Histone Methylation-Mediated Reproductive Toxicity to Consumer Product Chemicals in Caenorhabditis elegans: An Epigenetic Adverse Outcome Pathway (AOP)

Abstract

The significance of histone methylation in epigenetic inheritance underscores its relevance to disease and the chronic effects of environmental chemicals. However, limited evidence of the causal relationships between chemically induced epigenetic changes and organismal-level effects hinders the application of epigenetic markers in ecotoxicological assessments. This study explored the contribution of repressive histone marks to reproductive toxicity induced by chemicals in consumer products in Caenorhabditis elegans, applying the adverse outcome pathway (AOP) framework. Triclosan (TCS) and tetrabromobisphenol A (TBBPA) exposures caused reproductive toxicity and altered histone methyltransferase (HMT) and histone demethylase (HDM) activities, increasing the level of trimethylation of H3K9 and H3K27. Notably, treatment with an H3K27-specific HMT inhibitor alleviated reproductive defects and the transcriptional response of genes related to vitellogenin, xenobiotic metabolism, and oxidative stress. Comparison of points of departure (PODs) based on calculated benchmark concentrations (BMCs) revealed the sensitivity of histone-modifying enzyme activities to these chemicals. Our findings suggest that the ‘disturbance of HMT and HDM’ can serve as the molecular initiating event (MIE) leading to reproductive toxicity in the epigenetic AOP for TCS and TBBPA. The study extended the biological applicability of these enzymes by identifying model species with analogous protein sequences and functions. This combined approach enhances the essentiality, empirical support, and taxonomic domain of applicability (tDOA), which are crucial considerations for ecotoxicological AOPs. Given the widespread use and environmental distribution of chemicals in consumer products, this study proposes histone-modifying enzyme activity as an effective screening tool for reproductive toxicants and emphasizes the integration of epigenetic mechanisms into a prospective ERA.

Short abstract

This study verifies the contribution of epigenetic changes as sensitive markers to the toxicity of environmental chemicals in C. elegans and expands their cross-species applicability.

1. Introduction

Daily exposure to environmental chemicals in air, water, soil, food, and consumer products impacts the risk of developing various chronic diseases and environmental health. Notably, certain environmental agents can exert persistent and transgenerational adverse effects on health. In this regard, epigenome has emerged as a promising interface between the environment and the genome, shedding light on environmental toxicities that cannot be fully explained by genetic factors.¹⁻³ Epigenetics encompasses the full spectrum of transcriptional regulatory processes that alter gene function and cellular and phenotypic state.^4,5 Moreover, epigenetic modifications are plastic in response to environmental chemicals and can be heritable across life stages and generations, affecting disease susceptibility.^1,2,5−7 Numerous epidemiological and experimental studies have revealed the epigenetic effects of environmental chemicals and their implications in various chronic illnesses such as cancers, reproductive abnormalities, diabetes, obesity, and neurological disorders^2,8−12 Epigenetic mechanisms also can act in genotoxicity,¹³ endocrine disruption,^9,14 and metabolism.¹⁵

Until recently, epigenetic studies have focused more on DNA methylation, which is one of the primary epigenetic mechanisms. However, chromatin and histone modifications also play a pivotal role in epigenetic inheritance across meiosis and mitosis^5,16 and in the regulation of gene transcription through interplay with DNA methylation.¹⁷⁻²⁰ The most prevalent histone modifications affected by environmental pollutants are acetylation and methylation of lysine residues in the amino-terminal tails of histone 3 (H3) and histone 4 (H4).^3,8,20 Especially, post-translational histone methylation marks, such as H3 lysine 9 trimethylation (H3K9me3) and H3 lysine 27 trimethylation (H3K27me3), are linked to the repression of gene transcription, contributing to the formation of heterochromatin. These modifications are site-specifically deposited by histone-modifying enzymes or complexes such as histone methyltransferase (HMTs), histone demethylases (HDMs), and Polycomb repressive complex 2 (PRC2). Studies in higher eukaryotes reveal that only heterochromatin domains capable of inheritance involve H3K9me3 and H3K27me3.²¹

Although epigenetic modifications are common biochemical effects triggered by environmental toxicants,³ there are critical challenges in incorporating epigenetic biomarkers into (eco)toxicological assessments.^12,20,22,23 To utilize epigenetic changes as effective biomarkers for chemical risk assessment, more scientific evidence is needed on several key issues: (1) the causal relationships between epigenetic changes and phenotypic observations at the organismal level; (2) dose–response characterization of chemically induced epigenetic modifications at environmentally relevant concentrations; and (3) differences in the sensitivity of epigenetic responses to chemical exposures across ecological species. Although many agencies and working groups have endeavored to integrate chemical-induced epigenetic modifications into toxicological assessments and frameworks,^3,22,24,25 more concrete evidence of their biological and toxicological significance is required to enhance the applicability of epigenetic biomarkers. The Adverse Outcome Pathway (AOP) framework has been used to integrate toxicological data and illustrates the linkage between a direct molecular initiating event (MIE) and an adverse outcome (AO) for risk assessment. Once a putative AOP is constructed considering weight of evidence (experimented-based or knowledge-based) and the linkage between MIE and AO is validated, the MIE can be used to develop screening assays for toxicants and predict the adverse effects of chemicals.^3,22,25,26 Thus, incorporating epigenetic biomarkers into the AOP can facilitate the assessment of epigenetically modulating chemicals.

