inhibition of discoidin domain receptor 1 prevents stroma ... · tumor microenvironment inhibition...

12
Tumor Microenvironment Inhibition of Discoidin Domain Receptor 1 Prevents Stroma-Induced Peritoneal Metastasis in Gastric Carcinoma Hyejin Jin 1 , In-Hye Ham 1,2 , Hye Jeong Oh 1 , Cheong A Bae 1,2 , Dakeun Lee 3 , Young-Bae Kim 3 , Sang-Yong Son 1 , Yong-Joon Chwae 4 , Sang-Uk Han 1 , Rolf A. Brekken 5 , and Hoon Hur 1,2 Abstract Discoidin domain receptor 1 (DDR1) is activated by brillar (triple-helical) collagens and collagen IV, which are major components of tumor stroma; thus, DDR1 might be a critical mediator of communication between cancer cells and stroma. The aim of this study was to investigate the effect of DDR1 inhibition on stroma-induced perito- neal metastasis in gastric carcinoma. We analyzed by immunohistochemistry the correlation between DDR1 expression and the pattern of recurrence in gastric carci- noma tissues from a previously characterized and estab- lished gastric carcinoma patient cohort. We also cocultured human gastric carcinoma cell lines with gastric cancerassociated broblasts (CAF) and investigated DDR1 expression and activation. We evaluated CAF-induced tumorigenic properties of gastric carcinoma cell lines and the effect of a DDR1-specic inhibitor in organotypic cultures and in a peritoneal seeding xenograft animal model. The expression of DDR1 in gastric cancer tissues was positively associated with early recurrence (P ¼ 0.043) and a high incidence of peritoneal recurrence (P ¼ 0.036). We conrmed that coculturing with CAFs elevated DDR1 protein expression in gastric carcinoma cell lines and enhanced gastric carcinoma cell line spheroid formation in organotypic cultures in a tumor cell DDR1-dependent manner. Coimplantation of CAFs with gastric carcinoma cells enhanced peritoneal tumor formation in vivo, an effect that was sensitive to pharmacologic inhibition of DDR1. Implications: This study highlights that CAF-induced ele- vation of DDR1 expression in gastric carcinoma cells enhances peritoneal tumorigenesis, and that inhibition of DDR1 is an attractive strategy for the treatment of gastric carcinoma peritoneal metastasis. Mol Cancer Res; 16(10); 1590600. Ó2018 AACR. Introduction Gastric carcinoma is one of the most common malignant tumors and the third leading cause of cancer-related deaths worldwide (1). Although mortality from gastric carcinoma has gradually decreased in recent years, patients with late- stage disease still have poor prognosis due to tumor non- resectability or recurrence after resection. Large-scale clinical trials showed that adjuvant chemotherapy following curative resection could reduce the recurrence rate of stage II or III gastric carcinoma. However, approximately one fourth of all patients enrolled in those trials suffered recurrence during the follow-up period, and the peritoneum is one of the most common sites of recurrence (2, 3). To date, various studies have reported that patients with peritoneal metastasis have worse response to treatments than patients with hematoge- nous or lymphatic metastasis (46). However, specic modalities to prevent gastric carcinoma peritoneal metastases have not been established. Thus, new strategies based on specic molecular biomarkers related to peritoneal metas- tasis must be developed to prevent gastric carcinoma perito- neal recurrence. Several histologic features of primary gastric carcinomas have implicated the tumor stroma in the peritoneal metastasis of gastric carcinoma. First, it is well known that the likelihood of peritoneal recurrence increases when primary gastric carcino- mas invade the gastric wall deeply (6). During invasion, cancer cells that originate from the mucosal layer of the gastrointes- tinal tract are exposed to stromal tissues and activate various noncancerous cells, such as broblasts and inammatory cells (7). Second, the subtypes of gastric carcinomas with high proportions of tumor stroma preferentially metastasize to the peritoneum (6). Moreover, several studies have reported that cancer-associated broblasts (CAF) contribute to the peritoneal dissemination of gastric carcinoma (8, 9). Taken together, these ndings indicate that the interaction between cancer cells and stromal cells could be a therapeutic target to prevent peritoneal metastasis of gastric carcinoma. However, drugs targeting the 1 Department of Surgery, Ajou University School of Medicine, Suwon, Korea. 2 Brain Korea 21 Plus Research Center for Biomedical Sciences, Ajou University, Suwon, Korea. 3 Department of Pathology, Ajou University School of Medicine, Suwon, Korea. 4 Department of Microbiology, Ajou University School of Medicine, Suwon, Korea. 5 Division of Surgical Oncology, Department of Surgery, Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, Texas. Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/). Corresponding Author: Hoon Hur, Department of Surgery, Ajou University School of Medicine, 164 Worldcup-ro, Yeongtong-gu, Suwon-si, Gyeonggi-do 16499 Republic of Korea. Phone: 82-31-219-5200; Fax: 82-31-219-5575; E-mail: [email protected], [email protected] doi: 10.1158/1541-7786.MCR-17-0710 Ó2018 American Association for Cancer Research. Molecular Cancer Research Mol Cancer Res; 16(10) October 2018 1590 on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

Upload: others

Post on 14-Jul-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

  • Tumor Microenvironment

    Inhibition of Discoidin Domain Receptor 1Prevents Stroma-Induced PeritonealMetastasis in Gastric CarcinomaHyejin Jin1, In-Hye Ham1,2, Hye Jeong Oh1, Cheong A Bae1,2, Dakeun Lee3,Young-Bae Kim3, Sang-Yong Son1, Yong-Joon Chwae4, Sang-Uk Han1,Rolf A. Brekken5, and Hoon Hur1,2

    Abstract

    Discoidin domain receptor 1 (DDR1) is activated byfibrillar (triple-helical) collagens and collagen IV, whichare major components of tumor stroma; thus, DDR1 mightbe a critical mediator of communication between cancercells and stroma. The aim of this study was to investigatethe effect of DDR1 inhibition on stroma-induced perito-neal metastasis in gastric carcinoma. We analyzed byimmunohistochemistry the correlation between DDR1expression and the pattern of recurrence in gastric carci-noma tissues from a previously characterized and estab-lished gastric carcinoma patient cohort. We also coculturedhuman gastric carcinoma cell lines with gastric cancer–associated fibroblasts (CAF) and investigated DDR1expression and activation. We evaluated CAF-inducedtumorigenic properties of gastric carcinoma cell lines andthe effect of a DDR1-specific inhibitor in organotypiccultures and in a peritoneal seeding xenograft animal

    model. The expression of DDR1 in gastric cancer tissueswas positively associated with early recurrence (P ¼ 0.043)and a high incidence of peritoneal recurrence (P ¼ 0.036).We confirmed that coculturing with CAFs elevated DDR1protein expression in gastric carcinoma cell lines andenhanced gastric carcinoma cell line spheroid formationin organotypic cultures in a tumor cell DDR1-dependentmanner. Coimplantation of CAFs with gastric carcinomacells enhanced peritoneal tumor formation in vivo, an effectthat was sensitive to pharmacologic inhibition of DDR1.

    Implications: This study highlights that CAF-induced ele-vation of DDR1 expression in gastric carcinoma cellsenhances peritoneal tumorigenesis, and that inhibitionof DDR1 is an attractive strategy for the treatment ofgastric carcinoma peritoneal metastasis. Mol Cancer Res; 16(10);1590–600. �2018 AACR.

