Differences in junction-associated gene expression changes in three rat models of diabetic retinopathy with similar neurovascular phenotype

Diabetic retinopathy, also defined as microvascular complication of diabetes mellitus, affects the entire neuro-vascular unit with specific aberrations in every compartment. Neurodegeneration, glial activation and vaso-regression are observed consistently in models of diabetic retinopathy. However, the order and the severity of these aberrations varies in different models, which is also true in patients. In this study, we analysed rat models of diabetic retinopathy with similar phenotypes to identify key differences in the pathogenesis. For this, we focussed on intercellular junction-associated gene expression, which are important for the communication and homeostasis within the neurovascular unit. Streptozotocin-injected diabetic Wistar rats, methylglyoxal supple-mented Wistar rats and polycystin-2 transgenic (PKD) rats were analysed for neuroretinal function, vaso-regression and retinal expression of junction-associated proteins. In all three models, neuroretinal impairment and vasoregression were observed, but gene expression profiling of junction-associated proteins demonstrated nearly no overlap between the three models. However, the differently expressed genes were from the main classes of claudins, connexins and integrins in all models. Changes in Rcor1 expression in diabetic rats and Egr1 expression in PKD rats confirmed the differences in upstream transcription factor level between the models. In PKD rats, a possible role for miRNA regulation was observed, indicated by an upregulation of miR-26b-5p, miR-122-5p and miR-300-3p, which was not observed in the other models. In silico allocation of connexins revealed not only differences in regulated subtypes, but also in affected retinal cell types, as well as connexin specific upstream regulators Sox7 and miR-92a-3p. In this study, we demonstrate that, despite their similar phenotype, models for diabetic retinopathy exhibit significant differences in their pathogenic pathways and primarily affected cell types. These results underline the importance for more sensitive diagnostic tools to identify path-ogenic clusters in patients as the next step towards a desperately needed personalized therapy.