Role of Non-coding RNAs in the Pathogenesis of Endometriosis

Endometriosis is a disorder characterized by the presence of endometrial glands and stroma like lesions outside of the uterus. Although several hypothesis have tried to explain the underlying cause of endometriosis, yet the main cause remained obscure. Recent studies have shown contribution of non-coding RNAs in the pathogenesis of endometriosis. Two classes of these transcripts namely long non-coding RNAs (lncRNAs) and microRNAs (miRNAs) have mostly attracted attention of researchers. Several studies have reported aberrant expression of these transcripts in affected tissues from patients as well as animal models. Modulation of important signaling pathways such as PI3K/AKT, P38-MAPK, ERK1/2-MAPK and Wnt-β catenin by miRNAs and lncRNAs have potentiated these molecules as biomarkers or therapeutic agents in endometriosis. Single nucleotide polymorphisms with miR-126, miR-143 and miR-146b have been associated with risk of endometriosis. Moreover, miRNAs and lncRNAs control inflammatory responses, cell proliferation, angiogenesis and tissue remodeling, thus understanding the role of these transcripts in endometriosis is a possible way to develop novel diagnostic tests and therapeutic targets for this disorder.


INTRODUCTION
Endometriosis is a condition that endometrial glands and stroma like lesions are detected in organs outside of the uterus (1). These lesions can involve the peritoneum or being presented as superficial implants or cysts on the ovary, or deep infiltrating lesions (2). Although the main etiology of endometriosis is not clear, numerous hypotheses have tried to explain the development of this disorder. Among the most appreciated hypotheses is the retrograde menstruation which can be accompanied by possible hematogenous or lymphatic circulation, thus leading to seeding of endometrial tissue in ectopic places. Yet, this phenomenon is much more prevalent than the occurrence of endometriosis. Hence, other hormonal or immune-related factors contribute in implantation and persistence of lesions in the pelvic cavity (3). Imperfect differentiation or migration of Müllerian residues during fetal period or transdifferentiate of circulating blood cells are other popular hypotheses regarding development of endometriosis (3). Notably, several genomics studies have shown remarkable alterations in gene profile in endometriosis (4). The genetics basis of this condition is complex and has not been explored yet, though, most studies have reported a polygenic/multifactorial mode for its inheritance (4). Most recently, non-coding RNAs have been demonstrated to contribute in the pathogenesis of endometriosis (5). These transcripts have regulatory roles on expression of proteincoding genes, thus regulate several signaling pathways. They are classified into two main classes according to their length: long non-coding RNAs (lncRNAs) with sizes more than 200 nucleotides and microRNAs (miRNAs) with sizes about 20 nucleotides. LncRNAs can regulate the genetic information flow, through modulating chromatin structure, transcription, splicing, mRNA stability, mRNA accesibility, and post-translational alterations. They have interaction domains for DNA, mRNAs, miRNAs, and proteins which are specified by nucleotide sequence and secondary structure (6). NONCODE database has indicated the presence of at least 100,000 lncRNAs in the human genome (7) which significantly surpluses the number of protein coding genes. There are complex interaction networks between lncRNAs and miRNAs. While certain miRNAs can regulate the stability and half-life of lncRNA, lncRNAs can compete with miRNAs for binding with the mRNA target sites (6). Being mostly located in the cytoplasm, miRNAs constitute critical regulators of gene expression. They mostly exert their effects at post-transcriptional level through binding with their targets and subsequent mRNA degradation and/or translational repression. In addition, miRNAs have been shown to exert specific nuclear functions being emphasized by the miRNAguided transcriptional regulation of gene expression (8). The regulatory roles of miRNAs and lncRNAs in the expression of genes indicate their participation on the pathogenesis of human disorders. In the current study, we summarize the role of these transcripts in the pathophysiology of endometriosis.

