Background
Rhabdomyosarcoma is the most common soft tissue tumor in childhood, accounting for more than half of all soft tissue sarcomas in children [
1,
2]. The embryonal rhabdomyosarcoma subtype (ERMS) accounts for about 70 % of all rhabdomyosarcoma cases. In ERMS tumors, the Ras pathway is frequently mutated [
3]. Dysregulation of the Ras pathway may be a crucial event in muscle precursor cells leading to ERMS fate, as described in mice models [
4,
5].
Tumors contain a sub-population of cancer stem cells (CSCs) or cancer stem-like cells which are considered to be responsible for tumor initiation, propagation, invasiveness and metastasis [
6,
7]. Owing to the lack of universal markers for the isolation and identification of CSCs, enrichment of CSCs from tumors or cell lines through a non-adhesive culture system has been adopted as a means of characterizing their partial stemness phenotype [
8‐
10]. Several CSC markers have been identified in solid tumors including cell surface markers CD133, CD90, CD117, CXCR4 and CD166, soluble protein aldehyde dehydrogenase 1 (ALDH1), and transcription factor nanog [
6,
11,
12]. In particular, CD133 has been identified as a central marker of ERMS CSC [
13]. In stem cell (SC) medium, ERMS cell lines form spheres, named rhabdospheres, that are enriched in the CD133 positive population and have been shown to be more tumorigenic and more resistant to commonly used chemotherapies [
13]. CXCR4, which plays an important role in chemotactic and invasive responses in several solid tumors, increases in ERMS spheres [
14]. A high expression of CD133 in human ERMS samples also correlates with an unfavorable clinical outcome [
13]. Moreover, ALDH1 has been reported to be a potential marker of CSCs in ERMS [
15] and of muscle stem cells that spontaneously undergo myogenic differentiation [
16], as well as a marker of rapid isolation of the human myogenic progenitors for cell therapy [
17].
Signaling pathways in cancer stem cell biology are increasingly being used to investigate the mechanisms underlying the drug resistance, tumor relapse and dormant behavior exhibited by many tumors [
18,
19]. The inhibition of EGFR-mediated MEK/ERK signaling impairs stem cell self-renewal and reduces the propagation of the DU145 prostate cell line [
20]. Moreover, disruption of K-Ras or downstream signaling in colorectal cancer cell lines impairs CD133 expression [
21].
One of the main indicators of the sensitivity of cancer cells to chemotherapeutic agents is believed to be apoptosis, particularly via the intrinsic mitochondrial cascade. Various integrated signals converge on BAK, an important effector of intrinsic apoptosis. BAK is negatively regulated by BMX, a tyrosine kinase, which associates with and phosphorylates BAK, thereby contributing to its inactivation [
22]. BMX is often overexpressed in cancer cells to promote the survival of cancer.
It has been suggested in a previous work that MEK/ERK signalling is directly involved in the prevention of apoptosis [
23]. The authors discussed the mechanism underlying BAK-mediated mitochondrial apoptosis and MEK/ERK-mediated inhibition of tyrosine phosphatase, which affects BAK phosphorylation and activation, thereby contributing to maintain cell survival [
23].
Besides playing a role in the inhibition of apoptotic mechanisms, BMX is also required for maintenance of stem-like phenotypes in glioblastoma [
24].
In ERMS, the main pathways involved in CSC survival and growth in the tumor environment have not yet been clearly defined. The MEK/ERK pathway has been shown to play a critical role in controlling cell growth, radioresistance and differentiative signals in the RD [
25]. An interplay between ERKs and p38 mitogen-activated protein kinase (MAPK) has also been hypothesized [
26].
In this study, inhibition of MEK/ERK signaling by U0126 reduces the size and tumorigenicity of the stem-like RD cell population. Furthermore, U0126 treatment enhances the inhibitory effect of radiation on stem-like rhabdomyosarcoma cells by favoring apoptosis. These findings highlight the potential advantage of using MEK/ERK inhibitor to target embryonal stem-like rhabdomyosarcoma cells.
Methods
Embryonal rhabdomyosarcoma cell lines, RD and TE671 (HTL97021), were procured from the American Type Culture Collection and Interlab Cell Line Collection, respectively.
Alveolar RH30 was obtained from DSMZ (Braunschweig, Germany). Sphere-forming cells were obtained as described [
27]. Briefly, RD cells were cultured in anchorage-independent conditions (low attachment flasks or plates, Nunc) in SC-medium consisting in DMEM:F12 medium (Gibco-Invitrogen) with progesterone (2 μM), putresceine (10 μg/ml), sodium selenite (30nM), apo-transferrin (100 μg/ml) and insulin (50 mg/ml) (all from Sigma-Aldrich). Fresh human epidermal growth factor (20 ng/ml) and fibroblast growth factor (20 ng/ml) (PeproTech, London, UK) were added twice/week until cells formed floating spheres.
