In early development, embryonic and fetal programming orchestrates finely tuned processes that help establish long-term health and wellbeing. During this time, adverse events, such as exposure to environmental contaminants (Heindel et al, 2015), may disrupt these processes, leading to a higher risk of health effects later in life. This hypothesis, called the Developmental Origins of Health and Disease (DOHaD) (Barker, 2007), has been widely studied in epidemiological and animal models. The biological processes behind the hypothesis, however, are poorly understood.

One process that may underly the DOHaD hypothesis is disruption to early epigenetic programming (Barouki et al, 2018). The epigenome can be defined as mitotically (and sometimes meiotically) heritable marks that help regulate gene expression without altering the genome itself (Murrell et al, 2005). These can include marks on nucleotides (e.g., DNA methylation, hydroxymethylation), histone modifications (e.g., acetylation, phosphorylation, ubiquitination), or noncoding RNA molecules, such as long noncoding RNA (lncRNA) (Greally, 2018). One of the most studied epigenetic marks is DNA methylation. DNA methylation is established early in life. In development, the ova and sperm methylome are erased immediately after fertilization and gradually rewritten throughout gestation (Monk et al, 1987). A secondary phase of fetal programming takes place during primordial germ cell development and migration (Seisenberger et al, 2012). The writing of DNA methylation occurs via a cyclical process: first methylation marks are added to the nucleotides (often cytosines upstream of guanines, CpGs, when methylated, 5-mC) by DNA methyltransferases (DNMTs). Methyl marks can then be oxidized to hydroxymethylation (5-hydroxymethylcytosine, or 5-hmC) by ten-eleven translocation (TET) dioxygenases (Tahiliani et al, 2009), which can be further modified to 5-formylcytosine (5-fC) and 5-carboxylcytosine (5-caC). Both 5-fC and 5-caC can be removed and replaced with a naked cytosine via base excision repair (Moore et al, 2013).

DNA methylation and hydroxymethylation are stable and present throughout tissues in the body (Globisch et al, 2010; Li and Liu, 2011; Wu et al, 2011; Nestor et al, 2012) and likely have opposing regulatory effects on gene expression (Wu and Zhang, 2017). Traditionally, DNA methylation has been thought to have a gene “silencing” effect, but the exact effects are dependent on where the modifications are in the genome (Jones, 2012). Hydroxymethylation plays an essential role in normal development (Yan et al, 2023), helping regulate development in both the heart (Greco et al, 2016) and brain (Stoyanova et al, 2021). Especially in promotor regions, it may reverse the effects of methylation (Mellén et al, 2017). Throughout the genome, it can also act as a recruiter and signal for other epigenetic factors (Takai et al, 2014). While methylation and hydroxymethylation are inherently linked, the developmental programming of these marks do have some independence (Amouroux et al, 2016; Lopez et al, 2017; Yan et al, 2023). However, most environmental exposure studies evaluating the methylome and hydroxymethylome do not use methods that differentiate between these marks; instead, studies typically report on combined “total methylation” (Booth et al, 2012). Because adverse events during these processes could disrupt programming and alter the methylome and hydroxymethylome independently, it is essential to understand the unique response of each of these epigenetic marks.

Some environmental contaminants, such as the group of plasticizers known as phthalates, have developmental effects that have been recently identified. Exposure to phthalates is nearly universal (Woodruff et al, 2011; Zota et al, 2014) and has been linked with endocrine disrupting effects and an increased risk of metabolism and neurodevelopmental disorders and diseases (Braun, 2017). Other common contaminants, such as the metal lead (Pb), have been widely known as developmental toxicants for decades. Early life exposure to Pb can occur via drinking water, contaminated soil, or dust, which can subsequently disrupt brain development, slow growth, and impact the immune system (Bellinger et al, 2017). In animal models, other contaminants, like the plastic additive bisphenol A (BPA) may cause changes in hydroxymethylation (Kochmanski et al, 2018). Developmental exposure to both phthalates (Svoboda et al, 2020; Hsu et al, 2021; Liu et al, 2021) and Pb (Dou et al, 2019b; Wang et al, 2020; Hong et al, 2021; Svoboda et al, 2021) have been linked to differences in the epigenome-wide total methylation, but little is known about effects in the hydroxymethylome.