MIRNAS AND ENDOMETRIOSIS
Several studies have reported aberrant expression of miRNAs in affected tissues or peripheral blood samples of patients. Zhang et al. have extracted exosomes from the serum of patients with endometriosis and healthy subjects, then assessed expression miRNAs by miRNA microarrays. They reported differential expression of 24 miRNAs between these two sets of samples. As confirmed by qPCR, expression of miR-22-3p and miR-320a was increased in serum exosomes of patients compared with controls (9). Another study has shown that exosomal miR-22-3p isolated from peritoneal macrophages increases proliferation, migration, and invasion of ectopic endometrial stromal cells via modulation of the SIRT1/NF-κB signaling pathway (10). Others have assessed expression profile of miRNAs peritoneal or tissue samples obtained from these patients. For instance, Zhou et al. have used miRNA microarray technique to identify miRNA signature in the ectopic endometrium samples. They reported over-expression of miR-3154 and miR-3926 in theses tissues compared with normal endometrium (11). Zhang et al. have isolated mononuclear cells from peritoneal fluid of patients with endometriosis and assessed expression of miRNAs in the supernatant of peritoneal fluid. They also purified human endometrial stromal cells from both endometrial and endometriotic tissues of these patients. They reported up-regulation of miR-146b peritoneal fluid supernatant and CD14 + monocytes/Macrophages of peritoneal fluid in endometriosis patients. This miRNA could inhibit the M1 polarization of endometrial stromal cells cocultured macrophages (12). Table 1 shows the list of up-regulated miRNAs in samples obtained from patients with endometriosis.
Several miRNAs have been shown to be down-regulated during the pathogenic process of endometriosis. Rekker et al. have used fluorescence-activated cell sorting to endometrial stromal cells from paired endometrial and endometrioma biopsies. Subsequently, they profiled miRNAs in endometriotic stroma using high-throughput sequencing method. They reported downregulation of miR-375 in these cells compared to eutopic cells. This miRNA has been shown to regulate expression of the endothelin 1 (EDN1) gene (30). Yang et al. have shown down-regulation of miR-200b, miR-15a-5p, miR-19b-1-5p, miR-146a-5p, and miR-200c while up-regulation of VEGFA in endometriotic tissues. They have speculated that the higher angiogenic and proteolytic activities observed in the eutopic endometrium could assist the implantation of these cells at ectopic regions (39). Table 2 summarizes the function and characteristics of miRNAs that are down-regulated in samples obtained from patients with endometriosis.
Diagnostic power of several miRNAs has been assessed in endometriosis. Maged et al. have shown that serum miR-122 and miR-199a had a sensitivity of 95.6 and 100.0% and a specificity of 91.4 and 100%, respectively, for diagnosis of disease status in women. Thus, these miRNAs are putative serum biomarkers for endometriosis (24). Moustafa et al. have shown up-regulation of miR-125b-5p, miR-150-5p, miR-342-3p, and miR-451a, while down-regulation of miR-3613-5p and let-7b in serum samples of patients with endometriosis compared with controls. The area under curve (AUC) values in receiver operating characteristic (ROC) curves ranged from 0.68 to 0.92 for these miRNAs. Notably, a classifier combining these miRNAs provided an AUC of 0.94 as verified in the independent set of individuals not included in the training set. Importantly, neither phase of menstrual cycle nor use of hormonal medicines affected the expression levels in these miRNAs. Thus, authors concluded the potential of the miRNAs panel in detection of endometriosis in clinical setting (13). Table 3 summarizes the results of studies which reported diagnostic value of miRNAs in endometriosis.
