Epigenetic Alterations in Liver of C57BL/6J Mice after Short-Term Inhalational Exposure to 1,3-Butadiene

Background 1,3-Butadiene (BD) is a high-volume industrial chemical and a known human carcinogen. The main mode of BD carcinogenicity is thought to involve formation of genotoxic epoxides. Objectives In this study we tested the hypothesis that BD may be epigenotoxic (i.e., cause changes in DNA and histone methylation) and explored the possible molecular mechanisms for the epigenetic changes. Methods and Results We administered BD (6.25 and 625 ppm) to C57BL/6J male mice by inhalation for 2 weeks (6 hr/day, 5 days a week) and then examined liver tissue from these mice for signs of toxicity using histopathology and gene expression analyses. We observed no changes in mice exposed to 6.25 ppm BD, but glycogen depletion and dysregulation of hepatotoxicity biomarker genes were observed in mice exposed to 625 ppm BD. We detected N-7-(2,3,4-trihydroxybut-1-yl)guanine (THB-Gua) adducts in liver DNA of exposed mice in a dose-responsive manner, and also observed extensive alterations in the cellular epigenome in the liver, including demethylation of global DNA and repetitive elements and a decrease in histone H3 and H4 lysine methylation. In addition, we observed down-regulation of DNA methyltransferase 1 (Dnmt1) and suppressor of variegation 3–9 homolog 1, a histone lysine methyltransferase (Suv39h1), and up-regulation of the histone demethylase Jumonji domain 2 (Jmjd2a), proteins responsible for the accurate maintenance of the epigenetic marks. Although the epigenetic effects were most pronounced in the 625-ppm exposure group, some effects were evident in mice exposed to 6.25 ppm BD. Conclusions This study demonstrates that exposure to BD leads to epigenetic alterations in the liver, which may be important contributors to the mode of BD carcinogenicity.

The gaseous olefin 1,3-butadiene (BD) is a major high-volume industrial chemical monomer widely used in the production of synthetic rubber, resins, and plastics (Himmelstein et al. 1997;White 2007). The International Agency for Research on Cancer (IARC 2008) has recog nized BD as "known to be carcinogenic to humans." In rodents, BD causes tumor formation at several sites, including the hemato poietic system, lungs, heart, and liver (Himmelstein et al. 1997;Melnick and Sills 2001). BD is metabolized in the liver through oxidation by the family of cytochrome P450 mono oxygenases, a pathway that forms several epoxides, specifically 1,2-epoxy-3-butene, 1,2,3,4-diepoxybutane, and 3,4-epoxy-1,2butanediol (Filser et al. 2007;Himmelstein et al. 1997). The genotoxicity of BD-derived epoxides is considered to be a critical event in the initiation of tumorigenesis (Cochrane and Skopek 1994;Kemper et al. 2001).
Most of the research on the carcino genicity assessment of BD has focused on the formation of DNA adducts that lead to mutations and chromosomal aberrations (Goggin et al. 2009;Swenberg et al. 2000); however, even within classical genotoxic carcinogenesis models, the formation of DNA adducts is not the only mode of carcinogenesis. Recent data point to epigenetic alterations as important adverse biological effects caused by exposure to numerous environmental chemical and physical agents (Reamon-Buettner et al. 2008). Interestingly, data from human and animal studies have demon strated that many early indicators of exposure to environ mental toxi cants are epigenetic in nature (Baccarelli and Bollati 2009). It has been also suggested that epigenetic alterations, including whole-genome-and repetitive-element-associated hypo methylation and hyper methylation of the promoters of key genes (e.g., O 6 -methylguanine-DNA methyl transferase), may precede and/or provoke genetic alterations (Jacinto and Esteller 2007;Sawan et al. 2008). Furthermore, epigenetic changes not only may be important for under standing the molecular under pinnings of environmentally related disease but also may serve as biomarkers in toxicity and/or carcinogenicity assessment (Marlowe et al. 2009;Reamon-Buettner et al. 2008).
