A fine-scale dissection of the DNA double-strand break repair machinery and its implications for breast cancer therapy

Abstract

DNA-damage response machinery is crucial to maintain the genomic integrity of cells, by enabling effective repair of even highly lethal lesions such as DNA double-strand breaks (DSBs). Defects in specific genes acquired through mutations, copy-number alterations or epigenetic changes can alter the balance of these pathways, triggering cancerous potential in cells. Selective killing of cancer cells by sensitizing them to further DNA damage, especially by induction of DSBs, therefore requires careful modulation of DSB-repair pathways. Here, we review the latest knowledge on the two DSB-repair pathways, homologous recombination and non-homologous end joining in human, describing in detail the functions of their components and the key mechanisms contributing to the repair. Such an in-depth characterization of these pathways enables a more mechanistic understanding of how cells respond to therapies, and suggests molecules and processes that can be explored as potential therapeutic targets. One such avenue that has shown immense promise is via the exploitation of synthetic lethal relationships, for which the BRCA1-PARP1 relationship is particularly notable. Here, we describe how this relationship functions and the manner in which cancer cells acquire therapy resistance by restoring their DSB repair potential.

Figure 1.

Recruitment of early homologous recombination (HR) factors to double-strand breaks (DSBs). Proteins represented in different colours are recruited at different times: a) The MRN (MRE11–RAD50–NBS1) complex recognizes and binds to DSBs, which then recruits ATM and TIP60. b) Activated ATM phosphorates H2AX, leading to the formation of γH2AX that provides binding sites for MDC1. c) Next, two ubiquitin ligases RNF8 and RNF168 are recruited to catalyse polyubiquitination of γH2AX. This ubiquitination event is tightly controlled by various positive and negative regulators. d) Subsequently, BRCA1 (in the form of BRCA1-A complex) and 53BP1 are recruited; these two proteins play important roles in the balance between HR and NHEJ, wherein a variety of regulatory mechanisms are involved.

Figure 2.

A two-step model for the double-strand break (DSB) end resection. Proteins represented in different colours are recruited at different stages. a) The first step, 'initial resection', is carried out by the endonuclease activity of the MRN (MRE11–AD50–NBS1) complex and promoted by CtIP. Multiple regulatory mechanisms, especially the cell cycle-dependent regulation are involved. b) The second step, long-range resection, is performed by EXO1 or BLM in concert with DNA2. It remains unclear whether EXO1 and BLM work in parallel or interact.

Figure 3.

D loop formation and DNA repair synthesis. Proteins represented in different colours are recruited at different stages. a) The 3′ ssDNA overhang generated by DSB end resection is coated and stabilized by RPA, which is then displaced by RAD51 with the help of recombination mediators that promote both the formation and stability of RAD51-ssDNA filament. The balancing act of proteins involved in stability and dismantling of RAD51 filaments is depicted here as discussed in the text. Rad51 presynaptic filament performs homology searches with help of other proteins and invades nearby homologous duplex DNA template, resulting in the formation of the D loop structure. b) The invading strand is then elongated by copying missing genetic information from the template molecule, which involves the participation of several redundant DNA polymerases.

Figure 4.

The SDSA (synthesis-dependent strand annealing) and DSB repair sub-pathways. D loop formation and DNA repair synthesis can follow two different routes namely SDSA and DSBR to complete homologous recombination. In SDSA invading strand is displaced from D-loop and annealed with complementary strand associated with second end of the DSB. SDSA is preferred over DSBR during mitosis, and mainly results in a non-crossover product. In the DSBR pathway, the other end of the DSB is captured and double Holliday Junctions (dHJs) intermediate is formed which is then resolved to produce cross-over (mainly during meiosis) or non-crossover products.

Figure 5.

The single-strand annealing (SSA) sub-pathway of homologous recombination. This is a Rad51-independent sub-pathway of HR, which operates when there are regions of homology or direct repeats at both sides of the DSB. a) SSA is initiated by RAD52 that binds the 3′ ssDNA ends generated by DSB end resection. RAD52 then functions in concert with RPA to facilitate strand annealing between the two direct repeats. b) Next, the XPF–ERCC1 heterodimers remove the non-homologous 3′ single-stranded flaps between the two repeats. c) The two DSB ends are re-joined by DNA ligase III. d) The sequence continuity is restored.

Figure 6.

The canonical NHEJ (C-NHEJ). Proteins represented in different colours are recruited at different stages. a) The C-NHEJ pathway is initiated by the Ku70–Ku80 heterodimer. b) The Ku70–Ku80 dimer then recruits the DNA-PKcs kinase. c) In many instances ends of the breaks are not amenable to direct ligation and must be resected or filled in prior to ligation by end processing. d) The synthesis step is catalysed by DNA polymerase μ and λ. e) The gap after DNA repair synthesis is ligated by the XRCC4–LIG4–XLF complex. f) The sequence continuity is restored.

Figure 7.

The Alternative NHEJ (A-NHEJ). Proteins represented in different colours are recruited at different stages. In A-NHEJ, a) the broken ends are detected and bound by PARP1. b) This is followed by end-processing by MRN, CtIP and BRCA1, which is prohibited by 53BP1. c) The ligation step can be performed by either LIG3 in concert with XRCC1, or LIG1. d) The sequence continuity is restored.

Figure 8.

Strategy for synthetic lethality based cancer therapy: targeted inhibition of DNA-damage repair pathways in defined cancer cell populations to selectively kill cancer cells.

Figure 9.

Alternative model (198) centred on the unrestricted error-prone NHEJ as a cause of death in tumour cells. HR-deficient cells were found to be hypersensitive to PARP1 inhibition, but this effect was reversed by disabling C-NHEJ, verified through knockdown of Ku80 and Artemis. This suggests that C-NHEJ contributes to the toxicity of PARP1 inhibitors in HR-deficient cells, and therefore an active C-NHEJ is necessary for PARP inhibitor-based synthetic lethality.