Analysis of the Sub-Clonal Structure of Smoldering Myeloma over Time Provides a New Means of Disease Monitoring and Highlights Evolutionary Trajectories Leading to Myeloma

Background: Smoldering myeloma (SMM) is an asymptomatic plasma cell disorder, distinguished from monoclonal gammopathy of undetermined significance (MGUS) by a higher risk of progression to symptomatic multiple myeloma (MM). Studying the genetic makeup and sub-clonal architecture of bone marrow samples taken from the same case sequentially over time is an innovative strategy to define the evolutionary trajectory underlying myeloma initiation and progression through SMM to MM and may provide new strategies to identify progression and to intervene therapeutically before end organ damage develops.

Methods: Sequential samples from 9 SMM patients (53 samples) with a median follow-up of 7 years (range: 3.5 to 12.8 years) were analyzed. DNA was obtained from CD138+ cells from the bone marrow of SMM patients. 100 ng of DNA was fragmented, end-repaired, and adapters ligated, before hybridization using MedExomePlus (Nimblegen) with an additional capture for the IGH, IGK, IGL, and MYC loci. After PCR amplification hybridized libraries were sequenced on a NextSeq500 (Illumina) using 75 bp paired end reads. The median coverage was 93x (IQR 86-105) and 100x (IQR 95-103) for tumors and controls, respectively. Variant, translocations, and copynumber changes were called using Variant Effect Predictor (v.85), Manta (v0.29.6), and Sequenza respectively. Sub clonal architecture was determined using the Pyclone package and nNMF performed using the NMF package in R.

Results: The median number of mutations per sample was 79 (range: 34-236) and increased with time from diagnosis with a trend suggesting that the mutation rate of progressors (n=6) was higher than of the non-progressors (F=3.9, p=0.052). Samples with hyperdiploidy had a higher mutational rate than other subgroups (F=9, p=0.009) in relation to higher DNA contents. We previously defined a set of 63 genes that drive myeloma; 7/9 patients had a mutation in one of these genes, independently from progression status. Four patients had more than one driver mutation, which were in different clones in two patients and in the same clone in two patients. The acquisition of bi-allelic inactivation of myeloma drivers immediately before progression was seen in genes such as DIS3 and TRAF3 indicating a role in progression to an active disease state.

Translocations were detected in six patients from the initial time point. In one case, a t(8;14) was detected during follow-up, 5.9 years from diagnosis. Quantification of the rearranged MYC allele compared to the IGH rearranged locus was performed by ddPCR. This t(8;14) was not present at diagnosis, appeared in a small fraction (1%) 4.1 years after diagnosis and steadily increased over time reaching 45% in the last sample, 8.9 years from the initial diagnosis indicating growing dominance of a potentially progressive clone.

It was possible to reconstruct the sub-clonal structure and how it varied overtime for eight patients. This analysis identified a median number of seven sub-clones per patient, most of them related via branching evolutionary patterns (7/8). In one case a linear pattern was identified. Ninety-five percent of the tumor contents was occupied by five clones in 6/8 cases, and six in 2/8 cases. The median number of minor clone (<10% of tumor content) at diagnosis was estimated to be 3 (range: 1-5). In 7/8 patients a minor clone increased to more than at least 15% of tumor content and in 5/8 patients at least 20%. All patients that had more than 2 minor clones that increased to more than 15% progressed or had progressed (4/8). The only patient that progressed and did not display these clonal changes progressed within 4 months from the initial SMM sample, suggesting the clonal sweep had already occurred. Significant changes in sub-clonal structures were also seen in all samples at least one year prior to progression.

Conclusion: A comprehensive analysis of multiple SMM samples over time offers new insight into the mechanisms of progression of SMM to MM including the role of events we have identified previously associated with relapse e.g. MYC translocations, clonal sweeps, and biallelic deletions and changes in the clonal architecture. Changes in sub-clonal structure occurred before progression providing a new tool to monitor SMM.