PDK4 drives time-dependent physical endurance impairment during acute hypoxia by disrupting carbohydrate metabolism

Abstract

Background

Exposure to acute hypoxia severely impairs skeletal muscle function, but the temporal dynamics and underlying mechanisms remain unclear.

Methods

Male C57BL/6 mice were exposed to normobaric hypoxia (FiO₂ = 11.8%, simulating 4,500 m) for 0, 12, 24, 48, and 72 h. Exercise performance was evaluated through endurance and strength tests. Transcriptomic and metabolic adaptations were detected through RNA-seq and biochemical assays. Target validation was performed using adeno-associated virus (AAV) mediated gene knockdown and dichloroacetate (DCA).

Results

Endurance was most severely impaired after 12 h of hypoxic exposure and showed a trend toward recovery by 72 h, whereas muscle strength declined progressively. Hypoxia induced a metabolic reprogramming favoring oxidative phosphorylation, coupled with a marked upregulation of PDK4 at 12 h. This was associated with increased pyruvate dehydrogenase (PDH) phosphorylation, favoring the diversion of pyruvate away from oxidation and contributing to lactate accumulation. Both DCA treatment and AAV-mediated PDK4 knockdown improved endurance performance under hypoxic conditions.

Conclusion

These findings suggest that acute hypoxia impairs endurance, in part through PDK4-mediated alterations in pyruvate metabolism, providing a mechanistic basis for optimizing nutritional and therapeutic strategies under hypoxic stress.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12986-026-01155-8.

Introduction

Oxygen is the key substance that supports energy metabolism during high-intensity and endurance exercise. Natural high-altitude environments and artificially simulated hypoxic chambers both pose unique challenges to human physiology [1, 2]. In recent years, continuous advancements in sports science and medical research have significantly expanded our understanding of this field. Concurrently, the potential of pharmaceuticals, nutritional supplements, and other non-traditional interventions to mitigate hypoxic stress and prevent altitude sickness has received considerable attention [3–5]. Nevertheless, the detrimental effects of hypoxic environments on athletic performance, including acute reductions in exercise capacity, increased fatigue, and altitude sickness, remain significant [1, 3, 6]. Consistent research findings confirm that acute hypoxic exposure significantly diminishes endurance performance, as evidenced by prolonged completion times or reduced average power output in tests such as the 4-kilometer cycling time trial [7], 15-kilometer time trial [8], 5-kilometer time trial [9], and the longer 40-minute maximal cycling test [10]. Reports suggest that for every 1,000 m increase in altitude, maximum oxygen uptake decreases by around 6.5% [11]. Therefore, understanding how hypoxia affects athletic performance and exploring effective enhancement strategies is of significant theoretical and practical importance to athletes, military personnel, mountaineers, residents of high-altitude regions, and rescue workers.

The decline in exercise ability caused by acute hypoxia is a complex physiological process involving multiple systems throughout the body. Skeletal muscle accounts for 40–60% of adult body weight [12], and as the ultimate executive organ of exercise, it is an important target organ for improving exercise performance. Traditionally, it was thought that a low oxygen environment forced skeletal muscle energy metabolism to rely more on carbohydrate, as more ATP can be produced per unit oxygen consumption [13, 14]. However, recent studies have found that the intake and oxidation of carbohydrates by skeletal muscles are inhibited during acute hypoxia exposure. Young measured exogenous carbohydrate oxidation in a population exposed to an altitude of 4,300 m using 13C tracing and indirect calorimetry [15]. The results showed that exogenous carbohydrate oxidation decreased by 20–50% during the first 12 h of exposure to exercise. Margolis’ study also found that exogenous carbohydrate oxidation was reduced [16]. Midha detected glucose uptake and oxidation in different mouse tissues after hypoxic exposure, and the results showed that the uptake and oxidation of glucose in skeletal muscle significantly decreased during acute hypoxic exposure [17]. The molecular mechanisms behind this seemingly contradictory phenomenon are currently unclear. Therefore, exploring the mechanism by which acute hypoxia exposure leads to insufficient glucose uptake and utilization in skeletal muscles may be a key approach to improving hypoxia adaptation and exercise performance.

