Glycogen Synthase Kinase 3β Is a Novel Regulator of High Glucose- and High Insulin-induced Extracellular Matrix Protein Synthesis in Renal Proximal Tubular Epithelial Cells*

High glucose (30 mm) and high insulin (1 nm), pathogenic factors of type 2 diabetes, increased mRNA expression and synthesis of lamininβ1 and fibronectin after 24 h of incubation in kidney proximal tubular epithelial (MCT) cells. We tested the hypothesis that inactivation of glycogen synthase kinase 3β (GSK3β) by high glucose and high insulin induces increase in synthesis of laminin β1 via activation of eIF2Bϵ. Both high glucose and high insulin induced Ser-9 phosphorylation and inactivation of GSK3β at 2 h that lasted for up to 48 h. This was associated with dephosphorylation of eIF2Bϵ and eEF2, and increase in phosphorylation of 4E-BP1 and eIF4E. Expression of the kinase-dead mutant of GSK3β or constitutively active kinase led to increased and diminished laminin β1 synthesis, respectively. Incubation with selective kinase inhibitors showed that high glucose- and high insulin-induced laminin β1 synthesis and phosphorylation of GSK3β were dependent on PI 3-kinase, Erk, and mTOR. High glucose and high insulin augmented activation of Akt, Erk, and p70S6 kinase. Dominant negative Akt, but not dominant negative p70S6 kinase, inhibited GSK3β phosphorylation induced by high glucose and high insulin, suggesting Akt but not p70S6 kinase was upstream of GSK3β. Status of GSK3β was examined in vivo in renal cortex of db/db mice with type 2 diabetes at 2 weeks and 2 months of diabetes. Diabetic mice showed increased phosphorylation of renal cortical GSK3β and decreased phosphorylation of eIF2Bϵ, which correlated with renal hypertrophy at 2 weeks, and increased laminin β1 and fibronectin protein content at 2 months. GSK3β and eIF2Bϵ play a role in augmented protein synthesis associated with high glucose- and high insulin-stimulated hypertrophy and matrix accumulation in renal disease in type 2 diabetes.

Glycogen synthase kinase 3␤ (GSK3␤) 3 was originally identified as an enzyme required for the regulation of glycogen metabolism (1). However, it has been found to be involved in a variety of cellular responses including cytoskeletal regulation (2), cell cycle progression (3,4), apoptosis (5,6), and cell adhesion (7). GSK3␤ regulation of protein synthesis has not been completely understood. GSK3␤ acts as a switch that regulates both transcription and mRNA translation by controlling the activity of transcription factors and eukaryotic translation initiation factor 2B⑀ (eIF2B⑀), respectively. In the resting cell, GSK3␤ is unphosphorylated and active, inhibiting the activity of its substrates, e.g. eIF2B⑀ and glycogen synthase. Upon stimulation, GSK3␤ is phosphorylated on Ser-9 and inactivated, leading to release of inhibition on activity of its substrates. Several kinases are implicated in Ser-9 phosphorylation of GSK3␤ including p70S6 kinase (8), p90RSK (9), PKC, PKA (10 -12), and Akt (13,14).
Augmented protein synthesis contributes to two cardinal manifestations of diabetic kidney disease, i.e. renal hypertrophy and accumulation of extracellular matrix proteins (15). However, the role of GSK3␤ in these events has not been investigated. Signaling mechanisms favoring renal hypertrophy have received much attention (16 -25). In contrast, constitutive signaling mechanisms that counteract the prohypertrophic signaling mechanisms and check unrestricted progression of protein synthesis are not well understood. GSK3␤ appears to constitutively inhibit protein synthesis by its inhibition of eIF2B⑀. We hypothesized that inactivation of GSK3␤ by high glucose and high insulin induces an increase in synthesis of laminin ␤1 via activation of eIF2B⑀. We investigated if GSK3␤ is inactive in the kidney in type 2 diabetes. We employed both in vitro cell culture and in vivo animal models of type 2 diabetes in our investigation.
