Inr J Radmmn Oncolog)~ Bml Phys Vol. 22. pp. 765-768 Printed I” the U.S.A. All rights reserved.
Copyright
??Session F: Thiol Modulation
0360.3016192 $5.00 + .OO 0 1992 Pergamon Press plc
and Protection
HETEROGENEITY IN RESPONSE TO TREATMENT WITH BUTHIONINE SULFOXIMINE OR INTERFERON IN HUMAN MALIGNANT GLIOMA CELLS M. JOAN
PH.D.,
ALLALUNIS-TURNER,*
RUFUS S. DAY III, PH.D.,? Cross Cancer Institute,
KELLY
GERALDINE
M. BARRON, B.Sc,*
DOBLER, D1PL.t AND RAUL
I 1560 University
Avenue,
Edmonton,
C. URTASUN,
Alberta Canada
M.D.*
T6G 122
Two tumor cell lines were established from each of three human malignant glioma biopsy specimens (M059, M067, M071) and sensitivity to treatment with radiation or chemotherapeutic agents (BCNU, nitrogen mustard) was determined. The effects of recombinant human interferon-alpha (rIFN) on the radiation response and of buthionine sulfoximine (BSO) on the drug response were investigated as well. For tumor M059, two cell lines that differed significantly in radiosensitivity were isolated (surviving fractions at 2 Gy = 0.02 and 0.64). The chemosensitivity and response to chemical modification differed as well. Cell lines established from tumor MO71 differed in their response to rIFN only and were not sensitized by BSO. MO67 cell lines showed little difference and were not sensitized by either agent. These results suggest that differences may exist both within and among human malignant gliomas with regard to their sensitivity to drugs, radiation, and the ability of chemical agents to modify treatment responses. Malignant glioma, Tumor heterogeneity,
Interferon, Buthionine sulfoximine,
INTRODUCTION
Radiosensitizer,
Chemosensitizer.
can be isolated from biopsy specimens obtained prior to therapy. In addition, we wish to determine whether the sensitivity of these cells to treatment can be chemically modified and whether clones isolated from the same tumor respond to sensitizing agents in a similar manner. Preliminary results based on three human malignant glioma biopsies (M059, M067, and M071) are reported here.
Malignant gliomas are generally fatal irrespective of aggressive radiotherapy and/or chemotherapy. Studies that have evaluated the chemosensitivity of tumor cells obtained from untreated patients suggest that the drug resistant and sensitive cells coexist prior to therapy (13). It is possible that malignant gliomas also contain tumor clones that vary in radiosensitivity. However, the frequency of occurrence of such clones has yet to be fully documented. Because conventional therapy has done little to alter malignant glioma’s poor prognosis, various strategies using chemical modifiers to influence the effects of radiotherapy or chemotherapy have been undertaken. These have included the use of hypoxic cell sensitizers (3, 8), IUdR (9), and BUdR ( 12). In general, these interventions have failed to produce significant improvements in tumor control. The lack of sensitization may be the consequence of a number of factors. One possibility is that the response to chemical modifiers is heterogeneous and varies both within and among tumors. To investigate this, we have begun a study to determine whether tumor cell lines that differ in their inherent radiosensitivity or chemosensitivity
All procedures were approved in advance by the Institute’s Ethics Committee on Human Experimentation. Written informed consent was obtained from all patients prior to surgery. A portion of the diagnostic biopsy was obtained from three patients with grade 4 astrocytoma. One-half of each specimen was minced with scalpels then dissociated for 30 min at 37°C in an enzyme cocktail containing collagenase (0.025%), DNAase (0.04%), and Pronase (0.05%) followed by a 30 min holding at 4°C. Cells were then washed free of enzymes and cultured in DMEM/F12 medium supplemented with 10% fetal calf serum and 1 mM HEPES buffer. The remaining portion
Presented at the Seventh International Conference on Chemical Modifiers of Cancer Treatment, Clearwater, FL, 2-5 February, 199 I. * Dept. of Radiation Oncology. 1_Dept. of Medicine. Reprint requests to: Dr. Joan Allalunis-Turner.
Acknowledgements-The authors wish to thank Gina Kennedy for typing the manuscript. Supported by a grant from the National Cancer Institute of Canada and a Scholarship Establishment Award from the Alberta Heritage Foundation for Medical Research to Dr. R. S. Day III. Accepted for publication 26 July 199 1.
