The significance of Trk receptors in pancreatic cancer - SAGE Journals

8 downloads 61 Views 1MB Size Report
Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer ... range of 20%–25%.2 Pancreatic cancer remains a clinical challenge with ...
692256 research-article2017

TUB0010.1177/1010428317692256Tumor BiologyJohnson et al.

Original Article

The significance of Trk receptors in pancreatic cancer

Tumor Biology February 2017: 1­–13  © The Author(s) 2017 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1010428317692256 https://doi.org/10.1177/1010428317692256 journals.sagepub.com/home/tub

Matthew D Johnson1, Brandon Stone1, Bryan J Thibodeau2, Andrew M Baschnagel3, Sandra Galoforo1, Laura E Fortier2, Billie Ketelsen2, Samreen Ahmed2, Zakiya Kelley2, Alaa Hana1, Thomas G Wilson1, John M Robertson1, Robert P Jury4 and George D Wilson1

Abstract This study investigated the Trk receptor family as a therapeutic target in pancreatic ductal adenocarcinoma and assessed their prognostic significance. Global gene expression analysis was investigated in prospectively collected pancreatic ductal adenocarcinomas that had either undergone neoadjuvant chemoradiation or were treated by surgery. PANC1 and MIA-PaCa-2 cell lines were investigated to establish whether fractionated radiation altered expression of four neuroendocrine genes and whether this resulted in subsequent changes in radiosensitivity. A specific inhibitor of TrkA, B, and C, AstraZeneca 1332, was investigated in vitro and in vivo in combination with radiation. A tissue microarray was constructed from 77 pancreatic ductal adenocarcinoma patients who had undergone neoadjuvant chemoradiation and the Trk receptor, and neurogenic differentiation 1 expression was assessed and correlated with overall survival. A total of 99 genes were identified that were differentially expressed in the chemoradiation patients with neuroendocrine genes and pathways, in particular the neurogenic differentiation 1 and Trk receptor family, being prominent. Fractionated radiation upregulated the expression of neuroendocrine genes, and AstraZeneca 1332 treatment in vitro enhanced radiosensitivity. No added effect of AstraZeneca 1332 was observed in vivo. Trk receptor expression varied between isoforms but did not correlate significantly with clinical outcome. Radiation treatment upregulated neuroendocrine gene expression but the Trk receptor family does not appear to be a promising treatment target. Keywords Neuroendocrine gene expression, Trk receptors, radiation, AstraZeneca 1332 Date received: 8 July 2016; accepted: 16 August 2016

Introduction Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer mortality in the United States, and more than 46,000 patients are diagnosed annually.1 At the time of presentation, only 15% to 20% of patients are eligible for surgical resection which remains the only potentially curative treatment. Even for early stage disease, overall survival following resection at 5 years is in the range of 20%–25%.2 Pancreatic cancer remains a clinical challenge with considerable disagreement and lack of consensus on many aspects of treatment guidelines due to lack of sufficient evidence.3

1Department

of Radiation Oncology, William Beaumont Hospital, Beaumont Health System, Royal Oak, MI, USA 2Beaumont BioBank, William Beaumont Hospital, Beaumont Health System, Royal Oak, MI, USA 3Department of Human Oncology, University of Wisconsin–Madison, Madison, WI, USA 4Department of General Surgery, William Beaumont Hospital, Beaumont Health System, Royal Oak, MI, USA Corresponding author: George D Wilson, Department of Radiation Oncology, William Beaumont Hospital, Beaumont Health System, 3601 W. Thirteen Mile Road, Royal Oak, MI 48073, USA. Email: [email protected]

Creative Commons Non Commercial CC-BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 3.0 License (http://www.creativecommons.org/licenses/by-nc/3.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).

2 Patients presenting with borderline resectable disease, defined by limited extension into the peripancreatic volume, are more likely to involve positive surgical margins and recurrence if they undergo immediate resection. In this cohort of patients, neoadjuvant therapy plays a crucial role in the management to convert them to surgical candidates.4 However, despite the use of neoadjuvant therapy, many patients have inadequate disease response to proceed with surgery. One of the largest meta-analyses on neoadjuvant therapy with data from 4394 patients reported that 33% of patients with initially unresectable tumors went onto resection after neoadjuvant chemoradiation (CRT) treatment with gemcitabine (Gem) or 5-fluorouracil (5-FU) combined with standard radiation (45–50.4 Gy in 1.8 Gy fractions).5 Clearly, there is a need to improve upon current CRT regimens. Recent data from the metastatic setting have shown that multi-drug regimens such as FOLFIRINOX and Gem with nab-paclitaxel improve survival compared to monotherapy.6,7 However, treatment response rates for these regimens are modest at 31.6% and 23%, respectively, and these multi-drug regimens are also associated with increased toxicity and recommended only for patients with good performance status.6,7 These results underscore the need to find targeted therapies to increase response rates while minimizing toxicity, and this has been under intense investigation.8–12 A number of biological agents modulating different targets are currently in clinical development including targeting epidermal growth factor receptor and other core pathways deregulated in PDAC such as PI3K/mammalian target of rapamycin (mTOR), inhibiting angiogenesis, cell cycle, matrix metalloproteinases, DNA repair, cytokines, cyclooxygenase-2, proteasome, and immune checkpoint inhibitors, as well as other strategies targeting the stroma and vaccine-based therapeutics.9,10 However, despite this spate of investigations, very few phase-III trials have been conducted, and only one agent, erlotinib, has been approved for use by the Federal Drug Administration (FDA) in PDAC in combination with Gem in unresectable or metastatic cancer.13 In this study, we elected to utilize a different approach to identify possible targets for treatment. Most current targets are based on genomic and other biological characteristics identified from pre-treatment specimens of PDAC. Several critical signaling pathways have been identified including DNA repair, apoptosis, G1/S cell cycle transition, KRAS, Wnt, Notch and Hedgehog (Hh) signaling, transforming growth factor β (TGF-β), chromatin remodeling, and other cell invasion pathways14 which have become targets of precision medicine approaches in PDAC.11 However, another approach is to identify biological targets post-neoadjuvant therapies that exploit pathways which are differentially activated in the residual resistant cancer. In this study, using global gene expression analysis, we have identified a novel expression pattern of

