S100A12 in Vascular Smooth Muscle Accelerates Vascular Calcification in Apolipoprotein E–Null Mice by Activating an Osteogenic Gene Regulatory Program

Abstract

Objective

The proinflammatory cytokine S100A12 is associated with coronary atherosclerotic plaque rupture. We previously generated transgenic mice with vascular smooth muscle–targeted expression of human S100A12 and found that these mice developed aortic aneurysmal dilation of the thoracic aorta. In the current study, we tested the hypothesis that S100A12 expressed in vascular smooth muscle in atherosclerosis-prone apolipoprotein E (ApoE)–null mice would accelerate atherosclerosis.

Methods and Results

ApoE-null mice with or without the S100A12 transgene were analyzed. We found a 1.4-fold increase in atherosclerotic plaque size and more specifically a large increase in calcified plaque area (45% versus 7% of innominate artery plaques and 18% versus 10% of aortic root plaques) in S100A12/ApoE-null mice compared with wild-type/ApoE-null littermates. Expression of bone morphogenic protein and other osteoblastic genes was increased in aorta and cultured vascular smooth muscle, and importantly, these changes in gene expression preceded the development of vascular calcification in S100A12/ApoE-null mice. Accelerated atherosclerosis and vascular calcification were mediated, at least in part, by oxidative stress because inhibition of NADPH oxidase attenuated S100A12-mediated osteogenesis in cultured vascular smooth muscle cells. S100A12 transgenic mice in the wild-type background (ApoE+/+) showed minimal vascular calcification, suggesting that S100A12 requires a proinflammatory/proatherosclerotic environment to induce osteoblastic differentiation and vascular calcification.

Conclusion

Vascular smooth muscle S100A12 accelerates atherosclerosis and augments atherosclerosis-triggered osteogenesis, reminiscent of features associated with plaque instability.

Vascular calcification is a characteristic feature of atherosclerosis and is a predictor of cardiovascular events. Atherosclerotic plaques with nodules of calcium deposition or microcalcification of the thin fibrous cap are more susceptible to rupture. Plaque rupture causes acute vascular events, such as acute myocardial infarction, or, if clinically unnoticed, may lead to further plaque remodeling and progression.¹ Although intravascular calcification has been viewed as a passive degenerative process of calcium deposition in necrotic vascular cells, more recent studies have shown evidence of a highly regulated process involving a phenotypic switch of vascular cells into osteoblastic-like-cells. There is growing support that vascular smooth muscle cells (VSMC) and myofibroblast-like cells, because of their unique phenotypic plasticity, can respond to various stimuli by promoting expression of bone regulating proteins, including alkaline phosphatase, osteocalcin, osteopontin, bone morphogenic proteins, and collagen types I and II.^2,3

S100A12 (also known as EN-RAGE, calgranulin C) and other members of the S100/calgranulin family of calcium binding proteins, such as S100A8 (MRP8, calgranulin A) and S100A9 (MRP14, calgranulin B), are implicated in the regulation of a variety of intracellular and extracellular activities.⁴ We recently reported that S100A12 promotes a switch from contractile VSMC to VSMC that have reduced expression of contractile fibers, increased reactive oxygen species production, and increased interleukin-6 and transforming growth factor-β cytokine production.⁵ S100/calgranulins are endogenously expressed in granulocytes and myeloid cells and are not detectable in normal VSMC, but they are induced in VSMC in response to injury (such as endothelial cell wire injury⁶), in lipopolysaccharides,⁵ and in neovascular smooth muscle cell in the atherosclerotic vessel.⁷ Most importantly, Burke et al found strong expression of S100A12 in human coronary artery smooth muscle in ruptured plaques associated with sudden cardiac death, with the highest S100A12 expression observed in ruptured plaques of diabetic patients.⁸ These studies strongly suggest a relationship between the pathological expression of S100A12 in the vasculature and features of plaque instability.

