5'-AMP activated protein kinase α2 controls substrate metabolism during post-exercise recovery via regulation of pyruvate dehydrogenase kinase 4

Abstract

It is well known that exercise has a major impact on substrate metabolism for many hours after exercise. However, the regulatory mechanisms increasing lipid oxidation and facilitating glycogen resynthesis in the post-exercise period are unknown. To address this, substrate oxidation was measured after prolonged exercise and during the following 6 h post-exercise in 5´-AMP activated protein kinase (AMPK) α2 and α1 knock-out (KO) and wild-type (WT) mice with free access to food. Substrate oxidation was similar during exercise at the same relative intensity between genotypes. During post-exercise recovery, a lower lipid oxidation (P < 0.05) and higher glucose oxidation were observed in AMPKα2 KO (respiratory exchange ratio (RER) = 0.84 ± 0.02) than in WT and AMPKα1 KO (average RER = 0.80 ± 0.01) without genotype differences in muscle malonyl-CoA or free-carnitine concentrations. A similar increase in muscle pyruvate dehydrogenase kinase 4 (PDK4) mRNA expression in WT and AMPKα2 KO was observed following exercise, which is consistent with AMPKα2 deficiency not affecting the exercise-induced activation of the PDK4 transcriptional regulators HDAC4 and SIRT1. Interestingly, PDK4 protein content increased (63%, P < 0.001) in WT but remained unchanged in AMPKα2 KO. In accordance with the lack of increase in PDK4 protein content, lower (P < 0.01) inhibitory pyruvate dehydrogenase (PDH)-E1α Ser(293) phosphorylation was observed in AMPKα2 KO muscle compared to WT. These findings indicate that AMPKα2 regulates muscle metabolism post-exercise through inhibition of the PDH complex and hence glucose oxidation, subsequently creating conditions for increased fatty acid oxidation.

Figure 1. AMPKα₂ KO mice are exercise intolerant and display reduced FA oxidation during post‐exercise recovery

Maximal running speed test in AMPKα₂ KO and WT (A) and AMPKα₁ KO and WT mice (B). Respiratory exchange ratio (RER) during treadmill running at 50% of maximal running speed and in the following 6 h post‐exercise recovery period (C and D) and the average RER during 6 h of post‐exercise recovery (E and F) in AMPKα₂ KO and WT (C and E) and AMPKα₁ KO and WT (D and F). G, fatty acid oxidation calculated as described in Methods in AMPKα₂ KO and WT during 6 h of post‐exercise recovery. Data are presented as means ± SEM. n = 8. *P < 0.05, ***P < 0.001 significantly different from WT (main effect of genotype during recovery in C). WT, wild‐type; KO, knock‐out.

Figure 2. Signalling in AMPKα₂ KO and WT skeletal muscle

A, AMPKα₂ protein was absent in AMPKα₂ KO muscle. AMPKα₁ protein content (B), AMPK Thr¹⁷² phosphorylation (C) and ACC Ser²¹² phosphorylation/ACC protein (D) in quadriceps muscle at rest, immediately after exercise, and 6 h post‐exercise recovery. E, representative immunoblots. Data are presented as means ± SEM. n = 7–10. **P < 0.01, ***P < 0.001 significantly different from WT or main effect of genotype. ^# P < 0.05 main effect of exercise, ^### P < 0.001 significantly different from rest within WT. AU, arbitrary units; WT, wild‐type; KO, knock‐out. From one litter, WT and AMPKα₂ KO mice were allocated into a basal, resting group (Rest 1/R1) and an exercise group (Ex). From another litter, WT and AMPKα₂ KO were allocated into a basal, resting group (Rest 2/R2) and a recovery group (Rec).

Figure 3. Content of glycogen (A), malonyl‐CoA (B), free carnitine (C) and free CoA (D) in quadriceps muscle at rest, immediately after exercise, and 6 h post‐exercise recovery

Data are presented as means ± SEM. n = 7–10. ^# P < 0.05, ^### P < 0.001 main effect of exercise. ***P < 0.001 main effect of genotype. w.w., wet weight; d.w., dry weight. From one litter, WT and AMPKα₂ KO mice were allocated into a basal, resting group (Rest 1) and an exercise group. From another litter, WT and AMPKα₂ KO were allocated into a basal, resting group (Rest 2) and a recovery group.

Figure 4. AMPKα₂ KO mice have blunted exercise‐induced PDK4 protein expression and lower NADH content in muscle

PDK4 protein expression (A), PDH‐E1α Ser²⁹³ phosphorylation/PDH‐E1α protein (B), PDH‐E1α Ser³⁰⁰ phosphorylation/PDH‐E1α protein (C), NADH content (D) and acetyl‐CoA content (E) in quadriceps muscle at rest, immediately after exercise, and 6 h post‐exercise recovery. F, representative immunoblots. Data are presented as means ± SEM. n = 7–10. ^## P < 0.01, ^### P < 0.001 main effect of exercise (D) and significantly different from rest within WT (A); *P < 0.05, **P < 0.01 main effect of genotype; ***P < 0.001 significantly different from WT within exercise. AU, arbitrary units. From one litter, WT and AMPKα₂ KO mice were allocated into a basal, resting group (Rest 1/R1) and an exercise group (Ex). From another litter, WT and AMPKα₂ KO were allocated into a basal, resting group (Rest 2/R2) and a recovery group (Rec).

Figure 5. Similar exercise‐induced increase in PDK4 mRNA levels, p53 acetylation and decreased HDAC4 phosphorylation in AMPKα₂ KO and WT muscle

PDK4 mRNA levels (A), NAD⁺ content (B), p53 Lys³⁷⁹ acetylation/p53 protein (C), HDAC4 Ser⁶³² phosphorylation/HDAC4 protein (D), and miR‐107 gene expression (E) in quadriceps muscle at rest and immediately after 2 h of treadmill exercise. F, representative immunoblots. Data are presented as means ± SEM. n = 7–10. ^# P < 0.05, ^## P < 0.01, ^### P < 0.001 main effect of exercise. AU, arbitrary units; Ex, exercise; d.w., dry weight; Re, rest.

Figure 6. Scheme of proposed AMPK‐mediated regulation of fuel selection during post‐exercise recovery

Exercise‐induced increase in AMPKα₂ activity in skeletal muscle increases pyruvate dehydrogenase 4 (PDK4) protein content, and AMPKα₂ seems also to be crucial for NADH levels. Both inhibit pyruvate dehydrogenase (PDH) activity, whereby conversion of pyruvate to acetyl‐CoA is inhibited and consequently glucose oxidation (GLU OX). This enables increased fatty acid oxidation (FA OX).