Myelopoiesis Requires a Noncatalytic, Ras-Independent Function of SHP-2.

Activating mutations in PTPN11, which encodes the tyrosine phosphatase SHP-2, comprise the most frequent genetic lesion in juvenile myelomonocytic leukemia (JMML). Other etiologies of JMML include activating mutations in NRAS or KRAS2 and inactivation of the tumor suppressor NF1. These and other observations imply that PTPN11 functions in a common genetic pathway with RAS and NF1. Ras proteins are signal switch molecules that respond to extracellular stimuli by cycling between inactive GDP-bound and active GTP-bound conformations. Oncogenic alleles encode proteins that preferentially accumulate in the GTP-bound form. While NF1 encodes a GTPase activating protein for Ras that directly modulates Ras-GTP levels, the biochemical relationship between SHP-2 phosphatase activity and Ras signaling remains unclear. Most mammalian systems place SHP-2 upstream of Ras activation, but the mechanism is not known. Studies of Ptpn11 mutant embryos and of chimeric mice have shown that SHP-2 plays an essential role in hematopoietic development. We tested the hypothesis that the essential function of SHP-2 in primary hematopoietic cells is to activate Ras. To do this, we determined if Ras activation by expression of an oncogenic Kras2 allele could eliminate the requirement for SHP-2. We used conditional alleles of Kras2 (LSL-Kras^G12D) and Ptpn11 (Ptpn11^flox/flox) coupled with the inducible Mx1-Cre transgene. Juvenile mice were injected with polyI:polyC, resulting in expression of K-Ras^G12D and inactivation of Ptpn11. Although these mice uniformly developed fatal MPD similar to what we previously reported in Mx1-Cre, LSL-Kras^G12D mice (Braun et al., PNAS 101(2):597–602), myeloid progenitors invariably retained an intact Ptpn11 allele despite uniform activation of the conditional Kras^G12D allele. These data suggested that there was strong selective pressure to retain a functional Ptpn11 allele despite oncogenic K-Ras expression. To test this hypothesis directly, we enumerated myeloid progenitor colonies in methylcellulose medium immediately after inactivating Ptpn11 and activating Kras^G12D via retroviral transduction. This confirmed a strong dependence on SHP-2 for formation of myeloid colonies either in the presence or absence of Kras^G12D. Infecting Ptpn11^flox/flox, LSL-Kras^G12D cells with a Ptpn11-IRES-Cre virus fully restored the aberrant growth phenotype of Kras^G12D mutant cells. Remarkably, alleles encoding phosphatase-deficient SHP-2 proteins also rescued CFU-GM growth. These data indicate that SHP-2 is required for growth of both normal and neoplastic myeloid progenitors in vivo and in vitro. Our data support a model in which SHP-2 has essential hematopoietic functions that are independent of Ras activation and do not require SHP-2 phosphatase activity. The presence of protein-protein interaction domains in SHP-2 suggests that it may have a noncatalytic adaptor function. Because transformation by leukemogenic Ptpn11 alleles requires catalytic activity, our data imply that inhibition of SHP-2 catalysis will selectively target neoplastic hematopoietic progenitors.