Inf J Radrorwn Ondog~ Bml. Phm “01 Pruned I” the U.S.A All r,ghts reserved.
22. pp. 103-108 CopyrIght
036G3016192 $5.00 + 00 C 1991 Pergamon Press plc
0 Biology Original Contribution
HETEROGENEITY
OF RADIOSENSITIVITY
IN A HUMAN GLIOMA CELL LINE
X. YANG, ’ J. L. DARLING,* T. J. MCMILLAN,’ J. H. PEACOCKI AND G. G. STEEL’ ‘RadiotherapyResearch Unit, Institute of Cancer Research, Cotswold Road, Sutton, Surrey SM2 5NG; and ‘Gough-Cooper Department
of Neurological
Surgery, Institute of Neurology,
Queen Square, London WCIE 3BG, U.K.
Sixteen clones were isolated from an early-passage human glioma cell line (IN859) and have been found to show variation in several biological characteristics including DNA content, modal chromosome number, and morphology. In addition, heterogeneity of radiosensitivity was detected: the doses that gave a surviving fraction of 0.01 varied by a factor of approximately 1.5. The most sensitive (clone 6) and the most resistant (clone 9) clones were selected for further study; their surviving fractions at 2Gy (SF2) were 0.37 and 0.64, respectively. When compared at a fixed radiation dose the sensitive clone surprisingly demonstrated greater split-dose recovery than the resistant clone; it also showed greater low dose-rate sparing. Radiosensitivity,
Glioma, Heterogeneity,
Cellular recovery.
This line was not cloned prior to the current experiments, in which it was used in passages 8-30.
INTRODUCTION Heterogeneity of cellular characteristics is a common finding in both experimental tumour systems and in human tumours in situ (11). This has practical implications for clinical studies that are based on small tumor biopsies, for experimental studies based on clonal cell lines, and it also has important implications for the response of tumors to treatment. Clonal variation in sensitivity to chemotherapeutic agents has frequently been observed (10, 16, 23), and drug-resistant cells can be isolated by various selection procedures (8, 9, 15, 20). This is often implicated in the emergence of drug resistance during protracted chemotherapy. Clonal variation in tumors with respect to ionizing radiation sensitivity has been described in a number of reports: Hill et al. (12) in the B16 melanoma, De Wyngaert et al. (6) in a mammary adeno-carcinoma, Brouwner et al. (2) in Ll210 murine leukemia, Leith ef al. (14) in human colon and lung carcinomas, Welch er al. (26) in a murine mammary adeno-carcinoma, and Weichselbaum et al. (25) in a human epidermoid carcinoma. It was our aim in this study to study cellular heterogeneity in a human glioma cell line and to examine whether radiation resistant or sensitive cells exist in this system. This has then led to a study of cellular recovery in resistant and sensitive glioma sublines. METHODS
Culture conditions The cells used for experimentation were monolayer cultures grown in Ham’s F12 medium with 15% fetal calf serum and 20 kg/ml L-glutamine. They were gassed with 92% N,, 5% CO,, and 3% 0,.
Isolation of clones IN859 cells in log phase were trypsinized to form a single-cell suspension. The suspended cells were plated in 96well plates using dilutions down to a mean of 0.1 cells per well. After 2 weeks, wells containing a single clone were marked. After another week, the clones had reached confluence in the wells and the cells were then trypsinized and transferred into 25 cm2 plastic flasks for monolayer culture. A second cycle of limiting dilution was then performed to produce 16 independent clones. The studies reported here were performed in passages 2-12 of the clones.
Cell survival assay Cells in log phase were detached using trypsin: versene (0.05%: 0.02%), centrifuged, resuspended in full medium, diluted to the final required concentration, and plated in 25 cm2 plastic flasks. Cells were treated 2 hr after plating. Colonies with at least 50 cells were counted 2-3 weeks after plating. Survival was expressed as surviving fraction (S.F.) where:
AND MATERIALS
Cell line IN859 was derived at the Institute of Neurology from a grade IV astrocytoma which arose in a 72-year-old woman. Reprint requests to: T. J. McMillan. Acknowledgements-We are grateful to S. Stockbridge and R. Couch for preparing this manuscript and Dr. M. Ormerod and
Mr. P. Imrie for running cells on the flow cytometer. Accepted for publication 24 May 199 1. 103
104
I. J. Radiation
Oncology
Volume 22, Number 1, 1992
0 Biology 0 Physics
Table 1. Biological properties
Clone IN859 Clone Clone Clone Clone
Volume (pm3)
DNA index*
4060 4580 4080 2100 4370
2.1 1.65 1.41 1.38 1.35
4300 4800
1.35 1.56
1 3 6 8
Clone 9 Clone 13
of the clonal cell lines Doubling time (h)
P.E.?
In vitro morphological type*
58 67 55 52 60
39 19.4 72.7 30.0 24.2
0.19 0.27 0.09 0.17 0.38
Fasicular Fasicular Glial Epithelial Fibroblastic
61 65
26.9 50.2
0.34 0.22
Fibroblastic Epithelial
Modal chromosome number
* DNA Content Index = Ratio of Gl tumor cell peak to Gl sheep erythrocyte t P.E. = plating efficiency. $ Classification according to Bigner et al. (1).
S.F. =
Plating efficiency
of treated cells
Plating efficiency
of control cells
peak.
at or below 7 cGy/min, a 1.7TBq (45Ci) @ko source was used. Cells were carefully maintained at 37°C during irradiation.
