We previously showed that insulin has a profound effect to suppress pyruvate dehydrogenase kinase (PDK) 4 expression in rat skeletal muscle. In the present study, we examined whether insulin’s effect on PDK4 expression is impaired in acute insulin-resistant states and, if so, whether this change is accompanied by decreased insulin’s effects to stimulate Akt and forkhead box class O (FOXO) 1 phosphorylation. To induce insulin resistance, conscious overnight-fasted rats received a constant infusion of Intralipid or lactate for 5 h, while a control group received saline infusion. Following the initial infusions, each group received saline or insulin infusion (n = 6 or 7 each) for an additional 5 h, while saline, Intralipid, or lactate infusion was continued. Plasma glucose was clamped at basal levels during the insulin infusion. Compared with the control group, Intralipid and lactate infusions decreased glucose infusion rates required to clamp plasma glucose by ∼60% (P < 0.01), confirming the induction of insulin resistance. Insulin’s ability to suppress PDK4 mRNA level was impaired in skeletal muscle with Intralipid and lactate infusions, resulting in two- to threefold higher PDK4 mRNA levels with insulin (P < 0.05). Insulin stimulation of Akt and FOXO1 phosphorylation was also significantly decreased with Intralipid and lactate infusions. These data suggest that insulin’s effect to suppress PDK4 gene expression in skeletal muscle is impaired in insulin-resistant states, and this may be due to impaired insulin signaling for stimulation of Akt and FOXO1 phosphorylation. Impaired insulin’s effect to suppress PDK4 expression may explain the association between PDK4 overexpression and insulin resistance in skeletal muscle.
Pyruvate dehydrogenase kinase (PDK) phosphorylates and inactivates pyruvate dehydrogenase complex that catalyzes a rate-limiting step of glucose oxidation, i.e., the conversion of pyruvate to acetyl CoA. Previous studies (1–3) showed that starvation and experimental diabetes induce a stable increase in PDK activity in skeletal muscle, which explains decreased activity of pyruvate dehydrogenase complex and reduced glucose oxidation in these metabolic states. Skeletal muscle expresses two (i.e., PDK2 and PDK4) of the four PDK isoforms expressed in mammalian cells (PDK1–4) (4,5). The increases in muscle PDK activity with starvation and streptozotocin-induced diabetes have been shown to be due to a selective upregulation of PDK4 expression, which was attributed to increased circulating free fatty acids (FFAs) (6–9). However, our recent study (10) demonstrated that insulin had a strong effect to suppress PDK4 expression in skeletal muscle within 5 h, whereas changes in plasma FFAs had little effects. These data suggest that the increases in muscle PDK4 expression with starvation or diabetes may be due largely to insulin deficiency rather than to increases in circulating FFAs. Increased PDK4 expression, which would inhibit glucose oxidation, may be an important adaptive mechanism for glucose conservation in fasting states when glucose is scarce (11), and insulin appears to suppress it in fed states when glucose is abundant.
Insulin resistance (or insulin sensitivity) is associated with increased (or decreased) skeletal muscle PDK activity (1,6–8,12). PDK activity or PDK4 protein expression was shown to be increased in skeletal muscle of high-fat–fed rats (11,13–15), a well-established model of insulin resistance. PDK activity was also elevated in skeletal muscle of streptozotocin-induced diabetic animals (1–3), which are insulin resistant (16,17). On the other hand, increased insulin sensitivity in obese patients after malabsorptive bariatric surgery was associated with reduced muscle PDK4 expression (18). Furthermore, peroxisome proliferator–activated receptor (PPAR)γ agonists, which increase insulin sensitivity, decreased skeletal muscle PDK4 expression in Zucker diabetic fatty rats (19). One possibility is that increased PDK activity in skeletal muscle may play an important role in the development of insulin resistance by suppressing glucose oxidation (8,12). However, in light of our finding that insulin suppresses PDK4 expression (10), another possibility is that insulin’s ability to suppress PDK4 expression in skeletal muscle is impaired in insulin-resistant states, resulting in increased PDK4 expression and activity (12). One aim of the present study was to test this possibility.
