{"paper_id":"2904b0ae-c693-4f44-96e5-9e129a5485de","body_text":"INTRODUCTION\nEndometriosis is a common occurring gynaecological disease characterised by estrogen-dependent chronic inflammatory process that mainly affects chronic pelvic pain, dysmenorrhea and infertility [1]. At present, the main treatment strategies of endometriosis are surgical resection and hormone inhibition. However, due to many side effects and a high recurrence rate, there is no ideal treatment [2]. The mechanism underlying endometriosis progression remains not fully understood. Thence, it is crucial to reveal the underlying mechanism of endometriosis progression and to identify reliable molecular markers with diagnostic and therapeutic value for endometriosis.\nCurrently, with the advances of next-generation sequencing, especially microarray technology, a huge amount of data has been produced, among which the Gene Expression Omnibus (GEO) database is the largest comprehensive international public gene expression database [3]. Gene expression microarray technology has been widely applied in the study of gene expression profile of multiple human diseases, which provide an effective approach for the study of disease-related genes and the discovery of new therapeutic targets and biomarkers [4, 5]. In recent years, some studies have reported the mRNA expression profile of endometriosis [2, 6]. However, due to small sample size, different data processing methods and sample heterogeneity, the results are inconsistent. Integrated bioinformatics analysis can overcome the shortcomings of these single studies and combine different microarray datasets to obtain more stable results [4]. However, the key genes and signaling pathways in the progression of endometriosis remain unclear.\nMicroRNAs (miRNAs) are a class of small non-coding RNA with 18–25 nucleotide RNAs, which mainly inhibits mRNA expression at the transcriptional level [6]. In previous decades, miRNAs have been reported to be involved in many cellular processes, including proliferation and differentiation, and aberrant expression of miRNAs can also cause many diseases, including endometriosis [6, 7]. Although differences in miRNA expression profiles of endometriosis have been previously reported, few studies have focused on integrated analysis of miRNA-mRNA interactions. Herein, we evaluated the interaction between aberrant miRNAs and aberrant mRNAs in endometriosis, and built a network between them.\nObjectives\nThe purpose of our study was to identify differentially expressed mRNAs (DEmRNAs) and differentially expressed miRNAs (DEmiRNAs) in endometriosis, which might provide some candidate mRNAs and miRNAs for the treatment of endometriosis and provide a new theoretical basis for further experimental studies.\nMATERIAL AND METHODS\nRaw data collection\nThe miRNA/mRNA expression profiles in ectopic endometria (EC) and eutopic endometria (EU) tissue samples were downloaded from the GEO database with the keywords “endometriosis” and “Homo sapiens”. 4 mRNA datasets (GSE11691, GSE25628, GSE86534 and GSE105764) and 1 miRNA microarray dataset (GSE105765) were acquired. The GSE11691 series (GPL96 platform) included 19 (9 EC and 2 EU) specimens. The platform for GSE25628 was GPL571 including 6 EC and 7 EU samples. The platform for GSE86534, consisting of 4 EC and 4 EU samples, was GPL20115. GSE105764 included 8 EC and 8 EU tissue samples, and its platform was GPL20301. GSE105765 was in GPL11154 platform, which contained 8 EC and 8 EU tissue samples. The platform and serial matrix files were downloaded. The dataset information is listed in Table 1.