¹Children's Cancer Research Institute, Sydney Children's Hospital, Sydney, Australia

²Oncotech Inc, Irvine, California, USA

³Department of Pediatrics, Northwestern University Medical School, Chicago, Illinois, USA

⁴Department of Pediatrics, University of Illinois, Chicago, Illinois, USA

^aAuthor for correspondence

We have recently shown a close correlation between expression of the Multidrug Resistance-associated Protein (MRP) gene and the MYCN oncogene and provided evidence that high MRP expression is a powerful independent predictor of poor outcome in neuroblastoma (Norris et al., New Engl. J. Med., 334, 231 - 238, 1996). The effect of MYCN down-regulation on MRP expression and response to cytotoxic drugs was investigated in NBL-S neuroblastoma cells transfected wtih MYCN antisense RNA constructs. Concomitant with MYCN down-regulation, the level of MRP expression was decreased in the NBAS-4 and NBAS-5 antisense transfectants. These cells demonstrated significantly increased sensitivity to the high affinity MRP substrates vincristine, doxorubicin, sodium arsenate and potassium antimony tartrate, but not to the poor MRP substrates, taxol or cisplatin. Similarly, transfection of full-length MYCN cDNA into SH-EP neuroblastoma cells resulted in increased MRP expression and significantly increased resistance specifically to MRP substrates. The results provide evidence for the MYCN oncogene influencing cytotoxic drug response via regulation of MRP gene expression. Our data also provide a link between the malignant and chemoresistant phenotypes of this childhood malignancy.

neuroblastoma; MYCN oncogene; MRP gene; antisense; transfection; cytotoxicity

Introduction

Neuroblastoma, a disease which arises from embryonal neural crest cells, is the most common solid tumour of infancy (Brodeur et al., 1992). The majority of patients present with widely disseminated disease at diagnosis and despite intensive regimens, the prognosis for such patients is poor, with 5 year survival rates of less than 40% (Stram et al., 1996). A number of prognostic factors have been identified for this disease, with one of the most powerful being amplification of the MYCN oncogene (Brodeur et al., 1992). Thus patients whose tumours display MYCN gene amplification are at increased risk of early treatment failure, suggesting that the MYCN oncoprotein is closely linked to the drug resistant phenotype and disease progression of these tumours.

A candidate target gene for regulation by MYCN is the MRP gene (Cole et al., 1992) which encodes a transmembrane glycoprotein that is capable of conferring resistance to a wide range of structurally unrelated hydrophobic drugs by acting as a drug efflux pump (Zaman et al., 1994). We previously observed a close association between MYCN and MRP over-expression and hypothesized that MYCN regulates expression of this drug resistance gene (Bordow et al., 1994; Norris et al., 1996). The present study has investigated this hypothesis in neuroblastoma cells transfected with expression vectors containing either antisense or sense MYCN cDNA.

Results and discussion

Two independent patient-derived neuroblastoma clones, NBAS-4 and NBAS-5, were previously selected following transfection into the human neuroblastoma cell line NBL-S (Schmidt et al., 1994), to constitutively express high levels of a 1.2 kb antisense MYCN cDNA. Although parental NBL-S cells lack MYCN gene amplification, they nevertheless display high levels of MYCN gene expression, making them an ideal target cell line for studying the effects of inhibition of MYCN gene expression by antisense to MYCN. The effect of MYCN down-regulation on MRP gene expression was examined after stable antisense expression in these two clones. Following RT - PCR analysis, NBAS-4 and NBAS-5 demonstrated an approximately threefold decrease in MRP gene expression compared to both the parental NBL-S cells and NBV-1 vector-only control (Figure 1a). A similar result was obtained by Northern analysis, which showed clearly detectable MRP RNA in the control cell lines, but no detectable MRP bands in the two antisense clones (Figure 1b).

