Orchestrating the gut microbiota-mitochondrial-immune axis in gynecological diseases: mechanisms and dual-targeting therapeutic strategies

The interaction between gut microbiota and mitochondria shows specific clinical phenotypes in different gynecological diseases. Table 1 ( 17 , 37 , 41 , 83 – 96 ) summarizes representative alterations in microbiota composition, metabolites, and mitochondrial function across major gynecological diseases, including polycystic ovary syndrome(PCOS), endometriosis, premature ovarian insufficiency (POI), and gynecological malignancies. Integrated alterations in gut microbiota, metabolites, and mitochondrial function across gynecological diseases and their immunological consequences. [H], human observational/clinical studies; [A], animal studies; [I], in vitro mechanistic experiments. Alpha diversity refers to metrics (Shannon index, Chao1) derived from fecal samples in human observational studies. Peritoneal fluid and intratumoral microbiomes are compositionally distinct from fecal microbiota and cannot be directly compared for diversity metrics. Taxonomic shifts represent relative abundance changes and may vary across cohorts due to heterogeneity in study design, population characteristics (BMI, geography, diet), and analytical pipelines. Prevotella abundance in PCOS shows inconsistent directionality across cohorts and is not reported as a uniform increase. Mitochondrial phenotypes involving cGAS–STING activation and PINK1/Parkin-mediated mitophagy are primarily supported by experimental models; direct human tissue evidence is limited. TMAO elevation in POI is based mainly on animal model data; human metabolomics evidence remains limited. MR, Mendelian randomization. F/B ratio, Firmicutes/Bacteroidetes ratio. SCFAs, short-chain fatty acids; BAs, bile acids; LPS, lipopolysaccharide; mROS, mitochondrial reactive oxygen species; ΔΨm, mitochondrial membrane potential; TMAO, trimethylamine N-oxide; mtDNA, mitochondrial DNA; GDCA, glycodeoxycholic acid; TUDCA, tauroursodeoxycholic acid; HDAC, histone deacetylase; FXR, farnesoid X receptor; TGR5, Takeda G-protein-coupled receptor 5; Drp1, dynamin-related protein 1; TME, tumor microenvironment; MDA, malondialdehyde; NLRP3, NLR family pyrin domain-containing protein 3 inflammasome; cGAS–STING, cyclic GMP-AMP synthase–stimulator of interferon genes pathway; PGC-1α, peroxisome proliferator-activated receptor- γ coactivator 1α; PINK1/Parkin, PTEN-induced kinase 1/E3 ubiquitin ligase Parkin; MR, Mendelian randomization. However, when interpreting the above research results, we need to face the inherent heterogeneity of current microbiota research. There are often differences in microbiota signals among different population groups. Such differences are not statistical errors, but an important part of research evidence. Dietary pattern, body mass index, insulin resistance, region, medication history, and other confounding factors will significantly affect the host microbiota ( 97 ); Methodological differences such as sequencing processes, analysis methods, and sample type selection will also change the results of taxonomic classification ( 98 ). In addition, the mechanisms of cGAS–STING and PINK1/Parkin-mediated mitophagy are mostly based on experimental research, which should be interpreted accordingly. Therefore, the relationships summarized herein should be understood as context-dependent rather than universally applicable. To ensure the rigor of the research method and avoid exaggerating the clinical readiness, the following sections analyze the research evidence for each disease from three levels: clinical evidence and human observations, animal model mechanism support, and in vitro cell studies, and distinguish verified mechanisms and speculative hypotheses from those clinically documented, so as to provide a clear basis for subsequent translational research. PCOS is a complex endocrine disease, the core features of which are hyperandrogenism, ovulatory dysfunction, and serious metabolic problems such as insulin resistance ( 99 ). Figure 3 shows the multi-organ regulatory network related to gut microbiota dysbiosis, systemic metabolic disorder, and ovarian hyperandrogenism. Gut dysbiosis–mitochondria–insulin resistance axis in PCOS-associated hyperandrogenism. Gut dysbiosis reduces SCFA production, potentially weakening AMPK–PGC-1 α signaling and impairing mitochondrial biogenesis and fatty-acid oxidation (FAO). Increased gut permeability may elevate circulating LPS levels, thereby activating TLR4–NF-κB signaling and contributing to chronic low-grade inflammation. These metabolic and inflammatory inputs converge on mitochondrial dysfunction (↑mROS, ↓ATP, and altered mitochondrial dynamics) and impaired insulin signaling in muscle, liver, and adipose tissue, promoting systemic insulin resistance and hyperinsulinemia. Hyperinsulinemia may synergize with luteinizing hormone (LH) to enhance androgen synthesis in ovarian theca cells, while mitochondrial stress may impair steroidogenic regulation involving StAR, CYP11A1, and CYP19A1, thereby disrupting steroid balance. Excess androgen production may further aggravate metabolic dysfunction, forming a positive feedback loop relevant to PCOS progression. Clinical manifestations include increased HOMA-IR, elevated testosterone/free androgen index (FAI), menstrual irregularity, and polycystic ovarian morphology. Solid arrows indicate relatively well-supported mechanisms, whereas dashed arrows represent indirect or emerging pathways. PCOS, polycystic ovary syndrome; SCFAs, short-chain fatty acids; LPS, lipopolysaccharide; TLR4, Toll-like receptor 4; NF-κB, nuclear factor kappa B; mROS, mitochondrial reactive oxygen species; IR, insulin resistance; LH, luteinizing hormone; StAR, steroidogenic acute regulatory protein; HOMA-IR, homeostatic model assessment for insulin resistance; FAI, free androgen index; FAO, fatty-acid oxidation. Sequencing data from population cohorts showed that the gut microbiota composition of PCOS patients differed from that of healthy controls, mainly manifested by decreased microbial alpha diversity, reduced SCFA-producing taxa, decreased abundance of beneficial bacteria such as Bifidobacterium and Faecalibacterium prausnitzii , and altered Prevotella abundance across cohorts, which were associated with clinical indicators such as insulin sensitivity, blood lipid levels, and serum androgen concentration ( 100 , 101 ). Metabolomics analysis also found that PCOS patients had disordered metabolism of circulating secondary BAs and insufficient levels of SCFAs ( 102 ). Regarding clinical interventions, emerging evidence from recent randomized controlled trials (RCTs) and systematic reviews suggests that probiotic and synbiotic supplementation can modulate gut microbial composition and improve insulin resistance and hyperandrogenism in PCOS patients ( 103 , 104 ). A 2024 RCT further reported a significant reduction in total testosterone following synbiotic intervention combined with lifestyle modification ( 105 ). Recent clinical trials are also exploring dietary strategies, including Mediterranean diet interventions, for improving metabolic and endocrine outcomes in PCOS patients ( 106 ), while a head-to-head RCT by Borzan et al. indicated that probiotic supplementation significantly improved metabolic and endocrine parameters compared with metformin and placebo ( 107 ). However, despite these promising findings, direct clinical evidence linking gut-targeted interventions to restoration of mitochondrial bioenergetics remains lacking. Mechanistic pathways identified in animal models, such as the Escherichia coli Nissle 1917–IL-22–mitochondria axis, still await clinical validation ( 108 ). Therefore, larger standardized clinical trials are needed