{"paper_id":"79b6160f-13c2-4184-b8a3-f4a5a61e0ff0","body_text":"1\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nwww.nature.com/scientificreports\nGenome-wide Long Non-coding \nRNA Analysis Identified Circulating \nLncRNAs as Novel Non-invasive \nDiagnostic Biomarkers for \nGynecological Disease\nWen-Tao Wang1,*, Yu-Meng Sun1,*, Wei Huang1, Bo He2, Ya-Nan Zhao2 & Yue-Qin Chen1\nIncreasing evidence indicates that long non-coding RNAs (lncRNAs) play important roles in human \ndiseases. This study aimed to investigate the tissue and serum lncRNAs that are differentially expressed \nbetween patients with endometriosis, a gynecological disease, to evaluate the potential of these \nlncRNAs as non-invasive markers for the disease. The differentially expressed lncRNAs as competing \nendogenous RNAs (ceRNAs) were also analyzed to predict their functions in disease development. \nGenome-wide profiling of lncRNA expression patterns revealed that many lncRNAs were abnormally \nexpressed between sera and tissuesof the patient samples. A set of aberrant differentially expressed \nlncRNAs were further validated in a validation cohort of 110 serum and 24 tissue samples. Functional \nanalysis predicted that differentially expressed lncRNAs may participate in disease development \nthrough crosstalk between the ceRNAs of miRNAs and may be involved in a range of cellular pathways \nincluding steroid or hormone responses. We also found a unique set of lncRNAs that were associated \nwith disease severity and progression, and their diagnostic values were also investigated. Our study \ndemonstrated that lncRNAs could potentially serve as non-invasive biomarkers for the diagnosis of \nendometriosis and as important regulators in the progression of this disease.\nGenome-wide human transcriptional studies have revealed a large number of non-protein-coding RNAs (ncR-\nNAs), including short and long non-coding RNAs\n1,2. Emerging evidence has shown that long non-coding RNAs \n(lncRNAs), a less characterized class of molecules greater than 200 nucleotides (nt) in length, play important \nroles in a wide range of biological processes. LncRNAs, which are mRNA-like transcripts, are mainly transcribed \nby RNA polymerase II (RNA PII) and are polyadenylated, spliced, and primarily localized in the nucleus\n3,4. \nLncRNAs often form highly stable secondary structures, making it possible to quantitatively detect free RNAs in \nbody fluids, such as serum\n5,6. These characteristics suggested that lncRNAs might not only be potential biomark-\ners for clinical diagnosis of the disease but also be vital factors in disease development. In recent years, reports \nhave suggested that circulating lncRNAs exhibit a predictive value to serve as diagnostic biomarkers in prostate \ncancer\n7, gastric cancer8, B-cell neoplasms9, prenatal testing10,and heart failure11. More importantly, studies have \ndemonstrated that dysregulated expression of lncRNAs can lead to the occurrence and progression of a number \ntypes of diseases, including cancer\n12, leukemia13, and diabetes14. Additionally, lncRNAs might function as com-\npeting endogenous RNAs (ceRNAs) of miRNAs and may be involved in a range of cellular pathways. It has been \nknown that large numbers of miRNA binding sites exist on a wide variety of RNA transcripts, including lncRNAs, \nleading to the hypothesis that lncRNAs contain miRNA-binding sites can communicate with and regulate the tar-\nget mRNAs by competing specifically for shared miRNAs, thus acting as competing endogenous RNAs (ceRNAs) \nto protein coding mRNAs\n15. An example is the lncRNA cardiac hypertrophy related factor (CHRF) that directly \nregulates Myd88 expression as a ceRNA of miR-489, leading to cardiac hypertrophy16. HOTAIR and its targeted \n1Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory for Biocontrol, School of \nLife Science, Sun Yat-sen University, Guangzhou 510275, China. 2Dept of Obst & Gyn, Sun Yat-sen Memorial Hospital, \nSun Yat-sen University, Guangzhou 510120, China. *These authors contributed equally to this work. Correspondence \nand requests for materials should be addressed to Y.-Q.C. (email: lsscyq@mail.sysu.edu.cn)\nreceived: 19 August 2015\nAccepted: 02 March 2016\nPublished: 18 March 2016\nOPEN\n\nwww.nature.com/scientificreports/\n2\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nmiRNA miR-34a also functioned in the process of prostate cancer cell growth inhibited by genistein17. This evi-\ndence indicated there was crosstalk between lncRNAs and small non-coding RNAs in disease development.\nEndometriosis, a common estrogen-dependent gynecology disorder, affects 6 to 10% women of reproductive \nage, 50 to 60% of women and teenage girls with pelvic pain, and up to 50% of women with infertility. The disease is \ncharacterized by the presence of endometrium-like tissues outside the uterus, primarily on the pelvic peritoneum \nand ovaries18. This disease is diagnosed primarily by visualization during surgery, and the present gold standard \nfor the diagnosis of endometriosis is surgical assessment by laparoscopy. As a result, diagnosis and intervention \nare often delayed due to the lack of sensitive biomarkers in the early stages of the disease19. Thus, biomarkers with \nhigh sensitivity, high specificity and low trauma for the diagnosis of endometriosis are needed. In addition, the \npathogenesis of endometriosis is likely multifactorial, and several hypotheses have been suggested to explain the \npresence of ectopic endometrial tissue and stroma; these studies have provided novel biomarkers with potential \nuse for the diagnosis of and treatment strategy for the disease\n20,21. However, it is clear that the pathways involved \nin endometriosis are complicated, and the molecular mechanisms that underlie the process are largely elusive.\nIn this study, we applied genome-wide profiling to investigate the tissue and serum lncRNAs that were dif -\nferentially expressed between endometriosis patients and negative controls and to evaluate the potential of these \nlncRNAs as non-invasive diagnostic markers for the disease\n22. Furthermore, to better understand the potential \nroles of lncRNAs implicated in endometriosis progression, we further analyzed and predicted the functions of \nthese dysregulated lncRNAs.