Tumor Biology genes including chromogranin A (CHGA), chromogranin B (CHGB), secretogranin II (SCG2), secretogranin III (SCG3), and secretogranin V (SCG5) that were upregulated in the treated PDAC patients when compared to nontreated PDAC patients. In addition, we found neurogenic differentiation 1 (NEUROD1) to be a highly regulated expression target sub-network. NEUROD1 is a basic helix-loop-helix transcription factor important in pancreas development15 and has been found to be expressed in a subset of aggressive neuroendocrine tumors.16 The tropomyosin receptor kinase family, TrkA, TrkB, and TrkC, are tyrosine kinases that are downstream of NEUROD1, and Trk receptor function has been shown to mediate survival and differentiation of neurons during development and throughout adulthood.17 Aberrant expressions of these receptors have also been implicated in several neuronal and non-neuronal malignancies including neuroblastoma, melanoma, prostatic, head and neck, as well as pancreatic cancers.18 Earlier work, looking at 47 resected pancreatic adenocarcinoma specimens, revealed increased Trk receptor expression of 66% when compared with normal adjacent tissue.19 Another study showed that TrkB expression in resected pancreatic samples correlated with increased rates of perineural invasion, positive margin, and development of metastatic disease.20 Furthermore, established pancreatic cell lines have been shown to have increased growth and invasion when incubated with increasing amounts of the Trk receptor ligands.21,22 Based on our new findings and the existing literature, we have investigated the utility of AstraZeneca 1332 (AZD1332) which has a potent and selective effect on TrkA, TrkB, and TrkC in combination with radiation in pancreatic cancer cell lines and xenografts. In addition, we have examined whether Trk receptor status has prognostic significance in PDAC.

Materials and methods Patients and treatments Two cohorts of patients were studied. For the initial gene expression studies, patients were selected from a prospectively collected series of 270 pancreatic cancer patients consented, collected, and stored in the Beaumont Health BioBank. From this cohort, six patients were identified who received neoadjuvant CRT and who went onto surgical resection but had a poor pathological response to treatment. These patients were treated on a clinical trial using neoadjuvant CRT consisting of Gem given weekly for 7 weeks (1000 mg/m2) with erlotinib daily for 8 weeks (100 mg). Radiotherapy consisted of either 30 Gy or 36 Gy given over a 3-week period for a total of 15 fractions, to the gross tumor plus a 1-cm margin starting with the first dose of Gem.23 One patient only received neoadjuvant GemRT without erlotinib. These patients were age- and stage-matched with

3

Johnson et al. Table 1.  Patient and treatment characteristics. ID

Sex

Age (years)

TNM stage

Chemotherapy

Radiotherapy

Time to resection (months)

AX0640 AX1380 AX1260 AX0721 AX0907 AX1587 AX0001 AX1017 AX1249 AX1606 AX1649 AX1898 AX2351 AX0739 AX0808 AX0235

Female Male Female Female Female Female Female Male Male Female Female Male Female Female Male Female

54 60 42 51 70 76 55 65 78 65 55 82 70 51 69 53

T2N0M0 T3N0M0 T3N1M0 T3N0M0 T3N1M0 T3N0M0 T3N0M0 T3N1M0 T3N0M0 T3N0M0 T3N1M0 T3N1M0 T3N1M0 T3N1M0 T3N0M0 T3N1M0

Gemcitabine Gemcitabine and erlotinib Gemcitabine and erlotinib Gemcitabine and erlotinib Gemcitabine and erlotinib Gemcitabine and erlotinib None None None None None None None None None None

36 Gy in 15 Fx 30 Gy in 15 Fx 30 Gy in 15 Fx 30 Gy in 15 Fx 30 Gy in 15 Fx 36 Gy in 15 Fx None None None None None None None None None None

1.3 1.3 1.7 1.5 1.2 6.2                    

10 patients who underwent resection without neoadjuvant treatment. Table 1 lists the key clinical features of these patients. Fresh frozen material stored in the BioBank was retrieved for gene expression analysis. The second cohort of patients were retrospectively obtained from the pathology archive and consisted of 77 patients from a consecutive series of patients treated with resection followed by either full-dose gemcitabine-based chemoradiation (GemRT) or 5-FU-based chemoradiation (5FURT) for resectable pancreatic cancer between 1998 and 2012.24 For the 5FURT group, chemotherapy was typically started concomitantly with RT on day 1 and constituted a continuous infusion of 5-FU (200–300 mg/m2) administered Monday through Friday, or bolus 5-FU (500 mg/m2) on days 1–3 and days 29–31. For the GemRT group, concurrent gemcitabine was typically, a 30-min IV infusion at a dose of 1000 mg/m2 given on days 1, 8, and 15 of a 28-day cycle, with RT started on day 1.

