Presence of classical multidrug resistance and P-glycoprotein expression in human neuroblastoma cells

Annals of Oncology 9: 1009-1014, 1998. © 1998 Kluwer Academic Publishers. Primed in the Netherlands.

Original article Presence of classical multidrug resistance and P-glycoprotein expression in human neuroblastoma cells Ch. Kurowski & F. Berthold Department of Pediatric Hemalology and Oncology, University of Cologne, Germany

tures, SK-N-SH cells showed a relative resistance to vincristine and adriamycin (45.1 and 12.7-fold resp.) and reduced intraBackground: The role of P-glycoprotein (Pgp) associated multi- cellular accumulation of rhodamine-123 which could be nordrug resistance for neuroblastoma patients is controversial. malized by the Pgp blocker verapamil. Therefore we asked whether at all the typical functional features Pgp expression was detected by immunocytochemistry in 14 of the multidrug resistance phenotype could be found in out of 62 tumors (22.6%). No correlation was found to the neuroblastoma cells and studied the prognostic relevance of stage of the disease (P - 0.33), histopathological grading (P = Pgp expression. 0.82), N-myc oncoprotein expression (P = 0.76) or N-myc Patients and methods: Tumor touch preparations and tumor oncogene amplification (P = 0.20). Kaplan-Meier analysis of cell infiltrated bone marrow smears of 62 neuroblastoma event free survival for stage 4 tumors revealed a weak trend of patients were investigated. The expression of Pgp was deter- inferior survival for patients with Pgp positive tumors (logmined by a highly sensitive immunosandwich technique. Drug rank analysis: P = 0.069). resistance studies were performed by exposing cells to PgpConclusions: Though Pgp expression is detectable and funcdependent cytostatic drugs in tissue cultures. Intracellular tional in neuroblastoma cells, but its presence does not provide drug accumulation was examined by rhodamine-123 fluores- much information to the complex phenomenon of chemothercence microscopy. apy resistance in patients. Results: Pgp expression was demonstrable for the SK-N-SH cell line, but not detectable in CHP-100 and ten other neuroblastoma cell lines by immunocytochemistry. In tissue cul- Key words: multidrug resistance, neuroblastoma, p-glycoprotein

Introduction

Adrenal medulla represents the primary tumor site of neuroblastoma in 45 to 50% [9]. Therefore, neuroblastoma is supposed to express Pgp. Detection of Pgp in human neuroblastoma tumors has been reported [1015]. However, the impact on neuroblastoma tumor biology remains unclear: does Pgp represent a marker of cellular differentiation in the sense of a physiological transporter protein in adrenal medulla tissue or does Pgp rather confer the multidrug resistant phenotype to neuroblastoma cells? In the present study, we investigated the association of Pgp expression and the multidrug resistance phenotype in the human neuroblastoma cell lines. Furthermore, we analyzed the prognostic significance of Pgp expression in neuroblastoma tumor samples.

Classical multidrug resistance is one of the in vitro models for resistance in cancer chemotherapy. It is characterized by overexpression of the 170 kilodalton plasma membrane component P-glycoprotein (Pgp) [1]. Pgp possesses sequence homology to ATP-dependent bacterial transport proteins [2]. Therefore Pgp has been postulated to function as an ATP-dependent transmembrane drug efflux pump of broad specificity. Overexpression of Pgp is thought to confer cross resistance to different cytostatics by decreasing the intracellular drug accumulation [3]. Competition of substances at the substrate binding site of Pgp implies the possibility of chemosensitizing by reversing Pgp mediated drug efflux [4-6]. Occurrence of Pgp is not restricted to multidrug resistant tumor cells. A physiological expression has Patients and methods been demonstrated for several human tissue types [7, 8]. Among others, detection of Pgp has been reported for Cell cultures adrenal medulla cells. As well as in other human epiHuman neuroblastoma cell lines (SK-N-SH, CHP-100, CHP-134, thelial tissues, the microanatomic distribution of Pgp GI-CA-N, GI-ME-N, IMR-32. Kelly, LA-N-5, NMB, SK-N-LO, suggests an involvement into the excretory function of SK-N-MC) were maintained in monolayer culture in RPMI 1640 adrenal cells. medium (Gibco, Scotland) supplemented with 10% fetal calf serum

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

Summary

1010 (a) survival fraction 100

(a) 100

slant? fiadKE(SiJ

16 50

18

20

22

24

26

28

fen.1 (nin)

(b) staining fiaakn(%) 100 _

survival fraction (%)

ao eo.