Anthropogenic chemicals such as phthalates, flame retardants, and biocides are frequently and ubiquitously found in environmental matrices and humans due to their extensive use in industrial and consumer products and subsequent organic waste. Among these environmental chemicals, bisphenol A (BPA; intermediate, antioxidant, plasticizer), di-2-ethylhexyl phthalate (DEHP; plasticizer), hexabromocyclododecane (HBCD; flame retardant), tetrabromobisphenol A (TBBPA; flame retardant), and triclosan (TCS; antimicrobial) are common additives in various consumer products and plastics.²⁷⁻³⁰ These chemicals have raised significant concerns about their prevalence in the environment and potential adverse effects on human health and wildlife.^27,30−32 Known as endocrine-disrupting chemicals (EDCs), they can mainly cause reproductive and developmental toxicities, interfering with the metabolism of sex hormones, thyroid homeostasis, and the activities of antioxidative and detoxification enzymes, along with other systems.²⁸⁻³³

Caenorhabditis elegans is a suitable model organism for epigenetics studies due to its highly conserved epigenetic regulatory mechanisms and practical experimental features, such as a short life cycle and ease of manipulation, which are more advantageous than mammalian models.^34,35 Despite lacking DNA methylation and the associated enzymes, C. elegans studies have delineated the conservation of chromatin components and histone methylation-related transgenerational effects.^2,35 Furthermore, this model offers the advantages such as sensitivity to various contaminants comparable to that of other model organisms, functional relevance in ecosystems, and the availability of its genome information.^35,36 Given the importance of assessing the intergenerational effects of chemicals on organisms in real environments and the resulting population-level impacts, the strengths of C. elegans have significantly contributed to research on the reproductive, transgenerational, and multigenerational toxicity of environmental chemicals.^5,35,37,38 Consequently, it is gaining recognition as an effective tool for lower tiers of ecological risk assessment (ERA).³⁶

This study aimed to investigate the causal relationship between the reproductive toxicity of environmental chemicals and repressive histone marks in C. elegans. Additionally, we sought to assess the applicability of histone methylation marks in the ecotoxicological assessment of chemicals. To achieve these goals, we initially evaluated the reproductive toxicity and germline desilencing response to prevalent chemicals in consumer products (BPA, DEHP, HBCD, TBBPA, and TCS) using N2 and transgenic NL2507 strains. Subsequently, we measured histone-methylation-modifying enzyme activities, histone methylation levels (H3K9me3 and H3K27me3), and toxicity biomarker gene expressions across multiple concentrations to determine the benchmark concentrations (BMCs). Finally, we validated the association between histone methylation and reproductive toxicity of TCS and TBBPA using enzyme inhibitors in C. elegans and identified other model species likely susceptible to HMT and HDM activities. Figure 1 illustrates the overall workflow of this study.

Figure 1.

The overall workflow of this study. First, we exposed C. elegans to five chemicals commonly found in consumer products to evaluate their effects on reproduction and epigenetic changes. We then conducted an in-depth investigation combining a selective inhibitor assay and POD analysis, focusing on TCS and TBBPA, which displayed high reproductive toxicity and increased histone methylation. Finally, the experimental results were synthesized to construct epigenetic AOPs and identify the taxonomic domains of applicability for epigenetic MIEs.

2. Materials and Methods

2.1. C. elegans Strains and Cultivation

The strains used in this study were wild-type N2 and transgenic strain NL2507 (pkIs1582 [(let-858::GFP; rol-6(su1006)]), originally obtained from the Caenorhabditis Genetics Center (CGC) (University of Minnesota, MN, USA). The C. elegans strains were cultured at 20 °C, on nematode growth medium (NGM) agar plates, and seeded with a lawn of E. coli OP50. Gravid adults grown on NGM plates were collected for synchronization using a hypochlorite solution before exposure to chemicals (the detailed method provided in the Supporting Information).

2.2. Chemical Information and Preparation

Tables S1 and S2 provide information about the five chemicals investigated in this study, including their use category, regulatory status, and classification as emerging contaminants. These chemicals are mainly used as additives and intermediates with the functions of antioxidants, plasticizers, flame retardants, and biocides and are subject to regulation under chemical laws in Korea, Europe, and the United States. Notably, four of these chemicals, excluding DEHP, are identified as emerging contaminants based on environmental monitoring data.

For toxicity testing, chemicals were reconstituted in dimethyl sulfoxide (DMSO) to prepare stock solutions at concentrations 1,000 times higher than the exposure concentrations. These stock solutions were then diluted with complete S-medium containing OP50 to prepare the final chemical solutions. When exposing the chemicals with inhibitors, an inhibitor and chemical stock solutions, which were 2,000 times the exposure concentrations, were used. Both stock solutions were then mixed and diluted to the final exposure concentrations. This process ensured consistent DMSO concentrations (0.1%) across all exposure groups. The concentration of HMT inhibitors (GSK343 and BRD4770) was set at 0.5 μM based on the highest concentration of GSK343 that did not affect 72 h of reproduction (Figure S1).