The survival data were fitted using the linear quadratic equation (Log(SF) = - aD - f3D2) using the error minimizing program of Koschel (17). Irradiation A 33TBq (900Ci) @?o source was used to irradiate cell cultures at an acute dose rate of 1.5 Gy/min. For dose rates
Karyotype analysis Following treatment with 0.4kg/ml colcemid for 6 hr, cells were harvested, incubated in 0.075M potassium chloride for 5 min, and fixed in methanol/glacial acetic acid (3: 1). Fixed cells were dropped onto microscope slides and stained with 4% Giemsa for 10 min. At least 50 cells were analyzed for each cell line. DNA content analysis A suspension of single cells was prepared in 200~1 phosphate buffered saline (PBS), then added to 2ml ice-
1
Table 2. Parameters of the linear-quadratic equation fitted to the acute survival data on the cell clones
.I
P Clone
.Oi
1
.ooi
.oooi’ 0
I
I
I
I
1
2 4 6 8 10 Radiation Dose by)
Fig. 1. Cell survival curves for the parent glioma cell line IN859 and for 16 clones. Irradiation at high dose rate (1.5 Gy/min). Dashed line is IN859 parent cell line.
*Ortho Cytofluorograf
50 M.
2 3 5 6 I 8 9 10 12 13 14 16 17 18 19
(G;- ‘1 0.131 0.287 0.421 0.146 0.43 0.385 0.133 0.21 0.433 0.271 0.224 0.143 0.331 0.627 0.128 0.498
(GY-*)
0.0478 0.0311 0.0136 0.0542 0.051 0.0391 0.0419 0.032 0.0363 0.0302 0.0358 0.0507 0.0337 0 0.0388 0.0262
Heterogeneity of radiosensitivity 0 X.
YANG et al.
105 7 cGv/min
150 cGv/min
1.
s
zI= .I
. ?? .
.Ol
F
.z
z s
.Oi
.
.l
.d
A
2 cGv/min
,001
.OOOl
.
~ 0
e
i
16 Dose (Gyl
Fig. 3. Cell survival data for clone 6 irradiated at three different dose rates: 1.5 Gy/min, 7 cGy/min, 2 cGy/min. The solid lines through the data were simultaneously fitted with the IR model. For reference, the line fitting the high dose-rate data is reproduced as a dashed line.
.ooo 1
sicular morphology. However, the clones derived from it varied in their morphological appearance. All three of Bigner’s morphological groups were represented among the clones. Chromosome number varied widely among the clones, as did cell doubling time and cell volume. DNA content did not vary so widely, and all clones had a lower Gl DNA content than the parent line.
.00001 t
0
5 10 Radiation Dose (Gy)
Fig. 2. Cell survival data for clones 6 and 9 at high dose rate. Cell survival in the parent line is shown by the dashed line.
Sensitivity of clones to high dose-rate irradiation Cell survival of the clones following high dose-rate irradiation is shown in Figure 1. The data points have been omitted for clarity, but each curve is the fit through the combined data of a minimum of two experiments. A spectrum of sensitivities is seen; note that the parent line is near the middle of the range. The parameters of these curves are shown in Table 2. The most sensitive (clone 6) and the most resistant (clone 9) clones were selected for further study, The full data for these two clones are shown in Figure 2, where it can be seen that the data do not overlap at any dose level. The parameters of the survival curves are given in Table 3. The surviving fraction at 2 Gy (SF2) differed by almost a factor of 2 between the clones (0.37 and 0.64).
cold ethanol/PBS (70:30). The cells were centrifuged, resuspended in 800~1 PBS, 1OOpJ ribonuclease (1 mg/ml), and 100 ~1 propidium iodide (400 p,g/ml), incubated at 37°C for 30 min, and then analyzed by flow-cytometry.* Ethanol-fixed sheep erythrocytes were stained in parallel and run before and after the samples. RESULTS
Biological characterization of clones Biological characteristics of six of the clones and the parent line are given in Table 1. According to the classification of Bigner et al. (l), the IN859 parent line had a fa-
Table 3. Cell survival pararmeters
following
acute irradiation
P*
Cell line IN859 Clone 6 Clone 9
0.31 * 0.013 (0.16) 0.43 f 0.07 (0.49) 0.21 ? 0.033 (0.27)
Pm4
(GY-‘I
SF2’r
(GY-‘)
0.034 f 0.0016 (0.056) 0.051 k 0.009 (0.050) 0.032 k 0.004 (0.028)
0.50
0.026 k 0.005
0.37
0.049 k 0.006
0.64
0.033 t
0.011
* Values include standard errors. t SF2 = surviving fraction at 2Gy from calculated fit to acute survival data (Figure 2). $ p-value derived from multiple split-dose experiments (see text). Note: (Y and l3 values in parentheses were derived using the Incomplete Repair model (see text).
106
I. J. Radiation 150 cGy/min
Oncology
0 Biology 0 Physics
7 cGy/min
Dose
Volume 22, Number
I, 1992
2 cGy/min
IGy)
Fig. 4. Cell survival data for clone 9 at three different dose rates. See Figure 3.
Sensitiviq of the clones to low dose-rate irradiation The sensitivity of IN859 at different dose rates has been described previously (28); in the present study the sensitivity of clones 6 and 9 was assessed at the low dose rates of 7 and 2 cGy/min. These dose rates were chosen to allow opportunity for cellular recovery but not to allow a significant amount of cell proliferation to occur during radiation exposure. At 2 cGy/min the exposure time for 12 Gy was 14 hr, a relatively small duration in relation to the doubling times of the cell cultures (Table 1). Furthermore, it is likely that cell proliferation is inhibited by the radiation treatment, and especially at 7 cGy/min it may have been fully arrested (18). The results of the low dose-rate experiments are shown in Figures 3 and 4. Little dose-sparing was detected in clone 9 at 7 cGy/min, but significant sparing was seen at the lower dose rate and at both reduced dose rates in clone 6. The dose reduction factors (ratio of doses at low and high dose rate which reduce survival to 0.01, D,,,,) were 1.3 and 1.4 in clone 6 (7, 2 cGy/min, respectively) and 1.3 in clone 9 at 2 cGy/min. Surprisingly, therefore the more sensitive clone showed a slightly greater dose-sparing than the resistant clone when assessed in this way. The data in Figures 3 and 4 have been fitted by the Incomplete Repair model of Thames (24). The data at all dose rates are simultaneously fitted to identify the best values for the three parameters of the model: CY,l3, and the
10’
10’
kI
.
=
.
