Association of Homologous Recombination-DNA Damage Response Gene Mutations with Immune Biomarkers in Gastroesophageal Cancers

Abstract

The prevalence of homologous recombination-DNA damage response (HR-DDR) genetic alterations is of therapeutic interest in gastroesophageal cancers. This study is a comprehensive assessment of HR-DDR mutation prevalence across gastroesophageal adenocarcinomas and squamous cell carcinomas. Here we investigate the association of HR-DDR mutations with known predictors for immune-checkpoint inhibition [deficiency in mismatch-repair (dMMRP), tumor mutational burden (TMB), and programmed death ligand 1 (PD-L1)]. We confirmed HR-DDR mutations are present in a subset of gastroesophageal adenocarcinomas (23%) and gastroesophageal squamous cell carcinomas (20%). Biomarker expression of dMMRP (18% vs. 1%) and TMB-high with a cutoff of ≥10 mt/MB (27% vs. 9%) was significantly more prevalent in the DDR-mutated cohort compared with the non-DDR-mutated cohort. Mean combined positive score for PD-L1 in the total adenocarcinoma cohort was significantly higher in the DDR-mutated cohort compared with the non-DDR-mutated cohort (10.1 vs. 5.8). We demonstrated that alterations in ARID1A, BRCA2, PTEN, and ATM are correlated with dMMRP, TMB-high, and increased PD-L1 expression in gastroesophageal adenocarcinomas. Our findings show that a subset of gastroesophageal tumors harbor HR-DDR mutations correlated with established immune biomarkers. By better understanding the relationship between HR-DDR mutations and immune biomarkers, we may be able to develop better immunotherapy combination strategies to target these tumors.

Figure 1.

Schematic representation of HR in DDR. A DSB is induced on the DNA strand. The MRN complex (MRE11–RAD50–NBS1) recognizes and binds to the area of DNA damage. This leads to the recruitment and activation of the serine/threonine protein kinase ATM (which phosphorylates several key proteins in the HR-DDR pathway such as CHEK2, BRCA1, NBS1, and p53). ATM phosphorylates the histone variant γH2AX on Ser139, which subsequently leads to signal transduction. This leads to BRCA1 and BARD1 recruitment to the site of DNA damage. C-terminal interacting protein (CtIP) is recruited to the 5′ DNA end and begins resecting, whereas replication protein A (RPA) caps the exposed ssDNA. ATR and PALB2 are activated. BRCA1 recruits PALB2, FANCD2, BRCA2, and RAD51. RPA is exchanged for RAD51, which promotes nucleofilament formation and strand invasion to form the D-loop. DNA repair is mediated via the Holliday junction (red and black arrows represent crossover).

Figure 2.

Prevalence of patients with multiple DDR mutations stratified by (A) adenocarcinoma (AC) vs. squamous cell carcinoma (SCC; B) esophageal AC vs. esophageal SCC, and (C) gastric (combined AC and SCC) vs. gastroesophageal (combined AC and SCC).

Figure 3.

A, dMMRP/TMB biomarker prevalence across DDR cohorts, (B) mean CPSs across total adenocarcinoma, gastric adenocarcinoma, and pMMRP adenocarcinoma cohorts, and (C) TMB and PD-L1 CPS distribution across multiple DDR mutations.

Figure 4.

Percentage of cases with mutation in designated gene stratified by (A) esophageal adenocarcinoma vs. esophageal squamous cell carcinoma. Percentage of cases with mutation in designated gene stratified by (B) dMMRP, (C) TMB (≥10 mt/MB), and (D) PD-L1 (CPS ≥ 1) across total adenocarcinoma cohort (combined EC, GC, and GEJ).

Figure 5.

Protein changes across total adenocarcinoma DDR-M cohort in dMMRP vs. pMMRP (dMMRP gene prevalence cutoff 23.5%) involving ARID1A, BRCA2, PTEN, and CHEK2.

Figure 6.

Oncoprint depicting the prevalence of comutated DDR genes with ARID1A.