To confirm the stability of the chemical concentration during exposure, we measured the concentrations in the test medium using gas chromatography–mass spectrometry and high-performance liquid chromatography with diode array detection under specific conditions (Table S3). The chemical concentrations remained stable in the medium for 72 h of exposure, except for TBBPA, which was slightly lower than the nominal concentration up to 24 h but became comparable to the nominal concentration afterward (Table S4).

2.3. Reproduction Assay

Synchronized L1-staged worms were exposed to chemical solutions for 72 h at different concentrations (0, 0.01, 0.1, 1, 10, and 100 μM). Due to the insolubility and precipitation of HBCD at 100 μM, we limited the maximum exposure concentration to 10 μM for HBCD. Chemical solutions were prepared in complete S-medium containing 0.12% OP50. Approximately 200 nematodes were exposed to 1 mL of solution in a 24 well-plate for 50 h. Subsequently, worms from each group were transferred to clean NGM plates. L4 larvae were then picked and exposed to new 100 μL chemical solutions with 0.06% OP50 in a 96 well-plate for 24 h (eight worms per treatment, n = 3). The number of eggs and offspring produced by each adult worm was counted and recorded.

2.4. Measurement of Germline Desilencing, Enzymatic Activity, and Global Histone Methylation

GFP germline desilencing was assessed using the NL2057 strain as previously described.^39,40 Worms were synchronized and incubated on fresh NGM seeded with OP50 bacteria at 20 °C until they reached the L4 stage. They were then exposed to each chemical solution (10 μM) with OP50. After 24 h of exposure, the worms were allowed to settle on NGM plates for 1–2 h. Worms displaying the rolling phenotype were selected for analysis, placed on a slide, and covered with a coverslip. The number of individuals expressing germline GFP in the NL2507 population was counted under the microscope, and the percentage of desilencing was calculated (20 worms per treatment, n = 3). The worms exposed to the negative chemical DMSO (0.1%) exhibited an average GFP expression of 21%. Positive chemical dissolved in DMSO was employed and induced an average expression of 69%. Thus, germline desilencing was scored based on medians between the negative and positive controls (i.e., 45%), with values above marked as “strong” and below as “weak”.

For the measurement of enzymatic activity and global histone methylation, synchronized N2 worms were placed in a 6-well plate containing 3 mL of complete S-medium or a chemical solution with 0.5% OP50. After 72 h, worms were collected into a microcentrifuge tube and immediately frozen in liquid nitrogen. They were then stored at −80 °C until they were used for protein extraction. Additional methodological details are provided in the Supporting Information.

2.5. Gene Expression Analysis

Worms were exposed to different concentrations (0, 0.01, 0.1, 1, 10, and 100 μM) for 72 h under the same conditions used for protein extraction. Total RNA was extracted using an RNA extraction kit (NucleoSpin, MachereyeNagel, Düren, Germany), and RNA purity and quantity were measured using a Nanodrop instrument (NanoReady Touch, Life Real, China). After RNA extraction, gene expression analysis was performed by quantitative reverse transcription PCR (qRT-PCR). cDNA was synthesized by a reverse transcriptase (RT) reaction using Bio-Rad iScript cDNA Kits (Bio-Rad, CA, USA) and then amplified using a thermal cycler (Bio-Rad). Real-time PCR reactions were carried out on a CFX manager (Bio-Rad) using IQ SYBR Green SuperMix (Bio-Rad). Primers used in this study were designed using Primer3plus based on sequences available in NCBI or obtained from previous studies (listed with a description in Table S5). Three biological replicates were used for each qRT-PCR analysis, and the gene expression was normalized to that of actin as a housekeeping gene.

2.6. Calculation of Benchmark Concentration

Benchmark concentrations (BMCs) and lower confidence limits of benchmark concentrations (BMCLs) were calculated using BMDExpress v. 2.3. First, One-Way Analysis of Variance (ANOVA) was performed to select biological end points with significant dose–response behavior using a 1.2-fold change filter. BMCs and BMCLs were then estimated using the linear model with the following settings: maximum interaction, 250; confidence level, 0.95; BMR, 1.349 (10%) standard deviations; p-value cutoff, 0.05; and number of threads, 24.

2.7. Protein Sequence Similarity Assessment

We used the web-based Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS; https://seqapass.epa.gov/seqapass/) tool to calculate the quantitative protein sequence similarity across taxonomic groups and identify susceptible species. Three C. elegans HMTs [MET-2 (CCD73198.2), SET-25(NP_499738.3), and MES-2 (NP_496992.3)] and three HDMs [JMJD-2 (NP_496969.2), JMJD-1.2 (CDH93112.1), and UTX-1(NP_001309685.1)] were used as query sequences for Level 1 analysis to detect ortholog candidates based on primary amino acid sequences. Subsequently, Level 2 analysis was conducted to compare sequence similarity within selected functional domains of the target protein, specifically the SET domain (pfam00856) and JmjC domain, hydroxylase (pfam02373).