In order to assess whether introduction of MYCN antisense RNA into the NBL-S cell line had led to down-regulation of MRP protein expression in the NBAS-4 and NBAS-5 cell lines, flow cytometry was performed on the antisense transfectants. NBAS-4, NBAS-5 and NBV-1 cells were labelled by indirect immunofluorescence with mAb MRPr-1, which recognizes an internal epitope of human MRP (Flens et al., 1994). Consistent with the results of both RT - PCR and Northern analyses, MRP expression was detectable in the NBV-1 cells, while there was no reactivity of MRPr-1 over that seen with the isotype control antibody in either of the antisense clones (Figure 2a). Similarly, immunocytochemistry with mAb MRPr-1 demonstrated membrane staining in the parental NBL-S cells, while no staining was apparent in NBAS-5 cells (Figure 2b). Immunocytochemical analysis with the anti-MYCN mAb, NCM II 100, confirmed our earlier results (Schmidt et al., 1994) showing strong nuclear staining in the NBL-S parent cell line, but undetectable MYCN protein in NBAS-5 cells (Figure 2b). In contrast to the results demonstrating down-regulation of both MRP gene expression and MRP protein in the NBAS-4 and NBAS-5 clones, there was no apparent change in the levels of MDR1-P-glycoprotein in the antisense cells by comparison with vector-only controls (Figure 2c).

To determine whether the decreased level of MRP in the antisense transfectants was accompanied by a decrease in resistance to cytotoxic drugs, the cells were tested for their response to a range of compounds (Table 1). In comparison to NBV-1 control cells, NBAS-4 and NBAS-5 demonstrated a twofold increase in sensitivity to the anthracycline doxorubicin, and a 30-fold increase in sensitivity to the vinca alkaloid vincristine. NBAS-4 and NBAS-5 cells similarly demonstrated significantly increased sensitivity to the heavy metal anion compounds potassium antimony tartrate and sodium arsenate, which have been shown to be effective substrates for MRP (Cole et al., 1994). P-glycoprotein over-expressing cells, by contrast, display no cross-resistance to such heavy metal anions (Cole et al., 1994). No significant alteration in the response of the antisense transfectants to either taxol or cisplatin was observed by comparison with the vector-only control cells. While taxol is a highly effective substrate for P-glycoprotein, neither it nor cisplatin appears to be a good substrate for MRP based on the results of transfection studies (Cole et al., 1994; Grant et al., 1994). The pattern of results is thus consistent with the change in cytotoxic drug-response being mediated specifically by MRP.

Since down-regulation of MYCN gene expression had resulted in decreased MRP expression and increased sensitivity to cytotoxic drugs, we wished to study the effect of increased MYCN oncogene expression on both MRP expression and cytotoxic drug response. The SH-EP neuroblastoma cell line, lacking MYCN gene amplification and with negligible MYCN gene expression, was stably transfected with full-length MYCN cDNA (SH-EP/MYCN). Following RT - PCR analysis, SH-EP/MYCN cells demonstrated significantly increased MRP gene expression (mean±s.e. PCR ratio, 0.571±0.057) compared to both the parental SH-EP cells (mean PCR ratio, 0.398±0.031) and SH-EP/CMV vector-only control (mean PCR ratio, 0.401±0.038; P<0.05 for each comparison). To assess whether increased MRP expression in the MYCN transfectant cells was associated with increased drug resistance, the cells were tested for their response to various cytotoxic drugs (Table 2). By comparison with SH-EP/CMV control cells, SH-EP/MYCN demonstrated significantly increased resistance to three effective MRP substrates, doxorubicin, etoposide and sodium arsenate. No significant increase in resistance of SH-EP/MYCN cells to either cisplatin or taxol was observed by comparison with vector-only control cells. Consistent with the findings from the antisense transfectants, this pattern of drug response appears to be specifically mediated by MRP.

