Gab1 Mediates Neurite Outgrowth, DNA Synthesis, and Survival in PC12 Cells*

The Gab1-docking protein has been shown to regulate phosphatidylinositol 3-kinase PI3K activity and potentiate nerve growth factor (NGF)-induced survival in PC12 cells. Here, we investigated the potential of Gab1 to induce neurite outgrowth and DNA synthesis, two other important aspects of NGF-induced neuronal differentiation of PC12 cells and NGF-independent survival. We generated a recombinant adenovirus encoding hemagglutinin (HA)-epitope-tagged Gab1 and expressed this protein in PC12 cells. HA-Gab1 was constitutively tyrosine-phosphorylated in PC12 cells and induced the phosphorylation of Akt/protein kinase B and p44/42 mitogen-activated protein kinase. HA-Gab1-stimulated a 10-fold increase in neurite outgrowth in the absence of NGF and a 5-fold increase in NGF-induced neurite outgrowth. HA-Gab1 also stimulated DNA synthesis and caused NGF-independent survival in PC12 cells. Finally, we found that HA-Gab1-induced neuritogenesis was completely suppressed by pharmacological inhibition of mitogen-activated protein kinase kinase (MEK) activity and 50% suppressed by inhibition of PI3K activity. In contrast, HA-Gab1-stimulated cell survival was efficiently suppressed only by inhibition of both PI3K and MEK activities. These results indicate that Gab1 is capable of mediating differentiation, DNA synthesis, and cell survival and uses both PI3K and MEK signaling pathways to achieve its effects.

Nerve growth factor (NGF) 1 is a potent inducer of cell survival and differentiation in neurons and neural tumor cells (1,2). For example, NGF treatment of PC12 pheochromocytoma cells, a model system for sympathetic neuron differentiation, results in the elaboration of neurites, induction of a single round of DNA synthesis, and cell survival (3,4). NGF mediates its effects by binding to and stimulating the activity of the TrkA receptor tyrosine kinase (5,6). The activated TrkA receptor then recruits a series of cytoplasmic proteins to docking sites consisting of phosphorylated tyrosines in specific sequence contexts. These proteins include Shc, PLC-␥1, SHP-1, APS-1, SH2-B, and FRS-2 (1,7,8). The major function of substrate recruitment to TrkA is to couple this receptor to a number of intracellular signaling pathways. For example, the Shc binding site on TrkA is required for NGF-induced activation of Ras and PI3K (9,10), whereas both the Shc and PLC-␥1 binding sites on TrkA are required for maximal activation of MAPK (11). However, it has recently been found that the Gab1-docking protein, which can bind both GRB2, a regulator of Ras activity, and PI3K, associates with the majority of the activated PI3K activity in NGF-treated PC12 cells (12). Although Gab1 is phosphorylated on tyrosine in response to TrkA activation, Gab1 does not associate directly with TrkA (13). Thus, both Ras and PI3K signaling pathways, which are required for NGF-induced neurite outgrowth and survival, respectively, are regulated by direct receptor-substrate interactions through Shc, APS-1, SH2-B, and FRS-2 and by indirect receptor-Gab1 mechanisms.
