Selective spatiotemporal induction of matrix metalloproteinase … · Selective spatiotemporal induction of matrix metalloproteinase-2 and matrix metalloproteinase-9 transcription

Selective spatiotemporal induction of matrix metalloproteinase-2 and matrixmetalloproteinase-9 transcription after myocardial infarction

Rupak Mukherjee,1 Joseph T. Mingoia,1 James A. Bruce,1 Jeffrey S. Austin,2 Robert E. Stroud,1

G. Patricia Escobar,1 David M. McClister, Jr.,1 Claire M. Allen,1 Maria A. Alfonso-Jaume,3

M. Elizabeth Fini,2 David H. Lovett,3 and Francis G. Spinale1,4

1Division of Cardiothoracic Surgery, Medical University of South Carolina, and 4Ralph H. Johnson Veterans AdministrationMedical Center, Charleston, South Carolina; 2McKnight Vision Research Center, Bascom Palmer Eye Institute,University of Miami Miller School of Medicine, Miami, Florida; and 3Department of Medicine, VeteransAffairs Medical Center/University of California, San Francisco, California

Submitted 20 December 2005; accepted in final form 5 June 2006

Mukherjee, Rupak, Joseph T. Mingoia, James A. Bruce, Jef-frey S. Austin, Robert E. Stroud, G. Patricia Escobar, David M.McClister, Jr., Claire M. Allen, Maria A. Alfonso-Jaume, M.Elizabeth Fini, David H. Lovett, and Francis G. Spinale. Selectivespatiotemporal induction of matrix metalloproteinase-2 and matrixmetalloproteinase-9 transcription after myocardial infarction. Am JPhysiol Heart Circ Physiol 291: H2216–H2228, 2006. First publishedJune 9, 2006; doi:10.1152/ajpheart.01343.2005.—Myocardial remod-eling after myocardial infarction (MI) is associated with increasedlevels of the matrix metalloproteinases (MMPs). Levels of two MMPspecies, MMP-2 and MMP-9, are increased after MI, and transgenicdeletion of these MMPs attenuates post-MI left ventricular (LV)remodeling. This study characterized the spatiotemporal patterns ofgene promoter induction for MMP-2 and MMP-9 after MI. MI wasinduced in transgenic mice in which the MMP-2 or MMP-9 promotersequence was fused to the �-galactosidase reporter, and reporter levelwas assayed up to 28 days after MI. Myocardial localization withrespect to cellular sources of MMP-2 and MMP-9 promoter inductionwas examined. After MI, LV diameter increased by 70% (P � 0.05),consistent with LV remodeling. �-Galactosidase staining in MMP-2reporter mice was increased by 1 day after MI and increased furtherto 64 � 6% of LV epicardial area by 7 days after MI (P � 0.05).MMP-2 promoter activation occurred in fibroblasts and myofibro-blasts in the MI region. In MMP-9 reporter mice, promoter inductionwas detected after 3 days and peaked at 7 days after MI (53 � 6%,P � 0.05) and was colocalized with inflammatory cells at theperi-infarct region. Although MMP-2 promoter activation was simi-larly distributed in the MI and border regions, activation of theMMP-9 promoter was highest at the border between the MI andremote regions. These unique findings visually demonstrated thatactivation of the MMP-2 and MMP-9 gene promoters occurs in adistinct spatial relation with reference to the MI region and changes ina characteristic time-dependent manner after MI.

structure; remodeling; matrix metalloproteinases

REGIONAL AND GLOBAL CHANGES in left ventricular (LV) geome-try and structure, collectively termed remodeling, are commonsequelae after myocardial infarction (MI) (22, 29, 34). Signif-icant changes within the extracellular matrix (ECM) after MIcan facilitate adverse myocardial remodeling (8, 9, 33, 41).Structural changes within the MI region are initiated early(within hours), whereas later phases of myocardial remodeling

progress in a time-dependent manner through changes in thecomposition and conformation of the ECM (34). The matrixmetalloproteinases (MMPs) are endopeptidases that are presentwithin the myocardium, and changes in MMP abundance areassociated with ECM and LV remodeling after MI (9, 10, 26,28, 34, 35, 42). In previous studies, transgenic animal modelsor pharmacological MMP inhibition (12, 15–17, 20, 21, 26, 30,44) was used to demonstrate a causal role for MMPs in post-MILV remodeling. Nevertheless, it remained unknown whetherthese MMP species are expressed in regions of the myocar-dium critical to post-MI remodeling. The goal of this study wasto characterize the spatial and temporal patterns of expressionfor certain MMP species after MI.

A number of MMP species have been identified, and thepotential function of specific MMPs in post-MI LV remod-eling, as well as in LV hypertrophy, is an area of activeinvestigation (8, 10 –12, 17, 21, 22, 26, 28, 29, 34, 35). Twospecies of MMPs, MMP-2 and MMP-9, can degrade dena-tured collagens (gelatins) and basement membrane compo-nents (39). Previous studies have demonstrated that mRNAand protein levels of MMP-2 and MMP-9 are increased afterMI (7, 9, 15, 28, 35). The synthesis of MMP-2 and MMP-9is highly regulated at the transcriptional level. The genepromoter region of MMP-1, MMP-3, MMP-7, MMP-9,MMP-12, and MMP-13 includes a proximal activator pro-tein-1 (AP-1) site, which mediates an enhanced transcrip-tional response to a wide variety of cytokine and cellularstimuli (38). The MMP-2 promoter is notable for the ab-sence of this proximal AP-1 site but does include discreteenhancer elements more distally, which interact with AP-1components, as well as the YB-1 and p53 transcriptionfactors (2, 23). Thus the promoter elements of these verysimilar enzymes are quite distinct, suggesting that eachenzyme may operate within discrete cellular or stimulatorycontexts. The present study tested the hypothesis that dis-tinctly different spatiotemporal patterns of MMP-2 andMMP-9 expression would occur after MI, reflecting discretetranscriptional regulatory responses of the MMP-2 andMMP-9 genes to injury. To test this hypothesis, murineMMP-2 and MMP-9 reporter constructs were utilized toexamine expression of these MMP types after surgicalinduction of MI.

Address for reprint requests and other correspondence: R. Mukherjee,Cardiothoracic Surgery, Strom Thurmond Research Bldg., 770 MUSC Com-plex, Ste. 625, Medical Univ. of South Carolina, Charleston, SC 29425(e-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Am J Physiol Heart Circ Physiol 291: H2216–H2228, 2006.First published June 9, 2006; doi:10.1152/ajpheart.01343.2005.

http://www.ajpheart.orgH2216

MATERIALS AND METHODS

This study was designed to examine the temporal induction ofMMP-2 and MMP-9 gene promoters after MI. Transgenic mice, inwhich appropriate �-galactosidase (�-Gal) reporter constructs forMMP-2 or MMP-9 were inserted into the genome, were used. Allanimals were treated and cared for in accordance with the NationalInstitutes of Health Guide for the Care and Use of LaboratoryAnimals (Washington, DC: National Academy Press, 1996), and theprotocol was approved by the Institutional Animal Care and UseCommittee.

MMP-2 and MMP-9 reporter mice. MMP-2 reporter mice weregenerated using a 5-kb genomic fragment of the rat MMP-2 geneextending from �1686 (relative to the translational start site) to themiddle of the second exon and a plasmid (p1.2C) containing apolylinker 5� to a simian virus (SV40) polyadenylation signal. TheEscherichia coli �-Gal gene (lacz) was then isolated by PCR andligated between the MMP-2 promoter and the SV40 polyadenylationsequence to produce F8-�gal. Transgenic mice (CD-1 backgroundstrain) were generated using standard protocols and characterized byPCR screening of tail-clip DNA.

Mice with the MMP-9 reporter-lacz construct in the CD-1 back-ground strain were developed and described by Mohan et al. (24). Asinternal controls, transgene expression was ascertained by PCRscreening of tail-clip DNA and �-Gal elaboration [5-bromo-4-chloro-3-indolyl-�-D-galactopyranoside (X-gal) reacted] at the cut edges ofthe excised tail tips.

MI induction and experimental design. To provide equivalent sexdistribution, MI was surgically induced in male and female mice(108 � 3 days of age, 35.5 � 0.5 g body wt) in the MMP-2 andMMP-9 groups by ligation of the coronary artery, as describedpreviously (11). This technique has been documented to cause an MIof 35 � 4% of the LV (11). Five MMP-2 reporter mice and fiveMMP-9 reporter mice were included to serve as nonoperated, non-MIreference controls.

