Paper Details

ABSTRACT

Introduction: Multiple myeloma (MM) is a hematologic malignancy of terminally differentiated plasma cells driven by complex genetic alterations, yet the functional contribution of non-coding mutations remains poorly understood. Most genomic studies have focused on protein-coding alterations, overlooking potentially pathogenic mutations within cis-regulatory elements that may disrupt gene regulation. We therefore identified recurrent non-coding mutational hotspots in MM genome and their effects on epigenetic regulation.

Methods: Whole-genome sequencing data from 157 MM patient samples were analyzed to identify recurrent non-coding mutational hotspots. Somatic variants were filtered to exclude those mapping to repetitive, low-complexity, and centromeric regions, followed by iterative clustering to detect genomic regions containing at least five variants. These recurrent clusters were defined as mutational hotspots. Detected hotspots were prioritized using a negative binomial model accounting for hotspot length, mutation burden, and background mutation rate, followed by Monte Carlo-based genomic annotation enrichment analysis. Candidate hotspots were functionally characterized using a high-throughput lentiviral massively parallel reporter assay (Lenti-MPRA). Selected patient-derived variants were introduced into MM cell lines using homology-directed repair (HDR)-mediated CRISPR–Cas9 editing, followed by correlative genomic analyses and functional assays.

Results: A total of 1,284 recurrent hypermutated regions (mutational hotspots) were identified across the MM genome. Annotation enrichment analysis demonstrated significant enrichment of hotspots within MM super-enhancers (6.1-fold) and promoters (5.4-fold), supporting the preferential targeting of regulatory elements by recurrent non-coding mutations. As expected, 125 hotspots mapped to immunoglobulin loci with somatic hypermutation and class-switch recombination and were enriched for canonical and off-target AID mutational signatures; and hence excluded as physiological B-cell mutagenesis events. The remaining 1,159 hotspots included known MM coding driver genes, including NRAS, KRAS, BRAF, TP53 and DIS3, together with additional novel candidate regulatory regions.

Among the top 30 statistically significant hotspots, we identified an understudied hotspot overlapping the 3′UTRs of two neighboring genes, BMP6 and TXNDC5 (Padj = 3.13 × 10?77), which localized to predicted super-enhancer elements (SE) across multiple MM cell lines. Two additional hotspots were identified within the TXNDC5 first intron (Padj = 4.27 × 10?²⁵) and a downstream intronic region (Padj = 1 × 10?¹⁶), indicating TXNDC5 as a recurrently mutated non-coding driver in MM. Although TXNDC5 has been implicated in SE rearrangements, the biological impact of recurrent mutations at this locus remains unknown. In total, 40/157 samples had genetic alterations at this locus, including 10 with inter-chromosomal translocation partners such as MYC, resulting in SE shuffling between loci. The presence of mutations in the 30 samples without translocations may suggest an independent mechanism of gene expression dysregulation. Consistent with these findings, patient RNA-seq demonstrated lower TXNDC5 expression in mutated than in non-mutated samples (log2FC = −0.20, P = 0.036), supporting mutation-associated transcriptional deregulation.

To investigate the functional consequences of these mutations, four patient-derived variants were introduced into the KMS11 MM cell line by HDR CRISPR–Cas9 editing. TXNDC5 mutants had reduced TXNDC5 expression compared with wild-type cells at both the RNA (log2FC = −0.72, P = 0.0025) and protein levels. To determine whether decreased expression was due to epigenetic changes, we performed Micro-C and chromatin profiling and identified high-confidence reduced H3K27ac peaks within the TXNDC5 topologically associating domain (MACS2 bdgdiff score ≥30) in mutant KMS11 cells relative to wild-type controls, supporting disruption of local SE activity while preserving higher-order chromatin organization.

Conclusion: Our study identifies several recurrent non-coding regulatory hotspots in MM and demonstrates that mutations at the TXNDC5 super-enhancer disrupt gene regulation independently of structural rearrangements. Continued in vitro and patient-derived xenograft (PDX) studies will further establish its contribution to disease pathogenesis and evaluate its potential as a therapeutic vulnerability.