Wnt Inhibitor Dickkopf-1 As a Target for Passive Cancer Immunotherapy

Published OnlineFirst June 15, 2010; DOI: 10.1158/0008-5472.CAN-09-3879

Molecular and Cellular Pathobiology Cancer Research Wnt Inhibitor Dickkopf-1 as a Target for Passive Cancer Immunotherapy

Nagato Sato1,2, Takumi Yamabuki1, Atsushi Takano1,6, Junkichi Koinuma1, Masato Aragaki1, Ken Masuda1, Nobuhisa Ishikawa1, Nobuoki Kohno3, Hiroyuki Ito4, Masaki Miyamoto2, Haruhiko Nakayama4, Yohei Miyagi5, Eiju Tsuchiya5, Satoshi Kondo2, Yusuke Nakamura1, and Yataro Daigo1,6

Abstract Dickkopf-1 (DKK1) is an inhibitor of Wnt/β-catenin signaling that is overexpressed in most lung and esophageal cancers. Here, we show its utility as a serum biomarker for a wide range of human cancers, and we offer evidence favoring the potential application of anti-DKK1 antibodies for cancer treatment. Using an original ELISA system, high levels of DKK1 protein were found in serologic samples from 906 patients with cancers of the pancreas, stomach, liver, bile duct, breast, and cervix, which also showed elevated expression levels of DKK1. Additionally, anti-DKK1 antibody inhibited the invasive activity and the growth of cancer cells in vitro and suppressed the growth of engrafted tumors in vivo. Tumor tissues treated with anti-DKK1 displayed significant fibrotic changes and a decrease in viable cancer cells without apparent toxicity in mice. Our findings suggest DKK1 as a serum biomarker for screening against a variety of cancers, and anti-DKK1 antibodies as potential theranostic tools for diagnosis and treatment of cancer. Cancer Res; 70(13); 5326–36. ©2010 AACR.

Introduction growth factor (VEGF) used for the treatment of metastatic colorectal and lung cancers in combination with chemother- The concept of specific molecular targeting therapy has been apy (4). The results were promising when compared with applied to the development of innovative cancer-treatment responses of advanced cancers to conventional cytotoxic strategies. At present, two main approaches are available in agents, but the proportion of patients showing good response clinical practice: therapeutic monoclonal antibodies (mAb) is still limited, and in some cases, the medication caused se- and small-molecule agents (1). There is an increasing interest rious adverse effects including life-threatening ones. Hence, in the use of antibody-based immunotherapy for the treat- the development of new therapeutic antibodies targeting ment of malignant diseases, and some dramatic clinical transmembrane and/or secreted proteins, which show a responses have enhanced the activity in this field (1). For cancer-specific expression pattern and have an oncogenic example, rituximab (Rituxan) is the chimeric anti-CD20 anti- function, is eagerly awaited. body that revolutionized lymphoma treatment (2). Trastuzu- To develop new diagnostics and therapeutics improving mab (Herceptin) is the humanized Ab against the human cancer treatment, we have been screening the candidate tar- epidermal growth factor receptor (HER)/ERBB2 for the treat- get molecules by the following strategy: (a) identifying genes ment of patients with metastatic breast cancer in which HER/ upregulated in lung and esophageal cancers by genome-wide ERBB2 gene was highly overexpressed (3). Bevacizumab cDNA microarray system (5–11), (b) verifying the candidate (Avastin) is the humanized Ab against vascular endothelial genes for the absence of or very low level of expression in normal tissues (5), (c) validating biological significance of their overexpression by means of tissue microarray containing Authors' Affiliations: 1Laboratory of Molecular Medicine, Human hundreds of archived lung cancer samples and functional as- Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan; 2Department of Surgical Oncology, Hokkaido University says for examining their activity in cell growth and/or invasive Graduate School of Medicine, Sapporo, Japan; 3Department of activity (12–33), (d) evaluating them for usefulness as a serum Molecular and Internal Medicine, Graduate School of Biomedical diagnostic/prognostic biomarker for lung cancer by ELISA Sciences, Hiroshima University, Hiroshima, Japan; Divisions of – 4Thoracic Surgery and 5Molecular Pathology and Genetics, Kanagawa (29 33), if they are tumor-specific transmembrane or secretory Cancer Center, Yokohama, Japan; and 6Department of Medical proteins, and (e) screening the epitopes recognized by human Oncology, Shiga University of Medical Science, Otsu, Japan histocompatibility leukocyte HLA-*A0201/*A2402-restricted Note: Supplementary data for this article are available at Cancer Research CTL, if they are categorized into cancer testis antigens and Online (http://cancerres.aacrjournals.org/). are indispensable for cancer cell growth/survival (34–36). Corresponding Author: Yataro Daigo, Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Using this approach, we had identified Dickkopf-1 (DKK1) Japan. Phone: 81-3-5449-5457; Fax: 81-3-5449-5406; E-mail: ydaigo@ as a novel serologic and histochemical biomarker as well as ims.u-tokyo.ac.jp. a possible therapeutic target for lung and esophageal cancers doi: 10.1158/0008-5472.CAN-09-3879 (31). Our data indicated that DKK1 expression was associated ©2010 American Association for Cancer Research. with poor prognosis for patients with non–small cell lung

5326 Cancer Res; 70(13) July 1, 2010

DKK1 as an Immunotherapeutic Target

cancer (NSCLC) or esophageal squamous cell carcinoma tained at 37°C in humidified air with 5% CO2. Surgically (ESCC), and that serum DKK1 levels were significantly higher resected primary cancers from six pancreatic cancer pa- in lung and esophageal cancer patients than in healthy con- tients, five gastric cancer patients, seven HCC patients, trols; the proportions of the serum DKK1–positive cases was six bile duct cancers, six breast cancer patients, and five 74.1% for lung adenocarcinoma (83 of 112), 63.2% for lung cervical cancer patients were obtained with informed con- squamous cell carcinoma (43 of 68), 69.4% for small cell lung sent at Kanagawa Cancer Center (Yokohama, Japan). Pri- cancer (59 of 85), and 63.0% for ESCC (51 of 81; ref. 31). In mary bile duct cancers were obtained with informed addition, we identified that the exogenous expression of consent from six patients who had undergone surgery at DKK1 increased the invasive activity of mammalian cells. Hokkaido University Hospital (Sapporo, Japan). Primary The evidence further prompted us to focus on DKK1 as a lung cancers in the same set of 18 lung cancer cases whose potential target for therapeutic antibodies as well as a serum serum had been collected before surgery (nine patients biomarker for a wide range of human cancers. with DKK1-positive tumors and nine with DKK1-negative DKK1 encodes a secreted protein that plays a crucial role tumors) were obtained at Hiroshima University Hospital in head formation in vertebrate development (37), and is (Hiroshima, Japan) and Kanagawa Cancer Center. This known as a negative regulator of the Wnt-signaling path- study and the use of all clinical materials used were way in colon cancer cell lines (38). During specific period approved by individual institutional Ethical Committees. of embryogenesis, DKK1 is secreted and binds to the LRP5/6 coreceptor (39), resulting in the blocking of the Serum samples interaction with secreted Wnt protein, inducing degrada- Serum samples were obtained with informed consent from tion of β-catenin, and then reducing the expression of 207 healthy control individuals (168 males and 39 females; genes regulated by T-cell factor. This mechanism of DKK1 median age of 50.3 y with a range of 31–61 y) and from action is important in limb and head development (37). 179 pancreatic cancer patients (114 males and 65 females; Inhibition of the Wnt pathway by secreted DKK1 is also median age of 66 y with a range of 30–83 y), 101 gastric can- known to initiate cardiogenesis in vertebrate embryos (40). cer patients (71 males and 30 females; median age of 62 y On the other hand, the overexpression of DKK1 was de- with a range of 29–82 y), 168 HCC patients (146 males and scribed in multiple myeloma, hepatoblastoma, Wilms' 22 females; median age of 70 y with a range of 32–84 y), 107 tumor, prostate cancer, kidney cancer, and breast cancer bile duct cancer patients (75 males and 32 females; median as well as lung and esophageal cancers (31, 41–44). Serum age of 67 y with a range of 45–75 y), 169 breast cancer concentrations of DKK1 protein were reported to be in- patients (169 females; median age of 51 y with a range of creased in patients with multiple myeloma or osteosarcoma 27–72 y), and 182 cervical cancer patients (182 females; me- (41, 45). In spite of recent accumulating reports that indi- dian age of 48 y with a range of 31–77 y) who were registered cate the oncogenic potential of DKK1 activation and its in the Japanese Project for Personalized Medicine (BioBank usefulness as a cancer biomarker, there was no report indi- Japan). These serum samples from a total of 906 cancer pa- cating the direct antitumor effect of anti-DKK1 antibodies. tients were selected for the study on the basis of the follow- We here show evidence suggesting DKK1 as a diagnostic ing criteria: (a) patients were newly diagnosed and (b) their biomarker for a wide variety of cancers and results implicat- tumors were pathologically diagnosed as cancers (stages I–IV). ing the therapeutic potential of anti-DKK1 antibody to neu- Serum was obtained at the time of diagnosis and stored at tralize the activity of DKK1 function for cancer cell invasion −150°C. The sample size of each cancer group (pancreas, and growth. stomach, liver, bile duct, breast, and cervix) and a healthy control group provided 90% statistical power (α =0.05; Materials and Methods β = 0.1) to detect group differences in mean levels of serum DKK1. Two-tailed P value of <0.05 was considered statisti- Cell lines and tissue samples cally significant. The human cancer cell lines used in this study were as fol- lows: four NSCLC cell lines (A549, LC319, NCI-H2170, and Semiquantitative reverse transcription-PCR PC-14), 13 pancreatic cancer cell lines (Capan-1, Capan-2, A total of 3 μg aliquot of mRNA from each sample was re- HPAF-II, KLM-1, KP-1N, Miapaca-2, Panc02.03, Panc08.13, versely transcribed to single-stranded cDNAs using random PK-1, PK-59, PK-9, PL45, and SUIT-2), four gastric cancer cell primer (Roche Diagnostics) and SuperScript II (Invitrogen). lines (MKN1, MKN45, MKN7, and MKN74), seven hepato- Semiquantitative reverse transcription-PCR (RT-PCR) experi- cellular carcinoma (HCC) cell lines (HepG2, HUH-6, HUH-7, ments were carried out as previously described (31), with the SNU-398, SNU-423, SNU-449, and SNU-475), four bile duct following sets of synthesized primers specific for DKK1 or cancer cell lines (HuCCT1, TFK-1, RBE, and SSP-25), 14 breast with β-actin (ACTB)–specific primers as an internal control: cancer cell lines (BT-20, BT-474, BT-549, HCC1143, HCC1500, DKK1,5′-TAGAGTCTAGAACGCAAGGATCTC-3′ and HCC1937, MCF-7, MDA-MB-157, MDA-MB-231, MDA-MB- 5′-CAAAAACTATCACAGCCTAAAGGG-3′,andACTB, 453, MDA-MB-435S, SK-BR-3, T47D, and ZR-75-1), and two 5′-GAGGTGATAGCATTGCTTTCG-3′ and 5′-CAAGTCAGTG- cervical cancer cell lines (C33A and HeLa; Supplementary TACAGGTAAGC-3′. PCRs were optimized for the number of Table S1). All cells were grown in monolayer in appro- cycles to ensure product intensity to be within the linear priate media supplemented with 10% FCS and were main- phase of amplification.

www.aacrjournals.org Cancer Res; 70(13) July 1, 2010 5327

Sato et al.

