Early MT-1 MMP expression followingelastase exposure is associated withincreased cleaved MMP-2 activity inexperimental rodent aortic aneurysmsIndranil Sinha, BA, Kevin K. Hannawa, BS, Jonathan L. Eliason, MD, Gorav Ailawadi, MD,Michael P. Deogracias, Siddharth Bethi, MBBS, John W. Ford, BA, Karen J. Roelofs, DVM,Vladimir Grigoryants, MD, Peter K. Henke, MD, James C. Stanley, MD, andGilbert R. Upchurch, Jr, MD, Ann Arbor, Mich

Objective. The objective of this study was to determine the significance of membrane type 1 matrixmetalloproteinase (MT1-MMP) activation of MMP-2 in experimental abdominal aortic aneurysms.Methods. Rat aortas were perfused with either saline as a control or elastase, and harvested on 2, 4, or 7days after perfusion (n = 5 per treatment group/day). Aortic MT1-MMP and MMP-2 expression andprotein were determined by real time polymerase chain reaction and Western blotting, respectively. Aorticexplants were used to measure MMP-2 activity by zymography. Rat aortic smooth muscle cells in vitrowere exposed to increasing doses of elastase and analyzed for MT-1 MMP expression.Results. Aneurysms formed in 80% of the elastase-perfused aortas at 7 days, whereas none formed in thesaline-perfused aortas. Significantly increased MT1-MMP expression was observed only on day 4, whenlevels were 6.5-fold higher in elastase-perfused aortas compared with saline-perfused aortas (P < .01).By day 7, MT1-MMP protein was present only in the elastase-perfused aortas (P = .02). Byimmunohistochemistry, MT1-MMP was detectable only in the elastase-perfused group at day 7. CleavedMMP-2 activity (P = .045) was increased in elastase-perfused aortas compared with saline perfusedaortas at day 7. In rat aortic smooth muscle cells, MT-1 MMP expression increased in response toelastase (P = .02).Conclusion. The rodent aortic aneurysm model exhibits upregulation of MT1-MMP expression andprotein with subsequent increased conversion of MMP-2 from the latent to the cleaved form. (Surgery2004;136:176-82.)

From the Jobst Vascular Research Laboratories, Section of Vascular Surgery, Department of Surgery, Universityof Michigan, Ann Arbor, Mich

MATRIX METALLOPROTEINASES (MMPs), a family ofendopeptidases involved in degradation of extra-cellular matrix proteins, are of central relevance inthe pathogenesis of an abdominal aortic aneurysm(AAA).1-3 In particular, MMP-2 (gelatinase A) hasbeen shown to be important in the destructive pro-cess accompanying aneurysm development. Inhibi-

Presented at the 65th Annual Meeting of the Society ofUniversity Surgeons, St. Louis, Missouri, February 11-14, 2004.

Supported by NIH KO8 (HL67885-02), VonLeibig Award—Life-line Foundation, the Lifeline Medical Student Research Award,the Griswold and Margery H. Ruth Alpha Omega Alpha MedicalStudent Research Fellowship, and the Jobst Foundation.

Reprint requests: Gilbert R. Upchurch, Jr, MD, University ofMichigan Hospital, 2210 THCC, 1500 East Medical CenterDrive, Ann Arbor, MI 48109-0329.

0039-6060/$ - see front matter

� 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.surg.2004.04.010

176 SURGERY

tion of aneurysm formation was observed afterCaCl2 treatment of aortas in mice with a targeteddisruption of the MMP-2 gene.4

Most MMPs, including MMP-2, are producedand generally secreted in a latent form thatrequires stepwise activation through interactionswith a variety of proteins, including membrane type(MT)-MMPs.5-7 MMP-2 is secreted as an inactive72 kD proenzyme and becomes activated by pro-teolytic cleavage and removal of its terminalpropeptide domain, resulting in an activated 62kD form.8 MT-1 MMP is known to be critical in thisprocess.8 In this context, MT-1 MMP may berelevant to AAA pathogenesis as a necessary stepin activating MMP-2.

