“Let”-ing go with clonal expansion?

In this issue of Blood, Ikeda et al report the effects of the high mobility group A2 (Δhmga2) lacking the 3′ untranslated region (UTR) in hematopoietic tissues.¹  Overexpression of Δhmga2 in transgenic mice resulted in “big mice with big blood”: the mice weighed more and had a myeloproliferative phenotype with increases in peripheral blood counts, splenomegaly, a hypercellular bone marrow and erythropoietin-independent erythroid colony formation.

The HMGA2 gene encodes the HMGA2 chromatin remodeling protein, which binds to AT-rich regions of chromatin, alters DNA structure, and orchestrates the assembly of protein complexes to regulate gene expression. HMGA2 is expressed predominantly during embryonic development with low or undetectable expression in normal, differentiated tissues. In humans, HMGA2 is located at chromosome 12q13 in an area frequently involved in translocations and amplifications in benign, mesenchymal tumors. Emerging evidence also points to an important role for HMGA2 overexpression in malignant tumorigenesis, including poorly differentiated solid tumors and some cases of leukemia.^2,3  In normal development and differentiated tissues, HMGA2 protein translation is negatively regulated by the let-7 family of tumor suppressor microRNAs through multiple let-7 DNA binding sites in the 3′UTR. Indeed, loss of repression by let-7 appears to be an important mechanism whereby HMGA2 (and oncogenes such as cMYC, RAS, CCND1) are induced in some tumors. Thus, loss of let-7 tumor suppressor regulation enables diverse oncogenes to be expressed without the normal microRNA constraints. This could occur either by repression of let-7 tumor suppressors themselves or through genetic alterations involving the 3′UTRs where let-7s bind. There is evidence that both mechanisms (repression of let-7 or loss of 3′UTR binding sites) occur in human cancer.

In hematologic diseases, HMGA2 has been implicated in a variety of clinical scenarios. First, overexpression of HMGA2 was reported in patients with myeloproliferative neoplasias (MPNs).⁴  These patients all had translocations or inversions involving chromosomal bands 12q13-15 that resulted in overexpression of HMGA2. Second, chromosomal rearrangements causing a truncation in the 3′UTR of the HMGA2 gene has been reported in 2 PNH patients.⁵  These rearrangements delete the let-7 binding sites and cause overexpression of a full-length or truncated HMGA2 protein with preserved DNA binding capacity. Lastly, in human gene therapy trials, proviral insertion into the HMGA2 locus has occurred, removing suppression by let-7 microRNAs and leading to the clonal outgrowth of cells with this insertion.^6,7  Based on these intriguing findings, Ikeda and colleagues investigated the consequence of HMGA2 overexpression in a murine model. They engineered transgenic mice expressing full-length murine Hmga2 cDNA with a truncation of its 3′UTR. The transgenics express high levels of Hmga2 in the liver, bone marrow, spleen, and thymus. Similar to previous murine models misexpressing Hmga2, these mice weigh more than controls. Hematologically, the mice exhibit a myeloproliferative phenotype with increases in most cells (except lymphocytes) in the peripheral blood associated with a hypercellular bone marrow and hyperproliferative spleen. Hematopoietic cells from the transgenic mice showed a growth advantage over wild-type cells in competitive repopulation assays and serial bone marrow transplant experiments, indicating that forced expression of ΔHmga2 leads to a proliferative growth advantage in hematopoietic stem and progenitor cells. The authors speculate that overexpression of Hmga2 could explain the clonal expansion of PNH cells in selected patients with PNH, and may also contribute to clonal disease in patients with MPNs.

Ikeda and colleagues' careful and well-executed experiments represent an important advance in our understanding of the role of HMGA2 dysregulation on hematopoiesis. Their interesting findings also raise additional questions. Specifically, is there a role for HMGA2—or downstream pathways—in clonal expansion in PNH outside of the 2 previously described patients with chromosomal rearrangements involving HMGA2? PNH is a clonal hematopoietic stem cell disorder caused by a mutation of the X-linked gene PIGA.⁸  The PIGA gene mutations result in loss of glycosyl phosphatidylinositol (GPI)–anchored proteins on the surface of blood cells, which explains many of the clinical features of PNH, especially the hemolytic anemia; however, the mechanism of clonal dominance remains unclear. It appears likely that dysregulated HMGA2 expression contributed to clonal expansion in hematopoietic stem cells with the PIGA mutation in these 2 cases, although it is not clear that HMGA2 plays a role in clonal expansion in other PNH patients.⁹  Notably, recent attempts to detect aberrant HMGA2 expression in PNH patients have been unrevealing, suggesting that HMGA2 is not a major mediator of clonal expansion in these patients.⁹  Moreover, most PNH patients show evidence of bone marrow failure rather than a myeloproliferative phenotype. Splenomegaly and erythropoietin-independent erythroid colony formation are not typical features and most patients have pancytopenia. Because HMGA proteins function by modulating gene expression, it is possible that a common, downstream transcriptional network is associated with clonal expansion in PNH and other hematologic disorders. In fact, some downstream transcriptional targets of HMGA2 are known to be involved in embryonic development, stem cell functions, or cell-cycle progression, such as Snail, Slug, or Cyclin A. Here, Ikeda et al assessed gene expression profiles as well as STAT3 and AKT signaling at the protein level, and found evidence for increased expression of jak2 mRNA and protein in the hematopoietic stem cells from the ΔHmga2 mice, suggesting that jak2 signaling contributes to clonal expansion in this model.

Another pressing question raised by this and related studies is why HMGA2 appears to be a common integration site for retroviral/lentiviral vectors. Although it was originally thought that retro/lentiviral integration would be random, the emergence of nonrandom integrations and even leukemic transformation in gene therapy trials has taught us otherwise.^3,4  From these clinical events, it appears that the integration of retro/lentiviruses will depend, at least in part, on chromatin structure and it is possible that chromatin near the HMGA2 gene is conducive to viral integration in humans. Future studies are needed to address this important question. In addition, it will be interesting to determine how common dysregulation of HMGA2 and/or HMGA2 pathways are in MPNs and other hematologic diseases. Fortunately, the present study provides a useful model to further investigate HMGA2 function in hematopoiesis.