Abstract
Background/Aim: The use of ion radiation therapy is growing due to the continuously increasing positive clinical experience obtained. Therefore, there is a high interest in radio-biological experiments comparing the relative efficiency in cell killing of ions and photons as photons are currently the main radiation modality used for cancer treatment. This comparison is particularly important since the treatment planning systems (TPSs) used at the main ion therapy Centers make use of parameters describing the cellular response to photons, respectively ions, determined in vitro. It was, therefore, the aim of this article to compare the effects of high linear energy transfer (LET) ion radiation with low LET photons and determine whether the cellular response to low LET could predict the response to high LET irradiation. Materials and Methods: Clonogenic cell survival data of five tumor cell lines irradiated with different ion beams of similar, clinically-relevant, LET were studied in relation to response to low LET photons. Two mathematical models were used to fit the data, the repairable-conditionally repairable damage (RCR) model and the linear quadratic (LQ) model. Results: The results indicate that the relative biological efficiency of the high LET radiation assessed with the RCR model could be predicted based only on the response to the low LET irradiation. Conclusion: The particular features of the RCR model indicate that tumor cells showing a large capacity for repairing the damage will have the larger benefit from radiation therapy with ion beams.
The field of ion radiation therapy is under continuous development due to the growing positive clinical experience that keeps accumulating at the two main Centers using carbon ions, Heavy Ion Medical Accelerator (HIMAC) at Chiba, Japan and Heidelberger Ionenstrahl-Therapiezentrum (HIT) in Heidelberg, Germany. Besides these two main Centers only few more are already operational or in various stages of development (1). The clinical success of radiotherapy using ions relies on the physical properties of the ions in interaction with matter, which result in a sharp increase in the absorbed dose in the Bragg peak followed by a steep decrease beyond the dose maximum (2). However, in order to further increase the rate of cure using ion radiotherapy, the radiobiological advantages of ions should be further exploited for complementing the good conformity of the dose distribution to the target achievable due to their physical properties. Therefore, there is a continuous interest in radiobiological experiments comparing the relative efficiency in cell killing of ions and photons as the photons are currently the main radiation modality used for cancer treatment (3). Furthermore, the treatment planning systems (TPSs) used at the main ion therapy facilities in Japan and Germany, although using different approaches require as input data parameters describing the cellular response to photons, respectively ions, determined in vitro (4-6). Considerable effort has, therefore, been put on determining the relative biological efficiency (RBE) of ions in comparison to photons (7-12). In parallel, radiobiological models, with various degrees of complexity, have been developed and compared with respect to the ability of accurately fitting experimental cell survival data with the linear quadratic (LQ) model for cell survival initially developed for photons (4-6, 13-16). Until very recently, the LQ model was used at HIMAC in order to determine the corresponding biological dose leading to 10% survival fraction for experimental in vitro cell survival data of human salivary gland (HSG) tumor cells, which, in conjunction with previous knowledge and experience on the RBE of neutrons, was eventually used to calculate the biological dose for carbon ions. Currently, the microdosimetric kinetic (MK) model (4, 5) is used at HIMAC, which is, with some extent, similar, to the local effect model (LEM) (6) used at the German facilities, both heavily relying on the LQ model and the statistics and fluctuations of the microscopic dose deposition, or the local dose derived from an amorphous track structure model, respectively.
However, despite this considerable effort, there is still need for more experimental data on the efficiency in cell killing of ion radiation with high linear energy transfer (LET) in comparison to low LET photon radiation as well as for robust modelling able not only to accurately fit the data but also to allow for the explicit assessment of the relative efficiency in cell killing of high LET compared to low LET radiation. Therefore, the present article adds to the current experience by presenting the results of clonogenic cell survival experiments after low and high LET irradiation for a large variety of cell lines: small cell lung cancer (SCLC) cell line (U-1690), head and neck squamous cell carcinoma (HNSCC) cell line (FaDu), melanoma cell line (AA) and two prostate cell lines (PC-3 and DU-145).
