Imatinib mesylate potentiates topotecan antitumor activity in rhabdomyosarcoma preclinical models

Abstract

High levels of PDGFR expression in primary rhabdomyosarcoma (RMS) have been associated with disease progression. To date however, there are no reports on the activity of imatinib mesylate, a selective PDGFR inhibitor, in RMS preclinical models. A panel of 5 RMS cell lines was used to investigate the expression of PDGFRα and PDGFRβ, c-Kit and the multidrug transporter ABCG2 (also inhibited by imatinib). In vitro and in vivo experiments were performed using RD (embryonal) and RH30 (alveolar) cell lines to determine the efficacy of imatinib as single agent and in combination with topotecan (TPT). PDGFRβ was significantly expressed in all cell lines, with the highest levels in RD, while PDGFR α and ABCG2 were significantly expressed only in RH30 and RMZ-RC2. c-Kit was not detected. PDGFRβ signaling was active in RD but not in RH30, whilst PDGFRα signaling was not active in either cell lines. Significant ABCG2-mediated extrusion of Hoechst 33342 was demonstrated in RH30 but not in RD, and was inhibited by imatinib and the specific ABCG2 inhibitor Ko143. In vitro, imatinib was not active as a single agent at therapeutic concentrations, but significantly potentiated TPT antitumor activity in both cell lines. In vivo experiments using tumor xenografts confirmed the synergistic interaction in both cell lines. These results suggest that at least 2 different mechanisms—inhibition of ABCG2 and/or PDGFRβ—are involved in the synergistic interaction between imatinib and TPT, and support the use of this combination for the treatment of high-risk RMS patients. © 2006 Wiley-Liss, Inc.

Imatinib mesylate (STI571, Glivec®, Novartis Pharma AG, Basel, Switzerland) is an effective molecular targeting agent that, by reversibly competing with ATP for binding to the kinase domain, selectively inhibits at submicromolar concentrations several receptor and nonreceptor protein tyrosine kinases (PTKs): PDGFR α and β, c-Kit, c-Fms, Bcr-Abl, c-Abl and Arg.1, 2, 3 As such, imatinib mesylate has shown a potent therapeutic activity in malignancies where a PTK is constitutively activated as a result of either a reciprocal translocation, as for Bcr-Abl in chronic myeloid leukemia,4 or a gain-of function mutation, as for c-Kit or PDGFRα in gastrointestinal stromal tumor (GIST).5 However, increasing evidence suggests that imatinib mesylate can also be active, although to a lesser extent, in tumors in which PDGFRs and/or c-Kit are activated by autocrine or paracrine ligand-dependent stimulation, despite the absence of activating mutations.6, 7, 8, 9, 10, 11, 12

Recently, imatinib mesylate was also reported to selectively inhibit ABCG2 (also known as breast cancer resistance protein, BCRP), another ATP-dependent and plasma membrane-associated molecule, which belongs to the ATP binding cassette (ABC) transporter superfamily.13, 14 This glycoprotein confers clinical drug-resistance by actively extruding from tumor cells a variety of therapeutic compounds including methotrexate, mitoxantrone and topoisomerase I inhibitors belonging to both camptothecins-topotecan [TPT], irinotecan [CPT-11] and SN-38, the active metabolite of CPT-11- and indolocarbazoles.15 Therefore, when imatinib mesylate is combined with ABCG2-transported cytotoxic drugs, the possible role of ABCG2 should always be taken into account and investigated.

Metastatic rhabdomyosarcoma (RMS) remains one of the clinical challenges in pediatric practice with a 3-year overall survival (OS) at best of 55%.16 Additionally, the identification of poor-risk nonmetastatic patients is based on a multifactorial process whereby age at diagnosis, tumor size, primary site, stage and histology all contribute to clinical staging and grouping.17 However, a major influence upon treatment decision is still the presence of alveolar histology, which is associated with PAX3-FKHR or PAX7-FKHR gene fusions resulting from t(2;13)(q35;q14) or t(1;13) (p36;q14) translocations, respectively, and carries a poorer prognosis.18 This has highlighted the importance of molecular characterization in that fusion status is not only diagnostically helpful but is also associated with different clinical presentation and outcome among children with metastatic disease.19, 20 Administration of high-dose chemotherapy with hematopietic rescue has failed to achieve substantial improvements in survival of patients with metastatic RMS,21 [McDowell unpublished data]. Therefore, the search for alternative therapeutic targets and the development of novel therapeutic approaches are needed for the treatment of this subset of patients.

