Introduction

During exercise, fatty acids recruited from the blood and from triacylglycerol (TG) stores within the muscles are the major contributors to muscle fat oxidation. The importance of the endogenous muscle fat contribution to total fat oxidation during exercise is not yet fully understood primarily because of the heterogeneous muscle triacylglycerol storage and the problem of contamination from adipocytes and connective tissue around the muscle fibres. However, when prolonged exercise is performed, a majority of studies using different techniques, e.g. the muscle biopsy technique combined with chemical analysis or electron microscopy and the 1H NMR spectroscopy, have demonstrated muscle triacylglycerol depletion in the order 20–40% of the muscle triacylglycerol stores (Suter et al. 1995; Van Loon 2004). In addition, there is evidence that endurance-trained subjects have a higher storage and a higher recruitment of muscle triacylglycerol during exercise than untrained subjects (Suter et al. 1995). Type I muscle fibres apparently store more triacylglycerol (2–3-fold) than type II fibres (Essen 1977; Howald et al. 1985; Gaster et al. 2003). In a recent study, endurance-trained subjects exercised for 120 min at 60% VO2max, and muscle triacylglycerol was primarily recruited from type I muscle fibres and to a minor degree from type II muscle fibres (Van Loon et al. 2003). However, it is not known whether the relative contribution of type I and II muscle fibres, respectively, to triacylglycerol recruitment during exercise differs between trained and untrained subjects.

In skeletal muscles, stored TG is recruited by lipase activation during exercise. Hormone-sensitive lipase (HSL) is thought to be the most important lipase in human skeletal muscle, and the HSL activity is increased both by adrenaline stimulation and muscle contraction (Langfort et al. 1999, 2000; Donsmark et al. 2003; Watt et al. 2003b). In a series of studies, Watt et al. (2003a) colleagues found that the muscle HSL activity was increased after 1 h, but had returned to base line after 4 h of cycle exercise. Unfortunately, no data on muscle TG utilization was presented, and it is thus difficult to tell whether the change in muscle HSL activity played a role for muscle TG recruitment. It is also not known whether during exercise HSL activity differs between trained and untrained muscles. Therefore, this study investigated the muscle HSL activity and the depletion of TG in both type I and II muscle fibres during prolonged exercise in trained and untrained male subjects.

Methods

Subjects

Sixteen healthy young male subjects, aged 27 ± 1 years, height 181 ± 2 cm, weight 77 ± 2 kg, participated in the study. Subjects were fully informed of the nature and the possible risks associated with the study before they volunteered to participate. The study was approved by the ethical committee of the Copenhagen municipality and adhered to the principles of the Helsinki Declaration. Of the 16 subjects, eight were heavily engaged in endurance training and eight did not participate in any regular endurance-type training.

Experimental protocol

On 2 days, separated by at least 2 and not more than 14 days, subjects came to the laboratory. On day 1, subjects performed a standard progressive maximal oxygen uptake test with increments of 45 W min(1. Two days prior to the experimental day, subjects were asked to include carbohydrate-rich foods in their diet to ensure full muscle glycogen stores and on the last day subjects were asked to refrain from vigorous physical activity. On the experimental day, subjects came overnight fasted to the laboratory. Prior to exercise, body weight and height were measured. After a 15-min rest, a needle biopsy was obtained with suction from the vastus lateralis muscle (Bergström 1962). Subjects then started bicycle exercise at a workload estimated to elicit 60% of VO2peak. Throughout the exercise, pulmonary oxygen uptake, carbon dioxide excretion, and heart rate were frequently measured. Subjects had free access to water, and they were encouraged to drink regularly. After termination of exercise, subjects were immediately placed in a bed, and a biopsy was obtained with suction from the vastus lateralis muscle.

Analytical procedures

During exercise, heart rate was recorded continuously with a PE 3000 Sports Tester (Polar Electro, Finland). Pulmonary oxygen uptake and carbon dioxide excretion were measured by an automated on-line system (CPX, Medical Graphics, Spiropharma, Denmark). Gases of known composition were used to calibrate each system regularly.

