Increasing the susceptibility of the rat 208F fibroblast cell line to radiation-induced apoptosis does not alter its clonogenic survival dose-response.

Recent studies have suggested a correlation between the rate and incidence of apoptosis and the radiation response of particular cell lines. However, we found that increasing the rate of induction of apoptosis in the fibroblast line 208F, by transfecting it with human c-myc, did not lead to a change in its clonogenic survival dose-response for either gamma-irradiation or 125I-induced DNA damage. It was also found that expression of mutant (T24) Ha-ras in the 208F line appeared to decrease the level of apoptosis per mitosis after irradiation and inhibited the formation of nucleosomal ladders, but did not affect either the onset of the morphological features of apoptosis or the clonogenic survival dose-response of the cells to either gamma-irradiation or 125I-induced DNA damage. Our findings suggest that it may be incorrect to make predictions about the radiosensitivity of cells based only on knowledge of their mode of death. ImagesFigure 3Figure 4

The sensitivity of both normal and transformed cell types to killing by ionising radiation can differ markedly. For example, haemopoietic cell types are often radiation sensitive, while hepatocytes are relatively radiation resistant (Hendry, 1985). Similarly, lymphomas are typically responsive to irradiation, while melanomas often respond poorly (Steel, 1989). Although such generalisations about the dependence of radiation response on cell lineage have been useful in devising tumour radiotherapy regimens, a significant confounding factor is inter-individual differences in normal tissue and tumour response (Rofstad, 1986;Burnet et al., 1992).
Accordingly, there is considerable interest in understanding the mechanisms that produce such differences and thus devising assays that will more accurately predict tumour radioresponsiveness.
Recent work has focused attention on the possibility that susceptibility of normal or transformed cells to radiationinduced apoptosis may be an important indicator of radiosensitivity. Using a panel of mouse lymphoid or myeloid cell lines, all of which underwent apoptosis after irradiation, a correlation between the rapidity of induction of apoptosis and the clonogenic survival dose-response of a particular cell line was shown (Radford, 1994a). The greater sensitivity of these haemopoietic lines to radiation-induced DNA double-strand breaks as compared with fibroblast lines such as V79, which die by necrosis, suggested that radiosensitivity may be related to the mode of death (Radford, 1991(Radford, , 1994a. Similarly, in vivo studies with transplantable murine tumours showed that elevation of both spontaneous and radiationinduced apoptosis correlated positively with growth delay and negatively with TCD50 (dose to cure 50% of animals) (Meyn et al., 1993). However, these studies also showed a correlation between the incidence of apoptosis and tumour type, with several adenocarcinomas found to display increased apoptosis and radiosensitivity (longer growth delay and lower TCD%) when compared with several sarcomas.
Interestingly, one sarcoma was highly radiation sensitive in spite of showing no apoptosis following irradiation.
Because the cell lines used in the above studies have different origins, we were interested in investigating the relationship between mode and rapidity of cell death and radiosensitivity in cell lines with a common onrgin. Accordingly, we have examined the response to y-ray and DNAassociated "2I decay-induced damage of two transfectants of the rat lung fibroblast line 208F which express either human c-myc (cell line RBM7) or activated Ha-ras (cell line TI). Expression of these oncogenes has been shown to differentially alter susceptibility to both 'spontaneous' and serum withdrawal-induced apoptosis, with activated Ha-ras reducing the incidence of apoptosis over a 48 h period and c-myc increasing the incidence of apoptosis over this time (Arends et al., 1993). A similar pattern of high and low levels of apoptosis was seen in vivo using solid fibrosarcomas formed by the RBM7 and TI cell lines (Arends et al., 1994). Increased levels of apoptosis following growth factor withdrawal have also been demonstrated in Rat-l fibroblasts expressing c-myc (Evan et al., 1992) and in murine myeloid cells constitutively expressing c-myc (Askew et al., 1991). It was therefore of interest to determine how the expression of these genes affected the mode and rapidity of cell death following irradiation, and whether cellular sensitivity to DNA damage was altered.

