Cardiac PI3K-Akt Impairs Insulin-Stimulated Glucose Uptake Independent of mTORC1 and GLUT4 Translocation

Abstract

Impaired insulin-mediated glucose uptake characterizes cardiac muscle in humans and animals with insulin resistance and diabetes, despite preserved or enhanced phosphatidylinositol 3-kinase (PI3K) and the serine-threonine kinase, Akt-signaling, via mechanisms that are incompletely understood. One potential mechanism is PI3K- and Akt-mediated activation of mechanistic target of rapamycin (mTOR) and ribosomal protein S6 kinase (S6K), which may impair insulin-mediated activation of insulin receptor substrate (IRS)1/2 via inhibitory serine phosphorylation or proteasomal degradation. To gain mechanistic insights by which constitutive activation of PI3K or Akt may desensitize insulin-mediated glucose uptake in cardiomyocytes, we examined mice with cardiomyocyte-restricted, constitutive or inducible overexpression of a constitutively activated PI3K or a myristoylated Akt1 (myrAkt1) transgene that also expressed a myc-epitope-tagged glucose transporter type 4 protein (GLUT4). Although short-term activation of PI3K and myrAkt1 increased mTOR and S6 signaling, there was no impairment in insulin-mediated activation of IRS1/2. However, insulin-mediated glucose uptake was reduced by 50–80%. Although longer-term activation of Akt reduced IRS2 protein content via an mTORC1-mediated mechanism, treatment of transgenic mice with rapamycin failed to restore insulin-mediated glucose uptake, despite restoring IRS2. Transgenic activation of Akt and insulin-stimulation of myrAkt1 transgenic cardiomyocytes increased sarcolemmal insertion of myc-GLUT4 to levels equivalent to that observed in insulin-stimulated wild-type controls. Despite preserved GLUT4 translocation, glucose uptake was not elevated by the presence of constitutive activation of PI3K and Akt. Hexokinase II activity was preserved in myrAkt1 hearts. Thus, constitutive activation of PI3K and Akt in cardiomyocytes impairs GLUT4-mediated glucose uptake via mechanisms that impair the function of GLUT4 after its plasma-membrane insertion.

Regulation of glucose transport and use by insulin is central to the maintenance of whole-body glucose homeostasis (1). Insulin regulates glucose transport and use via a complex signaling cascade involving serial phosphorylation and activation of kinases such as phosphatidylinositol 3-kinase (PI3K) and the serine-threonine kinase Akt, also known as protein kinase B. The binding of insulin to the extracellular domain of the insulin receptor (IR) leads to phosphorylation and a conformational change in its intracellular tyrosine kinase domain that allows the IR to serve as a signaling scaffold and docking site for many proteins, including IR substrate (IRS) proteins (2). IRS proteins promote the assembly of regulatory and catalytic subunits of PI3K, which phosphorylates phosphatidylinositol 4,5-bisphosphate to generate phosphatidylinositol 3,4,5 triphosphate (PIP3) (3). Production of PIP3 recruits phosphoinositide-dependent kinase 1 to the plasma membrane (PM) for activation, which then phosphorylates Akt and other signaling intermediates (4). Activation of PI3K and Akt plays critical roles in the regulation of insulin-stimulated glucose uptake by promoting translocation of glucose transporter type 4 protein (GLUT4).

In cardiomyocytes, the two most highly expressed glucose transporters are GLUT1 and GLUT4, with GLUT4 being most abundant. GLUT1 mediates basal glucose transport, whereas GLUT4 is mainly responsible for insulin or contraction-mediated glucose transport (5). In quiescent myocytes, the majority of GLUT4 protein resides in a specialized vesicle population in an intracellular compartment. Upon insulin stimulation, GLUT4 vesicles are translocated to the PM via a multiple-step process, by which GLUT4 storage vesicles (GSVs) move to the PM, tether, dock, and ultimately fuse with the PM to expose GLUT4 proteins on the cell surface via a process involving modulation of cortical F-actin networks, snare proteins, and small GTPases of the Rab family (6). In addition to activation of PI3K and Akt, insulin regulates GLUT4 trafficking at multiple levels, such as phosphorylation of the Rab GTPase-activating protein Akt substrate 160 KD (AS160), which activates a family of small G proteins (Rabs) to promote movement of GSVs to the PM (7). At the PM, insulin activates PI3K and RalA, to promote tethering and docking of GSVs to an actin filament network (8). Insulin also activates Rac1 protein to promote actin remodeling that facilitates the GSV docking process (9). The rate of final GSV and PM fusion depends on physical properties of the PM and also requires activation of protein kinase C-ζ, both of which are regulated by insulin (10). In neonatal cardiomyocytes, an additional Ca²⁺ signal activates Akt to promote GLUT4 translocation (11).

