Ca2+ and calpain activity
In spite of the fact that the cytosolic [Ca2+] in these fibres is maintained within the 20–80 nm range (De Backer et al. 2002), a significant cleavage rate could be measured, suggesting either that this basal activity is Ca2+ insensitive, or, as already considered, the Ca2+ sensitivity of calpain is much higher because a significant proportion of μ-calpain is autolysed in living fibres. To test the first possibility, we studied several experimental conditions where the intracellular Ca2+ availability was either reduced or increased, in long-lasting, steady-state situations. We observed that the calpain activity could be changed by 30-fold within the range of experimental conditions studied (Fig. 2). Therefore, the calpain activity in situ was definitely Ca2+ sensitive. The second possibility was confirmed by the observation that a small but significant proportion of μ-calpain is present in the 78 kDa, autolysed form, in extracts from resting muscles (Fig. 7). Most probably, the basal calpain activity in resting fibres is due to this Ca2+-sensitive, autolysed form of μ-calpain. However, at [Ca2+]i of 100 nm, this activity would be, at most, 5% of its possible Vmax, as deduced from the activity–pCa2+ relationship established by Kapprell & Goll (1989). The observation by us and others that significant amount of the 78 kDa autolysed μ-calpain is present in resting condition is puzzling, as autolysis itself requires 30 μm Ca2+, whereas cytosolic [Ca2+] is about 50-fold lower. For the same reason, it seems unlikely that the Ca2+ stimulation of the cleavage rate could be due, even partially, to an increased proportion of the 78 kDa autolysed form, though we have no experimental way to exclude it.
In some experimental conditions tested here (change of [Ca2+]o to 15 mM, presence of a low concentration of a Ca2+ ionophore), calpain activity increased (Fig. 2) while it is documented that the bulk cytosolic [Ca2+] remained constant (De Backer et al. 2002). In the case of the ionophore, we previously showed that in spite of an increase of Ca2+ influx, the intracellular mechanisms of Ca2+ homeostasis were robust enough to maintain [Ca2+]i within the normal 20–80 nm range (De Backer et al. 2002). However, this bulk situation does not preclude the possibility that in microdomains, such as the submembranous space where Ca2+ channels open, the local [Ca2+] could be higher, in a dynamic equilibrium that depends on the size of the influx, on the one hand, and on the rate constants of the homeostatic mechanisms (diffusion, binding and active uptake), on the other hand. Thus, we propose that the observed changes of the cleavage rate reflected the activity of calpain located in the submembranous space where changes of Ca2+ influx could generate local changes of [Ca2+]. This is supported by the fact that the 78 kDa, autolysed form of μ-calpain is preferentially associated with the plasma membrane (Murphy et al. 2006b).
We further tested whether [Ca2+]i transients, as seen in excitation–contraction coupling, were able to increase calpain activity. In these experiments, [Ca2+]i peaked at a maximum of 500 nm with time courses ranging from tens of milliseconds to a few seconds. We found that neither series of action potentials, single (1 Hz) or in trains (100 Hz), nor K+-dependent depolarization affected the basal calpain activity, suggesting that the time-integral of these Ca2+ transients did not reach the threshold for further calpain activation. It must be recalled that fibres were treated with BTS so that no mechanical force was produced. The present results do not preclude the possibility that physiological contractions (when mechanical stress is developed) could activate the calpain system. However, it was reported that various types of exercise in humans did not increase autolysis of μ-calpain which requires [Ca2+] > 1 μm for several minutes (Murphy et al. 2006a).
We further examined the effect of [Ca2+]i transients of much slower time courses. The sequence of depletion and refilling of internal Ca2+ stores is associated with long-lasting [Ca2+]i transients evolving over several minutes, while [Ca2+]i peaks reached 250–350 nm. As our protocol allowed the separation in time of the depletion and the refilling, we observed that both phases were associated with a significant (1.5- to 2.5-fold) increase of the basal calpain activity. This increase occurred regardless of the origin of Ca2+ (i.e. the intracellular stores, in depletion and the extracellular milieu in refilling.
