Single-cell transcriptome analysis of the Akimba mouse retina reveals cell-type-specific insights into the pathobiology of diabetic retinopathy

Diabetic retinopathy is a common complication of diabetes and a leading cause of visual impairment and blindness. Despite recent advances, our understanding of its pathophysiology remains incomplete. The aim of this study was to provide deeper insight into the complex network of molecular and cellular changes that underlie diabetic retinopathy by systematically mapping the transcriptional changes that occur in the different cellular compartments of the degenerating diabetic mouse retina. Single-cell RNA sequencing was performed on retinal tissue from 12-week-old wild-type and Akimba (Ins2Akita×Vegfa+/–) mice, which are known to replicate features of clinical diabetic retinopathy. This resulted in transcriptome data for 9474 retinal cells, which could be annotated to eight distinct retinal cell types. Using STRING analysis, we studied differentially expressed gene networks in neuronal, glial and immune cell compartments to create a comprehensive view on the pathological changes that occur in the Akimba retina. Using subclustering analysis, we further characterised macroglial and inflammatory cell subpopulations. Prominent findings were confirmed at the protein level using immunohistochemistry, western blotting and ELISA. At 12 weeks, the Akimba retina was found to display degeneration of rod photoreceptors and presence of inflammatory cells, identified by subclustering analysis as monocyte, macrophage and microglial populations. Analysis of differentially expressed genes in the rod, cone, bipolar cell and macroglial compartments indicated changes in cell metabolism and ribosomal gene expression, gliosis, activation of immune system pathways and redox and metal ion dyshomeostasis. Experiments at the protein level supported a metabolic shift from glycolysis to oxidative phosphorylation (glyceraldehyde 3-phosphate dehydrogenase), activation of microglia/macrophages (isolectin-B4), metal ion and oxidative stress response (metallothionein and haem oxygenase-1) and reactive macroglia (glial fibrillary acidic protein and S100) in the Akimba retina, compared with wild-type mice. Our single-cell approach also indicates macroglial subpopulations with distinct fibrotic, inflammatory and gliotic profiles. Our study identifies molecular pathways underlying inflammatory, metabolic and oxidative stress-mediated changes in the Akimba mouse model of diabetic retinopathy and distinguishes distinct functional subtypes of inflammatory and macroglial cells. RNA-seq data have been deposited in the ArrayExpress database at EMBL-EBI ( www.ebi.ac.uk/arrayexpress ) under accession number E-MTAB-9061.