Expression of mRNA for PDGFRα and PDGFRβ has recently been reported in RMS primary tumors, with high levels of expression being associated with disease progression.22 Expression of c-Kit protein has also been described in RMS primary tumors.23 This pattern of PTK expression may thus provide a suitable preclinical setting in which to assess the activity of imatinib mesylate as single agent and in combination with TPT, the latter drug chosen because of its therapeutic activity in children with advanced RMS.24

Material and methods

Cell lines, culture conditions and drug formulation

RMS embryonal cell lines RD, RD18, CCA and alveolar cell lines RH30 and RMZ-RC2 were kindly provided by P.L. Lollini and have previously been described.25, 26, 27, 28, 29 All cell lines were maintained in DMEM medium containing 10% fetal calf serum (FCS) (Sigma, Dorset, UK), 1% L-glutamine and 1% streptomycin/penicillin. CCA and RMZ-RC2 cell lines were incubated in 7% CO₂, whilst the remaining cell lines in 5% CO₂. Imatinib mesylate for in vitro and in vivo studies was prepared by diluting with DMSO to make a stock solution of 10 mM. This was divided into aliquots and frozen at −20°C immediately. A stock solution of 10 mM TPT (Hycamtin®, GlaxoSmithKline, Crawley, UK) in sterile water was also prepared and stored in a similar manner. Individual aliquots were defrosted as required and the drugs were used immediately.

RNA extraction and real-time RT-PCR analysis

Total RNA from each cell line was isolated by TRIzol extraction reagent (Invitrogen Corporation, Carlsbad, CA) according to the manufacturer's instructions. Carryover DNA contamination was eliminated by treatment of total RNA with DNA-free kit (Ambion, Austin, TX) according to the manufacturer's instructions. Total RNA (0.5 μg) from each cell line was reverse transcribed in cDNA using the Retroscript kit (Ambion). Quantitative real-time PCR was carried out to detect β-actin expression that was utilized to normalize the amount of cDNA for each sample. β-actin primers were 5′CCTTCAACACCCCAGCCA3′ and 5′ACCCCTCGTAGATGGGCAC3′. Equal amounts of cDNA from each sample were amplified using the following primers to detect: PDGFRα—5′GGAGGATGATGATTCTGCCATT3′ and 5′TGTCGTAGGAGGCAGGTACCA3′; PDGFRβ—5′CTGGGCTAGACACGGGAGAA3′ and 5′ GGTGGGATCTGGCACAAAGA3′; c-kit—5′ATTTTCTCTGCGTTCTGCTCCTAC3′ and 5′CGCCCACGCGGACTATTA3′; ABC-G2—5′TGTCACAAGGAAACACCAATGG3′ and 5′CTTAACACAGCTCCTTCAGTAAATGC3′. Reaction linearity was checked by running serial dilutions of cDNA from RH30 (for PDGFRα and ABCG2), RD (for PDGFRβ) and HTLA230 human NB cell line30 (for c-kit) taken as positive controls. Two independent experiments were carried out in triplicate using the ABI Prism 7000 cycler (Applied Biosystems, Foster City, CA) with the Sybr®green fluorochrome. To evaluate whether the set of primers utilized in each reaction gave rise to an unique amplification product, the amplified products were run on an agarose gel to check the size of the resulting bands. In addition, during the quantitative real-time PCR, we ran the dissociation protocol associated to the ABI Prism 7000 cycler (Applied Biosystems). Results were analyzed by the ABI Prism 7000 SDS software.

Flow cytometry analysis

Flow cytometry analysis was carried out in all 5 cell lines. Cells were washed, suspended in phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin and incubated with phycoerytrin (PE)-conjugated monoclonal antibodies (mAbs) for 30 min at 4°C in the dark. The following mAbs were used: anti-PDGFRα (Becton Dickinson Pharmingen, San Diego, CA), anti-PDGFRβ (Becton Dickinson Pharmingen), anti-c-Kit (DakoCytomation, Milan, Italy) and anti-ABCG2 (clone 5D3) (Becton Dickinson Pharmingen). Negative controls were isotype-matched irrelevant mAbs. Two independent series of experiments, each in triplicate, were performed and in each experiment data from a minimum of 20,000 cells were acquired using a FACScan (Becton Dickinson) flow cytometer. Data analysis was performed by Cell Quest software (Becton Dickinson). Cells were electronically gated according to light scatter proprieties to exclude cell debris.

