Overview of miRNAs for the non-invasive diagnosis of endometriosis: evidence, challenges and strategies. A systematic review

ABSTRACT Objective The aim of the study was to assess the evidence on miRNAs as biomarkers for the diagnosis of endometriosis, as well as to provide insights into the challenges and strategies associated with the use of these molecules as accessible tools in clinical practice. Methods Systematic review conducted on PubMed®, Latin American and Caribbean Health Sciences Literature (LILACS), MEDLINE® and Web of Science databases using the search terms endometriosis (all fields) AND miRNA (all fields), evaluating all publication up to May 2019. Results Most miRNAs found to be dysregulated in this study were harvested from tissue samples, which precludes their use as a non-invasive diagnostic test. However, differential expression of 62 miRNAs was reported in samples that may be used for non-invasive diagnosis of endometriosis, such as blood, serum and plasma. Conclusion Despite the identification of several candidates, studies are investigatory in nature and have been conducted with small number of samples. Also, no particular miRNA has been validated for diagnostic purposes so far. Studies based primarily on biological samples and applicable to translational research are warranted. Large databases comprising information on sample type and the use of saliva and vaginal fluid for miRNAs identification may prove essential to overcome current barriers to diagnosis of endometriosis.


❚ INTRODUCTION
Endometriosis is a common disease that affects up to 10% of women of reproductive age (1,2) and is characterized by the presence of endometrial cells outside the uterine cavity. The disease has been the focus of many studies, however, the diagnosis is still very difficult. Clinical presentation varies widely, ranging from asymptomatic to severe, and no diagnostic biomarkers have been approved for routine clinical diagnosis of endometriosis to date. (1,3) Diagnostic imaging tests such as pelvic ultrasonography and magnetic resonance have been used, especially in deep endometriosis. However, examiner expertise has a strong impact on imaging findings, (4)(5)(6)(7) which ultimately makes the diagnosis difficult. In cases with no positive imaging findings, a final diagnosis of superficial endometriosis can only be made through histological analysis of the lesion, usually in samples obtained by laparoscopic surgery. (8,9) However, this procedure is invasive and requires general anesthesia.
The complexity of the disease, combined with the lack of precise and less invasive diagnostic methods, contributes to delayed diagnosis, which can take up to 11 years. (5,10,11) Therefore, there is great demand for accurate and less invasive diagnostic tests for endometriosis. (12)(13)(14)(15)(16) Different research groups have investigated the role of miRNAs (microRNAs or miR) in the regulation of known genes, given their association with processes involved in disease pathogenesis and progression. miRNAs are a class of small endogenous, non-coding RNA molecules involved in post-transcriptional regulation of gene expression. (17) These small molecules have also been found in peripheral blood and may therefore be potential diagnostic biomarkers for endometriosis. (18,19) This literature search was conducted to determine how close miRNAs are to being used as biomarkers for endometriosis. Findings of this review are expected to guide the next steps towards overcoming challenges associated with the use of miRNAs in clinical practice.

❚ OBJECTIVE
To determine which miRNAs are applicable to the diagnosis of endometriosis and to outline the challenges and strategies involved in the use of these molecules as accessible diagnostic tools in clinical settings.

❚ METHODS
To identify research articles addressing associations between endometriosis and miRNA, a search was conducted in PubMed ® , Latin American and Caribbean Health Sciences Literature (LILACS), MEDLINE ® and Web of Science databases using the search terms endometriosis (all fields) AND miRNA (all fields).
All publications listed up to May 2019 (automatically selected) were manually curated, and only those involving miRNA expression patterns, validated by polymerase chain reaction (PCR) in clinical samples of endometriosis, were discussed in this review. Articles published in languages other than English or based on cell culture, retracted articles and articles published in conference proceedings or inaccessible were excluded. Reports listed in more than one database were included only once in the pool of publications.
This study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for systematic reviews adopted by Hospital Israelita Albert Einstein (HIAE), located in São Paulo (SP), Brazil. Data were extracted in duplicate and independently by two different investigators, then compared for confirmation. miRNAs and their respective expression levels in different types of samples and patient populations were examined. Studies were also analyzed according to year and country of publication.