2.8. Statistical Analysis

The significance of differences between the groups was analyzed using R version 4.0.4. t tests or Wilcoxon rank-sum tests were used, depending on whether the data met the normality assumption and homogeneity of variances. Significant p-values are indicated with asterisks.

3. Results and Discussion

3.1. Effects of Chemicals on Reproduction and Germline Silencing

The reproductive toxicity of the five chemicals was ranked in the following order: TCS, TBBPA, HBCD, DEHP, and BPA, with TCS and TBBPA demonstrating steeper concentration–response curves (Figure S2). At a 100 μM concentration, worms exposed to TCS and TBBPA produced no offspring (Figure S2) and showed effects on growth from L4 larvae to adults. Therefore, the 100 μM concentration was excluded for further experiments. The hatching rate of eggs produced by adult worms was affected by 10 μM five chemicals after 72 h of exposure, although most eggs from all groups hatched after 120 h. Notably, only HBCD reduced the percentage of hatched eggs up to 96 h (Figures S3A-3C). Additionally, the number of both hatched and unhatched eggs decreased following exposure to DEHP, HBCD, TCS, and TBBPA after 72 h. However, after 96 h, only the number of hatched eggs decreased, with TCS and TBBPA continuing to reduce the number of hatched eggs up to 120 h. This suggests that TCS and TBBPA had a greater impact on the reproductive capacity of the worms than the egg hatching process (Figures S3D-3C).

Using a germline desilencing reporter, we captured the epigenetic impact of chemicals on repressive histone methylation in worms exposed to 10 μM (Figure 2A). The strain NL2507 carries repetitive transgene arrays of a GFP-tagged ubiquitously expressed gene, LET-858 (pkIs1582). This transgene is expressed in somatic cells, but it is transcriptionally silenced in the germline via H3K9me3 and H3K27me3 accumulation.⁴⁰⁻⁴² Germline GFP expression (desilencing) in strain NL2507 indicates that repressive histone marks diminished. BPA and DEHP exposure increased the pkIs1582 array expression, akin to S-adenosylmethionine (SAM) and EZH2 inhibitor 3-Deazaneplanocin A (DZnep) treatment. Conversely, TCS, TBBPA, and HBCD showed a “weak” prevalence of germline desilencing. Following treatment with TCS and TBBPA, the transgene expression remained suppressed to levels comparable to those of the DMSO control (21 ± 1.8%). Worms exposed to HBCD displayed higher germline expression (43 ± 9.6%) than those exposed to TCS and TBBPA. When comparing reproductive toxicity at a 10 μM concentration, the relatively more toxic TCS and TBBPA maintained a silenced transgene array in the germline, while the less toxic BPA and DEHP induced germline desilencing.

Figure 2.

Germline desilencing status, histone methylation, and histone methyltransferase activity in worms exposed to five chemicals. (A) Representative image of desilenced (DZNep; BPA; DEHP) and silenced (DMSO; TBBPA; HBCD; TCS) pkIs1582 array expression in the germline of NL2507 strains after exposure to 10 μM chemical additives. The percentage of worms displaying germline desilencing and the scoring results are shown on the left side of the images. (B) Representative Western blots of H3K9me3 and H3K27me3 in DMSO control and chemical-treated groups. (C) Fold changes in quantitative protein band intensity of H3K9me3 and H3K27me3 in chemical-treated wild type N2 compared to the DMSO control (the intensities of histone methylation markers were normalized to those of histone H3). H3K9 methyltransferase activity (D) and H3K27 methyltransferase activity (E) of the chemically treated groups relative to the control groups. Blue and red asterisks indicate a significant decrease and increase, respectively. All bars are means normalized to DMSO control ± SEM; * p < 0.05, ** p < 0.01, *** p < 0.001.

3.2. Effects of Chemicals on Repressive Histone Methylation

Numerous studies stated that environmental exposure modulates epigenetic modifications by influencing the activities of related enzymes.¹⁷ Some authors suggested that the interference of epigenetic enzyme activities could be potential MIEs for specific adverse outcomes.^25,26,43 Thus, we measured HMT activities and global levels of histone methylation. In a concentration-dependent manner, BPA, DEHP, and HBCD decreased H3K9 and H3K27 methyltransferase activities, while TCS and TBBPA increased them (Figures 2D-2E). For TCS and TBBPA, the most significant increases in HMT activity occurred at 10 μM, where the reproductive toxicity was severe. Protein expression analysis also revealed that TCS and TBBPA exposure heightened repressive histone mark levels (H3K9me3 and H3K27me3) (Figures 2B-2C). In contrast, BPA, DEHP, and HBCD exhibited opposing results. These results were comparable to pkIs1582 array expression in germline, suggesting that the desilencing of pkIs1582 array may serve as a relevant indicator of chromatin mark-regulated transcriptional modulation.

3.3. HMT-Mediated Reproductive Defects in C. elegans Exposed to TCS and TBBPA

Among the five chemicals, TCS, TBBPA, and HBCD displayed higher reproductive toxicity compared with BPA and DEHP. Notably, the alleviation of reproductive defects induced by TCS and TBBPA (10 μM) was evident following exposure to the H3K27-specific inhibitor, GSK343 (Figures 3A-3B). However, this recovery was not observed in HBCD-exposed worms (Figure 3C). As opposed to the effects observed in TCS and TBBPA-treated groups, we speculate that the inhibition of HMT activity by GSK343 can be ineffective in HBCD-exposed worms, given that HBCD exposure reduced the HMT activities and histone methylation levels.

