Inflammation: a culprit for vascular calcification in atherosclerosis and diabetes

Abstract

It is today acknowledged that aging is associated with a low-grade chronic inflammatory status, and that inflammation exacerbates age-related diseases such as osteoporosis, Alzheimer’s disease, atherosclerosis and type 2 diabetes mellitus (T2DM). Vascular calcification is a complication that also occurs during aging, in particular in association with atherosclerosis and T2DM. Recent studies provided compelling evidence that vascular calcification is associated with inflammatory status and is enhanced by inflammatory cytokines. In the present review, we propose on one hand to highlight the most important and recent findings on the cellular and molecular mechanisms of vascular inflammation in atherosclerosis and T2DM. On the other hand, we will present the effects of inflammatory mediators on the trans-differentiation of vascular smooth muscle cell and on the deposition of crystals. Since vascular calcification significantly impacts morbidity and mortality in affected individuals, a better understanding of its induction and development will pave the way to develop new therapeutic strategies.

Introduction

Paradoxically, whereas aging provokes a drop in bone formation [1], it is associated with vascular calcification (VC) [2, 3]. For instance, it was reported that post-menopausal women with the highest degree of aortic calcification had the lowest bone mineral density [4]. More generally, independent studies showed that in both men and women older than 50, the progression of aortic calcification was positively associated with the rate of decline in bone mineral density and with osteoporotic fractures [5]. One hypothesis to explain this relationship is that the two processes share a common etiology. As will be discussed in the last chapter of this manuscript, a potential shared etiology is inflammation [6].

Vascular calcification accompanies at least two common age-related diseases: atherosclerosis and type 2 diabetes mellitus (T2DM). During atherosclerosis, calcification develops in the arterial intima, whereas medial calcification is a hallmark of T2DM. In both diseases, vascular smooth muscle cells (VSMCs) play an important role in the formation of crystals. VSMCs indeed form in large arteries a bone-like tissue, which can result from endomembranous or endochondral ossification [7, 8], and eventually lead to a mature tissue, containing bone marrow [9]. As will be discussed in depth in the next chapters, arterial calcification is clearly associated with the mortality risk in individuals with atherosclerosis and diabetes. For instance, a large population-based cohort aged 30–89 years followed for more than 30 years showed a strong relationship between increasing severity of aortic arch calcification in women and men in middle age and risk of death [2]. In this study, aortic arch calcification was present in 1.9 % of men and 2.6 % of women, and was independently associated with older age. After adjustment for age, aortic arch calcification was associated with an increased risk of coronary heart disease in men and women, and with a 1.46-fold increased risk of ischemic stroke among women [2]. In this worrying context, deciphering the mechanisms involved in the initiation and development of calcification in these diseases appears crucial.

Many excellent reviews have been published on the mechanisms of VC. Our aim here is to focus on the role of age-related inflammation in the initiation of this process. Indeed, aging is linked to a chronic, low grade, pro-inflammatory state. This association between aging and inflammation has been referred to as “inflammaging” [10]. Inflammaging has been demonstrated in elderly individuals by the increase in the circulating levels of C-reactive protein (CRP) [11] and of cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-6 [12, 13]. Inflammaging may be conditioned in each individual by life exposition to various stimulators of inflammation, including infections and trauma [14], by psychological stress during aging or even prenatal and early life [15, 16], or by fat accumulation during aging [17–19]. A lack of clear resolution may explain that chronic inflammation accompanies aging. Although inflammatory markers are only increased 2- to 4-fold and thus far from the increases of acute inflammation [20], inflammaging is considered as the most common and important driving force of age-related pathologies, such as atherosclerosis, T2DM and osteoporosis [21]. Inflammation is likely to play a significant role in the development of VC. In a cohort free of clinically apparent cardiovascular disease, CRP levels were associated with coronary artery calcification in both men and women [22]. In addition, several recent convincing articles have demonstrated the role of inflammatory mediators in VC in vitro and in vivo. In this article, after a brief resume of the cellular and molecular mechanisms of ossification, we will focus on VC, beginning with atherosclerosis plaque calcification, and following with that of T2DM. We will deliberately not mention vascular calcification associated with chronic kidney disease (CKD). Indeed, although inflammation likely plays a role in the development of calcification in CKD patients, calcification is also strongly linked to dysregulated mineral metabolism characterized by long term elevation of serum phosphate levels as well as transient bouts of hypercalcemia, which is not the case in atherosclerosis and T2DM [23].

