Canadian Journal of Cardiology 30 (2014) 982e993
Review
Basic Mechanisms of Calcific Aortic Valve Disease Patrick Mathieu, MD, and Marie-Chloe Boulanger, PhD Laboratoire d’Études Mole culaires des Valvulopathies (LEMV), Groupe de Recherche en Valvulopathies (GRV), Que bec Heart and Lung Institute/Research Center, Department of Surgery, Laval University, Que bec, Que bec, Canada
ABSTRACT
RESUM E
Calcific aortic valve disease (CAVD) is the most common heart valve disorder. There is no medical treatment to prevent and/or promote the regression of CAVD. Hence, it is of foremost importance to delineate and understand the key basic underlying mechanisms involved in CAVD. In the past decade our comprehension of the underpinning processes leading to CAVD has expanded at a fast pace. Hence, our understanding of the basic pathobiological processes implicated in CAVD might lead eventually to the development of novel pharmaceutical therapies for CAVD. In this review, we discuss molecular processes that are implicated in fibrosis and mineralization of the aortic valve. Specifically, we address the role of lipid retention, inflammation, phosphate signalling and osteogenic transition in the development of CAVD. Interplays between these different processes and the key regulation pathways are discussed along with their clinical relevance.
La calcification de la valve aortique (CVA) est le trouble des valves quent. Aucun traitement me dical ne peut cardiaques le plus fre venir ou favoriser la re gression de la CVA. Par conse quent, il est pre finir et de comprendre les principaux particulièrement important de de canismes de base sous-jacents qui sont implique s dans la CVA. Au me cennie, notre compre hension des fondements cours de la dernière de très rapidement. Par des processus menant à la CVA a progresse quent, notre compre hension des processus biopathologiques conse s dans la CVA mènerait e ventuellement à fondamentaux implique laboration de nouveaux traitements pharmacologiques de la CAV. l’e culaires qui sont Dans cette revue, nous discutons des processus mole s dans la fibrose et la mine ralisation de la valve aortique. implique tention lipidique, de Particulièrement, nous traitons du rôle de la re l’inflammation, de la signalisation du phosphate et de la transition ogène dans le de veloppement de la CVA. Nous discutons des inoste rents processus et les voies de re gulation teractions entre ces diffe principales ainsi que leur pertinence clinique.
Calcific aortic valve disease (CAVD) encompasses a wide spectrum of clinical entities, from aortic sclerosis to severe aortic stenosis (AS). However, there is a general agreement that aortic sclerosis and AS are part of a same pathobiological process whereby the aortic leaflets undergo a progressive mineralization and fibrotic process. Hence, a group of leading experts, convened by the National Heart, Lung and Blood Institute, have developed a consensus statement and have recommended that the different clinical entities, aortic sclerosis and AS, should be termed CAVD.1 For many years, CAVD has been described as a “passive” process related to aging. However, in the past 15 years there has been a growing interest to decipher the cellular and molecular processes involved in CAVD. Studies underscored that CAVD shared many clinical risk factors with atherosclerosis. In addition, histopathological examination of surgically explanted stenotic aortic valves indicated that several
features such as lipid infiltration, inflammation, and calcification were commonly observed in CAVD.2 Moreover, several animal models of atherosclerosis also developed some degree of CAVD.3 Taken together, these findings suggested that CAVD was only another manifestation of an atherosclerotic process. However, 3 randomized trials with statins failed to demonstrate any benefit from a lipid-lowering approach in patients with a moderate-to-severe AS.4 These studies suggested that although CAVD has possibly some overlapping processes in common with atherosclerosis it has underlying specific mechanisms that lead to the mineralization of the aortic valve. To this effect, a growing number of studies have since shed more light on the underpinning processes at play in CAVD, which might open new therapeutic avenues. In this article, we review the latest discoveries related to CAVD, its basic molecular processes, and whenever possible we have tried to tie these basic discoveries with clinical observations, keeping in mind a translational approach.
Received for publication January 7, 2014. Accepted March 19, 2014. Corresponding author: Dr Patrick Mathieu, Institut de Cardiologie et de Pneumologie de Quebec/Quebec Heart and Lung Institute, 2725 Chemin Ste-Foy, Quebec, Quebec G1V 4G5, Canada. Tel.: þ1-418-656-4717; fax: þ1-418-656-4707. E-mail: [email protected] See page 990 for disclosure information.
Histopathology of CAVD and Basic Mechanism of Mineralization CAVD is characterized by the presence of mineralized nodules and fibrosis.5 Histological sections of explanted CAVD during surgeries revealed that mineralization starts in
0828-282X/$ - see front matter 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cjca.2014.03.029
Mathieu and Boulanger Basic Mechanisms of CAVD
the fibrosa layer and then extend through the tissue in distorting the normal architecture of the aortic valve. It is noteworthy that oxidized lipid species are present in the vicinity of mineralized nodules.2,6 Simple microscopic observations indicate that inflammatory cells, predominantly composed of macrophages, infiltrate the mineralized aortic valve and tend to form clusters.7 In these clusters of inflammatory cells an active process of neovascularization takes place. It has been reported that endothelial progenitor cells, which incidentally infiltrate the aortic valve, promote neoangiogenesis.8 Also, neovascularization of the aortic valve has been shown to be associated with the expression of heat shock protein 60.9 Hence, it suggests that inflammation and neovascularization are intimately linked to aortic valve mineralization and remodelling. To this effect, we recently reported in 285 stenotic aortic valves that the presence of inflammatory infiltrates is associated with a greater remodelling score of tissues.10 In some stenotic aortic valves, chondroosteogenic metaplasia is observed.11 Dense inflammatory infiltrates often accompany these metaplastic changes (Fig. 1). Thus, it is likely that, at least in a subgroup of valves, active transformation of resident cells into osteoblast-like cells occurs and might therefore participate in the mineralization of the aortic valve. Valve interstitial cells (VICs), the main cellular component of the aortic valve, are actively involved in the production of extracellular matrix and mineralization during CAVD. It should be pointed out that VICs represent a heterogeneous population of cells, which might thus exhibit different phenotypes. Li et al. proposed that VICs should be regrouped into 5 distinct phenotypes: (1) embryonic progenitor endothelial/mesenchymal cells; (2) quiescent VICs; (3) activated VICs; (4) progenitor VICs; and (5) osteoblastic VICs.12 VICs represent a cell population with a high plasticity, that might change phenotype interchangeably depending on cell context and cues delivered to the cells. As such, different VIC populations are present at different degrees from embryogenesis to later in life in the adult. For instance, although embryonic progenitor endothelial/ mesenchymal cells are involved in endothelial-mesenchymal transition (EndoMT) during valve formation from endocardial cushion, in adulthood the quiescent VICs ensure a normal valve function by promoting the maintenance of the extracellular matrix components.13 Activated VICs and progenitor VICs, which are involved in tissue repair, might also play a major role during different pathobiological processes, including CAVD. On this score, signals delivered to VICs play a key role in the development and progression of CAVD. For instance, VICs with a strong osteogenic potential have been identified in the aortic valve and therefore with appropriate cues this population might be activated and promote mineralization.14 It should be pointed out that, because of intensive research effort, several molecular cues and signalling cascades involved in pathologic mineralization and fibrosis have been brought to light in the recent years. Lipid Retention Process Small, dense, low-density lipoprotein Epidemiological studies emphasized that low-density lipoprotein (LDL) cholesterol is a risk factor for the
983
