Induction of Neurite Extension and Survival in Pheochromocytoma Cells by the Rit GTPase*

The Rit, Rin, and Ric proteins comprise a distinct and evolutionarily conserved subfamily of the Ras-like small G-proteins. Although these proteins share the majority of core effector domain residues with Ras, recent studies suggest that Rit uses novel effector pathways to regulate NIH3T3 cell proliferation and transformation, while the functions of Rin and Ric remain largely unknown. Since we demonstrate that Rit is expressed in neurons, we investigated the role of Rit signaling in promoting the differentiation and survival of pheochromocytoma cells. In this study, we show that expression of constitutively active Rit (RitL79) in PC6 cells results in neuronal differentiation, characterized by the elaboration of an extensive network of neurite-like processes that are morphologically distinct from those mediated by the expression of oncogenic Ras. Although activated Rit fails to stimulate mitogen-activated protein kinase/extracellular-signal-regulated kinase (MAPK/ERK) signaling pathways in COS cells, RitL79 induced the phosphorylation of ERK1/2 in PC6 cells. We also find that Rit-mediated effects on neurite outgrowth can be blocked by co-expression of dominant-negative mutants of C-Raf1 or mitogen-activated protein kinase kinase 1 (MEK1). Moreover, expression of dominant-negative Rit is sufficient to inhibit NGF-induced neurite outgrowth. Expression of active Rit inhibits growth factor-withdrawal mediated apoptosis of PC6 cells, but does not induce phosphorylation of Akt/protein kinase B, suggesting that survival does not utilize the phosphatidylinositol 3-kinase (PI3K)/Akt pathway. Instead, pharmacological inhibitors of MEK block Rit-stimulated cell survival. Taken together, these studies suggest that Rit represents a distinct regulatory protein, capable of mediating differentiation and cell survival in PC6 cells using a MEK-dependent signaling pathway to achieve its effects.

The Ras subfamily consists of the classical Ras (H-, K-, and N-), R-Ras, Ral, Rheb, TC21/R-Ras2, M-Ras/R-Ras3, and the newest members of the family Rit, Rin, and Ric (1). As molecular switches, the Ras-related GTPases respond to external signals by exchanging GTP for bound GDP, and the GTP-bound active proteins interact with specific downstream effectors thereby triggering intracellular signaling cascades through their interaction with a variety of target proteins (2). Guanine nucleotide exchange factors (GEFs) 1 and guanosine triphosphatase (GTPase)-activating proteins (GAPs) influence the relative proportion of molecules in the active and inactive conformations (3). GEFs promote activation by inducing the release of GDP, whereas GAPs inactivate Ras-like proteins by stimulating their intrinsic GTPase activity (4). In addition to this highly regulated GTPase cycle, membrane association of Rasrelated proteins is also essential to their biological activity.
While many studies have defined a critical role for the classical Ras proteins in the regulation of cell growth and differentiation, less is known about the other members of the Ras subfamily (1). This is particularly true for the Rit (Ras-like protein in all tissues), Rin (Ras-like protein in neurons), and Drosophila Ric (Ras-related protein that interacted with calmodulin) proteins whose cellular functions are only beginning to be characterized (5)(6)(7). These proteins share more than 50% sequence identity with Ras, including highly conserved core effector domains, and their GTP binding and hydrolysis activities has been confirmed (5). Unlike Ras, the C termini of Rit and Rin lacks a typical prenylation motif (CAAX, XXCC, or CXC) required for the association of Ras proteins with the plasma membrane. However, both proteins contain a well conserved series of basic amino acids at the C terminus, but the significance of this basic domain in their subcellular localization is not clear (6,7).
Recently, we have investigated the ability of Rit and Rin to regulate cell growth, transformation, and several signaling pathways used by other Ras family proteins (8). These studies demonstrated that Rit signals to Ras-responsive elements and transforms NIH3T3 cells to tumorigenicity, but fails to activate the ERK, JNK, p38, or PI3K/Akt kinases, indicating that Rit regulates growth control by different effector pathways than other transforming members of the Ras family (8). These biological activities may, at least in part, result from activation of the RalGEF/Ral signaling cascade. In support of this notion, we have demonstrated that Rit binds and activates the novel Ral-GEF, RGL3, to activate cellular Ral GTPases (9). However, while Rit and Rin interact with the same putative effectors in the two-hybrid system (5), constitutively active Rin does not demonstrate these same biological effects (8). Recently, we have demonstrated that Rin activation is regulated by growth factor-dependent signaling in neuronal cells, suggesting that Rin serves to regulate signaling cascades within the mature nervous system (10).
