Genetic mutations that affect mitochondrial function often cause skeletal muscle dysfunction. Here, we used mice with skeletal-muscle-specific disruption of the nuclear gene for mitochondrial transcription factor A ( Tfam ) to study whether changes in cellular Ca 2+ handling is part of the mechanism of muscle dysfunction in mitochondrial myopathy. Force measurements were combined with measurements of cytosolic Ca 2+ , mitochondrial Ca 2+ and membrane potential and reactive oxygen species in intact, adult muscle fibres. The results show reduced sarcoplasmic reticulum (SR) Ca 2+ storage capacity in Tfam KO muscles due to a decreased expression of calsequestrin-1. This resulted in decreased SR Ca 2+ release during contraction and hence lower force production in Tfam KO than in control muscles. Additionally, there were no signs of oxidative stress in Tfam KO cells, whereas they displayed increased mitochondrial [Ca 2+ ] during repeated contractions. Mitochondrial [Ca 2+ ] remained elevated long after the end of stimulation in muscle cells from terminally ill Tfam KO mice, and the increase was smaller in the presence of the cyclophilin D-binding inhibitor cyclosporin A. The mitochondrial membrane potential in Tfam KO cells did not decrease during repeated contractions. In conclusion, we suggest that the observed changes in Ca 2+ handling are adaptive responses with long-term detrimental effects. Reduced SR Ca 2+ release likely decreases ATP expenditure, but it also induces muscle weakness. Increased [Ca 2+ ] mit will stimulate mitochondrial metabolism acutely but may also trigger cell damage.
Mitochondrial diseases are caused by genetic mutations influencing the function of mitochondrial respiratory chain ( 1 , 2 ). The clinical presentation of mitochondrial diseases is highly variable with respect to organs or tissues that are affected and the age of disease onset, which may vary from early infancy to late adulthood ( 3 , 4 ). Nevertheless, symptoms generally occur in tissues with high-energy demand, such as brain, heart and skeletal muscle ( 5 ). Interestingly, mitochondrial defects are also implicated in many common age-related degenerative diseases, e.g. heart failure, Parkinson disease and cancer ( 6 ), as well as in the normal ageing process ( 7 , 8 ).
Mice with skeletal-muscle-specific disruption of the nuclear gene for mitochondrial transcription factor A ( Tfam KO mice) display important hallmarks of mitochondrial myopathy, i.e. ragged-red muscle fibres, accumulation of abnormally appearing mitochondria, progressively deteriorating respiratory chain function and reduced phosphocreatine concentration ( 9 , 10 ). Tfam is essential for mitochondrial DNA maintenance, transcription and replication ( 11 ), and disruption of this gene results in progressive impairment of respiratory chain function.
Muscle weakness and exercise intolerance are prominent symptoms in patients with mitochondrial myopathies ( 12 ). We have previously shown a reduced force production in the non-fatigued state in Tfam KO muscles ( 9 ), which in vivo would be manifested as a general muscle weakness and exercise intolerance, because muscles have to work at a higher fraction of their maximal capacity ( 13 ). The decreased force production in Tfam KO muscle may be due to severe dysfunction in some muscle fibres (e.g. the ragged-red fibres), resulting in almost complete loss of force in these. Alternatively, there may be a generalized force decrease in muscle fibres that, in principle, can be due to decreased ability of cross-bridges to produce force, decreased myofibrillar Ca 2+ sensitivity and/or decreased sarcoplasmic reticulum (SR) Ca 2+ release ( 14 ). In the present study, we distinguish between these alternatives by analysing single muscle fibres that allow simultaneous measurements of force and free cytosolic [Ca 2+ ] ([Ca 2+ ] i ).
Mitochondrial diseases have been suggested to involve three fundamental pathophysiological mechanisms: (i) impaired energy production, (ii) increased production of reactive oxygen species (ROS) resulting in oxidative stress and (iii) excessive accumulation of Ca 2+ in the mitochondria leading to cell damage or death, possibly involving opening of the mitochondrial permeability transition pore (mtPTP) ( 15 , 16 ). We have previously shown that the impaired energy production in the mitochondria of Tfam KO muscles is counteracted by a markedly increased mitochondrial mass ( 9 ). In the present study, we measure ROS levels, signs of oxidative stress and changes in [Ca 2+ ] in the mitochondrial matrix ([Ca 2+ ] mit ). Our results show that Tfam KO muscle cells have a decreased SR Ca 2+ content that leads to reduced tetanic [Ca 2+ ] i and force. Moreover, they do not display any signs of increased ROS production or oxidative stress, whereas they show a large increase in [Ca 2+ ] mit during repeated contractions.
Tetanic [Ca 2+ ] i is lower in Tfam KO skeletal muscle fibres
We measured [Ca 2+ ] i and force in single intact fast-twitch flexor digitorum brevis (FDB) muscle fibres. [Ca 2+ ] i in resting fibres was not different between Tfam KO and control cells [71 ± 6 ( n = 6) versus 75 ± 9 n m ( n = 7)]. Figure 1 shows [Ca 2+ ] i and force records from 100 Hz tetani produced in a Tfam KO (A) and a control (B) fibre, and mean data at different frequencies are shown in Figure 1 C and D. Tetanic [Ca 2+ ] i was significantly lower in Tfam KO than in control fibres at all stimulation frequencies, and force was significantly lower at 40–100 Hz ( P < 0.05). Tetanic stimulation in the presence of caffeine can be used to assess the SR Ca 2+ load and force at saturating [Ca 2+ ] i ( 17 ). Tetanic [Ca 2+ ] i during 100 Hz stimulation in the presence of 5 m M caffeine was significantly lower in Tfam KO than in control cells ( P < 0.05), and tetanic force tended to be lower in Tfam KO cells ( P = 0.06).
