PKCα-mediated ERK, JNK and p38 activation regulates the myogenic program in human rhabdomyosarcoma cells

Rhabdomyosarcoma (RMS), the most common paediatric soft tissue sarcoma,arises from muscle precursor cells. In RMS, a number of well known muscle-specific markers are expressed both in vivo and in vitro(Bouche et al., 1993;Merlino and Helman, 1999;Tonin et al., 1991). RD(rhabdomyosarcoma cells), the embryonal RMS tumor cell line, fails to differentiate in spite of the expression of myogenic-specific transcription factors such as myogenin and MyoD. RD cells are believed to gain a growth advantage through the action of a mutant N-ras oncogene(Chardin et al., 1985;Kong et al., 1995;Lassar et al., 1989;Olson et al., 1987) and/or mutated tumor suppressor p53 (Germani et al., 1994), which lead to uncontrolled proliferation and secretion of autocrine growth factors and myogenic inhibitors such as insulin-like growth factors-II (IGF-II) and transforming growth factor-β (TGF-β)(Bouche et al., 2000;De Giovanni et al., 1995;Derynck et al., 1987;Minniti et al., 1994). Nevertheless, the myogenic program can be rescued by fusing RD cells with fibroblasts, which suggests that a positive regulator of myogenic transcription factors is absent in RD cells but is present in the heterocaryon(Tapscott et al., 1993).

Unlike normal myoblasts, which respond to both the proliferative and differentiative action of insulin and IGFs(Coolican et al., 1997;Florini et al., 1991), RD cells do not differentiate in response to IGF. Recent reports have suggested that elevated cyclin D1 and CDK activities contribute to the inability of RD cells to arrest growth when cultured in mitogendeprived medium(Knudsen et al., 1998). Nevertheless, RD cells treated with PKC activators, such as the tumor promoter TPA, progressively acquire a more elongated shape, which is the typical myogenic morphology, and become unresponsive to mitogens and undergo growth arrest and myogenic differentiation(Aguanno et al., 1990). This suggests that activated PKC interferes with the transduction of mitogenic signals, thereby leaving myogenic transcription factors free to initiate muscle differentiation. Thus, knowledge of the downstream pathways of PKC in RD cells might help the search for agonists, other than tumor promoters,capable of reverting cells back to their non-transformed phenotype.

Growth factor signal transduction pathways involve several kinases,including those of the MAPK and P13K/Akt kinase families(Garrington and Johnson,1999). Within the MAPK family, ERKs, JNKs and p38 have been implicated in relaying extracellular signals to the nucleus. MAPK pathways, by regulating transcription factory activity(Hill and Treisman, 1995;Treisman, 1996), mediate specific responses, including proliferation, differentiation, apoptosis and stress (Minden and Karin,1997; Pang et al.,1995; Racke et al.,1997; Robinson and Cobb,1997; Traverse et al.,1992). In a typical pathway, Ras activation initiates a protein kinase cascade that leads to MAPK activation through the intervening protein kinases Raf and MEKs.

Although downstream effectors of MAPKs that elicit specific responses have yet to be fully identified, recent reports have suggested that MEF2 is a substrate of activated p38 kinase, a transcriptional target for signalling pathways that control skeletal myogenesis(Francisco and Eric, 1999;Zhao et al., 1999). In this regard, p38 has been reported to induce muscle differentiation in both L8 and C2C12 (Cuenda and Cohen, 1999;Zetser et al., 1999), whereas P13K/Akt is thought to be involved in differentiation and hypertrophy(Bodine et al., 2001;Jiang et al., 1998;Rommel et al., 2001) of L6 myogenic lines. Moreover, ERKs are activated in C2C12 when myogenesis is promoted by the removal of mitogens(Gredinger et al., 1998). Recently, a pivotal role for p38 has been documented in the pathological myogenic differentiation of a number of RMS lines, including RD cells transfected with the p38 upstream kinase MKK6(Puri et al., 2000). In addition, evidence has emerged indicating that crosstalk between MAPKs, such as the antagonistic effect of JNKs and p38 in the L6 and C2C12 myogenic cell lines, controls myogenesis (Meriane et al., 2000).

The involvement of both PKC-dependent and PKC-independent pathways of Raf/MEK activation in response to an agonist has been reported as a possible mechanism of MAPK activation (Qiu and Leslie, 1994; Schonwasser et al., 1998). The involvement of PKCs in Raf activation has been mainly studied in cells that activate ERKs in response to TPA(El Shemerly et al., 1997;Kaneki et al., 1999;Miranti et al., 1999;Racke et al., 1997). In fact,Raf and MEK phosphorylation have been observed in cells co-transfected with PKCα and β, which provides strong evidence for the involvement of PKCs in the regulation of these pathways(Marquardt et al., 1994;Sozeri et al., 1992). Furthermore, MEK-1 activation is mediated by conventional, novel and atypical PKC isoforms in a Raf-dependent and -independent manner(Schonwasser et al.,1998).