We now investigated the role of VSMC-expressed human S100A12 in atherosclerotic prone milieu, the apolipoprotein E (ApoE)–null mouse. We exploited the fact that S100A12 is not present in mice⁹ and used the previously generated C57BL/6J mice with VSMC-targeted expression of human S100A12. The S100A12 transgenic mice were now back-crossed into ApoE-null mice, also from the C57BL/6J background. In the absence of a high-fat diet, the presence of human S100A12 produced profound remodeling and calcification of atherosclerotic plaques in the S100A12/ApoE-null mice. An increase in osteogenic gene expression was noted in VSMC from prepathogenic mice, and this accelerated atherosclerosis was at least in part mediated by oxidative stress.

Methods

An expanded Methods section is available in the supplemental materials, available online at http://atvb.ahajournals.org. Briefly, C57BL/6J mice hemizygous for human S100A12 expressed in VSMC driven by the SM22α promoter were previously described.⁵ Hemizygous S100A12/C57BL/6J mice were mated with ApoE-null mice on a C57BL/6 background (The Jackson Laboratory). F3 generation S100A12/ApoE-null and wild-type (WT)/ApoE-null littermates not expressing the transgene were used for all experiments. All mice were genotyped for S100A12 and ApoE. All mice were housed at all times in specific pathogen–free barrier facilities and maintained on normal rodent chow with free access to food and water. All procedures were carried out with the approval of the institutional animal care and use committee of the University of Chicago.

Results

ApoE-Null Mice That Express Human S100A12 in VSMC Have Increased Vascular Calcification

To establish the role of S100A12 for vascular remodeling, we assessed the impact of S100A12 on atherosclerotic lesion in ApoE-null mice fed standard rodent chow. Serial sections of the proximal ascending aorta and of the proximal aortic arch at the junction of the innominate artery were examined in 10-month-old S100A12/ApoE-null and age-matched WT/ApoE-null littermate mice. We found that S100A12/ApoE-null mice showed a 1.4-fold increase in plaque area in the proximal ascending aorta and a 1.5-fold increase in plaque area in the innominate artery (Table). Remarkably, despite this rather small difference in overall plaque size between the 2 groups of mice, the atherosclerotic plaques in the S100A12/ApoE-null mice had markedly increased calcification on staining with alizarin red S, a stain for the presence of calcific deposition. In the S100A12/ApoE-null mice, we found that 45% of the innominate artery plaques and 18% of the aortic root plaques were calcified, compared with 7% and 10% in the WT ApoE-null littermate, respectively (P<0.05). Figure 1A shows a typical lesion in the innominate artery with extensive calcification and large necrotic zones. In contrast, we found minimal alizarin red–stained vasculature in age-and gender-matched WT/ApoE-null littermates, despite the presence of well-developed advanced atherosclerotic plaques with large lipid cores (Figure 1A, d to f). Furthermore, we noticed enhanced calcification of the medial layer most prominently seen in the aortic arch of S100A12/ApoE-null mice, even in the absence of large atherosclerotic lesions (representative example is shown in Figure 1B). This demonstrates that forced VSMC-targeted expression of S100A12 promotes calcification independent of atherosclerotic lesion size.

Table.

Compositional Analysis of Atherosclerotic Plaque in S100A12/ApoE-Null Mice and WT Littermate Mice

Aged 10 Months	Aortic Root		Innominate Artery
	WT/ApoE−/−	S100A12/ApoE−/−	WT/ApoE−/−	S100A12/ApoE−/−
Plaque area (×103μm2/section)	198.5±19	278.2±22.4* (1.4-fold increase)	125.7±21.4	188.8±18.0* (1.5-fold increase)
Calcified area, %	9.8±4.2	18.1±5.2†	7.3±4.4	45.4±6.4†
Macrophage, %	12.1±6.9	16.4±5.9	25.6±5.7	20.8±6.4
Medial expansion, μm	652±321	1403±453*	1543±654	3166±976*
Elastin degradation	Grade 1.5	Grade 3.4*	Grade 1.4	Grade 3.6*

Data are expressed as mean±SEM.

