dsd

REVIEWSG E N O M I C I N S TA B I L I T Y I N C A N C E R

DNA repair dysregulation from cancer driver to therapeutic targetNicola J. Curtin

Abstract | Dysregulation of DNA damage repair and signalling to cell cycle checkpoints, known as the DNA damage response (DDR), is associated with a predisposition to cancer and affects responses to DNA-damaging anticancer therapy. Dysfunction of one DNA repair pathway may be compensated for by the function of another compensatory DDR pathway, which may be increased and contribute to resistance to DNA-damaging chemotherapy and radiotherapy. Therefore, DDR pathways make an ideal target for therapeutic intervention; first, to prevent or reverse therapy resistance; and second, using a synthetic lethal approach to specifically kill cancer cells that are dependent on a compensatory DNA repair pathway for survival in the context of cancer-associated oxidative and replicative stress. These hypotheses are currently being tested in the laboratory and are being translated into clinical studiesA finely coordinated response of DNA repair and activation of cell cycle checkpoints has evolved to cope with everyday endogenous and environmental DNA damage (FIG.1). This DNA damage response (DDR) generally protects against genomic instability, which is an enabling characteristic for cancer development 1. DNA damage and replication stress is generally greater in cells in precancerous lesions, as well as in tumours, compared with those in normal tissues, suggesting that replication stress also contributes to genomic instability 2,3. Germline polymorphisms or mutations in DDR genes may predispose to cancer development, and somatic mutation and epigenetic silencing of DDR genes is common in cancer (TABLE1) and so these genes can be considered a subset of tumour suppressor genes. Loss of some elements of one DNA repair pathway may be compensated for by the increased activity of other elements or pathways. This presents both challenges and opportunities for cancer treatment. Upregulated DNA repair pathways can cause resistance to DNA-damaging chemotherapy and radiotherapy and so inhibitors of these pathways have the potential to sensitize cells to these therapies. This has been the primary rationale for the development of inhibitors of members in the DDR (TABLE2). Conversely, pathways that are lost represent vulnerabilities in the DNA repair capacity of the cancer cell that can be exploited by selecting the appropriate chemotherapy to induce unrepairable and hence more cytotoxic DNA damage. Indeed, the success of different chemotherapy regimens in some tumours but not in others, although arrived at empirically, may reflect the relative frequency of a particular DDR defect in that tumour. Therefore, DDR inhibitors have the potential to expand the range of tumour types that can be targeted with conventionaldrugs. Furthermore, tumour cells may be addicted to compensatory DDR pathways, upregulated or otherwise, for survival. The feasibility of exploiting this dependence, using the principle of synthetic lethality has been applied preclinically and now clinically with both checkpoint inhibitors and DNA repair inhibitors. Only a handful of classes of inhibitors have so far been evaluated as single agents but some exciting data are beginning to emerge. Inhibition of DNA repair is a new paradigm in cancer therapy, and there is heightened interest in the therapeutic potential of these inhibitors to selectively target tumours with minimal host toxicity. We are now entering a new era of cancer research in which patients may be stratified for appropriate therapy on the basis of the DDR status of their tumour, rather than on the tissue of origin. One challenge facing us is the accurate identification of exploitable defects. This Review describes the association of DDR defects with cancer, the use of inhibitors in combination with chemotherapy and radiotherapy, and as single agents, and the success, or otherwise, of translating these studies to theclinic.VOLUME 12 | DECEMBER 2012 | 801 2012 Macmillan Publishers Limited. All rights reserved

Cell cycle checkpointsPoints in the cell division cycle at which the cell may arrest in response to stress; for example, following DNA damage to prevent damage becoming inherited by daughter cells. The G1 checkpoint controls entry intoS phase, the S phase checkpoint halts the progression of S phase and theG2 checkpoint controls entry into mitosis.

Replication stressThe harmful effect of partially replicated DNA. It can be caused by oncogene-induced hyper-replication that activates origins more than once per Sphase, by nucleotide pool imbalance or by DNA damage; for example, by reactive oxygen species.

Newcastle University, Northern Institute for Cancer Research, Newcastle upon Tyne NE2 4HH, UK. e-mail: [email protected] doi:10.1038/nrc3399

NATURE REVIEWS | CANCER

REVIEWSAt a glance summaryTheDNAdamageresponse(DDR)coordinatestherepairofDNAandtheactivation ofcellcyclecheckpointstoarrestthecelltoallowtimeforrepair. DNAissubjecttoahighlevelofendogenousdamageandtheDDRisessentialforthe maintenanceofgenomicstabilityandsurvival. DysregulationoftheDDRcanleadtogenomicinstabilitythatpromotescancer developmentbutthatisexploitablewithbothconventionalcytotoxictherapyand DDRinhibitors.DownregulatedDDRpathwaysrenderthetumoursensitiveto specificcytotoxicsandsomeDDRinhibitors.UpregulatedDDRpathwaysconfer therapeuticresistance. InhibitorsoftheDDRhavebeendevelopedtoovercomeresistanceandtoaugment theactivityofconventionaltherapy. LossofaDDRpathwaycanleadtodependenceonacompensatorypathway,and targetingthissecondpathwaymayrenderendogenousDNAdamagecytotoxicbya processtermedsyntheticlethality,whichwillbetumour-specificbecausethenormal tissuesintheanimal(orperson)willhavefunctionalDNArepair. DespitepromisingpreclinicaldatacombiningDDRinhibitorswithconventional cytotoxicagents,thesecombinationshavebeenlesssuccessfulintheclinicandare oftenassociatedwithtoxicity.ExploitationofDDRdefectsbysyntheticlethalityisa morepromisingapproach.Clinicaldataontheuseofpoly(ADP-ribose)polymerase (PARP)inhibitorsinhomologousrecombinationrepair(HRR)-defectivetumoursare encouraging. RobustandvalidatedbiomarkerstoidentifyDDRdefectsthatareexploitableby bothconventionalcytotoxictherapyandagentstargetingtheDDRareneededto effectivelystratifypatients.

Synthetic lethalityCell death caused by inactivation (mutation or inhibition) of two genes or theirproducts or two pathways, when inactivation ofeither alone is not lethal.

AlkylationsTransfers of an alkyl group fromone molecule to another;for example, the transfer of a methyl group fromtemozolomide to guanine.

MyelosuppressionImpairment of bone marrow function resulting in reduced numbers of blood cells (red orwhite) or platelets.

PhaseIII trialsTest the efficacyof the novel drug, orcombination, and side effects compared with the standard-of-care for that tumour type, with patients randomized to two groups. Numbers of patients are larger (1,0003,000) than other Phase trials.

Therapeutic targeting of DNA damage repair Direct repair. The simplest form of DNA repair is the direct reversal of the lesion. O6-methylguanine DNA methyltransferase (MGMT) demethylates O6-methylguanine lesions, which are formed as a result of erroneous methylation by S-adenosylmethionine (SAM) and other alkylations at the O6-position of guanine that are induced by dietary nitrosamines or chemotherapy agents for example, temozolomide (TMZ), dacarbazine (DTIC) and nitrosoureas (such as, BCNU (also known as carmustine))4,5 (FIG. 1). Unrepaired O6-methylguanine is mutagenic because distorted pairing with cytosine or thymidine leads to G:C to A:T transitions on replication6. MGMT is ubiquitously expressed but levels vary in different tissues, and high MGMT expression in tumour cells is associated with resistance to BCNU and TMZ7,8 (TABLE1). MGMT is both a transferase and an acceptor, the reaction is stoichiometric and the resulting conformational change in the protein targets it for degradation. The higher levels of MGMT that are frequently observed in tumour tissue compared with normal tissue suggest that its depletion with pseudo-substrates that resemble O6-methylguanine might be a viable strategy to sensitize tumour cells to O6 alkylating agents. One such pseudo-substrate, O6-benzylguanine, developed in the 1990s, depleted MGMT and substantially increased nitrosourea and TMZ cytotoxicity preclinically but also increased bone marrow toxicity (reviewed in REF.9). The first clinical trial with O6-benzylguanine was reported in 1998 (REF.10), O6-benzylguanine was non-toxic at doses that depleted MGMT activity both in lymphocytes and in tumour tissue, including gliomas, confirming its ability to cross the bloodbrain barrier. However,

O6-benzylguanine enhanced TMZ- and BCNU-induced myelosuppression9. In general, these trials have shown only marginal clinical benefit, which has been limited by toxicities, presumably because O6-benzylguanine depleted MGMT in normal, as well as in tumour, tissues rendering normal tissues more sensitive to the companion cytotoxic, and the drug has not entered PhaseIII trials (TABLE2). Several other pseudo-substrates have been developed, but only lomeguatrib (also known as O6-(4-bromothenyl) guanine and PaTrin-2) has shown sufficiently promising preclinical activity to justify clinical evaluation. Lomeguatrib inhibited MGMT in lymphocytes and tumour biopsy samples, reduced the maximum tolerated dose (MTD) of TMZ by 25% and resulted in some initial responses11. However, subsequent PhaseII trials in combination with TMZ failed to show any substantial activity in patients with melanoma or colon cancer, although this may have been confounded by the high frequency of defects in the mismatch repair (MMR) pathway, which causes resistance to TMZ, in colon cancer or incomplete MGMT inhibition12 (TABLE2). A more promising approach may be the exploitation of reduced MGMT activity owing to epigenetic silencing in some cancers13. MGMT promoter methylation correlated with sensitivity to BCNU in patients with astrocytomas and also correlated with sensitivity to TMZ plus radiotherapy (the current standard-of-care) in patients with glioma14. Therefore, MGMT promoter methylation could be useful for stratifying patients for treatment with TMZ, with or without radiotherapy. Base excision repair. Single-strand breaks (SSBs) are the most common endogenous lesions, arising both directly from DNA sugar damage that is induced by reactive oxygen species (ROS) and indirectly from base excision repair (BER)-mediated enzymatic excision of damaged bases following their spontaneous deamination, oxidation (by ROS) or alkylation (by SAM)15 (FIG.1). ROS are produced endogenously as by-products of oxidative phosphorylation. The most common base oxidations are 8-oxoguanine (8-oxoG) and 5-hydroxycytosine, which mispair with adenine and thymine, respectively 15,16. Oxidative damage is generally increased in tumours: increased metabolism, oncogenic signalling and mitochondrial dysfunction result in 100-fold more 8-oxoG in cancer tissues than in normal tissues17. Inflammation promotes carcinogenesis and generates ROS in the tumour cell and its microenvironment 17. There is also a high level of spontaneous deamination of DNA bases, which causes mispairing and is thus potentially mutagenic if not rapidly and efficiently repaired15. Single-strand break repair (SSBR) and BER are often assumed to be synonymous because they involve the same components and are similar after the initial recognition step. Damaged bases are first removed by BER glycosylases toform abasic sites (also known as apurinic or apyrimidinic (AP) sites) and BER endonucleases then generate a nick (that is, an SSB), which, along with directly induced SSBs and those generated by topoisomerase I poisons, are the substrates for SSBR. Inwww.nature.com/reviews/cancer

