Increased Metabolic Muscle Fatigue Is Caused by Some but Not All Mitochondrial Mutations FREE

EXCESSIVE MUSCLE fatigue is a frequent report of patients with neuromuscular disorders, but also of patients with a disorder of the central nervous system.^{1 - 2} Yet, the causal mechanisms of both pathological and physiological muscle fatigue are not fully understood. Metabolic and nonmetabolic changes of muscle fibers, impairment of neuromuscular transmission, and changes within the central and peripheral nervous system may all contribute to muscle fatigue. However, it is recognized that specific biophysical changes that occur within the muscle fibers play a major role.^{2 - 6} Such changes may have a primary link to inadequate energy supply within the muscle fibers, as was suggested regarding patients with mitochondrial myopathies.⁷ In other patients, nonspecific changes secondary to a nonmuscular disorder may cause muscle fatigue, such as in spastic paraparesis or amyotrophic lateral sclerosis.^{4 ,8}

In patients with a mitochondrial myopathy, an intramuscular origin of excessive muscle fatigue has been identified⁷ ; however, the relative degree of muscle fatigue reported weakly correlated with the concentration of serum lactate.⁷ Because no patients with other neuromuscular diseases (ONMD) were studied, it was not clear to what extent the muscle fatigue in the patients studied was caused by impaired mitochondrial function or by other changes secondary to the weakness of the patients. The purpose of this study was to address this question by comparing twitch-fatigue properties of healthy subjects with those recorded in patients with either a mitochondrial disorder or ONMD. To minimize contributing influences of central fatigue, electrical stimulation of peripheral nerves was used to involuntarily activate the subjects' ankle dorsiflexor muscles.

The study was approved by the local ethics committee. All subjects, patients, and controls, were informed of the research character of the study and gave their consent before being tested.

In 21 patients, the diagnosis of mitochondrial encephalomyopathy (MEM) was based on the clinical picture and was confirmed if at least 5% ragged-red fibers were found in muscle specimens (14 patients), if there were abnormalities in the mitochondrial genome (18 patients), or both (Table 1). Eleven male and 10 female patients were studied. Their ages ranged from 14 to 62 years (mean, 43 years). Clinically, 13 patients had chronic progressive external ophthalmoplegia (CPEO) — 10 had deletions of the mitochondrial genome, and 3 had ragged-red fibers, but there were no abnormalities of the mitochondrial genome. The remaining 8 patients had mitochondrial point mutations. Their phenotypes were MEM, lactic acidosis, and strokelike episodes (MELAS) (n = 2); Leber hereditary optic neuropathy (n = 2); deafness and diabetes (n = 1); deafness alone (n = 1); painful muscle stiffness (n = 1)⁹ ; and myopathy (n = 1). Clinical and genetic data of the latter patient were already reported by Beyenburg et al¹⁰ and by Jaksch et al.¹¹

This ONMD group consisted of 14 male and 13 female patients with various neuromuscular disorders. The diagnoses were amyotrophic lateral sclerosis (ALS) (n = 6), facioscapulohumeral muscular dystrophy (n = 4), motor neuropathy with persistent conduction block (n = 1), polymyositis (n = 2), myotonic dystrophy (n = 3), oculopharyngeal muscular dystrophy (n = 1), myopathy with tubular aggregates (n = 1), proximal myotonic myopathy (PROMM) (n = 7), polyneuropathy (n = 1), and spinal muscular atrophy (n = 1). Their ages ranged from 21 to 71 years (mean, 48 years).

The control group consisted of 13 male and 12 female healthy volunteers. Their ages ranged from 19 to 62 years (mean, 34 years).

Subjects were at rest for at least 20 minutes before any force measurements were made. For the force measurements, a rigid apparatus was used to secure the subject's leg in a predetermined position that confined movement to ankle dorsiflexion (Figure 1).¹² An incorporated strain gauge was used to measure isometric torque. The innervating common peroneal nerve was electrically stimulated via surface electrodes (stimulator: Model DS7; Digitimer, Welwyn Garden City, England). The stimulus current delivered was increased until twitch torque was maximal (supramaximal stimulation). Next, the ankle angle was adjusted until the optimal dorsiflexor muscle length was identified (length-tension relationship). The surface electromyogram of the ankle dorsiflexors was recorded with a filter setting of 2 Hz to 2 kHz (amplifier: Type 780812; Toennies Co, Höchberg, Germany). Torque and surface electromyogram were digitized at a sampling rate of 5 kHz (DAQPad-1200; National Instruments, Austin, Tex). A LabVIEW program (Version 4.1; National Instruments) was used for data acquisition, storage on a personal computer, and data analysis.