Immunohistochemistry Matrigel were fixed and stained by Giemsa as directed by To investigate the expression status of the DKK1 protein the supplier (Becton Dickinson Labware). in clinical cancer tissues, we stained the tissue sections using a rabbit polyclonal antibody specific to human DKK1 (Santa Antibody-based neutralizing assay 3 Cruz) raised against the NH2-terminal portion (amino acids Cells were seeded onto 24-well plates (4.0 × 10 cells/well) 1–120) of DKK1, as previously described (31). The specificity for MTT assay (Cell Counting Kit-8; Dojindo) and Matrigel of anti-DKK1 antibody was confirmed by Western blot anal- chambers (5 × 104 cells/well) for cell invasion assay. In the ysis using lysates of DKK1-expressing NSCLC and ESCC cells culture medium containing 0.1% fetal bovine serum, 50 or (31), and those that were transfected with small interfering 100 nmol/L of an affinity-purified rabbit polyclonal anti- RNA duplexes against DKK1 or control small interfering human DKK1 antibody (Santa Cruz) was added to analyze RNAs (Supplementary Fig. S1A and B). its neutralizing activity against DKK1. Normal rabbit IgG (Santa Cruz) was used as a negative control. To remove ELISA the complement, these antibodies were purified by overnight Serum levels of DKK1 were measured by an ELISA sys- dialysis at 4°C using autoclaved PBS in Seamless Cellulose tem that had been originally constructed (31). First, a rabbit Tubing (Viskase Companies). Cell viability was determined polyclonal antibody specific to DKK1 (Santa Cruz) was by MTT assay after incubation for 10 days. Cell invasion ac- added to a 96-well microplate (Nunc) as a capture antibody tivity after a 22-hour incubation was determined by Matrigel and incubated for 2 hours at room temperature. After wash- invasion assay as described (31). In this condition, the anti- ing away any unbound antibody, 5% bovine serum albumin DKK1 antibody was confirmed by MTT assay to have no was added to the wells and incubated for 16 hours at 4°C suppressive effect on cell viability (data not shown). Termi- for blocking. After a wash, 3-fold diluted sera were added to nal deoxynucleotidyl transferase–mediated dUTP nick end the wells and incubated for 2 hours at room temperature. labeling (TUNEL) assay (In situ Cell Death detection kit, After washing away any unbound substances, a biotinylated Fluorescein, Roche Diagnostics), double staining with both polyclonal antibody specific for DKK1 using Biotin Labeling Calcein-AM and propidium iodide (Cellstain-Double Staining kit-NH2 (DOJINDO) was added to the wells as a detection kit, Dojindo), and immunofluorescent analysis using anti- antibody and incubated for 2 hours at room temperature. cleaved caspase-3 (Cell Signaling Technology) were per- After a wash to remove any unbound antibody enzyme re- formed at 7 days after the antibody treatment according agent, horseradish peroxidase-streptavisin was added to the to the supplier's protocol. wells and incubated for 20 minutes. After a wash, a sub- strate solution (R&D Systems, Inc.) was added to the wells Mice model andallowedtoreactfor30minutes.Thereactionwas The animal experiments were conducted according to the stopped by adding 100 μL of 2 N sulfuric acid. Color inten- institutional and national guidelines for the care and use of sity was determined by a photometer at a wavelength of laboratory animals, and approved by the institutional ani- 450 nm, with a reference wavelength of 570 nm. The cutoff mal use committee. DKK1-positive A549 or DKK1-negative level in this assay was set to provide optimal diagnostic NCI-H2170 cells (6.0 × 105) were s.c. implanted into the right accuracy and likelihood ratios (minimal false-negative and shoulder of 6-week-old male BALB/c nude mice (nu/nu). The false-positive results) for DKK1 (i.e., 14.7 U/mL), using re- six mice with tumor (50 mm3 volume on average) were ran- ceiver operating characteristic curves drawn with the data domized into two groups and i.p. administered with 100 μg/ of 346 lung and esophageal cancer patients, and 207 healthy 500 μL of a purified rabbit polyclonal anti-human DKK1 an- controls as previously described (31). tibody (Santa Cruz) or 100 μg/500 μL of normal rabbit IgG (control; Santa Cruz) at days 1, 4, 8, 11, 15, 18, 22, 25, 29, Matrigel invasion assay and 32 (a total of 10 injections). Tumor volume was mea- NIH3T3 and PC-14 cells transfected either with p3XFLAG- sured daily by using a caliper and applying the data to the tagged (COOH terminal) plasmids expressing DKK1 or with formula [volume = 0.52 × (width)2 × (length)] to calculate mock plasmids as well as lung cancer A549, NCI-H2170, and the volume of a spheroid. PC-14 cells were grown to near confluence in culture medium containing 10% FCS. The cells were harvested by Results trypsinization, washed in DMEM without addition of serum or proteinase inhibitor, and suspended in DMEM at concen- DKK1 expression in cancers originated from tration of 1 × 105 cells/mL. Before preparing the cell suspen- various tissues sion, the dried layer of Matrigel matrix (Becton Dickinson To examine a possible application of serum DKK1 levels as Labware) was rehydrated with DMEM for 2 hours at room a diagnostic biomarker for a wide range of human cancers, temperature. DMEM (0.75 mL) containing 10% FCS was we first investigated the expression level of DKK1 transcript added to each lower chamber in 24-well Matrigel invasion in cancers derived from the pancreas, stomach, liver, bile chambers, and 0.5 mL (5 × 104 cells) of cell suspension was duct, mammary gland, and uterus because our gene expres- added to each insert of the upper chamber. The plates of in- sion profile analysis implied the transactivation of DKK1 in serts were incubated for 22 hours at 37°C. After incubation, these tumors. By means of semiquantitative RT-PCR experi- the chambers were processed; cells invading through the ments, we confirmed its elevated expression in five of six

