2 Inhibitors As a Potential Therapy for ERCC1-Deficient Non-Small Cell

Home , RAD51

ORIGINAL ARTICLE A high-throughput screen identiﬁes PARP1/2 inhibitors as a potential therapy for ERCC1-deﬁcient non-small cell lung cancer

S Postel-Vinay1,2, I Bajrami1, L Friboulet2, R Elliott1, Y Fontebasso1, N Dorvault2, KA Olaussen2,3, F Andre´ 2,3, J-C Soria2,3, CJ Lord1 and A Ashworth1

Excision repair cross-complementation group 1 (ERCC1) is a DNA repair enzyme that is frequently defective in non-small cell lung cancer (NSCLC). Although low ERCC1 expression correlates with platinum sensitivity, the clinical effectiveness of platinum therapy is limited, highlighting the need for alternative treatment strategies. To discover new mechanism-based therapeutic strategies for ERCC1-defective tumours, we performed high-throughput drug screens in an isogenic NSCLC model of ERCC1 deficiency and dissected the mechanism underlying ERCC1-selective effects by studying molecular biomarkers of tumour cell response. The high-throughput screens identified multiple clinical poly (ADP-ribose) polymerase 1 and 2 (PARP1/2) inhibitors, such as olaparib (AZD-2281), niraparib (MK-4827) and BMN 673, as being selective for ERCC1 deficiency. We observed that ERCC1-deficient cells displayed a significant delay in double-strand break repair associated with a profound and prolonged G2/M arrest following PARP1/2 inhibitor treatment. Importantly, we found that ERCC1 isoform 202, which has recently been shown to mediate platinum sensitivity, also modulated PARP1/2 sensitivity. A PARP1/2 inhibitor-synthetic lethal siRNA screen revealed that ERCC1 deficiency was epistatic with homologous recombination deficiency. However, ERCC1-deficient cells did not display a defect in RAD51 foci formation, suggesting that ERCC1 might be required to process PARP1/2 inhibitor-induced DNA lesions before DNA strand invasion. PARP1 silencing restored PARP1/2 inhibitor resistance in ERCC1-deficient cells but had no effect in ERCC1-proficient cells, supporting the hypothesis that PARP1 might be required for the ERCC1 selectivity of PARP1/2 inhibitors. This study suggests that PARP1/2 inhibitors as a monotherapy could represent a novel therapeutic strategy for NSCLC patients with ERCC1-deficient tumours.

Oncogene (2013) 32, 5377–5387; doi:10.1038/onc.2013.311; published online 12 August 2013 Keywords: non-small cell lung cancer; ERCC1; PARP1; synthetic lethality; DNA repair

INTRODUCTION ERCC1 acts as a structure-specific DNA endonuclease, cutting DNA Non-small cell lung cancer (NSCLC) is the leading cause of cancer- on the 50 site of DNA lesions, a function that is thought to be rate- 10–13 related death worldwide, with 50% of patients presenting with limiting in the processing of platinum-induced DNA crosslinks. advanced or metastatic disease at diagnosis.1 Currently, less than Platinum salts are the cornerstone of NSCLC treatment, but their 15% of patients survive 5 years beyond diagnosis. The median administration is unfortunately restricted by cumulative haemato- overall survival in the metastatic setting is only 10–12 months, and neuro-toxicities, and has to be halted after 4–6 cycles even if despite aggressive treatments. Therefore, new therapeutic some degree of anti-tumour response is still observed. Moreover, a approaches for this disease are urgently needed. significant proportion of patients, who would otherwise benefit The recent advances in genome analysis have identified a from platinum-based therapy, are not eligible for such treatment, number of genetic alterations in NSCLC that could be therapeutically owing to organ dysfunction, co-morbidities or poor performance exploited as predictive biomarkers for guiding treatment decisions status. Approximately 50% of NSCLC patients display low levels of and customising therapy, eventually improving patient outcome.2,3 ERCC1 and could benefit from alternative approaches selectively 7 For example, the use of epidermal growth factor receptor inhibitors targeting ERCC1 deficiency. in EGFR-mutated tumours,4 as well as anaplastic lymphoma kinase In this study, we performed a high-throughput drug screening inhibitors in ALK-translocated diseases (10–15% and 5–7% of NSCLC, using an isogenic model of ERCC1-deficient NSCLC to discover respectively)5 represent significant advances in this area. new therapeutic approaches. Unfortunately, the vast majority of NSCLCs do not benefit from these treatments, highlighting the need for additional approaches. Excision repair cross-complementation group 1 (ERCC1), which RESULTS forms part of a key DNA repair enzyme active in the nucleotide A drug screen identifies poly (ADP-ribose) polymerase 1 and 2 excision repair pathway, is a promising predictive biomarker for inhibitors as being selectively toxic to ERCC1-deficient cells customised therapy.6–9 Low ERCC1 expression has been correlated In order to identify ERCC1-selective agents, we performed a drug with cisplatin sensitivity in several tumour types. As a heterodimer sensitivity screen by using a library of 80 drugs, either already with XPF (xeroderma pigmentosum complementation group F), used in oncology or in late-stage development.

1The Breakthrough Breast Cancer Research Centre and CRUK Gene Function Laboratory, Institute of Cancer Research, London, UK; 2De´partement de me´decine—Unite´ INSERM 981, Institut Gustave Roussy, Villejuif, France and 3Universite´ Paris-Sud XI, Paris, France. Correspondence: Dr CJ Lord or Professor A Ashworth, The Breakthrough Breast Cancer Research Centre and CRUK Gene Function Laboratory, Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, UK. E-mail: [email protected] or [email protected] Received 21 February 2013; revised 8 June 2013; accepted 10 June 2013; published online 12 August 2013 PARP1/2 inhibitors in ERCC1-deficient NSCLC S Postel-Vinay et al 5378 To maximise the potential for identifying ERCC1-selective selective effect as described below. Importantly, these selective effects, we used a recently generated isogenic panel14 of effects were consistent among all ERCC1-deficient clones and NSCLC-derived tumour lines,15 in which ERCC1 had been were observed using several different PARP1/2 inhibitors. inactivated by zinc finger-mediated gene targeting. In total, we used five isogenic NSCLC cell lines, including a parental ERCC1 Validation of the sensitivity of ERCC1-deficient cells to PARP1/2 wild-type A549 NSCLC cell line, one ERCC1-heterozygous inhibitors cell line (Ahez, expressing 65% of the original mRNA amount) As high-throughput screens often deliver false-positive results, we and three ERCC1-deficient clones (Ac216, Ac295 and Ac375, aimed to validate these results by assessing the PARP1/2 inhibitor expressing 6%, 18% and 15% of the originial protein amount, sensitivity of ERCC1-deficient cells in a number of different assay respectively) (Figure 1a and b). The clinical relevance of systems, including short-term assays (384- and 96-well assays) and these ERCC1-isogenic cell lines was confirmed by the extreme long-term colony-formation assays, utilising various PARP1/2 sensitivity to cisplatin of the ERCC1-deficient clones, which inhibitors. To minimise the potential for these ERCC1-selective were more than 100 times more sensitive than their ERCC1 wild- effects being due to the genetic drift often observed in tumour type and heterozygous counterparts (Figure 1c, Supplementary cell lines, we also assessed PARP1/2 sensitivity in ERCC1 wild-type Table S1). A549 cultures maintained at two different sites—the Institute of To identify ERCC1-selective effects, we plated each of the Cancer Research and Institut Gustave Roussy. Results obtained isogenic models in 384-well plates and exposed them to the drug with ERCC1 wild-type A549 cell lines from the Institute of Cancer library for 5 days. Each drug was represented at four concentra- Research and Institut Gustave Roussy were comparable tions (see Material and Methods). In total, we screened each of the (Supplementary Figure S2). Given the potential for using PARP1/2 isogenic cells in triplicate, combining this replica data in the final inhibitors in the clinical setting, we then focused on validating analysis. Only screens that met pre-defined quality criteria were these results by using two clinically relevant PARP1/2 inhibitors, considered for inclusion in this final data set (Supplementary namely olaparib (AZD-2281, Astra Zeneca, London, UK) and Figure S1A). To focus our analysis on ERCC1-selective effects, we niraparib (MK-4827, TesaroBio, Waltham, MA, USA). Both com- identified those drugs where there was a 415% difference in pounds displayed significant selectivity towards the ERCC1- surviving fraction between parental ERCC1-proficient and ERCC1- deficient clones, which were 10–100 times more sensitive to the deficient clones at two or more drug concentrations. This PARP1/2 inhibitors than their ERCC1-proficient counterpart (Figure approach identified 25 drugs for subsequent validation, including 2a and b; Supplementary Table S1). six different poly (ADP-ribose) polymerase 1 and 2 (PARP1/2) In addition to our original isogenic model, we also assessed the inhibitors (Supplementary Figure S1B and Supplementary Table generality of our findings by silencing ERCC1 by RNA interference. S2) that delivered ERCC1-selective effects among all ERCC1- Although we were able to generate ERCC1-deficient A549 clones deficient clones. Subsequent validation experiments, using the by gene targeting, siRNA-mediated silencing of ERCC1 in NSCLC same experimental procedure as for the high-throughput screen, models caused acute cytotoxicity (data not shown), precluding suggested that of the 25 drugs identified in the initial analysis, their use in siRNA experiments. However, we noted that Zhang only the PARP1/2 inhibitors showed a reproducible ERCC1- et al.16 had previously silenced ERCC1 in U2OS osteosarcoma cells with minimal cytotoxic effects. Using this system, we found that ERCC1 siRNA caused olaparib sensitivity compared with control 120 transfected cells. Strikingly, the effect of ERCC1 siRNA was comparable to the effect of BRCA2 siRNA (Figure 2c and d; 100 Supplementary Table S1).

80 A549 wt Ahez Ac216 Ac295 Ac375 ERCC1 isoform 202 rescues PARP1/2 inhibitor sensitivity in ERCC1-

parental cell line) ERCC1 60 deﬁcient NSCLC vs 40 As the use of isolated isogenic models do not always reﬂect the XPF impact of genetic heterogeneity on drug response, we examined 20 olaparib sensitivity in a panel of 14 NSCLC cell lines

Relative ERCC1 protein level (Supplementary Figure S3A). When comparing the olaparib 0 β expression (% actin sensitivity of NSCLC models with the expression of ERCC1 (as wt hez 216 295 375 detected by western blotting), we did not find a clear correlation Clone between reduced ERCC1 expression and olaparib sensitivity. 1.5 However, after examining cisplatin sensitivity in the same NSCLC A549 Ahez cell line panel (Supplementary Figures S3A and B), we found Ac216 that cisplatin sensitivity was significantly correlated to 1.0 Ac295 2 Ac375 olaparib sensitivity (r ¼ 0.5409, Po0.05, Pearson’s r correlation; Supplementary Figure S3B), despite the absence of any clear 0.5 correlation to the level of ERCC1 protein. ERCC1 is expressed as four distinct isoforms, 201, 202, 203 and Surviving Fraction 204.15 Isoforms 201, 203 and 204 lack amino acids encoded by 0.0 -2 -1 0 1 exons 10, 8 and 3, respectively, whereas ERCC1 isoform 202 is the [Cisplatin] Log μM only isoform to encompass the full XPA, XPF, MSH2, single-strand DNA and double-strand DNA binding domains.12,15 Very recent Figure 1. Isogenic model of ERCC1 deficiency. (a) Relative ERCC1 work has demonstrated that isoform 202 is a major determinant of expression quantified by MSD in the parental (wt) A549 cell line, the platinum sensitivity in NSCLC when compared with the other heterozygous ERCC1 (hez) cell line and three ERCC1-deficient clones. 15 (b) Western blot displaying ERCC1 and XPF expression for each cell isoforms. line. (c) Clonogenic survival experiment evaluating cisplatin sensi- We wanted to test whether the four distinct ERCC1 isoforms had tivity of ERCC1-deficient clones and ERCC1 heterozygous cell line differential effects on the PARP1/2 inhibitor response. compared with the parental cell line; error bars represent the s.d. By transfecting previously validated ERCC1 isoform cDNA expres- from the mean of three independent experiments. sion constructs15 into ERCC1-deficient A549 cells, we found that

Oncogene (2013) 5377 – 5387 & 2013 Macmillan Publishers Limited PARP1/2 inhibitors in ERCC1-deﬁcient NSCLC S Postel-Vinay et al 5379

Figure 2. Revalidation of the drug screen hits using multiple clinically relevant PARP1/2 inhibitors and multiple cellular models. (a) Hits revalidation using different clinical PARP1/2 inhibitors in 96-well plate (short-term assay). (b) Olaparib revalidation in colony-formation assay (long-term assay); DLD1 BRCA2 þ / þ and BRCA2 À / À are displayed as controls. Error bars represent the s.d. from the mean of three independent experiments. (c) Revalidation of olaparib sensitivity using siRNA silencing of ERCC1 in U2OS cells. Olaparib was added 48 h after reverse transfection and cells were exposed to the drug for 5 days. (d) Western blot showing ERCC1 and BRCA2 silencing after siRNA transfection in U2OS cells. (e) PARP1/2 inhibitor sensitivity is rescued by the reintroduction of the functional ERCC1 isoform.15 Error bars represent the s.d. from the mean of three independent experiments. (f) Clonogenic survival experiment evaluating BMN 673 sensitivity of ERCC1-deﬁcient clones. Error bars represent the s.d. from the mean of three independent experiments.

& 2013 Macmillan Publishers Limited Oncogene (2013) 5377 – 5387 PARP1/2 inhibitors in ERCC1-deficient NSCLC S Postel-Vinay et al 5380 the construct encoding isoform 202 restored PARP1/2 inhibitor mutation in other genes known to sensitise to PARP1/2 inhibitors resistance in ERCC1-deficient clones, whereas the other isoforms (data not shown). had no effect (Figure 2e; Supplementary Table S1), suggesting that similar to the response to cisplatin, the response to PARP1/2 The effect of ERCC1 deficiency on PARP1/2 inhibitor sensitivity is inhibitors was also determined by the ERCC1 isoform 202. epistatic with defects in genes that control the nuclear localisation of RAD51. To further investigate the mechanism by which ERCC1 deficiency led to PARP1/2 inhibitor sensitisation, we performed an ERCC1 deficiency in NSCLC sensitises cells to BMN 673, a novel olaparib siRNA sensitisation screen that simultaneously evaluated hyperpotent PARP1/2 inhibitor the effect of 911 different genes on the extent of PARP1/2 The majority of clinical PARP1/2 inhibitors have biochemical IC50 in inhibitor sensitivity in both ERCC1-proficient and -deficient clones. the nanomolar to micromolar range. BMN 673 is a highly potent For this screen, we used an siRNA library targeting kinase, tumour PARP1/2 inhibitor that selectively inhibits PARP1 at subnanomolar suppressor and DNA repair genes, an approach we have 17,18 concentrations, and is currently assessed in phase 1 clinical previously described elsewhere.19 We performed triplicate siRNA studies. We tested the effect of BMN 673 in our ERCC1-isogenic screens in the parental ERCC1-proficient A549 cell line and in the system and found that ERCC1-deficient clones were significantly two ERCC1-deficient clones that displayed the lowest levels of more sensitive to BMN 673 than their ERCC1-proficient ERCC1 expression (Supplementary Figure S5). We quantified the counterparts (Figure 2f; Supplementary Table S1). Consistent with effect of each siRNA in the library on olaparib sensitivity by the enhanced potency of this compound, the ERCC1-selective calculating drug effect (DE) Z-scores, where a Z-score of p À 2 was effect of BMN 673 was achieved at considerably lower concentra- used to define statistically significant olaparib sensitising effects. tions of PARP1/2 inhibitor than for the other clinical inhibitors By comparing sensitisation effects in ERCC1-proficient and (compare with Figures 2a and b; Supplementary Table S1). ERCC1-deficient clones, we found that siRNAs targeting well- established HR genes that control the localisation of RAD51 to the site of DNA damage, such as BRCA1, BRCA2, ATR and SHFM1 (aka Mechanistic dissection of NSCLC cell sensitivity to PARP1/2 DSS1) enhanced the olaparib sensitivity in ERCC1-proficient NSCLC inhibitors cells but not in ERCC1-deficient cells (Table 1). As a sign of the In order to understand the nature of PARP1/2 inhibitor sensitivity quality of the screens, BRCA2 siRNA were plated in duplicate in ERCC1-deficient NSCLC cells, we assessed a number of within the library, and both BRCA2 siRNA pools returned DE Z- molecular phenotypes associated with the response to PARP1/2 scores of o À 2 in the ERCC1-proficient cells but not in the ERCC1- inhibitors, namely; (i) the formation of nuclear RAD51 foci, (ii) the deficient clones, an effect we also independently validated effect of silencing homologous recombination (HR) genes, (iii) the (Figure 3; Supplementary Figure S6; Supplementary Table S1). formation and resolution of nuclear gH2AX foci, (iv) the effect on These observations suggested that ERCC1 deficiency and HR gene the cell cycle of PARP1/2 inhibitors and (v) the effect of PARP1 deficiency were in fact epistatic, such that the effect of modulating ablation on PARP1/2 inhibitor sensitivity. ERCC1 masked the phenotypic effect of modulating well-known HR genes. As observation of epistasis between genes is usually ERCC1-deficient NSCLC cells mount a PARP1/2 inhibitor-induced indicative of involvement in a shared process, it suggested that RAD51 response. The profound sensitivity of BRCA1 or BRCA2 although ERCC1 deficiency had no effect on the RAD51 response, mutant cells to PARP1/2 inhibitors is most likely caused by a ERCC1 function might be linked to the HR gene function in defect in the recruitment of the DNA recombinase RAD51 to sites response to PARP1/2 inhibitors. of DNA damage. In normal dividing cells, RAD51 recruitment (which can be monitored by visualising nuclear RAD51 foci using ERCC1-deficient cells display a delay in the repair of DNA damage immunocytochemistry) precedes DNA strand invasion as part of following PARP1/2 inhibitor exposure. We also investigated the the process of HR. We assessed whether ERCC1-deficient cells also ability of ERCC1-deficient cells to resolve DNA damage following displayed such a defect, by assessing the extent of nuclear RAD51 olaparib treatment. In addition to the formation of RAD51 nuclear foci formation in response to PARP1/2 inhibitor exposure. We foci, one of the other characteristics of exposure of PARP1/2 found that olaparib exposure elicited the formation of RAD51 foci inhibitors is the formation of nuclear gH2AX foci, a marker of the in both ERCC1-proficient and deficient models, and that ERCC1- phosphorylation of histone H2AX at the site of DNA double-strand deficient cell lines did not show the overt RAD51 defect found in breaks (DSBs) and stalled replication forks. We exposed ERCC1- BRCA-deficient models (Supplementary Figure S4). This suggested proficient and -deficient NSCLC cells to olaparib for 24 h, and that ERCC1 deficiency in NSCLC cells did not abrogate RAD51 monitored gH2AX foci formation after the drug had been function as a mechanism of PARP1/2 inhibitor sensitivity. Whole- removed from the culture media, using immunocytochemistry. exome sequencing of the clones also confirmed the absence of Before olaparib exposure, the frequency of cells with gH2AX foci in

Table 1. Results of a siRNA screen combined with olaparib in the ERCC1-isogenic model

siRNA A549a A549b Ac216 Ac295

BRCA2 À 4.00126919 À 9.945721698 À 0.731983661 À 0.604321416 BRCA2 À 3.160561281 À 11.63386025 1.333277082 À 0.302767388 ATR À 2.088426264 À 4.595545013 À 0.545177417 À 0.588415503 SHFM1 À 2.631896939 À 4.680863072 0.812441354 0.8319544 BRCA1 À 2.205340175 À 3.071777504 À 0.069539892 À 1.774590349 Abbreviations: DE, drug effect; ERCC1, excision repair cross-complementation group 1; siRNA, small interfering RNA. DE values of a siRNA screen evaluating 911 kinase, tumour suppressor and DNA repair genes, combined with olaparib treatment at non-toxic concentration (80% surviving fraction). The screen was performed in triplicate. The sensitising effect of each siRNA (DE) was estimated using a DE Z-score of p À 2 to define statistically significant olaparib sensitising effects. BRCA1, BRCA2, ATR and SHFM1 were the only siRNA sensitising the ERCC1-proficient cells only to olaparib, as assessed by a DE Z-score p À 2. A549a and A549b represent two individual screens independently performed in triplicate each in A549 cell line; Ac216: ERCC1-deficient clone 216; Ac375: ERCC1- deficient clone 375.

Oncogene (2013) 5377 – 5387 & 2013 Macmillan Publishers Limited PARP1/2 inhibitors in ERCC1-deﬁcient NSCLC S Postel-Vinay et al 5381

Figure 3. BRCA2 silencing is epistatic with ERCC1 deﬁciency in mediating PARP1/2 inhibitor sensitivity. Effect of BRCA2 knockdown by siRNA on sensitivity of ERCC1-isogenic cell lines to olaparib. Cells were reverse-transfected with BRCA2 siRNA and drug was added 48 h after transfection. Cells were exposed to the drug for 5 days. Error bars represent the s.d. from the mean of three independent experiments.

untreated ERCC1-proficient and -deficient clones was not PARP1 silencing causes PARP1/2 inhibitor resistance in ERCC1- significantly different, with all clones exhibiting B10% of cells deficient NSCLC cells. Several overlapping mechanisms have been with more than 10 gH2AX foci (data not shown). After 24 h of suggested to explain the cytotoxicity of PARP1/2 inhibitors, olaparib exposure, the frequency of cells with gH2AX foci including the formation of DNA DSBs subsequent to the failure increased, with 55–80% of cells exhibiting more than 10 gH2AX of single-strand break repair caused by PARP1 inhibition.14 More foci, regardless of the ERCC1 genotype (see time point T ¼ 0, recently, the observation that the cytotoxic response to small Figure 4a, frequency of cells with more than 10 gH2AX foci A549 molecule PARP1/2 inhibitors can be abrogated by the genetic vs Ac216: P ¼ 0.24; A549 vs Ac295: P ¼ 0.06; A549 vs Ac375: suppression of PARP1 levels has led to the hypothesis that PARP1 P ¼ 0.06, Student’s t-test). By contrast, the resolution of gH2AX foci trapped onto DNA as a result of its catalytic inhibition might be a after PARP1/2 exposure was significantly delayed in ERCC1- key cytotoxic DNA lesion.20,21 This observation is consistent with deficient clones when compared with the ERCC1-proficient the idea that auto-PARylation of PARP1 is required for the parental NSCLC cells, with 25–40% of ERCC1-deficient cells dissociation of this enzyme from damaged DNA and that, in the exhibiting more than 10 gH2AX foci, compared with only 8% in absence of a PARP1 substrate, the PARP1/DNA lesion is not the ERCC1-proficient parental clone at 76 h after drug removal formed, resulting in a minimisation of the effects of PARP1/2 (A549 vs Ac216: P ¼ 0.002; A549 vs Ac295: P ¼ 0.002; A549 vs inhibitors in certain contexts. Ac375: P ¼ 0.004, Student’s t-test, Figure 4a, Supplementary Figure We assessed whether silencing of PARP1 also minimised the S7). These observations were consistent with the hypothesis that cytotoxic effects of PARP1/2 catalytic inhibitors in ERCC1-deficient ERCC1-deficient cells displayed a defect in the resolution of DNA NSCLC cells. We found that PARP1 siRNA transfection rescued damage caused by PARP1/2 inhibitors. PARP1/2 inhibitor sensitivity in ERCC1-deficient clones, but PARP1 We also assessed the cell cycle response to olaparib exposure in depletion did not affect the sensitivity of ERCC1-proficient cells to ERCC1-deficient NSCLC cells. As in the previous experiment, PARP inhibition (Figure 5; Supplementary Figure S8; Supplementary we exposed cells to olaparib for 24 h and then monitored the Table S1). This suggested that the selective cytotoxicity of PARP1/2 changes in the cell cycle after the removal of olaparib, using flow inhibitors towards ERCC1-deficient cells may be primarily mediated cytometry. Although both ERCC1-deficient and ERCC1-proficient by the trapping of PARP1 onto the DNA. models exhibited a G2/M arrest in response to olaparib exposure, this arrest was much more profound and prolonged in ERCC1- deficient cells (Figure 4b, % cells in G2 at drug removal for A model for ERCC1-deficient NSCLC sensitivity to PARP1/2 A549 ¼ 26.7, Ac216 ¼ 51.8, Ac295 ¼ 51.8 and Ac375 ¼ 54.9; inhibitors Supplementary Table S3). This difference in G2/M arrest was most Our mechanistic dissection of PARP1/2 inhibitor sensitivity in pronounced 6 h after drug removal (% cells in G2 at 6 h after drug ERCC1-deficient NSCLC suggested the following: (i) ERCC1- removal for A549 ¼ 31.1, Ac216 ¼ 64.3, Ac295 ¼ 59.7 and deficient NSCLC cells are not profoundly deficient in terms of Ac375 ¼ 63.3), coinciding with the maximal formation of the RAD51 foci response or BRCA1/BRCA2 expression; (ii) ERCC1 gH2AX foci (Figure 4a and b), consistent with the hypothesis that deficiency is epistatic with HR gene silencing in terms of PARP1/2 the resolution of DNA damage in ERCC1-deficient clones was inhibitor sensitivity; (iii) gH2AX foci resolution in response to delayed in response to a PARP1/2 inhibitor, when compared with PARP1/2 inhibitors is delayed in ERCC1-deficient cells; (iv) ERCC1-proficient cells. persisting DNA damage is observed in ERCC1-deficient cells

& 2013 Macmillan Publishers Limited Oncogene (2013) 5377 – 5387 PARP1/2 inhibitors in ERCC1-deficient NSCLC S Postel-Vinay et al 5382 compared with that in wild-type counterparts following PARP1/2 replication fork restart and cell cycle progression. The necessity inhibitor exposure, which results in a delay in cell cycle for ERCC1 activity on the PARP1/DNA lesion before HR can restore progression; and (v) silencing of PARP1 before PARP1/2 inhibitor the replication fork is consistent with the epistasis observed treatment is able to minimise the ERCC1-selective effects of between HR genes and ERCC1 deficiency in terms of olaparib PARP1/2 inhibitors. sensitivity. Although a number of possible scenarios might explain these observations, the following proposed working model may be most consistent with the data (Figure 6): (i) PARP1 binds DNA in DISCUSSION response to a commonly occurring DNA insult but, in the presence One major challenge in the era of personalised medicine is the of a catalytic inhibitor, is trapped onto DNA (Figure 6a). This is identification of predictive biomarkers for drug response. ERCC1 consistent with recent data20 and the observation that silencing expression has previously been correlated with cisplatin response PARP1 by siRNA causes PARP1/2 resistance in ERCC1-deficient in NSCLC and other tumour types.6,7,22–28 Our screen identified NSCLC cells (Figure 5). (ii) When cells are in S phase, DNA-trapped PARP1/2 inhibitors as a potential novel therapeutic strategy for PARP1 stalls the oncoming replication fork (Figure 6b) and causes ERCC1-deficient NSCLC cells. We show that ERCC1-deficient NSCLC a gH2AX response (as demonstrated in Figure 4). In some cases, cell line models are not only sensitive to a range of clinical PARP1/ fork arrest leads to replication fork collapse and formation of a 2 inhibitors, but also that ERCC1 isoform 202, the isoform that DNA DSBs (Figures 6c and d), consistent with the formation of modulates cisplatin response, also causes PARP1/2 inhibitor RAD51 foci in both ERCC1-deficient and -proficient NSCLC cells resistance in NSCLC models. The epistasis between ERCC1 (Supplementary Figure S4). As stalling occurs upstream of the DNA dysfunction and HR gene silencing in terms of PARP1/2 inhibitor lesion, the creation of the DSB does not allow the removal of sensitivity, together with the lack of a profound RAD51 trapped PARP1. (iii) The DSB creation results in the formation of a dysfunction in ERCC1-deficient cells, suggests that the role of branched structure on the 50 side of the DNA, thereby creating a ERCC1 in the processing of PARP1/2 inhibitor-related DNA lesion substrate for the ERCC1/XPF DNA endonuclease, which excises might not be in HR itself but rather in the processing of the DNA and removes trapped PARP1 (Figure 6e and f). In the absence of lesion as a precursor to its final repair by RAD51-mediated HR. ERCC1, the DNA lesion is presumably not processed past point E; Together with previous data suggesting the nature of DNA lesions 20 cells remain trapped in S phase and display G2/M arrest, the caused by PARP1/2 inhibitors, we propose a model in which gH2AX response is still activated (both shown in Figure 4), and as PARP1 itself trapped on the DNA by PARP1/2 inhibitor might DNA DSBs are particularly lethal, cells either die at this point or use constitute a substrate lesion for ERCC1/XPF—before HR—which alternative forms of repair that are presumably suboptimal, thus would cause the selectivity observed. impairing their overall fitness. (iv) In ERCC1-proficient cells, gap PARP1/2 inhibitors have shown remarkable activity in BRCA- filling is performed after PARP1 excision by ERCC1/XPF via deficient breast and ovarian cancers.14 The present study provides conventional DNA polymerases (Figure 6g), which generates a evidence that PARP1/2 inhibitor-selective sensitivity may not be final substrate for HR (Figure 6h), eventually followed by limited to this population. Interestingly, ERCC1 also emerged as a

