Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
R. A. Harris, Richard Roudebush Veterans Affairs Medical Center, Research Service 151, Room D-3034, 1481 West Tenth Street, Indianapolis, IN 46202, USA Fax: +1 317 988 3180 Tel: +1 317 988 4544 E-mail: firstname.lastname@example.org
Although improving glucose metabolism by inhibition of pyruvate dehydrogenase kinase 4 (PDK4) may prove beneficial in the treatment of type 2 diabetes or diet-induced obesity, it may have detrimental effects by inhibiting fatty acid oxidation. Peroxisome proliferator-activated receptor α (PPARα) agonists are often used to treat dyslipidemia in patients, especially in type 2 diabetes. Combinational treatment using a PDK4 inhibitor and PPARα agonists may prove beneficial. However, PPARα agonists may be less effective in the presence of a PDK4 inhibitor because PPARα agonists induce PDK4 expression. In the present study, the effects of clofibric acid, a PPARα agonist, on blood and liver lipids were determined in wild-type and PDK4 knockout mice fed a high-fat diet. As expected, treatment of wild-type mice with clofibric acid resulted in less body weight gain, smaller epididymal fat pads, greater insulin sensitivity, and lower levels of serum and liver triacylglycerol. Surprisingly, rather than decreasing the effectiveness of clofibric acid, PDK4 deficiency enhanced the beneficial effects of clofibric acid on hepatic steatosis, reduced blood glucose levels, and did not prevent the positive effects of clofibric acid on serum triacylglycerols and free fatty acids. The metabolic effects of clofibric acid are therefore independent of the induction of PDK4 expression. The additive beneficial effects on hepatic steatosis may be due to induction of increased capacity for fatty acid oxidation and partial uncoupling of oxidative phosphorylation by clofibric acid, and a reduction in the capacity for fatty acid synthesis as a result of PDK4 deficiency.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Fibrates are clinically used to decrease plasma triacylglycerol and increase the level of high-density lipoprotein cholesterol . These actions of fibrates, which are thought to be due to altered expression of apolipoproteins [2,3] and peroxisomal and mitochondrial fatty acid oxidation enzymes , reduce the risk of cardiovascular disease, improve insulin sensitivity and reduce adiposity . Most of the effects of fibrates are mediated by activation of the nuclear receptor peroxisome proliferator-activated receptor α (PPARα), a transcription factor that is responsible for the regulation of genes involved in lipid metabolism. Activation of PPARα by fibrates also ameliorates hepatic steatosis in rodents [6,7].
Regulation of the activity of the pyruvate dehydrogenase complex (PDC) is important for glucose homeostasis and the control of fuel selection by tissues . When blood glucose levels are elevated by carbohydrate intake, PDC is relatively dephosphorylated and active, promoting glucose disposal and the synthesis and storage of fatty acids as triacylglycerols. When glucose levels are low, for example because of fasting, PDC is highly phosphorylated and inactive, inhibiting oxidation of gluconeogenic substrates (pyruvate, lactate and alanine) and promoting glucose synthesis and fatty acid oxidation. The phosphorylation state and therefore the activity of PDC  are determined by the relative activities of four pyruvate dehydrogenase kinases (PDKs) and two pyruvate dehydrogenase phosphatases. Of the enzymes that modulate PDC activity, PDK4 is of special interest because its expression is markedly increased by fasting , and studies with PDK4 knockout mice  have shown that PDK4 is important for maintaining fasting blood glucose levels.
Fibrates are effective for the treatment of dyslipidemia, one of the major problems of diabetes. The possibility that PDK4 inhibitors may also prove useful in the treatment of type 2 diabetes and obesity was suggested by previous studies [12,13]. As PPARα agonists increase PDK4 expression in muscle, heart, kidney and liver [14,15], which may mediate some of the effects of fibrates, there is concern that fibrates may be less effective in the presence of a PDK4 inhibitor. This study was designed to test whether the lipid-lowering effects of clofibric acid, a PPARα agonist, are affected by PDK4 deficiency. Surprisingly, PDK4 deficiency enhanced the beneficial effects of clofibric acid on hepatic steatosis, and did not prevent its hypolipidemic effects. These findings suggest the metabolic effects of clofibric acid are independent of the induction of PDK4 expression. Therefore, a PDK4 inhibitor and a fibrate could potentially be used in combination to lower blood glucose and ameliorate hepatic steatosis.