    IntroductionGastric carcinoma is one of the most common malignant

    tumors and the third leading cause of cancer-related deathsworldwide (1). Although mortality from gastric carcinomahas gradually decreased in recent years, patients with late-stage disease still have poor prognosis due to tumor non-resectability or recurrence after resection. Large-scale clinicaltrials showed that adjuvant chemotherapy following curativeresection could reduce the recurrence rate of stage II or IIIgastric carcinoma. However, approximately one fourth of all

    patients enrolled in those trials suffered recurrence during thefollow-up period, and the peritoneum is one of the mostcommon sites of recurrence (2, 3). To date, various studieshave reported that patients with peritoneal metastasis haveworse response to treatments than patients with hematoge-nous or lymphatic metastasis (4–6). However, specificmodalities to prevent gastric carcinoma peritoneal metastaseshave not been established. Thus, new strategies basedon specific molecular biomarkers related to peritoneal metas-tasis must be developed to prevent gastric carcinoma perito-neal recurrence.

    Several histologic features of primary gastric carcinomas haveimplicated the tumor stroma in the peritoneal metastasis ofgastric carcinoma. First, it is well known that the likelihood ofperitoneal recurrence increases when primary gastric carcino-mas invade the gastric wall deeply (6). During invasion, cancercells that originate from the mucosal layer of the gastrointes-tinal tract are exposed to stromal tissues and activate variousnoncancerous cells, such as fibroblasts and inflammatory cells(7). Second, the subtypes of gastric carcinomas with highproportions of tumor stroma preferentially metastasize to theperitoneum (6). Moreover, several studies have reported thatcancer-associated fibroblasts (CAF) contribute to the peritonealdissemination of gastric carcinoma (8, 9). Taken together, thesefindings indicate that the interaction between cancer cells andstromal cells could be a therapeutic target to prevent peritonealmetastasis of gastric carcinoma. However, drugs targeting the

    1Department of Surgery, Ajou University School of Medicine, Suwon, Korea.2Brain Korea 21 Plus Research Center for Biomedical Sciences, Ajou University,Suwon, Korea. 3Department of Pathology, Ajou University School of Medicine,Suwon, Korea. 4Department of Microbiology, Ajou University School ofMedicine, Suwon, Korea. 5Division of Surgical Oncology, Department of Surgery,Hamon Center for Therapeutic Oncology Research, University of TexasSouthwestern Medical Center, Dallas, Texas.

    Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

    Corresponding Author: Hoon Hur, Department of Surgery, Ajou UniversitySchool of Medicine, 164 Worldcup-ro, Yeongtong-gu, Suwon-si, Gyeonggi-do16499 Republic of Korea. Phone: 82-31-219-5200; Fax: 82-31-219-5575; E-mail:[email protected], [email protected]

    doi: 10.1158/1541-7786.MCR-17-0710

    �2018 American Association for Cancer Research.

    MolecularCancerResearch

    Mol Cancer Res; 16(10) October 20181590

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://crossmark.crossref.org/dialog/?doi=10.1158/1541-7786.MCR-17-0710&domain=pdf&date_stamp=2018-9-28http://mcr.aacrjournals.org/

  • tumor stroma have not been applied to treat gastric carcinomain a clinical setting.

    Discoidin domain receptor 1 (DDR1) is a member of thetransmembrane receptor tyrosine kinase family with a discoi-din homology domain in its extracellular region (10–12).Distinct from other receptor tyrosine kinases, DDR1 is activatedby collagens, including types I, II, III, IV, V, VIII, and XI, whichare major components of the extracellular matrix (ECM) insolid tumors. Various studies have reported that increasedDDR1 expression is associated with a poor prognosis in malig-nant tumors (13–16). Moreover, our previous report describedthat approximately half of the gastric carcinomas exhibiteda positive expression of DDR1, which was related to the depthof gastric carcinoma invasion. In addition, gastric carcinomapatients with DDR1 expression showed worse overall survivalrates than gastric carcinoma patients without DDR1 expres-sion. We also demonstrated that type I collagen could activateDDR1 signaling and enhance the aggressive phenotype ofgastric carcinoma cells (17). However, the significance ofDDR1 expression on the pattern of recurrence after resectionof primary gastric carcinoma was not evaluated. If the tumorstroma contributes to the peritoneal recurrence of gastric car-cinomas, DDR1 could be a key regulator that links peritonealmetastasis and tumor stroma in gastric carcinoma.

    Here, we found that DDR1 expression correlated with gastriccarcinoma peritoneal recurrence. The aim of this study was toevaluate if tumor stromal cells enhance the peritoneal metastasisof gastric carcinoma through the upregulation of DDR1, and ifinhibition of DDR1 expression and activity could be appliedtherapeutically to prevent peritoneal recurrence.

    Materials and MethodsHuman samples and immunohistochemical staining

    We analyzed the features of gastric carcinoma recurrence in202 patients who had undergone curative gastric resectionswith proper lymphadenectomy at the Department of Surgery,Ajou University Hospital, Korea, from May 2003 to December2005. We examined the incidence of tumor recurrence and itsfeatures in that cohort during the follow-up period. The sites ofrecurrence were classified into three groups: the peritoneum,hematogenous sites, and distant lymph nodes. The definitionof DDR1 positivity was described in our previous report (17).We analyzed the association of DDR1 expression in gastriccarcinomas with recurrence-free survival rates and the recur-rence rates at the specific sites.

    Cell lines and chemicalsWe purchased gastric carcinoma cell lines MKN74 (KCLB

    No. 80104) and MKN45 (KCLB No. 80103) from the KoreanCell Line Bank. These cells were cultured in RPMI-1640 medi-um (Invitrogen) supplemented with 10% fetal bovine serum(FBS; Equitech-Bio), 1% penicillin/streptomycin (Invitrogen),and 1% amphotericin B (Sigma-Aldrich). The cells were incu-bated at 37�C in a humidified atmosphere containing 5% CO2.

    Fibroblasts were isolated from fresh gastric carcinoma andpaired normal gastric specimens, as described in our previousstudy (18). In the present study, we used fibroblasts isolated fromsurgical specimens of a gastric carcinomapatient diagnosedwith a7.5-cm-sized moderately differentiated tumor adenocarcinoma.Stomach cancer specimens were obtained from patients under-

    going surgery at Ajou University Hospital (Suwon, Korea) whosepreoperative pathologic diagnosis was gastric carcinoma. Thisstudy was conducted in accordance with the ethics code of theWorld Medical Association (Declaration of Helsinki) and wasapproved by the Institutional Review Board of Ajou UniversityHospital (AJIRB-BMR-SMP-14-155).

    For cocultures, fibroblasts were seeded into the upper cham-bers of 6-well transwells, and gastric carcinoma cell lines wereseeded in 6-well tissue culture dishes.

    CRISPR/Cas9-mediated knockout (KO) of DDR1We established DDR1-deficient gastric carcinoma lines using

    the CRISPR-Cas9 system in order to investigate the role ofDDR1 in interactions between CAFs and gastric carcinoma celllines (Supplementary Fig. S1A). The target primer sequences forhuman DDR1 were #1 (forward, 50-CACCGGTGGAATGTCG-CTTCCGGCG-30; reverse, 50-AAACCGCCGGAAGCGACATTC-CACC-30) or #2 (forward: 50-CACCGCCCCCTAGGTTGTGGC-GCAT-30; reverse, 50-AAACATGCGCCACAACCTAGGGGGC-30). Oligonucleotides with BsmB1 restriction sites for guideRNAs were synthesized at Xenotech, then phosphorylated usingT4 kinase (NEB). The phosphorylated oligonucleotides werecloned into LentiCRISPR v2 and the sequences of the clonedplasmids that were extracted from several selected colonieswere confirmed at Cosmo Genetech.

    MKN74 cells were transfected with pLentiCRISPR-sgRNADDR1 using Lipofectamine 2000 (Invitrogen), according to themanufacturer's instructions. When the cells were confluent indishes, they were treated with 1 to 2 mg/mL puromycin forapproximately 2 weeks. The silencing of DDR1 in the survivingcells after puromycin selection was validated byWestern blotting.

