Tracking the Evolution of Therapy-Related Myeloid Neoplasms Using Chemotherapy Signatures

Abstract

Introduction Patients with clonal hematopoiesis (CH) are at risk for therapy-related Myeloid Neoplasm (tMN), especially patients with multiple myeloma (MM) undergoing high-dose melphalan and autologous stem cell transplantation (HDM-ASCT). Exposure to distinct chemotherapies, such as melphalan, introduces hundreds of mutations in human cells measurable as highly penetrant mutational signatures. The impact of these mutations on tMN development is poorly understood. Methods To characterize the genomic impact of distinct chemotherapies, we analyzed a cohort of 40 tMN whole genomes (WGS) from 39 patients exposed to cytotoxic chemo (and/or radiation). 13 of 16 patients who received HDM-ASCT had underlying MM. Results We compiled single nucleotide variants (SNV) and performed mutational signature analysis to identify the mutational processes in each tumor. Overall, only platinum and melphalan were associated with mutational signatures. All 10 tMN with prior platinum exposure bore a platinum signature (SBS31/SBS35) while only 6 of 16 post-HDM-ASCT (37.5%) bore the melphalan signature (SBS-MM1). As detection of chemo-signatures in bulk whole genomes is contingent on a single cell expanding to clonal dominance (Pich, Nat Gen 2019), this differential signature presence suggests that CH may escape exposure to melphalan via leukapheresis and be reinfused to expand to tMN. Support for this model comes from 1) confirmation via targeted sequencing (MSK-IMPACT) of pre-existent CH variants in 8/11 tested pre-ASCT samples/apheresis products that became clonal in tMN , 2) detection of platinum but not melphalan signatures in 3 neoplasms with sequential exposure to both agents, and 3) detection of SBS-MM1 in two tumors without an escape route (tMN post-oral melphalan and a newly sequenced post-HDM-ASCT transitional cell carcinoma). The mutational burden of melphalan signature-positive tMN was higher than those without (p=0.009). We next compared 298 de novo AML exomes (Beat AML) and 21 genomes (TCGA) to tMN genomes. De novo AML and tMN lacking chemo-induced mutagenesis shared a similar CNV landscape, while tMN harboring chemo-signatures were enriched for aneuploidies including deletions of 5q, 7q, 17p, and gains in 19p13.2 (FDR<0.01). Chromothripsis was enriched in tMN with chemo-signatures (p<0.001), with five patients having focal amplifications (median 7; range 5-12, copies) of the SMARCA4 gene locus. Furthermore, chemo-signature-positive cases had a higher prevalence of TP53 loss (10/16, 62.5%) compared to signature-negative cases (3/23, 13%; Fisher, p=0.002). Strikingly, among those receiving HDM-ASCT, all six cases with the SBS-MM1 signature had TP53 loss as compared to 2/10 (20%) without the signature (Fisher, p = 0.007). This stark difference indicates that loss of TP53 may allow precursors to survive direct exposure to a myeloablative dose of melphalan and acquire chemo-mediated aberrations, while precursors with unperturbed TP53 may preferentially be reinfused with ASCT and/or or be selected by milder chemotherapy. Finally, we used chemo-induced mutations duplicated with large chromosomal gains to establish if these events were either acquired or pre-existing (i.e., selected) in relation to chemotherapy. According to this new workflow, if chemo-mutations are duplicated within a chromosomal gain, they must have been present prior to the event (i.e., gain was acquired after exposure). Conversely, the presence of only non-duplicated clonal or subclonal chemo-mutations indicates the chromosomal gain occurred prior to exposure (i.e., gain was selected). We demonstrated that several aneuploidies and chromothriptic events are acquired following exposure to chemotherapy. Conclusion Only a fraction (37.5%) of tMN cases harbor the highly penetrant SBS-MM1 melphalan mutational signature as compared to 100% of post-platinum cases. This further elucidates a novel mode of malignant progression for tMN that is not reliant on direct mutagenesis or even direct exposure to chemotherapy where CH may be reinfused to expand to tMN post-myeloablative therapy. Conversely, in cases with chemo-induced mutagenesis, copy number gains and complex SV are acquired following cytotoxic therapy. For patients treated with HDM-ASCT, TP53-mutant CH-status appears to influence these routes of progression, providing insight into the interaction of HD-melphalan and CH for patients with MM. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal