Smooth muscle cell-specific matrix metalloproteinase 3 deletion reduces osteogenic transformation and medial artery calcification

Abstract

Aims

Vascular calcification is highly prevalent in atherosclerosis, diabetes, and chronic kidney disease. It is associated with increased morbidity and mortality in patients with cardiovascular disease. Matrix metalloproteinase 3 (MMP-3), also known as stromelysin-1, is part of the large matrix metalloproteinase family. It can degrade extracellular matrix components of the arterial wall including elastin, which plays a central role in medial calcification. In this study, we sought to determine the role of MMP-3 in medial calcification.

Methods and results

We found that MMP-3 was increased in rodent models of medial calcification as well as in vascular smooth muscle cells (SMCs) cultured in a phosphate calcification medium. It was also highly expressed in calcified tibial arteries in patients with peripheral arterial disease (PAD). Knockdown and inhibition of MMP-3 suppressed phosphate-induced SMC osteogenic transformation and calcification, whereas the addition of a recombinant MMP-3 protein facilitated SMC calcification. In an ex vivo organ culture model and a rodent model of medial calcification induced by vitamin D₃, we found that MMP-3 deficiency significantly suppressed medial calcification in the aorta. We further found that medial calcification and osteogenic transformation were significantly reduced in SMC-specific MMP-3-deficient mice, suggesting that MMP-3 in SMCs is an important factor in this process.

Conclusion

These findings suggest that MMP-3 expression in vascular SMCs is an important regulator of medial calcification and that targeting MMP-3 could provide a therapeutic strategy to reduce it and address its consequences in patients with PAD.

Graphical Abstract

Graphical Abstract.

Time of primary review: 33 days

See the editorial comment for this article ‘Matrix metalloproteinase-3 joins a growing list of proteases that regulate vascular calcification’, by C.M. Giachelli et al., https://doi.org/10.1093/cvr/cvae064.

1. Introduction

Vascular calcification is defined as the deposition of calcium hydroxyapatite crystals in the vessel wall. It is prevalent in patients with atherosclerosis, diabetes, and chronic kidney disease (CKD). In those with peripheral artery disease (PAD), it is associated with increased cardiovascular morbidity and mortality.^1–4 Arterial calcification has two major forms including one predominating in the intima, and one in the media. In the coronary and carotid arteries, intimal artery calcification is the major form and is found in the atherosclerotic plaque. It is often considered to be a maker of atherosclerotic plaque burden and is associated with atherosclerotic thickening and hardening of the artery.⁵ By contrast, Mönckeberg’s sclerosis or medial artery calcification is of primary importance in patients with diabetes and CKD, and it is frequently observed along with degraded elastin on histological evaluation.³ Cardiovascular disease is the most common cause of death in patients with diabetes and CKD, and it has been demonstrated that medial calcification strongly predicts cardiovascular risk, particularly in patients with end-stage renal disease.⁶ In lower extremity arteries, medial calcification can decrease arterial compliance and is thought to affect limb perfusion with a subsequent increased risk of amputation in patients with chronic limb-threatening ischaemia.⁷

Calcification was previously thought to be a passive phenomenon but is now known to be a highly and tightly regulated process with features similar to those of developing bone.^8,9 It is widely accepted that vascular smooth muscle cells (SMCs) play a pivotal role in calcification. In response to phosphate, these cells can change phenotype from a contractile to osteoblast-like and gain the ability to release matrix vesicles, ultimately leading to calcium deposition.¹⁰ Both intimal and medial calcification are thought to share basic mechanisms that involve the transformation of vascular SMCs into more bone-like cells.¹¹ Several mechanisms have been proposed to understand the development of calcification in arteries. Genetic data have shown that blood vessels normally express a number of inhibitors of mineralization, including pyrophosphate and matrix Gla protein, and lack of these molecules can cause spontaneous arterial calcification.¹² Moreover, calcification can be promoted by a number of bone proteins such as Runt-related transcription factor 2 (RUNX2), bone morphogenetic proteins (BMPs), and alkaline phosphatase (ALP).⁹ A mechanistic link between increased bone resorption and medial artery calcification in uraemia has been reported. Bone turnover leads to the release of minerals from bone tissues and may serve to transfer calcium microparticles to the blood. This has been suggested as one explanation for the known association between vascular calcification and osteoporosis in women after menopause.¹³ Recent studies have shown that cell necrosis and apoptosis may contribute to arterial calcification, especially in atherosclerotic intima calcification.¹⁴ Emerging evidence also suggests that calcium and phosphate may directly affect vascular cell mineralization by promoting nucleation and crystal growth of apatite.¹⁰ Additionally, more recent studies have demonstrated that microRNAs,¹⁵ the inflammasome,¹⁶ and autophagy¹⁷ may play synergistic roles in the development of vascular calcification.

Matrix metalloproteinases (MMPs), a class of zinc-dependent endopeptidases, are a superfamily that includes 28 enzymes. Accumulating evidence shows that MMPs not only function to degrade all types of extracellular matrix proteins such as elastin and collagens but also act as regulators of extracellular signalling networks.¹⁸ It has been demonstrated that MMPs are implicated in many pathophysiological processes such as tissue remodelling, wound healing, and angiogenesis.^18–20 Significant findings from several labs including ours have suggested that MMPs play a critical role in vascular calcification.^21,22 We have previously shown that the MMP pan inhibitors doxycycline and GM6001 could significantly suppress arterial medial calcification in rodent models, suggesting that they are important regulators of the process.²¹ Recently, using a PCR array for all MMPs, we found that MMP-3, also called stromelysin-1, was the most strongly induced of all MMPs in a widely used rodent model of arterial calcification.

In this study, we show that MMP-3 is highly induced in multiple calcifying conditions. Inactivation of MMP-3 suppresses phosphate-induced SMC osteogenic transformation and calcification. Global and SMC-specific deficiency of MMP-3 can reduce arterial medial calcification. These data suggest that MMP-3 is an important regulator of medial calcification and targeting MMP-3 could be a new strategy to decrease it in our PAD patient population.

