Matrix Metalloproteinase-14 Deficiency in Bone Marrow–Derived Cells Promotes Collagen Accumulation in Mouse Atherosclerotic Plaques

Rupture of atherosclerotic plaques in coronary arteries causes most fatal acute myocardial infarctions.^1,2 Such plaques characteristically have a thin fibrous cap, a large lipid-rich core, and abundant macrophages. Because interstitial collagen confers tensile strength on the fibrous cap, collagenolysis in the fibrous cap likely participates critically in plaque disruption.^3,4 Matrix metalloproteinases (MMPs) can degrade all components of arterial extracellular matrix, and considerable evidence supports their involvement in plaque remodeling. Three members of the MMP family denoted interstitial collagenases (MMP-1, MMP-8, and MMP-13) can cleave triple-helical fibrillar collagen at the neutral pH of the extracellular milieu.^5,6 In addition to these secreted, soluble enzymes, the membrane-anchored or membrane type 1 MMP, MMP-14, also can exhibit collagenase activity.^7–9 Human and animal studies have localized these collagenases in atherosclerotic plaques.^10–14 Mice genetically altered to express collagenase-resistant collagen (Col^R/R) have increased collagen content in atheromata in vivo.¹⁵ Mice lacking MMP-13 accumulate more collagen with a more organized supramolecular structure in plaque than those wild type for this key interstitial collagenase.¹⁶ MMP-14 localizes in human plaque¹⁷ and on peripheral blood monocytes during myocardial infarction.¹⁸ Experimental studies showed enhanced expression of MMP-14 during arterial remodeling after balloon injury¹⁹ and in the myocardium after ischemia/reperfusion.²⁰ However, the contribution of MMP-14 to collagen metabolism during atherosclerosis in mice remains unexplored. Mice genetically deficient in MMP-14 (Mmp14^−/−) typically die 3 weeks after birth of unknown causes. Before death, these mice show significant growth impairment and wasting.^21,22 These findings hamper the analysis of atherogenesis in compound mutant mice for Mmp14^−/− and low-density lipoprotein receptor–deficient mice (Mmp14^−/−/Ldlr^−/−). The present study circumvented this constraint by using bone marrow from Mmp14^−/− mice to examine the role of MMP-14 expressed by bone marrow–derived cells in atheroma formation. Macrophages, which arise from bone marrow, appear to express the bulk of MMP-14 in atheroma.²³ This approach permitted us to test the hypothesis that MMP-14 derived from bone marrow participates in collagen catabolism in plaques using lethally irradiated Ldlr^−/− mice reconstituted with bone marrow from Mmp14^−/− mice.

Methods

Animal Preparation

All experiments conformed to a protocol approved by the Standing Committee on Animals of Harvard Medical School. Mmp14^+/− mice²⁴ backcrossed 7 generations into congenic C57BL/6 mice were crossbred to generate Mmp14^−/− mice and Mmp14^+/+ littermates. Ldlr^−/− C57BL/6 mice 6 to 10 weeks of age (Jackson Laboratories, Bar Harbor, Me) were lethally irradiated (2 times at 700 rad 3 hours apart) and received bone marrow (5×10⁶ cells per mouse IV) derived from Mmp14^−/− (n=29) and Mmp14^+/+ (n=31) donor mice 2 to 3 weeks of age. After bone marrow reconstitution (5 weeks), mice consumed a high-cholesterol diet (Research Diets, New Brunswick, NJ; 1.25% cholesterol, 0% cholate) for 16 weeks (n=26) or 8 weeks (n=34). Mouse plasma was collected for cholesterol measurements at day 0, after bone marrow reconstitution, and at the time of death.¹² For the in vitro study, we used Mmp13^−/− mice with congenic C57BL/6 background.¹⁶ In separate experiments, we verified the effect of MMP-14 deficiency on reconstitution of bone marrow–derived cells by transplanting bone marrow–derived cells from Mmp14^+/+ or Mmp14^−/− mice (both CD45.2) into Ldlr^−/− mice (CD45.1; n=3 each group).²⁵ The reconstitution of peripheral blood monocytes was >92% in both groups (Figure 1).

