Mechanisms of stretch-induced muscle damage in normal and dystrophic muscle: role of ionic changes

Abstract

Muscle damage, characterized by prolonged weakness and delayed onset of stiffness and soreness, is common following contractions in which the muscles are stretched. Stretch-induced damage of this sort is more pronounced in the muscular dystrophies and the profound muscle damage observed in these conditions may involve similar pathways. It has been known for many years that damaged muscles accumulate calcium and that elevating calcium in normal muscles simulates many aspects of muscle damage. The changes in intracellular calcium, sodium and pH following stretched contractions are reviewed and the various pathways which have been proposed to allow ion entry are discussed. One possibility is that TRPC1 (transient receptor potential, canonical), a protein which seems to form both a stretch-activated channel and a store-operated channel, is the main source of Ca²⁺ entry. The mechanisms by which the changes in intracellular ions contribute to reduced force production, to increased protein breakdown and to increased membrane permeability are considered. A hypothetical scheme for muscle damage which incorporates these ideas is presented.

Any period of intense or prolonged muscle activity can cause a decline in performance which can be measured as muscle weakness. By definition, if the weakness is largely reversible in minutes or hours, it is described as fatigue. If the weakness is slowly or poorly reversible and associated with structural changes within the muscle, it is described as muscle damage. Muscle fatigue arises principally within the muscle and is thought to be caused by accumulation or depletion, either intracellular or extracellular, of various metabolites and ions (for review see Westerblad et al. 1991; Fitts, 1994; Allen et al. 1995). Muscle damage is particularly pronounced in muscles which are stretched during contraction (eccentric or stretched contractions), for instance the quadriceps group during downhill running. Characteristically, stretch-induced muscle damage causes an immediate weakness which can take a number of days to recover. In addition the affected muscles develop soreness, swelling and stiffness which typically become apparent 1–2 days after the initiating contractions. Sore and stiff muscles the day after unaccustomed and excessive exercise will be familiar to physically active individuals and this condition has been named delayed-onset muscle soreness (DOMS). Simultaneously there is often substantial release of soluble muscle proteins, such as creatine kinase, and inflammatory cells are attracted to the damaged muscle. The weakness and histological changes of muscle damage are normally gradually repaired over several weeks and can involve the regeneration of damaged skeletal muscle fibres from satellite cells (Hawke & Garry, 2001). While small numbers of stretched contractions can cause muscle damage, it is probable that much larger numbers of isometric or shortening contractions can also cause muscle damage (Jones et al. 1983; Gissel & Clausen, 2001).

The muscle damage which follows stretched contractions has important roles in exercise training (Colliander & Tesch, 1990) and sports injuries (Proske et al. 2004) and seems to stimulate a particular type of muscle redevelopment with new sarcomeres in series (Lynn & Morgan, 1994; Yu et al. 2003). Stretch-induced muscle damage is also more severe in older animals (Brooks & Faulkner, 1996) and may have a role in the decline of muscle function seen in the elderly. There is also considerable evidence that repeated and poorly reversed muscle damage plays a part in the profound muscle degeneration which characterizes muscular dystrophy. For instance, stretch-induced muscle damage is considerably more severe in the mdx mouse, which lacks dystrophin and is a model of Duchenne muscular dystrophy (Head et al. 1992; Petrof et al. 1993). Furthermore, transfection of dystrophin into mdx muscle causes a reduction in the magnitude of stretch-induced damage, showing that the absence of dystrophin exacerbates muscle damage (Deconinck et al. 1996; DelloRusso et al. 2002). Muscular dystrophy also involves ionic disturbances which appear to have a central role in the development of the muscle pathology (Turner et al. 1988; Gillis, 1999; Yeung et al. 2003b, 2005). For all these reasons interest in the ionic changes caused by stretch-induced damage is increasing rapidly at present.

The earliest structural change following stretch-induced muscle damage is sarcomere inhomogeneities characterized by a patchy distribution of over- and understretched sarcomeres and Z-line irregularities (Fridén et al. 1981). These morphological changes can be observed within the first stretched tetanus (Brown & Hill, 1991; Talbot & Morgan, 1996) and are thought to result from the instability of sarcomeres on the descending limb of the tension–length relation (Morgan, 1990). A recent review of stretch-induced muscle damage focused mainly on the mechanical consequences of stretched contractions (Proske & Morgan, 2001). In the present review we are concerned with the pathways which lead to muscle damage and, particularly, the role of ionic changes in triggering damage.

Ionic changes following stretched contractions

Calcium

There has been a long interest in the role of calcium in muscle damage. Early studies by Duncan (1978) showed that treatment of muscles with calcium ionophores, which increase intracellular calcium ([Ca²⁺]_i), cause muscle damage with initial contracture, swelling of mitochondria, loss of soluble muscle proteins and, eventually, degeneration of myofibrils. Duncan was particularly interested in the idea that Ca²⁺-activated damage might have a central role in the muscle damage observed in muscular dystrophy. Three outstanding issues dominate this field up to the present day. (i) What is the route of Ca²⁺ entry? (ii) Is the rise of Ca²⁺ simply a function of passive Ca²⁺ entry along its electrochemical gradient following membrane damage by some other route or is it a primary contributor to the damage pathway? (iii) What are the pathways by which elevated [Ca²⁺]_i causes muscle damage?Table 1 shows tabulated information regarding the main pathways for Ca²⁺ entry which have been proposed.