The biopsies were frozen in liquid nitrogen within 10–15 s of sampling, and the biopsies were stored at (80 °C until further analysis. Before biochemical analysis for enzyme activity, and triacylglycerol and glycogen contents, the muscle samples were freeze-dried and dissected free of connective tissues, visible fats, and blood using a stereomicroscope. The maximal activity of the enzymes β-hydroxy-acyl-CoA-dehydrogenase (HAD), citrate synthase (CS), phospho-fructo kinase (PFK), and lactate dehydrogenase (LDH) were determined fluorometrically according to Lowry and Passonneau (1972). HSL assays are based on measurements of release of [3H]oleic acid from 1(3)-mono[3H]oleoyl−2-oleylglycerol (a diacylglycerol analogue not hydrolysable at position 2), referred to as HSL (DG) activity, and from tri[3H]olein, referred to as HSL (TG) activity, under conditions optimal for HSL. In the basal state, the catalytic activity of adipose tissue HSL towards diacylglycerol (DG) substrates is about ten-fold higher than the activity against triacylglycerol (TG) substrates (Holm et al. 2000; Langfort et al. 1999; Straalfors et al. 1987). Upon phosphorylation by protein kinase A, the activity of adipose tissue and soleus muscle HSL towards TG increases markedly, whereas the HSL (DG) activity does not change significantly (Holm et al. 2000; Langfort et al. 1999; Straalfors et al. 1987). Furthermore, the increases in muscle HSL (TG) activity induced by adrenaline or contractions can be completely abolished by an anti-HSL antibody present during analysis (Langfort et al. 1999, 2000). In addition, also the increase in human muscle HSL (TG) activity induced by exercise can be totally abolished by an anti-HSL antibody present during analysis (Roepstorff et al. 2004). Accordingly, HSL (TG) activity is a measure of the activated form of HSL and the assay of choice for monitoring changes in the activation state of HSL, whereas HSL (DG) activity has been taken as a measure of total enzyme concentration(Holm et al. 2000; Langfort et al. 1999; Straalfors et al. 1987). Tri[3H]olein was obtained from Amersham Pharmacia Biotech and 1(3)-mono[3H]oleoyl-2-oleylglycerol (DG) was synthesized by Lennart Krabisch (Department of Cell and Molecular Biology, Lund University, Lund, Sweden). The TG and DG substrates were emulsified with phospholipids by sonication (Osterlund et al. 1996). BSA was used as a fatty acid acceptor. For measurements of HSL (TG) activity and HSL (DG) activity, 14 and 7 μl, respectively, of muscle supernatants (protein concentration ∼2.5 mg ml-1) were incubated for 30 min at 37 °C with 100 μl of 5 mM (available acyl chains) TG (1.25 × 104 c.p.m.) or DG (0.4 × 106 c.p.m.) substrate and enzyme dilution buffer (20 mM potassium phosphate, pH 7.0, 1 mM EDTA, 1 mM dithioerythritol, and 0.2 mg ml−1 BSA) in a total volume of 200 μl. Hydrolysis was stopped by addition of 3.25 ml of methanol–chloroform–heptane (10:9:7, by volume) followed, after vortexing for 10 s, by 1.1 ml of 0.1 M potassium carbonate and 0.1 M boric acid (pH 10.5). The mixture was vortexed vigorously for 10 s and centrifuged at 1,100 × g for 20 min. Then 1 ml of the upper phase containing the released fatty acids was mixed with 10 ml of scintillation liquid. Radioactivity was determined in a Tri-Carb 2200 CA (Packard) scintillation counter. One unit of enzyme activity is equivalent to the amount needed for 1 μmol of fatty acid to be released per minute at 37 °C.

Muscle glycogen content was analysed by the hexokinase method (Karlsson et al. 1970). Muscle triacylglycerol content was in principle analysed as described by Kiens and Richter (1996). In the final part of the analysis, the glycerol concentration was analysed in triplicate on a CMA analyzer (CMA 600 microdialysis analyzer, CMA/Microdialysis AB, Stockholm, Sweden). The histochemical determination of muscle TG and the methodological considerations are described in detail elsewhere (Gaster et al. 2003). In brief, 10 μm cryosections of skeletal muscle biopsies were first immunostained for slow myosin using the peroxidase (HRP) labelled streptavidin–biotin technique (LSAB) (Pfeiffer et al. 1996) to ensure fibre type identification. Subsequently, osmium staining was performed by incubating sections in 1% osmium solution (Bie-Berntsen, Copenhagen, Denmark) as osmium stains neutral lipids and cross-links proteins, thereby trapping and visualizing lipid droplets (Fig. 1). All sections used for quantification were stained simultaneously to ensure uniform technical procedures. The sections were blinded to the observer and analysed in a CAST GRID system (Olympus, Denmark). The number of osmium-stained fat droplets in slow and fast fibres were counted using the optical dissector principle (Gundersen 1986). The TG density in each fibre type could be expressed as the number of droplets per microlitre. The fraction of slow and fast fibres was calculated as the number of either slow or fast fibres divided by the total number of counted fibres.

