Transcriptomic analysis of cumulus cells shows altered pathways in patients with minimal and mild endometriosis

Endometriosis is a chronic inflammatory disorder that is highly associated with infertility. This association seems to be related to oocyte impairment, mainly in the initial stages of endometriosis (minimal and mild), where no distortions or adhesions are present. Nonetheless, invasive oocyte analyses are not routinely feasible; thus, indirect assessment of oocyte quality is highly desirable, and, in this context, cumulus cells (CCs) may be more suitable targets of analysis. CCs are crucial in oocyte development and could be used as an index of oocyte quality. Therefore, this prospective case–control study aimed to shed light on the infertility mechanisms of endometriosis I/II by analyzing the CCs’ mRNA transcription profile (women with endometriosis I/II, n = 9) compared to controls (women with tubal abnormalities or male factor, n = 9). The transcriptomic analyses of CCs from patients with minimal and mild endometriosis revealed 26 differentially expressed genes compared to the controls. The enrichment analysis evidenced some altered molecular processes: Cytokine-cytokine receptor interactions, Chemokine signaling, TNF signaling, NOD-like receptor signaling, NF-kappa B signaling, and inflammatory response. With the exception of CXCL12, all enriched genes were downregulated in CCs from patients with endometriosis. These findings provide a significant achievement in the field of reproductive biology, directing future studies to discover biomarkers of oocyte quality in endometriosis.

Endometriosis is a chronic inflammatory and estrogen-dependent disorder characterized by functional endometrial-like tissue outside the uterus 1 . According to the American Society for Reproductive Medicine, endometriosis can be classified into four stages: I-minimal, II-mild, III-moderate, and IV-severe 2 . It affects 6-10% of women of reproductive age 3 , and approximately 30-50% of them are estimated to be infertile 4 . The mechanisms underlying the etiopathogenesis of infertility in patients with endometriosis remain unclear, especially in the initial stages (minimal and mild), when no distortions and adhesions in the reproductive tract are present 5 .
Nonetheless, invasive oocyte analyses are not routinely feasible since human oocytes are rarely donated to research centers, and their application in invasive techniques precludes subsequent use in assisted reproduction procedures. Thus, the indirect evaluation of oocyte quality may contribute to understanding endometriosisrelated infertility [18][19][20] .
Indirect oocyte quality assessments are highly desirable, and, in this context, cumulus cells (CCs) may be more feasible targets of analysis. CCs are a specialized type of granulosa cells (GCs) 21

Consent for publication.
All authors have read the manuscript and approved its publication.

Results
Flowchart. During the recruitment period, a total of 54 patients were deemed eligible. However, seven patients did not agree to participate in the study. Thus, 47 patients provided written informed consent and began controlled ovarian stimulation for intracytoplasmic sperm injection. Ten patients did not undergo oocyte retrieval, whereas 37 did. There was no oocyte in 4 of them, and five patients exhibited few CCs, impeding donation for the study. Therefore, the obtained CCs were donated by 28 patients. Total RNA was isolated from the CCs, and RNA integrity was assessed, although it was unacceptable in 10 samples. We obtained 18 samples, nine controls, and nine patients with endometriosis I/II; the samples were clustered in pools of three patients each. The flowchart is depicted in Fig. 1.
Clinical variables. No significant differences between groups were observed regarding age, BMI, infertility time, and the number of oocytes ( Table 1). The majority of patients (83.3%; N:15) were submitted to the flexible antagonist protocol plus rFSH or menotropin. Only 3 (16.7%; 1 control and two endometriosis I/II patients) underwent the minimal stimulation protocol.

RNA next-generation sequencing.
Approximately 50 million reads per sample were obtained with RNA-Seq, and the average mapping quality was 92.8%. RNA-Seq provides the differential gene expression profiles in CCs of patients with endometriosis I/II compared to the controls. Such analysis enabled us to obtain a list of 26 DEGs (21 down-regulated and five up-regulated, in endometriosis I/II) with adjusted p < 0.05 as significant (Fig. 2).  (Table 2). Some genes were present in more than one pathway, all of which share multiple genes. Interestingly, all enriched genes underwent a negative regulation of www.nature.com/scientificreports/ expression in the CCs of the patients with endometriosis I/II compared to the controls, except for the CXCL12, which was up-regulated. The enrichment analysis also highlighted an essential biological process, the inflammatory response with 12 altered genes (CCL20, CXCL1, CXCL2, CXCL3, CXCL5, CXCL8, CXCL12, TNFAIP3, TNFRSF9, IL1A, IRAK2, and NFKB1) (p < 0.001).
Of the 26 DEGs, 25 were protein-coding genes. These genes encode 25 proteins, which were used to enrich protein-protein interaction networks (Fig. 3). The protein-protein interaction enrichment p value was < 1.0e−16.
The same pathways enriched using DEGs were found in the protein analysis (Fig. 3). The interaction enrichment showed that these proteins have significantly more interactions among themselves than expected for a random set of proteins of the same size and degree distribution drawn from the genome. Such enrichment indicates that the proteins are at least partially biologically connected as a group.