Table 1.

Properties of pathways allowing ion entry during damage in skeletal muscle

	Membrane tears	Ca2+ leak channel	Stretch-activated channel(SAC)	Gth factor-regulated channel (GRC)	Store-operated channel (SOC)	Prolonged exercise- induced pathway
Channel permeability (Ca2+)	—	10 pS (96 mm Ca2+ in pipette) [3]	13 pS (110 mm Ca2+ in pipette) [4]	8 pS (110 mm Ca2+ in pipette) [15]	8 pS (110 mm Ca2+ in pipette) [21], 11 pS [20]	—
Voltage sensitivity	—	No [3]	Activity increased by depolarization [4]	No [15]	No [21]	—
Permeable ions	All ions	Ba2+, Ca2+[3], Mn2+[8],not Na+[3,18]	Li+, Na+, Rb+, Cs+,Ca2+, Ba2+, K+[4]	Na+, K+, Ca2+, Ba2+[15]	Na+, Cs+, K+, Ca2+, Ba2+[20]; Mn2+[19, 21]	Na+, Ca2+[6]
Chemical activators	—	Bay K 8644, nifedipine [3]	Amphipaths, lipids [7]	—	—	Veratridine [6]
Chemical inhibitors	—	AN 1043 [1, 8]	Gd3+[23] Streptomycin [22] GsMTx [17]	Gd3+[15] Ruthenium red [9] SKF96365 [9]	La3+[21]; Gd3+, La3+, Ni2+[19]	Tetrodotoxin inhibits the veratridine-induced component [6] Nifedipine, minor effect [6]. Gd3+, Ni2+, no effect [6].
Permeable to large molecules	Yes [13]	No	No	No	No	Yes [6, 14]
Stretch-activated	Appear after stretched contractions [13]	Channel activity sometimes enhanced by pipette suction [3]	Increased or decreased channel activity by pipette suction [4, 5].	Increased activity by pipette suction [15]	Increased channel activity by pipette suction [20]	—
Store-depletion activation	—	Yes (6)	—	—	Yes (3)	—
Increased activity in mdx	Yes [16]	Yes [3]	Yes [5]	?; increased expression in δ-sarcoglycan-deficient hamster muscle [15]	Yes [20, 21]	—
Molecular identity	—	—	TRPC1 [11]	44% homology to TRPV1 [10]	TRPC 1 or 4 [21]	—
Physiological trigger	Mechanical damage prominent after stretched contractions [13, 16]	Required contractions and membrane damage to trigger in mdx muscle. Insertion in membrane involves raised [Ca2+]i and by proteolysis [2]	Activated by stretched contractions [24]. Activity can persist for 24 h [12]	Cyclically stretching leads to increased Ca2+ uptake and creatine kinase (CK) efflux blocked by Gd3+, SKF 96365, Ruthenium red [9]. Induced by IGF-1 which causes translocation to surface membrane [9, 10]; translocation blocked by Gd3+[9]	Functional role not explored.	Induced by repeated contractions. May be partly ischaemia induced [6, 14].

Tabulated list of properties of pathways allowing influx of ions into damaged muscles. References: 1, Alderton & Steinhardt (2000a); 2, Alderton & Steinhardt (2000b); 3, Fong et al. (1990); 4, Franco & Lansman (1990); 5, Franco-Obregon & Lansman (2002); 6, Gissel & Clausen (2001); 7, Hamill & McBride (1996); 8, Hopf et al. (1996); 9, Iwata et al. (2003); 10, Kanzaki et al. (1999); 11, Maroto et al. (2005); 12, McBride et al. (2000); 13, McNeil & Khakee (1992); 14, Mikkelsen et al. (2004); 15, Nakamura et al. (2001); 16, Petrof et al. (1993); 17, Suchyna et al. (2000); 18, Turner et al. (1991); 19, Tutdibi et al. (1999); 20, Vandebrouck et al. (2001); 21, Vandebrouck et al. (2002); 22, Winegar et al. (1996); 23, Yang & Sachs (1989); 24, Yeung et al. (2003a).

The first study of intracellular calcium in muscle as a consequence of stretched contractions used a 2 h period of downhill walking of rats (Duan et al. 1990). Muscles were isolated both immediately and 2 days after the exercise, mitochondria extracted and Ca²⁺ determined. Mitochondrial Ca²⁺ was increased about 3-fold immediately after exercise and about 6-fold 2 days later. Thus these experiments showed that exercise which included stretched contractions caused a substantial increase in muscle Ca²⁺ which accumulated in the mitochondria.

Intact single fibres can be stretched during contractions and resulting changes in force and [Ca²⁺]_i determined (Balnave & Allen, 1996). A relatively severe protocol of stretched contractions reduced tetanic [Ca²⁺]_i to about 50% of the pre-stretch control; in the presence of 10 mm caffeine this tetanic [Ca²⁺]_i could be returned to near the pre-stretch value and tetanic force showed a 10% recovery. These data demonstrate that part of the weakness of stretched muscle is caused by reduced sarcoplasmic reticulum (SR) Ca²⁺ release. Presumably much of the remaining force deficit is caused by the mechanical consequences of sarcomere inhomogeneity discussed above. Several studies have shown that resting [Ca²⁺]_i shows a small increase seen only after stretched contractions and not after isometric contractions (Balnave & Allen, 1995; Ingalls et al. 1998).