100
100
100 4
recovery half-time (see also Steel et al. (22)). The values for (Y and p obtained in this way are given in Table 3. They differ somewhat from the values derived from the acute survival data alone because they involve a compromise among the data at all three dose rates; their reliability depends upon the quality of fit by the Incomplete Repair model. To fit the data well, more dose-rates than have been used here are ideally required. It is therefore not surprising that in some cases the fit to the data is quite poor (e.g., clone 9 at 150 cGy/min) and that the derived recovery half time of 2.5 hr for clone 9 is longer than we have seen for many other tumor cells. Split-dose recovery Split-dose recovery has been evaluated in clone 6, clone 9, and the parent line at a number of dose levels, following the approach of Peacock et al. (19). In each experiment, time intervals of up to 200 min were allowed between two equal doses, and the maximum recovery ratio (RR,,,) was taken as the mean of at least three points between 120-200 min. The recovery ratio had always levelled off by 120 min. Analysis of the data was based on the prediction of the linear-quadratic cell survival equation that the relationship between RR,,, and dose is
. . :Ii;., .. Clone 9
Clone 6
Fig. 6. Survival curves (at 1.5 Gy/min) for 4 clones and the parent cell line in different passages.
.
Fig. 5. Relationship between the split-dose recovery ratio and dose for the 2 clones and for the parent cell line. The data are plotted as a function of 2d* where the dose split is (d + d). The full lines were obtained by linear regression.
where PRR is the quadratic term of the linear quadratic equation and d is the single dose in the split-dose experiment (19, 24). The suffix RR denotes that the l3 value was derived in this way rather than by fitting acute radiation cell survival curves; it is not certain that these two values are identical. A plot of In RR,,, against 2d2 should give a straight line of slope p. Figure 5 shows that our data on these three cell lines conform well to this relationship: in
Heterogeneity of radiosensitivity 0 X. YANGef al.
increased linearly with dose. The slope of each case RR,, the relationship for clone 6 is steeper than for the other two cell lines: at each dose level the recovery ratio was greater in the most sensitive clone. The values for PRR derived from these curves are given in Table 3. The stability of radiosensitivity of the clones The stability of radiosensitivity was tested in four clones (clones 1, 6, 9 and 13) and the parent line IN859 during repeated passage in vitro (Fig. 6). The cells were tested at passages 2, 7, and 12 for the clones and passages 8, 25, and 30 for parent line. The data for the parent line were particularly stable. Variation from one passage to another was greater in the clones, but there was no evidence for systematic trends in the data. DISCUSSION Clonal heterogeneity of drug response is a common finding with experimental tumors (10, 16, 23) but with respect to radiation sensitivity heterogeneity has less often been seen (14, 25). In this study we have demonstrated clonal variation in a number of cellular characteristics within a human glioma cell line, IN859. The cell cultures were seen to be morphologically heterogeneous and this was reflected in the variation among the isolated clones (Table 1). These experiments have produced evidence that the cell population of the original IN859 cell line was also heterogeneous with regard to radiosensitivity. Leith et al. (14) reported two human tumor cell lines, a lung and a colon carcinoma, and five subclones derived from each of them. They demonstrated significant differences in radiosensitivity between the clones and the parent lines. One surprising feature of their data, however, was that in each case the parent cell line had a sensitivity that was at one extreme of the range shown by the cell clones. The data produced here show, in contrast, that the parent line (IN859) had a radiosensitivity that was in the middle of the span of curves for the 16 cell clones (Fig. 1). The range of sensitivities found among the 16 clones was quite large. The doses required for a survival of lo-’ differed by a factor of 1.5. Some of this variation no doubt arose from experimental error, that is, from the scatter in the experimental points, but it seems unlikely that this gave rise to the whole of the variation. The repeat experiments on four clones and the parent line whose data are shown in Figure 6 support the view that the inter-experiment variation on a single cell clone was smaller than the difference between clones.
107
It has been claimed that tumor cell clones in vitro or in vivo show phenotypic drift in metastatic properties, biochemical properties, and sensitivity to various therapeutic modalities during growth, in contrast to mixtures of clones, which may remain stable for prolonged periods of growth (26). The present study has produced some evidence in support of this view (Fig. 6), but the difference was slight. Low dose-rate irradiation allows recovery from radiation damage to take place during irradiation. Previous work in this department has shown that in most human tumor cell lines this leads to considerable dose sparing (22). Although some radiosensitive cell lines, such as those derived from neuroblastomas, give little low-dose-rate sparing, the more resistant cell lines invariably are spared to an extent that is easily demonstrated. This is perhaps the first study in which the dose-rate effect has been examined in clones derived from a single human tumor cell line. Low dose-rate irradiation has been found to amplify the differences in sensitivity between cell lines. This was very clear in studies comparing tumors of different histological types (22). We had expected, therefore, that when the low dose-rate studies were done on the sensitive and resistant clones from IN859 the difference between these would also have been amplified. This does not seem to have been the case. It can be seen from Figures 3 and 4 that the difference between clones 6 and 9 was not greater at 2 cGy/min than at high dose rate, although it was still in the same direction. This implies that the difference between the clones can be described as a difference in the magnitude of the o-component of cell killing, a difference that amounts to a factor of about 1.8. A difference also in p is not ruled out by this, and indeed such a difference is demonstrated by the results shown in Table 2, derived in three different ways. The existence of clonal variation in radiation response presents the possibility that tumors may become radioresistant during treatment as the resistant clones are selected. This phenomenon has been suggested to occur in human tumors (25) but the evidence for it is scarce. Even in experimental systems radioresistance has proved to be difficult to develop (3). The exceptions to this are the studies by Courtenay (4, 5), Rynas and Newcombe (21), Dittrich et al. (7), and Juraskova and Drasil (13). Thus, even if clonal variation in radiosensitivity is a common feature of tumor cells, its influence on response to radiotherapy is still not clear. As pointed out by Yaes (27), the relatively small degree of resistance in clones such as clone 9 in the present study may explain the lack of cellular resistance during clinical treatment.
