The Targeted Disruption of Kdm6a/Utx in Mice Has Gender-Dependent Effects on Primitive Hematopoietic Cell Populations and Myeloid Progenitors

Putative inactivating mutations of EZH2 (the histone H3K27 methylase) and KDM6A (a histone H3K27 demethylase) both occur in myeloid malignancies, including acute myeloid leukemia (AML). The mechanism(s) by which genetic inactivation of KDM6A contributes to leukemogenesis is not clear, and the role of KDM6A in normal hematopoiesis is largely undefined. To address the role of KDM6A in hematopoiesis, we generated a conditional knockout mouse of the Kdm6a gene on the X chromosome (with LoxP sites flanking the 3^rd exon) and crossed these mice with Vav1-Cre transgenic mice to inactivate Kdm6a in hematopoietic stem/progenitor cells. Mice were born in expected Mendelian ratios with no aberrant phenotypic abnormalities. We characterized normal hematopoiesis from young (6 to 8 week old) male and female Kdm6a conditional KO mice crossed with Vav1-Cre mice. We included both male and female (both homozygous and heterozygous Kdm6a KO mice) animals, since Kdm6a can have gender dependent effects, and human UTY (the KDM6A homologue on the Y chromosome) does have H3K27 demethylase activity (Thieme S et al., Blood, 2012 and Walport, L.J. et al., J Biol Chem, 2014).

Young female Kdm6a null mice had a mild thrombocytopenia relative to all other cohorts with an average platelet count of 423 K/uL +/- 48 (n=10) vs. 794 K/ul +/- 76 for the WT littermates (n=10), p=0.006. We also observed mild splenomegaly in both the male and female Kdm6a null mice. The splenomegaly was not associated with extramedullary hematopoiesis or a shift in progenitor or mature lineage cell populations within the spleen. There was no difference in other blood cell counts, bone marrow cellularity, body weight or thymus weight between cohorts of young mice. We did not detect significant differences in the global levels of 21 histone H3 or 10 histone H4 modifications, using a multiplex colorimetric assay from lysates of whole bone marrow obtained from these mice. However, using western blotting, we did observe a decrease in H3K27 acetylation in both male and female Kdm6anull mice.  

We detected an aberrant self-renewal phenotype that may be relevant for leukemogenesis, which is in contrast to a previously reported impaired colony-forming ability using a knockdown approach of Kdm6a (Liu J et al., Exp Hematol, 2012). Using a serial replating assay with myeloid progenitor conditions, we found that both male and female Kdm6a null mice produced significantly more colonies in the second round of replating relative to control mice. Moreover, the homozygous female Kdm6a null mice had a significantly increased number of colonies at week 2, compared to hemizygous male mice (note that both are deficient for Kdm6a, but the male mice may compensate for it because of Uty). Flow cytometry revealed a slight myeloid skewing in the bone marrow of young female and male Kdm6a null mice with increased numbers of Gr-1+ and Cd11b+ cells. We did not detect differences in other lineages except for a slight decrease in erythroid precursors (as determined by Ter119 staining) in Kdm6a null mice. We also quantified the primitive hematopoietic and myeloid progenitor subpopulations from the bone marrow of these mice. Young female Kdm6a null mice had a significant decrease in the KLS population, which contained a lower frequency of short-term HSCs and multipotent progenitors. We also detected a significant decrease in MEPs (consistent with the observed thrombocytopenia). In young male Kdm6a null mice, the KLS population is not altered. A competitive transplant experiment validated the known engraftment defect in female Kdm6a null donor mice (Thieme S et al., Blood, 2012). Finally, we established a tumor watch consisting of female and male Kdm6a conditional KO mice and their littermate controls. After 18 months of follow-up, we did not observe the development of leukemia or other overt hematologic disease in either male or female mice deficient for Kdm6a, compared to a previous report that suggested that the rapid development of myelodysplasia (Thieme S et al., Blood, 2012). The difference in phenotype may be explained by cell-autonomous vs. non-autonomous effects (inactivation of Kdm6a in our model is essentially limited to the hematopoietic compartment). In sum, our data suggest that Kdm6a has a relatively subtle role in normal hematopoiesis, but the perturbations associated with its inactivation reveal insights into its role as a potential tumor suppressor in myeloid leukemogenesis.