^*P<0.05 and

^†P<0.01 compared with WT.

Figure 1.

Advanced calcification within the atherosclerotic lesions in the innominate artery (A) and medial calcification in the aortic arch (B and C) in 10-month-old S100A12/ApoE-null mice. A, Alizarin red stain shows large calcified nodules extending from the lumen (L) to internal elastic lamina in S100A12/ApoE mice (a) but not in advanced lesions of WT/ApoE-null littermates (d). Fibrofatty nodules (FFN) with cholesterol clefts are seen in both groups. S100A12/ApoE-null lesions showed layering of the lesions with presence of chondrocyte-like cells (marked *** in b and inset) and extensive erosion of the media (marked # # # in c) on Verhoeff-Van Giesen (VVG) stain. B, Alizarin red stain shows extensive medial calcification in the aortic arch of S100A12/ApoE-null mice, compared with calcifications located within the neointimal hyperplasia of a large plaque in aortic arch of WT/ApoE-null mice (C). With heavy vascular calcification, there was increased fragility of the vasculature in S100A12/ApoE-null mice during tissue processing, leading to artifacts (as seen in B).

S100A12/ApoE-null mice also had extensive degradation of elastic fibers in the innominate artery plaques (grades 3.6 and 1.4, respectively; P<0.01) and in the aortic root (grades 3.4 and 1.5, respectively; P<0.01) compared with WT/ApoE littermate mice (Table). This extensive erosion of the underlying media involved elastic fiber degradation and outward remodeling of the plaque (Figure 1C). Moreover, we found outward remodeling of the vasculature and a significant increase in circumference of the external elastic membrane in the innominate and aortic root in S100A12/ApoE-null mice (Table). These histopathologic data were also corroborated in vivo by high-frequency ultrasound. The in vivo vascular diameter was increased 17% at the level of the proximal ascending aorta, 20% at the distal ascending aorta, and 27% at the origin of the innominate artery in S100A12/ApoE-null mice compared with age- and gender-matched WT/ApoE littermate mice (P<0.05; Supplemental Figure I). To link S100A12 expression with accelerated vascular calcification and plaque remodeling, we examined transgene expression in the aorta of S100A12/ApoE-null and WT/ApoE-null mice. S100A12 was detected by immunofluorescence microscopy in the VSMC-rich medial layer of the S100A12 ApoE-null mice, but not in WT-ApoE-null mice (Supplemental Figure IIA). As expected, the expression of the S100A12 transgene was reduced in 8-month-old-mice compared with young S100A12/ApoE-null mice, because SM22α is a marker of smooth muscle cell maturation and differentiation and is known to be reduced in phenotypically modulated smooth muscle in atherosclerotic lesions of aged ApoE-null mice.¹⁰ This was further quantified by semiquantitative immunoblotting, and we found a significant reduction of S100A12 protein in the aorta of 8-month-old compared with 2-month-old S100A12/ApoE-null mice (Supplemental Figure IIB).

The features of calcified plaques did not arise from enhanced lipid levels because there were no significant differences in cholesterol and triglyceride levels between the 2 groups. Therefore, the increase in plaque size and changes in plaque composition did not derive from increased lipid levels (Supplemental Table I). In summary, these data show that forced expression of human S100A12 in VSMC promotes atherosclerotic plaque remodeling and nodular calcification in hyperlipedemic and proatherosclerotic environments, such as those present in ApoE-null mice.

