Selective Inhibition of ADAM28 Suppresses Lung Carcinoma Cell

Published OnlineFirst September 6, 2018; DOI: 10.1158/1535-7163.MCT-17-1198

Large Molecule Therapeutics Molecular Cancer Therapeutics Selective Inhibition of ADAM28 Suppresses Lung Carcinoma Cell Growth and Metastasis Satsuki Mochizuki1,2, Masayuki Shimoda1, Hitoshi Abe1, Yuka Miyamae1, Junko Kuramoto1, Noriko Aramaki-Hattori3, Ken Ishii4, Hideki Ueno2, Akira Miyakoshi5, Kanehisa Kojoh5, and Yasunori Okada1,6

Abstract

ADAM28 (a disintegrin and metalloproteinase 28) is over- inhibited IGF-1–stimulated cell proliferation of lung adeno- expressed by carcinoma cells in non–small cell lung carcino- carcinoma cell lines with ADAM28 expression, including PC-9 mas (NSCLC) and plays an important role in cancer cells, and promoted VWF-induced cell death in these cell lines. cell proliferation and metastasis by reactivation of insulin-like In lung metastasis models, antibody 211-14 significantly growth factor-1 (IGF-1) and escaping from von Willebrand reduced tumor growth and metastases of PC-9 cells and factor (VWF)–induced apoptosis through digestion of IGF- prolonged survivals in the antibody-treated mice compared binding protein-3 and VWF, respectively. To aim for new target with the control IgG-treated ones. Combination therapy of the therapy of NSCLC patients, we developed human neutralizing antibody and docetaxel was more effective than that of bev- antibodies 211-12 and 211-14 against ADAM28, which acizumab and docetaxel and showed further elongation of showed IC50 values of 62.4 and 37.5 nmol/L, respectively. survival time compared with monotherapy. No adverse effects Antibody 211-14 recognized the junctional region between were observed even after administration of 10-fold more than cysteine-rich domain and secreted-specific domain and effective dose of anti-ADAM28 antibody to normal mice. Our showed a KD value of 94.7 pmol/L for the epitope-containing data demonstrate that antibody 211-14 is a neutralizing anti- peptide. This antibody detected monkey and human secreted- body specific to ADAM28s and suggest that this antibody form ADAM28s, although it was not reactive with mouse may be a useful treatment remedy for NSCLC patients. membrane-anchored ADAM28m. Antibody 211-14 effectively Mol Cancer Ther; 17(11); 2427–38. 2018 AACR.

Introduction ADAM9, ADAM12, and ADAM28 have short-secreted splice var- iants besides prototype membrane-anchored form, and they are ADAMs (a disintegrin and metalloproteinases) are a family of named as follows: ADAM9s and ADAM9m (6), ADAM12s and transmembrane proteins that mediate cell-surface proteolysis and ADAM12m (7), and ADAM28s and ADAM28m (8). Our previous modulate cell–cell and cell–matrix interactions (1, 2). The human systematic studies on the expression of the proteolytic ADAMs ADAM gene family is composed of 13 proteolytic ADAMs that have indicated that both ADAM28m and ADAM28s forms are possess the characteristic reprolysin-type active site abundantly overexpressed predominantly by carcinoma cells in (HEXGHXXGXXHD) in the metalloproteinase domain and eight human non–small cell lung carcinoma (NSCLC) and breast nonproteolytic ADAMs (1). The typical structure of the proteolytic carcinoma tissues, showing positive correlations with carcinoma ADAMs consists of propeptide, metalloproteinase, disintegrin- cell proliferation and/or metastasis in these carcinomas (9, 10). like, cysteine-rich, epidermal growth factor (EGF)-like, transmem- We have also shown that serum levels of ADAM28 in NSCLC brane and cytoplasmic domains (1, 3–5). Some members such as patients signiﬁcantly increase with the tumor staging, lymph node metastasis, and carcinoma recurrence (11), and another group has suggested the possibility of ADAM28 as a biomarker for asbestos- 1 Department of Pathology, Keio University School of Medicine, Tokyo, Japan. related lung carcinomas (12). Concerning biological activities of 2Department of Surgery, National Defense Medical College, Saitama, Japan. ADAM28, we have demonstrated that ADAM28 contributes to 3Department of Plastic and Reconstructive Surgery, Keio University School of Medicine, Tokyo, Japan. 4Department of Orthopaedic Surgery, Keio University carcinoma cell proliferation through enhanced bioavailability of School of Medicine, Tokyo, Japan. 5GeneFrontier Corporation, Chiba, Japan. insulin-like growth factor-1 (IGF-1) by selective digestion of IGF- 6Department of Pathophysiology for Locomotive and Neoplastic Diseases, binding protein-3 (IGFBP-3) in the IGF-1/IGFBP-3 complex Juntendo University Graduate School of Medicine, Tokyo, Japan. (10, 13), and that ADAM28 enhances lung metastasis of lung Note: Supplementary data for this article are available at Molecular Cancer adenocarcinoma cells by escaping from von Willebrand factor Therapeutics Online (http://mct.aacrjournals.org/). (VWF)-induced apoptosis through digestion of VWF within blood Corresponding Authors: Yasunori Okada, Juntendo University Graduate School vessels (14). All these data have indicated that ADAM28 is of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113-8421, Japan. Phone: 81-3-5800- implicated in tumor growth and metastasis of NSCLC and breast 7531; Fax: 81-3-5800-7532; E-mail: [email protected]; and Satsuki carcinoma and suggest that ADAM28 may be a target molecule for Mochizuki, Department of Surgery, National Defense Medical College. Phone: therapy in these carcinoma patients. 81-4-2995-1637; Fax: 81-4-2996-5205; E-mail: [email protected] Lung carcinoma is one of the most common cancers in the doi: 10.1158/1535-7163.MCT-17-1198 world and the leading cause of cancer-related deaths in many 2018 American Association for Cancer Research. countries (15). NSCLC represents 80% to 85% of all lung cancers,

www.aacrjournals.org 2427

Mochizuki et al.