5328 Cancer Res; 70(13) July 1, 2010 Cancer Research

DKK1 as an Immunotherapeutic Target

clinical pancreatic cancers, in three of five gastric cancers, in potential of anti-DKK1 antibody, we investigated whether four of seven HCCs, in four of six bile duct cancers, in four of purified anti-DKK1 antibody (50 or 100 nmol/L) could six breast cancers, and in three of five cervical cancers. As inhibit the invasion of NIH3T3 cells transfected with expected from the expression profile analysis, the DKK1 tran- DKK1-expressing plasmid. Expectedly, cellular invasion script was hardly detectable in their corresponding normal caused by DKK1 overexpression was suppressed by the tissues (Fig. 1A). We also detected high levels of DKK1 anti-DKK1 antibody, and the number of DKK1-NIH3T3 cells expression in all of 13 pancreatic cancer cell lines, in 2 of 4 that invaded through Matrigel became almost equivalent to gastric cancer cell lines, in all of 7 HCC cell lines, in 2 of 4 bile that of parental NIH3T3 cells (Fig. 2A). duct cancer cell lines, 6 of 14 breast cancer cell lines, and 1 of We then investigated the effect of anti-DKK1 antibody 2 cervical cancer cell lines (Fig. 1B). To verify the overexpres- (50 or 100 nmol/L) on the lung cancer invasive activity sion of DKK1 in tumors, we performed semiquantitative through Matrigel using A549 cells, which showed high levels RT-PCR and immunohistochemical analyses using another of endogenous DKK1 expression. The cellular invasive abil- set of tumor tissue samples and anti-DKK1 antibody, and ity of A549 detected using Matrigel assays was suppressed confirmed that DKK1 was highly expressed in various clinical by the addition of anti-DKK1 antibody into their culture cancer materials, but hardly detectable in their adjacent nor- media, in a dose-dependent manner (P < 0.0001 for mal tissues examined (representative data of tumors were 100 nmol/L, P = 0.0003 for 50 nmol/L; each paired t test; shown in Supplementary Fig. S2A and B). Fig. 2B, left), whereas that of PC-14 or NCI-H2170 cells with a hardly-detectable level of DKK1 was not affected (repre- Serum levels of DKK1 in cancer patients sentative data of PC-14 was shown in Fig. 2B, right). Because DKK1 protein was secreted into the sera of Similarly, we investigated the effect of anti-DKK1 antibody patients with lung or esophageal cancer (31), we measured (50 or 100 nmol/L for 10 days) on the growth of A549 cells, by ELISA the levels of serum DKK1 in serologic samples from which showed high levels of endogenous DKK1 expression. patients with either of these six types of primary cancers that The growth of A549 cells was suppressed by the addition overexpressed DKK1. The mean (±1 SD) of serum DKK1 was of anti-DKK1 antibody into their culture media, in a dose- 23.8 ± 24.2 U/mL in 179 pancreatic cancer patients, 17.2 ± dependent manner (P = 0.0005 for 100 nmol/L, P = 0.0022 18.0 U/mL in 101 gastric cancer patients, 18.3 ± 16.9 U/mL for 50 nmol/L; each paired t test; Fig. 3A, left), whereas that in 168 HCC patients, 12.5 ± 12.2 U/mL in 107 bile duct cancer of PC-14 or NCI-H2170 cells with a hardly-detectable level of patients, 27.0 ± 22.4 U/mL in 169 breast cancer patients, and DKK1 was not affected (representative data of NCI-H2170 28.7 ± 28.0 U/mL in 182 cervical cancer patients (Fig. 1C). In was shown in Fig. 3A, right). To elucidate the molecular contrast, the mean (±1 SD) of serum DKK1 in 207 healthy mechanism of growth inhibition of cancer cells by anti- individuals were 6.1 ± 5.0 U/mL (31). The levels of serum DKK1 antibody, we microscopically examined the morpho- DKK1 protein were significantly higher in cancer patients logic change of A549 cells treated with anti-DKK1 antibody than in healthy volunteers (P < 0.001; Mann-Whitney U test; and found that the cellular morphology was altered an into Fig. 1C). The proportions of the serum DKK1–positive cases oval shape (Fig. 3B). Propidium iodide staining suggested was 50.8% for pancreatic cancer (91 of 179), 38.6% for gastric that most of them were dead cells (Supplementary Fig. S5). cancer (39 of 101), 53.0% for HCC (89 of 168), 29.9% for bile Additional microscopic analysis using anti–cleaved caspase-3 duct cancer (32 of 107), 65.1% for breast cancer (110 of 169), antibody and by TUNEL assay indicated that treatment of and 59.3% for cervical cancer (108 of 182). We further com- lung cancer cells with anti-DKK1 antibody induced apopto- pared the serum levels of DKK1 protein with its expression sis probably through proteolytic activation of caspase-3 levels in primary tumors in 18 lung cancer cases whose (Fig. 3C and D). serum had been collected before surgery (nine patients with DKK1-positive tumors and nine with DKK1-negative Inhibition of lung cancer cell growth by anti-DKK1 tumors). The serum levels of DKK1 protein showed good antibody in mice correlation with the positivity of DKK1 staining in primary We subsequently examined the effect of anti-DKK1 anti- tumors (Supplementary Fig. S3), further implying that body on in vivo cancer growth using a mice model. A549 DKK1 was secreted from the tumor to blood. The results cells with a high level of DKK1 or NCI-H2170 cells with a indicated the high specificity and the great potential of hardly-detectable level of DKK1 were s.c. implanted into serum DKK1 as a biomarker for the detection of a relatively the right shoulder of 6-week-old male BALB/c nude mice large proportion of common cancers. (nu/nu). The six mice with tumor (50 mm3 volume on average) were randomized into two groups and i.p. administered Inhibition of cell growth and invasive activity by with 100 μg/500 μL of a rabbit polyclonal anti-human DKK1 anti-DKK1 antibody antibody or 100 μg/500 μL of normal rabbit IgG at days 1, 4, We previously showed that lung and esophageal cancer 8, 11, 15, 18, 22, 25, 29, and 32 (a total of 10 injections). We patients with DKK1-positive tumors showed shorter cancer- observed a significant A549-derived tumor growth suppres- specific survival period than those with DKK1-negative sion by anti-human DKK1 antibody, compared with mice tumors, and that DKK1 has cellular invasive activity in vitro treated with the same dose of control IgG (P = 0.0013; each (representative data of PC-14 cells was also shown in Sup- paired t test; Fig. 4A and B). H&E staining using frozen sec- plementary Fig. S4; ref. 31). To examine the therapeutic tion of the resected tumors detected a fibrotic change and

www.aacrjournals.org Cancer Res; 70(13) July 1, 2010 5329

Sato et al.

Figure 1. DKK1 expression in primary cancers and serum levels of DKK1 in cancer patients. A, elevated expression of DKK1 transcript in clinical cancer tissues derived from six organs (pancreas, stomach, liver, bile duct, breast, and cervix). B, elevated expression of DKK1 transcript in cancer cell lines (pancreas, stomach, liver, bile duct, breast, and cervix). C, serum levels of DKK1 in patients with various types of cancer. Black lines, average serum level.