Endochondral and intramembranous ossification

During bone development and repair, two types of ossification can be encountered. In long bones, endochondral ossification involves the chondrocyte-mediated formation of a cartilaginous calcified template that is subsequently replaced by bone [7]. In contrast, flat bones develop through membranous ossification, which directly relies on the activity of osteoblasts.

Chondrocyte differentiation during endochondral ossification

The transcription factor SOX9 appears to play crucial roles in mesenchymal cell condensations and chondrocyte differentiation. Inactivation of SOX9 in limb buds using the Cre/loxP recombination system before the appearance of chondrogenic mesenchymal condensations results in the complete absence of condensations and subsequent cartilage formation [24]. In addition, SOX9 drives the production of an extracellular matrix (ECM) enriched in collagen type II, IX and XI, and glycosaminoglycans (Fig. 1) [7]. During chondrogenesis, in absence of growth plate vascularization, proliferative and differentiated chondrocytes experience hypoxia [25]. Hypoxia activates hypoxia inducible factor 1α (HIF-1α), which is involved in several processes. Firstly, HIF-1α is necessary for chondrocyte survival during hypoxia [25]. HIF-1α is also required for the expression of chondrocyte markers such as type II collagen [26]. Furthermore, HIF-1α induces the expression of vascular endothelial growth factor, which in turn, activates metaphyseal vascular invasion [25]. Following early chondrocyte differentiation, hypertrophic differentiation takes place under the control of the transcription factor RUNX2. RUNX2 expression increases with maturation of chondrocytes, which is severely disturbed in RUNX2-deficient mice [27]. Forced expression of RUNX2 in immature chondrocytes induces type X collagen and matrix metalloprotease-13 expression, tissue-nonspecific alkaline phosphatase (TNAP) activity and extensive matrix mineralization (Fig. 1) [28]. Eventually, mineralized cartilage vascularization leads to cartilage resorption and new bone apposition.

Fig. 1.

Transcription factors and markers involved in osteoblast or chondrocyte differentiation from mesenchymal progenitors. DMP dentin matrix protein, HIF hypoxia inducible factor, MMP matrix metalloprotease, TNAP tissue nonspecific alkaline phosphatase

Osteoblast differentiation in endomembranous ossification

In endomembranous ossification, osteoblasts differentiate from mesenchymal progenitors under the control of RUNX2, which notably stimulates the expression of a type I collagen rich matrix (Fig. 1) [29]. Then, osteoblasts express TNAP, which, like in cartilage, induces collagen mineralization [30, 31]. Finally, more mature osteoblasts secrete osteocalcin, a hormone regulating insulin levels and sensitivity [32]. A weak proportion of osteoblasts eventually become surrounded by the matrix they have calcified and give rise to osteocytes, the bone-resident cells. Among other markers, osteocytes secrete sclerostin, an inhibitor of several bone anabolic Wnt growth factors, and dentin matrix protein-1, a molecule controlling phosphatemia [33].

Molecular mechanisms of mineralization

Initiation of mineralization is likely governed by cell-derived matrix vesicles (MVs) [34, 35]. In the cartilage growth plate, their diameter ranges from 30 nm to 1 μm, with an average of approximately 200 nm [34]. It is believed that MVs initiate mineralization in cooperation with collagen fibrils, and that once the first calcium phosphate crystals have coalesced to form the mineralization front, the next steps occur essentially by crystal multiplication, independently of MVs. The membrane of MVs is markedly enriched with glycosylphosphatidylinositol-anchored TNAP compared with the cell membrane from which they originate [36]. In humans, TNAP is probably the protein that plays the most important role in inducing mineralization. Loss-of-function mutations in the gene encoding TNAP leading to hypophosphatasia were first documented in 1988 [37], and since then about 260 mutations have been reported [38]. The most severe form of hypophosphatasia manifests in utero with dramatic hypomineralization, and causes death at, or soon after, birth. It has been shown that TNAP stimulates crystal formation by hydrolyzing the mineralization inhibitor inorganic pyrophosphate (PP_i) (Fig. 2) [30, 31, 39]. PP_i is produced in the ECM by ankylosis protein homolog (ANKH) and ectonucleotide pyrophosphatase phosphodiesterase 1 (ENPP1). ANKH is a transmembrane protein that exports PP_i. Mutations in ANKH result in two distinct diseases associated with excessive mineralization: craniometaphyseal dysplasia [40, 41] and familial calcium pyrophosphate dihydrate deposition disease [42, 43]. ENPP1 is a membranous protein that hydrolyses ATP in the ECM to generate PP_i. Mutations in the ENPP1 gene are responsible for generalized arterial calcification in infancy (GACI) or idiopathic infantile arterial calcification [44, 45]. Calcification of large and medium-sized arteries and stenosis due to myointimal proliferation are significant features of the GACI phenotype and most affected children die in early infancy [46]. Causes of death are myocardial infarction, congestive heart failure, persistent arterial hypertension, or multi-organ failure [47]. In cartilage and bone, crystal formation resulting from decreased PP_i levels occurs in association with collagen fibrils and consist in a non-stoichiometric, calcium-deficient, carbonated apatite, whose composition vary with the age of crystals. In particular, aging of crystals is characterized by loss of acidic phosphate groups and incorporation of carbonate ions [48].