and surgical resection is curative for patients with localized followed by incubation with Protein G Sepharose 4 Fast Flow NSCLC. However, about 70% of NSCLC patients are in an beads (GE Healthcare Life Sciences) as previously described (26). advanced stage when diagnosed, and they are not candidates for Mouse anti-ADAM28 monoclonal antibody against the ADAM28 surgical interventions (16). Cytotoxic chemotherapy using plat- metalloproteinase domain (5 mg/mL; 297-2F3; ref. 10) was used inum-based anticancer drugs together with other agents such as as a positive control. Proteins electrotransferred onto polyviny- docetaxel was the only effective treatment to prolong survival and lidene difluoride (PVDF) membranes were immunoblotted with relax a symptom in these patients (17, 18). In recent years, the antibody (297-2F3; refs. 9, 10), and reacted with horseradish however, molecular targeted therapies have been introduced and peroxidase (HRP)-conjugated secondary antibody (#K4001, presented an alternative therapeutic approach for patients whose Dako), followed by incubation with chemiluminescence reagent tumors bear targeting oncogenic drivers, i.e., genomic alterations (Thermo Fisher Scientific). The bound antibody was detected by critical to cancer development and maintenance. They include at an imaging plate and BAS-2000 system (Fuji Photo Film Co.). least 10 drivers such as KRAS mutations, EGFR mutations and ALK rearrangements (19). Among them, target therapies for NSCLC Inhibition of ADAM28 activity by candidate antibodies patients with tumors harboring EGFR mutations or ALK rearran- Human ADAM28s (20 ng) was incubated with the Fabs (20 ng) gements have yielded high response rates, longer progression-free and then reacted with recombinant IGFBP-3 (200 ng; R&D survival and prolonged overall survival compared with traditional Systems) for 16 hours at 37C. The IGFBP-3 digestion was cytotoxic chemotherapies (19, 20). However, these mutations determined by immunoblotting with anti-IGFBP-3 antibody account for less than 25% of NSCLC patients and drug resistance (#sc-6003, Santa Cruz Biotechnology). For the titration study, frequently develops in the majority of patients (21). More recent- ADAM28s was incubated with different concentrations of anti- ly, immune-checkpoint inhibitors targeting the cytotoxic T lym- bodies 211-12 and 211-14 or control human IgG (#1-001-A, R&D phocyte-associated antigen 4 (CTLA-4) and programmed death Systems) and then subjected to IGFBP-3 digestion. Inhibitory receptor 1/programmed death ligand 1 (PD-1/PD-L1) axes have potency (IC50 values) was determined by densitometric analysis been developed and shown significant survival benefit in some of the protein bands of 18/22-kDa IGFBP-3 fragments and intact patients with advanced NSCLC (18). However, response rate is IGFBP-3 of 40/45-kDa using ImageJ software. approximately 20% of NSCLC patients (22, 23) and drug resistance is commonly observed after treatment with immune- Cross-reactivity of antibodies 211-12 and 211-14 with mouse checkpoint inhibitors, although the frequency of acquired resis- and macaque ADAM28 species tance to checkpoint blockade immunotherapies is undetermined Homogenate supernatants (10 mg protein/lane) of the testis (24). Therefore, new therapies besides traditional chemotherapy, and epididymis obtained from C57BL/6NCrl mice (Charles River targeted therapies for driver mutations, and immune-checkpoint Laboratories) and common marmosets (Callithrix jacchus; Central therapy are required. Institute for Experimental Animals) were subjected to immuno- In the current study, we developed human neutralizing precipitation with antibodies 211-12 and 211-14. ADAM28 in the antibody 211-14 to human ADAM28s by screening a human immunoprecipitates was detected by immunoblotting with anti- antibody library, characterized the specificity, inhibitory activ- body to mouse ADAM28 (sc-367055, Santa Cruz Biotechnology) ity, and efficiency of this antibody in vitro, and examined the and antibody 211-14 for the mouse and common marmoset effect on tumor growth and metastasis and survivals in mouse tissues, respectively. Anti-glyceraldehyde-3-phosphate dehydro- metastasis models. genase (GAPDH) antibody (ab125247, Abcam) was used for loading control. Materials and Methods Antibody characterization by immunoblotting and epitope Panning of phage display library and Fab generation mapping The Human Combinatorial Antibody Library (HuCAL GOLD ; Cross-reactivity of antibody 211-14 with main members of the MorphoSys AG and GeneFrontier Corporation) was used for the ADAM, ADAM with thrombospondin motifs (ADAMTS) and screening of antibodies. Recombinant proADAM28s and its acti- matrix metalloproteinase (MMP) families was tested by immu- vated form (13, 25) were biotinylated, immobilized on the noblotting using recombinant human ADAM9 (R&D Systems), streptavidin-coated beads and incubated with HuCAL GOLD . ADAM10 (R&D Systems), ADAM12 (Mochida Pharmaceutical), Bound Fab-displaying phages were enriched in three consecutive ADAM17 (R&D Systems), ADAMTS1 (R&D Systems), ADAMTS4 panning rounds. The pool of Fab genes was isolated and inserted (R&D Systems), ADAMTS5 (R&D Systems), MMP-1 (Kyowa into Escherichia coli expression vectors that lead to periplasmic Pharma Chemical), MMP-2 (Kyowa Pharma Chemical), MMP-3 expression of monovalent Fab equipped with Strep-tag II. The (Kyowa Pharma Chemical), MMP-7 (Kyowa Pharma Chemical), crude extracts of transformed colonies were tested in ELISA with MMP-9 (Kyowa Pharma Chemical), and MMP-13 (Millipore), all the immobilized antigens. Positive colonies were chosen for of which were also stained with silver stain kit (CosmoBio). The purification, and some of the Fabs were reformatted into whole specific reaction of antibody 211-14 with human ADAM28s was human IgG1. examined by immunoblotting of culture media obtained from PC-9 lung adenocarcinoma cells (1 mL/lane; ref. 14) and homo- Immunoprecipitation and immunoblotting genates of human lung adenocarcinoma tissues (20 mg/lane; Culture media containing human proADAM28s were prepared ref. 11). Immunoreactive domains of proADAM28s with anti- by infection of sf9 cells with pFASTBac1/proADAM28s (13, 25) body 211-14 were determined by immunoblotting using human and used for immunoprecipitation. The media were incubated recombinant ADAM28s proteins covering propeptide (Pro) and with candidate Fabs or nonimmune human IgG (negative control; metalloproteinase (MP) domains (Pro/MP) or disintegrin-like 5 mg/mL each) and reacted with anti-human Fab antibody, (Dis), cysteine-rich (CR) and secretory-specific (SS) domains

2428 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Human Antibody Inhibitor to ADAM28