Insulin-stimulated glucose transport is an important rate-limiting step for glucose metabolism in adipose tissue, cardiac and skeletal muscle, and this fundamental process is severely disrupted in type 2 diabetes and other insulin-resistant states (12). However, the mechanisms that lead to the development of insulin resistance and reduced insulin-stimulated glucose transport remain incompletely understood, particularly in the heart. Chronic insulin exposure desensitizes insulin signaling in many cell types and tissues. For example, in L6 myoblasts, chronic insulin treatment decreased Akt and MAPK phosphorylation and glucose uptake in response to subsequent insulin stimulation, via a mechanism involving reduced content of IRS1 and IRS2 proteins. Inhibition of PI3K by LY294002 restored IRS1 and IRS2 protein content and increased the responsiveness of Akt and MAPK toward subsequent insulin stimulation (13). Loss of function of the Tuberous sclerosis protein 1 (TSC1) and Tuberous sclerosis protein 2 (TSC2) complex leads to constitutive, unrestrained Rheb-mTORC1-S6K1 signaling. In TSC1^−/− or TSC2^−/− mouse embryonic fibroblasts, there was a reduction of IRS1 and IRS2 proteins, which resulted in a decrease of insulin signaling (14). Furthermore, deletion of ribosomal protein S6 kinase (S6K) protected mice from age- and diet-induced obesity and enhanced insulin sensitivity (15). Thus, an mTORC1-S6K1-IRS negative feedback pathway activated by abnormal PI3K-Akt activity may lead to insulin resistance.

In diabetic patients, basal and insulin-stimulated cardiac glucose transport are reduced (16, 17), but PI3K-Akt signaling is increased (17). Whether elevated PI3K-Akt signaling activates this mTORC1-S6K1-IRS negative feedback pathway in the heart leading to reduced glucose uptake is not known. This study sought to understand the mechanism by which activation of PI3K-Akt signaling in the heart may modulate cardiac glucose uptake in two mouse models overexpressing either a constitutively activated PI3K (caPI3K) or a myristoylated Akt1 (myrAkt1). We also tested whether the mTORC1 inhibitor rapamycin could restore the insulin-stimulated glucose uptake in these hearts. Our study reveals that activation of PI3K and Akt in the heart impairs glucose uptake despite intact GLUT4 translocation by mechanisms that are independent of mTORC1 activation.

Materials and Methods

Animals and rapamycin administration

caPI3K transgenic mice were previously generated by Seigo Izumo (18) by expressing a chimeric protein that contains the iSH2 domain of p85 fused to the N terminus of bovine p110α driven by the α-myosin heavy chain promoter. The caPI3K mice are on a pure FVB background.

myrAkt1 mice were previously generated by Kenneth Walsh (19). myrAkt1 mice have two transgenes: tetracycline transactivator (tTA) driven by α-myosin heavy chain promoter and a myrAkt1 transgene under the 'tet'O promoter. myrAkt1 mice were fed doxycycline (DOX) chow (1 g/kg) until the predetermined time (8 wk old) for myrAkt1 transgene induction. myrAkt1 mice are on mixed background.

myrAkt1 mice were mated to mice hosting a c-myc GLUT4 construct (20) driven by the 'tet'O promoter to generate cardiac-specific myrAkt1 and c-myc GLUT4 coexpressing mice (triple transgenic mice). Rapamycin at 2 mg/kg·d was administered by ip injection 1 d before withdrawal of DOX chow and continued until 1 d before harvest of the tissue. The solvent for rapamycin is 0.2% sodium carboxymethyl cellulose, 0.25% Tween 80, and 2% dimethylsulfoxide in water.

All animals described in this report were maintained and used in accordance with protocols approved by the Institutional Animal Care and Use Committee of the University of Utah.

Immunoprecipitation and Western blotting

Mice were fasted for 6 h before they were anesthetized with chloral hydrate (1 mg/g body weight) by ip injection. For insulin stimulation, human insulin (0.1 U/30 g body weight) was injected through the inferior vena cava (IVC) after anesthesia. Heart muscle was dissected and immediately frozen in liquid nitrogen 5 min after insulin injection.

Proteins were extracted from heart tissue with homogenization buffer containing 0.1% Triton X-100 with a protease and phosphatase inhibitor cocktail (Pierce, Rockford, IL). Protein concentration was then determined using the Micro BCA Protein Assay kit (Pierce).

For IRS1 immunoprecipitation, 0.5 mg of heart lysates were incubated with rabbit polyclonal antibodies against IRS1 for 3 h at 4 C on a rocker. Then protein A-Sepharose was added and incubated for 1 h at 4 C with gentle rocking. After being washed three times with buffer, the immunocomplexes were dissociated by boiling with 2× Novex Tris-glycine sodium dodecyl sulfate sample loading buffer (Invitrogen, Carlsbad, CA) at 95 C for 5 min.