Recent studies in human AAA tissue have shownan increase in both MMP-2 expression and activity.1

An increase of MT-1 MMP expression in AAApatients has been observed, as well when compared

SurgeryVolume 136, Number 2

Sinha et al 177

with nonaneurysmal aorta.9 These findings lead tothe hypothesis that MT-1 MMP may also be up-regulated in an experimental AAA model. Theobjective of this study was to identify the timecourse of MT-1 MMP and MMP-2 expression andactivity in an experimental rodent elastase-inducedAAA model.

METHODS

All rats were obtained from Charles RiverLaboratories (Wilmington, Mass). All experimentsand procedures were approved by the University ofMichigan Universal Committee on the Use andCare of Animals (UCUCA #8566).

Elastase perfusion. Pancreatic porcine elastaseperfusion was performed, as previously described.10

Briefly, male Sprague-Dawley rats (200-250 g) wereanesthetized under 2% isofluorane inhalation andthe infrarenal aorta was isolated. Digital images ofthe aorta were obtained using Spot Insight ColorOptical Camera (Diagnostic Instruments Inc,Sterling Heights, Mich) attached to an operatingmicroscope (Nikon, Melville, NY). Aortic diameterwas measured using Image Pro Express (MediaCybernetics Inc, Silver Spring, Md). Temporarycontrol of the aorta was obtained and an aortotomywas made near the bifurcation using a 30-gaugeneedle. The aorta was then cannulated with PE-10tubing and perfused with either 2 ml isotonic saline(n = 15) or 12 units/ml total of porcine pancreaticelastase in 2 ml saline (n = 15) for 60 minutes. Theaortotomy was repaired with 10-0 suture. Aorticdiameter measurements were repeated at time ofharvest. Rodent aortas were harvested at 2, 4, or 7days after perfusion (n = 5 for both the saline andelastase group at each day). All harvested aortaswere subjected to histology, immunohistochemistry,quantitative polymerase chain reaction (PCR), andzymography.

The excised infrarenal aorta was cut into 3 equalsegments. The first segment was washed once withPBS and immediately snap-frozen for TRIzolextraction of mRNA and intracellular proteins.The remaining segments were washed for 24 hoursin Dulbecco modified Eagle media (DMEM) with100 units/ml penicillin and 100 lg/ml streptomy-cin (Gibco, Rockville, Md). After the wash, thesecond segment of the aorta was fixed in 4%paraformaldehyde. The third segment was placedin DMEM for another 72 hours, after which themedia was collected for zymography.

Cell culture. Rat aortic smooth muscle cells (RA-SMCs) were derived from explants of excisedabdominal aortas of young, male Sprague-Dawley

rats. Isolated medial tissue was cut and plated in 60-mm diameter plastic tissue culture dishes usingbasement membrane Matrigel (CollaborativeResearch, Bedford, Mass). Culture media includedDMEM containing 10% fetal bovine serum(HyClone Laboratories, Logan, Utah), 100 units/ml penicillin, and 100 lg/ml streptomycin. Mediaand antibiotics were from Gibco. Cultures weremaintained at 378 C in a humidified, 5% CO2

atmosphere for 4 to 7 days, until spindle-shaped RA-SMCs were observed growing from the explant.After removing the explant, the remaining cellswere dispersed by Dispase (Collaborative Research),centrifuged, and resuspended in complete mediumthat was placed into T25 culture flasks. RA-SMCswere confirmed by staining with a monoclonalantibody against SMC-specific a-actin.

After RA-SMC growth to confluence, the cellswere incubated in 0, 0.05, 0.5, and 5 units/ml ofWorthington elastase (Lot #2279) for 5 minutes.The RA-SMCs were then washed twice in PBS.DMEM was then added, and the cells were in-cubated for 24 hours. The RA-SMCs were thenharvested and the mRNA isolated by extractionwith TRIzol (Life Technologies, Rockville, Md).