Skeletal muscle is the primary organ responsible for energy metabolism in the human body. Its ability to efficiently and flexibly utilize energy substrates during exercise is a key factor in determining exercise performance and endurance. The selection and conversion process of this energy substrate is called metabolic flexibility, and its core lies in precise regulation of carbohydrate oxidation pathways. The pyruvate dehydrogenase complex (PDC) is a key node connecting glycolysis and the mitochondrial tricarboxylic acid cycle (TCA) and is known as the gatekeeper of carbohydrate oxidation. Its activity directly determines whether pyruvate derived from glucose can enter the mitochondria for oxidation and energy production [18, 19]. Pyruvate dehydrogenase kinase (PDK), particularly the dominant skeletal muscle isoform PDK4, inhibits PDC activity by phosphorylating pyruvate dehydrogenase (PDH), thereby limiting carbohydrate oxidation [20]. An increase in PDK4 activity or upregulation of its expression enhances the phosphorylation of specific serine sites on the PDH E1α subunit. This leads to a decrease in the overall activity of the PDC and effectively closes the pathway for carbohydrates to enter mitochondrial oxidation [20, 21]. Therefore, the PDK4-PDH regulatory axis constitutes a molecular switch for skeletal muscle substrate selection. However, it has not been confirmed whether skeletal muscle PDK4 is expressed rapidly in response to acute hypoxia exposure or whether its upregulation is sufficient to explain the inhibition of glucose oxidation and the subsequent decrease in exercise endurance.

In this study, we investigated the impact of acute hypoxia on physical performance and skeletal muscle metabolism. We hypothesized that the impairment of exercise capacity induced by acute hypoxic exposure is associated with disturbances in glucose metabolism mediated by the PDK4-PDH axis in skeletal muscle.

Materials and methods

Reagents

The standard diet (XTI01WC-009), purchased from Jiangsu Xietong Pharmaceutical Bio-engineering Co. (Nanjing, Jiangsu, China), had a nutritional composition of 11.1% fat, 67.4% carbohydrate, and 21.5% protein. Adeno-associated virus (AAV) vectors were constructed by Heyuan Co. (Shanghai, China). Sodium dichloroacetate (DCA) was obtained from Shanghai Yuanye Bio-Technology Co. (Shanghai, China). Lactate test strips were procured from EKF Diagnostics (Magdeburg, Germany). Antibodies against PDK4 (12949-1-AP) and PDH E1 alpha (18068-1-AP) were purchased from Proteintech (Chicago, USA). The assay kits of L-lactate (BC2235), pyruvate (BC2205), glycogen (BC0345), and PDH activity (BC0380) were obtained from Solaibao (Beijing, China). The assay kit of glucose 6-phosphate (G0825F) was obtained from Grace (Jiangsu, China). Phospho-PDH (Ser293 of the E1-alpha subunit, bs-4036R) antibody and acetyl-CoA assay kit (AK340) were purchased from Bioss (Beijing, China).

Animals and experimental protocol

Male C57BL/6 mice (8 weeks old, 22–24 g) were obtained from Chongqing Tengxin Biological Co., Ltd. (China). Mice were housed under standard conditions (22–25 °C, 50–55% humidity, 12 h/12 h light/dark cycle) with free access to food and water. Body weight and food intake were recorded daily.

Experiment 1: Effects of hypoxia on physical performance and skeletal muscle metabolism

To determine the effects of hypoxia on physical performance and skeletal muscle metabolism, mice were randomly divided into five groups (n = 16): normoxic control group (CON) and four groups exposed to hypoxia for 12 h (H12h), 24 h (H24h), 48 h (H48h) and 72 h (H72h). After the corresponding hypoxic exposure, 8 mice from each group were randomly selected for blood collection from the tail vein to measure blood lactate concentrations under resting conditions, followed by tissue collection for subsequent analyses. The remaining 8 mice in each group were subjected to grip strength test, treadmill exhaustion test (as described by Huang et al. [22]; speed: 25 m/min; slope: 15°), and in situ isometric muscle function measurement to evaluate changes in physical performance. Notably, mice in the normoxic control group were tested under normoxic conditions (FiO₂ = 20.9%), whereas those in the hypoxic groups were tested in the hypoxic chamber (FiO₂ = 11.8%).

Experiment 2

Role of PDK4 in acute hypoxia-induced reduction of physical endurance

To investigate the role of PDK4 in the reduction of physical endurance induced by acute hypoxia, we employed adeno-associated virus (AAV)-mediated skeletal muscle-specific PDK4 knockdown and pharmacological intervention with the PDK inhibitor sodium dichloroacetate (DCA).

1. For the AAV treatment experiment, mice were randomly divided into two groups (n = 8) and received intramuscular injections of AAV vectors into the gastrocnemius and tibialis anterior muscles (total dose: 2 × 10¹¹ viral genomes per mouse, delivered across eight injection sites, 10 µL per site). The AAVs included the following constructs: pAAV-U6-shRNA(NC2)-CMV-mScarlet-WPRE (CON), and pAAV-U6-shRNA(Pdk4)-CMV-mScarlet-WPRE (AAV-shPdk4). Following a 4-week period for viral transduction and expression, mice were exposed to acute hypoxia for 12 h. After hypoxic exposure, the treadmill exhaustion test was performed under sustained hypoxic conditions. Immediately after exhaustion, blood glucose and lactate concentrations were measured via the tail vein, followed by anesthesia and tissue collection.

