Down-regulation of c-Myc following MEK/ERK inhibition halts the expression of malignant phenotype in rhabdomyosarcoma and in non muscle-derived human tumors

In order to determine whether c-Myc is a target of the MEK/ERK inhibitor U0126 in RD cells, we performed time course experiments with 10 μM U0126 followed by immunoblotting. As shown in Figure 1A, U0126 induced early (3 hours), drastic (1d-4d) c-Myc down-regulation that persisted throughout treatment (up to 4 days). Owing to ERK inhibition (see additional file 1) [28], the level of phosphorylated c-Myc was markedly reduced (Fig 1A) before (30 min) c-Myc down-regulation began. That ERKs are upstream kinases of c-myc in RD cells, as suggested by U0126 experiments, was further demonstrated by RNA interference experiment with ERK1-, ERK2-, ERK1/ERK2-siRNA in transient transfection. After 3 days of transfection, we observed a down-regulation of total and phospho-ERKs and a lack of c-Myc phosphorylation particularly in ERK2 and ERK1/ERK2 siRNA transfected cells (Fig 1B).

While the expression level of Max isoforms (21 and 22 Kda), which heterodimerize with c-Myc [1], was unaffected (Fig 1A), the amount of c-Myc associated with Max was dramatically reduced in U0126-treated cells, as shown by immunoprecipitation experiments (Fig 1C). Equal amounts of Max were detected in the immunocomplex (Fig 1C). Taken together, these results indicate that c-Myc is a down stream target of ERKs and MEK/ERK inhibition mediates loss of c-Myc and of the c-Myc/Max heterodimer, providing one possible molecular mechanism of growth arrest i.e. that induced by the MEK inhibitor U0126.

Since c-Myc expression is well known to be down-regulated during inhibition of cell growth [32, 33] we addressed whether the observed c-Myc down-regulation is simply a consequence of cessation of cell growth due to U0126 treatment. To this purpose we performed a time course experiment with or without U0126 treatment of RD cells that were subsequently processed for FACS and immunoblotting analysis. As shown in Figure 2A, while c-Myc level (grey bars) was already significantly reduced at 3 hrs of U0126 treatment the percentage of cells in G0/G1 (black bars) was unchanged compared to untreated cells. Subsequently at 12 hrs, the percentage of cells in G0/G1 phase was highly increased by U0126 treatment, thus following of several hours the c-Myc down-regulation (Fig 2A compare 3 hrs with 12 hrs). This result demonstrated that, since in the U0126-treated cells the loss of c-Myc preceded their withdrawal from cell cycle, c-Myc down-regulation is not a consequence of the cessation of cell growth but rather it might cause growth arrest. In light of these results we hypothesised that c-Myc-dependent cell cycle proteins expression was altered too. In fact, it has been suggested that c-Myc-mediated cell transformation involves modulation of cell cycle protein expression [13, 15, 17]. Thus, we investigated whether cell cycle proteins were modulated in U0126-treated RD cells by immunoblotting experiments. Figure 2B shows that U0126 treatment induced a decrease in cyclin E2, which was stronger than that observed in cyclin E1, from 12 hrs up to 4 days, and a decrease in cyclin A and B accumulation which started at 1 day and persisted thereafter (2–4 days). Moreover, a reduction in CDK2, which forms complexes with cyclin E, A and B, started one day after treatment (Fig 2B). Of note, we have recently shown that U0126 induced a decrease in cyclin D1 and an increase in CKI, p21^WAF1 and p27 [27]. Lastly, the expression profile of the cyclins, CDK and CKI was in agreement with the hypo-phosphorylated/active form of pRb, which was detected as early as 12 hrs after treatment started (Fig 2B). These results point to the existence of a pathway in which an U0126-mediated lack of c-Myc activity affects cell cycle protein expression and mediates G0/G1 cell cycle arrest in RD cells.