to determine whether the proposed gut–microbiota–mitochondrial axis can be functionally modulated in humans ( 104 , 109 ). Animal models of PCOS further supported the above clinical phenomena at the mechanistic level and suggest that changes in the microbiota may contribute to systemic endocrine disorders. SCFA-producing taxa in the model decreased, which was associated with suppression of the AMPK/PGC-1 α signaling axis and reduced mitochondrial biogenesis and fatty acid oxidation ( 110 ). At the same time, intestinal permeability increased, which promoted LPS to enter the circulation, thereby contributing to endotoxemia, activated the TLR4/NF-κB pathway, and formed a chronic low-grade inflammatory state ( 111 , 112 ). Inflammation may further aggravate systemic insulin resistance and lead to hyperinsulinemia. Hyperinsulinemia cooperated with luteinizing hormone (LH) to overactivate ovarian theca cells, leading to excessive androgen synthesis, forming a vicious cycle of “insulin resistance-hyperinsulinemia-hyperandrogenism” ( 112 , 113 ). In vitro experiments at the cellular level was associated with mitochondria are involved in local ovarian dysfunction. After metabolic or inflammatory stress treatment, granulosa cells and theca cells exhibit obvious mitochondrial energy metabolism disorders, characterized by decreased ATP production, damaged membrane potential, and a large amount of mitochondrial ROS accumulation ( 102 , 114 ). Excessive mitochondrial ROS directly inhibits cholesterol transport into mitochondria mediated by the STAR protein, reduces the activities of key steroidogenic enzymes such as CYP11A1 and CYP19A1, disrupts the processes of steroid production and follicular maturation, and aggravates the androgen phenotype of PCOS patients ( 115 ). Endometriosis is a chronic inflammatory disease. The core pathology is the ectopic growth of endometrial-like tissue, which is often accompanied by dysmenorrhea, chronic pelvic pain, and infertility ( 116 ). Existing studies show that the pathogenesis of endometriosis is closely related to gut microbiota dysbiosis, mitochondrial dysfunction, and amplification of the innate immune response in the peritoneal microenvironment, as illustrated in Figure 4 ( 16 ). Two-signal inflammasome activation model in endometriosis. Signal 1 (priming): gut dysbiosis with reduced SCFAs and barrier leak increases LPS translocation, activating TLR4–NF- κ B signaling in macrophages and dendritic cells and promoting the upregulation of NLRP3 and pro-IL-1β/pro-IL-18. Signal 2 (activation): lesion- and immune-cell–associated mitochondrial damage (↑mROS, ↓ΔΨm, and mtDNA release), together with ionic flux (K + efflux/Ca 2+ influx), promotes NLRP3 assembly and caspase-1 activation, resulting in IL-1β/IL-18 maturation, GSDMD pore formation, and pyroptosis. These inflammatory events may contribute to chronic inflammation, fibrosis, pain sensitization, and impaired endometrial receptivity associated with infertility. Therapeutic leverage points include microbiota restoration (probiotics/prebiotics), inhibition of the TLR4/NLRP3–caspase-1–GSDMD axis, and mitochondrial rescue strategies involving AMPK–PGC-1α activation and PINK1/Parkin-mediated mitophagy. Solid arrows indicate activation and signaling pathways, whereas dashed arrows represent therapeutic interventions. SCFAs, short-chain fatty acids; LPS, lipopolysaccharide; TLR4, Toll-like receptor 4; NF-κB, nuclear factor kappa B; IL, interleukin; mROS, mitochondrial reactive oxygen species; ΔΨm, mitochondrial membrane potential; mtDNA, mitochondrial DNA; GSDMD, gasdermin D. Clinical cohort studies found that the gut microbiota composition of patients changed significantly, mainly manifested by the proliferation of Proteobacteria (e.g., Escherichia coli ) and a decrease in dominant protective bacteria such as Lactobacillus ( 117 , 118 ). Recent studies further suggested that estrobolome alterations may contribute to estrogen metabolism imbalance in endometriosis, thereby influencing inflammatory activity and disease progression ( 119 ). Microbiota dysbiosis was associated with reduced levels of beneficial SCFAs in the systemic circulation and peritoneal cavity, as well as impaired intestinal barrier integrity and a possible “leaky gut” phenotype. These alterations may contribute to elevated endotoxin (LPS) levels in the circulation and peritoneal fluid, which have been associated with the severity of pelvic inflammation and disease stage ( 120 , 121 ). Regarding clinical intervention, evidence remains limited but is gradually emerging. Adjunctive probiotic or synbiotic therapy has been associated with reduced inflammatory markers and improved postoperative pain in endometriosis patients ( 122 ). A triple-blind RCT further showed that astaxanthin supplementation enhanced antioxidant capacity while reducing pro-inflammatory cytokines ( 123 ). Dietary interventions, including a low-FODMAP regimen, also demonstrated symptomatic improvement in randomized trials ( 124 ). However, current clinical evidence remains insufficient to determine whether gut-targeted interventions can directly modulate the peritoneal inflammatory microenvironment or alter disease progression in endometriosis. Recently published trials, including the ProMetrioS study, together with ongoing anti-inflammatory dietary intervention studies, may provide further translational insights into microbiota-based therapeutic strategies ( 125 , 126 ). Animal models of endometriosis provide an in vivo mechanistic basis for clinical observation. The endometriosis model supports that intestinal endotoxemia is a persistent cause of inflammation in the peritoneal microenvironment, and that LPS continuously activates the TLR4/NF-κB pathway in peritoneal immune cells ( 127 , 128 ). Experimental studies further suggested that mitochondrial stress-associated mtDNA release may contribute to cGAS–STING activation and amplification of peritoneal inflammation ( 37 ). Intervention on the gut-peritoneal axis by depleting the microbiota or blocking TLR4 significantly reduced the volume of ectopic lesions and decreased inflammatory cell infiltration, proving that the synergistic effect of intestinal PAMPs and local inflammation contributes to the colonization and angiogenesis of ectopic endometrial lesions ( 129 ). In vitro experiments at the cellular level using ectopic endometrial stromal cells and peritoneal macrophages clarified that mitochondrial stress and the NLRP3 inflammasome are the final effector factors of disease regulation ( 130 ). The microbial initiation signal aggravates mitochondrial dysfunction, leading to the accumulation of endogenous DAMPs such as decreased membrane potential, excessive production of mitochondrial ROS, and cytoplasmic release of mtDNA. Mechanistic experiments suggested that these endogenous danger signals may promote NLRP3 inflammasome activation, accompanied by caspase-1 activation, Gasdermin D (GSDMD) cleavage, pyroptosis, and increased release of IL-1β and IL-18, thereby contributing to a fibrotic and inflammatory local microenvironment ( 131 ). This microbiota-associated mitochondrial inflammatory signaling cascade inhibits ectopic cell apoptosis, aggravates local fibrosis, and increases pain sensitivity, which eventually leads to a decrease in endometrial receptivity associated with infertility in endometriosis patients ( 132 , 133 ). POI refers to premature depletion of the ovarian follicular pool before the age of 40, leading to gonadotropic hypogonadism and infertility. More and more studies have suppoeted that the gut microbiota–mitochondria axis is key to regulating ovarian lifespan, and that intestinal metabolic disorders amplify systemic inflammation and destroy ovarian mitochondrial quality control ( 88 , 134 ). Clinical data show that compared to the healthy population, the intestinal microbiota of POI patients has decreased alpha diversity and altered abundance of specific taxa ( 131 ). The increase in circulating microbial metabolites such as TMAO is positively correlated with the decline in ovarian reserve function ( 88 ). However, as these observations are largely correlational, it remains to be determined whether gut dysbiosis is a primary driver of POI or a secondary consequence of the altered hormonal environment. Clinical interventional evidence for POI remains particularly scarce. To date, no completed RCT has directly evaluated microbiota-targeted interventions for restoring ovarian reserve in POI patients; the closest human-level evidence is a Mendelian randomization study identifying genetically predicted causal links between specific gut taxa and POI risk, though the authors explicitly noted that clinical trials are still required to confirm these associations ( 89 ). Prospective interventional studies are urgently needed to determine whether gut-targeted approaches could offer a supportive strategy to preserve residual ovarian function in this population ( 91 , 134 ). Animal models of POI have provided preliminary mechanistic insights into the interaction between the gut microbiota and ovarian function. These studies showed that microbiota depletion induced by broad-spectrum antibiotics, together with exposure to POI-inducing toxins, was associated with follicular atresia and diminished ovarian reserve ( 135 , 136 ). At the mechanistic level, microbial metabolites and LPS translocation lead to systemic inflammation, thereby disrupting the follicular microenvironment. Experimental evidence further suggested that mitochondrial stress-associated signaling, including possible mtDNA-mediated cGAS–STING activation, may contribute to inflammatory amplification in the ovarian microenvironment ( 92 ). In addition, animal studies suggested that raising circulating SCFA levels can protect follicular development, indicating that microbial metabolic homeostasis can delay premature ovarian failure ( 137 , 138 ). At the cellular level, the POI disease phenotype may partly stem from the imbalance in the coupling relationship between the estrobolome and steroid production. The steroid production process of granulosa cells is extremely energy-consuming and highly dependent on mitochondrial ATP for energy supply ( 139 ). Experiments have supported that microbial metabolites such as TMAO and LPS can induce mitochondrial depolarization, excessive ROS production, and impaired mitophagy ( 140 , 141 ). These stress factors inhibit cholesterol transport mediated by the StAR protein, down-regulate the expression of CYP11A1 and CYP19A1, and hinder the synthesis of progesterone and estrogen ( 142 ). Abnormal mitochondrial-steroid production may form a pathological cycle: PINK1/Parkin-mediated mitophagy is impaired, leading to cytoplasmic release of mtDNA, activation of cGAS–STING inflammatory signaling, and further aggravation of granulosa cell aging and apoptosis ( 143 , 144 ). Continued activation of cGAS–STING promotes local formation of “inflammaging” in the ovary ( 92 ). This mitochondria-related chronic innate immune response induces granulosa cells to secrete aging-related phenotypes and accelerates follicular atresia and primordial follicle pool depletion ( 145 , 146 ). The gut microbiota–mitochondria axis triggers a sterile inflammatory cascade through mtDNA release, which represents a potential contributing mechanism in ovarian aging in POI patients. To summarize, the joint action of microbiota-induced inflammation and mitochondrial dysfunction is a key molecular target of POI ovarian lesions. Gynecological malignancies such as ovarian cancer and endometrial cancer exhibit metabolic plasticity. The gut microbiota–mitochondria axis is an important factor in regulating tumor metabolism, and can affect the tumor microenvironment (TME) by regulating metabolic reprogramming, immune cell infiltration, and endocrine sensitivity ( 93 ). Clinical cohort studies have found that specific gut microbiota composition characteristics are closely related to the progression of gynecological malignancies and the efficacy of chemotherapy. In the tumor microenvironment of patients with ovarian cancer and endometrial cancer, alterations in pro-inflammatory bacterial taxa, including Fusobacterium nucleatum , have been reported, which is associated with poor clinical prognosis and poor response to platinum-based chemotherapy ( 147 , 148 ). Metabolomics analysis also showed that the patients' circulating BAs and estrogen metabolites changed significantly, reflecting the status of the intestinal estrobolome and correlating with systemic inflammatory indices ( 8 , 148 ). However, a persistent translational challenge is distinguishing whether these microbial changes are primary drivers of tumorigenesis or secondary consequences of tumor-associated metabolic shifts and systemic anticancer therapies. Regarding clinical translation, research primarily focuses on mitigating treatment toxicity and optimizing immunotherapy outcomes. A meta-analysis indicated that probiotic supplementation in postoperative endometrial cancer patients enhances microbial diversity and alleviates gastrointestinal symptoms, though without clear oncological benefit ( 149 ). Additionally, fecal microbiota transplantation (FMT) combined with immune checkpoint inhibitors has shown promising response rates in solid tumors ( 150 ). For ovarian cancer, a randomized trial of probiotic intervention in patients receiving platinum-based chemotherapy is currently underway ( 151 ). Despite these efforts, direct evidence linking gut-targeted interventions to the modulation of the tumor microenvironment, mitochondrial bioenergetics, or improved clinical prognosis in gynecological malignancies remains lacking ( 149 , 152 ). Animal models of gynecological malignancies, including patient-derived xenografts and chemically induced tumor models, have provided mechanistic evidence supporting the involvement of the gut microbiota in shaping the tumor microenvironment. At the mechanistic level, intestinal metabolites such as SCFAs and secondary BAs act as systemic signaling molecules to regulate mitochondrial respiration of tumor cells. Altered bile acid signaling may also influence FXR/TGR5-mediated immune and metabolic regulation within the tumor microenvironment. Reduced SCFA availability may also impair HDAC-associated epigenetic regulation, thereby contributing to tumor-promoting metabolic and immune dysregulation within the tumor microenvironment ( 96 ). Endotoxemia caused by dysbiosis activates TLR4 signaling in the tumor microenvironment, thereby contributing to immunosuppressive cells and angiogenesis ( 153 ). At the same time, the gut microbiota–mitochondria axis can regulate the immune state of the tumor microenvironment; dysbiosis alters the mitochondrial function of tumor-infiltrating lymphocytes (TILs): insufficient intestinal SCFAs (especially butyrate) impair oxidative phosphorylation and memory formation in CD8+ T cells, leading to early T cell exhaustion and a decline in anti-tumor immune surveillance ( 154 ). Excessive production of mitochondrial ROS in the tumor microenvironment promotes tumor-associated macrophages (TAMs) to polarize towards the tumor-promoting M2 type, thus creating favorable conditions for tumor progression and metastasis ( 155 ). Certain bacteria, such as F. nucleatum have also been reported to enhance the metabolic plasticity of tumor cells, enabling them to adapt to the nutrient deficiency and hypoxic microenvironment common in gynecological malignancies ( 147 ). Experimental studies further suggested that dysregulated