Macrophage‐derived MMP‐9 enhances the progression of atherosclerotic lesions and vascular calcification in transgenic rabbits

Abstract Matrix metalloproteinase‐9 (MMP‐9), or gelatinase B, has been hypothesized to be involved in the progression of atherosclerosis. In the arterial wall, accumulated macrophages secrete considerable amounts of MMP‐9 but its pathophysiological functions in atherosclerosis have not been fully elucidated. To examine the hypothesis that macrophage‐derived MMP‐9 may affect atherosclerosis, we created MMP‐9 transgenic (Tg) rabbits to overexpress the rabbit MMP‐9 gene under the control of the scavenger receptor A enhancer/promoter and examined their susceptibility to cholesterol diet‐induced atherosclerosis. Tg rabbits along with non‐Tg rabbits were fed a cholesterol diet for 16 and 28 weeks, and their aortic and coronary atherosclerosis was compared. Gross aortic lesion areas were significantly increased in female Tg rabbits at 28 weeks; however, pathological examination revealed that all the lesions of Tg rabbits fed a cholesterol diet for either 16 or 28 weeks were characterized by increased monocyte/macrophage accumulation and prominent lipid core formation compared with those of non‐Tg rabbits. Macrophages isolated from Tg rabbits exhibited higher infiltrative activity towards a chemoattractant, MCP‐1 in vitro and augmented capability of hydrolysing extracellular matrix in granulomatous tissue. Surprisingly, the lesions of Tg rabbits showed more advanced lesions with remarkable calcification in both aortas and coronary arteries. In conclusion, macrophage‐derived MMP‐9 facilitates the infiltration of monocyte/macrophages into the lesions thereby enhancing the progression of atherosclerosis. Increased accumulation of lesional macrophages may promote vascular calcification.


| INTRODUC TI ON
Matrix metalloproteinases (MMPs), first described in 1962, 1 are a family of zinc-dependent endopeptidases with more than 23 kinds reported in humans. 2 MMPs play crucial roles in several physiological processes, including tissue remodelling, immune functions, reproduction and development, but are also involved in many pathological processes, such as tumour invasion, autoimmune diseases, neurodegenerative disorders and cardiovascular diseases. [3][4][5] MMPs are roughly divided into two major types according to their molecular properties: membrane-anchored and secreted. Furthermore, based on the hydrolysis substrates, secreted-type MMPs can be further classified into collagenases, gelatinases, stromelysins and matrilysins. [5][6][7] In addition to their enzymatic activities, MMPs also exhibit a number of other pathophysiological functions independent upon their ECM hydrolysis. [8][9][10] The gelatinases, including MMP-2 (72-kD) and MMP-9 (92-kD), are different from other MMPs due to a collagen-binding domain within the catalytic domain that is involved in the binding of collagenous substrates, elastin and thrombospondins. 11 MMP-9, also called gelatinase B, was first discovered as an enzyme involved in extracellular matrix (ECM) remodelling by degradation of denatured collagens (gelatins). 12 In the arterial walls, MMP-9 is synthesized and secreted by endothelial cells, smooth muscle cells as well as macrophages and is involved in the regulation of cell survival, migration, inflammation and angiogenesis. 2,13 Increased MMP-9 activity is associated with several diseases such as asthma, 14 systemic lupus erythematosus, 15 abdominal aortic aneurysms, 16 plaque rupture, 17 left ventricular hypertrophy 18 and stroke. 19 Accumulation of monocytes and macrophage-derived foam cells in the intima of large arteries is a hallmark of human and experimental animal atherosclerosis. 20 Macrophages and macrophage-derived foam cells in the arterial wall secrete substantial amounts of MMP-9, in concert with other MMPs, which have been considered to participate in the pathogenesis of atherosclerosis and plaque rupture. 17,21,22 However, it is not clear whether macrophage-derived MMP-9 is directly involved in the initiation or progression or both of atherosclerosis although MMP-9 was detected in the lesions of human atherosclerosis. 23,24 Functional polymorphism in the regulatory region of MMP-9 has been found to be associated with the severity of coronary atherosclerosis, 25 whereas high levels of plasma MMP-9 protein concentrations are closely correlated with acute coronary syndrome. 26,27 However, conflicting results have been reported in terms of MMP-9 function in the pathogenesis of atherosclerosis using apoE knockout (KO) mice. In one report, MMP-9 deficiency protected against cholesterol diet-induced atherosclerosis, 28 but in another, MMP-9 inactivation increased the atherosclerotic plaque growth and progression. 29 In addition, a report indicated that overexpressing human MMP-9 in macrophages increased collagen content in lesions but showed no effects on atherosclerotic lesions in apoE KO mice. 30 Recently, we performed an RNAseq analysis of the aortic lesions of both cholesterol-fed and WHHL rabbits and showed that MMP-9 along with MMP-1 and MMP-12 was predominately up-regulated compared with aortas of normal wild-type rabbits. 31 As MMP-9 is able to degrade several other ECMs in the arterial wall, we hypothesized that elevation of MMP-9 expression may participate in or mediate atherosclerotic lesion formation. To test this hypothesis, we generated transgenic (Tg) rabbits overexpressing rabbit MMP-9 gene specifically in the macrophage lineage and foam cells of atherosclerotic lesions. The rationale of using rabbits for this undertaking is twofold. First, compared with wild-type murine models, rabbits are more sensitive to a cholesterol diet and develop atherosclerosis rapidly. Therefore, it is possible to generate different types of atherosclerotic lesions in cholesterol-fed rabbits. Second, atherosclerotic lesions of cholesterol-fed rabbits are rich in macrophage-derived foam cells, which facilitates the analysis of macrophage functions in the arterial wall. 32 To the best of our knowledge, this is the first report to demonstrate that overexpression of MMP-9 in macrophages is not only involved in the progression of atherosclerosis but also increases vascular calcification.

