Unraveling the Mechanisms of Valvular Heart Disease to Identify Medical Therapy Targets: A Scientific Statement From the American Heart Association

Abstract

Valvular heart disease is a common cause of morbidity and mortality worldwide and has no effective medical therapy. Severe disease is managed with valve replacement procedures, which entail high health care–related costs and postprocedural morbidity and mortality. Robust ongoing research programs have elucidated many important molecular pathways contributing to primary valvular heart disease. However, there remain several key challenges inherent in translating research on valvular heart disease to viable molecular targets that can progress through the clinical trials pathway and effectively prevent or modify the course of these common conditions. In this scientific statement, we review the basic cellular structures of the human heart valves and discuss how these structures change in primary valvular heart disease. We focus on the most common primary valvular heart diseases, including calcific aortic stenosis, bicuspid aortic valves, mitral valve prolapse, and rheumatic heart disease, and outline the fundamental molecular discoveries contributing to each. We further outline potential therapeutic molecular targets for primary valvular heart disease and discuss key knowledge gaps that might serve as future research priorities.

Valvular heart disease (VHD), encompassing stenosis or regurgitation of the aortic, mitral, tricuspid, or pulmonary valves, causes significant cardiovascular morbidity and mortality worldwide. In 2017, the Global Burden of Disease Study reported an estimated 1.5 million disability-adjusted life-years for calcific aortic valve disease (CAVD), 0.87 million disability-adjusted life-years for degenerative mitral valve disease, and 0.14 million disability-adjusted life-years for other nonrheumatic VHDs.^1,2 Compared with other common cardiovascular conditions, the relative proportion of cardiovascular deaths attributable to VHD is small: 2.5% of cardiovascular deaths in 2019 compared with 49.2% of cardiovascular deaths due to ischemic heart disease, 17.7% of cardiovascular deaths due to ischemic stroke, and 6.2% due to hypertensive heart disease.³ However, in contrast to atherosclerotic cardiovascular disease, there are no available medical therapies to prevent or slow the progression of VHD. Instead, severe VHD is addressed with surgical or transcatheter valve replacement or repair procedures. With an aging population, morbidity and health care–associated costs attributed to VHD and related procedures continue to rise, highlighting an urgency to identify molecular targets for medical therapy.⁴ The purpose of this scientific statement is to outline the most current understanding of molecular and cellular mechanisms contributing to common causes of native VHD, to identify gaps in our understanding of VHD, and to discuss the next steps needed for the development of noninvasive therapies.

CARDIAC VALVE STRUCTURE AND DEVELOPMENTAL ORIGINS

The complex anatomy of the normal semilunar (aortic and pulmonic) and atrioventricular (mitral and tricuspid) valves allows unrestricted forward blood flow and prevents backflow with the changing intracardiac pressures resulting from ventricular contraction (Figure 1).⁵ At the tissue level, valve leaflets are made up of 3 layers: the collagen-rich fibrosa, the proteoglycan-rich spongiosa, and the elastin-rich ventricularis (in the semilunar valves) or atrialis (in the atrioventricular valves).^6–8 Normal aortic valves maintain circumferentially oriented collagen in the fibrosa, which provides tensile strength, and radially oriented elastic fibers in the ventricularis, which facilitate optimal coaptation of the aortic valve leaflets (Figure 2).^9,10 Normal mitral valves are translucent, smooth structures with a thin but dense central region of the leaflets with networks of collagens secreted by vimentin-positive fibroblasts in response to mechanical stimulation (Figure 3). The edge of the mitral valve leaflets is thick and rich in glycosaminoglycans, and collagen-rich chordae tendineae with high tensile strength connect the mitral valve leaflets to papillary muscles of the left ventricle, preventing prolapse into the left atrium.¹¹

Figure 1. Anatomical location of heart valves in diastole.

Netter illustration used with permission of Elsevier Inc. All rights reserved. www.netterimages.com

Figure 2. Semilunar valve leaflet structure.

A, Heart valves are coated in a layer of endothelial cells that encapsulate the valve interstitial cells in 3 layers defined by the extracellular matrix content. The ventricularis (V) is the elastin-rich layer of these structures connecting to the major arteries; the middle layer is the proteoglycan-rich spongiosa (S); and the outer layer is the collagen-rich fibrosa (F). Created with BioRender.com. B, Movat pentachrome staining of a calcified human aortic valve. Calcification is shown in purple. Bottom, Images of valves from a control and a fibrotic porcine valve. B, Reprinted with permission from Chen and Simmons.⁹

Figure 3. Atrioventricular valve leaflet structure.

A, The tricuspid and mitral valves bridge the chambers of the heart and are thus referred to as atrioventricular valves. Blood flows from the atrium to the ventricle. The atrialis abutting the atrium contains elastin fibers; the middle layers, the proteoglycan-rich spongiosa; and the outer layer, the collagen-rich fibrosa. Created with BioRender.com. B, Movat pentachrome stain (left; collagen in yellow, proteoglycans in blue-green, elastin in black) of a normal mitral valve and mitral valve with myxomatous degeneration. Picrosirius red staining (right) of normal and myxomatous mitral valves viewed under polarized light illustrates disruption birefringence of collagen fibers in myxomatous leaflets. Magnification ×100. B, Reprinted with permission from Rabkin et al.⁸

All cardiac valves maintain a layer of valvular endothelial cells (VECs), which are contiguous with the myocardial or aortic endothelium. VECs are derived from embryonic cardiac cushion endothelial cell populations. Valve interstitial cells (VICs) are interspersed throughout the 3 valve layers and are fibroblast-like cells originating from multiple embryonic sources, including the endocardial cushions, neural crest, and epicardium.¹² After birth, the primitive valves remodel into trilaminar leaflets as VIC subpopulations exposed to differential mechanical forces express compartment-specific extracellular matrix (ECM) genes.¹³ In mature valves, VICs reside in a quiescent state but can become activated in the setting of injury to facilitate repair.⁸ Activated VICs, sometimes called myofibroblastic, contribute to many common VHDs, including CAVD and mitral valve prolapse (MVP).¹² Recent single-cell analysis, phenotype-guided omics profiling, and network-based studies demonstrate that VICs are even more heterogeneous than previously thought, identifying specific disease-driver VIC populations capable of osteogenic differentiation in human CAVD.¹⁴

DISEASES OF THE AORTIC VALVE LEAFLETS

Definitions and Spectrum of Pathogenesis

Primary diseases of the aortic valve leaflets result in aortic stenosis (AS) and aortic regurgitation. The most common pathology of the aortic valve is CAVD, which encompasses the pathological fibrocalcific remodeling of the valve leaflets, resulting in calcific AS. Rheumatic heart disease (RHD) remains the highest prevalent VHD in low-income countries.¹ RHD invariably affects the mitral valve leaflets, but aortic valve involvement also is seen in 20% to 30% of cases. Congenital aortic valve abnormalities such as a unicuspid aortic valve or bicuspid aortic valve (BAV) typically result in earlier onset of calcific AS or less frequently in aortic regurgitation.