ApoE-Null Mice That Express S100A12 in VSMC Have Increased Gene Expression of Osteogenesis Regulating Genes in the Aorta

Diseases that have been linked to an increase of vascular calcification, such as hyperlipidemia, hypertension, diabetes mellitus, and dialysis-dependent end-stage renal disease, are associated with elevated oxidative stress (reviewed in Shao et al¹¹) and with increased serum S100A12 concentration.^12,13 It has been shown that oxidative stress and inflammation contributes to a phenotypic switch of VSMC with increased ossification that results in vascular calcification.^2,3,14–16 To test the hypothesis that S100A12 participates in vascular ossification characterized as increased expression of multiple osteogenesis-related genes, we examined the aorta of young mice to identify early changes preceding the pathogenic state of vascular calcification. In 3- to 4-month-old mice, we found well-developed complex atherosclerotic plaques, and on staining with alizarin red, we found patchy mural calcium deposition in S100A12/ApoE-null but not in WT/ApoE-null littermate (Figure 2A). At that age, there was no difference in atherosclerotic plaque area in S100A12/ApoE-null compared with WT/ApoE littermate mice at the aortic root (82.467±32.432 μm² and 79.378±34.875 μm², P=0.1) and at the innominate artery (16.986±12.482 μm² and 18.391±14.948 μm², P=0.4). Hence, an analysis of aortic gene expression at this age would detect genes over- or underrepresented that are important for the initiation of ossification and would not represent established calcification. Total RNA was prepared from pooled aorta samples (n=3 per group) and, after transcribing to cDNA, subjected to an osteogenesis RT² Profiler polymerase chain reaction array. Many genes were differentially expressed (data not shown), and we validated the results for Runx-2, bone morphogenic protein 2 (BMP-2), dentin matrix acidic phosphoprotein 1 (Dmp-1), and bone gla protein (BGLAP, also known as osteocalcin) using quantitative reverse transcription–polymerase chain reaction in independently derived total RNA. We found a 1.5-fold increase for Runx-2 and a 2.4- to 2.7-fold increase in gene expression of Bmp-2, Dmp-1, and BGLAP in the aorta of S100A12/ApoE-null mice compared with WT littermate ApoE-null mice, suggesting that S100A12 induces osteoblastic genes in the vasculature of atherosclerosisprone ApoE-null mice. Most interestingly, S100A12 transgenic mice in a WT background, lacking the proinflammatory/proatherosclerosis environment (S100A12/ApoE^+/+/C57BL6/J) showed no increase in gene expression of Runx-2, Bmp-2, Dmp-1 and BGLAP and were similar to WT littermate mice (Figure 2B). These S100A12/ApoE^+/+/C57BL6/J mice developed aortic dilation, as previously noted.⁵ We reexamined aged S100A12/ApoE^+/+/C57BL6/J mice using alizarin red S stain. We found small calcific nodules in the aortic arch in 2 of 10 S100A12/ApoE^+/+/C57Bl/6 mice; these regions colocalized to areas of medial degeneration with loss of VSMC and increased fibrosis. The most severe case is shown in Supplemental Figure III. We did not observe any calcification in the aortic arch of age-matched WT/ApoE^+/+/C57BL6 littermates. Taken together, these data show that S100A12 expressed in VSMC leads to augmented vascular calcification in a proatherosclerotic and hyperlipedemic environment such as that present in ApoE-null mice. However, in mice lacking this environment, S100A12 alone is not sufficient to promote significant vascular calcification.

Figure 2.

A, Atherosclerotic lesion in the innominate artery of 3-month-old S100A12/ApoE-null mice (a to c) and of age-matched WT/ApoE-null littermates (d to f) shows atherosclerotic lesions in both groups and small areas of calcium deposition in S100A12/ApoE-null mice. B. Quantitative reverse transcription–polymerase chain reaction shows increased gene expression of Runx-2, BMP-2, Dmp-1, and BGLAP in aortae harvested from S100A12/ApoE-null mice. GAPDH expression was used as internal control. Data are expressed as mean±SEM (*P<0.001 and **P=0.05 versus WT).