development of CAVD.15 Also, as previously highlighted when retrospective study results indicated that statins were associated with a slower hemodynamic progression rate of stenosis, 3 randomized controlled studies reported in contrast, that a lipid-lowering strategy neither resulted in lower aortic valve-related events nor in a slower progression rate of stenosis.4 Hence, considering that different active lipid species are present in the aortic valve, why did statins fail at preventing valve-related events? It has been argued that statins might have been started too late in the disease process when it was too advanced. Also, part of the answer might lie in the specific processes related to lipid retention and modification that occurs during CAVD. Oxidized LDL (ox-LDL) and their derived reactive lipid species are strong promoters of mineralization when assessed in isolated VICs.16 Studies indicate that circulating levels of ox-LDL are associated with the remodelling score of mineralized aortic valves.6 Moreover, Mohty et al. showed that the amount of ox-LDL within explanted CAVD tissues was associated with aortic valve inflammation (Fig. 2).17 In this study, the only lipid variable associated with the accumulation of tissue ox-LDL was the proportion of small, dense LDL. Small, dense LDLs have greater ability to infiltrate tissues and are prone to the oxidation process. The high proportion of small, dense LDL in patients with the metabolic syndrome might explain the faster progression rate of AS documented in this group of patients.18,19 In addition, it should be pointed out that although statins are efficient at lowering LDL-cholesterol they have, however, no or at best a modest effect on the proportion of circulating small, dense LDL. Lipoprotein-associated phospholipase A2 mediates mineralization of the aortic valve Uncoupling of nitric oxide (NO) activity in CAVD has been documented and it has been shown that it promotes the generation of reactive oxygen species, which might, in turn, promote the production of highly reactive lipid-derived oxidized species.20 We recently identified that lipoproteinassociated phospholipase A2 (Lp-PLA2) is overexpressed in tissues of CAVD. Lp-PLA2 converts ox-LDL into lysophosphatidylcholine (LPC), which is incidentally present in CAVD tissues.21 Moreover, LPC is a strong promoter of mineralization in isolated VICs through a cyclic adenosine monophosphate (cAMP)/protein kinase A pathway. Therefore, it is likely that Lp-PLA2 is produced locally, within the aortic valve, by macrophages and/or is transported in the aortic valve by LDL, particularly by small, dense LDL, and enhances lipid retention/modification. Recently, a single nucleotide polymorphism in the lipoprotein (a) locus (rs10455872) has been associated with aortic valve calcification.22 Moreover, the blood plasma level of lipoprotein (a) has also been associated with an increased risk of aortic valve stenosis.23 These findings are of interest considering that lipoprotein (a) transports oxidizedphospholipids, which are transformed by Lp-PLA2 into LPC.24 It should also be highlighted that other enzymes, expressed in stenotic aortic valve, might promote lipid retention/modification. In this regard, lipoprotein lipase is expressed in CAVD tissues where it colocalizes with oxidized lipids.25 Also, phospholipid transfer protein (PLTP) is overexpressed in CAVD tissues. Derbali et al. found that stimulation of
984
Canadian Journal of Cardiology Volume 30 2014
Figure 1. Histological sections of calcific aortic valve disease, that illustrate leukocyte infiltration (A, B); in (B) magnification of square in (A). The inflammatory infiltrates in calcific aortic valve disease is composed of macrophages (C) and T cells (D), which colocalize with osteogenic metaplasia et al.10 with kind and neovascularization (E, F); in (F) magnification of square in (E); asterisk indicates mineralized area. Reproduced from Côte permission from Springer Science and Business Media.
Toll-like receptor (TLR) 2 promoted the expression of PLTP by VICs.26 PLTP binds to high-density lipoproteins (HDLs) and might thus decrease their ability to perform reverse cholesterol transport. Biglycan, a proteoglycan, which binds to
LDL particles and promotes their retention through electrostatic interactions, is also a signalling molecule that can stimulate TLR-2 and, in doing so, promotes the expression of PLTP. Thus, there is in the aortic valve a highly efficient
Figure 2. In calcific aortic valve disease, the level of oxidized-LDL (ox-LDL) is associated with inflammatory infiltrates and the expression of tumour necrosis factor (TNF)-a. Reproduced from Mohty et al.17 with permission from Lippincott Williams and Wilkins/Wolters Kluwer Health. Copyright ª 2008.
Mathieu and Boulanger Basic Mechanisms of CAVD
mechanism that promotes and sustains the retention of lipids (Fig. 3). Is There a Role for HDLs in CAVD? HDLs promote the efflux of cholesterol from macrophages through adenosine triphosphate (ATP)-binding cassette transporters. Moreover, HDLs have antioxidative and antiinflammatory properties.27 Among other things, the expression of paraoxonase has been consistently shown to mediate some of the anti-inflammatory effects of HDLs. With regard to the pathologic mineralization process, HDLs have been shown to reduce the calcification of isolated vascular smooth muscle cells (VSMCs).28 Recently, Arsenault et al. showed that scavenger receptor class B type 1 (SR-B1) and ATPbinding cassette transporter A1 (ABCA1)-mediated cholesterol efflux was not affected in patients with CAVD.29 However, they documented that the efflux process was altered in patients with coronary artery disease. Within explanted specimens of CAVD apolipoprotein A1 of HDL is present in the vicinity of calcific nodules. Furthermore, it has been documented that apolipoprotein A1 contributes to the formation of amyloid proteins found in explanted human specimens of CAVD.30 Also, it was documented that amyloidderived proteins from CAVD promotes the mineralization of isolated VICs. Although not yet clearly elucidated it is thus possible that, in certain circumstances, HDLs are retained and modified in CAVD tissues. Hence, these altered HDLs might thus promote lipid retention and contribute to trigger mineralization by being transformed into amyloid substance. HDL is thus a heterogeneous group of lipoproteins and depending on
985
the context might be protective or detrimental. Studies in rodent models have shown that administration of HDL mimetic peptides prevented the mineralization of the aortic valve and even promoted the regression of CAVD.31 These in vivo findings suggest that targeting HDL is of potential therapeutic relevance. However, further work is necessary to understand the role of HDL in the pathogenesis of CAVD. Inflammation at the Centre Stage of Mineralization Matricellular proteins, inflammation, and neovascularization In a transcriptomic assay it was reported that approximately 15% of the 715 genes differentially regulated in human mineralized aortic valves were related to inflammation.32 In a histopathologic study of 285 human stenotic aortic valves, the presence of a chronic inflammatory infiltrate in 28% of the examined valves was documented.10 In these valves, the presence of chronic inflammatory infiltrates was associated with osteochondrogenic metaplasia and neovascularization (Fig. 1). Also, the density of leukocytes correlated with the expression of tumour necrosis factor (TNF)-a and the hemodynamic progression rate of stenosis documented preoperatively. Inflammation is accompanied by the expression of metalloproteinases (MMPs), which participate in the remodelling of tissues. MMP9, which is highly expressed within stenotic aortic valves, is closely associated with the expression of osteonectin/secreted protein, acidic and rich in cysteine/osteonectin (SPARC).8 Osteonectin, a matricellular protein, is secreted in the
Figure 3. Low-density lipoprotein (LDL) and Lp(a) transport phospholipids in the aortic valve. Lipids are trapped by biglycan (BGN) and decorin. Lipoprotein lipase forms complex with decorin, which helps retain lipoproteins. BGN entrains the production of phospholipid transfer protein (PLTP), which might enhance lipid retention by impeding on apolipoprotein A1 (ApoA1) function. Oxidative stress promoted by nitric oxide synthase (NOS) uncoupling induces oxidation of phospholipids. Lp-PLA2, which is transported by LDL and also secreted by macrophages promotes the production of lysophosphatidylcholine (LPC). In turn, LPC increases inflammation and the mineralization of valve interstitial cells (VICs). CE, cholesteryl ester; IL, interleukin; LPL, lipoprotein lipase; Lp-PLA2, lipoprotein-associated phospholipase A2; TNF, tumour necrosis factor.
