dna damage response_nature

8/2/2019 DNA Damage Response_nature

http://slidepdf.com/reader/full/dna-damage-responsenature 1/8

1The Breakthrough Breast Cancer Research Centre, The Institute of Cancer Research, Fulham Road, London SW3 6JB, UK. *These authors contributed equally to this work.

Genomic instability is one of the most pervasive characteristics of tumour cells and is probably the combined effect of DNA damage, tumour-specific DNA repair defects, and a failure to stop or stall the cell cycle before the damaged DNA ispassed on to daughter cells. Although these processes drive genomic instability and ultimately the disease process, theyalso provide therapeutic opportunities. A better understanding of the cellular response to DNA damage will not only informour knowledge of cancer development but also help to refine the classification as well as the treatment of the disease.

The DNA damage response

and cancer therapyChristopher J. Lord1* & Alan Ashworth1*

Towards the end of his life, the German biologist Theodor Boveri(1862–1915) published a prophetic and much-quoted manu-script on the genetic basis of cancer1,2 that in contemporary

terms2 would be hailed as a great example of ‘translation’ or ‘impact’. Inhis study of chromosome and cell behaviour in the sea urchin, Boverisuggested that the pathogenesis of cancer might be driven by a “specificand abnormal chromosome constitution”1,2. A century later, genomicinstability is recognized as a characteristic of most solid tumours andadult-onset leukaemias. In most cancers, the instability is obvious asalterations in chromosome number and structure, a phenotype termedchromosomal instability, and as changes to the structure of DNA, suchas nucleotide substitutions, insertions and deletions. When they occurin crucial ‘driver’ genes (of which there are probably fewer than tenper tumour) these mutations can alter cell behaviour, confer a selectiveadvantage and drive the development of the disease. Importantly, thesemutations can also influence how the tumour will respond to therapy.Alongside key driver mutations, emerging data from cancer genomesequencing suggests that a typical tumour may contain many thousandsof other genetic changes. These ‘passenger’ mutations do not contributedirectly to the disease, but are probably collateral damage from exposureto various environmental factors or defects in the molecular mecha-nisms that maintain the integrity of the genome.

DNA damage and the DNA damage responseGiven the potentially devastating effects of genomic instability, cellshave evolved an intricate series of interlocking mechanisms that main-tain genomic integrity 3. The size of this task is daunting; the integrity of DNA is continually challenged by a variety of agents and processesthat either alter the DNA sequence directly or cause mutation when

DNA is suboptimally repaired. For example, the ultraviolet componentof sunlight can cause up to 1 × 105 DNA lesions per cell per day 4, many of which are pyrimidine dimers. If left unrepaired, dimers that containcytosine residues are prone to deamination, which can ultimately resultin cytosine being replaced with thymine in the DNA sequence. Likewise,ionizing radiation (for example, from sunlight or cosmic radiation) cancause single-strand breaks (SSBs) and double-strand breaks (DSBs) inthe DNA double helix backbone. If misrepaired — for example, theinaccurate rejoining of broken DNA ends at DSBs — these breakscan induce mutations and lead to widespread structural rearrange-ment of the genome. Environmental agents that cause DNA damageand mutation include cigarette smoke (40 years of heavy smoking isestimated to cause as many as 1,000 DNA lesions per cell5), industrial

chemicals, mustard gases used in warfare and a wide variety of drugsused in chemotherapy. This final group of agents has been the mainstay of cancer therapy for decades and includes drugs that induce covalentcrosslinks between DNA bases (for example, cisplatin, carboplatin,oxaliplatin and mitomycin C), drugs that attach alkyl groups to bases(such as methyl methanesulphonate and temozolomide) and agentsthat cause SSBs or DSBs by trapping topoisomerase I or II enzymes onDNA (such as camptothecin and etoposide)3. Endogenous processesalso present a challenge to the integrity of the genome. For example,normal cellular metabolism generates reactive oxygen species that canoxidize DNA bases and cause SSBs. Similarly, errors in DNA replicationcaused by deoxyribonucleoside 5´-triphosphate (dNTP) misincorpora-tion are potentially mutagenic, as is base depurination and deamination,an intrinsic chemical liability of DNA3.

The variety and frequency of DNA lesions are matched by thecomplexity of mechanisms that counteract these threats to genomicintegrity. Collectively, these mechanisms are known as the DNA dam-age response (DDR). In general terms, the DDR can be divided intoa series of distinct, but functionally interwoven, pathways, which aredefined largely by the type of DNA lesion they process (Fig. 1). MostDDR pathways encompass a similar set of tightly coordinated processes:namely the detection of DNA damage, the accumulation of DNA repairfactors at the site of damage and finally the physical repair of the lesion.Most of the subtle changes to DNA, such as oxidative lesions, alkylationproducts and SSBs, are repaired through a series of mechanisms thatis termed base excision repair (BER). In BER, damaged bases are firstremoved from the double helix, and the ‘injured’ section of the DNAbackbone is then excised and replaced with newly synthesized DNA6.Key to this process are the enzymes poly(ADP-ribose) polymerase 1 and

2 (PARP1 and PARP2), which act as sensors and signal transducers forlesions, such as SSBs. In addition to BER, the pool of deoxynucleotides(deoxyadenosine triphosphate (dATP), deoxythymidine triphosphate(dTTP), deoxyguanosine triphosphate (dGTP) and deoxycytidinetriphosphate (dCTP)) that provide the building blocks of DNA can bechemically modified before they are incorporated into the double helix.The nucleotide pool is, therefore, continually ‘sanitized’ by enzymessuch as nudix-type motif 5 (NUDT5)7.

Whereas small base adducts are repaired by BER, some of the bulkiersingle-strand lesions that distort the DNA helical structure, such asthose caused by ultraviolet light, are processed by nucleotide excisionrepair (NER)8. NER is often subclassified into transcription-cou-pled NER, which occurs where the lesion blocks, and is detected

1 9 J A N U A R Y 2 0 1 2 | V O L 4 8 1 | N A T U R E | 2 8 7

doi:10.1038/nature10760REVIEW

© 2012 Macmillan Publishers Limited. All rights reserved



by elongating RNA polymerase; and global-genome NER, in whichthe lesion is detected not as part of a blocked transcription process butbecause it disrupts base pairing and distorts the DNA helix. Althoughthese processes detect lesions using different mechanisms, they repairthem in a similar way: DNA surrounding the lesion is excised and thenreplaced using the normal DNA replication machinery. Excision repaircross-complementing protein 1 (ERCC1) is key to this excision step.

