Abstract
BACKGROUND: Natural killer (NK) cells play a pivotal role in anti-tumor immunity; however, the prognostic significance of genetic variants in NK cell-related genes in gastric cancer (GC) remains largely uncertain. METHODS: We systematically evaluated 12,476 single-nucleotide polymorphisms (SNPs) in 151 NK cell-related genes for their associations with overall survival (OS) in 2,211 Chinese GC patients with pathological tumor-node-metastasis (TNM) stage I-III, who were enrolled in the Shanghai genome-wide association study (GWAS). Significant variants were further validated in an independent Jiangsu GWAS cohort comprising 1,049 GC patients with TNM stage I-III. The prognostic value of independent SNPs was evaluated. Bioinformatic annotations were performed through expression QTL, splicing QTL, histone modification QTL, and methylation QTL analyses, as well as transcription factor binding, differential expression, functional enrichment, and immune infiltration analyses. RESULTS: Three independent SNPs (CD160 rs9728526 A>G, MERTK rs114788905 G>A, and IL15 rs140007893 T>A) were significantly associated with GC OS, with adjusted hazard ratios of 1.16 (95% CI = 1.05-1.29, P = 0.006), 0.89 (95% CI = 0.81-0.99, P = 0.033), and 0.77 (95% CI = 0.62-0.97, P = 0.028), respectively. A combined analysis demonstrated a dose-dependent association between the number of unfavorable genotypes and poorer OS. Incorporating these SNPs improved the C-index, time-dependent AUC, net reclassification improvement (NRI), and integrated discrimination improvement (IDI) for OS prediction. Functional annotation indicated that the variant alleles exerted tissue- and immune cell-specific regulatory effects on gene expression, splicing, and transcription factor binding. The expression levels of CD160, MERTK, and IL15 were correlated with immune cell infiltration within the tumor microenvironment in GC. CONCLUSIONS: We identified three novel SNPs in NK cell-related genes that independently predict GC survival, providing potential prognostic biomarkers for risk stratification in GC, although further validation is warranted.