Muts Homologue 2 and the Long-Term Benefit of Adjuvant Chemotherapy in Lung Cancer

Published OnlineFirst February 15, 2010; DOI: 10.1158/1078-0432.CCR-09-2204

Clinical Imaging, Diagnosis, Prognosis Cancer Research MutS Homologue 2 and the Long-term Benefit of Adjuvant Chemotherapy in Lung Cancer

Nermine S. Kamal1, Jean-Charles Soria1,2, Jean Mendiboure3, David Planchard1, Ken A. Olaussen1, Vanessa Rousseau3, Helmut Popper5, Robert Pirker6, Pascale Bertrand7, Ariane Dunant3, Thierry Le Chevalier4, Martin Filipits6, and Pierre Fouret1,8 for the International Adjuvant Lung Trial-Bio investigators

Abstract Purpose: We sought to determine the long-term (median follow-up, 7.5 years) predictive power of human MutS homologue 2 (MSH2) immunohistochemical expression in patients who enrolled in the International Adjuvant Lung Trial. Experimental design: We tested the interaction between MSH2 and the allocated treatment (chemotherapy versus observation) in a Cox model adjusted on clinicopathologic variables. The significance level was set at 0.01. Results: MSH2 levels were low in 257 (38%) and high in 416 (62%) tumors. The benefit from chemotherapy was likely different according to MSH2 (interaction test, P = 0.06): there was a trend for chemotherapy to prolong overall survival when MSH2 was low [hazard ratio (HR), 0.76; 95% confidence interval (95% CI), 0.59-0.97; P = 0.03], but not when MSH2 was high (HR, 1.12; 95% CI, 0.81-1.55; P = 0.48). In the control arm, the HR was 0.66 (95% CI, 0.49-0.90; P = 0.01) when MSH2 was high. When combining MSH2 with excision repair cross-complementing group 1 (ERCC1) into four subgroups, the benefit of chemotherapy decreased with the number of markers expressed at high levels (P = 0.01). A similar decrease was noted when combining MSH2 and P27 (P = 0.01). Chemotherapy prolonged overall survival in the combined low MSH2/low ERCC1 subgroup (HR, 0.65; 95% CI, 0.47-0.91; P = 0.01) and in the combined low MSH2/low P27 subgroup (HR, 0.65; 95% CI, 0.46-0.93; P = 0.01). Conclusions: MSH2 expression is a borderline significant predictor of a long-term benefit from adjuvant cisplatin-based chemotherapy in patients with completely resected lung cancer. MSH2 combined with ERCC1 or P27 may identify patients most likely to benefit durably from chemotherapy. Clin Cancer Res; 16(4); 1206–15. ©2010 AACR.

Platinum compounds are a hallmark of chemotherapy (NER) pathway and the mismatch repair (MMR) pathway against cancer due to their DNA binding capacity, which (1). The role that DNA excision repair pathways play in results in DNA damage and cell death. A single DNA mediating platinum resistance has been studied for many cross-link, if not repairable, can be lethal. The processing years, first in the preclinical, and more recently, in the clin- of cross-links in mammalian cells is not clearly under- ical setting (2). The underlying hypothesis of these studies stood. However, it is known that their processing may in- is that a reduced capacity by tumor cells to excise plati- volve components belonging to different DNA excision num-DNA adducts increases cell sensitivity to chemother- repair pathways, including the nucleotide excision repair apy, which in turn may translate into a clinical benefit. In 2004, the International Adjuvant Lung Trial (IALT) was the first and still is the largest clinical trial to establish Authors' Affiliations: 1Translational Research Laboratory, 2Paris-Sud – University, Department of Medicine, 3Biostatistics and Epidemiology that patients with non smallcelllungcancer(NSCLC) Unit, and 4Department of Medicine, Gustave-Roussy Institute, Villejuif, benefited from adjuvant cisplatin-based chemotherapy af- 5 France; Institute of Pathology, Medical University of Medical School of ter their tumors have been surgically completely removed Graz,Graz,Styria,Austria;6Department of Internal Medicine, Medical University of Vienna, Vienna, Austria; 7Department of Radiobiology and (3). A benefit from platinum-based adjuvant chemother- Radiopathology, CEA, Fontenay-aux-Roses, France; and 8Pierre et apy was further validated by the results of the ANITA, Marie Curie University, Paris, France CALGB, and JBR10 trials, and by their meta-analysis Corresponding Author: Pierre Fouret, INSERM U985, Institut de – Cancérologie Gustave-Roussy, 39 rue Camille-Desmoulins, 94805 Villejuif (4 7). When the survival data in the IALT were updated Cedex, France. Phone: 33-1-42116510; Fax: 33-1-42116094; E-mail: in 2007, however, an improvement of the long-term over- [email protected]. all survival from chemotherapy could no longer be ob- doi: 10.1158/1078-0432.CCR-09-2204 served due to an excess of late deaths in patients who ©2010 American Association for Cancer Research. underwent chemotherapy (8).

