Modeling Myelodysplastic Syndromes In Mice By Altered Hoxa1 Spliceform Expression

Myelodysplastic syndromes (MDS) are a heterogeneous group of blood cell diseases with 30% of the patients developing acute myeloid leukemia (AML). The etiology of MDS is largely unknown and as a result there are no curative therapies. Mouse models that faithfully represent MDS are based on rare genetic abnormalities and currently represent only a very small subset of MDS patients.

Using in silico analysis we have identified that homeobox A1 (HOXA1) mRNA is significantly upregulated in 50% of MDS patients (1.8-fold increase, n=183 MDS, 17 controls, P<0.05). Importantly, the upregulation was observed across all subsets of patients, including those with a normal karyotype. Both human HOXA1 and mouse Hoxa1 are expressed as two different spliceforms generated by alternative splicing within exon 1 of the wildtype Hoxa1 (WT-Hoxa1). These spliceforms encode a full-length (Hoxa1-FL) and a truncated form (Hoxa1-T), the latter lacking the homeobox domain. Given that mutations in splicing machinery have been identified in up to 85% of MDS patients we hypothesised that deregulated HOXA1 spliceforms may contribute to MDS.

We identified that the Hoxa1 isoforms were differentially expressed in murine hematopoietic stem cells (HSCs) and progenitors. Retroviral overexpression of either WT-Hoxa1 (which generates both Hoxa1-FL and Hoxa1-T) or Hoxa1-T in murine bone marrow (BM) cells showed opposing effects on in vitro cell proliferation and colony forming cell (CFC) production respectively, suggesting that Hoxa1-T may negatively regulate Hoxa1-FL. We therefore generated a mutant Hoxa1 (MUT-Hoxa1), which expresses normal Hoxa1-FL but not Hoxa1-T, by oligomutagenesis at the splice site of Hoxa1-T. We transplanted recipient mice (n>28 per group) with bone marrow (BM) overexpressing either MUT-Hoxa1, WT-Hoxa1 or empty vector (MXIE) control.

All recipients of MUT-Hoxa1 BM developed peripheral blood (PB) macrocytic anemia (mean ± SEM: Hb (g/L): MUT-Hoxa1: 126±2.9**; WT-Hoxa1: 134±2.7; MXIE: 135±1.4; **P<0.01 vs MXIE), without leukocytosis. Furthermore, thrombocytopenia was also observed in MUT-Hoxa1 and WT-Hoxa1 recipients (PB platelets (x 10⁶/ml): MUT-Hoxa1: 738±88*; WT-Hoxa1: 777±57*; MXIE: 922 ± 37, *P<0.05 vs MXIE). Overexpression of MUT-Hoxa1 or WT-Hoxa1 was associated with significant apoptosis in BM erythroid cells (% apoptosis in GFP+ Ter119+ cells: MUT-Hoxa1: 20.7±4.9*; WT-Hoxa1: 25.3±5.2** MXIE: 13.6±2.5, *P<0.05, **P<0.01 vs MXIE). We also observed numerous dysplastic erythroid cells and micromegakaryocytes in the MUT-Hoxa1-overexpressing cells.

Strikingly, 88% of recipients of MUT-Hoxa1 BM, but no WT-Hoxa1 recipients, developed AML with a median time of 10.5 months post-transplant, independent of retroviral integration site (Kaplan-Meier survival curve, P<0.05 vs MXIE). Therefore, overexpression of MUT-Hoxa1 or WT-Hoxa1 in BM cells results in clinical features of MDS, with MUT-Hoxa1 overexpression resulting in a more aggressive MDS that spontaneously progresses to AML.

Analysis of BM from non-leukemic mice identified that MUT-Hoxa1 and WT-Hoxa1 recipients had significantly decreased HSC-containing lineage -ve, c-kit+, Sca-1+ (LKS+) cells (% of lin- GFP+ BM: MUT-Hoxa1 1.72 ± 0.28^#; WT-Hoxa1: 2.08 ± 0.37^#; MXIE: 5.09 ± 0.56%; ^#P<0.0001 vs MXIE, n>12). This was accompanied by increased progenitor cell-containing LKS- (% of lin- GFP+ BM: MUT-Hoxa1: 53.32 ± 5.65*; WT-Hoxa1: 47.42 ± 4.50; MXIE: 38.37 ± 4.13; *P<0.05 vs MXIE, n>12). Strikingly, granulocyte/monocyte progenitors were significantly reduced in the MUT-Hoxa1 LKS- cells, accompanied by a marked accumulation of megakaryocyte/erythroid progenitors (MEPs). We have identified deregulated gene pathways in the MUT-Hoxa1 BM MEPs by microarray analyses. We are currently testing drug candidates identified from these screens for their effects on BM obtained from our MUT-Hoxa1 and WT-Hoxa1 mouse models and MDS patients.

Taken together, these results suggest that overexpression of Hoxa1-FL results in MDS, and that concurrently inhibiting splicing of Hoxa1-T (MUT-Hoxa1) forms a more aggressive MDS phenotype. Preliminary analysis of CD34+ BM cells from MDS patients reveal significant numbers of patients with elevated HOXA1-FL mRNA compared to lower or absent levels of HOXA1-T. Our MUT-Hoxa1 and WT-Hoxa1 mouse models will therefore be highly valuable in identifying better therapies for a significant subset of MDS patients.