5330 Cancer Res; 70(13) July 1, 2010 Cancer Research

DKK1 as an Immunotherapeutic Target

Figure 2. Inhibition of cell-invasive activity by anti-DKK1 antibody. A, Matrigel invasion assay evaluating the effect of anti-DKK1 antibody (50 or 100 nmol/L; Y-axis) on the invasion of NIH3T3 cells transiently transfected with DKK1-expressing plasmids. Cellular invasion caused by DKK1 overexpression was suppressed by addition of anti-DKK1 antibody into their culture media. Each experiment was done in triplicate. B, Matrigel invasion assay evaluating the effect of anti-DKK1 antibody (50 or 100 nmol/L; Y-axis) on the invasion of endogenous DKK1-overexpressing NSCLC cell line A549 (left) and a non–DKK1- expressing NSCLC cell line PC-14 (right). The cellular invasion of A549 cells detected using Matrigel assays was suppressed by addition of anti-DKK1 antibody into their culture media, in a dose-dependent manner, whereas that of PC-14 cells was not affected. Each experiment was done in triplicate.

decrease of viable cancer cells in tumor tissues obtained at body treatment (Supplementary Fig. S6A). In addition, we 35 days after the beginning of anti-DKK1 antibody treat- observed no obvious toxicity including weight loss during/ ment, and a stronger fibrotic change and more significant after treatment or no histopathologic damage in six tis- decrease of viable cancer cells at 60 days (Fig. 4C). On the sues (lung, heart, liver, kidney, bone marrow, and bone) other hand, in vivo growth of DKK1-nonexpressing NCI- in any of three mice administered with anti-DKK1 anti- H2170 cells was not significantly affected by anti-DKK1 antibody (Supplementary Fig. S6B).

www.aacrjournals.org Cancer Res; 70(13) July 1, 2010 5331

Sato et al.

Discussion confirmed that DKK1 was highly expressed in cancers in various organs, but was hardly detectable in their adjacent Despite the development of modern cancer therapy, can- normal tissues. To elucidate the possible mechanism of cer is still one of the major causes of death in the world. DKK1 overexpression in cancer cells, we cultured lung can- Over the last two decades, several candidate molecular tar- cer A549 cells (high endogenous DKK1) and NCI-H2170 cells gets and biomarkers for cancer therapy have been reported. (hardly-detectable level of DKK1) in the presence or absence However, suppression of some of the target molecules can of 5-aza-deoxycytidine and examined DKK1 expression. provide survival benefits to a limited subset of the patients, Treatment of NCI-H2170 cells with 5-aza-deoxycytidine in- and a few useful biomarkers are presently available. In or- duced DKK1 mRNA expression in a dose-dependent man- der to identify molecules involved in pulmonary and esoph- ner, whereas that of A549 cells with 5-aza-deoxycytidine ageal carcinogenesis, and those useful as therapeutic targets did not affect the level of DKK1 expression (Supplementary and biomarkers for cancer, we have established an effective Fig. S7), implying that elevation of DKK1 expression in lung screening system and identified DKK1 as an oncoprotein cancer cells is likely to be caused by some epigenetic altera- whose upregulation is a frequent and important feature of tions including the loss of promoter methylation of the the malignant nature of human cancers. In this study, we DKK1 gene. In this study, we also found high levels of

Figure 3. Inhibition of cell growth by anti-DKK1 antibody in vitro. A, growth-suppressive effect of anti-DKK1 antibody. MTT assay and colony formation assay evaluating the effect of anti-DKK1 antibody (50 or 100 nmol/L; Y-axis) on the growth of DKK1-overexpressing NSCLC cell line A549 (left) and a non–DKK1-expressing NSCLC cell line NCI-H2170 (right). The cell growth of A549 cells detected using MTT assays was suppressed by addition of anti-DKK1 antibody into their culture media, in a dose-dependent manner, whereas that of NCI-H2170 was not affected. Each experiment was done in triplicate. B, morphologic change of A549 cells induced by anti-DKK1 antibody. C and D, induction of apoptosis of A549 cells treated with anti-DKK1 antibody. Immunofluorescent analysis revealed that the cells with morphologic change were stained by anti–cleaved caspase-3 antibody (C) and TUNEL assay (D).

5332 Cancer Res; 70(13) July 1, 2010 Cancer Research

DKK1 as an Immunotherapeutic Target

Figure 4. Growth suppressive effect of anti-DKK1 antibody on DKK1-expressing lung cancer cells that were transplanted to nude mice. A and B, average tumor volumes of three mice treated with anti-DKK1 antibody or IgG (control) were plotted. Animals were administered with each of the antibodies [100 μg/500 μL/animal at days 1, 4, 8, 11, 15, 18, 22, 25, 29, and 32 (a total of 10 injections)] by i.p. injection. Growth of grafted tumors derived from DKK1-expressing A549 cells was significantly suppressed by anti-DKK1 antibody. C, histopathologic examination of H&E-stained tumors (A549) treated with anti-DKK1 antibody. At day 35 after treatment with anti-DKK1 antibody, some decrease of viable cancer cells was observed in tumor tissues treated with anti-DKK1 antibody, compared with those with control IgG. At day 60, a fibrotic change and more significant decrease of viable cancer cells were observed.

