Aberrant expression of deoxyribonucleic acid methyltransferases DNMT1, DNMT3A, and DNMT3B in women with endometriosis

ENDOMETRIOSIS Aberrant expression of deoxyribonucleic acid methyltransferases DNMT1, DNMT3A, and DNMT3B in women with endometriosis Yan Wu, Ph.D.,a Estil Strawn, M.D.,b Zainab Basir, M.D.,c Gloria Halverson, M.D.,b and Sun-Wei Guo, Ph.D.a a

Department of Pediatrics, b Department of Obstetrics and Gynecology, and c Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin

Objective: Since endometriosis is a persistent disease with substantial gene dysregulation, there must be cellular memory of some sort that constitutes a unique cell identity for endometriotic cells. Epigenetic regulation, especially through DNA methylation, is a flexible, yet stable, mechanism for maintaining such a cellular memory. The aim of this study was to determine gene expression levels of DNMT1, DNMT3A, and DNMT3B, the three genes coding for DNA methyltransferases that are responsible for methylation. Design: Cross-sectional measurements of gene expression levels of DNMT1, DNMT3A, and DNMT3B on endometriotic tissue. Setting: Academic. Patient(s): Seventeen patients with laparoscopically confirmed endometriosis and 8 healthy women who underwent tubal sterilization who were free of endometriosis were recruited for the study. Intervention(s): Epithelial cells were harvested from tissue samples by laser capture microdissection and messenger RNA abundance was measured by quantitative real-time reverse transcription–polymerase chain reaction. Main Outcome Measure(s): The expression levels of these genes in epithelial cells from 13 ectopic endometrial tissue samples, 10 eutopic endometrial tissue samples taken from women with endometriosis, and 8 normal endometrial tissue samples from women without endometriosis. Result(s): The genes DNMT1, DNMT3A, and DNMT3B were over-expressed in the ectopic endometrium as compared with normal control subjects or the eutopic endometrium of women with endometriosis, and their expression levels were correlated positively with each other. Conclusion(s): The aberrant expression of these genes suggests that aberrant methylation may be rampant in endometriosis. This also provides a strong piece of evidence that endometriosis ultimately may be an epigenetic disease. (Fertil Steril威 2007;87:24 –32. ©2007 by American Society for Reproductive Medicine.) Key Words: DNA methyltransferase, endometriosis, epigenetics, gene expression, methylation

Aberrant gene expression has been demonstrated repeatedly and consistently in endometriosis. These aberrations range from the apparent inactivation of 17␤-hydroxysteroid dehydrogenase 2, progesterone receptor isoform B (PR-B), and HOXA10 (1–3) to the up-regulation of matrix metalloproteinases (4) and P450 aromatase (5) and to massive gene expression aberration uncovered by the microarray technology (6 –9). They are likely responsible for observed phenotypic aberrations at hormonal, biochemical, immunologic, and ultimately clinical levels in endometriosis, which are manifested as the excessive local production of

Received January 31, 2006; revised and accepted May 19, 2006. Supported by The Children’s Research Institute, Milwaukee, Wisconsin. Reprint requests: Sun-Wei Guo, Ph.D., Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Road, MS 756, Milwaukee, WI 53226-0509 (FAX: 414-456-6663; E-mail: [email protected]).

24

estrogen, inflammation, altered apoptotic mechanisms, and implantation failure. These aberrant expressions are seemingly quite stable and consistent in endometriosis, which indicates that cellular memory of some sort must be involved. Although constitutive aberrations such as a loss of heterozygosity and genomic alterations in endometriosis have been reported (10 –15) and may be responsible for gene expression aberrations, another important mechanism for gene regulation (i.e., epigenetics) has received scant attention so far. Epigenetics refers to heritable changes in DNA and chromatin that impact gene expression without changes in DNA sequence. There are 2 basic epigenetic regulatory mechanisms: DNA methylation and histone modifications. DNA methylation refers to the covalent modification of post-replicative DNA, which adds a methyl group to the

Fertility and Sterility姞 Vol. 87, No. 1, January 2007 Copyright ©2007 American Society for Reproductive Medicine, Published by Elsevier Inc.