Figure 3.

Effects of the H3K27 and H3K9-specific HMT inhibitors on reproductive toxicity and histone methylation in worms exposed to TCS, TBBPA, and HBCD. C. elegans exposed to TCS, TBBPA, and HBCD were cotreated with 0.5 μM concentration of the H3K27-specific HMT inhibitor GSK343 and the H3K9-specific inhibitor BRD4770 (A-C). Protein expression levels were measured in worms exposed to chemicals alone or a combination of chemicals and the HMT inhibitor (0.5 μM) using Western blots of H3K9me3 and H3K27me3 (D). The symbols * and # denote significant differences from the control group and significant differences between the group exposed to chemicals alone and the group coexposed to an inhibitor, respectively. Data are normalized to the DMSO control and presented means ± SEM; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ## p < 0.01, ### p < 0.001.

The H3K9-specific inhibitor BRD4770 did not alter the reproductive toxicity of TCS, TBBPA, or HBCD (Figure 3). This result was supported by changes in histone methylation levels following inhibitor treatment (Figure 3D). Treatment with GSK343 reduced H3K9me3 and H3K27me3 levels elevated by TCS or TBBPA, exhibiting a more potent inhibitory effect than that of BRD4770. These findings imply the roles of HMT activities and repressive histone modification in the adverse reproductive effects of TCS and TBBPA.

GSK343 mitigated the reproductive toxicity of TCS and TBBPA. In line with the result, we found the attenuation of both H3K9 and H3K27 methyltransferase activities was elevated by exposure to TCS and TBBPA with GSK343 treatment (Figures S4A-4B). GSK343 led to a greater decrease in activity of H3K27 methyltransferase compared to H3K9 methyltransferase, consistent with its role as an H3K27-specific inhibitor. These observations substantiate the correlation between HMT activation, increased repressive histone methylation markers, and the reproductive toxicity induced by TCS and TBBPA.

Both chemicals also elicited concentration-dependent reductions in H3K9-specific HDM activity, which were reversed by cotreatment with GSK343 (Figures S4C-4D). However, there was no statistical difference in H3K27 demethylase activity between the TCS exposure and inhibitor cotreatment groups, except at a concentration of 0.1 μM. For TBBPA exposure, GSK343 decreased H3K27 demethylase activity, contrasting with its effects on H3K9 demethylase activity. These results suggest that the inhibition of H3K9 demethylase activity also contributed to increased histone methylation due to the failure of methyl group removal from histone lysine residues.

Recent studies have emphasized the crosstalk between H3K9 and H3K27 methylations, orchestrated through physical and functional interactions between histone-modifying enzymatic machineries.^44,45 H3K9 methyltransferase G9a can form a superrepression complex with the H3K27 methyltransferase EZH2, leading to dual methylation of H3K9 and H3K27. Another mechanism is that G9a directly methylates H3K9 and indirectly increases H3K27 methylation via upregulating the polycomb-like 3 gene and promoting the chromatin recruitment of PRC2. H3K9 and H3K27 methylations are the well-understood heterochromatin marks involved in genetically distinct but highly conserved pathways of transcriptional repression.³⁴ The C. elegans genome harbors HMTs such as MET-2 (mammalian SETDB1 orthologue) and SET-25 (mammalian G9a/SUV39 SET orthologue) and PRC2-like complex proteins such as MES-2 (EZH2 ortholog). Although the G9a inhibitor BRD4770 did not reverse reproductive toxicity, exposure to TCS and TBBPA increased the activities of both histone-modifying enzymes, and their activities were suppressed by the H3K27-specific inhibitor. Overall, HMT and HDM can be major contributors to the adverse reproductive outcomes.

3.4. Transcriptional Responses of Toxicological Biomarker Genes Altered by Cotreatment with HMT Inhibitor, GSK343

To explore the effect of histone methylation on transcriptional regulation, we analyzed the expression of toxicological biomarker genes in response to TCS and TBBPA alone and in combination with GSK343 (Figure 4A, Tables S6–S7). Among the four genes associated with reproduction and steroidogenesis (see cluster f), vit-1, vit-3, and daf-12 were downregulated by both chemicals. Remarkably, GSK343 exposure upregulated vit-1 and vit-3 expressions, which were reduced in worms exposed to chemicals alone (Figures 4B-4C). Xenobiotic metabolism and stress response genes (clusters d and e) showed similar expression patterns, even though there were differences in levels of gene expressions. Their expression decreased at low concentrations and increased at a concentration of 10 μM, where severe toxicity occurred (Figure 3). For TCS exposure, the increased expression of cyp-35a2 and gst-4 genes, participated in xenobiotic metabolism phases I and II, was mitigated by cotreatment with HMT inhibitor (Figure 4B). TBBPA exposure also upregulated fmo-2 and ctl-2, involved in xenobiotic metabolism and oxidative stress response, and the two genes were downregulated by cotreatment with HMT inhibitor (Figure 4C). These findings suggest that TCS and TBBPA reduce vitellogenin gene expression and induce xenobiotic metabolism and oxidative stress responses, potentially regulated by histone methylation.