• Download figure
• Download PowerPoint

Fluorescence-Activated Cell Sorter Analysis

We performed fluorescence-activated cell sorter analysis (FACS) as previously described.²⁵ Briefly, 20 μL diluted mouse blood (1:1 in FACS buffer [2% BSA, 0.1% sodium azide in PBS]) and 1 μL Fc block (eBioscience, San Diego, Calif) were incubated at room temperature (15 minutes) before fluorescently labeled antibodies were added (15 minutes). After incubation with FACS lysis buffer, the cells were washed twice with FACS buffer and analyzed by FACS. CD45.2-FITC and CD45.1-PE antibodies and corresponding isotype controls were purchased from eBioscience.

Tissue Preparation and Histological Assays

The aortic roots were prepared as described previously.^26,27 Briefly, mice were perfused at physiological pressure with normal saline via the left ventricle, and the hearts and aortas were removed en bloc. The aortic root was embedded in optical cutting temperature compound (Sakura, Torrance, Calif). To evaluate intimal lesion size, frozen sections of aortic root were incubated with oil red O (0.5% in glycerol). Immunohistochemical studies used rat anti-mouse monoclonal antibody to Mac3, a macrophage marker (1:1000, BD PharMingen, San Diego, Calif), and smooth muscle cell (SMC) α-actin staining with primary antibody FITC-conjugated α-actin mouse monoclonal (1:500, Sigma, St Louis, Mo), followed by anti-FITC biotin–conjugated secondary antibody (1:400, Sigma) and a rabbit anti–MMP-14 polyclonal antibody (1:500, Chemicon, Temecula, Calif). We analyzed fibrillar collagen content using picrosirius red staining of sections, which were viewed under polarized light. Quantitative analyses used Image-Pro Plus Software (Media Cybernetics, Bethesda, Md). Two blinded observers recorded the percentage of the total area with positive color for each section.

Reverse-Transcription Polymerase Chain Reaction

Total RNA was extracted from whole mouse aortas (pooled, n=3 per group) and reverse transcribed. Real-time reverse-transcription polymerase chain reaction (RT-PCR) used SYBR Green PCR Master Mix and MyiQ Single Color Detection System (BioRad, Hercules, Calif). Oligonucleotide primers used to recognize mouse mRNAs included the following: MMP-2, 5′-GCA-CCC-TTG-AAG-AAG-TAG-CTA-TG-3′ and 5′-GCA-GGA-GAC-AAG-TTC-TGG-AGA-TA-3′; MMP-8, 5′-CAA-CCT-ATT-TCT-CGT-GGC-TG-3′ and 5′-TGC-AGG-TCA-TAG-CCA-CTT-AG-3′; MMP-9, 5′-AAC-ACA-CAG-GGT-TTG-CCT-TC-3′ and 5′-CGT-CGT-GAT-CCC-CAC-TTA-CT-3′; MMP-12, 5′-TTT-CTT-CCA-TAT-GGC-CAA-GC-3′ and 5′-GGT-CAA-AGA-CAG-CTG-CAT-CA-3′; MMP-13, 5′-TCC-CTT-GAT-GCC-ATT-ACC-AGT-C-3′ and 5′-AAA-AAG-AGC-TCA-GCC-TCA-ACC-TG-3′; MMP-14, 5′-AGG-GTT-CCT-GGC-TCA-TGC-3′ and 5′-ACA-GCG-GCC-GCA-CTC-ACA-3′; cathepsin K, 5′-CCA-GTG-GGA-GCT-ATG-GAA-GA-3′ and 5′-AAG-TGG-TTC-ATG-GCC-AGT-TC-3′; α1 procollagen I, 5′-TCT-TTC-TCC-TCT-CTG-ACC-G-3′ and 5′-AAG-GTG-CTG-ATG-GTT-CTC-C-3′; and GAPDH, 5′-TGG-GTG-TGA-ACC-ATG-AGA-AG-3′ and 5′-GCT-AAG-CAG-TTG-GTG-GTG-C-3′.