In vitro growth inhibition analysis

Cell growth studies were carried out in RD and RH30 cell lines following harvesting of cells from culture when in a logarithmic phase of growth. Aliquots of 4 × 10³ RD and 11 × 10³ RH30 cells were plated in 96-well plates (Cellstar, Greiner, Germany). After 24 hr, imatinib mesylate, TPT or control medium were added separately to the wells. Serial dilutions were made to span the range from ineffective concentrations to maximally effective concentrations. After cells were incubated for 72 hr with the drug indicated, cytotoxic effects were assessed using a standard 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) assay.31 Briefly, 150 μg of MTT were added to each well followed by 3-hr incubation; the culture supernatant was then removed, and cells were dissolved and mixed with 200 μl DMSO for each well. After thorough formazan solubilization, the absorbance of each well was measured at 540 nm (reference wavelength 690 nm) by using a microplate spectrophotometer (Victor 3°, Perkin-Elmer, MA). Growth inhibition was expressed as the ratio of the mean absorbance of treated cells to that of control cells. Two independent series of experiments, each in triplicate, were performed and the growth-inhibition rate was calculated as 30–90% inhibitory concentrations (ICs), using linear-log interpolation of the concentration-effect relations.

The pharmacological interaction between imatinib mesylate and TPT was evaluated using the CalcuSyn for Windows software package (Biosoft, Cambridge, UK) that is based on the Chou-Talalay median-effect equation method.32 This method defines the expected additive effect of 2 (or more) agents and quantifies synergism or antagonism by determining how much the combination effect differs from the expected additive effect. The type of interaction is expressed as combination index (CI) and categorized as synergistic, additive or antagonistic depending whether a CI value lower than, equal to or greater than 1 is found. Two further independent series of experiments, each in triplicate, using MTT assay were performed using a constant ratio (25:1) of each drug concentration in escalating doses, starting from the doses of 5.0:0.2 μM imatinib:TPT.

Protein analysis

Proteins were extracted from all 5 cell lines as previously described.33 Briefly, cells were lysed on ice in 50 mM Tris HCl, pH 7.4, 5 mM EDTA, 250 mM NaCl, 50 mM NaF, 0.1% Triton X-100, 0.1 mM Na₃VO₄, 1 mM phenylmethylsulfonylfluoride, 10 μg/ml leupeptin and 10 μg/ml pepstatin. Lysate was then centrifuged at 14,000g at 4°C for 10 min, the supernatant was collected and protein concentration was determined using a colorimetric assay (BioRad, Hercules, CA). Protein separation on SDS polyacrylamide gels and western blot analysis were carried out as previously described.33 Specific antibodies were anti-PDGFRα antibody (Upstate, Lake Placid, NY), anti-PDGFRβ antibody (Upstate), anti-phospho-tyrosine (Becton Dickinson Biosciences, Franklin Lakes, NJ) and anti-HSP70 (StressGen, San Diego, CA). Detection was performed by ECL Plus kit (Amersham Pharmacia, Piscataway, NJ) according to the manufacturer's instructions.

ABCG2-mediated extrusion of Hoechst 33342 analysis

Hoechst 33342 only becomes fluorescent when complexed with DNA and the rate of increase in cell fluorescence is directly related to the dye influx into cells. Since ABCG2 actively extrudes Hoechst 33342, the dye accumulation rate is slower in ABCG2-positive cells than in control negative cells and can be accelerated by incubation with ABCG2 inhibitors.34 Two independent experiments were carried out in triplicate according to a method previously described.14 Briefly, RD and RH30 cells were preincubated at 37°C in HPMI medium (120 mM NaCl, 5 mM KCl, 400 μM MgCl₂, 40 μM CaCl₂, 10 mM Hepes, 10 mM NaHCO₃, 10 mM glucose and 5 mM Na₂HPO₄) for 5 min, and then incubated with 1.0 μM Hoechst 33342 for 10 min. Imatinib mesylate at various concentrations (0.1–10.0 μM) was subsequently added, and the accumulation rate of Hoechst 33342 was measured for a further 10 min. The ABCG2 inhibition induced by imatinib mesylate was compared to that induced by the specific ABCG2 inhibitor Ko143 (generous gift from Alfred H. Schinkel, The Netherlands Cancer Institute, Amsterdam)35 when added at various concentrations (0.001–1.0 μM). Maximum Hoechst 33342 binding was obtained by the addition of digitonin 8.0 μM. Cell fluorescence induced by Hoechst 33342 was measured by using a fluorescence spectrophotometer (Applied Biosystems) at 350 nm (excitation)/460 nm (emission).

