Repair of DNA double-strand breaks and cell killing by charged particles.

It has been suggested that it is not simple double-strand breaks (dsb) but the non-reparable breaks which correlate well with the high biological effectiveness of high LET radiations for cell killing (Kelland et al., 1988; Radford, 1986). We have compared the effects of charged particles on cell death in 3 pairs of cell lines which are normal or defective in the repair of DNA dsbs. For the cell lines SL3-147, M10, and SX10 which are deficient in DNA dsb repair, RBE values were close to unity for cell killing induced by charged particles with linear energy transfer (LET) up to 200 keV/micrometer and were even smaller than unity for the LET region greater than 300 keV/micrometer. The inactivation cross section (ICS) increased with LET for all 3 pairs. The ICS of dsb repair deficient mutants was always larger than that of their parents for all the LET ranges, but with increasing LET the difference in ICS between the mutant and its parent became smaller. Since a small difference in ICS remained at LET of about 300 keV/micrometer, dsb repair may still take place at this high LET, even if its role is apparently small. These results suggest that the DNA repair system does not play a major role in protection against the attack of high LET radiations and that a main muse of cell death is non-reparable dsb which are produced at a higher yield compared with low LET radiations. No correlation was observed between DNA content or nuclear area and ICS.     

DNA double strand break production and rejoining in V79 cells irradiated with light ions.

Low energy protons and other densely ionizing light ions are known to have RBE>1 for cellular end points relevant for stochastic and deterministic effects. The occurrence of a close relationship between them and induction of DNA dsb is still a matter of debate. We studied the production of DNA dsb in V79 cells irradiated with low energy protons having LET values ranging from 11 to 31 keV/micrometer, i.e. in the energy range characteristic of the Bragg peak, using the sedimentation technique. We found that the initial yield of dsb is quite insensitive to proton LET and not significantly higher than that observed with X-rays, in agreement with recent data on V79 cells irradiated with alpha particles of various LET up to 120 keV/micrometer. By contrast, RBE for cell inactivation and for mutation induction rises with the proton LET. In experiments aimed at evaluating the rejoining of dsb after proton irradiation we found that the amount of dsb left unrepaired after 120 min incubation is higher for protons than for sparsely ionizing radiation. These results indicate that dsb are not homogeneous with respect to repair and give support to the hypothesis that increasing LET leads to an increase in the complexity of DNA lesions with a consequent decrease in their repairability.     

Correlation between cell death and induction of non-rejoining PCC breaks by carbon-ion beams.

We have shown a correlation between cell death and induction of non-rejoining chromatin breaks in two normal human cells and three human tumor cell lines irradiated by carbon-ion beams and X rays. Non-rejoining chromatin breaks were measured by counting the number of remaining chromatin fragments detected by the premature chromosome condensation (PCC) technique. Carbon-ion beams were accelerated by the Heavy Ion Medical Accelerator in Chiba (HIMAC). The cells were irradiated by two different mono-LET beams (LET = 13 keV/micrometer and 77 keV/micrometer ) and 200 kV X rays. The RBE values of cell death for carbon-ion beams relative to X rays were 1.1 to 1.4 for 13 keV/micrometer beams and 2.5 to 2.9 for 77 keV/micrometer beams. The induction rate of non-rejoining PCC breaks per cell per Gy was found to be highest for the 77 keV/micrometer beams for all of the cell lines.The results found in this study show that there is a good correlation between cell death and induction of non-rejoining PCC breaks for these human cell lines.     

LET dependence of cell death, mutation induction and chromatin damage in human cells irradiated with accelerated carbon ions.

We investigated the LET dependence of cell death, mutation induction and chromatin break induction in human embryo (HE) cells irradiated by accelerated carbon-ion beams. The results showed that cell death, mutation induction and induction of non-rejoining chromatin breaks detected by the premature chromosome condensation (PCC) technique had the same LET dependence. Carbon ions of 110 to 124keV/micrometer were the most effective at all endpoints. However, the number of initially induced chromatin breaks was independent of LET. About 10 to 15 chromatin breaks per Gy per cell were induced in the LET range of 22 to 230 keV/micrometer. The deletion pattern of exons in the HPRT locus, analyzed by the polymerase chain reaction (PCR), was LET-specific. Almost all of the mutants induced by 124 keV/micrometer beams showed deletion of the entire gene, while all mutants induced by 230keV/micrometer carbon-ion beams showed no deletion. These results suggest that the difference in the density distribution of carbon-ion track and secondary electron with various LET is responsible for the LET dependency of biological effects.     

