Epigenetic Regulation of ADME Genes during Development

Among the earliest investigations of epigenetic regulation of ADME genes were studies of the developmental trajectory of rat CYP2E1 expression conducted approximately 25 years ago (Umeno et al., 1988). Although not expressed in fetal hepatocytes, activation of CYP2E1 occurred within hours after birth. Studies using the cytosine methylation-dependent restriction endonucleases, HhaI and HpaII, associated the rapid postnatal increase in mRNA expression with DNA demethylation at the 5′-region of the gene (Umeno et al., 1988). In the mid-1990s, similar studies of CYP2E1 ontogeny in humans implicated progressive demethylation of specific CpG residues in exon1 and intron 1 with postnatal increases in CYP2E1 mRNA expression throughout the neonatal period (Vieira et al., 1996). Low levels of CYP2E1 mRNA in human fetal, neonatal, and adult kidney and lung tissue were also associated with extensive methylation of the same region of the gene (Vieira et al., 1998). A similar phenomenon has been reported for postnatal activation of Cyp1a2 in mice (Jin et al., 2004).

A more complex picture has begun to emerge with respect to regulation of the CYP3A developmental trajectory. The ontogeny of the CYP3A locus in humans is largely characterized by a “switch” from predominantly CYP3A7 expression in fetal hepatocytes to CYP3A4 expression in adult liver, whereas CYP3A5 expression is relatively stable throughout development, at least in those individuals with genotypes compatible with functional activity (Schuetz et al., 1994; Lacroix et al., 1997; Hines, 2008). Investigation of the mechanisms underlying this developmental pattern of expression has tended to focus on trans-acting factors, such as the age/developmental stage-specific change in the ratio of alternative transcripts of CCAAT/enhancer binding protein β (C/EBPβ), known as the LAP/LIP ratio, which function as transcriptional activators and repressors (Descombes and Schibler, 1991), and the nuclear receptors pregnane X receptor and constitutive androstane receptor (Vyhlidal et al., 2006). More recently, epigenetic regulatory mechanisms involving specific patterns of histone methylation have been demonstrated for an analogous switch from Cyp3a16 in neonatal mouse liver to Cyp3a11 in adult mouse liver (Li et al., 2009). DNA hypermethylation within the murine Cyp3a locus was not observed at any developmental stage; however, data recently have been presented indicating that silencing of CYP3A4 in human fetal liver is associated with DNA hypermethylation of individual CpG sites in close proximity to key regulatory elements, such as CLEM4, HNF4α, XREM, and C/EBP sites within the CYP3A4 regulatory region (Kacevska et al., 2012b). In addition, microRNA mmu-miR-298, which is decreased in wild-type mouse livers after puberty but increased in CYP3A4-transgenic mouse livers after puberty, has been shown to suppress CYP3A4 gene expression through direct and indirect targeting (Pan et al., 2009). In this issue, Xie et al. (2013) implicate a role for microRNA in cytochrome P450 (P450) ontogeny in their investigation of sexually dimorphic expression of Cyp2b9. For example, when the ability to generate microRNAs was compromised by knocking out Dicer1, the age-dependent decline in Cyp2b9 expression was abolished, providing indirect evidence for microRNAs and regulation of Cyp2b9 ontogeny.

