Introduction: Myeloproliferative neoplasms (MPN) are clonal hematopoietic stem cell (HSC) disorders characterized by overproduction of mature blood cells and increased risk of transformation to myelofibrosis (MF) and acute myeloid leukemia (AML), although the mechanisms driving disease progression remain elusive. Polycythemia vera (PV) is the most common clinical subtype caused by acquired mutations of JAK2. While PV generally presents as an indolent process, ~25% of patients will progress to MF or AML and JAK2 mutations alone do not account for MPN transformation. HMGA1 / 2 genes encode oncogenic chromatin binding proteins that are overexpressed in acute leukemia and drive clonal expansion in murine leukemia models. In preliminary studies, we found up-regulation in gene expression of both HMGA1/2 in HSC with disease progression in MPN. We therefore sought to test the hypothesis that HMGA proteins are critical drivers of transformation and rational therapeutic targets to treat or prevent PV progression.

Methods: We examined clonal evolution in HSC from 49 JAK2V617F -positive PV patients using standard and SNP-array karyotyping and a targeted resequencing panel of 163 genes linked to myeloid cancers. HSC clonal burden was assessed via JAK2V617F genotype analysis on single cells. We also measured HMGA1 / 2 in JAK2V617F -positive human AML cell lines from MPN patients (PV-AML cells), CD34+ cells from PV patients during chronic and transformation phases, and JAK2V617F transgenic murine models of PV (transgenic JAK2V617F) and PV-AML (transgenic JAK2V617F / MPLSV ; Blood 2015;126:484) using real-time quantitative RT-PCR. To elucidate the role of HMGA1/2, we silenced HMGA1 or HMGA2 via short hairpin RNA and assessed proliferation in human PV-AML cells. Finally, to investigate Hmga1 function in vivo, we crossed mice deficient in Hmga1 onto murine models of PV and PV-AML. We also compared leukemic engraftment in JAK2V617F -positive PV-AML cells with or without HMGA1/2 silencing.

Results: Both HMGA1/2 mRNA were up-regulated in all JAK2V617F -positive contexts. In primary human PV CD34+ cells, HMGA1 and HMGA2 were increased by 7- and 100-fold, respectively, compared to controls. There was a dramatic up-regulation in both HMGA1/2 in HSC of patients who transformed from PV to MF or AML compared to chronic phase PV when analyzed in cross-sectional samples from different patients or prospectively in selected patients. Overexpression of HMGA1 /2 also correlated with clonal dominance of JAK2V617F -homozygous stem cells, and additional mutations of epigenetic regulators, including EZH2 and SETBP1 . Similarly, when assessed in unfractionated bone marrow or tumor samples in transgenic mouse models for PV and PV-AML, Hmga1 / 2 were overexpressed compared to control littermates, with highest levels in the PV-AML mouse model. Silencing HMGA1 or HMGA2 in human PV-AML cells (DAMI, SET-2) dramatically halted proliferation. Moreover, leukemic engraftment was prevented in PV-AML cells after silencing HMGA1 or HMGA2. In preliminary studies with the transgenic PV-AML murine model, Hmga1 deficiency ameliorates the fulminant phenotype of aggressive leukemia and early mortality. PV-AML mice survived a median of 5 weeks (range 3-10; n=40) whereas PV-AML in Hmga1 deficient backgrounds survived a median of 8 weeks (range 7-16; n=7). The leukemic burden was decreased in mice with Hmga1 deficiency.

Conclusions: HMGA1 and HMGA2 are overexpressed in MPN, and higher levels associate with disease progression to MF and AML, in both human PV and in transgenic murine models of PV. Strikingly, silencing HMGA1 or HMGA2 halted proliferation and prevented leukemic engraftment in vivo . Preliminary data in Hmga1 deficient models suggest that Hmga1 is required for PV progression to AML. These data highlight a key role for HMGA proteins as critical drivers of PV transformation and suggest that HMGA1 /2 overexpression is a consequence of aberrant JAK/STAT signaling and epigenetic dysregulation. Our findings indicate that HMGA1 /2 overexpression may function as a necessary molecular switch for PV leukemic transformation. Therefore, HMGA proteins and their transcriptional pathways offer novel therapeutic targets aimed at the prevention of PV progression to MF and AML.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution