INTRODUCTION

The important cellular effects of exposure to ionizing radiation arise primarily from the induction of DNA double-strand breaks (DSBs) as a consequence of the nature of the energy deposition in DNA by ionizations that produce local clustered damage (1). Such DSBs can be repaired in mammalian cells by two recombinatorial pathways, nonhomologous end joining (NHEJ) and homologous recombination (HR) (2, 3). NHEJ is active throughout the cell cycle (predominating in G₀/G₁-phase cells) and is an error-prone process that does not require extensive regions of homology at the site of repair. Repair of DSBs by homologous recombination, on the other hand, is an error-free mechanism that occurs primarily in late S and G₂ phases of the cell cycle; HR uses homologous chromosome regions (the sister chromatid, in this case) as a template for repair. While the importance of homologous recombinational processes is clear and, in fact, essential (full knockout of many genes involved with HR is embryonic lethal), the relative contribution of HR to the processing and repair of DNA damage induced directly by ionizing radiation is still not clear (4).

In earlier studies, we reported a prominent bystander effect for the induction of sister chromatid exchanges (SCEs) after α-particle irradiation. Approximately 30% of the cells in confluent cultures showed increased levels of SCEs when 1% or fewer were traversed by an α particle (5). The bystander effect for the induction of HPRT mutations or chromosomal aberrations in wild-type cells was much smaller (6, 7). Interestingly, the mutations induced in bystander cells were almost entirely point mutations, in contrast to directly irradiated cells, in which total and partial gene deletions predominated (8). NHEJ-deficient cells, however, showed a very marked bystander effect for both chromosomal aberrations and HPRT mutations; mutations in these cells consisted primarily of deletions (7, 9). We hypothesize that the mutations induced in wild-type bystander cells are largely the result of oxidative base damage (10). While very few DSBs are produced by various reactive oxygen species (ROS), generated either directly by radiation or indirectly through a signal transduction process (10), they can arise from clustered lesions and may be increased transiently in irradiated normal cells at replication forks (11). Normally these DSBs are effectively repaired, but in NHEJ-deficient cells they are not, rendering bystander cells in these cultures more sensitive to the induction of large-scale mutations and chromosomal aberrations.

Since HR is known to play an essential role in the maintenance of genomic integrity, the present study was designed to examine the potential role of HR in the bystander effect. We examined the induction of chromosomal aberrations and specifically SCEs, since SCEs are thought to be a consequence of HR processes in which one possible outcome can be a crossover at the damage site resulting in the exchange of sister chromatids (12–14). We employed a hamster cell line deficient in RAD51C, a RAD51 paralog known to be involved in HR and Holliday junction processing in human cells (15).

MATERIALS AND METHODS

RESULTS

Table 1 summarizes the radiosensitivities as measured by clonogenic survival for the four cell lines examined in this study, along with their corresponding PRKDC (DNA-PKcs) and RAD51C genotypes. The V3/vec line is a clonal derivative of the V3 cell line, an NHEJ-deficient line originally derived from CHO cells (19). As confirmed by the immunoblot analysis shown in Fig. 1, V3/vec cells are deficient in DNA-PKcs. The HR-deficient line irs3 was derived from wild-type V79 cells (16), and the RAD51C deficiency is also confirmed by the immunoblot analysis shown in Fig. 1. As is characteristic for other cell lines deficient in NHEJ, the V3/vec cells are highly radiosensitive with respect to cell killing (Table 1) and are also hypersensitive for induction of chromosomal aberrations (see Fig. 4). On the other hand, the irs3 cells defective in HR showed no increased sensitivity for either end point of radiation damage.

Figure 2A and B shows the results of experiments measuring the induction of SCE in CHO and NHEJ-defective V3/vec cells (Fig. 2A) and V79 and HR-defective irs3 cells (Fig. 2B) by very low fluences of α particles. The mean doses for these experiments ranged from 0 to 2.6 mGy, corresponding to less than 1% of the cells experiencing a nuclear traversal by these α particles for the highest doses. The spontaneous levels of SCE per chromosome for unirradiated cells were approximately 0.34 for CHO and V79 cells, 0.40 for V3/vec cells, and only 0.16 for the HR-defective irs3 cells.

The induced levels of SCEs in wild-type CHO cells rose rapidly for the lower doses, reaching a plateau at approximately 1.3 mGy, where on average only one cell in 217 experienced a nuclear traversal. This dose–response curve is virtually identical to those reported earlier in two independent studies (5, 25). There appeared to be a much greater bystander effect in NHEJ-defective V3/vec cells (Fig. 2A), with the maximal level of induced SCEs occurring for a mean dose of 0.35 mGy. This dose is about one-quarter the dose for the same effect in wild-type CHO cells, a relative sensitivity similar to the ratio of doses for equal cell killing of these cells. The induction of SCEs for identical mean doses of α particles in wild-type V79 and HR-defective irs3 cells is shown in Fig. 2B. No significant induction of SCE in the irs3 cells was observed after α-particle irradiation at 0, 8 or 13 h after release from G₁, when the cells had reached late S/G₂ phase.