S100A12 in VSMC Promotes Osteoblastic Gene Expression When Cultured in a Proinflammatory Environment

To evaluate whether the osteoblastic gene expression observed in aortic tissue of S100A12/ApoE-null is directly linked to VSMC undergoing a phenotypic conversion, we isolated VSMC from aortae of 8-week-old mice. VSMC that express S100A12 (harvested from either S100A12/ApoE-null or S100A12/ApoE^+/+ mice) showed no spontaneous calcification nodules on alizarin red stain on culture for 7 days in normal Dulbecco’s modified Eagle’s medium (DMEM) and had morphology similar to VSMC from WT-ApoE aorta. However, when cultured in conditioned media of primary WT-ApoE-null macrophages supplemented with hyperlipedemic serum from WT/ApoE-null mice (5%), we found that VSMC that express S100A12 formed 3 to 4 calcifying nodules per ×10 power field (P<0.01), compared with no significant formation of calcification nodules in VSMC lacking S100A12 expression (Figure 3A and 3B). Use of hyperlipedemic media alone was not sufficient to induce calcified nodules within 7 days (data not shown). Because S100A12 is released from S100A12-VSMC into the cell culture supernatant when exposed to a proinflammatory milieu such as stimulation with low-dose lipopolysaccharide⁵ or conditioned media (data not shown), and the receptor for advanced glycation endproducts (RAGE) is expressed in VSMC,¹⁷ we examined the effect of soluble RAGE, the extracellular-ligand binding domain, for its ability to prevent VSMC calcification in vitro. As shown in Figure 3C, pretreatment of S100A12-VSMC cultured in conditioned media with soluble RAGE attenuated calcification by 65%. We next examined the gene expression of the osteogenic factors BMP-2, BGLAP, dentin matrix acidic phosphoprotein 1 (Drp-1), and Runx-2 in cultured VSMC. A 2- to 2.5-fold increase was seen when S100A12-VSMC were cultured in conditioned media. In contrast, gene expression was comparable in S100A12-VSMC cultured in regular DMEM and WT VSMC (Figure 3D). Importantly, pretreatment with soluble RAGE significantly attenuated the gene expression of Runx-2, BMP-2, BLAP, and Drp-1. In summary, these data show that S100A12 promotes osteoblastic gene expression in VSMC only when exposed to an inflammatory environment. Moreover, this effect is greatly attenuated when S100A12 is quenched by soluble RAGE, preventing activation of cell surface receptors such as RAGE.

Figure 3.

S100A12 promotes calcification of VSMC cultured in conditioned media from primary mouse macrophages. A, Alizarin red S stain of VSMC harvested from the aortae of S100A12/ApoE-null mice and cultured in conditioned medium developed mineralized nodules (arrow) and cellular calcium deposits (arrowhead). B, Quantification of calcified nodules by counting 5 random ×10 power fields (*P<0.003 versus DMEM and WT). C, Total cell alizarin red staining (osteogenesis assay, Millipore) of VSMC cultured in conditioned media and treated with soluble RAGE (sRAGE) as indicated (**P<0.05 compared with bovine serum albumin [BSA]). D, Quantitative reverse transcription–polymerase chain reaction for gene expression of BMP-2, BGLAP, Dmp-1, and Runx-2 in S100A12 VSMC and WT VSMC that were cultured in DMEM or conditioned media and treated with sRAGE as indicated. GAPDH expression was used as internal control, and data are expressed as mean±SEM of fold increase compared with WT VSMC cultured in DMEM (***P<0.03 compared with baseline condition and with WT VSMC, **P=0.05 compared with S100A12-VSMC in conditioned media with BSA).