The rat pheochromocytoma cell line PC12 has served as a model system for examining the mechanism of action of neuronal growth factor (NGF) (11). When treated with NGF these cells acquire a phenotype resembling sympathetic neurons, including the extension of neurites, cessation of cell division, electrical excitability, and expression of genes encoding specific neuronal markers (11). Using neurite outgrowth in PC12 cells as a measure of differentiation, studies have implicated the Ras/ERK pathway in NGF signaling. Ras activation results in activation of the protein kinases Raf-1, B-Raf, and MAP-ERK kinase kinase (MEKK) (12)(13)(14)(15) as well as the catalytic subunit of PI3K, and a series of RalGEFs (2). Introducing activated forms of upstream activators of ERK1/2, including Ras (16), Raf-1 (15), and MEK (17), induces neurite outgrowth in PC12 cells. The dominant-negative forms of Ras, Raf, and MEK block NGF-mediated neuritogenesis in PC12 cells (17)(18)(19). In addition, PI3K has been described to play an important role in PC12 neurite outgrowth (20), and a prominent role has been described for PI3K signaling in the regulation of NGF-dependent survival of PC12 cells and primary neurons (21)(22)(23). PI3K is directly stimulated by Ras (2) and promotes survival through stimulation of the Akt/PKB kinase in numerous cell systems (22,24,25).
In this study, we investigated the ability of the Rit GTPase to induce neurite outgrowth and protect pheochromocytoma cells from serum withdrawal-mediated cell death. To this end, we developed a recombinant adenovirus expressing constitutively active and dominant-negative mutants of Rit. Using these constructs we demonstrate that activated Rit supports robust neurite outgrowth and suppresses serum withdrawal-induced apoptosis in PC6 cells. Rit activates ERK1/2 but not Akt kinase activity and promotes cell survival and neuritogenesis through a MEK-dependent signaling cascade. These studies indicate that Rit functions as a potent inducer of cell survival and differentiation in neural tumor cells and suggest a role for Rit in regulating these same processes in neurons.

EXPERIMENTAL PROCEDURES
Cell Culture-PC6 is a subline of PC-12 cells that produces neurites in response to NGF but grows as well isolated cells rather than in clumps (26) (the generous gift of Thomas Vanaman, University of Kentucky, Lexington, KY). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma) containing 10% (v/v) heat-inactivated horse serum, 5% (v/v) fetal bovine serum (Invitrogen), and 50 g/ml gentamicin at 37°C in a humidified atmosphere of 5% CO 2 . HEK-293 cells were grown in DMEM supplemented with 5% (v/v) fetal bovine serum (Invitrogen) and 50 g/ml gentamicin at 37°C in a humidified atmosphere of 5% CO 2 .
Mammalian Expression Vectors, DNA Transfection, and Immunoblot Analysis-Mammalian expression vectors for wild type and mutant H-Ras and Rit have been described previously (5,8,9). HA-epitopetagged proteins were detected by immunoblotting using anti-HA monoclonal antibody (12CA5) followed by incubation with horseradish peroxidase-conjugated rabbit anti-mouse IgG (Zymed Laboratories Inc.) as described previously (9,(27)(28)(29). Rit expression was detected using anti-Rit monoclonal antibody R0003U (Gamma-1, Lexington, KY). To confirm the specificity of the anti-Rit antibody, monolayers of HEK-293 cells were transiently transfected with HA-tagged expression vectors encoding wild type Rit, Rin, or H-Ras as described previously (9). After a 48-h incubation, whole-cell lysates were prepared in buffer A (20 mM Tris, pH 7.6, 250 mM NaCl, 2.5 mM EDTA, 3 mM EGTA, 20 mM ␤-glycerol phosphate, 1 mM vanadate, 50 mM KF, and 1ϫ protease inhibitor mixture from Calbiochem) and the protein concentration determined by Bradford assay (Bio-Rad). An equal amount of protein from each lysate was separated on 10% SDS-polyacrylamide gels, using a standard SDS-PAGE protocol. After electrophoresis, the gel was transferred to nitrocellulose membrane and probed with either 1 g/ml anti-HA or 2 g/ml of anti-Rit antibody and developed using horseradish peroxidase-conjugated secondary antibodies (Zymed Laboratories Inc.) and chemilu-minescence as described previously (9). To analyze endogenous Rit expression in a variety of cell lines, whole-cell lysates were prepared and subjected to immunoblot analysis using 2 g/ml of anti-Rit monoclonal antibody (R003U) as described above. To examine the activation of ERK and Akt kinases, PC6 cells were infected with Ad-GFP or Ad-RitL79/GFP at m.o.i. values of 200 for 16 h. 48 h after infection, cells were serum-starved overnight and either harvested or pretreated with 100 ng/ml NGF for 10 min as indicated, before lysis in buffer A. Immunoblots were probed with anti-P-ERK (Promega), anti-P-Akt (Cell Signaling Technology), anti-ERK (Cell Signaling Technology), or anti-Akt (Cell Signaling Technology). Stripping and reprobing of membranes were performed as recommended by the manufacturer.
Transfection of PC6 cells was performed using Effectene (Qiagen). For the Ras activation assay, PC6 cells (7.5 ϫ 10 4 cells/60-mm plate) were transfected with the indicated mammalian expression plasmids mixed with 8 ml of Effectene enhancer and 10 l of Effectene reagent according to the manufacturer's protocol.
