Unravelling the epigenetic drivers of chip – MDS/CMML – AML progression through novel methylation-sensitive regulatory elements

Abstract Epigenetic reprogramming, particularly aberrant DNA methylation, drives many myeloid malignancies. While some patients have identifiable methylation modifier mutations (e.g. DNMT3a, TET2, ASXL1, or IDH1/2), many develop malignancies without such mutations. Disease progression is marked by widespread methylation changes that disrupt hematopoietic differentiation, yet the specific regulatory elements affected remain poorly defined. Furthermore, studies focusing on promoter CpG island (CGI) methylation have shown weak correlations with gene expression. This has hindered development of robust, clinically relevant methylation-based biomarkers associated with regulatory elements. Linking DNA methylation to function requires identifying functional regulatory elements. We recently reported Methylation Mesa (MM; preprint PMID: 39483908), a novel class of narrow-width (∼45–300 bp) regulatory elements with conserved methylation signatures in human and murine genomes. MM function as epigenetic switches for gene activation independent of CpG Islands. Focal core demethylation at a MM induces local histone modifications and chromatin reorganization, establishing new interaction loops inducing gene expression. Precise CRISPR-based demethylation, using CRISPR-DiR, mediated targeted demethylation reactivating silenced genes such as the tumor suppressor P16, confirming a regulatory role. To investigate epigenetic progression in Chronic Myelomonocytic Leukemia (CMML), we profiled MM in 27 adults from the PREACH clinical trial (ACTRN12621000223831), a phase 2/3 non-randomized, open-label study of newly diagnosed CMML patients stratified by mutation. Patients with RAS-pathway mutations (NRAS, KRAS, CBL) received 24 cycles of Azacitidine (Aza; 75 mg/m² SC, d1–7) and Lenzilumab (Lenz; 552 mg IV; d1 & d15 of cycle 1, then d1 per cycle). Matched whole-genome bisulfite sequencing (~33x coverage) and RNA-Seq (~78M paired reads) were performed at diagnosis and after 4 treatment cycles. Our analysis focused on Mesa with a methylation delta of ≥20% compared to flanking regions. Our findings suggest that the number of Mesa triggered post-treatment can stratify patient response. Responders exhibited significantly more triggered Mesa than non-responders (mean 314,000 vs 141,000; p = 0.01, Mann-Whitney). Mesa with ≥50% positional conservation within groups revealed distinct responder/non-responder signatures. Pathway enrichment identified differential regulation in hematopoiesis (CSF1/3, IL3), TP53 (BID, TP73), Warburg signaling (FOSL1, miR-23), and PTEN signaling (TP63, miR-26). Investigations utilizing our Epigenetic Engine (EEngine), which integrates 592 curated epigenetic datasets, will refine the identification of Mesas within functional grammar contexts. These refined Mesa will define a biomarker signature that predicts CMML progression by how patients will respond to treatment, setting the stage for functional validation in ex/in vivo studies. Alternatively, preliminary research in Acute Myeloid Leukemia (AML) using a THP1 DNMT3a-R882 dox-inducible model has revealed 3 key observations regarding Mesa dynamics following R882 induction: 1) Mesa Methylation Regulation: R882mut results in Mesa loci failing to maintain basal methylation despite stable flanking methylation, implicating DNMT3amut in Mesa-specific gene regulation. 2) Metabolic Rewiring: After 3 days Aza treatment, DNMT3amut samples exhibited distinct biosynthetic and salvage pathway rewiring. 3) Aberrant Growth and Cell Cycling: Signaling and metabolic circuits promoting aberrant proliferation were activated by day 3 and sustained through day 10 of treatment. EEngine analysis highlighted PAK and heme signaling components for prioritized ex and in vivo validation. In summary, we highlight the potential of Mesa signatures to identify novel functional epigenetic elements in myeloid and other malignancies. In CMML, distinct Mesa signatures correlate with patient response to treatment. These signatures could pinpoint epigenetic events necessary for effective therapy. Similarly, in MDS and AML, identification of Mesa elements could identify functional DNA methylation sensitive elements accounting for induction of disease as well as a potential therapeutic approach. Such research aims to uncover functional elements and epigenetic biomarker signatures that are directly linked to disease stage, progression, and responsiveness to treatment