Intensive activity of muscles causes a decline in performance, known as fatigue, that is thought to be caused by the effects of metabolic changes on either the contractile machinery or the activation processes. The concentration of inorganic phosphate (Pi) in the myoplasm ([Pi]myo) increases substantially during fatigue and affects both the myofibrillar proteins and the activation processes. It is known that a failure of sarcoplasmic reticulum (SR) Ca2+ release contributes to fatigue and in this review we consider how raised [Pi]myo contributes to this process. Initial evidence came from the observation that increasing [Pi]myo causes reduced SR Ca2+ release in both skinned and intact fibres. In fatigued muscles the store of releasable Ca2+ in the SR declines mirroring the decline in SR Ca2+ release. In muscle fibres with inoperative creatine kinase the rise of [Pi]myo is absent during fatigue and the failure of SR Ca2+ release is delayed. These results can all be explained if inorganic phosphate can move from the myoplasm into the SR during fatigue and cause precipitation of CaPi within the SR. The relevance of this mechanism in different types of fatigue in humans is considered.
There is a decline in the performance of muscles when they are used at near their maximum capacity. The changes in performance include reduced force production, decreased velocity of shortening, and slowed relaxation; the combination of these factors can lead to profound reductions in performance particularly for rapidly repeated movements. This decline of performance, or fatigue, is particularly obvious and well recognized in athletes who use their muscles very close to their maximum capacity. What is less appreciated is that many patients suffer serious consequences from muscle fatigue and could benefit from strategies to reduce fatigue. Patients with a variety of diseases that affect muscle directly (e.g. muscular dystrophy) or indirectly (e.g. strokes, multiple sclerosis) have reduced muscle function. If the muscles involved are used for vital functions such as breathing or walking, then the remaining functional muscles are used much nearer to their maximal capacity. Consequently essential muscles fatigue rapidly with potentially serious outcomes. Elderly humans suffer a gradual loss of muscle mass and consequently muscles may fatigue during everyday activities causing loss of mobility and independence. Heart failure patients suffer from breathlessness on exertion but also experience excessive muscle fatigue and both of these factors contribute to their reduced exercise capacity (Wilson, 1996; Lunde et al. 2001b).
In the present article we explain the reasons for thinking that increased concentration of inorganic phosphate in the myoplasm ([Pi]myo) has a central role in muscle fatigue including causing activation failure. We propose that inorganic phosphate (Pi), which increases substantially during fatigue, may enter the sarcoplasmic reticulum (SR), combine with Ca2+ and form an insoluble precipitate of calcium phosphate (CaPi), leading to reduced SR Ca2+ release and a consequent decline in muscle performance as originally suggested by Fryer et al. (1995). This mechanism provides a simple explanation for the failure of SR Ca2+ release in fatigue and may lend itself to therapeutic interventions in the future.
Mechanisms of fatigue
In muscles contracting at high work loads, stores of energy within the muscle are consumed and products of these reactions accumulate. Thus muscles become acid because glycogen is broken down anaerobically to lactic acid; Pi accumulates because phosphocreatine (PCr) is broken down to creatine (Cr) and Pi. In addition there are, of course, numerous other metabolic changes in fatigue (for review see Fitts, 1994).
In principle, the cellular mechanisms which control force include: (i) the Ca2+ concentration surrounding the myofilaments, (ii) the sensitivity of the myofilaments to Ca2+, and (iii) the force produced by the crossbridges characterized by the maximum Ca2+-activated force. Thus, at a cellular level, fatigue is generally thought to be caused by the cumulative effects of the various metabolite changes acting through these three mechanisms (for review see Westerblad et al. 1991; Allen et al. 1995). Historically by far the most popular hypothesis has been that intracellular acidosis associated with lactic acid accumulation is the major cause of fatigue acting through mechanisms (ii) and (iii). We have recently reviewed the evidence for this theory and outline why this theory is losing support, at least in mammals (Westerblad et al. 2001). As an alternative Pi has also been intensively investigated. Studies in intact muscles show that the resting [Pi]myo is 1–5 mm (Kushmerick et al. 1992) while during intense contraction it can rise to 30–40 mm (Cady et al. 1989) and the increased [Pi]myo may cause fatigue by acting through all three of the above mechanisms.