Fig. 2.

Induction of tissue mineralization by matrix vesicles. ANKH ankylosis protein homolog, ENPP1 ectonucleotide pyrophosphatase phosphodiesterase 1, PP _i inorganic pyrophosphate, TNAP tissue nonspecific alkaline phosphatase

Inflammation and vascular calcification in atherosclerosis

Atherosclerosis is an age-related disease. In humans, the early atherosclerotic lesions can usually be found in the aorta in the first decade of life. They consist of subendothelial accumulations of cholesterol-engorged macrophages, called “foam cells” (type I lesions) [49]. Then “fatty streak” lesions (type II lesions) are observed in the coronary arteries in the second decade and in the cerebral arteries in the third or fourth decades. Fatty streaks are not clinically significant, but are the precursors of intermediate lesions characterized by the accumulation of lipid-rich necrotic debris and VSMCs (type III lesions). Advanced atherosclerotic lesions can also be subdivided into three main histologically characteristic types: IV, V, and VI [50]. In type IV lesions, a dense accumulation of extracellular lipid occupies an extensive but well-defined region of the intima, enclosed by a “fibrous cap” consisting of VSMCs and ECM. This type of extracellular lipid accumulation is known as the lipid core. Type V lesions are defined as lesions in which prominent new fibrous connective tissue has formed. Morbidity and mortality from atherosclerosis is largely due to type IV and type V lesions in which disruptions of the lesion surface, hematoma or hemorrhage, and thrombotic deposits have developed [50]. Type IV or V lesions with one or more of these additional features are classified as type VI and may also be referred to as complicated lesions.

Calcification is a hallmark of atherosclerosis [9]. Coronary arterial calcification is indeed part of the development of atherosclerosis, occurs almost exclusively in atherosclerotic arteries, and is absent in the normal vessel wall [50, 51]. It has been estimated that over 70 % atherosclerotic plaques observed in the aging population are calcified [52]. Early calcifications are usually noted in stage III specimens, with intermediate and solid calcifications becoming increasingly prominent within advanced plaques, especially stages V and VI [53, 54]. The prevalent sites of calcification are located in the deeper regions of the intima and the atheroma [53].

The role of coronary calcification in plaque rupture has long been controversial [55–57]. Not all individuals with high coronary calcification develop myocardial infarction, and some cases of myocardial infarction occur in the absence of high levels of coronary arterial calcification [58]. Although there is a positive correlation between the site and the amount of coronary artery calcium and the percent of coronary luminal narrowing at the same anatomic site, the relation is nonlinear and has large confidence limits [59]. It has been proposed that atherosclerotic plaque proceeds through progressive stages where instability and rupture can be followed by calcification, perhaps to provide stability to an unstable lesion [60]. That is, patients who have calcified plaque are also more likely to have non-calcified or “soft” plaque that is prone to rupture and acute coronary thrombosis [61]. On the other hand, a recently growing body of evidence indicates that the direct assessment of coronary artery calcium can provide independent and incremental prognostic information over and above the traditional Framingham risk score. A number of recent publications have indeed reported on the incremental prognostic value of coronary artery calcification (CAC) in large series of patients including asymptomatic self-referred and population cohorts [62–66]. The American College of Cardiology Foundation/American Heart Association expert consensus proposed that measurement of CAC is reasonable for cardiovascular risk assessment in asymptomatic adults at intermediate risk (10–20 % 10-year risk), and low-to-intermediate risk (6–10 % 10-year risk) [67]. Early detection of CAC will allow to initiate lipid-lowering therapy and lifestyle modifications [68].