Furthermore, the present study also presents the findings of a modelling study using the Repairable-Conditionally Repairable damage model (RCR), a cell survival model previously developed by the same group (15, 17, 18) aiming to determine whether the sensitivity to low LET irradiation of the cell lines under investigation could be used as an indicator of the relative biological efficiency of ion therapy. The ability of the RCR model to provide this type of relationship was compared to the potential of the LQ model, which is still the reference model for cell survival, the one recently used in the TPS at HIMAC and the core of the LEM model used at HIT, to predict the same information.
Materials and Methods
Cell survival models. The Repairable-Conditionally Repairable describes the probability for clonogenic cell survival S after a dose D is given by:
[1]
where a, b and c are the parameters of the model. The first exponential term of the model describes the fraction of cells that have not been damaged by radiation and hence the parameter a gives the initial mean number of damage events per unit dose. The second term of the model describes the fraction of cells that have been damaged and subsequently correctly repaired. The parameter b describes the maximum amount of damage that can be repaired per unit dose and the c parameter is related to the probability that a potentially repairable event is actually repaired.
Given the underlining meaning and interpretation of the parameters of the model, the following constraints are used for fitting clonogenic cell survival data since the survival fractions cannot be greater than 1 and the total number of damage events has to be larger than or equal to the number of repairable damage events:
[2]
[3]
Since the first term in equation 1 gives the fraction of cells not being damaged and the second term describes the fraction of cells that have been damaged and subsequently repaired, the fraction of correctly repaired cells normalized to the fraction of cells that have been damaged, r, is described by the following equation:
[4]
The mean inactivation dose, 
, for a cell population which responds to radiation according to a given function describing the clonogenic cell survival, is mathematically described as the area under the survival curve which could be calculated by integrating the cell survival function in response to dose. If cell survival is described by the RCR model, it has been previously shown that 
can be calculated as (15):
[5]
The survival of the cells could also be calculated using the LQ model (13):
[6]
where α and β are the parameters of the model. Several explanations have been proposed for the α and β parameters, among the most common one being that β/α gives a measure of the repair capacity of the cells in the given population (19).
Cell lines and cell-culturing conditions. A panel of five different human tumor cell lines was used in this study. The human melanoma cell line AA, human small cell lung cancer (SCLC) cell line U-1690 (20), human head and neck SCC cell line FaDu isolated from a squamous cell carcinoma of the hypopharynx and two prostate cancer cell lines – PC-3 (isolated from bone metastasis) and DU-145 (isolated from brain metastasis) were used. The cells were cultured as monolayer in Eagle's MEM medium (Flow Laboratories, Rickmansworth, UK) containing 10% foetal bovine serum (FBS), 100 I.E./ml penicillin, 100 μg/ml streptomycin and 2.5 μg/ml fungizone (Gibco, Invitrogen, Inchinnan, UK) under standard conditions (37°C, 95% air and 5% CO2 atmosphere). Cells were passaged every week. The doubling time was approximately 24 h for all the cell lines. Cells in exponential growth phase were used in all experiments. Thus, cells were seeded at a density of 2-5·105 in 0.2 ml of medium, in the middle of 3-cm Petri dishes one day before the irradiation exposures.
The SF2 values from the different experiments.
Irradiation procedures. The cells were irradiated with low and high LET. For the low LET exposure 60Co or 137Cs photons were used. Cells were irradiated at Karolinska Institutet and Karolinska University Hospital, Stockholm, Sweden, at room temperature with doses of 0-12 Gy either from a 60Co therapy unit at a photon dose rate of <0.5 Gy/min and LET of 0.2 keV/μm or with a 137Cs source at a photon dose rate of 0.5 Gy/min.
For the high LET exposure doses in the range 0-4 Gy were used for various ions. Irradiations were performed at the Biomedical Unit at the Theodor Svedberg Laboratory (TSL), Uppsala, Sweden, with 33.7 MeV/u nitrogen ions (LET 85 keV/μm) and 24.8 MeV/u carbon ions (LET 80 keV/μm) (21, 22). Irradiations with 500 MeV/u argon ions (LET 90 keV/μm) and 290 MeV/u carbon ions (LET 80 keV/nm) were performed at the Heavy Ion Medical Accelerator (HIMAC), at the National Institute of Radiological Sciences (NIRS), Chiba, Japan.