In this study, we report that activation of the myogenic program of RD cells is dependent on PKCα-mediated ERK, JNK and p38 activation. The use of MAPK inhibitors and agonists allowed us to dissect TPA-mediated kinase activation and thereby correlate a cell response with specific kinase pathways. The results obtained show an anti-mitogenic role for activated JNKs and a role for activated p38 in the expression of muscle-specific genes. Moreover, ERK activation may be involved in the maintenance of the RD myogenic-related morphology.

The human rabdomyosarcoma cell line (ATCC, Rockville MD) was cultured as previously described (Germani et al.,1994). One day after plating, cells were treated with 10^-7, 10^-8 and 10^-9 M TPA to produce a dose-response curve, and 10^-7 M TPA was established as the dose that was effective for both MAPK activation and myogenic differentiation. For inhibitory studies, 10 μM U0126 (Promega) and 2.5 μM SB203580(Calbiochem) were added 5 minutes and 1 hour, respectively, before treatments were started. These concentrations are effective in dose-response experiments. For PKC inhibition, 60 nM Ro320432 (Calbiochem) was added 6 hours before any treatment was started. 10 ng/ml of anisomycin were added for 30 minutes and 5 days, 1 hour prior to TPA treatment. This concentration is effective for JNK and p38 activation without altering phospho-ERKs in a dose-response experiment.

For growth analysis, RD cells were harvested in trypsin-EDTA and counted in a hemocytometer chamber.

Total extracts were prepared by scraping cells in 2% SDS containing 2 mM pheny1 methy1 sulphony1 fluoride (PMSF), 10 μg/ml antipain, leupeptin and trypsin inhibitor, 10 mM sodium fluoride and 1 mM sodium orthovanadate and sonicating them for 30 seconds. Nuclear and membrane fractions were obtained by lysing cells in 10 mM Tris-HCl pH 7.5 containing protease and phosphatase inhibitors, homogenized by 20 strokes in a Dounce homogenizer and centrifuging them at 900 g. The nuclei-containing pellets were resuspended in 2%SDS containing proteases and phosphatases inhibitors. The 900 gsupernatants were centrifuged at 100,000 g to sediment the enriched membrane fractions. An aliquot of total lysates and nuclear and membrane fractions were used to evaluate the amount of proteins(Lowry et al., 1951). Equal amounts of total lysates or membrane and nuclear fractions (100 μg) were separated by 10% SDS-PAGE (Laemmli,1970) and transferred to a nitrocellulose membrane (Hybond C Extra Amersham) following standard procedures(Towbin et al., 1979). 8%SDS-PAGE was used for myosin detection. Immunoblottings were performed with the following antibodies: 1:1000 of anti-phospho-ERKs, anti-phospho-JNKs and anti-phospho-p38 (New England Biolabs, NEB), all of which recognise the phosphorylated/activated forms of these kinases; 1:250 antibodies which recognize the total ERKs, JNKs and p38 (Santa Cruz); 1:10 MF20 (supernatant of hybridoma kindly provided by D. Fischman, Cornell University); 1:1000 anti-PKCα and β1 (Transduction Laboratory); 1:500 anti-PKCγas previously described (Aquino et al.,1990); 1:500 anti-PCNA and 1:250 anti-hemagglutinin (Santa Cruz). Peroxidase-conjugated anti-mouse (1:3000) or anti-rabbit IgG (1:1500)purchased from Amersham or from Santa Cruz was used for ECL (Amersham)detection.

As the anti-phospho-MAPK antibodies detect each kinase only when dually phosphorylated, an increase in phosphorylation is an index of their activation. Densitometric analysis of bands, relative to both total and phosphorylated proteins, provided quantification (phospho-MAPK:total-MAPK) of TPA-induced ERK, JNK and p38 activation expressed as a fold increase over the control value, which was arbitrarily set at 1.

For the immunofluorescence study, untreated and treated cells were fixed with 4% paraformaldehyde (SIGMA) for 15 minutes at room temperature, washed three times, permeabilized in 0.2% Triton X-100 in PBS for 20 minutes and saturated for 1 hour with 1% of BSA (immunoglobulin free, SIGMA) in PBS. Undiluted MF20 was incubated for 1 hour and immunocomplexes were detected using (1:50) FITC-conjugated rabbit anti-mouse IgG (Zymed). For FACS analysis,untreated and treated cells were washed, and the pellets were resuspended in 50% FBS in PBS. Cells were fixed overnight after three parts of 70% cold ethanol had been added. Fixed cells were washed twice and resuspended in PBS additioned with 50 μg/ml of propidium iodide and 100 U/ml of DNAse-free RNAse (SIGMA). Cell cycle analysis of cells stained with propidium iodide was performed using a Coulter Epics XL flow cytometer (Beckman Coulter).

