of December 1, 2014.This information is current as

Gelatinases in the Pathogenesis of ColitisExperimental Colitis: Contrasting Role ofMetalloproteinase-2 Exacerbates Selective Ablation of Matrix

T. Gewirtz, Didier Merlin and Shanthi V. SitaramanBockbrader, Steven Epstein, Matam Vijay-Kumar, Andrew Pallavi Garg, Mauricio Rojas, Anupama Ravi, Katrina

http://www.jimmunol.org/content/177/6/4103doi: 10.4049/jimmunol.177.6.4103

2006; 177:4103-4112; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/177/6/4103.full#ref-list-1

, 24 of which you can access for free at: cites 58 articlesThis article

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2006 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on D

ecember 1, 2014

http://ww

w.jim

munol.org/

Dow

nloaded from

Selective Ablation of Matrix Metalloproteinase-2 ExacerbatesExperimental Colitis: Contrasting Role of Gelatinases in thePathogenesis of Colitis1

Pallavi Garg,* Mauricio Rojas,† Anupama Ravi,* Katrina Bockbrader,* Steven Epstein,*Matam Vijay-Kumar,‡ Andrew T. Gewirtz,‡ Didier Merlin,* and Shanthi V. Sitaraman2*‡

The matrix metalloproteinases (MMPs), MMP-2 and MMP-9, share structural and substrate similarities and are up-regulatedduring human as well as animal models of inflammatory bowel disease. We recently demonstrated that epithelial-derived MMP-9is an important mediator of inflammation and tissue damage in colitis. In this study, we examined the role of MMP-2 in acutecolitis. Colitis was induced using two models, administration of dextran sodium sulfate (DSS) and Salmonella enterica subsp.serovar Typhimurium (S.T.). Bone marrow chimeras were performed using bone marrow cells from wild-type (WT) andMMP-2�/� mice. Colitis was evaluated by clinical symptoms, myeloperoxidase assay, and histology. MMP-2 protein expressionand activity were up-regulated in WT mice treated with DSS or S.T. MMP-2�/� mice were highly susceptible to the developmentof colitis induced by DSS (or S.T.) compared with WT. During inflammation, MMP-2 expression was increased in epithelial cellsas well as in the infiltrating immune cells. Bone marrow chimera demonstrated that mucosa-derived MMP-2 was required for itsprotective effects toward colitis. Furthermore, we demonstrate that severe colitis in MMP-2�/� is not due to a compensatoryincrease in MMP-9. Finally, we show that MMP-2 regulates epithelial barrier function. In contrast to MMP-9, mucosa-derivedMMP-2 may be a critical host factor that is involved in the prevention or cessation of the host response to luminal pathogens ortoxins, an important aspect of healing and tissue resolution. Together, our data suggest that a critical balance between the twogelatinases determines the outcome of inflammatory response during acute colitis. The Journal of Immunology, 2006, 177: 4103–4112.

I nflammatory bowel diseases (IBD),3 which include ulcer-ative colitis and Crohn’s disease, are chronic, incurable in-testinal disorders. Ulcerative colitis and Crohn’s disease are

characterized by continuous or discontinuous mucosal inflamma-tion, respectively, with inflammatory cell infiltration, epithelial celldestruction, connective tissue defects, and ulceration of the mu-cosa of the major portion of intestine and/or colon. The etiopatho-genesis of IBD is poorly understood, but considered to be multi-factorial, involving genetic susceptibility and environmentalfactors that lead to the synthesis and release of inflammatory cy-tokines and recruitment of inflammatory cells culminating in tissuedestruction. Accumulating data from several studies indicate theinvolvement of matrix metalloproteinases (MMPs) in the patho-

genesis of IBD via their influence on the function and migration ofinflammatory cells, mucosal ulceration, as well as matrix deposi-tion and degradation (1–3). Clinical and experimental studies havedemonstrated an increase in the abundance of these matrix-degrad-ing proteases in IBD, and inhibition of MMP activation has beenshown to improve experimental colitis (4–7).

MMPs are a family of Zn2�-dependent extracellular matrix-de-grading endopeptidases that share common functional domainsand activation mechanisms (8–10). MMPs are the only mamma-lian enzymes capable of catalyzing cleavage of the interstitial col-lagen, types I, II, III, or IV, in their native triple helical form, at theneutral pH of the extracellular space (11). Depending on substratespecificity, amino acid similarity, and identifiable sequence mod-ules, the MMPs can be classified into four major subgroups: col-lagenases (MMP-1, -3, -8), gelatinases (MMP-2, -9), stromelysins(MMP-3, -7, -10, -12), and membrane-type metalloproteinases 1through 5. Although MMPs play important roles in normal tissueremodeling, dysregulated expression has been implicated in sev-eral pathological processes, such as arthritis (1, 12), atherosclero-sis, myocardial infarction (13), colorectal cancer, tumor invasion(14, 15), and IBD (4, 16). In many acute and chronic inflammatoryevents such as IBD, MMP levels have been shown to be up-reg-ulated (17–23).

Among the MMPs, MMP-2 and MMP-9 are the two knowngelatinases that degrade denatured collagen and are consistentlyup-regulated during active flares of IBD in human as well as inanimal models of colitis (17, 18, 23–26). Gelatinases differ fromother MMPs in terms of their structure as well as substrate spec-ificity. Both MMP-2 and MMP-9 have the prototypic structure thatis present in other MMPs: signal peptide, followed by the prodo-main, catalytic domain, hinge region, and the hemopexin domain,

*Division of Digestive Diseases, Emory University, Atlanta, GA 30322; †Division ofPulmonary, Allergy, and Critical Care Medicine, Department of Medicine, EmoryUniversity, Atlanta, GA 30322; and ‡Department of Pathology, Emory University,Atlanta, GA 30322

Received for publication April 14, 2006. Accepted for publication June 19, 2006.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institute of Diabetes and Digestive and Kid-ney Diseases Grant DK06411, Crohn’s and Colitis Foundation of America SeniorResearch Award and Elseivier Research Initiative Award (to S.V.S.), National Insti-tute of Diabetes and Digestive and Kidney Diseases Grant DK061941 and DK071594(to D.M.), and Digestive Disease Research Center Grant 5R24DK064399-02.2 Address correspondence and reprint requests to Dr. Shanthi V. Sitaraman, Divisionof Digestive Diseases, Room 201-F, 615 Michael Street, Whitehead Research Build-ing, Emory University, Atlanta, GA 30322. E-mail address: ssitar2@emory.edu3 Abbreviations used in this paper: IBD, inflammatory bowel disease; DSS, dextransodium sulfate; MMP, matrix metalloproteinase; MPO, myeloperoxidase; S.T., Sal-monella enterica subsp. serovar Typhimurium; TIMP, tissue inhibitor of metallopro-teinase; WT, wild type.

The Journal of Immunology

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

respectively (1). However, they differ from other MMPs by thepresence of fibronectin-like repeats in their catalytic domain. BothMMP-2 and MMP-9 share common substrates and are expressedby similar cell types (1, 11). A major structural difference betweenMMP-2 and MMP-9 is the lack of an additional type V collagen-like domain in MMP-2 (27). Little is known regarding the role ofthe MMP-2 or MMP-9 in inflammation or injury of colonduring IBD.

We recently demonstrated that MMP-9 activity and protein ex-pression were absent from normal colonic mucosa, but were up-regulated during experimental colitis in response to luminal toxin(dextran sodium sulfate; DSS) as well as bacteria Salmonella en-terica subsp. serovar Typhimurium (S.T.) (24). MMP-9�/� miceexposed to DSS or S.T. had significantly reduced inflammationand mucosal injury and were protected against acute colitis. Im-mune response to systemic administration of S.T. was also notaffected in MMP-9�/� mice. Finally, epithelial- but not immunecell-derived MMP-9 was required for tissue damage. The roleand function of MMP-2 in the pathogenesis of intestinal inflam-mation are not known. In this study, we sought to characterizethe role of MMP-2 in the pathogenesis of experimental acutecolitis.