Immunoblotting-Equal amounts of protein from cells were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and probed with primary antibody for 3 h (19,27). Primary antibodies were from Cell Signaling (Beverly, MA) if not otherwise mentioned. Laminin ␤1 and GSK3␤ antibody were from Santa Cruz Biotechnology, fibronectin antibody was from Sigma, and phospho-eIF2B⑀ (Ser-539) was purchased from Upstate, Lake Placid, NY. After washing, the membrane was incubated with peroxidase-conjugated secondary antibody (Jackson ImmunoResearch Laboratories Inc, West Grove, PA). Proteins were visualized by chemiluminescence using the ECL reagent (Pierce Biotechnology, Rockford, IL). Images of the bands were scanned by reflectance scanning densitometry, and the intensity of the bands was quantified using Scion Image software (27). For immunoblot analysis of 4E-BP1 phosphorylation, cell lysates were heated at 100°C for 7 min, and treated with 15% trichloroacetic acid. Immunoblotting was performed using an antibody specific for 4E-BP1 phosphorylated at Thr-37/46 (27).
GSK3␤ Immunokinase Activity Assay-Following treatment with high glucose or high insulin, cell lysates were prepared using the lysis buffer consisting of 50 mM Tris, 150 mM NaCl, 10% (v/v) glycerol, 0.5% (v/v) Nonidet P-40, 1 mM EDTA, 1 mM dithiothreitol, 500 M Na 3 VO 4 , 500 M NaF, 100 M ␤-glycerophosphate, 100 M sodium pyrophosphate, 10 g/ml leupeptin, and 1% (v/v) aprotinin. 300 g of total protein was incubated with 1 g of monoclonal anti GSK3␤ (sc-81462) in a rotor for 12 h at 4°C. The Equal amounts of protein from cell lysates extracted after treatment with or without high glucose and high insulin for the time shown were separated on SDS-PAGE. Following transfer to nitrocellulose membrane, the membranes were probed with specific antibodies for expression of matrix proteins laminin ␤1 (A and B) and fibronectin (C and D). Lower panels show immunoblotting for actin done to assess loading. Representative blots from four to five experiments are shown, and composite data are given in histograms. #, p Ͻ 0.05 and §, p Ͻ 0.01 versus control by ANOVA. FIGURE 2. High glucose and high insulin increase laminin ␤1 and fibronectin mRNA levels. Total RNA isolated from cells treated with or without high glucose and high insulin was reverse-transcribed to synthesize cDNA. 2 l of the cDNA subjected to quantitative real-time PCR using the SYBR Green method shows an increase in the transcript level of laminin ␤1 (A and B) and fibronectin (C and D) following incubation with high glucose or high insulin. GAPDH served as an internal control. The ratio of laminin ␤1 and fibronectin mRNA to GAPDH mRNA is shown. Histograms represent means Ϯ S.E. from three independent experiments. Significant differences between control and cells treated with high glucose and high insulin are indicated by #, p Ͻ 0.05; §, p Ͻ 0.01; and *, p Ͻ 0.001 by ANOVA. immune complexes were isolated by the addition of 25 l of a slurry of protein A/G-agarose (sc-2003) and incubation for 2 h at 4°C. Immunoprecipitates were washed twice with lysis buffer and twice with kinase reaction buffer (8 mM MOPS, pH 7.4, 0.2 mM EDTA, 10 mM magnesium acetate, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, 2.5 mM ␤-glycerophosphate, 10 g/ml leupeptin, and 1% (v/v) aprotinin). The kinase activity assays were performed in 40 l of total reaction buffer containing 62.5 M GSK3␤ substrate (BioMol, Plymouth, PA), 20 mM MgCl 2 , 125 M ATP, and 10 Ci [␥-32 P]ATP. The reaction mixture was allowed to proceed for 30 min at 30°C, and 15 l of the super-natant were spotted onto Whatman P81 phosphocellulose paper. The filter squares were washed five times for 5 min each in 0.75% phosphoric acid. The filters were then briefly rinsed in acetone, dried at room temperature, and subjected to liquid scintillation counting (28).
Real Time RT-PCR-MCT cells were lysed, and their total RNA was isolated by acid phenol extraction using TRIzol (Invitrogen, Carlsbad, CA). Equal amounts of total RNA (1 g) from MCT cells were converted to first-strand cDNA by using reverse transcriptase. PCR primer sequences for amplification of mouse laminin ␤1 (Primer Bank ID 21595540a2), mouse fibronectin (Primer Bank ID 4218966a1) and mouse GAPDH (Primer Bank ID 6679937a1) were obtained from Primer Bank, a public resource for PCR primers (29). 2 l of cDNA was amplified using SYBR Green PCR Master mix (Applied Biosystems, Foster City, CA) containing 100 nM forward and reverse primers. PCR amplification was performed using 7900HT Sequence Detection System (Applied Biosystems) using the manufacturer's protocol as recently described (21,27). Dissociation curve analysis was performed following PCR amplification to confirm the specificity of the primers. Relative mRNA expression was calculated using the ⌬⌬C t method.
