In Frontiers in medicine
Hepatocellular carcinoma (HCC) is a commonly diagnosed cancer with high mortality rates. The immune response plays an important role in the progression of HCC. Immunotherapies are becoming an increasingly promising tool for treating cancers. Advancements in scRNA-seq (single-cell RNA sequencing) have allowed us to identify new subsets in the immune microenvironment of HCC. Yet, distribution of these new cell types and their potential prognostic value in bulk samples from large cohorts remained unclear. This study aimed to investigate the tumor-infiltration and prognostic value of new cell subsets identified by a previous scRNA-seq study in a TCGA HCC cohort using CIBERSORTx, a machine learning method to estimate cell proportion and infer cell-type-specific gene expression profiles. We observed different distributions of tumor-infiltrating lymphocytes between tumor and normal cells. Among these, the CD4-GZMA cell subset showed association with prognosis (log-rank test, p < 0.05). We further analyzed CD4-GZMA cell specific gene expression with CIBERSORTx, and found 19 prognostic genes (univariable cox regression, p < 0.05). Finally, we applied Least absolute shrinkage and selection operator (LASSO) Cox regression to construct an immune risk score model and performed a prognostic assessment of our model in TCGA and ICGC cohorts. Taken together, the immune landscape in HCC bulk samples may be more complex than assumed, with heterogeneity and different tumor-infiltration relative to scRNA-seq results. Additionally, CD4-GZMA cells and their characteristics may yield therapeutic benefits in the immune treatment of HCC.
Li Lixing, Shen Lu, Ma Jingsong, Zhou Qiang, Li Mo, Wu Hao, Wei Muyun, Zhang Di, Wang Ting, Qin Shengying, Xing Tonghai
CIBERSORTx, HCC, LASSO, TILs, immune risk score, scRNA-seq