The gut-eye axis in blinding eye diseases: microbiota-driven immune dysregulation and immunomodulatory therapies

To synthesize recent (2020-2025) advances on how gut, oral, and ocular-surface microbiota contribute to major blinding eye diseases, dry eye disease (DED), non-infectious uveitis, glaucoma, optic neuropathy, age-related macular degeneration (AMD), and diabetic retinopathy (DR), and to evaluate the therapeutic potential of microbiome-based interventions.

PubMed and Web of Science were searched (January 2020-October 2025) using the terms "gut microbiota", "ocular diseases", and "immunomodulatory therapies". Eligible studies included original human and animal research demonstrating microbial dysbiosis or testing microbiome-directed therapies. Data were synthesized thematically across microbial composition, immune-metabolic mechanisms, and intervention outcomes.

Across all six diseases, dysbiosis was consistently characterized by depletion of anti-inflammatory taxa such as Akkermansia, Ruminococcaceae, and other short-chain fatty acid (SCFA) producers, with enrichment of pro-inflammatory bacteria including Proteobacteria, Staphylococcus, and Porphyromonas gingivalis. These changes were associated with increased intestinal permeability, systemic lipopolysaccharide (LPS) and trimethylamine N-oxide (TMAO), Th17 (T helper 17)/Treg (regulatory T cell) imbalance, and loss of SCFA-mediated neuroprotection. Probiotics containing Lactobacillus or Bifidobacterium improved tear stability and reduced inflammation in preclinical and pilot clinical studies, while high-fiber diets ameliorated lesions in age-related macular degeneration (AMD) and diabetic retinopathy (DR). Fecal microbiota transplantation confirmed microbial causality but revealed donor-dependent effects, and engineered Lactobacillus expressing angiotensin-converting enzyme 2 (ACE2) or Ang-(1-7) preserved retinal integrity in diabetic models.

Microbial dysbiosis acts as a common driver of immune-metabolic dysfunction in blinding eye diseases. Microbiome-targeted strategies show promising efficacy in experimental systems, but large, longitudinal human trials are needed for clinical translation.