Construction of a single nucleotide variant score-related gene-based prognostic model in hepatocellular carcinoma: analysis of multi-independent databases and validation in vitro.

Background: The accumulation of single nucleotide variants (SNVs) and the emergence of neoantigens can affect tumour proliferation and the immune microenvironment. However, the SNV-related immune microenvironment characteristics and key genes involved in hepatocellular carcinoma (HCC) are still unclear. We aimed to evaluate differences in the SNV-related immune microenvironment, construct a prognostic model and validate the key genes in vitro.

Methods: The categories of samples were defined by the expression of SNV score-related genes to evaluate the differences in mutational features, immune environment and prognosis. The survival model was constructed with survival-associated genes and verified in two independent test datasets. RCAN2, the key gene screened out for biofunction, was validated in vitro.

Results: IC2, among the three integrated clusters (IC1, IC2, IC3) classified by the 82 SNV score-related genes, was distinct from the rest in SNV score and immune cell infiltration, showing a better prognosis. Seven prognostic markers, HTRA3, GGT5, RCAN2, LGALS3, CXCL1, CLEC3B, and CTHRC1, were screened to construct a prognostic model. The survival model distinguished high-risk patients with poor prognoses in three independent datasets (log-rank P < 0.0001, 0.011, and 0.0068, respectively) with acceptable sensitivity and specificity. RCAN2 was inversely correlated with NK cell infiltration, and knockdown of RCAN2 promoted proliferation in HCC.

Conclusions: This study revealed the characteristics of the HCC SNV-associated subgroup and screened seven latent markers for their accuracy of prognosis. Additionally, RCAN2 was preliminarily proven to influence proliferation in HCC and it had a close relationship with NK cell infiltration in vitro. With the capability to predict HCC outcomes, the model constructed with seven key differentially expressed genes offers new insights into individual therapy.