REFERENCES 1. Bigner, D. D.; Bigner, S. H.; Ponten, J.; Westermark, B.; Mahaley, J. R.; 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. Neurol. 3:201-229; 1981.
108
I. J. Radiation Oncology 0 Biology 0 Physics
2. Brouwner, M.; Smets, L. A.; Jongsma, A. P. M. Isolation and characterization of subclones of Ll210 murine leukemia with different sensitivities to various cytotoxic agents. Cancer Res. 43:2884-2888; 1983. 3. Conger, A. D.; Luippold, H. J. Studies on the mechanism of acquired radioresistance in cancer. Cancer Res. 17:897-903; 1957. 4. Courtenay, V. D. The response to continuous irradiation of the mouse lymphoma L5178Y grown in vitro. Int. J. Radiat. Biol. 9581-591; 1965. 5. Courtenay, V. D. Radioresistant mutants of L5178Y cells. Radiat. Res. 38:186203; 1969 6. De Wyngaert, J. K.; Leith, J. T.; Peck, R. A.; Bliven, S. F.; Zeman, E. M.; Mariono, S. A.; Glicksman, A. S. Differential RBE values obtained for mammary adenocarcinoma tumour cell subpopulations after 14.8 MeV neutron irradiation. Radiat. Res. 88:118-131; 1981. 7. Dittrich, W.; Hohne, G.; Schubert, G. Development of a radio-resistant strain of Ehrlich carcinoma in mice. In: Mitchell, J. S., Holmes, B. E., Smith, C. L. eds. Progress in radiobiology. London: Oliver and Boyd; 1956. 8. Frondoza, C. G.; Trivedi, S. M.; Humphrey, R. L. Development and characterization of a cyclophosphamide-resistant mouse plasmacytoma cell line. Cancer Treat. Rep. 66:15351544; 1982. 9. Goldie, J. H.; Coldman, A. J. A mathematical model for relating the drug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat. Rep. 38:1727-1733; 1979. 10. Heppner, G. H.; Dexter, D. L.; Denucci, T.; Miller, F. R.; Calabresi, P. Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor. Cancer Res. 38:3758-3763; 1978. 11. Heppner, G. H.; Miller, B. E. Tumor heterogeneity: biological implications and therapeutic consequences. Cancer Metast. Rev. 2:5-23; 1983. 12. Hill, H. Z.; Hill, G. J.; Miller, C. F.; Kwang, F.; Purdy, J. Radiation and melanoma: Response of B16 mouse tumor cells and clonal lines to in vitro irradiation. Radiat. Res. 80: 259-276; 1979. V.; Drasil, V. Response of lymphosarcoma 13. Juraskova, LS/BL cells to continuous irradiation. Radiat. Res. 100:553563: 1984. 14. Leith, .I. T.; Dexter, D. L.; De Wyngaert, J. K.; Zeman, E. M.; Chu, M. Y.; Calabresi, P.; Glicksman, A. S. Differential responses to X-irradiation of subpopulations of two heterogeneous human carcinomas in vitro. Cancer Res.
Volume 22, Number 1, 1992 42:2556-2561; 1982. 15. McMillan, T. 3.; Stephens, T. C.; Steel, G. G. Development of drug resistance in a murine mammary tumour. Br. J. Cancer 52:823-832; 1985. 16. McMillan, T. J.; Stephens, T. C.; Steel, G. G. Clonal varation in the sensitivity of a murine mammary carcinoma to melphalan. Br. J. Cancer 53:753-759; 1986. 17. Millar, B. C.; Fielden, E. M.; Millar, J. L. Interpretation of survival curve data for Chinese hamster cells, line V79, using the multitargaret, multitarget with initial slope, and alpha, beta equations. Int. J. Radiat. Biol. 33:599-603; 1978. 18. Mitchell, J. B.; Bedford, J. S.; Bailey, S. M. Dose-rate effects on the cell cycle and survival of S3 HeLa and V79 cells. Radiat. Res. 79:520-536; 1979. 19. Peacock, J. H.; Cassoni, A. M.; McMillan, T. J.; Steel, G. G. Radiosensitive human tumour cell lines may not be recovery deficient. Int. J. Radiat. Biol. 54:945-953; 1988. 20. Ptiesler, H. D. Treatment failure in AML. Blood Cells 8:585-602; 1982. 21. Rhynas, P. 0. W.; Newcombe. H. B. A heritable change in radiation resistance of strain L mouse cells. Exp. Cell Res. 21:326-331; 1960. 22. Steel, G. G.; Deacon, J. M.; Duchesne, G. M.; Horwich, A.; Kelland, L. R.; Peacock, J. H. The dose-rate effect in human tumour cells. Radiother. Oncol. 9:299-310; 1987. 23. Stephens, T. C.; Peacock, J. H. Clonal variation in the sensitivity of B16 melanoma to m-AMSA. Br. 3. Cancer 45: 821-829; 1982. 24. Thames, H. D. An ‘incomplete-repair’ model for survival after fractionated and continuous irradiations. Int. J. Radiat. Biol. 47:319-339; 1985. R. R.; Beckett, M. A.; Dahlberg, W.; 25. Weichselbaum. Dritschilo, A. Heterogeneity of radiation response of a parent human epidermoid carcinoma cell line and four clones. Int. J. Radiat. Oncol. Biol. Phys. 14:907-912; 1988. 26. Welch, D. R.; Milas, L.; Tomasovic, S. P.; Nicolson, G. L. Heterogeneous response and clonal drift of sensitivities of metastatic 13762NF mammary adenocarcinoma clones to gamma-radiation in vitro. Cancer Res. 43:6-10; 1983. 27. Yaes, R. J. Tumor heterogeneity, tumor size, and radioresistance. Int. J. Radiat. Oncol. Biol. Phys. 17:993-1005; 1989. 28. Yang, X.; Darling, J. L.; McMillan, T. J.; Peacock, J. H.; Steel, G. G. Radiosensitivity, recovery and dose-rate effect in three human glioma cell lines. Radiother. Oncol. 19:4956; 1990.
