c-Kit is preferentially expressed in MYCN-amplified neuroblastoma and its effect on cell proliferation is inhibited in vitro by STI-571
Article first published online: 14 MAY 2003
Copyright © 2003 Wiley-Liss, Inc.
International Journal of Cancer
Volume 106, Issue 2, pages 147–152, 20 August 2003
How to Cite
Vitali, R., Cesi, V., Nicotra, M. R., McDowell, H. P., Donfrancesco, A., Mannarino, O., Natali, P. G., Raschellà, G. and Dominici, C. (2003), c-Kit is preferentially expressed in MYCN-amplified neuroblastoma and its effect on cell proliferation is inhibited in vitro by STI-571. Int. J. Cancer, 106: 147–152. doi: 10.1002/ijc.11187
- Issue published online: 5 JUN 2003
- Article first published online: 14 MAY 2003
- Manuscript Accepted: 21 FEB 2003
- Manuscript Revised: 27 JAN 2003
- Manuscript Received: 7 OCT 2002
- Italian Ministry of Health, Ricerca Finalizzata. Grant Number: 199/2001
- Associazione per la lotta al Neuroblastoma
- Associazione Italiana per la Ricerca sul Cancro (AIRC)
- Associazione per la Lotte Centro i Tumori Infantili (ALTI)
- Progetto Strategico 1998. Grant Number: L. 669/97-MURST
- c-Kit, Stem Cell Factor;
Coexpression for c-Kit receptor and its ligand stem cell factor (SCF) has been described in neuroblastoma (NB) cell lines and tumors, suggesting the existence of an autocrine loop modulating tumor growth. We evaluated c-Kit and SCF expression by immunohistochemistry in a series of 75 primary newly diagnosed neuroblastic tumors. Immunostaining for c-Kit was found in 10/75 and for SCF in 17/75, with 5/10 c-Kit–positive tumors also expressing SCF. For both, c-Kit and SCF staining were predominantly found in the most aggressive subset of tumors, i.e., those amplified for MYCN: c-Kit was detected in 8/14 amplified vs. 2/61 single copy (p<0.001), and SCF in 9/14 amplified vs. 8/61 single copy tumors (p<0.001). Furthermore, the association of c-Kit expression with advanced stage (3 or 4) (p=0.001) and of SCF expression with adrenal primary (p=0.03) was substantiated. The in vitro activity of the tyrosine kinase inhibitor STI-571 (imatinib mesylate, Gleevec, Glivec) on NB cell lines positive or negative for c-Kit was also assessed. When cells were grown in 10% fetal calf serum, the 4 c-Kit-positive cell lines tested were sensitive to STI-571 growth inhibition to a different extent (ranging from 30 to 80%); also the c-Kit-negative cell line GI-CA-N was slightly affected, suggesting that other STI-571 targets operate in regulating NB proliferation. In addition, c-Kit-positive cell lines SK-N-BE2(c) and HTLA230, grown in SCF only, remained sensitive (40 and 70% of growth inhibition, respectively), while, in the same conditions, proliferation of the c-Kit-negative cell line GI-CA-N was not affected. Immunoprecipitation of c-Kit from cell lysates of SK-N-BE2(c) and HTLA230 cells grown in SCF and subsequent western blot analysis of the immunoprecipitates revealed a sharp decrease of c-Kit phosphorylation after STI-571 treatment. These data demonstrate that both c-Kit and SCF are preferentially expressed in vivo in the most aggressive neuroblastic tumors and that their signaling is active in promoting in vitro NB cell proliferation that can be selectively inhibited by treatment with STI-571. © 2003 Wiley-Liss, Inc.
Neuroblastoma (NB), a childhood solid tumor derived from the sympatho-adrenal lineage of the neural crest,1 exhibits a remarkable heterogeneity in clinical behavior. This ranges from spontaneous differentiation to ganglioneuroma,2 the fully mature and benign counterpart, or spontaneous regression,3 to complete remission after minimal treatment,4 or to aggressive tumor growth and disease progression despite very intensive treatment.5 In the last group, which includes approximately 1/3 of all patients, the majority of which have metastatic disease at presentation, the identification of critical molecules to be targeted by rationally developed novel anticancer agents is urgently needed.