Animal Study-The C57BL6/KsJ leprϪ/Ϫ db/db mice, a model of type 2 diabetes, and its lean littermates (db/m) (Jackson Laboratory, Bar Harbor, ME), were maintained on regular laboratory chow. Blood glucose concentration was monitored FIGURE 3. High glucose and high insulin stimulate phosphorylation and dephosphorylation of GSK3␤ and eIF2B⑀, respectively. MCT cells were treated with or without high glucose or high insulin. Equal amounts of protein from cell lysates were separated by SDS-PAGE and immunoblotted with phosphospecific and total protein antibodies for GSK3␤ (A and B) and eIF2B⑀ (E and F). Representative blots from four experiments are shown. Histograms show composite densitometric data from four experiments. Statistical significance is shown as #, p Ͻ 0.05; §, p Ͻ 0.01; *, p Ͻ 0.001 by ANOVA. Immune complex assay for GSK3␤ activity was performed using cell lysates from control and high glucose-and high insulin-treated cells as described under "Materials and Methods" (C and D). Composite data from three experiments are shown in a histogram (#, p Ͻ 0.05; §, p Ͻ 0.01; and *, p Ͻ 0.001 by ANOVA).
for emergence of diabetes. In the present study, lean littermate control and diabetic mice were studied in the early phase, after 2 weeks of onset of hyperglycemia, and, in the late phase, at 2 months of hyperglycemia. Previously we have reported that db/db mice display renal hypertrophy at 2 weeks (23) and increase in laminin content at 2 months of diabetes (33). Mice were sacrificed at the end of each experimental period, and renal cortex was dissected out and processed for further analysis.
Statistical Analysis-All values are expressed as mean Ϯ S.E. obtained from at least three independent experiments. Statistical analysis was performed using one-way analysis of variance (ANOVA) for comparison between multiple groups and posthoc analysis using Student's Newman Keul's multiple comparison tests using Graph Pad Prism 4 software. Statistical comparisons between two groups were performed by the Student's t test. Statistical significance was assigned to values of p Ͻ 0.05.

High Glucose and High Insulin Stimulate Laminin and
Fibronectin Synthesis-Incubation of MCT cells with 30 mM glucose or 1 nM insulin stimulated synthesis of extracellular matrix proteins, laminin ␤1, and fibronectin, starting at 6 h and reaching a peak at 24 h ( Fig. 1, A-D). We investigated if induction of synthesis of laminin ␤1 and fibronectin was due to an increase in their mRNA levels. Quantitative real time RT-PCR showed an increase in their respective mRNA content upon treatment with high glucose and high insulin (Fig. 2, A-D). Preincubation with actinomycin D, a transcription inhibitor, inhibited laminin ␤1 synthesis induced by high glucose and high insulin at 24 h but not at 1 and 2 h, suggesting transcriptional regulation at later time points and, possibly, translational regulation at early time points. Actinomycin-D inhibited high glucose-and high insulininduced fibronectin synthesis at 2 and 24 h but not at 1 h, suggesting transcriptional regulation beyond 1 h of incubation (supplemental Fig. S1). We have previously reported that high glucose, high insulin, high glucose ϩ high insulin increased de novo protein synthesis by about 20 -25% (27). Immunoprecipitation of [ 35 S]me-thionine-labeled cell protein with laminin ␤1 antibody showed significant increment in laminin synthesis under the three conditions. These data showed that although high glucose, high insulin, and both together stimulated global protein synthesis in MCT cells, increments in laminin ␤1 synthesis were in addition individually stimulated (27).    A and B). Cells were transfected with empty vector or DN-p70S6K (C and D) prior to treatment with or without high glucose or high insulin. Equal amounts of lysates (30 g) from the MCT cells were separated on SDS-PAGE. Following transfer to nitrocellulose membrane, the membranes were probed the antibody against laminin ␤1 and Ser-9-phosphorylated GSK3␤. The membrane was stripped and reprobed with an antibody against actin to assess loading. Immunoblotting with anti-HA (hemagglutinin) antibody was done to show expression of transfectants. Representative blots from three independent experiments are shown. NOVEMBER 7, 2008 • VOLUME 283 • NUMBER 45

