Chondrolectin Is a Novel Diagnostic Biomarker and a Therapeutic Target for Lung Cancer

Published OnlineFirst October 20, 2011; DOI: 10.1158/1078-0432.CCR-11-0619

Clinical Cancer Imaging, Diagnosis, Prognosis Research

Ken Masuda1,4, Atsushi Takano1,2,3, Hideto Oshita1, Hirohiko Akiyama5, Eiju Tsuchiya6, Nobuoki Kohno4, Yusuke Nakamura1, and Yataro Daigo1,2,3

Abstract Purpose: This study aims to identify molecules that might be useful as diagnostic/prognostic biomarkers and as targets for the development of new molecular therapies for lung cancer. Experimental Design: We screened for genes that were highly transactivated in a large proportion of 120 lung cancers by means of a cDNA microarray representing 27,648 genes and found chondrolectin (CHODL) as a candidate. Tumor tissue microarray was applied to examine the expression of CHODL protein and its clinicopathologic signiﬁcance in archival non–small cell lung cancer (NSCLC) tissues from 295 patients. A role of CHODL in cancer cell growth and/or survival was examined by siRNA experiments. Cellular invasive effect of CHODL on mammalian cells was examined by Matrigel assays. Results: Immunohistochemical staining revealed that strong positivity of CHODL protein was associated with shorter survival of patients with NSCLC (P ¼ 0.0006), and multivariate analysis conﬁrmed it to be an independent prognostic factor. Treatment of lung cancer cells with siRNAs against CHODL suppressed growth of the cancer cells. Furthermore, induction of exogenous expression of CHODL conferred growth and invasive activity of mammalian cells. Conclusions: CHODL is likely to be a prognostic biomarker in the clinic and targeting CHODL might be a strategy for the development of anticancer drugs. Clin Cancer Res; 17(24); 7712–22. 2011 AACR.

Introduction molecule inhibitors of EGFR tyrosine kinase, such as gefitinib and erlotinib, have been investigated in clinical Lung cancer is one of the most common causes of cancer trials and/or were approved for clinical use. These agents death in the world, and non–small cell lung cancer are effective in the treatment of advanced NSCLC to a certain (NSCLC) accounts for nearly 80% of those cases (1, 2). In extent, but a proportion of patients who could receive a the last 2 decades, several newly developed cytotoxic agents survival benefit is still limited (4, 5). Hence, new therapeutic including paclitaxel, docetaxel, gemcitabine, vinorelbine, strategies, such as the development of more selective and and irinotecan have emerged to offer multiple therapeutic effective molecular targeted agents against lung cancer are choices for patients with advanced NSCLC; however, each eagerly awaited. of the new regimens can provide only modest improve- Genome-wide cDNA microarray analysis of cancers ments in survival and quality of life compared with tradi- enabled us to obtain their comprehensive gene expression tional cisplatin-based therapies (3). Recently, molecular profiles and to compare the gene expression levels with targeted agents, including anti-EGFR or anti-VEGF mono- clinicopathologic and biologic information of cancers (6). clonal antibody, cetuximab or bevacizumab, and small- This approach is also useful to identify unknown molecules involved in the carcinogenic pathway. Through the gene expression profile analysis of 120 lung cancers on a cDNA Authors' Affiliations: 1Laboratory of Molecular Medicine, Human Genome microarray consisting of 27,648 genes or expressed Center, Institute of Medical Science, The University of Tokyo, Tokyo; sequence tags (EST) coupled with purification of cancer 2Department of Medical Oncology; and 3Cancer Center, Shiga University of Medical Science, Otsu; 4Department of Molecular and Internal Medicine, cell population by laser microdissection, and subsequent Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima; comparison of the data with the expression profile of 31 5Department of Thoracic Surgery, Saitama Cancer Center, Saitama; and 6Kanagawa Cancer Center Research Institute, Yokohama, Japan normal human tissues, we identified a number of potential molecular targets for cancer diagnosis, treatment, and/or Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). choice of therapy (6–11). To verify the biologic and clinicopathologic significance of the respective gene products, we Corresponding Author: Yataro Daigo, Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, have also been carrying out high-throughput screening of Japan. Phone: 81-3-5449-5457; Fax: 81-3-5449-5406; E-mail: loss-of-function effects by means of the RNA interference [email protected] technique as well as tumor tissue microarray analysis of doi: 10.1158/1078-0432.CCR-11-0619 clinical lung cancer materials (12–41). This systematic 2011 American Association for Cancer Research. approach identified a set of molecules that seem to fall into