Quantitative (real-time) PCR. Expression of MT-1 MMP, MMP-2, and b-actin mRNA was determinedusing quantitative PCR. Messenger RNA was iso-lated by treatment of aortic segments with TRIzolreagent and reverse transcribed by incubating witholigo-dT primer (Life Technologies, Grand Island,NY) and M-MLV Reverse Transcriptase (LifeTechnologies, Grand Island, NY) at 948 C for 3minutes, followed by incubation at 408 C for 70minutes. The resultant cDNA was amplified by TaqPolymerase (Promega, Madison, Wis) in theSmartCycler quantitative polymerase chain reac-tion system (Cepheid, Sunnyvale, Calif). SYBRintercalating dye (Roche, Indianapolis, Ind) wasused in monitoring cDNA amplification for eachgene. SmartCycler quantification data are pres-ented as cycle threshold (Ct).

MT-1 MMP, MMP-2, and b-actin primer se-quences were derived using Primer Premiersoftware (Premier Biosoft International, Palo Alto,Calif) based on primary cDNA sequences fromGenBank (http://www.ncbi.nlm.nih.gov/Genbank/).The sequences for primers are 59-ATGGGTCAGAA-GGATTCCTATGTG-39 (b-actin sense), 59-CTTCAT-GAGGTAGTCAGTCAGGTC-39 (b-actin antisense),59-AGGCTCATTCATGGGTAGCG-39 (MT-1 MMPsense) and 59-TTCCAGTATTTGTTCCCTTTGTAG-39(MT-1 MMP antisense), 59-CATCGCTGCACCATCG-CCCATCATC-39 (MMP-2 sense) and 59-CCCAGGG-TCCACAGCTCATCATCATCA-39 (MMP-2 antisense).

SurgeryAugust 2004

178 Sinha et al

The results were normalized using the b-actingene. Quantification of mRNA levels used DCt

values calculated from the formula DCt = Ct target

gene e Ct b-actin. Expression of the target gene inratio to b-actin expression was calculated by theformula: target gene expression/b-actin expres-sion = 2-(DCt).

Substrate gel zymography. MMP distribution inharvested aortas was determined by zymographyusing reagents from Novex (San Diego, Calif) onexperimental medium, as previously described.11

Briefly, gelatin substrate zymograms were preparedusing precast 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels containing 1 mg/ml of gelatin.Experimental samples of equal volume were di-luted into 2X tris-glycine SDS sample buffer andseparated electrophoretically. Gel incubation per-iods varied to best quantify the desired MMP bandby Coomassie Blue staining. The relative molecularweight of each band was determined by compari-son of the bands against MMP-2 (72 kD) standards(Oncogene Research Products, Boston, Mass).MMP activity was measured by densitometry, aslater described.

Western blot analysis. Electrophoresis andWestern blot materials were from Bio-Rad(Hercules, Calif). Western blots for MT-1 MMPand MMP-2 were performed on intracellular pro-tein extraction from the TRIzol reagent. Proteinswere separated electrophoretically on a 10% SDS-PAGE and blotted to a nitrocellulose membrane.Nonspecific binding was blocked by incubating themembrane for 1 hour in 20 mM tris-HCl (pH 7.5)containing 0.5 M NaCl, 0.1% Tween 20 and 5%

Fig 1. Aortic diameter following either saline or elastaseperfusion. At days 4 and 7, the elastase-perfused aortaswere significantly larger than saline-perfused aortas. *P <.01. The dashed line approximately indicates the diameterat which the growth in aortic diameter is defined asaneurysmal. No saline-perfused aortas became aneurys-mal. In contrast, 80% of elastase-perfused aortas becameaneurysmal by the day 7. N = 5 per group at each day ofharvest.

nonfat milk (MMP-2 blots), and 5% BSA (MT-1MMP blots). MMP-2 antibodies were obtainedfrom Oncogene, and MT-1 MMP antibodies wereobtained from Santa Cruz Biotechnology (SantaCruz, Calif). Immunoreactive bands were visualizedusing peroxidase-coupled secondary antibodies andan ECL chemiluminescence detection kit fromAmersham (Piscataway, NJ). Each protein wasmeasured by densitometry.

Immunohistochemistry. Paraformaldehyde-fixed,paraffin-embedded samples underwent deparaffi-nization, rehydration, and unmasking using Trilogyfrom Cell Marque Corporation (Hot Springs, Ark)in an electric pressure cooker. Endogenous perox-idase activity was blocked using 3% hydrogenperoxide in methanol. Staining was performed forMMP-2 using a mouse antirat primary antibodyfrom Oncogene and a mouse immunoglobulin GVectastain ABC-AP kit from Vector Laboratories(Burlingame, Calif). MT-1 MMP staining was per-formed using a goat antirat MT1-MMP primaryantibody from Santa Cruz Biotechnology and a goatimmunoglobulin G Vectastain ABC-AP Kit.

