Single cell transcriptomic analysis reveals pathogenic cell heterogeneity and candidate inflammatory-associated markers in STZ-induced diabetic mouse retina

Background Diabetic retinopathy (DR) is driven by chronic hyperglycemia and involves coordinated vascular, inflammatory, and neuroglial dysfunction. Müller glia are central to retinal homeostasis, yet their cell-state heterogeneity and inflammatory response programs in DR mice remain incompletely characterized at single-cell resolution. Methods We reanalyzed a public scRNA-seq dataset of STZ-induced diabetic mouse retinas to characterize retinal cell populations and Müller glial states through clustering, perturbation, trajectory, functional enrichment, and co-expression network analyses, with selected targets further validated by Western blot. Results Integration and clustering of the scRNA-seq dataset identified the major retinal cell types as well as four distinct Müller glial subpopulations. Among annotated retinal cell populations, Müller glia showed the strongest transcriptional perturbation in the STZ group, indicating that they are among the most transcriptionally responsive retinal cell types under diabetic stress. Pseudotime analysis supported the presence of branch-dependent transcriptional programs among Müller subclusters and suggested that STZ conditions were associated with preferential progression toward a Müller substate enriched for photoreceptor-associated transcripts. Functional enrichment analysis showed that different Müller glial subclusters were associated with distinct biological processes, while sharing activation of inflammatory-response programs, and highlighted Cebpb as a candidate inflammation-associated factor. Co-expression network analysis further identified Müller glia-associated gene modules with subcluster- and condition-dependent activity patterns, including modules linked to photoreceptor-associated programs, stress responses, angiogenesis and inflammatory signaling. Protein-protein interaction analysis prioritized Junb as a highly connected candidate regulatory hub, and western blotting provided supportive tissue-level evidence for altered JUNB and CEBPB protein abundance in STZ versus control retinas. Conclusion We delineated the transcriptional heterogeneity of Müller glia and identified candidate state-associated modules and regulators linked to diabetic retinal stress responses. These findings support an active role for Müller glia in diabetic retinal remodeling through inflammatory, structural, and neuron-interactive programs, and provide a basis for future mechanistic and translational investigation of Müller glia-mediated pathology in diabetic retinopathy.