986
extracellular space and is cleaved by MMPs into a proangiogenic peptide. In turn, it is possible that neoangiogenesis accompanying inflammation might participate in the recruitment of more inflammatory cells in a vicious cycle. Mice deficient for chondromodulin-I, an antiangiogenic factor, develop mineralization of the aortic valve, which co-occurs with the expression of vascular endothelial growth factor A and neoangiogenesis.33 Thus, this mouse model support the hypothesis that neovascularisation of the aortic valve, which is closely associated with the inflammatory process, promotes the development of CAVD. Also, MMPs, by producing peptide fragments might directly promote the mineralization of the aortic valve. In this regard, MMP12, which is highly expressed in stenotic aortic valves, produces elastin-derived peptide fragments with osteogenic properties.34 Tumour necrosis family of cytokines and mineralization of the aortic valve Inflammatory cells produce different cytokines, that might promote the mineralization of the aortic valve. To this effect, the expression of TNF-a by macrophages promote the mineralization of VSMCs and VICs through several mechanisms. When stimulated, TNF receptor 1 recruits TNF receptor-associated death domain and activates Fas-associated protein with death domain, which results in activation of apoptosis through caspase 8 and 3.35 In turn, apoptosis is a strong promoter of ectopic valve mineralization. In VSMCs, TNF-a promotes the mineralization through a cAMP/protein kinase A pathway.36 Also, a study has underlined that TNF-a promoted the expression of Msx2 in VSMCs.37 Msx2, a homeobox transcription factor, has been shown to induce the mineralization of the aortic valve through the Wnt pathway.38 Recently, TNF-related apoptosis-inducing ligand has been shown to induce apoptosis-mediated mineralization of VIC cultures. TNF-related apoptosis-inducing ligand, a member of the TNF superfamily, binds to the death receptor 4, which is incidentally overexpressed by VIC during mineralization, and provokes programmed cell death through a caspase 8 pathway.39 Interleukin (IL)-1 receptor antagonist deficient mice, IL1-Ra/, have increased levels of circulating TNF-a and develop a thickening of the aortic leaflets. Of interest, the double knockout IL1-Ra/ TNF-a/ mice do not develop CAVD.40 Hence, it is likely that crosstalk between IL-1Ra and TNF-a is involved in the development of CAVD in mice. Upstream factors such as oxidized-LDLs are potent inducers of TNF-a expression. Different oxidized lipid species bind to pattern recognition receptors, such as the TLRs, and, in doing so, trigger inflammation. TLRs are expressed by inflammatory cells and also by VICs. In this regard, stimulation of TLR-2 in VICs promotes an osteogenic differentiation.41 Hence, expression of TLRs might link inflammation with mineralization during the development of CAVD. IL-6 Is a Key Player in Osteogenic Transition of VICs IL-6 is a pleiotropic cytokine, which is secreted in response to various stimuli. El Husseini et al. recently provided evidence for the first time that IL-6 is overexpressed in stenotic aortic valve and that VICs are an important source of IL-6.42 Moreover, it was documented that during
Canadian Journal of Cardiology Volume 30 2014
mineralization of VICs, IL-6 was a key signal in promoting the production of bone morphogenetic protein (BMP) 2. Of interest, IL-6 has also been shown to promote EndoMT in the aortic valve.43 Hence, it is possible that inflammatory processes increase the recruitment of mesenchymal cells within the aortic valve. However, whether mesenchymal cells derived from EndoMT participate actively in the pathobiology of CAVD remains to be determined. The Renin-Angiotensin System Hypertension is associated with the development of CAVD and a faster progression rate of stenosis.44 It should be pointed out that hypertension is often accompanied by the presence of visceral ectopic fat, which might contribute to activate the renin-angiotensin system (RAS), which has intricate links with inflammation and insulin resistance.45 Clinical retrospective studies have shown that inhibition of the RAS is associated with a slower progression rate of stenosis.46 In prehypertensive patients with CAVD, the circulating levels of angiotensin II are associated with inflammation and tissue remodelling of the aortic valves.47 In mice, the administration of angiotensin II results in a significant thickening of the aortic leaflets.48 Histological analyses of explanted human stenotic aortic valves revealed that angiotensin-converting enzyme (ACE) is expressed and colocalized with angiotensin II.49 It is believed that ACE could be transported in the aortic valve by LDL. Moreover, chymase, another angiotensin II-producing enzyme, has been found to be expressed in stenotic aortic valves.50 Hence, during the development of CAVD, an efficient system would locally promote the production of angiotensin II, which might participate in tissue fibrosis and remodelling of the aortic valve. Retrospective studies suggest that inhibition of the RAS with angiotensin receptor blockers (ARBs) and not with ACE inhibitors are associated with a slower hemodynamic progression rate of stenosis.44,51 We documented in 208 patients that patients taking ARBs had lower remodelling scores of stenotic aortic valves.52 In addition, patients taking ARBs had less inflammation and IL-6 expression in the aortic valve.53 It is possible that ACE inhibitors, which do not inhibit chymase, are less efficient at preventing the production of angiotensin II locally in the aortic valve. In the hypercholesterolemic rabbit, the administration of olmesartan prevented the development of CAVD.54 However, the role of the RAS and its inhibition in CAVD remains to be fully investigated in clinical trials before any conclusion can be drawn. Phosphate Signalling: The Ecto-Nucleotidase/ Purinergic System Phosphate is a crucial regulator of aortic valve mineralization The mineral found in stenotic aortic valves is mostly composed of hydroxyapatite of calcium. Mineralization in living organisms is dependent on the nucleation of calcium phosphate (Pi). In humans, the product of calcium and Pi reaches a saturation point and because of proteins that bind to calcium, the mineralization of soft tissues is prevented. For instance, fetuin A, an abundant plasma protein secreted by the
Mathieu and Boulanger Basic Mechanisms of CAVD
987
liver, prevents the formation of calcium-Pi crystals.55 Also, blood vessels and the aortic valves produce matrix Gla protein (MGP), which prevents mineralization when carboxylated through a vitamin K-dependent process.56 Among other things, the mechanism by which fetuin A and MGP prevent ectopic mineralization is by averting the nucleation of calcium-Pi crystals. Hence, it is likely that a delicate balance exists between factors that increase the Pi-calcium product and antimineralizing proteins, such as fetuin A and MGP, that prevent the nucleation process. Investigations in the past several years suggest that processes that increase systemic or local production of Pi are important drivers of the ectopic valve mineralization. In the Cardiovascular Health Study, the blood plasma level of Pi has been shown to be independently related to CAVD.57 In this regard, Pi is an important signalling molecule exerting promineralizing properties on VICs.58 El Husseini et al. recently identified that a type III sodium-dependent Pi cotransporter, Pit1 (SLC20A1), was overexpressed in stenotic aortic valves and promoted the mineralization of VIC cultures.59 In vitro, silencing of Pit1 prevented Pi-induced mineralization of VIC cultures. This study underlined that Pit1 is a key regulator by which Pi enters cells and promotes a loss of mitochondrial membrane potential and a release of cytochrome c in the cytosol, which entrains the activation of caspase 3. Hence, chronic exposure of VICs to slightly increased concentrations of Pi (2 mM) triggers apoptosis through a Pit1-dependent process, whereby mineralization is promoted. Hence, it is likely that mechanisms that sustain an elevated level of Pi, either systematically or locally, promote the mineralization of VICs.
receptor signal transduction in VICs involves the phosphoinositide 3-kinase/Akt pathway. Akt (PKB) inactivates Bax by a phosphorylation process on Ser184. Hence, a certain amount of ATP is necessary to prevent the apoptosis of VICs. When ATP level is depleted by an overactivity of the ectonucleotidases, less signalling is delivered by the P2Y2 receptor and programmed cell death is triggered. It is important to point out that apoptosis might promote/enhance mineralization by the production of apoptotic bodies, which contain membrane remnants along with Pi-producing enzymes. The importance of the P2Y2 receptor in preventing mineralization is supported by the fact that silencing of the receptor results in a hypermineralizing phenotype of VICs.60 We recently identified that among the different mechanisms that might increase the expression of ENPP1, mechanical strain plays a key role. In vitro, strain increases the mineralization of VIC cultures in a RhoA-dependent manner.63 On exposure to strain, VICs activate the small GTPase RhoA, which by the Rho kinase mediates the export of ENPP1-containing cargo vesicles to the plasma membrane. Therefore, increased expression of ENPP1 at the cell membrane promotes the production of Pi and the depletion of ATP, which when combined are strong promineralizing signals (Fig. 4). In rats, the administration of an ectonucleotidase inhibitor prevented warfarin-induced mineralization of the aortic valve.64 Taken together, these findings suggest that the ectonucleotidase/purinergic receptor system is one important effector in promoting mineralization of the aortic valve.