(Dis/CR/SS), both of which were equipped with FLAG tag (25). Analyses of tumor growth and metastases in early-stage and For epitope mapping, maltose binding protein (MBP)-fused advanced-stage metastasis mouse models peptides covering parts of the Dis, CR, and/or SS domain of PC-9ffLuc-cp156 cells, which express the fusion protein of ADAM28s and that of ectodomain of ADAM28m were expressed firefly luciferase and circularly permutated variant of Venus by Escherichia coli using pMAL-c4X vector (N8107S, New England (ffLuc-cp156; a variant of green fluorescent protein [GFP]), were Biolabs), and subjected to immunoblotting with antibody 211-14 established and characterized previously (14). The cells (1 106 or anti-MBP antibody (E8038S, New England Biolabs). cells/mouse) were injected into the tail vein of NOD/SCID male mice (Charles River Laboratories) under anesthesia. Tumor Immunocytochemistry in ADAM28 transfectants growth and metastasis were monitored by bioluminescence imag- Chinese hamster ovary (CHO-K1) cells, which were purchased ing using the In Vivo Imaging System (IVIS)-100 camera system from Health Science Research Resources Bank (Osaka) and val- (Xenogen; ref. 14). In the early-stage metastasis model, trap of idated by the Bank in 2008, were transfected with pIRES2-EGFP- PC-9ffLuc-cp156 carcinoma cells in the lungs was confirmed by ADAM28s, pcDNA3.1-ADAM28m/pIRES2-EGFP, or pIRES-EGFP imaging at 1 hour after intravenous injection, and the mice were empty vectors by Lipofectamine LTX reagent (Invitrogen). They randomly divided into two groups (n ¼ 10 each). At 1 day after were incubated with antibody 211-14 (20 mg/mL), mouse carcinoma cell injection, they received intraperitoneal injections monoclonal antibody (50 mg/mL; 297-2F3), or nonimmune of antibody 211-14 or nonimmune IgG (2 mg/kg body weight IgG (20 mg/mL) in Ham's F-12 medium (Wako Pure Chemical each) 5 times every other day. Some mice were sacrificed at 1 day Industries) containing 10% fetal bovine serum, fixed with 4% and 3 days after intravenous carcinoma cell injection, and the lung paraformaldehyde, and then reacted with secondary antibodies tissues were histologically examined. In addition, left lower lobes conjugated to Alexa Flour 555 (Thermo Fisher Scientific). of the lung were isolated from the mice at 1 day after the first After staining with Hoechst 33342 (Thermo Fisher Scientific), intraperitoneal injection of antibody 211-14, and the homoge- they were observed using a fluorescence microscope (Olympus nate supernatants were subjected to immunoblotting with anti- IX71; Olympus). cleaved caspase-3 antibody or anti–b-actin antibody (14). For advanced-stage metastasis model, the mice were left for 3 weeks Surface plasmon resonance interaction analysis after intravenous injection of PC-9ffLuc-cp156 cells. Lung metastases The MBP-fused Thr517-Thr526 peptide of ADAM28s, which of the mice were confirmed by IVIS, and they were randomly covers parts of the CR and SS domains, was covalently immobi- divided into two groups (n ¼ 10 each). The mice were received lized via amine coupling on CM5 sensor chip flow chambers, and injections of antibody 211-14 or nonimmune IgG (2 mg/kg body antibody 211-14 was injected to the chambers using Biacore 3000 weight each) 5 times every week. To test the effects of combination (GE Healthcare Life Sciences). The KD (the affinity) was calculated therapy of antibody 211-14 or human antibody against vascular from the determined Ka and Kd values. endothelial growth factor (bevacizumab; Chugai Pharmaceutical) with docetaxel (Sawai Pharmaceutical) on survival rates, the mice Cell proliferation and DNA fragmentation assays in lung in five groups (n ¼ 10 each) received intraperitoneal injections of carcinoma cell lines bevacizumab alone, nonimmune IgG alone, docetaxel alone, PC-9 lung adenocarcinoma cell line was purchased from antibody 211-14 plus docetaxel or bevacizumab plus docetaxel. Immuno-Biological Laboratories (Gumma) in 2014, and The antibodies and nonimmune IgG (2 mg/kg body weight each) H1975, H1650, H441, and Calu-3 lung adenocarcinoma cell were intraperitoneally injected 5 times every week, and docetaxel lines were from ATCC in 2016. Early-passage aliquots were frozen, (20 mg/kg body weight) was intraperitoneally administered once. and cell authentication was performed in January 2017 by STR Survival rates were analyzed by the Kaplan–Meier log-rank test. All analysis at Bex Co, Ltd. They were also confirmed as mycoplasma procedures of mouse experiments were performed according to negative with MycoAlert assay (Lonza Inc). The mRNA expression the guidelines for the Care and Use of Laboratory Animals of Keio of ADAM28s, ADAM28m, IGF-1, IGF-1 receptor, IGFBP-3, and University School of Medicine. b-actin in these cell lines was examined by RT-PCR using the primers for these molecules (Supplementary Table S1). Inhibitory Histology and IHC effect of antibody 211-14 on IGF-1 (100 ng/mL)–induced cell The lung tissues from the mice were fixed in buffered formalin, proliferation was examined by Cell Proliferation ELISA Kit and paraffin sections were stained with hematoxylin and eosin (#11647229001, Roche Applied Science; ref. 10). The half max- and thiosemicarbazide-periodic acid methenamine silver- imal effective concentration (EC50) in PC-9 cells was determined hematoxylin (TSC-PAM; ref. 14). They were also immunostained by measuring inhibition rate after incubation with different for Venus and Ki-67 with anti-GFP antibody (#598, Medical and concentrations of antibody 211-14 or control human IgG. The Biological Laboratory) and anti-Ki-67 antibody (MIB-1, DAKO), DNA fragmentation assay was performed by the Cell Death respectively, followed by reactions with ImmPRESS anti-mouse Detection ELISAPLUS kit (#11774425001, Roche Applied IgG (#MP-7402-50, Vector Laboratories) or anti-rabbit IgG sec- Science). Briefly, the cell lines (1 105 cells/mL) were incubated ondary antibody (#MP-7401-50, Vector Laboratories). Color was with 3 mg/mL VWF (a gift from Dr. Kenji Soejima in Kaketsuken, developed with 3, 30-diaminobenzidine tetrahydrochloride Kumamoto) in the presence or absence of 5 mg/mL antibody (10, 14), and the sections were counterstained with hematoxylin. 211-14 or control human IgG in serum-free medium for 48 hours at 37C (14). Supernatants of cell lysates were reacted with Toxicity study of antibody 211-12 in mice anti-histone (biotin-labeled) and anti-DNA (peroxidase- Because antibody 211-12 could inhibit ADAM28 activity and conjugated) antibodies, and then incubated with 2,20-azino-di cross-react with mouse ADAM28 (Figs. 1D and 2B), the antibody (3-ethylbenzthiazoline)-6-sulphuric acid (substrate solution). reformatted into mouse IgG was used for toxicity study. Male The absorbance was measured using microplate reader. C57BL/6NCrl mice (20 g; Charles River Laboratories) were

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2429

Mochizuki et al.

Figure 1. Screening of human neutralizing anti-ADAM28 antibodies. A, ELISA of candidate Fabs (211-02, 211-06, 211-12, 211-14, 211-26, and 213-09) selected by phage display of HuCAL. Human proADAM28s (left) or active ADAM28s (right) was coated on microplates and then reacted with the Fabs, followed by reactions 0 with HRP-conjugated antibody speciﬁc to human F(ab )2. As for a control, bovine serum albumin (BSA)-coated microplates were reacted with the Fabs. B, Immunoprecipitation of proADAM28s by candidate Fabs. Culture media containing human proADAM28s were subjected to immunoprecipitation (IMP) by candidate Fabs, anti-ADAM28 mouse monoclonal antibody (297-2F3; positive control) or control human IgG (Cont IgG; negative control), and immunoprecipitates were immunoblotted (IMB) with mouse anti-ADAM28 antibody (297-2F3). Cont, culture medium alone without IMP. C, Inhibition of ADAM28s activity by IGFBP-3 cleavage assay. Human active ADAM28s was reacted with candidate Fabs, anti-ADAM28 antibody (297-2F3) or Cont IgG, and then incubated with IGFBP-3. The IGFBP-3 digestion was monitored by immunoblotting with anti-IGFBP-3 antibody. Cont and None, IGFBP-3 incubated with buffer alone and ADAM28s

alone, respectively. Arrows and arrowheads, intact IGFBP-3 and its digestion fragments, respectively. D, Determination of IC50 of antibody 211-12 and antibody 211-14. Human ADAM28s was incubated with different concentrations of the antibodies or nonimmune IgG (Cont IgG) and subjected to the IGFBP-3 digestion, followed by immunoblotting.