GSK3␤ in Renal Matrix Protein Synthesis
High Glucose and High Insulin Promote Phosphorylation and Inactivation of GSK3␤ and Dephosphorylation of eIF2B⑀-In cells in the basal state, GSK3␤ is active and keeps eIF2B⑀ inactive. Both high glucose and high insulin induced phosphorylation of Ser-9 on GSK3␤, which is required for its inactivation, starting at about 2 h and lasting up to 48 h (Fig. 3, A and B). Combined incubation with high glucose and high insulin also increased GSK3␤ phosphorylation following the same time course as that of high glucose alone or high insulin alone (supplemental Fig. S2). We directly studied the effect of high glucose and high insulin on GSK3␤ activity. In an immunokinase activity assay, treatment of MCT cells with high glucose and high insulin significantly reduced [ 32 P]phosphate incorporation into the GSK3␤ substrate, demonstrating reduced activity of GSK3␤ (Fig. 3, C and D). These data show that the increase in Ser-9 phosphorylation of GSK3␤ correlates with reduction in its enzymatic activity. eIF2B⑀ mediates the antihypertrophic effects of GSK3␤. Upon inactivation of GSK3␤ eIF2B⑀ is dephosphorylated on Ser-539 and is free to promote translation initiation. The binding of eIF2 to the activated initiator methionyl tRNA is promoted by dephosphorylated eIF2B⑀ (34,35). High glucose and high insulin reduced eIF2B⑀ Ser-539 phosphorylation starting at 2 h and lasting for up to 48 h (Fig. 3, E and F). Thus, GSK3␤ inactivation and eIF2B⑀ dephosphorylation are temporally associated with high glucose-and high insulin-induced matrix protein synthesis.
High Glucose and High Insulin Regulate Initiation and Elongation Phases of Translation-Because translation is the rate-limiting step in gene expression culminating in peptide synthesis, we examined key steps in initiation and elongation phases of translation in the context of GSK3␤ regulation of matrix protein synthesis. During the initiation phase of mRNA translation, binding of the 43S ribosomal pre-initiation complex to the mRNA is facilitated by eIF4E in association with eIF4G and eIF4A (15). In the resting cell, eIF4E is held in an inactive complex by its binding protein 4E-BP1. Upon stimulation, 4E-BP1 is phosphorylated on several threonine and serine residues, including Thr-37/46 (36). This results in dissolution of the complex, freeing eIF4E to bind with mRNA cap to promote translation initiation. High glucose and high insulin augmented phosphorylation of 4E-BP1 (Fig. 4, A and B) and eIF4E (Fig. 4, C and D).
During the elongation phase of translation, amino acids are sequentially added to the nascent peptide chain by the participation of aminoacyl tRNA (34,37). Movement of the ribosome along the mRNA, resulting in translocation of aminoacyl tRNA from the A site to the P site on the ribosome, is regulated by eukaryotic elongation factor 2 (eEF2). Activation of eEF2 occurs following dephosphorylation at Thr-56 (38). High glucose and high insulin also induced dephosphorylation of Thr-56 on eEF2 (Fig. 4, E and F). Thus, high glucose or high insulin activate important steps in both initiation and elongation phases in association with inactivation of GSK3␤ and increased matrix protein synthesis in MCT cells.
Akt but Not p70S6 Kinase Regulates GSK3␤ Phosphorylation Induced by High Glucose and High Insulin-Next, we examined whether phosphorylation of Akt and p70S6 kinase is required for high glucose-and high insulin-induced phosphorylation of GSK3␤ and laminin ␤1 synthesis. In cells expressing dominantnegative Akt, high glucose and high insulin failed to stimulate GSK3␤ phosphorylation and laminin ␤1 synthesis (Fig. 6, A and B). However, in cells expressing kinasedead p70S6 kinase, only laminin ␤1 synthesis was blocked but not Ser-9 phosphorylation of GSK3␤ (Fig. 6, C  and D). Thus, stimulation of GSK3␤ phosphorylation by high glucose and high insulin is dependent on Akt but not p70S6 kinase activation in MCT cells, although both kinases are involved in the regulation of laminin ␤1 synthesis.