METHODS
AND MATERIALS
Biopsies
165
I. J. Radiation
766
Oncology
Table 1. Effect of radiation,
0 Biology 0 Physics
chemotherapeutic
Volume 22. Number 4, 1992
agents and chemical modifiers on human glioma cell lines
Cell line
DNA index
GSH*
SF2I
SER (rIFN)$
SER (BCNU + BSO)$
M059J M059K M067J M067K M07lJ M071K
1.66 2.89 1.64
17.5 15.5 4.5 7.4 8.8 9.7
0.02 0.64 0.36 0.30 0.19 0.14
0.7 1.9 1.0 I .o 1.4 2.0
1.o (S) 1.8 (S) I .O (R) I .O (R) 1 .o (S) 1.1 (S)
1.95 1.09 2.07
SER (HN2 + BSO)§ 1.8 1.3 1.1 I. I 1.1 1.2
(R) (R) (R) (R) (R) (R)
*GSH is expressed as nmol/106 cell. The error associated with these measurements was 5 10%. tSF2 values were derived from the computer fit of survival curves based on three or more separate irradiation experiments. $The sensitizer enhancement ratios (SER) = (dose to achieve 1% survival in control)/(dose to achieve 1% survival in rIFN-treated cells). Pooled results from three or more separate experiments were used for SER calculations. §The SER’s were calculated as the ratio of the values of the relative absorbance determined for control and BSO treated cells at doses of 100 FM BCNU or 5 PM HN2. Pooled results based on two or more separate experiments are shown. (S) indicates control cells were sensitive to drug; (R) indicates control cells were resistant to drug.
of the biopsy
was cut in fragments
h 1 mm3
and
placed
in 60 mm culture dishes. Coverslips with stopcock grease were gently pushed on top of the fragments. Biorich (DMEM or F12) medium plus 10% fetal calf serum was added. Cell lines arising from the enzyme and coverslip techniques are distinguished by the suffixes “J” and “K”, respectively. Cell lines established from tumors M059, M067, and MO71 were used at passage numbers 49-68, 13-23, and 7- 15, respectively.
Irradiation Exponential phase cells grown in medium only or pretreated with recombinant human interferon alpha (rIFN) ( 1000 p/ml X 48 hr) were irradiated in suspension and assayed for colony formation in monolayer cultures. The survival curves were fit to the linear quadratic model of cell kill and the surviving fraction at 2 Gy (SF2) was interpolated. The sensitizer enhancement ratios (SER) were calculated as the ratio at I % survival level of the radiation doses measured for control to that measured for rIFNtreated cells. Flow cytometry Cells were prepared for flow cytometric DNA analysis according to techniques previously described (6). Normal human lymphocytes served as a diploid control and were prepared similarly. Cells were analyzed using a flow cytometer* equipped with a 488 argon ion laser for excitation and the CellFIT cell cycle analysis program (Version 2.0). Chemosmsitivity Cells in complete culture medium were seeded into 96 well plates, and 48 hr later, buthionine sulfoximine (BSO) ( 100 FM) was added to selected wells. After 24 hr incu*FAScan@.
bation. BSO was removed and graded concentrations of freshly prepared BCNU or nitrogen mustard (HN2) were added for 2 hr. Following removal of the drugs, fresh medium was added and plates were incubated for an additional 7 days. Relative sensitivity to drug -t BSO was determined using a modification of the MTT assay previously described (1 1). Cells were judged to be sensitive to BCNU or HN2 if cell proliferation in the drug-treated wells was less than 60% of control values (14). The SER’s were calculated as the ratio of the values of the relative absorbances determined for control or BSO treated cells at doses of 100 PM BCNU or 5 PM HN2.
Exponential phase cells were assayed for intracellular glutathione (GSH) content using a biochemical assay ( 15).
RESULTS The results from all experiments are summarized in Table 1. Both tumor dissociation techniques gave rise to cell lines that contain one or more populations of aneuploid cells. The DNA indices ranged from 1.09 to 2.89. The GSH content of each cell line was also determined. Although the GSH values were similar for cell lines isolated from the same tumor, an approximate 4-fold difference was noted among the three tumors tested. For the two lines derived from tumor M059, significant differences in radio- and chemosensitivity were observed. The mean SF2 values of M059J and M059K cells were 0.02 and 0.64, respectively. The results of the MTT assay indicated that both cell clones were relatively sensitive to BCNU. The results obtained with the proliferation assay are thus consistent with our finding that these cells are unable to repair 06-methylguanine lesions in DNA. and
Human
malignant
glioma cells 0 M. J. ALLALIJNIS-TLJRNER e/ al.
both show the Merphenotype (Day, Dobler, AllalunisTurner, Fulton and Urtasun, unpublished data, June 1990). However, M059J cells were approximately five times more sensitive to BCNU than were M059K cells. Both M059J and M059K cells were resistant to HN2. In the chemical modification experiments, only M059K cells were sensitized to radiation by rIFN (SER = 0.7 for M059J and 1.9 for M059K). Similarly, when cells were pretreated with BSO prior to exposure to BCNU, only M059K cells were sensitized (SER = 1.8 vs. 1.0). In contrast, in experiments with HN2 + BSO, significant sensitization was observed with M059J cells (SER = 1.8). A lesser degree of sensitization was seen in M059K cells (SER = 1.3). Cell lines isolated from tumor MO67 expressed similar radio- and chemosensitivities which were not sensitized by rIFN. Both M067J and M067K were resistant to BCNU and HN2 in either the presence or in the absence of BSO. Cell lines from tumor MO7 1 had similar radiosensitivities, although MO7 1K cells showed a greater response to rIFN than did MO7 1J cells (SER = 2.0 vs 1.4). Both cell lines were resistant to HN2 in the presence or absence of BSO. Both MO7 1J and MO7 1K cells show the Merphenotype, although MO7 1K cells were slightly more sensitive to BCNU (- 1.2 fold). Neither MO7 1J nor MO7 1K cells were sensitized by BSO.
