The cancerocidal effect of radiotherapy is dependent on delivered dose. This is limited by normal tissue tolerance. However, at radiation dose-schedules well tolerated by most individuals, some patients develop late radionecrosis. The factors accounting for these interindividual differences are unknown (Safwat et al., 2002).

Four human radiosensitivity syndromes are related to defective DNA repair: ataxia telangiectasia (AT) (Taylor et al., 1975), Nijmegen breakage syndrome (NBS) (Taalman et al., 1983), ataxia telangiectasia-like disease (ATLD) (Stewart et al., 1999) and DNA ligase IV deficiency (O'Driscoll et al., 2001). All are rare autosomal recessive conditions. Affected individuals have often been children with severe, sometimes fatal, toxicity from conventional radiotherapy.

Most patients with late radionecrosis are adults with no preradiation phenotype. They may be intrinsically more radiosensitive, but form a continuum with the 'normal' population. Identifying these sensitive individuals may allow a higher radiation dose to be delivered to other patients, with improved chance of cure (Tucker et al., 1996). Studies of intrinsic radiosensitivity have shown a correlation with patient survival (West et al., 1997; Bjork-Eriksson et al., 1998) and late tissue morbidity (West et al., 1998; Alsbeih et al., 2000). They have not identified the underlying biochemical mechanisms. Two reports have suggested the importance of DNA repair proficiency. Herring et al. (1998) found a correlation between HAP1 expression and clonogenic survival after clinically relevant doses of radiotherapy. Alapetite et al. (1999) observed impaired strand-break rejoining in lymphocytes from six patients with RTOG grade 4 reactions. The control patients included individuals who had been irradiated several years previously, indicating the impaired rejoining was not consequent upon radiotherapy.

Taken together, these studies suggest that DNA repair activity affects clinical radiosensitivity. The predominant ionizing radiation-induced lethal DNA lesion is the double-strand break (DSB) (Iliakis, 1991). Its major repair pathway is nonhomologous end-joining (NHEJ). To explore the role of DNA repair in adult cancer patients with late radionecrosis, we derived cell lines from five patients with RTOG grade 4 late radionecrosis. We hypothesized that patients developing radionecrosis due to impaired DNA repair might be cancer-prone, and selected patients under age 50 years. Patients with necrosis but not fibrosis were selected because a cell-death end point was thought to be more likely related to impaired DNA repair than a proliferative end point such as fibrosis. Local Research Ethics Committee approval was obtained and patients provided written informed consent.

Initially, we investigated the expression and activity of enzymes involved in NHEJ in the five cell lines. We report two cell lines that exhibit reduced post-radiation viability, impaired DSB rejoining and reduced DNA-dependent protein kinase (DNA-PK) activity in vitro.

Cell line LB0003 was derived from a patient who presented with breast cancer at 40 years of age, treated by mastectomy and adjuvant radiotherapy. She developed chest wall necrosis requiring reconstruction after 17 years, contralateral primary breast cancer after 18 years and a presumed ipsilateral lung cancer 26 years later.

Cell line LB0004 was derived from a patient with FIGO stage Ib squamous cell cervix carcinoma at 43 years of age, treated with external beam and intracavitary radiotherapy. After 2 years, she developed vesicovaginal and ileovesical fistulae. Both fistulae were managed surgically and she remains disease-free 10 years later. Control cell lines used were all immortalized with Epstein–Barr virus (EBV) (Table 1).

Table 1 - Characteristics of control cell lines.

Full table

LB0003 and LB0004 exhibit moderately increased radiosensitivity

To determine whether these patients were predisposed to radionecrosis by increased cellular radiosensitivity, we assayed postradiation viability in LB0003, LB0004, positive and negative control cell lines. We chose a viability assay because of the difficulty associated with colony-forming clonogenic survival assays with lymphoblastoid cells in suspension. The number of proliferating cells after 0, 0.5, 1.0 and 2.0 Gy was determined for the cell lines LB0003 and LB0004, derived from patients with late radionecrosis, and for LB0001, Macsi, Wales and BD 2630. The postradiation viability curves for LB0003 and LB0004 were intermediate between those of the normal controls and the acute over-reactor (LB0001), and the AT cell line (Figure 1). After 2 Gy, the proportion of cells remaining viable was 38.1% for Wales, 38.2% for Macsi, 35.1% for LB0001, 21.0% for LB0003, 19.9% for LB0004 and 8.6% for BD2630.