Materials and MethodsExperimental animals

The Animal Care Committee of Emory University approved all proceduresperformed on animals and the procedures were in accordance with theGuide for the Care and Use of Laboratory Animals, published by the U.S.Public Health Service. MMP-2�/� mice (C57BL/6 background) were a giftfrom L. Matrisian (Vanderbilt University, Nashville, TN). The homozy-gous MMP-2-deficient mice used were progeny of heterozygous breedingpairs of C57BL/6 background with disruption of the MMP-2 gene thatwere backcrossed for more than six generations (28). These mice devel-oped normally and were fertile. To confirm the absence of MMP-2 geneexpression, genomic DNA was isolated and the disruption of the MMP-2gene was confirmed via PCR using primers designed to specifically detecthetero- and homozygote mice. For MMP-2 wild type (WT), we used prim-ers 5�-CAACGATGGAGGCACGAGTG-3� and 5�-CCGGGGAACTTGATCATGG-3� (29). For MMP-2�/� mice, we used primer 5�-TGCAAAGCGCATGCTCCAGA-3� and primer 5�-TGTATGTGATCTGGTTCTTG-3� (29). As described by Perez et al. (29), PCR conditions used were:one cycle of 94°C for 5 min, followed by 34 cycles of 94°C for 1 min, 59°Cfor 1 min, 72°C for 1 min and 30 s, and one cycle of 72°C for 7 min. Thelack of MMP-2 protein and activity was confirmed by Western blot andgelatin zymography, respectively, among MMP-2�/� mice. Age- and sex-matched WT and MMP-2�/� littermates used in the study were between 6and 8 wk old at the beginning of the experimental protocol and were main-tained under conditions as described previously (24).

Induction of DSS colitis

Colitis was induced in two groups of age- and sex-matched male and fe-male WT and MMP-2�/� littermates, by oral administration of DSS (ICNBiomedicals) at 3% (w/v) in tap water ad libitum for 6 days. Age-matchedmale and female WT and MMP-2�/� littermates receiving tap water servedas control. Mice were observed daily and evaluated for changes in bodyweight and development of clinical symptoms.

S.T. infection

Gut-restricted S.T. infection was induced, as described previously (24, 30).To prepare S.T. inocula, bacteria (S.T. SL3201) were grown overnight at37°C in 10 ml of Luria-Burtani broth in a 20-ml container with shaking(150 rpm) and were then used to inoculate fresh medium (1:100) and weregrown under the same conditions for 2–3 h until an OD at 550 nm of0.35–0.6 was reached. Bacterial cultures were then diluted in normal sa-line, and the CFU were enumerated by plating a dilution series of theinoculum. Water and food were withdrawn 4 h before treating with 7.5 mgof streptomycin (75 �l of sterile water containing streptomycin or 75 �l ofsterile water by gavage). Afterward, animals were supplied with food andwater ad libitum. At 20 h after streptomycin treatment, food and water werewithdrawn again for 4 h before mice were infected with 108 CFU of S.T.

(50-�l suspension in PBS) or treated with vehicle. Thereafter, food andwater were offered immediately. Mice were sacrificed by CO2 inhala-tion, and tissue samples were processed, as described for the DSS colitismodel (6).

Gelatin zymography

The activity of MMPs is measured by zymography under nonreducingconditions, as described previously (24, 31). Briefly, snap-frozen samplesof colon were homogenized with homogenizer and extracted in ice-coldnonreducing extraction buffer (20 �l/mg tissue). After centrifugation, thesupernatant was collected and the total protein concentration was measuredby Lowry method using protein assay reagent (Bio-Rad). Protein fromtissue samples was electrophoretically separated on 7.5% polyacrylamideagarose gel copolymerized with gelatin (1.5 mg/ml) as a substrate. The gelswere washed three times for 15 min each with 2.5% Triton X-100 (Sigma-Aldrich) to remove the SDS and to allow the electrophoresed enzymes torenature, before being incubated in zymography buffer (5 mmol/L CaC12

and 50 mmol/L Tris-HC1 (pH 7.5)) for 18 h at 37°C. The gels were thenstained with 0.5% Coomassie brilliant blue R-250 (Bio-Rad) and destainedwith methanol:acetic acid:water (v/v 4:1:5). Prestained standard high-rangeprotein markers (Bio-Rad) were used to determine the molecular weightsof the gelatinases. Clear digested regions representing MMP activity werequantified using an imaging densitometer (Camera Imaging Densitometer;Model 8300; Alpha Innotech).

Protein extraction and Western blot analysis

As described previously (24), for Western blot analysis, colon tissues ob-tained as above were homogenized and extracted with lysis buffer. Sampleswere then centrifuged at 12,000 rpm for 10 min at 4°C and the resultingsupernatant was used for assays. The total protein concentration of allsamples was measured by Lowry method using protein assay reagent (Bio-Rad). A total of 40 �g was boiled for 5 min in Laemmli’s sample buffer(Bio-Rad) and electrophoresed in 10% SDS-PAGE gels. Proteins weretransferred to nitrocellulose (Bio-Rad), and the membrane was thenblocked in 5% nonfat dry milk for 1 h. Incubation was performed overnightat 4°C with Abs for tissue inhibitor of metalloproteinase-1 (TIMP-1) (1:1000), TIMP-2 (1:1000), MMP-2 (5 �g/ml), and MMP-9 (1:2000)(Chemicon International). Subsequently, the membranes were washedwith Tris-NaCl-Tween 20 and incubated with a goat anti-rabbit(1:2500) or goat anti-mouse (1:4000) IgG HRP conjugate (Bio-Rad) for1 h at room temperature. Membranes were developed with WesternLightning Chemiluminescence Reagent Plus (PerkinElmer) and quan-tified by image analysis (32).

Clinical activity score

Assessment of body weights, stool consistency, and the presence of occult/gross blood by a guaiac test (Hemoccult Sensa; Beckman Coulter) weredetermined daily for each mouse. Colitis was quantified with a clinicalscore, as described by Cooper et al. (33), using the parameters of weightloss, stool consistency, and fecal blood. Briefly, no weight loss was con-sidered as 0 points, weight loss of 1–5% was scored 1 point, loss of 5–10%as 2 points, 10–20% weight loss as 3 points, and a loss of �20% of theweight was scored as 4. The stool character was characterized as normal(0), soft with well-formed pellets (1), soft without pellets (2), or diarrhea(4). For occult blood, no blood was scored 0, positive hemoccult scored as2 points, and gross bleeding was scored 4. The total score was added to geta clinical activity score ranging from 0 to 12. Six days after the inductionof colitis, mice were euthanized by CO2/hypothermia. The abdominal cav-ity was exposed by a midline laparotomy, and the entire colon was re-moved from the caecum to the anus. The colon was flushed with cold PBSand opened longitudinally for morphologic studies. The length and weightof the colon were measured, and tissue obtained from each colon wasprocessed for further assays.

Histological assessment of colitis

Colonic specimens obtained as above were fixed in formalin and coded forblind microscopic assessment of mucosal lesions (descending colon forDSS colitis and caecum for S.T. colitis). Sections were stained with H&E.Microscopic sections were analyzed and histologic scoring was performed,as described by Cooper et al. (33), based on three variables, according tothe severity of the induced damage. Briefly, for inflammation, rare inflam-matory cells in the lamina propria were counted as 0; increased numbers ofgranulocytes in the lamina propria as 1; confluence of inflammatory cellsextending into the submucosa as 2; and a score of 3 was given for trans-mural extension of the infiltrate. For crypt damage, intact crypt was scored0, loss of 1/3 basal counted as 1, loss of 2/3 basal was counted as 2, entire

4104 SELECTIVE ABLATION OF MMP-2 EXACERBATES EXPERIMENTAL COLITIS

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

crypt loss was scored as 3, change of epithelial surface with erosion as 4,and a score of 5 was given for confluent erosion. For evaluation of ulcers,an absence of ulcer was scored 0, 1–2 foci of ulcerations were scored as 1,3–4 foci of ulcerations were scored as 2, and confluent/extensive ulcerationwas scored 3. These values were added to give a total histological scoreof 11.