** 100 100 0 0 1 1 2 2 3-5 3-5 6-10 6-10 >10 50 >10 50 Foci Foci number H2AX foci (% of total) H2AX foci (% of total) number γ γ

0 0 Cells with Cells with T0 T4 T12 T28 T52 T76 T0 T4 T12 T28 T52 T76 A549 –Time post olaparib-removal (hrs) Ac216 – Time post olaparib-removal (hrs)

** 100 ** 100 0 0 1 1 2 2 3-5 3-5 6-10 6-10 >10 50 >10 50 Foci Foci H2AX foci (% of total) H2AX foci (% of total)

number γ number γ

0 Cells with 0 Cells with T0 T4 T12 T28 T52 T76 T0 T4 T12 T28 T52 T76 Ac295 – Time post olaparib-removal (hrs) Ac375 – Time post olaparib-removal (hrs) Figure 4. Kinetics of gH2AX foci formation and cell cycle analysis following olaparib treatment. (a) Quantification of gH2AX foci per cell following olaparib treatment in isogenic ERCC1-proficient and ERCC1-deficient cell lines. Cells were continuously exposed to 10 mM olaparib for 24 h and drug was removed (T0). Foci were counted at different time points following olaparib removal. The proportion of cells presenting more than 10 gH2AX foci was not significantly different between ERCC1-proficient and ERCC1-deficient cell lines at T0 (A549 vs Ac216: P ¼ 0.24; A549 vs Ac295: P ¼ 0.06; A549 vs Ac375: P ¼ 0.06, Student’s t-test); by contrast, this proportion was significantly higher at 76 h after drug removal in ERCC1-deficient clones as compared with the parental cell line (A549 vs Ac216: P ¼ 0.002; A549 vs Ac295: P ¼ 0.002; A549 vs Ac375: P ¼ 0.004) (b) Fluorescence-activated cell sorting profile of ERCC1-proficient and ERCC1-deficient cells before and after olaparib treatment at different time points. Cells were exposed to 10 mM olaparib for 24 h and drug was removed. Cell were stained with propidium iodide at several time points after drug removal (DR) for analysis of the DNA content and cell cycle phase. **Po0.005.

Oncogene (2013) 5377 – 5387 & 2013 Macmillan Publishers Limited PARP1/2 inhibitors in ERCC1-deﬁcient NSCLC S Postel-Vinay et al 5383 ERCC1-deficient cells ERCC1-proficient A549 cells Ac216 Ac295 Ac375 DMSO Drug Removal DR+6h DR+24h DR+5 days DR+4 days DR+3 days DR+30h

Figure 4. Continued. determinant of PARP1/2 inhibitor sensitivity in a wide siRNA unpublished data). Our ﬁndings add to the panel of clinically screen designed to identify modiﬁers or olaparib response, relevant DNA repair genes that modulate the cellular response to with a Z-score of À 2.248 (Postel-Vinay, Lord, Ashworth, these agents, including PTEN, ATM, ATR, CDK1, CHEK1 and CHEK2,

& 2013 Macmillan Publishers Limited Oncogene (2013) 5377 – 5387 PARP1/2 inhibitors in ERCC1-deﬁcient NSCLC S Postel-Vinay et al 5384

Figure 5. PARP1 silencing rescues PARP1/2 inhibitor sensitivity in the ERCC1-deﬁcient population. Effect of PARP1 knockdown by siRNA on sensitivity of ERCC1-isogenic cell lines to niraparib. Cells were reverse-transfected with PARP1 siRNA and drug was added 48 h after transfection. Cells were exposed to the drug for 5 days. Error bars represent the s.d. from the mean of three independent experiments. Silencing of PARP1 was checked by western Blot (Supplementary Figure S8B).