DISCUSSION Genotypic and phenotypic heterogeneity is a hallmark of human malignant gliomas (2). This variation in cellular and molecular characteristics suggests that sub-populations of tumor cells that differ in their inherent radiosensitivity or chemosensitivity may exist within a tumor prior to therapy. It is also possible that in a given tumor, subpopulations of cells may be differentially sensitive to the effects of chemical agents or cytokines that are used in an attempt to modify treatment response. In this study, we report the results of initial studies that are attempting to characterize heterogeneity in the response of human malignant glioma cells to treatment with radiation, BCNU, and HN2 as well as the chemical modifying agents, BSO, and rIFN. For these studies, two cell lines were established from each of three biopsy specimens by using different dissociation techniques (i.e., enzyme cocktail vs mechanical dissociation). Although it is possible that the dissociation techniques themselves may have selected for cells that are more or less sensitive to the agents tested, no consistent differences have been noted. For example, cell lines established from enzymatically dissociated tumor fragments were not consistently more drug- or radiationsensitive when compared to the companion line established from the same tumor through the use of mechanical techniques. Thus, the heterogeneity described in this report probably reflects geographic distribution of different populations of cells in different microregions of the tumor. In studies that assessed inherent radiosensitivity, little
167
difference in radiosensitivity was observed for the paired cell lines isolated from tumors MO67 and MO71 (SER = 0.36 and 0.30 for M067; 0.19 and 0.14 for MO7 1). However, a large difference in radiosensitivity was noted for the two cell lines from tumor MO59 (SF2 of 0.02 and 0.64 for M059J and M059K cells, respectively). In larger studies that have investigated the radiosensitivity of more radioresponsive tumor types (i.e., cervix, head and neck), intratumoral differences in SF2 values of this magnitude have not been observed (4, 16). Clearly, more gliomas must be studied before it can be ascertained whether the extreme differences in radiosensitivity observed for MO59 tumor cells commonly occur. Note that with the exception of M059K cells, all other cell lines tested for these studies were moderately radiosensitive. In other studies in our laboratory, we have characterized the SF2 values of 20 malignant glioma biopsies and have observed a mean (& SD) SF2 of 0.43 + 0.20 (Allalunis-Turner, Ban-on, Day, Fulton and Urtasun, unpublished data, December 1990). Thus. the glioma cell lines in this study are among the more radiosensitive that we have tested. Although malignant gliomas are clinically among the least radioresponsive of tumors, a large variation in inherent radiosensitivity is likely to exist as has been previously determined for other tumor types (4, 16). When the cell lines in this study were treated with rIFN prior to radiation, the radiosensitivity of cells from tumor MO67 was unaffected, whereas that of cells from tumors MO59 and MO71 was enhanced to varying degrees (range of SER = 0.7-2.0). Other investigators who have examined the effects of interferons on radiosensitivity have reported similar values (5). Studies by Chan and Keng have suggested that alteration in cell cycle distribution following rIFN treatment may contribute to its radiosensitizing effects (5). However, in the present study, no difference in cell cycles was noted in any of the six clones following exposure to rIFN. In addition, the degree of radiosensitization observed was independent of the antiproliferative effects of rIFN, as colony growth in all rIFN-treated cultures was reduced to a similar level (40-50s). Other investigators have suggested that rIFN’s sensitizing effects may result from altered ability to accumulate sublethal damage (7). Further work is required to test whether this is a factor in the sensitivity observed among these cell lines. In contrast to the limited information available concerning radiosensitivity, heterogeneity in chemosensitivity both within and among tumors has been well described for gliomas (13). Our observation that MO59 cells, for example, are both sensitive to BCNU but resistant to HN2 is consistent with this finding. We have also observed significant differences in response to the chemosensitizing effects of BSO among the tumor cell lines tested. Only cells from tumor MO59 were sensitized by BSO, although these lines differed in their response to BCNU and HN2. The four remaining cell lines were not sensitized by BSO. Other studies from our laboratory using early passage malignant glioma cell lines or freshly dissociated cells ob-
768
I. J. Radiation
Oncology
0 Bio!ogy 0 Physics
tained from glioma biopsies have demonstrated a similar variation in response to BSO’s chemosensitizing effects. These results have also suggested that chemosensitization by BSO is independent of initial GSH concentration. extent of GSH depletion, or Mer phenotype status (1) and Allalunis-Turner, Barron, Day, Fulton and Urtasun, unpublished data, June 1990). Other studies have reported variations in BSO’s effect on the chemosensitivity of a single cell line treated with a variety of drugs as well as its effects on different tumor types treated with the same agent (reviewed in 10). Thus, our results extend this observation and suggest that significant intratumoral heterogeneity may also exist with regard to BSO’s chemosensitizing effects. In conclusion, these preliminary results suggest that for
Volume 22. Number 4. 1992
human malignant gliomas, significant differences in radioand chemosensitivity can exist both within a single tumor and among different tumors. In addition, the ability of chemical agents to modify the treatment sensitivity may vary among the sub-populations of one tumor. It remains to be determined whether similar heterogeneity in treatment response, especially to chemical modifiers, occurs in other tumor types, or whether such differences are limited to tumors such as malignant gliomas in which heterogeneity is a “hallmark.” It will also be important to determine whether intratumoral differences in treatment sensitivities such as those described here are characteristic of tumors that are clinically radioresistant, such as gliomas and melanomas.