Myeloperoxidase (MPO) activity in the colon

Neutrophil infiltration into colon was quantified by measuring MPO activ-ity, as described previously (24, 34). Briefly, a portion of colon was ho-mogenized in 1:20 (w/v) of 50 mM phosphate buffer (pH 6.0) containing0.5% hexadecyltrimethyl ammonium bromide (Sigma-Aldrich), on ice us-ing a Polytron homogenizer. The homogenate was sonicated for 10 s,freeze-thawed three times, and centrifuged at 14,000 rpm for 15 min.Supernatant (14 �l) was added to 1 mg/ml o-dianisidine hydrochloride(Sigma-Aldrich) and 0.0005% hydrogen peroxide, and the change inabsorbance at 460 nm was measured. One unit of MPO activity was definedas the amount that degraded 1 �mol peroxidase per minute at 25°C. Theresults were expressed as absorbance per gram of tissue.

Immunofluorescence

Colon tissue from WT and MMP-2�/� mice embedded in paraffin wasobtained, as described by Castaneda et al. (24). The sections were rehy-drated using graded alcohols. Sections were treated with 0.5% TritonX-100 � 0.08% saponin in PBS at room temperature for 35 min. Thesections were rinsed in PBS and incubated with rhodamine-phalloidin (1/60) diluted in PBS for 40 min at room temperature. Sections were subse-quently blocked with 2% BSA for 1 h at room temperature. Sections wereincubated with primary Ab, anti-MMP-2 (Chemicon International; 15�g/ml in 2% BSA in PBS solution), or rabbit IgG (Sigma-Aldrich; con-trol), or anti-actin (Sigma-Aldrich; 1:50,000 in 2% BSA) for 1 h at roomtemperature. After washing three times with PBS, they were incubated withFITC-conjugated anti-rabbit secondary Ab (Bio-Rad; 1/100 in 2% BSA inPBS solution) or rhodamine-conjugated anti-mouse secondary Ab (Bio-Rad; 1/100 in 2% BSA in PBS solution), for 45 min at room temperatureand then mounted with Slow Fade (Molecular Probes) mounting mediumand examined using Zeiss LSM microscope.

Bone marrow transplantation

Bone marrow transplantation was performed, as described previously (24,35). Briefly, the femur and tibia were removed and stripped off all muscleand sinew, and bone marrow cells were harvested from both femurs andtibias by flushing the bone cavity with basal marrow medium (Iscove’smedium; Cambrex). After washing with PBS, bone marrow cells wereresuspended in basal marrow medium. Approximately 5 � 106 cells in 50�l were transplanted retro-orbitally. Four treatment groups with 10 animalsper group were used (WT3WT, WT3MMP-2�/�, MMP-2�/�3WT,and MMP-2�/�3MMP-2�/�). Mice were given neomycin at 2 mg/ml forthe first week of posttransplantation, after which they were switched to tapwater. Engraftment was verified by genotyping bone marrow cells usingspecific primers (29), as described previously in this work. Five weeksafter transplantation, we induced colitis by 3% DSS and mice wereassessed daily for rectal bleeding, weight loss, and diarrhea, and theclinical score was obtained. At the end of the experimental period (6days for DSS) (24), mice were sacrificed and colons were processed forhistology, MPO assay, and immunohistochemistry. Engraftment wasverified by performing genotyping on the bone marrow cells at the endof 4 wk after transplantation.

In vivo permeability

In vivo permeability assay to assess barrier function was performed usingan FITC-labeled dextran method, as described (36). Briefly, 8- to 10-wk-old WT and MMP-2�/� mice were used. Food and water were withdrawnfor 4 h and mice were gavaged with permeability tracer (60 mg/100 g bodyweight of FITC-labeled dextran, m.w. 4000; Sigma-Aldrich). Serum wascollected retro-orbitally; fluorescence intensity of each sample was mea-sured (excitation, 492 nm; emission, 525 nm; Cytofluor 2300; Millipore,Waters Chromatography); and FITC-dextran concentrations were deter-mined from standard curves generated by serial dilution of FITC-dextran.Permeability was calculated by linear regression of sample fluorescence(Excel 5.0; Microsoft).

Statistical analysis

The data are presented as mean � SEM. Statistical analysis was performedusing GraphPad Instat 3 software (�www.graphpad.com�). Groups were

compared using nonparametric tests, and significance of the differencesbetween groups was assessed using the Mann-Whitney U test or the Wil-coxon signed-ranks test for paired data. Values of p 0.05 were consid-ered statistically significant.

ResultsMMP-2 expression and activity are induced during DSS colitisin mice

DSS colitis model, a rapid and reproducible model of colitis thatmimics human IBD in several respects, was used to investigate theexpression of MMP-2 during colitis. WT C57BL/6 mice were ad-ministered 3% DSS in their drinking water (w/v) for 6 days. Micewere weighed daily, and stool was examined for consistency andthe presence of blood. Mice were euthanized after 6 days, andprotein lysates were prepared from the colon for zymographyand Western blot analysis, as described in Materials and Methods.Zymography was performed using a gelatin-impregnated PAGE toevaluate gelatinolytic activity. A representative zymogram and itsdensitometric analysis are shown in Fig. 1A. Pro-MMP-2 activitywas observed in colonic extracts prepared from WT mice that hadbeen fed with water (Fig. 1A, lanes 1–4). There was no activeMMP-2 detected in these mice. In contrast to mice fed water, co-lonic extracts from WT animals given DSS showed strong induc-tion of pro- and active MMP-2 activity (pro, 2 � 0.3-fold andactive, 7.5 � 3.9-fold over water-treated mice, respectively;mean � SE, 6 mice/group; Fig. 1A, lanes 5–9). MMP-2 activitysteadily increased over the course of DSS administration startingon day 1 (data not shown).

Western blot analysis of colonic protein extracts from WT micetreated with water or DSS was performed to confirm the identity ofgelatinolytic activity observed on zymography. Animals that hadnot been exposed to DSS showed constitutive MMP-2 protein ex-pression (Fig. 1B, lanes 1–4). Marked induction of MMP-2 proteinexpression was observed in colitic samples (Fig. 1B, lanes 5–8).WT mice exhibited immunoreactivity at 72 kDa consistent withpro-MMP-2, while the colitic samples showed immunoreactivity ata slightly lower molecular mass at 62 kDa consistent with activeMMP-2. Densitometric analysis revealed 8.2 � 4.1-fold increasein MMP-2 protein expression in extracts from mice exposed toDSS compared with mice fed water (mean � SE, 6 mice/group;p 0.01).

The proteolytic activity of MMPs is regulated by the balancebetween zymogen activation and enzyme inhibition through theirendogenous inhibitors, �-macroglobulins and TIMPs (16). Giventhe importance of TIMPs on the biological activities of MMPs, wenext examined whether the expression of TIMP-1 and TIMP-2 wasinfluenced differentially by DSS administration. Fig. 2 shows theexpression levels of TIMP-1 and TIMP-2 in WT mice fed withDSS or water. TIMP-1 levels were unchanged (Fig. 2), whileTIMP-2 levels were decreased in mice treated with DSS (Fig. 2,lanes 5–8). Together, these data suggest that MMP-2 activity andprotein expression are induced during DSS colitis along with asignificant decrease in TIMP-2.

MMP-2�/� mice developed severe inflammation in response toDSS and were unable to recover from DSS-induced colitis

To further investigate the role of MMP-2 in the pathogenesis ofcolitis, we used C57BL/6 mice with targeted deletion of MMP-2(28). These mice exhibit normal phenotype. We administered DSSin drinking water to age (8–10 wk)- and sex-matched C57BL/6WT and homozygous MMP-2�/� mice. The mice were comparedfor the clinical signs of disease, including weight changes, stoolconsistency, and occult blood, according to grading system previ-ously described (24, 33). Both WT and MMP-2�/� mice exposed

4105The Journal of Immunology

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

to DSS developed signs of colitis within 6 days after the admin-istration of DSS. However, colitis was significantly worse inMMP-2�/� mice given DSS, as evidenced by a clinical diseaseactivity score of 7.9 � 1 and 11.1 � 0.5, respectively (mean � SE,6 mice/group; p 0.01; Fig. 3A). Almost all of the MMP-2�/�

mice developed diarrhea after day 3. These mice were hemo-occultpositive starting from day 2 and exhibited frank bleeding on day 5after DSS administration. DSS-induced colitis is characterized bythe presence of inflammation of the colon manifested by cryptdestruction, mucosal damage, epithelial erosions, and infiltrationof inflammatory cells into the mucosal tissue. Tissues collectedfrom WT and MMP-2�/� mice exposed to DSS were examinedhistologically and compared with those from normal controls. Asshown in Fig. 3, B and C, a mean histological score of 9.7 � 0.7was observed in MMP-2�/� mice given DSS compared with amean score of 6.8 � 0.6 in WT mice given DSS (mean � SE, 6mice/group; p 0.01 compared with MMP-2�/� mice). Interest-ingly, ulceration was significantly increased in MMP-2�/� mice.The ulcers were not only more in number, but involved largersurface area compared with WT mice (Fig. 3C). Histological signsof inflammation were not detected in the water control groups(data not shown). Thus, the data obtained corroborated the resultsobtained from clinical analysis and confirmed the deleterious effectof targeted MMP-2 deletion toward the development of colitis.These data also suggest a potential role of MMP-2 as a protectiveMMP against colitis.

