Insulin-resistant states, including those involving obesity, are often associated with abnormal lipid handling, including increased tissue fatty-acid delivery and oxidation. The development of insulin resistance in such conditions often reflects elevated oxidation of lipid fuels and/or intracellular accumulation of lipid derivatives [e.g., triacyglycerol and long-chain acyl-coenzyme A (CoA)1]; these interfere with insulin signaling ((1, 2); reviewed in Ref. (3)). Increased mitochondrial fatty-acid oxidation suppresses the use of glucose as the main oxidative substrate (the glucose-fatty acid cycle) (reviewed in Ref. (4)), leading to impaired glucose disposal. Substrate competition between fatty acids and glucose occurs at the level of the pyruvate dehydrogenase complex (PDHC), which catalyzes the oxidative decarboxylation of pyruvate to form acetyl-CoA and so links glycolysis to the tricarboxylic acid cycle and adenosine triphosphate (ATP) production (reviewed in Ref. (5)). PDHC is normally active in most tissues in the well-fed state. However, because no pathway exists for the conversion of acetyl-CoA to glucose in mammals, suppression of PDHC activity is crucial for conserving 3-C compounds for glucose synthesis when glucose is scarce. Suppression of PDHC activity in skeletal muscle is of particular importance because of the large contribution of the muscle mass to whole-body glucose disposal (6). Conversely, inappropriate suppression of PDHC activity and flux in skeletal muscle, by abnormal partitioning of pyruvate between lactate (or alanine) production and oxidation, would be predicted to promote the development of hyperglycemia by fueling excessive gluconeogenesis.
The activity of PDHC is acutely controlled by allosteric inhibition by increasing the mitochondrial ratios of the reduced form of nicotinamide adenine dinucleotide/the oxidized form of nicotinamide adenine dinucleotide+ (NADH/NAD+) and acetyl-CoA/CoA concentration ratios. Such increases can be attributable to either an imbalance between PDHC activity and ATP requirement or a high rate of fatty-acid oxidation. Control of PDHC activity is also achieved through a reversible phosphorylation-dephosphorylation cycle involving the phosphorylation of up to 3 serine residues (designated sites 1, 2, and 3) on one of its component enzymes (pyruvate dehydrogenase; E1). Inactivation of PDHC by phosphorylation, primarily of site 1, is catalyzed by pyruvate dehydrogenase kinases (PDKs), which, although structurally similar to the prokaryotic histidine kinases, have a catalytic mechanism similar to the eukaryotic serine kinases (reviewed in Ref. (7)). Phosphorylation of site 1 renders PDHC completely devoid of activity, and thereby imposes an upper limit on PDHC flux in proportion to the percentage of the total number of PDHC molecules that are phosphorylated. PDK is subject to short-term regulation (activation) by mitochondrial acetyl-CoA and NADH, both products of the PDHC reaction and of fatty acid β-oxidation. In this manner, PDHC inactivation by phosphorylation is coordinated with PDHC allosteric inhibition. PDK is also subject to inhibition by pyruvate, which can be generated from glycolysis or from circulating lactate. Hence, in tissues that can rapidly modulate their rates of glycolysis, including fast-twitch muscle, which increases its glycolytic rate many-fold on recruitment to contraction, increased rates of pyruvate production can have a feed-forward action to augment glucose oxidation.