REFERENCES I. Allalunis-Turner, J.; Day, R. S.. 111;Urtasun. R. C.: Fulton. D. S.; Huyser-Wierenga, D. Mer phenotype, glutathione and inherent radiosensitivity as predictors of treatment response in patients with malignant glioma. (Abstr.). Proc. Amer. Assoc. Cancer Res. 30:288; 1989. 2. Bigner, D. D.; Bigner. S. H.: Ponten, J.: Westermark. B.; Mahaley, M. S.; Ruoslahti, E.: Herschman. H.; Eng, L. F.; Wikstrand. C. J. Heterogeneity of genotypic and phenotypic characteristics of fifteen permanent cell lines derived from human gliomas. J. Neuropathol. Exp. Neural. 15:201-229; 1981. 3. Bleehan. N. M.; Freedman. L. S.: Stenning, S. P. A randomized study of CCNU with and without Benznidazole in the treatment of recurrent grades 3 and 4 astrocytoma. lnt. J. Radiat. Oncol. Biol. Phys. 16:1077-1081; 1989. 4. Brock, W. A.; Baker, F. L.: Wike. J. L.; Sivon, S. L.; Peters, L. J. Cellular radiosensitivity of primary head and neck squamous cell carcinomas and local tumor control. Int. J. Radiat. Oncol. Biol. Phys. 18: l283- 1286: 1990. 5. Chang. A. Y. C.; Keng, P. C. Potentiation of radiation cytotoxicity by recombination interferons, phenomenon associated with increased blockage at the G2-M phase of the cell cycle. Cancer Res. 47:4338-434 I ; 1987. 6. Clevenger. C. Y.: Bauer, K. D.: Epstein, A. L. A method for simultaneous nuclear immunofluorescence and DNA content quantitation using monoclonal antibodies and flow cytometry. Cytometry 6:208-2 14: I985 7. Dritschilo. A.; Mossman, K.: Gray. M.; Sreevalsan, T. Potentiation of radiation injury by interferon. Amer. J. Clin. Oncol. 5:79-82; 1982. 8. Fulton. D. S.; Urtasun, R. C.: Shin, K. H.: Geggie, P. H. S.; Thomas. H.: Muller. P. J.; Moody, J.; Tanasichuk. H.: Mielke, B.: Johnson, E.; Curry, B. Misonidazole combined with hyperfractionation in the management of ma-
9.
IO.
I I.
12.
13. 14.
15.
16.
lignant glioma. Int. J. Radiat. Oncol. Biol. Phys. IO: I7091712: 1984. Kinsella. T. J.: Collins, J.: Rowland, J.; Klecker, R. Jr.; Wright, D.; Katz, D.; Steinberg, S. M. Pharmacology and phase I/II study of continuous intravenous infusion of iododeoxyuridine and hyperfractionated radiotherapy in patients with malignant glioblastoma multiforme. J. Clin. Oncol. 6:87 I-879: 1988. Jordan. J.; d’Arcy Doherty. M.; Cohen, G. M. Effects of glutathione depletion on the cytotoxicity of agents toward a human colonic tumor cell line. Br. J. Cancer 55:627-633: 1987. Mossman, T. Rapid calorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65:55-63: 1983. Phuphanich. S.: Levin. E. M.; Levin. V. A. Phase I study of intravenous bromodeoxyuridine used concomitantly with radiation therapy in patients with primary malignant brain tumors. Int. J. Radiat. Oncol. Biol. Phys. IO: 1769-1772; 1984. Rankin Shapiro, J.; Shapiro, W. R. Clonal tumor cell heterogeneity. Prog. Exp. Tumor Res. 27:49-66; 1984. Rosenblum, M.: Gerosa. M.; Dougherty, D. V.; Reese, C.: Barger, G. R.; Davis, R. L.; Levin, V. A.: Wilson, C. B. Agerelated chemosensitivity of stem cells from human malignant brain tumors. Lancet i:885-887; 1982. Tietze, F. Enzymatic method for determination of nanogram amounts of total and oxidized glutathione: application to mammalian blood and other tissues. Anal. Biochem. 27: 502-522; 1969, West, C. M. L.: Davidson, S. E.; Hunter, R. D. Evaluation of surviving fraction at 2 Gy as a potential prognostic factor for the radiotherapy of carcinoma of the cervix. Int. J. Radiat. Biol. 56:761-765; 1989.