For Western blotting, identical amounts of total protein in equivalent volumes were loaded and resolved by SDS-PAGE and transferred to PVDF (low fluorescence) or nitrocellulose membrane for immunoblot detection with specific antibodies. Detection was performed by measuring intensity of fluorescence from secondary antibodies using the Odyssey Infrared Imaging System, and quantification was done with their accompanying software (version 3.0; LI-COR Biosciences, Lincoln, NE).

Primary antibody list: actin and tubulin antibodies were purchased from Sigma (St. Louis, MO); IRS1 and IRS2 antibodies were purchased from Upstate (now Millipore, Billerica, MA); phospho-tyrosine, IRβ antibody, and pIRS1 Ser312 antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA); and all other antibodies were purchased from Cell Signaling (Danvers, MA).

Secondary antibody list: goat antirabbit 800 and goat antimouse 800 were purchased from LI-COR. Alexa Flour 488, Alexa Flour 597, and Alexa Flour 680 antibodies were purchased from Invitrogen.

Cardiomyocyte isolation

Mice were anesthetized and heparinized, then the hearts were cannulated and perfused with buffer containing 16 U/ml type IV collagenase for 10–12 min. The heart was then minced, and myocytes were dissociated by gentle pipetting. The cells were pelleted by centrifugation at 600 rpm and then resuspended in culture medium (21).

Insulin-stimulated 2-deoxyglucose (2-DG) uptake

Insulin-stimulated 2-DG uptake assays were performed in triplicate in 12-well culture plates as described before (21). In brief, isolated cardiomyocytes were attached to laminin-coated plates. Attached cardiomyocytes were then conditioned in modified DMEM low-glucose medium for 30 min before being fasted for 40 min in DMEM no glucose medium supplemented with 1 mg/ml BSA and 1 mm pyruvate. Glucose uptake was performed by adding 2-[1,2-³H]-deoxy-d-glucose in the presence of 0, 0.1, 1, and 10 nm insulin. Uptake is calculated by quantifying the radioactivity of 2-[1,2-³H]-deoxy-d-glucose-6-phosphate.

c-myc GLUT4 staining in nonpermeabilized cardiomyocytes

Cardiomyocytes were isolated and attached to laminin-coated chamber slides. Attached cells were conditioned and fasted as described for 2-DG uptake and were then treated with 10 nm insulin or saline for 15 min before being fixed with cold 4% paraformaldehyde in PBS. Fixed cells were washed and incubated with 100 mm glycine for 10 min. Then cells were blocked with 5% BSA in PBS for 15 min on ice followed by overnight incubation with primary antibody in a moist box at 4 C. The next day, cells were washed three times with PBS and then incubated with secondary antibody for 1 h at room temperature. Then cells were washed again three times with PBS and mounted with Prolong Gold Antifade with 4′,6-diamidino-2-phenylindole reagent (Invitrogen). The slides were imaged with an Olympus FV1000 confocal microscope (Olympus, Center Valley, PA).

Hexokinase (HK) activity

HK activity was determined by modifying the protocol provided by Worthington Biochemical Corp. (Lakewood, NJ) and previous publications (22, 23). The assay is based on fluorometric measurement of nicotinamide adenine dinucleotide⁺ in a coupled reaction with glucose-6-phosphate dehydrogenase. The HK assay buffer consisted of 50 mm Tris-HCl (pH 8.0), 10 mm MgCl₂, 111.7 mm glucose, 0.55 mm ATP, 0.267 mm nicotinamide adenine dinucleotide, and 1 unit/ml glucose-6-phosphate dehydrogenase (Worthington Biochemical Corp.); 15 μl of heart tissue lysates (∼120 μg) were added to each reaction, and the increase in absorbance at A340 was measured for 4 min. The initial linear portion of the standard curve was used for calculating the enzymatic activity. The difference in temperature sensitivities was used for the measurement of HK-I and HK-II activity (24). One aliquot of 15 μl heart tissue lysates was assayed to assess total HK activity, i.e. HK-I and HK-II. The second aliquot was heated for 1 h at 45 C and then assayed to assess “heat-stable” HK activity, which is predominantly HK-I activity. HK-II activity is calculated by subtracting HK-I activity from total HK activity.

Glycolysis measured by working hearts

Hearts were rapidly excised, cannulated, and perfused in the working mode at a constant pressure of 60 mm Hg with Krebs buffer containing 5 mm glucose and 0.4 mm palmitate prebound to 3% BSA at 37 C saturated with 95% O₂ and 5% CO₂ as previously described (21). Wild-type (WT) and caPI3K hearts were perfused in the presence or absence of 1 nm insulin. Glycolysis was determined by measuring the amount of ³H₂O released from the metabolism of [5-³H] glucose (PerkinElmer, Waltham, MA) as previously described (21).

