Objectives Alterations in natural killer (NK) cells activity cause damage to pancreatic islets in type 1 diabetes mellitus (T1DM). The aim of this study is to identify T1DM ketosis- or ketoacidosis-related genes in activated CD56+CD16+ NK cells. Methods Microarray datasets were downloaded from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) were analyzed using the GEO2R tool. Enrichment analyses were performed using Metascape online database and GSEA software. Cell-specific gene co-expression network was built using NetworkAnalyst tools. Cytoscape software was used to identify hub genes and construct co-expressed networks. Target miRNAs were predicted based on the DIANA-micro T, miRDB, and miRWalk online databases. Results A total of 70 DEGs were identified between T1DM patients recovered from ketosis or ketoacidosis and healthy control blood samples in GSE44314. Among the DEGs, 10 hub genes were screened out. The mature NK cell-specific gene co-expression network for DEGs in T1DM was built using NetworkAnalyst tools. DEGs between activated CD56+CD16+ NK cells and CD56brightCD16- NK cells were identified from GSE1511. After intersection, 13 overlapping genes between GSE44314 and GSE1511 microarray datasets were screened out, in which 7 hub genes were identified. Additionally, 59 target miRNAs were predicted according to the 7 hub genes. After validating with the exosome miRNA expression profile dataset of GSE97123, seven differentially expressed miRNAs (DEmiRNAs) in plasma-derived exosome were selected. Finally, a mRNA–miRNA network was constructed, which was involved in the T1DM ketosis or ketoacidosis process. Conclusion This work identified seven hub genes in activated CD56+CD16+ NK cells and seven miRNAs in plasma-derived exosome as potential predictors of T1DM ketoacidosis, which provided a novel insight for the pathogenesis at the transcriptome level.
