Selective inhibition of matrix metalloproteinase 13 (MMP-13) increases collagen content of established mouse atheromata

Abstract

Evidence has linked collagen loss with the onset of acute coronary events.

Objective

This study tested the hypothesis that selective MMP-13 collagenase inhibition increases collagen content in already established and nascent mouse atheromata.

Methods and Results

In vitro and in situ experiments documented the selectivity and efficacy of an orally available MMP-13 inhibitor (MMP13i-A). In vivo observations monitored macrophage accumulation and MMP-13 activity using molecular imaging. After 10 weeks of MMP13i-A treatment, apoE^-/- mice with evolving or established lesions exhibited reduced MMP-13 activity without affecting macrophage content, measured either by intravital microscopy or fluorescence reflectance imaging. Histological analysis indicated that MMP13-iA did not affect plaque size, or macrophage or smooth-muscle cell accumulation. Administration of MMP13i-A to mice with evolving or established atheromata substantially increased plaque interstitial collagen content in the intima and locally in the fibrous cap, compared to vehicle-treated controls. Analysis of collagen revealed thicker collagen fibers within the plaques of treated groups.

Conclusions

Pharmacological MMP-13 inhibition yields collagen accumulation in plaques (a feature associated in humans with resistance to rupture), even in established plaques. This study of considerable clinical relevance furnishes new mechanistic insight into regulation of the plaque's extracellular matrix, and validates molecular imaging for studying plaque biology.

Atherosclerosis and its thrombotic complications represent a major and still growing cause of death worldwide. Extensive studies have shown that most fatal coronary arterial thrombi result from physical disruption of atherosclerotic plaques. Atherosclerotic lesions that have ruptured and caused fatal acute myocardial infarction characteristically contain abundant macrophages underlying a thin and collagen-poor fibrous cap.¹ Extracellular matrix macromolecules, especially fibrillar interstitial collagens, confer tensile strength upon the plaque's fibrous cap.² We hypothesized that imbalance of collagen synthesis by smooth-muscle cells (SMCs) and its degradation by matrix-degrading enzymes, particularly matrix metalloproteinase (MMP) collagenases, regulates plaque collagen content.^{1, 3, 4} We previously have provided evidence in vivo supporting the contribution of collagenolysis to the control of collagen content in plaques. We reported the overexpression of several interstitial collagenases, and localized intermediates, in collagen breakdown in human atheromata using both biochemical and in situ morphological approaches.^5-8 Plaques in atherosclerosis-susceptible, apolipoprotein E–deficient (apoE^-/-) mice crossed with genetically altered mice that express collagenase-resistant collagen I (Col^R/R) contained more interstitial collagen than those in apoE^-/- control mice.⁸ Aortas of apoE^-/- Col^R/R mice showed increased tensile strength in a biomaterial assay, supporting a role for collagen content in determining the mechanical properties of atheromatous arteries. Among the MMP collagenases, MMP-1, MMP-8, and MMP-13 can cleave fibrillar collagen at neutral pH.^{7, 9, 10} In apoE^-/- mice, genetically-determined absence of MMP-13 — a major interstitial collagenase in this species — yielded increased plaque collagen content with thicker, more aligned, and more organized collagen fibers than in MMP-13 wild-type (WT) controls.⁵ Compensatory changes in protease expression in these mice with congenital collagenase lack, and the focus on the initial phase of lesion development limits the interpretation and generalizability of these results. These observations also raised the translational hypothesis that inhibition of MMP-13 reinforces the collagen content of the fibrous cap of established or evolving atherosclerotic plaques, in addition to nascent lesions. As non-specific MMP inhibitors can produce unwanted effects that have consistently limited their utility clinically and as mechanistic probes, we administered a highly selective MMP-13 inhibitor (MMP13i-A) orally to test this hypothesis.^11-13

The development of molecular imaging strategies to visualize plaque biology and progression represents a major opportunity for cardiovascular medicine, especially because many acute thrombotic events remain unpredictable and occur despite current optimum therapy. We and others have reported that imaging probes and modalities can visualize protease activity and macrophage content ex vivo and in vivo. ^14-19 To monitor plaque biology and actual MMP-13 inhibition in situ, this study used an activatable probe that is preferentially cleaved by MMP-13 and macrophage avid fluorescent iron nanoparticles, combined with intravital confocal microscopy and fluorescence reflectance imaging.

