Sex-dependent regulation of cytochrome P450 family members Cyp1a1, Cyp2e1, and Cyp7b1 by methylation of DNA

Abstract

Sexual differences are only partially attributable to hormones. Cultured male or female cells, even from embryos before sexual differentiation, differ in gene expression and sensitivity to toxins, and these differences persist in isolated primary cells. Male and female cells from Swiss Webster CWF mice manifest sex-distinct patterns of DNA methylation for X-ist and for cytochrome P450 (CYP; family members 1a1, 2e1m, and 7b1. Dnmt3l is differentially expressed but not differentially methylated, and Gapdh is neither differentially methylated nor expressed. CYP family genes differ in expression in whole tissue homogenates and cell cultures, with female Cyp expression 2- to 355-fold higher and Dnmt3l 12- to 32-fold higher in males. DNA methylation in the promoters of these genes is sex dimorphic; reducing methylation differences reduces to 1- to 6-fold differences in the expression of these genes. Stress or estradiol alters both methylation and gene expression. We conclude that different methylation patterns partially explain the sex-based differences in expression of CYP family members and X-ist, which potentially leads to inborn differences between males and females and their different responses to chronic and acute changes. Sex-differential methylation may have medical effects.—Penaloza, C.G., Estevez, B., Han, D.M., Norouzi, M., Lockshin, R.A., Zakeri, Z. Sex-dependent regulation of cytochrome P450 family members Cyp1a1, Cyp2e1, and Cyp7b1 by methylation of DNA.

susceptibility to disease is often a matter of sex (1–8). Women more frequently have multiple sclerosis, rheumatoid arthritis, systemic lupus erythematous, autoimmune thyroid disease, Graves' disease, and Hashimoto thyroiditis (1–8); and myocarditis, Wegener granulomatosis, idiopathic pulmonary fibrosis, gastritis, diabetes, and ankylosing spondylitis manifest more in men (3–5). These differences are usually attributed to sex hormones and, most recently, sex chromosomes (8). While sex hormones certainly account for some of the observed differences, nonhormonal factors, such as differential gene regulation, are major role players (9). Cells have biological sex, and in the absence of sex hormones, cells behave in a sex-dependent manner (9).

This difference has been demonstrated in several laboratories. Female embryonic day (ED)10.5 and ED17.5 mouse cells are significantly more sensitive to ethanol and camptothecin than their male counterparts (9). Cranial irradiation reduced the continuous addition of new neurons to the granule cell layer much more in females than males (10). Males produced less IFN-α in response to TLR7 and more cytokine IL10 after stimulation with TLR8 and TLR9; sex differences in cytokine production following viral stimulation may explain sex-related response to viral infections (11). Female rat kidneys were less impaired by HgCl₂ as a result of lower Oat1 and Oat3 (mercury uptake transporters; ref. 12). Undoubtedly, other differences are yet to be documented.

At the level of cells, sex differences are often attributed to hormones (9). However, male and female cells have innate differences set prior to the appearance of the fetal sex hormones (9). Differential sensitivity to ethanol and camptothecin precedes sex-specific hormonal influences (9). When cell cultures are supplemented with 17β-estradiol, cell sensitivity is affected, but sex differences persist, and the effects are also sex specific (9). 17β-Estradiol selectively protects female but not male neurons from glutamate-induced stress, with ERα being dimorphically neuroprotective (13). Female mice experience lower mortality and better preserve cardiac function after myocardial infarction and in simulated infarction in cardiomyocytes, possibly as a result of differences in ERα dominance in females (14).

Thus, sex hormones do not account for all the sex-based differences in stress response, cell survival, and metabolism. Such differences could create or arise from differential gene activity (15–22). Members of the cytochrome P450 (CYP) superfamily, responsible for metabolizing drugs and hormones, display sex-dependent gene regulatory patterns in mice, humans, and Drosophila (the latter having no known sex hormones; refs. 15–22); these differences might arise from hormones or epigenetic factors (23–28). For instance, higher Cyp1a1 expression in female patients with lung cancer increases the number of DNA adducts, reducing further breakdown of active carcinogens and increasing overall DNA damage (21). Females are less susceptible to the hyperoxia-induced oxidative stress because of differential Cyp2e1 expression (22). In male and female mouse cells, Cyp7b1, responsible for the aromatization of sex hormone intermediates, is dimorphic whether cells are stressed by ethanol or not, or in the presence of exogenous estrogen, indicating that hormones modulate preexisting differences in gene expression (9). Similarly, Cyp7b1 is higher in male ED10.5 brains, prior to the sex differentiation (23).

Since the CYP family consists of autosomal genes, it is likely that the differential regulation of these genes is epigenetic, such as by methylation of DNA, as is the case for X-chromosome inactivation, genomic imprinting, and regulation of tissue-specific gene expression (24–29). Developmental patterns in early embryonic development are regulated by de novo methylation and demethylation (25–27). DNA (cytosine-5)-methyltransferase 3 (Dnmt3) family members (30–33), which are methyltransferases highly expressed in embryonic stem cells and early developing germ cells (26–30), create and maintain these patterns, suppressing gene transcription by steric hindrance. We found Dnmt3-like (Dnmt3L) to be sex dimorphic in our system, (9); as a result, we further evaluated the expression of this subunit in this study. Other sex-dimorphic genes have not been evaluated.

We asked whether differential methylation of DNA could explain the differences in gene expression. We mapped the methylation of genes known to be sex dimorphic and tested whether inhibiting that methylation would eliminate the dimorphism. We find that the sex-dimorphic, stress-induced gene Cyp7b1 (9) and several others are differentially methylated. Reducing the differences in methylation decreases or eliminates the sex dimorphism in gene expression, while factors such as ethanol and estradiol, which are known to decrease the sex differential, also alter methylation patterns. Thus it is likely that differential methylation is at the root of endogenous sex differences; therefore, the origin of the differential methylation and its effects on disease are targets for further study.

