Combined immunodeficiency associated with increased apoptosis of lymphocytes and radiosensitivity fibroblasts.

Severe immunodeficiency characterized by lymphopenia was found in two siblings, one of whom was examined in detail. The calcium flux, pattern of tyrosine phosphorylation of proteins, and interleukin 2 (IL-2) production and proliferation in response to mitogens suggested that the peripheral blood T cells activated normally. The peripheral blood T cells were shown to have an activated phenotype with increased expression of CD45RO+ and CD95/Fas. Increased spontaneous apoptosis occurred in unstimulated lymphocyte cultures. The elevated apoptosis was not due to alterations in expression or to mutations in Bcl-2, Bcl-X(L), or Flip, nor could the spontaneous apoptosis be prevented by blocking Fas, suggesting that it was independent of Fas signaling. This is the first inherited combined immunodeficiency associated with impaired lymphocyte survival. Fibroblasts derived from the patient showed appreciable radiosensitivity in clonal assays, but apoptosis was not elevated. Our results show that the fibroblasts represent a new radiosensitive phenotype not associated with cell cycle checkpoint defects, V(D)J recombination defects, or elevated chromosome breakage. We suggest that the affected gene plays a role in an undetermined damage response mechanism that results in elevated spontaneous apoptosis in lymphoid cells and radiosensitivity in fibroblasts.

Hypersensitivity to low single doses and split-dose recovery: two manifestations of induced resistance that might be related.

PURPOSE: To study retrospectively the relationship between intrinsic radiosensitivity (SF2), and both the low-dose inducible response (alpha(s)/alpha(r)) and the amount of split-dose recovery (betaRR). MATERIALS AND METHODS: A total of 53 sets of experimental data obtained with 44 human cell lines were collected from the literature and the above relationships were studied. RESULTS: Analysis showed a statistically significant correlation between alpha(s)/alpha(r) and SF2 (p = 0.0023, 10 sets of data), and a statistically significant inverse correlation between betaRR and SF2 (p = 0.0005, 36 sets of data, AT excluded). Furthermore, the analysis of the relationship between the challenge dose SF2 (after a clinical-sized priming dose) and that of the single-dose SF2 (27 sets of data, AT excluded) showed a statistically significant correlation (p<0.0001), which deviates from, and becomes higher than, the one-to-one relationship for single-dose SF2<0.30, suggesting that the final response to fractionated irradiation in radiosensitive cells might not be predictable on the basis of simple reconstitution of survival from the single-dose treatment. CONCLUSION: The comparison between the two relationships: SF2/(alpha(s)/alpha(r)) and SF2/betaRR, suggests some parallelism indicating that these two phenomena may be inversely correlated and could be attributed to induced resistance mechanisms that might be triggered differently in sensitive and resistant cell lines.

Underestimation of the small residual damage when measuring DNA double-strand breaks (DSB): is the repair of radiation-induced DSB complete?

PURPOSE: To overcome the underestimation of the small residual damage when measuring DNA double-strand breaks (DSB) as fraction of activity released (FAR) by pulsed-field gel electrophoresis. MATERIALS AND METHODS: The techniques used to assess DNA damage (e.g. pulsed-field gel electrophoresis, neutral elution, comet assay) do not directly measure the number of DSB. The Blöcher model can be used to express data as DSB after irradiation at 4 degrees C by calculating the distribution of all radiation-induced DNA fragments as a function of their size. We have used this model to measure the residual DSB (irradiation at 4 degrees C followed by incubation at 37 degrees C) in untransformed human fibroblasts. RESULTS: The DSB induction rate after irradiation at 4 degrees C was 39.1+/-2.0 Gy(-1). The DSB repair rate obtained after doses of 10 to 80 Gy followed by repair times of 0 to 24 h was expressed as unrepaired DSB calculated from the Blöcher formula. All the damage appeared to be repaired at 24h when the data were expressed as FAR, whereas 15% of DSB remained unrepaired. The DSB repair rate and the chromosome break repair rate assessed by premature condensation chromosome (PCC) techniques were similar. CONCLUSION: The expression of repair data in terms of FAR dramatically underestimates the amount of unrepaired DNA damage. The Blöcher model that takes into account the size distribution of radiation-induced DNA fragments should therefore be used to avoid this bias. Applied to a normal human fibroblast cell line, this model shows that DSB repair is never complete.

