Screening of immune-related secretory proteins linking chronic kidney disease with calcific aortic valve disease based on comprehensive bioinformatics analysis and machine learning

Abstract Background Chronic kidney disease (CKD) is one of the most significant cardiovascular risk factors, playing vital roles in various cardiovascular diseases such as calcific aortic valve disease (CAVD). We aim to explore the CKD-associated genes potentially involving CAVD pathogenesis, and to discover candidate biomarkers for the diagnosis of CKD with CAVD. Methods Three CAVD, one CKD-PBMC and one CKD-Kidney datasets of expression profiles were obtained from the GEO database. Firstly, to detect CAVD key genes and CKD-associated secretory proteins, differentially expressed analysis and WGCNA were carried out. Protein-protein interaction (PPI), functional enrichment and cMAP analyses were employed to reveal CKD-related pathogenic genes and underlying mechanisms in CKD-related CAVD as well as the potential drugs for CAVD treatment. Then, machine learning algorithms including LASSO regression and random forest were adopted for screening candidate biomarkers and constructing diagnostic nomogram for predicting CKD-related CAVD. Moreover, ROC curve, calibration curve and decision curve analyses were applied to evaluate the diagnostic performance of nomogram. Finally, the CIBERSORT algorithm was used to explore immune cell infiltration in CAVD. Results The integrated CAVD dataset identified 124 CAVD key genes by intersecting differential expression and WGCNA analyses. Totally 983 CKD-associated secretory proteins were screened by differential expression analysis of CKD-PBMC/Kidney datasets. PPI analysis identified two key modules containing 76 nodes, regarded as CKD-related pathogenic genes in CAVD, which were mostly enriched in inflammatory and immune regulation by enrichment analysis. The cMAP analysis exposed metyrapone as a more potential drug for CAVD treatment. 17 genes were overlapped between CAVD key genes and CKD-associated secretory proteins, and two hub genes were chosen as candidate biomarkers for developing nomogram with ideal diagnostic performance through machine learning. Furthermore, SLPI/MMP9 expression patterns were confirmed in our external cohort and the nomogram could serve as novel diagnosis models for distinguishing CAVD. Finally, immune cell infiltration results uncovered immune dysregulation in CAVD, and SLPI/MMP9 were significantly associated with invasive immune cells. Conclusions We revealed the inflammatory-immune pathways underlying CKD-related CAVD, and developed SLPI/MMP9-based CAVD diagnostic nomogram, which offered novel insights into future serum-based diagnosis and therapeutic intervention of CKD with CAVD.