Studies in endometriosis have identified several miRNAs differentially expressed in ectopic versus eutopic endometrial tissues (Ohlsson Teague et al., 2009b
), endometrial cells (Pan et al., 2007
) and in women with endometriosis versus controls (Burney et al., 2009
). In general, few miRNAs were detected in more than one study suggesting complex relationships of miRNAs in gene regulatory networks during development of endometrial lesions. Although the relationship between miRNA and mRNA in endometriosis has been evaluated in two studies (Pan et al., 2007
; Ohlsson Teague et al., 2009b
), a link between miRNAs and effects of DNA variation at their target sites has not been previously investigated. We hypothesized that mutations in target binding sites for the differentially expressed miRNAs could alter transcriptional regulation resulting in changes in expression of the target gene(s) and modifying endometriosis risk.
We used stringent criteria to identify target genes for 22 miRNAs differentially expressed (Ohlsson Teague et al., 2009b
) in ectopic and eutopic endometrium. Once target genes were detected, we searched for SNPs in all corresponding sites for each miRNA within the target gene. This search identified 243 SNPs and we genotyped a subset of 102 SNPs in an Australian sample comprising 958 endometriosis cases and 959 controls. Forty-six percent of SNPs were not polymorphic in our sample. Fourteen SNPs showed evidence for rare minor alleles present in no more than three individuals. These were distributed roughly equally between being present in cases only, controls only or in both cases and controls.
Infertility is a major symptom associated with endometriosis and almost half of women with endometriosis are infertile (American Fertility Society, 1985
; Guo and Wang, 2006
). There is growing evidence that molecular dysfunction in eutopic endometrium contributes to endometriosis susceptibility and may also contribute to endometriosis-related infertility. A number of genes that code for proteins such as integrin (alpha V, beta 3), nitric oxide synthase, heat shock protein and interleukin-11, are reported to have an important role in endometrial receptivity (Agarwal et al., 2005
; Bondza et al., 2009
). The WHSC1
gene is located on chromosome 4p16.3 and is associated with Wolf–Hirschhorn syndrome. Patients with this gene defect suffer from craniofacial defects, intellectual disability, growth delay and 30% have genital anomalies including anomalies of the Mullerian system (female reproductive system). In a Whsc1
-deficient mouse model reported that surviving Whsc1
heterozygotes were mostly fertile, but homozygous embryos showed growth retardation and died within 10 days of birth (Nimura et al., 2009
). SNP rs14647 is located near a miRNA target site, between WHSC1
and in the 3′UTR region of both genes (the two genes are transcribed in opposite directions). UTRs are the regulatory elements of genes, and play important roles in RNA's translation and decay. The WHSC1
3′UTR region may be involved in dsRNA degradation when both genes are transcribed with overlapping UTRs. The current study is a genetic study in a large case–control sample set aiming to investigate the potential impact of 102 miRNA SNPs in women with endometriosis. Until now, most SNPs identified in genetic studies are not fully confirmed by functional study. However, it is well established that genetic studies not only attempt to identify functional SNPs but also ‘tag’ the approximate location of disease variants (Wang et al., 2010
). Another unknown SNP in high LD with rs14647 might either increase or decrease the accessibility of the miRNA target site, leading to inactivation of the target site by this or other unknown mechanisms. A recent study by Halvorsen et al.
showed that multiple specific SNPs in UTRs have a structural consequence and can result in a disease phenotype (Halvorsen et al., 2010
). We found the rs14647A allele in the WHSC1
gene was most strongly associated with risk of endometriosis-related infertility with an increased frequency up to 13-fold when compared with our controls, suggesting that the SNP may have local effects on the RNA structural consequence of the associated genes.