The major mechanisms that cope with DSBs are homologousrecombination9 and non-homologous end joining (NHEJ)10. Homol-ogous recombination acts mainly in the S and G2 phases of the cellcycle and is a conservative process in that it tends to restore the originalDNA sequence to the site of damage. Part of the DNA sequence aroundthe DSB is removed (known as resection) and the DNA sequence ona homologous sister chromatid is used as a template for the synthesisof new DNA at the DSB site. Crucial proteins involved in mediatinghomologous recombination include those encoded by the BRCA1,BRCA2, RAD51 and PALB2 genes. In contrast to homologous recom-bination, NHEJ occurs throughout the cell cycle. Rather than using ahomologous DNA sequence to guide DNA repair, NHEJ mediates repairby directly ligating the ends of a DSB together. Sometimes this processcan cause the deletion or mutation of DNA sequences at or around theDSB site. Therefore, compared with homologous recombination, NHEJ,

although mechanistically simpler, can often be mutagenic.Mismatch repair11 is crucial to the DDR. It deals primarily with dNTP

misincorporation and formation of ‘insertion and deletion’ loops thatform during DNA replication. These errors cause base ‘mismatches’ inthe DNA sequence (that is, non-Watson-Crick base pairing) that distortthe helical structure of DNA and so are recognized as DNA lesions. Therecognition of this distortion triggers a procession of events resulting inthe excision of newly synthesized DNA encompassing the mismatch siteand the resynthesis of DNA in its place. Key to the process of mismatchrepair are proteins encoded by the mutS and mutL homologue genes,such as MSH2 and MLH1.

Finally, translesion synthesis and template switching allow DNA tocontinue to replicate in the presence of DNA lesions that would oth-erwise halt the process. Translesion synthesis and template switching

are therefore usually considered to be part of the DDR. In translesion

synthesis, relatively high-fidelity DNA replication polymerases are tran-siently replaced with low-fidelity ‘translesion’ polymerases that are ableto synthesize DNA using a template strand encompassing a DNA lesion.Once the replication fork passes the site of the lesion, the low-fidelity DNA polymerases are normally replaced with the usual high-fidelity enzyme, which allows DNA synthesis to continue as normal. In templateswitching, the DNA lesion is bypassed at the replication fork by simply leaving a gap in DNA synthesis opposite the lesion. After the lesionhas passed the replication fork, the single-strand gap is repaired usingtemplate DNA on a sister chromatid, similar to the process used duringhomologous recombination.

Although sometimes considered distinct from the DDR, themechanisms that control the integrity of telomeric DNA at the end of each human chromosome also act as a barrier against genomic instabil-ity and mutation12. Rather than being an exposed DNA double-helixstructure at the end of the chromosome, telomeric DNA comprisesa series of guanine-rich, repetitive DNA sequences. These enable thetelomeric DNA to be bound in a loop-like structure with a series of proteins (telomere repeat binding factor 1 and 2 (TERF1 and TERF2),protection of telomeres 1 (POT1), TERF1-interacting nuclear factor2 (TINF2)) that form a shelterin complex. This ‘capping’ structureprevents the otherwise exposed ends of different chromosomes from

becoming fused together (a process known as end–end fusion) by theDDR. In most somatic cells, the length of telomeric DNA is reducedat each cell cycle, a process termed telomere attrition. Eventually, tel-omeres reach a crucial length that precludes the formation of an effec-tive shelterin complex. In normal cells, this failure in telomere cappinginduces a p53-mediated response that results in cellular senescence, amechanism that ultimately prevents unlimited cell proliferation. Whenthe p53 response is abrogated, end–end fusions occur, which leads tothe formation of chromosomes with two centromeres. At mitosis, eachcentromere in a dicentric chromosome attaches to an opposite spindlepole. The physical stress of opposing forces during chromosome segre-gation shears dicentric chromosomes, resulting in broken chromosomeends. This process is an ideal substrate for the chromosome translo-cation, focal DNA amplification and deletion events that potentially

drive tumorigenesis. Sheared chromosomes could also be the source of

Single-strandbreak

Double-strandbreak

Bulkyadducts

Base mismatches,insertions

and deletions

CH3

BER

Double-strand

break repair

Homologousrecombination

NHEJ NER Mismatchrepair

Direct reversal

PARP1

XRCC1

LIGASE 3

BRCA1

BRCA2

PALB2

ATM

CHEK1

CHEK2

RAD51

KU70/80

DNA-PK

ERCC4

ERCC1

MSH2

MLH1

MGMT

Breast, ovarian, pancreat ic Xerodermapigmentosa

Colorectal Gi loma

Proteins

PARP inhibitors, platinum salts Platinum salts Methotrexate Temozolomide

Tumour types

Drugs

Base alkylation

A

G

Figure 1 | A panoply of DNA repair mechanismsmaintains genomic stability. DNA is continually exposed to a series of insults that cause a range of lesions, from single-strand breaks (SSBs) to basealkylation events. The choice of repair mechanismis largely defined by the type of lesion, but factorssuch as the stage of the cell cycle also have a role.Key proteins involved in each DDR mechanism,the tumour types usually characterized by DDR

defects and the drugs that target these defects areshown. BER, base excision repair; NER, nucleotideexcision repair; NHEJ, non-homologous end- joining. Figure modified, with permission, from

ref. 72.

2 8 8 | N A T U R E | V O L 4 8 1 | 1 9 J A N U A R Y 2 0 1 2

REVIEWINSIGHT




the mutagenic ‘big bang’ or, chromothripsis13, that has been posited tooccur in some tumours. Whether or not this is the case, compared withhealthy cells, many pre-invasive and invasive cancers have significantly shorter telomeres, a situation in which end–end fusions and dicentricchromosome shearing could easily occur12.

The core DDR machinery does not work alone, but is coordinatedwith a set of complementary mechanisms that are also crucial to main-taining the integrity of the genome. For example, chromatin-remod-

elling proteins allow the DNA repair apparatus to gain access to thedamaged DNA14. Similarly, processes such as homologous recombina-tion are dependent on the close proximity of a sister chromatid, andprobably work in tandem with the chromosome cohesion machinery to ensure this is the case. Crucially, DDR core components interact withthe cell-cycle-checkpoint and chromosome-segregation machinery.These interactions allow DNA repair to occur before mitosis takes placeand ensure that the correct complement of genetic material is passedon to daughter cells15.

CancerOne of the most pervasive characteristics of human tumours is genomicinstability 16. Although the specific DDR defects are not known in mostcancers, there are several examples in which there is an incontrovertible

link between a particular DDR dysfunction and the neoplastic pheno-type. For example, 15% of sporadic colorectal tumours show an abnor-mal shortening or lengthening of dinucleotide repeat sequences. ThisDNA mutation pattern, known as microsatellite instability, is probably caused by an inability to repair DNA replication errors when mismatchrepair is defective. Microsatellite instability is observed not only in spo-radic colorectal tumours but also in a familial form of the disease knownas hereditary non-polyposis colorectal cancer (HNPCC). HNPCC isassociated with loss-of-function mutations in mismatch repair genes,such as MSH2 and MLH1

(ref. 11).Similarly, a subset of both sporadic and familial cancers seems to

be associated with defects in homologous recombination. Integrated(genomic, transcriptomic, methylomic) analysis of just under 500 sam-ples of high-grade serous ovarian adenocarcinoma has estimated thathomologous recombination is probably defective in approximately half.These homologous recombination defects are largely driven by DNAmutations or epigenetic silencing of genes such as BRCA1 and BRCA2

(ref. 17). Familial forms of breast, ovarian and pancreatic cancer are alsoassociated with loss-of-function mutations in homologous recombina-tion modifying genes, such as BRCA1, BRCA2, PALB2, ATM , RAD51C and RAD51D. A number of rare, syndromic conditions also exist inwhich the underlying genetic determinant is a loss-of-function muta-tion in a gene that modulates the response to DSBs9. These conditionsinclude ataxia telangiectasia (ATM), Bloom’s syndrome, Fanconi anae-mia, Rothmund–Thomson syndrome, Werner syndrome and Nijmegenbreakage syndrome, all of which have variable clinical phenotypes butare similar in that they cause extreme radiosensitivity, a characteristicof the inability to process DSBs effectively.

