In vivo experiments were performed in accordance with the National Institute of Health guide for the care and use of Laboratory animals (NIH Publication No. 8023, revised 1978), and were approved by the Cantonal Veterinary Department of Basel, Switzerland (License 1745). 24 male Sprague Dawley rats, weighing ~250 g (6 weeks), were obtained from Janvier Labs (Saint-Berthevin, France), and were acclimatized to the laboratory 1 week prior to the start of the study. The animals were housed in a temperature (20–22°C) and light (12 h light − 12 h dark cycle) controlled animal facility.

THP (N-trimethyl-hydrazine-3-propionate) was prepared by ReseaChem GmbH (Burgdorf, Switzerland) as described previously (Spaniol et al., 2001). After 1 week of acclimatization, 24 rats were randomly divided into two groups as follows: (1) animals treated with water (CTL, n = 12); (2) animals treated with THP (20 mg × 100 g⁻¹ body weight × day⁻¹ for 21 days) (THP; n = 12). THP was dissolved in water and the animals were treated for 3 weeks by oral gavage. Half of the animals of each group performed an exhaustive exercise before sacrifice. Water was provided ad libitum throughout the study. Food consumption and changes in body weight were examined every 3 days (Figures 1A,B). Based upon the findings of our previous study, the rats received a standard chow (Kliba Futter 3433, Basel, Switzerland) ad libitum (carnitine content 16.9 nmol × g⁻¹), as the effect of carnitine-deficient food upon carnitine homeostasis is minor in rats treated with THP (Spaniol et al., 2001).

After 3 weeks of treatment, the rats were submitted to an exhaustive exercise on a treadmill with previous acclimatization to the apparatus (Treadmill Control, Panlab®, Barcelona, Spain). Two days before the sacrifice, rats were accustomed to the treadmill in a daily session of 5-min training at a speed of 40 cm × s⁻¹ with a slope of +5°. On the day of sacrifice, the starting speed was 40 cm × s⁻¹ with a slope of +5°; after 40 min, the speed was increased by 5 cm × s⁻¹ every 2 min until exhaustion (Bouitbir et al., 2011). The rear of the treadmill was equipped with low-voltage, electric stimulating grids, to encourage the rats to run. The grid delivered 0.4 mA at a frequency of 2 Hz, which caused an uncomfortable shock but without injuring the animal. Exhaustion was defined as immobility for more than 5 s on the end lane grid despite electric stimulation (Bouitbir et al., 2011). Blood lactate concentrations were determined in a sample collected from the tip of the tail before, immediately after and 15 min after exercise using a lactate pro-LT device (Lactate Pro LT-1710, ARKRAY®).

Vertical mechanical work and power of the exercise session were calculated according to Isner-Horobeti et al. (2014): (1)Vertical mechanical work (J)=height×m×g (2)Vertical mechanical power (W)=vertical mechanical work/t Where: J = Joule; W = Watt; t = running time in s; height in meter calculated with the Pythagorean theorem; m = animal body weight in kg, and g = gravitational constant in m × s⁻².

After exhaustion, the rats were anesthetized with an intraperitoneal application of ketamine (160 mg × kg⁻¹) and xylazine (20 mg × kg⁻¹). The soleus (oxidative muscle composed predominantly with type I fibers), and the superficial part of the gastrocnemius (glycolytic muscle composed predominantly with type II fibers) were excised and conserved in ice-cold BIOPS buffer (10 mM Ca-EGTA buffer, 0.1 μM free calcium, 5.77 mM ATP, 6.56 mM MgCl₂, 20 mM taurine, 15 mM phosphocreatine, 0.5 mM dithiothreitol, and 50 mM K-MES, pH 7.1) until analysis. A part of the muscle samples was frozen in liquid nitrogen immediately after excision for further analysis.

Blood was obtained by intracardiac puncture or tail incision and collected into heparin-coated tubes. Plasma was separated by centrifugation at 3000 g for 15 min. Plasma and muscle samples were kept at −80°C until analysis.

