Table of Contents
Advances in Toxicology
Volume 2014, Article ID 984319, 13 pages
http://dx.doi.org/10.1155/2014/984319
Research Article

Chemical Exposure Generates DNA Copy Number Variants and Impacts Gene Expression

1School of Health Sciences, Purdue University, 550 Stadium Mall Drive, HAMP-1263D, West Lafayette, IN 47907, USA
2Institute of Neuroscience, University of Oregon, Eugene, OR 97403, USA

Received 29 August 2014; Revised 3 December 2014; Accepted 3 December 2014; Published 30 December 2014

Academic Editor: Mugimane Manjanatha

Copyright © 2014 Samuel M. Peterson and Jennifer L. Freeman. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

DNA copy number variation is long associated with highly penetrant genomic disorders, but it was not until recently that the widespread occurrence of copy number variation among phenotypically normal individuals was realized as a considerable source of genetic variation. It is also now appreciated that copy number variants (CNVs) play a role in the onset of complex diseases. Many of the complex diseases in which CNVs are associated are reported to be influenced by yet to be identified environmental factors. It is hypothesized that exposure to environmental chemicals generates CNVs and influences disease onset and pathogenesis. In this study a proof of principle experiment was completed with ethyl methanesulfonate (EMS) and cytosine arabinoside (Ara-C) to investigate the generation of CNVs using array comparative genomic hybridization (CGH) and the zebrafish vertebrate model system. Exposure to both chemicals resulted in CNVs. CNVs were detected in similar genomic regions among multiple exposure concentrations with EMS and five CNVs were common among both chemicals. Furthermore, CNVs were correlated to altered gene expression. This study suggests that chemical exposure generates CNVs with impacts on gene expression warranting further investigation of this phenomenon with environmental chemicals.

1. Introduction

Structural genetic variation in the human genome is present in many forms including single nucleotide polymorphisms (SNPs), variable tandem repeats (e.g., mini- and microsatellites), presence/absence of transposable elements, and structural alterations (e.g., deletions, duplications, and inversions). Until recently, SNPs were thought to be the predominant form of genomic variation and to account for much of the normal phenotypic variation [1]. Recent developments and applications of genome-wide technologies led to the discovery of thousands of copy number variants (CNVs) in the genomes of phenotypically normal humans [2, 3]. CNVs are defined as a duplication or deletion (i.e., a gain or loss of a genomic DNA segment relative to a reference sample) measuring greater than 1 kb in size [4]. Human genomic copy number variation has been studied for over 40 years, but it was assumed that CNVs were few in number, had a relatively limited impact on the total amount of human genetic variation, and were mainly associated with highly penetrant disease phenotypes. In 2004, two studies independently reported the widespread presence of CNVs in the genomes of phenotypically normal individuals [2, 3]. Following these initial studies, additional genome-wide analyses identified and characterized novel human CNVs (e.g., [5]). Widespread copy number variation in the human genome is now well documented with many CNVs spanning genes that are likely to affect gene networks [6]. CNVs result in various phenotypic effects including changes in gene expression levels, disruption of gene dosage and regulatory elements, and loss of regulatory elements [7].

Classical cytogenetics identified a variety of genomic variants that are related to disorders that are caused by a single variant (e.g., a deletion on chromosome 7 in Williams Beuren syndrome). CNVs that did not directly result in early-onset, highly penetrant genomic disorders were initially considered neutral in function, but CNVs are now appreciated to play a role in the onset of complex diseases including autism spectrum disorder (ASD), attention-deficit hyperactivity disorder (ADHD), and schizophrenia [8, 9]. In addition, CNVs are reported to influence late-onset diseases (e.g., Alzheimer’s disease and Parkinson’s disease). In addition to genetic factors, these diseases are also implicated to be influenced by environmental factors. The mechanisms by which environmental factors influence onset and pathogenesis of these diseases are not completely understood [10]. Current analysis of functional attributes of CNV regions is revealing enrichment for genes that are relevant to molecular-environmental interactions [3, 5]. Moreover, a study in postmortem brains of individuals with ASD indicates possible involvement of exposure to polychlorinated biphenyls (PCBs) with a duplication event on human chromosome 15 [11]. This study indicates an environmental link with CNVs and influence on complex disease, but it is not known if the copy number alteration was specifically generated by the environmental chemical exposure.

Exposure to environmental chemicals is one environmental factor that may contribute to the formation of CNVs (or copy number aberrations), but the ability of chemical exposure to generate CNVs has not been thoroughly investigated. With the development of genomic technologies including the array comparative genomic hybridization (CGH) technology and NextGen sequencing, copy number alterations are now efficiently detected throughout the genome. Previous assays and techniques applied to investigate the influence of chemical exposure on the genome were limited to detecting single nucleotide mutations or larger chromosomal aberrations (Figure 1). In addition, many of these assays had inefficient integration of structural DNA alterations with the reference genome sequence limiting further studies into the biological and functional significance of these DNA alterations. Thus, this class of DNA alteration was not thoroughly assessed in past genotoxicity studies. Three recent studies began to investigate the generation of CNVs with aphidicolin, hydroxyurea, and ionizing radiation in a cell culture system [1214], but no other agents have been investigated to date.