Actually, recent data indicate that two types of calcification may be distinguished in the atherosclerotic plaque. Microcalcifications may form first, consisting of crystals measuring less than 50 µm in diameter, often as small as a single calcified macrophage, and invisible in micro-computed tomography [69]. These microcalcifications have been mentioned in several intravascular imaging studies as a coronary calcification pattern that is extremely difficult to detect. Interestingly, macrophages have been reported to release mineralization-competent structures resembling to MVs and exosomes [70]. It is therefore possible that some microcrystals may precipitate independently on VSMCs, in association with inflammatory cells recruited in the intima. It has been proposed that the rupture of a plaque may be triggered by the explosive growth of small voids in the tissue in the vicinity of closely clustered microcalcifications [71]. In this model, the microcalcifications per se would not be dangerous, in contrary to their spacing and location in the cap relative to the location of the minimum cap thickness [71]. Vulnerable plaques therefore tend to be those with less extensive calcium deposits frequently seen in a spotty distribution [72], a finding supported by intravascular ultrasound studies of patients with acute coronary syndromes [73].

Subsequently to the formation of microcrystals, a propagation phase follows, due to the phenotypic conversion of VSMCs leading to ossification [74]. Indeed, human calcified plaques have been found to express mRNA of several markers of chondrocytes and/or osteoblasts such as RUNX2, TNAP and osteocalcin [75]. Similarly, in the apolipoprotein E-deficient (ApoE) ^−/− mouse model of atherosclerosis, calcified cartilage is formed in advanced plaques [76, 77]. In these mice, VSMCs are likely responsible for calcification since VSMC-specific RUNX2 deficiency prevents plaque calcification [78]. VSMCs are characterized by phenotypic plasticity that allows them to de-differentiate into mesenchymal precursors and/or trans-differentiate into chondrocytes. Interestingly, SM22 ^−/− mice develop medial chondrogenesis after carotid denudation with a fall in VSMC markers such as myocardin transcription factor and α-smooth muscle actin and concomitant enhanced expression of SOX9 and type II collagen [79]. In atherosclerotic plaques, VSMC-derived chondrocytes appear to release MVs [80] as do normal mineralizing chondrocytes. Their diameter ranges from 100 to 700 nm, and is therefore roughly the same as that of chondrocytes-derived MVs [34]. They are also enriched in TNAP [80]. Finally, like in growth plate cartilage, the crystals formed in human atherosclerotic plaques consist of a carbonated apatite [81].

Both microcalcifications and cartilage-like tissue formation may be triggered by inflammatory signals. In humans, focal arterial inflammation, as quantified by ¹⁸F-fluorodeoxyglucose/positron emission tomography (PET), was suggested to precede calcification within the same locations [82]. In this study, it is however, possible that microcalcifications were present earlier but were not detected. ¹⁸F-fluoride PET recently allowed the identification of macrophage infiltration and active microcalcification at the sites of plaque rupture [83]. Microcalcification may be induced by inflammation in more than one way, as discussed below. Additionally, osteogenesis in atherosclerotic plaques was also reported to associate with inflammation in ApoE ^−/− mice as revealed by in vivo imaging [84].

Role of tumor necrosis factor-α

Inflammatory mechanisms play a central role in mediating all phases of atherosclerosis, from initial recruitment of circulating leukocytes to the arterial wall to eventual rupture of the unstable plaque [85]. Many cytokines likely participate in atherosclerosis, including TNF-α, IL-1α and IL-1β, IL-4, interferon-γ (IFN-γ) [86]. Among these cytokines, TNF-α might be particularly important. TNF-α is predominantly expressed by macrophages in the plaque and mice deficient in both ApoE and TNF-α exhibit a 50 % reduction in aorta lesion size after 10 weeks of Western-style diet feeding [87, 88]. Among the multiple effects exerted by TNF-α at different atherosclerosis stages, which are likely to impact calcification indirectly, direct effects of TNF-α on crystal formation seem to occur. TNF-α indeed activates VSMCs to form calcium deposits (Table 1) [89–97]. TNF-α may exert its effects by stimulating the release of bone morphogenetic protein (BMP)-2, a potent bone anabolic factor [98]. TNF-α may also enhance BMP-2 activity by reducing the levels of its inhibitor, matrix Gla protein (MGP), in VSMCs [98, 99]. In parallel, TNF-α activates NF-κB to decrease expression of ANKH and reduce PP_i export [100]. Moreover, TNF-α decreases extracellular PP_i levels through TNAP activation [89]. Collectively, these effects of TNF-α likely contribute to activate VSMC trans-differentiation and trigger mineralization of the ECM (Fig. 3).