Recent studies (20–22) have implicated the forkhead transcription factor forkhead box class O (FOXO) 1 in the regulation of PDK4 expression by insulin. FOXO1 is expressed in insulin-sensitive tissues such as liver, skeletal muscle, and adipose tissue and is a major regulator of insulin action in these tissues (20–22). Kwon et al. (23) suggested that insulin’s effect to suppress PDK4 expression in HepG2 cells involves FOXO1. In addition, Furuyama et al. (24) reported that overexpression of FOXO1 increased PDK4 expression in C2C12 myotubes. These and other studies in cultured cells have suggested that insulin and other growth factors regulate the activities of FOXO transcription factors via their phosphorylation by the phosphatidylinositol 3-kinase (PI3K)-Akt pathway (25). FOXO transcription factors localize in the nucleus in the basal state and, upon stimulation with growth factors, get phosphorylated by Akt, leading to nuclear exportation and inhibition of FOXO-dependent transcription (26). Insulin stimulation of PI3K and/or Akt is often impaired in insulin-resistant states (27–30). If so, this would decrease insulin’s ability to regulate FOXO1 activity and thus PDK4 expression. A second aim of the present study was to test if insulin’s ability to stimulate Akt and FOXO1 phosphorylation is impaired, accompanying a possible impairment of insulin’s action on PDK4 expression, in insulin-resistant states.
RESEARCH DESIGN AND METHODS
Animals and catheterization.
Male Wistar rats weighing 275–300 g were obtained from Simonsen (Gilroy, CA) and studied ≥5 days after arrival. Animals were housed under controlled temperature (22 ± 2°C) and lighting (12-h light, 0600–1800; 12-h dark, 1800–0600) with free access to water and standard rat diet. At least 4 days before the experiment, animals were placed in individual cages with tail restraint as previously described (29,31), which was required to protect tail blood vessel catheters during experiments. Animals were free to move about and were allowed unrestricted access to food and water. Two tail vein infusion catheters were placed the day before the experiment, and one tail artery blood sampling catheter was placed in the morning of the experiment (i.e., 0600). All procedures were approved by the institutional animal care and use committee at the University of Southern California.
Experiments were conducted after an overnight fast; food was removed at 1700 on the day before the experiment. In the morning of the experiment, animals received a constant infusion of saline, Intralipid (triglyceride emulsion, 20% wt/vol, 0.9 ml/h; Baxter, Deerfield, IL) with heparin (40 units/h), or lactate (120 μmol · kg−1 · min−1; l-(+)-lactate sodium salt diluted in phosphate buffer, 0.06 mol/l NaH2PO4, and 0.0134 mol/l Na2HPO4, pH 4.5 [32,33]) for 5 h. Vettor et al. (32) reported that lactate infusion, at an infusion rate 50% higher than the present one, caused a slight increase in arterial pH from 7.37 to 7.59. Following the initial infusions, each group received a constant infusion of saline or porcine insulin (30 pmol · kg−1 · min−1; Eli Lilly, Indianapolis, IN) for an additional 5 h (n = 6–7 each), while the saline, Intralipid, or lactate infusion was continued. During the insulin infusion, plasma glucose was clamped at basal levels (∼6 mmol/l) by exogenous glucose infusion. At the end of the experiments, animals were anesthetized and gastrocnemius and soleus muscles were rapidly dissected, frozen immediately using liquid N2-cooled aluminum blocks, and stored at −80°C for later analysis. Blood samples were taken at various time points to measure plasma glucose, insulin, FFAs, and lactate.
Northern blot analysis for PDK2 and PDK4 mRNA levels.
