RESUMEN
Glutamine metabolism (GM) plays a critical role in hepatocellular carcinoma (HCC); however, a comprehensive methodology to quantify GM activity is still lacking. We developed a transcriptome-based GMScore to evaluate GM activity and investigated the association of GMScore with prognosis and therapeutic resistance. Two independent HCC cohorts with transcriptome data were selected from The Cancer Genome Atlas (TCGA, n = 365) and the International Cancer Genome Consortium (ICGC, n = 231). The expression of 41 GM-associated genes were used to construct and validate GMScore. Several genomic or transcriptomic biomarkers were also estimated. Tumor response to immune checkpoint inhibitors (ICIs) was predicted using the tumor immune dysfunction and exclusion algorithm. GMScore was closely correlated with patient characteristics, including stage, histology grade, alpha-fetoprotein level, and vascular invasion. High GMScore was an independent risk factor for overall survival (OS) in both cohorts (HR = 4.2 and 3.91, both p < 0.001), superior to clinical indices and other biomarkers. High GMScore presented transcriptome features to indicate cell growth advantages and genetic stability, which was associated with poor OS of patients who received transcatheter arterial chemoembolization (TACE). High GMScore was also related to high expression of immune checkpoint genes, increased infiltration of regulatory T cells, and decreased infiltration of M1 macrophages. More importantly, high GMScore indicated poor predicted responses to ICIs, which could be verified in an ICI-treated melanoma cohort. In conclusion, GMScore is a strong prognostic index that may be integrated into existing clinical algorithms. A high GMScore may indicate resistance to TACE and ICIs based on its transcriptome and immune features. Validations using other HCC cohorts, especially ICI-treated HCC cohorts, are necessary.