Tumor intrinsic pathways dictate the immune landscape and responses to immune checkpoint inhibitors in liver cancer

Abstract

Immune checkpoint inhibitors (ICIs) have become the standard of care for hepatocellular carcinoma (HCC) at advanced stages. ICIs have remarkable outcomes in activating T cells and providing durable responses but are only limited to a minority of HCC patients. Understanding the mechanisms that determine the efficacy of ICIs in HCC is beneficial to the stratification of patients who will likely respond or not respond to ICIs.

Emerging studies have revealed that tumor-intrinsic oncogenic pathways orchestrate the intraturmoral immune composition. To robustly study how genetic alterations shape HCC tumor immune microenvironment thereby affecting the response to ICI, a library of mouse HCCs with genotypes that closely simulate the genetic composition found in human HCCs were generated using genome-editing systems involving transposon and CRISPR-Cas9 systems by hydrodynamic tail-vein injection (HDTVi). Based on the CD8+ T cell infiltration level, tumors with different genotypes could be categorized into cold (low CD8+ T cell number) and hot (high CD8+ T cell number) 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 (Trp53KO/MYCOE) and a hot tumor (Keap1KO/MYCOE) were further characterized. RNA sequencing analysis of HCC tissues revealed Cxcl10 as a key chemokine to determine CD8+ T cell infiltration and responsiveness to anti-PD-1 in Keap1KO/MYCOE tumors. Single-cell RNA sequencing analysis of tumor-infiltrating T cells identified higher expression of proinflammatory cytokines and chemokines, and exhibited enrichment of progenitor exhausted CD8+ T cell phenotype in Keap1KO/MYCOE tumors when compared to Trp53KO/MYCOE HCC. Interestingly, the tyrosine kinase inhibitor (TKI), sorafenib, sensitized cold tumors to anti-PD-1 treatment.

By further studying the same anti-PD-1-resistant Trp53KO/MYCOE HDTVi model using mass cytometry, PD-1 blockade up-regulated another immune checkpoint, TIGIT, in tumor-infiltrating CD8+ T cells. TIGIT is known to bind ligands in the PVR family. Notably, PVRL1 was over-expressed and was associated with poorer survival in HCC patients. Mouse HCC tumors with Pvrl1 knockdown were smaller and infiltrated with more CD8+ T cells. Mechanistically, PVRL1 stabilized cancer cell’s surface PVR, which interacted with TIGIT to inactivate CD8+ T cells. Excitingly, combined treatment of anti-PD-1 and anti-TIGIT reduced tumor growth, prolonged survival, and thus overcame resistance to PD-1 inhibitor monotherapy in Trp53KO/MYCOE mouse HCC.

In sum, HCCs with genetic mutations driving hot tumors might be suitable for anti-PD-1 single treatment. HCCs with genetic mutations driving cold tumors might be suitable for combined treatment of anti-PD-1 plus TKI or dual ICI treatment. Furthermore, the study has exemplified the use of genome-editing systems to generate mouse HCCs to precisely simulate the same genetic makeup of human HCC to pave the way for precision medicine.