ADAM17 is a Tumor Promoter and Therapeutic Target in ...Cancer Therapy: Preclinical ADAM17 is a Tumor Promoter and Therapeutic Target in Western Diet–associated Colon Cancer Reba

Cancer Therapy: Preclinical

ADAM17 is a Tumor Promoter and TherapeuticTarget in Western Diet–associated Colon CancerReba Mustafi1, Urszula Dougherty1, Devkumar Mustafi2, Fatma Ayaloglu-Butun1,Michelle Fletcher1, Sarbani Adhikari1, Farhana Sadiq1, Katherine Meckel1, Haider I. Haider1,Abdurahman Khalil1, Joel Pekow1, Vani Konda1, Loren Joseph3, John Hart3,Alessandro Fichera4, Yan Chun Li1, and Marc Bissonnette1

Abstract

Purpose: Epidermal growth factor receptors (EGFR) arerequired for tumor promotion by Western diet. The metallopro-tease, ADAM17 activates EGFR by releasing pro-EGFR ligands.ADAM17 is regulated by G-protein–coupled receptors, includingCXCR4. Here we investigated CXCR4–ADAM17 crosstalk andexamined the role of ADAM17 in tumorigenesis.

Experimental Design: We used CXCR4 inhibitor, AMD3100and ADAM17 inhibitor, BMS566394 to assess CXCR4–ADAM17crosstalk in colon cancer cells. We compared the expression ofCXCR4 ligand, CXCL2, and ADAM17 in mice fed Western dietversus standard diet. Separately, mice were treated with marima-stat, a broad-spectrum ADAM17 inhibitor, or AMD3100 to assessEGFR activation by ADAM17 and CXCR4. Using Apc-mutant Minmice, we investigated the effects of ADAM17/10 inhibitorINCB3619 on tumorigenesis. To assess the effects of colonocyteADAM17, mice with ADAM17 conditional deletion were treatedwith azoxymethane (AOM). ADAM17 expression was also com-

pared in colonocytes from primary human colon cancers andadjacent mucosa.

Results: CXCL12 treatment activated colon cancer cell EGFRsignals, andCXCR4or ADAM17blockade reduced this activation.In vivo, Western diet increased CXCL12 in stromal cells and TGFain colonocytes. Marimastat or AMD3100 caused >50% reductionin EGFR signals (P < 0.05). InMinmice, INCB3619 reduced EGFRsignals in adenomas and inhibited intestinal tumor multiplicity(P < 0.05). In the AOM model, colonocyte ADAM17 deletionreduced EGFR signals and colonic tumor development (P < 0.05).Finally, ADAM17was upregulated >2.5-fold in humanmalignantcolonocytes.

Conclusions: ADAM17 is a Western diet–inducible enzymeactivated by CXCL12–CXCR4 signaling, suggesting the pathway:Western diet!CXCL12!CXCR4!ADAM17!TGFa!EGFR.ADAM17 might serve as a druggable target in chemopreventionstrategies. Clin Cancer Res; 23(2); 549–61. �2016 AACR.

IntroductionColon cancer is a major cause of cancer-related morbidity and

mortality in the Western world (1). Most colon cancers aresporadic, and Western diets are implicated in their pathogenesis(2). Western diets are low in digestible fiber and rich in animalfats. Increasing epidemics of obesity and diabetes mellitus in theUnited States are linked to Western diet, and are also associatedwith increased colon cancer risk (3, 4). Disturbingly, the world-wide incidence of colon cancer is increasing as more countriesadopt Western diets (1, 5). Thus, efforts to increase our under-

standing of the mechanisms by which Western diets promotecolon cancer are important to develop more effective strategies toprevent this disease.

Experimental models of colon cancer are widely employedto dissect diet-related influences in colonic tumorigenesis. Boththe genetic Apc-mutant Min mouse and the carcinogen-inducedazoxymethane (AOM) model of colon cancer are widely usedto investigate the roles of dietary factors in tumor development(6, 7). Western diet promotes tumorigenesis in the AOM andMin mouse model (8–10). In prior studies, using EGFR inhi-bitors, we showed that the signals from these receptors wererequired for efficient AOM-induced tumor development inrodents (11, 12). We also established that this receptor wasrequired for Western diet–induced tumor promotion in theAOM and azoxymethane/dextran sulfate sodium (AOM/DSS)models (9, 10).

EGFR is a member of the transmembrane ErbB tyrosinekinase receptor family that controls growth and differentiationof the colonic epithelium (13). EGFR is expressed on colonicepithelial cells and stromal cells including fibroblasts andendothelial cells (14, 15). This receptor is activated by numer-ous ligands, including TGFa and amphiregulin (AREG) that areproduced in the colon by epithelial cells and stromal cells.Upon ligand binding, EGFR homodimerizes or heterodi-merizes with other ErbB members. Dimerization stimulatesthe intrinsic receptor tyrosine kinase activity (13). In colon

1Department ofMedicine, University of Chicago, Chicago, Illinois. 2Department ofRadiology, University of Chicago, Chicago, Illinois. 3Department of Pathology,University of Chicago, Chicago, Illinois. 4Department of Surgery, University ofChicago, Chicago, Illinois.

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

R. Mustafi and U. Dougherty contributed equally to this article.

Corresponding Author: Marc Bissonnette, Department of Medicine, The Uni-versity of Chicago, KCBD Building, Rm 9112, 900 East 57th Street, Chicago, IL60637. Phone: 773-795-0833; Fax: 773-702-2281; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-15-3140

�2016 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 549

on June 21, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 3, 2016; DOI: 10.1158/1078-0432.CCR-15-3140

http://clincancerres.aacrjournals.org/

cancer, EGFR is most frequently activated by increases in ligandabundance (16). In this regard, we showed that Western dietalone increased colonic mucosal TGFa, whereas TGFa andamphiregulin were upregulated in colonic tumors (9).

ADAM enzymes (members of a disintegrin and metallopro-teinase family) regulate cleavage of EGFR ligands from mem-brane-bound pro-ligands. TGFa and AREG release are con-trolled by ADAM17, which was shown to be upregulated inhuman colon cancers (17). As a Western diet increased TGFa, itwas of interest to examine the effects of Western diet onADAM17 expression. In many cancers, ADAM17 is also aconvergence point downstream of G-protein–coupled recep-tors (GPCR; ref. 18) that can increase ADAM17 activity. In thisregard, GPCR CXCR4 and its ligand CXCL12 are implicated inthe progression of colon cancer (19). Using pharmacologicinhibitors, we asked whether Western diet might alter CXCL12in vivo and whether CXCL12–CXCR4 signals could regulateADAM17–EGFR signals in colon cancer cells in vitro. Marima-stat is a broad-spectrum metalloprotease inhibitor that blocksADAM17 activity, and AMD3100 is a potent CXCR4 inhibitor(20, 21). Here, using marimastat and AMD3100, we investi-gated the requirements for ADAM17 activity and CXCL12–CXCR4 signals in Western diet–promoted colonic EGFRsignals.

In fibroblasts, ADAM17 and ADAM10 regulate release of dif-ferent EGFR ligands (22). The availability of a dual ADAM17 andADAM10 inhibitor INCB3619 allowed us to examine the roles ofthese enzymes in intestinal tumorigenesis in theMinmouse (23).This model requires EGFR signals for robust tumorigenesis andresponds toWestern diet with increased tumor burden (8, 24, 25).To isolate ADAM17 effects and dissect the cell-specific role ofADAM17 in colon cancer development, we examined the effect ofdeleting ADAM17 from colonic epithelial cells in the AOMmodel. Thus, in the current study, we examined the role ofADAM17 in two different Western diet–promoted models ofcolon cancer. Using an ADAM17-specific inhibitor BMS566394,we also determined that CXCL12–CXCR4 signals activate EGFRvia ADAM17 in colon cancer cells. Taken together, the results of

this study expand our understanding of the CXCL12–CXCR4circuit activating ADAM17 as a potential link between EGFR andWestern diet in colon cancer.

Materials and MethodsFor details, see Supplemental Materials and Methods.