Creatine kinase activity in plasma was determined before and after exhaustive exercise with a kit according to the supplier's instructions (BioAssay Systems, Hayward, CA, USA). Free and total carnitine concentrations were determined as described previously (Morand et al., 2013).

Mitochondrial respiration measured in situ allows the characterization of functional mitochondria in their normal intracellular position and assembly, preserving essential interactions with other organelles. Mitochondrial oxygen consumption was studied in saponin-skinned fibers (Kuznetsov et al., 2008). Fibers were separated under a binocular microscope in BIOPS buffer at 4°C. After dissection, fibers were transferred into relaxing BIOPS buffer containing 50 μg × mL⁻¹ saponin and incubated at 4°C for 30 min while shaking to complete permeabilization of the sarcolemma. Permeabilized fibers were then washed in BIOPS buffer for 10 min under intense shaking to completely remove saponin. Before oxygraphic measurements, the fibers were washed twice for 5 min in MiR05 buffer (EGTA 0.5 mM, MgCl₂ 3 mM, taurine 20 mM, KH₂PO₄ 10 mM, HEPES 20 mM, D-sucrose 110 mM, BSA essentially fatty acid free 1 g/L, and lactobionic acid 60 mM) to remove any traces of high-energy phosphates. All oxygen measurements were performed at 37°C with an Oxygraph 2k apparatus equipped with the Datlab software (OROBOROS, Innsbruck, Austria).

Wet weight of the fibers (2–3 mg) was determined after permeabilization, which reduces osmotic variations of the water content as described (Pesta and Gnaiger, 2012).

Loosely connected fiber bundles were placed for 5 s onto dry filter paper. During this time period, we wiped off any liquid from the tip of the forceps with another filter paper. Then, we took the sample from the filter paper by a forceps and removed any trace of water on the surface on a dry area of the filter paper before weighing the sample. Immediately weight recording, the sample was transferred into 2.0 mL of MiR05 buffer and analyzed by high-resolution oximetry.

The rate of basal respiration was measured with the complex I substrates glutamate (10 mM) and malate (2 mM), before state 3 respiration was induced with ADP (2 mM). Since oxygen is the electron acceptor, enzyme complexes I, III and IV are used under these conditions. After inhibition of complex I with rotenone (0.5 μM), respiration was restored with addition of the complex II substrate succinate (10 mM). Under these conditions, enzyme complexes II, III, and IV were assessed. To verify the integrity of the outer mitochondrial membrane, cytochrome c (10 μM) was added. Inhibition of complex III with antimycin (2.5 μM) was followed by the addition of artificial substrates for complex IV, TMPD (0.5 mM) and ascorbate (2 mM). We performed the experiments in duplicate. Respiration rates are expressed in picomoles O₂ × s⁻¹ × mg⁻¹ wet weight.

Spectrophotometric analysis of the activity of individual enzyme complexes was performed as described by Spinazzi et al. (2012). In brief, frozen muscle samples were thawed and homogenized in water.

Complex I + III assay was performed in an assay mixture composed of (final concentrations) 50 mM potassium phosphate, 1 g/L BSA, 0.3 mM KCN, 50 μM ferricytochrome c, and 0.5 mM reduced nicotinamide adenine dinucleotide (NADH), pH 7.4. The increase in absorbance was followed at 550 nm. The rotenone-sensitive complex I + III activity was determined by subtracting the activity of identical wells containing in addition 1 μM rotenone.

Complex II activity was measured in an assay mixture containing (final concentrations) 50 mM potassium phosphate, 1 g/L BSA, 0.1 mM Na₂EDTA, 3.75 mM sodium azide, 0.3 mM KCN, 20 mM succinate, 100 μM dichlorophenolindophenol (DCIP), pH 7.2. Changes in absorbance were followed at 600 nm. Malonate sensitive activity was calculated by subtracting the activity of identical wells containing 10 mM malonate.

Complex IV activity was measured in an assay buffer containing (final concentrations) 50 mM potassium phosphate and 17 μM of freshly prepared ferrocytochrome c, pH 7.4. Changes in absorbance were followed at 550 nm. Specific complex IV was calculated by subtracting the activity of identical wells containing 0.3 mM KCN.