Neurite Outgrowth Studies-PC6 cells were seeded at 5 ϫ 10 4 cells on coverslips placed in six-well dishes. Coverslips were precoated with 5 g/ml laminin (Sigma) and 25 g/ml poly-L-lysine (Sigma) in PBS for 2 h. PC6 cells were transiently transfected using Effectene (Qiagen) with one of the following plasmids: pEGFP-C1, pEGFP-Rit, pEGFP-RitL79, pEGFP-H-RasL61, pEGFP-RitWT, pEGFP-H-RasWT, pEGFP-RitN35, pEGFP-RasN17, and pKH3-RitL79 either alone or in pairs. Protein expression was examined by epifluorescence microscopy for the GFP fusion proteins or by immunohistochemistry for the HA-tagged proteins. Counting the day of transfection as day 0, cells were either not subject to growth factor stimulation or were treated with NGF (100 ng/ml) on day 2 and then fixed on day 5. Cells were then washed three times in CM-PBS (1.26 mM CaCl 2 , 0.49 mM MgCl 2 , 0.91 mM MgSO 4 ) and fixed with 3.7% formaldehyde in CM-PBS for 20 min at room temperature. Cells were permeablized in 0.1% Triton X-100 in CM-PBS for 5 min, blocked for 30 min with 1% bovine serum albumin in CM-PBS, incubated with Texas Red-phalloidin (Molecular Probes) for 20 min, and washed extensively prior to mounting. For studies using HA-tagged proteins (Rit effector domain studies), slips were fixed and blocked as above, incubated with 2 g/ml anti-HA antibody for 20 min, washed with CM-PBS, and incubated with Texas Red-labeled anti-mouse (1: 1000 dilution, Vector Laboratories). Fixed and stained slips were mounted on glass slides with 12 l of Vectashield (Vector Laboratories) and examined under the appropriate illumination with a ϫ40 objective lens on an E600 microscope (Nikon). Cells were scored positive for neurite outgrowth if one or more neurites exceeded 1 cell body diameter in length.
For studies using recombinant adenovirus, PC6 cells were incubated for 16 h in the presence of virus, after which the medium was changed. RT-PCR from Neuronal Cultures-Primary dissociated cultures of sympathetic neurons were prepared from the superior cervical ganglia of embryonic day-21 rats as described previously (30,31), except that the nonneuronal cells were minimized by incubating the dissociated ganglia for 3 h on plastic culture dishes prior to plating onto collagencoated 35-mm dishes (ϳ5,000 cells/dish). Cultures were maintained in culture medium containing 90% minimal essential medium (Invitrogen), 10% fetal calf serum (Hyclone, Logan, UT), 20 M uridine, and 20 M fluorodeoxyuridine in the presence of 50 ng/ml 2.5 S NGF for either 1 (young) or 4 weeks (old).
Poly(A ϩ ) RNA was isolated from the cultured neurons, converted to cDNA, and specific cDNAs (Rit or NFM) were amplified by subjecting 2% of the cDNA to 25 PCR cycles, which is well within the linear range of PCR amplification for these specific genes and primer pairs as described previously (30,31). After amplification, cDNAs were separated by polyacrylamide gel electrophoresis on 12% gels, stained with SYBR Gold (Molecular Probes), and visualized by phosphoimaging technology (Fuji Medical Systems, Stamford, CT). The identity of the amplified cDNAs was confirmed by DNA sequencing.
Generation of Recombinant Adenovirus-Recombinant adenovirus co-expressing Rit79L and GFP (Ad-RitL79/GFP), Rit35N and GFP (Ad-RitN35/GFP), or GFP alone (Ad-GFP) were generated as described previously (32). Briefly, recombinant adenoviruses were generated through homologous recombination using the Escherichia coli strain BJ5183. pAdTrack constructs were linearized with PmeI and electroporated into BJ5183 bacteria along with the adenoviral plasmid pAdEasy-1 (32). Since, the human Rit cDNA contains an internal PmeI site, QuikChange site-directed mutagenesis (Stratagene) was used to destroy this restriction site, and each cDNA was subsequently cloned directionally into KpnI/XbaI-digested pAdTrack. All cDNAs were verified by DNA sequence analysis. The resulting recombinant adenoviral vector DNAs were digested with PacI and transfected into HEK-293 cells using LipofectAMINE (Invitrogen). Transfected cells were monitored for GFP expression. GFP-positive plaques formed after ϳ10 days in culture and nearly all cells expressed GFP by 14 -16 days, at which time the adenovirus was harvested by scrapping the cells from the plate and pelleting them along with all floating cells in the culture. After three cycles of freezing in a methanol/dry ice bath and rapid thawing at 37°C, 1 ml of viral lysate was used to infect 1 ϫ 10 6 HEK-293 cells. Virus was allowed to amplify until complete lysis of the cells, usually 4 -5 days. The viral lysate was collected and divided between 5-and 10-cm plates each containing 1-3 ϫ 10 6 HEK-293 cells. The efficiency of infection was followed with GFP, and viruses were harvested following complete lysis as described above. At this point, viral titers were often high enough to use for gene expression studies in PC6 cells. Higher titer stocks were generated as described, and CsCl banding purified the resultant viruses; final yields were generally 10 11 to 10 12 plaque-forming units (32,33). Infectious titer was determined by limiting dilution plaque assay on HEK-293 cells. All recombinant adenoviruses were screened for the presence of the wild type E1A gene by PCR in a nonpermissive cell line such as PC6 cells with HEK-293 cell DNA serving as a positive control for the E1A gene. Viruses were used at m.o.i. values of ϳ100 -200.