802 | DECEMBER 2012 | VOLUME 12 2012 Macmillan Publishers Limited. All rights reserved

REVIEWS SAM Nitrosated amines and bile acids Dietary nitrosamines 1030 O6-methylguanine ROS SAM Natural IR Base deamination or loss Trapped TOPOI ROS UV Tobacco smoke Aatoxin Replication errors SAM Base deamination

Endogenous or environmental

ROS Natural IR Trapped TOPOII

Unrepaired single strand lesions

Acrolein Crotonaldehyde

Lesion

10,000100,000 8-oxoguanine N7-meG N3-meA Uracil Hypoxanthine Xanthine SSB

1050 64 photoproducts Cyclopurines Bulky adducts Base mismatches DNA doublestrand breaks Stalled replication forks

ICL

Therapeutic

TMZ Alkylating agents Nitrosoureas

TMZ IR Radiomimetics TOPOI poisons Antimetabolites BER SSBR

Cisplatin Carboplatin Nitrosoureas

TMZ Nucleoside analogues

IR Radiomimetics TOPOI poisons

TMZ TOPOI poisons Antimetabolites

Cisplatin Carboplatin Nitrosoureas MMC

Repair pathway

Direct repair

NER

MMR

NHEJ

HRR

ICL repair

Figure 1 | Sources of DNA damage and their repair. Endogenous and environmental sources of DNA damage Nature Reviews | Cancer are shown in green boxes, with the lesions they cause in beige boxes (where known, the approximate number of the indicated type of lesion that occurs naturally in a cell each day is shown). Therapeutic DNA-damaging agents that cause the corresponding DNA lesion are shown in orange boxes. DNA repair pathways (blue boxes) repair DNA damage that is induced by endogenous and environmental DNA-damaging agents and thus protect the genome but they antagonize the efficacy oftherapeutic DNA-damaging agents (except for mismatch repair (MMR)). BER, base excision repair; HRR, homologous recombination repair; ICL, interstrand crosslink; IR, ionizing radiation; MMC, mitomycin C; NER, nucleotide excision repair; NHEJ, non-homologous end joining; ROS, reactive oxygen species; SAM, S-adenosyl methionine; SSB, single-strand break; SSBR, SSB repair; TMZ, temozolomide; TOPO, topoisomerase; UV, ultraviolet.

PhaseII trialsUse the maximum tolerated dose to treat groups of patients with a particular tumour type, these studies determine whether the drug, or drug combination, is effective in one or more tumour types and are used to further monitor drug safety. These generally involve small numbers of patients (100300).

DeaminationRemoval of an amide group; for example, deamination ofadenosine to inosine (adenine to hypoxanthine).

Oxidative phosphorylationMetabolic pathway for the generation of ATP occurring in the mitochondria of eukaryotes in which NADH and succinate from the Krebs cycle are oxidized by the electron transport chain. The process also generates reactive oxygen species.

practice, the entire pathway is usually referred to as BER. The pathway is subdivided into short patch BER (single nucleotide replacement; the predominant pathway) or long patch BER (two to 13 nucleotides are replaced), depending on the nature of the 5 and 3 ends and, possibly, ATP availability 18. The main components of the pathway are glycosylases, endonucleases, DNA polymerases and DNA ligases, with poly(ADP-ribose) polymerase 1 (PARP1) and PARP2 facilitating the process (FIG.2). PARP1 has a major role in BER, but it is also important in DNA double-strand break (DSB) repair (reviewed in REF.19). The binding of PARP1 to DNA breaks activates it, and this is necessary for the recruitment of XRCC1 and other BER proteins in the pathway 20. BER repairs DNA damage that is therapeutically induced by ionizing radiation (IR), DNA methylating agents, topoisomerase I poisons21 (for example, camptothecin, irinotecan and topotecan) and some antimetabolites. Germline and tumour-specific polymorphisms and mutations in BER genes are associated with cancer 22 (reviewed in REF.23) (TABLE1). The BER pathway is an attractive target for the modulation of chemosensitivity and radiosensitivity. Early inhibitors of DNApolymerase- (Pol ), flap endonuclease 1 (FEN1), ligase 1 and ligase 3 enhanced sensitivity to IR, TMZ and the alkylating agent methyl methanesulphonate (MMS), respectively (TABLE 2) . However, the most advanced drugs that target this

pathway are AP endonuclease 1 (APE1; also known as APEX1) inhibitors and PARP inhibitors. Both APE1 and PARP expression and/or activity are generally higher in tumours2426, which possibly reflects higher levels of endogenous DNA damage or DNA repair defects that are compensated for by higher BER activity. These increased levels cause chemotherapy and radiotherapy resistance, and inactivation of APE1 or PARP increases sensitivity to IR and alkylating agents (which cause DNA lesions that are substrates for BER) in the laboratory setting 24,27. There are two classes of APE1 inhibitor: methoxyamine, which binds the AP site in DNA (the APE1 substrate), and inhibitors of APE1 endonuclease activity. Preclinically, methoxyamine potentiates the cytotoxicity of TMZ28 and pemetrexed (an antifolate that causes base errors). In a PhaseI trial of methoxyamine, responses were seen in combination with pemetrexed, and there is a study currently ongoing with TMZ (TABLE2). Lucanthone, a topoisomerase II inhibitor, also inhibits APE1 endonuclease activity and potentiates the cytotoxicity of DNA-methylating agents in breast cancer cells29 and also potentiates the antitumour activity of radiotherapy in patients with brain metastases30. Novel, more specific, APE1 endonuclease inhibitors increased the persistence of AP sites invitro and increased the cytotoxicity of alkylating agents, but they have not yet moved into advanced preclinical or clinical evaluation31.VOLUME 12 | DECEMBER 2012 | 803

NATURE REVIEWS | CANCER 2012 Macmillan Publishers Limited. All rights reserved

REVIEWSTable 1 | Common DDR aberrations in cancerPathwayDirect repair BER

ProteinsMGMT OGG1

Importance in cancerHigher levels in tumours than in normal tissue confers resistance to DNA-alkylating agents. Methylation of the MGMT promoter is associated with better response to BCNU and TMZ in brain tumours Truncating mutations and R46Q variant (which has reduced activity) observed in renal cancers. OGG1 methylation is seen in various cancers and loss of expression is associated with poor prognosis in breast cancer. The polymorphism OGG1-S326C is associated with reduced activity and increased risk of developing lung cancer High APE1 expression in several tumour types is associated with drug and radiotherapy resistance The polymorphism XRCC1-R194W increased the efficiency of BER and is protective against cancer. The polymorphism XRCC1-R399Q reduced the efficiency of BER and predisposes to cancer Higher levels of expression occur in tumours. The polymorphism PARP1-V762A confers reduced activity and predisposes to various cancers 30% of tumours have Pol mutations that are not found in normal tissue, including frameshifts, del208-236, and K289M and I260M dominant-negative, transforming mutations XP (which is caused by defective global NER) is characterized by UV radiation sensitivity and skin cancer. XPC methylation occurs in bladder cancer. SNPs in XPA and XPC have been associated with lung and bladder cancer, and SNPs in XPG have been associated with sensitivity to chemotherapy ERCC1 polymorphisms are associated with skin and lung cancer. ERCC1 methylation occurs in glioma Aberrant expression observed in several tumour types MMR defects cause Lynch syndrome and HNPCC, which are associated with colorectal, stomach, ovarian and endometrial cancers. MLH1 promoter methylation is associated with spontaneous tumours in these tissues. MMR defects confer resistance to the DNA-methylating agents 6-thioguanine and cisplatin Nijmegen breakage syndrome (which is caused by defective NBS1) is characterized by chromosome instability, immunodeficiency, ionizing radiation sensitivity and cancer predisposition, especially lymphoma; heterozygous mutants are also cancer prone Point mutations are observed in ovarian cancers and shortening of the T(11) repeat microsatellite occurs in 93% of primary colorectal cancer probably as a result of MMR-induced microsatellite instability SNPs associated with breast cancer, and epigenetic silencing is associated with breast, colorectal and lungcancer Epigenetic silencing is associated with lung cancer

Refs14,150, 151 22,23

APE1 XRCC1 PARP1 Pol Global NER XP proteins

24 23,152 152,153 23 42,154

ERCC1 TLS MMR Pol H and Pol Q MSH2 and MLH1 NBS1

42 45 48,51

DSB repair

155,156

MRE11 NHEJ KU70 KU80

50, 157159 160 161 159, 162,163 128,164, 165 166 167 69,168, 169

DNA-PKcs, SNPs associated with glioma. SNPs may protect against breast and lung cancer. MiR-101 (which is induced by ligase 4 and NMYC) targets DNA-PKcs and thus reduces its expression XRCC4 HRR BRCA1 and BRCA2 XRCC2 RAD50 FANC proteins BRCA1 and BRCA2 mutation carriers have increased risk of breast, ovarian, prostate, pancreatic, melanoma and other gastrointestinal, gynaecological and haematological malignancies. Methylation of the BRCA1 promoter is common in spontaneous breast, ovarian and lung cancers XRCC2 is a RAD51 paralogue, frameshift mutation owing to microsatellite slippage that in MSI tumours confers sensitivity to crosslinking agents Frameshift mutations in the RAD50-associated microsatellite, which results in a truncated protein, occur in 31% of gastrointestinal cancers Fanconis anaemia is associated with haematological malignancies, especially MDS and AML, HNSCC, and oesophageal and gynaecological cancer. Most mutations occur in FANCA (65%), FANCC (15%) or FANCG (10%). FANCD1 (BRCA2), FANCN (PALB2) and FANCJ (BACH1; also known as BRIP1) are breast cancer susceptibility genes. Methylation of FANC genes is common in sporadic cancers; for example, FANCF is methylated in lung, ovarian and cervical cancer Ataxia telangiectasia (which is caused by defects in ATM) is associated with radiosensitivity and 100-fold increased cancer predisposition. Heterozygous germline mutations in ATM are associated with leukaemia and breast and pancreatic cancer. Epigenetic silencing of ATM and ATM polymorphisms are associated with breast, lung and colorectal cancer. MiR-421 and miR-101, which are induced by NMYC, both target ATM. CHK2 is a candidate tumour suppressor gene, inactivation is observed in multiple humantumours Frameshift mutations caused by deletions in the A(10) microsatellite as a consequence of MMR defects are associated with leukaemia, lymphoma and stomach and endometrial cancer