The force–[Ca 2+ ] i relationships of mean data from Tfam KO and control cells are plotted in Figure 1 E. The relationship for individual cells was established by plotting force ( P ) versus [Ca 2+ ] i in tetanic contractions of different frequencies, and data points were then fitted to the following Hill equation:m ) nor N (3.71 ± 0.55 versus 3.13 ± 0.50) differed significantly between Tfam KO and control cells, whereas Pmax was ∼15% lower in Tfam KO cells (366 ± 11 versus 429 ± 25 kPa; P < 0.05). Taken together, the dominating factor behind the lower force production in Tfam KO muscle fibres was decreased [Ca 2+ ] i during contractions. The effect of decreased [Ca 2+ ] i has the largest impact at lower stimulation frequencies, where the force–[Ca 2+ ] i relationship is steep. This is illustrated by the fact that in comparison with control cells, the mean force in Tfam KO cells was 43% lower at 20 Hz but only 18% lower at 100 Hz.
The slow decay (tails) of elevated [Ca 2+ ] i after the end of tetanic stimulation can be used to assess differences in SR Ca 2+ leakage and pumping ( 14 , 18 , 19 ). Figure 1 F shows mean [Ca 2+ ] i records obtained immediately after the end of 100 Hz tetani. The tails of elevated [Ca 2+ ] i were very similar in Tfam KO and control fibres, which indicates that the difference in tetanic [Ca 2+ ] i between the two groups is not due to differences in SR Ca 2+ leakage or pumping.
Tfam KO muscle fibres display decreased expression of calsequestrin-1
We used real-time PCR (RT-PCR) to quantify the expression of transcripts encoding for important proteins involved in SR Ca 2+ handling, and mean data are shown in Figure 2 A. The mRNA expression of the t-tubular voltage sensor [the dihydropyridine receptor (DHPR)], the SR Ca 2+ release channel (the ryanodine receptor 1; RyR1) and the RyR-associated proteins triadin and junction were similar in Tfam KO and control muscles. The mRNA expression of the dominant SR Ca 2+ pump in fast-twitch muscle (SERCA1) and the Na + /Ca 2+ exchanger (NCX) were also similar in the two groups. However, the mRNA expression of the SR luminal Ca 2+ binding protein calsequestrin-1 (CASQ1) was ∼60% lower in Tfam KO when compared with control muscles ( P < 0.05). This indicates a decreased SR Ca 2+ storage capacity in Tfam KO fibres, which is in accordance with the decreased tetanic [Ca 2+ ] i that we measured.
In agreement with the RT-PCR results, protein measurements with western blot showed a significantly lower expression of CASQ1 in Tfam KO than in control muscle (Fig. 2 B). To confirm that this did not reflect a general decrease in SR Ca 2+ handling proteins in Tfam KO muscle, we also measured the expression of DHPRs and found no difference between the two groups (Fig. 2 C).
Muscles of Tfam KO mice show a progressive deterioration in respiratory chain function over their ∼4-month life span ( 9 ). We measured the mRNA expression of CASQ1 in muscles of 1- and 2-month-old Tfam KO mice to elucidate whether the decrease in CASQ1 showed a similar progress. There was no difference in CASQ1 mRNA expression between Tfam KO and control muscles at 1 month (102 ± 2 versus 100 ± 3%; n = 3) and a tendency ( P = 0.08) for a decrease at 2 months (89 ± 4 versus 100 ± 2%; n = 3). Thus, a significant decrease in CASQ1 expression was only observed in muscles of 4-month-old Tfam KO mice.
Fatigue properties are not altered in Tfam KO muscle fibres
Figure 3 shows mean data obtained during fatigue induced by a series of 50 repeated tetanic stimulations. Tetanic [Ca 2+ ] i and force were generally lower in Tfam KO than in control fibres throughout the fatigue run (Fig. 3 A). However, the absolute difference between the two groups became smaller as fatiguing stimulation progressed. This is clearly observed in Figure 3 B, where data were normalized to the tetanic [Ca 2+ ] i and force in the first tetanus. It can then be seen that tetanic [Ca 2+ ] i was actually better maintained in Tfam KO cells, whereas the two groups showed virtually identical decreases in normalized tetanic force.
Fatigue is commonly accompanied by changes in the force–[Ca 2+ ] i relationship with reductions in maximum force and myofibrillar Ca 2+ sensitivity ( 13 ). Possible differences between Tfam KO and control cells with regard to fatigue-induced changes in the force–[Ca 2+ ] i relationship were assessed by plotting mean data of tetanic force versus [Ca 2+ ] i obtained during fatiguing stimulation. These plots show a similar pattern in the two groups with a marked fatigue-induced right-ward shift, indicative of decreased Ca 2+ sensitivity both in Tfam KO (Fig. 3 C) and control cells (Fig. 3 D).
The resting [Ca 2+ ] i increased ∼3-fold in both groups during fatigue, reaching 170 ± 17 n m in Tfam KO and 190 ± 31 n M in control fibres at the end of fatiguing stimulation. Recovery of tetanic [Ca 2+ ] i and force after fatiguing stimulation was similar in the two groups. At 30 min of recovery, tetanic [Ca 2+ ] i was 87 ± 6 and 71 ± 12% of the original in Tfam KO and control fibres, respectively ( P > 0.05); the corresponding tetanic forces were 74 ± 16 and 72 ± 9% ( P > 0.05). To sum up, we observed no signs of faster fatigue development in Tfam KO fibres when compared with control fibres.
