Activation of Glycogen Synthase Kinase 3β Ameliorates Diabetes-induced Kidney Injury*

Background: High glucose-induced matrix protein synthesis in renal cells requires glycogen synthase kinase 3β (GSK3β) inactivation. Results: Sodium nitroprusside (SNP) activated GSK3β and inhibited diabetes-induced kidney hypertrophy, matrix deposition, and albuminuria in mice without changing blood glucose. Conclusion: Activation of GSK3β by SNP ameliorates diabetic kidney injury. Significance: GSK3β may be a novel target for intervention in diabetic kidney disease. Increase in protein synthesis contributes to kidney hypertrophy and matrix protein accumulation in diabetes. We have previously shown that high glucose-induced matrix protein synthesis is associated with inactivation of glycogen synthase kinase 3β (GSK3β) in renal cells and in the kidneys of diabetic mice. We tested whether activation of GSK3β by sodium nitroprusside (SNP) mitigates kidney injury in diabetes. Studies in kidney-proximal tubular epithelial cells showed that SNP abrogated high glucose-induced laminin increment by stimulating GSK3β and inhibiting Akt, mTORC1, and events in mRNA translation regulated by mTORC1 and ERK. NONOate, an NO donor, also activated GSK3β, indicating that NO may mediate SNP stimulation of GSK3β. SNP administered for 3 weeks to mice with streptozotocin-induced type 1 diabetes ameliorated kidney hypertrophy, accumulation of matrix proteins, and albuminuria without changing blood glucose levels. Signaling studies showed that diabetes caused inactivation of GSK3β by activation of Src, Pyk2, Akt, and ERK; GSK3β inhibition activated mTORC1 and downstream events in mRNA translation in the kidney cortex. These reactions were abrogated by SNP. We conclude that activation of GSK3β by SNP ameliorates kidney injury induced by diabetes.


Increase in protein synthesis contributes to kidney hypertrophy and matrix protein accumulation in diabetes. We have previously shown that high glucose-induced matrix protein synthesis is associated with inactivation of glycogen synthase kinase 3␤ (GSK3␤) in renal cells and in the kidneys of diabetic mice. We tested whether activation of GSK3␤ by sodium nitroprusside (SNP) mitigates kidney injury in diabetes. Studies in kidneyproximal tubular epithelial cells showed that SNP abrogated high glucose-induced laminin increment by stimulating GSK3␤
and inhibiting Akt, mTORC1, and events in mRNA translation regulated by mTORC1 and ERK. NONOate, an NO donor, also activated GSK3␤, indicating that NO may mediate SNP stimulation of GSK3␤. SNP administered for 3 weeks to mice with streptozotocin-induced type 1 diabetes ameliorated kidney hypertrophy, accumulation of matrix proteins, and albuminuria without changing blood glucose levels. Signaling studies showed that diabetes caused inactivation of GSK3␤ by activation of Src, Pyk2, Akt, and ERK; GSK3␤ inhibition activated mTORC1 and downstream events in mRNA translation in the kidney cortex. These reactions were abrogated by SNP. We conclude that activation of GSK3␤ by SNP ameliorates kidney injury induced by diabetes.
Diabetes is the major cause of end stage renal disease (1,2). Diabetic kidney disease is characterized by excessive deposition of extracellular matrix in the form of thickening of glomerular and tubular basement membranes and increased amounts of mesangial matrix (glomerulosclerosis) and tubulo-interstitial fibrosis (3). We have previously reported that high glucose and high insulin, conditions associated with type 2 diabetes, increased protein synthesis, including matrix proteins in the renal proximal tubular epithelial (MCT) 3 cells (4). These changes were associated with inactivation of glycogen synthase kinase 3␤ (GSK3␤), a ubiquitously expressed and constitutively active serine/threonine kinase (5). GSK3␤ regulates a variety of cellular processes, including glycogen metabolism (6), gene transcription (7), apoptosis (8), and microtubule stability (9,10); our studies have shown that it serves as a constitutive inhibitor of protein synthesis in renal epithelial cells (5). Akt promotes protein synthesis by phosphorylating GSK3␤ at Ser-9, thereby inhibiting its activity (11,12). GSK3␤ activity can also be regulated by Tyr-216 phosphorylation (13,14). GSK3␤ regulates the activity of a broad range of substrates by phosphorylation (e.g. glucose metabolism by phosphorylation and inactivation of glycogen synthase (15) and protein synthesis by phosphorylation and inhibition of eukaryotic initiation factor 2B⑀ (eIF2B⑀)) (5,16). eIF2B is a heteropentamer; its catalytic ⑀ subunit promotes GDP/GTP exchange on eIF2, a key regulatory step in the initiation phase of mRNA translation (17). Augmented protein synthesis contributes to kidney hypertrophy and matrix protein increment seen in diabetic kidney disease (18). Although GSK3␤ is an inhibitor of high glucose-induced protein synthesis, whether its activation ameliorates diabetic kidney disease in vivo has not been studied. Sodium nitroprusside (SNP) is a GSK3␤ activator (19). In this study, we investigated whether activation of GSK3␤ by SNP ameliorates diabetes-induced kidney hypertrophy, albuminuria, and matrix protein accumulation.