and the FANC family of genes.19,20,29,30 The observation that only challenge—in classifying cell lines according to ERCC1 status is the ERCC1 isoform 202 was able to rescue the PARP1/2 inhibitor- absence of reliable assay, as the strong similarity among all selective effect also suggests that the processing of PARP1/2 isoforms precludes distinguishing ERCC1 isoform 202 (the unique inhibitor-generated DNA lesions might be more similar to the functional isoform) from other non-functional isoforms.15 molecular response to platinum adducts that was previously Expression of non-functional isoforms can therefore result in thought.15,31–33 Moreover, the relative correlation observed in the misclassification, and thus the development of functional assays, non-isogenic panel of 14 NSCLC cell lines between cisplatin such as the duolink technology that detects the ERCC1/XPF sensitivity and olaparib sensitivity also supports this hypothesis. heterodimer, will be crucial to overcome this hurdle and create a Taken together, these observations support the proposition that meaningful classification of ERCC1 functionality. With regards to platinum sensitivity could be a surrogate biomarker of PARP1/2 our isogenic model, the experiment we show here (Figure 2e), inhibitor sensitivity. Platinum administration has to be halted after where PARP1/2 inhibitor sensitivity rescue is observed when re- a few cycles, and platinum-sensitive patients could benefit from expressing ERCC1 functional isoform 202, provides evidence that ‘switch maintenance therapy’ (that is, introduction of a new agent PARP1/2 inhibitor sensitivity is very likely a direct consequence of following platinum-based therapy) in order to prolong tumour ERCC1 deficiency. shrinkage. Given the excellent tolerability profile of PARP1/2 Our mechanistic dissection of the sensitivity of ERCC1-deficient inhibitors as monotherapy, these agents could be evaluated as cells to PARP1/2 inhibitors revealed that ERCC1 deficiency was switch maintenance therapy in platinum-sensitive NSCLC patients. epistatic with HR deficiency towards PARP1/2 inhibitor sensitivity. In Furthermore, PARP1/2 inhibitors as monotherapy could be used as addition, we found that ERCC1-deficient cells displayed a significant first-line treatment (as an alternative to platinum) for NSCLC delay in DNA damage repair associated with a G2/M cell cycle arrest patients with ERCC1-deficient tumours, who are not eligible for following PARP1/2 exposure. A similar observation was previously platinum-based treatments for reasons such as poor performance described in ERCC1-null myoepithelial fibroblasts and embryonic status or comorbidities. stem cells following mitomycin C exposure in a study investigating The use of a relatively novel isogenic model of ERCC1 deficiency the role of ERCC1/XPF in the removal of DNA interstrand exemplifies the utility that such genetically controlled systems can crosslinks.35 Together with our observation that PARP1 silencing have in the identification of synthetic lethalities. Very recently, could rescue PARP1/2 inhibitor sensitivity, this suggests that ERCC1/ Cheng et al.34 reported the potential for using PARP1/2 inhibitors XPF may be involved in the removal of a lesion constituted of combined with platinum in ERCC1-low cells, using a non-isogenic PARP1 trapped onto the DNA by the PARP1/2 inhibitor.11,20 This panel of four NSCLC cell lines and two different PARP1/2 working model is consistent with the recent description of the inhibitors—namely veliparib (ABT888; Abbott, Abbott Park, IL, crystal structure of PARP1 bound to a DNA break:36,37 the major USA) and olaparib (AZD2281; Astra-Zeneca). This represents a bulk created by trapped PARP1 may support that removing PARP1 different but complementary approach to the approach we have from the damaged DNA strand is required for the DNA repair taken here; isogenic models have the advantage of limiting the machinery to have access to the intact DNA strand. Furthermore, number of genetic changes between wild-type and mutant cells, recent observations by Pommier and colleagues20 provide strong so that differences observed can largely be explained by changes evidence for PARP1 ‘trapping’ by PARP1/2 inhibitors. In addition, in the gene of interest. Although non-isogenic panels may better the limited double-helix distortion in the latter working model represent the impact of tumoural genetic and epigenetic favours that PARP1/2 sensitivity is related to the role of ERCC1 in heterogeneity, the results from non-isogenic analyses are often DSB repair rather than in nucleotide excision repair. more difficult to interpret given the number of genetic variables in In conclusion, high-throughput drug screens performed in an a non-isogenic cell line panel.29 Furthermore, a major pitfall—and isogenic model of ERCC1-deficient NSCLC cell lines identified

Oncogene (2013) 5377 – 5387 & 2013 Macmillan Publishers Limited PARP1/2 inhibitors in ERCC1-deﬁcient NSCLC S Postel-Vinay et al 5385

5’ 5’ ERCC1-deficient cells die PARP1 PARP1 RAD51 ERCC1 XPF DNA replication 5’ Excision of the lesion and processing of the DSB to prepare HR PARP1 5’

γH2AX response Replication fork RAD51 stalling 5’ Gap-filling

5’ PARP1

DS γ H2AX B RAD51 response Replication fork Homologous collapse recombination

5’ 5’ Replication restart PARP1 in ERCC1-proficient cells

Induction of RAD51 response

Recognition of the lesion by ERCC1/XPF Figure 6. Proposed model for explaining PARP1/2 inhibitors selectivity in ERCC1-deficient cells. (a) PARP1 binds DNA in response to a commonly occurring DNA damage, for example, following the formation of spontaneous single-strand break, but is trapped onto the DNA by the PARP1/2 inhibitor. (b) During DNA replication, PARP1 bound to DNA causes stalling of the replication fork. (c) This leads to fork regression and formation of a DSBs, which is ERCC1 independent. The corresponding gH2AX response can be detected by the formation of gH2AX foci. (d) The resulting structure creates a substrate for the ERCC1/XPF endonuclease. (e) ERCC1/XPF removes the lesion, while the DSB is processed to prepare for homologous recombination. In the absence of ERCC1, cells either die because of the toxicity of unresolved DSBs, or these DNA lesions are repaired by processes that ultimately impair cellular fitness. (f) In ERCC1-proficient cells, the PARP1 lesion is removed, gap filling is ensured by conventional DNA polymerase (g) and HR can then occur (h), allowing the restart the replication fork.

PARP1/2 inhibitors as being selectively toxic to ERCC1-deficient Protein analysis, western blotting and immunocytochemistry cells. Clinical trials in appropriately selected patients, associated Whole-cell protein extracts were prepared from cells lysed in NP250 buffer with translational studies to further examine the determinants of (20 mM Tris pH 7.6, 1 mM EDTA, 0.5% NP40, 250 mM NaCl) supplemented PARP1/2 sensitivity in this context, are warranted. with protease inhibitor cocktail tablets (Roche, West Sussex, UK). Western blots were carried out with precast Bis-Tris gels (Invitrogen, Paisley, UK) as described previously.38 Staining, visualisation and quantification of gH2AX MATERIALS AND METHODS and RAD51 foci by confocal microscopy was performed as described 14,39 Cell lines and compounds previously, after 24 h of treatment with 10 mM of olaparib. The generation of the ERCC1-deficient A549 cell lines using zinc finger nuclease gene targeting has been described previously, along with Cell-based assays methods for re-expressing different ERCC1 isoforms.15 A549 cells and U2OS cells were cultured respectively in Dulbecco’s modified Eagle’s and Short-term survival assays were performed in 96-well plates. Exponentially Mc Coy’s medium with 10% fetal calf serum. All cell lines were short growing cells were plated at a concentration of 400 (A549), 500 (Ac216, Ac295 tandem repeats DNA typing (STR-typed) in our institution using StemElite and Ac375) or 1000 (U2OS) cells/well. Drug was added 24 h after seeding and ID (Promega, Madison, WI, USA). Olaparib (AZD-2281) and MK-4827 cells were continuously exposed to the drug for 5 days, after which cell viability was estimated using CellTitre-Glo luminescence (Promega). (Niraparib; TesaroBio) were obtained from Selleck Chemicals (Houston, TX, 14,40 USA). BMN 673 was provided by Dr Jerry Shen and Len Post at BioMarin Clonogenic assay were performed as previously described. (Novato, CA, USA). Cells were seeded in 6-well plates (500 cells/well) and continuously exposed to drug 24 h after seeding for 14 days. Media containing fresh drug was replaced every 72 h. Cells were fixed with 10% trichloroacetic Antibodies acid and stained with sulphorhodamine B (Sigma-Aldrich, Gillingham, UK). Antibodies targeting the following epitopes were used: ERCC1 (3H11/sc- Colonies were counted manually and using a colony counting machine 53281), PARP1 (F2/sc-8007), RAD51 (H-92/sc-8349), actin (I-19/sc-1616) (all (ColCount, Oxford Optronix, Abingdon, UK). Survival fractions were from Santa Cruz, Dallas, TX, USA); gH2AX (phospho ser139, JBW301, from calculated and dose–response curves were plotted as previously Millipore, Watford, UK); XPF (ab85140, from Abcam, Cambridge, UK); BRCA2 described.14 All cell-based assays experiments were performed at least in (OP-95, from Calbiochem, Nottingham, UK). triplicate.