Terminal studies on the mice were performed at 1 h (acute) and 1,3, 7, 14, and 28 days after MI. At the assigned time after MI, the micewere deeply anesthetized by inhalation of isoflurane and weighed. Thethoracic cavity was opened, 0.2 ml of a 0.1 M CdCl2 solution wasinjected into the LV, and the heart was extirpated. The heart wasblotted dry, weighed, and sectioned along the long axis, such that theline of dissection would bisect the MI region, as visually determinedat the time of heart extirpation. One-half of the heart was used todetermine the extent of �-Gal staining (formalin fixation followed byreaction against the X-Gal substrate), and the other half was frozen inoptimal temperature cutting (OTC) embedding compound (Tissue-Tek). Since these studies were the first to examine the post-MIremodeling response in transgenic mouse lines with MMP promoterreporter constructs inserted into the genome, preliminary echocardio-graphic recordings were performed in a subset of the MMP-9 reportermice (n � 8) to determine whether post-MI LV remodeling pro-gressed in a manner similar to that reported previously (4–6, 11, 12,15, 17, 21, 30, 43). Specifically, echocardiographic determinations ofLV size and fractional shortening (15-MHz transducer, Sonos 5500,Agilent Technologies) were performed before MI induction and at 28days after MI, as previously described (11). Briefly, the parasternallong-axis view of the LV, together with a recording of the surfaceECG, was acquired to disk. LV volumes were determined by manualplanimetry of the LV endocardial border at end diastole (frame withR wave) and end systole (smallest LV area in the cardiac cycle) andapplication of a variant of Simpson’s algorithm [“method of disks”(27, 36), EnConcert, Agilent Technologies].

�-Gal staining and quantitation. MMP-2 or MMP-9 promoteractivation was assayed as a function of �-Gal elaboration in one-halfof each heart by overnight incubation in a substrate buffer (X-Gal,Sigma), as previously described (24). The hearts were photographed(model F707, Sony) with the atria and right ventricles attached. The

LV was separated, and the area and intensity of �-Gal staining on theLV epicardium, as well as the LV epicardial area, were determinedfrom the two-dimensional photographs by digital planimetry (Sig-maScan version 4.0, SPSS). The area of �-Gal staining was normal-ized to LV epicardial area and expressed as a percentage.

Spatial elaboration of �-Gal staining in each of the hearts wasquantitated with respect to the center of the MI region. Because ofdifferences in the extent of �-Gal staining on each heart, as well asheart-to-heart variability in post-MI LV geometry, preliminary studiesindicated that the center of mass of the MI region provided the mostconsistent, defined point of origin on which to base spatial measure-ments. The green channel from the color photographs of the �-Gal-stained hearts was extracted and inverted. Because �-Gal stainingusing the X-Gal substrate is in the green spectrum, the rationale forextracting only the green channel was based on the observation thatthe unstained myocardium converted to a more consistent light-graylevel after inversion. Moreover, conversion of the image to gray scaleprovided a more consistent means for quantitation of �-Gal stainingintensity than could be obtained from the color images. Therefore, inthe resultant gray-scale photographs, the region of �-Gal staining wasprojected as lighter areas on a darker background. The MI region wasdigitally planimetered for determination of the major axis of the MIregion as well as the center of mass, which was then defined as thecenter of the MI region. A five-pixel-wide region of interest, extend-ing from the center of the MI region to the remote myocardium alongthe major axis, was demarcated. With the center of the MI regiondefined as the origin, gray-scale intensities (0 � black, 255 � white)at 2.0 mm (border region) and 3.0 mm (remote region) were deter-mined.

Biochemistry. Before tissue processing for biochemical studies, thehearts that were frozen in OTC compound were thawed over ice,extracted, and washed stringently in ice-cold PBS before homogeni-zation. Thus biochemical studies were performed in the same hearts inwhich �-Gal staining was determined. �-Gal activity in the LV wasbiochemically determined from the frozen half of the hearts byisolation and homogenization of the peri-MI region and use of theFluoReporter lacZ/galactosidase quantitation kit (catalog no. F-2905,Molecular Probes, OR). Briefly, LV homogenates were incubated atroom temperature in a 96-well plate with a fluorogenic substrate(CUG, Molecular Probes). The reaction was stopped at 30 min, andfluorescence was measured (FLUOstar Galaxy, BMG Labtechnolo-gies). A calibration curve generated from a set of �-gal standards(1.6–100 ng/ml) was used to convert the fluorescence signal tonanograms per milliliter.

Gelatinase activity in the LV homogenates (20 �g total protein)was determined as a function of cleavage of a gelatinase substrate(ECM700 kit, Chemicon). Briefly, p-aminophenylmercuric acetate-reacted LV tissue homogenates and a biotinylated gelatinase substratewere added to wells of a 96-well plate. After 2 h of incubation, 100 �lof the sample/biotinylated gelatinase substrate were added to a biotin-binding plate and incubated for 30 min at 37°C. The wells werestringently washed, and a 1:3,000 dilution of streptavidin enzymeconjugate was added and incubated for 30 min at 37°C. After anotherwash cycle, the substrate solution was added, the plate was incubatedat room temperature for 20 min, and the absorbance at 450 nm wasrecorded (Multiskan MCC, Fisher Scientific). Increasing concentra-tions (3–200 ng/ml) of an activated MMP-2/MMP-9 standard (catalogno. CC073, Chemicon) were included in the assay.

Levels of the endogenous tissue inhibitor of metalloproteinses(TIMP)-1 and TIMP-4 were determined by immunoblotting (17).Briefly, LV extracts (10 �g) were loaded onto 4–12% Bis-Tris gels,electrophoretically separated, and transferred to a nitrocellulose mem-brane. After they were blocked and washed, the membranes wereincubated for 1 h with the primary antibody: Ab 8116 for TIMP-1 at1:5,000 dilution and Ab 816 for TIMP-4 at 1:5,000 dilution (bothfrom Chemicon). Recombinant TIMP-1 (catalog no. WBC022, R & D

H2217MMP PROMOTER INDUCTION WITH MI

AJP-Heart Circ Physiol • VOL 291 • NOVEMBER 2006 • www.ajpheart.org

Systems) and TIMP-4 (catalog no. HTIMP4, Triple Point Biologics)standards were added to each respective immunoblot.

Immunohistochemical staining. The fixed hearts previously stainedfor �-Gal were embedded in paraffin, and thin (4-�m) sections wereobtained. The sections were placed on slides, deparaffinized, andstained with eosin or used for immunohistochemistry. The eosin-stained sections in which LV morphology was maintained were usedto determine in situ LV geometrical indexes of endocardial diameter,endocardial perimeter, endocardial length of the MI segment, and LVwall thicknesses at the MI and remote regions by digital planimetry ofphotomicrographs (�1 objective, Zeiss Axioskop and SigmaScanversion 4.0). Immunohistochemical studies on the paraffin-embeddedsections were performed using primary antibodies directed against-smooth muscle actin (catalog no. A-2547, Sigma; 1:1,000 dilution)or the AP-1 transcription factors c-fos, c-jun, FosB, Fra1, and JunB(all 1:100 dilution; Santa Cruz Biotechnologies). In addition, immu-nohistochemical staining for MMP-2 (Ab 809 at 1:100 dilution;Chemicon) was performed for hearts from the MMP-2 reporter miceand against MMP-9 (Ab 19047, 1:100 dilution; Chemicon) for heartsfrom the MMP-9 reporter mice. Briefly, myocardial sections thatcontained the scar were blocked and then incubated for 60 min atroom temperature with primary antibodies. After incubation withsecondary antisera, MMP-2 and MMP-9 were detected by visualiza-tion of a 3�,3�-diaminobenzidine-hydrogen peroxide substrate (VectorLaboratories). The sections were imaged on an inverted microscope(Axioskop-2, Zeiss), and the images were digitized (AxioCam MRc,Zeiss). For detection of colocalization of cell types with �-Galstaining, cryosections (5 �m) were obtained from the OTC com-pound-embedded samples and immunostained for �-Gal (Ab 9361,Abcam) concomitantly with antibodies for one of the following:macrophages (MAC-3, CL8943A, Cedarlane; 1:200 dilution), lym-phocytes (CD4, CL012A, Cedarlane; 1:200 dilution), fibroblasts [dis-coidin domain receptor-2 (13), GEA4023–1, Genex; 1:250 dilution],and myofibroblasts (-smooth muscle actin, Ab 5694, Abcam;1:1,000 dilution). Secondary antibodies included Alexa 633-conju-gated anti-chicken antibodies for �-Gal (catalog no. A21103, Molec-ular Probes) and Alexa 488-conjugated secondary antibodies for theother markers (catalog nos. A11006 and A11008, Molecular Probes).A laser confocal microscope (Leica SP2, Hollings Cancer CenterMolecular Imaging Core Facility, Medical University of South Caro-

lina) was used to capture low-power (�10) differential interferencecontrast images and high-power (�63) fluorescent images of thesections. Negative controls for the fixed- and frozen-section immu-nohistochemistry included substitution with nonimmune antisera.