Effects of clofibric acid and PDK4 deficiency on body weight, tissue weights and food consumption of mice fed an HSF diet
As found in a previous study , PDK4 knockout mice gained less weight and had smaller livers than wild-type mice fed a high saturated fat (HSF) diet (Table 1). Clofibric acid also reduced body weight and epididymal fat pad weight, but, as expected, increased liver weight in both wild-type and PDK4 knockout mice (Table 1). Body weight gains were almost completely prevented by clofibric acid in both wild-type and PDK4 knockout mice (Table 1). Epididymal fat pad weights were likewise reduced to a similar extent in both groups. The reduction in liver size caused by PDK4 deficiency was completely overcome by clofibric acid (Table 1). Food consumption was not affected by clofibric acid in either the wild-type (1.8 ± 0.1 versus 2.1 ± 0.2 g per day for non-treated and clofibric acid-treated mice, respectively) or PDK4 knockout mice (1.8 ± 0.1 versus 2.0 ± 0.1 g per day for non-treated and clofibric acid-treated mice, respectively). Despite this, development of obesity was largely prevented by addition of clofibric acid to the diet of both the wild-type and PDK4 knockout mice (Fig. 1). Food efficiency in terms of weight gain per amount of food consumed was therefore markedly attenuated by clofibric acid in both the wild-type and PDK4 knockout mice.
Table 1. Body and tissue weights of fasted wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE (n =12). The diet of half of the wild-type mice and half of the PDK4 knockout mice was supplemented with clofibric acid after the first 14 weeks of feeding.
Body weight gain (g)
Final body weight (g)
Epididymal fat pads (% of body weight)
Liver (% of body weight)
First 14 weeks
Second 14 weeks
*P <0.05 relative to non-treated wild-type mice. †P <0.05 relative to non-treated PDK4 knockout mice.
15.0 ± 0.8
13.1 ± 1.3
48.3 ± 1.8
7.5 ± 0.3
3.8 ± 0.3
4.8 ± 1.7*†
36.3 ± 2.6*
4.8 ± 0.8*
5.2 ± 0.1*†
12.8 ± 0.9*
9.8 ± 0.8*
41.1 ± 1.5*
6.5 ± 0.2
3.0 ± 0.1*
2.6 ± 0.3*†
32.5 ± 1.3*†
3.8 ± 0.6*†
5.2 ± 0.2*†
As PPARα activation induces expression of uncoupling proteins (UCPs) in several tissues [16–20], dissipation of energy by heat as a consequence of uncoupling of oxidative phosphorylation may be responsible for the reduced food efficiency and lower weight gain in fibrate-treated rats [21,22]. Induction of expression of UCP3 by clofibric acid in the fed state in the liver of wild-type mice fed the HSF diet was confirmed by quantitative real-time PCR (Table 2). PDK4 deficiency alone had no effect, but supplementation of the diet of PDK4 knockout mice with clofibric acid produced an even greater relative amount of UCP3 to clofibric acid-treated wild type mice (Table 2). Therefore, uncoupling of oxidative phosphorylation and the resulting reduced food efficiency may contribute to the additive effects of clofibric acid treatment and PDK4 deficiency.
Table 2. UCP3 mRNA expression in livers of wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE relative to non-treated wild-type mice (n =4. UCP3 mRNA was measured by quantitative real-time PCR as described in Experimental procedures.
Liver (fold change)
*P <0.05 relative to non-treated wild-type mice. †P <0.05 relative to non-treated PDK4 knockout mice. #P <0.05 relative to clofibric acid-treated wild-type mice.
1.0 ± 0.1
59.0 ± 4.8*†
0.5 ± 0.2
82.2 ± 5.2*†#
Effects of clofibric acid and PDK4 deficiency on blood glucose levels and ketone bodies in mice fed an HSF diet
As expected , wild-type mice fed the HSF diet exhibited fasting hyperglycemia (Table 3), whereas PDK4 knockout mice remained euglycemic. Supplementing the diet with clofibric acid had no effect on blood glucose in the wild-type and PDK4 knockout mice (Table 3). As also expected , the levels of serum ketone bodies were significantly greater in PDK4 knockout mice than in wild-type mice (Table 3), and clofibric acid increased the level further in both groups. Thus, PDK4 deficiency and clofibric acid have additive effects on serum ketone bodies.
