Low-intensity exercise induces acute shifts in liver and skeletal muscle substrate metabolism but not chronic adaptations in tissue oxidative capacity

Abstract

Adaptations in hepatic and skeletal muscle substrate metabolism following acute and chronic (6 wk; 5 days/wk; 1 h/day) low-intensity treadmill exercise were tested in healthy male C57BL/6J mice. Low-intensity exercise maximizes lipid utilization; therefore, we hypothesized pathways involved in lipid metabolism would be most robustly affected. Acute exercise nearly depleted liver glycogen immediately postexercise (0 h), whereas hepatic triglyceride (TAG) stores increased in the early stages after exercise (0–3 h). Also, hepatic peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) gene expression and fat oxidation (mitochondrial and peroxisomal) increased immediately postexercise (0 h), whereas carbohydrate and amino acid oxidation in liver peaked 24–48 h later. Alternatively, skeletal muscle exhibited a less robust response to acute exercise as stored substrates (glycogen and TAG) remained unchanged, induction of PGC-1α gene expression was delayed (up at 3 h), and mitochondrial substrate oxidation pathways (carbohydrate, amino acid, and lipid) were largely unaltered. Peroxisomal lipid oxidation exhibited the most dynamic changes in skeletal muscle substrate metabolism after acute exercise; however, this response was also delayed (peaked 3–24 h postexercise), and expression of peroxisomal genes remained unaffected. Interestingly, 6 wk of training at a similar intensity limited weight gain, increased muscle glycogen, and reduced TAG accrual in liver and muscle; however, substrate oxidation pathways remained unaltered in both tissues. Collectively, these results suggest changes in substrate metabolism induced by an acute low-intensity exercise bout in healthy mice are more rapid and robust in liver than in skeletal muscle; however, training at a similar intensity for 6 wk is insufficient to induce remodeling of substrate metabolism pathways in either tissue.

NEW & NOTEWORTHY Effects of low-intensity exercise on substrate metabolism pathways were tested in liver and skeletal muscle of healthy mice. This is the first study to describe exercise-induced adaptations in peroxisomal lipid metabolism and also reports comprehensive adaptations in mitochondrial substrate metabolism pathways (carbohydrate, lipid, and amino acid). Acute low-intensity exercise induced shifts in mitochondrial and peroxisomal metabolism in both tissues, but training at this intensity did not induce adaptive remodeling of metabolic pathways in healthy mice.

INTRODUCTION

Substrate preference during exercise varies with intensity. At low intensities [<40% maximal oxygen consumption (V̇o_2max)] lipid is the predominant fuel, and as intensity increases the body starts to rely more heavily on carbohydrates (7, 8, 10, 30, 61). Peak fat oxidation rates are achieved at submaximal exercise (approximately 45–65% peak oxygen consumption), but as intensity increases above this point reliance on lipid diminishes and glucose becomes the predominant fuel, a phenomenon referred to as the crossover concept (7–9, 24, 59, 60). The source of substrate also varies with exercise intensity. In response to endurance exercise, adipose tissue undergoes lipolysis and hepatic glycogen stores are mobilized to provide fatty acids and glucose, respectively. During low-intensity exercise, skeletal muscle utilizes a greater percentage of these circulating substrates; however, as intensity increases, there is greater reliance on intramuscular triglyceride and glycogen stores (61). These findings emphasize that substrate mobilization and metabolism must be coordinated among multiple organ systems in response to exercise; therefore, understanding these relationships is important to maximize the health benefits. The interaction between skeletal muscle and liver is of particular interest.

In lean, healthy individuals, skeletal muscle represents ~40% of total body mass, whereas liver accounts for <3% body mass. Despite these large differences, at rest a relatively similar amount of cardiac output is delivered to both liver (~25%) and skeletal muscle (~20%), and each is predicted to account for nearly 20% of basal metabolic rate (17, 72). During maximal exercise, a 20-fold increase in whole body metabolic rate can occur, which is largely driven by skeletal muscle as ATP turnover can increase 100-fold and oxygen uptake in a single muscle group can increase 240-fold (2, 16, 57). To deliver sufficient oxygen and substrates to meet the increased energy demand, blood flow to skeletal muscle can increase ~30-fold (14, 18). Consequently, total blood flow to liver decreases during exercise, yet this is not matched by a corresponding decline in metabolic rate of liver (23). Indeed, the liver plays an important role in exercise tolerance by removing lactate through the Cori cycle and maintaining blood glucose via increasing glycogenolysis and gluconeogenesis (69, 73), both of which help delay exhaustion. Whereas the role of the liver in maintaining blood glucose supply during exercise has been studied in detail, less is known about how hepatic substrate metabolism itself changes in response to exercise.

Acute exercise increases mitochondrial respiration, TCA cycle flux, and fat oxidation in liver (3, 4, 19, 20, 29). High-fit rodents also have increased hepatic mitochondrial content and higher fatty acid oxidative capacity and are protected against high-fat-diet-induced metabolic diseases including hepatic steatosis, obesity, and insulin resistance (42–44, 46, 49, 68). Studies examining effects of exercise training in animal models of nonalcoholic fatty liver disease across a broad range of intensities (low, moderate, and high), modalities (voluntary wheel run and treadmill), and volumes (low vs. high) consistently demonstrate exercise improves hepatic mitochondrial function, increases fat oxidation, and reduces liver steatosis (5, 35, 36, 54, 56). These findings seem to translate to humans as exercise training appears to decrease hepatic lipid storage in male and female subjects with obesity (31, 32, 52). Clearly, exercise training has beneficial effects on liver health in models of metabolic disease, but less is known about adaptations in substrate metabolism in a healthy population. Also, to our knowledge, a comprehensive examination of adaptations in various substrate oxidation pathways in response to acute and chronic exercise in multiple tissues has not been described. The purpose of the present study was to address this by describing changes in oxidation of carbohydrate-, amino acid-, and fatty acid-derived fuels used by mitochondria as well as peroxisomal lipid oxidation in liver and skeletal muscle in response to both acute exercise and chronic exercise training in lean, healthy male mice. We were particularly interested in testing adaptations in lipid metabolism, so low-intensity exercise was chosen because 1) lipid is the primary fuel used at this intensity and 2) it emphasizes the contribution of circulating fuels to meet the increased energy demand of skeletal muscle. Within this paradigm, we hypothesized fatty acid metabolism pathways would be most robustly altered and that adaptations in skeletal muscle would outpace those in liver.

MATERIALS AND METHODS

Animals

C57BL/6J male mice (n = 54) were ordered from Jackson (stock no. 000664; Bar Harbor, ME) at 12 wk of age and studied at 20 wk of age. Mice were group-housed at room temperature under a 12:12-h light-dark cycle and allowed ad libitum access to food and water. A primary goal was to test adaptations in lipid metabolism pathways in response to low-intensity exercise, which emphasizes the relative contribution of fatty acids to energy production. Since even moderate-fat diets (25% kcal from fat) can induce obesity and significantly impact lipid metabolism pathways alone, mice were fed a relatively low-fat standard chow (Purina Rodent Chow 5001; Purina Mills, St. Louis, MO) that provides 28.5% kcal from protein, 13.5% kcal from fat, and 58% kcal from carbohydrate. This was done in an effort to maximize exercise-induced adaptations in lipid metabolism pathways while minimizing confounding adaptations that occur due to dietary lipid. This also has the advantage of being representative of the most common type of diet fed to rodents. At the end of the study, mice were anesthetized via intraperitoneal injection of ketamine-xylazine-acepromazine (16 mg/ml ketamine, 0.8 mg/ml xylazine, and 0.32 mg/ml acepromazine at a dose of 0.125 ml/20 g body wt). Serum was isolated from trunk blood, whereas tissues were collected and either 1) used fresh for substrate oxidation assays or 2) snap-frozen in liquid nitrogen and stored at −80°C until subsequent analyses could be performed. The Pennington Biomedical Research Center has an Association for Assessment and Accreditation of Laboratory Animal Care International-approved Comparative Biology Core facility and veterinary staff that monitor the health of the animals via a sentinel program and daily inspection. All studies were approved by the Institutional Animal Care and Use Committee.