Figure 4.

Alterations in the transcriptional response of biomarker genes under cotreatment with the HMT inhibitor GSK343. (A) A heatmap displaying gene expression in worms exposed to TCS or TBBPA and in worms exposed to both individual chemicals and HMT inhibitor GSK343. Fold changes in gene expression compared to the DMSO control were log2 transformed. Genes were categorized according to their functions (a: histone demethylase; b: histone methyltransferase; c: DNA damage repair; d: xenobiotic metabolism; e: stress response; f: yolk protein and steroid binding activity). Asterisks indicate a significant difference between the control group and the treated group (* p < 0.05, ** p < 0.01, *** p < 0.001). (B–C) The gene expression of selected genes recovered by cotreatment with the HMT inhibitor (see the genes in red in Figure 4A). The bars represent the mean gene expression in response to TCS and TBBPA with or without GSK343. Asterisks denote a significant difference from the respective chemical-treated group (* p < 0.05, ** p < 0.01, *** p < 0.001).

However, we found a discrepancy between mRNA expression levels and enzymatic activities, except for jmjd-1.2, which encodes a demethylase for H3K9me2 and H3K27me2. We postulate that TCS and TBBPA directly affect the enzymatic machineries (i.e., H3K9 demethylase and H3K9/H3K27 methyltransferase) (Figures 2 and S4), and the related genes may be regulated as a consequence of change in enzymatic activity. Enzymatic activity does not always correlate with changes in the mRNA expression. Wirbisky-Hershberger et al. identified that atrazine exposure reduced DNA methyltransferase 1 (DNMT1) activity and global DNA methylation levels but not dnmt1 gene expression in zebrafish.⁴⁶ Similarly, Torres et al. also stated that methylation/demethylation fluctuations at the gene level may not necessarily be translated into global patterns of methylation and that reversed effect on gene expression can be induced by compensatory mechanisms.⁴⁷ In our study, we found that chemical exposure decreased HDM activity but increased the expression of HDM-related genes (cluster a) and that the expression of HMT-related genes (cluster b) increased in worms exposed to both TBBPA and an HMT inhibitor. These results suggest that gene expression changes may be accompanied by chemical effects on epigenetic enzymes in a compensatory manner.

Even though the effective range of chemicals varies depending on species and exposure conditions, other studies have reported that TCS or TBBPA affects the transcriptional and biochemical levels of vitellogenin, reactive oxygen species (ROS), and oxidative stress responses in ecological organisms. Exposure to low concentrations of TCS (100–350 nM, mosquitofish) can increase vitellogenin gene expression,³³ while exposure to relatively higher concentrations decreases vitellogenin gene expression (40 μg/L, Daphnia magna)³³ and vitellogenin levels (200 μg/L, female zebrafish), impacting reproductive defects.⁴⁸ Furthermore, TCS induces oxidative stress in various aquatic organisms by elevating ROS levels, glutathione S-transferase (GST) and catalase (CAT) activities, and antioxidant-related gene expressions.^33,49,50 Lenz et al.⁵¹ also found that TCS caused reproductive impairment in C. elegans by inducing oxidative stress response and germline toxicity.

For TBBPA exposure, vitellogenin gene was upregulated in mosquitofish,⁵² which was a result opposite to that observed in our study. However, consistent with our findings, other researchers documented that oxidative stress was induced in C. elegans exposed to TBBPA. Liu et al.⁵³ demonstrated that exposure to TBBPA (10 and 100 μg/L) significantly increases ROS production and the expression of stress-related genes (sod-3, ctl-1, ctl-2, cyp-35a2, etc.). Specifically, loss-of-function mutations in ctl-2 caused a greater reduction in head thrash and body bend in response to TBBPA exposure compared to wild-type N2. Another study also identified that neurotoxicity and reproductive toxicity were caused by TBBPA, accompanied by elevated levels of biochemical indicators of oxidative stress and stress-related gene expressions.⁵⁴

Vitellogenins mainly function to transport lipids and micronutrients from adult tissues to oocytes. In C. elegans, vitellogenins are highly expressed, and abundant vitellogenesis can support postembryonic phenotypes and fertility in deleterious environment.⁵⁵ This gene regulation is influenced by various signaling pathways and physiological or environmental factors including oxidative stress. Additionally, oxidative stress and free radicals affect histone post-translational modifications. Particularly, histone H3 is vulnerable to the effects of oxidative stress-related mechanisms.⁵⁶ Concerning histone methylation, superoxide and hydroxyl radicals can inhibit HDM activity by oxidizing Fe(II) to Fe(III) in their catalytic center, leading to hypermethylation.⁵⁶

For example, the jumonji family proteins, a group of histone lysine-specific demethylases, demethylate histones via an oxidative mechanism, requiring Fe(II) and α-ketoglutarate as cofactors. However, the redox state of Fe at the enzyme active sites can interact with H₂O₂, thereby inhibiting their demethylation activities. In an indirect manner, glutathione (GSH) metabolism can affect the bioavailability of SAM, which is a precursor of GSH and a principal methyl donor.⁵⁶ Taken together, TCS and TBBPA-induced oxidative stress can trigger a change in vitellogenin gene expression and histone methylation as observed in this study.

