Amir Sonnenblick, Evandro de Azambuja, Hatem A. Azim Jr and Martine Piccart

An update on inhibitors moving to the adjuvant setting Amir Sonnenblick, Evandro de Azambuja, Hatem A. Azim Jr and Martine Piccart REVIEWS Abstract Inhibition of poly(adp-ribose) polymerase () enzymes is a potential synthetic lethal therapeutic strategy in cancers harbouring specific DNA-repair defects, including those arising in carriers of BRCA1 or BRCA2 mutations. Since the development of first-generation inhibitors more than a decade ago, numerous clinical trials have been performed to validate their safety and efficacy, bringing us to the stage at which adjuvant therapy with inhibitors is now being considered as a viable treatment option for patients with breast cancer. Nevertheless, the available data do not provide clear proof that these drugs are efficacious in the setting of metastatic disease. Advancement of a therapy to the neoadjuvant and adjuvant settings without such evidence is exceptional, but seems reasonable in the case of inhibitors because the target population that might benefit from this class of drugs is small and well defined. This Review describes the evolution of inhibitors from bench to bedside, and provides an up-to-date description of the key published or otherwise reported clinical trials of these agents. The specific considerations and challenges that might be encountered when implementing these compounds in the adjuvant treatment of breast cancer in the clinic are also highlighted. Sonnenblick, A. et al. Nat. Rev. Clin. Oncol. advance online publication 7 October 2014; doi:10.1038/nrclinonc.2014.163 BrEAST Data Centre, Department of Medicine, Institut Jules Bordet, Université Libre de Bruxelles, Boulevard de Waterloo 125, B 1000 Brussels, Belgium (A.S., E.d.A., H.A.A., M.P.). Correspondence to: M.P. martine.piccart@ bordet.be Introduction The poly(adp-ribose) polymerases (s) are a large family of enzymes with diverse functions, some of which are important for the repair of single-strand breaks (SSBs) in DNA via the base-excision repair (BER) pathway. 1 The role of s in DNA repair prompted the investigation of inhibitors as potential treatments for cancer, a disease in which DNA replication and, therefore, replication errors are prominent features, and deficiencies in DNA-repair pathways are common. In 2005, two simultaneously published seminal articles 2,3 demonstrated that breast cancer type 1 susceptibility protein (BRCA1) or breast cancer type 2 susceptibility protein (BRCA2) dysfunction resulted in a lack of homologous recombination and markedly sensitized cells to inhibition of enzymatic activity. In this context, inhibition caused chromosomal instability, cell-cycle arrest, and subsequent apoptosis, which seemed to be attributable to the persistence of DNA lesions that are normally repaired by homologous recombination. These findings illustrate that different pathways cooperate to maintain DNA integrity. In addition, these results exemplify the concept of synthetic lethality, which arises when simultaneous inhibition or ablation of two cellu lar pathways leads to cell death, whereas inhibition of either one of the two pathways alone does not. 2 Competing interests E.d.A. holds an advisory role for and has received research funding from GlaxoSmithKline. H.A.A. is a consultant for Celgene, Nanostring and Novartis, and has received honouraria from Celgene, GlaxoSmithKline, Nanostring and Novartis. M.P. is a consultant to, and has received honouraria and grant support from, Amgen, Bayer, Boehringer, Bristol-Myers Squibb, Roche and Sanofi; M.P. has also received grant support from Pfizer and honouraria from AstraZeneca. A.S. declares no competing interests. An obvious next step in development of -inhibitor therapies was validation of the synthetic lethality paradigm in patients with germline mutations in a BRCA1 or BRCA2 allele, who are known to have increased susceptibility to the development of cancers particularly of the breasts and ovaries. In addition, the use of inhibitors was predicted to be especially relevant to the treatment of triplenegative breast cancer (TNBC), which accounts for 15 20% of all breast cancers. 4 The potential utility of inhibitors as a targeted therapy in this setting was hypothesized based on the fact that TNBC shares multiple clinico pathological features with BRCA-mutated breast cancer, and harbours dysfunctional DNA-repair mechanisms. 5 The subsequent development of inhibitors to the stage at which they are now being evaluated as a viable therapeutic option in TNBC and BRCA-mutated solid cancers represents an exceptional example of a synthetic lethal therapeutic strategy that has been developed from bench to bedside in the past 10 years. Indeed, we are currently about to enter into an exciting era in which clinical trials of inhibitors in the adjuvant setting will be launched, despite the lack of clear evidence supporting the efficacy of these compounds in treating metastatic disease. Evaluating drugs as potential neoadjuvant and/or adjuvant therapies for localized disease without proven efficacy in the metastatic setting is exceptional. However, this course of action seems reasonable in the case of inhibitors, owing to the small and welldefined nature of the target population that might derive therapeutic benefit from such agents. Herein, we review the development of inhibitors to date, and discuss the considerations and challenges that might be encountered in the upcoming trials of this therapeutic approach. NATURE REVIEWS CLINICAL ONCOLOGY ADVANCE ONLINE PUBLICATION 1