Gene expression analysis Frozen pancreatic adenocarcinoma tissues from patient cohort 1 specimens stored at −80°C in RNAlater Stabilization Solution (Life Technologies, Carlsbad, CA) were homogenized with a gentleMACS Dissociator using gentleMACS M Tubes (Miltenyi Biotec Inc., Auburn, CA) prior to total RNA isolation using an Omega Total RNA Isolation Kit (Omega Bio-Tek 400, Norcross, GA). Following the manufacturer’s protocol, total RNA was purified using spin cartridge technology, quantified (NanoDrop 8000, Thermo Scientific, Waltham, MA), and stored at −80°C. RNA integrity was determined by Bioanalyzer analysis (Agilent, Santa Clara, CA) just prior to processing for expression microarray analysis. Using the Ambion® WT Expression Kit (Affymetrix, Santa

Clara, CA), total RNA was used to generate purified sense cDNA ready for fragmentation and labeling. Fragmentation and terminal labeling was done according to manufacturer’s protocols using the GeneChip® WT Terminal Labeling and Controls Kit. Hybridization to Affymetrix Human Exon 1.0 ST Arrays, washing and staining (with GeneChip Fluidics Station 450, Affymetrix), and scanning (with GeneChip Scanner 3000) were performed following manufacturer protocols. The CEL files containing the raw intensity data from the Affymetrix GeneChip arrays were imported into Partek® Genomics Suite™ (6.6 version 6.13.1224) and normalized using the robust multichip average with a guanine-cytosine content background correction, quantile normalization, log2-transformation, and median polish probeset summarization. Exons were then summarized to genes using the average of the probesets. Differentially expressed genes were identified using analysis of variance (ANOVA) with two factors: prognosis and scan date (random variable). Hierarchical clustering was carried out using Partek software (Partek Inc., St. Louis, MO). Hierarchical clustering analysis was performed using Euclidean distance as similarity measure and average linkage for the agglomerative method. Sub-network enrichment analysis (SNEA) and pathway analysis was performed using Pathway Studio 10.5.0.5 (Ariadne Genomics, Rockville, MD). The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE60646.

Cell lines and drugs The MIA PaCa-2 and PANC-1 cell lines were purchased from ATCC (Manassas, VA). Both are pancreas epithelial cell carcinoma lines. Cells were cultured and maintained

4 in Dulbecco’s Modified Eagle’s Medium (Gibco-Life Technologies, Grand Island, NY) supplemented with 10% fetal bovine serum. The selective TrkA/B/C inhibitor, AZD1332, was kindly provided by AstraZeneca (Boston, MA). For in vitro studies, a 10-mM solution was prepared in dimethyl sulfoxide (DMSO) and stored at −70°C. For in vivo experiments, AZD1332 was prepared daily as a suspension in 0.5% methylcellulose and administered by oral gavage.

Irradiation Cells were irradiated with a Xstrahl X-ray System, Model RS225 (Xstrahl, United Kingdom) at a dose rate of 0.29 Gy/min, tube voltage of 160 kVp, current of 4 mA, and filtration with 0.5 mM Al and 0.5 mM Cu. Cells in 25 cm2 flasks were irradiated (0.5–4 Gy) at 37°C. Animals were irradiated with a Faxitron Cabinet X-ray System, Model 43855F (Faxitron X-Ray, Wheeling, IL) at a dose rate of 0.69 Gy/min, tube voltage of 160 kVp and current of 4 mA.

Effect of fractionated irradiation of on posttreatment radiosensitivity To investigate whether a course of fractionated irradiation induced subsequent radiation resistance on surviving pancreatic cancer cells, MIA-PaCa-2 and PANC-1 cells were plated into T75 flasks at 2.5 × 105 cells/flask. The cells were then irradiated with 2 Gy/day, Monday to Friday, for 3 weeks for a total of 30 Gy. Cells were not passaged during radiation and media was exchanged two times per week. At the end of treatment, cell samples were then taken on a weekly basis to study gene expression of key neuroendocrine-related genes studies (described below). To assess whether the radiation treatment had any effect on the radiosensitivity of the cells, cells were plated into flasks at 5, 14, and 26 days after radiation treatment and treated with radiation (0–6 Gy) and processed for clonogenic survival. The control for these studies was cells from the same culture that the treated cells were started from. The controls were also irradiated (0–6 Gy) and processed for clonogenic survival.

Nucleic acid extraction and quantitative realtime polymerase chain reaction (RT-PCR) RNA was extracted from cells using Allprep DNA/RNA Mini Kit (Qiagen, Valencia, CA). SuperScript VILO cDNA Synthesis Kit (Invitrogen, Carlsbard, CA) was used for the synthesis of cDNA from RNA. The following predesigned TaqMan gene specific primers (Life Technologies) were used: CHGA (Assay ID: Hs00900375_m1), CHGB (Assay ID: Hs01084631_m1), GAPDH (Assay ID: Hs99999905_m1), SCG3 (Assay ID: Hs00203076_m1), and SST (Assay ID: Hs00356144_m1). Quantitative

Tumor Biology RT-PCR reaction mixture was prepared containing 2 uL cDNA (10 ng), 1x TaqMan Gene Expression Master Mix (Life Technologies), and 1x Gene Expression Assay (Life Technologies). Gene expression levels were quantified using the ViiA 7 Real-Time PCR system (Life Technologies). The following thermocycling condition was used: 50°C for 2 min, 95°C for 10 min, and 40 amplification cycles of 95°C for 15 s/60°C for 1 min. Relative gene expression levels were calculated by the 2−ΔΔCT method, where delta delta CT value (ΔΔCT) = (CTtarget − CThousekeeping) − (CTcontrol − CThousekeeping). All expression values were normalized to the housekeeping gene GAPDH. If gene expression was not detectable after 40 amplification cycles, the sample was excluded for further statistical analysis. Error bars are the relative quantification minimum (RQmin) and maximum (RQmax) which represent the standard error of the mean (SEM) expression levels.

The combination of radiation and AZD1332: MTT assay The MTT (3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) assay was used to assess effect of AZD1332, with or without radiation, on cell growth. Briefly, cells were plated at 1000 cells/well in normal growth media. On day 6 of culture, AZD1332, in varying concentrations, was added to the cells. After 24 h, MTT (5 mg/mL phosphate-buffered saline (PBS)) was added to each well and the plate returned to the CO2 incubator for 4 to 5 h. Media containing the MTT was then aspirated from the wells, and DMSO was added to the wells to dissolve the purple formazan that formed. Following 5-min incubation at 37°C, absorbance readings (at 560 and 670 nm) were taken on a Versamax multiplate reader (Molecular Devices, Sunnyvale, CA). The data were expressed as percent change relative to the control. Each experiment was repeated three or four times.