\nResults\nIdentification of differentially expressed lncRNAs between tissue and serum samples of endo-\nmetriosis patients. In an effort to identify lncRNAs that were differentially expressed between patients and \nnegative controls, we first performed a genome-wide lncRNA expression study using the Glue Grant Human \nTranscriptome Array\n23, which contained approximately 39,223 lncRNAs. The arrays were performed with 5 sets of \npooled samples, including a pool of 10 endometriosis serum samples, a pool of 10 control serum samples, a pool \nof 5 eutopic (EU) endometrium tissue samples, a pool of 5 ectopic (EC) endometrium tissue samples and a pool \nof 5 negative tissue controls. The array analysis identified 1682 lncRNAs with dysregulated expression (more than \n2-fold change) in the sera of patients with endometriosis compared with controls (Fig. 1A) and 1435 lncRNAs in \nthe ectopic endometrium compared with the eutopic endometrium (Fig. 1B). Furthermore, among the abnor -\nmally expressed lncRNAs, 125 lncRNAs were present in both the serum and tissue samples; 1557 lncRNAs were \npresent only in the serum set; and 1310 were present only in the endometriosis tissue set (Fig. 1C). Additionally, \namong the 125 deregulated lncRNAs in both serum and tissue, 55 lncRNAs showed the same expression pattern \n(for example, ENST00000544649, ENST00000529000 and ENST00000481067 were up-regulated in both serum \nand tissue), while 70 of the 125 deregulated lncRNAs presented an opposing expression pattern. For example, the \nexpression profiles of ENST00000426472, FR406817 and ENST00000477151 were increased in serum samples \nbut decreased in tissue samples. With further re-analysis of the differentially expressed lncRNAs from the array, \nwe classified the deregulated lncRNAs into different sets; i.e., retained introns, lincRNAs, or antisense RNAs. In \nthe study, we found that antisense RNAs were predominated (~82%; Fig. 1D), which may be generally closed to \nhost genes in the ensemble or NCBI database, indicating that they may present important roles in the process of \nthe disease. Figure 1E shows the top 65 differentially expressed lncRNAs in the tissue and serum samples, which \nclustered into their own biological subtypes. The results suggested that the expression pattern and function of \nlncRNAs in serum may be different from that in tissue, which is similar to that of the miRNAs in the disease\n24–26. \nHowever, further studies are necessary to investigate the origin of circulating lncRNAs.\nWe next investigated the expression profile of lncRNAs among different tissue samples: EC, EU and neg-\native endometrium controls. With unsupervised hierarchical clustering analysis, 60 lncRNAs had the most \ndifferential expression in these three types of tissue samples and clustered into their own biological subtypes \n(Figure S1).These differentially expressed lncRNAs may function in the development and processes of aeutopic \nendometrium, particularly those that are differentially expressed between the EU and control tissues, such as \nENST00000393610, NR_033688, and ENST00000482343. Furthermore, we also analyzed the mRNA expression \ndata in the comprehensive array and found large numbers of deregulated mRNAs (Figure S2), which may have the \npotential to serve as biomarkers for endometriosis\n27 and will also be important in our future studies.\nValidation of specific lncRNAs differentially expressed in serum and tissue and as potential \ndiagnostic biomarkers for the disease. We next endeavored to further validate the lncRNA array accu-\nracy and investigate the clinical application of serum lncRNA. Although relative quantification RT-PCR has been \nwidely used for mRNA and small non-coding RNA detection, this method requires a suitable internal control, \nand no stable, suitable and recognizable standard internal controls have been used for lncRNAs in body fluids. \nTherefore, an absolute quantitation method was proposed for further validation. We therefore first initiated and \ndeveloped a method for standard construction to quantify circulating lncRNAs. The lncRNA templates to con-\nstruct standard curves were designed and synthesized; for detailed procedures, see the Materials and Methods \nsection. The results showed the standard curve of these selected lncRNAs had good efficiency, R\n2 and slope28,29, \nindicating that the method was suitable for circulating lncRNA quantification (Figure S3).\nAccording to the method established, we chose 16 differentially expressed lncRNAs to validate the lncRNA \narray accuracy and investigate the clinical application of serum lncRNA. These selected lncRNAs presented sig-\nnificant deregulated expressions both in serum (serum NC/serum endometriosis group) and tissue (eutopic/\nectopic endometrium group), and all of these lncRNAs presented with highly significant differences. Among \nthose selected, 10 lncRNAs displayed the same expression patterns in both tissue and serum, and 6 lncRNAs \nshowed the opposite expression patterns in serum and tissue. Using the absolute qPCR method and the standard \ncurves specifically constructed for lncRNAs, we validated their expression in the serum sample set consisting \n\nwww.nature.com/scientificreports/\n3\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nof 59 endometriosis patients and 51 negative controls. Eight of the 16 selected lncRNAs could clearly distin -\nguish the disease samples from the control group with high confidence (P <  0.05). For example, the expression \nlevels of NR_038452 and ENST00000393610 were higher in endometriosis serum than in that of the controls, \nwhile the levels of ENST00000465368, NR_033688, ENST00000482343, NR_038395, ENST00000544649 \nand ENST00000529000 were lower in the disease patients than in the negative controls (Fig. 2A–H). We also \ninvestigated these particular serum lncRNAs in tissue samples, which consisted of9 paired EU and EC endo -\nmetrium samples and 6 negative control endometrium samples (Fig. 3). In the tissue samples, except for \nENST00000544649, 7 of the 8 lncRNAs were also found deregulated in endometriosis patients.\nTo explore whether these abnormally expressed lncRNAs could be useful for disease severity detection, we \nreanalyzed these lncRNAs in the subgroups of endometriosis patients serum samples at different stages, such as \nmild (stage I/II) and severe (stage III/IV). Figure 2I–M shows the four lncRNAs with different expression profiles \nin the subgroups. Notably, we found that the expression level of ENST00000482343 continued to decrease as the \ndisease progressed (P <  0.05), whereas the expression levels of NR_033688, NR_038452 and NR_038395 consist-\nently increased with disease severity, although no statistical significance was observed in this group. These results \nsuggested that the expression levels of these lncRNAs may be associated with the severity of the disease.\nInvestigation of circulating lncRNAs for use in the diagnosis of endometriosis.  