Article pubs.acs.org/est
Characteristic Molecular Signature for Early Detection and Prediction of Persistent Organic Pollutants in Rat Liver Kwang Hwa Jung,†,‡,§ Jeong Kyu Kim,†,‡,§ Min Gyu Kim,†,‡,§ Ji Heon Noh,†,‡ Jung Woo Eun,†,‡ Hyun Jin Bae,†,‡ Young Gyoon Chang,†,‡ Qingyu Shen,†,‡ Won Sang Park,†,‡ Jung Young Lee,†,‡ and Suk Woo Nam*,†,‡ †
Department of Pathology, College of Medicine and Functional RNomics Research Center, The Catholic University of Korea, Seoul, South Korea ‡ Functional RNomics Research Center, The Catholic University of Korea, Seoul, South Korea S Supporting Information *
ABSTRACT: Persistent organic pollutants (POPs) are degradation-resistant anthropogenic chemicals that accumulate in the food chain and in adipose tissue, and are among the most hazardous compounds ever synthesized. However, their toxic mechanisms are still undefined. To investigate whether characteristic molecular signatures can discriminate individual POP and provide prediction markers for the early detection of POPs exposure in an animal model, we performed transcriptomic analysis of rat liver tissues after exposure to POPs. The six different POPs (toxaphene, hexachlorobenzene, chlordane, mirex, dieldrin, and heptachlor) were administered to 11-week-old male Sprague−Dawley rats, and after 48 h of exposure, RNAs were extracted from liver tissues and subjected to rat whole genome expression microarrays. Early during exposure, conventional toxicological analysis including changes in the body and organ weight, histopathological examination, and blood biochemical analysis did not reflect any toxicant stresses. However, unsupervised gene expression analysis of rat liver tissues revealed in a characteristic molecular signature for each toxicant, and supervised analysis identified 2708 outlier genes that discerned the POPs exposure group from the vehicle-treated control. Combination analysis of two different multiclassifications suggested 384 genes as early detection markers for predicting each POP exposure with 100% accuracy. The data from large-scale gene expression analysis of a different POP exposure in rat model suggest that characteristic expression profiles exist in liver hepatic cells and multiclassification of POP-specific molecular signatures can discriminate each toxicant at an early exposure time. The use of these molecular markers may be more widely implemented in combination with more traditional techniques for assessment and prediction of toxicity exposure to POPs from an environmental aspect.
■
INTRODUCTION Persistent organic pollutants (POPs) are highly lipophilic chemicals that bioaccumulate in animal and human fats, and are residues created through industrial activities.1 Once released into various environmental compartments such as air, water, soil, sediment, and food, they can be dispersed on global scale and pose serious health and environmental risks. Because of chlorine, bromine, or fluoride groups on the hydrocarbon rings or chains, these substances are resistant to degradation both in the environment and in the human body.2 Thus, POPs have become some of the most important global environmental contaminants.3 © 2012 American Chemical Society
Many clinical, epidemiologic, and cell culture studies suggest that exposure to certain POPs might cause various adverse effects including biochemical deviations, carcinogenesis, neurological changes, reproductive abnormalities, behavioral abnormalities, and dysfunctions of the immunological network.4−6 The risks posed by POPs, especially organochlorine pesticides (OCPs) have become of increasing concern in international Received: Revised: Accepted: Published: 12882
June 20, 2012 November 7, 2012 November 15, 2012 November 15, 2012 dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
toxicants.20 The objective of this study is to identify the characteristic molecular signatures for hepatic effects that allow prediction of exposure to six different POPs (TO, HCB, CH, MI, DI, and HE) at a very early time during exposure by using whole genome expression analysis in rat liver tissue.