Stochastics of HZE-induced microlesions.

Fundamental biological experiments with bacteria, yeast, and mammalian cells irradiated with ions heavier than helium indicate that maximal probability of single-hit inactivation does not occur when the ion has LET below about 100-200 keV/micrometer. Theoretical treatments of cell inactivation data and the radiation chemistry in particle tracks are consistent with this finding. If a "microlesion" is defined as a linear array, within a tissue, of cells inactivated with maximum probability, surrounded by non-lethally damaged cells, then, by this definition, there must be an LET below which "microlesion" damage cannot be expected. In a retrospective survey of experimental literature in which single-particle effects in tissues were sought, it was found that little or no evidence has been reported supporting single-particle effects in tissues when LET was below 200 keV/micrometer, while some experimenters who irradiated tissues with particles having LET greater than 200 keV/micrometer reported effects that could be attributed to single-particle tracks.     

DNA fragmentation in mammalian cells exposed to various light ions.

Elucidation of how effects of densely ionizing radiation at cellular level are linked to DNA damage is fundamental for a better understanding of the mechanisms leading to genomic damage (especially chromosome aberrations) and developing biophysical models to predict space radiation effects. We have investigated the DNA fragmentation patterns induced in Chinese hamster V79 cells by 31 keV/micrometer protons, 123 keV/micrometer helium-4 ions and gamma rays in the size range 0.023-5.7 Mbp, using calibrated Pulsed Field Gel Electrophoresis (PFGE). The frequency distributions of fragments induced by the charged particles were shifted towards smaller sizes with respect to that induced by comparable doses of gamma rays. The DSB yields, evaluated from the fragments induced in the size range studied, were higher for protons and helium ions than for gamma rays by a factor of about 1.9 and 1.2, respectively. However, these ratios do not adequately reflect the RBE observed on the same cells for inactivation and mutation induced by these beams. This is a further indication for the lack of correlation between the effects exerted at cellular level and the initial yield of DSB. The dependence on radiation quality of the fragmentation pattern suggests that it may have a role in damage repairability. We have analyzed these patterns with a "random breakage" model generalized in order to consider the initial non-random distribution of the DNA molecules. Our results suggest that a random breakage mechanism can describe with a reasonable approximation the DNA fragmentation induced by gamma rays, while the approximation is not so good for light ions, likely due to the interplay between ion tracks and chromatin organization at the loop level.     

Morphometric studies of heavy ion damage in the brains of rodents.

As an approach to determining the relative biological effectiveness (RBE) of each of five different heavy ions for the mammalian brain, histological preparations of brains from mice exposed to various HZE particles at different doses and primary LETinfinity values were examined by means of semi-automated image analysis for volume changes in specific regions of the olfactory bulb. The mice were irradiated at 100 days of age and euthanatized about 500 days (16 months) later. Exposures were: 60Co gamma photons (LETinfinity = 1-2 keV/micrometer), 4He (LETinfinity = 6 keV/micrometer), 12C (LETinfinity = 80 keV/micrometer), 20Ne (LETinfinity = 150 keV/micrometer), 56Fe (LETinfinity = 180 keV/micrometer), and 40Ar (LETinfinity = 650 keV/micrometer). Animals receiving particle radiation were exposed in an extended Bragg peak region except for iron where the plateau region was used. The zones measured in the olfactory bulb were 1) the external plexiform layer (zone) and 2) an internal region consisting of the granule cells, internal plexiform layer, and layer of mitral cells. These studies indicated that volume changes did indeed occur, not only in absolute terms but also when expressed as the ratio of the structures to each other and to the bulb as a whole. Although this study is exploratory in character, the data obtained may nevertheless contribute to a determination of risk factors due to late effects from HZE articles.     