It is not clear whether exposure to BD causes epigenetic alterations in target tissues in addition to the induction of well-defined genotoxic changes. Thus, we tested the hypothesis that BD is also epi genotoxic (i.e., causes changes in DNA and histone methylation) and we explored the possible molecular mechanisms of these effects.
Background: 1,3-Butadiene (BD) is a high-volume industrial chemical and a known human carcinogen. The main mode of BD carcinogenicity is thought to involve formation of genotoxic epoxides. oBjectives: In this study we tested the hypothesis that BD may be epi genotoxic (i.e., cause changes in DNA and histone methylation) and explored the possible molecular mechanisms for the epi genetic changes. Methods and results: We administered BD (6.25 and 625 ppm) to C57BL/6J male mice by inhalation for 2 weeks (6 hr/day, 5 days a week) and then examined liver tissue from these mice for signs of toxicity using histopathology and gene expression analyses. We observed no changes in mice exposed to 6.25 ppm BD, but glycogen depletion and dys regulation of hepato toxicity biomarker genes were observed in mice exposed to 625 ppm BD. We detected N-7-(2,3,4-trihydroxybut-1-yl) guanine (THB-Gua) adducts in liver DNA of exposed mice in a dose-responsive manner, and also observed extensive alterations in the cellular epi genome in the liver, including demethylation of global DNA and repetitive elements and a decrease in histone H3 and H4 lysine methylation. In addition, we observed down-regulation of DNA methyltransferase 1 (Dnmt1) and suppressor of variegation 3-9 homolog 1, a histone lysine methyltransferase (Suv39h1), and up-regulation of the histone demethylase Jumonji domain 2 (Jmjd2a), proteins responsible for the accurate maintenance of the epigenetic marks. Although the epigenetic effects were most pronounced in the 625-ppm exposure group, some effects were evident in mice exposed to 6.25 ppm BD. conclusions: This study demonstrates that exposure to BD leads to epigenetic alterations in the liver, which may be important contributors to the mode of BD carcinogenicity. Determination of N-7 guanine adduct formation. Genomic DNA was isolated from mouse liver tissues by standard digestion of the tissue with proteinase K, followed by phenolchloroform extraction and ethanol precipitation. Levels of N-7-(2,3,4-trihydroxybut-1-yl) guanine (THB-Gua) were analyzed by liquid chromatography/tandem mass spectrometry as described by Koc et al. (1999) with minor modifications. Briefly, DNA (300 µg) from each sample was spiked with 500 fmol THB-Gua internal standard and adjusted with distilled water to 400 µL volume. Before injection into the ACQUITY UPLC column (Waters Corp., Milford, MA, USA) coupled to TSQ Quantum Ultra mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA), neutral thermal hydrolysis was performed at 95°C for 30 min and samples were centrifuged through a 10-kDa filter twice.
Quantitative reverse-transcription polymerase chain reaction (PCR). Total RNA was extracted from mouse liver tissues using TRI reagent (Ambion, Austin, TX, USA) according to the manufacturer's instructions. cDNA was synthesized from 5 µg total RNA using the RT 2 First Strand cDNA synthesis kit (SABiosciences, Frederick, MD, USA). We used Mouse Hepatotoxicity RT 2 Profiler PCR Arrays (SABiosciences), according to the manufacturer's protocol, to determine expression of 84 key genes implicated as potential biomarkers of liver toxicity. The rela tive level of mRNA for each gene was determined using the 2 ΔΔCt method (Schmittgen and Livak 2008). The results are presented as fold change for each mRNA in the liver of mice exposed to BD relative to those from control mice.
Determination of global DNA methylation status by methylation-sensitive cytosine extension assay. We evaluated the extent of global DNA methylation using a radiolabeled deoxycytidine-5´-triphosphate ([ 3 H]dCTP) extension assay as described elsewhere (Pogribny et al. 1999).

Methylation-sensitive quantitative PCR (qPCR) analysis of repetitive elements methylation.