Western Blotting and Gelatin Zymography

Whole aortas from Mmp14^−/−→Ldlr^−/− and Mmp14^+/+→Ldlr^−/− (n=3 each group) were harvested and snap-frozen in liquid nitrogen. After pulverization, samples were homogenized into radioimmunoprecipitation assay buffer (Boston Bioproducts, Boston, Mass) with EDTA-free protease inhibitor cocktail (Roche, Indianapolis, Ind) for 30 minutes at 4°C and centrifuged at 3000g for 30 minutes. For Western blotting, total protein (20 μg per well) was separated by standard SDS-PAGE and blotted to polyvinylidene difluoride membranes (Bio-Rad, Hercules, Calif) using a semidry blotting apparatus. Blots were blocked in 5% (wt/vol) defatted dry milk in PBS/0.1% Tween 20 (Sigma-Aldrich, St Louis, MO) overnight and incubated with the respective primary antibody (2 hours). The secondary peroxidase-conjugated antibody (1:10 000, Jackson Immunoresearch, West Grove, Pa) was added for another hour. Finally, immunoreactive proteins were visualized with the Western Lightning Chemiluminescence Reagent Plus (PerkinElmer Life Sciences, Boston, Mass). We used rabbit polyclonal antibody against MMP-14 (1:1000, Chemicon), goat polyclonal antibodies against MMP-13 (1:2000, Chemicon) and MMP-8 (1:100, R&D Systems, Minneapolis, Minn), and a rabbit polyclonal antibody against α-tubulin (1:500, Santa Cruz Biotechnology, Santa Cruz, Calif) as a loading control. For gelatin zymography, we used a previously described protocol.²⁸ Briefly, equal amounts of total protein were separated under nonreducing conditions by SDS-PAGE containing gelatin (1 mg/mL, BioRad). After washing with renaturation buffer (BioRad; 30 minutes), we incubated the gel in development buffer (24 hours) and then stained it with Coomassie brilliant blue 0.5% and destained in 25% methanol/10% acetic acid.

Macrophage Culture and In Vitro Collagenase Assay

Bone marrow–derived macrophages were harvested from the femurs and tibias of Mmp13^−/−, Mmp14^−/−, and wild-type mice. After incubation in red cell lysis buffer (ammonium chloride 0.155 mol/L in PBS), we centrifuged (20 minutes) the samples on Ficoll medium (LSM, ICN Biomedicals, Aurora, Ohio) and collected the monocytic cells. Macrophages were selected using medium with macrophage-colony stimulating factor (25 ng/mL, Cell Sciences, Canton, Mass) for 6 days. Determination of the collagenolytic capacity used cells maintained in serum-free medium overnight, unstimulated or stimulated with tumor necrosis factor-α (TNF-α; 10 ng/mL) and then incubated with fluorescein-labeled nondenatured collagen type-I (0.3 mg/mL, Calbiochem, La Jolla, Calif) for 48 hours in the presence or absence of 1,10-phenanthroline (0.1 mmol/L, Sigma), a broad metalloenzyme inhibitor. Digested collagen fragments were measured at 485-nm excitation and 530-nm emission in a fluorescent plate reader.

Statistical Analysis

Data are presented as mean±SD. Differences between groups were determined with the Mann-Whitney U test. Values of P<0.05 were considered significant.

The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.

Results

Characteristics of Mice

Body weight and plasma cholesterol did not differ between groups (Mmp14^+/+→Ldlr^−/− and Mmp14^−/−→Ldlr^−/−) consuming an atherogenic diet for 8 or 16 weeks. Total serum cholesterol level increased with the duration of the atherogenic diet and did not differ between groups (Table).