mitochondrial dynamics, particularly enhanced Drp1-mediated mitochondrial fission, may contribute to tumor proliferation and chemotherapy resistance, including reduced cisplatin sensitivity in gynecological malignancies ( 95 , 156 ). At the cellular level, the microbiota–mitochondria–apoptosis axis determines the survival and drug sensitivity of tumor cells. Estrobolome-mediated microbial metabolism regulates systemic estrogen levels through β-glucuronidase, up-regulates the glycolysis pathway of estrogen receptor-positive tumor cells, and strengthens mitochondrial fatty acid oxidation ( 8 , 157 ). At the same time, mitochondrial stress is a core mediating factor of chemotherapy resistance. In vitro experiments show that microbial signals can alter the mitochondrial membrane potential, regulate cytochrome C release, and increase the apoptotic threshold ( 158 , 159 ). Some microbial factors interfere with the quality control mechanism of mitophagy (PINK1/Parkin), potentially allowing tumor cells to evade mitochondrial-dependent apoptosis in the presence of taxane chemotherapy ( 160 ). The above research shows that targeting mitochondrial metabolic plasticity and microbiota-induced immunosuppression are important directions to overcome drug resistance in gynecological malignancies. At the molecular level, the interaction between mitochondrial stress and innate immune induction influences the efficacy of the tumor immune cycle. In theory, damaged mitochondria in gynecological tumor cells release mtDNA into the cytoplasm, which can activate the cGAS–STING pathway and recruit anti-tumor dendritic cells, but pathogenic bacteria such as F. nucleatum can inhibit this reaction and weaken the interferon-mediated immune killing effect ( 161 , 162 ). The interaction of microbial signals and mitochondrial dysfunction may represent an important mechanism contributing to tumor immune escape.

Gynecological diseases, such as polycystic ovary syndrome (PCOS), endometriosis, premature ovarian insufficiency (POI), and gynecological malignancies, seriously damage women's fertility and quality of life, and are closely related to endocrine disorders, abnormal energy metabolism, and chronic inflammation ( 1 , 2 ). In recent years, research on microbiology and mitochondrial biology has been continuously promoted. Many studies have supported that the interaction between gut microbiota and mitochondria is a common pathological mechanism in the occurrence and development of various gynecological diseases ( 3 , 4 ). Gut microbiota has been described as the “second genome” of the human body, which can produce short-chain fatty acids (SCFAs), bile acids (BAs), indole derivatives, trimethylamine N-oxide (TMAO) and other metabolites ( 5 , 6 ). It is noteworthy that the gut microbiota also constitute an “estrobolome”, that is, a collection of bacterial genes that regulate the level of estrogens in the whole body through the deconjugation and recycling of estrogens ( 7 , 8 ). These metabolites can modulate the energy metabolism and redox homeostasis of mitochondria, thereby regulating the processes of cell apoptosis, immune homeostasis, and hormone synthesis ( 3 , 9 ). At the same time, mitochondria are the core of cellular energy generation and stress signal transduction, and abnormal mitochondrial function will change the redox and inflammatory environment in cells, which in turn affects the composition and stability of gut microbiota ( 10 , 11 ). This two-way regulatory relationship constitutes the “gut microbiota–mitochondria axis”, and its role in metabolic and nervous system diseases has been increasingly recognized ( 12 , 13 ), and research in the field of gynecological diseases has gradually attracted attention. Existing research suggests that there is gut microbiota dysbiosis in PCOS patients, which reduces the production of SCFAs, aggravates endotoxin leakage, and further exacerbates mitochondrial oxidative stress and insulin resistance ( 14 , 15 ). In patients with endometriosis, the inflammatory reaction induced by lipopolysaccharide and mitochondrial DNA (mtDNA) released from damaged mitochondria can jointly activate the NLRP3 inflammasome and form a chronic inflammatory microenvironment ( 16 , 17 ). Microbial metabolite-induced mitochondrial apoptosis and metabolic reprogramming in patients with POI and gynecological malignancies lead to a decline in ovarian reserve function and tumor drug resistance, respectively ( 18 – 20 ). The above results propose that the interaction between microbiota and mitochondria is not a product of local lesions but a key phenotypic change in the progression of gynecological diseases. However, most current studies focus on gut microbiota dysbiosis or mitochondrial damage alone, and rarely integrate their related mechanisms, which not only limits the understanding of the pathogenesis of gynecological diseases but also hinders the development of targeted treatment programs. From the perspective of research methods, this paper is a narrative review, aiming to build an integrated theoretical framework. By Integrating evidence from population-based observational cohorts, animal models, and in vitro mechanistic studies, we summarize the basic biological laws of the interaction between gut microbiota and mitochondria. As illustrated in Figure 1 , we further analyze the specific processes of this regulator y axis across PCOS, endometriosis, POI, and gynecological malignancies. Critically, we also examine the current status of human clinical interventional evidence to date, identifying the limitations and translational challenges in this field. The purpose of this paper is to put forward a verifiable hypothesis for translational research and to provide new ideas for the study of the gut microbiota–mitochondria axis. Gut microbiome–mitochondria–immune–endocrine axis in gynecologic diseases. Gut dysbiosis reshapes microbial metabolites (SCFAs, bile acids, indoles, and TMAO) and may increase intestinal permeability, thereby elevating systemic exposure to LPS. These signals interact with host receptors and metabolic hubs (GPR41/43/109A, FXR, TGR5, and AhR) and converge on AMPK–PGC-1α and sirtuin pathways to regulate mitochondrial biogenesis, metabolic adaptation, and redox balance. LPS–TLR4–NF-κB signaling provides inflammatory priming, whereas mitochondrial stress-associated signals (mROS and mtDNA release) may further amplify innate immune activation through NLRP3 inflammasome signaling (caspase-1–mediated GSDMD pyroptosis) and cGAS–STING–type I IFN pathways. Mitochondrial quality control mechanisms, including antioxidant defense systems, mitochondrial dynamics, and PINK1/Parkin-mediated mitophagy, may function as regulatory “brakes”, while impairment of these processes may reinforce a positive feedback loop between inflammation and mitochondrial damage. An endocrine coupling loop links estrobolome-dependent estrogen recycling with the HPG axis and mitochondrial steroidogenesis, thereby contributing to disease-specific phenotypes in PCOS, endometriosis, POI, and gynecologic malignancies. Solid arrows indicate mechanisms supported by relatively consistent human and experimental evidence, whereas dashed arrows represent pathways that are predominantly supported by preclinical, indirect, or emerging evidence. Inhibitory T-shaped lines indicate regulatory or suppressive effects. SCFAs, short-chain fatty acids; TMAO, trimethylamine N-oxide; LPS, lipopolysaccharide; mROS, mitochondrial reactive oxygen species; mtDNA, mitochondrial DNA; AMPK, AMP-activated protein kinase; PGC-1 α , peroxisome proliferator-activated receptor-gamma coactivator-1 alpha; HPG, hypothalamic–pituitary–gonadal; PCOS, polycystic ovary syndrome; POI, premature ovarian insufficiency; TME, tumor microenvironment.