Total protein assay. Aortas were solubilized in0.1% SDS, and total cellular protein was deter-mined by a bicinchoninic acid protein assay(Pierce, Rockford, Ill).

Densitometry. Zymography and Western blotswere imaged using a FOTO/Analyst CCD CAMERA(Fotodyne, Hartland, Wis) and quantified usingGEL-Pro Analyzer software version 3.1 (MediaCybernetics, Silver Springs, Md). Optical densitieswere normalized to protein levels from thecorresponding protein assay.

Statistics. Data analysis used non-paired t testwith statistical significance assigned as P < .05.Analysis was performed using GraphPad Prism 3.0(GraphPad, San Diego, Calif).

RESULTS

Elastase-perfused aortas formed aneurysms by7 days after elastase perfusion. Abdominal aorticdiameters increased by a mean ( ± SEM) of 168%± 25% by 7 days after elastase perfusion, as

compared with 30% ± 5% in the saline-perfusedaortas (P = .0006). Of the elastase-treated aortas,80% formed an aneurysm by 7 days, as defined bya 100% growth in aortic diameter (n = 5). In salinecontrols, none of the aortas was aneurysmal by day7 (n = 5) (Fig 1).

Aortic wall MT-1 MMP is upregulated duringAAA pathogenesis. MT-1 MMP expression waselevated in the elastase-perfused aortas comparedwith the saline-perfused aortas at all time points

SurgeryVolume 136, Number 2

Sinha et al 179

Fig 2. A, MT-1 MMP/b-actin mRNA expression by real-time PCR demonstrated significantly elevatedexpression in the elastase-perfused aortas compared with the saline-perfused aortas at day 4. *P < .01, n = 5per group at each day of harvest. B, Representative Western blot demonstrating MT-1 MMP protein in theelastase-perfused aortas (E), with no MT-1 MMP in the saline-perfused aortas (S). C, Aortic MT1-MMPprotein quantity as measured by Western blots demonstrates that the protein is only detectable in theelastase-perfused aortas. *P = .02, n = 5 per group.

measured. By post-perfusion day 4, the level of MT-1 MMP expression was 6.5-fold higher (P < .01) inelastase-perfused aortas (transcription ratio of0.52 ± 0.13), compared with the saline-perfusedaortas (0.08 ± 0.01) (Fig 2, A). Increased MT-1MMP expression was reflected by increased proteinlevels by Western blotting of aortas on post-perfusion day 7. MT-1 MMP protein was presentonly in the elastase-perfused aortas, but could notbe detected in either group before day 7 (opticaldensity [OD] = 3.74 ± 1.01) (P = .02, Fig 2, B, C).

Immunohistochemistry revealed no positiveMT-1 MMP staining in the saline-perfused aortas(Fig 3, A). In contrast, positive MT-1 MMP stainingwas observed in the elastase-perfused aortas in boththe outer media and the adventitia (Fig 3, B).

Upregulation of MT-1 MMP is associated withincreased activation of MMP-2. MMP-2 expressionand protein levels (data not shown) were similarbetween the saline- and elastase-perfused aortas atall days of harvest (2, 4, and 7 days). Gelatin zymog-raphy at 2 and 4 days did not reveal significantdifferences in either pro- or cleaved-MMP-2 activityamong the 2 groups.

By immunohistochemistry at day 7, MMP-2 wasdetectable in the saline-perfused aortas (Fig 4, A).However, MMP-2 staining appeared moreprominent in the elastase perfused aortas (Fig 4,B). Staining was present in both groups primarily inthe outer media and adventitia, similar to MT1-MMP staining. Protein zymography at 7 days

showed similar levels of total MMP-2 activity (.0018OD/ng protein ± 0.0003 in the elastase groupcompared with .0017 OD/ng protein ± 0.0004 inthe saline group). In marked contrast, cleavedMMP-2 activity was 175% higher in the elastase-perfused aortas (.001 OD/ng protein ± 0.0001compared with .0006 OD/ng protein ± 0.0002 inthe saline-perfused aortas, P = .045) (Fig 5, A, B).

