Two Different Signal Transduction Pathways Are Implicated in the Regulation of Initiation Factor 2B Activity in Insulin-like Growth Factor-1-stimulated Neuronal Cells*

Eukaryotic initiation factor eIF-2B plays an important role in translation regulation and has been sug-gested to be implicated in the increased protein synthesis promoted in response to growth factors. We have used primary cultured neurons to delineate the signaling pathways by which insulin-like growth factor-1 (IGF-1), which plays a critical role in the survival of neuronal cells, promotes eIF-2B and protein synthesis activation. Treatment of cortical neurons with IGF-1 (100 ng/ml) for 30 min stimulates [ 3 H]methionine incorporation, and a parallel increase in eIF-2B activity was observed. Wortmannin and LY294002 reversed both effects, indicating that phosphatidylinositol 3-kinase mediates IGF-1-induced protein synthesis and eIF-2B activation. IGF-1 induced glycogen synthase kinase-3 (GSK-3) inactivation in a phosphatidylinositol 3-kinase-dependent fashion because it is inhibited by wortmannin and LY294002. By using GSK-3 immunoprecipitated from untreated and IGF-1-treated cells, we demonstrate the phosphorylation of eIF-2B coincident with its inactivation. The treatment of cortical neurons with IGF-1 also promoted the activation of mitogen-activated protein kinase (MAPK). The MAPK-activating kinase (MEK) inhibitor PD98059 inhibited MAPK activation and reversed IGF-1-induced protein synthesis and eIF-2B activation. These findings suggest that IGF-1-induced eIF-2B activation on neurons is promoted through phosphatidylinositol 3-kinase and GSK-3 kinase, and we report an IGF-1-induced MEK/MAPK activation pathway implicated in eIF-2B activation.

Translational control is pivotal in the intracellular action by which growth factors promote their effects. Protein synthesis is activated in different cell types by a variety of growth factors, which are determinant for cell growth, differentiation, and survival. The pathway through which this mechanism is exerted is not well characterized, at least in neurons of the central nervous system. Translational control starts at the level of initiation (1,2). The polypeptide chain initiation de-pends on initiation factor 2 (eIF-2), 1 which is required for ternary complex formation (eIF-2⅐GTP⅐Met-tRNA i ), as well as the binding of mRNA to the ribosomes and recognition of the initiator codon, and constitutes the steps subject to the finest regulation. The eIF-2 factor is required for priming each 40 S ribosomal subunit for every round of translation (3)(4)(5), but upon 80 S initiation complex formation, GTP is hydrolyzed and eIF-2 is released from the ribosome as a binary complex (eIF-2⅐GDP), which is functionally inactive. Initiation factor 2B (eIF-2B) promotes eIF-2 recycling by catalyzing the dissociation of GDP from eIF-2⅐GDP in the presence of GTP (4,6), making eIF-2 available to undergo a further round of initiation.
(PI3-K) (25). Recent findings have shown that the direct link between PI3-K activation and GSK-3 inactivation is provided by protein kinase B (PKB), which is located downstream of PI3-K and phosphorylate GSK-3 at the serine regulatory sites (26). So far, this signaling pathway has been found to be independent of mitogen-activated protein kinase (MAPK) (21,24,25,27).
Insulin-like growth factor-1 (IGF-1) plays a crucial role in growth and cell development regulation and exerts its action by activating multiple signal transduction pathways, notably Ras-MAPK and PI3-K pathways (28 -31). IGF-1 promotes survival of central nervous system neurons, where the PI3-K pathway via PKB activation seems to be critical and where the activation of MEK/MAPK pathway fails or has not been described (29,(31)(32)(33). So far, the possible role of eIF-2B in IGF-1-mediated response of neurons has not been established.
We have investigated the effect of IGF-1 on protein synthesis and eIF-2B activities in primary cultures of cortical neurons. Furthermore, by using specific inhibitors of PI3-K and MEK (MAPK/extracellular signal-regulated kinase-activating kinase) and studying GSK-3 and MAPK, we have delineated the two signal transduction pathways involved in eIF-2B activation. In the present paper, we first report two new findings. First, our findings provide the link between IGF-1-induced cellular growth and protein synthesis and eIF-2B activation. Secondly, we present evidence for the involvement of both PI3-K and MEK signaling pathways in IGF-1-dependent eIF-2B activation on neuronal cells.