Six trains of supramaximal electrical stimuli were given to each subject. The rates of stimulation were 0.2, 0.5, 1, 2, 3, and 5 per second. The duration of each train was 60 seconds, followed by an interval of 120 seconds without stimulation (Figure 2 and Figure 3).

For each stimulus, the amplitude of the compound muscle action potential (CMAP) and the amplitude of the force output generated by the muscle twitch were measured automatically (Figure 2, second row). For each train of stimuli, the amplitude of the first twitch (TA₀) and the mean amplitude of all twitches between 50 and 60 seconds after the first stimulus (TA₅₀) were measured. Rate-dependent twitch fatigue was defined as (TA₀ −TA₅₀)/TA₀ (Figure 4). Similarly, we measured the amplitude of the first evoked CMAP (A₀) and the mean amplitude of the CMAPs between 50 and 60 seconds after the first stimulus (A₅₀). The CMAP amplitude decrement was defined as (A₀−A₅₀)/A₀.

To assess potential increases of serum lactate levels under mild exercise, patients with mitochondrial disorders underwent a standardized test on a bicycle ergometer; cycle conditions were 30 W for 15 minutes.¹³ Patients were at rest for at least 30 minutes before exercise. A blood sample from the cubital vein was then taken, and exercise was started. Additional blood samples were taken after 5 minutes, 10 minutes, and 15 minutes during exercise, and again 15 minutes after the end of the exercise. Blood samples were immediately exposed to iced perchloric acid for the measurement of the lactate concentration, which was done within 2 hours. Relative changes in lactate levels were determined by dividing the lactate concentrations after 15 minutes of exercise, by the baseline values. Regarding lactate, area under the curve during exercise was determined by subtracting baseline lactate values from values after 5, 10, and 15 minutes of exercise, and computing the area under the resulting curve of lactate increase vs time by triangulation.

For identification of mitochondrial deletions, muscle DNA was digested with BamHI and PvuII prior to Southern blotting and hybridization to a mitochondrial DNA (mtDNA) probe of 16252 nucleotides (Expand Long Template PCR [polymerase chain reaction] System [Boehringer Mannheim, Mannheim, Germany], forward primer 15149-15174, and reverse primer 14831-14811), labeled with the ECL kit (Amersham Life Science, Buckinghamshire, England) and exposed to an x-ray film. In 4 patients, no muscle DNA was available. In these patients, blood DNA was used instead (Table 1).

According to Yamamoto,¹⁴ Zeviani et al,¹⁵ and Moorman et al,¹⁶ the mitochondrial point mutations A3243G, A8344G, and G11778A were identified by restriction analysis after PCR amplification of DNA extracted from muscle tissue. The mitochondrial point mutation G7497A was identified by sequencing of a PCR product of the region 7374-7626 of the mtDNA, and by cleavage with the restriction enzyme NcoI (loss of a restriction site). The fragments were fractionated on 8% polyacrylamide gel, stained with ethidium bromide, visualized on an ultraviolet transilluminator, and photographed.

The relative proportions of mutated to nonmutated mtDNA (heteroplasmy) on Southern blot analysis and restriction analysis were determined by ImageQuant software (Molecular Dynamics Inc, Sunnyvale, Calif).

Statistical analysis was done with SigmaStat (version 1.0 for Windows, SPSS Science, Chicago, Ill). Before parametric statistical tests were used, data were checked for normality of the distribution. P<.05 was considered significant.

In all individuals, both healthy subjects and patients, the amplitude of the first twitch of a stimulus train (TA₀) did not change during each test (Figure 5). Similarly, the amplitudes of the CMAPs were also maintained (ie, the amplitude decrement was less than 3% at stimulus rates of 2 per second or less, less than 5% at a stimulus rate of 3 per second, and less than 10% at 5 per second). Tetanic potentiation (Figure 2) was found only at 5 stimuli per second and at various degrees in both patients and controls.

Overall, twitch force, measured as TA₀, was greater in men than in women (t test: P = .03), and was not age dependent (Pearson product moment correlation: r = 0.23, P = .33). In controls, the decrement in twitch amplitude (ie, twitch fatigue) increased with the frequency of the stimulus train (Figure 6A). Relative twitch fatigue was greater in women than in men (t test: P = .02 at 2 stimuli per second) but was not age dependent (Pearson product moment correlation: r = 0.11, P = .61).

As in controls, in patients with MEM, twitch fatigue increased with the stimulus frequency (Figure 3, Figure 6B). Abnormally high twitch fatigue values relative to controls were found most frequently at a stimulus rate of 2 seconds (Table 2). Twitch fatigue was normal in all patients carrying the A3243G or the G11778A mutation, and in 2 patients with deletions (Table 1, patients 11 and 13). In contrast, relative twitch patterns were abnormal in 11 patients with CPEO and in the patient with a mitochondrial myopathy caused by the mitochondrial G7497A mutation (Table 1, patient 14).