MATERIALS AND METHODS

Animals, cell isolation, culturing and treatment

Swiss Webster mice were used for all experiments. While these experiments were performed on this outbred mouse strain, we find similar outcomes in C57BL/6 mice and (data not shown). Sex differences occur in DNA methylation in other species, including rat amygdala (34), and in GD-1 mouse germ lines (35). Thus, the sex differences appear to be general rather than strain specific. Female Swiss Webster mice were placed with males overnight; detection of vaginal plugs was designated ED0.5. Pregnant females were euthanized by CO₂ and cervical dislocation at ED10.5 and ED17.5, and adolescent mice were likewise euthanized at postnatal day (PN)4 and PN17. Dissected embryos and or tissues were placed in chilled sterile 1× phosphate-buffered saline (PBS). Each embryo was removed from its amniotic sac and placed in chilled Dulbecco's modified Eagle medium (DMEM). For sexing ED10.5 embryos, a small piece of tail was placed in a sterile polymerase chain reaction (PCR) tube and stored at −20°C for sexing by PCR as described below.

To isolate cells from whole embryos, embryos were homogenized by pipetting the embryos 10 times, using the 1-ml pipette tip. For embryonic and postnatal kidney, the tissues were sandwiched between sterile mesh (Sefar Nitex 03–150/38; Sefar AG, Heiden, Switzerland). The cells were separated from connective tissue by applied pressure to the sandwich. Suspended cells were centrifuged at 3000 rpm and resuspended in medium containing DMEM supplemented with 10% fetal bovine serum (FBS), 1% 100 U/ml penicillin, and 100 μg/ml streptomycin, divided among 35-mm plates, and incubated at 37°C in a humidified atmosphere with 5% CO₂. Cells were cultured to 70% confluence prior to treatment [ED10.5: 4 d, ED 17.5: 7 d, and PN17: 10 d]. Medium was changed every 3 d. Prior to treatment with the stressors, the cells were seeded at 1 × 10⁶ cells/plate and incubated for 24 h. To assess cell response, cultured cells were exposed to 400 μM ethanol or 5 nM 17β-estradiol dissolved in DMSO, in 5% FBS (FBS lowered to retard cell division). Treated cells were placed at 37°C for 24 h in ethanol medium, prior to the assessment of cell death.

Inhibition of DNA methylation by 5-aza-2′-deoxycytidine (5-Aza-dC)

ED10.5 cells, 5 × 10⁵/ml, were incubated for 5 consecutive passages at 3–4 d each in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin plus 50 μM 5-Aza-2′-dC dissolved in DMSO. Population doubling was calculated as: t_D = (t − t₀)log₂/(log N − log N₀), where t_D is doubling time, t and t₀ are the times at which the cells were counted, and N and N₀ are the cell numbers at times t and t₀, Medium was replenished every 3 d. At 24 h after the fifth passage, cells were suspended and assessed for DNA methylation.

DNA extraction and isolation by isopropanol precipitation

Cell pellets containing 5 × 10⁶ cells were lysed in lysis buffer (0.1 M Tris, pH 8; 0.2 M NaCl; 5 mM EDTA; and 0.4% SDS) with 0.2 mg/ml proteinase K. Pellets were incubated overnight at 55°C and recentrifuged. Supernatant was transferred into a new microfuge tube. Isopropanol and NaCl were used to precipitate DNA overnight at −20°C. DNA was cleaned and isolated using the Wizard DNA Clean-Up System (Promega, Madison, WI, USA).

Sodium bisulfite treatment of isolated DNA

DNA (2 μg/sample) was dissolved into 10 μl of H₂O. NaOH (1 μl, 6 N) was added to the DNA solution, which was incubated for 15 min at 37°C. To this, 120 μl of 4.04 M NaHSO₃ and 10 mM hydroquinone were added and cycled for 15 cycles: 30 s at 95°C and 15 min at 50°C. Samples were desalted using the Wizard DNA Clean-Up System. Samples were eluted in 20 μl of Tris-EDTA (TE) buffer. The purified product was then sequenced by Eurofins MWG Operon automated DNA sequencing service (Eurofins MWG Operon, Huntsville, AL, USA).

Cell viability by trypan blue exclusion

Dying cells ultimately become permeable to trypan blue. Since our results with trypan blue were completely consistent with other assays such as Live/Dead (Life Technologies, Carlsbad, CA, USA) and Hoechst 33258 (Sigma-Aldrich, St. Louis, MO, USA) staining (data not shown), we relied primarily on this simple and reliable assay. At the end of the treatments, cells were scraped, and the medium containing attached and floating cells was pelleted, resuspended in an equal volume of 0.4% trypan blue (Sigma-Aldrich) in 1× PBS, pH 7.4, and incubated at room temperature for 3 min. At least 200 cells (dead or alive) were counted under a microscope, and each treatment was done ≥3 times in triplicate. Blue cells were scored as nonviable.

The percentage cell death, 100 × dead cells/total cells, was normalized by subtracting the basal level of cell death observed in each control, ranging from 10 to 15%. Statistical significance of the results was calculated by standard t test; P < 0.05 counted as significant difference.