The finding of elevated resting [Ca²⁺]_i was confirmed in a study performed on mice following downhill running (Lynch et al. 1997). The measurements were made on surface fibres of whole muscles at 24 h and 48 h after the run. Increases in resting [Ca²⁺]_i were only noted at 48 h, a time at which histology showed many damaged cells infiltrated by inflammatory cells. Thus the elevated resting [Ca²⁺]_i in this study may well be a late consequence of an earlier damage pathway.

Recent studies from our own laboratory have made some progress in identifying pathways by which Ca²⁺ enters the cell following stretched contractions. These experiments have involved mdx mouse muscle, chosen because its sensitivity to stretch-induced damage is enhanced (Head et al. 1992; Petrof et al. 1993). We measured the [Ca²⁺]_i in single fibres following a protocol of 10 contractions (Yeung et al. 2005). Isometric contractions produced no detectable increase in resting [Ca²⁺]_i but following 10 stretched contractions resting [Ca²⁺]_i increased over 20–30 min. Over the same period tetanic [Ca²⁺]_i declined; both results are similar to earlier results in wild-type fibres. The rise in resting [Ca²⁺]_i could be prevented by streptomycin, Gd³⁺ or the spider venom toxin GsMTx4, each of which block stretch-activated channels (Hamill & McBride, 1996; Suchyna et al. 2000). In addition, removal of extracellular calcium prevented the rise in resting [Ca²⁺]_i confirming that it is caused by influx of Ca²⁺. These results suggest that stretch-activated channels (SACs) have a role in Ca²⁺ entry following stretched contractions.

Sodium

A series of stretched contractions also causes an increase in [Na⁺]_i in both wild-type mouse muscle fibres and mdx muscle fibres (Yeung et al. 2003a, b). The rise in [Na⁺]_i appears to follow the stretched contractions and takes 5–10 min to reach a peak. The rise in [Na⁺]_i is substantial (from 7 to 16 mm in wild-type fibres) and would be expected to be accompanied by osmotically equivalent water which would contribute to the swelling associated with stretched contractions (Foley et al. 1999). The rise in [Na⁺]_i should activate the Na⁺–K⁺ pump, increasing Na⁺ efflux which will be accompanied by osmotically associated water. This is the presumed reason for the appearance of vacuoles attached to the T-system following stretched contractions (Yeung et al. 2002a). Note that these vacuoles would constitute a region of the membrane which is distorted or stretched long after the precipitating stretched contractions. In the mdx mouse it appears that the resting [Na⁺]_i is elevated and this elevation is inhibited by Gd³⁺ and streptomycin, which also inhibit the rise of [Na⁺]_i following stretched contractions (Yeung et al. 2003a, b). These observations suggest that SACs may have an increased opening probability during normal activity in the mdx mouse and there is a further increase in opening probability triggered by stretched contractions.

pH

Protons also have an inward electrochemical gradient and are actively extruded to maintain a steady state. Resting intracellular pH was more acid by 0.17 pH units after a series of stretched contractions and no such effect was seen after isometric contractions. In addition removal of an acid load, which occurs principally on the Na⁺–H⁺ exchanger, was slowed. These findings suggest that the Na⁺–H⁺ exchanger is less effective although an increased inward leak of H⁺ has not been excluded (Yeung et al. 2002b). The increased [Na⁺]_i would provide part of the explanation for reduced effectiveness of the Na⁺–H⁺ exchanger; in addition damage to T-tubules might isolate some exchangers rendering them ineffective. Other studies have looked at longer term changes in pH regulation and show that 2 days after a period of stretched contractions pH buffering and activity of the lactate exchanger were reduced (Pilegaard & Asp, 1998).

Pathways for ionic entry

Membrane tears

The presence of overextended and underextended sarcomeres in muscles damaged by stretched contractions (Fridén et al. 1981) means that T-tubules, which are located at the A–I junction in mammalian fibres, can be severely distorted and may be damaged or torn (Allen, 2001). Mechanical damage of this, or other types, to the surface membrane would be expected to allow rapid equilibration of the intracellular and extracellular spaces causing increases in [Na⁺]_i and [Ca²⁺]_i, loss of soluble intracellular proteins, uptake of large molecular weight markers and depolarization of the membrane potential.

The existence of membrane tears was proposed by McNeil & Khakee (1992) from experiments involving downhill running by rats. They showed that about 21% of fibres in the triceps muscle had taken up albumin from the plasma indicating that the membrane had been damaged compared to 3% in sedentary controls. These results were obtained immediately after the 1 h period of exercise so that the time between stretched contractions and measurement would have varied from 0 to 1 h; after 24 h the number of fibres staining for albumin had fallen to 5%. Another technique for assessing membrane permeability involves the injection of the animals with the membrane-impermeant fluorescent dye Evans blue, and subsequently uptake of dye by damaged cells can be observed either macroscopically (Straub et al. 1997) or microscopically (Hamer et al. 2002). Uptake of Evans blue has also been used to assess the reversal of the susceptibility to muscle damage after dystrophin expression in mdx mice (Deconinck et al. 1996; Gregorevic et al. 2004). Alternatively impermeant fluorescent dyes, such as procion orange, can be perfused over isolated muscles subjected to stretched contractions and histological approaches used to assess the number of permeabilized fibres (Petrof et al. 1993). It is important to note that these experiments provide evidence of membrane permeability to large molecular weight markers; whether this permeability is caused by mechanical tears or some other mechanism remains to be determined.