Discussion
Endometriosis related to infertility seems to be associated with oocyte impairment, mainly in the initial stages of endometriosis, when no distortions or adhesions in the reproductive tract are present 6,42,43 . Exposure to hostile environments with macrophages, cytokines, and reactive oxygen species in the peritoneal and follicular fluid could lead to dysfunctional folliculogenesis and worsen oocyte quality 5,13,44,45 . Within this environment, the cross-talk between oocyte and CCs is crucial for oocyte development 24,25 . Therefore, CCs reflect oocyte status and could be used as an index of oocyte quality 10,12,27,28 . A large-scale analysis is essential to comprehend the biological changes in CCs. This study was the first to evaluate the transcriptome of CCs from infertile patients with endometriosis I/II compared to women without the disease.
The differential gene expression profile in the CCs of patients with endometriosis I/II showed 26 DEGs compared to the controls, demonstrating that endometriosis I/II is related to the deregulation of the CCs' transcriptome. Subsequently, enrichment analysis showed altered molecular mechanisms in the CCs of patients with endometriosis I/II; Cytokine-cytokine receptor interactions, Chemokine signaling, TNF signaling, NOD-like receptor signaling, and NF-kappa B signaling. These pathways are related to immunity, and, except for CXCL12, all enriched genes are downregulated in endometriosis CCs. Allegra et al. (2014) also found deregulated genes in all these pathways from CCs of women with severe endometriosis by microarray analysis 27 .
It is known that endometriosis is a chronic inflammatory disease that can cause excessive reactive oxygen species (ROS) accumulation and, consequently, intra-follicular oxidative stress, even in infertile women with endometriosis I/II 46 . Lin et al. (2020) found an increase in ROS in the granulosa cells of patients with endometriosis and suggested that this process induces cell senescence, contributing to endometriosis-associated infertility 47 . This proinflammatory and ROS-filled microenvironment can trigger immune system pathways like those found in our study.
NOD-like receptors are known as recognition receptors, responsible for recognizing pathogen-associated molecular patterns released by damaged cells 48 . The activation of these receptors leads to the transcription of several genes, including NFKB, which induces inflammatory cytokines and chemokines 49,50 . In the ovary, cytokines and chemokines promote leukocyte recruitment and activation, steroidogenesis, follicular growth, and ovulation 51,52 . In the literature, several cytokines are found down-regulated in endometriosis CCs when compared to the controls. Moreover, follicular fluid cytokines appear to be related to successful pregnancy following IVF treatments 52 . The TNF signaling pathway acts in several processes, including cell proliferation, differentiation, and apoptosis, in addition to the modulation of immune and inflammatory responses 53 . In part of this pathway, the gene TNFAIP6 plays an essential role in forming the extracellular matrix of the cumulus-oocyte complex 54,55 . Allegra et al. (2014) showed the down-regulation of the TNFAIP6 gene in CCs of patients with endometriosis 27 . All of these pathways are essential for ovulation, as well as fertilization. The expansion of the cumulus-oocyte complex can be improved by activating Toll-like receptors, followed by genes such as NFKB, besides cytokines and chemokines 50 . An inflammatory process marks the rupture of the follicle. Moreover, sperm induces the release of cytokines and chemokines from CCs, enhancing the fertilization process 56,57 . Therefore, alterations in these intricate molecular mechanisms may compromise oocyte quality and decrease fertilization rates.
The main limitation of this study was its small sample size, resulting from the strict eligibility criteria adopted and the low RNA integrity and concentrations in CCs. Also, pooling samples might not be beneficial when the gene expression levels display low variability and reduce samples. However, small RNA sample pools effectively reduce the variability and compensate for the loss of replicates 38 . Furthermore, the data obtained from studies www.nature.com/scientificreports/ using samples collected after different controlled ovarian stimulation protocols may not necessarily be extrapolated to natural cycles.
In conclusion, the present study shows, for the first time, that endometriosis I/II could promote alterations in the transcriptome of CCs. These results provide a better understanding of the mechanisms (Cytokine-cytokine receptor interactions, Chemokine, TNF, NF-kappa B, NOD-like receptor signaling, and inflammatory response) that may affect oocyte competence acquisition in patients with endometriosis I/II. These pathways share a variety of genes and cannot be considered an individualized process. This differential transcription profile provides a significant achievement in the field of reproductive biology, directing future studies with a larger cohort to discover biomarkers of oocyte quality in endometriosis, be they pathways or genes.