Calcific Aortic Stenosis

Classification and Epidemiology

AS is the most common VHD, with an estimated age-standardized prevalence of 116.3 cases of AS per 100 000 individuals according to the Global Burden of Disease 2019 study.³ This statistic is expected to rise with an aging population; the global prevalence of AS increased 124% between 1990 and 2017 alone.¹ Traditional risk factors for AS include older age; male sex; hypertension; type 2 diabetes; dyslipidemia, in particular elevated lipoprotein(a) [Lp(a)]; smoking; elevated body mass index; and end-stage renal disease.^10,15 Data are inconsistent on whether dietary vitamin D or calcium modifies risk of AS,¹⁶ with some data suggesting that supplemental calcium increases mortality among individuals with mild to moderate AS.¹⁷

The first sign of pathology in CAVD, aortic sclerosis, presents as leaflet thickening, representing focal lipid infiltration and inflammation, which is observable by ultrasound imaging. As the area of leaflet calcification enlarges, blood flow across the valve is obstructed, resulting in elevated flow velocities across the valve by Doppler ultrasound. One quarter of the population >65 years of age is estimated to have aortic sclerosis, and roughly 1.8% of individuals with aortic sclerosis progress to AS annually. In fact, the majority of individuals with aortic sclerosis never develop hemodynamically significant AS during their lifetime.¹⁸ Clinical risk factors for incident aortic sclerosis or aortic valve calcification are similar to those for AS and include dyslipidemia, the metabolic syndrome, smoking, and elevated body mass index. However, once aortic sclerosis is apparent, it remains unclear which risk factors promote faster progression to AS, and traditional risk factors, including Lp(a), are not clearly associated with progression of aortic valve calcification.^19,20 These data suggest that disease initiation and disease progression are likely different processes in CAVD, with resulting implications for timing and choice of medical therapy.

Molecular Mechanisms

The unique pathology manifest in CAVD is likely precipitated by a combination of shear stress, related to transvalvular blood flow, and mechanical stress, related to the aortic to left ventricular pressure difference and to valve opening and closing. These processes result in disruption of the VEC layer, which is a nidus for disease-initiating mechanisms within the leaflet, including inflammation, lipid infiltration, oxidative stress, eventual activation of fibrotic and osteogenic transcriptional programs, and the release of procalcifying extracellular vesicles (EVs), which synergistically promote pathological mineralization of the ECM surrounding the VICs (Figure 4). Mineralization typically starts in the fibrosa but expands to invade all 3 layers, with the noncoronary cusp generally the first to undergo adverse remodeling.^9,21 Disruption of the native ECM composition, whether the result of mechanical/shear stress or cell-specific mechanisms, affects aortic valve cellular phenotypes and function, which furthers ECM remodeling and consequent fibrocalcific changes and stiffening.^9,22–24 Fragmented ECM fibers additionally serve as loci for the nucleation of calcium and phosphate and the accumulation and aggregation of EVs.

Figure 4. Mechanisms driving calcification of the aortic valve.

Although clinical trials to date have not shown that statin therapy is beneficial for aortic stenosis/sclerosis, lipids contribute to the osteogenic transition of valve interstitial cells. Low-density lipoprotein (LDL) and lipoprotein(A) [Lp(a)] enter the fibrosa, triggering an inflammatory cascade that leads to increased oxidative damage and cellular death. Secreted cytokines act in a paracrine manner to trigger the activation of osteogenic transcriptional programs in valvular interstitial cells. Apoptotic bodies and necrotic cells may also serve as loci for the nucleation of minerals. An intact extracellular matrix (ECM) is critical for leaflet function, and the breakdown of ECM proteins not only prevents proper leaflet movement but also activates mechano-sensing pathways that stimulate osteogenic transcriptional programs. Damaged ECM also captures secreted vesicles that, when released from activated cells, contain the enzymes necessary to produce the calcium and phosphate building blocks of minerals, as well as microRNAs that may further alter cell homeostasis. Amyloid plaques can also accumulate along these broken ECM components. αSMA indicates α-smooth muscle actin; IFN, interferon; IL, interleukin; LPA, lysophosphatidic acid; MMP, matrix metalloproteinase; NF-κB, nuclear factor-κB; oxLDL, oxidized low-density lipoprotein; RANKL, receptor activator of nuclear factor-κB ligand; TGFβ, transforming growth factor-β; TNF-α, tumor necrosis factor-α; and VIC, valve interstitial cell. Created with BioRender.com.

Lipid deposits are observed in early pathological valve specimens, and the resulting inflammatory response is considered an early disease-initiating event in CAVD.^25,26 Strong epidemiological and genetic evidence demonstrates a causal role for many distinct apolipoprotein B cholesterol particles in CAVD, including low-density lipoprotein cholesterol (LDL-C),^27,28 remnant cholesterol,²⁹ and Lp(a).^26,30 Multivariable mendelian randomization of these apolipoprotein B particles suggests that LDL-C and Lp(a) may have independent causal roles in AS, although the association between LDL-C and AS is attenuated when also accounting for Lp(a). The association between triglycerides and AS becomes insignificant when accounting for other lipids.²⁷ The mechanisms by which Lp(a) promotes disease in AS are hypothesized to occur through the actions of oxidized phospholipids, which are covalently linked to Lp(a) and promote gene expression programs, resulting in calcific transformation of VICs.¹⁰ Supporting evidence includes robust data on autotaxin, an enzyme that metabolizes the high lysophosphatidylcholine content in Lp(a) associated with oxidized phospholipids. ATX gene expression is elevated in CAVD VICs; autotaxin is associated with circulating Lp(a); and autotaxin overexpression promotes mineralization of aortic valve tissue in murine models.³¹ Although these data strongly support a role for lipids and lipid metabolism in early CAVD pathogenesis, to date, lipid-modulating therapies such as statins have not shown benefit in attenuating CAVD progression.³²