To monitor the effect of lack of MMP-2 on the healing phase ofcolitis, we observed mice for 1 more additional wk after DSStreatment, during which time animals were given water. All of theMMP-2�/� mice died during this period, while only 25% of theWT mice died. Fig. 4A shows the percentile of change in bodyweight during healing phase and indicates that change in bodyweight of MMP-2�/� was 2-fold more than WTs. Fig. 4B showsthe survival curve of WT and MMP-2�/� mice, indicating higher

mortality among WTs compared with knockouts during recoveryphase. These data implicate the importance of MMP-2 in intestinalrecovery and healing following acute colitis.

MMP-2 mice exhibit increased susceptibility to Salmonella-induced colitis

As an alternate model of colitis, we used oral infection with S.T.,during which S.T. is administered after pretreatment of mice with

FIGURE 2. TIMP-2, but not TIMP-1, expression is down-regulatedduring DSS colitis. C57BL/6 mice were given water or 3% DSS (w/v)for 6 days, at which time mice were euthanized. Colon was harvestedand processed for Western blot. Representative Western blot of proteinfrom the colon of WT given water (lanes 1– 4) or DSS (lanes 5– 8)probed with anti-TIMP-1 and anti-TIMP-2 (1:1000) (Chemicon Inter-national), respectively. Each lane shows protein (30 �g/lane) from anindividual mouse with or without DSS treatment. Western blots werequantitated by scanning densitometry, and densities relative to water-treated controls are graphed. Values are representative of three indi-vidual experiments, mean � SE; n 5 for the DSS-treated groups, andn 4 for the control group in each of the experiments. Significantlydifferent from mice given water; �, p 0.05.

FIGURE 1. MMP-2 protein activity and expression are up-regulated during DSS colitis. C57BL/6 mice were given water or 3% DSS (w/v) for 6 days,at which time mice were euthanized. Colon was harvested and processed for gelatin zymography (A) and Western blot (B). A, Representative zymogramof protein (30 �g/lane) from the colon of an individual WT mouse with (A, lanes 5–9) or without (A, lanes 1– 4) DSS treatment shows gelatinolyticband in each lane. Zymogram was quantitated by scanning densitometry, and densities relative to vehicle-treated controls are graphed. Values arerepresentative of three individual experiments, mean � SE; n 5 for the DSS-treated groups, and n 4 for the control group in each of theexperiments. B, Representative Western blot of protein from the colon of WT given water (lanes 1– 4) or DSS (lanes 5– 8) probed with anti-MMP-2(5 �g/ml) (Chemicon International). Each lane shows protein (30 �g/lane) from an individual mouse with or without DSS treatment. Western blotswere quantitated by scanning densitometry, and densities relative to water-treated controls are graphed. Values are representative of three individualexperiments, mean � SE; n 5 for the DSS-treated groups, and n 4 for the control group in each of the experiments. Significantly different frommice given water; �, p 0.05; ��, p 0.001.

4106 SELECTIVE ABLATION OF MMP-2 EXACERBATES EXPERIMENTAL COLITIS

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

streptomycin. In this model, S.T. induces clinical and histologicalfeatures of enterocolitis predominantly involving the caecum (24,30). We chose this model because it recapitulates some aspects ofclinical and histological human infection as well as acute flares ofIBD, wherein mucosal-pathogen interaction is thought to play animportant role in the pathogenesis. The characteristic histologicalfeature of gut-restricted S.T. enteritis includes neutrophil infiltra-tion of the intestinal mucosa, the hallmark of infectious colitis, aswell as acute flares of IBD. Other histological features of S.T.colitis include epithelial ulceration, edema, and induction ofICAM-1. WT and MMP-2�/� mice were pretreated with strepto-mycin and then administered S.T. Mice were sacrificed at 24 and48 h after the administration of S.T., and colonic tissue was re-moved, weighed with the contents, and photographed. The caecumwas processed for histology and MPO activity. The caeca of all themice infected with S.T. appeared pale and shriveled to a small sizeand were filled with purulent exudates. MMP-2 activity was sig-nificantly up-regulated in WT mice treated with S.T. (Fig. 5A,lanes 3 and 4) similar to that seen in WT mice during DSS colitis.MMP-2 activity was not detected in the MMP-2�/� mice with(Fig. 5A, lanes 7 and 8) or without (Fig. 5A, lanes 5 and 6) treat-ment of S.T. MPO activity reflected the clinical observation (Fig.5B). MMP-2�/� mice showed significantly increased MPO activ-

FIGURE 3. MMP-2�/� develop severe DSS-induced colitis. Eight- to 10-wk-old WT C57BL/6 and MMP-2�/� mice were weighed and given water orDSS (3% w/v) for 6 days. Mice were euthanized, and colon was processed for histology. Disease severity was assessed and expressed in terms of clinicalactivity score in A. Clinical score was calculated using parameters of stool consistency, weight loss, and fecal blood. Each bar represents mean � SE; �,p 0.02. B, mean histological score; �, p 0.02 and #, p 0.05 compared with WT. C, representative sections from two individual experiments. n 5 for DSS group, and n 4 for control group. Colon specimens from descending colon were fixed in formalin, and histological score was calculated onH&E sections based on three variables: inflammatory cells, crypt damage, and ulcers. Both WT and MMP-2�/� mice developed inflammatory infiltrationand crypt damage. However, MMP-2�/� mice showed extensive ulceration compared with WT mice.

FIGURE 4. MMP-2�/� mice exhibit impaired recovery from DSS-in-duced colitis. WT and MMP-2�/� mice were given water or DSS for 6days, as described in Fig. 3 (n 4 per group). After 6 days, mice weregiven water and followed for weight loss (A) and mortality (B) duringrecovery phase. Each bar represents mean � SE; #, p 0.05.

FIGURE 5. MMP-2�/� mice showed increased susceptibility to Salmo-nella-induced colitis. WT and MMP-2�/� mice (C57BL/6) were pretreatedwith streptomycin before the administration of S.T. and sacrificed 48 hafter the administration of S.T. The caecum and colon were harvested.Western blot analysis (A) of protein from the caecum of WT given water(lanes 1 and 2) or S.T. (lanes 3 and 4) and MMP-2�/� given water (lanes5 and 6) or S.T. (lanes 7 and 8) probed with anti-MMP-2 was done. Eachlane shows protein (40 �g/lane) from an individual mouse with or withoutS.T. treatment. �-actin served as the loading control. Colons were snapfrozen in liquid nitrogen, and MPO activity was measured (B) as an indexof neutrophil infiltration into the injured tissue. WT mice are representedby f, and MMP-2�/� mice by o. Each bar represents mean � SE. n 3animals for each group. �, p 0.005. C, H&E sections based on threevariables: inflammatory cells, crypt damage, and ulceration showed thatboth WT and MMP-2�/� mice developed inflammatory infiltration andcrypt damage to S.T. induced colitis though knockouts showed more ul-ceration compared to WT.

4107The Journal of Immunology

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

ity compared with WT mice. Clinical symptoms and histologicalchanges appeared at 24 h in WT and MMP-2�/� mice. However,MMP-2�/� mice showed increased severity of inflammation com-pared with WT mice at 48 h (Fig. 5C) as marked by leukocyteinfiltration with loss of crypts as well as ulcerations at 48 h inMMP-2�/� mice treated with S.T. Taken together, these data sug-gest that MMP-2 is up-regulated during S.T.-induced colitis andMMP-2�/� mice develop severe S.T.-induced colitis comparedwith WT mice.