Figure 1.

Postradiation viability of cell lines. Data are shown for cell lines Macsi (__ . . __) Wales (__ . . __) LB0001 (- - -) LB0003 (. . . .) and LB0004 (__ . __) and BD2630 (____) from an individual with AT. EBV-immortalized lymphoblastoid cells were established by standard techniques (Godsen et al., 1986). Control lymphoblastoid cell lines (Macsi and Wales) from individuals with abnormal ocular development, but no history of cancer, were donated by Professor Veronica van Heyningen. BD 2630 was obtained from the European Collection of Cell Cultures (Salisbury, UK). Cells were grown in RPMI 1640 medium (Gibco BRL) with 10% FCS, 2 mM L-glutamine, penicillin and streptomycin. Exponentially growing cells were irradiated at 20°C with a 6 MV linear accelerator at 3.2 Gy/min. Cells were plated in triplicate at densities of 5 10⁴–5 10⁵ cells/ml and incubated at 37°C. A colorimetric proliferation assay using a compound (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS) reduced by metabolically active cells was used to determine viable cells (Promega Cell Titer 96 System). Similar assays have been shown to correlate well with clonogenic assays in estimating radiosensitivity (Price and McMillan, 1990). On days 1, 2, 3, 4 and 5 postirradiation, 20 l MTS was added to each well and incubated for 4 h. Optical density (OD) measurements were made at 450 nm, and triplicate data averaged after correction for background absorbance. Surviving fractions were estimated by dividing the mean OD of the irradiated sample wells by that of the nonirradiated sample wells from the day 5 assay results, at which time cell lines had returned to exponential growth following radiation-induced cell cycle delay, and before contact inhibition of cell growth within the wells develops. However, unirradiated cells had often ceased exponential growth by day 4, and control data were extrapolated to day 5 to allow calculation of the surviving fraction. This approach has been necessary when using the MTT proliferation assay (Price and McMillan, 1990)

Full figure and legend (17K)

Postradiation DSB repair is reduced in LB0003 and LB0004

To ascertain whether this reduced postradiation viability was due to impaired DNA repair, we investigated the induction and rejoining of DNA DSBs using PFGE. Unsynchronized cell cultures were used. Initial DSB induction by 30 Gy ionizing radiation was similar in LB0001, LB0003, LB0004 and Macsi (data not shown). Significantly fewer DSBs were repaired by LB0003 and LB0004 than by Macsi and LB0001 (Figure 2). Almost half (42.9% for LB0001 and 43.7% for Macsi) of DSB observed at 0 min were repaired in 30 min by the control cell lines, whereas only 8.6 and 5.4% were repaired by LB0003 and LB0004. In general, DSB rejoining was complete by 4 h, with normal controls restituting 72–81% of damage and the LB0003 and LB0004 only 35–53%. Even after 24 h, this difference persisted. This suggests that the two cell lines from individuals with late radionecrosis exhibit impaired DSB repair. The reduction in DSB repaired by the fast kinetic pathway is in keeping with previous studies of DSB-rejoining kinetics in DNA-PKcs-deficient cell lines (Allalunis-Turner et al., 1995b; DiBiase et al., 2000).

Figure 2.