n ¼ randomly divided into two groups ( 10 each), and the mice in Results each group were intraperitoneally injected 5 times every week with Fc-mousenized antibody 211-12 or normal mouse IgG at a Screening of human neutralizing anti-ADAM28 antibodies dose of 20 mg/kg body weight, which is 10-fold more than By screening the human antibody library HuCAL, we found fi effective dose of antibody 211-14. They were weighed every week that ve Fabs (antibodies 211-02, 211-06, 211-12, 211-14, and and observed for any death or changes in general behavior and 211-26) are immunoreactive with human proADAM28s and two other physical activities for 6 weeks. At 1 week after the final Fabs (antibodies 211-14 and 213-09) react with active ADAM28s injection of the antibody, they were anesthetized and blood was by ELISA (Fig. 1A). Immunoprecipitation showed that four anti- collected via cardiac puncture for hematologic parameters and bodies 211-02, 211-12, 211-14, and 211-26, besides control anti- biochemical estimations. At autopsy, whole organs were ADAM28 mouse monoclonal antibody (297-2F3; refs. 9, 10), are removed, weighed, and then subjected to histologic study. capable of immunoprecipitating proADAM28s from culture media containing human proADAM28s (Fig. 1B). Inhibition Statistical analyses study by IGFBP-3 cleavage assay indicated that the digestion of All statistical analyses were performed using a two-tailed Stu- intact IGFBP-3 protein by human ADAM28s is suppressed by dent t test. The significance of differences among more than three antibodies 211-12 and 211-14 in addition to anti-ADAM28 groups was determined by one-way analysis of variance followed mouse monoclonal antibody (297-2F3) (Fig. 1C). Thus, these by Bonferroni post hoc testing. P values less than 0.05 were two Fabs were reformatted into whole human IgG1 for further considered statistically significant. studies, and inhibitory potency (IC50) of antibodies 211-12 and

2430 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Human Antibody Inhibitor to ADAM28

Figure 2. Immunoblotting and cross-reactivity of antibody 211-12 and antibody 211-14. A, Immunoblotting of human proADAM28s and active ADAM28s by antibody 211-12 and antibody 211-14. Puriﬁed human proADAM28s, active ADAM28s, and culture medium containing human proADAM28s were subjected to immunoblotting with antibody 211-12, antibody 211-14, and mouse anti-ADAM28 antibody (297-2F3; positive control). B, Immunologic cross-reactivity of antibody 211-12 and antibody 211-14 with mouse ADAM28. Homogenate supernatants of the mouse testis (negative control) and epididymis were subjected to immunoprecipitation (IMP) with antibodies 211-12 and 211-14, and the immunoprecipitates were detected by immunoblotting (IMB) with antibody speciﬁc to mouse ADAM28. GAPDH, loading control. C, Immunological cross-reactivity of antibody 211-12 and antibody 211-14 with common marmoset ADAM28. Homogenate supernatants of the common marmoset testis (negative control) and epididymis were subjected to immunoprecipitation (IMP) with antibodies 211-12 and 211-14, and the immunoprecipitates were detected by immunoblotting (IMB) with antibody 211-14. GAPDH, loading control.

211-14 to human ADAM28s activity was determined to be 62.4 with monkey active ADAM28s (http://www.uniprot.org/uniprot/ and 37.5 nmol/L from the titration curves, respectively (Fig. 1D). A0A2K5X1T0) (Fig. 2C). Thus, these findings demonstrate that both antibodies 211-12 and 211-14 recognize monkey Immunoblotting analysis of antibodies 211-12 and 211-14 ADAM28s, and antibody 211-12 cross-reacts with mouse and their cross-reactivity with mouse and monkey ADAM28m. The immunoblotting data of antibody 211-12, show- ADAM28 species ing no bands for human proADAM28s and active ADAM28s By immunoblotting, antibody 211-14 recognized purified (Fig. 2A) or mouse epididymis active ADAM28m (Supplementary human proADAM28s of 61 kDa and active ADAM28s of 42 kDa Fig. S1A), suggest that this antibody may recognize the three- and specifically detected proADAM28s in the culture media, while dimensional structure of the ADAM28 species, although the antibody 211-12 showed no immunoreactivity with these species detailed analyses are necessary by future work. (Fig. 2A). Because the epididymis, but not the testis, expresses ADAM28 in mice and macaques as well as humans (27–29), we Determination of specificity and epitope of antibody 211-14 tested immunoreactivity of antibodies 211-12 and 211-14 with Because antibody 211-14 detected human ADAM28s by ADAM28 species in the tissue homogenates from mice or com- immunoblotting and neutralized the activity, we focused on this mon marmosets by immunoprecipitation. As shown in Fig. 2B, antibody for further studies, and found that the antibody antibody 211-12, but not antibody 211-14, immunoprecipitated recognizes human ADAM28 without cross-reactivity with mouse ADAM28 band of 80 kDa, which corresponds to active ADAM9, ADAM10, ADAM12, ADAM17, ADAMTS1, ADAMTS4, ADAM28m (27). By immunoblotting analysis, however, neither ADAMTS5, MMP1, MMP2, MMP3, MMP7, MMP9, or MMP13 antibody 211-12 nor antibody 211-14 showed positive bands in (Fig. 3A). In addition, the antibody specifically detected the the mouse tissue homogenates (Supplementary Fig. S1). On the 42-kDa active form of human ADAM28s in the homogenates of other hand, both antibodies immunoprecipitated ADAM28 of PC-9 cells and human lung adenocarcinoma tissues (Fig. 3B). 40 kDa from the common marmoset epididymis, which accords Immunoblotting of the recombinant NH2- and COOH-terminal

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2431

Mochizuki et al.

Figure 3. Speciﬁcity of antibody 211-14 to human ADAM28s and its epitope mapping. A, No cross-reactivity of antibody 211-14 with main members of the ADAM, ADAMTS, and MMP families by immunoblotting. Recombinant proteins of human ADAM28, 9, 10, 12, and 17; ADAMTS1, 4, and 5; and MMP1, 2, 3, 7, 9, and 13 (100 ng/lane) were subjected to SDS-PAGE and analyzed by silver stain and immunoblotting with antibody 211-14. B, Speciﬁc immunoreaction of antibody 211-14 with human active ADAM28s. Homogenate supernatants of PC-9 lung carcinoma cells and human lung adenocarcinoma tissues were subjected to immunoblotting with antibody 211-14 or nonimmune control IgG. C, Immunoreactivity of antibody 211-14 with the Dis/CR/SS domains of human ADAM28. Recombinant FLAG-tagged ADAM28 proteins corresponding to the Pro and Met domains (Pro/MP) or the Dis, CR, and SS domains (Dis/CR/SS) were subjected to immunoblotting with anti-FLAG antibody (left) or antibody 211-14 (right). D, Epitope mapping analysis of antibody 211-14. Maltose binding protein (MBP)-fused peptides covering the sequences of the ADAM28s Dis/CR/SS domains or the ectodomain of ADAM28m (A28m 517-526) were subjected to immunoblotting with antibody 211-14 or anti-MBP antibody. The epitopes showing positive reaction with antibody 211-14 are highlighted in bold.

