|Home | About | Journals | Submit | Contact Us | Français|
Glucocorticoid excess is characterized by increased adiposity, skeletal myopathy, and insulin resistance, but the precise molecular mechanisms are unknown. Within skeletal muscle, 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) converts cortisone (11-dehydrocorticosterone in rodents) to active cortisol (corticosterone in rodents). We aimed to determine the mechanisms underpinning glucocorticoid-induced insulin resistance in skeletal muscle and indentify how 11β-HSD1 inhibitors improve insulin sensitivity.
Rodent and human cell cultures, whole-tissue explants, and animal models were used to determine the impact of glucocorticoids and selective 11β-HSD1 inhibition upon insulin signaling and action.
Dexamethasone decreased insulin-stimulated glucose uptake, decreased IRS1 mRNA and protein expression, and increased inactivating pSer307 insulin receptor substrate (IRS)-1. 11β-HSD1 activity and expression were observed in human and rodent myotubes and muscle explants. Activity was predominantly oxo-reductase, generating active glucocorticoid. A1 (selective 11β-HSD1 inhibitor) abolished enzyme activity and blocked the increase in pSer307 IRS1 and reduction in total IRS1 protein after treatment with 11DHC but not corticosterone. In C57Bl6/J mice, the selective 11β-HSD1 inhibitor, A2, decreased fasting blood glucose levels and improved insulin sensitivity. In KK mice treated with A2, skeletal muscle pSer307 IRS1 decreased and pThr308 Akt/PKB increased. In addition, A2 decreased both lipogenic and lipolytic gene expression.
Prereceptor facilitation of glucocorticoid action via 11β-HSD1 increases pSer307 IRS1 and may be crucial in mediating insulin resistance in skeletal muscle. Selective 11β-HSD1 inhibition decreases pSer307 IRS1, increases pThr308 Akt/PKB, and decreases lipogenic and lipolytic gene expression that may represent an important mechanism underpinning their insulin-sensitizing action.
The pathophysiological effects of glucocorticoids are well described and impact upon almost all organ systems within the body. This is highlighted in patients with glucocorticoids excess, Cushing's syndrome characterized by central obesity, hypertension, proximal myopathy, insulin resistance, and in some cases overt type 2 diabetes. In addition, up to 2.5% of the population are taking prescribed glucocorticoids (1), and their side effects represent a considerable clinical burden for both patient and clinician.
Glucocorticoids induce whole-body insulin resistance (2); however, the precise molecular mechanisms that underpin this observation have not been defined in detail. In simple obesity and insulin resistance, circulating cortisol levels are not elevated (3), but in key insulin target tissues including liver, fat, and muscle, glucocorticoid availability to bind and activate the glucocorticoid receptor is controlled by 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1). 11β-HSD1 is an endo-lumenal enzyme that interconverts inactive (cortisone in humans and 11-dehydrocorticosterone [11DHC] in rodents) and active glucocorticoids (cortisol in humans and corticosterone [CORT] in rodents) (4). Critically, the directionality of 11β-HSD1 activity is cofactor (NADPH) dependent that is supplied by a tightly associated endo-lumenal enzyme, hexose-6-phopshate dehydrogenase (H6PDH). Decreases in H6PDH expression and activity decrease 11β-HSD1 oxo-reductase and increase dehydrogenase activity (5). Despite this bidirectional potential, the predominant direction of activity in liver, adipose, and muscle is oxo-reductase generating active glucocorticoid (cortisol and CORT), therefore, amplifying local glucocorticoid action.
Binding of insulin to its cell surface receptor leads to a conformational change and tyrosine autophosphorylation. Consequently, the insulin receptor substrate (IRS) family of adaptor proteins are recruited to the intracellular domain of the receptor and are phosphorylated at multiple tyrosine residues by the receptor tyrosine kinase to permit the docking of phosphatidylinositol-3-kinase (PI3K) and subsequent generation of PI(3,4,5)P3. Generation of this second messenger acts to recruit the Akt/PKB family of serine/threonine kinases to the plasma membrane where they are then activated (6). Further downstream, activated Akt1/protein kinase B (PKB) phosphorylates a rab-GAP (GTPase) protein, AS160, which is a crucial regulator of the translocation of GLUT4 GLUT storage vesicles to the plasma membrane (7). It is this mechanism that permits insulin-stimulated glucose entry into target tissues including skeletal muscle (8).