Materials and methods
Cell lines: growth conditions and p53 status The derivation of the 208F, TI and RBM7 cell lines used in this study is described elsewhere (Spandidos and Wilkie, 1984;Arends et al., 1993). Cells were grown in the alpha modification of Eagle's medium (ICN/Flow) supplemented with 10% fetal calf serum (Commonwealth Serum Laboratones, Melbourne, Australia). Cells were incubated at 37TC in sealed flasks that had been flushed with 5% carbon dioxide, 5% oxygen and 90% nitrogen. Cells were subcultured by treatment with pronase (Calbiochem). All experiments were performed with asynchronous cell cultures in log-phase growth. Under these conditions, the cell lines had population doubling times of 16 (208F), 16 (T1) and 17 h (RBM7).
The p53 status of each of the cell lines was examined by immunoprecipitation as described in Radford (1994b). A positive result was found for extracts of all three cell lines using PAb 421 antibody, which binds to both wild-type and mutant p53 protein, whereas a negative result was found with PAb 240 antibody, which binds specifically to mutant p53 (Gannon et al., 1990). y-Irradiation, "I labelling and clonogenic assay Cultures were irradiated at room temperature in a _3Cs source at a dose rate of approximately 0.9 Gy min-'. Prior to irradiation, cultures were rinsed to remove any dead cells and fresh growth medium was added. Cell monolayers for clonogenic assay were treated with 0.03% pronase in phosphatebuffered saline (PBS) plus 0.2 mM EDTA for 5 min at room temperature and then dispersed by pipetting. After washing, cell suspensons were counted using a Coulter counter and appropriate cell numbers were plated to give 50-100 colonies in each of five replicate Petri dishes. Colonies of > 50 cells were scored after 8 days' incubation at 37C. Mean (± s.e.) cloning efficiencies for 208F, TI and RBM7 were respectively 0.33 ± 0.03, 0.42 ± 0.05 and 0.40 ± 0.04. '25I labelling was performed by incubating cultures for approximately 24 h in growth medium containing around 2.5 kBq ml-' ['OIIiododeoxyuridine (NEN/DuPont) and 2.5 gLM thymidine. Cells were then washed and incubated in growth medium containing 20 pM thymidine and 20 gM deoxycytidine for 3 h. After chasing, incorporated "2I was measured by pelleting cells and counting radioactivity in a Compugamma CS (LKB) gamma counter. Pellets were then resupended in a known volume and aliquots were taken for counting to determine the number of "2I decays per cell per day. Cells were then aliquoted and frozen (at -1C per min in a controlled-rate freezer) in growth medium plus 10% dimethylsulphoxide (DMSO). Cell aliquots were removed from liquid nitrogen storage at various times, after known numbers of "2I decays per cell had occurred, and assayed for clonogenic survival. Further details are given in Radford (1991). Freezing and thawing did not significantly affect the cloning efficiency of these cell lines as evidenced by values of 0.33 ± 0.02, 0.35 ± 0.01 and 0.40 ± 0.02 for 208F, TI and RBM7 respectively.
A least-squares fit, using the criterion (SO -SE)2ISO, where SE is the estimated survival and So is the observed survival for each experimental point, was obtained for cell survival data using the KaleidaGraph (Abelbeck Software) program on a Macintosh computer. The tray survival data were fitted to the equation S= exp l-(aD + AD2)] where S is survival, D is dose and a and P are constants. '"I-decay survival data were fitted to the simple exponential S = AOexp(-DIDo), where S is survival, D is the number of 'lI decays and AO and Do are constants.