Complex activities were determined on a Tecan M200 Pro Infinity plate reader (Männedorf, Switzerland) using a 96-well plate. Values were normalized to citrate synthase activity.

Pieces of frozen muscle (5–10 mg wet weight) were homogenized with a vibrating microbead homogenizer (Mikro-Dismembrator, Sartorius®, Palaiseau, France) in a ratio (W/V) of 1/20 with a buffer containing 5 mM HEPES, 1 mM EGTA and 1 mM DTT, pH 8.7. The homogenate was then supplemented with 0.1% Triton X-100 and incubated on ice for 1 h. After centrifugation for 5 min at 3000 rpm, CS activity was determined in the supernatant by spectrophotometry (Tecan M200 Pro Infinity plate reader, Männedorf, Switzerland) using a 96-well plate as described by (Srere, 1969). Values were reported in μmol × min⁻¹ × g⁻¹ of wet weight.

Mitochondrial H₂O₂ production rates by permeabilized muscle fibers were determined as described previously using Amplex Red reagent (Invitrogen, Switzerland; Anderson and Neufer, 2006). Change in fluorescence per second (ΔF) was measured continuously with a spectrofluorometer (Fluoromax 4 Jobin Yvon®, Edison, New Jersey, USA). First, background ΔF (reactants only) was established before the reaction was initiated by addition of 2–3 mg permeabilized muscle fibers (from soleus or gastrocnemius to 600 μl of buffer Z containing K-methane sulfonate 110 mM, KCl 35 mM, EGTA 1 mM, K₂HPO₄ 10 mM, and MgCl₂ 3 mM with Amplex Red 5 μM and horseradish peroxidase 0.3 U/ml (Invitrogen, Switzerland). The following additions were then made sequentially: the complex I substrate glutamate-malate (5 and 2.5 mM, respectively), the complex II substrate succinate (5 mM), the complex I blocker rotenone (0.5 μM) and the complex III blocker antimycin A (2.5 μM).

The rates of H₂O₂ production were calculated using a standard curve established with the same experimental conditions, except that fibers were absent. The background activity was subtracted from the activity obtained in the presence muscle fibers. H₂O₂ measurements were expressed in pmoles × min⁻¹× mg⁻¹ wet weight.

H₂O₂ production and O₂ consumption were measured simultaneously in the same sample in the presence of glutamate + malate, rotenone, succinate, and ADP allowing to quantify the parameter under respiration supported by the complex II donor succinate under state 3 conditions. This allowed the calculation of the fraction of electrons out of sequence, which reduce O₂ to ROS in the respiratory chain (the percentage of free radical leak) instead of reaching cytochrome oxidase to reduce O₂ to water. Because two electrons are needed to reduce one mole of O₂ to H₂O₂, whereas four electrons are transferred in the reduction of one mole of O₂ to water, the percent of FRL was calculated according to Anderson and Neufer (2006) as the rate of H₂O₂ production divided by twice the rate of O₂ consumption, and then the result was multiplied by a factor of 100.

The GSH (reduced glutathione) content was determined using the luminescent GSH-Glo Glutathione assay (Promega, Wallisellen, Switzerland) according to the manufacture's manual. In brief, muscle tissue (10 mg) was homogenized in 1 mL PBS with 2 mM EDTA with a Mikro-Dismembrator for 1 min at 3000 rpm (Sartorius Stedim Biotech, Göttingen, Germany). After centrifugation, 50 μL supernatant was added to 50 μl GSH-Glo Reagent in a 96-well plate and incubated for 30 min. Afterwards, 100 μl Luciferin Detection Reagent was added to each well and the luminescence was measured after 15 min using a Tecan M200 Pro Infinity plate reader (Männedorf, Switzerland). The GSH standard curve was performed and the results were expressed in μmol × g⁻¹ of wet weight

Muscle tissue (50 mg) was homogenized with a Mikro-Dismembrator for 1 min at 2000 rpm (Sartorius Stedim Biotech, Göttingen, Germany). Muscle glycogen content was analyzed in alkaline (NaOH 0.1 mM) muscle extracts according to Harris et al. (1974). The lactate content was analyzed according to Olsen (1971).