The results \ndescribed above showed that endometriosis patients display a highly characteristic lncRNA expression pro -\nfile in both serum and tissue samples. We next endeavored to evaluate the diagnostic value of these aberrantly \nexpressed lncRNAs for endometriosis. Receiver operating characteristic (ROC) curve analysis was performed for \nthe expression of the lncRNAs mentioned above, and the associated area under the ROC curve (AUC), as well as \nthe sensitivity and specificity, was used to confirm the diagnostic potency. As shown in Fig. 4, the highest AUC of \na circulating lncRNA was for ENST00000482343, which reached 0.7159 [95% CI: 0.6176–0.8141, P <  0.001], with \n72.41% sensitivity and 71.74% specificity at the cutoff point. We also found that NR_038395 had the greatest sen-\nsitivity, which was 84.75% at the cutoff point among the specific lncRNAs, whereas ENST00000544649 revealed \nthe greatest specificity, which was 91.67% at the cutoff point. Previous studies have combined a single biomarker \nto improve the diagnostic power; therefore, we applied discriminant analysis to further investigate this possibility \nFigure 1. Abnormal expression of lncRNAs in serum and tissue from endometriosis and control patients. \n(A) Scatter plot of circulating lncRNA expression between the endometriosis and control samples; green spots \nshow a difference > 2. (B) Scatter plot of expression of lncRNAs between ectopic endometrium and eutopic \nendometrium; red spots show a difference > 2. (C) Set diagram showing dysregulated lncRNA expression \nbetween the serum and tissue samples; (D) The deregulated lncRNAs were classified into different sets: \nretained introns (4%), lincRNAs (12%), or antisense RNAs (82%);(E) Cluster analysis of lncRNA expression in \nendometriosis patient serum and serum controls, eutopic endometrium tissue samples, ectopic endometrium \ntissue samples and negative tissue controls. The 65 top-ranked, differentially expressed lncRNAs are displayed \n(fold-change >  2.0). The expression values are represented in red and green, indicating expression above and \nbelow the median expression value across all samples, respectively. Each pooled sample has an array of data in \nthe heatmap.\n\nwww.nature.com/scientificreports/\n4\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nFigure 2. Circulating lncRNAs differentially expressed in patients with or without endometriosis. The \nexpression levels of lncRNAs in serum samples from patients with endometriosis (n =  59) and control patients \n(n =  51) were detected with an absolute quantitative RT-PCR assay, and box plots illustrate the distinction. \nThe copy numbers of NR_038452 (A) and ENST00000393610 (B) were higher in endometriosis patients \nthan in the controls (P <  0.01), while the copy number of ENST00000465368 (C) P <  0.05), NR_033688 (D) \nP <  0.01), ENST00000482343 (E) P <  0.001), NR_038395 (F) P <  0.05), ENST00000544649 (G) P <  0.01) \nand ENST00000529000 (H) P <  0.05) were lower in endometriosis patientsthan in the controls. All P \nvalues were determined with a two-tailed Mann-Whitney U test. Additionally, altered expression levels of \nENST00000482343 (I), NR_033688 (J), NR_038452 (K) and NR_038395 (M) were observed in patients with \nvarying levels of endometriosis severity. Notably, the expression level of ENST00000482343 continued to \ndecrease as the disease progressed (P <  0.05). A Kruskal-Wallis test was implemented among the 3 groups,and \nmultiple comparisons were carried out using a LSD-t test. *, **, *** represents P <  0.05, P <  0.01, and P <  0.001, \nrespectively.\n\nwww.nature.com/scientificreports/\n5\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nby analyzing multiple dysregulated lncRNAs. As a result, we achieved an optimal combination of NR_038395, \nNR_038452, ENST00000482343, ENST00000544649 and ENST00000393610 to differentiate patients with and \nwithout endometriosis. The following discriminant equation was determined: predicted value of probabil -\nity (PVP)  =  0.832lnENST00000482343  +  0.230lnENST00000544649-0.536 lnENST00000393610-0.337 lnN-\nR_038395-0.124lnNR_038452-1.104. The AUC was as great as 0.8795 [95% CI: 0.8109–0.9482, P <  0.001], with \n89.66% sensitivity and 73.17% specificity, at the cutoff point of 0.3500 (Fig. 4I). Following our careful assessment \nof the diagnostic value of the selected serum lncRNAs listed above, we suggest that specific circulating lncRNAs \nmay have potential for detecting endometriosis.\nWe also examined if the abnormally expressed lncRNAs were associated with the clinical features of this dis-\nease, including pelvic adhesion and endometriosis with ovarian involvement. Retrospective analysis of lncRNAs \nfollowing Napierian logarithm transformation revealed that the expression levels of a set of lncRNAs varied; \nfor example, ENST00000482343, NR_038395 and ENST00000465368 were decreased in sera from patients with \npelvic adhesion caused by endometriosis (n =  37) compared with those without (n =  22; Fig. 5A–C). ROC curve \nanalysis showed that ENST00000482343 presented the highest AUC of 0.7469 [95% CI: 0.6230–0.8709, P <  0.01] \nwith 75.68% sensitivity and 63.64% specificity. Subsequently, a comparison of endometriosis with (n  =  45) or \nwithout (n =  14) ovarian involvement indicated that ENST00000482343, NR_038395, ENST00000465368 and \nENST00000529000 showed significantly decreased expression levels in patients with ovarian endometrioma \n(Fig. 5D–G). ENST00000482343 also had the greatest AUC of 0.7381 [95% CI: 0.5986–0.8776, P  <  0.01], with \n82.22% sensitivity and 57.14% specificity. Additionally, we correlated the expression of these circulatory lncRNAs \nwith the menstrual cycle. However, only NR_038452 (P =  0.029) showed a difference in expression level between \nthe follicular and luteal phases in patients with endometriosis. We also investigated other clinical features, such \nas infertility, and the degree of dysmenorrhea (mild, moderate, or severe); however, no statistically significant \ndifference was found (data not shown). Together, these results suggested that lncRNAs may have the potential to \ndetect endometriosis or distinguish the different pathological types of the disease.\nDifferentially expressed lncRNAs might function in disease development through ceRNA  \ncrosstalk.  