1206 Clin Cancer Res; 16(4) February 15, 2010

DNA Repair in Non–Small Cell Lung Cancer

function in the MMR pathway, MSH2 also recognizes and Translational Relevance binds to cisplatin-induced DNA interstrand cross-links, thereby initiating their excision and repair (17, 18). We report for the first time the ability of two DNA MSH2 is required to repair interstrand cross-links in mam- repair proteins MSH2 and ERCC1 and of the cyclin- malian cells, in which it physically interacts with the NER dependent kinase inhibitor P27 to predict the long- pathway component ERCC1 (19). During the recombina- term (median follow-up, 7.5 years) benefit from tional repair processing of interstrand cross-links, MSH2 adjuvant cisplatin-based chemotherapy in the Interna- cooperates with several components of DNA damage re- tional Adjuvant Lung Trial. As the magnitude of bene- pair pathways, including ERCC1, REV1, components of fit from chemotherapy decreased with time, the results the Fanconi anemia pathway, and homologous recombi- emphasize the importance of long-term follow-up in nation repair factors (20). Reduced MSH2 expression by clinical trials and translational studies. Our opinion tumor cells was reported in 10% to 58% of NSCLC (21– is that ERCC1, MSH2, and P27 are the three most 30). A recent study reported that the loss of MSH2 expres- promising markers of the benefit of adjuvant cisplat- sion in tumors from patients with advanced NSCLC led to in-based chemotherapy and should be tested in prior- higher rates of response to oxaliplatin-based chemothera- ity in validation studies. After validation, the most py, but not to cisplatin-based chemotherapy (30). In a efficacious markers will be used in the clinical practice larger study of genetic polymorphisms among patients either alone or in combination for recognizing patients with advanced NSCLC, the MSH2 gIV12-6T>C variant unlikely to benefit from conventional platinum-based was associated with low MSH2 expression and better re- therapies and to which new agents that are active in sponse to cisplatin, but the authors did not report a direct advanced stages of lung cancer can be proposed in relationship between MSH2 expression and chemotherapy the adjuvant setting. benefit (29). Using the long-term overall survival IALT data that were updated in 2007, we looked again at ERCC1 and studied for the first time the predictive power of MSH2 and the The IALT-BIO study was designed to examine whether a potential of combining MSH2 and ERCC1 to identify pa- limited number of markers predicted survival in relation- tients who benefited durably from chemotherapy. We also ship with chemotherapy. To reduce the risk of false- report the results of the combined analysis of MSH2 and positive results due to multiple comparisons, the signifi- P27 and discuss the potential usefulness of combining cance level was set at P = 0.01. Initially, 19 markers be- these markers. longing to five groups were studied: drug transporters, DNA repair, cell cycle regulators, signal transduction, and apoptosis. Only excision repair cross-complementa- Materials and Methods tion group 1 (ERCC1), a key NER protein also involved in homologous recombination (9) and maintenance of Patients and study design. All patients had participated in telomeres (10), significantly predicted (P = 0.009) the the IALT study (1,867 patients). The IALT-Bio study was short-term survival benefit of chemotherapy (11). Patients designed by a steering committee to examine whether tu- with low ERCC1 expression benefited from chemotherapy, mor markers assessed by immunohistochemical analysis whereas those with high ERCC1 expression did not. could be used to predict survival in relationship with che- Among the other markers, the closest to significantly pre- motherapy. The study was conducted according to a de- dict the benefit of chemotherapy was the cyclin-dependent tailed protocol that stressed the importance of collecting kinase inhibitor P27 (P = 0.02; ref. 12). Furthermore, all samples within the participating centers and required when ERCC1 and P27 were combined, the predictive va- a large number of tumor samples to ensure adequate pow- lues further increased (13). Other markers were either er for prognostic and predictive analyses. All 50 IALT cen- more weakly associated to the benefit of chemotherapy ters which had enrolled 10 or more patients (total 1,468 or not associated at all (14, 15). patients) were solicited to participate in the IALT-Bio The results with ERCC1 pointed to the potential of study. Twenty-eight centers in 14 countries agreed to par- studying the expression of DNA repair genes to tailor the ticipate. From their 1,045 patients, they contributed 867 use of adjuvant chemotherapy in lung cancer. Except for blocks, which represents an overall completeness rate of ERCC1, no DNA repair gene was included in the initial 83%. Approval was obtained from the local institutional IALT-Bio analysis. In a second phase of IALT-Bio, we thus review boards, according to the legal regulations in each tried to increase the predictive power of ERCC1 by exam- participating country. All tumors were reviewed centrally ining the expression of MutS homologue 2 (MSH2), a gene at the Centre Hospitalier Universitaire Albert Michallon, that is crucially involved in the repair of cisplatin-DNA according to the histopathologic classification system cross-links. MSH2, which is frequently mutated in heredi- adopted by the WHO in 2004 (31). After tissue collection tary nonpolyposis colon cancer, encodes a critical protein and a pathologic review, 783 tissue blocks were included of the MMR pathway (16). MSH2 binds to DNA mis- in the IALT-Bio study. The fixation procedure depended on matches, thereby initiating DNA repair. In addition to its each center and was not registered.

www.aacrjournals.org Clin Cancer Res; 16(4) February 15, 2010 1207

Kamal et al.