0015-0282/07/$32.00 doi:10.1016/j.fertnstert.2006.05.077

cytosine ring to form methyl cytosine. In mammals, this modification is found mainly in CpG dinucleotides (16) and is the most important epigenetic alteration in eukaryotes. We recently have reported aberrant methylation of HOXA10 in eutopic endometrium of women with endometriosis, which may be responsible for its aberrant expression (17). We also have found that the promoter region of PR-B has been hypermethylated in ectopic endometrium (18). The PR-B promoter hypermethylation may be responsible for reported down-regulation of PR-B (2) and might well explain the reason that approximately 9% of women with endometriosis do not respond well to progestin treatment (19). With these findings, we wondered whether the aberrant methylation that we observed is merely the tip of an iceberg. The molecular mechanisms underlying aberrant methylation are still not well understood (20). Over-expression of the enzymes that catalyze DNA methylation (i.e., DNA methyltransferase 1, 3A, and 3B) may be a prerequisite (21). DNMT1 is the most abundant methyltransferase in mammalian cells, with a preference for hemimethylated DNA and localization to the replication foci (22). DNMT3A and 3B, on the other hand, usually are involved in establishing patterns of DNA methylation during embryonic development (23). Hence, DNMT1 is often viewed as a “maintenance” methyltransferase; DNMT3A and 3B are viewed as de novo ones. Recent research, however, espouses the view that DNMT1 and DNMT3A and 3B cooperatively maintain DNA methylation (24, 25). In this study, we examined the expression of DNMT1 and DNMT3A and 3B in eutopic and ectopic endometrium of women with endometriosis and compared the expression levels with that in the endometrium of healthy women without endometriosis. We also carried out immunofluorescence staining in these tissues. MATERIAL AND METHODS Tissue Collection After informed consent was obtained, endometriotic and endometrial tissues were taken from 17 patients, aged 25–51 years, with endometriosis (cases) with surgically and histo-

logically confirmed American Fertility Society stages II–IV endometriosis. Among them, 6 patients had both endometrial and endometriotic tissue samples; the rest of the patients had either endometrial or endometriotic tissue sample. Endometrial biopsy specimens were obtained from 8 healthy women, aged 22– 42 years, who underwent tubal sterilization that was confirmed laparoscopically to be free of endometriosis and were used as controls. For each subject, the phase of menstrual cycle at the time of tissue harvesting was determined according to the criteria of Noyes et al (26) and recorded. All endometriotic and endometrial tissue samples were snap frozen on dry ice immediately after surgical dissection and then stored in a ⫺80°C freezer. This research was approved by the Institutional Review Board of Medical College of Wisconsin. Sample Preparation Tissue sections were prepared as previously described (9). Approximately 2,000 epithelial cells were harvested from each sample with the use of laser capture microdissection through a laser capture microscope system (Pixcell II; Arcturus Engineering, Mountain View, CA). RNA Isolation and Complementary DNA Synthesis Total RNA was isolated from epithelial cells with an RNA isolation kit from Arcturus Engineering. Ten nanograms of total RNA was treated with DNase I to remove potential DNA contamination and then reverse-transcribed in a final volume of 20 ␮L that contained 1 ⫻ 1st strand complementary DNA (cDNA) synthesis buffer, 5 mmol/L dNTPs (Invitrogen, Carlsbad, CA), 20 units of RNasin RNase inhibitor (Promega, Madison, WI), 200 units of Superscript II Reverse Transcriptase (Invitrogen). Samples were incubated at 42°C for 1 hour. The reactions were terminated at 94°C for 2 minutes. First-strand cDNA that was obtained from each RNA sample was stored at ⫺80°C until used. To measure the messenger RNA abundance of DNMT1, DNMT3A, and DNMT3B in each sample, semiquantitative polymerase chain reaction (PCR) was performed initially. The messenger RNA abundance of proliferating cell nuclear antigen (PCNA) and glyceraldehyde3-phosphate dehydrogenase (GAPDH) were used as prolif-

TABLE 1 Primers used in the real-time RT-PCR analysis. Gene DNMT1 DNMT3a DNMT3b PCNA GPDH

Forward

Reverse

Size (bp)

5=-TACCTGGACGACCCTGACCTC-3= 5=-TATTGATGAGCGCACAAGAGAGC-3= 5=-GGCAAGTTCTCCGAGGTCTCTG-3= 5=-CTCCAACTTCTGGGCTCAAG-3= 5=-CGACCACTTTGTCAAGCTCA-3=

5=-CGTTGGCATCAAAGATGGACA-3= 5=-GGGTGTTCCAGGGTAACATTGAG-3= 5=-TGGTACATGGCTTTTCGATAGGA-3= 5=-GTAAACGGACTGCTGGAGGA-3= 5=-AGGGGTCTACATGGCAACTG-3=

103 111 113 210 232

Wu. Aberrant expression of DNMTs in endometriosis. Fertil Steril 2007.