We have previously reported a close correlation between the levels of MYCN and MRP gene expression in both primary neuroblastoma samples and in cultured neuroblastoma cell lines following retinoic-acid induced differentiation (Bordow et al., 1994; Norris et al., 1997). Moreover, we have recently reported that the levels of MRP gene expression closely reflect the levels of MYCN gene expression in human neuroblastoma cell lines that have negligible-(SH-EP), intermediate-(NBL-S) and high-(BE(2)-C) level expression of the oncogene, respectively (Norris et al., 1998). The response of such cell lines to cytotoxic drugs which are MRP substrates correlates with their levels of MYCN and MRP expression (M Haber and M Norris, unpublished data). Similarly, the SK-N-SH cell line, with intermediate-level MRP gene expression (Norris et al., 1996), is significantly more resistant to cytotoxic drugs than its flat cell subclone, SH-EP (Spengler et al., 1986). While these data are consistent with the hypothesis that the oncogene influences the response of neuroblastoma cells to cytotoxic drugs by modulating MRP gene expression, the interpretation of such data is potentially confounded by alternate mechanisms of drug resistance, such as P-glycoprotein over-expression, operating in the various cell lines. In contrast, the present findings, in which specific up- or down-regulation of the MYCN oncogene in defined cell populations, results in concomitant modulation of both MRP gene expression and cytotoxic drug response, provides strong evidence in support of this hypothesis.

MYCN belongs to the class of basic helix - loop - helix leucine-zipper transcription factors which includes all the MYC family members and there is strong evidence that these oncoproteins are involved in regulating a range of cell cycle-related functions (Grandori and Eisenman, 1997). There is also a large body of evidence indicating that MYCN plays an important role in the neoplastic development and growth of neuroblastoma (Schwab et al., 1985; Weiss et al., 1997). Most recently, direct evidence of MYCN contributing to neuroblastoma tumorigenesis has been provided by the development of neuroblastoma in transgenic mice over-expressing human MYCN targeted to neural crest cells (Weiss et al., 1997). Although MYCN heterodimerizes with another protein, termed MAX, to form a complex with specific DNA-binding activity (Grandori and Eisenman, 1997), it is only MYCN which possesses a transcriptional activation domain in the N-terminal region of the protein (Wenzel and Schwab, 1995). The level of MAX in neuroblastoma, unlike MYCN, seems to be relatively constant and independent of tumour stage (Raschella et al., 1994). Despite the role that MYCN is believed to have as a transcriptional activator, and the abundant evidence for its central contribution to the malignant phenotype of neuroblastoma, the number of genes that have been identified as targets for this oncoprotein has to date been disappointingly small (Grandori and Eisenman, 1997). The present study suggests that this oncogene may regulate a gene involved in mediating drug resistance, thereby providing a link between the malignant and chemoresistant phenotypes of childhood neuroblastoma. Studies of the MRP gene promoter, involving deletion analysis of MRP promoter-luciferase vector constructs, are currently in progress to determine whether MYCN directly regulates transcription of MRP via one or more of the three E-box motifs which are present in the promoter sequence. Irrespective, since drugs such as doxorubicin, vincristine and etoposide are used in the treatment of children with this disease, the present findings, showing that the response to these drugs can be influenced by the MYCN oncogene, have direct clinical relevance.

In summary, the introduction of an MYCN antisense vector into neuroblastoma cells resulted in the down-regulation of MRP expression and increased drug sensitivity. This increased sensitivity cannot be attributed to reduced growth rate of the transfected cells associated with down-regulation of MYCN, since cells which cycle more slowly have been shown to be more, not less, resistant to drugs such as doxorubicin (Lee et al., 1997). The antisense transfectants demonstrated an MRP-specific response pattern to the cytotoxic drugs tested and moreover, no detectable change in the level of P-glycoprotein expression could be discerned in these cells. Conversely, the introduction of full-length MYCN cDNA into neuroblastoma cells otherwise expressing very low levels of MYCN, resulted in up-regulation of MRP expression and increased drug resistance. Given the close association between MYCN and MRP gene expression which has been demonstrated both in vitro and in vivo (Bordow et al., 1994; Norris et al., 1996), this study suggests a direct association between MYCN and the chemoresistant phenotype of neuroblastoma, and provides an explanation for the established association between MYCN gene amplification and poor outcome in this disease.