Determination of intracellular free glucose and glucose-6-phosphate

Gas chromatography-mass spectrometry (GC-MS) was used to measure free glucose and glucose-6-phosphate content of snap-frozen heart tissue isolated from WT and myrAkt1 transgenic mice, 14 d after DOX withdrawal, using previously described protocols (25) at the University of Utah Metabolomics Core Facility. In detail, heart tissue was homogenized and mixed with nine times volume of −20 C 90% methanol to achieve a final concentration of 80% methanol. The samples were incubated for 1 h at −20 C followed by centrifugation at 30,000 × g for 10 min to precipitate and remove proteins. The supernatant containing the extracted metabolites was then transferred to fresh disposable tubes and completely dried in vacuo.

GC-MS analysis of dried samples was performed with a Waters GCT Premier mass spectrometer (Waters, Milford, MA) fitted with an Agilent 6890 gas chromatograph (Agilent Technologies, San Jose, CA) and a GERSTEL MPS2 autosampler (GERSTEL, Linthicum, MD). Data were collected using MassLynx 4.1 software (Waters). Data analysis for glucose and glucose-6-phosphate was performed using QuanLynx, which identified the analytes and their peak area.

Statistical analysis

All values are shown as mean ± se. Datasets with two groups were compared with Student's t-test. Datasets with more than two groups were examined by one-way ANOVA followed by Bonferroni's multiple comparison test as a post hoc test. A statistic P value < 0.05 was considered significant. Student's t-test was performed using Microsoft EXCEL; ANOVA tests were performed using GraphPad Prism (GraphPad, San Diego, CA). Plots were drawn using GraphPad Prism.

Results

Expression of myrAkt1 impairs insulin-stimulated 2-DG uptake in isolated cardiomyocytes

Inducible Akt activation in the heart was achieved in transgenic mice expressing a myrAkt1 transgene fused to a 'tet'O operator sequence, which is governed by cardiomyocyte-restricted expression of a tTA protein (Tet-Off) (26). In the absence of DOX (a more stable tetracycline analog), tTA binds to the 'tet'O operator and induces target gene expression, which will normally be suppressed in the presence of DOX. Ten days after replacing DOX chow with normal chow, hemagglutinin (HA) tag was detected in myrAkt1 hearts, indicating expression of the transgene (Fig. 1A). Total Akt increased by 31% in myrAkt1 hearts compared with WT hearts and basal Akt Ser473 and Thr308 phosphorylation increased by 29.2 and 36.3%, respectively, in myrAkt1 hearts (Fig. 1A and Supplemental Figs. 1 and 2A, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org). Administration of insulin via the IVC markedly increased Akt Ser473 and Thr308 phosphorylation (>30-fold) in both WT and myrAkt1 hearts, and there was no difference in maximal insulin-stimulated Akt phosphorylation between WT and myrAkt1 groups (Fig. 1A and Supplemental Fig. 1). Basal S6 phosphorylation on Ser235/236 was also increased (Fig. 1A and Supplemental Fig. 1). Consistent with the increase in basal Akt phosphorylation, heart weight was increased by 8.7% when normalized to body weight and was further increased by 11.4% after 14 d of withdrawal from DOX (Supplemental Fig. 2B), which is also consistent with the further increase in HA levels, indicating increasing transgene expression (Supplemental Fig. 2C). myrAkt1 mice maintained on DOX chow exhibited basal levels of Akt Ser473 phosphorylation that were comparable with WT mice, indicating the absence of leaky transgene expression (Supplemental Fig. 2D) (27).

Fig. 1.

Short-term expression of myrAkt1 impairs insulin-stimulated 2-DG uptake in isolated cardiomyocytes. A, Expression of HA-tagged myrAkt1 transgene; basal and insulin-stimulated Akt Ser473 and Thr308 phosphorylation and S6 Ser235/236 phosphorylation in myrAkt1 hearts after 10-d withdrawal of DOX. B. Insulin-stimulated tyrosine phosphorylation of IRS1 protein in myrAtk1 hearts. Quantification is shown below the representative blot; AU, arbitrary unit; n = 3. C, Phosphorylation of IRS1 on multiple serine residues and total IRS1 and IRS2 protein levels in myrAkt1 hearts. D, Insulin-stimulated 2-DG uptake in isolated myrAkt1 cardiomyocytes; n = 4–5. *, P < 0.05; **, P < 0.01.

Despite a constitutive increase in the activation of the mTORC1 target S6 in myrAkt1 hearts, the ability of insulin to increase tyrosine phosphorylation of IRS1 protein was normal in myrAkt1 hearts (Fig. 1B). Furthermore, IRS1 serine phosphorylation at multiple sites, including the S6K1 target site Ser1101 (28), was not increased in myrAkt1 hearts (Fig. 1C and Supplemental Fig. 3), suggesting that the mTORC1-S6K1-IRS negative feedback pathway was not activated by the level of transgene induction achieved after 10 d of DOX withdrawal. Also, IRS1 and IRS2 protein levels were maintained in these myrAkt1 hearts (Fig. 1C). Acute insulin exposure did not increase IRS1 serine phosphorylation, which is consistent with the notion that IRS1 serine phosphorylation is a slow process compared with IRS1 tyrosine phosphorylation (Fig. 1C and Supplemental Fig. 3).