Investigations on mechanism (i) started with the classic work of Eberstein & Sandow (1963). They fatigued intact muscles with repeated tetani until force was greatly reduced and then increased the level of activation by increasing extracellular K+ or application of caffeine. Both these manoeuvres increased force substantially in the fatigued muscle suggesting that a reversible failure of activation was an important contributor to fatigue. A recent example of this approach is shown in Fig. 1 A which illustrates how caffeine can reverse much of the decline of force in a fatigued muscle. The rise in tetanic [Ca2+]myo, which activates the contractile proteins (Fig. 1 B (i)), initially increases (Fig. 1 B (ii)), but then declines during fatigue (Fig. 1 B (iii)). Agents such as caffeine, which increase the opening of the SR Ca2+ release channels (ryanodine receptors), can increase the amplitude of tetanic [Ca2+]myo (Fig. 1 B (iv)) and thus overcome much of fatigue. Thus the partial failure of SR Ca2+ release is accepted to be one of the causes of muscle fatigue (Allen et al. 1995; Williams & Klug, 1995; Favero, 1999).
The main focus of the present review is the role of Pi in the failure of SR Ca2+ release during fatigue. It is already established that increasing [Pi]myo reduces crossbridge force and Ca2+ sensitivity of the myofilaments (Cooke & Pate, 1985; Millar & Homsher, 1990) and probably contributes to the early fall in force (within 1 min) shown in Fig. 1 A. Inorganic phosphate also increases the open probability of the SR Ca2+ release channels (Fruen et al. 1994) and is a probable cause of the early increase in tetanic [Ca2+]myo illustrated in Fig. 1 B (ii). Inorganic phosphate also contributes to a slowing of the SR Ca2+ pump (Dawson et al. 1980a; Duke & Steele, 2000). These roles of increasing [Pi]myo in early fatigue are each supported by the finding that when creatine kinase is inoperative (see later section) the early fall of force, the early rise of tetanic [Ca2+]myo and the early slowing of the decline of [Ca2+]myo during relaxation are all abolished (Dahlstedt et al. 2001).
Dawson et al. (1978) initially suggested that slowing of the SR Ca2+ pump could explain the reduced force in fatigue. This could come about if the Ca2+, which is not returned to the SR, were sequestered in some organelle, e.g. mitochondria, or removed from the cell. However, the amount of Ca2+ sequestered by mitochondria during fatigue seems to be relatively small (Gonzalez-Serratos et al. 1978; Lännergren et al. 2001) and the rate of removal of Ca2+ to the extracellular space is also very slow (Balnave & Allen, 1998). Furthermore, studies in which the SR pump is slowed by pharmacological means show that the amplitude of the tetanic [Ca2+]myo becomes larger during pump inhibition presumably because the activity of the pump usually opposes the rise in tetanic [Ca2+]myo caused by SR Ca2+ release (Westerblad & Allen, 1994). Thus current evidence shows that slowing of the SR Ca2+ pump, while important to relaxation, is not the primary cause of reduced SR Ca2+ release.
Precipitation of calcium phosphate
None of the established effects of Pi described above appear capable of explaining the substantial changes in activation observed during fatigue. The first suggestion that increased [Pi]myo might have novel effects through precipitation of CaPi came from an important study by Fryer et al. (1995). They used a skinned fibre preparation with intact SR and tested the effects of increased [Pi]myo on the SR Ca2+ content, which was estimated from the force produced by caffeine. As background information it has long been known that the Ca2+ uptake of isolated SR vesicles can be greatly increased by agents such as oxalate or Pi (Hasselbach, 1964). These substances enter the SR, bind to Ca2+, and eventually precipitate with Ca2+. Consequently such substances buffer the Ca2+ within the SR and greatly increase the Ca2+ uptake capacity of the SR. Fryer et al. confirmed such effects of elevated [Pi]myo when the solution around the preparation was large. More importantly, they also showed that when the volume of solution around the fibre was restricted, elevation of [Pi]myo reduced the subsequent caffeine-induced contraction (Fig. 2 A). They proposed that Pi is capable of entering and leaving the SR leading to the events shown schematically in Fig. 2 B. It is known that the Ca2+ concentration in the SR ([Ca2+]SR) is of the order of 1 mm (Somlyo et al. 1981) with the total Ca2+ being much higher because much of the Ca2+ is bound to calsequestrin (CS), a Ca2+-binding protein with high capacity and low affinity (MacLennan & Holland, 1975). Ca2+ is thought to bind to, and dissociate from, calsequestrin relatively quickly so that it acts as a rapid buffer for Ca2+. The solubility product of CaPi is 6 mm2 (Fryer et al. 1995) so that with a [Ca2+]SR of 1 mm, if [Pi]SR exceeds 6 mm then CaPi will start to precipitate. The net effect is that [Ca2+]SR declines as an increasing fraction of the total SR Ca2+ becomes precipitated as CaPi. This mechanism can explain the reduced contraction caused by elevated [Pi]myo (Fryer et al. 1995).