One day after plating (3.5×10⁵ cells/ml), cells were transiently transfected with 0.5 μg of constitutively active PKCα-coding vector (A25E) or a dominant-negative version of PKCα(K368R), kindly provided by Dr Baier (University of Innsbruck)(Baier-Bitterlich et al., 1996)or with 0.5 μg of constitutively active MKK2 (KW71A), kindly provided by Dr Ahn (Howard Hughes Medical Institute)(Mansour et al., 1996). In the present paper, MKK2 is named MEK2, according to the current nomenclature(Cohen, 1997). For the in vivo c-Jun and Elk1 trans-reporting system (Stratagene), cells were transfected with the following plasmids: (i) 0.1 μg of pFA-c-Jun or pFA-Elk1 activator plasmids, which express proteins that consist of the DNA-binding domain of yeast GAL4 and the activation domains of c-Jun or Elk1; (ii) 1 μg of pFR-Luc plasmid, which contains a synthetic promoter with five tandem repeats of the GAL4-binding sites controlling expression of the luciferase gene. In this assay the fusion activators, which are phosphorylated and activated by JNKs and ERKs, transactivate the luciferase promoter. Thus, the extent of luciferase activity reflects the activation of a specific kinase and the corresponding signal transduction pathway. Lipofectamine Plus reagent was used as the transfectant according to the manufacturer's instructions (GIBCO BRL). One day after transfection, cells were treated with TPA, or left untreated,for 1 day. Total lysates from transfected cells were processed either for SDS-PAGE and immunoblotting or assayed for luciferase activity according to the manufacturer's instructions (Promega).

To assess the effect of TPA on ERK, JNK and p38 (MAPKs) activation,time-course experiments were performed by treating RD cells with 10^-7 M TPA, the concentration that we found to be effective upon early activation of ERKs, JNKs and p38 and late accumulation of sarcomeric myosin (Aguanno et al., 1990)(data not shown). Immunoblots of total cell lysates were performed with antibodies against phospho-active MAPKs and total MAPKs.

Analysis of MAPK activation (Fig. 1) shows, in control RD cells (C0), a high level of phospho-ERKs and phospho-p38 but a low level of phospho-JNKs. In culture, a drastic decrease in phospho-ERKs and phospho-JNKs and a significant increase in phospho-p38 occur over time (C2-C5, Fig. 1). Short TPA treatment (30 minutes to 7 hours) induces a rapid and sustained increase in both phospho-ERKs (2.3) and p38 (2.4) in comparison with the levels of control cells, whereas the increment in phospho-JNK/p46,though more marked (8.6), is less persistent.

After prolonged treatments (2-5 days), neither basal (C2, C5) nor TPA-induced phospho-JNKs are detectable, whereas a decrease in ERK phosphorylation of control cells causes more evident phospho-ERK stimulation(3.2-4.3), which is in contrast to the decrease in p38 stimulation following an increase in the basal phosphorylation level in control cells. Moreover, TPA does not alter the expression level of total kinases and the fold increases in ERK, JNK and p38 phosphorylation provide a quantification of kinase activation(see Materials and Methods). The ongoing activation of MAPKs is also demonstrated by a marked increase in the phospho-ATF2 and phospho-c-jun, the downstream targets of MAPKs (data not shown). These results demonstrate that TPA induces the concomitant activation of three distinct MAPK cascades.

We previously suggested that the effects of TPA on RD cell differentiation may be caused by the activation of PKCα rather than other isoforms,(γ and β1) (Bouche et al.,1995). In this paper, we demonstrate that there is a significant and rapid (30 minutes) TPA-induced selective PKCα membrane translocation that is not accompanied by either any expression of PKCγ or an increase in PKCβ1 membrane translocation (Fig. 2A).

PKCα-induced activation of ERKs, JNKs and p38 was then studied in transiently transfected RD cells with a constitutively active mutant form of PKCα (A25E) and with a control vector (CMV) by analysing the phosphorylated forms of ERKs, JNKs and p38 after immunoblottings of total lysates 24 hours after transfection. The results inFig. 2B show that the increase in PKCα expression (two-fold) in A25E-transfected cells over the level of control vector CMV-transfected cells induces a significant increase in ERK phosphorylation (1.7-fold) and an even more marked increase in the levels of phosphorylated JNKs (2.2-fold for p46 and 7.3-fold for p54) and p38(7.3-fold). It is noteworthy that a marked phosphorylation of JNK2/p54 is present in PKCα-transfected cells. Moreover, the extent of kinase activation in PKCα-transfected cells is consistent with that observed after 30 minutes to 7 hours of TPA treatment(Fig. 1).