DKK1 protein in serologic samples from 906 patients having imply that DKK1 is likely to be involved in critical steps of cancers at an early or advanced disease stage in either of tumor growth/metastasis. We also found the possible inhibi- six organs. Although further validation using a larger set tory effect of anti-DKK1 antibody on cell invasion and growth. of serum samples covering various clinical stages will be Moreover, systemic administration of anti-DKK1 antibody required, our data presented here also indicate a potential effectively suppressed growth of DKK1-positive tumors trans- clinical usefulness of DKK1 as a serologic biomarker for planted into nude mice without any obvious toxicity. NSCLC that could be widely used in clinical practice, such With the advances in understanding aberrant signaling as detection of cancer, prediction of the malignant potential pathways in various types of cancer, many pivotal regulators of tumor, and monitoring the disease condition after any of malignant behavior of cancer cells have emerged as candi- anticancer treatment. The evidence we showed here clearly dates for molecular target–based therapy. Both antibodies

www.aacrjournals.org Cancer Res; 70(13) July 1, 2010 5333

Sato et al.

and small-molecule compounds are promising tools for the phorylation that induces apoptosis in animal cap explants target-based cancer therapy (1). We in this study focused on (49). However, our assay showed that the level of noncano- the therapeutic application of anti-DKK1 antibody for cancer nical Wnt signal such as JNK phosphorylation as well as the treatment. Recently, a few groups reported the usefulness of level and subcellular localization of the canonical Wnt signal anti-DKK1 antibody for preventing bone absorption in some protein of β-catenin was not changed in DKK1-negative disease conditions. Another study showed that daily s.c. lung cancer NCI-H2170 cells after exogenous expression of injection of a neutralizing anti-DKK1 antibody in the area DKK1 protein, or in DKK1-positive A549 cells after the treat- surrounding myelomatous bone could suppress tumor- ment with anti-DKK1 antibody (Supplementary Fig. S8A–D). induced bone resorption and multiple myeloma growth To further examine the effect of exogenous DKK1 expression in vivo without obvious toxicity by ameliorating bone turn- or anti-DKK1 antibody treatment on canonical Wnt path- over presumably through increased osteoblastogenesis and way in lung cancer cells, we performed T-cell factor reporter reduction of osteoclastogenesis (46, 47). The role of DKK1 assays using DKK1-negative lung cancer NCI-H2170 cells in progression of bone lesions has also been studied in pros- that were transfected with pTOP-FLASH or pFOP-FLASH tate cancer; transfection of DKK1 into the osteoblastic pros- vector as well as DKK1 expression plasmid, or using tate cancer cell line C4-2B, which normally induces a mix of DKK1-positive A549 cells transfected with pTOP-FLASH or osteoblastic and osteolytic lesions, caused the cells to devel- pFOP-FLASH vector, which were also treated with anti- op osteolytic tumors in severe combined immunodeficient DKK1 antibody. Although canonical Wnt signal activity mice (43). Current evidence indicated that anti-DKK1 anti- detected as luciferase level was highly elevated in DKK1- body may be a less toxic therapeutic agent to inhibit cancer negative colon cancer DLD-1 cells (control), it was signifi- cell proliferation in certain types of human cancers over- cantly suppressed by exogenous DKK1 expression, which expressing DKK1, although further preclinical studies to was concordant with a previous report that DKK1 is a prove its efficacy and safety in individual cancer treatment negative regulator of Wnt signaling in colon cancer cells are required to qualify anti-DKK1 antibody for the develop- (Supplementary Fig. S9A and B; ref. 38). This suppressive ment of human clinical trials. Moreover, serum DKK1 could effect of exogenous DKK1 expression on Wnt signaling be a safe and less invasive biomarker for the selection of was neutralized by 100 nmol/L anti-DKK1 antibody treat- patients who should receive anti-DKK1 therapy. ment (Supplementary Fig. S9C). On the other hand, canon- Putative mechanisms of antibody-based cancer therapy can ical Wnt signal activity was not so high in DKK1-negative be classified into two categories (1). One is the direct action, lung cancer NCI-H2170 cells compared with DLD-1 cells, which can be further subcategorized into three modes of and it was not significantly affected by exogenous DKK1 ex- action, (a) blocking function, (b) stimulating function, and pression (Supplementary Fig. S9B). Furthermore, anti-DKK1 (c) targeting function. In blocking function, therapeutic anti- antibody treatment showed no detectable effect on Wnt body blocks the function of target signaling molecules or signal activity in DKK1-positive lung cancer A549 cells (Sup- receptors, by blocking ligand binding, inhibiting cell cycle pro- plementary Fig. S9D). The data implied that the induction of gression or DNA repair, inducing the regression of angiogene- apoptosis of lung cancer cells by anti-DKK1 antibody is sis, increasing the internalization of receptors, or reducing likely to be through the undefined signaling pathway. proteolytic cleavage of receptors (1). In stimulating function, Because DKK1 receptors relevant to cancer growth and in- mAb binds to membrane receptors and mimic the effect of vasion still remains unclear, further investigation of this their natural ligand, resulting in cell apoptosis (48). In targeting pathway is required to answer the questions about the function, mAb conjugated with toxins, radioisotopes, cyto- mechanism of anti-DKK1 antibody efficacy. Interestingly, kines, DNA molecules, or even small-molecule agents can selec- recent reports showed that DKK1 consists of conserved tively target tumor cells (1). The other mechanism of antibody NH2-terminal (N1) and COOH-terminal (C1) cysteine-rich therapy is the indirect action mediated by the immune system, regions. Antagonism of canonical Wnt-signaling occurs including complement-dependent cytotoxicity and antibody- through binding of C1 to LRP5/6 proteins on the surface dependent cytotoxicity (1). Our data showed that the treat- of the cell and subsequent disruption of the cell surface ment of lung cancer cells with anti-DKK1 antibody alone Wnt/LRP5/6/Frizzled signaling complex (49). N1 lacks induced apoptosis through the caspase-dependent pathway. the Wnt-antagonizing activity, and the C1 and N1 domains in Because DKK1 is a secretory protein, we may first speculate DKK1 are known to activate distinct signaling pathways. that anti-DKK1 antibody could bind to circulating extracellular The early embryologic activity of DKK1 indeed requires a DKK1 and neutralize its oncogenic function by inhibiting the novel activity that resides within the N1 domain (50). In fact, ligand-receptor interaction, as is the case of bevacizumab our polyclonal antibody to DKK1 was immunized against an (Avastin), a humanized recombinant mAb agent that targets NH2-terminal portion of DKK1 protein (codon 1–120). The the binding of secreted proangiogenic protein VEGF to its function of N1 may be associated with enhanced invasive receptors (VEGFR-1 and VEGFR-2). Another possibility of activity and antiapoptotic signaling pathway that could the antitumor effect of anti-DKK1 antibody is that a be neutralized by anti-DKK1 antibody. ligand-antibody complex comprising DKK1 and anti-DKK1 In summary, we have shown DKK1 as a biomarker for antibody may overdrive the receptor ligand signals and various kinds of common cancers and as a potential target change the balance of cell signals, which can induce cell for therapeutic antibodies. Anti-DKK1 antibody would be a death. DKK1 is known to enhance c-Jun-NH2-kinase phos- promising tool for cancer immunotherapy.

