Liquid Biopsy in Endometriosis: A Systematic Review

Despite laparoscopy being a standardized option to diagnose pelvic endometriotic implants, non-invasive biomarkers are necessary to avoid the discomfort of invasive procedures. Recent evidence suggests a potential role of microRNAs (miRNAs) as feasible biomarkers for the early diagnosis of endometriosis. Following the recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, we systematically searched PubMed, EMBASE, Scopus, Cochrane Library, and Science Direct in January 2023. We provided no restriction on the country and year of publication, and considered English published articles. We selected studies including patients with endometriosis and describing miRNA regulation in the context of endometriosis. Overall, 45 studies fulfilled the inclusion criteria, and 2045 patients with endometriosis and 1587 controls were screened. Patients were analyzed concerning miRNAs expression and sources, stage of disease, and symptoms, and compared to controls. Among DEMs, the ones with the widest delta between endometriosis patients and controls—Relative Expression ≥ 4 Log2(ratio)—were miR-145, miR-191, miR-195, miR-21-5p, miR-106b-5p, miR-195-5p, miR-451a, miR-200c, miR-20a-5p, and miR-15a-5p. Although the epigenetic regulation is partially unclear, miRNAs are valid biomarkers to diagnose endometriotic lesions in symptomatic and non-symptomatic women. MiRNAs modulation should be clarified, especially during therapies or relapse, to plan targeted management protocols.


Introduction
Endometriosis diagnosis in childbearing-age women is often delayed due to the lack of pathognomonic signs and symptoms [1][2][3]. Nowadays, transvaginal ultrasound (TVS) is the most cost-effective method to detect endometriotic lesions, but the gold-standard methodology for diagnosis is laparoscopy [4]. Laparoscopy is also considered the goldstandard treatment for endometriosis [2], while the best treatment option should consider the age of the patient, the symptoms, the desire to conceive, and previous surgeries [4]. In older women who underwent previous surgeries, medical treatment and In Vitro Fertilization may be discussed [4]. That emphasizes the necessity of non-invasive biomarkers to avoid the discomfort of laparoscopic procedures and simplify the diagnosis. Recent studies suggest the possibility of using microRNAs (miRNAs)-as reliable markers from different compartments (serum, plasma, endometrial biopsies, etc.)-in the early diagnosis of endometriosis [5][6][7]. Minimally invasive methods are necessary to assess the influence of bio-behavioral disruptors on the prognosis, treatment response, and recurrence.
The miRNAs are a class of small RNA molecules, composed of 15-22 nucleotides each, post-transcriptionally regulating genes [8]. MiRNAs hybridize into complementary mRNAs, which are involved in different cellular features, such as implantation, embryo developmental processes, tumor suppression, apoptosis, proliferation, angiogenesis, and metastasization [9][10][11]. Despite endometriosis showing a benign histological connotation, it develops and disseminates as a neoplastic process [12,13]. Moreover, there are pieces of evidence that endometriotic implants could transform into cancerous lesions [14]. In that context, endometriosis pathogenesis contains genetic, angiogenic, metabolic, and immunological alterations [12]. Endometriotic implants could undergo malignant transformation via altered molecular pathways, showing characteristics of atypia, invasivity, and diffusion, especially in E-cadherin-negative endometriotic cells [12]. Moreover, in 30% of cases, an endometriosis diagnosis may be linked to ovarian cancer detection [12]. Indeed, there is evidence that miRNAs are also involved in the pathogenesis of ovarian cancer, even in presentations linked to endometriosis [15][16][17]. One of the advantages of miRNAs is that they are accessible to sample. They could be found in multiple cellular compartments-in the context of different human systems-and be up-or downregulated [18][19][20]. It is estimated that miRNAs could be feasible biomarkers in the early diagnosis and management of endometriosis progression [20]. For example, endometriotic lesion development may depend on lower-expression cell adhesion and cytoskeleton molecules and decreased proteolysis [21][22][23]. In those contexts, the proliferation, migration, and stemness of endometrial stromal cells (ESCs) are increased in endometriosis through miRNAs' epigenetic transcription [24]. Most miRNAs can be found in the serum of endometriosis-affected patients, and they could be extracted through liquid biopsy [15,21]. Those data may pave the way for new strategies for early diagnosis of endometriotic implants. The present systematic review aimed to evaluate the distribution and regulation of the differently expressed miRNAs (DEMs) in the context of endometriosis.

Materials and Methods
The methods for this study were specified a priori based on the recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [25]. The present work is categorized on PROSPERO as: ID400389.

Search Method
We performed a systematic search for records about the expression of different miR-NAs in endometriosis-affected patients in PubMed, EMBASE, Scopus, Cochrane Library, and Science Direct in January 2023. We made no restriction on the country or year of publication, and considered only studies published entirely in English. We adopted the following string of keywords in each database to identify studies that fit to the topic of our review: "(Cell-Derived Microparticles OR MicroRNAs) AND Endometriosis".