Clonogenic survival assay. The cells were seeded before irradiation, as mentioned before, and reseeded after exposure at different concentrations depending on the used dose of radiation. The cells were plated in three Petri dishes (6 cm) per dose. The cell density in the dishes was also adjusted to compensate for the plating efficiency of the each cell line. Cells were cultured for 12-14 days, with a change of medium after 6-7 days. The resulting colonies were stained with 25% Giemsa stain. Colonies containing at least 50 cells were counted as survivors. The experiments were repeated one to four times.
For each absorbed dose, the surviving fraction was calculated as the ratio of the mean plating efficiency (PE) in dishes with irradiated cells over the PE in dishes with un-irradiated control cells. PE was calculated as the mean number of cell colonies over the number of plated cells. The mean plating efficiency (PE) ±SEM for non-irradiated cells was 89±7%, 69±11%, 62±14%, 37±7% and 58%±10% for U-1690, AA, PC-3, DU-145 and FaDu, respectively.
Relative biological effectiveness calculations. The relative biological effectiveness (RBE) values for the different tumor cells and irradiation exposures were calculated in two ways: either by dividing the photon doses to ion doses giving 10% survival (D10) or by dividing the corresponding 
values using both RCR and LQ survival models. As the experiments from photons and ions were not paired, the standard deviation of the RBE was calculated with the error propagation formula.
Results
The summary of the clonogenic cell survival experiments described as the surviving fraction of cells at 2 Gy, SF2, is given in Table I. The full set of clonogenic survival curves of the five cell lines exposed to photons and different accelerated ions is shown in Figure 1. All five cell lines were exposed to photons (Figure 1, left panel). All five cell lines were also exposed to 12C (80 keV/μm) and 36Ar (90 keV/μm) ions while two cells lines, AA and U-1690, were also exposed to 14N (85 keV/μm) ions (Figure 1, right panel). The curves in Figure 1 represent the RCR fit of the clonogenic survival data.
Table I and Figure 1 both show the expected differences in radiosensitivity to photon irradiation and to the various ions.
The corresponding RBE values for the different tumor cells calculated either as the ratio of doses giving 10% survival (D10) or by dividing the 
values using both RCR and LQ survival models are given in Table II. By comparing the RBE calculated in one way using the two cell line survival model for one single cell line, the differences between the resulting values are not significantly different, showing that the choice of the model is not critically affecting the result. If the comparison is performed for one cell line between the two ways of calculating the relative biological effectiveness, i.e. as ratio of D10 values or of 
values, the resulting values are not significantly different either proving that the choice of method for the RBE calculation is not critical.
In order to assess the differences in response of the various cell lines, the relative biological effectiveness calculated as the ratio of 
values was plotted against the parameters of the cell survival models used for fitting the experimental data (Figure 2). In the left panel of Figure 2, the 
ratio is shown as function of the β/α parameters of the LQ model, where the α and β parameters correspond to the
fitting of the cell survival data after photon irradiation. In the right panel of Figure 2 the 
ratio is shown as function of the b/a parameters of the RCR model fitting the cell survival after photon irradiation.
Difference in radiosensitivity between high and low LET irradiation. The clonogenic cell survival after exposure to photons (left panel) and ions beams (right panel) for all cell lines are shown. T stands for TSL and H for HIMAC. The data points show the mean survival from one to four repeat experiments with triplicate dishes ±SEM indicated when larger than the symbols.
For both cases shown in Figure 2 a linear relationship between the relative biological effectiveness and the cellular repair capacity was sought. For the right panel corresponding to the RCR fitting, a linear relationship with a positive slope was indeed obtained. The uncertainties are rather large but the slope of the indicated regression line has a significant value above one (p<0.001). The same plot (Figure 2, left panel) for the LQ model shows almost no correlation between the β/α values and the 
ratios.