A second indication that Pi has effects other than directly on the myofilaments came from a study in which Pi was directly injected into muscle cells (Westerblad & Allen, 1996). We were expecting to see reduced force and Ca2+ sensitivity due to the direct effects of Pi on the myofilaments but, to our surprise, these effects were hardly apparent and instead there was a drastic reduction in SR Ca2+ release which caused a fall in force. Since the expected effects of Pi on myofilaments were largely absent, we reasoned that most of the injected Pi had entered the SR, precipitated as CaPi, and consequently reduced SR Ca2+ release.
Role of calcium stores in fatigue
The studies cited above show that precipitation of CaPi can occur but it still has to be established that this process takes place in the more complex intracellular environment of fatigue. An important contribution to this debate arises from a landmark study by Gonzalez-Serratos et al. (1978). They fatigued single frog muscle fibres with repeated tetani and measured ionic concentrations by electron probe microanalysis. The electron microprobe could be focused down onto individual terminal cisternae of the SR and the total Ca2+ and phosphorus were measured. The important finding was that both total Ca2+ and phosphorus (which would be largely Pi) increased significantly in the fatigued muscle. The authors concluded that the Ca2+ store was not depleted during fatigue and hence the decline in tetanic [Ca2+]myo which occurs during fatigue has usually been interpreted as a failure of opening of the SR Ca2+ release channels.
Another approach to this problem is to try to measure the size of the SR Ca2+ store in intact muscle. This has frequently been done in cardiac myocytes in which caffeine causes rapid SR Ca2+ release. We have recently applied this approach to single cane toad fibres, as illustrated in Fig. 3 (Kabbara & Allen, 1999). Note that application of a drug which increases the opening probability of SR Ca2+ release channels, causes a large rise in [Ca2+]myo which is typically twice the size of a maximally activated tetanus. We call this the rapidly releasable SR Ca2+ content to indicate that Ca2+ bound to slowly releasable sites within the SR would not be detected by this approach. During fatigue it is clear from Fig. 3 that the rapidly releasable SR Ca2+ falls substantially; it also recovers after a rest period as does the tetanic [Ca2+]myo. In order to try to identify the whereabouts of the missing SR Ca2+ two types of experiments were performed. First, recovery from fatigue was performed in the absence of extracellular Ca2+; if the missing Ca2+ had been pumped out of the cell this would presumably have prevented recovery of SR Ca2+ release (Balnave & Allen, 1998). However, SR Ca2+ stores recovered completely in the absence of extracellular Ca2+ suggesting that the missing Ca2+ remained within the fibre. Second, recovery was performed in the presence of cyanide to delay the recovery of metabolic changes associated with fatigue (Fig. 3). It is known, for instance, that the elevated [Pi]myo recovers very slowly in the absence of oxidative metabolism (Dawson et al. 1980b). Cyanide prevented the recovery of the SR Ca2+ release, showing that oxidative metabolism is required for store recovery. These experiments show that the rapidly releasable SR Ca2+ declines during fatigue and are compatible with the hypothesis that the missing Ca2+ is within the SR as a precipitate of CaPi. Note that there is no conflict between these results, which show reduced rapidly releasable SR Ca2+ ([Ca2+]SR+[CaCS], see Fig. 2B), and the earlier results of Gonzalez-Serratos et al. (1978) which showed increased total SR Ca2+, which would include [Ca2+]SR+[CaPi]+[CaCS].
One weakness of the above approach is the possibility that some change in the SR Ca2+ release channels or their sensitivity to caffeine (Duke & Steele, 1998) is the cause of the reduced SR Ca2+ release in fatigue. Furthermore, we have no information about the time course of the changes in stored Ca2+. Both these weaknesses can be resolved by direct measurement of [Ca2+]SR. This has been achieved in a number of tissues (Golovina & Blaustein, 1997; Shmigol et al. 2001), and we have recently shown that it is possible in skeletal muscle (Kabbara & Allen, 2001). The essence of the method is to use a low affinity Ca2+ indicator loaded in its membrane-permeant form. Assuming there is esterase present in the SR, some indicator should be localized within the SR site and the indicator in the myoplasm will give little response to [Ca2+]myo because of the low Ca2+ sensitivity of the indicator. These studies showed that [Ca2+]SR declined throughout a period of fatiguing stimulation and recovered afterwards. These experiments do not allow unequivocal identification of the mechanism involved but they do support the previous studies in showing a reduction and recovery of [Ca2+]SR which presumably contribute to the changes in SR Ca2+ release.