Table 3. Blood glucose, triacylglycerol, free fatty acid and ketone body levels of fasted wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE (n =5 or 6).
Blood glucose (mg·dL−1)
Free fatty acid (mm)
Ketone bodies (mm)
*P <0.05 relative to non-treated wild-type mice. †P <0.05 relative to non-treated PDK4 knockout mice. #P <0.05 relative to clofibric acid-treated wild-type mice.
143 ± 5
48 ± 5
0.48 ± 0.10
0.20 ± 0.02
134 ± 7
29 ± 1*†
0.37 ± 0.07
0.58 ± 0.10*
90 ± 9*#
71 ± 9*
0.67 ± 0.09
0.46 ± 0.05*
105 ± 7*#
29 ± 1*†
0.29 ± 0.05†
1.09 ± 0.14*†#
Effects of clofibric acid and PDK4 deficiency on serum lipids and hepatic steatosis in mice fed an HSF diet
Supplementing the diet with clofibric acid lowers serum triacylglycerol (Table 3). In contrast, PDK4 deficiency caused a modest increase in serum triacylglycerol and a trend towards higher free fatty acid levels (Table 3). Despite this, the combination of clofibric acid and PDK4 deficiency significantly reduced both serum triacylglycerol and free fatty acids (Table 3). The beneficial effects of clofibric acid on serum lipids were therefore greater in PDK4 knockout mice than wild-type mice fed the HSF diet.
Hepatic steatosis induced by the HSF diet (Fig. 2A) was attenuated by PDK4 deficiency (Fig. 2B), as reported previously . Supplementing the diet of wild-type mice with clofibric acid also reduced the liver fat content (Fig. 2C). Combination of PDK4 deficiency and treatment with clofibric acid resulted in the lowest amount of liver fat (Fig. 2D). These findings were confirmed by quantitative analysis of the amounts of liver triacylglycerol (Table 4). Supplementing the diet with clofibric acid and PDK4 deficiency reduced liver triacylglycerol levels. Consistent with the histological analysis, the lowest amount of triacylglycerol was found in the liver of clofibric acid-treated PDK4 knockout mice.
Table 4. Concentration of triacylglycerol in livers and gastrocnemius muscles of overnight-fasted wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE (n =6). The concentration of triacylglycerol was measured as described in Experimental procedures.
Liver (mg·g−1 wet weight)
Gastrocnemius (mg·g−1 wet weight)
*P <0.05 relative to non-treated wild-type mice. †P <0.05 relative to non-treated PDK4 knockout mice.
235 ± 18
90 ± 8
114 ± 24*†
49 ± 8*
185 ± 9*
64 ± 7*
72 ± 11*†
33 ± 5*†
Similar to the findings for the liver, dietary clofibric acid and PDK4 deficiency decreased the amount of triacylglycerol in gastrocnemius muscles (Table 4). As with the liver, these effects were additive, with the lowest amount of triacylglycerol was found in the gastrocnemius muscle by the combination of clofibric acid and PDK4 deficiency. These findings show that this desirable effect of clofibric acid on hepatic steatosis is not dependent on up-regulation of PDK4.
Effects of clofibric acid and PDK4 deficiency on insulin sensitivity in mice fed an HSF diet
Wild-type mice fed the HSF diet for 20 weeks were insulin-resistant as determined by insulin levels (data not shown) and the insulin tolerance test (Fig. 3). As reported previously , PDK4 knockout mice fed the HSF diet were significantly more insulin-sensitive than wild-type mice. Supplementing the diet with clofibric acid also improved insulin sensitivity in the wild-type mice, but the effect was not additive with the increase in insulin sensitivity caused by PDK4 deficiency (Fig. 3).
Effects of clofibric acid and PDK4 deficiency on PDC activity in livers and gastrocnemius muscles of mice fed an HSF diet
As clofibric acid was expected to induce PDK4 in the liver, we anticipated a lower PDC activity state (percentage of the complex in the active dephosphorylated state) in the liver of clofibric acid-treated wild-type mice compared to non-treated mice. Surprisingly, clofibric acid significantly increased the PDC activity state (Table 5). In contrast to our previous study, which indicated no effect of PDK4 deficiency on the PDC activity in livers of mice fed a different high fat diet , PDK4 deficiency resulted in a significantly higher PDC activity state in the livers of mice fed the HSF used in this study (Table 5). The highest PDC activity state occurred in PDK4 knockout mice treated with clofibric acid (Table 5). These findings beg the question of whether expression or activities of the pyruvate dehydrogenase phosphatases may be affected by clofibric acid, but this was not investigated in this study. In contrast to the liver, neither PDK4 deficiency nor clofibric acid treatment increased the PDC activity state in the gastrocnemius muscles of wild-type and PDK4 knockout mice (Table 5).