fragments of proADAM28s, i.e., Pro/MP and Dis/CR/SS proteins, Thr517-Thr526 peptide by Biacore allowed us to determine asso- 5 1 1 showed that antibody 211-14 recognizes the Dis/CR/SS protein ciation and dissociation constants of Ka ¼ 2.05 10 M * S 5 1 (Fig. 3C), indicating that the Dis/CR/SS domains contain the and Kd ¼ 1.94 10 S , corresponding to an equilibrium 11 epitope. The epitope mapping analysis using 11 different peptides dissociation constant of KD ¼ 9.47 10 mol/L (Supplemen- covering the sequences of the Dis, CR, and SS domains of tary Fig. S2). human ADAM28s demonstrated that the antibody recognizes the Pro399-Arg540, Gln491-Arg540, Thr511-Pro530, Thr517-Thr526, and Immunocytochemical localization of ADAM28 in human Leu519-Pro528 peptides, but not other six peptides (Fig. 3D). On ADAM28-transfected cells by antibody 211-14 the other hand, the antibody did not react with the Thr517-Val526 The expression vectors containing cDNA fragments encoding peptide speciﬁc to the ectodomain of membrane-type ADAM28m human proADAM28s or proADAM28m were transfected in CHO- (ref. 30; Fig. 3D). These data suggest that T517ELWGPGRRT526, K1 cells, and the cells were subjected to immunocytochemistry which corresponds to the junctional region of the CR domain and using antibody 211-14, anti-ADAM28 mouse monoclonal anti- the SS domain of human ADAM28s, is the epitope of antibody body (297-2F3) or nonimmune IgG. As shown in Fig. 4A, anti- 211-14 (Fig. 3D), and thus this antibody is considered to be body 211-14 showed immunostaining signals both in the cyto- speciﬁc to secreted-type ADAM28s. Kinetic analysis for the plasm and on the cell membrane of proADAM28s transfectants, interaction between antibody 211-14 and the MBP-fused but not proADAM28m or mock transfectants. Mouse monoclonal

2432 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Human Antibody Inhibitor to ADAM28

Figure 4. Immunocytochemical localization of ADAM28 in ADAM28-transfected CHO-K1 cells by antibody 211-14. CHO-K1 cells transfected with pIRES2-EGFP-ADAM28s (ADAM28s/EGFP), pcDNA3.1- ADAM28m/pIRES2-EGFP empty (ADAM28m/EGFP), or pIRES2-EGFP empty (Mock EGFP) vectors were immunostained with antibody 211-14 and then incubated with goat anti-human antibody conjugated to Alexa Flour 555 (red) (A). As for positive and negative controls, these transfectants were immunostained with mouse anti-ADAM28 monoclonal antibody (297-2F3) (B)and nonimmune human IgG (C), respectively. They were observed by a ﬂuorescence microscope after nuclear staining with Hoechst 33342 dye. Bars, 20 mm.

antibody (297-2F3) showed the similar staining pattern in both cells showed negligible expression of ADAM28s or ADAM28m proADAM28s and proADAM28m (Fig. 4B) and only background (Fig. 5A). Because PC-9 cells appeared to strongly express staining was observed with nonimmune IgG (Fig. 4C). These data ADAM28, we examined the effect of antibody 211-14 on cell confirm the specificity of antibody 211-14 to ADAM28s and proliferation of IGF-1–treated PC-9 cells. As shown in Fig. 5B, the suggest the possibility that ADAM28s has the anchoring system antibody substantially decreased cell proliferation, and the half- on the cell membrane. maximal effective concentration (EC50) was calculated to be 14.7 nmol/L from the titration curve. When the effect of 5 mg/ Inhibition of IGF-1–induced cancer cell proliferation and mL antibody 211-14, a concentration corresponding to EC70 for promotion of VWF-induced cell death by antibody 211-14 IGF-1–induced cell proliferation, on IGF-1–induced cell prolif- We examined the effect of antibody 211-14 on IGF-1–induced eration of PC-9, H1975, H1650, H441, and Calu-3 cells was cell proliferation in five lung adenocarcinoma cell lines (PC-9, tested, significant inhibition was observed in all the cell lines H1975, H1650, H441, and Calu-3), among which only H1975 except for ADAM28 expression–negative H1975 cells (Fig. 5C).

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2433

Mochizuki et al.

Figure 5. Effects of antibody 211-14 on IGF-1–induced proliferation and VWF-induced cell death in lung adenocarcinoma cells. A, mRNA expression of ADAM28s, ADAM28m, IGF1 receptor (IGF1-R), IGF-1, and IGFBP-3 by RT-PCR in the ﬁve adenocarcinoma cell lines. b-Actin was used as a loading control. B, Half-maximal effective concentration

(EC50) of antibody 211-14 on inhibition of cell proliferation in PC-9 cells. IGF-1–stimulated PC-9 cells were cultured for 48 hours in the presence of antibody 211-14 (0.17–170 nmol/L) or control nonimmune human IgG (0.17–170 nmol/L), and then proliferation activity was detected by BrdUrd incorporation ELISA. C, Inhibition of IGF-1–induced cell proliferation by antibody 211-14 in lung adenocarcinoma cell lines. The cells were incubated with IGF-1 (100 ng/mL) alone or IGF-1 (100 ng/mL) plus antibody 211-14 (5 mg/mL), followed by BrdUrd incorporation ELISA. , P < 0.05; , P < 0.001. D, Promotion of VWF-induced cell death by antibody 211-14 in lung adenocarcinoma cell lines. The cells were incubated with VWF (3 mg/mL) alone or VWF (3 mg/mL) plus antibody 211-14 (5 mg/mL), and then analyzed for cell death by DNA fragmentation ELISA. , P < 0.01; , P < 0.001.

On the other hand, cell death was induced by treatment with VWF compared with the control IgG-treated mice (P < 0.001 only in ADAM28-negative H1975 cells, and other cell lines were each; Fig. 6B). When the lung tissue was immunohistochemically resistant to VWF effect (Fig. 5D). However, VWF-induced cell examined, Ki-67–positive cell index of the carcinoma cells was death emerged in these cell lines by treatment with the antibody similar in the antibody-treated mice (89.43% 3.02%; ratio of (Fig. 5D). The data indicate that antibody 211-14 inhibits IGF-1– Ki-67–positive cells to total carcinoma cells) compared with induced cell proliferation and enhances VWF-induced cell death control IgG-treated mice (84.91% 5.34%; Supplementary by inhibiting ADAM28s activity. Fig. S3). On the other hand, immunoblotting analysis of the expression of cleaved caspase-3 in the lung tissues showed a Effects of antibody 211-14 on lung metastasis of PC-9 cells in an significant increase in antibody-treated mice (4.30 1.30) than early-stage metastasis model control IgG-treated mice (1.00 0.98; P < 0.001; Fig. 6C). Survival At one day after the injection of PC-9ffLuc-cp156 lung carcinoma time of the mice was significantly prolonged in the antibody- cells, almost all carcinoma cells were localized within blood treated mouse group compared with the control IgG-treated vessels in the lung (Fig. 6A). Monitoring by IVIS indicated that group (P ¼ 0.0064; log-rank test; Fig. 6D). Intriguingly, 3 of 10 lung metastasis is significantly reduced in the antibody-treated mice treated with the antibody were completely free from carci- mouse group at 3 and 4 weeks after the antibody injection noma cells and survived for more than 1 year (Fig. 6D), suggesting

2434 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Human Antibody Inhibitor to ADAM28