Data analysis. From the in situ measurements of LV morphomet-rics, relative MI size was computed as the ratio of the endocardiallength of the MI segment to LV endocardial perimeter. A wallthickness ratio was computed as the ratio of the thickness at the MIregion to the maximum thickness remote from the MI (18). Changesin heart mass indexes, �-Gal staining, and �-Gal activity werecompared between the MMP-2 and MMP-9 reporter mice usingtwo-way analysis of variance, where the two main blocks werereporter group and time after MI. Pairwise comparisons were per-formed by Bonferroni’s adjusted t-test. For analysis of �-Gal activity,gelatinase activity, and TIMP-1 and TIMP-4 levels, values obtainedfrom the non-MI reference control samples were averaged and used tonormalize values obtained in the MI groups as well as in the non-MIgroup. Therefore, changes in these biochemical indexes were ex-pressed as percent difference from non-MI values and compared withnormalized non-MI control values using a two-sided t-test. Statisticaltests were performed using BMDP (version 7.0) or Stata (StataIntercooled, version 8.0) software packages. Values are means � SE.P � 0.05 was considered to be statistically significant.

RESULTS

MI was induced in a combined total of 124 mice thatelaborated the reporter construct for MMP-2 and MMP-9 genepromoters. A total of 13 [9 male and 4 female, P � 0.143 (by2 analysis)] mice died in the post-MI period: 6 died at 3 daysafter MI, 2 at 5 days after MI, 1 at 6 days after MI, 2 at 7 daysafter MI, and 1 each on 9 and 13 days after MI. The finalsample size of mice at each time after MI is provided in Table1. When indexed to body weight, heart mass was increasedfrom acute values by 1 day after MI and peaked between 7 and14 days after MI for the MMP-9 and MMP-2 reporter groups(Table 1). Echocardiographic determinations of LV geometryand function were determined in a subset of mice from theMMP-9 reporter group. Representative long-axis two-dimen-

Table 1. Time-dependent changes in heart mass indexes and intensity of positive �-Gal stainingin (MMP-2 and MMP-9 reporter mice after MI)

Days After MI

0 (Acute) 1 3 7 14 28

Heart mass, mgMMP-2 reporter 162�14 176�14 210�9b 215�14b 235�14b,c 229�10b,c

MMP-9 reporter 185�11 183�13 208�14 245�11b,c 252�12b,c,d 235�10b,c,d

Heart mass/body wt, mg/gMMP-2 reporter 4.5�0.1 5.4�0.2b 5.9�0.4b 6.1�0.2b 6.4�0.3b,c 5.6�0.2b

MMP-9 reporter 5.1�0.2* 6.1�0.2a,b 7.0�0.3*#� 8.4�0.6a,b,c,d 7.1�0.5b,c,e 7.4�0.4a,b,c,e

LV epicardial area, mm2

MMP-2 reporter 32.7�2.3 37.8�1.6 45.4�2.8b,c 44.8�2.8b 45.9�2.4b,c 46.9�2.7b,c

MMP-9 reporter 37.9�2.6 41.7�2.8 35.6�2.8a 41.2�2.9 44.5�2.0c 42.3�3.8Area of �-Gal staining, mm2

MMP-2 reporter 0.3�0.3 4.7�2.2b 16.1�2.9b,c 27.4�3.2b,c,d 28.0�3.6b,c,d 20.5�2.6b,c,f

MMP-9 reporter 0.4�0.3 0.2�0.2 7.1�2.9a,b,c 20.5�2.1b,c,d 13.2�2.5a,b,c,e 11.1�1.6a,b,c,e

Intensity of �-Gal staining, �103 pixels � luminosityMMP-2 reporter 84�84 1,291�630 3,225�573b,c 6,898�409b,c,d 7,911�221b,c,d 4,527�513b,c,e,f

MMP-9 reporter 114�80 71�53 1,949�729b,c 7,106�910b,c,d 4,133�929a,b,c,d,e 3,138�486b,c,e

nMMP-2 reporter 5 6 7 12 13 6MMP-9 reporter 9 6 10 14 12 11

Values are means � SE; n, number of mice. �-Gal, �-galactosidase; MMP, matrix metalloproteinase; MI, myocardial infarction; LV, left ventricular. aP �0.05 vs. MMP-2 reporter. bP � 0.05 vs. acute. cP � 0.05 vs. 1 day. dP � 0.05 vs. 3 days. eP � 0.05 vs. 7 days. fP � 0.05 vs. 14 days.

H2218 MMP PROMOTER INDUCTION WITH MI


sional-mode echocardiographic images from a mouse at base-line and 28 days after MI are provided in Fig. 1. In these mice,LV end-diastolic diameter increased from 4.2 � 0.1 mm beforeMI to 5.8 � 0.2 mm 28 days after MI (P � 0.05). Concomi-tantly, LV free wall thickness decreased from 0.77 � 0.01 mmbefore MI to 0.56 � 0.02 mm (P � 0.05) 28 days after MI.End-diastolic volume increased from 59 � 2 �l before MI to164 � 17 �l (P � 0.05) 28 days after MI, and LV ejectionfraction decreased from 53 � 2% before MI to 13 � 3% (P �0.05) 28 days after MI.

LV geometry was measured from in situ sections. In theMMP-2 reporter mice, endocardial dimension at the base of theLV was increased from acute values of 3.0 � 0.2 mm to 4.2 �0.4 mm 7 days after MI (P � 0.05) and remained increased(4.9 � 1.0 mm) 28 days after MI (P � 0.05 vs. acute).Similarly, in the MMP-9 reporter mice, LV endocardial dimen-sion increased from acute values of 3.2 � 0.2 mm to 4.6 � 0.4mm 7 days after MI (P � 0.05) and remained increased (5.2 �0.5 mm) 28 days after MI (P � 0.05 vs. acute). However, therewere no differences in any of the in situ morphometric mea-surements between the MMP-2 and MMP-9 reporter groups.Because the background strain (CD-1) of these mice was thesame and there was no difference in the morphometric mea-surement between the two groups, these data were pooled, andthe results are summarized in Table 2. The endocardial lengthof the MI segment was measured by determination of theextent of myocyte necrosis. In the sections obtained 1 h afterMI (acute), myocyte necrosis was not readily observed, and MIsize could not be computed. MI size was increased by 3 daysafter MI and increased further by 7 days after MI. LV wallthickness at the MI region was lower than acute values from 7days after MI, whereas wall thickness diametrically across theMI region was increased from acute values by 7 days after MI.Consequently, the ratio of wall thickness at the MI to wall

thickness at remote regions (18) decreased from acute values of0.86 � 0.06 to 0.69 � 0.05 by 7 days after MI (P � 0.05) andfurther decreased to 0.39 � 0.05 28 days after MI (P � 0.05 vs.acute and 7 days).

MMP-2 and MMP-9 promoter activation was assayed as theextent of elaboration of the �-Gal reporter. In the non-MIhearts from either group, �-Gal staining could not be visuallydetected on the LV. Representative photographs of �-Gal-stained hearts at each time after MI are shown in Fig. 2. �-Galstaining increased through 7 and 14 days after MI in theMMP-9 and MMP-2 reporter groups, respectively. Quantita-tive values for the area of positive �-Gal staining are providedin Table 1. �-Gal staining area normalized to LV epicardialarea in the MMP-2 reporter group was increased by 1 day afterMI, increased further through 7 days after MI, and fell there-after (Fig. 3, top). MMP-9 promoter activation increased fromacute and 1-day post-MI values by 3 days and peaked 7 daysafter MI. �-Gal activity was biochemically determined toprovide a sensitive assay for detection of post-MI MMP-2 andMMP-9 promoter activation. Compared with non-MI referencevalues, �-Gal activity in the MMP-2 reporter group increasedby 1 day after MI, further increased by 7 days after MI, andremained elevated thereafter. In the MMP-9 reporter group,�-Gal activity was increased from reference control and acutevalues by 3 days and peaked 7 days after MI. At 7 days afterMI, �-Gal staining was higher in the MI and border regionsthan in the remote region for the MMP-2 and MMP-9 reportergroups (Fig. 3, bottom). However, in the MMP-2 reportergroup, there was no difference in �-Gal staining between theMI and border regions. In contrast, �-Gal staining in thepost-MI hearts from the MMP-9 reporter group demonstratedhigher staining intensity at the border region than in the MIregion 7 days after MI (Fig. 3, bottom).