Cell cycle ATM and checkpoints CHK2

163,170, 171,172

ATR

79

AML, acute myeloid leukaemia; APE1, AP endonuclease 1; ATM, ataxia-telangiectasia mutated; ATR, ataxia-telangiectasia and Rad3-related; BACH1, BRCA1-associated C-terminal helicase 1; BER, base excision repair; BRIP1, BRCA1-interacting protein C-terminal helicase 1; DDR, DNA damage response; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; DSB, DNA double-strand break; FANC, Fanconi anaemia group protein; HNPCC, hereditary nonpolyposis colorectal cancer; HNSCC,head and neck squamous cell carcinoma; HRR, homologous recombination repair; MGMT, O6-methylgyuanine DNA methyltransferase; MDS, myelodysplastic syndrome; miR, microRNA; MMR, mismatch repair; MSI, microsatellite instability; NBS1, Nijmegen breakage syndrome protein 1; NER, nucleotide excision repair; NHEJ,non-homologous end joining; OGG1, 8-oxoguanine DNA glycosylase; PARP1, poly(ADP-ribose) polymerase 1; Pol, DNA polymerase; SNP, single nucleotide polymorphism; TLS, translesion synthesis; TMZ, temozolomide; UV, ultraviolet; XP, Xeroderma pigmentosum.


www.nature.com/reviews/cancer

REVIEWSIn preclinical studies, PARP inhibitors (PARPis) or genetic inactivation of PARP1 potentiated the cytotoxicity and antitumour activity of TMZ, topoisomerase I poisons and IR (reviewed in REF.27). There are subtle differences between the inhibition of PARP and the inhibition of APE1: APE1 inhibitors do not enhance the effects of topoisomerase I poisons and PARPis do not enhance the effects of antifolates32,33. This reflects differences in their substrates and the upstream role of APE1 in BER and the role of PARP after SSB generation. Complete, durable regressions observed when a PARPi was combined with TMZ in xenograft models34 led to the first clinical trial in 2003 of a PARPi (AG014699; also known as CO-338 and rucaparib) in combination with TMZ35. PARP inhibition was demonstrated in patient samples, the combination was tolerated and some clinical responses were observed. Modest clinical responses in a PhaseII trial were accompanied by myelosuppression that required dose reductions of TMZ, as reported in abstract form36. A further eight PARPis have undergone, or are currently undergoing, clinical evaluation and many are in combination with DNA-methylating agents, such as DTIC or TMZ, or the topoisomerase I inhibitors irinotecan and topotecan (TABLE2) combinations that are active in a variety of preclinical studies. Recent evidence suggests that the efficacy of PARPitopoisomeraseI poison combinations may be most effective in tumours that lack ERCC1XPF, which are involved in the nucleotide excision repair (NER) pathway (discussed below), as this represents an alternative repair pathway to overcome the damage induced by topoisomeraseI poisons37. Some of the current PARPi clinical trials use combinations that have not been shown to increase antitumour activity (or have not even been evaluated) preclinically, including combinations with paclitaxel, which is an antitubulin agent (see the ClinicalTrials. gov website; see Further Information). Clinical trials of PARPis have generally been disappointing owing to toxicity, which may be due to using a dose of PARPi that was established as safe when used as a single agent. In general, the preclinical data indicate that the MTD of single agent PARPi is much higher than when it is combined with a relevant cytotoxic agent (such as TMZ; compare the doses used in REFS38,39). This is because almost total inhibition of PARP is needed to render endogenous DNA damage cytotoxic, but this level of inhibition is not necessary to render the additional burden of deliberately introduced DNA damage cytotoxic, both in the tumour and in proliferating normal tissues. The abundant preclinical data demonstrating radiosensitization by PARP inhibition also support a role for this combination in cancer patients, particularly as the toxicities observed with the chemotherapy combinations might be avoided. Such trials are finally underway and interim results (in abstract form) have shown that veliparib was well tolerated with whole-brain radiotherapy in patients with brain metastases40. NER. NER removes helix-distorting adducts on DNA for example, those caused by ultraviolet (UV) radiation and tobacco smoke and contributes to the repair of intrastrand and interstrand crosslinks (ICLs); the xeroderma pigmentosum (XP) proteins and ERCC1 also have crucial roles in both the NER and ICL repair pathways41 (FIGS1,3,4). Hereditary defects in NER cause UV sensitivity and skin cancer development 42 (TABLE1). Deficiency in NER confers sensitivity to platinum agent therapy, which reflects a reduced capacity to repair ICLs43,44, and the measurement of levels of crucial NER enzymes could thus be used for patient stratification. There are currently no small-molecule inhibitors of NER, although cyclosporine and cetuximab might downregulate XPG and ERCC1XPF expression, respectively (TABLE2). Translesion synthesis. If damaged DNA bases or adducts are not repaired they may stall replication forks, which could thus contribute to genomic instability (reviewed in REF.45). Several DNA polymerases can synthesize DNA past DNA lesions. Such translesion synthesis (TLS) contributes to survival but, because these polymerases have no proofreading function, errors can occur and it should, therefore, be considered a DNA damage tolerance mechanism rather than a DNA repair mechanism. Defects in TLS polymerases contribute to carcinogenesis but also confer sensitivity to DNA-damaging agents (TABLE1), and inhibitors of these polymerases are starting to emerge46,47 (TABLE2). MMR. MMR corrects replication errors that cause the incorporation of the wrong nucleotide (a mismatch) and nucleotide deletions and insertions (FIGS1,5). Defective MMR increases mutation rates up to 1,000-fold, results in microsatellite instability (MSI) and is associated with cancer development48 (TABLE1). Several DDR genes have microsatellites for example, MRE11 (also known as MRE11A) and ataxia-telangiectasia mutated (ATM) and could be mutated in MSI-high cancer, potentially conferring sensitivity to some DNA-damaging agents49,50. However, defects in MMR cause tolerance to TMZ, platinum agents and some nucleoside analogues, which leads to drug resistance51,52, it would thus be counterproductive to develop inhibitors of MMR. Instead, research has focused on attempts to reactivate epigenetically silenced MLH1 (an MMR gene). After promising preclinical data that demonstrated chemosensitization53, clinical trials have shown adverse reactions (for example, NCT00748527, which was withdrawn) (TABLE2). Non-homologous end joining (NHEJ). DSBs are among the most difficult lesions to repair and are profoundly cytotoxic. Estimates suggest that about ten to50 DSBs per cell arise endogenously per day, mostly from ROS-induced DNA damage, which is increased in tumour cells54 (FIG.1). NHEJ ligates DSBs with minimal end processing, it is not error-free but it is active in all phases of the cell cycle, predominating in G0 phase and G1 phase55, and it is considered to be responsible for the rapid repair of up to 85% of IR-induced DSBs56,57. Therapeutically induced DSBs result directly from exposure to IR and to topoisomerase II poisons, and indirectly from the collision of replication forks with single-stranded lesions. The topoisomerase II poisons create a persistent protein-associated DSB58, and IR induces approximately one DSB for every 25 SSBs. TheVOLUME 12 | DECEMBER 2012 | 805 2012 Macmillan Publishers Limited. All rights reserved

Abasic sites(Also known as apurinic or apyrimidinic (AP) sites). A site on the DNA where there is no purine or pyrimidine base.

AntimetabolitesSubstances that resemble a normal precursor or cofactor, usually for DNA synthesis, which interfere with the normal metabolic process; for example, nucleoside analogues such as 6-thioguanine or folate analogues such as pemetrexed.

PhaseI trialDetermines pharmacokinetics (absorption, distribution, metabolism and excretion (ADME)) and the safe dose or maximum tolerated dose of a novel agent or combination. They are conducted in 2080 patients with a variety of tumour types who have often undergone treatment with standard-of-care but for whom no effective therapy options are available. These studies usually involve escalating the dose of the novel agent in successive cohorts of patients until unacceptable toxicities are seen.

Microsatellite instability(MSI). Microsatellites are sequences in DNA; for example, ten thymidines in direct sequence. MSI occurs when this sequence is shortened or lengthened owing to defects in the mismatch repair system.

NATURE REVIEWS | CANCER

REVIEWSTable 2 | Inhibitors of the DDR*PathwayDirect repair

TargetMGMT

InhibitorO -benzylguanine6

Current stageThe first clinical trial of O6-benzylguanine in combination with BCNU was reported in 1998 (REF.10). Currently in PhaseII clinical trials with toxicity issues and marginal benefit. Lomeguatrib plus TMZ combinations are in PhaseII trials but no positive data have been reported (dose and tumour type issues?)12 In vitro TMZ sensitization173 In vitro sensitization to IR and bleomycin174,175

and lomeguatrib

BER

FEN1 Pol Ligase 1 and ligase 3 APE1

NSC-281680 Pamoic acid, oleanolic acid and eicosapentaenoic acid L67 and L189 Methoxyamine Lucanthone CRT0044876

In vitro sensitization to MMS and IR176 Responses in combination with pemetrexed seen in PhaseI trial. PhaseII combinations with pemetrexed and with TMZ are underway177 PhaseI trial of combination with TMZ induced radiosensitization (also undergoing testing with a topoisomerase II poison)30 Preclinical evidence of TMZ sensitization31,178 The first PARPi in clinical trial. PhaseI and II trials with TMZ in patients with melanoma showed profound PARP inhibition and some clinical responses but increased myelosuppression in PhaseII trials, single agent PhaseII trials in patients with breast and ovarian cancer with BRCA mutations35,36. PhaseI trials with various cytotoxic combinations are underway Good single agent activity (40% response rate) demonstrated in BRCA mutation-associated ovarian cancer and 41% in breast cancer and 24% in patients with ovarian cancer without BRCA mutations at the MTD103. PhaseII single agent trial in selected patients104,179. Several PhaseI and II trials with a variety of drug combinations; however, there have been toxicity issues with topotecan combination and little benefit was observed with dacarbazine combination180,181 Phase 0 trial to determine active dose, PhaseI/II single agent and combination trials are ongoing in various solid and lymphoid tumours. Combinations with topotecan and cyclophosphamide were tolerated and evidence of inhibition of PARP and DNA repair was obtained142,143 PhaseII trial of TMZ combinations in melanoma and glioblastoma PhaseI trial of single agent in BRCA mutation-associated ovarian cancer PhaseI trial of TMZ combination in solid tumours PhaseI/II TMZ combination in solid tumours and melanoma PhaseI trial single agent in solid and haematological malignancies PhaseII/III trial of combination with gemcitabine and carboplatin. This is no longer considered a PARP inhibitor182 Preliminary invitro studies183 Preliminary invitro studies184 Preliminary invitro studies, aurintricarboxylic acid and ellagic acid displayed potent nanomolar activity46,47 Preclinical sensitization to cisplatin, TMZ and epirubicin53. PhaseII trial was terminated because of adverse reaction. Clinical trials are ongoing in combination with carboplatin and with TMZ Preclinical invitro and invivo enhancement of responses to IR and etoposide, exvivo sensitization of patient-derived CLL cells to mitoxantrone6165 Preclinical inhibition of DNA repair in radioresistant L5178Y cells185 Preclinical invitro inhibition of DSB repair and radiosensitization186 Dual DNA-PK and mTOR inhibitor in PhaseI clinical trial In vitro radiosensitization70,71 In vitro identification of RAD51 inhibition by a high-throughput screen of NIH compound library and inhibition of plasmid rejoining by HRR187. Imatinib inhibits RAD51 phosphorylation, DNA damage-induced RAD51 focus formation and sensitized cells to chlorambucil, MMC and IR72,73 CDK1 activates BRCA1. Preclinical invitro and invivo studies showed that AG024322 is synthetically lethal with PARP inhibition115. A clinical trial has been initiated with SCH727965 and ABT-888