5334 Cancer Res; 70(13) July 1, 2010 Cancer Research

DKK1 as an Immunotherapeutic Target

Disclosure of Potential Conflicts of Interest Grant Support

No potential conflicts of interest were disclosed. Yataro Daigo is a member of Shiga Cancer Treatment Project supported by Shiga Prefecture (Japan). The costs of publication of this article were defrayed in part by the payment Acknowledgments of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We thank BioBank Japan for providing the serum samples and the Kanagawa Cancer Research and Information Association for the arrangement of providing Received 10/21/2009; revised 04/09/2010; accepted 04/29/2010; published clinical tissue samples. OnlineFirst 06/15/2010.

References 1. Imai K, Takaoka A. Comparing antibody and small-molecule thera- like growth factor-II messenger RNA-binding protein 1 is associated pies for cancer. Nat Rev Cancer 2006;6:714–27. with tumor progression in patients with lung cancer. Clin Cancer Res 2. Maloney DG, Grillo-López AJ, White CA, et al. IDEC-C2B8 (Rituximab) 2007;13:434–42. anti-CD20 monoclonal antibody therapy in patients with relapsed low- 20. Suzuki C, Takahashi K, Hayama S, et al. Identification of Myc- grade non-Hodgkin's lymphoma. Blood 1997;90:2188–95. associated protein with JmjC domain as a novel therapeutic target 3. Carter P, Presta L, Gorman CM, et al. Humanization of an anti- oncogene for lung cancer. Mol Cancer Ther 2007;6:542–51. p185HER2 antibody for human cancer therapy. Proc Natl Acad Sci 21. Hayama S, Daigo Y, Yamabuki T, et al. Phosphorylation and activa- U S A 1992;89:4285–89. tion of cell division cycle associated 8 by aurora kinase B plays a 4. Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or significant role in human lung carcinogenesis. Cancer Res 2007;67: with bevacizumab for non-small-cell lung cancer. N Engl J Med 4113–22. 2006;355:2542–50. 22. Taniwaki M, Takano A, Ishikawa N, et al. Activation of KIF4A as a 5. Daigo Y, Nakamura Y. From cancer genomics to thoracic oncology: prognostic biomarker and therapeutic target for lung cancer. Clin discovery of new biomarkers and therapeutic targets for lung and Cancer Res 2007;13:6624–31. esophageal carcinoma. Gen Thorac Cardiovasc Surg 2008;56:43–53. 23. Mano Y, Takahashi K, Ishikawa N, et al. Fibroblast growth factor 6. Kikuchi T, Daigo Y, Katagiri T, et al. Expression profiles of non-small receptor 1 oncogene partner as a novel prognostic biomarker and cell lung cancers on cDNA microarrays: identification of genes for therapeutic target for lung cancer. Cancer Sci 2007;98:1902–13. prediction of lymph-node metastasis and sensitivity to anti-cancer 24. Kato T, Sato N, Hayama S, et al. Activation of HJURP (Holliday drugs. Oncogene 2003;22:2192–205. junction-recognizing protein) involved in the chromosomal stability 7. Kakiuchi S, Daigo Y, Tsunoda T, Yano S, Sone S, Nakamura Y. and immortality of cancer cells. Cancer Res 2007;67:8544–53. Genome-wide analysis of organ-preferential metastasis of human 25. Kato T, Sato N, Takano A, et al. Activation of placenta specific tran- small cell lung cancer in mice. Mol Cancer Res 2003;1:485–99. scription factor distal-less homeobox 5 predicts clinical outcome in 8. Kakiuchi S, Daigo Y, Ishikawa N, et al. Prediction of sensitivity of primary lung cancer patients. Clin Cancer Res 2008;14:2363–70. advanced non-small cell lung cancers to gefitinib (Iressa, ZD1839). 26. Dunleavy EM, Roche D, Tagami H, et al. HJURP is a cell-cycle- Hum Mol Genet 2004;13:3029–43. dependent maintenance and deposition factor of CENP-A at 9. Kikuchi T, Daigo Y, Ishikawa N, et al. Expression profiles of meta- centromeres. Cell 2009;137:485–97. static brain tumor from lung adenocarcinomas on cDNA microarray. 27. Hirata D, Yamabuki T, Ito T, et al. Involvement of epithelial cell trans- Int J Oncol 2006;28:799–805. forming sequence 2 (ECT2) oncoantigen in lung and esophageal 10. Taniwaki M, Daigo Y, Ishikawa N, et al. Gene expression profiles of cancer progression. Clin Cancer Res 2009;15:256–66. small-cell lung cancers: molecular signatures of lung cancer. Int J 28. Sato N, Koinuma J, Fujita M, et al. Activation of WD repeat and Oncol 2006;29:567–75. HMG-box DNA binding protein 1 in pulmonary and esophageal 11. Yamabuki T, Daigo Y, Kato T, et al. Genome-wide gene expression carcinogenesis. Clin Cancer Res 2010;16:226–39. profile analysis of esophageal squamous cell carcinomas. Int J Oncol 29. Ishikawa N, Daigo Y, Yasui W, et al. ADAM8 as a novel serological 2006;28:1375–84. and histochemical marker for lung cancer. Clin Cancer Res 2004;10: 12. Suzuki C, Daigo Y, Kikuchi T, Katagiri T, Nakamura Y. Identification 8363–70. of COX17 as a therapeutic target for non-small cell lung cancer. 30. Ishikawa N, Daigo Y, Takano A, et al. Increases of amphiregulin and Cancer Res 2003;63:7038–41. transforming growth factor-α in serum as predictors of poor response 13. Kato T, Daigo Y, Hayama S, et al. A novel human tRNA-dihydrouridine to gefitinib among patients with advanced non-small cell lung synthase involved in pulmonary carcinogenesis. Cancer Res 2005;65: cancers. Cancer Res 2005;65:9176–84. 5638–46. 31. Yamabuki T, Takano A, Hayama S, et al. Dickkopf-1 as a novel 14. Furukawa C, Daigo Y, Ishikawa N, et al. Plakophilin 3 oncogene as serologic and prognostic biomarker for lung and esophageal carci- prognostic marker and therapeutic target for lung cancer. Cancer nomas. Cancer Res 2007;67:2517–25. Res 2005;65:7102–10. 32. Ishikawa N, Takano A, Yasui W, et al. Cancer-testis antigen lympho- 15. Suzuki C, Daigo Y, Ishikawa N, et al. ANLN plays a critical role in cyte antigen 6 complex locus K is a serologic biomarker and a ther- human lung carcinogenesis through the activation of RHOA and by apeutic target for lung and esophageal carcinomas. Cancer Res involvement in the phosphoinositide 3-kinase/AKT pathway. Cancer 2007;67:11601–11. Res 2005;65:11314–25. 33. Takano A, Ishikawa N, Nishino R, et al. Identification of Nectin-4 16. Ishikawa N, Daigo Y, Takano A, et al. Characterization of SEZ6L2 oncoprotein as a diagnostic and therapeutic target for lung cancer. cell-surface protein as a novel prognostic marker for lung cancer. Cancer Res 2009;69:6694–703. Cancer Sci 2006;97:737–45. 34. Suda T, Tsunoda T, Daigo Y, Nakamura Y, Tahara H. Identification of 17. Takahashi K, Furukawa C, Takano A, et al. The neuromedin u-growth human leukocyte antigen-A24-restricted epitope peptides derived hormone secretagogue receptor 1b/neurotensin receptor 1 onco- from gene products upregulated in lung and esophageal cancers genic signaling pathway as a therapeutic target for lung cancer. as novel targets for immunotherapy. Cancer Sci 2007;98:1803–8. Cancer Res 2006;66:9408–19. 35. Mizukami Y, Kono K, Daigo Y, et al. Detection of novel cancer-testis 18. Hayama S, Daigo Y, Kato T, et al. Activation of CDCA1-2, members of antigen-specific T-cell responses in TIL, regional lymph nodes, and centromere protein complex, involved in pulmonary carcinogenesis. PBL in patients with esophageal squamous cell carcinoma. Cancer Cancer Res 2006;66:10339–48. Sci 2008;99:1448–54. 19. Kato T, Hayama S, Yamabuki Y, et al. Increased expression of insulin- 36. Kono K, Mizukami Y, Daigo Y, et al. Vaccination with multiple