Total RNA was extracted from muscle using a commercially available TRIzol Reagent kit (Invitrogen, Basel, Switzerland). Briefly, 50 mg of muscle were removed from the freezer and immediately immersed in 1 ml of TRIzol reagent. The aqueous and organic phases were separated using 200 μl of chloroform. Total RNA was precipitated using 600 μl of isopropyl alcohol, washed twice with ethanol, redissolved in RNAse free H₂O and stored at −80°C. The concentration and purity of the RNA were determined using a Nanodrop (Thermo Scientific, Wohlen, Switzerland) by measuring the absorbance at 260 and 280 nm. The reverse transcription (RT) was performed on 1 μg of total RNA using a commercially available kit (Invitrogen, Life Technologies, Switzerland). RT was performed in a thermal cycler with a profile of 65°C for 5 min, 42°C for 50 min, and 72°C for 15 min. All samples were run together. Following RT, samples were stored at −20°C until analysis. Real—time PCR measurement of individual cDNAs was performed in triplicate using SYBR green dye (Roche Diagnostics, Rotkreuz, Basel) containing 10 μM of each primer (sense and antisense). Primer sequences were designed using information contained in the public database in the GeneBank of the National Center for Biotechnology Information. The sequences of primer sets used are listed in Table 1. Amplification efficiency of each sample was calculated as previously described (Ramakers et al., 2003), and relative mRNA expression levels were calculated using the ΔΔCT method with beta actin gene as internal control.

To measure the mitochondrial DNA content, we determined the ratio of one mitochondrial gene to a nuclear gene using quantitative real-time RT-PCR as described previously with some modifications (Pieters et al., 2013). Total DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer's instructions (Quick-Start Protocol). The concentration of the extracted DNA was measured spectrophotometrically at 260 nm with the NanoDrop 2000 (Thermo Scientific, Wohlen, Switzerland). Afterwards, DNA was diluted in RNase free water to a final concentration of 10 ng × μL⁻¹. The analysis of the mitochondrial cytochrome b and the nuclear pyruvate kinase genes purchased from Microsynth (Balgach, Switzerland) was performed using SYBR Green real-time PCR (Roche Diagnostics, Rotkreuz, Switzerland) on an ABI PRISM 7700 sequence detector (PE Biosystems, Rotkreuz, Switzerland). Relative amounts of nuclear and mitochondrial DNA were determined by comparison of the amplified and quantified DNA from pyruvate kinase and cytochrome b (Primer sequences in Table 1). The amount of mitochondrial DNA was also normalized to citrate synthase activity.

Data is expressed as mean ± SEM. Comparisons of two groups were performed using the unpaired two-tailed Student's t-test. Multiple groups were compared using one-way ANOVA followed by a Tukey post-test. The software used was Prism version 6 (Graph Pad Software, San Diego, CA). Statistical significance was set at ^*p < 0.05.

As shown in Figure 1A, the body mass was similar between THP-treated and CTL rats during the 21 days of chow feeding. Each group linearly gained ~80 g body weight during the course of the study. Daily food consumption was also not different between the CTL and THP-treated rats (data not shown). The total carnitine concentration was 7.5 ± 0.9 and 55.7 ± 2.8 μmol × L⁻¹ and the plasma free carnitine concentration 4.7 ± 0.6 and 35.9 ± 1.6 μmol × L⁻¹ in THP-treated and control rats, respectively (Figure 1B).

The absolute muscle weight of tendon-to-tendon excised oxidative soleus muscle was reduced by 24% in the THP-treated when compared to CTL rats (143 ± 6 vs. 187 ± 7 mg wet muscle; p < 0.01; data not shown). This decrease in muscle weight was maintained when the soleus weight was expressed relative to the body weight (Figure 1C). In contrast, no significant reduction in absolute weight (1780 ± 60 vs. 1970 ± 60 mg wet muscle in THP-treated and CTL rats, respectively; data not shown) or muscle weight relative to body weight (Figure 1D) was observed for the mainly glycolytic gastrocnemius muscle of THP-treated vs. CTL rats.