Table 5. PDC activity in livers and gastrocnemius muscles of fasted wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE (n =5 or 6). The percentage active PDC was determined as described in Experimental procedures.
Liver (% active)
Gastrocnemius (% active)
*P <0.03 relative to non-treated wild-type mice. †P <0.01 relative to non-treated PDK4 knockout mice.
0.6 ± 0.2
1.6 ± 0.2
7.6 ± 0.5*†
1.5 ± 0.6
3.1 ± 0.4*
2.3 ± 0.3
9.0 ± 1.5*†
1.9 ± 0.4
Effects of clofibric acid on mitochondrial enzymes in livers of mice fed an HSF diet
In an effort to understand the basis for the increase in the PDC activity state in the liver of clofibric acid-treated mice, the protein levels for the four PDKs were measured by Western blot analysis. No effect of clofibric acid or PDK4 deficiency was observed on the amounts of PDK1, PDK2 or PDK3 protein (data not shown). As expected, PDK4 protein was totally absent from the livers of PDK4 knockout mice and was induced to high levels by clofibric acid in the liver of wild-type mice in the fed state (Fig. 4). However, treatment with clofibric acid did not increase the amount of PDK4 protein in the fasting state in the liver of wild-type mice fed the HSF diet (Fig. 4). The same was found for PDK4 expression in gastrocnemius muscles (data not shown). These findings may explain, at least in part, why the PDC activity was not suppressed in the livers of clofibric acid-treated wild-type mice in the fasting state. The mechanism by which treatment with clofibric acid increases the PDC activity remains to be elucidated.
Effects of clofibric acid and PDK4 deficiency on the liver capacity for fatty acid oxidation and synthesis in mice fed an HSF diet
To determine the mechanism by which clofibric acid reduces the hepatic steatosis induced by the HSF diet, the protein levels for selected enzymes involved in lipid metabolism were measured (Fig. 4). Clofibric acid increased the levels of medium-chain acyl CoA dehydrogenase (MCAD) in the livers of both wild-type and PDK4 knockout mice, consistent with the known stimulation of fatty acid oxidation by PPARα agonists . Consistent with this increase in fatty acid oxidation enzymes, the protein levels of estrogen-related receptor α (ERRα), a positive transcriptional factor for MCAD expression  and PDK4 expression , were significantly increased by clofibric acid in the livers of both genotypes (Fig. 4). As expected from our previous study , the amount of acetyl CoA carboxylase 1 (ACC-1), the rate-limiting enzyme for fatty acid synthesis, was decreased in the livers of PDK4 knockout mice with and without supplementation of the diet by clofibric acid (Fig. 4).
Effects of clofibric acid and PDK4 deficiency on acetyl CoA and CoA levels in livers of mice fed an HSF diet
Expansion of the CoA pool size is believed to be part of the mechanism by which clofibric acid promotes fatty acid oxidation in the liver [26,27]. Supplementing the HSF diet with clofibric acid increased the levels of CoA in both wild-type and PDK4 knockout mice fed the HSF diet (Table 6). Although not statistically significant, a trend towards higher acetyl CoA levels in the livers of clofibric acid-treated mice of both genotypes was observed (Table 6). It is also notable that PDK4 deficiency did not change the levels of CoA and acetyl CoA or the ratio of acetyl CoA to CoA with or without treatment with clofibric acid. These findings suggest that flux through PDC is probably not differentially regulated by the acetyl CoA to CoA ratio in the various treatments used in this study, and that expansion of the CoA pool in response to clofibric acid is not linked to up-regulation of PDK4 expression.
Table 6. Free CoA and acetyl CoA levels in the livers of fasted wild-type (WT) and PDK4 knockout (PDK4 KO) mice fed an HSF diet with and without clofibric acid. Values are means ± SE (n =5 or 6). Values are nmol·g−1 liver wet weight.