Repair of radiation-induced DNA double-strand breaks in human fibroblasts is consistent with a continuous spectrum of repair probability.

PURPOSE: To propose a novel interpretation of DNA double-strand break (dsb) repair based on the distribution of energy micro-deposition. MATERIALS AND METHODS: Double-strand break repair curves were studied either after irradiation at 4 degrees C or at 37 degrees C (low dose rate). Two human fibroblast cell lines were used: a control line, HF19, and an ataxia telangiectasia repair-deficient line, AT5BI. Irradiations were made with gamma-rays or alpha-particles (241Am). Repair data were fitted by the variable repair half-time (VRHT) model. Assuming that each dsb has its own inherent repair half-time (IRHT) and that the VRHT is the average of the IRHT at any time during repair, the distribution of the IRHT was calculated. RESULTS: At the end of the irradiation, the distribution was a continuous asymmetric curve with a maximum of dsb having a short IRHT. After 1 h of repair, the curve became bell-shaped. There is a striking similarity between the distribution of dsb repair half-times and that of energy micro-deposition described by Goodhead et al. (1993). CONCLUSION: This similarity suggests a possible causal relationship between the energy density deposition and the repair rate or the probability of dsb repair.

Tumour size in cancer of the cervix.

In a randomized trial comprising 204 patients with operable cervical carcinomas stages I and II, two low-dose rates in gynaecological brachytherapy were compared. Treatment consisted of Cs-137 uterovaginal application followed by surgery (either immediate or delayed). The results for the two dose rates have been published previously. The present paper concerns the correlation between outcome and tumour size. Tumour size was carefully estimated in two ways: by clinical examination under general anaesthesia and by measurements on the customized vaginal mould used for the brachytherapy. Ninety-one patients (45%) were classified as stage I, and 113 were classified as stage II proximal. The mean tumour size was 39 mm (range 15-64 mm). Cox's multivariate analysis indicated that the factors with a poor prognostic value were for survival: node involvement (N +) (p < 0.001), large tumour size (T +) (p < 0.001) and involvement of the endocervix (E +) (p <0.01); for event-free survival: N + (p <0.001), T + (p < 0.001); for local control; N + (p = 0.0001); for metastasis and regional relapse: N + (p < 0.001) and T + (p < 0.001). Stage was not a prognostic factor over the present range in either univariate or multivariate analysis. In this series tumour size is a powerful independent prognostic factor. It is therefore suggested that for the classification of cervical cancer and the indications for surgical staging and adjuvant treatment, tumour size should be taken into account.

Results of a European randomized trial of Etanidazole combined with radiotherapy in head and neck carcinomas.

PURPOSE: The aim of the study was to evaluate the efficacy and toxicity of Etanidazole, a hypoxic cell sensitizer, combined with radiotherapy in the treatment of head and neck squamous cell carcinoma. METHODS AND MATERIALS: A total of 374 patients from 27 European centers were included in this trial between 1987 and 1990. Treatment was either conventional radiotherapy alone (between 66 Gy in 33 fractions and 74 Gy in 37 fractions, 5 fractions per week), or the same radiotherapy dose plus Etanidazole 2 g/m2, three times weekly for 17 doses. A minimization procedure, balancing for center, site, and T stage (T1-T3 vs. T4) was used for randomization. RESULTS: Among the 187 patients in the Etanidazole group, 82% received at least 14 doses of the drug. Compliance to the radiotherapy protocol was 92% in the Etanidazole group and 88% in the control group; the main cause of deviation was acute toxicity, which was observed at an equal rate in the two treatment groups. Fifty-two cases of Grade 1 to 3 peripheral neuropathy were observed in the Etanidazole group vs. 5 cases, all of Grade 1, in the control group (p < 0.001). The 2-year actuarial loco-regional control rates were 53% in the Etanidazole group and 53% in the control group (p = 0.93), and the overall 2-year survival rates were 54% in each group (p = 0.99). CONCLUSION: Adding Etanidazole to conventional radiotherapy did not afford any benefit for patients with head and neck carcinoma. This study failed to confirm the hypothesis of a benefit for patients with N0-N1 disease, which had been suggested by the results of a previous study (10).

A DNA double-strand break defective fibroblast cell line (180BR) derived from a radiosensitive patient represents a new mutant phenotype.

