approaches such as three-dimensional (3-D) cultures have been designed to mimic liver in vivo and maintain metabolic activities for precise evaluation of drug metabolism in hepatocyte culture. In the present study, we focused on rat primary hepatocyte spheroids cultured on 3-D culture plate (ElplasiaTM, Kuraray Co., Ltd.). We compared CYP3A1 enzyme in terms of mRNA, protein levels and metabolic activity between spheroid culture and monolayer culture to clarify the mechanism of maintenance of metabolic activity. Furthermore, NADP and NADPH contents which are major cofactors related to CYP metabolism were also measured in both cultures. Methods: Rat primary hepatocytes were harvested from Sprague-Dawley male rat (7 weeks old) according to two-step collagenase perfusion procedure, and were inoculated for 9 days on ElplasiaTM. Hepatocyte viability after harvest was more than 90%. Spheroid formation of rat hepatocytes was achieved by inoculation for 5 days. CYP3A1 mRNA level and protein level were measured using real-time RT-PCR and immunoblotting. The metabolic activity level was evaluated using P450-GloTM CYP3A4 assay with Luciferin-IPA (Promega). NADP and NADPH contents were evaluated using NADP/NADPH-GloTMAssay (Promega). Results and Discussion: The mRNA level and metabolic activity were higher in spheroid culture than monolayer culture, and spheroid culture maintained them for 9 days. However, the protein levels maintained for 9 days in both cultures, and there was no difference between them. The correlation between CYP3A1 activity and protein levels was almost kept from day 0 to day 9 in spheroid culture, but not observed in monolayer culture. The contents of NADP and NADPH were higher in spheroid culture than monolayer culture. These results suggest that the loss of CYP3A1 activity in monolayer culture is not due to CYP3A1 protein levels, but rather due to other factors.

P270. DEVELOPMENT OF A HIGH CONTENT TOXICITY ASSAY USING HmREL CORPORATION HEPATIC CO-CULTURES James K. Morelli1, J. Dwyer1, E. Novik2, and Z. Jayyosi1 1 Dsar, Sanofi, Framingham, MA, USA and 2HmREL Corporation, North Brunswick, NJ, USA Technologies designed to de-risk hepatotoxicity in the early phase of drug discovery, using in vitro methods, have been continuously advancing for many years. The goal of these advances is to create an in vitro system which mimics in vivo liver function as closely as possible. We are collaborating in the evaluation of one such system, hepatocytes co-cultured with supporting stromal cells, developed by HmREL Corporation. HmREL co-cultures survive for much longer time periods compared to traditional primary hepatic cultures, allowing repeated treatments with compound for at least 14 days, mimicking animal study

166

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

treatment periods. In addition, these cultures maintain a number of markers of in vivo hepatic function which are lost in primary hepatic monocultures, including retention of metabolic enzyme expression for at least 14 days. The co-cultures are available in human, rat, dog and primate format allowing for interspecies drug metabolism and toxicity comparison. In this poster, we describe the development of a high content analysis (HCA) assay using HmREL human and rat co-cultures. As hepatocytes and not stromal cells are the target of toxicity analysis, it was necessary to develop a method to discriminate stromal cells from hepatocytes in HCA images. Once this was achieved, cells were treated with various test articles for 24 hours, including both compounds known to be hepatotoxic and negative control compounds. Cultures were then stained with fluorescent indicator dyes and evaluated for cytotoxicity, generation of reactive oxygen species (ROS) and alteration in mitochondrial membrane potential (MMP). Results include a rank ordering of compounds by cytotoxicity, with additional markers providing insight into potential mechanisms of toxicity.

P271. IN VITRO MODELING OF INFLAMMATORY RESPONSE ASSOCIATED WITH CHOLESTATIC INJURY IN HEPARG

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Rafal P. Witek, Jessica A. Bonzo, and Mukesh Kumar Department of Cell Biology, Thermo Fisher Scientific, Frederick, MD, USA HepaRG is a differentiated hepatic cell line that exists as two different phenotypes of hepatocyte-like and biliary-like cells. The two phenotypes of those cells coexist during culture in approximately one-to-one ratio. This bipotential differentiation behavior creates an organotypic-like co-culture system that is perfect for an in vitro modeling of effects of hepatic bile acid excretion on biliary cells during drug induced cholestasis or drug-induced cholangiopathy. The same system can also be used for modeling the effects of cytokines produced by biliary cells on hepatocyte function during cholastatic injury. As such, HepaRG represents a unique system that can be used to study liver injury associated with biliary excretion and cholestasis without the need of setting up more advanced organotypic models or 3D cultures. In this work, we use HepaRG to test its ability to predict immune mediated cholastatic injury utilizing clinically relevant compounds. All experiments were performed using terminally differentiated HepaRG cells. Cytokine 10- and 30-Plex Luminex Panel were used to asses cytokine production by HepaRG before and after treatment with drugs Cyclosporine A, Chlorpromazine, Flucloxacillin, and Troglitazone. All treatments were performed in presence or absence of LPS to determine the effect of those compounds on cytokine expression and inflammatory potential. Live/Dead assay, ALT, ATP, G6P-ase, Albumin and Urea production were used to assess cell viability, health and toxicity. Morphological and immunohistochemistry assessments were used to differentiate between hepatocyte and biliary cells, and to determine specificity of cells that were affected during the injury. As expected, following treatment with selected compounds, differential cytokine expression was observed. Chloropromazine and Troglitazone induced inflammatory like response that resulted in cytokine (IL6, TNFa, other) destabilization and up regulation. On the other hand, Flucloxacillin preferentially induced toxicity to biliary cells and had less effect on cytokine expression. Addition of LPS during compound treatment elevated the toxic effect and increased cytokine release suggesting that HepaRG by itself can be used to mimic the immune component that is involved in progression of liver injury. In conclusion, we show that HepaRG offers unique capability for an in vitro modeling of liver cholastatic injury and potentially enables early identification of drug candidates causing cholestasis or cholangiopathy.

P272. COMPARATIVE ANALYSIS OF DILI COMPOUNDS ON CYTOKINE PRODUCTION IN KUPFFER CELLS AND HEPATOCYTES Rafal P. Witek, Jessica A. Bonzo, Lauren E. Sangenario, and Mukesh Kumar Department of Cell Biology, Thermo Fisher Scientific, Frederick, MD, USA Presence of inflammatory cytokines in the liver is often associated with indirect Drug Induced Liver Injury (DILI) that results in liver parenchymal cell deterioration. This idiosyncratic toxicity is often overlooked due to lack of validated in vitro cell models pressuring development of new models that would elevate this problem. One of the recent developments utilizes an organotypic approach of co-culture of human hepatocytes and Kupffer cells to establish a model of liver inflammation. In this model, the presence of inflammatory cytokines modulates metabolic function of co-cultured hepatocytes. Although this model has been shown to predict Trovafloxacin toxicity that results from cytokine imbalance and accumulation of cytotoxic TNFa, to date there are no published studies examining the predictively of the system with other drugs known for idiosyncratic toxicity. For this reason, we performed comparative analysis of cytokine expression by Kupffer cells and co-cultures treated with panel of over 20 different toxicants know for inducing cytokine imbalance following their administration. The following list of compounds were used in the analysis: (dose dependent hepatotoxicity) Acetaminophen, Benzbromarone, Amiodarone; (dose independent hepatotoxicity) Fluconazole, Trovafloxacin, Phenacetin, Diclofenac, Quinine, Chlorzoxazone, Methotrexate, Chlorpromazine, Riluzole, Tamoxifen, Mexiletine HCL, Progesterone, Clomipramine, Nortriptyline HCL, Halothane; (minimally or not toxic, controls) Levofloxacin, Bupropion HCL, Ranitidine, Pioglitazone, Colchicine, Dexamethasone, Buspirone. Cell cultures (hepatocytes, co-cultures and Kupffer cells) were treated with various concentrations of each of the compounds in presence or absence of LPS. Following 24 hr incubation, media was collected and assayed for presence of cytokines using Cytokine 10- and 30-Plex Luminex Panel. In addition, RNA was isolated from treated cells and used for qRT-PCR analysis of major Phase I and II

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

167

enzymes. Obtained data indicates cytokine imbalance affecting hepatic functions that follows administration of various test compounds (Trovafloxacin, Diclofenac, Clomipramine, other) thus recapitulating previously published in vivo observations. The data, as shown by our comparative analysis, supports use of human hepatocytes co-cultured with Kupffer cells to determine potential imbalance in cytokine production that could result in hepatoxicity.

P273. 3D ORGANOTYPIC HUMAN HEPARG SPHEROID CULTURES FOR IN VITRO TOXICOLOGICAL STUDIES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Patrina Gunness1, Valery Schevchenko2, Christiane Guguen-Guillouzo2, Simon Messner1, Christophe Chesne2, and Jens Kelm1 1 InSphero AG, Schlieren, Switzerland and 2Biopredic International, Saint-Gregoire, France Drug-induced liver injury (DILI) remains the main cause of acute liver failure (ALF) and post-market drug withdrawals. The well-documented limitations of pre-clinical in vivo animal studies and in vitro two-dimensional (2D) human hepatic models hinder the accurate prediction of DILI in humans. Primary human hepatocytes (PHHs) remain the gold standard for in vitro toxicological studies; however, their use is often hindered by their limited availability and inter-donor variability. The human HepaRG cell line, which is often described as a surrogate for PHHs is known to possess liver- specific phenotype. The objective of the study was to characterize long-term 3D human HepaRG spheroid cultures in order to assess their suitability for in vitro toxicological studies. 3D HepaRG spheroid cultures were obtained using InSphero’s hanging drop technology. Following, spheroid formation, the cultures were maintained for 4 weeks and assessed for liver-specific morphology and function. The cultures were viable and maintained stable size and ATP content (diameter &250 mm; ATP content &15 pmol/HepaRG spheroid) over the culture period in the absence of a necrotic core. Results suggest that the cultures possess liver-specific phenotype evidenced by albumin and urea production. Furthermore, cultures possessed both basal and inducible CYP3A4 enzyme activity, which is one the most imperative enzymes in drug metabolism and toxicology. Therefore, taken all together, the results from the study suggest that the long-term organotypic human HepaRG spheroid cultures may be a promising in vitro tool for toxicological studies, including predictive and investigate studies.

P274. MODIFICATION OF CULTURE MEDIUM COMPOSITION AMELIORATES INTERCELLULAR BARRIER PROPERTY OF A HBMEC/CIb-BASED BLOOD–BRAIN BARRIER MODEL THROUGH PROMOTION OF CELLULAR DIFFERENTIATION Tomomi Furihata, Shinya Kawamatsu, Ryo Ito, Shota Suzuki, and Kan Chiba Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Chiba University., Chiba-Shi, Japan In vitro blood–brain barrier (BBB) models are important tools for central nervous system (CNS) drug development. The primary component of such models is the brain microvascular endothelial cells (BMEC) that form a cell monolayer, which mimics the BBB. One of most important functional features of the BBB is its remarkably strong intercellular adhesion, which firmly restricts penetration of diverse molecules into the brain. Therefore, the cells used in in vitro BBB models require the ability to bind together to form a tight intercellular barrier. Recently, in order to obtain valuable cells for use in development of in vitro BBB models, we established a new line of conditionally immortalized human brain microvascular endothelial cells (HBMEC/cib) (Kamiichi et al., 2012). It has been shown that a HBMEC/cib-based in vitro BBB model possesses intercellular barrier properties against sucrose or sodium fluorescein (Na-F) penetration to an extent that is similar to that of a primary BMEC monoculture model. Nevertheless, the functional integrity of the barrier needs to be improved. Therefore, the present study aimed at improving the barrier properties of HBMEC/cib via modifications to the culture medium composition. HBMEC/cib was cultured with either CSC-Complete recombinant medium supplemented with cultureboost-R (hereafter referred to HBMEC/CSC-cbR) or CSC-Complete recombinant medium supplemented with hydrocortisone (180 nM) (hereafter referred to HBMEC/CSC-HC). The results of Na-F permeability determination showed that HBMEC/CSC-HC exhibited significantly lower Na-F permeability compared with that of HBMEC/CSC-cbR (1.49 ± 0.38 vs. 0.94 ± 0.15 [103 cm/min]). Consistently, it was found that HBMEC/CSC-HC showed transendothelial electric resistance, while HBMEC/CSC-cbR did not (25.0 ± 5.6 vs. 2.5 ± 0.2 [V  cm2]). To obtain mechanistic insights, immunocytochemical analysis of adherens junction proteins was performed. The results showed that, while vascular-endothelial cadherin, b-catenin, and zonula occludens-1, were diffused in the cytoplasm in HBMEC/CSC-cbR, those expressed in HBMEC/CSC-HC were primarily localized at the cell border, which is considered likely to explain the barrier property improvement. Because it is known that tightening cellular adhesion is regarded as a hallmark of the endothelial cell differentiation process, and is often associated with decrease of cell mobility, an endothelial differentiation marker gene mRNA expression analysis and scratch assay were conducted using these cells. As expected, the results showed that von Willebrand factor mRNA level was approximately 5-fold higher in HBMEC/CSC-HC than that of HBMEC/CSC-cbR, and that HBMEC/CSC-HC shows significantly retarded migration ability compared with HBMEC/CSC-cbR. Based on the above findings, we conclude that the modification of the culture medium composition ameliorates the intercellular barrier property of our HBMEC/cib-based BBB model through promotion of cellular differentiation.

Reference Kamiichi A, Furihata T, Kishida S, et al. (2012). Establishment of a new conditionally immortalized cell line from human brain microvascular endothelial cells: A promising tool for human blood-brain barrier studies. Brain Res 1488:113–122.

168

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P275. THE USE OF STRAT Mä MEMBRANE FOR DERMAL PENETRATION STUDIES: EFFECT OF FORMULATION ON THE PERMEATION OF ANILINE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Andrew McEwen, Stuart G. Wood, Gill Ford, Hayley Leech, and Karen Wilson Quotient Bioresearch Limited, Rushden, UK The use of radiolabelled test compound and dermatomed human skin are the gold standard requirements for in-vitro dermal penetration studies. Whilst this provides valuable information for use in risk assessment studies for regulatory submission, the search for alternatives for use in the discovery phase of development continues. Often what is required is a comparative assessment of penetration between alternative drug products or drug formulations prior to selection of the lead compound. In this case the use of human skin is not required and decisions can be made based on permeation through a standard membrane. The use of a standard membrane would remove the effects of sample variability observed in biological samples. In a previous study we reported a comparative evaluation of the permeation of increasing dose concentration of [14C]-Aniline through Strat-MÔ membrane and human dermatomed skin (Wood et al., 2013). In this study we report a comparative evaluation of the effect of formulation on the permeation of [14C]-aniline through Strat-M membrane. Membranes were mounted in Franz static dermal penetration cells, the receptor fluid was filled with physiological saline with 5% bovine serum albumin (12 mL) and placed into a V9 magnetic stirrer block. The receptor fluid was maintained at a constant temperature of 32 ± 2  C using heated water jackets. The integrity of each sample was determined using tritiated water. The formulated test compound was applied as a single finite dermal application for a maximum of 24 hours. Concentrations tested were in the range 100–500 mg/mL. All cells were occluded following dosing. Receptor fluid was analysed for concentrations of radioactivity. To help with the interpretation of the results additional experiments were performed using [14C]-testosterone, one of the OECD reference compounds, for comparative purposes. Permeation constants (Kp) were calculated from the data obtained. For aniline the permeation constants were in the ranges 1.0E  03 to 9.0E04. For testosterone the permeation constants were 3.1E04 and 2.5E03 using Strat-M membrane. The permeation of aniline through Strat-M membrane has been investigated using different formulations. The permeation rates observed for testosterone were similar to those observed for aniline. The use of a standard membrane could enable compounds to be ranked in a reproducible manner.

Reference Wood et al. (2013). 10th International ISSX Meeting, Toronto.

P276. ENGINEERING HIGHLY FUNCTIONAL COCULTURES OF CRYOPRESERVED PRIMARY HUMAN HEPATOCYTES AND KUPFFER MACROPHAGES VIA MICROFABRICATION Christine Lin1, Scott Heyward2, Caitlin A. Brown2, and Salman R. Khetani1 1 School of Biomedical Engineering, Colorado State University, Fort Collins, CO, USA and 2Bioreclamation IVT, Baltimore, MD, USA Given differences between animal and human livers in drug metabolism pathways, cultures of primary human hepatocytes (PHHs) are used to evaluate drug metabolism and toxicity. However, confluent PHH monolayers on adsorbed collagen with or without an extracellular matrix (ECM) overlay display a rapid decline in hepatic functions, and lack Kupffer macrophages (KM) that are known to modulate hepatocyte responses to drugs via release of cytokines. Here, we sought to create a culture platform that can keep both cryopreserved PHHs and KMs highly functional for several weeks allowing modeling of chronic drug treatment and inflammatory stimuli effects. First, we compared the viability and attachment efficiencies of several lots of fresh and cryopreserved Kupffer macrophages (KMs). The effects of serum, culture medium supplements and ECM coating on KM viability and attachment were also tested. We found that cryopreservation did not significantly compromise the viability (i.e. 80– 82% for fresh vs. 78.8–82% for cryopreserved) or the attachment efficiency of the different KM lots. Furthermore, KMs could attach well (70–80% attachment efficiency) to un-coated or collagen-coated tissue culture polystyrene in the presence of bovine serum. Next, we introduced the aforementioned plateable cryopreserved KMs in a micropatterned co-culture (MPCC) model containing PHHs and 3T3-J2 murine embryonic fibroblasts. In particular, tissue culture polystyrene 24-well plates were subjected to soft lithography to pattern rat-tail collagen I into circular domains of empirically-optimized dimensions. PHHs selectively attached to the collagen domains and were then surrounded by the fibroblasts to create MPCCs. Thawed KMs were added at ratios of 1:1 or 1:0.4 (PHH:KM) onto MPCCs functionally stabilized over 7–9 days. The addition of various densities of un-stimulated KMs to MPCCs did not cause significant changes in PHH functions (albumin, urea, CYP450 activities) over at least 2 weeks. However, after stimulation with lipopolysaccharide (LPS) over 24 hours, CYP3A4 enzyme activities in PHHs decreased up to 3.5-fold in the presence of KMs. Secretions of TNF-a and IL-6 cytokines in culture supernatants as measured via ELISA increased substantially (up to 9-fold for TNF-a and 190-fold for IL-6) following stimulation with LPS and such increase in cytokine release was directly proportional to the number of KMs present in culture. KMs were present as assessed via CD68 and CD14 immunostaining and functional as assessed by the aforementioned cytokine release for at least 1 week on top of MPCCs. Finally, we evaluated modulation of clearance and toxicity of prototypical drugs in PHHs in MPCCs in the presence or absence of KMs. In conclusion, we have created a culture platform that can keep both cryopreserved PHHs and KMs highly

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

169

functional for prolonged durations in vitrotowards evaluating the effects of interactions between these cell types on drug disposition. In the future, we aim to incorporate other cell types of the liver in our model such as endothelial and stellate cells in order to better mimic in vivo outcomes.

P277. ABSTRACT WITHDRAWN

P278. EXPLORING DIFFERENCES IN DRUG MEMBRANE PERMEABILITY BETWEEN INTESTINE AND LIVER

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Hideaki Okochi1 and Leslie Z. Benet2 1 Department of Bipharmaceut Sciences, UCSF, San Francisco, CA, USA and 2Biopharmaceutical Sciences, UCSF, San Francisco, CA, USA Objective: In terms of the Biopharmaceutical Drug Disposition Classification System (BDDCS), both class-1 (BDDCS-1) and class-2 (BDDCS-2) drugs are extensively eliminated by drug metabolizing enzymes in the intestine and liver. Yet, BDDCS-2 drugs are known to be influenced by drug influx as well as efflux transporters in the liver and efflux transporters in the intestine, while BDDCS-1 drugs exhibit no clinically significant transporter effects in either tissue. To understand these differences, we characterized the passive diffusion and intrinsic uptake clearance of two BDDCS-1 and two BDDCS-2 drug exhibit higher permeabilities in rat intestine (enterocytes) and liver (hepatocytes). Methods: Drugs: Diltiazem (DTZ) and quinidine (QND) were used as BDDCS-1 drugs, while atorvastatin (ATV) and cyclosporine (CsA) were used as BDDCS-2 drugs. Enterocyte isolation: Enterocytes were isolated from jejunum of 300–350 g male SD rats by incubation with Matrispase solution (BD Biosciences) at 4  C for 6 hours. Hepatocyte isolation: Hepatocytes were isolated from male SD rats using the 2-step collagenase perfusion (37  C) technique. Drug uptake study: In 24-well plates, 0.4-mL of enterocytes (5  105 cells/mL) or hepatocytes (4  106 cells/mL) suspension were pre-incubated in an incubator-shaker (37  C, 120 rpm). Uptake studies were initiated by adding 400 mL of target drug into the cell suspension (6 different concentrations per drug, including 0.1 mCi/mL of 3H-drug). At predetermined times (0, 20, 40, 60, 90, 120, 300, 600 and 900 sec), 500 mL samples were taken and cells isolated by a rapid oilfiltration technique. Cells were resuspended and sonicated in 100 mL of H2O, and drug (liquid scintillation counting) and protein (microBCA kit) concentrations were determined. Data analysis: Data were fit to the Michaelis–Menten equation (V = Vmax  S/Km + Pdiff  S) and the passive diffusion component (Pdiff, mL/min/106 cells) and intrinsic uptake clearance (PSu,inf, mL/min/g liver) were estimated. Results. BDDCS-1 drugs showed the highest intracellular accumulations at the first point (20 sec) and stayed at plateau over 900 sec; in contrast, accumulations of BDDCS-2 drugs gradually increased and reached the highest levels at 120–300 seconds. Passive diffusion (Pdiff) into enterocytes was 32.0 (DTZ), 79.9 (QND), 1.17 (ATV) and 8.46 (CsA). Pdiff into hepatocytes was 9.57 (DTZ), 14.4 (QND), 0.0331 (ATV) and 0.149 (CsA). Intrinsic uptake clearance (PSu,inf) into enterocytes was 19.4 (DTZ), 48.6 (QND), 0.962 (ATV) and 3.71 (CsA). PSu,inf into hepatocytes was 601 (DTZ), 1369 (QND), 0.615 (ATV) and 0.133 (CsA). Conclusion. Both BDDCS-1 and BDDCS-2 drugs exhibited higher passive diffusion for enterocytes vs hepatocytes, but this was not true for intrinsic uptake clearances, which for BDDCS-1 drugs were 20-fold higher in hepatocytes than enterocytes. These are the first studies to compare the passive permeability and active uptake of drugs in enterocytes versus hepatocytes, here concentrating on the differences between BDDCS classes 1 and 2.

P279. A PERFUSED OXYGENATED BIO-ARTIFICIAL LIVER CONTAINING HEPARG CELLS IN A THREEDIMENSIONAL CONFIGURATION AS AN IMPROVED MODEL FOR DRUG HEPATOTOXICITY AND METABOLISM STUDIES Vincent A. van der Mark1, Martien van Wenum1, Dirk R. de Waart2, David Steen3, Valery Shevchenko4, Robert A.F.M. Chamuleau1, Christophe Chesne´4 and Ruurdtje Hoekstra1 1 Department of Experimental Surgery, Academic Medical Center, Amsterdam, the Netherlands, 2Tytgat Institute for Liver and Intestinal Research, Academic Medical Center, Amsterdam, the Netherlands, 3R & D, Biopredic International, Rennes, France, and 4Biopredic International, Saint-Gre´goire, France Introduction: Primary human hepatocytes (PHH) are the gold standard for determining in vitro hepatotoxicity and drug metabolism. Unfortunately, PHH have a limited life span and proliferative capacity. The human hepatic progenitor cell line HepaRG can be used as an alternative for long-term studies of drug metabolism and hepatotoxicity. However, for an optimal expression of liver-specific drug metabolism genes in HepaRG monolayers the addition of the xenobiotic compounddimethylsulfoxide (DMSO) is required, which in turn may interfere with the normal metabolism of other xenobiotic compounds. In addition, DMSO treatment causes cytotoxicity, rendering the DMSO-treated HepaRG monolayers less suitable for hepatotoxicity screening. In the past we developed a 9 mL perfused oxygenated bioreactor. This bio-artificial liver (BAL) was loaded with HepaRG cells which organized in three dimensions and showed increased liver-specific drug metabolism. Here we explored the possibilities to downscale the BAL and utilize it as a human liver model for screening drug metabolism and toxicity. Methods: A 2.5 mL BAL was developed and loaded with 2  108 HepaRG cells. We assayed several liver-specific functions, expression of hepatic genes and metabolism of acetaminophen in the BAL cultures and in HepaRG monolayers, both cultured without DMSO

170

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

until optimal differentiation (414 days for BAL and at 28 days for monolayer). Results: Liver-specific functions were higher in BAL cultures compared to monolayers, including ammonia elimination (9-fold), urea production (2-fold) and CYP3A4 activity (23-fold). Lactate production in monolayers turned to elimination in BALs; a hallmark of PHH. When compared with PHH, hepatic functions were relatively high in BAL cultures; ammonia elimination was 530% of PHH, lactate elimination 36%, urea production 36% and CYP3A4 activity 108%. mRNA expression of several phase 1, 2 and 3 drug metabolism genes in the BAL was compared with freshly isolated human liver biopsies; CYP2C9: 34%, CYP2C19: 40%, CYP3A4: 60%, UGT1A1: 29%, TPMT: 33%, GSTA1: 170%, ABCC2: 142%, OATP1B1: 14%, OATP2B1: 26%. Two BALs were chronically treated with a daily dose of 1.0 or 2.0 mMacetaminophen, respectively. After 9 days we observed glucuronidated (419 mM or 647 mM, respectively), sulfated (176 mM or 241 mM, respectively) and glutathione-conjugated (0.3 mM or 0.5 mM, respectively) metabolites in the culture medium after 24 hours. After 10 days ammonia elimination decreased to 4 or 34%, respectively, when compared to a non-treated BAL. After 17 days aspartate transaminase levels were elevated by 20–54% in respectively 1.0 and 2.0 mM acetaminophentreated BALs. Conclusion: HepaRG cells grown in the 2.5 mL BALs outperform those cultured in monolayer and approach hepatic functionality of PHH. The BAL system is suitable for long-term hepatotoxicity studies as observed during chronic treatment of acetaminophen.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P280. CYTOTOXIC AND APOPTOTIC PROPERTIES OF THE HEXANE EXTRACT OF MANGIFERA ZEYLANICA BARK AND IDENTIFICATION OF PHYTOCHEMICAL PROFILE EdiriweeraP MK1, TennekoonK H1, Samarakoon SR1, Thabrew I1, and de Silva ED2 1 Institute of Biochemistry, Molecular Biology and Biotechnology, Colombo, Sri Lanka and 2Department of Chemistry, Faculty of Science, University of Colombo, Colombo, Sri Lanka Mangifera zeylanica (family: Anacardiaceae) is a plant endemic to Sri Lanka. It has been used in the indigenous system of medicine for the treatment of cancers. Apoptosis plays a crucial role in life by maintaining homeostasis of organs and tissues in concert with differentiation, proliferation and growth. As cancer results from uncontrolled cell growth associated with hindered apoptosis, research has been carried out to isolate apoptosis inducing agents from natural products to use as anti-cancer drugs. The present study was aimed to assess whether cytotoxic activity of the hexane extract of Mangifera Zeylanica is mediated through induction of apoptosis in triple negative (MDA-MB-231) and oestrogen receptor positive (MCF-7) breast cancer cells, and ovarian cancer cells (SKOV3). Cytotoxic effect on all three cancer cells was evaluated by Sulforhodamine B assay at 24 h post incubation with the hexane extract. Normal mammary epithelial cells were used as controls. Apoptotic effect was evaluated by assessment of Caspase 3 and Caspase 7 activity using ApoTox-GloÔ triplex assay, fluorescent microscopy of morphology of the acridineorange/ethidium bromide stained cells and DNA fragmentation analysis by agarose gel electrophoresis. The results demonstrated that the hexane extract exhibits cytotoxicity against ovarian cancer cells, oestrogen receptor positive breast cancer cells and triple negative breast cancer cells with the IC50 values of 86.6 mg/ml, 87.64 mg/ml and 116.5 mg/ml, respectively. Hexane extract showed less cytotoxicity to mammary epithelial cells (IC50 217.2 mg/ml) when compared to the cancer cells. Caspase 3 and Caspase 7 proteins were significantly (p50.001) activated by the hexane extract in ovarian cancer cells and triple negative breast cancer cells in a dose dependent manner whereas no Caspase 7 activity was seen in oestrogen receptor positive breast cancer cells after 4 h incubation. DNA fragmentation and morphological changes observed under fluorescent microscope further confirmed the apoptosis induced by the hexane extract in all three cancer cells. GC-MS analysis of the hexane extract identified that it is rich in sterols and long chain hydrocarbons, including apoptosis inducing beta-sitosterol and alfa-amyrin (Chai, 2008; Oluwatoyin et al., 2012). Overall findings demonstrated that hexane extract mediates Caspase 3 and 7 dependent apoptosis in ovarian and triple negative breast cancer cells whereas apoptosis of oestrogen receptor positive breast cancer cells is independent of Caspase 7 activation. Cytotoxic and apoptotic effects exerted by the hexane extract of M. Zeylanica on three different cancer cells thus support the use of the plant in traditional medicine as an anticancer treatment. This work was supported by the National Research Council of Sri Lanka (Grant No 11-018).

References Chai JW, Kuppusamy UR, Kanthimathi MS. (2008). Beta-sitosterol induces apoptosis in MCF-7 cells. Malay J Biochem Mol Biol 16:28–30. Oluwatoyin A, Olubunmi A, Oluyomi SA, Gabriel A. (2012). Antioxidant and cytotoxicity of b-amyrin acetate fraction from Bridelia ferruginea Leaves. Asian Pacific J Tropical Biomed 2:S981–S984.