| Generation and characterization of Tg rabbits
Tg rabbits were generated by the methods established in our laboratory as reported previously. 33,34 The DNA construct used for microinjection was composed of rabbit MMP-9 cDNA under the control of the human scavenger receptor enhancer/promoter along with four augmented capability of hydrolysing extracellular matrix in granulomatous tissue.
Surprisingly, the lesions of Tg rabbits showed more advanced lesions with remarkable calcification in both aortas and coronary arteries. In conclusion, macrophage-derived MMP-9 facilitates the infiltration of monocyte/macrophages into the lesions thereby enhancing the progression of atherosclerosis. Increased accumulation of lesional macrophages may promote vascular calcification.

K E Y W O R D S
atherosclerosis, calcification, hypercholesterolaemia, macrophage, MMP-9, transgenic rabbits copies of the chicken β globin insulator ( Figure 1A), which prevents the position effect of transgene insulators. 35,36 In total, 1032 embryos were injected, and 879 embryos were implanted into 40 recipient female rabbits. Ten recipients gave birth to 26 pups, and among them, 2 pups were found to carry the transgenes by Southern blotting ( Figure 1B). Because mRNA expression levels of MMP-9 analysed by Northern blots were similar in both founders, Tg founder F1 rabbit was mated with wild-type rabbits to produce following progeny. In this study, rabbits at the age of 4 months were used unless otherwise specified. All rabbits were fed either with a chow diet or cholesterol-rich diet containing 0.5% cholesterol and 3% soybean oil (see below). 37 The rabbits were allowed access to the diet and water ad libitum. Plasma levels of total cholesterol (TC), triglycerides (TG) and HDL-cholesterol (HDL-C) were analysed weekly using the methods described previously. 38