PC6 Survival Assay-Cell survival was determined by assaying viable cell numbers using the Cell Counting Kit-8 according to the manufacturer's protocol (Dojindo Molecular Technologies). Ad-GFP-, Ad-RasV12-, and Ad-RitL79/GFP-infected naive PC6 cells were washed three times in serum-free DMEM, plated at 15,000 cells/well in polylysine-treated 96-well microtiter plates, and cell viability determined at the indicated times. To examine the survival adenovirus-infected differentiated cells, PC6 cells were stimulated with 100 ng/ml NGF for 7 days prior to viral infection and survival determined as described above. For pharmacological inhibition of PI3K or MEK, 50 M LY 294002 (New England Biolabs), or 50 M PD-98059 (New England BioLabs) were added to cells prior to plating to 96-well dishes, and additional drug was added with serum-free medium following 48 h of incubation. Viable cell counts were performed as above at the indicated times. Each experiment was repeated at least three times in triplicate.

Expression of Rit in Neurons and Pheochromocytoma Cells-
Although we, and others, have demonstrated that Rit mRNA is ubiquitously expressed, including expression within neuronal tissues, it is not known whether Rit is expressed in neurons (5-7). To this end, primary cultures of superior cervical ganglia were prepared from embryonic day 21 rats and incubated in the presence of NGF for 6 days (immature sympathetic neurons) or 30 days (mature neurons) and subjected to RT-PCR analysis. As shown in Fig. 1A, Rit mRNA was detected in both mature and immature cultures. To extend these results, and to identify a cultured cell system in which to examine the biological function of Rit, we examined Rit expression in a variety of cultured cell lines by Western blot analysis. Using an anti-Rit monoclonal antibody (Fig. 1B), a protein of ϳ28 kDa was detected in a variety of cultured cell lines (Fig. 1C). Endogenous Rit protein was abundantly expressed in both NIH3T3 and PC6 cells, and at much lower levels in C2C12 cells, but not in SH-SY5Y or MIN6 cells.
Neurite Outgrowth in PC6 Cells Is Induced by Rit Overexpression-Pheochromocytoma cells respond to the expression of activated Ras proteins by the cessation of cell growth and the extension of neurites (16,35). This differentiation process is characterized by rearrangements of the actin cytoskeleton at the plasma membrane, leading to the formation of lamellipodia, growth cones, and the subsequent extension of axon-like processes (11). To determine whether activated Rit could trigger this biological response, we used transient transfection to introduce expression plasmids that encoded GFP-tagged activated versions of Rit and Ras into PC6 cells and asked if expression of these fusion proteins induced neurite extension, the most prominent marker of PC6 cell differentiation. As shown in Fig. 2A, transfection of PC6 cells with a control plasmid expressing unfused GFP caused no detectable change in morphology relative to untransfected PC6 cells. These cells continued to proliferate and showed the same limited adherence and round shape characteristic of the parental PC6 cells. In contrast, PC6 cells transfected with DNA encoding the constitutively active Ras61L protein developed neurites that attained a length of greater than one cell bodies after ϳ36 h (see Fig. 2A). There were two or occasionally three of these axonlike extensions per cell (Fig. 2C), which continued to elongate over the next 48 -72 h. The processes were smooth and extended linearly, with a single growth cone and little secondary branching. In contrast, PC6 cells transiently transfected with an expression vector for constitutively active Rit (GFP-RitL79) developed multiple neurites ( Fig. 2A), which attained a length of greater than one cell body after ϳ48 h. Cells expressing GFP-RitL79 often produced four or more neurite extensions (Fig. 2C), with GFP-RitL79 localized predominantly to the plasma membrane including the neurite-like projections ( Fig.   FIG. 1. Rit expression in neurons and neuronal cell lines. A, the expression of Rit at the mRNA level was determined by performing RT-PCR on 2% of the cDNA generated from RNA isolated from embryonic day 21 primary rat superior cervical ganglia neurons as described under "Experimental Procedures." Neuronal cultures were incubated in the presence of 50 ng/ml 2.5 S NGF for either 6 days (immature neurons) or 30 days (mature neurons) before RNA isolation. B, wholecell lysates were prepared from human embryonic kidney 293 cells that were transiently transfected with expression vectors encoding HAtagged wild type Rit, Rin, or H-Ras. Fifty g of protein from each lysate was resolved by SDS-polyacrylamide gel electrophoresis on 10% gels and analyzed by immunoblotting with monoclonal anti-Rit (top panel) or monoclonal anti-HA (bottom panel) antibodies. The migration of molecular weight standards is indicated on the right. The blots were exposed to film for 30 s. C, fifty g of whole-cell lysate prepared from the indicated neuronal (PC6 and SH-SY5Y), fibroblast (NIH3T3), ␤-islet cell (MIN-6), or muscle (C2C12) cell lines was analyzed by immunoblotting with anti-Rit monoclonal antibody as described under "Experimental Procedures." The blot was exposed to film for either 30 s (top panel) or 4 min (bottom panel). These data are representative of those obtained in at least three separate experiments.