Ectonucleotidases regulate the levels of Pi and nucleotides/nucleosides: relevance for CAVD
Crosstalk between Notch and Wnt
Ectonucleotidases, which are membrane-bound enzymes that hydrolyze nucleotides and nucleosides, generate Pi and therefore might have an important effect on the mineralization of the aortic valve. A recent report identified that ectonucleotide pyrophosphatase/phosphodiesterase 1 is highly expressed in human stenotic aortic valves.60 Also, singlenucleotide polymorphisms in the 30 untranslated region and 50 untranslated region of ectonucleotide pyrophosphatase/ phosphodiesterase 1 (ENPP1) are significantly associated with CAVD. Of note, these single-nucleotide polymorphisms have been previously associated with diabetes, hypertension, and obesity. ENPP1 converts ATP, which is secreted by VICs, into adenosine monophosphate and pyrophosphate (PPi). PPi is a strong inhibitor of mineralization by binding to Pi and calcium crystals. Hence, a complete knockdown of ENPP1 in mice results in ectopic mineralization of tendons.61 In vitro experiments highlighted that overexpression of ENPP1 in VICs promoted the mineralization of cell cultures (Fig. 4). It has been documented that when overexpressed, ENPP1 exerts a promineralizing effect by several mechanisms. First, during mineralization of VIC cultures, alkaline phosphatase, another ectonucleotidase, is also highly expressed and promotes the hydrolysis of PPi into Pi.62 Hence, overall, the level of Pi surpasses PPi and, in doing so, promotes mineralization of VIC cultures. Second, a high level of ectonucleotidases depletes the extracellular level of ATP, which is an important survival signal for VICs through the P2Y2 receptor. P2Y2
Osteoblastic Transition: Notch, Wnt, and BMPs
One striking feature of stenotic aortic valves when examined using a microscope is that chondro-osteogenic metaplasia is present in approximately 15% of tissues.65 This suggests that a full transition toward osteogenic cells occurs during the development of CAVD, at least in a subset of patients. The clinical features that drive this response have not yet been identified, but the molecular pathways leading to osteogenic transdifferentiation have been under intense scrutiny in the past decade. The different pathways that lead to the production of osteoblast-like cells in the aortic valves have important overlap and cross talk and might thus respond to different stimuli, which results in transdifferentiation of VICs and/or mesenchymal cells into osteoblast-like cells (Fig. 5). Garg et al. reported that mutations of Notch1, a receptor involved in tissue patterning during embryogenesis, was associated with bicuspid aortic valve (BAV) and left ventricular outflow tract malformations in families.66 This suggested that Notch signalling might play a role in the development aortic valve pathologies. During the developmental process, cell-cell communications with Notch receptors (Notch1-4) and their agonists, such as the delta-like proteins and Jagged 1-2, ensure a proper tissue organization and determine cell fate. Engagement of Notch receptors by agonists results in the activation of the g secretase complex, which produces the Notch intracellular domain (NICD).67 The NICD next translocates to the nucleus where it promotes the activity of the recombining binding protein suppressor of hairless (RBPjk) and expression of the Hairy family of repressors. In VICs, one
988
Canadian Journal of Cardiology Volume 30 2014
Figure 4. The role of the ectonucleotidase and purinergic signalling in calcific aortic valve disease (CAVD). Nucleotides are secreted by VICs and hydrolyzed by ENPP1 and alkaline phosphatase (ALP), which increases the level of phosphate (Pi) (1). Pi enters cells through the Pi transporter Pit1/ SLC20A1, which increases the expression of ENPP1 in a positive feedback loop (2). A high level of ectonucleotidases in CAVD contributes to deplete the extracellular levels of nucleotides (3). In turn, decreased signalling through the P2Y2 receptor decreases Akt and promotes apoptosis-mediated mineralization (4). ADP, adenosine diphosphate; AMP, adenosine monophosphate; ATP, adenosine triphosphate; ENPP, ectonucleotide pyrophos et al.60 with permission from phatase/phosphodiesterase 1; mRNA, microRNA; PPi, pyrophosphate; VIC, valve interstitial cell. Adapted from Cote Elsevier.
important effect of the Hairy repressors is to prevent the expression of runt-related transcription factor 2 (Runx2) and BMP2. Runx2 is an osteoblast transcription factor, and BMP2 is a morphogen that promotes osteogenic differentiation. The Notch1þ/ mice given a Western diet develop calcification of the aortic valve but have a tricuspid anatomy.68 Similarly, the RBPjkþ/ mice have trileaflet aortic valves and exhibit mineralization and thickening of the aortic valves when given a cholesterol-rich diet.69 In mice with a gene knockout of periostin, a protein highly expressed in the endocardial cushion during embryogenesis, the overexpression of deltalike 1 homolog (Dlk1), a negative regulator of Notch1, is found. Knockout mice for periostin have dysmorphic BAVlike aortic valves and develop valve fibrosis/mineralization along with the expression of Runx2 and osteopontin.70 Hence, disturbances in the Notch signalling cascade leads to the production of pro-osteogenic signals, however, dysfunction in Notch signalling in animal models does not necessarily result in BAV. In humans, Ducharme et al. reported that mutations in Notch1 were associated with CAVD in patients
with trileaflet aortic valves.71 Taken together, these findings suggest that mutations of Notch1 signalling induce osteogenic transition of VICs and can be associated with different phenotypic traits (BAV and tricuspid aortic valve). The Wnt pathway is another important transduction system involved in cell fate and has also been implicated in the development of CAVD. Expression of Wnt3a is increased in stenotic aortic valves.72 In the canonical transduction pathway Wnt agonists bind to the coreceptor Lrp5/6 and Frizzled, which by still poorly understood mechanisms leads to inactivation of a complex formed by adenomatosis polyposis coli (APC), Axin, and glycogen synthase kinase 3 (GSK3). In absence of Wnt, the APC/Axin/GSK3 complex phosphorylates b-catenin, which is next targeted for destruction by the proteasome. Hence, when the coreceptor Lrp5/6 and Frizzled is stimulated, it promotes the activation and recruitment of Dishevelled, which in turn acts as an adaptor for APC/Axin/GSK3 and inhibits its activity.73 It then follows that degradation of b-catenin is prevented. The stabilized b-catenin translocates to the nucleus, where it modifies the transcriptional state of the cell by interacting with
Mathieu and Boulanger Basic Mechanisms of CAVD
989
microRNA-141 is just a reflection of increased mineralization often encountered with BAV or it is specific to the BAV anatomy and/or increased strain. Nonetheless, functional investigations in porcine VICs suggest that the microRNA-141 is a negative regulator of transforming growth factor (TGF)-b1induced osteogenic transition through BMP2. The complexity of noncoding genomes is just emerging and it is thus likely that the role of microRNAs and other noncoding RNAs in the pathobiology of CAVD will expand in the next several years.
Figure 5. Signalling and cross talk between the Notch and Wnt pathway in calcific aortic valve disease. Notch 1 signalling normally represses BMP2 and Runx2 expression and also might impede on bcatenin stabilization and signalling. In calcific aortic valve disease, increased expression of Wnt3a stimulates the coreceptor formed by Frizzled and Lrp5/6, which induces stabilization of b-catenin and promotes osteogenic transition of valve interstitial cells. BMP, bone morphogenetic protein; NICD, Notch intracellular domain; Runx2, runtrelated transcription factor 2.
the TCF family of transcription factors. In VICs, transclocation of b-catenin to the nucleus increases the transcriptional activity of BMP2 and promotes osteogenic transition.74 Of interest, Notch negatively regulates the Wnt signalling pathway. Although the molecular mechanism is still ill-defined, in vertebrates the NICD interferes with the transcriptional activity of b-catenin.75 Thus, although it remains to be investigated in the context of CAVD, it is possible that the reciprocal interactions between Notch and Wnt play a role in cell fate and transdifferentiation of VICs into osteoblast-like cells. Noncoding RNAs as regulators of BMP2 signalling in CAVD The Encyclopedia of DNA Elements (ENCODE) project has recently provided evidence that 90% of the genome is transcribed.76 Only 2% of the transcripts encode for proteins, whereas the others are noncoding RNAs. The small noncoding RNAs (
Review
Basic Mechanisms of Calcific Aortic Valve Disease Patrick Mathieu, MD, and Marie-Chloe Boulanger, PhD Laboratoire d’Études Mole culaires des Valvulopathies (LEMV), Groupe de Recherche en Valvulopathies (GRV), Que bec Heart and Lung Institute/Research Center, Department of Surgery, Laval University, Que bec, Que bec, Canada
ABSTRACT
RESUM E
Calcific aortic valve disease (CAVD) is the most common heart valve disorder. There is no medical treatment to prevent and/or promote the regression of CAVD. Hence, it is of foremost importance to delineate and understand the key basic underlying mechanisms involved in CAVD. In the past decade our comprehension of the underpinning processes leading to CAVD has expanded at a fast pace. Hence, our understanding of the basic pathobiological processes implicated in CAVD might lead eventually to the development of novel pharmaceutical therapies for CAVD. In this review, we discuss molecular processes that are implicated in fibrosis and mineralization of the aortic valve. Specifically, we address the role of lipid retention, inflammation, phosphate signalling and osteogenic transition in the development of CAVD. Interplays between these different processes and the key regulation pathways are discussed along with their clinical relevance.
