Original Paper Oncology I992;49:203- 208
Anne O'M eara3 A t sushi Imamuraa Pauline Johnson3 Raymond Balla Sean Rooneya Barry Kiercea Takashi Tsuruob Peter Dervanc
Reactivity of P-Glycoprotein Monoclonal Antibodies in Childhood Cancers
* Children's Research Centre, Our Lady's Hospital for Sick Children, Dublin. Ireland: 11 Division o f Experimental Chemotherapy, Cancer Chemotherapy Center, Japanese Foundation for Cancer Research, Tokyo, Japan; e Department of Pathology, Mater Misericordiae Hospital. Dublin, Ireland
Introduction The overall long-term disease-free survival of 60-65% in childhood cancer is attributable mainly to chemosensitivity of most paediatric tumours. Some patients, how'-
ever. become resistant to a w'ide range of chemotherapeut ic agents after an initial response and eventually die. 1n ad dition, a small percentage of tumours fail to respond to standard, albeit intensive, chemotherapy ab initio. This has prompted researchers to investigate the various me-
L)r. Anne O'M eara Children's Research Ccnlre Our Lady’s I lospital lor Sick Children Crumlin. Dublin 12 (Ireland)
< 1992 S. Karger AG. Basel 0030 2414 92 0493 0203 S 2.75 0
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
KeyWords Chemoresistance Childhood Cancers P-Glycoprotein JSB-I MRK16
Abstract P-Glycoprotein (P-gp), the product of the mdr-1 gene, is implicated in the de velopment of chemoresistance in a variety of, mostly adult, cancers. Its role in paediatric tumours, most of which arc non-epithelial in origin, has yet to be fully elucidated. A study was undertaken to investigate reactivity of two P-gp monoclonal antibodies (M Abs). JBS-I and M RK16, recognising cytoplasmic and surface epitopes, respectively, of the P-gp molecule, in a variety of newly diagnosed and relapsed childhood cancers. P-gp was not expressed in any of 36 tumours examined (neuroblastoma 13, nephroblastoma 12, rhabdomyosar coma 6, lymphoma 3. teratoma 1. Ewings 1), 14 of whom had chemoresistant disease. Reactivity to both MAbs was also investigated in patients with acute leukaemia. Out of 10 diagnostic acute lymphoblastic leukaemia (ALL) sam ples, a positive reaction with JSB-1was observed in 1patient who failed to remit on standard induction therapy and in 3 of 6 patients in ALL relapse, only 1 of whom showed low grade positivity with MRK16. Both MAbs reacted posi tively in 1 patient with acute non-lymphocytic leukaemia (AN I.L) at diagnosis who achieved remission with teniposideand cytosine arabi noside, but relapsed 7 months later and was again positive with both Mabs. JSB-1also showed vary ing degrees of positivity in 4 out o f4 other patients in ANLL relapse. It would therefore appear that P-gp is unlikely to mediate chemoresistance in most solid tumours of childhood, but may well play a major role in the development of chemoresistance in acute leukaemia.
chanisms of inherent and acquired chemoresistance in human cancers. Rodent and human multidrug-resistant cell lines have been established in vitro by multi-step drug selection [1,2], most of which have either amplification or over-expres sion of a gene termed mdr-1, which is transcribed into a 4.5-kb mRNA [3-5], the product of this gene being a 170kD membrane glycoprotein (P-gp) [6]. P-gp appears to act as an energy-dependent drug efflux pump which results in drugs being eliminated from cells before a cytotoxic effect can be exerted [7], A number of monoclonal antibodies (MAbs) have been raised to different epitopes on the P-gp molecule [8-10], and a pattern of reactivity has been esta blished for normal and neoplastic (mainly adult car cinoma) tissue [11], These authors have identified positiv ity with P-gp MAbs in normal epithelial cells making meaningful interpretation of various carcinomas (i.e. can cers arising from epithelial tissue) difficult. There is little information, on the other hand, on reactivity of these MAbs to paediatric tumours, most of which are non-epithelial in origin. The aim of this study was to screen a vari ety of childhood cancers with P-gp MAbs, thus enabling the role of this glycoprotein in the behaviour of childhood cancers to be established.
Results Normal kidney, adrenal, striated muscle, spleen, thy mus, lymph, lymph node and BM were screened with both M Abs: proximal tubules of kidneys reacted positively with both antibodies while glomeruli were negative. Adrenal cortex was strongly positive while medulla reacted weakly. Intensity of staining in all cases was greater with MRK16. Striated muscle, spleen, thymus, lymph node and BM were negative with both MAbs. Tumours
Samples Normal tissue and tum our samples from newly diagnosed un treated and partially treated or relapsed (grouped together as treated) patients were freshly frozen and stored at -195 C until analysis. The Adriamycin-resistant cell lines MCF-7Ad [ 12] and 2780-Ad [ 13] were used as positive controls. Bone marrow (BM) smears of patients with newly diagnosed and relapsed acute lymphoblastic leukaemia (ALL), acute non-lymphocytic leukaemia (ANLL), malignant disease with BM involvement and normal BM aspirates were fixed in acetone, air-dried and stored at -20 C until analysis. M ononuclear preparations of BM samples were processed immediately.
