|Home | About | Journals | Submit | Contact Us | Français|
Calcific aortic valve disease (CAVD) increasingly afflicts our aging population. One-third of our elderly have echocardiographic or radiological evidence of aortic valve sclerosis (CAVS), an early and subclinical form of CAVD. Age, gender, tobacco use, hypercholesterolemia, hypertension, and type II diabetes all contribute to the risk of disease that has worldwide distribution. Upon progression to its most severe form --- calcific aortic stenosis (CAS) --- CAVD becomes debilitating and devastating, and 2% of individuals over age 60 suffer from CAS to the extent that surgical intervention is required. No effective pharmacotherapies exist for treating those at risk for clinical progression. It is becoming increasingly apparent that a diverse spectrum of cellular and molecular mechanisms converge to regulate valvular calcium load; this is evidenced not only in histopathologic heterogeneity of CAVD but also from the multiplicity of cell types that can participate in valve biomineralization. In this review, we highlight our current understanding of CAVD disease biology, emphasizing molecular and cellular aspects of its regulation. We end by pointing to important biological and clinical questions that must be answered to enable sophisticated disease staging and the development of new strategies to medically treat CAVD.
Calcific aortic valve disease (CAVD) increasingly afflicts our aging population. One-third of our elderly have echocardiographic or radiological evidence of aortic valve sclerosis (CAVS), an early and subclinical form of CAVD1. However, even in middle age, approximately 10% exhibit CAVS by echocardiography2. Upon progression to its most severe form --- calcific aortic stenosis (CAS) --- CAVD becomes debilitating and devastating, and 2% of individuals over age 60 suffer from CAS to the extent that surgery is required to preclude death once symptoms occur2. Age, gender, tobacco use, hypercholesterolemia, and hypertension all contribute to the risk of disease that has worldwide distribution1, 2. Genetics plays a direct role in that bicuspid aortic valve – a congenital risk factor for precocious CAS – has a significant genetic diathesis3, 4. Recently, type II diabetes (T2DM) has emerged as a particularly relevant and worrisome metabolic risk factor for native CAVD5 as well as precocious degeneration of bioprosthetic valves6. It is becoming increasingly apparent that a diverse spectrum of cell-dependent mechanisms converge to regulate valvular calcium load; this is evidenced not only in histopathologic heterogeneity of CAVD7, 8 but also from the multiplicity of cell types –interstitial cells, endothelial cells, cardiac chondrocytes and circulating osteoprogenitors – that participate in valve biomineralization1, 9. This review highlights our current understanding of CAVD disease biology and its regulation, emphasizing cellular physiology and mechanisms. It ends with important biological and clinical questions that must be answered to enable sophisticated disease staging and the development of new strategies to medically treat CAVD.
The surgical view of CAVD is shaped by the gross anatomical appearance of rock-hard calcific nodules distorting and stiffening the normally pliant aortic leaflets7, 10. Detailed histopathological studies first established that much of the material in these calcified nodules is amorphous calcium phosphate – i.e., acellular aggregates not organized into any recognizable histological structure seen in functionally normal mineralizing tissues7. An elegant study by Mohler, Kaplan and colleagues demonstrate that mature bone – replete with marrow elements – can be identified in 13% of calcified valve specimens7 (Figure 1). Immunohistochemistry demonstrates evidence of inflammation and bone morphogenetic protein (BMP2) expression as occurs in atherosclerosis7. Fitzpatrick et al were amongst the first to identify woven and lamellar bone with osteoblast matrix production in CAS specimens11. Vascularization – a key component of orthotopic bone formation – was significantly more prominent in calcifying native aortic valves vs. bioprosthetic valves11. Nevertheless, as highlighted by Rajamannan, even in the absence of overt ectopic bone formation, osteogenic gene regulatory programs are elaborated in the injured valves undergoing biomineralization12. Thus, features of actively regulated calcified matrix metabolism are recruited and participate in biomineralization even in the absence of overt ossification.