Mitochondrial ROS production is not increased in Tfam KO skeletal muscle fibres
Changes in mitochondrial function have been suggested to increase ROS production and cause oxidative stress, which have been proposed to have a central role in the pathology of mitochondrial diseases ( 12 , 20 ). Redox centres in the electron transport chain leak electrons to oxygen, which results in the formation of superoxide (O 2*− ) ( 21 ). We used the fluorescent indicator MitoSOX Red to measure changes in mitochondrial [O 2*- ] ( 22 ). The MitoSOX Red fluorescence showed no significant increase during fatiguing stimulation in either Tfam KO or control muscle fibres (Fig. 4 A, left).
The lack of increase of MitoSOX Red fluorescence in Tfam KO fibres during severe metabolic stress induced by fatiguing stimulation was unexpected as increased ROS production is considered a hallmark in mitochondrial diseases ( 12 , 15 , 16 , 20 ). We therefore performed control experiments to ensure that MitoSOX Red could respond to increased ROS levels in our cells and that the response depended on mitochondrial respiration. The mitochondrial matrix contains a superoxide dismutase with manganese at the active site (MnSOD or SOD2), which converts O 2*− to hydrogen peroxide (H 2 O 2 ) ( 23 ). To assess the rate of mitochondrial O 2*− production, we performed experiments in which muscle fibres were exposed to 100 µ m H 2 O 2 , which would increase the mitochondrial [O 2*− ] by inducing product inhibition of SOD2 and thereby inhibit the conversion of O 2*− to H 2 O 2 ( 24 , 25 ). H 2 O 2 application resulted in a marked increase in the MitoSOX Red signal, and the amplitude of this H 2 O 2 -induced increase was ∼65% lesser in Tfam KO when compared with control fibres ( P < 0.01; Fig. 4 A, middle). Inhibition of mitochondrial complex I by acute exposure to rotenone has been shown to decrease mitochondrial ROS production under some conditions ( 26 ). Muscle fibres of control mice were exposed to H 2 O 2 in the presence of 2.4 µ M rotenone, and the H 2 O 2 -induced increase in MitoSOX Red fluorescence was then markedly reduced (Fig. 4 A, right).
Additional control experiments were performed with the general ROS indicator 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein (CM-H 2 DCF) ( 27 ). Application of 100 µ m H 2 O 2 resulted in an increase in CM-H 2 DCF fluorescence that was faster than with MitoSOX Red; the time for half-maximum increase was 3.1 ± 0.2 min ( n = 32) versus 5.7 ± 0.2 min ( n = 12; P < 0.001). The increase in CM-H 2 DCF fluorescence induced by H 2 O 2 was not significantly affected by rotenone [ F / F0 7.1 ± 0.7 without ( n = 20) and 5.4 ± 0.7 with ( n = 7) rotenone; P = 0.22]. Taken together, these control experiments show that, in contrast to the CM-H 2 DCF signal, changes in MitoSOX Red fluorescence depend on mitochondrial respiration and reflect the rate of mitochondrial O 2*− production. Thus, the results indicate that mitochondrial O 2*− production was, if anything, decreased in Tfam KO cells when compared with controls.
Oxidative stress can induce the expression of SOD2 and other ROS-detoxifying enzymes ( 28 ). However, SOD2 protein expression did not differ between Tfam KO and control muscles (Fig. 4 B). Furthermore, we measured the activity of SOD2, as well as the copper–zinc-dependent SOD1 and total SOD activity, and found no significant difference between Tfam KO and control muscles (Fig. 4 C), thus supporting the conclusion that oxidative stress is not increased.
Oxidative stress is associated with decomposition of polyunsaturated fatty acids, which leads to the formation of reactive carbonyl species, which can bind to protein ( 29 ). We used immunoblotting to detect protein binding of two such species, malondialdehyde (MDA) and 4-hydroxy-trans-2-nonenal (HNE), and the results show no increased binding in Tfam KO when compared with control muscles (Fig. 4 D and E). In summary, none of several different methods showed any signs of increased mitochondrial ROS production or oxidative stress in Tfam KO muscle cells.
Mitochondrial [Ca 2+ ] increases during fatigue in Tfam KO muscle fibres
Modest and transient increases in [Ca 2+ ] mit stimulate mitochondrial ATP production ( 30–32 ), whereas a prolonged increase may be a potent trigger of cellular damage ( 16 , 33 , 34 ). We used the fluorescent indicator Rhod-2 and a confocal microscope to measure [Ca 2+ ] mit during fatigue and recovery in single Tfam KO and control fibres. Figure 5 A shows original confocal images of Rhod-2 fluorescence obtained in a Tfam KO and a control fibre before, during and after fatigue induced by 50 repeated tetani. In contrast to the control fibre, the Tfam KO fibre showed a clear-cut increase in Rhod-2 fluorescence during fatigue, which had not returned to the baseline at 5 min after the end of stimulation. The striated fluorescence pattern is typical for mouse FDB fibres, in which the Rhod-2 containing mitochondria are arranged in rows on each side of the z-lines ( 35 , 36 ). Figure 5 B shows mean data and the control fibres displayed no change in [Ca 2+ ] mit , during fatigue or recovery, which is consistent with previous measurements in wild-type FDB fibres ( 36 , 37 ). In contrast, [Ca 2+ ] mit showed a 3–4-fold increase during fatigue in Tfam KO cells. [Ca 2+ ] mit recovered rather slowly in Tfam KO fibres of 4-month-old mice (the age used in other experiments in this study), and even 10 min after the end of fatiguing stimulation, it was still increased above the pre-fatigue value by ∼60% ( P < 0.05). We also studied Tfam KO fibres of 2-month-old mice, in whom the muscle pathology is markedly less severe than in the terminally ill 4-month-old animals. Interestingly, the recovery of [Ca 2+ ] mit after fatigue was much faster in Tfam KO fibres of these younger mice, and [Ca 2+ ] mit had recovered completely to its starting value at 10 min of recovery.
