| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print December 18, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Biochemistry and Biomedical Sciences,* the Henderson Research Centre, and the Department of Medicine, McMaster University, Hamilton, Ontario, Canada
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
It has become increasingly evident that disruption of endoplasmic reticulum (ER) homeostasis and/or the activation of the unfolded protein response (UPR) can play a significant role in the development and progression of atherosclerotic lesions.5 Independent cardiovascular risk factors including hyperhomocysteinemia, obesity, cholesterol, and diabetes mellitus have been linked to elevated levels of diagnostic markers of ER stress and the induction of the UPR.6-9 The mechanisms by which conditions of ER stress promote these downstream, pro-atherogenic changes are not yet fully understood. We have previously shown that ER stress-inducing agents can promote cellular responses that represent the hallmark features of atherosclerosis, including apoptosis, cholesterol accumulation, and activation of inflammatory pathways in cultured human aortic smooth muscle cells (HASMCs) and HepG2 cells.9,10 We have also shown that glucosamine, a downstream metabolite of glucose, is an ER stress-inducing agent and that the accumulation of glucosamine in vascular tissue is associated with ER stress and accelerated atherosclerosis in hyperglycemic mice.9
Glycogen synthase kinase (GSK)-3
and -3β are serine/threonine kinases that are regulated by mitogens and growth factors including insulin, insulin-like growth factor-1, and epidermal growth factor, as well as conditions of cellular stress including heat shock and oxidative stress. A partial list of substrates includes the translation initiation factor eIF2β
, transcription factors CREB, c-Jun, c-Myc, c-Myb, HSF-1, NFAT1, β-catenin, cyclin D1, ATP-citrate lysase, and the amyloid precursor protein tau.11,12
GSK-3 plays a central role in many cellular processes including glycogen metabolism, gene expression, cell cycle regulation, and cell proliferation. GSK-3 appears to play a fundamental role in the promotion of apoptotic signaling,13,14
and the GSK-3β isoform is essential for the hepatoprotective activation of nuclear factor-
B.15,16
Disruption and/or dysregulation of GSK-3 activity has been implicated in several pathophysiological processes and diseases including diabetes mellitus, cardiac hypertrophy, Alzheimer’s disease, bipolar disorder, and cancer.17-21
Recent findings suggest that GSK-3 may be involved in the cellular response to conditions of ER stress. First, GSK-3 is involved in cellular processes that are induced by ER stress including apoptosis and inflammation.15,22 Second, ER stress-inducing agents have been shown to increase GSK-3 activity in rat cerebral cortical cells, mouse insulinoma cells, and mouse embryonic fibroblasts (MEFs).22-26 Third, increased GSK-3 activity has been observed in human muscle and mouse β-cells under hyperglycemic conditions.20,27-29 Finally, we have shown that small molecule inhibitors of GSK-3 attenuate ER stress-induced apoptosis and lipid accumulation in vitro.10
We hypothesize that hyperglycemia accelerates atherogenesis by a pathway that involves ER stress-induced activation of GSK-3 and that inhibition of GSK-3 will slow the progression of lesion development. Here we investigate the role of GSK-3 and -3β in ER stress-induced unesterified cholesterol accumulation in vitro and examine the effect of one inhibitor of GSK-3, valproate, on the development of hepatic steatosis and accelerated atherosclerosis in a diabetic apolipoprotein E-deficient (apoE–/–) mouse model.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Wild-type, GSK-3–/–, and GSK-3β–/– MEFs were generous gifts from Dr. Bradley Doble and Dr. James Woodgett (McMaster University, Hamilton, Canada, and University of Toronto, Toronto, Canada). The cells were cultured in Dulbecco’s modified Eagle’s medium (Life Technologies, Burlington, Canada) containing 10% fetal bovine serum and maintained in a humidified incubator at 37°C with 5% CO2. Sodium valproate, glucosamine, and filipin were purchased from Sigma (Oakville, Canada). GSK-3 inhibitor II, 3-(3-carboxy-4-chloroanilino)-4-(3-nitrophenyl) maleimide, was purchased from Calbiochem (La Jolla, CA). All compounds were prepared fresh in culture medium, sterilized by filtration, and added to the cell cultures.
