Prognostic genes in hepatocellular carcinoma and their function: an analysis integrating high-throughput-based...

Identifying novel prognostic indicators for patients with hepatocellular carcinoma (HCC) is critical to improving patient outcomes. Mapping the spatial composition of HCC tumors remains challenging. Therefore, this study aimed to identify prognostic genes and assess their role in HCC by integrating spatial transcriptome sequencing (ST-seq) and single-cell RNA sequencing (scRNA-seq).

HCC-related datasets, including The Cancer Genome Atlas (TCGA)-HCC, International Cancer Genome Consortium (ICGC)-HCC, GSE149614, and GSE203612, were examined in this study. Single-cell and spatial transcriptome analyses were first conducted and followed by cell communication analysis. Target cell types with differential proportions and higher enrichment were identified. Least absolute shrinkage and selection operator (LASSO) Cox regression analysis was implemented to select prognostic genes and construct a risk model. Subsequently, a nomogram incorporating independent prognostic factors was developed, and immune microenvironment and immunotherapy response were analyzed. Finally, pseudotime analysis was performed to determine the developmental trajectory of the target cell types.

Among hepatocytes, T/natural killer (NK) cells, B cells, myeloid cells, fibroblasts, and endothelial cells, hepatocytes and T/NK cells were identified as target cell types. ST analysis delineated seven cell subclusters (C0-C6). Cell communication analysis indicated that hepatocytes interact more strongly with other cell types compared to T/NK cells. The number of interactions between C4 and other cell subclusters was higher than the interaction intensity. Of the seven prognostic genes, ADH4, IGFBP3, and LCAT were downregulated in HCC, while AKR1B10, GAGE2A, MAGEA6, and UCHL1 were upregulated. The patients deemed to have a higher risk according to the model had shorter survival. The nomogram that was created by combining the seven prognostic genes had a good ability to forecast the patients' probability of survival. In the immunoinfiltration analysis, UCHL1 expression had the highest positive correlation with T/NK cell abundance. Moreover, in silico analysis suggested a potential association between high-risk status and reduced sensitivity to immunotherapy. Pseudotime analysis indicated that the expression of UCHL1, GAGE2A, MAGEA6, ADH4, and LCAT displayed an inverted U shape during the differentiation of hepatocytes, with the expression of UCHL1 was elevated at the end of T/NK cell differentiation.

This study identified hepatocytes, T/NK cells, and seven prognostic genes in HCC, constructed a risk model and nomogram capable of predicting patient prognosis, and revealed through computational prediction that patients in the high-risk group might exhibit a poor response to immunotherapy, thereby providing new potential tools for the clinical treatment of HCC.