Animal Study-Our Institutional Animal Care and Use Committee approved the studies. Diabetes was induced in mice as described by the Diabetes Complications Consortium. C57Bl6 mice received daily intraperitoneal injections of streptozotocin (50 mg/kg) for 5 days (20); control mice received the citrate buffer. Following the determination of the optimal dose of SNP, control and diabetic mice received 200 g/kg SNP or its vehicle intraperitoneally daily for 3 weeks.
Albuminuria Estimation-Mice were individually placed in metabolic cages prior to sacrifice, and urine was collected for 24 h. Commercial kits were used to measure albumin (Bethyl Laboratories, Montgomery, TX) and creatinine (Enzo Life Sciences Inc., Farmingdale, NY) (21).
Immunoblotting-Primary antibodies were from Cell Signaling (Beverly, MA) if not otherwise indicated. Laminin-␤1 and GSK3␤ antibodies were from Santa Cruz Biotechnology, Inc.; fibronectin antibody was from Sigma; and phospho-eIF2B⑀ (Ser-539) was from Upstate Biotechnology, Inc. (Lake Placid, NY). Immunoblotting and scanning of the bands were done as described previously (20,22).
Immunohistochemical Analysis-Cryosections of snap-frozen kidney tissues from control and diabetic mice treated with or without SNP were subjected to immunoperoxidase staining with a polyclonal antibody against laminin trimer from Thermo-NeoMarker (Waltham, MA) as described (23). The area of laminin staining within the glomerulus was measured in digital images by selecting a lower and upper range of gray scale within the limits of background and the highest intensity of laminin immunoperoxidase staining. The circumference of each glomerulus was outlined, and the specific staining area was identified by pseudocoloring and calculated as a percentage of total glomerular area as described earlier (24). The measurements were made using a computer-based morphometric analysis system with Image-Pro Plus software (Media Cybernetics, Inc., Silver Spring, MD).
Statistical Analysis-All values are expressed as mean Ϯ S.E. Statistical analysis was performed using one-way analysis of variance for comparison between multiple groups and post hoc analysis using Newman-Keuls multiple comparison tests using GraphPad Prism version 4 software. Statistical comparisons between two groups were performed by Student's t test. Statistical significance was assigned to values of p Ͻ 0.05.
SNP Inhibits HG-induced Matrix Protein Expression and Upstream Regulators of GSK3␤-HG rapidly increased the expression of matrix proteins laminin ␤1 and laminin ␥1 that was blocked by SNP (Fig. 2, A and B). Activated Akt in response to high glucose inhibits GSK3␤ by phosphorylating it at Ser-9 (5,27). In addition, we have identified ERK1/2 MAPK as an upstream regulator of GSK3␤ Ser-9 phosphorylation in MCT cells (6). SNP pretreatment blocked HG-induced phosphorylation and activation of Akt and ERK (Fig. 2, C and D), suggesting that SNP activation of GSK3␤ is mediated by inhibition of Akt and ERK in HG-treated cells.