Mutation induction in human lymphoid cells by energetic heavy ions.

One of the concerns for extended space flight outside the magnetosphere is exposure to galactic cosmic radiation. In the series of studies presented herein, the mutagenic effectiveness of high energy heavy ions is examined using human B-lymphoblastoid cells across an LET range from 32keV/micrometer to 190 keV/micrometer. Mutations were scored for an autosomal locus, thymidine kinase (tk), and for an X-linked locus, hypoxanthine phosphoribosyltransferase (hprt). For each of the radiations studied, the autosomal locus is more sensitive to mutation induction than is the X-linked locus. When mutational yields are expressed in terms of particle fluence, the two loci respond quite differently across the range of LET. The action cross section for mutation induction peaks at 61 keV/micrometer for the tk locus and then declines for particles of higher LET, including Fe ions. For the hprt locus, the action cross section for mutation is maximal at 95 keV/micrometer but is relatively constant across the range from 61 keV/micrometer to 190 keV/micrometer. The yields of hprt-deficient mutants obtained after HZE exposure to TK6 lymphoblasts may be compared directly with published data on the induction of hprt-deficient mutants in human neonatal fibroblasts exposed to similar ions. The action cross section for induction of hprt-deficient mutants by energetic Fe ions is more than 10-fold lower for lymphoblastoid cells than for fibroblasts.     

DNA-lesion and cell death by alpha-particles and nitrogen ions.

When the natural logarithm of the surviving fraction is plotted against the dose of radiation, curves with shoulders at relatively high survival levels are obtained after gamma-rays. The curves were practically linear in case of HMV-I and HA-1 cells irradiated by charged particle beams. These cells were derived from human malignant melanoma and Chinese hamster cells, respectively. The amount of DNA single strand breaks (ssb) by gamma-rays or nitrogen-ions (LET=530KeV/micrometers) in HMV-I cells increases linearly with increment in dose, when the ssb is detected using the alkaline elution technique. There is no close relationship between the dose-response curve of the ssb and the dose-survival curves after gamma-rays or N-ions. The amount of DNA double strand breaks (dsb) by gamma-rays increases quadratically with increment of dose, in both HMV-I cells and HA-1 cells, when the dsb is detected using the neutral elution technique. The survival fraction for HA-1 cells is slightly higher than that for HMV-I cells, at the same dose, and the amount of dsb for HA-1 cells is considerably greater than that for HMV-I cells. These results suggest that the radiosensitivities to gamma-rays in different cell lines do not correspond to the number of DNA strand breaks. The amount of both non-repairable ssb and dsb also increases quadratically with increment of dose for gamma-rays and almost linearly with increment of dose for N-ions and alpha-particles (LET=36keV/micrometers for HA-1 cells and LET=77keV/micrometers for HMV-I cells). The dose-response curves for non-repairable dsb in case of these radiations seemed to mirror image the dose-survival curves for these radiations, in both cell lines. The number of non-repairable DNA strand breaks in the two cell lines, at the same level of survival was much the same. These results show the close relationship between the induction of non-repairable DNA strand breaks and cell killing.     

G2-chromosome aberrations induced by high-LET radiations.

We report measurement of initial G2-chromatid breaks in normal human fibroblasts exposed to various types of high-LET particles. Exponentially growing AG 1522 cells were exposed to gamma rays or heavy ions. Chromosomes were prematurely condensed by calyculin A. Chromatid-type breaks and isochromatid-type breaks were scored separately. The dose response curves for the induction of total chromatid breaks (chromatid-type + isochromatid-type) and chromatid-type breaks were linear for each type of radiation. However, dose response curves for the induction of isochromatid-type breaks were linear for high-LET radiations and linear-quadratic for gamma rays. Relative biological effectiveness (RBE), calculated from total breaks, showed a LET dependent tendency with a peak at 55 keV/micrometer silicon (2.7) or 80 keV/micrometer carbon (2.7) and then decreased with LET (1.5 at 440 keV/micrometer). RBE for chromatid-type break peaked at 55 keV/micrometer (2.4) then decreased rapidly with LET. The RBE of 440 keV/micrometer iron particles was 0.7. The RBE calculated from induction of isochromatid-type breaks was much higher for high-LET radiations. It is concluded that the increased production of isochromatid-type breaks, induced by the densely ionizing track structure, is a signature of high-LET radiation exposure.     