Key points The poly(adp-ribose) polymerases (s) are a large family of multifunctional enzymes that have a role in the repair of single-strand breaks in DNA BRCA1 or BRCA2 mutation, resulting in a lack of homologous recombination, sensitizes cells to inhibition of activity, which in turn leads to chromosomal instability, cell-cycle arrest, and subsequent apoptosis The most-compelling evidence of the efficacy of inhibitors in the treatment of cancer comes from studies that involved patients with BRCA1 or BRCA2 mutations Approximately 70% of BRCA1-mutant and 20% of BRCA2-mutant breast tumours present as triple-negative breast cancer, a disease with poor prognosis; no targeted treatments have been approved specifically in this setting In the adjuvant setting, a positive therapeutic effect of inhibitors is anticipated in the well-defined population of patients with high-risk BRCA mutated primary breast cancer Normal cell DNA repaired inhibitors Double-strand breaks Repair by homologous recombination Single-strand break BRCA-mutated cell No homologous recombination No repair Cell survival Cell death DNA repaired Figure 1 The role of inhibitors in synthetic lethality. enzymes play a key part in the repair of DNA damage. In particular, 1 binds to single-strand breaks in DNA and recruits other enzymes to repair the DNA damage. Failure to repair single-strand breaks can result in double-strand breaks during DNA replication; thus, inhibition can induce further DNA damage. However, DNA damage, which is a frequent occurrence during each cell cycle, can also be repaired through homologous recombination mechanisms. The BRCA1 and BRCA2 genes encode key components of these homologous recombination repair pathways and, therefore, BRCA-mutant tumours are inherently deficient in DNA repair. This vulnerability forms the foundation for selective targeted synthetic lethal therapy with inhibitors in patients with BRCA-mutant breast cancer. The DNA damage that occurs after inhibition of activity cannot be adequately repaired in these cancers and eventually results in chromosomal instability, cell-cycle arrest, and subsequent apoptotic cell death. As DNA-repair processes remain intact in noncancerous cells, which generally retain at least one functional copy of both BRCA1 and BRCA2, inhibition is hypothesized to selectively kill cancer cells, sparing normal tissue. Abbreviation:, poly(adp-ribose) polymerase. Overview of experiences from the bench DNA is continually damaged by environmental exposures and endogenous activities (such as DNA replication errors), which cause diverse lesions, including SSBs and double-strand breaks (DSBs). 6 1, a widely and abundantly expressed member of the family, facilitates DNA repair by binding to SSBs and recruiting DNA repair proteins to the site of damage. 7,8 Nevertheless, genetic ablation of 1 expression in mouse models did not increase the development or early onset of tumours. 9 Studies revealed that, in the absence of 1, spontaneous SSBs collapse replication forks and result in DSBs, thus triggering DNA repair by homologous recombi nation pathways; 9 this compensatory repair mechanism probably explains the normal susceptibility of Parp1-deficient mice to cancer. 9 However, in BRCAdeficient cells, DSBs that occur at collapsed replication forks cannot be repaired owing to deficiencies in homologous recombination, leading to cell death. 2 This observation provided the rationale for the hypothesis that treatment with inhibitors would be especially effective in patients with cancer that developed on a background of germline inherited defects in homologous recombination, such as tumours arising in carriers of BRCA1 and BRCA2 mutations. Moreover, it was theorized that the effects of inhibition could potentially be highly specific to tumour tissue because, in BRCA +/ carri ers, only cancer cells are assumed to be BRCA / and, therefore, defective in homologous recombination; thus, the cancer cells might be susceptible to -inhibitor induced cell death, whereas normal cells would probably be resistant and spared. The use of an inhibitor of a DNA-repair enzyme to selectively kill tumour cells with deficient homologous recombination in the absence of an exogenous DNAdamaging agent represented a new concept in cancer therapy. This concept is an example of synthetic lethality, a phenomenon that arises when combined mutation or blockade of two or more genes leads to cell death, whereas a mutation (or blockade) of only one of these genes does not affect viability. 10 The synthetic lethal treatment approach is built around a mutation that does not induce cell death (for example, in BRCA genes), although the mutation might confer a phenotype (such as defective homologous recombination); however, the underlying mutation provides the opportunity to therapeutically target additional pathways ( proteins) to achieve lethality (Figure 1). As mentioned, BRCA1 or BRCA2 dysfunction has been demonstrated to profoundly sensitize cells to inhibition, ultimately resulting in apoptosis. 3 Subsequently, attempts were made to explore other therapeutic opportunities for inhibitors. BRCAmutated and basal-like TNBC share clinicopathological characteristics, including negative receptor status, diminished BRCA protein expression (or truncation), TP53 mutation, and sensitivity to DNA-damaging chemo therapies, such as platinum-based platins; 4,11,12 thus, the phenotype of TNBC was suggested to possess BRCAness and, therefore, patients with TNBC could conceivably derive therapeutic benefit from inhibition of enzymes. Through a series of experiments, basallike, as well as BRCA1-mutated, breast-cancer cells were shown to be defective in BER of oxidative DNA damage, and this defect conferred sensitivity to inhibition. 13 These data suggested that the therapeutic benefit of inhibitors might also extend to tumours deficient in BER. In addition, basal-like TNBC cells were observed to be selectively sensitive to platinum and gemcitabine compared with non-tnbc cells, and these agents exhibited synergy with inhibitors in basal-like, but not in 2 ADVANCE ONLINE PUBLICATION www.nature.com/nrclinonc