Building on the current understanding summarized above, future research should focus on advancing clinical translation and therapeutic strategies in this field. In the future, research in this field can focus on analyzing the precise molecular mechanisms of the gut microbiota–mitochondria axis interaction, and promote related research from correlative description to causal exploration and high-resolution spatial localization studies. At present, most microbiota research in gynecological diseases is based on correlative analyses, such as changes in fecal microbial taxonomy, and it is necessary to explore causality with the help of advanced biological models. In the future, we can transplant patients' fecal microbiota into germ-free (GF) mouse models and observe whether mitochondrial dysfunction and disease phenotype are induced after colonization ( 86 ). At the same time, the construction of an in vitro system for co-culture of 3D organoids and immune cells from gynecological diseases is also worthy of further study ( 184 ). This kind of microphysiological system can verify whether specific strains or metabolites directly influence mitochondrial damage and activate the NLRP3 inflammasome, and offering a more robust platform for mechanistic studies. Traditional bulk sequencing technologies lose key information about tissue structure and the spatial microenvironment, and gynecological diseases such as the ovarian cancer TME and endometriosis lesions have strong spatial heterogeneity ( 185 ). Therefore, future research needs to analyze the interaction between in situ immunity and mitochondria with the help of spatial transcriptomics and spatial metabolomics ( 186 ). This kind of in situ study is at the frontier of mucosal immunology, and can precisely locate the distribution of microbial DNA and metabolites in tissues, clarify their spatial interactions with local macrophages, and simultaneously capture changes in host cell mitochondrial stress genes ( 187 ). An integrated cross-scale regulatory map of “microbiota–metabolite–mitochondrial state–tissue microenvironment” can be constructed by combining spatial omics and single-cell microenvironment technologies. The core bottleneck of current research, which is also the focus of peer review, is that the heterogeneity of methodologies is too high, leading to an inability to replicate research results. Factors such as diet, body mass index, and recent medication history can interfere with bioinformatics analysis results. Therefore, it is necessary to formulate and strictly implement standardized research norms in this field as soon as possible ( 188 ). In the future, a field consensus needs to clarify and unify operating procedures, standardize DNA extraction kits and analysis pipelines, and reduce batch effects ( 189 ). At the same time, it is necessary to clearly distinguish between fecal-resident microbiota and tissue-resident mucosal microbiota, as their immunomodulatory effects on mitochondria are significantly different ( 190 ). Adopting causal inference algorithms, controlling for multi-cohort confounding factors, and pre-registering study protocols could enhance the rigor of microbiota-immune research and facilitate the standardized development of the field ( 191 ).

The interaction between gut microbiota and mitochondria is not unidirectional conduction, but rather forms a two-way coupling system through metabolites, receptor signals, and immune regulation networks ( 21 , 22 ). On the one hand, intestinal SCFAs, secondary BAs, indole compounds and other metabolites can regulate mitochondrial energy metabolism, oxidative stress and apoptosis ( 23 , 24 ); On the other hand, excessive production of ROS, membrane potential depolarization and mtDNA cytoplasmic leakage caused by abnormal mitochondrial function will destroy epithelial hypoxia and change the mucosal immune state, and then reshape intestinal microbial composition ( 9 , 11 ). This mutually regulated “gut microbiota–mitochondria axis” is a core pathological hub of metabolic disorder and chronic inflammation in gynecological diseases ( 1 , 25 ). This section outlines the biological basis of gut microbiota–mitochondria interactions from three perspectives: microbial metabolites in host systemic signaling, microbiota-driven immune activation and amplification within the tissue microenvironment, and mitochondrial function and homeostasis. Intestinal microbial metabolites can regulate energy metabolism pathways, receptor signaling, oxidative stress reactions, and autophagy processes, promote mitochondrial biogenesis, maintain mitochondrial functional homeostasis, and achieve effective communication between exogenous nutritional signals and endocrine regulation ( 26 , 27 ). SCFAs are the main energy source for colonic epithelial cells; they can participate in the tricarboxylic acid cycle, activate intracellular energy-sensing pathways, improve energy utilization efficiency and insulin sensitivity, and promote mitochondrial biogenesis and fatty acid oxidation ( 28 ). At the same time, SCFAs can act on G protein-coupled receptors such as GPR41, GPR43, and GPR109A, and regulate intestinal hormone secretion and systemic metabolic rhythm ( 29 ). Similarly, the secondary BAs produced by gut microbiota can act as ligands for FXR and TGR5 receptors, regulating mitochondrial membrane potential, ROS generation, and systemic energy homeostasis ( 30 ). In addition to metabolic regulation, these receptor-mediated signals also affect estrogen synthesis and ovarian function. Gut-derived microbial enzymes, such as β-glucuronidase, regulate the ratio of active to inactive estrogens in the circulation through the estrobolome ( 8 , 31 ), and then affect the mitochondrial respiratory capacity and antioxidant defense function of the reproductive system, suggesting the relationship between intestinal microbial activity and mitochondrial energy metabolism ( 18 ). Tryptophan metabolites, especially indole derivatives, can regulate the activity and antioxidant capacity of respiratory chain complexes, and the kynurenine pathway can also act as a metabolic regulator to modulate mitophagy and biogenesis ( 24 , 32 ). To summarize, existing studies have supported that microbial metabolites can indirectly transmit exogenous nutritional signals to the endocrine network, while maintaining the stability of mitochondrial quality and function. In addition to metabolic regulation, an immune amplification loop can explain how gut microbiota and mitochondria jointly regulate chronic inflammation in the tissue microenvironment. After endotoxin translocation caused by intestinal