2A). Furthermore, as differentiation progressed, the extensions became more highly branched, wandered in various directions, and displayed numerous fine filopodia along their entire length. The extent of GFP-RitL79 mediated neurite extension was similar to that for GFP-RasL61, with more than 80% of transfected cells developing neurite extensions (Fig. 2B). Indeed, overexpression of wild type Rit was sufficient to induce similar morphological changes, although only ϳ40% of transfected cells developed neurites. The biological activity of Rit was therefore quite robust in pheochromocytoma cells, equal to the biological function of activated Ras, but resulted in morphological changes distinct from the axon-like extensions caused by expression of GFP-RasL61.
Rit-mediated Neurite Outgrowth Is ERK1/2-dependent-Before characterizing the signal transduction cascades used by Rit to mediate neuritogenesis, we developed an adenoviral expression system to allow efficient expression of Rit in PC6 cells as described under "Experimental Procedures." Adenovi-rus-mediated Rit expression was necessary, because repeated efforts to generate stable PC6 cell lines containing RitL79 under an inducible promoter system failed to produce stable PC6 cell lines demonstrating regulated RitL79 expression. The adenovirus system provides broad-spectrum infectivity, a low level of in vitro toxicity, and expression of GFP from a second promoter allows for infection efficiency to be determined using epifluorescence microscopy. In addition, high titer virus stocks are easily obtained and allow control over protein expression by altering the viral dose (32,33). To assess the ability of adenoviruses to infect and express recombinant genes, PC6 cells were infected at a m.o.i. (infectious viral particles/cell) from 10 to 200 and cultured for 48 h (Fig. 3). A majority (Ͼ80%) of the PC6 cells infected with RitL79/GFP recombinant adenovirus expressed both proteins, and expression of activated Rit induced cellular differentiation (Fig. 3B). We were not able to achieve 100% infectivity, likely due to limited cell division after infection. All further experiments were performed at m.o.i. greater Sustained activation of the MEK/ERK pathway has been suggested to play a critical role in oncogenic Ras-and NGFmediated differentiation of PC12 cells into a neuronal phenotype (36, 37). Since we have previously shown that activated Rit fails to induce ERK activation when expressed in COS cells (8), we assessed the ability of Rit to activate ERK1/2 when expressed in PC6 cells by examining their phosphorylation state. Antibodies that specifically recognize the active phosphotyrosine and phosphothreonine form of ERK1/2 were used in these studies. PC6 cells were infected with RitL79 or GFP adenoviruses and lysates from these cells were probed in immunoblots with anti-phospho-ERK. RitL79 expression stimulated the p44 and p42 ERK MAP kinases (Fig. 4A). Infection with control GFP-expressing adenoviruses did not result in ERK phosphorylation, although as expected, NGF stimulation of cells infected with GFP adenovirus resulted in potent ERK activation (Fig. 4A).
Since activation of the ERK kinases appeared to play a role in Rit-mediated neurite outgrowth, we next tested whether endogenous activity of the MEK/ERK cascade was required. Infection with adenovirus encoding dominant-negative intermediates in the Raf/MEK/ERK pathway was used to systematically interfere with Raf/ERK signaling (21). Fig. 4B shows that co-infection with either dominant-negative Raf1 or MEK1 adenoviruses, but not with GFP control adenovirus, potently suppressed constitutively active Rit-mediated neurite outgrowth. Indeed, these dominant-negative adenoviruses inhibited activated Rit and activated H-Ras-dependent differentiation to approximately the same extent, strongly supporting a role for MEK activity in Rit-mediated PC6 cell differentiation.
Expression of Dominant-negative Rit Inhibits NGF-mediated Neurite Outgrowth-Since expression of constitutively active Rit was alone sufficient to induce neurite outgrowth, we next examined the influence of Rit function on neurite outgrowth by expression of a dominant-negative Rit (RitN35) mutant. By analogy with other Ras-related GTPases, RitN35 is expected to be maintained in a conformational state that inhibits binding to downstream effector targets, but retains high affinity for its guanine nucleotide exchange factor (GEF), thus preventing activation of endogenous wild type Rit (38). After transient transfection with vectors expressing either GFP alone, GFP-RitN35, or GFP-H-RasN17, PC6 cells were treated with NGF, and 72 h later the morphology of GFP-expressing cells was examined. Expression of GFP did not alter NGF-induced neurite outgrowth (Fig. 5A). On the other hand, expression of both dominant-negative Rit and Ras inhibited neurite outgrowth. In these experiments, RitN35 could not completely suppress NGFinduced neurite outgrowth, although it was as effective as dominant-negative H-Ras, a proven inhibitor of this process (19). Similar results were obtained using RitN35 expressing recombinant adenovirus (data not shown). RitN35 and Ras17N had no significant effect on cell morphology or survival of the transfected cells compared with that of untransfected controls after 3 days (Fig. 3B), but reduced cell viability by approximately the same extent after 5 and 7 days in culture. Taken together, these results suggest that endogenous Rit function may be important to both PC6 cell viability and NGF-mediated neuritogenesis.