Exercise Studies

All mice were habituated on an Exer 3/6 treadmill (Columbus Instruments, Columbus, OH) for 3 days. The treadmill was set at 10° incline, and the habituation protocol consisted of 5-min stages at 0, 5, and 10 m/min followed by 2 min at 15 m/min. All mice responded well to habituation, so they were randomly assigned to either an exercise or sedentary control group. Mice are more reliant on glucose at a given exercise intensity than humans; however, a goal of this study was to use protocols that emphasized lipid utilization. In this regard, untrained C57BL/6 mice exercising at the speeds and incline used in the present study have a respiratory quotient slightly below 0.8 (39), lending credence to the notion that lipid was likely a primary fuel utilized herein. All exercise bouts were performed during the light cycle (6–10 AM) and were designed to recapitulate a low-exercise intensity, which induced a mild increase in blood lactate postexercise (Fig. 1B, inset). Metabolic adaptations were tested in response to 1) an acute exercise bout and 2) chronic exercise training, and greater experimental details are provided below. Aversive stimuli (puffs of air, light tapping with a brush, and light electrical shock) were used to motivate mice to run to ensure they completed each exercise bout. Mice were monitored continuously throughout each exercise bout, and the interventions were well-tolerated as none exhibited signs of exhaustion (no righting reflex and/or acceptance of light electrical shock for >5 s) or injury (awkward gait, foot injury, etc.).

Fig. 1.

Acute exercise (Ex) protocol and bloodwork. Timeline for the acute exercise study is depicted in A. For the acute exercise bout, the treadmill was set at a 10° incline and included a 5-min warmup (W) followed by a progressive increase in speed until minute 45, after which time the speed was reduced for 15 min (B). Blood lactate was measured before and after the exercise bout (B, inset) as a gauge of exercise intensity. Nonesterified fatty acids (NEFAs; C) and triglycerides (D) were measured in serum, and blood glucose was measured in conscious mice before tissue harvest (E). Sedentary controls (Sed), n = 10 mice; 0, 3, 24, and 48 h postexercise, n = 6 mice/time point. Statistical significance (P < 0.05) is represented as *pre- vs. postexercise, ^aindicated time point vs. Sed, ^bindicated time point vs. 0 h, and ^cindicated time point vs. 3 h.

Acute exercise study.

A total of 34 mice were used in the acute exercise study. The study design is shown in Fig. 1A, and the acute exercise protocol is illustrated in Fig. 1B. Mice exercised on a treadmill for 1 h, and tissues were harvested immediately postexercise (0 h) as well as 3, 24, and 48 h postexercise (n = 6 mice/time point). A habituated, nonexercised group served as sedentary controls (n = 10). As depicted in Fig. 1A, all exercise sessions occurred between 6 and 10 AM (light cycle) and tissues were collected at 10 AM after a 4-h food removal. The time of tissue collection and length of food removal were kept consistent among all groups to limit fluctuations in outcomes that could be mediated by shifts in circadian metabolism. Mice in the 24- and 48-h groups were returned to home cages immediately after the exercise bout with ad libitum access to food and water until 6 AM the morning of the tissue collection (Fig. 1A).

Chronic exercise training study.

To test exercise training adaptations at low intensity, mice ran (n = 10) on a treadmill for 6 wk, 5 days/wk, 1 h/day. All training sessions were performed during the light cycle between 7 and 11 AM. To account for adaptations in aerobic fitness, speed and/or duration of the exercise bouts were progressively increased weekly to increase workload (Supplemental Table S1; all supplemental material is available at https://doi.org/10.6084/m9.figshare.7732973.v1). Habituated, nonexercised mice served as sedentary controls (n = 10). Body mass and composition [fat, lean, and fluid mass; measured by Bruker minispec Live Mice Analyzer (LF50) Time Domain nuclear magnetic resonance (NMR)] were measured at the beginning and end of the 6-wk study. Tissues were collected 48 h after the final exercise bout, and food was removed 3 h before harvest. Joules were calculated for exercise bouts to account for work done in both horizontal and vertical planes at 10° incline using the following formula/s:

$$\text{joules\hspace{0.17em}(newton\hspace{0.17em}meters)}=\text{newtons}\times [\text{horizontal\hspace{0.17em}distance\hspace{0.17em}(m)}+\text{vertical\hspace{0.17em}distance\hspace{0.17em}(m)}]$$

$$\text{newtons\hspace{0.17em}(kg}\cdot \text{m}/{\text{s}}^2\text{)}=\text{body\hspace{0.17em}weight\hspace{0.17em}(kg)}\times 9.81\text{\hspace{0.17em}m}/{\text{s}}^2$$

$$\text{horizontal\hspace{0.17em}distance\hspace{0.17em}(m)}=\text{speed\hspace{0.17em}(m}/\text{min)}\times \text{duration\hspace{0.17em}(min)}\times \cos (10)$$

$$\text{vertical\hspace{0.17em}distance\hspace{0.17em}(m)}=\text{speed\hspace{0.17em}(m}/\text{min)}\times \text{duration\hspace{0.17em}(min)}\times \sin (10).$$

Substrate Oxidative Capacity

Fresh liver and mixed gastrocnemius (MG) skeletal muscle homogenates were prepared as described (49, 50, 74). With the use of established methods (49, 50, 74), liberation of ¹⁴CO₂ from homogenates incubated in the presence of [¹⁴C]radiolabeled substrates (American Radiolabeled Chemicals, St. Louis, MO) was used to measure complete oxidation (¹⁴CO₂) of the following: [1-¹⁴C]pyruvate (1 mM) was used to measure pyruvate dehydrogenase (PDH) activity, [2-¹⁴C]pyruvate (1 mM) was used to measure pyruvate oxidation, [U-¹⁴C]leucine (100 µM) was used to measure branched-chain amino acid oxidation, [1-¹⁴C]palmitate (200 µM) was used to measure fatty acid oxidation, and [1-¹⁴C]lignocerate (25 µM) was used to specifically measure peroxisomal fatty acid oxidation. Incomplete palmitate oxidation was assessed by measuring ¹⁴C-labeled acid-soluble metabolite (ASM) generation.

Citrate Synthase Activity

Citrate synthase activity was measured using the protocol of Srere (66). With the use of described methods (50), a citrate synthase activity assay was used to test mitochondrial intactness in liver and skeletal muscle homogenates from four additional C57BL/6J mice. Supplemental Fig. S1 shows mitochondrial intactness in MG homogenates was >75%, whereas liver homogenates had >95% intact mitochondria. The difference between tissues is likely due to the fact that skeletal muscle undergoes a more intense homogenization process than liver because it is more fibrous. Maintaining intact mitochondria within these homogenates is important as ¹⁴CO₂ liberation is completely abolished in lysed mitochondria; however, ASMs can still be generated, as previously shown (50).