DNA damage and repair in oncogenic transformation by heavy ion radiation.

Energetic heavy ions are present in galactic cosmic rays and solar particle events. One of the most important late effects in risk assessment is carcinogenesis. We have studied the carcinogenic effects of heavy ions at the cellular and molecular levels and have obtained quantitative data on dose-response curves and on the repair of oncogenic lesions for heavy particles with various charges and energies. Studies with repair inhibitors and restriction endonucleases indicated that for oncogenic transformation DNA is the primary target. Results from heavy ion experiments showed that the cross section increased with LET and reached a maximum value of about 0.02 micrometer2 at about 500 keV/micrometer. This limited size of cross section suggests that only a fraction of cellular genomic DNA is important in radiogenic transformation. Free radical scavengers, such as DMSO, do not give any effect on induction of oncogenic transformation by 600 MeV/u iron particles, suggesting most oncogenic damage induced by high-LET heavy ions is through direct action. Repair studies with stationary phase cells showed that the amount of reparable oncogenic lesions decreased with an increase of LET and that heavy ions with LET greater than 200 keV/micrometer produced only irreparable oncogenic damage. An enhancement effect for oncogenic transformation was observed in cells irradiated by low-dose-rate argon ions (400 MeV/u; 120 keV/micrometer). Chromosomal aberrations, such as translocation and deletion, but not sister chromatid exchange, are essential for heavy-ion-induced oncogenic transformation. The basic mechanism(s) of misrepair of DNA damage, which form oncogenic lesions, is unknown.     

Repair of DNA double-strand breaks and its effect on RBE.

DNA double-strand breaks (DSB) are induced linearly with absorbed dose both for sparsely and densely ionizing radiations. By enzymatic repair the linear relationship between the number of DSB and absorbed dose is converted into a non linear one. Furthermore, the RBE-values of high LET radiations for residual DSB increase with increasing amount of DSB repair especially in the low dose range. Unrepaired and/or misrepaired DSB are supposed to be responsible for chromosomal aberrations, cell killing, oncogenic cell transformation and gene mutation. At low doses, for these endpoints much higher RBE-values than those for initial DSB are observed. However, with increasing doses the RBE-values for these endpoints approach those for initial DSB. These observations are likely to be interpreted using the following two parameters of the energy deposition structure: 1. The distribution of clusters with respect to their size at the nm-scale and to the number of ionizations per cluster (cluster distribution). 2. The distribution of distances between clusters of definite size and with definite number of ionizations (distance distribution of clusters). For the induction of DSB solely the ionization density in clusters of nm-dimensions (i.e. the cluster distribution) is important. For unrepaired or misrepaired DSB (responsible for chromosome aberrations, cell killing, oncogenic cell transformation and gene mutation) both the cluster distribution and the distance distribution of clusters are relevant. At low doses the distance distribution of clusters along a single particle track determines the RBE-value. However, with increasing dose the distribution of clusters produced by all particles traversing the cell nucleus becomes increasingly determinant. Here, solely the cluster distribution is important as it is the case for the induction of DSB.     

DNA fragmentation by charged particle tracks.

High-LET (linear energy transfer) charged particles induce DNA double-strand breaks (DSB) in a non-random fashion in mammalian cells. The clustering of DSB, probably determined by track structure as well as chromatin conformation, results in an excess of small- and intermediate-sized DNA fragments. DNA fragmentation in normal human fibroblasts (GM5758) was analyzed by pulsed-field gel electrophoresis after irradiation with photons (60Co) or 125 keV/micrometers nitrogen ions. Compared to conventional DSB analysis, i.e. assays only measuring the fraction of DNA smaller than a single threshold, the relative biological effectiveness (RBE) for DSB induction increased with 100%. Further, the size distribution of DNA fragments showed a significant dependence on radiation quality, with an excess of fragments up to 1 Mbp. Irradiation of naked genomic DNA without histone proteins increased the DSB yields 25 and 13 times for photons and nitrogen ions, respectively. The results suggest possible roles of both track structure and chromatin organization in the distribution of DNA double-strand breaks along the chromosome.     