Stimulation of PC12 cells with NGF induces rapid and sustained phosphorylation of the ERK kinases (36). Therefore, we tested how dominant-negative Rit affected NGF induced ERK activation. PC6 cells were infected with GFP control and RitN35 recombinant adenovirus and 48 h later stimulated with NGF (100 ng/ml) for 10 min. Fig. 5B shows that expression of RitN35 resulted in an almost complete inhibition of ERK acti- vation as measured by anti-phospho-ERK Western blotting. Control adenovirus infection had no effect on NGF-stimulated ERK phosphorylation, controlling against possible effects of the adenovirus vector system.
The expression of activated Raf and MEK kinases cause terminal differentiation of PC12 cells, strongly supporting a central role for activation of the Raf/ERK pathway in neuritogenesis (15,17). Since dominant-negative Rit inhibited NGFinduced ERK stimulation (Fig. 5B), we next assessed the ability of dominant-negative Rit to inhibit activated Raf1-mediated neurite outgrowth. PC6 cells were co-infected with constitutively active Raf1 and either GFP expressing control, dominant-negative MEK1, or RitN35-expressing adenovirus, and 72 h later the infected cells were examined. Expression of dominant-negative Rit did not alter active Raf1-mediated neu-

FIG. 5. Rit effects NGF-mediated signaling in PC6 cells.
A, dominant-negative Rit suppresses NGF-mediated neurite outgrowth. PC6 cells cultured on polylysine-coated glass coverslips were transiently transfected with expression vectors encoding either GFP, GFP-RasN17, or GFP-RitN35, allowed to recover for 24 h, and then treated with 100 ng/ml NGF for 72 h as described under "Experimental Procedures." Cells were fixed, processed for epifluorescence, and random fields scored for neurite outgrowth as described in the legend to Fig. 2. GFP-expressing cells exhibiting neurites (1 cell body or greater in length) and cells without neurites were counted. Approximately 700 -1,000 cells/condition were counted from three separate experiments. B, dominant-negative Rit inhibits NGF-mediated ERK activation. PC6 cells (7.5 ϫ 10 5 cells/plate) were plated on polylysine-coated 60-mm dishes and transiently transfected with expression vectors encoding either pKH3 empty vector control (lanes 1 and 2) or pKH3-RitN35 (lanes 3 and 4). After 48 h, cells were serum-starved for 16 h and treated with 100 ng/ml NGF for 10 min as indicated (lanes 2 and 4). Fifty g of protein was resolved by SDS-polyacrylamide gel electrophoresis on 10% SDS-polyacrylamide gels and analyzed by immunoblotting with antiphospho-ERK (top panel), anti-HA (middle panel), or anti-ERK (bottom panel) antibodies. Similar results were obtained in several independent experiments. C, dominant-negative Rit does not inhibit activated Raf1meditated neurite outgrowth. PC6 cells were co-infected with recom-rite outgrowth; the population of cells with long neurites (exceeding two times the length of the cell body) was equal to that of control cells co-infected with active Raf1 and GFP adenoviruses (Fig. 5C). However, as expected, dominant-negative MEK1 potently inhibited neurite outgrowth. Taken together, these results suggest that dominant-negative Rit disrupts NGF-induced neurite outgrowth by inhibiting the MEK/ERK pathway at a level at or above that of the Raf kinases.
Constitutively Active Rit Stimulates PC6 Cell Survival without Activating PI3K-Since the expression of constitutively active Ras has previously been shown to prevent serum withdrawal-mediated apoptosis in PC12 cells (21), we next assessed the ability of activated Rit to promote the survival of naïve PC6 cells. As seen in Fig. 6A, adenovirus-mediated expression of activated Rit completely blocked apoptosis induced by growth factor deprivation. Epifluorescence microscopy demonstrated that RitL79-expressing cells showed a healthy profile with robust neurites despite the absence of growth factors. The Ad-GFP virus showed no toxicity in these experiments (data not shown). We also asked whether Rit expression was sufficient to mediate survival of differentiated PC6 cells. Following NGF-mediated neuronal differentiation, PC6 cells were infected with Ad-RitL79/GFP and cultured in serum-free medium without NGF and the percentage of surviving cells determined by MTT assay. We again found that RitL79 expression completely inhibited cell death (Fig. 6A).