Figure 6. Inhibitory effects of antibody 211-14 on the lung metastasis in an early-stage metastasis model. A, Histological demonstration of PC-9ffLuc-cp156 cells within blood vessels at 1 day after the intravenous injection into tail vein of NOD/SCID mice. Serial paraffin sections of the lung tissue were subjected to H&E staining, immunostaining for Venus with anti-GFP antibody, and TSC-PAM staining. Boxed region corresponds to areas shown by immunostaining and TSC-PAM on serial sections. Bars, 50 mm. B, Effect of antibody 211-14 on lung metastasis in NOD/SCID mice. The mice were intraperitoneally injected with antibody 211-14 or control nonimmune IgG (Cont IgG) every other day five times at 1 day after intravenous injection of the PC-9ffLuc-cp156 cells, and metastasis was monitored by IVIS imaging system for 4 weeks. Bars, mean SD. , P < 0.001. C, Immunoblotting analysis of cleaved caspase-3 in the lung tissue at 1 day after the injection of antibody 211-14 or control IgG. Representative immunoblotting data (left) and densitometric analysis of relative expression levels of cleaved caspase-3 (right; n ¼ 3). b-Actin, loading control. , P < 0.001. D, Survival of mice treated with antibody 211-14 or control IgG (Cont IgG) by Kaplan–Meier analysis. Note that 3 of 10 mice that received antibody 211-14 injections are free from carcinoma cells at 28 weeks after the carcinoma cell injection. possible complete cure for lung metastasis by the early treatment administration of antibody 211-14 and docetaxel and that of with the antibody. bevacizumab and docetaxel on mouse survivals in the advanced- stage metastasis model, the mice harboring multiple lung metas- Effects of antibody 211-14 on metastases of PC-9 cells in an tases were divided to five groups (n ¼ 10 each) and treated with advanced-stage metastasis model docetaxel once and antibodies five times. As shown in Fig. 7E, To assess the effects of antibody 211-14 in an advanced-stage combined therapy of antibody 211-14 and docetaxel showed metastasis model, we injected PC-9ffLuc-cp156 cells into tail veins of more than 2.3-fold prolongation of median and 20% survivals NOD/SCID mice and left them for 3 weeks. Multiple lung metas- compared with control IgG treatment (P ¼ 0.0001) or bevacizu- tases were histologically identified (Fig. 7A). The mice randomly mab monotherapy (P ¼ 0.0001; log-rank test). In addition, divided into two groups at 3 weeks after confirming the lung median and 20% survivals in the mice received concurrent metastases by IVIS were treated with antibody 211-14 or control administration of antibody 211-14 and docetaxel were signifi- IgG by intraperitoneal injections five times every week, and tumor cantly longer than those in mouse group treated with docetaxel growth and metastases were monitored by IVIS. As shown alone (P ¼ 0.0473) or bevacizumab and docetaxel (P ¼ 0.0232; in Fig. 7B and C, tumor growth and metastases were significantly log-rank test; Fig. 7E). suppressed at 8 and 10 weeks in the antibody-treated group compared with the control group (P < 0.001). The median and Toxicity study of antibody 211-12 administration in mice 20% survivals of the antibody-treated mouse group were approx- We lastly carried out toxicity study by intraperitoneal injec- imately 1.6-fold longer than the control IgG-treated group (P ¼ tions of Fc-mousenized antibody 211-12 or control mouse IgG 0.0145; log-rank test; Fig. 7D). To examine the effect of concurrent to normal male mice at a dose of 20 mg/kg body weight five

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2435

Mochizuki et al.

Figure 7. Inhibitory effects of antibody 211-14 on proliferation and metastasis of PC-9ffLuc-cp156 cells and survival analyses in an advanced-stage metastasis model. A, Histological demonstration of multiple metastases in the lung tissue. The lung samples were obtained at 3 weeks after intravenous injection of PC-9ffLuc-cp156 cells. Serial paraffin sections were subjected to H&E staining, immunostaining with anti-GFP antibody, or TSC-PAM staining. Boxed region on the left is magnified to right, and boxed region on the right corresponds to areas shown by immunostaining and TSC-PAM on serial sections. Dotted circles, metastatic foci. Bars, 1 mm (left) and 50 mm (right). B, Representative bioluminescence images of the mice before (3 weeks) and after antibody treatment (10 weeks). Antibody 211-14 or control nonimmune IgG (Cont IgG) was intraperitoneally injected five times every week by starting at 3 weeks after PC-9ffLuc-cp156 cell injection. C, Effect of antibody 211-14 on cell growth and metastases in the mice by IVIS. Bars, mean SD. , P < 0.001. D, Survival of mice treated with antibody 211-14 or Cont IgG. E, Survival of mice treated by combined therapy. Mice harboring multiple lung metastases were treated with antibody 211-14 plus docetaxel (Ab 211-14 þ DT), bevacizumab plus docetaxel (Beva þ DT), bevacizumab alone (Beva), docetaxel alone (DT), or control IgG alone (Cont IgG).

times once a week. Administration of the antibody and control Discussion IgG was well tolerated, and no treatment-related effects were noted on a comprehensive list of clinical observations, food Our characterization studies have indicated that antibody consumption, and body weights (Supplementary Fig. S4A). At 211-14 is immunoreactive with human ADAM28s without 1 week after the ﬁfth administration, they were subjected to cross-reactivity with main members of the ADAM, ADAMTS, autopsy. There was no difference in weights of major organs and MMP gene families. The antibody recognized the sequence between the two groups (Supplementary Fig. S4B). Blood corresponding to the junctional region between the CR and SS samples obtained showed no abnormalities for clinical hema- domains, and the epitope showed no sequence homology to tology or clinical chemistry parameters indicative of renal, other proteins according to the UniProt database (http://web. hepatic, or cardiac effects (Supplementary Table S2). Only the expasy.org/docs/swiss-prot_guideline.html). The CR and/or SS pathologic change found was spermatogenic granulomas in the domains are considered to interact with ADAM28s substrates epididymis in 1 of the 10 mice received the antibody (Supple- such as VWF (14). Thus, all these data demonstrate that this mentary Fig. S4C). antibody is speciﬁc to human ADAM28s and suggest that the

2436 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Human Antibody Inhibitor to ADAM28