2. For the DCA intervention, mice were randomly assigned to four groups (n = 8): normoxic control with saline (NC + Sal), normoxic control with DCA (NC + DCA), 12 h hypoxic exposure with saline (Hyp12 + Sal), and 12 h hypoxic exposure with DCA (Hyp12 + DCA). DCA was dissolved in saline, and mice received intraperitoneal injections of either DCA (50 mg/kg/d) or saline vehicle for two weeks. The Hyp12 + Sal and Hyp12 + DCA groups were then exposed to acute hypoxia for 12 h. After the exposure, resting blood glucose and lactate concentrations were measured from the tail vein. Subsequently, all mice underwent a treadmill exhaustion test, with the NC + Sal and NC + DCA groups tested under normoxic conditions and the Hyp12 + Sal and Hyp12 + DCA groups tested under sustained hypoxic conditions. Tissue samples were collected after the exercise test.

During the first week of the experiment, mice from all groups completed an adaptive training protocol on a motorized treadmill (speed: 15 m/min; duration: 10 min) [23]. Subsequently, body composition (including body weight, free water content, fat content, and muscle content) was assessed using an EchoMRI composition analyzer to ensure group homogeneity. Hypoxic conditions were achieved using a normobaric hypoxic chamber (HYPOXICO, USA) with an FiO₂ of 11.8%, simulating an altitude of 4,500 m. Before sampling, mice were anesthetized with Avertin and subsequently euthanized by cervical dislocation. Serum, gastrocnemius muscle, and quadriceps muscle were collected and stored at -80℃. All animal experiments were approved by the Animal Care and Use Committee of Third Military Medical University (Army Medical University), Chongqing, China (Approval No. AMUWEC20228026).

Grip strength test

Grip strength was measured using a grip strength meter (Ugo Basile, Italy). Each mouse underwent three consecutive tests, with a one-minute rest interval. The maximum grip strength was recorded as the average value of the three measurements. Relative grip strength was calculated as grip strength (g) divided by body weight (g).

In situ isometric muscle function measurement

In situ isometric muscle function of the gastrocnemius muscle was assessed using a muscle testing system (1300A 3-in-1 Whole Animal System, Canada). Mice were anesthetized and secured on the testing platform. A baseline stimulation frequency of 30 Hz was applied, and the stimulation current was gradually increased from 1 to 10 mA until the force output reached a plateau. The current at which the plateau was achieved was defined as the optimal stimulation intensity, which was subsequently fixed at 8 mA. To evaluate force production at different stimulation frequencies, electrical stimuli were applied at progressively increasing frequencies (10, 20, 40, 50, 60, 80, 100, 120, and 140 Hz). A 1-minute rest interval was allowed between each contraction. The maximum value recorded during the procedure was considered the maximal isometric force.

Biochemical analysis

Serum samples (40 µL) were diluted 1:4 with phosphate-buffered saline (PBS). Subsequently, the levels of glucose (GLU), alanine aminotransferase (ALT), aspartate transaminase (AST), blood urea, lactate dehydrogenase (LDH) and creatine kinase (CK) were measured using an automatic biochemical analyzer (Beckman, USA). The contents of several metabolites in muscle tissues were measured using specific commercial assay kits according to the manufacturers’ protocols.

Histological analysis

Gastrocnemius muscle samples were fixed in muscle fixation fluid (Servicebio, China) for 24 h. After dehydration and paraffin embedding, transverse sections of 4 μm thickness were prepared and stained with hematoxylin and eosin (H&E). The stained sections were scanned using a digital tissue section scanner (CQSQF-12, China) and quantitatively analyzed with ImageJ software. For immunofluorescence staining, paraffin embedded gastrocnemius tissue sections (4 μm) were subjected to antigen retrieval. The sections were placed in EDTA retrieval solution (Servicebio, China) in a water bath and heated at 90℃ for 30 min. The sections were then blocked with 3% (w/v) bovine serum albumin (BSA) in TBST for 30 min at room temperature and incubated overnight at 4℃ with a rabbit polyclonal anti-PDK4 antibody (Proteintech, 12949-1-AP) at a dilution of 1:400. After washing, the sections were incubated with the secondary antibody (GB21303, Servicebio) for 50 min at room temperature. Nuclei were counterstained with DAPI, and coverslips were sealed with anti-fluorescence quenching sealer (G1401, Servicebio).

Quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was extracted from gastrocnemius muscle tissues, and cDNA was synthesized using a PrimeScript RT Kit (Takara, Japan) according to the manufacturer’s instructions. Quantitative PCR was performed on a qTOWER 2.2 real-time PCR system (Analytik Jena, Germany) with a 20 µL reaction mixture containing 10 µL of 2 × Universal SYBR Green Fast qPCR Mix, 1 µL of cDNA template, 0.5 µL of 10 µM forward primer, 0.5 µL of 10 µM reverse primer and 8 µL of nuclease-free dH₂O. The thermal cycling protocol consisted of an initial denaturation at 95℃ for 3 min, followed by 40 cycles of denaturation at 95℃ for 5 s and annealing at 60℃ for 30 s. Gene expression levels were normalized to β-actin, and relative quantification was calculated using the 2^−ΔΔCT method. Oligonucleotide primers were synthesized by Sangon Biotech (Shanghai, China). The primer sequences are provided in Table 1.

Table 1.

Primer sequences for real-time RT-PCR

Gene	Forward Sequence (5’-3’)	Reverse Sequence (5’-3’)
Pdk4	AGGGAGGTCGAGCTGTTCTC	GGAGTGTTCACTAAGCGGTCA
Pdk1	GGACTTCGGGTCAGTGAATGC	TCCTGAGAAGATTGTCGGGGA
Pdk3	TCCTGGACTTCGGAAGGGATA	ACCTCTCTCATGGTGTTAGCC
Ldha	CGGCTGGGTCCTGGGAGAAC	ACCTCCTTCCACTGCTCCTTGTC
β-actin	TGAGAGGGAAATCGTGCGTGAC	CGCTCGTTGCCAATAGTGATGAC

Western blot

Total protein was extracted from gastrocnemius muscle tissues by homogenization in RIPA lysis buffer supplemented with protease and phosphatase inhibitors. Protein samples were separated by 10% SDS-PAGE and electrophoretically transferred onto PVDF membranes (Bio-Rad, USA). The membranes were blocked with 5% skim milk for 1 h at room temperature and then incubated with primary antibodies at 4℃ overnight. After washing, the membrane was incubated with the corresponding secondary antibodies for 1 h at room temperature. ImageJ software was used for quantitative analysis.

Transcriptome sequencing and bioinformatics analysis

Total RNA was extracted from the gastrocnemius muscles using TRIzol^® Reagent, followed by purification with an RNA purification kit (Majorbio, China). RNA integrity was assessed by electrophoresis on a regular agarose gel (Biowest, Spain). RNA concentration and purity were quantified using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), and the RNA integrity number (RIN) was determined using an Agilent 5300 system. Only high-quality RNA samples (OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, and RIN ≥ 6.5) were used for subsequent library preparation. Library preparation was performed, followed by sequencing on the NovaSeq X Plus platform (Illumina, USA) in paired-end mode with a read length of 150 bp (PE150). The sequencing depth was approximately 6 Gb of clean data per sample. The raw reads underwent quality control using Fastp software to obtain high-quality clean reads. All sequencing data from the samples demonstrated high quality, with Q20 values exceeding 98% and Q30 values exceeding 96%. The clean reads were aligned to the reference genome using the HiSat2 software with default parameters [24]. Gene expression levels were quantified based on the number of clean reads mapped to genomic regions (read counts) using RSEM software [25]. Differential expression analysis was conducted using the DESeq2 package in R, and genes satisfying the criteria of adjusted p-value < 0.05 (FDR) and |log2FC| > 1 were identified as differentially expressed genes (DEGs).

Statistical analysis

All data were obtained from at least three independent biological replicates. Statistical analyses were performed using GraphPad Prism 10.1.2. Due to the small sample size and the inability to assume normality, data were analyzed using nonparametric tests. Comparisons between two groups were performed using the Mann-Whitney U test, while comparisons among three or more groups were conducted using the Kruskal-Wallis H test followed by Dunn’s post hoc test. Data were presented as median and interquartile range (IQR). A p-value < 0.05 was considered statistically significant. Statistical significance was defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

Results

Hypoxia induces delayed force decline

To assess the physiological responses to acute hypoxia (FiO₂ = 11.8%), we monitored changes in body weight, food intake, and body composition in C57BL/6 mice. During the initial 24 h, hypoxia caused a significant reduction in food consumption and a concurrent decrease in body weight. Both parameters recovered to baseline levels after 72 h of hypoxia (Fig. 1A, B). Body composition analysis revealed that the early weight loss was primarily attributable to a reduction in free water content (Fig. 1C). In contrast, a significant loss of muscle content was observed only after 72 h of hypoxic exposure (Fig. 1D).

Fig. 1.