    Western blot analysisThe cells were washed with phosphate buffered saline and

    lysed in phospho-specific lysis buffer. Lysates were incubatedon ice for 20 minutes, then centrifuged at 13,000 � g for20 minutes at 4�C. Protein concentrations were determinedby Bradford assay (Bio-Rad). Equal amounts of protein fromeach sample were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto a poly-vinylidene difluoride membrane (Millipore. The immunoblotswere blocked by incubation in 5% skim milk, 250 mmol/L tris(hydroxymethyl) aminomethane–HCl (pH 8.0), 1500 mmol/LNaCl, and 0.1% Tween 20 for 1 hour at 25�C. Membraneswere incubated with the following primary antibodies: DDR1(1:1,000 dilution, #5583, Cell Signaling Technology), phos-phorylated-DDR1 (1:1,000 dilution, #11994, Cell SignalingTechnology), and b-actin (1:5,000 dilution, sc-47778, SantaCruz Biotechnology), followed by the corresponding HRP-conjugated secondary antibodies (Jackson ImmunoResearchLabs). Protein detection was performed with an enhancedchemiluminescence kit (Abclon).

    Cell viability testCells were seeded at 1� 104 cells per well in 96-well plates, and

    cell viability was measured using the highly sensitive water-soluble tetrazolium salt (WST)-based cell viability, cytotoxicityassay kit (DoGen). The EZ-Cytox solution (10 mL) was added toeach well, cells were incubated for 1.5 hours, and the absorbancewas measured by spectrophotometry at 450 nm. To evaluatethe effect of 7rh benzamide (provided by Dr. Ke Ding, Jinan

    DDR1 in Peritoneal Metastasis of Gastric Carcinoma

    www.aacrjournals.org Mol Cancer Res; 16(10) October 2018 1591

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • University) on cell viability, the cells were treated with differentconcentrations of 7rh benzamide over 72 hours.

    Reverse transcription PCRTotal RNA was prepared from gastric cancer cell lines and

    fibroblasts and converted to cDNA using 1 mg of each RNA in afinal volume of 20 mL. The cDNA mixture (3 mL) was used for 35cycles of PCR amplification (denatured at 94�C for 5 minutes,annealed at 62�C (DDR1), or 64�C (COL1A) for 30 seconds, andextended at 72�C for 5 minutes). The primers were as follows:DDR1 (forward, 50-CGAGCAGGTCA TCGAGAACG-30; reverse,50-CCCACCTGCAGTCTCACTG-30), COL1A (forward, 50-GTCA-CCCACCGACCAAGAAACC-30; reverse, 50-AAGTCCAGGCTGTC-CAGGGA TG-30), b-actin (ACTB) (forward, 50-TCAAGATCATT-GCTCCTCCTGAGC-30; reverse, 50-TGCTGTCACCTTCACCGTTC-CAGT-30).

    Organotypic culture and immunocytochemical stainingFor organotypic culture experiments, the artificial ECM con-

    sisted of a mixture of Matrigel (5 mg/mL, BD Biosciences) andcollagen I (1.5 mg/mL, BD Biosciences) with or without3,000 CAFs. Gastric cancer cell lines (3,000 cells/well) wereseeded on the artificial ECM. To measure the outcomes, wecounted the number of spheroids 200 mm or more in size andthe mean size of the 5 largest spheroids 7 days after seeding thecells under 40� magnification. Immunocytochemical stainingwas conducted after measuring outcomes approximately 1 weekafter the plating of the gastric carcinoma cell lines. Additionalmethodologic details for organotypic cultures and immunocy-tochemical staining were previously published by Dang andcolleagues (19).

    Secretome and transcriptome analysisWe performed secrotome analysis using Proteome Profiler

    Human Cytokine Array Kit (R&D Systems Inc.) in order toidentify upregulated secretory factors in culture supernatantsof cocultured MKN74 cells with CAFs relative to MKN74without CAFs. Conditioned medium from MKN74 cultureswas collected after 48 hours in serum-free medium and incu-bated with arrays containing 36 human cytokine-specific anti-bodies according to the manufacturer's instructions.

    Using RNeasy (Qiagen), RNA was extracted from MKN74cells cultured alone or together with CAFs. RNA quality wasassessed using a 2100 Bioanalyzer (Agilent), and cDNA wassynthesized using the GeneChip WT (Whole Transcript) Ampli-fication kit. Approximately 5.5 mg of labeled DNA target washybridized to Affymetrix GeneChip Human Gene 2.0 ST Arrays.Array data processing and analysis were performed using Affy-metrix GeneChip Command Console Software. Gene setenrichment analysis (GSEA) was used to identify upregulatedpathways in MKN74 cells cocultured with CAFs relative toMKN74 alone. The pathway categories were obtained fromMSigDB (https://software.broadinstitute.org). The cutoff valuefor statistical significance used in GSEA is a false discovery rate(FDR) of 25%. In order to reduce the likelihood of false positiveresults, this study used an FDR value of 5% as the cutoff levelfor enriched gene sets.

    Peritoneal xenograft mouse modelsAnimal care and handling procedures were performed in

    accordance with the Ajou University School of Medicine

    Institutional Animal Care and Use Committee guidelines,and all animal experiments described were approved by theAnimal Research Committee of the institution. The peritonealseeding animal model was established using 6- to 8-week-oldmale athymic nude mice (Orient Bio), weighing 16 to 18 g.We performed 3 types of experiments. In experiment 1, micewere randomly assigned to 3 groups. Mice in the first groupwere intraperitoneally injected with 1 � 107 cancer cells alone(experiment 1, group 1), or with 1 � 106 (experiment 1, group2) or 4 � 106 (experiment 1, group 3) CAFs. In experiment 2,mice were randomly assigned to 2 groups. Mice were intraper-itoneally injected with DDR1 KO cancer cell line only (exper-iment 2, group 1) or with 1� 106 CAFs (experiment 2, group 2).In experiment 3, mice were first assigned to 2 groups. Micewere intraperitoneally injected with 1 � 107 cancer cells only(experiment 3, group 1), or with 1 � 106 CAFs (experiment3 group 2). Each group was randomly divided into 2 sub-groups, and mice were orally administered the DDR1 inhibitor7rh (25 mg/kg) or vehicle every 2 days over 3 weeks, startingon day 1 after cell injection. The mice were sacrificed 3 weeksafter injection, and the numbers and sizes of the peritonealnodules were assessed.

    Statistical analysisRelapse-free survival rates were evaluated using log-rank tests,

    and survival curves were generated using the Kaplan–Meier meth-od. The differences between the expression of DDR1 and thepatterns of recurrence were evaluated using the c2 test. All exper-imental studies were performed independently 3 times. Statisticalanalysis was performed using IBM SPSS statistics (version 21 forMac OS X, IBM) and GraphPad Prism (version 6.0 for Mac OS X,GraphPad) software.

    ResultsDDR1 expression in primary gastric carcinoma tissuescorrelates with peritoneal recurrence

    In a previous report, DDR1 staining was observed at the cellmembrane or in the cytoplasm of tumor cells, and 50.5% ofgastric carcinoma tissues showed positive DDR1 expression(17). Gastric carcinoma tissues with DDR1 expression wereshown to be associated with deeper invasion and poorer overallsurvival, compared with tissues that did not express DDR1 (17).When we performed additional analyses with follow-up dataincluding recurrence, we found that patients with DDR1 expres-sion showed significantly shorter relapse-free survival timesthan DDR1-negative patients (P ¼ 0.043; Fig. 1A). During thefollow-up period, 51 patients (25.2%) presented with recur-rence, and 74.5% of all patients with recurrence had metastaticlesions in the peritoneum, the most common site of recurrence(Fig. 1B). Patients with DDR1 expression had significantlyhigher rates of peritoneal recurrence compared with thosenegative for DDR1 (P ¼ 0.036). However, hematogenous(P ¼ 0.101) or distant lymph node (P ¼ 0.285) recurrenceswere not significantly related to DDR1 expression (Table 1).