Table 1.

Chronological presentation of the most important reports of the effects of inflammatory cytokines on VSMC trans-differentiation and calcification in vitro

Cytokine	Cell source	Effect	References
TNF-α	Bovine VSMCs	The first report that TNF-α stimulates TNAP and calcium deposition in VSMCs. Mechanisms appear to involve PKA	[90]
TNF-α and OSM	Human VSMCs	THP-1 macrophages activate VSMCs to produce TNAP and mineralize through the secretion of TNF-α and OSM	[91]
IL-4	Human VSMCs	Stimulates calcium deposition through RUNX2	[97]
TNF-α	Mouse aortic myofibroblasts	Stimulates TNAP through canonical Wnt signaling	[154]
TNF-α	Rat VSMCs	Increases Msx2, RUNX2 and BMP-2 expression Stimulates TNAP and calcium deposition	[92]
TNF-α	Human VSMCs	Increases Msx2, RUNX2, osterix and BSP expression Stimulates TNAP and calcium deposition through NF-κB and Msx2	[93]
TNF-α and IL-1β	Human VSMCs	Stimulates TNAP through PPARγ inhibition	[89]
TNF-α	Human VSMCs	Increases calcium deposition through NF-κB-mediated decrease in ANKH expression and PPi export	[100]
TNF-α	Mouse aortic myofibroblasts	Increases calcium deposition through TNFR1, reactive oxygen species and Msx2	[95]
TNF-α	Rat aortic rings and hVSMCs	Increases BMP-2 and calcium deposition	[94]
TNF-α and IL-6	Mouse VSMCs	TNF-α decreases MGP expression TNF-α increases RUNX2, TNAP and calcium deposits IL-6 amplifies TNF-α effects on calcium deposition	[99]
TNF-α	Mouse MOVAS	Stimulates calcium deposition through PKA and ATF4	[96]
Inflammasome	Mouse VSMCs	Inflammasome inhibition prevents calcium deposition	[120]

Fig. 3.

Induction and pro-calcifying effects of inflammatory cytokines in atherosclerosis and type 2 diabetes mellitus (T2DM). Arrows references and words in brown concern atherosclerosis; arrows, references and words in blue concern more specifically T2DM, and arrows, references and words appearing in black concern cytokine effects on VSMC “trans-differentiation” and mineralization, which are likely to concern both atherosclerosis and T2DM

Role of IL-1 and the inflammasome

IL-1α and IL-1β are expressed in human atherosclerotic plaques, particularly by macrophages and endothelial cells [101]. In mouse, IL-1 receptor I deficiency [102], as well as overexpression of the IL-1 receptor antagonist IL-1Ra [103, 104], lead to less atherosclerosis. Conversely, IL-1Ra-deficient mice display exacerbated atherosclerosis [104, 105], and spontaneous aortitis [106, 107]. Several molecules have been reported to activate IL-1 release, through or independently from inflammasome activation. The inflammasome is a multiprotein complex that promotes IL-1β secretion after its cleavage by caspase-1. Cholesterol crystals, which form early during atherogenesis [108], stimulate macrophages to release IL-1β through inflammasome activation [108, 109], and IL-1α independently of it [110]. Fatty acids also likely contribute to generate vascular inflammation. In ApoE ^−/− mice fed a high-cholesterol diet, four fatty acids (palmitic acid, stearic acid, oleic acid and arachidonic acid) represented two-thirds of the fatty acid pool found in atherosclerotic plaques [111]. Of these, oleic acid induced IL-1α production independently from inflammasome activation [111]. Another report showed that palmitic acid triggers the release of IL-1β through AMP kinase inhibition and inflammasome activation [112]. Sphingolipids are also important inflammatory players in atherosclerosis that might increase IL-1 secretion [113]. Enhanced uptake of sphingolipid rich lipoproteins has been proposed, but increased local sphingolipid synthesis in the atherosclerotic plaque has also been observed and considered as a contributing mechanism [113]. Degradation of sphingomyelin on the LDL particle surface can promote aggregation by ceramide–ceramide interactions, thus promoting initiation of atherosclerosis [114, 115]. In ApoE-knockout mice, deficiency of acid sphingomyelinase, which generates ceramide from sphingomyelin, decreases early foam cell lesion areas by up to 50 % [114]. Interestingly, ceramide in particular activates the inflammasome to release IL-1β [116]. The inflammasome seems therefore to be a central component in the inflammation induced by cholesterol crystals, fatty acids and sphingolipids. Although like TNF-α, IL-1β likely exacerbates the whole development of atherosclerotic plaques and not specifically calcification, it may also directly stimulate calcification in VSMCs as TNF-α does. Indeed, in vitro studies reported that IL-1β stimulates TNAP activity and calcification in VSMCs [89, 117]. In addition, several inflammasome activators have been shown to enhance calcification. For instance, palmitic acid induces calcification in human VSMCs [118, 119]. Proposed mechanisms include reactive oxygen species generation [119], and the stimulation of BMP-2 expression [118]. Finally, it was recently reported that inhibition of inflammasome prevents calcification in cultured VSMCs [120].