Gab1 (growth-associated binder-1) was originally cloned as a GRB2-associated protein from a human glial tumor expression library and also through yeast two-hybrid screens using the Met receptor tyrosine kinase as bait (13,14). Gab1 belongs to family of docking proteins consisting of Gab2, IRS-1, IRS-2, and DOS that share sequence homology between their single pleckstrin homology domains and that have been shown to link cell surface receptors to intracellular signaling proteins (14,15). Gab1 has multiple SH2 and SH3 binding sites for PI3K and SHP-2 and for the adapter proteins GRB2, NCK, and CRK (13,14). Gab1 is tyrosine-phosphorylated by a variety of cell surface receptors, including the insulin, epidermal growth factor, c-Met, and c-erbB2 receptors, B-cell antigen, and the gp130 subunit of cytokine receptors (16 -19, 14, 20 -22). Gab1 can associate directly with activated receptors such as c-Met or indirectly through GRB2 or SHP-2 (13,14). Tyrosine-phosphorylated Gab1 then associates with downstream signaling proteins such as GRB2, the p85 subunit of PI3K and SHP-2 in various cell systems (12,14,(21)(22)(23)(24). A diverse set of biological responses are regulated by Gab1. For example, Gab1 mediates branching and alveolar morphogenesis of Madin-Darby canine kidney epithelial and EpH4 mammary epithelial cells (13,19,22), cell scattering of Madin-Darby canine kidney cells (13), transformation of epithelial and NIH 3T3 cells (25,26,14), and cell proliferation in NIH-3T3 (14). In addition, Gab1 reduces the concentration of NGF required for cell survival via the three PI3K binding sites found in the protein (12).
The observation that Gab1, through binding of GRB2 and PI3K, links TrkA to both MAPK-and PI3K-signaling pathways led us to ask whether Gab1 has the potential to induce other * This work was supported by the Academy of Finland and Finnish Cultural Foundation (to J. K.) and by grants from the National Cancer Institute of Canada and Medical Research Council (to D. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Generation of Polyclonal Antibody-Rabbit polyclonal antibody was raised against a C-terminal peptide of human Gab1, CRQSTESET-PAKSVK. The peptide shares only 40% homology with the C terminus of Gab2. The peptide was linked to keyhole limpet hemocyanin and used to immunize rabbits. Antibody that recognized the immunizing peptide by enzyme-linked immunosorbent assay and glutathione S-transferase-Gab1 and endogenous Gab1 by Western blotting was used for the experiments in this study. The specificity of the antibody for Gab1 was confirmed by incubating the antibody with immunizing peptide (1 M) for 4 h at 4°C before immunoprecipitation of glutathione S-transferase-Gab1 and endogenous Gab1.
Cell Culture-Rat PC12 pheochromocytoma cells (Ref. 5; from R. Kelly, University of California, San Francisco or from P. Barker, McGill University) were maintained in DMEM supplemented with 10% heatinactivated horse serum, 5% fetal calf serum, and antibiotics at 37°C in a 5% CO 2 atmosphere. For survival, proliferation, and differentiation assays, cells were washed three times in serum-free DMEM and plated onto rat tail collagen-coated 24-well culture dishes (at 2 ϫ 10 4 cells/well) containing serum-free DMEM supplemented with or without 100 ng/ml NGF. Cells were incubated with recombinant adenovirus for 16 h, after which medium was changed. Cell viability was determined at days 1 and 2 by MTT cell proliferation assay (Sigma; 27).
For [ 3 H]thymidine incorporation assays, PC12 cells were incubated with 0.5 Ci/well of [ 3 H]thymidine for 16 h. Cells were washed three times with ice-cold phosphate-buffered saline, and 5% trichloroacetic acid was added for 20 min at 4°C. Cells were washed once with 5% trichloroacetic acid, and a mixture of 0.1 N NaOH and 1% SDS was added for 10 min. Samples were counted by scintillation counter.
Neurite outgrowth experiments were carried out in serum-free medium with or without 100 ng/ml NGF. Two days after infection, PC12 cells with neurites longer than two cell bodies in diameter were counted as positive. For pharmacological inhibition of PI3K or MEK, 10 M LY 294002 (Biomol) and/or 20 M PD-98059 (Biomol) were added to cells on the day of adenovirus infection and when media was changed on the following day. Each experiment was repeated three times in triplicate.
For biochemical studies, PC12 cells were cultured in complete DMEM overnight in 12-well dishes (at 8 ϫ 10 4 cells/well). Cells were incubated in serum-free DMEM overnight and infected with Gab1 or LacZ recombinant adenovirus in 0.1% serum containing DMEM at a multiplicity of infection (m.o.i.) of 100. Infected PC12 cells were cultured for 24 h, incubated for 8 h in serum-free DMEM, and then treated with 100 ng/ml NGF for 1-20 min at 37°C.