Mitochondrial dysfunction has a central role in Ullrich congenital muscular dystrophy and Bethlem myopathy, which are myopathies with mutations in the genes encoding collagen VI. Recent studies show that the cyclophilin D-binding inhibitor cyclosporin A (CSA) normalizes mitochondrial function and decreases apoptosis in muscles of a mouse model and of patients with these myopathies; the positive effects of CSA were ascribed to the inhibition of inappropriate opening of the mtPTP ( 38–40 ). We studied the effect of CSA (1.6 µ m ) on the increase in [Ca 2+ ] mit observed during fatigue in Tfam KO and control fibres. The results show a ∼40% smaller contraction-induced increase in [Ca 2+ ] mit in Tfam KO fibres in the presence of CSA ( P < 0.05), whereas CSA had no effect in control fibres (Fig. 5 C).
As described earlier, mitochondria of Tfam KO fibres but not control fibres take up Ca 2+ during tetanic contractions. The lower tetanic [Ca 2+ ] i in Tfam KO fibres than in controls (Fig. 1 ) might then, in principle, be due to mitochondrial uptake of a significant amount of Ca 2+ released from the SR. To test this possibility, we measured tetanic [Ca 2+ ] i in Tfam KO fibres in the presence and absence of CSA, which inhibited the mitochondrial Ca 2+ uptake. [Ca 2+ ] i in 70 Hz tetani was similar in the presence [0.51 ± 0.07 µ m ( n = 10)] and absence [0.61 ± 0.09 µ M ( n = 9)] of CSA ( P > 0.05). Thus, inhibition of the mitochondrial Ca 2+ uptake with CSA had no significant effect on tetanic [Ca 2+ ] i in Tfam KO fibres, which indicates that the lower tetanic [Ca 2+ ] i was not due to mitochondrial Ca 2+ uptake.
Mitochondrial membrane potential in Tfam KO muscle fibres is not decreased during fatigue
Excessive increases in [Ca 2+ ] mit are frequently associated with a decreased mitochondrial membrane potential (ΔΨ m ) ( 16 ). We used the fluorescent indicator tetra-methyl rhodamine-ethyl ester (TMRE) to assess possible fatigue-induced changes in ΔΨ m in muscle cells of 4-month-old Tfam KO mice. After 50 fatiguing tetani, the TMRE fluorescence signal was 99 ± 3% ( n = 7) of the pre-fatigue value. Subsequent exposure to the mitochondrial uncoupler carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP; 4 µ m ) decreased the TMRE fluorescence to 31 ± 6% of the control in 10 min. Thus, there was no fatigue-induced decline in ΔΨ m in Tfam KO fibres.
Mterf3 KO muscle fibres with limited mitochondrial dysfunction have normal cellular Ca 2+ handling
The mitochondrial transcription termination factor 3 (MTERF3) protein is a negative regulator of mtDNA transcription initiation ( 41 ). Mice with conditional inactivation of the MTERF3 gene in heart and skeletal muscles develop fatal cardiomyopathy, whereas skeletal muscle function seems to be relatively well preserved at death ( 41 ). We used these Mterf3 KO to study whether mitochondrial Ca 2+ uptake occurs also in skeletal muscle cells that display limited mitochondrial abnormalities.
In the heart, Mterf3 KO causes aberrant mtDNA transcription, resulting in a decrease of mtDNA-encoded proteins and severe respiratory chain deficiency ( 41 ). Skeletal muscle of Mterf3 KO mice showed significant mitochondrial RNA and protein changes similar to those observed in the heart, although less severe. There was a general increase in levels of mitochondrial mRNAs transcribed from the heavy-strand promoter, whereas the only mRNA transcribed from the light-strand promoter, ND6, was decreased; the levels of the mitochondrial tRNA-phe and tRNA-glu were increased; the mtDNA-encoded complex IV subunit II protein levels were decreased, whereas the nuclear-encoded complex II subunit A protein levels were unchanged ( Supplementary Material, Fig. S1 ).
[Ca 2+ ] mit did not increase during 50 repeated tetani in FDB fibres of either Mterf3 KO or their littermate control mice (Fig. 6 A). Furthermore, the [Ca 2+ ] i and force–frequency relationships did not differ between Mterf3 KO and control muscle fibres in the unfatigued state ( P > 0.05; Fig. 6 B), and the two groups showed similar changes in tetanic [Ca 2+ ] i and force during fatigue induced by repeated 70 Hz tetani ( P > 0.05; Fig. 6 C). These findings indicate that the mild respiratory chain deficiency present in Mterf3 KO muscle does not cause aberrant Ca 2+ homeostasis and myopathy.
We demonstrate here that mitochondrial myopathy in the mouse causes major changes in the cellular Ca 2+ handling with decreased [Ca 2+ ] i during contraction causing reduced force production. Unexpectedly, there are no signs of increased mitochondrial ROS production or oxidative stress in Tfam KO cells. The findings raise important questions concerning the pathophysiology of the abnormal Ca 2+ handling in respiratory chain-deficient skeletal muscle.
How does the reduced force production relate to the abnormal Ca 2+ homeostasis?
We found lower force production in single FDB fibres of Tfam KO than control mice. However, the force was more decreased in whole extensor digitorum longus (EDL) muscles ( 9 ) than in single fibres; for instance, the force in 100 Hz tetani was about 40 and 20% lower than the control in whole muscles and single fibres, respectively. Thus, our results indicate that the reduced force in Tfam KO muscle is due to a combination of complete loss of force generation in some degenerated fibres and a generalized decrease in the remaining fibres.