Evidence is also building that the DDR is not only invoked but also

dysfunctional at an early stage in the development of neoplasia. Markersof DSBs, such as nuclear γH2AX foci (a histone phosphorylation eventthat occurs on chromatin surrounding a DSB), are markedly elevatedin some precancerous lesions18,19. One hypothesis19 proposes that theoriginal cause of these effects is oncogene activation. The activationof oncogenes such as MYC and RAS stimulates the firing of multiplereplication forks as part of a proliferative program. These forks rap-idly stall, collapse and form DSBs because they exhaust the availabledNTP pool or because multiple forks collide on the same chromosome.Regardless of the mechanism, stalled and collapsed forks normally invoke the DDR and cell-cycle checkpoints that enable DNA lesionsto be repaired before mitosis takes place. For precancerous lesions toprogress to mature tumours, it is thought that critical DSB signal trans-duction and cell-cycle checkpoint proteins, such as ATM, ataxia telangi-

ectasia and Rad3 related (ATR), and the master ‘gatekeeper’ protein p53

become inactivated. With these DDR components rendered dysfunc-tional, collapsed forks are not effectively repaired, and cells proceedthrough the cell cycle with DNA lesions intact, increasing the chanceof mutagenesis18,19.

Selection of cancer treatmentHistorically, DDR defects have already been exploited in the form of many commonly used chemotherapy and radiotherapy regimens.

One example is the use of platinum salts (carboplatin or cisplatin)frequently given in combination with the taxane paclitaxel in patientswith advanced ovarian cancer. Platinum salts cause DNA inter- andintrastrand crosslinks — lesions that are recognized by the DDR andrepaired by a combination of NER and homologous recombination.These agents may be effective in patients with ovarian cancer becausethese tumours commonly harbour defects in homologous recombina-tion20, especially high-grade serous ovarian cancers. Estimates vary,but somewhere between 37% and 55% of high-grade serous ovariancancers have germ line or somatic mutations in BRCA1 or BRCA2 andsome have epigenetic inactivation of BRCA1

(ref. 21). Cisplatin alsoimproves survival in patients with ERCC1-negative non-small-celllung cancers, but not in those with ERCC1-positive tumours22. TheNER pathway, and in particular the ERCC1–ERCC4 complex, is essen-

tial for the removal of platinum adducts, and low levels of ERCC1 areassociated with an accumulation of adducts and an increased apop-totic response to cisplatin. Finally, where alkylating agents work, they probably also exploit a DDR defect. Normally, O6-methyltransferase(MGMT) directly reverses DNA alkylation events caused by drugssuch as temozolomide when they occur on the O6 position of guanineresidues. However, in some tumours, the promoter of the MGMT geneis methylated, causing a reduction in MGMT expression. This defectcauses O6 alkylation events formed by temozolomide to persist so thatthey eventually pair incorrectly with thymidine residues during replica-tion. The resultant DNA mismatch is recognized by the mismatch repairapparatus, which ultimately stimulates cell death. MGMT methylationhas also been implicated in the response of tumours (such as gliomas)to temozolomide, and is now being used as a biomarker to direct theuse of therapy 23 (Table 1).

An alternative approach to chemotherapy is the design of drugs thattarget specific DDR components. Topoisomerase inhibitors, such asirinotecan (a topoisomerase I inhibitor) and etoposide (a topoisomer-ase II inhibitor), could be regarded as the first generation of DDR ‘tar-geted’ agents24. Topoisomerases catalyse the breaking and rejoining of the phosphodiester backbone as part of the process that unwinds thetorsional structure of DNA before processes such as transcription andDNA replication; agents that inhibit this function, such as irinotecanand etoposide, leave DNA breaks across the genome and have beenused effectively in some cancers (Table 1). The development of the nextgeneration of DDR inhibitors, particularly PARP inhibitors, has givenconsiderable impetus to this area.

PARP inhibitors as targeted therapy

PARP1 and PARP2 are members of the PARP protein superfamily.These enzymes catalyse the polymerization of ADP-ribose moietiesonto target proteins (PARsylation) using NAD+ as a substrate, releas-ing nicotinamide in the process. This modification often modulatesthe conformation, stability or activity of the target protein25. The bestunderstood role of PARP1 is in SSB repair ((SSBR) a form of BER).PARP1 initiates this process by detecting and binding SSBs througha zinc finger in the PARP protein. Catalytic activity of PARP1 resultsin the PARsylation of PARP1 itself and the PARsylation of a series of additional proteins, such as XRCC1 and the histones H1 and H2B; whenPARP activity is inhibited, SSB repair is compromised25. The develop-ment of PARP inhibitors began with the observation that nicotinamide,a product of PARP catalytic activity, is itself a weak PARP inhibitor, asare nicotinamide analogues, such as 3-aminobenzamide. The screen-

ing of chemical libraries and the subsequent chemical refinement


REVIEW INSIGHT




of compounds have led to second generation PARP inhibitors (forexample, the benzimidazole-4-carboxamide NU1025), and to a thirdgeneration of clinically usable inhibitors, such as the tricyclic lactamindole, AG014699, and phthalazinones, such as AZD2281 (olaparib)26.

For cancer therapy, PARP inhibitors are potent chemosensitizers. Forexample, 3-aminobenzamide enhances the cytotoxic effect of the DNA-methylating agent dimethyl sulphate, and several preclinical studieshave shown that clinical PARP inhibitors increase the effects of anotherDNA-methylating agent, temozolomide25,26. In 2005, potent PARPinhibitors were shown to selectively inhibit the growth of cells withdefects in either the BRCA1 or BRCA2 genes, suggesting a new use forthese agents27,28. PARP inhibition probably works by allowing the per-sistence of spontaneously occurring SSBs, or by inhibiting PARP releasefrom a DNA lesion. Whichever is the case, both of these DNA lesiontypes could conceivably stall and collapse replication forks, potentially creating lethal DSBs29. In normal cells, the effects of PARP inhibitionare buffered by homologous recombination, which repairs the result-ant DSB. However, effective homologous recombination is reliant onfunctioning BRCA1 and BRCA2, so when these genes are defective — asthey are in tumours of germline BRCA-mutant carriers — DSBs are leftunrepaired, and potent PARP inhibitors can cause cell death. In somein vitro models, cells with BRCA mutations are more than 1,000 timesmore sensitive to PARP inhibitors than cells with wild-type alleles27.This genotype-specific selectivity provides a strong argument for theseagents to be tested clinically as single agents, rather than in combina-tion with chemotherapies. Moreover, the sensitivity to PARP inhibi-tors seems to be defined more by the BRCA genotype of a cancer cellthan by its tissue of origin. Breast, ovarian and prostate cancers withBRCA mutations all seem to be profoundly sensitive to these drugs. The