Bloodwork

Blood glucose (ACCU-CHEK Aviva Plus Glucometer; Roche Diagnostics, Indianapolis, IN) and lactate (Lactate Plus Meter; Nova Biomedical, Waltham, MA) were measured via tail vein in conscious mice. Serum lipids were analyzed by measuring nonesterified fatty acids (Wako Chemicals, Richmond, VA) and triglycerides (Sigma, St. Louis, MO), whereas ELISA kits from Mercodia (Uppsala, Sweden) were used to measure insulin (cat. no. 10-1247-01) and glucagon (cat. no. 10-1271-01). Manufacturer’s recommended protocols were used for all measurements.

Tissue Substrate Storage

Glycogen and triglycerides (TAGs) were measured in liver and quadriceps.

Glycogen.

Powdered tissue (15–30 mg) was used to extract and measure glycogen using a commercially available kit (ab65620; Abcam, Cambridge, MA) according to the manufacturer’s instructions.

TAG extraction and quantification.

Powdered tissues (25–30 mg) were homogenized 2 × 15 s in 300-µl ice-cold 5% Nonidet P-40 using a handheld homogenizer (VWR, Radnor, PA). Lysates were slowly heated to 95°C for 5 min until the solution became white and then cooled at room temperature for ~15 min. This heat-cooling procedure was performed twice to solubilize tissue TAG. Homogenates were centrifuged (14,000 g, 2 min, room temperature), and supernatants were collected for TAG content determination. A commercially available triglyceride assay kit (cat. no. TR0100; Sigma) was used to measure glycerol as the indirect measure of total tissue triglycerides according to company instructions. This kit allows for determination of free glycerol, total triglycerides, and true triglycerides (total TAG minus free glycerol); all data are presented as true TAG.

Gene Expression

Gene expression was analyzed as described (47, 48, 65). Briefly, RNA was isolated from powdered MG and liver (approximately 15–20 mg) using an RNeasy Mini Kit (QIAGEN, Germantown, MD) with proteinase K digestion and on-column DNase treatment per manufacturer’s instructions. A NanoDrop ND-1000 (Thermo Fisher Scientific, Wilmington, DE) was used to measure RNA concentration and quality, and 1 µg was used to generate cDNA with an iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Real-time PCR was performed using an ABI PRISM 7900 HT Sequence Detection System (Life Technologies, Carlsbad, CA) using iTaq Universal SYBR Green Supermix (cat. no. 172-5124; Bio-Rad). Primer pairs were designed with Primer-BLAST software, and sequence information is provided in Supplemental Table S2.

Tissue Protein Analysis

Western blots were performed using standard SDS-PAGE. Briefly, tissue lysates were prepared using T-PER Tissue Protein Extraction Reagent (cat. no. 78510; Thermo Fisher Scientific) containing protease inhibitor and phosphatase inhibitor cocktails from Sigma. Twenty-five- to thirty-milligram powdered tissues were homogenized 2 × 30 s by a handheld homogenizer (VWR) in ice-cold buffer. Lysates were treated with 1% Triton X-100 (final concentration), sonicated 2 × 5 s, and placed on a shaker at 4°C for 1 h. Homogenates were centrifuged (12,000 g, 15 min, 4°C), and supernatants were analyzed for protein concentration by bicinchoninic acid assay (Thermo Fisher Scientific, Rockford, IL). Protein lysates were prepared in denaturing sample buffer and loaded in equal quantities (20–50 µg, depending on the protein of interest) into wells of precast polyacrylamide Tris·HCl gels (Bio-Rad). Following electrophoresis, proteins were transferred onto nitrocellulose membranes (Bio-Rad) and incubated overnight with the following primary antibodies: acetyl-CoA carboxylase (ACC; cat. no. 3676), phospho-ACC (p-ACC; cat. no. 11818), AMP-activated protein kinase-α (AMPKα; cat. no. 5831), p-AMPKα (cat. no. 2535), AMPKβ1/2 (cat. no. 4150), p-AMPKβ1 (cat. no. 4181), glycogen synthase (GS; cat. no. 3886), p-GS (cat. no. 3891), GSK-3β (cat. no. 12456), p-GSK-3β (cat. no. 5558), peroxisomal biogenesis factor 5 (PEX5; cat. no. 83020), S6 kinase (S6K; cat. no. 2708), pS6K (cat. no. 9234), and p-PKA substrate (cat. no. 9624) from Cell Signaling Technology; PDH (ab110330), p-PDH (ab92696), Catalase (ab16731), Total OXPHOS cocktail (ab110413), PEX19 (ab137072), and PMP70 (ab3421) from Abcam; and PEX14 (ABC142) and peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α; ST1202) from EMD Millipore. Horseradish peroxidase-linked secondary antibodies (anti-mouse IgG, cat. no. NXA931, and anti-rabbit IgG, cat. no. NA934V) were purchased from GE Healthcare (Piscataway, NJ), and proteins were detected using ECL chemistry. Reversible Protein Stain MemCode (cat. no. 24580; Thermo Fisher Scientific) was used to confirm equal transfer of proteins, and quantitation of these bands served as a loading control. A ChemiDoc Imaging System (Bio-Rad) was used to image bands, and Image Lab software (Bio-Rad) was used to quantitate band intensity.

Statistical Analysis

Blood lactate taken before and after exercise was analyzed using a paired t-test. All other data from the acute exercise time course were analyzed using a one-way ANOVA with Tukey post hoc analysis. Changes in body composition (NMR) in sedentary and exercise-trained mice before and after the 6-wk training protocol were analyzed using a two-way repeated-measures ANOVA with a Bonferroni post hoc analysis. All other data from the exercise training study were analyzed using a two-tailed t-test. GraphPad Prism software was used for all statistical analyses, and a P ≤0.05 was established a priori as representing a statistically significant difference.

RESULTS

Acute Exercise: Serum Response

The design of the acute exercise time course is shown in Fig. 1A, and the acute exercise bout is depicted in Fig. 1B. On conclusion of the exercise bout, a significant (~50%) increase in blood lactate was observed (Fig. 1B, inset), although this level would not exceed the lactate threshold (4–5 mM), thus confirming a low-intensity exercise stimulus was achieved. Immediately postexercise (0 h), the expected increase in circulating nonesterified fatty acids (NEFAs) was observed (Fig. 1C); however, there were no changes in circulating triglycerides (TAGs; Fig. 1D) and only a slight (nonsignificant) decrease in glucose (Fig. 1E). Interestingly, blood glucose levels were significantly higher 3 h postexercise than immediately after exercise cessation (Fig. 1E), despite the fact that mice were not provided food during this timeframe (Fig. 1A). Restoration of blood glucose 3 h postexercise did not seem to be related to changes in circulating levels of the glucose-regulating hormones insulin or glucagon (Supplemental Fig. S2), however, did coincide with decreased circulating TAGs (Fig. 1D).