We finally explored the biological process of deregulated lncRNAs, which may be considered \npotential markers for diagnostic endometriosis. Previous studies have hypothesized that numerous lncRNAs \nFigure 3. Aberrant expression profile of lncRNAs between the pairs of EU and EC endometriosis \npatient samples. The expression levels of special lncRNAs in the tissue samples, ENST00000482343 (A) \nENST00000393610 (B) NR_033688 (C) ENST00000465368 (D) ENST00000529000 (E) NR_038395 (F) and \nNR_038452 (G). The pairs of EU and EC endometriosis patient samples (n =  9) and controls (n =  6) were \naccessed with a quantitative RT-PCR assay, and the resultsare shown by dot graphs A one-way ANOV A test was \nimplemented when the comparison was among 3 groups of lncRNAs from tissue, and multiple comparisons \nwere performed using aLSD-t test. *, **, *** represent P <  0.05, P <  0.01, and P <  0.001, respectively.\n\nwww.nature.com/scientificreports/\n6\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\ncontaining many miRNA binding sites can act as competing endogenous RNAs (ceRNA) that involve the post-\ntranscriptional regulation of genes15,30,31; therefore, we constructed an lncRNA-miRNA crosstalk network using \nthe Target Scan database, which can predict the direct interactions between miRNAs and lncRNAs32,33. As shown \nin Figure S4, a large number of lncRNA-miRNA pairs were predicted to have direct interactions. In this crosstalk \nnetwork, an lncRNA that served as the ceRNA of a miRNA might have similar functions with the miRNA and its \ntargeted genes; thus, their functional categories were analyzed with the Database for Annotation, Visualization \nand Integrated Discovery (DAVID)34.\nWe next constructed an lncRNA-miRNA-mRNA network using 7 lncRNAs deregulated in both serum (Fig. 2) \nand tissue samples (Fig. 3), together with 28 miRNAs and their target genes 23,26,35,36 (Fig. 6A), which have been \nshown to play important roles in endometriosis in recent years. As shown in Fig. 6B, clustering of the lncRNAs, \nmiRNAs and their target genes involved in several functional processes occurred, including the processes of cell \nproliferation and growth, cell differentiation and migration, and steroid or hormone responses, which are closely \nrelated to the development of endometriosis. For example, ENST00000465368 is predicted to act as a ceRNA \nof miR-199a, which targets and inhibits the IKK β /nuclear factor-kappa B (NF- κ B) pathway35, and suppresses \nproliferation, migration and angiogenesis of endometrial mesenchymal stem cells by targeting the VEGFA 26, \nimplying that the lncRNA may have the ability to enhance endometrial stromal cell invasiveness and contribute to \nthe pathogenesis of endometriosis. Notably, a number of lncRNAs are predicted to be ceRNAs for many miRNAs; \nfor instance, NR_033688 for miR-10b, miR-29c, and miR-200c. Further,miR-10b inhibits epithelial endometri-\notic cell invasiveness by targeting Syndecan-1 (SDC1)36, suggesting that NR_033688 may associate with disease \nmigration, whereas, endometrial miR-200c influences many events during normal and disease progression, such \nas hormone response cellular transformation, inflammation, and angiogenesis, which indicates this lncRNA may \nbe involved with hormone mediated endometriosis progression. These results showed that lncRNAs might be \ninvolved in different ceRNA crosstalk, contributing to the development of endometriosis. Further studies are \nnecessary to confirm the crosstalk between lncRNAs and miRNAs in endometriosis pathogenesis.\nFigure 4. Assessment of the diagnostic accuracy of these special lncRNAs for endometriosis. Diagnostic \nvalue of serum lncRNAs for endometriosis: NR_038452 (A) ENST00000393610 (B) ENST00000465368 (C) \nNR_033688 (D) ENST00000482343 (E) NR_038395 (F) ENST00000544649 (G) and ENST00000529000 \n(H). The diagnostic power of the combination of the five specific lncRNAs (I): ENST00000482343, \nENST00000393610, ENST00000544649, NR_038395, and NR_038452 for endometriosis.\n\nwww.nature.com/scientificreports/\n7\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nDiscussion\nGenome-wide human transcriptional studies have shown that large numbers of lncRNAs are deregulated in the \ndisease process 1,2,37,38. Many dysregulated lncRNAs have been identified in tissues or in body fluids, and these \nlncRNAs were reported to play important roles in disease development or act as non-invasive biomarkers8–11,37–41. \nHowever, studies on the identification and functional characterization of lncRNAs in gynecological diseases, \nFigure 5. Association between lncRNA expression and clinical features of this disease. LncRNAs with \ndifferent expression levels in serum from patients with different clinical features, such as with or without pelvic \nadhesion (A–C, left) and with or without ovarian involvement (D–G). (A–G), right, shows the diagnostic value \nof specific lncRNAs for the pelvic adhesion and ovarian involvement of endometriosis.\n\nwww.nature.com/scientificreports/\n8\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nespecially in endometriosis, are limited. In this study, we investigated lncRNAs in the sera and tissues of endo -\nmetriosis patients and identified a set of lncRNAs that can discriminate severe vs. mild stages of the disease \nand other associated clinical features. Furthermore, we attempted to characterize the function of dysregulated \nlncRNAs in endometriosis development through the ceRNA crosstalk network. This study is the first to report on \ncirculating lncRNAs in gynecological disease, and it provides an understanding of lncRNAs that are associated \nwith endometriosis.\nIn recent years, non-coding RNAs, such as miRNAs, have been employed as biomarkers with high sensitivity \nand specificity24,46,47 and also as the key regulators in cell processes26, suggesting that non-coding RNA molecules \nhave potential roles in clinical diagnosis and in disease progression, for example, circulating miRNA let-7a–f and \nmiR-135a,b for endometriosis48. Several studies have also reported that lncRNAs, which are similar in length to \nmRNAs, are stable in serum or body fluids and can not only enhance or inhibit disease development but also serve \nas potential biomarkers for many diseases7–11,49. For instance, Trimarchi et al. found a specific Notch-regulated \nlncRNA, LUNAR1, can enhance IGF1R mRNA expression andsustain IGF1 signaling in efficient T-ALL growth, \nand confirmed that lncRNAs are important regulators of the oncogenic state in T-ALL13.Serum lncRNA LIPCAR \nis considered a novel biomarker of cardiac remodeling and is predictive of mortality in heart failure patients 11. \nIn this study, we revealed that lncRNAs could serve as non-invasive biomarkers for endometriosis and may also \ncontribute to the molecular pathogenesis of this disease.