P281. A DIY MODEL FOR GENERATION OF ROBUST 3D EPIDERMAL SKIN EQUIVALENTS COMPOSED OF NORMAL HUMAN PRIMARY EPIDERMAL KERATINOCYTES David T. Kuninger, Siddhita Gopinath, and Rhonda A. Newman Department of Cell Biology, Thermo Fisher Scientific, Frederick, MD, USA Three dimensional models composed of primary human cells enable in vitro modeling of complex human tissue and organ systems. In humans skin is the largest organ and is critical in maintaining an appropriate environmental barrier, wound healing and thermal regulation. Because of these properties physiologically relevant models of human skin are important in both basic

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

171

Downloaded by [112.134.238.102] at 07:30 10 February 2016

research and clinical applications. In addition, 3D organotypic models are increasingly being used to supplant animal testing in product development of cosmetic and consumer products, providing more accurate, reproducible, and cost effective solutions. Here, we present an ‘‘off the shelf’’ solution for the generation of 3D epidermal skin models composed of normal human primary epidermal keratinocytes derived from either adult or neonatal skin; pairing an optimized protocol with set of commercially available reagents. Experiments were conducted to modify the protocol published by Poumay et al. in 2004 to minimize donor to donor variation and provide workflow flexibility enabling direct seeding into 3D cell culture inserts with cryopreserved human epidermal keratinocytes (HEKs). Numerous parameters were evaluated, including different extracellular matrices, growth media and supplements, cell culture insert type and brand, cell seeding density, passage number as well as stratification conditions. From these experiments, we show that seeding of expanded neonatal HEK (HEKn) or adult HEK (HEKa) into NuncTM Cell Culture Inserts with Polycarbonate Membrane and cultured in EpiLife supplemented with Human Keratinocyte Growth Supplement in the presence of FGF7 Recombinant Human Protein, Ascorbic Acid, and CaCl2 resulted in consistent generation of 3D epidermal skin equivalents. Models produced using the optimized protocol display physiologically relevant morphology, displaying comparable number of cell layers and stratification to human epidermis. Basal cells are shown to express keratin 14, while suprabasal cells are shown to express keratin 10 and filaggrin. Furthermore, models were evaluated for their ability to correctly identify corrosivity and irritancy potential of a small panel of chemicals. Together, this protocol allows for generation of robust 3D skin model equivalents, which may be used for modeling of complex physiological processes in basic research.

P282. REPRODUCIBLE GENERATION OF 3D CORNEAL MODELS FROM NORMAL HUMAN LIMBAL DERIVED PROGENITOR CELLS Rhonda A. Newman and David T. Kuninger Department of Cell Biology, Thermo Fisher Scientific, Frederick, MD, USA The incorporation of primary human cells into 3-dimensional tissue models enables the in vitro recapitulation of complex human physiology, providing sophisticated biological models for research and product development. These types of tissue models can also be used to replace animal testing in many applications, which is coming under increased restriction for use in product development in the US and is banned in Europe for cosmetic and consumer product testing. Therefore, companies are moving toward the use of in vitro alternatives, such as 3D organotypic models that mimic the in vivo environment of human tissues to ensure compound safety and efficacy. For assessment of ocular irritancy there are currently three corneal models available, yet none of these models are derived from normal human primary corneal epithelial cells. Here, we present a reproducible system for generation of 3D models composed of normal human limbal progenitor cells. Using a series of Design of Experiments, follow-up titrations, and regression analyses we optimized key parameters for cell sourcing and expansion, growth medium, matrices, as well as selection of transwell culture inserts. Models produced using the optimized protocol display physiologically relevant morphology, with comparable number of cell layers and stratification to human corneal epithelia. Basal progenitor cells are shown to express p63alpha, while suprabasal cells are shown to express cytokeratins 3 and 12. Furthermore, expression of CD44v6, involved in intercellular junction formation, was shown to be present, and the apical surface was shown to express cell surface associated Mucin (MUC1). Compound irritancy screening using a panel of 26 chemicals demonstrated our corneal models effectively discriminate irritants from non-irritants, with 100% of the nine class I and 2 A irritants being identified, 60–100% of the five class 2B irritants being identified depending upon the donor. In order to ensure 3D model reproducibility and scalability, parameters were determined to screen for high quality donors and dissociated monolayer corneal cells. Selected donor material can be expanded to seed up to 20 000 twenty-four well inserts. Together the selection criteria and expansion combined with optimized seeding and stratification protocols enables the generation robust 3D corneal models composed of normal human primary corneal epithelial cells for assessment of ocular irritancy.

P283. IN VITRO BLOOD–BRAIN BARRIER MODEL USING RAT PRIMARY BRAIN ENDOTHELIAL CELLS AND ASTROCYTES Lei Ci, Guangqing Xiao, Elvana Veizaj, and Liang-Shang Gan Department of Drug Metabolism and Pharmacokinetics, Biogen Idec, Cambridge, MA, USA The blood–brain barrier (BBB) plays an important role in regulating the transport of endogenous substances and xenobiotic into the CNS. An in-vitro BBB model using rat primary brain endothelial cells and astrocytes was developed in order to understand the transport mechanism and to better predict BBB penetration. Rat primary brain endothelial cells were isolated from rat brains and cultured in the presence of puromycin for 3 days. The cells were then plated on transwell inserts and co-cultured with rat astrocytes that had already been seeded on the bottom of the transwell plate for 2 days prior. The co-culture conditions were optimized by monitoring the transendothelial electrical resistance (TEER) and the permeability of sucrose and digoxin across the cell monolayers. The expression of BBB tight junction proteins and brain transporters (e.g. Pgp) was analyzed with Western blot and immunofluorescent staining. Under the optimized culture conditions, the highest TEER was reached on da y 4 after the coculture, and expression of P-gp and tight junction proteins (ZO-1, Claudin-5 and Occludin) was observed. It was also found that

172

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

serum free condition could enhance the formation of the tight junctions. High TEER values were also associated with low sucrose permeability. Polarized transport of digoxin, quinidine and rhodamine-123 across the monolayers was observed, suggesting that Pgp was functional under the co-culture conditions. Evaluation of the expressions and transport activities of other brain transporters are under investigation. This in vitro BBB model can be an useful tool to study BBB properties and screen candidate compounds for neurotherapeutic projects.

P284. ASSESSMENT OF DRUG TOXICITY IN SMALL HISTOLOGICAL STRUCTURE OF THE EYE: APPLICATION OF MASS SPECTROMETRY IMAGING IN OPHTHALMIC CONTEXT

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Gregory Hamm, David Bonnel, Alain Heron, Fabien Pamelard, Raphael Legouffe, and Jonathan Stauber Imabiotech, Loos, France The eye is a highly complex tissue which consists of several small organs or structures. It can be basically divided into two regions, the anterior and the posterior parts, including the most important ocular structures. The anterior part consists of cornea, iridocorneal angle, trabecular meshwork, iris lens and aqueous humor. The posterior part contains the vitreous humor and the retina region (retina, sclera and choroid layers and the optic nerve, which is an extension of central nervous system). The two major eye diseases, Glaucoma and the Age-related Macular Degeneration (AMD), are closely correlated with these two histological regions of the eye. Mass spectrometry Imaging (MSI) applications to ophthalmic drug discovery have recently gained growing interest especially for preclinical studies in pharmacology or toxicology. In our study, MSI will be applied to assess the distribution and quantification of Benzalkonium chloride (BAK) compound (antiglaucoma eye drops preservative) in specific areas of the eye after instillation. The distribution of two BAK compounds (BAK C12 and BAK C14) were investigated in small specific histological regions of the eye (such as iridocorneal angle or sclera, choroid, retina regions) in order to estimate efficiency of action or adverse effects of the treatment. High spatial resolution images were performed at cells level (30 mm). Molecular distribution was also correlated to tissue histology using H&E staining or IHC. Local Drug concentration differences were observed according to histological area and position on the eye section (anterior, posterior, temporal or nasal side). In conclusion, MSI offers new insight in ocular therapeutic/pharmaceutical research, especially to give a better understanding of the drug candidate migration from the front to the back of the eye to assist drug efficiency or toxicity studies for specific tissue targeting eye diseases.

P285. PROTEIN TARGETS OF ISONIAZID REACTIVE METABOLITES: IMMUNOAFFINITY PRE-CONCENTRATION OF ADDUCTED PEPTIDES FOR PROTEOMIC ANALYSIS Yakov Koen1, Imir G. Metushi2, Jack P. Uetrecht3, Nadezhda Galeva1, and Robert P. Hanzlik1 1 Department of Medicinal Chemistry, University of Kansas, Lawrence, KS, USA, 2Department of Pharmacology and Toxicology, Faculty of Medicine, University of Toronto, Toronto, ON, Canada, and 3Faculty of Pharmacy, University of Toronto, Toronto, ON, Canada Isoniazid (INH) has been used for the treatment of tuberculosis since 1952. Appoximately 1% of INH patients experience idiosyncratic drug-induced liver injury (IDILI), but the incidence rises dramatically when INH is co-administered with rifampicin (a P450-inducer). The mechanisms underlying IDILI are unknown, but protein covalent binding (CVB) by chemically reactive metabolites is thought to be involved. CVB of INH to human liver microsomes was recently attributed to the acylation of lysine side chains by a diazohydroxide metabolite of INH, and a polyclonal antiserum to INA-lysine detected numerous INA-adducted proteins in livers of rats and mice treated with INH for five weeks (Metushi et al., Chem. Res. Toxicol. 25, 2567–2576, 2012). ‘‘Shotgun’’ proteomic analysis of liver S9 from INH-treated mice identified 466 individual proteins but only one INA-adducted peptide. To detect more INA-adducted peptides, immunospecific IgG was isolated from the antiserum and covalently attached to tosyl-activated M280 magnetic beads. The beads were then exposed to tryptic digests of lysozyme chemically modified with INA (INA-lysozyme), washed and eluted. LC-MS/MS analysis of the eluate showed 7 adducted peptides representing 5 adducted lysine residues (of 6 possible). No unadducted peptides were observed in the eluate. To mimic the expected low levels of adduction in biological samples, the INA-lysozyme digest was diluted 10- and 100-fold with a digest of native lysozyme and analyzed similarly. In all cases all five lysine positions were observed as INA-adducted peptides. When the S9-fraction from INH-treated mice (INH-S9)was analyzed similarly, 12 peptides were found in the Ab-M280 bead eluate (out of 4900 peptides observed in the whole unfractionated INH-S9 digest). Five peptides showed the presence of INA-adducts, and these peptides were assigned to five mouse proteins to a high degree of confidence. We also applied the undigested S9 fraction to Ab-M280 beads, eluted with 5 mM INH and digested the eluted proteins with trypsin. LC-MS/MS of this digest showed 4125 peptides (including two adducted peptides) representing 26 identifiable proteins. Since 11 of these proteins were also observed in S9 from untreated mice, we presume that the other 15 were true target proteins. Thus, using three analytical approaches, we observed eight INA-adducted peptides representing eight unequivocal target proteins, and indirect evidence for 13 other target proteins based on selective adsorption to Ab-M280 beads. The results suggest that Ab-M280 beads may be a useful tool for increasing sensitivity of analysis of INH-target proteome.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

173

This work was supported in part by NIH-GM21784 and Canadian Institutes for Health Research.

P286. METABOLIC ACTIVATION OF NEVIRAPINE AND INACTIVATION OF ITS ACTIVE METABOLITE BY GLUTATHIONE CONJUGATION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kenichiro Ogura, Tomohiro Ishii, Nanami Iwata, Asahi Miyazaki, Tomokazu Ohnuma, Takahito Nishiyama, and Akira Hiratsuka Drug Metabolism and Molecular Toxicology, Tokyo University of Pharmacy and Life Sciences, Hachioji-shi, Tokyo, Japan Purpose: Nevirapine (NVP), a non-nucleoside reverse transcriptase inhibitor that is widely used for the treatment of HIV infections. NVP is associated with serious clinically restrictive side effects such as skin reactions and hepatotoxicity. Although the mechanisms for the expression of adverse effects of NVP are still unclear, the involvement of metabolic activation to reactive electrophiles is supported by recent studies. Formation of a sulfate ester of 12-hydroxy-NVP (12-OH-NVP), one of P-450mediated oxidative metabolites of NVP, is supposed as one of the mechanisms of metabolic activation of NVP. In the present study, we determined human sulfotransferase (SULT) and glutathione S-transferase (GST) isoforms involving sulfation of 12OH-NVP and GSH conjugation of 12-OH-NVP sulfate. Methods: Human SULT isoforms. SULT1A1, 1A3, 1B1, 1C1, 1E1, 2A1 and 2B1, and human GST isoforms, hGSTA1-1, A2-2, A4-4, M1b-1b, M2-2, M4-4, P1-1, T1-1, T2-2, K1-1, O1-1, O2-2, Z1-1 were expressed in E. coli and purified for determination of enzymatic activities. Formation of 12-OH-NVP sulfate and GSH conjugate (12-SG-NVP) was analyzed by liquid-chromatography-mass spectrometry with LCMS-IT-TOF. Results and discussion: Of these SULT isoforms, SULT1A1, SULT1E1 and SULT2A1 showed sulfonating activity toward 12-OH-NVP. Enzymatically formed 12-OH-NVP sulfate was identified with a synthetic specimen by LCMS-IT-TOF. Kinetic analysis of the reaction indicated that SULT2A1 most efficiently catalyzed the sulfation of 12-OH-NVP. This result may follow a tendency of substrate specificity of SULT2A1 that has been recognized as alcohol or hydroxysteroid SULT. 12-OH-NVP sulfate nonenzymatically reacted slowly with GSH to form the conjugate 12-SG-NVP. Out of those tested GST isoforms, only human GSTT2-2 catalyzed GSH conjugation of the sulfate. This substrate specificity was in good accordance with the previous demonstration that hGSTT2-2 catalyzed GSH conjugation of reactive sulfate esters of carcinogenic arylmethanols, such as 5hydroxymethylchrysene and 7-hydroxymethyl-12-methylbenz[a]anthracene. Further kinetic analysis of the GSH conjugation is now in progress.

P287. IMPACT OF CYP3A5 GENETIC POLYMORPHISM ON MECHANISM-BASED INACTIVATION BY LAPATINIB James Chun Yip Chan, Devester Yun Ming Choo, and Eric Chun Yong Chan Department of Pharmacy, National University of Singapore, Singapore, Singapore Mechanism-based inactivation (MBI) of CYP450 enzymes results in irreversible and permanent loss of enzyme activity, increasing the risk of drug–drug interactions (DDI), leading to subtherapeutic effects or unexpected toxicity with the victim drug. Lapatinib is a dual tyrosine kinase inhibitor used to treat metastatic breast cancer; however it exhibits idiosyncratic hepatotoxicity. Primarily metabolized by CYP3A4/5, lapatinib is oxidized to a reactive quinoneimine intermediate suggested to cause MBI of CYP3A5. CYP3A5 is polymorphically expressed, where *1/*1 carriers possess CYP3A5 levels up to 50% of total CYP3A content, while *3/*3 carriers have negligible CYP3A5. While numerous efforts have been made to rationalize the structural features of xenobiotics that confer increased chemical reactivity, little is known about the biological determinants that may influence the vulnerability of CYP450 to inactivation. This study aims to explore the impact of CYP3A5 polymorphism in influencing individual susceptibility to lapatinib-induced CYP3A4/5 inactivation. In vitro inactivation of CYP3A4/5 by lapatinib was investigated in a panel of 12 CYP3A5-genotyped, single donor human liver microsomes (HLM). Lapatinib (0–75 mM) was pre-incubated with 0.6 mg/mL HLM for 0–30 minutes using a two-step incubation design, and residual enzyme activity determined using testosterone as the probe substrate to obtain the inactivation kinetics parameter kinact/KI. Reactive metabolite trapping studies using glutathione (GSH) as a trapping agent were performed in the same panel of HLM to determine the extent of quinoneimine formation as measured by lapatinib-derived GSH adduct levels. A two-fold higher kinact/KI ratio was observed which suggested higher potency of inactivation within CYP3A5 *1/*1 versus *3/*3 carriers (6.88 ± 2.09 and 3.57 ± 1.92 min1mM-1 respectively). Mean peak area ratios of lapatinib-derived GSH adduct were also two-fold higher in *1/*1 versus *3/*3 carriers (2.97 ± 1.11 and 1.27 ± 1.55 respectively, p40.05). Both these parameters correlated positively with testosterone 6Bhydroxylation activity, indicating that these observations are mediated primarily by CYP3A4/5. Although not statistically significant, the results suggest that CYP3A5 *1/*1 carriers generate elevated levels of reactive metabolite and may experience a greater magnitude of lapatinib-induced inactivation. Importantly, this is the first attempt to elucidate the influence of CYP450 polymorphisms on modulating the susceptibility of an individual to CYP450 inactivation and the ensuing clinical consequences, and the results highlight the multidimensional impact of CYP450 polymorphisms beyond affecting the clinical efficacy of xenobiotics.

Reference Chan EC, New LS, Chua TB, et al. (2012). Interaction of lapatinib with cytochrome P450 3A5. Drug Metab Disp 40:1414–1422.

174

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P288. CHARACTERIZATION OF CYP-MEDIATED METABOLIC PATHWAYS INVOLVED IN BENZBROMARONEINDUCED HEPATOTOXICITY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Minaka Shibuya, Masayuki Ishikawa, Takayuki Aizawa, Naoki Cho, Tomomi Furihata, Kaoru Kobayashi, and Kan Chiba Laboratory of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba-shi, Japan Benzbromarone (BBR), a uricosuric agent, has the potential to cause fatal hepatic failure in human beings. While the mechanisms underlying BBR-induced hepatotoxicity have yet to be fully clarified, cytochrome P450 (CYP) is assumed to be involved through BBR’s bioactivation to a reactive metabolite. Consistent with this, it has been shown that BBR is primarily metabolized to 6-hydroxy BBR (6-OH BBR) and 10 -hydroxy BBR (10 -OH BBR) by CYP2C9 and CYP3A4, respectively. However, the roles of CYP2C9 and CYP3A4 in BBR-induced cytotoxicity remain inconclusive. Therefore, in this study, we aimed to clarify the involvement of CYP2C9- and CYP3A4-mediated pathways in the cytotoxic potential of BBR in vitro. We began our study by developing human hepatoma FLC7 cells stably expressing CYP2C9, CYP3A4 or green fluorescent protein (GFP) (hereafter referred to as CYP2C9/FLC7, CYP3A4/FLC7, and GFP/FLC7, respectively). To examine CYP2C9 and CYP3A4 activities in these cells, diclofenac 4-hydroxylase and triazolam a-hydroxylase activities were measured by high performance liquid chromatography (HPLC). As expected, CYP2C9/FLC7 showed the highest diclofenac 4-hydroxylase activity levels among these cells, while CYP3A4/FLC7 showed the highest triazolam a-hydroxylase activity levels. Next, in order to examine the cytotoxic effects of BBR, the cells were treated with BBR for 24 hours in the glutathione-depleted condition, after which cell viability was determined by MTS assay. The results showed that BBR treatment caused a significant decrease in the viability of CYP2C9/FLC7 and CYP3A4/FLC7, but did not impact on GFP/FLC7, thus suggesting that both CYP2C9 and CYP3A4 were individually involved in BBR-induced cytotoxicity. Next, in order to advance characterization of CYP2C9- and CYP3A4-mediated pathways, we examined the cytotoxic effects of 6-OH BBR and 10 -OH BBR in those cells. The results showed that while 6-OH BBR treatment exclusively decreased the viability of CYP3C4/FLC7, 10 -OH BBR treatment caused a greater loss of cell viability in CYP2C9/FLC7 than in CYP3A4/FLC7 or GFP/FLC7. Therefore, it is considered likely that any functional collaboration between CYP3A4 and CYP2C9 would lie in the BBR-mediated cytotoxicity process. Because the effect of 10 -OH BBR on CYP2C9/FLC7 was much more extensive than that of 6-OH BBR on CYP3A4/FLC7, we sought to identify an 10 -OH BBR metabolite formed by CYP2C9 using electrospray ionization () liquid chromatography-tandem mass spectrometry (ESI-LC-MS/MS). The results showed that the previously unidentified BBR metabolite existed in the culture medium of 10 -OH BBR-treated CYP2C9/FLC7. Based on the results of its fragmentation profile, the metabolite was assumed to be the 10 ,6dihydroxy BBR. In conclusion, our results show that both CYP2C9 and CYP3A4 are independently and coordinately involved in BBR bioactivation, presumably through hydroxylation activity at positions 6 and 10 , respectively. Furthermore, our results are the first to demonstrate that the 10 -OH BBR metabolite (likely 10 , 6-dihydroxy BBR) was formed by CYP2C9, which may be a critical pathway to BBR-induced hepatotoxicity.

P289. IDENTIFICATION OF DRUG-INDUCED MITOCHONDRIAL ALTERATIONS USING THE HEPARGÕ CELL LINE David Steen1, Sandrine Camus1, Christophe Chesne2, Nelly Burton3, Mathieu Porceddu3, Claire Pertuiset3, and Annie Borgne-Sanchez3 1 R&D, Biopredic international, Rennes, France, 2CEO, Biopredic international, Rennes, France, and 3R&D, MITOLOGICS SA, Paris, France Drug-induced liver injury (DILI) is difficult to predict using classical in vitroassays and animal studies. To improve early prediction of DILI in humans, we developed a medium-throughput screening assay using isolated mouse liver mitochondria designed to detect loss of mitochondrial function and integrity. We next hypothesized that the human hepatic cell line HepaRG could be used for identification of drug-induced mitochondrial dysfunction involved in DILI in comparison to the classically used hepato-carcinoma cell line HepG2 cultured in glucose or galactose. We determined that the HepaRG cells’ CYP and mitochondrial expression profiles closely resemble those of primary human hepatocytes while those of HepG2 cells were much different, which gave us confidence that the HepaRG cells would exhibit functional advantages in detecting risk of mitochondrial toxicity from known toxicants. So we then investigated mitochondrial toxicity of Acetaminophen, Valproic acid and Zidovudine in HepaRG and HepG2 cells by measuring mitochondrial respiration, transmembrane potential and mtDNA content after a treatment of 2–12 days, as well as biotransformation of those compounds. Metabolic and mitochondrial functionality differed between the cell models, and importantly drug sensitivity differed as well, with the results derived from the HepaRG cells closer to the expected results for these well-studied agents. Our findings led us to conclude that the HepaRG cells may represent a valuable predictive model of human hepatoxicity due to mitochondrial damage.

P290. STRAIN AND SEX DIFFERENCES IN CARBON TETRACHLORIDE-INDUCED LIVER INJURY IN MICE Yasuhiro Masubuchi1 and Ayaka Ihara2 1 Faculty of Pharmaceutical Sciences, Chiba Institute of Science, Choshi, Chiba, Japan and 2Chiba Institute of Science, Choshi, Chiba, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

175

Strain and sex differences in drug-induced liver injury (DILI) are important information not only for the development of a suitable animal model for DILI but also for the identification of susceptibility factors of DILI. Carbon tetrachloride (CCl4) has been widely used to induce liver injury in various animal species such as mice, but limited information is available about the susceptible mouse strain and sex. In the present study, CCl4 hepatotoxicity was evaluated in different strains of male and female mice, along with hepatic cytochrome P450-depnendent enzyme activities and expression of inflammatory mediators. Two inbred (C57BL/6 and BALB/c) and one outbred (CD-1) mice of both sexes were used in this study. The mice were given CCl4 intraperitoneally at the doses of 0.5–10 mmol/kg. Hepatotoxicity was assessed by serum leakage of alanine aminotransferase (ALT) and histochemical analysis. Lipid peroxidation was measured as thiobarbituric acid-reactive substances (TBARS). Glutathione (GSH) levels were determined by the dithionitrobenzoic acid-glutathione disulfide reductase recycling assay. Microsomal oxidation activities were measured by HPLC as indices of specific P450 isoforms. Hepatic expression of cytokines and other inflammatory mediators were assayed by real-time RT-PCR. Treatment of male CD-1 mice with CCl4 resulted in higher serum ALT leakage than female mice in all dose groups, and the most prominent sex difference was observed at the lowest dose. The results were supported by the hematoxylin/eosin staining of liver sections. Similar sex differences in the CCl4 hepatotoxicity were observed in C57BL/6 and BALB/c mice. Interstrain comparison demonstrated that BALB/c mice were highly susceptible to the low dose of CCl4 as compared to C57BL/6 and CD-1 mice. CCl4 administration resulted in elevation of hepatic TBARS and decline in GSH levels only at the highest CCl4 dose, which may not be involved in pathogenesis of male-dominant toxicity by the low dose of CCl4. No major strain or sex differences were observed in hepatic microsomal CYP1A2, CYP2E1 or CYP3A11 activities, but those of CYP2D9 in male mice were higher than in female mice. Various gene expressions such as tumor necrosis factor-a, interleukin-6, heme oxygenase-1 were induced prior to induction of the hepatotoxicity, but the high susceptibility in males and in BALB/c mice could not be explained only by a single mediator. In conclusion, this study demonstrated major strain and sex differences in CCl4-induced liver injury, which is indicative of the involvement of multiple susceptibility genes. Particularly, male mice are highly susceptible, alternately suggesting that female mice can have estrogen-dependent protective processes against the liver injury.

P291. OAT-DEPENDENT CELLULAR ACCUMULATION OF P-CRESYL SULFATE, A UREMIC TOXIN, CAUSES VASCULAR ENDOTHELIAL AND SMOOTH MUSCLE CELL DAMAGES BY INDUCING OXIDATIVE STRESS Hiroshi Watanabe1, Yohei Miyamoto1, Yuki Enoki1, Yu Ishima1, Masafumi Fukagawa2, Masaki Otagiri3, and Toru Maruyama1 1 Department of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Kumamoto University, Kumamoto, Japan, 2 Division of Nephrology, Endocrinology and Metabolism, Tokai University School of Medicine, Kanagawa, Japan, and 3Faculty of Pharmaceutical Sciences Sojo University, Kumamoto, Japan Purpose: p-Cresyl sulfate (PCS), a uremic toxin, is considered to be a risk factor for cardiovascular disease in chronic kidney disease. However, our understanding of the vasucular toxicity induced by PCS and its mechanism is incomplete. The purpose of this study was to determine whether OAT-dependent cellular accumulation of PCS enhances the production of reactive oxygen species (ROS) in vascular endothelial and smooth muscle cells, resulting in cytotoxicity. Methods:The vascular toxicity of PCS was evaluated using human umbilical vein endothelial cells (HUVEC), human aortic smooth muscle cells (HASMC) and PCSoverloaded 5/6-nephrectomized rats. The 5/6-nephrectomized rats received an intraperitoneal injection with PCS at a dose of 50 mg/kg/day for 4 weeks. Results: PCS significantly increased ROS production in HUVEC and HASMC. The enhanced ROS production was slightly reduced in the presence of 4% HSA (80% (in the case of HUVEC) and 90% (in the case of HASMC)) compared to the absence of HSA. The ROS induced by PCS was significantly inhibited in the presence of various inhibitors, such as an NADPH oxidase inhibitor (DPI), a PI3K inhibitor (wortmannin) and a PKC inhibitor (calphostin C), suggesting the contribution of NADPH oxidase system. PCS also up-regulated the mRNA levels and the protein secretion of MCP-1 in HUVEC. In HASMC, PCS increased the mRNA levels of osteoblast specific protein such as ALP, osteopontin, Cbfa1 and ALP activity. The knockdown of Nox4, a subunit of NADPH oxidase, suppressed the cell toxicity induced by PCS. These vascular damage induced by PCS was largely suppressed in the presence of probenecid, an inhibitor of OAT. In PCS-overloaded 5/6-nephrectomized rats, plasma MCP-1 levels, osteopontin expression and ALP activity of the aortic arch were increased, accompanied by the induction of Nox4 expression. Collectively, the vascular toxicity of PCS can be attributed to its intracellular accumulation via OAT, which results in an enhanced NADPH oxidase activity and increased ROS production. Conclusion: PCS could play an important role in the development of cardiovascular disease by inducing the dysfunction of vascular systems.

P292. SYSTEMS PHARMACOLOGY MODELING PREDICTS HEPATOTOXIC POTENTIAL OF TROGLITAZONE AND PIOGLITAZONE Kyunghee Yang1, Jeffrey L. Woodhead2, Paul B. Watkins2, Brett A. Howell2, and Kim L. R. Brouwer3 1 Division of Pharmacotherapy and Experimental Therapeutics, The University of North Carolina at Chapel Hill Eshelman School of Pharmacy, Chapel Hill, NC, USA, 2Hamner-University of North Carolina Instititute for Drug Safety Sciences, Reseach Triangle Park, NC, USA, and 3Division of Pharmacotherapy & Experimental Therapeutics, The University of North Carolina at Chapel Hill Eshelman School of Pharmacy, Chapel Hill, NC, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

176

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Troglitazone (TGZ) caused delayed and life-threatening drug-induced liver injury (DILI) in diabetic patients, whereas the next in class, pioglitazone (PGZ), has rarely been associated with DILI. Inhibition of bile acid transport, which may result in accumulation of toxic bile acids in hepatocytes, is one proposed mechanism of TGZ-mediated hepatotoxicity. However, PGZ is a more potent inhibitor of the bile salt export protein (BSEP) than TGZ based on in vitro membrane vesicle transport studies. In the current study, the hepatotoxic potential of TGZ and PGZ due to interference with bile acid homeostasis was investigated using DILIsymÕ , a systems pharmacology model incorporating drug/metabolite disposition, bile acid physiology and pathophysiology, the hepatocyte life cycle, and liver injury biomarkers. Experimentally measured inhibition constants of TGZ, TGZ-sulfate, and PGZ for multiple bile acid transport proteins were employed to simulate the altered bile acid disposition and subsequent DILI in humans. To simulate the DILI responses at the population level, a human virtual population (SimpopsTM) that included variability in key model parameters describing bile acid and drug disposition, body weight, and sensitivity of ATP synthesis to hepatic bile acid accumulation was constructed. In the SimpopsTM, administration of 400–600 mg/day TGZ for 6 months resulted in delayed increases in serum alanine transaminase (ALT)43X upper limit of normal (ULN) in 2.4–4.2% of the population, with concomitant elevations in serum bilirubin42X ULN in 0.9–2.7% of the population. The simulated time to peak ALT was 108 ± 61 days. These results were similar to observations from the clinical trials where 200–600 mg/day TGZ elicited serum ALT elevations43X ULN in 1.9% of treated patients with time to peak ALT of 147 ± 86 days. Multiple regression analysis indicated that the maximum rate (Vmax) of lithocholic acid (LCA)-sulfate biliary excretion, Vmax of LCA synthesis in the intestinal lumen, the scaling factor governing the magnitude of farnesoid X receptor-mediated feedback regulation of canalicular bile acid efflux, the Vmax of chenodeoxycholic acid (CDCA) amide biliary excretion, biliary clearance of TGZsulfate, body weight, and the Vmax of CDCA amide basolateral efflux were significant predictors of TGZ-mediated DILI in the current model. No hepatotoxicity was predicted in the SimpopsTM after administration of clinically relevant doses of PGZ (15–45 mg/day) for 6 months, consistent with the clinical observations. In summary, mechanistic modeling based only on bile acid homeostasis adequately predicted the incidence and delayed presentation of TGZ hepatotoxicity, and correctly predicted lack of PGZ hepatotoxicity. These results demonstrate the utility of systems pharmacology models that integrate physiology and experimental data to evaluate DILI mechanisms and identify potential risk factors for DILI. Importantly, these mechanistic models may be useful to prospectively predict the hepatotoxic potential of new drug candidates. This work was supported by members of the DILI-sim Initiative and NIH R01GM41935

P293. DRUG-INDUCED LIVER INJURY CAUSED BY ENDOTHELIN RECEPTOR ANTAGONISTS: CONTRIBUTION OF MULTIPLE COMPOUND RELATED ADVERSE PROPERTIES Richard A. Thompson1, Gerry Kenna2, Simone Stahl3, Julie A. Eakins4, Alison Foster5, Linda Andersson1, and Marie Elebring6 1 Respiratory, Inflammation & Autoimmunity iMed DMPK, AstraZeneca R&D Molndal, Mo¨lndal, Sweden, 2Drug Safety Consultant, Cheshire, England, 3DMPK, Drug Safety and Metabolism, AstraZeneca, Macclesfield, Cheshire, UK, 4Department of Discovery Safety, Drug Safety and Metabolism, AstraZeneca, Macclesfield, UK, 5 Department of Translational Safety, Drug Safety and Metabolism, AstraZeneca, Macclesfield, UK, and 6DMPK, Cardiovascular & Metabolic Diseases iMed, AstraZeneca R & D-Molndal, Molndal, Sweden Drug-induced liver injury (DILI) has been observed in patients undergoing treatment for pulmonary arterial hypertension with the endothelin receptor antagonists (ERTA) sitaxentan and bosentan, but not following treatment with ambrisentan. Several cases of acute liver failure were seen with sitaxentan leading to its withdrawal in 2010, whereas bosentan has a FDA black box warning. We have previously shown the value of a multi-hazard approach in assessing the potential for adverse drug reactions (Thompson et al., 2012) and the aim of the present study was to use this approach to assess the possible role played by multiple contributory mechanisms in clinically relevant toxicity such as seen with the ERTAs. In the present study, inhibition of the Bile Salt Export Pump (BSEP) and Multidrug Resistance-associated Protein 2 (MRP2) was quantified using membrane vesicle assays. Inhibition of mitochondrial respiration in human liver-derived HuH7 cells was determined using a Seahorse XFe96 analyzer. Cytochrome P450 (CYP)-independent and CYP-mediated cell toxicity was assessed using transfected SV40-T-antigenimmortalized human liver epithelial (THLE) cells lines. Exposure adjusted assay ratios were calculated by dividing the maximum human drug plasma concentrations (Cmax,ss) by the IC50 or EC50 values obtained in vitro. Covalent binding (CVB) of radiolabelled drugs to human hepatocyte proteins was quantified and CVB body burdens were calculated by adjusting CVB values for fractional drug turnover in vitro and daily drug dose. Sitaxentan exhibited positive exposure adjusted signals in all 5 in vitro assays and a high CVB body burden. Bosentan, at its therapeutic dose, exhibited a positive exposure adjusted signal in one assay (BSEP inhibition) and a moderate CVB body burden. Ambrisentan exhibited no positive exposure adjusted assay signals and a low CVB body burden. The data indicate that multiple mechanisms could contribute to the rare but severe liver injury caused by sitaxentan in humans, provide a potential rationale for the markedly lower propensity of bosentan to cause liver injury, and highlight the relative safety of ambrisentan.