Multiple lines of evidence suggest that inflammation, along with and provoked by valvular lipid deposition, is also an early initiating event in CAVD. Inflammation is proposed to disrupt VIC and VEC homeostasis, promoting these cells to acquire osteoblast-like features and calcify.^9,33–38 Prior genome-wide association studies (GWASs) of common genetic variants in calcific AS implicate intronic single-nucleotide polymorphisms in the IL6 gene, encoding the proinflammatory cytokine interleukin-6,^27,28,39 which is a downstream cytokine of the NLRP3 inflammasome complex. Further supportive evidence relates to the observation that certain somatic variants in hematopoietic stem cells, a condition called clonal hematopoiesis of indeterminate potential, occur frequently in CAVD and portend worse prognosis after aortic valve replacement.^40,41 Clonal hematopoiesis of indeterminate potential promotes cardiovascular risk by augmenting NLRP3 inflammasome-mediated gene programs.⁴² Inflammation in CAVD also extends to innate immunity. Interferonopathies are a class of genetic diseases causing systemic elevation of type 1 interferons, which mimics constitutively active damage associated with molecular pattern signaling.⁴³ Several interferonopathies, including Singleton-Merten syndrome and ADAR-related type 1 interferonopathy, are associated with early and extreme aortic valve calcification phenotypes.^44–48

EVs participate in both physiological and pathological processes and are emerging mediators in the pathogenesis of CAVD, particularly in initiation and progression of calcification. Depending on size and type, EVs are classified as exosomes, microparticles, or apoptotic bodies. They are released from various cell types, including cells residing in valvular leaflets (eg, VECs, VICs, macrophages). An outer membrane of EVs protects their cargo, consisting of bioactive molecules, proteins, enzymes, microRNA, and other components of the parental cell. EVs can participate in cell-cell communications and act as messengers to distant tissues to control homeostasis and systemic responses. Several pathways have been implicated in EV calcification. First, sortilin, a multiligand sorting receptor also strongly associated with LDL-C metabolism⁴⁹ and a genome-wide significant risk locus in multiple GWAS for AS,^27,28,50 is demonstrated to contribute to the EV loading of the calcification enzyme TNAP (tissue-nonspecific alkaline phosphatase, ALPL gene) through intracellular trafficking mechanisms.⁵¹ When released from cells, EVs containing TNAP can then be trapped within collagen fibers and other ECM components and aggregate through annexin-1.⁵² Minerals generated from TNAP activity nucleate to form microcalcifications of hydroxyapatite,^53,54 which merge and progress to larger macrocalcifications that can be visualized with clinical imaging modalities.⁵⁵ The role of EVs in VHD and CAVD is largely understudied; moreover, explanted bioprosthetic heart valves have not yet been examined for the presence of EVs.

Potential Therapeutic Targets

There is notable overlap in clinical risk factors, epidemiology, and pathobiology of CAVD and atherosclerosis. However, multiple clinical trials of traditional preventive therapies for cardiovascular disease, including statins, have failed when applied to AS,^32,56 and no medically efficacious treatments remain. Robust prior genetic and experimental studies in CAVD, as described in this scientific statement, have generated numerous molecular targets and pathways (Table 1),^{1–3,6,13,21,23,26,35,38,42–46,52,56–64,67–70,72–95} which may, we hope, provide an opportunity for targeted drug development (reviewed in depth elsewhere^9,96). A handful of ongoing trials of medical therapy are targeting the progression of AS (Table 2). Perhaps most exciting is the recent success of Lp(a)-modifying therapies such as the siRNA molecule olpasiran, which effectively lowered plasma Lp(a) in a phase 2 clinical trial.⁹⁷ Active long-term studies will assess the impact of Lp(a)-modifying therapies on AS progression and outcomes (NCT05646381 [pelacarsan]). Additional opportunities for drug development exist in targeting pathways that are orthogonal to those apparent in atherosclerosis. These include mineral nucleation,⁹⁸ the NOTCH pathway signaling,⁵⁷ the renin-angiotensin-aldosterone signaling pathway,⁹⁹ and dipeptidyl peptidase 4,¹⁰⁰ among others. In an era of large, multiomics CAVD datasets on human DNA, RNA, and proteins,¹⁰¹ efforts might prioritize studying pathways and molecules that are expressed in the valve and have few putative off-target effects.

Table 1.

Molecular pathways in VHD and Potential Treatment Targets

Molecular pathway(s)	Potential medical targets	References
CAVD
Mechanical forces and shear stress	TGF-β, Notch1, cadherin-11	1–3, 21, 23, 35, 57–59
ECM	ECM protein organization, MMPs	13, 60
EVs	Annexin-A1	52, 61
Epigenetic reprogramming	TERT/STAT5, Notch1, ac–TGF-β1/H3K9	62–64
Lipid metabolism	Lp(a)/autotaxin, LDL-C, Lrp5	26, 27, 30, 67, 68
Oxidative stress	Impaired antioxidant enzymes	69, 70
Inflammation	IL-18, interferon signaling, IL-6, clonal hematopoiesis of indeterminant potential	26, 27, 38, 40, 42–46, 72
X-chromosome inactivation	BMX and STS	73
BAV
Notch1, eNos, Gata4/5/6, Smad6	Not medically accessible for congenital malformations	74, 75
MVP
ECM	ECM protein organization, MMPs	13, 60
Elastin deficit or fragmentation	MMPs (ie, MMP-1, MMP-2, MMP-13)	6, 76
TGF-β pathway	TGF-β pathway proteins	77, 78
Ciliogenesis	DZIP1	79
Cytoskeleton biology	FLNA, TNS1	77, 80
Cell polarity	DCHS1	81
Calcineurin/NFAT	Calcineurin inhibitors, LMCD1	82, 83
Rheumatic valve disease
Proinflammatory cytokines, adhesion molecules	IFN-γ, IL-17, TNF, TGF-β, VCAM-1	84–91
Ficolins/MBL pathway	MBL pathway of complement activation	89, 90, 92–95

