1 Introduction
Anaplastic lymphoma kinase (ALK) is a glycosylated, single-chain trans-membrane receptor tyrosine kinase (RTK) of the insulin receptor (InsR) superfamily. ALK was initially identified as part of an oncogenic fusion protein in anaplastic large-cell lymphoma, and it has since been shown to be involved in the embryonic nervous system and in tumor development [
1,
2]. However, the exact role of ALK in tumor initiation and progression remains elusive. While specific ligands are thought to bind and activate ALK, genetic alterations and protein overexpression have also been shown to result in ALK activation in a ligand-independent manner. Downstream signaling activates multiple pathways involved in cell proliferation, survival, and cell cycle progression, including PI3K/Akt, MEK/ERK, and the JAK-STAT/Cyclin D2 pathways [
3‐
5]. ALK expression has been implicated in several malignancies, including rhabdomyosarcoma (RMS), the most common pediatric soft tissue sarcoma (STS).
RMS is a highly malignant STS seen mostly in children, adolescents, and young adults [
6]. RMS has multiple histologically distinct subtypes, of which embryonal RMS (ERMS) and alveolar RMS (ARMS) are the most common among pediatric patients. ERMS is typically observed in young children and has a more favorable outcome compared to ARMS. ERMS frequently shows a gain of chromosomes 2, 8, and 12, loss of heterozygosity at 11p, and a higher mutational load. ARMS has a higher metastatic rate, can present itself throughout childhood and adolescence, and in the majority of cases a characteristic fusion of
PAX3 or
PAX7 on chromosome 2 to
FOXO1 on chromosome 13 occurs, resulting in the oncogenic fusion-protein PAX3/7-FOXO1 [
7‐
9]. A minority of ARMS lack the fusion, presenting with clinical and biological characteristics more resembling ERMS [
10]. Despite intensive therapy, the survival of metastatic patients remains poor. In addition, RMS survivors can suffer from immediate and late treatment-associated toxicities, underlining the need for new, targeted treatment strategies [
7,
11,
12].
Pillay et al. were among the first to assess the expression of ALK in RMS. ALK expression was seen in 23% of cases with the highest expression in ARMS (45%), compared to ERMS (15%) and other RMS subtypes (8%) [
13]. ALK expression was later associated with
ALK copy number gain, metastatic disease at diagnosis, and in some studies with a worse overall survival (OS) [
14‐
19]. Despite the
ALK copy number gain in ALK-positive tumors, true
ALK amplification was observed in only a small group of RMS patients [
14,
19]. In addition, until recently, no
ALK rearrangements or activating mutations were detected in RMS patients, and even now only one case of ERMS was shown to have an EML4-ALK fusion [
18]. In line with these findings, full-length ALK was observed in RMS tissue samples.
ALK exon deletions resulting in an immature ALK form have been reported. However, the influence of exon deletions on ALK functioning was not further examined [
14,
16].
A possible explanation for the higher expression of ALK in ARMS samples could be the high binding affinity of PAX3-FOXO1 for the 3rd intron of ALK and its influence on ALK transcription [
20]. Some controversy on the correlation of ALK expression with fusion status exists, and several studies could not find a difference in specific ALK gain or ALK expression between fusion-positive and fusion-negative RMS [
19,
14,
15]. Furthermore, specific ALK copy number gain could not be correlated with ALK expression or RMS subtype, and it was even suggested that the high expression levels might be the result of the high proliferative rate of the cells [
14,
15,
21]. In line with this hypothesis, no endogenous phosphorylation of ALK has been observed in RMS cell lines and activation of ALK could be induced only by the addition of phosphatase inhibitors or agonistic antibodies [
16,
22,
23], making a prominent role for ALK in RMS cell growth questionable. Silencing of
ALK did not influence RMS growth [
18,
24], whereas treatment with ALK inhibitors led to a reduction in cell proliferation and viability in several RMS cell lines [
25,
26]. The latter may be the result of multi-kinase activity of the inhibitors [
22,
24]. Nevertheless, independent of ALK activity, ALK might play a role in metastatic disease, since silencing of ALK led to reduced invasion of RMS cells in vitro, even in a non-phosphorylated state [
18].
The majority of clinical trials testing both FDA approved and experimental ALK inhibitors focus on tumor types harboring ALK aberrations. Studies with specific inclusion of ARMS or ERMS are scarce. To the best of our knowledge, only three ongoing studies have specified the inclusion of metastatic ARMS (NCT01548926), ERMS and ARMS patients (NCT02034981) or ALK-positive RMS patients (NCT01742286). Preliminary data of these trials have, however, not shown any response in the included RMS patients (Table
S1), which indicates that targeting ALK in RMS does not seem to be very promising in the clinic.
