Extracellular matrix dynamics in aortic valve health and disease: Insights into fibrocalcific remodeling and creation of biomimetic platforms

Abstract

The extracellular matrix (ECM) of the aortic valve plays a pivotal role in maintaining valve function and becomes profoundly altered during the progression of calcific stenosis of the native aortic valve (CAS). CAS involves fibrocalcific ECM remodeling characterized by increased proteoglycans and glycosaminoglycans, enhanced collagen deposition, and fragmentation of elastic fibers, all of which contribute to valve thickening, fibrosis, and calcification. In this brief review, we provide an overview of these ECM changes and discuss the relationship between aberrant ECM remodeling and other pathological features of CAS - namely, differentiation of the resident valve cell types, inflammatory activity, lipid deposition, and relative hypoxia. Sexual dimorphism in ECM dynamics and the creation of disease-inspired scaffold environments to mimic CAS fibrosis are also discussed. Overall, understanding the complex interplay between cell phenotypes and ECM remodeling is crucial for elucidating the pathophysiology of CAS and developing novel treatment strategies.

Cell phenotypes and ECM in the healthy aortic valve

The aortic valve is composed of three leaflets that allow unidirectional blood flow from the left ventricle into the aorta. Despite the thin nature of these valve leaflets, they exhibit a complex and sophisticated structure (Figure 1). With respect to cellular composition, the aortic valve is comprised of valvular endothelial cells (VECs) and valvular interstitial cells (VICs). VECs line the blood-contacting surfaces of the valve, where they help to maintain valve tissue homeostasis and regulate cell activation/proliferation,^1–3 while VICs populate the interior of the valve, acting to constantly remodel the valve in response to the harsh hemodynamic environment. Several subtypes of VICs can exist in the valve ⁴ ; within the healthy valve, however, VICs are predominantly found in a quiescent (qVIC) phenotype. ⁵ qVICs are typically identified as staining positively for vimentin, and are negative for alpha-smooth muscle actin (αSMA). ⁶

Figure 1.

Schematic of aortic valve extracellular matrix structure and cellular composition.

With respect to the extracellular environment of the healthy valve, the aortic valve possesses a trilaminar extracellular matrix (ECM) with unique biological and mechanical features. ⁷ On the aortic side lies the fibrosa layer, composed primarily of circumferentially oriented fibers of collagens I (∼75%) and III (∼25%). The fibrosa confers mechanical support and tensile strength to the leaflet, supporting valve function in the presence of extreme hemodynamic forces.^7,8 On the ventricular side of the valve lies the ventricularis, which is rich in elastin and provides the valve with the elasticity needed to regulate leaflet opening and closing. ⁹ Finally, sandwiched between the fibrosa and ventricularis is the spongiosa, which contains proteoglycans (PGs) and glycosaminoglycans (GAGs), providing lubricity and mechanical cushion to the leaflet. ¹⁰ In histological analyses of healthy valves, these layers are easily distinguishable from each other, and this complex arrangement of ECM is believed to be crucial in enabling both the mechanical and biological functionality of the valve.

Cell phenotypes and ECM in the stenotic aortic valve

Calcific stenosis of the native aortic valve (CAS) involves an active process of fibrocalcific remodeling, with advanced age as the primary risk factor. ¹¹ VECs and VICs remain present in the stenotic valve, but both of these cell types transdifferentiate into phenotypes that contribute to CAS progression. Specifically, VECs transform to a mesenchymal phenotype in a process known as Endothelial-to-Mesenchymal Transition (EndMT).^12–14 Meanwhile, cues such as altered tissue stiffness, inflammatory cytokines, oxidized phospholipids, and growth factors such as transforming growth factor-beta1 (TGF-β1) all act as powerful stimuli for VIC differentiation into activated, myofibroblastic VICs (aVICs) and osteoblastic VICs (obVICs), both of which contribute to aortic valve calcification.^2,15,16 Joining VICs and VECs in the stenotic valve are inflammatory cells, including macrophages, mast cells, leukocytes, and T cells.^17,18 As outlined in a subsequent section, all of these cell types play a role in regulating the extensive ECM disarray that is characteristic of CAS.

In CAS, the ECM becomes altered across all layers of the valve. In recent years, advancements in transcriptomics and proteomics have yielded information about dozens of ECM-related molecules whose expression becomes altered during CAS.^19–23 These analyses provide a more granular look at specific genes and molecules that are aberrantly expressed in CAS, and a recent review ²⁴ provides an in-depth discussion of the ECM in CAS. For the purposes of this brief review, we will focus on changes that occur in the major categories of the valve ECM, namely, collagens, elastin, and GAGs/PGs.