functions DNA repair Transcriptional regulation Chromatin modification Mitosis Cell death Functions in BER, DSB repair and NHEJ Component of the TLE1/Groucho corepressor complex involved in Wnt signalling implicated in transcriptional regulation of androgen receptor expression Maintenance of telomere length and chromosomal stability Involved in mitoticspindle formation Component of pathways mediating apoptosis Wnt C C T AC T G A AC GG GG TT GG G G T T G G G G G G A TG A CT T G CC C C A A C C C C A A C C C T C AT AG C G C GG T G G G TTGG G G T T C C C C A A Caspase-3 activation DVL GSK3β APC β-catenin Axin β-catenin cleavage Gene β-catenin transcription TLE1/Groucho Apoptosis Nucleus Figure 2 The different biological functions relevant to cancer. In addition to the classic activity of in BER, the family members have diverse functions in other biological processes, including transcriptional regulation, chromatin modification, mitosis (mitotic-spindle formation), and apoptosis, as well as intracellular trafficking, and energy metabolism (not shown). 1 and, to a lesser extent, 2 are important in maintaining telomere length and chromosomal stability. 1 also forms part of the Groucho/TLE1 co-repressor complex, and has been implicated as a transcriptional regulator of androgen receptor expression. Other s function in the repair of DSBs and in progression of mitosis (3), and some have potential roles in Wnt signalling and telomere maintenance (5 and 6). 1 is also a regulator of NHEJ, a mechanism of DSB repair. Abbreviations: APC, adenomatous polyposis coli protein; BER, base-excision repair; DSB, double-strand break; DVL, dishevelled homologue; GSK-3β, glycogen synthase kinase-3β; NHEJ, nonhomologous end joining;, poly(adp-ribose) polymerase; TLE1, transducin-like enhancer protein 1 (Groucho homologue). luminal, breast-cancer cell lines. 14 Indeed, some of these findings were incorporated into the design of clinical trials, providing rationale for investigations of iniparib (a candidate inhibitor) combined with gemcitabine and carboplatin, 15,16 although subsequent studies showed that iniparib does not inhibit s (as discussed in a following section). Other mechanisms to induce BRCAness have been suggested; for example, deficiency in the ataxia telangiectasia mutated (ATM) pathway, which is also involved in the repair of DSBs in DNA through homologous recombination, typically occurs in approximately 25% of chronic lymphocytic leukaemias. 17 Furthermore, cells deficient in individual components of the ATM pathway have been shown to be sensitive to inhibitors. 18 With more relevance to breast cancer, the tumour suppressor gene PTEN, which harbours mutations in 1 4% of breast cancers depending on the subtype, 19 was demonstrated to be important for maintenance of genome stability; PTEN deficiency was associated with defective homologous recombination in human tumour cells, which sensitized the cells to inhibitors, both in vitro and in vivo. 20 These data suggest that the clinical assessment of inhibitors might be extended to patients with PTEN mutant tumours. Although most inhibitors are known to inhibit 1 and 2, the family comprises around 15 other less-explored proteins (10 putative) that might all be influenced by such agents. Moreover, apart from the classic function of s in BER, the family members also have diverse biological functions, including transcriptional regulation, telomere cohesion, mitotic-spindle formation, intracellular trafficking, and energy metabolism (Figure 2). 21 23 1 and, to a lesser extent, 2 are primarily involved in BER; however, these enzymes are also important in maintaining telomere length and chromosomal stability. 24 In addition, as part of the Groucho/transducin-like enhancer protein 1 co-repressor complex, 1 has been implicated in transcriptional regulation of androgen receptor expression. 25 3 has a role in the repair of DSBs and in progression of mitosis, 26,27 whereas 5 and 6 are suggested to have roles in Wnt signalling and telomere maintenance. 28 Whether interference with these processes contributes to synthetic lethality after inhibition remains unclear. Nevertheless, evidence suggests that 1 is also a regulator of another mechanism of DSB repair: nonhomologous end joining (NHEJ). 29 As 1-mediated NHEJ might occur more selectively in cells deficient in homologous recombination, in which NATURE REVIEWS CLINICAL ONCOLOGY ADVANCE ONLINE PUBLICATION 3

repair mechanisms probably default to secondary pathways, inhibitors could ultimately prove especially useful in this setting. In 2012, a study demonstrated that inhibitors can act as poisons by trapping the 1 and 2 enzymes at sites of DNA damage, generating complexes that are more cytotoxic than the unrepaired SSBs that result from inactivation. 30 The study also found that certain inhibitors were more potent in trapping enzymes than others, and identified pathways other than homologous recombination that were essential for repairing the DNA complexes (such as post-replication repair, Fanconi anaemia, polymerase β, and FEN1 pathways). These findings, as well as those described previously, might provide the rationale for using inhibitors in cancers with deficiencies in DNA-repair proteins other than BRCA. BRCA-mutation, TNBC and prognosis Approximately 5% of breast cancers are associated with a germline mutation in the BRCA1 and/or BRCA2 genes. Owing to the small size of the BRCA-mutated patient subpopulation in breast cancer, comparisons of disease outcome between this group and the overall breast cancer patient population have only been performed in a number of small studies. 31,32 Nonetheless, the available data suggest that, once baseline prognostic factors and treatment are taken into account, patients with BRCA mutations have a similar outcome to their counterparts with sporadic breast cancers. No treatments have been approved specifically for patients with BRCA-mutated breast cancer and, therefore, these patients are treated according to hormone receptor and HER2 status. Approximately 70% of BRCA1-mutated and 20% of BRCA2-mutated breast cancers presents as triple-negative disease. 33 TNBC is associated with a poor prognosis, and the therapeutic options available for this disease are limited; thus, further development of effective targeted treatment represents an important unmet need. Patients with TNBC frequently develop early recurrence during the first few years after diagnosis, particularly early visceral metastases. 4 A general impression of the prognosis of patients with TNBC undergoing adjuvant therapy can be drawn from the BEATRICE study, 34 a randomized phase III trial that evaluated the impact of adjuvant bevacizumab therapy on invasive-diseasefree survival (IDFS). In the BEATRICE trial, 34 the 3 year IDFS for both the chemotherapy and the combined chemo therapy plus bevacizumab cohorts was approximately 83%. Of note, however, more than half of patients in the trial had tumours that were node negative, and one-third had tumours <2 cm. 34 Approximately half of all patients with TNBC receive neoadjuvant chemotherapy. The most important prognostic information to be obtained after neoadjuvant chemotherapy is whether patients have had a pathological complete response (pcr), commonly defined as the absence of invasive cancer in the resected breast specimen and sampled ipsilateral lymph nodes. Patients with TNBC in whom a pcr is achieved have a prognosis similar to that of patients with other breast cancer subtypes, with less than 10% developing recurrence at distant sites within 5 years. 