Sex determination

PCR was necessary for sexing anatomically undifferentiated ED10.5 embryos. DNA from a small piece of tail was isolated by digestion with 2 μl proteinase K (10 mg/ml in dH₂O) and 50 μl PCR-D buffer and left overnight at 65°C. The enzyme was denatured at 95°C for 10 min, and 1 μl was transferred into a PCR tube, to which we added 21 μl dH₂O, 25 μl Master Mix (Sigma-Aldrich), 2 μl MgCl₂, 1 μl Primer Mix (25 pmol/μl of Zfy and Zfxya, 12.5 pmol/μl of Zfx primers). Primers for Zfy were 5′-CTCCTGATGGACAAACTTTACGTCTC and 3′-GCTGAGCCTCTTTGGTATCTGAGAAA; primers for Zfxya were 5′-GAGAGCATGGAGGGCCATG and 3′-GAGTACAGGTGTGCAGCTC; Zfx primers were 5′-CTCTGAAGAAGAGACAAGTT and 3′-CTGTGTAGGATCTTCAATC. PCR was done for 40 cycles (94°C for 45 s, 60°C for 25 s, 72°C for 60 s) in the Eppendorf 2200 thermal cycler (Eppendorf, Hamburg, Germany). The amplified DNA was visualized on a 12% native polyacrylamide gel in TBE buffer using positive controls and viewed by UV illumination. Male samples showed a 124-bp band for gene Zfy and a 134-bp band for gene Zfx; females showed only the 134-bp Zfx band.

Sexing of other stages was by dissection.

Real-time quantitative reverse transcription PCR (qRT-PCR)

Small PCR products (100–200 bp) were amplified in quadruple on a Roche LightCycler 2.0 real-time PCR machine (Roche Diagnostics, Indianapolis, IN, USA), using universal PCR conditions (65 to 59°C touchdown, followed by 40 cycles of 15 s at 95°C, 10 s at 59°C, and 10 s at 72°C). cDNA (500 pg) was amplified in 20 μl reactions (0.3× Sybr-green, 3 mM MgCl₂, 200 μM dNTPs, 200 μM primers, and 0.5 U Platinum TaqDNA polymerase; Roche). Primer dimers were assessed by amplifying primers without cDNA. Primers were retained if they produced no primer dimers or only nonspecific signal after 40 cycles. Results were calculated as relative intensity compared to female expression. The last cycle was retained as baseline for comparison with absent genes. Data were plotted as cycle threshold (C_T) values, in which the C_T value represents the cycle at which fluorescence is first detected. A lower cycle number indicates a higher initial concentration of mRNA, and each decrease of one cycle indicates a doubling of initial concentration. Experiments were done in triplicate. Average C_T values are shown. Student's t test was performed, with values of P < 0.05 considered significant. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of C_T − 1.

RESULTS

Developmental timing and endogenous hormones affect sex-dimorphic expression of X-ist, Cyp1a1, Cyp2e1, Cyp7b1 , and Dnmt3l

Since littermates of inbred mouse strains are almost genetically identical except for the sex chromosomes, dimorphism in gene expression can be attributed to the sex chromosomes. To study the effects of sex on gene expression, we used homogenates from ED10.5 embryos, whole kidney, and lung tissue homogenates from ED17.5, PN17, and adult mice. At ED10.5, the fetuses have not been exposed to fetal sex hormones, as gonad development and onset of hormone production occurs at ∼ED13.5. At ED17.5, the fetuses have been exposed to the fetal sex hormones and have undergone secondary sex differentiation, while PN17 and adult mice exhibit adolescent and adult levels of hormones. Kidney and lung were selected as these are discrete, well-bounded, easy to acquire without contamination, and easy to maintain in culture for several population doublings. They are also bilateral, in case duplicate cultures are required. These organs are as well involved in several sexually dimorphic diseases, and CYP members are active in these organs.

Some members of the CYP family are sexually dimorphic in both normal and disease states and are candidates for assessment of the mechanism underlying sexually dimorphic disease. We therefore examined the expression of CYP family members by qRT-PCR using specific primers for each gene, to evaluate their baseline expression at these stages.

X-ist, inactivating transcript for the X chromosome, is expressed more in females, and therefore was a positive control; glyceraldehyde-3-phosphate dehydrogenase (Gapdh), a housekeeping glycolysis gene and not dimorphic, was a negative or loading control. As expected, samples from females possessed considerably more X-ist mRNA (Fig. 1A). Kidney tissue from ED17.5, PN17, and adult females expressed 10- to 71-fold more X-ist than males, while female lung tissues expressed 4- to 144-fold greater X-ist (Fig. 1A).

Figure 1.

CYP family members are expressed in a developmental-, tissue-, and sex-dependent manner in vivo. CYP expression was measured by qRT-PCR using specific primers to evaluate their baseline expression. Solid bars, male; shaded bars, females. A–F) Whole kidney and lung tissue homogenates from ED17.5, PN17, and adult mice. Samples were normalized at the RNA and cDNA levels for equal cDNA loading. Experiments were done in triplicate. Average C_T values are shown. Ordinate is average C_T of fluorescence detection for real-time PCR. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of CT − 1. Values below each set of bars represent the calculated fold difference (FD) between the male and female pair. A) Expression of X-ist is more prominent in females. B) Cyp1a1 expression is more prominent in females. C) Cyp2e1 expression is more prominent in females only at ED 17.5. D) Cyp7b1 expression is more prominent in females only at ED17.5 in kidney tissues and at ED17.5 and PN17 in lung tissues. E) Dnmt3l expression is more prominent in males. F) Gapdh is not sex dimorphic, except for PN17 in lung tissues. G) As a means of validating the data from homogenized PN 17 organs, whole-embryo homogenates from ED10.5 were assessed for CYP expression. Expression of X-ist, Cyp1a1, Cyp7b1, and Cyp2e1 is more prominent in females; that of Dnmt3l is more prominent in males; Gapdh expression is not sex dimorphic. *P < 0.05.