The PC12 cells were grown in 15% serum, a concentration to obtain maximal cell proliferation and survival. When PC12 cells were grown in complete medium, we observed a high basal level of Gab1 tyrosine phosphorylation, likely due to insulin/insulin-like growth factor-1 in the serum. We therefore incubated the cells overnight in serum-free medium to reduce this level. For adenovirus infections, we used 0.1% serum, an amount that helps to stabilize the adenovirus during infection, and that does not affect basal Gab1 tyrosine phosphorylation. After infection, the cells were again cultured in serum-free conditions in the absence or presence of NGF or hemagglutinin (HA)-Gab1 expression. We felt that it was important to assess NGF and HA-Gab1 responses in the absence of survival or proliferation factors present in serum.
Generation of Gab1 Recombinant Adenovirus-Gab1 cDNA was subcloned as an XbaI-HpaI fragment into the XbaI and ApaI site of pCDNA3 (Invitrogen). An HA tag (MYPYDVPDYASRGSPEF) was added onto the 5Ј end of the human Gab1 cDNA by polymerase chain reaction, and the cDNA was sequenced using automatic sequencing. Gab1 expression was verified by transient calcium phosphate transfections. The HA-Gab1 XbaI-ClaI fragment was subcloned into XbaI and NotI restriction site of the pADCMV5F adenovirus vector (Quantum Technologies). The HA-Gab1 plasmid was linearized with FseI and co-transfected with ready-cut adenovirus DNA (Quantum) into 293 human embryo kidney cells by calcium phosphate precipitation as described by Slack et al. (28). Recombinant adenovirus-containing plaques were purified three times by plaque assay, and the virus was amplified in 293 cells. Viruses were used both as crude and cesium chloride centrifugation-purified preparations. Infectious titer was determined by limiting dilution plaque assay on 293 cells. Viruses were used at m.o.i. values of 10, 50, 100, and 200.
Immunofluorescence-Infected PC12 cells were fixed with 4% paraformaldehyde, and immunofluorescence was performed with monoclonal HA antibody clone 12CA5 (Roche Molecular Biochemicals) using the method of Canfield et al. (29).
Western Blotting and Immunoprecipitation-After stimulation with 100 ng/ml NGF (Cederlane Laboratories Ltd, Canada), cells were lysed in lysis buffer (20 mM Tris (pH 7.4), 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 500 M sodium vanadate, aprotinin (5 g/ml), leupeptin (5 g/ml), and 1 mM phenylmethylsulfonyl fluoride). The lysates were incubated at 4°C for 20 min and were cleared by centrifugation at 10,000 ϫ g for 10 min. The lysates were incubated overnight with 2 g of monoclonal HA antibody or Gab1 antiserum. Protein A-Sepharose and/or protein G-Sepharose (Amersham Pharmacia Biotech) was added to the lysates for 2 h, and the immunoprecipitates were washed 3 times with lysis buffer without protease inhibitors and twice with distilled water. Proteins were eluted with Laemmli SDS loading buffer by boiling for 5 min, separated on a 7% polyacrylamide gel, and transferred to a nitrocellulose membrane (Schleicher & Schuell). The membranes were blocked with 3% bovine serum albumin for 1 h at room temperature, probed with anti-Gab1, anti-HA, or 4G10 anti-phosphotyrosine (Tyr(P)) (Upstate Biotechnology, Inc.) overnight at 4°C. Other antibodies used for Western blotting were anti-GRB2 (Santa Cruz Biotechnology), anti-p85/PI3K (N-SH2 domain; Upstate Biotechnology Inc), anti-NCK (Transduction Laboratories), anti-CRK (Transduction Laboratories), anti-p70 S6K (Santa Cruz), anti-phospho S473 Akt (New England Biolabs), anti-Akt (New England Biolabs), anti-active MAPK (Promega), and anti-extracellular signal regulated protein kinase-1 (Santa Cruz). The membranes were washed with Tris-buffered saline with Tween-20 and probed with peroxidase-conjugated anti-rabbit or anti-mouse (Roche Molecular Biochemicals) for 1 h at room temperature. The immune complexes were visualized by ECL (Amersham Pharmacia Biotech). Stripping and reprobing of membranes were performed as recommended by the manufacturer.
