Matrix gla protein (MGP) is a potent inhibitor of extracellular matrix (ECM) mineralization. MGP-deficiency in humans leads to Keutel syndrome, a rare genetic disease hallmarked by abnormal soft tissue calcification. MGP-deficient (Mgp–/–) mice show progressive deposition of hydroxyapatite minerals in the arterial walls and die within 2 months of age. The mechanism of antimineralization function of MGP is not fully understood. We examined the progression of vascular calcification and expression of several chondrogenic/osteogenic markers in the thoracic aortas of Mgp–/– mice at various ages. Although cells with chondrocyte-like morphology have been reported in the calcified aorta, our gene expression data indicate that chondrogenic/osteogenic markers are not upregulated in the arteries prior to the initiation of calcification. Interestingly, arterial calcification in Mgp–/– mice appears first in the elastic laminae. Considering the known mineral scaffolding function of elastin (ELN), a major elastic lamina protein, we hypothesize that elastin content in the laminae is a critical determinant for arterial calcification in Mgp–/– mice. To investigate this, we performed micro–computed tomography (µCT) and histological analyses of the aortas of Mgp–/–;Eln+/– mice and show that elastin haploinsufficiency significantly reduces arterial calcification in this strain. Our data suggest that MGP deficiency leads to alterations of vascular ECM that may in turn initiate arterial calcification. © 2014 American Society for Bone and Mineral Research.
Mineralization of cartilage, bone, and tooth extracellular matrix (ECM) is a physiologic process. In contrast, soft tissue calcification (ectopic mineralization), as seen in the intima or media of the blood vessels, is a pathologic condition. Intimal calcification of the atherosclerotic plaques can lead to plaque erosion and rupture, whereas medial calcification of large elastic arteries often increases arterial stiffness, leading to high systolic and pulse pressures.[1-3] The pathophysiology of these life-threatening diseases is still not well understood.
There is accumulating evidence suggesting that physiologic mineralization in skeletal hard tissues and pathologic calcification of blood vessel walls share multiple common determinants. These determinants include extracellular levels of the mineral ions calcium and inorganic phosphate (Pi), the presence of a scaffolding ECM for mineral deposition, and the relative amounts of mineralization inhibitors such as inorganic pyrophosphate (PPi) and/or matrix gla protein (MGP) present within the ECM microenvironment.[4, 5] In the hard tissues, specialized cells modulate these determinants to facilitate ECM mineralization. For example, in mineralizing bones, osteoblasts synthesize and assemble a collagen-rich scaffolding matrix and at the same time produce high levels of alkaline phosphatase (ALPL) to increase the extracellular Pi/PPi ratio.[4, 5]
In blood vessels, gain of promineralization functions is one way by which vascular smooth muscle cells (VSMCs) might initiate vascular calcification. Transdifferentiation of VSMCs into osteoblast-like cell types has been reported in the calcified tunica media of dialysis patients. VSMCs in these patients express an early chondrogenic/osteogenic transcription factor RUNX2 and subsequently, several other osteoblast markers such as type I collagen, osteopontin (OPN), bone sialoprotein, and ALPL. In a more mechanistic study, the upregulation of signaling through bone morphogenetic protein 2 (BMP2) via a homeobox transcription factor MSX2 has been shown to promote the osteogenic transdifferentiation of VSMCs and vascular calcification in a mouse model of diet-induced diabetes.
Although the above examples clearly suggest the chondrogenic/osteogenic differentiation of VSMCs as a major cause behind the mineralized blood vessels, additional factors might have effects on some forms of vascular calcification. Indeed, several studies demonstrated that ECM proteins, particularly those in the elastic laminae, play critical roles in the regulation of this process.[8-10] Elastin, a major elastic lamina protein produced by VSMCs, has been shown to act as a mineral nucleator both in vitro and in vivo.[9, 10] Elastin peptides have been shown to upregulate matrix metalloproteinase (MMP) expression.[11, 12] On the other hand, MMP-mediated degradation of elastin has been linked to medial calcification in several studies.[13, 14]
In this study, we investigated the effects of elastin content on the initiation and progression of medial calcification in MGP-deficient mice. MGP is a potent mineralization inhibitor secreted by VSMCs in the arteries and chondrocytes in all cartilaginous tissues.[15, 16] MGP deficiency in humans causes Keutel syndrome, a rare genetic disorder characterized by extensive soft tissue calcification, brachytelephalangia, pulmonary stenosis, and midfacial hypoplasia.[17, 18] Mgp–/– mice recapitulate many of these abnormalities, albeit with an acute vascular calcification phenotype.[16, 19] Because of the complications caused by the extensive mineral deposition in all arteries, most of the Mgp–/– mice die before 2 months of age.
