Stretch-induced activation of ERK in myocytes is p38 and calcineurin-dependent

Activation of speciﬁc mitogen-activated protein kinases (MAPKs) has been suggested to be involved in phenotype modulation of cells subjected to mechanical strain, which may be common to different mechano-sensitive cell types. We have submitted C2C12 myocytes to a static stretch and examined its effect upon the activation of ERK. Stretch induced a rapid but transient activation of ERK. This activation was however prevented when cells were pre-treated with inhibitors of p38 and calcineurin. The dependence of strain-induced ERK activation upon p38 suggests a cross-talk between these two pathways when mediating a response to a mechanical stimulus in this cell type. This suggests that cross relationships between these MAP kinases may be of crucial importance for myocyte phenotype modulation and differentiation in response to a mechanical stimulus. Copyright # 2008 John Wiley & Sons, Ltd.


INTRODUCTION
Mechanical stretch has been shown to produce both hypertrophy and changes in the expression of contractile and metabolic proteins in skeletal muscle cells in vivo. 1,2 The mechanisms by which mechanical stimuli can be transduced into phenotypic changes in this tissue are however poorly understood. We have recently carried out in vitro studies on mouse C2C12 skeletal muscle myocytes which demonstrated that alterations in contractile protein expression could be induced by passive stretch, similar to those induced in vivo, and that these changes were calcineurindependent. 3 In both cardiac myocytes and cells derived from the joint articular surface (AS) there is now a considerable body of evidence that signalling pathways involving the mitogen-activated protein kinases (MAPKs) are crucial to mediating the re-sponses to mechanical stimuli. 4,5 In particular it has been demonstrated that both p38 and extracellular regulated kinase (ERK) pathways can be co-activated in response to a mechanical stimulus and that physiological responses may involve cross-talk between these different pathways. 5 The role that these different MAPK pathways play both singly and in combination on the regulation of mechanically induced alterations in skeletal muscle phenotype is unclear.
Activation of ERK signalling has been reported in rodent skeletal muscle with treadmill running, in vitro contraction, electrically stimulated contraction in vivo, muscle overload and stretch, as well as in response to exercise regimes in human muscle. [6][7][8] These studies cannot however differentiate between the role of mechanical stimuli and that of nerve-induced contraction.
Recently we observed that passive stretch of C2C12 myocytes induces a rapid and sustained phosphorylation of p38 which was inhibited by the calcineurin/ NFAT inhibitor Cyclosporin-A (CsA). 9 We have also shown in these cells that stretch induces nuclear translocation of the transcription factor MEF2 which is known to be involved in the regulation of a number of muscle-specific genes. It was also observed in this study that stretch induced alterations in contractile protein expression. However, both the stretch-induced activation of MEF2 and altered contractile protein expression were prevented by inhibition of either the p38 or calcineurin pathways. In the present study we have examined the short-term effects of passive stretch upon the activation of ERK. A recent study in AS cells has shown that inhibition of p38-activation by the antagonist SB203580 inhibits activation of ERK in response to a mechanical stimulus. 5 In contrast, inhibition of the MEK/ERK pathway had no effect upon the activation of p38. We have therefore also examined in this study the effects of inhibiting the p38 pathway upon the levels of ERK activation in response to a mechanical stimulus. The role of the calcineurin pathway in mediating the effects of stretch upon ERK activation will also be examined due to the previously observed role of this pathway in inducing stretchinduced changes in phenotype. 3,9 MATERIAL AND METHODS

Cells and materials
C2C12 cells were cultured on silicon-based plates purchased from Flexcell, pre-coated with collagen Type1 (Sigma) and grown until 50% confluence in DMEM-GlutamaxI containing penicillin (50 mU ml À1 ) and streptomycin (50 mg À1 ml) supplemented with 10% FBS (Gibco BRL-Life Technologies) and maintained at 378C in 5%CO 2 . The medium was replaced with differentiation medium (2% Horse serum, Gibco BRL-Life Technologies) 2 days before the application of static stretch.

System used to stretch cells
To produce a continuous stretch we simply placed under the silicon basal membrane of each well to be stretched a glass sphere (marble) of diameter 1.6 cm at the centre of the well. To maintain a constant dilation we put a weight at the top of the plate applying a constant force (F) (Figure 1). In that way the silicone membrane surface area being stretched can be separated in two regions, S1 and S2. On the basis of the geometry of the system, (i.e. the axial symmetry around an axis perpendicular to the base of the well and passing through the centre of the marble), calculation using differential geometry and trigonometry can be used and show that the deformation applied on S1 is constant and does not depend on the distance between the centre and the periphery of the well (as is obvious after examination of the lateral view of the system in which S1 is represented by a line with a constant rate). Thus, the static stretch applied to C2C12 cells in the S1 region is constant and unidirectional from the edge of the well toward its centre. This is however not true for S2 as the silicone membrane on the top of the marble has a spheroid-like shape meaning that the stretch will depend on the distance from the centre to the periphery of this region. Nonetheless, S2 represents 5% of the silicone membrane surface area being stretched [S2/(S1 þ S2) Â 100 ¼ 5%], which remains negligible and can thus be neglected when compared to S1 which accounts for 95% of cells stretched. Thus, Western blot results are linked to a single cell population chiefly located on S1. Taking account of the dimensions of the well, calculus led us to determine that the percentage of dilation was about 8-9%. 9