7712 Clin Cancer Res; 17(24) December 15, 2011

CHODL as a Novel Cancer Biomarker

mas (SCC; LU61, NCI-H520, NCI-H1703, NCI-H2170, and Translational Relevance SK-MES-1), 1 large cell carcinoma (LCC; LX1), and 4 small cell lung cancers (DMS114, DMS273, SBC-3, and SBC-5). A Because there is a significant association of CHODL human bronchial epithelial cell line (BEAS-2B) was used as a positivity with poor clinical outcome of lung cancer control (Supplementary Table S1). All cells were grown in patients, CHODL positivity in resected specimens could monolayers in appropriate media supplemented with 10% be an index that provides valuable information to phy- fetal calf serum (FCS) and were maintained at 37 Cinan sicians in applying adjuvant therapy and intensive fol- atmosphere of humidified air with 5% CO . Primary lung low-up to the patients who are likely to relapse. Because 2 cancers were obtained with informed consent along with CHODL encodes a cancer-testis antigen, it may be useful adjacent normal lung tissue samples from patients, as for screening of human leukocyte antigen–restricted described previously (7, 11). All tumors were staged on the epitope peptides for cancer vaccine that can induce basis of the pathologic tumor-node-metastasis (pTNM) clas- specific immune responses by cytotoxic T cells against sification of the Unio Internationale Contra Cancrum (Ital- CHODL-positive cancers. Because CHODL plays impor- ian; UICC; ref. 44). Formalin-fixed primary NSCLCs and tant roles in cancer proliferation and invasion, targeting adjacent normal lung tissue samples used for immunostain- CHODL function by small-molecule compounds could ing on tissue microarrays had been obtained from 295 be a new therapeutic strategy that is expected to have a patients undergoing curative surgery at Saitama Cancer powerful anticancer activity with a minimal risk of Center (Saitama, Japan). To be eligible for this study, tumor adverse effects. samples were selected from patients who fulfilled all of the following criteria: (i) patients suffered from primary NSCLC with confirmed stage (only pT1 to pT3, pN0 to pN2, and pM0), (ii) patients underwent curative surgery but did not the category of cancer-testis antigens and in which upregu- receive any preoperative treatment, (iii) among them lation is a frequent and important feature of the malignant NSCLC patients with lymph node metastasis–positive nature of human cancer. These molecules identified can be tumors (pN1 and pN2) were treated with platinum-based regarded as potential targets for the development of highly adjuvant chemotherapies after surgery, whereas patients sensitive and specific biomarkers as well as being useful in with pN0 did not receive adjuvant chemotherapies, and (iv) the development of small molecular compounds, antibo- patients whose clinical follow-up data were available. This dies, and cancer vaccines that could have a more specific and study and the use of all clinical materials mentioned were efficient anticancer effect with minimal risk of adverse approved by individual institutional ethics committees. effects (12–41). Among them, Chondrolectin (CHODL) was commonly overexpressed in the great majority of lung Semiquantitative reverse transcriptase PCR analysis cancers. Total RNA was extracted from cultured cells and clini- CHODL, a human ortholog of mouse chondrolectin cal tissues with Trizol reagent (Life Technologies, Inc.) (chodl) encodes a single-pass transmembrane protein with according to the manufacturer’s protocol. Extracted one carbohydrate recognition domain of C-type lectins RNAs and normal human-tissue polyA RNAs were treated (42). C-type lectin family is classified into 17 groups accord- with DNase I (Roche Diagnostics) and then reverse tran- ing to their domain architecture. They are divided into 2 scribed using oligo (dT) primer and SuperScript II reverse major types: transmembrane C-type lectins and soluble transcriptase (Life Technologies, Inc.). Semiquantitative ones. These lectins have a wide variety of functions at reverse transcriptase PCR (RT-PCR) experiments were cytoplasm, cytoplasmic membrane, and extracellular space. carried out with synthesized CHODL gene–specific primers Mouse chodl was mainly expressed in embryonic tissues, (50-GGAAGGAAAGGAACTACGAAATC-30 and 50-GTTAA- and its expression level was tightly controlled during embry- AAGGAGCACAGGGACATA-30) or with b-actin (ACTB)- onic development. Chodl transcripts were present at the specific primers (50-ATCAAGATCATTGCTCCTCCT-30 and stage of 7 days postconception (dpc), and elevated up to 50-CTGCGCAAGTTAGGTTTTGT-30) as an internal control. the stage of 15 dpc, followed by decrease at the stage of 17 All PCR reactions involved initial denaturation at 95C for dpc (43). To date, the biologic function of CHODL and its 2 minutes followed by 22 (for ACTB) or 30 cycles (for activation in human cancers were not reported. CHODL) of 95C for 30 seconds, 54Cto60C for 30 sec- Here, we report evidences that CHODL might play a onds, and 72C for 60 seconds on a GeneAmp PCR system significant role in lung carcinogenesis and is a potential 9700 (Applied Biosystems). diagnostic and prognostic marker for lung cancer. Immunocytochemical analysis Materials and Methods Immunocytochemical analyses were done as previously described (13), with 1 mg/mL of a goat polyclonal anti- Cell lines and clinical tissue samples CHODL antibody (catalog no. sc-54867; Santa Cruz The human lung-cancer cell lines used in this study Biotechnology) for detecting endogenous CHODL and included 5 adenocarcinomas (ADC; A549, LC319, NCI- 4 mg/mL of a rabbit polyclonal anti-calreticulin antibody H1373, PC14, and NCI-H1781), 5 squamous cell carcino- (catalog no. SPA-600; StressGen) for detecting endogenous

www.aacrjournals.org Clin Cancer Res; 17(24) December 15, 2011 7713

Masuda et al.