Otto, O’Brien et al described the progression of CAVD in their landmark study of human aortic valve histopathology8. At the valve’s ventricular surface, subendothelial matrix and lipid accumulation with downward displacement and fragmentation of the subjacent elastic lamina represent the earliest pathological change. Plaque-like subendothelial deposits preferentially occur on the aortic surface (fibrosa) of the valve. When aortic valve and vascular disease are induced in LDL receptor-deficient mice, valve fibrosis also preferentially initiates on the aortic surface (Figure 2), potentially reflecting the non-laminar shear experienced by valve endothelial cells (VECs) on that surface1.
Endothelial dysfunction is likely to precede even these changes, reflecting the important role of the VEC as the initial barrier to metabolic, mechanical, and inflammatory insults8. Moreover, patients with AVS exhibit evidence of global endothelial dysfunction13. VECs can also contribute to calcifying vascular cell types via the endothelial-mesenchymal transition (EnMT) elicited response to cues that promote arteriosclerotic calcification14-16 (see below). As disease progresses, an exuberant inflammatory infiltrate consisting of T cells and macrophages is elaborated17. This inflammatory infiltrate is common both to CAVD of aging and the precocious biomineralization that arises with bicuspid aortic valves (BAV)17. Interestingly, very recent data suggest that an adaptive immune response may be activated in CAVD since clonally expanded effector-memory T cell populations are present both in the valve and in the circulation of patients with severe CAS18. The contribution of acquired immunity to CAVD remains to be determined. However, innate immune responses are clearly involved19; NLRP320, 21 and Toll-like receptor (TLR) signals22, 23 recognizing pathogen associated molecular patterns, oxidized LDL, and amorphous calcium phosphate are activated and pathogenic in CAVD22 (see below).
The earliest calcium deposition is observed in a stippled pattern associated with the base of lesions – viz., at the fibrosal interface with the fibrofatty expansion of the valve spongiosa8. Amorphous calcium phosphate deposits can form readily via epitaxial mineral deposition on a number of nidi, including cholesterol crystals24, 25, fragmented elastin fibers, and/or collagen26, 27. Murshed, Karesenty and colleagues were amongst the first to demonstrate that co-expression of collagen and alkaline phosphatase (ALP) in the elastin-rich environment of skin was adequate to drive mineralization28. By hydrolyzing inorganic pyrophosphate and de-phosphorylating osteopontin, natural inhibitors of calcium phosphate deposition, ALP enables robust tissue mineralization that would rapidly occur at the prevailing and physiological calcium phosphate concentrations that are fully permissive for mineral deposition29, 30. Rajamannan and colleagues deployed immunogold electron microscopy to demonstrate the increased presence of ALP protein in CAVD31.
The ALP ectoenzyme is not randomly secreted into the extracellular milieu but is shed from cells bound to matrix vesicles (MVs) – spherical lipid bilayers of ~ 30 to 100 nm diameter -- that organize mineralization in physical association with elastin and collagen fibrils32-35. The prescient bone biologist, H. Clarke Anderson, first demonstrated the presence of ALP-positive MVs in atherosclerotic plaques33. Kim, Schoen and colleagues went on to show the contributions of ALP – laden matrix vesicles to the valve sclerotic mineralization process in native and bioprosthetic valves36-38. A novel MV phosphatase discovered by Millan and colleagues, Phospho1, may also play a particularly important role in MV-mediated biomineralization39. However, although osteogenic gene expression including ALP can drive arterial calcification with or without overt ectopic bone formation, it has been recently appreciated that regulated biomineralization also can initiate even in conditions of relatively low ALP activity40, albeit more slowly41. MVs also contain annexins A5 (Anx5), A6 (AnxA6) and phosphatidyl serine – molecules that readily bind calcium and nucleate mineral deposition in the absence of ALP activity40, 42, 43. Of note, ubiquitin staining of CAVD specimens revealed evidence of enhanced autophagy and cell death with a relative paucity of TUNEL-positive cells44, suggesting that MVs released during autophagic death may be distinct from those of apoptosis. It should be noted that, once initiated, deposition of amorphous calcium phosphate may progress in the absence of MVs45 –particularly if inhibitors of mineralization such as pyrophosphate, phospho-osteopontin, and fetuin46 are deficienct. Indeed, the pro-inflammatory, thrombin-degraded form of osteopontin47 is observed at sites of severe CAVD48. The biochemical nature of the earliest mineral deposits in CAVD -- and the relative roles for ALP, phospholipids, annexins and other nucleators vs. mineralization inhibitors in disease progression -- remains to be fully characterized. Once mineral deposition has initiated, the arrival of circulating osteoprogenitors derived from myeloid cell lineage may play a particularly important role in programming stage-specific disease responses.