0.001

0.01

0.1

10

100

1000 10000 drug concentration (ng/ml)

Figure 1. Dose dependent survival fractions of cell lines SK-N-SH and CHP-100, (a) in the presence of vincristine, (b) in the presence of adriamycin.

Figure 2. Time-dependent rhodamine-123 accumulation in cells of cell lines SK-N-SH and CHP-100, (a) without Pgp-blocker, (b) in the presence of verapamil.

(Mannheim Boehringer, Germany), 2% L-glutamine (Mannheim Boehringer, Germany) and 0.2% glucose (Mannheim Boehringer, Germany). Controls for Pgp expression (gently provided by M.M. Gottesmann, NCI, Bethesda, Maryland, USA) included the drugsensitive parental carcinoma cell line KB-3-1 and the drug-resistant subline KB-8-5. Both cell lines were grown as multilayer cultures in Dulbecco's modified Eagle's medium (Gibco, Scotland) containing 10% fetal calf serum (Mannheim Boehringer, Germany) and 2% L-glutamine (Mannheim Boehringer, Germany). To prevent growth of revertants, cell line KB-8-5 was maintained continuously in the presence of colchicine (5ug/ml) (Sigma, Germany).

suspension were mixed with 5 ul of rhodamine-123 (Sigma, Germany) to a final concentration of 1.8 ug/1 and 5 ul of propidiumiodidemonohydrate (Sigma, Germany) to a final concentration of 30 ug/ml. After pipetting onto a microscopic slide, rhodamine-123 and propidiumiodide-monohydrate staining was evaluated at a fluorescence excitation of 365 nm. Every two minutes all rhodamine-123 stained and all rhodamine-123 non-stained cells of a constant viewfield were counted. Propidiumiodide-monohydrat staining allowed the exclusion of false-positive rhodamine-123 staining of dead cells. The rhodamine-123 staining fraction was determined as the percentage of rhodamine-123 stained cells compared with the number of all cells in the observed view-field. The period of time, that was necessary to attain a 50% staining fraction was calculated by linear regression with its 95% CI (due to the linear association of time and stained fraction of cells, see Figure 2). The described rhodamine-123 accumulation assay was performed for each cell line in the absence as well as in the presence of verapamil (Knoll AG, Germany) in a final concentration of20ul/ml.

Chemosensitivity of cell lines SK-N-SH and CHP-100 Cell lines SK-N-SH and CHP-100 were plated in cell culture flasks (Becton Dickinson, Germany) in growth medium at a density of 2 x 105 cells per flask. Twenty-four hours after plating growth medium was changed and supplemented with vincristine (Lilly Deutschland, Germany) or adriamycin (Farmitalia, Germany) in logarithmically increasing doses. After 96 hours of growth in the presence of cytostatics, cells were harvested with EDTA-trypsin (Sigma, Germany) and counted. The survival fraction was estimated as the percentage of surviving cells compared with the number of cells in untreated cultures. The 50% inhibition dose of cell growth (LD-50) was calculated with its 95% confidence interval (CI) by a non-linear regression model for each cell line and each cytostatic drug (due to the non-linear function of dose-response-curves, see Figure 1). The relative resistance of cell line SK-N-SH to one of the used drugs was estimated as the ratio of cell line specific LD-50 values.

