Abstract

Endometriosis is a common chronic inflammatory condition whose clinical manifestations extend beyond pelvic pain and infertility. The diversity of symptoms'including neuropathic pain, digestive disorders, and systemic comorbidities'as well as the frequent lack of correlation between symptoms and lesion distribution, suggest a disease with systemic expression, going beyond a purely local lesion-driven process. While the theory of retrograde menstruation has historically dominated as the primary hypothesis, it fails to explain why only 10% of women develop the disease. An alteration of the immune response appears necessary but not sufficient. Furthermore, the presence of bacteria or bacterial derivatives in the peritoneal and uterine cavities, together with well-established links between the microbiota, immunity, and immuno- inflammatory diseases, point toward a potential bacterial involvement in the pathophysiology of endometriosis. We hypothesized that certain bacterial species, through their interactions with mucosal tissues and the immune system, could contribute to lesion formation or promote the chronic inflammation characteristic of endometriosis. This thesis aimed to explore these relationships through a dual approach: an initial phase of critical synthesis of current knowledge on the digestive, vaginal, and uterine microbiota in endometriosis, followed by an experimental investigation using both murine models and clinical human data. First, a murine model of endometriosis was used to evaluate the impact of a Western-type pro-inflammatory diet. This diet led to a doubling in lesion volume compared to a standard diet and was associated with the disappearance of Akkermansia muciniphila, a bacterium known for its beneficial immunomodulatory properties. Moreover, the induction of endometriosis alone significantly altered the composition of the gut microbiota, independently of dietary changes, suggesting a bidirectional interaction between disease and microbiota. The second, clinical component focused on the largest cohort of endometrioma infections described to date. The data revealed a bacterial ecology distinct from that of upper genital tract infections, dominated by intestinal bacteria, primarily Escherichia coli. This particular tropism of E. coli for endometriotic tissue suggests a specific affinity for this tissue. MUC1 is increased in the endometrial stroma of patients with endometriosis. In vitro, exposure of endometrial stromal cells to E. coli lipopolysaccharides induces a pro-inflammatory response through activation of the PI3K/Akt pathway, a response not observed following exposure to A. muciniphila. Finally, a pilot 16S rRNA sequencing study of non-infected endometrioma contents identified specific bacterial signatures, including E. coli and A. muciniphila, absent from technical controls, suggesting intrinsic colonization. These results support a strong hypothesis: certain bacterial species act as key modulators of the immuno-inflammatory response in endometriosis. Some appear deleterious, others potentially protective. This emerging paradigm'integrating bacterial, immune, and environmental dimensions'opens promising avenues for prevention, diagnosis, and innovative treatments, particularly through targeted modulation of the microbiota.