Experimental animalsArchived colonic mucosal tissue for chronic diet studies. All animalstudies were approved by the Animal Care and Use Committee atthe University of Chicago and are fully compliant with the NIHguidelines for humane use of animals. Archived samples wereavailable from prior studies taken from control mice fed standarddiet (5% fat) or Western diet (20% fat) for 40 weeks (9). Samplesincluded distal colonic segments fixed in 10% formalin or frozenin optimal cutting temperature (OCT) medium as well as RNAand proteins from scrape-isolated colonic mucosa

Acute Western diet studiesMarimastat and AMD3100 experiments. We used A/J mice foracute dietary studies as we planned to study colon tumordevelopment in the AOM model that required an AOM-sus-ceptible mouse. Male A/J mice were implanted with Alzetpumps delivering 1.1 mmoles marimastat per day per mousein 50% polyethylene glycol (PEG) or 50% PEG alone (vehiclecontrol), or 0.126 mmoles AMD3100 per mouse per day orwater (vehicle control). There were 10 mice in each group.Marimastat is a broad-spectrum metalloprotease inhibitor, andAMD3100 is a CXCR4 inhibitor (21, 26). Mice were allowed 48hours to recover after pump implantation and then fed aWestern diet (9). Mice were sacrificed after 2 weeks and leftcolonic mucosa scrape-isolated and flash frozen. Circular seg-ments (3-mm length) were frozen in OCT or fixed in 10%buffered formalin and embedded in paraffin for immunostain-ing. For AMD3100 studies, mice were injected with bromo-deoxyuridine (BrdUrd) 1 mL/100 g bodyweight 2 hours beforesacrifice as recommended by the manufacturer.

Apc-mutant Minmouse studies. Apc-mutantMinmice on C57Bl6/Jbackground, containing a truncating mutation in Apc codon850, were obtained from The Jackson Laboratory (#002020).Mice were genotyped as recommended by The Jackson labo-ratory. At 6 weeks of age, Min mice (males and females) wererandomized and fed Western diet (n ¼ 20) or Western dietsupplemented with INCB3619 (n ¼ 20; 860 mg/kg chow).INCB3619 was added as a powder mixed with sucrose andprovided 120 mg/kg bodyweight/day. Western diet composi-tion was described previously (9). Mice were sacrificed at 7months of age and tumors >2 mm in size were collected in 4-cmsegments from duodenum, jejunum and ileum, and entirecolon as described previously (27). Tumors were collected forproteins and RNA, and segments fixed in formalin for immu-nostaining. Control proteins and RNA were prepared by scrapeisolation from tumor-free colonic and intestinal mucosa.

ADAM17 LoxP studiesAcute adeno-Cre ADAM17 LoxP studies.Mice with LoxP sequencesflanking ADAM17 exon 2 were obtained from The Jackson lab-oratory (stock number 009597) and genotyped as recommended.To confirm conditional ADAM17 deletion, we treated homozy-gous ADAM17LoxP mice with intrarectal adeno-Cre or adeno EV

Translational Relevance

Western diets are implicated in sporadic colon cancer.Epidermal growth factor receptors (EGFR) are increased inhuman colon cancer and required for tumor promotion byWestern diet in models of colon cancer. EGFR is activated byreceptor ligands released frommembrane-bound pro-ligands.ADAM17 and ADAM10 aremajor enzymes that regulate EGFRligand release. While ADAM17 is increased in sporadic coloncancer, no studies have directly tested whether this enzymepromotes tumor development. Moreover, as ADAM17 con-trols the release of many substrates, its role in colon cancerdevelopment is not predictable. In this study, using bothgenetic and chemical models of colon cancer, we demonstratefor the first time that ADAM17, and possibly ADAM10, aretumor-promoting. This establishes ADAM17 as a potentialtherapeutic target in colon cancer prevention. Furthermore, weuncovered a potentially important pathway linking ADAM17to Western diet–promoted colon cancer involving CXCL12–CXCR4 as upstream signals activating ADAM17.

Mustafi et al.

Clin Cancer Res; 23(2) January 15, 2017 Clinical Cancer Research550




(control) as described previously (28). After 7 days, mice weresacrificed anddistal colon segments frozen inOCT and stained forADAM17.

Chronic villin-Cre ADAM17 LoxP AOM tumor studies. Homozy-gousADAM17LoxPmice originally on aC57Bl6 backgroundwereinterbred 10 generations to AOM-susceptible A/J background andmated to villin-Cre–expressing mice that were also interbred 10generations to A/J background. Experimental villin-Creþ-homo-zygous ADAM17 LoxP (ADAM17D/D) mice and littermate con-trols homozygous ADAM17 LoxP mice lacking villin-Cre trans-gene were treated with AOM 10 mg/kg bodyweight weekly for 6weeks.Micewere then fed aWestern diet (9). Tumor developmentwas monitored by colonoscopy, and mice were sacrificed 24weeks after the first AOM injection (29). Colons were excised,and tumors graded histologically by a gastrointestinal pathologist(J. Hart). AOM-induced tumors, and scrape-isolated controlmucosa from left colon of genotype-matched ADAM17LoxPmicetreated with saline (AOM vehicle), were fixed in 10% formalin orsnap frozen inOCTorflash frozen for RNAandprotein extraction.

Isolation of colonocytes and stromal cells from colonicmucosaWe isolated colonocytes and stromal cells following previously

published methods with minor modifications (30, 31). Briefly,colons were removed and mucosa scrape-isolated and mincedwith razor blade into 2-mm pieces. Fragments were collected intubes containing 6-mL sterile ice-cold transport media, 50 IU/mLpenicillin, 50 mg/mL streptomycin, and 0.5 mg/mL gentamycin.Tissue was washed three times by gentle inversion and collectedby gravity sedimentation and resuspended in 10-mL chelatingbuffer (transport media plus 1 mmol/L EDTA and 1 mmol/LEGTA). Tissue was incubated on a shaker overnight at 4�C torelease colonocytes into the supernatant. The pellet was washedthree times with 3-mL ice-cold PBS, releasing residual colono-cytes, and colonocyte-containing supernatants were combined.Epithelial cell fractions (supernatants) and stromal cell fractions(pellets) were centrifuged at 400� g for 5minutes at 4�C, yieldingloosely packed pellets fromwhich proteinswere solubilized in 2�Laemmli buffer.

Cell culture and proliferationHCT116 and HT29 human colon cancer cells were obtained

within 6 months from ATCC and authenticated by ATCC usingshort tandem repeat DNA fingerprinting. Cells were cultured at37�C in a humidified atmosphere of 5% CO2–95% air underconditions recommended by ATCC. For EGFR transactivationexperiments, preconfluent cells were pretreated for 2 hours with20 mg/mL C225 antibodies, 5 mmol/L BMS566394, 10 ng/mLAMD3100, or PBS (vehicle) followed by 50 ng/mL CXCL12 orvehicle. Positive controls included cells treated with 10 ng/mLEGF. At indicated times, cells were broken and lysates probed forproteins by Western blotting.

Real-time PCRRNA was extracted from snap-frozen tissue using RNeasy Lipid

Tissue Mini Kit. Samples were homogenized in a bullet blenderand loaded onto an RNA-binding spin column, washed, anddigestedwithDNase I and eluted in 30-mL elution buffer. Sampleswere examined by Agilent chip for RNA purity and quantified byRibogreen. RNA (100ng)was reverse transcribed into cDNAusinga high capacity reverse transcription kit in 20-mL total volume.

Incubation conditions were 25�C for 10 minutes, 37�C for 120minutes, and 85�C for 5 minutes. The resulting first-strand com-plementary DNA (cDNA) was used as a template for quantitativePCR in triplicate using Fast SYBR Green Master Mix Kit. SeeSupplementary Methods for further details.