Neuroblastoma tumor samples

Tumor touch smears and tumor infiltrated bone marrow smears of 62 patients suffering from histologically diagnosed neuroblastoma were provided by children's hospitals collaborating in the German Neuroblastoma Study Group. Criteria for diagnosis and response met the INSS guidelines [17]. The patients were treated according to the German Neuroblastoma Study Group Protocols NB 85 and NB 90.

Immunocytochemical staining procedure for N-myc oncoprotein

Rhodamine-123 accumulation in cell lines SK-N-SH and CHP-100 Rhodamine-123 is a fluorochrome underlying the transport mechanism of Pgp. Therefore rhodamine-123 is considered as a marker substance of Pgp function. An accumulation assay was performed with modifications to a method of Efferth [16]. Briefly, 10 ul of a cell

Tumor touch smears and bone marrow smears were fixed in paraformaldehyd (2%) for 10 minutes followed by permeabilizalion of the cell membrane with Triton X (0.5%) (Sigma, Germany) for 10 minutes. Primary antibody MAK B8.4.B (gently provided by Prof. Dr. Schwab, DKFZ. Heidelberg, Germany) was applied in a final concentration of 80 ug/ml for 60 minutes at room temperature. A routinely used

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

(b) 100

500 5000 drug concentration (ng/ml)

1011

,9 .8 .7j

KB-8-5 cells are characterized by a three-fold relative resistance to adriamycine and a six-fold relative resistance to vinblastine.

Pgp-status: negative (n = 34) positive (n=10)

P-glycoprotein expression in neuroblastoma tumor samples statistical analysis

,5. ,4' ,2.

OA 0

200

400

600

800

1000

1200

time (days)

Associations of the Pgp status and known prognostic factors were tested for statistical significance by chi-square test. Due to the impact of tumor spread on the outcome of patients with neuroblastoma the statistical analysis of a prognostic relevance of Pgp expression was performed stage-related. The probability of relapse free survival by the Pgp status was evaluated by Kaplan-Meier analysis. The statistical difference of the curves was tested by the log-rank test.

Results streptavidin biotin peroxidase system (Amersham, England) was used for visualisation of primary antibody binding.

Immunocytochemical staining procedure for P-glycoprotein The preparation of tumor touch smears, tumor infiltrated bone marrow smears and cytospin preparations of cell lines included fixing in 10% buffered formalin for two hours, permeabilization of the cell membrane by dehydration in graded ethanol, incubation in xylene for 30 minutes and rehydratation in graded ethanol. Blocking of endogenous peroxidase was performed by a 10 minutes exposure to a mixture of 70% ethanol and 30% methanol supplemented with hydrogen peroxide in a final concentration of 3% followed by a 20 minute treatment with a hydrous solution of 3% hydrogen peroxide, 0.35% sodium azide and 0.15% periodic acid. Monoclonal antibody C-219 (Centocor Inc., USA) was used for immunocytochemical staining of Pgp. C-219 detects a highly conserved amino acid sequence of the Pgp isoforms I and III that resides in a cytoplasmic domain on the inner side of the cell membrane [18-20]. C-219 was applied at a concentration 1.2 ug/ml overnight in a humidified chamber at 4 °C. The initial layer of the used immunosandwich consisted of a peroxidase-conjugated monoclonal goat anti-mouse IgG antibody (Sigma, Germany) in a dilution of 1:100 in combination with a biotinylated monoclonal goat anti-mouse IgG-2a antibody (Amersham, England) in a dilution of 1:100. In a second step a mouse peroxidase-anti-peroxidase complex (Amersham, England) was applied in a dilution of 1:50. The immunosandwich was completed by an incubation with a mixture of peroxidase-conjugated monoclonal goat anti-mouse IgG antibody (Sigma, Germany) in a dilution of 1:50 and a peroxidase-conjugated streptavidin biotin complex (Amersham, England) in a dilution of 1:100. All different steps of the immunosandwich were performed in a humidified chamber at 4 °C for 30 minutes. Altogether the used immunosandwich technique theoretically resulted in a ratio of seven peroxidase molecules per one Pgp molecule. This appeared necessary to detect that small differences in Pgp expression like those between KB-8-5 and KB-3-1 cells. Peroxidase activity was visualized by incubation with 3.3'diaminobenzidine tetrahydrochlo-ride at a concentration of 0.5 mg/ml in 0.05 M tris(hydroxymethyl)aminomethane buffer (pH 7.6) and 0.015 hydrogen peroxide for 10 minutes at room temperature. Improvement of optical density of diaminobenzidine-polymers was performed with modifications to a method of Newman [21]: Cells were incubated with neutralised sodium sulphide (In) for five minutes. This step was followed by an exposure to a silver developing solution. This consisted of a hydrous solution of sodium carbonate, ammonium nitrate, silver nitrate, tungstosilicic acid and formaldehyde. The deposition of metallic silver to diaminobenzidine was stopped after 90 seconds by immersing the sections in 1% acetic acid. Finally, cells were counterstained with Meyer's Hamalaun (Merck, Germany). Staining comparable to KB-8-5 cells was considered positive for Pgp expression, non-staining comparable to K.B-3-1 cells considered as negative.