Kinetics of repair of DNA damage measured by PFGE assay. Unsynchronized cells from cell lines Macsi (––), LB0001 (), LB0003 (––) and LB0004 () were counted, aliquoted (4 10⁵) and irradiated on ice with 30 Gy using a caesium source (Gammacell 40 Exactor) at 1.06 Gy/min. Following irradiation, cells were incubated with fresh RPMI in 5% CO₂ at 37°C. Unirradiated cells and cells at each time point were harvested and washed. Each pellet was immediately resuspended in 0.5% Sea Plaque low melting temperature agarose at 38°C. The mixture was pipetted into prewarmed 75 l plug moulds (Bio-Rad) and allowed to set at 4°C. Cell plugs were transferred to 50 ml tubes containing three volumes of ice cold digestion cocktail (1% N-lauroylsarcosine (Sigma), 0.1 mg/ml Proteinase K (Sigma), 0.1 M EDTA, 0.01 M Tris-HCl, 0.02 M NaCl) and incubated for 1 h at 4°C, 2 h at 50°C and in two volumes of the digestion cocktail containing 0.1 mg/ml Proteinase K for 18 h at 50°C. Plugs were washed at room temperature in 50 volumes of Tris-EDTA (pH 7.6) with five changes of buffer over 3 h, in two volumes of Tris-EDTA (pH 7.6) containing 40 g/ml phenylmethylsulphonylfluoride for 30 min at 50°C, and in 50 volumes of Tris-EDTA (pH 7.6) with three changes of buffer and stored in the same solution at 4°C. Cell plugs were cut into 25 l samples and loaded into the wells of a 0.7% agarose (Seakem Gold agarose, BMA) gel prepared in 0.5 TBE (45 mM Tris, 1 mM EDTA, 45 mM boric acid, pH 8.5). A reference Saccharomyces cerevisiae chromosome molecular weight marker was added to each gel. The wells were sealed with 0.7% LMP agarose. The PFGE unit (Bio-Rad CHEF-DRII) was filled with 2 l of 0.5 TBE. The sample (at 10°C) was run at 45 V (1.7 V/cm) with a pulse time of 75 min for 40 h, followed by 100 V, 30 min pulse time, for 2 h. Following electrophoresis, the gel was dried, rinsed with double distilled water and placed in 150 ml SYBR Green I (BMA) (1 : 10 000) in 0.5 TBE pH 8.0 at 50°C for 3 h, cooled for 10 min on ice and analysed using a SynGene Gel Documentation System with a Syb-100 SYBR green filter. Gel images were analysed using the Gene Tools software (SynGene). The residual damage fraction was calculated as the ratio of the intensity of the material entering the gel lane to the total intensity of plug plus this material. All data were corrected by subtraction of the nonspecific migration detected in unirradiated control samples. The DSB remaining at a given time of postirradiation cell incubation was normalized to the incidence of DSB at time zero set at 100%. Three independent experiments were carried out for each cell line with two replicates per experiment

Full figure and legend (23K)

DNA-PK activity is reduced in cell-free extracts from LB0003 and LB0004

The DNA-PK complex plays a key role in NHEJ. We hypothesized that DNA-PK kinase activity might be one of the determining factors in DSB repair, and a surrogate for the major repair activity in these experiments, where the majority of cells were in G₁ or early S phase. We assayed DNA-PK activity in the cell lines LB0003 and LB0004 from patients with late morbidity, LB0001 from an acute over-reactor, and Macsi and Wales from individuals without cancer. To assay DNA-PK activity, DNA containing termini and all three complex subunits are required. Phosphorylation of a DNA-PK peptide substrate was measured in the presence of calf thymus DNA and corrected for background phosphorylation (in the absence of calf thymus DNA) by nonspecific kinases. We observed a 6–10-fold reduction in DNA-PK activity in the cell lines LB0003 and LB0004 compared to LB0001, Macsi and Wales (P<0.0001 for LB0003 and LB0004 vs Macsi; P<0.0005 for LB0003 and LB0004 vs Wales; and P<0.002 for LB0003 and LB0004 vs LB0001) (Figure 3).

Figure 3.