To further demonstrate the role of PKCα in MAPK activation, we tested the effect of both the wild-type (WT PKCα) and dominant-negative mutant of PKCα (DN PKCα) by using a c-Jun or Elk1 trans-reporting system designed for the assessment of in vivo MAPK activation (see Materials and Methods). In these experiments, RD cells were co-transfected with WT PKCα or DN PKCα with activator plasmid (pFA-c-Jun or pFA-Elk1) and a reporter plasmid pFR-Luc and were left untreated or were treated with TPA. The results in Fig. 2C show that Elk1- and c-Jun-driven luciferase activity (respectively, Elk-Luc and Jun-Luc) are induced by TPA in both control vector- (CMV+TPA) (4.7- and 2.1-fold respectively) and PKCα-transfected cells (WT PKCα + TPA)(4.2- and 3-fold respectively). By contrast, the ectopic expression of the DN-mutant form of PKCα decreases TPA-induced luciferase activity (DN PKCα +TPA) by about 61% for Elk-Luc and 38% for Jun-Luc. Taken together,these results indicate that activation of PKCα by TPA mediates MAPK pathway activation.

Evidence that PKCα is an upstream effector of MAPK activation was provided by the use of PKC inhibitor Ro320432 at a concentration (60 nM),which selectively inhibits the PKCα and β isoforms(Wilkinson et al., 1993). To analyse MAPK phosphorylation, RD cells were pre-treated with Ro320432 for 6 hours and were then left untreated or were treated with TPA for 30 minutes. Immunoblotting analysis shows that Ro320432 does not affect the level of phosphorylated MAPKs in untreated cells but completely prevents TPA-induced phosphorylation of ERKs, JNKs and p38 (Fig. 3A).

Likewise, to assess PKCα dependence on growth arrest and myogenic differentiation, growth curve (0-6 days) and late sarcomeric myosin heavy chain (MHC) (6 days) accumulation were analysed in RD cells pre-treated with Ro320432 and then treated with TPA. The results, shown inFig. 3B, clearly demonstrate that PKC inhibition prevents TPA-mediated growth arrest (Ro+TPA) without affecting the level of control cell proliferation (Ro).Fig. 3C shows that Ro320432 inhibits the TPA-induced sarcomeric MHC accumulation. It is noteworthy that,since PKCβ activation is not altered by TPA treatment, the results obtained by using Ro320432 can be ascribed to the selective inhibition of PKCα. Taken together, these results demonstrate that PKCα is an upstream effector of both TPA-induced MAPK activation and myogenic phenotype expression in RD cells.

Sustained ERK activation after prolonged treatment (2-5 days) with TPA may involve ERKs in the induction of late biological responses to TPA(Bennett and Tonks, 1997). We investigated the MEK1/2 inhibitor, U0126, which was used effectively to investigate the role of ERKs in regulating cellular responses owing to the fact that it inhibits the MEK/ERK pathway by preventing the activation of MEK1/2 and by blocking activated MEK1/2(Favata et al., 1998). We initially determined whether the simple inhibition of ERKs leads to an alteration in the basal and TPA-induced phosphorylation levels of p38 and JNKs. A time-course experiment was performed with 10 μM U0126, the minimal concentration required to obtain maximal inhibition of ERK phosphorylation in a dose-response experiment (data not shown).

Thus, the pattern of MAPK activation was analysed after immunoblotting of RD cells (Fig. 4A) pre-treated with 10 μM U0126 and left untreated (U) or treated with TPA for 30 minutes,2 and 5 days (U+TPA). Within 30 minutes, both in the presence and in the absence of TPA, a marked reduction in ERK and, unexpectedly, JNK phosphorylation is observed in U0126-treated cells(Fig. 4A). By contrast, p38 phosphorylation is instead increased after 30 minutes of U0126 treatment. After 2 and 5 days of treatment, the U0126-mediated inhibition of ERK phosphorylation is still present, both in the absence and presence of TPA,although to a lesser extent, whereas JNK phosphorylation is undetectable in both treated and untreated cells. After 2 and 5 days, phospho-p38 does not significantly change in U0126-treated cells, either with or without TPA, but does increase in control cells (C; Fig. 4A) as already seen in the experiment shown inFig. 1. It is noteworthy that p38 phosphorylation is rapidly (30 minutes) stimulated by U0126, which suggests that, upon MEK inhibition, either p38 is the only active MAPK or the activation of p38 is inversely correlated with the inactive state of ERKs or JNKs.

To rule out the possibility of a time-dependent inactivation of U0126 in the culture medium, we tested the 30 minute and 5 day conditioned U0126-containing media for their ability to inhibit ERK phosphorylation within 30 minutes of incubation of parallel RD cultures. Immunoblotting analysis shows that both these conditioned media retain their capacity to prevent ERK phosphorylation (Fig. 4B).