IHCFive-micron sections of formalin-fixed paraffin-embedded tis-

sue were cut and mounted on Vectabond-coated Superfrost Plusslides. Slides were heated to 60�C for 1 hour, deparaffinized by 5-minutewashes three times in xylene, hydrated in a graded series ofethanol washes, and rinsed with distilled water. Epitope retrievalswere achievedbymicrowaveheating for 15minutes in 10mmol/LTris-EDTAbuffer, pH9. For Ki67 staining, sectionswere incubatedin Ki67 antibodies (clone SP6) overnight at 4�C at 1:500 finaldilution (cat. # RM-9106-S0, Fisher Scientific). Ki67 and BrdUrdwere quantified using the Fiji image software (NIH, Bethesda,MD) with ImmunoRatio plug-in (32). Alcian blue staining wasperformed on 5-mm paraffin sections as recommended by themanufacturer. After deparaffinization and rehydration, sectionswere incubatedwithAlcianBlue solution for 30minutes, rinsed inwater, and counterstained with Nuclear Fast Red. For CXCL12staining, circular colonic segments (10 mm) in OCT were fixedwith cold acetone (�20�C) for 15 minutes and air dried at roomtemperature. Sections were blocked with Super Block (AAA500,ScykTek Laboratories) for 30 minutes at room temperature fol-lowed by overnight incubationwith Abcamprimary anti-CXCL12antibodies (1:35) at 4�C. See Supplementary Methods for furtherdetails.

In situ staining for TGFa transcriptsCircular left colon segments (12-mm) frozen in OCT were

prepared with a cryostat, placed on Superfrost Plus slides, fixedin 4%paraformaldehyde in PBS for 20minutes, andwashed twicewith DEPC-PBS for 5minutes. In situ hybridization and detectionof TGFa transcripts was performed using Exiqon miCURY LNAanti-sense oligonucleotide probes labeled 50- and 30- with digox-igenin (see Supplementary Methods for further details).

Western blottingProteins were extracted in SDS-containing Laemmli buffer,

quantified by RC-DC protein assay, and subjected to Westernblotting as described previously (9). Protein expression levelswere expressed as fold of indicated protein in scrape-isolatedtumor-free intestinal mucosa (Apcþ/Min mouse), or colonicmucosa from saline-treated genotype-matched ADAM17LoxPmice (mean � SD). Separate aliquots were probed for b-actin.

Statistical analysisWestern blotting densitometry data were normalized as fold

of vehicle-treated (cells) or normal mucosa (mice) and expressedasmeans� SD. Differences amongmultiple groups were assessedwith ANOVA, and differences between specific groups werethen comparedusingunpaired Student t test. Reverse transcriptasereactions were run in duplicate and assayed in triplicate, andCt values were averaged. UntransformedCt values were comparedbetween groups (33). Relative abundance, expressed as 2�DDCt,was calculated by exponentiating the estimated differences in Ct

and significance between groups calculated by Student t test.Tumor incidence was defined as the percent of mice with at leastone tumor and compared using Fisher exact test. Tumor

ADAM17 Is Deregulated in Colon Cancer

www.aacrjournals.org Clin Cancer Res; 23(2) January 15, 2017 551




multiplicity was defined as the average number of tumors pertumor-bearing mouse and compared using Mann–Whitney Utest. For all statistical analyses, P values �0.05 were consideredstatistically significant.

ResultsWestern diet upregulates CXCL12 and EGFR signalingcomponents

We previously investigated the effects of Western diet on coloncancer development (9). For the current study, we used archivedsamples from the control armof the prior study todissect effects ofdiet alone on EGFR signals. We examined transcript and proteinexpression levels of ADAM17, pro-TGFa, and K-Ras in colonicmucosa from control mice. In addition, we assessed the effects ofdiet on CXCL12, a cytokine implicated in colon cancer, which canactivate EGFR in other cancers. The protocol is shown in Fig. 1A.Western diet significantly increased transcripts (Fig. 1B) andprotein levels of ADAM17, pro-TGFa, and K-Ras, as well asCXCL12 in the distal colonic mucosa (Fig. 1C and D). In situhybridization of TGFa mRNA showed that transcripts were pre-dominantly expressed in colonic epithelial cells (Fig.1E and F),whereas CXCL12 was most abundant in stromal cells (Fig. 1Gand H). Western diets increased both TGFa (Fig. 1B–F) andCXCL12 levels (Fig. 1B–D, G, and H).

ADAM17mediates EGFR transactivation induced by CXCL12 incolon cancer cells

As colonic mucosal CXCL12 was increased by Western diet(Fig. 1) and shown to transactivate EGFR inother cancer cells (34),we asked whether CXCL12 could activate EGFR in colon cancercells. We pretreatedHT29 colon cancer cells with vehicle or EGFR-neutralizing C225 antibodies that block receptor signals by inhi-biting ligand binding. Cells were then treated with CXCL12 orvehicle. As shown in Fig. 2A, CXCL12 caused time-dependentincreases in phospho-active EGFR (pEGFR), pErbB2, a hetero-dimeric partner of EGFR, and CXCL12 increased ERK and AKTactivation. EGFR transactivation by CXCL12 required EGFRligand binding as assessed by suppression with anti-EGFR–spe-cific C225 antibodies. These results are quantified in Fig. 2B. Toassess whether EGFR transactivation by CXCL12 requiredADAM17, we pretreated cells with BMS566934, a specificADAM17 inhibitor (35). As shown in Fig. 2C, BMS566394suppressed EGFR signals induced by CXCL12 as assessed byreductions in phospho-EGFR, pErbB2, pAKT, and pERK. Theseresults are quantified in Fig. 2D. pEGFR and pErbB2 were greatestat 1 minute, while pAKT and pERK increases were greatest at 10minutes, consistent with the latter being downstream effectors ofEGFR. We obtained similar results in HCT116 cells (Supplemen-tary Fig. S1). As expected, AMD3100, a CXCR4 inhibitor, alsoreduced EGFR activation by CXCL12 (as assessed by reductions inpErbB2, pERK, and pAKT) in HT29 cells (Fig. 2E and F). Asignaling diagram showing colon cancer cell EGFR transactivationby CXCL12–CXCR4 is shown in Supplementary Fig. S2.

EGFR signals require metalloprotease activity and CXCR4signals in Western diet–fed miceMarimastat. As ADAM17 and several EGFR components wereincreased in colonic mucosa by Western diet, it was of interest toassess the role of ADAM17 in these diet effects. As specificADAM17 inhibitors were not commercially available, we chose

marimastat, a broad-spectrum metalloprotease inhibitor thatblocks ADAM17 (26), to assess a metalloprotease requirementfor EGFR signals in mice fed a Western diet. As shown in Sup-plementary Fig. S3, marimastat significantly inhibited EGFR sig-nals in colonic mucosa, with greater than 50% reductions inphospho-active EGFR (pEGFR), pErbB2, and pERK comparedwith Western diet alone. Thus, EGFR signaling in the colonicmucosa in Western diet–fed mice requires metalloprotease activ-ity. These results implicate ADAM17 and possibly ADAM10 asthey are major ADAMs responsible for regulating EGFR ligandrelease (22).

AMD3100. As CXCL12 was increased by Western diet (Fig. 1B–D,G, andH) and can transactivate EGFR in colon cancer cells (Fig. 2),we examined the effects of blocking CXCL12 signals in Westerndiet–fed mice using AMD3100, an inhibitor of G-protein–cou-pled receptor CXCR4 (Fig. 3A). As shown in Fig. 3B and C,AMD3100 significantly reduced phospho-active EGFR and down-stream effectors, indicating that EGFR signals are positively mod-ulated by CXCR4 signals underWestern diet conditions. To assessthe biological consequences of CXCR4 blockade, we measuredBrdUrd incorporation in the colonicmucosa. As shown in Fig. 3Dand E, AMD3100 significantly reduced BrdUrd incorporation incolonic crypts from12.8%� 6.7% to 6.3%� 2.9% (Fig. 3F, n¼ 4;P < 0.05).