communities, given their widespread use as insecticides, herbicides (i.e., weed killers), and fungicides. OCPs (pesticide-POPs) including aldrin, chlordane (CH), dichlorodiphenyltrichloroethane (DDT), dieldrin (DI), endrin, heptachlor (HE), hexachlorobenzene (HCB), mirex (MI), and toxaphene (TO) were identified in the Stockholm Convention.7,8 Of these, TO is a complex mixture of polychlorinated camphenes and bornanes, which was used as a pesticide in the agriculture and fishery sectors.9 TO can promote neurotoxicity, nephrotoxicity, hepatotoxicity, and endocrine toxicity.10 HCB accumulates in food chains and lipid rich tissues. Exposure of laboratory animals to HCB can elicit a variety of toxic effects including hepatic porphyria, reproductive dysfunctions, immunomodulation, and carcinogenicity.11 CH is a mixture of four main isomers (heptachlor, cis-chlordane, trans-chlordane, and trans-nonachlor) that are widely produced for agricultural and residential uses, and termite control.12 The insecticides CD and HE in human body fat and plasma are associated with carcinogenic activity in humans and experimental animals.13 According to the Agency for Toxic Substances and Disease Registry (ATSDR), MI is used as a pesticide to control insects and as a flame retardant additive in plastics, paper, and electrical goods. MI is a potent tumor promoter in laboratory animals, producing neoplastic nodules.14 DI has also been widely used agriculturally to control soil pests such as termites, grasshoppers, locusts, and beetles. DI exposure is positively associated with an increased incidence of Parkinson’s disease.15 Because of their variety of adverse effects, TO, HCB, CH, MI, and HE are classified by the International Agency for Research on Cancer (IARC) as group B2 carcinogens in humans. By contrast, DI is considered as a group 3 compound (“not classifiable as to carcinogenicity to humans”). Although POP-induced adverse effects have been proposed, the specialized effects of these chemicals remain to be elucidated in biological systems. Therefore, the toxicity profile of these exposures on a population-wide basis and the discovery of predictive markers associated with various POPs should be observed and verified. Recently, powerful functional genomic-based methods have been developed that provide new, mechanism-based assays for predicting toxic risks to humans. The expression genomic tool using the DNA microarray technology, developed for the concurrent analysis of a large number of genes, may be useful for risk assessment of toxic potency of various environmental toxicants.16 Therefore, to better understand the adverse effects of environmental toxic compounds for predicting the potential human health risk, many researchers initially focused on using gene expression alterations through microarray analysis. This approach has been dubbed “toxicogenomics”.17 Toxicogenomic analysis can be used to predict toxicological effects of compounds based on the similarities of the gene expression profiles, and its use has increased as a consequence of risks and effects on human health from exposure to environmental contaminants.18 The changes in gene expression associated with toxicants are more sensitive and occur before changes in clinical chemistry, histopathology, or clinical observations. Therefore, gene expression data might provide a window of opportunity and an early marker for preclinical diagnosis of possible toxic end point.19 The liver is one of the most important organs, because of its biological functions such as drug metabolism, amino acid metabolism, and lipid metabolism. Therefore, the liver is considered as the common target organ for environmental
■
EXPERIMENTAL SECTION Animals and Drug Treatments. Male Sprague−Dawley (SD) rats approximately 10−11-weeks-old were obtained from Orient (Seoul, Korea). Animals were housed in a controlled environment in the institutional animal facility with free access to water and pellet diet. The animal facility was operated at 21− 25 °C, a relative humidity of 50−70% with 12 h light/12 h dark cycle. During the study period, all animals were observed at least twice daily for mortality and morbidity. The experiments were performed in accordance with Institutional Animal Care and Use Committee guidelines. All animals were randomly assigned into treatment and control groups. TO (CAS Number: 8001−35−2), HCB (CAS Number: 118−74−1), CH (CAS Number: 12789−03−6), MI (CAS Number: 2385− 85−5), DI (CAS Number: 60−57−1), and HE (CAS Number: 76−44−8) were purchased from Supelco (Bellefonte, PA) and prepared following manufacturer’s protocol. Chemical toxic doses were selected at 20% (LD10) of the published LD50 for MI, DI, HE, TO, HCB, and CH, respectively. Rats were administered six POPs at aforementioned doses. For POPs exposure, each treatment group was following: seven animals for HCB, TO, and CH, six animals for MI, DI, and HE, respectively, but five of DI were subject for microarray study due to poor quality of RNA. For control group, 10 animals were subject for both general toxicity assays and microarray experiments. Rats were sacrificed and total RNA from rat liver were extracted at 48 h after oral administration of TO (10 mg/ kg), HCB (700 mg/kg), CH (50 mg/kg), MI (47 