The lower force in single FDB fibres of Tfam KO mice was mainly due to decreased [Ca 2+ ] i during contraction. A lower [Ca 2+ ] i in Tfam KO was also observed when 100 Hz stimulation was applied in the presence of caffeine. Caffeine facilitates SR Ca 2+ release ( 17 ), and hence a decreased tetanic [Ca 2+ ] i in the presence of caffeine reflects a reduced amount of SR Ca 2+ available for release. There was a specific down-regulation of the skeletal muscle isoform of calsequestrin, CASQ1, in Tfam KO muscle (Fig. 2 ). The classical role of CASQ1 is as a high capacity, low affinity Ca 2+ buffer which ensures that the free [Ca 2+ ] in the SR lumen remains close to 1 m m during contraction where Ca 2+ is first released from the SR and subsequently pumped back ( 42 ). More recent studies show that CASQ1 and SR Ca 2+ can also regulate the SR Ca 2+ release channels and in this way affect [Ca 2+ ] i and force during contraction ( 42 ). Thus, less CASQ1 in Tfam KO muscle can decrease SR Ca 2+ release by reducing the SR Ca 2+ storage and possibly by direct effects on the SR Ca 2+ release channels. In accordance with this, fast-twitch muscle cells of mice lacking CASQ1 show markedly smaller [Ca 2+ ] i increases in response to electrical stimulation and caffeine application ( 43 ).
Mitochondrial Ca 2+ uptake has been shown to buffer [Ca 2+ ] i transients in frog skeletal muscle fibres ( 36 ), in neurons ( 44 , 45 ) and in several other non-excitable cell types ( 46–48 ). This might imply that the increased mitochondrial Ca 2+ uptake in Tfam KO cells contributes to the decrease tetanic [Ca 2+ ] i seen in these cells. However, this appears not to be the case because: (i) increased Ca 2+ buffering results in slowed [Ca 2+ ] i transients in muscle fibres ( 36 , 49 ) and no such slowing was observed in Tfam KO fibres; (ii) altered Ca 2+ buffering would affect the tails of elevated [Ca 2+ ] i after tetanic stimulation ( 50 ) and there was no difference between Tfam KO and control cells in this respect; (iii) although [Ca 2+ ] mit increased during a series of repeated contractions in Tfam KO cells, tetanic [Ca 2+ ] i was actually better maintained in Tfam KO than in control muscle cells; (iv) CSA decreased the contraction-induced increase in [Ca 2+ ] mit but had no effect on tetanic [Ca 2+ ] i . Thus, these results indicate that mitochondrial Ca 2+ uptake does not contribute to the observed decrease in tetanic [Ca 2+ ] i in Tfam KO cells. Inline with this, mitochondria do not contain any high capacity Ca 2+ binding proteins, and Ca 2+ contained in the mitochondria is only a small fraction (<2%) of that in the SR ( 13 ).
Defective mitochondrial function induces complex changes in the expression of numerous proteins, including several proteins involved in cellular Ca 2+ homeostasis ( 51 , 52 ). It is then intriguing that we observed a specific down-regulation of CASQ1 in skeletal muscles of 4-month-old Tfam KO mice, and the mechanism underlying this change remains uncertain. The fact that CASQ1 mRNA expression was not significantly decreased in 1- and 2-month-old Tfam KO mice shows that the down-regulation is related to the progressive decline in respiratory chain function and/or to the prolonged increase in [Ca 2+ ] mit observed in the 4-month-old animals. Interestingly, skeletal muscle of CASQ1 KO mice shows a markedly increased mitochondrial content ( 43 ), which further illustrates the complex interplay between mitochondria and cellular Ca 2+ homeostasis.
Muscle weakness and exercise intolerance
Exercise intolerance is a key symptom in human mitochondrial myopathies ( 12 ). We observed no difference between the Tfam KO and control single fibres regarding the rate of force decrease during fatigue. This unexpected lack of effect on fatigue development in Tfam KO muscle can be explained by increased mitochondrial mass ( 9 ) and increased [Ca 2+ ] mit stimulating mitochondrial respiration (present results) that partially compensates for the respiratory chain defects. In addition, the energy consumption during fatiguing stimulation is reduced in Tfam muscle fibres due to their lower tetanic [Ca 2+ ] i and force.
Exercising humans experience fatigue when the intended physical activity can no longer be continued or when it is accompanied by the sensation of excessive effort and discomfort. During exercise in vivo , the muscle weakness observed in Tfam KO muscle would lead to early fatigue development because muscles always have to work at a higher fraction of their maximal capacity ( 13 ).
No signs of increased mitochondrial ROS production or oxidative stress
It is often assumed that increased ROS production and oxidative stress are important factors underlying the pathological changes in primary mitochondrial diseases ( 12 , 20 ). However, recent studies on tissue from mitochondrial disease models have failed to show an increase in ROS production or ROS-induced damage to cellular components. ( 8 , 53 ). The present measurements of mitochondrial O 2*− showed no signs of an increased production in Tfam KO cells either at rest or during fatigue. Increased ROS production and oxidative stress are associated with increased expression of ROS-detoxifying enzymes ( 28 ) and decomposition of polyunsaturated fatty acids, resulting in the formation of reactive carbonyl species, which can bind to protein ( 29 ). In Tfam KO cells, we observed neither an increase in SOD expression or activity nor an increase in carbonyl protein adducts. Taken together, several different methods to assess ROS metabolism show no signs of increased mitochondrial ROS production or oxidative stress in Tfam KO muscle cells.