The combination of radiation and AZD1332: clonogenic survival assay Cells were treated with 1 to 5 µM of AZD1332 for 1 h, irradiated (0–4 Gy), and then plated into flasks containing growth media and the corresponding concentration of AZD1332. Colonies were allowed to develop for 10– 14 days, stained with crystal violet, counted, and surviving fractions calculated. Data were normalized for plating efficiency, and survival curves were fitted using the linearquadratic equation.

Xenograft growth delay assay All animal experiments were approved by the Institute Animal Care and Use Committee. Xenografts were established in 8- to 10-week-old female NU(NCr)Foxn1nu athymic nude mice (Charles Rivers Laboratories,

5

Johnson et al. Wilmington, MA) by injecting MIA-PaCa-2 cells subcutaneously into the flank, at a density of 5 × 106 cells per 100 µL of Matrigel (Corning, Bedford, MA). Tumor volume was measured twice weekly by digital caliper and calculated using the formula (πab2) / 6 (a = largest diameter, b = smallest diameter). When tumors reached volumes of 200–300 mm3, animals were randomly assigned to experimental groups. The endpoint of the experiment was when tumor volumes were >2000 mm3 or 100 days post end of treatment. The MIA-PaCa-2 xenograft experiment consisted of four treatment groups (six mice/group): (1) control with sham oral gavage (0.1 cc, 0.5% methylcellulose, five times/week) for 3 weeks; (2) AZD1332 (20 mg/kg) in 0.1 cc 0.5% methylcellulose by oral gavage (five times/ week) for 3 weeks; (3) sham oral gavage followed 1 h later by radiation delivered as 2.0 Gy/day (five times/week) for 3 weeks; and (4) AZD1332 (20 mg/kg) in 0.5% methylcellulose by oral gavage followed 1 h later by radiation, 2.0 Gy/day (five times/week) for 3 weeks.

Tissue microarray (TMA) A TMA consisting of formalin-fixed paraffin-embedded cores from the 77 patients in cohort 2 was constructed using a semi-automated tissue microarrayer (Pathology Device, Westminster, MD). A pathologist marked representative areas of tumor tissue on hematoxylin and eosin stained sections from each tissue block. Two 1.0-mm tumor tissue cylinders were punched from each marked tumor areas of each tissue block and inserted onto the TMA.

Immunohistochemistry Immunohistochemical staining was performed by a horseradish-peroxidase (HRP) technique using either a DAKO Autostainer (DAKO, Hamburg, Germany) or Ventana Discovery XT (Ventana Inc., Tucson, AZ) according to the manufacturer’s recommendation. Paraffin-embedded tissue sections cut from the TMA were mounted on slides and treated as follows: (1) TrkA (Millipore, Billerica, MA) (clone MABN681)—1:200 dilution for 30 min, HRP flex secondary, and no antigen retrieval; (2) TrkB Millipore (clone AB9872)—1:600 dilution for 120 min, HRP flex secondary, and no antigen retrieval; (3) TrkC (Abcam, Cambridge, MA) (clone ab33656)—1:200 dilution for 120 min and anti-goat secondary with no antigen retrieval; (4) Pan-Trk (Abcam) (clone ab79770)—1:50 dilution for 15 min, Flex HRP secondary, and no antigen retrieval; and (5) NEUROD1 (Abcam) (clone ab60704)—monoclonal mouse, 1:800 dilution for 60 min, HRP flex secondary, and high pH (7) antigen retrieval. TrkA, B, and C and Pan-Trk antibodies were also double blocked with Avidin-Biotin and Background Sniper (Biocare Medical, Concord, CA) to reduce background. All antibodies were

then visualized with 3,3′-diaminobenzidine (DAB) chromogen (Vector Laboratories, Burlingame, CA), and hematoxylin was used for counterstaining.

Quantitative image analysis TMA slides were scanned and digitized using the Aperio ScanScope (Aperio Technologies Inc., Buffalo Grove, IL). Analysis was performed using Definiens Tissue Studio® software (Definiens Inc., Carlsbad, CA). The TMA cores were de-arrayed using a Grid template in the software and each core was detected and matched. In the Analysis Builder, a tissue–background separation was performed and each core analyzed using a customized algorithm. The algorithm was trained to detect regions of interest (ROIs) which were defined as viable tumor, non-tumor white space, and artifacts (e.g. histological stain deposits or tissue folds). Within each viable tumor ROI, individual nuclei were detected based on hematoxylin staining, and cells were simulated around the nucleus on the basis of the presence of a membrane. DAB staining thresholds were established to classify each cell based on both membrane and cytoplasmic intensity as negative, low, medium, or high. The solution was trained on the pan-Trk-stained TMA and applied to all the Trk subunits and NEUROD1stained cores with adjustments of thresholds based on the stain intensity for each subunit. Total stain intensity was recorded for areas of viable tumor.

Statistical analysis Pearson’s chi-square test was used to compare categorical variables. Student’s t-test was used to compare continuous data with a normal distribution. The Kaplan–Meier method and log rank test were used to calculate and compare actuarial rates of overall survival between groups of patients. A p-value of ⩽0.05 was considered statistically significant. Statistical analyses were performed using SPSS (version 22, IBM SPSS).