Apoptosis Assay-Gab1-and LacZ-infected PC12 cells were cultured in serum-free medium for 2 days and fixed with 4% paraformaldehyde for 5 min at RT. Cells were stained using a Fluorescein-FragEL DNA fragmentation detection kit (Oncogene Research Products) according to the manufacturer's protocol.

Characterization of Polyclonal Gab1
Antibody-A Gab1-specific polyclonal antibody was generated against a C-terminal peptide of human Gab1, as described under "Experimental Procedures." Anti-Gab1 recognized a protein of approximately 100 kDa in both NIH 3T3 and PC12 cells in Western blots (Fig.  1A). The Gab1 antibody also immunoprecipitated Gab1 protein from PC12 cells. Preincubation of the antibody with the immunizing C-terminal Gab1 peptide (1 M) prevented immunoprecipitation of Gab1 (Fig. 1A).
As previously reported (12), Gab1 was phosphorylated on tyrosine in NGF-treated PC12 cells. PC12 cells preincubated overnight in serum-free medium were stimulated with 100 ng/ml NGF, and Gab1 was immunoprecipitated with anti-Gab1 followed by Western blotting with anti-Tyr(P). Maximum tyrosine phosphorylation of Gab1 was achieved in 5 min, and phosphorylation persisted for 20 min (Fig. 1B). In non-serumstarved PC12 cells, Gab1 was constitutively tyrosinephosphorylated (data not shown). Adenovirus-expressed Gab1 Associates with p85/PI3K and GRB2-Gab1 has binding sites for the p85 regulatory subunit of PI3K and for the adaptor proteins GRB2, NCK, and CRK. To assess whether HA-Gab1 associated with p85/PI3K and GRB2, HA-Gab1-infected PC12 cells were treated with 100 ng/ml NGF for 5 or 10 min, lysed, and immunoprecipitated with monoclonal HA antibody. Western blots of the immunoprecipitated proteins were probed with p85/PI3K, GRB2, NCK, or CRK antibodies. HA-Gab1 constitutively associated with p85/PI3K and GRB2 (Fig. 3B) but not NCK or CRK (data not shown). NGF treatment stimulated the association of HA-Gab1 with p85/PI3K and GRB2 by approximately 2-fold. Anti-HA did not immunoprecipitate Gab1, p85/PI3K, or GRB2 from mock or control LacZ adenovirus-infected cells (data not shown).

Gab1 Expressed via Adenovirus Is Highly Expressed in PC12
Adenovirus-expressed Gab1 Enhances Akt/PKB and MAPK Phosphorylation and Induces p70 S6K Mobility Shift-To determine whether HA-Gab1 stimulated PI3K and MAPK signaling pathways, the activities of Akt/PKB and MAPK, which are regulated by PI3K and MEK activity, respectively, were assessed by examining their phosphorylation states. Antibodies that recognize activated, phosphorylated forms of Akt/PKB and MAPK were used in these experiments. PC12 cells were infected with the HA-Gab1 adenovirus and treated with 100 ng/ml NGF for various times. Lysates from these cells were probed in Western blots with anti-phospho-Akt or phospho-MAPK or with anti-Gab1, Akt/PKB, p70 S6K, or MAPK. The phosphorylation state of p70 S6K, which is regulated by PI3K activity, was also assessed by gel mobility shift. HA-Gab1 was constitutively tyrosine-phosphorylated in PC12 cells grown in 0.1% serum without NGF (Fig. 4A, top panel), whereas NGF treatment stimulated HA-Gab1 tyrosine phosphorylation (Fig.  4B, top right panel). HA-Gab1 expression stimulated the phosphorylation of Akt/PKB on serine 473 (a site required for Akt activation) and of p44 and p42 MAP kinases and induced a size shift of p70 S6K (Fig. 4, A and B). Infection with a control LacZ-expressing adenovirus did not increase Akt/PKB or MAPK phosphorylation or induce a mobility shift of p70 S6K (Fig. 4A).