carleticulin in the endoplasmic reticulum (ER) as a primary antibody was specific for CHODL protein by antigen block- antibody. The cells were incubated with these primary ing assays using CHODL antigen peptides (catalog no. sc- antibodies for 1 hour at room temperature, followed by 54867P; Santa Cruz Biotechnology) that were used for incubation with Alexa 488–conjugated donkey anti-goat immunization of goats to produce polyclonal anti-human secondary antibodies (Molecular Probe) and an Alexa 594– CHODL antibodies (Supplementary Fig. S2). Three inde- conjugated donkey anti-rabbit secondary antibodies pendent investigators assessed CHODL positivity semi- (Molecular Probe). Each stained specimen was mounted quantitatively without prior knowledge of clinicopathologic with Vectashield (Vector Laboratories, Inc.) containing 40, data. The intensity of CHODL staining was evaluated using 6-diamidino-2-phenylindole and visualized with Spectral the following criteria: strong positive (scored as 2þ), brown Confocal Scanning Systems (TSC SP2 AOBS; Leica Micro- staining in more than 50% of tumor cells completely systems). On immunocytochemical analyses, we confirmed obscuring cytoplasm; weak positive (1þ), any lesser degree that the antibody was specific for CHODL protein, using of brown staining appreciable in tumor cell cytoplasm; and NSCLC cell lines that endogenously expressed CHODL as absent (scored as 0), no appreciable staining in tumor cells. well as cell lines derived from NSCLC or bronchial epithelia Cases were accepted as strongly positive if 2 or more inves- that did not express it. We also confirmed the specificity of tigators independently defined them as such. the antibody by immunostaining NSCLC cells that were transfected with siRNAs against CHODL that could suppress Statistical analysis CHODL expression or control siRNAs that could not (Sup- Statistical analyses were conducted using the StatView plementary Fig. S1). statistical program (SAS) to compare patient characteristics with responses to therapy. Associations between clinico- Northern blot analysis pathologic variables and positivity for CHODL were com- Human multiple tissue blots (23 normal tissues includ- pared by the Fisher exact tests. Tumor-specific survival and ing heart, brain, placenta, lung, liver, skeletal muscle, kid- 95% CIs were evaluated with the Kaplan–Meier method, ney, pancreas, spleen, thymus, prostate, testis, ovary, small and differences between the 2 groups were evaluated with intestine, colon, leukocyte, stomach, thyroid, spinal cord, the log-rank test. Factors associated with the prognosis were lymph node, trachea, adrenal gland, bone marrow; BD evaluated using Cox’s proportional-hazard regression mod- Biosciences) were hybridized with a 32P-labeled PCR prod- el with a step-down procedure. Only those variables with uct of CHODL cDNA. The cDNA probe of CHODL was statistically significant results in univariate analysis were prepared by RT-PCR using primers CHODL-F (50-GGTGCA- included in a multivariate analysis. The criterion for remov- TAAACACTAATGCAGTC-30) and CHODL-R (50-GTTAAA- ing a variable was the likelihood ratio statistic, which was AGGAGCACAGGGACATA-30). Prehybridization, hybrid- based on the maximum partial likelihood estimate (default ization, and washing were carried out according to the P value of 0.05 for removal from the model). supplier’s recommendations. The blots were autoradio- graphed with intensifying screens at 80C for 7 days. Western blot analysis Cells were lysed in lysis buffer [50 mmol/L Tris-HCl (pH Tissue microarray and immunohistochemistry 8.0), 150 mmol/L NaCl, 0.5% NP40, 0.5% sodium deox- Tumor tissue microarrays were constructed with 295 ycholate, and Protease Inhibitor Cocktail Set III (EMD formalin-fixed primary lung cancers, as published previ- Biosciences, Inc.). Protein samples were separated by ously (45–47). The tissue area for sampling was selected SDS-polyacrylamide gels and electroblotted onto based on visual alignment with the corresponding hema- Hybond-ECL nitrocellulose membranes (GE Healthcare toxylin and eosin–stained section on a slide. Three, 4, or 5 Bio-Sciences). Blots were incubated with a mouse mono- tissue cores (diameter 0.6 mm; height 3–4 mm) taken from clonal anti-myc antibody. Antigen–antibody complexes a donor tumor block were placed into a recipient paraffin were detected with secondary antibodies conjugated to HRP block using a tissue microarrayer (Beecher Instruments. A (GE Healthcare Bio-Sciences). Protein bands were visual- core of normal tissue was punched from each case, and 5 mm ized by enhanced chemiluminescence Western blotting sections of the resulting microarray block were used for detection reagents (GE Healthcare Bio-Sciences). immunohistochemical analysis. To investigate the presence of CHODL protein in clinical RNA interference assay samples that had been embedded in paraffin blocks, we To evaluate the biologic functions of CHODL in lung stained the sections in the following manner. Briefly, 4 mg/ cancer cells, we used siRNA duplexes against the target mL of a goat polyclonal anti-human CHODL antibody genes. The target sequences of the synthetic oligonucleo- (catalog no. sc-54867; Santa Cruz Biotechnology) was tides for RNA interference were as follows: control 1 (si- added after blocking of endogenous peroxidase and pro- LUC: siRNA against Photinus pyralis luciferase gene), 50- teins. The sections were incubated with horseradish perox- CGUACGCGGAAUACUUCGA-30; control 2 (si-SCR: siRNA idase (HRP)–labeled anti-goat immunoglobulin G as the against chloroplast Euglena gracilis gene coding for 5S and secondary antibody. Substrate-chromogen was added and 16S (rRNAs), 50-GCGCGCUUUGUAGGAUUCG-30; si-CH- the specimens were counterstained with hematoxylin. On ODL-#A, 50-AAAGUGGCAUGGAAGUAUA-30; si-CHODL- immunohistochemical analyses, we confirmed that the #B, 50-GGUAUAAUUCCCAAUCUAA-30. Lung cancer cell

7714 Clin Cancer Res; 17(24) December 15, 2011 Clinical Cancer Research

CHODL as a Novel Cancer Biomarker

lines, SBC-5 and NCI-H2170, were plated onto 10-cm tative RT-PCR experiments in 7 of 15 additional lung cancer dishes (1.0 106 per dish) and transfected with either of tissues and 10 of 15 lung cancer cell lines (Fig. 1A and B). To the siRNA oligonucleotides (100 nmol/L) with 24 mLof determine the subcellular localization of endogenous Lipofectamine 2000 (Invitrogen) according to the manu- CHODL in lung cancer cells, we conducted immunocyto- facturers’ instructions. After 7 days of incubation, cell via- chemical analysis using anti-CHODL polyclonal antibo- bility was assessed by MTT assay. The number of colonies dies; CHODL protein was localized in cytoplasm with stained with Giemsa was also counted by colony formation granular appearance in A549 and NCI-H2170 cells, which assay using colony counting software (ImageJ software 1.42, expressed endogenous CHODL, but it was not detectable in NIH). CHODL-negative LC319 and BEAS-2B cells (Fig. 1C). Fur- thermore, we confirmed that CHODL protein was mainly Cell growth assay colocalized with calreticulin that was used as an ER marker COS-7, LC319, and BEAS-2B cells which scarcely (Fig. 1D), but not colocalized with Golgi and endosome expressed endogenous CHODL were plated at densities of markers (data not shown). We confirmed that the positive 1 106 cells/10-cm dish and transfected with plasmids signal by anti-CHODL antibody obtained in lung cancer designed to express myc-tagged CHODL or mock plasmids. A549 cells was markedly reduced by suppression of CHODL Cells were selected in medium containing 0.4 mg/mL of expression by siRNA against CHODL, which proved the geneticin (Invitrogen) for 7 days, and cell viability was specificity of antibody to CHODL protein (Supplementary assessed by MTT assay (Cell Counting Kit-8 solution; Fig. S1). Dojindo Laboratories). Northern blot analysis identified a 2.8-kb CHODL transcript to be highly expressed in the testis; however, it was Matrigel invasion assay hardly detectable in other normal tissues examined (Fig. COS-7, LC319, and BEAS-2B cells transfected either with 2A). We subsequently examined expression of CHODL plasmids expressing myc-tagged CHODL or with mock protein in 5 normal tissues (heart, liver, lung, kidney, and plasmids were grown to near confluence in Dulbecco’s testis), as well as lung cancers using anti-CHODL antibody, Modified Eagle’s Medium or RPMI containing 10% FCS. and found that it was hardly detectable in the former 4 The cells were harvested by trypsinization, washed in medi- tissues, whereas positive CHODL staining was detected at um without addition of serum or protease inhibitor, and cytoplasm of testis and lung cancer tissues (Fig. 2B). We suspended in medium at a concentration of 2 105 per confirmed that the positive signal by anti-CHODL antibody milliliter. Before preparing the cell suspension, the dried obtained in lung cancer tissues was markedly diminished by layer of Matrigel matrix (Becton Dickinson Labware) was preincubation of the antibody with the CHODL antigen rehydrated with medium for 2 hours at room temperature. peptide used for immunization of goat, which also inde- Medium (0.75 mL) containing 10% FCS was added to each pendently indicated its specificity to CHODL protein (Sup- lower chamber in 24-well Matrigel invasion chambers, and plementary Fig. S2). The expression levels of CHODL pro- 0.5 mL (1 105 cells) of cell suspension was added to each tein in lung cancer were significantly higher than those in insert of the upper chamber. The plates of inserts were testis. We also evaluated CHODL staining in lung cancers incubated for 22 hours at 37C, in which condition, no and adjacent normal lung tissues and confirmed CHODL growth promoting effect by exogenous CHODL expression protein to be positively stained in the majority of NSCLC was also confirmed by MTT assay. Then the chambers were tissues, but not in their corresponding normal lungs (Fig. processed, and cells invading through the Matrigel were 2C). Furthermore, CHODL protein was hardly detectable fixed and stained by Giemsa as directed by the supplier in benign lung disease of emphysema (Supplementary (Becton Dickinson Labware). Fig. S3).