35 However, patients with TNBC who do not experience a pcr have a considerably poorer prognosis, 35 and a meta-analysis of neoadjuvant chemotherapy trials reported a pcr rate of only 31% in patients with TNBC. 36 This meta-analysis provided additional evidence that patients in whom a pcr was not achieved were at higher risk of earlier recurrence compared with patients who experience a pcr (HR 6.02, 95% CI 3.92 9.25). 36 These data highlight the need to develop further treatment options in order to improve the prognosis of patients in whom a pcr is not observed after neoadjuvant chemotherapy as well as patients with a high-risk of recurrence in the adjuvant setting. Clinical experience of inhibitors Since the publication of the two landmark articles 2,3 that described the sensitivity of BRCA-mutated cells to inhibition, 127 US National Cancer Institute (NCI)- registered clinical trials have been launched to evaluate a range of compounds with -inhibitor activity. At the time of writing (January 2014), these trials were in different phases of clinical development, with more than 30 having been published or otherwise reported (in abstract form, for example). We provide a general summary of the major observations and the developmental status of each compound evaluated to date in Table 1. The focus of many of the trials of inhibitors is the patient population with BRCA1-mutated and BRCA2-mutated breast cancer, or other cancers with BRCAness features, such as TNBC and ovarian serous papillary carcinoma. However, in an effort to increase the population of patients with cancer who might benefit from inhibitors, several pharmaceutical companies are attempting to integrate these agents with c hemotherapy regimens. In tumours in non-brca-mutation carriers Several inhibitors have been evaluated in combination with cytotoxic drugs in patients without germline BRCA mutations. For example, in a phase I study in patients with relapsed or refractory multiple myeloma, 37 the combination of the inhibitor veliparib and the proteasome inhibitor bortezomib seemed to be well toler ated, with evidence of considerable antitumour activity.in another phase I study, 38 performed to determine the maximal tolerated dose (MTD) of a regimen comprising the topoisomerase inhibitor topotecan plus the inhibitor olaparib, haematological toxicity became limiting at a subtherapeutic dose (Table 2). Thus, this combination was not explored further. Olaparib has also been combined with the alkylating agent dacarbazine for the treatment of patients with mela noma; although this combination was well tolerated, no therapeutic responses were observed and this t reatment approach has not been further pursued. 39 Outside the context of TNBC or BRCA-mutation carriers, noteworthy experiences with inhibitors also 4 ADVANCE ONLINE PUBLICATION www.nature.com/nrclinonc

Table 1 The range of inhibitors in clinical use and development inhibitor Olaparib (AZD2281) Veliparib (ABT 888) Rucaparib (AG 014,699; CO 338) BMN 673 CEP 9722 Niraparib (MK4827) Company Target Route of administration Findings of key trials, if available, and/or phase of development AstraZeneca 1/2/3 Oral Series of phase II trials demonstrated efficacy in BRCA-mutation carriers 52,54 Currently being evaluated in the adjuvant setting in patients with TNBC 70 Abbott Laboratories 1/2 Oral Acceptable safety profile and promising antitumour activity, especially in BRCA-deficient patients with ovarian cancer, was observed in a series of phase I trials 71 Phase II studies are ongoing 72 Clovis Oncology 1/2 Oral or intravenous BioMarin Pharmaceutical Teva Pharmaceutical Industries Currently being evaluated in a phase II/III study in BRCA-mutation carriers with mbc or advanced-stage ovarian cancer 73,74 1/2 Oral BMN 673 has shown impressive antitumour activity in patients with BRCA mutations 48 Currently, phase II III studies in patients with germline BRCA mutations are ongoing 75,76 1/2 Oral Clear evidence of inhibition has been demonstrated in preclinical studies, 77 and early studies in patients 42 Merck 1/2 Oral Phase I II studies have revealed antitumour activity, especially in patients with germline BRCA mutations 49 At present, phase II III studies in such patients are ongoing 78,79 Abbreviations: mbc, metastatic breast cancer;, poly(adp-ribose) polymerase; TNBC, triple-negative breast cancer. come from a phase II trial of iniparib (BSI 201) in combination with gemcitabine and carboplatin in patients with platinum-sensitive recurrent ovarian cancer. 40 The preclinical studies 13,14 we have described and a pivotal trial in metastatic TNBC 16 provided the rationale for the design of this trial. The combination of gemcitabine and carboplatin with iniparib was associated with an overall response rate (ORR) of 70.6% in the first 17 patients treated with this regimen, and the safety profiles were consistent with those observed in previous clinical studies of iniparib combined with gemcitabine and carboplatin. 16,40 The study authors concluded that iniparib plus gemcitabine and carboplatin demonstrated activity in patients with platinum-sensitive recurrent ovarian cancer, with an increased ORR compared with the rate observed in previous studies of gemcitabine plus carboplatin (70.6% versus 47.2%). 40 However, iniparib is no longer considered a classic inhibitor, and phase III trials of this combination therapy were d isappointing; 15 therefore, this regimen has been abandoned. Maintenance treatment represents an interesting alternate use of inhibitors that has been evaluated in a randomized, double-blind, placebo-controlled phase II study (Table 2). 41 In particular, maintenance treatment with olaparib was investigated in patients with platinum-sensitive, relapsed, serous ovarian cancer who had responded to a platinum-based chemo therapy regimen. 