The quality of slides after H&E staining, referred included every factor used in the stratified randomization throughout the article as slide quality, was semiquantita- (center, tumor stage, and type of surgery) plus clinical and tively classified into two categories (average and good) histologic prognostic factors (age, sex, WHO performance taking into account staining contrast as well as cell integ- status, nodal status, lymphoid infiltration, and the revised rity and morphologic appearance. histopathologic type). All other factors that were statisti- Tissue microarray construction. Three representative tu- cally related to the biomarker status in the multivariate mor areas were selected for each case. Cores measuring logistic model (P < 0.05) were added to the Cox model. 0.6 mm in diameter and 5 mm in length (spots) were ar- The predictive value of the biomarker was studied by rayed following a map. Among the 783 IALT-Bio tissue testing the interaction between the biomarker status and blocks that contained tumor material, triplicate spots were the allocated treatment (chemotherapy versus observa- obtained in 768 (98%) cases. Random spots were com- tion) in the same Cox model. Tests of homogeneity of pared with the original blocks to verify the agreement be- the hazard ratios were done within the Cox model. For tween their coordinates and the original samples. the joint analysis of MSH2 and ERCC1, the Cox model in- MSH2 immunostaining. Immunostaining was done fol- cluded the same variables as above together with three lowing a standard procedure using the Vectastain Elite kit variables indicating the status of the combination of with NovaRED (Vector Laboratories) as the substrate and MSH2 and ERCC1 and a different treatment variable for Mayer's hematoxylin as the counterstain. The primary an- each of the four combinations. All reported P values were tibody was the mouse monoclonal antibody FE11 (Cal- two sided. In the IALT-Bio analysis plan, P values below biochem) raised against the COOH-terminal fragment of 0.01 were considered as statistically significant to limit human MSH2. For epitope retrieval, slides were heated at the risk of false-positive results. Survival rates were esti- 98°C for 1 h in 10 mmol/L citrate buffer (pH 7.3). Sec- mated using the Kaplan-Meier method (all P values indi- tions were incubated at room temperature for 90 min with cated, besides Kaplan-Meier curves, were adjusted P values the FE11 antibody at a dilution of 1:50. The freshly cut corresponding to the Cox analysis). tissue microarray (TMA) sections were manually immu- All analyses were done using the SAS software, version nostained in a single experiment that included a tonsil sec- 9.1 (SAS Institute, Inc.) and curves were drawn with the tion as an external control. Tigre software. Evaluation of MSH2 immunostaining. Two investigators (N.S.K. and P.F.), who were blinded to clinical data, inde- pendently evaluated MSH2 nuclear reactivity. The TMA Results slides were scanned at high resolution (VM3 virtual scan- ner, Ziemens), enabling the study of an identical high- MSH2 expression. Among the 768 patients whose tumor quality image at ×20 magnification for each spot for was included in the TMA, no tumor material could be an- detailed evaluation. alyzed after immunohistochemistry in 34 (4%) cases. After The spots were carefully examined for reactive lung or excluding the cases without valid internal controls, the stromal cells (endothelial cells and fibroblasts), which H-scores were evaluated in 673 of 768 (88%) patients. served as an internal positive control. Spots without a val- The H-scores were as follows: 0.2 in 1 case (0%), 1 in 52 id internal control were discarded. Cases for which no val- cases (8%), 2 in 363 cases (54%), and 3 in 257 cases id tumor spot could be evaluated were excluded. (38%). The median H-score was 2, which was chosen to Staining intensity was graded on a scale of 0 to 3, using separate high MSH2 cases (H-score = 3) and low MSH2 the expression level in fibroblasts or endothelial cells as a cases (H-score < 3). There were 257 (38%) high MSH2 reference (defined a priori as a score of 2). The percentage cases and 416 (62%) low MSH2 cases. Figure 1 shows of reactive tumor cells was graded on a scale of 0% to examples of high MSH2 and low MSH2 cases. 100%. A proportion score was assigned to the percentage of reactive tumor nuclei: 0 was assigned if 0% of reactive tumor nuclei were found; 0.1 was assigned if 1% to 9% were found; 0.5 was assigned if 10% to 49% were found; and 1 was assigned if ≥50% of reactive tumor nuclei were present. This proportion score was multiplied by the staining intensity to obtain a histology score (H-score) for each patient (11). All discordant cases were reviewed to reach a consensus. Statistical analysis. Long-term IALT survival data were used, with median follow-up of 7.5 years (8). A logistic model stratified by center was used to compare patients with high MSH2 and low MSH2 tumors.

The prognostic values of the biomarker status and che- Fig 1. Examples of high MSH2 and low MSH2 NSCLS. A, high MSH2. motherapy for overall survival were studied using the Cox The staining was intense and diffuse. B, low MSH2. Arrow, positive model. As in the main IALT analysis (3), the Cox model stromal cells (internal control).

1208 Clin Cancer Res; 16(4) February 15, 2010 Clinical Cancer Research

DNA Repair in Non–Small Cell Lung Cancer

Table 1. Patient characteristics

Characteristic Patients with high MSH2 Patients with low MSH2 All patients P* n tumors (n = 257) tumors (n = 416) ( = 673) Number Percent Number Percent Number Percent

Sex 0.05 Male 221 86 326 78 547 81 Female 36 14 90 22 126 19 Age 0.99 <55 years 70 27 127 31 197 29 55-64 years 116 45 181 44 297 44 >64 years 71 28 108 26 179 27 Pathological TNM stage† 0.79 Stage I 83 32 150 36 233 35 Stage II 53 21 102 25 155 23 Stage III 121 47 164 39 285 42 Tumor 0.83 T1 36 14 63 15 99 15 T2 145 56 263 63 408 61 T3 74 29 82 20 156 23 T4 2 1 8 2 10 1 Nodes 0.99 N0 122 47 188 45 310 46 N1 69 27 125 30 194 29 N2 66 26 103 25 169 25 Histological type 0.002 Adenocarcinoma 51 20 154 37 205 30 Squamous-cell carcinoma 174 68 217 52 391 58 Other NSCLC 32 12 45 11 77 11 Surgery 0.34 Pneumonectomy 118 46 161 39 279 41 Lobectomy or segmentectomy 139 54 255 61 394 59 Performance status scored‡ 0.37 0 147 57 222 53 369 55 1 84 33 165 40 249 37 2 26 10 29 7 55 8 Lymphoid infiltration 0.20 Not intense 235 91 360 87 595 88 Intense 22 9 56 13 78 12 Pleural invasion 0.19 No 232 90 386 93 618 92 Yes 25 10 30 7 55 8 Vascular invasion 0.44 No 181 70 295 71 476 71 Yes 76 30 121 29 197 29 Lymphatic invasion 0.19 No 71 28 137 33 208 31 Yes 186 72 279 67 465 69 Quality after H&E staining 0.01 Average 37 14 28 7 65 10 Good 220 86 388 93 608 90

NOTE: Percentages may not total 100 because of rounding. *P values testing the difference between positive and negative tumors were calculated using logistic regression stratified on centre. †TNM denotes tumor-node-metastases. ‡World Health Organization scores.

www.aacrjournals.org Clin Cancer Res; 16(4) February 15, 2010 1209

Kamal et al.

