Whole Exome Sequencing and Targeted Sequencing Reveal the Heterogeneity of Genomic Evolution and Mutational Profile in Smoldering Multiple Myeloma

Introduction: Recent data shows that multiple myeloma (MM) almost always arises from precursor states called Monoclonal Gammopathy of Undetermined Significance (MGUS) and Smoldering Multiple Myeloma (SMM), but not all patients with MGUS or SMM develop MM. Risk factors of progression for SMM patients are largely based on tumor load as represented by an M-protein ≥ 3 g/dL, a free light chain (FLC) ratio outside the range of 0.125 to 8, and ≥ 10% plasma cells in the BM. However, the genetic lesions that underlie progression, the molecular factors that cause rapid versus slow progression, and the factors that distinguish the relatively indolent MGUS from SMM are not well known. Further, the genomic landscape of SMM is not well characterized. One potential factor is MYC overexpression. Bergsagel et al. have found that MYC levels increase when comparing MGUS, SMM and overt MM. Other frequently altered pathways in MM are NF-kB, MAPK and DNA damage. In addition, limited studies of paired SMM and MM samples show that in many cases, the aggressive subclones can already be detected, in small cell fractions, before overt MM develops. However, the cause of progression to MM is unclear, in large part because sequential genomic studies of MGUS/SMM progression have yet to be undertaken. To address these questions, in this study we examine clinically-annotated samples from patients with SMM.

Methods: We performed whole exome sequencing (WES) (mean target coverage 50X/100X) on 49 germline-tumor matched samples from patients with SMM (DNA from bone marrow CD138+ plasma cells matched with germline DNA from peripheral blood mononuclear cells). Libraries were constructed using Agilent SureSelect XT2 library prep kit, and hybridized to Agilent's whole exome V5+UTR capture probes and then sequenced on HiSeq 2500 (Illumina). We also performed targeted deep sequencing using a custom enrichment bait set on 25 samples of progressor (n=12) and non-progressor (n=13) SMM samples. Libraries were also constructed with Agilent SureSelect XT2 library prep kit and enriched by hybridizing to an in-house designed customized target bait, then sequenced on HiSeq 2500. Sequencing data were analyzed using previously established analytic pipelines including MuTect, RecapSeg, GISTIC, MutSig, and ABSOLUTE.

Results: The number of Somatic Single Nucleotide Variants (SSNVs) seen in SMM ranged from 1 to 98 nonsilent mutations with an average of 1.14 mutations/Mb, which is slightly lower than MM (1.6 mutations/Mb) from previous studies (p-value=0.05). This large and varying range of mutational load among samples suggests that SMM is likely a heterogenous entity where some patients are closer to MGUS and others closer to MM. We identified likely drivers in SMM in about ~32% of the samples, including mutations in MM candidate driver genes such as NRAS, KRAS and PTPN11(overall 36 events were present in COSMIC). SMM also had somatic CNAs in about ~50% of SMM samples, such as hyperdiploidy, gain of chromosome 1q, deletion of 13p and 17p, which match the hallmark chromosome changes seen in MM.

Comparing deep targeted sequencing of 100 genes (mean target coverage 361X) in samples from 12 SMM patients who progressed to myeloma vs. 13 SMM patients who did not, we found non-synonymous mutations exclusive to progressors, suggesting that with more samples we may find genetic alterations that predict progression in SMM.

Conclusion: This study demonstrates that WES and targeted sequencing of SMM patients can detect MM candidate driver genes as well as hallmark CNAs seen in MM patients. Further, there may be potential different mutational features between progressors and non-progressors. Thus, this approach can be used to identify genetic drivers of clonal progression from MGUS/SMM to MM that may present opportunities for early therapeutic intervention and prevention of disease progression.