To understand differences in the hydroxymethylome after developmental exposure to both the phthalate, di(2-ethylhexyl) phthalate (DEHP), and Pb, a longitudinal mouse model was used. This study was conducted as a part of the National Institute of Environmental Health Sciences (NIEHS) Toxicant Exposures and Responses by Genomic and Epigenomic Regulators of Transcription II (TaRGET II) Consortium, which aims to determine how the environment affects disease susceptibility across the life course through changes to the epigenome. We hypothesized that perinatal DEHP and Pb exposure would result in tissue- and sex-specific changes in DNA hydroxymethylation in adulthood.

While the epigenome is comprised of many different modifications and molecules, there has been a strong focus on DNA methylation in studies on the effects of exposures to common chemicals. However, these studies typically use methods that do not differentiate between types of DNA modifications, even though methylation and hydroxymethylation act in biologically opposite ways. In this study, we found that the genome-wide mouse hydroxymethylome is affected by developmental exposure to both DEHP and Pb, with differences in hydroxymethylation observed in adulthood.

In the brain, hydroxymethylation accounts for 33%–50% of DNA modifications, a much higher proportion than in other tissues (Cui et al, 2020). This high occurrence may indicate its importance in normal brain function, such as memory formation (Kremer et al, 2018). Brain hydroxymethylation also plays a role in response to injury (Morris-Blanco et al, 2019; Madrid et al, 2021; Moyon et al, 2021) and oxidative stress (Delatte et al, 2015), likely in a region-specific manner (Doherty et al, 2016). Its role in injury and oxidative stress responses may also be why hydroxymethylation across the body appears to be more sensitive to environmental assaults than other epigenetic marks (Chatterjee et al, 2020).

Exposure to DEHP or other phthalates has been associated with differences in bulk measures of hydroxymethylation in both human urine (Pan et al, 2016) and rat testes (Abdel-Maksoud et al, 2015). Presently, we found that perinatal exposure to DEHP was associated with later-in-life differences in region-specific blood hydroxymethylation in males and females and brain hydroxymethylation in males. There were much fewer regional differences in blood compared to brain, which could be related to the brain’s high levels of hydroxymethylation compared to other tissues. Perinatal DEHP exposure may also alter hydroxymethylation in a tissue specific manner, like with total DNA methylation in our same model (Wang et al, 2020; Liu et al, 2021; Svoboda et al, 2021). In the male brain, there were several regions in gene promoters responsible for microtubule control/cell division (Kif2b), cell transport (Slc12a7), neural-specific endoplasmic reticulum and Golgi functions (Creg2), and potassium channel function (Kctd19). These differences were also linked to gene ontology pathway enrichment in general cellular processes and in pathways related to androgen signaling and bone development. In humans, disruptions in androgen signaling have been one of the primary health effects of concern after developmental DEHP exposure. High exposures have been associated with decreased anogenital distances in males (Swan et al, 2005; Li and Ko, 2012) and long-term changes in growth and metabolism (Tsai et al, 2016; Tsai et al, 2018). More recently there have been concerns about bone growth and development in both animal models (Bielanowicz et al, 2016; Chiu et al, 2018) and humans (Heilmann et al, 2022).

Pb-exposed females showed no differences in hydroxymethylation; only Pb-exposed male brains had increases in hydroxymethylation in the brain. Hydroxymethylation differences after Pb exposure have been explored in various human tissues, including childhood blood (Rygiel et al, 2021), cord blood (Sen et al, 2015; Okamoto et al, 2022), toenail and placenta (Tung et al, 2022), and sperm (Zhang et al, 2021). Epigenome-wide differences were only assessed in toenails/placenta using the EPIC array (Tung et al, 2022) and in sperm using hMeDip-seq (Zhang et al, 2021). In toenails and placentas, most of the differentially hydroxymethylated sites were higher with Pb exposure (Tung et al, 2022), whereas sperm showed mostly lower hydroxymethylated regions (Zhang et al, 2021). Presently, all male cortex regions had higher hydroxymethylation with Pb exposure. All three epigenome-wide studies (Tung et al, 2022; Zhang et al, 2021, and the present study) identified differential hydroxymethylation in calcium genes or pathways. Because Pb is a bioanalogue of calcium, these similarities are expected. These three studies also consistently reported differences in pathways related to nervous system development and synapse function, even in non-neural tissues. This link should be further explored, as small, sparse, and local differences in hydroxymethylation have been associated with gene expression (Marion-Poll et al, 2022), potentially representing a “fine tuning” mechanism of gene regulation that is linked with transcription factor recruitment (Lercher et al, 2014). Because the brain is the primary target organ of Pb toxicity, collective results may be revealing differences in hydroxymethylation patterns that underlie the link between developmental Pb exposure and later life neurotoxic effects.