demonstration that BRCA-mutant cells are sensitive to PARP inhibi-tors has now become the emblematic ‘poster child’ for exploiting syn-thetic lethality in cancer therapy. Synthetic lethality, a genetic conceptdeveloped by Dobzhansky 30, Lucchesi31 and others and championed asa therapeutic approach for cancer by Hartwell and colleagues32 and Kae-lin33, is the process in which different single gene defects are compatiblewith cell viability, but the synthesis or combination of these gene defectsresults in cell death. In this case, BRCA dysfunction is synthetically lethal when PARP is inhibited. Furthermore, the demonstration thatone DDR pathway (homologous recombination) could be targeted by inhibition of another (PARP-mediated BER and SSB repair) suggeststhat DDR pathways functionally buffer one another. This hypothesishas long been established in model organisms (such as yeast) in whichsynthetic lethalities can be easily identified using genetic screens. These

studies and the BRCA/PARP paradigm in human tumour cells has

inspired the search for other DDR synthetic lethalities that could beexploited in the treatment of cancer, including other synthetic lethalapproaches for BRCA-mutant tumour cells34, as well as for tumourcells with other gene defects. For example, the mismatch repair genedefects found in colorectal cancer are synthetically lethal with inhibi-tion of genes controlling the response to oxidative DNA damage, suchas POLB

35 , POLG

35 and PINK1 (ref. 36). The first therapeutic test of thehypothesis that mismatch-repair defective cancers can be treated witha synthetic lethality approach37 is now in a proof-of-concept clinicaltrial (http://www.clinicaltrials.gov, NCT00952016).

The preclinical work described above has paved the way for subsequentclinical studies of PARP inhibitors (Table 1). Phase I studies have estab-lished the safety of olaparib, a potent PARP inhibitor, as a single agentand shown that significant and durable antitumour responses can beestablished in patients with BRCA-mutant breast, ovarian or prostatetumours38. Furthermore, olaparib does not seem to cause many of theside effects associated with standard chemotherapies. In subsequentphase II clinical studies, 40% of patients with breast or ovarian cancerwith germline BRCA mutations had a favourable response to the drug.This is a particularly high response given that the patients in these tri-als had been heavily pretreated and had become resistant to a rangeof chemotherapies. Such work has led to a series of additional PARPinhibitors being tested in clinical trials39.

Recently, attention has focused on clinical trials assessing iniparib.Unlike drugs such as olaparib and AG014699, iniparib was originally proposed to be a non-competitive inhibitor of PARP1 that disrupts theinteraction between PARP1 and DNA40. In a phase II trial, iniparibelicited a significant response in patients with triple negative (oestro-gen-receptor, progesterone-receptor and HER2-amplification nega-

tive) breast cancer when used in combination with gemcitabine andcarboplatin. However, a subsequent phase III trial failed to deliverthe predefined improvements in progression-free or overall survival.The reasons for this failure are unclear, although differences in thepatient populations and study design may have contributed. However,the clinical activity of iniparib may not be related to PARP inhibition41;this drug has very low PARP inhibition in vitro, and its mechanismof action in vivo is unclear40,41. We therefore believe that there is lit-tle evidence to regard the relative failure of iniparib in phase III trialsas a PARP-inhibitor class effect. In fact, the clinical observations withpotent catalytic-site PARP inhibitors have been promising when usedin appropriate populations39.

Alongside the clinical trials of PARP inhibitors, preclinical researchhas extended our understanding of how these agents work and how they

might be used most effectively in the clinic. Since the initial reports of

Table 1 | Examples of DDR inhibitors in clinical use or in development

Class Name Stage of development Cancer type

Topoisomerase Iinhibitor

Irinotecan Licensed for use Mainly colorectal cancers

Topoisomerase IIinhibitor

Etoposidephosphate

Licensed for use For example, testicular cancer, Ewing’s sarcoma, lung cancer, lymphoma,lymphocytic leukaemia and glioblastoma multiforme

DNA protein kinaseinhibitor

CC-115 Phase I study under way Advanced solid tumours, non-Hodgkin’s lymphoma and multiple myeloma

PARP inhibitor Olaparib 2 phase II studies completed, 9 additional

phase II studies under way

Ovarian, breast, gastric, colorectal and pancreatic tumours and a range of

other solid tumours

Veliparib 16 phase II studies under way Ovarian, breast, gastric, colorectal and pancreatic tumours and a range of

other solid tumours

Rucaparib 2 phase II studies recruiting Breast and other solid tumours

MK4827 Phase I study completed, phase Ibongoing

Glioblastoma multiforme, melanoma and other solid tumours andleukaemias

CHEK1 inhibitor AZD7762 Phase I study completed Range of solid tumours

MGMT inhibitor Lomeguatrib Phase I study under way Colorectal cancer and other solid tumours

Telomerase inhibitor Imetelstat 4 phase II studies under way Solid tumours and leukaemias


REVIEWINSIGHT




BRCA/PARP synthetic lethality, several other determinants of PARP-inhibitor sensitivity have been identified. As expected, most of thesedeterminants cause PARP-inhibitor sensitivity by modulating homolo-gous recombination (for example, RAD51, ATRX, SHFM1, RPA1, NBN,ATR, ATM, CHEK1, CHEK2, miR-182, CDK1, SWI5–SFR1, USP1/UAF1 and several Fanconi anaemia proteins42–47). Of particular interestis the observation that cyclin dependent kinase 1 (CDK1) inhibitionsuppresses homologous recombination and may therefore elicit PARP-

inhibitor sensitivity both in vitro and in vivo

46

. More extensive studiesare now required to identify the patient subgroup, and specifically thegenetic context, in which this combination could be used. In addition,mutations in the tumour-suppressor gene PTEN

48 and the TMPRSS2–

ERGfusion49 can sensitize cells to clinical PARP inhibitors; both of thesechanges are relatively common in human tumours so could be used asbiomarkers to direct the use of these agents. In the case of the PTEN/PARP synthetic lethality, one model suggests that PTEN deficiency causes a downregulation in RAD51 levels, therefore disrupting homol-ogous recombination49. Finally, several PARP-inhibitor-resistance-causing mechanisms have been identified, particularly in BRCA1- orBRCA2-mutant tumour models. In two studies, PARP-inhibitor resist-ance was caused by secondary mutations in the BRCA1 or BRCA2 genesthemselves50,51. These mutations restored the open reading frame of theBRCA1

orBRCA2

gene, enabled the generation of functional proteinsand restored the ability to repair DNA damage caused by PARP inhibi-tors. Importantly, these secondary mutations were observed not only in in vitro models but also in ovarian tumours from patients who hadbecome resistant to platinum salts such as carboplatin50–53. These drugstarget cells with deficient homologous recombination, albeit with lessefficiency than PARP inhibitors. Subsequent studies have suggestedthat these reversion events are relatively common in BRCA-mutantcancers that do not respond to platinum salts. This is especially truein ovarian cancer, in which more than 42% of the carboplatin-resistanttumours tested had secondary mutations that restored the BRCA1 orBRCA2 open reading frames54. Taken together, these observations couldhelp to explain why some platinum-refractory BRCA-mutant patientsin PARP-inhibitor clinical trials failed to respond and strengthen theargument for the inclusion of biomarkers that monitor BRCA1 andBRCA2 sequence and homologous recombination function in futuretrials. Furthermore, the restoration of homologous recombination anddrug resistance by secondary mutations provides strong confirmationthat, at least in this setting, PARP inhibitors and platinum salts requiredefective homologous recombination for antitumour activity. Otheridentified mechanisms of PARP-inhibitor response include the upregu-lation of genes that encode P-glycoprotein efflux pumps55 and the lossof TP53BP1 (a DDR protein) in BRCA1-deficient cells, suggesting thatthese mechanisms should also be assessed in clinical biopsies56,57.