Results Association of CHODL protein expression with poor clinical outcome of NSCLC patients CHODL expression in lung cancers and normal tissues To verify the biologic and clinicopathologic significance To search for novel target molecules for the development of CHODL, we also examined the expression of CHODL of therapeutic agents and/or diagnostic biomarkers for protein by means of tissue microarrays containing lung NSCLC, we first screened genes that showed more than a cancer tissues from 295 NSCLC patients who underwent 5-fold higher level of expression in cancer cells than in surgical resection. We classified a pattern of CHODL expres- normal cells, in more than 50% of the lung cancers analyzed sion on the tissue array ranging from absent (scored as 0) to by cDNA microarray. Among 27,648 genes or ESTs weak/strong positive (scored as 1þ or 2þ; Fig. 2D). As screened, we identified that the expression of CHODL shown in Table 1, the number of NSCLC tissues scored as transcript showed 5-folds or higher in 63% of NSCLCs 2þ,1þ, and 0 were 98 (33.2%), 131 (44.4%), and 66 compared with normal lung (control). Moreover, our cDNA (22.4%), respectively. We then evaluated the association microarray database indicated that CHODL was hardly between CHODL status and clinicopathologic variables detectable in normal tissues except testis. Therefore, we among the 295 patients. The median survival time of determined to select CHODL as a good candidate for further NSCLC patients with strong CHODL staining (score 2þ) analyses. We confirmed its transactivation by semiquanti- was significantly shorter than that with absent/weak

www.aacrjournals.org Clin Cancer Res; 17(24) December 15, 2011 7715

Masuda et al.

A Clinical lung cancers Normal ADC SCC SCLC lung LC 1 LC2 LC3 LC4 LC5 LC6 LC7 LC8 LC9 LC10 LC11 LC12 LC13 LC14 LC15 CHODL

ACTB

Lung cancer cell lines B Airway ADC SCC LCC SCLC epithelia NCI-H1781 NCI-H1373 LC 319 A549 PC14 SK-MES-1 NCI-H520 NCI-H1703 NCI-H2170 LU61 LX1 DMS114 DMS273 SBC-3 SBC-5 BEAS-2B CHODL Figure 1. CHODL expression in lung cancers. A, expression of CHODL in ACTB clinical lung cancers examined by semiquantitative RT-PCR. B, expression of CHODL in lung CHODL positive CHODL negative C cancer cell lines examined by Airway epithelia semiquantitative RT-PCR. C, subcellular localization of A549 NCI-H2170 LC319 BEAS-2B endogenous CHODL protein in lung cancer cells. CHODL was stained in the cytoplasm of cells with granular appearance in CHODL-positive A549 and NCI-H2170 cells, CHODL CHODL whereas CHODL was not stained in LC319 and BEAS-2B cells that did not express endogenous CHODL. D,colocalization of endogenous CHODL with calreticulin as ER marker in A549 cells. Merge Merge

D Calreticulin CHODL Merge A549

CHODL expression (score 0 to 1þ; P ¼ 0.0006 by log-rank In multivariate analysis of prognostic factors, CHDOL pos- test; Fig. 2E). We also used univariate analysis to evalu- itivity, age, larger tumor size, and lymph node metastasis ate associations between patient prognosis and other were significant and independent prognostic factors for factors including age (65 vs. >65 years), gender (male vs. NSCLC patients. female), histologic classification (non-ACD vs. ACD), smoking status (current/former smoker vs. never smoker), Growth-promoting effect of CHODL pathologic T stage (T2 þ T3 vs. T1), pathologic N stage Using siRNA oligonucleotides for CHODL, we attempted (N1 þ N2 vs. N0), and CHODL positivity (strong positive to knock down the expression of endogenous CHODL in vs. weak positive/absent; Table 2). Among those para- lung cancer cell lines SBC-5 and NCI-H2170, which showed meters, strong CHODL positivity, elderly, male gender, high levels of endogenous CHODL expression. Two non-ACD histology, advanced pT stage, and advanced CHODL-specific siRNAs (si-CHODL-#A and si-CHODL- pN stage were significantly associated with poor prognosis. #B) significantly suppressed expression of CHODL

7716 Clin Cancer Res; 17(24) December 15, 2011 Clinical Cancer Research

CHODL as a Novel Cancer Biomarker

Normal tissues A

(Kb) H B P L L S K P S T P T O S C L S T S L T A B 9.5 7.5 4.4

2.4 1.35

0.24

Figure 2. Expression of CHODL in normal tissues and lung tumors. A, Heart Liver Lung Kidney Northern blot analysis of the CHODL transcript in 23 normal adult human tissues. B, immunohistochemical evaluation of CHODL protein using anti–CHODL antibody in lung ADC, lung SCC, lung LCC, and 5 normal tissues. C, Testis Lung ADC Lung SCC Lung LCC immunohistochemical staining of CHODL protein using anti-CHODL C CASE1 CASE2 CASE3 CASE4 antibody in 4 representative paired lung tumors and adjacent normal Lung lung tissues. D, cancer immunohistochemical evaluation of CHODL expression on tissue microarrays (100). Examples are shown of strong, weak, absent, and normal lung tissue. E, Kaplan–Meier Normal analysis of tumor-speciﬁc survival lung in patients with NSCLCs according to CHODL expression (P ¼ 0.0006; log-rank test). Number of cases at risk is shown in Supplementary D Lung adenocarcinoma Table S2. Strong positive Weak positive Absent Normal lung

E Lung cancer mortality 100 Weak CHODL positive 80 or absent (n = 197)

40 P Survival rate (%) = 0.0006 20 Strong CHODL positive (n = 98) 0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 Postoperative days

www.aacrjournals.org Clin Cancer Res; 17(24) December 15, 2011 7717

Masuda et al.