\nLncRNAs are a class of molecules greater than 200 nt in length1,2, which might contain more genetic informa-\ntion than miRNAs, which are only 19~24 nt in length50,51. Because of this fact, circulating lncRNAs may present \nmore information in serum when serving as non-invasive markers. We have found that the optimal combination \nof NR_038395, NR_038452, ENST00000482343, ENST00000544649, and ENST00000393610 can differentiate \npatients with and without endometriosis. These lncRNAs might have the potential for disease detection. In this \nstudy, we also found the expression levels of certain lncRNAs were related to the clinical features of this disease. \nAmong these lncRNAs, ENST00000482343 was abnormally expressed in samples representing all of the clinical \nindicators, such as pelvic adhesion. A previous study has showed an association between lncRNA H19 expression \nduring the menstrual cycle and the differentiation state of the human female reproductive tract52. In this study, we \ninvestigated the association of the differentially expressed circulatory lncRNAs with the menstrual cycle, however, \nonly NR_038452 showed a difference in expression level between the follicular and luteal phases in patients with \nendometriosis. This may be the small sample sizes used or it may be the reason that the stages of the menstrual \ncycle may not affect the expression of most of lncRNAs in serum. Further study is necessary to validate the diag-\nnostic value of circulating lncRNAs in a large cohort of samples.\nRecently, both serum5,6,24 and plasma8 are used to extract circulating RNAs. In the plasma, there are differ -\nent anticoagulants, such as EDTA, sodium oxalate, heparin, and trisodium citrate. Some of them, like EDTA, \ncan affect the efficiency of PCR reaction 42,43. So we used serum in our study and chose the Glue Grant Human \nTranscriptome Array microarrays for identification of the circulating lncRNA because the technique has been \nFigure 6. Graphical view of lncRNA-miRNA-mRNA network for lncRNAs. (A) Cluster analysis of the \nexpression data of target genes using the GG-H array. Cluster analysis of lncRNA expression in endometriosis \npatient serum and serum controls, eutopic endometrium (EU) tissue sample, ectopic endometrium (EC) \ntissue sample and healthy tissue control (C). The 68 target genes are displayed. The expression values are \nrepresented in yellow and blue to show expression above and below the median expression value across all \nsamples, respectively. (B) Graphical view of lncRNA-miRNA-mRNA network for 7 candidate lncRNAs. Boxes \ncorrespond to lncRNAs, diamonds correspond to miRNAs, circles correspond to mRNAs, and the edges \ncorrespond to direct interaction links. The most significant regions are marked with background colors, and the \nlabels describe the main functions assigned.\n\nwww.nature.com/scientificreports/\n9\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\ncomprehensively designed to interrogate various aspects of the transcriptome, including gene expression, alter-\nnative splicing, and non-coding transcription. We also pooled samples to improve the products of circulating \nRNAs in serum. Previous reports have validated that pooled samples are good for circulating RNA research44,45. \nThe disadvantage is that the expression pattern of circulating RNAs in each of sample cannot be obtained from the \narray data. Thus verification of the expression profile of circulating RNAs by qPCR in a set of samples is necessary. \nThere is a challenge to quantify these molecules with a normal PCR method. One limitation of this approach is \nlack of stable, suitable and recognizable standard internal controls for lncRNAs. Thus, it is difficult to use rela-\ntive quantification PCR to identify the differential expression patterns of circulating lncRNAs, and an absolute \nquantification method might be appropriate for lncRNA detection. Due to a shortage of synthetic analogues of \nlncRNA from commercial companies\n53, designing and obtaining a suitable lncRNA template for standard curve \nconstruction is essential. In this study, we have designed and cloned these selected circulating lncRNA templates, \nwhich were validated with good efficiency, R\n2 and slope. These synthetic analogues can be used to identify circu-\nlating lncRNAs and provide a source for novel lncRNA detection in body fluids.\nIt has been shown that the expression profile of lncRNAs presents spatial and temporal patterns 1,37,54,55 .\nTherefore, in this study, we examined the expression patterns of lncRNAs in normal, eutopic, and ectopic endo-\nmetrium samples. A number of lncRNAs were differentially expressed in normal, eutopic, and ectopic endome-\ntrium samples, suggesting that they might function in the development and progression of endometriosis\n56,57. \nMore importantly, we also found a number of lncRNAs that act as ceRNAs of miRNAs, and these lncRNAs were \nclustered according to many their biological processes, including cell proliferation and growth\n26,36, cell differenti-\nation andmigration26,35, and steroid or hormone responses36,which are closely related to the development of endo-\nmetriosis. For example, ENST00000465368 was suggested to have the related function of miR-199a, which has the \nability to suppress the invasiveness, proliferation, migration and angiogenesis of endometrial mesenchymal stem \ncells\n26,35.NR_033688 may associate with disease migration for interaction with miR-10b, which inhibits epithelial \nendometriotic cell invasiveness by targeting Syndecan-1 (SDC1)36. These candidate lncRNAs that are aberrantly \nexpressed both in serum and tissue and act as ceRNAs might provide new insight into the molecular mechanism \nof the disease. Further studies are necessary to validate the regulatory network between lncRNAs and miRNAs, as \nwell as the target genes of the miRNAs associated with this disease.\nIn conclusion, we investigated the expression profile of lncRNAs in serum and tissue samples from patients \nwith or without endometriosis. We also established a standard curve that had good efficiency for the quantifica-\ntion of circulating lncRNAs. Using the absolute qPCR method with the standard curves we obtained, we observed \nthat the combination of five circulating lncRNAs, including NR_038395, NR_038452, ENST00000482343, \nENST00000544649 and ENST00000393610, were potential non-invasive biomarkers for endometriosis. Our \nstudy also presented a possible candidate pool of lncRNAs in tissue for future functional studies associated \nwith endometriosis. With the construction of a ceRNA crosstalk network, these candidate lncRNAs clustered \nin relation to various biological processes, suggesting that they may play important roles in the progression of \nendometriosis.