Primary Neuronal Cultures-Primary cultures of cells from cerebral cortex were prepared from 16-day-old fetuses removed from timed pregnant Harlan Sprague-Dawley rats. Fetuses were placed in Leibovitz L-15 medium for brain dissection. The cerebral cortex was separated from the rest of the brain using iridectomy scissors, and the meningeal membranes were carefully removed. The resulting pieces were then dissociated using Pasteur pipette and 20 -21G needles into a homogeneous cell suspension. Trypan blue exclusion was used to count the living cells. Neurons was seeded on plastic multidishes precoated with 0.05 mg/ml poly-D-lysine at a density of 2-2.5 ϫ 10 5 cells/cm 2 and cultured at 37°C with 5.5% CO 2 in air, in Dulbecco's medium with 15% fetal calf serum. After 24 h, cultures were placed and maintained in serum-free medium Dulbecco's/Ham's F-12 (1:1, v/v) (D:F medium) supplemented with 1.8 mg/ml glucose, 100 g/ml transferrin, 100 M putrescine, 20 nM progesterone, and 30 nM sodium selenite. 6 -7-day-old neurons in culture were used in the experiments. The neuronal content, as determined by immunocytochemistry with antibodies against the neuron-specific protein ␤-tubulin isotype III, was found to be more than 90%. Cells were maintained in D:F medium without supplements for 16 h before treatments and then placed in the same medium in the absence or presence of additives. When inhibitors were used, cells were treated with the inhibitors for 1 h before and during the treatment. Cells were washed with ice-cold phosphate-buffered saline before harvesting.
Protein Synthesis Rate Measurement-Protein synthesis rate was assessed in 16-mm multidishes where the medium was aspirated and replaced with 0.25 ml of fresh medium containing 0.35 Ci/mmol [ 3 H]methionine (115 M). After incubation, the dishes were washed twice with phosphate-buffered saline, and the cells were harvested in 0.25 ml of 0.25% Nonidet P-40 in buffer 20 mM Tris-HCl, pH 7.6, 10 mM potassium acetate, 1 mM dithiothreitol, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine and centrifuged for 30 min at 12,000 ϫ g. The protein present in the supernatant was precipitated with 10% trichloroacetic acid, and radioactivity was determined by liquid scintillation.
eIF-2B Activity Measurement-Both untreated and treated cells cultured on 35-mm multidishes were lysed for 10 min in hypotonic buffer (10 mM Tris-HCl, pH 7.6, 10 mM KCl, 1 mM dithiothreitol, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 g/ml leupeptin, pepstatin, and antipain, 2 mM ␤-glycerophosphate, 2 mM sodium molybdate, and 0.2 mM sodium orthovanadate). The lysate was made to 4 mM magnesium acetate and 140 mM potassium acetate and was centrifuged for 10 min at 12,000 ϫ g. All the steps were carried out at 4°C. A binary complex eIF-2⅐[ 3 H]GDP was formed as described (35). eIF-2B activity present in cell extracts (60 g of protein) was measured for its capacity to exchange eIF-2-bound [ 3 H]GDP for free GDP (35). 1 pmol of eIF-2⅐[ 3 H]GDP was used as a substrate, and the GDP exchange reaction was incubated for 5 min. eIF-2B activity was expressed as a percentage or pmol of [ 3 H]GDP released from binary complex.
eIF-2␣ Phosphorylation Determination-Both untreated and treated cells cultured on 35-mm multidishes were harvested, lysed in 8.5 M urea, 5% 2-mercaptoethanol, 5 mM sodium phosphate, 70 mM sodium chloride (10 ϫ 10 6 cells/ml), and kept for 30 min at room temperature. The cell lysate was then centrifuged at 18,500 ϫ g for 30 min, and the supernatant was resolved in horizontal isoelectric focusing slab gels. The bands corresponding to eIF-2␣ and eIF-2␣P proteins were stained on immunoblots with a polyclonal antibody (1:100) purified by immunoaffinity and quantified as described (36) with an image analyzer with software package (DiversityOne, PDI, New York).
eIF-2B Level Determination-eIF-2B levels were determined in cell extracts, prepared in the same way as that for eIF-2B assay, from both untreated and treated cells by SDS-polyacrylamide gel electrophoresis (PAGE) and Western blots. The blots were stained with a monoclonal antibody (1:500) raised against the ⑀ subunit of eIF-2B (eIF-2B⑀), a generous gift from Dr. Kimball (University of Hershey, Hershey, Pennsylvania). Quantification of bands was as described above.