In the advanced CAVD lesions described by Fitzpatrick11, Mohler7, and Rajamannan31, both amorphous calcium phosphate concretions and ectopic woven bone formation were observed. This points to the molecular and cellular heterogeneity of the biomineralization processes that control valvular calcium accrual7 (Figure 3). Only recently have the potential reasons for this heterogeneity been mechanistically considered – likely relating in part to the recruitment of circulating osteoprogenitors capable of recapitulating the complete bone microenvironment49, 50. Khosla51, Pignolo52, and colleagues have identified circulating osteoprogenitor cell populations that participate in ectopic osteogenic injury responses during fracture repair and valve ossification. The origin of these circulating cells is almost certainly the bone marrow since COP (circulating osteogenic precursor) cells appear in higher numbers during growth phase and young adulthood51, 53. Indeed, the well-known relationship between younger age and bioprosthetic valve failure may relate to this observation54. In the ectopic bone seen in 13% of patients with CAS52, Mohler and Pignolo described sequential phases of inflammatory infiltration, fibroproliferation, neovascularization, cartilage formation and endochondral ossification. Type I collagen(+) CD45(+) COP cells appear in human valves at the fibroproliferative (70%) and neovascularization (30%) phases of disease, while CD45(+) cells were identified in both ossifying and non-ossifying valve segments52. Since COP cells were only observed in segments adjacent to valves with true heterotopic ossification, the authors suggest that this cell population participates in the late-stage histopathological evolution of CAVD52.
More recently, the myeloid calcifying cell (MCC) was described by Fadini and colleagues55. These cells are positive for both bone alkaline phosphatase / ALP and osteocalcin. Studies of gender-mismatched bone marrow transplant recipients and studies of individuals afflicted with chronic myelogenous leukemia confirmed bone marrow derivation of MCC osteoprogenitors55. Importantly, circulating MCCs are increased in the setting of T2DM – and may help explain the increases in CAVS arising in both native and bioprosthetic valves in this setting55. Very recently, Drake and colleagues provided evidence that in murine models that CD45(+)CD133(+)CD34(−) bone marrow derived cells contribute to the VIC population on the ventricular side of semilunar valves in the absence of valve injury during normal valve homeostasis56. A subset of these cells located near sites of valve coaptation expressed myeloid markers56. The precise relationships between this latter cell type, COP cells, MCCs, and the fibrocyte – first described by Bucala in 199457 – have yet to be firmly established. However, Pignolo speculates that the COP cells his group identified likely represent members of the myeloid-derived fibrocyte lineage58. As rigorous functional assays are co-registered with more sophisticated cell-surface lineage markers, these relationships will no doubt emerge. Furthermore, as described below, the mechanical and paracrine milieu that regulates the osteofibrogenic potential of valve tissue-autonomous cell types may be significantly altered by valve accumulation of myeloid-derived lineages including both COP cells and macrophages.
The aortic valve forms from both neural crest derived and endothelial- viz., endocardial -- cushion progenitors59, with migratory neural crest cells programmed in part via interactions with cells of the secondary heart field59, 60. Many of the key paracrine signaling molecules of the Wnt61, 62, and Notch 63, 64signaling family as well as of the TGF-beta superfamily (including BMP)65, which are critical for control of bone formation66, also play vital roles in the earliest stages of aortic valve morphogenesis. Moreover, as Lincoln and Yutzey first highlighted, a number of transcriptional regulators necessary for cartilage, bone and tendon formation – including Sox9, NFATc1, Msx1, Msx2, and scleraxis (Scx) -- are also highly expressed in developing aortic valves67. While this may not appear immediately germane to CAVD, the roles played by these molecules in skeletal morphogenesis are enlightening when viewed in the context of diseased valve biology.