DNA-PK activity in cell lines. Exponentially growing cells from the cell lines Macsi, Wales, LB0001, LB0003 and LB0004 were harvested, and whole-cell extracts prepared (Manley et al., 1983). Briefly, 0.8–1 l of cells at a density of 0.5–1.0 10⁶ cells/ml were pelleted and washed. The pellet was resuspended in four packed cell volumes (PCV) hypotonic lysis buffer (10 mM Tris-HCl pH 8, 1 mM EDTA, 5 mM DTT and protease inhibitors) and left on ice for 20 min. The cells were disrupted in a Dounce homogenizer, and an additional 4 PCV 50 mM Tris-HCl pH8, 10 mM MgCl₂, 2 mM DTT, 25% sucrose and 50% glycerol added, followed by one PCV ammonium sulphate. After 30 min, the extract was centrifuged at 38 000 r.p.m. for 3 h at 4°C. The supernatant was removed, and 0.33 g/ml ammonium sulphate added. The extract was centrifuged at 10 000 g for 30 min, and the precipitate recovered and resuspended in buffer containing 25 mM HEPES pH 7.9, 100 mM KCl, 12 mM MgCl₂, 1 mM EDTA and 17% glycerol. This was dialysed against two changes of the same buffer for 6 h, centrifuged at 10 000 g for 10 min, and the supernatant snap frozen and stored at -70°C. DNA-PK activity was assayed by measuring the phosphorylation of a biotinylated p53-derived peptide substrate (Promega SignaTECT DNA-Dependent Protein Kinase (DNA-PK) Assay System) by cell extracts in the presence or absence of calf-thymus DNA. Incorporation of -[³²P] into the peptide was analysed with a phosphorimager (confirmed by scintillation counting) and DNA-PK activity calculated with a correction for background kinase activity. Results were based on at least three experiments using extracts prepared on at least two separate occasions

Full figure and legend (29K)

To confirm that other DNA damage and signalling pathways, including ATM which may also be stimulated by DNA strand breaks were intact (and hence the reduction in protein kinase activity in the presence of DNA ends was due to reduced DNA-PK activity), we measured the levels of p53 after 4 Gy ionizing radiation to confirm that normal upregulation of the protein occurred after radiation of these cell lines. This was normal in all cell lines tested, indicating normal ATM activity in these cell lines (data not shown).

A correlation between DSB rejoining, radiosensitivity and DNA-PK activity has been reported in lung cancer (Sirzen et al., 1999) and squamous cell cancer cell lines (Polischouk et al., 1999), and between radiosensitivity and DNA-PK activity in oesophageal cancer cell lines (Zhao et al., 2000). In contrast, one group reported no correlation between DNA-PK activity and radiosensitivity in glioma cell lines (Allalunis-Turner et al., 1995a). This may indicate that other factors have more effect on radiosensitivity in gliomas, which are highly radioresistant.

The reduced DNA-PK activity we observed in EBV-immortalized cell lines does not seem to be an artefact of transformation or of previous radiotherapy. Normal levels of activity were seen in all the control EBV-transformed cell lines, one of which (LB0001) was derived from a patient who had received radiotherapy over 2 years previously. This strongly suggests that the association of reduced DNA-PK activity with reduced postradiation viability in these cell lines is a reflection of an underlying DNA repair deficiency in vivo, and may have contributed to the injury in these patients. Another group has used EBV-transformed cells to investigate radiosensitivity, and has also shown an association between increased cellular radiosensitivity and the risk of late injury (Alapetite et al., 1999).

All DNA-PK subunits are present in cell-free extracts from LB0003 and LB0004

The residual DNA-PK activity observed in the cell lines LB0003 and LB0004 suggested that these cell lines were unlikely to be totally deficient in DNA-PKcs, Ku70 or Ku80. To confirm this, we performed immunoblots against these proteins in the cell lines with reduced DNA-PK activity, LB0003 and LB0004, and controls (LB0001, Macsi and Wales). We could not detect a significant difference in the level of Ku 80 and 70 proteins among any of the cell lines (Figure 4a and b). DNA-PKcs was present in both the cell lines with reduced activity and the control cell lines (Figure 4c).

Figure 4.