Results Gene expression differences in neoadjuvanttreated PDACs There were 99 genes that were differentially expressed (p ⩽ 0.01 and twofold) in the neoadjuvant chemoRT-treated samples compared to the no-treatment PDAC (Supplemental Table 1). Of particular interest, there were multiple granin family genes upregulated in the chemoRT cohort of patients compared to the surgical specimens (Table 2). Using the Fisher exact test, the list of 99 differentially expressed genes was examined for different categories that were highly represented. The most represented Gene Ontology (GO) biological process were “synaptic transmission” followed by “regulation of insulin secretion,” “energy reserve

6

Tumor Biology

Table 2.  Key genes, cell processes, and expression targets that are upregulated in the chemoradiation-treated tumors using a p ⩽ 0.01 and twofold cutoff using the Fisher exact test. Gene symbol

Gene name

ChemoRT versus surgery—fold change

p-value

CHGA CHGB SCG2 SCG3 SCG5 SCGN

Chromogranin A Chromogranin B Secretogranin II Secretogranin III Secretogranin V Secretagogin, EF-hand calcium binding protein

3.69 5.28 2.24 3.26 3.37 3.91

3.18E−03 3.37E−03 5.08E−03 2.06E−03 3.94E−02 1.03E−02

Overlap

p-value

23 46 14 18 10 52 10

1.33E−09 1.94E−09 2.02E−08 6.12E−08 7.70E−08 1.26E−07 7.43E−07

8 5 7 10 3 7 4 4 5 3 8 10

1.97E−07 3.15E−07 3.66E−07 4.90E−07 7.09E−07 7.56E−07 6.59E−06 7.95E−06 1.60E−05 4.88E−05 9.89E−05 3.73E−04

  Genes regulating cell processes of   Transmission of nerve impulse   Cell differentiation   Neurotransmitter secretion  Exocytosis   Hormone secretion   Cell proliferation   Synaptic plasticity Expression targets of  Neurohormone  MAFA  NEUROD1  ADCYAP1   Chromogranin B (CHGB)  PDX1  HNRNPU  POUF42  NEUROG3  MYT1   Brain-derived neurotrophic factor (BDNF)   Nerve growth factor (NGF)

metabolic process,” “protein polymerization,” “regulation of cell migration,” and “neurotransmitter secretion” (Table 2). The neural theme was also highly represented in the most upregulated expression targets which included neurohormone, MAFA, NEURODI, ADCYAP1, GHGB, PDX1, NEUROG3, MYT1, brain-derived neurotrophic factor (BDNF), and NFG (Table 2). Pathway signaling analysis found the Trk receptor family to be mediated by neuroendocrine signaling, NEUROD1, BDNF, and NFG (Figure 1) indicating that these receptors may be a potential treatment target.

Does fractionated irradiation alter graninrelated genes and induce post-treatment radioresistance in vitro? To investigate whether a course of fractionated irradiation could induce granin-related gene expression in vitro, we studied MIA-PaCa-2 and PANC-1 cell lines before and after receiving 15 fractions of 2 Gy over 3 weeks. RT-PCR was used to study the gene expression levels of CGHA,

CGHB, SCG3, and SST weekly up to 6 weeks after the course of radiation (Figure 2). CHGA and SST were not detectable in the MIA-PaCa-2 cell line either before or after irradiation. In PANC-1 cells, SCG3, SST, and CHGB all showed a significant elevation that persisted, although with progressive decline, for 6 weeks post-irradiation. The effect of radiation on CHGA was not significant. CHGB showed a modest increase within the first 2 weeks postirradiation which did not reach significance while SCG3 was barely detectable in MIA-PaCa-2 cells and also showed minimal changes post-irradiation. To assess whether exposure to a fractionated irradiation schedule induced subsequent radioresistance, MIAPaCa-2 and PANC-1 cell lines were exposed to 15 fractions of 2 Gy over 3 weeks as described above. Following the fractionated schedule, the surviving cells were assessed for radiosensitivity by standard clonogenic survival assay (re-irradiation with single, graded doses of radiation) at 5, 14, and 26 days after the end of the fractionated treatment (Figure 3). Both cell lines showed a similar response with a more radioresistant response at 5 days post fractionation

7

Johnson et al. but no difference, compared to the unirradiated controls, at days 14 and 26 after the end of fractionated treatment.

Do Trk inhibitors modulate the radiation response in vitro? Figure 4(a) shows the data for the effect of AZD1332 alone on the growth of PANC-1 and MIA-PaCa-2 cells, and the

combination of drug with radiation is shown in Figure 4(b) and (c). MIA-PaCa-2 cells were more sensitive to the growth inhibitory effects of AZD1332 compared with PANC-1; the IC50 for MIA-PaCa-2 was 1.65 µM while it was estimated to be 9.70 µM for PANC-1. For the combination of the drug with radiation, a dose of 2.5 µM was selected, and its interaction was studied at 0.5, 1.0, 2.0, and 4.0 Gy. Both cell lines were relatively resistant to radiation alone but the addition of AZD1332 resulted in a dose modifying factor (DMF) of 1.24 for PANC-1 (Figure 4(b)) but had a greater effect in MIA-PaCa-2 where the DMF was 2.06 (Figure 4(c)).

Does a specific Trk inhibitor increase the effect of radiation in vivo? The MIA-PaCa-2 cell line was chosen to assess the effect of AZD1332 on the radiation response in vivo due to its greater interaction in vitro and the limited availability of the agent. A dose of 20 mg/kg delivered daily for 3 weeks was chosen based on existing pre-clinical data;25 there was no existing data combining AZD1332 with radiation. Figure 5 shows that there was a very minimal effect of the drug alone on the growth of MIA-PaCa-2 xenografts but unequivocally no interaction with radiation in vivo.

Figure 1.  Regulation of Trk receptors by neuroendocrine signaling pathways. Genes in red are upregulated in the chemoRT group, blue are downregulated, and gray show no change.