To evaluate the primary sphere formation, cells from sub-confluent (70–80 %) monolayer cultures were plated at a density of 100, 500 or 1000 cells in a 24-well culture plate (Corning Inc, Corning, NY, USA). For the sphere formation assay, the number of primary tumorspheres was counted. The primary spheres were mechanically dissociated and re-plated together with residual cell aggregates to obtain the second generation of spheres (Additional file
1: Figure S1).
MEK/ERK inhibitor U0126 (Promega, Madison, WI, USA) and MAPK p38 inhibitor SB203580 (Calbiochem, Nottingham, UK) were dissolved in dimethylsulfoxide (DMSO; Sigma-Aldrich) and used at the concentrations indicated. For a dose–response curve, RD cells, plated at a density of 1000 cell/well, as described above, were treated with varying concentrations of U0126 (1–20 μM) (3 wells per treatment) and spheres were counted. SB203580 was used at 2.5 μM, according to previous tests [
26]. TE671 and RH30 were treated with 10, 20 or 40 μM U0126.
Radiation was delivered at room temperature using an x-6 MV photon linear accelerator, as previously described [
28]. The total single dose of 4 Gy was delivered with a dose rate of 2 Gy/min using a source-to-surface distance (SSD) of 100 cm. A plate of Perspex thick 1.2 cm was positioned below the cell culture flasks in order to compensate for the build-up effect. Tumor cells were then irradiated placing the gantry angle at 180°. Non-irradiated controls were handled identically to the irradiated cells with the exception of the radiation exposure. The absorbed dose was measured using a Duplex dosimeter (PTW).
Flow cytometer analysis
Stem cell markers in rhabdomyosarcoma cells were evaluated by staining with monoclonal antibodies conjugated with phycoerythrin (PE) anti–CD133 anti–CD90, anti–CXCR4, anti−CD105, and with allophycocyanin (APC) anti-CD117(all from BD Biosciences, Buccinasco, Italy).
Appropriate isotype controls for non-specific binding were used for each antibody. A minimum of 50,000 events were acquired for each sample by a flow cytometer (FACSCalibur, BD Biosciences) using CellQuest software (BD Biosciences) for data acquisition and analysis.
Cell cycle analysis
A DNAcon3 kit (Dako, Glostrup, Denmark) was used for DNA staining. Briefly, 1 ml propidium iodide solution was added to each test tube containing dehydrated buffer mixture. After 10 min, cells were added to each tube and incubated at 4 °C for 1 h. Analysis was performed with FACScalibur, and the cell-cycle distribution was analyzed using Mod-Fit software (Verity Software House, Topsham, ME, USA).
Aldefluor assay
The stem cell population expressing ALDH enzymatic activity was assessed by means of the Aldefluor™ kit (StemCell Technologies, Vancouver, BC, Canada), according to the manufacturer’s instructions. Briefly, 1 × 105 cells were resuspended in Aldefluor assay buffer containing ALDH-substrate, and incubated for 45 min at 37 °C; a set of cells was stained using identical conditions with diethylaminobenzaldehyde, a specific ALDH inhibitor, as a negative control. Samples were analyzed by means of FACSCalibur, and the resulting fluorescence profiles were compared.
Immunoblot analysis
Cells were lysed in Tris–HCl 10 mM pH 7.5, 1 % SDS containing phosphatase and protease inhibitors (Roche, Mannheim, Germany). Proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane (Schleicher & Schuell, BioScience, Germany) by electroblotting. Immunoblotting was performed with the following antibodies: anti-Nanog, anti-ERK, anti-phospho-ERK1/2,anti- myogenin, anti-αtubulin, anti-GAPDH, anti-DNAPKcs, anti-Rad51, anti-BMX (all from SantaCruz Biotechnology, Santa Cruz, CA), anti-phospho-p38 (Cell Signaling Technology, Danvers, MA, USA) and anti- myosin heavy chain (MHC) (MF20 supernatant of hybridoma). Anti-mouse or anti-rabbit HRP-conjugated antibodies (Bethyl Laboratories Inc., Montgomery, TX, USA) were used for ECL (GE Health Life Sciences, Piscataway Township, NJ, USA) detection. Signals from protein bands were digitally acquired and quantified using the Chemidoc XRS system (BIORAD, Brossard, QC, Canada).