22. pp. 103-108 CopyrIght
036G3016192 $5.00 + 00 C 1991 Pergamon Press plc
0 Biology Original Contribution
HETEROGENEITY
OF RADIOSENSITIVITY
IN A HUMAN GLIOMA CELL LINE
X. YANG, ’ J. L. DARLING,* T. J. MCMILLAN,’ J. H. PEACOCKI AND G. G. STEEL’ ‘RadiotherapyResearch Unit, Institute of Cancer Research, Cotswold Road, Sutton, Surrey SM2 5NG; and ‘Gough-Cooper Department
of Neurological
Surgery, Institute of Neurology,
Queen Square, London WCIE 3BG, U.K.
Sixteen clones were isolated from an early-passage human glioma cell line (IN859) and have been found to show variation in several biological characteristics including DNA content, modal chromosome number, and morphology. In addition, heterogeneity of radiosensitivity was detected: the doses that gave a surviving fraction of 0.01 varied by a factor of approximately 1.5. The most sensitive (clone 6) and the most resistant (clone 9) clones were selected for further study; their surviving fractions at 2Gy (SF2) were 0.37 and 0.64, respectively. When compared at a fixed radiation dose the sensitive clone surprisingly demonstrated greater split-dose recovery than the resistant clone; it also showed greater low dose-rate sparing. Radiosensitivity,
Glioma, Heterogeneity,
Cellular recovery.
This line was not cloned prior to the current experiments, in which it was used in passages 8-30.
INTRODUCTION Heterogeneity of cellular characteristics is a common finding in both experimental tumour systems and in human tumours in situ (11). This has practical implications for clinical studies that are based on small tumor biopsies, for experimental studies based on clonal cell lines, and it also has important implications for the response of tumors to treatment. Clonal variation in sensitivity to chemotherapeutic agents has frequently been observed (10, 16, 23), and drug-resistant cells can be isolated by various selection procedures (8, 9, 15, 20). This is often implicated in the emergence of drug resistance during protracted chemotherapy. Clonal variation in tumors with respect to ionizing radiation sensitivity has been described in a number of reports: Hill et al. (12) in the B16 melanoma, De Wyngaert et al. (6) in a mammary adeno-carcinoma, Brouwner et al. (2) in Ll210 murine leukemia, Leith ef al. (14) in human colon and lung carcinomas, Welch er al. (26) in a murine mammary adeno-carcinoma, and Weichselbaum et al. (25) in a human epidermoid carcinoma. It was our aim in this study to study cellular heterogeneity in a human glioma cell line and to examine whether radiation resistant or sensitive cells exist in this system. This has then led to a study of cellular recovery in resistant and sensitive glioma sublines. METHODS
Culture conditions The cells used for experimentation were monolayer cultures grown in Ham’s F12 medium with 15% fetal calf serum and 20 kg/ml L-glutamine. They were gassed with 92% N,, 5% CO,, and 3% 0,.
Isolation of clones IN859 cells in log phase were trypsinized to form a single-cell suspension. The suspended cells were plated in 96well plates using dilutions down to a mean of 0.1 cells per well. After 2 weeks, wells containing a single clone were marked. After another week, the clones had reached confluence in the wells and the cells were then trypsinized and transferred into 25 cm2 plastic flasks for monolayer culture. A second cycle of limiting dilution was then performed to produce 16 independent clones. The studies reported here were performed in passages 2-12 of the clones.
Cell survival assay Cells in log phase were detached using trypsin: versene (0.05%: 0.02%), centrifuged, resuspended in full medium, diluted to the final required concentration, and plated in 25 cm2 plastic flasks. Cells were treated 2 hr after plating. Colonies with at least 50 cells were counted 2-3 weeks after plating. Survival was expressed as surviving fraction (S.F.) where:
AND MATERIALS
Cell line IN859 was derived at the Institute of Neurology from a grade IV astrocytoma which arose in a 72-year-old woman. Reprint requests to: T. J. McMillan. Acknowledgements-We are grateful to S. Stockbridge and R. Couch for preparing this manuscript and Dr. M. Ormerod and
Mr. P. Imrie for running cells on the flow cytometer. Accepted for publication 24 May 199 1. 103
104
I. J. Radiation
Oncology
Volume 22, Number 1, 1992
0 Biology 0 Physics
Table 1. Biological properties
Clone IN859 Clone Clone Clone Clone
Volume (pm3)
DNA index*
4060 4580 4080 2100 4370
2.1 1.65 1.41 1.38 1.35
4300 4800
1.35 1.56
1 3 6 8
Clone 9 Clone 13
of the clonal cell lines Doubling time (h)
P.E.?
In vitro morphological type*
58 67 55 52 60
39 19.4 72.7 30.0 24.2
0.19 0.27 0.09 0.17 0.38
Fasicular Fasicular Glial Epithelial Fibroblastic
61 65
26.9 50.2
0.34 0.22
Fibroblastic Epithelial
Modal chromosome number
* DNA Content Index = Ratio of Gl tumor cell peak to Gl sheep erythrocyte t P.E. = plating efficiency. $ Classification according to Bigner et al. (1).
S.F. =
Plating efficiency
of treated cells
Plating efficiency
of control cells
peak.
at or below 7 cGy/min, a 1.7TBq (45Ci) @ko source was used. Cells were carefully maintained at 37°C during irradiation.
The survival data were fitted using the linear quadratic equation (Log(SF) = - aD - f3D2) using the error minimizing program of Koschel (17). Irradiation A 33TBq (900Ci) @?o source was used to irradiate cell cultures at an acute dose rate of 1.5 Gy/min. For dose rates
Karyotype analysis Following treatment with 0.4kg/ml colcemid for 6 hr, cells were harvested, incubated in 0.075M potassium chloride for 5 min, and fixed in methanol/glacial acetic acid (3: 1). Fixed cells were dropped onto microscope slides and stained with 4% Giemsa for 10 min. At least 50 cells were analyzed for each cell line. DNA content analysis A suspension of single cells was prepared in 200~1 phosphate buffered saline (PBS), then added to 2ml ice-
1
Table 2. Parameters of the linear-quadratic equation fitted to the acute survival data on the cell clones
.I
P Clone
.Oi
1
.ooi
.oooi’ 0
I
I
I
I
1
2 4 6 8 10 Radiation Dose by)
Fig. 1. Cell survival curves for the parent glioma cell line IN859 and for 16 clones. Irradiation at high dose rate (1.5 Gy/min). Dashed line is IN859 parent cell line.