The molecular mechanisms underpinning insulin resistance are complex and variable. Serine/threonine phosphorylation of IRS1 (in particular Ser307 phosphorylation) has been shown to negatively regulate insulin signaling through multiple mechanisms including decreased affinity for the insulin receptor and increased degradation (9,10).
The interaction of glucocorticoids and the insulin signaling cascade has only been examined in a small number of studies that have offered variable explanations for the induction of insulin resistance (11–14). Importantly, the role of serine phosphorylation and the impact of prereceptor glucocorticoid metabolism have not been explored. The 11β-HSD1 knockout mouse is relatively insulin sensitive (15), and specific inhibitors of 11β-HSD1 improve lipid profiles, glucose tolerance, and insulin sensitivity and have considerable potential as therapeutic agents (16–18). However, the molecular mechanisms that underpin these observations remain to be defined.
Therefore, we have characterized the impact of glucocorticoids upon the insulin-signaling cascade and analyzed the expression, activity, and functional impact of 11β-HSD1 in vitro and using in vivo mouse models.
Mouse skeletal muscle cell line, C2C12 myoblasts (ECACC, U.K.), were grown in DMEM (PAA, U.K.) supplemented with 10% FBS (37°C, 5% CO2). Cells were grown to 60–70% confluence before differentiation (initiated by replacing growth media with DMEM with 5% horse serum). After 8 days, myoblasts had fused to form multinucleated myotubes. Before treatment, cells were washed out with serum-free media for 4 h. For experiments examining Ser24 phosphorylation, C2C12 cells stably overexpressing myc-rIRS1 were generated and used as described previously for 3T3-L1 adipocytes (19).
Primary human myoblasts were obtained from PromoCell (Heidelberg, Germany). Myoblasts were cultured to confluence, as per the manufacturer's guidelines using the supplied media. Once confluent, media was changed to a chemically defined media (PromoCell, Germany) including 2% horse serum and cells differentiated into myotubes for 11 days. After differentiation, cells were incubated with serum-free media for 4 h before treatment.
Unless otherwise stated, for all cell culture experiments investigating insulin-signaling cascade protein phosphorylation, media was spiked with human insulin (0.1 μg/ml, Sigma, U.K.) for the final 15 min of the treatment period to achieve insulin stimulation. In experiments using the glucocorticoid receptor antagonist, RU38486, cells were pretreated with RU38486 (10 μmol/l) for 10 min before adding Dexamethasone (Dex). Treatments and reagents were supplied by Sigma, Poole, U.K. unless otherwise stated. Selective 11β-HSD1 inhibitors (A1 and A2, >95% and >99% purity, respectively) were provided through material transfer agreements with AstraZeneca (Macclesfield, U.K.). Inhibitor properties are presented within the results section.
Total RNA was extracted from cell lysates using the Tri-Reagent system and from whole-tissue explants using RNeasy Fibrous Tissue Mini Kit (Qiagen). Integrity, concentration, and reverse transcription were performed as we have previously described (20). Specific mRNA levels were determined using an ABI 7500 sequence detection system (Perkin-Elmer Applied Biosystems, Warrington, U.K.). Reactions were performed in 20 μl volumes on 96-well plates in reaction buffer containing 2 × TaqMan Universal PCR Master mix (Applied Biosystems, Foster City, CA). Probes and primers for all genes were supplied by applied biosystems ‘assay on demand’ (Applied Biosystems). All reactions were normalized against 18S rRNA as an internal housekeeping gene.
Data were obtained as ct values (ct = cycle number at which logarithmic PCR plots cross a calculated threshold line) and used to determine Δct values with Δct = (ct of the target gene) – (ct of the housekeeping gene). Data are expressed as arbitrary units using the following transformation (expression = 105 × [2−Δct] arbitrary units [AUs]). When used, fold changes were calculated using the following equation: [fold increase = 2−difference in Δct].