| Analysis of MMP-9 expression
To examine the tissue expression of MMP-9 in Tg and non-Tg rabbits, Northern blotting was performed as described previously. 39 In brief, total RNA was isolated from bone marrow, lung, spleen, liver, aorta, heart, kidney, adrenal, pancreas, and both alveolar and peritoneal macrophages (see below) using TRIzol reagent (Life Technologies, Inc) and 10 μg of RNA was denatured in the presence of dimethyl sulfoxide and glyoxal was subjected to electrophoresis in a 1.2% formaldehyde agarose gel, and then transferred to a Nitran nylon membrane and then treated with UV light in the UV crosslinker. The membrane was hybridized with the 32 P-labelled a rabbit MMP-9 cDNA probe synthesized with a Prime-It II random primer labelling kit (Stratagene). The blot was rehybridized with a 32 P-labelled human β-actin probe (Clontech Laboratories) to confirm that equal amounts of RNA were loaded in each lane. To evaluate MMP-9 protein expression and enzymatic activity, we collected elicited peritoneal macrophages from the peritoneal cavity 4 days after injection of 4% Brewer's thioglycollate broth, as described previously. 40 In brief, rabbits were anaesthetized by intramuscular injection of ketamine Com) (50 ng/mL) for 48 hours, and the conditioned media were then collected for Western blotting and gelatin zymographic analysis. 32 The same aliquots of the conditioned media from each group were fractionated by electrophoresis on 10% SDS-polyacrylamide gels, transferred onto a nitrocellulose membrane, then incubated with monoclonal antibody (mAb) (Table S1) against human MMP-9, which cross-reacted with rabbit MMP-9. To evaluate gelatinase activity, we first performed gelatin gel zymographic analysis using the method reported previously. 32 Briefly, a piece of fresh aortic arch was homogenized in ice-cold RIPA. 20 μg proteins were separated by electrophoresis through 8% SDS-polyacrylamide gels containing gelatin (1 mg/mL) (Wako Pure Chemical Industries Ltd., Osaka, Japan) under non-denaturing and non-reducing conditions. After washing in 2.5% After that, sections were washed with PBS to remove traces of OCT.
Quenched fluorescein-conjugated gelatin (Molecular Probes) was used as the substrate. The gelatin (20 µg/mL), which was dissolved in reaction buffer with 10% low gelling agarose (Sigma-Aldrich), was dropped on the top of sections and incubated with reaction buffer in a dark humidity chamber at 37°C for 6 hours. 5 mmol/L EDTA in reaction buffer was used to block gelatinase activity. The sections were then rinsed with PBS and stained with DAPI for staining nuclei.
Slides were then mounted with a cover glass using an aqueous agent.

| Experimental design
Tg rabbits along with sex-and age-matched non-Tg littermates were used for the following experiment. In the first experiment, rabbits (male group: n = 6 for non-Tg and n = 7 for Tg, female group: n = 3 for non-Tg and n = 7 for Tg) were fed a diet containing 0.5% cholesterol and 3% soybean oil for 16 weeks. The lesions of atherosclerosis formed in these rabbits were mainly those of so-called early-stage lesions (fatty streaks), and thus, it was amenable to evaluate the effects of increased MMP-9 expression on the initiation of atherosclerosis. In the second experiment, rabbits (male group: n = 10 for non-Tg and n = 12 for Tg, female group: n = 10 for non-Tg and n = 16 for Tg) were fed the same cholesterol-rich diet for 28 weeks.
Long-duration hypercholesterolaemia in the second experiment in comparison with the first experiment was expected to induce more advanced atherosclerotic lesions, such as fibrous plaques and complicated plaques, in these rabbits. 41,42 Therefore, we were able to address the influence of MMP-9 expression on the progression of atherosclerosis.