More detailed analyses of the results reported in Fig. 2B are shown in Fig. 3A–D, in which the distributions of frequencies of both spontaneous SCEs per chromosome (0 mGy) and total (spontaneous plus induced) SCEs per chromosome after irradiation with 1.3 mGy in V79 and irs3 cells are plotted. The frequency distributions of SCEs per chromosome on a per cell basis were approximately Poisson. Thus not only did the RAD51C-deficient irs3 cells have significantly lower spontaneous SCE levels, but α particles did not significantly induce SCEs in HR-defective bystander cells, at least in this dose range. As was shown previously for CHO cells (5, 25), the steep increase in mean SCE frequencies after low-dose α-particle irradiation in wild-type V79 cells is the result of SCEs occurring in bystander cells; less than 1% of the cells would be traversed by an α particle after a dose of 1.3 mGy. On the other hand, no significant shift in mean SCE frequencies occurred in irs3 cells (Fig. 3C, D); the mean SCE per chromosome was 0.160 ± 0.008 (SEM) for unirradiated cells and 0.171 ± 0.006 (SEM) after a mean dose of 1.3 mGy.

We have previously shown a greatly enhanced bystander effect for chromosomal aberrations in cells defective in the NHEJ pathway (7). In that study, cells of a mouse DNA-PKcs knockout cell line irradiated with a mean dose of 1.7 mGy were shown to have a 10-fold higher number of chromosomal aberrations per α-particle track than wild-type cells, reflecting a 10-fold increase in the fraction of cells with chromosomal aberrations. As can be seen in Fig. 4A, a very similar result was found in the present study for the V3/vec hamster cell line deficient in DNA-PKcs. On the other hand, irs3 cells demonstrate a dose–response curve for the induction of chromosomal aberrations that is identical to that of wild-type V79 cells, both for very low doses (Fig. 4B) and for doses up to 400 mGy when each cell received multiple hits (data not shown). It should be noted that in all these low-dose experiments, chromatid-type aberrations predominated, consistent with the hypothesis that the majority of DNA damage generated in bystander cells is single-stranded (e.g. base/sugar damage, single-strand breaks), not double-stranded (i.e. damage that would generate chromosome-type aberrations).

DISCUSSION

The data presented in Fig. 4 confirm our previous findings (7, 9) and indicate that bystander cells deficient in the NHEJ pathway are highly sensitive to the induction of chromosomal aberrations. High concentrations of reactive oxygen species have been reported in bystander cells (10, 26). We have hypothesized previously that complex DSBs resulting from the processing of single-strand breaks and closely opposed oxidative lesions, in addition to those occurring spontaneously, are efficiently repaired in wild-type cells, whereas they remained unrejoined or misrejoined in NHEJ-deficient cells (7). Such lesions would be highly clastogenic compared with oxidative base damage, which is presumably the major DNA lesion in wild-type bystander cells. This is consistent with the observation that mutations in wild-type bystander cells are almost entirely point mutations, whereas deletions are the predominant lesion in NHEJ-deficient cells (8, 9). Since chromosome-type aberrations are produced in mitosis after irradiation in G₁ phase, such complex DSBs must also be induced in G₁. The fact that the majority of aberrations induced in wild-type bystander cells were chromatid types also suggests that base damage is a principal lesion underlying the observed effects (27).

This hypothesis would presume that unrepaired and/or misrepaired DNA damage occurring in NHEJ-deficient bystander cells is not adequately repaired by the HR pathway. This conclusion is consistent with the observation that the bystander response for chromosomal aberrations in cells defective in the error-free HR pathway irradiated 0–13 h after subculture was small and identical to that of wild-type cells (Fig. 4). It is furthermore consistent with the radiosensitivity of these lines (Table 1); that is, the marked bystander response for chromosomal aberrations in the highly radiosensitive NHEJ-defective line compared to the very small bystander effect for wild-type and HR-deficient cells, both of which showed normal radiosensitivity.

These conclusions are also supported by the findings for SCE induction. A marked bystander effect for SCE induction was observed for both wild-type cell lines, whereas a much greater sensitivity to the induction of SCEs in bystander cells was apparent in the NHEJ-defective line as manifested by the steep rise in SCEs at very low doses (Fig. 2A). It has been postulated that exchanges between sister chromatids occur as one possible outcome of a homologous recombination process involving the formation of double Holliday junctions designed to correct errors present in the DNA prior to cell division (14). The enhanced bystander response for SCE induction could imply that an enhanced number of DNA lesions were present in NHEJ-deficient cells as they progressed into late S and G₂, stimulating an increased level of homologous recombination between sister chromatids in an attempt to repair these lesions. One could further surmise that this repair was not very successful, as reflected by the enhanced bystander response for chromosomal aberrations. Since all accounts to date suggest that SCEs arise through one possible outcome (crossover) of the homologous recombination process, the greatly diminished bystander response for the RAD51C-deficient irs3 cells and the low frequency of spontaneous SCEs are consistent with their inability to carry out homologous recombination. To further address this issue, additional hamster cell lines mutated in other key HR proteins are currently being investigated

Regarding the above interpretation suggesting base damage as a critical lesion underlying SCE induction in bystander cells, it is also noteworthy that large increases in spontaneous SCEs have been reported in cells from patients with Bloom's syndrome. The very high SCE levels for Bloom's syndrome cells have been attributed to a defect in BLM-TOPO IIIa that is responsible for resolution of double Holliday junctions arising from base damage repair. As already mentioned, without an intact HR pathway, no SCEs would be expected. While interpretations at the level of molecular mechanisms are always attractive, it is also worth noting that branch migrations of double Holliday junctions over tens of Mb along the chromosome would be necessary to explain the sizes of the SCEs seen at the cytogenetic level.