antibody acts as an exosite inhibitor and blocks the ADAM28s independent on the mutation of EGFR in the carcinoma tissues activity by preventing the enzyme from interaction with the (M. Shimoda and colleagues, unpublished observations). There- substrates. fore, antibody 211-14 may be a novel therapeutic option in most ADAM28s promotes carcinoma cell proliferation through reac- NSCLC patients regardless of driver gene mutations and have a tivation of IGF-1 by digestion of IGFBP-3 in the IGF-1/IGFBP-3 clinical impact by the combination therapy with chemotherapeu- complex (13) and prevents VWF-induced apoptosis of carcinoma tic drugs such as docetaxel. cells in blood stream (14), leading to enhanced carcinoma cell In the pilot toxicity study, we have demonstrated that mouse- growth and progression. In this study, we have demonstrated that nized antibody 211-12 that cross-reacts with mouse ADAM28 the antibody affects the ADAM28s-expressing lung carcinoma shows no or negligible adverse effects on the mice. However, cells by suppressing IGF-1–induced cell proliferation and pro- because this study was carried out without information about moting VWF-induced cell death probably through acting as an affinity of antibody 211-12 to mouse ADAM28, we recently antibody inhibitor to ADAM28s proteinase activity. The cell- performed toxicity study of antibody 211-14 using macaques surface localization of ADAM28s shown by immunocytochem- and found no or negligible changes in the blood tests or pathology istry using this antibody may confer the effective inhibitor func- at autopsy (M. Shimoda and colleagues, unpublished observations of the antibody by facilitating to act on the cell membrane as tions). Our recent study on the expression of ADAM28 in human well as in the extracellular milieu. normal organs has indicated that ADAM28 is expressed, albeit at In our early-stage metastasis model, antibody 211-14 remark- lower levels compared with NSCLC tissue, by the epithelial cells in ably suppressed the PC-9 carcinoma cell metastasis and growth in several organs, including the epididymis, bronchus, and stomach the lung, leading to extension of survival time of the mice. The (29). Because the epithelial cells and blood vessels are surrounded increased expression of cleaved caspase-3 in the lung tissue from by basement membrane with a firm structure (34), it seems the antibody-treated mice, albeit no changes in MIB-1–positive unlikely that our antibody against ADAM28 readily permeates cell index of the carcinoma cells, compared with control IgG- from blood vessels and reaches the ADAM28-expressing epithelial treated mice suggests that inhibition of ADAM28 activity by the cells in the normal organs. antibody may contribute to blockade of the metastatic process by The present study has potential limitations. Because our metas- interfering with carcinoma cell survival within blood stream tasis models were made by intravenously injecting PC-9 cells in through acceleration of VWF-induced apoptosis as we previously NOD/SCID mice, they do not reflect the spontaneous metastasis proposed (14). Intriguingly, the antibody treatment cured metas- initiated by intravasation of invading carcinoma cells from the tasis in 30% of the experimental mice, enabling them to accom- primary tumors and immune responses cannot be evaluated in plish their life free from cancer cells. Recent report on a phase II our models. Nevertheless, monitoring tumors by IVIS in the dose-escalation trial of desmopressin in breast cancer patients has advanced-stage metastasis model showed reduction in distant demonstrated that administration of desmopressin, which stimu- organ metastases by treatment with the antibody, suggesting lates the release of VWF from endothelial cells (31), reduces the suppressive effect of the antibody on the metastasis from the circulating cancer cells by increasing blood levels of VWF (32). lung tumors. Based on the in vitro data, we speculated the Altogether, it is plausible to think that our antibody is useful for possible involvement of inhibition of IGF-1–mediated cell prevention of micrometastases by extinguishing circulating cancer growth and promotion of VWF-induced apoptosis by the anti- cells, which are commonly detected in cancer patients, including body in the metastasis models. However, the molecular mecha- NSCLC patients (33). nism of the antibody effects in vivo remains to be determined Clinical trials of investigational anticancer agents are usually by further studies, which may include analysis of tumor cell conducted in patients with advanced-stage cancer. Therefore, growth and metastasis in patient-derived tumor xenografts in we examined the antitumor effects of antibody 211-14 in the recently developed NSG (NOD.Cg-PrkdcSCIDIl2rgtm1Wjil/SzJ) advanced-stage metastasis model, which may represent the mice. Although the exact molecular mechanism should be clar- patients with stage IV lung cancer or advanced lung cancer. Our ified prior to clinical trials, antibody 211-14 targeting ADAM28s data indicated that the antibody effectively attenuates tumor might be a possible remedy for prolongation of survival of growth and distant metastases from the lungs, resulting in sig- patients with NSCLC. nificant elongation of the survival. Because most patients with advanced-stage NSCLC harboring neither EGFR mutations nor Disclosure of Potential Conflicts of Interest ALK rearrangements are treated with chemotherapeutic drugs, we S. Mochizuki has ownership interest (including stock, patents, etc.) in a tested the effect of antibody 211-14 combined with docetaxel and patent. M. Shimoda has ownership interest (including stock, patents, etc.) in a found that the combination therapy significantly elongates the patent. A. Miyakoshi is a researcher at GeneFrontier Corporation. Y. Okada has ownership interest (including stock, patents, etc.) in a patent. No potential survival time compared with the concurrent therapy of bevaci- conflicts of interest were disclosed by the other authors. zumab and docetaxel as well as monotherapies. Therapies targeting EGFR mutations or ALK rearrangements Authors' Contributions have been successful in the patients with NSCLC harboring these Conception and design: S. Mochizuki, N. Aramaki-Hattori, Y. Okada driver gene mutations (19–21). However, because the prevalence Development of methodology: M.Shimoda,N.Aramaki-Hattori, of the gene alterations is confined to less than 25% of total NSCLC A. Miyakoshi, K. Kojoh, Y. Okada patients (20, 21), and drug resistance develops for the majority of Acquisition of data (provided animals, acquired and managed patients within one year of commencing treatment (18, 21), the patients, provided facilities, etc.): S. Mochizuki, M. Shimoda, H. Abe, N. Aramaki-Hattori, H. Ueno, A. Miyakoshi development of new therapies other than the target therapy to Analysis and interpretation of data (e.g., statistical analysis, biostatistics, driver genes is required. Approximately 70% of NSCLC patients computational analysis): M. Shimoda, H. Abe, K. Ishii, Y. Okada show moderate- or high-level expression of ADAM28 (9, 11) and Writing, review, and/or revision of the manuscript: S. Mochizuki, M. Shimoda, our preliminary study shows that the ADAM28 expression is A. Miyakoshi, Y. Okada

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2437

Mochizuki et al.

Administrative, technical, or material support (i.e., reporting or organizing Institute, Kanazawa University (to S. Mochizuki and M. Shimoda), and Adapt- data, constructing databases): S. Mochizuki, H. Abe, Y. Miyamae, J. Kuramoto, able and Seamless Technology Transfer Program through Target-driven R&D of K. Ishii, H. Ueno, Y. Okada JST (A-STEP) AS2614150Q (to S. Mochizuki, M. Shimoda and Y. Okada). Study supervision: H. Ueno, Y. Okada The costs of publication of this article were defrayed in part by the Acknowledgments payment of page charges. This article must therefore be hereby marked We thank Ms. Chiho Yoshinaga of the Juntendo University for her technical advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate assistance and Dr. Toyoshirou Ito of the Central Institute for Experimental this fact. Animals for giving us marmoset tissues. This work was supported by JSPS KAKENHI grant number 15K08409 Received December 1, 2017; revised April 13, 2018; accepted August 29, 2018; (to S. Mochizuki), External Collaborative Research Grant of Cancer Research published ﬁrst September 6, 2018.