However, a notable finding of this study is that the HMT inhibitor restored vitellogenin gene expression, which was reduced by chemical exposure. In honey bee (Apis mellifera L.), exposure to RG108, a DNMT inhibitor, resulted in decreased DNA methylation, increased expression of vitellogenin gene and protein, and increased lifespan, suggesting possible epigenetic regulation of vitellogenin.⁵⁷ Another study reported that BPA exposure in zebrafish increased the expression of the follicle stimulating hormone receptor (fshr) gene and reduced H3K27me3 enrichment in fshr, which influences the synthesis of 17β-estradiol and vitellogenin, thereby leading to oocyte growth.⁵⁸ Based on these results, we infer that TCS and TBBPA may directly or indirectly inhibit vitellogenin gene transcription via repressive histone marks, H3K27me3 and H3K9me3, by perturbing HTM/HDM activity.

3.5. Building a Putative Epigenetic AOP for TCS and TBBPA Using BMC Analysis

We compared the BMCs of TCS and TBBPA for enzyme activity, gene expression, and reproductive toxicity (Tables S8 and S9). This comparative analysis supports that epigenetic enzymatic changes were more sensitive than reproductive end points and some gene expressions, suggesting that these epigenetic modifications have the potential to be used as early biochemical markers for apical outcomes. The TCS-exposed group showed significant concentration–response behavior in the expression of 12 biomarker genes, three enzyme activities, and reproductive toxicity, with the HDM activity being the most sensitive (Figure 5A). Similarly, the TBBPA-exposed group displayed significant concentration–response in the expression of 15 genes, four enzyme activities, and reproductive toxicity, with HDM/HMT activity being the most sensitive, followed by gene expression and reproductive toxicity (Figure 5B). Notably, vit-1 and vit-3 expressions, which were restored by GSK343, exhibited changes in both TCS- and TBBPA-exposed groups at concentrations exceeding those affecting the enzymatic activity. Based on our findings, we propose a putative epigenetic AOP, considering biological plausibility, the essentiality of KEs, and empirical evidence (Figure 5C). The essentiality of events was determined by whether chemically induced effects were reversed by the HMT inhibitor, and empirical evidence guided the positioning of upstream and downstream events within the AOP.

Figure 5.

The benchmark concentration (BMC) comparison between biological end points and the putative epigenetic AOP for histone methyltransferase-mediated ecotoxicity of TCS and TBBPA in C. elegans. Comparative analysis of the BMCs of TCS (A) and TBBPA (B) for enzymatic, transcriptional, and apical end points. The range plots present lower and upper confidence limit values. The number of genes and enzyme activities with statistically significant concentration–response behavior are depicted for each chemical. (C) The putative epigenetic AOP, constructed considering biological plausibility, the essentiality of the events, and empirical evidence. Essentiality was determined based on whether the effects of the chemical were reversed by the HMT inhibitor, and empirical evidence was used to position upstream and downstream components within the AOP based on BMCs (refer to Tables S8 and S9). Solid boxes and arrows represent events and relationships identified experimentally in this study, while dashed boxes and arrows represent events and relationships expanded by reviewing other literature data (refer to the Results and Discussion section) and AOP wiki data (refer to Figure S5). The KE IDs that exist in the AOP wiki (https://aopwiki.org/) are shown in each box. MIE: molecular initiating event; KE: key event; AO: adverse outcome; BMC: benchmark concentration.

3.6. Cross-Species Extrapolation Using SeqAPASS

Histone-modifying enzyme activities for H3K9 and H3K27 methylations were identified as the most sensitive markers. We further evaluated the protein sequence similarity of these enzymes by using the SeqAPASS tool. HMTs typically feature a conserved catalytic Su(var.)3–9, Enhancer of zeste, and Trithorax (SET) domain, which contains the SAM-binding sites. HDMs with Jumonji C (JMJC) domain have dioxygenase activity, acting on all trimethylation states.³⁴ Thus, we focused on the well-understood C. elegans SET domain-containing HMTs (SET-25, MET-2, and MES-2) and JmjC domain-containing HDMs (JMJD-2, JMJD-1.2, and UTX-1).

Both Level 1 and Level 2 analyses revealed high protein sequence similarities and a number of orthologs across taxa. Among the six proteins, MET-2, MES-2, JMJD-2, and UTX-1 have more orthologs than SET-25 and JMJD-1.2 (Tables S10 and S11). In addition to wildlife species, 167 common model species for ecotoxicological assessment were identified as susceptible species with high protein similarities to C. elegans (Figures S6, S7, and 6). A list of model organisms included in each taxonomic class is provided in Table S12. Actinopteri contained the most species across all four proteins (Figure 6A). Actinopteri, Bivalvia, Branchiopoda, Eurotatoria, Hexanauplia, Collembola, and Insecta were common classes across all enzymes (Figure 6B and Table S13). Collectively, protein similarity prediction expands the tDOA, indicating the possibility that epigenetic enzymatic responses to TCS and TBBPA may be sensitive in different species with analogous sequences and functions of proteins.