Compound Mutant Ldlr^−/− Mice With Mmp14^−/− Bone Marrow–Derived Cells Expressed Less MMP-14 Than Those Receiving Mmp14^+/+ Bone Marrow–Derived Cells

Our experimental approach supposes that bone marrow–derived macrophages furnish most of the MMP-14 in the plaque. In the mice constructed for this study, only the bone marrow–derived cells from Mmp14^−/− mice lack MMP-14. Many arterial SMCs and endothelial cells do not originate from bone marrow²⁹ but could produce MMP-14, especially in inflammatory environments such as atheromata.^30,31 Macrophages localized in the atherosclerotic intima of Ldlr^−/− mice reconstituted from Mmp14^−/− donors did not contain MMP-14, whereas in intimal lesions of Ldlr^−/− mice receiving Mmp14^+/+ bone marrow, macrophages showed strong immunostaining for MMP-14 (Figure 2A). Quantitative analysis of aortic extracts demonstrated that the aortic wall of Ldlr^−/− mice receiving Mmp14^−/− bone marrow contained significantly less MMP-14 mRNA (Figure 2B) and protein (Figure 2C) compared with mice reconstituted with Mmp14^+/+ bone marrow.

• Download figure
• Download PowerPoint

MMP-14 Deficiency in Bone Marrow–Derived Cells Does Not Influence Atherosclerotic Lesion Size or Cellular Composition

Aortic root lesion size was similar in Ldlr^−/− mice receiving Mmp14^+/+ bone marrow and Ldlr^−/− mice receiving Mmp14^−/− bone marrow (Figure 3A). Quantitative image analysis revealed similar intimal areas in Ldlr^−/− mice reconstituted with Mmp14^+/+ or Mmp14^−/− after 8 weeks (0.13±0.08 and 0.12±0.11 mm², respectively) or 16 weeks (0.33±0.10 and 0.25±0.12 mm², respectively) on atherogenic diet (Figure 3B). These lesions contained similar macrophage and SMC content at 16 weeks (Figure 3C). Quantitative image analysis confirmed a similar percentage of positive area for SMCs (α-actin⁺) and macrophages (Mac-3⁺) in plaques of mice receiving Mmp14^+/+ and Mmp14^−/− bone marrow cells at 8 and 16 weeks (2.16±2.21% versus 2.46±2.03% and 5.38±1.89% versus 4.89±2.14% at 8 and 16 weeks, respectively, for SMCs; 5.26±2.2% versus 4.08±2.13% and 7.12±3.39% versus 6.08±2.13% at 8 and 16 weeks, respectively, for macrophages) (Figure 3D). Our results agree with our previous study that found similar plaque burden and cell content in Mmp13^−/−/ApoE^−/− and Mmp13^+/+/ApoE^−/− mice.¹⁶

• Download figure
• Download PowerPoint

MMP-14 Deficiency in Bone Marrow–Derived Cells Increases Collagen Content in the Atherosclerotic Intima

Aortic atheromata of Ldlr^−/− mice receiving Mmp14^−/− bone marrow cells displayed greater accumulation of fibrillar collagen, as shown by picrosirius red staining under polarized light (Figure 4A). Quantitative analysis showed that MMP-14 deficiency in bone marrow–derived cells increases collagen content expressed as the percentage of intimal area (Figure 4B). As observed in infarcted hearts of MMP-9–deficient mice,³² targeted deletion of a single MMP may cause “compensatory” changes in expression of other MMPs or other enzymes involved in collagen turnover. After 16 weeks of atherogenic diet, RT-PCR analysis of mRNA expression in aortic extracts detected similar levels of all tested MMPs (MMP-2, -8, -9, -12, and -13) and a cysteine proteinase, cathepsin K, also implicated in arterial wall remodeling.³³ Nor did MMP-14 deficiency affect interstitial collagen gene expression because aortas of both experimental groups had similar procollagen Iα-mRNA levels (Figure 5).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

MMP-14 Deficiency Decreases Collagenase Activity of Bone Marrow–Derived Macrophages

Because macrophages furnish most MMPs in atheromata, further experiments compared the collagenase activity of bone marrow–derived macrophages from Mmp13^−/−, Mmp14^−/−, and wild-type mice using a fluorescein-labeled collagen type I as substrate to monitor collagen degradation (Figure 6). Bone marrow–derived macrophages from Mmp14^−/− mice had significantly lower collagenolytic activity than those from wild-type or Mmp13^−/− mice. Although treatment with the proinflammatory cytokine TNF-α increased collagenolysis by macrophages, collagen-degrading activity remained higher in macrophages from wild-type mice compared with Mmp14^−/− and Mmp13^−/− mice. After addition of 1,10-phenanthroline, a nonselective metalloenzyme inhibitor, degradation of collagen in the 3 different conditions decreased almost to the level obtained without stimulation, suggesting that MMPs account for most of the TNF-α–induced collagenolytic activity.

