Assistant Professor Memorial Sloan Kettering New York City, New York
Introduction: The introduction of chimeric antigen receptor T cells (CART) has transformed treatment for relapsed/refractory multiple myeloma (RRMM). Despite high response rates, the mechanisms underlying primary resistance and treatment failure remain poorly defined.
Methods: To investigate the genomic mechanisms underlying primary refractoriness to anti-BCMA CART, we analyzed 76 whole genomes and 10 whole exomes (WES) from 74 patients treated with anti-BCMA CART (ide-cel, n=58; cilta-cel, n=16).
Results: Genomic complexity (i.e., Maura et al., JCO 2024) correlated with inferior outcomes after CART (p=0.03). This was validated in an independent cohort of 35 cilta-cel–treated patients profiled with MSK-IMPACT sequencing (p=0.0016).
Three additional genomic groups were associated with worse outcomes. First, loss of genes essential for plasma cell differentiation (e.g. CD38, XBP1) was linked to poor prognosis. Importantly, XBP1 and CD38 expression correlated with TNFRSF17 in RNAseq data from 1719 MM patients. Second, alterations in NFkB pathway genes (i.e., CYLD, PRKD2, NFKB2, MAP3K14) were associated with resistance to anti-BCMA CART. To validate the link between NFkB and plasma cell differentiating genes and suboptimal response to anti-BCMA immunotherapy, we investigated 5 Vk*MYC mice treated with anti-BCMA T-cell engagers. The two refractory lines showed loss of Xbp1, Map3k14 fusion, and downregulation of Tnfrsf17.
The third group associated with poor outcomes included events known to confer high-risk disease such as 1q gain, TP53 mutation, loss of CDKN2C, RPL5, CCSER1, and MAF/MAFB translocations. Using the RNA-seq data set, we observed a correlation between TP53 and TNFRSF17 expression (R2 0.18; p< 0.0001). To further explore this, we analyzed single-cell RNA sequencing (scRNA-seq) data from clonal plasma cells in 11 patients collected prior to CAR-T therapy, of which 5 also had matched WGS data, observing increased cell cycle activity and reduced plasma cell features in non-responders - including lower TNFRSF17 expression. Finally, we assessed Tnfrsf17 expression in the newly developed Bclγ1 mouse models, with and without monoallelic loss of Trp53 (Larrayoz et al ASH 2024). Similar to human data, this model showed that monoallelic Trp53 loss was associated with impaired plasma cell differentiation and reduced Tnfrsf17 expression. Finally, co-occurrence of genomic complexity, loss of plasma cell differentiation genes, and NFκB pathway alterations defined a group of patients with primary refractoriness to anti-BCMA CART. Importantly, the prognostic impact of these genomic alterations was independent of the clinically derived MyCARe risk score, presence of EMD, or CART products.
Conclusions: Overall, these findings suggest that comprehensive genomic profiling can more accurately predict clinical outcomes in MM patients treated with anti-BCMA CART than current clinical models, and may serve as a valuable tool for guiding personalized treatment strategies.
