Abstract / Summary
Abstract Background Gamma-aminobutyric acid (GABA), a metabolite that promotes metastasis, has been linked to hepatocellular carcinoma (HCC) progression and metastasis. However, GABAergic communication among non-parenchymal cells during HCC progression and metastasis remains to be fully elucidated at the single-cell level. Methods We integrated single-cell (GSE149614) and TCGA-LIHC bulk data, annotated immune/stromal subsets. GABA source and receptor scores were defined to construct a communication potential matrix, and pseudotime analysis was performed(Monocle3). Differentially expressed genes were functionally enriched (GO/KEGG), and used to construct a prognostic model (LASSO-Cox), externally validated in GSE14520 and clinical HCC samples by qPCR, IHC, and IF. Results We mapped immune and stromal subsets across normal, tumor, PVTT, and metastatic lymph node (MLN) tissues. GABA scores were elevated in tumor, PVTT, and MLN compared with normal. TCGA-LIHC analysis suggested enhanced GABA synthesis and reduced degradation, suggesting a transcriptional shift favoring GABA biosynthesis and signaling in HCC. Macrophages were activated in tumors and exhibited functional alterations at metastatic sites. T/NK cytotoxicity progressively declined from normal to metastatic tissues, correlating with elevated GABA and macrophage scores. A three-gene prognostic model ( DNASE1L3 , STMN1 and SPP1 ) was validated and showed predictive value for HCC prognosis, and its expression trends in 20 clinical HCC samples were consistent with predictions. Conclusion This study provides a single-cell map of GABAergic communication in HCC, suggesting that GABA signaling may promote immunosuppression and metastasis via macrophage-NK crosstalk. A three-gene model was developed for risk stratification, and targeting the GABA pathway may be a promising direction for future investigation.