\n|\nTable 1. Details of Gene Expression Omnibus (GEO) endometriosis data |\n||||||\n|\nGEO ID |\nSamples (EU:EC) |\nType |\nPlatform |\nYear |\nAuthor |\nType |\n|\nGSE11691 |\n7:9 |\nmRNA |\nGPL96 |\n2008 |\nD.Stephen Charnock-Jones |\nTissue |\n|\nGSE25628 |\n7:6 |\nmRNA |\nGPL571 |\n2013 |\nRaffaele A Calogero |\nTissue |\n|\nGSE86534 |\n4:4 |\nmRNA |\nGPL20115 |\n2019 |\nChen Zhang |\nTissue |\n|\nGSE105764 |\n8:8 |\nmRNA |\nGPL20301 |\n2018 |\nLuyang Zhao |\nTissue |\n|\nGSE105765 |\n8:8 |\nmiRNA |\nGPL11154 |\n2018 |\nLuyang Zhao |\nTissue |\n|\nEU — eutopic endometrial; EC — ectopic endometrial |\nDifferential expression analysis\nDifferential expression analysis of mRNAs was performed using the limma and metaMA package analysis. Inverse normal method was utilized to combine the p value. Benjamini and Hochberg method was used to analyze individual p values. The multiple comparison-corrected false discovery rate (FDR) was then obtained. The FDR < 0.05 and |Combined.ES| > 2 was considered as the cut-off criterion for DEmRNAs. To obtain DEmiRNAs between EC and EU tissue sample, the DESeq2 package in R was used for analysis with the cut-off criteria of FDR < 0.05 and |log2FC| > 3. The volcano map and hierarchical clustering analysis were produced using the R package.\nFunctional enrichment analysis\nWe used Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis to uncover the biological functions of the DEmRNAs. The GeneCodis3 (http://genecodis.cnb.csic.es) was used for the GO enrichment and KEGG pathway enrichment analysis. The cut-off values for the significant of GO and KEGG terms was FDR < 0.05.\nDEmiRNA-DEmRNA interaction analysis\nThe DEmiRNA-DEmRNA interaction pairs in EC vs EU were identified. First, six bioinformatic algorithms (RNA22, miRanda, miRDB, miRWalk, PICTAR2 and Targetscan) were utilized to predict the supposed target DEmRNAs of DEmiRNAs. Then, the DEmiRNA-DEmRNA pairs were recorded by ≥ 4 algorithms. Those DEmRNAs negatively correlated with DEmiRNAs were retained for further investigation. Ultimately, DEmiRNA-DEmRNA interaction networks were built by using Cytoscape 3.7.1 (http://www.cytoscape.org/).\nQuantitative real time-polymerase chain reaction (QRT-PCR)\nFour patients diagnosed as endometriosis and four normal controls were included in this study. This study has been approved by the ethics institute of Binzhou medical university hospital (2020-LW-025). The signed informed consent of all the participants was obtained. Total RNA was isolated using a TRIzol reagent (Invitrogen, USA). Complementary DNA reverse transcription and QRT-PCR were performed by Fast Quant RT Kit (Invitrogen, USA) and Super Real PreMix Plus SYBR Green (Invitrogen, Waltham, MA, USA), respectively. The human GAPDH and hsa-U6 was used as endogenous controls for mRNA and miRNA expression in analysis, respectively. The 2−ΔΔCt method was used to process the data.\nExpression validation and diagnostic analysis of DEmRNAs in online dataset\nIn order to further validate the expression of identified DEmRNAs, GSE7305 and GSE23339 datasets of endometriosis were used for expression analysis. In addition, diagnostic analysis of identified DEmRNAs was performed in the GSE23339 dataset.