\nMethods\nPatient and serum samples. The samples and clinicopathologic data were collected from the Department \nof Obstetrics and Gynecology, Sun Y at-sen Memorial Hospital (Guangzhou, China) in 2014. All the patients \nbrought into this research were suffered from severe dysmenorrhea, pelvic mass or infertility. The negative con-\ntrols were confirmed to be fallopian tubal diseases through laparoscopy and hysteroscopy, with neither endome-\ntriosis nor endometrial lesions. And for the positive cases, the inclusion criteria were as follows: 20–50 years old; \nno hormone therapy for at least 3 months; non-smoker; and no coexisting inflammatory disease. Women suf-\nfering from malignancy, benign ovarian cyst except endometrioma, severe pelvic inflammation observed during \nsurgery, known chronic, systemic, metabolic, or endocrine disease including polycystic ovarian syndrome, were \nexcluded from this study.\nThe study included 59 serum samples from patients diagnosed with peritoneal and/or ovarian endometriosis \nby laparoscopic and pathological examination and 51 control samples from patients primarily diagnosed with \ntubal factor infertilityand confirmed absence of endometriosis during their surgical procedure. Additionally, \nwe examined 9 paired eutopic and ectopic endometrium samples from endometriosis patients and 6 negative \nendometrium controls from patients without endometriosis. The detailed clinical parameters of the cohort are \npresented in Table 1. Table S1 lists the pooled samples of the endometriosis and non-endometriosis patients in \nthe array groups. No significant differences in age and BMI were found. Finally, all patients provided informed \nconsent, and the study was approved by the ethics committee of Sun Y at-sen University. The sample collection and \ntreatment were carried out in accordance with the approved guidelines.\nSerum and tissue processing and RNA isolation. The clinical blood samples from donors who fasted \novernight were left for clotting at room temperature after collection and were then centrifuged within 1 h at \n3000 rpm at 4 °C for 10 min to harvest the serum. Extraction of total RNA from 1  ml of the serum samples was \nachieved using the mirVana PARIS Kit (Ambion, TX). RNA was eluted with 100 μl of 95 °C pre-heated Elution \nSolution. Total RNA was isolated from tissue samples with TRizol (Invitrogen) according to the manufacturer’s \ninstructions. The quantity and quality of total RNA was acceptable when there was an obvious absorbance peak at \n260 nm, measured with a NanoDrop (Thermo Fisher, USA), and then, approximately 500 ng (~10 μg/μl) of RNA \nwas obtained from 1 ml of serum. No difference in the amount of extracted RNA in a unit of serum was found \nbetween the control and endometriosis samples. The total RNA from tissue samples was used only if the ratio of \nthe absorbance at 260 nm and 280 nm (A260/A280) was between 1.8 and 2.2\n44,45. All RNA samples were stored at \n80 °C until further use.\n\nwww.nature.com/scientificreports/\n10\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nFor the Glue Grant Human Transcriptome Array, which included 39,223lncRNAs (Affymetrix, \nUSA),circulating RNA was extracted from two pooled samples from 10 endometriosispatients or 10 \nnon-endometriosis controls (each serum sample was 500 μl, and each pool contained 5 mL)44,45. The array data \nhave been submitted to the NCBI GEO Archive (the accession number is GSE77182). Based on the results of \nlncRNA microarray analysis, specific primers for lncRNAs (Table S2) were synthetized and used to quantify \nlncRNAs in both serum and tissue samples. The reverse transcription of 2 μl of total RNA was carried out using a \nReverTra Ace qPCR RT Kit (Toyobo, Japan). The levels of lncRNAs were measured in triplicate by SYBR Premix \nEx Taq II-based (Takara, Japan) quantitative real-time PCR with ABI Stepone plus (ABI, American). The Cq \nvalue, which ranged from 15 to 35, was identified as applicable. Non-RT-PCR and no cDNA templates served as \nnegative controls, and we therefore preformed an absolute quantitation method.\nStandard curve construction. Due to the current lack of stable, suitable and recognizable standard inter-\nnal controls for lncRNAs in serum, we applied an absolute quantitation method 27,28.In this study, we obtained \nthe pure and accurate lncRNA standards in five steps. The first step was to harvest the target lncRNA template. \nThe standards were analyzed in parallel with the clinical samples under identical qPCR conditions to calculate \nthe start copies of clinical samples in a 20 μl SYBR reaction system. Reverse transcription was carried out using \ntotal RNA templates from a common cell line with specific primers. Target lncRNA sequences with ideal melt \ncurves and sizes were identified using SYBR qPCR and 2.0% agarose gel electrophoresis. Second, to sequence the \nlncRNA templates, the target lncRNAs sequences were cloned and transformed into competent E. coli and were \nsubsequently sequenced by Life Technologies (Thermo Fisher, USA), and the sequences that were 100% aligned in \nBLAST (Basic Local Alignment Search Tool) were considered acceptable. Bacteria carrying the desired sequences \nwere cultivated, and the plasmids were extracted using the Plasmid Plus Midi Kit (QIAGEN, Genman). Third, to \npurify and retrieve the standards. Extracted by the AxyPrep DNA Gel Extraction Kit (Axygen, USA), standard \nsamples were obtained following PCR using plasmid templates and 2.0% agarose gel electrophoresis. Standards \nwere purified by 3 M sodium acetate and alcohol (overnight), and the A260/280 ratio of absorbance was in the \nrange of 1.8–2.0, and the A260/230 ranged between 2.0 to 2.2. Fourth, to ensure exactness of the standards, the \namplification efficiency, R\n2 and the slope 27,28 were used to evaluate the standard curves of these lncRNAs. The \nresults showed the standard curve of these selected lncRNAs had good efficiency, R 2 and slope, suggesting the \nmethod was suitable for circulating lncRNA quantification. The expression levels of lncRNAs in serum were \nquantified by establishing standard curves with a set of serially diluted standard samples, the starting concentra-\ntion of which was determined by spectrophotometry. Finally, we validated the standard curves within the circu-\nlating RNA samples, and almost all detectable signals of serum samples were on the standard curves, indicating \nthat all of the standards were correct. We ensured the standards were exact and the method was reliable using the \nabovementioned methods.