Gab1 expression has previously been reported to decrease the amount of NGF necessary to inhibit apoptosis of PC12 cells (12). To determine whether HA-Gab1 expression could enhance or prolong NGF-regulated signaling responses, we compared the phosphorylation states of Akt/PKB, p70 S6K, and MAPK in NGF-stimulated control or HA-Gab1 expressing cells. HA- Gab1-expressing PC12 cells stimulated by NGF showed a rapid and sustained phosphorylation of Akt/PKB and MAPK and more pronounced Akt/PKB phosphorylation as compared with uninfected cells (Fig. 4B) Adenovirus-expressed Gab1 Enhances Neurite Outgrowth of PC12 Cells-HA-Gab1 was expressed in PC12 cells, and neurite outgrowth was assessed as compared with mock or control LacZ adenovirus-infected cells. In Gab1-infected PC12 cells cultured for 48 h in serum-free medium without NGF, the cells appeared flat with neurite projections (Fig. 5). Fifty percent of PC12 cells expressing HA-Gab1 had neurites at least two cell bodies in diameter, as compared with approximately 5% of mock or LacZ control adenovirus-infected cells (Fig. 6A). When the results of all experiments from both PC12 subclones obtained from different sources were quantified, PC12 cells infected at an m.o.i. of 100 had 2.5 times more neurite projections, and at an m.o.i. of 200, had 10 times more neurites than did noninfected PC12 cells or cells infected with a LacZ control adenovirus (Fig. 6B). In PC12 cells treated with 100 ng/ml NGF, HA-Gab1 induced virtually all of the cells to elaborate neurites, as compared with 25% of LacZ-infected cells (Fig. 6C). When the results of all experiments from both PC12 subclones were quantified, PC12 cells increased 6-fold the number of neurites as compared with NGF-treated cells expressing a LacZ control adenovirus (Fig. 6D). Thus, HA-Gab1 expression is sufficient both to induce neuritogenesis in the absence of NGF treatment of cells and to enhance neuritogenesis induced by NGF.
Inhibition of MEK and PI3K Activities Suppress Gab1-induced Neurite Outgrowth-We next assessed the contribution of MEK and PI3K signal transduction pathways to Gab1-mediated neurite outgrowth using PD-98059, a MEK-selective inhibitor, and LY 294002, a PI3K-selective inhibitor. Previous results using MEK and PI3K inhibitors indicated that both of these enzymes play roles in neurite outgrowth from PC12 cells (30,31). HA-Gab1-infected PC12 cells were treated with 10 M LY 294002 and/or 20 M PD-98059 inhibitors at the time of infection. PD-98059 completely inhibited neurite outgrowth from untreated HA-Gab1-expressing cells (Fig. 6C) and decreased the number of neurite-bearing cells by 4-fold in NGFtreated HA-Gab1-expressing cells (Fig. 6D). LY 294002 also completely suppressed neurite outgrowth from untreated HA-Gab1-expressing cells (Fig. 6C). In NGF-treated HA-Gab1-expressing cells, however, LY 294002 only partially inhibited neurite outgrowth, by 2-fold (Fig. 6D). Thus, Gab1-induced neurite outgrowth requires both MEK and PI3K activities, with MEK activity being more important for Gab1-mediated neuritogenesis in NGF-treated cells.