Induction and rejoining of DNA double-strand breaks in CHO cells after heavy ion irradiation.

DNA double-strand breaks (DSBs) are the crucial events ultimately leading to cell inactivation. Aimed at understanding the biological action of the charged particle component of cosmic radiation, the induction of DSBs and their repairability was evaluated in Chinese hamster ovary (CHO-K1) cells after exposure to accelerated particles. Irradiations were performed with various ion species including O, Ni and Ca, covering a LET range from 20 to 2000 keV/micrometer. DSBs were determined for plateau-phase cells using the electrophoretic elution of radiation-induced DNA fragments in a static electric field combined with fluorescence scanning of ethidium bromide stained gels. Assuming a DSB yield of 22 DSB per Gy per cell, as derived from X-irradiation, cross-sections for DSB production were calculated from the corresponding fluence-effect curves at a fraction of 0.7 of DNA retained. The same ordinate was used as a reference for the calculation of relative biological efficiency (RBE) for DSB induction. At low LETs (< or = 20 keV/micrometer) RBE values slightly above unity were obtained, but a decrease of RBE was observed with increasing LET. In the region of 100-200 keV/micrometer the RBE for initial DSB induction was clearly below unity. Rejoining of DSBs was assessed by measuring the fraction of DNA retained following post-irradiation incubation of cells under culture conditions. After exposure to Ca ions, DSB rejoining was considerably impaired compared to X-rays.     

The effect of track structure on cell inactivation and chromosome damage at a constant LET of 120 keV/micrometer.

The influence of track structure on chromosome damage and cell inactivation are being investigated. Plateau-phase normal human fibroblast cultures were irradiated with gamma rays, and He, Ne and Ar ions. Particle velocities were chosen so that all beams had an LET of 120 keV/micrometer. In this constant-LET experimental design, the radial distribution of excitations and ionizations about the particle track is the most significant variable. Using premature chromosome condensation, chromatin breaks were measured at two time points, promptly after irradiation and after a prolonged incubation to allow for repair. These measurements give an indication of both initial chromosomal damage and also residual damage that is either not repaired or is misrepaired. Survival was measured under the same conditions. Results indicate that the RBEs for both cell inactivation and, to a lesser extent, chromosome damage decrease as particle energy increases.     

The role of DNA repair on cell killing by charged particles.

It can be noted that it is not simple double strand breaks (dsb) but the non-reparable breaks that are associated with high biological effectiveness in the cell killing effect for high LET radiation. Here, we have examined the effectiveness of fast neutrons and low (initial energy = 12 MeV/u) or high (135 MeV/u) energy charged particles on cell death in 19 mammalian cell lines including radiosensitive mutants. Some of the radiosensitive lines were deficient in DNA dsb repair such as LX830, M10, V3, and L5178Y-S cells and showed lower values of relative biological effectiveness (RBE) for fast neutrons if compared with their parent cell lines. The other lines of human ataxia-telangiectasia fibroblasts, irs 1, irs 2, irs 3 and irs1SF cells, which were also radiosensitive but known as proficient in dsb repair, showed moderated RBEs. Dsb repair deficient mutants showed low RBE values for heavy ions. These experimental findings suggest that the DNA repair system does not play a major role against the attack of high linear energy transfer (LET) radiations. Therefore, we hypothesize that a main cause of cell death induced by high LET radiations is due to non-reparable dsb, which are produced at a higher rate compared to low LET radiations.     