The methyla tion status of major and minor satellites, long interspersed elements 1 (LINE1), and short interspersed nuclear elements (SINE) B1 and B2 repetitive elements was determined by methylation-sensitive McrBC-qPCR assay (Martens et al. 2005). Briefly, genomic DNA (1 µg) was digested overnight with the methylation-specific restriction enzyme McrBC (New England Biolabs, Ipswich, MA, USA) and then analyzed by qPCR on an ABI 7900 Real-time PCR System (Applied Biosystems, Forrest City, CA, USA). The threshold cycle (C t ) was defined as the fractional cycle number that surpassed the fixed threshold. The C t values were converted into the absolute amount of input DNA using an absolute standard curve method. An increased amount of input DNA after digestion with McrBC was indicative of hypo methylation, whereas a decreased amount of input DNA was indicative of hyper methylation.

Methylated DNA immunoprecipitation (MeDIP)-qPCR analysis of LINE1 methylation.
We used a MeDIP assay combined with qPCR to assess quantitatively the methylation status of LINE1 repetitive sequences in liver of control and BD-exposed mice. MeDIP was performed as described by Weber et al. (2005). Briefly, 5 µg genomic DNA was randomly sheared to an average length of 0.2-1.0 kb by sonication and divided into immuno precipitated and input portions. DNA from the immuno precipitated portions was incubated overnight at 4°C with a monoclonal antibody (5 µg) against 5-methyl cytosine (Abcam, Cambridge, MA, USA), followed by overnight incubation with Pan-mouse IgG Dynal magnetic beads (Invitrogen, Carlsbad, CA, USA) at 4°C. The methylated DNA/ antibody complexes were then digested with proteinase K, and enriched DNA was recovered by phenol-chloroform extraction followed by ethanol precipitation. Purified DNA from immuno precipitated DNA and input DNA samples were analyzed by qPCR on an ABI 7900 real-time PCR system as described above. The relative changes in the extent of LINE1 methylation were determined by measuring the amount of DNA in immuno precipitated DNA after normalization to the input DNA.
Statistical analyses. Results are presented as mean ± SD. Comparisons between control and BD-exposed mice were made by Student's t-test; p-values < 0.05 were considered significant.

Effects of BD exposure on body weight, serum alanine aminotransferase (ALT) levels, and liver histopathology.
We evaluated changes in body weight and serum ALT activity, and Figure 1. Effects of BD exposure on body weight, liver histopathology, and DNA damage. (A) Body weights (mean ± SD) of control mice at the beginning and end of the study and of mice exposed to 6.25 ppm BD or 625 ppm BD for 2 weeks (n = 5). (B) Representative images of hema toxylin and eosin-stained liver sections from control mice and mice exposed to 625 ppm BD (original magnification, 400×). (C) Amounts (mean ± SD) of THB-Gua-BD adducts (total and racemic and meso forms) in liver DNA from mice exposed to 0, 6.25, or 625 ppm BD (n = 5). *p < 0.05 compared with control mice at day 12. # p < 0.05 compared with control.

Amount of THB-Gua adducts/10 7 nucleotides
Racemic Meso histo morphological alterations in the liver of C57BL/6J mice exposed to 0, 6.25, or 625 ppm BD. We observed a small (> 10%) yet significant body weight loss only in mice exposed to 625 ppm BD ( Figure 1A). There were no significant changes in the liver sample weights between the control group and either treatment group. Serum ALT activity did not change in mice exposed to BD (data not shown). Histopathological evaluation of the liver sections revealed that the hepatocytes of control mice manifested minimal to mild hepato cellular cyto plasmic vacuolization morphologically characteristic of glycogen ( Figure 1B, left). In contrast, the hepato cytes of the BD-treated mice showed diffuse glycogen depletion characterized by the lack of cyto plasmic vacuoles ( Figure 1B, right).

Effect of BD exposure on expression of hepatotoxicity-related biomarker genes.