MSH2 expression and baseline clinical characteristics. The tients with low MHS2 tumors, median overall survival relationships between MSH2 expression and the clinical tended to be 16 months longer in the chemotherapy characteristics are provided in Table 1. The proportion of arm (58 months) than in the control arm (42 months). adenocarcinomas was lower (P =0.002)forhighMSH2 In the high MSH2 group, there was no difference in the cases (20%) than for low MSH2 cases (37%). The mor- overall survival between the chemotherapy arm and the phologic slide quality after H&E staining was associated control arm (adjusted hazard ratio for death, 1.12; 95% with MSH2 expression (P = 0.01). CI, 0.81-1.55; P = 0.48; Table 2, Fig. 2B). The 5-year overall In a logistic model adjusted on sex and slide quality, survival rates among patients with high MSH2 tumors histology differed according to MSH2 expression, with were 42% (95% CI, 34-51%) in the chemotherapy arm fewer adenocarcinomas among patients with high MSH2 and 49% (95% CI, 40-58%) in the control arm. The 8-year tumors (P = 0.006). overall survival rates among patients with high MSH2 tu- The 673 cases included in the MSH2 analysis differed from mors were 34% (95% CI, 25-43%) in the chemotherapy the 95 cases excluded in terms of histology (P < 0.001), type arm and 39% (95% CI, 31-49%) in the control arm. of surgery (P = 0.02), and slide quality (P = 0.01; there were Prognostic effect of MSH2 expression on overall survival. fewer squamous cell carcinomas, fewer pneumonectomies, In the control arm, high MSH2 compared with low and lower slide quality in excluded cases). MSH2 was associated with an adjusted hazard ratio for Overall survival and adjuvant chemotherapy. For the death of 0.66 (95% CI, 0.49-0.90; P = 0.01). Median over- group of patients included in the MSH2 analysis (673 pa- all survival tended to be 16 months longer in the high tients), the adjusted hazard ratio for death associated with MSH2 group (58 months) than in the low MSH2 group chemotherapy compared with observation was 0.88 [95% (42 months). confidence interval (95% CI), 0.72-1.07; P = 0.21]. The 5- In the chemotherapy arm, there was no difference in year overall survival rates were 47% (95% CI, 41%-52%) overall survival between high MSH2 and low MSH2 tu- in the chemotherapy arm, and 44% (95% CI, 39-49%) in mors (adjusted hazard ratio for death, 0.99; 95% CI, the control arm. The 8-year overall survival rates were 36% 0.74-1.32; P = 0.93). (95% CI, 31-42%) in the chemotherapy arm and 37% Effect of chemotherapy on overall survival according to (95% CI, 32-43%) in the control arm. MSH2 and ERCC1 expression. The expression levels of Effect of chemotherapy on overall survival according to MSH2 and ERCC1 were both available for 658 patients MSH2 expression. Overall, the test for interaction between (84% of the 783 patients included in IALT-Bio). In a sub- chemotherapy and MSH2 gave a P value of 0.06 in the Cox analysis of these 658 patients, the test for interaction model including slide quality and histologic type. In the between chemotherapy and the markers gave borderline low-MSH2 group, there was a trend for overall survival P values (P = 0.05 for MSH2 and P = 0.04 for ERCC1). to be longer in the chemotherapy arm than in the control In the low MSH2 group (404 patients), the adjusted arm (adjusted hazard ratio for death, 0.76; 95% CI, 0.59- hazard ratio for death associated with chemotherapy 0.97; P = 0.03; Table 2; Fig. 2A). The 5-year overall survival versus observation was 0.75 (95% CI, 0.58-0.98; P = rates among patients with low MSH2 tumors were 49% 0.03; Table 3). In the high MSH2 group (254 patients), (95% CI, 43-56%) in the chemotherapy arm and 41% there was no difference in the overall survival between (95% CI, 34-48%) in the control arm. The 8-year overall the chemotherapy arm and the control arm (adjusted survival rates among patients with low MSH2 tumors were hazard ratio for death, 1.14; 95% CI, 0.82-1.57; P = 0.44). 38% (95% CI, 32-45%) in the chemotherapy arm and Among these 658 patients, the adjusted hazard ratio for 36% (95% CI, 30-43%) in the control arm. Among pa- death associated with chemotherapy versus observation

Table 2. Overall survival according to allocated treatment and MSH2 status

Chemotherapy group Control group Hazard ratio for P (No deaths / No patients) (No deaths / No patients) death (95% CI)*

Patients with low MSH2 tumors 131 / 215 130 / 201 0.76 (0.59-0.97) P = 0.03 N = 416 Patients with high MSH2 tumors 83 / 131 75 / 126 1.12 (0.81-1.55) P = 0.48 N = 257 Hazard ratio for death (95% CI)† 0.99 (0.74-1.32) 0.66 (0.49-0.90) - - P Value P = 0.93 P = 0.01 - P = 0.06‡

*Hazard ratios are for the comparison of the chemotherapy group with the control group. †Hazard ratios are for the comparison of patients with MSH2 positive tumors with those with MSH2 negative tumors. ‡The P value is for the interaction between MSH2 expression and treatment.

1210 Clin Cancer Res; 16(4) February 15, 2010 Clinical Cancer Research

DNA Repair in Non–Small Cell Lung Cancer

Fig. 2. Kaplan-Meier estimates of the probability of overall survival according to treatment with chemotherapy. Survival rates were estimated with the use of the Kaplan-Meier method. The P values beside the curves are those of the adjusted Cox model. A, overall survival according to treatment in patients with low MSH2 tumors. The adjusted hazard ratio for death in the chemotherapy group, compared with the control group, was 0.76 (95% CI, 0.59-0.97; P = 0.03). B, overall survival according to treatment in high MSH2 tumors. The adjusted hazard ratio for death in the chemotherapy group, compared with the control group, was 1.12 (95% CI, 0.81-1.55; P = 0.48). C, overall survival according to the combined low MSH2/low ERCC1 subgroup. The adjusted hazard ratio for death in the chemotherapy group, compared with the control group, was 0.65 (95% CI, 0.47-0.91; P = 0.01). D, overall survival according to the combined high MSH2/high ERCC1 subgroup. The adjusted hazard ratio for death associated with chemotherapy was 1.32 (95% CI, 0.88-1.99; P = 0.19). E, overall survival according to the combined low MSH2/high ERCC1 subgroup. The hazard ratio for death in the chemotherapy group, compared with the control group, was 1.32 (95% CI 0.88-1.99; P = 0.71). F, overall survival according to the combined high MSH2/low ERCC1 subgroup. The hazard ratio for death in the chemotherapy group, compared with the control group, was 0.90 (95% CI, 0.52-1.55; P = 0.70).

www.aacrjournals.org Clin Cancer Res; 16(4) February 15, 2010 1211

Kamal et al.