BAV indicates bicuspid aortic valve; BMX, cytoplasmic tyrosine-protein kinase BMX; CAVD, calcific aortic valve disease; DCHS1, dachsous cadherin-related 1; DZIP, DAZ interacting zinc finger protein 1; ECM, extracellular matrix; EV, extracellular vesicle; FLNA, filamin A; H3K9, histone 3 lysine 9; IFN-γ, interferon-γ; IL, interleukin; LDL-C, low-density lipoprotein cholesterol; LMCD1, LIM and cysteine rich domains 1; Lp(a), lipoprotein(a); Lrp5, low-density lipoprotein receptor-related protein 5; MBL, mannose-binding lectin; MMP, matrix metalloproteinase; MVP, mitral valve prolapse; STAT5, signal transducer and activator of transcription 5; STS, steroid sulfatase; TERT, telomerase reverse transcriptase; TGF, transforming growth factor; TGF-β, transforming growth factor beta; TNF, tumor necrosis factor; TNS1, tensin 1; VCAM-1, vascular cell adhesion molecule 1; and VHD, valvular heart disease.

Table 2.

Ongoing Clinical Trials Evaluating Medical Therapies for Primary VHD

Disease	Target	Drug	Trial name	Enrollment criteria	Enrollment (n)	Primary outcome	NCT number
AS	Endothelial dysfunction	Evogliptin (DPP4 inhibitor)	EVOID-AS (A Study to Evaluate the Efficacy and Safety of DA-1229 (Evogliptin) in Patients With Mild to Moderate Calcific Aortic Valve Disease With Mild to Moderate Aortic Stenosis)	Mild to moderate AS	867	Change in aortic valve calcium score by CT scan	NCT05143177
AS	Lp(a)	Pelacarsen	A Multicenter Trial Assessing the Impact of Lipoprotein(a) Lowering With Pelacarsen (TQJ230) on the Progression of Calcific Aortic Valve Stenosis	Mild to moderate AS with Lp(a) ≥50 mg/dL	502	Change in peak aortic valve velocity; change in aortic valve calcium score by CT scan	NCT05646381
AS	LDL-C	PCSK9 inhibitor	Effect of PCSK9 Inhibitors on Calcific Aortic Valve Disease	Mild to moderate AS with moderate to high cardiovascular risk and LDL-C >70 mg/dL on statin therapy or Lp(a) >50 mg/dL	160	Change in aortic valve calcium score by CT scan	NCT04968509
AS	Inflammation	Colchicine	CHIANTI (Colchicine and Inflammation in Aortic Stenosis)	Moderate AS	150	Change in aortic valve calcium score by CT scan	NCT05162742
AS	Renin-angiotensin system	ARBs	ARBAS (Angiotensin Receptor Blockers in Aortic Stenosis)	Mild to moderate AS and normal LVEF (LVEF ≥50%)	144	Change in aortic valve calcium score by CT scan	NCT04913870
AS	PPAR-γ	Pioglitazone	Effects of Pioglitazone in Calcific Aortic Valve Disease	Mild to moderate AS	100	3-y mortality	NCT05875675
AS	Inflammation	Colchicine	COPAS pilot (Effects of Colchicine on the Progression of Aortic Valve Stenosis: A Pilot Study)	Mild to moderate AS	24	Change in aortic valve calcification activity measured by NaF uptake on PET/CT at 6 mo	NCT05253794
RHD	Fibrosis	Dapagliflozin	Dapa-Rhemis (Dapagliflozin Effect on Rheumatic Mitral Stenosis)	Severe mitral stenosis, heart failure NYHA class II-III	36	Mitral valve mean pressure gradient, Kansas City Cardiomyopathy Questionnaire	NCT05618223
MVP	…	…	…	…	…	…	…
BAV	…	…	…	…	…	…	…

ARB indicates angiotensin receptor blocker; AS, aortic stenosis; BAV, bicuspid aortic valve; CT, computed tomography; DPP4, dipeptidyl peptidase 4; LDL-C, low-density lipoprotein cholesterol; Lp(a), lipoprotein(a); LVEF, left ventricular ejection fraction; MVP, mitral valve prolapse; NYHA, New York Heart Association; PCSK9, proprotein convertase subtilisin/kexin type 9; PET, positron emission tomography; PPAR-γ, peroxisome proliferator-activated receptor-γ; RHD, rheumatic heart disease; and VHD, valvular heart disease.

CONGENITAL BICUSPID AND UNICUSPID VALVES

Classification and Epidemiology

Congenital malformation of the aortic valve can result in a BAV, the most common valve malformation with a prevalence of ≈1.5%, or a unicuspid valve, which is much less common but likely underdiagnosed because differentiation from a bicuspid valve can be challenging.¹⁰² The 2 leaflets of the BAV result from fusion or absence of valve leaflet progenitors during development and have distinct classifications based on orientation of the leaflets.¹⁰³ BAV is often diagnosed in adulthood because congenitally malformed aortic valves typically function normally until leaflet opening is limited by superimposed fibrocalcific changes. However, patients with a congenital BAV are predisposed to aortic valve dysfunction at a younger age than people with trileaflet aortic valves, and almost all will require valve intervention during their lifetime.¹⁰⁴ Overall, approximately half of all individuals undergoing valve replacement for AS have a BAV, with a higher prevalence of BAV disease in patients <60 years of age.¹

Molecular Mechanisms

Genetic linkage studies have reported evidence of heritability of BAV in up to 89% of cases,^75,105 but in most cases, the genetic basis has not been identified, suggesting complex pathology. Heritable BAV has been linked to Turner, Marfan, and Loeys-Dietz syndromes, and rare monogenic variants have been identified in NOTCH1, GATA4/5/6, and SMAD6, as well as in genes associated with primary cilia and endothelial mesenchymal transition.^74,75 Regardless of initial causation, BAV disease includes structural abnormalities in the valve leaflets, sometimes associated with aortic dilation, leading to abnormal flow patterns and abnormal stressstrain of the valve leaflets. These alterations in valve leaflet structure and biomechanical stress likely contribute to accelerated leaflet degeneration and calcification in patients with a bicuspid compared with trileaflet aortic valves.^74,106,107

