Repair of radiation-induced DNA double-strand breaks in mammalian cells has been thought to process mainly through the NHEJ, with DNA-PK being the central repair protein complex for this pathway. Nevertheless, HR, the most important repair pathway for Dsb in yeast, also contributes to the repair of Dsb in vertebrate cells (Sonoda et al, 2001
) and is essential for their viability. Inactivation of the key HR protein Rad51 in chicken lymphocytes leads to an accumulation of cells in the G2/M with subsequent increase of chromosome aberrations prior to cell death (Sonoda et al, 1998
). In the same way, murine cells deficient in Rad51 are nonviable (Tsuzuki et al, 1996
). In comparison, cells overexpressing Rad51 showed higher recombination frequencies (Lambert and Lopez, 2001
), reduced DNA break frequencies, translocations, chromosome aberrations, decreased induction of apoptosis (Grandy et al, 2002
; Raderschall et al, 2002b
), stimulated tumorigenesis (Bertrand et al, 2003
) and increased resistance to IR (Vispe et al, 1998
To evaluate the role of Rad51 in the radiosensitivity of NSCLC cell lines and the rejoining process of radiation-induced Dsb, expression of Rad51 was modulated with As-ODN targeting Rad51. Since complete loss of Rad51 gene expression leads to cell death in vertebrate cells, the use of As-ODN represents a rational tool and an opportunity to analyse the impact of the modulation of Rad51-dependent homologous recombination pathway in the radiosensitivity of human NSCLC cell lines. After a moderate downregulation of Rad51 expression to about 60% of the nontransfected cells, the formation of Rad51 foci, overall rejoining of Dsb, induction of apoptosis and clonogenic survival were evaluated after exposure of NSCLC cell lines to IR. In addition, γ-H2AX foci formation as a measure for DNA double-strand breaks and its co-localisation with Rad51 were also evaluated.
The present data demonstrate a dose- and time-dependent increase in the percentage of cells with Rad51 foci in NSCLC cell lines. Cell lines with nonfunctional p53, in comparison to the cell lines with functional p53, show a higher expression level of Rad51 at the mRNA level by a factor of about 2 and 5 in G1- and G2-enriched populations, respectively. It has been previously shown that wild-type p53 protein downregulates the level of Rad51 expression in primary cells (Xia et al, 1997
). Reduced expression of Rad51 and reduced spontaneous formation of Rad51 foci have also been found in p53+/+
in comparison to p53−/−
primary mouse embryonic fibroblasts (Kumari et al, 2004
). The activity of Rad51 was shown to be regulated by a transcription-independent direct protein–protein interaction with p53 (Linke et al, 2003
). In cases where p53 is absent or mutated, Rad51 has been shown to have an increased recombination activity. In the present study, however, the high expression level of Rad51 found in cell lines with nonfunctional p53 is not reflected by a high fraction of Rad51 foci-presenting cells. Instead, the proliferation status of tumour cell lines primarily determines the formation of Rad51 foci after IR. Higher expression of Rad51 by a factor of about 2 and a high fraction of Rad51 foci in proliferating (>40%), in comparison to G1-enriched, confluent populations (<5%) was also found in NSCLC cell lines, irrespective of the p53 status. A clear cell-cycle-dependent expression of Rad51 foci has also been found in previous studies with higher fraction of Rad51 foci-presenting cells in the S- and G2-phases in comparison to the G1-phase (Yamamoto et al, 1996
; Yuan et al, 2003
). The highest fraction of Rad51 foci was found at 4–6
h after irradiation, a time where most of the initially radiation-induced Dsb are rejoined via the DNA-PK dependent non-HR mechanism. After this time, the percentage of cells with Rad51 foci gradually decreases. A high fraction of cells with persisting Rad51 foci at 24
h after irradiation was found in radiosensitive cell lines. These persisting rad51 foci possibly reflect Dsb rejoining products sensing the recruitment of HR pathway, obviously with different efficacy in the NSCLC cell lines studied.
In order to explore the possibility that residual Rad51 foci reflect nonrepaired DNA double-strand breaks, phosphorylation of histone γ
-H2AX in response to DNA double-strand breaks produced by IR has been measured, which is proposed to concentrate repair factors at sites of DNA damage (Celeste et al, 2003
). Although differences in the loss of γ
-H2AX foci have been found to be related in part to the intrinsic radiosensitivity of cervical cancer cell lines (Banath et al, 2004
), our data did not support these observations. A co-localisation frequency of more than 80% for Rad51 and γ
-H2AX foci elucidate that persisting Rad51 foci mostly contains DNA double-strand breaks. However, the Rad51-positive foci, in comparison to the H2AX-positive but Rad51-negative foci, obviously have more correlative power for the radiation sensitivity of lung cancer cell lines.