Fertility and Sterility姞

25

erative and endogenous controls, respectively. The details of cDNA primers that were designed for subsequent PCR were described (27), and the primers are listed in Table 1. Real-Time PCR Quantitative real-time reverse transcriptase–PCR (RT-PCR) was carried out on a Sequence Detection System (ABI 7900; Applied Biosystems, Foster City, CA) and monitored by 1x SYBR Green master mix (Qiagen, Valencia, CA). Four microliters of first-strand cDNA was added to a total volume of 25 ␮L that contained 1x SYBR Green master mix (Qiagen) and 0.2 ␮mol/L each primer. The PCR reaction was performed with 45 cycles at 95°C for 15 seconds and 60°C for 1 minute initiated with 95°C for 10 minutes. The PCR products of the expected size were visualized on a 0.8% agarose gel. The relative messenger RNA level of each gene was calculated with relative quantitation of gene expression (Applied Biosystems). The means of 3 replicated measurements were calculated and shown as mean (S.D.). Immunofluorescence Staining Expression of DNMT1 in epithelial cells were detected with rabbit anti-DNMT1 and fluorescein isothiocyanate conjugate conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology Inc, Santa Cruz, CA). Mouse monoclonal anti-cytokeratin (8/18) (NCL-5D3; Novcastra, Newcastle-upon-Tyne, UK) and fluorescein isothiocyanate conjugated goat anti-mouse IgG (Biomeda, Foster City, CA) were used to distinguish epithelial cells from stromal cells in glands. The frozen tissue sections were fixed initially in 100% cold acetone for 2 minutes. After being washed with cold phosphate-buffered saline solution, sections were blocked with phosphate-buffered saline solution that contained 10% normal goat serum (blocking solution) for 2 hours at room temperature. The primary antibody for DNMT1 was diluted at 1:50; the antibody for cytokeratin (8/18) was diluted at 1:25 dilution in blocking solution, and they were both incubated on the sections for overnight at 4°C. Sections that had been treated with rabbit IgG, but without primary antibody, served as negative controls and nonspecific binding controls for primary antibody reaction. After being washed with cold phosphate-buffered saline solution, fluorescein isothiocyanate conjugated secondary antibody diluted at 1:500 for DNMT1 and 1:200 for cytokeratin (8/18) were added to the section and incubated for 2 hours at room temperature. After being washed with phosphate-buffered saline solution, the sections were mounted by coverslips with anti-fade mounting medium (Biomeda). Fluorescence staining was visualized with a microscope (Nikon E-600 Epi-Fluorescence Microscope; Nikon, Melville, NY). Statistical Analysis For semiquantitative PCR data, unpaired t-test with unequal variances was used to test whether there was a difference in 26

Wu et al.

the mean between ectopic or eutopic endometrium and the control group; paired t-test was used for the paired ectopic and eutopic data. Because of the skewed distribution of data, a randomization test was used for quantitative real-time PCR data to test whether the group means are different. All randomization tests were based on 20,000 permutations under the null hypothesis of no difference. Sample median of each group was presented because sample median is more statistically robust than sample means. To test the significance of Pearson’s correlation coefficient, the t-test based on Fisher’s Z transform was used. All data were square-root transformed before the calculation because of skewed distribution. All computations were carried out in R 2.2.0, a language and environment for statistical computing and graphics (www.r-project.org). The significance level was set at 5%. RESULTS The characteristics of recruited subjects, along with information on which patients’ ectopic or eutopic or both were used for analysis, are listed in Table 2. We initially examined the expression levels of DNMT1, DNMT3A, and DNMT3B by semiquantitative RT-PCR analysis for 6 paired ectopic and eutopic endometrial tissue samples; 6 normal endometrial tissue samples from 6 loosely age-matched, endometriosis-free women who underwent tubal ligation were used for controls. The mean ages of the cases and control subjects were 35.5 ⫾ 6.4 years old and 30.5 ⫾ 7.4 years old, respectively. Hence, the age distributions in the 2 groups were comparable. The 2 groups had roughly comparable distributions of menstrual phases; the percentage of tissues in the proliferative, secretory, and menses phases were 83.3%, 16.7%, and 0 for the cases and 62.5%, 25.0%, and 12.5% for the control group, respectively. The 2-sided Fisher’s exact test revealed that there was no difference at all in the distributions of menstrual phases between any 2 of the 3 groups (all probability values were 1.0). We harvested epithelial cells from these samples using laser capture microdissection (LCM). PCR products that corresponded to DNMT1 (103 base pair [bp]), DNMT3A (111 bp), DNMT3B (113 bp), and GAPDH (232 bp) from the LCM captured epithelial cells were generated (Fig. 1A). The expression levels of DNMT1, DNMT3A, and DNMT3B relative to GAPDH were determined (Fig. 1B). The expression level of DNMT1 was significantly higher in ectopic endometrium (P⬍.0001) and eutopic endometrium (P⫽.03) as compared with control subjects (mean relative expression level ⫾ SD: ectopic, 0.96 ⫾ 0.08; eutopic, 0.60 ⫾ 0.10; control, 0.45 ⫾ 0.45). The difference between ectopic and eutopic endometrium also was statistically significant (P⬍.0001). The expression level of DNMT3A was higher (P⫽.001) in ectopic endometrium, as compared with its paired counterpart or controls (ectopic, 2.0 ⫾ 0.24; eutopic, 1.54 ⫾ 0.53; control, 1.35 ⫾ 0.27). Similarly, the expression level of DNMT3B was significantly higher (P⫽.028) in ectopic endometrium, as compared with its paired eutopic

Aberrant expression of DNMTs in endometriosis

Vol. 87, No. 1, January 2007

TABLE 2 Characteristics of women with endometriosis and control subjects. Patient P1 P2 P3b P4b P5b P6 P7b P8 P9 P10 P11 P12b P13b P14c P15c P16c P17c C1 C2 C3 C4 C5 C6 C7 C8

Age (y)

