Integrative Analysis of the Genomic and Transcriptomic Landscape of Relapsed/Refractory Multiple Myeloma Patients Treated With Venetoclax in Combination With Bortezomib and Dexamethasone: Biomarker Analyses From the Phase 3 BELLINI Study

Background: Multiple myeloma (MM) is a clonal plasma cell malignancy characterized by a complex genomic landscape with molecular and clinical heterogeneity. Venetoclax (Ven) is a selective, potent, oral BCL-2 inhibitor currently under investigation as a potential targeted therapy for the treatment of relapsed/refractory (RR) MM. Using a next-generation sequencing (NGS)-based approach, the mutational landscape and gene expression profile (GEP) of patients (pts) with RRMM in BELLINI was evaluated for potential impact on observed clinical outcomes in key biomarker subgroups.

Methods: BELLINI (NCT02755597) is a randomized, double-blind, multicenter Phase 3 study of bortezomib + dexamethasone (Bd) + either Ven or placebo (Pbo) in pts with RRMM who received 1–3 prior therapies and were sensitive or naïve to proteasome inhibitors. Pts were randomized 2:1 to receive Ven 800 mg/d or Pbo + Bd. The following biomarker analyses were performed by central laboratory assessments of CD138-enriched bone marrow mononuclear cells collected at baseline: BCL2 gene expression by quantitative PCR (qPCR), cytogenetic abnormalities by interphase fluorescence in situ hybridization (FISH), identification of somatic mutations by whole-exome sequencing (WES), and transcriptomic analysis by RNA sequencing (RNAseq). Correlation between mutational status, BCL2^high status (Harrison et al. Blood. 2019:134 suppl:142), cytogenetics, GEP, and survival outcomes were examined by pairwise t-test and by hazard ratio (HR) from Cox proportional hazard model.

Results: As of the data cutoff (13 Sep 2019), 291 pts were randomized (194 to Ven and 97 to Pbo); 262 pts (90.0%) were evaluable by FISH, 257 pts (88.3%) by qPCR, 197 pts (67.7%) by WES, and 184 pts (63.2%) by RNAseq. The mutational landscape of evaluable pts was highly heterogeneous (Figure 1A). MAPK pathway mutations were the most prevalent (62/197 pts; 31.5%); however, recurrent mutations in the NF-kB pathway (TRAF3 or CYLD) were also observed (12/197 pts; 6.1%). Within the MAPK pathway, the most frequently mutated genes were KRAS (39/197 pts; 19.8%), NRAS (8/197 pts; 4.1%), BRAF (8/197 pts; 4.1%), and NF1 (12/197 pts; 6.1%). Although a higher rate of pts with a MAPK mutation was observed in the t(11;14) and BCL2^high subgroup compared to pts without these biomarkers (39/101 pts [38.6%] vs 23/95 pts [24.2%]; P=.04), improvement in progression-free survival (PFS) with Ven+Bd was observed in pts with (HR=0.08; 95% CI, 0.02–0.31; P<.001) or without (HR, 0.31 [95% CI, 0.16–0.63]; P=.001) a MAPK pathway mutation.

High somatic mutation loads are associated with increased genomic instability, resistance to therapy, and decreased overall survival (OS) in MM. High tumor mutation burden (TMB), defined as ≥20 mutations/megabase of sequenced DNA, was observed in 12/197 pts (6.1%). Although there were a limited number of high TMB pts, reduced PFS (HR, 4.55 [95% CI, 2.33–8.88]; P<.001) and OS (HR, 3.89 [95% CI, 1.76–8.56]; P<.001) compared to low TMB pts was observed regardless of treatment arm. WES detected a recurrent BCL2 mutation (Ala43Thr) in 15/197 pts (7.6%). This BCL2 mutation has not previously been identified in MM and did not appear to affect survival outcomes among pts in either arm.

Supervised clustering of RNAseq data grouped pts into respective MM GEP (Zhan et al. Blood. 2006;108:2020) at the following frequencies: CD-1 (2.7%), CD-2 (8.2%), HY (40.2%), MF (1.6%), MS (10.9%), LB (5.4%), and PR (31.0%). As shown in Figure 1B, 16/20 pts (80%) allocated to the CD-1/CD-2 cluster had a t(11;14) mutation, but 17/20 pts (85%) and 2/3 pts (67%) in MS and MF GEP groups had t(4;14) or t(14;16) mutations, respectively. The HY GEP group was primarily (56/74 pts; 76%) hyperdiploid by FISH. Though a higher rate of pts with a HY GEP (50/98 pts [51.0%] vs 24/86 pts [27.9%]; P<.001) and a lower rate of PR GEP (11/98 pts [11.2%] vs 46/86 pts [53.5%]; P<.001) were observed in the t(11;14) and BCL2^high subgroups compared with non-t(11;14) and BCL2^low subgroups, respectively, improvement in PFS with Ven+Bd was independent of GEP classification.

Conclusion: Using an integrative approach of genomic and transcriptomic profiling, response to Ven+Bd in key biomarker subgroups was shown to be independent of concurrent mutations in key oncogenic signaling pathways and GEPs known to be prognostic in MM; however, high mutation load appears to be a clinically relevant risk factor that negatively affects survival in MM.