In vivo growth inhibition analysis

The separate effects of imatinib mesylate, TPT and a combination of these 2 agents were explored in RD and RH30 cell lines growing as xenografts in nude mice CD/1 (nu/nu). Cells (5 × 10⁶) were injected subcutaneously in both flanks and treatment was started when tumors were palpable. A series of preliminary experiments was performed to optimize the doses of imatinib mesylate (100 and 200 mg/kg/day) and TPT (0.5, 1.0 and 1.5 mg/kg/day) when administered in combination by oral gavage. As a result of the high toxic death rate observed at the higher doses (data not shown), in all the subsequent experiments, imatinib mesylate was administered p.o. at 100 mg/kg/day in 2 divided doses each day and TPT at 0.5 mg/kg/day for 5 consecutive days for 3 consecutive weeks. Each drug-treated group and control group (only treated with carrier p.o.) consisted of 10 mice. Tumor weight (TW) and total weight were monitored every 3 days. Tumor weight was calculated by the formula: Tumor weight (TW) = d² × D/2 (where d and D represent the shortest and the longest diameter, respectively). Antitumor activity was evaluated according to 2 criteria: (i) Tumor weight inhibition (TWI) in treated (T) vs. control (C) mice according to the formula: 100 − (T/C × 100) and (ii) Log₁₀ cell kill (LCK) according to the formula: (T − C)/3.32 × DT (where T is the mean time [days] required for treated tumors and C for the control to reach an established weight, and DT is the mean doubling time of control tumors). Mice were observed for 3 weeks after the treatment stopped to evaluate time to progression. Wilcoxon exact test (2-tailed) was applied to compare in vivo tumor growth between treated and control groups using GraphPad Prism 4 for Windows (GraphPad Software, San Diego, CA). The level of significance was set at p < 0.05. Xenograft studies were approved by the “Istituto Superiore di Sanità” (National Institute of Health) and the Ethical Committee of Catholic University, and all animal care were in accordance with local institutional guidelines.

Results

RNA expression for c-Kit, PDGFRα and PDGFRβ and ABCG2

Results obtained from quantitative (real-time) RT-PCR are listed in Table I. Significant levels of PDGFRβ were found in CCA, RMZ-RC2 and RD, with the highest levels being detected in the embryonal RD cell line. Significant levels of PDGFRα mRNA were only detected in alveolar RH30 and RMZ-RC2 cell lines, whilst only low levels were demonstrated in CCA. No significant levels of c-kit mRNA expression were detected. Expression of ABCG2 mRNA was rather similar to that of PDGFR alpha, with the highest levels being found in RH30 and RMZ-RC2, whilst lower levels were found in RD18 among the embryonal cell lines.

Table I. PDGFR Alpha, PDGFR Beta, C-Kit and ABCG2 mRNA Expression in the RMS Cell Lines as Determined by Real-Time RT-PCR

Genes	Cell lines
	RD18	RH30	CCA	RMZ-RC2	RD	HTLA-230
The relative expression level of each gene was calculated by comparing the values of each cell line with that of the cell line used to build the calibration curve taken as 100 (see Material and methods section). The c-kit-expressing human NB cell line HTLA-230 was only used for the calibration curve to determine the expression of c-kit. Values are reported ± standard deviation. Abbreviations: n.d., not detected.
PDGFR α	0.14 ± 0.04	100.00 ± 0.00	24.25 ± 0.16	73.11 ± 7.79	2.25 ± 0.70
PDGFR β	40.44 ± 0.98	47.31 ± 0.34	66.52 ± 1.92	75.78 ± 7.85	100.00 ± 0.00
c-kit	n.d.	n.d.	7.91 ± 3.38	7.50 ± 3.33	n.d.	100.00 ± 0.00
ABCG2	23.51 ± 6.27	100.00 ± 1.04	3.30 ± 0.94	79.47 ± 30.97	4.06 ± 0.51

Protein expression for PDGFRα and PDGFRβ, c-Kit and ABCG2

Flow cytometry was utilized to confirm the presence of the receptor PTKs and ABCG2 in the panel of RMS cell lines (Fig. 1). Expression of PDGFRβ on cell surface was detected in all cell lines and was rather homogeneous within each cell line but considerably heterogeneous among the different cell lines, paralleling the levels of mRNA expression as measured by real-time RT-PCR. Expression of PDGFRα and ABCG2 on the cell surface also varied among the 5 cell lines, with 1/5 and 2/5 lines expressing significant levels of the protein, respectively. Expression of c-Kit was very low in all 5 cell lines.

Figure 1. Expression of PDGFRα, PDGFRβ, c-Kit and ABCG2 proteins in the RMS cell lines as assessed by flow cytometry. Cells were stained with PE-conjugated mAbs against PDGFR alpha, PDGFRβ, c-Kit and ABCG2 and analyzed by flow cytometry. Two independent experiments were performed in triplicate as described in Material and methods section.

In vitro activity of imatinib mesylate and topotecan

Based on the pattern of expression of the target proteins and on the favorable characteristics of in vitro and in vivo cell growth, embryonal RD cell line and alveolar RH30 cell line were chosen for the subsequent experiments.

Initially, the activity on tumor growth of imatinib mesylate and TPT as single agents was evaluated. When the 2 cell lines in the logarithmic phase of growth were treated with imatinib mesylate (0–50.0 μM) or TPT (0–5.0 μM), the proliferation of each cell line was inhibited in a concentration-dependent manner (Table II). MTT assays showed that the imatinib mesylate concentrations inhibiting cell growth by 50% (IC₅₀) after a 72 hr continuous exposure were 22.6 μM for RD and 29.0 μM for RH30; TPT ICs₅₀ were 0.52 μM for RD and 1.32 μM for RH30.