*P <0.01 relative to non-treated wild-type mice. †P <0.01 relative to non-treated PDK4 knockout mice.
158 ± 6
134 ± 14
0.8 ± 0.1
235 ± 7*†
179 ± 16
0.8 ± 0.1
169 ± 9
130 ± 19
0.8 ± 0.1
225 ± 10*†
177 ± 24
0.8 ± 0.1
We originally thought that up-regulation of PDK4 by clofibric acid may be responsible for the beneficial effects of clofibric acid on serum and hepatic lipid levels. This was based on previous findings showing that clofibric acid and other PPARα ligands, e.g. WY-14643, increase PDK4 expression in major tissues of the body [14,15], and that PDK4 deficiency increases PDC activity, which increases pyruvate oxidation at the expense of fatty acid oxidation . However, the data presented here for PDK4 knockout mice show that induction of PDK4 expression is not necessary for the beneficial effects of clofibric acid on lipid levels. Indeed, clofibric acid is more effective in PDK4 knockout mice, suggesting that the induction of PDK4 that normally occurs in response to clofibric acid may counter the beneficial effects of this compound on lipid levels.
As expected, wild-type mice fed an HSF diet accumulated large amounts of fat and became obese. Clofibric acid prevented development of obesity by the mice. As reported previously , PDK4 deficiency also attenuated the weight gain induced by the HSF diet. Although not statistically significant, a trend toward further attenuation in weight gain was found in clofibric acid-treated PDK4 knockout mice compared to clofibric acid-treated wild-type mice. In agreement with these effects on body weight, clofibric acid, PDK4 deficiency and a combination of both factors reduced epididymal fat pad weight. As mice in the four groups consumed the same amounts of food, the lower weight gain in clofibric acid-treated wild-type and PDK4 knockout mice was not due to lower calorie intake. Indeed, food efficiency as determined by weight gain relative to food intake was markedly reduced in mice fed the diet supplemented with clofibric acid. As UCPs are induced by PPARα activation [16–21] and clofibric acid induced UCP3 significantly in wild-type mice and even more significantly in PDK4 knockout mice, dissipation of energy by heat as a consequence of uncoupling of oxidative phosphorylation may be responsible for controlling weight gain in the clofibric acid-treated mice.
Activation of PPARα by PPARα agonists increases the enzymatic capacity for fatty acid oxidation . Established mechanisms include induction of mitochondrial and peroxisomal fatty acid oxidation enzymes , induction of malonyl CoA decarboxylase, which relieves inhibition of carnitine palmitoyl transferase 1(CPT-1) by malonyl CoA , and expansion of the size of the CoA pool . Evidence that clofibric acid acts at least in part via these mechanisms was found in this study. However, there is also evidence that PPARα agonists can increase the capacity for de novo lipogenesis . Therefore, clofibric acid may also dissipate energy via a futile cycle of de novo fatty acid synthesis followed by fatty acid oxidation. However, loss of energy via this futile cycle would be reduced in clofibric acid-treated PDK4 knockout mice because fatty acid synthesis enzymes are reduced in PDK4 knockout mice .
Both clofibric acid and PDK4 deficiency reduced liver fat accumulation, and their combination further reduced such accumulation, suggesting that the effects of PDK4 deficiency and clofibric acid on hepatic steatosis are additive, with independent mechanisms. As expected, clofibric acid increased the level of MCAD, a key mitochondrial fatty acid oxidation enzyme. In agreement with the histological observations, PDK4 deficiency did not reduce the increased amounts of MCAD in the livers of clofibric acid-treated mice. Moreover, the level of ERRα, which is known to increase MCAD expression , was also increased in the livers of clofibric acid-treated wild-type and PDK4 knockout mice. As PDK4 deficiency reduces fat accumulation in the liver at least in part by decreasing the amounts of enzymes involved in fatty acid synthesis , the amount of ACC-1 was measured. Clofibric acid did not alter the amount of ACC-1, but PDK4 deficiency caused a significant reduction. The level of ACC-1 was also reduced in clofibric acid-treated PDK4 knockout mice. Both clofibric acid treatment and PDK4 deficiency increased the levels of serum ketone bodies. In the case of clofibric acid, this probably reflects the increase in enzymatic capacity for fatty acid oxidation. In PDK4 deficiency, the mechanism is less certain, but may be due to stimulation of ketone body formation from fatty acids as a consequence of inhibition of the citric acid cycle. Previous studies have shown that PDK4 deficiency limits the synthesis of oxaloacetate in the liver . PDK4 deficiency causes greater PDC activity, resulting in greater oxidation of pyruvate by peripheral tissues, thereby reducing the release of three-carbon compounds into the blood and thus the amount of pyruvate available for synthesis of oxaloacetate in the liver. Limiting the citric acid cycle by the availability of oxaloacetate directs acetyl CoA produced by fatty acid oxidation into ketone body formation. As oxidation of fatty acids via the citric acid cycle produces more ATP than oxidation of fatty acids to ketone bodies, larger amounts of fatty acids have to be oxidized to meet the energy needs of the liver in PDK4 knockout animals.