Reference Thompson RA, Isin EM, Li Y, et al. (2012). In vitro approach to assess the potential for risk of idiosyncratic adverse reactions caused by candidate drugs. Chem Res Toxicol 25:1616–1632.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

177

P294. REFINING CONCRETE MULTISCALE CASCADES TO EXPLAIN MULTIPLE ASPECTS OF ACETAMINOPHEN HEPATOTOXICITY AND ITS INHIBITION

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Andrew K. Smith1, Glen E.P. Ropella2, and C. Anthony Hunt1 1 Department of Bioengineering and Therapeutic Sciences, University of California, San Francisco, San Francisco, CA, USA and 2 Tempus Dictum, Inc., Milwaukie, OR Our goal is to validate concrete, plausible in silico explanations for how hepatic zonations of networked intracellular mechanisms orchestrate to cause the characteristic features of acetaminophen (APAP) induced liver injury in mice during APAP exposure. We instantiate and challenge mechanistic hypotheses capable of generating the characteristic spatial and temporal injury patterns. We use discrete event simulation and Monte Carlo sampling of agent-based, multiscale, models. We refer to them as biomimetic analogs. They are designed to evolve iteratively. Our five-stage approach combines the scientific method and good software engineering practices. (1) Assemble a large, diverse set of ‘‘targeted attributes’’ (TAs) that we eventually wish to explain. Specify the subset that will be the validation targets for the current work. (2) Formulate in silico mechanisms. Phenomena to be generated during execution will be products of temporal component interactions. (3) Refactor and expand both components and mechanisms of already validated in silico analogs and instantiate the formulated mechanisms. Record and measure phenomena generated during in silico experiments. (4) Compare simulated and literature-reported results using quantitative and qualitative ‘‘similarity measures’’ (SMs). When several SM criteria are achieved, the analog has achieved a degree of validation and can stand as a plausible, concrete, valid explanation of the targeted phenomena. (5) Challenge/falsify that Stage 4 mechanism by moving one or more TAs from the large Stage 1 set to the subset of validation targets. We cycle through this Iterative (mechanism) Refinement Protocol until all validation targets (new + previous) are achieved. This work focuses on results of Stage 5 experiments. Stage 4 validation targets already achieved at different granularities include APAP pharmacokinetics, hepatic clearance, and metabolite ratios; lobule zonation, zonation of metabolizing enzymes, glutathione depletion, damage caused by the reactive metabolite, repair of damage, and hepatocyte necrosis occurring first adjacent to the central vein. For this work we added dynamic attributes characterizing degrees of inhibition of hepatocyte necrosis caused by a single dose of SP600125 (1,9-pyrazoloanthrone) given two hours after a toxic APAP dose. In one in silico experiment, thousands of objects (each maps to an amount of APAP) percolate through analog lobules and cells. Simulating inhibition by SP600125 is challenging because so doing requires 1) a separate set of SP600125 objects that can percolate simultaneously and 2) giving analog components from mouse level to intracellular the ability to distinguish between APAP and SP600125, and autonomously adjust their interaction protocols to one, the other, or both. Achieving expanded sets of validation targets forces analog mechanisms to become increasingly fine grain and thus more cogent. We demonstrate achieving quantitative SP600125 absorption, distribution, and clearance validation targets as well as toxicity inhibition targets. So doing strengthens our case that analog mechanisms stand as a plausible, concrete, valid explanation of how APAP toxicity emerges within and across biological scales.

P295. COMPARISON OF EPITHELIAL–MESENCHYMAL TRANSITION INDUCED BY TGF-BETA1 AND METHOTREXATE IN HUMAN ALVEOLAR EPITHELIAL CELL LINE Masashi Kawami, Mioka Miyamoto, Ryoko Yumoto, and Mikihisa Takano Department of Pharmaceutics and Therapeutics, Graduate School of Biomedical & Health Sciences, Hiroshima University, Hiroshima, Japan Methotrexate (MTX) is a drug widely used for the treatment of certain cancers as well as rheumatoid arthritis. Unfortunately, it often induces serious interstitial lung diseases such as pulmonary fibrosis. Recently, it has been recognized that epithelialmesenchymal transition (EMT) plays an essential role in the fibrosis of various organs including lungs. Transforming growth factor-b1 (TGF-b1) is a potent fibrogenic cytokine and induces EMT via its signaling cascade. However, the mechanism underlying MTX-induced lung injury remains unclear. Therefore, in this study, we examined the effect of MTX on the phenotype of cultured alveolar epithelial cell line A549, and compared with that of TGF-b1, focusing on EMT-like changes of the cells. A549 cells were treated with TGF-b1 (10 ng/mL) for 72 hours or with MTX (300 nM) for 144 hours. Morphological changes of A549 cells induced by TGF-b1 and MTX were observed by phase-contrast microscopy and confocal laser scanning microscopy after staining the actin filament and nucleus with fluorescent probes. The expression level of epithelial and mesenchymal marker mRNAs was measured by real-time PCR analysis, and phosphorylation of Smad2, the protein involved in TGF-b1 signaling cascade, was analyzed by Western blotting. Both TGF-b1 and MTX induced morphological changes in A549 cells from epithelial-like morphology to elongated fibroblast-like morphology. Remodeling of actin filaments was also observed in A549 cells treated with these compounds. In addition, mRNA expression of epithelial markers such as cytokeratin 19 was decreased, while that of mesenchymal markers such as a-smooth muscle actin was increased by TGF-b1 and MTX. Like TGF-b1, phosphorylation of Smad2 was enhanced by MTX treatment, suggesting that EMT-like phenotypic changes similar to those by TGF-b1 were induced by MTX. In order to clarify the role of TGF-b1 in MTX-induced EMT, the effect of SB431542, a specific inhibitor for TGF-b receptor kinase, was examined. SB431542 tended to suppress not only the morphological changes of the cells but also altered expression of epithelial and mesenchymal marker mRNAs induced by MTX treatment. Interestingly, the expression of TGF-b1 mRNA was increased by the treatment of A549 cells with MTX. In conclusion, MTX would induce EMT

178

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

similar to that induced by TGF-b1 in A549 cells, which may be related to the in-vivo lung injury after MTX administration. In addition, MTX-induced EMT may partly be mediated by TGF-b1 produced in the alveolar epithelial cells.

P296. KINECTIN REGULATES GAP JUNCTION PROTEINS IN HEPATOCYTE-IMPLICATIONS IN RESPONSE TO HEPATOTOXICITY

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Pornteera Pawijit1,2, Yu Yu2, Yan Zhou3, and Hanry Yu2 1 NUS Graduate School for Integrative Sciences and Engineering, National University of Singapore, Singapore, Singapore, 2 Department of Physiology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore, and 3 Mechanobiology Institute, National University of Singapore, Singapore, Singapore Kinectin is a ubiquitous endoplasmic reticulum (ER) membrane protein expressed in almost every cell in the body. It has been implicated in many diseases, including hepatocellular carcinoma, but with our current understanding of kinectin, we are unable to comprehend how kinectin might be involved. From in vitro studies, it has been shown that kinectin binds mainly to kinesin, an anterograde motor protein and elongation factor-1d, a protein involved in protein synthesis. The few functions of kinectin uncovered include the importance of kinectin in the dynamics of the ER, transportation of organelles along microtubules, anchoring of elongation factor to the ER and the role of ER in focal adhesion maturation. However, how all these in vitro functions translate in vivo is what we would like to find out. Here, we show that kinectin plays a protective role in thioacetamide-induced liver injury. Kinectin knockout mice (KTN/) show more pronounced liver injury in comparison to the wildtype (WT) mice, which is reflected in the higher AST and ALT levels (1.5 fold of WT), as well as the higher degree of centrilobular necrosis. This higher susceptibility to thioacetimide-induced liver damage is due to the deregulation of the gap junction proteins, channels that allow solutes to pass between cells. In response to acute liver injury, there is usually a reduction in gap junction protein, which would limit the area of damage. However, in KTN/ hepatocytes, there is no such reduction, leading to uncontrolled propagation of injury. How kinectin regulates levels of gap junction proteins is not yet understood. Understanding this mechanism might help illuminate hepatocyte protective mechanisms in drug-induced liver injury. Does the role that kinectin play in this setting the same as the already known functions of kinectin? Or will it be a function that has not been discovered before?

P297. CYTOTOXICITY AND APOPTOTIC EFFECTS OF MICROCYSTIN-LR ON HUMAN EMBRYONIC KIDNEY (HEK 293) AND HUMAN KIDNEY ADENOCARCINOMA (ACHN) CELL LINES M. A. P. C. Piyathilaka1, M. M. Pathmalal1, K. H. Tennekoon2, B. G. D. N. K. De Silva1, S. S. R. Samarakoon2 and S. Chanthirika2 1 Department of Zoology, University of Sri Jayewardenepura, Nugegoda, Sri Lanka and 2Institute of Biochemistry, Molecular Biology, and Biotechnology, University of Colombo, Colombo, Sri Lanka Cyanobacterial toxins, especially the microcystins, are found in eutrophied waters all over the world and their potential to impact on human and animal health is a cause for concern. Microcystin-LR (MC-LR) is one of the common toxic microcystin congeners and occurs frequently in diverse water systems. Recent work suggested that apoptosis plays a major role in the toxic effects induced by MC-LR in hepatocytes. Although, MC-LR has been reported as a potent hepatotoxin, there is now increasing evidence implicating it as a nephrotoxin. Possible nephrotoxicity of MC and whether apoptosis mediates such a toxic effect is poorly understood. Present study was carried out to evaluate cytotoxic and apoptotic effects of MC-LR on human embryonic kidney cells (HEK-293) and human kidney adenocarcinoma cells (ACHN). Cells were treated with MC-LR (1.0–200 mM) for 24 h and cytotoxicity was evaluated by 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and Sulphorhodamine B (SRB) assays. MC-LR induced apoptotic effects were evaluated by (a) assessing morphological changers by light and fluorescence microscopy (b) estimating expression of selected apoptotic related genes (P53, Bax, and Survivin) by real time PCR analysis and (c) estimating level of caspase 3 and caspase 9. Results of MTT and SRB assays showed that MC-LR induced cytotoxicity in both cell lines at 24 h post-incubation. Morphological changes observed under light and fluorescence microscopes confirmed that MC-LR induced apoptosis in both cell types. Exposure to a higher concentration of MC-LR (10 mM) up-regulatedSurvivin gene expression in both cell lines, while a 10 fold lower concentration of MC-LR down regulated Survivin expression. Furthermore, activity of caspase 3 and 9 which mediate induction of apoptosis significantly (p50.001, One-Way ANOVA with Dunnett’s post test) increased in both types of cell lines treated with MC-LR for 24 h. Expression of pro-apoptotic Bax and tumor suppressor p53 genes increased significantly (p50.001, One-Way ANOVA with Dunnett’s post test) in a dose dependant manner in both cell lines. Overall findings of the present study demonstrate that the MC-LR exerts cytotoxic and apoptotic effects in–vitro on kidney cells confirming that MC-LR is a nephrotoxin.

P298. SYSTEMS BIOLOGICAL ANALYSIS OF CARDIOTOXICITY ASSOCIATED WITH KINASE INHIBITORS Masashi Honma, Yoshiaki Kariya, and Hiroshi Suzuki Department of Pharmacy, The University of Tokyo Hospital, Tokyo, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

179

Molecularly targeted drugs, including targeted kinase inhibitors, are an important class of therapeutics widely used to treat different types of cancer, and a large number of new compounds are under clinical development. However, their performance with respect to improving clinical outcomes are often hampered by their limited tolerability. Particularly, kinase inhibitors can affect cardiac tissues and cardio vascular disease is one of the leading causes of death in patients who survived with anticancer therapies. This type of effect is not always expressed as immediate physiological responses, making it difficult to identify during the drug development process. Systems biology and systems pharmacology are useful in analyzing the mechanisms for the development of human diseases and the molecular mechanisms of drug actions. Especially where the polypharmacological property of a compound obscures the mechanisms of physiological drug actions, systems-based approaches to integrate information at the molecular level provide a more complete understanding. Thus, we aimed to achieve systems-based understandings of cardiomyocyte toxicity associated with kinase inhibitors. As an initial step, we have constructed a map covering signaling pathways at the downstream of receptor tyrosine kinases and apoptosis-related pathways, taking account of expression profiles of related molecules in cardiomyocytes. The map was constructed using CellDesigner, which is one of the most popular modeling software, and includes 922 entities with 632 reactions. Next, we performed a static analysis for the network structure of the constructed map, and determined the importance of each molecule in the network. The analysis was performed using Cytoscape, a major tool for analyzing biological pathways, using ‘‘betweenness centrality’’ score as the indicator for the importance of each node. As a result, 55 kinases, including JNK1, PAK2, Akt, and p38, are considered to have impacts on the behavior of the network. Finally, we estimated the influence of kinase inhibitors on cardiomyocyte cell death as the total value of the inhibition rate for each kinase corrected by the betweenness centrality score of each kinase. Results showed that the frequency of cardiac disorders associated with clinical use of each kinase inhibitor is reasonably correlated with its influence score (r = 0.85). These approaches will be widely applicable for systems-based approaches.

P299. ABSTRACT WITHDRAWN

P300. RISK ASSESSMENT OF ACYL GLUCURONIDES IN EARLY DRUG DISCOVERY SETTING Mithat Gunduz, Amanda Cirello, and Amin Kamel Department of Analytical Sciences & Imaging, Metabolite Identification Group, Novartis Institutes for Biomedical Research Inc., Cambridge, MA, USA Acyl glucuronide related reactivity/toxicity has been widely reported in the literature. This work was aimed to understand if the risk assessment of acyl glucuronides could be done without a radiolabeled material and/or an authentic standard. The acyl glucuronide standards were purchased from American Custom Chemicals Corporation (San Diego, CA). The aglycones used for the biosynthesis of the acyl glucuronides were obtained from Novartis Institutes for Biomedical Research sample bank. Human liver microsomes were purchased from CellzDirect (Durham, NC). In drug discovery setting, often time, the authentic standards of metabolites and/or radiolabeled drugs are not available. Our aim in this assay was to investigate if biosynthesized acyl glucuronides would have similar reactivity/stability as the authentic standards in potassium phosphate (KH2PO4 0.1 M; pH: 7.4) buffer, hence that would eliminate the time consuming chemical synthesis. In our assay, we investigated the acyl glucuronide authentic standards reactivity/stability in phosphate buffer. Then, the acyl glucuronides were biosynthesized in human liver microsomes and the stability/reactivity of the biosynthesized acyl glucuronides were determined in phosphate buffer. Results were compared to determine if the biosynthesized acyl glucuronides would have the same reactivity/stability as the authentic standards in phosphate buffer. The reactivity/stability assay for the acyl glucuronide standards were carried out by incubating each standard (20 uM) with phosphate buffer in shaking water bath at 37  C. Time points (0, 1, 2 and 4 hour) were obtained directly from the incubate and quenched with equal amount of acetonitrile containing 1% formic acid. The biosynthesis of acyl glucuronides were carried out by the incubations of the aglycones (20 uM), in human liver microsomes (1 mg/mL) fortified with KH2PO4 (0.1 M; pH: 7.4) buffer, MgCl2 (5 mM), alamethicin (50 ug/mL), saccharolactone (1 mg/mL) and UDPGA (3.2 mg/mL). Time points were taken directly from the incubation at 0, 1 and 2 hours and quenched with equal amount of acetonitrile containing 1% formic acid to determine the % of acyl glucuronide formed. The remaining samples were quenched with equal amount of acetonitrile containing 1% formic acid and centrifuged. The supernatants were removed and dried completely under gentle stream of nitrogen gas. The pellets were then dissolved in KH2PO4 (0.1 M; pH: 7.4) buffer and time points (0, 1, 2 and 4 hour) were obtained directly from the incubate and quenched with equal amount of acetonitrile containing 1% formic acid. All analyses were carried out on a Thermo-Finnigan LTQ-OrbitrapÕ with accurate mass measurement and MSn capabilities. All samples were analyzed to determine the % remaining for Acyl Glucuronides and aglycones. Our data highly suggested that the stability of biosynthesized acyl glucuronides were mostly in agreement with the stability of the of the acyl glucuronide authentic standards. The risk assessment of the carboxylic acid containing drugs were carried out by comparing the stability of the acyl glucuronide form of the drug candidate to those well studied and characterized drugs. In the risk assessment experiment the acyl glucuronides were chosen based on their categories as ‘‘safe’’, ‘‘warning’’ and ‘‘withdrawn’’.

180

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P301. URINARY EXCRETIONS OF NEWLY DEVELOPED BIOMARKERS FOR RENAL TUBULAR DAMAGE DURING A REPEATED HIGH-SALT INTAKE IN SPONTANEOUSLY HYPERTENSIVE RATS Keiko Hosohata, Daisuke Yoshioka, and Akio Fujimura Division of Clinical Pharmacology, Department of Pharmacology, School of Medicine, Jichi Medical University, Shimotsuke, Japan

Downloaded by [112.134.238.102] at 07:30 10 February 2016

A high-salt intake exacerbates hypertension and develops renal damage in hypertensive patients. However, the characteristics of renal tubular changes are not fully evaluated. Objective of this study was to clarify the time course of renal tubular alterations, and urinary excretions of conventional and new biomarkers for renal damage during a repeated high-salt intake in spontaneously hypertensive rats (SHR) and Wistar Kyoto rats (WKY). Male SHR and WKY received a regular (0.8% NaCl) or high (8% NaCl) salt diet from 9 to 17 weeks of age. At 10 weeks of age, a high-salt intake histologically caused renal tubular damages severely in SHR accompanied with increased blood pressure (BP), and mildly in WKY without high BP. Although albuminuria increased at 14 weeks of age only in salt-loaded SHR, a high-salt diet significantly elevated urinary excretions of vanin-1 and neutrophil gelatinase-associated lipocalin (NGAL) both in SHR and WKY from 10 weeks of age. Conversely, urinary kidney injury molecule-1 showed a significant elevation at 15 weeks of age only in salt-loaded SHR. The mRNA expression of endocytic receptor, by which NGAL is reabsorbed in renal tubules, was significantly reduced in salt-loaded SHR and WKY, indicating that the increase in urinary NGAL depended on the impaired renal reabsorption. These findings suggested that a high-salt intake caused renal tubular damage, accompanied with elevated urinary vanin-1 and NGAL before albuminuria. Thus, urinary vanin-1 and NGAL might be useful for early detection of renal injury caused by a high-salt intake.

P302. IN VITRO EVALUATION ON RENAL TOXICITY OF COPPER OXIDE NANOPARTICLES ¨ zhan1 Mahmoud Abudayyak1, Tuba Gurkaynak Altincekic2, and Gu¨l O 1 Department of Pharmaceutical Toxicology, Istanbul University, Istanbul, Turkey and 2Department of Chemical Engineering, Process and Reactor Design, Istanbul University, Faculty of Engineering, Istanbul, Turkey Copper oxide nanoparticles (CuO-NPs) are increasingly used in various industrial applications as catalyst, in gas sensor, heat transfer fluids, micro-electronics due to their special properties like high electrical and thermal conductivities. CuO-NPs have led to major concerns regarding the potential hazards for human health due to the high occupational and environmental exposure. Generally, nanotoxicity research focused on respiratory tract exposures for assessing the health effects of metal nanoparticles. However, nano-sized particles can cross membranes and further distribute into blood, brain, lung, heart, kidney, spleen, liver, intestine and stomach. According to some researchers, kidney is one of the targets for CuO-NPs causing renal tubules necrosis in the mice (Chen et al., 2006; Meng et al., 2007) There are few significant studies evaluating toxicity of CuO NPs in renal system. Therefore, we investigated the cyto- and genotoxic effects of CuO-NPs (TEM: 450 nm) in normal rat kidney epithelial cells (NRK-52 E) at 3.75–30 mg/ml concentrations. Alkaline comet assay was used for their genotoxic effects while lactate dehydrogenase (LDH), neutral red uptake (NRU) and mitochondrial succinate dehydrogenase (MTT) assays were used for their cytotoxic activities. Our results showed that CuO-NPs decreased highly the cell viability (IC50  19.6 mg/ml), and increased significantly DNA damage in dose-dependent manner. For the highest exposure concentration, tail intensity increased approximately 8.8-fold compared to the control group. In future, it will be better to evaluate the cellular morphological change and oxidative stress incidence. The present study highlights on the in vitro renal toxicity of CuO-NPs. References Chen Z, Meng H, Xing G, et al. (2006). Acute toxicological effects of copper nanoparticles in vivo. Toxicol Lett 163:109–120. Meng H, Chen Z, Xing G, et al. (2007). Ultrahigh reactivity provokes nanotoxicity: Explanation of oral toxicity of nano-copper particles. Toxicol Lett 175:102–110.

P303. INVOLVEMENT OF MONOCARBOXYLATE TRANSPORTER 4 IN STATIN-INDUCED CYTOTOXICITY Masaki Kobayashi, Yurika Kikutani, Shotaro Sasaki, and Ken Iseki Faculty of Pharmaceutical Sciences, Hokkaido University, Sapporo, Japan 3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors such as statins are the most widely used cholesterollowering agents for prevention of obstructive cardiovascular events. However, severe adverse events, i.e. skeletal muscle abnormalities, including myopathy and rhabdomyolysis, associated with lipophilic statins sometimes limit the lipid-lowering therapy with these agents. Lovastatin, one of statins increases exercise-induced skeletal muscle injury. On the other hand, exercise induces protein contents of monocarboxylate transporter 4 (MCT4), is expressed strongly in skeletal muscle and is thought to play a major role in the transport of metabolically important monocarboxylates such as L-lactic acid. Since statins also have monocarboxylate structures within the molecules, we hypothesized that statin-induced muscle injury is associated with MCT4 expression and function. In order to verify this hypothesis, we examined the effect of statins on L-lactic acid uptake using MCT4-expressed cells and to clarify the relationship between stain-induced cytotoxicity and MCT4 expression in RD cell line as a model of in vitro skeletal muscle. Lipophilic statins significantly inhibited [14C] L-lactic acid uptake in a concentration-

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

181

dependent manner in MCT4-expressed cells. Moreover, lipophilic statins reduced the number of viable cells and caused dramatic morphological changes and caspase-3/7 activation, and induced MCT4 expression levels in RD cells. Two siRNAs (10 nM) for MCT4 significantly decreased MCT4 expression at 72 h after transfected to RD cells. Lipophilic statin-induced RD cell injury was suppressed by MCT4 siRNAs transfected to RD cells. These results suggest that the mechanism of statin-induced muscle injury is associated with MCT4 expression.

P304. INCREASE OF RISK OF ACUTE RENAL FAILURE AND SERUM CREATININE LEVEL BY USE OF PIPERACILLIN/ TAZOBACTAM

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Masami Kawahara1, Sho Wada2, Hiroki Uda1, Satoko Hatano2, Takeo Nakanishi2, and Ikumi Tamai2 1 Department of Pharmacy, Kanazawa Municipal Hospital, Kanazawa, Japan and 2Department of Membrane Transport and Biopharmaceutics, Faculty of Pharmaceutical Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan A combination product of antibacterial piperacillin (PIPC) and beta-lactamase inhibiting drug, tazobactam (TAZ), is effective for various infectious diseases and its renal toxicity is reported to be rare (0.4% incidence). However, we experienced the higher incidence of the acute renal toxicity in the patients not with PIPC alone but with PIPC/TAZ therapy. Among them, a few patients exhibited an increase of serum creatinine (Scr) without increase of blood urea nitrogen (BUN). The observation might be explained by a specific change in creatinine clearance without renal failure. We hypothesized that the interaction of PIPC/TAZ with secretory transporters for creatinine may cause some effect on the phenomenon. The purpose of this study is to elucidate the cause of renal toxicity by PIPC/TAZ. Based on medical records of the patients (April 2010 to March 2014), the frequencies of adverse renal effects were surveyed retrospectively. The plasma concentrations of PIPC and TAZ in seven patients were measured. The effect of PIPC and TAZ on renal creatinine transporters was examined using cells transfected with transporter genes. PIPC and TAZ were measured by LC-MS/MS. Retrospective study resulted in 3.6% incidence of acute renal failure (true kidney failure) in 798 PIPC/TAZ-administered patients, which was significantly higher than the results reported in the product’s Interview Form in 2014. In two of seven patients whose plasma concentrations were measured, PIPC and TAZ levels were two-times higher than others with an increase in Scr, while no elevation of BUN was observed. OCT2- and MATE1-mediated transport of creatinine was inhibited by PIPC but not by TAZ. In addition, TAZ inhibited MRP4, which transports PIPC as substrate, suggesting that PIPC accumulates in renal tubular epithelial cells when TAZ is coadministered, by decreasing MRP4mediated renal secretion of PIPC. Hence, it is possible that Scr could be increased by specific interaction of renal secretory transporters for creatinine with PIPC, of which concentration could be increased by coadministered TAZ, even when there is no renal failure (false kidney failure). In conclusion, PIPC/TAZ may clinically have higher possibility of acute renal failure than considered generally. However, some patients may show increase in Scr by specific interaction with PIPC on creatinine transporters in kidney without renal failure. Accordingly, there may be higher possibility to observe true and false kidney failure by PIPC/TAZ through an interaction of PIPC with TAZ on transporters, than PIPC alone therapy.

P305. EVALUATION OF REACTIVE METABOLITES INDUCED MITOCHONDRIAL TOXICITY AS A CAUSATIVE RISK FACTOR IN DRUG INDUCED LIVER INJURY Shuichi Sekine1, Maho Shirakawa1, Ayaka Tanaka1, Toshiharu Horie2, and Kousei Ito1 1 The Laboratory of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan and 2 Faculty of Pharmaceutical Sciences, Teikyo Heisei University, Tokyo, Japan Purpose: The risk of idiosyncratic drug toxicity (IDT) is of great concern to the pharmaceutical industry. IDT is believed to be orchestrated by multiple events, such as reactive metabolites (RMs) formations, oxidative stress and mitochondria toxicity. Many of IDT drugs (e.g. Troglitazone, Dicrofenac, Isoniazid) are known to induce liver mitochondrial injury. However, only little information is available for (reactive) metabolite-dependent DILI risks associated with mitochondrial dysfunction, because of difficulties in purification of instable RMs. In an effort to address this question, we constructed one pot in vitro assay system which enables us to assess the possibility of (reactive) metabolite-dependent mitochondrial toxicity without any purification process. In this study, (1) we evaluated the validity of this assay method by using benzbromarone (BBR) which is known to cause mitochondrial toxicity and to be metabolized in the liver, and 2) we applied the assay system for 48 test drugs (2) Withdrawn, 13 Black box warning, 11 Warning, 17 hepatotoxic drug, 5 Safe Drug) to evaluate if it is useful to detect the potential risk of (reactive) metabolite-dependent mitochondrial toxicity. Methods: (1) Evaluation of RM-dependent mitochondrial toxicity by using BBR; Liver microsome was isolated from male Sprague-Dawley rat (6–7weeks) pretreated with dexamethasone (Dex; 100 mg/kg/day) for 3 days. BBR (54 mM) was incubated with the isolated Dex treated rat liver microsome (DRLM) for 30 min in the absence or presence of NADPH and/or cytochrome P450 (CYP) inhibitors. Then, mitochondria isolated from normal rat liver were added and mitochondrial toxicity was assessed by monitoring the decrease of absorbance at 540 nm (mitochondrial membrane permeability transition). The formation of known metabolites, including 10 -OH BBR, 6-OH BBR and 10 ,6-OH BBR were measured by LC-MS/MS. (2) Application of the assay system for other drugs; 48 test drugs were evaluated at concentration range of 1–100 times Cmax. Results and discussion: (1) Mitochondrial toxicity was enhanced by BBR pre-incubated in DRLM with NADPH, but was not by that without NADPH. Time-dependent formation of the metabolites (10 -OH BBR, 6-OH BBR,

182

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

10 ,6-OH BBR) was observed during the pre-incubation, and non-specific CYP inhibitors including clotrimazole, SKF-525A, and ketoconazole all inhibited formation of these 3 metabolites. On the other hand, enhancement of the mitochondrial toxicity by BBR pre-incubated in DRLM with NADPH was inhibited by clotrimazole and SKF-525A but not by ketoconazole. These results suggested that putative metabolite(s) other than these 3 metabolites might be produced in a NADPH-dependent manner to enhance mitochondrial toxicity, and that its formation was inhibited by clotrimazole and SKF-525A but not by ketoconazole. (2) Among 48 test drugs, 14 test drugs showed NADPH-dependent enhancement of mitochondrial toxicity. Among 14 drugs, productions of reactive/toxic metabolites have been reported in 10 drugs (Amodiaquine, BBR, Chlorpromazine, Flutamide, Indomethacin, Nortriptyline, Quinidine, Tamoxifen, Ticlopidine, Warfarin). Conclusion: The assay system can assess the putative (reactive) metabolite-dependent mitochondrial toxicity without any purification process. This will help to improve the screening efficiency of DILI risk compound particularly at early drug development stage.