barrier damage, pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS) will activate innate immune receptors ( 33 ), providing a key initiating signal for the activation of the NLRP3 inflammasome ( 34 , 35 ). At the same time, ROS and mtDNA released from damaged mitochondria, as endogenous damage-associated molecular patterns (DAMPs), cannot be effectively eliminated; instead, they promote the assembly of the NLRP3 inflammasome and the release of pro-inflammatory factors ( 36 , 37 ), further amplifying the initial inflammatory response induced by microbiota ( 38 ). Exogenous PAMPs and endogenous DAMPs work together to form a continuous excessive inflammatory cycle ( 39 ). In endometriosis and other diseases, this kind of synergistic immune cascade reaction is considered an important contributing factor that triggers a vicious circle of pain, local inflammation, and fibrosis ( 17 , 40 , 41 ). Mitochondrial quality control is key to avoiding excessive inflammation and metabolic disorder ( 42 ). Stress-induced electron leakage and inflammatory activation of the electron transport chain will first increase the production of ROS, and beneficial metabolites such as SCFAs can alleviate this problem by strengthening the antioxidant system ( 43 ). However, excessive accumulation of ROS beyond the scavenging capacity will accelerate lipid peroxidation and eventually lead to the collapse of the mitochondrial membrane potential ( 44 ). Mitochondrial dynamics also plays an important role. Under inflammatory and oxidative conditions, excessive mitochondrial fission leads to structural fragmentation, decreased ATP production, and impaired steroid production, which results in abnormal follicular development ( 45 ); Conversely, enhanced mitochondrial fusion and cristae remodeling help maintain functional stability ( 46 ). Finally, mitophagy can specifically remove damaged mitochondria, prevent mtDNA and ROS from being released into the cytoplasm, and block the vicious circle caused by NLRP3 inflammasome activation ( 47 ). Damage to this clearance pathway will not only aggravate the inflammatory reaction but also lead to defects in tissue repair and drug resistance ( 48 ). To sum up, the interaction between microbiota and mitochondria is reflected in three interrelated levels: systemic metabolic mediation, microenvironmental immune amplification, and mitochondrial homeostasis maintenance, which together constitute a “metabolism-immunity-endocrine” regulatory network. The intracellular signaling pathways and molecular targets that play regulatory roles will be elaborated in detail later.

In summary, this study puts forward a new closed-loop model of “microbiota-immunoamplification-mitochondria-endocrine”, which changes the research paradigm of gynecological pathophysiology from a single-organ perspective to a systemic multi-system interactive perspective. It clarifies that gut microbiota dysbiosis is considered a key contributing factor, and that immune cells, especially macrophages, act as amplification mediators to induce mitochondrial dysfunction, which in turn leads to endocrine disorders and promotes the occurrence and development of diseases such as endometriosis, polycystic ovary syndrome, and gynecological malignancies. Interventions targeting this regulator y axis, such as probiotics and mitochondria-targeted drugs, have good translational prospects, but the first task at present is to rigorously elucidate the relevant pathway mechanisms through spatial multi-omics and standardized causal experimental models. Analyzing the complex interaction between microbiota and mitochondria can provide new theoretical support for the research and development of individualized and precise diagnostic and therapeutic technologies for female reproductive system diseases.

Based on the framework of “metabolic input–immune amplification–oxidative stress/mitophagy–endocrine regulation” schematically summarized in Figure 5 , intervention strategies targeting the microbiota are fundamental means to regulate the gut-ovarian axis, and the core translational goals are to restore microbial diversity, correct metabolite imbalance, and reduce the systemic endotoxin load. Integrated microbiota–mitochondria co-targeting strategy for gynecologic diseases. Module 1 outlines microbiota-directed interventions (diet, probiotics/prebiotics, and FMT) together with mitochondria-targeted approaches, including antioxidant, autophagy/mitophagy, and mitochondrial dynamics–related strategies. Module 2 depicts the restoration of microbial homeostasis (eubiosis) and metabolite outputs (SCFAs, bile acids, and indoles) alongside a reduced endotoxin (LPS) burden. Module 3 highlights key sensing and signaling pathways, including the inhibition of LPS–TLR4–NF-κB signaling, activation of FXR/TGR5 and AhR, and estrobolome-dependent estrogen recycling. Module 4 integrates these signals at the mitochondrial hub, demonstrating how AMPK–PGC-1α–associated mitochondrial biogenesis, stable mitochondrial homeostasis, and PINK1/Parkin-mediated mitophagy act collectively to limit mROS and mtDNA release, thereby dampening NLRP3 inflammasome activity. Module 5 summarizes the potential improvements in metabolic, endocrine, inflammatory, and therapeutic-response phenotypes across PCOS, endometriosis, POI, and gynecologic malignancies. Solid arrows indicate signaling or activation pathways, blue T-shaped lines indicate inhibitory effects, and dashed arrows represent intervention-associated modulation and delivery. FMT, fecal microbiota transplantation; FOS, fructooligosaccharides; GOS, galactooligosaccharides; SCFAs, short-chain fatty acids; FXR, farnesoid X receptor; TGR5, Takeda G protein-coupled receptor 5; AhR, aryl hydrocarbon receptor; LC3, microtubule-associated protein 1A/1B-light chain 3; mROS, mitochondrial reactive oxygen species; mtDNA, mitochondrial DNA; PCOS, polycystic ovary syndrome; POI, premature ovarian insufficiency. Probiotics, such as Lactobacillus spp., can regulate the function of the gut-ovarian axis by competitively inhibiting pathogenic bacteria and strengthening the integrity of the mucosal barrier ( 163 ). Preclinical studies have shown that this type of intervention can regulate the LPS–TLR4–NF-κB axis, reduce NLRP3 inflammasome activation, and support mitochondrial biogenesis ( 164 ). Prebiotics such as fructooligosaccharides and inulin can selectively promote the proliferation of SCFA-producing taxa ( 165 ). SCFAs can enhance antioxidant defense, promote PINK1/Parkin-mediated mitophagy, and reduce mtDNA cytoplasmic release ( 166 ). However, the intervention effects of probiotics and