Taqman 348-well custom arrays were purchased from Applied Biosystems containing 45 custom genes and 3 housekeeping genes. Five hundred nanograms of cDNA were mixed with Taqman universal PCR master mix (Applied Biosystems), and the array was run on an ABI 7900HT Fast Real-Time PCR System (Applied Biosystems). Data were obtained as ct values and fold changes calculated. Results were validated with standard Taqman RT-PCR.
Cells were scraped into 100 μl RIPA buffer (50 mmol/l Tris pH 7.4, 1% NP40, 0.25% sodium deoxycholate, 150 mmol/l NaCl, 1 mmol/l EDTA), 1 mmol/l phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (Roche, Lewes, U.K.), incubated at −80°C (10 min) on ice (30 min), and centrifuged at 4°C (10 min, 14,000 rpm). The supernatant was transferred to a fresh tube and total protein concentration determined by a commercially available assay (Bio-Rad Laboratories, Hercules, CA).
Twenty micrograms of protein was resolved on an SDS-PAGE gel (acrylamide percentage varied according to protein size). Proteins were transferred onto nitrocellulose membrane, Hybond ECL (GE Healthcare, Chalfont St. Giles, U.K.). Primary (anti-IRS1, anti-pSer307 IRS1, anti-IRS2, and anti-AS160 from Upstate, Dundee, U.K.; anti-PKB/akt, anti-pSer473 PKB/akt [recognizing isoforms 1 and 2], and anti-Thr308 Akt/PKB from R&D Systems, Abingdon, U.K.; and anti-pTyr608 IRS1 from Biosource, Nivelles, Belgium) and secondary antibodies (Dako, Glostrop, U.K.) were used at a dilution of 1/5,000. Membranes were reprobed for β-actin and primary and secondary antibodies used at a dilution of 1/5,000 (Abcam plc, Cambridge, U.K.). Bands were visualized using ECL detection kit (GE Healthcare, Chalfont St. Giles, U.K.) and quantified with Genesnap by Syngene (Cambridge, U.K.).
Glucose uptake activity was analyzed by measuring the uptake of 2-deoxy-d-[3H] glucose as described previously (21). After treatment, cells were washed three times with Krebs-Ringer-Hepes (KRP) buffer and incubated with 0.9 ml KRP buffer at 37°C for 20 min. Insulin (0.5 μg/ml) was then added, and the cells were incubated at 37°C for 20 min. Glucose uptake was initiated by the addition of 0.1 ml KRP buffer and 37MBq/l 2-deoxy-d-[3H] glucose (GE Healthcare) and 7 mmol/l glucose as final concentrations. After 60 min, glucose uptake was terminated by washing the cells three times with cold PBS. Cells were lysed and radioactivity retained by the cell lysates determined by scintillation counting.
Briefly, intact cells were incubated with 100 nmol/l 11DHC and tritiated tracer for 2–24 h dependent upon assay system. In studies using selective 11β-HSD1 inhibitors, cells were incubated for 24 h with inhibitor before 11β-HSD1 assay being performed. Steroids were then extracted using dichloromethane, separated using a mobile phase consisting of ethanol and chloroform (8:92) by thin layer chromatography, and scanned using a Bioscan 3000 image analyzer (Lablogic, U.K.). Protein levels were assayed using a commercially available kit (Bio-Rad, Richmond, CA).
The selective 11β-HSD1 inhibitor A2 was used in two separate mouse protocols. All experimental procedures were conducted in accordance with the Animal Scientific Procedures Act 1986, Animal Welfare Act 2006, and local guidelines. First, male C57Bl6/J mice (6 weeks of age) were maintained for 20 weeks on a diet comprising 48 kcal% fat/10 kcal% fructose (Research Diets RD06072801, New Brunswick, NJ). Mice were housed with a standard light cycle (06:00 h on/18:00 h off), and A2 was administered (20 mg/kg b.i.d) by oral gavage for 28 days and compared against vehicle-control and pair-fed, vehicle-treated animals (n = 9–10). Food intake was assessed over the 28-day period, and on day 24 after 12-h fast, an oral glucose tolerance test (OGTT, 2 g/kg) was performed. Conscious tail-prick samples were analyzed for glucose using the Accu-Chek Aviva system (Roche) and for insulin using Ultra Sensitive Mouse Insulin ELISA kits (Crystal Chem #90800, Downers Grove, IL).