Outside of the skull, bone formation largely occurs via endochondral ossification, a process whereby an initially avascular cartilaginous template provisionally mineralizes via matrix vesicle and chondrocyte apoptotic mechanisms68. Osteoclast-mediated calcified matrix remodeling and vascular invasion with vascular-associated osteoprogenitors results in true bone formation69. The transcription factor Sox9 plays multiple vital and sequential roles in endochondral bone formation, starting with the specification of mesechymal progenitors that form the first cartilaginous anlage70. The valve interstitial cell (VIC) population contains abundant Sox9-positive mesenchymal cells that are osteogenic progenitors70, 71. As Yutzey and colleagues demonstrated, Cre-mediated conditional deletion of Sox9 from EC and mesenchymal cells demonstrated its critical role in valve morphogenesis, VIC proliferative expansion, proteoglycan expression and CAVD71. As in the skeleton, while specifying a cell type with the capacity for biomineralization, valve Sox9 demarcates and maintains the VIC as a structurally synthetic but non-mineralizing mesenchymal cell with self-renewing potential71. As a transcription factor important for elaboration of specific extracellular matrix programs, Sox9 binds WWCAAWGX(N)CWTTGWW (W = A or T) DNA cognates, then directs the expression and ultimate secretion of chondrocytic type II, IX, and XI collagens, and multiple proteoglycans including aggrecan, biglycan, and fibromodulin72. The latter two are particularly important since biglycan and fibromodulin organize the extracellular “niche” environment for progenitors capable of creating dense regular connective tissue73. Loss of Sox9 expression in valve interstitial cells is accompanied by progression to biomineralization71 – similar to the programming one expects from studies of the endochondral growth plate in the developing skeleton70. Sox9 binding to cognates in Runx2 and Osx genes – two transcriptional regulators necessary for osteogenic mineralization – has been proposed to inhibit elaboration of the osteoblast phenotype during chondrocyte specification72. Thus, a model emerges in which sustained Sox9 expression and activity in VICs preserves the proliferative and synthetic chondrocytic phenotype necessary for valve function and durability, and prevents the elaboration of osteogenic mineralization programs71, 74. Of note, since Msx2 binds to Sox9 and inhibits its function75, some of the pro-osteogenic actions of this homeodomain transcription factor may in fact relate to antagonism of Sox9 actions in VICs and other osteochondroprogenitors above and beyond its activation of canonical Wnt signaling76.
Another molecule familiar to skeletal biologists – CNP or C-type natriuretic peptide – is also important in the biology of valve calcinosis. CNP functions to promote endochondral bone formation via activation of B-type CNP receptors – a transmembrane guanylate cyclase that signals in part via downstream cyclic GMP dependent protein kinases PGK-I and PGK-II77, 78. Simmons and colleagues first identified that VECs on the ventricular surface express very high levels of CNP, as compared to the aortic VECs on the aortic surface of the aortic valve79. They went on to show that VICs located in the ventricular portion of the valve interstitium also expressed high levels of CNP80. Since the ventricular face of the aortic valve is resistant to the procalcific mechanical and metabolic milieu that drives CAVD8, they examined the effects of CNP on VIC-dependent mineralization. CNP and PKG signaling were shown to prevent VIC osteogenic nodule formation and mineralization80. Thus, paracrine interactions between ventricular surface VECs and subjacent VIC populations may serve to restrict osteogenic potential of valve cells, and may explain in part the temporospatial evolution of CAVD.
Other pathways important in skeletal morphogenesis are also involved in the regulation of cardiac valve morphogenesis and CAVD. TGF-beta, BMPs, and Wnt polypeptides act via corresponding ALK- and LRP- receptor signaling complexes to promote bone formation, mineralization, and skeletal homeostasis throughout vertebrate life66. Rajamannan and colleagues first identified the contributions of Wnt/beta-catenin signaling to CAVD12. Insightful studies demonstrated expression of the osteogenic Wnt receptor LRP5 and upregulation of signaling mediator beta-catenin in calcifying human aortic valve specimens12. Subsequent analysis of apoE−/−;LRP5−/− mice demonstrated that global deficiency of this canonical Wnt signaling system reduced aortic valve calcification81. Of note, circulating inhibitors of Wnt signaling are elevated in patients with CAVD82; since many of these inhibitors are upregulated by canonical Wnt pathway activation as mechanisms for feedback inhibition, these molecules may serve as useful biomarkers for disease severity82.