MMP-2 localizes to epithelial cells and immune cells

We next determined the localization of MMP-2 in normal andinflamed colon. Immunofluorescence of colon sections performedwith affinity-purified MMP-2 Ab directed against the catalytic do-main of mouse MMP-2 confirmed the presence of this enzymemainly in epithelial cells (Fig. 6A). There was no staining in thelamina propria immune cells (Fig. 6A, boxed). In mice fed withDSS, MMP-2 staining was strongly induced and was seen in thecrypt epithelial cells (Fig. 6B). Strong MMP-2 staining was alsodetected in the infiltrating immune cells (Fig. 6C). Control sectionstained with isotype Ab (anti-rabbit IgG) instead of primary Abshowed no staining. There was no staining visualized in MMP-2�/� mice (Fig. 6D). Thus, the expression of MMP-2 is induced inmice exposed to DSS and is localized to epithelial and immunecells.

Mucosa-derived MMP-2 mediates its protective effect

To understand the role of MMP-2 in the pathogenesis of colitis, wefirst determined the relative contribution of mucosal vs immunecell-derived MMP-2 in the development of colitis. To explore thisquestion, we generated bone marrow chimeras of WT or MMP-2

mice. Four treatment groups with 10 animals per group (D:WT3R:WT, D:WT3R:MMP-2�/�, D:MMP-2�/�3R:WT,D:MMP-2�/�3R:MMP-2�/�; D, donor; R, recipient) were in-cluded. Bone marrow transplantation was done, as described inMaterials and Methods. All mice recovered uneventfully frombone marrow transplantation, except 20% of the MMP-2�/

�3MMP-2�/� group died. Five weeks after bone marrow trans-plantation, mice were fed 3% DSS in drinking water. The micewere compared for the clinical signs of weight loss, stool consis-tency, and occult blood, and mice were sacrificed after 6 days ofDSS. Colitis was worst in all MMP-2�/� mice whether they re-ceived WT or MMP-2�/� bone marrow (Fig. 7, A and B). Thesemice had extensive crypt damage, epithelial ulceration, crypt ab-scess formation, and infiltration of inflammatory cells and neutro-phils in the mucosa (Fig. 7C). Fig. 7A shows the clinical score ofbone marrow chimeras, indicating that MMP-2�/� mice receivingbone marrow cells from MMP-2�/� or WT mice had higher scores(9.1 � 0.3 and 11.4 � 0.4, respectively) compared with WT micereceiving bone marrow cells from WT or MMP-2�/� mice (6.4 �0.5 and 7.1 � 0.5, respectively). MPO assay (Fig. 7B) reflected thesame fact that MMP-2�/� mice that received MMP-2�/� or WTbone marrow exhibited significantly increased MPO activity andobvious signs of colonic inflammation and tissue damage com-pared with WT mice that received WT or MMP-2�/� bone mar-row. To verify the efficiency of bone marrow reconstitution, weperformed immunofluorescence staining of DSS-treated chimericmice. As shown in Fig. 7D, MMP-2�/� recipients of WT bonemarrow showed MMP-2 staining restricted to immune cells, whileWT recipient of MMP-2�/� bone marrow showed staining re-stricted to the epithelium. These data suggest that reconstitution ofthe transplanted bone marrow was efficient. Collectively, thesedata demonstrate that immune cell-derived MMP-2 is neither re-quired for migration nor sufficient to induce tissue damage in DSS-induced colitis. Furthermore, these data suggest that mucosalMMP-2 expression during colitis is required for its protective ef-fect on the development of colitis.

Severe colitis in MMP-2�/� mice is not due to a compensatoryincrease in MMP-9

We recently demonstrated that MMP-9 is up-regulated during ex-perimental colitis and mucosa-derived MMP-2 mediates tissuedamage during colitis. MMP-9�/� are protected from DSS- andS.T.-induced colitis (24). To verify whether a compensatory in-crease in MMP-9 among MMP-2�/� mice is contributing to theseverity of colitis seen in these mice, we performed Western blotin the colonic tissue lysates of WT and MMP-2�/� mice admin-istered water or DSS. Fig. 8, A and B, shows a representativeWestern blot of WT and MMP-2�/�, respectively. In contrast toWT mice fed water (Fig. 8A, lanes 2–4), colonic extracts from WTmice given DSS (Fig. 8A, lanes 5 and 6) showed increased MMP-9expression. Interestingly, MMP-9 expression was not increased inMMP-2�/� mice fed with water (Fig. 8B, lanes 2–4) and admin-istration of DSS increased the expression of MMP-9 (Fig. 8B,lanes 5 and 6) similar to WT mice given DSS. These data suggestthat a compensatory increase in MMP-9 does not play a role in theseverity of colitis seen with the deletion of MMP-2.

MMP-2�/� mice exhibit compromised epithelial barrier function

Because our data show that epithelial-derived MMP-2 plays a pro-tective role in the development of colitis, we hypothesized thatMMP-2 may play a role in epithelial barrier function and barrierfunction may be decreased in MMP-2�/� mice, rendering themsusceptible to injury and inflammation. Loss of barrier functionprovided by epithelial cells is thought to be the initial inciting

FIGURE 6. MMP-2 localizes to the epithelial and immune cells.C57BL/6 mice were weighed and randomized into four groups: WT wateror DSS (3% w/v) and MMP-2�/� water or DSS (3% w/v) for 6 days. Micewere euthanized, and colon was processed for immunofluorescence usingMMP-2-specific Ab (Chemicon International; 15 �g/ml). ImmunoreactiveMMP-2 was revealed by FITC-conjugated secondary Ab. Shown here arerepresentative sections from three individual experiments. n 5 for DSSgroup, and n 4 for control group. MMP-2 staining in WT mice fed water(A), WT mice fed 3% DSS showing epithelial staining (B), and infiltratingimmune cells (C). Boxed area in A, lamina propria immune cells. MMP-2�/� mice showed no staining (D).

4108 SELECTIVE ABLATION OF MMP-2 EXACERBATES EXPERIMENTAL COLITIS

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

event that underlies injury and inflammation in many intestinaldisorders, including IBD (37, 38). Such barrier defects result in themigration of antigenic material, previously confined to the intes-tinal lumen, into the submucosa, exposing lamina propria immunecells to naive Ags, eliciting inflammatory response and epithelial

FIGURE 8. Severe colitis in MMP-2�/� mice is not due to a compensatoryincrease in MMP-9 expression. WT (A) and MMP-2�/� (B) mice (C57BL/6)were given water or 3% DSS (w/v) for 6 days, at which time mice wereeuthanized. Colon was harvested and processed for Western blot. Represen-tative Western blot of protein from the colon of WT and MMP-2�/� micegiven water (lanes 2–4) or DSS (lanes 5 and 6) probed with anti-MMP-9 isshown in A and B, respectively. Purified MMP-9 (4 ng/lane) was loaded inlane 1 in both A and B. Each lane shows protein (40 �g/lane) from an indi-vidual mouse with or without DSS treatment. �-actin served as the loading control.

FIGURE 9. MMP-2�/� mice exhibit compromised epithelial barrier func-tion. WT and MMP-2�/� mice (C57BL/6) were gavaged FITC-dextran (m.w.4000), as described in Materials and Methods. Serum was obtained retro-orbitally and processed for fluorescence. Data are represented as mg FITC/�gprotein/hour. Each bar represents mean � SE. n 10; p 0.01.

FIGURE 7. Bone marrow chimeras showed that mu-cosa-derived MMP-2 mediates its protective effect.Bone marrow transplantation was performed, as de-scribed in Materials and Methods (D, donor; R, recip-ient; and KO, MMP-2�/� mice). Mice were euthanized,and colon was processed for assessment of colitis, MPOassay, histology, and immunofluorescence. A, Clinicalscore showing that KO mice had the worst colitiswhether they have received WT or KO bone marrow.Each bar represents mean � SE. n 10, a or c vs b:p 0.01; c vs a: p 0.05. B, Shows the measure ofMPO activity that is an index of neutrophil infiltrationinto the injured tissue, and was performed from colonsof bone marrow chimeras, as described in Materialsand Methods. C, Histological assessment of colitis afteradministration of 3% DSS for 5 days, indicating exten-sive crypt damage, epithelial ulceration, crypt abscessformation, and infiltration of inflammatory cells andneutrophils in the mucosa of KO mice receiving bonemarrow from WT or KO mice compared with WT micereceiving bone marrow cells from WT or KO mice. Toppanel, �20 magnification; middle panel, �40 magni-fication; and lower panel, �60 magnification. D, Im-munofluorescence staining of DSS-treated chimericmice using MMP-2-specific Ab. ImmunofluorescentMMP-2 was revealed by rhodamine-conjugated sec-ondary Ab. Sections were counterstained with anti-actin and revealed by FITC-conjugated secondary Ab.