Gaps in Our Understanding

Many knowledge gaps in the pathobiology of primary aortic valve disease remain that may serve as future research priorities (Table 3). Of particular interest is the study of early CAVD, or aortic sclerosis. Although not yet clinically actionable, ≈25% of individuals ≥65 years of age and 50% of individuals ≥80 years of age have aortic sclerosis.¹⁸ Emerging data illustrate that aortic sclerosis is a biologically active disease state. ¹⁸F-sodium fluoride (¹⁸F-NaF) PET imaging, which is a marker of early calcification, demonstrates increased uptake among valves of individuals with aortic sclerosis compared with normal valves,¹⁰⁸ and new calcium deposition is observed to occur at areas of increased ¹⁸F-NaF uptake on serial scans over 2 years. It is unknown why some individuals progress more rapidly from aortic sclerosis to hemodynamically significant AS. Future research might focus on clinical and genetic factors predicting the onset or progression of aortic sclerosis and AS. Although the risk factors associated with AS are similar to those for general cardiovascular disease, the acute initial events that trigger a valve cell to reprogram to a calcifying state are not fully understood. In vitro disease modeling illustrates that stimuli such as biomechanical stress, inflammation, and oxidative stress promote the osteogenic reprogramming of VICs.^109,110 Moving forward, understanding the mechanisms by which these stresses initiate calcification programs and identifying novel pathways driving the transition of healthy valve cells into calcifying ones could subsequently be leveraged for the development of noninvasive therapies to prevent, halt, or even reverse valve calcification.

Table 3.

Knowledge Gaps in the Molecular Mechanisms of VHD

Knowledge gap	Specific research questions	Suggested research approach
CAVD
Mechanisms underlying sex differences	What is the role of hormones in driving osteogenic differentiation? Are other mechanisms besides X chromosome inactivation at play?	Large human studies with biological material that are adequately powered to study men and women separately
Calcification initiation and progression	Why does the noncoronary cusp calcify first? Why does calcification initiate in the fibrosa? Do progenitor-type cells reside in the valve leaflets or sinus?	Longitudinal population-based studies with serial imaging and biosample collection (ie, plasma and DNA)
Ex vivo, in vivo models that recapitulate human CAVD pathogenesis and progression	What materials would support valve organoid development? Can we humanize a mouse model of aortic valve calcification to more accurately mimic a human phenotype?	Funded basic science infrastructure supporting the development of organic and inorganic models of valvular heart tissue
BAV disease
Developmental mechanisms of valve leaflet abnormalities	Are there preserved developmental pathways mediating congenital valve disease?	Longitudinal examination of mouse models of BAV with high penetrance of abnormalities Evaluation of rare loss-of-function variations in relevant developmental pathways in large human populations
Pathogenesis of BAV relative to trileaflet aortic valve calcific disease	Is the pattern of calcification observed in bicuspid AS the same as observed in trileaflet AS?	Hemodynamic/molecular studies of BAVs vs trileaflet aortic valves
MVP
Mechanisms driving the initiation and evolution of myxomatous disease vs fibroelastic deficiency	What are the cell (ie, VIC/VEC) and gene pathways differentiating classic myxomatous disease from fibroelastic deficiency?	Basic studies with access to human tissue and cell culture from both normal and diseased (Barlow disease and fibroelastic deficiency) samples
RHD
Mechanisms underlying sex differences in disease	Why is the prevalence of RHD higher in women?	Study of epidemiological and biological factors that may predispose women to disease (eg, exposure to streptococcus, access to antibiotics, biological variables)
Mechanisms driving left-sided valve predilection	Why are left-sided heart valves more likely to be affected by RHD?	Basic studies focused on physiological differences between left- and right-sided heart valves (pressure, shear, oxygen tension) and how those predispose to inflammation
Mechanisms driving the initiation and evolution of myxomatous disease vs fibroelastic deficiency	What are the cell (ie, VIC/VEC) and gene pathways differentiating classic myxomatous disease from fibroelastic deficiency?	Basic studies with access to human tissue and cell culture from both normal and diseased (myxomatous and fibroelastic deficiency) samples

BAV indicates bicuspid aortic valve; CAVD, calcific aortic valve disease; MVP, mitral valve prolapse; RHD, rheumatic heart disease; VEC, valvular endothelial cell; VHD, valvular heart disease; and VIC, valve interstitial cells.

AS manifests notable epidemiological differences between certain populations without clear biological explanation. Men with CAVD tend to calcify more often than women, who develop a greater degree of valvular fibrosis.¹¹¹ This is clinically manifest by women having more severe hemodynamic AS with less aortic valve calcification on imaging.¹¹¹ We do not know whether CAVD pathogenesis differs between men and women from disease onset or whether there is divergence in disease progression.¹¹² Using porcine and human VICs, investigators found that X-chromosome inactivation contributed to sex differences, whereby several genes escaping X-chromosome inactivation increased myofibroblast activation preferentially in women.⁷³ BAV is similarly differentially prevalent between sexes; at birth, BAV is more prevalent in male babies (7.1/1000 individuals) than female babies (1.9/1000 individuals), and this 3- to 4-fold difference in prevalence persists to adulthood. The underlying biology explaining this difference remains unknown; however, reduced dosage of genes that escape X-chromosome inactivation has been proposed as a contributing mechanism.¹¹³ In addition, non-White individuals have a lower reported prevalence of AS than White individuals,¹¹⁴ although studies are often biased by differential VHD-related health care interactions for non-White populations.¹¹⁵ To date, most of the large population studies on CAVD have focused on White individuals, and GWASs have not been adequately powered to pursue sex-stratified analysis. Future research efforts should target diverse populations so that clinical and biological insights can best represent the general population.

Equally important to our understanding of CAVD pathogenesis is the development of ideal models that recapitulate the human condition. Lineage-tracing murine models demonstrate that mouse valve leaflets do not calcify in the same way as human leaflets.¹¹⁶ Murine aortic valves contain pigment-producing melanocytes,¹¹⁷ which are easily confused for positive von Kossa staining (which identifies calcium deposits) because both appear as brown precipitates. Melanocytes in VICs help to produce the elastin of the leaflet^118,119; whether a similar population exists in humans in unknown. Ovine, porcine, and bovine models better recapitulate the CAVD observed in humans, but large-animal models entail high cost and space requirements.¹²⁰ The field of bioengineering has made excellent use of biomaterial to recapitulate valve leaflet stiffness, for example, with 3-dimensional bioprinting, which enables evaluation of mechanosensitive cellular responses,^121,122 but we do not yet have a suitable tool for modeling valve disease ex vivo.