Invasion assay
Invasion assay was used to assess the invasive potential of the cells, according to the standard protocol. Briefly, cells were plated in the upper chamber of a 24-well Transwell plate (8 μm pore size filter; Corning Inc., Corning, NY, USA) at a density of 80,000 cells/well in 200 μl of SC medium. 750 μl of SC medium containing 10 % FBS was added to the lower chamber as a chemoattractant, or SC medium alone as a negative control. After 24 h at 37 °C, non-invading cells were removed from the upper surface of inserts with a cotton swab and invaded cells were fixed with 4 % paraformaldehyde and stained. The number of cells that invaded the filter was counted using a bright-field microscope. Ten randomly selected fields were counted for each filter and the experiments were carried out twice in triplicates.
Apoptosis assay
The Annexin/V-PI assay was carried out using the Annexin V-FITC Apoptosis Detection Kit (MERK Millipore). Rhabdospere RD cells were harvested and the pellets were immediately resuspended in the binding buffer provided. Cells were stained with 5 μl of FITC Annexin V and 5 μl of PI. The mixture was left to incubate at room temperature for 15 min and then was acquired by FACSCalibur (BD Biosciences) and analyzed using CellQuest software (BD Biosciences).
NOD/SCID mice transplantation
NOD/SCID mice were bred and maintained under defined conditions at the Experimental Animal Welfare Sector of the Istituto Superiore di Sanità, Rome, Italy. All animal procedures complied with the European Community Directive on the welfare of experimental animals (Directive 2010/63/EU) upon approval of the protocol by the Institutional Animal Experimentation Committee. Equal numbers (2× 10
6 cells) of adherent RD parental cells or rhabdosphere-derived cells suspended in 100 μl of phosphate buffered saline were subcutaneously injected into four- to six-week-old NOD/SCID female mice. Viability of the injected cells was confirmed by trypan blue (Sigma) staining prior to injection. Intraperitoneal injections of U0126 started when tumors reached a volume of 80–100 mm
3. U0126 solution was prepared in DMSO as a stock solution of 10 mmol/L, and the amount of drug (25 μmol/Kg/mouse) to be injected into a set of mice was diluted with carrier solution (40 % DMSO in physiologic solution). The U0126 dose used here had previously been tested and found to be non-toxic in mice and to down-regulate ERK1/2 in tumors [
25]. U0126 was administered 3 times per week. This protocol was chosen because full inhibition of ERK activation is guaranteed in vivo after 24 h and was documented after this time [
25]. Four weeks after the beginning of treatment, the mice were killed by cervical dislocation and the tumors removed and weighed.
For U0126 pre-treatment, RD cells were cultured in SC medium in the presence of 10 μM U0126 for 15 days, followed by the transplantation of 2 × 106 cells into the flank of NOD/SCID mice. Following cell injection, mice did not receive U0126 for the rest of the in vivo period. Four weeks after tumor appearance, the mice were killed by cervical dislocation and the tumors removed and weighed.
Assessment of in vivo response to treatments
The effects on tumor growth of different treatments were evaluated by measuring the following: (1) tumor volume measured during and at the end of the experiment. Tumors were measured with a vernier caliper every 4 days and their volume, expressed in mm3, was calculated as length x (width)2/2. (2) Tumor weight measured at the end of experiment; (3) tumor progression, defined as an increase greater than 100 % of the tumor volume at the beginning (80–100 mm3) of the U0126 treatment delivered intraperitoneally; (4) time to progression. In the experiments in which the incidence of tumor development was studied, the occurrence of this event was defined as the appearance of a measurable (80–100 mm3) subcutaneous tumor lesion at the site of cell injection.
Dissociation of tumor into single-cell suspension
For the FACS analysis, cells from xenograft tumors were obtained by means of the tissue dissociation protocol that combines mechanical and enzymatic approaches. Isolated tumor was minced with a sterile scalpel and digested with a solution of 1.5 mg/ml collagenase II (Gibco,) and 20 μg/ml DNAse I (Sigma) added to the tumor during mincing to facilitate tumor dissociation. After 2 h of incubation at 37 °C, cells were dissociated, washed and processed for FACS analysis.