ADAM17/10 inhibitor INCB3619 suppresses ADAM17 activityand reduces tumor development in Apc-mutant Min mice

We next employed INCB3619, an inhibitor with greater spec-ificity, to block EGFR ligand release. INCB3619 is orally bioavail-able and highly specific for ADAM17 and ADAM10, which reg-ulate release of EGFR ligands (22, 23). In initial experiments, weconfirmed that INCB3619blocked recombinantADAM17activityin vitro (Supplementary Fig. S4A). INCB3619 supplementationin vivo (120 mg/kg bodyweight/day) also inhibited intestinalmucosal ADAM17 activity, decreasing Vmax by 50% (Supplemen-tary Fig. S4B). To dissect the role of ADAM17/10 on tumorgrowth, we next examined the effects of INCB3619 on intestinaltumorigenesis in the Apc-mutant Min mouse. Mice on Westerndiet supplemented with INCB3619 grew normally, with growthcurve closely paralleling mice fed a Western diet alone (Fig. 4A).As shown in Fig. 4B, INCB3619 did not alter tumor incidence inthe small intestine, but it appeared to reduce tumors in thecolon, with 25% colonic tumor incidence in the Western dietplus INCB3619-treated group, compared with 50% in Westerndiet alone group (P ¼ 0.07). In the small intestine, INCB3619significantly reduced median tumor multiplicity from 67.0 �11.0 in the Western diet group to 38.5 � 5.8 in the group givenINCB3619 (Fig. 4C, n ¼ 20 Western diet alone, and n ¼ 20Western diet þINCB3619; P < 0.01). As shown in Fig. 4D,INCB3619 also decreased small intestinal tumor size from 3.3� 0.6 mm to 2.4 � 0.2 mm (P < 0.05).

In agreement with prior studies (25), EGFR signals were upre-gulated in Min tumors as shown in Fig. 5A and quantified in Fig.5B. Compared with levels in mice on Western diet alone,INCB3619 significantly reduced activation levels of several ofthese proteins in adenomas and adjacent normal-appearing intes-tinal mucosa (Fig. 5A and B). EGFR effectors c-Myc, cyclin D1 andCox-2 and Notch1 effector Hes1 were also increased in tumors,and INCB3619 significantly reduced these proteins in Min ade-nomas (Fig. 5C andD). Todetermine the biological consequences

Mustafi et al.





of these signaling changes, we assessed the effects of INCB3619oncell proliferation in Min adenomas. As shown in Fig. 5E andquantified in Fig. 5F, INCB3619 significantly reduced Ki67 stain-

ing in adenomas from 58.1� 10.3 to 13.3� 3.0 (�, P < 0.001; n¼10 adenomas frommice onWestern diet and n¼ 11 adenomas inmice on Western diet plus INCB3619). Proliferation in normal-

Figure 1.

Effects ofWestern diet on CXCL12 andEGFR signaling components in colonicmucosa. Flash-frozen colonic mucosaltissues from control mice fed standardor Western diet (WD) for 40 weekswere used from a prior study (9).Tissues were extracted for proteinsand RNA, and real-time PCR andWestern blotting were performed.A, Experimental protocol. B, CXCL12,ADAM17, TGFa, and K-Ras transcripts[� , P < 0.05; †, P < 0.005; z, P < 0.001,compared with levels in mice fedstandard (Std) diet; n ¼ 4 mice foreach diet]. C, Representative Westernblots for indicated proteins.D, Expression levels for indicatedproteins (� , P < 0.05; z, P < 0.001compared with levels in standarddiet–fed mice; n ¼ 4 mice for eachdiet). E and F, In situ TGFa mRNA(white arrows) under standard diet (E)and Western diet conditions (F);magnification, 10�. The relativefluorescence intensity of TGFa mRNAwas 2.8 � 1.4 in Western dietcompared with standard diet(P < 0.05). G and H, CXCL12immunostaining (white arrows)magnification, 10�. Colon sections inOCT frommice on standard diet (G) orWestern diet (H) were stained forCXCL12. Note: CXCL12 ispredominantly expressed in stromalcells in the pericryptal mucosa andsubmucosa and increased in Westerndiet–fed mice. In situ TGFa mRNAhybridization and CXCL12-immunostained sections arerepresentative of 3 mice in eachdietary group.






Mustafi et al.





appearing small intestinal mucosa was also significantly reducedby INCB3619 (Supplementary Fig. S5). As Notch signals in theintestine are controlled by ADAM10 in wild-type mice (36), weexamined the effect of INCB3619 on goblet cells in Min mousesmall intestine, as a read out of Notch activity. As shown inSupplementary Fig. S6, goblet cells were increased by INCB3619consistent with blockade of Notch signals. Notch signals havebeen shown to contribute to intestinal tumorigenesis in the Minmouse (37). Pro-TNFa was also increased in small intestine byINCB3619, consistent with the predicted effect of inhibited cleav-age of pro-TNFa (Supplementary Fig. S7).

Although Min mice have substantially fewer colonic tumorscompared with small intestinal tumors, we also examined theeffects of INCB3619 on colonic crypt cell proliferation andtumors. INCB3619 reduced colonic crypt cell proliferation to16.7% � 1.7%, compared with 24.1% � 3.3% for Western dietalone as assessed by Ki67 staining (n ¼ 10 mice per group, P <0.005). There were 25 colonic tumors in 10 mice in the Westerndiet alone group (N ¼ 20 mice) and 9 colonic tumors in 5 micein the INCB3619 supplemented group (N ¼ 20 mice). Tumorincidence was, therefore, 50% in unsupplemented group and25% in the INCB3619 supplemented group (P ¼ 0.07, Fisher

Figure 3.

CXCR4 inhibitor AMD3100 suppressesEGFR signals and proliferation incolonic mucosa of mice fed Westerndiet (WD). Mice were implanted withAlzet pumps and received vehicle orAMD3100 (5 mg/kg bodyweight/day)and were fed a Western diet. After2 weeks, mice were sacrificed and leftcolons were harvested. Indicatedproteins were measured by Westernblotting. A, Treatment protocol.B, EGFR signals. Western blots ofindicated proteins from scrape-isolatedmucosa in left colon from mice onWestern diet (WD) alone or Westerndiet þ AMD3100 (þ). C, Quantitativeexpression levels (� , P < 0.001;†, P < 0.005 compared with Westerndiet alone, n ¼ 4 mice in each group).D, BrdUrd incorporation in vehicle-treatedmouse. E,BrdUrd incorporationin AMD3100-treated mouse. F, BrdUrdquantitation (� , P < 0.05, comparedwith BrdU in Western diet alone;n ¼ 4 mice per condition).

Figure 2.EGFR activation by CXCL12 requires EGFR ligand binding and is mediated by ADAM17 and inhibited by AMD3100.A, Anti-EGFR antibody blocks EGFR activation byCXCL12. HT29 cells were pretreated for 2 hourswith PBS (vehicle, Ctl) or 20 mg/mLC225, an anti-EGFR antibody. Cells were then stimulatedwith CXCL12 (50 ng/mL)for the indicated times (1, 5, 10minutes) and cell lysates probed for phospho-EGFR (pEGFR), pErbB2, pAKT andpERK, aswell as for panEGFR, panErbB2, panAKT, andpanERK and b-actin as loading control. Separate wells were treated with vehicle or EGF (10 ng/mL) as a positive control. Note that C225 antibodies suppressedEGFR signals, indicating that EGFR activation by CXCL12 involves an EGFR ligand-dependent mechanism. B, Quantified phospho-active EGFR signals in cellspretreatedwith vehicle or C225 normalized to vehicle-treated (control) cells. C,ADAM17mediates the effect of CXCL12 on EGFR activation. Cells were pretreated for2 hours with vehicle (�) or 5 mmol/L BMS566394 (þ, BMS), an ADAM17-specific inhibitor, and then stimulated with 50 ng/mL CXCL12 for the indicated times (1, 5,10 minutes). Cell lysates were probed for the indicated phospho-active proteins (EGFR cascade). MWS: lane containing molecular weight standards. Note thatBMS566394 inhibits CXCL12-induced EGFR signals, indicating that ADAM17mediates EGFR transactivation by CXCL12. D, Indicated phospho-protein signals in cellspretreated with BMS566394 normalized to vehicle-treated (control) cells. E, AMD3100 inhibits EGFR activation by CXCL12 as assessed by pErbB2, pERK, and pAKT.HT29 cells were pretreated with 10 ng/mL AMD3100 or PBS (vehicle) for 2 hours, and cells were then stimulated with 100 ng/mL CXCL12 or vehicle. At the indicatedtimes (1, 5, 10 minutes), cell lysates were probed for pErbB2, pERK, and pAKT. F, Phospho-active protein signals in cells pretreated with vehicle or AMD3100normalized to signals in vehicle-treated cells. Results in A–F are representative of two independent platings with comparable findings.






exact test). Tumor multiplicities were 2.5 and 1.8 tumors pertumor-bearing mouse, respectively (P ¼ 0.23, Mann–WhitneyU test). The average colonic tumor size was 2.9 mm in theWestern diet group and 2.2 mm in the INCB3619-supplemen-ted group (P < 0.05).