mg/kg), DI (7.8 mg/kg), and HE (8 mg/kg) with vehicle (corn oil). Blood Biochemistry. Rat whole blood for clinical chemistry measurements was collected at 2 days after administration. POP-induced serum levels of glutamic oxaloacetic transaminase (GOT), glutamic pyruvic transaminase (GPT), and total bilirubin (TBIL) as hepatic enzyme markers were determined using standard blood biochemistry techniques. Whole-Genome Rat Oligoarray and Formulation. High-density rat whole-genome oligo nucleotide microarrays were manufactured at the array core facility of the Functional RNomics Research Center, The Catholic University of Korea. Array-Ready Oligo Set (AROS) oligosets were purchased from OPERON (Huntsville, AL). This set contains a total of 26 963 probes targeting rat genes. Fifty-, 60-, or 70 mers of synthesized oligos were robotically printed and processed.21 RNA Isolation, Microarray Hybridization, and Data Acquisition. RNA was isolated using TRIzol (Invitrogen, Carlsbad, CA), and quality control was performed using RNA StdSens Chips on an Experion bioanalyzer (Bio-Rad Laboratories, Hercules, CA). Then, the sample with poor RNA quality was excluded from microarray experiment. Universal Rat Reference RNA (RRR; Stratagene, La Jolla, CA) was used as reference RNA for the microarrays. Twenty micrograms of total RNA was used to prepare the DNA targets, as previously described.22 The reference RNA was labeled with Cyanine 3dUTP (NEN Life Science, Boston, MA), and test samples were labeled with Cyanine 5-dUTP (NEN Life Science). In brief, cDNA targets were generated by reverse-transcription from 12883
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
addition, serum biochemical analyses indicated that POP exposure to rats for 48 h did not cause liver damage (see SI Table S2). These results suggested that general toxicity assessment for the early exposure to POPs did not provide sufficient information for the early prediction or detection of POS exposure in rat model. Thus, for early detection of environmental toxicant damage, it was evidently necessary to configure molecular substances capable of direct response to these stresses and dynamic change at the intracellular level. Characteristic Gene Expression Profiling Analyses of POPs Provides Molecular Signature Discerning Each Toxicant in Rat Liver. A previous result indicated that the characteristic molecular signatures of toxicants in liver tissue could discriminate the molecular basis of hepatotoxicity at 72 h of early exposure.26 This suggested that the transcriptomic response to chemical exposure was specific to the toxicant, and characteristic molecular signatures could be used as early prognostic biomarkers. This toxicogenomic expression approach has been proven to be useful for predicting hepatotoxicity early during exposure. Therefore, we adopted this approach to identify a characteristic molecular signature for early prediction of POP environmental toxicants using a rat model. Initially it was investigated whether characteristic molecular profiles for POPs could be detected in liver tissues. Rats were administered six POPs at 20% (LD10) of the published LD50 doses. After 48 h of oral administration, total RNAs were extracted from liver tissues and subjected to DNA microarrays containing 26 963 genetic elements. In general, transcriptomic regulation occurs through dynamic intracellular responses to multiple biological impacts, such as physiologic or pathophysiologic stresses. Thus, characteristic gene expression profiles reflect these factors. To investigate whether a large-scale gene expression profile in liver tissues reflects different exposure to POPs, we performed unsupervised hierarchical clustering analysis. Among 26 963 genetic elements, 5366 genetic elements that passed minimal filtering criteria (% of data present ≥100) resulted in whole transcriptome of rat liver. As shown in Figure 1, hierarchical clustering analysis of these gene sets resulted in two main clusters in dendrogram. Each cluster was then subclustered by different gene expression profiling on dendrogram. Within dendrogram of clustering analysis, the left cluster included MI-, HE-, and DI-treated groups, and the right cluster included TO-, CH-, and HCB-treated groups. No individual POP-treated samples were mixed with vehicletreated control group and other subclusters (Figure 1B). This result suggested that a characteristic molecular signature for each POP occurs and can be detected in rat liver transcriptome, and that this characteristic molecular signature could discern each POP based on molecular classification of gene expression data. Based on these results, we next sought to identify outlier genes that discriminated POPs from vehicle-treated controls. Supervised analysis of the Welch’s t test (P < 0.005) algorithm resulted in 2708 outlier genes (see SI Figure S1A) that exactly disunited into two groups based on expression profiles (see SI Figure S1B). These POPs outlier genes were further validated by prediction confidence analysis by using leave-one-out cross validation (LOOCV). A summary of the frequencies of class assignments using high-accuracy classifier (2708 genes) is provided in SI Figure S1C, with 100% of prediction for two classes (control and POPs).