Role of increased mitochondrial [Ca 2+ ]
Mitochondrial Ca 2+ uptake during repeated contractions has previously been shown not to occur in FDB fibres of wild-type mice ( 36 , 37 ). In agreement with this, we did not observe any increase in [Ca 2+ ] mit during repeated tetanic stimulation in FDB fibres of control mice. In contrast, fibres of Tfam KO mice showed a marked increase in [Ca 2+ ] mit during the repeated contractions. Myopathy symptoms get worse over time in Tfam KO mice ( 9 ). Muscle cells of 2-month-old Tfam KO mice showed an increase in [Ca 2+ ] mit during the repeated contractions, which were rapidly reversed after the cessation of stimulation. Mitochondrial respiration is stimulated by an increase in [Ca 2+ ] mit ( 30–32 ) and hence the observed increase can be seen as an appropriate adaptive response early in the disease process. Muscle cells of 4-month-old Tfam KO mice also showed increased [Ca 2+ ] mit during the period with contractions, but at this stage, [Ca 2+ ] mit was still elevated 10 min after the end of stimulation. Such prolonged increases in [Ca 2+ ] mit are generally accepted as having deleterious effects on cell function ( 16 , 33 ).
CSA that binds to the mitochondrial protein cyclophilin D has been shown to have a protective effect in collagen VI myopathies (Ullrich congenital muscular dystrophy and Bethlem myopathy), where dysfunctional mitochondria have a central role in the disease process ( 38 , 39 ). Indeed, it was recently shown that mitochondrial function was largely normalized and apoptosis was decreased in muscle biopsies of patients with collagen VI myopathies after 1 month of oral treatment with CSA ( 40 ). The action of CSA is generally ascribed to its ability to inhibit opening of mtPTP in stressed and Ca 2+ -overloaded mitochondria and, in this way, prevent mitochondrial Ca 2+ release and depolarization leading to pathological cell damage and death ( 16 , 33 ). Our results show a novel action of CSA: it significantly inhibited the contraction-induced increase in [Ca 2+ ] mit in Tfam KO fibres (Fig. 5 C). [Ca 2+ ] mit is tightly controlled in intact cells of adult mammalian skeletal muscles and when [Ca 2+ ] i increases (e.g. during contraction), the electrochemical gradient is in the direction of Ca 2+ flowing into the mitochondria ( 36 , 54 ). Mitochondria of Tfam KO cells displayed no obvious signs of Ca 2+ overload in the rested state, and the experiments with TMRE showed that they do not depolarize during fatiguing stimulation. This means that the electrochemical gradient would drive Ca 2+ into the mitochondria during fatiguing stimulation, which can explain the increase in [Ca 2+ ] mit during in Tfam KO cells. The fact that [Ca 2+ ] mit did not increase during fatigue in control cells indicates that the permeability of the mitochondrial membrane for Ca 2+ is very limited in these. Thus, Ca 2+ can pass the mitochondrial membrane in Tfam KO fibres and this occurs, at least partially, via a CSA/cyclophilin D-dependent pathway. The beneficial effect of CSA in myopathies with mitochondrial engagement may involve inhibition of mitochondrial Ca 2+ accumulation ( 38–40 , 55 ), thus limiting deleterious effects of prolonged increases in [Ca 2+ ] mit ( 16 , 33 , 34 ).
In addition to Tfam KO mice, we also used another mouse mitochondrial myopathy model, Mterf3 KO mice ( 41 ). Skeletal muscles of Mterf3 KO mice show significant changes in mitochondrial RNA and protein levels. However, Mterf3 KO muscle fibres show no changes in [Ca 2+ ] i and force production during control conditions or fatigue (Fig. 6 ). Interestingly, Mterf3 KO fibres did not display any increase in [Ca 2+ ] mit during repeated tetanic stimulation. Thus, mitochondrial Ca 2+ accumulation is not a mandatory feature in skeletal muscles with aberrant mtDNA transcription, because it is not observed in the Mterf3 KO muscle cells that display mild mitochondrial defects and no contractile deficiency. This suggests that the increase in [Ca 2+ ] mit observed in Tfam KO cells occurs as a response to the deteriorating respiratory chain function, although a more direct effect of the absence of Tfam cannot be excluded.
Tfam KO skeletal muscle cells experience a progressive deterioration of mitochondrial respiratory chain function with a threatening collapse in the cellular ATP supply. To counteract this, the cell adapts by triggering measures to enhance ATP production and reduce energy expenditure. Increased mitochondrial mass improves ATP production capacity ( 9 ), and increased [Ca 2+ ] mit can stimulate ATP production in the mitochondria. Decreased SR Ca 2+ release may reduce energy expenditure through decreased ATP consumption from both cross-bridge cycling and SR Ca 2+ pumping. However, these adaptations come with a price. Long-term increase in [Ca 2+ ] mit can effectuate cellular damage, and reduced SR Ca 2+ release causes muscle weakness and locomotor problems.
MATERIALS AND METHODS
Skeletal muscle Tfam KO (Tfam loxP/loxP , Mlc1f-Cre) mice were generated as described elsewhere ( 9 ) and compared with Tfam loxP/loxP littermate controls. Note that the Tfam disruption was induced by expressing cre-recombinase under the control of the Mlc1f promoter, which is predominantly active in fast-twitch type II muscle fibres ( 56 ), and hence fast-twitch muscles are more affected than slow-twitch muscles. Heart- and skeletal-muscle-specific Mterf3 knockout mice were generated as described previously ( 41 ). Mice were euthanized by rapid neck disarticulation and the muscles were excised. All animal experiments were approved by the Stockholm North Local Animal Ethics Committee.