Study Selection
The study selection was made independently by I.I. and P.F. In the case of a discrepancy, C.R. decided on inclusion or exclusion. Inclusion criteria were (1) studies including patients with endometriosis; (2) studies describing differently expressed miRNAs (DEMs) and their regulation in the context of endometriosis signs and symptoms; and (3) peer-reviewed articles, published originally. We excluded non-original studies, pre-clinical trials, animal trials, abstract-only publications, and articles in a language other than English. If possible, we tried to contact the authors of studies that were published as conference abstracts via e-mail and asked them to provide their data. We assessed all included studies concerning potential conflicts of interest.

Extraction and Quantification of miRNAs
Liquid biopsy is a minimally invasive procedure to extract microvesicles from serum [12]. Extraction, amplification, and quantization of miRNAs are based on different procedures.
Other authors described miRNA isolation in endometrial stromal cells from biopsies of ectopic endometrial lesions or eutopic endometria, which were placed into two halves in buffered formaline for storage and molecular examination. The RNA quality was first evaluated according to the integrity of the strains in the samples, whereas further analysis was performed based on the RNA minimum degradation in each sample [65].
Among the DEMs isolated, only those with AUC (Area Under the Curve) > 0.6 and significant allele and genotype distribution frequencies (p < 0.05) were considered in the present study.

Studies' Characteristics
We mention the studies selected and all reasons for exclusion in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart ( Figure 1). After the database search, 124 articles matched the search criteria. After removing records without full text, duplicates, and wrong study designs (e.g., reviews), 64 were eligible. Overall, 45 matched the inclusion criteria and were included in the systematic review. The countries where the studies were conducted, the year range, the studies' design, and the number of participants are summarized in Table 1. Overall, the publication years ranged from 2013 to 2022. In total, 2045 patients with endometriosis and 1587 controls were analyzed.

Outcomes
A total of 2045 patients were included in the review. Regarding miRNA sources, miRNAs in endometrial stromal cells (ESCs) were extracted from the biopsies of ectopic endometrial lesions and/or eutopic endometria. Otherwise, miRNAs were extracted from serum, plasma, follicular fluid, and cumulus cells. Those data are summarized in Tables 2 and 3.

Early Diagnosis
In total, 58 miRNAs were upregulated in endometriosis-affected patients, whereas, 67 miRNAs were downregulated. Those data are summarized in Tables 2 and 3. Except for 8, the other 37 studies revealed the stage of disease. In total, 18 records involved patients with ASRM (American Society for Reproductive Medicine) stages III-IV of disease; 16 records involved patients with ASRM stages I-IV of disease; whereas only 3 studies included patients with low or intermediate stages of disease (I-III). In particular, Wang et al. enrolled 30 patients with stages I-II of disease and revealed upregulation of miR-20a-5p through liquid biopsy [71]. In parallel, liquid biopsy showed downregulation of miR-30c-5p, miR127-3p, miR-99b-5p, and miR-15b-5p in the same cohort [71]. Liu et al. isolated miRNAs from ESCs, both in eutopic and ectopic endometria, demonstrating that miR-449b-3p was downregulated in the early stages of endometriosis-affected women [74]. Petracco et al. enrolled patients with stages II-III of endometriosis [80]. MiRNA was isolated both from eutopic and ectopic endometrial samples, and miR-135a/b was downregulated [80]. In the last two studies, the difference in the relative expression of miRNAs between patients and controls was <2 Log2(ratio) [74,80].