Hypoxia induces muscle and strength loss. (A) Body weight. (B) Food consumption. (C) Free water content. (D) Muscle content. (E) Relative grip strength. (F) Isometric force-frequency curve. (G) Maximum isometric force. (H) Mean fiber diameter of the gastrocnemius muscle. (I) Cross-sectional area of the gastrocnemius muscle. (J) Representative hematoxylin and eosin (H&E) staining of the gastrocnemius muscle. Scale bar: 20 μm. (K) Serum urea levels. (L) Serum creatine kinase levels. (M) Serum lactate dehydrogenase levels. Data are presented as median (IQR)

To explore the effect of hypoxia on muscle strength, we evaluated limb grip strength and contractile function in mice. Grip strength remained unchanged during the early phase of hypoxia, but decreased significantly after 72 h (Fig. 1E). Similarly, the force-frequency relationship measured by the muscle testing system showed a decreasing trend in both isometric force and maximal force after 72 h of hypoxia, although no statistically significant differences were observed (Fig. 1F, G). To assess structural changes, we performed H&E staining on the gastrocnemius muscles (Fig. 1J). The results showed a significant decrease in the mean diameter and cross-sectional area of muscle fibers after 48 and 72 h of hypoxia (Fig. 1H, I). Consistent with the changes in muscle fiber area, serum biochemistry revealed a significant increase in blood urea after 48 h of hypoxia (Fig. 1K). Notably, these functional and structural changes occurred in the absence of overt muscle damage, as serum levels of the muscle injury markers CK and LDH remained unchanged at all time points (Fig. 1L, M). In summary, acute hypoxia initially induced weight loss primarily due to decreased free water, followed by loss of muscle mass and strength after 72 h.

Acute hypoxia impairs endurance performance

To determine the effect of hypoxia on endurance, a treadmill exhaustion test was conducted on mice. Endurance was severely impaired at 12 h but gradually recovered to near-normoxic levels by 72 h (Fig. 2A). Before the treadmill test, we measured lactate levels in the tail vein after exposure. The results showed that the level was highest in the H12h group (Fig. 2B). Similarly, lactate levels in gastrocnemius muscle showed the same trend (Fig. 2C). To investigate the effect of hypoxia on skeletal muscle glucose metabolism, we detected several key metabolites. The results showed no significant changes in muscle glycogen or glucose 6-phosphate (G-6-P) levels between groups (Fig. 2D, E). Similarly, there were no differences in liver glycogen levels between groups (Fig. 2F). Notably, muscle pyruvate concentrations increased after 12 h of hypoxia (Fig. 2G). At this time point, skeletal muscle acetyl-CoA content decreased significantly (Fig. 2H). Blood glucose levels showed a trend toward an increase after 12 h of hypoxic exposure, although no statistically significant differences were observed (Fig. 2I).

Fig. 2.

Acute hypoxia induces carbohydrate metabolism disorders and endurance impairment. (A) Exhaustion time during treadmill test. Mice in the CON group were tested under normoxic conditions, while those in the hypoxic groups (H12h-H72h) were tested under sustained hypoxia. (B) Blood lactate levels under resting conditions. (C) L-lactate content in gastrocnemius muscle. (D) Glycogen content in quadriceps muscle. (E) Glucose 6-phosphate content in quadriceps muscle. (F) Glycogen content in liver. (G) Pyruvate content in gastrocnemius muscle. (H) Relative acetyl-CoA levels in gastrocnemius muscle. (I) Serum glucose levels. Data are presented as median (IQR)

Hypoxia induces transcriptomic signatures of metabolic adaptation in the gastrocnemius

To elucidate the molecular mechanisms underlying hypoxia-induced exercise impairment, we performed RNA sequencing of gastrocnemius muscles from CON, H12h, and H72h groups. Principal component analysis (PCA) showed clear segregation among the three groups (Fig. 3A). KEGG enrichment analysis indicated that genes upregulated in the H12h group were enriched in ribosome and oxidative phosphorylation pathways, while downregulated genes were primarily linked to fatty acid synthesis (Fig. 3B). A Venn diagram identified 174 genes that were consistently upregulated under hypoxic conditions (Fig. 3C). Subsequent enrichment analysis of this gene set revealed oxidative phosphorylation (OXPHOS) as a significantly enriched pathway (Fig. 3D). These transcriptomic findings indicate that acute hypoxia triggers a metabolic reprogramming in skeletal muscle, characterized by a shift toward enhanced OXPHOS.

Fig. 3.