Elastase exposure upregulates MT-1 MMP ex-pression in RA-SMC. To ascertain if elastase affectsRA-SMC MT-1 MMP expression in vitro, a dose-response to elastase was performed (n = 4 pergroup). MT-1 MMP expression increased withtransient exposure to increasing doses of elastase.Exposure of RA-SMCs to an elastase dose of 5units/ml resulted in a 64.1% (P = .024) increase inMT-1 MMP expression (2.21 ± 0.25) comparedwith the saline-treated cells (1.34 ± .15) (Fig 6).

DISCUSSION

This study suggests that MT-1 MMP expression isinduced in experimental AAAs and this may beimportant in MMP-2 activation. MT-1 MMP expres-sion significantly increased at day 4 in elastase-perfused aortas as compared with saline-perfusedaortas. By day 7, MT-1 MMP expression in theelastase-perfused aortas had decreased to expres-sion levels similar to the saline-perfused aortas,suggesting that the increased MT-1 MMP expres-sion is only transiently expressed during aneurysmformation. Increased cleaved MMP-2 activity also

SurgeryAugust 2004

180 Sinha et al

became significant on day 7, suggesting thatincreased protein levels of MT-1 MMP may benecessary for this conversion to proceed. Similarfindings in humans have been noted by Daviset al.12 Concurrent in vitro experiments revealedMT-1 MMP to be upregulated when RA-SMCs wereexposed to elastase at concentrations that inducerodent experimental AAAs in vivo. This suggeststhat such a cellular response may be a plausiblemechanism of increase in MT-1 MMP levels in the

Fig 3. Immunohistochemistry (20x) revealed little de-tectable MT-1 MMP in the saline-perfused aortas (A),compared with prominent MT-1 MMP staining in theelastase-perfused aortas (B), primarily in the outer mediaand adventitia.

intact aorta. In addition, a substantial influx ofinflammatory cells may also play a role in increasedMT-1 MMP production.

Elastase-induced abdominal aneurysms inrodents share certain key features with humanAAAs.10 Elastase infusion in rats leads to sub-sequent aortic dilatation, collagen and elastindegradation, MMP upregulation, and an extensiveinflammatory cell infiltrate in the outer media andadventitia of the aortic wall.9,13-15 Some haveargued that elastase perfusion of the aorta is anacute AAA model that may not mimic human AAAdevelopment, especially owing to variability ofexperimental aneurysm formation with varying lotsof elastase.16 All AAAs in the present study werecreated using a single lot of elastase. Unfortunately,another available experimental aneurysm modelof local CaCl2 application, as described by Gertzet al.17 produces only modest 30% to 50% increasesin aortic diameter at 10 weeks.4 Thus, the rodentelastase model serves as the best currently availableexperimental method to study molecular and cel-lular changes accompanying aneurysm formation.

Recent descriptive studies in humans implicatea role for MT-1 MMP in AAA pathogenesis. Onestudy noted increased MT-1 MMP expression inAAA wall smooth muscle cells and macrophages, ascompared with normal aortic tissue.9 A secondhuman study has implicated MT-1 MMP activity inAAA based on its activity of cleaving the latent 72-kD MMP-2 to its activated 62-kD form.12 This laterstudy demonstrated a significant proportion of proMMP-2 in the normal aorta and a greater pro-portion of activated MMP-2 in diseased aorta.Furthermore, in vitro studies in which MT-1 MMPwas depleted resulted in insignificant amounts ofMMP-2 activation compared with normal, empha-sizing the role of MT-1 MMP in the activation ofMMP-2.18,19

The pivotal function of MT-1 MMP in matrixremodeling, a central point of dysregulation inAAA formation, is evidenced by the fact that MT-1MMP-deficient mice develop numerous connectivetissue disorders from inadequate collagen turn-over.20 Experimental studies using MT-1 MMP-/-

mice might help to delineate further the role ofMT-1 MMP in experimental aneurysm formation,but such mice are not healthy enough to undergoelastase perfusion. In addition, MT-1 MMP in-hibitors may also more clearly define the role ofMT-1 MMP. However, specific MT-1 MMP inhibitorsare not yet available. Although existing data suggestthat MT-1 MMP may have a role in aneurysmpathology, it remains to be determined if MT-1MMP has a role in AAA progression.21

SurgeryVolume 136, Number 2

Sinha et al 181

Fig 4. MMP-2 by immunohistochemistry (20x) was detectable in the saline-perfused aortas (A), but wasmore prominent in the elastase-perfused aortas (B).