GSK-3 Activity and eIF-2B⑀ Phosphorylation Determinations-Cell extracts (10 g), prepared in the same way as that for eIF-2B assay, from both untreated and treated cells were assayed for GSK-3 activity in 50 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 100 M ATP, 1 Ci of [␥-32 P]ATP (0.5 Ci/mmol), with 12 g of RRAAEELDSRAGS(P)PQL eIF-2B⑀-based peptide used as a substrate, or with 12 g of RRAAEELDSRAGAPQL as a negative control as described (37), in a volume of 25 l. After 15 min at 30°C, 10 l of the sample was spotted onto P81 phosphocellulose paper, rinsed three times with 3% (v/v) phosphoric acid, and counted. The other 10-l sample of the reaction mixture was analyzed by SDS-PAGE run at a pH of 7.8 (to avoid the peptide exclusion) followed by autoradiography and quantified with an image analyzer as described above.
For eIF-2B⑀ phosphorylation by GSK-3, 50 g of cell extract was immunoprecipitated with 3 l of anti-GSK-3␤ and 25 l of protein G-Sepharose as described (38). One-half of the immunoprecipitated sample was incubated in 10 mM Hepes, pH 7.4, 0.2 mM EDTA, 10 mM MgCl 2, 25 M ATP, 7 Ci of [␥-32 P]ATP, and 1 g of purified eIF-2B in a volume of 25 l for 15 min at 30°C. The reaction was stopped by adding 12.5 l of SDS sample buffer and then resolved by SDS-PAGE. The dried gel was stained and exposed to film. The resulting autoradiographs were quantified as described above. The other fraction was incubated in the same conditions without radioactive ATP and with 0.3 g of purified eIF-2B and assayed for eIF-2B activity as described above.
Phospho-specific GSK-3 and MAPK Western Blots-Cell extracts, prepared in the same way as that for eIF-2B assay, was analyzed by SDS-PAGE. After electrophoresis, the gels were transferred to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech), and the blots were visualized by using specific polyclonal antibody that recognizes the phosphorylated GSK-3␤(Ser 9 ) (inactive form of the kinase). The total amount of the kinase was detected with a monoclonal antibody, recognizing both unphosphorylated and phosphorylated forms of GSK-3␤. To detect the ERK1 and 2 MAP kinases, phosphorylation protein immunoblot analysis was performed by using anti-ERK1 and 2 diphosphorylated monoclonal antibody, which recognizes neither the nonphosphorylated nor the monophosphorylated forms of the enzyme (inactive forms). Equal loading of samples was checked with an anti-ERK1 and 2 polyclonal antibody.
Statistical Analysis-Results are expressed as the mean Ϯ S.E. values for independent experiments. Statistical analysis was performed using t test for paired and unpaired data versus control values or analysis of variance and Dunnett's post-test for comparisons between treated groups.

