Deletion of HoxA Genes Inhibits Proliferation and Differentiation of Myeloid Progenitors.

Mutants for single homeobox genes are showing neglible to mild phenotypes in the hematopoietic compartment. This is likely due to functional redundancy between Hox genes that has been demonstrated in embryonic development by compound mutant models. Interestingly, our previous data showed that hematopoietic stem cells (HSCs) lacking the majority of the HoxB genes (Hoxb1-Hoxb9) remained fully competent in replenishment of the hematopoietic compartment (Bijl, 2006). In addition we reported the expression of the majority of the HoxA cluster genes in HSC enriched E14.5 fetal liver fractions at levels at least one log higher than for HoxB cluster genes, suggesting that not HoxB but HoxA genes have a physiological role in hematopoiesis at that stage in the ontogeny. To investigate further whether HoxA genes are required for definitive hematopoiesis we used a conditional knock-out mouse model, in which the HoxA cluster was flanked by loxP sites (HoxA^flox/flox). To excise the HoxA genes Cre recombinase in conjunction with GFP was retrovirally expressed in HoxA^flox/flox fetal liver cells (Ly5.2). Flow cytometry showed transduction efficiencies in total fetal liver of 40% for Cre-GFP and 60% for MSCV-GFP control vector. Transduced cell composites were both plated in myeloid colony forming unit assays and transplanted in congenic (Ly5.1) mice to evaluate the transduction rate in progenitor populations and their capacity to proliferate and differentiate. Cre⁺ progenitor cells were able to grow out in colonies, and cultures showed a transduction efficiency of 69% for Cre-GFP and 99% for control vectors at the level of progenitors. The excision of the HoxA locus was confirmed by PCR for two Cre-GFP colonies. However, Cre-GFP⁺ colonies were much smaller than control colonies, indicating a defect in the proliferation potential of Cre⁺ myeloid progenitors. Further expansion of GM colonies in liquid culture was less efficient for Cre⁺ (1 out of 15), than for controls (5 out of 24), underlining a major defect in proliferation of early myeloid progenitors with granulocytic and monocytic differentiation potential in the absence of HoxA genes. All the colony types (GEMM, GM-CFU, G-CFU, M-CFU and BFU-E) were present in the Cre⁺ culture, but GM-CFUs and M-CFUs were proportional less represented than in control cultures (17% vs. 26% and 40% vs. 48%, respectively), indicating that HoxA genes have a function in early myeloid differentiation. Moreover, all colonies were in general poorly differentiated, showing that myeloid differentiation is sensible to reduced HoxA levels particularly in progenitors for the granulocytic/monocytic lineage. The capacity of HoxA^−/− progenitors to induce small colonies might be due to some residual levels at the time of excision or in the case of mature progenitors a reduced requirement for HoxA genes, which is in agreement with the decreased expression of Hox genes with maturation of blood cells. No progeny of Cre-GFP⁺ cells was observed in irradiated recipients 4 weeks post-transplantation, which demonstrate complete failure of HoxA^−/− progenitors to repopulate. However, it is too early to conclude whether deletion of HoxA genes affects the activity of HSCs. Altogether our in vitro and in vivo data show that HoxA genes are essential for the proliferation and differentiation of early myeloid progenitors that seems not to be compensated by other Hox gene members. The effect of HoxA deletion on HSC activity is currently under investigation.