Figure 1: Toxicity assays interrogating DNA sequence alterations by size. Classic cytogenetic methodologies routinely identify whole genome, whole chromosomal, and microscopic structural chromosomal aberrations. At the opposite end of the spectrum, mutation assays are optimized to detect single base pair mutations. Development of genome-wide technologies including array-based assays (e.g., array CGH) and whole genome sequencing now permit efficient detection of DNA structural alterations of an intermediate size including copy number alterations and enable direct integration with a reference genome sequence. As a result, the ability of chemical exposure to induce this type of DNA alteration was not thoroughly investigated in the past and is just now beginning to be addressed.

The importance of using genomics to identify environmental chemical influence on the human genome is now recognized [15] and the specific influence of CNVs is recognized as an emerging environmental health issue (http://nas-sites.org/emergingscience/meetings/genomic-plasticity/). In this study, a proof of principle experiment using a zebrafish cell line was completed to test the hypothesis that chemical exposure will result in CNVs detectable with the use of array CGH technology to set the stage for future analysis into the influence of environmental chemical exposure in generating CNVs. The zebrafish is a prominent model vertebrate system in a variety of biological disciplines. A finished genome sequence and conserved genetic function between the zebrafish and human genomes permit translation of molecular mechanisms of toxicity observed in the zebrafish model system to humans [16, 17]. Several zebrafish orthologs are reported to play a key role in human disease and large-scale mutant screens demonstrate that mutations in some of these orthologs display phenotypes similar to those present in human diseases [18, 19]. In addition, the zebrafish has been used for many years as a toxicological model [20] and a model for DNA repair mechanisms (e.g., [21]). A CNV map is established for the zebrafish genome and confirms the plasticity of the zebrafish genome permits CNV formation [22] and suitability for application in this study. Two genotoxic chemicals commonly used as reference chemicals, ethyl methanesulfonate (EMS), and cytosine arabinoside (Ara-C) were included in this proof of principle experiment at multiple exposure concentrations to assess dose-response and differences among the two chemicals. In addition, global gene expression analysis was completed to correlate CNVs caused by chemical exposure and alterations in gene expression.

2. Materials and Methods

2.1. Cell Line and Toxicity Assay

A zebrafish fibroblast cell line established from approximately 100 embryos of the AB zebrafish strain that is described in detail in Freeman et al. [23] was used in this study. This zebrafish cell line was used in this proof of principle study since it is well-characterized, is routinely monitored for cytogenetic changes, and has been used in previous zebrafish cytogenetic studies [23]. In addition, use of this zebrafish cell line will provide ease in moving into in vivo studies with in vivo zebrafish in future studies. Two reference chemicals routinely used in genotoxicity assays, ethyl methanesulfonate (EMS; CAS 62-50-0; Sigma, St. Louis, MO) and cytosine arabinoside (Ara-C; CAS 147-94-4; Sigma, St. Louis, MO), were investigated for potential to generate CNVs. A cell confluency assay was first completed to identify the toxicity of EMS and Ara-C in this cell line. This assay is modified from Plewa et al. [24]. Briefly, cells were harvested from cell culture flasks following a standard trypsin protocol and cell concentration determined. The assay was set up in 96-well plates with 7,000 cells per well in an appropriate volume of media and chemical stock to achieve desired chemical test concentrations. Plate set-up included a first column blank and a second column negative control. Each plate contained four subsample wells per chemical concentration. Following set-up, plates were placed in an incubator at 28°C and 5% CO2. After 72 hours, the cells were fixed in 50% methanol and stained with 1% crystal violet in 50% methanol, and excess crystal violet solution was washed from the plate. The cells were then treated with 1% SDS to bring the crystal violet back into solution. The absorbance of each well was read on a microplate reader at 595 nm and readings from the four subsample wells of each test concentration averaged. A percent negative control value was calculated for each test concentration. This value represents the confluency of the cells grown in the presence of the test compound as compared to the unexposed control cells. Three replicate plates were completed and the average percent negative control values of the three replicate plates for each test concentration calculated, plotted, and fit a sigmoidal curve. The 50% and 75% confluency value of the negative control was calculated for each chemical to determine test concentrations for array CGH analysis. These values were chosen to be able to compare if CNVs would be generated from exposure treatments that ranged from 50% impacts on cell confluency to exposure treatments that did not alter cell confluency in the EMS experiment and to then choose exposure treatments that did not impact cell confluency in the Ara-C experiment.

2.2. Copy Number Analysis

A zebrafish-specific oligonucleotide platform was designed and printed in conjunction with Roche NimbleGen (Madison, WI) for this study to analyze copy number changes. The zebrafish oligonucleotide platform was manufactured using Roche NimbleGen’s proprietary Maskless Array Synthesizer technology using photomediated synthesis chemistry. DNA probes were selected using a proprietary probe screening system. -balanced probe selection was coupled with heuristic and Al predictive methods derived from their experimental database. Probe sets were selected to represent the genomic target and to have excellent hybridization characteristics. Specifically for this design, segmental duplications (i.e., regions of the genome with up to 5 close matches) were included as some copy number alterations are reported to be associated with these genomic segments [5] and highly repetitive sequences were excluded. In addition, a number of standards were also included throughout the array. A number of self to self-hybridizations were first conducted with this platform to assess the performance of the array platform and to determine resolution of platform.