Furthermore, prolonged U0126 treatment (3-6 days) induces drastic morphological changes in both untreated and TPA-treated cells, appear as round-shaped (Fig. 5), although no cell detachment is observed even after 15 days of treatment. These results show that inhibition, by U0126, of the MEK/ERK pathway in RD cells parallels inhibition of JNK phosphorylation and is accompanied by a drastic morphological change.

Interestingly, the U0126-mediated JNK downregulation suggests that MEK1/2 are upstream activator kinases of JNKs, just as U937 cells have been reported to be (Franklin and Kraft,1995). To verify this hypothesis, RD cells were transfected with constitutively active HA-tagged-MEK2 or empty vector; subsequent immunoblotting analysis of total lysates shows a significant increase in the level of both phospho-JNK and phospho-ERK in MEK2-transfected cells (MEK2)when compared with control cells (CMV)(Fig. 6A).

To assess the in vivo activation of the JNK pathway, parallel RD cell cultures were co-transfected with either constitutively active HA-tagged-MEK2(CA MEK2) or control vector (CMV) and with both pFA-c-Jun activator plasmid and a reporter plasmid pFR-Luc. A dramatic increase in luciferase activity is detected in MEK2-transfected cells when compared with control vector-transfected cells (Fig. 6B).

Taken together, these results demonstrate that, in RD cells, MEK2 is an upstream activator kinase of JNK.

The concomitant deletion of MAPK activation, growth arrest and myogenic-specific marker expression induced by the PKC inhibitor(Fig. 3) led us to investigate whether ERK, JNK and p38 pathways play distinct roles in these effects by using a selective MAPK agonist or inhibitor. To study possible changes in growth potential of RD cells, we used anisomycin, which is reported to act as a true signaling agonist of JNKs and p38(Hazzalin et al., 1998),U0126, shown here to inhibit ERKs and JNKs(Fig. 4), and SB203580, which inhibits p38.

We first performed a time-course experiment to investigate whether long-lasting anisomycin treatment induces persistent JNK activation without altering cell viability. Ten ng/ml of anisomycin is the dose that induces maximal JNK activation and is ineffective on ERKs either in the presence or absence of TPA (data not shown) (Cano et al., 1994). For the time-course experiment, cells were left untreated or were treated, for different periods of time, with 10 ng/ml of anisomycin, added 1 hour before TPA treatment. Immunoblots of total lysate show that anisomycin persistently stimulates (30 minutes-5 days)phosphorylation of p38 and JNKs, though it does this to a lesser extent at 5 days (Fig. 7A). Moreover, 3 day anisomycin pre-treatment does not impair the activation of myosin expression during chase in the presence of TPA (Fig. 7B) even at doses as high as 50 ng/ml. Notably,anisomycin-pre-treated cells synthesize more myosin than control cells during chase in the presence of TPA (Fig. 7B). These results rule out the possibility that prolonged anisomycin treatment affects cell viability.

We therefore investigated whether cell proliferation is affected by prolonged JNK or p38 activation by treating cells with anisomycin, in the presence of the p38 inhibitor SB203580, to exclude the contribution of p38. 2.5 μM of SB203580 is the minimal concentration that does not affect either ERK and JNK activation or, as recently reported, PKB/Akt(Lali et al., 2000) (data not shown). Cells were treated and the number of cells was counted after 2 to 6 days of treatment. Fig. 8Ashows that anisomycin induces drastic growth arrest, which is not modified by SB203580. Similarly, TPA, which concomitantly activates ERKs, JNKs and p38,also induces growth arrest, which is not modified by SB203580 even after 4 days of treatment (Fig. 8B). However, a 30% growth increase, after 6 days of treatment, occurs in SB203580-treated cells with or without TPA. These results led us to hypothesize that the growth arrest pathway is triggered by highly activated JNKs, rather than p38, to counterbalance the mitogenic effects of a threshold level of active ERKs.

To verify this hypothesis, we used U0126, which, by downregulating ERK and JNK pathways (Fig. 4A),drastically alters the critical MAPK pathways balance, which is, in turn,likely to be the cause of the transduction of mitogenic signals. Cells were treated with U0126 for different periods of time, and the number of cells was counted after 2 to 6 days of culture. Fig. 8C shows that U0126 induces drastic growth arrest, which is not reversed by p38 inhibition for up to 4 days of culture, whereas a 38% growth recovery occurs between days 4 and 6 of SB203580 treatment.

Moreover, in order to investigate whether the decrease in cell numbers(Fig. 8A-C) observed during the various treatments was a result of withdrawal from the cell cycle, we analysed the nuclear distribution of proliferating cell nuclear antigen (PCNA), which is known to be downregulated in growth arresting cells(Mercer et al., 1991). The distribution of RD cells in the cell cycle by flow cytometric analysis was also analysed. Immunoblots of nuclear extracts from untreated cells and from cells treated with TPA, U0126 and anisomycin for 3 days(Fig. 8D) shows a consistent reduction in nuclear PCNA in TPA-treated cells, whereas nuclear PCNA in U0126-and anisomycin-treated cells is undetectable. Furthermore, the results of the flow cytometric analysis point to G1 arrest in TPA-, anisomycin- and U0126-treated cells between days 1 and 4 of treatment(Table 1).