Epigenetic Regulation of Tissue-Specific Expression of ADME Genes

DNA methylation and histone modifications represent important mechanisms for tissue-specific gene expression; it has even been suggested that it may be possible to identify specific cells/tissues based on their unique DNA methylation profiles (Shiota, 2004). In general, hypomethylation of a gene promoter region combined with expression of tissue-specific regulatory factors is permissible for driving cell- or tissue-specific gene expression. For example, experiments involving demethylation of cytosine residues and overexpression of C/EBPδ, a lung-enriched form of C/EBP, implicate epigenetic modulation of C/EBP binding sites as a mechanism underlying the expression of human CYP2A13 in the olfactory mucosa of transgenic mice and in human lung cancer cells (Ling et al., 2007). A series of studies investigating kidney- and liver-specific expression of multiple transporters reveals a pattern of tissue-dependent differentially methylated regions (T-DMRs) that influence the interaction of hepatocyte nuclear factor 1 (HNF1). HNF1 is an integral transactivating factor for organic anion transporters (OATs) in liver and kidney, with its cognate binding site in the promoter region of targeted transporters. In the case of amino acid transporters and OATs, hypomethylation of Hnf1α binding sites leads to transporter expression in kidney whereas hypermethylation silences gene expression in the liver. In contrast, Hnf1α binding sites are hypermethylated in cerebrum and are not available to drive transporter expression in that tissue (Kikuchi et al., 2010). The opposite scenario (hypomethylated T-DMRs in liver and hypermethylated in kidney) is operative for liver-specific expression Oatp transporters (Imai et al., 2013). In this issue, Oda et al. (2013) provide evidence that decreased expression of UGT1A1 in kidney is a function of decreased binding of HNF1α due to both DNA hypermethylation and histone hypoacetylation, implicating multiple levels of epigenetic regulation in tissue-specific expression of UGT1A1.

Epigenetic Regulation of Sexual Dimorphism

Sexual dimorphism refers to phenotypic differences between males and females of the same species. Studied extensively, sex differences in the expression of P450s and other drug-metabolizing enzymes in liver have been attributed to the differences in the temporal pattern of growth hormone release by the pituitary (pulsatile in males and continuous in females) that are more pronounced in rodents than in humans (Waxman and Holloway, 2009). Although the results of early studies of male-specific CYP2C11 expression in liver implied a role for epigenetic regulation (Ström et al., 1994), Yokomori et al. provided more definitive evidence for epigenetic regulation by demonstrating that the Cyp2a4 promoter was differentially demethylated in female CD-1 mice, and the Cyp2d9 promoter was preferentially demethylated in male mice (Yokomori et al., 1995). Furthermore, binding of GABP was sensitive to the methylation status of the Cyp2d9 promoter, confirming the association between promoter methylation status, transcription factor recruitment, and sex-specific expression (Yokomori et al., 1995). In this special issue, Xie et al. report that several microRNAs are associated with the hepatic expression of Cyp2b9 in female C57Bl/6J mice, and that the expression of these microRNAs is lower in female mice than in male mice. Furthermore, Cyp2b9 expression is increased in male Dicer1 knockout mice, providing evidence that regulation by microRNA can be added to the list of mechanisms underlying sexually dimorphic P450 expression.

Contribution of Epigenetics to Interindividual Variability in Drug Disposition and Response

Interindividual variability in drug biotransformation has been subject to extensive investigation at multiple levels, including genetic and nongenetic factors as well as multiple in vitro and in vivo approaches (Zanger and Schwab, 2013). In addition to variations in the regulatory and coding regions of P450 and other genes themselves, genetic variation in transcription factors, such as pregnane X receptor and constitutive androstane receptor (Chai et al., 2013), and in the electron donor cytochrome P450 oxidoreductase (Pandey and Flück, 2013) also contributes to interindividual variability in ADME gene expression and function. There is also growing interest in studying epigenetic factors, which may be additional causes of large variations in drug disposition and response. As an example of how epigenetic regulation of ADME genes may influence drug response, this issue’s report by Pan et al. (2013) provides data linking hsa-miR-1291-directed downregulation on ABCC1 to an increase in the intracellular accumulation of doxorubicin, which was accompanied by an increase in chemosensitivity to the effects of the drug. From a more clinical perspective, the current hope/hype is that changes in epigenetic modifications and other factors may account for some of the interindividual variability in epigenetic markers, either as a consequence of genetic variation, environmental exposures (Choudhuri et al., 2010), or other factors. In doing so, these factors may also account for some of the “missing heritability” observed in large population studies (Mill and Heijmans, 2013). Epigenetic mechanisms therefore may be used along with genetic, environmental, and other factors to guide drug and dose selection toward an improved treatment outcome.