*Ortho Cytofluorograf
50 M.
2 3 5 6 I 8 9 10 12 13 14 16 17 18 19
(G;- ‘1 0.131 0.287 0.421 0.146 0.43 0.385 0.133 0.21 0.433 0.271 0.224 0.143 0.331 0.627 0.128 0.498
(GY-*)
0.0478 0.0311 0.0136 0.0542 0.051 0.0391 0.0419 0.032 0.0363 0.0302 0.0358 0.0507 0.0337 0 0.0388 0.0262
Heterogeneity of radiosensitivity 0 X.
YANG et al.
105 7 cGv/min
150 cGv/min
1.
s
zI= .I
. ?? .
.Ol
F
.z
z s
.Oi
.
.l
.d
A
2 cGv/min
,001
.OOOl
.
~ 0
e
i
16 Dose (Gyl
Fig. 3. Cell survival data for clone 6 irradiated at three different dose rates: 1.5 Gy/min, 7 cGy/min, 2 cGy/min. The solid lines through the data were simultaneously fitted with the IR model. For reference, the line fitting the high dose-rate data is reproduced as a dashed line.
.ooo 1
sicular morphology. However, the clones derived from it varied in their morphological appearance. All three of Bigner’s morphological groups were represented among the clones. Chromosome number varied widely among the clones, as did cell doubling time and cell volume. DNA content did not vary so widely, and all clones had a lower Gl DNA content than the parent line.
.00001 t
0
5 10 Radiation Dose (Gy)
Fig. 2. Cell survival data for clones 6 and 9 at high dose rate. Cell survival in the parent line is shown by the dashed line.
Sensitivity of clones to high dose-rate irradiation Cell survival of the clones following high dose-rate irradiation is shown in Figure 1. The data points have been omitted for clarity, but each curve is the fit through the combined data of a minimum of two experiments. A spectrum of sensitivities is seen; note that the parent line is near the middle of the range. The parameters of these curves are shown in Table 2. The most sensitive (clone 6) and the most resistant (clone 9) clones were selected for further study, The full data for these two clones are shown in Figure 2, where it can be seen that the data do not overlap at any dose level. The parameters of the survival curves are given in Table 3. The surviving fraction at 2 Gy (SF2) differed by almost a factor of 2 between the clones (0.37 and 0.64).
cold ethanol/PBS (70:30). The cells were centrifuged, resuspended in 800~1 PBS, 1OOpJ ribonuclease (1 mg/ml), and 100 ~1 propidium iodide (400 p,g/ml), incubated at 37°C for 30 min, and then analyzed by flow-cytometry.* Ethanol-fixed sheep erythrocytes were stained in parallel and run before and after the samples. RESULTS
Biological characterization of clones Biological characteristics of six of the clones and the parent line are given in Table 1. According to the classification of Bigner et al. (l), the IN859 parent line had a fa-
Table 3. Cell survival pararmeters
following
acute irradiation
P*
Cell line IN859 Clone 6 Clone 9
0.31 * 0.013 (0.16) 0.43 f 0.07 (0.49) 0.21 ? 0.033 (0.27)
Pm4
(GY-‘I
SF2’r
(GY-‘)
0.034 f 0.0016 (0.056) 0.051 k 0.009 (0.050) 0.032 k 0.004 (0.028)
0.50
0.026 k 0.005
0.37
0.049 k 0.006
0.64
0.033 t
0.011
* Values include standard errors. t SF2 = surviving fraction at 2Gy from calculated fit to acute survival data (Figure 2). $ p-value derived from multiple split-dose experiments (see text). Note: (Y and l3 values in parentheses were derived using the Incomplete Repair model (see text).
106
I. J. Radiation 150 cGy/min
Oncology
0 Biology 0 Physics
7 cGy/min
Dose
Volume 22, Number
I, 1992
2 cGy/min
IGy)
Fig. 4. Cell survival data for clone 9 at three different dose rates. See Figure 3.
Sensitiviq of the clones to low dose-rate irradiation The sensitivity of IN859 at different dose rates has been described previously (28); in the present study the sensitivity of clones 6 and 9 was assessed at the low dose rates of 7 and 2 cGy/min. These dose rates were chosen to allow opportunity for cellular recovery but not to allow a significant amount of cell proliferation to occur during radiation exposure. At 2 cGy/min the exposure time for 12 Gy was 14 hr, a relatively small duration in relation to the doubling times of the cell cultures (Table 1). Furthermore, it is likely that cell proliferation is inhibited by the radiation treatment, and especially at 7 cGy/min it may have been fully arrested (18). The results of the low dose-rate experiments are shown in Figures 3 and 4. Little dose-sparing was detected in clone 9 at 7 cGy/min, but significant sparing was seen at the lower dose rate and at both reduced dose rates in clone 6. The dose reduction factors (ratio of doses at low and high dose rate which reduce survival to 0.01, D,,,,) were 1.3 and 1.4 in clone 6 (7, 2 cGy/min, respectively) and 1.3 in clone 9 at 2 cGy/min. Surprisingly, therefore the more sensitive clone showed a slightly greater dose-sparing than the resistant clone when assessed in this way. The data in Figures 3 and 4 have been fitted by the Incomplete Repair model of Thames (24). The data at all dose rates are simultaneously fitted to identify the best values for the three parameters of the model: CY,l3, and the
10’
10’
kI
.
=
.