PRIMARY MITRAL VALVE DISEASE

Definitions and Spectrum of Pathogenesis

Primary diseases of the mitral valve result in mitral stenosis and mitral regurgitation (MR). The most common cause of mitral valve disease worldwide is rheumatic mitral stenosis.¹ The mitral valve and the annular ring are also subject to calcification, which may lead to stenosis and accumulation of fibrosis in the context of atherosclerotic disease burden. However, primary calcific mitral stenosis is rare.¹²³ The most common cause of primary MR is myxomatous valve disease, also called MVP.

Myxomatous Mitral Valve Disease

Classification and Epidemiology

MVP is charactered by redundant and thickened valve leaflets with abnormal sagging of 1 or both of the leaflets into the left atrium during systole.¹²⁴ MVP is common, with an estimated prevalence of 1 in 40 individuals. Most patients never develop significant regurgitation or require intervention.¹²⁵ However, in up to 25% of cases, MVP leads to significant MR,¹²⁶ sometimes requiring surgical repair, generally as a result of gradually increased leaflet prolapse and regurgitation over many years. MVP can rarely present as sudden rupture of chordae tendineae, resulting in acute severe regurgitation.

Molecular Mechanisms

Suspected triggering mechanisms of MVP are partially genetically determined and may be accelerated in response to mechanical stress and consequences of normal aging in the mitral valve, which include a decrease in cellularity, disoriented collagen fibers, and an increase in elastin fibres.¹²⁴ There are 2 forms of mitral valve disease: Barlow disease, characterized by myxomatous degeneration, and fibroelastic deficiency.¹²⁷ In Barlow disease, valves are thick and present an excess of connective tissue and amorphous ECM. They are rich in proteoglycans and fragmented elastin and poor in collagens, including at the level of the chordae tendineae. Fibroelastic deficiency results in thin leaflets and less content in proteoglycans, collagen, and elastin.

The cause of MVP is only partially understood (Figure 5). In addition to primary mitral valve disease, prolapse occurs as a manifestation of syndromic conditions, including Marfan syndrome (caused by heterozygous variations in FBN1¹²⁸), Loeys-Dietz syndrome (caused by heterozygous variations in genes encoding transforming growth factor-β [TGF-β] receptor components¹²⁹), and vascular Ehlers-Danlos syndrome (caused by variations in COL3A1¹³⁰), and is seen in a subset of patients with BAV disease. Impairments in ECM composition as observed in Marfan syndrome (eg, fibrillins) or vascular Ehlers-Danlos syndrome (eg, collagens) result in increased secretion of matrix metalloproteinases, which drive collagen and elastin fragmentation. The TGF-β pathway, mutated in Loeys-Dietz syndrome, plays a central role during valve development and in response to mechanical stress in that it promotes further cell proliferation and myofibroblast differentiation.

Figure 5. Pathogenesis of MVP.

A central role for transforming growth factor-β (TGF-β) signaling has emerged from genetic studies and is likely augmented by tissue damage from mechanical stress. TGF-β contributes to activation of valve interstitial cells (VICs). Other genes with possible roles in mechano-sensing/mechano-transduction include FLNA and TNS1. Serotonin signaling has been implicated in mitral valve prolapse (MVP) pathogenesis with an impact on TGF-β signaling and extracellular matrix (ECM) composition. DZIP1 and COL3A1 contribute to alteration of ECM integrity. Not pictured are genes involved in valvular development, including DCHS1. Created with BioRender.com.

The study of familial nonsyndromic MVP provides additional mechanisms related to cytoskeleton and planar cellular biology. A rare and X-linked form of polyval-vulopathy that comprises MVP is caused by variations in the filamin-A gene (FLNA), which encodes an actin-binding protein that crosslinks actin filaments to membrane glycoproteins.⁷⁷ Exome-sequencing studies in families have identified loss-of-function variations in DCHS1, encoding dachsous cadherin-related-1 gene, a member of the cadherin superfamily involved in cell adhesion.⁸¹ DCHS1 deficiency in VICs alters migration and cellular patterning from patients and provokes loss of cell polarity during valve development in mice and zebrafish models.⁸¹ Recently, DAZ interacting zinc finger protein 1 (DZIP1), a primary cilia gene, was reported as causal in familial MVP. Loss of primary cilia genes is hypothesized to modify ECM deposition during valve development and to enhance progressive myxomatous degeneration.⁷⁹

At the population level, the study of common genetic variation using GWASs in MVP has provided novel biological insights and confirmed those reported in rare syndromes and familial forms. An early meta-analysis of MVP GWASs revealed a long-range regulatory element for the TNS1 gene, which encodes a focal adhesion protein involved in cytoskeleton organization.^131,132 This work also reported 5 additional risk loci, including a variant in LMCD1, which encodes a repressor of GATA6, an important regulator for cardiac development and activator of calcineurin/NFAT pathway.¹³¹ It is interesting that pathway-driven enrichment analyses¹² and a follow-up MVP GWAS involving >4500 patients with MVP¹³³ stressed the pivotal role of cardiac development genes and TGF-β signaling in sporadic MVP, including TGFβ2 and LTBP2, and β-spectrin, a TGF-β signaling/scaffolding protein. Of note, several cardiomyopathy genes were mapped as potential causal genes in MVP risk loci, including ALPK3, BAG3, and RBM20. The mechanistic link with MVP revealed by these associations is still to be determined. We note that biological insights into fibroelastic deficiency forms of MVP are still missing. Recently, a link between serotonin and mitral valve remodeling was made. Polymorphisms in the serotonin transporter solute carrier family 6 member 4 (SLC6A4), also known as SERT, are associated with an earlier requirement for mitral valve surgery. Castillero et al¹³⁴ reported that decreased SERT activity contributes accelerated mitral valve remodeling and progression to MR.

Potential Therapeutic Targets

The long latency period between diagnosis and the need for surgical intervention makes MVP a good candidate for medical therapy that could slow or halt disease progression. Such a long-term therapy, however, must also be safe or ideally free from on-target side effects and toxicities. Several lines of evidence from the study of syndromic and sporadic forms of MVP point to the TGF-β pathway. The TGF-β pathway is involved in fundamental biology across multiple organ systems with roles in development, response to injury, and pathologies, including cancer. Therefore, it has been a popular therapeutic target for diseases as diverse as systemic sclerosis and focal segmental glomerulosclerosis to multiple forms of cancer.¹³⁵ Given the pleiotropic roles of TGF-β across such a range of biological settings, any intervention to modulate the pathway will require subtlety and likely tissue specificity to avoid undesirable side effects and toxicities.