total RNA in the presence of Cy-3 or Cy-5 deoxyuridine 5′triphosphate using SuperScipt II enzyme (Invitrogen). Residual dye was removed using a Microcon YM-30 column (Millipore, Billerica, MA). Cy-5-labeled cDNA targets were hybridized with Cy-3-labeled reference in 32 ul of hybridization solution consisting of 50% DIG Easy-Hyb (ROCHE, Basel, Switzerland) and 40 ug herring sperm DNA for 18 h at 42 °C in a humidified conventional hybridization chamber. After hybridization, washing was carried out as follows: (1) 2X SSC, 0.1% sodium dodecyl sulfate at room temperature (RT) for 2 min, (2) 1X SSC at RT for 2 min, (3) 0.2X SSC at RT for 2 min, (4) 0.05X SSC, and (5) distilled water at RT for 2 min. Washed slides were scanned using a GenePix 4000B scanner (Axon Instruments, Union City, CA), and Cy-3/Cy-5-signals were measured using GenePix Pro 5.0 microarray analysis software (Axon Instruments). Scanning and Data Analysis. Arrays with hybridized targets were scanned using an Axon scanner and scanned images were analyzed using GenePix Pro 5.0 software. Spots of poor quality, as determined by visual inspection, were excluded from further analysis. Data collected from each array was submitted to GenPlex software (Istech, Seoul, Korea). Flagged spots and spots with a signal-to-noise Ratio (SNR) < 1.5 were excluded from the analysis, unless specified otherwise. Signal intensities within each array were normalized using the LOWESS algorithm (0.2 of data fraction and iteration No. 3) according to 48 pin groups and variances between the arrays were normalized. Then, we used a data set of genes that satisfied the filtering criteria (genes having more than 75% of log-transformed ratio values presenting in across all arrays). Hierarchical clustering of log ratios was performed using Cluster and TreeView 2.3 (Stanford University, Palo Alto, CA). Pearson’s correlation, Manhattan distance, mean centering, and complete linkage were applied during all clustering applications. Molecular Pathway Mining and Gene Set Enrichment Analysis. To identify molecular pathways associated with differentially expressed genes exposed to POPs, we used the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database (http://www.ncbi.nih.gov/geo/). To investigate underlying mechanisms of each tested PAH, gene set enrichment analysis (GSEA) was performed with deregulated genes from the microarray gene expression data.23 We used the Gene Ontology Biological Process (http://www.broadinstitute. org/gsea) as our gene set database.
■
RESULTS General Toxicology Assessment Does Not Reflect POP Exposure and Discerns POPs from Non-Treated Control at the Early Exposure Time Point. Previous studies on predictive and discernible molecular markers for environmental toxicants24,25 led us to consider whether the characteristic expression signature of each POP offered a basis for predictive profiles of unforeseen toxic compounds to humans. To this end six different POPs (TO, HCB, CH, MI, DI, and HE) were selected as most representative POPS and were administered to SD rats. After 48 h, liver, kidney, lung, and whole blood were collected. Serum samples were subjected to clinical biochemical analysis and then body and organ weights were measured for general toxicology assessment of POPs at early exposure. As shown in Table S1 of the Supporting Information (SI), POP exposure to rats for 48 h did not affect whole body weight change or changes in liver, lung, and kidney weight compared to vehicle-treated normal controls. In 12884
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
Identification of Predictive and Detective Molecular Markers for Environmental Toxicant POPs. The molecular markers that could determine and predict individual toxicants from POPs were then identified. To retrieve the maximum number of genes that precisely discriminated each POP, two different multiclassification algorithms (Volcano plot method and Kruscal-Wallis H-test) were used in a commercial analytical program,24 followed by whole computation [gene selection algorithm, the ratio of between group to within-group sums of squares (BSS/WSS)] to find the genes that classified each toxicant with 100% accuracy. Finally, the outlier genes from these two methods were combined (Figure 2A). Briefly, to identify large-scale gene expression changes by each toxicant exposure (each POP), we applied the Volcano plot method with high stringent cut off value (P < 0.001 and 2-fold change, 360 genes; GenePlex, package tool, Istech) to each toxicant. We next queried for those genes that successfully classified seven different groups (including control) using the Kruskal−Wallis H-test (P = 0.0001, 340 genes). This resulted in maximum number of genes that satisfied with 100% prediction accuracy using LOOCV. By a combination of these outliers, 384 genes were identified as the maximum number of genes that determined each toxicant within the POPs exposure group (Figure 2A). Hierarchical clustering of 384 gene expression could discriminate all six different POPs on dendrogram by its own expression profile (Figure 2B; see SI Table S4 for 384 gene list). These 384 outlier genes were further validated by prediction confidence analysis, classified in 100 random partitions by LOOCV. A summary of the frequencies of class
Figure 1. Unsupervised hierarchical clustering analysis of expression profiling in rat liver tissues after administration of POPs. (A) Variation in expression of 5366 genes selected on the basis of the minimal filtering criteria in six different toxicants. The data are presented after two-dimensional hierarchical clustering, which organizes transcripts and treatments on the basis of similarity. Each row represents a single transcript and each column an experimental treatment. The red and green indicate increased and decreased expression relative to the control, respectively. The color saturation reflects the difference in expression between the sample liver tissue RNA and the common reference RNA. (B) The dendrogram derived from clustering using 5 366 genes set between experimental treatments.
