Tumor-Intrinsic and Immune Predictors of Response to Ixazomib, Lenalidomide, and Dexamethasone in High-Risk Smoldering Multiple Myeloma

Background

Early treatment of patients with high-risk smoldering myeloma (HRSMM) has been shown to delay progression to multiple myeloma (MM), but not all patients respond well. Identifying biological predictors of response and resistance to treatment can help optimize treatment selection for patients and improve outcomes. Here, we identify tumor and immune biomarkers of response to ixazomib, lenalidomide, and dexamethasone treatment (I-PRISM study) that can be measured at diagnosis.

Methods

We performed single-cell RNA sequencing (scRNAseq), coupled with B cell receptor and T cell receptor sequencing, on CD138+ tumor and CD138- immune cells as well as whole genome sequencing (WGS) on tumor cells from the bone marrow of 22 HRSMM patients (14 biochemical progressors, 8 non-progressors) enrolled in the I-PRISM study and 11 healthy donors. Among progressors, 5 patients went on to develop myeloma by the SLiM-CRAB criteria.

Results

The median time to biochemical progression was 33 months and the median follow-up for non-progressors was 48 months. 35 patients were successfully cytogenetically classified by FISH; in the remaining patients (n=10), we inferred the presence of IgH translocations and copy number variants (CNVs) using scRNAseq and WGS. Translocations and CNVs inferred from scRNAseq showed excellent concordance with WGS data, suggesting that scRNAseq-based cytogenetic classification is reliable. Notably, scRNAseq identified additional CNVs (n=7) that were missed by FISH, highlighting its greater sensitivity. No significant association was observed between cytogenetics and response to treatment.

Differential expression analysis of tumor cells identified by BCRseq revealed higher MHC-I gene expression in non-progressors compared to progressors, suggesting a potential link between MHC-I expression and response to therapy. We validated this observation using data from the PADIMAC study, where MM patients were treated with a different proteasome inhibitor (PI; Bortezomib) in a combination that did not include lenalidomide. This suggests a broader link between tumor-intrinsic MHC-I expression and response to proteasome inhibition (PI). Gene set enrichment analysis (GSEA) further revealed upregulated oxidative phosphorylation pathways in tumor cells from progressors, suggesting enhanced energy metabolism as a driver of progression post-treatment with PIs. Since cytotoxic CD8+ T cells recognize and eliminate cells presenting endogenous antigens on their surface via the MHC-I pathway, we investigated whether they show differences between progressors and non-progressors. By comparing the clonal expansion rate of two CD8+ T cell subsets, the earlier, memory-like granzyme GZMK+ compartment and the more terminally differentiated GZMB+ compartment, we found that progressors exhibited greater clonal expansion in the GZMK+ compartment than non-progressors. Notably, we previously showed that the GZMK+ compartment is the main source of PD-1 expression, a key T cell exhaustion marker, in patient bone marrow. Additionally, clonally expanded T cells from progressors showed a reduced proportion of the GZMB+ phenotype, suggesting their clonally expanded T cells may have a less mature cytotoxic profile compared to non-progressors. This may be related to the lower levels of MHC-I expression observed in tumor cells from progressors, which may indicate impaired tumor antigen presentation and cytotoxic T cell activation.

Conclusions

Our results highlight the role of the immune microenvironment and its complex interplay with tumor cells, potentially involving tumor antigen presentation and cytotoxic T cell activation, in response to PI-based regimens. Immune profiling may help to improve risk stratification of patients with HRSMM and MM and inform the selection of therapy for specific patients.