Fig. 1. Representative long-axis echocardiographic imagesof left ventricle (LV) and aorta (Ao) before induction ofmyocardial infarction (MI) and 28 days after MI at enddiastole. Arrows, wall thinning in MI region. Scale bar, 1mm for x- and y-axes.

Table 2. Time-dependent changes in LV morphometrics pooled from MMP-2 and MMP-9 reporter mice after MI

Days after MI

0 (Acute) 1 3 7 14 28

LV diameter at base, mm 3.0�0.2 3.2�0.4 3.7�0.4 4.4�0.2a 4.7�0.4a,b 5.1�0.4a,b,c

LV endocardial perimeter, mm 19.5�1.2 22.8�0.9 29.1�1.4a,b 32.8�2.1a,b 33.9�2.1a,b 38.4�3.7a,b,c,e

MI size, % 20�4 26�2b 33�2b,c 36�3b,c 37�3b,c

Wall thickness, mmMI 0.70�0.09 0.67�0.08 0.64�0.05 0.59�0.06a 0.52�0.06a,b 0.48�0.04a,b,c

Remote 0.81�0.07 0.82�0.09 0.92�0.10 1.10�0.06a 1.32�0.08a,b,c 1.62�0.18a,b,c,d

n 11 9 15 23 16 15

Values are means � SE; n, number of mice. aP � 0.05 vs. acute. bP � 0.05 vs. 1 day. cP � 0.05 vs. 3 days. dP � 0.05 vs. 7 days. eP � 0.05 vs. 14 days.



Fig. 2. Regions of positive �-galactosidase (�-Gal) staining in hearts from matrix metalloproteinase (MMP)-2 and MMP-9 gene promoter reporter mice 1 h (acute) and1, 3, 7, 14, and 28 days after MI induction. In MMP-2 reporter mice, areas of �-Gal staining were evident by 1 day after MI and increased further 3 and 7 days afterMI. In MMP-9 reporter mice, �-Gal staining was observed after 3 days and peaked at 7 days after MI. RV, right ventricle; LA, left atrium. Scale grid represents four1 mm � 1 mm squares.

Fig. 3. Top: area of positive �-Gal staining normalized toLV epicardial area. Sample sizes for MMP-2 and MMP-9reporter mice at each post-MI time point are presented inTable 1. Bottom: spatial localization of MMP-2 and MMP-9gene promoter activation with respect to the center of theMI region 7 days after MI. In MMP-2 and MMP-9 reportergroups, �-Gal staining was higher in MI and border regionsthan in respective MI regions. In MMP-2 reporter group,there was no difference in �-Gal staining between theborder region and the MI region. �-Gal staining was higherat the border region than within the MI region for MMP-9gene promoter reporter group. #P � 0.05 vs. acute (1 h afterMI). �P � 0.05 vs. 1 day after MI. aP � 0.05 vs. 3 daysafter MI. bP � 0.05 vs. 7 days after MI. *P � 0.05 vs.MMP-2 reporter. rP � 0.05 vs. remote region. dP � 0.05 vs.border region.



The gelatinase assay used provided a summation of allgelatinase activity, including that of MMP-2 and MMP-9. Inthe MMP-2 and MMP-9 reporter groups, low levels of gelati-nase activity were detected in the non-MI reference controlsamples (5.8 � 1.2 and 5.1 � 0.8 ng/ml, respectively, P � 0.33between groups). Although gelatinase activity increased afterMI in both groups, there was no difference between groups atany time after MI (lowest P � 0.19 at 3 days after MI).Therefore, the results were pooled and are summarized in Fig.4 (top). Gelatinase activity was increased from non-MI refer-ence control values and acute post-MI values by 1 day after MI(17.9 � 6.6 ng/ml, P � 0.05) and further increased 7 days afterMI (44.1 � 5.8 ng/ml, P � 0.05). At 28 days after MI,gelatinase activity (30.1 � 4.2 ng/ml) remained higher thannon-MI values (P � 0.05) but was lower than values recorded7 days after MI (P � 0.05).

Robust immunoblotting signals (Fig. 4, bottom) for TIMP-1and TIMP-4 were detected at all times after MI and in thenon-MI reference control samples. In the non-MI samples or atany time after MI, there were no group-dependent differencesin TIMP-1 or TIMP-4 levels. Accordingly, the data at eachtime after MI were pooled between the groups, and the resultsare summarized in Fig. 4 (bottom). TIMP-1 levels were in-creased compared with non-MI values 3 days after MI but werelower than non-MI and 3-day post-MI values 14 and 28 daysafter MI. TIMP-4 levels were consistently lower than non-MIor acute values at all times after MI.

Representative photomicrographs of eosin-stained sections 7and 14 days after MI are provided in Fig. 5. Distribution of�-Gal in the MMP-2 reporter mice was detected within the MIregion as well as viable myocardium. In the MMP-9 reportermice, �-Gal staining was localized at the border between theMI and viable myocardial regions. For localization of �-Galelaboration to specific myocardial cell types, as well as nuclearactivation factors (Figs. 6–9), immunoblotting was performedon paraffin-embedded or frozen sections. The specificity ofeach of the primary antibodies was determined by 1) substitu-tion with nonimmune antisera from the same species in whichthe primary antibody was raised and 2) immunohistochemicalstaining of known cell types in non-MI samples. For example,a strong immunohistochemical signal against -smooth muscleactin (Fig. 6) was detected in smooth muscle cells aroundblood vessels in non-MI samples but not when the primaryantibody was substituted with nonimmune antisera. In thepost-MI sections from the MMP-2 reporter group, positivestaining for -smooth muscle actin was detected around somecells, presumably myofibroblasts, that colocalized with �-Galstaining (Fig. 5). A similar colocalization of �-Gal and-smooth muscle actin staining was not readily evident in thepost-MI sections from the MMP-9 reporter group. Immunohis-tological staining for MMP-2 protein in myocardial sectionsfrom the MMP-2 reporter mice and staining for MMP-9 proteinin the MMP-9 reporter mice could be localized adjacent toareas of �-Gal staining 7 and 14 days after MI (Fig. 7).Colocalization of staining for �-Gal and markers for specificcell types was undertaken in double-immunolabeled sections 7and 14 days after MI (Fig. 7). Because �-Gal staining peaked7 days after MI for the MMP-2 and MMP-9 reporter groups,colocalization of �-Gal staining and AP-1 activation factorswas examined 7 days after MI (Fig. 9). A minimum of 20randomized, but separate, high-power fields (HPF) were qual-

itatively assayed for coincident staining of the cell marker orAP-1 activation factor with �-Gal. Coincident staining wasgraded as follows: absent (in �20% of HPF), possible (in20–80% of HPF), and definite (in �80% of HPF). A summaryof cell types and AP-1 activation factors that colocalized with�-Gal staining 7 days after MI is provided in Table 3. In the

Fig. 4. Top: changes in gelatinase activity after MI for MMP-2 and MMP-9reporter groups. Values from non-MI reference control samples were averagedand used to normalize gelatinase activity at each post-MI time point and forreference control samples (horizontal dashed line at 100%). Bottom: represen-tative immunoblots for tissue inhibitor of MMP (TIMP)-1 and TIMP-4 innon-MI reference control (C) samples and at different times after MI. Robustsignals for TIMP-1 and TIMP-4 were detected in all samples. Values obtainedfrom non-MI reference control samples were averaged and used to normalizeTIMP-1 or TIMP-4 levels at each time after MI and for reference controlsamples (horizontal dashed line at 100%). #P � 0.05 vs. non-MI controls andacute (1 h after MI). �P � 0.05 vs. 1 day after MI. aP � 0.05 vs. 3 days afterMI. bP � 0.05 vs. 7 days after MI. cP � 0.05 vs. reference control (100%).