Materials and Methods

Detailed Methods are available online at http://atvb.ahajournals.org.

Animal Preparation

All experiments conformed to a protocol approved by the Standing Committee on Animals of Harvard Medical School. We studied the impact of MMP-13 inhibition on atherosclerosis-susceptible apoE^-/- mice with congenic c57bl/6 background. Male mice 8–10 weeks of age (n=12) consumed an atherogenic diet (semi-purified chow containing 1.25% cholesterol and 0% cholate, Research Diets, New Brunswick, NJ) for 10 weeks, together with oral administration of indicated doses of the MMP-13 inhibitor MMP13i-A (Amgen, Thousand Oaks, CA). To test the effects of MMP-13 inhibition on already established atheromata, mice were fed an atherogenic diet for 10 weeks before starting the 10-week treatment (n=12) (Supplemental Fig. I). For both studies, MMP13i-A and corresponding vehicle were administered by gavage twice a day, spaced by 10–12 hours. Controls were age-matched and treated with vehicle (methylcellulose 0.5%, Convergent Bioscience, Toronto, Canada). Age-matched MMP13^-/- apoE^-/- deficient mice (c57bl/6) consumed a high-fat diet for 10 weeks and were used to assess imaging agent specificity (n=3).⁵ All mice were maintained in animal facilities at Harvard Medical School. Animal care and procedures were reviewed and approved by the Institutional Animal Care and Use Committees and performed in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care and the National Institutes of Health.

Human Tissues

Endarterectomy specimens of human carotid plaques (n=5) were obtained by protocols approved by the Human Investigation Review Committee at Brigham and Women's Hospital. Surgical tissue samples were embedded in OCT compound and stored at -80°C until use. Serial cryo-sections were cut, air-dried on to microscope slides, and used in parallel for in situ zymography (ISZ) and immunohistochemistry.

Molecular Imaging

Molecular imaging agents

We used fluorescent nanoparticles based on an iron oxide core (Ex753/Em773) for the detection of macrophage accumulation. MMPsense (Ex680/Em700) fluorescence reflected MMP-13 activity.

Intravital confocal microscopy (IVM)

The carotid artery was exposed at the level of the external and internal carotid bifurcation. Anesthetized mice were then placed under the confocal microscope to detect the fluorescence emitted by the probes in plaques (Olympus FV1000, Tokyo, Japan). Image collection of different channels was done sequentially to avoid crosstalk between channels, as we described previously.²⁰

Ex vivo fluorescence reflectance imaging (FRI)

After euthanasia, aortas were imaged and fluorescence was quantified using a FRI system equipped with multichannel filter sets (Carestream, Rochester, NY). The sum of the fluorescence levels in the aortic root and arch region was subtracted to the background level, calculated in a similar area size with no aorta exposed.

Cell Culture

Human primary monocyte-derived macrophages

Monocytes were cultured and differentiated into macrophages after 10 days in RPMI containing 5% human serum and 1% penicillin–streptomycin. Incubation with recombinant tumor necrosis factor-α (TNFα) and interleukin-1β (IL1β) elicited macrophage activation.

MMP Activity Assay

MMP-13 activity was measured using a fluorescent MMP-13 substrate, as previously described.²¹ Selectivity assays were performed using catalytic domains of human MMPs and omniMMP fluorescent peptide (Enzo Life Sciences) as substrate.