(a-c) Immunoblot of subunits of DNA-PK in indicated cell lines. Exponentially growing cells from cell lines Macsi, Wales, LB0001, LB0003 and LB0004, were lysed in 50 mM Tris pH 7.5, 1 mM EDTA, 10 mM DTT, 0.2% Triton X-100 and a protease inhibitor cocktail (Sigma). Aliquots (20 l) of each sample containing equivalent amounts of protein were separated by SDS–PAGE. Following transfer to nitrocellulose membrane (Amersham), Ponceau's staining was used to confirm protein loading and transfer. After blocking with 5% milk, the membranes were probed with (a) anti-Ku80 (Serotec AHP318), (b) anti-Ku 70 (Serotec AHP316) and (c) anti-DNA-PKcs (Ab-2, Neomarkers, raised against the N-terminus of the protein). Detection was performed using ECL (Amersham). The immunoblots were performed at least twice using separate extracts in each case. The marker was from Amersham (RPN 800)

Full figure and legend (185K)

Conclusion

No humans with a DNA-PKcs or Ku null phenotype have been described. While the two cell lines described here exhibit substantially reduced DNA-PK activity, both have some residual function and express all three protein components of DNA-PK. It is possible that this reduced level of activity is sufficient for function under normal conditions, but insufficient to deal with the extent of DNA repair required following radiotherapy.

Mutations described in transgenic mice and cell lines have shown protein truncations in one of the DNA-PK components, impaired protein-to-protein interactions between the DNA-PK subunits and reduced kinase activity. Our data exclude the former possibility, since we have observed normal size product of all three proteins in immunoblots. However, we cannot, on the current evidence, exclude mutations within the genes encoding one of the three DNA-PK subunits, affecting either protein–protein interactions, DNA-binding or the kinase activity itself, or in genes encoding other unidentified proteins, which interact with or post-translationally modify one of the subunits of DNA-PK. One cell line in the XRCC7 complementation group (XR-C2) has been described with normal levels of DNA-PKcs, but no detectable kinase activity. This cell line exhibited significantly increased radiosensitivity and impaired DSB repair. Sequence analysis of the PRKDC gene in XR-C2 demonstrated a point mutation producing a single amino-acid substitution six residues from the C-terminus of the DNA-PKcs protein (Woods et al., 2002), but sequencing of the cell lines reported here has not shown a similar mutation.

This is the first report of a biochemical abnormality in cell lines derived from patients with late radionecrosis at conventional doses; that this occurs in two of the five patients investigated in a pathway so central to recovery from radiation damage in human cells raises the possibility that mild deficiencies in the dominant pathway of DSB repair, perhaps related to single nucleotide polymorphisms in the DNA-PK genes, might be a common cause of late radionecrosis.