Fig. 1
figure 1

Sections of vastus lateralis muscle double stained with osmium and slow myosin. Fibre type was visualized by an immunoperoxidase reaction and TG droplets by an osmium staining. Representative pictures of muscle before (a) UT, (c) T and after (b) UT, (d) T a 3-h exercise bout in untrained (UT, n = 8) and trained (T, n = 8) subjects exercising at 58% of VO2max. The stained slow fibres present more osmium stained TG than the unstained fast fibres

Full size image

Statistics

Results are given as mean ± S.E.M., if not otherwise stated. One-way and two-way analyses of variance (ANOVAs) with repeated measures for the time factor were performed to test for changes due to training and/or time. In the case of significant main effects or interactions, Student–Newman–Keuls post hoc test was performed to locate statistical differences. In all cases, p < 0.05 was taken as the level of significance in a two-tailed test.

Results

The age, body weight, height, body mass index (BMI), maximal heart rate, muscle LDH activity, and muscle PFK activity were similar between the untrained and the trained groups (Table 1). However, the differences in maximal oxygen uptake, citrate syntheses activity, and HAD activity confirm that training status and muscle oxidative capacity were markedly different between groups (Table 1). The prolonged exercise was performed at a workload of 159 ± 7 and 218 ± 6 W in the untrained and trained groups, respectively (p < 0.05). This was equivalent to relative exercise intensities of 58 ± 1 and 58 ± 1% of VO2max, respectively, and a lower (p < 0.05) oxygen utilization (l min−1) in untrained than in trained subjects. In both groups, five of eight subjects were able to complete the full 180 min of exercise at the given workload. For the remaining three subjects in each group, the workload had to be decreased markedly in order for them to complete the last of the 3 h of exercise. Albeit these subjects exercised from 15 to 60 min at lower relative exercise intensity, they did not exhibit muscle data different from those of the rest of their respective group.

Table 1 Anthropometric data, maximal oxygen uptake, maximal heart rate, and muscle (vastus lateralis) enzyme activity in eight untrained and eight trained men
Full size table

During the exercise, the respiratory exchange ratio (RER) decreased (p < 0.05) steadily across the exercise bout and there was no significant difference between the trained and the untrained groups (Fig. 2a). Carbohydrate oxidation was not significantly different between untrained and trained subjects during exercise (Fig. 2b), whereas a significantly higher fat oxidation was present throughout the exercise in the trained compared with the untrained group (Fig. 2c). Muscle glycogen was not significantly different between trained and untrained subjects before or after the exercise, and despite a numerical difference of approx. 32% in glycogen breakdown, this was also not statistically significant (Fig. 3a). Muscle triacylglycerol content determined biochemically was similar before the exercise, but after the exercise significantly decreased only in the trained subjects (Fig. 3b). Muscle triacylglycerol breakdown was significant only in the trained and not in the untrained group (Fig. 3b).

Fig. 2
figure 2

The RER (a), the carbohydrate oxidation (b), and the fat oxidation (c) during a 3-h submaximal exercise bout at the same relative exercise load in an untrained (n = 8) and a trained (n = 8) groups. Two-way RM ANOVA main effects are indicated on the graphs. *p < 0.05 trained vs. untrained. † p < 0.05 vs. prior sample. ‡ p < 0.05 vs. 30 min

Full size image
Fig. 3
figure 3

Muscle glycogen (a) and triacylglycerol (b) content before and after 3-h submaximal exercise bout as well as breakdown of these substrates during exercise in an untrained (n = 8) and a trained (n = 8) groups exercising at 58% of VO2max. *p < 0.05 trained vs. untrained. † p < 0.05 before vs. after

Full size image

The muscle triacylglycerol was also determined by a histochemical technique (Fig. 1). In type I fibres, there was no difference between groups prior to the exercise (Fig. 4a). The TG stores were higher (p < 0.05) in type I than type II muscle fibres in both trained and untrained subjects (Fig. 4). The percent of type I muscle fibres was similar between groups (Fig. 4c). After the exercise, muscle triacylglycerol in type I muscle fibres was significantly decreased in both groups, and the breakdown did not differ significantly between the groups (Fig. 4a). In type II fibres, muscle triacylglycerol was similar before the exercise and decreased in both groups after the exercise (Fig. 4b). The depletion of muscle TG was not different between the fibre types in untrained muscles, but tended to be higher (p = 0.07, two-way ANOVA interaction) in type I compared with type II fibres in trained muscles.

