Integrative study of gene expression datasets in retinal samples of Diabetic Retinopathy

Sharmila Rajendran,Shanmuganathan Seetharaman,Umashankar Vetrivel,Kaviarasan Kuppan

DOI: https://doi.org/10.1016/j.exer.2022.109194

Abstract:Diabetic Retinopathy is prevalent among patients with uncontrolled hyperglycemia resulting in vision loss. Despite numerous challenges to create a link among these conditions, the characterization of pathological neovascularization causing retinal damage due to the prognosis of early non-proliferative diabetic retinopathy to late proliferative diabetic retinopathy needs deep understanding. In this study, meta-analysis-based integration of gene expression datasets for the fibrovascular membrane of PDR and neural retina of NPDR were compared, to investigate the differentially expressed genes involved in retinal angiogenesis. Human samples with gene expression profiling of the same experiment type and platform with sufficient information for analysis were included in the study. The studies from cell lines and non-human studies, human samples that include serum, cornea, lens, and/or other ocular tissues or fluids, and studies that lack basic information for analysis were excluded. The microarray datasets available in the Gene Expression Omnibus database of the early and late stages in DR were screened to find common gene expression profiles. Using the INMEX bioinformatics tool, significantly upregulated and downregulated genes in the neural retina of Non-Proliferative Diabetic Retinopathy and fibrovascular membrane of Proliferative Diabetic Retinopathy were compared and studied by the combine effect size method. Using the STRING database PPI network, 50 upregulated and 50 downregulated genes were used to find the key candidate genes involved in retinal disease/degeneration in eye/retinal tissues. In the extensive gene expression meta-analysis performed using INMEX bioinformatics tool, overall, 7935 differentially expressed genes were identified and the respective heatmap was created by using the visualization tools of INVEX. STRING database PPI network identified Retinol Binding Protein 3, Neural Retina Leucine Zipper, S-Antigen Visual Arrestin, Peripherin 2, and Aryl Hydrocarbon Receptor Interacting Protein Like-1 to be the most highly ranked hub genes. The newly discovered potential genes related to retinal angiogenesis causing FVM formation in DR may provide insight into the cellular pathogenesis of NPDR to PDR.