In this regard, the recent development of new agents specifically aimed at blocking the activity of tyrosine kinase receptors has opened a novel area of therapeutic intervention for targeting their signaling pathways and effectively inhibiting tumor cell growth.6 A noticeable example is STI-571 (imatinib mesylate, Gleevec/Glivec, Novartis Pharmaceuticals, Basel, Switzerland), a selective inhibitor of several structurally related receptor tyrosine kinases including c-Abl, Bcr-Abl, c-Kit and the platelet-derived growth factor receptor (PDGF-R). This molecule has recently been approved in many Western countries for the treatment of chronic myeloid leukaemia (CML) and gastrointestinal stromal tumors (GIST), and its potential therapeutic activity in other c-Kit-positive malignancies such as mastocytosis, seminoma and acute myelogenous leukemias is presently being investigated.7
In NB, expression of mRNA and/or protein for c-Kit and its ligand stem cell factor (SCF) has been reported both in continuous cell lines8, 9 and in primary tumors,8, 9, 10, 11 thus suggesting the existence of an autocrine loop modulating clonogenicity and tumor growth. Therefore, the present study was aimed at i) determining the expression of c-Kit in a clinical series of neuroblastic primary tumors at diagnosis, ii) evaluating whether the in vitro stimulation of c-Kit with SCF can affect proliferation of NB cell lines, and iii) assessing whether the in vitro treatment with STI-571 can prevent such c-Kit-induced alterations.
MATERIAL AND METHODS
Tumor samples from primary site were obtained by surgery before any treatment from 75 newly diagnosed children with NB or ganglioneuroblastoma (GNB), aged 1 day to 192 months (median, 19 months), and admitted to the Department of Pediatrics at La Sapienza University and to the Division of Oncology at Bambino Gesù Children's Hospital in Rome, and to the Oncology Unit at RLC NHS Trust, Alder Hey, in Liverpool. No selection criteria were applied except for the availability of frozen tumor tissue for the immunohistochemical and molecular analyses. Histopathological examination was carried out by the local pathologist and the tumors were classified as NB (number 58) or GNB (number 17) according to the International Neuroblastoma Pathology Committee (INPC) guidelines.12 The final clinico-pathological diagnosis fulfilled the International Criteria for Neuroblastoma Diagnosis.13 The primary site was adrenal in 37 patients, abdominal nonadrenal in 20, thoracic in 16 and cervical in 2. Stage was assessed according to the International Neuroblastoma Staging System13 and was as follows: 8 patients were at stage 1, 17 at stage 2, 15 at stage 3, 25 at stage 4 and 10 at stage 4S. Institutional informed written consent was obtained from the patients' parents or guardians.
Immunohistochemical and molecular analysis of primary tumors
Tumor tissue obtained at surgery was immediately snap frozen in liquid nitrogen and then stored at −80°C until analysis. Four-micrometer thick cryostat sections from each sample were stained with 0.1% toluidine blue in phosphate-buffered saline (0.1 M sodium phosphate buffered saline, pH 7.2) in order to assess the tumor cell content, and additional sections were used for immunoperoxidase staining after fixation in cold absolute acetone for 10 min. The anti–c-Kit MoAb 1.D29.3D6 (Boehringer Mannheim Biochemica, Mannheim, Germany) and the anti–SCF MoAb 10E514 were used. Indirect avidin-biotin immunoperoxidase staining was performed under standard conditions using commercially available reagents (Vectastain, Vector Laboratories, Burlingame, CA). Slides were then counterstained with Mayer's hematoxylin. Sections incubated with isotype matched control immunoglobulins were used as controls. Immunohistochemical findings were evaluated independently by 2 investigators (M.R.N. and P.G.N.). Stain was scored “positive” when the totality of the lesion was stained although with variable intensity, “weakly positive” when the intensity of the stain was homogenously weak and “variable” when positive and negative tumor areas, the latter not accounting for more than 70% of the lesion, were present. Lesions displaying dubious and negative stain were scored “negative.” MYCN gene copy number was assessed by Southern blot as previously described.10
Human continuous NB cell lines RN-GA,15 HTLA230,16 KCNR,17 SH-EP,18 GI-CA-N,19 SK-N-BE2,8 SK-N-AS,20 SK-N-BE2(c) (a subclone of SK-N-BE2)21 and LAN-522 were cultured in RPMI 1640 medium (Euroclone, Devon, UK) complemented with 10% foetal calf serum (FCS) (Hyclone Laboratories, Inc., Logan, UT) at 37°C in 5% CO2. HTLA230, GI-CA-N, SK-N-BE2(c) cells were treated with STI-571 (Novartis, London, UK) 0.1, 1.0 or 10.0 μM for 48 hr in 10% FCS or with STI-571 10.0 μM in serum-deprived medium supplemented with 50 ng/ml SCF (Pepro Tech EC, Ltd., London, UK) for the same time. GI-CA-N and SK-N-BE2(c) cells were treated with 0.1, 1.0 or 5.0 μM all trans-retinoic acid (ATRA) (Sigma Chemical Co., St. Louis, MO) and 10.0 μM STI-571 for 48 hr in 10% FCS.