Array CGH analysis was performed similarly as described in Peterson and Freeman [25]. Zebrafish cells were exposed to three chemical concentrations of EMS calculated from the cytotoxicity curve to represent (1) a concentration 50% of the negative control value, (2) a concentration 75% of the negative control value, and (3) a concentration where no cytotoxic impacts were observed for a dose-response assessment and a corresponding negative control without chemical exposure. Two exposure concentrations with limited cytotoxicity were included for Ara-C. Cells were harvested from maintenance cultures and cell concentration determined. Appropriate volume of media and chemical stock was added to each petri dish to achieve the desired test concentrations (i.e., 0 mM, 0.5 mM, 2 mM, and 5 mM for EMS and 0 μM, 0.1 μM, and 1 μM for Ara-C). 7.5 × 105 cells were initially seeded into each dish. After set-up, petri dishes were placed in an incubator at 28°C and 5% CO2 for 72 hours (the equivalent of 1.5 cell cycle lengths). After 72 hours, cells were harvested and genomic DNA was isolated following a standard phenol: chloroform isolation method as described in Freeman et al. [26]. Genomic DNA quantity and quality were determined using a NanoDrop ND-1000 and gel electrophoresis. The negative control was the reference sample for each replicate and was cohybridized with each treatment (i.e., the negative control and the three treatment concentrations) using a two-color hybridization strategy. Genomic DNA samples were labeled and subsequently hybridized upon the zebrafish array CGH platform using the protocol outlined in the Roche NimbleGen User’s Guide (Roche NimbleGen, Madison, WI). For each array hybridization, 1 μg of test DNA and 1 μg of reference DNA were fluorescently labeled with dye-labeled 9 mers (i.e., the test DNAs were labeled with Cy3 and the reference DNA was labeled with Cy5). Dye-incorporation and DNA quality and quantity were assessed using a NanoDrop ND-1000 spectrophotometer. Cy3-labeled test DNA and Cy5-labeled reference DNA was combined into one tube for each test concentration and injected into a mixer attached to the array CGH chip as described in the NimbleGen Array User’s Guide. The chip was placed in a bay of the NimbleGen Hybridization System and DNA hybridized for 16 hours at 42°C. Following hybridization, the arrays were washed in solutions supplied in the Roche NimbleGen wash buffer kit followed by spin drying of the slides in a microfuge slide dryer.

Hybridized arrays were scanned using two-color scanning for Cy3 and Cy5 at 5 microns on a GenePix 4000B (Molecular Devices, Sunnyvale, CA). Scans were optimized for Cy3 and Cy5 signal intensities in the same range and for ~1% of the features saturated. Array image data was extracted using the NimbleScan software program (Roche NimbleGen, Madison, WI). The Cy3 and Cy5 signal intensities were normalized to one another using qspline normalization, a simple and robust nonlinear method of normalization for two-color experiments [27]. Normalized signal intensity files were generated by NimbleScan. Internal control probes and overall variation of signal intensity were used to assess the quality of each array CGH experiment. The NimbleScan data was then exported into the Nexus Copy Number software to calculate DNA sequence regions that deviated from the expected 1 : 1 molar ratio of the test to reference DNA (log2 ratio of 0) similar to as reported previously [22]. Called regions represent CNVs as a result of chemical exposure. Genomic locations and magnitude of gain/loss were compared among the experiments and among the chemical treatments. Genomic locations of CNVs were integrated with the zebrafish reference sequence for characterization. Genomic location of CNVs was compared among samples and overlapping segments determined. Two separate experiments were completed for each chemical.

2.3. Global Gene Expression Analysis

To investigate the impacts of CNVs caused by chemical exposure on gene expression, global gene expression analysis was performed with RNA isolated from a 2 mM EMS exposure and a control treatment following similar procedures as described previously [28]. The 2 mM EMS treatment was chosen as an exposure that had a minor effect on confluency of the cells and at which a number of CNVs were detected. Three biological replicates were included that consisted of three separate control samples and three separate samples treated with 2 mM EMS. Microarray analysis was performed similarly to as described in Peterson et al. [28] with the zebrafish 385 K expression platform (Roche NimbleGen, Madison, WI) using the one-color hybridization strategy. As such, six different microarrays were hybridized for this analysis. This platform contains 385,000 60-mer probes interrogating 37,157 targets with up to 12 probes per target. Following hybridization, arrays were washed and scanned at 5 microns using a GenePix 4000B array scanner (Molecular Devices, Sunnyvale, CA). Array image data was extracted using the NimbleScan software program (Roche NimbleGen, Madison, WI). Fluorescence signal intensities were normalized using quantile normalization [29] and gene calls generated using the Robust Multichip Average algorithm [30] following manufacturer recommendations. Further statistical processing of the array data was performed with Array Star (DNASTAR, Inc., Madison, WI) and Ingenuity Pathway Analysis software (Ingenuity Systems, Redwood City, CA) to identify specific genes altered following EMS exposure. A robust and reproducible list of differentially expressed genes using recommendations from the Microarray Quality Consortium [31, 32] was first determined by genes consistently expressed (Students -test, ) and substantially altered with a fold change of ±2.0. Genomic location of genes with altered expression was compared to the genomic location of CNVs and gene ontology analysis and molecular pathway analysis completed using UCSC Genome Browser (http://www.genome.ucsc.edu/) and Ingenuity Pathway Analysis (IPA) software following similar parameters as in previous experiments [28]. All genes were converted and reported as human homologs.