Activation of mechanosensitive Ca2+ channels offered another opportunity to increase Ca2+ availability and activate calpain. This was obtained by submitting the fibres to hypo-osmotic swelling that produced a 175% increase of calpain activity (Fig. 5D, left). This effect was directly related to an increased of [Ca2+]i, resulting from an increased influx of external Ca2+ (Fig. 5C, left) through the channels: it was suppressed when Ca2+ was removed from the external medium and when the channels were blocked by the specific GsMTx-4 toxin (Fig. 6A). Thus calpain activation was not produced by the osmotic swelling per se and the resulting membrane stress and deformation. Here again, [Ca2+]i transients evolved over several minutes and individual peaks rarely exceeded 120 nm (Fig. 5C)
Our results showed that [Ca2+]i transients must attain a certain amplitude and time course combination to activate calpain, suggesting the presence of some integrative mechanisms. This was indeed demonstrated in experiments on purified μ-calpain (from erythrocytes) subjected to repetitive (1–50 Hz) Ca2+ pulses producing 10 μm peaks of [Ca2+] (Tompa et al. 2001). Our results suggest that such an integrative mechanism might operate in situ for [Ca2+]i transients, the amplitudes of which remained well below 1 μm (i.e. similar to those attained by most Ca2+ signals in physiological conditions). Recent findings suggest that the ‘integrative mechanism’ proposed above might involve calpain phosphorylation by Ca2+ and/or by stress-activated kinases. Indeed, m- and μ-calpains can be phosphorylated and activated by protein kinase Cι and by extracellular signal-regulated kinases 1/2 (ERK1/2) in human lung cancer cells (Xu & Deng, 2004, 2006) It is interesting that in skeletal muscle these latter kinases are activated by mechanical stretch (hence the involvement of mechanosensitive Ca2+ channels) and by physical exercise (Kumar et al. 2004; Nakamura et al. 2005). Alternatively, calpain phosphorylation might provide an independent activation mechanism.
Is the basal calpain activity measured in resting fibres still sensitive to the calpastatin inhibition? The basis of this inhibitory effect is calpastatin binding to calpain, which is itself Ca2+ sensitive; it shows a much higher Ca2+ sensitivity for the autolysed forms (both m- and μ-calpain). At the cytosolic [Ca2+] of resting fibres, this binding would be ∼60% complete for autolysed μ-calpain (Kapprell & Goll, 1989). Thus the basal activity we measured most probably reflects that of the calpastatin-free and autolysed μ-calpain. In fibres from transgenic mice over-expressing calpastatin (300-fold increase), we observed that the basal calpain activity was further reduced by 40% (Fig. 2). Most probably the high concentration of calpastatin increased the relative importance of the calpain–calpastatin complex, but the very low [Ca2+]i prevented a complete inhibition. Moreover, one cannot exclude a differential calpain and calpastatin localization, as the latter has been observed to be confined to aggregates (in neuroblastoma; De Tullio et al. 1999), so that, in spite of an elevated calpastatin content, the formation of the calpain–calpasatin complex and the inhibition of calpain activity would be marginal or moderate in our resting conditions. This would not preclude an important inhibitory effect as [Ca2+]i increases by the combined effect of calpastatin solubilization (De Tullio et al. 1999) and increased binding of calpain to calpastatin. At 0.5 μm Ca2+, the calpain–calpastatin complex would amount to 90% (Kapprell & Goll, 1989). The fact that all our protocols intended to increase [Ca2+]i, produced, at most, a 3- to 5-fold increase of the basal calpain activity (Figs 2 and 4D) might reflect a self-limiting process, resulting from two antagonistic effects of Ca2+: activation of the enzyme and binding of its inhibitor calpastatin.
Calpain activity in situ was affected by several compounds documented as specific inhibitors that reduced the activity to 40–50% of the resting values, provided they were in contact with the fibres for at least 30 min before measurements. None completely suppressed calpain activity. They were less efficient than diffusible EGTA which reduced [Ca2+]i to undetectable levels (∼10−9m) and produced the more potent inhibition (Fig. 2). The absence of effect of bortezomid, a specific inhibitor of the proteasome, indicates that this proteolytic system did not contribute to the cleavage rate of Boc-Leu-Met-CMAC.
Calpain activity in dystrophin-lacking fibres (mdx)
Dystrophin-lacking fibres isolated from the mdx mouse, showed a 1.5-fold increase of calpain activity in resting conditions (Fig. 8), while the cytosolic [Ca2+] remained within normal values (Fig. 5C, right). We found no evidence that this effect reflected a significant increase of autolysed μ-calpain (Fig. 7). This elevated calpain activity could be normalized by a 10-fold reduction of [Ca2+]o as anticipated from previous results (Turner et al. 1988). As an elevated activity of the voltage-independent/mechanosensitive Ca2+ channels had been observed in mdx fibres, the increase of calpain activity seems directly related to the increased Ca2+ influx through these channels. When the latter are further stimulated by hypo-osmotic swelling, calpain activity was also increased (Fig. 5D, right), an effect that was suppressed by the specific channel blocker GsMTx-4. As discussed above, a higher Ca2+ influx is expected to increase the submembrane [Ca2+] and stimulate the autolysed μ-calpain preferentially located there, as already discussed for normal fibres. Submembrane [Ca2+] has been reported to be ∼3-fold higher in mdx fibres (Mallouk & Allard, 2000), but this observation was recently challenged using a membrane-bound Ca2+ indicator (Han et al. 2006) (for a detailed discussion of this controversial point, see Gillis, 2007). However, increased calpain phosphorylation could also contribute to the elevated cleavage rate in mdx fibres, as a higher level of activation of ERK1/2 has been observed in the mdx muscle in response to stretch, which is a Ca2+-dependent process (Kumar et al. 2004).