41 Progression-free survival (PFS) was significantly prolonged by olaparib treatment compared with placebo (median 8.4 months versus 4.8 months P <0.001), although interim analysis showed no overall survival benefit, possibly due to confounding related to treatment of a proportion of the control cohort with inhibitors after disease progression; in the overall patient population, no olaparib-treated patients but 16 of the 129 (12.4%) placebo-treated patients went on to receive a inhibitor in other clinica l studies after specific requests for unblinding after disease progression. 41 In addition, CEP 9722, another oral 1/2 inhibitor (Table 1), has been evaluated as both a single-agent therapy and in combination with temozolomide in a phase I dose-escalation study 42 that evaluated its safety, pharmacokinetics, and pharmacodynamics in patients with advanced-stage solid tumours. In combination with temozolomide, CEP 9722 was well tolerated and signals indicative of inhibition were demonstrated. 42 Overall, however, a therapeutic benefit for the strategy of combining chemotherapy with inhibitors in the general cancer population has not been clearly proven in the studies performed to date (Table 2). In breast cancer irrespective of BRCA status As one of the most-common tumours manifest in carriers of germline BRCA mutations is breast cancer, and because of the BRCAness of basal-like TNBC, the use of TNBC irrespective of BRCA status as a setting for clinical trials to evaluate the different inhibitors in development is tempting. Although no longer considered a inhibitor, iniparib was the first drug of this class to be assessed in a phase III clinical trial. 15 The promising results of a phase II trial in patients with TNBC, which was initially presented in 2009 and subsequently published in 2011, 16 elicited a great deal of excitement over this agent. In this trial, a total of 123 patients were randomly assigned to receive gemcitabine and carbo platin with or without iniparib on days 1, 4, 8, and 11 of every 21-day treatment cycle. 16 The addition of iniparib to gemcitabine and carboplatin improved the rate of clinical benefit that is, the percentage of patients with a complete response, partial response or stable disease for 6 months from 34% to 56% (P = 0.01) and the ORR rate from 32% to 52% (P = 0.02). 16 In addition, compared with the gemcitabine and carboplatin cohort, prolongation of the median PFS duration from 3.6 months to 5.9 months, and of the median overall survival duration from 7.7 months to 12.3 months was also observed with the addition of iniparib to the treatment NATURE REVIEWS CLINICAL ONCOLOGY ADVANCE ONLINE PUBLICATION 5

Table 2 Clinical trials of inhibitors in solid tumours (published or reported) inhibitor evaluated and NCT identifier Olaparib NCT00516438 NCT00516802 NCT00572364 NCT00710268 NCT01081951 Regimens and doses Setting Important efficacy and/or safety data Topotecan: 0.5 or 1 mg/m 2 daily for 3 days Olaparib: 50, 100, or 200 mg BID for six cycles Dacarbazine: 600 800 mg/m 2 on day 1 of every 21-day cycle Olaparib: 20 200 mg on days 1 7 of every 21-day cycle Olaparib: dose escalation from 100 mg to 400 mg BID Bevacizumab: 10 mg/kg every 2 weeks Olaparib: 100, 200, or 400 mg BID Paclitaxel: 175 mg/m 2 from day 1 for six cycles Carboplatin: AUC4* from day 1 for six cycles Olaparib: 200 mg BID on days 1 10 of every 21-day cycle for six cycles, then 400 mg BID as maintenance therapy NCT01063517 Paclitaxel: 80 mg/m 2 intravenously on days 1, 8, 15 of every 28-day cycle Olaparib: 100 mg BID (tablet form) Phase I study to determine the safety and tolerability of olaparib combined with topotecan for the treatment of patients with advanced-stage solid tumours (n = 19) 38 Phase I study of the safety and tolerability of olaparib and dacarbazine in patients with advanced-stage solid tumours (n = 40) 39 Phase I dose-finding and pharmacokinetic study of olaparib in Japanese patients with advanced-stage solid tumours (n = 12) 80 Phase I study that assessed the safety and tolerability of olaparib in combination with bevacizumab in patients with advanced stage solid tumours (n = 12) 81 A randomized, open-label phase II study of olaparib plus P/C followed by olaparib maintenance treatment in patients with platinum-sensitive recurrent serous ovarian cancer (n = 162) 82 Randomized, double-blind phase II study of olaparib plus paclitaxel in patients with recurrent or metastatic gastric cancer (n = 124) 83 NCT00753545 Olaparib: 400 mg BID Randomized, double-blind, placebocontrolled phase II study of olaparib as a maintenance treatment in relapsed patients with platinum-sensitive serous ovarian cancer (n = 265) 41 NCT00678132 Cisplatin: 60 mg/m 2 on day 3 Gemcitabine: 500 mg/m 2 on days 3 and 10 Olaparib: 100 mg BID on days 1 4 Veliparib Phase I study of olaparib with cisplatin and gemcitabine in adults with solid tumours (n = 23) 84 NCT00387608 Veliparib: single oral dose of 10, 25, or 50 mg Phase 0 evaluation of veliparib in patients with advanced-stage malignancies (n = 13) 85 NCT00553189 NCT01063816 NCT01051596 NCT01085422 Topotecan: varying schedules Veliparib: 10 mg BID Carboplatin: AUC4 on day 1 of every 21-day cycle Gemzar: 800 mg/m 2 on day 8 of every 21-day cycle Veliparib: 250 mg BID (MTD) Temozolomide: 150 mg/m 2 orally daily Veliparib: 40 mg orally BID Temozolomide: on days 1 5 of a 28-day cycle Veliparib: 40 mg BID between days 1 7 NCT01205828 Temozolomide: 150 mg/m 2 on days 1 5 in a 28 day cycle Veliparib: 40 mg daily on days 1 7 NCT01445522 Cyclophosphamide: 50 mg daily on a 21-day cycle Veliparib: 20 60 mg daily for 7, 14, or 21 days Phase I study of veliparib in combination with topotecan in adults with refractory solid tumours and lymphomas (n = 24) 86 Phase I study of veliparib in combination with G/C in patients with advanced-stage solid tumours (n = 59) 87 Study to evaluate response in patients with metastatic, unresectable colorectal cancer that had progressed on all standard therapies (n = 47) 88 Study of veliparib with temozolomide in patients with metastatic castration-resistant prostate cancer (n = 26) 89 Evaluation of veliparib with temozolomide in patients with liver cancer refractory to sorafenib (n = 16) 