Table 1. Association between CHODL-status in NSCLC and patients' characteristics (n ¼ 295)

Total CHODL strong CHODL weak CHODL absent P strong vs. n ¼ 295 n ¼ 98 n ¼ 131 n ¼ 66 weak or absent

Gender Male 209 75 87 47 0.1372 Female 86 23 44 19 Age, y <65 130 45 55 30 0.7091 65 165 53 76 36 Histologic type ADC 176 62 79 35 0.3818a SCC 85 24 43 18 Others 34 12 9 13 Smoking status Never 78 23 41 14 0.484 Smoker 217 75 90 52 pT factor T1 120 33 53 34 0.1018 T2 þ T3 175 65 78 32 pN factor N0 193 57 88 14 0.0699 N1 þ N2 102 41 43 18

NOTE: Others, LCC plus adenosquamous cell carcinoma. aADC vs. non-ADC.

transcripts compared with control siRNAs (si-LUC and si- To further disclose a potential role of CHODL in tumor- SCR; Fig. 3A). MTT and colony formation assays revealed igenesis, we prepared plasmids designed to express either that reduction of CHODL expression by the 2 siRNAs CHODL (pcDNA3.1/myc-His vector) or mock vector and signiﬁcantly suppressed the growth of SBC-5 and NCI- transfected them into COS-7, LC319, and BEAS-2B cells H2170 cells (Fig. 3B and C; Supplementary Fig. S4). which scarcely expressed endogenous CHODL. Cells that

Table 2. Cox's proportional hazards model analysis of prognostic factors in patients with NSCLCs

Variables HR (95% Cl) Unfavorable/favorable P

Univariate analysis CHODL 1.841 (1.294–2.620) Strong/weak or absent 0.0007a Age, y 1.699 (1.188–2.431) 65/<65 0.0037a Gender 1.639 (1.092–2.463) Male/female 0.0171a Histologic type 1.558 (1.108–2.190) Non-ADC/ADC 0.0108a Smoking status 1.310 (0.872–1.968) Smoker/never smoker 0.1929 pT factor 2.985 (1.984–4.484) T2 þ T3/T1 <0.0001 pN factor 2.451 (1.745–3.448) N1 þ N2/N0 <0.0001 Multivariate analysis CHODL 1.660 (1.161–2.372) Strong/weak or absent 0.0054a Age, y 1.869 (1.303–2.680) 65/<65 0.0007a Gender 0.829 (0.530–1.298) Male/female 0.413 Histologic type 1.195 (0.821–1.739) Non-ADC/ADC 0.3519 pT factor 2.288 (1.495–3.497) T2 þ T3/T1 0.0001a pN factor 2.160 (1.522–3.058) N1 þ N2/N0 <0.0001

aP < 0.05 (Fisher exact test).

7718 Clin Cancer Res; 17(24) December 15, 2011 Clinical Cancer Research

CHODL as a Novel Cancer Biomarker

A SBC-5 NCI-H2170 Anti-CHODL siRNA Control siRNA Anti-CHODL siRNA Control siRNA si-#A si-#B si-LUC si-SCR si-#A si-#B si-LUC si-SCR CHODL CHODL

ACTB ACTB

P < 0.0001 P < 0.0001 Figure 3. Growth effect of CHODL B P < 0.0001 P < 0.0001 expression. A–C, response of SBC- 1.5 1.5 5 and NCI-H2170 cells to si-CHODL#A, si-CHODL#B, or control siRNAs (si-LUC or si-SCR). 1.0 1.0 A, the level of CHODL mRNA expression detected by 0.5 semiquantitative RT-PCR in cells 0.5 (490/630nm) (490/630nm)

transfected with either control Relative absorbance siRNAs or si-CHODLs. B,the effect Relative absorbance of siRNA against CHODL on cell 0 0 si-#A si-#B si-LUC si-SCR si-#A si-#B si-LUC si-SCR viability, evaluated by MTT assay. C, the effect of siRNA against C CHODL on colony formation, evaluated by colony formation assays. All assays were conducted thrice and in triplicate wells. D and si-#A si-#B si-LUC si-SCR E, growth promoting effects of si-#A si-#B si-LUC si-SCR CHODL. D, transient expression of CHODL in COS-7, LC319, and D COS-7 LC319 BEAS-2B BEAS-2B cells detected by CHODL Mock CHODL Mock CHODL Mock Western blot analysis. E,effect of CHODL on growth of COS-7, Anti-myc Anti-myc Anti-myc LC319, and BEAS-2B cells. At each Anti-ACTB Anti-ACTB Anti-ACTB time point, cell viability was evaluated by MTT assay. E COS-7 LC319 BEAS-2B P P < 0.0001 1.5 < 0.0001 1.2 P = 0.02478 1.0 1.0 ) ) 0.8 ) 1.0 0.8 0.6 0.6 0.4 0.5 0.4 490/630 nm 490/630 nm 490/630 nm ( ( ( 0.2 0.2 Relative absorbance Relative absorbance Relative absorbance 0 0 0 CHODL Mock CHODL Mock CHODL Mock

