Causal relationship between gut microbiota and adenomyosis: metagenomics sequencing and Mendelian randomization
Abstract
<h4>Background</h4>Emerging evidence implicates the gut microbiota in the pathogenesis of adenomyosis (AM); however, whether this association is causal and through which mechanisms it operates remain largely unknown.<h4>Methods</h4>To interrogate potential causal relationships, we performed a two-sample Mendelian randomization (MR) analysis leveraging inverse-variance weighting (IVW) as the primary estimator, complemented by MR-Egger, weighted median, and weighted mode approaches, to evaluate the causal effects of gut microbial taxa and microbiota-derived metabolic pathways on AM. We further conducted mediation analyzes to delineate the role of circulating immune-cell phenotypes in this process. In parallel, in an independent clinical cohort, 22 patients with AM and 23 age-matched healthy controls recruited from the health-screening center of our institution were enrolled according to stringent inclusion and exclusion criteria (including antibiotic-use history and long-term local residency) and subjected to shotgun metagenomic sequencing. Significant differences in the types of bacterial communities were observed between the AM group and the control group. Subsequently, the results were cross-compared with those of the MR study using the Linear Discriminant Analysis Effect Size (LEfSe) method, and further verified using the ANCOM-BC method to determine the common microbial characteristics.<h4>Results</h4>MR analysis identified ten microbial taxa and ten metabolic pathways with evidence of potential causal associations with AM. Of these, nine taxa and five pathways were associated with a reduced risk of AM, including <i>Alistipes indistinctus</i> (OR = 0.847, 95% CI = 0.754-0.951, <i>p</i> = 0.005, <i>p</i>~FDR~ > 0.05), <i>Ruminococcus torques</i> (OR = 0.818, 95% CI = 0.712-0.941, <i>p</i> = 0.005, <i>p</i>~FDR~ > 0.05), class <i>Deltaproteobacteria</i> (OR = 0.780, 95% CI = 0.629-0.967, <i>p</i> = 0.024, <i>p</i>~FDR~ > 0.05), family <i>Desulfovibrionaceae</i> (OR = 0.780, 95% CI = 0.629-0.967, <i>p</i> = 0.024, <i>p</i>~FDR~ > 0.05), order <i>Desulfovibrionales</i> (OR = 0.780, 95% CI = 0.629-0.967, <i>p</i> = 0.024, <i>p</i>~FDR~ > 0.05), <i>Parasutterella excrementihominis</i> (OR = 0.875, 95% CI = 0.784-0.977, <i>p</i> = 0.017, <i>p</i>~FDR~ > 0.05), <i>Ruminococcus bromii</i> (OR = 0.836, 95% CI = 0.718-0.972, <i>p</i> = 0.020, <i>p</i>~FDR~ > 0.05), <i>Bacteroides finegoldii</i> (OR = 0.919, 95% CI = 0.855-0.987, <i>p</i> = 0.020, <i>p</i>~FDR~ > 0.05), and the genus <i>Parasutterella</i> (OR = 0.886, 95% CI = 0.797-0.986, <i>p</i> = 0.026, <i>p</i>~FDR~ > 0.05); the five protective pathways comprised dTDP-L-rhamnose biosynthesis (OR = 0.819, 95% CI = 0.674-0.995, <i>p</i> = 0.045, <i>p</i>~FDR~ > 0.05), lactose and galactose degradation (OR = 0.818, 95% CI = 0.689-0.972, <i>p</i> = 0.022, <i>p</i>~FDR~ > 0.05), the reductive TCA cycle (OR = 0.919, 95% CI = 0.851-0.993, <i>p</i> = 0.032, <i>p</i>~FDR~ > 0.05), allantoin degradation to glyoxylate (OR = 0.907, 95% CI = 0.830-0.991, <i>p</i> = 0.030, <i>p</i>~FDR~ > 0.05), and glycolysis I (from glucose-6-phosphate) (OR = 0.850, 95% CI = 0.747-0.967, <i>p</i> = 0.013, <i>p</i>~FDR~ > 0.05).Conversely, one taxon and five pathways were associated with an increased risk of AM: the genus <i>Lactobacillus</i> (OR = 1.083, 95% CI = 1.008-1.164, <i>p</i> = 0.030, <i>p</i>~FDR~ > 0.05), degradation of glucose and glucose-1-phosphate (OR = 1.202, 95% CI = 1.056-1.369, <i>p</i> = 0.005, <i>p</i>~FDR~ > 0.05), peptidoglycan biosynthesis (in Enterococcus faecium) (OR = 1.138, 95% CI = 1.007-1.285, <i>p</i> = 0.039, <i>p</i>~FDR~ > 0.05), pyruvate fermentation to acetone (OR = 1.118, 95% CI = 1.001-1.248, <i>p</i> = 0.048, <i>p</i>~FDR~ > 0.05), glycerol degradation to butanol (OR = 1.118, 95% CI = 1.011-1.237, <i>p</i> = 0.031, <i>p</i>~FDR~ > 0.05), and <i>de novo</i> pyrimidine deoxyribonucleotide biosynthesis (OR = 1.216, 95% CI = 1.063-1.390, <i>p</i> = 0.004, <i>p</i>~FDR~ > 0.05).Mediation analysis revealed that the immune phenotype "CD24 on CD24<sup>+</sup>CD27<sup>+</sup> B cells" mediated the pathway from <i>Ruminococcus bromii</i> to AM, accounting for 32.91% of the total effect (<i>p</i> = 0.020).Shotgun metagenomic profiling of the clinical cohort demonstrated no significant differences in α-diversity or β-diversity between the AM and control groups. At the phylum level, the relative abundance of Desulfobacterota was significantly decreased in the AM group (<i>p</i>< 0.05), and at the genus level, <i>Alistipes</i> was similarly reduced (<i>p</i>< 0.05). LEfSe analysis further indicated enrichment of <i>Escherichia</i> and <i>Clostridium</i> in the AM group, whereas <i>Desulfobacterota</i> and <i>Rikenellaceae</i> were enriched in the Control group. Matching the aforementioned results with the Mendelian randomization (MR) outcomes revealed that <i>Desulfovibrionales</i> and <i>Desulfovibrionaceae</i> constituted the shared microbial taxa. This finding was subsequently re-validated and confirmed using the ANCOM-BC method.<h4>Conclusions</h4>Integrating genetic causal inference with clinical metagenomic validation, this study provides convergent evidence that specific gut microbial taxa, their associated metabolic pathways, and immune-cell-mediated mechanisms may be causally implicated in the development of AM. These findings offer a framework for future microbiota-targeted preventive and therapeutic strategies against AM.