In this study, we demonstrate that inducible genetic deletion of the only non-redundant element of the Hh cascade, Smo, was unable to affect adult hematopoiesis, specifically at the level of the HSC. Smo-deficient HSCs display normal abilities to differentiate, self-renew and regenerate the immune system. In agreement with these phenotypic and functional studies, gene expression profiling analysis demonstrated that HSC-specific gene expression “signature” was preserved in Smo-deficient HSCs. Interestingly, the simultaneous ablation of both the Hh and Notch pathways was also unable to affect HSC differentiation and function. Moreover, using a gain-of-function model, we found that Hh hyper-activation did not lead to expansion of the HSC compartment. Finally, Smo deletion had no effect on the ability of the Notch1 oncogene to transform early hematopoietic stem cells and progenitors and to induce T-ALL. All of these findings are of unique importance as they directly question the current consensus on the role of Hh signaling in adult hematopoiesis.
Our studies are in contrast to a recent report by Zhao et al. that also used a conditional Smo
allele deletion (Zhao et al., 2009
). One possible explanation for this discrepancy is the utilization of a distinct mode of deletion. Zhao et al. use the Vav-cre deleter strain that appears to be hematopoietic specific; however it is able to delete the Smo
alleles in both adult and fetal hematopoiesis. Indeed it was previously shown that the Vav promoter can efficiently drive Cre-recombinase expression in e.d. 13.5 fetal liver HSC (Stadtfeld and Graf, 2005
). It is possible that the reported HSC defects in the Vav-creSmoF/F
model reflect Hh signaling functioning not in adult but in fetal HSC function and hematopoiesis. Although future work is required to identify putative Hh roles in fetal hematopoiesis, our data clearly demonstrate that Hh signaling is dispensable for adult HSC function.
Our observations suggest that Hh hypeactivation is unable to expand bone marrow stem cells and progenitors, a conclusion that is inconsistent with a report by Trowbridge and colleagues (Trowbridge et al., 2006
). A potential reason for this discrepancy could be the utilization of different animal models. In the germline Ptch+/-
model both HSC and/or the HSC “niche” could contribute to phenotype, whereas in the inducible SmoM2 model expression of the activated allele is largely restricted to the hematopoietic compartment. Moreover, putative differences on the effect of Hh hyper-activation on HSC/LSK cell cycle progression could be explained by the differential analysis performed. Indeed, Trownbridge et al. study the cell cycle status of Ptch+/-
LSKs after transplantation while we study steady- state LSKs shortly after SmoM2 activation. Finally, it is important to note that the gain-of-function
of a pathway effector (SmoM2) may well engender a different hematopoietic phenotype than the loss-of-function
of a negative regulator (Ptch
) that may have effects on other signaling pathways that could influence hematopoiesis.
Our analyses also failed to demonstrate a significant effect of Smo
deletion on T cell differentiation, as proposed previously by several studies including one from our own laboratory, in which Smo
was deleted in early T cell progenitors using the Lck-cre strain (Crompton et al., 2007
; El Andaloussi et al., 2006
). This discrepancy could be due to the differential mode of Cre-recombinase activation and pathway deletion. Indeed, Lck-cre is only active in early thymocytes and it ensures deletion in both fetal and adult thymus, suggesting again that fetal and adult hematopoiesis has unique and distinct Hh signaling requirements. Another reason for the phenotypic discrepancy could be that the Lck-cre-driven thymic effect was only partial. It is thus possible that our current studies, aimed mainly at HSC function, were not quantitative enough to reveal slight alterations of early T cell differentiation. It is more difficult to explain the differential effects on thymic size and progression of T cell maturation of the Mx1-cre-mediated Smo deletion reported by El Andaloussi et al. The timing of the analysis could provide a potential explanation. In this study thymi were analyzed at week 4 and 16 post-deletion, while El Andaloussi et al analyzed mice only one week after the last polyI:polyC injection. It is thus possible that the outcome of these studies was dictated by the timing of the analysis, especially as the thymus is a tissue with enormous regenerative capacity. Additional explanations could also include background differences as the mice studied here (and by Hoffman et al.) are C57Bl/6 SmoF/F
while El Andaloussi et al. utilized 129X1/SvJ SmoF/null
animals. It is thus possible that the effects on T cell development were influenced by the genetic background of the analyzed mice. Future studies that directly compare T cell development in the different Hh deficient strains are necessary to address the extent of Hh function in T cell development.
Is there any role for Hh in hematopoiesis? The strongest evidence supporting a pivotal role of Hh signaling in hematopoiesis came from the study of zebrafish embryo Hh mutants. (Gering and Patient, 2005
). As zebrafish hematopoiesis shares striking similarities to the mammalian fetal blood development, it is possible that the Hh pathway, as previously suggested, plays a more prominent role during fetal blood development. Moreover, it is possible that Hh function is masked by the synergistic function of other signaling pathways. Indeed, pathways such as Notch and Wnt, which have been previously shown to be capable of interacting with Hh (Hallahan et al., 2004
; Mak et al., 2006
; Yang and Niswander, 1995
; Yokota et al., 2004
), could collaborate with each other to ensure self-renewal and specify differentiation (Duncan et al., 2005
). In this report, we showed that deletion of both RBPJ and Smo did not affect HSC function, suggesting Notch and Hh signaling do not play synergistic roles. However, we cannot exclude potential redundancy with other signaling cues.
Two recent reports (Dierks et al., 2008
; Zhao et al., 2009
) have identified Smo as a drug target for the targeting of BCR-ABL+
human leukemic stem cells, introducing the notion that the Hedgehog pathway could be important for malignant hematopoiesis and the maintenance of leukemia. In the light of these seminal findings, our results are of further importance as they prove that pharmacological targeting of Hedgehog in leukemia is feasible as physiological HSC function and progression of hematopoiesis remains unaffected. They also suggest that not all blood malignancies can be treated using similar therapeutic protocols as the progression of T-ALL is not affected by the silencing of Hedgehog function.