Fig. 4
figure 4

Histochemically determined muscle triacylglycerol content and breakdown in type I fibres (a), type II fibres (b) before and after a 3-h exercise bout and type I muscle fibre content (c) in an untrained and a trained groups exercising at 58% of VO2max. † p < 0.05 before vs. after

Full size image

HSL (TG) activity, which is a measure of the activated form of HSL, was similar at rest and after the exercise in untrained compared with trained subjects and never changed with the exercise (Fig. 5a). Also HSL (DG) activity, which has been taken as a measure of total enzyme concentration, was similar at rest and after the exercise in untrained compared with trained subjects and never changed with the exercise (Fig. 5b).

Fig. 5
figure 5

HSL activity measured with triacylglycerol as substrate (a) and with diacylglycerol as substrate (b) before and after a 3-h submaximal exercise bout at 58% of VO2max in eight untrained and eight trained men

Full size image

Discussion

In this present study, both HSL (TG) and HSL (DG) activities did not change with the exercise and was similar in untrained and trained human skeletal muscles both before and after the prolonged exercise. Nevertheless, exercise muscle triacylglycerol was utilized and the recruitment of this store tended to be higher in trained compared with untrained subjects. This suggests a difference in the regulation of HSL activity or of the activity of other lipases in skeletal muscles during the exercise in trained compared with untrained human subjects.

In contrast to our hypothesis, neither a difference in HSL expression nor a difference in HSL activation between trained and untrained subjects was observed either before or after the prolonged exercise. The finding of no difference between pre-exercise and post-exercise HSL activities is consistent with published data, where the muscle HSL activity was similar before and after 4 h of submaximal exercise (Watt et al. 2003a). However, in this study by Watt et al. an increase in HSL activity was observed after 1-h exercise. Accordingly, a transient increase in HSL activity might also have been found in this study, if more muscle biopsies had been taken.

However, the finding in this study of similar HSL (TG) and HSL (DG) activities between trained and untrained subjects before and after the exercise in the present study is a novel observation. Unchanged muscle HSL (DG) activity has also been observed in endurance-trained rats (Enevoldsen et al. 2001). Previous studies have demonstrated that the increase in HSL (TG) activity during muscle contraction and exercise can be fully ascribed to HSL (Langfort et al. 2000; Roepstorff et al. 2004), whereas no changes in HSL (DG) activity have been observed during adrenalin stimulation or muscle contraction (Langfort et al. 1999, 2000). Taking these findings into account and by the fact that we found an increased muscle TG breakdown in trained compared with untrained subjects, we should expect to find an increased HSL (TG) activity during the exercise. Nevertheless, we may have missed a difference in HSL (TG) activity between trained and untrained subjects during the exercise bout due to the low frequency of muscle biopsy sampling. However, other mechanisms than increased chemical HSL (TG) activity may also account for the increased TG breakdown during the prolonged exercise. One possibility is that HSL translocates to the lipid droplets to a greater extent in endurance-trained subjects than in sedentary subjects and thereby increase the TG breakdown. Another possibility is that other lipases, particularly the adipose triglyceride lipase (Zimmermann et al. 2004; Villena et al. 2004; Jenkins et al. 2004), could be influenced by training and play a role in muscle TG lipolysis (Zechner et al. 2005).

As expected, we observed a markedly higher fat oxidation in the trained than in the untrained subjects. The contribution of the muscle fat store to total fat oxidation during the exercise has been extensively studied and no breakdown (Starling et al. 1997; Helge et al. 2001; Bergman et al. 1999) as well as breakdown of triacylglycerol in exercising muscle (Hurley et al. 1986; Coyle et al. 1997) have been reported. However, based on the entire literature two recent thorough reviews conclude that muscle TG is recruited, particularly during the prolonged exercise and in trained subjects (Watt et al. 2002; Van Loon 2004). In line with this conclusion, we observed a depletion of histochemically determined triacylglycerol in both type I and II fibres during the exercise. Furthermore, in trained compared with untrained subjects, this depletion tended to be higher in type I fibres, and the chemically determined muscle triacylglycerol breakdown was significantly higher in the trained subjects. In line with our previous studies (Helge et al. 1996), the chemically determined muscle TG recruitment in untrained subjects showed more variation and thus did not become significant. The presence of a significant TG breakdown in trained subjects is consistent with an increased muscle TG storage located around the mitochondria, which makes the triacylglycerol-derived fatty acids more accessible to oxidation (Vock et al. 1996; Hoppeler 1986).

Conclusion

This study has demonstrated that HSL activity is similar in untrained and trained human skeletal muscles before and after 3 h of the submaximal exercise. Nevertheless, during the exercise muscle triacylglycerol depletion tended to be higher in trained compared with untrained subjects, indicating that during the exercise the HSL activity or the activity of other lipases may be regulated differently in trained compared with untrained human muscles.