Discussion
From a functional perspective, miRNAs are involved in intercellular crosstalk, both in eutopic endometrial tissue and endometriotic implants [70]. Scientific literature highlighted the potential role of DEMs as biomarkers for endometriosis-affected women. The expression and modulation of miRNAs are wide and heterogeneous, and we considered in our study only DEMs with the highest AUC (>0.6) and significant allele and genotype distribution frequencies (p < 0.05). Hypothetically, miRNAs may indirectly represent the cellular microenvironment that leads to the formation of endometriotic implants. Therefore, their research could help intercept endometriosis before macroscopic lesions are identifiable on an ultrasound. Although it is extremely difficult to determine the most sensitive and specific miRNAs in endometriosis pathogenesis, we have underlined miRNAs expression in symptomatic patients. For example, specific miRNAs are overexpressed in dysmenorrheic or infertile women. The presence of the symptom can help us in a twofold way. It can help us identify patients for further investigation by liquid biopsy. It can also give us information about how patients evolve to this symptomatology, helping us to understand the molecular mechanisms underlying the development of the symptomatology. This consideration is also interesting from the perspective of infertility symptoms without organic pelvic distorting lesions. In these cases, infertility is likely related to the uterine microenvironment corrupted by a chronic inflammatory state. The study of miRNAs in these patients can identify conditions invisible to the eye. Farsimadan et al. isolated miRNAs in infertile endometriotic patients without any declared symptoms, and they highlighted the upregulation of miR-146a rs2910164 and miR-149 rs2292832, which showed an intermediate delta of relative expression between patients and controls, e.g., ≥2 > 4 Log2(ratio) [20]. Further, Li et al. isolated miR-451 as upregulated miRNA in infertile endometriotic patients, but they did not declare whether those patients were suffering from other symptoms, such as dysmenorrhea [78]. However, their analysis revealed an intermediate difference in miR-451 relative expression between patients and controls [78]. On the other hand, there is little evidence about miRNA expression in endometriotic patients suffering from dysmenorrhea, but without infertility-related problems. Bendifallah et al. showed upregulation of hsa-miR-29b-1-5p, hsa-miR-4748, hsa-miR-515-5p, hsa-miR-548j-5p and hsa-miR-6502-5p, whereas Dabi et al. revealed upregulated miR-124-3p, even though neither of the studies specified the incidence of infertility in their cohorts [99,100]. Moreover, neither of the studies disclosed the difference in the relative expression of miRNAs between patients and controls [99,100]. Farsimadan et al. isolated miRNAs in infertile endometriotic patients without any declared symptoms, and they highlighted the upregulation of miR-146a rs2910164 and miR-149 rs2292832, which showed an intermediate delta of relative expression between patients and controls, e.g., ≥2 > 4 Log2(ratio) [20]. Further, Li et al. isolated miR-451 as an upregulated miRNA in infertile endometriotic patients, but they did not declare whether those patients were suffering from other symptoms, such as dysmenorrhea [78]. However, their analysis revealed an intermediate difference in miR-451 relative expression between patients and controls [78]. In our opinion, given the inaccurate definition of the signs and symptoms in different studies-mainly due to the heterogeneity of endometriosis presentation-it would be appropriate to focus on DEMs with the widest range in relative expression between patients and controls to avoid high false-positive rates during miRNA isolation. The real clinical use of miRNAs should lie in implementing the diagnostic capabilities of early forms. With this in mind, the different ways in which miRNA assays can be obtained should be emphasized. The site of expression probably influences miRNA modulation. Liquid and incisional endometrial biopsy may be valid options for miRNA extraction, even if eutopic endometrial tissue seems to have more defined expression profiles [68,83]. Surely, liquid biopsy through patients' serum or plasma would be more feasible and cost-effective as a screening method in women suffering from dysmenorrhea, infertility, and dyschezia. An endometrial tissue biopsy may be a valid option in diagnosing endometriotic implants in patients with a suspected transvaginal ultrasound. Moreover, serum and plasma miRNAs are supposed to have a distant effect involving systemic organs. That suggests a potential role of liquid biopsy in detecting endometriotic implants of unknown location [20,106]. Although isolation strategies from saliva also revealed positive biomarkers for endometriosis detection, liquid biopsy could find a place in the application of screening programs on the fertile population, even in the complete absence of symptoms [106,107]. Unfortunately, this perspective needs more investigation of miRNA fluctuations in the early stages. For those reasons, one of the best candidates for early diagnosis of endometriosis may be miR-145 [21]. In their study, Bashti et al. showed an important advantage of miR-145-neither of the patients enrolled suffered from infertility or dysmenorrhea [21]. In addition, miR-145 expression is mostly upregulated in stages I-II of disease, suggesting its crucial role in tracking recent endometriotic lesions [21]. Finally, a hypothetical function of miRNAs could be related to the follow-up of patients on medical therapy. Any fluctuations could directly represent the response of endometriosis tissue to medical therapy, optimizing the chronification of treatment. Our review investigated all present literature on the topic. This represents its strength and weakness, related to the enormous heterogeneity of the data in the literature. The hope is that it will represent the basis for further investigation of the most interesting miRNAs useful for the clinical management of patients with endometriosis, with targeted studies specifically designed to investigate the various aspects.

Conclusions
Most recent evidence shows that intercellular crosstalk has a critical role in endometriosis pathogenesis, although there is heterogeneity of the data. In that context, specific molecular signatures could mark the homeostasis of endometrial tissue. Focusing on endometrial tissue, a pattern of miRNAs may be useful for the early diagnosis and management of endometriosis-affected women, mainly through liquid biopsy. Although there is a lack of data regarding DEMs in response to different therapeutic regimens, that analysis could be performed in women, administered with laparoscopy, oral contraceptives, Non-Steroidal Anti-Inflammatory Drugs (NSAIDs), or a GnRH antagonist during the FU period. To date, although the mechanism of epigenetic regulation remains unclear, the assessment of miRNAs' expression could be a promising and cost-effective tool to detect the presence of endometriotic implants in symptomatic and non-symptomatic patients. Further studies are needed to clarify miRNAs' modulation during treatment or the recurrence of disease in order to predict disease development and plan targeted management options.

Conflicts of Interest:
The authors declare no conflict of interest.