In situ Zymography (ISZ)

To estimate MMP collagenase activity in situ, we incubated 6–8 μm sections with DQ-Collagen (Invitrogen), in the presence of MMP13i-A (10 μM), the broad-spectrum MMP-inhibitor Ilomastat (10 μM), or ortho-phenanthroline (0.5 mM), for 48 hours.

Histological Assays

For histological evaluations of the aortic root, we used the method of Paigen et al.²² Plaque size and composition was characterized with anti-α-smooth-muscle actin (Sigma-Aldrich) and anti-Mac3 (BD Pharmingen) antibodies. Quantification was performed on ImagePro Plus 5.1 (Media Cybernetics).^{8, 23-25}

Collagen Fiber Morphology and Quantification

Detection of interstitial collagen used picrosirius red staining with linear polarized light, as described previously.^{7, 23, 26, 27} As fiber thickness increases, the color shifts from green to red. To discriminate the green and red birefringent fibers, green and red optic filters (HQ535/50m, D605/55m, Chroma) were disposed under polarized light. Images were analyzed using image analysis software (NIS-Elements, Nikon). The relative amount of each fiber color was expressed as a percentage of the total amount of collagen in the region.

Real-Time qRT-PCR

Total RNA was extracted from abdominal aortas (n=5 in each group). Real-time qRT-PCR used SYBR green PCR Master Mix, QuantiTect Oligonucleotides for mouse MMP-8, MMP-12, MMP-13, MMP-14, cathepsin K, and GAPDH (Qiagen), and MyiQ Real-Time PCR Detection System (Biorad, Hercules, CA).

Statistical Analysis

Differences between the treated and untreated groups were determined with the Mann-Whitney U test.

Results

Highly selective MMP-13 inhibition by MMP13i-A

Pre-clinical trials with broad MMP inhibitors have persistently failed because of toxicity or limited efficacy.^11-13 We therefore conducted this series of experiments with a novel putatively selective MMP-13 inhibitor (MMP13i-A) (Supplemental Fig. II). To test the selectivity of the nonhydroxamic acid-based MMP13i-A compound against MMP-13, various MMP catalytic domains were incubated with a broad MMP fluorogenic substrate in the presence of MMP13i-A at pH 7.5, in the optimum range for MMPs. MMP13i-A inhibited with nanomolar potency (50% inhibition concentration [IC₅₀]) recombinant active MMP-13, but not MMP-1, MMP-2, MMP-7, MMP-9, MMP-12, or MMP-14 (Fig. 1A). MMP13i-A reduced MMP-13 activity by 80% in a concentration-dependent manner using a specific MMP-13 cleavable substrate (Fig. 1B). Moreover, we tested whether MMP13i-A could also block MMP-13 released by activated macrophages, the most abundant producers of this MMP in atherosclerotic plaques. Incubation of macrophage supernatants with MMP13i-A reduced MMP-13 activity by 53% (human, p<0.001) and 76% (mouse, p<0.001) (Fig. 1C and 1D). In vitro incubation of this substrate with various MMP catalytic domains affirmed the selectivity of this substrate for MMP-13 (Fig. 1E).

Figure 1.

Effect of MMP13i-A inhibitor on recombinant MMP activity. 50 percent inhibition concentration (IC₅₀) was based on evaluation of six different concentrations of MMP13i-A (0–20,000nM) (A). Inhibition of MMP-13 activity of recombinant active form of MMP-13 (B), supernatant of primary human (C) and murine macrophages (D). Selective cleavage of MMP-13 substrate by various recombinant catalytic domains of human MMPs (E). Means ± SEM are represented.