Immunoblot Analysis
Antibodies to calreticulin (SPA-600) and the anti-KDEL monoclonal antibody (SPA-827) that recognizes GRP78/BiP were purchased from StressGen Biotechnologies (Victoria, Canada). Anti-GADD153/CHOP (sc-7351) monoclonal antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-β-actin (AC-15) antibodies were purchased from Sigma. MEFs were treated with 5 mmol/L glucosamine, 10 µg/ml tunicamycin, or 7 µmol/L A23187 for 18 hours. Total protein lysates (40 µg) from MEFs were solubilized in sodium dodecyl sulfate-polyacrylamide gel electrophoresis sample buffer and separated on sodium dodecyl sulfate-polyacrylamide gels under reducing conditions. After incubation with the appropriate primary and horseradish peroxidase-conjugated secondary antibodies (Life Technologies), the membranes were developed using the Immobilon Western chemiluminescent horseradish peroxidase substrate (Millipore, Billerica, MA).
Unesterified Cholesterol Staining
The accumulation of unesterified cholesterol was determined by filipin staining.30 MEFs, grown on coverslips, were treated with 4 mmol/L valproate, or 200 nmol/L GSK-3 inhibitor II for 1 hour and challenged with glucosamine (5 mmol/L) for an additional 18 hours. Cells were washed three times with medium 1 (150 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L CaCl2, 20 mmol/L HEPES pH 7.4, 2 g/L glucose), fixed with 4% paraformaldehyde for 20 minutes at room temperature and then incubated for 2 hours with 50 µg/ml of filipin in medium 1 at room temperature. Cells were washed three times with medium 1 and then filipin-bound unesterified cholesterol complexes were visualized by fluorescence microscopy with excitation at 335 to 385 nm (emission at 420 nm). The accumulation of unesterified cholesterol in primary mouse hepatocytes was quantified using Sigma Plot (Systat Software, Inc., Chicago, IL) as described previously.10
Analysis of Unesterified Cholesterol Staining by Fluorescence-Activated Cell Sorting (FACS)
MEFs were grown on 100-mm-diameter culture dishes to 50% confluency and treated with the appropriate drugs as described above. The cells were washed three times with medium 1, fixed in 4% paraformaldehyde at room temperature for 20 minutes, and incubated with 50 µg/ml filipin for 2 hours at room temperature. Cells were washed with phosphate-buffered saline (PBS), trypsinized, and stored in PBS containing 1% fetal bovine serum. Filipin fluorescence was measured using a FACS VantageSE instrument (Becton Dickinson, San Jose, CA) with 350 to 70 nm UV laser and emission detection at 420 to 460 nm. Data were analyzed using FlowJo software (Treestar, Ashland, OR).
Mouse Models
Five-week-old female apoE–/– (B6.129P2-ApoEtm1Unc) mice were placed on a defined chow diet (TD92078; Harlan Teklad, Madison, WI) and randomly divided into two groups (n = 24 per group). To induce hyperglycemia, one group was injected intraperitoneally with multiple low doses of streptozotocin (STZ) (40 mg/kg) as previously described.9,31 After 1 week, half of the mice in each group were switched to the control diet supplemented with 625 mg/kg of sodium valproate (TD02165). All mice had unrestricted access to both food and water throughout the study. Mice were sacrificed at 15 weeks of age and blood and tissues were collected for further analysis. To test the effect of lithium, mice with STZ injections were given the control diet supplemented with 4 g of LiCO3/kg chow (TD05217) (Harlan Teklad, Madison, WI). The McMaster University Animal Research Ethics Board approved all procedures.
Plasma Analysis
Random and fasting (12 hours) whole blood glucose levels were measured using a DEX glucometer (Bayer, Pittsburgh, PA). Plasma glucose and lipid levels were determined in nonfasting mice using the colorimetric diagnostic kits for total cholesterol, triglycerides, and glucose purchased from Thermal DMA Inc. (Melbourne, Australia) Plasma valproate concentrations were determined using an AxSYM system (Abbott Laboratories, Mississauga, Canada).
