While exogenous DSBs are induced by IR or drugs such as bleomycin or etoposide, endogenous DSBs arise as byproducts of normal intracellular metabolism. It has been estimated that the spontaneous rate of endogenous DSBs may be as high as 50 breaks per cell per cell cycle (Vilenchik and Knudson, 2003
). For example, DSBs can be detected when replication forks stall and collapse, a process that is thought to occur frequently during the S-phase. The cell can repair and/or restart replication forks by multiple mechanisms (Haber and Heyer, 2001
), but the major mechanism that deals with replication-associated DSBs is HR. Intriguingly, replication intermediates can be deliberately broken or cleaved by the Mus81 endonuclease complex and the resulting DSB allows homology-mediated strand invasion, damage bypass, and reconstitution of the replication fork. The essential role of HR in replication is illustrated by the pronounced proliferative defect and embryonic lethality of mice with knockouts of genes that control HR, including the Rad51 recombinase or the breast cancer susceptibility genes BRCA2 or BRCA1 (Powell and Kachnic, 2003
). Indeed, it has been suggested that the primary reason for the existence of HR is the maintenance of functional replication (Klein and Kreuzer, 2002
). It is now clear that the cell takes advantage of the HR machinery to repair exogenous DSBs as well. Chromatid breaks in the S and G2 cell-cycle phases may be predominantly repaired by using the sister chromatid as a template. Therefore, genetic defects in HR can lead to both impaired DNA replication and enhanced IR sensitivity (Thompson and Schild, 2001
; Powell and Kachnic, 2003
). Moreover, impaired HR is also associated with, typically pronounced, hypersensitivity to DNA inter-strand crosslinks (ICLs), which result from cellular exposure to cytotoxic agents such as platinum compounds or mitomycin C.
However, the relationship between the control of replication and repair of IR-induced DSBs is likely not straightforward because DNA lesions caused by IR may not always be ideal substrates for repair by HR mechanisms that are normally employed during replication restart. IR typically generates clusters of ionisations, each containing at least 10 ionisations within a diameter of perhaps five or more nanometers (Steel, 1996
). If such an event impinges on DNA, with the diameter of the double helix being approximately 2.5
nm, it would be expected to cause considerable local damage, including DSBs, single-strand breaks, and base damage. The repair of such a clustered damage site may be complex and/or slow and perhaps under a genetic control that has only partial overlap with the removal of endogenous DSBs. Indeed, mutations in genes that are involved in HR and replication often cause only modest or no radiation hypersensitivity. For example, the anti-recombinogenic effects of the BLM helicase, mutated in Bloom's Syndrome, or the p53 tumour suppressor are not thought to significantly modulate cellular radiation resistance (Dahm-Daphi, 2000
; Thompson and Schild, 2001
; Böhnke et al, 2004
). Mutations in other genes that act in HR have been reported to increase radiation sensitivity by only ~2-fold or less (Thompson and Schild, 2001
). In general, mutations in NHEJ genes lead to greater radiation hypersensitivity than mutations in HR genes, suggesting that NHEJ is the dominant pathway for the removal of IR-induced DSBs. However, it is possible that this relationship changes when IR is combined with radiosensitising chemotherapy, which is the case in many cancer treatments.
Is there also a role for NHEJ in the repair of replication-associated DSBs? While it is difficult to envision how NHEJ should contribute to the restart of collapsed replication forks, NHEJ has been suggested to contribute to the repair of DSBs in the S-phase (Rothkamm et al, 2003
). It is likely that the balance of NHEJ and HR in the removal of DSBs depends on the type and location of the lesion, among other factors. Of note, NHEJ is inherently error-prone and mutagenic, because this process is unable to faithfully restore the original DNA sequence, as opposed to HR, and because NHEJ itself can introduce sequence changes during repair. Still, NHEJ is suggested to be the dominant DSB repair pathway in mammalian cells, which is in part related to the fact that only a very small fraction of the genome is coding for genes and regulatory elements, and that small sequence changes are tolerated by the cells.
Replication, recombination, and repair are intricately linked and cannot be studied separately (Klein and Kreuzer, 2002
; Alberts, 2003
). It follows that it should be of crucial importance to understand the genetic determinants and molecular mechanisms of replication in tumour cells, in order to predict the effectiveness and outcome of cancer therapy. For example, the suggested primary role of the BRCA1/2-defined pathway is to facilitate HR in the bypass of stalled replication forks. As to the question of how cancer arises in cells with deficiencies in BRCA1 or BRCA2, it seems likely that the impaired function of HR is a key step, since this is the only established defect in BRCA2-deficient cells (Powell et al, 2002
; Powell and Kachnic, 2003
). A mutator phenotype can be triggered by a defect in the regulation of HR, either by making the process error-prone or by shunting repair events into a NHEJ pathway that is inherently mutagenic. Additional yet unknown mutational steps to bypass the proliferative block are also necessary so that net growth in the number of cancer cells can occur. It remains unsolved why this described defect in HR leads to the tissue-specific cancers of the breast and ovary. Proliferation in breast or ovarian epithelium may be associated with higher levels of endogenous DNA damage, relative to other tissues, which leads to replication stalling and requires HR. It is likely that defective HR in BRCA1/2-mutant tumour cells also underlies their hypersensitivity to IR and DNA crosslinking agents, which may ultimately impact on the chances of combination cancer therapy to achieve tumour eradication. However, to date there is no consistent evidence that BRCA1/2-mutant tumours carry a higher likelihood of radiocurability, but this is in large part due to the fact that residual tumour burden is the critical determinant of local control. Disentangling the complex relationship between tumour cell replication, genomic instability, cellular hypersensitivity to DNA-damaging agents, and clinical tumour control will be an area of great research interest in the years to come.