Inhibition of creatine kinase prevents the rise of Pi
A prerequisite for precipitation of CaPi in the SR to occur is that [Pi]myo increases. The Pi accumulating in the myoplasm during fatigue mainly stems from PCr breakdown. Creatine kinase (CK) holds the following reaction near equilibrium:
Thus the net reaction when ATP is rapidly consumed is as follows:
Consequently when the CK reaction is inhibited, fatigue occurs without any major increase in [Pi]myo (Dahlstedt et al. 2000) and then precipitation of CaPi within the SR should be avoided. On these grounds it might be hypothesized that tetanic [Ca2+]myo will remain unchanged during fatiguing stimulation in muscle fibres with inoperative CK. This hypothesis has recently been tested both with genetically engineered mice, which have no detectable CK in their skeletal muscles (CK−/−) (Dahlstedt et al. 2000), and with pharmacological inhibition of the CK reaction using 2,4-dinitro-1-fluorobenzene (DNFB) (Dahlstedt & Westerblad, 2001). The basic results of these two studies were similar: there was a markedly slower decline of tetanic [Ca2+]myo during fatigue after CK inhibition (Fig. 4). Both these models have limitations: CK−/− muscle displays adaptations with, for instance, an increased mitochondrial content (Steeghs et al. 1997) and DNFB has non-specific inhibitory effects on muscle function (for discussion see Dahlstedt & Westerblad, 2001). Since the limitations of the two models are quite disparate and they both cause a slower decline of tetanic [Ca2+]myo during fatigue, it appears safe to conclude that the preserved tetanic [Ca2+]myo in fatigue is due to inhibition of CK. Thus, the results from muscle fibres with inoperative CK strengthen the proposal that CaPi precipitation is an important mechanism underlying the failure of SR Ca2+ release in fatigue.
Pi appears to enter the SR via ATP-sensitive anion channels
The results of an early study looking at Pi transport over the SR membrane indicate that Pi enters the SR via an ATP-dependent Pi transporter (Stefanova et al. 1991; Fryer et al. 1997). However, the results of a subsequent study suggest that Pi entry into the SR is a passive process, which was inhibited in the presence of ATP (Posterino & Fryer, 1998). It has now become clear that the SR membrane contains small conductance chloride channels, which may conduct Pi (Ahern & Laver, 1998; Laver et al. 2001). Interestingly the open probability of these channels increases at low [ATP]myo confirming the above study on intact SR (Posterino & Fryer, 1998). This dependence on [ATP]myo can explain one apparent weakness of the hypothesis that raised [Pi]myo causes CaPi precipitation in the SR: [Pi]myo increases relatively early during fatiguing stimulation while the decline of tetanic [Ca2+]myo generally occurs quite late. Moreover, in mouse fast-twitch fibres the decline of tetanic [Ca2+]myo temporally correlates with an increase in [Mg2+]myo, which presumably stems from a net breakdown of ATP (Westerblad & Allen, 1992), and it is not obvious why CaPi precipitation in the SR should show a temporal correlation with ATP breakdown. The ATP dependence of the presumed SR Pi channels can explain both why Pi enters the SR with a delay and why there is a temporal correlation between declining [ATP]myo and declining tetanic [Ca2+]myo.
Relevance to exercise in humans
The evidence is now reasonably strong that precipitation of CaPi in the SR can contribute to the decline of SR Ca2+ release in muscles and is therefore one important cause of fatigue. In what kinds of exercise is this mechanism likely to be important? Clearly a number of processes are involved and these place some limit on the time course over which such a mechanism could develop. The first process is breakdown of PCr which is the main source of Pi. This can occur quickly (∼1 min) if the muscle is maximally and continuously activated (Kushmerick & Meyer, 1985; Cady et al. 1989). The second process is that the Pi has to move from the myoplasm to the SR presumably by the Pi permeable channel in the SR membrane noted above. To calculate the time course of entry of Pi into the SR requires knowledge of the number of channels and their Pi permeability. An upper limit can be obtained by assuming the channels exhibit their maximum permeability and that [Pi]myo= 20 mm and these assumptions suggest an initial flux of 0.7 mm s−1 (Laver et al. 2001). Assuming further that [Pi]SR= 2 mm, it would only take 6 s for the [Pi]SR to exceed 6 mm. Fryer et al. (1997) measured Pi flux as 0.05 mm s−1 which would give a much longer equilibration time though this time would be shorter if [Pi]SR is higher than 2 mm. Thirdly, in many systems it is common for precipitation to fail to occur immediately the solubility product threshold is exceeded; a nucleus for precipitation is required and the solubility product may be exceeded transiently before precipitation occurs. Since these processes are in series it would seem unlikely that precipitation of CaPi would occur in fatiguing activities shorter than at 1–2 min. During recovery all these processes have to be reversed and this may be one factor why Pi recovery often precedes force recovery by some minutes (Cady et al. 1989).