La calcification de la valve aortique (CVA) est le trouble des valves quent. Aucun traitement me dical ne peut cardiaques le plus fre venir ou favoriser la re gression de la CVA. Par conse quent, il est pre finir et de comprendre les principaux particulièrement important de de canismes de base sous-jacents qui sont implique s dans la CVA. Au me cennie, notre compre hension des fondements cours de la dernière de très rapidement. Par des processus menant à la CVA a progresse quent, notre compre hension des processus biopathologiques conse s dans la CVA mènerait e ventuellement à fondamentaux implique laboration de nouveaux traitements pharmacologiques de la CAV. l’e culaires qui sont Dans cette revue, nous discutons des processus mole s dans la fibrose et la mine ralisation de la valve aortique. implique tention lipidique, de Particulièrement, nous traitons du rôle de la re l’inflammation, de la signalisation du phosphate et de la transition ogène dans le de veloppement de la CVA. Nous discutons des inoste rents processus et les voies de re gulation teractions entre ces diffe principales ainsi que leur pertinence clinique.
Calcific aortic valve disease (CAVD) encompasses a wide spectrum of clinical entities, from aortic sclerosis to severe aortic stenosis (AS). However, there is a general agreement that aortic sclerosis and AS are part of a same pathobiological process whereby the aortic leaflets undergo a progressive mineralization and fibrotic process. Hence, a group of leading experts, convened by the National Heart, Lung and Blood Institute, have developed a consensus statement and have recommended that the different clinical entities, aortic sclerosis and AS, should be termed CAVD.1 For many years, CAVD has been described as a “passive” process related to aging. However, in the past 15 years there has been a growing interest to decipher the cellular and molecular processes involved in CAVD. Studies underscored that CAVD shared many clinical risk factors with atherosclerosis. In addition, histopathological examination of surgically explanted stenotic aortic valves indicated that several
features such as lipid infiltration, inflammation, and calcification were commonly observed in CAVD.2 Moreover, several animal models of atherosclerosis also developed some degree of CAVD.3 Taken together, these findings suggested that CAVD was only another manifestation of an atherosclerotic process. However, 3 randomized trials with statins failed to demonstrate any benefit from a lipid-lowering approach in patients with a moderate-to-severe AS.4 These studies suggested that although CAVD has possibly some overlapping processes in common with atherosclerosis it has underlying specific mechanisms that lead to the mineralization of the aortic valve. To this effect, a growing number of studies have since shed more light on the underpinning processes at play in CAVD, which might open new therapeutic avenues. In this article, we review the latest discoveries related to CAVD, its basic molecular processes, and whenever possible we have tried to tie these basic discoveries with clinical observations, keeping in mind a translational approach.
Received for publication January 7, 2014. Accepted March 19, 2014. Corresponding author: Dr Patrick Mathieu, Institut de Cardiologie et de Pneumologie de Quebec/Quebec Heart and Lung Institute, 2725 Chemin Ste-Foy, Quebec, Quebec G1V 4G5, Canada. Tel.: þ1-418-656-4717; fax: þ1-418-656-4707. E-mail: [email protected] See page 990 for disclosure information.
Histopathology of CAVD and Basic Mechanism of Mineralization CAVD is characterized by the presence of mineralized nodules and fibrosis.5 Histological sections of explanted CAVD during surgeries revealed that mineralization starts in
0828-282X/$ - see front matter 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cjca.2014.03.029
Mathieu and Boulanger Basic Mechanisms of CAVD
the fibrosa layer and then extend through the tissue in distorting the normal architecture of the aortic valve. It is noteworthy that oxidized lipid species are present in the vicinity of mineralized nodules.2,6 Simple microscopic observations indicate that inflammatory cells, predominantly composed of macrophages, infiltrate the mineralized aortic valve and tend to form clusters.7 In these clusters of inflammatory cells an active process of neovascularization takes place. It has been reported that endothelial progenitor cells, which incidentally infiltrate the aortic valve, promote neoangiogenesis.8 Also, neovascularization of the aortic valve has been shown to be associated with the expression of heat shock protein 60.9 Hence, it suggests that inflammation and neovascularization are intimately linked to aortic valve mineralization and remodelling. To this effect, we recently reported in 285 stenotic aortic valves that the presence of inflammatory infiltrates is associated with a greater remodelling score of tissues.10 In some stenotic aortic valves, chondroosteogenic metaplasia is observed.11 Dense inflammatory infiltrates often accompany these metaplastic changes (Fig. 1). Thus, it is likely that, at least in a subgroup of valves, active transformation of resident cells into osteoblast-like cells occurs and might therefore participate in the mineralization of the aortic valve. Valve interstitial cells (VICs), the main cellular component of the aortic valve, are actively involved in the production of extracellular matrix and mineralization during CAVD. It should be pointed out that VICs represent a heterogeneous population of cells, which might thus exhibit different phenotypes. Li et al. proposed that VICs should be regrouped into 5 distinct phenotypes: (1) embryonic progenitor endothelial/mesenchymal cells; (2) quiescent VICs; (3) activated VICs; (4) progenitor VICs; and (5) osteoblastic VICs.12 VICs represent a cell population with a high plasticity, that might change phenotype interchangeably depending on cell context and cues delivered to the cells. As such, different VIC populations are present at different degrees from embryogenesis to later in life in the adult. For instance, although embryonic progenitor endothelial/ mesenchymal cells are involved in endothelial-mesenchymal transition (EndoMT) during valve formation from endocardial cushion, in adulthood the quiescent VICs ensure a normal valve function by promoting the maintenance of the extracellular matrix components.13 Activated VICs and progenitor VICs, which are involved in tissue repair, might also play a major role during different pathobiological processes, including CAVD. On this score, signals delivered to VICs play a key role in the development and progression of CAVD. For instance, VICs with a strong osteogenic potential have been identified in the aortic valve and therefore with appropriate cues this population might be activated and promote mineralization.14 It should be pointed out that, because of intensive research effort, several molecular cues and signalling cascades involved in pathologic mineralization and fibrosis have been brought to light in the recent years. Lipid Retention Process Small, dense, low-density lipoprotein Epidemiological studies emphasized that low-density lipoprotein (LDL) cholesterol is a risk factor for the
983