Monoclonal Antibodies The IgGI MAb JSB-1 [8] and the IgG2A MAb M R K I6[9j re cognising cytoplasmic and surface epitopes, respectively, of P-gp were employed. JSB-1 was used at 1:50 dilution of ascitic fluid and M R K 16 at a concentration of 4 pg/ml.
Methods JSB-1 reactivity was measured on BM smears and cytospin prepa rations using an APA AP technique; M RK. 16 reactivity of viable cells was measured by an indirect method using FITC-conjugated rabbit anti-mouse IgG F(ab'), fragment and analysed by FACScan (Becton Dickinson) flow cytometry. Cell lines were also analysed by fluores
Tumour samples from 36 patients in total were screened (table 1). Most tumours were screened with JSB-1 and M RK 16 simultaneously; both MAbs gave identical results with one exception, an untreated rhabdomyosarcoma which showed diffuse positivity with JSB-1, but was nega tive with MRK16 (table 1). This patient was sensitive to chemotherapy which included vincristine, actinomycin D and cyclophosphamide and remains off treatment with no evidence of disease 3 years from diagnosis. All other tu mours, including those with residual or recurrent disease following chemotherapy (both grouped together as ‘treat ed’ in table 1), were negative. Chemotherapy for children with neuroblastoma (NBL) had included vincristine, cyclophosphamide, cisplatinum and an epipodophyllotoxin, as per European Neuroblastoma Study Group protocol, while patients with Wilms’ tumour (WT) had re ceived vincristine, actinomycin D and Adriamycin, ac cording to the United Kingdom Children’s Cancer Study Group protocol. A consistent observation however was the strong degree of positivity observed in endothelial cells of blood vessels which was most apparent in NBL tu mours.
O'Meara/Imamura/Johnson/Ball/Rooney/ Kierce/Tsuruo/Dervan
P-gp in Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Materials and Methods
204
cence microscopy in order to overcome any misinterpretation o f data by flow cytometry due to cell clumping. 5-pm sections o f freshly frozen tissue were cut using a cryostat microtome. Sections were fixed in acetone at 4 C for 10 min, then flooded with 10% species-specific normal serum to minimise non specific antibody binding. Endogenous peroxidase activity was inac tivated by 0.3% hydrogen peroxide in methanol for 30 min. Sections were then incubated with primary antibody in a moist chamber at 4 C for 16 h. The streptavidin-biotin complex (Dakopatts) method was applied for immunohistochemistry. Second antibody was bioti nylated F(ab'), fragment rabbit anti-mouse IgG (Dakopatts). Sec tions were developed in 0.025% diaminobenzidine solution. Spec ificity of immunohistochemical staining was confirmed by control staining with non-immune mouse serum. All slides were counterstained with 1% methyl green.
Leukaemias The Adriamycin-resistant cell lines MCF-7Ad and 2780-Ad reacted strongly with both MAbs (fig. 1). JSB-1 andM RK 16 reactivity was examined on BM aspirates of patients with ALL and AN LL (table 2). In the ALL group, a diagnostic aspirate from 1 patient with T cell ALL who failed to remit on standard induction therapy, consisting of vincristine, Adriamycin, L-asparaginase and pred nisolone, as per a modified ALL-BFM 81 protocol [14], showed >50% positivity with JSB-1, but was negative with MRK16; this patient eventually achieved haematological remission with alternative chemotherapy, which included amsacrine, etoposide and cytosine arabinoside. Three out of 6 patients in BM relapse, all of whom de veloped BM relapse on maintenance treatment consisting of mercaptopurine and methotrexate, reacted with one or both Mabs: Patient 1 was negative with both antibodies on BM re lapse, failed to achieve second remission with induction therapy as stated above, and showed up to 50% positivity in BM with JSB-1 prior to death; 5% of lymphoblasts reacted with MRK16. Patient 2 was similarly negative on BM relapse, failed to achieve second remission also with the above-mentioned induction therapy, which was later changed to etoposide
Table 1. Summary of childhood solid tumours screened
Tumour
P-gP positive
p-gp negative
Chemoresistant disease
0 0
6 7
1 5
0 0
7 5
2 4
la 0
4 1
1 0
0 0
2 i
0 1
0 0
i i
0 0
WT Untreated (7) Treated (5)
Rhabdomyosarcoma Untreated (5) Treated (1)
Marrow status
JSB-1
MRK.16
9(1 H) 13(0) 6( 1H. 2M)
7(0) 8(0) 3 (1L)
ALL Diagnosis (n = 10) Remission (n = 14) Relapse (n = 6)
ANLL Diagnosis (n = 1) Remission (n = 1) Relapse (n = 5)
1(H) 1(0) 5 (2H. 2M. 1L)
1(H) 1(0) 2(1 H)
Miscellaneous
Lymphoma Untreated (2) T reated ( 1)
Fig. 1. JSB-1 staining of MCF-7Ad cell line using an APAAP technique (63 x 1 .2 5 x 12.5). Note variable intensity of staining among cells which is maximal at cytoplasmic aspect o f cell mem brane.