RNA extraction and Northern blot analysis were performed as previously described (10). Briefly, total RNA was extracted from frozen muscles using Tri Reagent from Molecular Research Center (Cincinnati, OH) according to the manufacturer’s instructions. Electrophoresis was performed using 25 μg each of total RNA preparations in 1% denaturing gel. RNA was then capillary transferred onto a positive-charged nylon membrane (BrightStar-Plus; Ambion, Austin, TX). cDNA probes for rat PDK2 and PDK4 were obtained by RT-PCR using an Advantage One Step RT-PCR kit from Clontech (Palo Alto, CA) and the following primers: PDK-4, 5′-CGTCGCCAGAATTAAAGCTC and 3′-CTGCCAGTTTCTCCTTCGAC and PDK-2, 5′-GTCAGCTAGGGGCCTTCTCT and 3′-CAGGACTATGCAGGCAGTGA. The cDNA probes were labeled with [32P]dCTP (Perkin Elmer) using a DECAprime DNA labeling kit (Ambion). Hybridization was carried out at 42°C in Ultrahyb solution (Ambion). Relative densities from the autoradiographs were quantified using the Bio-Rad Molecular Analyst. To control RNA loading, all blots were quantified for glyceraldehyde-3-phosphate dehydrogenase by use of a probe included in the DECAprime DNA labeling kit (Ambion).
Western blot analysis for total and phosphorylated protein levels of Akt and FOXO1.
Frozen muscles (∼50 mg) were homogenized using a Tekmar homogenizer (Cincinnati, OH) at half-maximum speed (1 min, on ice) in 500 μl of buffer (20 mmol/l Tris, pH 7.5; 5 mmol/l EDTA; 10 mmol/l Na4P2O7; 100 mmol/l NaF; 2 mmol/l Na3VO4; 1% NP-40; 1 mmol/l phenylmethylsulfonyl fluoride; 10 μg/ml aprotinin; and 10 μg/ml leupeptin) (10). Muscle lysates were further solubilized by incubating with continuous rotation at 4°C for 1 h. Total protein was obtained by centrifugation at 14,000g at 4°C for 20 min. The supernatants (50 μg protein) were resolved by SDS-PAGE followed by electrophoretic transfer of proteins onto Hybond-P membranes (Amersham, Piscataway, NJ). The membranes were then probed with rabbit antibodies against Akt (Cell Signaling Technology, Beverly, MA), phospho-Akt (Ser307 specific; Cell Signaling Technology), or FOXO1. Two different antibodies were used for FOXO1: one from Santa Cruz Biotechnology (Santa Cruz, CA) for total FOXO1 protein level and the other from Cell Signaling Technology for FOXO1 phosphorylation. The anti-FOXO1 antibody from Cell Signaling Technology has an epitope that includes one (i.e., Ser319) of the three major phosphorylation sites (i.e., Thr24, Ser256, and Ser319) affected by Akt. This antibody recognizes unphosphorylated FOXO1, and therefore, signal with this antibody is inversely related to FOXO1 phosphorylation in the absence of change in total protein level (34). Cell Signaling Technology also provides phospho-FOXO1–specific antibodies (i.e., Thr24 and Ser256), but these antibodies did not work well with our muscle extracts, presumably due to the presence of interfering proteins. After the incubation with primary antibodies, the membranes were incubated with a secondary antibody (i.e., horseradish peroxidase–conjugated anti-rabbit IgG; Amersham). Signals were then detected by an enhanced chemiluminescence method and quantified using the Bio-Rad Molecular Analyst.
Plasma glucose was analyzed by a glucose oxidase method on a Beckman Glucose Analyzer II (Beckman, Fullerton, CA). Plasma insulin was measured by radioimmunoassay using a kit from Linco Research (St. Charles, MO). Plasma FFAs were measured using an acyl-CoA oxidase-based colorimetric kit (Wako Chemicals, Richmond, VA). Plasma lactate was measured by a lactate oxidase method on a YSI lactate analyzer (Yellow Springs Instruments, Yellow Springs, OH).
Data are expressed as means ± SE. The significance of the differences in mean values among different treatment groups was evaluated using one-way ANOVA followed by ad hoc analysis using the Tukey test. Pearson’s correlation was used to evaluate univariate correlation. P < 0.05 was considered statistically significant.
Acute induction of insulin resistance with Intralipid or lactate infusion.