| Analysis of aortic atherosclerosis
At the end of cholesterol diet feeding, all rabbits were euthanized by injection of sodium pentobarbital solution (100 mg/kg BW) through an ear vein. The aortas were en face stained by Sudan IV solution for quantitative analysis of the gross atherosclerotic lesion area as described previously. 43 For microscopic quantification of the lesion area, each segment of the aorta arch from all rabbits was cut into cross sections as reported previously. 37,44 Then, all specimens were embedded in paraffin and sections (3 μm) were stained with haematoxylin and eosin (HE) and elastica van Gieson (EVG). For further microscopic evaluation of cellular components and MMP-9 expression in the lesions, serial paraffin sections of the aorta arch were immunohistochemically stained with mAbs against rabbit macrophages (RAM11), α-smooth muscle actin (HHF35), MMP-9 and caspase-3 (Table S1). The following antigen retrieval method was used. Citrate buffer was prepared by mixing 0.1 mol/L citric acid with 0.1 mol/L sodium citrate hydrate solution as 1:4. Then, paraffin-embedded section slides were immersed in the citrate buffer and autoclaved at 120°C for 10 minutes. After that, slides were washed with PBS once and blocked with 10% goat serum at room temperature for 30 minutes. Abs were diluted in 10% goat serum, and slides were incubated with each first Ab at 4°C for overnight and followed by F I G U R E 1 Transgenic construct for generation of transgenic rabbits. The DNA construct used for microinjection was composed of rabbit MMP-9 cDNA under the control of a human scavenger receptor A enhancer/promoter region (5 kb) along with four copies of the chicken β globin insulator (A). Two Tg founders (designated as F1 and F2) were identified by Southern blot analysis using a rabbit MMP-9 cDNA probe (B). Northern blotting analysis revealed that endogenous MMP-9 expression was mainly in the bone marrow of non-Tg rabbits, whereas Tg MMP-9 was expressed in isolated macrophages, bone marrow and lungs. GAPDH was used as an internal control (C). Peritoneal macrophages were isolated from both Tg and non-Tg rabbits (3-mon), and incubated in serum-free medium for 48 h with or without PMA. Conditioned media were collected and analysed for MMP-9 protein and enzymatic activity by Western blotting and zymographic analysis. Each lane represents a sample from an individual animal and n = 3 for each group (D) peroxidase-conjugated goat-antimouse IgG (Histofine Sab-Po(M), Nichirei Bioscience, Inc) for 1 hour at room temperature. Amino-9ethylcarbazole (AEC) (Nichirei Bioscience) was used as a substrate for visualizing the antigen signals and nuclei were stained with haematoxylin. To evaluate Ab specificity, the slides were incubated with mouse non-specific IgG or PBS to replace the first Ab (Figures S11 and S12). Aortic lesions were histologically classified into early-stage lesions (type II lesions: either fatty streaks with foam cells > 60% of the lesions or fibrotic lesions mainly composed by SMCs and ECM with foam cells < 50%) or advanced lesions (type IV atheroma or V fibroatheroma containing typical lipid or necrotic cores with calcification) according to the AHA classification. 45 The lengths of each lesion on each section were measured and quantified as reported previously. 46 In addition, the severity of the aortic calcification was evaluated by measuring the calcification area on each section of the aortic arch based on von Kossa staining. All section images for microscopic quantification were taken with an Olympus BX51 light microscope equipped with a DP70 digital camera (Olympus) and quantified with Lumina Vision V2.04 image analysis software (Mitani Co.). For this undertaking, we defined a colour pixel threshold of immunostaining intensity to detect the AEC-stained red colour by selecting areas first, and then, we used the same threshold to measure colour intensity in each specimen. For analysis of Mϕ, SMC and calcification in cellular distribution, we measured and showed the real positive area. Atherosclerotic quantification was performed by two independent observers blindly. Aortic lesions were also collected and homogenized for gelatinase activity and Western blotting analysis using mAbs against MMPs and tissue inhibitors of matrix metalloproteinase (TIMPs) shown inTable S1. 32 In brief, aortic arch was homogenized in ice-cold RIPA buffer (Thermo Fisher Scientific) supplemented with a proteinase inhibitor cocktail (1% v/v) (Sigma-Aldrich). After centrifugation, the supernatant was collected for measuring protein content by a Bio-Rad protein assay kit. The supernatant was aliquoted for performing Western blotting, zymography and enzymatic assay. For Western blotting analysis, equal amounts of crude protein (30 μg) were fractionated by electrophoresis on 10% SDS-ployacrylamide gels under reducing condition. After that, the proteins were transferred to Bio-Rad's 0.2 μm pore-size nitrocellulose membranes and these membranes were incubated with each Ab (Table S1) at 4°C overnight, and then washed three times with PBST (0.1% Tween 20 in 1X PBS) and reacted with horseradish peroxidase-conjugated secondary Abs, followed by enhanced chemiluminescence detection.