Acute hypoxia induces metabolic reprogramming in gastrocnemius muscle. (A) Principal component analysis (PCA) of transcriptomic profiles from the CON, H12h, and H72h groups. (B) Top enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for genes differentially expressed in the H12h group compared to the CON group. (C) Venn diagram showing the overlap of differentially expressed genes in the H12h and H72h groups relative to the CON group. (D) KEGG pathway enrichment analysis of the 174 genes consistently upregulated under hypoxia

Acute hypoxia increases PDK4 expression and pyruvate-to-lactate conversion

PDK4 is a critical regulator of skeletal muscle energy metabolism by phosphorylating the E1α subunit of PDC, thereby inhibiting the conversion of pyruvate into acetyl-CoA and its subsequent oxidation in the tricarboxylic acid cycle [26]. Transcriptomics analysis revealed that acute hypoxia (12 h) significantly upregulated Pdk4 mRNA expression, which returned to normoxic levels after 72 h (Fig. 4B, C). Consistent with this, qPCR analysis confirmed a similar temporal pattern of Pdk4 expression in skeletal muscle (Fig. 4D). Immunofluorescence staining further demonstrated elevated PDK4 protein levels after 12 h of hypoxia (Fig. 4A). Correspondingly, immunoblotting revealed that both PDK4 expression and PDH E1α phosphorylation peaked at 12 h and gradually decreased with prolonged hypoxia (Fig. 4E-G). Collectively, these findings demonstrate that acute hypoxia activates the PDK4-PDH axis, which suppresses pyruvate oxidation and contributes to the early impairment of endurance performance.

Fig. 4.

Acute hypoxia upregulates PDK4 and enhances PDH phosphorylation. (A) Representative immunofluorescence images of PDK4 (red) in gastrocnemius muscle sections. Nuclei were counterstained with DAPI (blue). Scale bar: 4 μm. (B) Volcano plot of differentially expressed genes from transcriptomic analysis (H12h vs. CON). (C) Pdk4 expression from transcriptomic analysis. (D) Relative mRNA expression of Pdk4 in gastrocnemius muscle quantified by qPCR. (E-G) Representative western blots (E) and quantitative analysis of PDK4 protein (F) and phosphorylated PDH (p-PDH, Ser293) levels (G) in gastrocnemius muscle. Data are presented as median (IQR)

PDK4 inhibition restores hypoxic exercise capacity

To investigate the functional role of PDK4, we injected AAV empty vector (CON) or AAV vectors encoding PDK4-targeting shRNA (AAV-shPdk4) into the gastrocnemius and tibialis anterior muscles of mice prior to hypoxic exposure. Compared with mice injected with the control virus, mice with PDK4 knockdown showed a 55% increase in treadmill exhaustion time (Fig. 5A). Meanwhile, blood lactate (Fig. 5B) and glucose levels (Fig. 5C) were significantly decreased after the exhaustion test. Quantitative PCR analysis confirmed the specificity of the intervention by showing decreased Pdk4 and Ldha mRNA levels but no change in Pdk1 or Pdk3 expression (Fig. 5D-G). Western blotting further demonstrated reduced PDK4 protein abundance and decreased phosphorylation of PDH E1α (Fig. 5H-J). Moreover, PDH activity was significantly increased in the AAV-shPdk4 group (Fig. 5K). Collectively, these results suggest that PDK4 is a key mediator of hypoxia-induced endurance deficits and that its targeted inhibition improves exercise capacity.

Fig. 5.

Genetic knockdown of PDK4 improves exercise endurance under acute hypoxia. (A) Treadmill exhaustion time in CON and AAV-shPdk4 mice following 12 h of hypoxia. (B, C) Blood lactate (B) and glucose (C) levels measured immediately after treadmill exhaustion test. (D-G) Relative mRNA expression of Pdk1 (D), Pdk3 (E), Pdk4 (F), and Ldha (G) in gastrocnemius muscle. (H-J) Representative western blots (H) and quantitative analysis of PDK4 protein (I) and phosphorylated PDH (p-PDH, Ser293) levels (J) in gastrocnemius muscle. (K) PDH activity in gastrocnemius muscle. CON mice were injected with an AAV empty vector; AAV-shPdk4 mice were injected with an AAV vector expressing shRNA targeting PDK4. Data are presented as median (IQR)

DCA-mediated PDK4 inhibition restores physical endurance

To evaluate the therapeutic potential of PDK4 inhibition, we administered the pharmacological inhibitor dichloroacetate (DCA) prior to hypoxic exposure in mice. Although DCA treatment had no effect on the time to exhaustion under normoxic conditions, it significantly reduced resting blood lactate levels under hypoxia and prolonged the time to exhaustion by approximately 30% (Fig. 6B, C). Furthermore, DCA attenuated the hypoxia-mediated decline in skeletal muscle acetyl-CoA content (Fig. 6D). Immunofluorescence revealed that 12-hour hypoxic exposure notably increased PDK4 protein levels, an effect that was attenuated by DCA treatment (Fig. 6A). Consistent with this, immunoblot analysis confirmed that DCA effectively suppressed both the hypoxia-induced upregulation of PDK4 protein and the phosphorylation of PDH E1α (Fig. 6E-G). Taken together, these results identify PDK4 as a critical mediator of endurance impairment under acute hypoxia and establish its pharmacological inhibition as a viable strategy for restoring pyruvate oxidation and improving physical performance.