Table 3. Overall survival according to allocated treatment and MSH2/ERCC1 subgroups

Patients with low Patients with high Total MSH2 tumors MSH2 tumors

Patients with low ERCC1 tumors Number of death / number 162 / 250 57 / 104 of patients Hazard ratio for death (95% CI) 0.65 (0.47-0.91)* 0.90 (0.52-1.55)* 0.73 (0.55-0.96)* P Value P = 0.01 P = 0.70 P = 0.03 Patients with high ERCC1 tumors Number of death / number of patients 90 / 154 99 / 150 Hazard ratio for death (95% CI) 0.92 (0.60-1.42)* 1.32 (0.88-1.99)* 1.11 (0.83-1.50)* P Value P = 0.71 P = 0.19 P = 0.49 Total Hazard ratio for death (95% CI) 0.75 (0.58-0.98)* 1.14 (0.82-1.57)* - P = 0.03 P = 0.44

*Hazard ratios are for the comparison of the chemotherapy group with the control group.

was 0.73 (95% CI, 0.55-0.96; P = 0.03) in the low ERCC1 0.65-1.28; P = 0.60). When we defined a score by counting group (354 patients; Table 3). In the high ERCC1 group thenumberofhighMSH2andERCC1markersinthe (304 patients), there was no difference in overall survival above-defined subgroups (allocating a score of 0 for no between the chemotherapy arm and the control arm (ad- marker at high levels, 1 for 1 marker at high levels, and 2 justed hazard ratio for death, 1.11; 95% CI, 0.83-1.50; P = for 2 markers at high levels), the test for trend of a decreas- 0.49). ing effect of chemotherapy with an increasing score was sig- Relationships between MSH2 and ERCC1 expression. The nificant (Ptrend = 0.01; Pheterogeneity = 0.02). proportion of high MSH2 cases was 29% (104 of 354 Predictive value of combining MSH2 and P27. Like for cases) in the low ERCC1 group and 49% (150 of 304 MSH2 and ERCC1, the effect of chemotherapy was exam- cases) in the high ERCC1 group (P < 0.001). The associa- ined in the following subgroups defined by combining tion of MSH2 with ERCC1 remained in a model adjusted MSH2 and P27: low MSH2/low P27, high MSH2/high on slide quality and histologic type (P < 0.001). P27, low MSH2/high P27, and high MSH2/low P27. The Predictive value of combining MSH2 and ERCC1. The ef- test for heterogeneity when considering these four sub- fect of chemotherapy was examined in the following subgroups was not significant (P = 0.11). The differential ef- groups defined by combining MSH2 and ERCC1: low fect of chemotherapy for patients with high MSH2 tumors MSH2/low ERCC1, high MSH2/high ERCC1, low MSH2/ compared with those with low MSH2 tumors was similar high ERCC1, and high MSH2/low ERCC1 (Fig. 2C-F). The among patients with high P27 and among patients with test for heterogeneity when considering these four sub- low P27 tumors (P = 0.91), i.e., the ratio of the hazard ra- groups was not significant (P = 0.07). The differential ef- tios associated with MSH2 among patients with low P27 fect of chemotherapy for patients with high MSH2 tumors (0.92/0.65) was equal to that among patients with high compared with those with low MSH2 lesions was similar P27 tumors (1.31/0.88). among patients with high ERCC1 and among patients In the combined low MSH2/low P27 subgroup, the haz- with low ERCC1 tumors (P = 0.94), i.e., the ratio of the ard ratio for death was reduced by chemotherapy (adjusted hazard ratios associated with MSH2 among patients with hazard ratio for death, 0.65; 95% CI, 0.46-0.93; P = 0.01). low ERCC1 (0.90/0.65) was equal to that among patients In the high MSH2/high P27 subgroup, the hazard ratio for with high ERCC1 tumors (1.32/0.92; Table 3). death associated with chemotherapy was 1.31 (95% CI, In the combined low MSH2/low ERCC1 subgroup, 0.85-2.01; P = 0.22). The adjusted hazard ratio for death overall survival was 21 months longer in the chemothera- associated with chemotherapy when one marker was ex- py arm (55 months) than in the control arm (34 months; pressed at high levels (either MSH2 or P27 but not both) adjusted hazard ratio for death, 0.65; 95% CI, 0.47-0.91; was 0.90 (95% CI, 0.67-1.21; P = 0.50). When we defined a P = 0.01; Fig. 2C). In the high MSH2/high ERCC1 sub- score by counting the number of high MSH2 and P27 mar- group, the hazard ratio for death associated with chemo- kers in the above-defined subgroups (allocating a score of 0 therapy was 1.32 (95% CI, 0.88-1.99; P = 0.19; Fig. 2D). for no marker at high levels, 1 for 1 marker at high levels, The adjusted hazard ratio for death associated with chemo- and 2 for 2 markers at high levels), the test for trend of a therapy when one marker was expressed at high levels (ei- decreasing effect of chemotherapy with an increasing score ther MSH2 or ERCC1 but not both) was 0.91 (95% CI, was significant (Ptrend = 0.01; Pheterogeneity = 0.05).