Table II. In Vitro Activity of Imatinib Mesylate and TPT as Single Agents as Assessed by MTT Assay in RD and RH30 Cell Lines

Cell line	Inhibitory concentrations (ICs)
	Imatinib mesylate (μM)				TPT (μM)
	IC30	IC50	IC70	IC90	IC30	IC50	IC70	IC90
RD	15.0 ± 2.5	22.6 ± 2.8	30.3 ± 3.5	41.0 ± 4.2	0.17 ± 0.06	0.52 ± 0.09	0.92 ± 0.27	1.61 ± 0.31
RH30	20.8 ± 3.2	29.0 ± 3.4	35.8 ± 3.7	44.8 ± 4.8	0.67 ± 0.18	1.32 ± 0.38	2.50 ± 0.37	3.90 ± 0.42

Subsequently, the activity and pharmacological interactions of imatinib mesylate and TPT in combination were investigated in the same cell lines. In this case, MTT assays were carried out using a constant ratio (25:1) of each drug concentration in escalating doses to determine the CI values (Table III). The median-effect plots for the 2 cell lines based on CalcuSyn analysis of cytotoxicity data are shown in Figures 2a and 2b. Mean CI values for IC₅₀, IC₇₀ and IC₉₀ demonstrated a synergistic interaction ranging from a moderate synergism (CI: 0.7–0.85) at all the ICs tested for RH30 and at IC₅₀ for RD, to a stronger synergism (CI: 0.3–0.7) at IC₇₀ and IC₉₀ for RD.

Figure 2. In vitro activity of imatinib mesylate and TPT in combination. Median-effect plot of cytotoxicity data developed by CalcuSyn software for RD (a) and RH30 (b) cell lines. The drugs were combined at a fixed 25:1 ratio. Each plot is representative of 2 independent experiments in triplicate—Fa, affected fraction; Fu, unaffected fraction; D, concentration of drug used.

Table III. In Vitro Activity of Imatinib Mesylate and TPT in Combination as Assessed by MTT Assay in RD and RH30 Cell Lines: Mean Combination Index (CI) Values ± SD at Different IC Values

Cell lines	CI12
	IC50	IC70	IC90
1 CI at constant 25:1 ratio for both cell lines. 2 Moderate synergism: CI 0.7–0.85; synergism: CI 0.3–0.7.
RD	0.74 ± 0.12	0.49 ± 0.03	0.47 ± 0.05
RH30	0.79 ± 0.06	0.74 ± 0.06	0.71 ± 0.05

Phosphorylation status of PDGFRα and PDGFRβ

Because both RD and RH30 cell lines were sensitive to imatinib mesylate and expressed significant levels of PDGFRβ (whilst RH30 also expressed significant levels of PDGFR alpha), the ability of imatinib mesylate to inhibit PDGFRα and PDGFRβ phosphorylation in both cell lines was investigated (Fig. 3). Equivalent basal levels of phosphorylation for PDGFR were detected at very low levels in RD and RH30 cells cultured in 0.1% FCS. In RD cells, treatment with 100 ng/ml recombinant PDGF BB (but not with recombinant PDGF AA at the same concentration) produced a slight increase in PDGFRβ phosphorylation, which was in turn partially inhibited by pretreatment with 15.0 μM imatinib mesylate. In RH30 cells, no increase in either PDGFRα or PDGFRβ was demonstrated after treatment with recombinant PDGF AA or BB, respectively. These findings suggest that PDGFRα signaling is not active in RD and RH30, whilst PDGFRβ signaling is only detectable in RD.

Figure 3. Phosphorylation status of PDGFRα and PDGFRβ in RH30 and RD cell lines. Cells were cultured in serum starvation conditions for 24 hr. After starvation, recombinant PDGF AA or PDGF BB (100 ng/ml) was added for 10 min before harvesting. Some cultures were pretreated with 15.0 μM Glivec for 1 hr before PDGF AA or BB administration. Cell extracts were separated on PAGE and immunoblotting analysis was carried out using antiPDGFR alpha, antiPDGFRβ and antiphosphotyrosine antibodies. Abbreviations: IM, Imatinib mesylate; Un, serum starved control.

