Integrated single cell transcriptomic analysis of kinic-related genes in cervical cancer reveals diagnostic value and microenvironment regulation

Cervical squamous cell carcinoma (CESC) is one of the most common malignancies of the female reproductive system, and its development involves complex gene regulatory networks. Kinic, a novel post-translational modification of proteins, has recently garnered increasing attention for its role in regulating tumor metabolism and signal transduction; however, its functional role and clinical significance in cervical cancer remain unclear. We integrated multiple CESC transcriptomic datasets from TCGA and GEO (GSE39001, GSE63514, GSE9750, TCGA-CESC, GSE52903) and corrected for batch effects using ComBat. We screened for genes in modules associated with clinical traits using WGCNA, and identified candidate genes by taking the intersection of these modules with differentially expressed genes (DEGs) and the Kinic -related gene set (KRG). We selected core genes through cross-validation using three machine learning algorithms—LASSO, SVM-RFE, and Random Forest—and constructed diagnostic receiver operating characteristic (ROC) curves to evaluate their performance. Biological pathways were analyzed using GSEA/GSVA, while immune microenvironment characteristics were examined via CIBERSORT, and core-gene-expression-based molecular subtyping was performed via consensus clustering. Single-cell RNA sequencing was used to determine the cellular origins of genes, and regulatory functions were explored using Monocle 2 pseudo-time series analysis, scTenifoldKnk virtual knockout, and CellChat cell-cell communication analysis. The intersection of DEGs, WGCNA module genes, and KRGs yielded 8 candidate genes; after cross-validation using machine learning, CDKN2A, XPO1, and TST were ultimately identified as core diagnostic genes. The AUC of the three-gene diagnostic panel reached 0.952 (95% CI: 0.916–0.979). These three genes were primarily enriched in cell cycle, DNA replication, and metabolic pathways. Analysis of the immune microenvironment revealed increased infiltration of M0/M1 macrophages and activated memory CD4 + T cells, along with depletion of quiescent mast cells and memory CD4 + T cells; consensus clustering identified three core-gene-expression-based molecular subtypes: C1, C2, and C3. Single-cell sequencing revealed that CDKN2A and TST are primarily expressed in epithelial cells, while XPO1 is most highly expressed in macrophages; pseudo-time-series analysis showed that the expression of these three genes dynamically changes throughout development. Virtual knockout simulation suggested that CDKN2A and XPO1 were associated with differential regulation of approximately 2.0% and 2.8% of genes, respectively, while knocking out TST had no significant effect. Cell-cell communication analysis revealed that the network was more complex in the CDKN2A low-expression group, while communication was more active in the XPO1 and TST high-expression groups; the MIF signaling pathway ligand-receptor pairs (APP–CD74, MIF–[CD74 + CXCR4]) showed significant differences across the groups. CDKN2A, XPO1, and TST are key genes associated with Kinic in CESC; their expression patterns are linked to tumor progression and may contribute to cell cycle regulation, metabolic reprogramming, and remodeling of the immune microenvironment. The diagnostic multivariable logistic regression model based on these three genes demonstrates high accuracy, providing new evidence and potential biomarkers for the molecular diagnosis, immunophenotyping, and targeted therapy of CESCs.