    CAFs enhance the peritoneal tumorigenic potential of gastriccarcinoma

    At first, we screened for the expression of DDR1 and mesen-chymal markers, such as a-SMA and Vimentin in gastric carcino-ma cell lines and fibroblasts, and we identified that DDR1 was

    Jin et al.

    Mol Cancer Res; 16(10) October 2018 Molecular Cancer Research1592

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    https://software.broadinstitute.orghttp://mcr.aacrjournals.org/

  • only expressed in gastric carcinoma cell lines (SupplementaryFig. S1B).

    To identify if CAFs enhance the tumorigenesis of gastriccarcinoma cell lines, we conducted organotypic cultures ofgastric carcinoma cell lines (Fig. 2A and B). We found that thenumber of spheroids and the mean size of the 5 largest spher-oids in MKN74 culture were significantly higher in the presenceof CAFs relative to cells cultured on ECM only (Fig. 2C). Inaddition, we established the peritoneal seeding xenograft mousemodel to assess the effects of CAFs on the peritoneal tumori-genesis of gastric carcinoma cell lines in vivo (Fig. 2D). Asexpected, the number of tumors and mean weight of the largestnodules were significantly increased in groups injected withCAFs, and these factors depended on the number of CAFs(Fig. 2E and F). These results indicated that CAFs enhance theperitoneal tumorigenesis of gastric carcinoma cell lines.

    CAFs induced upregulation of DDR1 in gastric carcinomacell lines

    We evaluated the phosphorylation of DDR1 in organotypiccultures of MKN74 through immunocytochemistry (Fig. 3A).Moreover, we identified, through immunohistochemistry, theupregulation of DDR1 in the peritoneal seeding xenograftmouse model, in the groups that received coinjection ofMKN74 and CAFs (Fig. 3B).

    We used a transwell coculture system to investigate CAF-induced changes in DDR1 expression. In gastric carcinoma celllines, MKN74 and MKN45 cells that were cocultured with CAFsusing the transwell coculture system showed increased expressionof DDR1 relative to cultures without CAFs by Western blottinganalysis (Fig. 3C). However, the coculture with fibroblasts iso-lated from normal gastric tissues (normal tissue-associated fibro-blasts: NAFs) did not enhance the expression of DDR1 in gastriccarcinoma cells (Supplementary Fig. S2A). In addition, DDR1

    gene expression in MKN74 cells was not increased despite cocul-ture with CAFs (Supplementary Fig. S2B). These results indicatedthat CAF upregulated DDR1 at the posttranscriptional level andCAF-induced upregulation of DDR1 was involved in the perito-neal spread of gastric carcinoma.

    CAF-induced STAT3 activation may be involved in DDR1upregulation

    We investigated the release of 36 different cytokines, growthfactors and other secreted proteins in conditioned culture mediaof MKN74 with/without CAFs (Supplementary Fig. S3A).Results after 48 hours of culture are shown in SupplementaryFig. S3B. IL6, IL8, and C–C motif chemokine ligand 2 (CCL2)were readily detected in conditioned medium when coculturedwith CAFs. When we analyzed signal transduction pathwaysbased on the KEGG pathway database (https://www.genome.jp/kegg/pathway.html), they were all related to the Janus kinase(JAK1)-signal transducer and activator of transcription (STAT3)signaling pathway (Supplementary Fig. S3B).

    We examined the gene expression changes in MKN74 cellsinduced by coculture with CAFs (Supplementary Fig. S3A).GSEA was used to determine potentially altered signalingpathways and the results identified that the IL6/JAK/STAT3signaling-related genes were enriched in MKN74 cells cocul-tured with CAFs [enrichment score: 1.83, nominal P < 0.001,FDR Q ¼ 0.001, family wise error rare (FWER) P ¼ 0.006;Supplementary Fig. S3C]. To validate these results, we perform-ed Western blotting analysis for the MKN74 cell line coculturedwith/without CAFs. Along with the upregulation and phos-phorylation of DDR1, phosphorylation of STAT3 was signifi-cantly upregulated in MKN74 cocultured with CAFs relativeto MKN74 cells cultured alone (Supplementary Fig. S3D).Meanwhile, the phosphorylation status of AKT was notaltered in coculture with CAFs. These results suggested that

    Figure 1.

    Pattern of recurrence in gastriccarcinoma patients according to DDR1expression. A, DDR1 expression inprimary gastric carcinoma tissuescorrelates with early recurrence in theKaplan–Meier curve, analyzedwith thelog-rank test (P ¼ 0.043). B, Venndiagram of the sites of recurrence ingastric carcinoma patients,demonstrating that the peritoneum isthe most common site of recurrence.

    Table 1. Pattern of recurrence according to DDR1 expression in GCs

    n DDR1 negative (n ¼ 100) DDR1 positive (n ¼ 102) PTotal recurrence 51 19 (19.0%) 32 (31.4%) 0.043Peritoneal recurrence 38 13 (13.0%) 25 (24.5%) 0.036Hematogenous recurrence 19 6 (6.0%) 13 (12.7%) 0.101Distant lymph node recurrence 14 5 (5.0%) 9 (8.8%) 0.285

    DDR1 in Peritoneal Metastasis of Gastric Carcinoma

    www.aacrjournals.org Mol Cancer Res; 16(10) October 2018 1593

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    https:&sol;&sol;www.genome.jp&sol;kegg&sol;pathway.htmlhttps:&sol;&sol;www.genome.jp&sol;kegg&sol;pathway.htmlhttp://mcr.aacrjournals.org/

  • CAFs-induced cytokines such as IL6 activated the JAK/STAT3pathway in gastric carcinoma cells, and this may be related tothe upregulation of DDR1. However, the exact mechanisms ofCAF-induced upregulation of DDR1 need to be clarified throughfuture studies.

    DDR1 inhibition using the CRISPR/Cas9 system suppressesthe effects of CAFs

    We conducted organotypic cultures for gastric carcinomacell lines using DDR1-knockout MKN74. We found that theDDR1-KO MKN74 cells generated equivalent numbers ofspheroids with a mean size similar to the 5 largest spheroidsin the ECM culture only and ECM with CAF coculture condi-tions (Fig. 4A and B). Moreover, we investigated the activationof DDR1 in the DDR1-KO MKN74 cells from the organotypiccultures through immunocytochemistry. As expected, the ECMwith CAFs condition did not enhance the activation of DDR1in the DDR1-KO MKN74 cells (Fig. 4C). When we establishedthe peritoneal seeding xenograft using DDR1-KO MKN74, thenumber of tumors in the groups with DDR1-KO cell line alonewas not significantly different from DDR1-KO cells coinjectedwith CAFs (P ¼ 0.751). There was no expression of DDR1 inthe xenograft tumors of the DDR1-KO cell line (Fig. 4D). Thesedata suggested that the genetic inhibition of DDR1 in gastriccarcinoma cells suppressed the tumor promoting effectsof CAFs.