Fig. 6.

Pharmacological inhibition of PDK4 by DCA rescues hypoxia-impaired endurance. (A) Representative immunofluorescence images of PDK4 (red) in the gastrocnemius muscle. Nuclei were counterstained with DAPI (blue). Scale bar: 4 μm. (B) Treadmill exhaustion time in mice under normoxia or following 12 h of hypoxia with or without DCA treatment. (C) Blood lactate levels measured under resting conditions. (D) Relative acetyl-CoA levels in gastrocnemius muscle. (E-G) Western blot analysis (E) and quantitative analysis of PDK4 protein (F) and phosphorylated PDH (p-PDH, Ser293) levels (G) in gastrocnemius muscle. Data are presented as median (IQR)

Discussion

Hypoxia is a significant characteristic of high-altitude environments. Acute exposure to low oxygen environments, such as high-altitude travel, aviation flights, or certain clinical pathological conditions, can lead to a significant decline in human physical activity. This phenomenon presents a significant challenge for military personnel, mountaineers, athletes, and patients with cardiovascular diseases [1, 27, 28]. It is well established that acute hypoxia significantly reduces endurance capacity [11]. The underlying physiological mechanisms are multifaceted, involving dysregulation across the cardiopulmonary oxygen transport chain [29, 30], central nervous system [31, 32], and peripheral skeletal muscles [33, 34]. As the primary effector of movement, skeletal muscle is a critical target for therapeutic intervention. Under hypoxia, it must contend with diminished oxygen supply alongside sustained or increased metabolic demand. Consequently, elucidating the metabolic alterations induced by acute hypoxia in skeletal muscle is essential for enhancing adaptation to hypoxia and improving exercise performance.

The conventional understanding suggests that hypoxia shifts skeletal muscle energy metabolism toward greater carbohydrate reliance, as carbohydrates yield more ATP per unit of oxygen consumed [13, 14]. However, evidence indicates that carbohydrate supplementation fails to enhance endurance performance under hypoxia. For instance, Caris et al. reported that carbohydrate supplementation did not alter time to exhaustion during running under normobaric hypoxia (FiO₂ = 13.5%) [35]. Similarly, Liao et al. found no performance improvement from carbohydrate intake during a 40-km cycling time trial under normobaric hypoxia (FiO₂ = 16.3%) [36]. Furthermore, even with optimized exogenous carbohydrate oxidation using fructose-glucose mixtures, Bradbury et al. observed no performance benefit under hypobaric hypoxia (460 mmHg) compared to a placebo [37, 38]. This contrasts with the well-established effect of carbohydrate supplementation at sea level [36, 37, 39]. The molecular mechanisms behind this contradictory phenomenon remain unclear. To address this knowledge gap, we established a mouse model of acute hypoxia exposure using a normobaric hypoxia chamber to elucidate how acute hypoxia disrupts skeletal muscle metabolism and impairs exercise performance.

Our findings indicate that acute hypoxia induces temporally distinct impairments in skeletal muscle physical performance and metabolic homeostasis. Endurance performance decreased most markedly during the acute phase (12 h), a decline that correlated with elevated PDK4 expression and associated metabolic alterations. By 72 h of exposure, endurance performance had returned to near-normoxic levels, coinciding with the normalization of key pyruvate metabolic parameters. In addition to the metabolic adaptations, the partial recovery of aerobic performance after 72 h of hypoxic exposure may also involve compensatory adjustments in the respiratory and cardiovascular systems, including increased ventilation and cardiac output, which help maintain oxygen delivery to working muscles during sustained hypoxia. In contrast, muscle strength progressively declined, reaching statistical significance at 72 h, an effect likely attributable to enhanced skeletal muscle proteolysis [40]. Collectively, these results suggest that acute hypoxic exposure impairs endurance and strength performance in a time-dependent manner, with PDK4-mediated metabolic reprogramming in skeletal muscle playing a contributing role.

Transcriptome analysis showed that acute hypoxia induces metabolic reprogramming in skeletal muscle, characterized by an upregulation of oxidative phosphorylation (OXPHOS) and a downregulation of fatty acid biosynthesis. This reprogramming is thought to optimize ATP production and conservation under oxygen-limited conditions. However, despite the upregulation of the OXPHOS pathway, the oxidative metabolism of glucose is paradoxically hindered. This is because acute hypoxia markedly upregulates PDK4, which in turn increases PDH phosphorylation, thereby inhibiting PDH activity. The inhibition of PDH disrupts pyruvate metabolism, leading to pyruvate accumulation and its diversion toward lactate production. Consequently, the efficiency of carbohydrate utilization in skeletal muscle is substantially compromised, which exacerbates energy deficiency and impairs endurance performance [41]. This mechanism likely explains why carbohydrate supplementation fails to improve exercise performance during acute hypoxia.