Figure 2. Identification of predictable and discernible gene-based biomarkers for each different POP toxicant. Differential gene expression profiles were selected as genes that were significantly deregulated by the six different toxicants. (A) Analysis method for the identification of minimum discernible molecular biomarkers. (B) Characteristic expression profiles of the 384 predictive molecular biomarkers. (C) Accuracy test using the leave-one-out cross validation method. All POPs-treated groups were categorized into seven different classes. 12885
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
■
assignments using high-accuracy classifier (384 genes) is provided in Figure 2C with 100% of prediction for each class. Biological Process of Predictive Molecular Markers That Are Differentially Expressed in Rat Liver by POPs. To gain comprehensive insight into the molecular basis of predictive molecular markers that are differentially expressed in rat liver by POPs, we identified the molecular mechanisms in which these 384 genes participated using the Kyoto Encyclopedia of Genes and the Genomes (KEGG) pathway database (http://www.genome.jp/kegg/) pathway mining tool. KEGG pathway mapping is the process to map molecular data sets, especially large-scale data sets in genomics. The results of the biological processes are given using bar graph in Figure 3
Article
DISCUSSION
Exposure to environmental pollutants is a major health risk globally. The list of potentially toxic compounds to which humans are exposed continues to grow. Associations between environmental toxic compounds, such as volatile organic compounds (VOCs), and POPs including polycyclic aromatic hydrocarbons (PAHs) and health outcome are difficult to define, given their complexity and often poor characterization.27 Especially, despite international agreements intended to limit the release of POPs, these compounds still persist in the environment and food chain,28 so that most human populations are still exposed to POPs through consumption of fatcontaining foods such as fish, dairy products, and meat.29 Augmented awareness of environmental compounds and the potential human health risk posed by chemical stressors remains a major challenge for toxicologists. However, it is hard to predict the toxic risks in humans. Predicting potential human health risks by the use of microarrays that measure the responses of toxicologically relevant genes and identifying selective and sensitive biomarkers of toxicity are major challenges for toxicologists in predicting and discovery toxicology. Microarray-based predictive toxicology may allow screening and prediction of safety evaluations due to chemical toxic effects and environmental compounds toxicology in the environment and the human health. We previously reported that transcriptomic responses of circulating rat blood cells reflect the exposure to environmental toxicants, such as VOCs and PAHs, and the characteristic molecular signatures can discriminate and predict the type of toxicant at an early exposure time.24,25 These findings suggest that the blood expression signature could be used as a predictable and discernible surrogate marker for detection and prediction of environmental toxicants, and that the use of these molecular markers may be more widely implemented in combination with more traditional techniques for assessment and prediction of toxicity exposure to VOCs or PAHs from an environmental aspect. However, the ultimate goals of predicting environment toxicology are to accurately and selectively predict exposure of environment toxicants at a very early time of exposure in relevant concentrations of environment toxicants from the human body. To these ends, detection of the transcriptomic response to toxicants from select organ tissues is advantageous, since this will provide more relevant information to human systems. In general, the liver is involved in the breakdown of carbohydrates, protein, and fats, splitting them into components the body can use with a series of chemical reactions. The liver plays a critical role in transforming and dismantling chemicals, and is very susceptible to toxicity from these agents. In a previous report, we demonstrated that the characteristic molecular signatures of toxicants in liver tissue could discriminate the molecular basis of hepatotoxicity at 72 h of early exposure.26 This suggested that the transcriptomic response to chemical exposure is specific to the toxicant, and the characteristic molecular signatures could be used as early prognostic biomarkers. This expression toxicogenomic approach has proven to be useful for predictive hepatotoxicity at an early exposure time. Thus, gene expression profiling from liver exposed to POPs that examines critical, toxicologically relevant gene and signal response pathways has the potential to improve the risk assessment and safety evaluation of POPs.
Figure 3. Molecular function category through functional classification of the 384 outlier genes using KEGG pathway database.
and gene list in Table S3 of the SI. From KEGG pathway mapping, we identified 13 pathways that contain at least four gene elements mapped in pathways from 384 molecular markers. As shown in Figure 3, the major signaling pathways affected in the rat liver exposed to POPs were identified as focal adhesion, peroxisome proliferator-activated receptor signaling pathway, cytokine-cytokine receptor interaction, metabolism of xenobiotics by cytochrome P450, fatty acid metabolism, oxidative phosphorylation, androgen and estrogen metabolism, arachidonic acid metabolism, renal cell carcinoma, porphyrin and chlorophyll metabolism, complement and coagulation cascades, Alzheimer’s disease, and ribosome pathway. In addition, we performed gene set enrichment analysis (GSEA) from the deregulated genes by POPs to dissect signaling pathways that are enriched by POPs in rat liver. From this analysis, it was found that Chang_Liver_Cancer_Subclass_Proliferation_Up signaling pathway resulted in commonly activated and Subsrate_Specific_Transmembrane_Transporter_Activity pathway was commonly down-regulated by MI, DI and HE (Figure 4 and SI Table S5). Additionally, we were able to identify the each POP-specific enriched signaling pathway that selectively up-regulated or down-regulated in POP-treated liver tissues (see SI Figures S2−S7 and Tables S6−S11). These results support our finding that POP may affect on the liver cell proliferation and/or various signaling cascades in liver metabolism. Further studies on these signaling pathways will provide information how POP-specific signatures relate to specific biochemical pathways or how they differ or common in terms of the impact on adverse biological effects in an environment. 12886
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
Figure 4. Gene set enrichment analysis (GSEA) of the POPs-response gene set in a ranked list of differentially expressed genes in POPs exposure. (A-F) GSEA enrichment plots and corresponding heat map images of the Chiang liver cancer subclass proliferation up gene set and Substrate specific transmembrane transporter activity gene set in each POP exposures (MI, DI, HE) versus healthy control are shown, respectively. Genes in heat maps are shown in rows; the sample is shown in one column. Expression level is represented as a gradient from high (red) to low (blue).