MMP-13 inhibition decreased collagenolysis in plaques

To test further the efficiency of MMP13i-A and the overall impact of MMP-13 inhibition on collagenolysis in plaques, we performed in situ zymography for collagenolysis using a fluorescent substrate. MMP13i-A blocked collagenolysis in murine and human atheromata using a collagen I-based fluorescent substrate assay (Fig. 2A and 2B and Supplemental Fig. III). The non-specific metalloenzyme inhibitor ortho-phenanthroline and the general MMP inhibitor Ilomastat served as controls. MMP13i-A blocked this activity to a greater extent in mouse than in human lesions. As adult mice do not express MMP-1, these data suggest that MMP-13 plays a prominent role in type I collagen degradation in this species, and that this process depends less upon MMP-8 and MMP-14 in mice.

Figure 2.

In situ zymography for collagenolytic activity in murine (B) and human (C) atheromata. Macrophage staining was performed on adjacent sections. Ortho-phenanthroline and the broad MMP inhibitor Ilomastat were used as controls. All images represent n ≥ 3 subjects/experiments.

Fluorescent nanoparticles and MMP-13 activatable probes allowed imaging of plaque burden and MMP-13 activity in atherosclerotic mice

Molecular imaging in vivo has considerable potential in diagnostics and in directing therapy, but it also may help to probe pathophysiological mechanisms. To image atherosclerotic plaques, we used near-infrared fluorescent nanoparticles based on an iron oxide core. We tested selective uptake by macrophages — a prominent cellular component of atheromatous plaques — by incubating nanoparticles for 4 hours with either murine macrophages or primary murine SMCs (Fig. 3A). Macrophages internalized nanoparticles in a dose-dependent manner, but we observed no uptake by SMCs. Fluorescent labeled LDL (DiO-Ac-LDL) served as a positive control for selective macrophage uptake.

Figure 3.

Fluorescent nanoparticle uptake by murine macrophages, but not by primary murine smooth-muscle cells (SMC) in vitro (A). Fluorescent LDL was used as a control for macrophage phagocytic activity. Representative ex vivo fluorescent reflectance imaging (FRI) of MMP13^-/-apoE^-/- mice and MMP13^+/+ apoE^-/- mice injected with MMP-13 probe (B). Images and quantification of ex vivo FRI on isolated aortas from WT mice or atherosclerotic apoE^-/- mice coinjected with MMP-13 probe (left) and fluorescent nanoparticles (right) (C and D). Means ± SEM are represented. (*p<0.05, **p<0.01, ***p<0.001).

We employed this imaging strategy primarily to follow and verify in vivo the ability of MMP13i-A to inhibit MMP-13 activity, using a quenched fluorescent activatable probe that increases in fluorescence by several logs when cleaved by active MMP-13. We tested the selectivity of this MMP-13 probe by comparing ex vivo FRI of aortas from apoE^-/- and apoE^-/- MMP-13^-/-atherosclerotic mice. After 10 weeks on an atherogenic diet, mice received an MMP-13 probe intravenously 48 hours before sacrifice and tissue collection. While FRI detected the fluorescent signal in the aortic root or arch and in the abdominal portion of aortas of apoE^-/- mice, the signal intensity decreased substantially in apoE^-/- mice crossed with MMP-13–deficient mice, indicating the selectivity of the proteinase probe for this enzyme (Fig. 3B) ⁵. Tissue from apoE^-/- mice that did not receive the substrate showed no fluorescence. Moreover, to estimate how much MMP-13 activity increases and macrophages accumulate in atherosclerotic plaques versus healthy tissue, we co-administered the MMP-13 probe and the macrophage-targeted nanoparticles to apoE^-/- mice on an atherogenic diet and to WT mice on regular chow. Moreover, administration of the MMP-13 probe and the macrophage-targeted nanoparticles to apoE^-/- (atherogenic diet) mice and WT (chow) mice revealed substantial increases of both MMP-13 activity and macrophage accumulation in atherosclerotic plaques versus healthy tissue (Fig. 3C and 3D). Quantification of FRI showed a 407% increase in MMP-13 signal and a 344% increase in macrophage phagocytic activity in the arteries of the hypercholesterolemic mice (p<0.05).