Colonocyte ADAM17 deletion suppresses AOM-inducedcolonic tumorigenesis

The INCB3619 studies established that concomitant inhibi-tion of ADAM17 and ADAM10 reduced Min tumorigenesis. Todirectly test the role of ADAM17 in tumor development andisolate epithelial cell–specific contribution, we deletedADAM17 from colonocytes by interbreeding mice with anADAM17 exon 2 flanked by LoxP sequences to transgenic miceexpressing Cre-recombinase under a villin promoter (38, 39).Both the villin-Cre transgenic mice and the ADAM17 floxedmice were first intercrossed 10 generations with A/J mice toachieve an AOM-susceptible A/J background (C57Bl6 mice arerelatively resistant to AOM). We examined the ability of villin-Cre to delete ADAM17 from colonocytes. Colons from control(no AOM treatment) homozygous ADAM17LoxP mice andvillin-Cre-homozygous ADAM17LoxP mice (ADAM17DD) wereharvested, and colonocytes and stromal cells prepared asdescribed in the Materials and Methods. ADAM17 was detected

by Western blotting in fractions from the indicated mice in Fig.6A. Stromal cells were detected by vimentin (VIM) expressionand colonocytes by cytokeratin 20 (CK20) expression. Notethat ADAM17 was readily detected in stromal cells of bothgenotypes. In contrast, ADAM17 was absent in colonocytesfrom mice expressing villin-Cre. These results confirmed thatADAM17 was deleted from crypt epithelial cells by villin-Crerecombinase.

To assess the role of colonocyte ADAM17 in colonic tumordevelopment, we treated villin-Cre-homozygous ADAM17LoxPmice (ADAM17D/D) and control homozygous ADAM17LoxPmice (no Cre) with AOM or saline (AOM vehicle) weekly for 6weeks. Animals were sacrificed after 28 weeks when screeningcolonoscopy revealed the presence of tumors. ADAM17 wasabundant in all cell types in tumors from homozygousADAM17LoxP mice as shown by ADAM17 immunostainingin Supplementary Fig. S8. In contrast, ADAM17 was detectedonly in stromal cells of tumors from villin-Cre-homozygousADAM17LoxP mice (Supplementary Fig. S8). Deletion ofADAM17 from colonocytes significantly reduced the incidenceof tumors (adenomas þ cancers) as well as cancers alone(Fig. 6B). ADAM17 deletion from colonocytes also significantlysuppressed tumor multiplicity (P < 0.05) as shown in Fig. 6C,from a median of 5 tumors/mouse in homozygous

Figure 4.

ADAM17/ADAM10 inhibitor INCB3619suppresses tumor development inApc-mutant Min mice. A, Effect ofINCB3619 on mouse growth.(� , <P < 0.05 compared with Westerndiet alone). B, Effects of INCB3619 onsmall intestine (SI) and colon tumorincidence. C, INCB3619 suppressessmall intestinal tumor multiplicity[n ¼ 20 mice on Western diet (WD)and 20 mice on Western diet þINCB3619; � , P < 0.01, Mann–WhitneyU test). D, INCB3619 suppresses smallintestinal tumor growth (mean � SD,P < 0.05).

Mustafi et al.





Figure 5.

INCB3619 suppresses EGFR signals and proliferation in adenomas from Apc-mutant Min mice. Mice were treated as described in the Materials and Methods. At 7months of age, mice were sacrificed, and intestinal adenomas harvested. Diet-matched adjacent normal-appearing intestinal mucosa was scrape-isolated andproteins solubilized in Laemmli buffer. A,Western blots of indicated EGFR signals. B, Densitometry. Values represent 6–8 mice per group [� , P < 0.05; �� , P < 0.005,phospho-protein levels in adenomas from mice on Western diet (WD) alone (tumor), compared with adjacent normal-appearing mucosa (Nl); †, P < 0.05,phospho-protein levels in normal-appearingmucosa in group receiving INCB3619 (Nlþ INCB), compared with levels in intestinal mucosa frommice onWestern dietalone (Nl); z, P < 0.05; zz, P < 0.005, phospho-protein levels in adenomas frommice on INCB3619 (Tumorþ INCB), compared with adenomas frommice onWesterndiet alone (tumor)]. C,Western blots of indicated EGFR and Notch1 effectors. D, Densitometry [� , P < 0.05; �� , P < 0.005, compared with levels in normal-appearingintestinal mucosa; z, P < 0.05, compared with levels in adenomas from mice on Western diet alone; n ¼ 6–8 mice per group]. E, Ki67 immunostaining inrepresentative adenoma from mouse on Western diet alone (left) and adenoma from mouse on Western diet þ INCB3619 (right). F, Quantitation of Ki67 staining(� , P < 0.001 compared with Western diet alone, n¼ 10 adenomas from mice onWestern diet alone, n¼ 11 adenomas from mice onWestern dietþ INCB3619. Eachadenoma is representative of a different mouse).






ADAM17LoxP mice (n ¼ 32) versus 2 tumors per mouse invillin-Cre-homozygous ADAM17LoxP mice (n ¼ 29). Deletionof colonocyte ADAM17 also reduced tumor cell proliferation(Fig. 6D) and tumor size (Fig. 6E). We next examined the effectsof colonocyte ADAM17 deletion on EGFR and Notch1 signals.As shown in Fig. 6F and quantified in Fig. 6G, EGFR signals(pErbB2, pERK) and Hes1 were significantly reduced in tumorsfrom mice with ADAM17DD, compared with tumors from micewith ADAM17LoxP.

To assess ADAM17 changes in human colon cancers, weisolated colonocytes and stromal cells from colon cancers andadjacent normal-appearing mucosa. As shown in Supplemen-tary Fig. S9, ADAM17 appeared to be increased in malignantcolonocytes compared with adjacent nontransformed colono-cytes (P ¼ 0.06, n ¼ 4 tumors). Interestingly, three of the 4patients summarized in the right panel of Supplementary Fig.S9 had diabetes mellitus and/or hyperlipidemia—two meta-bolic disorders frequently associated with Western style diets.In an additional 6 intact human cancers, we found that

ADAM17 expression was 3.1 � 1.8–fold higher in tumorscompared with adjacent mucosa (P < 0.05, n ¼ 6). These resultsare consistent with prior findings showing increased ADAM17immunostaining in tumors with staining predominantly inmalignant colonocytes (17).

DiscussionIn the current study, we demonstrated that the CXCL12–

CXCR4 axis activated ADAM17–EGFR signals in vitro in coloncancer cells and in vivo in colonic mucosa of mice fed a Westerndiet. Furthermore, we showed that pharmacologic inhibition ofADAM17 and ADAM10 and genetic deletion of colonocyteADAM17 suppress Western diet–promoted tumor develop-ment, establishing an oncogenic role for ADAM17, and possi-bly ADAM10, in Western diet–promoted colon cancer. Otherpreclinical studies showed that ADAM17 haploinsufficiencyalters energy metabolism and protects mice against obesityand diabetes, suggesting a further benefit for targeting this

Figure 6.