2. Methods

2.1. Animals

All animal care and procedures were conducted in accordance with animal protocols (IACUC 2022–20279) that were approved by the University Animal Care and Use Committee at Yale University and Beth Israel Deaconess Medical Center. Animals for the experiments were anaesthetized with 2.5% isoflurane via a vaporizer and then perfused with saline. Animals that were not used for experiments and the retired breeding pairs underwent euthanasia by CO₂, followed by cervical dislocation. The global MMP-3 knockout mice were generously provided by Lynn Matrisian at Vanderbilt University. This mouse line has been backcrossed with C57BL/6 mice for at least 9 generations. To generate SMC-specific MMP3-deficient mice. CRISPR/Cas9 technology was used to generate MMP-3 floxed mice (Taconic Biosciences, Germantown, NY, USA). MMP-3 floxed mice were then bred with SMC-specific Cre mice smooth muscle myosin heavy chain (SMMHC)-CreER^T2 (JAX, B6.FVB-Tg (Myh11-cre/ERT2)1Soff/J, stock No, 019079) to produce SMMHC^CreERT2/MMP3 flox^+/+ mice. All animals were fed with a normal rodent diet ad libitum. To induce SMC-specific MMP-3 conditional knockout, SMMHC^CreERT2/MMP3 flox^+/+ mice were Intraperitoneally injected with tamoxifen (TMX) for 5 consecutive days followed by 10 days of rest.

2.2. Reagents

Cholecalciferol (VitD₃), MMP-3 inhibitor, corn oil, tamoxifen, sodium phosphate dibasic heptahydrate (Na₂HPO₄.7H₂O), sodium phosphate monobasic monohydrate (NaH₂PO₄.H₂O), and calcium chloride (CaCl₂) were purchased from Millipore Sigma (St. Louis, MO, USA). Human MMP-3 Recombinant Protein, Dulbecco’s Modified Eagle’s Medium (DMEM), and fetal bovine serum (FBS) were obtained from Thermo Fisher Scientific (Waltham, MA, USA). Other reagents were from Millipore Sigma (St. Louis, MO).

2.3. Vascular aortic SMC culture and calcification

Rat aortic SMCs were isolated from the aortas of male Sprague-Dawley (SD) rats (250 ± 50 g) using an enzymatic method as previously described.²³ Rat aortic SMCs were cultured in a DMEM medium containing 10% FBS and 1% pen/strep in a humidified, 37°C, 5% CO₂ incubator and were used less than passage 8. Human aortic SMCs were obtained from Cell Biologics and were cultured in a DMEM medium containing 20% FBS and 1% pen/strep. To induce calcification in rat aortic SMCs, the confluent SMCs were cultured in a calcification medium containing 3.5 mM Pi and 3 mM CaCl₂ for 7 days. The calcification medium was replaced every 2–3 days. To induce calcification in human aortic SMCs, the confluent human aortic SMCs were cultured in a calcification medium containing 3.2 mM Pi for 6 days.

2.4. Medial calcification rodent model

Medial calcification was induced by VitD₃ injection as described previously.²¹ Briefly, for the calcification rat model, SD rats were subcutaneously (SC) injected with 3 × 10⁵IU/kg VitD₃ for three consecutive days, and then harvested after 11 days. For the mouse calcification model, C57BL/6J mice were injected with 5 × 10⁵IU/kg VitD₃ for three consecutive days and then harvested after 7 days. VitD₃ solution was prepared by dissolving VitD₃ powder in denatured alcohol, Kolliphor EL, and D-glucose solution, successively. The vehicle solution is similar to VitD₃ solution but without VitD₃. At the end of the experiment, the animals were perfused with saline and the aortas were harvested.

2.5. Aortic organ calcification ex vivo

Thoracic aortas were gently dissected from mice and cleared of fat tissues, and then cut into 3 mm-long aortic rings. The ring segments were immediately cultured in a DMEM medium containing 10% FBS and 1% pen/strep in a humidified incubator (37°C, 5% CO₂). Calcification was induced by adding 2.6 mM phosphate (Pi) in the culture medium and cultured for 12 days. Calcification medium was replaced every 2–3 days.

2.6. Calcification assessment

Calcium assay was performed using the o-cresolphthalein complexone method as described previously.^21,24 Briefly, SMCs, aortic segments, or dried aortas were placed in 0.6 N HCl for 2 days before assessment of calcium content. For SMC calcification, the concentration of protein concentration was determined using a Pierce™ bicinchoninic acid (BCA) Protein Assay Kit (Thermo Fisher Scientific, 23225) and calcification was normalized to total protein and expressed as µg/mg protein. For aortic ring segments, calcification was shown as mg/mm aorta. For medial calcification, the extent of calcification was presented as mg/mg dry aorta tissue. Additionally, Von Kossa staining was performed to examine calcium deposits in the aortas. Meanwhile, Verhoeff-Van Gieson’s (VVG) staining was used to examine the integrity of elastin.

2.7. siRNA transfection

Rat MMP-3 siRNA and control siRNA were synthesized by Millipore Sigma. Rat aortic SMCs were transfected with 50 nM MMP-3 siRNA or control siRNA for 2 days using Lipofectamine RNAi/MAX reagent (Thermo Fisher Scientific) and then replaced with a calcification medium containing 3.5 mM Pi and 3 mM CaCl₂ for 7 days. The calcification medium was replaced every 2–3 days.

2.8. RNA isolation and quantitative real-time PCR

Total RNA was isolated from SMCs or aorta tissues using RNeasy Mini Kit (Qiagen, 74104). cDNA was synthesized using iScript cDNA Synthesis Kit (Bio-Rad Laboratories, 170–8890). Quantitative real-time PCR (qPCR) was conducted using PowerUp™ SYBR^® Green Master Mix (ThermoFisher Scientific, A25742) in an Applied Biosystems 7500 Fast Real-Time PCR machine. The levels of mRNA were obtained using the comparative Ct method and normalized with internal control glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The primers used for qPCR were listed in Supplementary material online, Table S1.

