Poly(ADP-ribose) polymerase (PARP1) facilitates DNA repair by binding to DNA breaks and attracting DNA repair proteins to the site of damage. Nevertheless, PARP1-/- mice are viable, fertile and do not develop early onset tumours. Here, we show that PARP inhibitors trigger gamma-H2AX and RAD51 foci formation. We propose that, in the absence of PARP1, spontaneous single-strand breaks collapse replication forks and trigger homologous recombination for repair. Furthermore, we show that BRCA2-deficient cells, as a result of their deficiency in homologous recombination, are acutely sensitive to PARP inhibitors, presumably because resultant collapsed replication forks are no longer repaired. Thus, PARP1 activity is essential in homologous recombination-deficient BRCA2 mutant cells. We exploit this requirement in order to kill BRCA2-deficient tumours by PARP inhibition alone. Treatment with PARP inhibitors is likely to be highly tumour specific, because only the tumours (which are BRCA2-/-) in BRCA2+/- patients are defective in homologous recombination. The use of an inhibitor of a DNA repair enzyme alone to selectively kill a tumour, in the absence of an exogenous DNA-damaging agent, represents a new concept in cancer treatment.

Recent studies have indicated the existence of tumorigenesis barriers that slow or inhibit the progression of preneoplastic lesions to neoplasia. One such barrier involves DNA replication stress, which leads to activation of the DNA damage checkpoint and thereby to apoptosis or cell cycle arrest, whereas a second barrier is mediated by oncogene-induced senescence. The relationship between these two barriers, if any, has not been elucidated. Here we show that oncogene-induced senescence is associated with signs of DNA replication stress, including prematurely terminated DNA replication forks and DNA double-strand breaks. Inhibiting the DNA double-strand break response kinase ataxia telangiectasia mutated (ATM) suppressed the induction of senescence and in a mouse model led to increased tumour size and invasiveness. Analysis of human precancerous lesions further indicated that DNA damage and senescence markers cosegregate closely. Thus, senescence in human preneoplastic lesions is a manifestation of oncogene-induced DNA replication stress and, together with apoptosis, provides a barrier to malignant progression.

DNA repair pathways can enable tumour cells to survive DNA damage that is induced by chemotherapeutic treatments; therefore, inhibitors of specific DNA repair pathways might prove efficacious when used in combination with DNA-damaging chemotherapeutic drugs. In addition, alterations in DNA repair pathways that arise during tumour development can make some cancer cells reliant on a reduced set of DNA repair pathways for survival. There is evidence that drugs that inhibit one of these pathways in such tumours could prove useful as single-agent therapies, with the potential advantage that this approach could be selective for tumour cells and have fewer side effects.

SummaryFaithful DNA replication is essential to all life. Hydroxyurea (HU) depletes the cells of dNTPs, which initially results in stalled replication forks that, after prolonged treatment, collapse into DSBs. Here, we report that stalled replication forks are efficiently restarted in a RAD51-dependent process that does not trigger homologous recombination (HR). The XRCC3 protein, which is required for RAD51 foci formation, is also required for replication restart of HU-stalled forks, suggesting that RAD51-mediated strand invasion supports fork restart. In contrast, replication forks collapsed by prolonged replication blocks do not restart, and global replication is rescued by new origin firing. We find that RAD51-dependent HR is triggered for repair of collapsed replication forks, without apparent restart. In conclusion, our data suggest that restart of stalled replication forks and HR repair of collapsed replication forks require two distinct RAD51-mediated pathways.

Nature 414, 165-166 (2001) In Fig. 1b of this Brief Communication, the legend and text did not make it clear that two different groups of ten sheep were used in the study to give overall n 5 20. A reanalysis of the data using a post-hoc Tukey test (rather than a paired t-test, as originally stated) revealed some small errors that altered the significance values slightly; however, there is no overall change in the results. The maximum retest interval was 801 rather than 800 days, and 100-500 trials were conducted for 1-6 weeks (not 400-500 trials for 4-6 weeks, as published). A revised version of Fig. 1b showing the corrected statistical changes and an expanded legend incorporating further methodological detail are available as Supplementary Information to this Corrigendum.Supplementary Information is linked to the online version of this Corrigendum at www.nature.com/nature.

Cancers have dysfunctional redox regulation resulting in reactive oxygen species production, damaging both DNA and free dNTPs. The MTH1 protein sanitizes oxidized dNTP pools to prevent incorporation of damaged bases during DNA replication. Although MTH1 is non-essential in normal cells, we show that cancer cells require MTH1 activity to avoid incorporation of oxidized dNTPs, resulting in DNA damage and cell death. We validate MTH1 as an anticancer target in vivo and describe small molecules TH287 and TH588 as first-in-class nudix hydrolase family inhibitors that potently and selectively engage and inhibit the MTH1 protein in cells. Protein co-crystal structures demonstrate that the inhibitors bind in the active site of MTH1. The inhibitors cause incorporation of oxidized dNTPs in cancer cells, leading to DNA damage, cytotoxicity and therapeutic responses in patient-derived mouse xenografts. This study exemplifies the non-oncogene addiction concept for anticancer treatment and validates MTH1 as being cancer phenotypic lethal.

The collective somatic mutations observed in a cancer are the outcome of multiple mutagenic processes that have been operative over the lifetime of a patient. Each process leaves a characteristic imprint--a mutational signature--on the cancer genome, which is defined by the type of DNA damage and DNA repair processes that result in base substitutions, insertions and deletions or structural variations. With the advent of whole-genome sequencing, researchers are identifying an increasing array of these signatures. Mutational signatures can be used as a physiological readout of the biological history of a cancer and also have potential use for discerning ongoing mutational processes from historical ones, thus possibly revealing new targets for anticancer therapies.

Cancer chromosomal instability (CIN) results in an elevated rate of change of chromosome number and structure and generates intratumour heterogeneity1,2. CIN is observed in the majority of solid tumours and is associated with both poor prognosis and drug resistance3,4. Therefore, understanding a mechanistic basis for CIN is paramount. Here we find evidence for impaired replication fork progression and elevated DNA replication stress in CIN+ colorectal cancer (CRC) cells relative to CIN− CRC cells, with structural chromosome abnormalities precipitating chromosome missegregation in mitosis. We identify three novel CIN-suppressor genes (PIGN (MCD4), RKHD2 (MEX3C) and ZNF516 (KIAA0222)) encoded on chromosome 18q, which is subject to frequent copy number loss in CIN+ CRC. 18q loss was temporally associated with aneuploidy onset at the adenoma-carcinoma transition. CIN-suppressor gene silencing leads to DNA replication stress, structural chromosome abnormalities and chromosome missegregation. Supplementing cells with nucleosides, to alleviate replication-associated damage5, reduces the frequency of chromosome segregation errors following CIN-suppressor gene silencing and attenuates segregation errors and DNA damage in CIN+ cells. These data implicate a central role for replication stress in the generation of structural and numerical CIN, which may inform new therapeutic approaches to limit intratumour heterogeneity.