CYP family member Cyp1a1, which metabolizes xenobiotics and drugs, differs markedly according to sex, with little variation among tissues or by age. Female Cyp1a1 expression was 16- to 256-fold higher than in male equivalent tissues (Fig. 1B). Cyp2e1, responsible for ethanol metabolism, is similar but not as dramatic. At ED17.5, but not later, females express 4× more in kidney and 11× more in lung (Fig. 1C). Cyp7b1, important in synthesis of steroid hormones, manifests moderate differences in expression at ED17.5. This difference disappears in PN17 and adult kidney, but female ED17.5 kidney and lung and PN17 lung express 7- to 14-fold more Cyp7b1 (Fig. 1D).

We considered that the differences in expression could arise from overall differences in methylation. Dnmt3l is a methyltransferase subunit, known to interact with histone deacetylase 1 (HDAC1), thus regulating the expression of various genes. Male Dnmt3l expression was 4- to 97-fold higher than expression in females in all tissues and developmental stages assessed (Fig. 1E). Our negative (constitutively expressed) control, Gapdh, is expressed independently of sex, with the exception of PN17 when females express 7-fold more Gapdh in lung tissue. Otherwise, Gapdh expression does not vary (Fig. 1F).

Gene expression in homogenates of whole embryos, prior to appearance of fetal hormones (ED10.5), was similar to that in later development: female expression of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1was 4–32× higher in ED10.5 whole embryos. In contrast, Dnmt3l expression was 32× higher in males; no difference was found in Gapdh (Fig. 1G). Overall, as others have occasionally noted, expression of the CYP genes is specific to tissue type, developmental timing, and chromosomal sex (Fig. 1B–D). Nevertheless, we conclude that chromosomal sex can regulate expression.

Sex differences identified in whole tissue were maintained in cell culture

To assess the dimorphic regulation of gene expression, we used the in vitro system, in which we showed sex dimorphic gene expression (9). We used ED10.5 mixed cell cultures, as well as mixed cell cultures from kidneys of ED17.5 and PN17 animals. Cells were maintained in culture as described in Materials and Methods. The expression of X-ist, Cyp1a1, Cyp2e1, Cyp7b1, Dnmt3l, and Gapdh was assessed by qRT-PCR for each of the cell types to obtain baseline expression.

For ED10.5 embryos, sex differences in gene expression were maintained, as in whole-tissue homogenates seen in Fig. 1, which resulted in 8- to 45-fold statistically higher expression of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1 in females, and 16-fold higher expression of Dnmt3l in males, with no difference in Gapdh (Fig. 2A). Except for Cyp7b1, the differences in expression were maintained in ED17.5 kidney mixed cell cultures. Females showed 6- to 125-fold higher expression of X-ist, Cyp1a1, and Cyp2e1, with no differences in the expression of Cyp7b1 or Gapdh between the sexes, and males showed 12×-fold higher expression of Dnmt3l in ED17.5 kidney cells (Fig. 2B).

Figure 2.

Results from cells cultured from organs as indicated, in contrast to Fig. 1, which represents measurements from whole embryos and homogenates of organs. CYP family members are expressed in a developmental-, tissue-, and sex-dependent manner in vitro. CYP expression was measured by qRT-PCR using specific primers to evaluate their baseline expression. Solid bars, male; shaded bars, females; ED10.5 mixed cells (A), ED17.5 kidney cells (B), and PN17 kidney cells (C). Samples were normalized at the RNA and cDNA levels for equal cDNA loading. Experiments were done in triplicate. Average C_T values are shown. Ordinate is average C_T of fluorescence detection for real-time PCR. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of C_T − 1. Values below each set of bars represent the calculated fold difference between the male and female pair. A) Expression of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1 is more prominent in female ED10.5 mixed cells, with Dnmt3l expression being more prominent in male ED10.5 mixed cells and no difference in Gapdh expression. B) Expression of X-ist, Cyp1a1, and Cyp2e1 is more prominent in female ED17.5 kidney cells, with Dnmt3l expression more prominent in male ED17.5 kidney cells and no difference in Gapdh and Cyp7b1 expression. C) Expression of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1 is more prominent in female PN17 kidney cells, with Dnmt3l expression more prominent in male PN17 kidney cells and no difference in Gapdh expression. *P < 0.05.

In cultures from PN17 kidney, females expressed significantly (5- to 355-fold) more X-ist, Cyp1a1, Cyp2e1, and Cyp7b1, while Dnmt3l expression was 32× higher in males, with no differences in the expression of Gapdh (Fig. 2C). Thus, with the possible exception of Cyp7b1, our cell culture system closely reflects the in vivo expression, confirming sex-dimorphic gene expression profiles in vitro. We can therefore use cell culture to explore the origin of the sex differences.

Promoter region methylation of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1 is sex specific

The male/female difference in expression suggests sex-specific targets of transcriptional regulation, such as those typically found in the promoter regions. Therefore, we searched for consensus sequences in the genomic sequence of X-ist and several members of the CYP family. We found cAMP response elements (CREBs), estrogen response elements (EREs), and glucocorticoid response elements (GREs) in the promoter regions of Cyp1a1, Cyp2e1, Cyp3A4, Cyp4A11, Cyp5A1, Cyp7b1, and Cyp8A1 (Table 1) and selected these genes for further study. Cyp1a1, Cyp5A1, and Cyp7b1 contain EREs; Cyp1a1, Cyp3A4, Cyp5A1, and Cyp8A1 contain GREs; none contain CREBs; but 5 of the 7 assessed sequences possess CpG islands > 200 bp, potential sites of methylation (Table 1).

Table 1.