MMP-13 activity colocalized with macrophage burden in vivo

Various cell types can produce MMP-13, but macrophages likely are the major source of MMP-13 in human atherosclerotic lesions.⁷ To test this hypothesis, we coinjected apoE^-/- mice consuming an atherogenic diet for 10 weeks with the MMP-13 probe and phagocytosable nanoparticles. After 48 hours, anesthetized mice underwent in vivo confocal microscopy of carotid arteries. Sequential fluorescence imaging established that macrophages and MMP-13 activity colocalized in the carotid plaques. As expected, MMP-13–dependent fluorescence appeared more diffuse than the macrophage signal in vivo because MMP-13 is a secreted enzyme. Moreover, both signals colocalized with plaques visible under white light (Supplemental Fig. IV-A). To characterize better the spatial correlation between MMP-13 activity and macrophages in the plaque, we collected aortic roots/arches after in vivo confocal microscopy to detect fluorescence precisely on cryo-sections of plaques. Fluorescent nanoparticles colocalized with a specific macrophage marker (mac3), indicating their selective uptake by this cell type (Supplemental Fig. IV-B). In addition, MMP-13 activity colocalized tightly with macrophages, further suggesting that this cell type contributes most of the MMP-13 activity in mouse atheromata.²⁸

MMP13i-A treatment experimental design

We used in vivo molecular imaging with the MMP-13 activatable probe to determine the sufficient dosage of MMP13i-A to block efficiently the fluorescence reflecting MMP-13 activity in plaques. A preliminary dose-response study allowed us to select 40 mg/kg/day as the minimal dosage needed to reduce of MMP-13 activity in vivo (Supplemental Fig. IX). Pharmacokinetic analysis of the compound in rodents estimates a peak circulating concentration of ~20μM that effectively inhibits MMP-13 (79%), while only modestly affecting the activity of other MMPs (<IC₅₀) in vitro (Fig. 1, Supplemental Figs. VIII and XI). To assess the impact of MMP-13 inhibition on plaque biology, age-matched atherosclerotic susceptible apoE^-/- mice (males) were treated for 10 weeks with MMP13i-A while consuming an atherogenic diet. We also administered the MMP-13 inhibitor for 10 weeks to another group of apoE^-/- male mice that had already consumed the atherogenic diet for 10 weeks, to assess the effect of MMP13i-A on established lesions — a more clinically relevant situation (Supplemental Fig. I). In neither study did MMP13i-A treatment at 40 mg/kg/day induce visible toxicity or alter body weight (Supplemental Table A). Previous preclinical studies with MMP inhibitors often failed because of painful musculoskeletal syndrome (MSS) toxicity ^11-13, which associated with an accumulation of collagen in and around joints. MMP13i-A–treated mice did not exhibit any apparent joint swelling, and all mice appeared normal without apparent pain or distress.

MMP13i-A treatment decreased MMP-13 activity but did not alter macrophage accumulation

After 10 weeks, MMP-13 inhibitor–treated mice had markedly less MMP-13 activity in carotid plaques than did control mice, as determined by intravital confocal microscopy (Fig. 4A). In contrast, the treatment did not modulate macrophage accumulation in plaques. Ex vivo FRI of the aortic root/arch indicated a 35.6% (p=0.0048) decrease in MMP-13 activity in the aortic arch of treated mice compared with control mice. In contrast, macrophage accumulation did not change significantly (Fig. 4B and 4C). These results demonstrate that oral administration of the MMP13i-A compound inhibits MMP-13 activity in atherosclerotic plaques in vivo without substantial change in the cellular composition of lesions.

Figure 4.

IVM of the left carotid artery from MMP13i-A–treated and untreated apoE^-/- mice, coinjected with MMP-13 activatable probe and nanoparticles for macrophage detection (A). Images are representative of n=6 animals per group. After IVM, dissected aortas were subjected to ex vivo FRI (B). The upper panel displays aortas in white light. The lower panel shows fluorescence only of the aortas and aortic arch region. Means ± SEM are represented. (*p<0.05, **p<0.01, ***p<0.001).