The present study is, to the best of our knowledge, the first approach to the comparison of genome-wide expressions by direct exposure to six different POPs in rat liver. Exposure to these POPs elicited transcriptomic changes, which could be used to identify characteristic molecular signatures for each POP in liver tissues. Unsupervised analysis of differential gene expression of POPs revealed that liver transcriptome reflected POP exposure at an early exposure time and could be the basis of early detection or prediction molecular markers for exposure to POPs (Figure 1). Although conventional toxicologic analysis including blood biochemistry and body weight changes did not provide information for the exposure of POPs at this early time of exposure (48 h), transcriptomic responses provided numerous data points to reflect these chemical stresses to rat liver tissue. In addition, simple nonparametric analysis resulted in 2708 gene signatures that could discern early exposure of POPs in the rat model system (see SI Figure S1). These POPs are prevalent among environmental contaminants because they are resistant to common modes of chemical, biological, or photolytic degradation, and cause adverse effects to human circulatory system.30 Thus, 2708 outlier genes may represent a common epigenetic responses associated with POPs-specific stress in rat liver.
Evidence of the benefits of microarrays in toxicology to estimate and predict toxicity is steadily accumulating. Predictive toxicology relies mainly on class prediction, the methods of which are based on the postulation that gene expression profiles of known toxicants from representative toxicologic classes are able to predict the toxicologic effects of compounds based on the similarities between gene expression profiles.18 The measurement of gene expression levels upon exposure to a chemical can both provide information about the mechanism of action of toxicants and form a sort of “genetic signature” from the pattern of gene expression changes it elicits both in vitro31 and in vivo.32 To explore this, we designed two different serial analyses of multiclassification, and combined them to robustly augment the prediction potential of each POP. The combination of Kruscal-Wallis H-test and Volcano plot analysis resulted in 384 genes that could exactly annotate seven different classes, including the control group, with 100% accuracy (Figure 2 and SI Table S4). This result suggests that transcriptomic responses of hepatic liver tissues could be very specific to each environmental toxicant, and each toxicantspecific molecular signature may contribute to provide information on developing predictive toxicogenomics. Furthermore, as we expected, underlying molecular pathways in the 384 genetic elements included the focal adhesion, the 12887
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
consequences of DDT use. Environ. Health Perspect. 2009, 117 (9), 1359−67. (7) El-Shahawi, M. S.; Hamza, A.; Bashammakh, A. S.; Al-Saggaf, W. T. An overview on the accumulation, distribution, transformations, toxicity and analytical methods for the monitoring of persistent organic pollutants. Talanta 2010, 80 (5), 1587−97. (8) Tsai, W. T. Current status and regulatory aspects of pesticides considered to be persistent organic pollutants (POPs) in Taiwan. Int. J. Environ. Res. Public Health 2010, 7 (10), 3615−27. (9) Saleh, M. A. Toxaphene: Chemistry, biochemistry, toxicity and environmental fate. Rev. Environ. Contam. Toxicol. 1991, 118, 1−85. (10) de Geus, H. J.; Besselink, H.; Brouwer, A.; Klungsoyr, J.; McHugh, B.; Nixon, E.; Rimkus, G. G.; Wester, P. G.; de Boer, J. Environmental occurrence, analysis, and toxicology of toxaphene compounds. Environ. Health Perspect. 1999, 107 (Suppl 1), 115−44. (11) Giribaldi, L.; Chiappini, F.; Pontillo, C.; Randi, A. S.; Kleiman de Pisarev, D. L.; Alvarez, L. Hexachlorobenzene induces deregulation of cellular growth in rat liver. Toxicology 2011, 289 (1), 19−27. (12) Manar, R.; Vasseur, P.; Bessi, H. Chronic toxicity of chlordane to Daphnia magna and Ceriodaphnia dubia: A comparative study. Environ. Toxicol. 