Adenovirus-expressed Gab1 Induces DNA Synthesis-NGF has been reported to induce a single round of DNA synthesis in PC12 cells (32). To determine whether HA-Gab1 could induce a similar response, PC12 cells were infected with HA-Gab1 or LacZ adenovirus in the presence of showed much smaller (7.5-fold) increases in [ 3 H]thymidine incorporation as compared with cells infected for 24 h (Fig. 7B). Total cell counting did not show significant differences in cell number between noninfected and HA-Gab1 adenovirus-infected PC12 cells cultured in serum-free medium, indicating that HA-Gab1 induced DNA synthesis and not cell proliferation (data not shown). Adenovirus-expressed LacZ, unlike HA-Gab1, did not induce DNA synthesis. The LacZ control adenovirus, but not the HA-Gab1 adenovirus, showed some cytopathic effects at the 48 h but not at the 24-h time point, as assessed visually and by MTT assay (Fig. 7 and data not shown).
Adenovirus-expressed Gab1 Stimulates PC12 Cell Survival-Overexpression of Gab1 in PC12 cells has previously been shown to reduce the concentration of NGF necessary to prevent apoptosis (12). We asked whether Gab1 expression alone is sufficient to mediate the survival of PC12 cells grown in the absence of any survival factors. PC12 cells were infected with HA-Gab1 and cultured in serum-free medium without NGF, and the percentage of surviving cells was determined by MTT assay. 24 h after infection, PC12 cells expressing HA-Gab1 (m.o.i. of 200) survived 5-fold better than uninfected or LacZ adenovirus-infected cells (Fig. 8A). The LacZ adenovirus showed no toxicity in these experiments (Fig. 8A). HA-Gab1 increased cell survival in a dose-responsive manner, with survival at 200 m.o.i. greater than that observed with 100 ng/ml NGF (Fig. 8A). 48 h after infection, HA-Gab1-expressing PC12 cells also showed 8-fold increases in surviving cells as compared with uninfected or LacZ adenovirus-infected cells (Fig.  8B). As reported by Holgado-Madruga et al. (12), HA-Gab1 enhanced NGF-induced PC12 cell survival by 2-fold (Fig. 8B). Thus, Gab1 is not only capable of enhancing NGF-induced cell survival but can effectively stimulate cell survival in the absence of NGF.
Adenovirus-expressed Gab1 Protects PC12 Cells from Apoptosis-To confirm that HA-Gab1 inhibits apoptosis of PC12 cells, a DNA fragmentation assay was performed. HA-Gab1 and LacZ adenovirus-infected PC12 cells were cultured in serum-free medium for 48 h after infection, and apoptosis was assessed using a Fluorescein-FragEL DNA fragmentation de-tection kit. Cells were counterstained with Hoechst nuclear stain. Although 70% of noninfected PC12 cells were apoptotic when cultured without NGF, only 22% of HA-Gab1-expressing PC12 cells were apoptotic (Fig. 8, C and D). 18% of uninfected PC12 cells treated with NGF were apoptotic (Fig. 8, C and D). Cells infected with the LacZ control adenovirus did not show significant differences in apoptosis as compared with uninfected cells.
Inhibition of MEK and PI3K Activities Suppresses HA-Gab1induced Cell Survival-The experiments of Fig. 6 indicate that both MEK and PI3K activities are required for HA-Gab1-mediated neurite outgrowth. To determine the requirement for these activities in HA-Gab1-mediated survival, cells were treated with the MEK inhibitor PD-98059 (20 M), or the PI3K inhibitor LY 294002 (10 M). Inhibition of MEK or PI3K activities suppressed HA-Gab1-mediated survival by only 10 or 35%, respectively (Fig. 9A). However, treatment of HA-Gab1infected cells with both MEK and PI3K inhibitors, either in the presence or absence of NGF treatment, completely suppressed cell survival (Fig. 9, A and B). These results suggest that MEK and PI3K act together to mediate Gab1-induced cell survival. To confirm that the PD-98059 and LY 294002 inhibitors suppressed the activities of MEK and PI3K pathways, respectively, the phosphorylation states of Akt/PKB and MAPK and the mobility of p70 S6K in gels were examined in HA-Gab1expressing cells in the absence of NGF treatment. PD-98059 inhibited HA-Gab1-induced activation of MAPK but not of Akt/ PKB, whereas LY 294002 inhibited HA-Gab1-induced phosphorylation of Akt/PKB and p70 S6K but not of MAPK (Fig.  9C). LY 294002 treatment also reduced the tyrosine phosphorylation of HA-Gab1, suggesting that PI3K activity stimulates or maintains Gab1 activity (Fig. 9C). The LY 294002-induced reduction in Gab1 tyrosine phosphorylation may occur at specific sites on Gab1 other than those responsible for MAPK activation, such as the GRB2 association site.