PARP

AG014688 (also known as CO-338 and rucaparib) AZD2281 (also known as olaparib)

ABT-888 (also known as veliparib)

INO-1001 MK4827 CEP-9722 GPI-21016 (also known as E7016) BMN673 BSI-201 (also known as iniparib NER TLS XPG ERCC1XPF Pol Cyclosporine Cetuximab 3-O-methylfunicone, aurintricarboxylic acid and ellagic acid DAC (reactivation) NU7026, NU7441, IC86621 and IC87361 OK-135 SU11752 CC-115 HRR MRE11 RAD51 Mirin B02, A03, A10 and imatinib AG024322 and SCH727965 (CDK1 inhibitors)

MMR NHEJ

MLH1 DNA-PKcs

BRCA1



REVIEWSTable 2 (cont.) | Inhibitors of the DDR*Pathway Target InhibitorKU55933, KU60019 and CP466722 Caffeine, shisandrin B, NU6027 and VE821 MK-1775

Current stagePreclinical invitro sensitization to IR, etoposide and camptothecin81,188,189 Preclinical invitro chemosensitization and radiosensitization9295 Preclinical invitro and invivo chemosensitization and radiosensitization190,191 and patient-derived sarcoma explants exvivo and as a single agent192. Evidence of activity in clinical trials193 IRC-083864 has activity in pancreatic and prostate cancer xenografts194 and has entered clinical trial under the name of Debio 0931 (REF.195) but no data are available CHK1 and CHK2 (UCN-01 is a pan-kinase inhibitor): PhaseI/II trials as a single agent and in combinations, trials were stopped owing to toxicities91 CHK1 and CHK2: PhaseI combinations with gemcitabine and with irinotecan CHK1: PhaseI combination with gemcitabine CHK1: PhaseI various drug combinations in leukaemia and lymphoma CHK1 and CHK2: PhaseI in combination with gemcitabine CHK1: PhaseI single agent trial CHK2: invitro sensitization of topoisomerase I poisons and IR196

Checkpoints ATM ATR WEE1

CDC25 CHK1 and CHK2

Several, including IRC-083864 (Debio 0931) UCN-01 AZD7762 PF00477736 SCH900776 XL9844 LY2606368 PV1019

APE1, AP endonuclease 1; ATM, ataxia-telangiectasia mutated; ATR, ataxia-telangiectasia and Rad3-related; BER, base excision repair; CDC25, cell division cycle 25; CDK1, cyclin-dependent kinase 1; CLL, chronic lymphocytic leukaemia; DAC, decitabine; DDR, DNA damage response; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; DSB, DNA double-strand break; FEN1, flap endonuclease 1; HRR, homologous recombination repair; IR, ionizing radiation; MGMT, O6-methylguanine DNA methyltransferase; MMC, mitomycin C; MMR, mismatch repair; MMS, methyl methanesulphonate; MTD, maximum tolerated dose; NER, nucleotide excision repair; NHEJ, non-homologous end joining; NIH, US National Institutes of Health; PARP, poly(ADP-ribose) polymerase; PARPi, PARP inhibitor; Pol, DNA polymerase; TLS, translesion synthesis; TMZ, temozolomide; XP, Xeroderma pigmentosum. *Where no reference is given information may be found on the ClinicalTrials.gov website (see Further information).

RadiomimeticsDrugs that introduce the same DNA damage as ionizing radiation.

Fanconi anaemiaA rare genetic disorder that results in aplastic anaemia, leukaemia and cancer susceptibility, and hypersensitivity to DNA crosslinking agents. The pathway is responsible for therepair of DNA interstrand crosslinks and overlaps somewhat with homologous recombination repair.

Crosslinking agentsMolecules with two reactive groups (for example, bifunctional alkylating agents and cisplatin) that can react with two groups in DNA on thesame strand to form intrastrand crosslinks or opposite strands to form interstrand crosslinks.

core NHEJ proteins are KU70 (also known as XRCC6), KU80 (also known as XRCC5), DNA-dependent protein kinase catalytic subunit (DNA-PKcs) KU70KU80 DNA-PKcs form a complex known as DNA-PK artemis, XRCC4XLF and ligase 4 (FIG.4). DSB recognition by ATM and the MRN complex (which comprises MRE11, RAD50 and Nijmegen breakage syndrome protein 1 (NBS1) (described in more detail below)) may also be an early event in NHEJ. Few defects in members of the NHEJ pathway have been identified as being associated with cancer (TABLE1). DNA-PKcs is a member of the PI3K-related protein kinase (PIKK) family of enzymes that also includes ATM, ataxia-telangiectasia and Rad3-related (ATR) and mTOR. PI3K inhibitors, such as wortmanin and LY294002, also inhibit DNA-PKcs, and in proof-of-concept studies these drugs hindered DSB rejoining and enhanced the cytotoxicity of DSB-inducing agents59,60. More potent and specific DNA-PKcs inhibitors have been developed61,62 (TABLE2) that substantially slow DSB repair and increase the cytotoxicity and antitumour activity of IR, radiomimetics and topoisomerase II poisons in cells and xenografts63,64. DNA-PKcs levels and activity were higher in patientderived B cell chronic lymphocytic leukaemia (B-CLL) cells that were stratified as having poor prognosis than in good prognosis B-CLL cells, and the DNA-PKcs inhibitors NU7026 and NU7441 enhanced the sensitivity of the poor prognosis B-CLL cells to topoisomerase II poisons65. Although none of these agents has progressed to clinical trials, a dual mTOR and DNA-PKcs inhibitor, CC-115, is undergoing early clinical evaluation (TABLE2). Combinations of DNA-PKcs inhibitors with radiotherapy would seem to be the most attractive on the basis of the

profound radiosensitization seen in a number of preclinical models and the reduced likelihood of toxicities that might be predicted from a combination with systemic chemotherapy. Homologous recombination repair. The homologous recombination repair (HRR) pathway of DSB repair is a highly complex process that involves multiple proteins and occurs during the S and G2 phases of the cell cycle55. The broken DNA ends of a DSB are resected to allow invasion of the single strands into the sister chromatid, which functions as a template for accurate resynthesis of the damaged DNA (FIG.4). Fortunately, this copy is accessible as the two chromatids are held together by the cohesin protein complex until mitosis66. HRR is inextricably linked to the S and G2 checkpoints (discussed below). Although this pathway repairs only a minor proportion of DSBs it may be the most crucial as it is high fidelity, it deals with stalled and collapsed replication forks, as well as single-ended DSBs, and also (in cooperation with the NER and Fanconi anaemia pathways) the processing of ICLs67,68 (FIGS1,3). HRR is crucial for the maintenance of genomic stability, and the function of the entire pathway can be compromised by mutation in one or more genes. Many tumour suppressors participate in this pathway, including BRCA1, BRCA2 and ATM (TABLE1). Tumours with HRR defects are highly sensitiveto crosslinking agents (such as cisplatin, carboplatin and nitrosoureas) and DSBs that are induced by IR and topoisomerase I poisons. Owing to the role of HRR in replication fork restart HRR-defective cells are hypersensitive to certain antimetabolites that induce base lesions and/or replication fork stalling (for example,VOLUME 12 | DECEMBER 2012 | 807


REVIEWSTOPOI 5 3 NEIL OGG1 APE1 P APE1 5 3 PARP XRCC1 TDP OH PNPK OH P OH P Pol or Pol PCNA

dNTPs PARP XRCC1 Pol LIG3 FEN1 LIG1

PARP XRCC1

gemcitabine)69. The fairly high frequency of HRR defects in tumours may underlie the efficacy of cytotoxic therapy and provide a rationale for the use of inhibitors of HRR to sensitize tumours with functional HRR to conventional chemotherapy. Moreover, HRR-defective cells are selectively killed by PARPis (discussed below). There are few HRR inhibitors, but mirin is an inhibitor of MRE11 endonuclease activity and thus of HRR function70. However, MRE11 is also upstream of NHEJ, and so mirin inhibits NHEJ71 and its effects are not specific to HRR. This makes it difficult to determine whether inhibition of HRR or inhibition of NHEJ by mirin contributes most to sensitization and hence which pathway should be inhibited for the greatest clinical benefit. Activating phosphorylation of RAD51 and RAD51 focus formation important steps in HRR are dependent on the proto-oncogene ABL1; and the BCR-ABL1 inhibitor imatinib sensitized cells to DNA crosslinking agents and IR72,73. Other prototype RAD51 inhibitors have been identified (TABLE2) but the most common way to target HRR is by inhibition of the ATMCHK2 or ATRCHK1 pathways (discussed below).

Short patch BER Upregulated in cancer

Long patch BER Inhibitor in preclinical development Inhibitor in clinical trial

Mutated or silenced in cancer Polymorphism associated with cancer

FocusThe accumulation of a substance (usually protein) thatmay be identified and visualized (usually by a fluorescently tagged antibody) in one spot.

Nature Reviews Cancer Figure 2 | Base excision repair. In the first step of| base excision repair (BER) the oxidized, deaminated and alkylated bases are removed by specific glycosylases; 8-oxoguanine DNA glycosylase (OGG1) or members of the Nei-like protein (NEIL) family are examples. The resulting apurinic or apyrimidinic (AP) site is then hydrolysed by an AP endonuclease, such as APE1. The nick in the DNA is then repaired by short patch BER (the predominant mode) or long patch BER, depending on the nature of the 5 and 3 ends and, possibly, ATP availability. Polynucleotide kinase phosphatase (PNKP; a 3 DNA phosphatase and 5 DNA kinase) may be necessary to modify the broken ends for replacement and/or rejoining. Proliferating cell nuclear antigen (PCNA), the 9-1-1 complex and flap endonuclease 1 (FEN1) are required for the processing of long patches. In short patch repair the single nucleotide is replaced by DNA polymerase- (Pol ) and the gap is rejoined by ligase 3 (LIG3), and in long patch repair up to 13 nucleotides are replaced by Pol or Pol and rejoining is completed by LIG1 (REF.143). Poly(ADP-ribose) polymerase 1 (PARP1) and XRCC1 facilitate repair by recruiting repair enzymes and providing the scaffold for short patch and long patch BER. The high negative charge of the ADP-ribose polymers in the vicinity of the break is necessary for the recruitment of XRCC1 (REF.20), which in turn recruits PNPK and DNA polymerase, and also facilitates the loosening of chromatin to facilitate repair. Single-strand breaks associated with trapped topoisomerase I involve the removal of topoisomerase I by tyrosyl-DNA phosphodiesterase 1 (TDP1) followed by PNPK activity, both of which rely on PARP1 and XRCC1 for their recruitment.