The absence of MGP has been implicated in BMP-mediated chondrogenic/osteogenic transdifferentiation of VSMCs. Interestingly, our recent data indicate that this may not be the primary cause of vascular calcification in Mgp–/– mice. Instead, our data suggest that the alterations of vascular ECM in the absence of MGP may result in the observed medial calcification phenotype. MGP is colocalized with elastin in the arterial elastic lamina, which is the first site of ectopic mineral deposition in Mgp–/– mice.[21, 22] Based on this observation and the published data demonstrating a mineral scaffolding role for elastin, we hypothesize that elastin content in the elastic laminae is a critical determinant for arterial calcification in Mgp–/– mice. In support of this hypothesis, herein, we show that elastin haploinsufficiency decreases mineral accumulation in the arteries of Mgp–/–;Eln+/– mice.
Subjects and Methods
Generation of Mgp–/– and Col2a1-lacZ mice has been described.[16, 23] Eln–/– mice were originally generated by Dr. Dean Li at the University of Utah (Salt Lake City, UT, USA); Alpl–/– and Thbs1–/– mice were provided by Drs. Jose Luis Millan at Sandford-Burnham Medical Research Institute (La Jolla, CA, USA), and Jack Lawler at Harvard University (Boston, MA, USA), respectively.[24-26] All the knockout mice were whole-body knockouts. Mgp–/– mice lack exon 1 to 3 and part of exon 4 of the Mgp gene; Eln–/– mice lack the promoter and exon 1 of the Eln gene, and Alpl–/– mice lack exons 1 through 6. As reported earlier, none of these strains produce the mRNA or the functional protein encoded by the targeted gene. Experiments were performed on mice with mixed backgrounds (C57BL/6 and SvJ/129) using gender-matched littermate controls. Mice were maintained in a pathogen-free standard animal facility, and the experimental procedures were performed following an animal use protocol approved by the Animal Care Committee of McGill University. Genotypes were determined by PCR on genomic DNAs isolated from tail biopsies. The sequences of the primers used for genotyping are presented in Supplemental Table 1.
Skeletal preparation and histological analysis
Skeletal tissues from newborn and adult mice were fixed overnight in 95% ethanol, stained in 0.015% Alcian blue dye (Sigma-Aldrich, Oakville, Ontario, Canada) in a 1:4 solution of glacial acetic acid and absolute ethanol for 24 hours. Tissues were then treated with 2% potassium hydroxide for another 24 hours (or until the soft tissues were dissolved) and then stained by 0.005% Alizarin red (Sigma-Aldrich) solution in 1% potassium hydroxide. Finally, the stained skeletal tissues were clarified in 1% potassium hydroxide/20% glycerol for ≥2 days. For plastic sectioning, aortas were fixed overnight in 4% paraformaldehyde (PFA)/PBS, embedded in methyl methacrylate, and sectioned (7-µm-thick), and von Kossa and van Gieson or Hart's staining were applied. For elastin immunohistochemistry, 6-µm-thick frozen sections were cut from wild-type (WT) and Mgp–/– aortas, fixed in 4% PFA in PBS (pH 7.4) and treated with 4 µg/mL proteinase K in Tris-EDTA buffer at 37°C for 15 minutes. Insoluble alpha elastin was detected using a rabbit polyclonal antibody (Abcam; 21607) following the supplier's instructions. Images were taken at room temperature using a light microscope (DM200; Leica) with 20× (numerical aperture of 0.40) and 40× (numerical aperture of 0.65) objectives. All histological images were captured using a camera (DP72; Olympus), acquired with DP2-BSW software (XV3.0; Olympus), and processed using PhotoShop (Adobe, San Jose, CA, USA).