DISCUSSION
Gab1, an IRS-1 family member, has previously been shown to be phosphorylated on tyrosine in NGF-treated cells and to bind the majority of NGF-activated PI3K (12). In addition, overexpression of Gab1 reduced the concentration of NGF required for PC12 cell survival in a manner dependent upon the integrity of the PI3K binding sites (12). Holgado-Madruga et al. (12) propose that Gab1 plays a major role in linking TrkA to PI3K activation and cell survival responses. In this report, we demonstrate that Gab1 overexpression in PC12 cells not only enhances NGF-dependent survival but on its own, induces cell survival, neurite outgrowth, and DNA synthesis. Thus Gab1 has the potential to regulate these three most evident phenotypic effects of NGF. The differences between the effects of Gab1 reported here and the previous report from Holgado-Madruga et al. (12) are most likely due to the higher levels of Gab1 expression that we achieve using recombinant adenovirus, which we estimate as approximately 10-fold more than endogenous Gab1 expression (at an m.o.i. of 100). Our study therefore assesses the potential of Gab1 to induce NGF-like responses in PC12 cells. Our results suggest that Gab1 can play major roles in cell survival, neuritogenesis, and DNA synthesis responses.
Gab1, expressed using recombinant adenovirus, was tyrosine-phosphorylated in the absence of NGF addition to cells. Furthermore, adenovirus-expressed Gab1 was capable of inducing the tyrosine phosphorylation or activation of signaling proteins, cell survival, and neurite outgrowth in untreated PC12 cells. This may be due to tyrosine phosphorylation of exogenously expressed Gab1 by factors in serum such as insulin-like growth factor-1 and epidermal growth factor (14) or to limiting amounts of a tyrosine phosphatase that may function to keep Gab1 in an "off" position. We have recently observed that the SHP-1 tyrosine phosphatase may play such a role for the TrkA. 2 Constitutive, ligand-independent tyrosine phosphorylation of Gab1 and its association with signaling proteins such as PI3K has been previously observed in Madin-Darby canine kidney cells overexpressing Gab1 (13) and in PC12 cells (14).
How does Gab1 induce effects as diverse as cell survival, neuritogenesis, and DNA synthesis? Gab1 has been shown to bind a series of signaling proteins in vitro and in vivo, including GRB2, PLC-␥1, PI3K, and SHP-2. We and others have shown that these proteins can mediate cell survival, neurite outgrowth, and DNA synthesis in PC12 cells and other cell types (1,33). Cell survival in PC12 cells is regulated by PI3K (34 -36). Holgado-Madruga et al. (12) demonstrate that the ability of Gab1 to enhance NGF-induced survival is dependent upon Gab1 association with PI3K. The phosphorylated YMXM motifs in Gab1 presumably bind and activate PI3K, analogous to PI3K association and activation by insulin-activated IRS-1 (37). We show that cell survival by Gab1 requires, in part, PI3K activity and that Gab1 causes the phosphorylation of two of the major targets of NGF-induced PI3K activity, the serine/threo-2 N. Marsh, C. Dubrevill, and D. R. Kaplan, unpublished data. nine kinases Akt/PKB and p70 S6K. In sympathetic and cerebellar neurons and in a hippocampal cell line, PI3K mediates neurotrophic factor-induced cell survival by activating Akt/ PKB (38,39,40). Our results suggest that Gab1 induces cell survival by stimulating both Akt/PKB and MEK activity. Akt/ PKB may in turn promote cell survival by suppressing GSK-3 activity (41) or by inducing the phosphorylation and inactivation of the BAD, procaspase, or forkhead apoptotic proteins (42). Our unpublished results indicate that Akt/PKB activity is required for NGF and serum-induced survival of PC12 cells and that Akt/PKB promotes survival by inhibiting the activity of an MEKK-JNK cell death pathway. 3 The survival pathway regulated by MEK/MAPK is not yet defined in PC12 cells. Although NGF has been reported to induce survival solely through PI3K (34,43), Pá rrizas et al. (35) note that insulin-like growth factor-1 can promote PC12 cell survival in a MEK and PI3K-dependent manner.