P306. PLATINUM-DNA DAMAGE-INDUCED TRANSCRIPTIONAL INHIBITION IN CULTURED RAT DRG CELLS CONTRIBUTES TO NEURONAL CELL BODY ATROPHY AND NEUROTOXICITY INDUCED BY OXALIPLATIN

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Fang Yan, Virginia Ip, Stephen M. F. Jamieson, and Mark J. McKeage Department of Pharmacology and Clinical Pharmacology and Auckland Cancer Society Research Centre, The University of Auckland, Auckland, New Zealand Background: Platinum-based anticancer drugs, including oxaliplatin, cisplatin and carboplatin, have long been used to combat a wide range of cancers, by forming platinum-DNA adducts that inhibit DNA replication and transcription of cancer cells. However, peripheral neurotoxicity is a dose-limiting side effect of oxaliplatin and cisplatin, which has been associated with the cell body atrophy of sensory neurons in dorsal root ganglia (DRG) through a not yet fully understood mechanism. In this study, cultured rat DRG cells and adult Wistar rats were used to determine the role of platinum-DNA damage-induced transcriptional inhibition in DRG neuronal cell body atrophy induced by oxaliplatin and other platinum drugs. Methods: Platinum binding to DNA was measured by inductively coupled plasma mass spectrometry. Level of RNA synthesis was fluorescently visualised by RNA-incorporation of 5-ethynyl uridine (EU), and was measured by either fluorescence intensity quantification of RNA-incorporated EU or scintillation counting of RNA-incorporated [3H]uridine. Total RNA content was quantified by spectrophotometry. Neuronal cell body size was determined by measuring planimetric area of each neuron in phase-contrast or haematoxylin and eosin staining images. Results: The level of RNA-incorporation of EU in cultured rat DRG neurons correlated with their cell body size, but both were reduced by oxaliplatin treatment in a time- and concentration-dependent manner. Platinum binding to the DNA of cultured rat DRG cells occurred before the oxaliplatin-induced inhibition of RNA synthesis and depletion of total RNA content, which were all followed by the development of neuronal cell body shrinkage. Both sodium thiosulfate and cimetidine reduced platinum binding to DNA, and protected cultured rat DRG cells from both inhibition of RNA synthesis and neuronal cell body shrinkage induced by oxaliplatin. Four different platinum-based anticancer drugs, oxaliplatin, cisplatin, carboplatin and ormaplatin, showed different amounts of platinum binding to DNA, which corresponded to their effects on RNA synthesis and neuronal cell body size of cultured rat DRG cells, and to their clinical neurotoxicity. A model transcriptional inhibitor, actinomycin D, also caused neuronal cell body atrophy following its inhibition of RNA synthesis of cultured rat DRG cells. In addition, in vivo administration of oxaliplatin reduced the total RNA content, concurrently with its induction of neuronal cell body atrophy, in rat DRG tissue. Conclusions: Platinum-DNA damage-induced transcriptional inhibition contributes to DRG neuronal cell body atrophy and neurotoxicity induced by oxaliplatin and other platinum-based anticancer drugs.

P307. METABOLIC PROFILING OF GEISSOSCHIZINE METHYL ETHER, ALKALOID OF UNCARIA HOOK, IN HUMAN LIVER MICROSOMES USING LC/MS/MS Hirotaka Kushida1, Takashi Matsumoto1, Junko Watanabe1, Yasushi Igarashi2, Hiroaki Nishimura2, Kazuya Maemura1, and Yoshio Kase1 1 Tsumura Research Laboratories, Tsumura & Co., Inashiki-gun, Japan and 2Kampo Formulation Development Center, Tsumura & Co., Inashiki-gun, Japan Purpose: Geissoschizine methyl ether (GM) is an indole alkaloid found in Uncaria hook, a crude drug constituting the Kampo formulation Yokukansan. GM is thought to be one of the active components in Yokukansan with pharmacological effects. We previously reported that GM decreases aggressiveness and improves social behavior in isolation-induced mice through a partial agonist action on the 5-HT1A receptor. On the other hand, studies in which the pharmacokinetics of Uncaria hook alkaloids were investigated in rats following oral administration of Yokukansan detected GM in the plasma and brain (Kushida et al., 2013). Nevertheless, the metabolic profile of GM has not been sufficiently investigated. GM is structurally similar to the pharmacologically active yohimbine and geisoschizine (de-methylated GM). Considering these compounds may be produced as GM metabolites, it is important to examine the metabolic profile of GM. Therefore, we examined the metabolic profile of GM in human liver microsomes using LC/MS/MS. Methods: GM (final volume 20 mM) and human liver microsomes (final volume 5 mg/mL) were allowed to react at 37  C for 60 min. The reaction mixture was extracted with ethyl acetate. IDA measurements of the extracted samples were performed with the API4000 LC/MS/MS system. GM metabolites were analyzed with Light Sight software. Results and discussion: Peaks with m/z values of 383.2, 353.2, 365.2, 381.2, and 385.2 were detected in the extracted ion

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

183

chromatogram. These peaks were attributed to hydroxylated, de-methylated, dehydrogenated, methylated, and water-adduct forms, respectively, based on their fragmentation patterns in MS2 spectra. The indole skeleton was speculated to be hydroxylated in the hydroxylated metabolite. Modifications in other metabolites were presumed to occur in the remaining part of the structure. As for the de-methylated metabolites, 2 types were presumably generated, based on the structure of GM, which included geisoschizine. Currently, we are attempting to synthesize the two de-methylated forms for definitive identification. Conclusions: We identified GM metabolites in human liver microsomes using LC/MS/MS and revealed the metabolic profile of GM in humans.

Reference Kushida H, Fukutake M, Tabuchi M, et al. (2013). Simultaneous quantitative analyses of indole and oxindole alkaloids of Uncaria Hook in rat plasma and brain after oral administration of the traditional Japanese medicine Yokukansan using highperformance liquid chromatography with tandem mass spectrometry. Biomed Chromatogr 27:1647–1656.

P308. IMPACT OF GENDER, AGE AND FED/FASTED STATE OF RATS ON THEIR SERUM HYDROPHILIC METABOLITES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Keiko Maekawa1, Kosuke Saito1, Kirk Pappan2, Masaki Ishikawa1, Masayo Urata1, Yoko Tajima1, Mayumi Murayama1, Yuji Kumagai3, and Yoshiro Saito1 1 National Institute of Health Sciences, Tokyo, Japan, 2Metabolon, Inc., Durham, NC, USA, and 3Clinical Trial Center, Kitasato University East Hospital, Kanagawa, Japan Background and purpose: Metabolomics is a powerful tool for exploring biomarkers to evaluate disease states, therapeutic efficacy, and toxicity of drugs because it provides a comprehensive view of biological processes in the body. However, metabolomic profiles are known to be influenced by study design factors such as sample population and sample collection conditions – which may lead to confounding results. This study aimed to elucidate how the differences in age, gender, and fed/ fasted state of rats affect hydrophilic metabolite levels in serum. Methods: Sera were collected from Sprague-Dawley rats divided into seven groups (n = 10 in each group), depending on their gender (male vs. female), age (10 vs. 30 weeks old), fed/fasted state (16 hr-, 22 hr- fasted, or chow-fed at 10 weeks old) and blood collection time (morning vs. afternoon). Hydrophilic compounds with low molecular weight were extracted with methanol and measured using both liquid chromatography-mass spectrometry (LC/MS) and gas chromatography/MS. Relative levels of the detected metabolites were statistically compared among groups, and significant levels were set at p50.05. Results and discussion: A total of 314 metabolites, such as amino acids, carbohydrates and nucleotides, were identified. Of these, 38% and 49% of metabolites showed gender-related differences in their levels in young (10 weeks old) and old (30 weeks old) animals, respectively, while 35% and 30% showed age-related differences in male and female rats, respectively. Thus, both age and gender contributed to the differences of a large number of metabolites in rat serum. For example, markers of oxidative stress and redox defense such as methionine, 2-aminobutyrate, oxidized glutathione, and ophthalmate were varied by both gender and age. Furthermore, differences in fed/fasted state strongly impacted the levels of serum metabolites, resulting in statistically significant differences in 50% of metabolites between the 16 hour-fasting and nonfasting conditions. For example, levels of numerous markers of protein hydrolysis such as C-glycosyltryptophan, 3methylhistidinewere elevated in fasting conditions, suggesting that fasting invoked muscle protein hydrolysis. In addition, increased levels of the stress hormone corticosterone and its precursor 11-dehydrocorticosterone were observed in fasting conditions, suggesting that fasting induces stress responses. On the other hand, levels of a smaller number of metabolites were influenced by the duration of fasting (16 h vs. 22 h) and the sample collection time (morning vs. afternoon). Conclusion: Metabolites with different levels among gender or age-biased samples should be paid attention in biomarker exploration and diagnosis using them. Our fundamental information on sample collection for measuring blood hydrophilic metabolites is useful for ensuring the quality of biomarker discovery and qualification process at preclinical phase.

P309. PRECLINICAL METABOLISM OF DACLATASVIR, AN INHIBITOR OF THE HEPATITIS C VIRUS NS5A REPLICATION COMPLEX Benjamin M. Johnson1, Omar D. Lopez2, Xiaohong Liu1, Xiaohua Huang1, John E. Leet3, Weiping Zhao4, Anjaneya Chimalakonda1, Aberra Fura1, Jay O. Knipe1, Kimberley J. Snow1, R. Marcus Fancher4, Van Nguyen2, Jason Goodrich2, Makonen Belema2, Lawrence G. Hamann2, Min Gao5, Yue-Zhong Shu4, and W. Griffith Humphreys4 1 Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Company, Wallingford, CT, USA, 2Department of Discovery Chemistry, Bristol-Myers Squibb Company, Wallingford, CT, USA, 3Synthesis and Analysis Technology Team, Bristol-Myers Squibb Company, Wallingford, CT, USA, 4Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Company, Lawrenceville, NJ, USA, and 5 Department of Discovery Infectious Diseases, Bristol-Myers Squibb Company, Wallingford, CT, USA Daclatasvir (BMS-790052; dimethyl (2S,20 S)-1,10 -((2S,20 S)-2,20 -(5,50 -(biphenyl-4,40 -diyl)bis(1H-imidazole-5,2-diyl))bis(pyrrolidine-2,1-diyl))bis(3-methyl-1-oxobutane-2,1-diyl)dicarbamate) is a first-in-class, potent and selective inhibitor of the hepatitis C virus nonstructural protein 5A replication complex. In support of preclinical studies during discovery, liquid chromatography– tandem mass spectrometry and NMR were used in connection with synthetic approaches to investigate the biotransformation of

Downloaded by [112.134.238.102] at 07:30 10 February 2016

184

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

daclatasvir in vitro and in the cynomolgus monkey, dog, and rat. The results of these studies indicated that cytochromes P450 were the main enzymes involved in the metabolism of daclatasvir. Oxidative pathways that were active against daclatasvir included regioselective hydroxylation of the valine methyl group; decarbamoylation of the valine amino moiety; b-, g-, and d-oxidations of the pyrrolidine group; and associated multi-step pathways. Of particular interest, d-oxidation of the pyrrolidine moiety resulted in ring opening to a putative aminoaldehyde intermediate followed by an intramolecular reaction between the aldehyde and the proximal imidazole nitrogen atom. This reaction yielded a stable metabolite that contained a new 5,6,7,8tetrahydroimidazo[1,2-a]pyridine ring system and was inactive against NS5A. Although it eluted as a single analyte during HPLC, multiple NMR experiments (1H, ROESY, gHMBC) were used to show that this metabolite was comprised of two equally abundant but non-interchangeable diastereomers with the hydroxyl group oriented differently between them. Despite robust formation of this metabolite in multiple systems including glutathione-supplemented liver microsomes, glutathione conjugates were not observed, suggesting that the intramolecular rearrangement described above was favored over intermolecular binding as a means of quenching the aldehyde intermediate. Furthermore, overall rates of covalent binding to protein associated with metabolism of daclatasvir were low (50–70 pmol/mg/h) in NADPH-supplemented liver microsomes from human, monkey, dog, and rat. The metabolite profile in plasma of preclinical species demonstrated that daclatasvir was the main drug-related compound in circulation and was accompanied by relatively modest levels of metabolites. Disposition in preclinical species occurred mainly via direct biliary excretion of daclatasvir, direct intestinal excretion of daclatasvir, and biliary excretion of metabolites – predominantly those formed by pyrrolidine d-oxidation (rat and monkey) and decarbamoylation (dog). This biotransformation profile supported the continued development of daclatasvir, which has since advanced through clinical trials and is presently under regulatory review in multiple countries for the treatment of chronic HCV infection in combination with other direct-acting antivirals.

P310. COMPARISON OF PLASMA METABOLITE PROFILES DETERMINED BY CONVENTIONAL LIQUID SCINTILLATION COUNTING AND ACCELERATOR MASS SPECTROMETRY Mark R. Kagan1 and Stephen Dueker2 1 Department of Drug Metabolism and Pharmacokinetics Novartis Pharmaceuticals, East Hanover, NJ, USA and 2Eckert & Ziegler Vitalea Science, Davis, CA, USA The metabolic profile of a drug candidate in human plasma is almost always needed to evaluate the drug candidate’s safety and is sometimes needed to develop rigorous PK/PD relationships. Traditionally, these plasma profiles have been determined using [14C]-labeled drug and liquid scintillation counting (LSC) for radioactivity measurements. The sensitivity of this analytical method typically requires doses containing 50–100 mCi of radioactivity. Due to the ethical considerations associated with radioactivity doses of this magnitude, these studies are typical conducted in a minimum number of subjects (4–6) with each subject receiving a single dose of the drug. Alternatively, if the much more sensitive technique of Accelerator Mass Spectrometry (AMS) is used, a 1000-fold reduction in the radioactivity dose could be achieved. Such a reduction would alleviate the ethical concerns associated with multiple dosing of larger groups of subjects. Additionally, such an enhancement in sensitivity might represent the only practical means for obtaining plasma profiles for drugs administered at very low doses such as those used in ocular, dermal, and some inhalation therapies. Whereas the determination of plasma profiles using LSC methods are readily performed in-house by the Drug Metabolism group at Novartis, determinations involving AMS, due to the cost and complexity of the instrumentation involved, require the analyses to be outsourced to an outside provider. In order to gain confidence in the ability of AMS to accurately reproduce the results of a conventional metabolite profiling study, a collaborative evaluation was performed with Vitalea Science, a leading commercial provider of AMS services. In this evaluation, control human plasma was simultaneously spiked with five [14C]-radiolabeled Novartis compounds at radioactivity levels suitable for LSC (850 dpm/mL/ compound). The resulting metabolite profile was determined by Novartis using traditional LSC methods. The sample was then diluted 1000-fold with additional control human plasma and sent to Vitalea Science for AMS analysis. Prior to sending the sample to Vitalea Science, it was fortified with sufficient amounts of the unlabeled compounds to insure that the overall amounts were similar in the diluted and undiluted samples. For all five of the compounds analyzed, the retention times measured by the two laboratories agreed within 5%. For four of the five compounds analyzed, the relative amounts determined by the two laboratories agreed within 25%. The larger differences in relative amounts obtained for the fifth compound was believed to be an artifact that would not have occurred if the samples had been sent directly to Vitalea Sciences from the clinical site as would have been done in a real study. The overall conclusion of the evaluation was favorable with respect to AMS being able to generate plasma metabolite profiles that closely agree with those determined by LSC.

P311. METABOLITE EXPOSURE ASSESSMENT STRATEGY IN PRECLINICAL SPECIES IN ABSENCE OF CHEMICALLY SYNTHESIZED STANDARD USING LIQUID CHROMATOGRAPHY AND MASS SPECTROMETRY Ranjith Kumar Ponnamaneni1, Arun kumar Manoharan1, Nageswara Rao Muddana1, Gopinadh Bhyrapuneni1, and Ramakrishna Nirogi2 1 Department of Drug Metabolism and Pharmacokinetics, Suven Life Sciences Ltd, Hyderabad, India and 2Department of Discovery Research, Suven Life Sciences Ltd, Hyderabad, India

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

185

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Identification, characterization and obtaining quantitative information of drug metabolites in preclinical species is becoming prerequisite in the early identification of active metabolites and also in the selection of most appropriate toxicological species for the safety demonstration of major circulatory metabolites in human plasma. Regulatory agencies (FDA in 2008, ICH M3(R2) in 2009 and 2012) issued guidelines on metabolites in safety testing (MIST) by setting threshold for circulatory human metabolite. Overall the regulatory guidelines indicate the safety coverage of major circulatory human metabolites in at least one of the preclinical toxicology species. As the synthesis of all metabolite standards may not be possible, thus imposing analytical challenges in drug development for the quantitative comparison of metabolite exposure levels between human and animals. Right from the in-vitro studies until the safety coverage in in-vivo preclinical plasma samples alternate bioanalytical strategy is required to provide fit-for-purpose quantitative comparison of AUC levels that can permit critical decisions related to metabolite safety. The objective here is to demonstrate bioanalysis strategy to gain quantitative information in terms metabolite exposures in preclinical subject using midazolam and its two metabolites. With time averaged AUC pooling of plasma samples parent and metabolites AUCs are calculated from the full scan chromatographic peak areas and critical factors that would alter mass spectrometer (MS) ionization and in turn effects on instrument response are discussed. Analytical factors to be considered to assess the data with reliability are addressed. The experimental strategy is demonstrated by assessing the exposures of midazolam and its two metabolites 1-OH-midazolam and 4-OH-midazolam in plasma samples after oral dosing of midazolam to Wistar rat. The exposure levels obtained by this plasma pooling method with chromatographic peak area ratio are further evaluated by determining AUCs from the pharmacokinetic samples analysis at each time point using standard bioanalytical method employing linearity curves of metabolites standards.

P312. FORMATION OF THE GLUTATHIONE CONJUGATE METABOLITES OF DICLOFENAC IN CHIMERIC TK-NOG MOUSE WITH HUMANIZED LIVER Kohei Nozawa1, Satoshi Ito1, Hiroyuki Chijiwa1, Shota Nakamura1, Hiroshi Suemizu2, Yasuyuki Ohnishi2, Hidetaka Kamimura1, and Shin-ichi Ninomiya1 1 Sekisui Medical Co., Ltd., Tokai, Ibaraki, Japan and 2Department of Biomedical Research, Central Institute for Experimental Animals, Kawasaki, Japan The exposure to drug metabolites in humans and preclinical species during drug development needs to be determined to ensure that the safety of drug-related components in humans is adequately assessed in the standard toxicology studies. An early understanding of key metabolites of drugs offers time and resource sparing advantages to ascertain those metabolite exposures. Although recent analytical technique provides effective strategy to profile drug metabolites, more respective humanized-animal model is expected to put forward such the human metabolism prediction. Here, we investigate the drug metabolism after administration of diclofenac in chimeric mice (TK-NOG mouse; highly immunodeficient mouse strain (NOG) expressing a thymidine kinase transgene) with humanized liver. Glutathione conjugates of diclofenac are recently receiving attention with respect to relations with its hepatic toxicity, and some of them are human specific. In two different literatures, incubation of diclofenac with human liver microsomes generated different glutathione-conjugates, 20 -hydroxy-30 -(glutathione-S-yl)-monoclofenac (M4) and 40 -hydroxy-20 -(glutathioneS-yl)-monoclofenac (M5) with almost the same mass spectra, probably generated via epoxide and quinone-imine type intermediates, respectively. The aims of our study were to investigate whether these metabolites were generated in humanized chimeric mice produced from TK-NOG mice. After singe oral administration of diclofenac to the mouse model, we found a metabolite with a single HPLC peak giving the same mass spectrum in the bile of humanized chimeric mice together with some other glutathione conjugates. This in vivo monoclorinated metabolite was also found after oral administration of 40 -hydroxy diclofenac. Although it is not clear whether this conjugate corresponds to M4 or M5, chimeric mice with humanized liver were considered to be a functional tool for the study of drug metabolism in humans. We are planning further strucure elucidation of these glutathionyl hydroxy monoclofenac and the other glutahione conjugates of diclofenac.

P313. IN VITRO METABOLISM OF THE MAYTANSINOID PAYLOADS USED FOR ANTIBODY DRUG CONJUGATES John A. Davis, Dan A. Rock, and Brooke M. Rock Department of Pharmacokinetics and Drug Metabolism, Amgen Inc., Seattle, WA, USA 0

0

Antibody drug conjugates (ADCs) containing the cytotoxic tubulin-targeting maytansinoid derivatives, N2 -deacetyl-N2 -(30 0 mercapto-1-oxopropyl)-maytansine (DM1) and N2 -deacetyl-N2 -(4-mercapto-4-methyl-1-oxopentyl)-maytansine (DM4) are referred to as antibody maytansinoid conjugates (AMCs). Several AMCs are currently being evaluated in clinical trials as targeted chemotherapeutic agents. These agents are routinely conjugated either through ‘‘cleavable’’ di-sulfide or ‘‘noncleavable’’ thioether linkages whereupon tumor specific binding leads to intracellular uptake and degradation liberating either free maytansinoid or lysine-linker-maytansinoid conjugates, respectively. Free maytansinoids have been reported in circulation in clinical studies where AMCs were dosed regardless of the linker used (Lapusan et al., 2012; Shen et al., 2012). Despite the low levels detected in these studies, the observation of hepatotoxicity observed from clinical studies performed with AMCs warrants investigation into the metabolism of the maytansinoids. This may help in the elucidation of the mechanism(s) contributing to the observed toxicity. Previous reports show free maytansinoids undergo oxidative metabolism in the liver and are reversible and

186

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

time-dependent inhibitors of cytochrome P450 (P450) enzymes 2D6 and 3A, respectively (John et al., 2012; Sun et al., 2011). The current work aims to better define and understand the metabolism and clearance of these highly potent moieties. To this end, in vitro metabolites of DM1 and DM4 formed in microsomes and hepatocytes from human, non-human primate, and rat were determined. In addition, the human hepatic enzymes responsible for metabolite formation were determined. Results indicated the maytansinoids are cleared by a variety of metabolic pathways. The respective sulfinic and sulfonic acids were the major oxidative metabolites formed from P450s. Experiments performed with microsomal, S-9, and cytosolic liver fractions demonstrated that S-methylation was also an important metabolic pathway. Subsequent metabolism of the S-methyl metabolites of DM1 and DM4 to their respective sulfoxides and sulfones were shown to be mediated by P450s and to a lesser extent FMO1. In addition, the spontaneous formation of mixed disulfide ‘‘metabolites’’ including glutathione, cysteinylglycine, and cysteine adducts along with homo dimer formation (DM1 only) was observed in all matrices including plasma. In order to obtain adequate selectivity to identify ‘‘drug related’’ material, mass defect filtered (MDF) data were generated with high resolution mass spectrometry. This information was then leveraged to develop highly sensitive triple quadrupole methods. Combined, these two techniques allowed the characterization of both DM1 and DM4 metabolites from multiple experiments. Understanding the metabolism of these maytansiods will aid in our overall understanding of the disposition of AMCs.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

References Davis JA, Rock DA, Wienkers LC, Pearson JT. (2012). In vitro characterization of the drug–drug interaction potential of catabolites of antibody-maytansinoid conjugates. Drug Metab Disposit 40:1927–1934. Lapusan S, Vidriales MB, Thomas X, et al. (2012). Phase I studies of AVE9633, an anti-CD33 antibody-maytansinoid conjugate, in adult patients with relapsed/refractory acute myeloid leukemia. Investigational New Drugs 30:1121–1131. Shen BQ, Bumbaca D, Saad O, et al. (2012). Catabolic fate and pharmacokinetic characterization of trastuzumab emtansine (T-DM1): an emphasis on preclinical and clinical catabolism. Curr Drug Metab 13:901–910. Sun X, Widdison W, Mayo M, et al. (2011). Design of antibody-maytansinoid conjugates allows for efficient detoxification via liver metabolism. Bioconjug Chem 22:728–735.

P314. CHARACTERIZATION OF HEPATIC LIPID PROFILES OF NONALCOHOLIC STEATOHEPATITIS (NASH) AND FOLLOWING FIBROSIS USING A MOUSE MODEL Masaki Ishikawa1, Kosuke Saito1, Takashi Uebansou2, Keiko Maekawa1, Yuya Senoo1, Mayumi Murayama1, Yoko Tajima1, Tomoko Nishimaki-Mogami3, Hiroki Nakanishi4, Kazutaka Ikeda5, Makoto Arita6, Ryo Taguchi7, Masato Fujii8, Yuichiro Shibazaki8, Hiroyuki Yoneyama8, Takao Nammo2, Kazuki Yasuda2, and Yoshiro Saito1 1 Division of Medicinal Safety Science, National Institute of Health Sciences, Tokyo, Japan, 2Department of Metabolic Disorder, National Center for Global Health and Medicine, Tokyo, Japan, 3Division of Novel Foods and Immunochemistry, National Institute of Health Sciences, Tokyo, Japan, 4Research Centre for BioSignal, Akita University, Akita, Japan, 5Institute for Advanced Biosciences, Keio University, Tsuruoka, Japan, 6Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan, 7College of Life and Health Sciences, Chubu University, Kasugai, Japan, and 8Stelic Insititute & Co, Tokyo, Japan Introduction: Nonalcoholic fatty liver disease (NAFLD), including simple steatosis, nonalcoholic steatohepatitis (NASH) and following fibrosis/cirrhosis, is now a major health problem in the world. Alteration of lipid metabolism has been well associated with the progression of NASH as well as steatosis. However, previous reports only focused on free fatty acids, and overall lipid profiles in each stage of NAFLD have not been characterized in detail yet. In this study, we aimed to make a comprehensive analysis on lipid profiles of steatosis, NASH and following fibrosis in a mouse model. Methods: We used ‘‘STAM mouse’’ (Fujii et al., 2013) as the NASH model, which was treated by streptozotocin in the neonatal period and fed a high fat diet (HFD) from 4 weeks old. This model is unique in that it demonstrates progression from steatosis to NASH followed by fibrosis and hepatocellular carcinoma, which is very similar to NASH in humans. We used HFD mice as controls. Control HFD mice (n = 5) remained steatosis up to 12 weeks old, while STAM mice (n = 5) showed NASH and fibrosis at 8 and 12 weeks old, respectively. We performed a lipidomic analysis using liquid chromatography-mass spectrometry (LC-MS) and LC-MS/MS for liver tissues of STAM and HFD mice. Results: We determined in total 257 lipid species, including 103 phospholipids, 17 sphingolipids, 112 neutral lipids, 12 free fatty acids, 8 acylcarnitines and 5 eicosanoids from liver samples of STAM and control HFD mice. The highlighted differences in lipid molecules between steatosis and NASH are as follows: (1) low percent composition of palmitate and high of stearate in phosphatidylcholines and phosphatidylethanolamines in NASH; (2) high levels of sphingolipids such as sphingomyelins, ceramides and hexysosyl-ceramides in NASH; (3) low levels of diacylglycerols (DGs) and triacylglycerols in NASH; (4) high levels of cyclo-oxygenase metabolites from arachidonate in NASH. On the other hand, the highlighted changes in lipid molecules from 8 to 12 weeks old are as follows: (1) an increase in acyl-carnitines and DGs in STAM mice; (2) an agedependent increase in phosphatidylinositols in HFD mice, but its absence in STAM mice. Conclusion: Our present study delineates the change of lipid profiles upon progression from steatosis to NASH as well as NASH to following fibrosis, and would help to clarify the underlying mechanisms of NASH progression. We are planning to analyze the plasma from these models to detect potential biomarkers which may reflect lipidomic changes in the NASH liver.

Reference Fujii M, Shibazaki Y, Wakamatsu K, et al. (2013). A murine model for non-alcoholic steatohepatitis showing evidence of association between diabetes and hepatocellular carcinoma. Med Mol Morphol 46:141–152.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

187

P315. METABOLISM OF EIGHT MODEL PHARMACEUTICAL COMPOUNDS IN RAT- AND HUMAN-HEPATOPACTM VERSUS LIVER MICROSOMES AND SUSPENSION HEPATOCYTE PLATFORMS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Julius O. Enoru1, William DeMaio1, Adiba Watanyar1, Kenneth Draper1, Amanda Moore2 and Okey Ukairo2 1 Drug Safety & Metabolism, Ricerca Biosciences LLC, Concord Twp, OH, USA and 2Hepregen Corporation, Medford, MA, USA Metabolism evaluations for compounds in early development are typically conducted in vitro using liver microsomes and suspension hepatocytes. While these systems have served the drug development community for many years, they have limitations and result in infrequent ‘‘metabolite surprises’’, a term used to describe the observation of human metabolites that were previously undetected in preclinical studies. Such surprises can have a significant impact on the timeline of a drug development program and has triggered investigations to develop in vitro systems that more accurately predict human in vivo metabolism. In this study, the metabolite profiles of eight model pharmaceutical compounds with various biotransformation reactions were investigated in parallel using a functionally stable model of primary hepatocytes [micropatterned co-cultures (MPCCs), referred to as HepatopacTM] and the traditional liver microsomal and suspension hepatocyte platforms (Khetani & Bhatia, 2006). Incubations were performed for 0 and 60 min in liver microsomes; 0 and 2 hin suspension hepatocytes; and 0, 4 hr, 2 days, and 7 days in HepatopacTM. Metabolites were identified by LC/MS employing a narrow range of chromatographic conditions, which were representative of drug metabolism screening parameters used in an early drug development setting. Our results show that HepatopacTM are metabolically more active than the traditional platforms, as judged from levels of test article disappearance and formation of metabolites. Formation of metabolites from non-P450 mediated reactions, especially UGTs, and secondary metabolites appear to be more predominant in HepatopacTM. Results from this investigation also suggest that use of HepatopacTM for metabolite profiling studies will be broadly useful in minimizing ‘‘metabolite surprises’’. Use of the HepatopacTM platform also provides cells that maintain metabolic activity over an extended period in culture; a difference that likely explains the increased abundance of secondary metabolites in this system.

Reference Khetani SR, Bhatia SN. (2006). Engineering tissues for in vitro applications. Curr Opin Biotechnol 17:524–531.

