*3.2. Transcriptional Evidence for lncRNA Involvement in Endometriosis*

To date, there have been around 14 genome-wide profiling transcriptome studies of human patient samples and animal models for endometriosis (Table 1). These studies are mainly descriptive in nature, with experimental validation limited to a few selected candidates. These studies often include in silico analysis examining the relationship between differentially expressed protein coding and lncRNA transcripts. Further bioinformatic approaches, such as gene ontology analysis, are then often used to predict the relevance of the identified targets to biological pathways involved in the pathogenesis of endometriosis. Initial studies were mostly conducted using microarray platforms to assess the expression of lncRNAs and mRNAs. Using a human lncRNA expression microarray, Sun et al. [33] were the first to assess genome-wide the relationship between lncRNA and mRNA expression in ovarian endometriosis lesions compared to paired autologous eutopic endometrial samples. They identified 948 lncRNAs and 4088 mRNAs as differentially expressed in the study cohort and validated the differential expression of the top 10 lncRNA candidates using qRT-PCR. Based on co-expression with mRNAs, lncRNAs in this study were predicted to take part in tissue adhesion, angiogenesis, estrogen production, and immune response, all processes known to be associated with the pathogenesis of endometriosis. Co-expression and genomic proximity were then used to predict 49 *cis*-regulating lncRNAs and their protein coding targets, while *trans*-regulating lncRNAs were predicted by co-expression with transcription factors and network analysis. This indicated that the top candidates were in a network with MYC, CTCF, and E2F4. Another study from Cai et al. (2019) used microarrays to profile lncRNA and mRNA expression in an endometriosis rat model, resulting in the identification of 115 upregulated and 51 downregulated lncRNAs together with 182 differentially expressed protein coding mRNA transcripts [34]. Co-expression analysis revealed five lncRNAs (*LOC102551276*, *NONRTT006252*, *LOC103691820*, *LOC102546604,* and *NON-RATT003997*) that show a similar expression pattern to four protein coding mRNAs (*Adamts7*, *P2ry6*, *Dlx3,* and *TP53*), indicating a possible functional relationship in endometriosis.

Using a transcriptome array, Wang et al. [35] investigated the expression profile of lncR-NAs in serum and tissue samples from patients with and without endometriosis. Following qRT-PCR validation of differentially expressed lncRNAs, they identified a combination of five circulating lncRNAs (*NR\_038395, NR\_038452, ENST00000482343, ENST00000544649,* and *ENST00000393610*) that they proposed could act as non-invasive biomarkers for the disease.

In the last few years, high-throughput RNA sequencing technology has surpassed microarrays as the preferred genome-wide technology for the identification of differentially expressed lncRNAs in endometriosis. Unlike microarrays, RNA sequencing is not biased by prior knowledge, enabling the discovery and characterization of lncRNAs that may play a role in the disease. A number of RNA sequencing studies have now been conducted comparing tissues from either patients or endometriosis animal models (Table 1). For example, using RNA sequencing, Cui et al. identified 86 differentially expressed lncRNAs and 1228 differentially expressed mRNAs in patients with ovarian endometriosis lesions compared to eutopic endometrial controls [36]. Pathway and gene ontology analysis showed that differentially expressed lncRNAs were involved in the regulation of cell proliferation, adhesion, migration, steroidogenesis, and angiogenesis, all processes implicated in endometriosis lesion formation and survival at ectopic sites of implantation.

**Table 1.** Genome-wide studies that identified differentially expressed lncRNAs in endometriosis.



### **Table 1.** *Cont.*


**Table 1.** *Cont.*

EM, endometriosis; HEECs, human endometrial endothelial cells; LMD, laser microdissection; DIE, deep infiltrating endometriosis. 1, small sample size (less than 30 per group); 2, no comprehensive clinical information (i.e., American Fertility Society (rAFS) disease stage, lesion entities, menstrual cycle phase); 3, no validation in an adequate independent cohort; 4, EM-free controls are not appropriate (i.e., CIN patients, no laparoscopic proof); 5, combining heterogeneous datasets (i.e., different lesion entities, cycle phases, stages, cell types); 6, not all relevant tissues analyzed (i.e., eutopic tissues of EM-free controls, eutopic and ectopic tissues of EM patients).

> In spite of the advantages listed above, RNA sequencing studies have some limitations. Due to their high cost, they are often limited to a small number of samples and lack extensive validation. However, they can form the basis for further studies that perform functional validation of candidate genes and investigate their value as diagnostic and prognostic markers in endometriosis. These validation studies usually have cohorts with a larger sample size and may include functional experiments that attempt to uncover the

molecular mechanism of lncRNA action (Table S1). For example, AFAP1-AS1 was identified as one of the most differentially expressed lncRNAs in the microarray study of Sun et al. [33]. In a new study, Lin et al. [47] could confirm and extend these findings in a cohort of *n* = 36 patients. They show that *AFAP1-AS1* is overexpressed in ectopic endometrium of women with endometriosis (*n* = 18), compared to paired eutopic endometriosis tissue (*n* = 18) and normal endometrium of women without the disease (*n* = 10). In vitro gene targeting assays in primary human endometriotic stroma cells (ESCs) indicated that this lncRNA regulates epithelial–mesenchymal transition (EMT) in endometriosis by regulating transcription of the EMT-related transcription factor ZEB1. Furthermore, experiments in a xenograft mouse model indicated that AFAP1-AS1 was required for the growth of ectopic tissue. In this case, a shRNA knockdown of AFAP1-AS1 in the Ishikawa endometrial cancer cell line led to reduction of the subcutaneous tumor size, compared to animals injected with a non-targeting shRNA-transfected cell line [47].

Liu et al. analyzed the value of lncRNA H19 expression as an endometriosis biomarker [48]. They found that H19 expression in the both the ectopic and eutopic endometrium of endometriosis patients was significantly higher than in the normal endometrium. Overexpression of H19 lncRNA in endometriosis lesions was associated with infertility, recurrence of disease, bilateral ovarian lesions, an increased CA125 level and with progression in the revised American Fertility Society (rAFS) disease stage. Further multivariate logistic regression analysis showed that H19 overexpression in endometriosis lesions is a prognostic factor for endometriosis recurrence.

In summary, recently there has been an effort by a number of studies using genomewide high-throughput technologies to identify differentially expressed lncRNAs in endometriosis not only for their potential clinical application as diagnostic or prognostic biomarkers of the disease, but also in order to gain insights into the pathogenesis of the disease (Table 1). The lncRNAs where further work has been done to validate a role in endometriosis are listed in Table S1 and summarized in Figure 2.

**Figure 2.** Differentially expressed (DE) lncRNAs in endometriosis based on validation studies: (**a**) Up- and downregulated lncRNAs in endometriosis tissues and/or cells (**b**) in body fluids or lesion microenvironment.

32