Taken together, these results indicate that JNK activation or complete ERK and JNK downregulation are sufficient to induce growth arrest, whereas p38 activity is likely to contribute to maintaining a steady state in RD cells already tending towards myogenic differentiation.

The differentiated phenotype of myogenic lines, including RD, has recently been ascribed to selective p38 activation(Puri et al., 2000;Wu et al., 2000a). In this study, since activated p38 was detected in the absence (U0126) or in the presence of different levels of activated ERKs and JNKs (TPA) or in the presence of highly and persistently activated JNKs (anisomycin), we investigated whether activated p38 is necessary and sufficient to induce myogenic phenotype expression independently of other activated MAPKs. For this purpose, analysis of sarcomeric MHC expression was performed by immunoblotting of lysates (Fig. 9A) and by immunofluorescence (Fig. 9B) in RD cells treated with TPA, U0126, in the presence and in the absence of SB203580, and with TPA in the presence of anisomycin for 6 days. As shown inFig. 9A, both U0126-treated and TPA-treated cells accumulate more sarcomeric MHC (U) than control proliferating RD cells (C); moreover, U0126 potentiates the effect of TPA on MHC expression (U+TPA). By contrast, both SB203580 and anisomycin inhibit TPA-mediated accumulation of sarcomeric MHC (SB+TPA, AN+TPA), and SB203580 also inhibits U0126-mediated MHC accumulation (U+SB).

Immunofluorescence analysis confirmed the results of the immunoblotting and shows that filamentous fluorescence is present in TPA-treated cells, whereas unassembled diffused myosin staining is observed in U0126-treated round cells both in the presence and in the absence of TPA(Fig. 9B). It is noteworthy that myosin accumulation in U0126-treated cells rules out any toxic effect of this inhibitor, which suggests that U0126-treated cells are metabolically active.

These data indicate that p38 plays a pivotal role in the activation of the myogenic program and that persistently activated JNK might antagonise the differentiating effects of activated p38. To further verify the latter hypothesis, we treated U0126-pre-treated RD cells, in which both JNKs and ERKs are completely downregulated, with anisomycin for 30 minutes, 3 and 5 days, to reactivate JNK in the absence of activated ERKs. We then analysed the myosin expression and the level of phospho-active-JNKs by immunoblotting experiments. The results in Fig. 10 show that TPA- and U0126-mediated myosin expression is strongly inhibited by anisomycin at day 3 as well as at day 5. Interestingly, U0126 induces MHC accumulation earlier (3 days) than TPA treatment does, whereas anisomycin alone (AN) does not induce MHC expression in spite of sustained p38 activation(Fig. 7). As expected, besides activating p38, anisomycin potently and persistently (from 30 minutes to 5 days) re-activates JNK in U0126-pre-treated cells at all times of treatment(Fig. 10).

These results suggest that JNKs counteract the p38-induced myogenic program, thereby indicating that p38 activation is required but is not sufficient to induce the expression of myogenic markers when sustained activation of JNKs occurs.

Growth arrest and tissue-specific gene expression are distinct stages in myogenic differentiation that are likely to be regulated by distinct signaling pathways. We have identified the ERK, JNK and p38 cascades as targets of the signal transduction pathway induced by PKCα activation in RD cells.

Analysis of MAPK pathways in proliferating RD cells reveals a high level of activated ERKs and a low level of activated JNKs. These findings, which are in keeping with data in the literature, suggest that activated ERKs correlate with proliferation, whereas activated JNKs correlate with cellular responses to stress (Iordanov et al.,1998; Rosette and Karin,1996). Surprisingly, though myogenic phenotype is induced by p38 activation (Puri et al., 2000;Wu et al., 2000a), the consistent increase in phospho-p38 in cultured control RD cells fails to induce myogenesis, suggesting that p38 downstream pathways are inhibited. TPA,which promotes the myogenesis process in RD cells, induces a rapid and sustained increase in phospho-active ERKs and p38 and a transient activation in JNKs. Concomitant activation of the three MAPKs is not a common event in other myogenic cell lines or in other cell types, although we have previously observed similar results in an inflammatory-like response induced by TNF-α treatment in Sertoli cells (De Cesaris et al., 1998; De Cesaris et al., 1999). It has also been reported that TPA-induced macrophagic differentiation of U937 leukemic cells requires ERK, JNK and p38 activation (Franklin and Kraft,1997).

