Abstract
Background: Diabetic nephropathy (DN) was considered a severe microvascular complication of diabetes, which was recognized as the second leading cause of end-stage renal diseases. Therefore, identifying several effective biomarkers and models to diagnosis and subtype DN is imminent. Necroptosis, a distinct form of programmed cell death, has been established to play a critical role in various inflammatory diseases. Herein, we described the novel landscape of necroptosis in DN and exploit a powerful necroptosis-mediated model for the diagnosis of DN. Methods: We obtained three datasets (GSE96804, GSE30122, and GSE30528) from the Gene Expression Omnibus (GEO) database and necroptosis-related genes (NRGs) from the GeneCards website. Via differential expression analysis and machine learning, significant NRGs were identified. And different necroptosis-related DN subtypes were divided using consensus cluster analysis. The principal component analysis (PCA) algorithm was utilized to calculate the necroptosis score. Finally, the logistic multivariate analysis were performed to construct the necroptosis-mediated diagnostic model for DN. Results: According to several public transcriptomic datasets in GEO, we obtained eight significant necroptosis-related regulators in the occurrence and progress of DN, including CFLAR, FMR1, GSDMD, IKBKB, MAP3K7, NFKBIA, PTGES3, and SFTPA1 via diversified machine learning methods. Subsequently, employing consensus cluster analysis and PCA algorithm, the DN samples in our training set were stratified into two diverse necroptosis-related subtypes based on our eight regulators' expression levels. These subtypes exhibited varying necroptosis scores. Then, we used various functional enrichment analysis and immune infiltration analysis to explore the biological background, immune landscape and inflammatory status of the above subtypes. Finally, a necroptosis-mediated diagnostic model was exploited based on the two subtypes and validated in several external verification datasets. Moreover, the expression level of our eight regulators were verified in the singe-cell level and glomerulus samples. And we further explored the relationship between the expression of eight regulators and the kidney function of DN. Conclusion: In summary, our necroptosis scoring model and necroptosis-mediated diagnostic model fill in the blank of the relationship between necroptosis and DN in the field of bioinformatics, which may provide novel diagnostic insights and therapy strategies for DN.