Histone Deacetylase Inhibitors as Therapeutics for Endometriosis

Xin Li; Xishi Liu; Sun-Wei Guo


Expert Rev of Obstet Gynecol. 2012;7(5):451-466. 

In This Article

Endometriosis as an Epigenetic Disease

In the last 7–8 years, evidence has been slowly but progressively accumulating in support of the notion that endometriosis is an epigenetic disease. The relevant findings published so far are summarized in the following sections.

HOXA10 Hypermethylation

The very first piece of evidence suggesting that endometriosis may be an epigenetic disease came from a study showing that, compared with women without endometriosis, the putative promoter of HOXA10 is hypermethylated in the endometrium of women with endometriosis.[14]HOXA10 is expressed in the endometrium, and its expression is considerably elevated during the midsecretory phase of the menstrual cycle, which corresponds to the time of implantation and increase in circulating progesterone.[15] This suggests that HOXA10 may be involved in establishing conditions necessary for implantation, that is, uterine receptivity.[16]

However, HOXA10 gene expression is significantly reduced in the endometrium of women with endometriosis, indicating possible defects in uterine receptivity,[17,18] likely responsible for reduced fertility in women with endometriosis. Since promoter hypermethylation is often associated with gene silencing, the reported HOXA10 promoter hypermethylation may be viewed as a proximate cause for reduced HOXA10 gene expression in the endometrium of women with endometriosis.[14] Indeed, it has recently been reported that HOXA10 expression in the eutopic endometrium is reduced and that the HOXA10 promoter is hypermethylated in infertile women with minimal endometriosis.[19]

Within 2 years, HOXA10 promoter hypermethylation has been confirmed in baboon models of endometriosis, which coincides with reduced HOXA10 expression.[20] Similar to the baboon study, the surgical induction of endometriosis in mouse models also resulted in the downregulation of HOXA10 gene expression as well as hypermethylation at the HOXA10 promoter.[21] Thus, endometriosis appears to cause HOXA10 promoter hypermethylation in the eutopic endometrium, leading to its downregulation and possibly to a disturbance in uterine receptivity.

PR-B Hypermethylation

PR-B promoter hypermethylation is the second piece of evidence.[22] Remarkably, this hypermethylation goes hand in hand with reduced PR-B gene expression, providing evidence for the role of epigenetic aberration in the downregulation of PR-B. In endometriosis, there is a general tendency of progesterone resistance.[1] It is also known that PR-B expression is reduced in both eutopic and ectopic endometria.[23] Since progesterone is mediated through its receptors, including PR-B, PR-B downregulation may be responsible, at least in part, for progesterone resistance. Yet, the exact mechanism underlying persistent PR-B downregulation is still unknown.[1,24] PR-B promoter hypermethylation provides a simple yet theoretically plausible explanation as to why there is a general tendency towards progesterone resistance in endometriosis.

In human adenomyosis, the PR-B promoter is also reported to be hypermethylated.[24] More remarkably, treatment with HDACs and the demethylation agent (DMA) inhibits proliferation of adenomyotic stromal cells and their cell cycle progression.[24] Unfortunately, the mouse model of endometriosis did not find PR-B hypermethylation.[25] Whether the mouse model truly recapitulates the human condition is unclear.

Aberrant Expression of DNA Methyltransferases

Given the finding of HOXA10 and PR-B hypermethylation in endometriosis, one naturally may wonder as to whether such aberrations are isolated events or representative of a more general phenomenon. The report that DNMT1, DNMT3A and DNMT3B – the three genes encoding DNA methyltransferases (DNMTs) that are involved in the methylation of genomic DNA – are all overexpressed in endometriosis thus provides an important piece of evidence that these aberrations are not isolated events.[26] However, another study reports that DNMT1 and DNMT3B protein expression in ectopic endometria was significantly lower than that in control endometria.[27] A recent study found that DNMT1 immunoreactivity, but not gene expression, is elevated in eutopic endometria with endometriosis-associated infertility.[28]

These conflicting reports are likely due to the use of different materials: the first study used endometriotic epithelial cells harvested through laser capture microdissection while the second used tissue culture that consists of several mixed cell types. The third study used eutopic endometrial cells. Regardless, these studies suggest aberrant expression of one or more DNMTs in endometriosis. It should be noted that, in view of the report that DNMT1 expression can be regulated post-transcriptionally,[29] it may not be surprising to see elevated DNMT1 immunostaining but no changes in DNMT1 expression. Incidentally, aberrant immunoreactivity of DNMT1, DNMT3A and DNMT3B is found in adenomyosis as well.[30]

Since DNMTs are known to be involved in de novo and maintenance methylation, their dysregulation indicates that aberrant methylation may be rampant in endometriosis. In addition, considering that methylation is closely associated with chromatin remodeling, the dysregulation of these genes suggests that there should be aberrant epigenetic changes other than methylation (such as aberrant histone modification) in endometriosis.