Voltage-independent/mechanosensitive Ca2+ channels display a higher activity in the absence of dystrophin. In resting conditions, the mechanisms of intracellular Ca2+ homeostasis are robust enough to cope with the increased Ca2+ influx and to maintain [Ca2+]i within normal values (Fig. 5C, right). Notwithstanding, the increased Ca2+ influx is able to stimulate some calpain activity (Fig. 5D, right) probably located near the plasma membrane. However, in response to stimulation of the channel activity by mechanical stress, [Ca2+]i exceeded normal values and further stimulated calpain activity (Fig. 5C and D, right). Assuming that fibre swelling imposes on the plasma membrane a stress that simulates the one generated by contraction, the present results suggest that mechanosensitive Ca2+ channels could be stimulated by contractile activity, and that this stimulation, amplified in the absence of dystrophin, could result in greater calpain activity in mdx fibres. In particular, his would be the case in eccentric contractions (Allen et al. 2005). The moderate calpain activation observed in mdx fibres is, however, not a specific feature of dystrophinopathy, as it was observed in calpain-3-deficient fibres (present results) and in δ-sarcoglycan-lacking fibres from the mutant hamster (Bartoli et al. 2006).
Calpain activation: a step towards fibre necrosis or a fibre protection mechanism?
A widely held view is that calpain activation contributes to muscle wasting in dystrophinopathy (Turner et al. 1988). This view is supported by the fact that calpain inhibition alleviates dystrophic disorders (Spencer & Mellgren, 2002). However, the only clear cause–effect relationship between calpain activation and a structural/functional defect comes from the recent work of Murphy et al. (2006b) on normal rat fibres. They showed that application of pre-activated exogenous μ-calpain to stretched (200%), skinned fibres produces a sharp decline in passive tension resulting from proteolysis of the tension-bearing filaments of titin. This required [Ca2+] of1 μm. The same effect was obtained by dipping the fibres in a 5 mm Ca2+ solution. These effects resulted from a massive activation of calpain and were predominantly due to the high proteolytic activity of the 76 kDa autolysed form (see Fig. 7 for the various forms of calpain). The possibility cannot be excluded that the experimental protocol artificially increased the susceptibility to calpain by exposing regions of the titin molecule that are recoiled at normal muscle length. By contrast, the situation in unstretched, resting muscles, normal and mdx, is far from these experimental conditions: [Ca2+]i is maintained at 100 nm, the 76 kDa form of μ-calpain is not detected and resting tension is negligible. Light microscopic observations of collagenase-isolated fibres do not reveal structural alterations (the striation pattern remained very sharp and regular) even after stimulation by Ca2+ of the calpain activity of the amplitude reported here.
Recently, it was shown that a localized wound of the plasma membrane of a muscle fibre reseals spontaneously in a matter of a few tens of seconds. This requires dysferlin and its activation by Ca2+ (Bansal et al. 2003). Moreover, resealing involves membrane fusion and a local remodelling of the cytoskeleton for which calpain activation by Ca2+ is essential for the degradation of talin and vimentin; resealing and thus cell survival is highly compromised in calpain-null mutant cells or in the presence of calpain inhibitors or EGTA (Mellgren et al. 2006). The presence of the highly Ca2+-sensitive 78 kDa form of μ-calpain close to the membrane would provide a ready-to-work system for membrane repair. In this context, the activation of Ca2+ leak channels in the surroundings of microlesions in myotubes (McCarter & Steinhardt, 2000) may be seen as a way to provide a channel-controlled influx of Ca2+ needed to activate both the dysferlin and the calpain systems. Loss-of-function mutations of dysferlin are responsible for limb girdle muscular dystrophy type 2B, suggesting unexpectedly that membrane wounds and repairs are common events in a healthy muscle fibre. Thus, instead of being seen as deleterious, calpain activation, at the level observed here, may be considered as playing an important function in maintaining fibre integrity. In these circumstances, calpain inhibition could have adverse effects. This view could be extended to mdx fibres where the slightly higher (∼1.5-fold) calpain activity may be seen as an adequate response to a higher occurrence of wounds in a plasma membrane made fragile by the loss of dystrophin and its associated glycoproteins. However, if membrane damage allows [Ca2+]i to rise and remain far above physiological values then a massive activation of calpain would occur with structural/functional damage as reported by Murphy et al. (2006b). They showed that the threshold [Ca2+]i for damage was in the 1–10 μm range, and full effect occurred at much higher Ca2+ concentrations. This probably occurs during eccentric contractions to which mdx fibres are highly susceptible (Moens et al. 1993). Indeed, a very recent study demonstrated that high calpain activity was specifically detected in fibres exhibiting structural damage after extensive downhill run (Bartoli et al. 2006). In these circumstances, calpain inhibition (e.g. by high levels of calpastatin or pharmacological compounds) would be beneficial.