Acute Exercise: Hepatic Substrate Oxidative Capacity

Ability to use carbohydrate-derived substrates was tested by measuring pyruvate dehydrogenase (PDH) activity and pyruvate oxidation (Fig. 2A). Although changes were subtle, the trend was for PDH activity and pyruvate oxidative capacity to decrease in the early stages (0–3 h) after exercise cessation; however, by 48 h postexercise, both parameters were significantly higher than those of nonexercised mice. PDH phosphorylation (Supplemental Fig. S3A) and expression of the PDH kinase 4 (PDK4) gene (Fig. 2C) were largely unaltered after acute exercise, and the patterns generally coincided with PDH activity and pyruvate oxidation. To test adaptations in amino acid metabolism, we measured oxidative capacity to catabolize the branched-chain amino acid (BCAA) leucine (Fig. 2A). Unlike carbohydrate metabolism, hepatic leucine oxidation gradually increased after exercise cessation and peaked 24 h postexercise; however, it returned to sedentary control levels 48 h postexercise. This pattern was not mirrored by genes linked to BCAA metabolism (Fig. 2C). Exercise-induced changes in hepatic lipid metabolism were tested by measuring the oxidative capacity to catabolize a long-chain fatty acid (palmitate; C16:0) to test mitochondrial fat utilization, whereas a very long-chain fatty acid (lignocerate; C24:0) was used to assess peroxisomal lipid catabolism (Fig. 2A). The ability to completely catabolize both lipid substrates (CO₂) increased immediately postexercise (0 h), was significantly reduced 3 h postexercise, and generally returned to sedentary levels by 24–48 h postexercise. A similar pattern was observed when measuring incomplete (ASM) and total palmitate oxidation (Supplemental Fig. S3B). Moreover, changes in hepatic lipid metabolism after acute exercise exhibited much more fluctuation than either carbohydrate-derived or amino acid substrates. Interestingly, changes in complete palmitate oxidation were negatively correlated with blood glucose levels (Fig. 2B). The changes in fat oxidation rates did not appear to coincide with alterations at the molecular level as genes involved in mitochondrial and peroxisomal lipid metabolism were largely unaltered postexercise (Fig. 2C). Additionally, genes related to mitophagy, mitochondrial fission, and mitochondrial fusion (Supplemental Fig. S3C) were largely unaltered. Interestingly, PGC-1α gene expression increased nearly fivefold in the liver immediately postexercise (Fig. 2D) and PGC-1α protein levels were slightly elevated (Fig. 2E); however, alterations in PGC-1α did not coincide with substantial remodeling of the oxidative capacity of the tissue. Indeed, the only change observed in proteins involved in the oxidative phosphorylation system of the mitochondrial electron transport chain was a decrease in complex I in the early stages postexercise (0–3 h), yet expression of this protein returned to baseline levels by 24–48 h after exercise cessation (Fig. 2F). Decreased protein content of a complex I subunit is interesting, and although we cannot entirely rule out the possibility of poor antibody recognition of complex I contributing to this finding, if the results are real this could be due to exercise-induced shifts in several regulated pathways (i.e., incorporation into supercomplexes, posttranslational modification that either targets the protein for catabolism or interferes with the antibody recognition site, etc.). It is, however, worth noting that it did not seem to limit substrate oxidative capacity at these time points.

Fig. 2.

Effect of acute exercise on hepatic substrate oxidative capacity. [¹⁴C]radiolabeled substrates were used to measure pyruvate dehydrogenase (PDH) activity as well as oxidation (Ox) rates of pyruvate, leucine, palmitate, and lignocerate in liver homogenates (A). Acute exercise induced shifts in hepatic palmitate oxidation that had a significant negative correlation with changes in blood glucose (B). Expression of genes involved in carbohydrate, amino acid, as well as mitochondrial (Mito) and peroxisomal (Perox) lipid metabolism was assessed by RT-PCR (C). Genes involved in mitochondrial biogenesis were measured by RT-PCR (D). Western blot analyses were used to measure protein levels of peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α; E) and electron transport chain complexes (F), and quantitated band intensity is normalized to total protein stain using MemCode (MemC). All electron transport chain complex proteins were run on a single gel but are presented as cropped images due to different exposure times for Complex I (30 s) vs. Complexes II–V (1 s). All data are expressed relative to the average of sedentary (Sed) controls. Sed, n = 10 mice; 0, 3, 24, and 48 h postexercise, n = 6 mice/time point. Statistical significance (P < 0.05) is represented as ^aindicated time point vs. Sed, ^bindicated time point vs. 0 h, ^cindicated time point vs. 3 h, and ^dindicated time point vs. 24 h. AU, arbitrary units; CI-NDUFB8, complex I, NADH:ubiquinone oxidoreductase subunit B8; CII-SDHB, complex II, succinate dehydrogenase complex iron sulfur subunit B; CIII-UQCR2, complex III, coenzyme Q - cytochrome c oxidoreductase; CIV-MTCO1, complex IV, mitochondrially encoded cytochrome c oxidase I; CV-ATP5A, complex V, ATP synthase F₁ subunit-α.

Acute Exercise: Hepatic Substrate Storage

Intrahepatic glycogen stores were dramatically reduced immediately postexercise (Fig. 3A), but by 3 h postexercise levels were restored despite no access to food during this period. Glycogen synthase promotes glycogen storage, so it is noteworthy that decreased hepatic glycogen occurred despite increased total GS content; however, it appears GS was kept predominantly in the inactive phosphorylated form at 0 h postexercise (Fig. 3B). Glycogen synthase kinase-3β (GSK-3β) inhibits GS by phosphorylating the enzyme, so GSK-3β protein was examined. Results show that neither total GSK-3β nor phosphorylated GSK-3β (inactive state) change in liver after exercise (Fig. 3C). Also, gluconeogenic genes were increased immediately postexercise (Fig. 3D) and the pattern mirrors that of PGC-1α (Fig. 2D), which is consistent with the notion that PGC-1α has a primary role to drive gluconeogenesis (76). Recently, ribosomal S6 kinase (S6K) was reported to dissociate the gluconeogenic effects of PGC-1α from its role in mitochondrial biogenesis (37), so the activation status of S6K was tested. Neither total nor phosphorylated (active) S6K was altered in liver after acute exercise (Supplemental Fig. S4A). Together, these results suggest the liver mobilizes most of its glycogen during low-intensity exercise; however, as glycogen stores are depleted, adaptations occur (increased GS content, PGC-1α, and gluconeogenic genes) that facilitate complete restoration of glycogen content within 3 h after completing the exercise bout, despite having no access to food.

Fig. 3.

Effect of acute exercise on hepatic stored substrate and signaling cascades. Intrahepatic glycogen stores were measured in sedentary (Sed) controls as well as at various time points (0, 3, 24, and 48 h) postexercise (A). Western blots were performed to measure total and phosphorylated (p) forms of glycogen synthase (GS; B) and glycogen synthase kinase-3β (GSK3β; C), whereas genes linked to gluconeogenesis were measured by RT-PCR (D). Intrahepatic triglycerides (True TAG; E) as well as genes involved in triglyceride synthesis (F) and storage (G) were measured. Western blots were used to assess phosphorylated protein kinase A (Phospho-PKA) substrates (H) as well as total and phosphorylated forms of AMP-activated protein kinase, α-isoform (AMPKα; I) and AMP-activated protein kinase, β1-isoform (AMPKβ1; J). Quantitated band intensity of all Western blot images is normalized to total protein stain using MemCode (MemC). Protein and gene expression data are expressed relative to the average of the sedentary control group. Sed, n = 10 mice; 0, 3, 24, and 48 h postexercise, n = 6 mice/time point. Statistical significance (P < 0.05) is represented as ^aindicated time point vs. Sed, ^bindicated time point vs. 0 h, and ^cindicated time point vs. 3 h. AU, arbitrary units.