Overview of publications on miRNA and endometriosis
A total of 449 research articles addressing associations between endometriosis and miRNA were found in databases selected for this review. Most (185) were retrieved from PubMed ® , followed by LILACS and MEDLINE ® (158) and Web of Science (106). Out of einstein (São Paulo). 2021;19:1-12 this publication pool, 46 matched final selection criteria and were selected for further discussion in this review ( Figure 1). SNP: single-nucleotide polymorphisms; LncRNAs: long non-coding ribonucleic acids; circRNAs: circular ribonucleic acids; PCR: polymerase chain reaction. The number of publications investigating dysregulated miRNAs in women with endometriosis increased sharply since 2009, with approximately half of selected articles (23) published in the last 3 years. China and the United States were the countries with the largest number of publications (21 and 9 articles, respectively).
A total of 33 miRNAs were examined in more than one study. Of these, 13 miRNAs were analyzed in the same types of samples. miRNAs identified in more than one study and body fluid are described in table 2.
were also found to be dysregulated in other types of tissue, such as EC and eutopic endometrium, and in the peritoneal fluid. Of these, 35% were detected in the same type of tissue in more than one study, including

❚ DISCUSSION
Endometriosis can be a debilitating disease and may lead to poor quality of life. (65) The disease is associated with dysmenorrhea, deep dyspareunia, chronic pelvic pain and infertility (66,67) and is considered a public health concern, given the impact on patient physical and psychological health, and the socioeconomic impact of diagnosis, treatment and clinical control costs. (68) The final diagnosis of endometriosis is currently based on histological analysis of the lesion, usually in samples obtained by laparoscopic surgery. (69) However, imaging modalities are important non-invasive diagnostic alternatives for ovarian and deep endometriosis. Both surgical and non-surgical approaches require considerable professional skill and availability of specific data, which may represent a huge economic and health burden in developing countries. (4)(5)(6)(7)(8)(9) In the last three decades, researchers worldwide have tried to identify a non-invasive test that could shorten the turnaround time for diagnosis of endometriosis. CA-125 can be detected in blood or peritoneal fluid  and is one of the best studied biomarkers. In some case studies, measurement of CA-125 levels was deemed promising, especially for diagnosis of more invasive endometriosis, provided measurements are made in the beginning of the menstrual cycle. (70)(71)(72) In spite of conflicting results regarding the value of CA-125 as a final and important biomarker reported in recent reviews, according to Socolov et al., CA-125 is still the most recommended biomarker for endometriosis diagnosis and monitoring. (73) In a more recent Cochrane review published in 2016, Nisenblat et al. compared the accuracy of any combination of noninvasive diagnostic tests to surgical diagnosis of pelvic endometriosis, using randomized controlled trials or cross-sectional studies published until early 2015 as a reference standard. Authors concluded that none of the biomarkers investigated (including CA-125) could be duly evaluated due to insufficient or poor-quality evidence, given the high heterogeneity and risk of bias in selected studies. (15) CA-125 is most elevated in advanced stages of endometriosis. Therefore, the sensitivity of this marker is limited. Its specificity is also thought to be poor, since it is upregulated in other gynecological conditions. (74) In this context, the search for novel and effective noninvasive biomarkers capable of improving endometriosis diagnosis, management and monitoring remains high on the priority list.
Circulating miRNAs, first identified as noninvasive serological markers of tumors in 2008, (75)(76)(77) are promising alternative candidates. The high stability of circulating miRNAs in human plasma and their resistance to multiple sample handling procedures has been emphasized in these pioneer studies.