Statistical analysis
Continuous variables were summarized as the mean and S.D. or 95 % CI for the mean. Statistical comparisons between controls and treated groups were established by carrying out the Student’s t test for unpaired data (for two comparisons). Dichotomous variables were summarized by absolute and/or relative frequencies. For dichotomous variables, statistical comparisons between control and treated groups were performed by means of the exact Fisher’s test. The incidence of tumor development and tumor progression were analyzed by using Kaplan-Meier curves and Gehan’s generalized Wilcoxon test. Curves were compared by means of the log rank test and determination of the hazard ratio (HR). All the tests were two-sided and were determined by Monte Carlo significance. P values <0.05 were considered statistically significant. SPSS (statistical analysis software package, IBM Corp., Armonk, NY, USA) version 10.0 and StatDirect (version. 2.3.3., StatDirect Ltd, Altrincham, Manchester, UK) were used for the statistical analysis and graphic presentation.
Discussion
Cancer stem cell research is becoming increasingly important in the investigation of the development, spread, resistance to chemo- and radio-therapy and relapse of cancer.
We previously demonstrated, both in vitro and in vivo, the responsiveness of the RD cell line to MEK/ERK inhibition, which induces growth arrest, myogenic differentiation, radiotherapy sensitization and tumor growth impairment [
25,
28]. On the basis of these data, we decided to assess the contribution of the MEK/ERK pathway in controlling the cancer stem-like compartment in the ERMS cell system.
It is generally agreed that tumorspheres enriched in cancer stem-like cells are highly tumorigenic [
13,
27,
31,
32]. By culturing RD cells in SC-medium, we obtained rhabdospheres enriched in positive CD133, CXCR4, ALDH and Nanog stem-like cells that are highly tumorigenic in vivo. These results are consistent with those reported in previous studies [
13‐
15].
Since Ras/ERK is an upstream pathway of CD133 expression [
21,
33], the use of U0126, a MEK/ERK inhibitor, proved useful to study the dependence of a stem cell-like population on the MEK/ERK pathway. We used U0126 to demonstrate, for the first time, the critical contribution made by MEK/ERK signaling to the cancer stem-like phenotype in the RD cell line. Indeed, sphere formation is inhibited by U0126, in a dose-dependent manner, in the embryonal RD cell line though not in the Ras negative alveolar RH30 cell line. The RH30 cell line does not exhibit persistent ERK inhibition by U0126 whereas it does exhibit very low levels of CD133, thus suggesting that other pathways underlie the alveolar stem cell-like phenotype.
In rhabdospheres derived from RD cells, the stem cell markers CD133, CXCR4 and Nanog were enhanced and were dramatically inhibited by U0126 treatment, whereas ALDH activity was increased by MEK/ERK inhibition. This finding is in agreement with a recent report showing that high ALDH1 activity is related to the myogenic potential of muscle precursors [
34]. It has also been demonstrated that the induction of myogenic differentiation occurs spontaneously in myogenic precursors that highly express ALDH1 [
16]. The increased ALDH activity observed under MEK/ERK inhibition in the system we adopted may indeed be related to myogenic differentiation. In this regard, the differentiative markers myogenin and MHC are enhanced by the U0126 treatment of RD cells in SC-medium. We may therefore speculate that MEK/ERK pathway inhibition induces molecular reprogramming, which rescues the myogenic precursor phenotype in rhabdomyosarcoma cancer stem-like cells.
Together these data strongly suggest that the differentiation boost resulting from MEK/ERK inhibition turns off cancer stem-like phenotype expression.
Published studies by us and other authors have shown that MAPK p38 plays a pivotal role in myogenic differentiation [
26,
35]. Here we found that inhibition of the differentiative action of MAPK p38 significantly enhances the expression of Nanog and phospho-active ERK1/2, correlates with an increased S phase of the cell cycle and accelerates sphere formation. The size of the CD133 positive cell population was not affected markedly under MAPK p38 inhibition in this study, and that of the CXCR4 positive population actually decreased. By contrast, Nanog and phospho-active ERK1/2 were significantly enhanced, thereby suggesting that they play a major role in rhabdosphere formation. These results agree with those recently reported on the role of Nanog in ERMS as an inducer of sphere formation, as an important gene for tumor promoting properties and as a prognostic marker for ERMS patients [
36].
The positive effects of MAPK p38 inhibition on the stem-like phenotype are reverted by MEK/ERK inhibition when treatment with inhibitors is concomitant, thereby demonstrating that chronic MEK/ERK inhibition strongly impairs the stem-like phenotype in embryonal rhabdomyosarcoma.