We performed an exhaustive treadmill exercise in a subgroup of THP-treated and CTL rats. As shown in Figure 2A, the distance that was covered by animals treated with THP was lower compared to the untreated animals (−39%; p < 0.05; Figure 2A). Since the animals should have covered 960 m during the first 40 min (speed of the treadmill 40 cm × s⁻¹), the data show that the majority of the animals of both groups were exhausted already before 40 min (before increasing the speed of the treadmill). When we calculated the vertical mechanical power of the animals, we observed a significant decrease in the THP group compared to the CTL group (−42%; p < 0.05; Figure 2B).

As shown in Figure 2C, the plasma lactate concentrations increased significantly with exercise in both treatment groups, indicating that the rats exercised above the lactate threshold. The plasma lactate concentration immediately after exercise was significantly higher in the THP compared to the control group (p < 0.05). After 15 min of recovery, the plasma lactate concentration had almost reached pre-exercise values in both treatment groups without a significant difference between the groups. As expected, the plasma creatine kinase activity increased significantly in both THP-treated and CTL rats with exercise (Figure 2D). After exercise, the plasma creatine kinase activity was significantly higher in THP-treated as compared to CTL rats.

Before exercise, the glycogen content was numerically lower in soleus and gastrocnemius muscle of THP-treated compared to CTL rats, but without reaching statistical significance (Figures 3A,B). Exhaustive exercise was associated with a significant decrease in the glycogen content in soleus and gastrocnemius muscle of both THP-treated and CTL rats, but there was no significant difference between the two groups. The lactate content in soleus significantly increased in THP-treated and CTL rats with exercise, reaching significantly higher values for THP-treated compared to CTL rats (Figure 3C). In the gastrocnemius, the lactate content increased in both THP-treated and CTL rats with exercise but without a significant difference between the two groups (Figure 3D).

In order to find out possible reasons for the reduced physical performance of THP-treated rats, we determined the mitochondrial respiration of the electron transport chain in mitochondria of soleus and gastrocnemius muscle. As shown in Figure 4, the maximal respiration rate in the soleus muscle, a red muscle mainly composed of oxidative fibers, was higher than in the gastrocnemius, which consists mainly of glycolytic fibers. In the soleus, treatment with THP was associated with decreased succinate and TMPD/ascorbate-supported respiration, compatible with an impairment of the complexes II and IV, but possibly also III. In comparison, in gastrocnemius, treatment with THP was associated with impaired TMPD/ascorbate-supported respiration (indicating impaired activity of complex IV), whereas complex I, II and III were not affected.

In order to confirm the results obtained by oximetry, we determined the activity of individual enzyme complexes of the electron transport chain in homogenate of soleus and gastrocnemius muscle and related them to the activity of citrate synthase (Table 2). The activity of citrate synthase, a marker of the mitochondrial content (Larsen et al., 2012), was similar in both groups in both muscle types. As shown in Table 2, treatment with THP was associated with a decrease of mitochondrial complex II and IV in soleus, whereas the activity of complex I + III of the electron transport chain were not affected, confirming the results obtained by oximetry. In gastrocnemius, the activities of the enzyme complexes were generally decreased by treatment with THP, but without reaching statistical significance.

Taken together the results obtained by oximetry and spectrophotometry indicate that carnitine deficiency associated with THP specifically impairs the function of complex II and IV in the oxidative, and of complex IV in the glycolytic muscle.

Figure 5A shows the results of H₂O₂ production obtained in permeabilized fibers from the skeletal muscles in response to sequential addition of substrates and inhibitors. Net H₂O₂ release by fibers was low in the absence of respiratory substrates in the both treatment groups, but increased by approximately 2-fold after addition of the complex I substrates glutamate and malate. In line with previous observations (Anderson and Neufer, 2006), the highest rates of H₂O₂ release under basal conditions were observed in the presence of the complex II substrate succinate. The interruption of normal electron flow with addition of the complex I inhibitor rotenone and of the complex III inhibitor antimycin A was associated a substantial increase in H₂O₂ release in both treatment groups. In comparison to CTL rats, H₂O₂ production in THP-treated rats was similar under all states investigated for both soleus and gastrocnemius muscle (Figures 5A,B).