Liver TAGs increased immediately postexercise (0 h) and remained elevated for ≥3 h after exercise; however, changes in TAG content subsided by 24–48 h postexercise (Fig. 3E). Since hepatic fatty acid oxidative capacity and TAG storage both changed after acute exercise, factors involved in lipid storage were examined. As shown, acute exercise did not significantly alter expression of key genes that regulate TAG synthesis (Fig. 3F) or the perilipin family of lipid droplet-associated proteins (Fig. 3G). Acetyl-CoA carboxylase (ACC) produces malonyl-CoA, which inhibits mitochondrial lipid entry and is a precursor in de novo lipogenesis; thus ACC activation promotes lipid storage. As shown, neither total (active) nor phosphorylated (inactive) ACC was robustly altered postexercise (Supplemental Fig. 4B) nor was expression of genes involved in regulating malonyl-CoA markedly altered (Supplemental Fig. S3C). Overall, these findings suggest changes in palmitate oxidative capacity (Fig. 2A), and hepatic TAG levels (Fig. 3E) are not likely related to shifts in ACC function in this study.

Direct regulators of glycogen and TAG content did not seem to account for acute exercise-induced changes in liver substrate storage; thus upstream factors that regulate substrate turnover in response to exercise were tested. The increase in serum NEFAs immediately postexercise (Fig. 1C) is consistent with an adrenergic response driving lipolysis in adipose tissue. Adrenergic stimulation in liver activates protein kinase A (PKA), which favors glycogenolysis and lipolysis (63). To test whether PKA activation may mediate exercise-induced shifts in intrahepatic substrate storage, tissue lysates were probed with an antibody that recognizes consensus phosphorylation sites that are PKA substrates. With the use of this as a proxy for PKA activity, quantitation of the entire lane for each exercise time point did not reveal broad-scale PKA activation (Fig. 3H). Glucagon has also been reported to be a primary driver of PKA activity in the liver in response to exercise (69), thus a relative lack of PKA induction is consistent with unchanged circulating glucagon levels (Supplemental Fig. S2) and suggests the exercise stimulus was not sufficient to induce these pathways. Collectively, these observations indicate hepatic glycogen was completely restored within 3 h after exercise despite 1) limited activation of glucagon signaling and 2) no access to food during this time period. Acute exercise also activates AMP-activated protein kinases (AMPK), which similarly promote utilization of intrahepatic glycogen and TAG as fuel (71); thus AMPK regulation was tested. Low-intensity exercise did not alter expression or activation (phosphorylated form) of either AMPKα (Fig. 3I) or AMPKβ1 (Fig. 3J) isoforms in liver. Since AMPK inhibits hepatic gluconeogenesis (71), a lack of induction of AMPK likely contributes to maintenance of blood glucose during low-intensity exercise.

Acute Exercise: Skeletal Muscle Substrate Oxidative Capacity

Since skeletal muscle is heavily recruited during exercise, the effect of an acute, low-intensity exercise bout on substrate oxidative capacity was tested in mixed gastrocnemius (MG) homogenates (Fig. 4A). Despite a significant increase in PDK4 3 h postexercise (Fig. 4B), PDH-E1α phosphorylation was unaltered (Supplemental Fig. S5A) and metabolism of carbohydrate-derived substrates changed very little in response to acute exercise, with the exception being a significant increase in pyruvate oxidation 48 h postexercise (Fig. 4A). Likewise, leucine oxidation was similar across all time points (Fig. 4A) and expression of genes related to BCAA metabolism did not change (Fig. 4B), indicating BCAA metabolism in MG was unaffected by acute exercise. Regarding mitochondrial fat metabolism, a significant increase in carnitine-acylcarnitine translocase (Cact), which controls acyl-carnitine transport into mitochondria, was observed between 0 and 24 h postexercise (Fig. 4B). Palmitate oxidation rates generally followed a similar pattern (Fig. 4A and Supplemental Fig. S5B); however, the net effect was not sufficient to suggest acute exercise induced substantial changes in mitochondrial fatty acid oxidative capacity. Peroxisomal genes were unaltered by acute exercise (Fig. 4B); however, peroxisomal lipid oxidation exhibited significant shifts. Specifically, exercise-induced adaptations in lignocerate (C24:0) oxidation exhibited a delayed response as rates were highest between 3 and 24 h after exercise; however, this elevation was not sustained 48 h postexercise (Fig. 4A). The notion of skeletal muscle having a delayed response to the exercise bout appeared to be partially supported by the fact that PGC-1α gene expression was significantly increased 3 h postexercise (Fig. 4C). However, no differences in PGC-1α protein content (Fig. 4D), oxidative phosphorylation complexes of the electron transport chain (Fig. 4E), or remodeling of mitochondrial fission, fusion, or mitophagy at the gene level (Supplemental Fig. S5C) were observed.

Fig. 4.

Effect of acute exercise on skeletal muscle substrate oxidative capacity. [¹⁴C]radiolabeled substrates were used to measure pyruvate dehydrogenase (PDH) activity as well as oxidation (Ox) rates of pyruvate, leucine, palmitate, and lignocerate in homogenates prepared from mixed gastrocnemius skeletal muscle (A). Expression of genes involved in carbohydrate, amino acid, as well as mitochondrial (Mito) and peroxisomal (Perox) lipid metabolism was assessed by RT-PCR (B). Genes involved in mitochondrial biogenesis were measured by RT-PCR (C). Western blot analyses were used to measure protein levels of peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α; D) and electron transport chain complexes (E), and quantitated band intensity is normalized to total protein stain using MemCode (MemC). All electron transport chain complex proteins were run on a single gel but are presented as cropped images due to different exposure times for Complex I (60 s) vs. Complexes II–V (1 s). All data are expressed relative to the average of sedentary (Sed) controls. Sed, n = 10 mice; 0, 3, 24, and 48 h postexercise, n = 6 mice/time point. Statistical significance (P < 0.05) is represented as ^aindicated time point vs. Sed, ^bindicated time point vs. 0 h, ^cindicated time point vs. 3 h, and ^dindicated time point vs. 24 h. AU, arbitrary units; CI-NDUFB8, complex I, NADH:ubiquinone oxidoreductase subunit B8; CII-SDHB, complex II, succinate dehydrogenase complex iron sulfur subunit B; CIII-UQCR2, complex III, coenzyme Q - cytochrome c oxidoreductase; CIV-MTCO1, complex IV, mitochondrially encoded cytochrome c oxidase I; CV-ATP5A, complex V, ATP synthase F₁ subunit-α.