In order to verify whether in RD cells loss of c-Myc might cause growth arrest (Fig 1A) in the absence of MEK/ERK inhibition by U0126, we stably transfected RD cells with vector expressing MadMyc chimera, a strong antagonist of c-Myc activity [15]. RD cells stably transfected with c-Myc expressing vector and vector alone were also prepared. The efficiency of MadMyc chimera and c-Myc transfections was assessed by immunoblotting of transient and stably transfected RD cells with c-Myc antibody, which recognizes both c-Myc and MadMyc chimera (Fig 3A). Phospho-ERK immunoblotting revealed that there were more phospho-ERKs in MadMyc stably transfected cells than in either c-Myc- or CMV-transfected cells (Fig 3A), whereas no changes were detected in transiently transfected samples. The stably transfected polyclonal populations were also analysed for growth potential (Fig 3B). Proliferation of MadMyc-expressing cells was reduced after plating by 33.7% on day 2, and up to 68.4% by day 4, thereby indicating that MadMyc chimera expression blocked RD cell proliferation. By contrast, c-Myc over-expressing cells proliferated more than control cells (CMV) from day 3, attaining a 43.6% increase over the level of control cells by day 4. In MadMyc chimera stably transfected cells, expression of cyclin D1, A and B as well as the faster migrating form of CDK2 [21], which is present in CMV, were markedly reduced, whereas CDK4 expression was not (Fig 4). Moreover, increased p21^WAF1 expression occurred in MadMyc-expressing cells (Fig 4). These data demonstrate that c-Myc pathway disruption (MadMyc chimera expression) determines a molecular pattern resembling that induced by the MEK/ERK inhibitor (Fig 2). However, cyclin E1, E2 and p27 were not altered by MadMyc expression (Fig 4), suggesting that cyclin E down-regulation and p27 enhanced expression by U0126 might be due to ERK depletion [34] in RD cells. Taken together, these data demonstrate that c-Myc pathway disruption alone (MadMyc chimera expression) establishes a molecular pathway for growth arrest in RD cells.

We have previously shown that RD cell growth inhibition can be induced by phorbol ester TPA and U0126 through different mechanisms mediated by ERK activation and inhibition respectively [27, 28].

We therefore investigated whether the growth inhibitory function of U0126 and TPA was accompanied by a decreased anchorage independent growth, as determined by a colony-forming assay in soft agar. No colony formation was observed in U0126-treated cells after 2 weeks, whereas numerous, large colonies were present in both the untreated and TPA-treated RD cells (Fig 5A). These data demonstrate that U0126, though not TPA, inhibits anchorage-independent growth in RD cells.

In order to explain the different effects of U0126 and TPA in regulating anchorage-independent growth, we investigated whether cells growing without substrate attachment were still responsive to growth arrest signals induced by U0126 and TPA. For this purpose we performed an experiment in which RD cells were grown either in suspension (see Methods) or in adherent cultures, in the presence or absence of U0126 or TPA. Growth was monitored at 1, 2 and 4 days. U0126 in both suspension and adherent cultures inhibited growth (Fig 5B), whereas TPA did not induce growth arrest in suspension, as instead occurred in adherent cultures (Fig 5B). These results demonstrated that the growth potential of RD cells can be inhibited by both TPA and U0126, whereas anchorage-independent growth is abolished by U0126 only. We then investigated whether the different effects of U0126 and TPA on growth without a substrate attachment can be correlated with the modulation of c-Myc phosphorylation and expression levels as observed in U0126-treated cells (Fig 1). We found that c-Myc expression and phosphorylation were enhanced after TPA treatment and were instead down-regulated by U0126 (Fig 6 and see additional file 2). We also analysed the levels of p21^WAF1 and cyclin D1, previously shown to be up-regulated in response to TPA-mediated growth arrest signal [27]. While U0126 down-regulated cyclin D1, TPA, according to its inability to arrest growth in suspension culture, failed to increase the levels of p21^WAF1 but still induced cyclin D1 increased expression. It is noteworthy that, in this culture condition, TPA only slightly stimulated ERK phosphorylation whereas U0126 still inhibited phospho-ERK levels. The results of growth with or without substrate attachment (Fig 5) and of the biochemical analysis (Fig 6), demonstrate that the mere growth potential inhibition in adherence condition is not a requisite for the reversal of anchorage independent growth.

The results of Figure 5 and 6 suggested to assess whether c-Myc pathways played by itself a role in the reversal of anchorage-independent growth in the absence of MEK/ERK inhibition. To this purpose c-Myc- and MadMyc chimera-transfected clones were grown in soft agar. The results of the soft agar assay of CMV-, c-Myc- and MadMyc chimera-stably transfected cells demonstrated that MadMyc chimera expression abolished, whereas c-Myc expression enhanced, colony formation, when compared with CMV-transfected cells (Fig 7). Furthermore, c-Myc overexpression sensibly rescued the anchorage-independent growth in U0126-treated cells (Fig 7). The results of MadMyc chimera indicate that the disruption of the c-Myc can be responsible of the reversion of transformation potential in RD cells.