TGF-beta superfamily / ALK receptor/ Smad signaling cascades, initiated by ALK1 and ALK5 activation in ECs and ALK5 in pericytes, are indispensible for normal vasculogenesis from the very earliest stages of vertebrate development83, 84. Likewise, active Wnt/LRP receptor/beta-catenin signaling is necessary for the function of EC progenitors62, 85 and -- with BMP / ALK receptor/ Smad pathways – it promotes the endothelial-mesenchymal transition (EnMT) necessary to create cardiac valves86 and epicardial fibroblasts87. These same morphogenetic pathways appear to participate in the pathobiology of cardiovascular calcification and post-natal valve homeostasis. Miller, Weiss, Heistad and colleagues were amongst the first to demonstrate that hypercholesterolemia-driven CAVD activated osteogenic BMP/Smad1/5 and fibrogenic TGF-beta/Smad2 signaling in a hemodynamically significant murine disease model88. Intriguingly, once initiated, cholesterol lowering was capable of reducing calcification but neither fibrosis nor Smad2 signaling in diseased valves. Jo and colleagues related activation of this fibro-osteogenic signaling cascade in human aortic valves to selective down-regulation of Smad6, an intracellular inhibitor of BMP and TGF-beta signaling89.
Bostrom and colleagues elegantly demonstrated that inhibition of this BMP-activated pathway may be safely achieved to limit cardiovascular calcification by autocrine/paracrine elaboration of matrix Gla- protein (MGP)90. MGP is a vitamin K-dependent, noggin-like, faux receptor inhibitor of BMP2 and BMP491. While a detailed analysis of aortic valve function was not undertaken, it is intriguing to note that exposure to warfarin – an inhibitor of MGP gamma-carboxylation and function – is significantly associated with human CAVD risk92. Using an ex vivo model, Ferrari and colleagues demonstrated that cyclical stretch and BMP4 could activate VIC-mediated calcification of human aortic valve leaflets – and that noggin inhibited this process93. Similarly, Yu and colleagues demonstrated that Fc-ALK3, an engineered BMP faux receptor, could also safely reduce cardiovascular calcification in LDLR−/− mice94. The ALK2/ALK3 inhibitor LDN-193189 - a compound with lesser inhibitory potency for ALK5 – was able to recapitulate the effects of Fc-ALK3 administration in this model94. A full toxicology study was not performed and, unfortunately, valvular structure and function were not addressed.
However, other studies have recently pointed to important roles for ALK signaling in valve homeostasis and highlight the therapeutic challenges. The TGF-beta receptor ALK5 has been targeted by post-natal pharmacological inhibition in hopes of mitigating cardiopulmonary, renal and hepatic fibrosis95. Moreover, as discussed below, mechanically challenged VICs become hyper-responsive to TGF-beta, elaborating myofibroblast phenotype and osteogenic potential96. However, Anderton et al recently demonstrated that ALK5 inhibition with either AZ12601011 or AZ12799734 -- small molecule inhibitors of ALK5 – induced cardiac valve inflammation with neutrophil infiltration, VIC proliferation, and hemorrhagic degeneration in adult rats97. These data point to the critical role for ALK5 activity in maintenance of valve integrity throughout life – but highlights the difficulties of direct targeting of ALK receptors for treatment of cardiovascular disease95, 97. Strategies that electively modify ALK signaling “tone” via extracellular inhibitors of specific ligands appear to hold therapeutic promise.