Regulatory consensus across CYP family members

Gene	ERE	CREB	GRE	CpG > 200 bp
Cyp1a1	+	–	+	+
Cyp2e1	–	–	–	+
Cyp3A4	–	–	+	+
Cyp4A11	–	–	–	–
Cyp5A1	+	–	+	+
Cyp7b1	+	–	–	+
Cyp8A1	–	–	+	–

EREs were found in the promoter region of Cyp1a1, Cyp5A1, and Cyp7b1; no CREBs were found in the promoter region; GREs were found in the promoter region of Cyp1a1, Cyp3A4, Cyp5A1, and Cyp8Al; and 5 (Cyp1a1, Cyp2e1, Cyp3A4, Cyp5A1, and Cyp7b1) of the 7 assessed sequences possess CpG islands > 200 bp.

Since the CYP genes show sex-specific regulation and consistently possess large CpG islands, we evaluated methylation as a potential regulator. Dnmt3l, a DNA methyltransferase gene, is expressed dimorphically; X-ist is subject to epigenetic regulation (19). We therefore asked whether DNA methylation explained the dimorphic expression of X-ist. We identified >100 CpG islands spanning the X chromosome and ≥3 spanning the promoter region of X-ist that are >100 CpG dinucleotides (Table 2). We focused on the region spanning from −2393 to −2011 on the promoter region of X-ist, which contained the largest CG-rich region, as well as several promoter regions from Cyp genes.

Table 2.

CpG island identification and selection for select CYP family members

Gene	Location	CpGs	CpGs > 200 bp	CpG islands analyzed	CpG island position	Size of region
Cyp1a1	Chr 9: 57,545,420–57,551,629	6	2	1	−873 to −566	307
Cyp2e1	Chr 7: 147,949,638–147,960,874	12	1	1	−1231 to −982	249
Cyp7b1	Chr 3: 17,971,950–18,143,338	3	1	1	−667 to −402	265
X-ist	Chr X: 100,655,714–100,678,556	>3	>3	1	−2393 to −2011	382
Gapdh	Chr 12: 71,021,068–71,022,717	0	0	NA	NA	NA

We identified 6, 12, 3, >3, and 0 CpG islands spanning the promoter regions of Cyp1a1 (positions −873 to −566), Cyp2e1 (positions 1231 to −982), Cyp7b1 (positions 667 to −402), X-ist (positions −2393 to −2011) and Gapdh respectively.

We selected single CpG islands spanning >200 bases. We identified positions −873 to −566 for Cyp1a1, −1231 to −982 for Cyp2e1, and −667 to −402 for Cyp7b1, representing CpG islands of 307, 249, and 265 bases, respectively (Table 2). We asked whether DNA methylation played a role in dimorphic gene expression. We used sodium bisulfite, which converts all unmethylated cytosines to uracils, allowing us to identify site-specific methylation. Subsequently, we purified the treated DNA and amplified the region spanning the CpG island on the promoter of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1. The amplicons were again purified, repurified, and submitted for sequencing of 5 replicates of each condition (Eurofins MWG Operon). Each replicate represented an individual experiment, diluted and amplified as a separate condition. The data represent the mean methylation status for each spanning region. Of 5 individual rounds of sequencing for each condition, a positive methylation site represents 5 of 5 sequenced samples sharing the methylation pattern. There existed variations in methylation sites, but these were present 20% of the time or less frequently. Any ambiguous results prompted a new round of validating experiments. In Fig. 3, the sequences presented represent the regions where modifications were identified. A solid circle on the cytosine in Fig. 3 identifies where the base was methylated. We describe below the methylation status of these specific promoter regions:

Figure 3.

Sex-dimorphic DNA methylation under normal conditions. Genomic DNA extracted from male and female cells was treated with sodium bisulfite, and specific promoter regions were amplified and sequenced. Presented are the parental DNA sequences, with notations on variations between the sexes. Solid circles indicate methylated sites; shaded circles indicate unmethylated sites. Five individual rounds of amplification and sequencing are averaged into the representation, with each site resulting from ≥4 of 5 replicates. DNA methylation status is sex dimorphic under normal conditions, as observed with X-ist (5 methylation sites in males; 3 in females), Cyp1a1 (3 methylation sites in males; 4 in females), Cyp2e1 (5 methylation sites in males and females, 3 of which are common), and Cyp7b1 (4 methylation sites in males; 2 in females).

X-ist

Five sites of methylation were found in males; 3 in females. One was common to both sexes (Fig. 3). Two were unique to females and 4 to males. This difference has not been previously reported.

Cyp1a1

Three sites of methylation were found in males; 4 in females; of these, only one was common. Three were unique to females and 2 to males.

Cyp2e1

Males and females each had 5 methylated sites. Of these, 3were common to both males and females (Fig. 3). Two sites were female-specific; 2 were specific to males.

Cyp7b1

Males had 4 methylation sites and females 2, of which 1 is common. One site was unique to females and 3 to males.

Dnmt3l and Gapdh

Dnmt3l and Gapdh were not analyzed, as these sequences do not possess detectable CpG islands.

Thus, the promoters from differentially expressed genes also are methylated differentially. These differences suggest an explanation for sex-dimorphic gene expression. We tested this hypothesis by inhibiting methylation.

Inhibition of DNA methylation reduces sex differences in methylation and gene expression for X-ist, CYP family members, and Dnmt3l

To examine the relationship between expression and methylation, we blocked the de novo transfer of methyl groups, using 10 μM 5-Aza-dC, which prevents the faithful copy of methylated sites to the daughter strand from the parental strand. This inhibition was maintained for 5 rounds of replication as measured by population doublings, which should reduce the methylated strands to 3% of the cell population. The 5-Aza-dC was not toxic, and, in fact, the cells did better and delayed or evaded senescence when exposed to chronic 5-Aza-dC (Fig. 4). Inhibition of methylation reduces differential methylation of CYP P450 members, as well as the elimination or reduction of sex differences, as described below:

Figure 4.