The relevance of the active MEK/ERK pathway in cancer stem-like cells with tumor initiating properties is demonstrated by the significant delay in tumor development (11.4 vs 18.1 weeks) and reduced tumor size displayed by xeno-transplanted RD cells pre-cultured in the presence of the MEK/ERK inhibitor. It is noteworthy that our data showing that rhabdospheres express high ALDH activity, which is further increased by U0126 treatment, are only partially in agreement with those of other authors [
15], who reported that ALDH1 is a marker of cancer stem cells in ERMS. This suggests that ALDH activity is sensitive to induced signaling. Indeed, U0126-treated cells that strongly express ALDH do not appear to play a role in early tumor initiation and the development of tumor masses, which would be expected to be larger than those of untreated rhabdospheres. The fact that tumor development is delayed and the tumors themselves are smaller may mean that the subpopulation that expresses a high degree of ALDH activity does not contribute to the tumorigenicity of cancer stem-like cells if the active ERK pathway is absent but undertakes the myogenic precursor program. This hypothesis is supported by the low expression level of CD133 and CXCR4 in U0126-pre-treated cells and correlates strongly with the delay in tumor development that occurred without any further U0126 being added. It is noteworthy that, at the end point, the number of CD133 and CXCR4 positive cells in this tumor population was two folds that in xenografts induced by rhabdospheres. This finding appears to be in contrast to the delay in tumor development and warrants further investigation. Moreover, the intraperitoneal U0126 treatment of mice xeno-transplanted with rhabdosphere cells inhibits tumor growth by about 50 %. The responsiveness of rhabdosphere-derived tumors to the MEK/ERK inhibitor in developing xenografts might be consequent to the reduction in the size of the CD133 population even in in vivo conditions. On the basis of all these in vivo data, continuous treatment with the MEK/ERK inhibitor might help to maintain low levels of CD133 and CXCR4 positive cells.
Following Ras activation, the MAPK pathway has been reported to contribute to the invasive potential of cancer cells [
37]. Other authors have demonstrated that increased Caveolin 1 expression enhances ERK pathway activation and potentiates invasiveness of RD cells [
38]. Furthermore, it is worth recalling data from others [
39‐
41] that suggest the role of CD133 and CXCR4 content in sustaining high metastatic capacity in in vitro and in vivo model of some tumor types. However, the metastatic activity of cancer stem cells is a multistep process, that includes the invasiveness, but is not completely performable in vitro. The reduced invasion potential of U0126-treated cells in SC medium compared with rhabdosphere cells is in keeping with the reduced expression level of CXCR4 and CD133 in U0126-treated cells. Our result indicates that in RD cancer like stem cells invasion potential is a property that depends on ERK pathway.
Targeting MEK/ERK pathway to reduce the chemo- and radio-resistant CD133 positive cell population [
13] may have important implications in the treatment of ERMS.
Indeed, the radioresistance phenotype of cancer stem cells might be the cause of cancer relapse [
42,
43]. The reduction of CD133 positive population might have a beneficial effect on radiation efficacy given that in some cases in CD133 positive cells active ERKs is enhanced by radiation [
44].
A combination of radiotherapy and chemotherapy is one approach currently being used to treat rhabdomyosarcoma. Within this context, the rationale underlying the treatment of rhabdospheres using U0126 was the hypothesis that MEK/ERK inhibition enhances radiosensitivity in the presence of an enriched cancer stem-like population. Radiation or U0126 treatment on their own modify the integrity of rhabdospheres, alter the percentage of stem cell markers and reduce the DNA machinery components levels, though radiation alone is less effective. Combined therapy induces a more pronounced dismantling of rhabdospheres and inhibits the expression of stem cell markers and Rad51. Bearing in mind that Rad51 expression is highly sensitive to MEK/ERK inhibitor [
45], the markedly reduced levels of Rad51 and DNA-PKcs observed in U0126-treated rhabdospheres indicate that MEK/ERK inhibition impairs DNA repair mechanisms, thereby rendering RD cells more sensitive to radiation. The ability of MEK/ERK pathways to orchestrate the complex mechanism of survival in tumor cells, including resistance to radiation, is also demonstrated here by the MEK inhibitor-mediated down-regulation of BMX, whose absence is known to relieve cells from the negative regulation of apoptosis [
22]. Therefore, the MEK/ERK-dependent inhibition of BMX expression [
23] may be involved in the enhanced sensitivity of RD cells to radiation.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CC and FV concepted and designed the study, analyzed data, and wrote the manuscript; LM performed experiments, analyzed data, and prepared figures; GG performed statistical analysis, reviewed the data, and revised the manuscript; FM performed immunoblotting experiments, and analyzed data; GM performed experiments and analyzed data on the animals; AG performed experiments, analyzed data and revised the manuscript; VT designed radiotherapy experiments, and revised critically the manuscript; HJH concepted the study, reviewed the data, and revised the manuscript; VDP performed and evaluated the invasion assay; BMZ concepted and supervised the study, and wrote and revised the manuscript. All authors read and approved the manuscript.