    Blocking DDR1 signaling through 7rh represses the effects ofCAFs on gastric carcinomas

    We next evaluated the effect of 7rh, a specific inhibitor ofDDR1 (20), which may be a new candidate therapeutic agent.First, we demonstrated that low-dose 7rh does not affect theviability of gastric carcinoma cell lines or fibroblasts (Fig. 5A),and that it significantly reduces DDR1 expression in MKN74cells compared with vehicle treatment (Fig. 5B). In addition, weconfirmed that the COL1A gene encoding COL1A, a ligand ofDDR1, was only expressed in fibroblasts (Supplementary FigS4A). Western blotting analysis for MKN74 cocultured withCAFs showed that low-dose 7rh can reduce the phosphoryla-tion of DDR1 as well as upregulation of DDR1 (SupplementaryFig S4B).

    Then, we conducted organotypic cultures of gastric carcinomacell lines with 7rh. The organotypic cultures revealed that 7rhreduced the number of spheroids and the mean size of the 5largest colonies despite coculture with CAFs (Fig. 5C and D).Immunocytochemistry for the organotypic cultures showed that7rh suppressed the phosphorylation of DDR1 in MKN74 cells(Fig. 5E).

    Moreover, to further clarify the effects of 7rh, we performedin vivo experiments using the peritoneal seeding xenograftmouse model of gastric carcinoma (Fig. 6A). As expected,7rh significantly reduced the number of tumor nodules in theCAF coinjection group (Fig. 6B). Immunohistochemistry on

    Figure 2.

    Peritoneal tumorigenesis of gastric carcinoma cells stimulated by CAFs. A to C, Organotypic culture of gastric carcinoma cell lines. A, Schematic figureof organotypic culture and B, representative photos of MKN74 colonies (scale bar, 200 mm). C, The number of colonies larger than 200 mm and themean sizes of the 5 largest MKN74 colonies were significantly increased in the presence of CAFs (� , P < 0.05, independent t test). D to F, Peritonealtumorigenesis of MKN74 gastric cancer cell line in BALB/c nude mice (n ¼ 8, each group). D, Representative photos of peritoneal nodules (arrow) 18 daysafter injection of MKN74 with/without CAFs into the abdominal cavity. E, The number of tumor nodules in each group and F, mean weight of thelargest nodules (� , P < 0.05; �� , P < 0.01, independent t test).

    Jin et al.

    Mol Cancer Res; 16(10) October 2018 Molecular Cancer Research1594

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • the harvested tumor nodules revealed that 7rh reduced theexpression of DDR1 in peritoneal tumors (Fig. 6C). However,the weight of mice in each group did not show differencesduring the treatment (Fig. 6D). Thus, 7rh suppresses the CAF-induced peritoneal tumorigenesis of gastric carcinomas with-out side effects through the inhibition of DDR1 signaling.

    DiscussionThis study reveals the potential of DDR1 as a novel thera-

    peutic target to prevent stroma-induced gastric carcinoma peri-toneal recurrence after resection. In our cohort of patients, whounderwent curative resection for gastric carcinomas, we found

    that DDR1 expression in primary tumors could be a biomarkerto predict peritoneal recurrence. The in vitro and in vivo experi-ments demonstrated that DDR1 was upregulated in gastriccarcinoma cells stimulated with CAFs. Furthermore, inhibitionof DDR1 induced significant antitumorigenic effects in CAF-stimulated gastric carcinoma cells in organotypic cultures andin a peritoneal xenograft mouse model. Our findings corrobo-rate the favorable evidence for the use of DDR1 inhibitors toprevent peritoneal recurrence after curative resection for gastriccarcinomas and are summarized in Fig. 6E.

    Peritoneal metastasis is a distinctive pattern of recurrence ingastric carcinomas and is associated with poorer outcomes thanhematogenous or lymphatic metastasis (21, 22). In particular,

    Figure 3.

    CAF-induced upregulation of DDR1 in gastric cancer cell lines. A, Immunocytochemistry revealed that DDR1 phosphorylation was induced to a greaterextent on the ECM in the presence of CAFs than on the ECM alone (scale bar, 200 mm; �, P < 0.05, independent t test). B, Immunohistochemistry ofharvested intraperitoneal nodules showed that MKN74 xenografts mixed with CAFs had greater expression of DDR1 and accumulation of fibroblastsexpressing a-SMA than xenografts with only MKN74 cells (scale bar, 50 mm). C, Western blotting showed that coculture with CAFs enhanced DDR1expression in MKN74 and MKN45 gastric carcinoma cells. The DDR1 and b-actin bands were quantified with densitometry using ImageJ and then DDR1 bandintensities were normalized to b-actin band intensity. Relative expression of DDR1 in coculture with CAFs over 48 hours was significantly highercompared with MKN74 cells cultured alone (� , P < 0.05, independent t test).

    DDR1 in Peritoneal Metastasis of Gastric Carcinoma

    www.aacrjournals.org Mol Cancer Res; 16(10) October 2018 1595

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • small peritoneal seeding nodules in the peritoneal cavity are notadequately detected in imaging studies such as abdominal com-puted tomography, so peritoneal recurrence is diagnosed at latestages and associatedwith poor outcomes. Therefore, a predictionof peritoneal metastasis through the analysis of the resectedprimary tumor is required to suggest additional modalities toprevent metastasis after curative resection for gastric carcinomas.However, reliable markers for site-specific recurrence have onlyrarely been investigated in gastric carcinomas. Recently, a tran-scriptome analysis of curatively resected gastric carcinoma tissuesrevealed that SYT8 could be a candidate to predict peritoneal

    recurrence during the follow-up period (23). However, theresearchers obtained transcriptome data from only 16 patients,so the statistical power of the data might be insufficient to con-firm SYT8 as a reliable marker. In addition, therapeutic agentstargeting SYT8 protein are not available. Others reported thatthe infiltration of S100-positive dendritic cells could be areliable marker for peritoneal recurrence of gastric carcinomas(24).However, the studies lacked supportive experimental data todemonstrate a correlation between dendritic cells and the peri-toneal tumorigenesis of gastric carcinoma cells. In the presentstudy, we analyzed the clinical data of 202 gastric carcinoma

    Figure 4.

    Genetic inhibition of DDR1 suppressesthe effects of CAFs. A to C,Organotypic cultures of DDR1-knockout MKN74 cells. A,Representative photos of colonies(scale bar, 200 mm). B, No differenceswere observed in the number ofcolonies larger than 200 mm or themean size of the 5 largest coloniesfrom DDR1-knockout MKN74 cells inthe presence or absence of CAFs. C,Immunocytochemistry showed thatDDR1-knockout MKN74 cells did nothave enhanced DDR1 phosphorylationon the ECM in the presence of CAFs,unlike DDR1-positive MKN74 cells(scale bar, 200 mm). D, Peritonealtumorigenesis of DDR1-knockoutMKN74 cells mixed with or withoutCAFs inBALB/c nudemice (n¼6, eachgroup). There was no difference in thenumber of intraperitoneal tumors(P ¼ 0.751), and we confirmed noexpression of DDR1 in xenografts withDDR1-knockout MKN74 cells.

    Jin et al.

    Mol Cancer Res; 16(10) October 2018 Molecular Cancer Research1596

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • patients, whose primary carcinoma tissues were processed forDDR1 expression by immunohistochemistry. These resultsshowed that DDR1 expression in patients was related to perito-neal recurrence (the most common recurrence), but not to hema-togenous or lymphatic metastasis. Moreover, our experimentalstudies indicated that stroma-induced DDR1 upregulation ingastric carcinoma cell lines enhances their peritoneal tumorigen-esis, whereas the inhibition of DDR1 reduces it. Our findingsprovide support for the use of DDR1 expression as a predictivebiomarker for peritoneal recurrence in patients with gastric car-cinomas, and for the application of 7rh, a DDR1 inhibitor, for theprevention of peritoneal recurrence after curative resection.