transiently expressed exogenous CHODL revealed signifi- proportion of patients showing good response is still cant growth promotion compared with the mock-trans- limited. Therefore, further development of molecular fected cells (Fig. 3D and E). targeting drugs for cancer is urgently awaited. We have screened diagnostic and therapeutic target molecules Enhanced cell invasion by CHODL by whole-genome gene expression profile analysis as well Because strong expression of CHODL in NSCLC tissues as subsequent tissue microarray and functional analyses was associated with poor prognosis of patients, we next (7–41). Using this approach, we have shown here that examined a possible role of CHODL in cellular invasion by CHODL is frequently overexpressed in clinical lung can- Matrigel assays using COS-7, LC319, and BEAS-2B cells. cer samples and cell lines, and that its gene product plays Transfection of CHODL cDNA into either of the cells indispensable roles in the growth and progression of lung significantly enhanced their invasive activity (Fig. 4A–C). cancer cells. This result also suggests that CHODL could contribute to CHODL appears to belong to C-type lectin–containing the more malignant phenotype of cells. protein family that is classified into 17 groups. They are involved in diverse processes including cell recognition and Discussion communication, cell–cell adhesion, and extracellular matrix (ECM)–cell interactions (48). For example, some proteogly- Several molecular targeting drugs have been developed cans that belong to soluble C-type lectin family member and proved their efficacy in cancer therapy. However, the without transmembrane domain play an important role in

www.aacrjournals.org Clin Cancer Res; 17(24) December 15, 2011 7719

Masuda et al.

A COS-7 LC319 BEAS-2B CHODL Mock CHODL Mock CHODL Mock Anti-myc Anti-myc Anti-myc

Anti-ACTB Anti-ACTB Anti-ACTB Figure 4. Enhancement of cellular invasiveness by CHODL. A, transient expression of CHODL in COS-7 LC319 BEAS-2B COS-7, LC319, and BEAS-2B cells B P < 0.0001 P < 0.0001 120 P = 0.0056 detected by Western blot analysis. 120 50 B and C, assays showing the 100 100 invasive nature of COS-7, LC319, 40 and BEAS-2B cells in Matrigel 80 80 matrix after transfection with 30 60 60 expression plasmids for human CHODL 20 . The relative number of 40 40 cells migrating through the ﬁ 20 10 20 Matrigel-coated lters (B) and

Number of invated cells Number of invated cells Number of invated cells Number Giemsa staining (C; magniﬁcation, 0 0 0 100). Assays were done thrice and CHODL Mock CHODL Mock CHODL Mock in triplicate wells. C COS-7 LC319 BEAS-2B

CHODL Mock CHODL Mock CHODL Mock

the structure and stability of the ECM and the interaction information about cancer cell proliferation and cancer between the ECM and the cell (49). Some lectins are also progression. important in embryonic development and immune To date, several prognostic biomarkers for lung cancer responses. They function as recognition molecules in impor- were reported, however, it is difficult to measure the malig- tant processes in the immune system: for example, the nant level of tumors in individual patients using only a selectines, which belong to C-type lectin with a transmem- single marker. To quantitatively predict the prognosis of brane domain, function at cell surface in leukocyte–leuko- NSCLC patients who had undergone curative surgery, we cyte and leukocyte–endothelial interactions (49). Some next attempted combined assays as a pilot study using the transmembrane C-type lectins such as calnexin play impor- positivity data of CHODL and the available data of 2 other tant roles as molecular chaperones like calreticulin in the prognostic biomarkers for NSCLC (FGFR1OP and DLX5) ER. These proteins bind to misfolded proteins and pre- which had been previously identified by our group using the vent them from being exported from the ER to the Golgi same set of lung cancer tissues used in this study (23, 25). apparatus (50). Because we confirmed that CHODL was NSCLC patients were divided into 4 groups: group 1, all 3 partly localized in the ER of cancer cells, it might have a markers were scored as strong positive in each patient; similar function in control system for some oncogenic pro- group 2, either of 2 markers was strong positive; group 3, teins in the ER. either of one marker was strong positive; and group 4, no In this study, we showed that CHODL was expressed only marker was strong positive. There is a significant correlation in testis among the 23 normal tissues examined and was between the increased number of strong positive markers highly expressed in surgically resected samples from NSCLC and poorer prognosis of NSCLC patients (Supplementary patients, indicating CHODL to be a typical cancer-testis Fig. S5). Although further screening and validation analyses antigen. Clinicopathologic evidences obtained through our that determine the best combination of promising prog- tissue microarray experiments indicated that NSCLC nostic biomarkers providing the best separation of long patients with strong CHODL-positive tumors had shorter survivors and short ones are necessary, the results show that cancer-specific survival periods than those with weak/ this type of diagnostic approach combined with more absent expression tumors. Functional assays by knockdown quantitative scoring system of immunohistochemical stain- of CHODL expression with siRNA or exogenous expression ing may be a useful and practical index in the clinic for of CHODL revealed that CHODL was important for cellular application of adjuvant therapy to the patients who are growth and invasion. Although the precise molecular mech- likely to have poor prognosis after surgery. anism underlying our observations is unknown, the com- In summary, we have shown that overexpressed CHODL bined results strongly suggest that CHODL is likely to is likely to be one of essential contributors to a growth- function as an oncogenic protein in lung carcinogenesis promoting pathway and to aggressive features of lung and tumor progression. Elucidation of the mechanism cancers. CHODL is a possible prognostic biomarker in the implied by these observations should reveal important new clinic.

7720 Clin Cancer Res; 17(24) December 15, 2011 Clinical Cancer Research

CHODL as a Novel Cancer Biomarker

Disclosure of Potential Conﬂicts of Interest Japan Society for the Promotion of Science. Y. Daigo is a member of Shiga Cancer Treatment Project (Shiga Prefecture, Japan). The costs of publication of this article were defrayed in part by the No potential conﬂicts of interest were disclosed. payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Grant Support this fact.

This work was supported, in part, by Grant-in-Aid for Scientiﬁc Research Received March 8, 2011; revised September 9, 2011; accepted October 8, (B) and Grant-in-Aid for Scientiﬁc Research on Innovative Areas from The 2011; published OnlineFirst October 20, 2011.