3. Results

3.1. Cell Toxicity

The toxicity of EMS and Ara-C in the zebrafish cell line was first investigated and results were used to determine the exposure concentrations for the array CGH analysis. Test concentrations were calculated at the 50% negative control value, at 75% negative control value, and at a concentration where no impacts on cell confluency were observed for a dose-response assessment in the array CGH analysis for EMS (Figure 2(a)). 5 mM, 2 mM, and 0.5 mM, were chosen as test concentrations for EMS, respectively. In addition, two concentrations with limited impacts to cell confluency were included for Ara-C (0.1 μM and 1 μM; Figure 2(b)).

Figure 2: Cell toxicity assay. The toxicity profile of (a) EMS and (b) Ara-C was first completed in the zebrafish cell line to determine impacts on cell confluency in this specific cell line. From this analysis chemical concentrations were determined for array CGH experiments.
3.2. CNVs following Chemical Exposure

The zebrafish oligonucleotide array CGH platform contains 385,000 probes, approximately 50 to 75 nucleotides in length, tiling the zebrafish genome with a median spacing of ~3.2 kb (Table 1). Four self to self-hybridization experiments were first conducted to assess the performance of the platform and to determine the resolution of the platform. No calls were found to present in these self to self-hybridization experiments greater than 5 consecutive probes in length (~12.8 kb) and the resolution of the platform was estimated at 16 kb (6 consecutive probes in length). All oligonucleotide array-based platforms have some degree of background noise, which varies among each specific platform. As a result there is generally a lack of confidence in single probe calls. Evaluating a series of self to self-hybridizations (in which no copy number alterations should be observed) and confirmatory experiments for calls observed on this platform, we determined that calls containing at least 6 consecutive probes have a high degree of confidence (Figure 3). Using these calling parameters, the number of false positive calls was significantly decreased. In addition, calls had an average segmentation mean ±0.075 or greater in magnitude.

Table 1: Genome coverage of the zebrafish array CGH platform.
Figure 3: Self to self-hybridization assessment. A series of four self to self-hybridizations were conducted to assess the performance of the array CGH platform to determine the number of consecutive probes that are needed to have high confidence in a true call. From these experiments, it was determined that high confidence is attained in calls in which at least six consecutive probes significantly deviate from the expected 1 : 1 molar ratio. As a result, resolution of this platform is approximately 16 kb.

For assessment of EMS, two separate experiments were performed using the two-color hybridization strategy. Genomic DNA from each chemical treatment was cohybridized with the respective negative control as the reference sample. In the first experiment, 5, 17, and 1 CNVs were called in the 0.5, 2, and 5 mM treatments, respectively (Table 2). In the second experiment, 10, 0, and 11 CNVs were called in the 0.5, 2, and 5 mM treatments, respectively (Table 2). In total 44 CNVs were called with a loss in copy number in 28 regions and a gain in copy number in 16 regions. CNVs ranged from 19.8 to 7,069.8 kb in size. The number of CNVs did not increase with dose, but there were 11 CNVs with overlapping genomic locations among the EMS chemical treatments (Table 3). All overlapping CNVs agreed in their respective loss or gain in copy number furthering confidence in these calls. In addition, magnitude of change was also similar among the overlapping CNVs. Overall 8 overlapping CNVs had a loss in copy number, while 3 overlapping CNVs had a gain in copy number. The overlapping genomic segments were refined regions with a size ranging from 33.5 to 1,283.8 kb. Three of the overlapping CNVs were in three different samples including a loss on chromosome 4, a loss on chromosome 5, and a gain on chromosome 14 (Figure 4), while the remaining were present in two samples (Table 3). Considering overlapping CNVs, a total of 29 different copy number variable regions were present among the EMS treatments (Figure 5(a)). When CNVs generated by EMS exposure were compared to CNVs found in the AB strain of zebrafish (the strain from which the cell line was derived) 39% of the CNVs were found to overlap (Table 2).

Table 2: CNVs generated by EMS exposure.
Table 3: Common copy number variable regions among EMS treatments.
Figure 4: CNVs in common genomic locations among EMS treatments. Multiple CNVs mapped to common genomic locations following the EMS treatments including a gain in copy number in all three concentrations on chromosome 14. This overlapping genomic segment was 120.2 kb in size and mapped to base pair region 16,612,394–16,732,619 on chromosome 14. (a) 0.5 mM, (b) 2 mM, and (c) 5 mM EMS treatments.
Figure 5: CNVs generated by chemical exposure. (a) Exposure to EMS resulted in 44 CNVs among all treatments and consisted of 11 CNVs with overlapping genomic locations. Considering genomic location overlap, 29 different copy number variable regions were identified and included both gains (12; green bars) and losses (17; red bars). Length of bar indicates frequency. (b) Ara-C exposure resulted in 18 CNVs among the treatments including 13 gains (green bars) and 5 losses (red bars). There were no overlapping CNVs among the Ara-C treatments.

Similar to the EMS experiments, two separate experiments were completed for Ara-C. In the first experiment, 1 CNV was called in the 0.1 μM treatment, while 2 CNVs were called in the 1 μM treatment. In the second experiment no CNVs were called in the 0.1 μM treatment and 15 CNVs were called in the 1 μM treatment (Table 4). Of the 18 total CNVs, 5 were losses and 13 were gains in copy number and ranged from 28.9 to 1,505.2 kb in size (Figure 5(b)). When CNVs generated by Ara-C exposure were compared to CNVs found in the AB strain of zebrafish 44% of the CNVs were found to overlap (Table 4). There was no overlapping CNVs among the two concentrations or among the two experiments, but 5 CNVs did overlap with CNVs in the EMS experiment (Table 5). The length of these CNVs in the Ara-C and EMS treatments was similar. Two CNV regions on chromosomes 9 and 14 had consistent gains in both chemical treatments. Two CNV regions on chromosomes 5 and 6 had a loss in the EMS treatment and a gain in copy number following Ara-C treatment, while the final CNV region on chromosome 21 had a gain in the EMS treatment and a loss in the Ara-C treatment.