RNA Isolation and Quantitative Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)
Total RNA was extracted from apoE–/– mouse livers using commercially available TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. The concentration and purity of the RNA were determined by measuring the absorbance at 260 and 280 nm. cDNA was synthesized from 2 µg of total RNA using a commercially available kit (High Capacity cDNA reverse transcription kit; Applied Biosystems, Foster City, CA) following the guidelines provided by the manufacturer. Real-time PCR was conducted at the following settings: one cycle at 95°C for 10 minutes, followed by 40 cycles at 95°C for 1 minute, and at 60°C for 1 minute using a GeneAmp 7300 sequence detection system (Applied Biosystems,) and SYBR GreenER qPCR SuperMix for ABI Prism (Invitrogen). Template cDNAs were amplified using the following primers: for sterol regulatory element binding protein (SREBP)-2, the forward primer was 5'-GCGTTCTGGAGACCATGGA-3' and reverse was 5'-ACAAAGTTGCTCTGAAAACAAATCA-3', the amplicon size is 131 bp; for SREBP-1c, the forward primer was 5'-GGAGCCATGGATTGCACATT-3' and reverse was 5'-GCTTCCAGAGAGGAGGCCAG-3', the amplicon size is 
170 bp; for 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMG-CoAR), the forward primer was 5'-CTTGTGGAATGCCTTGTGATTG-3' and reverse was 5'-AGCCGAAGCAGCACATGAT-3', the amplicon size is 76 bp; for fatty acid synthase (FAS), the forward primer was 5'-GCTGCGGAAACTTCAGGAAAT-3' and reverse was 5'-AGAGACGTGTCACTCCTGGACTT-3', the amplicon size is 84 bp; for β-actin, the forward primer was 5'-GGCACCACACCTTCTACAATG-3' and the reverse was 5'-GGGGTGTTGAAGGTCTCAAAC-3'; the amplicon size is 132 bp. For GRP78, the forward primer was 5'-ACCTGGGTGGGGAAGACTTT-3' and reverse was 5'-TCTTCAAATTTGGCCCGAGT-3', the amplicon size is 232 bp; for PDI, the forward primer was 5'-AACGGGAGAAGCCATTGTA-3' and reverse was 5'-AGGTGTCATCCGTCAGCTCT-3', the amplicon size is 156 bp; for spliced XBP-1, the forward primer was 5'-GTCCGCAGCAGGTGCAG-3' and reverse was 5'-GGGAGTTCCTCCAGACTAGC-3', the amplicon size is 177 bp; for CHOP, the forward primer was 5'-GTCCCTAGCTTGGCTGACAGA-3' and the reverse was 5'-TGGAGAGCGAGGGCTTTG-3'. The standard method was used for the quantification of the mRNA expression in the livers using β-actin as a normalization control gene.
Immunohistochemical Analysis
Mice were euthanized, hearts were flushed with 1x PBS, and perfusion-fixed with 10% neutral buffered formalin. After removal, hearts, including the aortic root, were cut transversely and embedded in paraffin. Serial sections (4 µm) of aortic root were collected on precoated glass slides for measurement of lesion size (hematoxylin and eosin staining) or immunohistochemical staining.32 The Vectastain ABC System (Vector Laboratories, Burlingame, CA) was used for immunohistochemical analysis. Sections were stained with primary antibodies and visualized using appropriate biotinylated secondary antibodies and Nova Red. Nonspecific staining was controlled for using a similar section and pre-immune IgG. Images were captured with a charge-coupled device color video camera (Sony, Tokyo, Japan) and analyzed using Northern Exposure (Empix, Tonawanda, NY) and SigmaScan Pro software.
Lipid Staining
Liver cryosections, 7 µm thick, were collected on slides and fixed in formal calcium for 30 minutes. Neutral lipids were visualized in sections that were rinsed and incubated in a saturated, filtered solution of Oil Red O (Sigma) for 7 minutes. After rinsing with H2O, nuclei were stained with hematoxylin, and slides were mounted in Crystalmount.33 Relative amounts of lipid staining were quantified using Image J 1.35 Software (National Institutes of Health, Bethesda, MD).