S100A12/ApoE-Null Mice Have Increased Oxidative Stress, and Inhibition of NADPH Oxidase Attenuates Calcification in Cultured VSMC

Increased oxidative stress is a major finding in advanced atherosclerosis and is one mechanism by which S100A12 might alter VSMC function. We have previously shown that S100A12 increases oxidative stress in VSMC,⁵ and S100 proteins have been linked to enhanced oxidative stress in various mammalian cell lines.¹⁸ Byon et al first demonstrated that H₂O₂ initiates calcification of VSMC dependent on oxidative stress–activated Runx-2.¹⁶ To establish a role for oxidative stress in the S100A12/ApoE-null mouse model of enhanced vascular calcification, we measured urine 8-isoprostane, which serves as a marker of lipid peroxidation and has been validated in several mouse models of oxidative stress. We found a marked increase in urinary 8-isoprostane in 4- and in 10-month-old S100A12/ApoE-null mice (2.3-and 3.7-fold, respectively; P<0.05; Figure 4A). Although clearly many factors are involved in the initiation and progression of atherosclerotic lesions, our findings suggest that at least one mechanism by which S100A12 augments atherosclerosis and vascular calcification is by promoting oxidative stress.

Figure 4.

Effect of S100A12 on oxidative stress and ossification. A, 8-Isoprostane in the urine of WT/ApoE-null and S100A12/ApoE-null at 4 and 10 months (M) of age. (*P<0.05, **P<0.001). B, Intracellular hydrogen peroxide levels measured by 2′7-dichlorofluorescence diacetate (DCFDA) microscopy in cultured VSMC (*P<0.05, **P<0.001). Cond. indicates conditioned. C, Effect of apocynin and DPI on calcifying nodule formation in cultured VSMC from S100A12/ApoE-null aorta. Calcification nodules were quantified after staining with alizarin red (*P=0.05). D, Effect of apocynin and DPI on gene expression of Runx-2, BMP-2, BGLAP, and dentin matrix acidic phosphoprotein 1 (Drp-1) in VSMC from S100A12/ApoE-null aorta cultured in conditioned media (*P<0.001 compared with control dimethyl sulfoxide [DMSO]–treated cells).

We then examined markers of oxidative stress in cultured VSMC. Similar to our previous findings of increased reactive oxygen species in VSMC harvested from aortae of S100A12/C57BL6/J/ApoE+/+,⁵ we found increased cellular H₂O₂ production in VSMC harvested from S100A12/ApoE-null mice compared with WT VSMC on fluorescence microscopy with 2′7-dichlorofluorescence diacetate dye. At baseline, S100A12-VSMC had a 1.24-fold increase in fluorescence intensity, and in the presence of conditioned media there was a 3.7-fold increase, compared with a 1.7-fold increase in WT VSMC cultured with conditioned media (P<0.01, Figure 4B).

Previously, it was shown that oxidative stress induced by S100 proteins in VSMC is mediated at least in part by the NAPDH oxidase system.¹⁷ We therefore sought to investigate whether inhibiting NADPH oxidase could attenuate the calcification we observed in cultured VSMC. VSMC pretreated with 1 μmol/L apocynin or 10 μmol/L diphenylene iodonium (DPI), 2 compounds known to inhibit NADPH oxidase activity in VSMC,^19,20 showed significant reduction in the formation of calcified nodules in S100A12-VSMC cultured in conditioned media for 7 days (Figure 4C). We next examined the effect of NADPH oxidase inhibitors on gene expression of Runx-2, BMP-2, BGLAP, and Drp-1. Both apocynin and DPI attenuated gene expression of all 4 osteogenic factors by 85% to 30% in S100A12-VSMC cultured for 1 to 7 days in conditioned media (Figure 4D). These findings support that S100A12-induced calcification in VSMC is mediated, at least in part, by oxidative stress involving NADPH oxidases.