Materials and methods

Cell lines

NBL-S, a MYCN unamplified human neuroblastoma cell line and NBL-S-derived MYCN antisense transfectant cell lines NBAS-4 and NBAS-5, and their vector-only transfectant control, NBV-1, have been described previously (Schmidt et al., 1994). The antisense cell lines were derived by co-transfection of NBL-S cells with a 1.2 kb MYCN antisense cDNA inserted into a pCMV₂ vector together with the pSV2neo plasmid, followed by isolation of neomycin-resistant clones. The SH-EP/MYCN cell line was derived by co-transfection (Lipofectamine, Life Technologies, Gaithersberg, MD, USA) of SH-EP cells (Ross et al., 1983) with a 1.6 kb cDNA containing the complete coding sequence of MYCN inserted into the pCMV₂ vector, (a kind gift from Dr Geoffrey Krystal, Medical College of Virginia, Richmond, VA, USA), together with pSV2neo, followed by isolation of neomycin-resistant clones. The vector-only transfectant control, SH-EP/CMV, was similarly derived using pCMV₂ lacking MYCN sequences. Increased MYCN expression, by comparison with SH-EP/CMV, was confirmed by Northern and Western analyses. The MRP over-expressing human breast cancer cell line, MCF7/VP (Schneider et al., 1994), was generously supplied by Dr E Schneider (Wadsworth Center for Laboratories and Research, Albany, NY, USA). All lines were maintained in DMEM supplemented with 10% foetal calf serum and were free of mycoplasma. Transfectant cell lines were intermittently passaged in neomycin to ensure maintenance of constructs.

RT - PCR and Northern analysis

The synthesis of cDNA using Moloney murine leukemia virus reverse transcriptase and random hexanucleotide primers, and the RT - PCR assay have been described previously (Bordow et al., 1994; Norris et al., 1996). A cDNA equivalent of approximately 50 ng was amplified for 30 cycles using primers to the MRP target sequence and a control gene sequence (₂-microglobulin) together in the same tube. Primer pairs, described elsewhere (Bordow et al., 1994; Norris et al., 1996), were selected on the basis that they spanned an intron - exon boundary, and were tested to ensure that they did not amplify genomic DNA. Following polyacrylamide gel electrophoresis, the level of expression of MRP was densitometrically determined using the Gel Doc 1000 Gel Documentation System (Bio Rad Laboratories, Sydney, Australia) and expressed relative to the level of the control ₂-microglobulin gene expression. Experiments were performed in triplicate. For Northern analysis, total RNA (10 g) was subjected to electrophoresis on 1% agarose gels containing 1´MOPS buffer (20 mM morpholino-propanesulphonic acid, 5 mM sodium acetate, 1 mM EDTA (pH 7.0)) and 6% formaldehyde. Following capillary transfer onto Gene Screen Plus membrane, hybridization with a ³²P-labelled MRP probe generated by PCR amplification with the MRP-specific primers described above, and autoradiography were performed as described previously (Haber et al., 1989). To control for loading and transfer of RNA, the membrane was reprobed with the constitutively expressed gene, glyceraldehyde-3'-phosphate dehydrogenase.

Flow cytometry and immunocytochemistry assays

Flow cytometric analysis by indirect immunofluorescence labelling was carried out with monoclonal antibody (mAb) MRPr-1 (Kamiya Biomedical Co., Tukwila, WA, USA) and mAb 4E3 (Signet Lab., Inc, Dedham, MA, USA), their respective unrelated isotope controls mAb rat IgG2a (Accurate Chemical and Scientific Corp., Westbury, NY, USA) and mAb mouse IgG2a (Becton Dickinson, San Jose, CA, USA), and their respective secondary antibodies, goat anti-rat IgG-phycoerythrin and goat anti-mouse IgG-phycoerythrin. Samples were washed and concentrated to 5´10⁵ cells/ml, prior to fixing and permeabilization (Caltag Lab, San Francisco, CA, USA). Cells were stained with mAb MRPr-1 or 4E3 and indirectly labelled with the appropriate secondary antibody. For immunocytochemistry analysis, cells were cytocentrifuged and fixed in acetone. Cells were stained with NCM II 100 (Oncogene Research Products, Cambridge, MA, USA) or MRPr-1, as indicated, using the Biotin-streptavidin Amplified Detection System (Biogenex, San Ramon, CA, USA).