Electron microscopy and gel electrophoresis At 24 h intervals after irradiation, non-adherent cells were collected by giving flasks several sharp taps, removing the growth medium, and then rinsing the monolayer once with PBS-EDTA. The growth medium and PBS-EDTA wash were pooled and centrifuged and the cell pellets were then fixed on ice for 30 min in growth medium (without serum) containing 0.25% glutaraldehyde and 45 min in 2.5% glutaraldehyde before post fixation with osmium tetroxide. Cells were then embedded in Spurr's resin and sectioned. Sections of control cells were obtained by fixing and embedding cells in situ on glass coverslips. DNA degradation samples were processed as described previously . Approximately 2 Lg of DNA from each sample was electrophoresed on a 1.5% agarose gel using SPP-l/EcoRI DNA (Bresatec, Adelaide, Australia) as size markers. Pulsed-field gel electrophoresis (PFGE) was carried out on a CHEF apparatus with a hexagonal array of electrodes (Chu et al., 1986) for 24 h at 150V with a pulse time of 80s. Plugs were prepared by mixing 1 x 10' (unless otherwise stated) non-adherent or adherent cells (removed from flasks by pronase treatment as descnbed above) in growth medium with an equal volume of 1% low gelling temperature agarose (SeaPlaque, FMC) in balanced salt solution (BSS). Plugs were then placed in NDS (10mm Tris, 0.5 EDTA, 1% lauroylsarcosine, pH 9.5) plus 1 mg ml-' proteinase K for 60 min on ice followed by overnight incubation at 37C before eectrophoresis. Yeast chromosomes from Saccharomyces cerevisiae strain YP148 (Pyle et al., 1988) and A DNA (BRL) were used as markers. Following electrophoresis, gels were stained with ethidium bromide.
Meaurement of apoptosis, population expansion and mitotic fraction Non-adherent cells were collected over 24 h intervals as described above. They were then pelleted and counted using a haemocytometer. The level of apoptosis was determined by mixing equal volumes of non-adherent cell suspension and growth medium containing #g ml-1 ethidium bromide and 3 ig ml-' acridine orange and then scoring approximately 500 cells for apoptotic nuclear morphology by fluorescence microscopy. The number of adherent cells present was determined by treating the monolayer with pronase and Coulter counting the suspension. The fraction of non-adherent cells which was apoptotic was multiplied by the total number of non-adherent cells, divided by the total number of cells in both adherent and non-adherent fractions, and then multiplied by 100 to give the percentage apoptosis.
After irdiation, all flasks were rinsed and had fresh growth medium added at 24 h intervals. The non-adherent cells collected thus represented cells released from the monolayer over a 24 h period. A combination of 24 h medium changes and seeding flasks at appropriate cell numbers ensured that cell death was attributable to the effects of irradiation and not to medium depletion. Population expansion was defined as the number of cells in the monolayer at a given time divided by the number at the time of irradiation.
The level of post-irradiation mitotic activity was determined by incubation of cultures with the mitotic spindle poison nocodazole at 0.1 agml-' for 3h and then scoring the fraction of metaphase-arrested cells. Further details are given in .

Resus
Clonogenic suival dose-response is not markedly changed by over-expression of c-myc or T24-ras Clonogenic survival curves were obtained for the parent and transfected cell lines exposed to either 7-irradiation or DNAassociated "2I decays. The -tray survival curves suggest that the overexpression of c-myc or an activating mutation of Ha-ras in these cells does not markedly affect their radiosensitivity ( Figure 1). The similarity in the response of these cell lines is particularly evident at high doses. At low doses, the Ha-ras transfectant TI may show increased resistance leading to a slightly larger shoulder region as compared with the parent line. An activated ras-induced change in the shoulder region of the survival curve of a transfected line has been reported by others (Hermens and Bentvelzen, 1992).
In order to quantify more readily the level of DNA damage required for cell kIilling, the sensitivity of the three cell lines to DNA-associated "tI decays was measured. Radioactive decay of "zI atoms incorporated into cellular DNA produces high lnear energy transfer (LET)-type DNA damage (Charlton, 1986) and results in approximately one DNA double-strand break per decay event (Krisch and Sauri, 1975). The "2I data shown in Figure 2 suggest that there is no significnt difference in the number of DNA doubl-strand breaks required to produce a lethal event in each of the cell lines. Do values of 50 ± 2.5, 53 + 1.4 and 56 ± 2.5 "2I decays were obtained for 208F, TI and RBM7 respectively. These Do values suggest that around 50-56 "2I The data were fitted to a simple exponential function of accmulated decays and Do values of 50 ± 3, 53 ± 1 and 56 ± 3 "2I decays were obtained for 208F, TI and RBM7 respectively. Each data set was obtained from at least three independent experiments.