90 Study to establish the safety and MTD of veliparib combined with cyclophosphamide and to characterize the regimen (n = 35) 91 Safety: dose-limiting haematological adverse events at subtherapeutic MTD Safety: study defined the tolerable dose of olaparib in combination with dacarbazine Efficacy: no advantage over single-agent dacarbazine Safety: olaparib was well tolerated up to the 400 mg BID dose Efficacy: PR in one patient and SD in four patients Safety: the combination was generally well tolerated with no DLTs Efficacy: olaparib plus P/C followed by maintenance olaparib showed a significant improvement in PFS versus P/C alone (median 12.2 months versus 9.6 months; HR 0.51; P = 0.0012) Safety: the combination was well tolerated Efficacy: improved OS, but not PFS, in the olaparib groups versus paclitaxel alone Safety: well tolerated Efficacy: improvement in PFS (8.4 months versus 4.8 months), but not OS, observed Safety: myelosuppression was observed at relatively low doses Efficacy: PR in 2/21 patients Safety: veliparib was well tolerated Efficacy: marked decrease in levels was observed in tumour biopsy tissues and peripheral blood mononuclear cells Safety: MTD determined Efficacy: inhibition shown to modulate the capacity to repair topoisomerase 1-mediated DNA damage Safety: the combination was acceptably tolerated with a safety profile similar to that observed for G/C alone Safety: combination was very well tolerated Efficacy: disease control rate of 23% Safety: combination was well tolerated Efficacy: 4/25 patients had stable disease for a minimum of 4 months Safety: the combination was fairly well tolerated Efficacy: minimal therapeutic activity observed Safety: treatment was well tolerated Efficacy: PR in 7/35 patients 6 ADVANCE ONLINE PUBLICATION www.nature.com/nrclinonc

Table 2 (cont.) Clinical trials of inhibitors in solid tumours (published or reported) inhibitor evaluated and NCT identifier CEP 9722 Regimens and doses Setting Important efficacy and/or safety data NCT00920595 Temozolomide: 150 mg/m 2 on days 1 5 CEP 9722: 150 1,000 mg on days 1 5 of a 14-day cycle Rucaparib NCT01526928 Carboplatin: AUC3 5 on day 1 Rucaparib: starting dose of 80 mg orally on days 1 14 of every 21-day cycle Phase I dose-escalation study to evaluate the safety, pharmacokinetics, and pharmacodynamics of CEP 9722 in patients with advanced-stage solid tumours (n = 26) 42 Phase I study to assess the safety and tolerability of escalating oral doses of rucaparib in combination with carboplatin in patients with advanced-stage solid tumours (n = 23) 92 NCT01482715 Rucaparib: 40 480 mg BID in a 21-day cycle Study evaluating the safety profile of escalating doses of continuous daily oral rucaparib (n = 29) 93 NA Temozolomide: 200 mg/m 2 on days 1 5 of every 28-day cycle Rucaparib: intravenously at 12 mg/m 2 every 28 days Study investigating safety and efficacy in patients with advanced-stage metastatic melanoma (n = 46) 94 Safety: the combination was well tolerated Efficacy: inhibition was demonstrated by pharmacodynamic analysis Safety: combination was safe Efficacy: PR and durable stable disease were observed in 4/7 heavily pretreated patients with ovarian cancer Safety: well tolerated up to 360 mg BID Efficacy: PR observed in some patients Safety: Myelosuppression occurred in 54% Efficacy: RR = 17%; median PFS of 3.5 months *In the P/C only arm, in the chemotherapy alone arm the carboplatin dose was AUC6. In an interim analysis at 38% maturity. Abbreviations: AUC, area under the concentration-versus-time curve; BID, twice daily; DLTs, dose-limiting toxicities; G/C, gemcitabine plus carboplatin; MTD, maximum tolerated dose; NA, not applicable; NCT, National Cancer Institute National Clinical Trials database; OS, overall survival;, poly(adp-ribose) polymerase; P/C, paclitaxel plus carboplatin; PFS, progression-free survival; PR, partial response; RR, response rate; SD, stable disease. regimen. 16 The phase III trial using the same design and the co-primary end points of overall survival and PFS was initiated to confirm these result, and included 519 patients; 15 disappointingly, the results of this study did not meet the criteria for significance of the co-primary end points. 15 The failure of iniparib stimulated a back-to-the-bench phase of research that led to some intriguing observations. Patel et al. 43 compared the actions of three inhibitors, iniparib, olaparib, and veliparib. Consistent with earlier reports, these researchers demonstrated that olaparib and veliparib selectively inhibited colony formation and induced apoptosis in cells lacking expression of functional BRCA2 or ATM. 43 Furthermore, similar to findings with earlier-generation inhibitors, olaparib and veliparib sensitized cells to the topoisomerase I inhibitors camptothecin and topotecan. 43 By contrast, iniparib exhibited limited or no capacity to selectively induce apoptosis in cells deficient in homologous recombination. 43 In other experiments, iniparib also failed to sensitize cells to cisplatin, gemcitabine, or paclitaxel. 43 Thus, Patel et al. 43 concluded that, although high concentrations (>40 μmol/l) of iniparib can induce the death of normal and neo plastic cells, the effects of this agent probably do not reflect inhibition and should not be used to guide decisions regarding other agents that selectively inhibit s. Opinions of iniparib therapy evolved from careful optimism to clear pessimism; nevertheless, additional studies were performed to evaluate i nhibitors p rimarily olaparib in patients with metastatic TNBC without BRCA1 or BRCA2 mutations (Table 3). In a phase II, multi centre, open-label, nonrandomized study, 44 women with advanced-stage ovarian carcinoma or TNBC received 400 mg of olaparib twice daily. This study enrolled 91 patients (26 with breast cancer), who were stratified according to whether or not they had BRCA1 or BRCA2 mutations, and the primary end point was ORR. 44 No confirmed objective responses were reported in the groups of patients with breast cancer, and the most-frequent adverse event was fatigue (reported by 13 [50%] of the patients with breast cancer). Currently, therefore, clear evidence of efficacy from trials using inhibitors in TNBC outside the context of BRCA mutation, as in solid cancers, is lacking. In patients with BRCA mutations Considering preclinical observations regarding the actions of inhibitors, patients with BRCA mutations are the most investigated population in clinical trials of different inhibitors. As these patients have inherent sensitivity to inhibitors in accordance with the concept of synthetic lethality, most trials focusing on this population used single-agent inhibitor therapy without assessing the potential synergism with cytotoxic agents. Different inhibitors including olaparib, niraparib, BMN 673, and rucaparib (Table 1) have shown promising evidence of therapeutic activity as monotherapies in patients with cancer who have BRCA1 or BRCA2 mutations (Table 4). 