Cells are capable of resealing artificially produced membrane defects and this is a process which requires the presence of extracellular Ca²⁺ and seems to involve the production of intracellular vesicles which subsequently fuse, forming a patch over the membrane defect (McNeil et al. 2000). Resealing can be inhibited by botulinum or tetanus toxin suggesting that the fusion of vesicles is analogous to exocytosis (for review see McNeil & Steinhardt, 2003).

A valuable new approach to understanding membrane repair is to cause local damage to the membrane by means of an intense laser pulse (Bansal et al. 2003). In the presence of the dye FM 1-43, which fluoresces when in contact with cell membrane, it appears that normal fibres repair these holes in less than 1 min whereas repair in mice with absent dysferlin, which causes a mild type of muscular dystrophy, is greatly slowed. Interestingly, the rate of repair in mdx fibres was normal.

These studies show clearly that membrane defects are a normal consequence of muscle activity, particularly when it involves stretched contractions. Membrane defects can be repaired and, at least for some types of injury, the repair can be very rapid. Inevitably such membrane defects would also allow equilibration of ions across the membrane.

Given this evidence of membrane defects, we have searched for localized regions of elevated ions which might be expected to accumulate on the intracellular side of such a defect. Imaging single fibres for [Ca²⁺]_i in the 30 min period following stretched contractions failed to identify any such areas (Balnave et al. 1997). This might be because Ca²⁺ entry is rapidly sequestered in the SR and/or mitochondria or because the defects were too few or too small to be detected. To eliminate the possibility that SR or mitochondrial uptake disguised entry, we imaged [Na⁺]_i in fibres during the period when [Na⁺]_i was rising but again failed to identify localized areas of elevated [Na⁺]_i (Yeung et al. 2003a). These data, combined with the observation that blockers of stretch-activated channels prevent the early rise in [Na⁺]_i and [Ca²⁺]_i, suggest that the early ion entry is through channels rather than membrane defects.

Leak channels

Calcium-specific leak channels were first described by Fong et al. (1990) in mouse and human muscle fibres. These channels have a permeability of ∼10 pS, open spontaneously at rest, are not voltage dependent and allow entry of Ca²⁺, Ba²⁺ and Mn²⁺ but not Na⁺. Their open probability is higher in mdx muscle and muscles from humans with Duchenne muscular dystrophy (Fong et al. 1990). Paradoxically, their opening probability was increased by the L-type Ca²⁺ channel blocker nifedipine but blocked by the nifedipine analogue AN 1043 (Alderton & Steinhardt, 2000a). This channel is activated by store depletion using cyclopriazonic acid judged both by patch-clamp estimates of channel open time and by the rate of Mn²⁺ influx and quenching of fura-2 fluoresence (Hopf et al. 1996). The function of these channels in normal muscle is unclear because contractile activity does not appear to influence their activity (Alderton & Steinhardt, 2000a) but in dystrophic muscle these channels appear to develop around the site of artificially induced membrane damage and the appearance of these channels was inhibited by the calpain inhibitor leupeptin (McCarter & Steinhardt, 2000). Thus the sequence of events proposed to lead to damage in dystrophic muscle is that membrane tears lead to localized Ca²⁺ entry which instigates the repair of the defect. Activation of proteolysis is thought essential for the activation of Ca²⁺ leak channels which lead to further Ca²⁺ entry and Ca²⁺-activated proteolysis which is the final common pathway of damage (Alderton & Steinhardt, 2000b).

Stretch-activated channels

SACs were first described in fetal skeletal muscle by Guharay & Sachs (1984). Characteristically these channels increase their opening probability when negative pressure is applied to the patch pipette. The channels, observed in adult skeletal muscle by several groups (Winegar et al. 1996; Vandebrouck et al. 2001), are non-selective cation channels permeable to Na⁺, K⁺, Ca²⁺ and Ba²⁺ and have a conductance of 13 pS when the pipette is filled with 110 mm Ca²⁺ (Franco & Lansman, 1990). At rest their open probability is low but it is increased steeply by reduced pressure. Typically these channels turn on or off within 1 s when pressure is applied or removed (Yeung et al. 2005). These channels are blocked by Gd³⁺ at 10–20 μm and by aminoglycoside antibiotics such as streptomycin at 100–200 μm (for review see Hamill & McBride, 1996). Neither of these agents is specific for SACs but the spider venom toxin GsMTx4 is both more potent (effective at 5–10 μm) and appears to be more specific (Suchyna et al. 2000). These channels are more prevalent in neonatal tissues and myotubes grown from satellite cells and appear to decline in frequency as the skeletal muscle becomes more mature (Haws & Lansman, 1991). Also, these channels seem to be expressed more frequently in mdx mouse muscle (Haws & Lansman, 1991) and human Duchenne muscular dystrophy myotubes (Vandebrouck et al. 2001).