& 2013 Macmillan Publishers Limited Oncogene (2013) 5377 – 5387 PARP1/2 inhibitors in ERCC1-deficient NSCLC S Postel-Vinay et al 5386 Drug screen REFERENCES We used an in-house drug library encompassing 80 drugs either used in 1 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics 2009. Cancer J clinical practice or in late-stage development. Each compound was Clin 2009; 59: 225–249. dissolved in 100% dimethyl sulphoxide (DMSO) to give 5 mM stocks and 2 Andre F, McShane LM, Michiels S, Ransohoff DF, Altman DG, Reis-Filho JS et al. then diluted to 0.5, 0.05, 0.005 and 0.0005 mM stocks in 96-well two- Biomarker studies: a call for a comprehensive biomarker study registry. Nat Rev dimensional matrix plates. Daughter plates in 384-well format were Clin Oncol 2011; 8: 171–176. prepared from these 96-well two-dimensional matrix racks using the 3 Curtin NJ. DNA repair dysregulation from cancer driver to therapeutic target. Nat Hamilton Microlab Star robotic platform. Compounds were stored under a Rev Cancer 2012; 12: 801–817. nitrogen atmosphere using a StoragePod (Roylan Developments, Leather- 4 Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, Brannigan BW et head, UK). al. Activating mutations in the epidermal growth factor receptor underlying Cell lines were seeded (500 cells/well) into 384-well plates using a responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 2004; 350: MultiDrop Combi Dispenser (Thermo Fisher Scientific, Leicestershire, UK) 2129–2139. and incubated overnight at 37 1C, 5% CO2. Replicate cell plates were then 5 Camidge DR, Bang YJ, Kwak EL, Iafrate AJ, Varella-Garcia M, Fox SB et al. Activity loaded onto Microlab Star screening platform and drug plates were serially and safety of crizotinib in patients with ALK-positive non-small-cell lung cancer: diluted in Dulbecco’s modified Eagle’s medium before being added to the updated results from a phase 1 study. Lancet Oncol 2012; 13: 1011–1019. cell plates. The final drug concentrations used for each drug were 1000, 6 Olaussen KA, Dunant A, Fouret P, Brambilla E, Andre F, Haddad V et al. DNA repair 100, 10 and 1 nM. The final DMSO concentration in all wells was 0.2% (v/v). by ERCC1 in non-small-cell lung cancer and cisplatin-based adjuvant Controls included 0.2% (v/v) DMSO and 10 mM staurosporine (Sigma- chemotherapy. N Engl J Med 2006; 355: 983–991. Aldrich). After incubation in drug-containing media for 4 days, cell viability 7 Postel-Vinay S, Vanhecke E, Olaussen KA, Lord CJ, Ashworth A, Soria JC. The was quantified with CellTiter-Glo (Promega) using a Victor X5 Multilabel potential of exploiting DNA-repair defects for optimizing lung cancer treatment. plate reader luminescence protocol (Perkin Elmer, Waltham, MA, USA). Nat Rev Clin Oncol 2012; 9: 144–155. Luminescence data from each well was normalised to the median signal 8 Simon GR, Schell MJ, Begum M, Kim J, Chiappori A, Haura E et al. Preliminary from DMSO-containing wells to calculate the survival fraction. indication of survival benefit from ERCC1 and RRM1-tailored chemotherapy in patients with advanced nonsmall cell lung cancer: evidence from an individual siRNA silencing patient analysis. Cancer 2012; 118: 2525–2531. 9 Vilmar AC, Santoni-Rugiu E, Sorensen JB. ERCC1 and histopathology in advanced All siRNA silencing experiments were performed using a SMARTpool of NSCLC patients randomized in a large multicenter phase III trial. Ann Oncol 2010; four distinct siRNA species targeting different sequences of the target 21: 1817–1824. transcript (Dharmacon, Thermo Fisher Scientific, Leicestershire, UK). Cells 10 Fagbemi AF, Orelli B, Scharer OD. Regulation of endonuclease activity in human were reverse-transfected using RNAimax (Invitrogen) transfection reagent. nucleotide excision repair. DNA Repair 2011; 10: 722–729. Transfection efficacy was assessed by independently transfecting cells 11 Kirschner K, Melton DW. Multiple roles of the ERCC1-XPF endonuclease in DNA concomitantly with PLK1 siRNA, which produced more than 95% cell repair and resistance to anticancer drugs. Anticancer Res 2010; 30: 3223–3232. growth inhibition. Validation of RNAi gene silencing was performed by 12 Tripsianes K, Folkers G, Ab E, Das D, Odijk H, Jaspers NG et al. The structure of the western blotting from pools of concomitantly transfected cells, as human ERCC1/XPF interaction domains reveals a complementary role for the two described above. proteins in nucleotide excision repair. Structure 2005; 13: 1849–1858. 13 Bergstralh DT, Sekelsky J. Interstrand crosslink repair: can XPF-ERCC1 be let off the siRNA screen with olaparib hook? Trends Genet 2008; 24: 70–76. An siRNA library (784 kinases and tumour suppressor genes, and 127 DNA 14 Farmer H, McCabe N, Lord CJ, Tutt AN, Johnson DA, Richardson TB et al. Targeting repair genes) was purchased from Dharmacon. Each well contained a the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature 2005; SMART pool of four distinct siRNA species targeting different sequences of 434: 917–921. the target transcript. Each plate was supplemented with negative 15 Friboulet L, Olaussen KA, Pignon JP, Shepherd FA, Tsao M, Graziano S et al. ERCC1 siCONTROL (12 wells; Dharmacon) and positive control (4 wells, siPLK1, isoform expression and DNA repair in non-small cell lung cancer. N Engl J Med Dharmacon). RNAi screening conditions were optimised and raw CellTitre- 2013; 368: 1101–1110. Glo (Promega) luminescent viability readings were performed as previously 16 Zhang YW, Regairaz M, Seiler JA, Agama KK, Doroshow JH, Pommier Y. Poly(ADP- described.41 Drug or vehicle (DMSO) was added 48 h after transfection at ribose) polymerase and XPF-ERCC1 participate in distinct pathways for the repair of topoisomerase I-induced DNA damage in mammalian cells. Nucleic Acids Res 1 mM concentration in media and cells were exposed to olaparib for 5 days. Statistical analysis of the siRNA screen was performed as described 2011; 39: 3607–3620. elsewhere.38 17 Wang B, Chu D, Feng Y, Xin Y, Myers P, Post L et al. Novel PARP inhibitors with potent antitumor activity as single-agent and combination therapies. Mol Cancer Ther 2009; 8(Suppl 12): A121. Flow cytometry analysis 18 Patterson M, Murray J, Curtin NJ. Stability of PARP inhibition by BMN 673 in Cells were plated in 10-cm dishes and exposed to olaparib 10 mM 24 h after human PBMCs and Leukaemic Cell Cultures. Eur J Cancer 2012; 48(Suppl 6): 106 plating. After 1 day of drug exposure, cells were collected, fixed with (Abstract 348). ethanol and stained using propidium iodide solution (20 mg/ml PI and 19 Turner NC, Lord CJ, Iorns E, Brough R, Swift S, Elliott R et al. A synthetic lethal 100 mg/ml RNase A in phosphate-buffered saline). Total DNA content was siRNA screen identifying genes mediating sensitivity to a PARP inhibitor. Embo J quantified and analysed by flow cytometry on a fluorescence-activated cell 2008; 27: 1368–1377. scan cytometer. 20 Murai J, Huang SY, Das BB, Renaud A, Zhang Y, Doroshow JH et al. Trapping of PARP1 and PARP2 by Clinical PARP Inhibitors. Cancer Res 2012; 72: 5588–5599. 21 Kedar PS, Stefanick DF, Horton JK, Wilson SH. Increased PARP-1 association with CONFLICT OF INTEREST DNA in alkylation damaged, PARP-inhibited mouse fibroblasts. Mol Cancer Res 2012; 10: 360–368. AA and CJL are inventors on patents describing the use of PARP inhibitors and may 22 Yin M, Yan J, Martinez-Balibrea E, Graziano F, Lenz HJ, Kim HJ et al. ERCC1 benefit from the ICR ‘Rewards to inventors scheme’. All other authors declare no and ERCC2 polymorphisms predict clinical outcomes of oxaliplatin-based conflicts of interest. chemotherapies in gastric and colorectal cancer: a systemic review and meta- analysis. Clin Cancer Res 2011; 17: 1632–1640. 23 Langer CJ. Exploring biomarkers in head and neck cancer. Cancer 2012; 118: ACKNOWLEDGEMENTS 3882–3892. We thank David Roberston and Marieke Aarts for assistance with confocal microscopy 24 Vilmar AC, Sorensen JB. Customising chemotherapy in advanced nonsmall cell and FACS analysis, and Jerry Shen and Len Post at Biomarin for the provision of BMN lung cancer: daily practice and perspectives. Eur Respir Rev 2011; 20: 45–52. 673. We also acknowledge NHS funding to the NIHR Royal Marsden Hospital BRC. This 25 Metzger R, Bollschweiler E, Holscher AH, Warnecke-Eberz U. ERCC1: impact in work was supported by grants from Cancer Research UK and The European Union as multimodality treatment of upper gastrointestinal cancer. Future Oncol 2010; 6: part of the FP7 teams ‘DDResponse’ and ‘Eurocan’. SPV is supported by a 1735–1749. Translational Research Fellowship from the European Society of Medical Oncology 26 Steffensen KD, Waldstrom M, Jakobsen A. The relationship of platinum resistance (2011 and 2012) and by funding from the Institut National du Cancer: Bourse pour la and ERCC1 protein expression in epithelial ovarian cancer. Int J Gynecol Cancer Formation a` la Recherche Translationnelle (2011) 2009; 19: 820–825.