\nStatistical analysis. All statistical calculations and figures were performed using SPSS PASW Statistics (ver-\nsion 17.0) and GraphPad Prism (version 5.0). The GraphPad analysis was not only used for figure generations but \nEndometriosis (n = 59) Normal control (n = 51)\nAge, mean ± SD 32.34 ±  7.277 29.56 ±  4.841\nDysmenorrhea 31 22\nMain Diagnosis \n(Besides \nEndometriosis)\nLeiomyoma and Adenomyosis Fallopian Tube Disease\nStage of the menstrual cycle\nFollicular phase 50 44\nLuteal phase 9 7\nPelvic adhesion Caused by endometriosis Caused by inflammation\nPresent 37 24\nAbsent 22 27\nr-AFS Stage\nStage I 12 NA\nStage II 2 NA\nStage III 30 NA\nStage IV 15 NA\nDistribution of Endometriosis\nOvarian Endometrioma 45 NA\nPeritoneal Lesion 14 NA\nDIE status\nWith DIE lesions 4 NA\nWithout DIE lesions 55 NA\nTable 1.  Clinical characteristics of all samples used in the study. NA, not applicable; DIE, deep infiltrating \nendometriosis; r-AFS, revised American Fertility Society.\n\nwww.nature.com/scientificreports/\n11\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\nalso used to perform a Fisher’s exact test and Mann-Whitney U test, which were used to determine the signifi-\ncance of differentially expressed circulating lncRNA levels between the two groups. The Kruskal-Wallis test and a \none-way ANOV A were used when the comparison was made among 3 groups of lncRNAs from serum and tissue, \nand multiple comparisons were made with aLSD-t test. Additionally, a parametric test was used with one-way \nANOV A tests when the comparison was made among 3 groups for lncRNAs from tissue. There weretwo reasons \nfor using non-parametric tests to address the circulating lncRNAs data. On the one hand, the inherent variation \nwasmuch greater, which didnot meet the condition for aparametric test; on theother hand, there was no statis-\ntical significance (P <  0.05) when the parametric test was used, such as anunpaired t test. SPSS PASW Statistics \nwas used for ROC curve analysis, Y ouden’s index and discriminant analysis.The source code of the TargetScan \ndatabase, which was used to search for the candidate targets of conserved 8mer and 7mer sites that matched the \nseed region of miRNA\n32,33 was used for searching the lncRNA that served as the ceRNA of miRNA. The DAVID34 \nwas used to construct the gene network. All P values were two-tailed, and a P <  0.05 was considered statistically \nsignificant. All of the data were analyzed following Napierian logarithm transformation, and unpaired t tests and \nROC curves were performed to determine the diagnostic utility of serum lncRNAs. The optimal cutoff point was \nchosen as the point at which Y ouden’s index was maximal.\nReferences\n1. Guttman, M. et al. Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in mammals. Nature 458, \n223–227 (2009).\n2. Ponting, C. P ., Oliver, P . L. &Reik, W . Evolution and functions of long noncoding RNAs. Cell 136, 629–641 (2009).\n3. Wang, K. C. & Chang, H. Y . Molecular mechanisms of long noncoding RNAs. Mol Cell 43, 904–914 (2011).\n4. Mattick, J. S. The genetic signatures of noncoding RNAs.Plos Genet 5, e1000459 (2009).\n5. Reis, E. M. &Verjovski-Almeida, S. Perspectives of Long Non-Coding RNAs in Cancer Diagnostics. Front Genet 3, 32 (2012).\n6. Dong, L. et al. Circulating CUDR, LSINCT-5 and PTENP1 long noncoding RNAs in sera distinguish patients with gastric cancer \nfrom healthy controls. Int J Cancer 137, 1128–1135 (2015).\n7. Ren, S. et al. Long non-coding RNA metastasis associated in lung adenocarcinoma transcript 1 derived miniRNA as a novel plasma-\nbased biomarker for diagnosing prostate cancer. Eur J Cancer 49, 2949–2959 (2013).\n8. Zhou, X., Yin, C., Dang, Y ., Y e, F . & Zhang, G. Identification of the long non-coding RNA H19 in plasma as a novel biomarker for \ndiagnosis of gastric cancer. Sci Rep 5, 11516 (2015).\n9. Isin, M. et al. Investigation of circulating lncRNAs in B-cell neoplasms. Clin Chim Acta 431C, 255–259 (2014).\n10. Tsui, N. B. et al. Maternal plasma RNA sequencing for genome-wide transcriptomic profiling and identification of pregnancy-\nassociated transcripts. Clin Chem 60, 954–962 (2014).\n11. Kumarswamy, R. et al. The Circulating Long Non-Coding RNA LIPCAR Predicts Survival in Heart Failure Patients. Circ Res 114, \n1569–75 (2014).\n12. Y ang, L. et al. lncRNA-dependent mechanisms of androgen-receptor-regulated gene activation programs. Nature  500, 598–602 \n(2013).\n13. Trimarchi, T. et al. Genome-wide mapping and characterization of Notch-regulated long noncoding RNAs in acute leukemia. Cell \n158, 593–606 (2014).\n14. Reddy, M. A. et al. Regulation of inflammatory phenotype in macrophages by a diabetes-induced long noncoding RNA. Diabetes 63, \n4249–4261 (2014).\n15. Tay, Y ., Rinn, J. & Pandolfi, P . P . The multilayered complexity of ceRNA crosstalk and competition.Nature 505, 344–352 (2014).\n16. Wang, K. et al. The long noncoding RNA CHRF regulates cardiac hypertrophy by targeting miR-489. Circ Res  114, 1377–1388 \n(2014).\n17. Chiyomaru, T. et al. Genistein Inhibits Prostate Cancer Cell Growth by Targeting miR-34a and Oncogenic HOTAIR. Plos One 8, \ne70372 (2013).\n18. Giudice, L. C. Clinical practice. Endometriosis. N Engl J Med 362, 2389–2398 (2010).\n19. Rogers, P . A. et al. Defining future directions for endometriosis research: workshop report from the 2011 World Congress of \nEndometriosis In Montpellier, France. Reprod Sci 20, 483–499 (2013).\n20. Pavone, M. E. & Bulun, S. E. Clinical review: The use of aromatase inhibitors for ovulation induction and superovulation. J Clin \nEndocrinol Meta b 98, 1838–1844 (2013).\n21. Huhtinen, K. et al. Intra-tissue steroid profiling indicates differential progesterone and testosterone metabolism in the endometrium \nand endometriosis lesions. J Clin Endocrinol Metab 99, E2188–E2197 (2014).\n22. Brosens, I., Puttemans, P ., Campo, R., Gordts, S. &Brosens, J. Non-invasive methods of diagnosis of endometriosis. Curr Opin Obstet \nGynecol 15, 519–522 (2003).\n23. Xu, W . et al. Human transcriptome array for high-throughput clinical studies. Proc Natl AcadSci USA 108, 3707–3712 (2011).\n24. Wang, W . T., Zhao, Y . N., Han, B. W ., Hong, S. J. & Chen, Y . Q. Circulating microRNAs identified in a genome-wide serum \nmicroRNA expression analysis as noninvasive biomarkers for endometriosis. J Clin Endocrinol Metab 98, 281–289 (2013).\n25. Santamaria, X. & Taylor, H. MicroRNA and gynecological reproductive diseases. Fertil Steril 101, 1545–1551 (2014).\n26. Hsu, C. Y . et al. miRNA-199a-5p regulates VEGFA in endometrial mesenchymal stem cells and contributes to the pathogenesis of \nendometriosis. J Pathol 232, 330–343 (2014).\n27. Fassbender, A., Burney, R. O., O, D. F ., D’Hooghe, T. & Giudice, L. Update on Biomarkers for the Detection of Endometriosis. \nBiomed Res Int 2015, 130854 (2015).