1212 Clin Cancer Res; 16(4) February 15, 2010 Clinical Cancer Research

DNA Repair in Non–Small Cell Lung Cancer

Discussion emerged as a borderline significant predictor of the short- term benefit of chemotherapy (P = 0.02). Here, we also Complete resection of tumor followed by adjuvant plat- report that MSH2 could be combined with P27 to predict inum-based chemotherapy plays a central role as a cura- the long-term benefit of chemotherapy, although the tests tive treatment for NSCLC. However, in close relationship for heterogeneity for the combinations of MSH2 and P27 with pathologic stage, only 23% to 67% of patients are (P = 0.10 and P = 0.05) could not establish that they were still alive at 5 years after the initial treatment with curative independent. Data showing that ERCC1 and P27 could intent (32). Recurrences that account for mortality occur also be combined for a better prediction of the benefit most commonly at distant extrathoracic sites. Beyond 5 of chemotherapy have been previously shown by our years, the mortality rate decreases strongly, such that sur- group (13). All together, our results are consistent with vivors may thereafter entertain a substantial hope to be the view that ERCC1, MSH2, and P27 function in a com- cured. Here, we report that the expression by tumors of mon pathway in response to cisplatin-DNA adducts, as two crucial DNA excision repair proteins, MSH2 and cell cycle arrest may be required for DNA repair to occur. ERCC1, were associated, albeit not significantly consider- The numbers for the joint analysis of MSH2 with P27 and ing the significance level that was set at 0.01, with patients that of MSH2 with ERCC1 were similar. With regard to the who benefited durably (with median follow-up of over 7 future use of predictive biomarkers alone or in combina- years) from cisplatin-based chemotherapy in the largest tion in the clinical practice, the choice between the three clinical trial of adjuvant chemotherapy for lung cancer. most promising markers that were uncovered by IALT-Bio Considering the early and possible late toxicity of plati- will be determined by the results of validation studies. num-based chemotherapy (3, 8), the evaluation of Before our report, there were only few indirect hints that MSH2 and ERCC1 expression may have a strong potential MSH2 expression could predict the benefit of platinum- to be useful to tailor individual chemotherapy after com- based chemotherapy (29, 30). The discovery of predictive plete surgical removal of NSCLC. markers requires their study in a large cohort of patients In the present study, we could show that the combina- enrolled in a randomized trial such as the IALT. Then, tion of MSH2 and ERCC1 was superior to the use of either the markers can be correlated with sufficient statistical one of the marker alone for the prediction of the long- power to survival data and to treatment. Here, the results term chemotherapy benefit. When low levels of MSH2 were strengthened by the use of two-sided statistical tests or ERCC1 were considered separately, there was a trend in a Cox model that was adjusted on cancer-specific and for chemotherapy to reduce the hazard for death by trial-specific factors, including potential confounding vari- ∼25% for each of the two markers. However, when ables, such as histologic type, slide quality, and center. The MSH2 and ERCC1 were considered jointly and when the relative protein expression levels were assessed by directly proteins were both expressed at low levels, the hazard for comparing the levels in tumor cells to those in surround- death was reduced by 35% in favor of chemotherapy. In ing pulmonary or stromal cells. The primary antibody for this subgroup (38% of patients) defined by both low MSH2 is a well-known reagent used in the clinical setting MSH2 and low ERCC1, patients who underwent chemo- to help for the pathologic diagnosis of MMR defect in co- therapy had a median survival of 55 months, whereas lon cancer. Our hypothesis was that low levels of MSH2 those who were left to observation after surgery had a me- would predict a benefit from chemotherapy because our dian survival of 34 months, which translates into a gain of previous analysis of ERCC1 using the short-term IALT sur- 21 months in favor of chemotherapy (P = 0.01). The tests vival data have shown such relationship between low le- for heterogeneity for the two markers combined were not vels of the DNA repair protein and the short-term significant, but consistent with their independence (P = benefit from chemotherapy. However, it is known that 0.07 and P = 0.02), suggesting that the loss of MSH2 MMR-defective tumors such as hereditary nonpolyposis and that of ERCC1 may act cumulatively. Because assess- colon cancer may be resistant to treatment with alkylating ments of the function of the proteins are not possible in agents (34). The MMR defect results in microsatellite insta- archival material, we can only assume that the low protein bility that inactivates many proapoptotic and DNA dam- levels were indeed associated with impaired DNA repair age response genes (35). The frequent loss of MSH2 in activity. Although experimental evidence supports that lung cancer cells is well documented (21–30) but MMR MSH2 and ERCC1 proteins cooperate in a common pro- defects and the resulting microsatellite instability appear cess of excision and repair of cisplatin-induced DNA ad- rare in lung cancer (36). The consequences of MSH2 loss ducts (19, 20), the broad substrate range of the MMR are organ specific (37), which may explain why MSH2 loss and the NER, on the other hand, may suggest that one may have opposite consequences on sensitivity to chemo- pathway may take care of lesions that are not repaired therapy in lung and colon cancers. by the other (33). The combined use of the two markers The specificity of the 8F1 monoclonal antibody has been may give a better evaluation of the repair of cisplatin- debated (38). We have shown (HeLa and A549) that 8F1 induced cross-links. The study of other DNA repair factors could discriminate cells in which ERCC1 was expressed and may still improve the accuracy of the prediction. isogenic strains in which ERCC1 expression was knocked Among the 19 protein markers that were initially down by small interfering RNA (39). Recently, Bhagwat planned for study in the first phase of IALT-Bio, P27 had et al. (40) showed that 8F1 recognized a second antigen

www.aacrjournals.org Clin Cancer Res; 16(4) February 15, 2010 1213

Kamal et al.

in skin fibroblasts from patients with XPF deficiency—an Several predictors for the benefit from adjuvant chemo- antigen that migrated as a second band closely above the therapy for lung cancer have now been published, includ- ERCC1 band in Western blot. Unfortunately, the authors ing factors involved in DNA synthesis (43) or regulation of did not bring further data that might clearly identify the the cell cycle (12) as well as mutations in KRAS and TP53 nature of the additional band, and therefore, the possi- (44), and a lung metagene model reported by Potti and bility that it is an unknown ERCC1 isoform remains colleagues (45). A comprehensive view of all these predic- open. Further, it was proposed, because HeLa cells do tors is needed together with a validation of the most inter- not have appreciable amounts of the additional band, esting ones on other completed trials. The most promising that these cells would be inappropriate for use as either markers will be incorporated in future trials to select pa- positive or negative control for validating 8F1. Therefore, tients unlikely to benefit from conventional platinum- in addition to A549 and HeLa, we investigated several based therapies and to which new agents that are active other carcinoma cell lines and found similar results in advanced stages of lung cancer can be proposed in (the data are part of an ongoing investigation), meaning the adjuvant setting (41). In our opinion, expression of that the additional band with a slightly higher molecular genes that are directly involved in the repair of cisplatin- weight in skin fibroblasts, and discussed in Bhagwat et al. DNA adducts such as MSH2 and ERCC1 as well as genes (40), actually might have limited interest in the study of involved in cell events that are required for DNA repair to most lung carcinomas. occur, including genes involved in cell cycle arrest such as As previously mentioned, when the IALT cohort was P27 and DNA damage response genes, should be tested in studied with median follow-up of 5 years, chemotherapy priority in future trials. reduced the mortality by 4% (3), and ERCC1 low levels were found significantly (P < 0.01) associated with a re- Disclosure of Potential Conflicts of Interest duction of the hazard for death of 35% in favor of chemo- No potential conflicts of interest were disclosed. therapy (11). Regarding the long-term benefit from chemotherapy, the numbers for MSH2 were almost exactly Acknowledgments those seen for ERCC1 (Table 3), suggesting that the long- term predictive abilities of MSH2 and ERCC1 are similar. We are indebted to Gorana Tomasic for the helpful discussion about the As the magnitude of benefit from chemotherapy decreases use of the FE11 antibody, Estelle Taranchon who did the immunostaining, David Jaminé (European Organization for Research and Treatment of with time, the predictive ability of the markers seems like- Cancer, Brussels) who scanned the TMA slides, and Lorna Saint Ange for wise to decrease. Possible unexplained late chemotherapy- editing. related mortality and other factors may be involved (8). The results emphasize the importance of long-term Grant Support follow-up of patients enrolled in clinical trials and trans- Comité du Val de Marne de la Ligue contre le cancer (P. Fouret), INCa lational studies (41). projet libre (P. Fouret, J.C. Soria), INCa PNES Poumon (P. Fouret, J.C. Soria), It is also noteworthy that in the control arm, patients unrestricted research grant from Eli Lilly (T. Le Chevalier), and a DUERCC whose tumors expressed high levels of the DNA excision fellowship (N.S. Kamal). The costs of publication of this article were defrayed in part by the repair proteins survived longer than those whose tumors payment of page charges. This article must therefore be hereby marked did not. A high capacity to repair DNA seems to act as a advertisement in accordance with 18 U.S.C. Section 1734 solely to barrier against the occurrence of a more aggressive cancer indicate this fact. phenotype but may not allow patients to benefit from Received 8/14/09; revised 11/10/09; accepted 12/10/09; published DNA damaging drugs such as cisplatin (42). OnlineFirst 2/9/10.