A cellular feature common to virtually all clinically significant forms of macrovascular calcification – including CAVD – is inflammation7, 17. TNF98, 99, IL1-beta100, advanced glycosylation end products101, 102, IL6103, and oxidized LDL cholesterol (oxLDL)104, 105 have all been shown to activate vascular biomineralization and vascular osteogenic signaling processes. Valve calcification is increased in mice lacking IL1RA, an inhibitor of IL1 signaling106. Since concomitant TNF deficiency reduces calcification arising with IL1R deficiency106, TNF bioactivity appears to be an important component – potentially related to pro-osteogenic oxidative stress signaling107. Furthermore, osteogenic BMP and Wnt signaling cascades are entrained to TNF activity in the vessel wall.98 However, with respect to CAVD, the actions of oxLDL deserve special attention. oxLDL is a pro-inflammatory pathogenic component of dyslipidemia, arising from spontaneous extracellular chemical oxidation of LDL cholesterol108. oxLDL signals in part via heteromeric TLR complexes – the receptors for viral and bacterial pathogens that activate innate immune responses to non-specifically fight infection109. Elevated levels of oxLDL are associated with worsened fibrocalcific responses in CAVD23. Meng and colleagues recently identified that vascular BMP2 expression was inducible by TLR4 agonists110 and by a biglycan-TLR2 relay in VICs22. In this model, BMP2 and ALP induction by oxLDL was reduced by RNAi-mediated TLR2 “knockdown.” Mathieu confirmed biglycan signaling via TLR2 in human CAVD111. However, López and colleagues also showed that TLR4 agonists can activate human VICs112. The relative in vivo contributions of TLR2, TLR4, and other co-receptors in this pathway to CAVD remain to be delineated.
A key component of inflammation is intracellular oxidative stress, which produces signaling cascades that generate reactive oxygen species (ROS) such as hydrogen peroxide and superoxide. Miller, Heistad and colleagues were the first to demonstrate the pro-osteogenic, pathogenic role of ROS generation in CAVD113. Using a combination of human histochemistry, histopathology, and murine disease models, they showed that hydrogen peroxide generation downstream of NOS uncoupling played a vital role in disease biology. Moreover, they demonstrated that enzymatic defenses that dissipated oxidative stress were down-regulated in diseased valves113. In VSMCs and vascular myofibroblasts, hydrogen peroxide activates both osteogenic Cbfa1/Runx2114 and Msx2/Wnt107 signaling cascades to promote mineralization. Both of these regulatory cascades were shown by Miller et al to be activated in calcifying human aortic valves113. Liberman and colleagues noted elevated hydrogen peroxide levels adjacent to ectopic calcification in a rabbit model of CAVD115, independently confirming the relationship with oxidative stress. Moreover, oxLDL increases VIC production of Wnt3a, a morphogen that drives osteogenic differentiation via LRP5116. The VIC population arises in part from circulating hematopoietic stem cells56, 117; thus, proposed by Rajamannan116 an osteogenic “stem cell niche” -- replete with marrow elements7 -- may be ectopically created in aortic valves by the sustained paracrine Wnt signals elaborated in response to oxLDL.
Recently, Ferrari and colleagues reported an exciting new component of osteogenic ROS signaling in VICs118. They showed that hydroxgen peroxide activated the DNA damage response, with Akt activation mediating the upregulation of the osteogenic transcription factors Runx2 and Msx2.118 Transduction of VICs with adenovirus expressing catalase prevented osteogenic transcription factor induction118. Since hydrogen peroxide also mediates TNF induction of Msx2 and Wnt7b in calcifying vascular myofibroblasts107, modulation of ROS signaling and the osteogenic phase of the DNA damage response may provide new therapeutic options for both valve and vascular sclerosis. Moreover, because T2DM and the metabolic syndrome result in significant mitochondrial dysfunction that propagates cellular ROS accumulation119, targeting this osteogenic “intracrine” signaling cascade may prove to be most therapeutically effective in that clinical setting. These data converge with intriguing insights from Chau et al, demonstrating a role for BMP-Smad1 signaling in the DNA damage response120. Smad1 not only promotes osteogenic Msx2 and Runx2 gene expression121 but also functions as a potent activator of Runx2-directed transcription122. Thus, a potential “intracrine” signaling cascade emerges whereby the DNA damage response may promote osteogenic calcium accrual in VICs as relevant to CAVD (Figure 4).
Another morphogenetic signaling pathway that is central to bone formation, valve morphogenesis, and CAVD is the Jagged/Notch pathway. In the skeleton, Notchl signaling maintains proliferative expansion of bone marrow mesenchymal cells while inhibiting precocious osteoblast differentiation and depletion of this osteoprogenitor pool.123 In the developing heart, Jaggedl signals provided by ECs support Notch1-mediated EnMT necessary for cardiac valve morphogenesis124. Notch1 deficiency in mice and humans cause a spectrum of aortic valve diseases including bicuspid aortic valve and CAVD125, 126. Ex vivo, Notch suppresses osteogenic Runx2 signaling and mineralization in VICs74, 127, 128 . Moreover, Garg et al demonstrated that Notch1 sustains expression of Sox9 in VICs74, inhibiting osteogenic mineralization as predicted from the model of Yutzey (Figure 5). Of note, in this respect, VIC genomic responses to Notch1 activation diverge from those elicited in skeletal chondrocytes -- where Sox9 is a direct target of Notch1 and RBPJk-mediated suppression129. Whether Notch1 signaling is altered by metabolic and mechanical stimuli that promote cardiovascular calcification remains to be determined. However, haploinsufficiency for a RBPJk, a key transcriptional mediator of Notch1 signaling, predisposes to diet-induced CAVD in mice127.