IGF-1 Effect on Protein
Synthesis-Cultured neurons were treated with IGF-1 for 30 min, 60 min, or 2 h, and the protein synthesis rate was determined in the presence of 0.35 Ci/mmol [ 3 H]methionine (115 M). As shown in Fig. 1A, amino acid incorporation was linear (r 2 , 0.994 both) at all the times assayed. Protein synthesis was significantly increased after exposure to 10 g/ml insulin (not shown) and 100 ng/ml IGF-1 1.25-and 1.3-fold over control, respectively (Fig. 1A). A concentration of 0.1 M wortmannin, a PI3-K inhibitor, was enough to block the activation of [ 3 H]methionine incorporation produced by IGF-1 (Fig. 1B), suggesting that PI3-K pathway is necessary for protein synthesis activation. Moreover, PI3-K inhibitor LY294002 (30 M) reduced protein synthesis rate below 0.1 M wortmannin level. The MEK inhibitor PD98059 (30 M) inhibited protein synthesis rate below control levels. This finding suggests a role for MEK/MAPK kinases activation in growth factor-induced protein synthesis in our cultures. eIF-2B Activation by IGF-1-Neuronal cells were treated for 30 min with growth factors and processed to measure eIF-2B activity as described above (see "Experimental Procedures"). Insulin (10 g/ml) increased eIF-2B activity by 32% (not shown), whereas IGF-1 (100 ng/ml) induced a 45% increase (Fig. 2) and nerve growth factor (100 ng/ml) did not have any effects on eIF-2B activity (not shown). Fig. 2A shows the linearity (r 2 , 0.959) of the reaction between 0.5 and 1.5 pmol of substrate (binary complex, eIF-2⅐[ 3 H]GDP). Wortmannin, PD98059, and LY294002 inhibited IGF-1-induced activation of eIF-2B (Fig. 2B), paralleling the effects produced on protein synthesis (Fig. 1B). The inhibitors did not produce any effects on untreated control cells or eIF-2B activity assay at the concentrations tested (not shown).
eIF-2␣ Phosphorylation State and eIF-2B Levels-IGF-1-induced increased eIF-2B activity, which might be due to decreased levels of their physiological inhibitor (eIF-2␣P), or might be a reflection of increased cellular levels of the factor. Consequently, untreated and IGF-1-treated cells for 30 min were lysed and analyzed by horizontal isoelectric focusing slab gels and protein immunoblot to determine eIF-2␣ and eIF-2␣P levels or by SDS-PAGE and protein immunoblot to achieve eIF-2B⑀ levels. The bands corresponding to eIF-2␣ and phosphorylated eIF-2␣ (eIF-2␣P) were identified by a polyclonal antibody recognizing both forms of eIF-2␣ (Fig. 3A). eIF-2B protein detection was carried out by a monoclonal antibody recognizing eIF-2B⑀, and the amount of eIF-2B⑀ was considered as being representative of the total amount of eIF-2B (Fig. 3B). Quantification of the bands in three independent experiments showed no differences between untreated and IGF-1-treated cells in the phosphorylated form of eIF-2␣ (12.4 Ϯ 1.3% versus 11.5 Ϯ 0.8% of total eIF-2␣) or eIF-2B levels (62.4 Ϯ 3.8 versus 61 Ϯ 2.1, in arbitrary units), suggesting that increased eIF-2B activity in the latter was not due to either the change in eIF-2␣ phosphorylation status or eIF-2B levels. GSK-3 Activity in IGF-1-treated Cells-To determine GSK-3 involvement in IGF-1-induced eIF-2B activation, kinase activity was measured in untreated and IGF-1-treated neuronal cells. As shown in Fig. 4A, IGF-1 treatment induced a 31% (from 34 to 22 arbitrary units) decrease in the phosphorylation of the specific peptide used as a substrate of the reaction, which correlates with the observed activation of both eIF-2B activity and protein synthesis rate (Figs. 1 and 2). IGF-1-induced GSK-3 inactivation was blocked by wortmannin and LY294002 but not by PD98059 (Fig. 4B), indicating the known upstream regulation of GSK-3 by PI3-K. PI3-K inhibition by wortmannin and LY294002 promoted GSK-3 activation and eIF-2B and protein synthesis inhibitions. PD98059 did not exert any effects on GSK-3 inactivation by IGF-1, suggesting the independent regulation of GSK-3 by MEK/MAP kinases. eIF-2B Phosphorylation by GSK-3-To further assess the participation of GSK-3 in IGF-1 signal cascade, both phosphorylation and activity of exogenous added eIF-2B were simultaneously determined in GKS-3 immunoprecipitates from IGF-1treated and untreated neurons. eIF-2B⑀ phosphorylation by GSK-3 from IGF-1-treated cells was 32% lower (from 17 to 12 arbitrary units) than eIF-2B⑀ phosphorylation produced by GSK-3 in untreated cells, suggesting higher activity of the enzyme in the latter (Fig. 5A). As a consequence of lower eIF-2B phosphorylation levels in IGF-1-treated immunoprecipitates, eIF-2B activity was higher in these samples (Fig. 5B). Both findings confirm that IGF-1-induced eIF-2B activation is promoted via a decrease in GSK-3-dependent eIF-2B⑀ phosphorylation.