In addition to dissecting the mechanisms behind PARP-inhibitorsensitivity and resistance, there has been considerable debate regard-ing the potential of PARP inhibitors in non-familial cancers. Germlinemutations in BRCA1 or BRCA2 are common in hereditary breastcancer, but relatively rare in sporadic breast cancer. However, many

sporadic tumours have molecular and histopathological features incommon with BRCA-deficient tumours, a concept originally termedBRCAness20. For example, BRCAness may be caused by methylationof the BRCA1 gene promoter (which causes epigenetic silencing of BRCA1 expression) or even somatic mutation of genes that modulatehomologous recombination. Furthermore, traditional histopathologi-cal methods and gene-expression profiling approaches have suggestedsome level of phenotypic or molecular overlap between triple-negativebreast cancers, basal-like breast cancers and BRCA1-mutant familialbreast cancers20. Although the overlap is not absolute, it has led to thehypothesis that there may be a subset of sporadic breast cancers thatexhibit features of BRCAness and may respond favourably to PARPinhibitors20. This hypothesis has yet to be fully tested experimentally,in part because existing tumour-cell models of triple-negative breast

cancer do not encompass the molecular and functional heterogeneity

of triple-negative breast cancer. In the clinical setting, patients withtriple-negative breast cancer do not show a frequent response to eitherolaparib or veliparib39, although the number of patients tested is stillsmall. Needless to say, larger studies are needed, as are clinical trialsthat lead to a full drug licence for PARP inhibitors in BRCA-mutation-related breast cancer.

Future issues

More research is needed into how DDR activity affects the clinicalefficacy of DNA-damaging chemotherapies, and here we highlight someareas that need to be addressed in the future.

Pharmacological targeting Although the DDR has been the subject of intense study, the biochemicaland genetic interactions between the various DNA repair pathways isnot completely understood. This understanding would be facilitatedby the availability of potent and specific DDR-pathway small-moleculeinhibitors. Past efforts have focused on the protein kinases that trans-duce DDR signals (such as ATM, ATR and DNA-PK) and on the PARPsuperfamily members that detect damage and act as signal transductionmediators. For example, synthetic lethalities with ATM and DNA-PKsuggest that ATM inhibitors could increase the potential of DNA-dam-

aging chemotherapy inTP53

-mutant tumours, and DNA-PK inhibi-tors could be used in combination with chemotherapy in ATM -mutanttumours58. A number of other enzyme classes have also been identifiedas having a role in the DDR, including enzymes that mediate SUMOyla-tion, histone modification, acetylation and ubiquitylation59. The avail-ability of pharmacological inhibitors of these proteins would facilitatethe search for additional synthetic lethal interactions. An example isthe identification of small-molecule inhibitors (such as MLN4924) thatmodulate neddylation, a post-translational modification process analo-gous to ubiquitylation60. Inhibition of neddylation, which is induced by MLN4924, causes stabilization of the DNA-replication licensing factorCDT1, DNA synthesis in the absence of cell-cycle progression and acti-

vation of the DDR60. These agents could potentially target tumour cellsthat have existing DNA repair defects or are under replicative stress.

The clinical use of DDR-pathway drugs remains a challenge. Additionof such agents to standard chemotherapy regimens will probably lead togreater side effects, and therefore more work in this area is needed on theappropriate dosing and scheduling of inhibitors. These studies shouldalso address the possibility that targeting the DDR could also increasethe rate of second cancers or leukaemias, something that has already been observed for some chemotherapies. In principle, these iatrogeniceffects might be avoided if DDR inhibitors could be administered fora relatively short duration, a hypothesis that should be tested clinically as well as in preclinical models. Finally, combining DDR inhibitors withother targeted agents, such as signal transduction inhibitors or targetedradiation, has been underexplored and could form the basis of effectivetreatment regimens.

Cancer stem cells and the tumour micro-environment

Many human cancers contain cells that share some of the self-renewaland differentiation properties of stem cells. These cancer stem cellsare thought to act as tumour-initiating cells that, after treatment, may potentially repopulate the therapy-resistant tumour. With this role inmind, understanding the response to DNA-damaging therapy in cancerstem cells may prove essential in understanding whether a treatmentwill ultimately work. Evidence is accumulating that the DDR has dis-tinct properties in cancer stem cells. These may include controversialand unproven concepts such as the ‘immortal strand’ model in whichparental, but not newly synthesized DNA, is retained in stem cells aftermitosis; retaining only the parental strand of DNA is thought to limitthe impact of DNA replication errors to the stem cell population61.Whether this is the case or not, a number of studies have suggestedthat resistance to therapy is mediated by altered DDR activity in cancer

stem cells and tumour-initiating cells. For example, a comparison of


REVIEW INSIGHT




breast tumour biopsies before and after chemotherapy treatment hasshown an enrichment of cancer stem cells, which suggests an intrinsicresistance of these cells. In mouse models, cancer stem cells deficientin p53 showed accelerated DNA repair activity 62, as well as high levelsof AKT and WNT signalling, both of which promoted cancer-stem-cell survival after exposure to ionizing radiation 63. In glioblastomamultiforme, cancer stem cells have been characterized on the basis of expression of prominin (CD133; now known as PROM1)64. CD133+

cells are enriched after irradiation of human glioblastoma multiformexenografts, suggesting that cancer stem cells are necessary for tumourregrowth after radiotherapy. Cancer stem cells showed more robustactivation of the DDR and higher intrinsic DNA repair activity. Block-ing the kinases CHEK1 and CHEK2 led to enhanced killing of cancerstem cells by ionizing radiation65, suggesting that a pharmacological

approach may overcome this intrinsic resistance. Much more detailed

and systematic knowledge of the consequences of DNA damage andthe DDR in stem cells is required to exploit this important area.