Copyright
??Session F: Thiol Modulation
0360.3016192 $5.00 + .OO 0 1992 Pergamon Press plc
and Protection
HETEROGENEITY IN RESPONSE TO TREATMENT WITH BUTHIONINE SULFOXIMINE OR INTERFERON IN HUMAN MALIGNANT GLIOMA CELLS M. JOAN
PH.D.,
ALLALUNIS-TURNER,*
RUFUS S. DAY III, PH.D.,? Cross Cancer Institute,
KELLY
GERALDINE
M. BARRON, B.Sc,*
DOBLER, D1PL.t AND RAUL
I 1560 University
Avenue,
Edmonton,
C. URTASUN,
Alberta Canada
M.D.*
T6G 122
Two tumor cell lines were established from each of three human malignant glioma biopsy specimens (M059, M067, M071) and sensitivity to treatment with radiation or chemotherapeutic agents (BCNU, nitrogen mustard) was determined. The effects of recombinant human interferon-alpha (rIFN) on the radiation response and of buthionine sulfoximine (BSO) on the drug response were investigated as well. For tumor M059, two cell lines that differed significantly in radiosensitivity were isolated (surviving fractions at 2 Gy = 0.02 and 0.64). The chemosensitivity and response to chemical modification differed as well. Cell lines established from tumor MO71 differed in their response to rIFN only and were not sensitized by BSO. MO67 cell lines showed little difference and were not sensitized by either agent. These results suggest that differences may exist both within and among human malignant gliomas with regard to their sensitivity to drugs, radiation, and the ability of chemical agents to modify treatment responses. Malignant glioma, Tumor heterogeneity,
Interferon, Buthionine sulfoximine,
INTRODUCTION
Radiosensitizer,
Chemosensitizer.
can be isolated from biopsy specimens obtained prior to therapy. In addition, we wish to determine whether the sensitivity of these cells to treatment can be chemically modified and whether clones isolated from the same tumor respond to sensitizing agents in a similar manner. Preliminary results based on three human malignant glioma biopsies (M059, M067, and M071) are reported here.
Malignant gliomas are generally fatal irrespective of aggressive radiotherapy and/or chemotherapy. Studies that have evaluated the chemosensitivity of tumor cells obtained from untreated patients suggest that the drug resistant and sensitive cells coexist prior to therapy (13). It is possible that malignant gliomas also contain tumor clones that vary in radiosensitivity. However, the frequency of occurrence of such clones has yet to be fully documented. Because conventional therapy has done little to alter malignant glioma’s poor prognosis, various strategies using chemical modifiers to influence the effects of radiotherapy or chemotherapy have been undertaken. These have included the use of hypoxic cell sensitizers (3, 8), IUdR (9), and BUdR ( 12). In general, these interventions have failed to produce significant improvements in tumor control. The lack of sensitization may be the consequence of a number of factors. One possibility is that the response to chemical modifiers is heterogeneous and varies both within and among tumors. To investigate this, we have begun a study to determine whether tumor cell lines that differ in their inherent radiosensitivity or chemosensitivity
All procedures were approved in advance by the Institute’s Ethics Committee on Human Experimentation. Written informed consent was obtained from all patients prior to surgery. A portion of the diagnostic biopsy was obtained from three patients with grade 4 astrocytoma. One-half of each specimen was minced with scalpels then dissociated for 30 min at 37°C in an enzyme cocktail containing collagenase (0.025%), DNAase (0.04%), and Pronase (0.05%) followed by a 30 min holding at 4°C. Cells were then washed free of enzymes and cultured in DMEM/F12 medium supplemented with 10% fetal calf serum and 1 mM HEPES buffer. The remaining portion
Presented at the Seventh International Conference on Chemical Modifiers of Cancer Treatment, Clearwater, FL, 2-5 February, 199 I. * Dept. of Radiation Oncology. 1_Dept. of Medicine. Reprint requests to: Dr. Joan Allalunis-Turner.
Acknowledgements-The authors wish to thank Gina Kennedy for typing the manuscript. Supported by a grant from the National Cancer Institute of Canada and a Scholarship Establishment Award from the Alberta Heritage Foundation for Medical Research to Dr. R. S. Day III. Accepted for publication 26 July 199 1.