Targeting cell cycle checkpoints The DDR requires the integration of cell cycle control via checkpoint signalling to allow time for repair to prevent DNA damage being made permanent by replication and mitosis. The PIKKs ATM and ATR have crucial roles by signalling DNA damage to cell cycle checkpoints and DNA repair pathways (reviewed in REFS74,75) (FIG.6). The ATMCHK2 pathway primarily responds to DSBs to induce G1 arrest, and the ATRCHK1 pathway triggers S and G2 phase arrest. ATM promotes HRR by recruiting BRCA1 to DSBs but can also antagonize BRCA1 and promote NHEJ by recruiting p53 binding protein 1 (53BP1), and these antagonistic functions may be cell cycle regulated76. ATR is activated by DNA single-strand double-strand junctions that arise as intermediates in NER, at stalled replication forks and at resected DSBs, it phosphorylates CHK1 to activate S and G2 arrest, interacts with the MMR machinery and stabilizes and re-starts stalled replication forks via HRR77. ATR and CHK1 phosphorylate a number of proteins involved in HRR and ICL repair, including BRCA2, RAD51, Fanconi anaemia group D2 (FANCD2) and FANCE 78. ATM mutations, epigenetic silencing and polymorphisms are associated with various types of cancer, and a high frequency of frameshift mutations of the A(10) repeat microsatellite of ATR, which results in a truncation, is associated with various tumours with MSI79 (TABLE1). As checkpoint activation is a common feature of the DDR to a variety of DNA-damaging agents, and as ATM and ATR have multiple downstream targets, checkpoint inhibitors are likely to sensitize a broad range of DNA-damaging agents75,80. The first selective ATM inhibitor, KU55933, inhibited IR-induced ATM-dependent phosphorylation events and sensitized cancer cells to IR and topoisomerase inhibitors81. This, and other inhibitors (TABLE2), are at an early stage of development, possibly because the role of ATM in promoting oncogene-induced senescence and inhibiting tumorigenesis82 may have been awww.nature.com/reviews/cancer


REVIEWSdeterrent from developing inhibitors. However, DDR pathways are generally tumour suppressive and ATM is still a potentially viable therapeutictarget. Targeting the S and G2 checkpoints is particularly attractive for cancer therapy because loss of G1 checkpoint control is a common feature of cancer cells83, making them more reliant on the S and G2 checkpoints to prevent DNA damage triggering cell death. Proof-of-principle genetic studies showed that inhibiting the S and G2 checkpoints by inactivation of ATR or CHK1 abrogated DNA damage-induced G2 checkpoint arrest and sensitized cancer cells to a variety of DNAdamaging chemotherapeutic agents8486. Furthermore, oncogenic replicative stress may render cancer cells sensitive to inhibitors that prevent the S and G2 checkpoints as single agents (discussed below). WEE1 and cell division cycle 25 (CDC25) inhibitors have been developed, and some have entered into clinical trials (TABLE2) but clinical data are limited. Most research has focused on the development of CHK1 inhibitors, which have entered clinical studies (TABLE2). These include the nonspecific staurosporin analogue, UCN-01, and its derivatives, and highly potent selective CHK1 inhibitors (reviewed in REF.87). In preclinical studies most of these inhibitors had little or no impact on cell cycle distribution or viability perse but they did prevent cell cycle arrest and increased cytotoxicity induced by DNA-damaging agents. Xenograft studies, mostly in combination with gemcitabine, confirmed sensitization8890. UCN-01 was the first of this type of inhibitor to enter clinical trials, but after PhaseII trials it was discontinued owing to dose-limiting toxicities and a lack of convincing efficacy that was probably due to poor specificity and pharmacokinetics. The newer, more specific inhibitors of CHK1 have generally been combined with gemcitabine in PhaseI studies, in which myelosuppression was the major toxicity that led to the termination of the trials, and no efficacy data have yet been presented (reviewed in REFS87,91). For a long time, the only available ATR inhibitor was caffeine, which lacks potency and specificity but that has nevertheless provided useful preclinical radiosensitization data92. Recently, a high-throughput cell-based screen has identified potent ATR inhibitors that were cytotoxic to cells with replicative stress93, and other groups have identified the novel ATR inhibitors, VE-821 and NU6027 (REFS94,95). Both drugs sensitized cells to a variety of DNA-damaging agents, which reflects the role of ATR in the response to multiple DNA lesions, but differed in their dependence on p53 status. As these reports are the first to use novel ATR inhibitors, and different cell lines were investigated, no conclusions can be drawn regarding the role of p53 in determining response. NU6027 also inhibited RAD51 focus formation (indicative of HRR suppression), leading to synthetic lethality with BER inactivation95 (discussed below).

TC-NER RNA Pol II CSA CSB TFIIH XPF ERCC1 XPA XPG

GG-NER

5 3

5 3 XPC DDB RAD23B XPE

RPA RFC PCNA Polymorphism associated with cancer Mutated or silenced in cancer Inhibitor in preclinical development Pol or Pol

LIG3

Figure 3 | Nucleotide excision repair. Preferential repair of lesions that stall Nature Reviews | Cancer transcription on the coding strand is by transcription-coupled nucleotide excision repair (TC-NER); the entire genome is repaired by global NER (GG-NER)42. These pathways differ in their initial steps, TC-NER involves Cockayne syndrome WD repeat protein A (CSA) and CSB, whereas in GG-NER recognition is dependent on Xeroderma pigmentosum group C-complementing protein (XPC)RAD23B and DNA damage-binding protein (DDB); XPA, replication protein A (RPA) and TFIIH are involved in both pathways. Thereafter the steps are common, with excision of the damaged oligonucleotide by XPG and ERCC1XPF, then resynthesis of the intact oligonucleotide and ligation are accomplished by DNA polymerase- (Pol ) or Pol and DNA ligase 3 (LIG3). PCNA, proliferating cell nuclear antigen; RFC, replication factor C.

Synthetic lethality Perhaps the most promising prospect for the future of cancer treatment is the exploitation of dysregulated DDR by the synthetic lethality approach. This term describes the process by which defects in two different genes or pathways together result in cell death but independently do not affect viability (FIG.7). The term was first applied to explain the selective killing of cancer cells with particular molecular defects by some agents96. For example, hyperproliferation in tumour cells that is caused by defects in the RB pathway may elevate topoisomerase II activity, making these cells particularly sensitive to topoisomeraseII poisons, such as doxorubicin97. However, doxorubicin has marked side effects on normal tissues and the aim of synthetic lethal approaches is to combine tumour-associated molecular defects with corresponding selectively lethal drugs or to combine drugs that confer selective lethality to tumour cells and not to normaltissue. Therefore, it was exciting to discover that cells and xenograft tumours defective in HRR (for example, through BRCA1 or BRCA2 mutation) could be killed by concentrations of PARPis that were non-toxic to HRRfunctional cells and normal tissues98,99. One proposed mechanism of this synthetic lethality is that the high levels of endogenous SSBs (FIG.1) remain unrepaired when PARP is inhibited and cause stalled replicationVOLUME 12 | DECEMBER 2012 | 809


REVIEWSaMMR Base mismatch 5 3 Deletion or insertion

b Direct repairPol or Pol Pol MSH2 MSH3 MLH1 MLH3 Endonuclease 1 Degradation MMR Repairs newly synthesized strand leaving damaged base on template strand MGMT CH3 CH3 O6meG or 6TG

MSH2 MSH6

PMS2

RPA PCNA RFC Pol or Pol FEN1 LIG1

Cell death in subsequent S phase

Upregulated in cancer Mutated or silenced in cancer Inhibitor in preclinical development Inhibitor in clinical trial

Figure 4 | Mismatch repair. a | DNA mismatches resulting from the insertion of a mispaired or fraudulent nucleotide arerecognized by MSH2MSH6 heterodimers, whereas deletions and insertions are recognized by MSH2MSH3 | Cancer Nature Reviews heterodimers. Downstream processing requires PMS2 and MLH1MLH3 heterodimers. Components of other DNA repair pathways such as endonuclease 1, replication protein A (RPA), proliferating cell nuclear antigen (PCNA), replication factor C (RFC), DNA polymerase- (Pol ) or Pol and flap endonuclease 1 (FEN1) have also been implicated in excision and resynthesis past the lesion. Importantly, mismatch repair (MMR) is strand-specific, correcting the daughter strand. Therefore, this pathway is crucial for the repair of replication errors inserted opposite the correct template strand under normal circumstances. b | However, when damage on the template strand, such as O6-methylguanine (O6meG) or 6-thioguanine (6TG), causes mispairing at replication the MMR machinery attempts to repair the newly synthesized strand, rather than the damaged one, which results in a DNA double-strand break during the subsequent S phase or causes apoptosis owing to signalling by the MMR machinery to ataxia-telangiectasia and Rad3-related (ATR)CHK1 (REF.51). LIG1, DNA ligase 1; MGMT, O6-methylguanine DNA alkyltransferase.

forks that require HRR for their resolution. The trapping of PARP1 at these breaks (PARPis prevent catalytic activity but not DNA binding) may further impede repair and require HRR for resolution. Alternatively, PARP1 similar to HRR may have a role in restarting stalled replication forks100. In HRR-defective cells these stalled forks persist or are erroneously repaired, which leads to cell death (FIG.4b). HRR defects are not limited to the classic examples of BRCA1 and BRCA2 defects and their association with inherited breast and ovarian cancer syndrome. Epigenetic silencing of BRCA1 also rendered cells and xenografts sensitive to PARP inhibition, demonstrating the potential of PARPis in sporadic cancer 38. Moreover, hypoxia causes the downregulation of several HRR genes, resulting in contextual synthetic lethality with PARPis, and it has recently been suggested that hyperthermia induces BRCA2 degradation101,102.810 | DECEMBER 2012 | VOLUME 12

The prospect of non-toxic therapy has real clinical potential, with multiple clinical trials of PARPis as single agents underway (TABLE2). Data with the PARPi olaparib are very encouraging. Good responses were seen in patients with BRCA-associated breast and ovarian cancer, and even in unselected patients with high-grade serous ovarian cancer 103,104, which is a cancer type that is now known to have a high frequency of HRR defects. Furthermore, knockdown of various other genes involved in HRR also confers sensitivity to PARPis105, expanding the range of potential targets for PARPi therapy. Reliable biomarkers to identify tumours carrying these defects that would render them sensitive to PARPis are needed (discussed below). Similarly, inhibition of BER with APE1 inhibitors is also synthetically lethal in cells with HRR dysfunction106. One would predict that inhibitors of Pol (another BER component) would have similar activity.www.nature.com/reviews/cancer