ED10.5 recovery of male and female cells after exposure to 5-Aza-dC. Inhibition of DNA methylation by 5-Aza-dC results in increased cell recovery after each round of passage of ED10.5 cells. *P < 0.05.

X-ist

Inhibition of DNA methylation almost eliminates the sex differences in both methylation and expression of X-ist. Blocking de novo DNA methylation results in a partial loss of DNA methylation sites on the promoter region of X-ist in males and females. Males lose 3 sites of methylation (Fig. 5, open circles) but maintain 2 sites. Females lose 1 site of methylation but maintain 2 (Fig. 5). 5-Aza-dC does not block all de novo methylation. However, the reduction of DNA methylation resulted in fewer differences in DNA methylation between the sexes, from 6 in control conditions to 2 in 5-Aza-dC-treated cells (Fig. 5), with 1 male-specific and 1 female-specific site. The expression of X-ist was increased by reduction of methylation in both sexes, but more in males, so that the ratio dropped from 8:1 (female to male ratio) to near unity in cells exposed to 5-Aza-dC (Fig. 6).

Figure 5.

5-Aza-dC reduces sex differences in methylation. Extracts of genomic DNA from male and female cells were treated with sodium bisulfite, and specific promoter regions were amplified and sequenced. Presented are the parental DNA sequences, with notations on variations between the sexes. Solid circles indicate methylated sites; open circles indicate loss of methylation; shaded circles indicate unmethylated sites; shaded circles with borders indicate newly gained methylation. Five rounds of amplification and sequencing are averaged into the representation, with each site resulting from ≥4 of 5 replicates. 5-Aza-dC reduces differential methylation, as observed with X-ist and CYP family members Cyp1a1, Cyp2e1, Cyp7b1, and Dnmt3l.

Figure 6.

5-Aza-dC reduces sex differences in expression. Sex dimorphism of X-ist decreases from 8 to 1; Cyp1a1, from 23 to 6; Cyp2e1, from 8 to 2; Cyp7b1, from 16 to 6; Dnmt3l, from 8 to 3; and Gapdh from 4 to 1). Expression was measured by qRT-PCR using specific primers to evaluate baseline expression. Solid bars, male; shaded bars, females; ED10.5 mixed cells. Samples were normalized at the RNA and cDNA levels for equal cDNA loading. Experiments were done in triplicate. Average C_T values are shown. Ordinate is average C_T of fluorescence detection for real-time PCR. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of C_T − 1. Values below each set of bars represent the calculated fold difference (FD) between the male and female pairs. *P < 0.05.

Cyp1a1

The promoter for Cyp1a1 contains 2 male-specific sites of methylation, while in females there are 3, and 1 of these is common. When methylation is inhibited, 2 sites survive in males and females, of which 1 is unique to each sex (Fig. 5). This reduction of differences in DNA methylation sites (5 to 2) resulted in a drop from 23-fold to 6-fold difference in the expression of Cyp1a1 (Fig. 6).

Cyp2e1

The promoter for Cyp2e1 has 5 methylated sites in males and 5 in females. Of these, 2 are unique to males and 2 to females. When methylation is blocked, the number of methylated sites is reduced to 3 in males and 4 in females, with only 1 unique site in females (Fig. 5). This reduction of methylation differences resulted in a reduction in differential Cyp2e1 expression from 8-fold to 2-fold (Fig. 6).

Cyp7b1

The promoter for Cyp7b1 has 4 sites methylated in males and 2 in females; 3 of the male sites are unique, with only 1 being common. After methylation is blocked, only 1 site, the male-specific site, persists in males and 0 in females (Fig. 5). Similar to the other CYP members, loss of DNA methylation sites results in a reduction of sex differences in Cyp7b1 expression, dropping from 16- to 6-fold (Fig. 6).

Dnmt3l

The promoter for Dnmt3l did not possess a CpG island, and therefore, differences in methylation status were not identified; gene expression varied minimally as a result of 5-Aza-dC exposure, indicating that it is not regulated by DNA methylation (Fig. 6).

Gapdh

Methylation status for Gapdh was not analyzed, as Gapdh did not have any long CpG islands. Expression of Gapdh (negative control) did not vary significantly as a result of loss of de novo DNA methylation, (Fig. 6).

In all cases, blocking methylation reduced both the number of sites methylated as well as the number of sex-unique sites (Fig. 5). Expression of X-ist, Cyp1a1, Cyp2e1, and Cyp7b1 increased significantly more in males than females. These results suggest that methylation patterns establish the transcriptional regulation of Cyp1a1, Cyp2e1 and Cyp7b1 (Fig. 6). Normally these patterns—presumptively established by the difference in sex chromosomes—likely generate the sex-differential expression. Blocking methylation destroys or reduces these differences and results in the loss of sex differences in expression for these genes (Figs. 5 and 6).

Ethanol-induced stress produces differences in gene expression and methylation status for X-ist, Dnmt3l, and CYP family members

When male and female cells are exposed to stressors such as ethanol or camptothecin, female cells are more sensitive to the stress, and the sex dependence of gene expression likewise responds to stress (9). Gene expression in response to stress is not unidirectional; stress can induce expression of a gene in one sex and suppress it in the other sex (9). To examine the relationship between differential DNA methylation and expression in response to stress, we exposed cells to 400 μM ethanol, representing LD₅₀. The cells were maintained for 24 h and assessed for gene expression and DNA methylation. After 24 h exposure to ethanol, gene expression and DNA methylation differences were as follows:

X-ist

In the short promoter region assessed for X-ist, females lost 2 sites of methylation (Fig. 7, open circles). These changes resulted in 6 sites differing, 5 unique male sites and 1 unique female site, compared to 6 differences in controls. These changes to the sites of methylation resulted in an increase of sex differences in X-ist expression from 16- to 32-fold, with females having higher expression after exposure to ethanol (Fig. 8). Although the difference between sexes remains significant, the doubling of the difference is not a significant change.