Do these channels contribute to ion entry after stretched contractions? There is no direct patch clamp data establishing that these channels show increased opening after stretched contractions and if muscle stretch is comparable to negative pressure in the patch clamp, then channel activity should revert to normal within seconds of the stretch. However, the [Ca²⁺]_i and [Na⁺]_i increase over 10–20 min following stretched contractions and if SACs are involved their opening would need to be controlled by a slow process which was a consequence of stretched contractions. One possibility is that cytoskeletal elements are damaged by the stretched contraction (Lieber et al. 1996) and then influence the opening of SACs (Guharay & Sachs, 1984). As noted above, the various blockers of SACs are capable of preventing the rise of [Ca²⁺]_i and [Na⁺]_i which provides strong evidence that the influx is through these channels (Yeung et al. 2003b, 2005).

Further evidence supporting a role of SACs following stretched contractions arises from measurements of the membrane potential. McBride et al. (2000) showed that after stretched contractions the membrane potential was depolarized by about 10 mV for at least 24 h. This depolarization could arise from increased permeability to Ca²⁺ and Na⁺ either through SACs or membrane tears but this depolarization was partly prevented by either oral streptomycin, fed to the animals for several days beforehand, or by Gd³⁺, added to the muscle bath at the time of measurement. Thus SAC blockers prevent some of the depolarization which strongly suggests that SACs are responsible for one component of the depolarization and the increased ion permeability which underlies it.

A very important recent development is the first molecular identification of SACs (Maroto et al. 2005). These authors studied a mechanosensitive channel (MscCa) in frog oocytes with characteristics which are very similar to the SACs identified in mammalian muscle. Proteins extracted from oocytes were run on HPLC and protein peaks could be extracted and reconstititued in liposomes to give channels with identical characteristics. This protein had a molecular weight of 80 kDa and bound antibodies to TRPC1. Human TRPC1 was expressed in Chinese hamster ovary (CHO-K1) cells, chosen for low endogenous expression of MscCa, and produced a 10-fold increase in channel activity with similar properties to the frog channel. In CHO-K1 cells TRPC1-induced channel activity was inhibited by 3 μm Gd³⁺. TRPC1, which is widely expressed in tissues including skeletal muscle (Vandebrouck et al. 2002), has previously most frequently been characterized as a store-operated channel (Beech et al. 2004). Thus it now seems likely that the SAC described in muscle is encoded by TRPC1.

Growth factor-regulated channels

These channels, originally described as CD20 in lymphocytes, are Ca²⁺-permeable channels activated by insulin-like growth factor (IGF-1) (Kanzaki et al. 1997). Later they were renamed growth factor-regulated channels (GRCs) and shown to have considerable homology with the TRPV family (Kanzaki et al. 1999). These channels may underlie store-operated channel behaviour (Ju et al. 2003) but also appear to be stretch activated (Nakamura et al. 2001) with properties similar to the stretch-activated channel described in wild-type and mdx mice (Franco & Lansman, 1990). Iwata et al. (2003) showed that in the δ-sarcoglycan-deficient Syrian hamster, which has a cardiomyopathy and muscular dystrophy, these channels are overexpressed and can be activated by stretch, allowing Ca²⁺ entry which appears to trigger loss of creatine kinase.

Store-operated channels

Store-operated channels (SOCs) have been described a number of times in skeletal muscle though their function is unclear (Kurebayashi & Ogawa, 2001; Launikonis et al. 2003). Tutdibi et al. (1999) used the Mn²⁺ quench method to identify SOCs in muscle and showed that their activity was about twice as high in mdx compared to wild-type and was blocked by Gd³⁺, La³⁺ and Ni²⁺. The distinction between SOCs and SACs in skeletal muscle has been called into question by Vandebrouck et al. (2002) who showed that a channel apparently identical to the muscle SAC was opened by store depletion. They showed that both wild-type and mdx fibres expressed TRPC1, 4 and 6 in the membrane fraction and that antisense mRNA for TRPC1 and 4 was able to reduce the activity of spontaneous Ca²⁺ channels in mdx fibres while having no effect on voltage-dependent Na⁺ channels. This study suggests that both SAC and SOC activity could arise from a single channel encoded by the TRPC1 or 4 gene.

The above sections describe four channels leading to ionic entry associated with stretched contractions. This raises the question as to whether some of these channels are really the same? There are a number of intriguing similarities between the various channels identified but also significant differences. The various channels all have quite similar permeabilities (Table 1). Another similarity is that both the leak channel and the GRC have been reported to be both store operated and stretch activated (see references in Table 1). Furthermore, both the muscle SOC and SAC appear to be encoded by either the TRPC1 or 4 gene (Vandebrouck et al. 2002). Given the recent proposal that the vertebrate SAC is encoded by the TRPC1 gene (Maroto et al. 2005) it now seems likely that the muscle SAC and SOC are closely related and composed of TRPC1 protein. Although the functional properties of the GRCs appear very similar to SACs, their molecular identity showed the closest homology to TRPV1 (Kanzaki et al. 1999; Iwata et al. 2003). One clear distinction between the various channels is that the Ca²⁺ leak channel is the only channel reported as not permeable to Na⁺ (Fong et al. 1990). Given these differences it seems unlikely that the molecular composition of all the channels is identical. One possibility is that the various channels described are all TRP channels but exhibit the heterogeneity of subunit composition which is a feature of TRP channels (Beech et al. 2004).