Few studies have reported association between single nucleotide polymorphisms (SNPs) within miRNA coding genes and risk of endometriosis. For instance, Sepahi et al. have genotyped the rs4636297 of miR-126 in 157 endometriosis patients and 252 healthy subjects. G allele of this SNP has been shown to protect against endometriosis. Moreover, significant association has also been detected between the A allele and severity of endometriosis (72). Zhang et al. have shown association between the CT/CC genotypes of miR-146b rs1536309 and the risk of pain symptom of endometriosis. Moreover, they detected lower levels of the miR-146b and higher pro-inflammatory functions in macrophages from CT/CC genotype carriers (12). Nimi-Hoveidi et al. have genotyped miR-143 rs41291957 and rs4705342 SNPs in infertile women with endometriosis and matched healthy subjects. They Let-7b-5p, The family of let-7 in the serum shows a dysregulation in endometriosis.     ---Mentioned-microRNAs could be considered as biomarkers for the diagnosis of endometriosis. (29) reported association between the C allele of rs4705342 and increased risk of endometriosis. In addition, the A allele of rs41291957 polymorphism was associated with susceptibility to endometriosis (73). Table 4 shows the results of studies which assessed association between miRNA SNPs and endometriosis.

Number of cases and controls Variant References
Endometriosis patients (n = 157) and healthy controls (n = 252) miR-126 rs4636297 is associated with endometriosis risk and its severity. For ir-126 rs4636297 in allele (G vs. A) and genotype (GG vs. AA genotype), there was significant protection against endometriosis (72) Endometriosis patients (n = 74) and healthy controls (n = 23) miR-146b rs1536309 C > T polymorphism is associated with the risk of pain symptom of endometriosis. rs1536309 CT/CC frequency is involved in increased pain susceptibility. miR-146b rs1536309 C > T polymorphism by regulating miR-146b expression was associated with the M1 polarization of macrophages (12) Infertile women (n = 77) with endometriosis and healthy controls (n = 226) Among the groups of the study, there was a significant difference in the genotype distribution and allele frequency of miR-143 rs41291957 and miR-143 rs4705342 polymorphism. C allele and TC genotype were associated with an increased risk of endometriosis (73) compared with paired eutopic endometrium tissues in the majority of patients. They also demonstrated higher serum levels of this lncRNA after treatment. Notably, serum levels of UCA1 on the day of discharge were remarkably lower in patients with recurrence compared with patients without recurrence. Based on these results, authors concluded that UCA1 participates in the pathogenesis of ovarian endometriosis and may be a putative diagnostic and prognostic marker for this condition (76). Tables 5, 6 show up-and down-regulated lncRNAs in the endometriotic samples, respectively. Among down-regulated lncRNAs is H19 whose role in the pathogenesis of endometriosis has been shown in Figure 1.   The expression pattern of MALAT1 has been assessed in a number of studies among them is the study by Liang et al. that reported down-regulation of this lncRNA in the endometriosis (48). Figure 2 depicts the molecular mechanism of involvement of MALAT1 in this disorder.

INTERACTION BETWEEN MIRNAS AND LNCRNAS IN THE PATHOGENESIS OF ENDOMETRIOSIS
Based on the significant roles of lncRNAs and miRNAs in the pathogenesis of endometriosis and the presence of functional interactions between these two sets of transcripts, it is expected that lncRNA/miRNA pairs could regulate certain aspects of endometriosis. LncRNAs can act as a competing endogenous RNA (ceRNA) for miRNAs to affect their bioavailability of these transcripts. Assessments in the endometrial tissues have led to identification of a number of miRNAs that are inhibited by the lncRNA H19. For instance, Ghazal et al. have shown that H19 serves as a molecular sponge to decrease the availability of let-7. This interaction leads to over-expression of the downstream target of let-7, IGF1R, thus increasing the proliferation of endometrial stroma cells. They also demonstrated downregulation of H19 in the eutopic endometrium of patients with endometriosis and speculated that the subsequent decrease in the IGF1R activity might diminish endometrial stromal cell proliferation and negatively influence the endometrial receptivity for pregnancy (92). In addition, Xu et al. have demonstrated the role of the estrogen-modulated H19/ACTA2/miR-216a-5p axis in the regulation of invasion and migration of eutopic endometrial stromal cells in subjects with endometriosis (78). Liu et al. have reported the significance of H19/miR-342-3p/IER3 axis in suppression of Th17 cell differentiation and decreasing the risk of endometriosis (93). A recent high throughput study of RNA profile of the ectopic and eutopic endometrium of patients has led to construction of the ceRNA network. Assessment of the RNA interaction network in endometriosis has resulted to identification of the role of miRNAs and lncRNAs that are associated with cyclin-dependent kinase 1 (CDK1) and proliferating cell nuclear antigen (PCNA). These genes regulate the growth and apoptosis of endometrial stromal cells, thus are involved in the pathophysiology of endometriosis. Taken together, the RNA interactive network has critical role in this disorder (102).