• Download figure
• Download PowerPoint

MMP-14 Deficiency Alters Activation of MMP-13 but Not of MMP-2 or MMP-8

We previously demonstrated a key role of MMP-13 in the regulation of plaque collagen content.¹⁶ Western blot analysis revealed that the level of latent MMP-13 (≈57 kDa) increased in aortic extracts of mice receiving Mmp14^−/− bone marrow cells, whereas levels of active MMP-13 (≈45 kDa) decreased considerably. The level of a truncated form of MMP-13 (≈20 kDa) was similar in both groups (Figure 7A). Gelatin zymography demonstrated similar levels of latent and active MMP-2 in both groups (pro-Mmp2, ≈72 kDa; active Mmp2, ≈60 kDa) (Figure 7B), indicating that MMP-14 influences MMP-13 activation either directly or indirectly during atherogenesis. In the atherosclerotic aorta, however, MMP-14 does not appear critical to MMP-2 activation under these conditions. Western blot analysis revealed similar levels of active MMP-8 between the groups (Figure 7C).

• Download figure
• Download PowerPoint

Discussion

This study demonstrates that MMP-14 influences the collagen level in mouse atherosclerotic plaque. Moreover, MMP-14–mediated interstitial collagenase activity in plaque depends substantially on bone marrow–derived cells. Indeed, bone marrow–derived inflammatory cells provide most of the MMP-14 within the arterial wall under atherogenic conditions.^30,34 Vascular SMCs also express considerable MMP-14 during mouse development.^35,36 However, the role of Mmp14 in these cells will have to await the development of conditionally inactivated Mmp14 mice, which are presently unavailable. In the meantime, successful transplantation of bone marrow—a critical step in this study—enabled analysis of the role of bone marrow–derived cells in experimental atherogenesis. An earlier study showed that MMP-9 participates importantly in the recruitment and maturation of bone marrow stem cells.³⁷ The bone marrow reconstitution experiments reported here indicate that trafficking of bone marrow cells to the atherosclerotic plaque does not require MMP-14 expression.

Deficiency of MMP-14 in macrophages does not affect the size or cellular content of plaques. Earlier work in Mmp14^−/− mice indicated that MMP-14 acts as a collagenase during osteogenesis, soft tissue remodeling,^21,38 and degradation of the basement membrane to initiate neoangiogenesis.²² Furthermore, MMP-14 has collagenolytic activity in vitro.^39,40 The present study establishes in vivo that MMP-14 influences collagen content of atherosclerotic plaques. A previous study compared the atherosclerotic lesion in the aorta between Ldlr^−/− mice and littermate Ldlr^−/− mice receiving total irradiation and bone marrow reconstitution.⁴¹ The collagen layer was significantly thicker in untreated Ldlr^−/− mice than in Ldlr^−/− mice receiving total irradiation and bone marrow reconstitution, suggesting that bone marrow reconstitution may have influenced SMC involvement in plaque fibrosis.

Interpretation of the present data requires careful consideration of several potential confounders. Indeed, unexpected decreases in myocardial collagen after myocardial infarction in Mmp9^−/− mice likely resulted from a compensatory increase in MMP-13 in infarcted tissue.³² Therefore, we measured the RNA level of the other interstitial collagenases MMP-8 and MMP-13 and the potentially collagenolytic cysteine proteinase cathepsin K. MMP-14 deficiency affected neither the level of the other enzymes tested nor the level of procollagen-I mRNA. Thus, collagen accumulation in the atherosclerotic plaques of mice receiving Mmp14^−/− bone marrow–derived cells did not appear to result from compensatory changes in other collagenases or in interstitial collagen gene expression. In vitro study of bone marrow–derived macrophages documented decreased collagen breakdown by Mmp14^−/− cells compared with wild-type and Mmp13^−/− cells under unstimulated conditions or after TNF-α stimulation (Figure 6).