Static stretch application
To allow a stretch of the C2C12 cells, we used specific 6-well plates containing a silicon-made basal membrane (Flexcell) coated with collagen. The medium was renewed the day before the initiation of experiments with fresh serum. Then, the following day CsA and SB203580 (Sigma and Calbiochem, respectively) were added when necessary to a final concentration of Figure 1. Schematic representation of the system used to apply a static stretch to C2C12 cells. To allow continuous stretch application we placed a marble of radius 0.8 cm (r) beneath the silicon membrane of initial radius 4 cm (R). To maintain the central location of the marble beneath the membrane, the marble was placed on a plastic support of thickness 0.7 cm (h). A calculation can demonstrate that under these conditions the deformation of the silicon membrane surface area is 8-9% 4 mM and 1 mM, respectively. The C2C12 cells were stretched for up to 3 h as described previously. 3,9 Western blotting Incubations were terminated by washing with ice-cold phosphate buffered saline (PBS) containing orthovanadate (0.4 mM) and whole cell lysates prepared in lysis buffer as described previously. 3,9 The protein content was measured using the Bradford test (Sigma) and equal quantities of protein were resolved using 10% SDS-PAGE. The gel was then transferred onto Hybond-P membrane (Amersham) that was then blocked with non-fat dry milk (5% (w/v À1 )) in PBS-Tween (1/1000 (v/v À1 )). The primary ERK antibody (cell signalling technology, concentration 1/1000 (v/v À1 )) was incubated in PBS-Tween with the membrane for 1 h. The membrane was subsequently washed 5 Â 10 min in PBS-Tween. HRP-antibody (concentration of 1/10000 (v/v À1 )) in PBS-Tween was incubated for 1 h prior to washing the membrane and chemiluminescence reaction (ECL-plus, Amersham).

Statistical analysis
Data are expressed as ERK phosphorylation levels relative to basal values, mean AE SD, n ¼ 6. Data were analysed by two-way ANOVA and Benferroni's multiple comparison test was performed. Statistical significance is denoted by Ã ( p < 0.05) and ÃÃ ( p < 0.005).

RESULTS
Passive stretch induced a rapid phosphorylation (activation) of ERK that peaked at 5 min relative to the untreated control basal value, i.e. in the absence of stretch and inhibitors (Figure 2A, B, p < 0.005). The phosphorylation of this signalling factor rapidly declined to, or below, control levels by 30 min ( p < 0.05) There was no change in total ERK protein concentrations over the time course measured (Figure 2A). Inhibition of p38 activation through the addition of SB203580 prevented the stretchinduced activation of ERK. Similarly the addition of CsA (which inhibits the action of calcineurin) produced a similar blunting of the stretch-induced ERK activation ( Figure 2B).

DISCUSSION
We have shown that static stretch of C2C12 muscle cells in vitro leads to a rapid but transient ERK activation (Figure 2). The rapid activation of this signalling cascade suggests a response directly elicited by the mechanical stimulus and not a secondary response to paracrine growth factor action (as discussed below). It is also remarkably similar to the time course for ERK activation observed during stretch-induced hypertrophy of cardiac myocytes in vitro. 4 In both cases the phosphorylation of ERK reaches a peak within 30 min ( p < 0.005) followed by a rapid deactivation to control levels over the subsequent 30 min ( p < 0.05).
As stated above we have also recently shown that passive stretch of C2C12 myocytes induces a rapid phosphorylation of p38 similar to that observed for ERK. 9 In contrast to the latter, however, the activation of p38 was not transient but was sustained over the 3 h time course examined. We therefore examined the suppression of p38 activation by the inhibitor SB203580 and found this completely inhibited ERK phosphorylation. This suggests cross-talk between the p38 and ERK/MAPK pathways in mediating the response of C2C12 myocytes to passive stretch. A previous study in endothelial cells has shown an agonist-specific cross-talk between these pathways and interestingly in AS cells it has been shown that suppression of p38 inhibits ERK activation in response to a mechanical stimulus but not in response to growth factor or pervanadate-induced accumulation of active ERK. 5,10 These data together with those in the present study suggest that ERK activation is a common response to mechanical stimuli in a number of mechano-sensitive cell types and that the response to this stimulus is also dependent upon activation of p38. This suggests that the intracellular events mediated by mechanical stretch may be common to many cell types, in contrast to those induced by other stimuli such as the action of specific cytokines. Therefore there is a real need to examine whether this response to a mechanical stimulus is similar in other cell types.
We also observed that inhibition of the calcineurin pathway through the addition of CsA prevented stretch-induced activation of ERK. This could be either through the direct action of this pathway upon ERK activation or through p38 inactivation which we have shown previously. 9 It has been demonstrated at the molecular level that CsA and FK506 (Tacrolimus-another inhibitor of calcineurin unrelated to CsA) inhibits p38 independently of ERK, which suggests that calcineurin is very likely involved in p38 hypo-phosphorylation (reviewed in Reference 11). Thus these experimental results suggest the ERK activation is likely mediated by p38 activation. However, to confirm this, further studies need to be carried out to determine the precise roles of the different MAPK signalling pathways (and their interaction) in producing phenotypic alterations in skeletal muscle in response to mechanical stimuli.
The role of calcineurin in mediating stretch responses is poorly understood and appears likely to act through downstream targets other than simply the NFAT transcription factors. Furthermore while the calcineurin pathway plays a major role in cellular hypertrophy in cardiomyocytes this does not appear to be the case in skeletal myocytes. While it is known that CsA acts to inhibit the activation of NFATs it cannot be said that this pathway is definitely involved in ERK activation in response to stretch as this inhibitor may act upon other pathways.