Force and [Ca 2+ ] i measurements
Single fibres were mechanically dissected and suspended between an adjustable hook and an Akers AE801 force transducer in the perfusion channel of a muscle bath placed on the stage of an inverted microscope. The fibre was superfused by a Tyrode solution of the following composition (millimolar): NaCl 121, KCl 5.0, CaCl 2 1.8, MgCl 2 0.5, NaH 2 PO 4 0.4, NaHCO 3 24.0, EDTA 0.1, glucose 5.5 and 0.2% fetal calf serum. Tetanic stimulation was achieved by supramaximum current pulses with a duration of 0.5 ms delivered via platinum plate electrodes lying parallel to the fibre. Force was measured as the mean over 100 ms where it was maximal and expressed relative to the cross-sectional area. [Ca 2+ ] i was measured with the fluorescent Ca 2+ indicator indo-1 (Molecular Probes/Invitrogen). Indo-1 was mixed in a buffer (150 m m KCl, 10 m m HEPES, pH 7.1) to a final concentration of 10 m m and microinjected into fibres. The dye was excited with light at 360 ± 5 nm, and the light emitted at 405 ± 5 and 495 ± 5 nm was measured with two photomultiplier tubes. The 405/495 ratio ( R ) was translated to [Ca 2+ ] i using the following equation:57 ).
The [Ca 2+ ] i – and force–frequency relationships were obtained by producing tetani at 20–100 Hz (350 ms duration) at 1 min intervals. Fatigue was produced by a total of 50 tetani (70 Hz, 350 ms duration) given at 2 s intervals.
Measurements of mitochondrial [Ca 2+ ]
Isolated FDB fibres were incubated in 5 µ m Rhod-2 AM (Invitrogen/Molecular Probes) for 90–120 min at room temperature ( 54 ). Following loading, fibres were washed for at least 30 min. Measurements were performed using a BioRad MRC 1024 confocal unit with a krypton–argon mixed gas laser attached to a Nikon Diaphot 200 inverted microscope. Rhod-2 was excited with 568 nm light, and the emitted light was collected through a 585 nm long-pass filter. Images were analysed using ImageJ ( http://rsb.info.nih.gov/ij/ ) and data are expressed as F / F0 , where F is the fluorescence intensity at any given time point and F0 is the value at the start of the experiment. Changes in [Ca 2+ ] mit were measured during fatigue induced by 50 repeated tetani (as described above); the stimulation was stopped for ∼6 s at regular intervals as previously described ( 54 ) and the confocal images were obtained in these pauses. The recovery of [Ca 2+ ] mit after fatigue was followed by obtaining images at regular intervals for 10 min. [Ca 2+ ] mit was measured in either single dissected fibres using a 40× (N.A. 1.2) oil-immersion objective (Fig. 5 A and B) or enzymatically isolated fibres ( 58 , 59 ) using a 20× (N.A. 0.75) dry objective (Figs 5 C and 6 A). The two methods gave very similar results (cf. 4 months Tfam KO and control fibres not exposed to CSA in Fig. 5 B and C).
Measurements of mitochondrial superoxide production
Changes in mitochondrial O 2*− production were monitored using MitoSOX Red (Invitrogen/Molecular Probes). Fibres were incubated with MitoSOX Red (5 µM) for 30 min at room temperature and experiments started after 5 min of washing. MitoSox Red was excited with 488 nm light and emitted light was collected through a 585 nm long-pass filter. Single dissected FDB fibres were used to monitor changes in mitochondrial O 2*− production during fatigue induced by 50 repeated tetani (as described above) with regular pauses to obtain confocal images.
Basal mitochondrial O 2*− production was assessed using both single dissected and enzymatically isolated FDB fibres; the results showed no difference between the two modes of cell isolation, and data have therefore been pooled. A confocal image of fibres at rest was obtained, and then fibres were paced continuously at 1 Hz and images were taken after 5 and 10 min. Then, 100 µ m H 2 O 2 was applied to induce product inhibition of SOD2 ( 24 , 25 ), and confocal images were obtained every minute as long as the fibre was viable.
Control experiments were performed with the general ROS indicator 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein (CM-H 2 DCF; 5 µ m ), which was loaded into enzymatically isolated muscle fibres in the acetyl ester form (CM-H 2 DCFDA; Invitrogen/Molecular Probes). In these experiments, loading and washing time, stimulation and H 2 O 2 application were identical to those used MitoSox Red. CM- H 2 DCFDA was excited using 488 nm light, and the emitted light was collected through a 515 nm long-pass filter.
Measurements of mitochondrial membrane potential (ΔΨ m )
ΔΨ m was measured with the fluorescent indicator TMRE (Invitrogen/Molecular Probes). Isolated FDB fibres were incubated with 1 µ m TMRE for 15 min, followed by 15 min of wash-out. Confocal images of TMRE fluorescence were obtained by excitation at 568 nm and measuring emission through a bandpass filter of 605 ± 32 nm. Cells were fatigued according to the protocol described above, and images were obtained before and immediately after the 50 fatiguing tetani. Thereafter, 4 µ m FCCP (Sigma) was added to fully depolarize the mitochondria.