Table 2. BM reactivity
NBL Untreated (6) Treated (7)
and cytosine arabinoside, and survived in relapse for 6 months; 50% of lymphoblasts reacted with JSB-l and 17% with M R K 16, 1 months prior to death. Patient 3 was also negative with both Mabs on first BM relapse; a second remission was achieved with etoposide
N B L inB M
- Diagnosis (n = 3) - Relapse (n = 1) Hodgkins positive BM (n = 1) ITP (n = 1) Infection (n = 2) Normal (n = 17)
2(0) 1(H) 1(0) 0 0 13(0)
3(0) 0 0 1(0) 2(0) 6(0)
Miscellaneous
a
Numbers in parentheses refer to the total number o f patients. JSB-l positiveand MRK16 negative.
n = Total number of patients; ITP = idiopathic thrombocytopaenic purpura. Number in parentheses = number of patients with positive reaction graded as H = > 50% cells positive; M = 20-50%, and L = 10-20%.
205
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Teratoma: untreated (1) Ewing’s: untreated (1)
and cytosine arabinoside. but a second BM relapse oc curred 4 months later. JSB-1 at that time reacted with more than 50% of lymphoblasts (fig. 2) and M R K 16 was nega tive. Third BM remission was not achieved. Among the ANLL patients, only one diagnostic as pirate was available for analysis; a high degree of positivity was noted with both M Abs (fig. 3). This patient w ho had trisomy 21 and an M2 (FAB classification) morphology, achieved haematologic remission with a teniposide and cytosine arabinoside induction schedule; remission BM
was negative w'ith both MAbs. Seven months from diag nosis, while still on a maintenance schedule of teniposide and cytosine arabinoside, she developed a BM relapse; JSB-1and M R K 16 demonstrated strong positivity at that time. Four other ANLL relapses demonstrated varying degrees of positivity with JSB-1, only 1 of which was also screened with MRK 16 and was negative. All of these pa tients had chemoresistant disease. One patient with relapsed NBL showed 80% positivity of NBL cells in BM with JSB-1; MRK 16 analysis was not performed.
Discussion
Fig. 2. JSB- 1 reactivity in relapsed ALL (63 x 1.25 x 12.5). In tense staining observed in scanty cytoplasm of lymphoblasts (1).
It would appear from the above data that expression of P-gp is, at best, an infrequent phenomenon in both NBL and nephroblastoma (WT). Epithelial cells within the kid ney, especially proximal tubules, demonstrated strong reactivity with both MAbs, but WT, either untreated diag nostic material or chemoresistant relapsed tissue, was con sistently negative. The pattern of reactivity in adrenal gland was similar to what Sugawara et al. [ 15] have report ed, i.e. homogeneous staining of cortex with patchy stain ing of medulla. All NBL tumours however, including those with chemoresistant disease, were negative with both antibodies; this observation is in agreement with others [16], Phaeochromocytoma, another tumour of adrenal medulla, has also been found to be negative [11,
3oon
206
O’Meara Imamura Johnson Ball Rooney Kierce/Tsuruo Dervan
P-gp :n Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Fig. 3. Flow cytometric analysis (FACScan)of MRK 16 reactivity at diagnosis in I patient with ANLL (shaded peak); clear peak represents FITC-conjugated rabbit antimousc IgG F(ab')2control.
16]. Increased mdr-1 transcript levels have however been reported in relapsed NBL [17, 18]; one would therefore have expected increased mRNA would imply more pro tein expression, but this is evidently not the case. Endo thelial cells of blood vessels, on the other hand, showed marked positivity and may well act, in this situation, by creating pharmacological sanctuaries within these tu mours as others have reported [19], Most rhabdomyosar comas in this study were investigated at diagnosis, most of whom had chemosensitive disease. Van der Valket al. [11] have reported weak positivity ofM RK16,JBS-landC219 with striated/skeletal muscle, which would explain the fal se-positive result obtained with one chemosensitive tu mour to JSB-1. Chan et al. [20] have observed statistically significant differences in P-gp expression in soft tissue sar coma of childhood between patients who relapsed and those who did not; of 9 patients in that series where P-gp was detected, 5 were negative at first presentation. The pattern of reactivity in leukaemia samples was par ticularly interesting. Reactivity of P-gp antibodies only de veloped with progression of disease, and JSB-1 positivity was observed earlier and to a greater extent than M R K 16. The fact that JSB-1 recognises a cytoplasmic epitope which is generated earlier in protein synthesis than the sur face epitope which M R K 16 recognises would explain this observation, and data from others [21] would support this theory. This may well explain the failure of Ito et al. [22] to demonstrate P-gp expression in acute leukaemia patients using M R K 16, while Ma et al. [23] showed positivity using C219 M Ab. Based on this data, it would appear advisable
to employ M Abs recognising both internal and cell surface epitopes before regarding samples as negative. Only 1 pa tient with ANLL was investigated at diagnosis and de monstrated a high degree of positivity with both MAbs. She achieved haematological remission with a teniposide, Cytosar induction schedule, but relapsed 7 months from diagnosis. While teniposide is one of the naturally occur ring agents susceptible to P-gp action, the synergistic ac tion with Cytosar may have initially overcome the other wise likely evolution of an mdr phenotype. Conversely, 3 patients with relapsed ALL did not react with either MAb. Clearly alternative mechanisms of chemoresistance are operational in some malignancies. In conclusion, this study has shown that in acute leu kaemias of childhood, P-gp may be a major mechanism in the development of multidrug resistance. There was no evidence for P-gp-mediated chemoresistance in even the most resistant of solid tumours. It may be that mdr transcript levels, rather than P-gp expression, are more re levant in the development of chemoresistance. Data from haematological studies however [22, 23], including this study, would appear to belie this suggestion. Larger scale studies, particularly with serial monitoring, are needed to address this issue. Acknowledgements We are grateful to F. Breatnach for permission to study his pa tients, R. J. Fitzgerald for providing tumour samples and O. Reddy for secretarial assistance.