Colonocyte ADAM17 deletion suppresses EGFR signals and reduces colonic tumor development in AOM-treated mice. Villin-Cre and ADAM17 LoxP/þ mice wereinterbred to an AOM-susceptible A/J background. Villin-Cre-homozygous ADAM17LoxP (ADAM17DD) and homozygous ADAM17LoxP mice were then treated withAOM or vehicle (saline). Two weeks after the sixth AOM injection, mice were begun on a Western diet (WD, 20%) as described in the Materials and Methods.Twenty-eight weeks after the first AOM treatment, mice were sacrificed and tumors harvested. A, Cell-specific ADAM17 expression. Colonocytes and stromal cellswere isolated from scraped mucosa from left colon of villin-Creþ - homozygous ADAM17LoxP (villin-Creþ) and homozygous ADAM17LoxP mice (villin-Cre �) asdescribed in the Materials and Methods. (Mice in A received no AOM treatment). The indicated cell fractions were probed byWestern blotting for ADAM17, vimentin(VIM, stromal cell marker) and cytokeratin20 (CK20, colonocyte marker). Note that in villin-Creþ - homozygous ADAM17LoxP mice (ADAM17DD), ADAM17was deleted from colonocytes, but not from stromal cells. B, Colon tumor incidence. Tumors were classified histologically as adenomas or carcinomas (� ,†, P < 0.05compared with control homozygous ADAM17LoxP mice; n ¼ 32 homozygous ADAM17LoxP mice and n ¼ 29 ADAM17DD mice). C, Colon tumor multiplicity.The median number of tumors was significantly lower in ADAM17DD group compared with the homozygous ADAM17LoxP group (2.0 vs. 5.0 tumors per tumor-bearingmouse, respectively; � ,P <0.05, Mann–WhitneyU test).D,Ki67 staining (�,P<0.05, n¼ 14 tumors in each genotype). E,Tumor size (� ,P <0.05; n¼ 35 tumorsfrom homozygous ADAM17LoxP mice and n ¼ 12 tumors from ADAM17DD mice). F, EGFR and Notch1 signals in AOM tumors (T) and genotype-matchedscrape-isolated left colonic mucosa from vehicle-treated mice (N). Blots are representative of 6 tumors per group. G, Quantitative densitometry (� , P < 0.05compared with homozygous ADAM17LoxP mucosa; †, P < 0.05 compared with adenomas in the homozygous ADAM17LoxP group).

Mustafi et al.





protease (40, 41). ADAM17 and ADAM10 regulate cleavage ofmany substrates, including several pro-EGFR ligands (22).While the roles of other ADAM17 and ADAM10 substratesremain uncertain with respect to tumor development, growingevidence indicates that EGFR ligands contribute importantly tocolon cancer development.

In the current study, we also demonstrated that a Westerndiet increased ADAM17, colonocyte TGFa, and stromal cellCXCL12 in the colonic mucosa (Fig. 1). Furthermore, CXCL12activated EGFR signals in HT29 cells and HCT116 cells, pro-totypes of chromosomal-unstable and microsatellite-unstablecolon cancer cells, respectively (Fig. 2 and SupplementaryFig. S1). EGFR transactivation required ligand binding to EGFreceptors and was mediated by ADAM17. As expected,AMD3100 suppressed EGFR activation by CXCL12 in coloncancer cells (Fig. 2). ADAM10 has also been implicated in EGFRtransactivation by CXCR4 in other cancers (42), but we havenot investigated this pathway in colon cancer cells. Mechanismsby which CXCR4 activates ADAM17 and possibly ADAM10remain largely unknown (43).

Broad-spectrum metalloprotease inhibitor marimastat andCXCR4 inhibitor AMD3100 reduced colonic EGFR signals(Fig. 3 and Supplementary Fig. S2). These studies uncover apotentially important role for ADAM17 and CXCL12–CXCR4signals in Western diet–promoted EGFR activation in the colon.We did not investigate potential antitumor effects of AMD3100in cancer models, but our studies showing reductions inEGFR transactivation and colonic proliferation suggest thatAMD3100 might exert antitumor effects in these models.AMD3100 was also shown to inhibit experimental colitis(44). A clinical trial is currently in progress that will testAMD3100 antitumor efficacy in advanced human cancers(Clinical Trials NCT02179970).

AsADAM17andADAM10control releaseof EGFR ligands (22),it was of interest to examine the antitumor effects of INCB3619,an inhibitor that blocks both these metalloproteases. In priorstudies, INCB3619 demonstrated antigrowth effects in severaltumor xenograft models (45). We confirmed that INCB3619inhibited recombinant ADAM17 in vitro and intestinal mucosalADAM17 in vivo (Supplementary Fig. S4). INCB3619 also reducedintestinal tumor burden in Western diet–fed Min mice (Fig. 4)and suppressed EGFR downstream signals in tumors andadjacent normal-appearing intestinal mucosa (Fig. 5). Inter-estingly, INCB3619 reduced pEGFR levels in normal appear-ing mucosa, but not in tumors, suggesting that the inhibitormay suppress tumor initiation more than tumor progression.Supporting this hypothesis regarding an effect on tumorinitiation, INCB3619 also suppressed intestinal and colonicepithelial cell proliferation.

In the clinic, broad-spectrum metalloproteases, such as mar-imastat, were poorly tolerated because of debilitating periarticularinflammation and tendonitis (26). In contrast, INCB3619 waswell tolerated (45). A newer generation ADAM17/ADAM10inhibitor INCB7839 was also well tolerated in a phase I trial(46). In another clinical trial, administration of this analogue,along with trastuzumab to patients with advanced HER2-positivebreast cancers, was well tolerated and induced responses in 50%of the evaluable patients (47). Ongoing development of moreselective agents to inhibit ADAM17 and possibly ADAM10 sug-gest, therefore, that this may be a useful strategy in individuals atincreased risk for colon cancer.

ADAM17 has more than 40 identified substrates, includingNotch1 and Notch ligands and TNFa. Blocking TNFa and Notchsignaling may contribute to the antitumor effects of INCB3619.Pro-TNFa was more abundant in Min adenomas from micesupplemented with INCB3619, consistent with reduced pro-TNFa cleavage by ADAM17 (Supplementary Fig. S7). INCB3619also suppressed Notch1 signaling as assessed by decreased Hes1(Fig. 5C and D). ADAM10 is the major regulator of Notchsignaling in intestinal stem cells under non-stress conditions(36). In preliminary studies, we found that Alcian blue stainingwas increased by INCB3619 in Apcþ/Min mouse small intestine,consistent with ADAM10 blockade (Supplementary Fig. S6). Inprimary human colon cancers and colon cancer cells, however,ADAM17 has also been shown to regulate Notch signaling byparacrine and autocrine mechanisms involving release of Notchligands (48, 49).

Mice with colonocyte ADAM17 deletion have a normal phe-notype in the absence of stress, but develop more severe DSS-induced colitis (50, 51). To dissect the role of colonocyteADAM17 in colon cancer development, we showed that ADAM17deletion in colonic epithelial cells significantly reduced tumordevelopment (Fig. 6). These studies established an importantoncogenic role for colonocyte ADAM17 in a Western diet–pro-moted model of colon cancer. Despite colonocyte ADAM17deletion, however, tumors did emerge, albeit with lower efficien-cy. This might reflect persistence of ADAM17 in stromal cells,releasing stromal-derived EGFR ligands or other ADAM17 sub-strates that can activate colonocytes by a paracrine mechanism. Inthis regard, as in the case of INCB3619 studies, deletion ofcolonocyte ADAM17did not significantly reduce pEGFR in colon-ic tumors, but did decrease pErbB2, pERK, and Hes1, suggestingcontributions from other ADAM17-sensitive pathways such asNotch1. It is possible that INCB3619 inhibited aberrant Notchsignaling by blocking ADAM10, whereas colonocyte ADAM17deletion inhibited release of Notch ligands such as Jagged1 orJagged2. In this regard, in addition to EGFR-dependent growth-inhibitory effects in colon cancer cells, studies with potent anti-ADAM17 antibodies have uncovered EGFR-independent growth-inhibitory effects consistent with contributions from TNFa andNotch1 (52). Our human studies emphasize the potential onco-genic importance of increased ADAM17 signaling in malignantcolonocytes as this enzyme was upregulated in transformedhuman colonocytes (Supplementary Fig. S9) in agreement withprior studies in intact tumors (17).