We examined the role of PI 3-kinase, Erk, and mTOR in GSK3␤ Ser-9 phosphorylation. Akt activation by high glucose and high insulin is PI 3-kinase-dependent (data not shown) (43) and Akt activation results in GSK3␤ inhibition by phosphorylation of Ser-9 (44). Preincubation of MCT cells with LY294002, a selective inhibitor of PI 3-kinase, U0126, a MEK inhibitor, and, rapamycin, a specific inhibitor of mTOR, blocked GSK3␤ phosphorylation (Fig. 7, A-F) and laminin ␤1 synthesis (Fig. 7, G and H) induced by high glucose and high insulin at 24 h. As rapamycin data suggested, mTOR regulates GSK3␤, we studied their interaction. Immunoprecipitation with anti-mTOR antibody and immunoblotting with anti-GSK3␤ antibody suggested that the two proteins exist in a complex; the intensity of binding did not seem to change following exposure to high glucose or high insulin (supplemental Fig. S3). Because PI 3-kinase-Akt-mTOR belong to a linear pathway, it is not surprising that inhibitors of PI 3-kinase (LY294002) and mTOR (rapamycin) inhibited GSK3␤ phosphorylation. We explored if Erk activation was under the regulation of PI 3-kinase. LY294002, abrogated high glucose-and high insulin-induced phosphorylation of Erk1/2 MAP kinase (supplemental Fig. S4), suggesting the latter was under the control of PI 3-kinase. These data confirm our previous report that Erk is downstream of PI 3-kinase in insulin-treated MCT cells (45). Because PI 3-kinase regulates both mTOR and Erk activation, we see complete inhibition of Ser-9 phosphorylation of GSK3␤ with each of the three inhibitors.

Expression of Kinase Inactive (K85A) and Constitutively Active (S9A) Constructs of GSK3␤ Alter High Glucose-and High
Insulin-induced Changes in Laminin ␤1 Synthesis-Next, we investigated the requirement of GSK3␤ inactivation for high glucose-and high insulin-induced synthesis of laminin ␤1 in  NOVEMBER 7, 2008 • VOLUME 283 • NUMBER 45

GSK3␤ in Renal Matrix Protein Synthesis
MCT cells. Cells were transiently transfected with either a kinase inactive (K85A) or a constitutively active (S9A) construct of GSK3␤. Expression of kinase-dead K85A-GSK3␤ resulted in increased laminin ␤1 synthesis in control cells (Fig.  8, A and B). However, inactive GSK3␤ did not add to the incre-mental effect of high glucose and high insulin on laminin ␤1 synthesis. Expression of constitutively active GSK3␤ (S9A), which cannot be phosphorylated and therefore cannot be inhibited by stimuli was found to be effective in abolishing both high glucose-and high insulininduced laminin ␤1 synthesis in MCT cells (Fig. 8, C and D); however, S9A mutant did not affect constitutive laminin ␤1 synthesis. Expression of mutant constructs was confirmed by immunoblotting the cell lysates using an anti-HA antibody. Because eIF2B⑀ is a direct substrate of GSK3␤, we examined its phosphorylation status in cells expressing the mutants. Overexpression of kinase-inactive K85A GSK3␤ construct caused dephosphorylation of eIF2B⑀ in control cells (supplemental Fig. S5). Conversely overexpression of constitutively active S9A construct of GSK3␤ caused phosphorylation of eIF2B⑀ in control cells (supplemental Fig. S5). We described the data above (Fig. 6, C and D) that showed p70S6 kinase was not an upstream regulator of GSK3␤ phosphorylation. We further explored whether GSK3␤ could regulate p70S6 kinase phosphorylation when exposed to high glucose and high insulin in cells expressing the K85A and S9A mutants. With either construct, high glucose and high insulin were able to stimulate Thr-389 phosphorylation of p70S6 kinase, similar to cells expressing control plasmid vector (supplemental Fig. S6). These data indicate that Thr-389 phosphorylation of p70S6 kinase is not under control of GSK3␤.