Inflammatory cytokines are induced by apoptosis and necrosis

The presence of apoptotic macrophages and VSMCs in atherosclerotic plaques has been confirmed by a number of studies [121]. Apoptotic indices are low in early lesions, but increase in frequency as lesions develop, in both the necrotic core and fibrous cap. Macrophage and VSMC apoptosis is induced by oxidized LDL [122], or oxysterols [123]. Interestingly, whereas in vascular aging VSMC apoptosis doesn’t induce inflammation, it does in atherosclerosis. This is because the clearance of apoptotic bodies is impaired in atherosclerotic plaques [124], where apoptotic cells are subject to secondary necrosis [125]. Necrotic cells release the alarmin IL-1α, which likely contributes to initiate calcification. Alternatively, uncleared apoptotic bodies themselves may represent a nidus for calcification, with charged phosphatidylserine-bearing membranes able to nucleate crystals [126]. Moreover, whereas living cells carefully separate calcium (in the sarcoplasm and the mitochondria) from P_i (in the cytoplasm), both ions may be incorporated into apoptotic bodies, and contribute to the formation of crystals. In cultures of VSMCs [126], and interestingly also in cultures of chondrocytes [127], inhibition of apoptosis by the broad-range caspase inhibitor ZVAD-FMK reduces the extent of mineralization. Since this inhibitor blocks caspase-1, it would be interesting to determine whether it acts through inhibition of apoptosis and/or that of the inflammasome [120]. Finally, crystals may activate a vicious cycle associating apoptosis and inflammation. Indeed, crystals induce apoptosis in VSMCs, after phagocytosis and dissolution in lysosomes [128], and activate macrophages to secrete more TNF-α and IL-1β [129].

Inflammatory cytokines are induced by VSMC senescence

During aging, cell senescence is likely to induce VMSC trans-differentiation and calcification. Replicative senescence of VMSCs enhances their trans-differentiation and calcification [130]. Senescence can result from lamin A accumulation in the nucleus due to abnormal maturation; and it was proposed that lamin A interferes with DNA damage repair signaling [131]. The most sever form of laminopathy due to mutations in the gene encoding lamin A is the Hutchinson-Gilford progeria syndrome (HPGS) [132]. Patients with HPGS suffer premature aging, including premature atherosclerosis and calcification [133–135], leading to myocardial infarction or stroke at an age of 13 years [132, 136–138]. Prelamin A accumulates in calcifying VSMCs in vivo and its overexpression in vitro promotes trans-differentiation and calcification [139]. Liu et al. [139] reported that prelamin A accumulation induces DNA damage, that in turn triggers the secretion of pro-mineralizing factors. Interestingly, DNA damage induces the secretion of IL-6, whereas inhibition of the DNA damage response inhibits IL-6 secretion by VSMCs, suggesting that senescence exerts paracrine pro-inflammatory effects. Finally, the factors that drive prelamin A accumulation in aging VSMCs remain to be identified.

Inflammation and vascular calcification associated with type 2 diabetes mellitus

Like atherosclerosis, T2DM is an age-related disease with a chronic pro-inflammatory state [140]. In T2DM, chronic elevation of circulating nutrients such as glucose and free fatty acids contributes to the induction of inflammatory processes observed within various tissues, including pancreatic islets and the vascular wall [140, 141]. T2DM is associated with calcification of the media [9]. In 4553 subjects studied in a 20-year longitudinal study, calcification was shown to first appear in the feet and develop proximally [142]. In this study, diabetic patients with medial arterial calcification, compared with diabetic patients without medial arterial calcification, had 5.5-fold the rate of amputations [142]. Moreover, in another study of 1059 patients with T2DM, femoral arterial calcification was also found to predict lower limb amputation [143]. Even more worrying, compared to the general population, patients with diabetes tend to have more atherosclerotic plaques on coronary computed tomography angiography than patients without diabetes, notably more calcified plaques [144, 145], and are significantly more likely to develop coronary heart disease [142, 146, 147]. In adults with diabetes, measurement of CAC has recently been considered helpful for cardiovascular risk assessment in asymptomatic adults with diabetes [67].

