Endometriosis is a gynecological inflammatory disease affecting women of reproductive age (1, 2). Approximately 200 million women worldwide, 10% to 15% of women of reproductive age and 2.5% of postmenopausal women are affected by endometriosis (3, 4). Three main forms of endometriosis are described: (i) ovarian endometriosis, (ii) superficial peritoneal endometriosis and (iii) deep infiltrating endometriosis (DIE) (1, 5), the latter being recognized as the most severe form (6, 7). Endometriosis is defined by implantation and invasive growth of endometrial tissue in extra-uterine locations, causing chronic pelvic pain, dyspareunia, dysmenorrhea, menorrhagia, bowel symptoms, and infertility (1, 6, 8). Endometriosis symptoms are associated with substantial reductions in quality of life (9, 10). Living with severe cyclic or continuous pelvic pain can lead to stress, anxiety, depression and absenteeism from work (11, 12).

Endometriosis is a multifactorial disease, with complex pathophysiological mechanisms, of which genetic and environmental components are still poorly evaluated (13, 14). The etiology of endometriosis is not completely understood. Several hypotheses have been put forward concerning the histological origins of endometriosis; the most accepted theory being Sampson’s theory of retrograde menstruation, which involved fragments of menstrual endometrium being disseminated through the fallopian tubes (7, 9, 13). However, this phenomenon is observed in nearly 90% of women, suggesting that immune and hormonal dysfunctions may add to the observed fragmentation, yielding the adhesion, survival and proliferation of the lesions (1, 2, 10). Several mechanisms such as exacerbated production of growth and pro-inflammatory factors, an increase in estradiol expression combined with progesterone resistance, and an overexpression of reactive oxygen species might be involved in the development of endometriosis (10, 14, 15). Distinct immunological abnormalities involving angiogenesis, vasculogenesis and inflammation have been well described. These processes involve molecules including the VEGF factor that triggers angiogenesis, tumor necrosis factor (TNF)-α, which plays an essential role in increasing proliferative potential and acts primarily as a precursor to initiating an inflammatory response by activating a cascade of other cytokines, such as IL-1, IL-6 and CXCL8 (8, 16–18). The ineffectiveness of using anti-inflammatory agents to treat endometriosis shows that the disease is more related to the loss of balance between pro- and anti-inflammatory molecules (4). It appears clearly today that to understand the complexity of this disease, it is necessary to study the processes involved as a whole, as well as potential interactions between their components.

There is currently no treatment for endometriosis. On the other hand, the time between the development of the first lesions and the diagnosis is estimated between 7 and 10 years (2, 10). Thus, two major challenges must be met: identification of early diagnostic biomarkers, and that of potential therapeutic targets. An increasing number of studies are based on the search for biomarkers involved in endometriosis in order to develop less invasive diagnostic methods (i.e., urine tests, blood tests) (19, 20). The biological complexity under endometriosis has been previously addressed using computational biology approaches on the basis endometriosis-related alternations (e.g. immune cell infiltration) (21) or the development and progression of endometriosis (22). However, to the best of our knowledge, no study had used endometriosis-related symptoms to build up its underlying biomolecular processes.

Finding useful information in the genetic data generated for endometriosis is very challenging and computational biology can be helpful. Our study used tools of computational biology to identify biological processes and protein networks underlying the symptomatology of endometriosis. We combined text-mining, functional enrichment and protein-protein interaction analyses to suggest some biomarkers or therapeutic targets deserving further exploration.

This study provided a text-mining approach that tapped data from bioinformatics banks, with the aim of investigating the protein network related to the symptomatology of endometriosis. A holistic approach was favored to understand the associated complex biological mechanisms behind this symptomatology. Expressed protein may support the advent of some symptoms, which should help health professionals and clinicians in their investigations. As an attempt to address the knowledge gaps surrounding this disease, a special feature of our approach relies on the interest in all the genes collected in the literature as being related to the symptomatology of endometriosis. These genes are often detected based on the results of isolated experimentations.

Thirty-five key proteins were identified in the current study as potential modulators of the symptomatology of endometriosis. Inflammation in endometriosis, widely supported by existing literature, is reflected by an overexpression of inflammatory cytokines, inhibition of endothelial function and hormonal dysregulations (10, 33). A pro-inflammatory cytokine such as Interleukin-1 β (IL-1β) may enhance the proliferation of endometriotic cells. IL-1β can also trigger the production of IL-6 and IL-8 (other pro-inflammatory cytokines), which are involved in more proliferation and the decrease of apoptotic rate (14). Inflammation state not only leads to dysmenorrhea, dyspareunia and infertility (34) but also cause oxidative stress connected to poor‐quality embryos and immature oocytes (14). The over-activation of macrophages and mast cells in endometriosis produces IL-1β, TNF-α, IL-6, and IL-10 (35) found in our analysis. Even if the involvement of the IL-10 pathway in endometriosis is still poorly understood, the IL-10 signaling pathway has an ability to block cytokines and chemokines from macrophages being the root of inflammatory processes (36, 37). Moreover, anti-inflammatory signaling pathways IL-4 and IL-13 are involved in the cellular immune response, particularly T helper type 2 (38). IL-4 and IL-13 type I and II receptor signaling pathways are linked to Signal Transducer and Activator of Transcription 6 (STAT6) expression (39), responsible for mediating IL-4 and IL-13 immune signaling, increase proliferation, adhesion and viability of endometriosis lesions (38). Suppressor Of Cytokine Signaling protein 1 (SOCS1) inhibits STAT6 expression and has a protective effect by activating cell apoptosis (39). SOCS1 dysregulation may exacerbate the IL-4 and IL-13 properties associated with endometriosis.