In the current study, we focused on Western diet effects ontumor development as our prior studies showed that EGFR wasrequired for tumor promotion under Western diet, but notstandard diet conditions (9). Our initial observations, however,were carried out in hybrid C57Bl6 � A/J mice that are relativelyresistant to AOM. It is possible that colonocyte ADAM17 deletionin AOM-susceptible A/J mice might also suppress tumor devel-opment under standard diet conditions. This could suggest anevenwider role for ADAM17blockade in colon cancer prevention.

In conclusion,wehave identified aCXCL12–CXCR4 circuit thatactivates ADAM17, linking Western diet to colonic EGFR signals,and we have shown that ADAM17 plays an important role intumor development in the AOMandMinmousemodels of coloncancer. A schema summarizing these findings is shown in Sup-plementary Fig. S2. These studies support further research target-ing colonic ADAM17 in individuals at increased risk for coloncancer development.






Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: R. Mustafi, A. Fichera, Y.C. Li, M. BissonnetteDevelopment of methodology: R.Mustafi, U. Dougherty, F. Sadiq, H.I. Haider,Y.C. Li, M. BissonnetteAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): R. Mustafi, U. Dougherty, D. Mustafi, F. Ayaloglu-Butun, M. Fletcher, K. Meckel, A. Khalil, J. Pekow, J. Hart, A. Fichera,M. BissonnetteAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): R. Mustafi, D. Mustafi, M. Fletcher, S. Adhikari,J. Pekow, L. Joseph, Y.C. Li, M. BissonnetteWriting, review, and/or revision of the manuscript: R. Mustafi, D. Mustafi,J. Pekow, V.J.A. Konda, L. Joseph, A. Fichera, Y.C. Li, M. BissonnetteAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): R. Mustafi, U. Dougherty, K. Meckel, A. Khalil,M. BissonnetteStudy supervision: R. Mustafi, M. Bissonnette

AcknowledgmentsWe gratefully acknowledge the generous gift of ADAM17 inhibitor

BMS566394 from Bristol Myers Squibb (Bristol-Myers Squibb Company)provided by Dr. Bruce Walcheck (University of Minnesota, Minneapolis, MN)and INCB3619 from the Incyte Corporation. We also thank Dr. T.-C. He forproviding the adenovirus and adeno-Cre virus.

Grant SupportThis work was supported by the NIH grants (CA164124; to M. Bissonnette,

CA180087; to Y.C. Li), P30DK42086 (Digestive Disease Research Core Center),and the Samuel Freedman GI Cancer Laboratory Fund at the University ofChicago.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received January 4, 2016; revised July 18, 2016; accepted July 25, 2016;published OnlineFirst August 3, 2016.

References1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer

statistics. CA Cancer J Clin 2011;61:69–90.2. Weisburger JH. Dietary fat and risk of chronic disease: mechanistic insights

from experimental studies. J Am Diet Assoc 1997;97:S16–23.3. La Vecchia C, Negri E, Decarli A, Franceschi S. Diabetes mellitus

and colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 1997;6:1007–10.

4. Frezza EE,WachtelMS, Chiriva-InternatiM. Influence of obesity on the riskof developing colon cancer. Gut 2006;55:285–91.

5. Sung JJ, Lau JY, Goh KL, Leung WK. Asia Pacific Working Group onColorectal C. Increasing incidence of colorectal cancer in Asia: implicationsfor screening. Lancet Oncol 2005;6:871–6.

6. Kwong LN, Dove WF. APC and its modifiers in colon cancer. Adv Exp MedBiol 2009;656:85–106.

7. TakahashiM,Wakabayashi K. Genemutations and altered gene expressionin azoxymethane-induced colon carcinogenesis in rodents. Cancer Sci2004;95:475–80.

8. Wasan HS, Novelli M, Bee J, Bodmer WF. Dietary fat influences on polypphenotype in multiple intestinal neoplasia mice. Proc Natl Acad Sci U S A1997;94:3308–13.

9. Dougherty U, Cerasi D, Taylor I, Tekin U, Badal S, Aluri L, et al.Epidermal growth factor receptor is required for colonic tumor promo-tion by dietary fat in the azoxymethane/dextran sulfate sodium model:roles of transforming growth factor-{alpha} and PTGS2. Clin CancerRes 2009;15:6780–9.

10. ZhuH, Dougherty U, Robinson V,Mustafi R, Pekow J, Kupfer S, et al. EGFRsignals downregulate tumor suppressorsmiR-143 andmiR-145 inWesterndiet-promotedmurine colon cancer: role of G1 regulators. Mol Cancer Res2011;9:960–75.

11. Fichera A, Little N, Jagadeeswaran S, Dougherty U, Sehdev A, Mustafi R,et al. EGFR signaling is required for microadenoma formation in themouse azoxymethane model of colonic carcinogenesis. Cancer Res2007;67:827–35.

12. Dougherty U, Sehdev A, Cerda S, Mustafi R, Little N, Yuan W, et al.Epidermal growth factor receptor controls flat dysplastic aberrant cryptfoci development and colon cancer progression in the rat azoxymethanemodel. Clin Cancer Res 2008;14:2253–62.

13. Prenzel N, Fischer OM, Streit S, Hart S, Ullrich A. The epidermalgrowth factor receptor family as a central element for cellularsignal transduction and diversification. Endocr Relat Cancer 2001;8:11–31.

14. Kuwai T, Nakamura T, Sasaki T, Kim SJ, Fan D, Villares GJ, et al. Phos-phorylated epidermal growth factor receptor on tumor-associated endo-thelial cells is a primary target for therapy with tyrosine kinase inhibitors.Neoplasia 2008;10:489–500.

15. Yoo J, Rodriguez Perez CE, NieW, Edwards RA, Sinnett-Smith J, RozengurtE. TNF-alpha induces upregulation of EGFR expression and signaling inhuman colonic myofibroblasts. Am J Physiol Gastrointest Liver Physiol2012;302:G805–14.

16. Barber TD, Vogelstein B, Kinzler KW, Velculescu VE. Somatic mutations ofEGFR in colorectal cancers and glioblastomas. N Engl J Med 2004;351:2883.

17. Blanchot-Jossic F, JarryA,MassonD, Bach-NgohouK, Paineau J,DenisMG,et al. Up-regulated expression of ADAM17 in human colon carcinoma: co-expression with EGFR in neoplastic and endothelial cells. J Pathol2005;207:156–63.

18. Fischer OM, Hart S, Gschwind A, Ullrich A. EGFR signal transactivation incancer cells. Biochem Soc Trans 2003;31:1203–8.

19. Akishima-Fukasawa Y, Nakanishi Y, Ino Y, Moriya Y, Kanai Y, Hirohashi S.Prognostic significance of CXCL12 expression in patients with colorectalcarcinoma. Am J Clin Pathol 2009;132:202–10.

20. de Meijer VE, Le HD, Meisel JA, Sharma AK, Popov Y, Puder M. Tumornecrosis factor alpha-converting enzyme inhibition reverses hepatic stea-tosis and improves insulin sensitivity markers and surgical outcome inmice. PLoS One 2011;6:e25587.

21. Rosenkilde MM, Gerlach LO, Jakobsen JS, Skerlj RT, Bridger GJ, SchwartzTW. Molecular mechanism of AMD3100 antagonism in the CXCR4 recep-tor: transfer of binding site to the CXCR3 receptor. J Biol Chem 2004;279:3033–41.

22. Sahin U, Weskamp G, Kelly K, Zhou HM, Higashiyama S, Peschon J, et al.Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of sixEGFR ligands. J Cell Biol 2004;164:769–79.

23. Liu X, Fridman JS,WangQ, Caulder E, YangG, CovingtonM, et al. Selectiveinhibition of ADAM metalloproteases blocks HER-2 extracellular domain(ECD) cleavage and potentiates the anti-tumor effects of trastuzumab.Cancer Biol Ther 2006;5:648–56.

24. Roberts RB, Min L, Washington MK, Olsen SJ, Settle SH, Coffey RJ, et al.Importance of epidermal growth factor receptor signaling in establishmentof adenomas and maintenance of carcinomas during intestinal tumori-genesis. Proc Natl Acad Sci U S A 2002;99:1521–6.