Cytotoxicity assays

Cytotoxic drugs were purchased from manufacturers as follows: vincristine and etoposide from Sigma (St. Louis, MO, USA); doxorubicin from Farmitalia (Sydney, Australia); potassium antimony tartrate and sodium arsenate from Baker (Biolab Scientific, Sydney, Australia); taxol from Calbiochem (Sydney, Australia); cisplatin from Sigma (St. Louis, MO, USA). Cells were seeded in 96-well plates and cell growth inhibition was determined after 72 h continuous exposure in vitro to various concentrations of cytotoxic drugs, using a microtitre-based assay with the Alamar Blue^TM reagent (Astral Scientific, Sydney, Australia). Determination of ID₅₀ values and statistical analysis was performed as described previously (Haber et al., 1989). The statistical significance between ID₅₀ values from different cell lines was assessed by t-tests, using 2-sided P values. In addition, all significant findings were confirmed using the non-parametric Wilcoxon two-sample rank-sum test.

Acknowledgements

This work was supported by grants to M Haber, GM Marshall and MD Norris from the National Health and Medical Research Council, Australia and from the New South Wales State Cancer Council, Australia. ML Schmidt was supported by the Schweppe Foundation, (Chicago, IL, USA) and the Children's Cancer Research Fund, (Los Angeles, CA, USA). SB Bordow and J Gilbert are the recipients of Australian Postgraduate Awards. The Children's Cancer Research Institute is incorporated as the Children's Cancer Institute Australia for Medical Research. The authors are grateful to R Garcia for expert technical assistance.

Bordow SB, Haber M, Madafiglio J, Cheung B, Marshall GM and Norris MD. (1994). Cancer Res. 54, 5036-5040. MEDLINE

Brodeur GM, Azar C, Brother M, Hiemstra J, Kaufman B, Marshall H, Moley J, Nakagawara A, Saylors R, Scavarda N, Schneider S, Wasson J, White P, Seeger RC, Look T and Castleberry R. (1992). Cancer 70, 1685-1694. MEDLINE

Cole SPC, Bhardwaj G, Gerlach JH, Mackie JE, Grant CE, Almquist KC, Stewart AJ, Kurz EU, Duncan AM and Deeley RG. (1992). Science 258, 1650-1654. MEDLINE

Cole SPC, Sparks KE, Fraser K, Loe DW, Grant CE, Wilson GM and Deeley RG. (1994). Cancer Res. 54, 5902-5910. MEDLINE

Flens MJ, Izquierdo MA, Scheffer GL, Fritz JM, Meijer CJ, Sheper RJ and Zaman GJ. (1994). Cancer Res. 54, 4557-4563. MEDLINE

Grandori C and Eisenman RN. (1997). Trends Biochem. Sci. 22, 177-181. Article MEDLINE

Grant CE, Valdimarsson G, Hipfner DR, Almquist KC, Cole SP and Deeley RG. (1994). Cancer Res. 54, 357-361. MEDLINE

Haber M, Norris MD, Kavallaris M, Bell DR, Davey RA, White L and Stewart BW. (1989). Cancer Res. 49, 5281-5287. MEDLINE

Lee J-S, Scala S, Matsumoto Y, Dickstein B, Robey R, Zhan Z, Altenberg G and Bates SE. (1997). J. Cell. Biochem. 65, 513-526. MEDLINE

Norris MD, Bordow SB, Marshall GM, Haber PS, Cohn SL and Haber M. (1996). N. Engl. J. Med. 334, 231-238. MEDLINE

Norris MD, Bordow SB, Haber PS, Marshall GM, Kavallaris M, Madafiglio J, Cohn SL, Salwen H, Schmidt ML, Hipfner DR, Cole SPC, Deeley RG and Haber M. (1997). Eur. J. Cancer 33, 1911-1916. MEDLINE

Norris MD, Madafiglio J, Gilbert J, Marshall GM and Haber M. (1998). Eur. J. Cancer in press.