Does Trk receptor expression have any prognostic significance? Results obtained from immunohistochemistry (IHC) using the specific antibodies against TrkA, B, and C and

Figure 2.  Long-term gene expression analysis of SCG3, SST, CHGA, and CHGB in response to a 3-week course of fractionated ) and MIA-Paca-2 ( ) cell lines. Data are expressed as the mean ± SEM of the fold difference relative irradiation in PANC-1 ( to the PANC-1 control (normalized to GAPDH).

8

Tumor Biology

Figure 3.  The effect of a course of fractionated radiation on the subsequent radiosensitivity of PANC-1 (a) and MIA-PaCa-2 (b) cells at 5, 14, and 26 days after the end of the treatment.

Figure 4.  The effect of AZD1332 on the growth of (a) PANC-1 and MIA-PaCa-2 cell lines assessed by the MTT assay. The effect of AZD1332 on the radiation response of (b) PANC-1 and (c) MIA-PaCa-2 cells assessed by clonogenic survival.

the pan-Trk antibody are shown in Figure 6. For all antibodies, there was membranous and diffuse cytoplasmic staining with varying staining intensity between the antibodies and among different PDAC cores. TrkB showed the highest expression levels within the PDAC tumors for the individual receptors with a median intensity of 0.37 (range: 0.31–0.46), while the pan-Trk antibody exhibited the most intense staining pattern and intensity distribution (median: 0.49, range: 0.33–0.66). TrkC showed the least staining within this group of patients (median intensity: 0.17, range: 0.11–0.38), while TrkA was an intermediate

between TrkB and C with a median intensity of 0.23, and range of 0.12–0.32. NEUROD1 showed weak diffuse membranous/cytoplasmic staining (mean: 0.19, range: 0.12–0.59). Receptor expression was dichotomized using the median intensity value, and higher and lower expression correlated with overall survival in Figure 7. None of the individual receptors or NEUROD1 showed a significant relationship with overall survival. The pan-Trk antibody showed a trend for the lower expressing tumors to have improved survival but this did not reach significance

Johnson et al.

Figure 5.  The effect of AZD1332 and radiation on the growth of MIA-PaCa-2 xenografts. The tumor volumes were normalized to the individual tumor volume at the start of treatment. Six mice were analyzed for each treatment arm, and the data represent the mean normalized volume ± SEM.

(p = 0.212). There was no correlation between Trk or NEUROD1 expression and the incidence of distant metastases.

Discussion This study was based on our initial findings that several components of neuroendocrine signaling pathways were upregulated after CRT in patients treated for adenocarcinoma of the pancreas. Figure 1 highlights the relationships between some of the key genes we have studied and shows the central role of the Trk receptor family in mediating these signaling pathways. The three Trk receptor isoforms, TrkA, TrkB, and TrkC, function as high-affinity growth factor receptors for the neurotrophin family of soluble ligands that include nerve growth factor (NGF) and BDNF and are associated with neuronal maintenance and survival during development.26 Trk receptors have also been shown to act as oncogenes with roles in malignant transformation, metastasis, and survival signaling in human tumors.27,28 Several independent mechanisms of Trk pathway activation have been described including constitutive oncogenic fusions, autocrine signaling, and point mutations. The ETV6-NTRK3 gene fusion has been implicated as a transforming factor in multiple cell lineages29 with particular importance in breast cancer.30 However, pathway activation has also been reported in thyroid, lung, prostate, neuroblastoma, medulloblastoma, acute myelogenous leukemia, and pancreatic cancer18,19 signifying that this pathway may be more generally involved in cancer development and tumor cell survival. Several reports suggest that the neurotrophin family of growth factors and their cognate receptors (TrkA, B and C) may play a significant

9 role in the propensity for PDACs to invade perineural tissue.19,31,32 The initial data that stimulated this study were the observations of upregulation of a series of neuroendocrine-related genes in the post-CRT surgical specimens of PDAC patients in comparison with patients who underwent surgery as the primary treatment. Neuroendocrine genes and pathways have not figured in the core signaling pathways revealed by global genomic analysis in PDAC14 but those analyses have been carried out on matched normal and advanced pancreatic adenocarcinomas. Indeed, in this study there was also a group (not matched) of normal controls (data not shown). When comparing the normal pancreas to either the surgical specimens or the CRT patients, we identified a different set of genes and pathways that included many of the core signaling pathways described by Jones et al.,14 in particular apoptosis, cell cycle, cell migration, Wnt signaling, and TGF-β signaling. There was a remarkable difference when the CRTtreated PDACs were compared with the non-treated PDAC cases (Table 2) highlighting that the differential gene expression was due to the treatment and presumably the selection of resistant clones of cells that survived CRT. Interestingly, we found little evidence for genes involved in stem cell signaling or epithelial to mesenchymal transition to have more prevalent expression following CRT apart from a 2.66-fold increase in aldehyde dehydrogenase 1A3 (ALDH1A3). To study whether this observation could be replicated in vitro, we investigated the effect of a fractionated course of radiation on subsequent long and short-term gene expression of some key genes which were upregulated in CRT patients including CHGA (3.69-fold), CHGB (5.28fold), SCG3 (3.26-fold), and SST (4.85-fold). The PANC-1 cell line showed the most striking changes in gene expression following the fractionated schedule with significant changes in SCG3, SST, and CHGB which were elevated immediately following radiation treatment and remained elevated for up to 6 weeks. CHGB was also upregulated in the MIA-PaCa-2 cell line whereas the other genes were undetectable before or after irradiation. There is no literature on the direct effect of radiation on chromogranins, secretogranins, and somatostatin although radiation has been shown to induce somatostatin receptors in small cell lung cancer at low doses33,34 in the context of radionuclide therapy. These results support the observations in the PDAC tumors. However, we were not able to demonstrate that the persistent upregulation of these genes (and other genes not tested) led to an increase in radioresistance. Subtle changes in radiosensitivity were seen within a few days of the completion of the fractionation schedule which were probably due to residual effects of the recovering population of surviving cells but this was lost at 14 and 26 days after the end of treatment (Figure 3).