Pgp expression in neuroblastoma cell lines Pgp expression was detected only in one out of eleven neuroblastoma cell lines (SK-N-SH). This line and CHP 100 were used exemplary for further experiments. The N-myc expression was found in all six cell lines with known N-myc amplification and additionally in SK-NMC and CHP-100. Chemosensitivity of cell lines SK-N-SH and CHP-100 Neuroblastoma cell line SK-N-SH and neuroepithelioma cell line CHP-100 showed a dose-dependent inhibition of cell growth in the presence of vincristine (Figure la) and adriamycin (Figure lb). The LD-50 values for both cell lines differed statistically significant under the influence of vincristine (11.3 ng/ml for SK-N-SH versus 0.3 ng/ml for CHP-100) and adriamycin (141.3 ng/ml for SK-N-SH versus 11.2 ng/ml for CHP-100). Thus, SK-N-SH cells showed a 45.1-fold relative resistance to vincristine and a 12.7-fold relative resistance to adriamycin compared to CHP-100 cells. Rhodamine-123 accumulation in cell lines SK-N-SH and CHP-100 Cell lines SK-N-SH and CHP-100 showed a continuous, time dependent increase in rhodamine-123 stained cells (Figure 2a). A 100% staining was observed in CHP-100 cells after 18 minutes, the 50% value was 7.0 (6.6-7.4) minutes. The presence of verapamil did not change these characteristics (50% at 7.3 (6.9-7.7) minutes, Figure 2b). SK-N-SH cells showed a 100% staining after 28 minutes and 50% at 12.1 (11.7-12.5) minutes. In the presence of verapamil the accumulation of rhodamine was significantly accelerated (50% value at 8.0 (7.6-8.5) minutes. This resulted in now equal 50% values for both cell lines. P-glycoprotein expression in human neuroblastoma tumors Sixty-eight tumor samples of 62 patients suffering from histologically diagnosed neuroblastoma were evaluated

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

Figure 3. Kaplan-Meier analysis for event free survival in stage 4 neuroblastoma by the Pgp status.