4109The Journal of Immunology

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

injury that characterize these diseases (37, 38). To test our hypoth-esis, we studied barrier function in WT and MMP-2�/� mice usinga FITC-labeled dextran method, as described in Materials andMethods. Mice were administered FITC-dextran by gavage, andfluorescence was quantitated in the serum at 4 and 24 h after theadministration of FITC-dextran. As shown in Fig. 9, WT miceshowed an FITC-dextran of 1.05 � 0.1 mg of FITC/�g protein/h.In comparison, there was �2-fold increase in FITC-dextran levelsin MMP-2�/� mice (1.8 � 0.03 mg of FITC/�g protein/h), sug-gesting decreased barrier function in these mice.

DiscussionIn this study, we demonstrate that although MMP-2 and MMP-9are both gelatinases that originate from similar epithelial sourcesas well as share structural and substrate similarities, they haveopposing effects on the development of acute colitis. Consistentwith other studies, we show that MMP-2, like MMP-9, is up-reg-ulated during colitis (4, 16, 24, 39). Using two models of acutecolitis, chemical- and bacteria-induced colitis, we demonstrate thatMMP-2�/� mice are highly susceptible to the development of co-litis compared with their WT counterparts. These results are incontrast to the effect of MMP-9 on colitis; MMP-9�/� are pro-tected from the development of chemical- or bacteria-induced co-litis. These results were surprising given that MMP-2 and MMP-9belong to the family of gelatinases that share structural and matrixsubstrate similarities. Both MMP-2 and MMP-9 possess the fi-bronectin repeat characteristic of the gelatinases. However,MMP-2 lacks the collagen V-like domain present in MMP-9, po-tentially changing their nonmatrix substrate specificity. For exam-ple, McQuibban et al. (40) reported contrasting effect of MMP-2and MMP-9 on MCP-3. MMP-2 degrades MCP-3, while MMP-9has no effect on the protein. The authors extrapolated the differencein MCP-3 degradation between MMP-2 and MMP-9 to the colla-gen V domain. The effect of MMP-2 on MCP-3 degradation andpossibly degradation of other chemoattractants is thought to un-derlie the protective effect of MMP-2 on inflammatory arthritis,myocardial damage, and corneal injury (11, 41–49). In addition todegradation of chemoattractants or chemokines, a role for MMP-2in tissue remodeling has been shown in wounded corneal epithelialcells. In this study, MMP-9 impaired remodeling, while MMP-2was required for proper tissue remodeling and wound healing.These authors further showed that fibroblast-derived MMP-2 wasrequired for stromal remodeling that is integral to corneal woundhealing. Along with the observations that myofibroblasts are animportant source of MMP-2 during intestinal inflammation (19)and are integral to epithelial wound healing, it is likely thatMMP-2 may play a role in tissue remodeling that is required forproper wound healing (50, 51).

Our data show that cellular constituents of intestinal mucosa,including epithelial cells as well as immune cells, express MMP-2in response to DSS or S.T. These observations are consistent withpublished data that MMP-2 is highly expressed in the intestinalepithelia during human IBD. However, to our knowledge, the roleof MMP-2 in the pathogenesis of colitis has not been demon-strated. We demonstrate that MMP-2 may be involved in the reg-ulation of epithelial barrier function. MMP-2�/� mice exhibitedbarrier dysfunction, as evidenced by increased FITC-dextran trans-location in these mice compared with their WT counterpart. Datafrom human and animal studies have demonstrated a central rolefor epithelial dysfunction in the pathogenesis of intestinal inflam-mation (37, 38). The current paradigm for the pathogenesis of IBDsupported by evidence from human and animal studies as well ascultured cell models proposes three key components to be neces-sary for the initiation and progression of disease: 1) disruption of

the epithelial barrier (direct effect on barrier or impaired healing inresponse to injury); 2) access of luminal contents to the laminapropria, that is, immune cells; and 3) an abnormal immune re-sponse (37, 38). Epithelial dysfunction is seen in patients muchearlier than histologic or clinical manifestation of IBD. It isthought that barrier dysfunction allows contents of the intestinallumen to mix freely with the contents of the lamina propria, elic-iting an immune/inflammatory response, which characterizes in-flammatory diseases of the intestine. Our data show that barrierintegrity is compromised in MMP-2�/� mice, which may possiblylead to their increased susceptibility to luminal toxins (DSS) orpathogens (S.T.).

There are several possible mechanisms by which MMP-2might affect barrier function. One possibility is that MMP-2 canmodule tight junction directly by associating with tight junctionproteins. A series of studies have shown that MMP-2 intimatelyassociates with claudins (52–54). For example, it has been dem-onstrated that MMP-2 localizes to tight junctions in Madin-Darby canine kidney cells and ectodomain of claudin-1 inter-acts with the catalytic domain MMP-2 (52). Claudin familyproteins are the major constituents of tight junction strands,which are directly involved in paracellular sealing (barrier func-tion) as well as in membrane domain differentiation (fencefunction) in epithelial and endothelial cell (55, 56). Hence, theassociation of MMP-2 with claudin may modulate paracellularpermeability. In support of this association, our data show thatoverexpression of MMP-2 increases barrier function measuredusing FITC translocation (our unpublished data). In this con-text, it is interesting that MMP-9, unlike MMP-2, has been as-sociated with increased paracellular permeability by cleavingextracellular domain of occludin while having no effect on clau-din (57). Such disparate effect of MMP-2 and MMP-9 on per-meability may account for their opposite effects on the patho-genesis of colitis.

Given our previous observation that MMP-9 mediates in-flammatory response and tissue damage, a possible explanationfor the severe colitis seen in MMP-2�/� mice could be relatedto a compensatory increase in MMP-9 in MMP-2�/� mice.Such a compensatory increase in MMP-9 has been shown tomediate severe inflammatory response in an autoimmune en-cephalitis model in MMP-2�/� mice (45). The potential role ofMMP-2 as a modulator of MMP-9 expression via the mem-brane-type 1-MMP/TIMP-2 complex has been suggested (45,58). However, we did not observe any increase in MMP-9 ex-pression or activity in MMP-2�/� mice fed water, and the in-crease in MMP-9 was modest after DSS colitis compared withWT mice fed DSS. Thus, expression of MMP-9 may be partiallymodulating the development of colitis in MMP-2�/� mice,while the loss of MMP-2 expression may play a crucial role inthe severity of colitis. However, further studies are required todelineate the precise mechanism by which MMP-2 modulatesits protective effects.

In summary, we demonstrate that MMP-2, a gelatinase thatshares structural and substrate similarities with MMP-9, plays aprotective role toward the development of acute colitis. We alsoshow that mucosal MMP-2 contributes to its protective effect, andimmune cell-derived MMP-2 is not required for its migration orprotective effect. Lastly, we show that MMP-2 is required for in-tact epithelial barrier function. Together, our data suggest thatMMP-2 plays a protective role in the development of murine co-litis possibly by contributing to barrier function. Furthermore, acritical balance between the two gelatinases determines the out-come of inflammatory response during acute colitis. Our data are

4110 SELECTIVE ABLATION OF MMP-2 EXACERBATES EXPERIMENTAL COLITIS

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

relevant in designing gelatinase-based therapeutic strategies for thetreatment of intestinal inflammation.

AcknowledgmentsWe thank Dr. Lynn Matrisian (Vanderbilt University, Nashville, TN) forproviding us C57BL/6 MMP-2�/� mice, and Dr. J. Xu (Emory University,Division of Pulmonary, Atlanta, GA) for helping with bone marrowchimeras.