References 1. Edwards DR, Handsley MM, Pennington CJ. The ADAM metalloprotei- 19. Kris MG, Johnson BE, Berry LD, Kwiatkowski DJ, Iafrate AJ, Wistuba II, et al. nases. Mol Aspects Med 2008;29:258–89. Using multiplexed assays of oncogenic drivers in lung cancers to select 2. Murphy G. Regulation of the proteolytic disintegrin metalloproteinases, targeted drugs. JAMA 2014;311:1998–2006. the 'Sheddases'. Semin Cell Dev Biol 2009;20:138–45. 20. Barlesi F, Mazieres J, Merlio JP, Debieuvre D, Mosser J, Lena H, et al. Routine 3. Seals DF, Courtneidge SA. The ADAMs family of metalloproteases: molecular profiling of patients with advanced non-small-cell lung cancer: multidomain proteins with multiple functions. Genes Dev 2003;17:7–30. results of a 1-year nationwide programme of the French Cooperative 4. Blobel CP. ADAMs: key components in EGFR signalling and development. Thoracic Intergroup (IFCT). Lancet 2016;387:1415–26. Nat Rev Mol Cell Biol 2005;6:32–43. 21. Forde PM, Ettinger DS. Targeted therapy for non-small-cell lung cancer: 5. Mochizuki S, Okada Y. ADAMs in cancer cell proliferation and progression. past, present and future. Expert Rev Anticancer Ther 2013;13: Cancer Sci 2007;98:621–8. 745–58. 6. Fry JL, Toker A. Secreted and membrane-bound isoforms of protease 22. Brahmer J, Reckamp KL, Baas P, Crino L, Eberhardt WE, Poddubskaya E, ADAM9 have opposing effects on breast cancer cell migration. Cancer Res et al. Nivolumab versus docetaxel in advanced squamous-cell non-small- 2010;70:8187–98. cell lung cancer. N Engl J Med 2015;373:123–35. 7. Okada A, Mochizuki S, Yatabe T, Kimura T, Shiomi T, Fujita Y, et al. 23. Garon EB, Rizvi NA, Hui R, Leighl N, Balmanoukian AS, Eder JP, et al. ADAM-12 (meltrin alpha) is involved in chondrocyte proliferation via Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J cleavage of insulin-like growth factor binding protein 5 in osteoarthritic Med 2015;372:2018–28. cartilage. Arthritis Rheum 2008;58:778–89. 24. Restifo NP, Smyth MJ, Snyder A. Acquired resistance to immunotherapy 8. Mochizuki S, Okada Y. ADAM28 as a target for human cancers. Curr Pharm and future challenges. Nat Rev Cancer 2016;16:121–6. Des 2009;15:2349–58. 25. Shimoda M, Hashimoto G, Mochizuki S, Ikeda E, Nagai N, Ishida S, 9. Ohtsuka T, Shiomi T, Shimoda M, Kodama T, Amour A, Murphy G, et al. et al. Binding of ADAM28 to P-selectin glycoprotein ligand-1 enhances ADAM28 is overexpressed in human non-small cell lung carcinomas and P-selectin-mediated leukocyte adhesion to endothelial cells. J Biol correlates with cell proliferation and lymph node metastasis. Int J Cancer Chem 2007;282:25864–74. 2006;118:263–73. 26. Inoki I, Shiomi T, Hashimoto G, Enomoto H, Nakamura H, Makino K, 10. Mitsui Y, Mochizuki S, Kodama T, Shimoda M, Ohtsuka T, Shiomi T, et al. et al. Connective tissue growth factor binds vascular endothelial growth ADAM28 is overexpressed in human breast carcinomas: implications for factor (VEGF) and inhibits VEGF-induced angiogenesis. FASEB J carcinoma cell proliferation through cleavage of insulin-like growth factor 2002;16:219–21. binding protein-3. Cancer Res 2006;66:9913–20. 27. Howard L, Maciewicz RA, Blobel CP. Cloning and characterization 11. Kuroda H, Mochizuki S, Shimoda M, Chijiiwa M, Kamiya K, Izumi Y, et al. of ADAM28: evidence for autocatalytic pro-domain removal and ADAM28 is a serological and histochemical marker for non-small-cell lung for cell surface localization of mature ADAM28. Biochem J 2000; cancers. Int J Cancer 2010;127:1844–56. 348:21–7. 12. Wright CM, Larsen JE, Hayward NK, Martins MU, Tan ME, Davidson MR, 28. Jury JA, Perry AC, Hall L. Identification, sequence analysis and expression of et al. ADAM28: a potential oncogene involved in asbestos-related lung transcripts encoding a putative metalloproteinase, eMDC II, in human and adenocarcinomas. Genes Chromosomes Cancer 2010;49:688–98. macaque epididymis. Mol Hum Reprod 1999;5:1127–34. 13. Mochizuki S, Shimoda M, Shiomi T, Fujii Y, Okada Y. ADAM28 is activated 29. Miyamae Y, Mochizuki S, Shimoda M, Ohara K, Abe H, Yamashita S, by MMP-7 (matrilysin-1) and cleaves insulin-like growth factor binding et al. ADAM28 is expressed by epithelial cells in human normal protein-3. Biochem Biophys Res Commun 2004;315:79–84. tissues and protects from C1q-induced cell death. FEBS J 2016;283: 14. Mochizuki S, Soejima K, Shimoda M, Abe H, Sasaki A, Okano HJ, et al. 1574–94. Effect of ADAM28 on carcinoma cell metastasis by cleavage of von 30. Roberts CM, Tani PH, Bridges LC, Laszik Z, Bowditch RD. MDC-L, a novel Willebrand factor. J Natl Cancer Inst 2012;104:906–22. metalloprotease disintegrin cysteine-rich protein family member expressed 15. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin by human lymphocytes. J Biol Chem 1999;274:29251–9. 2012;62:10–29. 31. Kaufmann JE, Vischer UM. Cellular mechanisms of the hemostatic effects 16. Besse B, Adjei A, Baas P, Meldgaard P, Nicolson M, Paz-Ares L, et al. 2nd of desmopressin (DDAVP). J Thromb Haemost 2003;1:682–9. ESMO consensus conference on lung cancer: non-small-cell lung cancer 32. Weinberg RS, Grecco MO, Ferro GS, Seigelshifer DJ, Perroni NV, Terrier FJ, first-line/second and further lines of treatment in advanced disease. et al. A phase II dose-escalation trial of perioperative desmopressin Ann Oncol 2014;25:1475–84. (1-desamino-8-d-arginine vasopressin) in breast cancer patients. Spring- 17. Khuri FR, Herbst RS, Fossella FV. Emerging therapies in non-small-cell lung erplus 2015;4:428. cancer. Ann Oncol 2001;12:739–44. 33. Hamilton G, Rath B. Detection of circulating tumor cells in non-small cell 18. Hirsch FR, Suda K, Wiens J, Bunn PA Jr.New and emerging targeted lung cancer. J Thorac Dis 2016;8:1024–8. treatments in advanced non-small-cell lung cancer. Lancet 2016;388: 34. LeBleu VS, Macdonald B, Kalluri R. Structure and function of basement 1012–24. membranes. Exp Biol Med (Maywood) 2007;232:1121–9.

2438 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Selective Inhibition of ADAM28 Suppresses Lung Carcinoma Cell Growth and Metastasis

Satsuki Mochizuki, Masayuki Shimoda, Hitoshi Abe, et al.

Mol Cancer Ther 2018;17:2427-2438. Published OnlineFirst September 6, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-17-1198

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2018/09/06/1535-7163.MCT-17-1198.DC1

Cited articles This article cites 34 articles, 6 of which you can access for free at: http://mct.aacrjournals.org/content/17/11/2427.full#ref-list-1

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://mct.aacrjournals.org/content/17/11/2427. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.