1012 The probability of event-free survival dependent on the Pgp status was estimated with regard to tumor stage. Since case numbers were small for all other tumor stages, Kaplan-Meier analysis was performed for stage 4 tumors only. The analysis of event-free survival for the Pgp negative group resulted in a median survival time of 986 days (95% CI: 571-1401 days) versus 478 days (95% CI: 274-681 days) for the Pgp positive group. Comparison of the estimated probabilities for event-free survival demonstrated a small trend of superior survival for the group of patients with Pgp negative tumors (log-rank test, P = 0.069). Within the group of stage 4 neuroblastoma Pgp status was correlated with several clinical data considered as independent prognostic factors by the German Neuroblastoma Study Group [22]. There was no statistical significant correlation of Pgp expression and any of these prognostic factors including age, N-myc amplification status, serum ferritin level, serum lactate dehydrogenase activity, initial white blood count and resectability of the tumor. Discussion The multidrug resistant phenotype of SK-N-SH cells P-glycoprotein is the molecular hallmark of classical multidrug resistance. Occurrence of Ppg in human neuroblastoma cell lines including SK-N-SH has been reported [23]. However, Pgp expression in neuroblastoma cell lines has not been correlated to drug resistance so far. Immunocytochemistry confirmed the qualitative different Pgp expression of SK-N-SH and CHP-100 cells. Cell growth under the influence of adriamycin and vincristine revealed a relative resistance of SK-N-SH cells in comparison to cell line CHP-100 for both drugs. Intracellular accumulation of rhodamine-123, a drug underlying the Pgp-dependent efflux transport, was diminished for cell line SK-N-SH. The latter phenomenon was reversible in the presence of verapamil, known to compete for the Pgp-binding site [5]. Thus, we demonstrated in SK-N-SH cells an association of Pgp expression and the phenotype of multidrug resistance including reduced intracellular accumulation of drugs and increased resistance to cytostatics. Two different transmembrane transport mechanisms may underlay this observation: Pgp and the multidrug resistance-associated protein (MRP) [23, 24]. In contrast to Pgp, verapamil does not interfere with the MRP-associated transmembrane transport [25, 26]. Since the reduced intracellular accumulation of rhodamine-123 was reversible by verapamil, a Pgp-dependent multidrug resistance phenotype of SK-N-SH cells can be assumed. P-glycoprotein expression in human neuroblastoma tumors

Pgp expression in human human neuroblastoma tumors has been reported controversial: Occurrence of Pgp has

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

for Pgp expression. Forty-five tumor samples resulted from diagnostic procedures at the time of diagnosis, 12 from diagnostic procedures during tumor progress or tumor recurrence, four samples were obtained during delayed surgery, one at second look surgery. Tumor specimens of 48 patients did not show any immunoreactivity for Pgp (77.4%). Staining for Pgp was observed in tumor samples of 14 patients (22.6%). The staining intensity was comparable to the staining reactivity of control cell line KB-8-5 in all tumors. The positive staining was restricted to tumor and not seen in stroma cells. No differences were found in the Pgp-status between primary tumors and metastases (three cases) and in consecutive samples before and after chemotherapy (two cases, both stage 4). Pgp expression was detectable in non-localized as well as in localized tumors. The frequency of Pgp positive tumors was 22% for non-localized (stage 4: 10 of 44; 4S: 1 of 5) and 23% for localized tumors (3 of 13). The two patients with Pgp positive stage 1 tumors underwent surgical resection of the tumors only and were in complete remission after a follow-up of five and 18 months, respectively. The patient with the Pgp positive stage 2 tumor received in addition to surgical treatment chemoand radiotherapy and was in complete remission after a follow-up of 16 months. The patient with the Pgp positive stage 4S tumor achieved a short remission status after surgical tumor resection. Further on the tumor progressed to stage 4 (N-myc amplification status not known). Pgp was expressed in tumors of all histopathological grades with nearly the same incidence (grade 1: 4 of 15 (27%), grade 2: 3 of 15 (20%), grade 3: 5 of 21 (24%), P = 0.82; n.s.). Also no correlation was observed between Pgp status and primary tumor site. Pgp expression occurred in tumors arising from the adrenal medulla as well as in tumors arising from abdominal and mediastinal ganglia. Out of 62 tumors, in 57 N-myc oncogene expression was analyzed. The incidence of Pgp expression was similar for N-myc positive (4 of 17; 24%) and N-myc negative (11 of 40; 28%) tumors (P - 0.76; n.s.). Data on the N-myc oncogene amplification status were available for 44 tumors. N-myc expression was found in all N-myc amplified and additionally in 7/36 (19.4%) non-amplified tumors. Pgp expression was not associated with N-myc amplification (P - 0.20). Forty-nine tumor samples were taken before chemotherapy. Ten of these tumors demonstrated Pgp expression (20.4%). Thirteen tumor samples were collected after chemotherapy. Five of these tumors were characterized by detectable levels of Pgp (38.5%) (P = 0.19). Of 22 patients achieving complete remission during chemotherapy, six children (27.3%) had Pgp positive tumors. In the group of patients with partial response only one out of 12 tumors demonstrated Pgp expression. Two patients showed no response to the therapeutic approaches including first-line chemotherapy. None of these tumors was characterized by Pgp expression.

