Single line or combination immune checkpoint inhibition (ICI) therapies in metastatic melanoma have shown high response rates and durable survival. However, while at least 50% of patients show positive response to ICI treatments, ∼60% of treated patients develop severe irAEs with high morbidity, substantially reducing treatment benefits. To date, no clinical or molecular indicators have been identified to predict irAEs.
Using data from our germline whole-exome sequencing scan of 69 anti-CTLA-4 (ipilimumab; IPI) treated patients, 30 with grade 3-5 irAEs and 39 with grade 0-2 irAEs, we assessed the association of exonic variants with irAEs by logistic regression analysis. Next, we cross-referenced the signification associations with irAEs against 1,140 risk variants previously found in GWAS on autoimmunity. Finally, pathway analyses of germline associations have been performed to identify biological networks involved in the susceptibility of IPI-related irAEs.
We found most significant associations with increased risk of severe irAEs for two germline variants: rs504963 (OR = 2.57, p = 0.005697) in 3'UTR of FUT2, previously associated in GWAS with multiple autoimmune traits, including psoriasis, lupus, rheumatoid arthritis and celiac disease; and rs1738074 (OR = 2.209, p = 0.02528) in TAGAP found in GWAS for association with celiac disease and multiple sclerosis. While additional variants were also identified as significant, pathway analyses have found enrichment of associated variants in chemotaxis biological processes (p = 0.04).
Our approach provides the first evidence that germline variants previously associated with autoimmune risk modulate the susceptibility to irAEs in patients treated by ICI. This includes associations with severe irAEs for FUT2, a protein involved in H-antigen production and linked with multiple autoimmune diseases. We have also found enrichment of variants associated with irAEs in chemotaxis processes, critically important in migration of dendritic cells upon treatment with anti-CTLA4 ICI. Upon validation in larger patient subsets, these findings suggest novel personalized biomarkers predictive of IPI-related toxicty, potentially extending to other ICI treatments.
Clinical trial identification
Legal entity responsible for the study
J.S. Weber: Consulting for Bristol-Myers Squibb, Merck, AstraZeneca and Genetech. All other authors have declared no conflicts of interest.