In summary, the findings reported here suggest that clofibric acid ameliorates body weight gain and hepatic steatosis by partially uncoupling oxidative phosphorylation and increasing the capacity for fatty acid oxidation. Up-regulation of PDK4 is not required for this effect. Indeed, PDK4 deficiency complements the action of clofibrate by decreasing the capacity for fatty acid synthesis, and perhaps also by promoting the oxidation of fatty acids to ketone bodies. These findings suggest that a PDK4 inhibitor and a fibrate could potentially be used in combination to reduce body weight and prevent hepatic steatosis in patients with type 2 diabetes or diet-induced obesity.
Experimental protocols were approved by the Animal Care and Use Committee of the Indiana University School of Medicine. PDK4 knockout mice were generated as previously described . At 5 weeks of age, groups of 12 male wild-type and 12 male PDK4 knockout mice were housed in cages (two mice per cage, both mice in each cage were of the same genotype) under controlled temperature (23 ± 2 °C) and a 12 h light/dark cycle (lights on at 7 am and off at 7 pm). The mice were fed a high saturated fat (HSF) diet (catalog number D12330; Research Diets, New Brunswick, NJ, USA) that was low in carbohydrate and high in saturated fat (58% by calories). According to the supplier, the fatty acids present in the HSF diet were 93.3% saturated, 2.4% monounsaturated and 4.3% polyunsaturated. After 14 weeks on the HSF diet, 0.5% clofibric acid (catalog number 197777; Sigma-Aldrich, St Louis, MO, USA) was added to the diet of half of the mice in each group. Based on their measured food consumption, the mice consumed approximately 10 mg of clofibric acid per day. The body weights of the mice were determined weekly. Food consumption was monitored during the 16th week of the feeding period. The experiment was terminated after 28 weeks of feeding. After overnight fasting (from 3 pm to 8 am), blood was taken from the tail for measurement of glucose, and the mice were anesthetized using pentobarbital (60 mg·kg−1 of body weight). Blood was drawn from the inferior vena cava to measure serum metabolites. Gastrocnemius muscles, liver, heart and epididymal fat pads were harvested as rapidly as possible in the order given, immediately freeze-clamped using Wollenberger tongs at the temperature of liquid nitrogen, powdered under liquid nitrogen using a mortar and pestle, and stored at −85 °C for analysis. Small pieces of the liver were also collected and snap-frozen in liquid nitrogen for histological analysis.
Measurement of metabolite concentrations
The levels of triacylglycerols in sera or tissues were determined using the L-type TG H assay kit from Wako Chemical (Richmond, VA, USA), and the levels of free fatty acids were determined using the Half Micro Assay kit from Roche Diagnostics [(catalog number 1 383 175) Indianapolis, IN, USA]. Tissue levels of CoA and acetyl CoA were determined as described by Michal and Bergmeyer .
Measurement of the PDC activity
Pulverized tissues were homogenized in an extraction buffer  containing 30 mm Hepes/KOH, pH 7.5, 0.5 mm thiamine pyrophosphate, 3% v/v Triton X-100, 5 mm EDTA, 2% v/v bovine serum, 5 mm dithiothreitol, 10 μm tosyl phenylalanyl chloromethyl ketone, 10 μg·mL−1 trypsin inhibitor, 1 μm leupeptin, 2 mm deoxycholic acid and 50 mm potassium fluoride. Supernatants were obtained by centrifugation at 10 000 g for 10 min at 4 °C. Total and actual PDC activities were measured as previously described . PDC activity is expressed as the percentage of actual activity relative to total activity.