1013

P-glycoprotein expression and tumor cell differentiation Pgp expression was investigated for correlation with the main biological characteristics of tumor evolution: histological grading and stage of disease. The incidence of Pgp positive tumors ranged from 20% to 30% without any statistical difference within all three different types of histopathological grading. No significant association was observed between Pgp status and stage of the disease: 12 out of 53 non-localized neuroblastoma showed elevated levels of Pgp (22,6%), three out of nine localized tumors were demonstrated to express Pgp (33%). In contrast to other previous studies Nakagawara reported the occurrence of Pgp expression in well differentiated, localized neuroblastoma33. Hereby we confirm this observation. However, Pgp expression was neither restricted to nor its incidence increased within the group of stage 1-3 neuroblastoma. Furthermore, there was no statistical significant correlation of Pgp expression with N-myc oncogeneprotein expression as well as with N-myc oncogene amplification. Therefore, our results suggest, that Pgp expression per se does not indicate a high grade of tumor cell differentiation. To our knowledge, it is the first time that a stage 4S tumor was demonstrated to express Pgp at an elevated level. However, in this case the course of the disease ended after a short remission status with a progress into stage 4 (N-myc amplification status unknown). P-glycoprotein expression and classical multidrug resistance Pgp expression did not correlate with the patients remission status after first line chemotherapy: We found no evidence that Pgp expression might be involved in mechanisms causing primary chemotherapy resistance as it has been reported by Chan et al. [11].

The incidence of Pgp positive tumors was not different between tumors treated by chemotherapy and the group of untreated tumors. This oberservation contrasts with previous reports [11, 29]. Furthermore, tumor samples of two patients tested for Pgp expression at different times of the clinical course showed no qualitative change of the initial negative status under or after chemotherapy. Therefore, our results do not confirm, that Pgp expression evolves by time with the use of cytostatics in the sense of cellular salvage pathway. In stage 4 patients Kaplan-Meier analysis for eventfree survival revealed a small trend of inferior survival of patients with Pgp positive tumors (P = 0.069). Taking into account, that the Pgp status showed no statistical correlation to any other risk factor, our results suggest, that occurrence of elevated Pgp levels might slightly worsen the outcome of children with stage 4 neuroblastoma. However, we cannot conclude that occurrence of Pgp predicts the success or failure of therapy for stage 4 neuroblastoma in contrast to the study of Chan et al. [11]. Therefore it remains questionable whether the detection of Pgp justifies the use of chemosensitizing agents capable of reverting Pgp mediated multidrug resistance. Other cellular mechanisms have been demonstrated to be of relevance for neuroblastoma drug resistance including MRP [24] and further membrane transport proteins [14, 25, 34]. For example, Norris et al. [24] found 57% five-year survival in patients with high compared to 94% with low MRP expression (P < 0.001). In conclusion, based on our results chemotherapy resistance appears to be a complex process and Pgp expression does not provide much information on this phenomenon in'neuroblastoma. Acknowledgement This work was supported by a grant of the German Cancer Aid (Deutsche Krebshilfe,T 7/93/Be 4).