Residual chromatin breaks as biodosimetry for cell killing by carbon ions

We have studied the relationship between cell killing and the induction of residual chromatin breaks on various human cell lines and primary cultured cells obtained by biopsy from patients irradiated with either X-rays or heavy-ion beams to identify potential bio-marker of radiosensitivity for radiation-induced cell killing. The carbon-ion beams were accelerated with the Heavy Ion Medical Accelerator in Chiba (HIMAC). Six primary cultures obtained by biopsy from 6 patients with carcinoma of the cervix were irradiated with two different mono-LET beams (LET= 13 keV/μm, 76 keV/μm) and 200kV X rays. Residual chromatin breaks were measured by counting the number of non-rejoining chromatin fragments detected by the premature chromosome condensation (PCC) technique after a 24 hour postirradiation incubation period. The induction rate of residual chromatin breaks per cell per Gy was the highest for 76 keV/μm beams on all of the cells. Our results indicated that cell which was more sensitive to the cell killing was similarly more susceptible to induction of residual chromatin breaks. Furthermore there is a good correlation between these two end points in various cell lines and primary cultured cells. This suggests that the detection of residual chromatin breaks by the PCC technique may be useful as a predictive assay of tumor response to cancer radiotherapy.     

Dose protraction studies with low- and high-LET radiations on neoplastic cell transformation in vitro.

A major objective of our heavy-ion research is to understand the potential carcinogenic effects of cosmic rays and the mechanisms of radiation-induced cell transformation. During the past several years, we have studied the relative biological effectiveness of heavy ions with various atomic numbers and linear energy transfer on neoplastic cell transformation and the repair of transformation lesions induced by heavy ions in mammalian cells. All of these studies, however, were done with a high dose rate. For risk assessment, it is extremely important to have data on the low-dose-rate effect of heavy ions. Recently, with confluent cultures of the C3H10T1/2 cell line, we have initiated some studies on the low-dose-rate effect of low- and high-LET radiation on cell transformation. For low-LET photons, there was a decrease in cell killing and cell transformation frequency when cells were irradiated with fractionated doses and at low dose rate. Cultured mammalian cells can repair both subtransformation and potential transformation lesions induced by X rays. The kinetics of potential transformation damage repair is a slow one. No sparing effect, however, was found for high-LET radiation. There was an enhancement of cell transformation for low-dose-rate argon (400 MeV/u; 120 keV/micrometer) and iron particles (600 MeV/u; 200 keV/micrometer). The molecular mechanisms for the enhancement effect is unknown at present.     

Neoplastic cell transformation by energetic heavy ions and its modification with chemical agents.

For many years we have been interested in understanding the potential carcinogenic effects of cosmic rays. We have studied the oncogenic effects of cosmic rays with accelerator-produced heavy particle radiation and with a cultured mammalian cell system--C3H10T1/2 cells. Our quantitative data obtained with carbon, neon, silicon, and iron particles showed that RBE is both dose and LET dependent for neoplastic cell transformation. RBE is higher at lower dose, and RBE increases with LET up to about 200 keV/micrometer. In nonproliferation confluent cells, heavy-ion induced transformation damage may not be repairable, although a dose modifying factor of about 1.7 was observed for X-ray radiation. Our recent studies with super-heavy high-energy particles, e.g., 960 MeV/U U235 ions (LET = 1900 keV/micrometer), indicate that these ions with a high inactivation cross-section can cause neoplastic cell transformation. The induction of cell transformation by radiation can be modified with various chemicals. We have found that the presence of DMSO (either during or many days after irradiation) decreased the transformation frequency significantly. It is, therefore, potentially possible to reduce the oncogenic effect of cosmic rays in space through some chemical protection.     

Biochemical mechanisms and clusters of damage for high-LET radiation.

Using mechanisms of indirect and direct radiation, a generalized theory has been developed to account for strand break yields by high-LET particles. The major assumptions of this theory are: (i) damage at deoxyribose sites results primarily in strand break formation and (2) damage to bases leads to a variety of base alterations. Results of the present theory compare well with cellular data without enzymatic repair. As an extension of this theory, we show that damage clusters are formed near each double strand break for high-LET radiation only. For 10 MeV/n (LET = 450 keV/micrometer) neon ions, the results show that on average there are approximately 3 additional breaks and approximately 3 damaged bases formed near each double strand break. For 100 MeV/n helium ions (LET = 3 keV/micrometer), less than 1% of the strand breaks have additional damage within 10 base pairs.