The 180BR cell line was derived from an acute lymphoblastic leukemia patient who overresponded to radiation therapy and died following radiation morbidity. 180BR cells are hypersensitive to the lethal effects of ionizing radiation and are defective in the repair of DNA double-strand breaks (DSBs). The levels and activity of the proteins of the DNA-dependent protein kinase complex are normal in 180BR cells. To facilitate a measurement of V(D)J recombination, we have characterized 180BRM, a SV40-transformed line derived from 180BR. 180BRM retains the radiosensitivity and defect in DSB repair characteristic of 180BR. The activities associated with DNA-dependent protein kinase are also normal in 180BRM cells. The ability to carry out V(D)J recombination is comparable in 180BRM and a reference control transformed human cell line, MRC5V1. These results show that 180BR and 180BRM differ from the rodent mutants belonging to ionizing radiation complementation groups 4, 5, 6, and 7 and, therefore, represent a new mutant phenotype, in which a defect in DNA DSB rejoining is not associated with defective V(D)J recombination. Furthermore, we have shown that 180BR can arrest at the G1-S and G2-M cell cycle checkpoints after irradiation. These results confirm that 180BR can be distinguished from ataxia telangiectasia.

Hypersensitivity of ataxia telangiectasia fibroblasts to ionizing radiation is associated with a repair deficiency of DNA double-strand breaks.

We have studied the intrinsic radiosensitivity, repair of potentially lethal damage (PLD) and the repair rate of radiation-induced DNA double-strand breaks (DSB) in 11 non-transformed human fibroblast cell lines, four of which were homozygous for the A-T mutation and two that were heterozygous (A-TH). All the experiments were done on cells in plateau phase of growth (97-99% of cells in G0/G1). With a dose of 30 Gy delivered at 4 degrees C, the A-T cell lines had faster repair rates of up to 6 h, after which the repair curve crossed that of the control so that the residual damage at 24 h was higher in the A-T cells. Irradiation at 37 degrees C at low dose rate 1 cGy.min-1) produced even more marked differences between the A-T cells and controls: the residual DSB level was always higher in A-T cells than controls at doses of 5-40 Gy, due to defective repair of a small fraction of DSB in A-T cells. The two protocols showed DSB repair rates for the A-TH cell lines that were intermediate between those of the A-T and control cells. There was a quantitative relationship between the residual DSB after irradiation at 37 degrees C and the intrinsic radiosensitivity, and with the extent of PLD repair. There were very few apoptotic cells in the non-transformed control and A-T cell line, both before and after irradiation. In combination, these result support the contention that the defective repair of DSB is a mechanism of the hypersensitivity linked to the A-T mutation.

Radiation-induced DNA double-strand breaks and the radiosensitivity of human cells: a closer look.

A large number of reports suggest that DNA double-strand breaks (DSB) play a major role in the radiation-induced killing of mammalian cells. However, the arguments supporting the relationship between DSB and radiosensitivity are generally indirect. Furthermore, care must be taken to allow for the possible impact of the techniques and of the experimental protocols on the relationship between DSB and cell death. The recent data on DSB induction, repair and misrepair in human cell lines and their correlation with intrinsic radiosensitivity are reviewed.

Dose-rate effects on the survival of irradiated hypersensitive and normal human fibroblasts.

The linear-quadratic model describes cell killing by radiation as due to two processes defined by the linear (alpha) component and by the quadratic (beta) component. As alpha and beta are interdependent, it is difficult to evaluate accurately the alpha component (which characterizes the intrinsic radiosensitivity). It has been suggested that irradiation at low dose-rate (around 1 cGy/min) allows the disappearance of the beta component and thus gives a direct measure of alpha. The present results verify this hypothesis with plateau phase cells. The survival of five human fibroblast cell lines in exponentially growing and density-inhibited, confluent cultures maintained at 37 degrees C following exposure to 60Co gamma-rays at dose-rates of 0.33-100 cGy/min followed by delayed plating (only for plateau phase cells) was monitored. Three of these cell lines are considered to be 'normal' and two are derived from hypersensitive individuals. The mean inactivation doses (D) of the five cell lines for acute doses with immediate plating were 173, 163, 136, 107 and 67 cGy. (D) increased with delayed plating recovery for 4 of the 5 cell lines and the survival of the 5 cell lines increased after low dose-rate exposure (1 cGy/min) without altering the ranking. The differences between cell lines (absolute values of (D) increased with decreasing the dose-rate. Analysis of the survival curves with the General Linear Quadratic (GLQ) model gave repair half-times for each cell line which were not correlated with the intrinsic radiosensitivities. Surprisingly, the alpha component decreased with decreasing dose-rate for all 5 cell lines (only in plateau phase). Thus low dose-rates do not allow direct measurement of the alpha component; the decrease in alpha could be interpreted as adaptive radioresistance.