SNP Blocks HG-induced Activation of Src and Pyk2, Kinases
That Regulate ERK and Akt Activity-To identify the link between hyperglycemia and GSK3␤, we explored Src and Pyk2, which have been reported as upstream regulators for Akt and ERK (22,28,29). HG increased Tyr-416 autophosphorylation of Src and Tyr-402 phosphorylation of Pyk2, a calcium-dependent proline-rich non-receptor tyrosine kinase (Fig. 3, A and B). SNP inhibited HG-induced phosphorylation of Src and Pyk2 (Fig. 3, A and B). Next we examined whether activation of Src and Pyk2 is required for HG-induced GSK3␤ inactivation and laminin ␥1 Preincubation with PP2 (a Src inhibitor) and BAPTA/AM (a Pyk2 inhibitor) blocked HG-induced GSK3␤ phosphorylation (P) at Ser-9 (C and D) and synthesis of laminin ␥1 (E and F). HG-induced activation of Akt and ERK was blocked by PP2 (G and I) and BAPTA/AM (H and J). Immunoblotting using antibody against the respective total proteins or actin was done to assess loading. Representative blots and histograms of composite data from 3-4 experiments are shown. Error bars, S.E. synthesis. Preincubation with either PP2, a Src inhibitor, or BAPTA/AM, a calcium chelator that inhibits calcium-dependent Pyk2, blocked HG-induced phosphorylation of GSK3␤ (Fig. 3, C and D, respectively) and laminin ␥1 synthesis (Fig. 3, E and F, respectively). Furthermore, activation of Src and Pyk2 was required for HG-induced activation of Akt and ERK (Fig. 3, G-J). These data show that HG activates Src and Pyk2 as upstream regulators of Akt and ERK, leading to inhibition of GSK3␤ culminating in laminin synthesis; SNP inhibits these HG-induced events.

SNP Abrogates Stimulatory Effects of HG on the Initiation and Elongation
Phases of mRNA Translation-Because mRNA translation is rate-limiting for peptide generation, we examined whether SNP affected HG stimulation of translation events. mTOR complex 1 (mTORC1) regulates the initiation and elongation phases of mRNA translation by inactivation of 4EBP1 and activation of p70S6 kinase. Eukaryotic initiation factor 4E (eIF4E) is held in an inactive complex by its binding protein 4EBP1; phosphorylation of the latter releases eIF4E, which undergoes phosphorylation on Ser-209 and associates with  DECEMBER 19, 2014 • VOLUME 289 • NUMBER 51 eIF4G to facilitate the initiation phase of translation. p70S6 kinase phosphorylates Ser-366 of eukaryotic elongation factor 2 (eEF2) kinase, resulting in its inactivation; reduced activity of eEF2 kinase contributes to reduction in Thr-56 phosphorylation of eEF2, which facilitates the elongation phase of translation (12,24). HG-induced activation of mTORC1, indicated by increased Thr-389 phosphorylation of p70S6 kinase and its downstream target ribosomal S6 protein (rpS6), was abrogated by SNP (Fig. 4, A and B). HG-induced changes in phosphorylation of eIF4E, eIF4G, eEF2, and eEF2 kinase were also inhibited by SNP pretreatment (Fig. 4, C-F).

GSK3␤ Activation Ameliorates Diabetic Nephropathy
Because SNP is an NO donor, we examined whether NO contributes to GSK3␤ activation by employing a structurally dissimilar NO donor. NONOate abrogated HG-induced phosphorylation of GSK3␤ and dephosphorylation of eIF2B⑀ (data not shown). Thus, SNP recruits the nitric oxide (NO) pathway to mitigate high glucose-induced GSK3␤ Ser-9 phosphorylation that results in its activation. Together, these data demonstrate that HG-induced GSK3␤ inhibition and stimulation of events in mRNA translation are blocked by SNP, thereby leading to reduced laminin synthesis in MCT cells. Based on these in vitro data, we tested whether SNP can block kidney matrix protein accumulation in type 1 diabetes.
SNP Administration in Diabetic Mice-To determine the optimal dose of SNP that activated GSK3␤ but did not reduce blood pressure, we administered 200, 400, and 800 g/kg of the agent in saline daily intraperitoneally for 6 days. At a 200 g/kg dose of SNP, the blood pressure measured on alternate days was unaffected; however, at a higher dose of 800 g/kg, the blood pressure fell (Fig. 5A). At a 200 g/kg dose, there was activation of GSK3␤ in the kidney cortex, as shown by a reduction in Ser-9 phosphorylation of GSK3␤ and increase in Ser-535 phosphorylation of eIF2B⑀ (Fig. 5B). Thus, the 200 g/kg/day dose of SNP was chosen and administered for 3 weeks. SNP did not affect blood glucose concentration in control or diabetic mice (Table 1). Body weight was significantly reduced, and the mean arterial pressure was elevated in the diabetic mice; these parameters were unaffected by SNP (Table 1).