Effects of MT-1 MMP on MMP-2 may beimportant in AAA formation considering thatMMP-2 is unique in its ability to cleave both elastinand fibrillar-collagen.4,22 Longo and his colleagueshave previously demonstrated MMP-2edeficient

Fig 5. A, Representative zymogram demonstrating in-creased cleaved MMP-2 (62 kD) in the elastase-perfusedaortas (E) compared with saline-perfused aortas (S). B,Zymography demonstrated no significant difference intotal MMP-2 activity at 7 days. However, there wassignificantly higher cleaved MMP-2 activity in theelastase-perfused aortas by zymography. *P = .045, n = 5per group at each day of harvest.

mice were unable to form aneurysms. However, ithas not yet been demonstrated whether the latentform of MMP-2 alone is sufficient in producingAAA, or if it requires activation.4 MT-1 MMP mayalso play a role in the activation of MMP-8 andMMP-13.23,24 Although the role of MMP-8 isunknown in AAA pathogenesis, MMP-13 expres-sion is upregulated in human AAA tissue.25

This present study documents a role for MT-1MMP in experimental AAA formation, possiblymediated by increased MMP-2 activation. An un-derstanding of MT-1 MMP and its role in AAA mayultimately provide a pharmacologic modality bywhich to inhibit AAA development and progressionin humans.

Fig 6. MT1-MMP expression in rat aortic smooth musclecells increased in a dose-dependent manner in responseto transient elastase exposure. *P = .024, n = 4 pergroup.

SurgeryAugust 2004

182 Sinha et al

REFERENCES

1. McMillan WD, Patterson BK, Keen RR, Pearce WH. In situlocalization and quantification of 72 kD type IV collagenasein aneurysmal, occlusive, and normal aorta. J Vasc Surg1995;22:295-305.

2. McMillan WD, Patterson BK, Keen RR, Shively VP,Cipollone M, Pearce WH. In situ localization and quanti-fication of mRNA for 92 kD type IV collagenase and itsinhibitor in aneurysmal, occlusive, and normal aorta.Atheroscler Thromb Vasc Biol 1995;15:1139-44.

3. Thompson RW, Holmes DR, Mertens RA, Liao S, BotneyMD, Mecham RP, et al. Production and localization of 92 kDgelatinase in abdominal aortic aneurysms; an elastolyticmetalloproteinase expressed by aneurysm infiltrating mac-rophages. J Clin Invest 1995;96:318-26.

4. Longo MG, Xiong W, Greiner TC, Zhao Y, Fiotti N, BaxterTB. Matrix metalloproteinases 2 and 9 work in concert toproduce aortic aneurysms. J Clin Invest 2002;110:625-32.

5. Ben-Yosef Y, Lahat N, Shapiro S, Bitterman H, Miller A.Regulation of endothelial matrix metalloproteinase-2 byhypoxia/reoxygenation. Circ Res 2002;90:784-91.

6. Borden P, Heller RA. Transcriptional control of matrixmetalloproteinases and tissue inhibitors of metalloprotei-nases. Crit Rev Eukaryot Gene Expr 1997;7:159-78.

7. Nagase H. Activation mechanisms of matrix metalloprotei-nases. Biol Chem 1997;378:151-60.

8. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA,Goldberg GI. Mechanism of cell surface activation of 72 kDtype-IV collagenase. J Biol Chem 1995;270:5331-8.

9. Nollendorfs A, Greiner TC, Nagase H, Baxter TB. Theexpression and localization of membrane type-1 matrixmetalloproteinase in human abdominal aortic aneurysms.J Vasc Surg 2001;34:316-22.