Fuge 3 Ekctron micrographs of control (unirradiated) and -tirradiated (12 Gy plus incubation at 3TC for 24 h) cells. decay-induced DNA double-strand breaks are required to produce a lethal event in each of these cell lines.

All cell lines used show apoptotic death following irradiation
The morphology associated with cell death in each of the three cell lines, following exposure to a -tray dose (12 Gy) that would reduce clonogenic survival to 0.5% of the control value, was determined. Both light microscopy of stained sections and fluorescence microscopy (after staining with ethidium bromide and acridine orange) suggested that dying cells showed cytoplasmic shrinkage and condensation and margination of nuclear chromatin (data not shown). These features are diagnostic for apoptosis (Arends and Wylie, 1991). The concusion that radiation induces apoptosis in each of the three cell lines was confirmed by electron microscopy ( Figure 3). These eltron micrographs showed characteristic features of apoptosis such as chromatin condensation and margmation to the nuclear periphery (Figure 3f), convolutions of the nuclar membrane ( Figure 3b) and eventual cellular break-up into apoptotic bodies (Figure 3d). Cells colcted over 0-24 or 48-72 h after irradiation were also examined by electron microscopy for features characteristic of necrotic cell death, such as cellular or organelle swelling.
Over these time intervals, no evidence indicative of necrotic cell death was found in any of the three cell lines.
The echanism of cell death after iradiation was further examined by analysis of the pattern of DNA dgradation occurring in dying cells. One of the features often associted with apoptosis is the digestion of nuclear DNA into fragments that are multiples of the 180 bp nucleosome unit (Wyllie, 1980;Arends et al., 1990). When eletrophoretically separated on an agarose gel, these DNA fragments produce a chaaeristic 'ladder' pattern. At various time points after irradiation, DNA was extrcted from control (non-irradiated) monolayers and separately from the adherent and nonadherent cell fractions of iradiated culures. The parent ine 208F and the c-myc-transfected ine RBM7 both showed clear DNA ladders charactersti of apoptosis in the nonadherent cell fraction 24 h after irradiation (Figure 4a). However, the Ha-ras-transfected ie TI showed no ladder after 24h (Figure 4a). DNA extracted and analysed from nonadherent Tl cells at 36 or 48 h after irradiation still did not show laddering of the DNA (Figure 4b). DNA was also extracted from all three cell lines at time points prior to 24 h post irradiation, but ladders were not d (data not shown). DNA degradation in T cells was then examined further usng PFGE of cells incubated for 24 or 48 h at 3TC after 12 Gy of t-irradiation. II each case, the DNA from 1 x 106 cells was examined, except for the 24 h non-adherent fraction where only 0.5 x 10' ceils were available. The adherent fraction of the irradiated cell cultures, and to a lesser extent the control sample, showed a broad range of DNA sizes, indicating random gradation. DNA from the nonadherent fraction of irradiated TI cultures incubated for 48h, showed more marked degradation and the possible appearance of a lOO1kbp intermediate that is subsequently degraded further (Figure 4c). Elevated expresmion of c-myc increases the incidence of apoptosis following 7-trradiation Previous work using these cell lines had shown a difference in the incidence of apoptosis at 48 h after serum withdrawal (Arends et al., 1993). It was therefore of interest to examine the relative time of onset of apoptosis for asynchronous cultures of each of the cell lines following a 0.5% clonogenic survival ?