Table 4: CNVs generated by Ara-C exposure.
Table 5: Common copy number variable regions in EMS and Ara-C treatments.
3.3. Comparative Gene Expression Analysis

To elucidate impacts of CNVs on gene expression, global gene expression analysis was completed with the 2 mM EMS treatment. After removal of redundant probes and accounting for gene orthology a total of 1,146 genes were mapped with altered expression. 979 genes were downregulated and 167 genes were upregulated (see Supplementary Table 1 availablein supplementary material online at http://dx.doi.org/10.1155/2014/984319). Gene ontology and pathway analysis with IPA indicated enrichment with genes associated with diseases and disorders, molecular and cellular functions, and physiological system development and function (Table 6).

Table 6: Gene ontology analysis for enrichment of biological function with 2 mM EMS treatment.

59% of CNV regions (10/17 regions) resulted in a direct impact on gene expression for genes mapping within the CNVs. Five of the ten regions contained genes orthologous to human genes (Table 7). Three CNVs were correlated with a single gene, while two CNVs were correlated with two genes. There were 86% positive associations (a copy number gain associated with increased expression or a copy number loss with decreased expression) and 14% negative associations (a gain associated with decreased expression or a loss associated with increased expression).

Table 7: Genes with altered expression mapping within CNVs in the 2 mM EMS treatment.

4. Discussion

Current knowledge on the role of chemical exposure in the generation of CNVs is limited. In this study, we applied a zebrafish array CGH platform to investigate this phenomenon. CNV identification in the zebrafish genome was recently completed and confirmed the plasticity of the zebrafish genome permits CNV formation [22]. In this proof of principle experiment a zebrafish cell line was initially used to investigate CNV generation associated with chemical exposure with the intention to translate these findings in future in vivo studies using this model.

Two genotoxic chemicals, EMS and Ara-C, were tested for their ability to generate CNVs. EMS is routinely used as reference chemical in genotoxicity assays that is reported to directly produce random point mutations in genetic material by direct alkylation and is often used as a chemical mutagen in studies with model organisms (e.g., [33]). Although alkylating agents are thought to primarily generate point mutations, EMS is also reported to cause other genetic alterations including DNA strand breaks [34, 35]. A range of EMS chemical treatments from those that resulted in a 50% decrease in cell confluency to no impacts on cell confluency were included in this experiment. An increase in the number of CNVs was not observed with increasing dose, but CNVs were detected in similar genomic regions among the multiple test concentrations of EMS indicating a potential hotspot of genomic instability and a nonrandom genotoxic mechanism for this chemical. Moreover, 39% of the CNVs generated by EMS exposure overlapped with known CNVs in the genome of the AB strain of zebrafish [22] indicating these regions may be more susceptible to genomic rearrangements than other regions. Each of these experiments was started from the same batch of cells with similar cytogenetic structure to alleviate detection of background CNVs and thus support CNVs observed in this study were due to the chemical exposure.

Ara-C was included as a comparative chemical to assess if CNVs are chemical-specific and to further assess CNVs at concentrations with limited alterations on cell confluency. Ara-C is also often used as a reference chemical in genotoxicity assays, is a chemotherapy agent, interferes with DNA synthesis, and results in chromosomal aberrations [36, 37]. In the Ara-C treatments, a higher number of CNVs were generated at the higher exposure treatment and no overlapping CNVs were detected between the two treatments, but 44% of the CNVs generated by Ara-C exposure overlapped with CNVs in the genome of the AB strain of zebrafish [22]. While a number of chemical-specific CNVs were called, 5 CNVs were generated in similar genomic regions in both the EMS and Ara-C treatments indicating these regions may be more susceptible to genomic alterations in a nonchemical-specific manner. Two of the five regions were consistent in gain and/or loss in copy number in the specific genomic region. These findings are similar to the patterns of CNVs generated by exposure to aphidicolin and hydroxyurea in a study where CNVs were distributed among the genome with some hotspots of formation [13].

Overall a dose-response was observed with the two treatments of Ara-C, but not among the EMS treatments. It is hypothesized that this difference is due to the effects on cell confluency among the two ranges of test concentrations in the EMS versus the Ara-C experiments (i.e., the EMS exposure treatments ranged from those that resulted in 50% cell confluency to no impact on cell confluency, while both of the Ara-C treatments did not impact cell confluency).