References 1. Juliano RL, Ling V. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochim Biophys Acta 1976; 455: 152-62. 2. McGrath JP, Varshavsky A. The yeast STE6 gene encodes al homologue of the mammalian multidrug resistance P-glycoprotein. Nature 1989; 340: 400-4. 3. Cornwell MM, Pastan I, Gottesmann MM. Binding of drugs and ATP by P-Glykoprotein and transport of drugs by vesicles from human multidrug resistant cells. In Roninson IB (ed): Molecular and Cellular Biology of Multidrug Resistance in Tumor Cells 1991. 4. Bellamy WT, Dalton WS, Dorr RT. The clinical relevance of multidrug resistance. Cancer Invest 1990; 8: 547-62. 5. Salmon SE, Dalton WS, Grogan TM et al. Multidrug-resistant myeloma: Laboratory and clinical effects of verpamil as a chemosensitizer. Blood 1991; 78: 44-50. 6. Sonneveld P. Durie BGM, Lokhorst HM et al. Modulation of multidrug-resistant multiple myeloma by cyclosporin. Lancet 1992; 340: 255-9.

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

been denied, attributed to cellular differentiation or linked up with classic multidrug resistance [10-15, 26]. In this study, we investigated tumor samples of 62 neuroblastoma for Pgp expression by a highly sensitive peroxidase-immunosandwich technique in order to detect as small differences of three- to six-fold relative resistance (KB-8-5 versus KB-1-3). By this way Pgp was detectable in 14 tumors. This confirms previous studies, which reported that Pgp expression belongs to the biological repertoire of human neuroblastoma tumor cells [10-15]. Thus, in vivo levels of Pgp in human neuroblastoma seem to be noticeably low suggesting that routinely used immuncytochemical techniques may fail to demonstrate typical levels of Pgp in human neuroblastoma cells [28] and explaining the results of a previous study, where Pgp was only demonstrated in tumor stroma but not in neuroblastoma cells [26]. In accordance to other studies that used highly sensitive methods for the detection of multidrug resistance gene mRNA or Pgp, the incidence of Pgp positive tumors in our study was 22.4% [11, 13, 29-34].

1014

24. 25. 26. 27. 28.

29. 30.

31. 32. 33. 34. 35.

by retinoic acid-induced differentiation. Mol Cell Biol 1989; 9: 4337-44. Norris MD, Bordow AB, Marshall GM et al. Expression of the gene for multidrug-resistance-associated protein and outcome in patients with neuroblastoma. N Engl J Med 1996; 334: 231-8. Cole SPC, Bhardwaj G, Gerlach JH et al. Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line. Science 1992; 258: 1650-3. Favrot M, Combaret V, Goillot E et al. Expression of P-glycoprotein restricted to normal cells in neuroblastoma biopsies. Br J Cancer 1991; 64: 233-8. Shen DW, Fojo A, Chin JE et al. Human multidrug-resistant cell lines: increased mdrl expression can precede gene amplification. Science 1986; 232: 643-5. Chan HSL, Bradley G, Thorner P et al. A sensitive method for immunocytochemical detection of P-glycoprotein in multi-drugresistant human ovarian carcinoma cell lines. Lab Invest 1988; 59: 870-5. Bourhis J, De Vathaire F, Wilson GD et al. Combined analysis of DNA ploidy index and N-MYC genomic content in neuroblastoma. Cancer Res 1991; 51: 33-6. Corrias VM, Cornaglia-Ferraris P, Di Martino D et al. Expression of multiple drug resistance gene, MDR1, and n-myc oncogene in an Italian population of human neuroblastoma patients. Anticancer Res 1990; 10: 897-902. Goldstein LJ, Galski A, Fojo A et al. Expression of a multidrug resistance gene in human cancers. J Natl Cancer Inst 1989; 81: 116-24. Goldstein LJ, Fojo AT, Ueda K. et al. Expression of the multidrug resistance, MDR-1, gene in neuroblastomas. J Clin Oncol 1990; 8: 128-36. Nakagawara A, Kadomatsu K, Sato S et al. Inverse correlation between expression of multidrug resistance gene and N-myc oncogene in human neuroblastomas. Cancer Res 1990; 51' 3043-7. Ramani D, Dewchand H. Expression of mdrl/P-glycoprotein and pllO in neuroblastoma. J Pathol 1995; 175: 13-22. Reynolds CP, Tomayko MM, Donner L et al. Biological classification of cell lines derived from human extra-cranial neural tumors. Prog Clin Biol Res 1988; 271: 291-306.

