Biochemical and Transcriptomic Effects of Herring ... - ACS Publications

Article pubs.acs.org/est

Biochemical and Transcriptomic Effects of Herring Gull Egg Extracts from Variably Contaminated Colonies of the Laurentian Great Lakes in Chicken Hepatocytes Doug Crump,*,† Kim L. Williams,† Suzanne Chiu,† Robert J. Letcher,† Luke Periard,† and Sean W. Kennedy†,‡ †

Ecotoxicology and Wildlife Health Division, Environment Canada, National Wildlife Research Centre, Carleton University, Ottawa, ON, Canada K1A 0H3 ‡ Department of Biology, University of Ottawa, Ottawa, ON, Canada K1N 6N5 S Supporting Information *

ABSTRACT: Determining the effects of complex mixtures of environmental contaminants poses many challenges within the field of ecotoxicology. In this study, graded concentrations of herring gull egg extracts, collected from five Great Lakes breeding colonies with variable burdens of organohalogen contaminants (OHCs), were administered to chicken embryonic hepatocytes to determine effects on 7-ethoxyresorufin-O-deethylase (EROD) activity, porphyrin accumulation, and mRNA expression. EROD activity and porphyrin accumulation permitted the ranking of colonies based on the efficacy of eliciting an aryl hydrocarbon receptor-mediated response. An avian ToxChip polymerase chain reaction (PCR) array provided more exhaustive coverage in terms of potential toxicity pathways being affected, including xenobiotic and lipid metabolism and the thyroid hormone pathway. Herring gull eggs from Channel Shelter Island (CHSH, Lake Huron) and Gull Island (GULL, Lake Michigan) had among the highest OHC burdens, and extracts elicited a biochemical and transcriptomic response greater than that of extracts from the other three, less polluted colonies. For example, EROD EC50 values and porphyrin ECthreshold values were lower for CHSH and GULL extracts than for the other colonies. Extracts from CHSH and GULL altered 15 and 13 of 27 genes on the PCR array compared to no more than eight genes for the less contaminated sites. The combination of a wellestablished avian in vitro assay, two well-characterized biochemical assays, and the avian ToxChip PCR array permitted the geographical discrimination of variably contaminated herring gull eggs from the Great Lakes. Such high-throughput assays show potential promise as cost-effective tools for determining toxic potencies of complex mixtures in the environment.

■

INTRODUCTION A majority of ecotoxicological studies focus on exposure and effects of single chemicals and do not accurately reflect the environmental reality for organisms exposed to complex mixtures of chemical stressors. The ecotoxicological effects of industrial mixtures were initially investigated in a series of studies in The Netherlands in the 1980s.1 Three general approaches have emerged to assess whole mixture toxicities and identify the causative components: (1) effect-directed analysis (EDA), (2) toxicity identification and evaluation (TIE), and (3) whole effluent toxicity (WET) testing.1 Furthermore, Teuschler et al. proposed that cutting edge research approaches and methodologies (e.g., genomics, proteomics, biochemistry, and bioinformatics) should be used to support the scientific advances in mixture toxicology.2 EDA can be applied to both in vitro and in vivo studies to detect toxicologically active fractions and compounds in various environmental matrices.1 Extracts derived from avian eggs represent one such environmental matrix. Powell et al. collected Published 2015 by the American Chemical Society

double-crested cormorant (Phalacrocorax auritus) eggs from breeding colonies near Green Bay, Lake Michigan, an area that had high levels of organohalogen contaminants (OHCs) and historical evidence of compromised reproduction.3 Cormorant egg extracts were prepared and administered to chicken eggs via injection. A 1 egg equivalent [∼322 pg of toxic equivalents, relative to 2,3,7,8-tetrachloro-dibenzo-p-dioxin (2,3,7,8-TCDD) (TEQs) per gram] caused significantly greater chicken embryonic mortality (77%) compared to vehicle-injected embryos.3 The same cormorant extract was injected into fertilized cormorant embryos collected from an uncontaminated site, and the mortality rate was identical between the vehicle control and the 1 egg equivalent extract, highlighting Received: Revised: Accepted: Published: 10190