To further assess the influence of CNVs generated by chemical exposure, global gene expression analysis was conducted for the 2 mM EMS treatment. Overall gene expression alterations were identified to be associated with developmental disorders, skeletal muscular disorders, infectious diseases, connective tissue disorder, and cardiovascular disease. In addition, alterations were associated with genes involved in cellular movement, amino acid metabolism, small molecular biochemistry, cellular assembly and organization, and cellular function and maintenance. Expression alterations were also enriched for genes associated with tissue morphology, organismal survival, embryonic development, organismal development, and organ morphology. Furthermore, a direct comparison of genomic regions harboring CNVs with the genomic location of genes with altered expression in the 2 mM EMS treatment indicates that CNVs generated by chemical exposure impact gene expression. This analysis identified both direct associations and negative associations. The negative association may be regulatory in nature. Genes with altered expression associated with CNVs include genes involved in SNAP receptor activity (STX16), the initiation of transcription (MED12), the initiation of protein synthesis (EIF2S), acetyl-CoA transport (SLC33A1), the de novo synthesis of purine nucleotides (GMPS), DNA binding (ARID5B), and signal transduction (CAPN5). Genes are also associated with various diseases including a deletion in STX16 with autosomal dominant pseudohypothyroidism [38], a decrease in expression of EIF2S with uveal melanoma [39], a polymorphism in ARID5B with an increased risk of MLL rearrangements in early childhood leukemia [40], and mutations in CAPN5 with autosomal dominant neovascular inflammatory vitreoretinopathy [41]. In addition, ARID5B is essential for adipogenesis and liver development, while CAPN5 plays an important role in developmental processes [42]. Moreover, it is likely the CNVs are also linked to altered expression of other genes as CNVs are reported to have global influence on the transcriptome [43]. It should also be recognized that exposure to the genotoxic chemicals can result in single nucleotide mutations. Thus, some of the detected gene expression changes may be due to single nucleotide mutations and/or other DNA alterations, but overall the data indicates that the CNVs generated by the chemical treatments are likely to contribute to gene expression changes and support further studies in the functional effects of the CNVs generated by the chemical exposures.

The zebrafish genome is reported to have gone through two rounds of whole genome duplication during the course of evolution with a third event occurring before the last teleost radiation [44]. The duplicated genome did lead to some difficulty in mapping of the zebrafish genome compared to the rodent and human genomes, but a finished reference sequence is now available [17]. These duplication events may influence the presence of segmental duplications and is suggested to lead to an increase in CNVs in the zebrafish genome [22]. As such, the genome duplication may also influence the frequency at which CNVs generated by chemical exposures may be observed. Additional studies will be needed with other model systems to investigate and further understand the influence of chemical exposure on generating CNVs.

5. Conclusions

Overall this study indicates that chemical exposure results in CNVs that alter gene expression. This study is setting the stage for future investigations into the specific mechanism of formation and expansion to assess environmental chemicals, additional structural genomic alterations using sequencing technologies, and inclusion of in vivo systems to study the biological and functional significance of the CNVs and their influence on disease pathways.

Conflict of Interests

The authors declare no conflict of interests regarding the publication of this paper.

Acknowledgments

This work was supported by a Purdue Research Foundation Faculty Grant (JLF), a Purdue OVPR Incentive Grant Program: Encouraging and Stimulating Purdue’s Emerging Research: Category II Grant (JLF), and a Purdue Research Foundation Research Grant X1 (JLF, SMP).