\nImmune cell infiltration analysis in the tissue of endometriosis\nThe CIBERSORT algorithm (https://cibersort.stanford.edu/) was applied to estimate the burden of immune infiltration subtypes in the tissue of endometriosis, based on a linear support vector regression algorithm with mRNA expression data with systems-level mRNA expression data in the R toolkit.\nRESULTS\nDEmRNAs and DEmiRNAs in endometriosis\nFour datasets (GSE11691, GSE25628, GSE86534 and GSE105764) were used to identify DEmRNAs according to the criteria of FDR < 0.05 and |Combined.ES| > 2. A total of 835 DEmRNAs were obtained between EC and EU, of which 508 mRNAs were up-regulated and 327 mRNAs down-regulated. Top ten up- and down-regulated DEmRNAs between EC and EU are listed in Table 2. Then, the volcano map of all DEmRNAs and the clustering heat map of the top 100 DEmRNAs were established, as presented in Figure 1A and Figure 1B, respectively. GSE105765 dataset was utilized to obtain DEmiRNAs with DESeq2 package in R. Under the threshold of FDR < 0.05 and |log2FC| > 3, 39 DEmiRNAs were identified, including 15 up-regulated and 24 down-regulated miRNAs. The all DEmiRNAs between EC and EU are listed in Table 3. Figure 1C and Figure 1D displayed the volcano map and heat map of all DEmiRNAs, respectively.\n|\nTable 2. The top 10 up- and down-regulated DEmRNAs |\n|||||\n|\nID |\nSymbol |\nCombined ES |\np value |\nFDR |\nUp/down |\n|\n165 |\nAEBP1 |\n5.167795 |\n0 |\n0 |\nUp |\n|\n185 |\nAGTR1 |\n3.486139 |\n0 |\n0 |\nUp |\n|\n358 |\nAQP1 |\n3.694497 |\n0 |\n0 |\nUp |\n|\n633 |\nBGN |\n3.555798 |\n0 |\n0 |\nUp |\n|\n684 |\nBST2 |\n3.953571 |\n0 |\n0 |\nUp |\n|\n730 |\nC7 |\n5.401551 |\n0 |\n0 |\nUp |\n|\n1296 |\nCOL8A2 |\n5.345685 |\n0 |\n0 |\nUp |\n|\n1359 |\nCPA3 |\n3.413792 |\n0 |\n0 |\nUp |\n|\n1475 |\nCSTA |\n4.274204 |\n0 |\n0 |\nUp |\n|\n1809 |\nDPYSL3 |\n4.468443 |\n0 |\n0 |\nUp |\n|\n1832 |\nDSP |\n–3.70545 |\n0 |\n0 |\nDown |\n|\n3216 |\nHOXB6 |\n–4.9245 |\n0 |\n0 |\nDown |\n|\n3664 |\nIRF6 |\n–4.44489 |\n0 |\n0 |\nDown |\n|\n5357 |\nPLS1 |\n–3.57045 |\n0 |\n0 |\nDown |\n|\n5774 |\nPTPN3 |\n–4.15988 |\n0 |\n0 |\nDown |\n|\n7163 |\nTPD52 |\n–3.42889 |\n0 |\n0 |\nDown |\n|\n7368 |\nUGT8 |\n–3.97117 |\n0 |\n0 |\nDown |\n|\n9053 |\nMAP7 |\n–4.37345 |\n0 |\n0 |\nDown |\n|\n10053 |\nAP1M2 |\n–3.80742 |\n0 |\n0 |\nDown |\n|\n10653 |\nSPINT2 |\n–3.57607 |\n0 |\n0 |\nDown |\n|\nES — effect size; FDR — false discovery rate |\n|\nTable 3. All DEmiRNA between ectopic endometrial (EC) and eutopic endometrial (EU) |\n|||||\n|\nSymbol |\nBase mean |\nLog2FC |\np value |\nFDR |\nUp/down |\n|\nhsa-miR-449b-5p |\n1086.373 |\n–6.97688 |\n1.76E-52 |\n1.07E-49 |\nDown |\n|\nhsa-miR-141-3p |\n86518.7 |\n–6.10501 |\n4.94E-44 |\n1.09E-41 |\nDown |\n|\nhsa-miR-202-5p |\n2763.408 |\n9.463128 |\n5.37E-44 |\n1.09E-41 |\nUp |\n|\nhsa-miR-34c-3p |\n107.9907 |\n–4.91143 |\n2.73E-39 |\n4.15E-37 |\nDown |\n|\nhsa-miR-34b-5p |\n1164.497 |\n–5.14853 |\n1.62E-38 |\n1.64E-36 |\nDown |\n|\nhsa-miR-200c-3p |\n6715.331 |\n–4.8971 |\n1.50E-38 |\n1.64E-36 |\nDown |\n|\nhsa-miR-375-3p |\n9495.127 |\n–7.56061 |\n3.29E-36 |\n2.86E-34 |\nDown |\n|\nhsa-miR-449a |\n2592.821 |\n–6.34734 |\n3.65E-35 |\n2.78E-33 |\nDown |\n|\nhsa-miR-708-5p |\n5194.45 |\n3.202732 |\n5.47E-35 |\n3.69E-33 |\nUp |\n|\nhsa-miR-200a-3p |\n5035.966 |\n–5.29445 |\n7.57E-35 |\n4.60E-33 |\nDown |\n|\nhsa-miR-141-5p |\n623.3556 |\n–5.33507 |\n1.31E-34 |\n7.23E-33 |\nDown |\n|\nhsa-miR-449c-5p |\n2901.445 |\n–6.88749 |\n2.80E-34 |\n1.42E-32 |\nDown |\n|\nhsa-miR-34c-5p |\n8806.229 |\n–4.67436 |\n4.91E-31 |\n2.30E-29 |\nDown |\n|\nhsa-miR-196b-3p |\n43.35577 |\n–3.92688 |\n4.16E-29 |\n1.81E-27 |\nDown |\n|\nhsa-miR-200b-5p |\n259.7849 |\n–4.57665 |\n7.06E-27 |\n2.68E-25 |\nDown |\n|\nhsa-miR-200a-5p |\n240.5646 |\n–4.72025 |\n1.53E-26 |\n5.47E-25 |\nDown |\n|\nhsa-miR-34b-3p |\n134.4835 |\n–4.1418 |\n4.63E-21 |\n1.57E-19 |\nDown |\n|\nhsa-miR-216a-5p |\n33.17238 |\n4.37693 |\n1.56E-20 |\n5.00E-19 |\nUp |\n|\nhsa-miR-10a-5p |\n233418.6 |\n–3.43782 |\n1.26E-19 |\n3.84E-18 |\nDown |\n|\nhsa-miR-196b-5p |\n11000.64 |\n–4.02737 |\n1.49E-19 |\n4.31E-18 |\nDown |\n|\nhsa-miR-675-3p |\n162.5129 |\n–3.12799 |\n2.13E-19 |\n5.62E-18 |\nDown |\n|\nhsa-miR-10a-3p |\n778.103 |\n–3.5977 |\n1.10E-18 |\n2.56E-17 |\nDown |\n|\nhsa-miR-223-3p |\n2210.115 |\n3.033881 |\n4.21E-17 |\n8.53E-16 |\nUp |\n|\nhsa-miR-217-5p |\n52.13757 |\n4.165841 |\n6.62E-17 |\n1.30E-15 |\nUp |\n|\nhsa-miR-200b-3p |\n6496.317 |\n–4.06437 |\n5.63E-16 |\n1.04E-14 |\nDown |\n|\nhsa-miR-514a-3p |\n250.334 |\n6.071821 |\n1.92E-14 |\n3.08E-13 |\nUp |\n|\nhsa-miR-187-3p |\n189.1282 |\n–4.31383 |\n5.58E-13 |\n7.89E-12 |\nDown |\n|\nhsa-miR-615-3p |\n117.1398 |\n3.897824 |\n2.16E-12 |\n2.79E-11 |\nUp |\n|\nhsa-miR-509-3p |\n1752.883 |\n6.502398 |\n2.30E-12 |\n2.91E-11 |\nUp |\n|\nhsa-miR-133b |\n15.98328 |\n3.316177 |\n3.28E-12 |\n4.06E-11 |\nUp |\n|\nhsa-miR-203a-3p |\n1936.05 |\n–3.40336 |\n1.27E-11 |\n1.49E-10 |\nDown |\n|\nhsa-miR-513c-5p |\n103.668 |\n6.525797 |\n3.85E-10 |\n3.77E-09 |\nUp |\n|\nhsa-miR-7974 |\n13.6423 |\n–3.3685 |\n5.38E-09 |\n4.36E-08 |\nDown |\n|\nhsa-miR-508-5p |\n33.38197 |\n5.767593 |\n7.19E-08 |\n4.86E-07 |\nUp |\n|\nhsa-miR-873-5p |\n167.9197 |\n–3.52944 |\n1.27E-07 |\n8.04E-07 |\nDown |\n|\nhsa-miR-506-3p |\n24.203 |\n4.292988 |\n2.11E-07 |\n1.26E-06 |\nUp |\n|\nhsa-miR-509-5p |\n30.98636 |\n4.127607 |\n3.95E-07 |\n2.19E-06 |\nUp |\n|\nhsa-miR-513b-5p |\n38.95121 |\n5.924017 |\n6.81E-07 |\n3.67E-06 |\nUp |\n|\nhsa-miR-137-3p |\n20.66388 |\n3.527503 |\n1.60E-05 |\n7.01E-05 |\nUp |\n|\nFC — fold change; FDR — false discovery rate |\nFunctional enrichment analysis of DEmRNAs\nAs showed in Figure 2A, GO analysis results indicated that DEmRNAs were mainly enriched in signal transduction and multicellular organismal development in the biological process. In the cellular component analysis, DEmRNAs were primarily enriched in the cytoplasm and membrane. Molecular function analysis indicated that DEmRNAs mainly were enriched in protein binding, metal ion binding, nucleotide binding, receptor activity and ATP binding. KEGG pathway enrichment analysis displayed that DEmRNAs were markedly enriched in cell adhesion molecules, pathways in cancer, focal adhesion and phagosome (Fig. 2B).