P-231: Whole genome sequencing reveals lower APOBEC-mutational activity in multiple myeloma patients with African Ancestry compared with those of non-African Ancestry

Introduction: It remains unknown if multiple myeloma (MM) in patients with African Ancestry (AA) has a different spectrum of genomic aberrations when compared to patients with non-African Ancestry. Previous studies in patients with AA suggest an excess of t(11;14) and a deficiency of TP53 mutations.

Methods: Together with the New York Genome Center / Polyethnic-1000 consortium, we performed whole genome sequencing (WGS) in patients with diverse ethnic backgrounds. Combined with our previously published WGS and CoMMpass, the cohort comprised 1488 WGS with 208 self-declared AA. We used a comprehensive pipeline for somatic variant calling (mgp1000), together with admixture analysis to estimate ancestry using geographically-distinct reference populations (Africa; AFR, Europe; EUR, America; AMR, East Asia; EAS and South Asia; SAS).

Results: We identified 5 clusters corresponding to single continent-level reference population, together with a cluster characterized by highly admixed individuals with no dominant ancestry. 5% self-identifying as AA were assigned to other clusters, while only 34% self-identifying as Hispanic were assigned to the AMR-dominant group.

The prevalence of t(11;14), t(4;14) and t(14;16) were not significantly different between clusters. For somatic mutations, the tumor mutational burden (TMB) was lower in AFR cluster (median 2.21 mutations per Mb) than the EUR cluster (2.94, FDR adj. p=4.3x10-6), without differences in high-incidence mutations including TP53. There were also no significant differences in copy number aberration.

The most striking difference was observed in mutational signatures; AFR had lower contribution from clock-like signatures (SBS1/SBS5) and lower APOBEC-mutational activity (SBS2/SBS13), which were responsible for the difference in TMB. Mutational timing analysis, using the corrected ratio between duplicated and non-duplicated mutations to estimate the timing of chromosomal gains, did not reveal differences in the estimated timing of cancer initiation between clusters. As no differences in age, tumor purity or coverage were observed, this suggests a different mutational rate between clusters.

Interestingly, WGS from the Pancancer Analysis of Whole Genomes dataset in bladder and cervical cancer, (associated with APOBEC activity in pathogenesis), also had lower APOBEC-activity in AFR-dominant clusters. Further, leveraging both single cell WGS (DLP+ platform) and a validation cohort from EUR-dominant patients (GMMG-HD6 trial), we found that 34% of AFR lacked any APOBEC activity at either clonal or subclonal levels, compared with only 16% in EUR (p=0.03) and < 5% in the GMMG-HD6 dataset.

Conclusions: In the largest series of WGS from racially diverse MM patients to date, integrating genomic data with comprehensive ancestry information, MM in patients with AFR-dominant admixture emerged as biologically different. Differential mutational signature contribution, particularly APOBEC- mutational activity, appears to result in distinct genomic evolution over time.