prebiotics are greatly influenced by strain specificity, host gut microbiota composition, and dosage. At present, the heterogeneity of clinical research is high, and the evidence is insufficient. It is necessary to carry out high-quality longitudinal clinical trials and clarify the applicable population and optimal preparation regimen to achieve clinical translation and application. FMT is a technique to quickly restore intestinal ecological balance and metabolic homeostasis by transplanting complex microbial communities ( 167 ). Theoretically, FMT can reconstruct the metabolic profile of SCFAs and BAs, relieve systemic inflammation, and reset the gut microbiota–mitochondria axis ( 168 ). Animal models and early population-based exploratory studies show that this technology can improve metabolic and inflammatory indices, but it is still in the experimental stage in the field of gynecological diseases ( 169 ). The standardization of donor screening, long-term safety, the risk of drug-resistant gene transfer, and the durability of the transplantation effect have not yet been resolved ( 170 ). Therefore, FMT can only be used in strictly monitored clinical trials or specific compassionate use scenarios that meet ethical and regulatory requirements. Oxidative stress and impaired mitochondrial quality control are central links between gut microbiota dysbiosis and endocrine disorders, so therapeutic strategies targeting mitochondrial homeostasis have important research value, with the core being to alleviate oxidative damage and restore the dynamic balance of mitochondrial biogenesis and renewal. MitoQ and other targeted mitochondrial antioxidants can selectively enrich in the mitochondrial matrix, scavenge mitochondrial ROS and maintain membrane potential. Preclinical studies of ovarian cancer and PCOS models show that MitoQ can alleviate lipid peroxidation and improve insulin signaling or chemotherapy-induced apoptosis ( 171 , 172 ). Metabolic modulators such as NAD+ precursors can improve the efficiency of the tricarboxylic acid cycle, strengthen oxidative phosphorylation, and repair energy metabolism disorders ( 173 , 174 ). However, the translational value of these drugs in the field of gynecology still needs to be verified, and the optimal dosage, tissue-specific delivery mode, and interaction with standard chemotherapy regimens need to be clarified before clinical application. Repairing mitochondrial quality control function, especially mitophagy and dynamic regulation, can reduce the release of DAMPs such as mtDNA, and is an important adjunctive treatment direction. Drugs such as urolithin A and spermidine have been shown in preclinical studies to promote PINK1/Parkin-mediated mitophagy, inhibit NLRP3 inflammasome assembly, and restore cellular homeostasis ( 90 , 175 ). Regulating mitochondrial dynamics, for example by inhibiting excessive fission via Drp1, can promote mitochondrial fusion and improve energy metabolism efficiency in tumor cells or granulosa cells ( 95 , 176 ). However, this type of drug is still in the research and development stage, and problems such as difficulty in systemic delivery and risk of off-target effects need to be handled cautiously. The pathological cycle of gynecological diseases is complex, and it is difficult to completely block it with a single intervention method. Therefore, as schematically summarized in Figure 5 , an integrated “dual-targeting” framework combining upstream microbiota regulation with downstream mitochondrial repair may represent a promising translational approach. The core logic of this integrated approach is to block the innate immune amplification reaction at two levels: upstream microecological intervention (prebiotics, probiotics, microbiota-modulating diet) can optimize metabolic input, reduce systemic endotoxemia, and reduce the exogenous trigger signals for cell surface receptors such as TLR4 ( 177 ); Downstream targeted mitochondrial intervention can scavenge mitochondrial ROS, enhance PINK1/Parkin-mediated mitophagy, and prevent the release of DAMPs such as mtDNA ( 178 ). From a mechanistic point of view, this dual blockade can break the activation mode of the NLRP3 inflammasome: upstream intervention weakens the LPS-induced NF-κB transcriptional initiation signal, and downstream intervention inhibits the activation signal caused by mitochondrial oxidative stress and structural collapse ( 179 ). Applying this theoretical framework to translational research requires multi-dimensional evaluation rather than empirical treatment. Follow-up preclinical models and population-based exploratory research need to simultaneously monitor changes in intestinal microecology, systemic immune indicators, and easily accessible mitochondrial function indicators in peripheral blood mononuclear cells ( 180 ). The integrated model provides a hypothetical basis for follow-up studies and facilitates the understanding of the interactions among gut microbiota dysbiosis, cellular energy metabolism, and endocrine lesions. The “microbiota-mitochondrial dual targeting framework” provides a feasible idea for blocking innate immune amplification reactions, but individual heterogeneity should be fully considered in clinical translation. Factors such as body mass index, insulin resistance, diet, region, and medication history of patients with gynecological diseases will affect the basal microecology and metabolic homeostasis ( 181 ), which is also the reason for the inconsistency of microbial signals in existing cohorts, and it also shows that a single fixed intervention program cannot be applied to all patients. Therefore, the clinical translation of this framework needs to implement individualized and precise stratification strategies. There are differences in the basal microbial communities among hosts, and their responses to microecological regulators such as probiotics and prebiotics and to mitochondrial-targeted drugs show obvious individual differences ( 182 ). To enhance the translational value of the dual-targeting strategy, future clinical trials need to shift from single biomarker monitoring to multi-omics-driven mechanistic phenotyping. For example, through integrative omics analysis, patients could be stratified into an endotoxin/inflammation-driven type and a mitochondrial intrinsic damage type, and matched with microbiota- or mitochondria-targeted intervention programs ( 183 ). Acknowledging and systematically studying individual heterogeneity will not reduce the mechanistic value of the framework, but will ensure that subsequent translational research is carried out precisely, safely, and individually.