S100A12 constitutes 2% to 5% of neutrophil cytosolic protein²¹ and has been linked to antimicrobial and antiparasitic activity.²² Because the respiratory burst in neutrophils is mediated by NADPH oxidases with gp91^phox(Nox2) as a main catalytic subunit, we queried whether S100A12 interacted with the smooth muscle gp91^phox homologs to mediate oxidative stress in VSMC. In VSMC, NADPH oxidase activity derives from gp91^phox homologs called Nox1 and Nox4.^23,24 VSMC were transiently transfected with a plasmid encoding S100A12 fused to a myc epitope, followed by stimulation with angiotensin II (Ang II). As shown in Figure 5A, Nox1 is upregulated in VSMC in response to Ang II. Using anti-myc magnetic Micro Beads, S100A12-myc was isolated from VSMC-membrane extracts as shown by immunoblotting with α-S100A12 IgG. Most importantly, Nox-1 was pulled down from in the Ang II–stimulated and S100A12-transfected VSMC (lane 2 in Figure 5B). The converse experiment, immunoprecipitation with an anti-Nox1 IgG, similarly yielded S100A12 protein (Figure 5C, left lane). These findings demonstrate that S100A12 complexes with Nox1 in VSMC.

Figure 5.

S100A12 associates with Nox1. A, Ang II stimulated Nox-1 expression in rat VSMC (A7r5) in cells transiently transfected with S100A12-myc and in control cells. B, Membrane extracts from S100A12-myc or mock-transfected rat VSMC stimulated with and without Ang II were immunoprecipitated with antimyc IgG. Immunoblotting shows S100A12 in transfected cells (two left lanes) and Nox1 in S100A12-transfected and Ang II–stimulated VSMC (second lane from left). C, Membrane extracts were immunoprecipitated with α-Nox1 IgG or nonimmune rabbit IgG, followed by immunoblotting for S100A12. IB indicates immunoblotting; IP, immunoprecipitation.

To evaluate whether increased S100A12 occurs in human atherosclerotic disease, we examined 5 aortic tissue blocks. The tissue was obtained from surgery for aortic aneurysms and was diagnosed as arterial wall with atherosclerosis. We found strong expression of S100A12 and nearby medial necrosis in all 5 cases. As shown in Supplemental Figure IV, S100A12 was expressed in VSMC, and also colocalized to cells that stained positive for myeloperoxidase, a marker for granulocytes, macrophages, and other inflammatory cells. In contrast, VSMC in normal aorta showed no S100A12 (Supplemental Figure IVG and IVH).

Discussion

Vascular calcification is a pathological condition that occurs in atherosclerosis and is particular accelerated in patients with diabetes and end-stage renal disease. Both conditions have been associated with increased cardiovascular events, as well as increased serum concentration of S100A12.^12,13 Specifically, S100A12 is expressed in VSMC of ruptured coronary artery plaques in victims of sudden death,⁸ and transcriptional profiling of platelet derived mRNA identified S100A8/9 as a regulator of atherothrombosis in patients presenting with ST elevation myocardial infarction.²⁵

S100/calgranulins activate RAGE,²⁶ RAGE promotes atherosclerosis in several animal models,^27–30 and ApoE-null mice that either lack RAGE,²⁷ or lack S100A9³¹ have attenuated atherosclerosis. Despite this large circumstantial evidence that S100/calgranulins promote vascular disease, direct evidence of S100A12 accelerating atherosclerosis has been missing. We found here that the combination of S100A12 and the ApoE-null background was sufficient to induce calcification conversion of the atherosclerotic plaques and medial calcification. This degree of calcific atherosclerosis exceeds what is reported for chow-fed ApoE-null mice of similar age³² and is similar to mouse models that use a combination of severe hyperlipidemia and hyperglycemia.³³ Our results demonstrate that expression of S100A12 in VSMC is pathological and, in the appropriate environment, can lead to plaque morphologies considered “vulnerable,” with increased elastic fiber breakdown, large necrotic cores, outward remodeling with increased elastic media expansion, and formation of calcified nodules.

