S82 EXABS-186-CML Epigenetic Dysregulation of CML S Tiong Ong, MD1,2,3,* 1Cancer & Stem Cell Biology Programme, Duke-NUS Medical School, Singapore 2Department of Hematology, Singapore General Hospital, Singapore 3Division of Hematologic Malignancies and Cellular Therapy, Department of Medicine, Duke University Medical Center, Durham, NC, USA. *Corresponding author: [email protected] Keywords Biomarkers, blast crisis, drug resistance, DNA methylation, polycomb repressive complex Introduction Disease progression in CML is characterized by the acquisition of recurrent genetic events that are thought to contribute to drug resistance and eventual blast crisis (BC) transformation.1 Because of the large number of reported gene mutations, as well as their pleiotropic effects on gene expression (GE) and cell behavior, this model poses significant challenges for detecting early-stage progression and the development of therapeutic interventions using targeted therapies. In contrast to genetic events, BC transcriptomes are remarkably similar, and converge on a common set of hallmark GE signatures encompassing increased stemness, inflammation, relative quiescence, and altered survival, which we term the core resistance transcriptome (CRT).2 Together, these observations suggest that epigenetic dysregulation,3 rather than genetic mutations per se, are the major contributor to hallmark GE signatures in BC, and the ensuing adverse biologic and clinical outcomes. Accordingly, understanding the epigenetic events which shape CML drug-resistance and progression transcriptomes are key to developing comprehensive biomarkers for identifying at-risk patients requiring closer monitoring and changes in treatment.4 Material and Methods To comprehensively identify genetic and epigenetic events which contribute to the CRT in an unbiased manner, we performed integrated multiomics analysis of 74 chronic phase and BC samples using whole genome and exome sequencing, transcriptomic and DNA methylation profiling, as well as ChIP-sequencing.2 Results We first confirmed that genetically heterogeneous BC progenitors shared a common transcriptome that transcended individual somatic mutations. Interestingly, pathway-based analysis of genetic events demonstrated statistically significant enrichment for alterations of multiple members of the polycomb repressive complexes (PRC) 1 and 2, implying that events affecting PRC function confer clonal fitness under in vivo conditions which select for BCR::ABL1-independence, i.e. during TKI treatment. Notably, we found recurrent gain- and loss-of-function events affecting PRC1 and PRC2 components respectively, which we confirmed using gene set enrichment analysis. Next, analysis of the BC DNA methylome revealed a convergent methylome that was also agnostic to genetic mutations, but which was enriched for hypermethylation of established PRC target genes, which we confirmed by ChIP-seq. Our combined results suggested a model for disease progression characterized by altered PRC2 activity that selects PRC2 (EZH2)-bound genes (regulating differentiation and tumor suppressor activity within the CRT) for ‘permanent’ silencing by DNA methylation. In parallel, we document increases in PRC1 activity regulating a set of overlapping and distinct genes from those subject to EZH2-directed DNA methylation. This model predicts that: (1) A combination of PRC1 and DNA methylation inhibitors would be particularly effective in targeting the CRT and BC cells; and (2) Individuals with CP who are destined to transform will harbour GE signatures enriched for targets of altered PRC and DNA methylation activity. Consistent with our model, we found the combination of PRC1 (PTC-209) and DNA methylation (decitabine) inhibitors was particularly effective against BC progenitors, including extinguishing their stemness function. We also found that, using two independent CP datasets, specific CRT genes (silenced by PRC and DNA methylation in BC) were predictive for poor outcome at the time of CP. Next, using our recently-generated single cell atlas of diagnostic CP bone marrow, we confirmed that CRT GE signatures could be recovered in pseudo-bulked CD34+ cells in CP patients who subsequently underwent BC transformation.5 Turning to the effect of genetic mutations on GE, we also assessed the contribution of RUNX1 and other recurrently mutated genes to the CRT, and identified mutation-specific functional effects that were ‘add-ons’ to the CRT.6 Ongoing work in our group is using single cell approaches to identify cell states which give rise to drugresistant clones, as well as the molecular processes at play within putative BC ‘cells of origin’. Conclusions and Discussion We developed a molecular model of drug resistance and disease progression in CML that facilitated the identification of effective combination strategies in BC, as well as novel biomarkers to identify early-stage drug resistance in CP patients. This model highlights the contribution of epigenetic dysregulation to drug resistance transcriptomes, while also incorporating the contribution of genetic mutations. Our results also suggest that epigenetic therapy, given early enough in the course of disease progression, may also prevent the emergence of drug-resistant clones and fullblown BC.
RkJQdWJsaXNoZXIy MTk3NTQxMg==