Inhibition of ABCG2-mediated extrusion of Hoechst 33342 by imatinib mesylate

In cells expressing ABCG2, Hoechst 33342 accumulation rate directly reflects the activity of ABCG2 protein, and ABCG2 inhibition results in an increase in the dye accumulation rate. In the present study, Hoechst 33342 accumulation was measured directly in intact RD and RH30 cells with and without imatinib mesylate. For comparison, ABCG2 inhibition induced by Ko143 was also determined. Only a marginal inhibition of Hoechst 33342 extrusion was observed in RD cells following incubation with either Ko143 or imatinib mesylate (Fig. 4a), and this was in agreement with the very low level of ABCG2 mRNA and protein expressed by this cell line. On the contrary, a significant inhibition of Hoechst 33342 extrusion was obtained by both compounds in the intensely ABCG2-positive RH30 cells, although in a range of concentrations spanning 4 orders of magnitude (Fig. 4b). Ko143 was already effective at nanomolar concentrations, showing a half-maximal inhibitory activity at concentrations of 0.006 μM. Imatinib mesylate showed inhibitory activity at higher concentrations but still in the range of levels therapeutically achievable in vivo, with a half-maximal inhibitory effect at 1.1 μM.

Figure 4. Inhibition of ABCG2-mediated Hoechst 33342 extrusion induced by Ko143 and imatinib mesylate in RD (a) and RH30 (b) cells. Two independent experiments were performed in triplicate as described in Material and methods section. Standard deviation did not exceed 5% of each value.

In vivo activity of imatinib mesylate and topotecan

The in vivo activity of imatinib mesylate and TPT as single agents and in combination was determined by administering the 2 drugs orally to mice with palpable RD or RH30 xenografts. A series of preliminary experiments (data not shown) demonstrated that the optimal dose of the 2 drugs when administered in combination was 100 mg/kg/day in 2 divided doses each day for imatinib mesylate and 0.50 mg/kg/day for TPT. Consequently, the same doses were used when the 2 drugs were administered as single agents. When the 2 drugs were administered in combination at higher doses (see Material and methods section), a high rate of toxic deaths was observed, with the majority of mice dying on treatment between days 5 and 8, and the remaining mice appearing severely wasted. No gross abnormalities were found on autopsy of these animals. In contrast, no deaths or wasting were observed in control mice treated with carrier only.

Imatinib mesylate given as single agent for 5 consecutive days for 3 consecutive weeks induced a marginal tumor growth inhibition only in RH30 xenograft (p = 0.03), while TPT alone had a marginal effect on tumor growth only in RD xenograft (p = 0.03) (Figs. 5a and 5b). However, the 2 drugs in combination showed a more significant activity than as single agents both in RD (p = 0.03) and RH30 (p = 0.01) (Figs. 5a and 5b). Xenografts treated with the combination grew at a significant slower rate throughout the course of treatment: in RD xenografts TWI (%) was 84 and LCK 0.76; in RH30 xenografts TWI (%) was 90 and LCK 0.9. Toxicity in terms of progressive weight loss and physical activity was not observed in any of the treated mice.

Figure 5. In vivo activity of imatinib mesylate (100 mg/kg/day) and TPT (0.50 mg/kg/day) as single agents and in combination in RD (a) and RH30 (b) xenografts. Experiments were performed in triplicate as described in Material and methods section. Marginal tumor growth inhibition was induced by TPT in RD (p = 0.03) and by imatinib mesylate in RH30 (p = 0.03), whilst the combination showed more significant activity in both xenografts, RD (p = 0.03) and RH30 (p = 0.01). Symbols: ▪ imatinib mesylate; ▴ TPT; ♦ imatinib mesylate + TPT; • control.

Discussion

The increasing knowledge of molecular mechanisms underlying cancer development and progression has led to the identification of several classes of potential targets and to the development of molecules that can selectively modulate these targets, thus allowing novel therapeutic strategies. One such target class is represented by PTKs, critical components of the cellular signaling apparatus and signal transduction pathways that regulate cell proliferation, differentiation, survival and motility by catalyzing protein phospyhorylation and dephosphorylation. Two main types of TKs are known: (i) receptor TKs are transmembrane proteins with a ligand-binding extracellular domain and a catalytic intracellular domain; (ii) nonreceptor TKs lack a transmembrane domain and are found at the inner surface of the cell membrane, in the cytoplasm or in the nucleus.36 In normal cells, the enzymatic activity of both types of TKs is tightly regulated, whilst in tumor cells their constitutive activation caused by gain-of function mutation or fusion with a partner protein, or their abnormal autocrine or paracrine stimulation by the ligand can result in dysregulated cell growth.36, 37 Constitutive activation of receptor TKs has been demonstrated in several solid tumors: c-met mutants have been reported in papillary renal carcinoma,38 childhood hepatocellular carcinoma39 and small cell lung cancer40; c-kit mutants in gastrointestinal stromal tumors (GISTs),41, 42 testicular seminoma,43 small cell lung cancer44 and melanoma45; FGFR3 mutants in bladder cancer,46, 47 cervical cancer46 and oral squamous cell cancer48; PDGFRα mutants in GISTs49; EGFR mutants in glioblastoma multiforme,50 non-small cell lung cancer,51, 52, 53 colorectal cancer,54 ovarian cancer,55 oesophageal and pancreatic adenocarcinoma.56