-ray dose. At 24 h periods following irradiation, non-adherent cells were colleted, counted, and the percentage of cells undergoing apoptotic death was scored by fluorescence microscopy. Apoptotic cells were not found in adherent cell populations and non-adherent, non-apoptotic cells were found to account for less than 3.5% of non-apoptotic cells. The results of this experiment showed that the c-myc- A ad daim DR Alindge et a * of radiation-induced apoptosis, over the time interval examined, than the other lines ( Figure 5a). However, the expression of activated Ha-ras (TI) did not alter the frequency of apoptosis compared with the parent line (208F) up to 96 h post irradiation (Figure 5a). Adherent cells were also counted and RBM7 was found to maintain a similar number of cells on the monolayer as the parent line in spite of its higher rate of apoptosis. Cell numbers of TI, however, continued to increase up to 72 h post irradiation (Figure 5b). In order to determine the level and time of resumption of mitotic activity, replicate cultures of the three cell lines were incubated with the mitotic spindle poison nocodazole for successive 3 h intervals following 12 Gy of y-irradiation (Figure 5c). Following an initial marked depression of mitotic activity, a wave of mitosis was noted, which suggested a partial cell cycle-synchronising effect of irradiation on each of the lines. Cultures of both the myc-and rastransfected lines (RBM7 and TI) resumed cycling more rapidly and showed a higher fraction of cells entering mitosis than the parental 208F line. In loto, the data in Figure 5 suggest that, relative to the parental cell line, the level of apoptosis per mitosis is similar in the myc-transfected line but might be dereased in the ras-transfected line. Assuming that radiation-induced apoptosis occurs after mitosis in these cell lines, the increased incidence of apoptosis in the myctransfected line would then be a consequence of its higher (relative to the parental line) level of post-irradiation mitotic activity. Time-lapse cinemicroscopy studies will be required in order to confirm these conclusions. 575 0-2 hr 24-4en h 4-2n h 2-9 n The 208F rat fibroblast line and its m1'cand ras-transfected derivatives RBM7 and TI all undergo radiation-induced apoptosis as evidenced by morphology. This response is broadly in keeping with previous observations of 'sponb taneous' cell death of these cell lines in culture by apoptosis, 5 (albeit at high and low levels respectively for RBM7 and TI) (Arends et al., 1993) and also the pattern of death observed 4in solid tumours: RBM7 fibrosarcomas showed a high level of apoptosis and very little necrosis, but TI demonstrated widespread necrosis with low levels of apoptosis (Arends et  3 In an effort to explain differences in response to irradiation between cell types, many groups have looked at the possible link between oncogene expression and radiosensitivity (reviewed in Kasid et al., 1993). The consequences of trans- IV3JWL. L^_FIV33V U MUaVS IH 41UaLr WC W VI overexpression of normal ras have been reported to increase the radioresistance of mouse 3T3 fibroblasts and of rat rhabdomyosarcoma cells (Sklar, 1988;Samid et al., 1991;Hermens and Bentvelzen, 1992), while studies using transfected normal or immortalised human cells found that expression of activated ras does not, by itself, lead to an increase in radioresistance (Mendonca et al., 1991;Su and Little, 1992). We found that expression of activated Ha-ras in the TI derivative of the 208F fibroblast line had no marked effect on its clonogenic survival dose-responses for y-ray or  (1993), who showed that this endonuclease(s) is downregulated by the expression of actvted ras. Despite lacking detectable nucleosomal ladder-producing endon se activity, irradiated TI cells showed extesive DNA degradation to relatively high molkular weight fragments and normal apoptotic morphology. Apoptosis without the appearance of nucleosomal ladders has been reported previously (Cohen et al., 1992), and it has been shown that the characteristic apoptotic nuclear morphology is associated with inital cleavage of nuclear DNA into 300 kbp fragments and then into 50 kbp fragmets and does not require the production of 180 bp nudeosomal ladders (Brown et al., 1993;Oberhammer et al., 1993).