References 1. Wu Q, Christensen LA, Legerski RJ, Vasquez KM. Mismatch repair plus carboplatin compared with observation in stage IB non-small- participates in error-free processing of DNA interstrand crosslinks in cell lung cancer: CALGB 9633 with the Cancer and Leukemia Group human cells. EMBO Rep 2005;6:551–7. B, Radiation Therapy Oncology Group, and North Central Cancer 2. Martin LP, Hamilton TC, Schilder RJ. Platinum resistance: the role of Treatment Group Study Groups. J Clin Oncol 2008;26:5043–51. DNA repair pathways. Clin Cancer Res 2008;14:1291–5. 7. Pignon JP, Tribodet H, Scagliotti GV, et al. Lung adjuvant cisplatin 3. Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, evaluation: a pooled analysis by the LACE Collaborative Group. J Vansteenkiste J. Cisplatin-based adjuvant chemotherapy in pa- Clin Oncol 2008;26:3552–9. tients with completely resected non-small-cell lung cancer. N Engl 8. Le Chevalier T DA, Arriagada R, Bergman B, et al. IALT Collaborative J Med 2004;350:351–60. Group. Long-term results of the International Adjuvant Lung Cancer 4. Winton T, Livingston R, Johnson D, et al. Vinorelbine plus cisplatin Trial (IALT) evaluating adjuvant cisplatin-based chemotherapy in vs. observation in resected non-small-cell lung cancer. N Engl J Med non-small cell lung cancer (NSCLC). J Clin Oncol 2008;26:398s. 2005;352:2589–97. 9. Bergstralh DT, Sekelsky J. Interstrand crosslink repair: can XPF- 5. Douillard JY, Rosell R, De Lena M, et al. Adjuvant vinorelbine plus ERCC1 be let off the hook? Trends Genet 2008;24:70–6. cisplatin versus observation in patients with completely resected 10. Zhu XD, Niedernhofer L, Kuster B, Mann M, Hoeijmakers JH, de stage IB-IIIA non-small-cell lung cancer (Adjuvant Navelbine Interna- Lange T. ERCC1/XPF removes the 3′ overhang from uncapped tel- tional Trialist Association [ANITA]): a randomised controlled trial. omeres and represses formation of telomeric DNA-containing double Lancet Oncol 2006;7:719–27. minute chromosomes. Mol Cell 2003;12:1489–98. 6. Strauss GM, Herndon JE II, Maddaus MA, et al. Adjuvant paclitaxel 11. Olaussen KA, Dunant A, Fouret P, et al. DNA repair by ERCC1 in