In this study, we demonstrate that PKCα is an upstream kinase of MAPK cascades, which regulate growth arrest and myogenic differentiation. Notably,the PKC inhibitor Ro320432 prevents MAPK phosphorylation as well as growth arrest and myogenic differentiation. The involvement of PKCα in the activation of the three MAPK pathways is strongly supported by transient transfection experiments. In fact, ectopic expression of the constitutively active form of PKCα, but not its dominant-negative form, induces ERK,JNK and p38 activation. In conclusion, these data suggest that PKCα is an essential activator of the three MAPK cascades which, in turn, play relevant roles in growth arrest and myogenic differentiation in RD cells(Fig. 11).

Although it has been demonstrated that normal and pathological myogenesis is dependent on p38 activation, the pathways inducing growth arrest, which enable myogenic tumor cells to activate myogenic-specific programs, require further investigation.

The use of MAPK inhibitors allowed us to dissect the TPA-mediated kinase activation and, hence, to show that ERKs and JNKs are involved in the regulation of withdrawal from the cell cycle and that p38 is required for the initiation of the myogenic differentiation program. The MEK1/2 inhibitor,U0126, which drastically and persistently inhibits ERK and, unexpectedly, JNK activation, induces rapid (1 day) and drastic growth arrest. This is demonstrated by a decreased growth potential (2-6 days), an increased number of G1-arrested cells as shown by FACS analysis, as well as by a drastic decrease in nuclear PCNA. These effects are not modified after prolonged U0126 treatment, although the extent of ERK inhibition decreased slightly between days 2 and 5, and this decrease is not due to the instability of the drug in the culture medium. Moreover, the inhibition of JNK activation is not a result of a non-specific effect of U0126, since cells transfected with a constitutively active form of MEK2 express a much higher level of activated JNKs than control cells transfected with empty vector do.

Besides preventing ERK and JNK activation, U0126 also induces an increase in activated p38; thus, growth inhibition may be due either to the downregulation of ERKs and/or JNKs or to the activation of p38. By using the p38 inhibitor SB203580 together with the MEK inhibitor, we demonstrate that inhibition of p38 can only revert growth arrest after prolonged incubation times (6 days). Similarly, in TPA-treated cells, in which the three MAPKs are activated, p38 inhibition does not reverse growth arrest before 6 days. These data point to a role of p38 in the maintenance, rather than in the induction,of growth arrest.

Since there is no specific JNK inhibitor with which to gain an insight into the role of JNKs in growth arrest, we used anisomycin, a potent JNK and p38 agonist. The fact that anisomycin does not affect ERK activity permits us to study the induction of strong JNK activation in the absence of ERK modulation. We found that persistent activation of JNKs in anisomycin-treated RD cells induces growth arrest even in the presence of the p38 inhibitor, as shown by the reduction in growth potential and in nuclear PCNA, as well as by an increased number of G1-arrested cells.

All these data indicate that a critical level of activated ERKs sustains deregulated growth of RD cells (control cells) and that a high level of activated JNKs is required to counteract the proliferative action of ERKs(TPA-treated cells). When the ERK pathway is abrogated (U0126-treated cells),the JNK pathway is concomitantly downregulated, and growth arrest occurs following the lack of proliferative action of ERKs. We conclude that growth arrest of RD cells comes from two alternative pathways triggered either by downregulation of ERKs or by activation of JNKs, which thus confer a growth arresting function on JNKs (Fig. 11).

Interestingly, the MEK inhibitor induces a round as opposed to the typical spindle-shaped morphology, without detachment or apoptotic events; this suggests that activated ERK or JNK may play a role in maintaining myogenic-related morphology. It has recently been reported that in the C2C12 myogenic cell line, MAPK-specific phosphatase (MKP-1) overexpression downregulates ERK activity sufficiently to rescue myogenic-specific gene expression but prevents mature myotube formation. These results point to a role for activated ERKs during early and late myogenic differentiation(Bennett and Tonks, 1997). Furthermore, the role of activated ERKs in late myogenesis has been confirmed by other studies in which the MEK inhibitor PD98059 prevents the myoblast fusion process without affecting the expression of muscle-specific genes in C2C12 cells (Gredinger et al.,1998). By comparing the data from those studies with our results,which show that the MEK inhibitor prevents the acquisition of a myogenic-related morphology, we suggest that a critical level of activated ERKs, which probably affects cytoskeleton or sarcomeric organization, may be responsible for the morphological phenotype of RD cells. Indeed, the recovery of ERK phosphorylation, during U0126 treatments, may reflect a requirement of the ERK pathway during the late myogenic process. However, we cannot rule out the possibility that the acquisition of a myogenic-related morphology requires functional interaction between the three MAPKs.