\n28. Li, Y . et al. Serum circulating human mRNA profiling and its utility for oral cancer detection. J Clin Oncol 24, 1754–1760 (2006).\n29. Donia, D., Divizia, M. & Pana’ , A. Use of armored RNA as a standard to construct a calibration curve for real-time RT-PCR. J Virol \nMethods 126, 157–163 (2005).\n30. Cesana, M. et al. A long noncoding RNA controls muscle differentiation by functioning as a competing endogenous RNA. Cell 147, \n358–369 (2011).\n31. Liz, J. et al. Regulation of pri-miRNA Processing by a Long Noncoding RNA Transcribed from an Ultraconserved Region. Mol Cell \n55, 138–147 (2014).\n32. Lewis, B. P ., Burge, C. B. &Bartel, D. P . Conserved seed pairing, often flanked by adenosines, indicates that thousands of human \ngenes are microRNA targets. Cell 120, 15–20 (2005).\n33. Friedman, R. C., Farh, K. K., Burge, C. B. &Bartel, D. P . Most mammalian mRNAs are conserved targets of microRNAs. Genome Res \n19, 92–105 (2009).\n34. Dennis, G. J. et al. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol 4, P3 (2003).\n35. Dai, L., Gu, L. & Di, W . MiR-199a attenuates endometrial stromal cell invasiveness through suppression of theIKKbeta/NF-kappaB \npathway and reduced interleukin-8 expression. Mol Hum Reprod 18, 136–145 (2012).\n36. Schneider, C. et al. Targeting of syndecan-1 by micro-ribonucleic acid miR-10b modulates invasiveness of endometriotic cells via \ndysregulation of the proteolytic milieu and interleukin-6 secretion. Fertil Steril 99, 871–881 (2013).\n\nwww.nature.com/scientificreports/\n12\nScientific  RepoRts  | 6:23343 | DOI: 10.1038/srep23343\n37. Guttman, M. & Rinn, J. L. Modular regulatory principles of large non-coding RNAs. Nature 482, 339–346 (2012).\n38. Wang, Y ., Li, Y ., Y ang, Z., Liu, K. & Wang, D. Genome-Wide Microarray Analysis of Long Non-Coding RNAs in Eutopic Secretory \nEndometrium with Endometriosis. Cell Physiol Biochem 37, 2231–2245 (2015).\n39. Qiu, J. J. et al. Overexpression of long non-coding RNA HOTAIR predicts poor patient prognosis and promotes tumor metastasis in \nepithelial ovarian cancer. Gynecol Oncol 134, 121–128 (2014).\n40. Liao, L. M. et al. Low expression of long noncoding XLOC_010588 indicates a poor prognosis and promotes proliferation through \nupregulation of c-Myc in cervical cancer. Gynecol Oncol 133, 616–623 (2014).\n41. Zhou, C. et al. The differential expression of mRNAs and long noncoding RNAs between ectopic and eutopic endometria provides \nnew insights into adenomyosis. Mol Biosyst. 12, 362–370 (2016).\n42. Kirstein, L. M. et al. Effects of anticoagulant, processing delay, and assay method (branched DNA versus reverse transcriptase PCR) \non measurement of human immunodeficiency virus type 1 RNA levels in plasma. J Clin Microbiol 37, 2428–2433 (1999).\n43. Duvigneau, J. C. et al. Heparin and EDTA as anticoagulant differentially affect cytokine mRNA level of cultured porcine blood cells. \nJ Immunol Methods 324, 38–47 (2007).\n44. Y ang, Y . et al. Plasma long non-coding RNA, CoroMarker, a novel biomarker for diagnosis of coronary artery disease. Clin Sci 129, \n675–685 (2015).\n45. Y an, Y . et al. Influence of a high-altitude hypoxic environment on human plasma microRNA profiles. Sci Rep 5, 15156 (2015).\n46. Du, M. et al. Circulating miR-497 and miR-663b in plasma are potential novel biomarkers for bladder cancer. Sci Rep 5, 10437 \n(2015).\n47. Wang, J. et al. Genome-wide circulating microRNA expression profiling indicates biomarkers for epilepsy. Sci Rep 5, 9522 (2015).\n48. Cho, S., Mutlu, L., Grechukhina, O. & Taylor, H. S. Circulating microRNAs as potential biomarkers for endometriosis. Fertil Steril \n103, 1252–1260 (2015).\n49. Liu, Z. et al. Microarray Profiling and Co-Expression Network Analysis of Circulating lncRNAs and mRNAs Associated with Major \nDepressive Disorder. Plos One 9, e93388 (2014).\n50. Zamore, P . D. & Haley, B. Ribo-gnome: the big world of small RNAs. Science 309, 1519–1524 (2005).\n51. Wang, W . T. et al. Differentially expressed microRNAs in the serum of cervical squamous cell carcinoma patients before and after \nsurgery. J Hematol Oncol 7, 6 (2014).\n52. Ariel, I. et al. Genomic imprinting and the endometrial cycle. The expression of the imprinted gene H19 in the human female \nreproductive organs. Diagn Mol Pathol 6, 17–25 (1997).\n53. Bustin, S. A. & Mueller, R. Real-time reverse transcription PCR (qRT-PCR) and its potential use in clinical diagnosis. Clin Sci 109, \n365 (2005).\n54. Lee, J. T. Epigenetic regulation by long noncoding RNAs. Science 338, 1435–1439 (2012).\n55. Sun, P . R., Jia, S. Z., Lin, H., Leng, J. H. & Lang, J. H. Genome-wide profiling of long noncoding ribonucleic acid expression patterns \nin ovarian endometriosis by microarray. Fertil Steril 101, 1038–1046 (2014).\n56. Ghazal, S. et al. H19 lncRNA alters stromal cell growth via IGF signaling in the endometrium of women with endometriosis. Embo \nMol Med 7, 996–1003 (2015).\n57. Lee, G. H. et al. Association of CDKN2B-AS and WNT4 genetic polymorphisms in Korean patients with endometriosis. Fertil Steril \n102, 1393–1397 (2014).\nAcknowledgements\nWe thank our colleagues at Sun Y at-sen Memorial Hospital for their valuables suggestions. We also thank all \nparticipants for agreeing to take part in our study. This work was supported by funds from the National Science \nFoundation of China (30500578) and Guangdong province (No. 2014T70833).\nAuthor Contributions\nW .-T.W ., Y .-M.S. carried out the experiments and data analysis and wrote the manuscript. W .H. participated in \nthe experiments, and B.H., Y .-N.Z. participated in the data analysis and sample collection. Y .-Q.C. designed and \nwrote the manuscript.\nAdditional Information\nSupplementary information accompanies this paper at http://www.nature.com/srep\nCompeting financial interests: The authors declare no competing financial interests.\nHow to cite this article: Wang, W .-T. et al. Genome-wide Long Non-coding RNA Analysis Identified \nCirculating LncRNAs as Novel Non-invasive Diagnostic Biomarkers for Gynecological Disease. Sci. Rep. 6, \n23343; doi: 10.1038/srep23343 (2016).\nThis work is licensed under a Creative Commons Attribution 4.0 International License. The images \nor other third party material in this article are included in the article’s Creative Commons license, \nunless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, \nusers will need to obtain permission from the license holder to reproduce the material. To view a copy of this \nlicense, visit http://creativecommons.org/licenses/by/4.0/","source_license":"CC0","license_restricted":false}