Despite a modest increase in basal Akt phosphorylation, basal glucose uptake was not increased in cardiomyocytes isolated from myrAkt1 hearts. Remarkably, and despite robust activation of Akt phosphorylation by insulin, insulin-stimulated 2-DG uptake was blunted in myrAkt1 cardiomyocytes. Relative to identically treated WT cardiomyocytes, 2-DG uptake was reduced by 80% in myrAkt1 cardiomyocytes after stimulation with 10 nm insulin (Fig. 1D). It is well established that acute activation of Akt will stimulate GLUT4 translocation and increase glucose uptake (29). However, the gradual decline in systemic levels of DOX by chow removal may render it difficult to detect an early increase in basal glucose uptake in our mouse model. Taken together, these data suggest that moderate and short-term activation of Akt1 impairs insulin-stimulated 2-DG uptake by a mechanism that may be independent of mTORC1-S6K1 inhibitory feedback signaling to IRS proteins.

Expression of a caPI3K impairs insulin-stimulated 2-DG uptake

We next examined an alternative mouse model with cardiac-restricted expression of a caPI3K. caPI3K mice had a 26% increase of heart weight when normalized to body weight at the age of 5 wk. Similar to myrAkt1 hearts, basal Akt phosphorylation on Ser473 and Thr308 sites in caPI3K hearts was increased by 4.4- and 3.3-fold, respectively, and was further increased after in vivo insulin administration to a level comparable with that of insulin-stimulated WT hearts (Fig. 2A). Basal S6K1 and S6 phosphorylation was also increased by 146.0 and 65.8%, respectively, but insulin further stimulated S6K1 in caPI3K hearts to levels that were equivalent to that of similarly treated WT hearts. Basal glycogen synthase kinase 3-β (Gsk3β) phosphorylation was not increased in caPI3K hearts but was stimulated by insulin (Fig. 2B).

Fig. 2.

Short-term expression of a caPI3K does not impair insulin-stimulated Akt and downstream Gsk3β and S6K1 phosphorylation. A, Basal and insulin-stimulated Akt phosphorylation on Ser473 and Thr308 sites in caPI3K hearts. Densitometric quantification is shown below the blots. B, Basal and insulin-stimulated phosphorylation of Akt downstream targets and densitometric analysis as indicated. AU, Arbitrary unit. *, P < 0.05; **, P < 0.01.

In 5-wk-old caPI3K hearts, the content of IRS1 and IRS2 proteins was unchanged (Fig. 3A). However, despite an increase in basal PI3K-Akt signaling and a further increase in the phosphorylation of Akt and its downstream targets after insulin stimulation, cardiomyocytes isolated from 5-wk-old caPI3K mice did not exhibit increased basal 2-DG uptake, and insulin-stimulated 2-DG uptake was blunted. Relative to identically treated WT cardiomyocytes, 2-DG uptake was reduced by 47.5% in caPI3K cardiomyocytes after stimulation with 10 nm insulin (Fig. 3B). A similar pattern was observed when glycolysis rates were measured in ex vivo isolated working hearts. In WT hearts, 1 nm insulin increased glycolysis by 73.9%. However, this increase in insulin-stimulated glycolysis was absent in caPI3K hearts (Fig. 3C), confirming the impaired insulin-stimulated glucose uptake observed in caPI3K cardiomyocytes. Thus, activation of PI3K in the heart also impairs insulin-stimulated 2-DG uptake via mechanisms that may be independent of mTORC1-S6K1 inhibitory feedback signaling to IRS proteins.

Fig. 3.

Short-term expression of a caPI3K impairs insulin-stimulated 2-DG uptake in the absence of any change in IRS protein content. A, IRS1 and IRS2 protein levels in 5-wk-old caPI3K hearts. Quantification is shown below the blots. AU, Arbitrary unit. B, Insulin-stimulated 2-DG uptake in isolated caPI3K cardiomyocytes; n = 5. C, Basal and insulin-stimulated glycolysis in control and caPI3K hearts perfused with saline or 1 nm insulin, determined ex vivo in isolated working hearts; n = 4. *, P < 0.05.

Prolonged myrAkt1 expression decreases IRS2 protein and rapamycin treatment normalizes IRS2 protein content but fails to restore insulin-stimulated 2-DG uptake in myrAkt1 cardiomyocytes

The data presented thus far suggest that short-term activation of PI3K or Akt in the heart desensitizes insulin-mediated glucose uptake independently of inhibitory feedback by mTORC1-S6K1 to IRS proteins. To explore whether longer-term activation of this signaling pathway will eventually reduce IRS signaling, we examined 8-wk-old caPI3K mice and observed 48.8 and 48.6% reduction in IRS1 and IRS2 proteins, respectively (Fig. 4A), and the ability of insulin to increase Akt phosphorylation was blunted (Fig. 4B). Similar changes were observed in mice in which the myrAkt1 transgene was activated for 14 d, in that IRS2 protein was reduced by 68%, although IRS1 protein levels were not changed (Fig. 4C). Treatment of 14-d induced myrAkt1 mice with rapamycin normalized IRS2 protein content (Fig. 4C). Thus, longer-term activation of PI3K and Akt will eventually reduce IRS signaling in the heart in an mTORC1-dependent manner.

