Using mouse liver cancer models based on somatic genome editing to predict immune checkpoint inhibitor responses

Tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs) are the only two classes of FDA-approved drugs for individuals with advanced hepatocellular carcinoma (HCC). While TKIs confer only modest survival benefits, ICIs have been associated with remarkable outcomes but only in the minority of patients who respond. Understanding the mechanisms that determine the efficacy of ICIs in HCC will help to stratify patients likely to respond to ICIs. This study aims to elucidate how genetic composition and specific oncogenic pathways regulate the immune composition of HCC, which directly affects response to ICIs.

A collection of mouse HCCs with genotypes that closely simulate the genetic composition found in human HCCs were established using genome-editing approaches involving the delivery of transposon and CRISPR-Cas9 systems by hydrodynamic tail vein injection. Mouse HCC tumors were analyzed by RNA-sequencing while tumor-infiltrating T cells were analyzed by flow cytometry and single-cell RNA-sequencing.

Based on the CD8⁺ T cell-infiltration level, we characterized tumors with different genotypes into cold and hot tumors. Anti-PD-1 treatment had no effect in cold tumors but was greatly effective in hot tumors. As proof-of-concept, a cold tumor (Trp53^KO/MYC^OE) and a hot tumor (Keap1^KO/MYC^OE) were further characterized. Tumor-infiltrating CD8⁺ T cells from Keap1^KO/MYC^OE HCCs expressed higher levels of proinflammatory chemokines and exhibited enrichment of a progenitor exhausted CD8⁺ T-cell phenotype compared to those in Trp53^KO/MYC^OE HCCs. The TKI sorafenib sensitized Trp53^KO/MYC^OE HCCs to anti-PD-1 treatment.

Single anti-PD-1 treatment appears to be effective in HCCs with genetic mutations driving hot tumors, while combined anti-PD-1 and sorafenib treatment may be more appropriate in HCCs with genetic mutations driving cold tumors.

Genetic alterations of different driver genes in mouse liver cancers are associated with tumor-infiltrating CD8⁺ T cells and anti-PD-1 response. Mouse HCCs with different genetic compositions can be grouped into hot and cold tumors based on the level of tumor-infiltrating CD8⁺ T cells. This study provides proof-of-concept evidence to show that hot tumors are responsive to anti-PD-1 treatment while cold tumors are more suitable for combined treatment with anti-PD-1 and sorafenib. Our study might help to guide the design of patient stratification systems for single or combined treatments involving anti-PD-1.