25. Moran AE, Hunt DH, Javid SH, RedstonM, Carothers AM, Bertagnolli MM.Apc deficiency is associated with increased Egfr activity in the intestinalenterocytes and adenomas of C57BL/6J-Min/þ mice. J Biol Chem2004;279:43261–72.

26. RasmussenHS,McCannPP.Matrixmetalloproteinase inhibition as anovelanticancer strategy: a review with special focus on batimastat and marima-stat. Pharmacol Ther 1997;75:69–75.

27. Moser AR, Pitot HC, Dove WF. A dominant mutation that predis-poses to multiple intestinal neoplasia in the mouse. Science 1990;247:322–4.


Mustafi et al.




28. Hung KE, MaricevichMA, Richard LG, ChenWY, RichardsonMP, Kunin A,et al. Development of a mouse model for sporadic and metastatic colontumors and its use in assessing drug treatment. Proc Natl Acad Sci U S A2010;107:1565–70.

29. Dougherty U, Mustafi R, Wang Y, Musch MW, Wang C-Z, Konda VJ, et al.American ginseng suppresses Western diet-promoted tumorigenesis inmodel of inflammation-associated colon cancer: role of EGFR. BMCComplement Altern Med 2011;11:111.

30. Seidelin JB, Horn T, Nielsen OH. Simple and efficient methodfor isolation and cultivation of endoscopically obtained humancolonocytes. Am J Physiol Gastrointest Liver Physiol 2003;285:G1122–8.

31. Khalil H, Nie W, Edwards RA, Yoo J. Isolation of primary myofibroblastsfrom mouse and human colon tissue. J Vis Exp 2013;80:e50611.

32. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T,et al. Fiji: an open-source platform for biological-image analysis. NatMethods 2012;9:676–82.

33. Livak KJ, Schmittgen TD. Analysis of relative gene expression data usingreal-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods2001;25:402–8.

34. Porcile C, Bajetto A, Barbieri F, Barbero S, Bonavia R, Biglieri M, et al.Stromal cell-derived factor-1alpha (SDF-1alpha/CXCL12) stimulates ovar-ian cancer cell growth through the EGF receptor transactivation. Exp CellRes 2005;308:241–53.

35. Romee R, Foley B, Lenvik T, Wang Y, Zhang B, Ankarlo D, et al. NK cellCD16 surface expression and function is regulated by a disintegrin andmetalloprotease-17 (ADAM17). Blood 2013;121:3599–608.

36. Tsai YH, VanDussen KL, Sawey ET, Wade AW, Kasper C, Rakshit S, et al.ADAM10 regulates Notch function in intestinal stem cells of mice. Gas-troenterology 2014;147:822–34 e13.

37. Fre S, Pallavi SK, HuygheM, La�eM, Janssen K-P, Robine S, et al. Notch andWnt signals cooperatively control cell proliferation and tumorigenesis inthe intestine. Proc Natl Acad Sci U S A 2009;106:6309–14.

38. Horiuchi K, Kimura T, Miyamoto T, Takaishi H, Okada Y, Toyama Y, et al.Cutting edge: TNF-alpha-converting enzyme (TACE/ADAM17) inactiva-tion in mouse myeloid cells prevents lethality from endotoxin shock.J Immunol 2007;179:2686–9.

39. Madison BB, Dunbar L, Qiao XT, Braunstein K, Braunstein E, GumucioDL.Cis elements of the villin gene control expression in restricted domains ofthe vertical (crypt) and horizontal (duodenum, cecum) axes of the intes-tine. J Biol Chem 2002;277:33275–83.

40. SerinoM,Menghini R, Fiorentino L,AmorusoR,MaurielloA, LauroD, et al.Mice heterozygous for tumor necrosis factor-alpha converting enzyme are

protected from obesity-induced insulin resistance and diabetes. Diabetes2007;56:2541–6.

41. Gelling RW, Yan W, Al-Noori S, Pardini A, Morton GJ, Ogimoto K, et al.Deficiency of TNFalpha converting enzyme (TACE/ADAM17) causes alean, hypermetabolic phenotype in mice. Endocrinology 2008;149:6053–64.

42. Kasina S, Scherle PA, Hall CL, Macoska JA. ADAM-mediated amphiregulinshedding and EGFR transactivation. Cell Prolif 2009;42:799–812.

43. Blobel CP. ADAMs: key components in EGFR signalling and development.Nat Rev Mol Cell Biol 2005;6:32–43.

44. Xia XM, Wang FY, Zhou J, Hu KF, Li SW, Zou BB. CXCR4 antagonistAMD3100modulates claudin expression and intestinal barrier function inexperimental colitis. PLoS One 2011;6:e27282.

45. Fridman JS, Caulder E, HansburyM, Liu X, Yang G,WangQ, et al. Selectiveinhibition of ADAMmetalloproteases as a novel approach for modulatingErbB pathways in cancer. Clin Cancer Res 2007;13:1892–902.

46. Infante J, Burris H, Lewis N, Donehower R, Redman J, Friedman S, et al. Amulticenter phase Ib study of the safety, pharmacokinetics, biologicalactivity and clinical efficacy of INCB7839, a potent and selective inhibitorof ADAM10 and ADAM17. Breast Cancer Res Treat 2007;106:S269.

47. Newton R, Bradley E, Levy R, Doval D, Bondarde S, Sahoo TP, et al. Clinicalbenefit of INCB7839, a potent and selective ADAM inhibitor, in combi-nation with trastuzumab in patients with metastatic HER2þ breast cancer.J Clin Oncol 2010;28:15s, (suppl; abstr 3025).

48. Wang R, Ye X, Bhattacharya R, Boulbes DR, Fan F, Xia L, et al. A disintegrinand metalloproteinase domain 17 regulates colorectal cancer stem cellsand chemosensitivity via notch1 signaling. Stem Cells Transl Med 2016;5:331–8.

49. Lu J, Ye X, Fan F, Xia L, Bhattacharya R, Bellister S, et al. Endothelial cellspromote the colorectal cancer stem cell phenotype through a soluble formof Jagged-1. Cancer Cell 2013;23:171–85.

50. Feng Y, Tsai YH, Xiao W, Ralls MW, Stoeck A, Wilson CL, et al. Loss ofADAM17-mediated tumor necrosis factor alpha signaling in intestinalcells attenuatesmucosal atrophy in amousemodel of parenteral nutrition.Mol Cell Biol 2015;35:3604–21.

51. Shimoda M, Horiuchi K, Sasaki A, Tsukamoto T, Okabayashi K, HasegawaH, et al. Epithelial cell-derived a disintegrin and metalloproteinase-17confers resistance to colonic inflammation through EGFR activation.EBioMedicine 2016;5:114–24.

52. Rios-Doria J, Sabol D, Chesebrough J, Stewart D, Xu L, Tammali R, et al. Amonoclonal antibody to ADAM17 inhibits tumor growth by inhibitingEGFR and Non-EGFR-Mediated pathways. Mol Cancer Ther 2015;14:1637–49.






2017;23:549-561. Published OnlineFirst August 3, 2016.Clin Cancer Res Reba Mustafi, Urszula Dougherty, Devkumar Mustafi, et al.

associated Colon Cancer−Diet ADAM17 is a Tumor Promoter and Therapeutic Target in Western

Updated version

10.1158/1078-0432.CCR-15-3140doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2016/08/03/1078-0432.CCR-15-3140.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/23/2/549.full#ref-list-1

This article cites 51 articles, 23 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/23/2/549.full#related-urls

This article has been cited by 5 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://clincancerres.aacrjournals.org/content/23/2/549To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-15-3140

http://clincancerres.aacrjournals.org/content/suppl/2016/08/03/1078-0432.CCR-15-3140.DC1

http://clincancerres.aacrjournals.org/content/23/2/549.full#ref-list-1

http://clincancerres.aacrjournals.org/content/23/2/549.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/23/2/549


Documents

ADAM17 is a Tumor Promoter and Therapeutic Target in ...Cancer Therapy: Preclinical ADAM17 is a Tumor Promoter and Therapeutic Target in Western Diet–associated Colon Cancer Reba