As mentioned above, TGF-beta receptor I and ALK3 signaling -- activated by TGF-beta superfamily members -- is vital to the earliest stages of embryonic vasculogenesis, vascular remodeling, and cardiovascular morphogenesis. Even at this very earliest stage of vascular physiology, mechanical forces alter vascular remodeling and cellular potential directed by paracrine morphogen signaling130. Of note, VICs exhibit profound sensitivity to mechanical cues, selectively elaborating osteogenic potential under the correct mechanical environment131. To maintain a quiescent, non – osteogenic, non-myofibrogenic phenotype, VICs must experience a matrix stiffness below 10 kPa96. The mineralizing osteogenic phenotype is most apparent when VICs are exposed to a matrix stiffness of 25 kPa131, 132, a value that approximates that experienced by bone-forming osteoblast in unmineralized osteoid133. At a stiffness > 100 kPa, VICs undergo myofibroblastic differentiation with increases in apoptotic cell death and mineralization via apoptotic body calcification132. Upon adoption of the myofibroblast phenotype (smooth muscle cell alpha actin, type I collagen production), VICs become hypersensitive to TGF-beta signaling134, 135 -- and must experience a matrix stiffness under 10 kPa to resume the original quiescent state96. Of note, independent studies implementing cyclical stretch of aortic valve cusps confirm that mechanical forces enhance VIC responses to TGF-beta136.
Unfortunately, most of our ex vivo model systems for studying VIC functions utilize culture on plastic and glass – and expose cells substrate stiffness levels of 10,000 kPa or greater131. Thus, much of our understanding of VIC physiology and signaling has been performed under conditions that mimic the mechanical microenvironment experienced by VICs in advanced CAVD; calcium phosphate concretions and mineralized bone exhibit this degree of mechanical stiffness. More sophisticated culture approaches will be required to provide an integrated understanding of how mechanical environment, cell-cell interactions, and the neuroendocrine/metabolic milieu interact to increase the risk for CAVS and CAVD progression to clinically significant CAS131. The reader is referred to an outstanding review in this series, authored by Anseth and colleagues, that details the interplay between valve biomechanics and the pathobiology of CAVD137.
As mentioned above, the endothelial-mesenchymal transition (EnMT) plays a central role in cardiac valve development and the generation of fibroblasts from epicardium during ischemic repair87. Studies by Bischoff, Aikawa, Schoen, and colleagues first demonstrated co-expression of CD31 and smooth muscle cell alpha-actin in a subset of ovine VICs16. They went on to show that human pulmonary VEC were capable of responding to TGF-beta to elaborate a VIC phenotype. Human pulmonary VEC and ovine aortic VEC clones behaved similarly, with EnMT downstream of TGF-beta dependent upon Notch signaling65. The osteogenic potential of clonal ovine VEC populations was also clearly demonstrated15. Ex vivo, mechanical strains on the order of 10% to 20% were shown to enhance EnMT via TGF-beta/Smad and Wnt/beta-catenin signaling cascades, respectively14. Strains applied orthogonal to anisotropic VEC alignment (mediated by extracellular fibronectin) exerted the greatest stimulus for EnMT, as functionally assessed by development of a contractile response to endothelin14. The extent to which EnMT contributes to CAVD in vivo has yet to be determined. Intriguingly, activating mutations in human ALK2 have been demonstrated to promote heterotopic bone formation in fibrodysplasia ossificans progressiva – occurring in great part via osteoprogenitor development from the endothelial – mesenchymal transition138. However, while these patients develop massive ectopic bone deposits in muscle in association with minimal trauma, cardiac valve calcification is not characteristic of this disorder139.