www.aacrjournals.org Cancer Res; 70(13) July 1, 2010 5335

Sato et al.

peptides derived from novel cancer-testis antigens can induce osteoblastic bone metastases through Wnts. Cancer Res 2005; specific T-cell responses and clinical responses in advanced esoph- 65:7554–60. ageal cancer. Cancer Sci 2009;100:1502–9. 44. Wirths O, Waha A, Weggen S, et al. Overexpression of human 37. Glinka A, Wu W, Delius H, et al. Dickkopf-1 is a member of a new Dickkopf-1, an antagonist of wingless/WNT signaling, in human family of secreted proteins and functions in head induction. Nature hepatoblastomas and Wilms' tumors. Lab Invest 2003;83:429–34. 1998;391:357–62. 45. Lee N, Smolarz AJ, Olson S, et al. A potential role for Dkk-1 in the 38. Niida A, Hiroko T, Kasai M, et al. DKK1, a negative regulator of Wnt pathogenesis of osteosarcoma predicts novel diagnostic and treat- signaling, is a target of the β-catenin/TCF pathway. Oncogene 2004; ment strategies. Br J Cancer 2007;97:1552–9. 23:8520–6. 46. Yaccoby S, Ling W, Zhan F, et al. Antibody-based inhibition of DKK1 39. Semënov MV, Tamai K, Brott BK, et al. Head inducer Dickkopf-1 is a suppresses tumor-induced bone resorption and multiple myeloma ligand for Wnt coreceptor LRP6. Curr Biol 2001;11:951–61. growth in vivo. Blood 2007;109:2106–11. 40. Marvin MJ, Di Rocco G, Gardiner A, et al. Inhibition of Wnt activity 47. Fulciniti M, Tassone P, Hideshima T, et al. Anti-DKK1 mAb (BHQ880) induces heart formation from posterior mesoderm. Genes Dev 2001; as a potential therapeutic agent for multiple myeloma. Blood 2009; 15:316–27. 114:371–9. 41. Politou MC, Heath DJ, Rahemtulla A, et al. Serum concentrations of 48. Congy-Jolivet N, Probst A, Watier H, et al. Recombinant therapeutic Dickkopf-1 protein are increased in patients with multiple myeloma monoclonal antibodies: mechanisms of action in relation to structural and reduced after autologous stem cell transplantation. Int J Cancer and functional duality. Crit Rev Oncol Hematol 2007;64:226–33. 2006;119:1728–31. 49. Niehrs C. Function and biological roles of the Dickkopf family of Wnt 42. Forget MA, Turcotte S, Beauseigle D, et al. The Wnt pathway regulator modulators. Oncogene 2006;25:7469–81. DKK1 is preferentially expressed in hormone-resistant breast tumours 50. Korol O, Gupta RW, Mercola M. A novel activity of the Dickkopf-1 and in some common cancer types. Br J Cancer 2007;96:646–53. amino terminal domain promotes axial and heart development inde- 43. Hall CL, Bafico A, Dai J, et al. Prostate cancer cells promote pendently of canonical Wnt inhibition. Dev Biol 2008;324:131–8.

5336 Cancer Res; 70(13) July 1, 2010 Cancer Research

Wnt Inhibitor Dickkopf-1 as a Target for Passive Cancer Immunotherapy

Nagato Sato, Takumi Yamabuki, Atsushi Takano, et al.

Cancer Res 2010;70:5326-5336. Published OnlineFirst June 15, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-09-3879

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2010/06/14/0008-5472.CAN-09-3879.DC1

Cited articles This article cites 50 articles, 26 of which you can access for free at: http://cancerres.aacrjournals.org/content/70/13/5326.full#ref-list-1

Citing articles This article has been cited by 9 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/70/13/5326.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://cancerres.aacrjournals.org/content/70/13/5326. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.