American Fertility Society stage

Location of the lesion

Menstrual phasea

34 43 44 36 33 33 36 51 30 31 33 28 36 46 25 28 36 36 26 39 42 25 27 27 22

III IV IV IV III IV III III II III III III III III III IV III — — — — — — — —

Right ovarian lesion Right ovarian lesion Left ovarian lesion Left ovarian lesion Peritoneal lesion Peritoneal lesion Peritoneal lesion Peritoneal lesion Pelvic lesion Ovarian lesion Ovarian lesion Ovarian lesion Peritoneal lesion Ovarian lesion Right ovarian Ovarian lesion Ovarian lesion N/A N/A N/A N/A N/A N/A N/A N/A

Secretory Proliferative Proliferative Proliferative Proliferative Secretory Proliferative Proliferative Proliferative Secretory Proliferative Menstrual Proliferative Secretory Secretory Proliferative Proliferative Proliferative Proliferative Proliferative Proliferative Secretory Proliferative Early secretory Menstrual

Note: P ⫽ Patient; C ⫽ control subject; N/A ⫽ not applicable. a Early and late proliferative phases have been grouped into the proliferative phase. b Both endometrial and endometriotic samples were available for analysis. c Only eutopic endometrial tissue samples were available for analysis. Wu. Aberrant expression of DNMTs in endometriosis. Fertil Steril 2007.

endometrium or controls (ectopic, 0.99 ⫾ 0.51; eutopic, 0.45 ⫾ 0.21; control, 0.35 ⫾ 0.18). The expression levels of DNMT3A and DNMT3B in eutopic and control were similar (P⫽.46 and .42, respectively). Hence, the semiquantitative RT-PCR analysis found that the expression of all 3 genes were significantly higher in the ectopic endometrium than in the eutopic endometrium or normal control endometrium, but the expression of only DNMT1 in eutopic endometrium was significantly higher than control subjects. We further investigated DNMT expression using real-time quantitative PCR, a more sensitive and accurate method than the semiquantitative PCR. We examined a total of 13 ectopic endometrial samples and 10 eutopic endometrial samples from endometriotic women with various menstrual phases. Eight normal endometrial tissue samples from endometriosisfree women were used as controls. Total RNA was isolated from laser-captured epithelial cells. Fertility and Sterility姞

For the ectopic endometrium, eutopic endometrium, and the control group, the mean ages were 36.0 ⫾ 6.4, 34.8 ⫾ 6.7, and 30.5 ⫾ 7.4 years, respectively. The 3 groups had comparable distributions of menstrual phases: the percentage of tissues in the proliferative, secretory, and menses phases were 69.2%, 23.1%, and 7.7% for the ectopic endometrium group, respectively, 70.0%, 20.0%, and 10.0% for the eutopic endometrium group, respectively, and 62.5%, 25.0%, and 12.5% for the control group, respectively. Two-sided Fisher’s exact test revealed that there was no difference at all in the distributions of menstrual phases between any 2 of the 3 groups (all probability values were 1.0). Therefore, the 3 groups were fairly comparable in terms of age and menstrual phase. To adjust for possible differences in proliferation during different phases of the menstrual cycle, the expression levels of all DNMTs were normalized against PCNAs 27

FIGURE 1 Semiquantification by RT-PCR of messenger RNA expression of DNMT1, DNMT3A, and DNMT3B in 6 pairs of ectopic and eutopic endometrial epithelium and in 6 control samples. (A) Lanes 1– 6: ectopic endometrial epithelium; Lanes 7–12: eutopic endometrial epithelium; Lanes 13–18, normal endometrial epithelium from endometriosis-free women. Lane 19, no template as a negative control for PCR. GAPDH served as endogenous control. (B) Expression levels of DNMTs relative to GAPDH in 18 tissue samples.

Wu. Aberrant expression of DNMTs in endometriosis. Fertil Steril 2007.

(28), because its expression has been shown to be increased in the proliferative phase in humans (29). We found that relative expression of all DNMTs were significantly higher in the ectopic endometrium as compared with control subjects, with probability values being .021, .005, and .043 for DNMT1, DNMT3A, and DNMT3B, respectively. The fold changes in ectopic endometrium as compared with control subjects were 2.54, 3.40, and 4.16 for DNMT1, DNMT3A, and DNMT3B, respectively. For eutopic endometrium, the corresponding fold changes were 0.85, 1.93, and 2.70, respectively, but only the DNMT3A expression level was significantly higher in the eutopic endometrium than the control subjects (Fig. 2A). Although the expression levels of DNMT3A and DNMT3B were similar in ectopic endometrium and eu28

Wu et al.

topic endometrium groups (P⬎.05), that of DNMT1 was significantly higher in the former group than the latter group (P⫽.015; Fig. 2A). Normalization with GAPDH as a reference standard yielded similar results (Fig. 2B), except for DNMT3B, in which no statistically significant difference was found between the ectopic endometrium and the control subjects. Similar to the PCNA normalization, only the DNMT1 expression level was significantly higher in the eutopic endometrium as compared with control subjects. In addition, the expression levels of DNMT1 and DNMT3B were similar in ectopic endometrium and eutopic endometrium groups (P⬎.05), but that of DNMT3A was significantly higher in the former group than the latter group (P⫽.007). The fold changes in ectopic endometrium as compared with controls were 2.33,

Aberrant expression of DNMTs in endometriosis

Vol. 87, No. 1, January 2007

FIGURE 2 Median gene expression levels in the ectopic and eutopic endometrium of women with endometriosis and the control group, as normalized by PCNA (A) and GAPDH (B). The pvalues listed in the figure represent results of randomization tests.