Acute Exercise: Intramuscular Substrate Storage

During exercise, skeletal muscle consumes more fuel to meet increased energy demand. Data above suggest the exercise bout mobilized lipids from adipose tissue (Fig. 1C) and glucose from hepatic glycogen stores (Fig. 3A) to help meet the increased energy requirement; however, it is also important to examine potential shifts in intramuscular substrate stores in response to exercise. Acute exercise did not alter intramuscular glycogen (Fig. 5A) or TAGs (Fig. 5D). Additionally, acute exercise did not appear to activate signaling cascades that control glycogen and TAG turnover in muscle including GS (Fig. 5B), GSK-3β (Fig. 5C), PKA (Fig. 5G), or AMPKα (Fig. 5H). Genes involved in TAG synthesis (Fig. 5E) and storage were also largely unaltered with the possible exception being a generalized decrease in a subset of the perilipin genes (Plin2, Plin4, and Plin5) 48 h after exercise (Fig. 5F). Finally, no significant changes were noted in ACC signaling (Supplemental Fig. S6A) or expression of Acc1, Mcd, or Scd1 genes (Supplemental Fig. S6B), suggesting acute low-intensity exercise was not sufficient to modulate pathways that regulate fat oxidation and storage in skeletal muscle of lean, healthy mice.

Fig. 5.

Effect of acute exercise on skeletal muscle stored substrate and signaling cascades. Intramuscular glycogen stores were measured in sedentary (Sed) controls as well as at various time points (0, 3, 24, and 48 h) postexercise (A). Western blots were performed to measure total and phosphorylated (p) forms of glycogen synthase (GS; B) and glycogen synthase kinase-3β (GSK3β; C). Intramuscular triglycerides (True TAG; D) as well as genes involved in triglyceride synthesis (E) and storage (F) were measured. Western blots were used to assess phosphorylated protein kinase A (Phospho-PKA) substrates (G) as well as total and phosphorylated forms of AMP-activated protein kinase, α-isoform (AMPKα; H). Quantitated band intensity of all Western blot images is normalized to total protein stain using MemCode (MemC). Protein and gene expression data are expressed relative to the average of the sedentary control group. Sed, n = 10 mice; 0, 3, 24, and 48 h postexercise, n = 6 mice/time point. Statistical significance (P < 0.05) is represented as ^aindicated time point vs. Sed, ^bindicated time point vs. 0 h, ^cindicated time point vs. 3 h, and ^dindicated time point vs. 24 h. AU, arbitrary units.

Exercise Training: Evidence of Adaptations at Low Intensity

To examine the effects of training at low-exercise intensity, mice were run on a treadmill for 6 wk, 5 days/wk, 1 h/day using the protocol detailed in Supplemental Table S1. Both sedentary and trained mice lost weight over the 6-wk intervention, but an interaction effect indicates the loss in body mass was more robust in exercise-trained mice (Fig. 6A). As shown in Fig. 6B, sedentary mice gained significant amounts of lean and fluid mass during the 6-wk period, but this effect was negated in exercise-trained mice. Unfortunately, food intake was not measured in this study, but it seems likely that the shifts in body mass and composition are due primarily to exercise as most literature in this area suggests food intake is either unchanged or increased in response to low-to-moderate-intensity exercise (1, 40). Heart weight was not different in exercise-trained mice (Fig. 6C), but when expressed relative to body weight a training effect was observed (Fig. 6D). Although glycogen levels in mice are much lower than rats or humans (33), the exercise training protocol did increase glycogen content in skeletal muscle (Fig. 6E), which is consistent with an exercise training response (21). Also, the exercise training protocol decreased tissue TAGs in both liver and skeletal muscle (Fig. 6F). Collectively, these findings confirm the 6-wk exercise training protocol was sufficient to induce training adaptations.

Fig. 6.

Evidence for exercise training adaptations after 6-wk training at low intensity. Body weight (A) was measured in sedentary (n = 10 mice) and trained (n = 10 mice) mice before and after the 6-wk intervention. Changes in body composition (fat, lean, and fluid mass) are expressed as values after the 6-wk intervention minus baseline values (B). Heart weight was obtained (C) and expressed relative to body weight (D). Liver and quadriceps skeletal muscle samples were analyzed for stored glycogen (E) and triglyceride (True TAG) levels (F). Statistical significance (P < 0.05) is represented as #Pre- vs. Post-6-wk intervention, *Sedentary vs. Trained, and ^interaction effect.

Exercise Training: Liver Adaptations

Despite dynamic changes in hepatic substrate oxidation rates in response to acute exercise (Fig. 2A), training for 6 wk at a similar intensity did not significantly alter oxidation rates of carbohydrate-, amino acid-, or lipid-derived fuels in liver homogenates pulled 48 h after the last exercise bout (Fig. 7A and Supplemental Fig. S7A). Likewise, expression of genes involved in carbohydrate, BCAA, and lipid metabolism was largely unchanged in liver from exercise-trained mice (Fig. 7B). To test this further, Western blot analyses were performed to assess peroxisomal proteins, and results indicate that PEX5 protein content was significantly reduced, whereas other peroxisomal proteins were unaltered (Fig. 7C). Since PGC-1α can mediate exercise-induced remodeling of metabolic pathways (including mitochondrial biogenesis and gluconeogenesis) and levels of this transcriptional coactivator were increased immediately following acute exercise in liver (Fig. 2D), we measured PGC-1α and mitochondrial markers. Exercise training did not alter hepatic PGC-1α gene (Fig. 7D) and protein (Fig. 7E) content. Likewise, few differences were detected when analyzing genes involved in mitochondrial fission, fusion, or mitophagy (Supplemental Fig. S7B) nor were differences in citrate synthase activity (Fig. 7F) or electron transport chain complexes observed with the exception being a small increase in complex III protein (Fig. 7G). Decreased expression of the acyl-CoA synthetase long-chain family member 1 (Acsl1) gene (Supplemental Fig. S7C) did not associate with changes in fatty acid oxidative capacity but did coincide with lower intrahepatic TAGs (Fig. 6F); however, other genes related to TAG storage were not altered (Supplemental Fig. S7C). Finally, as shown in Supplemental Fig. S7, D–H, no differences in phosphorylation status of AMPKα, AMPKβ1, ACC, GS, or GSK-3β were observed. Overall, these results suggest the training stimulus was insufficient to produce substantial metabolic remodeling in the liver.

Fig. 7.

Exercise training for 6 wk at low intensity is not sufficient to remodel hepatic oxidative capacity. [¹⁴C]radiolabeled substrates were used to measure pyruvate dehydrogenase (PDH) activity as well as oxidation (Ox) rates of pyruvate, leucine, palmitate, and lignocerate in homogenates prepared from liver of sedentary (n = 10 mice) and trained (n = 10 mice) mice (A). Expression of genes involved in carbohydrate (CHO), amino acid, as well as mitochondrial (Mito) and peroxisomal (Perox) lipid metabolism was assessed by RT-PCR (B). Peroxisomal proteins were detected by Western blot (C). Effects of exercise training on mitochondrial biogenesis were tested by measuring expression of peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) and Tfam genes (D), PGC-1α protein (E), citrate synthase activity (F), and protein levels of electron transport chain complexes (G). Quantitated band intensity of all Western blots images is normalized to total protein stain using MemCode (MemC). All data are expressed relative to the average of the sedentary group. Statistical significance (P < 0.05) is represented as *P < 0.05, Sedentary vs. Trained. AU, arbitrary units; CI-NDUFB8, complex I, NADH:ubiquinone oxidoreductase subunit B8; CII-SDHB, complex II, succinate dehydrogenase complex iron sulfur subunit B; CIII-UQCR2, complex III, coenzyme Q - cytochrome c oxidoreductase; CIV-MTCO1, complex IV, mitochondrially encoded cytochrome c oxidase I; CV-ATP5A, complex V, ATP synthase F₁ subunit-α; PEX, peroxisomal biogenesis factor.