References 5 Riordan JR. Deuchars K. Kartner N. et al: Amplification of P-glycoprotein genes in multidrug resistant mammalian cell lines. Nat ure 1985;316:816-819. 6 Ueda K. Cornwell M M, Gottesman M M. et al: The mdr-1 gene responsible for multidrug resistance codes for P-glycoprotein. Biochem Biophys ResCommun 1986;141:956-962. 7 Gottesman MM. Pastan I: The multidrug transporter, a double edged sword. J Biol Chem 1988;263:12163-12166. 8 Scheper RJ, Bulte JWM, Brakke JG P. et al: Monoclonal antibody JSB-1 detects a highly conserved epitope on the P-glycoprotein asso ciated with multidrug resistance. Int J Cancer 1988;42:389-394.
9 Hamada H. Tsuruo T: Functional role for the 170- to 180-kDa glycoprotein specific todrugresistant tumor cellsas revealed by monoclonal antibodies. Proc Natl Acad Sci USA 1986:83: 785-7789. 10 Kartner N. Fverndcn-Porclle D. Bradley G .ct al: Detection of P-glycoprotein in mdrcell lines by monoclonal antibodies. Nature 1985:316: 820-823. 11 Vandcr Valk P, Van KalkenCK, Ketelaars II. et al: Distribution of multidrug resistance asso ciated P-glycoprotein in normal and neoplastic human tissues. Ann Oncol 1990; 1:56—64.
207
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
1 Juliano RJ, Ling V: A surface glycoprotein modulating drug permeability in Chinese ham ster ovary cell mutants. Biochcm Biophys Acta 1976;455:152-162. 2 FojoA,AkiyamaS.Gottesman M M.PastanI: Reduced drug accumulation in multiple drugresistant human KB carcinoma cell lines. Can cer Res 1985:45:3002 3007. 3 Roninson IB. Chin JE, Choi KG. et al: Isola tion of human mdr cDNA sequences amplified in multidrug resistant KB carcinoma cells. Proc Natl Acad Sci USA 1986;83:4538-4542. 4 Scotto KW, Bicdler JL. Melera PW: Am plification and expression of genes associated with multidrug resistance in mammalian cells. Science 1986;232:751-755.
16 Cordon-CardoC, O'Brien JP, BertinoJ.etal: Expression of the human multidrug resistance gene (P-glycoprotein) in human tumours: An Immunopathologie study. US Can Acad Pa thol 1990;22A:124. 17 Bourhis J, Bcnard J, Hartman O, et al: Correla tion of mdrl gene expression with chemo therapy in neuroblastoma. J Natl Cancer Inst 1989;81:1401-1405. 18 Fojo AT, Ueda K. Slamon DJ, et al: Expres sion of a multidrug resistance gene in human tumours and tissues. Proc Natl Acad Sci USA 1987;84:265-269. 19 Cordon-CardoC, O'Brien JP, Casals D, et al: Immunoanatomic and immunopathologie ex pression of the multidrug resistance gene pro duct. Cancer Cells 1989;7:87-93.
20 Chan HSL, Thorner PS, Haddad G, et al: Immunohistochemical detection of P-glyco protein: Prognostic correlation in soft tissue sarcoma of childhood. J Clin Oncol 1990;8: 689-704. 21 Spier CM, List AF, Cline A, et al: Drug resis tance in the myelodysplastic syndromes. US Can Acad Pathol 1990;95A:557. 22 Ito Y, Tanimoto M, Kumazawa T, et al: In creased P-glycoprotein expression and multidrugresistant gene (mdrl ) amplification are in frequently found in fresh acute leukaemia cells. Sequential analysis of 15 cases at initial pre sentation and relapsed stage. Cancer I989;63: 1534-1538. 23 Ma DDF, Davey RA, Harman DW .etal: De tection of a multidrug resistant phenotype in acute non-lymphoblastic leukaemia. Lancet 1987;i: 135 137.
208
O’Meara/Imamura/Johnson/Ball/Rooney/ Kierce/Tsuruo/Dervan
P-gp in Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
12 FairchildCR. IvySP. Kao-ShanC.etal: isola tion of amplified and overexpressed DNA se quences from Adriamycin-resistant human breast cancer cells. Cancer Res 1987;47:514. 13 Rogan AM, Hamilton TC. Young RD. et al: Reversal of Adriamycin resistance by verapa mil in human ovarian cancer. Science 1984; 224:994. 14 Vandenberghe E. Staines A. Brcatnach F, O'Meara A: Recent experience with intensive combination chemotherapy for treatment of childhood acute lymphoblastic leukaemia. Ir J MedSci 1989;158:97-101. 15 Sugawara I. Nakahama M. Hamada H. et al: Apparent stronger expression in the human adrenal cortex than in the human adrenal medulla of M, 170,000-180,000 P-glycoprotein. Cancer Res 1988:48:4611— 4614.