Fig. 4.

Chronic PI3K-Akt activation of longer duration reduces IRS protein through a mechanism involving mTORC1. A, IRS1 and IRS2 protein levels in 8-wk-old caPI3K hearts. Quantification is shown on the right. B, Insulin-stimulated Akt phosphorylation in 8-wk-old caPI3K hearts. Quantification is shown on the right. C, Fourteen-day induction of myrAkt1 reduces IRS2 protein levels, which was normalized by rapamycin treatment. Quantification is shown below the blots. AU, Arbitrary unit. *, P < 0.05; **, P < 0.01.

Given our earlier findings that Akt activation impaired glucose uptake before any reduction in IRS proteins, we reasoned that restoration of IRS proteins with rapamycin treatment would not restore insulin-mediated glucose uptake in mice with longer-term activation of Akt. As shown in Fig. 5A, rapamycin completely abolished S6K and S6 phosphorylation in both WT and myrAkt1 hearts but also decreased Akt Ser473 phosphorylation via inhibition of mTORC2, a kinase complex that directly phosphorylates Akt on Ser473 (Fig. 5A and Supplemental Fig. 4) (30). Although attenuated, insulin increased Akt Thr308 phosphorylation in WT hearts treated with rapamycin and to a lesser extent in rapamycin-treated myrAkt1 hearts. A similar pattern was observed with Gsk3β phosphorylation (Fig. 5A and Supplemental Fig. 4). Rapamycin treatment did not inhibit insulin-stimulated glucose uptake in WT cardiomyocytes but failed to restore insulin-mediated glucose uptake in myrAkt1 cardiomyocytes (Fig. 5B). Moreover, treatment of 5-wk-old caPI3K mice or myrAkt1 mice after 10 d of DOX withdrawal with rapamycin also abolished the increase in S6 phosphorylation but did not restore glucose uptake (Supplemental Figs. 5 and 6). Thus, inhibition of mechanistic target of rapamycin (mTOR) signaling in mice with constitutive activation of PI3K or Akt does not restore insulin-mediated glucose uptake. Although the inhibitory impact of rapamycin on insulin-mediated Akt signaling could be a basis for the lack of rescue with rapamycin, the fact that rapamycin does not impair glucose uptake in WT mice makes this possibility less likely. Taken together, these findings support the hypothesis that chronic Akt activation may inhibit glucose uptake via mTORC1-independent mechanisms.

Fig. 5.

Fourteen-day rapamycin treatment does not restore insulin-stimulated 2-DG uptake in myrAkt1 cardiomyocytes. A, The effect of rapamycin treatment on basal and insulin-stimulated phosphorylation of Akt and its downstream kinases in myrAkt1 hearts after 14 d of DOX withdrawal. B, The effect of rapamycin treatment on insulin-stimulated 2-DG uptake in WT and myrAkt1 cardiomyocytes; n = 4–8. a, P < 0.05 vs. WT vehicle (Vehi); b, P < 0.05 vs. WT rapamycin (Rapa).

Impairment of 2-DG uptake in myrAkt1 cardiomyocytes occurs despite normal GLUT4 translocation

GLUT1 and GLUT4 are two major glucose transporters in cardiomyocytes. Decreased 2-DG uptake could result from reduction in GLUT1 or GLUT4 proteins. Expression of myrAkt1 reduced GLUT4 protein by 16%, whereas GLUT1 expression was increased by 62% after 10 d of DOX withdrawal (Fig. 6A). Those effects seem, at least in part, to be regulated at transcriptional levels as evidenced by a 2-fold rise in GLUT1 mRNA and a 40% drop in GLUT4 mRNA in a microarray (our unpublished data). However, a 16% reduction of GLUT4 protein cannot account for an 80% reduction of maximal insulin-stimulated 2-DG uptake. Also, rapamycin treatment restored the GLUT4 content (Fig. 6A) but failed to restore insulin-stimulated 2-DG uptake, confirming that the change in GLUT4 protein cannot account for impaired 2-DG uptake in myrAkt1 cardiomyocytes.

Fig. 6.

GLUT4 and GLUT1 levels and AS160 phosphorylation in myrAkt1 hearts. A, GLUT1 and GLUT4 protein levels in WT and myrAkt1 hearts treated with or without rapamycin. Quantification is shown below the blots. AU, Arbitrary unit. B, Insulin-stimulated AS160 phosphorylation in WT and myrAkt1 hearts. Quantification is shown on the right. *, P < 0.05; **, P < 0.01.