One limitation of our model is the use of forced expression of S100A12 in smooth muscle, because under physiological conditions S100A12 is expressed in high levels mainly in myeloid cells and particular in neutrophilic granulocytes. However, it has been shown that S100A12 is induced in pathological conditions such as coronary artery plaque rupture,⁸ thoracic aneurysm associated with MYH11 mutation,⁵ and atherosclerotic aortic aneurysms (Supplemental Figure IV). A pathological role for S100A12 in VSMC is further supported by the findings by Johansson et al demonstrating the expression of the murine homolog mS100A9 in the disruption-prone shoulder region within atherosclerotic plaques in LDL receptor–deficient mice.³³ This suggests, indeed, that S100 proteins expressed in vascular cells other than the “physiological” S100A12-positive-myeloid cells are positioned to play an important role in atherosclerosis.

The aggressive vascular remodeling observed in the S100A12/ApoE-null mice was associated with increased urinary 8-isoprostane levels, as a marker of increased oxidative stress. Increased oxidative stress is a hallmark of advanced atherosclerosis and is multifactorial in its origin. We previously showed that S100A12 causes increased reactive oxygen species in VSMC, and we expanded these findings by demonstrating that S100A12-triggered reactive oxygen species production in VSMC is at least in part dependent on NADPH oxidases because inhibition of NADPH oxidase reduced the expression of osteoblastic genes in S100A12 expressing VSMC. Moreover, quenching of S100A12 using soluble RAGE, the extracellular-ligand binding domain, limited access of S100A12 to cell surface RAGE and attenuated mineralization in cultured VSMC. Our data strongly suggest that S100A12 increases the susceptibility of VSMC to dedifferentiate in favor of gaining an expression profile of genes characteristic of osteoblasts in a proinflammatory environment, such as those present in ApoE-null mice. The atherosclerotic and inflammatory environment is required for this process because we found only minimal vascular calcification in C57BL/6J mice that express S100A12 in VSMC but lack atherosclerosis. We speculate that S100A12 accelerates dysfunction in VSMC, particularly under conditions of increased Nox1 expression, such as stimulation by Ang II^17,24 or advanced glycation end products.¹⁹ We found that S100A12 associates with Nox1, although future studies are necessary to define the implication of this interaction. In addition, inhibition of NADPH oxidase in cell culture using apocynin or DPI markedly reduced reactive oxygen species production, osteoblastic gene expression, and the formation of calcified nodules in VSMC that express S100A12. Of note, we did not observe expression of Nox-2 in cultured VSMC harvested from WT or S100A12/ApoE-null aortae (data not shown), although Nox-2 has recently been implicated in superoxide production and atherosclerotic plaque formation in ApoE-null mice.³⁴

S100A12 is actively secreted from neutrophilic granulocytes on activation of protein kinase C,³⁵ and serum concentrations of S100A12 are often increased in inflammatory diseases associated with neutrophil activation. Therefore, previous studies linking increased S100A12 serum concentration to vascular disease are unable to establish a direct and causal role for S100A12. Goyette et al recently found increased serum S100A12 from patients with coronary artery disease undergoing coronary angiography, and most importantly, a significant gradient across the coronary circulation when blood was simultaneously sampled in the coronary sinus and in the aorta.³⁶ This suggests that S100A12 elevation is specific to the coronary vasculature in patients with acute coronary syndrome. However, the same group could not find any release of proinflammatory cytokines or expression of matrix metalloproteases in human peripheral blood mono-nuclear cells or human monocyte–derived macrophages in response to recombinant S100A12 protein.³⁶ This may be related to a nonphysiological structure of the recombinant S100A12 protein, whose formation and biological function is greatly affected by zinc and calcium concentration and by the oxidation state.³⁷ Alternatively, S100A12 may have a greater impact on vascular cells other than macrophages, such as VSMC or endothelial cells. Our data support the notion that S100A12 affects function and biology of VSMC and promotes oxidative stress and expression of osteoblastic genes in an appropriate, proatherosclerotic, and hyperlipedemic environment.