Hypersensitivity to very-low single radiation doses: its relationship to the adaptive response and induced radioresistance.

There is now little doubt of the existence of radioprotective mechanisms, or stress responses, that are upregulated in response to exposure to small doses of ionizing radiation and other DNA-damaging agents. Phenomenologically, there are two ways in which these induced mechanisms operate. First, a small conditioning dose (generally below 30 cGy) may protect against a subsequent, separate, exposure to radiation that may be substantially larger than the initial dose. This has been termed the adaptive response. Second, the response to single doses may itself be dose-dependent so that small acute radiation exposures, or exposures at very low dose rates, are more effective per unit dose than larger exposures above the threshold where the induced radioprotection is triggered. This combination has been termed low-dose hypersensitivity (HRS) and induced radioresistance (IRR) as the dose increases. Both the adaptive response and HRS/IRR have been well documented in studies with yeast, bacteria, protozoa, algae, higher plant cells, insect cells, mammalian and human cells in vitro, and in studies on animal models in vivo. There is indirect evidence that the HRS/IRR phenomenon in response to single doses is a manifestation of the same underlying mechanism that determines the adaptive response in the two-dose case and that it can be triggered by high and low LET radiations as well as a variety of other stress-inducing agents such as hydrogen peroxide and chemotherapeutic agents although exact homology remains to be tested. Little is currently known about the precise nature of this underlying mechanism, but there is evidence that it operates by increasing the amount and rate of DNA repair, rather than by indirect mechanisms such as modulation of cell-cycle progression or apoptosis. Changed expression of some genes, only in response to low and not high doses, may occur within a few hours of irradiation and this would be rapid enough to explain the phenomenon of induced radioresistance although its specific molecular components have yet to be identified.

A new model describing the curves for repair of both DNA double-strand breaks and chromosome damage.

A review of reports dealing with fittings of the data for repair of DNA double-strand breaks (DSBs) and excess chromosome fragments (ECFs) shows that several models are used to fit the repair curves. Since DSBs and ECFs are correlated, it is worth developing a model describing both phenomena. The curve-fitting models used most extensively, the two repair half-times model for DSBs and the monoexponential plus residual model for ECFs, appear to be too inflexible to describe the repair curves for both DSBs and ECFs. We have therefore developed a new concept based on a variable repair half-time. According to this concept, the repair curve is continuously bending and dependent on time and probably reflects a continuous spectrum of damage repairability. The fits of the curves for DSB repair to the variable repair half-time and the variable repair half-time plus residual models were compared to those obtained with the two half-times plus residual and two half-times models. Similarly, the fits of the curves for ECF repair to the variable repair half-time and variable half-time plus residual models were compared to that obtained with the monoexponential plus residual model. The quality of fit and the dependence of adjustable parameters on the portion of the curve fitted were used as comparison criteria. We found that: (a) It is useful to postulate the existence of a residual term for unrepairable lesions, regardless of the model adopted. (b) With the two cell lines tested (a normal and a hypersensitive one), data for both DSBs and ECFs are best fitted to the variable repair half-time plus residual model, whatever the repair time range.

Potential doubling time and clinical outcome in head and neck squamous cell carcinoma treated with 70 GY in 7 weeks.