In childhood solid tumors, except hepatocellular carcinoma39 and a minority of GISTs,57, 58 no such constitutive activation has been reported so far. However, autocrine and/or paracrine activation of receptor TKs has been described in a few of them, providing a tool for the successful utilization of selective inihibitors in preclinical models. Neuroblastoma express PDGFRα and PDGFRβ,59 c-Kit60, 61 and EGFR,62, 63 and their selective inhibition can reduce tumor growth both in vitro and in vivo.10, 11, 12, 63 Likewise, the selective inhibition of c-Kit has been reported to reduce in vitro and in vivo tumor growth in Ewing's sarcoma.9 With regard to RMS, immunohistochemical detection of c-Kit has been reported in a small series of primary tumors,23 and to date there is only 1 report on PDGFRα and PDGFRβ expression in primary tumors, which points to a treatment role for imatinib mesylate in this malignancy.22

The 5 RMS cell lines investigated in the present study were representative of both embryonal and alveolar subtypes, and both known translocations. Notable levels of PDGFRβ expression were demonstrated in all cell lines, with the highest levels in the embryonal RD cell line. On the contrary, significant PDGFRα expression was only found in the alveolar RH30 and RMZ-RC2 cell lines. However, when the phosphorylation status of the 2 receptors in RD and RH30 cells was investigated in basal conditions and after selective stimulation, an increased phosphorylation of PDGFRβ was demonstrated in RD cells but not in RH30, whilst no change in PDGFRα phosphorylation could be detected in either cell lines. These findings suggest that (i) PDGFRα signaling is not active in RD and RH30 cell lines despite the significant levels of expression found in the former and (ii) an active PDGFRβ signaling is only detectable in RD. Significant expression of c-Kit was not detected in any of the 5 cell lines.

Given the possibility that the multidrug transporter ABCG2 might also play a role in imatinib mesylate activity,13, 14 the expression and activity of this molecule were also investigated. High levels of expression were found in RH30 and RMZ-RCA, both alveolar subtype, whilst lower but still notable levels were only demonstrated in RD18 among the embryonal cell lines. In vitro functional efflux studies using Hoechst 33342 as a fluorescent agent whose accumulation rate into cells is indicative of ABCG2 activity showed that in RH30 cells imatinib mesylate is able to inhibit ABCG2 activity in a concentration-related manner, with an inhibitory activity at concentrations in the range of levels therapeutically achievable in vivo.64 Incubation of the same cell lines with Ko143, a selective inhibitor of ABCG2,35 exhibited a similar inhibition pattern as obtained with imatinib mesylate, lending further support to the fact that imatinib mesylate has a blocking effect on ABCG2 activity. In RD cells on the contrary, the inhibition of ABCG2 activity by either imatinib mesylate or Ko143 was able to inhibit only partially (∼10%) Hoechst 33342 extrusion and this is in agreement with the low levels of expression of ABCG2 in these cells.

On the basis of the pattern of expression of the target proteins and on the favorable characteristics of in vitro and in vivo cell growth, 2 cell lines were chosen for the experiments aimed at evaluating the in vitro and in vivo antitumor activity of imatinib mesylate as single agent and in combination with topotecan: the embryonal RD cell line and the alveolar RH30 cell line. When utilized as single agent, imatinib mesylate inhibited in vitro tumor growth only at concentrations not therapeutically achievable in vivo, with an IC₅₀ of 22.6 μM and 29.0 μM for RD and RH30, respectively. Similarly, when imatinib mesylate was administered in vivo at the usual doses and using a conventional schedule,12 marginal tumor growth inhibition was only observed in RH30 xenograft but not in RD. Taken together, these findings suggest that imatinib mesylate as single agent has marginal therapeutic activity in RMS tumor growth, even in those cell lines such as RD in which a low level of activity of PDGFRβ signaling is detectable.

Previous studies in leukemia and solid tumor preclinical models assessing the combination of imatinib mesylate with conventional cytotoxic drugs have indicated that a significant synergism may occur between these 2 mechanistically distinct therapeutic strategies.65, 66, 67, 68, 69, 70 In this study, TPT was chosen as its significant antitumor activity has been described in RMS preclinical models,71, 72 and subsequently proven in the clinical setting.24 The Chou-Talalay median-effect equation method32 was employed to quantify at different levels of concentrations whether antagonistic, addictive or synergistic effects for the combination TPT and imatinib mesylate were in action in vitro. Such analysis has been shown to be useful for identifying effective drug combinations underpinning clinical therapeutic protocols.73 In 2 further independent series of MTT assays, a synergistic interaction between imatinib mesylate and TPT was demonstrated at all the ICs tested, ranging from a moderate synergism (CI: 0.7–0.85) at the lower concentrations in both RD and RH30 cells, to a stronger synergism (CI: 0.3–0.7) at the higher concentrations in RD but not RH30 cells.