If we are to optimize therapies that exploit the DDR, then the impactof the tumour microenvironment and the non-tumour cells that sitwithin a neoplastic lesion must be considered. For example, non-tumour stromal cells form at least half of the cellular complement of pancreatic ductal adenocarcinomas. This stromal compartment con-tains pancreatic stellate cells, which, in in vitro systems at least, cause

tumour cells to become resistant to ionizing radiation; this radiopro-tective effect could explain the relative radioresistance of pancreaticductal adenocarcinomas. Already, there is a preliminary understandingof how stromal and tumour cells interact and communicate to protectagainst ionizing radiation66, but the understanding of how the micro-environment modulates the DDR is still limited. Given the complexity of studying molecular mechanisms in vivo and of effectively model-ling cell–cell interactions in vitro, this is not surprising. These techni-cal problems are not insurmountable, but they do raise an importantissue: historically, much of our knowledge of the DDR has originatedfrom studies of cells in vitro. Understanding the DDR in vivo is criticalif cancer therapies are to be modelled correctly.

The genomic scar

Historically, cancer taxonomy has mainly been based on the anatomicalsite of a tumour and its clinical and histopathological features. Cancersare now increasingly classified according to their underlying pattern of pathogenic mutations, a process which is being facilitated by exome orwhole-genome sequencing of each cancer67. Treatment decisions canthen be made on the basis of pharmacologically targetable drivers of the disease. In addition to identifying the driver genes affected, exam-ining the collateral genomic damage caused by DNA repair deficiency may help in the classification of cancer and could provide biomarkersto inform therapy. In essence, the mutational spectrum of a tumour,revealed by its DNA sequence, is a sequence ‘scar’ that reflects boththe mutagens that a tumour has been exposed to and the repair pro-cesses that have operated to mitigate their impact67. Already, there aresome precedents to suggest that this approach could have merit. Forexample, homologous-recombination deficient tumours have a char-acteristic mutational spectrum defined by the use of error-prone DSBrepair mechanisms68,69. Therefore, the wider tumour genome sequenceor sequence scar in a tumour with defective homologous recombina-tion might be used to decide which patients should be treated withPARP inhibitors. Microsatellite instability is another example. In many cases, microsatellite instability in a tumour reveals both the mutagenicnature of DNA replication and the failure to reverse these sequenceerrors when mismatch repair is defective11. Given the impact that theDDR can have on both disease development and response to treatment,it seems reasonable to consider categorizing tumours according to theirDNA repair defects and to use this information to personalize treat-ment (Fig. 2). As a result of the speed at which in-depth, whole-genometumour sequences are now being generated, we will soon have a firstdraft of the mutational landscape of all human cancers, which will allow

these possibilities to be addressed.To match these molecular signatures of DNA mutation and repair,

we require functional biomarkers that identify DDR defects and aidin the selection of therapy. Again, there are preliminary signs that thiscould be achievable: a mechanistic determinant of PARP-inhibitorsensitivity, RAD51 foci deficiency, can be measured by immunohis-tochemical staining in tumour biopsies70, and the ability to detectother DNA repair complexes by immunohistochemistry is improv-ing71. In principle, functional biomarkers should be identifiable formost DDR pathways even though significant technical and practicalissues need to be resolved before such tests can be used in the clinic.Because the DDR pathway is fully engaged only after DNA is dam-aged, standard tumour biopsies or specimens are not particularly useful for measurement of DDR-pathway activity. A functional test

of DDR-pathway deficiency might therefore need to include some

Transitions

Transversions

Small and simple mutations

Insertions

Deletions

Frequency, genome position

and sequence context of each

mutation type

Large and complex mutations

Amplication/gene duplication

Large deletions

TranslocationsChromothripsis or

chromoanasynthesis-derived

complex lesions

Loss of heterozygosity

0

20

40

60

80

M i c r o s a t e l l i t e

i n s t a b i l i t y

T r a n s i t i o n

T r a n s v e r s i o n

A m p l i c a t i o n

T r a n s l o c a t i o n

D e l e t i o n a n k e d

b y t a n d e m r

e p e a t s

0

20

40

60

T r a n s i t i o n

T r a n s v e r s i o n

A m p l i c a t i o n

T r a n s l o c a t i o n

D e l e t i o n a n k e d

b y t a n d e m r

e p e a t s

Tumour biopsy Genome sequencing

Patient A Patient B

Treatment 1 Treatment 2

X

M i c r o s a t e l l i t e

i n s t a b i l i t y

Prognosis and

treatment options

Treatment 1 Treatment 2

X

Prognosis and

treatment options

Tumour classication

Identication of DNA defect

Figure 2 | Genomic scars in cancer. Modern sequencing approaches meanthat the nucleotide sequence and mutational spectrum of tumour biopsiescan be rapidly determined. This spectrum, or sequence scar, could be usedto classify each tumour. Classifiers could take into account the frequency and location of mutations, as well as their predominant sequence context(for example, insertions and deletions that are flanked by small regions of microhomology 13,73). Such classifiers could then be used to estimate theprobable outcome for the patient and to select the most appropriate therapy.The choice of therapy is defined by the balance of sequence scars found intumour genome biopsies.


REVIEWINSIGHT




form of genotoxic stress, followed by a measurement of the molecularresponse. For example, to test the homologous recombination defectin tumour cells by measuring the RAD51 response, cells must firstbe exposed to an agent that causes DNA damage. These agents couldinclude ionizing radiation, chemotherapy or the PARP inhibitoritself. Clinically, this could mean that patients need to be treatedwith a DNA-damaging agent before biopsy for the purposes of bio-marker assessment or alternatively tumour biopsies would have to

be exposed to DNA-damaging agents ex vivo. Both scenarios havepractical or logistical issues. For instance, tumour cells may notrespond in the same way ex vivo as they do in vivo, and pretreatinga patient to assess tumour response might not always be feasiblefrom a clinical perspective. Nevertheless, molecular and functionalassays could mean that a tumour could be defined as, for example,‘deficient for homologous recombination and potentially sensitive toPARP inhibitor’ before treatment rather than simply being classif iedas a ‘non-familial, triple-negative breast cancer’. The development of functional biomarkers of other DDR pathways would considerably advance this area.

The genomic instability of tumour cells is one of their most pervasivecharacteristics and presents a therapeutic opportunity. Already, DNA-damaging chemotherapies have provided the core of cancer treatment

for the past half-century and, if used appropriately, targeted therapiesthat exploit the DDR could deliver better therapeutic responses in thefuture. Nevertheless, there are still significant unexplored territories forthe field; an understanding of the DDR in vivo, the definition of clinicalbiomarkers that allow the selection of appropriate therapy and an under-standing of how the DDR interacts with the other homeostatic systemsin the cell are relative unknowns. We believe that the crucial factors thathave previously allowed this field to progress — the comprehensive study of the basic biology of the DDR combined with a desire to apply thisinformation in the clinic — will drive significant advances in the future. ■

1. Boveri, T. Zur Frage der Entstehung Maligner Tumoren (Gustav Fischer,1914).

2. Boveri, T. Concerning the origin of malignant tumours by Theodor Boveri.Translated and annotated by Henry Harris. J. Cell Sci. 121, 1–84 (2008).A translation of Boveri’s original work that proposed the hypothesis that a

form of genomic dysregulation characterizes tumour cells.3. Ciccia, A. & Elledge, S. J. The DNA damage response: making it safe to play

with knives. Mol. Cell 40, 179–204 (2010).4. Hoeijmakers, J. H. DNA damage, aging, and cancer. N. Engl. J. Med. 361,

1475–1485 (2009).5. Phillips, D. H., Hewer, A., Martin, C. N., Garner, R. C. & King, M. M. Correlation

of DNA adduct levels in human lung with cigarette smoking. Nature 336, 790–792 (1988).

