HLA class I expression and ratios as diagnostic and predictive biomarkers in hepatocellular carcinoma: A multi-omics study

Background Human leukocyte antigen class I (HLA-I) genes, including HLA-A, HLA-B, and HLA-C, are essential for immune responses against hepatocellular carcinoma (HCC) by presenting peptides to cytotoxic T lymphocytes. This study aimed to investigate the diagnostic and predictive value of HLA-I genes in HCC through the integration of bulk and single-cell RNA sequencing, spatial transcriptomics, and mutational profiling. Materials and methods Bulk RNA-seq and microarray datasets were analyzed to assess differential expression, clinical associations, survival outcomes, and diagnostic potential of HLA-I genes. Single-cell RNA sequencing data were integrated to evaluate cell-type-specific expression patterns, co-expression dynamics, and immune evasion using pseudo-bulk approaches. Spatial transcriptomics was analyzed at both pooled and patient levels to characterize compartment-specific heterogeneity. Additionally, allele-specific expression and peptide-MHC binding affinities were examined, and somatic mutation profiles of HLA-I and antigen-processing machinery genes were obtained from The Cancer Genome Atlas Liver Hepatocellular Carcinoma cohort (TCGA-LIHC). Comprehensive statistical and modeling analyses were performed to ensure robust evaluation across datasets. Results Elevated HLA-I expression levels and expression ratios were associated with advanced clinicopathological features. Higher HLA-A/HLA-B expression ratios correlated with poor clinical outcomes. HLA-I gene expression was higher in Sorafenib non-responders, whereas the expression ratios showed divergent patterns: HLA-A/HLA-B was lower, while HLA-A/HLA-C and HLA-B/HLA-C were higher in non-responders. In a separate diagnostic analysis distinguishing HCC from non-tumor liver tissue, the combined model incorporating all three HLA-I expression ratios achieved an area under the receiver operating characteristic (ROC) curve (AUC) of 0.90. Single-cell RNA sequencing revealed reduced HLA-I expression in most tumor microenvironment cell types, with a subset of malignant cells showing apparent loss of expression. Expression dysregulation and binding affinity differences were observed among high-prevalence HLA-I alleles. Mutation analysis indicated that approximately 7% of patients harbor non-synonymous alterations in HLA-I or antigen-processing machinery genes. Additional evaluation demonstrated stable predictive performance and clinical utility beyond ROC-based discrimination. Conclusion HLA-I gene expression patterns, expression ratios, and mutation status show potential as diagnostic and predictive biomarkers of Sorafenib response in HCC patients, highlighting their relevance for precision oncology applications.