Figure 6.

Analysis of quantitative functional protein domain similarity for C. elegans histone methyltransferase and histone demethylase across common model species. (A) Distribution of model species across enzymes and biological classes. The numbers indicate the number of model species included in each node. The color gradient of nodes represents a distribution rate of model species within each level of the tree. (B) Mean protein sequence similarity of model species by biological class. The dots are colored according to the dominant habitat of the species.

3.7. Potentials and Perspectives for Ecotoxicological Assessment

As outlined in the Introduction, utilizing epigenetic changes as early markers of transgenerational effects for ecotoxicological assessment requires further research on their causal link to individual-level effects, dose–response characterization, and applicability across ecological species. This study addresses these issues by identifying the causality between chemical-induced reproductive toxicity and repressive histone methylation and by characterizing concentration–response relationships across biochemical, transcriptional, and organismal end points in C. elegans. Our findings propose disturbance of HMT and HDM activity as the potential MIEs of TCS and TBBPA toxicity. Additionally, the study expands the tDOA for the MIE by identifying model species with high enzyme protein sequence similarity to C. elegans, suggesting potential sensitivity in HMT and HDM activities across taxa. The combined approach of selective inhibitor assays, BMC analysis, and protein conservation assessment aids in establishing a direct or indirect link between epigenetic changes and transcriptional and phenotypic responses and developing AOPs. However, further investigations are necessary to strengthen this epigenetic AOP in ecotoxicological assessments. These include the following:

• (1)Assessing repressive histone methylation at vitellogenin gene regions: A site-specific histone methylation analysis via chromatin immunoprecipitation (ChIP) can pinpoint whether transcriptional repression of vitellogenin genes is specifically due to methylation or other molecular perturbations.
• (2)Integrating additional toxicological data on TCS and TBBPA into the AOP: TCS and TBBPA are endocrine disruptors that together cause oxidative stress, inflammatory response, and metabolic disruption in ecological model species.^{32,33,48,49,53,54,59} The epigenetic and transcriptional changes identified in this study may be linked to other toxicological mechanisms of these chemicals. Biochemical molecules involved in methylation (e.g., SAM, folate, α-ketoglutarate, etc.) could also be integrated into the AOP.
• (3)Multigenerational observation: Heterochromatin components are major carriers of stable epigenetic information across cell divisions or generations.⁵ Additionally, parental methylated nucleosomes can be randomly distributed to the daughter molecules during DNA replication and generate newly methylated ones through the propagation of methylation (from segregated old histones to neighboring histones) by chromatin binding proteins and HMTs.^5,21,60 Thus, changes in repressive histone marks can contribute to the adaptive and transgenerational effects of TCS and TBBPA.^{33,49,53,54,59} Beyond a single generation, it is necessary to assess if chemical-induced histone methylation alterations persist across multiple generations and influence individual or population-level characteristics.

4. Environmental Implication

Chemicals in consumer products are ubiquitously distributed in the human body and the environment due to their widespread use. Among these chemicals, EDCs constitute a significant proportion and have a high potential to impact the reproductive system and epigenetic states. This study proposed a qualitative and quantitative approach to evaluate the causality between epigenetic markers and reproductive toxicity induced by chemicals using the alternative model organisms C. elegans. This approach involves assessing the recovery of downstream events following exposure to enzyme inhibitors and identifying sensitive end points through comparison of toxicity PODs. Additionally, the experimentally derived results were integrated into the AOP framework, expanding the tDOA of epigenetic MIEs. Ultimately, this study highlights the integration of epigenetic mechanisms into prospective ERA of chemicals by proposing “HMT/HDM enzyme activity” as an effective screening tool, through elucidating the association of repressive histone marks with reproductive toxicity.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.4c04061.

• Materials and methods for C. elegans synchronization, enzymatic activity assays, and Western blot; chemical information on products, regulatory status, and nominal and measured concentrations; gene primers used for quantitative RT-PCR and relative mRNA expression results; effect of histone methyltransferase inhibitors on reproduction; reproduction and hatching rates in C. elegans exposed to BPA, DEHP, HBCD, TCS and TBBPA; effects of the H3K27-specific histone methyltransferase inhibitor (GSK343) on enzymatic activities in C. elegans exposed to TCS and TBBPA; calculated benchmark concentrations of TCS and TBBPA; existing adverse outcome pathways from AOP Wiki; ortholog detection based on protein sequence similarity; quantitative functional protein domain similarity for histone methyltransferases and demethylases across taxonomic groups; model organisms with high sequence similarity to the functional domain sequences of C. elegans histone methyltransferases and demethylases (PDF)

Author Contributions

Jiwan Kim: Conceptualization, Investigation, Methodology, Software, Data curation, Visualization, Writing–original draft, Writing–review and editing. Jinhee Choi: Conceptualization, Methodology, Funding acquisition, Writing–review and editing, Supervision. All authors have read and approved the manuscript.

This work was supported by the National Research Foundation of Korea grant (NRF-2020R1A2C3006838) funded by the Ministry of Science and ICT, South Korea.

The authors declare no competing financial interest.