RNA extraction and RT-PCR
Total RNA was prepared by TRIzol extraction reagent (Gibco-BRL, Gaithersburg, MD). Carryover DNA contamination was eliminated by treatment of total RNA with DNA-free kit (Ambion, Austin, TX) according to the manufacturer's instructions. RNA was reverse transcribed with the first strand cDNA synthesis kit for RT-PCR (Roche, Mannheim, Germany) using an input of 500 ng for each reaction. Subsequent PCR amplifications were carried out for the indicated number of cycles at the appropriate annealing temperature for each pair of primers:
c-kit primer up 5′GCCCACAATAGATTGGTATTT3′
SCF primer up 5′GATTCTCACTTGCATTTATCTTC3′
β-actin primer up 5′TCATCACCATTGGCAATGAG3′
Cell proliferation was measured using the colorimetric cell proliferation kit WST-1 (Roche) based on the colorimetric detection of a formazan salt. In each well 4×104 cells were seeded in RPMI 1640 medium supplemented with 10% FCS or with 50 ng/ml SCF (Pepro Tech EC, Ltd.). STI-571 and/or ATRA were added at the indicated concentration and the colorimetric reading at 450 nm was carried out after 48 hr according to the manufacturer's instructions. Background absorbance of each sample at 630 nm was subtracted from the readings at 450 nm.
Cellular proteins were extracted as previously described.23 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 described elsewhere.24 Specific antibodies were: anti–c-Kit antibody C14 (Santa Cruz Biotechnology, Santa Cruz, CA), anti–phospho-tyrosine PY20 (BD Transduction Laboratories) and anti–HSP70 (StressGen, San Diego, CA). Detection was performed by ECL Plus kit (Amersham Pharmacia, Piscataway, NJ) according to the manufacturer's instructions. Immunoprecipitations were carried out with 800 μg of total cell lysates as previously described24 except that protein-G was used to bind the antigen-antibody complex.
c-Kit and SCF expression in NB primary tumors
Detectable levels of expression of the c-Kit receptor were observed in 10 out of 75 tumors tested (13%). Six lesions scored positive (Fig. 1a) and 4 variable. In 4 positive cases the stain was clearly associated to the tumor cell membrane and in 3 also with the cytoplasm. Expression of SCF could be demonstrated in 17 out of 75 tumors assayed (23%). Eight lesions scored positive, 8 weakly positive and 1 variable. The immunostaining had a fine ground glass pattern confined to the cell cytoplasm (Fig. 1b). Interestingly, 5/10 c-Kit–positive tumors also expressed SCF. As far as c-Kit expression was concerned, no association with the histopathological categories (NB vs. GNB) or clinical features such as age at diagnosis (<1 year vs. =1 year) and primary site (adrenals vs. nonadrenals) was substantiated (Table I). A significant association was observed instead for c-Kit to be expressed in advanced stages: 10/40 cases (25%) at stages 3/4 vs. 0/35 cases at stages 1/2/4S (p=0.001). As far as SCF expression was concerned, no association with the histopathological categories or clinical features such as age at diagnosis and stage was observed (Table I). A significant association was demonstrated however for SCF to be expressed in adrenal primary site: 12/37 adrenal primaries (32%) vs. 5/38 nonadrenal primaries (13%) (p=0.03).