METHODS
AND MATERIALS
Biopsies
165
I. J. Radiation
766
Oncology
Table 1. Effect of radiation,
0 Biology 0 Physics
chemotherapeutic
Volume 22. Number 4, 1992
agents and chemical modifiers on human glioma cell lines
Cell line
DNA index
GSH*
SF2I
SER (rIFN)$
SER (BCNU + BSO)$
M059J M059K M067J M067K M07lJ M071K
1.66 2.89 1.64
17.5 15.5 4.5 7.4 8.8 9.7
0.02 0.64 0.36 0.30 0.19 0.14
0.7 1.9 1.0 I .o 1.4 2.0
1.o (S) 1.8 (S) I .O (R) I .O (R) 1 .o (S) 1.1 (S)
1.95 1.09 2.07
SER (HN2 + BSO)§ 1.8 1.3 1.1 I. I 1.1 1.2
(R) (R) (R) (R) (R) (R)
*GSH is expressed as nmol/106 cell. The error associated with these measurements was 5 10%. tSF2 values were derived from the computer fit of survival curves based on three or more separate irradiation experiments. $The sensitizer enhancement ratios (SER) = (dose to achieve 1% survival in control)/(dose to achieve 1% survival in rIFN-treated cells). Pooled results from three or more separate experiments were used for SER calculations. §The SER’s were calculated as the ratio of the values of the relative absorbance determined for control and BSO treated cells at doses of 100 FM BCNU or 5 PM HN2. Pooled results based on two or more separate experiments are shown. (S) indicates control cells were sensitive to drug; (R) indicates control cells were resistant to drug.
of the biopsy
was cut in fragments
h 1 mm3
and
placed
in 60 mm culture dishes. Coverslips with stopcock grease were gently pushed on top of the fragments. Biorich (DMEM or F12) medium plus 10% fetal calf serum was added. Cell lines arising from the enzyme and coverslip techniques are distinguished by the suffixes “J” and “K”, respectively. Cell lines established from tumors M059, M067, and MO71 were used at passage numbers 49-68, 13-23, and 7- 15, respectively.
Irradiation Exponential phase cells grown in medium only or pretreated with recombinant human interferon alpha (rIFN) ( 1000 p/ml X 48 hr) were irradiated in suspension and assayed for colony formation in monolayer cultures. The survival curves were fit to the linear quadratic model of cell kill and the surviving fraction at 2 Gy (SF2) was interpolated. The sensitizer enhancement ratios (SER) were calculated as the ratio at I % survival level of the radiation doses measured for control to that measured for rIFNtreated cells. Flow cytometry Cells were prepared for flow cytometric DNA analysis according to techniques previously described (6). Normal human lymphocytes served as a diploid control and were prepared similarly. Cells were analyzed using a flow cytometer* equipped with a 488 argon ion laser for excitation and the CellFIT cell cycle analysis program (Version 2.0). Chemosmsitivity Cells in complete culture medium were seeded into 96 well plates, and 48 hr later, buthionine sulfoximine (BSO) ( 100 FM) was added to selected wells. After 24 hr incu*FAScan@.
bation. BSO was removed and graded concentrations of freshly prepared BCNU or nitrogen mustard (HN2) were added for 2 hr. Following removal of the drugs, fresh medium was added and plates were incubated for an additional 7 days. Relative sensitivity to drug -t BSO was determined using a modification of the MTT assay previously described (1 1). Cells were judged to be sensitive to BCNU or HN2 if cell proliferation in the drug-treated wells was less than 60% of control values (14). The SER’s were calculated as the ratio of the values of the relative absorbances determined for control or BSO treated cells at doses of 100 PM BCNU or 5 PM HN2.
Exponential phase cells were assayed for intracellular glutathione (GSH) content using a biochemical assay ( 15).
RESULTS The results from all experiments are summarized in Table 1. Both tumor dissociation techniques gave rise to cell lines that contain one or more populations of aneuploid cells. The DNA indices ranged from 1.09 to 2.89. The GSH content of each cell line was also determined. Although the GSH values were similar for cell lines isolated from the same tumor, an approximate 4-fold difference was noted among the three tumors tested. For the two lines derived from tumor M059, significant differences in radio- and chemosensitivity were observed. The mean SF2 values of M059J and M059K cells were 0.02 and 0.64, respectively. The results of the MTT assay indicated that both cell clones were relatively sensitive to BCNU. The results obtained with the proliferation assay are thus consistent with our finding that these cells are unable to repair 06-methylguanine lesions in DNA. and
Human
malignant
glioma cells 0 M. J. ALLALIJNIS-TLJRNER e/ al.
both show the Merphenotype (Day, Dobler, AllalunisTurner, Fulton and Urtasun, unpublished data, June 1990). However, M059J cells were approximately five times more sensitive to BCNU than were M059K cells. Both M059J and M059K cells were resistant to HN2. In the chemical modification experiments, only M059K cells were sensitized to radiation by rIFN (SER = 0.7 for M059J and 1.9 for M059K). Similarly, when cells were pretreated with BSO prior to exposure to BCNU, only M059K cells were sensitized (SER = 1.8 vs. 1.0). In contrast, in experiments with HN2 + BSO, significant sensitization was observed with M059J cells (SER = 1.8). A lesser degree of sensitization was seen in M059K cells (SER = 1.3). Cell lines isolated from tumor MO67 expressed similar radio- and chemosensitivities which were not sensitized by rIFN. Both M067J and M067K were resistant to BCNU and HN2 in either the presence or in the absence of BSO. Cell lines from tumor MO7 1 had similar radiosensitivities, although MO7 1K cells showed a greater response to rIFN than did MO7 1J cells (SER = 2.0 vs 1.4). Both cell lines were resistant to HN2 in the presence or absence of BSO. Both MO7 1J and MO7 1K cells show the Merphenotype, although MO7 1K cells were slightly more sensitive to BCNU (- 1.2 fold). Neither MO7 1J nor MO7 1K cells were sensitized by BSO.