Immunoblot Analysis of Mouse Liver Proteins
Cytoplasmic proteins were isolated from mouse liver using a NE-PER kit (Pierce, Rockford, IL) and total protein lysates were prepared from mouse liver solubilized in RIPA buffer. Equivalent amounts of total protein were separated on sodium dodecyl sulfate-polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA), as described previously.10,33 After incubation with the appropriate primary and horseradish peroxidase-conjugated secondary antibodies, the membranes were developed using the Immobilon Western chemiluminescent horseradish peroxidase substrate (Millipore). Specific bands were quantified using the Typhoon 9410 Imaging System (GE Healthcare, Waukesha, WI).
GSK-3β Activity Assay
GSK-3β was immunoprecipitated from total protein lysates prepared from mouse liver. Kinase activity was measured by monitoring the incorporation of 32P onto phospho-glycogen synthase peptide-2 (pGS-2; Upstate, Billerica, MA) as previously described.34
Statistical Analysis
Results are presented as the mean ± SD. One-way analysis of variance was used for repeated measurements of the same variable, and Duncan’s multiple range t-test was used to assess differences between experimental groups and controls. Probability values of <0.05 were considered statistically significant.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Conditions of ER stress have previously been associated with increased GSK-3 activity in various rodent cells.23-26
However the effect of GSK-3 deficiency on the ability of a cell to initiate the UPR is not known. To determine whether GSK-3 /β activity is directly involved in the UPR, we challenged wild-type MEFs and MEFs that are deficient in GSK-3 or GSK-3β with ER stress-inducing agents, including tunicamycin (10 µg/ml), the calcium ionophore A23187 (7 µmol/L), and glucosamine (5 mmol/L) for 18 hours and measured the induction of specific UPR proteins (Figure 1
and see Supplemental Figure S1 at http://ajp.amjpathol.org). The results indicate that each of the MEF cell lines respond to various ER stress-inducing agents by significantly increasing the levels of GRP78 and GADD153. Basal levels of calreticulin are relatively high in the MEFs, and the fold induction under conditions of ER stress is not significant. These findings suggest that GSK-3 and GSK-3β do not play essential roles in the induction of the UPR in MEFs.
|
and -3β Deficiency on ER Stress-Induced Cholesterol Accumulation
It has been established that conditions of ER stress can promote cholesterol accumulation in specific cell types including hepatocarcinoma cells.10
We examined the potential role of GSK-3 activity in this response using the GSK-3-deficient MEFs. Filipin staining of MEFs treated with glucosamine was analyzed using FACS and fluorescence microscopy (Figure 2)
. Results show that exposure to 5 mmol/L glucosamine induces significant unesterified cholesterol accumulation in wild-type MEFs but not in GSK-3–/– or GSK-3β–/– MEFs. This effect appears to be ER stress-specific because the GSK-3-deficient MEFs do accumulate unesterified cholesterol when exposed to U18666A, a chemical that blocks intracellular trafficking of cholesterol, thereby activating cholesterol biosynthesis (see Supplemental Figure S2 at http://ajp.amjpathol.org).35
These findings suggest that both the and β GSK-3 isoforms play a specific role in ER stress-induced unesterified cholesterol accumulation.
|
We have previously shown that the branched chain fatty acid, valproate, can inhibit both recombinant and intracellular GSK-3/β activity.10,34
We have also shown that valproate and other GSK-3 inhibitors attenuate ER stress-associated unesterified cholesterol accumulation and apoptosis in cultured HASMCS and HepG2 cells (see Supplemental Figure S3 at http://ajp.amjpathol.org).10
Here we examine the effect of valproate, and a potent maleimide-based GSK-3 inhibitor, on glucosamine-induced cholesterol accumulation in wild-type MEFs. FACS analysis of filipin-stained wild-type MEFs indicates that glucosamine-induced unesterified cholesterol accumulation is significantly lower in cells pretreated with valproate or GSK-3 inhibitor II (Figure 3)
. This result is consistent with the previous findings suggesting that GSK-3 activity plays a role in ER stress-associated unesterified cholesterol accumulation.