10. Anidjar S, Salzmann JL, Gentric D, Lagneau P, Camilleri JP,Michel JB. Elastase-induced experimental aneurysms in rats.Circulation 1990;82:973-81.

11. Upchurch GR Jr, Ford JW, Weiss SJ, Knipp BS, Peterson DA,Thompson RW, et al. Nitric oxide inhibition increasesmatrix metalloproteinase-9 expression by rat aortic smoothmuscle cells in vitro. J Vasc Surg 2001;34:76-83.

12. Davis V, Persidskaia R, Baca-Regen L, Itoh Y, Nagase H,Persidsky Y, et al. Matrix metalloproteinase-2 production andits binding to the matrix are increased in abdominal aorticaneurysms. Arterioscler Thromb Vasc Biol 1998;18:1625-33.

13. Koch AE, Haines GK, Rizzo RJ, Radosevich JA, Pope RM,Robinson PG. Human abdominal aortic aneurysms: immu-nophenotypic analysis suggesting an immune-mediatedresponse. Am J Pathol 1990;137:1199-213.

14. Brophy CM, Reilly JM, Smith GJW, Tilson MD. The role ofinflammation in nonspecific abdominal aortic aneurysmdisease. Ann Vasc Surg 1991;5:229-33.

15. Parums DV, Dunn DC, Dixon AK, Mitchinson MJ. Charac-terization of inflammatory cells in patients with chronicperiaortitis. Am J Cardiovasc Pathol 1990;3:121-9.

16. Curci JA, Thompson RW. Variable induction of experimen-tal abdominal aortic aneurysm model with different prep-arations of porcine pancreatic elastase. J Vasc Surg 1999;29:385.

17. Gertz SD, Kurgan S, Eisenberg D. Aneurysm of the rabbitcommon carotid artery induced by periarterial applicationof calcium chloride in vivo. J Clin Invest 1998;81:649-56.

18. Rajavashith TB, Xu XP, Jovinge S, Meisel S, Xu XO, ChaiNN, et al. Membrane type-1 matrix metalloproteinaseexpression in human atherosclerotic plaques: evidence foractivation by proinflammatory mediators. Circulation 1999;99:3103-9.

19. Ruangpanit N, Chan D, Holmbeck K, Birkadel-Hansen H,Palarek J, Yang C, Bateman JF, Thompson EW. Gelatinase A(MMP-2) activation by skin fibroblasts: dependent onMT1-MMP expression and fibrillar collagen form. MatrixBiol 2001;20:193-203.

20. Holmbeck K, Bianco P, Caterina J, Yamada S, Kromer M,Kuznetsov SA, et al. MT-1 MMP deficient mice developdwarfism, osteopenia, arthritis, and connective tissuedisease due to inadequate collagen turnover. Cell 1999;99:81-92.

21. Freestone T, Turner RJ, Coady A, Higman DJ, GreenhalghRM, Powell JT. Inflammation and matrix metalloproteinasesin the enlarging abdominal aortic aneurysm. ArteriosclerThromb Vasc Biol 1995;15:1145-51.

22. Aimes R, Quigley J. Matrix metalloproteinase-2 is aninterstitial collagenase. J Biol Chem 1995;270:5872-6.

23. Holopainen JM, Moilanen JA, Sorsa T, Kivela-Rajamaki M,Tervahatiala T, Vesaluoma MH, et al. Activation of matrixmetalloproteinase-8 by membrane-type 1 MMP and theirexpression in human tears after photorefractive keratec-tomy. Invest Ophthalmol Vis Sci 2003;44:2550-6.

24. Knauper V, Will H, Lopez-Otin C, Smith B, Atkinson SJ,Stanton H, et al. Cellular mechanisms for human procolla-genase-3 (MMP-13) activation. Evidence that MT1-MMP(MMP-14) and gelatinase A (MMP-2) are able to generateactive enzyme. J Biol Chem 1996;271:17124-31.

25. Mao D, Lee JK, VanVickle SJ, Thompson RW. Expression ofcollagenase-3 (MMP-13) in human aortic aneurysms andvascular smooth muscle cells in culture. Biochem BiophysRes Commun 1999;261:904-10.