Like in atherosclerotic plaques, it would seem that VC associated with T2DM recapitulates endochondral ossification. Studies of human medial calcification in T2DM and aging showed that the VSMCs lose the expression of calcification inhibitors, such as MGP, and begin to express differentiation markers of chondrocytes such as type II collagen [148]. Moreover, histological examination of femoral arteries from two patients with long-term T2DM revealed the presence of type II collagen in foci of cartilaginous metaplasia, providing evidence for endochondral ossification [149]. In addition, in low-density lipoprotein receptor mutant (Ldlr ^−/−) mice, T2DM accelerated cartilage formation and calcification in the aorta [150].

Role of TNF-α and IL-1β in medial calcification in T2DM

Cytokine stimulation of VSMCs probably plays an important role in calcification associated with T2DM. It is well-known that in humans, glucose induces the secretion of several inflammatory mediators such as TNF-α and IL-6 [151]. Hyperglycemia is likely to induce vascular inflammation. Indeed, in mouse, overexpression of GLUT1 in VSMC induces arterial wall inflammation [152]. In human and rodent VSMCs also, glucose induces the secretion of TNF-α [153]. As we mentioned above, TNF-α is able to induce calcification in cultures of VSMC [89, 90], and in vivo, the TNF-α inhibitor infliximab has been shown to prevent medial calcification in ldlr ^−/− diabetic mice, without reducing obesity, hypercholesterolemia and hyperglycaemia [154]. However, the fact that only 30 % of calcium deposits were prevented in the latter study suggests that other inflammatory factors are involved in the pro-calcifying effects of glucose. IL-1β may be one of these inflammatory factors, although in vivo studies have only reported yet that IL-1β worsens T2DM in general but not vascular calcification specifically. Evidence of its importance in the progression of metabolic disorders comes from the encouraging results of a phase 3 clinical trial with T2DM patients treated with the IL-1Ra anakinra [155]. Several articles have highlighted the role of IL-1β in the development of insulin resistance. Mice deficient in NLRP3 (NOD-like receptor family, pyrin domain containing 3), which lack inflammasome-dependent release of IL-1β show improved glucose tolerance and insulin sensitivity [156, 157]. Accordingly, caspase-1 inhibition increases insulin sensitivity [158]. Activators of the inflammasome-mediated release of IL-1β in T2DM include glycemia, ceramides, and islet amyloid polypeptide [159]. Glucose itself stimulates the release of IL-1β from pancreatic β-cells [160]. Recently, glucose has been shown to enhance the levels of thioredoxin-interacting protein, a protein that interacts with the NLRP3 inflammasome to trigger IL-1β secretion [156].

Glucose treatment of VSMCs has been shown to accelerate calcification, and up-regulate expression of osteoblast/chondrocyte markers (Fig. 3). For instance, bovine VSMCs incubated with high levels of glucose expressed more RUNX2 and osteocalcin, displayed higher alkaline phosphatase activity and mineralization capacity compared to cells cultured in normal glucose levels [161]. In addition, in culture of human aortic smooth muscle cell, glucose increased the expression of BMP-2 and RUNX2 [162]. Furthermore, in rat VSMCs also, high glucose levels increased alkaline phosphatase activity, mineralization and osteocalcin expression [163, 164]. Finally, medial calcification appears to relate to the degree of glycemic control [165]. Collectively, these data showing that (1) glucose induces inflammation in VSMCs, (2) activates the inflammasome-dependent release of IL-1β, (3) IL-1β stimulates VSMC to calcify, and (4) inflammasome inhibition prevents calcification suggest that glucose may stimulate calcification in part through IL-1β activation.

Can we reconcile the opposite role of inflammation on vascular calcification and bone formation?