As in controls and in patients with MEM, in patients with ONMD fatigue increased with stimulus rate (Figure 6C). Fatigue in this group of patients was significantly higher than in controls at all stimulus rates above 0.5 per second, but it was not different from that in the group of patients with mitochondrial disorders. Abnormally high twitch fatigue values were observed at all stimulus rates above 0.5 per second. There was no apparent correlation between diagnosis or type of disorder (ie, neurogenic vs myopathic) and presence of abnormal fatigue.

When pooled data of all subjects studied were analyzed for correlations, muscular twitch fatigue at a stimulus rate of 2 per second was found to be associated with muscle force and the presence of an MEM. No other variables with an influence on twitch fatigue were identified (Table 3). At stimulus rates of 0.5, 1, and 3 per second, fatigue significantly depended on muscle force only (P = .004, P<.001, and P = .005, respectively). The correlation with presence of an MEM was positive but not significant at these rates.

The dependence of twitch fatigue on muscle force was nonlinear. At the stimulus rate of 2 seconds, there was a strong nonlinear correlation between fatigue and TA₀ (Figure 7A). Fatigue values were corrected for the influence of TA₀ by subtraction of the results of the nonlinear regression equation; in other words, the residuals of regression were determined. These values were called "specific fatigue." After this elimination of the influence of the muscle force, patients with an MEM had significantly higher specific fatigue than patients with ONMD (t test: P<.001) and controls (t test: P<.001; Figure 7B). In patients with an MEM, those with CPEO had significantly higher specific fatigue than the rest (t test: P = .004) (Table 4) and than patients with ONMD (t test: P = .001). The specific fatigue of the patient with a G7497A mutation was similar to that of the patients with CPEO. The specific fatigue of the patients with an A3243G or a G11778A mutation was not different from the specific fatigue of controls and of patients with ONMD (t test: P>.80). Within the control group, the specific fatigue was not different between men and women (t test: P = .86).

The influence of gender on fatigue that was seen in the control group, was not found when statistical analysis was performed on the pooled data of all subjects studied (Table 3).

In all patients with MEM, twitch fatigue correlated with the absolute concentration of serum lactate after standardized bicycle exercise (r = 0.59, P = .008), with the area under the lactate curve (r = 0.55, P = .02), and with the relative increase of serum lactate levels (r = 0.59, P = .008). There was no significant correlation between twitch fatigue and serum lactate at rest. Specific fatigue correlated with the increases of serum lactate levels on exercise (r = 0.52, P = .02). All of these correlations were closer when calculated only for the patients with CPEO (r = .73, P = .01; r = 0.69, P = .02; r = .66, P = .03; and r = 0.74, P = .009, respectively), while none of these correlations was found in the patients without CPEO. No correlations were found between any of the enzyme activities and muscle force or fatigue.

No significant correlations were found between heteroplasmy and serum lactate levels associated with the bicycle tests, twitch force, or twitch fatigue (r range, 0.05-0.35; P = .09 to P = .93).

The measurements described in this article (Figure 5) confirm the high reproducibility and straightforwardness of the muscle-force assessment method used in this study.¹² Although the train frequencies used in this study caused a relative discomfort, these studies were well tolerated by all individuals.

The constancy of the amplitude of the CMAPs of both patients and controls shows that the twitch fatigue measured in this study was of intramuscular origin and not caused at the neuromuscular junction or within the muscle membrane.

To date, 2 types of intramuscular fatigue have been noted and have been labeled as metabolic fatigue and nonmetabolic fatigue.^{2 - 6} Metabolic fatigue is observed after short duration of exercise (1-2 minutes) and shows a rapid recovery. In contrast, nonmetabolic fatigue is observed after longer duration of the exercise and is characterized by a slower recovery.⁶ The stimulation protocol and the complete recovery of TA₀ within 120 seconds after each train of stimuli both show that fatigue measured in this study was primarily or solely of the metabolic type.

Dengler et al⁷ compared metabolic muscle fatigue in a group of 20 patients with CPEO vs a control group and found a significant difference. The results presented here are confirming, but in that study, only 65% of the individual patients had an abnormal test result, while all but one of the patients with CPEO of the present study (92%) had abnormal results. An explanation may be that the fatigue test used in that study required voluntary contraction, and, consequently, the resulting twitch force values were more scattered in both patients and controls than the values obtained here.