Phosphorylation of AS160 was also increased in myrAkt1 hearts (Fig. 6B), implying that GSV translocation might not be impaired by myrAkt1 expression. To determine whether GLUT4 protein was normally translocated and inserted into the PM, immunofluorescence detection of c-myc GLUT4 in nonpermeabilized cardiomyocytes was used. For these studies, transgenic mice that harbor a DOX-regulated c-myc GLUT4 fusion transgene in which c-myc is fused to the exofacial loop of GLUT4 (20) were generated. c-myc GLUT4 mice were crossed to myrAkt1 double transgenic mice to generate inducible myrAkt1 and c-myc GLUT4 coexpressing mice (triple transgenic mice). In the absence of insulin stimulation, negligible levels of c-myc GLUT4 fluorescence were detected in nonpermeabilized c-myc GLUT4 cardiomyocytes. After stimulation with 10 nm insulin, the c-myc GLUT4 fluorescence was greatly increased, confirming insulin-stimulated GLUT4 translocation in WT cardiomyocytes. Overexpression of myrAkt1 in cardiomyocytes increased c-myc GLUT4 fluorescence even in the absence of insulin, indicating that overexpression of myrAkt1 was sufficient to target GLUT4 protein to the cell surface. The intensity of fluorescence could not be further increased after insulin stimulation (Fig. 7). Taken together with the observation that basal 2-DG glucose uptake is not increased in myrAkt1 cardiomyocytes, the increase of basal c-myc fluorescence suggests that GLUT4 intrinsic activity or glucose phosphorylation is significantly impaired in myrAkt1 cardiomyocytes.

Fig. 7.

c-myc immunofluorescence staining in nonpermeabilized c-myc GLUT4 and myrAkt1 coexpressing cardiomyocytes. c-myc immunofluorescence staining in nonpermeabilized cardiomyocytes from c-myc GLUT4 heart (control) and c-myc GLUT4/myrAkt1 coexpressing hearts (triple transgenic) treated with saline or insulin for 15 min. Line profiles for green fluorescence were shown adjacent to each picture. Scale bar, 10 μm. The image is representative of three independent experiments in three sets of mice.

After being transported across the cell membrane, glucose is rapidly phosphorylated by HK to glucose-6-phosphate, before entering the glycolytic pathway. The abundance of HK-II protein, the predominant form of HK in the heart, was the same in WT and myrAkt1 hearts (Fig. 8A), and the activity of HK-II was not reduced in myrAkt1 hearts when treated with saline or insulin compared with WT hearts (Fig. 8B). Intracellular free glucose was increased by 59% upon insulin stimulation in WT hearts but declined by 31% in myrAkt1 hearts, which could reflect increased conversion to glucose-6-phosphate or reduced insulin-stimulated glucose transport. Glucose-6-phosphate was increased by 4-fold in WT hearts treated with insulin but was increased by only 35% in myrAkt1 hearts. Taken together with the reduction in free glucose, this limited increase in glucose-6-phosphate is likely a result of reduced substrate (free glucose) available for phosphorylation in myrAkt1 hearts vs. reduced HK-II activity. Importantly, the ratio of glucose to glucose-6-phosphate in myrAkt1 hearts was similar to that in control hearts (Fig. 8C), suggesting normal HK-II function in myrAkt1 hearts. Thus, the defect of insulin-stimulated 2-DG uptake in myrAkt1 cardiomyocytes likely resulted from a defect in GLUT4 activation. p38MAPK activity has been implicated in GLUT4 activation after PM translocation (31). However, relative to controls, no impairment in p38MAPK phosphorylation was observed in myrAkt1 transgenic hearts (Supplemental Fig. 7).

Fig. 8.

HK-II activity is not impaired in myrAkt1 hearts. A, HK-II protein levels measured by Western blotting, quantification is shown below the blot. B, HK-II activity in WT and myrAkt1 hearts isolated 5 min after saline or insulin injection via the IVC; n = 4. C. Free glucose, glucose-6-phosphate, and glucose/glucose-6-phosphate (G/G6P) ratio measured by GC-MS in WT and myrAkt1 hearts isolated 5 min after saline or insulin injection via the IVC. Data are expressed as fold change relative to saline-treated samples; n = 8. *, P < 0.05; **, P < 0.01.

