Pantothenate kinase 4 controls skeletal muscle substrate metabolism

Abstract

Metabolic flexibility in skeletal muscle is essential for maintaining healthy glucose and lipid metabolism, and its dysfunction is closely linked to metabolic diseases. Exercise enhances metabolic flexibility, making it an important tool for discovering mechanisms that promote metabolic health. Here we show that pantothenate kinase 4 (PanK4) is a new conserved exercise target with high abundance in muscle. Muscle-specific deletion of PanK4 impairs fatty acid oxidation which is related to higher intramuscular acetyl-CoA and malonyl-CoA levels. Elevated acetyl-CoA levels persist regardless of feeding state and are associated with whole-body glucose intolerance, reduced insulin-stimulated glucose uptake in glycolytic muscle, and impaired glucose uptake during exercise. Conversely, increasing PanK4 levels in glycolytic muscle lowers acetyl-CoA and enhances glucose uptake. Our findings highlight PanK4 as an important regulator of acetyl-CoA levels, playing a key role in both muscle lipid and glucose metabolism.

Fig. 1. PanK4 is an exercise target abundant in skeletal muscle and associated with metabolic dysregulation.

A PanK4 Ser63 phosphorylation (p-PanK4^Ser63) and its regulation by exercise or muscle contraction in skeletal muscle from humans and rodents indicated by phosphoproteomics^,. p-PanK4^Ser63 in human vastus lateralis before (Pre) and post-exercise indicated time-points into recovery from vigorous cycling exercise (B, n = 6), prolonged endurance cycling, sprint cycling and strength exercises (C, n = 8). D Pank4 mRNA in indicated male mouse tissues (n = 6). E mRNA abundance of indicated Pank isoforms determined by whole genome transcriptome analysis in male mouse skeletal muscle (n = 8). F abundance of PanK1, PanK2 and PanK4 protein in indicated tissues from male mice and representative Western blots (n = 4). D–F data are presented as mean values +/− SEM. Statistic: A significantly regulated phosphopeptides (adjusted P-values less than 0.05) were identified in larger data sets using LIMMA’s moderated t-statistics with the eBayes method (see Refs. ^, for more detail). B repeated measures one-way ANOVA and C repeated measures two-way ANOVA with log2-transformed data and Šidák post hoc testing. Source data are provided as a Source Data file. AU = arbitrary units. Schematic in Fig. 1A was created in BioRender. Kleinert, M. (2024) BioRender.com/t99l144.

Fig. 2. Skeletal Muscle PanK4 regulates fatty acid oxidation.

A PanK4 protein abundance in indicated tissues from male PanK4 wildtype (WT) and muscle-specific PanK4 knockout (mKO) mice at age 28 weeks. B body composition at age 26 weeks (n = 8 for WT and n = 6 for mKO) of male PanK4 WT and PanK4 mKO mice. C, D, volcano plot of 508 detected metabolites and heatmap of indicated metabolites in TA muscle from male glucose-stimulated PanK4 WT and PanK4 mKO mice at age 28 weeks (n = 8 for WT and n = 6 for mKO). E overview of the Coenzyme A biosynthesis pathway. F palmitate oxidation in ex vivo basal or contracted soleus muscles from ad libitum fed male PanK4 WT and PanK4 mKO mice at age 12-19 weeks (n = 10). Citrate synthase activity and muscle triglyceride (TG) concentrations in soleus (G) and gastrocnemius (H) muscles from male glucose-stimulated PanK4 WT and PanK4 mKO mice at age 28 weeks (n = 8 for WT and n = 6 for mKO). Statistic: B two-tailed unpaired t-test within fat or lean mass; C, D two-tailed welch test; F two-way ANOVA; G, H, two-tailed unpaired t-test. Statistical analyses were conducted with log2-transformed data. B, F–H, data are presented as mean values +/− SEM. PanK1-3 pantothenate kinase 1-3, PPCS phosphopantothenate-cysteine ligase, PPCDC phosphopantothenoylcysteine decarboxylase, PPAT phosphopantetheine adenylyltransferase, DPCK = 3’-Dephospho-CoA kinase. Source data are provided as a Source Data file.

Fig. 3. PanK4 regulates acetyl-CoA and metabolic flexibility in skeletal muscle.