10

Tumor Biology

Figure 6.  Trk receptor expression in a tissue microarray of 77 PDACs. The left-hand panel shows the staining intensity distribution obtained from image analysis for each antibody studied, and the right-hand panel shows a representative core stained for all four antibodies.

11

Johnson et al.

Figure 7.  Kaplan–Meier’s overall survival analysis of the patient cohort as a function of Trk receptor expression: (a) pan-Trk, (b) TrkA, (c) TrkB, and (d) TrkC. The groups were dichotomized by the median staining intensity as either below (——) or above (- - - - -) median.

Inhibition of the Trk pathway as a therapeutic strategy has been studied using both small-molecule kinase inhibitors and monoclonal antibodies mainly for central nervous system disorders but also for cancer therapy. The most frequently studied compound has been CEP-701 otherwise known as lestaurtinib,35–37 and a phase I trial of lestaurtinib in combination with Gem has been reported in advanced PDAC.38 Unfortunately, the trial accrued few patients and was terminated early with no evidence of radiological responses. A confounding feature of lestaurtinib is that it is a multikinase inhibitor with activity against PDGFR, RET, VEGF, PKC, Flt3, and JAK2 kinases. We chose to study AZD1332 because it is a selective

small-molecule inhibitor of the Trk tyrosine kinase family with potent ATP-competitive inhibition of the three Trk isoforms.25,39 AZD1332 has not been studied in the context of PDAC previously nor has it been previously combined with radiation. The in vitro data were very promising (Figure 3) where both cell lines showed significant growth inhibition in response to increasing concentrations of drug with MIA-PaCa-2 being more sensitive. Even more encouraging was the combination with radiation in the clonogenic assays. There was obvious interaction with radiation again particularly in the MIA-PaCa-2 cell line where we observed a DMF of 2. Disappointingly the in vitro data did not translate into efficacy in vivo. Using a

12 3-week fractionated schedule of 2 Gy combined with daily AZD1332, we failed to demonstrate any significant effect of the drug alone or its combination with radiation. The dosing schedule we used was based on other pre-clinical data25 and should have shown efficacy if it was active in this xenograft model. We also failed to see detectable effects on treatment-induced gene expression changes in the xenograft model (data not shown). Finally, we studied the correlation between Trk receptor and NEUROD1 expression and clinical outcome in a TMA consisting of 77 PDAC tumors treated with surgery followed by CRT. We were unable to demonstrate any clear cut relationships with overall survival or incidence of distant metastases and Trk receptor or NEUROD1 expression. There was a trend for tumors with lower expression of Trk receptors, as identified by the pan-Trk antibody, to have better survival but this was not significant. In conclusion, we have identified a novel gene expression pattern in patients who have undergone neoadjuvant CRT for PDAC that implicates neuroendocrine-related pathways and shown that radiation does upregulate those pathways in pancreatic cancer cell lines although this did not lead to radioresistance. In addition, we demonstrated that a specific inhibitor of the Trk receptor family, which are key receptors involved in the upregulated pathways, has activity against the pancreatic cell lines and interacted to enhance their radiation sensitivity in vitro. However, we were not able to demonstrate any in vivo activity of the drug or its combination with radiation. Further research is required to understand the significance of radiation interaction with neuroendocrine signaling pathways but Trk receptors do not seem to be a promising target for treatment in PDAC. Acknowledgements The authors are grateful to AstraZeneca for the provision of AZD1332 for this study.

Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding Part of this research was supported by the Mopper Fund and part by a Herb & Betty Fisher Seed Research Grant.

References 1. Siegel R, Ma J, Zou Z, et al. Cancer statistics. CA Cancer J Clin 2014; 2014(64): 9–29. 2. Winter JM, Brennan MF, Tang LH, et al. Survival after resection of pancreatic adenocarcinoma: results from a single institution over three decades. Ann Surg Oncol 2012; 19: 169–175. 3. Takaori K, Bassi C, Biankin A, et al. International Association of Pancreatology (IAP)/European Pancreatic