Diabetes-induced Changes in GSK3␤ in Type 2 Diabetes-Mice with spontaneous onset of type 2 diabetes were studied at 2 weeks (Early diabetes) and 2 months (Late diabetes) of hyperglycemia when they manifest renal hypertrophy (23), and, increase in laminin content, respectively (33). The renal cortex from diabetic mice showed an increase in phosphorylation of GSK3␤ at Ser-9 (Fig.  9, A and B) and dephosphorylation of Ser-539 on eIF2B⑀ (Fig. 9, C and D) both at early and late phases of diabetes compared with renal cortex from control db/m mice; contents of GSK3␤ FIGURE 9. Phosphorylation of GSK3␤ and p70S6K is increased and eIF2B⑀ phosphorylation is decreased in the renal cortex of mice with type 2 diabetes. C57BL6/KsJ db/db mice were sacrificed at 2 weeks (early db/db) and 2 months (late db/db) of diabetes. Equal amounts of renal cortical homogenates from control and diabetic mice were separated by SDS-PAGE and immunoblotted with antibodies against phospho-GSK3␤ (A and B), phospho-eIF2B⑀ (C and D) and phospho-p70S6 kinase (E and F). Immunoblotting with antibodies against GSK3␤, eIF2B⑀, p70S6 kinase, and tubulin was done to show equal loading. Composite mean Ϯ S.E. data from four mice in each group are shown in a histogram. Statistical significance between the two groups is shown as #, p Ͻ 0.05; §, p Ͻ 0.01; *, p Ͻ 0.001, by Student's t test, db/db versus control. and eIF2B⑀ were unaffected. We also observed increased Thr-389 phosphorylation of p70S6 kinase in the renal cortex of mice with diabetes when compared with their lean littermates at both time points (Fig. 9, E and F) without significant changes in the content of the kinase. Inactivation of GSK3␤ by phosphorylation and activation of eIF2B⑀ by dephosphorylation correlated with the increment of laminin ␤1 chain in diabetic renal cortex (33).

DISCUSSION
Our study showed that high glucose and high insulin increased mRNA and protein expression of laminin ␤1 and fibronectin in the renal proximal tubular epithelial cells and that it required inactivation of GSK3␤. We have earlier reported that high glucose and high insulin selectively induce synthesis of laminin ␤1 chain matrix protein (but not type IV collagen or fibronectin) within 5 min without any change in its mRNA level. The regulatory mechanism involves mRNA translation and not transcription (27). Our present data indicate that full expression of the laminin and fibronectin under high glucose and high insulin stimulation requires enhanced mRNA translation as well as transcription.
In the quiescent cell, GSK3␤ is unphosphorylated and active thereby inhibiting its molecular targets that promote cell growth. After stimulation with growth factors or induction of canonical Wnt pathway, GSK3␤ is inactivated by Ser-9 phosphorylation, releasing its pro-growth substrates from tonic inhibition (46). Growth factors regulate GSK3␤ via kinases that directly phosphorylate the kinase to control its activity. Akt, Erk, and possibly p70S6 kinase have been suggested as upstream controllers of GSK3␤ (8,42). We observed that high glucose-and high insulin-induced GSK3␤ inactivation by phosphorylation at Ser-9 was dependent on activation of Akt. Abrogation of high glucose and high insulin-induced changes in GSK3␤ phosphorylation by LY294002, suggested that PI 3-kinase was also involved in our system. As mTOR inhibition with rapamycin abolished high glucose-and high insulin-induced Ser-9 phosphorylation of GSK3␤, we investigated if it involved p70S6 kinase. Expression of dominant negative p70S6 kinase did not affect high glucose and high insulin regulation of Ser-9 phosphorylation of GSK3␤. However, mTOR and GSK3␤ appear to be constituents of a complex suggesting the possibility that some other substrate of mTOR or mTOR itself may play a part. Erk inhibitor, U0126, also inhibited high glucose and high insulin stimulation of GSK3␤ phosphorylation suggesting it was another upstream regulator of GSK3␤. Abolition of high glucose-and high insulin-induced Erk phosphorylation by LY294002 suggests that PI 3-kinase is upstream of Erk, confirming previous report in insulin-treated MCT cells (45). Thus, PI 3-kinase appears to regulate both Akt-mTOR and Erk pathways in MCT cells treated with high glucose and high insulin (Fig. 10). In MCT cells employed in this study, high insulin stimulation of Erk appears to be dependent on PI 3-kinase (45). The precise interactions among these kinase pathways in the context of high glucose and high insulin regulation of GSK3␤ activity and protein synthesis need to be investigated further.