To determine whether inhalational exposure to BD resulted in the molecular changes indicative of liver toxicity, we examined expression of a panel of genes known to serve as biomarkers of toxicity. Livers from mice exposed to 6.25 ppm BD showed mini mal changes in gene expression, whereas those from mice exposed to 625 ppm BD showed changes in gene expression that were more pronounced (Table 1). This was evident by a greater number of differentially expressed genes and greater magnitude of gene expression in mice exposed to 625 ppm BD compared with mice exposed to 6.25 ppm BD (Table 1). Most of the genes, including carbonic anhydrase 3 (Car3), leucine-rich repeat-containing G protein-coupled receptor 5 (Lgr5), and sterol regulatory element binding transcription factor 1 (Srebf1), were down-regulated (Table 1). Interestingly, decreased expression of Car3, Lgr5, and Srebf1 genes is a common response to liver injury (Hsieh et al. 2009;Kojima et al. 2002). In contrast, the expression of heme oxygenase-1 (Hmox1) was up-regulated nearly 18-fold (Table 1). Hmox1 is one of several coordinately regulated proteins involved in protecting against liver injury induced by a variety of hepato toxicants and has been suggested to be a reliable indicator of hepatotoxicity (Choi et al. 2003;Morio et al. 2006).
Hepatic THB-Gua adduct levels. We assessed amounts of THB-Gua adducts in hepatic DNA from control mice and mice exposed to 6.25 and 625 ppm BD using mass spectrometry ( Figure 1C). We observed a dose-dependent increase in THB-Gua adducts in hepatic DNA, with the levels significantly increased after both 6.25 and 625 ppm BD exposures.
Effect of BD exposure on the levels of global DNA and repetitive elements methyla tion. BD exposure resulted in substantial decreases in the levels of global DNA methyla tion, to similar extents in both 6.25 and 625 ppm groups ( Figure 2A). The status of methylation of major and minor satellites and of LINE1 and SINE B1 and B2 repetitive elements is a sensitive indicator of the degree of global DNA methylation (Yang et al. 2004). Major and minor satellites and SINE B1 and B2 repetitive elements were demethylated in a dose-dependent manner after BD exposure ( Figure 2B,C), whereas LINE1 methylation decreased only in mice exposed to 625 ppm [ Figure 2C; see also Supplemental Material, Figure 1  Effect of BD exposure on histone modifications. Alterations in histone-controlled chromatin structure are important epigenetic changes elicited by many toxicants (Bagnyukova et al. 2008;Moggs and Orphanides 2004); therefore, we examined the effect of BD on the extent of methylation of histones H3K9, H3K27, and H4K20. We found no changes in histone trimethylation in the livers of mice exposed to 6.25 ppm BD (Figure 3). In contrast, exposure to 625 ppm BD caused a significant decrease in the levels of H3K9me3, H3K27me3, and H4K20me3: 30%, 19%, and 54% lower, respectively, than in control mice. Exposure to either 6.25 or 625 ppm BD had no effect on the extent of histone H3K9 Table 1. Genes significantly differentially expressed (> 2-fold) in mouse liver after BD exposure compared with control animals (p < 0.05).

Figure 2.
Effects of BD exposure on DNA methylation in mouse liver. (A) Loss of global DNA methylation in livers of BD-exposed mice measured using the cytosine extension DNA methylation assay (Pogribny et al. 1999). Methylation status of major and minor satellites (B) and of SINE B1 and B2 and LINE1 repetitive elements (C) assessed using the methylation-sensitive McrBC-qPCR assay; data are mean ± SD (n = 5) fold change relative to control.