|
Hyperglycemia was induced in female apoE–/– mice by multiple low-dose intraperitoneal injections of STZ.9,31
Blood glucose levels were monitored weekly in all experimental groups. All STZ-injected mice were hyperglycemic by the time of the last injection. At the time of sacrifice (15 weeks of age), 12-hour fasting plasma glucose concentration in STZ-injected mice was 7.6 ± 2.1 mmol/L versus 4.3 ± 0.7 mmol/L (P < 0.05) in control mice (Table 1)
.
|
Supplementation of defined mouse chow with 625 mg/kg of sodium valproate increased plasma valproate levels to 37.5 ± 14.9 µmol/L in control apoE–/– mice (Table 1)
. Plasma valproate levels were not significantly different in hyperglycemic apoE–/– mice relative to normoglycemic mice. Valproate supplementation had no significant effect on body or liver weight in any of the experimental groups relative to age-matched mice fed control diet. In addition there was no significant effect on fasting and fed plasma glucose [10 weeks (data not shown) or 15 weeks (Table 1)
], plasma cholesterol (Table 1)
, or lipid profile (data not shown) as a result of valproate supplementation. Consistent with previous studies, the induction of hyperglycemia did not significantly affect plasma cholesterol levels in 15-week-old apoE–/– mice.9
These findings indicate that this level of valproate supplementation is subsymptomatic, which is consistent with previous preclinical analysis of valproate toxicity in rodents.36
When dietary levels of sodium valproate were increased to 1250 mg/kg chow, corresponding plasma valproate levels increased to 131.4 ± 10.1 µmol/L. This level of supplementation did correspond to a significant decrease in plasma glucose levels in STZ-treated apoE–/– mice (from 23.2 ± 5.6 to 13.1 ± 2.0 mmol/L, P < 0.05). The diet containing the lower (625 mg/kg) concentration of sodium valproate was used in all subsequent experiments.
Effects of Hyperglycemia and Valproate on Hepatic Levels of O-Linked Protein Glycosylation
Glucosamine is a product of glucose flux through the hexosamine pathway and an ER stress-inducing agent that we have previously linked to diabetic atherogenesis.9,10
Alterations in intracellular glucosamine concentration were determined by evaluating the extent of O-linked glycosylation, as described previously.9,13
Hyperglycemic mice have elevated levels of O-linked N-acetyl glycosylation (GlcNAc) in aortae relative to control mice (Figure 4)
. The supplementation of diet with sodium valproate did not affect the observed levels of O-linked glycosylation in cells of the vascular lesions (Figure 4A)
. This result is supported by quantification of O-linked glycosylated protein p62 in liver by immunoblot analysis (Figure 4B)
.13
These findings are consistent with the similar blood glucose concentrations observed in supplemented and nonsupplemented mice.
|
Liver and aortic cross sections from mice in each of the treatment groups were immunostained with an anti-KDEL antibody, directed against GRP78/94, or an anti-phospho-PERK antibody. Consistent with our previous findings, hyperglycemia correlates with increased protein levels of KDEL and phospho-PERK staining in smooth muscle cells and macrophages that make up the vascular lesion (Figure 5A)
. Dietary supplementation with sodium valproate did not significantly affect the protein levels of these ER stress markers in atherosclerotic lesions of control or hyperglycemic mice. Similar results were observed by immunohistochemical analysis in the livers from the same experimental groups (Figure 5B)
. The expression levels of hepatic ER stress response genes were also examined using real time RT-PCR (see Supplemental Figure S4 at http://ajp.amjpathol.org). Results indicate that mRNA transcripts encoding ER stress response genes are significantly elevated in hyperglycemic mice and that supplementation with sodium valproate attenuates this effect.
