Chromothripsis drives high-risk multiple myeloma by copy number abnormalities and rewiring the epigenome to activate oncogenic programs.

Abstract

Abstract Background: Chromothripsis, a catastrophic event of chromosomal shattering and reassembly, is a key driver of oncogenesis and is associated with poor prognosis in multiple myeloma (MM). However, the specific functional consequences and the molecular mechanisms by which it rewires the MM regulatory landscape remained largely unknown. Methods: Using previous definitions, we identified chromothripsis (≥10 breakpoints, ≥7 CN oscillations, ≥200 kb segment sizes), chromothripsis-like (5-9 breakpoints and 3-6 CN oscillation) and non-chromothripsis (WT) samples. We utilized multiomic data from the CoMMpass dataset (n=1,066) and an independent Indiana University (IU) dataset (n=134), as well as 14 patient-derived xenograft (PDX) models with short-/long-read whole genome sequencing (WGS, average depths 92x and 16x) to identify single nucleotide variations (SNVs), copy number (CN) abnormalities and structural variation (SV), as well as expression (>124 million reads), chromatin states (Cut&Tag-IT, >10M reads, H3K27ac, H3K4me1, H3K4me3, H3K9me3, H3K27me3, and H3K36me3), and Micro-C/LinkPrep (Dovetail Genomics; >900M reads) to identify 3D chromatin architecture/folding inside the cell and enhancer-target interactions. Results: We identified 417 chromothripsis events in 268 patients (25%) from the CoMMpass dataset. Events were identified more frequently in relapsed and refractory (RRMM) samples compared to newly diagnosed (NDMM) samples (30% vs. 22%), suggesting an association between chromothripsis and disease progression. The number of events found in chromosomes highly correlated with chromosome sizes (r=0.63, Pearson correlation). However, events were particularly enriched on chromosomes 1, and 17 (0.26 and 0.30 events per million bps). Breakpoints of these events were enriched on 17q and 1p (odds ratio=2.2 and 1.7, p=0.0003 and 0.001). Chromothriptic samples were associated with inferior progression-free (PFS; 970 vs. 1227 days, p=0.003) and overall survival (2176 vs. not reached, p=0.006) compared to WT. In chromothriptic samples, 34% exhibited breakpoints on >1 chromosome and an higher number of affected chromosomes correlated with worse PFS (≥3 vs. 2, p=0.03; ≥3 vs. 1, p=0.02). Meanwhile, CN oscillations plus breakpoints was an independent factor associated with inferior PFS(p=0.05,: 700 vs. 1358 days). Del(TENT5C), del(TP53)and t(4;14) were enriched in the chromothripsis group (32%, 18% and 20%, odds ratio=2.6, 3.4 and 2.4, p<0.001) and are associated with genomic instability. We applied the same approach and further identified 19 chromothriptic samples from the IU dataset, where t(4;14) and del(TP53) were also enriched (odds ratio=2.6 and inf.) We identified differentially expressed genes (DEGs) between the chromothripsis group and a subset of WT group with balanced frequencies of high-risk events (e.g. t(4;14) & del(TP53)) to generate a chromothripsis signature. Pathway analysis on 37 consistent DEGs between the two datasets identified ‘G2/M checkpoint’ and ‘interferon α/γ responses‘(p<0.05), suggesting selection for a growth advantage and activation of the cGAS/STING pathway. A neuron network model for predicting chromothripsis was trained with DEGs and reached performance of AUC=0.90 and AUPR=0.87. We utilized multi-omics data from established PDX models to further examine the functional consequences of chromothripsis events. Among 9 events detected in 7 samples, 5 events exhibited significantly enriched APOBEC mutational signature and kataegis within events. Copy number change alone did not account for changes in gene expression within the chromothripsis regions, with 71% of genes dysregulated through neo-TAD formation and rewiring of the epigenome, specifically through juxtaposition of super-enhancers and allele-specific expression associated with DNA methylation changes. Key dysregulated genes included MYCN (Log2FC=6.3), the tumor suppressor PMAIP1 (Log2FC=-4.9), and cell adhesion proteins (EPCAM/CD326, Log2FC=3.5) that were rarely expressed in other MM samples. Conclusion: We revealed novel molecular characteristics of chromothripsis and the mechanisms caused it. We determined that the number of events and CN abnormalities are associated with poor outcome and that epigenetic states are reprogrammed within the events. These findings advance the clinical risk stratification and reveal unique therapeutic vulnerabilities tied to chromothripsis-associated epigenetic dysregulation.