*p < 0.05 compared with control. **p < 0.05 compared with 6.25 ppm BD. H4K20me2 was decreased by 39% in mice exposed to 625 ppm BD (see Supplemental Material, Figure 2). Effect of BD exposure on the levels of Dnmt1, Suv39h1, Suv4-20h1, and Jmjd2a proteins. Because BD exposure had an effect on DNA and histone methylation, we also investigated possible mechanisms of these epi genetic aberrations. Dnmt1 is the main enzyme responsible for maintaining faithful genomic methylation in somatic mammalian cells (Goll and Bestor 2005). Therefore, we assessed whether the observed loss of global DNA and LINE1 methylation in the livers of BD-exposed mice may be attributed to alterations in Dnmt1 expression. Exposure of C57BL/6J mice to 625 ppm BD resulted in a significant 47% decrease in the levels of Dnmt1 protein ( Figure 4). Likewise, this concentration of BD led to a significant down-regulation of Suv39h1, the main histone methyl transferase responsible for H3K9 tri methyla tion (Rea et al. 2000). In contrast, levels of Suv4-20h1, a methyl transferase that catalyzes trimethylation of histone H4K20 (Benetti et al. 2007), were not altered in the livers of BD-exposed (625 ppm) mice, despite a substantial decrease in H4K20me3 (Figure 3).
To explain this discrepancy and to investigate other mechanisms that may cause the decrease in H3K9me3, we examined the levels of Jmjd2a, the histone demethylase that catalyzes demethylation of trimethylated H3K9 and H4K20 (Couture et al. 2007). Indeed, BD exposure (625 ppm) caused a significant 68% increase in the protein level of Jmjd2a in the livers (Figure 4). Levels of Dnmt1, Suv39h1, and Jmjd2a proteins did not change in mice exposed to 6.25 ppm BD (data not shown).

Discussion
Environmental chemicals are considered to be geno toxic if they, or products of their metabolism, are shown to interact directly with DNA, causing damage and mutations and ultimately leading to tumor formation (Shuker 2002). Agents that are carcinogenic yet are not known to be DNA reactive or test negative in the genotoxicity assays are classified as non genotoxic. Although this traditional classi fi cation has been useful in both mechanistic toxicology and regulatory decision making, mounting evidence suggests that despite differences in a chemical's DNA reactivity, both classes of agents may lead to prominent epi genomic alterations in tissues that are targets for carcino genesis. Recent work on epigenetic effects of various chemical agents has led to a suggestion that some carcinogens may be epi genotoxic (Pogribny et al. 2008).
The present study demonstrates that even short-term exposure of mice to BD by inhalation not only led to formation of DNA adducts but also produced prominent epigenetic changes in the liver, thus establishing BD as both a genotoxic and an epigenotoxic chemical. Specifically, we observed a pronounced loss of global DNA and LINE1 methylation; substantial decreases in histone H3K9 trimethylation, histone H4K20 dimethyla tion, and histone H4K20 trimethyla tion, and altered expression of proteins responsible for the accurate maintenance of these epigenetic marks. Even though it is well established that BD-induced cancer development involves formation of geno toxic reactive metabolites of BD and subsequent formation of DNA adducts, especially the most abundant THB-Gua adducts (Goggin et al. 2009;Koc et al. 1999), our study shows that DNA damage may not be the only mode of carcinogenic action for BD.
One of the most pronounced changes observed in the present study was BD-induced loss of global DNA and LINE1 methylation. Several mechanisms may be responsible for global DNA demethylation, including a) formation of adducts with 2´-deoxyguanosine, especially in GC-rich DNA domains; b) DNA damage and repair; and c) altered expression of the main cellular DNA methyltransferase Dnmt1. Indeed, BD exposure induced profound DNA damage, especially the formation of N-7 guanine adducts, whose presence may profoundly compromise the methylation capacity of Dnmt1, as has been observed with other guanine adducts (Valinluck and Sowers 2007). Additionally, activation of repair-mediated DNA demethyla tion (Barreto et al. 2007) and decreased expression and/or functioning of Dnmt1 caused by a number of factors, including direct effects of BD and its metabolites on Dnmt1 protein, aberrant Figure 3. Effects of BD exposure on histone trimethylation in mouse liver as determined by Western blotting. (A) H3K9me3, H3K27me3, and H4K20me3 levels assessed by immuno staining using specific anti bodies against trimethylated histones; equal sample loading was confirmed by immuno staining against histone H3 ("Loading" row) and histone H4 (data not shown). (B) Densitometry analysis of the immuno staining results shown as change in methylation relative to control after correction for the total amount of each histone in the individual samples; data are presented as mean ± SD (n = 5).