Current studies examining the preclinical effects of ALK in RMS focused on the use of the first-generation ALK inhibitors, which were associated with a more multi-kinase targeting profile [
22,
24]. In addition, so far these agents were tested solely in vitro. The second-generation ALK inhibitor ceritinib has clinical activity in patients with ALK-positive, metastatic non-small cell lung cancer (NSCLC) and was capable of overcoming resistance to the first-generation ALK inhibitor crizotinib [
27,
28]. In order to further elucidate the potential of ALK inhibition in RMS and to determine the efficacy of second-generation ALK inhibitors in these tumors, the effects of ceritinib were determined in ARMS and ERMS models in vitro and on growth of ARMS models in vivo. In addition, we studied whether combined treatment of ceritinib and dasatinib was capable of enhancing the monotherapeutic effects of ALK inhibition in RMS.
2 Materials and Methods
2.1 Reagents
The ALK inhibitor ceritinib (LDK378) was provided by Novartis (Basel, Switzerland). The multi-tyrosine kinase inhibitor dasatinib (Abl/Src family kinases) was purchased from SelleckChem (Munich, Germany). Compounds were diluted in DMSO or 0.5% HPMC/0.5% Tween80 for in vitro and in vivo experiments, respectively.
2.2 Cell Lines and Cell Culture
PAX3-FOXO1-positive ARMS (Rh30, Rh41) and -negative ERMS (Rh18 and RD) cell lines were generously provided by Dr. Peter Houghton. The Aska cell line was generously provided by Dr. Kazuyuki Itoh. Both ARMS cell lines were cultured in RPMI 1640 medium, RD and Aska cells in DMEM medium, and Rh18 cells in McCoy’s 5A medium (all Lonza, Breda, The Netherlands). Media were supplemented with 10% fetal bovine serum (Gibco, ThermoFisher, Breda, The Netherlands) and 1% Penicillin/Streptomycin (Lonza). Cells were cultured in a humidified atmosphere of 5% CO2/95% air at 37 °C.
2.3 Cell Viability Assay
Cell viability was assessed in MTT assays as previously described [
29]. RMS cells were seeded at 5000 cells (RD, Rh30, and Rh41), 3000 cells (Rh18), or 10,000 cells (Aska) in 100 μl/well. Experiments were repeated in triplicate and IC
50-values were calculated with GraphPad Prism Version 5.03 software.
2.4 Wound Healing Assay and Cell Cycle Analysis
Wound healing assays and cell cycle analysis were performed as previously described [
29,
30]. For the wound healing assays, RMS cells were seeded at 0.6–1.5 × 10
6 cells/2 ml/well, and the cell cycle analysis was performed on a BD FACSCalibur (Franklin Lakes, NJ, USA) flow cytometer.
2.5 PathScan Analysis
The human PathScan RTK signaling antibody array kit (Cell Signaling Technology, cat.#7949, Danvers, MA, USA) was used according to the manufacturer’s protocol. Lysates of cells treated with or without ceritinib (2.5 μM) for 24 h were used. Images were analyzed with the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA) and Odyssey Application Software (version 3.0.30). Kinase phosphorylation was calculated with the following equation:\( {\left(\mathrm{Et},1\; or\;\mathrm{Et},2/\overline{x}\mathrm{Epc}\right)}^{\ast }100\% \). Both spots on the array of a particular target (Et,1 and Et,2) were individually divided by the average of all positive control spots on the array (\( \overline{x} \)Epc) and representative percentages were calculated. Phosphorylation is presented as relative phosphorylation.
2.6 Western Blot
Western Blot analysis was performed as previously described [
29]. A detailed list of monoclonal antibodies used in this study can be found in the
electronic supplementary material. The ALK-rearranged synovial sarcoma cell line Aska was used a positive control for ALK and pALK analysis [
23]. Images were analyzed with the Odyssey Infrared Imaging System and Odyssey Application Software.
2.7 siRNA Experiments
Cells were transfected with 25 nM of a scrambled control siRNA, or specific anti-ALK or anti-IGF1R siRNA (Dharmacon, Lafayette, CO, USA) using Lipofectamine 2000 transfection reagents and opti-MEM medium according to the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA). Following transfection, the cells were either lysed for Western Blot analysis or cell viability was determined by MTT assay. Cell viability experiments were performed in triplicate and the p-value was calculated (Student’s t-test) with GraphPad Prism Version 5.03 software. *p < 0.05, **p < 0.01, ***p < 0.001.