100
100
100 4
recovery half-time (see also Steel et al. (22)). The values for (Y and p obtained in this way are given in Table 3. They differ somewhat from the values derived from the acute survival data alone because they involve a compromise among the data at all three dose rates; their reliability depends upon the quality of fit by the Incomplete Repair model. To fit the data well, more dose-rates than have been used here are ideally required. It is therefore not surprising that in some cases the fit to the data is quite poor (e.g., clone 9 at 150 cGy/min) and that the derived recovery half time of 2.5 hr for clone 9 is longer than we have seen for many other tumor cells. Split-dose recovery Split-dose recovery has been evaluated in clone 6, clone 9, and the parent line at a number of dose levels, following the approach of Peacock et al. (19). In each experiment, time intervals of up to 200 min were allowed between two equal doses, and the maximum recovery ratio (RR,,,) was taken as the mean of at least three points between 120-200 min. The recovery ratio had always levelled off by 120 min. Analysis of the data was based on the prediction of the linear-quadratic cell survival equation that the relationship between RR,,, and dose is
. . :Ii;., .. Clone 9
Clone 6
Fig. 6. Survival curves (at 1.5 Gy/min) for 4 clones and the parent cell line in different passages.
.
Fig. 5. Relationship between the split-dose recovery ratio and dose for the 2 clones and for the parent cell line. The data are plotted as a function of 2d* where the dose split is (d + d). The full lines were obtained by linear regression.
where PRR is the quadratic term of the linear quadratic equation and d is the single dose in the split-dose experiment (19, 24). The suffix RR denotes that the l3 value was derived in this way rather than by fitting acute radiation cell survival curves; it is not certain that these two values are identical. A plot of In RR,,, against 2d2 should give a straight line of slope p. Figure 5 shows that our data on these three cell lines conform well to this relationship: in
Heterogeneity of radiosensitivity 0 X. YANGef al.
increased linearly with dose. The slope of each case RR,, the relationship for clone 6 is steeper than for the other two cell lines: at each dose level the recovery ratio was greater in the most sensitive clone. The values for PRR derived from these curves are given in Table 3. The stability of radiosensitivity of the clones The stability of radiosensitivity was tested in four clones (clones 1, 6, 9 and 13) and the parent line IN859 during repeated passage in vitro (Fig. 6). The cells were tested at passages 2, 7, and 12 for the clones and passages 8, 25, and 30 for parent line. The data for the parent line were particularly stable. Variation from one passage to another was greater in the clones, but there was no evidence for systematic trends in the data. DISCUSSION Clonal heterogeneity of drug response is a common finding with experimental tumors (10, 16, 23) but with respect to radiation sensitivity heterogeneity has less often been seen (14, 25). In this study we have demonstrated clonal variation in a number of cellular characteristics within a human glioma cell line, IN859. The cell cultures were seen to be morphologically heterogeneous and this was reflected in the variation among the isolated clones (Table 1). These experiments have produced evidence that the cell population of the original IN859 cell line was also heterogeneous with regard to radiosensitivity. Leith et al. (14) reported two human tumor cell lines, a lung and a colon carcinoma, and five subclones derived from each of them. They demonstrated significant differences in radiosensitivity between the clones and the parent lines. One surprising feature of their data, however, was that in each case the parent cell line had a sensitivity that was at one extreme of the range shown by the cell clones. The data produced here show, in contrast, that the parent line (IN859) had a radiosensitivity that was in the middle of the span of curves for the 16 cell clones (Fig. 1). The range of sensitivities found among the 16 clones was quite large. The doses required for a survival of lo-’ differed by a factor of 1.5. Some of this variation no doubt arose from experimental error, that is, from the scatter in the experimental points, but it seems unlikely that this gave rise to the whole of the variation. The repeat experiments on four clones and the parent line whose data are shown in Figure 6 support the view that the inter-experiment variation on a single cell clone was smaller than the difference between clones.
107
It has been claimed that tumor cell clones in vitro or in vivo show phenotypic drift in metastatic properties, biochemical properties, and sensitivity to various therapeutic modalities during growth, in contrast to mixtures of clones, which may remain stable for prolonged periods of growth (26). The present study has produced some evidence in support of this view (Fig. 6), but the difference was slight. Low dose-rate irradiation allows recovery from radiation damage to take place during irradiation. Previous work in this department has shown that in most human tumor cell lines this leads to considerable dose sparing (22). Although some radiosensitive cell lines, such as those derived from neuroblastomas, give little low-dose-rate sparing, the more resistant cell lines invariably are spared to an extent that is easily demonstrated. This is perhaps the first study in which the dose-rate effect has been examined in clones derived from a single human tumor cell line. Low dose-rate irradiation has been found to amplify the differences in sensitivity between cell lines. This was very clear in studies comparing tumors of different histological types (22). We had expected, therefore, that when the low dose-rate studies were done on the sensitive and resistant clones from IN859 the difference between these would also have been amplified. This does not seem to have been the case. It can be seen from Figures 3 and 4 that the difference between clones 6 and 9 was not greater at 2 cGy/min than at high dose rate, although it was still in the same direction. This implies that the difference between the clones can be described as a difference in the magnitude of the o-component of cell killing, a difference that amounts to a factor of about 1.8. A difference also in p is not ruled out by this, and indeed such a difference is demonstrated by the results shown in Table 2, derived in three different ways. The existence of clonal variation in radiation response presents the possibility that tumors may become radioresistant during treatment as the resistant clones are selected. This phenomenon has been suggested to occur in human tumors (25) but the evidence for it is scarce. Even in experimental systems radioresistance has proved to be difficult to develop (3). The exceptions to this are the studies by Courtenay (4, 5), Rynas and Newcombe (21), Dittrich et al. (7), and Juraskova and Drasil (13). Thus, even if clonal variation in radiosensitivity is a common feature of tumor cells, its influence on response to radiotherapy is still not clear. As pointed out by Yaes (27), the relatively small degree of resistance in clones such as clone 9 in the present study may explain the lack of cellular resistance during clinical treatment.