PURPOSE: To study the predictive value of pretreatment potential doubling time and labeling index, as measured by flow cytometry in patients with head and neck squamous cell carcinoma treated with conventional radiotherapy. METHODS AND MATERIALS: 70 patients with a squamous cell carcinoma of the oropharynx and 4 patients with another involved head and neck site were entered in this prospective study. The duration of the S phase (TS), the labeling index (LI), and the potential doubling time (Tpot) were obtained by flow cytometry measurements of a tumor biopsy obtained after i.v. injection of 200 mg bromodeoxyuridine to the patient. The treatment consisted of 70 Gy in 7 weeks, 2 Gy per fraction and five fractions per week. RESULTS: The mean and median LI were 7.7% (standard deviation, SD: 5.0) and 6.3%, respectively. The mean and median TS were 9.3 h (SD: 3.6) and 8.3 h, respectively. The mean and median Tpot were 5.6 days (SD: 5.4) and 4.6 days, respectively. No significant relationship was found between the Tpot or LI and the tumor stage (T), nodal status (N), histological grade, and the site of the primary within the oropharynx. The only parameter significantly associated with an increased risk of local relapse was the tumor stage (p < 0.001). The mean Tpot for the group of tumors that relapsed locally was 5.3 days (SD: 3.3), compared to 6.1 days (SD: 4.08) for those who did not relapse locally (NS). Two parameters were significantly associated with a decrease in disease-free (DFS) and overall survival, namely the tumor stage (p < 0.005, and p < 0.001, respectively, for DFS and overall survival) and nodal involvement (p = 0.02 and (p < 0.005, respectively, for DFS and overall survival). The TS, LI, DNA index, and Tpot were not significantly associated with local relapse, DFS, and survival, either in the univariate or in the multivariate analysis. CONCLUSIONS: The method used to evaluate tumor cell kinetics did not provide clinically relevant kinetic parameters for this type of cancer. The classic prognostic factors (tumor stage and nodal status) were strongly associated with clinical outcome.

Might intrinsic radioresistance of human tumour cells be induced by radiation?

Survival measurements were made on six human tumour cell lines in vitro after irradiation with single doses of X rays. Doses up to 5 Gy were used giving surviving fractions down to 20%, but the majority of the measurements were made at doses < 1 Gy. These six cell lines have very different intrinsic radiosensitivities: HT29, Be11, and RT112 are radioresistant with surviving fractions at 2 Gy (SF2) between 60 and 74%, while MeWo, SW48, and HX142 are radiosensitive (SF2 = 3-29%). For all the cell lines, response over the dose range 2-5 Gy showed a good fit to a Linear-Quadratic (LQ) model. However, HT29, Be11, and RT112 cells showed a significant increase in X-ray radiosensitivity at doses below < 1 Gy compared with the prediction extrapolated from a LQ model fitted to the data at higher doses. The LQ model also slightly underpredicted the effect of low-dose X rays in MeWo cells, but the response of SW48 and HX142 cells was well described by the LQ model at all doses, with no evidence of increased low-dose effectiveness. The most plausible explanation for this phenomenon is that it reflects an induced radioresistance so that low doses of X-rays in vitro are more effective per Gy than higher doses, because only at higher doses is there sufficient damage to trigger repair systems or other radioprotective mechanisms. It follows that variation in the amount of inducible radioresistance might explain, in part, differences in intrinsic radiosensitivity above > 1 Gy between cell lines: cells would be intrinsically radiosensitive because they have a diminished inducible response.

High split-dose recovery in hypersensitive human fibroblasts: a case of induced radioresistance?

We studied the extent of split-dose recovery in seven non-transformed human fibroblast cell lines of different intrinsic radiosensitivity (HF19, 1BR3, 149BR, 84BR, GM739, 180BR and AT2EM). Experiments were performed on both growing and plateau-phase cells. The seven cell lines displayed a wide range of intrinsic radiosensitivity. The D of plateau phase cells ranged from 0.56 (AT2EM) to 3.02 Gy (HF19). The recovery ratios (RR) of the three non-ataxic hypersensitive cell lines (84BR, GM739, and 180BR) were significantly higher than those predicted from the single-dose survival curves of both growing and plateau-phase cells. In addition, in these three hypersensitive cell lines the challenge dose survival curve generated after different priming doses showed a reduction in the intrinsic radiosensitivity; the high RRs observed were due both to beta and a reduction in alpha. This suggests that a protective mechanism may be triggered by the first irradiation leading to induced radioresistance. For growing cells, the relationship between ln RR and 2D2 was well fitted by linear regression. With plateau phase cells, RR appeared to be dose dependent in a more complex fashion. Thus, no single value of beta RR was representative of the split-dose recovery. With the ataxic cell line AT2EM, the split-dose studies detected a limited capacity to recover in spite of the beta value of the single dose survival curve being nil.

Dose-rate effect on radiation-induced DNA double-strand breaks in the human fibroblast HF19 cell line.