    Our previous report using the same cohort as that of the presentstudy indicated that the proportion of DDR1-positive cells inprimary gastric carcinoma tumors increased among the tumorcells that had more deeply invaded the gastric wall (17). Otherstudies, which evaluated the clinical relevance of DDR1 expres-sion for gastrointestinal malignancies, including gastric andesophageal carcinomas, showed similar results (25, 26). Carci-

    noma of the gastrointestinal tract usually originates from epithe-lial cells in the mucosal layer, which subsequently invade thesubmucosa, the muscle, and the serosal layers. The initial step ofcancer invasion is the degradation of the basement membrane,followed by increased cross-talk between the invading carcinomacells and stromal tissue (27). Stromal tissue consists of variouscomponents, including noncancerous cells and ECM. Amongthose components, fibroblasts are essential to contribute to ECMremodeling by secreting proteases such as metalloproteinases,which effectively degrade the ECM (28). In addition, fibroblastscan produce various soluble autocrine and paracrine secretoryproteins that regulate the activity of the surrounding cells as wellas themselves (29). In the present study, we speculated that gastriccarcinoma cells invading the gastric wall upregulate DDR1through interaction with CAFs. Our experimental data supporta role for CAFs in enhancing the upregulation of DDR1 incocultures, organotypic cultures, and peritoneal xenografttumors, whereas DDR1 gene transcription was not altered aftercoculture with CAFs. These findings are consistent with a recent

    Figure 5.

    Effects of 7rh on gastric cancer cell lines. A, Low-dose 7rh did not affect the viability of gastric carcinoma cell lines or fibroblasts. B, Western blottingrevealed that 7rh reduced DDR1 expression in MKN74 cells. The DDR1 and b-actin bands were quantified with densitometric program ImageJ and thenrelative DDR1 band intensities were normalized to b-actin band intensity. C to E, Organotypic cultures of MKN74 cells with 7rh treatment. C, Representativephotos of colonies (scale bar, 200 mm). D, The number of colonies larger than 200 mm and the mean size of the 5 largest colonies from MKN74 cellswere markedly decreased by 7rh treatment on the ECM with CAFs (� , P < 0.05; �� , P < 0.01, independent t test). E, Immunocytochemistry showed that DDR1phosphorylation in MKN74 was suppressed by 7rh treatment on the ECM in the presence of CAFs (scale bar, 200 mm).

    DDR1 in Peritoneal Metastasis of Gastric Carcinoma

    www.aacrjournals.org Mol Cancer Res; 16(10) October 2018 1597

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • report that insulin-like growth factor 1 (IGF-1) induces upregula-tion of DDR1 by suppressing miR-119 through the PI3K/AKTpathway, andnot by enhancing the transcriptionofDDR1 (30). Inthe present study, we suggested several CAF-produced factors,including IL6, IL8, and CCL2, that act as activators to enhanceupregulation of DDR1 in gastric carcinoma cells through secre-tome analysis. Moreover, those factors can activate the JAK/STAT3signaling pathway, which iswell known to be related to cancer cellsurvival (31, 32). CAF-induced JAK/STAT3 signal activation wasvalidated through transcriptome analysis andWesternblotting forgastric carcinoma cells cocultured with CAFs; however, we couldnot discern a direct relationship between JAK/STAT3 and the

    upregulation of DDR1, so the exact mechanisms of CAF-inducedupregulation of DDR1 in gastric carcinomas needs to be morefully explored in the future.

    In the present study, we hypothesized that the upregulationof DDR1 is essential for the tumorigenic ability of CAF-stimulated, free-floating cancer cells in the peritoneum. Ourdata support this hypothesis, as DDR1-knockout cells did notundergo CAF-stimulated increases in colony number or clonesize. There are several studies describing the ability of DDRsto promote tumor growth in metastatic sites like the boneand liver (33, 34). However, there had been no proof of therole of DDR1 in the peritoneal metastasis of cancer cells. In

    Figure 6.

    Effects of 7rh in mouse xenograft model. A to D, Peritoneal tumorigenesis of gastric cancer cell lines in BALB/c nude mice with 7rh treatment (n ¼ 6,each group). Mice were treated with 25 mg/g 7rh 3 times per week for 3 weeks. A, Representative photomicrographs of peritoneal nodules (arrow) frommice that were treated with 7rh for 18 days after MKN74 injection into the abdominal cavity with/without CAFs. B, The number of tumor nodules wasreduced by 7rh treatment in mice injected with MKN74 and CAFs, whereas 7rh treatment did not statistically affect the number of tumor nodules in miceinjected with MKN74 cells only. The effects on the MKN45 tumor nodules were the same as those for MKN74. C, Immunohistochemistry on harvestedintraperitoneal nodules showed that 7rh reduced DDR1 expression in MKN74 xenograft tumors (scale bar, 50 mm). D, Weight of mice showed a similarpattern among each group during treatment. E, Schematic displaying the proposed consequences of peritoneal metastasis of gastric carcinomas,which include invasion of cancer cells from the mucosal layer to serosa, and the activation of cancer cells by CAFs during invasion. Activated cancercells increase DDR1 expression and are exfoliated into the peritoneal cavity as free-floating cells. These cells can form tumor nodules at the peritonealsurface. 7rh can reduce tumorigenesis at the peritoneal surface through inhibition of DDR1.

    Jin et al.

    Mol Cancer Res; 16(10) October 2018 Molecular Cancer Research1598

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • the early stage of peritoneal metastasis, the floating cells dis-seminated from gastric carcinoma primary tumors must survivein the peritoneal cavity (35). In some tumor cells, DDR1directly enhances prosurvival mechanisms through inter-actions with Notch1 or the activation of NF-kB and the down-stream effector COX-2 (36, 37). Finally, the disseminated cellsaggregate into metastatic nodules on the peritoneal surface.DDR1 is known to play a critical role in supporting cell–celladhesion in a variety of noncancerous cells (38–40) andmalignant cell lines (41). Studies suggest that DDR1 stabilizesE-cadherin at the cell surface and promotes cell–cell aggrega-tion (38, 40). Although we have not proposed an exact mech-anism, our results support the hypothesis that the upregulationof DDR1 in gastric carcinoma cells is a critical step in theformation of tumor nodules from disseminated cancer cellson the peritoneal surface.

    The efficacy of adjuvant intraperitoneal chemotherapy hasbeen studied for the prevention of peritoneal recurrence aftercurative resection in patients with locally advanced gastriccarcinomas. Cisplatin and mitomycin C have been used ascytotoxic agents for intraperitoneal administration, but theirsurvival benefits have been controversial (42). A meta-analysisshowed that the risk of postoperative abscess in the abdominalcavity was increased by 137% in patients with intra-abdominalchemotherapy relative to those receiving surgery alone (43).Those findings have resulted in physicians doubting the ben-efits of intraperitoneal chemotherapy for the prevention ofperitoneal recurrence in gastric carcinoma s. Based on thefindings of the present study, we suggest that the novel, orallyavailable drug 7rh may prevent the tumorigenesis of free-floating cancer cells in the abdominal cavity. In experimentsperformed by Gao and colleagues, 7rh potentially inhibited theproliferation, adhesion, and tumorigenicity of cancer cellsexpressing high levels of DDR1 and had a good pharmacoki-netic profile with oral bioavailability of 67.4% (44). Anothergroup described that 7rh could attenuate DDR1 expression innasopharyngeal cancer cells as seen in our data for gastriccarcinoma cells, and the authors suggested JAK1–STAT3 sig-naling pathway as the downstream target of DDR1 (45). In thepresent study, we also confirmed that 7rh potentially attenuatesCAF upregulation and phosphorylation of DDR1 using acoculture system. In addition, our in vivo studies revealed noadverse effects of the oral administration of 7rh, and we also

    confirmed that 7rh reduced the expression of DDR1 in xeno-graft tumors. These results suggested that 7rh could be orallyadministered to patients after resection for primary gastriccarcinomas in order to prevent peritoneal recurrence in patientsat high risk through the downregulation of DDR1 levels.