P316. IN VIVO RAT METABOLISM OF AN ACUTE MYELOID LEUKEMIA (AML) DRUG CPX-351 (CYTARABINE:DAUNORUBICIN) LIPOSOME INJECTION Heasook Kim-Kang1, Yijun Yi1, Peter L. Wang1, Paul Tardi2, Sherwin Xie2, and Lawrence Mayer2 1 Drug Metabolism and Biotransformation, XenoBiotic Laboratories, Inc., Plainsboro, NJ, USA and 2Celator Pharmaceuticals Corp., Vancouver, Canada CPX-351 liposome Injection is a liposomal formulation of a synergistic 5:1 molar ratio of cytarabine and daunorubicin, two agents commonly used to treat hematologic malignancies, particularly acute myeloid leukemia (AML). CPX-351 has been reported to be highly active in every subgroup of older patients with newly diagnosed AML and produced the greatest benefit in patients with high risk and poor prognosis. Sprague Dawley male and female rats, both intact and bile duct-cannulated (BDC), were administered a single IV bolus dose of 15 units/kg (100 mCi/kg) of [14C]CPX-351. In addition, the rats were administered with a cocktail mixture of cytarabine and daunorubicin in saline solution (i.e. in free form). Blood, plasma, urine, feces, and bile were collected for mass balance, metabolite profiling, and identification. The metabolite profiles in plasma, bile, urine, and feces collected from the rats administered with [14C]CPX-351 as liposomal injection or a free form were compared. The elimination was slower when rats were administered in a liposomal form (vs free form), which is an important contributor to the enhancement of the efficacy of the drug. Unchanged daunorubicin (with daunorubicinol as a minor component) or unchanged cytarabine (with N-arabinosyluracil as a minor component) were the major components in plasma, urine and bile. In general, the percent distributions of daunorubicin or cytarabine and their major metabolites were similar between the two different dose groups (i.e. liposome injection vs. saline injection). In addition, no significant gender difference was observed in metabolite profiles. Although daunorubicin was eliminated mainly in bile as an unchanged form, the fecal metabolites indicated its transformation to other compounds likely by intestinal flora, which were then excreted since none of fecal metabolites were detected in plasma, urine, or bile. The fecal bacterial transformation involved the deconjugation of daunorubicin, followed by oxidation, dehydration and O-acetylation. Both daunorubicin and cytarabine in CPX-351 remained intact with minimal metabolism. The fecal metabolites derived from daunorubicin were formed through intestinal flora and eliminated without any evidence of recirculation.

P317. LIPIDOMIC PROFILING IN SKIN FROM DB/DB AND NORMAL MICE USING DIRECT-INFUSION NANOELECTROSPRAY-TANDEM MASS SPECTROMETRY Jong-Cheol Shon1, Zhexue Wu1, Eung Ho Choi2, and Kwang-Hyeon Liu1 1 College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, South Korea and 2Department of Dermatology, Yonsei University Wonju College of Medicine, Wonju, South Korea

188

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Diabetes mellitus (DM) is a metabolic disease with multiple complications that cause various diseases over the years. DM has known to induce many pathophysiologic changes in skin stratum corneum. This study investigated the differences in lipidomic profiling between db/db and normal mice skin (n = 5). Lipidomic profiling was analyzed direct-infusion nanoelectrospraytandem mass spectrometry. The db/db and normal groups were clearly discriminated from each other on PLS-DA and OPLS-DA score plot and major metabolites contributing to the discrimination were assigned as ceramides. The db/db mice showed highe levels of N-acylsphingosine (NS ceramide), and lower levels of N-acylphytosphingosine (NP) and N-acyldihydrosphingosine (NdS) than normal mice. Additionally, the pH in skin of db/db mice is significantly higher than that of normal mice (p50.05). In PLS analysis, NS ceramide showed positive correlation with skin pH, whereas NdS and NP ceramide showed negative correlation with skin pH. On the basis of these metabolites, we proposed a metabolic pathway related to diabetes-related diseases. These ceramides can be used to better understand diabetes and related diseases. Furthermore, the level changes of these ceramides can be used to assess the risk of diabetes dermatitis and the therapeutic effect of diabetes dermatitis management.

P318. THE PHARMACOKINETICS OF DICLOFENAC AND ITS METABOLITES USING THE CHIMERIC TK-NOG MOUSE WITH HUMANIZED LIVER

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Satoshi Ito1, Kohei Nozawa1, Hiroyuki Chijiwa1, Shota Nakamura1, Hiroshi Suemizu2, Yasuyuki Ohnishi2, Hidetaka Kamimura1, and Shin-ichi Ninomiya1 1 Sekisui Medical Co., Ltd., Tokai, Ibaraki, Japan and 2Biomedical Research, Central Institute for Experimental Animals, Kawasaki, Japan Interspecies differences in drug metabolism have made it difficult to use preclinical animal testing data to predict the drug metabolites that will occur in humans. Chimeric mice with humanized liver are very useful for the prediction of circulating human metabolites before entering the first-in-man trials (Kamimura et al. 2010). Recently, chimeric TK-NOG mice with humanized livers have been developed. In this new system, a herpes simplex virus type 1 thymidine kinase transgene was expressed in the liver of a highly immunodeficient mouse strain (NOG) (Ito et al., 2002) to produce the NOG mouse expressing a thymidine kinase transgene (TK-NOG) (Hasegawa et al. 2011). Brief exposure to a nontoxic dose of ganciclovir causes a rapid and temporally controlled ablation of mouse liver cells expressing the transgene, which enabled the transplanted human liver cells to develop into a mature human organ with 3-dimensional architecture and a gene expression pattern including many human drug-metabolizing enzymes and transporters. The purpose of this study was to evaluate the applicability of this novel model to predicting human metabolites of drugs using diclofenac as model compounds. 30 -Hydroxy-40 -methoxydiclofenac (VI) is known to accumulate in human plasma after repeated administration of diclofenac sodium (Faigle et al., 1988). After single oral administration of diclofenac to the chimeric TK-NOG mouse, the unchanged drug in plasma peaked at 0.25 h and then declined with t1/2 of 2.4 h. 40 Hydroxydiclofenac (II) and 30 -hydroxydiclofenac also peaked at 0.25 h and disappeared within 24 h. On the contrary, metabolite VI reached the maximum at 8 h and declined with t1/2 of 13 h. This metabolite VI was hardly detected in plasma after oral administration to control TK-NOG mouse. When diclofenac was given once per day for 14 days, peak and trough levels of VI plateaued within 3 days. To clarify the metabolic route of forming VI, II was orally administered to a control TK-NOG and chimeric TK-NOG mouse. The plasma concentration of II reached the maximum at 0.25 h and declined rapidly in both animals, but tended to be higher in the chimeric TK-NOG mouse. When VI was analyzed, the chimeric TK-NOG mouse exhibited a similar plasma profile to that after administration of diclofenac suggesting that VI was generated via II as an intermediate. TK-NOG chimeric mice with humanized liver were considered to be a functional tool for the study of drug metabolism in humans.

References Faigle et al. (1988). Xenobiotica 18:1449–1455. Hasegawa M, Kawai K, Mitsui T, et al. (2011). The reconstituted ‘humanized liver’in TK-NOG mice is mature and functional. Biochem Biophys Res Commun 405:405–410. Ito M, Hiramatsu H, Kobayashi K, et al. (2002). NOD/SCID/gamma(c)(null) mouse: An excellent recipient mouse model for engraftment of human cells. Blood 100:3175–3182. Kamimura H, Nakada N, Suzuki K, et al. (2010). Assessment of chimeric mice with humanized liver as a tool for predicting circulating human metabolites. Drug Metab Pharmacokinet 25:223–235.

P319. INDIVIDUAL DIFFERENCE IN METABOLIC CAPACITY BY CYP3A5 GENETIC POLYMORPHISM AFFECTS CROSS-REACTIVITY IN VARIOUS TACROLIMUS ASSAYS Tetsuya Kaneko1, Takashi Fujioka1, Yosuke Suzuki1, Toshiaki Nagano2, Fuminori Sato3, Yuhki Sato1, Shunji Asakura2, Hiromitsu Mimata3, and Hiroki Itoh1 1 Department of Clinical Pharmacy, Oita University Hospital, Oita, Japan, 2Department of Pharmacy, Oita Red Cross Hospital, Oita, Japan, and 3Department of Urology, Oita University Faculty of Medicine, Oita, Japan Introduction: The calcineurin inhibitor tacrolimus has a narrow therapeutic index and shows considerable interindividual variability in pharmacokinetics. One contributing factor to the inconsistent pharmacokinetics of tacrolimus may be the variable expressions of functional CYP3A4, CYP3A5, and P-glycoprotein efflux pumps. CYP3A5 genetic polymorphism is

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

189

known to affect the pharmacokinetics of tacrolimus in stable transplantation recipients. Carriers of a CYP3A5*1-allele exhibit substantially higher tacrolimus clearance resulting in markedly higher tacrolimus dose requirement compared with homozygous CYP3A5*3/*3 patients. Therapeutic drug monitoring (TDM) of tacrolimus is important to optimize efficacy and minimize toxicity and intolerance. The aim of this study is to evaluate cross-reactivity in tacrolimus antibodyconjugated magnetic immunoassay (ACMIA), chemiluminescent enzyme immunoassay (CLIA) and enzyme multiplied immunoassay technique (EMIT), and to analyze the effect of CYP3A5 genetic polymorphism on the cross-reactivity. Methods: Thirty-nine Japanese kidney transplant recipients who underwent transplantation more than 180 days prior to the study were included. Blood samples were obtained from routine TDM of tacrolimus, and tacrolimus concentrations were measured by ACMIA, CLIA, EMIT and liquid chromatography-tandem mass spectrometry (LC-MS/MS). CYP3A5 genetic polymorphism (CYP3A5*3) was analyzed by TaqMan genotyping assays. The ratios of tacrolimus concentrations measured by ACMIA, CLIA or EMIT to LC-MS/MS for different CYP3A5 genotypes were compared using independent-samples ttest. This study was approved by the ethics committee of Oita University. Each subject received information about the scientific purpose of the study, and gave written informed consent. Results and Discussion: Seventeen patients were CYP3A5*1 allele carriers and 22 patients were homozygous (CYP3A5*3/*3) for the CYP3A5*3 allele. Blood tacrolimus concentrations measured by ACMIA, CLIA and EMIT correlated significantly with those by LC-MS/MS (r = 0.91, p50.001, r = 0.92, p50.001 and r = 0.78, p50.001, respectively). The ACMIA/LC-MS/MS and CLIA/LC-MS/MS ratios were significantly different between patients with CYP3A5*1 allele and CYP3A5*3/*3 (p = 0.019 and 0.039), while the EMIT/LC-MS/MS ratio was not significantly different (p = 0.092), although a tendency of influence by CYP3A5 genetic polymorphism was observed. [CONCLUSION] These results suggest that the high metabolic capacity associated with the CYP3A5*1 genotype may affect the cross-reactivity in ACMIA, CLIA and EMIT.

P320. CURRENT INDUSTRY PRACTICES IN THE IN VIVO ASSESSMENT OF HUMAN DRUG METABOLISM: A SURVEY BY THE DRUG METABOLISM AND CLINICAL PHARMACOLOGY LEADERSHIP GROUPS OF THE IQ CONSORTIUM Cornelis E. C. A. Hop1, J. G. Slatter2, L. J. Christopher3, T. Edeki4, P. D. Gorycki5, J. Keirns6, M. L. Marshall7, S. D. Oliver8, C. Prakash9, D. K. Spracklin10, R. Subramanian11, Z. Zeng12, N. G. Agrawal13, and S. K. Chowdhury14 1 Drug Metabolism & Pharmacokinetics, Genentech, South San Francisco, CA, USA, 2Amgen, Seattle, WA, USA, 3Bristol Myers Squibb, Princeton, NJ, USA, 4AstraZeneca, Wilmington, DE, USA, 5GlaxoSmithKline, King of Prussia, PA, USA, 6Astellas, Northbrook, IL, R.D. Lee, Takeda, Deerfield, IL, USA, 7Drinker Biddle & Reath LLP, Washington, DC, USA, 8AstraZeneca, Macclesfield, UK, 9Biogen Idec, Cambridge, MA, USA, 10Pfizer, Groton, CT, USA, 11Amgen, Thousand Oaks, CA, USA, 12 Sanofi, Bridgewater, NJ, USA, 13Merck, West Point, PA, USA, and 14Takeda, Cambridge, MA, USA Background: The Safety Testing of Drug Metabolites, and ICH M3(R2) guidances and new technologies such as accelerator mass spectrometry & high resolution LCMS/MS (HR-LCMS/MS) have changed how the pharmaceutical industry investigates clinical drug metabolism. The International Consortium for Innovation and Quality in Pharmaceutical Development (IQ Consortium) undertook a survey of industry practices related to in vivo human drug metabolism and its implications for preclinical drug metabolism studies. Twenty-six member companies, comprising a majority of large and medium pharmaceutical companies, responded and the survey results will be presented in detail in this poster. Methods: The electronic survey was distributed in April 2013. The company affiliations were removed prior to data compilation to ensure anonymity. Results: Clinical Radiolabeled Study Metrics: In 2008–12, 26 companies conducted 230 clinical radiolabeled studies. An oral dose was used in the majority (78%) of studies, but a relatively large (19%) number of intravenous studies were performed as well. Typically, 5–6 subjects were dosed with 75–125 mCi of 14C for the oral studies. Most of the studies were in healthy volunteers, but 11% of the studies involved subjects with relevant disease states. The majority (88%) of the companies outsourced the human ADME study, but the metabolite profiling was mostly done in house (66–75%). Microtracer Studies by AMS: A microtracer (1 mCi of 14C) radiodose was used in 21% of studies (48 studies in total). Of these, 48% were absolute bioavailability and 43% were metabolism/excretion studies; 10% of the studies were done under an exploratory IND. Timing of clinical and preclinical studies: In general, release of the MIST guidance had not changed or slightly delayed the timing of preclinical and clinical metabolism and excretion studies although some respondents indicated that they performed the human studies earlier. Non-radioactive Human Metabolite Profiling: Most companies leveraged HR-LCMS/MS to profile metabolites in Phase 1 studies and these data appear to drive decision making for identification of potential disproportionate metabolites. Bioanalysis: HR-LCMS and AMS techniques and sample pooling strategies are commonly used. Several companies have successfully used semi-quantitative non-radiolabeled plasma metabolite data to make a case to regulators that a major human metabolite was not disproportionate and did not require additional studies. MIST Guidance Strategy: Many companies have initiated a dialogue with regulators regarding potential disproportionate metabolites and about 60% of companies apply further diligence on metabolites that are 410% of total-drug related material to proactively address potential MIST issues. Conclusion: Companies routinely conduct human 14C studies, and now most companies also routinely profile plasma metabolites by HR-LCMS in Phase 1 SAD and MAD studies. Since the 2008 MIST guidance, new technologies have enabled new approaches to plasma metabolite profiling and provide relevant metabolism information earlier in clinical development.

190

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P321. THE USE OF RADIOLABELS IN STUDYING ADME PROPERTIES OF BIOTHERAPEUTIC PROTEINS: EFFECT OF LABELLING ON PHYSICAL PROPERTIES OF PROTEINS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Andrew McEwen, Stuart Wood, Stephanie Geoffroy, Stephen Harris, and Claire Henson Quotient Bioresearch Limited, Rushden, UK The development of protein based pharmaceuticals has stimulated interest in both the analytical characterization of large proteins and in developing an understanding of their heterogeneity and degradation pathways. Development of these molecules requires a basic understanding of their pharmacokinetic properties and biodistribution. The gold standard for studies of adsorption, distribution, metabolism and excretion (ADME) remains the use of radiolabelled drugs. Unlike small molecule ADME studies which traditionally utilise carbon-14 labelled material the use of radiolabels to study large molecules is less well established. The radiolabel needs to be located in a stable position within the molecule, provide a high specific activity and have no impact upon the biological activity of the test compound. A large number of enzymatic modifications as well as chemical and physical degradations may occur in vivo and an understanding of the relative importance these processes is key to interpreting the results obtained. Modifications to the original structure can influence disposition and half-life based on conformational changes, alteration of the binding efficiency/specificity and changes in surface charge. Proteins such as albumin or monoclonal antibodies may also be used as drug carriers thus taking advantage of their exceptionally long half-life of serum albumin or IgG. The long half-lives observed are attributed to interaction with the neonatal Fc receptor (FcRn) and thus any modifications to the native structure may affect the in vivo behavior of the product. Low-molecular weight drugs and bioactive proteins have therefore been conjugated to albumin, or encapsulated into albumin nanoparticles to improve in vivo stability and alter the tissue distribution of the therapeutic agent. In addition ADME studies in biotherapeutics generally use chemical linkers to add the radiolabel to the protein. This study was performed to provide information on the effects of radiolabelling on the biological activity, surface charge and pharmacokinetic properties of proteins following radiolabelling. Proteins (apomyoglobin, albumin and myoglobin) were labeled using both [3H]- and [14C]-N-Succinimidyl Propionate, a lysine specific reagent and [14C]-N-Ethyl Maleimide, a cysteine specific reagent. Proteins were purified using a size exclusion column and checked by HPLC using BioSep SEC-2000 column (300  7.8 mm) with mobile phases 100 mM Sodium phosphate pH6.8 and 300 mM Sodium Chloride. Surface charge was measured for proteins both pre- and post-labelling using ion exchange (IEX) chromatography with a ProteinPak HiRes Q (anion exchange) column at a temperature of 30  C. The mobile phases were: (A) 20 mM tris-buffer, pH 8–9 and (B) 20 mM tris-buffer, 1 M NaCl, pH 8–9. The aim of the current study was to investigate the effect of the radiolabelling method on surface charge and physical properties of proteins. It has been shown that the radiolabelling methods employed had minimal effect upon the physical properties of the proteins studied thus giving confidence in the results of subsequent ADME studies performed.

P322. APPLICATION OF LIQUID SCINTILLATION COUNTER TO PHARMACOKINETICS STUDIES: EVALUATION OF PLASTIC SCINTILLATOR FOR MEASURING RADIOLABELED COMPOUNDS IN BLOOD Yusuke Tomozoe1, Yuka Kato1, Kiyoshi Ogiwara1, Tomoyuki Yoshimura1, and Etsuko Furuta2 1 Measuring System Engineering Department, Hitachi Aloka Medical, Ltd., Tokyo, Japan and 2Ochanomizu University Graduate School, Tokyo, Japan Purpose: Liquid Scintillation Counter (LSC) has often been used to measure radiolabeled compounds in blood in the pharmacokinetics studies. Although the measurement sensitivity of LSC is high, there are problems that radioactive waste solvent is generated after every measurement and that the measured sample is not able to be reused. To solve these problems, Furuta et al. proposed the use of the plastic scintillator (PS) for LSC measuring instead of liquid scintillator and showed that it can measure 3H and 14C with high counting efficiency without chemical quenching (Furuta et al., 2009). In this study, to investigate whether this method can be applied to measurement of radiolabeled compounds in blood, we evaluated the count rate linearity and the quenching status in 14C measuring in serum. Methods: In order to evaluate the count rate linearity and the quenching status in serum, 14C-labelled methionine was added to mouse serum, and the mixture was used as sample (approx. 0.5, 2, 20, 60, 100 Bq/10 mL). Hydrogen peroxide was used as bleaching agent. PS used was BC-400 (Saint-Gobain); 0.5 mm thickness sheet was cut in the size of 13 mm in width and 45 mm in length. Ten microliter of samples were applied to the PS sheet and incubated for 2 h at room temperature for dry. After dry-up, another new PS sheet was put on the sheet and the pair of PS sheets was put in a 20 mL glass vial. Samples were measured by LSC-6101B (Hitachi Aloka Medical, Ltd.) and spectrum was analyzed. Results and discussions: In the count rate linearity evaluation, both serum-free and serum sample showed good linearity between 2 and 100 Bq/10 mL. In serum sample the count decreased and the spectral shape became different to serumfree sample. This may be caused by color quenching which occurs after emitting scintillation light. The serum sample, to which hydrogen peroxide was added, showed the same level of the count and the same spectral shape to serum-free sample. This indicates that color quenching can be eliminated by bleaching reagent. Conclusion: LSC measuring by PS sheet showed good count rate linearity in 14C in serum. Color quenching occurs in serum sample, but this can be eliminated by hydrogen peroxide. In this method sample volume can be lowered to sub microliter and sample can be reused. Thus, LSC measuring by PS sheet was shown to have a potential to be applied to measurement of radiolabeled compounds in blood.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

191

Reference Furuta E, Yokota S, Watanabe Y. (2009). Identification of beta nuclides measured by using plastic scintillator and liquid scintillation counter. Radiocarbon LSC 2008:19–26.

P323. APPLICATION OF LOW BACK GROUND LIQUID SCINTILLATION COUNTER TO PHARMACOKINETICS STUDIES: SENSITIVE QUANTIFICATION OF RADIOCARBON (14C) IN HUMAN URINE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Yuka Kato1, Masanobu Horie2, Mashiho Yanagi2, Tomoyuki Yoshimura1, and Shigeo Baba2 1 Measuring System Engineering Department, Hitachi Aloka Medical, Ltd., Tokyo, Japan and 2Tokyo University of Pharmacy and Life Sciences, Tokyo, Japan Purpose: To promote new drug development efficiency, microdose clinical trials of administering radiolabeled ultra-trace compounds to human at early stage, have been performed. This is for selecting a development candidate compound that exhibits optimal pharmacokinetics. In microdose clinical trials, Accelerator Mass Spectrometry (AMS) is recommended to use since radioactivity of the administered compound is quite low. Then, we have proposed the use of Low Background Liquid Scintillation Counter (Low BG LSC) for microdose clinical trials because there is no memory effect to the equipment and sample preparation is so simple. Wehave studied the ultrasensitive radiocarbon measurement in human urine using a Low BG LSC (Baba et al., 2009; Horie et al., 2010). Here, we will report the measurement result using 145 mL vial and the variations of 40K and endogenous 14C such as intra-day, inter-day and individual. Method: Human urine sample was collected from 5 volunteers (adult male) in accordance with the time schedule of each. Five milliliter or 20 mL of human urine was added to 15 mL or 80 mL of liquid scintillator (Ecoscinti Ultra), respectively and mixed with the amount of 14C defined in each experiment. For internal standard or linearity examination, 14C (1 Bq) or 14C (0.02, 0.04, 0.1, 0.2, 0.4 Bq) was used respectively. Mock urine prepared using methyl orange was used as background sample with water sample. To evaluate the effect of color quenching hydrogen peroxide was used as bleaching agent. The correction curve for double labeling (14C and 40K) was created using a methyl orange as a quencher. Samples were measured for 100 min by Low BG LSC (LSC-LB5B or LSC-LB7, Hitachi Aloka Medical, Ltd) and correction curve samples were measured for 60 min. Results and discussions: Measurement accuracy is greatly improved by using 145 mL vial in the optimal window and good linearity was shown between 0.001 and 0.02 Bq/mL. In human urine BG variations were shown (intra-day, inter-day and individual) and found to be caused by endogenous 40K. These variations became ignorable using the optimal window and the internal standard method without consideration of the color quenching effect. Further the double labeling correction curve made it possible to measure14C and 40K separately and a good correction result was obtained up to 0.05 Bq/mL. Conclusion: Low BG LSC was able to measure the sample ranged from 102 to 106 Bq/vial without the memory effect and double-labeled sample measurement was effective. Moreover, it has advantages that sample preparation is simple, urine sample varied from few uL to about 40 mL can be measured, and running cost is much lower than AMS. Low BG LSC was shown to be highly effective for microdose clinical trials.

References Baba S, Horie M, Yanagi M, et al. (2009). Application of low background liquid scintillation counting method to pharmacy: Part 1 – Measurement of ultra-low level 14C activity in urine. Radioisotopes 58:169–178. Horie M, Yanagi M, Baba S, et al. (2010). Application of low background liquid scintillation counting method to pharmacy: Part 2. Variation of endogenous 14C in human urine. Radioisotopes 59:599–605.

P324. UTILIZATION OF STABLE ISOTOPE LABELING TO FACILITATE THE IDENTIFICATION OF VERY POLAR METABOLITES IN RAT URINE Su-Er W. Huskey, Hongmei Li, Alexandre Catoire, Zhigang Jian, Jin Zhang, Jimmy Flarakos, Tapan Ray, and James Mangold Department of Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, East Hanover, NJ, USA Identification of polar metabolites of small molecule drug candidates in drug development is often challenging. Several major polar metabolites of compound Y (partial structure shown below) were detected in rat urine from the radiolabeled ADME study, but could not be characterized by LC-MS/MS analysis. In such instances, a strategy chosen by many investigators is to use a high specific activity radiolabeled version of the test compound, having the enriched isotope abundance ratio (i.e. [monoisotopic mass]:[mass with 1 atom of radioactive isotope]) serve as a mass filter. In this case, however, [14C]compound Y was uniformly labeled, such that a distribution of x = 1–4 14C atoms were present in the radiolabeled compound Y molecule, and the [all 12C] to [14Cx = 1–4] isotope abundance ratio was only 0.1 or lower. This ratio was insufficient to be useful as a mass filter despite having a high specific activity (120 mCi/mg). Since a stable-label version of compound Y was already available for use as an internal standard of compound Y, we decided to use it to enhance the isotope ratio as a mass filter. Accordingly an oral dose was designed, with a mixture of [13C6]compound Y, compound Y, and trace [14C] compound Y (specific activity 3.65 mCi/mg), such that the ratio of [all 12C] to [13C6] was 0.9. In addition, a new separation method was developed to retain the polar metabolites longer on the HPLC column. Using this strategy, four polar metabolites (M1–M4) were successfully identified and characterized

192

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

in rat urine. Overall this alternative approach was shown to be practical and valuable for the identification of polar metabolites by LC-MS/MS. The details of this approach are discussed in this presentation. Partial structure of compound Y

P325. DIFFERENCES FOR BOVINE SERUM ALBUMIN EFFECT ON GLUCURONIDATION ACTIVITY IN RAT, MONKEY, AND HUMAN LIVER MICROSOMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Fumihiro Nakamori1 and Yoichi Naritomi2 1 Analysis & Pharmacokinetics Research Labs., Astellas Pharma Inc., Osaka-shi, Japan and 2Analysis & Pharmacokinetics Research Labs., Astellas Pharma. Inc., Tsukuba-shi, Japan Background: Bovine serum albumin (BSA) stimulation of glucuronidation activity in liver microsomes has recently gained attention as a method of preventing underestimation of in vitro glucuronidation. Although the effect of glucuronidation in human liver microsomes has been investigated, whether or not glucuronidation in liver microsomes is stimulated by BSA and if this effect is subject to species differences is unknown. Here, we investigated BSA stimulation of glucuronidation activity in liver microsomes of rats, monkeys, and humans. Methods: Buprenorphine, propofol, and zidovudine, which are mainly metabolized by human UGT1A1, 1A9, or 2B7, were selected as model substrates. These compounds were incubated with rat, monkey, or human liver microsomes in the presence or absence of BSA, and in vitro glucuronidation clearance (CLint, in vitro) values were determined. Results: When incubated with buprenorphine, BSA increased CLint, in vitro values 15.5-fold in liver microsomes of rats, 14.9-fold in those of monkeys, and 4.1-fold in those of humans. When incubated with propofol, BSA increased CLint, in vitro values 33.9-fold in liver microsomes of rats, 3.0-fold in those of monkeys, and 65.9-fold in those of humans. When incubated with zidovudine, BSA increased CLint, in vitro values 4.7-fold in liver microsomes of rats, 8.6-fold in those of monkeys, and 6.1fold in those of humans. Conclusions: BSA increased glucuronidation activities in the liver microsomes of rats, monkeys, and humans, albeit with marked species differences. Further investigation is required to clarify the causes of these differences between species and substrates in effect of BSA on glucuronidation in liver microsomes.

P326. AN INTEGRATED STRATEGY TO DETERMINE THE ABSORPTION, DISTRIBUTION, METABOLISM, AND EXCRETION, AND THE ABSOLUTE BIOAVAILABILITY OF BEVENOPRAN (CB-5945) IN HEALTHY MALE SUBJECTS Adedayo Adedoyin1, Robert Shibuya1, Stuart Mair2, Lloyd Stevens2, Ian Nisbet3, Stuart Wood4, Iain Shaw2, and Lee Techner1 1 Cubist Pharmaceuticals, Lexington, MA, USA, 2Quotient Clinical, Nottingham, UK, 3Project Management, Quotient Clinical, Nottingham, UK, and 4Quotient Bioresearch Ltd, Rushden, UK Introduction: Bevenopran (CB-5945) is a peripherally acting mu opioid receptor antagonist being developed to treat opioidinduced constipation (OIC) in patients with chronic non-cancer pain. This integrated two part cross-over study assessed the absorption and disposition of bevenopran after intravenous (IV) and oral administrations to determine mass balance, absolute bioavailability, routes of excretion, fraction absorbed and metabolism in six healthy male subjects. The ability to formulate and manufacture both IV and oral drug products at microtracer and therapeutic doses respectively; along with clinical conduct of the study and the ability to correctly implement accelerator mass spectrometry (AMS) to provide the differential analysis required for 14C IV drug product is critical for the success of this study approach. Subjects (or Materials) and Methods: This was a crossover study in which 6 healthy male volunteers were administered IV and oral radiolabeled doses of bevenopran. In Part 1, subjects received a single oral dose of 0.5 mg bevenopran solution followed after 1 hour by a 15 minute IV infusion of 5 mg [14C]-bevenopran (containing not more than [NMT] 1000 nCi 14C) In Part 2, subjects received a single oral dose of 0.5 mg [14C]-bevenopran solution (containing NMT 811 mCi 14C). Blood, urine, feces and bile samples were collected and assayed for total radioactivity and parent drug. Results: An integrated approach that met all regulatory requirements was established to enable the conduct of this study. This strategy enabled the characterization of the IV and oral pharmacokinetics of bevenopran. Mass balance data showed greater than 90% of the administered dose was excreted in urine and feces. Bevenopran showed good oral bioavailability and was well tolerated in normal healthy male volunteers. Details of the results will be described. Conclusion: This integrated study design is a highly efficient mechanism in terms of material, time and cost for generating definitive IV PK, absolute bioavailability, mass balance and human metabolism data from a single radiolabelled clinical protocol.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

193

P327. THE EFFECT OF BUFFER IONIC STRENGTH OR VARIOUS MEDIA ON THE IN VITRO METABOLISM OF CYTOCHROME P450 SUBSTRATES IN POOLED HUMAN LIVER MICROSOMES AND CRYOPRESERVED HUMAN HEPATOCYTES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Faraz Kazmi1, Phyllis Yerino1, and Andrew Parkinson2 1 XenoTech, LLC, Lenexa, KS, USA and 2XPD Consulting, Shawnee, KS, USA In our previous work, we reported that neither cofactor availability nor membrane permeability accounted for the much slower in vitro clearance of midazolam in suspended cryopreserved human hepatocytes (CHH) compared with human liver microsomes (HLM). We posited the difference was possibly an effect of ionic strength (Kazmi et al., 2013a, 2013b). In the present study we evaluated the effects of buffer ionic strength and various media on CYP1A1, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and multiple CYP3A4/5 activities in HLM and CHH. CYP activities were measured at three ionic strengths (5, 50 and 200 mM phosphate buffer) and/or in five commonly used cell culture media (KHB, MCM + , DMEM + HEPES, Waymouth’s and William’s E + HEPES) in both HLM and CHH. NADPH-fortified pooled HLM (n = 200) at 0.1 mg/mL were incubated for 5 min with phenacetin, coumarin, bupropion, amodiaquine, diclofenac (tolbutamide for CHH), S-mephenytoin, dextromethorphan, chlorzoxazone, midazolam, nifedipine, alfentanil, verapamil, testosterone and atorvastatin at their approximate Km values. In the case of pooled CHH (n = 50), incubations were conducted at 1 million cells/mL for 10– 60 min. Reactions were terminated with an equal volume of organic solvent containing internal standard, followed by protein precipitation and analysis by LC/MS/MS. In HLM, phenacetin (CYP1A2), coumarin (CYP2A6), bupropion (CYP2B6), amodiaquine (CYP2C8), diclofenac (CYP2C9), S-mephenytoin (CYP2C19), and dextromethorphan (CYP2D6) activities all were highest at 50 mM phosphate buffer. For chlorzoxazone (CYP2E1) and midazolam (CYP3A4/5), the enzymatic activities were highest at 200 mM phosphate buffer. When incubations in HLM were conducted with various media, MCM + medium was found to support the highest activity of CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C19 and CYP2D6. For CYP2C9, CYP2E1 and CYP3A4/5, William’s E and KHB media were found to support the highest enzymatic activity levels, though the activity was approximately half of 50 mM phosphate buffer activity. With CHH there were modest differences in enzymatic activity between the five media with the exception of CYP2C19, where the highest activity was supported by MCM + medium, and CYP3A4/5, where the highest activity was supported by William’s E medium. When multiple substrates of CYP3A4/5 were assessed in HLM, all of them had maximal activity at 200 mM phosphate buffer. In contrast to midazolam, the five other CYP3A4/5 substrates (nifedipine, alfentanil, verapamil, testosterone, and atorvastatin) were all metabolized by HLM to a similar extent in all five media. The results of this study suggest that the slower clearance of midazolam in CHH relative to HLM is perhaps unique to midazolam because no such marked difference was observed with five other CYP3A4/5 substrates. CYP3A4/5 activity in HLM was found to be highly dependent on ionic strength, with William’s E and KHB media in both HLM and CHH providing the best midazolam activity levels. These findings have implications for optimizing in vitro conditions to measure drug clearance and shed further light on the unusual behavior of midazolam clearance.