Although a compelling argument can be made for aberrant DNA methylation in the pathogenesis of various cancers (Baylin and Jones, 2011), a causal role for epigenetic dysregulation of cellular processes leading to complex diseases is less well established. A quick review of the list of studies supporting a role for epigenetic factors in the regulation of ADME genes (Kacevska et al., 2011) reveals that the majority were conducted in tumor cell lines. To begin exploring the role of epigenetic mechanisms in vivo, Rieger et al. (2013) describe the variability in expression of 56 microRNAs selected on the basis of their potential relevance for regulation of ADME genes in a panel of 92 (40 male and 52 female) human liver samples. They identified a set of six microRNAs that demonstrated considerable (1000- to 30,000-fold) variability in the samples, and another set of four microRNAs that varied <10-fold. Furthermore, significant correlations between microRNA expression and age as well as disease processes (e.g., inflammation and cholestatic liver disease) were observed; no significant correlation between expression and sex were observed. However, attempts to validate previously reported associations (Yokoi and Nakajima, 2013) between various microRNAs and CYP mRNA, protein, and activity levels with this dataset met with varying degrees of success. This finding serves as a reminder that the results from in vitro studies conducted using optimized cell-based model systems may differ from “population-based” studies; the latter involves more confounding factors, such as disease processes, exposure to concurrent medications, lifestyle factors (e.g., smoking, alcohol use), and other environmental exposures. In other words, study designs best suited for mechanistic insights come at a cost in terms of generalizability to populations of human patients. On the other hand, studies designed to characterize population variability may result in hypothesis-generating associations, whereas they are limited with respect to apparent or casual associations. Both types of studies nevertheless provide valuable information that must be interpreted in the proper context.

Epigenetic Regulation of ADME Genes in the Adaptive Response to Xenobiotic Exposure

In this issue, Ramamoorthy et al. (2013) pursue sources of interindividual variability in ADME gene expression further by characterizing the induction of microRNA expression by rifampin in a set of seven primary human hepatocyte cultures. In contrast to the targeted approaches used by Rieger et al., this group adopted a global profiling strategy and identified 33 microRNAs that were upregulated by rifampin and a set of 35 microRNAs that were downregulated by rifampin. As expected, they observed negative correlations between various microRNAs and expression of some P450s (including orphan P450s), reflecting suppression of target gene expression by microRNA. Nevertheless, some positive correlations were also noted, indicating that changes in the expression of microRNAs in response to perturbing factors, such as enzyme-inducing agents, might have indirect effects on gene expression by affecting multiple components of the gene regulatory machinery and thereby influence downstream phenotypes. The studies by Rieger et al. (2013) and Ramamoorthy et al. (2013) provide a glimpse of the complexity of microRNAome in a highly dynamic system among which the diseases (e.g., inflammation and cholestasis) and other perturbing factors (e.g., concurrent medications) have the potential to alter the microRNA repertoire and might further influence ADME processes. The potential complexity of the regulatory process is further illustrated in the paper by Shukla et al. exploring the role of microRNAs in insulin-mediated downregulation of CYP2E1 (2013). In primary rat hepatocytes, insulin treatment increased the expression of miR-132 and miR-122, which target the 3′untranslated region of CYP2E1, and mechanistic studies implicated the phosphatidyl inositol-3-kinase/Akt/mammalian target of rapamycin pathway in the induction of target microRNAs by insulin.

To fully understand the impact of epigenetic regulation (especially microRNAs) on disease pathogenesis and on drug disposition and response, more comprehensive in vivo studies are required. Prior to integration of epigenetic factors in clinical research, the stability of microRNAs in biologic specimens obtained in the course of pharmacokinetic studies and clinical trials needs to be determined. In a Short Communication in this issue, Benson and Skaar (2013) demonstrate, using Mir-16 and Mir-223 as examples, that microRNAs are stable in whole blood at room temperature for up to at least 12 hours, supporting the use of circulating microRNAs as potential biomarkers in diagnosis and prognosis. However, additional investigation is required to establish whether the stability patterns reported by Benson and Skaar are generalizable to other plasma microRNAs, or whether stability is microRNA-dependent.