Abstract
Abstract
Background Checkpoint inhibitors (CPIs) benefit only a proportion of patients and may be associated with severe adverse events (AEs) which cannot be predicted. We hypothesized that the host genetics could be used as predictive biomarkers for CPI response and AE prediction. Therefore, we conducted a study based on single nucleotide polymorphisms (SNPs) from genes affiliated with immune response and tumor-microenvironment interaction.
Methods Germinal DNA was obtained from advanced cancer patients treated with anti PD-1/PD-L1 CPIs in the Centre Antoine Lacassagne (Nice, France) from July 2012 to January 2018. DNA was genotyped on the MassARRAY system (Agena Bioscience®) using a custom panel of 166 SNPs covering 86 preselected immunogenetic-related genes (Minor allele frequency MAF>0.05 in Caucasians). All tested SNPs were in Hardy-Weinberg equilibrium. Univariate analysis was performed to select the significant SNPs (p<0.05) by either Fisher or Ki2 tests. Treatment outcome prediction was based on an elastic-net penalized logistic regression with 5-fold cross validation. The predictive ability model was performed using a concordance (c)-index (c-index > 0.5 being considered as good prediction). Computational analysis using a GTEX portal was used to determine potential eQTL (expression Quantitative Trait Loci) in tissues.
Results 94 patients were identified, with median age 68 (32-85), 67% male, with a majority (51%) having advanced non-small cell lung cancer. Median follow-up was 16.3 months (95% CI: 12.5-18.3). Overall response rate (ORR) was observed in 49/94 (54%) of patients, with adverse events (grade 3-4) observed in 15/94 (16%) of patients. ORR was significantly predicted by tumor microenvironment related gene polymorphisms (CCL2, NOS3, IL1RN, IL12B, CXCR3, IL6R). In contrast, grade 3-4 AEs were linked to target-related gene SNPs (UNG, IFNW1, CTLA-4, PD-L1, IFNL4). The predictive (c)-index was 0.81 (95% CI: 0.72-0.9) for response and 0.89 (95% CI: 0.76-1.00) for toxicity. In silico functionality exploring (GTEX portal) pointed IL6R (rs4845618) and CTLA4 (rs3087243) as impacting gene expression.
Conclusion Our data strongly support the role of distinct SNPs in immunogenetic related genes to predict efficacy and safety of anti PD1/PD-L1 therapies. These data support the notion that patient-specific, germinal biomarkers may supplement tumor-specific biomarkers in predicting response to CPI therapy, and that additional germinal biomarkers may predict grade 3-4 AEs.
Citation Format: Sadal REFAE, Jocelyn GAL, Nathalie EBRAN, Josiane OTTO, Delphine BORCHIELLINI, Frederic Peyrade, Emmanuel CHAMOREY, Patrick Brest, Gerard Alain Milano, Esma SAADA-BOUZID. Germinal immunogenetics predicts treatment outcome for PD1 PD-L1 checkpoint inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1370.