2012, 27 (2), 90−7. (13) Cassidy, R. A. Cancer and chlordane-treated homes: A pinch of prevention is worth a pound of cure. Leuk. Lymphoma 2010, 51 (7), 1363−4. (14) Porter, K. L.; Chanda, S.; Wang, H. Q.; Gaido, K. W.; Smart, R. C.; Robinette, C. L. 17beta-estradiol is a hormonal regulator of mirex tumor promotion sensitivity in mice. Toxicol. Sci. 2002, 69 (1), 42−8. (15) Corrigan, F. M.; Wienburg, C. L.; Shore, R. F.; Daniel, S. E.; Mann, D. Organochlorine insecticides in substantia nigra in Parkinson’s disease. J. Toxicol. Environ. Health A 2000, 59 (4), 229−34. (16) Duggan, D. J.; Bittner, M.; Chen, Y.; Meltzer, P.; Trent, J. M. Expression profiling using cDNA microarrays. Nat. Genet. 1999, 21 (1 Suppl), 10−4. (17) Nuwaysir, E. F.; Bittner, M.; Trent, J.; Barrett, J. C.; Afshari, C. A. Microarrays and toxicology: The advent of toxicogenomics. Mol. Carcinog. 1999, 24 (3), 153−9. (18) Maggioli, J.; Hoover, A.; Weng, L. Toxicogenomic analysis methods for predictive toxicology. J. Pharmacol. Toxicol. Methods 2006, 53 (1), 31−7. (19) Ulrich, R.; Friend, S. H. Toxicogenomics and drug discovery: Will new technologies help us produce better drugs? Nat. Rev. Drug Discov 2002, 1 (1), 84−8. (20) Kaplowitz, N. Drug-induced liver disorders: iImplications for drug development and regulation. Drug Saf. 2001, 24 (7), 483−90. (21) Park, J. Y.; Kim, S. Y.; Lee, J. H.; Song, J.; Noh, J. H.; Lee, S. H.; Park, W. S.; Yoo, N. J.; Lee, J. Y.; Nam, S. W. Application of amplified RNA and evaluation of cRNA targets for spotted-oligonucleotide microarray. Biochem. Biophys. Res. Commun. 2004, 325 (4), 1346−52. (22) Nam, S. W.; Park, J. Y.; Ramasamy, A.; Shevade, S.; Islam, A.; Long, P. M.; Park, C. K.; Park, S. E.; Kim, S. Y.; Lee, S. H.; Park, W. S.; Yoo, N. J.; Liu, E. T.; Miller, L. D.; Lee, J. Y. Molecular changes from dysplastic nodule to hepatocellular carcinoma through gene expression profiling. Hepatology 2005, 42 (4), 809−18. (23) Subramanian, A.; Tamayo, P.; Mootha, V. K.; Mukherjee, S.; Ebert, B. L.; Gillette, M. A.; Paulovich, A.; Pomeroy, S. L.; Golub, T. R.; Lander, E. S.; Mesirov, J. P. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U.S.A. 2005, 102 (43), 15545−50. (24) Jung, K. H.; Noh, J. H.; Eun, J. W.; Kim, J. K.; Bae, H. J.; Xie, H.; Jang, J. J.; Ryu, J. C.; Park, W. S.; Lee, J. Y.; Nam, S. W. Molecular signature for early detection and prediction of polycyclic aromatic hydrocarbons in peripheral blood. Environ. Sci. Technol. 2011, 45 (1), 300−6. (25) Kim, J. K.; Jung, K. H.; Noh, J. H.; Eun, J. W.; Bae, H. J.; Xie, H. J.; Jang, J. J.; Ryu, J. C.; Park, W. S.; Lee, J. Y.; Nam, S. W. Identification of characteristic molecular signature for volatile organic compounds in peripheral blood of rat. Toxicol. Appl. Pharmacol. 2011, 250 (2), 162−9.
peroxisome proliferator-activated receptor signaling pathway, the cytokine-cytokine receptor interaction, and the metabolism of xenobiotics by cytochrome P450, which are very well-known intracellular signaling pathways in liver hepatocytes (Figure 3). In addition, GSEA analysis provided us signaling pathways that highly enriched by tested-POPs as underlying molecular mechanisms that these POPs induce their toxic responses or as gene signature defining their mode of action. (Figure 4 and SI Figures S2−S7). In conclusion, characteristic molecular signatures for POPs exist in liver hepatic cells, and multiclassification of this POPspecific signature could discriminate each toxicant at an early exposure time. These expression signatures could be used as predictable and discernible surrogate markers for the detection of biological responses to environmental POP exposure. Further work on these gene signatures is needed to provide a solid mechanistic insight into the toxicological effects on the human system.