Since c-Myc over-expression efficiently inhibits myogenesis [35], we investigated whether the functional inactivation of c-Myc rescued the myogenic program. For this purpose, RD cells were transiently co-transfected with MadMyc chimera- or c-Myc-expressing vector together with a reporter plasmid (pMyo84-luc) containing the MEF and E-box binding sites of human myogenin promoter [36]. We observed a four-fold increase in the myogenin promoter transactivation as a result of MadMyc chimera expression (Fig 8A). By contrast, c-Myc over-expression (c-Myc) led to a significant (two-fold) reduction in basal myogenin promoter activity (CMV). Moreover, no changes in myogenin or MyoD expression levels occurred in either MadMyc chimera- or c-Myc-transfected cells (Fig 8B), suggesting that MadMyc chimera expression led to the rescue of myogenic transcription factor transactivating functions. Expression of myogenic-specific markers, such as sarcomeric myosin heavy chain (MHC), occurred as a result of the restored function of myogenic transcription factors (Fig 8B).

Furthermore, MadMyc chimera stably-expressing cells predominantly displayed an elongated myotube-like cell morphology, as shown in the immunofluorescence experiment with MHC antibody (Fig 8C). Lastly, in order to ascertain whether the over-expression of c-Myc overcame the differentiative effect of U0126, RD cells transiently transfected with c-Myc or empty (CMV) vectors were treated with U0126, or were left untreated, for 4 days, and were analysed for c-Myc, phospho-ERK, myogenin and sarcomeric myosin expression. The results demonstrated that U0126 inhibited phospho-ERKs in both CMV- and c-Myc transfected cells, markedly down- regulated c-Myc, and increased myogenin and myosin expression in CMV-transfected cells. By contrast, c-Myc forced expression attenuated U0126-mediated c-Myc down-regulation, myogenin and myosin increased expression (Fig 9). This result suggested that the U0126-mediated effects on the myogenic program were counteracted by the high c-Myc level.

Taken together, these results demonstrate that the mere inhibition of c-Myc can rescue the myogenic program in RD cells by myogenic transcription factor activation, MHC expression and myogenic-like phenotype acquisition.

To investigate whether the anti-growth and anti-oncogenic effects of MEK/ERK inhibition are peculiarity of soft tissue-derived tumor cell lines, such as RD, we used IGR39 melanoma-, SW403 colon adenocarcinoma-, PC3 prostate-derived human tumor cell lines, C2C12 and NI3T3 as control untransformed muscle and non-muscle cell lines. We first investigated, in time course experiments either with or without U0126, the effects of MEK/ERK inhibition on the c-Myc phosphorylation level and expression. As shown in Figure 10A (see additional file 3), U0126 efficiently inhibited ERK phosphorylation in all the tumor cell lines tested and induced a decrease in c-Myc expression as well as in its phosphorylation throughout the treatment period (6 hrs-4 days). In the normal cell lines, such as C2C12 and NIH3T3, phospho-ERK was markedly inhibited by U0126 at early treatments (6 hrs), but recovered at longer treatments (1–4 days). U0126 treatment did not alter c-Myc expression in either C2C12 or NIH3T3 (Fig 10B). The analysis of growth potential (Fig 11A) demonstrated that U0126 treatment reduced, as in RD cells, the number of cells by 71% in IGR39, 65% in SW403 and 81% in PC3 cells. Normal untransformed cell lines were less sensitive to the growth inhibiting effects of U0126, with the number of cells dropping by 12% in C2C12 and 18% in NIH3T3. These results indicated that in normal untransformed cell lines U0126 inhibited growth slightly, while failed to induce long-lasting phospho-ERK inhibition.

Moreover, the colony-forming assay in soft agar showed that the colony formation of the IGR39, SW403 and PC3 tumor cell lines was abolished by U0126, whereas numerous, large colonies were present in the untreated cells (Fig 11B).

These data show that cell transformation of different tumor-derived cell lines is halted by inhibition of MEK/ERK pathway followed by c-Myc down-regulation.

The embryonal Rhabdomyosarcoma (RD), the prostate carcinoma PC3 (ATCC, Rockville MD), the melanoma IGR39 and colon adenocarcinoma SW403 (DSMZ, Braunschweig, Germany) human cancer cell lines were cultured in Dulbecco modified Eagle medium (DMEM), supplemented with glutamine, gentamycin (GIBCO-BRL Gaithersburg, MD) and 10% (RD, PC3 and SW403) or 15% (IGR39) heat-inactivated foetal bovine serum (Hyclone, Logan UT). C2C12 and NIH3T3 (ATCC) were grown in DMEM supplemented with glutamine, gentamycin and 10% heat-inactivated foetal bovine serum. One day after plating, cells were treated with 10 μM U0126 kinase inhibitors (Promega, Madison, WI) or 10^-7 M TPA (Sigma, St Louis, MO) for the times shown in the figures.