June 5, 2015 July 16, 2015 July 20, 2015 July 20, 2015 DOI: 10.1021/acs.est.5b02745 Environ. Sci. Technol. 2015, 49, 10190−10198

Environmental Science & Technology

■

species-specific lethality responses between chickens and cormorants.4 End points other than LD50 values can be incorporated into mixture-based studies. For example, 7-ethoxyresorufin-Odeethylase (EROD) activity is often used as a surrogate measure for the identification of hazardous aromatic compounds in EDA.5 In the 1990s, a chicken embryonic hepatocyte (CEH) screening assay was used to determine effects of complex mixtures of OHCs extracted from herring gull eggs from the Great Lakes and great blue heron eggs from British Columbia on EROD activity and porphyrin accumulation.6,7 All extracts, comprising polychlorinated biphenyls (PCBs), polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs), and other nonpolar OHCs, led to EROD induction and porphyrin accumulation. Lorenzen et al. exposed CEH to OHC extracts prepared from common tern eggs collected from three Great Lakes sites and one reference site from the east coast of Canada.8 Tern extracts from Hamilton Harbour, Lake Ontario, had the highest concentrations of ∑PCBs and ∑PCDD/Fs and an EROD EC50 significantly lower than that of extracts prepared from a non-Great Lakes reference site.8 More recent studies have incorporated the use of genomics for the characterization of whole effluent or mixture toxicity in aquatic organisms. For example, the effects of whole effluent on hepatic gene expression in common carp (Cyprinus carpio) were determined using a cDNA microarray,9 and Falciani et al. aimed to differentiate six variably contaminated sites of origin of European flounder (Platichthys f lesus) by screening hepatic gene expression profiles.10 Christiansen et al. used a transcriptomics-based approach to screen for effects of environmental mixtures in largescale suckers (Catostomus macrocheilus) collected from three variably contaminated areas of the Columbia River.11 Expression of 72 probes correlated significantly with liver concentrations of various organochlorines (OCs), PCBs, and polybrominated diphenyl ether (PBDE) flame retardants and included genes associated with drug and lipid metabolism, oxidative stress, and apoptosis.11 To the best of our knowledge, such studies have not been conducted with avian species. The objectives of this study were (1) to characterize the concentration of a suite of legacy and emerging OHCs in herring gull egg extracts collected from five geographically distinct Great Lakes colonies and (2) to utilize an EDA approach based on a well-established, high-throughput avian in vitro assay, two well-characterized biochemical assays, and a more recently developed avian ToxChip polymerase chain reaction (PCR) array to determine spatial differences in effects among colonies. The ultimate goal of the study was to determine if this EDA approach could permit the geographical discrimination of variably contaminated herring gull egg extracts from the Great Lakes. Graded concentrations of herring gull egg extracts were administered to CEH, and EROD activity and porphyrin accumulation were determined to compare and rank colonies in terms of efficacy of eliciting an aryl hydrocarbon receptor (AhR)-mediated response. An avian ToxChip PCR array12 was used to determine transcriptomic effects of the five colonies on 27 genes covering a range of toxicologically relevant pathways, including xenobiotic and lipid metabolism, the thyroid hormone pathway, apoptosis, and oxidative stress.