DISCUSSION
Several studies have assessed expression profile of lncRNAs and miRNAs in tissues/blood samples obtained from patients with endometriosis. Association between genomic variants of miRNAs and endometriosis has also been another research avenue. However, the latter field has been less explored for lncRNAs. Considering the presence of myriads of SNPs within lncRNA coding genes that modulate their expression and regulatory functions on their targets, assessment of their association with the risk of endometriosis is a necessary step for identification of the role of these transcripts. Noncoding RNAs have fundamental roles in the development of endometriosis. Their role in this process has been highlighted not only by the studies which reported their aberrant expression in patients' samples, but also by the investigations which showed modulation of their expression by therapeutic agents. For instance, Quercetin (3,3 ′ ,4 ′ ,5,7-pentahydroxyflavone) as a phytochemical agent with antioxidant, anti-inflammatory and antiangiogenic characteristics has been shown to suppress the proliferation and induce cell cycle arrest in VK2/E6E7 and End1/E6E7 cells. Moreover, it has exerted antiproliferative and anti-inflammatory impacts on endometriosis autoimplanted mouse models. This effect has been accompanied by induction of miR-503-5p, miR-1283, miR-3714 and miR-6867-5p in both cell lines and stimulation of miR-503-5p and miR-546 expression in the animal model (103). Saponin extract, as another natural therapeutic agent has been shown to reduce expression of miR-21-5p in the human endometrial stromal cells from patients with endometriosis. Suppression of this miRNA could induce apoptosis in these cells. These results imply that he therapeutic Frontiers in Oncology | www.frontiersin.org

FIGURE 1 | (A)
A pro-endometriotic microenvironment produced by an existing endometriotic lesion provides the appropriate micro-environment for the progression of this disorder. After the buildup of cells by a previously established lesion, these elements show distinctive features that destroy immune surveillance (96). (B) H19 levels have been shown to be decreased in the PBMCs of patients with endometriosis. This is accompanied with an increase in miR-342-3p levels. This miRNA binds with the 3 ′ UTR of IER3, thus inhibiting its expression (93). IER3 participates in proteasomal degradation of IF-1. Decrease in the levels of IF-1 leads to increase in reactive oxygen species (ROS) levels (97). ROS increases active extracellular TGF-β levels. This cytokine influences RORγt, thus activating transcription of IL-17 and leading to differentiation of TH17 cells (98).
effect of saponin is exerted through modulation of specific miRNAs (69). Expressions of miRNAs have been assessed in different samples from patients with endometriosis such as endometrium, peripheral blood and peritoneal fluid. There are some cases of inconsistency between these studies. For instance, expression of miR-451a has been shown to be up-regulated in serum (13), exosomes (20) and endometriosis lesions of patients with endometriosis (23) as well as samples obtained from mouse models of endometriosis (22). However, another study has reported downregulation of miR-451 in the eutopic endometrial tissues of patients with endometriosis compared with control tissues (53). The lncRNA UCA1 has been reported to be upregulated in ectopic endometrium tissues compared with paired eutopic endometrium tissues in the majority of patients using qRT-PCR (76). On the other hand, a microarray analysis showed down-regulated of this lncRNA in ovarian ectopic endometrial tissue compared with paired eutopic endometrial tissue (75). Similar discrepancy has been observed for MALAT1. While it has been upregulated in endometrial tissues from patients with endometriosis compared with normal controls (78), it was downregulated in granulosa cells from endometriosis patients compared with controls (91). The heterogeneity of samples and the method of expression analysis can partly explain the inconsistency of these results.