Separately, the impact of A2 upon skeletal muscle gene and protein expression in an additional hyperglycemic model were determined (male KK/Ta Jcl mice aged 7 weeks; CLEA Japan, Tokyo, Japan). Animals had free access to water and irradiated RM3 (E) diet composed of 11.5 kcal% fat, 27 kcal% protein, and 62 kcal% carbohydrate (Special Diets Services, Witham, U.K.). Compound A2 (20 mg/kg) or vehicle (10 ml/kg) was administered by oral gavage at 08:00 and 20:00 h for 4 consecutive days. After the administration of the third dose of A2, mice were anesthetized by inhalation of 2–3% v/v isoflurane (vaporized by oxygen at 1.6 l/min flow rate) and a 5 mg slow-release cortisone pellet (~8 mg · kg−1 · day−1 in a 29 g mouse) implanted subcutaneously in the lateral aspect of the neck (Innovative Research of America, Sarasota, FL). On day 4, 1–2 h after the seventh oral dose of A2, a rising-dose carbon dioxide concentration was used to humanely kill mice in the fed state and femoral quadriceps muscles were removed and snap frozen in liquid nitrogen.
Where data were normally distributed, unpaired Student's t tests were used to compare single treatments to control. If normality tests failed, then nonparametric tests were used. One-way ANOVA on ranks was used to compare multiple treatments, doses, or times (SigmaStat 3.1; Systat Software, Point Richmond, CA). Statistical analysis on real-time PCR data was performed on mean Δct values.
In agreement with published data, treatment with the synthetic glucocorticoid Dex (1 μmol/l, 24 h) decreased insulin-stimulated glucose uptake (2.1 ± 0.3 vs. 1.7 ± 0.2 dpm × 105, P < 0.05, n = 4) consistent with the induction of insulin resistance in differentiated rodent C2C12 skeletal myocytes.
Using real-time PCR, Dex (1 μmol/l, 24 h) decreased mRNA expression of IRS1 (8.7 ± 0.7 vs. 4.5 ± 0.4 AU, P < 0.05, n = 5), an effect that was blocked by the glucocorticoid antagonist RU-38486 (Table 1). In contrast, IR, Akt/PKB2, PI3K, AS160, and GLUT4 expression all increased after Dex treatment. RU38486 reversed only the effects of Dex upon GLUT4 and AS160 expression. Although Dex treatment did not alter 11β-HSD1 expression, H6PDH expression increased (0.10 ± 0.01 vs. 1.30 ± 0.07 AU, P < 0.001) and was reversed by coincubation with RU38486 (0.18 ± 0.02 AU, P < 0.001 vs. Dex). Absolute mRNA expression data after Dex treatment with and without the glucocorticoid receptor antagonist RU38486 are presented in Table 1.
At the protein level, treatment with Dex decreased IRS1 total protein expression (0.4-fold, P < 0.05) that was recovered by coincubation with RU38486 (Fig. 1A and B). Insulin-stimulated, activating Tyr608 phosphorylation of IRS1 was unchanged with Dex treatment (Fig. 1A and B). However, we observed enhanced inactivating Ser307 phosphorylation (3.3-fold, P < 0.05) after Dex treatment that was prevented by RU-38486 (Fig. 1A and B). Total IRS2 protein expression was increased by Dex (1.7-fold, P < 0.001) (Fig. 1A and C). Further downstream, Akt/PKB protein expression did not change with Dex treatment, but activating Ser473 phosphorylation decreased (0.5-fold, P < 0.05) (Fig. 1A and D). Dex treatment increased AS160 expression 1.5-fold (P < 0.05), which was blocked by coincubation with RU38486 (Fig. 1A and E).