development of CAVD.15 Also, as previously highlighted when retrospective study results indicated that statins were associated with a slower hemodynamic progression rate of stenosis, 3 randomized controlled studies reported in contrast, that a lipid-lowering strategy neither resulted in lower aortic valve-related events nor in a slower progression rate of stenosis.4 Hence, considering that different active lipid species are present in the aortic valve, why did statins fail at preventing valve-related events? It has been argued that statins might have been started too late in the disease process when it was too advanced. Also, part of the answer might lie in the specific processes related to lipid retention and modification that occurs during CAVD. Oxidized LDL (ox-LDL) and their derived reactive lipid species are strong promoters of mineralization when assessed in isolated VICs.16 Studies indicate that circulating levels of ox-LDL are associated with the remodelling score of mineralized aortic valves.6 Moreover, Mohty et al. showed that the amount of ox-LDL within explanted CAVD tissues was associated with aortic valve inflammation (Fig. 2).17 In this study, the only lipid variable associated with the accumulation of tissue ox-LDL was the proportion of small, dense LDL. Small, dense LDLs have greater ability to infiltrate tissues and are prone to the oxidation process. The high proportion of small, dense LDL in patients with the metabolic syndrome might explain the faster progression rate of AS documented in this group of patients.18,19 In addition, it should be pointed out that although statins are efficient at lowering LDL-cholesterol they have, however, no or at best a modest effect on the proportion of circulating small, dense LDL. Lipoprotein-associated phospholipase A2 mediates mineralization of the aortic valve Uncoupling of nitric oxide (NO) activity in CAVD has been documented and it has been shown that it promotes the generation of reactive oxygen species, which might, in turn, promote the production of highly reactive lipid-derived oxidized species.20 We recently identified that lipoproteinassociated phospholipase A2 (Lp-PLA2) is overexpressed in tissues of CAVD. Lp-PLA2 converts ox-LDL into lysophosphatidylcholine (LPC), which is incidentally present in CAVD tissues.21 Moreover, LPC is a strong promoter of mineralization in isolated VICs through a cyclic adenosine monophosphate (cAMP)/protein kinase A pathway. Therefore, it is likely that Lp-PLA2 is produced locally, within the aortic valve, by macrophages and/or is transported in the aortic valve by LDL, particularly by small, dense LDL, and enhances lipid retention/modification. Recently, a single nucleotide polymorphism in the lipoprotein (a) locus (rs10455872) has been associated with aortic valve calcification.22 Moreover, the blood plasma level of lipoprotein (a) has also been associated with an increased risk of aortic valve stenosis.23 These findings are of interest considering that lipoprotein (a) transports oxidizedphospholipids, which are transformed by Lp-PLA2 into LPC.24 It should also be highlighted that other enzymes, expressed in stenotic aortic valve, might promote lipid retention/modification. In this regard, lipoprotein lipase is expressed in CAVD tissues where it colocalizes with oxidized lipids.25 Also, phospholipid transfer protein (PLTP) is overexpressed in CAVD tissues. Derbali et al. found that stimulation of
984
Canadian Journal of Cardiology Volume 30 2014
Figure 1. Histological sections of calcific aortic valve disease, that illustrate leukocyte infiltration (A, B); in (B) magnification of square in (A). The inflammatory infiltrates in calcific aortic valve disease is composed of macrophages (C) and T cells (D), which colocalize with osteogenic metaplasia et al.10 with kind and neovascularization (E, F); in (F) magnification of square in (E); asterisk indicates mineralized area. Reproduced from Côte permission from Springer Science and Business Media.
Toll-like receptor (TLR) 2 promoted the expression of PLTP by VICs.26 PLTP binds to high-density lipoproteins (HDLs) and might thus decrease their ability to perform reverse cholesterol transport. Biglycan, a proteoglycan, which binds to
LDL particles and promotes their retention through electrostatic interactions, is also a signalling molecule that can stimulate TLR-2 and, in doing so, promotes the expression of PLTP. Thus, there is in the aortic valve a highly efficient
Figure 2. In calcific aortic valve disease, the level of oxidized-LDL (ox-LDL) is associated with inflammatory infiltrates and the expression of tumour necrosis factor (TNF)-a. Reproduced from Mohty et al.17 with permission from Lippincott Williams and Wilkins/Wolters Kluwer Health. Copyright ª 2008.
Mathieu and Boulanger Basic Mechanisms of CAVD
mechanism that promotes and sustains the retention of lipids (Fig. 3). Is There a Role for HDLs in CAVD? HDLs promote the efflux of cholesterol from macrophages through adenosine triphosphate (ATP)-binding cassette transporters. Moreover, HDLs have antioxidative and antiinflammatory properties.27 Among other things, the expression of paraoxonase has been consistently shown to mediate some of the anti-inflammatory effects of HDLs. With regard to the pathologic mineralization process, HDLs have been shown to reduce the calcification of isolated vascular smooth muscle cells (VSMCs).28 Recently, Arsenault et al. showed that scavenger receptor class B type 1 (SR-B1) and ATPbinding cassette transporter A1 (ABCA1)-mediated cholesterol efflux was not affected in patients with CAVD.29 However, they documented that the efflux process was altered in patients with coronary artery disease. Within explanted specimens of CAVD apolipoprotein A1 of HDL is present in the vicinity of calcific nodules. Furthermore, it has been documented that apolipoprotein A1 contributes to the formation of amyloid proteins found in explanted human specimens of CAVD.30 Also, it was documented that amyloidderived proteins from CAVD promotes the mineralization of isolated VICs. Although not yet clearly elucidated it is thus possible that, in certain circumstances, HDLs are retained and modified in CAVD tissues. Hence, these altered HDLs might thus promote lipid retention and contribute to trigger mineralization by being transformed into amyloid substance. HDL is thus a heterogeneous group of lipoproteins and depending on
985
the context might be protective or detrimental. Studies in rodent models have shown that administration of HDL mimetic peptides prevented the mineralization of the aortic valve and even promoted the regression of CAVD.31 These in vivo findings suggest that targeting HDL is of potential therapeutic relevance. However, further work is necessary to understand the role of HDL in the pathogenesis of CAVD. Inflammation at the Centre Stage of Mineralization Matricellular proteins, inflammation, and neovascularization In a transcriptomic assay it was reported that approximately 15% of the 715 genes differentially regulated in human mineralized aortic valves were related to inflammation.32 In a histopathologic study of 285 human stenotic aortic valves, the presence of a chronic inflammatory infiltrate in 28% of the examined valves was documented.10 In these valves, the presence of chronic inflammatory infiltrates was associated with osteochondrogenic metaplasia and neovascularization (Fig. 1). Also, the density of leukocytes correlated with the expression of tumour necrosis factor (TNF)-a and the hemodynamic progression rate of stenosis documented preoperatively. Inflammation is accompanied by the expression of metalloproteinases (MMPs), which participate in the remodelling of tissues. MMP9, which is highly expressed within stenotic aortic valves, is closely associated with the expression of osteonectin/secreted protein, acidic and rich in cysteine/osteonectin (SPARC).8 Osteonectin, a matricellular protein, is secreted in the
Figure 3. Low-density lipoprotein (LDL) and Lp(a) transport phospholipids in the aortic valve. Lipids are trapped by biglycan (BGN) and decorin. Lipoprotein lipase forms complex with decorin, which helps retain lipoproteins. BGN entrains the production of phospholipid transfer protein (PLTP), which might enhance lipid retention by impeding on apolipoprotein A1 (ApoA1) function. Oxidative stress promoted by nitric oxide synthase (NOS) uncoupling induces oxidation of phospholipids. Lp-PLA2, which is transported by LDL and also secreted by macrophages promotes the production of lysophosphatidylcholine (LPC). In turn, LPC increases inflammation and the mineralization of valve interstitial cells (VICs). CE, cholesteryl ester; IL, interleukin; LPL, lipoprotein lipase; Lp-PLA2, lipoprotein-associated phospholipase A2; TNF, tumour necrosis factor.