Article

EXPERIMENTAL SECTION

Egg Collection. Herring gull eggs (five per colony) were collected between late April and early May 2012 from Gull Island, Lake Michigan (GULL; 45°42′10.39″N, 85°50′15.45″W), Chantry Island, Lake Huron (CHT; 44°29′21.89″N, 81°24′6.90″W), Channel Shelter Island, Lake Huron (CHSH; 43°40′10.22″N, 83°49′26.98″W), Leslie Street Spit, Lake Ontario (LSS; 43°37′21.81″N, 79°20′26.57″W), and Strachan Island, St. Lawrence River (STR; 45°1′16.28″N, 74°48′42.07″W). A total of five fertilized, unincubated white leghorn chicken eggs were obtained in 2013 from the Canadian Food Inspection Agency (CFIA, Ottawa, ON). For both herring gull and chicken eggs, the five eggs were pooled on an equal wet weight (ww) basis and stored at −20 °C prior to the preparation of OHC extracts. Sample Preparation for Contaminant Analysis and Hepatocyte Administration. Herring gull and chicken egg homogenates were freeze-dried and prepared for contaminant analysis and hepatocyte administration according to details provided in the Supporting Information. Briefly, 6.0−24.0 g (freeze-dried weight) of the egg homogenates and triplicate 1 g (ww) samples of 1947 NIST fish SRM were mixed with sodium sulfate and extracted with 6 volumes of 50 mL of a 1:1 DCM/ hexane mixture. Each extract volume of 300 mL was rotary evaporated and concentrated to a volume that resulted in a concentration of 0.5 g ww of egg sample/mL. Experimental blanks and the SRM fish extracts were concentrated to 0.1 g ww/mL. PCBs, OCs, PBDEs, and non-PBDE flame retardants (FRs) were extracted according to previously published methods.13,14 Briefly, three 1 mL samples, equivalent to 0.5 g ww, were taken from the original egg pool extracts. For the NIST 1947 SRM fish extracts, the 1 mL volume was equivalent to 0.1 g ww. These subsamples were spiked with 25 μL of an internal standard solution containing BDE-30, BDE-156, 13C10-synDechlorane Plus (DDC-CO), 13C10-anti-DCC-CO, 13C12-BDE209, six 13C12-PCB congeners, and 19 13C6−14-OCs at concentrations ranging from 0.5 to 1.16 ng/μL (Table S1). A 10% portion of the spiked sample was transferred to a preweighed aluminum dish for the gravimetric determination of lipid content. Bulk lipid removal of sample extract was conducted using gel permeation chromatography (GPC) for two sample extract portions: (1) the remaining 0.5 g ww sample/mL of the nonspiked total egg extract for CEH administration and (2) the remaining spiked subsamples for contaminant analysis (not including PCDDs/Fs). As in previous analysis of the target contaminants in eggs of herring gulls from the Great Lakes,13 the spiked GPC sample extracts were further cleaned up using a LC-Si solid phase extraction (SPE) and concentrated in 200 μL of isooctane for GC−MS analysis. The GPC extract from the total unspiked herring gull and chicken egg homogenate was separated into subportions in hexane of 1 g ww/mL of original egg pool sample and subjected to SPE as described above. SPE extracts were evaporated to dryness and reconstituted in 0.5 mL of dimethyl sulfoxide (DMSO) for subsequent administration to CEH. This final sample fraction represented >93% of the original weight of the egg pool homogenate. Determination of OHC Concentrations. Sample fractions were analyzed for 74 PCB congeners, 20 OCs, 47 BDE congeners, and 24 non-PBDE flame retardant compounds 10191