|
GSK-3β was immunoprecipitated from hepatic protein lysates from normoglycemic and hyperglycemic mice and kinase activity was determined as previously described.34
GSK-3β kinase activity was significantly elevated in the liver of hyperglycemic mice fed control diet compared to normoglycemic mice (Figure 6)
. Hepatic GSK-3β activity in hyperglycemic apoE–/– mice fed valproate-supplemented diet was similar to the levels found in normoglycemic mice and significantly lower than hyperglycemic mice fed the control diet. This indicates that valproate can inhibit GSK-3β in vivo.
|
Cryosections of liver isolated from each of the treatment groups were stained with Oil Red O (Figure 7A)
. Consistent with our previous findings, conditions of hyperglycemia promote the accumulation of hepatic lipids in apoE–/– mice.9
Significantly reduced levels of hepatic lipid were present in hyperglycemic mice that received dietary supplementation with valproate relative to nonsupplemented controls (Figure 7B)
. This finding is supported by an observed increase in hepatic mRNA transcripts encoding proteins involved in lipid biosynthesis, including SREBP-1c, FAS, and HMG-CoA reductase, in hyperglycemic mice fed control diet. Hyperglycemic mice fed the valproate-supplemented diet had significantly lower levels of these transcripts (Figure 7C)
.
|
Hyperglycemia significantly increased mean atherosclerotic lesion volume at the aortic root of apoE–/– mice relative to control apoE–/– mice (2.3-fold) (Figure 8
and Table 1
), consistent with reported data.37
Dietary supplementation with sodium valproate correlated to a significant decrease in cross-sectional lesion area and total mean lesion volume in hyperglycemic mice (4.98 ± 0.76 versus 8.62 ± 1.23 x 10–3mm3, P < 0.05) relative to nonsupplemented controls (Figure 8
and Table 1
). Hyperglycemic apoE–/– mice fed a diet supplemented with lithium (4 g LiCO3/kg chow), a selective GSK-3 inhibitor, also had reduced lesion size at the cross section of aortic root compared to control diet-fed mice (see Supplemental Figure S5 at http://ajp.amjpathol.org). In hyperglycemic mice the reduced lesion volume corresponded with a significant reduction in the number of Mac-3-positive macrophages/foam cells within the atherosclerotic plaque (Table 1)
. A nonsignificant trend toward smaller lesion volume was observed in normoglycemic apoE–/– mice fed the valproate diet. In all mice, atherosclerotic lesion development was confined to a region within 120 µm of the aortic root and no lesions were observed in the descending aorta or the aortic arch of any of the treatment groups (data not shown).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In this study we begin to identify the molecular links by which ER stress may promote atherosclerosis and present evidence that GSK-3 plays a central role in the transmission of pro-atherogenic signals under conditions of hyperglycemia. In vitro, we show that MEFs deficient in GSK-3 have a similar UPR but are resistant to ER stress-induced cholesterol accumulation relative to wild-type MEFs, supporting a role of GSK-3 subsequent to ER stress but preceding cellular dysfunctions. In vivo we show that STZ-induced hyperglycemic apoE–/– mice have increased GSK-3 activity, elevated ER stress levels, hepatic steatosis, and significantly enlarged atherosclerotic lesion volume relative to normoglycemic controls. Dietary supplementation with valpoate, a molecule with GSK-3 inhibitory properties,17,34 normalizes hepatic GSK-3 activity, reduces hepatic steatosis, and attenuates accelerated atherosclerosis in this model. Together these findings suggest that conditions of hyperglycemia may activate ER stress response pathways that signal through GSK-3 to dysregulate cholesterol metabolism and perhaps activate other pro-atherosclerotic pathways.
The present study is based on previous results showing that elevated concentrations of glucose and glucosamine can promote ER stress in cultured hepatocytes, vascular smooth muscle cells, and macrophages.9
We and others have demonstrated that ER stress can lead to caspase activation and programmed cell death in cultured endothelial cells,38,39
and SREBP activation and subsequent lipid accumulation in HASMCs and hepatocytes.33,40
Exposure of cultured HASMCs, Thp-1, and HepG2 cells to glucosamine promotes a robust ER stress response and the induction of nuclear factor-B and SREBP activity.9,10
The molecular mechanisms by which glucosamine disrupts ER homeostasis and the pathways by which ER stress activate pro-atherogenic responses are not fully understood.