Cells were harvested in phosphate buffered saline, sedimented and lysed in 10 mM Tris pH 7, 50 mM NaCl, 1% NP40, 1 mM ZnCl2, additioned with protease and phosphatase inhibitors. Protein extracts were clarified by centrifugation. Supernatant, normalized as equal amounts of proteins, were incubated with Max antibody (H-2) (Santa Cruz Biotechnology, Santa Cruz CA) at 4°C for 3 hrs. 30 μl of protein-G Plus (Santa Cruz Biotechnology) were added to collect immunocomplexes. Protein G-bound immunocomplexes were washed 6 times with extraction buffer and processed for SDS-PAGE and immunoblotting.

Cells were lysed in 2% SDS containing phosphatase and protease inhibitors sonicated for 30 sec. Proteins of whole cell lysates were assessed using the Lowry method [46], and equal amounts were separated on SDS-PAGE. The proteins were transferred to a nitrocellulose membrane (Schleicher & Schuell, BioScience GmbH, Germany) by electroblotting. Immunoblottings were performed with the following antibodies: anti-c-Myc polyclonal (N-262) or monoclonal (9E10), anti-phospho c-Myc (Thr 58/Ser 62-R), anti-Max (H-2), anti-phospho ERK1/2 (E-4), anti-ERK2 (C-14 positive also for ERK1), anti-p21^WAF1 (C-19), anti-p27 (F-8), anti-Cyclin-E (HE12), -A (H-432), -D1 (M 20) and -B (H-20), CDK2 (M2) and 4 (H-22), -pRb (C-15), anti-myogenin (F-D5), a-tubulin (B-7), MyoD (C-20) (all from Santa Cruz Biotechnology) and anti-MHC (MF20, gift from Fichman D). Peroxidase-conjugate anti-mouse or anti-rabbit IgG (Amersham-Pharmacia Biotech, UK or Santa Cruz) were used for enhanced chemiluminescence (ECL) detection.

One day after plating, RD cells were transfected with plasmids using Lipofectamine Plus reagent (Invitrogen, Italy) according to the manufacturer's instructions. For the luciferase assay, the CMV or the c-Myc (kindly provided by Dr. L.G. Larsson) or MadMyc chimera plasmid (kindly provided by Dr. R. Bernards) were co-tranfected in RD cells together with pMyo84-luc (kindly provided by B.M. Scicchitano described in [36]). Total lysates were processed for luciferase activity according to the manufacturer's instructions (Promega Italia).

RD stably transfected cells were obtained transfecting cells with a plasmid encoding c-Myc, MadMyc chimera or empty vector CMV, all carrying G418-neomycin resistance. Polyclonal populations of CMV, c-Myc and MadMyc chimera expressing cells were selected using 0.4 mg/ml of G418-neomycin (Sigma) for three weeks. RNA interference experiments were performed with siRNA for ERK1 and ERK2 (Sancta Cruz Biotechnology) using Lipofectamine 2000 reagent (Invitrogen, Italy), according to the manufacturer's instructions. Briefly, cells were plated at 40–50% confluence and transfected after 24 hr with 100 nM siRNA, which we ascertained was sufficient to detect maximum fluorescence using fluorescein-conjugated control siRNA.

Cells were fixed in 4% paraformaldehyde and washed; non-specific binding sites were blocked with 3% BSA in PBS for 20 min at room temperature. Cells were then incubated for 1 hr at RT with a 1:100 dilution of the anti-MHC (MF20), specific mouse monoclonal antibody. After rinsing with PBS, the cells were incubated with anti-mouse IgG-Cy3 and DAPI (Zymed, Invitrogen, Italia).

RD cells were initiated as adherent cultures, detached and seeded in 50 ml Falcon tube at 5 × 10⁴ cells/ml in a total volume of 12 ml of same medium as adherent cultures and after 1 day additioned with TPA or U0126. All tubes were placed in an orbital shaker (~120 rpm) in a 37°C humidified incubator with 5% CO₂.

Colony-forming assays were based on standard methods. Briefly, 2 × 10⁴ cells were resuspended in 4 ml of 0.33% special Noble agar (Difco, Detroit, MI) and plated (6 cm plate) in growth medium-containing 0.5% soft agar. Colonies were photographed 14 days after plating.

Cells from adherent and suspension culture were counted using hemocytometer, and tested for exclusion of trypan blue. Data are expressed as average of triplicate + standard error.

For FACS analysis cells were harvested by trypsin-EDTA and washed; pellets were resuspended in 0,3 ml 50% FCS in PBS, additioned with 0,9 ml 70% ethanol and left O/N in the dark at 4°C before FACS analysis (Coulter Epics XL Flow Cytometer, Beckman Coulter CA, USA).