References

1. Alapetite C, Thirion P, de la Rochefordiere A, Cosset JM and Moustacchi E. (1999). Int. J. Cancer, 83, 83–90. | Article | PubMed | ISI | ChemPort |
2. Allalunis-Turner J, Lintott LG, Barron GM, Day RS and Lees-Miller SP. (1995a). Cancer Res., 55, 5200–5202. | PubMed | ChemPort |
3. Allalunis-Turner J, Zia PK, Barron GM, Mirzayans R and Day III RS. (1995b). Radiat. Res., 144, 288–293. | PubMed | ChemPort |
4. Alsbeih G, Malone S, Lochrin C, Girard A, Fertil B and Raaphorst GP. (2000). Int. J. Radiat. Oncol. Biol. Phys., 46, 143–152. | Article | PubMed |
5. Bjork-Eriksson T, West CM, Karlsson E, Slevin NJ, Davidson SE, James RD and Mercke C. (1998). Br. J. Cancer, 77, 2371–2375. | PubMed |
6. DiBiase SJ, Zeng ZC, Chen R, Hyslop T, Curran Jr WJ and Iliakis G. (2000). Cancer Res., 60, 1245–1253. | PubMed | ISI | ChemPort |
7. Godsen CM, Davidson C and Robertson M. (1986). Human Cytogenetics A Practical Approach, Rooney DE, Czepulowski BH (eds). IRL Press: Oxford, pp. 31–54.
8. Herring CJ, West CM, Wilks DP, Davidson SE, Hunter RD, Berry P, Forster G, MacKinnon J, Rafferty JA, Elder RH, Hendry JH and Margison GP. (1998). Br. J. Cancer, 78, 1128–1133. | PubMed | ISI | ChemPort |
9. Iliakis G. (1991). Bioessays, 13, 641–648. | Article | PubMed | ISI | ChemPort |
10. Manley JL, Fire A, Samuels M and Sharp PA. (1983). Methods Enzymol., 101, 568–582. | Article | PubMed | ISI | ChemPort |
11. O'Driscoll M, Cerosaletti KM, Girard PM, Dai Y, Stumm M, Kysela B, Hirsch B, Gennery A, Palmer SE, Seidel J, Gatti RA, Varon R, Oettinger MA, Neitzel H, Jeggo PA and Concannon P. (2001). Mol. Cell, 6, 1175–1185. | Article |
12. Polischouk AG, Cedervall B, Ljungquist S, Flygare J, Hellgren D, Grenman R and Lewensohn R. (1999). Int. J. Radiat. Oncol. Biol. Phys., 43, 191–198. | Article | PubMed |
13. Price P and McMillan TJ. (1990). Cancer Res., 50, 1392–1396. | PubMed | ISI | ChemPort |
14. Safwat A, Bentzen SM, Turesson I and Hendry JH. (2002). Int. J. Radiat. Oncol. Biol. Phys., 52, 198–204. | Article | PubMed |
15. Sirzen F, Nilsson A, Zhivotovsky B and Lewensohn R. (1999). Eur. J. Cancer, 35, 111–116. | Article | PubMed | ISI | ChemPort |
16. Stewart GS, Maser RS, Stankovic T, Bressan DA, Kaplan MI, Jaspers NG, Raams A, Byrd PJ, Petrini JH and Taylor AM. (1999). Cell, 99, 577–587. | Article | PubMed | ISI | ChemPort |
17. Taalman RD, Jaspers NG, Scheres JM, de Wit J and Hustinx TW. (1983). Mutat. Res., 112, 23–32. | Article | PubMed | ISI | ChemPort |
18. Taylor AM, Harnden DG, Arlett CF, Harcourt SA, Lehmann AR, Stevens S and Bridges BA. (1975). Nature, 258, 427–429. | Article | PubMed | ISI | ChemPort |
19. Tucker SL, Geara FB, Peters LJ and Brock WA. (1996). Radiother. Oncol., 38, 101–113. | Article |
20. West CM, Davidson SE, Elyan SE, Swindell R, Roberts SA, Orton CJ, Coyle CA, Valentine H, Wilks DP, Hunter RD and Hendry JH. (1998). Int. J. Radiat. Biol., 73, 409–413. | Article | PubMed |
21. West CM, Davidson SE, Roberts SA and Hunter RD. (1997). Br. J. Cancer, 76, 1184–1190. | PubMed | ISI | ChemPort |
22. Woods T, Wang W, Convery E, Errami A, Zdzienicka MZ and Meek K. (2002). Nucleic Acids Res., 19, 5120–5128. | Article |
23. Zhao HJ, Hosoi Y, Miyachi H, Ishii K, Yoshida M, Nemoto K, Takai Y, Yamada S and Suzuki N. (2000). Clin. Cancer Res., 6, 1073–1078. | PubMed | ISI | ChemPort |

Acknowledgements

Our thanks to Professor Trevor McMillan and colleagues in Professor Malcolm Dunlop's group for discussion and advice, and to Professor Malcolm Paterson for a critical reading of the manuscript. This work was supported by grants from the Singapore National Cancer Centre, Singapore Cancer Society, the Singapore National Medical and SingHealth Research Councils, the Melville Trust for the Care and Cure of Cancer and the Scottish Hospitals Endowments Research Trust.