Exercise Training: Skeletal Muscle Adaptations

Although acute exercise did not markedly modify metabolic pathways in skeletal muscle (Fig. 4), exercise training did alter intramuscular glycogen and TAG storage (Fig. 6, E and F); thus adaptations in substrate metabolism pathways in skeletal muscle were tested. Like liver, low-intensity exercise training did not significantly change oxidative capacity for carbohydrate-, amino acid-, or lipid-derived substrates in muscle (Fig. 8A and Supplemental Fig. S8A). Likewise, genes involved in carbohydrate, BCAA, and lipid metabolism (mitochondrial and peroxisomal) as well as peroxisomal proteins were unaltered (Fig. 8, B and C). Markers indicative of mitochondrial biogenesis (Fig. 8D), fission, fusion, and mitophagy (Supplemental Fig. S8B) were also virtually unchanged at the gene level. Likewise, PGC-1α protein (Fig. 8E), citrate synthase activity (Fig. 8F), and oxidative phosphorylation protein content (Fig. 8G) were not altered in trained mice. No differences in phosphorylation status of AMPKα, ACC, GS, or GSK-3β were detected (Supplemental Fig. S8, D–G). Finally, although exercise training reduced intramuscular TAGs (Fig. 6F), no adaptations in genes linked to TAG synthesis were found (Supplemental Fig. S8C), which is consistent with a lack of adaptation in fatty acid oxidative capacity. Collectively, these results suggest that despite findings showing the training stimulus was sufficient to alter intramuscular substrate storage, it was not sufficient to induce remodeling of the mitochondrial or peroxisomal oxidative pathways.

Fig. 8.

Exercise training for 6 wk at low intensity is not sufficient to remodel skeletal muscle oxidative capacity. [¹⁴C]radiolabeled substrates were used to measure pyruvate dehydrogenase (PDH) activity as well as oxidation (Ox) rates of pyruvate, leucine, palmitate, and lignocerate in homogenates prepared from mixed gastrocnemius of sedentary (n = 10 mice) and trained (n = 10 mice) mice (A). Expression of genes involved in carbohydrate (CHO), amino acid, as well as mitochondrial (Mito) and peroxisomal (Perox) lipid metabolism were assessed by RT-PCR (B). Peroxisomal proteins were detected by Western blot (C). Effects of exercise training on mitochondrial biogenesis were tested by measuring expression of peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) and Tfam genes (D), PGC-1α protein (E), citrate synthase activity (F), and protein levels of electron transport chain complexes (G). Quantitated band intensity of all Western blots images is normalized to total protein stain using MemCode (MemC). All data are expressed relative to the average of the sedentary group. AU, arbitrary units; CI-NDUFB8, complex I, NADH:ubiquinone oxidoreductase subunit B8; CII-SDHB, complex II, succinate dehydrogenase complex iron sulfur subunit B; CIII-UQCR2, complex III, coenzyme Q - cytochrome c oxidoreductase; CIV-MTCO1, complex IV, mitochondrially encoded cytochrome c oxidase I; CV-ATP5A, complex V, ATP synthase F₁ subunit-α.

DISCUSSION

The exercise response requires coordinated adaptations among several organ systems. The goal of this study was to better define the coordinated regulation of substrate metabolism between skeletal muscle and liver in response to low-intensity exercise, and we were particularly interested in testing adaptations in lipid metabolism pathways. Low-intensity exercise was chosen because 1) lipid is the primary fuel used at this intensity and 2) it emphasizes contributions of circulating fuels derived from nonmuscle organs to meet the increased energy demand of skeletal muscle. Substrate storage as well as oxidative capacity to use carbohydrate, amino acid, and fatty acid fuels (mitochondrial and peroxisomal) were measured in liver and skeletal muscle in response to acute exercise and chronic exercise training in lean, healthy mice. To our knowledge, this is the most comprehensive assessment to date describing adaptations in substrate metabolism pathways in skeletal muscle and liver in response to acute and chronic low-intensity exercise.

Developing rodent exercise protocols to match relative intensity to those used in human studies offers technical challenges. Human studies often define and monitor relative exercise intensity through 1) breath gas measurements based on V̇o_2max, 2) heart rate reserve, and/or 3) rating of perceived exertion. These methods are either very challenging (V̇o_2max and heart rate reserve) or impossible (rating of perceived exertion) to use in rodent studies. Alternatively, testing blood lactate is commonly used to gauge exercise intensity in humans and can be readily measured in rodents; thus the current study assessed blood lactate levels sampled immediately postexercise to objectively assess exercise intensity. In this regard, low-intensity exercise has been reported to increase circulating lactate 60–80% in humans (10) and mice (6), which is similar to the ~50% increase observed herein. Furthermore, mice in this study had similar outcomes as humans after acute low-intensity exercise as we found significant increases in circulating NEFAs, induction of PGC-1α gene expression, and little evidence of mobilization of intramuscular substrates (10, 61). Additionally, AMPK activity increases in proportion to increasing exercise intensity. Rodent studies often show low-intensity exercise increases AMPK activity (6, 15, 25); however, this is not universally reported (12, 67, 70). Importantly, most human studies report low-intensity exercise does not alter AMP levels (27) or p-AMPK status (10, 41, 53, 62, 75) and that an exercise intensity >60% V̇o_2max is required to increase AMPK activity in human skeletal muscle (58). With this in mind, findings from our study showing acute exercise did not alter AMPK phosphorylation are consistent with human literature using low-intensity exercise. Overall, these findings provide confidence that 1) results from these studies likely translate well to lean, healthy humans and 2) a low-intensity exercise protocol was achieved and that testing blood lactate may be a practical and reliable tool to gauge exercise intensity in mice.

In the current study, alterations in circulating substrates without changes in intramuscular substrate stores are consistent with reports indicating skeletal muscle primarily uses circulating substrates during low-intensity exercise (61, 64). Because of the reliance on circulating fuels, we predicted substrate oxidation rates in skeletal muscle homogenates would closely track with blood glucose and/or fatty acids; however, this was not the case. Indeed, with the exception of peroxisomal lipid utilization, substrate oxidative capacity generally did not shift in response to acute exercise. It is undeniable that in vivo fuel metabolism by skeletal muscle increases substantially during exercise; however, our findings suggest the oxidative capacity of muscle in lean, healthy mice was sufficient to meet the energy demand of the low-intensity exercise protocol used, thus remodeling metabolic pathways was not necessary. These findings are in line with some other reports that show low-intensity exercise does not consistently alter mitochondrial content in lean, healthy subjects (6, 38). Alternatively, it is likely that exposing subjects who are older or with obesity to a similar training regimen or increasing the training volume in healthy mice similar to that which occurs with unrestricted access to a running wheel would have resulted in expansion of metabolic pathways by low-intensity exercise training as previously reported (34, 51, 55, 64).

