Abstract 1230

C/EBPα is required for the formation of the granulocyte-monocyte progenitor; however, its role in subsequent lineage specification remains uncertain. We previously compared G-CSF with M-CSF signaling in murine marrow, finding that G-CSF more potently induces phosphorylation of SHP2 whereas M-CSF more potently activates ERK. Knockdown (KD) of SHP2 impaired granulopoiesis relative to monopoiesis and unexpectedly led to reduced C/EBPα RNA. A. Cantor and colleagues recently reported that SHP2 dephosphorylates Src kinase-modified RUNX1 and increases the ability of RUNX1 to induce megakaryopoiesis, and we demonstrated that in myeloid cells RUNX1 binds and activates the CEBPA gene, via its promoter and via a conserved 450 bp enhancer located at +37 kb containing 4 RUNX1-binding sites. Thus, G-CSF or other cytokines potentially direct SHP2 to activate RUNX1 which then increases C/EBPα transcription in myeloid progenitors to favor granulopoiesis. As a test of this model, we have now transduced lineage-depleted murine marrow cells, cultured in SCF, TPO, and Flt3L to minimize myeloid differentiation, with shRNAs targeting CEBPA or with a control lentiviral vector. Western blotting indicated 2-fold KD by shRNAs B9 or B11 after 2 days of transduction followed by 2 days of puromycin selection. As this analysis was done on a mixed population, greater KD may occur in rare progenitors. Upon transfer to IL3, IL6, and SCF to favor myeloid maturation, vector-transduced cells produced 4-fold more CFU-G than CFU-M in methylcellulose, whereas CEBPA KD resulted in 4- or 5-fold more CFU-M than CFU-G. FACS analysis of pooled CFUs demonstrated significant reduction of Mac-1+Gr-1+ granulocytes and an increase in Mac-1+Gr-1 monocytes, and examination of cell morphology after Wright's-Giemsa staining further confirmed a shift towards monopoiesis in response to reduced C/EBPα. Similar findings were obtained for cells placed in liquid culture. In addition, on day 3 of liquid culture Q-RTPCR analysis showed that CEBPA shRNAs induced 2- or 3-fold CEBPA RNA KD, reduced granulocytic MPO, NE, PR3, and GCSFR RNAs, increased monocytic MCSFR and CD14 RNAs, and increased expression of EGR1, IRF8, and c-Jun RNAs, encoding transcription factors that help direct monopoiesis. RUNX1 RNA was unchanged, and PU.1 RNA was reduced <2-fold. To determine the effect of CEBPA KD in vivo, transduced, puromycin-selected CD45.2+ marrow cells were transplanted into lethally irradiated, syngeneic CD45.1+ recipients together with CD45.1+ carrier cells. At day 28, FACS analysis indicated a shift towards monopoiesis amongst shRNA B9- or B11-transduced marrow cells, and CFU assay of sorted CD45.2+ cells demonstrated significantly reduced CFU-G with increased CFU-M. Besides affecting myeloid lineage determination, CEBPA KD increased the rate of cell proliferation in liquid culture, associated with blasts evident upon morphologic analysis and an increase in Mac-1Gr-1 cells, which mainly consisted of c-Kit+Sca-1 myeloid progenitors. In addition, CEBPA KD by shRNA B9, which impaired granulopoiesis more effectively than shRNA B11, increased CFU replating to 8 generations, compared with 4 generations for control cells. Analysis on day 3 of liquid culture did not reveal an alteration in the proportion of cells in G1, S, or G2/M or increased apoptosis, and c-Myc and Bcl-2 RNA levels were unaffected. We propose that CEBPA KD not only strongly impairs granulopoiesis, but also modestly impedes monopoiesis to produce a transient accumulation of cells with blast morphology and the ability to proliferate. Overall, our findings suggest that relative to monopoiesis, granulopoiesis requires increased level or increased activity of C/EBPα. Offering a potential mechanism, we previously found that C/EBPα :AP-1 leucine zipper heterodimers are more abundant than C/EBPα homodimers during monopoiesis compared with granulopoiesis, in part due to c-Fos, c-Jun, and JunB induction, and that C/EBPα :AP-1 heterodimers can direct marrow monopoiesis. Perhaps higher levels of C/EBPα favor homodimer formation, with reduced levels still able to heterodimerize with AP-1 proteins. ERK modification of C/EBPα on S21, as occurs downstream of M-CSF, reduces its ability to direct granulopoiesis; phospho-C/EBPα (S21) potentially retains the ability to direct AP-1 proteins to hybrid DNA elements recently identified in numerous monocytic enhancers.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.