2.9. Western blot

SMCs were lysed in radioimmunoprecipitation assay buffer with protease inhibitor Cocktail (Millipore Sigma, Louis, MO, USA). The concentrations of total proteins were measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). Cell lysates were loaded on SDS-PAGE gel and transferred into the polyvinylidene difluoride membrane. The membrane was then incubated with a primary antibody and a horseradish peroxidase (HRP)-conjugated secondary antibody. The western blot signals were obtained using Amersham ECL Western Blotting Detection Reagent (GE Healthcare Life Science, Amersham, UK) and a Kodak film professor. The primary antibodies used in this study were anti-MMP-3 (Abcam, Ab53015) and anti-GAPDH (Millipore Sigma, MAB374).

2.10. Histological staining

Animals were perfused with saline and 10% neutral buffered formalin (NBF) and the abdominal aortas were dissected, fixed, embedded, and paraffin sections were cut. Haematoxylin and eosin (H&E) staining was used to examine tissue morphology. Verhoeff-Van Gieson (VVG) staining was used to assess the integrity of elastin. Von Kossa staining was performed to examine calcium phosphate deposition.

2.11. Immunofluorescent staining

Aortas were isolated and fixed in 10% phosphate-buffered formalin (NBF), and the paraffin cross-sections were prepared. The sections were then deparaffinized, followed by antigen retrieval. Next, the sections were blocked with Dako serum-free blocking solution and incubated with primary antibodies and fluorescent dye-conjugated secondary antibodies Alexa Fluor 555 or 488 (ThermoFisher Scientific, A-11008 or A-11001). The primary antibodies anti-MMP-3 (Abcam, Ab53015) and anti-SM-a-actin (Dako, M0851) were used. 4',6-diamidino-2-phenylindole was used to stain nuclei. The images were captured by Zeiss LSM 880 confocal microscope.

2.12. Immunohistochemistry

Human calcified and uncalcified tibial arteries were obtained from CHTN (Cooperative Human Tissue Network, Rockville, MD, USA). The procedures were conducted in accordance with the IRB protocol (IRB#: 2011P-000121) approved by Beth Israel Deaconess Medical Center Committee on Clinical Investigation. The tissues were fixed in 10% NBF then embedded, and paraffin sections were cut. The sections were deparaffinized, antigen-retrieved, blocked with Dako serum-free blocking solution (Dako, X090930), and incubated with primary antibody and biotinylated secondary antibody. The ABC kit (Vector Laboratories, PK-4000) and 3, 3'-diaminobenzidine substrate kit (Vector Laboratories, SK-4100) or Vector Red Substrate Kit were used for the detection. Haematoxylin counterstaining was performed. The primary antibodies were anti-MMP-3 (Abcam, ab53015), anti-RUNX2 (Medical and Biological Laboratories, D130-3), anti-BMP2 (Abcam, Ab14933), anti-SM-MHC (Thermo Fisher Scientific, 21404-1-AP), and anti-SM22α (Abcam, ab14106).

2.13. Statistical analysis

Statistical analyses were conducted using GraphPad Prism 7 and the statistical analyses were performed using t-test or one-way analysis of variance (ANOVA) two-way ANOVA with multiple comparisons. Data are presented as mean ± SD or mean ± SE. P < 0.05 was considered statistically significant.

3. Results

3.1. MMP-3 is induced during calcification in vitro and in vivo

Our previous study has demonstrated that MMPs are important in vascular calcification using MMP pan inhibitors.²¹ To identify the specific MMP that primarily contributes to medial artery calcification, we injected rodent with sublethal doses of VitD₃, the models thought to recapitulate human medial calcification.^21,25 As shown in Figure 1A, in VitD₃ injection mouse model, qPCR results showed that MMP-3 was the most highly induced MMP among MMPs in calcified arteries. Western blotting data also confirmed that MMP-3 was increased in calcified arteries (Figure 1B and C). We also examined MMP-3 levels in VitD₃-injected rats. As shown in Figure 1D (upper panel), Von Kossa staining showed calcium deposits in the medial layer (Primarily SMCs) of vessels following VitD₃ injection. Concomitantly, immunohistochemistry staining showed that MMP-3 was highly expressed and colocalized with calcifying areas (Figure 1D, bottom panel). To further confirm this finding, qPCR was performed. Consistently, MMP-3 was increased in the aortas of VitD₃-injected rats compared with vehicle controls (Figure 1E). To further examine if the levels of MMP-3 were induced in the calcify conditions in vitro, rat aortic SMCs were cultured in a calcification medium containing 3.5 mM Pi and 3 mM Ca for 7 days. As shown in Figure 1F, Alizarin Red S staining showed increased calcium deposits in Pi/Ca-treated rat SMCs compared with non-calcifying cells. Accordingly, the expression and activity of MMP-3 were highly induced in the calcifying SMCs (Figure 1G–I). Additionally, we examined whether MMP-3 was also induced in human calcifying SMCs. As shown in Figure 1J–L, as expected, Pi treatment significantly increased MMP-3 in both mRNA and protein levels. All these data suggested a potential role for MMP-3 in regulation of SMC calcification.

Figure 1.