These same studies also established the concept of disease diagnosis based on specific cell-free miRNA signatures. Since then, miRNAs have been validated as noninvasive diagnostic markers for several diseases, including oncologic, inflammatory, cardiovascular, metabolic and reproductive disorders. miRNAs proved to be ideal diagnostic markers in oncology, as shown by differential circulating miRNA expression patterns in lung, ovarian, colorectal, prostate and breast cancer patients relative to healthy controls. (78) In the female reproductive system in particular, dysregulated miRNA expression has been studied in uterine leiomyomata, in several gynecologic cancers (including adenocarcinomas), and in pregnancy disorders, such as preeclampsia and preterm birth. (79)(80)(81)(82)(83) These small noncoding molecules associated with several diseases have been proposed as useful diagnostic candidates for endometriosis. (84) In this review, miR-145 was the miRNA found to be differentially expressed in the largest number of studies (six articles). In the 46 studies analyzed, most miRNAs found to be dysregulated in endometriosis were harvested from tissue samples. Bodily fluids were seldom investigated, even though they may be used as non-or minimally invasive diagnostic tools. Also, most studies compared miRNA expression differences between the eutopic and EC of patients with endometriosis and only a few compared the endometrium of patients with endometriosis, suggesting that examinations based endometrial biopsies are difficult.
As regards dysregulated miRNAs in endometriosis patients compared in this review, 30 were found in the blood, 27 in the serum and 18 in the plasma of women with endometriosis relative to control populations. Differences in the molecular composition of serum and plasma have been well-documented. (85,86) When comparing the miRNA spectrum between serum and plasma, Wang et al., (19) detected several differences in RNA levels driven by the release of certain miRNAs and other RNAs during the coagulation process, and suggested that use of plasma as the sample of choice for studying circulating miRNAs, since RNAs released during coagulation may alter the true repertoire of circulating miRNAs.
Differential expression of six miRNAs was detected in the peritoneal fluid of endometriosis patients relative to non-affected women. Hence, some miRNAs found in peritoneal fluid may play a role in the pathogenesis of endometriosis. However, given the nature of this fluid, its use is limited by the need for surgical (i.e., invasive) collection.
Some points are worthy of note and should be emphasized in these studies: conflicting results. They have been reported in studies investigating miR-145, -424, -199a, -29c, -126, -16, -195 and -18a expression in the same type of sample. Major characteristics of these studies are described below.
Upregulation of miR-145 was found in the serum, in a study with 24 stages III and IV endometriosis and 24 control patients, (51) and in plasma, in a study with 55 stages I and II endometriosis and 23 control patients. (58) In contrast, the same miRNA was found to be downregulated in the serum in a study including 60 cases and 25 controls, (19) in which disease stage was not reported.
miR-424 was downregulated in blood in a study with four patients with mild endometriosis and einstein (São Paulo). 2021;19:1-12 three controls. (62) However, it was also found to be upregulated in the serum of 30 patients with minimalmild endometriosis relative to 20 control individuals. (52) miR-199a was upregulated in the serum of patients with endometriosis in two studies, one with 60 stages III and IV endometriosis and 25 control patients, (19) and another with 45 endometriosis and 35 control patients. (50) However, the same miRNA was found to be downregulated in the serum in a different study with 40 endometriosis and 25 control patients. (54) A study with 15 clinical cases and 11 controls revealed miR-29c upregulation in the EC of women with endometriosis relative to the eutopic endometrium in the Control Group. (39) This finding was further confirmed in a study including 51 women with endometriosis and 32 control women, (25) in the proliferative and secretory phases of the menstrual cycle. However, conflicting results suggesting miR-29c downregulation in the EC of 20 women with endometriosis relative to the eutopic endometrium of ten control patients, (27) all of them in the proliferative phase of the cycle, have been reported by a different researcher.
miR-126 was found to be upregulated in the ectopic compared to the eutopic endometrium of eight women with stages III to IV endometriosis (45) in the proliferative and secretory phases of the menstrual cycle. However, miR-126 downregulation was reported in the ectopic compared to the eutopic endometrium in 31 women with stages III to IV endometriosis, (20) all of them in the secretory phase of the menstrual cycle. miR-16 and miR-195 were found to be upregulated in plasma of 33 women with endometriosis relative to 20 control patients. (60) However, another study identified both downregulated in the blood of four patients with mild endometriosis relative to three controls. (62) miR-18a was upregulated in serum of 24 women with stage III and IV endometriosis compared to 24 control patients. (51) However, it was found to be downregulated in the blood of four patients with mild endometriosis compared to three controls. (62) Conflicting results emphasize the relevance of criteria such as menstrual cycle phase, disease stage, type of sample and type of test procedure, and the need for studies with larger sample size to develop novel diagnostic tests for endometriosis.