Clearly, CAVD is both common and complicated. While we know so much more about the pathobiology than we did a decade or so ago, we still lack key clinical tools and non-surgical therapeutic options. Novel diagnostics, biomarkers, and therapeutic strategies are needed. Not everyone with echocardiographically - defined CAVS progresses to end-stage CAS that requires aortic valve replacement surgery. Although we know some of the risk factors, including echocardiographic estimates of valve calcium load140, in early stage disease we cannot predict with any certainty who will and will not progress. Development of robust biomarkers for disease progression will not only provide the clinical tools necessary for patient risk stratification and assessment of response to therapeutic intervention, but may also illuminate novel pathways that can be targeted medically to prevent disease progression. The need to capture this risk of disease progression is particularly acute in patients with bioprosthetic valves6 – but may be fundamentally different from patients with native valve sclerosis11. The seminal realization by Demer that oxylipids and cellular ROS signals are common to both bone loss and arterial calcium accrual141, 142 suggests that cellular and/or molecular signatures indicative of vascular and skeletal oxidative stress will prove useful. Moreover, a “point of no return” will exist where medical strategies will prove to be insufficient1. This likely contributed to the failure of statin-based strategies to mitigate CAVD progression in several large studies143, 144 even though preclinical and epidemiological studies indicate the important role for cholesterol metabolism in CAVD pathogenesis1. Being able to identify early on those patients that are at greatest risk -- and who is and is not a candidate for medical intervention -- would be of great clinical utility. While embarking on the quests for biomarker and medical therapeutics, it must be remembered that discoveries forthcoming from studies of arterial calcification may or may not be applicable to CAVD; the vascular histoanatomic distinctions between CAVD, atherosclerotic calcification, medial calcification, calcific uremic arteriolopathy, and the mineral and bone disorder of chronic kidney disease portend overlapping yet distinct disease biology145.
BMP and TGF-beta signaling clearly participate in vascular calcium metabolism, including CAVD1, 9. Future studies will potentially assess the capacity of biologic vs. small molecule modulation of ALK2/ALK3 signaling to safely limit CAVD as an extension of preclinical studies of vascular sclerosis94– and may help delineate the relative contributions of EnMT vs. circulating osteoprogenitor cells in CAVD initiation and progression. However, the potential risks associated with targeting the ALK receptor kinases became apparent with the ALK5 inhibitors97, indicating that more pharmacological tools and studies are required. Moreover, Kalluri very recently demonstrated that small peptide agonists for ALK3 actually promoted the reversal of renal fibrosis and renal regeneration146. Thus, technologies are needed to ensure tissue – specific targeting of agents modulating morphogenetic signaling pathways. Of note, in a model of uremic cardiovascular calcification, Aikawa and colleagues demonstrated that inhibition of cathepsin S could mitigate progressive CAVD147. Because elastin fragments liberated by cathepsin S protease activity are bioactive peptides that promote osteogenic differentiation in concert with TGF-beta148, novel strategies such as these may afford effective therapeutic mechanisms with acceptable safety profiles.
A few important clues lie within the clinical literature with respect to potential medical strategies for CAVD treatment. Older patients with osteoporosis treated with amino-bisphosphonates experience less aortic valve and aortic calcium accrual149. Moreover, bisphosphonates “reprogram” the circulating osteoprogenitors, suggesting that certain osteotropic drugs can favorably impact the bone-vascular axis150, 151. However, younger women treated with bisphosphonates experience increases in aortic valve calcium load149, indicating that the rate-limiting mechanisms regulating vascular calcium metabolism change with age. Once again, a better understanding of the mechanisms controlling the initiation and progression of CAVS is needed. As the morphogenetic, metabolic, mechanical, inflammatory, and neuroendocrine regulation of CAVD takes shape from ongoing research, new diagnostic and therapeutic strategies will emerge that will better help us address the needs of our patients with valvular heart disease.
D.A.T. is supported by grants HL088651, HL088138, HL069229, and HL114806 from the National Institutes of Health, and transitional research funds from the Sanford-Burnham Medical Research Institute at Lake Nona, FL.
Disclosures – D.A.T. serves as a consultant for Merck and Co.