Surprisingly, myosin expression is detected in MEK-inhibitor-treated cells in spite of their round morphology. The finding that myogenic differentiation of RD cells occurs both after TPA-mediated ERK and JNK activation and after their downregulation, by the MEK inhibitor, seems contradictory. It is notworthy that both treatments induce an increase in p38 activation. Similarly, p38 activation, following ERK downregulation by the MEK inhibitor PD98059, has already been reported in non-myogenic cell lines, but has been found to be associated with apoptosis(Berra et al., 1998). Furthermore, growth arrest, an obligatory step for differentiation as well as for the myogenic process, is dramatically impaired in RD cells, probably because of the predominance of ERK-mediated mitogenic signals. Thus, an explanation for our contradictory data comes from the finding that, in RD cells, the attainment of the myogenic process can occur either when the ERK pathway is abrogated or when it is antagonised by JNKs. Growth arrest is necessary but not sufficient to induce the myogenic phenotype, although both these events can be dissociated. In fact, the p38 inhibitor SB203580 prevents myosin accumulation in both growth arrested TPA- and U0126-treated cells. Moreover, this finding clearly demonstrates that p38 is responsible for promoting the myogenic program, which is in agreement with data in the literature, thereby showing that p38 mediates myogenic-specific gene expression (Wu et al.,2000a).

Interestingly, no myosin expression is found in cells treated with TPA and anisomycin, the latter inducing long-lived p38 and JNK activation. The failure of TPA to induce myogenic differentiation in RD cells in the presence of anisomycin may be because of persistent JNK activation, which can inhibit the p38-mediated myogenic pathway (Fig. 11). JNKs activate c-Jun, which may antagonise MyoD function by impairing myogenic differentiation (Bengal et al., 1992). In addition, a role for JNKs has recently been proposed in the loss of Myf5 nuclear localization, an event which occurs early in the myogenic process (Meriane et al.,2000). It is noteworthy that during anisomycin removal, cells progressively re-acquire the ability to activate the TPA-mediated myogenic program, which suggests that the removal of the JNK agonist permits myogenic-specific gene expression. In addition, RD cells pretreated with a high anisomycin concentration (50 ng/ml) synthesized, during chase, more myosin that cells treated with lower doses. This result supports the finding that growth inhibition, caused by persistently active JNK, although inhibitory for myogenic-specific gene expression, does not irreversibly impair the myogenic process but allows RD cells to be more responsive to differentiation signals induced by TPA. Moreover, since transient JNK activation, following TPA treatment, does not impair myosin expression, it could be postulated that transient activation of JNKs does not interfere with the p38-induced myogenic program whereas persistent activation does. The pathway that induces myogenic markers in RD cells might, therefore, be dependent on a balance between JNK and p38 activities. In agreement with this, the U0126-induced myosin expression is also drastically inhibited by anisomycin treatment, which persistently re-activates JNKs even after long incubation times, thus supporting the inhibitory role of JNK on the myogenic gene expression program.

The authors of one recent noteworthy paper found that p38 plays a role in both growth arrest and myogenic phenotype expression in RD cell lines stably transfected with the p38 upstream kinase MKK6(Puri et al., 2000). The partial discrepancy between those results and ours, concerning the role of p38 in growth arrest, may depend on the different experimental approaches used to induce myogenic differentiation in RD cells. The forced expression of a constitutively active isoform of MKK6 kinase, when used to study downstream kinase activation, may impair the temporal sequence of responses related to kinase modulatory events. By contrast, in this study, treatment with agonists that activate endogenous kinase cascades expanded the temporal scale of the differentiation process, thereby allowing a more detailed characterization of events.

It may be hypothesized that RD cells, unlike untransformed satellite cells,which are their normal precursors, develop the transformed phenotype because of their failure to respond to physiological differentiative signals, for example, IGFs (Coolican et al.,1997). Differentiative signals inducing the satellite cell myogenic program have been reported to be mediated by the MAPK p38, which is involved in the induction of myogenic-specific marker expression(Wu et al., 2000b).

Interestingly, here we demonstrate that the coordinated activation of ERKs,JNKs and p38, all playing distinct roles in growth arrest and expression of myogenic-specific markers in RD cells, is controlled by activated PKCα. Knowledge of pathways that induce growth arrest in RD cells provides important information concerning the role of PKCα in the balance between proliferation and differentiation in the pathological myogenic process. Moreover, our data point both to a crucial role for MAPK activation length and to a drastic change in the scenario of activated kinases that induce RD cells to attain the growth arrest state alone or to move on towards myogenic differentiation. We believe that it may be possible to exploit these results for future studies of novel therapeutic approaches.

We are grateful to S. Adamo for his helpful discussion and C. Giacinti for her skillful collaboration. We are particularly indebted to A. Floridi for his generous help and support in the course of this work. We also thank Lewis Baker for reviewing the English in the manuscript. This work was supported by grants from MURST 40%, AIRC, Telethon (D107) and ASI.