To induce insulin resistance in skeletal muscle, overnight-fasted rats received a constant infusion of Intralipid or lactate for 10 h, while the control group received saline infusion. Each of the infusions was performed with or without insulin (and glucose clamp) during the final 5-h period for determination of insulin’s effects on PDK4 expression. Intralipid infusion raised plasma FFA concentration from ∼0.7 (control group) to ∼1.2 mmol/l (Fig. 1, time 0; P < 0.05), and lactate infusion raised plasma lactate concentration from ∼0.5 to >2 mmol/l (P < 0.05). Interestingly, lactate infusion also raised plasma glucose from ∼6 to ∼9 mmol/l (P < 0.05). It is unclear whether this increase was due to increased hepatic glucose production and/or decreased peripheral glucose clearance. Insulin infusion during the final 5-h period raised plasma insulin levels similarly to ∼600 pmol/l in all groups. During the insulin infusion, plasma glucose was clamped at the same level of ∼6 mmol/l by exogenous glucose infusion (hyperinsulinemic-euglycemic clamp). In the lactate group, plasma glucose was clamped at this, rather than its own, basal level to compare insulin action at the same glucose concentrations among the different infusion groups. Glucose infusion rate (GIR) required to clamp plasma glucose, reflecting insulin’s action to promote glucose uptake and to inhibit hepatic glucose production, increased rapidly and reached a steady state within 2 h of insulin infusion in the control group. Compared with the control group, Intralipid and lactate infusions decreased GIR resulting in ∼60% lower rates at the end (P < 0.01), demonstrating an induction of insulin resistance. These results are similar to those in our previous studies (29,31), which also showed that these decreases in GIR with Intralipid and lactate infusion were largely due to decreases in insulin-stimulated glucose uptake.
Effects of acute insulin resistance on insulin’s ability to suppress PDK4 mRNA expression.
At the end of the experiments, animals were anesthetized and gastrocnemius (and soleus) muscle samples were taken and analyzed for PDK mRNA levels. PDK2 mRNA level was not altered in gastrocnemius muscle by the 10-h lactate or Intralipid infusion (Fig. 2; P > 0.05). In addition, PDK2 mRNA level was not altered by the final 5-h insulin infusion in all three (i.e., saline, Intralipid, and lactate) infusion groups (P > 0.05). In contrast, PDK4 mRNA level was suppressed by insulin by >80% in the saline control group (P < 0.05), consistent with our previous report (10). This effect of insulin was significantly reduced with the Intralipid and the lactate infusions, although the change was less dramatic with the lactate infusion. As a result, PDK4 mRNA levels with insulin were higher in the insulin-resistant (i.e., Intralipid and lactate) groups compared with the control group (P < 0.05). It is interesting to note that the 10-h Intralipid infusion had a small (∼25%) but statistically significant effect to decrease basal PDK4 mRNA levels (P < 0.05), since previous studies suggested that increased plasma FFA levels increase muscle PDK4 expression (see discussion). Similar results were obtained in soleus muscle (Fig. 3), which contains mostly slow-twitch oxidative fibers (versus fast-twitch fibers in gastrocnemius muscle), suggesting that the insulin effect on PDK4 mRNA expression and its alteration in insulin-resistant states are independent of muscle fiber type.
Our previous study (10) demonstrated that a 5-h insulin infusion had a profound (∼72%) effect to decrease PDK4 mRNA level, but insulin’s effect was only modest (∼21%) at the protein level, suggesting that the 5-h insulin infusion period was not a sufficient time for decreased PDK4 transcription to be fully reflected in protein level. In the present study, we did not determine PDK4 protein level, as detecting a change in the small insulin effect on PDK4 protein level would be practically infeasible or otherwise require a huge number of experiments.
Effects of acute insulin resistance on insulin’s ability to increase Akt and FOXO1 phosphorylation.