DISCUSSION Genotypic and phenotypic heterogeneity is a hallmark of human malignant gliomas (2). This variation in cellular and molecular characteristics suggests that sub-populations of tumor cells that differ in their inherent radiosensitivity or chemosensitivity may exist within a tumor prior to therapy. It is also possible that in a given tumor, subpopulations of cells may be differentially sensitive to the effects of chemical agents or cytokines that are used in an attempt to modify treatment response. In this study, we report the results of initial studies that are attempting to characterize heterogeneity in the response of human malignant glioma cells to treatment with radiation, BCNU, and HN2 as well as the chemical modifying agents, BSO, and rIFN. For these studies, two cell lines were established from each of three biopsy specimens by using different dissociation techniques (i.e., enzyme cocktail vs mechanical dissociation). Although it is possible that the dissociation techniques themselves may have selected for cells that are more or less sensitive to the agents tested, no consistent differences have been noted. For example, cell lines established from enzymatically dissociated tumor fragments were not consistently more drug- or radiationsensitive when compared to the companion line established from the same tumor through the use of mechanical techniques. Thus, the heterogeneity described in this report probably reflects geographic distribution of different populations of cells in different microregions of the tumor. In studies that assessed inherent radiosensitivity, little
167
difference in radiosensitivity was observed for the paired cell lines isolated from tumors MO67 and MO71 (SER = 0.36 and 0.30 for M067; 0.19 and 0.14 for MO7 1). However, a large difference in radiosensitivity was noted for the two cell lines from tumor MO59 (SF2 of 0.02 and 0.64 for M059J and M059K cells, respectively). In larger studies that have investigated the radiosensitivity of more radioresponsive tumor types (i.e., cervix, head and neck), intratumoral differences in SF2 values of this magnitude have not been observed (4, 16). Clearly, more gliomas must be studied before it can be ascertained whether the extreme differences in radiosensitivity observed for MO59 tumor cells commonly occur. Note that with the exception of M059K cells, all other cell lines tested for these studies were moderately radiosensitive. In other studies in our laboratory, we have characterized the SF2 values of 20 malignant glioma biopsies and have observed a mean (& SD) SF2 of 0.43 + 0.20 (Allalunis-Turner, Ban-on, Day, Fulton and Urtasun, unpublished data, December 1990). Thus. the glioma cell lines in this study are among the more radiosensitive that we have tested. Although malignant gliomas are clinically among the least radioresponsive of tumors, a large variation in inherent radiosensitivity is likely to exist as has been previously determined for other tumor types (4, 16). When the cell lines in this study were treated with rIFN prior to radiation, the radiosensitivity of cells from tumor MO67 was unaffected, whereas that of cells from tumors MO59 and MO71 was enhanced to varying degrees (range of SER = 0.7-2.0). Other investigators who have examined the effects of interferons on radiosensitivity have reported similar values (5). Studies by Chan and Keng have suggested that alteration in cell cycle distribution following rIFN treatment may contribute to its radiosensitizing effects (5). However, in the present study, no difference in cell cycles was noted in any of the six clones following exposure to rIFN. In addition, the degree of radiosensitization observed was independent of the antiproliferative effects of rIFN, as colony growth in all rIFN-treated cultures was reduced to a similar level (40-50s). Other investigators have suggested that rIFN’s sensitizing effects may result from altered ability to accumulate sublethal damage (7). Further work is required to test whether this is a factor in the sensitivity observed among these cell lines. In contrast to the limited information available concerning radiosensitivity, heterogeneity in chemosensitivity both within and among tumors has been well described for gliomas (13). Our observation that MO59 cells, for example, are both sensitive to BCNU but resistant to HN2 is consistent with this finding. We have also observed significant differences in response to the chemosensitizing effects of BSO among the tumor cell lines tested. Only cells from tumor MO59 were sensitized by BSO, although these lines differed in their response to BCNU and HN2. The four remaining cell lines were not sensitized by BSO. Other studies from our laboratory using early passage malignant glioma cell lines or freshly dissociated cells ob-
768
I. J. Radiation
Oncology
0 Bio!ogy 0 Physics
tained from glioma biopsies have demonstrated a similar variation in response to BSO’s chemosensitizing effects. These results have also suggested that chemosensitization by BSO is independent of initial GSH concentration. extent of GSH depletion, or Mer phenotype status (1) and Allalunis-Turner, Barron, Day, Fulton and Urtasun, unpublished data, June 1990). Other studies have reported variations in BSO’s effect on the chemosensitivity of a single cell line treated with a variety of drugs as well as its effects on different tumor types treated with the same agent (reviewed in 10). Thus, our results extend this observation and suggest that significant intratumoral heterogeneity may also exist with regard to BSO’s chemosensitizing effects. In conclusion, these preliminary results suggest that for
Volume 22. Number 4. 1992
human malignant gliomas, significant differences in radioand chemosensitivity can exist both within a single tumor and among different tumors. In addition, the ability of chemical agents to modify the treatment sensitivity may vary among the sub-populations of one tumor. It remains to be determined whether similar heterogeneity in treatment response, especially to chemical modifiers, occurs in other tumor types, or whether such differences are limited to tumors such as malignant gliomas in which heterogeneity is a “hallmark.” It will also be important to determine whether intratumoral differences in treatment sensitivities such as those described here are characteristic of tumors that are clinically radioresistant, such as gliomas and melanomas.