    Although this study has several limitations such as thenonvalidation of clinical data with an external cohort and theusage of only two types of gastric carcinoma cell lines, we canconclude that intratumoral stromal cells, particularly CAFs,enhance the upregulation of DDR1 in gastric carcinomas, andthat DDR1 is a key factor in CAF-induced tumorigenesis. Wesuggest that treatment with 7rh, a specific inhibitor of DDR1, inan adjuvant setting after curative resection might prevent peri-toneal recurrence of gastric carcinomas.

    Disclosure of Potential Conflicts of InterestR.A. Brekken has ownership interest (including stock, patents, etc.) in and is

    a consultant/advisory board member for Tuevol Therapeutics. No potentialconflicts of interest were disclosed by the other authors.

    Authors' ContributionsConception and design: R.A. Brekken, H. HurDevelopment of methodology: H. Jin, I.-H. HamAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): H.J. Oh, C.A. Bae, D. Lee, S.-U. HanAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): H. Jin, S.-Y. Son, H. HurWriting, review, and/or revision of the manuscript: H. Jin, S.-Y. Son,R.A. Brekken, H. HurAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): H. Jin, Y.-J. ChwaeStudy supervision: Y.-B. Kim, S.-U. Han, H. Hur

    AcknowledgmentsThis research was supported by the Basic Science Research Program

    through the National Research Foundation of Korea (NRF), funded by theMinistry of Education (2016R1D1A1B03933083). This research was alsosupported by a grant from the Korea Health Technology R&D Projectthrough the Korea Health Industry Development Institute (KHIDI), fundedby the Ministry of Health and Welfare, Republic of Korea (HI17C0364).

    The authors thank Professor Ding Ke at Jinan University provision of 7rhand Dave Primm for help in editing this article.

    Received December 13, 2017; revised April 11, 2018; accepted May 30, 2018;published first June 4, 2018.

    References1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al.

    Cancer incidence and mortality worldwide: sources, methods and majorpatterns in GLOBOCAN 2012. Int J Cancer 2015;136:E359–86.

    2. Bang YJ, Kim YW, Yang HK, Chung HC, Park YK, Lee KH, et al. Adjuvantcapecitabine and oxaliplatin for gastric cancer after D2 gastrectomy(CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet2012;379:315–21.

    3. Sakuramoto S, SasakoM, Yamaguchi T, Kinoshita T, Fujii M, Nashimoto A,et al. Adjuvant chemotherapy for gastric cancer with S-1, an oral fluor-opyrimidine. N Engl J Med 2007;357:1810–20.

    4. Ahmed A, Ukwenya AY, Makama JG, Mohammad I. Management andoutcome of gastric carcinoma in Zaria, Nigeria. Afr Health Sci 2011;11:353–61.

    5. Maehara Y, Hasuda S, Koga T, Tokunaga E, Kakeji Y, Sugimachi K.Postoperative outcome and sites of recurrence in patients following cura-tive resection of gastric cancer. Br J Surg 2000;87:353–7.

    6. Roviello F, Marrelli D, de Manzoni G, Morgagni P, Di Leo A, Saragoni L,et al. Prospective study of peritoneal recurrence after curative surgery forgastric cancer. Br J Surg 2003;90:1113–9.

    7. Clark AG, Vignjevic DM. Modes of cancer cell invasion and the role of themicroenvironment. Curr Opin Cell Biol 2015;36:13–22.

    8. Kurashige J, Mima K, Sawada G, Takahashi Y, Eguchi H, Sugimachi K, et al.Epigenetic modulation and repression of miR-200b by cancer-associatedfibroblasts contribute to cancer invasion and peritoneal dissemination ingastric cancer. Carcinogenesis 2015;36:133–41.

    9. YamaguchiH, YoshidaN, TakanashiM, Ito Y, FukamiK, YanagiharaK, et al.Stromal fibroblasts mediate extracellular matrix remodeling and invasionof scirrhous gastric carcinoma cells. PLoS One 2014;9:e85485.

    10. Alves F, Vogel W, Mossie K, Millauer B, Hofler H, Ullrich A. Distinctstructural characteristics of discoidin I subfamily receptor tyrosinekinases and complementary expression in human cancer. Oncogene1995;10:609–18.

    DDR1 in Peritoneal Metastasis of Gastric Carcinoma

    www.aacrjournals.org Mol Cancer Res; 16(10) October 2018 1599

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • 11. Valiathan RR, Marco M, Leitinger B, Kleer CG, Fridman R. Discoidindomain receptor tyrosine kinases: new players in cancer progression.Cancer Metastasis Rev 2012;31:295–321.

    12. Vogel W. Discoidin domain receptors: structural relations and functionalimplications. FASEB J 1999;13:S77–82.

    13. FordCE, Lau SK, ZhuCQ, Andersson T, TsaoMS, VogelWF. Expression andmutation analysis of the discoidin domain receptors 1 and 2 in non-smallcell lung carcinoma. Br J Cancer 2007;96:808–14.

    14. Hou G, Vogel W, Bendeck MP. The discoidin domain receptor tyrosinekinase DDR1 in arterial wound repair. J Clin Invest 2001;107:727–35.

    15. Quan J, Yahata T, Adachi S, Yoshihara K, Tanaka K. Identification ofreceptor tyrosine kinase, discoidin domain receptor 1 (DDR1), as apotential biomarker for serous ovarian cancer. Int J Mol Sci 2011;12:971–82.

    16. Yang SH, Baek HA, Lee HJ, Park HS, Jang KY, Kang MJ, et al. Discoidindomain receptor 1 is associated with poor prognosis of non-small cell lungcarcinomas. Oncol Rep 2010;24:311–9.

    17. Hur H, Ham IH, Lee D, Jin H, Aguilera KY, Oh HJ, et al. Discoidin domainreceptor 1 activity drives an aggressive phenotype in gastric carcinoma.BMC Cancer 2017;17:87.

    18. Lee D, Ham IH, Son SY, Han SU, Kim YB, Hur H. Intratumor stromalproportion predicts aggressive phenotype of gastric signet ring cell carci-nomas. Gastric Cancer 2017;20:591–601.

    19. Dang TT, Prechtl AM, Pearson GW. Breast cancer subtype-specific interac-tions with the microenvironment dictate mechanisms of invasion. CancerRes 2011;71:6857–66.

    20. Aguilera KY, Huang H, Du W, Hagopian MM, Wang Z, Hinz S, et al.Inhibition of discoidin domain receptor 1 reduces collagen-mediatedtumorigenicity in pancreatic ductal adenocarcinoma. Mol Cancer Ther2017;16:2473–85.

    21. Dicken BJ, Bigam DL, Cass C, Mackey JR, Joy AA, Hamilton SM. Gastricadenocarcinoma: review and considerations for future directions. Ann Surg2005;241:27–39.

    22. VanCutsemE, Sagaert X, Topal B,HaustermansK, PrenenH.Gastric cancer.Lancet 2016;388:2654–64.

    23. Kanda M, Shimizu D, Tanaka H, Tanaka C, Kobayashi D, HayashiM, et al. Significance of SYT8 for the detection, prediction, andtreatment of peritoneal metastasis from gastric cancer. Ann Surg2018;267:495–503.

    24. Saito H, Osaki T, Murakami D, Sakamoto T, Kanaji S, Ohro S, et al.Prediction of sites of recurrence in gastric carcinoma using immunohis-tochemical parameters. J Surg Oncol 2007;95:123–8.