Another interesting lead from the genetics of MVP is LMCD1. As a cofactor for GATA6, this complex inhibits DNA binding for lung and cardiac-specific promoters.⁸² However, LMCD1 also activates calcineurin/NFAT,⁸³ a known signaling pathway in heart valve development.¹³⁶ Whether LMCD1 functions in 1 or both of these roles in MVP is yet to be determined. Reduced LMCD1 expression is linked to increased MVP risk, consistent with the observation that knockdown of LMCD1 in zebrafish increased atrioventricular valve regurgitation.¹³¹ If LMCD1 is acting in MVP through its role in modulating calcineurin signaling, it would be through reduced calcineurin activity. Calcineurin inhibitors, including cyclosporin A and FK-506, are approved potent immunosuppressants, but calcineurin activators are not in clinical use and may be more challenging to develop for therapeutic purposes.

Other possible therapeutic targets in mitral valve disease include the focal adhesion protein tensin-1 and the transcription factor glis1, but the biology of their roles in MVP requires further elucidation. Overall, the need for a long-term therapy imposes a requirement for safety that, combined with the pleiotropic nature of the targets identified to date, makes a medical therapy for MVP a challenging goal.

Rheumatic Heart Disease

Classification and Epidemiology

Acute rheumatic fever results from infection with group A streptococcus, typically pharyngitis. Susceptibility is influenced by both bacterial strain and human host genetic characteristics. RHD, a long-term consequence of acute rheumatic fever, arises as a result of an immunemediated attack on the cardiac valves and other cardiac structures. RHD remains prevalent in regions of the world with limited access to resources, including diagnostic tests and penicillin.^84,85,137 The Global Burden of Disease study determined that in 2019 there were ≈2.8 incident cases and >40 million prevalent cases of RHD, representing a 1.5-fold increase in incident cases and 1.7-fold increase in prevalent cases since 1990.¹³⁸

Molecular Mechanisms

RHD is predominantly a disease of women, who account for 80% of cases. A recent proteomics study identified the protein ProTα (prothymosin-α) in the pathogenesis of RHD and suggested a possible role in the sex predisposition of this disease because ProTα is shown to contribute to CD8⁺ T-cell cytotoxicity associated with estrogen receptor α activity.¹³⁹

Immune activation in RHD is believed to occur through molecular mimicry. Specifically, linked CD4⁺ T-cell and B-cell reactivity to streptococcal antigens, including M protein and the group A carbohydrate epitope N-acetyl-b-D-glucosamine, leads to the production of autoantibodies that target cross-reactive epitopes in proteins present in the valve endothelium and endocardium, including laminin and cardiac myosin, among others.^84,137,140 This mimicry hypothesis is supported by studies in animal models, particularly the Lewis rat model, in which immunization with inactivated group A streptococcus or recombinant M5 protein with adjuvant induces production of anti-myosin antibodies and valvular carditis.^141,142 It is important to note that newer proteomic approaches are facilitating identification of additional autoantibody specificities.¹⁴³ For example, ProTα was recently observed to facilitate CD8⁺ T-cell recognition of human type 1 collagen, which shows molecular mimicry with Streptococcus pyogenes and capability to elicit immune responses in these cells.¹³⁹

Studies from both human subjects and animal models suggest that binding of autoantibodies to the valve surface increases the expression of vascular cell adhesion molecule 1 and related molecules that promote infiltration of T cells and innate immune cells, particularly macrophages, into the valve interstitium.^{84,137,144,145} The early inflammatory phase of RHD is dominated by proinflammatory cytokines produced by T cells and macrophages, including interferon-γ, interleukin-17, and tumor necrosis factor-α.^86,87 With chronic inflammation, valve fibrosis develops, driven by canonical fibrotic mediators, including transforming growth factor-β.⁸⁸

Consistent with the involvement of CD4⁺ T cells in pathogenesis, GWASs have identified HLA class II major histocompatibility complex risk alleles for RHD; the particular alleles vary according to the population studied.^84,85,146 Other genetic risk loci have been identified, nearly all of which influence expression of genes encoding proteins of both the innate and adaptive immune systems (eg, CTLA4, FCGR2A, FCN1, FCN2, FCN3, IGH, IL10, IL1RN, MBL2, TLR2, TNF, TGFB1).^{84,85,89–91} The FCN1-3 genes encode ficolins, a family of pattern recognition receptors that can bind to lipoteichoic acid and other constituents of the group A streptococcus cell wall.^89,90,92,93 This binding can trigger activation of the mannose-binding lectin pathway of complement activation, consistent with MBL2 also representing a susceptibility locus.^94,95 The RHDGen Network study (Genetics of Rheumatic Heart Disease) recently collected samples from 2548 patients with RHD and 2261 control subjects from 8 African countries.¹⁴⁷ This study replicated a previously reported genetic association with the immunoglobulin heavy chain locus (IGH)¹⁴⁸ and revealed several suggestive loci, including a novel susceptibility locus on chromosome 11 exclusive to Black African individuals.¹⁴⁹

Potential Therapeutic Targets

Targeted therapy for RHD remains limited. Primary treatment of acute group A streptococcus infection and secondary prophylaxis with penicillin are critical.^84,150 Historically, nonspecific immune modulators, including nonsteroidal anti-inflammatory drugs, corticosteroids, and intravenous immunoglobulin, have been used, but meta-analyses do not support their benefit.¹⁵¹ Heart failure that arises as a result of chronic RHD is treated with conventional approaches, including diuretics, fluid restriction, and angiotensin-converting enzyme inhibitors. Many patients eventually require valve replacement surgery.

For many immune-mediated diseases, biological therapeutics targeting T-cell activation, proinflammatory cytokines, cell adhesion molecules, and complement activation are in widespread use in resource-rich areas of the world. From our understanding of the immunopathogenesis of RHD, it is reasonable to expect that these agents could be useful for patients with RHD, particularly during the early inflammatory phase of disease. Unfortunately, practical considerations of cost, distribution, storage, and administration of these new therapies pose substantial barriers in the regions of the world where RHD remains so prevalent.