\nDEmiRNA-DEmRNA interaction network\nDEmiRNA-DEmRNA interaction network analysis was performed for up-regulated and down-regulated miRNAs, respectively. For the up-regulated miRNAs, we obtained 397 DEmiRNA-DEmRNA pairs including 15 DEmiRNAs and 188 DEmRNAs. The DElncRNA-DEmRNA interaction network was consisted of 203 nodes and 397 edges (Fig. 3). Among the down-regulated miRNAs, a total of 937 DEmiRNAs-DEmRNAs pairs including 24 DEmiRNAs and 330 DEmRNAs were identified. The DElncRNA-DEmRNA interaction network consisted of 329 nodes and 937 edges (Fig. 4).\nFunctional enrichment analysis of DEmRNAs targeted by DEmiRNAs\nAs showed in Figure 5A, GO analysis results indicated that DEmRNAs targeted by DEmiRNAs were mainly enriched in signal transduction, multicellular organismal development and cell adhesion in the biological process. In the cellular component analysis, DEmRNAs targeted by DEmiRNAs were primarily enriched in the cytoplasm, integral to membrane and plasma membrane. In molecular function analysis, DEmRNAs targeted by DEmiRNAs were mainly enriched in protein binding, metal ion binding, nucleotide binding, receptor activity and ATP binding. KEGG pathway enrichment analysis displayed that DEmRNAs targeted by DEmiRNAs were markedly enriched in cell adhesion molecules, pathways in cancer, leukocyte transendothelial migration, cytokine-cytokine receptor interaction and MAPK signaling pathway (Fig. 5B).\nValidation by QRT-PCR\nTo confirm the results of our integrated analysis, the expression of 3 mRNAs (IRF6, PTGER3 and NTRK2) and 3 miRNAs (hsa-miR-133b, hsa-miR-449a and hsa-miR-873-5p) were measured by quantitative real-time PCR in the blood sample of patients with endometriosis and normal controls. As showed in Figure 6, PTGER3 and NTRK2 were up-regulated and IRF6, hsa-miR-133b, hsa-miR-449a and hsa-miR-873-5p were down-regulated in endometriosis compared with normal control. Except for hsa-miR-133b, QRT-PCR results of other mRNAs were consistent with those of the GEO analysis.\nExpression validation and diagnostic analysis of DEmRNAs\nGSE7305 and GSE23339 datasets of endometriosis were used for expression analysis of IRF6, PTGER3 and NTRK2 (Fig. 7). The result showed that IRF6 was significantly down-regulated, PTGER3 and NTRK2 were significantly up-regulated, which was in line with bioinformatics analysis. In addition, diagnostic analysis of IRF6, PTGER3 and NTRK2 was performed in the GSE23339 dataset (Fig. 8). The AUC value of IRF6, PTGER3 and NTRK2 was more than 0.8, which suggested that they had a potential diagnostic value.\nImmune cell infiltration analysis in endometriosis\nThe CIBERSORT algorithm was used to estimate the burden of immune infiltration subtypes in the tissue of endometriosis (Fig. 9). The percentages of 22 kinds of immunocyte subtypes between ectopic endometria and eutopic endometria were detected. Interestingly, infiltration of macrophages M2 (p = 1.3e-08) was the most significantly increased, the infiltration of NK cells activated (p = 3.8e-11) was the most obviously reduced in ectopic endometria.\nDISCUSSION\nEndometriosis is a usual gynecological disease in women’s reproductive stage, and is defined as the hormone-dependent abnormal growth of cells outside the uterus including ovary, peritoneum, intestines and vagina [8]. A small number of endometriosis patients may lead to further tumor formation [9]. Therefore, identification of DEmRNAs and DEmiRNAs is crucial for uncovering the pathogenesis of endometriosis.\nIn the current study, bioinformatics analysis was used to analyze the mRNA expression profile in the GSE11691, GSE25628, GSE86534 and GSE105764 datasets. 