REFERENCES 1. Bigner, D. D.; Bigner, S. H.; Ponten, J.; Westermark, B.; Mahaley, J. R.; 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. Neurol. 3:201-229; 1981.
108
I. J. Radiation Oncology 0 Biology 0 Physics
2. Brouwner, M.; Smets, L. A.; Jongsma, A. P. M. Isolation and characterization of subclones of Ll210 murine leukemia with different sensitivities to various cytotoxic agents. Cancer Res. 43:2884-2888; 1983. 3. Conger, A. D.; Luippold, H. J. Studies on the mechanism of acquired radioresistance in cancer. Cancer Res. 17:897-903; 1957. 4. Courtenay, V. D. The response to continuous irradiation of the mouse lymphoma L5178Y grown in vitro. Int. J. Radiat. Biol. 9581-591; 1965. 5. Courtenay, V. D. Radioresistant mutants of L5178Y cells. Radiat. Res. 38:186203; 1969 6. De Wyngaert, J. K.; Leith, J. T.; Peck, R. A.; Bliven, S. F.; Zeman, E. M.; Mariono, S. A.; Glicksman, A. S. Differential RBE values obtained for mammary adenocarcinoma tumour cell subpopulations after 14.8 MeV neutron irradiation. Radiat. Res. 88:118-131; 1981. 7. Dittrich, W.; Hohne, G.; Schubert, G. Development of a radio-resistant strain of Ehrlich carcinoma in mice. In: Mitchell, J. S., Holmes, B. E., Smith, C. L. eds. Progress in radiobiology. London: Oliver and Boyd; 1956. 8. Frondoza, C. G.; Trivedi, S. M.; Humphrey, R. L. Development and characterization of a cyclophosphamide-resistant mouse plasmacytoma cell line. Cancer Treat. Rep. 66:15351544; 1982. 9. Goldie, J. H.; Coldman, A. J. A mathematical model for relating the drug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat. Rep. 38:1727-1733; 1979. 10. Heppner, G. H.; Dexter, D. L.; Denucci, T.; Miller, F. R.; Calabresi, P. Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor. Cancer Res. 38:3758-3763; 1978. 11. Heppner, G. H.; Miller, B. E. Tumor heterogeneity: biological implications and therapeutic consequences. Cancer Metast. Rev. 2:5-23; 1983. 12. Hill, H. Z.; Hill, G. J.; Miller, C. F.; Kwang, F.; Purdy, J. Radiation and melanoma: Response of B16 mouse tumor cells and clonal lines to in vitro irradiation. Radiat. Res. 80: 259-276; 1979. V.; Drasil, V. Response of lymphosarcoma 13. Juraskova, LS/BL cells to continuous irradiation. Radiat. Res. 100:553563: 1984. 14. Leith, .I. T.; Dexter, D. L.; De Wyngaert, J. K.; Zeman, E. M.; Chu, M. Y.; Calabresi, P.; Glicksman, A. S. Differential responses to X-irradiation of subpopulations of two heterogeneous human carcinomas in vitro. Cancer Res.
Volume 22, Number 1, 1992 42:2556-2561; 1982. 15. McMillan, T. 3.; Stephens, T. C.; Steel, G. G. Development of drug resistance in a murine mammary tumour. Br. J. Cancer 52:823-832; 1985. 16. McMillan, T. J.; Stephens, T. C.; Steel, G. G. Clonal varation in the sensitivity of a murine mammary carcinoma to melphalan. Br. J. Cancer 53:753-759; 1986. 17. Millar, B. C.; Fielden, E. M.; Millar, J. L. Interpretation of survival curve data for Chinese hamster cells, line V79, using the multitargaret, multitarget with initial slope, and alpha, beta equations. Int. J. Radiat. Biol. 33:599-603; 1978. 18. Mitchell, J. B.; Bedford, J. S.; Bailey, S. M. Dose-rate effects on the cell cycle and survival of S3 HeLa and V79 cells. Radiat. Res. 79:520-536; 1979. 19. Peacock, J. H.; Cassoni, A. M.; McMillan, T. J.; Steel, G. G. Radiosensitive human tumour cell lines may not be recovery deficient. Int. J. Radiat. Biol. 54:945-953; 1988. 20. Ptiesler, H. D. Treatment failure in AML. Blood Cells 8:585-602; 1982. 21. Rhynas, P. 0. W.; Newcombe. H. B. A heritable change in radiation resistance of strain L mouse cells. Exp. Cell Res. 21:326-331; 1960. 22. Steel, G. G.; Deacon, J. M.; Duchesne, G. M.; Horwich, A.; Kelland, L. R.; Peacock, J. H. The dose-rate effect in human tumour cells. Radiother. Oncol. 9:299-310; 1987. 23. Stephens, T. C.; Peacock, J. H. Clonal variation in the sensitivity of B16 melanoma to m-AMSA. Br. 3. Cancer 45: 821-829; 1982. 24. Thames, H. D. An ‘incomplete-repair’ model for survival after fractionated and continuous irradiations. Int. J. Radiat. Biol. 47:319-339; 1985. R. R.; Beckett, M. A.; Dahlberg, W.; 25. Weichselbaum. Dritschilo, A. Heterogeneity of radiation response of a parent human epidermoid carcinoma cell line and four clones. Int. J. Radiat. Oncol. Biol. Phys. 14:907-912; 1988. 26. Welch, D. R.; Milas, L.; Tomasovic, S. P.; Nicolson, G. L. Heterogeneous response and clonal drift of sensitivities of metastatic 13762NF mammary adenocarcinoma clones to gamma-radiation in vitro. Cancer Res. 43:6-10; 1983. 27. Yaes, R. J. Tumor heterogeneity, tumor size, and radioresistance. Int. J. Radiat. Oncol. Biol. Phys. 17:993-1005; 1989. 28. Yang, X.; Darling, J. L.; McMillan, T. J.; Peacock, J. H.; Steel, G. G. Radiosensitivity, recovery and dose-rate effect in three human glioma cell lines. Radiother. Oncol. 19:4956; 1990.