It has been previously reported that ER stress-inducing agents, including thapsigargin and brefeldin A, increase GSK-3 activity in human neuroblastoma cells,22 mouse insulinoma cells,23 MEFs,24 and rat cerebral cortical cells.25,26 The mechanism by which GSK-3 is activated possibly involves the dephosphorylation of phospho-Ser-9 of GSK-3β, an event that switches GSK-3 from an inactive to an active kinase.22,23,25 ER stress-induced GSK-3 activity promotes caspase-dependent apoptosis in neuronal cells, and pretreatment with lithium or small molecule inhibitors of GSK-3 is sufficient to protect cells from ER stress-induced programmed cell death.22,25,26
Valproate is a widely prescribed drug in the treatment of epilepsy and bipolar disorder; however, the mechanism(s) by which valproate alleviates convulsions and modifies behavior remain controversial.18,41 Valproate has many intracellular effects that may be responsible for its clinical efficacy. These include its ability to potentiate GABA-mediated postsynaptic inhibition,42 to promote the depletion of intracellular inositol,43 to inhibit histone deacetylase (HDAC) activity,44 to promote the expression of protein chaperones,10,45 and to directly and/or indirectly inhibit GSK-3 activity.17,19,34 We have shown that exposure to valproate can protect cultured cells, including HepG2 and HASMCs, from ER stress-induced lipid accumulation and apoptosis.10 The ability of valproate-derived compounds to attenuate ER stress-induced lipid accumulation correlates directly with their ability to inhibit GSK-3 (see Supplemental Figure S3 at http://ajp.amjpathol.org). The ability of other established GSK-3 inhibitors, such as alsterpaullone and a synthetic bisaryl-maleimide,26 to attenuate ER stress-induced cellular dysfunction supports a role for GSK-3 in ER stress signaling.
Increased GSK-3β activity has been reported in skeletal muscle from humans with type 2 diabetes27
and adipose tissue from C57BL/6 mice with diet-induced diabetes.28
We have determined that hepatic GSK-3β activity in STZ-induced hyperglycemic apoE–/– mice is more than twofold higher than normoglycemic controls. This increase corresponds to an elevation of intracellular glucosamine levels [1.8-fold (P < 0.05) increase in O-linked GlcNAc] and a similar increase in hepatic ER stress levels as well as a twofold (P < 0.05) elevation in hepatic lipid content. Hepatic steatosis is characteristic of hyperglycemic rodent models9,46
and is also a complication associated with diabetes mellitus in humans.47
We propose that glucosamine-induced ER stress promotes hepatic lipid accumulation by a pathway that involves GSK-3 activation. This model is supported by the findings that GSK-3–/– MEFs do not accumulate unesterified cholesterol under conditions of ER stress (Figure 2)
and that hyperglycemic mice fed a valproate-supplemented diet have a significant reduction in hepatic GSK-3β activity (Figure 6)
and hepatic lipid levels (Figure 7)
, but no change in ER stress levels (Figure 5)
relative to nonsupplemented hyperglycemic mice.
It is well established that STZ-induced hyperglycemia promotes accelerated atherosclerosis in apoE–/– mice.9,31
As with hepatic steatosis, accelerated atherosclerosis is associated with increased levels of glucosamine and ER stress in the cells of the vascular lesion (Figures 4 and 5)
.9
Here we show that dietary supplementation with valproate is associated with significantly reduced lesion volume in hyperglycemic mice but not in normoglycemic mice.