|c-Kit (+/all)||SCF (+/all)|
|p = n.s.||p = n.s.|
|Age <1 year||2/30||9/30|
|p = n.s.||p = n.s.|
|Age ≥1 year||8/45||8/45|
|p = n.s.||p = 0.03|
|Stages 1, 2 and 4S||0/35||7/35|
|p = 0.001||p = n.s.|
|Stages 3 and 4||10/40||10/40|
|MYCN single copy||2/61||8/61|
|p < 0.001||p < 0.001|
MYCN amplification (>10 copies per haploid genome) was demonstrated by Southern blot in 14/75 tumors (19%). For both c-Kit and SCF proteins, staining was predominantly found in MYCN amplified tumors (Table 1): expression for c-Kit was detected in 8/14 (57%) amplified vs. 2/61 single copy tumors (3%) (p<0.001), and expression for SCF in 9/14 (64%) amplified vs. 8/61 (13%) single copy tumors (p<0.001).
c-Kit and SCF expression in NB cell lines
c-kit and SCF expression were tested in 9 human NB cell lines by RT-PCR (Fig. 2a). c-kit and SCF were detectable in 6 and 8 cell lines, respectively, after 35 PCR cycles. Five cell lines co-expressed the ligand and the receptor. These data confirm that the expression of c-Kit and SCF is a common finding in NB cells growing in vitro. Next, we analyzed the expression of c-Kit protein by western blotting. c-Kit expression was detectable in only 2 cell lines (Fig. 2b), both of which exhibited the highest c-Kit levels also in RT-PCR. It should be noticed that the lower band visible in all samples but RN-GA is a spurious signal detected also by other antibodies against c-Kit.25 The apparent contrast between western blotting and RT-PCR results can be explained by a low level of c-Kit expression in several NB cells that can only be detected with sensitive RT-PCR or by the existence of a post-transcriptional control which prevents the expression of c-Kit protein. c-Kit and SCF expression and MYCN amplification in each cell line used are summarized in Table II.
|Cell line||MYCN amplification||c-Kit expression||SCF expression|
Effect of STI-571 on NB cell lines proliferation
To test possible effects of STI-571 on cell proliferation, we chose 3 cell lines: HTLA230, which was c-Kit positive in RT-PCR and western blot analyses, SK-N-BE2(c), which was positive only by RT-PCR, and GI-CA-N, which was negative for a preliminary experiment to define the most effective STI-571 concentration. Three concentrations of STI-571 (0.1, 1.0 and 10.0 μM) were used to treat the cell cultures for 48 hr. Proliferation was determined with a colorimetric assay taking the same cells, which were kept in basal growth conditions (medium + 10% FCS) as controls. To verify the efficacy of the STI-571 batch used, we treated a BCR-abl expressing cell line at the concentration of 1 μM. We obtained a partial inhibition of cell proliferation only at the highest concentration of STI-571 in SK-N-BE2(c) and in GI-CA-N (Fig. 2b). No effect was detectable in HTLA230 cells in these conditions. The growth of BCR-abl positive cell line was completely inhibited.
We carried out another experiment extending STI-571 (10 μM) treatment for 96 hr and including 2 additional cell lines (KCNR and LAN-5), which were both positive for c-Kit expression by RT-PCR. Interestingly, a more prolonged treatment with STI-571 revealed an inhibitory effect on cell proliferation also in the HTLA230 cell line, which was not inhibited at 48 hr (Fig. 3b). On the other hand, the c-Kit-negative GI-CA-N cells, after a slight decrease at 48 hr of treatment, did not exhibit inhibition at 96 hr (Fig. 3b). The other cell lines used were responsive to STI-571 reaching a plateau of inhibition at 48 hr (Fig. 3b).