508 up-regulated and 327 down-regulated mRNAs were obtained. In addition, GSE105765 dataset was utilized to identify DEmiRNAs with DESeq2 package in R, and 39 (15 up-regulated and 24 down-regulated) DEmiRNAs were identified in endometriosis, which were further used to build a DEmiRNA-DEmRNA interaction network. Finally, the expression of three mRNAs (IRF6, PTGER3 and NTRK2) and three miRNAs (hsa-miR-133b, hsa-miR-449a and hsa-miR-873-5p) were measured by QRT-PCR.\nThere is increasing evidence that the has-miR-449 is silenced or down-regulated in various tumor cells and plays vital roles in tumor cell proliferation, migration and invasion [10, 11]. Zhao et al. [6] have found that hsa-miR-449a is declined in patient with endometriosis. Mu et al. [12] have revealed that hsa-miR-449a may be one of mechanisms by which acupuncture therapy help promotes endometrium receptivity and prepare for in vitro fertilization and embryo transplantation. Given that endometriosis shares many common characteristics with malignant diseases, including invasion and migration, the low expression of hsa-miR-449a is likely to play a crucial role in the pathogenesis of endometriosis. Then, QRT-PCR confirmed that the expression level of hsa-miR-449a was significantly decreased in endometriosis and targeted prostaglandin-endoperoxidase synthase 3 (PTGER3). The expression level of PTGER3 is significantly increased in endometriosis [13]. In this study, PTGER3 was up-regulated in both GEO integration analysis and QRT-PCR verification. Moreover, PTGER3 had a potential diagnostic value for endometriosis. Hence, we speculate that hsa-miR-449a-PTGER3 interaction is part of the pathogenesis of progression of endometriosis.\nHsa-miR-873 is observably down-regulated in ectopic pregnancy than in viable intrauterine pregnancy and spontaneous abortion patients, and it is considered to be a valuable noninvasive and stable biomarker for early detection of ectopic pregnancy [14]. In our study, hsa-miR-873-5p was reduced in both GSE105765 dataset and QRT-PCR verification. Neurotrophic receptor tyrosine kinase 2 (NTRK2) was targeted by hsa-miR-873-5p. NTRK2, a member of the tropomyosin receptor kinase family, has previously been associated with multiple cancers, including lung cancer, breast cancer and pilocytic astrocytoma [15–17]. Importantly, NTRK2 was up-regulated in both GEO integration analysis and QRT-PCR verification. Besides, KEGG pathway enrichment analysis found that NTRK2 was enriched in the MAPK pathway. It is noted that NTRK2 could be considered as a diagnostic biomarker for endometriosis. We infer that hsa-miR-873-5p-NTRK2 interaction plays a pivotal role in the development and progression of endometriosis by regulating MAPK pathway.\nHsa-miR-133b, located on chromosome 6p12.2, is highly expressed in muscle development and initially considered as a muscle-specific miRNA [18]. Subsequent reports have shown that hsa-miR-133b can be widely detected in a variety of tissues. Xiao et al. [19] have found that hsa-miR-133b is up-regulated in oocytes treated with IGF-1 compared with the negative control, and hsa-miR-133b may play important roles in the growth and maturation of oocytes. In our study, hsa-miR-133b was elevated in endometrium and targeted interferon regulatory factor 6 (IRF6). IRF6 