FIG. 8. HA-Gab1 enhances PC12 cell survival and potentiates NGF-induced survival. PC12 cells were cultured and infected as described in Fig. 5 and incubated in serum-free medium for 24 (A) or 48 (B) h, and MTT assays performed. Cells were treated with 100 ng/ml NGF at the time of infection as indicated. C and D, terminal dUTP nick-end labeling (TUNEL) assay for apoptosis. Infected PC12 cells were cultured in serum-free medium for 48 h, stained using Fluorescein-FragEL DNA fragmentation detection kit, and counterstained with Hoechst nuclear stain. Positively stained apoptotic cells, and the total number of Hoechst-stained nuclei were counted from five separate views/well. The stimulation of MAPK activity in this report was in contrast to that obtained by Holgado-Madruga et al. (14). The later study showed that overexpression of Gab1 in NIH-3T3 cells inhibited the activity of the MAPK pathway, possibly due to Gab1 competing with SOS for binding to GRB2, and thus preventing GRB2/SOS/Ras/Raf-1-induced activation of MEK/ MAPK. Three other reports, however, show that the MAPK pathway is stimulated by Gab1 overexpression, including those of Weidner et al. (13) in Madin-Darby canine kidney cells, Nishida et al. (Nishida) in 293T cells, and Takahashi-Tezuka et al. (20) in lymphoid cells. The ability of Gab1 to activate or suppress MAPK activity may be a function of whether GRB2 is present in sufficient quantities to allow GRB2 coupling to both Gab1 and SOS.
Gab1 appeared to induce a single round of DNA synthesis without stimulating cell proliferation in serum-free medium, an observation similar to that of NGF treatment of PC12 cells (32). Similarly, Klippel et al. (33) demonstrate that overexpression of PI3K activity in rat embryo fibroblasts induced DNA synthesis and activation of cdk2 and -4 but not cell proliferation. We therefore suggest that Gab1-stimulated PI3K activity regulates the induction of DNA synthesis and that a Gab1/ PI3K pathway may be responsible for the mitogenic effects of NGF that are observed immediately after NGF treatment of PC12 cells (48).
Our results indicate that Gab1 is capable of mediating neurite outgrowth, DNA synthesis, and cell survival responses and uses both PI3K-and MEK-signaling pathways to achieve its effects. Gab1 could either function as a primary Trk effector or provide an additional mechanism for activating PI3K and MEK/MAPK that normally occurs through the Shc and PLC-␥1 binding sites on Trk. The binding of GRB2 to the Shc site is thought to be a primary means used by Trk to activate PI3K and Ras (9, 10), whereas the PLC-␥1 binding site, together with the Shc binding site, participates in the activation of MAPK (11). The tyrosine phosphorylation of Gab1 by NGF-activated Trk and the subsequent activation by Gab1 of PI3K and MEK/ MAPK-signaling pathways may be a major alternative mechanism used by NGF to stimulate cell survival and neuritogenesis.