Raschella G, Romeo A, Negroni A, Pucci S, Dominici C, Castello MA, Bevilacqua P, Felsani A and Calabretta B. (1994). Cancer Res. 54, 2251-2255. MEDLINE

Ross RA, Spengler BA and Biedler JL. (1983). J. Natl. Cancer Inst. 71, 741-749. MEDLINE

Schmidt ML, Salwen HR, Manohar CF, Ikegaki N and Cohn SL. (1994). Cell Growth Differ. 5, 171-178. MEDLINE

Schneider E, Horton JK, Yang C-H, Nakagawa M and Cowan KH. (1994). Cancer Res. 54, 152-158. MEDLINE

Schwab M, Varmus HE and Bishop JM. (1985). Nature 316, 160-162. MEDLINE

Spengler BA, Ross RA and Biedler JL. (1986). Cancer Treat Rep. 70, 959-965. MEDLINE

Stram DO, Matthay KK, O'Leary M, Reynolds CP, Haase GM, Atkinson JB, Brodeur GM and Seeger RC. (1996). J. Clin. Oncol. 14, 2417-2426. MEDLINE

Weiss WA, Aldape K, Mohapatra G, Feuerstein BG and Bishop JM. (1997). EMBO J. 16, 2985-2995. MEDLINE

Wenzel A and Schwab M. (1995). Eur. J. Cancer 31A, 516-519. MEDLINE

Zaman GJR, Flens MJ, van Leusden MR, de Haas M, Mulder HS, Lankelma J, Pinedo HM, Scheper RJ, Baas F, Broxterman HJ and Borst P. (1994). Proc. Natl. Acad. Sci. USA 91, 8822-8826. MEDLINE

Figure 1 (a) MRP gene expression in parental NBL-S cells, vector-only control NBV-1 cells and MYCN-antisense transfected clones, NBAS-4 and NBAS-5. Following independent, competitive RT - PCR analysis performed on at least three separate occasions, ratios between MRP and control ₂-microglobulin gene products were densitometrically determined for each sample. Columns, mean; bars, s.e. The level of MRP gene expression in each of the MYCN antisense clones was significantly lower (P<0.05) than that in either the NBL-S or NBV-1 cells. (b) Northern blot analysis of MRP gene expression. Aliquots of RNA (10 g) from the MRP-overexpressing positive control cell line MCF7/VP, parental NBL-S cells, vector-only control NBV-1 cells and MYCN-antisense transfected clones, NBAS-4 and NBAS-5, were electrophoresed in 1% agarose prior to Northern hybridization with a ³²P-labelled PCR-amplified fragment of the MRP gene (upper panel). The same membrane was reprobed with the control gene, encoding glyceraldehyde-3'-phosphate dehydrogenase (lower panel)

Figure 2 (a) Flow cytometric analysis of MRP using mAb MRPr-1. Vector-only control NBV-1 and MYCN antisense transfected clones, NBAS-4 and NBAS-5, were stained with MRP-specific mAb MRPr-1 and an isotype control mAb rat IgG2a, indirectly labelled in each case with phycoerythrin. (b) Immunocytochemical detection of MYCN and MRP, using mAbs NCM II 100 and MRPr-1, respectively, in cytocentrifuge preparations of parental NBL-S (upper panels) cells and MYCN antisense-transfected (lower panels) NBAS-5 cells. (c) Flow cytometric analysis of P-glycoprotein using mAb 4E3. Vector-only control NBV-1 and MYCN antisense transfected clones, NBAS-4 and NBAS-5, were stained with P-glycoprotein-specific mAb 4E3 and an isotype control mouse IgG2a mAb, indirectly labelled in each case with phycoerythrin

Table 1 Table 1

Table 2 Table 2