belongs to a family of nine transcription factors and has a highly conserved helix-Transshelix-helix DNA binding domain and a less conserved protein binding domain [20]. IRF6 has been reported to be associated with endometriosis [21]. Interesting, IRF6 was one of the top 10 DEmRNAs and down-regulated in both GEO integration analysis and QRT-PCR verification. In addition, IRF6 had a potential diagnostic value for endometriosis. These findings indicate that hsa-miR-133b-IRF6 may be involved in regulating the progression of endometriosis.\nEndometriosis is a chronic inflammatory disease related to the impaired immune response at the site of lesion implantation [22]. The abnormal distribution of immune cells has been found in the abdominal cavity of endometriosis patients [23, 24]. In this study, the CIBERSORT algorithm was used to estimate the burden of immune infiltration subtypes in the tissue of endometriosis. The percentages of 22 kinds of immunocyte subtypes between ectopic endometria and eutopic endometria were detected. We found that the infiltration of macrophages M2 was the most significantly increased, the infiltration of NK cells activated was the most obviously reduced in ectopic endometria. Macrophages belong to the mononuclear phagocyte system that is part of both the innate immune system and the adaptive immune system [25]. It is found that M2 macrophages are associated with resolution of inflammation and promotion of tissue repair [26]. Previous report has confirmed that M2 macrophages infiltrates endometriotic lesions and promotes angiogenesis [27, 28]. In patients with endometriosis with pelvic pain, the proportion of NK cells in the peripheral blood is significantly higher than that of women with stable disease without pelvic pain [29]. Suppressing the cytotoxic activity of NK cells may help to reduce the progression of pelvic pain in endometriosis patients. This suggested that macrophages M2 and NK cells activated may play an important role in endometriosis.\nCONCLUSIONS\nBased on integrated analysis, we obtained 835 DEmRNAs and 39 DEmiRNAs in endometriosis. In addition, a miRNA-mRNA interaction network was constructed to find the crucial DEmiRNA-DEmRNA pairs that play an important role in the pathogenesis of endometriosis. Moreover, key DEmiRNA-DEmRNA pairs were also verified using QRT-PCR. These identified DEmRNAs and DEmiRNAs may be involved in the pathogenesis of endometriosis. However, there are limitations to our study. Firstly, this study is only based on integrated analysis and QRT-PCR validation. Further large sample studies are needed to validate the expression of identified mRNAs and miRNAs in endometriosis. Secondly, some functional experiment is further needed to investigate the deeper molecular mechanism of identified mRNAs and miRNAs in endometriosis.\nArticle information and declarations\nAvailability of data and materials\nThe datasets used and analyzed during the current study are available from public database Gene Expression Omnibus repository. Accession numbers of the datasets used in this study are GSE11691, GSE25628, GSE86534 and GSE105764 in Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo).\nAcknowledgements\nNone.\nConflict of interest\nThe authors declare that they have no conflict of interest.","source_license":"CC0","license_restricted":false}