Muscle damage following repeated isometric muscle contractions

Early work established that repeated isometric contractions could cause damage characterized by irreversible loss of force and release of soluble intracellular enzymes (Jones et al. 1983). This type of damage was accelerated in anoxia, could be produced in resting muscles by metabolic inhibitors, required extracellular calcium and could be produced by elevating [Ca²⁺]_i with a calcium ionophore (Jones et al. 1984; Duncan & Jackson, 1987). Thus it was thought that metabolic depletion caused a rise in [Ca²⁺]_i involving entry of extracellular Ca²⁺ and that Ca²⁺-activated intracellular pathways were critical to damage.

Gissel & Clausen have developed a model of damage in which isometric rat extensor digitorum longus (EDL) muscles are subjected to prolonged stimulation. In some studies 1 Hz twitches were repeated for 4 h (Gissel & Clausen, 2003); in others tetani (10 s at 40 Hz followed by 30 s rest) continued for 1 h (Mikkelsen et al. 2004). Developed force was only about 5% of the original level after 1 h stimulation and there was virtually no recovery. These muscles show damage judged by gradually increasing Ca²⁺ uptake and gradually increasing lactate dehydrogenase release. There were also changes in cell Na⁺, K⁺ and sucrose spaces which were compatible with a small fraction of the cells becoming damaged and therefore permeable to these substances. Other studies from this laboratory (Gissel & Clausen, 2003) have shown that electroporation or application of the calcium ionophore can also simulate the increase in LDH release only if extracellular Ca²⁺ was present suggesting that Ca²⁺ entry from the extracellular space is a key aspect of the subsequent damage.

The rat EDL muscle is around 2 mm in diameter so that in vitro preparations will inevitably develop a hypoxic core during intense activity; thus part of the damage in these studies may represent ischaemia and/or hypoxia affecting in particular the central cells. This idea is supported by studies showing that the hypoxia is capable of simulating the damage observed by raising [Ca²⁺]_i in muscle cells and that it is the central cells in a muscle which exhibit the most striking signs of damage (Bannister & Publicover, 1995). It is not yet clear whether these studies have identified a damage pathway caused by isometric contractions or whether ischaemia, which may not be a feature of fatiguing contractions in vivo, is the cause of the damage.

Consequences of ionic changes

Reduced force production

The causes of the early decline of force after stretched contractions fall into several categories. Muscles subjected to stretched contractions exhibit regions of over- and understretched sarcomeres (Fridén et al. 1981) and these cause a shift in the peak of the tension–length relation to longer lengths (Talbot & Morgan, 1998). Because of the shift of the peak of the tension–length relation, force at the original optimum length declines and there is a recovery of force as the muscle is stretched to the new optimum length (Wood et al. 1993; Brockett et al. 2001). Thus the degree of recovery of force as it is stretched to the new optimum length can be regarded as indicating the degree of damage arising from this cause (for review see Proske & Morgan, 2001).

A second cause of the early decline of force is that excitation–contraction coupling is impaired. Armstrong's group were the first to specifically suggest that changes in excitation–contraction coupling might have an important role in the muscle weakness after stretched contractions (Warren et al. 1993). They reached this conclusion from the observation that while 10 or 20 stretched contractions caused a substantial reduction in tetanic tension, the caffeine contractures were little affected. Assuming that caffeine caused a direct release of Ca²⁺ from the SR, they deduced that in stretched contractions failure of SR Ca²⁺ release was an important contributor. This conclusion was confirmed when it was shown that tetanic [Ca²⁺]_i declined in single fibres subjected to stretched contractions (Balnave & Allen, 1995). Some of the possible pathways leading to reduced force following stretched contractions are shown as a flow chart in Fig. 1.

Figure 1. Pathways involved in stretch-induced muscle damage.

Possible intracellular pathways by which stretched contractions cause reduced force production and increased membrane permeability. Dashed box indicates hypothetical mechanisms which may be involved in activating channels for Ca²⁺ entry. Dashed arrow indicates positive feedback pathway that would occur when increased membrane permeability causes elevated [Ca²⁺]_i. For further details see text. ROS, reactive oxygen species; creatine kinase (CK).

There is a range of possible causes of the reduced Ca²⁺ release following stretched contractions. One possibility is the membrane depolarization described by McBride et al. (2000). This depolarization was blocked by Gd³⁺ or streptomycin suggesting that it arose from increased activity in stretch-activated channels. In contrast another study found no evidence of depolarization following stretched contractions (Warren et al. 1993). If depolarization occurs, it would cause inactivation of Na⁺ channels and reduce the amplitude of the action potential and hence Ca²⁺ release. The rise in [Na⁺]_i described above would also reduce the amplitude of the action potential and Ca²⁺ release. The acidosis following stretched contractions could also contribute to reduced force since acidosis has been shown to reduce both the Ca²⁺ sensitivity and the maximum Ca²⁺-activated force of the myofibrillar proteins (Fabiato & Fabiato, 1978). However, these pH effects are probably relatively small at 37°C (Pate et al. 1995).