Insulin tolerance test
At the 20th week of feeding, insulin tolerance tests were performed after food had been withheld from the mice for 5 h (9 am to 2 pm). Insulin (Humulin R, 1 U·kg−1 body weight; Eli Lilly, Indianapolis, IN, USA) was given by intraperitoneal injection. Tail blood glucose levels were measured at 0, 15, 30 and 60 min.
Histological examination of the liver was performed by the Immunohistochemistry Laboratory of Indiana University School of Medicine. The liver sections were stained using Oil Red O.
Western blot analysis
Tissue powder prepared under liquid nitrogen was homogenized with radio-immunoprecipitation assay buffer containing 50 mm Tris/HCl (pH 7.4), 150 mm NaCl, 0.25% v/v deoxycholic acid, 1% v/v Nonidet P-40 [(catalog number I-8896) Sigma-Aldrich, St Louis, MO, USA], 1 mm EDTA, 10 μm tosyl phenylalanyl chloromethyl ketone, 10 μg·mL−1 trypsin inhibitor, 1 mm phenylmethylsulfonyl fluoride, 2 μg·mL−1 aprotinin, 3.5 mm bis-benzamidine, 50 mm potassium fluoride and 0.4 mm sodium orthovanadate. Tissue extracts were obtained by centrifugation at 13 400 g for 10 min at 4 °C. Protein concentration was determined by the Bio-Rad assay [(catalog number 500-0006) Bio-Rad, Hercules, CA, USA]. Equal amounts of protein were separated by SDS/PAGE, transferred to a nitrocellulose membrane by the wet blotting method, and probed with antibodies directed against estrogen-related receptor α (ERRα; catalog number 07-662; Millipore, Billerica, MA, USA), medium-chain acyl CoA dehydrogenase (MCAD; catalog number sc-50587, Santa Cruz Biotechnology, Santa Cruz, CA, USA), acetyl CoA carboxylase 1 [ACC-1, (catalog number 07-439) Upstate, Lake Placid, NY, USA], tubulin [(catalog number ab6046) Abcam, Cambridge, MA, USA], PDK1 [(catalog number KAP-PK112) Stressgen, Plymouth Meeting, PA, USA], PDK2 (catalog number sc-100534, Santa Cruz Biotechnology), PDK3 (catalog number H00005165-M01; Abnova, Walnut, CA, USA) and PDK4 . The amounts of bound antibodies were assessed by the peroxidase activity of horseradish peroxidase-conjugated secondary antibody as detected by chemiluminescence using Lumi-light Western blotting substrate (Roche Diagnostics).
Quantitative real-time PCR analysis
Total RNA was isolated from frozen liver samples using an RNAqueous®-4PCR kit [(catalog number AM1914) Applied Biosystems/Ambion, Austin, TX, USA]. RNA integrity was checked by agarose gel electrophoresis. Validated primer sets for DNA amplification were obtained from SA Biosciences (Frederick, MD, USA). The Genbank accession numbers for uncoupling protein 3 and tubulin β1 are NM_009464 and NM_001080971, respectively.
Using Brilliant II SYBR® Green QRT-PCR, AffinityScript Master Mix, 2-Step reagent [(catalog number 600834) Stratagene, La Jolla, CA, USA], 3 μg of RNA from each sample were reverse-transcribed to cDNA, and then an equal amount of cDNA (determined using a standard curve generated for each gene analyzed) was used as template for quantitative PCR using the Mx3005P quantitative PCR system (Stratagene). A threshold cycle (Ct) value was determined for each amplification curve. After normalization to the Ct value obtained for tubulin β1 in the same samples, the resulting ΔCt values were used to obtain the relative fold change in expression of the indicated gene in each treated group relative to that in the control group.
Values are mean ± SE for the indicated number of independent samples. The statistical significance of differences between groups was determined by one-way ANOVA corrected by the Tukey–Kramer method or by Student’s t test. P values <0.05 were considered to be statistically significant.
We thank Dr N.H. Jeoung at the Department of Fundamental Medical and Pharmaceutical Sciences in the Catholic University of Daegu (South Korea) for helpful discussion. This work was supported by grants to R.A.H. from the National Institutes of Health (DK47844) and a VA Merit Review Grant.