Previous studies (24,23) have implicated FOXO1 in the regulation of PDK4 gene expression by insulin. Insulin is known to phosphorylate and inhibit FOXO activity via the PI3K/Akt pathway (35). We examined whether insulin’s effects to stimulate Akt and FOXO1 phosphorylation are impaired in the insulin resistant states induced with Intralipid or lactate. As expected, insulin had a profound effect to increase Akt phosphorylation in gastrocnemius muscle of the control group (Fig. 4). This effect of insulin was significantly reduced with Intralipid or lactate (∼45 and ∼60%, respectively; P < 0.05). Total Akt protein level was not altered by Intralipid or lactate infusion or by insulin. Similarly, total FOXO1 protein level in gastrocnemius muscle was not altered by any of the infusions (Fig. 5). FOXO1 phosphorylation was quantified using the anti-FOXO1 antibody from Cell Signaling Technology (see research design and methods), which recognizes active, unphosphorylated FOXO1 (34). We found that insulin significantly decreased the signal for unphosphorylated FOXO1 (thus increased FOXO1 phosphorylation; P < 0.05). This effect of insulin was less with Intralipid infusion or absent with lactate infusion. There was a significant, negative correlation between Akt phosphorylation and unphosphorylated FOXO1 or PDK4 mRNA level and a positive correlation between unphosphorylated FOXO1 and PDK4 mRNA level (Fig. 6). These data support the notions that insulin suppresses PDK4 mRNA expression by stimulating Akt and FOXO1 phosphorylation and reduced insulin’s ability to suppress PDK4 mRNA expression in insulin-resistant states may be due to impaired insulin stimulation of Akt.
Increased PDK4 expression in muscle has been observed in insulin resistant states, such as high-fat–fed (11,13–15,25) or streptozotocin-induced diabetic (1–3) animals. However, it was unclear whether PDK4 overexpression causes insulin resistance or vice versa. The present finding that insulin’s effect to suppress PDK4 expression was impaired in insulin-resistant states, resulting in increased PDK4 mRNA levels with insulin, suggests that insulin resistance may cause PDK4 overexpression. However, we cannot exclude the possibility that PDK4 overexpression can also cause insulin resistance by suppressing glucose oxidation (8,12). In regard to this, it is interesting to note that a high-fat diet increased PDK4 expression in human skeletal muscles as early as after 1 day (15), presumably before insulin resistance develops. In rats, the development of skeletal muscle insulin resistance took >3 days of high-fat feeding (36,37). Our preliminary data showed that muscle PDK4 mRNA level increased significantly within 2 days of high-fat feeding (data not shown). Taken together, these data raise the possibility that muscle PDK4 activity increases early during high-fat feeding, preceding and possibly causing insulin resistance. Thus, it may be possible that the cause-and-effect relationship between PDK4 overexpression and insulin resistance exists in both directions to result in a vicious cycle of PDK4 overexpression and insulin resistance. Whether PDK4 overexpression can indeed cause insulin resistance remains to be tested.
Earlier studies (1–3) suggested that increased circulating FFA level is responsible for the upregulation of PDK4 expression in starvation and experimental diabetes. FFAs are an endogenous ligand for the PPAR-α (38,39), which is known to induce muscle PDK4 expression (6,35). However, our previous study (10) demonstrated that changes in PDK4 expression that occur during refeeding of fasted rats could not be explained by changes in plasma FFAs since suppression of plasma FFAs for 5 h similar to those during refeeding had no effect on PDK4 mRNA expression. In the present study, elevation of plasma FFAs for 10 h via Intralipid infusion (in the absence of changes in plasma insulin) did not increase PDK4 mRNA levels. In fact, we observed a 25% decrease in PDK4 mRNA level in gastrocnemius muscles with the Intralipid infusion. Thus, these data strengthen our suggestion that plasma FFA level may not play a major role in the upregulation of PDK4 expression in starvation and diabetes (10). In addition to PPARα, FFAs stimulate PPARγ. PPARγ was shown to antagonize FOXO1 activity (40). It is conceivable that elevated plasma FFAs stimulate PPARγ and antagonize FOXO1 binding to PDK4 promoter, which may be responsible for the decrease in PDK4 mRNA level with Intralipid infusion.