2.8 In Vivo Experiments
Female Balbc/nude mice (6–8 weeks) were subcutaneously implanted with Rh41 RMS xenografts, and experiments were started at a tumor size of 0.1–0.6cm3 (1–2 weeks after implantation). Ceritinib was administered daily at 25 mg/kg or 50 mg/kg by oral gavage. Each therapy group (n = 7) was treated for 4 weeks. Tumor growth was monitored by caliper measurements in three dimensions (length (l), width (w) and height (h); all maximum diameter) twice weekly. Tumor size was calculated using the formula:4/3π x l/2 x w/2 x h/2. Mice were euthanized on day 28. If treatment was scheduled on day 28, mice were euthanized 2–4 h post injections. Tumor sizes are depicted as relative tumor volume (RTV = tumor volume at any time (Vt)/ tumor volume at t = 0 (Vt0)). All applicable international, national, and institutional guidelines for the care and use of animals have been followed. All procedures performed in studies involving animals were in accordance with the ethical standards of the animal ethical committee of the Radboud University, Nijmegen, The Netherlands.
2.9 Immunohistochemistry (IHC)
IHC of tumor xenografts was performed to evaluate vascularization (CD34), proliferation (Ki67), apoptosis (Caspase-3), and effects on the PI3K/Akt and MEK/ERK pathways (pAkt/pERK). A detailed description of the methodology and analysis can be found in the
supplementary material.
2.10 Combination Indices
Drug synergy of combined ceritinib and dasatinib treatment was assessed by calculation of the combination index (CI) and dose reduction index (DRI). The CI and DRI were calculated with CompuSyn software according to the manufacturer’s recommendation [
31,
32]. Concentrations equal to 1/4xIC50 - 4xIC50-values of ceritinib and dasatinib were combined in a constant ratio. The monotherapy and combination therapy effects on cell viability were determined in three independent experiments and an average fraction of cell viability affected (FA)-value was used for further calculations. Synergism of the combination treatment is represented in an isobologram and Table
2. The line represents an additive effect (CI = 1) of the combination at the given FA-value. Points below or above the line represent synergism (CI < 1.0) and antagonism (CI > 1.0), respectively. Points on the x-axis (ceritinib) and y-axis (dasatinib) represent the dose of the monotherapy necessary to generate the FA-value. DRI values >1 represents a favorable dose reduction in the combination treatment compared to the monotherapy dose.
4 Discussion
In this study, we aimed to elucidate the mechanism of action and the potential of ALK inhibition in RMS. The underlying reason for this is that although several clinical studies are evaluating the effects of ALK-targeting in RMS patients, the exact role of ALK in these tumors remains questionable. No effects have been reported in RMS in the clinic so far, urging the need to critically re-evaluate the potential of ALK as a target for RMS therapy. Previous research into the role of ALK in RMS has led to the discovery of full-length ALK in RMS, with a predilection for the ARMS phenotype. Accordingly, a similar preference for ARMS was observed in our study with higher sensitivity to ALK-targeted therapy. However, instead of the expected decrease in ALK activity upon treatment, nearly undetectable endogenous levels of pALK were observed. This lack of intrinsic ALK activity was previously observed in ARMS cell lines [
22,
23]. Peron et al. further showed that the activation of ALK is dependent on the induction of autophosphorylation by agonistic antibodies and the activity of cytoplasmic phosphatases [
22]. However, the cytotoxicity seen upon treatment does suggest an alternative target of ceritinib may be present in ARMS and ERMS cell lines.
In line with previous findings [
33], ceritinib led to a decrease in IGF1R activity in all cell lines. Cellular signaling activity was decreased by inhibition of the PI3K/Akt/S6 pathway with the exception of one ERMS cell line. In this cell line a decrease in pAkt was accompanied by an increase in S6 activity. pS6 is regulated by both the PI3K/Akt and the MEK/ERK pathways [
34,
35]. The ERMS cell lines showed an increased ERK activity post ceritinib treatment, possibly explaining the observed increase in pS6. The role of IGF1R in the mechanism of action of ceritinib was further investigated by silencing of IGF1R and ALK. IGF1R knockdown reduced viability in the RMS cell lines, whereas knockdown of ALK had no effect. Of note, the Aska cell line with constitutively active ALK did show reduced viability following ALK siRNA treatment. This further shows that the reduction in cell viability in RMS cell lines upon ceritinib treatment is likely dependent on decreased IGF1R activity, rather than ALK activity.