References

  1. The International HapMap Consortium, “A haplotype map of the human genome,” Nature, vol. 437, pp. 1299–1320, 2005. View at Publisher · View at Google Scholar
  2. A. J. Iafrate, L. Feuk, M. N. Rivera et al., “Detection of large-scale variation in the human genome,” Nature Genetics, vol. 36, no. 9, pp. 949–951, 2004. View at Publisher · View at Google Scholar · View at Scopus
  3. J. Sebat, B. Lakshmi, J. Troge et al., “Large-scale copy number polymorphism in the human genome,” Science, vol. 305, no. 5683, pp. 525–528, 2004. View at Publisher · View at Google Scholar · View at Scopus
  4. J. L. Freeman, G. H. Perry, L. Feuk et al., “Copy number variation: new insights in genome diversity,” Genome Research, vol. 16, no. 8, pp. 949–961, 2006. View at Publisher · View at Google Scholar · View at Scopus
  5. R. Redon, S. Ishikawa, K. R. Fitch et al., “Global variation in copy number in the human genome,” Nature, vol. 444, no. 7118, pp. 444–454, 2006. View at Publisher · View at Google Scholar · View at Scopus
  6. D. F. Conrad, D. Pinto, R. Redon et al., “Origins and functional impact of copy number variation in the human genome,” Nature, vol. 464, no. 7289, pp. 704–712, 2010. View at Publisher · View at Google Scholar · View at Scopus
  7. B. E. Stranger, M. S. Forrest, M. Dunning et al., “Relative impact of nucleotide and copy number variation on gene expression phenotypes,” Science, vol. 315, no. 5813, pp. 848–853, 2007. View at Publisher · View at Google Scholar · View at Scopus
  8. J. Sebat, B. Lakshmi, D. Malhotra et al., “Strong association of de novo copy number mutations with autism,” Science, vol. 316, no. 5823, pp. 445–449, 2007. View at Publisher · View at Google Scholar · View at Scopus
  9. T. Walsh, J. M. McClellan, S. E. McCarthy et al., “Rare structural variants disrupt multiple genes in neurodevelopmental pathways in schizophrenia,” Science, vol. 320, no. 5875, pp. 539–543, 2008. View at Publisher · View at Google Scholar · View at Scopus
  10. C. E. G. Bruder, A. Piotrowski, A. A. C. J. Gijsbers et al., “Phenotypically concordant and discordant monozygotic twins display different DNA copy number variation profiles,” The American Journal of Human Genetics, vol. 82, no. 3, pp. 763–771, 2008. View at Publisher · View at Google Scholar · View at Scopus
  11. M. M. Mitchell, R. Woods, L.-H. Chi et al., “Levels of select PCB and PBDE congeners in human postmortem brain reveal possible environmental involvement in 15q11-q13 duplication autism spectrum disorder,” Environmental and Molecular Mutagenesis, vol. 53, no. 8, pp. 589–598, 2012. View at Publisher · View at Google Scholar · View at Scopus
  12. M. F. Arlt, J. G. Mulle, V. M. Schaibley et al., “Replication stress induces genome-wide copy number changes in human cells that resemble polymorphic and pathogenic variants,” American Journal of Human Genetics, vol. 84, no. 3, pp. 339–350, 2009. View at Publisher · View at Google Scholar · View at Scopus
  13. M. F. Arlt, A. C. Ozdemir, S. R. Birkeland, T. E. Wilson, and T. W. Glover, “Hydroxyurea induces de novo copy number variants in human cells,” Proceedings of the National Academy of Sciences of the United States of America, vol. 108, no. 42, pp. 17360–17365, 2011. View at Publisher · View at Google Scholar · View at Scopus
  14. M. F. Arlt, S. Rajendran, S. R. Birkeland, T. E. Wilson, and T. W. Glover, “Copy number variants are produced in response to low-dose ionizing radiation in cultured cells,” Environmental and Molecular Mutagenesis, vol. 55, no. 2, pp. 103–113, 2014. View at Publisher · View at Google Scholar · View at Scopus
  15. C. L. Yauk, J. Lucas Argueso, S. S. Auerbach et al., “Harnessing genomics to identify environmental determinants of heritable disease,” Mutation Research—Reviews in Mutation Research, vol. 752, no. 1, pp. 6–9, 2013. View at Publisher · View at Google Scholar · View at Scopus
  16. W. B. Barbazuk, I. Korf, C. Kadavi et al., “The syntenic relationship of the zebrafish and human genomes,” Genome Research, vol. 10, no. 9, pp. 1351–1358, 2000. View at Publisher · View at Google Scholar · View at Scopus
  17. K. Howe, M. D. Clark, C. F. Torroja et al., “The zebrafish reference genome sequence and its relationship to the human genome,” Nature, vol. 496, no. 7446, pp. 498–503, 2013. View at Publisher · View at Google Scholar
  18. J. F. Amatruda and L. I. Zon, “Dissecting hematopoiesis and disease using the zebrafish,” Developmental Biology, vol. 216, no. 1, pp. 1–15, 1999. View at Publisher · View at Google Scholar · View at Scopus
  19. K. Dooley and L. I. Zon, “Zebrafish: a model system for the study of human disease,” Current Opinion in Genetics and Development, vol. 10, no. 3, pp. 252–256, 2000. View at Publisher · View at Google Scholar · View at Scopus
  20. A. J. Hill, H. Teraoka, W. Heideman, and R. E. Peterson, “Zebrafish as a model vertebrate for investigating chemical toxicity,” Toxicological Sciences, vol. 86, no. 1, pp. 6–19, 2005. View at Publisher · View at Google Scholar · View at Scopus
  21. C. L. Bladen, S. Navarre, W. S. Dynan, and D. J. Kozlowski, “Expression of the Ku70 subunit (XRCC6) and protection from low dose ionizing radiation during zebrafish embryogenesis,” Neuroscience Letters, vol. 422, no. 2, pp. 97–102, 2007. View at Publisher · View at Google Scholar · View at Scopus
  22. K. H. Brown, K. P. Dobrinski, A. S. Lee et al., “Extensive genetic diversity and substructuring among zebrafish strains revealed through copy number variant analysis,” Proceedings of the National Academy of Sciences of the United States of America, vol. 109, no. 2, pp. 529–534, 2012. View at Publisher · View at Google Scholar · View at Scopus
  23. J. L. Freeman, A. Adeniyi, R. Banerjee et al., “Definition of the zebrafish genome using flow cytometry and cytogenetic mapping,” BMC Genomics, vol. 8, article 195, 2007. View at Publisher · View at Google Scholar · View at Scopus
