Development of hyperdiploidy starts at an early age and takes a decade to complete

Abstract

Nearly half of patients with multiple myeloma (MM) have hyperdiploidy (HMM) at diagnosis. Although HMM occurs early, the mutational processes before and after hyperdiploidy are still unclear. Here, we used 72 whole-genome sequencing samples from patients with HMM and identified pre- and post-HMM mutations to define the chronology of the development of hyperdiploidy. An MM cell accumulated a median of 0.56 mutations per megabase before HMM, and for every clonal pre-HMM mutation, 1.21 mutations per megabase accumulated after HMM. This analysis using mutations before and after hyperdiploidy shows that hyperdiploidy happens after somatic hypermutation. Prehyperdiploidy mutations are activation-induced cytidine deaminase and age/clock-like signature driven, whereas posthyperdiploidy mutations are from DNA damage and APOBEC. Interestingly, the first hyperdiploidy event occurred within the first 3 decades of life and took a decade to complete. Copy number changes affecting chromosomes 15 and 19 occurred first. Finally, mutations before initiating event affected chromosomes at different rates, whereas post-initiating event mutational processes affect each chromosome equally.

Graphical abstract

Figure 1.

Pre- and Post-HMM mutations. (A) Workflow diagram. (B) Mutational distribution from a sample at diagnosis on hyperdiploid chromosome. The x-axis shows VAFs of mutations, and the y-axis shows the density of mutations for a given VAF. Orange section represents the clonal mutations accumulated before HMM and amplified with copy number gain. Purple and green sections are combining clonal mutations accumulated before and after hyperdiploid event. Mutations in gray area represent subclonal mutations. Panel on right represents how each mutation contributes to the distributions on the left. VAFs given below right panel assuming 100% tumor purity and no sequencing errors. Distributions on the left account for the tumor purity and 80% confidence intervals. (C) Ratio between all clonal mutations accumulated in post- and pre-HMM. The x-axis shows ratios on HMM chromosomes, and the y-axis shows the ratio. (D) Predicted number of mutations per megabase accumulated (y-axis) before HMM (orange), after HMM (clonal; blue), and subclonal (gray; x-axis). (E) Absolute (top) and relative (bottom) contribution of 5 mutational signatures detected in patients with HMM. Contributions from each signature is color coded, and legend is shown on the top. The y-axis is absolute counts (top) and scaled proportions (bottom); each bar is representing 1 sample (x-axis). (F) Relative (y-axis) mutational signature contributions (x-axis) by categories (color-coded box plots). Chr3, chromosome 3; Dup., duplications; QC, quality control; SNV, single nucleotide variants.

Figure 2.

Timing of hyperdiploid events in MM. (A) Estimated age (x-axis) for first (blue) and last (red) HMM copy number gains in each patient. Patients are in y-axis, ordered from the earliest estimate (bottom) to late (top). Age at diagnosis is shown with black “+” sign on the right side. (B) Estimated age (x-axis) for each HMM chromosome. Similar to panel A, patients are ordered from bottom to top by the earliest to latest HHM timing. Chromosomes are color coded and shown with different colors, as shown in the legend. (C) Relative timing of each copy number gain on hyperdiploid chromosomes. Relative time 0 refers to the earliest possible time, and 1 refers to the latest. Gain segment sizes may be different for each gain event on each chromosome. (D) Number of mutations per megabase (x-axis) on hyperdiploid chromosomes (y-axis) for patients with HMM at diagnosis. Overall chromosome length was used for correcting the mutations per megabase. Chromosome or local copy numbers are not included in the analysis. (E) Number of mutations per megabase (y-axis) contributed by AID (left) and APOBEC (right) for each chromosome (x-axis). (F) Model of mutational increase from the first HMM event to diagnosis. MM growth separated into 3 stages, colored by different background colors. Age and number of mutations were used from mean age and estimated mutational loads shown in figures. Mutational processes active in each stage are shown on the top with arrows.