RNA isolation and quantitative PCR
Total RNA from control and Tfam KO mouse EDL muscle tissue samples were isolated using the RNeasy Mini Kit (Qiagen). cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit (MBI Fermentas), and quantitative PCR reactions were performed with the ABI 7700 Sequence Detection System (Applied Biosystems, USA) using TaqMan chemistry. The forward and reverse primer sequences were as follows: skeletal muscle DHPR (Cacna1s) 5′-CGAGGAGGTTGGCCAGG-3′ and 5′-GCATAGAGAAGCCAAAGTTGTCG-3′ (GenBank accession no. XM_358335); ryanodine receptor 1 (RyR1) 5′-ACAGGACACTCTTGTATGGCCAC-3′ and 5′-AGGCAACTCAGGTACATACGACTG-3′ (NM_009109); triadin 5′-GTCTGTCACCGAAGACATTGTGA-3′ and 5′-TGTGATAATCAGAGCGATGACAAG-3′ (BC034343); junction 5′-GAGAAAGAACTGAAGGCCTGTCA-3′ and 5′-AGATTTCCTCTGCCCTTTAGCA-3′ (AF289490); calsequestrin 1 (CASQ1) 5′-TGCGGCTGGCACTGCT-3′ and 5′-ATCTTCCCCCTGGACCCC-3′ (NM_009813); skeletal muscle sarco/endoplasmic reticulum Ca 2+ -ATPase 1 (SERCA1, Atp2a1) 5′-CAGGAAGGGAGCACAATGGA-3′ and 5′-CCCCGAAATAGGACAAACATTC-3′ (NM_007504) and sodium-calcium exchanger (NCX) 5′-TTGTTTTCCCATGTTGACCATATAA-3′ and 5′-GAGCCAGTACATTCAGTGGTTTCA-3′ (AF004666). The fluorogenic probes were: DHPR, 5′-Fam-CCCAACCACGGCATCACCCACT-Tamra-3′; RyR1, 5′-Fam-CCATCCTGCTCCGGCACGC-Tamra-3′; triadin, 5′-Fam-CAGCTCCCCTGCAGCCTGGC-Tamra-3′; junction, 5′-Fam-AATGGCTCACCTTCCTGACTCCAGTTCA-Tamra-3′; CASQ1, 5′-Fam-TTTGTACTGGTGCTAGGGACGCCCA-Tamra-3′; SERCA1, 5′-Fam-CCGCGCACTCCAAGTCCACAGA-Tamra-3′ and NCX, 5′-Fam-TGCAGATACAGAGGCAGAAACAGGAGGAA-Tamra-3′. The results were normalized to 18S rRNA quantified from the same samples using the forward and reverse primers 5′-TGGTTGCAAAGCTGAAACTTAAAG-3′ and 5′-AGTCAAATTAAGCCGCAGGC-3′ and the fluorogenic probe 5′-Vic-CCTGGTGGTGCCCTTCCGTCA-Tamra-3′.
Isolation of RNA and northern blot analysis of skeletal muscle from Mterf3 KO mice were carried out as described previously ( 41 ).
Protein expression in Tfam KO muscle was assessed by western blotting. Muscles were then homogenized in ice-cold lysis buffer [pH 7.6, HEPES (20 m m ), NaCl (150 m m ), EDTA (5 m m ), glycerol (20% v/v), potassium fluoride (25 m m ), sodium vanadate (1 m m ), protease inhibitor coctail (Roche), Triton X-100 (0.5% v/v)] and centrifuged. Supernatant protein content was determined using Bradford assay (Bio-Rad). Samples were diluted 1:1 with Laemmli sample buffer (Bio-Rad) containing 5% β-mercaptoethanol and boiled for 10 min. Twenty micrograms of supernatant protein was separated by NuPAGE Novex 4–12% Bis–Tris Gels (Invitrogen) and transferred onto polyvinylidene fluoride membranes (Bio-Rad). Membranes were blocked in 5% (w/v) non-fat milk TBS-T, followed by overnight incubation at 4 °C with a primary antibody. Anti-HNE membranes were reduced with 10 m m NaBH 4 in TBS prior to blocking with 7.5% non-fat milk TBS-T. Antibodies were made up in 1% (CASQ1) or 5% (all other) non-fat milk TBS-T, and dilutions of 1:1000 were used for MDA (rabbit anti-MDA, Academy Bio-Medical, catalogue no. MD20A-R1a), HNE (rabbit anti-HNE-Michael Adducts, Calbiochem, no. 393207) and SOD2 (rabbit anti-SOD2, upstate no. 06-984), 1:3000 for CASQ1 (rabbit CASQ1, Santa Cruz Biotechnology, catalogue no. sc-28274), 1:500 for DHPR (mouse anti-DHPR alpha 2 subunit, Abcam, no. ab2864-100). Membranes were washed and incubated for 1.5 h at room temperature with the appropriate secondary antibody: goat anti-mouse HRP, 1:1000 (Bio-Rad, no. 170-6516) or donkey anti-rabbit HRP and 1:5000 (Bio-Rad, no. 170-6515). Immunoreactive bands were visualized using enhanced chemiluminescence (Super Signal; Pierce, Rockford, IL, USA). Band densities were analysed with ImageJ (NIH, USA; http://rsb.info.nih.gov/j/ ).
Isolation of protein and western blot analysis of skeletal muscle from Mterf3 KO mice were performed as previously described ( 41 ).
SOD activity was measured as described previously ( 60 ). Briefly, muscles were homogenized, centrifuged and SOD activities in the supernatant were measured with a kit based on the inhibition of the reduction of tetrazolium salt induced by superoxide production (Cayman Chemicals, USA, lot no. 706002). SOD2 was determined by inhibition of the copper–zinc-dependent SOD (SOD1) with 1 m m sodium cyanide. One unit of SOD activity (U) was defined as the amount of enzyme required to inhibit 50% of tetrazolium salt reduction induced by superoxide production. SOD activities are expressed relative to the protein concentration, which was determined using the Bradford assay (Bio-Rad).
Values are given as means±SEM. Statistically significant differences ( P < 0.05) were determined using unpaired or paired t -test.
This work was supported by the Muscle Dystrophy Association, the United Mitochondrial Disease Foundation, the Swedish Research Council, the Swedish Foundation for Strategic Research (Functional Genomics and INGVAR) and Funds of the Karolinska Institutet.
Conflict of Interest statement . None declared.