Diacylglycerol (DAG)-dependent protein kinase C (PKC) isoforms have been implicated in the phosphorylation of IRS1 at Ser24, and this has been linked to the pathogenesis of insulin resistance (19). We examined the effect of Dex using C2C12 cells stably overexpressing IRS1. Although Dex increased Ser307 phosphorylation in this model, there was no detectable impact upon Ser24 phosphorylation or ectopic IRS1-myc protein levels (Fig. 1F).
To determine whether our observations with synthetic glucocorticoids were applicable in a physiological context, further experiments were performed using endogenous rodent glucocorticoids (11DHC and CORT). Consistent with our observations using Dex, CORT caused a dose- and time-dependent decrease in IRS1 total protein expression (dose: 1.0 [control] vs. 0.59-fold [100 nmol/l], P < 0.01, 0.47-fold [250 nmol/l], P < 0.05, 0.44-fold [500 nmol/l], P < 0.01, 0.38-fold [1,000 nmol/l], P < 0.01; time: 1.0 [control] vs. 0.19 ± 0.04 [48 h], P < 0.05) (Fig. 2A and B). This was accompanied by a dose- (1.0 [control] vs. 2.80-fold [250 nmol/l], P < 0.01, 3.99-fold [500 nmol/l], P < 0.01, 4.37-fold [1,000 nmol/l], P < 0.001) (Fig. 2A) and time- (1.0 [control] vs. 3.0 ± 0.08 [48 h], P < 0.05) (Fig. 2B) dependent increase in Ser307 phosphorylation.
11β-HSD1 mRNA was highly expressed in C2C12 cells. Expression was also detected in whole-tissue explants of mouse quadriceps muscle, although levels were lower than those seen in liver and adipose tissue (Table 2). In all systems (C2C12 cells, human primary cultures, and whole-tissue explants from mice), functional, bidirectional 11β-HSD1 activity was demonstrated with predominant oxo-reductase activity (Fig. 3A and B). In mouse quadriceps explants, oxo-reductase activity was significantly decreased after coincubation with the nonselective 11β-HSD inhibitor, glycyrrhetinic acid (2 μmol/l, 2 h) (114.0 ± 5.7 vs. 44.6 ± 11.1 pmol · g−1 · h−1, P < 0.05) (Fig. 3A and B). A1 and A2 are selective 11β-HSD1 inhibitors provided through an investigator-led collaboration with AstraZeneca. A1 has a half-maximal inhibitory concentration (IC50) for human recombinant 11β-HSD1 of 0.3 nmol/l and for rat 637 nmol/l, mouse 33 nmol/l, and human 11β-HSD2 >15 μmol/l. A2 has an IC50 for human recombinant 11β-HSD1 of 7 nmol/l and for rat 94 nmol/l, mouse 26 nmol/l, and human 11β-HSD2 >15 μmol/l. Treatment with A1 (1 μmol/l, 24 h), significantly decreased oxo-reductase activity in mouse quadriceps whole-tissue explants, differentiated C2C12 cells, and primary cultures of differentiated human skeletal myocytes (Fig. 3C).
Paralleling our observations with Dex and CORT, 11DHC (250 nmol/l, 24 h) decreased IRS1 total protein expression (0.5-fold, P < 0.05) and increased pSer307 IRS1 (2.0-fold, P < 0.05) in C2C12 cells. Coincubation with the nonselective 11β-HSD inhibitor glycyrrhetinic acid (2.5 μmol/l, 24 h) reversed 11DHC-induced pSer307 IRS1 to levels seen in control untreated cells (1.1-fold, P = 0.56 vs. control) (Fig. 4A). Glycyrrhetinic acid treatment alone was without effect (data not shown). Similarly, observations with the selective 11β-HSD1 inhibitor A1 (2.5 μmol/l, 24 h) mirrored those with glycyrrhetinic acid, completely blocking the effects of 11DHC to decrease total IRS1 expression and increase pSer307 IRS1 (Fig. 4B). Extending these findings, in primary cultures of human skeletal muscle, cortisone (250 nmol/l, 24 h) decreased insulin-stimulated pThr308 Akt/PKB without altering total Akt/PKB protein expression. These observations were completely abolished following coincubation with A1 (Fig. 4C).