Tumor Biology Club (EPC) consensus review of guidelines for the treatment of pancreatic cancer. Pancreatology 2016; 16(1): 14– 27. 4. Mehta VK, Fisher G, Ford JA, et al. Preoperative chemoradiation for marginally resectable adenocarcinoma of the pancreas. J Gastrointest Surg 2001; 5: 27–35. 5. Gillen S, Schuster T, Meyer Zum Buschenfelde C, et al. Preoperative/neoadjuvant therapy in pancreatic cancer: a systematic review and meta-analysis of response and resection percentages. PLoS Med 2010; 7: e1000267. 6. Conroy T, Desseigne F, Ychou M, et al. Folfirinox versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 2011; 364: 1817–1825. 7. Von Hoff DD, Ervin T, Arena FP, et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 2013; 369: 1691–1703. 8. Danovi SA, Wong HH and Lemoine NR. Targeted therapies for pancreatic cancer. Br Med Bull 2008; 87: 97–130. 9. Zagouri F, Sergentanis TN, Chrysikos D, et al. Molecularly targeted therapies in metastatic pancreatic cancer: a systematic review. Pancreas 2013; 42: 760–773. 10. Chiorean EG and Coveler AL. Pancreatic cancer: optimizing treatment options, new, and emerging targeted therapies. Drug Des Devel Ther 2015; 9: 3529–3545. 11. Frank TS, Sun X, Zhang Y, et al. Genomic profiling guides the choice of molecular targeted therapy of pancreatic cancer. Cancer Lett 2015; 363: 1–6. 12. Seicean A, Petrusel L and Seicean R. New targeted therapies in pancreatic cancer. World J Gastroenterol 2015; 21: 6127–6145. 13. Senderowicz AM, Johnson JR, Sridhara R, et al. Erlotinib/ gemcitabine for first-line treatment of locally advanced or metastatic adenocarcinoma of the pancreas. Oncology 2007; 21: 1696–1706; discussion 1706–1699, 1712, 1715. 14. Jones S, Zhang X, Parsons DW, et al. Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 2008; 321: 1801–1806. 15. Westerman BA, Poutsma A, Maruyama K, et al. The proneural genes NEUROD1 and NEUROD2 are expressed during human trophoblast invasion. Mech Dev 2002; 113: 85–90. 16. Osborne JK, Larsen JE, Gonzales JX, et al. NEUROD1 regulation of migration accompanies the differential sensitivity of neuroendocrine carcinomas to TrkB inhibition. Oncogenesis 2013; 2: e63. 17. Tessarollo L. Pleiotropic functions of neurotrophins in development. Cytokine Growth Factor Rev 1998; 9: 125– 137. 18. Thiele CJ, Li Z and McKee AE. On Trk—the TrkB signal transduction pathway is an increasingly important target in cancer biology. Clin Cancer Res 2009; 15: 5962–5967. 19. Miknyoczki SJ, Lang D, Huang L, et al. Neurotrophins and Trk receptors in human pancreatic ductal adenocarcinoma: expression patterns and effects on in vitro invasive behavior. Int J Cancer 1999; 81: 417–427. 20. Sclabas GM, Fujioka S, Schmidt C, et al. Overexpression of tropomysin-related kinase b in metastatic human pancreatic cancer cells. Clin Cancer Res 2005; 11: 440–449. 21. Zhu Z, Kleeff J, Kayed H, et al. Nerve growth factor and enhancement of proliferation, invasion, and tumorigenicity

Johnson et al. of pancreatic cancer cells. Mol Carcinog 2002; 35: 138– 147. 22. Zhu ZW, Friess H, Wang L, et al. Nerve growth factor exerts differential effects on the growth of human pancreatic cancer cells. Clin Cancer Res 2001; 7: 105–112. 23. Robertson JM, Margolis J, Jury RP, et al. Phase I study of conformal radiotherapy and concurrent full-dose gemcitabine with erlotinib for unresected pancreatic cancer. Int J Radiat Oncol Biol Phys 2012; 82: e187–192. 24. Huang J, Robertson JM, Margolis J, et al. Long-term results of full-dose gemcitabine with radiation therapy compared to 5-fluorouracil with radiation therapy for locally advanced pancreas cancer. Radiother Oncol 2011; 99: 114–119. 25. Thress K, Macintyre T, Wang H, et al. Identification and preclinical characterization of AZ-23, a novel, selective, and orally bioavailable inhibitor of the Trk kinase pathway. Mol Cancer Ther 2009; 8: 1818–1827. 26. Huang EJ and Reichardt LF. Neurotrophins: roles in neuronal development and function. Annu Rev Neurosci 2001; 24: 677–736. 27. Nakagawara A. Trk receptor tyrosine kinases: a bridge between cancer and neural development. Cancer Lett 2001; 169: 107–114. 28. Rubin JB and Segal RA. Growth, survival and migration: the Trk to cancer. Cancer Treat Res 2003; 115: 1–18. 29. Lannon CL and Sorensen PH. ETV6-NTRK3: a chimeric protein tyrosine kinase with transformation activity in multiple cell lineages. Semin Cancer Biol 2005; 15: 215–223. 30. Euhus DM, Timmons CF and Tomlinson GE. ETV6 NTRK3-Trk-ing the primary event in human secretory breast cancer. Cancer Cell 2002; 2: 347–348.

13 31. Sakamoto Y, Kitajima Y, Edakuni G, et al. Expression of Trk tyrosine kinase receptor is a biologic marker for cell proliferation and perineural invasion of human pancreatic ductal adenocarcinoma. Oncol Rep 2001; 8: 477–484. 32. Schneider MB, Standop J, Ulrich A, et al. Expression of nerve growth factors in pancreatic neural tissue and pancreatic cancer. J Histochem Cytochem 2001; 49: 1205–1210. 33. Oddstig J, Bernhardt P, Nilsson O, et al. Radiation-induced up-regulation of somatostatin receptor expression in small cell lung cancer in vitro. Nucl Med Biol 2006; 33: 841–846. 34. Oddstig J, Bernhardt P, Nilsson O, et al. Radiation induces up-regulation of somatostatin receptors 1, 2, and 5 in small cell lung cancer in vitro also at low absorbed doses. Cancer Biother Radiopharm 2011; 26: 759–765. 35. Hexner EO, Mascarenhas J, Prchal J, et al. Phase I dose escalation study of lestaurtinib in patients with myelofibrosis. Leuk Lymphoma 2015; 56: 2543–2551. 36. Minturn JE, Evans AE, Villablanca JG, et al. Phase I trial of lestaurtinib for children with refractory neuroblastoma: a new approaches to neuroblastoma therapy consortium study. Cancer Chemother Pharmacol 2011; 68: 1057–1065. 37. Shabbir M and Stuart R. Lestaurtinib, a multitargeted tyrosine kinase inhibitor: from bench to bedside. Expert Opin Investig Drugs 2010; 19: 427–436. 38. Chan E, Mulkerin D, Rothenberg M, et al. A phase I trial of cep-701 + gemcitabine in patients with advanced adenocarcinoma of the pancreas. Invest New Drugs 2008; 26: 241–247. 39. Wang T, Lamb ML, Scott DA, et al. Identification of 4-aminopyrazolylpyrimidines as potent inhibitors of Trk kinases. J Med Chem 2008; 51: 4672–4684.