*p < 0.05 compared with control.  Figure 4. Western blot analysis of proteins responsible for DNA and histone lysine methylation in the livers of control and BD-exposed mice. (A) Dnmt1, Jmjd2a, Suv39h1, and Suv4-20h1 proteins assessed by immuno staining using specific anti bodies; equal sample loading was confirmed by immunostaining against β-actin ("Loading" row). (B) Densitometry analysis of the immuno staining results shown as change relative to control; data are presented as mean ± SD (n = 5).
*p < 0.05 compared with control. Change from control (%) * * * expression of microRNAs (e.g., miR-29b, miR-148, and miR-152), and expression of chromatin-modifying proteins (Garzon et al. 2009;Huang et al. 2010;Vire et al. 2006;Wang et al. 2009), may further contribute to the loss of DNA methylation. BD-induced hypomethylation of DNA was also accompanied by a decrease in trimethylation of histones H3K9, H3K27, and H4K20. The accurate maintenance of the status of chromatin methylation depends on proper functioning and inter dependent coopera tion between DNA methylation and histone methylation machinery. For instance, direct interaction between Dnmt1 and Suv39h1 histone methyl transferase coordinates both DNA and histone lysine methylation (Esteve et al. 2006;Fuks et al. 2003). Because we observed downregulation of Dnmt1 and Suv39h1 after exposure to 625 ppm BD, these proteins may not only induce demethylation of DNA but also lead to a decrease in histone lysine trimethylation. Correspondingly, expression of Jmjd2a, an enzyme that possesses lysine demethylase activity with specificity to H3K9 (Yamane et al. 2006), was up-regulated, which also may be one of the factors explaining the mechanisms for the observed effects of BD exposure on histones. Finally, the BD epoxides may also bind to methylation or acetylation sites on H3 and H4 so that they cannot be methylated or acetylated. This phenomenon has been shown for formaldehyde (Lu et al. 2008); however, the relevance of this mechanism for BD needs to be established. Importantly, altered expression and/or functioning of Dnmt1, Suv39h1, Suv4-20h1, and Jmjd2a proteins has been attributed to destabilization of chromatin assembly and genome stability (Dion et al. 2008;Fodor et al. 2006;Peng and Karpen 2009;Peters et al. 2001). Specifically, it has been demonstrated that loss of histone H3K9, H3K27, and H4K20 trimethylation markedly impairs chromatin structure, diminishes the ability of the cells to regulate and maintain accurately the cell cycle, disrupts the balance between cell proliferation and differentiation, and severely reduces cell viability (Rea et al. 2000;Yang and Mizzen 2009). H4K20 dimethylation plays a crucial role in DNA repair by attracting the checkpoint protein Crb2 to the sites of DNA damage (Greeson et al. 2008). In addition, a decrease in H4K20me2 impairs cell survival after genotoxic insult and disrupts the ability of cells to maintain checkpoint-mediated cellcycle arrest (Sanders et al. 2004).

Conclusions
This study shows that short-term inhalational exposure of mice to the genotoxic carcinogen BD, in addition to inducing the formation of DNA adducts, also caused prominent epigenetic alterations in the liver. These results highlight the significance of epigenetic events in the mechanism of BD toxicity and carcinogenicity. Epigenetic alterations are not only important features of cancer cells but also play a major role in the etiology of cancer. Furthermore, alterations in the cellu lar epigenome, which may be termed epi genotoxicity, elicited by various genotoxic and non genotoxic agents may result in the emergence of epigenetically reprogrammed proliferating cells with a growth-advantage pheno type and a high potential for activation of mutator pathways (Pogribny et al. 2008). We also posit that early appearance of epigenetic alterations and their association with altered expression of genes that are indicators of liver toxicity strongly suggest that changes in DNA and histone methylation may be useful predictive markers for safety assessment.