Several studies show exercise training obese, insulin-resistant rodents across a broad range of intensities, volumes, and modalities increases liver oxidative capacity and improves hepatic steatosis (5, 35, 36, 54, 56). Herein, we report 6 wk of training at low intensity did not alter hepatic substrate oxidative capacity in young, lean, healthy male mice fed a low-fat diet; however, the training regimen was sufficient to lower hepatic TAGs, supporting a model of improved hepatic health. The lack of training adaptations in substrate oxidative capacity was curious since acute low-intensity exercise induced rapid and robust alterations in hepatic substrate oxidation rates. Of interest is that liver lipid metabolism was negatively correlated with circulating glucose, which is likely best explained in the context of previous findings showing hepatic fat oxidation during exercise plays an important role in the exercise response by 1) delivering ketones into the circulation for use as metabolic fuel and 2) providing energy necessary to drive gluconeogenesis (11, 13, 73). With this in mind, the most significant increase in hepatic fat oxidation occurred immediately postexercise, which was a time at which liver glycogen was significantly reduced. Recalling that skeletal muscle relies heavily on circulating substrates during low-intensity exercise and that intramuscular glycogen did not change in response to acute exercise, these results stress the importance of the liver as a primary source of glucose during low-intensity exercise in mice.

Hepatic glycogenolysis clearly contributed to maintenance of blood glucose in this study; however, exercise also activated gluconeogenesis (13). Evidence suggests gluconeogenesis increased after acute exercise as gluconeogenic genes were elevated immediately postexercise and hepatic glycogen was completely restored within 3 h after exercise despite no access to food. Glucagon has been reported to drive a gluconeogenic response during exercise (73); however, data suggest the gluconeogenic response in the present study occurred independent of glucagon as circulating levels were not altered and there was little evidence of PKA activation. We speculate that despite the fact that low-intensity exercise mobilized lipids from white adipose tissue, neither the adrenergic stimulus nor the drop in blood glucose was sufficient to invoke a response from pancreatic α-cells. If true, this suggests alterations in hepatic lipid metabolism occur independently from, and seemingly precede, glucagon regulation of hepatic substrate metabolism during low-intensity exercise; however, this concept is speculative and needs to be more thoroughly tested. Alternatively, PGC-1α is involved in hepatic exercise adaptations as it can 1) increase mitochondrial biogenesis and substrate oxidative capacity (45) and 2) regulate gluconeogenesis (76). Results herein support previous findings (22) showing rapid and robust induction of hepatic PGC-1α immediately postexercise, but these effects are transient and return to basal levels 3 h postexercise. PGC-1α induction in liver occurred at a time point when 1) hepatic glycogen content was greatly reduced, 2) gluconeogenic genes were increased, 3) circulating glucose slightly declined, 4) circulating NEFAs were elevated, and 5) hepatic fat oxidation rose despite few changes in genes/proteins involved in oxidative metabolism. Additionally, the negative correlation between hepatic palmitate oxidation and blood glucose after acute exercise is consistent with reports indicating increased fat oxidation in the liver that results from elevated NEFA delivery provides energy necessary to drive gluconeogenesis (73). Also, activation of S6 kinase inhibits the gluconeogenic function of PGC-1α while maintaining its effects on mitochondrial biogenesis (37); however, our findings indicate S6 kinase activation is unaltered, thus the gluconeogenic potential of PGC-1α would be retained. Therefore, we speculate these adaptations represent a physiological state where the primary response of the liver to low-intensity exercise is to modulate metabolism to maintain blood glucose and that the oxidative capacity in lean, healthy mice is sufficient to meet this need. This is supported by findings showing that despite the fact that acute low-intensity exercise increased PGC-1α expression and induced rapid changes in hepatic substrate oxidation rates, exercise training at low intensity was not sufficient to induce persistent remodeling of metabolic pathways. Collectively, these findings show that alterations in hepatic metabolism in response to acute low-intensity exercise are rapid and emphasize that liver plays an important role in the exercise response in mice.

Although hepatic palmitate oxidation is elevated immediately postexercise, it is not sufficient to prevent excess intrahepatic TAG accumulation at this time point. Increased hepatic TAGs persisted between 0 and 3 h postexercise, which is consistent with previous reports (26, 28). Acute exercise provides a strong lipolytic stimulus, and liberation of NEFA from white adipose tissue exceeds the metabolic demand of skeletal muscle (61). As a result of the mismatch between lipid supply and demand, it is apparent that the liver compensates by increasing incorporation of circulating lipids into the hepatic TAG pool. In this regard, although the importance of increasing hepatic fat oxidation during exercise as discussed above should not be understated, it is likely that a primary role of the liver in lipid metabolism during exercise is to remove excess lipids from the circulation and store them in the neutral lipid pool (26). It is, however, important to note that this effect is transient and hepatic TAGs return to levels similar to sedentary controls by 24–48 h postexercise. Moreover, the increase in hepatic TAGs in the early stages after an acute exercise bout does not promote steatosis as evidenced by the fact that chronic training at a similar intensity for 6 wk reduced hepatic TAGs, suggesting beneficial effects on long-term hepatic health.

Overall, although 6 wk of low-intensity exercise training was sufficient to alter substrate storage in liver and skeletal muscle, the stimulus did not appear to be strong enough to induce remodeling of mitochondrial or peroxisomal metabolic pathways. In contrast, alterations in substrate metabolism pathways after acute low-intensity exercise provided more interesting observations. Mitochondrial substrate oxidation changed little in skeletal muscle after acute exercise; however, peroxisomal lipid oxidation was elevated between 3 and 24 h postexercise. This may indicate peroxisomal lipid metabolism plays a role in exercise recovery, but this is speculative and requires further study. Although adaptations in skeletal muscle were delayed, perhaps the most interesting results indicated adaptations in substrate metabolism pathways after acute exercise were more rapid in liver. A swift transition of the liver toward fat utilization likely protects against exercise-induced hypoglycemia, and we predict this high degree of metabolic flexibility in liver of healthy mice plays an important role in the acute exercise response, but this remains to be further tested. As stated earlier, several key observations in this study are consistent with findings in humans, which provides confidence that results herein translate favorably to lean, healthy people. Generally speaking, most evidence in the literature suggests low-intensity training is quite likely to induce metabolic remodeling in cases of metabolic disease; however, our results suggest a more intense training regimen may be required to induce metabolic remodeling in lean, healthy individuals, but this requires additional study. Finally, it is important to note that only male mice were tested; thus examining similar questions in females will likely provide fruitful future directions to determine whether sex-specific differences to low-intensity exercise exist.

GRANTS

This work used Pennington Biomedical Research Center core facilities (Genomics and Animal Metabolism and Behavior) that are supported in part by Centers of Biomedical Research Excellence (3-P30-GM-118430) and Nutrition Obesity Research Centers (2P30-DK-072476) grants from the National Institutes of Health. This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant 1-R01-DK-103860 (to R. C. Noland). S. E. Fuller was supported by a T32 fellowship (AT004094).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.E.F. and R.C.N. conceived and designed research; S.E.F., T.-Y.H., J.S., H.M.B., N.M.E., M.C.S., C.M.W., J.M.B., S.J.B., and R.C.N. performed experiments; S.E.F., T.-Y.H., J.S., S.J.B., J.J.C., and R.C.N. analyzed data; S.E.F., T.-Y.H., J.S., S.J.B., J.J.C., and R.C.N. interpreted results of experiments; S.E.F., T.-Y.H., J.S., S.J.B., J.J.C., and R.C.N. prepared figures; S.E.F., T.-Y.H., and R.C.N. drafted manuscript; S.E.F., T.-Y.H., and R.C.N. edited and revised manuscript; S.E.F., T.-Y.H., and R.C.N. approved final version of manuscript.