Micro–computed tomography analysis
Micro–computed tomography (µCT) analyses of the skeletal samples were performed at the Centre for Bone and Periodontal Research Core Facility at McGill University using a high-resolution µCT scanner (SkyScan 1072). For all µCT analyses, the X-ray source was operated at 45 kV and at 222 µA. Images were captured using an air-cooled X-ray CCD-camera (1024 × 1, 024 pixels). Samples were scanned at a magnification resulting in a pixel size of 11.25 µm. A rotation step of 0.9 degrees/180 degrees and an exposure time of 2800 ms for each step were used. Threshold was set at 33% of the maximum and the cross-sections along the specimen's long-axis were reconstructed using NRecon (v188.8.131.52;SkyScan). Each cross-section was reduced to half-size to facilitate the analysis, giving a pixel size of 22.50 µm. CT-Analyser (v184.108.40.206) and ANT 3D Creator (v2.4) software (both from SkyScan) were used to analyze and to perform 3D rendering, respectively. The latter software was used to generate the pseudocolor images. The regions of interest were drawn manually.
Tissue samples obtained from calcified and noncalcified arteries were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer. After washing with buffer and a solution of 100% ethyl alcohol-H2O, samples were dehydrated with 100% ethyl alcohol, and embedded in low-viscosity, thermally-curing Epon resin. Ultrathin sections (70–100 nm) were cut from the resin blocks by using a Reichert-Jung Ultracut E ultramicrotome with a Diatome (Biel, Switzerland) diamond knife. The sections were transferred onto 200-mesh Cu transmission electron microscopy (TEM) grids with Formvar support film for image analysis in a FEI Tecnai 12 transmission electron microscope equipped with a Gatan (Pleasanton, CA, USA) Bioscan CCD camera Model 792 and at an accelerating voltage of 120 kV in bright-field mode. Structural and chemical analysis of mineralized materials were obtained from nonstained sections using a FEI Tecnai G2 F20 cryo-scanning transmission electron microscope (STEM) equipped with a Gatan Ultrascan 4000 4k × 4k CCD camera Model 895 at an accelerating voltage of 200 kV. Chemical composition was determined by energy dispersive spectroscopy using the EDAX Genesis microanalytical system.
Gene expression analysis
Gene expression analyses were performed using a quantitative real-time PCR (qRT-PCR) system (Model 7500; Applied Biosystems, Mississauga, ON, Canada). Total RNA was extracted from different tissues with TRIZOL reagent (Invitrogen, Burlington, ON, Canada) and subjected to DNase I (Invitrogen) treatment. The first-strand cDNA synthesis and qRT-PCR were performed using a high-capacity cDNA reverse-transcription kit (Applied Biosystems) and Maxima SYBR green quantitative PCR master mix (Fermentas, Burlington, ON, Canada), respectively. Relative gene expression was analyzed by SDS software (Applied Biosystems) using comparative CT and hypoxanthine guanine phosphoribosyl transferase (Hprt, a housekeeping gene) expression as an endogenous control. In order to calculate the delta cycle threshold (ΔCT) value, the mean CT value of the expression of a gene in a sample was first normalized to the mean CT value of Hprt expression in that sample. The ΔCT value of the calibrator sample was subtracted from that of the sample of interest to obtain the ΔΔCT value. The relative expression was reported as 2–ΔΔCT. The data presented were obtained from 3 individual mice of each genotype. The primer sequences are presented in Supplemental Table 2.
Serum and tissue biochemistry
Serum calcium and Pi levels were measured using commercially available kits (Diagnostic Chemicals Limited, Charlottetown, PEI, Canada). To measure tissue ALPL activity, aortic protein extracts were prepared with 1× Passive Lysis Buffer (Promega, WI, USA) and total proteins were measured by the Pierce Coomassie Plus Protein Assay (Thermo Scientific). ALPL activity was measured using para-nitrophenylphosphate (ρ-nitrophenyl phosphate) substrate (Sigma-Aldrich) and then normalized by the respective protein concentration in the extracts as described. For collagen and elastin analysis, specimens were cut into small pieces and digested with high-purity bacterial collagenase (C0773; Sigma, Germany; 1 U/mL, 37°C, 12 hours). After centrifugation, the soluble fraction containing collagen was subjected to hydrolysis and amino acid analysis. The residual fraction was extracted by hot alkali (0.1 N NaOH, 95°C, 45 minutes). After centrifugation the supernatant containing noncollagenous/non-elastin proteins (NENCs) and the insoluble residue containing insoluble elastin were subjected to hydrolysis. Hydrolysis was performed in 6 N HCl at 110°C for 24 hours. Dried hydrolyzates were redissolved in sodium citrate loading buffer (pH 2.2) and amino acid analysis was performed by ion exchange chromatography with postcolumn derivatization with ninhydrin (Biochrom 30; Biochrom, United Kingdom). The amino acid profile of the supernatant after collagenase digestion showed a profile typical for collagen, whereas the supernatants after hot alkali treatment were free of hydroxyproline. The amino acid profile of the insoluble fraction after hot alkali extraction was typical for elastin. The content of collagen, elastin, and of non-elastin/noncollagenous protein was expressed per total protein (mg/mg).