2.50, and 2.67 for DNMT1, DNMT3A, and DNMT3B, respectively. For eutopic endometrium, the corresponding fold changes were 0.66, 2.33, and 2.60, respectively. The fold changes with PCNA normalization and GAPDH normalization were significantly correlated (r ⫽ 0.83; P⫽.041). We also calculated and tested the statistical significance of Pearson’s correlation coefficients of expression levels among the 3 genes. With PCNA as a normalizing measurement and square-root transformation, the correlation coefficients between DNMT1 and DNMT3A, between DNMT1 and DNMT3B, and between DNMT3A and DNMT3B were, respectively, 0.996 (P⬍2.2 ⫻ 10⫺6), 0.828 (P⫽3.0 ⫻ 10⫺8), and 0.817 (P⫽6.4 ⫻ 10⫺8). Similar correlation coefficients, which were calculated on the basis of GAPDH normalization and square-root transformation, were 0.953 (P ⬍ 2.2 ⫻ 10⫺16), 0.444 (P⫽.012), and 0.431 (P⫽.015), respectively. This indicated that the expression levels of these 3 genes are correlated positively. We also performed immunostaining to detect the intracellular expression of DNMT1. Fig. 3 shows the results from control, eutopic, and ectopic endometrium samples. Although the control sample showed negative or focal weak staining for DNMT1, ectopic and eutopic endometrial glands exhibited stronger or focally stronger staining for DNMT1. The result was consistent with quantitative RT-PCR analysis. Because of the unavailability of tissue samples, we did not perform similar experiments for DNMT3A and 3B. DISCUSSION Methylation of DNA refers to the covalent modification of cytosine to 5=-methyl cytosine in the genome. In vertebrates, methylation occurs in the context of cytosines followed by guanine, or the so-called CpG sites. In humans, DNA methylation is a crucial epigenetic modification of the genome that plays an important role in the regulation of gene expression and genomic imprinting (30), cell differentiation (31), and regulation of many cellular processes (32). Aberrant DNA methylation has been recognized as an important mechanism in tumorigenesis (33, 34). In the last 5 years, there has been an increasing recognition that aberrant methylation is associated with a growing list of human diseases (35, 36).

Wu. Aberrant expression of DNMTs in endometriosis. Fertil Steril 2007.

Fertility and Sterility姞

We have found that the expression levels of 3 genes that code for DNA methyltransferase (DNMT1, DNMT3A, and DNMT3B) are over-expressed in the epithelial component of endometriotic implants, as compared with normal control or eutopic endometrium of women with endometriosis. There was indication that DNMT3A is up-regulated in eutopic endometrium of women with endometriosis. We also have found that DNMT1, DNMT3A, and DNMT3B expression levels are correlated positively with each other. Unfortunately, the limited amount of tissue samples available for laboratory analysis, after being halved for histologic exam29

FIGURE 3 Expression of DNMT1 examined by immunofluorescence staining. The ectopic epithelium showed stronger DNMT1 immunoreactivity than eutopic epithelium and normal epithelium from tubal ligation women. Cytokeratin (8/18) was used to distinguish epithelium structures in endometrial glands. Original magnification: ⫻20.

Wu. Aberrant expression of DNMTs in endometriosis. Fertil Steril 2007.

ination, severely limited our effort to carry out immunostaining analysis for more samples and for DNMT3A and 3B as well. The limited sample tissues also precluded us from examining the stromal component of ectopic implants, a topic that warrants further investigation. It should be noted, however, that the gene expression of DNMT1, 3A, and 3B often correlates with their protein expression (37). Our limited immunostaining study on DNMT1 is consistent with this notion. Our results, which were based on PCNA normalization and GAPDH normalization were consistent overall, but they differed for DNMT3B (P⫽.043 in the former vs P⬎.05 in the latter). Factored in the concordance in fold changes with PCNA and GAPDH normalization and the semiquantitative RT-PCR results, which were based on GAPDH, we inferred that the difference was statistically different. This minor discrepancy between PCNA and GAPDH normalization has been observed in other studies (37). For eutopic endometrium, semiquantitative PCR analysis indicated that DNMT1 was over-expressed; quantitative 30

Wu et al.

real-time RT-PCR suggested that DNMT3A was overexpressed. This discrepancy may be attributed to the use of different tissue samples from different patients or other factors yet to be identified. The moderate sample size also may have insufficient statistical power. Given these somewhat incongruent results and uncertainty, it is too early to speculate on the relevance of our current finding with our previous finding of aberrant methylation of HOAX10. Further research is warranted to resolve the incongruence and uncertainty. Ideally, comparison between ectopic and eutopic endometrium would be best made within the same patients. As in many studies that involve human samples, heterogeneity or some less than congruent results are often present. This study is probably no exception. We suspect that the incongruence might be due to, in part, different patients and different methods used and the moderate sample size. The 3 genes coding for DNA methyltransferases that catalyze cytosine methylation have been reported to be overexpressed in numerous cancers (27, 28, 38, 39) and may be