SNP Ameliorates Diabetes-induced Kidney Hypertrophy and Albuminuria-The kidney/body weight ratio was increased significantly in diabetic mice demonstrating kidney hypertrophy (Fig. 6A), a cardinal early manifestation of kidney injury in diabetes; SNP significantly reduced renal hypertrophy (Fig. 6A), although the parameter was still higher than in non-diabetic mice (Fig. 6A). Diabetes caused an increase in albuminuria when compared with non-diabetic mice (Fig. 6B); it was partially but significantly inhibited by SNP (Fig. 6B).
SNP Inhibits Diabetes-induced Increase in Matrix Proteins-Renal tissue fibrosis due to accumulation of matrix proteins is a major contributor to kidney failure in diabetes; the profibrogenic cytokine TGF␤ plays an important role in stimulating renal fibrosis in diabetes (30). Diabetes increased the kidney FIGURE 5. Determination of optimal dose of SNP. A, mice were injected intraperitoneally with saline or SNP (200, 400, and 800 g/kg) in a similar volume daily for 6 days. Blood pressure was measured on alternate days by the tail cuff method in conscious mice using the CODA non-invasive blood pressure system (Kent Scientific Corp., Torrington, CT). Mice were trained for 1 week on the restrainer placed on a warm platform. Blood pressure (systolic, diastolic, and mean arterial pressure) in each mouse was recorded over 30 cuff inflations after 10 training cuff inflations (51). B, mice treated as above were sacrificed on day 6. Renal cortical homogenates were immunoblotted with antibodies against phospho (P)-Ser-9 GSK3␤, phospho-Ser-535 eIF2B⑀, GSK3␤, or eIF2B⑀. The histogram shows quantitative data from three mice in each group. Error bars, S.E. content of fibronectin, laminin ␤1, and laminin ␥1; these changes were significantly inhibited by SNP (Fig. 7, A-C). Diabetic mice showed higher expression of TGF␤ in the renal cortex, which was associated with an increase in SMAD3 phosphorylation, suggesting activation of the TGF␤ signaling pathway; SNP inhibited both of these parameters (Fig. 7, D and E). Immunoperoxidase staining for laminin trimer was performed in kidney cortical sections to assess matrix expansion (Fig. 8). Diabetes was associated with increased glomerular size and increased laminin deposition in the glomerular mesangium and in the tubulo-interstitium when compared with kidney tissues from control mice given vehicle alone (Fig. 8, C versus A). SNP abolished renal mesangial expansion and laminin accumulation in diabetic mice (Fig. 8, D versus C). Fig. 8, E and F, show morphometric quantification of the glomerular tuft area and fractional area of laminin staining in the glomerular tuft, respectively. The diabetes-induced significant increase in glomerular tuft area and laminin deposition was abrogated in mice treated with SNP.

SNP Inhibits Diabetes-induced Changes in Phosphorylation of GSK3␤ and Other
Signaling Proteins-We examined the signaling mechanisms involved in amelioration of kidney injury by SNP in diabetic mice. The diabetes-induced increased Ser-9 phosphorylation of GSK3␤ was abrogated by SNP (Fig. 9A), suggesting that SNP had reactivated the kinase. SNP tended to restore eIF2B⑀ phosphorylation in the renal cortex of diabetic mice (Fig. 9B). Interestingly, the increase in Ser-21 phosphorylation in GSK3␣ induced by diabetes was unaffected by SNP (Fig. 9A), implying selectivity in the regulation of GSK isoforms by SNP. We explored whether SNP regulated Akt and ERK, kinases that mediate HG stimulation of GSK3␤ phosphorylation in MCT cells (5). Diabetes increased the phosphorylation of Akt and ERK in the kidney cortex, showing their activation; both parameters were partly inhibited by SNP (Fig. 9, C and D). In order to identify the upstream regulators of ERK and Akt, we investigated the status of Src and Pyk2. Diabetes increased the phosphorylation of Src at Tyr-416 and Pyk2 at Tyr-402 that was abolished by SNP (Fig. 9, E and F). These data suggest that SNP exerts its inhibitory effect on GSK3␤ phosphorylation by blocking the activation of its upstream kinases.