Thoracic vertebrae together with the descending aorta were dissected and fixed in 2% formalin and 0.2% glutaraldehyde in PBS containing 5 mM EGTA and 2 mM MgCl2 for 5 minutes. Fixed tissues were rinsed three times for 5 minutes in rinse buffer (PBS containing 2 mM MgCl2 and 0.2% IGEPAL CA 630) and stained overnight at 37°C in the same buffer supplemented with 5 mM of each of K3Fe(CN)6 and K4Fe(CN)6.3H2O and 25mg/mL X-gal.
Arterial tissue extracts were separated on 10% polyacrylamide resolving gel containing 1.5% gelatin. Electrophoresis was performed at 100 V and the gels were incubated at 37°C in incubation buffer (50 mM CaCl2, 0.5 M Tris-HCl pH 7.6). Gels were stained with Coomassie blue and destained in 10% methanol and 20% glacial acetic acid.
All results are shown as standard deviation of the mean. Statistical analyses were performed by Student's t test or one-way ANOVA followed by Bonferroni post hoc test, with p < 0.05 considered significant as indicated by a single asterisk and p < 0.001 by double asterisks.
No upregulation of chondrogenic/osteogenic markers in Mgp–/– aortas prior to the initiation of calcification
We first examined the initiation and progression of arterial calcification in young MGP-deficient mice. µCT analysis of Mgp–/– thoracic aortas showed that ectopic calcification did not occur until at least 6 days after birth. Thereafter, sparse mineral deposits were visible in the aorta of 9-day-old MGP-deficient mice. By the 14th day, the thoracic aortas appeared to be fully mineralized and within 28 days mineralized intercostal arteries were clearly visible in the µCT images (Fig. 1A).
As is the case in some human pathologic conditions, arterial calcification in MGP-deficient mice can be caused by chondrogenic/osteogenic differentiation of VSMCs. In order to investigate this, we performed qRT-PCR expression analysis of several major chondrocyte and osteoblast marker genes in unmineralized aorta samples from 6-day-old Mgp–/– mice. Interestingly, we did not observe any significant upregulation of Runx2, encoding an early transcription factor, essential for chondrogenic/osteogenic commitment of mesenchymal stem cells (Fig. 1B). Also, the expression of chondrogenic markers SRY-box containing gene 9 (Sox9), collagen type II alpha 1 (Col2a1), aggrecan (Agc1), and collagen type X alpha 1 (Col10a1), and osteogenic markers osterix (Osx), collagen type I alpha 1 (Col1a1), alkaline phosphatase liver/bone/kidney (Alpl), and bone gla protein/osteocalcin (Bglap) were not upregulated in Mgp–/– aorta samples when compared to those from the WT control mice (Fig. 1B). We next examined the expression of these marker genes in fully calcified aortas of 14-day-old Mgp–/– mice. A significant upregulation of Sox9 and Agc1 was observed, although the expression of all other marker genes remained largely unaffected (Fig. 1C).
We then followed a transgenic approach to determine whether there is a chondrogenic commitment of VSMCs in Mgp–/– aortas. We first crossed Mgp+/– mice with Col2a1-lacZ mice expressing bacterial β-galactosidase under the control of a Col2a1 promoter fragment highly expressed in proliferating chondrocytes. The resultant Mgp+/–;Col2a1-lacZ mice were backcrossed with Mgp+/– mice to generate Mgp–/–;Col2a1-lacZ mice. We analyzed the thoracic vertebrae together with the descending aortas of these mice at 14 days of age, first by X-gal and then by Alizarin red staining. As expected, the cartilaginous tissues in the growth plates of the vertebrae were positive for the characteristic blue stain generated by the breakdown of X-gal by Col2a1-driven β-galactosidase activity. However, we did not observe any detectable presence of this blue stain in the aortas of Mgp–/–;Col2a1-lacZ mice, further supporting our qRT-PCR data that the Col2a1 promoter was not active in the arterial tissues of 14-day-old Mgp–/– mice. Alizarin red stain confirmed that the arterial tissues were fully mineralized in Mgp–/–;Col2a1-lacZ mice (Fig. 2A).