References Kazmi F, Yerino P, Parkinson A. (2013a). Drug Metab Rev 45:162. Kazmi F, Yerino P, Parkinson A. (2013b). Drug Metab Rev 45:163.

P328. EFFECT OF BUFFER CONDITIONS ON CYP2C8-MEDIATED 6a-HYDROXYLATION OF PACLITAXEL BY HUMAN LIVER MICROSOMES Toshiyuki Kudo, Yuya Ozaki, Eri Hotta, Aya Matsubara, and Kiyomi Ito Department of Clinical Pharmacokinetics, Musashino University, Tokyo, Japan Purpose: In in vitro metabolism studies using human liver microsomes, the metabolic activities of various cytochrome P450 (CYP) enzymes have been reported to depend on the buffer components and their concentration. While the buffer concentration differentially affects the metabolic activity of CYP enzymes such as CYP1A2, CYP2A6, CYP2C9, CYP2C19, CYP2D6, and CYP3A4, no study has reported whether that of CYP2C8 depends on buffer conditions. The present study investigated the effect of buffer component and concentration on the metabolic activity of CYP2C8 using paclitaxel 6a-hydroxylation as a marker reaction. Methods: Paclitaxel (2.5–50 mM) was incubated at 37  C for 30 minutes with pooled human liver microsomes (0.25 mg/ mL) in the presence of NADPH-regenerating system. Potassium phosphate buffer or Tris-HCl buffer was used with the concentrations of 10, 50, 100 or 200 mM. The reaction was stopped by mixing with ice-cold ethyl acetate and, after centrifugation, the amount of 6a-hydroxypaclitaxel in the upper layer was determined by HPLC-UV. The 6a-hydroxypaclitaxel formation rates were fitted to the Michaelis–Menten equation to estimate the maximum velocity (Vmax) and Michaelis constant (Km) for respective buffer conditions, and the intrinsic clearance (CLint) was calculated as Vmax divided by Km. Results and Discussion: The estimated Km values were 1.2- to 2.3-fold higher for Tris-HCl buffer than for phosphate buffer when the values were compared at the same buffer concentrations. While the Vmax value showed a dramatic 4.4-fold increase with increasing the concentration of Tris-HCl buffer from 10 mM to 200 mM, the value was maximum at 50 mM when phosphate buffer was used. About 3-fold higher CLint was obtained with low concentrations (10, 50 mM) of phosphate buffer compared with Tris-HCl

194

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

buffer, while the CLint values were comparable between phosphate buffer and Tris-HCl buffer at high concentrations (100, 200 mM). The buffer concentration dependency was also observed for the CLint value; in particular, more than 2-fold difference was observed between 50 mM and 100 mM Tris-HCl buffer, which are both widely used in in vitro metabolism studies. The observed effects of buffer conditions on paclitaxel 6a-hydroxylation were different from those reported for CYP3A4-mediated aand 4-hydroxylation of midazolam. Conclusion: The metabolic activity of CYP2C8 by human liver microsomes was suggested to depend on both component and concentration of the buffer used, with a pattern which is different from that for other CYP enzymes. These findings suggest the need to optimize the buffer conditions in the in vitro–in vivo scaling of metabolic clearances and also when estimating the contribution of each CYP enzyme in a drug metabolism involving multiple enzymes.

P329. IDENTIFICATION OF A NOVEL GLUTATHIONE CONJUGATE OF KDU691

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Helen Gu, Alexandre Catoire, Lai Wang, Heidi Einolf, and James Mangold Department of Drug Metabolism and Pharmacokinetics, Novartis, East Hanover, NJ, USA KDU691 is being developed for the treatment of malaria. However, liver safety findings were observed in early safety studies in animals. Accordingly, the potential role of metabolism was investigated to help explain what may be responsible. In vitro metabolism of KDU691 in mouse, rat, dog, monkey and human liver microsomes was investigated and a novel glutathione (GSH) conjugate of KDU691 was detected in all species suggesting bioactivation to form a reactive intermediate. The evidence for this pathway was seen in vivo, with this GSH conjugate also detected in rat plasma after a single oral dose of 500 mg/kg of KDU691. The structural characterization of the GSH conjugate was achieved by liquid chromatography-tandem mass spectrometry and NMR analyses. A tentative bioactivation mechanism has been proposed, involving cytochrome P450 (CYP) mediated hydrogen atom abstraction from the amide nitrogen of KDU691, with subsequent trapping of the nitrogen-centered radical by GSH. This mechanistic hypothesis was further tested by incubation of the desmethyl analog of KDU691 with liver microsomes. This analog would be expected to not as readily form the proposed nitrogen-centered radical species. The incubation results showed that, as predicted, the desmethyl analog did not form the GSH conjugate. Additional investigations showed that the GSH conjugate formation was primarily catalyzed by human CYP3A4 and CYP2D6, with lesser involvement of CYP1B1, CYP1A1 and CYP2J2. Subsequent studies showed that in animals, rat liver CYP3A1 and 3A2, and monkey liver CYP3A8 were able to catalyze this bioactivation in each case leading to the GSH conjugate. In light of the above observations, new back up analogs are being investigated to avoid this apparent bioactivation liability.

P330. HSP90 INHIBITOR DRUG CONJUGATES (HDC): IN-VITRO PROPERTIES AND CHARACTERIZATION OF ENZYMES INVOLVED IN THE METABOLISM OF THE CONJUGATE Ritu Singh1, Takayo Inoue1, Noriaki Tatsuta1, Dinesh Chimmanamada2, and Weiwen Ying2 1 Department of Drug Metabolism and Pharmacokinetics, Synta Pharmaceuticals, Lexington, MA, USA and 2Department of Chemistry, Synta Pharmaceuticals, Lexington, MA, USA It is well established that Hsp90 inhibitors selectively target cancer cells and are retained at high concentrations for several days. Based on this property, we designed Hsp90 inhibitor Drug Conjugates (HDCs) as a novel strategy to deliver anticancer agents directly to cancer cells. HDCs are small-molecule drugs consisting of an Hsp90 inhibitor (targeting moiety) joined to an anti-cancer agent (payload) via a cleavable chemical linker optimized for controlled release of the payload inside cancer cells. The advantage of this approach is to improve the safety profile of a cytotoxic drug and overcome drug resistance. More than 40 unique payloads have been conjugated with Hsp90 inhibitors to date. STA-12-8666 is an Hsp90 inhibitor conjugated by a cleavable linker to SN-38, the highly potent metabolite of irinotecan. STA-12-8666 has shown superior efficacy compared to irinotecan in pre-clinical cancer models with multiple cell-lines (Gerber, 2014; Sanghani et al., 2003). The purpose of this investigation was to characterize the DMPK properties of the conjugate in vitro. Efflux transport function in C2BBe1 (wild type) and MDR1, BCRP, or MRP2 knockout cell lines have shown that STA-12-8666 is not a substrate of these efflux transporters. Uptake transport studies using HEK293 cells transiently overexpressing OATP1B1/1B3 have shown that this conjugate is not a substrate or inhibitor of these transporters. In-vitro CYP450 inhibition and induction studies have shown that STA-12-8666 does not inhibit or induce these enzymes. Incubation of human liver microsomes with Phase I and II co-factors has demonstrated hydrolysis of the conjugate resulting in SN-38 (pay-load) and Hsp90 inhibitor as well as formation of two HDC–glucuronide conjugates. To better understand the metabolism of the conjugate, the carboxylesteraseenzymes involved in the hydrolysis and the UGT enzymes involved in the glucuronidation were identified. All three isoforms of carboxylesterase enzymes, CES1b, CES1c and CES2, were able to release similar levels of SN-38 from STA-128666. These enzymes are known to be expressed at various levels in tumor tissue (Sanghani et al., 2003). Using recombinant UGT 1 A enzymes, UGT1A1 and UGT1A3 were found to be the major contributors in the metabolism of STA-12-8666, with minor contribution from UGT1A6, UGT1A7, UGT1A8, UGT1A9 and UGT1A10. Overall, data from this study suggests that STA-12-8666 has better DMPK properties than irinotecan. This characterization provides a basis for understanding the mechanism of action of these novel compounds which will aid in the design and optimization of conjugates for cancer therapy.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

195

References Gerber DE. (2014). HSP90 inhibitor drug conjugate (HDC) platform. Presented at IASLC 14th annual targeted therapies of the treatment of lung cancer meeting, 22 February 2014, Santa Monica, CA. Sanghani SP, Quinney SK, Fredenburg TB, et al. (2003). Carboxylesterases expressed in human colon tumor tissue and their role in CPT-11 hydrolysis. Clin Cancer Res 9:4983–4991.

P331. ABSORPTION, METABOLISM, AND EXCRETION OF ALISERTIB (MLN8237), AN AURORA A KINASE INHIBITOR, IN PATIENTS WITH ADVANCED SOLID TUMORS OR LYMPHOMAS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Sandeepraj Pusalkar1, Xiaofei Zhou2, Yuexian Li1, Lawrence Cohen1, Mingxiang Liao1, Johnny Yang1, Saeho Chong1, Karthik Venkatakrishnan2, and Swapan Chowdhury1 1 Department of Drug Metabolism and Pharmacokinetics, Takeda Pharmaceuticals International Co., Cambridge, MA, USA and 2 Department of Clinical Pharmacology, Takeda Pharmaceuticals International Co., Cambridge, MA, USA Absorption, metabolism, and excretion of the sodium salt of alisertib (4-{[9-chloro-7-(2-fluoro-6-methoxyphenyl)-5Hpyrimido[5,4-d][2]benzazepin-2-yl]amino}-2-methoxybenzoic acid), an investigational Aurora A kinase inhibitor, was studied in 3 patients with advanced solid tumors or lymphomas. Following a single oral dose of [14C]-alisertib at 80 mCi/35 mg, an average of 88% of the total administered radioactivity was recovered in feces whereas 3% was recovered in urine over 14-days postdose. Alisertib AUCinf and t1/2 in plasma were 16802 nM*h and 22.7 h, respectively, whereas those of the total radioactivity in plasma were 47758 nM*h and 51.2 h, respectively. [14C]-alisertib was the major drug-related component in the 0–192 h AUCrepresentative pooled plasma samples, accounting for 48% of the total plasma radioactivity on average. The mean percentages of alisertib, alisertib acyl glucuronide (M1), O-desmethyl alisertib (M2), and hydroxy alisertib (M3) in 0–192 h pooled plasma were 48%, 12%, 35%, and 6%, respectively. The mean relative percentages of M1, M2, and M3 compared with alisertib were 23%, 74%, and 12% of alisertib AUC0–192h, respectively. Metabolites M1 and acyl glucuronide of hydroxy alisertib (M9) were the major drug-related components in 0–192 h pooled urine. Unchanged alisertib and M3, at 26% and 21% of the dose, respectively, were the major drug-related components in 0–192 h pooled fecal extracts. A majority of the metabolites identified in feces resulted from further metabolism of primary phase I oxidative metabolites M2 and M3. Based on the excretion data and metabolite profiles in 0–192 h pooled urine and feces, 68% of alisertib dose is excreted as oxidative metabolites. Based on in vitro CYP phenotyping, 58% of the administered dose is estimated to be cleared via CYP3A4 catalysis. Therefore, there is a potential for drug–drug interactions (DDI) when coadministered with CYP3A4 inhibitors. Additionally, based on the metabolite profiles in AUC-representative 0–192 h pooled plasma samples, M1 and M2 represent 410% of total drug-related material and only M2 is 425% of the parent drug’s exposure. Accordingly, M1, M2, and alisertib were evaluated as potential perpetrators of CYP/transporter-based DDI in vitro. Alisertib was not an inhibitor of CYP1A2, 2C9, 2C19, 2D6, or 3A4/5, but was an inhibitor of CYP2C8 (IC50 = 16.3 mM). The major circulating metabolites, M1 and M2, were not potent reversible or time-dependent inhibitors of CYP1A2, 2B6, 2C8, 2C9, 2C19, 2D6, and 3A4/5 isozymes. Alisertib was an inhibitor of P-gp, OCT2, OAT1, OAT3, and OATP (IC50 = 4.0, 6.2, 29.9, 2.9, and 5.1 mM, respectively), and a weak inhibitor of BCRP and MRP2. M1 was not a P-gp and OAT1 inhibitor (IC504100 mM), but did inhibit BCRP, OCT2, OAT3, and OATP (IC50 = 21.2, 69.4, 37.3, and 8.9 mM, respectively). M2 was not an inhibitor of P-gp, BCRP, OAT1, or OATPs up to 50 mM, but inhibited OCT2 and OAT3 (IC50 = 15.8 and 2.7 mM, respectively). Based on in vitro inhibition studies on M1 and M2 with a panel of CYP450 isozymes and transporters, there is minimal potential, at the therapeutic dose of alisertib, for M1 and M2 perpetrating DDI with coadministered drugs.

P332. IDENTIFICATION OF THE MAJOR METABOLITES OF SYL-927, A SELECTIVE SPHINGOSINE 1-PHOSPHATE RECEPTOR 1 AGONIST IN RATS AND HUMAN LIVER MICROSOMES Yan Li1, Shu Yang1, Jinping Hu1, Xiaojiang Wang2, Qiong Xiao2, Yulin Tian2, and Dali Yin2 1 Department of Drug Metabolism, Institute of Material Medica, Chinese Academy of Medical Sciences, Beijing, China and 2 Department of Synthetic Medicinal Chemistry, Institute of Material Medica, Chinese Academy of Medical Sciences, Beijing, China SYL-927(2-Amino-2{2-(40 -(2-ethyloxazol-4-yl)-[1,10 -biphenyl]-4-yl)ethyl} propane-1, 3-diol hydrochloride), is a new selective sphingosine 1-phosphate receptor 1 agonist in the development for the treatment of autoimmune disorders. Previous study showed that SYL-927 underwent the reversible biotransformation to an active phosphate (SYL-927-P), and the excretion of parent drug and SYL-927-P were considerably low in feces and urine of rats after oral administration, so SYL927 could be eliminated via other metabolic pathways. The aim of current study was to further identify the oxidative metabolites of SYL-927, locations of metabolites formation and the enzymes involved in the hydroxylation and phosphorylation. The excreta (urine, feces and bile) were collected for the metabolites characterization after a single oral administration of SYL-927 (3 mg/kg) in male SD rats. Metabolites formation of SYL-927 in different tissue homogenates, RLMs /HLMs and blood (erythrocytes and platelets) were also investigated with selected inhibitors and specific antibodies. A panel of recombinant human CYPs was used to further identify the enzymes responsible for the metabolism of SYL-927. Parent drug and

196

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

metabolites in biological samples were determined by LC-MS/MS analysis. Besides parent drug and its phosphate, two detectable metabolites-M1 (hydroxylate) and M2 (sulfate conjugate of M1) were found in blood, urine, feces and bile of rats after oral dosing. The formation of M1 in RLMs /HLMs was NADPH-dependent, with a Km and Vmax of 0.35 mM and 22.25 pmol/min/mg protein in RLMs and 207.3 mM and 149.9 pmol/min/mg protein in HLMs, respectively. Among the incubation of SYL-927 with heart, liver, spleen, lung, kidney, intestine, lymph nodes homogenates and blood, M1 was mostly generated in liver and intestine homogenates, less in other tissues, not in blood. In 15 recombinant human P450 enzymes, CYP1A1 and CYP2J2 were mostly responsible for the production of M1, less by CYP 1A2, 2C19, 2D6. In addition, M1 ´ -naphthaflavone for 1A1, formation could be inhibited by CYP2J2 and CYP1A1 antibody and selective inhibitors (100 mM jA 50 mM astemizole for 2J2), which further confirmed the involvement of two isozymes. The formation of SYL-927-P mostly occurred in rat and human blood. Both selected inhibitors of SPHK1 (DMS 10 mM) and SPHK2 (FTY720 10 mM) could significantly inhibit SYL-927-P production in rat and human erythrocytes/platelets. In conclusion, SYL-927 was metabolized into SYL-927-P, M1 and M2 in rats. Both SPHK1 and SPHK2 were primarily responsible for SYL-927-P formation in blood. CYP1A1 and CYP2J2 were the major enzymes involved in the formation M1.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P333. ASSESSMENT UNDER INITIAL RATE CONDITIONS OF THE SELECTIVITY AND TIME COURSE OF CYTOCHROME P450 INACTIVATION IN POOLED HUMAN LIVER MICROSOMES AND HEPATOCYTES: OPTIMIZATION OF INHIBITOR CONDITIONS USED FOR REACTION PHENOTYPING STUDIES Faraz Kazmi, Phyllis Yerino, Brian W. Ogilvie, Etsuko Usuki, Jennifer Chladek, and David B. Buckley XenoTech, LLC, Lenexa, KS, USA The identification of the metabolic pathways (reaction phenotyping) for new drug candidates is an essential component of drug development, as these studies increase the understanding of how a drug is cleared and the potential for the drug to be subject to drug–drug interactions with co-administrated medications. One approach used to identify which cytochrome P450 (CYP) enzymes are responsible for the metabolism of a drug is the use of specific chemical inhibitors in test systems such as pooled human liver microsomes (HLM) or pooled cryopreserved human hepatocytes (CHH). The design of these studies is critical because factors such as metabolic depletion of the inhibitor, protein binding, and insufficient enzyme inactivation can produce convoluted and misleading results. In the present study, we examined the selectivity and duration of CYP inactivation, with a range of protein concentrations and incubation times, for a variety of commonly used CYP inhibitors all under initial rate probe substrate conditions. Briefly, NADPH-fortified pooled HLM (n = 200) at 0.1, 0.5 and 1 mg/mL or pooled CHH (n = 100) at 1 million cells/mL, were pre-incubated for 30 min at 37  C with various CYP inhibitors, namely furafylline (10 mM), phencyclidine (up to 30 mM), gemfibrozil glucuronide (100 mM), tienilic acid (20 mM), esomeprazole (10 mM), quinidine (5 mM), paroxetine (1 or 5 mM), ketoconazole (1 or 4 mM), CYP3cide (2.5 mM), mibefradil (1 mM), and troleandomycin (50 mM). Following the preincubation step, marker substrate (&Km) incubations were performed for up to 5 min (10 min for CHH) at three post preincubation time points (0, 30, 120 min; simulating the time course of an unknown drug) to determine CYP1A2 (phenacetin), CYP2B6 (bupropion), CYP2C8 (amodiaquine), CYP2C9 (diclofenac), CYP2C19 (S-mephenytoin), CYP2D6 (dextromethorphan), CYP2E1 (chlorzoxazone) and CYP3A4/5 (midazolam) residual activities. Metabolite formation was determined by LC-MS/MS analysis. Results indicated that, in both HLM and CHH, inhibition of specific CYP enzymes can be accomplished over the duration of 120 minutes after pre-incubation with an inhibitor. Specifically, furafylline selectively inactivated CYP1A2 activity by 67–87%; phencyclidine inactivated CYP2B6 activity by 48–91% (with moderate CYP2D6 inhibition at higher concentrations); gemfibrozil glucuronide selectively inactivated CYP2C8 activity by 33–98%; tienilic acid selectively inactivated CYP2C9 activity by 97–100%; paroxetine and quinidine selectively inactivated CYP2D6 activity by 55–90% and 40–94% respectively; and mibefradil, CYP3cide, troleandomycin and ketoconazole inactivated CYP3A4/5 activity by 63–91%, 81–91%, 92–95%, and 78–97% respectively. Esomeprazole was found to potently inhibit CYP2C19 with some moderate inhibition of CYP2E1 and CYP3A4/5. In summary, these findings demonstrate the suitability and optimized conditions for the use of common chemical inhibitors in CYP reaction phenotyping studies, with implications on how these studies are designed.

P334. IN-VITRO PHENOTYPING OF TICLOPIDINE METABOLISM IN HUMAN LIVER MICROSOMES: EVIDENCE FOR THE ROLE OF FMO IN METABOLISM AND FORMATION OF OXIDATIVE METABOLITES Raghava Choudary Palacharla1, Venkatesham Uthukam1, Vijaya Bhargava Kanamarlapudi1, Arun Kumar Manoharan1, Ranjit kumar Ponnamaneni1, Ilayaraja Kalaikadiban2, Rajesh Kumar Boggavarapu2, and Ramakrishna Nirogi3 1 Department of Drug Metabolism and Pharmacokinetics, Suven Life Sciences Ltd, Hyderabad, India, 2Department of Bioanalytical Research, Suven Life Sciences Ltd, Hyderabad, India, and 3Department of Discovery Research, Suven Life Sciences Ltd, Hyderabad, India Ticlopidine is an anti-platelet drug approved in 1979 for prevention of thrombotic stroke with recommended twice daily dose of 250 mg1. Most of the orally administered dose is absorbed and peak systemic concentrations were reached within 2 h. The clearance of ticlopidine decreased significantly after repeated dosing exhibiting non-linear pharmacokinetics. Ticlopidine is extensively metabolized in liver and its metabolites were characterized and reported in the literature. The main metabolites

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

197

identified in humans include 2-oxoticlopidine, hydroxy ticlopidine, ticlopidine N-oxide, and thienopyridinium form. CYP2C192 and CYP2B63 were shown to contribute to the biotransformation of ticlopidine. However, the enzymes involved in the formation of oxidative and active metabolite were not completely elucidated. The objective of the current study is to identify CYP450 and non-CYP450 enzymes involved in the metabolism of ticlopidine in human liver microsomes and recombinant enzymes. Further, the enzymes involved in each of the biotransformation pathways of ticlopidine were characterized. Phenotyping assay for ticlopidine was conducted in human liver microsomes in the presence of specific chemical inhibitors of CYP450 enzymes and in recombinant enzymes. Alpha-Napthoflavone, tranylcypromine, PPP, montelukast, sulfaphenazole, N-3-benzyl nirvanol, quinidine, 4-methyl pyrazole, and azamulin were used as specific inhibitors of CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4 respectively. Substrate depletion was monitored in both test systems. The role of non-CYP enzymes was investigated by preheating the microsomal incubation for 2 min at 50  C. Qualitative metabolite profiling was performed in the absence and presence of chemical inhibitors, preheated microsomal incubations, recombinant CYP450 and FMO enzymes. Results from the chemical inhibition study indicate that CYP2B6 is the major CYP450 enzyme involved in the biotransformation. The decrease in the metabolism of ticlopidine in preheated liver microsomal sample indicated that FMO also plays a major role. This is further confirmed by measuring the ticlopidine depletion in recombinant human FMO3. The involvement of CYP450 enzymes and FMO in biotransformation pathways of ticlopidine is discussed.

P335. STUDIES ON THE METABOLIC ACTIVATION OF THE CRTH2/DP DUAL ANTAGONIST AMG 009

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Mark P. Grillo and Michelle Tadano Lohr Department of Pharmacokinetics & Drug Metabolism, Amgen, South San Francisco, CA, USA AMG 009 (2-(4-(butylcarbamoyl)-2-(2,4-dichlorophenylsulfonamido)phenoxy)-3-methoxyphenyl)acetic acid) is a dual antagonist of chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2) and prostanoid D receptor (DP) designed for the treatment of inflammation related to asthma and rhinitis. The development of AMG 009 was suspended after unanticipated increases in hepatic ALT/AST levels were observed in healthy volunteers. A mechanism for this observation was proposed where the metabolism of AMG 009 by cytochrome P450 and/or UGT enzymes resulted in the formation of chemicallyreactive intermediates that bind covalently to protein. In the present studies, we investigated the metabolic activation of AMG 009 in incubations with human liver microsomes (HLM) and with rat and human hepatocytes. Results showed that when radiolabeled [14C]AMG 009 (10 mM) was incubated with HLM (1 mg/mL, 37  C), time-dependent covalent binding to protein occurred in an NADPH-dependent fashion leading to 50 ± 16 pmol equivalents bound/mg protein/hr. SDS-PAGE analysis of the adducted-proteins revealed unidentified major protein-adducts at 31 and 45 kDa. Covalent binding to protein was abolished to background levels in incubations performed in the presence of glutathione (GSH, 10 mM). Analysis of HLM incubation extracts by LC-MS/MS and LC/radioactivity analysis showed the formation of 4 different GSH-adducts of AMG 009. The identity of the GSH-adducts was determined from tandem LC-MS/MS analysis and showed 3 of the GSH-adducts originated from metabolic activation of the 2,4-dichlorophenyl moiety, and one adduct generated from bioactivation on the n-butyl amide group. The major GSH-adduct originated from bioactivation of the n-butyl amide to a proposed reactive imine-amide intermediate. Incubations with rat and human hepatocytes (2  106 cells/mL, 2-h) resulted in covalent binding of [14C]AMG 009 (10 mM) occurring in a time-dependent fashion reaching 24 and 22 pmol/mg protein, respectively. Mechanistic experiments in rat hepatocytes showed that inhibition of AMG 009 acyl glucuronide formation had no effect on covalent binding, whereas inhibition of P450 led to a complete inhibition. In summary, results from the present studies showed AMG 009 to be metabolically activated by cytochrome P450 on the n-butyl amide- and dichlorobenzyl-moieties to imine-amide and epoxide-type reactive intermediates, respectively, which mediated the covalent binding of AMG 009 to protein in liver microsomes and in hepatocytes. The results from these metabolic activation studies with AMG 009 assisted subsequent discovery efforts to successfully identify potent backup CRTH2/ DP antagonists lacking cytochrome P450-mediated bioactivation.

P336. THE STUDY FOR THE IN VITRO/IN VIVO DISCREPANCY OF HUMAN-DISPROPORTIONATE METABOLITE OF S-777469 Kazutaka Sekiguchi1, Takushi Kanazu1, Kazuya Fukumura2, Hiroshi Hasegawa1, and Yoshitaka Yamaguchi1 1 Department of Drug Metabolism & Pharmacokinetics, Research Laboratory for Development, Shionogi & Co., Ltd, Osaka, Japan and 2Department of Clinical Pharmacology & Pharmacokinetics, Clinical Research Department, Shionogi & Co., Ltd, Osaka, Japan S-777469, which exhibits potent and selective agonism of Cannabinoid receptor 2, has been developed as an anti-pruritic and anti-atopic dermatitis agent. In human hepatocytes, the main metabolite was S-777469 hydroxymethyl (5-HM) and the metabolite profiles were qualitatively similar to those in rat and dog hepatocytes. Furthermore, the metabolite observed in rat and dog hepatocytes were also observed in these species of plasma, suggesting that S-777469 would not generate disproportionate metabolites in clinical (Sekiguchi et al., 2013). However, the mass balance study of [14C]-S-777469 revealed that 5carboxymethyl (5-CA) was major circulating metabolite and the AUCinf ratio of 5-carboxymethyl (5-CA), which was observed in human hepatocytes as a minor metabolite, was 24% to total radioactivity in humans. The elimination half-life values of 5-CA and S-777469 in human plasma were comparable with each other, suggesting that there was in vitro and in vivo discrepancy in the

198

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

formation of 5-CA. Thus, the objective of this study was to identify human enzymes involved in the formation of 5-CA for the investigation of this discrepancy. The in vitro studies using human liver microsomes and recombinant CYP isoforms revealed that the responsible enzyme for 5-HM formation was CYP2C9. Next, the formation of 5-CA was investigated in human liver fraction. In liver S9 fraction, 5-CA was mainly formed from 5-HM in the presence of NADPH. In addition, CYP2C9 was also predominant enzyme responsible for the 5-CA formation. Therefore, S-777469 was thought to be sequentially metabolized to 5CA solely by CYP2C9. Since CYP2C9 activity can be inhibited by some fatty acids existed in in vitro systems (Rowland et al., 2008), the kinetic parameters of the formation of 5-HM and 5-CA in the absence or presence of BSA, which binds fatty acids, were determined. The unbound intrinsic clearance of 5-HM and 5-CA were increased by 241% and 404% (from 4.24 ml/min/mg to 10.2 ml/min/mg for 5-HM, from 0.05 ml/min/mg to 0.21 ml/min/mg for 5-CA) in the presence of 1% BSA, respectively. This data suggested that CYP2C9-mediated oxidative metabolism of S-777469 might be suppressed in in vitro systems such as microsomes and hepatocytes. The influence of fatty acids on oxidative metabolism of S-777469 is also under investigation.

References

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Rowland A, Elliot DJ, Knights KM, et al. (2008). The ‘‘albumin effect’’ and in vitro-in vivo extrapolation: Sequestration of long-chain unsaturated fatty acids enhance phenytoin hydroxylation by human liver microsomal and recombinant cytochrome P450 2C9. Drug Metab Dispos 36:870–877. Sekiguchi K, Kanazu T, Takeuchi M, et al. (2013). Non-clinical evaluation of the metabolism, pharmacokinetics and excretion of S777469, a new cannabinoid receptor 2 selective agonist. Xenobiotica 44:48–58.