2012 Macmillan Publishers Limited. All rights reserved

REVIEWSICL Stalled replication forks DSB PARP BRCA1 53BP1 CtIP EXO1 HRR RPA RAD51 BRCA2 RAD52 RAD54 SSB XPF ERCC1 NHEJ MRN ATR FANC complex BRCA2 RAD51 RPA Unhooking

ATM

ICL repair

KU70 KU80 DNA-PKcs End synapsis Artemis

H2AX P

Upregulated in cancer Mutated or silenced in cancer

Sister chromatid XRCC4 XLF LIG4 Resolvases ligase

Polymorphism associated with cancer Inhibitor in preclinical development Inhibitor in clinical trial

Ligation

G1 or G0

S or G2

Figure 5 | DNA double-strand break and interstrand crosslink repair. An early step of DNA double-strand break (DSB) Nature Reviews | Cancer repair is the recruitment of the MRN nuclease complex (comprised of MRE11, RAD50 and Nijmegen breakage syndrome 1 (NBS1)). In non-homologous end joining (NHEJ) binding of the KU70KU80 heterodimer and the DNA-dependent protein kinase catalytic subunit (DNA-PKcs) to form DNA-PK ensures synapsis of the DNA ends144. DNA-PKcs phosphorylates histone H2AX, and crucially also itself, which allows dissociation. Artemis processes the DNA ends, which are then ligated by DNA ligase 4 (LIG4) and stabilized by the XRCC4XRCC4-like factor (XLF) complex56. In homologous recombination repair (HRR), BRCA1 and possibly poly(ADP) ribose polymerase 1 (PARP1) facilitates recruitment of the MRN complex, which together with CtBP-interacting protein (CtIP) and exonuclease 1 (EXO1) resect the DNA ends145,146. The MRN complex recruits and activates ataxia-telangiectasia mutated (ATM), which stimulates MRE11, NBS1, CtIP and EXO1 by phosphorylation. ATM also phosphorylates histone H2AX, which aids recruitment of p53 binding protein 1 (53BP1) and BRCA1 (REF.147). The single-stranded DNA overhang is rapidly coated with replication protein A (RPA), which prevents it from being degraded. This recruits the ataxia-telangiectasia and Rad3-related (ATR)ATR-interacting protein (ATRIP) complex, which signals via phosphorylation of CHK1 to induce S and G2 arrest (not shown). ATM and ATR phosphorylate BRCA1, which stimulates its E3 ubiquitin ligase activity. ATR also phosphorylates RPA2 and the kinase CHK1, which in turn phosphorylates RAD51. RAD51 is then delivered by BRCA2 to displace RPA to form the nucleoprotein filament that can invade the complementary duplex DNA, forming a Holliday junction148,149. The invading strand is extended by DNA polymerase and rejoins the end of the DSB, to form a crossover or non-crossover repair product. Stalled replication forks primarily activate ATR rather than ATM77. The Fanconi anaemia (FANC) proteins also promote HRR at stalled replication forks resulting from interstrand crosslinks (ICLs). Recruitment of the Blooms syndrome helicase complex leads to signalling to cell cycle checkpoints via ATRCHK1 and repair proteins including RPA, BRCA1, FANCN and BRCA2, which are important components of HRR.

Excitement about the synthetic lethal approach involving PARPis must be tempered, as resistance to PARPis can develop owing to secondary mutations in BRCA1 or BRCA2 that restore their function107,197. In addition, even in BRCA-mutant cells, HRR function and PARPi resistance can be restored if 53BP1 or DNA-PKcs are also inactivated76,108, a phenomenon that is known as synthetic viability. These observations suggesting that NHEJ and HRR compete to repair DSBs, but that survival is dependent on DSB repair by HRR, are analogous to studies in Fanconi anaemia pathway-defective human cells in which the sensitivity to ICL-induced DSBs can be rescued by the inhibition of NHEJ109. 53BP1 is commonly lost in BRCA1-mutant and triple-negative breast cancer and in lung cancer 110, which could compromise the activity of PARPis against these types of tumours in clinicaltrials.NATURE REVIEWS | CANCER

Other examples of synthetic lethality of HRR defects and PARPis are emerging. For example, HRR dysfunction owing to ATM loss in haematological malignancies or to mutation of MRE11 (secondary to MMR defects) confers sensitivity to PARPis111113. It is possible that such secondary defects in MRE11 may also contribute to the sensitivity of MSH2-defective cells to Pol (a component of BER) inhibition114. The principle of tumour-selective synthetic lethality is to exploit defects found only in the tumour, and the use of two agents targeted at two compensatory pathways would be predicted to be generally toxic rather than tumour-specific. However, the cyclin-dependent kinase CDK1 activates BRCA1, and CDK1 inhibition in combination with PARP inhibition was cytotoxic in lung cancer models but spared normal tissues115. Although these data were only published recently, clinical trials using this combination have already been initiated.VOLUME 12 | DECEMBER 2012 | 811

2012 Macmillan Publishers Limited. All rights reserved

REVIEWSDSB Replication stress

NER Resected DSB

ATRIP ATM ATR

p53

CHK2

CHK1

CDC25A

CDC25A

CDC25C

WEE1

CDK4 or CDK6 Cyclin D

CDK2 Cyclin E

CDK2 Cyclin A or cyclin E

CDK1 Cyclin B

G1/S Stop Cell cycle progression

S phase Stop

G2/M Stop

genomic integrity. ATM knockdown, or inhibition with KU55933, was synthetically lethal in cells with defects in the Fanconi anaemia pathway, members of which are commonly mutated or lost in cancers118, and the CHK1 inhibitor Go6976 reduced cell survival and profoundly increased cisplatin sensitivity in cells with a defective Fanconi anaemia pathway 119, thus raising the potential for targeted therapy with ATM and CHK1 inhibitors. Hyperactive growth factor signalling and oncogeneinduced replicative stress increase DNA breakage that activates the ATRCHK1 pathway, and some examples of the synthetic lethality of checkpoint or DNA repair inhibitors in cells harbouring activated oncogenes have been identified. ATR knockdown was synthetically lethal in cells transformed with mutant KRAS120, and inhibition of CHK1 and CHK2 significantly delayed disease progression of transplanted MYC-overexpressing lymphoma cells invivo 121. Similarly, the oncogenic ETS fusion genes that drive several cancer types, including prostate cancer, Ewings sarcoma and acute myeloid leukaemia, increase DNA damage and confer hypersensitivity to PARPis and DNA-PKcs inhibitors; the combination of olaparib and TMZ also resulted in the complete regression of Ewings sarcoma xenografts122124. Taken together, these data demonstrate synthetic lethal interactions between the HRR and BER pathways, between checkpoint signalling and the Fanconi anaemia pathway, and between activated oncogenes and the disruption of S and G2 checkpoints and DNA repair. As the arsenal of inhibitors that target components of these pathways expands the potential for using these synthetic lethal interactions increases, as long as the exploitable defects in the tumour can be identified with suitable biomarkers.

Figure 6 | Signalling DNA damage to cell cycle checkpoints. Ataxia-telangiectasia Nature Reviews | G1 mutated (ATM) is activated by DNA double-strand breaks (DSBs) and triggers theCancer checkpoint, by phosphorylating and hence activating CHK2 and p53. Ataxiatelangiectasia and Rad3-related (ATR) is primarily activated by junctions of single-stranded DNA and double-stranded DNA, which arise at stalled replication forks and resected DSBs and are nucleotide excision repair (NER) intermediates. This triggers the intra-S phase and the G2 checkpoints via phosphorylation of CHK1, which in turn phosphorylates WEE1 (which activates this kinase) and cell division cycle 25 (CDC25) phosphatases (which inhibits it) to inhibit cell cycle progression through the coordinate suppression of cyclin-dependent kinase (CDK) activity74. It is important to note that there is crosstalk between the ATMCHK2 and ATRCHK1 pathways and that they share many substrates. ATRIP, ATR-interacting protein. Dashed arrows indicate secondary targets.

PharmacodynamicThe physiological or biochemical effect of a drug on the body. A pharmacodynamic biomarker is a measure of this effect; for example, the product of an enzyme reaction may be reduced by a drug that inhibits the enzyme.

Analogously, with the increased sensitivity to PARP inhibition when HRR is compromised, inhibition of HRR reduces the survival of cells in which BER is compromised; for example, indirect inhibition of HRR with CHK1 and/or CHK2 inhibition is synthetically lethal with PARPis, particularly in MYC-overexpressing cells that are in a hyperproliferative state116,117. The ATR inhibitor NU6027 was also more profoundly cytotoxic to BER-defective cells and in BER-functional cells treated with a PARPi, reflecting the complementarity of HRR and BER95. Synthetic lethality between checkpoint signalling and the Fanconi anaemia pathway is emerging, suggesting complementary roles in the maintenance of

Biomarkers Laboratory and clinical evidence clearly demonstrates that defects in specific DDR pathways can render cancer cells particularly vulnerable to certain therapeutic agents, be they targeted agents or conventional cytotoxic agents. Pathway defects may be more common in one tumour type than another, probably contributing to the empirical selection of one drug over another in tumours of different tissue origin, but, as TABLE1 shows, such pathway defects may occur in other tumour types and could be exploitable with the appropriate agent. Predictive biomarkers of these defects are needed to be able to exploit them for a stratified approach, rather than the conventional cancer type-specific approach. Pharmacodynamic biomarkers that monitor the biological effects of the drug are also needed to guide the trials of molecularly targeted agents in which target inhibition, rather than toxicity, will determine the selection of the dose to be administered.Predictive biomarkers. When the inactivation of a single gene has been identified as a crucial determinant of sensitivity it may then be used to select patients for the appropriate therapy. For example, low levels of the NER endonuclease ERCC1 correlate with cisplatin sensitivity in non-small-cell lung cancer (NSCLC), bladder cancerwww.nature.com/reviews/cancer


REVIEWSNormal cell DNA damage DNA damage DNA damage Tumour cell DNA damage

Survival

Survival

Survival

Survival

defects: gene expression profiling, methylation-specific arrays, IHC analysis of tissue microarrays (TMAs) and copy number aberrations by array comparative genomic hybridization (aCGH), which have varying degrees of predictive power 129132. An alternative approach is to assess HRR function in fresh viable tumour material by measuring the number of RAD51 foci (FIG.4) following exvivo DNA damage induction133135. RAD51 foci may also be measured in formalin-fixed tumour biopsy samples taken after chemotherapy treatment 136. Owing to the invasive procedures that are needed to obtain tumour material, except in the case of haematological malignancies, circulating tumour cells offer the best hope of routinely obtaining suitable material137. Pharmacodynamic biomarkers. A general marker of DNA damage is the phosphorylation of histone H2AX (referred to as H2AX) by ATM, ATR and DNA-PK. H2AX foci, formed at sites of DSBs, or increased levels of H2AX, may be measured by immunofluorescence microscopy, flow cytometry or immunoblotting and used to detect DNA damage138. The increase and/or persistence of H2AX can be used to demonstrate the inhibition of PARP, DNA-PK, ATR and CHK1 (as this would indicate the increase or persistence of DNA damage). To measure the effect of an agent that directly causes DNA DSBs in all phases of the cell cycle (for example, etoposide) patient-derived lymphocytes can be used139. However, proliferating tissue is needed to measure the effect of an agent that causes replication fork stalling. Hair follicles have been used to detect H2AX foci in clinical trials of PARPis140. To directly measure the effect of a molecularly targeted agent, immunological methods may be used to detect the product. For example, activation of DNA-PK and ATM in response to DNA damage can be determined by measuring their autophosphorylation with phospho-specific antibodies, and PARP activity may be measured by immunodetection of the ADP-ribose polymer product, to guide PARPi clinical trials35,141143.

