Next to their function in embryogenesis the clustered Hox-homeobox genes are also master regulators of hematopoietic differentiation and development. Perturbed HOX expression can be found in a significant percentage of acute leukemia. Recently we have shown that several members of the HOXA cluster transform primary hematopoietic cells and induce leukemia in vivo . In an attempt to define Hox downstream targets important for leukemogenesis we determined the gene expression pattern in hematopoietic precursor cells transformed by conditional HOXA1 and HOXA9 derivatives. Interestingly, amongst the several hundred genes controlled by each HOX protein individually we could identify a significant overlap of genes that responded to both. This group included c-Myb that has been shown before to be important for Hox-mediated leukemogenesis indicating that this approach is able to identify a candidate set of genes important for transformation by HOX proteins.
Within this collection the gene for the guanine exchange factor Vav2 caught our interest because it had been originally identified in a screen for bona fide oncogenes. Moreover it is known that Vav2 controls activity of the small Rho-type GTPase Rac1 that in turn is an important mediator of (leukemic) stem cell mobility and engraftment. ChIP experiments confirmed HOXA9 binding in the putative vav2 promoter region. In addition a luciferase reporter placed under control of the vav2 promoter showed Hox-responsive behavior indicating that vav2 is a direct Hox-target. Concomitant with vav2-levels also Rac1 activity was under strict control of Hox factors. This was true also for HOX+Meis double transformed cells. To investigate the importance of vav2 for Hox-mediated leukemogenesis hematopoietic cells from vav2 knock-out mice were investigated in detail. vav2−/−cells could be transformed by HOXA9 (+Meis) in vitro with identical efficiency as their wild-type counterparts. Colony formation in replating assays, the capability to induce outgrowth of permanent myeloid precursor lines and the differentiation state of these lines did not differ significantly amongst the wt and vav2−/− samples. However, no active Rac1 could be detected in vav2−/− cell lines transformed by HOXA9. Importantly, retroviral re-expression of vav2 could rescue part of this defect. Obviously, Vav2 constitutes the major GEF controlling Rac1 activity in myeloid precursors despite the presence of the homologous proteins Vav1 and Vav3. Because Rac1 has been implicated in surface receptor recycling and in stem-cell homing we checked the level of the SDF1 “homing-cytokine” receptor CxCR4 (CD184) on Hox-transformed wt and vav2−/−cells and could confirm a complete absence of CD184 on the latter. Again reintroduction of vav2 rescued part of this phenotype. Transplantation experiments are underway and preliminary results indicate that loss of vav2 is accompanied by an impaired ability for extramedullary hemopoiesis, an import determinant of leukemia aggressiveness. Our results explain how Hox genes that are highly expressed in stem- and precursor cells can regulate the important process of stem cell homing. Rac1 inhibitors are known and these substances have been tested  in mice with special success on MLL fusion induced leukemias that are a paradigma for Hox-induced leukemogenesis. Therefore our results do not only provide the molecular framework for this observation but they are also able to identify a potential patient collective that may benefit from pharmacological Rac1 inhibition.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.