6. David, S. S., O’Shea, V. L. & Kundu, S. Base-excision repair of oxidative DNAdamage. Nature 447, 941–950 (2007).

7. Nakabeppu, Y. et al. Mutagenesis and carcinogenesis caused by theoxidation of nucleic acids. Biol. Chem. 387, 373–379 (2006).

8. Cleaver, J. E., Lam, E. T. & Revet, I. Disorders of nucleotide excision repair:the genetic and molecular basis of heterogeneity. Nature Rev. Genet. 10, 756–768 (2009).

9. Moynahan, M. E. & Jasin, M. Mitotic homologous recombination maintainsgenomic stability and suppresses tumorigenesis. Nature Rev. Mol. Cell Biol. 11, 196–207 (2010).

10. Lieber, M. R. NHEJ and its backup pathways in chromosomal translocations.Nature Struct. Mol. Biol. 17, 393–395 (2010).

11. Jiricny, J. The multifaceted mismatch-repair system. Nature Rev. Mol. CellBiol 7, 335–346 (2006).

12. Artandi, S. E. & DePinho, R. A. Telomeres and telomerase in cancer.Carcinogenesis 31, 9–18 (2010).

13. Stephens, P. J. et al. Massive genomic rearrangement acquired in a singlecatastrophic event during cancer development. Cell 144, 27–40 (2011).

14. Bell, O., Tiwari, V. K., Thoma, N. H. & Schubeler, D. Determinants anddynamics of genome accessibility. Nature Rev. Genet. 12, 554–564 (2011).

15. Warmerdam, D. O. & Kanaar, R. Dealing with DNA damage: relationshipsbetween checkpoint and repair pathways. Mutat. Res. 704, 2–11 (2010).

16. Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation.Cell 144, 646–674 (2011).

17. Cancer Genome Atlas Research Network. Integrated genomic analyses ofovarian carcinoma. Nature 474, 609–615 (2011).

18. Bartkova, J. et al. Oncogene-induced senescence is part of thetumorigenesis barrier imposed by DNA damage checkpoints. Nature 444, 633–637 (2006).

19. Halazonetis, T. D., Gorgoulis, V. G. & Bartek, J. An oncogene-induced DNA

damage model for cancer development. Science 319, 1352–1355 (2008).References 18 and 19 describes how oncogenes are able to induce theDDR early in oncogenesis.

20. Turner, N., Tutt, A. & Ashworth , A. Hallmarks of ‘BRCAness’ in sporadiccancers. Nature Rev. Cancer 4, 814–819 (2004).

21. Banerjee, S., Kaye, S. B. & Ashworth, A. Making the best of PARP inhibitorsinovarian cancer. Nature Rev. Clin. Oncol. 7, 508–519 (2010).

22. Friboulet, L. et al. Molecular characteristics of ERCC1-negative versusERCC1-positive tumors in resected NSCLC. Clin. Cancer Res. 17, 5562–5572 (2011).

23. Weller, M. et al. MGMT promoter methylation in malignant gliomas: ready

for personalized medicine? Nature Rev. Neurol. 6, 39–51 (2010).24. Pommier, Y., Leo, E., Zhang, H. & Marchand, C. DNA topoisomerases and

their poisoning by anticancer and antibacterial drugs. Chem. Biol. 17, 421–433 (2010).

25. Rouleau, M., Patel, A., Hendzel, M. J., Kaufmann, S. H. & Poirier, G. G. PARPinhibition: PARP1 and beyond. Nature Rev. Cancer 10, 293–301 (2010).

26. Zaremba, T. & Curtin, N. J. PARP inhibitor development for systemic cancertargeting. Anticancer Agents Med. Chem. 7, 515–523 (2007).

27. Farmer, H. et al. Targeting the DNA repair defect in BRCA mutant cells as atherapeutic strategy. Nature 434, 917–921 (2005).

28. Bryant, H. E. et al. Specific killing of BRCA2-deficient tumours withinhibitors of poly(ADP-ribose) polymerase. Nature 434, 913–917 (2005).References 27 and 28 describe the synthetic lethal interaction betweenBRCA mutations and PARP inhibition.

29. Lord, C. J. & Ashworth, A. Targeted therapy for cancer using PARPinhibitors. Curr. Opin. Pharmacol. 8, 363–369 (2008).

30. Dobzhansky, T. Genetics of natural populations. Xiii. Recombinationand variability in populations of drosophila pseudoobscura. Genetics 31, 269–290 (1946).

31. Lucchesi, J. C. Synthetic lethality and semi-lethality among functionallyrelated mutants of Drosophila melanogaster . Genetics 59, 37–44 (1968).

32. Hartwell, L. H., Szankasi, P., Roberts, C. J., Murray, A. W. & Friend, S. H.Integrating genetic approaches into the discovery of anticancer drugs.Science 278, 1064–1068 (1997).

33. Kaelin, W. G. Jr. Synthetic lethality: a framework for the development ofwiser cancer therapeutics. Genome Med. 1, 99 (2009).

34. Issaeva, N. et al. 6-Thioguanine selectively kills BRCA2-defective tumorsand overcomes PARP inhibitor resistance. Cancer Res. 70, 6268–6276(2010).

35. Martin, S. A. et al. DNA polymerases as potential therapeutic targets forcancers deficient in the DNA mismatch repair proteins MSH2 or MLH1.Cancer Cell 17, 235–248 (2010).

36. Martin, S. A., Hewish, M., Sims, D., Lord, C. J. & Ashworth, A. Parallelhigh-throughput RNA interference screens identify PINK1 as a potentialtherapeutic target for the treatment of DNA mismatch repair-deficientcancers. Cancer Res. 71, 1836–1848 (2011).

37. Martin, S. A. et al. Methotrexate induces oxidative DNA damage and is

selectively lethal to tumour cells with defects in the DNA mismatch repairgene MSH2. EMBO Mol. Med. 1, 323–337 (2009).

38. Fong, P. C. et al. Inhibition of poly(ADP-ribose) polymerase in tumors fromBRCA mutation carriers. N. Engl. J. Med. 361, 123–134 (2009).This article describes the results of the first clinical trial to demonstratethe clinical potential of a synthetic lethal approach for therapy.

39. Balmaña, J., Domchek, S. M., Tutt, A. & Garber, J. E. Stumbling blocks on thepath to personalized medicine in breast cancer: the case of PARP inhibitorsfor BRCA1/2-associated cancers. Cancer Discov. 1, 29–34 (2011).

40. Mendeleyev, J., Kirsten, E., Hakam, A., Buki, K. G. & Kun, E. Potentialchemotherapeutic activity of 4-iodo-3-nitrobenzamide. Metabolic reductionto the 3-nitroso derivative and induction of cell death in tumor cells inculture. Biochem. Pharmacol. 50, 705–714 (1995).