  24. M. J. Plewa, Y. Kargalioglu, D. Vankerk, R. A. Minear, and E. D. Wagner, “Mammalian cell cytotoxicity and genotoxicity analysis of drinking water disinfection by-products,” Environmental and Molecular Mutagenesis, vol. 40, no. 2, pp. 134–142, 2002. View at Publisher · View at Google Scholar · View at Scopus
  25. S. M. Peterson and J. L. Freeman, “Cancer cytogenetics in the zebrafish,” Zebrafish, vol. 6, no. 4, pp. 355–360, 2009. View at Publisher · View at Google Scholar · View at Scopus
  26. J. L. Freeman, C. Ceol, H. Feng et al., “Construction and application of a zebrafish array comparative genomic hybridization platform,” Genes Chromosomes and Cancer, vol. 48, no. 2, pp. 155–170, 2009. View at Publisher · View at Google Scholar · View at Scopus
  27. C. Workman, L. J. Jensen, H. Jarmer et al., “A new non-linear normalization method for reducing variability in DNA microarray experiments,” Genome Biology, vol. 3, no. 9, pp. 1–16, 2002. View at Google Scholar
  28. S. M. Peterson, J. Zhang, G. Weber, and J. L. Freeman, “Global gene expression analysis reveals dynamic and developmental stage-dependent enrichment of lead-induced neurological gene alterations,” Environmental Health Perspectives, vol. 119, no. 5, pp. 615–621, 2011. View at Publisher · View at Google Scholar · View at Scopus
  29. B. M. Bolstad, R. A. Irizarry, M. Åstrand, and T. P. Speed, “A comparison of normalization methods for high density oligonucleotide array data based on variance and bias,” Bioinformatics, vol. 19, no. 2, pp. 185–193, 2003. View at Publisher · View at Google Scholar · View at Scopus
  30. R. A. Irizarry, B. Hobbs, F. Collin et al., “Exploration, normalization, and summaries of high density oligonucleotide array probe level data,” Biostatistics, vol. 4, no. 2, pp. 249–264, 2003. View at Publisher · View at Google Scholar · View at Scopus
  31. L. Guo, E. K. Lobenhofer, C. Wang et al., “Rat toxicogenomic study reveals analytical consistency across microarray platforms,” Nature Biotechnology, vol. 24, no. 9, pp. 1162–1169, 2006. View at Publisher · View at Google Scholar · View at Scopus
  32. L. Shi, L. H. Reid, W. D. Jones et al., “The MicroArray Quality Control (MAQC) project shows inter- and intraplatform reproducibility of gene expression measurements,” Nature Biotechnology, vol. 24, no. 9, pp. 1151–1161, 2006. View at Publisher · View at Google Scholar
  33. P. F. Swann and P. N. Magee, “Nitrosamine-induced carcinogenesis. The alkylation of N-7 of guanine of nucleic acids of the rat by diethylnitrosamine, N-ethyl-N-nitrosourea and ethyl methanesulphonate,” Biochemical Journal, vol. 125, no. 3, pp. 841–847, 1971. View at Google Scholar · View at Scopus
  34. K. Mogami, P. T. O'Donnell, S. I. Bernstein, T. R. Wright, and C. P. Emerson Jr., “Mutations of the Drosophila myosin heavy-chain gene: effects on transcription, myosin accumulation, and muscle function,” Proceedings of the National Academy of Sciences of the United States of America, vol. 83, no. 5, pp. 1393–1397, 1986. View at Publisher · View at Google Scholar · View at Scopus
  35. S. G. Whittaker, S. F. Moser, D. H. Maloney, W. W. Piegorsch, M. A. Resnick, and S. Fogel, “The detection of mitotic and meiotic chromosome gain in the yeast Saccharomyces cerevisiae: effects of methyl benzimidazol-2-yl carbamate, methyl methanesulfonate, ethyl methanesulfonate, dimethyl sulfoxide, propionitrile and cyclophosphamide monohydrate,” Mutation Research/Genetic Toxicology, vol. 242, no. 3, pp. 231–258, 1990. View at Publisher · View at Google Scholar · View at Scopus
  36. J. L. Freeman and A. L. Rayburn, “In vivo genotoxicity of atrazine to anuran larvae,” Mutation Research—Genetic Toxicology and Environmental Mutagenesis, vol. 560, no. 1, pp. 69–78, 2004. View at Publisher · View at Google Scholar · View at Scopus
  37. G. E. Pantelias and S. Wolff, “Cytosine arabinoside is a potent clastogen and does not affect the repair of X-ray-induced chromosome fragments in unstimulated human lymphocytes,” Mutation Research, vol. 151, no. 1, pp. 65–72, 1985. View at Publisher · View at Google Scholar · View at Scopus
  38. F. M. Elli, L. De Sanctis, E. Peverelli et al., “Autosomal dominant pseudohypoparathyroidism type Ib: a novel inherited deletion ablating STX16 causes loss of imprinting at the A/B DMR,” Journal of Clinical Endocrinology and Metabolism, vol. 99, no. 4, pp. E724–E728, 2014. View at Publisher · View at Google Scholar · View at Scopus
  39. A. Linge, S. Kennedy, D. O'Flynn et al., “Differential expression of fourteen proteins between uveal melanoma from patients who subsequently developed distant metastases versus those who did not,” Investigative Ophthalmology and Visual Science, vol. 53, no. 8, pp. 4634–4643, 2012. View at Publisher · View at Google Scholar · View at Scopus
  40. M. Emerenciano, T. C. Barbosa, B. A. Lopes et al., “ARID5B polymorphism confers an increased risk to acquire specific MLL rearrangements in early childhood leukemia,” BMC Cancer, vol. 14, no. 1, article 127, 2014. View at Publisher · View at Google Scholar · View at Scopus
  41. V. B. Mahajan and J. H. Lin, “Lymphocyte infiltration in CAPN5 autosomal dominant neovascular inflammatory vitreoretinopathy,” Clinical Ophthalmology, vol. 7, pp. 1339–1345, 2013. View at Publisher · View at Google Scholar · View at Scopus
  42. T. Franz, L. Winckler, T. Boehm, and T. N. Dear, “Capn5 is expressed in a subset of T cells and is dispensable for development,” Molecular and Cellular Biology, vol. 24, no. 4, pp. 1649–1654, 2004. View at Publisher · View at Google Scholar · View at Scopus
  43. C. N. Henrichsen, E. Chaignat, and A. Reymond, “Copy number variants, diseases and gene expression,” Human Molecular Genetics, vol. 18, no. 1, pp. R1–R8, 2009. View at Publisher · View at Google Scholar · View at Scopus
  44. I. G. Woods, C. Wilson, B. Friedlander et al., “The zebrafish gene map defines ancestral vertebrate chromosomes,” Genome Research, vol. 15, no. 9, pp. 1307–1314, 2005. View at Publisher · View at Google Scholar · View at Scopus