Cell growth regulated by GSK3␤ could involve either cell proliferation or cell hypertrophy. Hypertrophy requires activa-tion of signaling pathways that transduce extracellular signals to their intracellular targets causing modification of the translational apparatus in response to these signals (47). Molecular mechanisms that regulate GSK3␤ in the context of protein synthesis are not well understood. Because mRNA translation is the rate-limiting step in gene expression culminating in protein synthesis (34), we examined the role of GSK3␤ in the initiation and elongation phases of translation. Upon stimulation by high glucose and high insulin important events in both phases were triggered as observed by the activation of mTOR-4EBP1/Erk-eIF4E pathway and mTOR-p70S6K-eEF2 pathway (Figs. 4 and 5). One of the critical steps in the initiation of mRNA translation is the binding of eIF2 to the activated initiator methionyl tRNA (met-tRNA) and subsequent formation of a ternary complex that binds to the 40S ribosomal subunit. This process requires activities of eIF2B⑀, a guanine nucleotide exchange factor, in order to stimulate the GDP/GTP exchange reaction of eIF2. GSK3␤ phosphorylates eIF2B⑀ at Ser-539 (Ser-535 in the rat sequence) and inactivates it (48). Because stimulation of matrix protein synthesis by high glucose and high insulin was associated with inactivating Ser-9 phosphorylation of GSK3␤ and reduction in Ser-539 phosphorylation of eIF2B⑀, the latter appears to be the mediator of protein synthesis regulation by GSK3␤.
There are other reported instances of GSK3␤ regulating cell hypertrophy. A recent report suggests that inhibition of the GSK3␤/eIF2B⑀ translational control pathway contributes to airway smooth muscle hypertrophy in vitro and in vivo (49). GSK3␤ has been shown as an important negative regulator of cardiac hypertrophy (40,50,51). Overexpression of a constitutively active mutant GSK3␤ S9A in neonatal rat cardiac myocytes markedly reduced hypertrophy induced by G␣ q -coupled receptor stimulation (40  eIF2B⑀ mutated at its GSK3␤ phosphorylation site preventing its inactivation, spontaneous hypertrophy developed, and the anti-hypertrophic effects of GSK3␤ overexpression were abolished. Various reports have shown that there is a marked increase in GSK3␤ activity in heart and skeletal muscle of type 2 diabetes (53)(54)(55). These data are consistent with elevated GSK3 activity in muscle in type 2 diabetes (56,57). These changes could be a cause or consequence of development of insulin resistance and type 2 diabetes. In contrast to the insulinresistant non-renal tissues such as the liver, kidney remains sensitive to insulin in db/db mice model of type 2 diabetes (58). However, the activity of GSK3␤ in the kidneys in type 2 diabetes has not been examined. We observed that in the db/db mice, renal cortex homogenates exhibit increased Ser-9 phosphorylation of GSK3␤ concomitant with decreased phosphorylation of eIF2B⑀ during early stage of type 2 diabetes. As GSK3␤ is active in the basal state of the tissue and normally functions as repressor of protein synthesis and transcription, the phosphorylation and inactivation of GSK3␤ will lead to increased protein synthesis. At this stage, both renal hypertrophy and increase in laminin ␤1 content are evident in the renal cortex of db/db mice (21). The increment in laminin ␤1 protein is associated with unchanged mRNA content suggesting its regulation by translation (ibid). Thus, GSK3␤ could have a role in both renal hypertrophy and accumulation of matrix protein laminin ␤1 in the diabetic kidney. One could anticipate that renal hypertrophy and laminin ␤1 accumulation induced by hyperglycemia and hyperinsulinemia would be rescued by activation of GSK3␤ or by overexpression of constitutively active form of GSK3␤ in the kidney in the db/db mouse.
Similar to GSK3␤, AMP-activated protein kinase (AMPK) is also a physiological inhibitor of protein synthesis; stimulation of AMPK activity ameliorated renal hypertrophy in rodents with type 1 diabetes (17). Overexpression of S9A mutant of GSK3␤, which cannot be inactivated, inhibited laminin ␤1 synthesis induced by high glucose and high insulin suggesting that this could be a useful strategy to reduce matrix accumulation in diabetic kidney disease. However, because GSK3␤ inactivation is required for glycogen synthesis in response to high glucose and insulin secretion in insulin responsive tissues, modulation of GSK3␤ activity will have to be selective in the kidney tissue in diabetes.