MMP-13 inhibition increased collagen content in evolving and established plaques

MMP-13 inhibition by MMP13i-A led to a significant increase in collagen accumulation in atherosclerotic lesions after 10 weeks of treatment. MMP13i-A–treated mice had higher levels of lesional fibrillar collagen stained by picrosirius red and quantified by polarized light microscopy in both evolving and established plaques, as compared to vehicle-treated mice (8.1±1.1% vs. 13.9±1.7%, p=0.005; 22.2±1.4% vs. 30.7±1.9%, p=0.003, Fig. 5A and 5B). Consistent with the imaging results, the content of macrophages did not differ between treated mice and control mice (Fig. 5B). MMP13i-A treatment did not alter plaque size or SMC accumulation, the main source of interstitial collagens in arteries. Nor did MMP-13 inhibition in established lesions affect apoptosis or calcification (Supplemental Figs. V and VI).

Figure 5.

Representative picrosirius red staining viewed under white light and linearly polarized light to show fibrillar collagen in the aortic intima of apoE^-/- mice with developing or established atheromata, treated with vehicle or MMP13i-A (n=10 for each group) (A). Quantification of collagen content, plaque size, macrophage, and smooth-muscle cell content in neointima of plaques (B). Bars represent mean ± SEM. Qualitative analysis of collagen content using red and green filtered polarized light (C). *p reflects statistical significance between MMP13i-A–treated and respective untreated groups. (*p<0.05, **p<0.01, ***p<0.001).

Qualitative analysis revealed an accumulation of red collagen fibers in plaques under polarized and red or green filtered light, indicating thicker, larger collagen fibrils (Fig. 5C).^{5, 29} Because collagen accumulation at the base of atheroma might not improve plaque stability, we tested whether collagen content increased more particularly in the critical fibrous cap area. Fibrous caps contained 26% more collagen in MMP13i-A–treated mice with established lesions versus respective controls (Fig. 6A and 6B). In addition to the increase in collagen in fibrous caps, inhibition of MMP-13 also yielded significantly larger and thicker fibrous caps — a feature associated with human plaques that have not ruptured (Fig. 6C and 6D). Thus the plaques of MMP13i-A–treated mice exhibit characteristics associated with “stable” human plaques. MMP13i-A treatment did not affect the expression of mRNAs that encode other known collagenolytic enzymes, such as MMP-8, MMP-14, MMP-12, or cathepsin K in lesions, indicating that increased plaque collagen with MMP-13 blockade in vivo arose without compensatory alterations in these collagenases in vivo (Supplemental Fig. VII).

Figure 6.

Representative image and quantification of picrosirius red staining of fibrous caps of established lesions from mice treated with vehicle or MMP13i-A (n=9 for each group) (A and B). Quantification of the fibrous cap area (%) and minimal thickness (C and D). Bars represent mean ± SEM. *p reflects statistical significance between MMP13i-A–treated and respective untreated groups. (*p<0.05, **p<0.01, ***p<0.001).

Discussion

The thrombotic complications of atherosclerosis remain major causes of death worldwide, highlighting the need to unravel the underlying pathophysiology and develop new therapeutic approaches that can address the residual burden of events that occur in at-risk patients despite the current standard of care. Fracture of the fibrous caps of plaques causes most fatal acute myocardial infarctions. Human atheromata that have ruptured and caused such events characteristically have a thin, collagen-poor fibrous cap and a macrophage-rich lipid core, and bear other hallmarks of local inflammatory activation.^{30, 31} We and others have furnished evidence supporting active regulation of collagen breakdown in plaque in humans and mice using in vitro and in situ analyses, and in experiments using genetically altered mice.⁷ Notably, our prior mouse experiments have established increased collagen content of experimental atheromata in animals with collagen substrate mutated to resist breakdown by MMP collagenases, and in those with germ-line inactivation of MMP-13, a major interstitial collagenase in mice.^{5, 8}