Received 11 May 1998; accepted 7 July 1998. Correspondence to:

F. Berthold, MD Department of Pediatric Hematology and Oncology University of Cologne Joseph-Stelzmann-Str. 9 D50924 Cologne Germany E-mail: [email protected]

Downloaded from https://academic.oup.com/annonc/article-abstract/9/9/1009/222346 by guest on 06 September 2018

7. Biedler JL. Genetic aspects of multidrug resistance. Cancer 1992; 70: 1799-809. 8. Roninson IB. Structure and evolution of P-glykoproteins. In Roninson IB (ed): Molecular and Cellular Biology of Multidrug Resistance in Human Tumor Cells 1991. 9. Berthold F. Overview: Biology of Neuroblastoma. In Pochedly C (ed): Neuroblastoma: Tumor Biology and Therapy 1990. 10. Bates ES, Shieh CY, Tsokos M. Expression of mdr-1/P-glycoprotein in human neuroblastoma. Am J Pathol 1991; 139: 305—15. 11. Chan HSL, Haddad G, Thorner PS et al. P-glycoprotein expression as a predictor of the outcome of therapy for neuroblastoma. N Engl J Med 1991; 325: 1608-14. 12. Cordon-Cardo C, O'Brien JP, Boccia J et al. Epxression of the multidrug resistance gene product (P-Glycoprotein) in human normal and tumor tissues. J Histochem Cytochem 1990; 38: 1277—87. 13. Haddad G, Thorner PS, Bradley G et al. A sensitive multilayer immunoalkaline phosphatase method for detection of leukemic and tumor cells in the bone marrow. Lab Invest 1994; 71: 595-603. 14. Oda Y, Rose I, Radig K et al. Expression of MDR1 /p-glycoprotein and multidrug resistance-associated protein in childhood solid tumors. Virchows Arch 1997; 430: 99-105. 15. Obana K, Hashizume K. Expression of multidrug-resistance related P-glycoprotein shows good prognosis in neuroblastoma. J Pediatr Surg 1997; 32: 420-2. 16. Efferth T, Lohrke J, Volm M. Reciprocal correlation between expression of P-Glycoprotein and accumulation of Rhodamine 123 in human tumors. Anticancer Res 1989; 9: 1633-8. 17. Brodeur GM, Seeger RC, Barret A et al. International criteria for diagnosis, staging, and response to treatment in patients with neuroblastoma. J Clin Oncol 1988; 6: 1874-81. 18. Georges E, Bradley G, Gariepy J, Ling V. Detection of P-glycoprotein isoforms by gene-specific monoclonal antibodies. Proc Natl Acad Sci USA 1990; 87: 152-6. 19. Kartner N, Evernden-Porelle D, Bradley G, Ling V. Detection of P-glycoprotein in multidrug resistant cell lines by monoclonal antibodies. Nature 1985; 316: 820-3. 20. Meyers MB, Rittmann-Grauer L, O'Brien JP, Safa AR. Characterization of monoclonal antibodies recognizing a M 180.000 P-glycoprotein: Differential expression of the M 180.000 and M 170.000 P-glycoproteins in multi-drug resistant human tumor cells. Cancer Res 1989; 49: 3209-14. 21. Newman GR, Jasani B, Williams ED. The visualisation of trace amounts of diaminobenzidine (DAB) polymer by a novel goldsulphide-silver method. J Microsc 1983; 132: RP1-2. 22. Berthold F. Gesellschaft fur Padiatrische Onkologie - Multizentrische Therapiestudie zur Behandlung von ICindern und Jugendlichen mit Neuroblastom (Neuroblastomstudie NB90), Studienprotokoll, 1990. 23. Bates SE, Mickley LA, Chen Y-N et al. Expression of a drug resistance gene in human neuroblastoma cell lines: Modulation