| Analysis of coronary lesions
To assess coronary atherosclerosis, all hearts of both male and female rabbits of the second experiment were cut into five blocks and the block-I, which contains the main trunk of the left coronary artery, was used for coronary lesion analysis. The whole protocol for heart sections was described in details elsewhere 42 and also can be obtained from the Appendix A. Supplementary data published through the following website: http://www.scien cedir ect.com/scien ce/artic le/pii/S0163 72581 4001855). The blocks were cut into 4 serial sections (3 μm thick) at 50 μm intervals, and in total, 4 cuts were conducted. For 4 sections collected from each cut were stained for HE, EVG, and von Kossa or immunohistochemically stained with mAb (RAM11) against macrophages. The lesion size was quantified by EVG-stained specimens and expressed as stenosis % (lesion area/ coronary lumen area). Calcification and Mϕ areas were quantified as aortic lesion analysis described above and expressed as mm 2 .

| Tartrate-resistant acid phosphatase (TRAP) staining
Sections of aorta arch were deparaffinized and rehydrated through graded ethanol to distilled water. The slides were then incubated with TRAP staining solution which contained 0.1 mg/mL naphthol AS-MX (Sigma-Aldrich) and 0.08 mg/mL fast red violet LB salt (Sigma-Aldrich) in the presence of 8 mmol/L sodium tartrate and 30 mmol/L sodium acetate (pH 5.0) at 37°C for 2 hours. After rinsing in distilled water, the slides were mounted with aqueous mounting medium.

| Chemotaxis analysis and carrageenaninduced granuloma
To examine the effects of MMP-9 enzymatic activity on macrophage migration and matrix digestion, we performed chemotaxis analysis and granuloma assays. Chemotaxis assays for alveolar macrophages isolated from Tg and non-Tg rabbits (n = 4 for each group) were performed using Biocoat cell culture inserts coated with laminin (Becton Dickinson Labware, Bedford, MA). The lower compartments were loaded with the same medium containing human recombinant monocyte chemoattractant protein-1 (MCP-1) (Pepro Tech EC) at 10 ng/mL. After 48 hours of incubation (37°C, 5% CO 2 ), the number of macrophages that penetrated the membranes was counted in 10 high-power fields randomly from each well.
We generated a carrageenan-induced granuloma model that mimics the condition of macrophages and foam cells accumulated in atherosclerosis lesions as previously reported, 40  staining. The total area of collagen staining was calculated using the image analysis system as above described.

| Statistical analysis
Statistical analysis was performed using SPSS 16.0 software. All data are expressed as mean ± SEM. Statistical analysis was performed for parametric data by Student's t test and for non-parametric data (calcification study) by the Mann-Whitney U test. In all cases, significance was set as a P-value <.05.

| Generation of Tg rabbits
We generated 2 Tg founder rabbits as shown by Southern blotting analysis ( Figure 1B). One founder rabbit was bred to provide F1 progeny for the current study. Tg rabbits did not exhibit any visible abnormalities in the whole-body appearance and histological examinations did not reveal any pathological changes in heart, lung, spleen, adrenal and liver. Peripheral blood white cells in Tg rabbits were not different from those of non-Tg rabbits (data not shown).
Northern blotting analysis revealed that Tg rabbits expressed prominent MMP-9 mRNA in peritoneal and alveolar macrophages and macrophage-rich organs (lung and bone marrow), whereas in non-Tg rabbits, endogenous MMP-9 mRNA expression was extremely low and only detected in the bone marrow, presumably in bone marrow macrophages ( Figure 1C). To analyse MMP-9 proteins and enzymatic activity, we isolated elicited macrophages from the peritoneal cavity of rabbits. On Western blotting analysis of the conditioned media, the peritoneal macrophages from Tg rabbits secreted approximately 10-fold more (without PMA treatment) and fourfold more (after PMA treatment) MMP-9 protein than control macrophages, as calculated by OD values ( Figure 1D). Gelatin zymography demonstrated that MMP-9 proteins secreted by macrophages are enzymatically active ( Figure 1D).

| Effects of MMP-9 on the initiation of atherosclerosis
We first examined whether MMP-9 overexpression affected the early-stage lesion formation of atherosclerosis in rabbits fed a cholesterol diet for 16 weeks. As shown in Figure S1, both male and female Tg and non-Tg rabbits developed similar hypercholesterolaemia throughout the experiment. The aortic lesions were mainly found in the aortic arch, and the gross lesion area was not significantly different between Tg and non-Tg rabbits of both male and female ( Figure   S2). Histological examinations revealed that the lesions of aortic atherosclerosis in Tg and non-Tg rabbits were predominantly composed of type II lesions (fatty streaks), which were characterized by foam cell accumulation intermingled with small numbers of smooth muscle cells and ECM; however, the number of lesional macrophages was significantly increased by 2.8-fold in male ( Figure 2) and 1.3fold in female Tg rabbits ( Figure S3). In these areas, MMP-9 immunoreactive proteins were co-localized with macrophages in Tg rabbits, whereas MMP-9 was barely stained in the lesions of non-Tg rabbits ( Figure 2 and Figure S3).