Food intake decreased within the first 48 h in the A2-treated animals in comparison with vehicle-treated controls. However, by day 4 and for the remainder of the 28-day protocol, food intake did not differ between the groups (day 4: 15.4 ± 0.7 vs. 16.5 ± 0.7 kcal/day [A2 vs. control], P = 0.08). At day 28, fasting blood glucose and insulin levels were lower in the A2-treated animals compared with both vehicle-treated and pair-fed controls (glucose: 6.8 ± 0.3 vs. 7.4 ± 0.35 vs. 7.7 ± 0.3 mmol/l, P < 0.05; insulin: 0.60 ± 0.10 vs. 0.82 ± 0.14 vs. 0.91 ± 0.11 ng/ml, P < 0.05, A2 vs. vehicle vs. pair-fed vehicle). Similarly, homeostasis model assessment values were lower (4.2 ± 0.9 vs. 6.0 ± 0.98 vs. 7.1 ± 0.9 [A2 vs. vehicle vs. pair-fed vehicle], P < 0.05) as was insulin secretion (area under curve, AUC) across an OGTT (2.29 ± 0.23 vs. 2.95 ± 0.37 vs. 2.94 ± 0.21 ng · ml−1 · h−1 [A2 vs. vehicle vs. pair-fed vehicle], P < 0.05). Glucose levels across the OGTT (AUC) did not change significantly (22.4 ± 0.46 vs. 24.2 ± 0.42 vs. 22.9 ± 0.45 mmol · l−1 · h−1 [A2 vs. vehicle vs. pair-fed vehicle], P = not significant).
Cortisone pellet–implanted KK mice treated with A2 for 4 consecutive days had increased total IRS1 protein expression, decreased Ser307 phosphorylation, and increased Thr308 phosphorylation of Akt/PKB in whole-tissue quadriceps explants. Tyr608 phosphorylation did not change (Fig. 5A). Genecard analysis of quadriceps mRNA expression following A2 treatment is shown in Table 3. Positive findings were endorsed with real-time PCR (Fig. 5B). 11β-HSD1 expression decreased (0.48-fold) (Table 3, Fig. 5B) without effect on glucocorticoid receptor or H6PDH expression. In agreement with our protein expression data, IRS1 mRNA expression increased after selective 11β-HSD1 inhibition (Fig. 5B). The regulatory subunit of PI3K p85 decreased 0.25-fold after treatment with A2 with no change in catalytic subunit (p110) expression (Table 3, Fig. 5B). PI3K activity was not measured as part of this study. A2 treatment decreased expression of key target genes involved in lipogenesis (ACC1 0.3-fold, DGAT 0.4-fold), lipolysis (HSL 0.3-fold, ATGL 0.39-fold) and lipid oxidation (ACC2 0.6-fold) (Table 3, Fig. 5B). In addition, PDK4 increased 1.7-fold (Fig. 5B).
In this study, we have characterized in detail the impact of both synthetic and endogenous glucocorticoids upon insulin signaling in the rodent skeletal muscle cell line, C2C12. In addition, we have characterized expression and activity of 11β-HSD1 in both rodent and human skeletal muscle, ascribed a functional significance to its activity in terms of insulin sensitization, and have begun to explore the mechanisms by which this occurs.