MGP has been suggested to be a negative regulator of BMP-mediated signaling in VSMCs. We investigated the expression of inhibitor of DNA binding 1 (Id1), a direct target of BMP signaling, in the thoracic aortas of 14-day-old Mgp–/– mice. We observed that the expression of this gene is actually downregulated in the absence of MGP (Fig. 2B). We next examined MGP-deficient aortas for the activity of a BMP/WNT-inducible enzyme, ALPL, a marker for mineralizing hypertrophic chondrocytes in the growth plate cartilage and osteoblasts in bone. As shown in Fig. 2C, we did not detect any increase of ALPL enzymatic activity in the aortas in the absence of MGP. To further investigate any involvement of ALPL during the initiation of calcification in Mgp–/– aortas, we generated Mgp–/–;Alpl–/– mice. Sera from these mice did not show any detectable ALPL activity as indicated by their inability to cleave ρ-nitrophenylphosphate in an in vitro enzymatic assay (Fig. 2D). When examined by skeletal preparations and Alizarin red staining, the calcified aortas of Mgp–/–;Alpl–/– mice stained comparably to that of control Mgp–/–;Alpl+/– mice (Fig. 2E).
Finally, in order to investigate whether MGP-deficient calcified arteries deposit cartilaginous matrix at a later age, we stained the thoracic aorta sections from 14- and 28-day-old Mgp+/– and Mgp–/– mice with Safranin-O (stains proteoglycans). Considering that Mgp+/– mice are phenotypically indistinguishable from WT mice, we used them interchangeably as controls in our experiments. As shown in Fig. 2F and G, Safranin-O staining was undetected in 14-day-old aorta sections, although the vertebral cartilage sections on the same slide (positive control) showed characteristic bright red staining. In agreement with the published data showing the presence of cartilaginous matrix in the calcified Mgp–/– aortas, we detected faint Safranin-O stain in the media of the 28-day-old Mgp–/– mice (Fig. 2G).
Mgp–/– aortas show altered ECM remodeling
Hart's histochemical staining (stains elastin) of the thoracic aorta sections prepared from 10-day-old nonperfused Mgp–/– mice showed a complete loss of the characteristic waviness of the elastic laminae (Fig. 3A). Histological analysis by light microscopy also showed areas with disrupted elastic laminae and an overall widening of the walls of the calcified aorta. We performed von Kossa and van Gieson staining of the consecutive arterial sections and observed that the initial mineral deposition was localized only along the elastic laminae (Fig. 3B). This observation was further supported by transmission electron microscopy of the ultrathin sections of arterial samples from 10-day-old Mgp–/– mice, which showed the presence of needle-like mineral crystals within the elastic lamina (Fig. 3C and C1).
Our light microscopy data (Fig. 3A, B) suggested an altered cellular phenotype in the arteries of MGP-deficient mice. In order to investigate this further, we performed qRT-PCR to compare the expression of several key VSMC markers in the arterial tissues from 6-day-old Mgp–/– and Mgp+/– mice. As shown in Fig. 4A, actin alpha 2 (Acta2) and collagen type III alpha 1 (Col3a1) expression were downregulated. However, we did not observe any detectable change in fibulin 1 (Fbln1), fibulin 2 (Fbln2), fibulin 5 (Fbln5), fibrillin 1 (Fbn1), and thrombospondin 1 (Thbs1) expression in MGP-deficient arterial walls prior to the initiation of calcification. In addition to that, there was a significant upregulation of Eln in the Mgp–/– aorta (Fig. 4A). We reproduced this interesting finding in a separate qRT-PCR experiment using cDNAs prepared from four different arterial samples from 6-day-old Mgp–/– mice and their WT littermates (Fig. 4B). Additionally, immunohistochemistry using a polyclonal antibody against insoluble elastin showed more intense staining in frozen aorta sections from 9-day-old Mgp–/– mice in comparison to those from their WT littermates (Supplemental Fig. 1).
We next performed gelatin zymography of arterial protein extracts that were prepared from 7-, 10-, and 21-day-old Mgp–/– and control WT mice. Although there was no detectable alteration of gelatinase activity in the extracts prepared from 7-day-old Mgp–/– mice, we detected a remarkable increase of gelatinase activity in the extracts prepared from 10- and 21-day-old mutants (Fig. 4C). Our subsequent gene expression analysis by qRT-PCR showed that the expression of Mmp3, Mmp9, and Mmp13 were upregulated in the arterial samples from 10-day-old Mgp–/– mice (Fig. 4D). Additionally by 21 days of age, we observed an upregulation of Mmp2 in MGP-deficient arteries (Fig. 4E).