MMP-3 is increased under calcifying conditions in vitro and in vivo. (A) MMP-3 was the most highly induced MMP in a model of medial calcification induced by high doses of vitamin D₃ (Vit D₃). 10-week-old male C57BL/6J mice were SC injected with 5 × 10⁵IU/kg for three consecutive days, and the aortas were harvested after 7 days. n = 3 mice. (B and C). Western blotting data showing increased MMP-3 expression in the aortas of VitD₃-injected mice. n = 3 mice. (D and E). The expression of MMP-3 was increased in the calcified arteries of VitD₃-injected rats. Male SD rats were subjected to SC injection with 3 × 10⁵IU/kg Vit D₃ for three consecutive days, and the aortas were harvested after 11 days. (D) Immunohistochemistry staining of MMP-3. Von Kossa staining shows the deposition of calcium phosphate. Scale bar, 100 μm. E, qPCR results. n = 3 rats. (F) Alizarin Red S staining for calcium deposits showed that calcification occurred in calcifying SMCs in vitro. Confluent rat aortic SMCs were cultured in a calcification medium containing 3.5 mM phosphate (Pi) and 3 mM calcium (Ca) for 7 days. (G and H) MMP-3 mRNA and protein levels were increased in response to Pi/Ca in rat aortic SMCs. n = 3 independent experiments. (I) MMP-3 activity was increased in response to Pi/Ca in rat aortic SMCs. The activity of MMP-3 was examined using an MMP-3 Activity Assay Kit. n = 3 independent experiments. (J–L) MMP-3 mRNA and protein levels were also increased in calcifying human aortic SMCs. Confluent human aortic SMCs were cultured in a calcification medium containing 3.2 mM Pi for 6 days. qPCR and western blotting results were normalized using GAPDH. n = 3 independent experiments. Data were analyzed by t-test. Data were analyzed by t-test. Values are mean ± SD. *P < 0.05, P < 0.01, ***P < 0.001.

3.2. MMP-3 is highly expressed in calcified arteries from vascular patients

To examine whether MMP-3 expression is also increased in human calcified arteries, tibial arteries from deceased donors without PAD or from patients undergoing amputation due to end-stage diabetic vascular disease were paraffin-embedded and cross-sectioned. Immunohistochemical staining for osteogenic markers RUNX2 and BMP2, SMC marker SM-a-actin, and MMP-3 was performed. As shown in Figure 2A, although the arterial tissues were decalcified due to the sectioning requirement, we were still able to identify that calcification was not present in normal arteries but prominent in the medial layers of the arteries from patients with PAD. Moreover, RUNX2 and BMP2 were increased but SM-a-actin was decreased in the calcified arteries compared with normal vessels (Figure 2C–E), suggesting that osteogenic transformation was increased in calcified arteries. As expected, MMP-3 expression was largely increased in the calcified areas of the tibial arteries in patients with PAD (Figure 2B). Notably, MMP-3 is also colocalized with osteogenic markers BMP2 and RUNX2. These data suggest that MMP-3 could be critical in vascular osteogenic transformation and medial calcification in vascular patients.

Figure 2.

MMP-3 is highly expressed in calcified arteries from patients with arterial disease. Human tibial arteries were fixed and decalcified, and paraffin sections were prepared. (A) Immunohistochemistry staining showed that MMP-3 and osteogenic markers RUNX2 and BMP2 were increased but SM-α-actin was decreased in calcified arteries. Scale bar, 200 μm. (B–E) The quantitative data. Data were analyzed by t-test. Values are mean ± standard error (SE). n = 4 specimens. *P < 0.05, P < 0.01, ***P < 0.001. L, lumen. Outlined area: calcified area.

3.3. MMP-3 mediates SMC calcification and osteogenic transformation in vitro

To determine whether MMP-3 plays a role in vascular calcification, rat SMCs were treated with the specific MMP-3 siRNA or control siRNA and then cultured in a calcification medium containing Pi/Ca. As shown in Figure 3A and Supplementary material online, Figure S1A, the addition of Pi/Ca largely increased calcium content in the control siRNA-transfected SMCs, however, this pro-calcification effect was significantly suppressed in MMP-3 siRNA-transfected cells. Consistently, we also observed that a small-molecule MMP-3 inhibitor could dose-dependently inhibit Pi-induced SMC calcification (Figure 3B). To examine whether the activated MMP-3 is capable of inducing calcification, we employed a human recombinant MMP-3 protein. As shown in Figure 3C, the treatment with MMP-3 recombinant protein dose-dependently enhanced Pi/Ca-induced SMC calcification. These data suggest that MMP-3 is an important regulator of SMC calcification. To examine if MMP-3 is important for SMC osteogenic transformation, the osteogenic markers Osterix (Sp7) and ALP (TNSALP), and SMC markers SM-22a (TAGLN) and SM-MHC (MYH11) were assessed by qPCR. As shown in Figure 3D–G, in response to Pi/Ca, the levels of Sp7 and NSALP were markedly increased, whereas TAGLN and MYH11 were decreased, indicating increased osteogenic transformation. However, inhibition of MMP-3 significantly suppressed this effect. These data suggest that MMP-3 is capable of mediating SMC osteogenic transformation. To determine whether MMP-3 plays a similar role in human SMC calcification, human SMCs were first transfected with human MMP-3 siRNA or control siRNA and then cultured in Pi calcification medium, As shown in Figure 3H and Supplementary material online, Figure S1B, transfection of MMP-3 siRNA markedly suppressed Pi-mediated calcification in human SMCs. Likewise, knockdown of MMP-3 inhibited Pi-induced osteogenic transformation. Additionally, we performed a transition experiment using mouse SMCs isolated from MMP-3-WT and MMP-3-KO mice. As shown in Figure 3M, deficiency of MMP-3 in SMCs from MMP-3-KO mice significantly blocked Pi-induced calcification compared with that from MMP-3-WT mice. These data suggest that MMP-3 plays a critical role in SMC calcification.

Figure 3.