| Effects of MMP-9 on the progression of atherosclerosis
In the second experiment, we fed rabbits with a cholesterol diet for 28 weeks to examine whether increased MMP-9 expression influenced the advanced lesion formation. Compared with the experiment 1 in which the lesions were smaller in size and mainly consisted of the fatty streaks, the lesions of the experiment 2 were more extensive and predominately by advanced lesions, such as fibrous plaques and atheroma (types IV-V lesions). 40 The total cholesterol levels of Tg and non-Tg rabbits were essentially similar ( Figure S4 Figure 5B). In macrophage-rich regions, we could also observe many so-called microcalcification-generating matrix vesicles ( Figure 5C). Many lesional macrophages associated with calcification were assumptively apoptotic because they could be stained by both TUNEL and caspase-3 staining ( Figure   S7). Interestingly, marked calcification was also found in coronary arteries: 21.6-fold increase of calcification areas in male Tg rabbits ( Figure 6) and 12.9-fold increase in female Tg rabbits ( Figure S8). In these calcified areas, marked macrophage accumulation was often observed: 8.1-fold increase in male Tg rabbits and 2.5-fold increase in female Tg rabbits compared with non-Tg rabbits. Coronary stenosis was increased in male Tg rabbits even though not significantly different (ns).
In addition to the severe advanced lesions associated with prominent calcification in Tg rabbits described above, macrophages frequently expanded the deep parts of the tunica media or even the adventitial layers ( Figure S9). Each dot represents the data of an individual animal. *P < .05, **P < .01 vs non-Tg rabbits found that carrageenan-induced granulomas of Tg rabbits contained significantly less extracellular matrix content (47% decrease, P < .01) than those of non-Tg rabbits ( Figure 7D). In addition, Western blotting showed that increased MMP-9 was actually accompanied by significant increase of MMP-2, MMP-12, TIMP-1 and TIMP-2 in Tg rabbits ( Figure 7E).

| MMP-9 enhances macrophage migration and digestion of extracellular matrix
To clarify the mechanisms responsible for increased macrophage accumulation in the lesions of Tg rabbits, we compared the capacity of alveolar macrophages isolated from Tg and non-Tg rabbits to invade an immobilized ECM in vitro. In response to the presence of the chemoattractant MCP-1, the number of gel-invading macrophages from Tg rabbits was 2.3-fold greater than that from non-Tg rabbits (P < .01) ( Figure 7F).