REFERENCES I. Allalunis-Turner, J.; Day, R. S.. 111;Urtasun. R. C.: Fulton. D. S.; Huyser-Wierenga, D. Mer phenotype, glutathione and inherent radiosensitivity as predictors of treatment response in patients with malignant glioma. (Abstr.). Proc. Amer. Assoc. Cancer Res. 30:288; 1989. 2. Bigner, D. D.; Bigner. S. H.: Ponten, J.: Westermark. B.; Mahaley, M. S.; Ruoslahti, E.: Herschman. H.; Eng, L. F.; Wikstrand. C. J. Heterogeneity of genotypic and phenotypic characteristics of fifteen permanent cell lines derived from human gliomas. J. Neuropathol. Exp. Neural. 15:201-229; 1981. 3. Bleehan. N. M.; Freedman. L. S.: Stenning, S. P. A randomized study of CCNU with and without Benznidazole in the treatment of recurrent grades 3 and 4 astrocytoma. lnt. J. Radiat. Oncol. Biol. Phys. 16:1077-1081; 1989. 4. Brock, W. A.; Baker, F. L.: Wike. J. L.; Sivon, S. L.; Peters, L. J. Cellular radiosensitivity of primary head and neck squamous cell carcinomas and local tumor control. Int. J. Radiat. Oncol. Biol. Phys. 18: l283- 1286: 1990. 5. Chang. A. Y. C.; Keng, P. C. Potentiation of radiation cytotoxicity by recombination interferons, phenomenon associated with increased blockage at the G2-M phase of the cell cycle. Cancer Res. 47:4338-434 I ; 1987. 6. Clevenger. C. Y.: Bauer, K. D.: Epstein, A. L. A method for simultaneous nuclear immunofluorescence and DNA content quantitation using monoclonal antibodies and flow cytometry. Cytometry 6:208-2 14: I985 7. Dritschilo. A.; Mossman, K.: Gray. M.; Sreevalsan, T. Potentiation of radiation injury by interferon. Amer. J. Clin. Oncol. 5:79-82; 1982. 8. Fulton. D. S.; Urtasun, R. C.: Shin, K. H.: Geggie, P. H. S.; Thomas. H.: Muller. P. J.; Moody, J.; Tanasichuk. H.: Mielke, B.: Johnson, E.; Curry, B. Misonidazole combined with hyperfractionation in the management of ma-
9.
IO.
I I.
12.
13. 14.
15.
16.
lignant glioma. Int. J. Radiat. Oncol. Biol. Phys. IO: I7091712: 1984. Kinsella. T. J.: Collins, J.: Rowland, J.; Klecker, R. Jr.; Wright, D.; Katz, D.; Steinberg, S. M. Pharmacology and phase I/II study of continuous intravenous infusion of iododeoxyuridine and hyperfractionated radiotherapy in patients with malignant glioblastoma multiforme. J. Clin. Oncol. 6:87 I-879: 1988. Jordan. J.; d’Arcy Doherty. M.; Cohen, G. M. Effects of glutathione depletion on the cytotoxicity of agents toward a human colonic tumor cell line. Br. J. Cancer 55:627-633: 1987. Mossman, T. Rapid calorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65:55-63: 1983. Phuphanich. S.: Levin. E. M.; Levin. V. A. Phase I study of intravenous bromodeoxyuridine used concomitantly with radiation therapy in patients with primary malignant brain tumors. Int. J. Radiat. Oncol. Biol. Phys. IO: 1769-1772; 1984. Rankin Shapiro, J.; Shapiro, W. R. Clonal tumor cell heterogeneity. Prog. Exp. Tumor Res. 27:49-66; 1984. Rosenblum, M.: Gerosa. M.; Dougherty, D. V.; Reese, C.: Barger, G. R.; Davis, R. L.; Levin, V. A.: Wilson, C. B. Agerelated chemosensitivity of stem cells from human malignant brain tumors. Lancet i:885-887; 1982. Tietze, F. Enzymatic method for determination of nanogram amounts of total and oxidized glutathione: application to mammalian blood and other tissues. Anal. Biochem. 27: 502-522; 1969, West, C. M. L.: Davidson, S. E.; Hunter, R. D. Evaluation of surviving fraction at 2 Gy as a potential prognostic factor for the radiotherapy of carcinoma of the cervix. Int. J. Radiat. Biol. 56:761-765; 1989.
