We demonstrated that G-CSF signaling in lineage-negative marrow myeloid progenitors induces C-terminal SHP2 tyrosine phosphorylation more potently than does M-CSF signaling and that SHP2 knockdown in marrow or in the 32Dcl3 cell line reduces Cebpa gene transcription and impairs granulopoiesis.1,2 We also found that Runx1 directly activates Cebpa transcription via promoter elements and via a +37 kb enhancer and that Runx1 gene deletion reduces Cebpa mRNA and impairs granulopoiesis.3 Runx1 is phosphorylated by Src kinase on five tyrosines and dephosphorylated by SHP2, and RUNX1(5F) enhances megakaryopoiesis and rescues CD8 T cell development, whereas RUNX1(5D) is ineffective.4 Mutation to F prevents while change to D or E mimics phosphorylation. These findings suggested that in myeloid cells there exists a G-CSF to SHP2 to RUNX1 to Cebpa activation pathway directing granulopoiesis. We compared wild-type (WT) murine RUNX1b and a series of variants, Y260/375/378/379/386F (5F), Y260F (1F), Y375/378/379/386F (4F), 5D, 1D, and 4D for activation in 293T cells of a reporter with four RUNX1-binding sites linked to the TK promoter and the luciferase cDNA. Expression was compared by Western blotting. WT stimulated the reporter 5-fold on average, relative to empty CMV vector, 5D activated the reporter 15-fold, while 5F was nearly inactive. 4F was also nearly inactive, while 4D had WT activity, as did the 1F and 1D variants. Both 4F and 5F were expressed at or above WT levels; interestingly, 5D, but not 4D or 1D, was expressed at much higher levels than the other variants. A similar pattern was obtained with the MCSFR-Luc reporter. To further query the activity of the four C-terminal tyrosines, we evaluated the activity of Y375/378F (2F*), Y379/386F (2F), 2E, Y375F, Y378F, Y379F, and Y386F. The single residue mutations retained WT activity, whereas both the 2F* and 2F variants demonstrated significant reduction despite expression at or above WT levels. The 2E variant expressed at levels below WT but had increased activity. We then transfected the RUNX1 reporter with activated Src, Runx1, or both – in several experiments Src stimulated the reporter approximately 3-fold, on average, RUNX1 4-fold, and the combination 35-fold. In the same experiments, Src + 5F led to only 6.6-fold activation, and synergy was also lost if the reporter lacked the RUNX1-binding sites. These findings, together with the observation that SHP2 directly activates Src family kinases downstream of G-CSF receptor signaling,5 raises the possibility that in myeloid cells there actually is a G-CSF to SHP2 to Src to Runx1 to Cebpa activation pathway. Consistent with this idea, we find that the 5D variant has reduced interaction with co-expressed HDAC3 relative to WT, as assessed by co-IP, whereas the 5F variant has increased interaction. We previously found that RUNX1-ER stimulates endogenous Cebpa transcription when activated in 32Dcl3 cells. As the ER segment has trans-activating activity, to evaluate RUNX1 variants we utilizing the zinc-inducible MT promoter. Preliminary analysis indicates that both WT RUNX1b and its 5D variant can induce C/EBPa protein and RNA in this context. We previously found that exogenous RUNX1-ER can rescue granulopoiesis when transduced into marrow from Runx1(flox/flox);Mx1-Cre mice subjected 4 weeks earlier to pIpC injections. We now have constructed MIGC retroviral vectors expressing RUNX1, 5F, or 5D and GFP-Cre. These, or the empty MIG or MIGC vectors, were packaged and transduced into marrow isolated from RUNX1(flox/flox) mice. Three days later, lineage-depleted, GFP+ cells were cultured in methylcellulose with IL-3, IL-6, and SCF. Compared with MIG, MIGC reduced the percent CFU-G amongst CFU-G+CFU-M, from 55+/-1% to 10+/-2%; WT RUNX1 or 5D partially rescued granulopoiesis, to 31+/- 8% or 23+/-9% respectively, whereas 5F transduction led to only 4+/-1% CFU-G. In summary, in 293T cells and potentially in myeloid cells Src kinase activates Runx1 and this may favor granulopoiesis.1. Jack et al. Blood 114:2172, 2009. 2. Zhang and Friedman. Blood 118:2266, 2011. 3. Guo et al. Blood 119:4408, 2012. 4. Huang et al. Genes Dev. 26:1587, 2012. 5. Futami et al. Blood 118:1077, 2011. Disclosures:Cantor:AMGEN: Membership on an entity's Board of Directors or advisory committees.
Read full abstract