Recent studies have shown that differential expression of PDK isoforms and PDK isoform shifts in individual tissues may modulate PDHC activity according to sustained changes in the tissue requirements for ATP production or lipid clearance. To date, four PDK isoenzymes have been identified in humans and rodents (designated PDK1, PDK2, PDK3, and PDK4) (7, 8). In several oxidative tissues, including liver and skeletal muscle, PDK activity is modified in response to altered nutritional and endocrine status. Studies in animal models have indicated that this can often reflect increased expression of PDK2 and/or PDK4. Increased lipid delivery and/or handling invariably results in the increased expression of the specific PDK isoenzyme PDK4. Both PDK activity and the level of PDK4 expression increase after prolonged food withdrawal (9, 10, 11, 12, 13) in response to a diet high in saturated fat (14, 15) and in insulin-deficient diabetes (9). Longer-term increases in PDK4 protein may be important for increasing the phosphorylation not only of site 1 of E1, causing PDHC inactivation, but also of site 2 of E1. Although phosphorylation of site 2 does not cause PDHC inactivation, it may lock PDHC into an inactive form. Reactivation of PDHC in liver, heart, and skeletal muscle is much slower when refeeding after prolonged starvation, when PDK4 expression is enhanced, than after acute starvation (16, 17, 18). In addition, studies with recombinant PDK isoforms have demonstrated that PDK4 activity is more responsive to activation by increased NADH/NAD+ ratios and less responsive to suppression by the pyruvate analogue dichloroacetate (8). Decreased sensitivity of PDHC activity to increasing pyruvate shows a close correlation with increased PDK4 protein expression in skeletal muscle (14, 19). This potentially means that increased glycolytic flux may exert a relatively diminished influence to increase glucose oxidation, resulting in increased pyruvate availability for lactate formation and regeneration of NAD+ for continuation of glycolysis. It could also be hypothesized that, on initiation of anaerobic exercise, fatty acids could compete more effectively with glucose for a better share of a diminished oxygen supply. In such a way, changes in PDK4 expression may modulate metabolic protection against subtotal anoxia by optimizing provision of glycolytic ATP. However, as a downside, the oxidation of fatty acids requires more oxygen than that of glucose (which may be critical during suboptimal tissue perfusion), and increased glycolysis could cause intracellular acidosis (and thus, poor exercise tolerance) and increased generation of 3-C precursors, which, in turn, could contribute to increased rates of glucose production.
The concept that diminished effects of insulin, characteristic of type 2 diabetes mellitus and obesity, might be linked to PDK expression was first seen in nondiabetic Pima Indians, in whom obesity and type 2 diabetes mellitus is prevalent. In this genetically homogeneous group, muscle PDK mRNA expression correlates positively with fasting plasma insulin and negatively with insulin-mediated glucose uptake (20). In the study by Rosa et al. in this issue of Obesity Research (21), evidence is presented that environmental, as well as genetic, factors may be important. Their data show that, at the molecular level, skeletal muscle PDK remains in the “starvation” mode, even though plentiful nutrients are available, as signaled by the development of obesity and the accumulation of muscle triacylglycerol. Even more significantly, changes in PDK4 mRNA expression in skeletal muscle were shown to correlate with an improvement in insulin sensitivity in morbidly obese patients who had undergone malabsorptive bariatric surgery (bilio-pancreatic diversion), showing a link between PDK4 expression and fat mass. Importantly, there was a significant negative correlation between whole-body glucose uptake and PDK4 mRNA expression in skeletal muscle. This study, therefore, demonstrates, for the first time, that the expanded fat mass in obesity leads to metabolic adaptations in muscle that would be predicted to promote fatty-acid oxidation in this tissue, at the expense of uncoupling of nonoxidative and oxidative glucose degradation. Enhanced PDK4 expression may, therefore, be viewed as a potential compensatory mechanism to counter excessive formation of intracellular lipid and exacerbation of impaired insulin sensitivity. The good news is that reversal of obesity and insulin resistance can allow a decline in PDK4 expression in skeletal muscle, thereby facilitating glucose clearance and allowing an appropriate balance between glucose and lipid use. With improved glucose oxidation, it would be predicted that proton production during exercise would be reduced, allowing greater exercise tolerance as well as modest oxygen-sparing effect. This study, therefore, reinforces the concept that environmental, or possibly pharmacological, interventions that modulate PDK4 expression could have significant beneficial outcomes for the control of nutrient handling in obesity.