Development of a novel Ikaros/Aiolos transcriptional signature and correlation with patient response in multiple myeloma patients treated with CELMoDs

Background: Multiple myeloma (MM) is a neoplasm characterized by the clonal proliferation of malignant plasma cells in the bone marrow. Immunomodulatory drugs and distinct CELMoD™ agents modulate Cereblon to induce the degradation of the hematologic transcription factors, Ikaros (Ik) and Aiolos (Ai), resulting in anti-proliferative and pro-apoptotic cell intrinsic effects on multiple myeloma tumor cells. While Ik/Ai degradation in MM cells can be semi-quantified using immunohistochemistry (IHC), clinical responses are not always correlated. Here, we devise a quantitative RNA-seq based Ik/Ai activity score in MM cells utilizing preclinical next-generation sequencing data and validate our approach using clinical data. Methods: CUT&RUN and ChIP-seq experiments were performed on H929 MM cell line to identify binding targets of Ik. To refine the target list, we investigated transcriptional changes of the Ik targets when exposed to lenalidomide (LEN), pomalidomide (POM), and mezigdomide (MEZI) at 6, 24, and 48-hour timepoints, and iberdomide (IBER) at a 24-hr timepoint. Differential expression (DE) analysis between compound-treated and DMSO-treated samples was performed using DESeq2, and significantly DE genes were determined. Ik binding target results were overlapped with the DE genes to generate a high-confidence Ik regulon: 115 up-regulated and 95 down-regulated targets. Results: Gene Set Variation Analysis was performed on Ik regulon to derive an Ik activity score. The score was first applied to RNA-seq paired samples (N=7) at screening and at cycle (C) 2 day (D) 15 from patients treated with MEZI in the CC-92480-MM-001 study. On-treatment samples had significantly lower activity score compared to screening. The Ik activity score was then applied to IBER-treated patients from the CC-220-MM-001 study, comparing RNA-seq paired samples (N=41) at screening and at C2D15. On-treatment samples, similarly, had significantly lower activity score compared to screening. Responders (R) of IBER and MEZI showed a slight trend of greater change in activity score.To determine if the change in activity score could be used as a biomarker of response with progression-free survival (PFS), the IBER dataset was investigated. Using an optimal cutpoint, significant association was observed with higher change in activity score favoring longer PFS. Association analysis using IBER IHC paired, at screening and at C2D15 samples (N=55), similarly showed significant association with higher percentage of Ik degradation also favoring longer PFS. Validation using IBER samples (N=22) having both RNA-seq and IHC data resulted in significant association between PFS and RNA-seq based and IHC based scores. Conclusions: We compiled Ik targets and devised an RNA-seq based Ik activity score that is significantly associated with patient outcome and is comparable with the IHC score.