DOI: 10.1021/acs.est.5b02745 Environ. Sci. Technol. 2015, 49, 10190−10198


Article

based on methods described elsewhere.13,14 The specific identity of all contaminants analyzed is available in the Supporting Information along with the GC and MS operating parameters. All PCB congeners and OCs in the spiked sample fractions were determined by gas chromatography−singlequadrupole mass spectrometry operated in electron impact mode [GC−MS(EI)]. BDE congeners and non-PBDE FRs were determined by gas chromatography−single-quadrupole mass spectrometry operated in electron caption negative ionization mode [GC−MS(ECNI)]. Quantification of all compounds was performed using an internal standard method based on the relative response factor (RRF) of the corresponding internal standard and target compounds. Data analysis was performed using Agilent G1701EA GC/ MSD ChemStation software. A calibration curve with five concentrations of target compounds and constant concentrations of internal standards was first generated. The calibration curves showed good linearity for PCBs (r2 > 0.99), OCs (r2 > 0.98), and FRs (r2 > 0.98). All contaminant concentrations are reported on a nanogram per gram ww basis. A blank sample was included with each colony- and chicken control-specific batch of triplicate egg homogenate subsamples and triplicate NIST 1947 SRM fish homogenate samples. Working standards and solvent blanks were used to calibrate the GC−MS instrument for every batch of seven samples. On the basis of the spiked internal standards, the mean percent recoveries of all analytes from egg or fish samples were generally >70%. Method limits of quantification (MLOQs) were based on 10 times the signal-to-noise ratio, and method limits of detection (MLODs) were based on 3 times the signal-to-noise ratio. MLOQs for PCBs, OCs, and FRs in egg homogenate and fish samples ranged from 0.1 to 3 ng/g ww. MLODs for PCBs, OCs, and FRs ranged from 0.04 to 0.6 ng/g ww. See the Supporting Information for the MLOD and MLOQ concentrations for all individual compounds analyzed (Tables S2−S7). Finally, 5 g aliquots of the herring gull and chicken egg homogenates (pre-freeze-drying) were sent to RPC Laboratories (Fredericton, NB) for analysis of polychlorinated dibenzop-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and non-ortho-substituted PCBs using gas chromatography− high-resolution mass spectrometry (GC−HRMS). Methods were based on U.S. EPA Methods 1613B and 8290A for dioxins and furans and on U.S. EPA Method 1668C for non-ortho PCBs. Reference materials, blanks, and duplicates were also analyzed for quality assurance purposes. Data were analyzed by RPC laboratories, and TCDD-equivalent concentrations (TEQs) were calculated on the basis of the WHO avian toxic equivalency factors (TEFs).15 Preparation and Dosing of CEH Cultures. Fertilized, unincubated white leghorn chicken eggs were obtained from the CFIA and artificially incubated (Petersime Model XI) at 37.5 °C and 60% relative humidity. On incubation day 19, embryos (n = 25) were euthanized by decapitation and livers were removed and pooled. Cultured hepatocytes were prepared by collagenase digestion and filtration, as described previously.16,17 Hepatocytes were distributed into 48-well plates by adding 25 μL of the cell suspension to 500 μL of supplemented medium. The cells were incubated at 37.5 °C and 5% CO2 for 24 h and then treated with the DMSO vehicle control or serial dilutions of “neat” extracts prepared from one of the five herring gull egg pools or the chicken egg pool (n = 3 wells/ treatment for EROD plates, and n = 6 wells/treatment for the

PCR array). The extract dilutions used for EROD plates were 1 (“neat”), 0.3, 0.1, 0.03, 0.01, 0.003, 0.001, 0.0003, and 0.0001. For PCR array analysis, extract dilutions of 1 (“neat”) and 0.01 were included. A positive control, consisting of a nominal TCDD concentration of 300 nM, was also included on the EROD plates (n = 3 wells). CEH were incubated for 24 h following extract administration; the medium was then aspirated, and CEH were frozen at −80 °C for subsequent RNA isolation. Plates used for EROD assays were rinsed with 200 μL/well of phosphate-buffered saline ethylenediaminetetraacetic acid prior to being flash-frozen in powdered dry ice and stored at −80 °C. EROD and Porphyrin Assays. EROD assays were conducted as described previously.16,17 Reagents were obtained from Sigma-Aldrich (St. Louis, MO) unless another supplier is indicated. Briefly, CEH were incubated at 37.5 °C in the presence of nicotinamide adenine dinucleotide phosphate (NADPH, reduced) and 7-ethoxyresorufin for 7 min. The reaction was stopped by cold acetonitrile containing fluorescamine (0.15 mg/mL). Resorufin and protein [bovine serum albumin (BSA)] standard curves were prepared on each plate.17 Plates were analyzed for both EROD activity (excitation wavelength of 530 nm and emission wavelength of 590 nm) and total protein concentration (excitation wavelength of 400 nm and emission wavelength of 460 nm) using a fluorescence plate reader (Cytofluor 2350, Millipore, Bedford, MA). Total porphyrin analysis was conducted according to previously reported methods.17 Briefly, 400 μL of uroporphyrin standard solutions was added to the wells that contained resorufin and BSA standards, and HCl (400 μL, 3 M) was added to wells that contained hepatocytes. Plates were incubated at room temperature for 30 min followed by analysis of total porphyrins using the Cytofluor 2350 fluorescent plate reader (excitation wavelength of 400 nm and emission wavelength of 645 nm). PCR Array. Total RNA was extracted from CEH exposed to two concentrations (1 or “neat” and 0.01) of the five herring gull egg extract preparations using the Qiagen RNeasy 96-kit according to the manufacturer’s instructions (Qiagen, Mississauga, ON). To ensure a sufficient RNA yield for PCR array analysis, two replicate wells of hepatocytes were combined prior to RNA extraction (n = 3 technical replicates/treatment group). The resulting RNA was treated with DNase to remove genomic DNA using Turbo DNA-free kits (Ambion, Austin, TX), and RNA quality and concentration were determined using a Nanodrop 2000 (Thermo Scientific). Total RNA (∼200 ng) was reverse transcribed to cDNA using the RT2 First-Strand cDNA synthesis kit (Qiagen) without the genomic DNA elimination step. The custom chicken RT2 Profiler PCR Array was built by SABiosciences (Valencia, CA) according to our specifications (Table S8). Unlike EROD activity or porphyrin accumulation, the PCR array permits a more exhaustive coverage in terms of potential toxicity pathways being impacted beyond the AhR pathway. However, the array does contain an AhR-responsive gene (Cyp1a4), thus permitting comparative verification between the transcriptomic and enzymatic activity of the protein. Each 96-well array contained three identical sets of 27 target genes and five control genes, allowing three technical replicates to be screened per plate. The five control genes included two internal control genes, a positive PCR control, a reverse transcription control, and a genomic DNA contamination control. cDNA was added directly to the RT2 SYBR 10192