Although previous studies have demonstrated a critical role for the MEK/ERK signaling cascade in NGF-mediated survival, these data have identified PI3K/Akt as the primary mediators of this protective effect in both PC12 and primary neurons (21,39). Since we have shown that Rit does not stimulate Akt activity in COS cells (8), we analyzed Rit-mediated Akt activation in PC6 cells by immunoblotting with a phosphorylationspecific Akt antibody, a direct correlate to Akt kinase activity in sympathetic neurons (40). Akt phosphorylation was robustly increased upon NGF stimulation of control adenovirus (GFP) infected PC6 cells (Fig. 6B). In contrast, activated Rit expression did not result in Akt phosphorylation. Taken together, these studies indicate that RitL79 is capable of effectively stimulating survival in both naïve and differentiated PC6 cells without activating the PI3K/Akt pathway. Fig. 4 suggests that MEK activity may be necessary for RitL79-mediated cell survival. However, previous studies have demonstrated that activated Ras-mediated cell survival depends predominantly upon PI3K/Akt signaling pathway (reviewed in Ref. 39). To determine the requirement for MEK and PI3K activities in RitL79-mediated survival, adenovirus infected cells were treated with the MEK inhibitor PD-98059 (50 M) or the PI3K inhibitor LY 294002 (50 M). Rit-mediated cell survival was sensitive to MEK inhibition, although as reported previously, Ras-mediated survival was MEK independent (Fig. 7A). Thus, MEK-dependent pathways are necessary for Rit-mediated, but not Ras-mediated, cell survival. Moreover, treatment of infected cells with PI3K inhibitor completely suppressed cell survival mediated by both activated Rit and Ras (Fig. 7B). Since activated Rit does not stimulate the PI3K/Akt signaling pathway, these results suggest that some level of endogenous PI3K activity is required for PC6 cell survival.

Inhibition of MEK Activity Suppresses Rit-induced Cell Survival-The experiments of
To confirm that the PD-98059 and LY 294002 inhibitors suppressed the activities of MEK and PI3K pathways, respectively, the phosphorylation states of Akt and ERK1/2 were examined in PC6 cells following infection with adenovirus expressing RitL79 or RasV12. PD-98059 inhibited Rit-induced phosphorylation of ERK1/2 but did not inhibit RasV12-induced activation of Akt (Fig. 7, A and B), whereas LY 294002 treatment inhibited RasV12-induced phosphorylation of Akt (Fig.  7B). Taken together, these studies suggest that Rit-mediated survival signaling is MEK-dependent and distinct from survival mechanisms controlled by H-Ras. DISCUSSION The Rit, Rin, and Drosophilia Ric proteins constitute a novel branch of the Ras subfamily, whose conservation from flies to humans suggests conservation of important physiological functions. We have previously shown that Rit can signal to Rasresponsive elements and transform NIH3T3 cells to tumorigenecity without activating the ERK, JNK, p38, or PI3/Akt kinases, indicating that Rit regulates growth control by different effector pathways than other members of the Ras subfamily  (8). In this report, we demonstrate that (1) overexpression of constitutively active Rit induces cell survival and neurite outgrowth in PC6 cells, 2) that Rit is capable of activating ERK, but not PI3K/Akt, signaling pathways in PC6 cells, and 3) that MEK activity is required for both Rit-mediated cell survival and neuritogenesis. These studies provide new insight into Rit function and suggest that Rit might play a role in regulating cell survival and differentiation within the nervous system.
Rit, like most Ras family members, is expressed in a wide variety of tissues, which may reflect its role in regulating a ubiquitous signal transduction cascade(s). Before beginning to investigate the physiological function of Rit in neuronal cells, we examined its expression in primary neurons. Rit mRNA was detected in both immature and NGF aged cultures of primary superior cervical ganglia cells (Fig. 1A), confirming the neuronal expression of Rit. However, Western blot analysis utilizing monoclonal anti-Rit antibody detected Rit expression in only a subset of mammalian tissue culture cells (Fig. 1C), suggesting that Rit function might not be essential to the survival of all cell types.
The expression of activated Rit is capable of inducing potent neurite outgrowth in PC6 cells (Fig. 2). Thus, Rit joins a select subset of Ras subfamily GTPases, including the classical Ras proteins (H-Ras, N-Ras, and K-Ras) (16) and TC21/R-Ras2 (41) in its ability to stimulate neurite outgrowth in pheochromocytoma cells. However, while our previous studies have defined Rit as a weakly transforming GTPase in NIH3T3 cells (8), RitL79-mediated neuritogenesis was remarkably potent. When transiently transfected into PC6 cells, GFP-RitL79 induced neurite outgrowth in ϳ80% of all cells, while ϳ65% of cells expressing GFP-H-RasL61 demonstrated neuritogenesis (Fig.  2B). In addition, the morphology of Rit-and H-Ras-transfected PC6 cells were strikingly different. PC6 cells expressing activated H-Ras developed an average of two axon-like extensions/ cell (Fig. 2, A and C). In contrast, activated Rit induced a distinct morphology; cells extended an average of four to five extensions/cell, and as the differentiation of these cells progressed, the extensions became more highly branched and developed fine filopodia along their length (Fig. 2, A and C). The molecular mechanisms regulating the cytoskeletal changes necessary for neurite outgrowth are only beginning to be understood (42). Rho family GTPases have the ability to alter various actin-derived morphologies and have been ascribed roles in axonal growth and guidance, dendritic formation, neurite formation, and for causing serum-dependent neurite retraction in N1E-115 cells (42,43). Whether Rit modulates the activity of a Rho family member, or the activity of another protein involved in neurite outgrowth, remain to be addressed.