MMP-2 reporter mice, �-Gal staining was prominent in fibro-blasts, myofibroblasts, and, to a lesser degree, in CD4� cells.In the MMP-9 reporter mice, �-Gal staining was colocalizedwith macrophages, CD4� lymphocytes, and fibroblasts (Fig.7). Areas of �-Gal staining colocalized with nuclei that stainedpositive for FosB, Fra1, and JunB in the MMP-2 reporter groupand with nuclei that stained positive for c-jun, c-fos, and FosBin the MMP-9 reporter group (Fig. 9).

DISCUSSION

In the present study, the time-dependent activation ofMMP-2 and MMP-9 gene promoters after MI was examinedusing transgenic reporter constructs (2, 24). The unique andimportant findings of the present study are twofold. 1) Post-MILV remodeling was associated with distinct patterns of MMP-2and MMP-9 promoter activation. Specifically, although induc-tion of the MMP-2 and MMP-9 gene promoters peaked 7 daysafter MI, MMP-2 promoter induction remained higher thanMMP-9 promoter induction from 7 days after MI. Moreover,there were distinct differences in the spatial activation ofMMP-2 and MMP-9 gene promoters with respect to the local-ization in the post-MI myocardium. Specifically, distribution ofthe reporter protein was similar in the MI and border regions inthe MMP-2 gene promoter reporter mice but was higher in theborder region than in the MI region in the MMP-9 reportermice. Moreover, MMP-2 or MMP-9 promoter activation wascolocalized with cells that expressed AP-1 nuclear transcrip-tion factors, which bind to consensus sequences on MMP-2 orMMP-9 gene promoters. 2) The cellular sources and myocar-dial localization for MMP-2 and MMP-9 promoter activationafter MI were different. Specifically, 7 days after MI, MMP-2promoter activation occurred in fibroblasts and myofibroblasts,whereas MMP-9 promoter activation was highest in macro-phages and lymphocytes but was also detected in fibroblasts.These findings provide important insight into the post-MIinduction of the MMP-2 and MMP-9 gene promoters.

In the present study, induction of MMP-2 and MMP-9promoters was assayed as a function of �-Gal elaboration.Therefore, interpretation of findings is premised on the syn-chronal transcription of lacZ and the MMP genes. There are atleast two reasons for the parallel transcription of lacZ andendogenous MMP-2 and MMP-9. 1) The design of the target-ing vectors did not disrupt the MMP-2 or MMP-9 promoterregions, in that only the open reading frames were replaced bythe lacZ cassette (2, 24). 2) In vitro induction of MMP-9caused parallel increases in �-Gal elaboration and endogenousMMP-9 levels (24). Cells derived from mice were capable ofactivating the rabbit MMP-9/�-Gal promoter and the endoge-nous MMP-9 promoter identically (24). Consistent with thesefindings, the post-MI localization of MMP-2 and MMP-9proteins occurred at areas of �-Gal elaboration in the MMP-2and MMP-9 reporter mice, respectively. Thus changes in �-Gallevels likely provided a reliable surrogate index of de novoMMP-2 or MMP-9 transcription after MI.

The gelatinases MMP-2 and MMP-9 have been uniformlydemonstrated to contribute to post-MI remodeling, andchanges in mRNA and protein levels of MMP-2 and MMP-9occur after MI (7, 9, 28, 32, 35). Importantly, causal roles forMMP-2 and MMP-9 in post-MI LV remodeling have beenestablished (12, 15, 16, 21). For example, Ducharme et al. (12)demonstrated that post-MI LV dilation was attenuated inMMP-9-deficient mice, and Hayashidani et al. (15) and Mat-sumura et al. (21) reported a reduction in post-MI mortality inMMP-2-deficient mice. However, MMP levels and activitychange as a function of post-MI duration, and unique temporalprofiles of protein and mRNA expression after MI have beenidentified for several MMP types (7, 9, 15, 28, 35). Consistentwith the time-dependent changes in gelatinase activity ob-served in the present study, Cleutjens et al. (9) reported thattotal gelatinase zymographic levels were highest in rat hearts 7days after MI. However, it must be recognized that a numberof proteolytic systems within the myocardium possess gela-

Fig. 5. Representative photomicrographs of myo-cardium spanning peri-infarct and MI regionsfrom MMP-2 and MMP-9 promoter reporter mice7 and 14 days after MI. In MMP-2 reporter mice,�-Gal staining (blue) was detected within MIregion and was also distributed between viablemyocytes. In MMP-9 reporter mice, �-Gal stain-ing appeared at border regions 7 days after MI. At14 days after MI, �-Gal staining was prominentin cells at perivascular regions (arrowheads).Scale bar, 50 �m.



tinolytic activity. These gelatinolytic systems include the ge-latinases MMP-2 and MMP-9, other MMPs, and members ofthe “disintegrin and metalloproteinase” family (31, 39). Todetermine relative contributions of MMP-2 and MMP-9 inpost-MI remodeling, Tao et al. (35), using gelatin zymography,reported that although MMP-9 zymographic levels were in-creased by 1 day and achieved a peak at 2 days, MMP-2zymographic levels did not increase until �4 days after MI. Inthe present study, MMP-2 promoter activation preceded thedetection of MMP-9 promoter activation. There are severalreasons for the apparent disparity between the results of thepresent study and those reported by Tao et al. 1) Becausesample processing required for zymography results in unfold-ing of the proform, as well as the active form, of the enzymes,

the zymographic levels reported by Tao et al. likely reflectedchanges in total MMP-2 and MMP-9 protein levels, in contrastto “recruitable” gelatinase activity measured in the p-amino-phenylmercuric acetate-activated samples in the present study.The zymographic technique, which allows for the electro-phoretic separation of MMPs on the basis of molecular weight,can differentiate between the heavier proform and the activeform. Such a separation would not be readily appreciated in thegelatinase assay employed in the present study. 2) In the studyby Tao et al., the early post-MI appearance of MMP-9 proteinwas coincident with the appearance of neutrophils (35) thatrelease preformed MMP-9 (19). Therefore, it is possible thatthe early post-MI increase in MMP-9 zymographic levels wasdue to release from existent stores, and not to activation of theMMP-9 gene promoter. Moreover, the myocardial ECM canalso serve as a store for latent MMPs (25, 37), which may beactivated in the early post-MI period. Thus activation of theselatent MMPs, which would not require immediate synthesis,may be a potential explanation for the apparent temporal lagbetween the early post-MI increase in zymographic MMPlevels (35) and the MMP gene promoter activation (presentstudy). 3) MMP-2 is constitutively expressed in myocytes andmyocardial fibroblasts (2, 40), and an early post-MI upregula-tion of MMP-2 transcription from these myocardial cell typescannot be ruled out. 4) As documented by Peterson et al. (28),a temporal discordance between mRNA expression and proteinformation may have occurred after MI. These differencesnotwithstanding, the present study builds on the existingknowledge of post-MI MMP regulation in two important ways.1) This study demonstrated that at least one cause of changesin MMP levels is activation of MMP gene promoters. 2) Thefindings of this study demonstrate distinct spatial differenceswith respect to induction of the MMP-2 and MMP-9 genepromoters after MI.

In general, activation of the promoter region for a particulargene should serve as a surrogate marker of mRNA transcrip-tion of that gene. In the present study, activation of the MMP-9gene promoter, assayed as a function of the elaboration of the�-Gal marker protein, was detected 3 days after MI and peaked7 days after MI. In contrast, Chen et al. (7) reported thatMMP-9 mRNA levels peaked 2 days after MI. However, theremay be several reasons for the discordant transcription/trans-lation of a reporter gene construct and mRNA expression forthe protein of interest. 1) The two gene products, the promoterreporter construct and the native gene, are distinctly different,and the end-point measurements for transcription of these twogene products were different. Specifically, Chen et al. deter-mined the message signal as a readout for MMP-9 transcrip-tion, whereas levels of a surrogate marker protein were mea-sured in the present study. 2) It must be recognized that thesurrogate protein marker must be transcribed and translated.Therefore, posttranscriptional/posttranslational regulation ofthe reporter gene product could potentially introduce latency,i.e., a time delay, with respect to detection of MMP-9 promoteractivation. 3) The MMP-9 reporter construct used in thepresent study, albeit functional, was truncated at �522 basesfrom the start of the transcription site (24). Therefore, it ispossible that transcription factors that bind at more distal sitesmay be responsible for the rapid transcription of endogenousMMP-9 mRNA. In light of the differences between the find-ings of the present study and those reported by Chen et al., a

Fig. 6. Representative low- and high-power photomicrographs demonstratingstaining for -smooth muscle actin in sections from MMP-2 and MMP-9reporter groups 7 and 14 days after MI and in sections from non-MI referencecontrols. Sections that stained positive for -smooth muscle actin (brown) areshown at top, and negative controls, in which nonimmune serum was substi-tuted for primary antibody, are shown at bottom. Red inset squares inlow-power micrographs designate area from which high-power views wereobtained. Scale bars, 20 and 100 �m for high- and low-power photomicro-graphs, respectively.



study that examines the relation between MMP-9 mRNAexpression and elaboration of the �-Gal marker after MI iswarranted.