The irregular calcification of the thoracic and abdominal aorta segments in Mgp–/– mice correlates with the local variations of the elastin content
Our µCT analysis showed differential amounts of deposited minerals in the thoracic and abdominal segments of Mgp–/– aortas (Fig. 5A). Although arterial tissue and lumen volumes were comparable in these two areas, the mineral deposition was at least 50% lower in the abdominal aorta (Fig. 5B). In order to examine whether there is a correlation between arterial elastin content and the amount of minerals deposited, we compared the amount of elastin, collagen, and NENCs in the thoracic and abdominal aorta segments of WT mice. We observed that the elastin levels were at least 2.5-fold lower in the abdominal aorta in comparison to the thoracic aorta (Fig. 5C). Interestingly, the total amount of collagen was significantly higher in the abdominal segment in comparison to the thoracic segment of the aorta, whereas NENC protein levels were comparable in both segments (Fig. 5C). We next examined the expression of several VSMC markers in the thoracic and abdominal segments of calcified aortas of 14-day-old Mgp–/– mice. Our qRT-PCR analysis showed differential expression of Acta2 (encodes alpha smooth muscle actin) and Eln in the thoracic and abdominal aorta—Acta2 expression was significantly higher and Eln expression was significantly lower in the abdominal aorta in comparison to the thoracic aorta. We did not detect any alterations of Col3a1 or Col1a1 expression in these two segments (Fig. 5D). Finally, we compared the expression of progressive ankylosis gene (Ank) and ectonucleotide pyrophosphatase/phosphodiesterase 1 (Enpp1) in the thoracic and abdominal aorta segments of Mgp–/– mice. These two genes are involved in the maintenance of extracellular levels of PPi, a potent mineralization inhibitor. We observed lower expression of both Ank and Enpp1 in the abdominal aorta of Mgp–/– mice, suggesting that increased PPi level is not the cause of reduced mineral deposition in this arterial segment (Fig. 5E).
Decrease in Eln gene dosage impedes the progression of arterial calcification in Mgp–/– mice
We next performed a genetic experiment to investigate the role of elastin in ectopic calcification in MGP-deficient arteries. It was shown earlier that the complete ablation of elastin in Eln–/– mice leads to perinatal death. However, despite an approximately 50% reduction of the tissue elastin content, Eln+/– mice have a normal lifespan. We first mated Mgp+/– mice with Eln+/– mice to generate Mgp+/–;Eln+/– mice, which were then inbred to generate Mgp–/–;Eln+/– mice. We compared the progression of arterial calcification in Mgp–/– and littermate Mgp–/–;Eln+/– mice by µCT at 14 and 28 days after birth. As shown in Fig. 6A, by 14 days Mgp–/– aortas were almost completely mineralized, whereas Mgp–/–;Eln+/– aortas only had a sporadic deposition of minerals. Intercostal arteries were clearly mineralized in Mgp–/– mice by 28 days of age. However, Mgp–/–;Eln+/– arteries showed significantly reduced mineral deposition; no mineralized intercostal arteries were evident at this time point. These observations were further confirmed by von Kossa and van Gieson staining of the thoracic aorta sections from both Mgp–/– and Mgp–/–;Eln+/– mice (Fig. 6B). We did not observe any significant difference of VSMC markers in the arteries of 14-day-old Mgp–/– and Mgp–/–;Eln+/– mice (Fig. 6C). Interestingly, however, Acta2 and Eln expression were significantly decreased in the arteries of both Mgp–/– and Mgp–/–;Eln+/– mice in comparison to WT mice. The expression of Col3a1 and Col1a1 were not significantly altered in the arteries of these mice in comparison to WT mice. We next measured Pi, calcium, and ALPL levels in the serum of Mgp+/–, Mgp–/–, and Mgp–/–;Eln+/– mice. These known determinants of ECM mineralization were not altered in the two latter genotypes (Fig. 6D). We then examined whether the ablation of another elastic lamina-associated protein could reduce mineral deposition in Mgp–/– mice. For this control experiment, we generated Mgp–/–;Thbs1–/–mice lacking both MGP and thrombospondin 1. As shown in Fig. 6E, the absence of thrombospondin 1 failed to recapitulate the reduced arterial calcification seen in Mgp–/–;Eln+/– mice. The delay of arterial calcification in Mgp–/–;Eln+/– mice significantly prolonged their lifespan, with one Mgp–/–;Eln+/– mouse surviving up to 7 months and siring two litters (Fig. 6F).