The free radical leak (FRL), described as a good index for ROS production (Anderson and Neufer, 2006), was significantly higher in the THP group compared to controls in both muscles in the presence of the complex II substrate succinate (Figures 5C,D). Interestingly, carnitine deficiency enhanced FRL in gastrocnemius muscle (+210%, Figure 5D), whereas the FRL increased robustly in soleus (+309%, Figure 5C) with an increase significantly higher in soleus muscle compared with gastrocnemius.

To study the antioxidative defense system, we analyzed the mRNA content of cytosolic SOD1 and mitochondrial SOD2 by q-RT-PCR as well as the GSH content in soleus and gastrocnemius muscle. As shown in Figures 6A,B, the SOD1 mRNA expression level was similar in both treatment groups in both muscles. Interestingly, the SOD2 mRNA expression level was increased in soleus of THP-treated as compared to CTL rats, reaching borderline significance (Figure 6A; p = 0.057). In gastrocnemius, the SOD2 mRNA expression level was also numerically higher in the THP-group compared to the CTL group (Figure 6B; p = 0.0794). The GSH content, which is an important defense against ROS accumulation, was decreased in soleus muscle of THP-treated compared to CTL rats (Figure 6C; p < 0.05) but was similar between the treatment groups in gastrocnemius (Figure 6D). Since we did not determine oxidized glutathione (GSSG), this finding could be explained by both increased oxidation of GSH and/or impaired GSH biosynthesis.

Impaired physical performance in THP-treated rats could not only be a consequence of mitochondrial damage, but also of an impaired mitochondrial biogenesis. We therefore determined PGC-1α and PGC-1β mRNA expression levels, both important transcriptional co-activators of numerous genes coding for mitochondrial proteins. mRNA expression of PGC-1α and PGC-1β were lower in the soleus of THP-treated than CTL rats (p < 0.01 and p < 0.05, respectively; Figure 7A). Consequently, mitochondrial TFAm mRNA expression, which regulates the amount of mtDNA, was also lower in THP-treated compared to the CTL rats (p < 0.05; Figure 7A). In contrast, in the gastrocnemius, PGC-1β mRNA expression was increased in THP-treated compared to CTL rats (p < 0.05; Figure 7B), whereas the mRNA expression of PGC-1α and of TFAm were not different between the two groups.

In soleus muscle, in agreement with the mRNA expression of PGC-1α and PGC-1β, both the amount of mitochondrial DNA expressed per nuclear DNA (p < 0.05; Figure 7C) and per citrate synthase activity (p < 0.05; Figure 7E) was decreased in THP-treated compared to control rats. In comparison, in gastrocnemius muscle, the mitochondrial DNA content was not affected by treatment with THP (Figures 7D,F).