Glucocorticoids impair insulin signaling at multiple levels, importantly decreasing total IRS1 protein expression and increasing Ser307 phosphorylation. IRS-1 serine phosphorylation at this site has been reported to decrease the affinity of IRS1 for the insulin receptor and increased IRS1 degradation (9,10), and this may account for the decrease in total protein expression that we observed. It is also sufficient to account for the glucocorticoid-induced decrease of insulin-stimulated glucose uptake. The pivotal role of IRS1 in skeletal muscle insulin signaling is highlighted by IRS1 knockout mice (22–24) that develop marked insulin resistance. Serine phosphorylation of IRS1 at numerous residues has been implicated in the development of insulin resistance (19,25–27). Specifically, Ser307 phosphorylation is a negative regulator of IRS1 function. Inflammatory cytokines including tumor necrosis factor-α and C-reactive protein increase Ser307 phosphorylation (28,29), and insulin itself has been described to have similar effects (30,31). Although several kinases have been implicated in serine phosphorylation of IRS1 (32), PKCθ is believed to have a critical role, notably after free fatty acid (FFA) exposure (33,34). PKCθ knockout animals resist lipid-induced skeletal muscle insulin resistance (35), and rodent models with muscle-specific targeted serine-to- alanine substitutions at residues within IRS1 including Ser307 resist fat-induced insulin resistance (36). Glucocorticoid induction of lipolysis and consequent FFA generation (37) as well as increased FFA uptake will activate PKCθ, and this may well be an important contributor to the insulin resistance induced by glucocorticoid. This is endorsed by our gene expression analysis of lipogenic/lipolytyic genes in rodent muscle after 4-day treatment with the selective 11β-HSD1 inhibitor A2 (see discussion below). In addition, the lack of Ser24 phosphorylation may also add weight to this hypothesis. PKCθ does not contribute to phorbol-12-myristate-13-acetate–induced Ser24 phosphorylation but instead is dependent upon PKCα activation (19).
Other studies have also highlighted the pivotal role of IRS1 in glucocorticoid-associated insulin resistance, although serine phosphorylation has not been examined. The results of these studies do show a degree of variability and some but not all have shown decreased activating tyrosine phosphorylation of IRS1 (11–13). Others have reported changes in insulin receptor expression and activation, PI3K activity and expression, and IRS2 expression and phosphorylation (12,38,39). The explanation for these inconsistencies is not entirely clear but may reflect differences between animal and cell models and specific investigative protocols.
Downstream of IRS1, we observed decreased activating pSer473 Akt/PKB and pThr308 Akt/PKB, and we propose that this may be a direct consequence of reduced insulin signaling capacity through enhanced IRS1 inactivation. In addition, we observed an increase in AS160 protein expression. AS160 is a recently identified protein with Rab-guanosine triphosphate (GTP)ase activity. Under basal conditions, it is resident within GLUT4-containing vesicles and limits the GTP availability that is necessary for vesicle translocation to the cell membrane to permit glucose entry. Upon phosphorylation by activated Akt/PKB, AS160 dissociates from the vesicle, allowing GTP to bind to Rab proteins and vesicle translocation to the cell membrane to occur (7). While regulation of AS160 phosphorylation at differing sites by growth factors including IGF-1 and EGF has been described (40), glucocorticoid regulation has not been explored. Our data show that glucocorticoids increase both AS160 protein and mRNA expression in a glucocorticoid receptor–dependent mechanism. These observations are interesting and point toward a separate mechanism of regulation rather than simply a downstream consequence of decreased IRS1 activation.
Although the net effect of glucocorticoid was to induce insulin resistance, we did observe an increase in IRS2 mRNA and protein expression. In addition IR, PI3K (p85 subunit), and GLUT4 mRNA expression increased, although protein expression was not examined in this study nor was PI3K-specific activity. It is possible that this represents a compensatory mechanism to preserve insulin sensitivity in an attempt to compensate for the inhibition of signaling through IRS1. However, overall the effect of glucocorticoid exposure is to limit insulin-stimulated glucose uptake.