These previous experiments proved that collagenolysis participates in the regulation of the collagen content of plaques, but they also raised several new important issues. First, unmeasured compensatory changes in the substrate or enzymes in animals with congenital defects might unwittingly confound their interpretation. Second, germline modification of a collagenase does not determine whether the enzyme participates in early or later stages of lesion development. Third, and highly clinically relevant, any efforts to develop or test strategies to inhibit collagenolysis therapeutically require demonstration that intervention on established lesion initiation can influence collagen content and plaque structure. The present study aimed to address these key outstanding questions that arose from our prior work using novel tools: an orally active, selective MMP-13 inhibitor, and molecular imaging to ascertain MMP-13 inhibition in vivo at the doses tested.

After validating the efficacy and selectivity of the MMP13i-A compound for MMP-13, we co-localized MMP-13 activity with macrophages in plaques, affirming in vivo the important roles of these inflammatory cells as sources of this interstitial collagenase, and the relevance of therapeutic strategies that target macrophages. By exploiting the fluorescent properties of MMP and macrophage imaging probes, we demonstrated directly through in vivo and ex vivo imaging, the inhibition of MMP-13 activity under the conditions of the experiments by MMP13i-A without affecting macrophage content within lesions, as attested by immunohistochemical analysis. Quantitative and qualitative analysis of the plaque collagen demonstrated that oral administration of a selective MMP-13 inhibitor significantly increased the overall content of collagen and yielded thicker fibers, to a similar extent as observed previously in early lesion formation in mice with congenital loss of MMP-13 due to germline manipulation.⁵ As MMP13i-A administration did not significantly increase the content of either SMCs or colIα1 transcript levels (data not shown) in lesions, the collagen accumulation in plaque likely results from reduced degradation (MMP-13 being the prominent collagenase in mouse plaques), rather than from an increase in collagen production. These alterations in the extracellular matrix of a plaque's fibrous cap would likely lessen the chance of rupture in human lesions.³²

While adding collagen to the fibrous cap of atherosclerotic lesions should render them more resistant to rupture by increasing thickness or strength, changing plaque matrix composition might have multiple effects, and could even render lesions more fragile. Adding organized fibrillar collagen in the fibrous cap would add tensile strength to that region, but adding collagen nearby — but not within the actual rupture location — could increase stress in the rupture region. An analogous “double-edged sword” may apply to plaque calcifications. While increased calcium may strengthen plaques, very local microcalcifications may prove “destabilizing”. Increasing fibrillar collagen only at the base of the atheroma, near the arterial media, without strengthening the fibrous cap, might also not improve overall stability and could increase local shear stresses. Indeed, we have shown that genetically driven increases in adventitial collagen enhance aneurysm formation in atherosclerotic aortas in mice. We further showed with formal biomechanical testing that these aortas can exhibit increased susceptibility to fracture.³³ It therefore was particularly important that mice receiving the MMP-13 inhibitor accumulated collagen in the fibrous caps. MMP13i-A administration also significantly increased fibrous cap area and thickness in plaques from MMP13i-A–treated mice compared to those from vehicle-treated control mice. These observations further indicate that plaques from mice treated with a highly selective MMP-13 inhibitor show reinforcement of the characteristics of human plaques thought to confer resistance to rupture.

In mice, the interstitial collagenases MMP-8 and MMP-13 can cleave helical collagen fibers at the neutral pH that prevails in the extracellular space. The present results affirm a major role for MMP-13 in collagen degradation in mouse atheromata. Using an in situ assay for collagenolysis, selective inhibition of MMP-13 substantially limited collagenolysis in plaques to an extent similar to that produced by broad-spectrum MMP inhibitors. This result also indicates that in mice, MMP-8 contributes less to plaque collagenolysis than does MMP-13.