Figure 7.

Ethanol alters methylation patterns for X-ist, Dnmt3l, Cyp1a1, Cyp2e1, and Cyp7b1.Extracts of genomic DNA from male and female cells were treated with sodium bisulfite, and specific promoter regions were amplified and sequenced. Presented are the parental DNA sequences, with notations on variations between the sexes. Solid circles indicate methylated sites; open circles indicate loss of methylation; shaded circles indicate unmethylated sites; shaded circles with borders indicate newly gained methylation. Five rounds of amplification and sequencing are averaged into the representation, with each site resulting from ≥4 of 5 replicates.

Figure 8.

Ethanol (EtOH) alters dimorphism of expression. Dimorphism of X-ist increases from 16 to 32; Cyp1a1, decreases from 128 to 3; Cyp2e1, increases from 13 to 32; Cyp7b1, decreases from 16 to 4; Dnmt3l, increases from 8 to 16; and Gapdh decreases from 2 to 1. Expression was measured by qRT-PCR using specific primers to evaluate their baseline expression. Solid bars, male; shaded bars, females; ED10.5 mixed cells. Samples were normalized at the RNA and cDNA levels for equal cDNA loading. Experiments were done in triplicate. Average C_T values are shown. Ordinate is average C_T of fluorescence detection for real-time PCR. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of C_T − 1. Values below each set of bars represent the calculated fold difference (FD) between the male and female pairs. *P < 0.05.

Cyp1a1

One site is lost in males (Fig. 7, open circle) and 2 sites in females (Fig. 7, open circles), leaving 2 sites in each sex, all unique. This reduction of differences in DNA methylation, from 5 to 4 unique sites, resulted in a marked reduction of sex differences in the expression of Cyp1a1, from 128- to 3-fold after ethanol exposure (Fig. 8).

Cyp2e1

The number of methylated sites is reduced to 3 in males by loss of two sites (Fig. 7, open circles) and increased to 8 in females by gain of 3 sites (Fig. 7, shaded circles with borders), resulting in 9 unique sites and one common. This increase of methylation differences, in response to ethanol exposure, results in an increase from 13- to 32-fold differences between the sexes in expression of Cyp2e1 (Fig. 8).

Cyp7b1

Only 2 male-specific sites persist in males, by loss of two sites in males (Fig. 7, open circles); females lost 2 sites (Fig. 7, open circles), resulting in 2 male-specific sites. As above, loss of DNA methylation sites results in a reduction of sex differences in Cyp7b1 dropping from 16- to 4-fold after ethanol exposure (Fig. 8).

Exposure of cells to ethanol, a form of stress, alters methylation and results in both a reduction of sex differences in methylation and in expression of Cyp1a1 and Cyp7b1 genes, whereas for Cyp2e1, an increase in difference in methylation sites is reflected in an increased difference in expression between the sexes.

Exposure of cells to 17β-estradiol results in differences in gene expression and methylation for X-ist and CYP family members

Similar to ethanol, 17β-estradiol modulates sex differences in gene expression. When male and female cells are exposed to estradiol or estradiol in combination with ethanol, cell survival is sex dimorphic, and gene expression is sex dependent (9). 17β-Estradiol can induce genes in one sex and suppress the genes in the other sex (9). To examine the relationship between the differential DNA methylation and expression in response to sex hormones, we exposed cells to 5 nM 17β-estradiol and maintained them for another 24 h. DNA methylation profiles (Fig. 9) and gene expression (Fig. 10) were assessed:

Figure 9.

Estradiol alters methylation sites of X-ist, Cyp1a1, Cyp2e1, Cyp7b1, and Dnmt3l. Extracts of genomic DNA from male and female cells were treated with sodium bisulfite, and specific promoter regions were amplified and sequenced. Presented are the parental DNA sequences, with notations on variations between the sexes. Solid circles indicate methylated sites; open circles indicate loss of methylation; shaded circles indicate unmethylated sites; shaded circles with borders indicate newly gained methylation. Five individual rounds of amplification and sequencing are averaged into the representation, with each site resulting from ≥4 of 5 replicates.

Figure 10.

Estradiol alters expression levels. In the presence of 17-β estradiol (E2), the sex dimorphic ratio of X-ist is reduced from 16 to 0; Cyp1a1, from 128 to 4; Cyp2e1, from 13 to 2; Cyp7b1, from 16 to 4; and Dnmt3l, from 8 to 1, with no change for Gapdh. Expression was measured by qRT-PCR using specific primers to evaluate their baseline expression. Solid bars, male; shaded bars, females; ED10.5 mixed cells. Samples were normalized at the RNA and cDNA levels for equal cDNA loading. Experiments were done in triplicate. Average C_T values are shown. Ordinate is average C_T of fluorescence detection for real-time PCR. Each increment in C_T indicates a starting transcript concentration ∼0.5× that of C_T − 1. Values below each set of bars represent the calculated fold difference (FD) between the male and female pairs. *P < 0.05.

X-ist

Compared to control after exposure to 17β-estradiol, the short promoter region lost 4 sites of methylation (Fig. 9, open circles) and gained 1 site in males (Fig 9, shaded circle with border), whereas females lost 1 site. These changes resulted in 4 sex-unique sites, 2 male and 2 female (Fig. 9). These changes to the sites of methylation, from 6 to 4 unique sites, resulted in a reduction of differences in X-ist expression from 16-fold to 0 (Fig. 10).

Cyp1a1

In comparison to control, 1 site survives in each sex, as a result of a loss of 2 male and 3 female sites (Fig. 9, open circles). This reduction from 5 to 2 unique sites resulted in a reduction from 128- to 4-fold of sex differences in the expression of Cyp1a1 (Fig. 10).