Promoter Hypomethylation at SF-1 & ER

Consistent with the view that aberrant methylation may be rampant in endometriosis, several recent studies provide more evidence for epigenetic aberrations that may be involved in endometriosis. In particular, steroidogenic factor-1 (SF-1) is a transcriptional factor that is essential for the activation of several steroidogenic genes involved in estrogen biosynthesis. In normal endometrial stromal cells, SF-1 is usually undetectable; however, aberrant expression has been demonstrated in endometriotic stromal cells. Notably, a study by Xue et al. showed that the promoter of SF-1 has increased methylation in endometrial cells but is hypomethylated in endometriotic cells.[31] The same group also demonstrated that the ERβ promoter is hypomethylated in endometriotic cells, likely accounting for its overexpression.[32]

Epigenetic Aberrations in Other Genes

Aromatase is not normally expressed in endometrial stromal cells. Izawa et al. report that the treatment of endometrial stromal cells with a DMA, 5-aza-deoxycytidine, was associated with a dramatic increase in expression of the aromatase gene.[33] They also highlighted that a stretch of CpG demethylation occurs within a nonpromoter CpG island of the aromatase gene in endometriotic cells while the same stretch is heavily methylated and associated with methyl-CpG-binding proteins in endometrial cells.[34] Thus, the increased expression of the aromatase gene in endometriosis is likely attributable to the apparent DNA hypomethylation in a nonpromoter CpG island.[34]

It is reported that endometriotic cells lack expression of the intercellular adhesion protein E-cadherin, a known metastasis-suppressor protein in epithelial tumor cells whose deregulated expression has also been lined with invasiveness of endometriotic cells.[35,36] E-cadherin was found to be hypermethylated in two immortalized endometriotic cell lines, and the treatment with trichostatin A (TSA), a histone deacetylase inhibitor, resulted in reactivation of E-cadherin with concomitant attenuated cellular invasion.[37] These findings would suggest that, at least in endometriotic cell lines, silenced E-cadherin, likely by hypermethylation, is associated with increased invasiveness. Szczepanska et al. recently reported that the HOXA11 promoter appears to be hypermethylated in eutopic endometrium from women with endometriosis-associated infertility, along with its reduced expression.[28]

A theme of post-translational modification and histone deacetylation is also suggested by an in silico study incorporating large-scale gene expression profiling data from paired ectopic and eutopic endometria,[38] supporting the role of epigenetic aberration in endometriosis. Consistent with this view, a recent study based on the whole-genome scanning of methylation status in 25,500 promoters compared endometriotic cells in three subtypes of endometriosis (superficial endometriosis, ovarian endometriomas and deep infiltrating endometriosis) with their respective eutopic endometrium, and reported that, with a predetermined threshold of 1.5 or 0.66 (exceeding 1.5 would be considered hypermethylated while lower than 0.66 would be considered as hypomethylated), there are 153 (14), 29 (53) and 19 (20) hypermethylated (hypomethylated) promoters in superficial endometriosis, ovarian endometriomas and deep infiltrating endometriosis, respectively, compared with eutopic endometrium.[39] There are 11 hypermethylated and nine hypomethylated chromosomal regions common to all three subtypes of endometriosis. Hypermethylated regions appear to be located at the ends of chromosomes while hypomethylated regions are found to be randomly distributed along the chromosomes.[39] While these findings are novel, their significance, if any, is unclear. In any case, these whole-genome studies are consistent with a more recent study demonstrating the involvement of the epigenome in the endometrium and its abnormalities in the endometrium in women with endometriosis.[40]

Aberrant Expression of Genes Involved in Histone Modifications

Compared with the work on DNA methylation, there have been few reports on aberrant histone modification in endometriosis. Kawano et al. recently reported decreased acetylated histone H3 and H4 in endometriotic stromal cells compared with normal endometrial stromal cells.[41] Recently, class I HDACs are reported to be overexpressed in endometriosis[42] and adenomyosis.[43]

The authors found that immunoreactivity to lysine-specific demethylase 1 (LSD1 or KDM1A), which can demethylate mono- and di-methylated lysines – specifically histone 3, lysines 4 and 9 (H3K4 and H3K9) – is elevated in endometriotic lesions as compared with normal endometrium [Ding et al., Unpublished data]. In addition, it was found that SIRT1, a class III histone deacetylase, and EZH2, a histone methyltransferase that methylates H3K27, have increased immunoreactivity in endometriotic lesions compared with normal endometrium [Ding et al., Unpublished data].[44] Table 1 compiles a list of epigenetic aberrations in endometriosis identified so far.