Anne O'M eara3 A t sushi Imamuraa Pauline Johnson3 Raymond Balla Sean Rooneya Barry Kiercea Takashi Tsuruob Peter Dervanc
Reactivity of P-Glycoprotein Monoclonal Antibodies in Childhood Cancers
* Children's Research Centre, Our Lady's Hospital for Sick Children, Dublin. Ireland: 11 Division o f Experimental Chemotherapy, Cancer Chemotherapy Center, Japanese Foundation for Cancer Research, Tokyo, Japan; e Department of Pathology, Mater Misericordiae Hospital. Dublin, Ireland
Introduction The overall long-term disease-free survival of 60-65% in childhood cancer is attributable mainly to chemosensitivity of most paediatric tumours. Some patients, how'-
ever. become resistant to a w'ide range of chemotherapeut ic agents after an initial response and eventually die. 1n ad dition, a small percentage of tumours fail to respond to standard, albeit intensive, chemotherapy ab initio. This has prompted researchers to investigate the various me-
L)r. Anne O'M eara Children's Research Ccnlre Our Lady’s I lospital lor Sick Children Crumlin. Dublin 12 (Ireland)
< 1992 S. Karger AG. Basel 0030 2414 92 0493 0203 S 2.75 0
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
KeyWords Chemoresistance Childhood Cancers P-Glycoprotein JSB-I MRK16
Abstract P-Glycoprotein (P-gp), the product of the mdr-1 gene, is implicated in the de velopment of chemoresistance in a variety of, mostly adult, cancers. Its role in paediatric tumours, most of which arc non-epithelial in origin, has yet to be fully elucidated. A study was undertaken to investigate reactivity of two P-gp monoclonal antibodies (M Abs). JBS-I and M RK16, recognising cytoplasmic and surface epitopes, respectively, of the P-gp molecule, in a variety of newly diagnosed and relapsed childhood cancers. P-gp was not expressed in any of 36 tumours examined (neuroblastoma 13, nephroblastoma 12, rhabdomyosar coma 6, lymphoma 3. teratoma 1. Ewings 1), 14 of whom had chemoresistant disease. Reactivity to both MAbs was also investigated in patients with acute leukaemia. Out of 10 diagnostic acute lymphoblastic leukaemia (ALL) sam ples, a positive reaction with JSB-1was observed in 1patient who failed to remit on standard induction therapy and in 3 of 6 patients in ALL relapse, only 1 of whom showed low grade positivity with MRK16. Both MAbs reacted posi tively in 1 patient with acute non-lymphocytic leukaemia (AN I.L) at diagnosis who achieved remission with teniposideand cytosine arabi noside, but relapsed 7 months later and was again positive with both Mabs. JSB-1also showed vary ing degrees of positivity in 4 out o f4 other patients in ANLL relapse. It would therefore appear that P-gp is unlikely to mediate chemoresistance in most solid tumours of childhood, but may well play a major role in the development of chemoresistance in acute leukaemia.
chanisms of inherent and acquired chemoresistance in human cancers. Rodent and human multidrug-resistant cell lines have been established in vitro by multi-step drug selection [1,2], most of which have either amplification or over-expres sion of a gene termed mdr-1, which is transcribed into a 4.5-kb mRNA [3-5], the product of this gene being a 170kD membrane glycoprotein (P-gp) [6]. P-gp appears to act as an energy-dependent drug efflux pump which results in drugs being eliminated from cells before a cytotoxic effect can be exerted [7], A number of monoclonal antibodies (MAbs) have been raised to different epitopes on the P-gp molecule [8-10], and a pattern of reactivity has been esta blished for normal and neoplastic (mainly adult car cinoma) tissue [11], These authors have identified positiv ity with P-gp MAbs in normal epithelial cells making meaningful interpretation of various carcinomas (i.e. can cers arising from epithelial tissue) difficult. There is little information, on the other hand, on reactivity of these MAbs to paediatric tumours, most of which are non-epithelial in origin. The aim of this study was to screen a vari ety of childhood cancers with P-gp MAbs, thus enabling the role of this glycoprotein in the behaviour of childhood cancers to be established.
Results Normal kidney, adrenal, striated muscle, spleen, thy mus, lymph, lymph node and BM were screened with both M Abs: proximal tubules of kidneys reacted positively with both antibodies while glomeruli were negative. Adrenal cortex was strongly positive while medulla reacted weakly. Intensity of staining in all cases was greater with MRK16. Striated muscle, spleen, thymus, lymph node and BM were negative with both MAbs. Tumours
Samples Normal tissue and tum our samples from newly diagnosed un treated and partially treated or relapsed (grouped together as treated) patients were freshly frozen and stored at -195 C until analysis. The Adriamycin-resistant cell lines MCF-7Ad [ 12] and 2780-Ad [ 13] were used as positive controls. Bone marrow (BM) smears of patients with newly diagnosed and relapsed acute lymphoblastic leukaemia (ALL), acute non-lymphocytic leukaemia (ANLL), malignant disease with BM involvement and normal BM aspirates were fixed in acetone, air-dried and stored at -20 C until analysis. M ononuclear preparations of BM samples were processed immediately.
Monoclonal Antibodies The IgGI MAb JSB-1 [8] and the IgG2A MAb M R K I6[9j re cognising cytoplasmic and surface epitopes, respectively, of P-gp were employed. JSB-1 was used at 1:50 dilution of ascitic fluid and M R K 16 at a concentration of 4 pg/ml.
