Targeting HMGA1-driven leukemic transformation in myeloproliferative neoplasms with pacritinib

Abstract Background: The transformation of myeloproliferative neoplasms (MPNs) to secondary acute myeloid leukemia (sAML) is a lethal event with a median survival of under six months. A critical unmet need exists for robust early biomarkers to identify high-risk patients and innovative therapies to intercept this progression, particularly given the limited efficacy of first-generation JAK inhibitors. High-mobility group A1 (HMGA1), an architectural chromatin factor, has been implicated in MPN pathogenesis, yet its precise role in leukemic transformation and its clinical utility remain incompletely defined. Methods We conducted a comprehensive investigation integrating multi-omics profiling (scRNA-seq, CITE-seq, ATAC-seq, CUT&Tag, RNA-seq) with analyzing 240 longitudinal patient biopsies across MPN stages. The clinical utility of HMGA1 was validated in independent cohorts (including OHSU BeatAML). Functional studies utilizing gain and loss-of-function leukemia models and patient-derived xenografts (PDX) elucidated HMGA1-mediated leukemogenic mechanisms and assessed therapeutic vulnerabilities to JAK inhibitors, including the next-generation JAK2/FLT3/IRAK1 inhibitor pacritinib. Results We identified HMGA1 as a critical driver and potent predictor of MPN-sAML transformation. HMGA1 protein expression, readily detectable by immunohistochemistry (IHC), demonstrated a stepwise increase from chronic MPN phases to sAML (P<0.0001). HMGA1 IHC exhibited exceptional diagnostic accuracy for sAML (AUC=0.96), outperforming conventional markers (CD34/CD117). Crucially, elevated HMGA1 (>40.78%) in non-blast phase MPN predicted impending leukemic transformation within 12 months (47% conversion rate), offering a median lead time of 6 months before overt blast crisis. Single-cell analyses confirmed that HMGA1 upregulation occurs early and persists across heterogeneous malignant populations, overcoming the limitations of transient stem cell markers. Clinically, high HMGA1 defined a high-risk transcriptional subtype and served as an independent predictor of poor overall survival in sAML (In-house HR=7.18, P=0.0019; BeatAML HR=3.04, P=0.0096). Furthermore, elevated HMGA1 correlated with resistance to first-generation JAK inhibitors (e.g., ruxolitinib) and lower complete remission rates. Mechanistically, HMGA1 expression is driven by synergistic cooperation between JAK2V617F and TP53 mutations. Integrative chromatin analysis (CUT&Tag, ATAC-seq) revealed that HMGA1 directly binds and maintains open chromatin at key cell cycle regulatory loci, including E2F targets and G2/M checkpoint genes (e.g., E2F1, CDK2, CCNB1/2). This transcriptional reprogramming promotes proliferation, blocks differentiation, and accelerates leukemic progression in vivo. We discovered that HMGA1 overexpression confers resistance to ruxolitinib and fedratinib by sustaining E2F/G2-M programs. Notably, pacritinib effectively overcame this resistance, suppressing these oncogenic networks, inducing cell cycle arrest and DNA damage (γ-H2AX) irrespective of HMGA1 status. In HMGA1-overexpressing xenograft models, pacritinib significantly reduced tumor burden (8.0-fold decrease in bioluminescence) and doubled median survival (38 to 79 days, P=0.0006). Conclusion Our findings establish HMGA1 as a superior, readily implementable biomarker for early predicting leukemic transformation and risk stratification in MPN. We identify HMGA1 as a central driver of aggressive disease and therapeutic resistance, and validate pacritinib as a rational, clinically available strategy to target HMGA1-driven malignancy. These results strongly advocate for integrating HMGA1 assessment into clinical practice and the prospective evaluation of pacritinib for high-risk MPN-sAML.