MMP-3 mediates SMC calcification and osteogenic transformation. (A) Knockdown of MMP-3 inhibited calcification in rat aortic SMCs. Rat aortic SMCs were transfected with 100 nM control siRNA or MMP-3 siRNA for 2 days and then cultured in a calcification medium containing 3.5 mM Pi/3 mM Ca for 7 days. Calcium content was examined by calcium assay. (B) The small-molecule MMP-3 inhibitor dose-dependently suppressed Pi/Ca-induced SMC calcification. Confluent rat aortic SMCs were treated with various concentrations of MMP-3 inhibitor and then cultured in 3.5 mM Pi/3 mM Ca calcification medium for 7 days. The culture media and inhibitor were replaced every other day. (C) Activated MMP-3 promoted Pi/Ca-mediated SMC calcification. Confluent rat aortic SMCs were pre-treated with indicated doses of human recombinant MMP-3 protein and then cultured in 3.5 mM Pi/3 mM Ca calcification medium for 7 days. (D–G) Inhibition of MMP-3 suppressed SMC osteogenic transformation in rat aortic SMCs. The osteogenic markers Osterix (Sp7) and ALP (TNSALP) and SMC markers SM-MHC (MYH11) and SM22α (TAGLN) were determined by qPCR and normalized with GAPDH. (H) Knockdown of MMP-3 significantly inhibited calcification in human aortic SMCs. Human aortic SMCs were transfected with 100 nM control siRNA or MMP-3 siRNA for 2 days and then cultured in a calcification medium containing 3.2 mM Pi for 6 days. (I) MMP-3 inhibitor dose-dependently suppressed Pi-induced calcification in human aortic SMCs. (J–L) MMP-3 inhibitor suppressed osteogenic transformation in human aortic SMCs. Confluent human aortic SMCs were pre-treated with MMP-3 inhibitor and then cultured in 3.02 mM Pi calcification medium for 6 days. (M) Deletion of MMP-3 blocked Pi-mediated mouse aortic calcification. Mouse aortic SMCs isolated from MMP-3-WT and MMP-3-KO mice were cultured in a calcification medium containing 3.5 mM Pi for 12 days. The culture media and inhibitor were replaced every other day. Data were analyzed by one-way ANOVA or two-way ANOVA with multiple comparisons. One-way ANOVA analysis was corrected with a post hoc test. Values are mean ± SD. n = 3–4 independent experiments *P < 0.05, **P < 0.01, ***P < 0.001.

3.4. Global deficiency of MMP-3 reduces medial calcification ex vivo and in vivo

To investigate whether MMP-3 plays a role in medial calcification, we first used an ex vivo aortic ring organ culture model.^24,26 This model has been used to distinguish local from systemic effects of factors on medial calcification. Aortic ring segments were isolated from MMP-3-WT and MMP-3-KO mice and then cultured in calcification medium with or without 2.6 mM Pi for 8 days. As shown in Figure 4A, the calcium assay showed no calcium accumulation in the cultured aortic rings from MMP-3-WT and MMP-3-KO mice under non-calcifying conditions. By contrast, Pi treatment markedly increased calcium content in the aortic rings of MMP-3-WT mice. However, this effect was significantly reduced in aortic rings from MMP-3-KO mice. These data further suggest that local MMP-3 expression is most likely to play an important role in the development of medial calcification. To further determine the role of MMP-3 in arterial medial calcification in vivo, we used high-dose VitD₃ injection in mice to create a pro-calcification environment. As shown in Figure 4B, after vehicle injection, neither MMP-3-WT nor MMP-3-KO mice showed detectable arterial calcification. By contrast, VitD₃ injection robustly increased aortic calcification in MMP-3-WT mice, however, this effect was significantly reduced in MMP-3-KO mice. Likewise, Von Kossa staining showed that MMP-3-KO mice had fewer calcium deposits in the medial layers than MMP-3-WT mice (Figure 4C). To further study whether MMP-3 was capable of regulating osteogenic transformation in vivo, qPCR was performed for the osteogenic markers RUNX2 and ALP and SMC markers SM22a and SMMHC. As shown in Figure 4D–H, VitD₃ injection significantly increased the expression of RUNX2 and BMP2, whereas it decreased SMMHC and SM22a in MMP-3-WT mice. VitD₃-induced osteogenic transformation, however, was significantly reduced in MMP-3-KO mice. These data suggest that MMP-3 is pivotal in osteogenic transformation and medial calcification.

Figure 4.

Global deficiency of MMP-3 reduces medial calcification ex vivo and in vivo. (A) Deletion of MMP-3 inhibited aortic calcification ex vivo. Aortas were isolated from MMP-3-WT and MMP-3-KO mice and then cut into aortic rings, followed by incubation in a calcification culture medium containing 2.6 mM Pi for 12 days. The calcification medium was changed every 2–3 days. n = 3–6 mice. (B and C). Global deletion of MMP-3 reduced medial calcification in vivo. MMP-3-WT and MMP-3-KO mice were subjected to the injection with 5 × 10⁵IU/kg VitD₃ for 3 consecutive days. 7 days later, the aortas were harvested. (B) calcium assay results. (C) Von Kossa staining results. Vehicle, n = 5; VitD₃, n = 15. Scale bar, 100 μm. (D–H) Deletion of MMP-3 inhibited osteogenic transformation in vivo. Immunohistochemistry staining and qPCR were performed for osteogenic markers RUNX2 (RUNX2) and BMP2 (BMP2), and SMC markers SMMHC (MYH11) and SM22α (TAGLN). The results were normalized with GAPDH. n = 3 mice. Scale bar, 100 μm. Data were analyzed by two-way ANOVA with multiple comparisons. Values are mean ± SE. **P < 0.01, ***P < 0.001.