Aberrant expression of DNMTs in endometriosis

Vol. 87, No. 1, January 2007

involved in aberrant methylation that is observed commonly in cancers (34, 40). In endometriosis, gene dysregulation appears to be wide-spread and is responsible for most, if not all, phenotypic aberrations (6 –9, 41– 43). Because endometriosis is a persistent disease, there must be cellular memory of some sort that constitutes a unique cell identity for endometriotic stromal or epithelial cells. Epigenetic regulation, especially through DNA methylation, is a flexible yet stable mechanism for maintaining such a cellular memory, because methylation is heritable in somatic cells and largely irreversible. As Rhee et al. (44) demonstrated, disruption of either DNMT1 or DNMT3B did not change gene-specific methylation and associated gene silencing in vitro. However, when both enzymes were disrupted, methyltransferase activity was nearly eliminated, which caused extensive genomic demethylation. Their study demonstrates convincingly that DNMT1 and DNMT3B together are responsible for most of the DNA methyltransferase activities in cancer cells and presumably in other cells. The positive correlation in gene expression levels among the 3 genes, as demonstrated in this study, indicates a common regulatory pathway because they may be synchronized to be up-regulated in endometriosis. The simultaneous up-regulation of all 3 genes in endometriosis appears to indicate an orchestrated action of both de novo and maintenance methyltransferases, which is likely to further result in aberrant methylation that, in turn, leads to persistent gene dysregulation in endometriosis. It should be noted that, although aberrant methylation so far has been found almost always in cancers, it is not exclusive to cancers. In view of the aberrant expression of DNMT1, DNMT3A, and DNMT3B, there is reason to believe that aberrant methylation may be more rampant than we thought previously. In conjunction with the evidence of aberrant methylation at HOXA10 in the eutopic endometrium of women with endometriosis (17) and at PR-B in the ectopic endometrium, the aberrant expression of the 3 DNMT genes strongly suggests that endometriosis is an epigenetic disease. As such, there is a hope that DNA demethylating agents may be used to restore methylation aberrations (34, 45). Along this line, the identification of DNA methylation aberration as a biomarker for diagnostic and prognostic purposes may also be promising, given the lack of any reliable biomarkers in endometriosis (46, 47). In comparing gene expression levels between different groups, we did not make any adjustment for age, because there has been no report on the effect of age on expression of DNMTs. We did not adjust for pelvic pain or infertility, which are 2 major symptoms of patients with endometriosis. Hence the 2 symptoms, especially the former, are inseparable attributes of endometriosis. On the other hand, the control subjects that were recruited to this study were, by necessity, women without endometriosis who were apparently normal were free of pelvic pain or infertility. Because the prevailing view is that endometriosis is causal, at least in Fertility and Sterility姞

part, for chronic pelvic pain (48) and, to a lesser degree, infertility (49, 50) (instead of the other way around), we believe that pelvic pain or infertility is not a cofounder in this case and does not need to be controlled for. In fact, most, if not all, studies have used similar control subjects without controlling for pain or infertility (51). In calculating the correlation coefficients of DNMT expression levels, we included all tissue samples (endometriotic tissues, eutopic endometrium, and normal endometrium), for the following reasons: First, by definition, the correlation coefficient quantifies the linear relationship, if any, between 2 variables (in our case, the gene expression levels of 2 genes of interest). The gene expression levels of the 3 genes were measured on the same tissue from the same individuals, and it is likely that they are co-regulated by some other factors or genes. If there are other intrinsic factors or genes that regulate the expression levels of these genes, the correlation coefficient should reflect the change as a result of the change in these intrinsic factors. Hence, correlation coefficients that are calculated across all tissue samples appear to be more sensible than calculated within each tissue group. Second, because there will be 3 groups ⫻ 3 genes, there would be 9 correlation coefficients; if we calculated them separately, we would run into problems of much reduced sample size (and thus statistical power) and of multiple comparisons. Six more correlation coefficients would also make it more difficult to grasp the whole picture. In summary, we have demonstrated, to the best of our knowledge, for the first time that DNMT1, DNMT3A, and DNMT3B are over-expressed in endometriotic tissues. This provides another piece of evidence that endometriosis may be ultimately an epigenetic disease. This may open up a new avenue for the development of novel therapeutics for endometriosis and provide a potentially promising way for epigenetic reprogramming that restores dysregulated gene expression. Acknowledgment: We thank 2 anonymous reviewers for their helpful comments on an earlier version of this article.