Importantly, our study suggests that PDK4 inhibition represents a promising therapeutic strategy for counteracting hypoxia-induced metabolic disturbances and endurance impairment. We show that genetic knockdown of skeletal muscle PDK4 via AAV-shPdk4 significantly improves endurance capacity under hypoxia. Pharmacologically, we employed DCA, a pyruvate analog and pan-PDK inhibitor known to restore PDH complex activity [42]. Consistent with its reported mechanism, for example, in reducing PDH phosphorylation in colon cancer cells [43], we found that DCA treatment attenuated the hypoxia-induced upregulation of PDK4 and consequent PDH phosphorylation. This intervention was associated with the redirection of glycolytic pyruvate into the TCA cycle, accompanied by reduced lactate accumulation and improved endurance performance.

However, several limitations should be noted in this study. We employed a normobaric hypoxic (NH) model, in which the physiological stimuli were achieved by adjusting the fraction of inspired oxygen to mimic hypobaric hypoxia (HH) [44, 45]. Although the NH model is widely used in hypoxia research, accumulating evidence suggests that physiological responses to NH and HH are not entirely equivalent [46, 47]. Therefore, future studies are warranted to validate the applicability of these findings to populations undergoing rapid ascent to high altitude, thereby facilitating the translation of these results to the field of high-altitude medicine.

Conclusions

In summary, this study investigated the impact of acute hypoxia on physical performance and skeletal muscle metabolism (Fig. 7). Acute hypoxia induces temporally distinct metabolic and functional adaptations in skeletal muscle. In the early phase (12 h), PDK4-mediated suppression of pyruvate oxidation contributes to impaired endurance performance, whereas prolonged exposure (72 h) leads to loss of muscle and subsequent reduction in muscle strength. Importantly, inhibition of PDK4 with DCA improved endurance performance under hypoxic conditions, accompanied by reduced lactate accumulation. These findings provide insights for the maintenance of physical performance in low-oxygen environments.

Fig. 7.

The mechanism of time-dependent exercise impairment induced by acute hypoxia and its therapeutic targeting. Acute hypoxic exposure impairs exercise capacity in a time-dependent manner: endurance is most severely compromised at 12 h, whereas grip strength declines most significantly by 72 h. Mechanistically, under normoxia, glycolytic pyruvate is primarily converted to acetyl-CoA by pyruvate dehydrogenase (PDH) to fuel oxidative phosphorylation (OXPHOS) for efficient ATP production. During the early hypoxic phase (12 h), despite a compensatory upregulation of OXPHOS pathways, a marked induction of pyruvate dehydrogenase kinase 4 (PDK4) in skeletal muscle increases PDH phosphorylation. This modification is associated with reduced pyruvate oxidation, shunting pyruvate away from mitochondrial oxidation toward lactate accumulation and impairing endurance performance. Pharmacological inhibition of PDK4 with dichloroacetate (DCA) alleviates this early impairment by restoring pyruvate flux into the tricarboxylic acid (TCA) cycle. Following prolonged hypoxia (72 h), Pdk4 expression subsides and glucose metabolism partially normalizes. However, sustained hypoxia ultimately leads to muscle loss and the associated decline in strength performance

Abbreviations

AAV

Adeno-associated virus

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

BSA

Bovine serum albumin

CK

Creatine kinase

DCA

Sodium dichloroacetate

DEGs

Differentially expressed genes

GLU

Glucose

G-6-P

Glucose 6-phosphate

HH

Hypobaric hypoxia

H&E

Hematoxylin and eosin

IQR

Interquartile range

LDH

Lactate dehydrogenase

NH

Normobaric hypoxia

OXPHOS

Oxidative phosphorylation

PCA

Principal component analysis

PDC

Pyruvate dehydrogenase complex

PDH

Pyruvate dehydrogenase

PDK4

Pyruvate dehydrogenase kinase 4

TCA

Tricarboxylic acid cycle

Author contributions

MM, LY conceived and designed the research plan and provided guidance on the experimental procedures. HL offered additional experimental guidance. PW carried out most experiments, performed the data analyses, and drafted the manuscript. YX, MZ, XR, SM took part in animal experiments. All authors have read and approved the final version.

Funding

The author(s) declare that no financial support was received for the research and/or publication of this article.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.