DisclosuresThe authors have no financial conflict of interest.

References1. Sternlicht, M. D., and Z. Werb. 2001. How matrix metalloproteinases regulate

cell behavior. Annu. Rev. Cell. Dev. Biol. 17: 463–516.2. Parks, W. C., and S. D. Shapiro. 2001. Matrix metalloproteinases in lung biology.

Respir. Res. 2: 10–19.3. Parks, W. C., C. L. Wilson, and Y. S. Lopez-Boado. 2004. Matrix metallopro-

teinases as modulators of inflammation and innate immunity. Nat. Rev. Immunol.4: 617–629.

4. Naito, Y., and T. Yoshikawa. 2005. Role of matrix metalloproteinases in inflam-matory bowel disease. Mol. Aspects Med. 26: 379–390.

5. Naito, Y., T. Takagi, M. Kuroda, K. Katada, H. Ichikawa, S. Kokura, N. Yoshida,T. Okanoue, and T. Yoshikawa. 2004. An orally active matrix metalloproteinaseinhibitor, ONO-4817, reduces dextran sulfate sodium-induced colitis in mice.Inflamm. Res. 53: 462–468.

6. Medina, C., S. Videla, A. Radomski, M. W. Radomski, M. Antolin, F. Guarner,J. Vilaseca, A. Salas, and J. R. Malagelada. 2003. Increased activity and expres-sion of matrix metalloproteinase-9 in a rat model of distal colitis. Am. J. Physiol.284: G116–G122.

7. Di Sebastiano, P., F. F. di Mola, L. Artese, C. Rossi, G. Mascetta, H. Pernthaler,and P. Innocenti. 2001. Beneficial effects of Batimastat (BB-94), a matrix met-alloproteinase inhibitor, in rat experimental colitis. Digestion 63: 234–239.

8. Philip, S., A. Bulbule, and G. C. Kundu. 2004. Matrix metalloproteinase-2: mech-anism and regulation of NF-�B-mediated activation and its role in cell motilityand ECM-invasion. Glycoconj. J. 21: 429–441.

9. Corbel, M., C. Belleguic, E. Boichot, and V. Lagente. 2002. Involvement ofgelatinases (MMP-2 and MMP-9) in the development of airway inflammation andpulmonary fibrosis. Cell Biol. Toxicol. 18: 51–61.

10. Vu, T. H., and Z. Werb. 2000. Matrix metalloproteinases: effectors of develop-ment and normal physiology. Genes Dev. 14: 2123–2133.

11. Matsubara, M., J. D. Zieske, and M. E. Fini. 1991. Mechanism of basementmembrane dissolution preceding corneal ulceration. Invest. Ophthalmol. VisualSci. 32: 3221–3237.

12. Martignetti, J. A., A. A. Aqeel, W. A. Sewairi, C. E. Boumah, M. Kambouris,S. A. Mayouf, K. V. Sheth, W. A. Eid, O. Dowling, J. Harris, et al. 2001. Mu-tation of the matrix metalloproteinase 2 gene (MMP2) causes a multicentric os-teolysis and arthritis syndrome. Nat. Genet. 28: 261–265.

13. Hemdahl, A. L., A. Gabrielsen, C. Zhu, P. Eriksson, U. Hedin, J. Kastrup,P. Thoren, and G. K. Hansson. 2006. Expression of neutrophil gelatinase-asso-ciated lipocalin in atherosclerosis and myocardial infarction. Arterioscler.Thromb. Vasc. Biol. 26: 136–142.

14. Mook, O. R., W. M. Frederiks, and C. J. Van Noorden. 2004. The role of gela-tinases in colorectal cancer progression and metastasis. Biochim. Biophys. Acta1705: 69–89.

15. Turpeenniemi-Hujanen, T. 2005. Gelatinases (MMP-2 and -9) and their naturalinhibitors as prognostic indicators in solid cancers. Biochimie 87: 287–297.

16. Gao, Q., M. J. Meijer, F. J. Kubben, C. F. Sier, L. Kruidenier, W. van Duijn,M. van den Berg, R. A. van Hogezand, C. B. Lamers, and H. W. Verspaget. 2005.Expression of matrix metalloproteinases-2 and -9 in intestinal tissue of patientswith inflammatory bowel diseases. Dig. Liver Dis. 37: 584–592.

17. Baugh, M. D., M. J. Perry, A. P. Hollander, D. R. Davies, S. S. Cross, A. J. Lobo,C. J. Taylor, and G. S. Evans. 1999. Matrix metalloproteinase levels are elevatedin inflammatory bowel disease. Gastroenterology 117: 814–822.

18. Baugh, M. D., G. S. Evans, A. P. Hollander, D. R. Davies, M. J. Perry, A. J. Lobo,and C. J. Taylor. 1998. Expression of matrix metalloproteases in inflammatorybowel disease. Ann. NY Acad. Sci. 859: 249–253.

19. Kirkegaard, T., A. Hansen, E. Bruun, and J. Brynskov. 2004. Expression andlocalization of matrix metalloproteinases and their natural inhibitors in fistulae ofpatients with Crohn’s disease. Gut 53: 701–709.

20. Von Lampe, B., B. Barthel, S. E. Coupland, E. O. Riecken, and S. Rosewicz.2000. Differential expression of matrix metalloproteinases and their tissue inhib-itors in colon mucosa of patients with inflammatory bowel disease. Gut 47:63–73.

21. Louis, E., and J. Belaiche. 2000. Genetics of Crohn’s disease behavior. ActaGastroenterol. Belg. 63: 377–379.

22. Bailey, C. J., R. M. Hembry, A. Alexander, M. H. Irving, M. E. Grant, andC. A. Shuttleworth. 1994. Distribution of the matrix metalloproteinases strome-lysin, gelatinases A and B, and collagenase in Crohn’s disease and normal in-testine. J. Clin. Pathol. 47: 113–116.

23. Pender, S. L., S. P. Tickle, A. J. Docherty, D. Howie, N. C. Wathen, andT. T. MacDonald. 1997. A major role for matrix metalloproteinases in T cellinjury in the gut. J. Immunol. 158: 1582–1590.

24. Castaneda, F. E., B. Walia, M. Vijay-Kumar, N. R. Patel, S. Roser,V. L. Kolachala, M. Rojas, L. Wang, G. Oprea, P. Garg, et al. 2005. Targeteddeletion of metalloproteinase 9 attenuates experimental colitis in mice: centralrole of epithelial-derived MMP. Gastroenterology 129: 1991–2008.

25. Tarlton, J. F., C. V. Whiting, D. Tunmore, S. Bregenholt, J. Reimann,M. H. Claesson, and P. W. Bland. 2000. The role of up-regulated serine proteasesand matrix metalloproteinases in the pathogenesis of a murine model of colitis.Am. J. Pathol. 157: 1927–1935.

26. Seifert, W. F., T. Wobbes, and T. Hendriks. 1996. Divergent patterns of matrixmetalloproteinase activity during wound healing in ileum and colon of rats. Gut39: 114–119.

27. Opdenakker, G., P. E. Van den Steen, and J. Van Damme. 2001. Gelatinase B: atuner and amplifier of immune functions. Trends Immunol. 22: 571–579.

28. Itoh, T., T. Ikeda, H. Gomi, S. Nakao, T. Suzuki, and S. Itohara. 1997. Unalteredsecretion of �-amyloid precursor protein in gelatinase A (matrix metalloprotein-ase 2)-deficient mice. J. Biol. Chem. 272: 22389–22392.

29. Perez, S. E., D. A. Cano, T. Dao-Pick, J. P. Rougier, Z. Werb, and M. Hebrok.2005. Matrix metalloproteinases 2 and 9 are dispensable for pancreatic islet for-mation and function in vivo. Diabetes 54: 694–701.

30. Barthel, M., S. Hapfelmeier, L. Quintanilla-Martinez, M. Kremer, M. Rohde,M. Hogardt, K. Pfeffer, H. Russmann, and W. D. Hardt. 2003. Pretreatment ofmice with streptomycin provides a Salmonella enterica serovar Typhimuriumcolitis model that allows analysis of both pathogen and host. Infect. Immun. 71:2839–2858.