Discussion
The present study investigated the relationship of the biological effectiveness of high LET radiation on the capacity to repair photon-induced damage. When analyzing the cell survival curves, the five tumor cell lines with all their underlying genetic differences showed a large variation in their response to photons but a smaller variation when exposed to different ion beams. The most likely explanation to this reduced variability in sensitivity is that high LET irradiation induces more severe and difficult damage to be repaired and the intrinsic sensitivity becomes less critical.
The biological effect, measured as RBE at a 10% survival level or 
ratios was almost independent of the methods for calculations i.e. the two survival models used in this study. However, the RBE will differ at different survival levels and be more pronounced in the region where low dose hypersensitivity is seen (Figure 1). The results showed that there is no significant difference in the calculated RBE values from the ratio of the D10 doses and the 
ratios, irrespective of whether the LQ and the RCR models have been used, even though the survival levels corresponding to the 
values were much larger than 10%. They vary between 36% and 44% for the photons and between 28% and 40% for the different ions and cell types with the lower values for the most radiosensitive cell line FaDu. To avoid the low-dose hypersensitivity the analysis was limited to the survival values from 1 Gy and above and which is also a clinically relevant dose range.
RBE (
ratios) plotted against the cellular repair capacity. The mean of 
ratios ±SEM, indicated when larger than the symbols, are plotted against the parameter ratio from the photon exposures (β/α, left panel) and (b/a, right panel). Regression lines and the corresponding parameters are shown for both plots.
The RBE values calculated from the D10 doses and the 
ratios (using the RCR and the LQ models). RBERCR/LQ denotes the relative biological effectiveness calculated by dividing the photon and the ion dose corresponding to 10% survival when the fitting of the experimental data was performed using the RCR or the LQ model, respectively. ratioRCR/LQ denotes the relative biological effectiveness calculated by dividing the total inactivation dose, 
, when the fitting of the experimental data was performed using the RCR or the LQ model.
In the RCR model the parameter b is interpreted to be connected with the repair capacity of the cells and b/a will then be an estimate of the fraction of damaged cells that are repaired. Thus, plotting the biological effect in terms of 
ratios against the b/a values for the photon response, an increased biological effect was expected and indeed observed for larger b/a values. A b/a value determined for low LET beams may accordingly be used to predict the response to ion beams, as indicated by the regression line in the right panel in Figure 2.
The similar analysis performed using the LQ model showed that the β/α parameters did not retain the capacity of predicting the expected increased radiobiological effectiveness of ion therapy on resistant cell lines. This suggests that care should be taken when parameters derived by assuming a linear-quadratic dependence for cell killing are used in the TPSs together with RBE values for determining the biological dose corresponding to a given physical dose for ions.
Nevertheless, the correlation between RBE and the ratio b/a of the RCR model indicates that tumor cells showing a large capacity for repairing the damage will benefit the most from radiation therapy with ion beams.
Acknowledgements
We gratefully thank our colleagues at the division of Biomedical Radiation Sciences, Rudbeck Laboratory, Uppsala University (Uppsala, Sweden) for skilful help with the carbon and nitrogen ions irradiation procedures and at the Research Centre for Charged Particle Therapy and HIMAC, NIRS (Chiba, Japan) for help with argon ions irradiation procedures. We thank Annelie Meijer for help with the photon experiments.
Footnotes
-
↵* These Authors contributed equally to the manuscript.
-
Funding
This work has been supported by ‘the Swedish Cancer Society’; ‘the Stockholm Cancer Society’; ‘the Funds of the Karolinska Institutet’ and ‘the European Community's Seventh Framework Programme’ [FP7/2007-2013] under ‘[grant agreement no: 215840-2]’. This study was supported in part by IOL (International Open Laboratory) of NIRS.
-
Conflicts of Interest
None to declare.
- Received February 4, 2014.
- Revision received April 26, 2014.
- Accepted April 28, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.