SNP Blocks Diabetes-induced Changes in mRNA Translation-We explored whether SNP activation of GSK3␤ in diabetes leads to inhibition of mTORC1 and downstream events in mRNA translation that participate in matrix protein synthesis. Diabetes led to kidney parenchymal activation of mTORC1, as shown by an increase in Thr-389 and Ser-240/242 phosphorylation of p70S6 kinase and ribosomal S6 protein, respectively; both of the parameters were inhibited by SNP (Fig. 10, A and B). In addition, the diabetes-induced increase in phosphorylation of eIF4E and eEF2 kinase was also blocked by SNP (Fig. 10, C  and D). Thus, SNP ameliorated diabetes-induced mTORC1 activation (Fig. 10, A and B) and stimulation of initiation and elongation phases of translation (Fig 10, C and D), which are crucial for augmented protein synthesis involved in kidney hypertrophy and matrix protein increment.

DISCUSSION
Our findings show that in in vitro and in vivo models of diabetes-induced kidney injury, hyperglycemia leads to inhibition of GSK3␤ activity, allowing mTORC1 activation and stimulation of events in initiation and elongation phases of mRNA translation. These effects augment synthesis of proteins, including matrix proteins. Stimulation of GSK3␤ with SNP inhibits mTORC1 activation and protein synthesis, ameliorating matrix accumulation and albuminuria in mice with type 1 diabetes.
GSK3␣ and -␤ share 98% sequence identity in their catalytic domain (16). They seem to have the same substrate specificity  and are thought to phosphorylate glycogen synthase at a similar rate (31). GSK3␣ and -␤ appear to play a redundant role in mixed lineage proto-oncogene-driven leukemias, (32) and in maintaining the ␤-catenin levels in resting cells (33). Generation of mice in which phosphorylation of Ser-21/9 is not possible, due to knockin of alanine residues resulting in inactivation of GSK3␣ and -␤, has shed light on the functional importance of GSK3 mammalian physiology. GSK3␣/␤ S21A and S9A knock-in mice have fewer glomeruli and increased albuminuria, indicating that the kinase is important for glomerular development and integrity of barrier function against proteinuria (34).
Recent investigations have shown that the GSK3␣ and -␤ isoforms may have distinct roles. Whereas GSK3␤ knockout is embryonically lethal (35), GSK3␣ knock-out mice are viable, although they manifest accelerated aging and cardiac hypertrophy (36). An interesting finding to emerge from our studies is that GSK3␤ is preferably inactivated in diabetic kidney injury; SNP stimulated GSK3␤ rather than the GSK3␣ isoform, which was sufficient to ameliorate diabetic kidney injury. Additional investigation into the distinct roles of these isoforms in mediating or mitigating diabetic kidney injury will require the employment of individual Ser-21 and Ser-9 phosphorylation-deficient knock-in mice. We examined potential upstream mechanisms that may be responsible for the protective effect of SNP on diabetes-induced kidney injury via GSK3␤ activation. In addition to Akt (37), our previous studies had identified both ERK and mTOR/ p70S6 kinase as upstream kinases of GSK3␤ in mouse proximal tubular epithelial cells (5,37). There appears to be some selectivity in upstream kinases among kidney cells because ERK and p90Rsk but not Akt phosphorylate GSK3␤ in renal interstitial fibroblasts upon stimulation with tissue plasminogen activator (38). SNP led to the inhibition of ERK, Akt, and downstream targets of mTOR pathways in both in vitro and in vivo models employed in this study. In the present study, proximal events involved in the activation of Akt and ERK in diabetes were examined in the renal cortices and in proximal tubular epithelial cells in culture. We found that diabetes activated non-receptor tyrosine kinases Src and Pyk2, which led to activation of ERK and Akt that inactivate GSK3␤. Pyk2 is abundantly expressed in renal tubules; the kinase contributes to matrix accumulation in the kidney because Pyk2 Ϫ/Ϫ mice have decreased renal fibrosis following ureteral obstruction (39). In contrast to our data, Src activates GSK3␤ by tyrosine phosphorylation in prostate cancer (40), suggesting cell-and contextspecific regulation of GSK3␤ by its upstream kinases. Further studies are needed to understand the mechanisms involved in diabetes-induced activation of Src and Pyk2 in the kidney.
The other factor that could lead to activation of renal ERK and Akt in diabetes is TGF␤, a fibrogenic cytokine that facilitates synthesis of general proteins and matrix proteins contributing to renal hypertrophy and fibrosis (41). TGF␤ acts via dimerization of type I and type II receptors, leading to association with SMAD2 and -3 and activating them by phosphorylation (42,43). Diabetic mice showed an increase in TGF␤ expression and stimulation of its signaling via SMAD3 in this study; SNP administration abrogated it, suggesting that SNPinduced reduction in Akt, ERK, and mTORC1 may be in part due to inhibition of TGF␤ signaling.