3.5. Generation of SMC-specific MMP-3 conditional knockout mouse

It is known that vascular SMCs play a critical role in the development of calcification. We next attempted to investigate if MMP-3-regulated calcification was attributed to vascular SMC-specific MMP-3 expression. First, we generated MMP-3 floxed mice using CRISPR/Cas9-mediated genome engineering technology. As shown in Figure 5A, Exons 2–4 were selected as the conditional knockout regions. Cas9, guide RNA, and targeting vector were co-injected into fertilized eggs, and the pups were genotyped by PCR using specific primers, followed by sequencing analysis. As shown in Figure 5B, four positive MMP-3 floxed F1 mice including two males and two females were identified using specific primers and passed health RapidCHECK. To validate these MMP-3 floxed mice, we first used a Cre-lox strategy in vitro. SMCs were isolated from MMP-3 floxed mice, cultured, and infected with the adenovirus coding LacZ (Adeno-LacZ) or Cre (Adeno-Cre). As shown in Figure 5C, Adeno-Cre infection substantially decreased MMP-3 mRNA levels compared with Adeno-LacZ infection. Consistently, Adeno-Cre infection also reduced MMP-3 protein levels (Figure 5D and E), indicating that MMP-3 floxing is successful. To create SMC-specific MMP-3 knockout mice, MMP-3 floxed mice (MMP3 flox^+/+) were first bred with SMC-specific Cre mice SMMHC-CreER^T2 to produce SMMHC^CreERT2/MMP-3 flox^+/+ mice, and then SMMHC^CreERT2/MMP3 flox^+/+ mice were injected with Oil or tamoxifen (TMX) to produce wild-type and SMC-specific MMP-3 knockout mice. To validate if MMP-3 was specifically deleted in SMCs in vivo, we isolated the aortas from Oil and TMX-injected SMMHC^CreERT2/MMP3 flox^+/+ mice and then removed the adventitial layers and scraped endothelial cells. The remaining medial layers primarily containing SMCs were utilized for MMP-3 expression analysis. As shown in Figure 5F, qPCR results showed that the medial layers from TMX-injected SMMHC^CreERT2/MMP3 flox^+/+ mice have much lower MMP-3 levels than those from Oil-injected mice, indicating that MMP-3 was deficient in SMCs derived from TMX-injected SMMHC^CreERT2/MMP-3 flox^+/+ mice. To further examine if MMP-3 was exclusively deficient in SMC-enriched arterial medial layers of MMP-3 knockout mice, immunofluorescence staining was performed. As shown in Figure 5G, in Oil-injected mice, MMP-3 was expressed in the medial layers, however, TMX injection largely reduced MMP-3 staining in these areas, indicating that TMX injection is capable of inducing SMC-specific deficiency in SMMHC^CreERT2/MMP-3 flox^+/+ mice.

Figure 5.

SMC-specific MMP-3 knockout mouse generation. (A) MMP-3 conditional knockout targeting strategy. The exon 2-4 of MMP-3 gene were selected as the knockout regions and the targeting vector was synthesized. CRISPR/Cas9 technology was used for MMP-3 conditional knockout mouse production. (B) The data from agarose gel electrophoresis showed that four MMP-3 floxed F1 mice including two males and two females were identified using the specific primers. (C and E) Validation of MMP-3 floxed mice. The results from qPCR and Western blotting showed that the adenovirus coding Cre robustly decreased MMP-3 expression compared with LacZ controls in SMCs. Mouse aortic SMCs were isolated from MMP-3 floxed mice using the enzymatic method. Three to five passage SMCs were used for this study. The cultured SMCs were infected with 100 multiplicity of infection adenovirus coding LacZ (Adeno-LacZ) or Cre (Adeno-Cre) for 4 days. (C) qPCR results; (D) Representative Western blotting images; (E) Quantitative data for Western blotting. n = 3 independent experiments. (F) SMC-specific MMP-3-KO mouse was generated. MMP-3 floxed mice were first bred with SMC-specific Cre mice SMMHC-CreER^T2 to produce SMMHC^CreERT2/MMP-3 flox^+/+ mice. SMMHC^CreERT2/MMP3 flox^+/+ mice were then injected with oil or tamoxifen (TMX, 75 mg/kg, IP) for five consecutive days, followed by 10 days of rest, to produce wild-type and SMC-specific MMP-3-KO mice. To confirm that MMP-3 was exclusively expressed in SMCs, the aortas were dissected from TMX or Oil-injected SMMHC^CreERT2/MMP3 flox^+/+ mice and then digested with collagenase to remove the adventitial layers. The endothelial cells were also scraped using tiny cotton swabs. The remained medial layers mainly containing SMCs are used for qPCR. qPCR results showed that MMP-3 expression in the medial layers of TMX-injected mice was extremely low compared with Oil-injected mice. n = 4 mice. (G) Immunofluorescence staining showed that MMP-3 expression was largely decreased in the medial layer of the carotid arteries of SMMHC^CreERT2/MMP3 flox^+/+ mice after TMX injection compared with Oil injection. The expression of MMP-3 and SMC marker SM-a-actin were examined by immunofluorescence staining. Scale bar, 40 μm. Data were analyzed by t-test. Values are mean ± SD. **P < 0.01, ***P < 0.001. L, lumen; M, media. A, adventitial.

3.6. SMC-specific MMP-3 deletion reduces osteogenic transformation and medial calcification

To explore if SMC-specific MMP-3 expression plays a critical role in mediating medial calcification, SMC-specific promotor-driven MMP-3 floxed (SMMHC^CreERT2/MMP-3 flox^+/+) mice were first injected with Oil or TMX and then subjected to VitD₃ injection. As shown in Figure 6A, the extent of calcium was very low in the aortas of both Oil and TMX-injected SMMHC^CreERT2/MMP3 flox^+/+ mice treated with vehicle. Following VitD₃ injection, calcification was robustly increased in Oil-injected mice, however, this effect was significantly reduced in TMX-injected mice. Additionally, Von Kossa staining was performed to visualize calcium deposits in the vessels. As shown in Figure 6B, there was no positive Von Kossa staining detected in both Oil and TMX-treated mice after the injection of the vehicle. After VitD₃ injection, although calcium deposits were increased in the medial layer of both Oil and TMX-treated mice, the extent of calcification in TMX-treated mice was significantly lower than that in Oil-treated mice. These data suggest that SMC-specific MMP-3 is crucial for the development of medial calcification. It has been demonstrated that elastin is implicated in vascular calcification.^27,28 MMP-3 can degrade extracellular matrix such as elastin. Therefore, VVG staining for elastin integrity was performed. As expected, VitD₃ injection resulted in extensive degradation of elastin in Oil-injected but not in TMX-injected SMMHC^CreERT2/MMP-3 flox^+/+ mice (Figure 6C–E), suggesting that MMP-3-mediated elastin degradation could play a role in medial calcification. To further determine the role of SMC-specific MMP-3 in osteogenic transformation, the osteogenic markers RUNX2 and BMP2 and SMC markers SMMHC and SM22α were examined by immunohistochemistry staining. As shown in Figure 6F–J, SMC-specific MMP-3 deletion in TMX-injected mice significantly decreased the expression of RUNX2 and BMP2 and increased SMMHC and SM22α compared with Oil-injected mice after VitD₃ injection. These data suggest that MMP-3 in vascular SMCs is an important regulator in osteogenic transformation and medial calcification.