Hence, the findings from the current study pinpoint that both pro-inflammatory and anti-inflammatory pathways are involved in the symptomatology of endometriosis. This was assumed to play a paradoxical role in acute and chronic phases of the disease due to pathological immune imbalance (38). Interestingly, a study on the mouse model has shown that pre-existing peritoneal inflammation did not contribute to the development of endometriosis and could even reduce it (33). This supports the need to provide understanding of the precise role of inflammation in this disease.

Some previously identified targets in the literature (i.e. IL-1, IL-6, IL-4, and VEGF) were emphasized by our analysis. The IL-1 pathway is a pro-inflammatory cytokine that activated NF-kB inflammatory pathways (34). Dysregulation of cytokines and NF-kB factor induces both an inflammatory process and an immune system dysfunction involved in endometriosis (40). The NF-kB biological pathway has several subunits such as the p65 involved in the regulation of cell survival. The pathway expressed by p65 also plays a role in the inflammatory response and contributes to angiogenesis and metastasis survival (41–43). NF-kB1 subunit promotes the expression of inflammatory cytokines (44, 45).

Similar to the existing literature, the implication of TNF-α is again emphasized in this study on Figure 4 , portraying the most important protein interactions. TNF-α, which plays an essential role in increasing proliferative potential, acts primarily as a precursor in initiating an inflammatory response by activating a cascade of other cytokines, such as IL-1, IL-6 and the vascular endothelial growth factor (VEGF) (7, 46). Abnormal VEGF levels may impair the process of angiogenesis, which is favorable to embryonic implantation, thus justifying the high miscarriage rate (47, 48). The current study also highlighted some lesser-known targets (i.e. PGE1, AVP) (49–51), which may require deeper investigations. Excessive expression of Prostaglandin E receptor 1 (PGE1) is linked to an inflammatory reaction as for tumor protein 53 (p53). Unlike our finding, no link between endometriosis and arginine vasopressin (AVP) or Caspase-3 has been reported in the literature, and this may require more exploration (46–48, 52). Furthermore, the KRAS protein, stressed in the protein network ( Figure 4 ) is a factor that is overexpressed in skeletal muscle and myocardium, uterus, adrenal cortex and some bone marrow stem cells (53). A mutation of KRAS can cause hyperplasia. This may explain the occurrence of endometrial hyperplasia in endometriosis (49, 54). TGF-β1 also plays a role in muscle diseases by inhibiting myogenesis and regeneration (55). This involves the development of fibrosis or atrophy of the muscle skeleton (56, 57). Finally, our study pointed out Neurotrophic Receptor Tyrosine Kinase 1 (NTRK1) protein, which may be involved in neuropathic pain. Nerve growth factor (NGF) is the main ligand of NTRK1 (26). It was shown that NGF is directly associated with pelvic pain and more specifically with dysmenorrhea, dyspareunia, painful bladder syndrome and irritable bowel syndrome (58, 59). These pains reach the nervous system by provoking a nociception. Such a damage to the nervous system would lead to neuropathic or neuroinflammatory pain (59, 60). In summary, this apparent crosstalk between immune cells, nerves, and central pain pathways is providing an opportunity to develop more targeted therapies against endometriosis and its symptoms (20).

Although the proteins identified should be taken with caution, given the heterogeneity of the studied tissues, the different forms of the disease and techniques used, the current study provides an overall picture of the proteins involved in the symptomatology of endometriosis. In future studies, it would be interesting to examine the involvement of these proteins given the stage of endometriosis and the phases of the menstrual cycle. It is noteworthy that protein interactions that are not already known in the STRING database may lead to discarding key targets involved in the symptomatology of endometriosis. Some proteins have been deeply investigated in the literature while others are rarely studied proteins, and thus may not be detected during text-mining. Also, the network of proteins obtained does not allow weighing to them according to the importance of their involvement in the disease. Thus this approach proves to be complementary to other studies exploring the completeness of the genome (e.g. Genome-wide association studies, GWAS) or using the Gene Expression Omnibus database (GEO) to study differentially expressed genes (61–63). Our study still remains an exploratory analysis that was based on the general symptomatology of endometriosis. We do not get any information on either the different forms of endometriosis or the different stages of the disease. Potentially, this approach can be replicated at a higher level of detail to allow comparing the biological implications of eutopic and ectopic endometrium. In the same way, the query can be refined to select articles from different stages of the disease (early or advanced stage) or from different phases of the menstrual cycle. Hence, the comparison of the signaling pathways related to the early stage of endometriosis (with a search for genes in the articles involved in the early stage of the disease) and the more advanced stage warrants attention. Finally, because Pubtator tools may not have perfect discriminative ability in distinguishing between genes and alleles, some overlapping cannot be completed ruled out.

Taken together, our data highlight that pathways associated to endometriosis symptomatology have sometimes paradoxical roles, certainly resulting in a loss of balance, and these may be time-dependent. Then, developing strategies to enhance their protective effects or to combat their pathological responses at specific stages of the disease could prove therapeutic potential for endometriosis. In conclusion, our study identifies 35 interrelated key proteins with the highest ability to control pathways associated to endometriosis symptomatology. While some proteins such as IL-1β, IL-6, IL-4, and VEGF are largely evaluated, our data are suggestive of further investigation on proteins such as PGE1 and AVP. Our study prioritizes potential biomarkers and key targets, and further assessing them in endometriosis could help for the development of diagnostic tools and therapeutic strategies for endometriosis.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

FE and MF designed the study, contributed to the methodology, data curation and execution of the study, analyzed the data and wrote the manuscript. KB and ED-D contributed to the methodology and reviewed the manuscript. AL, ML, and AN reviewed the manuscript. BG and DZ designed the study and revised the manuscript providing their expertise. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.