There are many suggestions that elevated resting [Ca²⁺]_i can affect the excitation–contraction coupling process (for review see Westerblad et al. 2000). Typically, situations which caused an elevation of resting [Ca²⁺]_i produce a prolonged reduction in tetanic [Ca²⁺]_i which reduces force particularly at low stimulation frequencies where the force is most sensitive to changes in [Ca²⁺]_i (Westerblad et al. 1993; Chin & Allen, 1996). Probably the clearest evidence arises from skinned fibres with intact T-tubular and SR connections in which elevation of resting [Ca²⁺]_i, for example to 2.5 μm for 1 min, leads to a substantial reduction in the ability of T-tubular depolarization to trigger SR Ca²⁺ release (Lamb et al. 1995). Under these circumstances caffeine is still capable of releasing the SR Ca²⁺ content indicating that the Ca²⁺-sensitive step lies between the T-tubule and the SR. Recent work from the same laboratory (Verburg et al. 2005) has shown that skinned muscles contain calpain 3, which is activated by the rise of [Ca²⁺]_i, and is capable of cleaving titin as well as disrupting SR Ca²⁺ release. The likelihood that calpain 3 underlies the Ca²⁺-dependent uncoupling is suggested by recent work showing that calpain 3 can be activated by submicromolar [Ca²⁺]_i (Branca et al. 1999) and by showing that leupeptin, a calpain 1, 2 and 3 inhibitor, blocked uncoupling while calpastatin, which inhibits calpain 1 and 2 but not calpain 3 (Ono et al. 2004), did not prevent uncoupling. Further evidence for the rise in resting [Ca²⁺]_i having a role in reduced tetanic [Ca²⁺]_i comes from a recent study of mdx fibres in which either removal of extracellular Ca²⁺ or blocking SACs partially reversed the reduction in tetanic [Ca²⁺]_i (Yeung et al. 2005). On the other hand a study of the stretch-induced force deficit in wild-type mouse EDL found that a cocktail of protease inhibitors (including leupeptin) was not capable of reversing the force deficit (Warren et al. 2002). This study does not support a role for calpain 3 but access of the protease inhibitors to the centre of an isolated muscle is one possible concern.

Yet another possible cause of reduced force after stretched contractions arises from recent work showing that reactive oxygen species (ROS) produced by tetanic contractions can reduce the myofibrillar Ca²⁺ sensitivity (Moopanar & Allen, 2005). Previous studies have shown that stretched contractions can reduce Ca²⁺ sensitivity (Balnave & Allen, 1996) and there are also suggestions that ROS production is increased by stretched contractions (Best et al. 1999) although only after a 24 h delay.

Protein breakdown

Intense exercise leads to increased protein breakdown both in the whole animal and in isolated muscles (Belcastro et al. 1998). Although lysosomal and ubiquitin–proteasome pathways are involved, there is a consensus that Ca²⁺-activated proteases or calpains are one of the pathways involved (Belcastro et al. 1998). Muscle contains the ubiquitous μ-calpain (calpain 1) and m-calpain (calpain 2) and, in addition, calpain 3 which is relatively specific to skeletal muscle (Goll et al. 2003). Calpains 1 and 2 are activated by [Ca²⁺]_i in the micromolar and millimolar range and thus would not be expected to be activated under physiological conditions unless additional, as yet unidentified, factors contribute to their activation. Calpain 3, while initially thought to be Ca²⁺ insensitive, has recently been shown to be extremely [Ca²⁺]_i sensitive in the nanomolar range but is normally maintained inactive while bound to specific sites in muscle, notably titin (Branca et al. 1999). The calpains are cysteine proteases and cleave only a limited range of intracellular proteins including titin, nebulin, desmin, troponin, tropomyosin and many kinases and signalling molecules. Thus their physiological role, while still uncertain, appears to involve remodelling of cells, cell fusion and the proteolytic degradation of various signalling molecules (Goll et al. 2003). The recent study of skinned toad muscle described above (Verburg et al. 2005) provides more direct evidence that calpain 3 could have a role in damage to titin and excitation–contraction coupling when quite moderate rises in [Ca²⁺]_i occur.

It is likely that calpains contribute to muscle damage following stretched contractions (for review see Belcastro et al. 1998). In support of this concept is the rise in resting [Ca²⁺]_i following stretched contactions (see earlier section), the increased activity of calpain following prolonged treadmill running (Belcastro, 1993), and the rapid breakdown of the cytoskeletal protein desmin, which is a target of calpain, following stretched contractions (Lieber et al. 1996). It has also been shown that the calpain inhibitor leupeptin can reduce the proteolysis of proteins induced by elevating intracellular calcium (Zeman et al. 1985).