REFERENCES 1. Zeitoun K, Takayama K, Sasano H, Suzuki T, Moghrabi N, Andersson S, et al. Deficient 17beta-hydroxysteroid dehydrogenase type 2 expression in endometriosis: failure to metabolize 17beta-estradiol. J Clin Endocrinol Metab 1998;83:4474 – 80. 2. Attia GR, Zeitoun K, Edwards D, Johns A, Carr BR, Bulun SE. Progesterone receptor isoform A but not B is expressed in endometriosis. J Clin Endocrinol Metab 2000;85:2897–902. 3. Taylor HS, Bagot C, Kardana A, Olive D, Arici A. HOX gene expression is altered in the endometrium of women with endometriosis. Hum Reprod 1999;14:1328 –31. 4. Osteen KG, Yeaman GR, Bruner-Tran KL. Matrix metalloproteinases and endometriosis. Semin Reprod Med 2003;21:155– 64. 5. Noble LS, Simpson ER, Johns A, Bulun SE. Aromatase expression in endometriosis. J Clin Endocrinol Metab 1996;81:174 –9. 6. Kao LC, Germeyer A, Tulac S, Lobo S, Yang JP, Taylor RN, et al. Expression profiling of endometrium from women with endometriosis

31

7.

8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20. 21. 22.

23.

24. 25.

26. 27.

28.

32

reveals candidate genes for disease-based implantation failure and infertility. Endocrinology 2003;144:2870 – 81. Matsuzaki S, Canis M, Vaurs-Barriere C, Pouly JL, Boespflug-Tanguy O, Penault-Llorca F, et al. DNA microarray analysis of gene expression profiles in deep endometriosis using laser capture microdissection. Mol Hum Reprod 2004;10:719 –28. Matsuzaki S, Canis M, Vaurs-Barriere C, Boespflug-Tanguy O, Dastugue B, Mage G. DNA microarray analysis of gene expression in eutopic endometrium from patients with deep endometriosis using laser capture microdissection. Fertil Steril 2005;84(suppl 2):1180 –90. Wu Y, Kajdacsy-Balla A, Strawn E, Basir Z, Halverson G, Jailwala P, et al. Transcriptional characterizations of differences between eutopic and ectopic endometrium. Endocrinology 2006;147:232– 46. Gogusev J, Bouquet de Joliniere J, Telvi L, Doussau M, du Manoir S, Stojkoski A, et al. Detection of DNA copy number changes in human endometriosis by comparative genomic hybridization. Hum Genet 1999;105:444 –51. Bischoff FZ, Heard M, Simpson JL. Somatic DNA alterations in endometriosis: high frequency of chromosome 17 and p53 loss in late-stage endometriosis. J Reprod Immunol 2002;55:49 – 64. Shin JC, Ross HL, Elias S, Nguyen DD, Mitchell-Leef D, Simpson JL, et al. Detection of chromosomal aneuploidy in endometriosis by multicolor fluorescence in situ hybridization (FISH). Hum Genet 1997;100: 401– 6. Kosugi Y, Elias S, Malinak LR, Nagata J, Isaka K, Takayama M, et al. Increased heterogeneity of chromosome 17 aneuploidy in endometriosis. Am J Obstet Gynecol 1999;180:792–7. Goumenou AG, Arvanitis DA, Matalliotakis IM, Koumantakis EE, Spandidos DA. Microsatellite DNA assays reveal an allelic imbalance in p16(Ink4), GALT, p53, and APOA2 loci in patients with endometriosis. Fertil Steril 2001;75:160 –5. Guo SW, Wu Y, Strawn E, Basir Z, Wang Y, Halverson G, et al. Genomic alterations in the endometrium may be a proximate cause for endometriosis. Eur J Obstet Gynecol Reprod Biol 2004;116:89 –99. Bird AP. CpG-rich islands and the function of DNA methylation. Nature 1986;321:209 –13. Wu Y, Halverson G, Basir Z, Strawn Y, Yan P, Guo SW. Aberrant methylation at HOXA10 may be responsible for Its aberrant expression in the endometrium of patients with endometriosis. Am J Obstet Gynecol 2005;192:371– 80. Wu Y, Strawn E, Basir Z, Halverson G, Guo SW. Promoter hypermethylation of progesterone receptor isoform B (PR-B) in endometriosis. Epigenetics 2006;1(2):106 –11. Vercellini P, Cortesi I, Crosignani PG. Progestins for symptomatic endometriosis: a critical analysis of the evidence. Fertil Steril 1997;68: 393– 401. Ushijima T, Okochi-Takada E. Aberrant methylations in cancer cells: Where do they come from? Cancer Science 2005;96:206 –11. Robertson KD. DNA methylation, methyltransferases, and cancer. Oncogene 2001;20:3139 –55. Lei H, Oh SP, Okano M, Juttermann R, Goss KA, Jaenisch R, et al. De novo DNA cytosine methyltransferase activities in mouse embryonic stem cells. Development 1996;122:3195–205. Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 1999;99:247–57. El-Osta A. DNMT cooperativity: the developing links between methylation, chromatin structure and cancer. Bioessays 2003;25:1071– 84. Ting AH, Jair KW, Suzuki H, Yen RW, Baylin SB, Schuebel KE. Mammalian DNA methyltransferase 1: inspiration for new directions. Cell Cycle 2004;3:1024 – 6. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Fertil Steril 1950;1:3–25. Girault I, Tozlu S, Lidereau R, Bieche I. Expression analysis of DNA methyltransferases 1, 3A, and 3B in sporadic breast carcinomas. Clin Cancer Res 2003;9:4415–22. Xiong Y, Dowdy SC, Xue A, Shujuan J, Eberhardt NL, Podratz KC, et al. Opposite alterations of DNA methyltransferase gene expression in

Wu et al.