■
ASSOCIATED CONTENT
S Supporting Information *
Summary of the changes in body and organ weight, biochemical analysis of blood serum, and the lists of differentially expressed genes and enriched gene sets by POPs exposure in rat liver are provided as Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Phone: 82-2-2258-7314; fax: 82-2-537-6586; E-mail: swnam@ catholic.ac.kr. Author Contributions §
These authors contributed equally to this study.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Korean Ministry of the Environment “The Converging-Technology Project” (Grant No. 212 101 003), and by the Korean Science and Engineering Foundation via the “Cancer Evolution Research Center” at The Catholic University of Korea (Grant No. 2012047939).
■
REFERENCES
(1) Mullerova, D.; Kopecky, J. White adipose tissue: Storage and effector site for environmental pollutants. Physiol. Res. 2007, 56 (4), 375−81. (2) Yu, G. W.; Laseter, J.; Mylander, C. Persistent organic pollutants in serum and several different fat compartments in humans. J. Environ. Public Health 2011, 2011, 417980. (3) Jones, K. C.; de Voogt, P. Persistent organic pollutants (POPs): State of the science. Environ. Pollut. 1999, 100 (1−3), 209−21. (4) Porta, M.; Puigdomenech, E.; Ballester, F.; Selva, J.; Ribas-Fito, N.; Llop, S.; Lopez, T. Monitoring concentrations of persistent organic pollutants in the general population: The international experience. Environ Int 2008, 34 (4), 546−61. (5) Carpenter, D. O. Polychlorinated biphenyls (PCBs): Routes of exposure and effects on human health. Rev. Environ. Health 2006, 21 (1), 1−23. (6) Eskenazi, B.; Chevrier, J.; Rosas, L. G.; Anderson, H. A.; Bornman, M. S.; Bouwman, H.; Chen, A.; Cohn, B. A.; de Jager, C.; Henshel, D. S.; Leipzig, F.; Leipzig, J. S.; Lorenz, E. C.; Snedeker, S. M.; Stapleton, D. The Pine River statement: Human health 12888
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
Environmental Science & Technology
Article
(26) Eun, J. W.; Ryu, S. Y.; Noh, J. H.; Lee, M. J.; Jang, J. J.; Ryu, J. C.; Jung, K. H.; Kim, J. K.; Bae, H. J.; Xie, H.; Kim, S. Y.; Lee, S. H.; Park, W. S.; Yoo, N. J.; Lee, J. Y.; Nam, S. W. Discriminating the molecular basis of hepatotoxicity using the large-scale characteristic molecular signatures of toxicants by expression profiling analysis. Toxicology 2008, 249 (2−3), 176−83. (27) Briggs, D. Environmental pollution and the global burden of disease. Br. Med. Bull. 2003, 68, 1−24. (28) Ruzzin, J.; Petersen, R.; Meugnier, E.; Madsen, L.; Lock, E. J.; Lillefosse, H.; Ma, T.; Pesenti, S.; Sonne, S. B.; Marstrand, T. T.; Malde, M. K.; Du, Z. Y.; Chavey, C.; Fajas, L.; Lundebye, A. K.; Brand, C. L.; Vidal, H.; Kristiansen, K.; Froyland, L. Persistent organic pollutant exposure leads to insulin resistance syndrome. Environ Health Perspect 2010, 118 (4), 465−71. (29) Fisher, B. E. Most unwanted. Environ Health Perspect 1999, 107 (1), A18−23. (30) Casals-Casas, C.; Desvergne, B. Endocrine disruptors: From endocrine to metabolic disruption. Ann. Rev. Physiol. 2011, 73, 135− 62. (31) Waring, J. F.; Ciurlionis, R.; Jolly, R. A.; Heindel, M.; Ulrich, R. G. Microarray analysis of hepatotoxins in vitro reveals a correlation between gene expression profiles and mechanisms of toxicity. Toxicol. Lett. 2001, 120 (1−3), 359−68. (32) Hamadeh, H. K.; Bushel, P. R.; Jayadev, S.; Martin, K.; DiSorbo, O.; Sieber, S.; Bennett, L.; Tennant, R.; Stoll, R.; Barrett, J. C.; Blanchard, K.; Paules, R. S.; Afshari, C. A. Gene expression analysis reveals chemical-specific profiles. Toxicol. Sci. 2002, 67 (2), 219−31.
12889
dx.doi.org/10.1021/es302480v | Environ. Sci. Technol. 2012, 46, 12882−12889