References 1. Ahmedin J, Rebecca S, Elizabeth W, Yongping H, Jiaquan X, Michael 19. Kato T, Hayama S, Yamabuki Y, Ishikawa N, Miyamoto M, Ito T, et al. JT. Cancer Statistics, 2009. CA Cancer J Clin 2009;59:225–49. Increased expression of IGF-II mRNA-binding protein 1 is associated 2. Miki D, Kubo M, Takahashi A, Yoon KA, Kim J, Lee GK, et al. Variation in with the tumor progression in patients with lung cancer. Clin Cancer TP63 is associated with lung adenocarcinoma susceptibility in Jap- Res 2007;13:434–42. anese and Korean populations. Nat Genet 2010;42:893–6. 20. Suzuki C, Takahashi K, Hayama S, Ishikawa N, Kato T, Ito T, et al. 3. Schiller JH, Harrington D, Belani CP, Langer C, Sandler A, Krook J, Identification of Myc-associated protein with JmjC domain as a novel et al. Eastern Cooperative Oncology Group. Comparison of four therapeutic target oncogene for lung cancer. Mol Cancer Ther chemotherapy regimens for advanced non-small-cell lung cancer. N 2007;6:542–51. Engl J Med 2002;346:92–8. 21. Hayama S, Daigo Y, Yamabuki T, Hirata D, Kato T, Miyamoto M, et al. 4. Dowell J, Minna JD, Kirkpatrick P. Erlotinib hydrochloride. Nat Rev Phosphorylation and activation of cell division cycle associated 8 by Drug Discov 2005;4:13–4. aurora kinase B plays a significant role in human lung carcinogenesis. 5. Pal SK, Pegram M. Epidermal growth factor receptor and signal Cancer Res 2007;67:4113–22. transduction: potential targets for anti-cancer therapy. Anticancer 22. Taniwaki M, Takano A, Ishikawa N, Yasui W, Inai K, Nishimura H, et al. Drugs 2005;16:483–94. Activation of KIF4A as a prognostic biomarker and therapeutic target 6. Daigo Y, Nakamura Y. From cancer genomics to thoracic oncology: for lung cancer. Clin Cancer Res 2007;13:6624–31. discovery of new biomarkers and therapeutic targets for lung and 23. Mano Y, Takahashi K, Ishikawa N, Takano A, Yasui W, Inai K, et al. esophageal carcinoma. Gen Thorac Cardiovasc Surg 2008;56:43–53. Fibroblast growth factor receptor 1 oncogene partner as a novel 7. Kikuchi T, Daigo Y, Katagiri T, Tsunoda T, Okada K, Kakiuchi S, et al. prognostic biomarker and therapeutic target for lung cancer. Cancer Expression profiles of non-small cell lung cancers on cDNA micro- Sci 2007;98:1902–13. arrays: identification of genes for prediction of lymph-node metastasis 24. Kato T, Sato N, Hayama S, Yamabuki T, Ito T, Miyamoto M, et al. and sensitivity to anti-cancer drugs. Oncogene 2003;22:2192–205. Activation of holliday junction recognizing protein involved in the 8. Kakiuchi S, Daigo Y, Tsunoda T, Yano S, Sone S, Nakamura Y. chromosomal stability and immortality of cancer cells. Cancer Res Genome-wide analysis of organ-preferential metastasis of human 2007;67:8544–53. small cell lung cancer in mice. Mol Cancer Res 2003;1:485–99. 25. Kato T, Sato N, Takano A, Miyamoto M, Nishimura H, Tsuchiya E, et al. 9. Kakiuchi S, Daigo Y, Ishikawa N, Furukawa C, Tsunoda T, Yano S, et al. Activation of placenta specific transcription factor distal-less homeo- Prediction of sensitivity of advanced non-small cell lung cancers to box 5 predicts clinical outcome in primary lung cancer patients. Clin gefitinib (Iressa, ZD1839). Hum Mol Genet 2004;13:3029–43. Cancer Res 2008;14:2363–70. 10. Kikuchi T, Daigo Y, Ishikawa N, Katagiri T, Tsunoda T, Yoshida S, et al. 26. Dunleavy EM, Roche D, Tagami H, Lacoste N, Ray-Gallet D, Nakamura Expression profiles of metastatic brain tumor from lung adenocarci- Y, et al. HJURP is a cell-cycle-dependent maintenance and deposition nomas on cDNA microarray. Int J Oncol 2006;28:799–805. factor of cenp-a at centromeres. Cell 2009;137:485–97. 11. Taniwaki M, Daigo Y, Ishikawa N, Takano A, Tsunoda T, Yasui W, et al. 27. Hirata D, Yamabuki T, Miki D, Ito T, Tsuchiya E, Fujita M, et al. Gene expression profiles of small-cell lung cancers: molecular signa- Involvement of epithelial cell transforming sequence 2 (ect2) oncoanti- tures of lung cancer. Int J Oncol 2006;29:567–75. gen in lung and esophageal cancer progression. Clin Cancer Res 12. Suzuki C, Daigo Y, Kikuchi T, Katagiri T, Nakamura Y. Identification of 2009;15:256–66. COX17 as a therapeutic target for non-small cell lung cancer. Cancer 28. Sato N, Koinuma J, Fujita M, Hosokawa M, Ito T, Tsuchiya E, et al. Res 2003;63:7038–41. Activation of WD repeat and high-mobility group box DNA binding 13. Kato T, Daigo Y, Hayama S, Ishikawa N, Yamabuki T, Ito T, et al. A novel protein 1 in pulmonary and esophageal carcinogenesis. Clin Cancer human tRNA-dihydrouridine synthase involved in pulmonary carcino- Res 2010;16:226–39. genesis. Cancer Res 2005;65:5638–46. 29. Sato N, Koinuma J, Ito T, Tsuchiya E, Kondo S, Nakamura Y, et al. 14. Furukawa C, Daigo Y, Ishikawa N, Kato T, Ito T, Tsuchiya E, et al. Activation of an oncogenic TBC1D7 (TBC1 domain family, member 7) Plakophilin 3 oncogene as prognostic marker and therapeutic target protein in pulmonary carcinogenesis. Genes Chromosomes Cancer for lung cancer. Cancer Res 2005;65:7102–10. 2010;49:353–67. 15. Suzuki C, Daigo Y, Ishikawa N, Kato T, Hayama S, Ito T, et al. ANLN 30. Nguyen MH, Koinuma J, Ueda K, Ito T, Tsuchiya E, Nakamura Y, et al. plays a critical role in human lung carcinogenesis through the activa- Phosphorylation and activation of cell division cycle associated 5 by tion of RHOA and by involvement in the phosphoinositide 3-kinase/ mitogen-activated protein kinase play a crucial role in human lung AKT pathway. Cancer Res 2005;65:11314–25. carcinogenesis. Cancer Res 2010;70:5337–47. 16. Ishikawa N, Daigo Y, Takano A, Taniwaki M, Kato T, Tanaka S, et al. 31. Ishikawa N, Daigo Y, Yasui W, Inai K, Nishimura H, Tsuchiya E, et al. Characterization of SEZ6L2 cell-surface protein as a novel prognostic ADAM8 as a novel serological and histochemical marker for lung marker for lung cancer. Cancer Sci 2006;97:737–45. cancer. Clin Cancer Res 2004;10:8363–70. 17. Takahashi K, Furukawa C, Takano A, Ishikawa N, Kato T, Hayama S, 32. Ishikawa N, Daigo Y, Takano A, Taniwaki M, Kato T, Hayama S, et al. et al. The neuromedin u-growth hormone secretagogue receptor 1b/ Increases of amphiregulin and transforming growth factor-alpha in neurotensin receptor 1 oncogenic signaling pathway as a therapeutic serum as predictors of poor response to gefitinib among patients target for lung cancer. Cancer Res 2006;66:9408–19. with advanced non-small cell lung cancers. Cancer Res 2005;65: 18. Hayama S, Daigo Y, Kato T, Ishikawa N, Yamabuki T, Miyamoto M, 9176–84. et al. Activation of CDCA1-KNTC2, members of centromere protein 33. Yamabuki T, Takano A, Hayama S, Ishikawa N, Kato T, Miyamoto M, complex, involved in pulmonary carcinogenesis. Cancer Res et al. Dickkopf-1 as a novel serologic and prognostic biomarker for lung 2006;66:10339–48. and esophageal carcinomas. Cancer Res 2007;67:2517–25.