Figure 6.

SMC-specific MMP-3 deletion reduces osteogenic transformation and medial calcification. 7-week-old male SMMHC^CreERT2/MMP-3 flox^+/+ mice were IP injected with 75 mg/kg tamoxifen (TMX) or Oil for five consecutive days and waited for 10 days of rest, followed by subcutaneous injection with 5 × 10⁵IU/kg VitD₃ or vehicle for 3 consecutive days. Aortas were collected after 7 days. (A) Calcium assay showed that deficiency of SMC-specific MMP-3 suppressed VitD₃-induced medial calcification in vivo. (B) Quantitative data from Von Kossa staining showed that SMC-specific MMP-3 deficiency decreased calcium deposits in the abdominal aortas. (C) Representative images of HE, Von Kossa, and VVG staining. Scale bar, 100 μm. (D and E) Deficiency of SMC-specific MMP-3 blocked VitD₃-mediated elastin degradation. The integrity of elastin was examined by VVG staining. The elastin area and elastin break were analyzed. (F) Deletion of SMC-specific MMP-3 inhibited osteogenic transformation in vivo. Representative immunohistochemistry staining images for osteogenic markers RUNX2 and BMP2, and SMC marker SMMHC. Scale bar, 100 μm. (G–J) Quantitative results. Data were analyzed by one-way ANOVA or two-way ANOVA with multiple comparisons. One-way ANOVA analysis was corrected with a post hoc test. Values are mean ± SE. Vehicle, n = 6; VitD₃, n = 16–22. **P < 0.01, ***P < 0.001. L, lumen.

4. Discussion

We herein provide data suggesting that matrix metalloproteinase 3 (stromelysin-1) plays a central role in promoting medial calcification. We show that MMP-3 is increased in calcifying rodent and human SMCs and that it is highly expressed in a mouse model of arterial calcification as well as in diseased tibial arteries from patients with PAD. Inhibition of MMP-3 using a small-molecule inhibitor decreased calcification in SMCs exposed to calcifying medium, whereas addition of MMP-3 recombinant protein enhanced calcification. Deficiency of either global or SMC-specific MMP3 in mice significantly reduced medial calcification and osteogenic transformation. These findings suggest that targeting vascular MMP-3 expression could be a successful strategy to reduce vascular calcification and improve outcomes in patients with PAD.

In patients with PAD, higher levels of medial calcification are associated with increased morbidity and mortality. It can predict amputation and death in patients with chronic limb-threatening ischaemia, and it has incremental value in risk prediction over standard risk factors and the ankle-brachial index.²⁹ We have previously shown that broad spectrum MMP inhibitors such as the antibiotic doxycycline and the hydroxamic acid zinc chelator GM6001 can prevent medical calcification in rodent models. In the present study, we first showed that MMP-3 was the most strongly upregulated of all MMPs investigated in a rodent model of medial calcification caused by high doses of vitamin D₃. The relationship between this model and human medial calcification is unclear. Histological evaluation, however, shows calcific patterns that share similarities with those identified in tibial arteries from patients with diabetes who require amputation as we show in Figures 1D and 2A. One advantage of the vitamin D₃ calcification model is that it results in relatively consistent levels of calcium accumulation. Other models such as the 5/6 nephrectomy plus phosphate model or the adenine model that induces renal failure may provide further insights. Interestingly, patients with chronic kidney disease are typically treated with an active form of vitamin D, overriding their deficiency of the vitamin’s activity that is related to multiple factors.³⁰ MMP-3 expression was not seen in tibial arteries from autopsy specimens taken from patients without evidence of peripheral artery disease while it was significantly expressed in calcified vessels from patients undergoing amputation due to diabetes. The specific factors that induce MMP-3 expression in calcified arteries, however, remain unknown.

In normal arteries, the medial layer mainly consists of elastin-surrounded smooth muscle cells. Both human arteries and animal models show that medial calcification mostly occurs along the elastic lamellae in the medial layer, especially in the areas with extensive degradation.^3,25 The first phase of calcification is associated with small calcium deposits on the elastin fibres, but the late phase involves the accumulation of mineral deposits as well as extensive changes in elastin morphology along the entire medial layer. Elastin fibres become thin and broken, and vessels are dilated due to loss of plasticity. Studies have shown that elastin degradation can accelerate vascular calcification.^28,31 In this study, we performed VVG staining for elastin integrity and observed that both global and SMC-specific MMP-3 deficiency markedly reduced elastin degradation and medial calcification, suggesting that MMP-3-mediated elastin degradation could be one of the major mechanisms through which MMP-3 regulates medial calcification, although there is no direct evidence showing that elastin degradation contributes to calcium deposition during medial calcification. It is also possible that calcium deposits, especially in the late phase of calcification, could provoke sterile inflammation to secrete extracellular matrix-degrading enzymes, and ultimately degrade elastin fibres.