The second objective of this review was to provide new directions for future studies aimed to identify a miRNA which may be used as a reliable biomarker and an accurate diagnostic tool for endometriosis. Sadly, according to this critical literature review no particular miRNA or miRNA combination has been validated for improved diagnosis of endometriosis to date. This may reflect the heterogeneity of the disease and resultant differences in tissue composition. (87) Thus, we support the World Endometriosis Research Foundation (WERF) and Endometriosis Phenome and Biobanking Harmonization Project (EPHect) initiatives. Endometriosis research teams worldwide must join forces in order to develop large databases comprising data derived from samples obtained from patients with well-characterized endometriosis. This is an important tool for identification and validation of biomarkers and may play a key role in biomarker investigation in future endometriosis studies. (88) The inclusion of a large global pool of clinical samples collected from endometriosis patients is vital for the advancement of medical knowledge, and could be a key factor in the implementation of targeted therapies, which may enhance treatment effectiveness and improve the quality of life of endometriosis patients.
No studies investigating miRNA expression profile in the vaginal fluid were found in this literature review. This body fluid can be easily collected during gynecological examinations and, in spite of high rates of bacterial colonization, appears to be a promising source of diagnostic material. (89,90) The value of differential miRNA expression in vaginal fluid as potential screening test for HPV has been examined, with interesting results. (91)(92)(93)(94)(95) Likewise, none of the papers examined investigated miRNAs in saliva. To date, there are no scientifically proven salivary biomarkers for endometriosis. Saliva is a suitable and desirable medium for biomarker detection (96,97) and its applicability to the diagnosis of endometriosis has been explored previously. (98,99) Saliva is widely available and can be easily collected in a non-invasive manner, at low cost and with minimal discomfort. Therefore, it is an ideal fluid for biomarker investigation and is attracting great interest in the public health sector. The use of saliva for miRNA identification could be a potential non-invasive solution to overcome current barriers to the diagnosis of endometriosis.
This study has some limitations. When evaluating papers with contrasting results, it was not possible to tease out the factors underlying such different outcomes. Reasons explaining miRNA heterogeneity were also not found.
In this review, different studies investigating miRNA expression in endometriosis patients were einstein (São Paulo). 2021;19:1-12 discussed. Most of these studies were based on pooled or small samples. Large, well-designed clinical trials aimed to validate endometriosis-related miRNAs are needed in order to develop accurate, low-invasiveness diagnostic methods for endometriosis. The clinical impact of scientifically proven miRNA biomarkers for endometriosis will translate into better access to care and less health disparities, with potential impacts on global health. The diagnosis of endometriosis at earlier stages of the disease may lead to dramatic reduction in health costs and provide significant benefits for patients through improved health and quality of life.

❚ CONCLUSION
Differential miR-145 expression was reported in the largest number of studies (six articles). Most dysregulated miRNAs were harvested from tissue samples.
No particular miRNA or miRNA combination has been validated for improved diagnosis of endometriosis to date. This may have reflected the heterogeneity of the disease and resultant differences in tissue composition. Large databases comprising data derived from samples collected from patients with well-characterized endometriosis may play a key role in biomarker investigation in future studies. The use of saliva and vaginal fluid samples for miRNA identification could be a potential non-invasive solution to overcome current barriers to the diagnosis of endometriosis.

❚ ACKNOWLEDGEMENTS
To Carlos Augusto Cardim de Oliveira for his instructions about the best practices for a good systematic review.