To uncouple the specific effect of c-Kit from that of other proliferative pathways operating in the NB cell lines used, we carried out a similar experiment using SCF exogenously added as the only growth factor. Cell cultures were added with 50 ng/ml of SCF and treated with 10 μM STI-571 for 48 hr. Interestingly, the c-Kit positive cell lines [HTLA230 and SK-N-BE2(c)] were sensitive to growth inhibition by STI-571 (Fig. 3c). On the contrary, the c-Kit negative cell line GI-CA-N was not inhibited. Taken together these data demonstrate that c-Kit stimulation is active in promoting the proliferation of NB cells and that STI-571 treatment is able to inhibit c-Kit–induced cell growth.
c-Kit stimulation is involved in proliferation and differentiation control in other cell types.26 In NB, the differentiation therapy using retinoids is commonly used.27 To evaluate if the inhibition of c-Kit signal could influence the differentiation effect of retinoids, we treated GI-CA-N, SK-N-BE2(c), HTLA230, KCNR and LAN-5 cell lines grown in 10% FCS with 0.1, 1.0 and 5.0 μM ATRA with or without STI-571 (10.0 μM) for 48 hr. ATRA alone did not cause any significant decrease in cell proliferation. On the contrary, the SK-N-BE2(c) cells underwent a proliferative burst which was already described in some cell lines during the early phases of neural differentiation.28 The association of ATRA with STI-571 caused a decrease of cell proliferation of about 20% independently from the concentration of ATRA (Fig. 3d). Of interest, the proliferation burst observed in ATRA-treated SK-N-BE2(c) cells was abolished by STI-571 (Fig. 3d). On the other hand, the combination of ATRA and STI-571 did not produce any synergistic or additive effect on cell proliferation respect to the treatment with STI-571 alone (compare % of proliferation in Fig. 3b with that in Fig. 3d for the same cell lines).
STI-571 inhibits c-Kit phosphorylation
To understand if the effects of STI-571 on the NB cell lines tested were associated with specific inhibition of tyrosine phosphorylation, we carried out a western blot analysis in nonreducing conditions using an anti-phosphotyrosine antibody on cell lysates from HTLA230 and SK-N-BE2(c) cells grown in 10% FCS or in 50 ng/ml SCF and treated with STI-571 (10 μM) for 48 hr. A specific band with an apparent molecular weight in excess of 160 kDa was down-regulated in HTLA230 and SK-N-BE2(c) cells after treatment with STI-571 (indicated with an arrow in Fig. 4a).
Next, we carried out an immunoprecipitation using a specific anti-c-Kit antibody on cell lysates from HTLA230 and SK-N-BE2(c) cells grown in presence of SCF (50 ng/ml) and treated with STI-571 (10 μM). Immunoprecipitates were run on a SDS-PAGE and western blot analysis to detect phospho-tyrosine was performed. After stripping, c-Kit was immunodetected on the same filter. A sharp decrease in a specific band detected by the anti-phospho-tyrosine antibody was observed in both cell lines after treatment with STI-571. Of interest, c-Kit immunodetection revealed a band of the same molecular weight (Fig. 4b). These data strongly suggest that STI-571 caused a specific inhibition of the c-Kit phosphorylation.
The c-Kit receptor is expressed in several human tumors.29, 30 Previous studies in NB have reported the simultaneous expression of c-Kit and its ligand SCF both in continuous cell lines8, 9 and in primary tumors.8, 11, 31 Interestingly, cell proliferation and colony-forming ability of c-Kit-positive NB cell lines were inhibited after incubation with blocking antibodies to c-Kit,8 thus demonstrating the existence of a SCF/c-Kit autocrine loop modulating tumor growth.8 In other cell types, the signal transduction by an active c-Kit receptor has been associated with increased proliferation and survival.32 As far as its signaling is concerned, c-Kit is phosphorylated at Tyr568 and Tyr570 following ligand stimulation and phosphorylation is inhibited by tyrosine kinase inhibitors.32
In our study, we confirmed in a larger series of primary tumors and cell lines that neuroblastic tumors expresses c-Kit and SCF. Noteworthy, c-Kit and SCF expression were preferentially found in the most aggressive subset of tumors, i.e., those harboring genomic amplification of MYCN oncogene, a well-recognized hallmark of biological and clinical aggressiveness.13 We cannot exclude however that very low levels of expression for c-Kit, undetectable by means of immunohistochemistry, could be present in a part of the remaining tumors. This appears reasonable because, at least in vitro, the expression of c-Kit seems to be low in several cases as suggested by the lower number of c-Kit positive cell lines by western blot analysis (2 out of 10 cell lines) compared to those positive by the more sensitive RT-PCR technique (6 out of 10).