We also found significant evidence for association between SLC22A23
haplotypes with SNPs in miR-125a-3p binding sites and predisposition to endometriosis. SLC22A23
belongs to a large family of transmembrane proteins that function as uniporters, symporters and antiporters to transport organic ions across cell membranes (Jacobsson et al., 2007
). The SLC22A23
target site harbours five miRNA target SNPs: rs3813486, rs1127473, rs7742745, rs35091219 and rs3211066. Mutation in the functional SNP rs7742745 was only detected in three cases and not in 959 controls, suggesting a possible role in risk for endometriosis in these patients. Haplotype analysis for three SNPs (rs3813486, rs1127473 and rs3211066), excluding the rare variants, supports a role for variation in SLC22A23
in endometriosis risk. Haplotypes in SLC22A23
were strongly associated with infertility and more severe stage of the disease. The three SNPs are in one major haplotype block, with the common haplotype AGC representing 80.3% of alleles in subfertility cases and 86.1% of alleles in controls. The minor haplotype AGG accounted for 2.2% of alleles in all cases and 0.9% of alleles in all controls. The association detected with these haplotypes suggests that these variants may play a role in expression of the SLC22A23
transcript levels. The association results remained significant after correction of multiple testing, but need to be replicated in independent samples. Comprehensive genotyping in large, powerful, samples with appropriate data on disease subphenotypes will be required to ultimately confirm and characterise the involvement of the gene variation in the pathogenesis of endometriosis.
To evaluate the miRNA target SNPs that potentially contribute to molecular pathways in endometriosis, we performed pathway analysis using IPA software. By assessing 41 miRNA SNPs with the genotype data, IPA identified several overlapping common pathways constituted by two major molecular networks: TNF
, both are centrally located in the cellular growth and proliferation pathways. TNF
is a pro-inflammatory cytokine and is known to play important roles in the pathological processes involved in endometriosis (Barcz et al., 2000
; Braun et al., 2002
is an important tumour suppressor gene and was found to be less expressed in endometriosis lesions when compared with normal controls from immunohistochemical studies (Cameron et al., 2002
; Goumenou et al., 2006
). The functional relationship between TNF/NF-κB
(nuclear factor of kappa light polypeptide gene enhancer in B-cells) signalling and CDKN2A
in tumorigenesis has been explored in a mouse model (Altomare et al., 2009
), but their functions and relationships in the pathogenesis of endometriosis are unclear. In women with endometriosis, activated macrophages secrete TNF
-α, which then activates NF-κB
and other inflammatory mediators to promote epithelial cell proliferation and in turn contributes to the disease progress. Two reports from same group provided evidence that inhibition of TNF
reduces endometriosis. Treatment with recombinant human TNFRSF1A
(tumour necrosis factor receptor superfamily, member 1A) or anti-TNF
antibody in baboons, effectively inhibited the development of endometriosis (D'Hooghe et al., 2006
; Falconer et al., 2006
-α has been shown to inhibit asbestos-induced cytotoxicity via a NF-κB
-dependent pathway (Yang et al., 2006
) and NF-κB
inhibition reduces cell proliferation, but increases apoptosis of endometriotic lesions (Gonzalez-Ramos et al., 2008
). A previous study provided the information that supports the gene-to-function relationship between binding of promoter fragment from the human CDKN1A
gene and human WHSC2
protein occurs in Hct 116 cells (Gomes et al., 2006
). We found evidence for association with functional miRNA SNPs/SNP haplotypes in WHSC1
loci. Both loci have not previously been reported to be associated with endometriosis. Given that these miRNA SNPs impact molecular networks and their potential targets in the pathogenesis of endometriosis, our findings may provide valuable insights for the future investigations.
Although the present work supports our hypothesis that miRNA target SNPs might be involved in the development of endometriosis, it remains to be determined whether the association between these SNPs and endometriosis risk is due to an effect on miRNA binding. Functional studies in ectopic and eutopic endometrial tissues would be required, but the current results need first to be replicated in independent large sample sets to confirm and characterise the involvement of the gene variation in the pathogenesis of endometriosis.