45 48 The results of a first-in-human trial of the novel oral inhibitor BMN 673 were presented at the 2013 ASCO Annual Meeting; 48 the pharmacokinetics, pharmacodynamics, safety and antitumour activity of this agent were evaluated in a two-stage dose-escalation study in patients with solid tumours, with 3 6 patients treated at each dose level. 48 In the second stage of this trial (expansion of the group treated at the MTD), only patients with tumours defective in DNA repair were recruited. 48 Among patients with tumours with deleterious BRCA1 NATURE REVIEWS CLINICAL ONCOLOGY ADVANCE ONLINE PUBLICATION 7

Table 3 Published or reported clinical trials of inhibitors in breast cancer independent of BRCA status inhibitor used and NCT identifier Olaparib Regimens and doses Setting Important efficacy and/or safety data NCT00679783 Olaparib: 400 mg twice a day Phase II study of olaparib in patients with TNBC (n = 26 with breast cancer; 65 patients with ovarian cancer were also included) 44 NCT01116648 Cediranib: 20 30 mg daily (or to DLT) Olaparib: 100 400 mg BID NCT00707707 Paclitaxel: weekly at 90 mg/m 2 for 3 weeks per 4 week cycle Olaparib: 200 mg BID continuously NA Olaparib: 10, 30, 100, 200 or 400 mg BID (capsule formulation) for the 4 5 days preceding breast-cancer surgery Veliparib NCT01009788 Temozolomide: 150 mg/m 2 of days 1 5 Veliparib: 40 mg BID on days 1 7 Study of olaparib in combination with the antiangiogenic agent cediranib in recurrent TNBC (n = 8; 20 patients with ovarian cancer were also included) 95 Study of olaparib, in combination with paclitaxel in patients with metastatic TNBC (n = 19) 96 Study to identify an effective biological dose of olaparib for future trials (n = 60) 97 Study to evaluate whether combining veliparib with temozolomide would be active in metastatic breast cancer (n = 41) 98 Efficacy: no confirmed objective responses were reported Safety: the combination had haematological DLTs Safety: diarrhoea, nausea and neutropenia observed in 58 63% of patients Efficacy: 37% (7 patients) had a confirmed PR Safety: tolerability was consistent with previous studies Efficacy: unexpectedly low olaparib exposure (poor pharmacokinetics) Safety: common grade 3 4 toxicities included thrombocytopenia and neutropenia Efficacy: best responses for the 24 evaluable patients included one CR, two PRs, seven instances of SD Abbreviations: BID, twice daily; CR, complete response; DLT, dose-limiting toxicity; NA, not applicable; NCT, National Cancer Institute National Clinical Trials database;, poly(adp-ribose) polymerase; PFS, progression-free survival; PR, partial response; SD, stable disease; TNBC, triple-negative breast cancer. or BRCA2 mutations, objective responses occurred in almost 65% of the patients with BRCA-mutant ovarian and/or peritoneal cancer and >30% of the patients with BRCA-mutant breast cancer. 48 Moreover, in a phase I trial published in 2013 (Table 4), 49 niraparib was well tolerated among the 100 patients treated and showed promising antitumour activity, especially in BRCAmutation carriers (40% and 50% ORR in ovarian and breast cancer, respectively). Olaparib is the most-investigated inhibitor in BRCA-mutation carriers (Table 4). Fong et al. 46 conducted a phase I trial of olaparib that enrolled 60 patients, of whom 22 were carriers of a germline BRCA1 or BRCA2 mutation. Reversible dose-limiting toxicity was seen in one of eight patients who received 400 mg of olaparib twice daily (grade 3 mood alteration and fatigue), and two of five patients treated with a 600 mg dose of the drug twice daily (grade 4 thrombocyto penia and grade 3 somnolence). 46 This manageable safety profile prompted the investigators to enrol another cohort of patients comprising only carriers of BRCA1 or BRCA2 mutations, who were treated with olaparib at a dose of 200 mg twice daily. 46 No obvious increase in toxicity was seen in this additional population, although antitumour activity was reported only in mutation carriers with ovarian, breast or prostate cancers; 46 however, the ORR was 47%, which is impressive considering that the population studied comprised heavily p retreated patients. 46 Two phase II proof-of-concept trials have also been performed to evaluate single-agent olaparib therapy in women with confirmed BRCA1 or BRCA2 mutations and advanced-stage breast 50 or ovarian cancer. 51 In the breast cancer study, 50 two cohorts each comprising 27 patients were treated continuously with olaparib at doses of either 100 mg or 400 mg twice daily; the ORRs were 41% and 22% in the high-dose and low-dose schedules, respectively, and nausea and fatigue mainly at grade 1 2 were the most-common adverse events. 50 Similar results were obtained in the study involving patients with ovarian cancer; the ORRs were 33% (11 of 33 patients) in the cohort assigned to 400 mg olaparib twice daily and 13% (3 of 24 patients) in the cohort assigned to the 100 mg twice daily dosage regimen. 51 The findings of these two phase II studies provided positive proof-of-concept for the efficacy and tolerability of genetically targeted treatment with olaparib in patients with BRCA-mutated advanced breast and ovarian cancers, and d emonstrated a dose response relationship. Since the publication of these two trials, 50,51 olaparib has been investigated in several more clinical trials with different designs and varying patient populations (Table 4). In an open-label phase II, multicentre, noncomparative study presented at the 2013 ASCO Annual Meeting, 52 tumours in 298 BRCA1/2-mutation carriers were evaluated for responsiveness to olaparib, regardless of tumour type. In this study, heavily pretreated patients with advanced-stage cancer refractory to standard therapy nearly all of the patients with breast cancer included had received more than three different prior chemotherapies for metastatic disease received olaparib 400 mg twice daily until disease progression (median duration of treatment was 5.5 months; range <1 month to 28.5 months). 52 The most commonly reported adverse events were fatigue (59%), nausea (59%), and vomiting (37%). 52 A clinical benefit of olaparib therapy was demonstrated for patients with prostate and pancreatic cancers, and antitumour activity in patients with ovarian and breast cancer was also confirmed. 52 8 ADVANCE ONLINE PUBLICATION www.nature.com/nrclinonc