    25. Lu X, Nemoto S, Lin A. Identification of c-JunNH2-terminal protein kinase(JNK)-activating kinase 2 as an activator of JNK but not p38. J Biol Chem1997;272:24751–4.

    26. Xie R, Wang X, Qi G, Wu Z, Wei R, Li P, et al. DDR1 enhances invasion andmetastasis of gastric cancer via epithelial-mesenchymal transition. TumourBiol 2016;37:12049–59.

    27. Glentis A, Gurchenkov V, Matic Vignjevic D. Assembly, heterogeneity, andbreaching of the basement membranes. Cell Adh Migr 2014;8:236–45.

    28. Chang HY, Chi JT, Dudoit S, Bondre C, van de Rijn M, Botstein D, et al.Diversity, topographic differentiation, and positional memory in humanfibroblasts. Proc Natl Acad Sci U S A 2002;99:12877–82.

    29. Nakamura T, Matsumoto K, Kiritoshi A, Tano Y, Nakamura T. Induction ofhepatocyte growth factor in fibroblasts by tumor-derived factors affects

    invasive growth of tumor cells: in vitro analysis of tumor-stromal inter-actions. Cancer Res 1997;57:3305–13.

    30. Mata R, Palladino C, Nicolosi ML, Lo Presti AR,Malaguarnera R, RagusaM,et al. IGF-I induces upregulation of DDR1 collagen receptor in breastcancer cells by suppressing MIR-199a-5p through the PI3K/AKT pathway.Oncotarget 2016;7:7683–700.

    31. Johnson DE, O'Keefe RA, Grandis JR. Targeting the IL-6/JAK/STAT3 sig-nalling axis in cancer. Nat Rev Clin Oncol 2018;15:234–48.

    32. Huynh J, Etemadi N, Hollande F, Ernst M, Buchert M. The JAK/STAT3 axis:a comprehensive drug target for solid malignancies. Semin Cancer Biol2017;45:13–22.

    33. Badiola I, Olaso E, Crende O, Friedman SL, Vidal-Vanaclocha F. Discoidindomain receptor 2 deficiency predisposes hepatic tissue to colon carcino-ma metastasis. Gut 2012;61:1465–72.

    34. Valencia K, Ormazabal C, Zandueta C, Luis-Ravelo D, Anton I, Pajares MJ,et al. Inhibition of collagen receptor discoidin domain receptor-1 (DDR1)reduces cell survival, homing, and colonization in lung cancer bonemetastasis. Clin Cancer Res 2012;18:969–80.

    35. KandaM, Kodera Y.Molecular mechanisms of peritoneal dissemination ingastric cancer. World J Gastroenterol 2016;22:6829–40.

    36. Das S, Ongusaha PP, Yang YS, Park JM, Aaronson SA, Lee SW. Discoidindomain receptor 1 receptor tyrosine kinase induces cyclooxygenase-2and promotes chemoresistance through nuclear factor-kappaB pathwayactivation. Cancer Res 2006;66:8123–30.

    37. Kim HG, Hwang SY, Aaronson SA, Mandinova A, Lee SW. DDR1 receptortyrosine kinase promotes prosurvival pathway through Notch1 activation.J Biol Chem 2011;286:17672–81.

    38. Eswaramoorthy R,Wang CK, ChenWC, TangMJ, HoML, Hwang CC, et al.DDR1 regulates the stabilization of cell surface E-cadherin and E-cadherin-mediated cell aggregation. J Cell Physiol 2010;224:387–97.

    39. Wang CZ, Yeh YC, Tang MJ. DDR1/E-cadherin complex regulates theactivation of DDR1 and cell spreading. Am J Physiol Cell Physiol 2009;297:C419–29.

    40. Yeh YC, Wu CC, Wang YK, Tang MJ. DDR1 triggers epithelial cell differ-entiation by promoting cell adhesion through stabilization of E-cadherin.Mol Biol Cell 2011;22:940–53.

    41. Hidalgo-Carcedo C, Hooper S, Chaudhry SI, Williamson P, Harrington K,Leitinger B, et al. Collective cell migration requires suppression of acto-myosin at cell-cell contacts mediated by DDR1 and the cell polarityregulators Par3 and Par6. Nat Cell Biol 2011;13:49–58.

    42. Feingold PL, Kwong ML, Davis JL, Rudloff U. Adjuvant intraperitonealchemotherapy for the treatment of gastric cancer at risk for peritonealcarcinomatosis: A systematic review. J Surg Oncol 2017;115:192–201.

    43. Yan TD, Black D, Sugarbaker PH, Zhu J, Yonemura Y, Petrou G, et al. Asystematic review and meta-analysis of the randomized controlled trialson adjuvant intraperitoneal chemotherapy for resectable gastric cancer.Ann Surg Oncol 2007;14:2702–13.

    44. Gao M, Duan L, Luo J, Zhang L, Lu X, Zhang Y, et al. Discovery andoptimization of 3-(2-(Pyrazolo[1,5-a]pyrimidin-6-yl)ethynyl)benza-mides as novel selective and orally bioavailable discoidin domain receptor1 (DDR1) inhibitors. J Med Chem 2013;56:3281–95.

    45. LuQP, ChenWD, Peng JR, Xu YD, Cai Q, FengGK, et al. Antitumor activityof 7RH, a discoidin domain receptor 1 inhibitor, alone or in combinationwith dasatinib exhibits antitumor effects in nasopharyngeal carcinomacells. Oncol Lett 2016;12:3598–608.

    Mol Cancer Res; 16(10) October 2018 Molecular Cancer Research1600

    Jin et al.

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/

  • 2018;16:1590-1600. Published OnlineFirst June 4, 2018.Mol Cancer Res Hyejin Jin, In-Hye Ham, Hye Jeong Oh, et al. Stroma-Induced Peritoneal Metastasis in Gastric CarcinomaInhibition of Discoidin Domain Receptor 1 Prevents

    Updated version

    10.1158/1541-7786.MCR-17-0710doi:

    Access the most recent version of this article at:

    Material

    Supplementary

    http://mcr.aacrjournals.org/content/suppl/2018/06/02/1541-7786.MCR-17-0710.DC1

    Access the most recent supplemental material at:

    Cited articles

    http://mcr.aacrjournals.org/content/16/10/1590.full#ref-list-1

    This article cites 45 articles, 10 of which you can access for free at:

    Citing articles

    http://mcr.aacrjournals.org/content/16/10/1590.full#related-urls

    This article has been cited by 1 HighWire-hosted articles. Access the articles at:

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

    Subscriptions

    Reprints and

    [email protected]

    To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

    Permissions

    Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

    .http://mcr.aacrjournals.org/content/16/10/1590To request permission to re-use all or part of this article, use this link

    on August 26, 2020. © 2018 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

    Published OnlineFirst June 4, 2018; DOI: 10.1158/1541-7786.MCR-17-0710

    http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-17-0710http://mcr.aacrjournals.org/content/suppl/2018/06/02/1541-7786.MCR-17-0710.DC1http://mcr.aacrjournals.org/content/16/10/1590.full#ref-list-1http://mcr.aacrjournals.org/content/16/10/1590.full#related-urlshttp://mcr.aacrjournals.org/cgi/alertsmailto:[email protected]://mcr.aacrjournals.org/content/16/10/1590http://mcr.aacrjournals.org/

    /ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages false /GrayImageMinResolution 200 /GrayImageMinResolutionPolicy /Warning /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages false /MonoImageMinResolution 600 /MonoImageMinResolutionPolicy /Warning /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 900 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False

    /CreateJDFFile false /Description > /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ > /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MarksOffset 18 /MarksWeight 0.250000 /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /NA /PageMarksFile /RomanDefault /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /LeaveUntagged /UseDocumentBleed false >> > ]>> setdistillerparams> setpagedevice