P337. CARBON–CARBON BOND CLEAVAGE AND INTRAMOLECULAR CYCLIZATION OF 2-CYANOETHYL CARBAMATE DERIVATIVES IN HUMAN LIVER MICROSOMES Jayaprakasam Bolleddula1, Steven Sullivan2, Lisa Feng3, Jennifer Jiang2, Paul R. Fatharee2, and Patrick J. Brassil1 1 Department of Drug Metabolism and Pharmacokinetics, Theravance Biopharma US, Inc., South San Francisco, CA, USA, 2 Department of Medicinal Chemistry, Theravance Biopharma US, Inc., South San Francisco, CA, USA, and 3Independent Consultant, South San Francisco, CA, USA Drug metabolizing enzymes such as the cytochromes P450 primarily catalyze oxidation, reduction and hydrolysis reactions of xenobiotics. Carbon-carbon bond formation and carbon-carbon bond cleavage reactions are rarely reported in the literature and are considered as an exception in xenobiotic metabolism. In one of our lead discovery programs, 2-cyanoethyl carbamate derivatives were designed as inhibitors of an inflammatory pathway. Incubation of chiral 2-cyanoethyl carbamate derivatives in human liver microsomes yielded diastereomeric 4-hydroxy-oxazolidinones, M1 and M2 as major metabolites. The mechanism of formation of M1 and M2 is proposed as a hydroxylation of the apha-carbon to the nitrile functional group to form unstable cyanohydrin which undergoes carbon–carbon bond cleavage to afford an aldehyde with concomitant elimination of hydrogen cyanide. The aldehyde intermediate reacts with the carbamate nitrogen via intramolecular nucleophilic attack to form 4-hydroxyoxazolidinones, M1 and M2. The chiral 2-cyanoethyl carbamate derivatives studied contain one stereocenter with ‘‘R’’ configuration and the formation of aldehyde and subsequent cyclization results in the formation of an additional stereocenter. The biotransformation thus yields two diastereomeric metabolites with RRand RS configurations. In vitro studies suggest that metabolites, M1 and M2 are formed disproportionately in human liver microsomes compared to rat and dog. Reaction phenotyping experiments with chemical inhibitors and recombinant enzymes revealed that CYP2C19 is the primary enzyme that catalyzes the formation of M1 and M2. It is thought that subsequent to CYP2C19 hydroxylation, the hydrogen cyanide elimination and cyclization steps are non-enzymatic reactions. The data suggest that oxidative elimination of the nitrile group and intramolecular cyclization may be a common metabolic pathway for cyanoethyl carbamate derivatives with a secondary carbamate nitrogen.

P338. METABOLISM OF VITAMIN K IN HUMAN LIVER MICROSOMES AND HEPG2 CELLS Amanda L. Johnson1, Katheryne Z. Edson2, Aaron M. Teitelbaum1, and Allan E. Rettie1 1 Department of Medicinal Chemistry, University of Washington, Seattle, WA, USA and 2Department of Pharmacokinetics and Drug Metabolism, Amgen, Thousand Oaks, CA, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

199

Vitamin K is an essential cofactor for physiological functions that include blood coagulation, bone metabolism/mineralization, and inhibition of vascular calcification. A specific form of vitamin K2 (menaquinone-4; MK4) is emerging as a key K vitamer for the extra-hepatic functions of this vitamin. While recycling of vitamin K1 (phylloquinone; PK) through its epoxide metabolite has been studied extensively, little is known about vitamin K catabolism, save that both PK and MK4 are ultimately excreted in urine and bile as side-chain shortened metabolites, K acid 1 and K acid 2. Catabolism of structurally similar vitamin E and ubiquinone suggest a common pathway to these chain-shortened acidic metabolites that involves an initial P450-mediated !-hydroxylation and subsequent b-oxidation. Our laboratory previously identified CYP4F2 as a major enzyme responsible for the initial !-hydroxylation of both PK and MK4. However, the P450-dependent metabolism of MK4 is likely much more complex than that of PK due to its unsaturated isoprenyl side-chain. Therefore, the goals of the present study were to define the catabolic pathways of MK4 in human liver, and the effect of the common CYP4F2*3 variant on MK4 metabolism in HepG2 cells. Metabolic comparisons were also made between MK4 and PK. Vitamin K metabolite analysis was performed by negative ion LC-MS/MS, typically in multiple reaction monitoring mode. Enzyme sources included pooled human liver microsomes, recombinant P450s and stably transfected HepG2 cells. Human liver microsomal incubations with MK4 yielded three major NADPH-dependent products, M1, M2 and M3. M1 was definitively identified as !-hydroxy MK4 using an authentic chemical standard. CYP102 formed mainly M2 and M3, suggesting that these metabolites were oxidized internally on the unsaturated side-chain. We also observed altered MK4 catabolism upon characterization of the uptake, depletion, and metabolite formation of K vitamers in HepG2 cells stably transfected with lentiviral constructs of wild-type CYP4F2, its common V433M variant, and an EGFP transfection control. Uptake of a given K vitamer was similar across all cell lines, although MK4 was taken up by HepG2 cells much more rapidly than PK. However, vitamin K depletion and metabolite formation differed considerably in cell lysates from the three cell lines. Vitamin K depletion was most rapid in the wild-typeCYP4F2 cell line, where we observed formation of the vitamin K !-hydroxy metabolite, as well as M2. In contrast, depletion of MK4 and PK was much slower and less extensive in both the EGFP and V433M cell lines, and formation of the !-hydroxylated products were negligible. Additional data will be presented for the analysis of downstream K acids in these cell lines, as well as further characterization of M2 and M3. These data demonstrate the complexity of hepatic MK4 catabolism relative to PK in liver-derived cell preparations and suggest the possibility that the common CYP4F2*3 variant has catabolic influence on physiological effects driven by MK4.

P339. ABSORPTION, METABOLISM, AND EXCRETION OF [14C]AXELOPRAN, A PERIPHERALLY-RESTRICTED NEUTRAL k-OPIOID RECEPTOR ANTAGONIST, AFTER ORAL AND INTRAVENOUS ADMINISTRATION TO HEALTHY HUMAN VOLUNTEERS Jayaprakasam Bolleddula1, Carla Washington2, Pierre-Jean Colson3, Aniruddha Amrite1, Ross Vickery2, Philip Worboys4, and David L. Bourdet1 1 Department of Drug Metabolism and Pharmacokinetics, Theravance Biopharma US, Inc., South San Francisco, CA, USA, 2 Department of Clinical Pharmacology and Experimental Medicine, Theravance Biopharma US, Inc., South San Francisco, CA, USA, 3Process R & D, Theravance Biopharma US, Inc., South San Francisco, CA, USA, and 4Theravance, Inc., South San Francisco, CA, USA Axelopran (formerly TD-1211) [3-{(1R,3R,5S)-8-(2-{cyclohexylmethyl[(2S)-2,3-dihydroxypropanoyl]amino}ethyl)-8-azabicyclo[3.2.1]octan-3-yl}benzamide] is a peripherally restricted, multivalent inhibitor of the m-opioid receptor designed with the goal of alleviating gastrointestinal side effects of opioid therapy without impairing analgesia. The absorption, metabolism and excretion of [14C]axelopran were investigated in ten healthy human volunteers following a single intravenous (1.6 mg; 80 mCi) and oral (15 mg; 300 mCi) administration of [14C]axelopran in a fixed sequence crossover design. Mean recovery of the administered dose was approximately 90% after IV and oral administration. After IV administration, a mean of 41% of the administered dose was recovered in urine and 49% recovered in feces. After oral administration, fecal excretion was the predominant elimination pathway with 77% of the dose recovered in feces and 13% recovered in urine. The mean absolute oral bioavailability of [14C]axelopran was 21%. Unchanged [14C]axelopran was the major drug-related component (73% of total radioactivity) in human plasma after IV administration. [14C]Axelopran was metabolized extensively after oral administration. M3 (N-dealkylation), M6 (hydroxylation), and M20 (hydrolysis and hydroxylation) were the major circulating metabolites in plasma, and each represented greater than 10% of total drug-related material after oral administration. Unchanged [14C]axelopran was the major drug-related component in urine after intravenous (32% of dose) or oral (6% of dose) administration. M14 was a major metabolite in feces accounting for approximately 10% and 27% of the administered dose after IV and oral administration, respectively. M14 is formed by hydrolysis of axelopran. High interindividual variation in the formation of M14 was observed. Overall, the major metabolic pathways of [14C]axelopran include hydrolysis (M14), hydroxylation (M6), N-dealkylation (M3), and a hydrolysis with hydroxylation (M20).

P340. A PHASE ONE OPEN-LABEL SINGLE-RADIOLABELED DOSE STUDY TO INVESTIGATE THE ABSORPTION, METABOLISM, AND EXCRETION OF [14C]PD-0332991 IN HEALTHY MALE VOLUNTEERS Cho-Ming Loi1, Alfin Vaz2, Justin Hoffman3, Melissa O’Gorman2, Leonid Kirkovsky3, Diane Wang3, and Deepak Dalvie3 1 Pfizer Research and Development, San Diego, CA, USA, 2Pfizer Inc, Groton, CT, USA, and 3Pfizer Inc, San Diego, CA, USA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

200

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

Background and objectives: Palbociclib (PD-0332991) is an orally active inhibitor of cyclin-dependent kinases (CDK)4 and CDK6. An open-label study was conducted to evaluate the mass balance and pharmacokinetics (PK) of palbociclib in 6 healthy male subjects after oral administration of a single 125 mg dose of palbociclib containing 100 mCi of [14C]palbociclib. The objectives of this study were to: (1) characterize the primary routes of elimination of palbociclib and drug-related material in humans, (2) characterize the PK of total radioactivity and compare with the PK of palbociclib and PF-05089326 (lactam metabolite), and (3) identify the metabolites of palbociclib in plasma, urine, and/or feces. Methods: After oral administration of [14C]palbociclib, serial blood samples were collected at specified times over 9 days. Urine and fecal sample collections were made before dosing (‘‘blank’’) and in 24 hour intervals or as passed from time of dosing until the end of the study. Through the morning of Day 9 and beyond, daily excreta sample collections continued until either (a) the amount of radioactivity recovered in excreta was at least 90% of administered radioactive dose, or (b) less than 1% had been recovered in excreta from 2 consecutive days. Total radioactivity in urine and feces were analyzed using liquid scintillation counting, and accelerator mass spectrometry was used to measure plasma total 14C content. Concentrations of palbociclib (in plasma and urine) and PF-05089326 (in plasma) were determined with validated methods using LC/MS/MS. Metabolite identification was performed using LC/MS. Results: Recovery of drug-derived radioactivity in feces and urine were 74.1% and 17.5% of dose, respectively, for a median total recovery of 91.6%. Estimate of total exposure (AUCinf) for plasma total radioactivity was 5-fold higher than that for plasma palbociclib, indicating the presence of circulating metabolic product(s). In plasma, palbociclib was the most abundant drugrelated entity, accounting for 23.3% of circulating radioactivity. The most abundant circulating metabolite was a glucuronide conjugate of palbociclib (M22), accounting for 14.8% of plasma radioactivity. Other metabolites, individually, constituted510% of plasma radioactivity. The ratios of metabolite to parent Cmax and AUCinf values of PF-05089326 (palbociclib lactam, M17) were 0.18 and 0.10, respectively. Two major metabolites, the sulfamic acid of palbociclib (M11) and a carboxylic acid (M16), were recovered in the feces, accounting for 25.8% and 14.2% of dose, respectively. Recovery of unchanged palbociclib in the feces was minimal (2.3% of dose). The primary drug-related entity recovered in the urine was unchanged palbociclib (6.9% of dose), with M22 accounting for only 1.5% of dose. Single oral 125 mg dose of palbociclib containing 100 mCi of [14C]palbociclib was safe and well-tolerated in the healthy adult males evaluated in this study. Conclusions: [14C]Palbociclib administered orally to 6 healthy male volunteers was extensively absorbed, metabolized, and excreted mainly in feces. Oxidation and sulfonation were the major primary metabolic pathways for [14C]palbociclib.

P341. METABOLITE PROFILING AND IDENTIFICATION IN CRYOPRESERVED HUMAN HEPATOCYTES USING THERMO SCIENTIFIC COMPOUND DISCOVERER SOFTWARE Jonathan N. Bauman1 and Caroline Ding2 1 Department of Pharmacokinetics, Dynamics and Metabolism, Pfizer Inc., Groton, CT, USA and 2Thermo Fisher Scientific, San Jose, CA, USA A consistent challenge of the drug discovery project team is to make rapid impactful decisions on limited data sets to move a program forward. Traditional methods of understanding metabolic clearance issues include screening half-lives in in vitro systems and trying to understand structure activity relationships, is rapid although imprecise. Alternately, determining metabolic soft-spots empirically can be time consuming with spectral interpretation being rate limiting. The combination of several different technologies: cryopreserved hepatocytes, high-resolution mass spectrometry and data analysis and interpretation software, can provide more rapid and accurate results to help teams make the right decisions. Cryopreserved hepatocytes provide in vivo-like enzyme expression levels, proper cell morphology, and increased ability to draw in vitro/in vivo correlations to make sound decisions regarding a compound’s fate. High-resolution mass spectrometry itself has improved the quality of metabolite identification, but also has enabled the advent of software that can take advantage of the precision of accurate mass instruments and combine that with computational power and techniques to further enhance the value of this data. This poster will share the results of a series of marketed drugs that were incubated with cryopreserved human hepatocytes, profiled by high resolution mass spectrometry and analyzed using Compound Discoverer software (Thermo Scientific).

P342. A QUANTITATIVE GLUTATHIONE TRAPPING ASSAY COUPLED WITH STRUCTURAL CHARACTERIZATION FOR EVALUATION OF REACTIVE METABOLITES Ichiro Kino, Keigo Kosaka, Kaori Murakoshi, Atsushi Iwamura, Takesada Shimura, Hidefumi Kaji, and Toshiyuki Kume DMPK Research Laboratories, Research Division, Mitsubishi Tanabe Pharma Corporation, Toda-shi, Japan Purpose: Glutathione (GSH) trapping assay is commonly used to evaluate reactive metabolites of drug candidates in pharmaceutical companies. The quantitative GSH trapping assay is useful for the prioritization of candidate compounds, whereas the qualitative GSH trapping assay using LC-MS/MS can provide valuable information on the reactive site. The aim of this study is to develop the evaluation system having both advantages of quantitative and qualitative assays. Method: We performed the covalent binding (CB) assay, [3H]GSH trapping assay, and structural analysis on compounds including drugs withdrawn from the market and informed with warnings attributed to severe hepatotoxicity or agranulocytosis in prescribing

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

201

information. We incubated 22 radiolabeled drugs with hepatocytes and quantified the CB by liquid scintillation counting. We incubated 40 drugs with NADPH-fortified human liver microsomes in the presence of GSH (mixture of natural GSH/ stable isotope [13C215N] / radioisotope [3H]). Then the incubation mixtures were analyzed by radio-LC-MS/MS, for obtaining the extent of adduct formation and the mass information. As necessary, MS/MS spectra were subsequently obtained for the structural analysis. Results and discussion: The amount of CB ranged from 1.1 to 160.4 pmol/h/mg, and the extent of adduct formation ranged from 0 to 2908 pmol/h/mg. Good correlation was observed between the amount of CB and the extent of adduct formation. By evaluating relationships between the extent of adduct formation and clinical toxicity, we confirmed that the drugs with higher doses and higher extent of adduct formation show a higher risk. By structural analysis, we could classify adducts into [3H]GSH (g-glutamylcysteinylglycine) adducts and [3H]cysteinylglycine (CysGly) adducts, that are thought to be generated metabolically from [3H]GSH adduct or directly from free [3H]CysGly. Vesnarinone and Rosiglitazone were detected only as [3H]CysGly adducts in this assay. After performing a structural analysis by LC-MS/MS, the adducted position was estimated to be around tertiary amines. We considered this technique, quantitative GSH trapping assay coupled with structural characterization, is useful for ranking the risky structure (the reactive part) as a quantitative manner in the screening of drug candidates.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P343. AN E. COLI-BASED BIOREACTOR FOR CYP3A4-DERIVED METABOLITES Anne E. Hagen1, Jonathan N. Bauman1, Loretta M. Cox1, Brenda Ellerbrock2, Carlos Martinez3, Richard P. Schneider1, Raman Sharma1, Sanjeewa G. Rupasinghe4, George T. Tkalcevic5, Alfin Vaz1, and Gregory S. Walker1 1 Department of Pharmacokinetics, Dynamics and Metabolism, Pfizer Worldwide Research and Development, Groton, CT, USA, 2 Akorn Pharmaceuticals, Decatur, IL, USA, 3Department of Pharmaceutical Sciences, Pfizer Worldwide Research and Development, Groton, CT, USA, 4Department of Pharmaceutical Sciences, BioResource International Inc., Durham, NC, USA and 5Diversified R&D – Diagnostics, Zoetis, Kalamazoo, MI, USA It has become an expectation in the pharmaceutical industry to have an earlier understanding of the role of metabolites in discovery. For a variety of reasons metabolites are often difficult to synthesize via classical medicinal chemistry methods. Alternatively, these metabolites can be produced in small amounts from a variety of biological incubations. Once these metabolites are produced in these biological matrices they can then be isolated to a pure form by liquid chromatography. Quantitative NMR (q-NMR) enables the concentration of these isolated metabolite solutions to be determined without the need of a gravimetrically prepared reference standard. Once isolated, characterized and quantitated these metabolites can be used as substrate in a variety of drug metabolism and pharmacology studies. The impact of these studies may ultimately result in drug candidates with improved metabolic stability, a more detailed understanding of metabolite pharmacological activity, and/or a metabolite standard that could be used as a bioanalytical standard. In order to have an impact in early discovery the structural characterization of a metabolite must be completed in a time frame that is quicker than the chemistry cycle time. Typically this time frame is two weeks from the initial observation of a metabolite in the screening process to the completion of the structural characterization. In vitro syntheses with microbes or biomimetics are options commonly used to generate metabolites. However, the use of these systems carries an additional burden of structural proof since in the absence of a chemically characterized authentic metabolite standard; these systems rely only on chromatographic and mass spectral similarities of the product generated to be equal to that of the metabolite. When the reaction phenotype of a metabolite is known, biosynthesis of the metabolite using recombinant enzymes to generate sufficient quantities of the metabolite for NMR characterization is also commonly employed. This approach is cost intensive in terms of biological reagents used. In order to access authentic metabolites of NCEs from CYP3A4-derived metabolism we have generated E. Coli expressing human CYP3A4 and Oxidoreductase as a cost effective ‘‘off the shelf reagent’’ for the production of metabolites. Herein we discuss some of the advantages and limitations of the use of such constructs. Several new chemical entities were selected with low, moderate, and high intrinsic clearance in human liver microsomes, and a known CYP3A4 component. Substrate concentrations ranging from 20–50 mM were incubated with E. coli suspended in M9 minimal media with 20% glucose in an orbital shaker at 37  C with high aeration for 3–4 hours. Metabolites were isolated using standard chromatographic techniques followed by 2-D structure analysis and quantitation by NMR. The moderate and high clearance compounds yielded qNMR standards in the 3–15 mM (40 mL) range, but generation of metabolites from compounds with low intrinsic clearance through 3A4 were not successful. Under ideal circumstances metabolites were biosynthesized, isolated, and characterized by NMR in three days.

P344. FLEXIBLE STRATEGIES FOR THE CONDUCT OF HUMAN METABOLISM STUDIES, WITH ONCOLOGY MOLECULES Iain Shaw, Lisa Clarke-Lens, Peter Scholes, and Lloyd Stevens Quotient Clinical, Nottingham, UK Clinical ADME studies are an important and necessary part of the drug development process. Performing ADME studies for oncology molecules presents unique challenges regarding subject population and manufacture and supply of 14C drug product for dosing to the subjects. Three alternative approaches are described to support these studies using an integrated approach to drug

Downloaded by [112.134.238.102] at 07:30 10 February 2016

202

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

product development, GMP manufacturing and clinical testing. For ADME studies in oncology molecules these options provide tailored solutions to gain the critical data needed. Requirements: Safety assessment and toxicological review to confirm suitability of operator handling processes and dosing to selected study population at desired dose level 14C drug product development 14C drug product manufacture Option 1 Study in healthy volunteers with clinical or sub-clinical 14C doses  Single dosing period  Standard sample collection/analysis  Efficient use of 14C API  Real time mass balance generation  Metabolite profiling/identification from pooled study samples Option 2 Study in healthy volunteers with 14C microdose  Single dosing period  Standard sample collection  AMS analysis  Efficient use of 14C API  Delayed but consolidated mass balance data generation  Metabolite profiling from pooled study samples  Metabolite identification from correlation with unlabelled samples from other studies Option 3 Study in patient population with clinical 14C dose  Intermittent dosing periods  Data in defined target patient population  ‘Per patient’ 14C drug product supply  Efficient use of 14C API  Real time but sporadic mass balance data generation  Metabolite profiling/identification for individual subjects Results: Case Study 1: Molecule X (molecular targeting drug) – Phase II development. Based on a full and successful safety and toxicological review the molecule was dosed in healthy volunteers at therapeutic dose. A simple 14C labelled suspension was used with 7 days shelf life. Product was manufactured and dosed within 2 days of manufacture – single site. Case Study 2: Elacytarabine, (Cytotoxic drug) – Mass balance data needed to support a market application. Normal approaches could not be taken due to the toxicity of the drug and logistical challenges of recruiting patients. A 14C labelled microdose of the drug was given to healthy volunteers to gain the data needed. Study synopsis was pre-approved with FDA. Case Study 3: Molecule Z, (Cytotoxic drug) – Required mass balance and metabolism data to be acquired in oncology patients using a 14C intravenous (IV) solution. Product was manufactured on a ‘per patient’ basis, QP released and shipped to a specialist European clinic for dosing based on a 7 day shelf life. In all three case studies 14C drug products were made ‘on demand’ under full GMP controls with a ‘fit for purpose’ CMC data package and shelf life. Conclusions: Conduct of the clinical ADME study in healthy volunteer studies offers significant time and cost savings whilst offering key expertise of accredited clinical site. Drug product can also be manufactured ‘on demand’ for oncology patients at specialist centres based on recruitment rates which reduces shelf life requirements and CMC investments in a ‘one-off’ drug product.

P345. ABSTRACT WITHDRAWN

P346. HEPATOCYTE-BASED MATHEMATICAL MODELING REVEALS A NOVEL METABOLIC PATHWAY OF BOSENTAN IN HUMAN LIVER Norikazu Matsunaga1, Naomi Kaneko2, Angelina Yukiko Staub1, Takeo Nakanishi1, Kenichi Nunoya2, Haruo Imawaka2, and Ikumi Tamai3 1 Department of Membrane Transport and Biopharmaceutics, Kanazawa University, Kanazawa, Japan, 2Pharmacokinetic Research Laboratories, Ono Pharmaceutical Co., Ltd., Tsukuba, Japan, and 3Faculty of Pharmaceutical Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan Hepatic disposition is governed by four factors: basolateral uptake, basolateral efflux, intracellular metabolism, and biliary excretion. The contribution of each factor varies depending on different types of diseases, genetic mutations, and/or concomitant drugs. Therefore, understanding of hepatic disposition of a compound helps to elucidate the inter-individual variability of

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

203

Downloaded by [112.134.238.102] at 07:30 10 February 2016

pharmacokinetics and to predict drug–drug interaction risks. Quantitative assessment of dynamics leads to more adequate understanding of a whole figure of drug disposition. Accordingly, quantitative mathematical modeling in hepatocytes is a powerful tool to clarify the fates of a parent drug and its metabolite(s) in the liver. In the present study, we analyzed the metabolic/hepatobiliary transport of bosentan to construct a mathematical model of hepatic disposition, including parent bosentan and its metabolites in human hepatocytes. It has been known that bosentan is metabolized by CYP3A4 and CYP2C9 to three metabolites: Ro 48-5033 (hydroxylation at t-butyl group of bosentan), Ro 47-8634 (O-demethylation of the phenolic methylester of bosentan), and Ro 64-1056 (combination of hydroxylation and O-demethylation of bosentan). In the conventional model of bosentan and its three metabolites, however, intracellular Ro 64-1056 was overestimated compared with the observed values, while others are well explained. Then, we made the following three hypotheses; (i) Ro 64-1056 is further metabolized, (ii) metabolic rate from Ro 48-5033 to Ro 64-1056 is overestimated because an unidentified metabolic pathway is involved, and (iii) metabolic rate from Ro 47-8634 to Ro 64-1056 is likewise overestimated for the same reason. To evaluate the above hypotheses, we performed in vitro metabolism study. As a result, we found a novel metabolic pathway of Ro 47-8634 in the human liver and succeeded in determining the metabolite’s structure. Furthermore, we re-established a mathematical model incorporating the novel metabolic pathway, which greatly improved the predicted values of Ro 64-1056 in the cell compartment without significantly affecting any others. Our findings demonstrated that mathematical modeling of in vitro hepatic disposition based on hepatocytes is useful in finding novel metabolic pathway as well as for more quantitative prediction and translation to in vivo disposition.

P347. THE NET LOSS OF A CARBON ATOM FROM A TETRAHYDRO-TRIAZOLPYRAZINE HETEROCYCLE IS THE MAIN METABOLIC CLEARANCE PATHWAY OF A PARP INHIBITOR IN VITRO AND IN RAT IN VIVO Ru¨diger Kaspera1, Chungang Gu1, Jeffrey Johannes2, James Atherton3, Minli Zhang3, and Marie-Eve Beaudoin1 1 iMed Oncology DMPK, AstraZeneca Pharmaceuticals, Waltham, MA, USA, 2iMed Oncology Chemistry, AstraZeneca Pharmaceuticals, Waltham, MA, USA, and 3iMed Infection, AstraZeneca Pharmaceuticals, Waltham, MA, USA Nitrogen heterocycles represent an important class of compounds in cancer drug development. Designing candidates of metabolic stability (St. Jean & Fotch, 2012) or diversifying metabolism with regard to drug interactions has been of particular interest (Isoherranen et al., 2012). In preclinical investigations, a tetrahydrotriazolpyrazine heterocycle was developed as PARP-inhibitor for selective targeting of tumor cells with centrosome amplification during mitosis. This inhibitor is a low clearance compound and seems to be metabolically cleared through a predominant route in vitro and in vivo. During in vitro metabolism studies, this tetrahydrotriazolpyrazine heterocycle (a 3-(1,1-difluoroethyl)-6-methyl-5,6,7,8-tetrahydro-[1,2,4]triazolo[4,3-a]pyrazine-7-carbonyl-derivative) underwent oxidation to yield a predominant metabolite by the net loss of a carbon from the tetrahydrotriazolpyrazine heterocycle (the loss of 12 dalton as determined with accurate mass). This predominant metabolism was observed across species in mouse, rat, dog and human hepatocytes, with rodents exhibiting a higher rate of metabolite formation relative to human and dog. The compound revealed a low in vivo clearance of 20% liver blood flow in rat (17 mL min1 kg1) with minimal biliary and renal clearance of unchanged parent drug, indicating metabolism as main pathway to clear the compound. In addition, this particular ‘‘net carbon-loss’’-metabolite was found to be the main drug related material excreted in rat bile and urine. Mechanistically, the net carbon-loss was hypothesized to proceed through ringoxidation and subsequent decarboxylation. Additional studies using different chromatographies and identification of intermediates, positive and negative mode fragmentation and deuterium exchange experiments supported this mechanism. Interestingly, the loss of carbon was dependent on the substituent of the electron deficient bicyclic 6-carbonyl tetrahydrotriazolpyrazine. Replacement of the attached difluoroethyl residue with an electron donating alkyl moiety such as an isopropyl shifted metabolism away from decarboxylation and the subsequent net carbon loss. In summary, modulating the electron density in this tetrahydrotriazolpyrazine heterocycle could be a potential design route to circumvent and diversify this single route metabolism and enable stability of this moiety.

References St. Jean D, Fotch C. (2012). Mitigating heterocycle metabolism in drug discovery. J Med Chem 55:6002–602. Isoherranen N, Lutz JD, Chuang SP, et al. (2012). Importance of multi-P450 inhibition in drug–drug Interactions: Evaluation of incidence, inhibition magnitude, and prediction from in virto data. Chem Res Toxicol 25:2285–2300.

204

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

P348. IN VITRO METABOLISM OF HOMOEGONOL IN HUMAN LIVER MICROSOMES AND CHARACTERIZATION OF CYTOCHROME P450 AND UDP-GLUCURONOSYLTRANSFERASE ENZYMES

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Soon Sang Kwon, Hyeon-Uk Jeong, Ju Hyun Kim, and Hye Suk Lee Drug Metabolism and Bioanalysis Laboratory, College of Pharmacy, The Catholic University of Korea, Bucheon, South Korea Homoegonol is a biologically-active neolignan isolated from Styrax species with cytotoxic, antimicrobial, anti-inflammatory, and anti-asthma activities. Homoegonolshowed the high systemic clearance (133 mL/min/kg), a short half-life (0.49–1.33 h), low urinary recovery (0.005% of the dose), and low oral bioavailability (6.0%) in rats, indicating the extensive metabolism of homoegonol. However, there was no report on the metabolism of homoegonol. We aimed to identify the metabolites ofhomoegonol formed from in vitro incubation of homoegonol with human liver microsomes by using liquid chromatographymass spectrometry (LC-MS), and to characterize the cytochrome P450 (CYP) and UDP-glucuronosyltransferase (UGT) enzymes responsible for homoegonol metabolism in human liver microsomes. Incubation of homoegonol with human liver microsomes in the presence of NADPH and UDP-glucuronic acid resulted in the formation of 4 metabolites, namely, 3-Odesmethylhomoegonol, hydroxyhomoegonol, homoegonol glucuronide, and 3-O-desmethylhomoegonol glucuronide. The specific CYP isozymes responsible for the formation of 3-O-desmethylhomoegonol and hydroxyhomoegonol from homoegonol were identified using a combination of correlation analysis in different human livermicrosomes and studies on the immune-inhibition properties using CYP1A2 and CYP3A4 antibodies and metabolic characteristics in human cDNA-expressed CYP enzymes. CYP1A2 and CYP3A4 were identified as the major enzymes responsible for the formation of 3-O-desmethylhomoegonol. CYP2J2 and CYP3A4 played the predominant roles in the formation of hydroxyhomoegonol. Using cDNA-expressed human UGT isozymes, the formation of homoegonol glucuronide might be mediated by UGT2B7, UGT1A4, UGT1A3, and UGT1A1 enzymes. UGT1A8, UGT1A1, and UGT1A10 enzymes were responsible the formation of 3-O-desmethylhomoegonol glucuronide from 3-O-desmethylhomoegonol. In conclusion, homoegonol was metabolized by O-demethylation to 3-O-desmethylhomoegonol, hydroxylation to hydroxyhomoegonol, and glucuronidation to homoegonol glucuronide and 3-O-desmethylhomoegonol glucuronide. The pharmacokinetics ofhomoegonol may be affected by CYP1A2, CYP2J2, CYP3A4, UGT1A1, UGT1A3, UGT1A4, UGT1A8, UGT1A10, and UGT2B7 isozymes responsible for the metabolism of homoegonol.