In addition to the effect of exogenous glucocorticoids, we have shown that prereceptor metabolism of endogenous glucocorticoid by 11β-HSD1 is a crucial regulator of insulin sensitivity in skeletal muscle. 11β-HSD1 is expressed and biologically active in human skeletal muscle (41). Overexpression of 11β-HSD1 has been described in rodent skeletal muscle in models of diabetes (42) and myotubes isolated from patients with insulin resistance and type 2 diabetes (43,44). However, this is not a consistent finding (45), and its precise contribution to metabolic and muscle phenotype is still to be clarified. Selective 11β-HSD1 inhibitors are currently in development; in rodents and primates they limit local glucocorticoid availability and improve glucose tolerance, lipid profiles, and insulin sensitivity (16,46). Very recently, studies using in vitro differentiated primary in human myoblasts have shown that 11β-HSD1 inhibition (pharmacological or siRNA) can limit cortisone- but not cortisol-induced changes in glucose uptake, glycogen synthesis, and palmitate oxidation. However, in this model, an insulin-sensitizing action could not be demonstrated (47). In our study, we have clearly shown expression and activity of 11β-HSD1 in human and rodent skeletal muscle that is blocked by selective and nonselective 11β-HSD1 inhibitors. This is functionally important and not only restores IRS1 protein levels to control values but also decreases pSer307 IRS1, enhances Akt/PKB activation, and may represent an important insulin-sensitizing mechanism of selective 11β-HSD1 inhibitors. These observations appear consistent in both our in vitro and rodent in vivo models. A2 has an insulin-sensitizing action as evidenced by decreased fasting glucose and insulin levels, decreased homeostasis model assessment scores, and reduced insulin secretion across an OGTT. Unfortunately, clamp studies were not performed as part of this protocol but have been reported elsewhere with other selective 11β-HSD1 inhibitors (16). In addition to the actions described above, A2 administration to mice in vivo decreased lipogenic gene expression (ACC1, FAS, and DGAT) and increased FFA utilization (decreased ACC2 leading to a decrease in the malonyl CoA-mediated inhibition of β-oxidation) in agreement with published observations (48). Furthermore, decreased HSL and ATGL will afford decreases in FFA and DAG generation. Interestingly, we also observed a 1.7-fold increase in PDK4 expression with A2 treatment in vivo. PDK4 is a negative regulator of the pyruvate dehydrogenase complex, limiting acetyl CoA generation. Rodents with deletion of PDK4 have increased glucose oxidation (49), and in cell culture systems, glucocorticoids increase PDK4 expression (47). The discrepancies with our data, where A2 increased PDK4 expression, almost certainly reflect the complexities of whole-animal versus cell culture models (we too have observed decreased PDK4 expression after Dex treatment in C2C12 myotubes [data not shown]). Importantly, the increase in PDK4 with selective 11β-HSD1 inhibitors may further serve to drive lipid oxidation at the expense of glucose oxidation. The net effect of all these observations will be to decrease intramyocellular lipid accumulation as well as local FFA and DAG generation. Consequently, PKCθ activation will be decreased, and this may be responsible for the reduction in Ser307 phosphorylation after selective 11β-HSD1 inhibition. However, this hypothesis remains to be proven and needs to be addressed in future studies.
In conclusion, Ser307 phosphorylation of IRS1 is a novel mechanism of glucocorticoid-induced insulin resistance. The prereceptor modulation of glucocorticoid availability is an important regulator of glucocorticoid action in skeletal muscle. Selective 11β-HSD1 inhibitors enhance insulin action, and we propose that this may predominantly be through modulation of lipid metabolism within skeletal muscle. Clinical data utilizing 11β-HSD1 inhibitors are beginning to emerge in obese patients and those with type 2 diabetes (50); it is likely that their efficacy in muscle will provide an additional pharmacological benefit in the treatment of type 2 diabetes and insulin resistance.
Funding for this study has been provided by the Wellcome Trust, U.K. (program grant to P.M.S. and Clinician Scientist Fellowship to J.W.T.), a case award studentship from the Biotechnology and Biological Sciences Research Council in conjunction with AstraZeneca (to S.A.M.), the Medical Research Council (Clinical Research Training Fellowship to M.S.), and Diabetes U.K. (project grant to J.S.).
C.L., D.L., A.Y., G.C., and D.M.S. are employed by AstraZeneca, U.K., which manufactures pharmaceuticals related to the treatment of diabetes and its complications. In addition, those indicated are also stockholders in AstraZeneca. No other potential conflicts of interest relevant to this article were reported.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.