SHP2 Tyrosine Phosphatase Enhances CEBPA Gene Expression, STAT3 Activation, and Cytokine-Dependent Granulopoiesis.

Abstract 1550

Culture of individual murine granulocyte-monocyte progenitors (GMP) with myeloid cytokines demonstrates that M-CSF or G-CSF help direct monocyte versus granulocyte lineage specification (T. Schroeder and colleagues, Science, 2009). We evaluated signaling pathways induced in murine, lineage-negative marrow cells to demonstrate that M-CSF more potently induces ERK activation via PLCγ and PKC whereas G-CSF more potently activates STAT3 and induces SHP2 phosphorylation on tyrosines 542 and 580, potentially increasing SHP2 activity or altering its substrate specificity (Jack et al 2009). In the same study, we found that chemical inhibition of ERK reduces CFU-M but not CFU-G, whereas exposure of marrow cells to the SHP1/SHP2 inhibitor NSC-87877 reduces CFU-G but not CFU-M. SHP2 indirectly activates ERK, and NSC-87877 prevented ERK phosphorylation induced by G-CSF but not M-CSF, underscoring the specificity of G-CSF for SHP2 activation. We postulate that constitutively active SHP2 found in JMML cases excessively activates ERK to enable a monocytic phenotype. To further investigate the role of SHP2 in normal granulopoiesis, we screened a panel of lentiviral vectors expressing SHP2 shRNAs for their ability to reduce SHP2 protein expression in the murine 32Dcl3 granulocyte progenitor cell line. shRNA77 reduced SHP2 expression 5-fold in pooled transductants, whereas the other shRNAs produced at most 3-fold inhibition. Five 32Dcl3-shRNA77 subclones were obtained, two with 90%, one with 80%, and two with approximately 65% reduction in SHP2 levels. Proliferation in IL-3 correlated inversely with the degree in SHP2 knockdown. Expression of myeloperoxidase and lactoferrin RNAs were reduced 24 hrs after transfer from IL-3 to G-CSF in each subclone, in comparison to cells transduced with the empty pLKO.1 lentiviral vector, again correlated with the degree of SHP2 knockdown. C/EBPα RNA and protein levels were markedly reduced in all five subclones, whereas PU.1 RNA or protein or RUNX1 protein levels were unaffected. As expected, knockdown of SHP2 reduced ERK activation, and STAT3 activation was also reduced in the two shRNA77 sublcones evaluated. We also developed two 32Dcl3 subclones containing >90% knockdown of both ERK1 and ERK2, using separate shRNAs, and demonstrated that these lines each retain normal levels of C/EBPα protein, normal STAT3 activation, and normal induction of MPO, PR3, and LF by G-CSF, despite have markedly reduced proliferation. Thus, reduced C/EBPα RNA and protein expression or STAT3 activation as a consequence of SHP2 knockdown is not due to loss of ERK activation or to reduced growth rate. Lineage-negative murine marrow cells cultured in TPO, Flt3 ligand, and SCF were transduced with shRNA77 or empty vector. Cell numbers expanded approximately 7-fold in each group during this period. Upon transfer to IL3, IL6, and SCF, control cells continued to expand whereas shRNA77-transduced cells showed little proliferation in liquid culture. SHP2 protein levels were reduced 3-fold. Quantitative RT-PCR analysis after 48 hrs in IL3/IL6/SCF demonstrated that the granulocytic markers MPO, PR3, LF, NE, and C/EBPε were significantly reduced (p < 0.05) and the monocytic marker CD68 was significantly increased (p < 0.05) in cells transduced with shRNA77 compared to empty vector, and there was a significant increase in Mac-1+Gr1- monocytes compared with Mac-1+Gr1+ granulocytes, each in three independent experiments. SHP2 shRNA77 reduced CFU-G from 75% to 50% of total CFU-G + CFU-M, again in three experiments (p = 0.005). Knockdown of SHP2 in marrow cells with shRNA75 also reduced C/EBPα protein levels and expression of granulocytic markers. In summary, SHP2 RNA knockdown extends prior results obtained using a chemical inhibitor to support the conclusion that SHP2 is required for granulocyte lineage commitment and maturation. Reduced C/EBPα expression and/or STAT3 activation provides a mechanistic basis for this finding, with a novel SHP2 substrate rather than ERK activation likely responsible.