Methods JSB-1 reactivity was measured on BM smears and cytospin prepa rations using an APA AP technique; M RK. 16 reactivity of viable cells was measured by an indirect method using FITC-conjugated rabbit anti-mouse IgG F(ab'), fragment and analysed by FACScan (Becton Dickinson) flow cytometry. Cell lines were also analysed by fluores
Tumour samples from 36 patients in total were screened (table 1). Most tumours were screened with JSB-1 and M RK 16 simultaneously; both MAbs gave identical results with one exception, an untreated rhabdomyosarcoma which showed diffuse positivity with JSB-1, but was nega tive with MRK16 (table 1). This patient was sensitive to chemotherapy which included vincristine, actinomycin D and cyclophosphamide and remains off treatment with no evidence of disease 3 years from diagnosis. All other tu mours, including those with residual or recurrent disease following chemotherapy (both grouped together as ‘treat ed’ in table 1), were negative. Chemotherapy for children with neuroblastoma (NBL) had included vincristine, cyclophosphamide, cisplatinum and an epipodophyllotoxin, as per European Neuroblastoma Study Group protocol, while patients with Wilms’ tumour (WT) had re ceived vincristine, actinomycin D and Adriamycin, ac cording to the United Kingdom Children’s Cancer Study Group protocol. A consistent observation however was the strong degree of positivity observed in endothelial cells of blood vessels which was most apparent in NBL tu mours.
O'Meara/Imamura/Johnson/Ball/Rooney/ Kierce/Tsuruo/Dervan
P-gp in Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Materials and Methods
204
cence microscopy in order to overcome any misinterpretation o f data by flow cytometry due to cell clumping. 5-pm sections o f freshly frozen tissue were cut using a cryostat microtome. Sections were fixed in acetone at 4 C for 10 min, then flooded with 10% species-specific normal serum to minimise non specific antibody binding. Endogenous peroxidase activity was inac tivated by 0.3% hydrogen peroxide in methanol for 30 min. Sections were then incubated with primary antibody in a moist chamber at 4 C for 16 h. The streptavidin-biotin complex (Dakopatts) method was applied for immunohistochemistry. Second antibody was bioti nylated F(ab'), fragment rabbit anti-mouse IgG (Dakopatts). Sec tions were developed in 0.025% diaminobenzidine solution. Spec ificity of immunohistochemical staining was confirmed by control staining with non-immune mouse serum. All slides were counterstained with 1% methyl green.
Leukaemias The Adriamycin-resistant cell lines MCF-7Ad and 2780-Ad reacted strongly with both MAbs (fig. 1). JSB-1 andM RK 16 reactivity was examined on BM aspirates of patients with ALL and AN LL (table 2). In the ALL group, a diagnostic aspirate from 1 patient with T cell ALL who failed to remit on standard induction therapy, consisting of vincristine, Adriamycin, L-asparaginase and pred nisolone, as per a modified ALL-BFM 81 protocol [14], showed >50% positivity with JSB-1, but was negative with MRK16; this patient eventually achieved haematological remission with alternative chemotherapy, which included amsacrine, etoposide and cytosine arabinoside. Three out of 6 patients in BM relapse, all of whom de veloped BM relapse on maintenance treatment consisting of mercaptopurine and methotrexate, reacted with one or both Mabs: Patient 1 was negative with both antibodies on BM re lapse, failed to achieve second remission with induction therapy as stated above, and showed up to 50% positivity in BM with JSB-1 prior to death; 5% of lymphoblasts reacted with MRK16. Patient 2 was similarly negative on BM relapse, failed to achieve second remission also with the above-mentioned induction therapy, which was later changed to etoposide
Table 1. Summary of childhood solid tumours screened
Tumour
P-gP positive
p-gp negative
Chemoresistant disease
0 0
6 7
1 5
0 0
7 5
2 4
la 0
4 1
1 0
0 0
2 i
0 1
0 0
i i
0 0
WT Untreated (7) Treated (5)
Rhabdomyosarcoma Untreated (5) Treated (1)
Marrow status
JSB-1
MRK.16
9(1 H) 13(0) 6( 1H. 2M)
7(0) 8(0) 3 (1L)
ALL Diagnosis (n = 10) Remission (n = 14) Relapse (n = 6)
ANLL Diagnosis (n = 1) Remission (n = 1) Relapse (n = 5)
1(H) 1(0) 5 (2H. 2M. 1L)
1(H) 1(0) 2(1 H)
Miscellaneous
Lymphoma Untreated (2) T reated ( 1)
Fig. 1. JSB-1 staining of MCF-7Ad cell line using an APAAP technique (63 x 1 .2 5 x 12.5). Note variable intensity of staining among cells which is maximal at cytoplasmic aspect o f cell mem brane.