Article

had extremely low concentrations of ∑18PCBs, ∑9OCs, ∑10FRs, and total TEQs compared to extracts derived from gull eggs (Tables 1 and 2). Herring gull colonies were ranked

Green Mastermix (Qiagen), and 25 μL of this mixture was added to each well containing a set of primers at preoptimized concentrations. Arrays were run using the Stratagene MX3005P PCR system (Agilent Technologies, Santa Clara, CA) with the following thermal profile: 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min and ending with a dissociation curve segment of 95 °C for 1 min, 55 °C for 30 s, and 95 °C for 30 s. Biochemical and mRNA Expression Data Analysis. EROD activity data were fit to a modified Gaussian curve as described previously.17 For each egg extract preparation, three EROD curves were generated from data derived from separate cell culture plates. EC50 values are presented as the mean values of replicates ± the standard error (SE). For porphyrin concentration−response curves, data were fit to a logistic curve using GraphPad Prism version 5.02. ECthreshold (ECthr) values, which represent the concentrations of an extract that elicit a porphyrin response significantly greater than that of DMSO-treated cells, were determined from concentration− response curves averaged from three replicate curves using a one-way analysis of variance (ANOVA) (p < 0.05) and Tukey’s HSD post hoc test. PCR array data were analyzed using MxPro version 4.10 (Agilent Technologies). An example of raw cycle threshold data is included in Table S9 to show the variability in amplification among replicates. Cycle threshold (Ct) data were normalized to two internal control genes [elongation factor 1-α (Eef1a1) and ribosomal protein L4 (Rpl4)], and the fold change of target gene mRNA abundance relative to the vehicle control was calculated using the 2−ΔΔCt method.18 Significant fold change differences (p < 0.05) between the DMSO vehicle control and the two concentrations of herring gull egg extract administered to hepatocytes (i.e., 1 or “neat” and 0.01) were calculated by one-way ANOVA with Tukey’s HSD post hoc test (GraphPad Prism version 5.02). Hierarchical clustering of the fold change data was performed using R Statistics (R Development Core Team) for the highest concentration of extract administered (i.e., 1 or “neat”). Nonsignificant fold changes and those of 80% of the ∑PCB concentration. For the 20 OCs analyzed, HCB, octachlorostyrene, oxychlordane, heptachlor epoxide, transnonachlor, cis-nonachlor, p,p′-DDE, dieldrin, mirex, and photomirex were quantifiable. Of the 47 BDE congeners analyzed, six congeners (BDE 47, 99, 100, 153, 183, and 209) were quantifiable (i.e., above the MLOQ) and constituted >90% of the ∑PBDE concentration. α- Hexabromocyclododecane (HBCD) was detectable at low nanogram per gram ww levels in eggs from LSS, CHT, and STR. syn- and anti-DCCCO were quantifiable only in LSS egg homogenates at lownanogram per gram ww concentrations. Concentrations of this suite of OHCs were determined in egg extracts to compare levels in eggs collected from spatially distinct colonies in the Great Lakes with those in domestic chicken eggs and, further, to determine variable biochemical and molecular-level effects of these “chemical cocktails” following exposure to CEH. As expected, colony variability was evident in terms of OHC burdens, and chicken egg extracts