Because MMP transcription generally occurs “on de-mand” (24), transcriptional regulation represents an impor-tant step in MMP synthesis. The promoter regions of manyof the MMP genes contain binding sites for a number oftranscription factors, including AP-1, Sp-1, and NF- B (2,24). Activation of these transcription factors likely initiatesthe synthesis of a number of proteins, including the MMPs.Yoshiyama et al. (45) reported that, in a rat MI model,activities of AP-1, Sp-1, and NF- B remained similar tocontrol values 1 day after MI, progressively increasedthrough 5 days after MI, and remained elevated through 14days after MI. In the present study, MMP-2 and MMP-9promoter activation followed a time course similar to thatdescribed for the transcription factors by Yoshiyama et al.Therefore, it is likely that transcription factor induction andpost-MI activation of the gene promoters for these MMPswere temporally coincident. However, this issue remainsspeculative, and a study characterizing the role of transcrip-tion factor activation in post-MI MMP-2 or MMP-9 pro-moter activation would be appropriate.

Despite similar substrate specificities, MMP-2 andMMP-9 are separate gene products, and synthesis of thesesimilar enzymes is likely dictated by a differential activationof transcription factors. Specifically, the structure of theMMP-2 promoter differs from that of MMP-9, in that itlacks a proximal AP-1 binding site, previously shown tointeract with the AP-1 components c-fos and c-jun (2).Consequently, until recently, MMP-2 was considered to bean AP-1-unresponsive gene. However, a recent in vitrostudy demonstrated that MMP-2 transcription was inducibleby neurohormones and cytokines and that this response was

magnified by hypoxia (2). Promoter mapping studies local-ized the hypoxia response to a distal AP-1 site that bindsFosB/JunB heterodimers (2). In the present study, post-MIMMP-2 promoter activation colocalized with the AP-1 fac-tors FosB, Fra1, and JunB, whereas MMP-9 promoter acti-vation colocalized with c-fos, c-jun, and FosB. Thus thetemporal differences between MMP-2 and MMP-9 promoteractivation in the early post-MI period were likely due todifferential activation of specific AP-1 factors and stimula-tion of cell-specific signaling cascades. Future studies usingtransgenic AP-1 binding site deletion constructs are requiredto elucidate potential differential effects of AP-1 activationof the MMP-2 and MMP-9 promoters.

In the present study, localization of MMP-9 promoter acti-vation, compared with that of the MMP-2 promoter, occurredat the border region of the MI. It must be recognized that thecellular sources for MMP-2 and MMP-9 are different. Forexample, MMP-2 has been reported to be expressed in cardiacfibroblasts, myofibroblasts, and myocytes, and MMP-9 ismainly expressed in inflammatory cells, including macro-phages and leukocytes (12, 20, 35). The present study builds onthe findings of these previous reports by demonstratingpost-MI activation of the MMP-2 or MMP-9 gene promoters inthe same cell types previously shown to express MMP-2 orMMP-9 proteins. Because fibroblasts and myofibroblasts arepresent within the MI region at later phases of post-MI remod-eling (8, 10, 34), the persistence of MMP-2 promoter activationmay have been due to the continual activation by fibroblastsand myofibroblasts within the MI region and myocytes in theviable remote myocardium. The early post-MI expression ofMMP-9 is considered to play a crucial role in facilitatingcellular migration to the MI region (10). Specifically, Heymanset al. (16) reported a reduction in the number of infiltratingleukocytes 7 days after MI in MMP-9-deficient mice. The

Fig. 7. Representative photomicrographs of MI and border regions from hearts of MMP-2 and MMP-9 promoter mice 7 and 14 days after MI stained usingantibodies against MMP-2 and MMP-9, respectively. In MMP-2 reporter mice, positive staining for MMP-2 protein (brown) could be localized at and aroundcells that elaborated �-Gal (blue-green) 7 and 14 days after MI. In MMP-9 reporter mice, staining for MMP-9 protein (brown) was also observed at and aroundcells that elaborated �-Gal. Negative controls were those in which the primary antibody was substituted for nonimmune sera, and representative photomicro-graphs represent remote myocardium 14 days after MI for MMP-2 reporter group and peri-MI region 14 days after MI for MMP-9 reporter group. Scale bars,20 �m.



relative reduction in MMP-9 promoter activation during thelater post-MI periods likely reflects the resolution of the in-flammatory response within the myocardium (35). Thus thedistinct spatiotemporal patterns for MMP-2 and MMP-9 pro-moter activation after MI were likely due to dynamic changeswith respect to the cellular composition of the MI and peri-infarct regions.

The endogenous TIMPs are important regulators ofMMP-mediated tissue remodeling (10, 11, 17, 28, 39). Ofthe four TIMP types identified, TIMP-1 remains the bestcharacterized, whereas the distribution of TIMP-4 has beendemonstrated to occur at higher concentration in the myo-cardium than in other tissues (14). Therefore, post-MI al-terations in TIMP-1 and TIMP-4 levels were examined inthe present study. Changes in the levels of TIMP-1 andTIMP-4 have been reported within the post-MI myocardium(26, 28, 42). Moreover, this laboratory has reported exac-erbated post-MI LV dilation in TIMP-1-deficient mice (17),suggesting a key role for the TIMPs in attenuating post-MIremodeling. In the present study, an early increase inTIMP-1 levels followed by a reduction to below non-MIvalues was observed in the MMP-2 and MMP-9 reportergroups. TIMP-4 levels were consistently lower than non-MI

or acute values at each subsequent time after MI. Thesignificance of these findings is twofold. 1) The similartime-dependent changes in TIMP-1 and TIMP-4 levels inthe MMP-2 and MMP-9 reporter groups suggest that thegenomic insertion of the reporter constructs did not idiosyn-cratically alter the post-MI remodeling phenotype. 2) Theconcomitant increase in gelatinase activity at the laterpost-MI time points suggests that a stoichiometric imbal-ance between MMPs and TIMPs, due to the persistentactivation of the MMP-2 and MMP-9 promoters and a lossof TIMP-mediated inhibition of the MMPs, contributed tothe progression of post-MI LV remodeling. Nevertheless,TIMP levels were determined from LV homogenates, and aspatial pattern for the distribution of these TIMP typesor cellular sources could not be determined. Future studiesare needed to examine colocalization of TIMP-1, TIMP-4,and the other TIMPs with MMP-2 or MMP-9 promoteractivation.