GAPS IN OUR UNDERSTANDING

There are many gaps in the understanding of human MVP. One major question facing the field currently is whether malignant forms of MVP can be prospectively identified and how they should be best managed clinically. Although the aberrant signaling pathways in MVP are slowly coming to light, a lack of faithful animal models for common MVP has limited our ability to probe these pathways for therapeutic potential. Spontaneous MVP occurs with high frequency in certain canine breeds. The Cavalier King Charles Spaniel is one such breed, but many small dog breeds have a high incidence. The genetics and relevance of canine MVP to human MVP are yet to be established, but this model could provide a ready resource for therapeutic development and discovery.

Although older studies using M-mode diagnostic criteria suggested a higher prevalence of MVP among women, larger community-based epidemiological evaluations such as that in the Framingham Heart Study demonstrate that the prevalence of MVP is similar between sexes.¹²⁵ However, there are apparent anatomical differences between men and women such that women have less posterior prolapse, less flail, and more leaflet thickening.¹⁵² MVP-relevant outcomes are also different between sexes, with women having higher mortality and lower surgery rates for severe MR than men. Biological mechanisms accounting for differences in anatomy and outcomes remain unknown; however, it is thought that the disparity in sex-specific surgical rates for severe MR may be due to differences in height and weight between sexes, leading to differences in interpretation of left ventricular enlargement.

Key questions about the pathogenesis of RHD remain unanswered. For instance, improved understanding is needed about why RHD is 2-fold more prevalent in women than in men^153–155; the impact of this discrepancy is underscored by the particular risks that RHD poses during pregnancy.¹⁵⁶ Similarly, improved understanding of the mechanisms accounting for the propensity of RHD to affect the left-sided cardiac valves, particularly the mitral valve, could lead to insights more broadly applicable to other forms of valvular disease; hypothesized mechanisms include differences in pressure/shear stress, oxygen tension, and other physiological parameters. Last, for studies of group A streptococci–induced rheumatic carditis, the Lewis rat model predominates. However, the number of genetic tools available in the rat is limited. Autoimmune valvular carditis occurs in the K/BxN mouse model of autoantibody-dependent arthritis, and this model has increasingly been proven valuable to dissect the immunopathogenic mechanisms at play.^142,157

RIGHT-SIDED VHD

In contrast to left-sided VHD, the majority of right-sided VHDs in adults are secondary disorders resulting from left-sided heart disease, pulmonary disease or pulmonary vascular disease. Mild tricuspid regurgitation is present in up to 80% of healthy individuals, but more severe regurgitation is present in older adults, often in association with atrial fibrillation, heart failure, or mitral valve disease or related to the position of a transvenous pacer lead.¹⁵⁸ A minority of individuals with tricuspid regurgitation have primary disease of the tricuspid valve, which either is congenital or may be due to acquired causes such as RHD, endocarditis, or metastatic carcinoid disease.¹⁵⁹ One of the more common congenital lesions causing primary tricuspid valve disease is the Ebstein anomaly, which occurs in ≈1 of every 200 000 live births and is characterized by apical displacement of the tricuspid annulus, an “atrialized” right ventricle, and an often fenestrated anterior tricuspid valve leaflet with the septal and posterior leaflets adhered to the myocardium.¹⁶⁰ There are few described, known genetic variations implicated in Ebstein anomaly, with small family-based studies identifying roles for NKX2.5 and MYH7.^161,162

Primary pulmonary valve stenosis or regurgitation is almost always due to congenital lesions. Specific causes of congenital pulmonary stenosis include tetralogy of Fallot and Noonan syndrome. It remains unknown why the aortic valve, which is also semilunar, manifests calcific stenosis, whereas the pulmonary valve does not, even for individuals with pulmonary hypertension.

FUTURE DIRECTIONS AND GAPS IN KNOWLEDGE

Despite advances in identifying mechanisms underlying the development of VHD, significant gaps in knowledge remain in several key areas (Table 3). First and foremost, identification of the primary drivers in the early development of VHD is necessary, particularly if factors amenable to preventive intervention can be defined. In addition to the mechanisms highlighted here, exogenous factors such as radiation-induced valve disease¹⁶³ drive maladaptive processes and lead to valve impairment. Improvement of existing imaging and diagnostic modalities might allow more rapid clinical implementation of potential medical therapy, with end points defined by imaging evidence of slowed disease progression in the valve leaflets because clinical end points such as mortality and valve replacement would require large sample sizes and long follow-up intervals. In both aortic and mitral valve disease, research highlights the possibility for medical therapies to slow or halt the progression of disease, but continued efforts are needed to identify new molecular targets and translate them into viable medicines. We highlight in Table 4 several research priorities informed by these gaps in knowledge that we hope can serve as incentives for future research funding in VHD.

Table 4.

Priorities for Future Research in VHD

	Research priorities
CAVD	Sex-differences in CAVD initiation and progression Bio-realistic disease models of CAVD, including 3D bioprinting to better recapitulate the multicellular and mechanical complexity of VHD Factors contributing to accelerated disease progression Genetic contributions to disease Determination of drug targets
BAV disease	Genetic screening/identification of heritable BAV loci Animal models with high frequency/penetrance of BAV for analysis of developmental and disease mechanisms
MVP	Additional genetic risk factors in MVP Biological translation of GWAS risk loci Determination of drug targets Mechanism of response to mechanical stress Specific genetic factors and mechanisms of MVP-associated arrythmias Specific genetic factors and mechanisms of fibroelastic deficiency
RHD	Pathogenic pathways amenable to small-molecule inhibition

3D indicates 3-dimensional; BAV, bicuspid aortic valve; CAVD, calcific aortic valve disease; GWAS, genome-wide association study; MVP, mitral valve prolapse; and RHD, rheumatic heart disease.

To achieve these goals, future research efforts will require translational study designs bridging data from clinical, molecular, and animal model system studies. Science should incorporate novel technologies, including imaging, systems biology, single-cell sequencing, and other new omics and computational platforms to enhance the identification of early disease processes and new molecular pathways. Careful review of candidate drug targets is needed with consideration of factors such as cost, distribution, storage, and administration, all of which affect equitable access to the prevention and treatment of VHD.