Inflammation may thus be a leading cause of VC [166]. Both CRP and inflammatory cytokines are associated with increased CAC [22, 166, 167]. On the opposite, inflammation is a well-recognized inhibitors of bone formation [1]. For instance, using a cohort of 168 randomly selected men and women with a mean age of 63 years, Ding and collaborators reported that higher levels of TNF-α, IL-6 and CRP are associated with increased bone loss over 3 years of follow-up [168]. An inverse association between CRP levels and bone mineral density has also been observed in a sample of 2807 elderly females [169]. Moreover, in healthy individuals above 40, CRP levels represent a strong risk predictor of nontraumatic fractures, and its predictive significance extend to levels commonly regarded as low grade inflammation [170]. In the Geelong Osteoporosis study, the fracture risk was increased 24–32 % for each SD increase in CRP levels in elderly women [171]. Finally, in healthy individuals over 70, Cauley and collaborators have shown that in addition to CRP levels, high serum levels of inflammatory markers IL-6, TNF-α, and TNF receptors, can also predict a higher incidence of non-traumatic fractures [172]. These effects of inflammation are probably due to increased bone resorption and decreased bone formation [1]. In cultured osteoblasts, TNF-α and IL-1β inhibit the expression of collagen type I, which is expressed early during differentiation, and the secretion of osteocalcin, a marker of mature cells [173]. The inhibition of collagen may account for a great part in the inhibition of bone formation, since collagen is the most abundant protein in bone, where it represents the framework for crystal deposition. In osteoblasts, TNF-α inhibitory effects are due to decreased expression of RUNX2 [173], and increased Smurf1-mediated RUNX2 degradation [174]. The molecular effects of IL-1β on osteoblasts have been much less investigated but it seems that in human osteoblasts, IL-1β has the same effects on RUNX2 activity and on collagen and osteocalcin expression [173].

Concerning the effects of inflammation on endochondral ossification, it is well-known that children with chronic inflammatory diseases experience impaired longitudinal growth rate as compared to healthy ones. In cultured whole rat metatarsal bones, both IL-1β and TNF-α impair metatarsal longitudinal growth, decrease the proliferation of growth plate chondrocytes, and increase chondrocyte apoptosis [175]. In cultured chondrocytes, both cytokines decrease TNAP activity and mineralization [89]. Strikingly, whereas inflammasome may be involved in the induction of VC [120], its activation in vivo causes growth retardation and osteopenia [176].

Therefore, whereas TNF-α and IL-1β seem to accelerate VC in atherosclerosis and T2DM, they are potent inhibitors of osteoblast and chondrocyte differentiation and mineralization. These contradictory effects are also observed in ankylosing spondylitis (AS), an inflammatory disease that affects the axial skeleton and the peripheral joints, where excessive tendon and ligament ossification accompanies systemic bone loss [177]. In AS, these opposite patterns of ossification may be reconciled by the observation that ossification proceeds in locations where inflammation has resolved [178]. That resolution of inflammation is a prerequisite for ossification is also a condition for bone fracture healing. Indeed, a necessary early inflammation phase takes place during bone repair, whereas inflammation slows bone formation later on [179]. One can therefore hypothesize that inflammation plays an activating role in the initiation of ossification and VC by promoting the expression of osteogenic growth factors, and that these factors efficiently stimulate mineralization when and where inflammation has resolved. In vascular cells, TNF-α and/or IL-1β increase the levels of BMP-2 and osteogenic Wnt family members [98, 154]. Atherosclerosis and T2DM are chronic low-grade inflammatory diseases, during which periods of inflammation are separated by periods of resolution. It is therefore conceivable that active calcification proceeds when inflammation has resolved under the action of BMP and Wnt factors.

In conclusion, the data reviewed here indicate that atherosclerosis and T2DM, two age-related diseases with an inflammatory character, develop vascular calcification in response to inflammatory signals. Efforts to determine the nature and role of these signals in vascular calcification in vivo are therefore warranted.

Abbreviations

ANKH

Ankylosis protein homolog

BMP

Bone morphogenetic protein

CAC

Coronary artery calcification

CKD

Chronic kidney disease

CRP

C-reactive protein

ECM

Extracellular matrix

ENPP1

Ectonucleotide pyrophosphatase phosphodiesterase 1

FDG

Fluorodeoxyglucose

GACI

Generalized arterial calcification in infancy

HIF

Hypoxia inducible factor

HPGS

Hutchinson–Gilford progeria syndrome

IFN

Interferon

IL-1Ra

Interleukin-1 receptor antagonist

MGP

Matrix Gla protein

MMP

Matrix metalloprotease

MVs

Matrix vesicles

NLRP3

NOD-like receptor family, pyrin domain containing 3

PET

Positron emission tomography

PP_i

Inorganic pyrophosphate

T2DM

Type 2 diabetes mellitus

VSMCs

Vascular smooth muscle cells

TNAP

Tissue nonspecific alkaline phosphatase

TNF

Tumor necrosis factor