In Pb-exposed males, an imprinted gene (Rian) and several lncRNA genes were identified as differentially hydroxymethylated. Genomic imprinting is an epigenetically regulated process in which a gene is expressed from one allele in a parent of origin-specific manner. This phenomenon occurs in nearly 1% of the protein coding genes and includes maternally-expressed genes that are paternally imprinted, or vice versa (Barlow and Bartolomei, 2014; Kanduri, 2016). Imprinted genes are typically clustered to form imprinted domains, which also include the genetic code for at least 1-2 lncRNA. Imprinted domains often also contain gametic differentially methylated regions, one of which controls the entire domain to serve as an imprinting control region. The Rian lncRNA is part of a large imprinted domain on chromosome 12 (in mice). The human ortholog, MEG8, regulates vascular smooth muscle cell proliferation, migration, and apoptosis via miRNA interactions with the peroxisome proliferator activator receptor alpha (PPARα) (Zhang et al, 2019). Differential hydroxymethylation of imprinted genes was also identified in mice after developmental exposure to BPA (Kochmanski et al, 2018), which may confer broader patterns of disruption in imprinted genes after developmental exposure to common environmental contaminants.

While most identified regions were unique between exposures, there were four genes that overlapped between the DEHP and Pb in the male cortex, including Camkmt, Galnt2, Ccdc192, and Asic2. Camkmt encodes for a methyltransferase that assists in calcium dependent signaling, Galnt2 encodes for a glycotransferase linked with metabolism functions, Ccdc192 encodes for a long-noncoding RNA, and Asic2 encodes for a protein in an ion channel, with high prevalence in the central nervous system. In the same cohort of mice from the present study, differential total methylation was identified in the Galnt2 gene in heart tissue after DEHP and Pb exposure (Svoboda et al, 2021). Differential total methylation of the human ortholog of Galnt2 was also identified in a human cohort exposed to Pb (Svoboda et al, 2021), representing either a potential biomarker of exposure or, due to the Galnt2’s large size, general epigenetic differences that broadly occur after environmental exposures. There was also one gene that overlapped in brain/blood across exposures–Phactr1, a gene that encodes for a phosphatase that regulates the actin cell structure. These overlapping genes may represent common areas of the hydroxymethylome that are particularly sensitive to environmental toxicants. Alternatively, as hydroxymethylation generally confers genome instability (Supek et al, 2014), these could be chance overlaps due to tertiary genome structure or other factors.

Hydroxymethylation is different between sexes and across ages, typically increasing throughout development (Cisternas et al, 2020). In our study, there were large differences between sexes, which may be a due to normal sex differences or due to sex differences in responses to environmental exposures. Additional studies with larger sample sizes and other model strains/species should be conducted, and the inclusion of physiological and behavioral endpoints should be emphasized, especially given the growing body of evidence on sex-specific responses to toxicants (Gade et al, 2021). Our smaller sample size could also result in false negatives. Validation studies with larger sample sizes should aim to include multiple developmental timepoints and tissues to understand hydroxymethylation differences in varying ages and across the body. Future validation studies could also consider PCR or nanopore techniques to confirm hMeDIP-seq results. Because we didn’t measure individual animal dietary intake of either DEHP or Pb, future studies should also measure toxicant exposures in offspring to better estimate dose-response relationships. Additionally in the brain, it is important to consider the impacts of non-cytosine hydroxymethylation (Ma et al, 2017), the parallel differences in methylation, and the cell-type specific patterns in epigenetic modifications which were not presently assessed.

Overall, this study demonstrates that sex-specific hydroxymethylation is different in adulthood in response to developmental environmental exposures, in not just the brain (specifically cortex), but also blood. The differences, however, were limited and should be validated with future work. Results were observed using a method that is able to detect regional differences across >95% of the genome, mostly in areas with lower cytosine-guanine base density (Beck et al, 2022). Because environmental epigenetic studies have traditionally only used methods that capture the collective DNA modifications (e.g., methylation + hydroxymethylation) in high density cytosine-guanine areas of the genome, many environmental effects in the hydroxymethylome may be masked. Future research needs to distinguish DNA methylation and hydroxymethylation functions and responses to environmental exposures, with the aim of revealing potential biomarkers or interventions in the exposure-disease pathway. The hydroxymethylome has had some promising studies in these areas already (Kim et al, 2018; Morris-Blanco et al, 2019; Morris-Blanco et al, 2021). Going forward, its role and responses to the environment should be a major research focus, sharing the spotlight with the diverse modifications and molecules of the epigenome.