Recent investigations have demonstrated that in addition to MMP-3, both MMP-2 and MMP-9 are capable of mediating calcification in CKD.^32,33 MMP-9 also contributed to atherosclerotic calcification in ApoE knockout mice.³⁴ Additionally, a more recent study showed that MMP-10 was increased in the calcified aortic valves and suggested that it could play an important role in the progression of aortic valve calcification and stenosis.³⁵ However, the detailed mechanisms by which these matrix-degrading enzymes are able to regulate vascular calcification remain unknown. SMC osteogenic transformation is one of the core elements of vascular calcification in pre-clinical models. Similar to bone cell differentiation, SMCs can change their phenotype from contractile to bone-like cells. Several signalling pathways are involved in this phenotypic switch. For example, wingless-related integration site (Wnt) and bone morphogenic proteins (BMPs) signalling pathways have been shown to play an important role in this process.^36,37 The Wnt and BMP cascade can target RUNX2 which is primarily responsible for this phenotypic change. Our present study showed that MMP-3 expression was increased in calcifying arteries from rodents and patients with calcified tibial arteries. We further showed that MMP-3 expression was colocalized with osteogenic markers. Knockdown or deficiency of MMP-3 was able to suppress SMC osteogenic transformation in vitro and in vivo. These data suggest that MMP-3 can mediate calcification at least in part, by regulating SMC osteogenic transformation. However, the detailed mechanisms by which MMP-3 regulates SMC osteogenic transformation need further investigation.

Calcification is a complex process and involves overlapping mechanisms. To determine whether the effects of MMP-3 on calcification were based on systemic changes, we used an ex vivo organ culture calcification model. In this model, aortas isolated from MMP-3-WT and MMP-3-KO mice were directly cultured in a Pi-containing calcification medium. This effectively excludes the effects of circulating factors. We observed that loss of MMP-3 in ex vivo aortic segments significantly decreased calcification compared with wild-type controls, suggesting that MMP-3 mediates calcification most likely through local effects within the vasculature. It is known that the vascular wall contains multiple cell types that may contribute to the initiation and control of calcification. Emerging evidence suggests that multipotent vascular stem cells also exist in the medial layer and these cells have multilineage potential for osteogenic differentiation, a potentially permissive factor in vascular mineralization.^38,39 Studies have also shown that macrophages can promote calcification by regulating vascular cell osteogenic transformation.^40,41 Additionally, endothelial cells may also be involved in vascular calcification by regulating an endothelial-to-mesenchymal transition (EndMT).^42,43 Importantly, in this study, we used SMC-specific inducible MMP-3 knockout mice to define the role of SMC MMP-3 in medial calcification. We observed that the deletion of MMP-3 in SMCs significantly reduced both medial calcification and osteogenic transformation, suggesting that SMC-specific MMP-3 is critical for regulating medial calcification. Although the potential role of MMP-3 expression from other cellular sources in the arterial wall, such as macrophages, endothelial and adventitial cells has yet to be investigated, it is noteworthy that global MMP-3 deficiency and SMC-specific MMP-3 deletion resulted in a similar inhibitory effect on calcification, suggesting that SMC MMP-3 likely plays a major role.

Recent studies have demonstrated that MMP-3 could function as an upstream activator of other MMPs, such as MMP-9, and therefore contribute to the development of vascular diseases.^44,45 It has been reported that MMP-3 can activate MMP-9, which mediates vascular SMC proliferation and pathological remodelling.⁴⁴ MMP-3 can also play a major role in placental leukocytes by regulating the MMP-9 activation process.⁴⁶ In our supplementary material online, Figure S3, we found that although knockdown of MMP-3 did not affect MMP-2, it indeed slightly decreased MMP-9. However, the data from casein and gelatine zymography revealed that inhibition of MMP-3 did not change the activities of MMP-2/9. These data suggest that MMP-3 mediates SMC calcification through a mechanism that does not include the activation of MMP-2 and MMP-9. We have previously shown that proliferation occurs in areas of calcification.²⁵ The role of proliferation and the relationships between MMP-3, proliferation, and calcification will require further investigation.

In recent clinical studies, we showed that long-term doxycycline treatment did not slow the progression of arterial calcification in patients with small aortic aneurysms.⁴⁶ However, the drug treatment dose was unable to decrease MMP-3 levels in serum. These data along with the present findings suggest that a more focused strategy using selective MMP inhibitors may be more effective.

In summary, we have shown that MMP-3 is highly expressed in calcifying arteries in rodents and humans. MMP-3 inhibition significantly suppresses Pi-induced SMC calcification in vitro and deletion of MMP-3 reduces medial calcification in vivo. Knockdown or inhibition of MMP-3 suppresses SMC calcification in vitro, and MMP-3 deficiency through either global deletion or SMC-specific knockout can reduce medial calcification in vivo. The most likely mechanisms by which MMP-3 mediates medial calcification are through effects on elastin degradation, osteogenic transformation of vascular SMCs, or both. These findings suggest that MMP-3 inhibition strategies may reduce calcification or slow its progression. Further investigations are needed to determine whether the beneficial effects of MMP-3 inhibition on limiting medial calcification can result in improved outcomes for patients with PAD.

Translational perspective.

Increasing calcification in lower extremity arteries strongly predicts adverse cardiovascular events including amputation, and this occurs in a manner that is independent of atherosclerotic occlusive disease. Reducing arterial calcification is now considered a potential strategy to improve outcomes in vascular patients. We present data showing that the matrix-degrading enzyme MMP-3 is strongly upregulated in calcifying cells and organ culture, as well as in calcified arteries from rodent models and patients. We show that smooth muscle cell-specific MMP-3 regulates artery calcification suggesting that inhibition strategies may slow the progression of calcific arteriopathy and improve outcomes in patients with peripheral artery disease.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Funding

This work was supported by the National Heart, Lung, and Blood Institute, NIH, US grants R01HL138357 (to R. Guzman) and R01HL157111 (to Y. Cai; R. Guzman).

Data availability

The authors declare that all supporting data are available within the article and its Data Supplement.