Genomics of Multiple Myeloma Influences the Expression of CAR T-Cell Targets

Multiple Myeloma (MM) is an incurable disease, with a particularly poor prognosis for patients with refractory/relapsed MM or high-risk cytogenetics. Chimeric Antigen Receptor (CAR) T-cell therapy targeting BCMA can induce deep responses in highly pretreated RRMM; however, remissions are not sustained, and the majority of patients eventually relapse. We hypothesized that genomic determinants of MM play a role in dictating the expression of surface targets that can be of use for immune targeting.

We analyzed the gene expression of 24 immunotherapeutic targets in a combined dataset of 1900 MM patients from three independent expression datasets obtained from the Multiple Myeloma Research Foundation CoMMpass study and Gene Expression Omnibus. Given that CAR T-cell therapy may be especially important for patients with high-risk myeloma, we defined the expression of each target in high-risk MM patients by stratifying the patient population based on several genomic features impacting prognosis. In high-risk patients with t(14;16), we found that IGF1R, ITGB7, GPRC5D, CD70, TACI and ICAM1 were overexpressed compared to the standard risk patients with t(11;14). In high-risk patients with t(14;20), ITGB7, GPRC5D and TACI were overexpressed, and in high-risk patients with t(4;14), IGF1R, GPRC5D, CD70, CD44, BCMA, CD138, FUT3, SLAMF7, CD56 and CD200 were overexpressed compared to the standard risk patients with t(11;14). We found that ITGB7, CD86, CD81 and IGF1R were overexpressed in non-hyperdiploid patients and IGKC, CD138, BCMA, CD74, CD47, BST2, LY9, CD200, CD56, CD1D and ICAM1 were overexpressed in hyperdiploid patients. In stratifying patients using a quantification of drivers that include mutated genes, copy number events and chromosomal level changes (Walker et al. Blood, 2018), we found that GPRC5D, ITGB7, IGF1R and CD70 expression were increased in patients bearing 5-9 and/or 10+ drivers compared to 0-4 drivers. Additionally, we conducted a gene co-expression network analysis and identified 30 gene modules highly correlated with 16 cell surface targets from our panel, further suggesting that genetic determinants of MM may shape a targetable cell surfaceome. In order to determine whether targeting any of these candidate antigens might cause major toxicity to normal cells, we utilized several repositories providing protein data (Perna et al. Cancer Cell, 2017) to annotate their expression in several normal cell types.

Our work provides a means of target selection for precision CAR therapy, by considering both patient genomic backgrounds and cancer cell surface profiles. Furthermore, our results provide a roadmap for immunotherapy of MM by unbiasedly comparing the expression of top MM cell surface targets in patient data and normal cells and suggest that the genetic landscape of MM may predict the expression of specific targets for precision immunotherapy. The quest for novel MM targets for immunotherapies remains open, and CAR target discovery driven by specific genetic events remains an active area of investigation.