| D ISCUSS I ON
Accumulation of macrophage-derived foam cells in the intima of large arteries is a hallmark of atherosclerosis. Macrophages secrete a variety of bioactive substances, such as MMPs and cytokines, thereby playing an important role in lesion formation and progression. 32 In the current study, we focused on MMP-9, an important gelatinase, and generated Tg rabbits that overexpressed MMP-9 in the macrophage lineage to examine whether increased MMP-9 expression affects the initiation and progression of atherosclerosis. For this undertaking, we fed Tg and non-Tg rabbits a cholesterol diet for 16 and 28 weeks.
In the first experiment, we did not see any significant influence of MMP-9 expression on the fatty streak size but the macrophage number was indeed increased in the lesions of Tg rabbits. In this aspect, MMP-9 is similar to other MMPs such as MMP-1 (collagenase) and MMP-12 (elastase) because Tg rabbits expressing either MMP-1 23 or MMP-12 47 in macrophages also failed to increase the fatty F I G U R E 5 Different microscopic patterns of aortic calcification in Tg rabbits. Calcified materials were observed in the centre of fatty streak type lesions (possibly associated with apoptotic macrophages) (the arrows point the calcified area) (A). Calcification was associated with fragile or unstable lesions which have a large lipid core and a thin cap (B). (Note: the surface of the lesions looks 'erupted' possibly caused by artefact during the specimen preparation). Calcium particles (~1 μm) within macrophage accumulation are possibly those of socalled microcalcification-generating matrix vesicles (von Kossa staining on the right) (C) F I G U R E 6 MMP-9 expression increases coronary calcification in Tg rabbits after feeding a cholesterol diet for 28 weeks. Serial paraffin sections of left coronary artery lesions of male non-Tg and Tg rabbits were stained with H&E, mAb against macrophages or von Kossa. The coronary stenosis = lesion area/total lumen area × 100(%) was measured and is expressed as a percentage, and the macrophages (Mϕ) positive area and calcification area in lesions were quantified. Data are expressed as mean ± SEM, n = 8 for each group. Each dot represents the data of an individual animal. *P < .05, **P < .01 vs non-Tg rabbits F I G U R E 7 Demonstration of increased MMP-9 expression in the lesions of Tg rabbits. Proteins isolated from aortic samples were fractionated on 10% SDS-PAGE and immunoblotted with MMP-9 and β-actin was used as an internal control (A, top). Zymography was performed using gelatin as substrate as described in the Methods (A, bottom). n = 3 for each group. To measure the gelatinase enzymatic activity directly in the lesions, we performed in situ zymography as described in the Methods. The gelatinase activity on the sections of aortic lesions was visualized as green, and nuclei were stained as blue (B). White dotted lines indicate the internal elastic lamina. For quantification of gelatinase activity, the homogenate of aortic arch was mixed with quenched fluorescein-conjugated gelatin, and fluorescent density was measured and expressed as the gelatinase activity (C). High gelatinase activity of Tg rabbits was inhibited in the presence of MMP-2/9 inhibitors (C). Values are mean ± SEM, n = 4 for each group. Carrageenan-induced granuloma sections were prepared as described in the Methods (D). Paraffin sections of granulomas were stained by Masson trichrome staining, and the positive areas for collagen fibres (blue or purple) were quantified with an image system (D). Ten high-power fields (HPF) of each section were randomly selected and calculated. Data are expressed as mean ± SEM, n = 4 for each group. **P < .01 vs non-Tg rabbits. Proteins isolated from aortic arch were fractionated on 10% SDS-PAGE and immunoblotted with each mAb against MMP-2, MMP-12, TIMP-1 TIMP-2 and β-actin (E). Data are expressed as mean ± SEM and n = 4 for each group. Alveolar macrophage chemotaxis assays were performed as described in the Methods. Results are expressed as the average number of cells that migrated through immobilized ECM film after 48 h incubation (F). Data are expressed as mean ± SEM and n = 4 for each group. **P < .01 vs non-Tg rabbits steak formation. Meanwhile, macrophage lineage overexpressing MMP-9 in apoE KO mice showed no effects on lesion size. 41 It has been reported that MMP-9 deficiency appears to enhance plaque growth in apoE/MMP-9 double KO mice 30 although MMP-9 deficiency protected against cholesterol diet-induced atherosclerosis in the same double KO mice. 29 In the second experiment, rabbits were fed a cholesterol diet for 28 weeks and developed greater and more advanced lesions (than experiment one) which enabled us to examine whether increased MMP-9 affects the progression of atherosclerosis. Although the aortic arch lesions stained by Sudan IV became almost saturated at 28 weeks, gross lesion areas were significantly increased in female (not male) Tg rabbits. The molecular mechanism for this gender effect of MMP-9 is still unknown but it will be interesting to investigate whether sex hormones interact with MMP-9 in future.