Dengler et al⁷ also found a correlation between muscle fatigue and the serum concentration of lactate at rest and after exercise. The present study confirms this finding as well. This suggests that the metabolic changes due to mitochondrial dysfunction constitute increased metabolic muscle fatigue. However, because no patients with ONMD had been studied by Dengler et al,⁷ it was not known whether nonspecific factors, such as altered contractile apparatus, played a probably more important causal role.

In this study, abnormal fatigue was found not only in patients with MEM, but also in patients with ONMD. This confirms what had been suspected⁷ (ie, that muscle fatigue of the metabolic type to some extent is not specific for CPEO or MEM).

The strong nonlinear dependency of muscle force and fatigue found in this study stresses that fatigue of the metabolic type should be considered a nonspecific phenomenon. The correlation between muscle force and fatigue had not been previously identified when 20 patients with mitochondrial myopathies were studied,⁷ and this is most likely because fewer subjects were studied, and because the results were more scattered than in the present study, for the previously mentioned reasons. This confirms the usefulness of the method for quantitative assessment of involuntary muscle force used in this study.

Because there was no fatigue at train frequencies of 0.2 stimuli per second, no correlation between fatigue and twitch force could be found. While at stimulus rates between 0.5 per second and 3 per second a correlation was found between fatigue and twitch force, there was no such correlation at a stimulus rate of 5 per second. This is most likely because at 5 stimuli per second, tetanic potentiation occurred. Thus, measurement of twitch force at train frequencies of 5 stimuli per second is unsuitable for the assessment of muscle fatigue. It should be noted that such high train frequencies are also relatively uncomfortable.

Because high fatigue nonspecifically accompanied low muscle force, the fatigue data that were corrected for the influence of low muscle force were called specific fatigue (Figure 7B). The analysis of the specific fatigue data clearly showed that muscular twitch fatigue was higher in patients with MEMs than in patients with ONMD, even when the same degree of weakness was detected. This excessive fatigue, however, was only found in patients with a CPEO phenotype (deletions) or pure myopathy (G7497A mutation), and not in those with Leber hereditary optic neuropathy (G11778A mutation) or patients with the A3243G mutation. This suggests that excessive metabolic fatigue is caused by specific metabolic changes that are caused by some but not all mitochondrial mutations. The close correlation between specific fatigue and the increase of serum lactate in the patients with CPEO indicates that muscle fatigue in these patients is associated with the degree of the mitochondrial dysfunction.

Neither in all patients nor in the subgroup of patients with CPEO were there correlations between heteroplasmy and serum lactate levels or muscular force measures. This suggests that the degree of muscle involvement is more influenced by the type of mutation than by heteroplasmy. In patients with CPEO, the muscle involvement was not influenced by heteroplasmy at all. Because no data for patients with an A3243G and high heteroplasmy, or for patients with a G7497A mutation and low heteroplasmy were available, the influence of heteroplasmy on the degree of muscle involvement in these conditions remains unknown.

Interestingly, exercise intolerance is a frequent complaint of patients with the typical phenotype of the A3243G mutation (ie, MELAS).¹⁷ In contrast to patients with CPEO or mitochondrial myopathy, patients with MELAS more frequently have signs or symptoms of an encephalopathy. The fatigue data of the present study support the hypothesis that exercise intolerance in MELAS patients is caused by central abnormalities and not by impaired muscle metabolism.

Because it is capable of objectively measuring both muscle force and fatigue, the test reported here is of potential value in assessing treatments for mitochondrial disorders and ONMD. Patient discomfort will be less than in the present study, as stimulus rates of 3 or more per second are not necessary for that purpose.

In conclusion, it was established that, generally, muscle fatigue of the metabolic type accompanies muscular weakness. Specifically, some but not all mitochondrial mutations cause excessive metabolic muscle fatigue.

Corresponding author: Wilhelm J. Schulte-Mattler, MD, Klinik und Poliklinik für Neurologie, Universität Regensburg, Im Bezirksklinikum, Universitätsstraße 84, 93053 Regensburg, Germany (e-mail: wilhelm.schulte-mattler@klinik.uni-regensburg.de).

Accepted for publication September 9, 2002.

Study concept and design (Drs Schulte-Mattler, Müller, Iaizzo, and Zierz); acquisition of data (Drs Schulte-Mattler, Müller, and Deschaur); analysis and interpretation of data (Drs Schulte-Mattler, Gellerich, Iaizzo, and Zierz); drafting of the manuscript (Dr Schulte-Mattler); critical revision of the manuscript for important intellectual content (Drs Schulte-Mattler, Müller, Deschauer, Gellerich, Iaizzo, and Zierz); statistical expertise (Drs Schulte-Mattler and Iazzo); Administrative, techical, and material support (Drs Schulte-Mattler, Iaizzo, and Zierz); study supervision (Drs Schulte-Mattler, Gellerich, and Zierz).