We measured DNA double-strand breaks (dsbs) immediately after exposure of a non-transformed human fibroblast cell line (HF19) to gamma-rays (0-40 Gy) at four dose-rates (10, 1, 0.1, and 0.01 Gy/min) at 37 degree C using clamped homogeneous electric field (CHEF) gel electrophoresis. The shape of the dose-response curves, which could be approximated by a straight line over the range 0-20 Gy for irradiation at 4 degree C, became curvilinear when irradiation was carried out at 37 degree C at 10, 1, 0.1, and 0.01 Gy/min and reached a plateau at 10 Gy after irradiation at 0.01 Gy/min. We present a mathematical analysis that predicts the results of irradiation at 37 degree C from dsb induction and repair data obtained at 4 degree C, followed by incubation for repair at 37 degree C. The model assumes that the rate of dsb rejoining changes continuously with repair time and that it is independent of dose and dose-rate in the range 10-40 Gy. The model also assumes a linear induction of dsb with dose at 4 degree C and dsb induction is independent of dose-rate and of temperature during irradiation. Independent measurements of dsb induction at 4 degree C and of repair rate accurately predict the dsb levels after irradiation at 37 degree C, during which both phenomena occur simultaneously.

Induction and rejoining of DNA double-strand breaks and interphase chromosome breaks after exposure to X rays in one normal and two hypersensitive human fibroblast cell lines.

The aim of this work was to measure simultaneously and in a quantitative manner double-strand breaks (DSBs), interphase chromosome breaks and cell lethality either immediately after irradiation, or at various times thereafter (up to 24 h), in cells of three nontransformed human fibroblast cell lines of widely different intrinsic radiosensitivity. We wished to assess initial damage, repair kinetics and residual damage at the DNA and the chromosome level, and to correlate these parameters with cell killing. We employed HF19 cells, a normal fibroblast cell line, AT2 cells, a radiosensitive cell line from a patient suffering from ataxia telangiectasia (AT), and 180BR cells, a radiosensitive cell line from a patient with no clinical symptoms of AT. AT2 and 180BR cells, in addition to being radiosensitive, also display a reduced ability to repair potentially lethal damage compared to HF19 cells. The yield of DSBs, as measured by pulsed-field gel electrophoresis, is similar in all three cell lines (slopes correspond to 1.6-1.7% Gy-1 of DNA-associated radioactivity released from the gel well into the lane). In contrast, residual DSBs measured 24 h after irradiation are almost zero for HF19 cells (0.1% confidence interval = 0-1.4%), but are 12.5% (+/- 2.3%) and 43.8% (+/- 1.2%) of those measured immediately after irradiation in AT2 and 180BR cells, respectively. Residual interphase chromosome breaks are 11.6% (+/- 1.6%), 29.7% (+/- 5.7%) and 41.4% (+/- 2.2%) of those measured immediately after irradiation in HF19, AT2 and 180BR cells, respectively. Neither the initial yield of DSBs nor that of excess interphase chromosome breaks can explain the differences in radiosensitivity between the three cell lines; however, there is a correlation between residual DSBs, rate of DSB rejoining at 24 h, residual interphase chromosome breaks on the one hand and cell survival on the other hand.

The beta component of human cell survival curves and its relationship with split-dose recovery.

In principle, alpha and beta can be obtained from single-dose survival curves using standard linear-quadratic fitting; however, alpha and beta being interdependent, it is difficult to evaluate them together with good precision. On the assumption that full recovery from a split-dose treatment gives a result that is the product of the single-dose surviving fraction, it has been suggested that the measurement of split-dose recovery should provide a method to measure beta alone using the formula: beta RR = lnRR/2d2. Most of the studies published to date have been carried out on cancer cell lines or transformed normal cells. We have systematically tested the above proposal on two normal human fibroblast cell lines (HF19 and 1BR3) in two different situations: growing cells, and plateau-phase cells. Two different protocols were used to assess both the potential influence of a priming dose on the surviving cells and the extent of the split-dose recovery. The survival curves generated after different priming doses did not show any significant change in comparison with those achieved without previous irradiation. In addition, the split-dose survival was not different from the square of the corresponding single-dose survival (model free). In these conditions, beta RR's obtained by a linear regression of the recovery ratio data were very similar to the beta's obtained by single doses. However, a curvilinear regression (with a very small negative term at high doses) appears to be more appropriate for cells in plateau phase. This has the result that, as the dose increases, the cell survival curves tend to become less bending than would be expected from the linear-quadratic model; however, the linear-quadratic fitting is still a reasonable characterization of the radiation response since the in vitro colony formation method does not allow measurement of survival < 10(-4).