IGF-1 Induces GSK-3 and MAPK Phosphorylation-By using an antibody recognizing the Ser 9 residue in the phosphorylated form of GSK-3␤ (inactive form), and unable to detect nonphosphorylated GSK-3␤ (active form) we found increased inactive GSK-3 form upon stimulation of neuronal cells with IGF-1 (Fig. 6, A and B, top panels). This phosphorylation was reversed by 0.1 M wortmannin and completely blocked by 1 M of the inhibitor (Fig. 6A). The MAP kinases ERK1 and 2 were activated by dual phosphorylation on threonine and tyrosine residues (e.g. Thr 183 and Tyr 185 in ERK2). Interestingly, by using an antibody against these regulatory sites of active ERK that does not recognize either the nonphosphorylated or the monophosphorylated form of the enzyme, we also detected ERK2 activity activation induced by IGF-1-treatment on neuronal cells (Fig. 6B, middle panel). As expected, PD98059 treatment did not affect GSK-3 phosphorylated status and did reverse ERK2 activation (Fig. 6B, top and middle lanes). To verify that the different lanes contained comparable amounts of sample, all immunoblots were reprobed with the correspond-ing antibodies recognizing both the phosphorylated and nonphosphorylated form of the kinases (Fig. 6, A and B, bottom  panels). These control antibodies were raised in a different host animal to avoid signal cross-reactivity. DISCUSSION The findings reported here support the possibility that in cortical neurons two signaling components are simultaneously activated by IGF-1. Interestingly, both signal pathways, PI3-K and MEK, are implicated in protein synthesis and eIF-2B activation. The straight correlation existing between IGF-1 action and eIF-2B activation suggests eIF-2B participation in IGF-1-mediated survival described in this type of cells (29,31,32).
Insulin stimulates protein synthesis in mammalian cells, and at least two signaling pathways, the PI3-K and Ras-MEK/ MAPK cascades, may be involved in the hormone action. Several initiation or elongation factors involved in insulin action, eIF-4E binding protein dephosphorylation, S6 ribosomal protein phosphorylation, eIF-2B activation, and eEF-2 elongation factor inactivation have been reported (23). On the contrary, in neuronal cells insulin fails to activate MAPK and, at superphysiological concentrations, promotes survival via PI3-K signaling (29). It has been postulated that in this system the effects of insulin are consequence of insulin-IGF-1 receptor cross-reactivity (39). In agreement with these results, we found FIG. 3. eIF-2␣ phosphorylation and eIF-2B levels are not affected by IGF-1 treatment. Cultured neurons were untreated (lanes 1) or treated with IGF-1 (100 ng/ml) (lanes 2) for 30 min. A, cells were lysed in 8.5 M urea, and fractions (60 g of protein) were resolved in horizontal isoelectric focusing slab gels. The bands corresponding to eIF-2␣ and phosphorylated eIF-2␣ (eIF-2␣P) proteins were detected on immunoblots developed by a polyclonal antibody purified by immunoaffinity. B, cells were processed, and fractions (40 g of protein) were analyzed by SDS-PAGE and Western blots. eIF-2B protein was stained with a monoclonal anti-eIF-2B⑀ antibody.
FIG. 4. Effect of wortmannin, PD98059, and LY294002 on IGF-1-induced GSK-3 inactivation. GSK-3 activity was determined in extracts from untreated (control) or IGF-1 (100 ng/ml) treated cells for 30 min, using a peptide substrate as described under "Experimental Procedures." A, representative experiment of the kinase assay subject to SDS-PAGE and autoradiography. The numbers express the quantification of the phosphorylated peptide band in arbitrary units and represent the average of three independent experiments, performed in duplicate. B, untreated cells were incubated with IGF-1 in the absence or presence of wortmannin, PD98059, and LY294002 for 30 min. Control value corresponds to 622 cpm/g protein and was considered as being 100%. The data were obtained from three independent experiments; error bars indicate S.E. *, p Ͻ 0.05 versus control; ***, p Ͻ 0.001 versus control. higher activation of protein synthesis and eIF-2B in IGF-1treated than that of insulin-treated neurons.