MGP deficiency in humans leads to Keutel syndrome, a rare autosomal recessive genetic disorder hallmarked by severe soft tissue calcification.[17, 18] Several recent clinical studies identified altered MGP levels as a risk factor for vascular calcification and further reinforced the importance of MGP in the prevention of ectopic calcification in vertebrate soft tissues.[31, 32] In transgenic mice, Mgp misexpression in osteoblasts results in severe bone mineralization defects. Although this study shows that gamma carboxylation of glutamic acid residues in MGP is critical for its antimineralization function, a complete understanding of its mechanism of action in vascular tissues is still lacking.
MGP has been identified as a negative regulator of BMP, an early signaling molecule necessary for skeletal development.[22, 33, 34] It has been proposed that in the absence of MGP, an augmented BMP signaling leads to chondrogenic/osteogenic transdifferentiation of the VSMCs in the blood vessel wall. Emerging data, however, suggest that this may not be the main reason behind the vascular calcification in MGP-deficient mice. Using cultured rat aortas treated with warfarin that prevents vitamin K-dependent gamma carboxylation of gla proteins, Lomashvili and colleagues showed that BMP-2 inhibitor noggin did not prevent arterial calcification in this tissue culture system. This finding suggests that MGP prevents arterial calcification independent of its proposed inhibitory effect on BMP-2–driven osteogenesis. Our data presented here provides further in vivo evidence supporting this inference.
VSMCs and cells in the mineralized tissues, eg, chondrocytes and osteoblasts, share a common developmental precursor—the mesenchymal stem cell. This common developmental origin may facilitate the transdifferentiation of VSMCs into chondrocyte/osteoblast-like cells in response to a variety of stimuli, such as increased serum levels of calcium and Pi, lipid accumulation, and inflammatory responses, causing the induction of pro-chondrogenic/-osteogenic signaling pathways.[34, 36] Indeed, transdifferentiation of VSMCs into osteoblast-like cell types has been reported in the mineralized tunica media of dialysis patients.
In order to investigate whether chondrogenic/osteogenic transdifferentiation of VSMCs is a prerequisite for medial calcification in Mgp–/– mice, we first monitored the expression of key chondrogenic/osteogenic markers in the arteries before and after the initiation of mineral deposition. We did not detect any upregulation of the expression of major chondrogenic/osteogenic markers in Mgp–/– aortas prior to the initiation of medial calcification. To further investigate the possible chondrogenic transdifferentiation of MGP-deficient VSMCs, we designed an in vivo reporter gene assay using the Mgp–/–;Col2a1-lacZ mice. This set of experiments stemmed from the reasoning that in the event that MGP deficiency leads to chondrogenic differentiation of VSMCs, it should also induce β-galactosidase activity driven by chondrocyte-specific Col2a1 promoter in these cells. However, there was no detectable presence of β-galactosidase activity in the calcified arteries of 14-day-old Mgp–/–;Col2a1-lacZ mice. Taken together, these findings raise the possibility that the initiation of vascular calcification in Mgp–/– mice does not require any pro-chondrogenic/-osteogenic alteration of the VSMC phenotype.
Luo and colleagues showed a late-stage appearance of chondrocyte-like cells in the heavily calcified Mgp–/– aorta. In agreement with this finding, we also observed a weak but distinct Safranin-O staining in the arterial media of 28-day-old “knockout”, suggesting the presence of proteoglycans in the fully calcified vascular ECM. However, we believe that this late-stage alteration of ECM does not influence the initiation of medial calcification because there is a considerable delay between the first detection of mineral crystals along the elastic laminae and the appearance of cartilaginous matrix in the MGP-deficient arteries.
We have reported that forced Alpl expression in the arterial ECM led to ectopic calcification of blood vessel walls in Col1a2-Alpl transgenic mice. Alpl expression, a marker of hypertrophic chondrocyte and osteoblast differentiation, is induced by BMP signaling. In agreement with our gene expression data, we did not detect any increase of ALPL activity in the arteries of Mgp–/– mice. Additionally, the absence of ALPL activity in Mgp–/–;Alpl–/– compound mutants did not prevent arterial calcification.