P349. AZOLES AS NOVEL INHIBITORS OF MONOAMINE OXIDASE (MAO) A AND B ISOZYMES Abdul Naveed Shaik1, Mahamad Yunnus Mahat2, Karthikeyan Kandasamy3, Barbara W. LeDuc1, and Ansar Ali Khan4 1 Department of Pharmaceutical Sciences, MCPHS University, Boston, MA, USA, 2Jawaharlal Nehru Technological University, Hyderabad, Hyderabad, India, 3Manipal College of Pharmaceutical Sciences, Manipal, Karnataka, India, and 4Drug Metabolism and Pharmacokinetics,Vimta Labs Ltd, Hyderbad, India Introduction: Azoles are a class of compounds that contain a five-membered aromatic ring with two heteroatoms. The use of two azoles viz 1-amino benzotriazole (1-ABT) and ketoconazole as standard probe inhibitors to decipher the CYP and/or nonCYP mediated metabolism of xenobiotics is well appreciated by researchers and regulators. In the present investigation, for the first time, we have characterized azoles as novel, mixed function inhibitors of monoamine oxidases (MAO) A and B. Methods: 1-ABT, ketoconazole, astemizole, fluconazole, and sulphaphenazole were selected and screened as potential MAO inhibitors using kynuramine as a standard, non-specific MAO probe substrate in mouse, rat and human liver S9 fractions. Inhibition of enzyme activity was assessed by measuring the formation of metabolite (4-OH-quinoline) using LC/MS/MS. Further, 1-ABT and ketoconazole were screened against human recombinant MAO-A and B isozymes. The in vivo consequence of MAO inhibition was determined by evaluating the oral and intravenous pharmacokinetics of sumatriptan, a known MAO substrate, using ketoconazole, in Swiss albino mice. Results: Ketoconazole and 1-Amino benzotriazole strongly inhibited MAO in mouse, rat, human liver S9 fractions with IC50 values 0.7, 7.9, 16.2 and 0.07, 0.12, 6.05 mM respectively. Astemizole moderately inhibited MOA, whereas fluconazole and sulphaphenazole did not inhibit MAO in mouse, rat and human liver S9 fractions. Additional studies with expressed human MAO A and B isozymes revealed that ketoconazole predominantly inhibited MAO A in comparison to MAO B. In agreement with the in vitro findings, oral ketoconazole (50 mg/ kg, 1 h prior to sumatriptan) in mice significantly decreased the systemic plasma clearance of intravenous sumatriptan (44%) and increased the AUCLast by approximately 45%. Further, sumatriptan oral Cmax and AUCLast in presence of ketoconazole increased by 47 and 39%, respectively. Conclusion(s): In summary, the results described herein clearly demonstrate that 1ABT and ketoconazole are novel and mixed function inhibitors of MAO A and B across the species tested. Further, ketoconazole increased the oral and intravenous sumatriptan plasma exposures by decreasing the systemic plasma clearance, thereby demonstrating the translation of in vitro MAO inhibition to in vivo. These findings should be considered when using 1-ABT/ketoconazole as tool compounds to evaluate CYP/non-CYP mediated metabolism, particularly when MAO-dependent metabolism is also involved.

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

205

P350. IN VITRO METABOLISM OF MAM-2201 IN HUMAN LIVER MICROSOMES AND HEPATOCYTES AND CHARACTERIZATION OF CYTOCHROME P450

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Eun Nam Kim1, Hyeon-Uk Jeong1, Ju Hyun Kim1, Hee Seung Kim2, Jin Young Kim2, Moon Kyo In2, and Hye Suk Lee1 1 Drug Metabolism and Bioanalysis Laboratory, College of Pharmacy, The Catholic University of Korea, Bucheon, South Korea and 2Forensic Chemistry Laboratory, Forensic Science Division, Supreme Prosecutor’s Office, Seoul, South Korea MAM-2201, (1-(5-fluoropentyl)-1H-indol-3-yl)(4-methyl-1-naphthalenyl)-methanone, is a new synthetic cannabinoid. Nothing has been published on the pharmacology, side effects, potential overdose cases, deaths of MAM-2201, though it presumably has similar properties to the closely related AM-2201 and JWH-122, which are both full agonists and unselectively bind to CB1 and CB2 cannabinoid receptors with low nanomolar affinity. There was no report on the pharmacokinetics and metabolism of MAM2201. We aimed to identify the metabolites of MAM-2201 formed from in vitro incubation of MAM-2201 with human liver microsomes andhepatocytes by using liquid chromatography-high resolution mass spectrometry, and to characterize the cytochrome P450 (CYP) enzymes responsible for MAM-2201 metabolism. Incubation of MAM-2201 with human liver microsomes in the presence of NADPH resulted in the formation of ten metabolites including 4-hydroxypentyl-MAM-2201 (M2, identified by authentic standard), four hydroxy-MAM-2201 (M1, M3, M4, M5), three dihydroxy-MAM-2201 (M6, M7, M8), 5hydroxypentyl-MAM-2201 (M10), and hydroxyl-5-hydroxypentyl-MAM-2201 (hydroxy-M10, M9). Using cDNA-expressed human CYP isozymes, the formation of 4-hydroxypentyl-MAM-2201 (M2) was mediated CYP2B6, CYP2C8, and CYP2J2. A major metabolite, hydroxy-MAM-2201 (M3) was formed by CYP1A2, CYP3A5, CYP2C9, CYP2C19, CYP2D6, and CYP2J2. The formation of dihydroxy-MAM-2201 (M6), a major metabolite, was mediated by CYP2C19 and CYP3A5. The formation of 5-hydroxypentyl-MAM-2201 (M10) might be mediated by CYP2C8 with minor contribution of CYP2B6 and CYP2J2. The formation of hydroxy-M10 (M9) might be mediated by CYP2C8, CYP2C9, CYP2C19, CYP1A2, and CYP3A5. CYP3A5 was involved but CYP3A4 was not involved in the metabolism of MAM-2201. After incubation of MAM-2201 in human hepatocytes for 2 h, MAM-2201 and five metabolites such as hydroxy-MAM-2201 (M3 and M4), 5-hydroxypentyl-MAM-2201 (M10), hydroxy-M10 (M9), and M10 glucuronide. In conclusion, MAM-2201 was metabolized by hydroxylation to five hydroxy-MAM2201 (M1-M5), three dihydroxy-MAM-2201 (M6-M8), and hydroxy-M10 (M9), and defluorination to 5-hydroxypentyl-MAM2201 (M10) in human liver microsomes. CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2J2, and CYP3A5 were responsible for the metabolism of MAM-2201 to 10 metabolites.

P351. ABSTRACT WITHDRAWN

P352. ABSTRACT WITHDRAWN

P353. PREDICTION OF HUMAN-SPECIFIC METABOLITES OF COMPOUND X USING CHIMERIC MICE WITH HUMANIZED LIVER Masahiro Yahata1, Takao Watanabe1, Masaaki Tagawa2, and Masashi Yabuki1 1 Preclinical Research Laboratories, Sumitomo Dainippon Pharma Co., Ltd., Osaka, Japan and 2Medical Affairs, Sumitomo Dainippon Pharma Co., Ltd., Tokyo, Japan Characterization of drug metabolites in human plasma is crucial for understanding their contributions to drug efficacy and toxicity. The purpose of this study was to assess whether chimeric mice with humanized liver adequately predict the formation of human-specific metabolites of compound X. After a single oral administration of compound X to healthy volunteers, plasma concentrations of compound X and its metabolites were determined using LC-MS/MS. Compound X was primarily metabolized via hydrolysis of the amide bond to M1, and the ratio between the mean area under the curve of M1 to that of the unchanged drug (AUCM1/AUCUD) was 324% in human plasma. Furthermore, M1 was metabolized to oxidized metabolites (M2, M3, M4, and M5) and the AUCmet/AUCUD ratio for each was 1.5%, 6.6%, 5.8% and 6.0%, respectively. These oxidized metabolites of M1 were not detected in rats, dogs, or monkeys, indicating that they were human-specific. Next, a single oral dose of 14C-labeled compound X (20 mg/kg) was administered to PXB mice with humanized liver (PhoenixBio, Co., Ltd, Hiroshima, Japan) and control mice (SCID mice). Metabolic profiling of plasma samples collected at 1 and 6 h after administration was conducted using Radio-LC-MS/MS. At 1 h after administration, compound X and M1 were detected as major components in plasma of PXB mice, accounting for 25% and 42% of the total peak area determined by the radiochromatogram, respectively. The ratio of the plasma concentration of M1 to the unchanged drug (CpM1/CpUD) was 173%. M4 and M5 were also detected in plasma of PXB mice, accounting for 0.7% and 1.4% of the total peak areas, respectively (CpM4/CpUD and CpM5/CpUD were 2.9% and 5.7%, respectively). Although they were not found to be radioactive, trace levels of M2 and M3 were detected via LC-MS/MS in the plasma of PXB mice. At 6 h after administration, M4 was detected in the plasma of PXB mice, accounting for 0.5% of the total

206

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

peak area (CpM4/CpUD was 36%), M5 was detected at trace levels via LC-MS/MS, and M2 and M3 were not detected at all, even by LC-MS/MS. In contrast, no human-specific metabolites (M2, M3, M4, or M5) were detected in the plasma of SCID mice, even though M1 was detected and accounted for 20% of the total peak area. The rank order of plasma exposure of humanspecific metabolites in PXB mice (M4 and M544M2 and M3) was different from that in humans (M3, M4, and M54M2). Therefore, PXB mice could not fully predict plasma exposure of human-specific metabolites in humans. These results suggested that PXB mice could be useful as an in vivo model for qualitative prediction of human-specific metabolites in human plasma.

P354. KINETIC DEUTERIUM ISOTOPE EFFECT ON METABOLISM AND PHARMACOKINETIC PROPERTIES OF AN ALDEHYDE OXIDASE SUBSTRATE

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Anthony J. Lee, Michael Serby, Vincent Peterkin, and Jill Wetter Drug Metabolism and Pharmacokinetics, AbbVie, North Chicago, IL, USA Kinetic deuterium isotope effect has been utilized to alter DMPK properties of drug molecules by decreasing rate of metabolism through replacement of C–H bonds with C–D bonds at metabolic soft spots. There have been efforts to deuterate marketed drugs to improve pharmacokinetic properties. Aldehyde Oxidase (AO) substrates have been targeted for deuterium substitution since the proposed mechanism involves a C–H bond breaking as a rate determining step. A001 (H), which contains pyrrolotriazolopyrazine moiety, was metabolically stable in rat and human liver microsomes but underwent oxidative metabolism and subsequent glucuronidation in hepatocytes. The structure of oxidative metabolite M1 was identified and confirmed to be a dihydropyrrolopyrazinone by LC-MS and proton NMR spectroscopy. Oxidation to M1 was mediated by AO, confirmed by incubating [3H]A001 in liver cytosol in the presence or absence of AO inhibitor raloxifene. After an hour of incubation in human and rat cytosol, M1 was detected with 21 and 22% of total radioactivity, respectively, while M1 formation was nearly completely inhibited in the presence of raloxifene. Following a single intravenous administration of [3H]A001 to bile duct cannulated rats at 5 mg/kg, 72.2% of dosed radioactivity was recovered as M1 and its glucuronide in bile and urine while 16.2% was as parent drug, indicating AO mediated oxidation as a major elimination pathway. Deuterated analog, A001-d4 (D), was prepared with nearly 100 atom % D at the carbon of M1 oxidation. To determine kinetic isotope effect in vitro, H and D were incubated in hepatocytes for four hours and depletion of compounds over time was assayed by LC-MS. In human and rat hepatocytes, H to D ratio of scaled intrinsic clearance was 3.2 and 3.5, respectively, showing 43-fold decrease in the rate of in vitro metabolism. Following intravenous administration of 1:1 mixture of H and D at 1 mg/kg to Sprague Dawley rats, plasma clearance values were 1.7 and 1.0 L/h/kg, respectively, and half-lives were 0.7 and 1.3 hours, respectively. AUC values after 3 mg/kg oral administration of 1:1 mixture were 0.52 and 0.80 mg  hr/mL for H and D, respectively, showing 1.5–1.9 fold changes in pharmacokinetic parameters. In summary, replacement of hydrogen with deuterium at the site of metabolism of A001, an AO substrate, decreased the rate of metabolism in vitro and improved pharmacokinetic parameters in rat in vivo.

P355. ABSTRACT WITHDRAWN

P356. IN VITRO METABOLISM OF DOXAPRAM AND ITS ACTIVE METABOLITE KETODOXAPRAM BY CYP3A4/5 Hitoshi Ohno1, Noboru Kamada1, Takuro Endo1, Makoto Murakami1, and Mamoru Kobayashi2 1 Pharmacokinetics Research, Kissei Pharmaceutical Co., Ltd., Matsumoto-City, Japan and 2Development Research, Kissei Pharmaceutical Co., Ltd., Azumino-City, Japan Purpose: Doxapram (DOX) has been used as a respiratory stimulant in the case of postanesthesia, drug-induced CNS depression, and chronic pulmonary disease associated with acute hypercapnia. Previous report has shown that administered DOX via I.V. was rapidly eliminated from blood and that the excreted fraction of DOX in urine was small in humans (1). Small amount of unchanged DOX was excreted into urine and bile in rats (2). DOX was metabolized to ketodoxapram (Keto-DOX), an active metabolite of DOX, which was eliminated from blood as rapidly as DOX (1). While DOX is suggested to be eliminated mainly through metabolism according to these reports, little is known about the metabolism of DOX, which means that the enzymes remain to be identified responsible for the metabolism of DOX and Keto-DOX and even for the formation of Keto-DOX. In this study, we examined in vitro metabolism of DOX as well as Keto-DOX to identify these enzymes and qualitatively explored metabolites of DOX. Method: DOX and Keto-DOX were incubated with human hepatic subcellular fraction in the presence or absence of enzyme cofactors (NAD and NADPH). In inhibition assay, selective inhibitors of each cytochrome P450s (CYPs) isoform were also co-incubated in human liver microsomes. Recombinant CYPs were used in the presence of NADPH. Concentrations of DOX and Keto-DOX were analyzed by LC-MS/MS. The exploration of DOX metabolites was performed by LC-QTOF. Results & Discussion: Both DOX and Keto-DOX were rapidly metabolized in human liver microsomes in an NADPH-dependent manner, but not metabolized at all in human liver cytosol regardless of enzyme cofactors. Ketoconazole, a selective inhibitor of CYP3A, inhibited not only the clearance of DOX and Keto-DOX but also Keto-DOX formation from DOX in human liver microsomes. Consistent with these results, recombinant CYP3A4 and CYP3A5 exhibited high activities for both DOX and Keto-DOX metabolism and also for the formation of Keto-DOX from DOX. Qualitative exploration of DOX

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

207

metabolites demonstrated that major metabolites of DOX formed in recombinant CYP3A4 and CYP3A5 were roughly in accordance with metabolites formed in human liver microsomes. These metabolites included the metabolites reported to be detected in dog urine and human fetal liver. Conclusion: In summary, these findings indicate that CYP3A4 and CYP3A5 are dominantly responsible for the rapid clearance of DOX as well as Keto- DOX and for the formation of Keto-DOX in vitro.

References Robson RH, Prescott LF. (1979). A pharmacokinetic study of doxapram in patients and volunteers. Br J Clin Pharmacol 7:81–87. Takeda T. (1973). Studies on metabolism of doxapram hydrochloride in rats. Shinshu Igaku Zasshi 21:549–560.

P357. EXTRAPOLATION OF HEPATIC CLEARANCE FOR THE DRUGS METABOLIZED BY MONOAMINE OXIDASE FROM IN VITRO DATA

Downloaded by [112.134.238.102] at 07:30 10 February 2016

Kazuki Hayashi, Yusuke Masuo, Noritaka Nakamichi, and Yukio Kato Faculty of Pharmacy, Kanazawa University, Kanazawa, Japan In vitro–in vivo extrapolation (IVIVE) approach has been widely used to predict clearance in vivo from in vitro data in the research and development of therapeutic agents. Information on IVIVE approach for the drugs metabolized by non-CYP enzymes is relatively limited, compared with those metabolized by CYP. Monoamine oxidase (MAO) is not only involved in metabolism of physiological substrates including noradrenaline and serotonin, but also a certain drug molecules. MAO is localized in mitochondrial outer membranes and expressed in liver and other organs. This may hinder the usual IVIVE approach using hepatocytes and liver microsomes. In the present study, we attempted to establish IVIVE approach to predict hepatic clearance in vivo for MAO-metabolized drugs from in vitro data. Both tyramine and phenylephrine were selected as typical substrates metabolized by MAO. Both compounds were incubated with human cryopreserved hepatocytes or liver microsomes and disappearance of both compounds during the incubation mixture were quantified by LC-MS/MS. Unbound fraction of these compounds in human serum was also determined using equilibrium dialysis method. The concentration in the medium of both of the compounds linearly decreased during the incubation with human cryopreserved hepatocytes. In vitro hepatic intrinsic clearance was calculated from such disappearance curves. On the other hand, unbound oral clearance in vivo of both compounds in human was determined from literature data (Pt´a˘cek et al., 2007; VanDenBerg et al., 2003) and measured unbound fraction. Considering the physiological scaling factor, in vitro-in vivo difference was approximately 5.3- and 19-fold for tyramine and phenylephrine, respectively. Such tendency was compatible with the previous report for another MAO substrate, zolmitriptan (Di et al., 2012), and one of the possible reasons for the underestimation is the contribution of extrahepatic MAO. Disappearance of phenylephrine was also observed in the incubation mixture containing human liver microsomes, and this disappearance was completely inhibited by MAO inhibitor, pargyline. Considering the physiological scaling factor, in vitro–in vivo difference based on liver microsomes data was approximately 28-fold for phenylephrine, and underestimation of in vivo clearance was still observed. Further studies are now performed to more quantitatively estimate the in vivo–in vitro difference. Hepatic clearance for MAO substrates may be quantitatively estimated from in vitro experimental data. Appropriate scaling factor could be needed to quantitatively estimate hepatic clearance for the MAO substrates from in vitro data. The validity of the utilization of hepatocytes and liver microsomes in the IVIVE approach for MAO substrates should be further evaluated.

References Di L, Trapa P, Obach RS, et al. (2012). A novel relay method for determining low-clearance values. Drug Metab Dispos 40:1860–1865. Pt´a˘cek P, Klı´ma J, Macek J. (2007). Development and validation of a liquid chromatography–tandem mass spectrometry method for the determination of phenylephrine in human plasma and its application to a pharmacokinetic study. J Chromatograph B 858:263–268. VanDenBerg CM, Blob LF, Kemper EM, et al. (2003). Tyramine pharmacokinetics and reduced bioavailability with food. J Clin Pharmacol 43:604–609.

P358. NON-CYP METABOLISM OF COMPOUND A AND STRATEGY OF METABOLITE EVALUATION Kazuhiro Yoneda1, Kenichi Nunoya2, Tsuyoshi Yokoi3, and Haruo Imawaka4 1 Pharmacokinetic Research Laboratories, Ono Pharmaceutical Co., Ltd., Tsukuba, Japan, 2Ono Pharmaceutical Co., Ltd., Tsukuba, Japan, 3Drug Safety Sciences, Nagoya University Graduate School of Medicine, Naogya, Japan, and 4Pharmacokinetic Research Laboratories, Ono Pharmaceutical Co., Ltd., Ibaraki, Japan In drug development, it is important to evaluate the efficacy and safety of metabolites, and drug interaction. However, metabolite evaluation may disturb the drug development when human metabolites are unlikely to be forecasted from non-clinical studies. In the present study, we report the difficulties in prediction of metabolites in human based on the non-clinical study data that we experienced and future strategy of metabolite evaluation. Compound A, which is we are developing, showed no differences in microsomal metabolism between animals and humans. Therefore, its development was moved to clinical stage. Then, a metabolite analysis in Phase I study revealed that 42 metabolites, including the ones that could not be detected in non-clinical studies, were present in human plasma. To discuss their safety, the exposure levels of those metabolites were compared between humans and animals although authentic samples were unavailable. As a result, the exposure of 8 metabolites was higher in

208

Poster Abstracts

Drug Metab Rev, 2015; 47(S1): 41–297

humans than in animals. Then, we immediately conducted a human mass balance study using carbon-14 labeled Compound A and analyzed all of the human metabolites to clarify the proportions of respective exposure levels to the total exposure. The quantification result showed that there were no metabolites whose exposure exceeded 10% of the total exposure, the criterion for safety assessment. Further, we elucidated the metabolite structures by an accurate mass spectrometry analysis and found out the contribution of Cytochrome P450s, esterases, UDP-glucuronosyltransferases, and reductases to Compound A metabolism. We also found that Compound A was mainly metabolized by reductases, and the metabolites were further metabolized by multiple enzymes. The reductases were estimated to be aldo-keto reductase and carbonyl reductase by an in vitro metabolism study using inhibitors; however, detection of the multiple step metabolism was difficult in microsomes and other in vitro samples. This was considered owing to the species differences in reductase activity and no evaluations in cytosols or hepatocytes that contain reductases in the pre-clinical period. Since non-CYP metabolism (e.g. aldehyde oxidase) is considered important recently, we evaluate species differences in metabolism using hepatocytes that contain some kind of non-CYP metabolic enzymes as our current strategy for metabolite evaluation. In addition, exposure of metabolite is one of the key factors to discuss the efficacy and safety of metabolites, and drug interaction. Our strategy includes a clarification of metabolite exposure in human by semiquantification of Phase I study samples while postponing the full validation and human mass balance study to a later date. This would allow us to confirm the efficacy of developing drugs at an early stage.

Downloaded by [112.134.238.102] at 07:30 10 February 2016

P359. SPECIES DIFFERENCES IN PYRIDOXAL OXIDATION ACTIVITY Daisuke Kobayashi, Atsushi Johno, Yasuka Ohmi, Kyousuke Misawa, Miki Morita, MIka Ishikawa, Teruki Yoshimura, and Keiji Wada Department of Food and Chemical Toxicology, School of Pharmaceutical Sciences, Health Sciences University of Hokkaido, Ishikari-Tobetsu, Hokkaido, Japan Purpose: Vitamin B6 is an essential nutrient for metabolism of amino acids and neurotransmitters, and is clinically used to treat and prevent vitamin B6deficiency. The main metabolite of vitamin B6 is pyridoxic acid which is produced by oxidation of pyridoxal, and is excreted in urine. It was reported that enzymes responsible for pyridoxal oxidation were NAD-dependent enzyme, aldehyde dehydrogenase, and NAD-independent enzyme, aldehyde oxidase which is involved in metabolism of drug such as, famcyclovir. However, species differences in pyridoxal oxidation activity and rate of contribution of each enzyme to the oxidation in liver have not been reported. In the present study, species differences in pyridoxal oxidation activities were investigated by using hepatic cytosol of human, mouse, rat and guinea pig. Method: Human liver cytosol was purchased from BD Biosciences. Hepatic cytosols of ddY mouse, Wistar/ST rat and Hartley guinea pig were prepared according to a standard protocol. Formed pyridoxic acid was measured by HPLC with fluorescence detector after incubation of pyridoxal with cytosol in the presence or absence of cofactors and aldehyde oxidase inhibitors. Results and discussion: A marked difference of pyridoxal oxidation activity was observed among species examined. The pyridoxal oxidation activity in human and mouse liver cytosol were much higher than that in rat liver cytosol. Characterization of enzymes involved in pyridoxal oxidation in human and mouse liver cytosol was done. Incubation studies in the presence or absence of NAD revealed that NAD-dependent enzyme was not involved in pyridoxal oxidation in both species. On the other hand, aldehyde oxidase inhibitors such as menadione and raloxifene inhibited pyridoxal oxidation activities in human and mouse, suggesting that aldehyde oxidase was principal enzyme for pyridoxal oxidation in both species. However, inhibition effects of norharman, which was an isoform specific inhibitor of aldehyde oxidase, were different between in human cytosol and mouse cytosol. It was therefore presumed that kinds and natures of aldehyde oxidase isoforms responsible for pyridoxal oxidation, or rate of contribution of each isoform were different between human and mouse.

P360. IN VITRO SCREENING OF NON-CYP PHASE I METABOLISM: ASSESSMENT OF ALDO-KETO REDUCTASES Phil Butler, Katie Plant, Joanne Shaw, and Clive Dilworth Cyprotex, Macclesfield, UK Aldo-keto reductases (AKRs) are a cytosolic superfamily of enzymes which primarily catalyse aldehyde or ketone redox transformations. The preferred AKR-mediated transformation is reduction and detoxification of aldehydes and ketones to primary and secondary alcohols respectively, although AKR-mediated oxidation has also been reported. Many of these reactions are a NADPH-dependent reduction although a small number of AKRs are capable of binding NADH and show dual NADH/ NADPH specificity. AKRs catalyse a broad range of substrates including xenobiotics, carcinogens, prostaglandins and steroids. The absence of AKRs in microsomes limits the application of typical in vitro stability screens for carbonyl-containing xenobiotics, whilst the broad substrate selectivity of AKRs may lead to difficulties in predicting metabolism from the structure of the compound. As such identification of the AKR-mediated metabolism may be challenging. Methods exist which assess carbonyl-reducing activity by monitoring oxidation of NADPH at 340 nm but are not sensitive or specific to AKRs. This work focuses on the investigation of AKR-mediated metabolism utilising recombinant AKRs (AKR1A1, 1B1, 1B10, 1C1, 1C2, 1C3, 1C4 and 1D1) and human liver cytosol to assess substrate specificity and inhibitor selectivity. This may allow development of an isoform-specific phenotyping screen suitable for use in human liver cytosol. Reported AKR substrates acetohexamide, progesterone, dolasetron and naloxone were incubated in recombinant AKRs or human liver cytosol with a range of reportedly

DOI: 10.3109/03602532.2015.1071933

Poster Abstracts

209

selective inhibitors or a cocktail of inhibitors. Intrinsic clearance was determined from monitoring substrate depletion by LCMS/MS over a 120 min time-course in human liver cytosol and recombinant enzymes, plus and minus inhibitors supplemented with NADPH and initiated with the addition of substrate. In many cases, reportedly specific substrates were metabolised by more than one isoform of AKR in the expressed enzymes. There was also an apparent lack of inhibitor selectivity at the single inhibitor concentration tested. For example, indomethacin (a reportedly specific AKR1C3 inhibitor) was a potent inhibitor of dolasetron metabolism in AKR1B10. However, identification and confirmation of AKR substrate activity in the expressed enzymes proved a useful tool for confirming AKR involvement in the metabolism of substrates in human liver cytosol. Additionally, using these techniques to identify inhibitor selectivity will prove useful for investigating the contribution of AKR metabolism to in vitro hepatic clearance in more complex systems such hepatocytes.

P361. INHIBITION OF ALDEHYDE OXIDASE ACTIVITY BY NATURAL PRODUCT CONSTITUENTS: ASSESSING CLINICAL RELEVANCE AND IDENTIFYING KEY ENZYME–LIGAND INTERACTIONS

Downloaded by [112.134.238.102] at 07:30 10 February 2016

John T. Barr1, Jeffrey P. Jones2, Nicholas H. Oberlies3, and Mary F. Paine1 1 Department of Experimental and Systems Pharmacology, Washington State University, Spokane, WA, USA, 2Department of Chemistry, Washington State University, Pullman, WA, USA, and 3Department of Chemistry and Biochemistry, The University of North Carolina at Greensboro, Greensboro, NC, USA Predicting the inhibitory effect of a xenobiotic on drug metabolizing enzyme activity based solely on chemical structure is a fundamental tool in drug discovery. Aldehyde oxidase (AO) is emerging as an important drug metabolizing enzyme, yet the relationship between chemical structure and inhibitory activity has not been evaluated systematically. We hypothesized that an in silico molecular model for human AO, coupled with ligand docking, could be used to link chemical structure with inhibition potency. This hypothesis was tested with a small library of natural product constituents (n = 24), a convenient source of closely related yet chemically diverse structures. The aims of this work were to (1) screen the library for AO inhibitors; (2) determine inhibition potency (Ki) of selected constituents; (3) identify constituents as potential perpetrators of clinically relevant interactions with conventional medications; and (4) relate structural features to Ki and identify key ligand binding interactions with the enzyme. The AO-catalyzed oxidation of O6-benzylguanine, an anti-cancer agent in Phase I clinical trials, was used as the index reaction. Incubation mixtures consisted of O6-benzylguanine (125 mM), human liver cytosol (0.1 mg/mL), constituent (10, 100 mM), and potassium phosphate buffer (25 mM, pH 7.4); after 4 min, reactions were quenched with acetonitrile (3 volumes) and analyzed for 8-oxo-benzylguanine by LC-MS/MS. Based on the screening data, 17 constituents were selected for Ki determination using a 6  6 matrix of substrate and inhibitor concentrations. Reversible inhibition models were fit with the data via nonlinear least-squares regression. Three constituents were described best by a competitive model; 14 were described best by a mixed-mode model, which consists of two Ki terms: Kis, the affinity between inhibitor and free enzyme, and Kii, the affinity between inhibitor and enzyme-substrate complex. Kis and Kii ranged from 0.26–73 and 0.80–120 mM, respectively. EGC and EGCG, prominent constituents in green tea, were the most potent, with Kiss of approximately 0.3 mM. The static model comparing systemic inhibitor concentration to Kis ([I]/Kis) suggested moderate risk (0.3–1.3) of an interaction involving ECG and EGCG as perpetrators. A model for human AO was developed based on the crystal structure of AOX3, a highly homologous murine AO with 62% identity. Constituents were docked into the enzyme active site; docking scores correlated with Kis (r2 = 0.5). Key ligand binding interactions were identified within the active site, including molybdenum within the cofactor and residues Phe885, Phe 923, and Glu 882. These results provide a framework for developing in silico models to predict the AO inhibitory activity of a xenobiotic based solely on chemical structure.

P362. MECHANISM-BASED INHIBITION OF HUMAN CARBOXYLESTERASE 2 BY IRINOTECAN Masaki Mikamoto1, Katsuyuki Fukuda1, and Kazutomi Kusano2 1 Department of Drug Metabolism and Pharmacokinetics Japan, Eisai Co., Ltd., Tsukuba, Ibaraki, Japan and 2Department of Global Drug Metabolism and Pharmacokinetics, Eisai Co., Ltd., Tsukuba, Japan Irinotecan (CPT-11) is a water-soluble prodrug of SN-38, which is widely used for treatment of several cancers. Carboxylesterase 2 (CES2) is a key enzyme that catalyzes irinotecan hydrolysis to produce the active metabolite, SN-38, and the counterpart product, 4-piperidinopiperidine (4PP), in humans. Due to its poor bioconversion efficiency, plasma concentration of SN-38 is much lower than that of irinotecan after intravenous administration to humans.