The RBM7 cell ine, which overexpresses c-myc, showed an increase in the incidence of radiation-induced apoptosis, over the time period examined, with respect to the parent line. This finding is consistent with the observations of Arends et al. (1993Arends et al. ( , 1994) that the RBM7 line shows increased susceptibility to serum withdrawal-induced apoptosis and increased intrinsic apoptosis when growing as a solid fibrosarcoma or in culture, and with the hypothesis that c-myc expression induces a state in which cells are 'primed' for apoptosis (Arends and Wylie, 1991). The increase in apoptosis may reflect a more rapid post-irradiation resumption of mitotic activity in the RBM7 cell line.
However, despite their inceased rate of i-irradiationinduced apoptotic death, the sensitivity of RBM7 cells to either ?-irradiation or DNA-associated '2I decay-induced cell killing is not significantly different from that of the parent line. This result contrasts with previous data from mouse lymphoid lines, which revealed a correlation between rate of radiation-induced apoptotic death and radiosensitiity (Radford, 1994a). However, it should be noted that, even in the RBM7 line, apoptosis is induced considerably more slowly in irradiated fibroblasts than in the more radiosensitive lymphoid lines. For example, apoptotic cells were not det in RBM7 cultures until at least 8 h after irradiation (data not shown), as compared with 1-2 h in radiosensitive lymphoid ines (Radford, 1994a). The comparatively lengthy time period available to irradiated RBM7 cells, prior to possible induction of apoptosis, may be adequate for DNA repair.
The reason for the difference in the rate of induction of apoptosis between irradiated fibroblasts and some lymphoid lines is currently unclear. Studies with mouse lymphoid ines have s ed that rapid induction of apoptosis after iradiation is dependent upon the presence of wild-type p53 protein (Radford, 1994b). Although all three fibroblast lines used in this study appear to contain non-mutant p53 protein, a definitive conclusion awaits DNA sequencing (see Materials and methods). Studies by other investigators, although not defining the mode of cell death occurring, have also generally concluded that overexpression of c-myc does not alter radioresistance (reviwed in Kasid et al., 1993).
Tese findings lead us to question the hypothesis that a cell's radiosensitivity can be directly related to its mode of death. Indeed, the number of "5I decay-induced DNA double-strand breaks required to produce a lethal event in the 208F ine and its transfected derivatives was smilar to the nunmber required to kill the V79 fibroblast line (Do = 61 ± 2), which undergoes necrotic cell death (Radford, 1991). This sugests that differences in radiosensitivity between cell lines may be related more to the intrinsic characteristics of the cell type of origin than to mode of death and that radiosensitivity is a phenotypic property distinct from susceptibility to apoptosis and that it may be independently genetically influenced. However, there is a need to study the relationship between radiosensitivity and mode of cell death in a wider range of cell types before such conclusions can be confirmed. It should also be noted that the significant difference in residual post-irradiation proliferation between the cell lines studied may complicate the extrapolation of our results to the in vivo situation.  (1986). Flow cytometry and DNA dgradation ha isi of two types of ceDl death. FEDS Lett.,194,[347][348][349][350] ARENDS MJ AND WYLLIE AH. (1991). Apoptosis: nisms and roes in pathology. It. Rev. Exp. Padl.,32,. ARENDS M, MORRIS RG AND WYLLIE AHl (1990). Apoptosis: the role of the endonuelas. Am. J. Pathal., 136, 593-608. ARENDS MJ, MCGREGOR AH, TOFT NJ, BROWN EJH AND WYLLIE AL (1993). Susceptibility to apoptosis is differentiajly regulated by c-myc and mutated Ha-ras oncognes and is aiated with endonue availability. Br. J. Cancer, 60, 1127-1133. ARENDS MJ, MCGREGOR AH AND WYLLIE All. (1994). Apoptosis is inversely relatei to necrosis and determine net growth in tumours bearing constitutively expressed myc, ras and HPV oncognes. Am. J. Padul., 144, 1045-1057. ASKEW DS, ASHMUN RA, SIMMONS BC AND CLEVELAND JL.