In fatigue of a very long duration (> 1 h), muscle activation is generally quite low and the final decline of performance correlates well with glycogen depletion (Bergström et al. 1967). In this kind of exercise [Pi]myo is only moderately raised (Vøllestad et al. 1988) and it is unclear whether CaPi precipitation would occur. Furthermore, in muscles which have been glycogen depleted by repeated bursts of activity in the absence of glucose, fatigue occurs very rapidly but is still caused by a failure of SR Ca2+ release (Chin & Allen, 1997; Kabbara et al. 2000). Interestingly, such muscles do not show reduction in SR Ca2+ stores at the point of fatigue (Kabbara et al. 2000) possibly because the muscles fatigue so quickly that the series of events outlined above cannot occur. These results suggest that failure of SR Ca2+ release in glycogen-depleted muscles involves a different mechanism. For instance, studies on skinned fibres suggest that glycogen may, in addition to its energy storage role, have a structural role which is essential for normal excitation-contraction coupling (Stephenson et al. 1999).
Patients with heart failure experience rapid and disabling fatigue in their skeletal muscles and the mechanism of this fatigue has been the subject of considerable debate (Wilson, 1996; Lunde et al. 2001b). Obviously cardiac output, particularly during exercise, is reduced in heart failure patients so that blood supply to muscles will tend to be smaller. Consequently for a given level of exercise the metabolic changes are usually larger; for instance, in heart failure subjects the resting [Pi]myo in muscles is higher in some studies (Massie et al. 1987; van der Ent et al. 1998) and the increase in [Pi]myo on exercise is greater in most studies (Wilson et al. 1985; Arnolda et al. 1991). However, the changes are not large and it seems unlikely that the direct effects of Pi on the myofilaments are sufficient to explain the increased degree of fatigue. Therefore we suggest that raised [Pi]myo may contribute to the more rapid fatigue by causing precipitation of CaPi in the SR. In addition there are many changes in the protein composition of skeletal muscles in heart failure (for review see Lunde et al. 2001b) and there have been attempts to discover whether isolated skeletal muscles taken from animals with heart failure fatigue more rapidly. Perreault et al. (1993) used bundles of fast muscle and showed that those from an animal with heart failure fatigued more rapidly and had a larger decline in tetanic [Ca2+]myo. This result suggested that an activation defect develops in skeletal muscle in animals with heart failure and that reduced blood flow is not the sole or major cause. However, more recent studies on single fibres from a similar model of heart failure did not show any major increase in fatiguability and only modest changes in [Ca2+]myo (Lunde et al. 2001a). If changes in activation are modest in isolated single fibres this may be because the oxygen supply is more than adequate in this preparation and consequently aerobic metabolism can provide for much of the metabolic requirements. Conversely when muscles are in situ in heart failure subjects, the oxygen supply will be more limited so that anaerobic metabolic changes, such as a rise in [Pi]myo, will occur earlier and may lead to activation changes by means of the CaPi precipitation mechanism.
Failure of SR Ca2+ release has been shown to be a major contributor to muscle fatigue. Increasing evidence supports the hypothesis that precipitation of CaPi in the SR contributes to the failure of SR Ca2+ release. We suggest that this mechanism may be important in high intensity activities which lead to fatigue in > 1–2 min but other mechanisms are probably more important in lower intensity activities which cause fatigue in > 1 h. CaPi precipitation in the SR may contribute to the disabling fatigue common in many muscle disabilities and is especially likely to contribute in heart failure where [Pi]myo levels are higher.
Work from the authors' laboratory was supported by the National Health and Medical Research Council of Australia (to D.G.A.) and the Swedish Medical Research Council (Project Number 10842), the Swedish National Centre for Sports Research and funds at the Karolinska Institutet (to H.W.).