Nevertheless, lesional macrophages were significantly increased in both aortic and coronary atherosclerotic lesions in both male and female Tg rabbits compared with non-Tg rabbits. Accumulation of macrophages in the lesions even extended to the deeper medial tunica or even the adventitia. Although it is still unknown whether these adventitial macrophages come from the inside or outside, it is most likely that they develop as a result of 'outside in' rather than 'inside out' migration. Regardless of this assumption, emigration of monocytes/macrophages requires the degradation of the ECM and increased MMP-9 activity in Tg macrophages undoubtedly leads to enhancement of ECM degradation thereby facilitating the influx of monocyte/macrophages into the lesions. 28 This contention was further supported by our chemotaxis study and granuloma model assay showing that macrophages from Tg rabbits exhibited higher infiltrative activity towards a chemoattractant, MCP-1 in vitro and augmented capability of hydrolysing ECM in granulomatous tissue ( Figure 7). Previous studies using apoE KO mice demonstrated that local expression of catalytic form of MMP-9 by adenoviral vectors enhances plaque rupture and haemorrhage. 16,48 In the current study, we observed so-called "vulnerable sites" in the aortic lesions of Tg rabbits but did not find any plaque ruptures, suggesting pro-MMP-9 activation is a key process of MMP-9-derived plaque rupture.
In spite of this, one of the most striking findings observed in the current study was the demonstration of remarkable calcification in the aortic and coronary lesions in Tg rabbits, which mimicked human advanced atherosclerotic lesions with calcification. 17,49 This finding was initially unexpected and surprising, but it is possible that MMP-9 may mediate vascular calcification through several possible molecular mechanisms. 50,51 First, it has been reported that apoptotic macrophages in the lesions can release microcalcification-generating matrix vesicles into the ECM, eventually resulting in calcification 52 which is supported by our finding that macrophage-rich lesions contained microcalcification. High expression of caspase-3 in the lesions of Tg rabbits allowed us to speculate that increased cell death may be also involved in the process of vascular calcification. Given the fact that overexpression of MMP-9 led to the incremental infiltration of macrophages in the lesions of Tg rabbits, it is presumable that apoptotic macrophages would be increased. It is not clear; however, whether increased MMP-9 expression changes the calcium milieus in the intima, or stimulates macrophages to release more microcalcification-generating matrix vesicles. 53 Second, it is well known that macrophages can differentiate into osteoclasts that help the active resorption of the ECM in the arterial wall. If this process was retarded, calcium deposition in the lesions would be enhanced. This notion was supported by the observation that TRAP staining, a histochemical marker for osteoclast differentiation 54 were significantly reduced in atherosclerotic lesions of Tg rabbits compared with non-Tg rabbits ( Figure S10). Finally, besides macrophages, vascular smooth muscle cells constitute another cellular source that participates in the vascular calcification. For example, increased MMP-9 can lead to incremental hydrolysis of elastin and such soluble elastin-derived peptides from MMP-9 hydrolysis promote osteogenic differentiation of SMCs in arterial wall. 55 In addition to MMP-9, the lesional MMP-2 expression was simultaneously increased in Tg rabbits possibly caused by increased macrophages in the lesions, which may also affect lesion development and vascular calcification. 56,57 It should be pointed out that in addition to MMPs, cathepsins or other proteinases in the lesions may be also involved in the lesion development. 58,59 It will be interesting to investigate how these proteinases play an interactive role in the development of atherosclerosis. Nevertheless, vascular calcification in the atherosclerotic lesions apparently affects arterial stiffness and contractility thus impairing blood flow which leads to ischaemia in the organs. Our current results may support the notion that inhibition of vascular macrophage-derived MMP-9 may become a new potential therapeutics to prevent vascular calcification, 60,61 while this hypothesis remains to be verified in future.
In conclusion, we have successfully created Tg rabbits that overexpress MMP-9 specifically in macrophage lineage. Increased expression of MMP-9 in macrophages not only enhances the formation of advanced atherosclerotic lesions in Tg rabbits, but also leads to a marked vascular calcification. Therefore, macrophage-derived MMP-9 in the arterial wall may exert other physiological functions beyond its classical gelatinase activity. Although the molecular mechanisms remain yet to be clarified, it will be interesting to investigate in the future whether inhibition of MMP-9 functions as a therapeutic method to inhibit vascular calcification.

ACK N OWLED G EM ENTS
We thank Dedong Kang and Yoko Nakagawa for their participation and help in this study.

AUTH O R CO NTR I B UTI O N S
JF and Y.EC designed the study; KN, SK and FM created and bred Tg rabbits. YC, ABW, BN, LC, HS, TK, YY, JZ, JL and EL performed experiments, collected, analysed and interpreted the data. YC and JF wrote the paper. All authors had final approval of manuscript submission.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon reasonable request.