From background diabetic retinopathy to its proliferative stage. What is the role of gut microbiota in the trajectory of DR? a Mendelian randomization study with mediation analysis

Abstract Background Growing evidence suggests that gut microbiota (GM) plays a role in diabetic retinopathy (DR), but the causal microbial drivers and their stage-dependent roles during DR progression remain poorly characterised. Using genetic causality methods, we aim to depict a longitudinal GM ma...

Full description

Saved in:
Bibliographic Details
Main Authors: Yifan Zhou, Jialong Dong, Zhenyu Wang, Chen Huang, Xiaotong Yu, Xinjun Wang, Bo Yang, Yuechen Wu, Qing Peng
Format: Article
Language:English
Published: BMC 2025-08-01
Series:Diabetology & Metabolic Syndrome
Subjects:
Online Access:https://doi.org/10.1186/s13098-025-01813-6
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Background Growing evidence suggests that gut microbiota (GM) plays a role in diabetic retinopathy (DR), but the causal microbial drivers and their stage-dependent roles during DR progression remain poorly characterised. Using genetic causality methods, we aim to depict a longitudinal GM mapping and stage-stratified GM signatures across the DR trajectory, spanning initial background DR (BDR) through non-proliferative form (NPDR), to advanced proliferative stage (PDR). Methods GWAS data of 207 GM taxa (from phylum to species) were acquired from the Dutch Microbiome Project (N = 7,824), and DR from FinnGen (over 300,000 individuals). A bidirectional two-sample Mendelian Randomization (TSMR) analysis was conducted to elucidate directional causality between GM and DR. Multiple sensitivity evaluations were performed for pleiotropy, heterogeneity, and stability. Additionally, two-step MR and multivariable MR (MVMR) were performed to dissect causal GM-DR relationships using 1400 candidate circulating metabolite level/ratio data from a Canadian cohort (N = 8,299). Results We identified 11 causal GM taxa (1 family, 3 genera, and 7 species) during the progression of DR. Notably, species_Bacteroides_dorei and species_Dorea_longicatena demonstrated pan-stage pathogenicity (BDR and PDR, all OR>1, P IVW <0.05), while family_Clostridiaceae (OR = 1.540, 95%CI: 1.110–2.135), genus_Clostridium (OR = 1.473, 95%CI: 1.086–1.997), and species_Eubacterium_ramulus (OR = 1.382, 95%CI: 1.000-1.911) specifically promoted NPDR. Six causal protective GM taxa, comprising four species, Bifidobacterium_longum (OR = 0.540, 95%CI: 0.370–0.788), Bacteroides_stercoris (OR = 0.633, 95%CI: 0.407–0.986), Ruminococcus_torques (OR = 0.771, 95%CI: 0.627–0.948), Roseburia_hominis (OR = 0.652, 95%CI: 0.501–0.849), and two genera, Escherichia (OR = 0.799, 95%CI: 0.638–0.999) and Flavonifractor (OR = 0.825, 95%CI: 0.685–0.993), mitigate NPDR or PDR risks. Among the 76, 90, and 86 causal DR stage-specific metabolites, the mannose-to-hydroxyproline ratio was the only metabolite universally linked to all DR stages, with the other 31 metabolites influencing dual phases. Mediation analysis validated five metabolites (asparagine, cystine, gamma-glutamylglycine, ximenoylcarnitine (C26:1) levels, and androsterone glucuronide to etiocholanolone glucuronide ratio) as key mediators bridging causal GM-DR links, with considerable mediating proportions of 6.75%, 14.90%, 3.03%, 7.25%, and 11.97%, respectively. Conclusion This study pioneers causal longitudinal mapping of GM dynamics across DR progression through integrated genetic prediction and metabolomic mediation analyses, delineating stage-specific microbial drivers, pan-stage pathogens, and metabolite mediators that collectively orchestrate DR pathogenesis. The identified GM-metabolite-DR axis establishes an actionable roadmap for targeted microbiome modulation and metabolite-based therapeutic strategies, bridging observational associations to mechanistic intervention opportunities.
ISSN:1758-5996