In PC12 cells, constitutive Ras activity promotes cell cycle arrest and differentiation into a neuronal phenotype (16,35). Differentiation appears to require Ras-induced Raf activation, since constitutively active versions of Raf and MEK can mimic, at least in part, the effects of activated Ras (15,17). Moreover, inhibition of the Ras-MEK-ERK pathway blocks stimulus-induced neurite outgrowth (19). Thus, neurite outgrowth in PC12 cells requires the activity of MEK and ERKs (17) and perhaps p38 (44). Our previous studies have shown, at least in COS cells, that Rit does not activate these kinase pathways (8). However, immunoblot analysis using anti-phospho-ERK antibodies confirm that activated Rit up-regulates phosphorylated ERKs in PC6 cells (Fig. 4A). In addition, RitL79-induced neuritogenesis was strongly suppressed by dominant-negative Raf1 and MEK1 (Fig. 4B), suggesting that MEK/ERK activity is responsible for transmitting Rit-mediated neurite outgrowth signals. Additional support comes from the observation that dominant-negative Rit inhibits NGF-induced neurite out-growth, at least in part, by potently suppressing ERK activity (Fig. 5). The finding that dominant-negative Rit fails to inhibit neurite outgrowth induced by expression of constitutively active C-Raf (Fig. 5C) indicates that dominant-negative Rit functions to inhibit NGF-mediated activation of the MEK/ERK cascade at or above the level of Raf. Taken together, these results suggest that Rit activates ERKs via a MEK-dependent pathway. A greater understanding of Rit-dependent kinase pathways will be the topic of future investigation.
How Rit expression might result in cell type-specific activation of the ERK kinases is unclear, but several possibilities can be envisioned. Although initial yeast two-hybrid studies failed to detect a direct interaction between Rit and any member of the Raf kinase family (5), it is possible that Rit regulates a tissue restricted Raf-dependent mechanism to activate ERK1/2. Indeed, the recent finding that the Ras-related GTPase Rap1 activates ERK activity in cells expressing B-Raf, which is predominantly expressed in neuronal cell types (45), but antagonizes ERK activity in cells which do not express B-Raf is particularly intriguing (46 -48). A similar signaling mechanism might explain the cell-specific coupling of Rit to ERK1/2 activation. Alternatively, Rit might activate a non-Raf member of the MEKK family, such as MEKK-2 or MEKK-3 (45) or regulate ERK function via a MEK-independent mechanism (49). It is also possible that an adaptor or effector molecule necessary for Rit-mediated regulation of MEK function might demonstrate a restricted distribution. The regulation of Raf proteins is complex, and a number of mechanism have been characterized that might play a role in this process, including the action of various protein kinases and protein scaffolds (reviewed in Ref. 45). Finally, cell type-dependent variability in the subcellular localization of Rit and Raf/MEKK proteins could also account for the differences we have observed. Clearly, additional studies will be needed to clarify the signaling events involved in Rit-dependent stimulation of ERK1/2 activity.
Studies reported here also demonstrate that the Rit requires, in part, MEK activity (Fig. 7), but not PI3K/Akt signaling (Figs. 6B and 7) to promote PC6 cell survival. Thus, RitL79-mediated, but not RasL61-induced, PC6 cell survival was strongly suppressed by pharmacological inhibition of MEK activity, indicating that MEK activity is responsible for transmitting Rit-mediated survival signals. However, RitL79-induced cell survival was also potently suppressed by inhibition of PI3K activity (Fig. 7B). Since Rit does not stimulate phospho-Akt in PC6 cells, these results indicate that PI3K/Akt signaling plays an essential role in PC6 cell survival. Evidence for the role of PI3K/Akt pathways in NGF-mediated survival of neuronal cells is well documented (21,23,34,40). PI3K/Akt regulates cell survival by inhibiting the activities of the cell death proteins BAD (24), the transcription factor Forkhead (50), and the JNK/p53 cell death pathway (51). In contrast, the role of the MEK/ERK pathway in mediating anti-apoptotic effects of neurotrophins is less clear (reviewed in Ref. 39). Many studies have suggested that ERK is not a major mediator of the neuroprotection afforded by neurotrophins or membrane depolarization in both PC12 cells and a variety of primary neurons following trophic withdrawal (21,34,(51)(52)(53). However, MEK activity can mediate cell survival in neurotrphin-regulated systems, including differentiated PC12 cells (54) and cultured cerebellar granule neurons (55). Indeed, it has been shown recently that TrkB uses both PI3K and MEK signaling pathways to promote neuronal survival (56) and that sympathetic neuronal survival promoted by activated Ras occurs partially through the MEK/ERK pathway (51). This survival pathway may function by activating the transcription factor CREB, which is a critical regulator of NGF-mediated neuronal survival (57).
Sensory and sympathetic neurons undergo apoptosis when deprived of trophic factor support in vitro. Neurotrophins potently stimulate neuronal survival in part by activating Ras. The results of this study suggest that Rit, acting through a MEK-dependent signaling pathway, regulates signaling pathways that control neuronal differentiation and survival mechanisms. Rit might function as a primary target of neurotrophic factor signaling or provide an additional mechanism for activating MEK/ERK in a trophic factor-independent manner (58,59). Therefore, it will be important to determine whether there are specific GEFs and GAPs that regulate Rit function and whether extracellular stimulus-mediated signaling pathways in neurons regulate cellular Rit-GTP levels. The recombinant adenovirus developed in this paper will be excellent reagents to test the role of Rit-mediated signal transduction in primary neurons.