on the basis of the sum of various groups of chemicals as follows: (a) ∑18PCBs, CHSH > GULL > LSS > STR > CHT; (b) ∑10OCs, LSS > GULL > CHSH > STR > CHT; (c) ∑10FRs, CHT > LSS > GULL > CHSH > STR (Table 1). In terms of dioxin-like TEQs, the rank order of colonies was as follows: (a) ∑non-ortho PCBs, CHSH ≫ GULL > LSS > CHT ≈ STR; (b) ∑dioxins, CHSH ≫ LSS > GULL > CHT ≈ STR; (c) ∑furans, CHSH ≫ GULL > LSS ≈ CHT > STR (Table 2). The high concentrations of several OHCs at CHSH, a colony located near Saginaw Bay, Lake Huron, were not surprising given its traditionally elevated levels of these OHCs compared to those of other Great Lakes sites.19,20 On the other hand, Chantry Island (CHT) has historically been used as a Great Lakes reference site given the low concentrations of most OHC classes.19,20 For example, ∑18PCB concentrations were 4.78 and 1.59 μg/g and ∑furan concentrations were 29.4 and 2.18 pg of TEQ/g at CHSH and CHT, respectively (Tables 1 and 2). Interestingly, ∑FR concentrations were actually greatest at CHT [0.81 μg/g vs 0.37 μg/g at CHSH (Table 1)] among the 10193



Article

five colonies, indicating the importance of looking at a suite of OHCs and not just legacy organochlorines. The suite of OHCs measured in this study does not represent the full complement of contaminants in herring gull eggs from the Great Lakes. Other studies have reported additional OHCs in herring gull eggs collected from these various colonies such as organophosphate flame retardants,21 perfluoroalkyl acids,22 and other (brominated) contaminants whose source origin is unclear (e.g., methoxylated polybrominated diphenoxybenzenes23). In addition, there are likely many other unknown pollutants present that have yet to be discovered and/or measured. However, the spatial variability observed for the OHCs that were measured allowed us to determine if the biochemical and molecular techniques employed were sensitive enough to link the site of origin with a subtle biological measurement. Effects of Egg Extracts on EROD Activity and Porphyrin Accumulation. Herring gull egg extracts induced EROD activity and porphyrin concentrations in a concentration-dependent manner in CEH (Figure 1A,B). The

