Abstract

Immunomodulatory drugs (IMiDs) are commonly used in first-line therapy to treat multiple myeloma (MM), a plasma cell malignancy. Previous studies have shown that cereblon (CRBN) is critical for the efficacy of IMiDs, such as lenolidomide, pomalidomide and thalidomide, in MM. In a small cohort of 9 patients with lenalidomide-resistant MM, CRBN expression decreased 20-90% after treatment. In a MM cell line, acquired deletion of CRBN was the primary genetic event conferring IMiD resistance. Although CRBN expression is copy number (CN) dependent, CN aberrations affecting CRBN are thought to be rare, with less than 2% of patients exhibiting monoallelic deletion of CRBN. We analyzed clinical and molecular data from the Multiple Myeloma Research Foundation (MMRF) CoMMpass trial (NCT145429) IA11 release to better understand the molecular basis underlying CRBN-related IMiD resistance.

IMiDs were used in first-line therapy for 75% (728/971) of patients in the CoMMpass cohort. We identified five CRBN mutations in 5/891 (<1%) patients with available whole exome sequencing data at baseline, before therapeutic intervention. Four baseline CRBN mutations were nonsynonymous; two occurred in exon 3, and one each in exon 10 and 11, which are involved in forming the IMiD binding pocket. None of the four patients with both follow up data and CRBN mutations at baseline had relapsed, with follow up ranging from 18-30 months. Despite IMiD treatment in 68/84 (81%) patients between serial timepoints, no patient developed a CRBN mutation at relapse, suggesting that acquired mutations in CRBN do not play a large role in the development of IMiD resistance in treated patients.

RNAseq data was available for 32 serial patients of which 23 (72%) were IMiD treated. Only a subset of IMiD-treated patients (13/23, 57%) exhibited a decrease (ranging from 5-75%) in CRBN expression at relapse, however the decrease from baseline (mean = 41.4 transcripts per million (TPM)) to relapse (mean = 26.8 TPM) was significant (p=0.004) in this group. We analyzed the change in transcript expression of nine noncoding and four protein-coding CRBN transcripts from baseline to relapse in this patient subset and found four differentially expressed non-coding transcripts and two differentially expressed protein coding transcripts; CRBN-001 ENST00000231948 (17.4 versus 10.7 mean TPM; p = 0.04) which codes the full length CRBN protein, and CRBN-010 ENST00000450014 (3.3 versus 2.1 mean TPM; p = 0.02) which is missing exons 10 and 11. Reduced expression of CRBN-001 plays the greatest role in the overall reduction of CRBN and since it contains the IMiD binding motifs, likely also plays a greater role in CRBN resistance in IMiD-treated patients. Of the 13 patients who exhibited a decrease in CRBN expression at relapse, 10 had whole genome sequencing (WGS) CN data at both timepoints and 2/10 (20%) patients experienced CRBN CN loss from baseline to relapse. This suggests that CRBN CN loss may account for acquired IMiD resistance in a larger subset of patients than previously reported.

Across the cohort, 14/871 (2%) patients with WGS data exhibit CRBN CN loss at baseline, while 342/871 (39%) patients exhibit CRBN CN gain, primarily whole chromosome gains related to trisomies of chr3 in a subset of hyperdiploid patients at baseline. Survival analyses revealed that IMiD-treated patients with CRBN CN gains at baseline exhibit better progression free survival (PFS) (median = 43 months) and overall survival (OS) (median = 55 months) than non IMiD-treated patients with CRBN CN gains (median PFS = 24 months, p<0.0001, median OS not reached, p<0.001). Although not significant, IMiD-treated patients with CRBN CN gains exhibit better PFS than IMiD-treated CN-normal patients (median = 39 months), suggesting that patients with CRBN CN gains at baseline may benefit from an IMiD based treatment regimen. Overall, these results highlight that chr3/CRBN CN gain can be a predictor of good outcome in IMiD treated patients and demonstrate the clinical benefit of obtaining CRBN CN and expression data throughout a MM patient's disease course.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.