Oncogene (2013) 5377 – 5387 & 2013 Macmillan Publishers Limited PARP1/2 inhibitors in ERCC1-deficient NSCLC S Postel-Vinay et al 5387 27 Chen S, Zhang J, Wang R, Luo X, Chen H. The platinum-based treatments for 34 Cheng H, Zhang Z, Borczuk A, Powell CA, Balajee AS, Lieberman HB et al. PARP advanced non-small cell lung cancer, is low/negative ERCC1 expression better inhibition selectively increases sensitivity to cisplatin in ERCC1-low non-small cell than high/positive ERCC1 expression? A meta-analysis. Lung cancer 2010; 70: lung cancer cells. Carcinogenesis 2013; 34: 739–749. 63–70. 35 Niedernhofer LJ, Essers J, Weeda G, Beverloo B, de Wit J, Muijtjens M et al. The 28 Lord RV, Brabender J, Gandara D, Alberola V, Camps C, Domine M et al. Low structure-specific endonuclease Ercc1-Xpf is required for targeted gene replace- ERCC1 expression correlates with prolonged survival after cisplatin plus gemci- ment in embryonic stem cells. Embo J 2001; 20: 6540–6549. tabine chemotherapy in non-small cell lung cancer. Clin Cancer Res 2002; 8: 36 Langelier MF, Planck JL, Roy S, Pascal JM. Crystal structures of poly(ADP-ribose) 2286–2291. polymerase-1 (PARP-1) zinc fingers bound to DNA: structural and functional 29 Rehman FL, Lord CJ, Ashworth A. Synthetic lethal approaches to breast cancer insights into DNA-dependent PARP-1 activity. J Biol Chem 2011; 286: therapy. Nat Rev Clin Oncol 2010; 7: 718–724. 10690–10701. 30 Johnson N, Li YC, Walton ZE, Cheng KA, Li D, Rodig SJ et al. Compromised CDK1 37 Langelier MF, Planck JL, Roy S, Pascal JM. Structural basis for DNA damage- activity sensitizes BRCA-proficient cancers to PARP inhibition. Nat Med 2011; 17: dependent poly(ADP-ribosyl)ation by human PARP-1. Science 2012; 336: 728–732. 875–882. 38 Bajrami I, Kigozi A, Van Weverwijk A, Brough R, Frankum J, Lord CJ et al. Synthetic 31 Sijbers AM, van der Spek PJ, Odijk H, van den Berg J, van Duin M, Westerveld A et al. lethality of PARP and NAMPT inhibition in triple-negative breast cancer cells. Mutational analysis of the human nucleotide excision repair gene ERCC1. Nucleic EMBO Mol Med 2012; 4: 1087–1096. Acids Res 1996; 24: 3370–3380. 39 Graeser M, McCarthy A, Lord CJ, Savage K, Hills M, Salter J et al. A marker of 32 Dabholkar M, Vionnet J, Parker R, Bostickbruton F, Dobbins A, Reed E. Expression homologous recombination predicts pathologic complete response to neoadju- of an alternatively spliced ercc1 messenger-RNA species, is related to reduced vant chemotherapy in primary breast cancer. Clin Cancer Res 2010; 16: 6159–6168. DNA-repair efficiency in human T-lymphocytes. Oncol Rep 1995; 2: 209–214. 40 Edwards SL, Brough R, Lord CJ, Natrajan R, Vatcheva R, Levine DA et al. Resistance 33 Sun Y, Li T, Ma K, Tian Z, Zhu Y, Chen F et al. The impacts of ERCC1 gene exon VIII to therapy caused by intragenic deletion in BRCA2. Nature 2008; 451: 1111–1115. alternative splicing on cisplatin-resistance in ovarian cancer cells. Cancer Invest 41 Lord CJ, Martin SA, Ashworth A. RNA interference screening demystified. J Clin 2009; 27: 891–897. Pathol 2009; 62: 195–200.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)

& 2013 Macmillan Publishers Limited Oncogene (2013) 5377 – 5387