and ECthreshold values are listed in Table 3. The EROD EC50 values were significantly lower for CHSH and GULL than for the other three colonies. Porphyrin ECthr varied among colonies, with the lowest observed for CHSH, GULL, and LSS (Figure 1B and Table 3). EROD activity and porphyrin accumulation have been used previously to determine dioxin-like effects of individual compounds and complex environmental mixtures in cell culture studies.8,24−26 The significant yet variable response of these biochemical measures in this study indicates that DLCs are present at variable concentrations in the extracts (verified by chemical analysis) and that herring gull colonies can be spatially organized in part, upon the induction of these end points. While the extracts administered to CEH contained all PCDDs, PCDFs, PCBs, chlorinated pesticides (e.g., p,p′-DDE, dieldrin, and mirex), and other OHCs (e.g., PBBs, PBDEs, and other flame retardants), it is possible that OHCs other than those quantified in the extracts could have contributed to CYP1A induction. Regardless, the CEH assay provided a useful means of screening for dioxin-like effects of the complex mixtures. Extracts derived from herring gull and great blue heron eggs induced EROD activity in a dose-dependent manner in CEH and illustrated how the bioassay could be used to estimate TEQ concentrations in environmental samples contaminated with complex mixtures.6 Lorenzen et al. exposed CEH to common tern extracts prepared from four different colonies (three Great Lakes and one non-Great Lakes reference site) and observed variable alterations of EROD activity and porphyrin accumulation.8 Tern extracts from Hamilton Harbour, Lake Ontario, had the highest concentrations of ∑PCBs, ∑PCDDs, and ∑PCDFs and a significantly lower EROD EC50 compared to the values of extracts prepared from the non-Great Lakes reference site. Porphyrins are the building blocks of heme biosynthesis, and it has been shown that several OHCs can deregulate this process, leading to an increase in liver porphyrin concentrations. A good linear correlation was observed between highly carboxylated porphyrin (HCP) concentrations in the liver of herring gulls from the Great Lakes and their respective PCB concentrations.26 In addition, Kennedy et al. administered nonpolar contaminant extracts prepared from livers of adult herring gulls from the Great Lakes to CEH and observed a positive correlation with HCPs.26 HCP concentrations have therefore been identified as an individual-level adverse outcome and represent an end point that could be measured in fieldcollected avian species, especially from colonies that had the lowest EROD EC50 and/or porphyrin ECthr in the current study (i.e., CHSH and GULL). Avian ToxChip PCR Array. In addition to EROD activity and porphyrin analysis, the custom-designed Avian ToxChip PCR array was used to screen for mRNA expression effects in CEH at two concentrations of herring gull extracts (1 or “neat” and 0.01) compared to a DMSO solvent control. Given the lack of response for EROD activity and porphyrin accumulation following exposure of CEH to chicken egg extracts (Figure 1) and a limitation in the number of plates available, those extracts were not screened using the PCR array. The array contained two internal control genes, Eef1a1 and Rpl4, and their expression was not affected by treatment with any of the extract preparations tested. There was no amplification observed in the genomic DNA contamination control, and the reverse transcription control and positive PCR control met the appropriate quality control and assurance criteria. The

Figure 1. Concentration-dependent effects of herring gull and chicken egg extracts on 7-ethoxyresorufin-O-deethylase (EROD) activity (A) and porphyrin concentration (B) in cultured chicken embryonic hepatocytes. Herring gull eggs were collected from Channel Shelter Island (CHSH), Gull Island (GULL), Leslie Street Spit (LSS), Strachan Island (STR), and Chantry Island (CHT). Data points for EROD activity and porphyrin concentration represent the mean values obtained from triplicate cell culture plates ± the standard error of the mean. On the basis of their respective curve fits, extract concentration values are represented on a log scale for EROD activity and a linear scale for porphyrin concentration.

decrease in EROD activity at higher extract dilutions was expected on the basis of earlier studies24,25 and is due to the competitive inhibition of the reaction by the inducer(s). Neither EROD activity nor porphyrin concentrations were induced in CEH exposed to chicken egg extract because concentrations of dioxin-like compounds (DLCs) were too low to elicit a biochemical response (Tables 1 and 2). EROD EC50 10194



Article

Table 3. Effective Concentration 50 (EC50) and ECthreshold (ECthr) Values for 7-Ethoxyresorufin-O-deethylase (EROD) Activity (picomoles per minute per well) and Porphyrin Concentrations (picomoles per well) in Chicken Embryonic Hepatocyte Cultures Exposed to Graded Concentrations of Herring Gull and Chicken Egg Extracts for 24 ha EROD EC50 (extract dilution) EROD ECthr porphyrin ECthr (extract dilution)