Several studies in humans and animals have associated high local and systemic levels of MMP-2 and/or MMP-9 with atherothrombotic events,^34-37 suggesting a role in plaque destabilization. MMP-2 and MMP-9 are important in matrix remodeling, particularly through the further degradation of fragmented collagen molecules following primary cleavage by MMP collagenases that possess the rare ability to cleave intact interstitial fibrillar collagen. Our results presented here therefore imply that blockade of MMP-13 collagenase, and consecutive blockade or decrease in primary cleavage of interstitial fibrillar collagen, would prevent or delay its further degradation by gelatinases — even at high levels — as found in plaques considered prone to rupture.

Major differences between mouse and human atherogenesis limit the ready extrapolation of this and other studies. Mature mice do not express MMP-1, considered a major interstitial collagenase in humans, which we and others have previously localized in human plaques. Moreover, mouse atheromata seldom rupture and cause thrombotic complications, except under extreme conditions or after major manipulations. Thus, these experiments cannot prove that MMP-13 inhibition “stabilizes” plaques or prevents thrombotic complications, but they provide novel mechanistic insight into the regulation of the lesional extracellular matrix.

Broad-spectrum MMP inhibitors have failed consistently in clinical trials due to lack of efficacy (e.g., in metastatic cancer) and to deleterious side effects (most prominently, musculoskeletal toxicity).^{11, 13, 38-41} Some MMP-13 inhibitors appear to limit such toxicity and show beneficial clinical outcomes in arthritic diseases which may also involve MMP-13.^42-44 In our study, the highly selective MMP13i-A compound did not induce any apparent musculoskeletal toxicity at 40 mg/kg/day, an issue that warrants further in-depth study. As humans express three MMP interstitial collagenases (MMP-1, MMP-8, and MMP-13), selective MMP-13 inhibition might obviate the undesired effects encountered in the clinic with broad-spectrum MMP inhibitors that have impeded their use in humans. Our preclinical results obtained in mice or human tissues not only furnish mechanistic insight into plaque biology, but also provide proof of concept that merits consideration for further clinical development.

Molecular imaging for atherosclerosis promises to offer new tools for testing pathophysiologic hypotheses in humans, for choosing effective doses of novel therapeutics to guide trials with clinical endpoints, and for obtaining early signals of in vivo efficacy in altering plaque biology. This study used molecular imaging in vivo and ex vivo to colocalize macrophages with MMP-13 activities in human and murine atheromata. In mice, intravital confocal microscopy imaged macrophages and MMP-13 in plaques, and helped establish the proper dosage to achieve MMP-13 blockade in vivo (data not shown) and to monitor and validate MMP-13 inhibition during MMP13i-A treatment. These experimental findings illustrate the potential utility of imaging inflammation and proteinase activity in dose ranging in the design and conduct of clinical trials. The potential combination of imaging and targeting technologies with therapeutic compounds (e.g., MMP-13 inhibitor) also raises the challenging prospect of “theranostic” strategies that could associate selective targeting and imaging-based assessment of biochemical efficacy in vivo.

In conclusion, this study advances the mechanistic conjecture that collagenases participate decisively in plaque biology in several important ways, and points a path toward clinical translation of this concept. The results establish that interstitial collagenase inhibition can alter the collagen content of evolving and established atheromata, a key unresolved issue in plaque biology and in consideration of this strategy for therapeutic intervention. We further demonstrate that a highly selective, orally administered interstitial collagenase inhibitor might have efficacy in this regard, avoiding the adverse effects encountered with broad-spectrum MMP inhibitors in the clinic. Finally, the results of this study illustrate how molecular imaging approaches might aid clinical development of MMP inhibitors by permitting dose ranging and providing biochemical evidence of efficacy in situ, as a prelude to trials of clinical efficacy.