Further examination of the mechanism of action of ceritinib showed a clear distinction between ARMS and ERMS. Ceritinib treatment predominantly increased RTK phosphorylation in ERMS compared to decreased RTK phosphorylation in ARMS. In addition to the increased phosphorylation, the lack of cell cycle arrest and apoptosis strengthens the idea that upon ceritinib treatment, ERMS cells increase phosphorylation of compensatory pathways to prevent cell death. ARMS, on the other hand, showed a clear induction of apoptosis-associated PARP cleavage and a cell cycle arrest in one cell line (Rh30). No cell cycle arrest was observed in Rh41. This difference might be explained by the p53-status of the ARMS cell lines. Rh41 harbors a p53 (c. 700_712del) mutation, while Rh30 is p53 wild-type. p53 is the main regulator of the G1 checkpoint, and loss of p53 leads to the loss of the G0/G1 checkpoint [
36]. Hypothetically, the altered oncogenic signaling caused by ceritinib can lead to cellular stress and induce the expression of p53 in Rh30, resulting in a G0/G1 cell cycle arrest and induction of apoptosis [
36,
37]. In Rh41, a G0/G1 arrest cannot be induced and the cell cycle is continued. However, upon cellular stress the lack of p53 can lead to premature exit of the cells from the G2 to M phase. This premature exit of the cell might prevent proper mitosis and lead to the induction of cell death [
36,
38].
Based on the current results, it can be said that ceritinib monotherapy has a greater effect on ARMS tumor cell growth compared to ERMS. In vivo, ARMS xenograft growth was inhibited by ceritinib treatment with a significant increase in caspase-3 levels. The detected increase of caspase-3, in addition to the in vitro observed increased PARP cleavage, suggest that the effect of ceritinib on cell viability is mediated by caspase-dependent apoptosis [
39]. Due to large variations in tumor volume in all treatment groups, the mechanism of action of ceritinib in vivo was difficult to examine. Interestingly, when only the outliers of the control and 50 mg/kg treatment groups were further evaluated, a significant decrease in proliferation and a trend towards higher caspase-3 levels was observed, indicating that ceritinib treatment in vivo can reduce tumor growth in a manner similar to that established in vitro. However, no complete tumor loss was observed upon treatment, making monotherapy with ceritinib an unlikely treatment option for ARMS.
In order to determine whether combination treatment could enhance the anti-tumor effect of ceritinib, combined treatment with the Abl/Src inhibitor dasatinib was examined for both ARMS and ERMS in vitro. Src is known to modulate the activity of multiple signaling pathways by interaction with both receptors at the cell membrane and downstream signaling proteins [
40]. Even though it was one of the first identified tyrosine kinases, the exact role of Src in oncogenesis is still not clear [
41]. Src activity is mediated via the phosphorylation of multiple tyrosine kinase domains with opposing functions. Phosphorylation at Tyr527/Tyr530 leaves Src in an inactive conformation, while dephosphorylation of this domain leads to conformational changes and allows the autophosphorylation at Tyr416/Tyr419 and the binding and phosphorylation of Src substrates [
41]. Since all cell lines showed high levels of pSrc (Tyr416), even following ceritinib treatment, and Src is known to be of importance in sarcoma and was previously shown to increase following IGF1R treatment in RMS, the combination of ceritinib with dasatinib was considered of interest [
42,
43]. The synergistic effects and favorable dose reduction indices suggest the combination of dasatinib and ceritinib may be favorable in both ARMS and ERMS cell lines.
The current study contributes to the hesitations concerning ALK targeting in RMS. Despite the clear distinction between ARMS and ERMS in sensitivity to ceritinib and the in vivo decrease in tumor growth, the lack of intrinsic ALK activity and the lack of cell growth reduction following ALK knockdown cast doubt on the role of ALK in RMS tumorigenesis. Moreover, ceritinib was shown to inhibit IGF1R activity. Unfortunately, clinical trials testing IGF1R inhibitors in RMS have shown limited results as well, making a monotherapeutic treatment with either ceritinib or more specific IGF1R inhibitors undesirable [
44,
45]. However, the synergistic effect of the in vitro combination of ceritinib and the Abl/Src kinase family inhibitor dasatinib does underline the potential of combined IGF1R and Src targeted therapy in both ARMS and ERMS.