It is not clear at this time if the vascular and hepatic effects of valproate supplementation are mechanistically linked or are independent phenomena. Hepatic GSK-3β activity is clearly reduced in supplemented mice relative to controls (Figure 6)
; however, it was not possible to determine vascular activity because of the limited amount of tissue available. Therefore, it is possible that valproate indirectly inhibits atherogenesis by altering hepatic lipid metabolism, although the fact that plasma lipid levels are not significantly affected suggests that the effects of valproate are a result of localized action. It is important to note that the plasma valproate concentration attained in this mouse model is 10 times lower than clinically relevant plasma concentrations for the treatment of epilepsy in humans (0.28 to 0.7 mmol/L48
). There is no clear correlation between valproate use and cardiovascular disease or incidence of myocardial infarction in human populations, although any association is likely complicated by weight gain that is observed in humans but not in rodents taking valproate.48,49
Together these findings suggest that the observed hepatic, and perhaps vascular, effects of dietary supplementation result from the ability of valproate to attenuate the increase in GSK-3 activity observed under conditions of hyperglycemia. We hypothesize that conditions of hyperglycemia accelerate atherogenesis by a pathway involving ER stress-induced GSK-3 activity and that valproate attenuates this effect by directly or indirectly inhibiting GSK-3 signaling.
Although the data presented here is consistent with this hypothesis, there are several other potential explanations for the observed effect. First, it has been shown that GSK-3 inhibition by lithium can have some insulin mimetic effects involving stimulation of glucose uptake and glycogen synthesis in adipose and muscle cells.50,51
However, we observed no significant changes in plasma glucose levels or intracellular O-linked glycosylation in mice supplemented with 625 mg/kg of valproate suggesting that this mechanism is not playing a major role. Second, the recognized ability of valproate to affect other intracellular enzymes and pathways could produce effects that slow or block atherogenesis. Valproate is a known HDAC inhibitor and it is therefore possible that valproate confers athero-protection by altering gene expression patterns through the inhibition of HDAC.44,52
The HDAC inhibitor, trichostatin (TSA), has been shown to inhibit vascular smooth muscle cell proliferation53
but actually exacerbates atherosclerosis in an LDL receptor-deficient mouse model.54
We observed no significant difference in hepatic histone acetylation in any of the mouse treatment groups suggesting that HDAC activity was unaffected by hyperglycemia and the concentration of plasma valproate induced by supplementation (data not shown). Exposure to valproate is known to increase the expression of cellular chaperones including HSP70, GRP78, HSP47, protein disulfide isomerase (PDI), and calreticulin in rat neurons and in cultured HepG2 cells.10,45
Overexpression of GRP78 or calreticulin can protect certain cells from stress-induced dysfunction.33,55
Furthermore, a recent report has shown that small molecule chaperones can alleviate ER stress and improve glucose homeostasis and insulin sensitivity in a mouse model of type 2 diabetes.7
However, no significant increase in ER chaperone expression was observed in liver or aortae of mice fed the valproate diet, and we have shown in vitro10
and in vivo that exposure to valproate does not significantly affect the protein levels of diagnostic markers of ER stress in cells and tissue (Figure 5)
. We conclude that valproate must confer its cytoprotective properties by acting subsequent to the induction of ER stress and the UPR. Finally, it must be noted that valproate can be metabolized into a complex array of intermediates,56
any one of which may be acting as the direct effector on one or more of these pathways. Further studies will be required to more clearly determine which, if any, of these factors and mechanisms are directly responsible for the observed athero-protective effect.
Identification of the factors and pathways that mechanistically link diabetes mellitus to cardiovascular disease is vital to our understanding of the molecular mechanisms of disease development and progression. The validation of a pro-atherogenic role for ER stress and the specific involvement of GSK-3 will facilitate the identification of other potential targets and small molecule interventions toward the goal of developing effective therapies to treat and prevent cardiovascular disease.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Heart and Stroke Foundation of Ontario (grant NA5556/T6104), the Canadian Institutes of Health Research (grant MOP-62910), the Heart and Stroke Foundation of Canada (New Investigators grant to G.H.W.), the Natural Sciences and Engineering Research Council of Canada (scholarship to A.J.B.), and the Canadian Diabetes Association (scholarship to A.J.B.).
Supplemental material for this article can be found on http://ajp.amjpathol.org.
Accepted for publication September 11, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-agonist) attenuates atherogenesis with no effect on hyperglycaemia in a combined diabetes-atherosclerosis mouse model. Diabetes Obes Metab 2003, 5:45-50[CrossRef][Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