www.aacrjournals.org Clin Cancer Res; 17(24) December 15, 2011 7721

Masuda et al.

34. Ishikawa N, Takano A, Yasui W, Inai K, Nishimura H, Ito H, et al. Cancer- toxic T lymphocytes reactive to tumor cells. Cancer Sci 2011;102:697– testis antigen lymphocyte antigen 6 complex locus K is a serologic 705. biomarker and a therapeutic target for lung and esophageal carcino- 42. Weng L, Smits P, Wauters J, Merregaert J. Molecular cloning and mas. Cancer Res 2007;67:11601–11. characterization of human chondrolectin, a novel type I transmem- 35. Takano A, Ishikawa N, Nishino R, Masuda K, Yasui W, Inai K, et al. brane protein homologous to C-type lectins. Genomics 2002;80:62–7. Identification of nectin-4 oncoprotein as a diagnostic and therapeutic 43. Weng L, Hubner€ R, Claessens A, Smits P, Wauters J, Tylzanowski P, target for lung cancer. Cancer Res 2009;69:6694–703. et al. Isolation and characterization of a novel C-type lectin, Chon- 36. Sato N, Yamabuki T, Takano A, Koinuma J, Aragaki M, Masuda K, et al. drolectin (Chodl), which is predominantly expressed in muscle cells. Wnt inhibitor Dickkopf-1 as a target for passive cancer immunother- Gene 2003;308:21–9. apy. Cancer Res 2010;70:5326–36. 44. Sobin LH, Witterkind C, editors. TNM classification of malignant 37. Suda T, Tsunoda T, Daigo Y, Nakamura Y, Tahara H. Identification of tumours, 6th ed. New York: Whiley-Liss; 2002. human leukocyte antigen-A24-restricted epitope peptides derived 45. Chin SF, Daigo Y, Huang HE, Iyer NG, Callagy G, Kranjac T, et al. A from gene products upregulated in lung and esophageal cancers as simple and reliable pretreatment protocol facilitates fluorescent in situ novel targets for immunotherapy. Cancer Sci 2007;98:1803–8. hybridization on tissue microarrays of paraffin wax embedded tumour 38. Mizukami Y, Kono K, Daigo Y, Takano A, Tsunoda T, Kawaguchi Y, samples. Mol Pathol 2003;56:275–9. et al. Detection of novel cancer-testis antigen-specific T-cell 46. Callagy G, Cattaneo E, Daigo Y, Happerfield L, Bobrow LG, Pharoah responses in TIL, regional lymph nodes, and PBL in patients with PD, et al. Molecular classification of breast carcinomas using tissue esophageal squamous cell carcinoma. Cancer Sci 2008;99:1448–54. microarrays. Diagn Mol Pathol 2003;12:27–34. 39. Harao M, Hirata S, Irie A, Senju S, Nakatsura T, Komori H, et al. HLA- 47. Callagy G, Pharoah P, Chin SF, Sangan T, Daigo Y, Jackson L, et al. A2-restricted CTL epitopes of a novel lung cancer-associated cancer Identification and validation of prognostic markers in breast cancer testis antigen, cell division cycle associated 1, can induce tumor- with the complementary use of array-CGH and tissue microarrays. J reactive CTL. Int J Cancer 2008;123:2616–25. Pathol 2005;205:388–96. 40. Kono K, Mizukami Y, Daigo Y, Takano A, Masuda K, Yoshida K, et al. 48. Zelensky AN, Gready JE. The C-type lectin-like domain superfamily. Vaccination with multiple peptides derived from novel cancer-testis FEBS Lett 2005;272:6179–217. antigens can induce specific T-cell responses and clinical responses in 49. Drickamer K, Taylor ME. Biology of animal lectins. Annu Rev Cell Biol advanced esophageal cancer. Cancer Sci 2009;100:1502–9. 1993;9:237–64. 41. Tomita Y, Imai K, Senju S, Irie A, Inoue M, Hayashida Y, et al. A novel 50. Ellgaard L, Helenius A. Quality control in the endoplasmic reticulum. tumor-associated antigen, cell division cycle 45-like can induce cyto- Nat Rev 2003;4:181–91.

7722 Clin Cancer Res; 17(24) December 15, 2011 Clinical Cancer Research

Chondrolectin Is a Novel Diagnostic Biomarker and a Therapeutic Target for Lung Cancer

Ken Masuda, Atsushi Takano, Hideto Oshita, et al.

Clin Cancer Res 2011;17:7712-7722. Published OnlineFirst October 20, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-11-0619

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2011/10/20/1078-0432.CCR-11-0619.DC1

Cited articles This article cites 49 articles, 23 of which you can access for free at: http://clincancerres.aacrjournals.org/content/17/24/7712.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/17/24/7712.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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/17/24/7712. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.