Cyp2e1

The number of methylated sites is reduced from 5 to 3 in males and 4 in females, with only 1 unique, as a result of 2 male and 1 female site losses (Fig. 9, open circles). This reduction of differences in methylation results in a reduction from 13- to 2-fold differences between the sexes in the expression of Cyp2e1 (Fig. 10).

Cyp7b1

The promoter for Cyp7b1 has 4 sites methylated in males and 2 in females; 3 of the male sites are unique, and 1 is common. After exposure to estradiol, only 1 male site persists and 3 in females, resulting from 3 sites lost in males (Fig. 9, open circles) and a site gained in females (Fig. 9, shaded circle with border). Similar to the other CYP members, restructuring of DNA methylation sites as a result of estrogen (4 unique sites, but different from no hormone) results in a reduction of sex differences in Cyp7b1 between the sexes, dropping from 16- to 4-fold after blockage of DNA methylation (Fig. 10).

Exposure of cells to 17β-estradiol alters methylation status and results in reduction of sex differences in both methylation and gene expression. These results suggest that DNA methylation selectively regulates autosomal genes such as CYP in a sex-dependent manner, which can in part explain the many instances where gene expression is sexually skewed, presumptively resulting in differences at the physiological level. Such differences could persist throughout life and are relevant to evaluation of normal physiology and pathology in men and women.

DISCUSSION

We have previously shown that cells differing only in chromosomal sex respond differently to various stressors (9). Here we address the mechanism of this dimorphism. A previously unrecognized sex-differential gene regulation pattern persists in isolated primary cells. Male and female cells from similar tissues have sex-distinct patterns of DNA methylation, with the magnitude of the difference somewhat proportional to the differential in expression. This conclusion is valid for CYP family members Cyp1a1, Cyp2e1, and Cyp7b1, as well as X-ist but not Dnmt3l, which shows differential expression but no known difference in methylation; nor for Gapdh, which shows no difference in methylation and almost no sex differential in expression. Inhibition of DNA methylation or exposing cells to ethanol or 17β-estradiol (both of which reduce differences in methylation) reduces or eliminates the differential expression. Thus, reduction or elimination of sex differences in gene expression can largely be attributed to a loss of differences in DNA methylation.

In the absence of stress, male and female cells have different gene expression profiles alterable by stress, though often the cells maintain sex specificity (9). Such differential gene regulation can generate differential resistance to stress. The drug detoxifying enzymes of the CYP family are known to behave in a sex-specific manner (15–23). The most consistent regulatory element in the promoters of CYP family members and X-ist was CpG islands, which were present in most sex-dimorphic CYP family members (Table 2). Of these, Cyp1a1, Cyp2e1, and Cyp7b1 are expressed in a sexually dimorphic manner and have CpG islands of at least 200 bp. We previously reported that Dnmt3l was sex dimorphic in the absence of stress (9).

Sex differences in the expression of the CYP genes in vivo exist regardless of tissue type or developmental time (Fig. 1). These differences in gene expression persist or are enhanced in culture (Fig. 2). Likewise, under normal conditions, the inherited DNA methylation patterns for CYP family members Cyp1a1, Cyp2e1, and Cyp7b1 are sex dimorphic (Fig. 3). We therefore first attempted to disrupt this pattern using 5-Aza-dC (33–36), which was not toxic (Fig. 4), even though both sexes lost methylation sites, resulting in fewer differences DNA methylation (Fig. 5). Thus, we would expect reduced sex dimorphism in expression.

Aberrant promoter methylation causes loss of gene expression that can provide selective growth advantage to neoplastic cells. Jones and Baylin (37) argue that the promoter hypermethylation-associated loss of DNA repair capacity can predispose cells to mutation and may provide a selective advantage to neoplastic cells. The loss of mismatch repair capacity causes increased cell survival by decreased apoptotic response and increased resistance to chemotherapeutic drugs (37). We find that exposing cells to 5-Aza-dC results in an increase of cell recovery, potentially as a result of aberrant patterning of methylation.

5-Aza-dC increased all gene expression and reduced or eliminated sex differences, even with X-ist (Fig. 6). Ethanol changed methylation patterns within 24 h (Fig. 7), modestly affecting gene expression, most characteristically increasing male expression to sharply decrease sex differential (Cyp1a1 and Cyp7b1, Fig. 8). Prolonged exposure to ethanol was lethal, which prevented us from assessing the long-term effect of ethanol on DNA methylation. However, 24 h exposure to ethanol elicited differential gene expression and modified the methylation profile of dimorphically expressed genes (Fig. 7). Likewise, estradiol modulated DNA methylation, influencing the expression profiles of these genes (Figs. 9 and 10).

The female sex hormone 17-β estradiol had markedly different effects on male and female cells. When ED10.5 cells were cultured for 3 d without exogenous estradiol and the last 24 h in the presence of the hormone, both male and female methylation patterns were changed (Fig. 9). Typically, expression was increased in both sexes, but substantially more in males, thus reducing sex-based differences (Fig. 10).

Thus, all our experiments lead us to the conclusion that differences in patterns of methylation can at least partially explain the sex-based differences in expression of CYP family members and X-ist. Whether we block methylation by 5-Aza-dC, stress cells with ethanol, or expose them to a female sex hormone, changes in methylation differences coincide with changes in expression and difference in expression, approximately proportionate to the number of methylation differences. While much remains to be learned about the origins of these differences and the manner in which additives and other factors affect each site of methylation, the evidence points to the argument that inborn differences between males and females and their different responses to chronic and acute changes in their lives could derive from these differences in methylation. The medical effects of differential methylation will become clearer as we learn more.