Pathway C Death

Pathway C

Pathway C

Pathway C

Figure 7 | Synthetic lethality. Normal cells have functional DNA repairReviews | Cancer Nature pathways so that they respond appropriately to damage by activating either pathway A or pathwayB. There may also be an inappropriate pathway C that would incorrectly activate cell death in response to the damage, which is suppressed. In the presence of an inhibitor of pathway B the normal cell can still rely on pathway A for survival. In a tumour cell there may be a mutation or silencing of a component of pathway A, compromising its ability to process the damage. In this case the tumour is reliant on pathway B for survival so that when a pathway B inhibitor is applied the cell dies owing to either the persistence of DNA damage that can no longer be repaired or the activation of the inappropriate pathway C. When no external cytotoxic agent is applied and the combination of pathway A and pathway B inhibition is highly cytotoxic there is synthetic lethality; when the combination is less cytotoxic but confers hypersensitivity to low doses of a cytotoxic agent, this can be termed synthetic sickness. For example, the working hypothesis for the synthetic lethality of poly(ADP) ribose polymerase (PARP) inhibition (which inhibits base excision repair (BER)) in cells with defects in homologous recombination repair (HRR). When both BER and HRR are inactive not only does this mean that replication forks that collide with unrepaired single-strand breaks (SSBs) result in lesions that cannot be repaired (as this is usually repaired by HRR) but also that NHEJ becomes activated and contributes to their erroneous and lethal repair.


Pathway A

Pathway A

Pathway A

Pathway A

Pathway B

Pathway B

Pathway B

Pathway B

Death

Death

Death

and ovarian cancer 44,125. Alternatively, a downstream effect can be used to identify any defect in a pathway (for example, MSI to identify MMR defects)48. Because defective MMR causes resistance to cisplatin, the determination of MSI and ERCC1 levels could be used to decide whether treatment with cisplatin would be appropriate. Immunohistochemistry (IHC) analysis of formalinfixed paraffin-embedded samples may be a useful tool for identifying DDR defects in order to stratify patients, and there is some limited evidence that NHEJ components can be used as predictive biomarkers. For example, KU80 expression was an independent predictor for poor response to radiotherapy in patients with head and neck squamous cell carcinoma (HNSCC)126. There is considerable interest in identifying HRR defects because of their potential exploitation by PARPis, as well as conventional agents. For example, patients with BRCA-mutated breast cancer responded better to cisplatin than to conventional breast cancer therapy 127. Mutational screening for germline BRCA1 and BRCA2 defects only identifies a subset of HRR-defective tumours and does not detect epigenetic silencing of these genes in sporadic cancers128. Several studies report methods to identify tumours with non-germline HRR

Final conclusion and perspective Cancer-specific dysregulation of the DDR is ripe for therapeutic exploitation with the appropriate conventional cytotoxic and/or DDR inhibitor targeted at a compensatory pathway. Current treatment may not always be optimum. For example, BRCA-mutant breast cancers respond much better to cisplatin than to standard therapy 127. Ideally, if anticancer therapy could be selected on the basis of the molecular phenotype of the cancer, rather than the tissue in which it arose, it could improve response, spare patients from unnecessary side effects and reduce healthcare costs of administering drugs that are likely to have little effect. Reliable biomarkers to identify DDR defects and potential vulnerabilities in minute quantities of tissue will be key to the success of this approach. Chemosensitization and radiosensitization with DDR inhibitors is promising preclinically, but careful titration of both the cytotoxic agent and the DDR inhibitor is needed to avoid increased toxicity in the clinicalVOLUME 12 | DECEMBER 2012 | 813

REVIEWSsetting. The exploitation of DDR defects through synthetic lethality represents a more exciting approach to truly tumour-specific therapies with reduced toxicities. Clinical proof-of-principle data that this approach can work are provided by the PARPis in BRCA-defective cancers. There are probably more synthetic lethalities to be discovered and exploited, and there is also a pressing need to identify these combinations along with biomarkers to discover exploitable defects in cancer in patients.

1.

2. 3.

4.

5.

6. 7.

8.

9.

10.

11.

12. 13.

14.

15.

16.

Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646674 (2011). An excellent review highlighting enabling characteristics, as well as emerging hallmarks, of cancer. Bartkova, J. etal. DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature 434, 864870 (2005). Gorgoulis, V. G. etal.Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions. Nature 434, 907913 (2005) References 2 and 3 show that early in tumorigenesis oncogenic stress activates the DDR, which acts as a temporary barrier to progression, but defects in the DDR eventually arise allowing full-blown cancer to develop. Tubbs, J.L., Pegg, A.E. & Tainer, J.A. DNA binding, nucleotide flipping, and the helix-turn-helix motif in base repair by O6-alkylguanine-DNA alkyltransferase and its implications for cancer chemotherapy. DNA Repair 6, 11001115 (2007). Tricker, A.R. & Preussmann, R. Carcinogenic N-nitrosamines in the diet: occurrence, formation, mechanisms and carcinogenic potential. Mutat. Res. 259, 277289 (1991). Saffhill, R., Margison, G.P. & OConnor, P.J. Mechanisms of carcinogenesis induced by alkylating agents. Biochim. Biophys. Acta 823, 111145 (1985). Zaidi, N.H., Liu, L. & Gerson, S.L. Quantitative immunohistochemical estimates of O6-alkylguanineDNA alkyltransferase expression in normal and malignant human colon. Clin. Cancer Res. 2, 577584 (1996). Pegg, A.E. Mammalian O6-alkylguanine-DNA alkyltransferase: regulation and importance in response to alkylating carcinogenesis and therapeutic agents. Cancer Res. 50, 61196129 (1990). Rabik, C.A., Njoku, M.C. & Dolan, M.E. Inactivation of O6-alkylguanine DNA alkyltransferase as a means to enhance chemotherapy. Cancer Treat. Rev. 32, 261276 (2006). Friedman, H.S. etal. Phase I trial of O6-benzylguanine for patients undergoing surgery for malignant glioma. J.Clin. Oncol. 16, 35703575 (1998). The first clinical trial of a DNA repair inhibitor supported by pharmacodynamic measurement of the impact on O6-methylguanine levels in tumour material. Ranson, M. etal. Lomeguatrib, a potent inhibitor of O6-alkylguanine-DNA-alkyltransferase: phase I safety, pharmacodynamic, and pharmacokinetic trial and evaluation in combination with temozolomide in patients with advanced solid tumors. Clin. Cancer Res. 12, 15771584 (2006). Watson, A.J. etal. Tumor O6-methylguanine-DNA methyltransferase inactivation by oral lomeguatrib Clin. Cancer Res. 16, 743749 (2010). Esteller, M., Hamilton, S.R., Burger, P.C., Baylin, S.B. & Herman, J.G. Inactivation of the DNA repair gene O6-methylguanine-DNAmethyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res. 59, 793797 (1999). Hegi, M.E. etal. MGMT gene silencing and benefit from temozolomide in glioblastoma. N.Engl. J.Med. 352, 9971003 (2005). The demonstration in a study of >200 patients that patients with epigenetic silencing of MGMT had a significantly better response to TMZ and radiotherapy. Lindahl, T. Instability and decay of the primary structure of DNA. Nature 362, 709715 (1993). Review of the extent of endogenous and environmental damage to DNA, linking it to ageing and cancer. van Loon, B., Markkanen, E. & Hbscher, U. Oxygen as a friend and enemy: how to combat the mutational potential of 8-oxo-guanine. DNA Repair 9, 604616 (2010).

17. Wiseman, H. & Halliwell, B. Damage to DNA by reactive oxygen and nitrogen species: role in inflammatory disease and progression to cancer. Biochem. J. 313, 1729 (1996). 18. De Vos, M., Schreiber, V. & Dantzer, F. The diverse roles and clinical relevance of PARPs in DNA damage repair: current state of the art. Biochem. Pharmacol. 84, 137146 (2012). 19. Schreiber, V., Dantzer, F., Ame, J.C. & de Murcia, G. Poly(ADP-ribose): novel functions for an old molecule. Nature Rev. Mol. Cell Biol. 7, 517528 (2006). 20. El-Khamisy, S.F., Masutani, M., Suzuki, H. & Caldecott, K.W. A requirement for PARP-1 for the assembly or stability of XRCC1 nuclear foci at sites of oxidative DNA damage. Nucleic Acids Res. 31, 55265533 (2003). An elegant study demonstrating that ADP-ribose polymer formation at the site of the DNA break is necessary to recruit XRCC1, the scaffold protein of BER. 21. Plo, I. etal. Association of XRCC1 and tyrosyl DNA phosphodiesterase Tdp 1 for the repair of topoisomerase I-mediated DNA lesions. DNA Repair 2, 10871100 (2003). 22. Larsen, E., Meza, T.J., Kleppa, L. & Klungland, A. Organ and cell specificity of base excision repair mutants in mice. Mutat. Res. 614, 5668 (2007). 23. Sweasy, J.B., Lang, T. & DiMaio, D. Is base excision repair a tumour suppressor mechanism? Cell Cycle 5, 250259 (2006). 24. Abbotts, R. & Madhusudan, S. Human AP endonuclease 1 (APE1): from mechanistic insights to druggable target in cancer. Cancer Treat. Rev. 36, 425435 (2010). 25. Hirai, K., Ueda, K. & Hayaishi, O. Aberration of poly(adenosine diphosphate-ribose) metabolism in human colon adenomatous polyps and cancers. Cancer Res. 43, 34413446 (1983). 26. Zaremba, T. etal. Poly(ADP-ribose) polymerase-1 polymorphisms, expression and activity in selected human tumour cell lines. Br. J.Cancer 21, 256262 (2009). 27. Rouleau, M., Patel, A., Hendzel, M.J., Kaufmann, S.H. & Poirier, G.G. PARP inhibition: PARP1 and beyond. Nature Rev. Cancer 10, 293301 (2010). 28. Taverna, P. etal. Methoxyamine potentiates DNA single strand breaks and double strand breaks induced by temozolomide in colon cancer cells. Mutat. Res. 485, 269281 (2001). The early identification that inhibition of APE1 endonuclease could increase TMZ cytotoxicity. 29. Luo, M. & Kelley, M.R. Inhibition of the human apurinic/apyrimidinic endonuclease (APE1) repair activity and sensitization of breast cancer cells to DNA alkylating agents with lucanthone. Anticancer Res. 24, 21272134 (2004). 30. Del Rowe, J.D. etal. Accelerated regression of brain metastases in patients receiving whole brain radiation and the topoisomerase II inhibitor, lucanthone. Int. J.Radiat. Oncol. Biol. Phys. 43, 8993 (1999). 31. Mohammed, M.Z. etal. Development and evaluation of human AP endonuclease inhibitors in melanoma and glioma cell lines. Br. J.Cancer 104, 65366563 (2011). 32. McNeill, D.R. & Wilson, D.M. A dominant-negative form of the major human abasic endonuclease enhances cellular sensitivity to laboratory and clinical DNA-damaging agents. Mol. Cancer Res. 5, 6170 (2007). 33. Bowman K. J., White, A., Golding, B. T., Griffin, R. J. & Curtin, N. J. Potentiation of anticancer agent cytotoxicity by the potent poly(ADP-ribose) polymerase inhibitors, NU1025 and NU1064. Br. J.Cancer 78, 12691277 (1998). 34. Calabrese, C.R. etal. Preclinical evaluation of a novel poly(ADP-ribose) polymerase-1 (PARP-1) inhibitor, AG14361, with significant anticancer chemo- and radio-sensitization activity. J.Natl Cancer Inst. 96, 5667 (2004).