Mechanistically, lncRNAs can sponge miRNAs, regulate expression of inflammatory factors, alter cell proliferation, migration and apoptosis of endometrial cells. They might also affect implantation process (104,105). Several transcription factors and signaling pathways have been regulated by lncRNAs in the endometrial tissues. Examples are HOX genes, N-cadherin, snail, slug, TCF8/ZEB1, matrix metalloproteinase, apoptosis related genes such as caspases and autophagy-related genes such as Beclin1.
The advent of next generation sequencing has enhanced the pace of identification of dysregulated non-coding RNAs in all human diseases including endometriosis. This technique has been applied by Khalaj et al. to identify signature of these transcripts in extracellular vesicles (EVs) obtained from endometriosis patient tissues and plasma samples compared with controls. Authors have demonstrated the presence of distinctive signatures of miRNAs and lncRNAs indicating their participation in the pathogenesis of endometriosis. Dysregulated transcripts were enriched in the pathways related to immune and  (48). MALAT1 increase expression of NF-κB which in turn binds with PGE-2 to enhance its expression (99). PGE-2 activates BCL2/BAX through interaction with EP2/EP4 receptor and suppresses intrinsic apoptotic pathway (100). PGE2 also activates cell proliferation through EP2/EP3 (101). In addition, PGE-2 suppresses MMP2, CD36 and annexin A2 in macrophages, thus inhibits phagocytic activity of macrophages. These effects facilitate implantation and growth of endometrial tissue in the peritoneum (101). PGE-2 influences angiogenic activity and cell cycle progression through increasing expression of VEGF and inhibiting PTEN, respectively (99). MALAT1 can enhance MMP9 levels. MMP9 increases production of the truncated isoform of Steroid receptor coactivator-1. MALAT1 also increases transcription of ZEB1/ZEB2, therefore induces mesenchymal cell phenotype which is accompanied by enhancement of cell migration (48). metabolic functions. Their results indicated that endometriosisassociated EVs transport distinctive cargo and influence the disease course by modulation of inflammation, angiogenesis and proliferation (23). Moreover, exosomal miRNAs isolated from peritoneal macrophages have been shown to increase proliferation, migration, and invasion of ectopic endometrial stromal cells (10). Thus, these transcripts have fundamental roles in the pathogenesis of endometriosis. Taken together, these studies have opened a new research era for identification of the pathophysiology of endometriosis. Another technical development which has facilitated identification of this process has been the cell sorting technique. This technique has paved the way for cell-type-specific analysis of ectopic tissues to recognize the interactions between different cell types during the course of disease (30).
Considering the unavailability of affected tissues in the endometriosis except through invasive methods, identification of biomarkers in the serum of patients has a practical significance. Recent studies have demonstrated appropriate diagnostic power and sensitivity and specificity values for several miRNAs in this regard. Several miRNAs panels are expected to be applied in the clinical settings with high diagnostic power values. In spite of the presence of these supporting results, there is no consensus on a panel for the diagnosis of endometriosis, since most of studies have been conducted in small samples sizes of patients and their results have not been verified in independent samples. Besides, based on the differences in the source of controls, the applied techniques and the biological sources, meta-analysis of the obtained data is complicated. The diagnostic power of lncRNAs in the endometriosis has been less studied. Thus, future studies are needed to assess this aspect as well.
Taken together, based on the results of human and animal investigation, both miRNAs and lncRNAs participate in the pathogenesis of endometriosis. A more comprehensive assessment of these transcripts using the high throughput methods and identification of the functional links between these two sets of transcripts would facilitate identification of the pathogenesis of endometriosis and recognition of possible therapeutic targets in this regard.