Limitations and conclusions. Several limitations of thepresent study must be recognized. 1) The transgenic reportermouse lines were useful in detecting the spatiotemporalinduction of MMP-2 or MMP-9 transcription vs. MMP-2 orMMP-9 functionality. In addition, despite the existence of

Fig. 8. Representative double-immunostained confocal photomicrographs of MI region from MMP-2 and MMP-9 reporter mice 7 and 14 days after MI for �-Galand for markers of myofibroblasts [-smooth muscle actin (-SMA)], fibroblasts [discoidin domain receptor-2 (DDR-2)], macrophages (MAC-3), andlymphocytes (CD4). �-Gal staining is shown as red and staining of respective cell markers as green; areas of colocalization are orange/yellow. Low-powerphotomicrographs using differential interference contrast imaging are shown at left; areas enclosed in red boxes are areas from which high-powerphotomicrographs were obtained. Scale bars, 20 and 100 �m for high- and low-power micrographs, respectively.



an associative relation between AP-1 activation and changesin MMP-2 and MMP-9 levels in humans (1), there may bespecies-dependent differences in MMP transcriptional reg-ulation between humans and mice. 2) In the present study,post-MI changes in total gelatinase activity were deter-mined. In light of the findings of this study, a future studythat differentiates between the temporal changes in MMP-2and MMP-9 activity and relates these changes to post-MIMMP-2 or MMP-9 gene promoter activation would beappropriate. 3) Male and female mice were included in thepresent study. Although more male than female mice diedafter MI, the sex-based difference in post-MI mortality didnot attain statistical significance. Nevertheless, sex differ-ences with respect to post-MI remodeling and rupture ratesare becoming increasingly apparent (3– 6, 43). Cavasin et al.(4 – 6) reported a greater degree of post-MI LV dilation inmale mice. In addition, Bridgman et al. (3) recently dem-

onstrated that sex-based differences in post-MI LV remod-eling were associated with higher post-MI levels of atrialnatriuretic peptide in the female mice. Because sample sizesat each time after MI are relatively small in the presentstudy, subset analyses to determine sex-based differenceswith respect to post-MI activation of the MMP-2 or MMP-9gene promoters would likely be underpowered and were notperformed. 4) Colocalization of MMP-2 and MMP-9 pro-moter activation was examined in only a specific subset ofcell types and at specific times after MI. Because cellularsources for MMP-2 and MMP-9 may differ during thecourse of post-MI LV remodeling, a study in which MMP-2and MMP-9 promoter activation is examined in a morecomplete portfolio of cell types during the early and latephases of post-MI remodeling is warranted. Nevertheless,this study provided unique visual evidence for post-MIinduction of the MMP-2 and MMP-9 gene promoters in aregion- and time-specific manner. Specifically, MMP-2 pro-moter activation increased through 7 days after MI andremained elevated through 28 days after MI. Moreover,MMP-2 promoter activation was not limited to the MIregion but was also identified in the remote myocardium. Incontrast, MMP-9 promoter activation peaked 7 days afterMI and was detected most prominently at the border be-tween the MI and the remote regions. Taken together, theseunique findings demonstrate that activation of the MMP-2and MMP-9 gene promoters occurs in distinct spatial rela-tion with reference to the MI region and follows character-istic time-dependent patterns after MI.

ACKNOWLEDGMENTS

The authors acknowledge the assistance of Sarah S. Camens, Brooke S.Kaplan, Amy H. Leonardi, Julie E. McLean, I. Matthew Mains, and Stuart M.Saunders.

GRANTS

This study was supported by National Institutes of Health Grants HL-66029(R. Mukherjee), HL-45024, HL-97012, and PO1-HL-48788 (F. G. Spinale),EY-12651 and EY-14801 (M. E. Fini), and PO1-HL-68738 and DK-29776(D. H. Lovett). M. E. Fini also received grant funding from the Walter G. RossChair Foundation.

Fig. 9. Representative photomicrographs of MI region from MMP-2 andMMP-9 reporter mice 7 days after MI counterstained for activator protein-1(AP-1) transcription factors c-fos, c-jun, FosB, Fra1, and JunB. �-Gal stainingis blue-green. In MMP-2 reporter mice, �-Gal elaboration was colocalizedwith positive staining for FosB, Fra1, and JunB. In MMP-9 reporter mice,�-Gal elaboration was colocalized to positive staining for c-fos, c-jun, FosB,and JunB. Insets: higher-magnification images of highlighted portions onrespective micrographs. Scale bars, 10 �m.

Table 3. Colocalization of MMP-2 and MMP-9 promoteractivation with myocardial cell types and AP-1 transcriptionfactors 7 days after MI

MMP-2 Reporter MMP-9 Reporter

Cell typesMacrophages � �CD4� lymphocytes �/� �Fibroblasts � �Myofibroblasts � �

AP-1 Transcription factorsc-fos � �c-jun �/� �FosB � �Fra1 � �/�JunB � �/�

�, Definite colocalization [�80% of randomized high-power fields (HPF)];�, no colocalization (coincident staining in �20% of HPF); �/�, coincidentstaining in 20–80% of HPF.



REFERENCES

1. Aljada A, Ghanim H, Mohanty P, Hofmeyer D, Tripathy D, andDandona P. Hydrocortisone suppresses intranuclear activator-protein-1(AP-1) binding activity in mononuclear cells and plasma matrix metallo-proteinase 2 and 9 (MMP-2 and MMP-9). J Clin Endocrinol Metab 86:5988–5991, 2001.

2. Bergman MR, Cheng S, Honbo N, Piacentini L, Karliner JS, andLovett DH. A functional activating protein 1 (AP-1) site regulates matrixmetalloproteinase 2 (MMP-2) transcription by cardiac cells through inter-actions with JunB-Fra1 and JunB-FosB heterodimers. Biochem J 369:485–496, 2003.

3. Bridgman P, Aronovitz MA, Kakkar R, Oliverio MI, Coffman TM,Rand WM, Konstam MA, Mendelsohn ME, and Patten RD. Gender-specific patterns of left ventricular and myocyte remodeling followingmyocardial infarction in mice deficient in the angiotensin II type 1areceptor. Am J Physiol Heart Circ Physiol 289: H586–H592, 2005.

4. Cavasin MA, Sankey SS, Yu AL, Menon S, and Yang XP. Estrogen andtestosterone have opposing effects on chronic cardiac remodeling andfunction in mice with myocardial infarction. Am J Physiol Heart CircPhysiol 284: H1560–H1569, 2003.

5. Cavasin MA, Tao Z, Menon S, and Yang XP. Gender differences incardiac function during early remodeling after acute myocardial infarctionin mice. Life Sci 75: 2181–2192, 2004.

6. Cavasin MA, Tao ZY, Yu AL, and Yang XP. Testosterone enhancesearly cardiac remodeling after myocardial infarction, causing rupture anddegrading cardiac function. Am J Physiol Heart Circ Physiol 290: H2043–H2050, 2006.

7. Chen J, Tung CH, Allport JR, Chen S, Weissleder R, and Huang PL.Near-infrared fluorescent imaging of matrix metalloproteinase activityafter myocardial infarction. Circulation 111: 1800–1805, 2005.

8. Cleutjens JP and Creemers EE. Integration of concepts: cardiac extra-cellular matrix remodeling after myocardial infarction. J Card Fail 8:S344–S348, 2002.

9. Cleutjens JP, Kandala JC, Guarda E, Guntaka RV, and Weber KT.Regulation of collagen degradation in the rat myocardium after infarction.J Mol Cell Cardiol 27: 1281–1292, 1995.

10. Creemers EE, Cleutjens JP, Smits JF, and Daemen MJ. Matrixmetalloproteinase inhibition after myocardial infarction: a new approachto prevent heart failure? Circ Res 89: 201–210, 2001.

11. Creemers EE, Davis JN, Parkhurst AM, Leenders P, Dowdy KB,Hapke E, Hauet AM, Escobar PG, Cleutjens JP, Smits JF, DaemenMJ, Zile MR, and Spinale FG. Deficiency of TIMP-1 exacerbates LVremodeling after myocardial infarction in mice. Am J Physiol Heart CircPhysiol 284: H364–H371, 2003.

12. Ducharme A, Frantz S, Aikawa M, Rabkin E, Lindsey M, Rohde LE,Schoen FJ, Kelly RA, Werb Z, Libby P, and Lee RT. Targeted deletionof matrix metalloproteinase-9 attenuates left ventricular enlargement andcollagen accumulation after experimental myocardial infarction. J ClinInvest 106: 55–62, 2000.

13. Goldsmith EC, Hoffman A, Morales MO, Potts JD, Price RL, Mc-Fadden A, Rice M, and Borg TK. Organization of fibroblasts in theheart. Dev Dyn 230: 787–794, 2004.

14. Greene J, Wang M, Liu YE, Raymond LA, Rosen C, and Shi YE.Molecular cloning and characterization of human tissue inhibitor ofmetalloproteinase 4. J Biol Chem 271: 30375–30380, 1996.

15. Hayashidani S, Tsutsui H, Ikeuchi M, Shiomi T, Matsusaka H,Kubota T, Imanaka-Yoshida K, Itoh T, and Takeshita A. Targeteddeletion of MMP-2 attenuates early LV rupture and late remodeling afterexperimental myocardial infarction. Am J Physiol Heart Circ Physiol 285:H1229–H1235, 2003.

