Reconstructing the Genomic History of Multiple Myeloma Precursor Disease: Novel Insights and Clinical Applicability

Introduction

Early indolent stages of Multiple Myeloma (MM), which include monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM), are heterogeneous diseases with overall rates of progression of 1% to 10% per year.

Mutations present in MGUS and SMM have been characterized by our group and others and improve current risk models. However, whether any genomic events are sufficient for an MGUS or SMM to transform to MM remains an open fundamental question.

To address this, we sequenced the largest cohort of MGUS and SMM genomes to date starting with a low-input DNA template, enriching for low burden of disease compared to previous studies. We report the translational implications of a large-scale characterization of indolent stages of MM after 2 years of follow-up.

Methods

Bone marrow (BM) samples were collected from 130 patients with MGUS and SMM after informed consent for research use (IRB #14-174). Plasma cells were selected from BM samples using CD138+ magnetic beads before DNA purification and library construction. Tumor and matched germline libraries were whole-genome sequenced (WGS) to 60X and 30X coverage respectively on NovaSeq 6000 S4 flowcells at the Genomics Platform of the Broad Institute. Analysis was performed with methods from the Cancer Program of the Broad Institute. All structural variants were manually reviewed and compared to fluorescence in situ hybridization (FISH) clinical laboratory reports.

Results

The median age of participants was 62 (range, 37 to 87), with 52% female. At the time of BM sampling, all patients were untreated; 24 patients diagnosed with MGUS (18%) and 106 diagnosed with SMM (82%) were homogeneously distributed across the IMWG 2/20/20 risk stratification stages (Low-risk: 32%, Intermediate-risk: 34%, High-risk: 34%).

Somatic mutation burden increased with progressive disease staging, with a median of 2,666 single-nucleotide variants in MGUS (range 429-6546), and 3,233, 4,080, and 5,111 in low, intermediate, and high-risk SMM, respectively (range 463-11083, 161-8408, and 2549-7218, respectively, P=0.007). Compared to non-progressors, progressors were found to have an increase of molecular clock signature, symbolizing continuous accumulation of mutations over time (Fold change 2.40, P=0.002), and this difference was independent of patient age and disease stage. Confirming previous studies, APOBEC signature was strongly associated with patients with MAF alterations (t(14;16), t(14;20), median in MAF, 24%, non-MAF, 5.8%, P=3E-5), and uniquely identified in one patient with MMSET t(4;14) translocation.

Catastrophic structural variants (SV) such as chromothripsis and chromoplexy were found in 7% of patients across 2/20/20 stages. Complex SVs were strongly associated with clinical progression to symptomatic MM (Odds Ratio (OR) 9.83, CI_95%[1.91, 53]). Being statistically independent of risk staging, these events are a potential new risk factor not captured by current clinical models, but successfully characterized by an unbiased genomic assay.

In our cohort, MYC was found rearranged with IgL, IgH, and other partners in 9% of patients. Strikingly, all but one case also harbored a RAS hotspot mutation, showing a strong non-random cooperation between RAS and MYC (P=0.005), not previously observed, which translated into increased risk of clinical progression (OR 7.75, CI_95%[1.26, 48.8]). Moreover, most MYC-abnormal tumors (translocation or complex rearrangement) were missed by clinical FISH testing.

WGS detected >99% of translocations and copy number abnormalities reported by FISH. In clinical reports with unknown translocations, WGS could identify translocation partners, including IgH-MYC and rare IgH-MAFA. In 82% of patients, an increased diagnostic yield was found by WGS compared to FISH using existing models, showing clinical applicability to replace molecular cytogenetics.

Conclusion

Analysis of the largest cohort of MGUS and SMM genomes to date was able to decipher the genomic history of MM, unraveling events including molecular clock signature, APOBEC, complex SVs, and MYC/RAS cooperation. This can further define a genomic risk model of disease progression independent of clinical tumor burden markers such as 2/20/20. This highlights the potential of WGS to solve the “missing predictability” of current clinical models of progression.