As previously shown, membrane type 1 MMP (MMP-14) participates in the activation of the latent forms of MMP-2 (progelatinase A) and MMP-13.^42–44 To test whether MMP-14 acts in the plaque directly as a collagenase or also acts by processing pro–MMP-13 and/or MMP-2, we examined the activation of MMP-13 and MMP-2 in the atherosclerotic aortas of mice reconstituted with MMP-14–deficient or wild-type bone marrow. Interestingly, deficiency of MMP-14 decreased the levels of activated MMP-13 but did not affect MMP-2 activation in atherosclerotic lesions, suggesting a dual role of MMP-14 in collagenolysis and plaque stabilization. Notably, MMP-14 deficiency did not affect MMP mRNA levels, including MMP-13, as demonstrated by real-time RT-PCR. Although MMP-14 deletion decreased accumulation of a cleaved form of MMP-13 in mouse atheromata, lack of MMP-14 did not affect MMP-8 activation. This result indicates that the situation in atheromata in vivo may differ from results of an in vitro study on human tear fluid MMP-8 during wound healing after acute eye injury⁴⁵ that suggested MMP-14–dependent activation of MMP-8. Because our previous analysis indicated a role for Mmp13 activity in the atherosclerotic plaque, the present findings suggest that MMP-13 activation may mediate, at least in part, the effect of MMP-14 on collagen metabolism during atherogenesis. Unlike humans, mice lack MMP-1; hence, MMP-13 appears to subserve in mice the functions of MMP-1 in humans. These considerations illustrate that the pathophysiological principles demonstrated here may not apply directly to humans. Although several tissue and cell types require MMP-14 for activation of pro–MMP-2, our present data suggest that alternate mechanisms operate in atheromata.^21,24 In addition, a recent study described how MMP-14 had developmental effects that did not depend on its role in pro–MMP-2 activation during lung and submandibular gland maturation.⁴⁶ Taken together, these data suggest that effects of macrophage-derived MMP-14 other than pro–MMP-2 activation dominate in atherosclerosis. Because furin activates MMP-14 intracellularly in the trans-Golgi, MMP-14 can exert its proteolytic activity in the pericellular space as soon as it anchors in the cell membrane.⁴⁷ A recent study showed that the furin-like proconvertase PC5 also can activate MMP-14 in vascular SMCs.⁴⁸

Human coronary artery plaques that have caused fatal thrombosis typically have a thin fibrous cap, reduced levels of intact interstitial collagen, and abundant levels of MMP-13¹² and MMP-14¹⁷ and display biochemical signatures of collagenolysis in situ. The present study found that MMP-14 deficiency in bone marrow–derived cells (primarily macrophages) does not influence atherosclerotic lesion size or cellular composition but does, with time, substantially increase lesional content of fibrillar collagen. These results agree with our previous findings that collagenase resistance or MMP-13 deficiency promoted collagen accumulation, a key molecular determinant of plaque stability, but not atheroma burden.^15,16

The present study used cholesterol-fed Ldlr^−/− mice, an established mouse model of hypercholesterolemia, to induce accumulation of plaque macrophages, a major source of MMPs. Macrophage infiltration in hypercholesterolemia promotes MMP production in arteries.^1–4 Indeed, we reported early on that oxidatively modified low-density lipoprotein could promote MMP-14 expression in vascular cells in culture.³⁰ Our own work and studies of others have demonstrated that lipid lowering reduces MMP expression and yields collagen accumulation in atheromata.³

Taken together with studies using Mmp13^−/− and Col^R/R mice,^15,16 the present work further supports the involvement of collagenases from the MMP family in arterial collagen remodeling and illustrates a novel aspect of collagen metabolism in atherosclerosis. These results shed new mechanistic light on the molecular and cellular mechanisms that promote collagen degradation and thus may influence the biomechanical properties of plaques. The role of bone marrow–derived cells, principally macrophages, in regulating collagen accumulation demonstrated here underscores the role of inflammation in clinically critical aspects of plaque biology.

Footnotes