Scaffolding molecules are increasingly recognized as integral components of many signal transduction cascades. These are proteins without enzymatic function but contain membrane localization sequences, motifs for binding SH3 domains and tyrosine phosphorylation sites for binding SH2 domains. In hematopoietic cells, the Grb2-associated binder (Gab) family of scaffolding molecules couple to a plethora of cytokine receptors and receptor tyrosine kinases. There are three mammalian Gabs: Gab1, Gab2 and Gab3. Gab1 is ubiquitously expressed while the expression of Gab2 and Gab3 is more restricted. Each of the three Gabs has been knocked out: Gab1−/− mice are embryonically lethal, Gab2−/− mice are viable and show selective deficiencies in allergic responses and osteoclastogenesis while Gab3−/− mice are healthy. As most tissues and cell types contain multiple members, it is not clear if the limited phenotypes of the Gab2 and Gab3 knockout mice reflect redundant roles of the Gab proteins. We wish to elucidate what functions, if any, are played by the Gabs in Colony Stimulating Factor-1 (CSF-1)-dependent signaling and proliferation. CSF-1 is a critical growth, survival and differentiation factor for myelomonocytic cells. The IL-3-dependent myeloid progenitor cell line, 32D, is well characterized in our lab. Hematopoietic cells express predominantly Gab2 and Gab3. Real-time RTPCR data revealed that Gab2 and Gab3 are expressed at equivalent levels when copy numbers were determined. To evaluate the contribution of Gab2 and Gab3, we used an RNA interference approach. Multiple small hairpin (sh) RNAs targeted to different parts of the coding region or 3′UTR of Gab2 and Gab3 were screened by reporter assays (Mao and Lee, J. Cell Biol. 170:305) and the most effective shRNAs were chosen for further studies. To select for cells in which both Gab2 and Gab3 are silenced, shGab2 and shGab3 were cloned into a GFP- or dsRed-expressing lentiviral vector respectively. 32D cells were doubly infected with GFP and dsRed lentiviruses and doubly positive cells isolated by FACS sorting. Cell lines were established in which Gab2 and Gab3 were either silenced individually or in combination. Cells in which Gab2 alone was silenced showed a 50–60% reduction in proliferation after 4 d in the presence of maximal CSF-1 doses (10 nM) when compared to cells expressing an irrelevant shRNA, directed against a frog protein. In contrast, no difference was observed in response to maximal doses of IL-3 (10 ng/ml). CSF-1 provoked Akt, but not Erk activation, was significantly reduced in Gab2-silenced cells. When Gab3 was knocked down, the effect on CSF-1 or IL-3 mediated proliferation was not significant as was the effect on Akt and Erk activation. Surprisingly, when Gab2 and Gab3 were silenced together, we observed an increase of 30% in CSF-1-mediated proliferation after only 2 d in culture whereas proliferation in IL-3 was modestly reduced. Similar findings were observed in a clonal assay where clones were scored for size distribution. Inhibitor studies showed that doubly silenced cells were relatively insensitive to the PI3-kinase inhibitor, LY294002. This insensitivity was not observed with rapamycin, which inhibits mTOR. In summary, Gab2 appears to have a positive role in CSF-1 dependent proliferation, correlating with Akt activation while Gab3 appears to be largely dispensable for this function. However simultaneous suppression of Gab2 and Gab3 revealed that these proteins are also involved in modulating PI3-kinase-independent pathways that antagonize proliferation. Current studies are aimed at further characterizing this negative regulation.

Author notes

Corresponding author

Sign in via your Institution