35.

36.

37.

38.

39.

40.

41.

42.

43. 44.

45. 46.

47.

48.

49.

50.

51.

The first demonstration that a potent PARPi could enhance the antitumour activity of IR and topoisomerase I poisons in xenograft models, with complete durable regressions observed in combination with TMZ. Plummer, R. etal. Phase I study of the poly(ADPribose) polymerase inhibitor, AG014699, in combination with temozolomide in patients with advanced solid tumors. Clin. Cancer Res. 14, 79177923 (2008). The first clinical trial of a PARPi in cancer patients. Plummer, R. etal. First and final report of a phaseII study of the poly(ADP-ribose) polymerase (PARP) inhibitor, AG014699, in combination with temozolomide (TMZ) in patients with metastatic malignant melanoma (MM). J.Clin. Oncol. Abstr. 24, 8013 (2006). Zhang, Y.W. etal. Poly(ADP-ribose) polymerase and XPF-ERCC1 participate in distinct pathways for the repair of topoisomerase I-induced DNA damage in mammalian cells. Nucleic Acids Res. 39, 31283140 (2011). Drew, Y. etal. Therapeutic potential of poly(ADPribose) polymerase inhibitor AG014699 in human cancers with mutated or methylated BRCA1 or BRCA2. J.Natl Cancer Inst. 103, 334346 (2011). Thomas, H.D. etal. Preclinical selection of a novel poly(ADP-ribose) polymerase (PARP) inhibitor for clinical trial. Mol. Cancer Ther. 6, 945956 (2007). Mehta, M.P. etal. Phase I safety and pharmacokinetic (PK) study of veliparib in combination with whole brain radiation therapy (WBRT) in patients (pts) with brain metastases. J.Clin. Oncol. Abstr. 30, 2013 (2012). Naegeli, H. & Sugasawa, K. The xeroderma pigmentosum pathway: decision tree analysis of DNA quality. DNA Repair 10, 673683 (2011). Andressoo, J.O., Hoeijmakers, J.H. & de Waard, H. Nucleotide excision repair and its connection with cancer and ageing. Adv. Exp. Med. Biol. 570, 4583 (2005). Koberle, B. etal. DNA repair capacity and cisplatin sensitivity of human testis tumour cells. Int. J.Cancer 70, 551555 (1997). Usanova, S. etal. Cisplatin sensitivity of testis tumour cells is due to deficiency in interstrandcrosslink repair and low ERCC1-XPF expression. Mol. Cancer 9, 248 (2010). Lange, S.S., Takata, K. & Wood, R.D. DNA polymerases and cancer. Nature Rev. Cancer 11, 96110 (2011). Mizushina, Y. etal. 3-O-methylfunicone, a selective inhibitor of mammalian Y-family DNA polymerases from an Australian sea salt fungal strain. Mar. Drugs 7, 624639 (2009). Dorjsuren, D. etal. A real-time fluorescence method for enzymatic characterization of specialized human DNA polymerases. Nucleic Acids Res. 37, e128 (2009). Umar, A. etal. Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J.Natl Cancer Inst. 96, 261268 (2004). Wu, X., Xu, Y., Chai, W. & Her, C. Causal link between microsatellite instability and hMRE11 dysfunction in human cancers. Mol. Cancer Res. 9, 14431448 (2011). Ham, M.F. etal. Impairment of double-strand breaks repair and aberrant splicing of ATM and MRE11 in leukemia-lymphoma cell lines with microsatellite instability. Cancer Sci. 97, 226234 (2006). Demonstration of the impact of MMR defects on the function of HRR pathways. Karran, P. & Bignami, M. DNA damage tolerance, mismatch repair and genome instability. Bioessays 16, 833839 (1994).



REVIEWS52. Fordham, S.E., Matheson, E.C., Scott, K., Irving, J.A. & Allan, J.M. DNA mismatch repair status affects cellular response to Ara-C and other anti-leukemic nucleoside analogs. Leukemia 25, 10461049 (2011). 53. Plumb, J.A., Strathdee, G., Sludden, J., Kaye, S.B. & Brown, R. Reversal of drug resistance in human tumor xenografts by 2-deoxy-5-azacytidine-induced demethylation of the hMLH1 gene promoter. Cancer Res. 60, 60396044 (2000). 54. Vilenchik, M.M. & Knidson, A.G. Endogenous DNA double-strand breaks: production, fidelity of repair, induction of cancer. Proc. Natl Acad. Sci. USA 100, 1287112876 (2003). 55. Shrivastav, M., De Haro, L.P. & Nickoloff, J.A. Regulation of DNA double-strand break repair pathway choice. Cell Res. 18, 134147 (2008). 56. Mahaney, B.L., Meek, K. & Lees-Miller, S.P. Repair of ionizing radiation-induced DNA doublestrand breaks by non-homologous end-joining. Biochem. J. 417, 639650 (2009). 57. Beucher, A. etal. ATM and Artemis promote homologous recombination of radiation-induced DNA double-strand breaks in G2. EMBO J. 28, 34133427 (2009). 58. McClendon, A.K.& Osheroff, N. DNA topoisomerase II, genotoxicity, cancer. Mutat. Res. 623, 8397 (2007). 59. Rosenzweig, K.E., Youmell, M.B., Palayoor, S.T. & Price, B.D. Radiosensitization of human tumor cells by the phosphatidylinositol3-kinase inhibitors wortmannin and LY294002 correlates with inhibition of DNA-dependent protein kinase and prolonged G2-M delay. Clin. Cancer Res. 3, 11491156 (1997). Early demonstration that PI3K inhibitors that also inhibit purified and cellular DNA-PKcs are radiosensitizers. 60. Boulton, S., Kyle, S. & Durkacz, B.W. Mechanisms of enhancement of cytotoxicity in etoposide and ionising radiation-treated cells by the protein kinase inhibitor wortmannin. Eur. J.Cancer 36, 535541 (2000). 61. Hardcastle, I.R. etal. Discovery of potent chromen4-one inhibitors of the DNA-dependent protein kinase (DNA-PK) using a small-molecule library approach. J.Med. Chem. 48, 78297846 (2005). 62. Shinohara, E.T. etal. DNA-dependent protein kinase is a molecular target for the development of noncytotoxic radiation-sensitizing drugs. Cancer Res. 65, 49874992 (2005). 63. Zhao, Y. etal. Preclinical evaluation of a potent novel DNA-dependent protein kinase inhibitor NU7441. Cancer Res. 66, 53545362 (2006). The first DNA-PKcs inhibitor that was demonstrated to induce chemosensitization in tumour-bearing mice. 64. Munck, J.M. etal. Chemosensitisation of cancer cells by KU-0060648; a dual inhibitor of DNA-PK and PI-3K. Mol. Can. Ther. 11, 17891798 (2012). 65. Willmore, E. etal. A novel DNA-dependent protein kinase inhibitor, NU7026, potentiates the cytotoxicity of topoisomerase II poisons used in the treatment of leukemia. Blood 103, 46594665 (2004). 66. Feeney, K.M., Wasson, C.W. & Parish, J.L. Cohesin: a regulator of genome integrity and gene expression. Biochem. J. 428, 147161 (2010). 67. Saleh-Gohari, N. etal. Spontaneous homologous recombination is induced by collapsed replication forks that are caused by endogenous DNA singlestrand breaks. Mol.Cell Biol. 25, 71587169 (2005). 68. Deans, A.J. & West, S.C. DNA interstrand crosslink repair and cancer. Nature Rev. Cancer 11, 467480 (2011). 69. Evers, B., Helleday, T. & Jonkers, J. Targeting homologous recombination repair defects in cancer. Trends Pharmacol.Sci. 31, 372380 (2010). 70. Dupr, A. etal. A forward chemical genetic screen reveals an inhibitor of the Mre11-Rad50-Nbs1 complex. Nature Chem. Biol. 4, 119125 (2008). 71. Rass, E. etal. Role of Mre11 in chromosomal nonhomologous end joining in mammalian cells. Nature Struct. Mol. Biol. 16, 819824 (2009). 72. Aloyz, R. etal. Imatinib sensitizes CLL lymphocytes to chlorambucil. Leukemia. 18, 409414 (2004). 73. Choudhury, A. etal. Targeting homologous recombination using imatinib results in enhanced tumor cell chemosensitivity and radiosensitivity. Mol. Cancer Ther. 8, 203213 (2009). 74. Dernheimer, F.A. & Kastan, M,B. Multiple roles of ATM in monitoring and maintaining DNA integrity. Febs Letts. 584, 36753681 (2010). 75. Kastan, M.B. & Bartek, J. Cell cycle checkpoints and cancer. Nature 432, 316323 (2004). 76. Bouwman, P. etal. 53BP1 loss resc

dsd

Documents

replication

homologous

nucleic acids

homologous

nucleotide

contextual

collapsed

base excision