29.

30. 31.

32. 33. 34. 35.

36. 37.

38.

39.

40.

41. 42.

43.

44.

45. 46.

47.

48. 49.

50. 51.

Aberrant expression of DNMTs in endometriosis

endometrioid and serous endometrial cancers. Gynecol Oncol 2005;96:601–9. Narkar M, Kholkute S, Chitlange S, Nandedkar T. Expression of steroid hormone receptors, proliferation and apoptotic markers in primate endometrium. Mol Cell Endocrinol 2006;246:107–13. Reik W, Dean W, Walter J. Epigenetic reprogramming in mammalian development. Science 2001;293:1089 –93. Ehrlich M. Expression of various genes is controlled by DNA methylation during mammalian development. J Cell Biochem 2003;88:899 – 910. Robertson KD, Wolffe AP. DNA methylation in health and disease. Nat Rev Genet 2000;1:11–9. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002;3:415–28. Esteller M. Aberrant DNA methylation as a cancer-inducing mechanism. Annu Rev Pharmacol Toxicol 2005;45:629 –56. Scarano MI, Strazzullo M, Matarazzo MR, D’Esposito M. DNA methylation 40 years later: its role in human health and disease. J Cell Physiol 2005;204:21–35. Robertson KD. DNA methylation and human disease. Nat Rev Genet 2005;6:597– 610. Xiong Y, Dowdy SC, Podratz KC, Jin F, Attewell JR, Eberhardt NL, et al. Histone deacetylase inhibitors decrease DNA methyltransferase-3B messenger RNA stability and down-regulate de novo DNA methyltransferase activity in human endometrial cells. Cancer Res 2005;65: 2684 –9. Langer F, Dingemann J, Kreipe H, Lehmann U. Up-regulation of DNA methyltransferases DNMT1, 3A, and 3B in myelodysplastic syndrome. Leuk Res 2005;29:325–9. Jin F, Dowdy SC, Xiong Y, Eberhardt NL, Podratz KC, Jiang SW. Up-regulation of DNA methyltransferase 3B expression in endometrial cancers. Gynecol Oncol 2005;96:531– 8. van Doorn R, Gruis NA, Willemze R, van der Velden PA, Tensen CP. Aberrant DNA methylation in cutaneous malignancies. Semin Oncol 2005;32:479 – 87. Giudice LC, Kao LC. Endometriosis. Lancet 2004;364:1789 –99. Matsuzaki S, Canis M, Darcha C, Fukaya T, Yajima A, Bruhat MA. Increased mast cell density in peritoneal endometriosis compared with eutopic endometrium with endometriosis. Am J Reprod Immunol 1998; 40:291– 4. Arimoto T, Katagiri T, Oda K, Tsunoda T, Yasugi T, Osuga Y, et al. Genome-wide cDNA microarray analysis of gene-expression profiles involved in ovarian endometriosis. Int J Oncol 2003;22:551– 60. Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, Schuebel KE, et al. DNMT1 and DNMT3b cooperate to silence genes in human cancer cells. Nature 2002;416:552– 6. Miyamoto K, Ushijima T. Diagnostic and therapeutic applications of epigenetics. Jpn J Clin Oncol 2005;35:293–301. Kitawaki J, Ishihara H, Koshiba H, Kiyomizu M, Teramoto M, Kitaoka Y, et al. Usefulness and limits of CA-125 in diagnosis of endometriosis without associated ovarian endometriomas. Hum Reprod 2005;20: 1999 –2003. Hornstein MD, Gleason RE, Orav J, Haas ST, Friedman AJ, Rein MS, et al. The reproducibility of the revised American Fertility Society classification of endometriosis. Fertil Steril 1993;59:1015–21. Howard FM. The role of laparoscopy in chronic pelvic pain: promise and pitfalls. Obstet Gynecol Surv 1993;48:357– 87. D’Hooghe TM, Debrock S, Hill JA, Meuleman C. Endometriosis and subfertility: Is the relationship resolved? Semin Reprod Med 2003;21: 243–54. Witz CA, Burns WN. Endometriosis and infertility: Is there a cause and effect relationship? Gynecol Obstet Invest 2002;53(suppl 1):2–11. Wieser F, Vigne JL, Ryan I, Hornung D, Djalali S, Taylor RN. Sulindac suppresses nuclear factor-kappaB activation and RANTES gene and protein expression in endometrial stromal cells from women with endometriosis. J Clin Endocrinol Metab 2005;90:6441–7.

Vol. 87, No. 1, January 2007