IGF-1 effect on eIF-2B activity cannot be explain in terms of an increased factor level or decreased eIF-2␣P levels. Instead, the straight correlation observed between PI3-K inhibition, GSK-activation, and eIF-2B inhibition suggests that eIF-2B activity could be regulated through phosphorylation. This mechanism has been proposed to explain eIF-2B activation in response to nerve growth factor and epidermal growth factor in PC12 cells (21). Although eIF-2B activation has been reported in vivo in different types of cells following stimulation with several treatments including growth factors, eIF-2B phosphorylation status in vivo has not been demonstrated as yet (13).
IGF-1 has become significant because it is known to promote important cellular responses that vary with the different types of cell. For instance, IGF-1 promotes hypertrophy by growth and differentiation in many types of cells (30,40), mediates anabolic and cardiovascular actions of growth hormone in vivo (30), and induces cerebellar neuron survival (29,31,32). It has very recently established the important role of IGF-1 in the treatment of neurodegenerative disorders (33,41). Our results delineate IGF-1-induced activation signaling pathway for protein synthesis in neuronal cells, which implies PI3-K activa-tion, as shown with the inhibitors wortmannin and LY29002, phosphorylation, and consequently inactivation of GSK-3 and eIF-2B activation. Although GSK-3 is a substrate in vitro for three insulin-stimulated protein kinases, (p70 S6 kinase, p90 S6 kinase (p90 RSK ), and PKB), current evidence points out that PKB is the enzyme responsible for GSK-3 inactivation in vivo by insulin (26,27). PKB is a protein kinase downstream of PI3-K and has been established to mediate IGF-1 effects on neuronal survival (29). Because PKB is sensitive to wortmannin and independent of MAPK pathway (26,27), PKB activation may be the link between PI3-K activation and GSK-3 inactivation in IGF-1-induced protein synthesis and eIF-2B activation in cortical neurons. A similar pathway has been proposed to explain nerve growth factor and epidermal growth factor-induced protein synthesis and eIF-2B activation in PC12 cells (21).
Insulin and IGF-1 have failed to activate the Ras-MEK/ MAPK pathway in certain types of cells, including cortical and cerebellar neurons (29,(31)(32)(33). Our findings clearly show the existence of an IGF-1-induced increase in MAPK (ERK2) phosphorylation (active form of the enzyme), specifically inhibited by PD98059, which demonstrates the activation of this cascade. Our finding that IGF-1 activates MAPK and increases eIF-2B activity, in a MEK pathway-dependent fashion, raised the possibility that MAPK regulates eIF-2B factor. The fact that the inhibition of PI3-K by wortmannin promotes GSK-3 activation and eIF-2B and protein synthesis inhibitions, whereas PD98059 does not affect GSK-3 activity, suggests an MEKindependent regulation of GSK-3. We can conclude that IGF-1 induces protein synthesis activation and eIF-2B activity in a PI3-K-and MEK kinase-dependent fashion.
Here we report findings supporting the possibility that IGF-1 promotes activation of two different signaling pathways in cortical neurons; interestingly both of them promote protein synthesis and eIF-2B activation. The observed parallelism between eIF-2B activation and protein synthesis suggests eIF-2B involvement in protein synthesis regulation following IGF-1 addition. IGF-1-induced eIF-2B activation occurs in a PI3-Kdependent fashion by inhibiting GSK-3; whereas MEK-induced activation of eIF-2B might be mediated by casein kinase-induced phosphorylation of the factor or other unknown eIF-2Bspecific protein kinase. The two pathways seem to be independent each other, because MEK inhibitor PD98059 did not affect GSK-3 activity (see above), nor did PI3-K inhibitor LY294002 affect MAPK (ERK1 and 2) phosphorylation (42), 2 but both mechanisms are necessary to maintain eIF-2B activation. Further studies are necessary to disclose the steps implicated in MEK-induced activation of eIF-2B and to fully establish the potential role of eIF-2B in neuron survival.