16. Heymans S, Luttun A, Nuyens D, Theilmeier G, Creemers E, MoonsL, Dyspersin GD, Cleutjens JP, Shipley M, Angellilo A, Levi M, NubeO, Baker A, Keshet E, Lupu F, Herbert JM, Smits JF, Shapiro SD,Baes M, Borgers M, Collen D, Daemen MJ, and Carmeliet P. Inhibi-tion of plasminogen activators or matrix metalloproteinases preventscardiac rupture but impairs therapeutic angiogenesis and causes cardiacfailure. Nat Med 5: 1135–1142, 1999.

17. Ikonomidis JS, Hendrick JW, Parkhurst AM, Herron AR, EscobarPG, Dowdy KB, Stroud RE, Hapke E, Zile MR, and Spinale FG.Accelerated LV remodeling after myocardial infarction in TIMP-1-defi-cient mice: effects of exogenous MMP inhibition. Am J Physiol Heart CircPhysiol 288: H149–H158, 2005.

18. Jugdutt BI, Joljart MJ, and Khan MI. Rate of collagen depositionduring healing and ventricular remodeling after myocardial infarction inrat and dog models. Circulation 94: 94–101, 1996.

19. Kjeldsen L, Bainton DF, Sengelov H, and Borregaard N. Structural andfunctional heterogeneity among peroxidase-negative granules in humanneutrophils: identification of a distinct gelatinase-containing granule sub-set by combined immunocytochemistry and subcellular fractionation.Blood 82: 3183–3191, 1993.

20. Lindsey ML, Gannon J, Aikawa M, Schoen FJ, Rabkin E, Lopresti-Morrow L, Crawford J, Black S, Libby P, Mitchell PG, and Lee RT.Selective matrix metalloproteinase inhibition reduces left ventricular re-modeling but does not inhibit angiogenesis after myocardial infarction.Circulation 105: 753–758, 2002.

21. Matsumura S, Iwanaga S, Mochizuki S, Okamoto H, Ogawa S, andOkada Y. Targeted deletion or pharmacological inhibition of MMP-2prevents cardiac rupture after myocardial infarction in mice. J Clin Invest115: 599–609, 2005.

22. McKay RG, Pfeffer MA, Pasternak RC, Markis JE, Come PC, NakaoS, Alderman JD, Ferguson JJ, Safian RD, and Grossman W. Leftventricular remodeling after myocardial infarction: a corollary to infarctexpansion. Circulation 74: 693–702, 1986.

23. Mertens PR, Steinmann K, Alfonso-Jaume MA, En-Nia A, Sun Y, andLovett DH. Combinatorial interactions of p53, activating protein-2, andYB-1 with a single enhancer element regulate gelatinase A expression inneoplastic cells. J Biol Chem 277: 24875–24882, 2002.

24. Mohan R, Rinehart WB, Bargagna-Mohan P, and Fini ME. GelatinaseB/lacZ transgenic mice, a model for mapping gelatinase B expressionduring developmental and injury-related tissue remodeling. J Biol Chem273: 25903–25914, 1998.

25. Montfort I and Perez-Tamayo R. The distribution of collagenase innormal rat tissues. J Histochem Cytochem 23: 910–920, 1975.

26. Mukherjee R, Brinsa TA, Dowdy KB, Scott AA, Baskin JM, De-schamps AM, Lowry AS, Escobar GP, Lucas DG, Yarbrough WM,Zile MR, and Spinale FG. Myocardial infarct expansion and matrixmetalloproteinase inhibition. Circulation 107: 618–625, 2003.

27. Nichols K, Lefkowitz D, Faber T, Folks R, Cooke D, Garcia EV, YaoSS, DePuey EG, and Rozanski A. Echocardiographic validation of gatedSPECT ventricular function measurements. J Nucl Med 41: 1308–1314,2000.

28. Peterson JT, Li H, Dillon L, and Bryant JW. Evolution of matrixmetalloprotease and tissue inhibitor expression during heart failure pro-gression in the infarcted rat. Cardiovasc Res 46: 307–315, 2000.

29. Pfeffer MA and Braunwald E. Ventricular remodeling after myocardialinfarction. Experimental observations and clinical implications. Circula-tion 81: 1161–1172, 1990.

30. Rohde LE, Ducharme A, Arroyo LH, Aikawa M, Sukhova GH,Lopez-Anaya A, McClure KF, Mitchell PG, Libby P, and Lee RT.Matrix metalloproteinase inhibition attenuates early left ventricular en-largement after experimental myocardial infarction in mice. Circulation99: 3063–3070, 1999.

31. Roy R, Wewer UM, Zurakowski D, Pories SE, and Moses MA. ADAM12 cleaves extracellular matrix proteins and correlates with cancer statusand stage. J Biol Chem 279: 51323–51330, 2004.

32. Seeland U, Kouchi I, Zolk O, Itter G, Linz W, and Bohm M. Effect oframipril and furosemide treatment on interstitial remodeling in post-infarction heart failure rat hearts. J Mol Cell Cardiol 34: 151–163, 2002.

33. Sun Y, Cleutjens JP, Diaz-Arias AA, and Weber KT. Cardiac angio-tensin converting enzyme and myocardial fibrosis in the rat. CardiovascRes 28: 1423–1432, 1994.

34. Sutton MG and Sharpe N. Left ventricular remodeling after myocardialinfarction: pathophysiology and therapy. Circulation 101: 2981–2988,2000.

35. Tao ZY, Cavasin MA, Yang F, Liu YH, and Yang XP. Temporalchanges in matrix metalloproteinase expression and inflammatory re-sponse associated with cardiac rupture after myocardial infarction in mice.Life Sci 74: 1561–1572, 2004.

36. Tortoledo FA, Quinones MA, Fernandez GC, Waggoner AD, andWinters WL Jr. Quantification of left ventricular volumes by two-dimensional echocardiography: a simplified and accurate approach. Cir-culation 67: 579–584, 1983.

37. Tyagi SC, Matsubara L, and Weber KT. Direct extraction and estima-tion of collagenase(s) activity by zymography in microquantities of ratmyocardium and uterus. Clin Biochem 26: 191–198, 1993.



38. Vincenti MP. The matrix metalloproteinase (MMP) and tissue inhibitor ofmetalloproteinase (TIMP) genes. Transcriptional and posttranscriptionalregulation, signal transduction and cell-type-specific expression. MethodsMol Biol 151: 121–148, 2001.

39. Visse R and Nagase H. Matrix metalloproteinases and tissue inhibitors ofmetalloproteinases: structure, function, and biochemistry. Circ Res 92:827–839, 2003.

40. Wang W, Schulze CJ, Suarez-Pinzon WL, Dyck JR, Sawicki G, andSchulz R. Intracellular action of matrix metalloproteinase-2 accounts foracute myocardial ischemia and reperfusion injury. Circulation 106: 1543–1549, 2002.

41. Weber KT, Sun Y, Katwa LC, Cleutjens JP, and Zhou G. Connectivetissue and repair in the heart. Potential regulatory mechanisms. Ann NYAcad Sci 752: 286–299, 1995.

42. Wilson EM, Moainie SL, Baskin JM, Lowry AS, Deschamps AM,Mukherjee R, Guy TS, John-Sutton MG, Gorman JH III, Edmunds

LH Jr, Gorman RC, and Spinale FG. Region- and type-specific induc-tion of matrix metalloproteinases in post-myocardial infarction remodel-ing. Circulation 107: 2857–2863, 2003.

43. Wu JC, Nasseri BA, Bloch KD, Picard MH, and Scherrer-Crosbie M.Influence of sex on ventricular remodeling after myocardial infarction inmice. J Am Soc Echocardiogr 16: 1158–1162, 2003.

44. Yarbrough WM, Mukherjee R, Brinsa TA, Dowdy KB, Scott AA,Escobar GP, Joffs C, Lucas DG, Crawford FA Jr, and Spinale FG.Matrix metalloproteinase inhibition modifies left ventricular remodelingafter myocardial infarction in pigs. J Thorac Cardiovasc Surg 125:602–610, 2003.

45. Yoshiyama M, Omura T, Takeuchi K, Kim S, Shimada K, Yam-agishi H, Teragaki M, Akioka K, Iwao H, and Yoshikawa J.Angiotensin blockade inhibits increased JNKs, AP-1 and NF- B DNA-binding activities in myocardial infarcted rats. J Mol Cell Cardiol 33:799 – 810, 2001.



Selective spatiotemporal induction of matrix metalloproteinase … · Selective spatiotemporal induction of matrix metalloproteinase-2 and matrix metalloproteinase-9 transcription

Documents