Treatment of exponentially growing cells with As-Rad51 significantly reduces the percentage of cells with Rad51 foci in all NSCLC cell lines, but did not change the fate of Dsb rejoining in the FAR assay. Therefore, no indication of an involvement of Rad51 in the direct rejoining process of radiation-induced Dsb is evident from these data. Even under the condition where the NHEJ pathway is disabled with wortmannin, As-Rad51 has no measurable effect on Dsb rejoining. However, at concentrations resulting in a significant inhibition of DNA-PK-dependent NHEJ, wortmannin also inhibits ATM (Sarkaria et al, 1998
). It has been shown that ATM regulates other proteins known to play a role in DNA-repair, cell-cycle regulation and radiation sensitivity, including p53, CHK2, BRCA1 and NBS1 (Canman et al, 1998
; Cortez et al, 1999
), and is required for the assembly of Rad51 foci (Chen et al, 1999
; Yuan et al, 2003
). Therefore, wortmannin should not only affect NHEJ through inhibition of DNA-PK activity in NSCLC cell lines (Sak et al, 2002
), but also HR via inhibition of ATM-dependent formation of Rad51 foci. Conflicting results regarding the effect of wortmannin on HR has been published, with an increase (Delacote et al, 2002
) and a decrease of HR (Paull et al, 2000
; Allen et al, 2003
). Our results, however, show an efficient inhibition of Rad51 foci formation after treatment with wortmannin, which in part explains its higher effect on the clonogenic survival of NSCLC cell lines when compared to a treatment schedule using As-ODN targeting DNA-PKcs (Sak et al, 2002
). Inhibition of Rad51 foci with wortmannin is not a result of reduced Rad51 expression, but is obviously regulated at the post-translational level, because treatment with wortmannin leads to an increased expression of Rad51 mRNA. The observed effect of wortmannin on NHEJ, HR and cell survival is obviously the result of noncompetitive and covalent binding of wortmannin to the kinase region of DNA-PKcs protein (Izzard et al, 1999
). The formation of covalent adducts may therefore prevent the dissociation of DNA-PKcs from the DNA ends, and thereby block both HR and NHEJ. Alternatively, but not exclusively, wortmannin may inhibit HR through inactivation of ATM, and thereby prevents Rad51 foci formation. Nevertheless, a selective inactivation of NHEJ via As-ODN targeting DNA-PKcs, the key repair protein of NHEJ, also did not change the Dsb rejoining response of NSCLC cell lines to As-Rad51 treatment. Therefore, no contribution of Rad51-dependent HR in the direct removal of radiation-induced Dsb as measured by the FAR assay is evident from these data. In this respect, Wang et al (2001)
also have shown that HR did not measurably contribute to the removal of radiation-induced Dsb, even under conditions where NHEJ is compromised.
Despite the obvious missing of an involvement of Rad51 in the direct rejoining process of radiation-induced Dsb, loss of Rad51 has a deleterious effect on cell survival (Tsuzuki et al, 1996
; Sonoda et al, 1998
). In the same manner, downregulation of Rad51 foci formation has been shown to increase apoptotic cell death (Raderschall et al, 2002b
). Our data support these observations and show that treatment of NSCLC cell lines with As-Rad51 increases radiation-induced apoptosis. Interestingly, the most apoptosis-proficient cell lines after As-Rad51 treatment are H520 and H460, both cell lines with high Rad51 foci remaining at 24
h after irradiation.
Even though Rad51 is expressed at higher levels in tumour cells as compared with normal cells (Raderschall et al, 2002a
), modulation of Rad51 expression has been shown to be more susceptible in tumour cells (Russell et al, 2003
). The increased expression of Rad51 found in tumour cells can be partially attributed to the higher proliferating fraction of tumour cells, because a higher expression of Rad51 mRNA by a factor of about 2 in exponentially growing, in comparison to confluent, NSCLC cell lines is evident from the present study. Nevertheless, the different regulatory processes of Rad51 in normal and tumour cells (Russell et al, 2003
) and its role in radiosensitivity offer the possibility for a selective targeting of radiotherapy.
In conclusion, these data demonstrate that human NSCL cell lines have the same percentage of cells with Rad51 foci at early times after IR, albeit the significant differences in Rad51 expression. Therefore, differences in Rad51 expression alone did not determine the proficiency of Rad51 foci formation. Instead, cell cycle distributions at the irradiation time primarily affect the formation of IR-induced Rad51 foci. Downregulation of Rad51 expression within a specific cell line decreases the fraction of Rad51 foci-presenting cells and significantly increases radiation-induced apoptotic cell death. The relative level of persisting Rad51 foci measured 24
h after irradiation neither correlates to the relative level of persisting γ
-H2AX foci nor to the residual level of DNA double-strand breaks, but was significantly correlated with SF2.