Targeted anticancer therapies are aimed at tumor eradication without or with minimal interference with the survival of normal cells.33 The tyrosine kinase inhibitor STI-571 has dramatically changed the therapeutic achievements in the cure of CML by specifically targeting the BCR/abl function.34 Although STI-571 possesses a rather selective activity on BCR/abl, few other tyrosine kinases, including c-Kit, have been demonstrated to be responsive to the inhibitory activity of this agent.24, 34 Accordingly, the effectiveness of STI-571 has also been established in vitro and in vivo in c-Kit–expressing GIST,35 thus broadening STI-571 field of clinical application to solid malignancies. Tumor phenotyping for c-Kit has in fact demonstrated the expression of the receptor in a significative fraction of metastatic melanoma36 and NSCLCs37 among others.
In NB cell lines, STI-571 was able to inhibit cell proliferation albeit less efficiently than in a BCR-abl positive cell line used as control. STI-571-driven growth inhibition was associated with a decrease in c-Kit phosphorylation as demonstrated by immunoprecipitation experiments. Of interest, STI-571 inhibition of proliferation was also detectable, at least during the first 48 hr of treatment, in a c-Kit negative cell line (GI-CA-N) grown in complete medium (with 10% FCS). The latter finding suggests that although STI-571 is able to inhibit the auto-phosphorylation of c-Kit in the c-Kit positive cell lines, other STI-571–sensitive pathways may exist in NB.
In clinical practice, cancer chemotherapy relies on the use of drug combinations chosen with the aim of producing additive or synergistic therapeutic effects. Neuroblastic tumors frequently maintain a certain degree of ability to differentiate under a variety of stimuli in vitro and in vivo. Differentiation therapy using retinoids is thus routinely employed in NB treatment.26 Therefore, we investigated in vitro the existence of a possible synergism between STI-571 and ATRA. In NB cell lines, however, the effect of STI-571 in combination with ATRA was not increased as compared to that of STI-571 alone. On the other hand, we noticed that STI-571 abolished the proliferative burst detected in SK-N-BE2(c) brought about by ATRA alone.
In conclusion, c-Kit is mainly expressed in the most aggressive neuroblastic tumors and STI-571 is able to inhibit NB proliferation in vitro through c-Kit inhibition. Although we demonstrated a specific effect of STI-571 on c-Kit receptor, our data do not rule out the possibility that STI-571 targets other than c-Kit are also active in this tumor. Identification of these molecules could help to tailor additional specific tyrosine kinase inhibitors that could be used in combination with or in alternative to STI-571 for a more effective NB treatment.
We thank Novartis Pharma AG for the STI-571 used in our study and C. Mancini for technical assistance. R.V. is recipient of a fellowship from Bambino Gesù Children's Hospital. V.C. is recipient of a fellowship from Fondazione Adriano Buzzati-Traverso.
- 5Induction chemotherapy in metastatic neuroblastoma: does dose influence response? A critical review of published data standards, options and recommendations (SOR) project of the National Federation of French Cancer Centres (FNCLCC). Eur J Cancer 2000; 36: 1808–15., , , , .
- 26A novel mechanism of cooperation between c-Kit and erythropoietin receptor: stem cell factor induces the expression of Stat5 and erythropoietin receptor, resulting in efficient proliferation and survival by erythropoietin. J Biol Chem 2001; 276: 1099–6.and .
- 31Expression of c-Kit and stem cell factor is associated with N-myc gene amplification in neuroblastoma. Proc Am Assoc Cancer Res 1996; 37: 220–1., , , , , , .