Whole Genome Sequencing of Unique Paired SMM/MGUS Progressing to MM Samples Reveals a Genomic Landscape with Diverse Evolutionary Pattern

Symptomatic multiple myeloma (MM) is preceded by an indolent expansion of clonal plasma cells, known as monoclonal gammopathy of undetermined significance (MGUS) and/or smoldering MM (SMM). Despite initial efforts using karyotype, copy number arrays and whole exome sequencing, the full catalogue of genomic events associated with SMM/MGUS progression to MM has not been investigated before in a comprehensive way. MATERIALS AND METHODS: We utilized paired bone marrow samples collected from 11 patents at initial presentation (SMM n=10 and MGUS n= 1) and subsequently at the time of their progression to MM. DNA from purified CD138+ plasma cells from these samples along with a matched normal sample was subjected to whole genome sequencing (WGS) using HiSeq X Ten machine. Somatic mutations, structural variation and copy number changes were analyzed using standard analysis pipeline at the Wellcome Trust Sanger Institute. RESULTS: The total number of somatic variants (substitutions and indels) observed at the asymptomatic stage [5780 (range 2599-7760)] was not significantly different than what observed in the corresponding MM samples [5954 (ranges 2824-8227), p-value = 0.1]. Clonal mutations in driver genes such as NRAS, BRAF and DIS3 were found in SMM/MGUS samples, suggesting they can represent early lesions in MM pathogenesis. Similarly, IGH translocations (3/11), hyperdiploidy (7/11) and recurrent chromosomal aneuploidies (i.e del1p and gain 1q) were found at asymptomatic stage with similar frequencies to MM, suggesting that SMM may share many genomic features with MM. To have a reliable estimate of the evolution of the subclonal structure of samples we used a Bayesian hierarchical Dirichlet process after correcting the VAF for the percentage of contaminating normal cells (Bolli et al, Nature Comms 2014) to group mutations with similar adjusted VAF into clusters that reflect the sub-clonal composition of the tumor. Using this approach, we found that all patients presented one or more clusters of sub-clonal variants, reflecting spontaneous evolution of the disease before diagnosis, even in SMM/MGUS phases of the disease. In the 11 patients we observed two main patterns of progression to MM. In the first, the sub-clonal architecture shifted significantly because of loss of sub-clone(s) and gain of other(s), in a branching evolution pattern likely driven by spontaneous acquisition of new genetic lesions conferring competitive proliferative advantage to a previously minor sub-clone. In the second pattern, there was no change of the sub-clonal architecture of the sample despite a clinical progression to a symptomatic stage. Interestingly, the pattern of change of genomic rearrangements and copy number events was in broad concordance with that of point mutations, suggesting that these classes of events act in a concerted way during tumor evolution. Overall, we found that 7/11 SMM/MGUS progressed to MM showing spontaneous evolution of the subclonal structure of the tumor, and those where generally characterized by a longer time to progression as compared to samples showing no evolution. Interestingly, we did not observe any correlation between the total number of substitutions, rearrangements, indels, copy numbers changes and time to transformation. CONCLUSIONS: We observed two different models of SMM/MGUS progression to MM. Cases characterized by no significant change in the clonal architecture of the tumor likely represented disease which had already acquired the required mutational change/s for disease proliferation, where only time was needed for the disease to reach the burden required to manifest as clinically symptomatic disease. Early identification of such cases could provide a rationale for early treatment in the SMM phase that could be clinically beneficial. Conversely, cases showing subclonal change representing an example of spontaneous Darwinian evolution where the clinical progression is driven by spontaneous acquisition of further genomic lesions in competing subclones. Here, the effects of early treatment would have to be weighed against the risk of hastening the outgrowth of chemoresistant clones.