As far as the molecular events underlying this synergism are concerned, besides the involvement of other not yet identified molecules, at least 2 different mechanisms are involved in relation to the synergistic interaction between imatinib mesylate and TPT in in vitro RMS models, both at receptor TK and multidrug transporter level, depending on the cellular context. Given that in RD cells, PDGFRβ signaling was found to be low but active and was partially inhibited by pretreatment with imatinib mesylate, whilst only low levels of ABCG2 expression and activity were found, it seems reasonable to assume that in this cell line imatinib mesylate was mainly effective through the inhibition of PGDFRβ antiapoptotic signaling. However, in the in vivo setting, a higher level of activity could be present as a result of the different microenvironment, which could positively influence this signal transduction. In RH30 cells instead, because of the high levels of expression and activity of ABCG2, the mechanism of action was more likely to be mediated through the inhibition of ABCG2 and consequent increase of TPT intracellular concentrations.

Subsequently, further experiments were carried out to evaluate the antitumor activity of imatinib mesylate as single agent and in combination with TPT in RD and RH30 xenografts. Prior to these, a series of preliminary experiments were undertaken to ascertain the optimal doses of the 2 agents when administered in combination. Two dose levels for imatinib mesylate (100 and 200 mg/kg/day)12 and 3 dose levels (0.5, 1.0 and 1.5 mg/kg/day) for TPT71, 72 were investigated. However, an unexpectedly high rate of toxic deaths was observed at the higher dose levels, possibly attributable to the inhibition of ABCG2 activity in gut epithelial cells and in cells lining the biliary tree, leading to greater intestinal absorption and reduced hepatobiliary excretion of oral TPT as previously described in relation to the coadministration of ABCG2 inhibitors.74, 75 In the subsequent experiments, imatinib mesylate was therefore administered at 100 mg/kg/day and TPT at 0.50 mg/kg/day. In agreement with the in vitro findings, imatinib mesylate as a single agent only induced marginal tumor growth inhibition in RH30 xenografts. However, when imatinib mesylate was administered in combination with TPT, a significant therapeutic synergism was demonstrated in both xenografts, with a tendency for a greater therapeutic increase in RH30 xenograft. Significantly, this therapeutic synergism was not associated with any detectable increase of systemic toxicity.

Previous studies have shown that imatinib mesylate inhibits ABCG2 activity13, 14 and can reverse resistance to TPT in vitro in cells engineered to over-express ABCG2.13 The present study confirms and extends these findings, providing evidence that imatinib mesylate can potentiate TPT activity both in vitro and in vivo also in cells expressing constitutive levels of ABCG2. In addition to the direct inhibition of PGDFRβ and ABCG2 in tumor cells active in vitro, at least 2 other mechanisms both mediated by the imatinib-induced inhibition of PDGFR signaling in normal stromal cells could also account for this in vivo synergism: (i) the increase of TPT uptake caused by the reduction of tumor interstitial fluid pressure and increase of transcapillary transport; (ii) the inhibition of tumor angiogenesis.76 Further additional molecular mechanisms, however, cannot be ruled out based on the present findings.

This study is the first one to investigate in RMS the use of imatinib mesylate alone and in combination with a cytotoxic drug. Several potentially important conclusions can be extrapolated. First, imatinib mesylate alone is only marginally effective in RMS preclinical models, at least in the range of concentrations therapeutically achievable. Second, with regard to TPT, imatinib mesylate appears to be an effective chemosensitizer for RMS cells both in vitro and in vivo, and this enhancement seems to be related to the inhibition of ABCG2 and/or PDGFRβ at consitutive levels of expression. Third, the doses utilized for each individual drug at which the combination is effective are lower than expected, and this may need to be taken into account in the clinical setting. The theoretical and preclinical rationale to investigate in RMS the combination of TPT and imatinib mesylate warrants further study in the clinical setting, especially in ABCG2-expressing tumors.

Acknowledgements

Dr. H.P.M. was a recipient of a Cancer Research UK Clinicians bursary. The authors thank Novartis Pharma AG (Basel, Switzerland) for providing imatinib mesylate utilized in this study, Dr. E. Buchdunger for the fruitful critical discussions, Dr. F.M. Perla for the help in cell cultures, Ms. O. Mannarino and Ms. G. Cusano for the skilful technical assistance.