The role of calpains in dystrophic muscle damage has also received considerable attention. Early studies by Turner et al. (1988) showed that [Ca²⁺]_i was elevated in mdx muscles and that proteolysis was increased but could be decreased by lowering extracellular (and presumably intracellular) calcium. Two studies have directly tested the role of calpain activation in muscular dystrophy by the use of calpain inhibitors. Badalamente & Stracher (2000) injected the calpain inhibitor leupeptin into the limbs of mdx mice for 30 days. This treatment reduced the calpain activity of limb muscles which also showed reduced central nuclei and larger muscle fibres suggesting that fibre damage and regeneration had been reduced by the treatment. Spencer & Mellgren (2002) crossed mdx mice with a transgenic mouse overexpressing calpastatin, an endogenous calpain inhibitor which inhibits calpain 1 and 2 but not 3. The mdx mice which overexpressed calpastatin showed reductions in histological signs of muscle damage and reduced regeneration of fibres. However, membrane damage assessed by procion orange uptake and creatine kinase release was unaffected. These two studies support a role for calpains in the muscle damage in mdx mice but the exact pathway is unclear and it seems that the membrane damage might involve a different pathway or perhaps involve calpain 3 rather specifically.

Increased membrane permeability

A characteristic feature of stretch-induced muscle damage is an increase in membrane permeability judged by loss of creatine kinase and entry of membrane-impermeant dyes into muscle cells (McNeil & Khakee, 1992). However, the mechanism involved remains the subject of continuing debate. The idea that membrane tears occur as a result of contractile activity and especially stretched contractions has been advocated (McNeil & Khakee, 1992; Petrof et al. 1993). It is clear that the membrane permeability is greater in mdx muscle and this has been attributed to greater membrane ‘fragility’ in the absence of dystrophin and/or the dystroglycan-related complex (Petrof et al. 1993). Support for the role of dystrophin comes from expression of dystrophin in mdx muscle and the finding that membrane permeability is thereby reduced (Deconinck et al. 1996). However, against this hypothesis is the failure to measure increased ‘fragility’ of the membrane in mdx mice (Hutter et al. 1991) and the failure to observe localized areas of elevated ion concentrations at the putative sites of tears (Balnave et al. 1997; Yeung et al. 2003a). An alternative hypothesis is that the increase in membrane permeability is a secondary consequence of Ca²⁺ entry and this possibility is supported by studies which show that elevating [Ca²⁺]_i causes increased membrane permeability (Duncan & Jackson, 1987; Howl & Publicover, 1990; Gissel & Clausen, 2003).

We recently re-examined this issue in isolated mdx muscles using uptake of procion orange as the marker of increased membrane permeability (N. P. Whitehead, M. Streamer & D. G. Allen, unpublished observations). Stretched contractions were imposed and the percentage of permeabilized cells determined histologically. In the absence of stretches 1–2% of fibres were permeabilized. Immediately after the stretched contractions, when one might expect that mechanical tears would be most apparent, 5% of cells were permeabilized. However, the number of permeabilized cells increased progressively to 12–15% by 2 h. This suggests that there is an early component of increased permeability, which might be caused by tears or other rapid processes, followed by a slower process. This conclusion was reinforced by using either streptomycin or GsMTx4 to block SACs and both of these treatments reduced the number of permeabilized cells at 1 h to around 5%. Since streptomycin and GsMTx4 prevent the rise of [Ca²⁺]_i after stretched contractions (Yeung et al. 2005) this suggests that a Ca²⁺-dependent mechanism is involved in the slow increase in permeability.

By what mechanism does elevated [Ca²⁺]_i cause increased membrane permeability? One possibility is that elevated [Ca²⁺]_i causes increased ROS production by mitochondria (for review see Brookes et al. 2004) and that peroxidation of membrane lipids then causes membrane defects to develop, perhaps by changes in membrane fluidity (Mason et al. 1997; Child et al. 1998). The importance of ROS was suggested by the effectiveness of deferioxamine, dithiothreitol and reduced glutathione in reducing [Ca²⁺]_i-induced membrane permeability (Howl & Publicover, 1990). There is also evidence that the elevated [Ca²⁺]_i activates phospholipases, particularly phospholipase A2, based on the observation that inhibition of phospholipase A2 reduces [Ca²⁺]_i-induced membrane permeability (Duncan & Jackson, 1987; Howl & Publicover, 1990). Phospholipases lead to the production of lysophospholipids and free fatty acids both of which can disrupt membrane structure.

In contrast there seems little evidence that the [Ca²⁺]_i-induced activity of calpains is involved in the increased membrane permeability. For instance leupeptin did not affect the Ca²⁺-induced increase in membrane permeability (Jackson et al. 1984). In mdx mice the possible role of Ca²⁺-activated proteolysis in the increased membrane permeability was tested by crossing a calpastatin-overexpressing mouse line with mdx mice (Spencer & Mellgren, 2002). As noted earlier, calpastatin overexpression reduced muscle necrosis and regeneration but membrane damage, assessed by creatine kinase loss and procion orange uptake, was unaffected. This suggests that calpains 1 and 2 are not directly involved in membrane damage.

Conclusions

Muscle damage is a frequent accompaniment to normal activities and muscle is capable of complete repair over a period of 1–2 weeks. In normal muscle there is evidence that Ca²⁺ entry is a part of the stretch-induced damage pathway and in the mdx mouse and the δ-sarcoglycan-deficient hamster it appears that increased Ca²⁺ entry contributes to the greater damage these muscles exhibit. The pathway(s) for the increased Ca²⁺ entry in both wild-type and dystrophic muscles are still uncertain but TRP channels are emerging as a possibility. Although membrane tears have been widely advocated there is little detailed evidence for their presence. The downstream pathways for damage are still poorly defined though roles for calpains and reactive oxygen species are emerging.