Table 2. BM reactivity
NBL Untreated (6) Treated (7)
and cytosine arabinoside, and survived in relapse for 6 months; 50% of lymphoblasts reacted with JSB-l and 17% with M R K 16, 1 months prior to death. Patient 3 was also negative with both Mabs on first BM relapse; a second remission was achieved with etoposide
N B L inB M
- Diagnosis (n = 3) - Relapse (n = 1) Hodgkins positive BM (n = 1) ITP (n = 1) Infection (n = 2) Normal (n = 17)
2(0) 1(H) 1(0) 0 0 13(0)
3(0) 0 0 1(0) 2(0) 6(0)
Miscellaneous
a
Numbers in parentheses refer to the total number o f patients. JSB-l positiveand MRK16 negative.
n = Total number of patients; ITP = idiopathic thrombocytopaenic purpura. Number in parentheses = number of patients with positive reaction graded as H = > 50% cells positive; M = 20-50%, and L = 10-20%.
205
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Teratoma: untreated (1) Ewing’s: untreated (1)
and cytosine arabinoside. but a second BM relapse oc curred 4 months later. JSB-1 at that time reacted with more than 50% of lymphoblasts (fig. 2) and M R K 16 was nega tive. Third BM remission was not achieved. Among the ANLL patients, only one diagnostic as pirate was available for analysis; a high degree of positivity was noted with both M Abs (fig. 3). This patient w ho had trisomy 21 and an M2 (FAB classification) morphology, achieved haematologic remission with a teniposide and cytosine arabinoside induction schedule; remission BM
was negative w'ith both MAbs. Seven months from diag nosis, while still on a maintenance schedule of teniposide and cytosine arabinoside, she developed a BM relapse; JSB-1and M R K 16 demonstrated strong positivity at that time. Four other ANLL relapses demonstrated varying degrees of positivity with JSB-1, only 1 of which was also screened with MRK 16 and was negative. All of these pa tients had chemoresistant disease. One patient with relapsed NBL showed 80% positivity of NBL cells in BM with JSB-1; MRK 16 analysis was not performed.
Discussion
Fig. 2. JSB- 1 reactivity in relapsed ALL (63 x 1.25 x 12.5). In tense staining observed in scanty cytoplasm of lymphoblasts (1).
It would appear from the above data that expression of P-gp is, at best, an infrequent phenomenon in both NBL and nephroblastoma (WT). Epithelial cells within the kid ney, especially proximal tubules, demonstrated strong reactivity with both MAbs, but WT, either untreated diag nostic material or chemoresistant relapsed tissue, was con sistently negative. The pattern of reactivity in adrenal gland was similar to what Sugawara et al. [ 15] have report ed, i.e. homogeneous staining of cortex with patchy stain ing of medulla. All NBL tumours however, including those with chemoresistant disease, were negative with both antibodies; this observation is in agreement with others [16], Phaeochromocytoma, another tumour of adrenal medulla, has also been found to be negative [11,
3oon
206
O’Meara Imamura Johnson Ball Rooney Kierce/Tsuruo Dervan
P-gp :n Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
Fig. 3. Flow cytometric analysis (FACScan)of MRK 16 reactivity at diagnosis in I patient with ANLL (shaded peak); clear peak represents FITC-conjugated rabbit antimousc IgG F(ab')2control.
16]. Increased mdr-1 transcript levels have however been reported in relapsed NBL [17, 18]; one would therefore have expected increased mRNA would imply more pro tein expression, but this is evidently not the case. Endo thelial cells of blood vessels, on the other hand, showed marked positivity and may well act, in this situation, by creating pharmacological sanctuaries within these tu mours as others have reported [19], Most rhabdomyosar comas in this study were investigated at diagnosis, most of whom had chemosensitive disease. Van der Valket al. [11] have reported weak positivity ofM RK16,JBS-landC219 with striated/skeletal muscle, which would explain the fal se-positive result obtained with one chemosensitive tu mour to JSB-1. Chan et al. [20] have observed statistically significant differences in P-gp expression in soft tissue sar coma of childhood between patients who relapsed and those who did not; of 9 patients in that series where P-gp was detected, 5 were negative at first presentation. The pattern of reactivity in leukaemia samples was par ticularly interesting. Reactivity of P-gp antibodies only de veloped with progression of disease, and JSB-1 positivity was observed earlier and to a greater extent than M R K 16. The fact that JSB-1 recognises a cytoplasmic epitope which is generated earlier in protein synthesis than the sur face epitope which M R K 16 recognises would explain this observation, and data from others [21] would support this theory. This may well explain the failure of Ito et al. [22] to demonstrate P-gp expression in acute leukaemia patients using M R K 16, while Ma et al. [23] showed positivity using C219 M Ab. Based on this data, it would appear advisable
to employ M Abs recognising both internal and cell surface epitopes before regarding samples as negative. Only 1 pa tient with ANLL was investigated at diagnosis and de monstrated a high degree of positivity with both MAbs. She achieved haematological remission with a teniposide, Cytosar induction schedule, but relapsed 7 months from diagnosis. While teniposide is one of the naturally occur ring agents susceptible to P-gp action, the synergistic ac tion with Cytosar may have initially overcome the other wise likely evolution of an mdr phenotype. Conversely, 3 patients with relapsed ALL did not react with either MAb. Clearly alternative mechanisms of chemoresistance are operational in some malignancies. In conclusion, this study has shown that in acute leu kaemias of childhood, P-gp may be a major mechanism in the development of multidrug resistance. There was no evidence for P-gp-mediated chemoresistance in even the most resistant of solid tumours. It may be that mdr transcript levels, rather than P-gp expression, are more re levant in the development of chemoresistance. Data from haematological studies however [22, 23], including this study, would appear to belie this suggestion. Larger scale studies, particularly with serial monitoring, are needed to address this issue. Acknowledgements We are grateful to F. Breatnach for permission to study his pa tients, R. J. Fitzgerald for providing tumour samples and O. Reddy for secretarial assistance.