A acetyl-CoA determined by non-targeted metabolomics in indicated skeletal muscle from male PanK4 WT and PanK4 mKO mice weeks (n = 8 for WT and n = 6 for mKO). B targeted analysis of acetyl-CoA in gastrocnemius muscle from fasted and fed male (circles) and female (squares) PanK4 WT and PanK4 mKO mice (n = 5 for fast/WT, n = 4 for fed/WT, n = 4 for fast/mKO n = 4 for fed/mKO). C representative Western blots and quantification of pan-lysine acetylation in gastrocnemius muscle from male (circles) and female (squares) PanK4 WT and PanK4 mKO mice (n = 5). D–F targeted analysis of citrate and malonyl-CoA in gastrocnemius muscle from fasted and fed male (circles) and female (squares) (E, n = 5 for fast/WT, n = 4 for fed/WT, n = 4 for fast/mKO n = 4 for fed/mKO; F, n = 5) and in indicated muscle from fed male (G, n = 12) PanK4 WT and PanK4 mKO mice. G representative Western blots and quantification of pan-lysine malonylation in gastrocnemius muscle from male (circles) and female (squares) PanK4 WT and PanK4 mKO mice (n = 5). H activity of pyruvate dehydrogenase in the active form (PDHa) activity in gastrocnemius muscle from male (circles) and female (squares) PanK4 WT and PanK4 mKO mice (n = 7 for fast/WT, n = 7 for fed/WT, n = 6 for fast/mKO n = 7 for fed/mKO). A–H data are presented as mean values +/- SEM. Statistic: A–H, two-way ANOVA. Statistical analyses were conducted with log2-transformed data. Šidák post hoc testing was performed. Source data are provided as a Source Data file. AU = arbitrary units.

Fig. 4. Skeletal muscle PanK4 regulates whole-body insulin sensitivity and muscle glucose uptake.

A, B glucose tolerance in male PanK4 WT and PanK4 mKO mice at indicated ages (n = 8 for WT and n = 6 for mKO). C fasting plasma insulin concentrations in male PanK4 WT and PanK4 mKO mice (n = 8 for WT and n = 6 for mKO). D ex vivo basal- or insulin (3.0 nM)-stimulated glucose uptake in skeletal muscle (EDL) from male PanK4 WT and PanK4 mKO mice (n = 8 for WT/basal, n = 8 for WT/insulin, n = 8 for mKO/basal, n = 7 for mKO/insulin). Maximal running speed (E) and maximal distance (F) during treadmill running of male PanK4 WT and PanK4 mKO (n = 9). G glucose uptake into indicated skeletal muscle during 20 min of rest or treadmill running at 75% of maximal running capacity in male PanK4 WT and PanK4 mKO mice (rest: n = 9 for WT/quad, n = 9 for WT/triceps, n = 10 for WT/gastroc, n = 7 for mKO; running: n = 9 for WT, n = 8 for mKO/quad, n = 6 for mKO/triceps, n = 7 for mKO/gastroc). H acetyl-CoA levels determined by targeted analyses in quadriceps muscle from male PanK4 WT and PanK4 mKO mice that had ran on the treadmill (n = 11 for WT and n = 10 for mKO). Male C57BL/6J mice aged 12–16 weeks were treated with recombinant adeno‐associated virus serotype 6 encoding PanK4 (rAAV6:PanK4) injected into  tibialis anterior (TA) muscle, while contralateral TA was injected with rAAV6:MCS as a control: PanK4 protein abundance (I) and glucose uptake (J, n = 13) in TA muscles were determined 14 days later in fed mice. K acetyl-CoA detected by non-targeted metabolomics in fasted or refed male C57BL6/J mice overexpressing PanK4 in one TA and MCS in contralateral TA (n = 4 for MCS/fast, n = 6 for MCS/fed, n = 4 for PanK4/fast, n = 6 for PanK4/fed). A–H, J, K data are presented as mean values +/− SEM. Statistic: A, B repeated measures two-way (time x genotype) ANOVA; C two-tailed student’s t-test; D two-way (genotype x insulin) ANOVA; G two-way ANOVA (genotype x muscle) within rest or run; H two-tailed student’s t-test. J two-tailed student’s t-test; K two-way (genotype x feeding) ANOVA. Statistical analyses were conducted with log2-transformed data. Šidák post hoc testing. Source data are provided as a Source Data file. Schematic in A was created in BioRender. Kleinert, M. (2024) BioRender.com/m45y371; schematic in I was created in BioRender. Kleinert, M. (2024) BioRender.com/l81e607.