In the present study, insulin resistance was acutely induced by Intralipid or lactate infusion. These models of acute insulin resistance have been well characterized by us (29,31,41) and others (36,42). The changes in GIR with Intralipid or lactate are similar to those previously reported (29,31). Our previous studies (29,31,41) have demonstrated that these changes are largely due to changes in insulin-stimulated glucose uptake in skeletal muscle. In addition, a 5-h infusion of lactate (29) or Intralipid (F.N.L., J.H.Y., unpublished data) in vivo, at doses identical to those in the present study, decreased insulin-stimulated glucose transport activity, as assessed in vitro in isolated muscles following the infusions. Since changes in insulin signaling pathways in these insulin resistance models have been well documented (29,42), in the present study we determined only insulin-stimulated Akt phosphorylation, which is directly related to FOXO1 phosphorylation. The changes in insulin-stimulated Akt phosphorylation are similar to those described in previous studies (29,42), which demonstrated that these changes are caused by impaired insulin stimulation of PI3K, an upstream event.
Recent studies (23,24) have implicated FOXO1 in the regulation of PDK4 expression by insulin. Studies in cultured cells have shown that insulin and other growth factors regulate the activities of FOXO transcription factors via phosphorylation by the PI3K-Akt pathway (25). Thus, FOXO transcription factors localize in the nucleus in the basal state and, upon stimulation with growth factors, get phosphorylated by Akt, leading to nuclear exportation and inhibition of FOXO-dependent transcription (26). In the present study, we found a significant positive correlation between unphosphorylated (active) FOXO1 and PDK4 mRNA levels and a significant negative correlation between Akt phosphorylation and unphosphorylated FOXO1 levels, suggesting that the above mechanisms for insulin regulation of PDK4 expression via Akt and FOXO1 phosphorylation operate in intact rat skeletal muscle. However, we cannot exclude the possibility that FOXO1-independent mechanisms are also involved in insulin regulation of PDK4 mRNA expression (43). Recent studies (44,45) have suggested a role of PPARγ coactivator 1α in the regulation of PDK4 gene expression via FOXO1-independent mechanisms. Whether PPARγ coactivator 1α contributes to the acute insulin regulation of PDK4 gene expression remains to be studied.
FOXO1 phosphorylation was determined indirectly using the anti-FOXO1 antibody from Cell Signaling Technology, taking advantage of the fact that this antibody recognizes unphosphorylated FOXO1 (34). Thus, in the absence of changes in total FOXO1 protein levels, the signal detected by this antibody is inversely related to FOXO1 phosphorylation. Insulin significantly decreased this signal in the control group, without altering total FOXO1 protein levels, indicating a decrease in unphosphorylated FOXO1 and thus an increase in phosphorylated FOXO1. Alkaline phosphatase treatments of muscle lysates (for protein dephosphorylation) increased the anti-FOXO1 signal and resulted in no difference between basal and insulin-stimulated muscle (data not shown), confirming that total FOXO1 protein was not changed by insulin. The Cell Signaling Technology anti-FOXO1 antibody has been the most popular antibody for quantifying FOXO1 protein level. The fact that this antibody recognizes only (or preferentially) unphosphorylated FOXO1 (around Ser319) was previously observed (34) but not widely recognized. Many studies have used the Cell Signaling Technology anti-FOXO1 antibody for quantification of FOXO1 protein level. The present and Wang et al.’s (34) studies suggest that the data with Cell Signaling Technology anti-FOXO1 antibody need to be carefully interpreted as they could reflect changes in FOXO1 phosphorylation rather than total protein levels. Phospho-specific antibodies from Cell Signaling Technology did not work well with our muscle homogenates, presumably due to the presence of interfering proteins.
In conclusion, our results demonstrate that the ability of insulin to suppress PDK4 mRNA expression in skeletal muscle, observed in our previous study (10), was impaired in insulin-resistant states acutely induced in rats with Intralipid or lactate infusion. Our data also indicate that this impairment was accompanied by impaired insulin stimulation of Akt and FOXO1 phosphorylation, suggesting a major role of these molecules in the regulation of PDK4 expression by insulin in skeletal muscle.
Y.I.K. and F.N.L. contributed equally to this work.
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.