1214 Clin Cancer Res; 16(4) February 15, 2010 Clinical Cancer Research

DNA Repair in Non–Small Cell Lung Cancer

non-small-cell lung cancer and cisplatin-based adjuvant chemother- cancer. A tissue microarray study. Biomed Pap Med Fac Univ Pa- apy. N Engl J Med 2006;355:983–91. lacky Olomouc Czech Repub 2006;150:255–9. 12. Filipits M, Pirker R, Dunant A, et al. Cell cycle regulators and out- 27. Kanellis G, Chatzistamou I, Koutselini H, et al. Expression of DNA come of adjuvant cisplatin-based chemotherapy in completely re- mismatch repair gene MSH2 in cytological material from lung cancer sected non-small-cell lung cancer: the International Adjuvant Lung patients. Diagn Cytopathol 2006;34:463–6. Cancer Trial Biologic Program. J Clin Oncol 2007;25:2735–40. 28. Scartozzi M, Franciosi V, Campanini N, et al. Mismatch repair system 13. Pirker R, Filipits M, Dunant A, et al. IALT-Bio: a challenging research (MMR) status correlates with response and survival in non-small cell to improve adjuvant chemotherapy of completely resected NSCLC. J lung cancer (NSCLC) patients. Lung Cancer 2006;53:103–9. Thorac Oncol 2007;2:S397–8. 29. Hsu HS, Lee IH, Hsu WH, Kao WT, Wang YC. Polymorphism in the 14. Filipits M, Haddad V, Schmid K, et al. Multidrug resistance proteins hMSH2 gene (gISV12-6T > C) is a prognostic factor in non-small cell do not predict benefit of adjuvant chemotherapy in patients with lung cancer. Lung Cancer 2007;58:123–30. completely resected non-small cell lung cancer: International Adju- 30. Cooper WA, Kohonen-Corish MR, Chan C, et al. Prognostic signifi- vant Lung Cancer Trial Biologic Program. Clin Cancer Res 2007; cance of DNA repair proteins MLH1, MSH2 and MGMT expression in 13:3892–8. non-small-cell lung cancer and precursor lesions. Histopathology 15. Brambilla E, Soria J, Haddad V, et al. Prognostic and predictive value 2008;52:613–22. of apoptosis related factors Fas, FasL and survivin in non small cell 31. Brambilla E, Lantuejoul S, Dunant A. IALT (International Adjuvant lung carcinoma patients enrolled in the IALT Trial. J Thorac Oncol Lung Cancer Trial):quality assessment and histopathological review 2007;2:S444–5. according to the WHO 2004 classification and assessment of 16. Fishel R, Lescoe MK, Rao MR, et al. The human mutator gene ho- prognostic and predictive role of pathological criteria. Lung Cancer molog MSH2 and its association with hereditary nonpolyposis colon 2005;49;suppl2:S44. cancer. Cell 1993;75:1027–38. 32. Mountain CF. Revisions in the International System for Staging Lung 17. Duckett DR, Drummond JT, Murchie AI, et al. Human MutSα recog- Cancer. Chest 1997;111:1710–7. nizes damaged DNA base pairs containing O6-methylguanine, O4- 33. Yang W. Structure and mechanism for DNA lesion recognition. Cell methylthymine, or the cisplatin-d(GpG) adduct. Proc Natl Acad Sci Res 2008;18:184–97. U S A 1996;93:6443–7. 34. Claij N, te Riele H. Microsatellite instability in human cancer: a prog- 18. Zdraveski ZZ, Mello JA, Farinelli CK, Essigmann JM, Marinus MG. nostic marker for chemotherapy? Exp Cell Res 1999;246:1–10. MutS preferentially recognizes cisplatin- over oxaliplatin-modified 35. Duval A, Hamelin R. Mutations at coding repeat sequences in mis- DNA. J Biol Chem 2002;277:1255–60. match repair-deficient human cancers: toward a new concept of tar- 19. Lan L, Hayashi T, Rabeya RM, et al. Functional and physical interac- get genes for instability. Cancer Res 2002;62:2447–54. tions between ERCC1 and MSH2 complexes for resistance to cis- 36. Forgacs E, Wren JD, Kamibayashi C, et al. Searching for microsatel- diamminedichloroplatinum(II) in mammalian cells. DNA Repair (Amst) lite mutations in coding regions in lung, breast, ovarian and colorec- 2004;3:135–43. tal cancers. Oncogene 2001;20:1005–9. 20. Zhang N, Liu X, Li L, Legerski R. Double-strand breaks induce ho- 37. Weinberg RA. The biology of cancer. Garland Science Eds 2007: mologous recombinational repair of interstrand cross-links via coop- 463–556. eration of MSH2, ERCC1-XPF, REV3, and the Fanconi anemia 38. Niedernhofer LJ, Bhagwat N, Wood RD. ERCC1 and non-small- pathway. DNA Repair (Amst) 2007;6:1670–8. cell lung cancer. N Engl J Med 2007;356:2538–40, author reply 21. Xinarianos G, Liloglou T, Prime W, et al. hMLH1 and hMSH2 expres- 40–1. sion correlates with allelic imbalance on chromosome 3p in non- 39. Olaussen KA, Fouret P, Kroemer G. ERCC1-specific immunostaining small cell lung carcinomas. Cancer Res 2000;60:4216–21. in non-small-cell lung cancer. N Engl J Med 2007;357:1559–61. 22. Brooks KR, To K, Joshi MB, et al. Measurement of chemoresistance 40. Bhagwat NR, Roginskaya VY, Acquafondata MB, Dhir R, Wood RD, markers in patients with stage III non-small cell lung cancer: a novel Niedernhofer LJ. Immunodetection of DNA repair endonuclease approach for patient selection. Ann Thorac Surg 2003;76:187–93; ERCC1-XPF in human tissue. Cancer Res 2009;69:6831–8. discussion 93. 41. Chhatwani L, Cabebe E, Wakelee HA. Adjuvant treatment of re- 23. Wang YC, Lu YP, Tseng RC, et al. Inactivation of hMLH1 and hMSH2 sected lung cancer. Proc Am Thorac Soc 2009;6:194–200. by promoter methylation in primary non-small cell lung tumors and 42. Gazdar AF. DNA repair and survival in lung cancer-the two faces of matched sputum samples. J Clin Invest 2003;111:887–95. Janus. N Engl J Med 2007;356:771–3. 24. Zhu LQ, Diao LM, Chen DJ, et al. [Relationship between hMSH2 and 43. Zheng Z, Chen T, Li X, Haura E, Sharma A, Bepler G. DNA synthesis FHIT gene expression in non-small cell lung cancer]. Ai Zheng 2003; and repair genes RRM1 and ERCC1 in lung cancer. N Engl J Med 22:571–4. 2007;356:800–8. 25. Hsu HS, Wen CK, Tang YA, et al. Promoter hypermethylation is the 44. Tsao MS, Aviel-Ronen S, Ding K, et al. Prognostic and predictive im- predominant mechanism in hMLH1 and hMSH2 deregulation and is a portance of p53 and RAS for adjuvant chemotherapy in non small- poor prognostic factor in nonsmoking lung cancer. Clin Cancer Res cell lung cancer. J Clin Oncol 2007;25:5240–7. 2005;11:5410–6. 45. Potti A, Mukherjee S, Petersen R, et al. A genomic strategy to refine 26. Skarda J, Fridman E, Plevova P, et al. Prognostic value of hMLH1 prognosis in early-stage non-small-cell lung cancer. N Engl J Med and hMSH2 immunohistochemical expression in non-small cell lung 2006;355:570–80.

www.aacrjournals.org Clin Cancer Res; 16(4) February 15, 2010 1215

MutS Homologue 2 and the Long-term Benefit of Adjuvant Chemotherapy in Lung Cancer for the International Adjuvant Lung Trial-Bio investigators, Nermine S. Kamal, Jean-Charles Soria, et al.

Clin Cancer Res 2010;16:1206-1215. Published OnlineFirst February 15, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-09-2204

Cited articles This article cites 43 articles, 13 of which you can access for free at: http://clincancerres.aacrjournals.org/content/16/4/1206.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/16/4/1206.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/16/4/1206. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.