31. Snoek-van Beurden, P. A., and J. W. Von den Hoff. 2005. Zymographic tech-niques for the analysis of matrix metalloproteinases and their inhibitors. Bio-Techniques 38: 73–83.

32. Sitaraman, S. V., D. Merlin, L. Wang, M. Wong, A. T. Gewirtz, M. Si-Tahar, andJ. L. Madara. 2001. Neutrophil-epithelial crosstalk at the intestinal lumenal sur-face mediated by reciprocal secretion of adenosine and IL-6. J. Clin. Invest. 107:861–869.

33. Cooper, H. S., S. N. Murthy, R. S. Shah, and D. J. Sedergran. 1993. Clinico-pathologic study of dextran sulfate sodium experimental murine colitis. Lab.Invest. 69: 238–249.

34. Gewirtz, A. T., L. S. Collier-Hyams, A. N. Young, T. Kucharzik, W. J. Guilford,J. F. Parkinson, I. R. Williams, A. S. Neish, and J. L. Madara. 2002. Lipoxin A4

analogs attenuate induction of intestinal epithelial proinflammatory gene expres-sion and reduce the severity of dextran sodium sulfate-induced colitis. J. Immu-nol. 168: 5260–5267.

35. Rojas, M., J. Xu, C. R. Woods, A. L. Mora, W. Spears, J. Roman, andK. L. Brigham. 2005. Bone marrow-derived mesenchymal stem cells in repair ofthe injured lung. Am. J. Respir. Cell Mol. Biol. 33: 145–152.

36. Furuta, G. T., J. R. Turner, C. T. Taylor, R. M. Hershberg, K. Comerford,S. Narravula, D. K. Podolsky, and S. P. Colgan. 2001. Hypoxia-inducible factor1-dependent induction of intestinal trefoil factor protects barrier function duringhypoxia. J. Exp. Med. 193: 1027–1034.

37. Dignass, A. U., D. C. Baumgart, and A. Sturm. 2004. Review article: the aetio-pathogenesis of inflammatory bowel disease: immunology and repair mecha-nisms. Aliment Pharmacol. Ther. 20(Suppl. 4): 9–17.

38. Clayburgh, D. R., L. Shen, and J. R. Turner. 2004. A porous defense: the leakyepithelial barrier in intestinal disease. Lab. Invest. 84: 282–291.

39. Pirila, E., N. S. Ramamurthy, T. Sorsa, T. Salo, J. Hietanen, and P. Maisi. 2003.Gelatinase A (MMP-2), collagenase-2 (MMP-8), and laminin-5 �2-chain expres-sion in murine inflammatory bowel disease (ulcerative colitis). Dig. Dis. Sci. 48:93–98.

40. McQuibban, G. A., J. H. Gong, J. P. Wong, J. L. Wallace, I. Clark-Lewis, andC. M. Overall. 2002. Matrix metalloproteinase processing of monocyte chemoat-tractant proteins generates CC chemokine receptor antagonists with anti-inflammatory properties in vivo. Blood 100: 1160–1167.

41. Itoh, T., H. Matsuda, M. Tanioka, K. Kuwabara, S. Itohara, and R. Suzuki. 2002.The role of matrix metalloproteinase-2 and matrix metalloproteinase-9 in anti-body-induced arthritis. J. Immunol. 169: 2643–2647.

42. Jackson, C., M. Nguyen, J. Arkell, and P. Sambrook. 2001. Selective matrixmetalloproteinase (MMP) inhibition in rheumatoid arthritis: targetting gelatinaseA activation. Inflamm. Res. 50: 183–186.

43. Matsusaka, H., M. Ikeuchi, S. Matsushima, T. Ide, T. Kubota, A. M. Feldman,A. Takeshita, K. Sunagawa, and H. Tsutsui. 2005. Selective disruption of MMP-2gene exacerbates myocardial inflammation and dysfunction in mice with cyto-kine-induced cardiomyopathy. Am. J. Physiol. 289: H1858–H1864.

44. Matsumura, S., S. Iwanaga, S. Mochizuki, H. Okamoto, S. Ogawa, andY. Okada. 2005. Targeted deletion or pharmacological inhibition of MMP-2prevents cardiac rupture after myocardial infarction in mice. J. Clin. Invest.115: 599 – 609.

45. Esparza, J., M. Kruse, J. Lee, M. Michaud, and J. A. Madri. 2004. MMP-2 nullmice exhibit an early onset and severe experimental autoimmune encephalomy-elitis due to an increase in MMP-9 expression and activity. FASEB J. 18:1682–1691.

46. McQuibban, G. A., J. H. Gong, E. M. Tam, C. A. McCulloch, I. Clark-Lewis, andC. M. Overall. 2000. Inflammation dampened by gelatinase A cleavage of mono-cyte chemoattractant protein-3. Science 289: 1202–1206.

47. Corry, D. B., A. Kiss, L. Z. Song, L. Song, J. Xu, S. H. Lee, Z. Werb, andF. Kheradmand. 2004. Overlapping and independent contributions of MMP2 andMMP9 to lung allergic inflammatory cell egression through decreased CC che-mokines. FASEB J. 18: 995–997.

4111The Journal of Immunology

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from

48. McCawley, L. J., and L. M. Matrisian. 2001. Matrix metalloproteinases: they’renot just for matrix anymore! Curr. Opin. Cell Biol. 13: 534–540.

49. Ito, A., A. Mukaiyama, Y. Itoh, H. Nagase, I. B. Thogersen, J. J. Enghild,Y. Sasaguri, and Y. Mori. 1996. Degradation of interleukin 1� by matrix met-alloproteinases. J. Biol. Chem. 271: 14657–14660.

50. McKaig, B. C., D. McWilliams, S. A. Watson, and Y. R. Mahida. 2003. Expres-sion and regulation of tissue inhibitor of metalloproteinase-1 and matrix metal-loproteinases by intestinal myofibroblasts in inflammatory bowel disease.Am. J. Pathol. 162: 1355–1360.

51. Powell, D. W., P. A. Adegboyega, J. F. Di Mari, and R. C. Mifflin. 2005. Epi-thelial cells and their neighbors. I. Role of intestinal myofibroblasts in develop-ment, repair, and cancer. Am. J. Physiol. 289: G2–G7.

52. Miyamori, H., T. Takino, Y. Kobayashi, H. Tokai, Y. Itoh, M. Seiki, and H. Sato.2001. Claudin promotes activation of pro-matrix metalloproteinase-2 mediatedby membrane-type matrix metalloproteinases. J. Biol. Chem. 276: 28204–28211.

53. Agarwal, R., T. D’Souza, and P. J. Morin. 2005. Claudin-3 and claudin-4 ex-pression in ovarian epithelial cells enhances invasion and is associated with in-creased matrix metalloproteinase-2 activity. Cancer Res. 65: 7378–7385.

54. Siu, M. K., and C. Y. Cheng. 2004. Interactions of proteases, protease inhib-itors, and the �1 integrin/laminin �3 protein complex in the regulation ofectoplasmic specialization dynamics in the rat testis. Biol. Reprod. 70:945–964.

55. Tsukita, S., and M. Furuse. 2000. Pores in the wall: claudins constitute tightjunction strands containing aqueous pores. J. Cell Biol. 149: 13–16.

56. Tsukita, S., and M. Furuse. 2000. The structure and function of claudins, celladhesion molecules at tight junctions. Ann. NY Acad. Sci. 915: 129–135.

57. Pflugfelder, S. C., W. Farley, L. Luo, L. Z. Chen, C. S. de Paiva, L. C. Olmos,D. Q. Li, and M. E. Fini. 2005. Matrix metalloproteinase-9 knockout confersresistance to corneal epithelial barrier disruption in experimental dry eye.Am. J. Pathol. 166: 61–71.

58. Strongin, A. Y., I. Collier, G. Bannikov, B. L. Marmer, G. A. Grant, andG. I. Goldberg. 1995. Mechanism of cell surface activation of 72-kDa type IVcollagenase: isolation of the activated form of the membrane metalloprotease.J. Biol. Chem. 270: 5331–5338.

4112 SELECTIVE ABLATION OF MMP-2 EXACERBATES EXPERIMENTAL COLITIS

by guest on Decem

ber 1, 2014http://w

ww

.jimm

unol.org/D

ownloaded from