CHSH

GULL

LSS

STR

CHT

chicken

0.0033 ± 0.0006 a 0.003 0.1

0.0037 ± 0.0006 a 0.003 0.1

0.0090 ± 0.0011 b 0.003 0.1

0.0147 ± 0.0018 c 0.01 0.3

0.0093 ± 0.0021 bc 0.01 0.3

NC NC NC

Mean EC50 and ECthr values were derived from three replicate cell culture plates ± the standard error. Superscript letters indicate significant differences among treatments (p < 0.05). Abbreviations: NC, not calculated; CHSH, Channel Shelter Island, Lake Huron; GULL, Gull Island, Lake Michigan; LSS, Leslie Street Spit, Lake Ontario; STR, Strachan Island, St. Lawrence River; CHT, Chantry Island, Lake Huron.

a

Cyp1a4 was the most responsive transcript on the PCR array and was upregulated in a concentration-dependent manner by the five herring gull egg extracts based on the two extract dilutions administered [1 and 0.01 (Figure 2 and Table S10)]. This corroborates the findings of the EROD and porphyrin assays. Following the binding of DLCs with the AhR, the DLC−AhR complex enters the nucleus, binds to the xenobiotic response element on DNA, and interacts with transcriptional cofactors to alter the expression of numerous genes, including cytochrome P450s.27 In vitro measures of avian AhR activation (e.g., Cyp1a4 induction) have been shown to be good predictors of overt DLC toxicity in avian embryos;28 however, non-DLCs have also been shown to alter this gene in part because of the promiscuity of the AhR for structurally divergent compounds. For example, several organic FRs, including tetrabromobisphenol A (TBBPA), TBBPA-bis(2,3-dibromopropyl ether), tris(methylphenyl) phosphate (TMPP), and allyl 2,4,6-tribromophenyl ether, as well as the bisphenol A replacement alternative, bisphenol S, upregulated Cyp1a4 in CEH.12,29 Therefore, the presence of additional chemicals, including flame retardants, in the complex extracts could have contributed to the observed Cyp1a4 mRNA induction. Cyp3a37, another phase I metabolizing enzyme, and Alas1, which controls the production of an enzyme involved in cellular heme biosynthesis (i.e., δ-aminolevulinate synthase I), were upregulated in CEH exposed to four of the five herring gull extracts (Figure 2). The co-induction of these two gene targets was also observed in CEH following administration with BPS and sunlight-irradiated flame retardants, including BDE-209 and tetradecabromo-1,4-diphenoxybenzene (TeDB-DiPhOBz).29,30 Dogra et al. reported co-induction of a related P450 enzyme, Cyp2h1 (not included on the ToxChip array), and Alas1 in the liver of chicken embryos following exposure to cycloheximide and phenobarbital.31 The induction of these genes following exposure to phenobarbital indicates activation of the chicken xenobiotic receptor (CXR), and in mammals, both Cyp3a and Alas1 are regulated by the pregnane X receptor (analogous to CXR).32 Cyp3a5 expression was upregulated in wild largescale suckers exposed to high concentrations of environmental contaminants in the Columbia River,11 and Cyp3a4 was induced in fathead minnows caged in rivers near highly agricultural areas.33 In zebrafish exposed to municipal effluent and other xenobiotics, Cyp3a65 was induced as was a Cyp3a gene in wild male turbot from polluted sites along the California coast.34 Collectively, these studies highlight the potential importance of certain key molecular markers that relate to contaminant burdens in a biologically relevant manner. Normal growth in birds is controlled by the interplay between thyroid hormone, growth hormone (GH), and insulinlike growth factor 1 (Igf1), and there is considerable cross-talk between the hypothalamic−pituitary−thyroid and GH axes.35

complete list of gene targets and their respective fold changes for the two extract concentrations of each herring gull colony are included in the Table S10. The hierarchical clustering of herring gull extracts based on their transcriptional alterations in CEH resulted in a heat map with distinct branches (Figure 2). Extracts from CHSH and

Figure 2. Heat map depicting significant fold changes of genes on the Avian ToxChip PCR array clustered by herring gull colony at the highest administered concentration (“neat” or 1). Red and green hues indicate significant up- and downregulation, respectively, and fold change values of