986
extracellular space and is cleaved by MMPs into a proangiogenic peptide. In turn, it is possible that neoangiogenesis accompanying inflammation might participate in the recruitment of more inflammatory cells in a vicious cycle. Mice deficient for chondromodulin-I, an antiangiogenic factor, develop mineralization of the aortic valve, which co-occurs with the expression of vascular endothelial growth factor A and neoangiogenesis.33 Thus, this mouse model support the hypothesis that neovascularisation of the aortic valve, which is closely associated with the inflammatory process, promotes the development of CAVD. Also, MMPs, by producing peptide fragments might directly promote the mineralization of the aortic valve. In this regard, MMP12, which is highly expressed in stenotic aortic valves, produces elastin-derived peptide fragments with osteogenic properties.34 Tumour necrosis family of cytokines and mineralization of the aortic valve Inflammatory cells produce different cytokines, that might promote the mineralization of the aortic valve. To this effect, the expression of TNF-a by macrophages promote the mineralization of VSMCs and VICs through several mechanisms. When stimulated, TNF receptor 1 recruits TNF receptor-associated death domain and activates Fas-associated protein with death domain, which results in activation of apoptosis through caspase 8 and 3.35 In turn, apoptosis is a strong promoter of ectopic valve mineralization. In VSMCs, TNF-a promotes the mineralization through a cAMP/protein kinase A pathway.36 Also, a study has underlined that TNF-a promoted the expression of Msx2 in VSMCs.37 Msx2, a homeobox transcription factor, has been shown to induce the mineralization of the aortic valve through the Wnt pathway.38 Recently, TNF-related apoptosis-inducing ligand has been shown to induce apoptosis-mediated mineralization of VIC cultures. TNF-related apoptosis-inducing ligand, a member of the TNF superfamily, binds to the death receptor 4, which is incidentally overexpressed by VIC during mineralization, and provokes programmed cell death through a caspase 8 pathway.39 Interleukin (IL)-1 receptor antagonist deficient mice, IL1-Ra/, have increased levels of circulating TNF-a and develop a thickening of the aortic leaflets. Of interest, the double knockout IL1-Ra/ TNF-a/ mice do not develop CAVD.40 Hence, it is likely that crosstalk between IL-1Ra and TNF-a is involved in the development of CAVD in mice. Upstream factors such as oxidized-LDLs are potent inducers of TNF-a expression. Different oxidized lipid species bind to pattern recognition receptors, such as the TLRs, and, in doing so, trigger inflammation. TLRs are expressed by inflammatory cells and also by VICs. In this regard, stimulation of TLR-2 in VICs promotes an osteogenic differentiation.41 Hence, expression of TLRs might link inflammation with mineralization during the development of CAVD. IL-6 Is a Key Player in Osteogenic Transition of VICs IL-6 is a pleiotropic cytokine, which is secreted in response to various stimuli. El Husseini et al. recently provided evidence for the first time that IL-6 is overexpressed in stenotic aortic valve and that VICs are an important source of IL-6.42 Moreover, it was documented that during
Canadian Journal of Cardiology Volume 30 2014
mineralization of VICs, IL-6 was a key signal in promoting the production of bone morphogenetic protein (BMP) 2. Of interest, IL-6 has also been shown to promote EndoMT in the aortic valve.43 Hence, it is possible that inflammatory processes increase the recruitment of mesenchymal cells within the aortic valve. However, whether mesenchymal cells derived from EndoMT participate actively in the pathobiology of CAVD remains to be determined. The Renin-Angiotensin System Hypertension is associated with the development of CAVD and a faster progression rate of stenosis.44 It should be pointed out that hypertension is often accompanied by the presence of visceral ectopic fat, which might contribute to activate the renin-angiotensin system (RAS), which has intricate links with inflammation and insulin resistance.45 Clinical retrospective studies have shown that inhibition of the RAS is associated with a slower progression rate of stenosis.46 In prehypertensive patients with CAVD, the circulating levels of angiotensin II are associated with inflammation and tissue remodelling of the aortic valves.47 In mice, the administration of angiotensin II results in a significant thickening of the aortic leaflets.48 Histological analyses of explanted human stenotic aortic valves revealed that angiotensin-converting enzyme (ACE) is expressed and colocalized with angiotensin II.49 It is believed that ACE could be transported in the aortic valve by LDL. Moreover, chymase, another angiotensin II-producing enzyme, has been found to be expressed in stenotic aortic valves.50 Hence, during the development of CAVD, an efficient system would locally promote the production of angiotensin II, which might participate in tissue fibrosis and remodelling of the aortic valve. Retrospective studies suggest that inhibition of the RAS with angiotensin receptor blockers (ARBs) and not with ACE inhibitors are associated with a slower hemodynamic progression rate of stenosis.44,51 We documented in 208 patients that patients taking ARBs had lower remodelling scores of stenotic aortic valves.52 In addition, patients taking ARBs had less inflammation and IL-6 expression in the aortic valve.53 It is possible that ACE inhibitors, which do not inhibit chymase, are less efficient at preventing the production of angiotensin II locally in the aortic valve. In the hypercholesterolemic rabbit, the administration of olmesartan prevented the development of CAVD.54 However, the role of the RAS and its inhibition in CAVD remains to be fully investigated in clinical trials before any conclusion can be drawn. Phosphate Signalling: The Ecto-Nucleotidase/ Purinergic System Phosphate is a crucial regulator of aortic valve mineralization The mineral found in stenotic aortic valves is mostly composed of hydroxyapatite of calcium. Mineralization in living organisms is dependent on the nucleation of calcium phosphate (Pi). In humans, the product of calcium and Pi reaches a saturation point and because of proteins that bind to calcium, the mineralization of soft tissues is prevented. For instance, fetuin A, an abundant plasma protein secreted by the
Mathieu and Boulanger Basic Mechanisms of CAVD
987
liver, prevents the formation of calcium-Pi crystals.55 Also, blood vessels and the aortic valves produce matrix Gla protein (MGP), which prevents mineralization when carboxylated through a vitamin K-dependent process.56 Among other things, the mechanism by which fetuin A and MGP prevent ectopic mineralization is by averting the nucleation of calcium-Pi crystals. Hence, it is likely that a delicate balance exists between factors that increase the Pi-calcium product and antimineralizing proteins, such as fetuin A and MGP, that prevent the nucleation process. Investigations in the past several years suggest that processes that increase systemic or local production of Pi are important drivers of the ectopic valve mineralization. In the Cardiovascular Health Study, the blood plasma level of Pi has been shown to be independently related to CAVD.57 In this regard, Pi is an important signalling molecule exerting promineralizing properties on VICs.58 El Husseini et al. recently identified that a type III sodium-dependent Pi cotransporter, Pit1 (SLC20A1), was overexpressed in stenotic aortic valves and promoted the mineralization of VIC cultures.59 In vitro, silencing of Pit1 prevented Pi-induced mineralization of VIC cultures. This study underlined that Pit1 is a key regulator by which Pi enters cells and promotes a loss of mitochondrial membrane potential and a release of cytochrome c in the cytosol, which entrains the activation of caspase 3. Hence, chronic exposure of VICs to slightly increased concentrations of Pi (2 mM) triggers apoptosis through a Pit1-dependent process, whereby mineralization is promoted. Hence, it is likely that mechanisms that sustain an elevated level of Pi, either systematically or locally, promote the mineralization of VICs.
receptor signal transduction in VICs involves the phosphoinositide 3-kinase/Akt pathway. Akt (PKB) inactivates Bax by a phosphorylation process on Ser184. Hence, a certain amount of ATP is necessary to prevent the apoptosis of VICs. When ATP level is depleted by an overactivity of the ectonucleotidases, less signalling is delivered by the P2Y2 receptor and programmed cell death is triggered. It is important to point out that apoptosis might promote/enhance mineralization by the production of apoptotic bodies, which contain membrane remnants along with Pi-producing enzymes. The importance of the P2Y2 receptor in preventing mineralization is supported by the fact that silencing of the receptor results in a hypermineralizing phenotype of VICs.60 We recently identified that among the different mechanisms that might increase the expression of ENPP1, mechanical strain plays a key role. In vitro, strain increases the mineralization of VIC cultures in a RhoA-dependent manner.63 On exposure to strain, VICs activate the small GTPase RhoA, which by the Rho kinase mediates the export of ENPP1-containing cargo vesicles to the plasma membrane. Therefore, increased expression of ENPP1 at the cell membrane promotes the production of Pi and the depletion of ATP, which when combined are strong promineralizing signals (Fig. 4). In rats, the administration of an ectonucleotidase inhibitor prevented warfarin-induced mineralization of the aortic valve.64 Taken together, these findings suggest that the ectonucleotidase/purinergic receptor system is one important effector in promoting mineralization of the aortic valve.