Table 4 Published or reported clinical trials of cancer therapy with inhibitors in BRCA-mutation carriers inhibitor used and NCT identifier Olaparib NCT00516373 NCT00494234 NCT00494442 NCT00628251 Regimens and doses Setting Important efficacy and/or safety data Olaparib: varied from 10 mg daily for 2 weeks in every 3-week cycle to 600 mg BID continuously Olaparib: 100 mg or 400 mg BID Olaparib: 100 mg or 400 mg BID Intravenous liposomal doxorubicin: 50 mg/m 2 on a 28 day cycle Olaparib: 200 mg or 400 mg BID Phase I study investigating inhibition of in tumours in carriers of BRCA mutations (22/60 patients enrolled) 46 Study of olaparib in patients with BRCA1/2 mutations and advanced breast cancer (n = 54) 50 Evaluation of olaparib in patients with BRCA1/2 mutations and recurrent ovarian cancer (n = 57) 51 Phase II, open-label, randomized, multicentre study comparing the efficacy and safety of olaparib and pegylated liposomal doxorubicin in patients with BRCA1/2 mutations and recurrent ovarian cancer (n = 97) 54 NCT01078662 Olaparib: 400 mg BID Open-label phase II study of olaparib monotherapy in patients with advanced cancer and a germline BRCA1/2 mutation (n = 298) 52 NCT00753545 Olaparib: 400 mg BID Preplanned subgroup analysis of BRCA status in a study of olaparib maintenance therapy in platinum-sensitive relapsed ovarian cancer (BRCA status was known for n = 254) 53 NCT01445418 BMN 673 NCT01286987 Niraparib Carboplatin: AUC3 5 on a 21-day cycle Olaparib: 100 400 mg every 12 h on days 1 7 of a 21-day cycle BMN 673: 25 1100 μg per day Evaluation of whether olaparib can be administered safely when combined with carboplatin in patients with breast (n = 8) and ovarian cancer (n = 37) 99 First-in-human trial of novel oral inhibitor BMN 673 in patients with solid tumours (17/39 patients enrolled had BRCA mutations) 48 NCT00749502 Niraparib: 30 400 mg Phase I study to determine drug toxicity and tolerability (24/100 patients enrolled had BRCA mutations) 49 Rucaparib NCT01482715 Rucaparib: 18 mg/m 2 on days 1 5 of a 21 day cycle Efficacy study in confirmed BRCA-mutation carriers with advanced ovarian and/or advanced breast cancer (n = 41) 47 Safety: olaparib treatment was associated with a limited number of the adverse effects of conventional chemotherapy Efficacy: radiological or tumour-marker responses or meaningful disease stabilization (stable disease for a period of 4 months or more) in cancer associated with BRCA1/2 mutations Safety: toxicity in women with BRCA1/2 mutations was similar to that seen in those without BRCA mutations Efficacy: ORR was 41% among 27 patients in the 400 mg BID cohort Safety: toxicity was tolerable (mild in severity) Efficacy: ORR was 33% among 33 patients in the 400 mg BID cohort and 13% at 100 mg BID doses Efficacy: no statistically significant difference in PFS between the treatment groups Safety: well tolerated Efficacy: RR of 31% and 13% in ovarian and breast cancer, respectively Safety: toxicity was mild in severity Efficacy: patients with germline mutations in BRCA had the greatest PFS benefit with olaparib maintenance therapy versus placebo (11.2 months versus 4.3 months) Safety: well tolerated Efficacy: activity in carriers of germline BRCA mutations; 50% PR rate and 1 patient had a complete response Safety: well tolerated Efficacy: RECIST and/or CA 125 responses occurred in 11/17 BRCA-mutation carriers Safety: well tolerated Efficacy: 40% ORR (8/20 patients) in BRCA1/2-mutation carriers with ovarian cancer and 50% ORR (2/4 patients) in BRCA1/2-mutation carriers with breast cancer Safety: no DLT occurred at any dose Efficacy: ORR was 5%; however, 26% of patients achieved SD for 4 months Abbreviations: BID, twice daily; CA-125, cancer antigen 125; DLT, dose-limiting toxicity; NCT, National Cancer Institute National Clinical Trials database; ORR, overall response rate;, poly(adp-ribose) polymerase; PFS, progression-free survival; PR, partial response; RECIST, Response Evaluation Criteria In Solid Tumours; SD, stable disease. As previously described, maintenance treatment with olaparib in platinum-sensitive relapsed ovarian cancer was associated with a significant improvement in PFS. 41 In a preplanned subgroup analysis of this randomized trial data (Table 4), 53 patients with a germline BRCA mutation had the greatest PFS benefit with olaparib maintenance therapy compared with placebo (11.2 months versus 4.3 months). Importantly, 19 patients had received olaparib for >3 years, 53 demon strating a unique long-lasting response in a subset of patients. Overall, median overall survival for patients with a germline BRCA mutation was 3 months longer than patients without germline BRCA mutations among the study population that received olaparib, although this difference was not statistically significant; 23% of patients who initially received placebo subsequently received a inhibitor, which represents a potential confounding factor. 53 Together, the results of these trials reveal that olaparib therapy is likely to be of benefit to patients with deleterious BRCA mutations. However, not all trials of olaparib in patients with germline BRCA mutations have reported positive results. A phase II, open-label, randomized, multicentre study that compared olaparib to pegylated liposomal doxorubicin (PLD) in 97 patients with ovarian cancer and BRCA1 or BRCA2 mutations who experienced tumour recurrence within 12 months of platinum-based therapy NATURE REVIEWS CLINICAL ONCOLOGY ADVANCE ONLINE PUBLICATION 9