41. Liu, X. et al. Iniparib non-selectively modifies cysteine-containing proteinsin tumor cells and is not a bona fide PARP inhibitor. Clin. Cancer Res. advance ahead of print http://dx.doi.org/10.1158/1078-0432.CCR-11-1973 (29 November 2011).

42. McCabe, N. et al. Deficiency in the repair of DNA damage by homologousrecombination and sensitivity to poly(ADP-ribose) polymerase inhibition.Cancer Res. 66, 8109–8115 (2006).

43. Peasland, A. et al. Identification and evaluation of a potent novel ATRinhibitor, NU6027, in breast and ovarian cancer cell lines. Br. J. Cancer 105, 372–381 (2011).

44. Murai, J. et al. The USP1/UAF1 complex promotes double-strand breakrepair through homologous recombination. Mol. Cell. Biol. 31, 2462–2469(2011).

45. Moskwa, P. et al. miR-182-mediated downregulation of BRCA1 impacts DNArepair and sensitivity to PARP inhibitors. Mol. Cell 41, 210–220 (2011).

46. Johnson, N. et al. Compromised CDK1 activity sensitizes BRCA-proficientcancers to PARP inhibition. Nature Med. 17, 875–882 (2011).

47. Akamatsu, Y. & Jasin, M. Role for the mammalian Swi5−Sfr1 complex inDNA strand break repair through homologous recombination. PLoS Genet. 6, e1001160 (2010).

48. Mendes-Pereira, A. M. et al. Synthetic lethal targeting of PTEN mutant cellswith PARP inhibitors. EMBO Mol. Med. 1, 315–322 (2009).

49. Brenner, J. C. et al. Mechanistic rationale for inhibition of poly(ADP-ribose)polymerase in ETS gene fusion-positive prostate cancer. Cancer Cell 19,

664–678 (2011).


REVIEW INSIGHT




An important paper suggesting that ETS gene fusions may confer PARP-inhibitor sensitivity in prostate cancer.

50. Edwards, S. L. et al. Resistance to therapy caused by intragenic deletion inBRCA2. Nature 451, 1111–1115 (2008).

51. Sakai, W. et al. Secondary mutations as a mechanism of cisplatin resistancein BRCA2-mutated cancers. Nature 451, 1116–1120 (2008).References 50 and 51 are the first demonstrations of synthetic lethalresistance to PARP inhibitors and platinum drugs in BRCA-mutant cells.

52. Sakai, W. et al. Functional restoration of BRCA2 protein by secondary BRCA2 mutations in BRCA2-mutated ovarian carcinoma. Cancer Res. 69, 6381–6386(2009).

53. Swisher, E. M. et al. Secondary BRCA1 mutations in BRCA1-mutated ovariancarcinomas with platinum resistance. Cancer Res. 68, 2581–2586 (2008).

54. Norquist, B. et al. Secondary somatic mutations restoring BRCA1/2 predictchemotherapy resistance in hereditary ovarian carcinomas. J. Clin. Oncol. 29, 3008–3015 (2011).

55. Rottenberg, S. et al. High sensitivity of BRCA1-deficient mammary tumors tothe PARP inhibitor AZD2281 alone and in combination with platinum drugs.Proc. Natl Acad. Sci. USA 105, 17079–17084 (2008).

56. Bouwman, P. et al. 53BP1 loss rescues BRCA1 deficiency and is associated withtriple-negative and BRCA-mutated breast cancers. Nature Struct. Mol. Biol. 17, 688–695 (2010).

57. Bunting, S. F. et al. 53BP1 inhibits homologous recombination in Brca1-deficient cells by blocking resection of DNA breaks. Cell 141, 243–254 (2010).

58. Jiang, H. et al. The combined status of ATM and p53 link tumor developmentwith therapeutic response. Genes Dev. 23, 1895–1909 (2009).

59. Galanty, Y.et al. Mammalian SUMO E3-ligases PIAS1 and PIAS4 promoteresponses to DNA double-strand breaks. Nature 462, 935–939 (2009).

60. Soucy, T. A. et al. An inhibitor of NEDD8-activating enzyme as a new approachto treat cancer. Nature 458, 732–736 (2009).This paper provides proof of the concept that targeting additional elementsof the DDR and replication machinery, such as neddylation complexes, couldprovide an alternative approach to cancer therapy.

61. Lansdorp, P. M. Immortal strands? Give me a break. Cell 129, 1244–1247(2007).

62. Diehn, M. et al. Association of reactive oxygen species levels and radioresistancein cancer stem cells. Nature 458, 780–783 (2009).

63. Zhang, M., Atkinson, R. L. & Rosen, J. M. Selective targeting of radiation-resistanttumor-initiating cells. Proc. Natl Acad. Sci. USA 107, 3522–3527 (2010).

64. Singh, S. K. et al. Identification of human brain tumour initiating cells. Nature 432, 396–401 (2004).

65. Bao, S. et al. Glioma stem cells promote radioresistance by preferentialactivation of the DNA damage response. Nature 444, 756–760 (2006).

66. Mantoni, T. S., Lunardi, S., Al-Assar, O., Masamune, A. & Brunner, T. B. Pancreaticstellate cells radioprotect pancreatic cancer cells through beta 1-integrinsignaling. Cancer Res. 71, 3453–3458 (2011).

67. Stratton, M. R. Exploring the genomes of cancer cells: progress and promise.Science 331, 1553–1558 (2011).

68. Stephens, P. J. et al. Complex landscapes of somatic rearrangement in humanbreast cancer genomes. Nature 462, 1005–1010 (2009).

69. Barber, L. J. et al. Comprehensive genomic analysis of a BRCA2 deficient humanpancreatic cancer. PLoS ONE 6, e21639 (2011).

70. Graeser, M. et al. A marker of homologous recombination predicts pathologiccomplete response to neoadjuvant chemotherapy in primary breast cancer.Clin. Cancer Res. 16, 6159–6168 (2010).

71. Polo, S. E. & Jackson, S. P. Dynamics of DNA damage response proteins at DNAbreaks: a focus on protein modifications. Genes Dev. 25, 409–433 (2011).

72. Hoeijmakers, J. H. Genome maintenance mechanisms for preventing cancer.Nature 411, 366–374 (2001).

73. Liu, P. et al. Chromosome catastrophes involve replication mechanismsgenerating complex genomic rearrangements. Cell 146, 889–903 (2011).

Acknowledgements We thank Cancer Research UK, The Wellcome Trust,Breakthrough Breast Cancer, AACR, The Komen Foundation, The Breast CancerResearch Foundation, the European Union and The Breast Cancer Campaignfor funding the work in our laboratory. We also acknowledge NHS funding to theRoyal Marsden Hospital NIHR Biomedical Research Centre.

Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare competing financial interests: details accompanythe full-text HTML version of the paper at www.nature.com/nature. Readers arewelcome to comment on the online version of this article at www.nature.com/nature.Correspondence should be addressed to C.J.L. ([email protected]) or A.A.([email protected])


REVIEWINSIGHT

dna damage response_nature

Documents