Ectonucleotidases regulate the levels of Pi and nucleotides/nucleosides: relevance for CAVD
Crosstalk between Notch and Wnt
Ectonucleotidases, which are membrane-bound enzymes that hydrolyze nucleotides and nucleosides, generate Pi and therefore might have an important effect on the mineralization of the aortic valve. A recent report identified that ectonucleotide pyrophosphatase/phosphodiesterase 1 is highly expressed in human stenotic aortic valves.60 Also, singlenucleotide polymorphisms in the 30 untranslated region and 50 untranslated region of ectonucleotide pyrophosphatase/ phosphodiesterase 1 (ENPP1) are significantly associated with CAVD. Of note, these single-nucleotide polymorphisms have been previously associated with diabetes, hypertension, and obesity. ENPP1 converts ATP, which is secreted by VICs, into adenosine monophosphate and pyrophosphate (PPi). PPi is a strong inhibitor of mineralization by binding to Pi and calcium crystals. Hence, a complete knockdown of ENPP1 in mice results in ectopic mineralization of tendons.61 In vitro experiments highlighted that overexpression of ENPP1 in VICs promoted the mineralization of cell cultures (Fig. 4). It has been documented that when overexpressed, ENPP1 exerts a promineralizing effect by several mechanisms. First, during mineralization of VIC cultures, alkaline phosphatase, another ectonucleotidase, is also highly expressed and promotes the hydrolysis of PPi into Pi.62 Hence, overall, the level of Pi surpasses PPi and, in doing so, promotes mineralization of VIC cultures. Second, a high level of ectonucleotidases depletes the extracellular level of ATP, which is an important survival signal for VICs through the P2Y2 receptor. P2Y2
Osteoblastic Transition: Notch, Wnt, and BMPs
One striking feature of stenotic aortic valves when examined using a microscope is that chondro-osteogenic metaplasia is present in approximately 15% of tissues.65 This suggests that a full transition toward osteogenic cells occurs during the development of CAVD, at least in a subset of patients. The clinical features that drive this response have not yet been identified, but the molecular pathways leading to osteogenic transdifferentiation have been under intense scrutiny in the past decade. The different pathways that lead to the production of osteoblast-like cells in the aortic valves have important overlap and cross talk and might thus respond to different stimuli, which results in transdifferentiation of VICs and/or mesenchymal cells into osteoblast-like cells (Fig. 5). Garg et al. reported that mutations of Notch1, a receptor involved in tissue patterning during embryogenesis, was associated with bicuspid aortic valve (BAV) and left ventricular outflow tract malformations in families.66 This suggested that Notch signalling might play a role in the development aortic valve pathologies. During the developmental process, cell-cell communications with Notch receptors (Notch1-4) and their agonists, such as the delta-like proteins and Jagged 1-2, ensure a proper tissue organization and determine cell fate. Engagement of Notch receptors by agonists results in the activation of the g secretase complex, which produces the Notch intracellular domain (NICD).67 The NICD next translocates to the nucleus where it promotes the activity of the recombining binding protein suppressor of hairless (RBPjk) and expression of the Hairy family of repressors. In VICs, one
988
Canadian Journal of Cardiology Volume 30 2014
Figure 4. The role of the ectonucleotidase and purinergic signalling in calcific aortic valve disease (CAVD). Nucleotides are secreted by VICs and hydrolyzed by ENPP1 and alkaline phosphatase (ALP), which increases the level of phosphate (Pi) (1). Pi enters cells through the Pi transporter Pit1/ SLC20A1, which increases the expression of ENPP1 in a positive feedback loop (2). A high level of ectonucleotidases in CAVD contributes to deplete the extracellular levels of nucleotides (3). In turn, decreased signalling through the P2Y2 receptor decreases Akt and promotes apoptosis-mediated mineralization (4). ADP, adenosine diphosphate; AMP, adenosine monophosphate; ATP, adenosine triphosphate; ENPP, ectonucleotide pyrophos et al.60 with permission from phatase/phosphodiesterase 1; mRNA, microRNA; PPi, pyrophosphate; VIC, valve interstitial cell. Adapted from Cote Elsevier.
important effect of the Hairy repressors is to prevent the expression of runt-related transcription factor 2 (Runx2) and BMP2. Runx2 is an osteoblast transcription factor, and BMP2 is a morphogen that promotes osteogenic differentiation. The Notch1þ/ mice given a Western diet develop calcification of the aortic valve but have a tricuspid anatomy.68 Similarly, the RBPjkþ/ mice have trileaflet aortic valves and exhibit mineralization and thickening of the aortic valves when given a cholesterol-rich diet.69 In mice with a gene knockout of periostin, a protein highly expressed in the endocardial cushion during embryogenesis, the overexpression of deltalike 1 homolog (Dlk1), a negative regulator of Notch1, is found. Knockout mice for periostin have dysmorphic BAVlike aortic valves and develop valve fibrosis/mineralization along with the expression of Runx2 and osteopontin.70 Hence, disturbances in the Notch signalling cascade leads to the production of pro-osteogenic signals, however, dysfunction in Notch signalling in animal models does not necessarily result in BAV. In humans, Ducharme et al. reported that mutations in Notch1 were associated with CAVD in patients
with trileaflet aortic valves.71 Taken together, these findings suggest that mutations of Notch1 signalling induce osteogenic transition of VICs and can be associated with different phenotypic traits (BAV and tricuspid aortic valve). The Wnt pathway is another important transduction system involved in cell fate and has also been implicated in the development of CAVD. Expression of Wnt3a is increased in stenotic aortic valves.72 In the canonical transduction pathway Wnt agonists bind to the coreceptor Lrp5/6 and Frizzled, which by still poorly understood mechanisms leads to inactivation of a complex formed by adenomatosis polyposis coli (APC), Axin, and glycogen synthase kinase 3 (GSK3). In absence of Wnt, the APC/Axin/GSK3 complex phosphorylates b-catenin, which is next targeted for destruction by the proteasome. Hence, when the coreceptor Lrp5/6 and Frizzled is stimulated, it promotes the activation and recruitment of Dishevelled, which in turn acts as an adaptor for APC/Axin/GSK3 and inhibits its activity.73 It then follows that degradation of b-catenin is prevented. The stabilized b-catenin translocates to the nucleus, where it modifies the transcriptional state of the cell by interacting with
Mathieu and Boulanger Basic Mechanisms of CAVD
989
microRNA-141 is just a reflection of increased mineralization often encountered with BAV or it is specific to the BAV anatomy and/or increased strain. Nonetheless, functional investigations in porcine VICs suggest that the microRNA-141 is a negative regulator of transforming growth factor (TGF)-b1induced osteogenic transition through BMP2. The complexity of noncoding genomes is just emerging and it is thus likely that the role of microRNAs and other noncoding RNAs in the pathobiology of CAVD will expand in the next several years.
Figure 5. Signalling and cross talk between the Notch and Wnt pathway in calcific aortic valve disease. Notch 1 signalling normally represses BMP2 and Runx2 expression and also might impede on bcatenin stabilization and signalling. In calcific aortic valve disease, increased expression of Wnt3a stimulates the coreceptor formed by Frizzled and Lrp5/6, which induces stabilization of b-catenin and promotes osteogenic transition of valve interstitial cells. BMP, bone morphogenetic protein; NICD, Notch intracellular domain; Runx2, runtrelated transcription factor 2.
the TCF family of transcription factors. In VICs, transclocation of b-catenin to the nucleus increases the transcriptional activity of BMP2 and promotes osteogenic transition.74 Of interest, Notch negatively regulates the Wnt signalling pathway. Although the molecular mechanism is still ill-defined, in vertebrates the NICD interferes with the transcriptional activity of b-catenin.75 Thus, although it remains to be investigated in the context of CAVD, it is possible that the reciprocal interactions between Notch and Wnt play a role in cell fate and transdifferentiation of VICs into osteoblast-like cells. Noncoding RNAs as regulators of BMP2 signalling in CAVD The Encyclopedia of DNA Elements (ENCODE) project has recently provided evidence that 90% of the genome is transcribed.76 Only 2% of the transcripts encode for proteins, whereas the others are noncoding RNAs. The small noncoding RNAs (