At the cellular level, intestinal metabolites mainly regulate mitochondrial function through the AMPK/PGC-1α signaling hub. After SCFAs (especially butyrate) enter cells, the intracellular AMP/ATP ratio is altered, and then AMPK is phosphorylated. In vitro and in vivo studies have supported that activated AMPK can directly phosphorylate PGC-1α, and at the same time activate SIRT1, which deacetylates PGC-1α ( 49 ). This synergistic post-translational modification promotes mitochondrial biogenesis, strengthens fatty acid β-oxidation, and regulates cellular energy metabolism ( 50 ). Therefore, the SCFAs/AMPK/PGC-1α signaling cascade is a core molecular bridge that alleviates lipotoxicity and an important defense mechanism against metabolic disorders ( 51 ). In the same way, secondary BAs transform microbial signals into metabolic adaptation reactions through signaling pathways mediated by specific receptors. BAs bind to the membrane G protein-coupled receptor TGR5, which increases intracellular cAMP levels and further promotes AMPK activation ( 52 ). At the same time, BAs activate the nuclear receptor FXR, regulating lipid metabolism and maintaining the transcription of related target genes ( 53 , 54 ). Experimental models show that abnormal regulation of the BA-receptor axis leads to mitochondrial energy metabolism disorder, which is also a common molecular phenotype in PCOS and intrahepatic cholestasis of pregnancy ( 55 , 56 ). Tryptophan metabolism is another important intracellular regulatory pathway. Microbial indole derivatives can up-regulate the activity of the respiratory chain complex ( 57 ), and the kynurenine pathway is a double-edged sword for metabolic regulation. Mechanistic studies show that physiologically relevant concentrations of kynurenine metabolites can maintain mitochondrial number via the AMPK/SIRT1 axis ( 58 , 59 ), but excessive accumulation of kynurenine disrupts the redox balance and inhibits mitophagy ( 60 ). Cohort studies of PCOS and endometriosis patients found that increased kynurenine concentration was positively correlated with disease severity, and that it exacerbates mitochondrial damage through a pathological positive feedback loop ( 61 ). To summarize, various metabolic signals converge on the AMPK/PGC-1 α and SIRT1 regulatory networks, transforming intestinal microbial changes into long-term transcriptional programs and regulating mitochondrial number and cellular homeostasis. The synergistic effect of intestinal microbial signals and mitochondrial stress is mainly realized through the convergence of innate immune cascade pathways, and the core follows a “double-hit” activation mode. The first step is exogenous signal transduction. After intestinal barrier damage, intestinal LPS enters the circulation ( 62 ). As a typical PAMP, it binds to TLR4 on the surface of macrophages, triggering NF-κB nuclear translocation, providing a key initiating signal, and greatly up-regulating the transcription of NLRP3 inflammasome components and pro-inflammatory cytokine precursors ( 63 , 64 ). Subsequently, mitochondrial dysfunction releases endogenous DAMPs, providing key activation signals. Under oxidative stress, the integrity of the mitochondrial membrane is damaged, leading to the leakage of mtDNA and ROS into the cytoplasm ( 65 ). In vitro experiments with macrophages have shown that free mtDNA in the cytoplasm can be directly recognized by cGAS, which activates the cGAS–STING pathway, induces phosphorylation of IRF3 and NF-κB, and promotes the expression of type I interferons and broad-spectrum inflammatory mediators ( 66 ). At the same time, oxidized mtDNA together with mitochondrial ROS can bind to and activate the NLRP3 inflammasome protein, triggering its oligomerization to form an active inflammasome, activating caspase-1, cleaving mature IL-1β and IL-18, and finally causing pyroptosis ( 47 , 67 ), driving the inflammatory amplification cascade illustrated in Figure 2 . Amplification valve: PAMP–DAMP synergy drives innate immune escalation relevant to gynecologic inflammation. Panel A illustrates gut barrier leak–associated PAMP priming. Increased intestinal permeability may elevate systemic LPS exposure, thereby activating TLR4–NF-κB signaling and promoting transcriptional priming of inflammasome-associated components, including NLRP3 and pro-IL-1β/pro-IL-18. Panel B depicts mitochondrial stress–associated DAMP release. Increased mROS and reduced mitochondrial membrane potential (↓ΔΨm) may promote cytosolic mtDNA release, which can contribute to NLRP3 inflammasome assembly and, in parallel, engage cGAS–STING signaling to induce type I interferons and inflammatory cytokines. Inflammasome activation promotes caspase-1–dependent IL-1β/IL-18 maturation and GSDMD cleavage, leading to pore formation and pyroptosis. Pyroptosis and inflammatory amplification may further aggravate barrier dysfunction and mitochondrial injury, thereby promoting a self-reinforcing inflammatory loop relevant to chronic pelvic inflammation and tissue remodeling. Mitochondrial quality-control (“QC”) mechanisms, including PINK1/Parkin-mediated mitophagy, antioxidant defenses, and balanced fusion–fission dynamics, may function as regulatory “brakes” that limit mtDNA/mROS release and dampen this amplification process. Solid arrows denote well-established mechanisms, whereas dashed arrows indicate emerging, indirect, or predominantly preclinical pathways. T-shaped lines represent inhibition. PAMP, pathogen-associated molecular pattern; DAMP, damage-associated molecular pattern; LPS, lipopolysaccharide; mROS, mitochondrial reactive oxygen species; ΔΨm, mitochondrial membrane potential; mtDNA, mitochondrial DNA; cGAS, cyclic GMP-AMP synthase; STING, stimulator of interferon genes; GSDMD, gasdermin D; QC, quality control. The molecular interaction between the TLR4-NF-κB and mtDNA-cGAS/NLRP3 inflammasome axes forms an irreversible positive feedback cycle of excessive inflammation. The continuous activation of the NLRP3 inflammasome and the sustained release of IL-1β will not only aggravate local tissue damage but also disrupt the adaptive immune response, promote macrophage polarization towards the pro-inflammatory M1 type, overactivate Th17 cells, and inhibit the function of regulatory T cells ( 68 – 70 ). This complex intracellular signaling network is the core molecular basis for the formation of a chronic inflammatory microenvironment and the disruption of immune tolerance in reproductive system diseases ( 71 ). Intestinal metabolites and innate immune signals can regulate oxidative stress, and the molecular execution mechanism of mitophagy is the ultimate key to mitochondrial quality control. ROS is not only a downstream injury marker but also an important regulatory factor in the gut microbiota–mitochondria axis regulatory system. Physiological concentrations of ROS can act as signaling molecules to initiate antioxidant defense, but continuous gut microbiota dysbiosis leads to LPS translocation, which causes the oxidative load to far exceed the cellular clearance capacity, leading to severe lipid peroxidation and collapse of the mitochondrial membrane potential (ΔΨm) ( 72 , 73 ). In vitro mechanistic studies have supported that loss of mitochondrial membrane potential is a critical trigger for the activation of the PINK1/Parkin signaling pathway ( 74 ). After mitochondrial depolarization, PINK1 can no longer be degraded by the proteasome but stably accumulates on the outer mitochondrial membrane ( 75 ). PINK1 accumulation phosphorylates ubiquitin molecules and recruits the E3 ubiquitin ligase Parkin to damaged mitochondria ( 76 ). Activated Parkin ubiquitinates outer mitochondrial membrane proteins to form recognizable ubiquitin chain signals, and then autophagy adaptor proteins bind to the ubiquitinated mitochondrial proteins and connect them with the autophagy protein LC3, thus achieving selective encapsulation of damaged mitochondria and lysosomal degradation ( 77 , 78 ). It is worth noting that the integrity of the PINK1/Parkin-mediated clearance pathway determines the cellular immunophenotype. Experimental models show that beneficial microbial metabolites such as SCFAs and hydrogen sulfide (H 2 S) can indirectly support mitophagy and maintain mitochondrial function ( 79 , 80 ); On the other hand, chronic dysbiosis-induced inflammation will inhibit or damage the PINK1/Parkin pathway, resulting in massive accumulation of damaged mitochondria and release of large amounts of mtDNA and ROS, which may further promote cGAS–STING and NLRP3 inflammasome activation, as illustrated in Figure 2 ( 81 ). Therefore, the PINK1/Parkin axis serves as a key molecular checkpoint that may influence whether transient microbiota-induced stress is resolved or progresses into persistent and potentially irreversible inflammatory lesions ( 82 ).