Taken together, our data from different lines of experiments unambiguously indicate that chondrogenic/osteogenic differentiation of VSMCs is not a prerequisite for the elastic lamina calcification in Mgp–/– mice. A possible explanation for the discrepancies between our findings and that of the initial study showing MGP as a regulatory protein for BMP-2 could be that the former used a cell culture model whereas we used in vivo models for our experiments. In a later study, Yao and colleagues showed increased SMAD phosphorylation as an indication of upregulated BMP signaling in the Mgp–/– aorta. However, the structural changes in the histology sections presented in their work clearly shows that these analyses were performed in the calcified aorta; therefore, an indirect stimulatory effect of calcification on SMAD phosphorylation cannot be fully ruled out.
Our alternative hypothesis suggests that in the absence of MGP, an alteration of vascular ECM, particularly the elastin-rich elastic lamina, may act as a catalyst for the initiation of ectopic calcification. In fact, the importance of mineral scaffolding type I collagen ECM in bone mineralization has been shown by several in vitro and in vivo experiments. For example, deficiency of ATF4, a CREB family transcription factor, results in reduced type I collagen synthesis by osteoblasts, which in turn inevitably leads to a reduction of mineralized bone mass in Atf4–/– mice. This later observation led us to examine the effects of reduced elastin content on the amount of minerals deposited within and alongside the arterial elastic laminae.
The role of elastin, the most abundant elastic lamina protein, as a mineral nucleator in the vasculature is supported by its degradation during the initial steps of medial calcification in humans. Also, in calcified blood vessels mineral deposits are commonly found on the elastic laminae in humans and in animal models including warfarin-treated rats.[40, 41] Our data shows that vascular mineral deposition in Mgp–/– mice is more acute in the thoracic aorta than in the abdominal aorta. In agreement with our hypothesis, this irregular mineral deposition correlates with the differential elastin levels in different arterial segments. Finally, we show that a reduced Eln gene dosage in Mgp−/−;Eln+/− mice significantly delays the initiation and progression of vascular calcification. This genetic experiment further validates our working hypothesis.
It is possible that MGP, through a direct interaction, protects mineral nucleating sites on elastin and prevents spontaneous calcification of the elastic laminae. Alternatively, the loss of MGP-elastin interaction may lead to an abnormal assembly and/or remodeling of the vascular ECM, which may in turn initiate ectopic calcification. We observed an upregulation of Eln expression in MGP-deficient aortas prior to the initiation of calcification. Interestingly, however, Eln expression was significantly decreased in the calcified aortas. This downregulation of Eln expression could be due to the phenotypic changes of Mgp–/– VSMCs in response to the arterial mineral deposition. Further evidence supporting the alterations of ECM properties in Mgp–/– arteries came from our observation that several MMPs are upregulated in the aortas of 10-day-old Mgp–/– mice. The upregulation of MMP expression coincides with the initiation of elastic laminae calcification. At this point it is not clear whether the initiation of calcification induces MMP expression in Mgp–/– aortas or the latter event induces the former. It is likely that the alterations of matrix remodeling by MMPs further contribute to the progression of mineral deposition in the vascular walls.
In the current study, we provide genetic proofs that the initiation of arterial calcification in Mgp–/– mice does not require an upregulation of Alpl or other chondrogenic/osteogenic markers. We show that the elastin content in the arterial walls acts as a critical determinant for the observed medial calcification in Mgp–/– mice. Further work is needed to determine how MGP deficiency affects the assembly, remodeling and the physical properties of the elastic laminae to cause medial elastocalcinosis.
All authors state that they have no conflicts of interest.
This work was supported by an operating grant from Canadian Institutes of Health Research (CIHR; Funding Reference Number 123310) to MM. HV acknowledges financial support from the Natural Sciences and Engineering Research Council (NSERC) of Canada. E.C.D. is a Canada Research Chair and MM receives salary support from the Fonds de la Recherche en Santé du Québec. The authors thank Hazem Eimar for help on statistical analyses, Garthiga Manickam and Marijana Kanisek for critical reading of the manuscript and Mia Esser for the maintenance of mouse colonies.
Authors' roles: Study design: MM. Study conduct, data collection and analysis: HR, ZK, JL, SL, HV, JB, ECD, and MM. Data interpretation: HV, JB, ECD, and MM. Drafting manuscript: ZK and MM. Revising manuscript content: HV, JB, ECD, and MM. Approval of the final version of the manuscript: HR, ZK, JL, SL, HV, JB, ECD, and MM. MM takes responsibility for the integrity of the data analysis.