GSK3␤ regulates several cellular events, including glucose metabolism and protein synthesis, by phosphorylation of its FIGURE 8. Immunohistochemical analysis of renal laminin expression. Immunoperoxidase staining with an antibody against laminin trimer showed increased deposition in glomerular mesangium in kidneys from diabetic mice (C) compared with non-diabetic control mice (A). SNP (D) ameliorated diabetesinduced changes in renal laminin expression when compared with kidney from diabetic mice (C). SNP did not affect laminin content in non-diabetic control mice (B). The histogram shows composite averages of renal glomerular tuft area (E) and immunoperoxidase staining intensities of laminin (F) in control and diabetic mice treated with or without SNP. For each mouse, 25-40 glomeruli were analyzed (n ϭ 4 in each group). Error bars, S.E.
substrates glycogen synthase and eIF2B⑀, respectively. eIF2B⑀ is a guanine nucleotide exchange factor that stimulates the GDP/ GTP exchange reaction of eIF2 during the initiation phase of mRNA translation (44). The signaling mechanism by which SNP inhibits diabetes-induced protein synthesis appears to involve inhibition of eIF2B⑀ by activation of GSK3␤. In addition, by inhibiting ERK activation in the kidney in diabetic mice, SNP also affected other events in the initiation phase, such as activation of eIF4E, the mRNA cap binding protein. Furthermore, SNP inhibited diabetes-associated stimulation of FIGURE 9. Sodium nitroprusside inhibits diabetes-induced signaling reactions in the kidney cortex. Equal amounts of protein from renal cortical lysate were immunoblotted with corresponding antibodies to assess changes in phosphorylation (P) of GSK 3␣/␤ (A) and eIF2B⑀ (B). Immunoblotting was done to identify upstream regulators of GSK 3␤ phosphorylation: Akt (C), ERK (D), Src (E), and Pyk2 (F). Immunoblotting with antibodies against respective total protein was done to assess loading. Representative blots and histograms are shown for 5-8 mice in each group. Error bars, S.E. mTORC1 and led to inhibition of p70S6 kinase. The latter not only facilitates the initiation phase but also directly regulates the elongation phase of mRNA translation. This is achieved by inhibition of eEF2 kinase, which contributes to dephosphorylation/activation of eEF2, leading to stimulation of the elongation phase. Our data show that negative regulation of mRNA translation by GSK3␤ could be an important therapeutic intervention in diabetic kidney injury because we have observed translational regulation of extracellular matrix proteins, such as laminin, in that disease (4,5,24,45). Together, these data show that SNP activation of GSK3␤ affects key regulatory steps in protein synthesis, resulting in amelioration of cardinal manifestations of diabetic kidney injury.
In contrast to our current and previous finding that diabetes is associated with GSK3␤ inactivation in the kidney (5), others have reported stimulation of GSK3␤ in the diabetic kidney and amelioration associated with inactivation of the kinase (46). However, that study employed diabetic mice at a more advanced age; previous investigators have reported changes in signaling pathways in the kidneys at a longer duration of diabetes (47). GSK3␤ has diverse roles in kidney pathology. Activation of GSK3␤ is associated with apoptosis in acute tubular necrosis, and administration of a single dose of lithium, a GSK3␤ inhibitor, is protective (48). In contrast, GSK3␤ seems to play a protective role in renal fibrosis associated with unilateral ureteral obstruction (49) and in a model of polycystic kidney disease (50). Thus, whether GSK3␤ plays an ameliorative or adverse role in kidney disease is context-specific.
The limitations of our study include the relatively short duration of observation; long term observation is needed to explore whether regulation of GSK3␤ changes with time. Our data do not address the efficacy of SNP in ameliorating kidney injury in a model of type 2 diabetes. Additional work is needed to explore whether kidney inflammation is affected by stimulation of GSK3␤ in diabetic mice and whether agents that only stimulate GSK3␤ would be as beneficial as SNP, which also affected the upstream kinases Akt and ERK. Future studies are planned to address these issues. In conclusion, our results suggest that inhibition of GSK3␤ contributes to kidney injury during the initial stages of diabetes. It provides evidence that the GSK3␤/ eIF2B⑀ axis may be a therapeutic target to reduce renal extracellular matrix protein accumulation in diabetes.