Discussion

There is an emerging consensus that many pathways conspire to induce insulin resistance in muscle and adipose tissue. Although impaired insulin-stimulated glucose uptake is a consistent feature, this may occur in the absence of a coordinate reduction in insulin-mediated IRS1 phosphorylation or Akt activation (32). Recent studies in cardiac muscle have also suggested that in models of insulin resistance, persistent activation of Akt signaling remains, despite a reduction in glucose uptake (17). In some of these studies, the impairment in glucose uptake was attributed to defects in GLUT4 translocation or to a reduction in GLUT4 protein or both, via mechanisms that are incompletely understood (17, 33). Given that persistent activation of PI3K and Akt in the heart may be a consequence of hyperinsulinemia in insulin-resistant states, the present study was designed to determine whether constitutive activation of PI3K and Akt signaling in the heart could independently impair insulin-mediated glucose uptake and GLUT4 translocation. The study was predicated on the hypothesis that chronic activation of PI3K and Akt signaling would desensitize insulin-mediated glucose uptake via mechanisms that were secondary to mTORC1-mediated activation of S6K, which would desensitize IRS signaling via serine phosphorylation of IRS proteins and IRS protein degradation (Fig. 9). Unexpectedly, the present study indicates that in cardiomyocytes, PI3K and Akt activation may impair myocardial glucose uptake before any evidence of S6K-mediated inhibition of IRS1/2. Moreover, the defect in glucose uptake occurs despite normal translocation of GLUT4 and preservation of HK activity. These data indicate that chronic activation of PI3K and Akt may impair glucose uptake via mechanisms that may impair the intrinsic activity of the GLUT4 transporter (Fig. 9).

Fig. 9.

Insulin signaling and glucose transport in normal hearts and in insulin resistance and type 2 diabetes. In a healthy heart, insulin activates PI3K-Akt signaling and promotes GLUT4 translocation to the cell surface to facilitate glucose transport. Insulin resistance and type 2 diabetes is characterized by hyperinsulinemia. A persistent increase in circulating insulin activates PI3K-Akt leading to mTOR-S6K negative feedback signaling to IRS1/2. Activation of PI3K-Akt signaling in the heart rapidly impairs insulin-stimulated glucose uptake before inactivation of IRS by this negative feedback pathway. Although GLUT4 translocation may be maintained or increased, glucose uptake is decreased, suggesting the existence of a GLUT4 postinsertion activation mechanism, which is impaired in hearts with chronic activation of PI3K and Akt.

To evaluate the possibility that impaired glucose uptake could be a consequence of a defect in GLUT4 translocation, we used a myc-epitope-tagged GLUT4 construct, in which myc is inserted in the exofacial loop of GLUT4. Activation of Akt was associated with phosphorylation of AS160 and clear evidence of gain in surface GLUT4 in the absence of insulin, to a level equivalent to that observed in insulin-stimulated control cells. Treatment with insulin did not further increase surface GLUT4. Yet despite clear evidence of Akt-mediated GLUT4 translocation and increased cardiomyocyte GLUT1 content, we obtained no evidence for augmented basal or insulin-stimulated glucose uptake. This could involve either lack of activation of GLUT4 (or GLUT1) (34) and/or inhibition of HK. However, the possibility that expression of myrAkt1 inhibits HK was excluded (Fig. 8), leaving the lack of activation of GLUT4 (or GLUT1) the most plausible explanation for impaired 2-DG uptake.

In cardiomyocytes, the signals regulating GLUT4 traffic include PI3K, Akt, and Ca2⁺ (11). Although much progress has been made in elucidating many of the signaling pathways activated by insulin that mediate GLUT4 translocation and the molecular mechanisms governing docking and fusion of GLUT4-containing vesicles (35), far less is known about the regulation of the specific activity of GLUT4 after its translocation and insertion into the PM. Previous studies in cultured adipocytes and muscle cells suggested that activation of GLUT4 may play an important role in glucose transport after transporter insertion (31, 34). Interestingly, although direct delivery of PIP3 to adipocytes promoted GLUT4 translocation in cultured adipocytes and L6 myocytes, there was no associated increase in glucose uptake, suggesting the existence of PI3K-independent mechanisms that regulate the activation of GLUT4 transporters after insertion into the PM (36). The present study shows that although persistent activation of PI3K and Akt may lead to insertion of GLUT4 into the sarcolemmal membrane of cardiomyocytes, GLUT4 activation is desensitized. Although previous studies have implicated p38MAPK in GLUT4 activation after PM translocation (31), relative to controls, no impairment in p38MAPK phosphorylation was observed in myrAkt1 transgenic hearts, which argues against a role for this pathway in the Akt-mediated desensitization of glucose uptake. In addition to translocation, insulin also regulates the postfusion dispersal and spatial distribution of GLUT4 in the PM, which may correlate with the full activation of glucose uptake (37). Whether chronic activation of Akt1 in the heart leads to a defect of GLUT4 dispersion needs to be further studied.

In conclusion, the present study sheds additional mechanistic insight into the rapid impairment of insulin-mediated myocardial glucose uptake that characterizes insulin-resistant states. Using models of constitutive activation of PI3K and Akt signaling pathways, which mimic the impact of diabetes and obesity on myocardial insulin signaling, we show that Akt activation impairs glucose uptake via mechanisms that impair the activation of GLUT4 transporters independent of activation of mTORC1.