References 5 Riordan JR. Deuchars K. Kartner N. et al: Amplification of P-glycoprotein genes in multidrug resistant mammalian cell lines. Nat ure 1985;316:816-819. 6 Ueda K. Cornwell M M, Gottesman M M. et al: The mdr-1 gene responsible for multidrug resistance codes for P-glycoprotein. Biochem Biophys ResCommun 1986;141:956-962. 7 Gottesman MM. Pastan I: The multidrug transporter, a double edged sword. J Biol Chem 1988;263:12163-12166. 8 Scheper RJ, Bulte JWM, Brakke JG P. et al: Monoclonal antibody JSB-1 detects a highly conserved epitope on the P-glycoprotein asso ciated with multidrug resistance. Int J Cancer 1988;42:389-394.
9 Hamada H. Tsuruo T: Functional role for the 170- to 180-kDa glycoprotein specific todrugresistant tumor cellsas revealed by monoclonal antibodies. Proc Natl Acad Sci USA 1986:83: 785-7789. 10 Kartner N. Fverndcn-Porclle D. Bradley G .ct al: Detection of P-glycoprotein in mdrcell lines by monoclonal antibodies. Nature 1985:316: 820-823. 11 Vandcr Valk P, Van KalkenCK, Ketelaars II. et al: Distribution of multidrug resistance asso ciated P-glycoprotein in normal and neoplastic human tissues. Ann Oncol 1990; 1:56—64.
207
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
1 Juliano RJ, Ling V: A surface glycoprotein modulating drug permeability in Chinese ham ster ovary cell mutants. Biochcm Biophys Acta 1976;455:152-162. 2 FojoA,AkiyamaS.Gottesman M M.PastanI: Reduced drug accumulation in multiple drugresistant human KB carcinoma cell lines. Can cer Res 1985:45:3002 3007. 3 Roninson IB. Chin JE, Choi KG. et al: Isola tion of human mdr cDNA sequences amplified in multidrug resistant KB carcinoma cells. Proc Natl Acad Sci USA 1986;83:4538-4542. 4 Scotto KW, Bicdler JL. Melera PW: Am plification and expression of genes associated with multidrug resistance in mammalian cells. Science 1986;232:751-755.
16 Cordon-CardoC, O'Brien JP, BertinoJ.etal: Expression of the human multidrug resistance gene (P-glycoprotein) in human tumours: An Immunopathologie study. US Can Acad Pa thol 1990;22A:124. 17 Bourhis J, Bcnard J, Hartman O, et al: Correla tion of mdrl gene expression with chemo therapy in neuroblastoma. J Natl Cancer Inst 1989;81:1401-1405. 18 Fojo AT, Ueda K. Slamon DJ, et al: Expres sion of a multidrug resistance gene in human tumours and tissues. Proc Natl Acad Sci USA 1987;84:265-269. 19 Cordon-CardoC, O'Brien JP, Casals D, et al: Immunoanatomic and immunopathologie ex pression of the multidrug resistance gene pro duct. Cancer Cells 1989;7:87-93.
20 Chan HSL, Thorner PS, Haddad G, et al: Immunohistochemical detection of P-glyco protein: Prognostic correlation in soft tissue sarcoma of childhood. J Clin Oncol 1990;8: 689-704. 21 Spier CM, List AF, Cline A, et al: Drug resis tance in the myelodysplastic syndromes. US Can Acad Pathol 1990;95A:557. 22 Ito Y, Tanimoto M, Kumazawa T, et al: In creased P-glycoprotein expression and multidrugresistant gene (mdrl ) amplification are in frequently found in fresh acute leukaemia cells. Sequential analysis of 15 cases at initial pre sentation and relapsed stage. Cancer I989;63: 1534-1538. 23 Ma DDF, Davey RA, Harman DW .etal: De tection of a multidrug resistant phenotype in acute non-lymphoblastic leukaemia. Lancet 1987;i: 135 137.
208
O’Meara/Imamura/Johnson/Ball/Rooney/ Kierce/Tsuruo/Dervan
P-gp in Chemoresistance of Childhood Cancers
Downloaded by: University of Chicago Library 205.208.116.24 - 4/13/2018 4:21:49 PM
12 FairchildCR. IvySP. Kao-ShanC.etal: isola tion of amplified and overexpressed DNA se quences from Adriamycin-resistant human breast cancer cells. Cancer Res 1987;47:514. 13 Rogan AM, Hamilton TC. Young RD. et al: Reversal of Adriamycin resistance by verapa mil in human ovarian cancer. Science 1984; 224:994. 14 Vandenberghe E. Staines A. Brcatnach F, O'Meara A: Recent experience with intensive combination chemotherapy for treatment of childhood acute lymphoblastic leukaemia. Ir J MedSci 1989;158:97-101. 15 Sugawara I. Nakahama M. Hamada H. et al: Apparent stronger expression in the human adrenal cortex than in the human adrenal medulla of M, 170,000-180,000 P-glycoprotein. Cancer Res 1988:48:4611— 4614.