The evaluation of the exercise capacity showed that both the distance covered until exhaustion and vertical mechanical work were reduced in THP-treated rats. Regarding skeletal muscle performance, carnitine has two main functions. First, it is an obligatory intermediate for the transport of long-chain fatty acids across this inner mitochondrial membrane (Fritz and Mc, 1959; Bremer, 1983) and, second, carnitine stimulates the activity of pyruvate dehydrogenase (Wall et al., 2011) by buffering the mitochondrial CoASH pool (Brass and Hoppel, 1980; Friolet et al., 1994). As indicated by the increased lactate blood and skeletal muscle concentrations and by the reduced skeletal muscle glycogen stores, rats exercised above the lactate threshold during the physical performance experiment, suggesting that mainly glucose and to a minor extent fatty acids have been the main fuel for skeletal muscle under these conditions (Gollnick et al., 1974; Roberts et al., 1996). It can therefore be assumed that the effects of carnitine on glucose metabolism have been more important in the current situation than those on fatty acid metabolism. Accordingly, the observed decrease in physical performance is probably mainly due to an impaired function of glycolytic and to a minor extent of oxidative muscles. In a previous study using the same animal model of carnitine deficiency, we have indeed shown in in vitro experiments that the maximal contraction of the mainly glycolytic extensor digitorum longus (EDL) muscle is impaired (Roberts et al., 2015). This observation fits well with the current findings. The most probable explanation for this finding was a marked decrease in the EDL muscle free CoA content in rats treated with THP, leading to a reduced flux through pyruvate dehydrogenase (Roberts et al., 2015). In the current studies, we additionally found a significant impairment in the activity of complex IV in the gastrocnemius muscle of THP treated rats, offering an additional explanation regarding the observed reduction in exercise capacity. Furthermore, we found an increased free radical leak, possibly as a consequence of the impaired activity of the mitochondrial electron transport chain (Balaban et al., 2005). The cellular glutathione content was not affected in gastrocnemius muscle of THP-treated rats, suggesting that the increase of the free radical leak was not large enough to impair the antioxidative defense system.

In comparison to the glycolytic gastrocnemius muscle, the effects of carnitine depletion were more pronounced on the oxidative soleus muscle. Nevertheless, as discussed above, these findings may contribute, but cannot entirely explain the observed decrease in exercise performance in THP-treated rats, since the rats were exercising above the lactate threshold. Similar to the gastrocnemius, carnitine depletion was associated with increased lactate content after exercise in the soleus muscle. Since, in our previous study (Roberts et al., 2015), the free CoA content was only numerically decreased and the ex vivo pyruvate dehydrogenase activity maintained in soleus muscle of THP-treated rats, this finding is explained best by the observed impairment of complex II and IV of the electron transport chain, which was more pronounced in the oxidative soleus compared to the glycolytic gastrocnemius muscle. Impaired activity of the electron transport chain is associated with an increase in the mitochondrial NADH/NAD ratio, which could in vivo impair the activity of pyruvate dehydrogenase (Patel et al., 2014). Similar to the gastrocnemius muscle, we observed an increase in ROS production relative to oxygen used (increased free radical leak), which was associated with an increase in the mRNA expression of SOD2 and with a decrease in the cellular GSH content. A possible interpretation of these findings is that the ROS production in soleus muscle was high enough to reduce the cellular glutathione pool and to induce SOD2 mRNA expression. Increased mRNA expression of SOD2 as a consequence of enhanced ROS production has been described previously (Kensler et al., 2007; Felser et al., 2013). Since carnitine has been described to possess antioxidative properties (Gülçin, 2006), carnitine deficiency could decrease the antioxidative capacity, leading to oxidative stress already at normally tolerated ROS production levels.

Increased ROS production has been described not only to up-regulate the antioxidative defense system including the Nrf2 pathway, and SOD2 (Kensler et al., 2007), but also to stimulate mitochondrial proliferation (Piantadosi and Suliman, 2012). Accordingly, in the gastrocnemius muscle, the mRNA expression of mitochondrial proliferation regulators (PGC-1α, PGC-1β, TFAm) was numerically increased, with PGC-1β reaching statistical significance. This increase in the mRNA expression of mitochondrial proliferation regulators did, however, not translate into a significant increase in mtDNA. In contrast, in the oxidative soleus muscle, which is rich in mitochondria, mRNA expression of the above mentioned regulators of mitochondrial proliferation was significantly decreased in THP-treated rats with a consequent significant decrease in the mtDNA related to nuclear DNA or citrate synthase activity. The reasons for the different reaction of the oxidative soleus vs. the glycolytic gastrocnemius muscle to carnitine depletion are currently not clear, but may be related to the more accentuated mitochondrial damage and to muscle atrophy that we observed in soleus muscle in the current as well as in our previous study (Roberts et al., 2015). In our previous study, we could demonstrate that soleus muscle atrophy in THP-treated rats is a consequence of apoptosis, which is triggered most likely by mitochondrial dysfunction (Roberts et al., 2015).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.