Calcific aortic valve stenosis is largely driven by excessive deposition of a wide array of ECM molecules, including collagens (I, II, III, IV, V, VI), PGs (lumican, biglycan, versican, aggrecan, decorin), and varied other matricellular proteins such as laminins, fibronectin, periostin, and prolargin. ²⁴ One of the earliest hallmarks of CAS is increased production of PGs and GAGs, including decorin, hyaluronic acid, and versican. ²⁵ This increase in PGs and GAGs is responsible for the valvular thickening that occurs early in CAS, ²⁶ in addition to providing an environment that enables sequestration of oxidized phospholipids (Ox-PLs).^24,25 Following this initial PG/GAG enrichment, significant increases in fibrillar collagen deposition and remodeling start to occur. Fibrillar collagen is the primary indicator of fibrosis, and alterations to collagen become more pronounced as CAS progresses. ²⁷ Increased deposition of fibrillar collagens occurs throughout the valve, most notably in the spongiosa, where the fiber density increases by more than two-fold. ²⁸ The architecture of this collagen is also altered, as collagen fibers in stenotic valves are wider, which may be due to increased expression of lysyl oxidase (LOX), a potent crosslinker of collagen fibers that is found to be elevated in stenotic valves.^28,29 Additionally, these collagen fibers are shorter, likely due to increased expression of collagen-degrading matrix metalloproteinases (MMPs) by VICs.^30–32 Numerous MMPs have shown a positive association with progression of CAS, including MMP-1, MMP-9, MMP-10, MMP-12, and MMP-28.^33–35 In the ventricularis, elastic fibers continue to become more fragmented and disorganized, ²⁷ again due to the increased activity of MMPs^30,31 as well as multiple cathepsins.^36,37

At this point, the original ECM organization of the healthy valve has become disrupted, and, as a consequence, the valve has lost its ability to properly function in terms of biomechanics. The valve has now thickened by as much as 6-fold ³⁸ and significantly increased in fibrillar collagen density, thereby increasing leaflet stiffness. The fragmentation of elastin further causes a loss of leaflet elasticity, rendering the valve incapable of proper closure during diastole. The inability of the fibrotic valve to open and close correctly is a direct result of the excessive ECM deposition and remodeling described above.

Aortic valve fibrosis is followed by calcification, wherein nodule formation takes place via deposition of calcium salts. As described by others, ³⁹ this calcification occurs via both dystrophic and ossific mechanisms, with dystrophic calcification preceding osteogenic pathway activation. Briefly, dystrophic calcification emerges from the apoptosis of myofibroblastic VICs and subsequent deposition of calcium and phosphates at these nucleation sites; this process sets the stage for subsequent ossification, which results from more traditional osteogenic mechanisms. The majority of mineralization in CAS is believed to come from dystrophic calcification. ²⁴ However, ECM disruptions in the vicinity of nodules do not appear to vary with calcification mechanism. Regardless of origin, nodules are commonly surrounded by increased concentrations of several GAGs and PGs (eg, decorin, biglycan, versican, and HA).^40–42 These GAGs/PGs remain elevated in the valve, and their expression is correlated with the extent of stenosis.

Valvular disease hallmarks and their contributions to ECM remodeling

As described above, aortic valves become stenotic due to extensive fibrosis and calcification. The ECM changes involved in this pathological remodeling are the result of multiple other events in the valve, including: VEC and VIC differentiation, lipid deposition, inflammation, and hypoxia (Figure 2). In the following sections, we describe these elements of valve pathology, with specific attention to how they relate to the valve ECM and fibrosis.

Figure 2.

Overview of cellular and molecular events that contribute to aberrant extracellular matrix accumulation, and, ultimately, valvular stenosis.

VEC and VIC differentiation

In healthy valves, VECs and VICs coordinate to maintain homeostatic levels of ECM. However, VEC- and VIC-derived phenotypes are also responsible for driving much of the aberrant ECM remodeling that ultimately leads to fibrosis and calcification. Throughout the body, a healthy endothelium serves as a semipermeable barrier responsible for mediating the movement of molecules between the blood stream and the underlying tissue ⁴³ ; dysregulation of this barrier function has been linked to numerous cardiovascular pathologies, including CAS. ⁴⁴ In EndMT, ECs develop morphological features and cellular functions that are typically attributed to mesenchymal cells, thereby becoming more fibroblastic in nature. ¹² Valvular EndMT can be induced by disturbed shear stress, inflammation, oxidative stress, lipid deposition, and biomolecular cues such as interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and TGF-β1. ⁴⁵ In the aortic valve, shear stress-dependent EndMT occurs preferentially on the fibrosa side, where valvular fibrosis originates.^46,47 VEC EndMT involves loss of EC-specific markers such as CD-31, VE-Cadherin, and von Willebrand factor^43,48 and gain of a phenotypic profile similar to that of myofibroblastic VICs, with increased expression of αSMA, collagen deposition, and increased contractility.^45,49,50 By essentially serving as an additional source of valvular fibroblasts, these differentiated VECs contribute to the ECM deposition that is characteristic of valvular fibrosis.^51,52 VECs that have undergone EndMT may also continue to differentiate into an osteoblastic phenotype. ⁵³

Meanwhile, VICs populate the bulk of the valve leaflet interior, and, as described earlier, can differentiate into myofibroblastic (aVIC) or osteoblastic (obVIC) phenotypes in response to mechanical and biochemical stimuli. In a healthy valve, aVICs account for fewer than 2.5% of the cells in the valve. ⁵⁴ However, aVICs become abundant during the onset of CAS and are thought to be the major producer of the fibrillar collagens that lead to development of fibrosis. Substrate stiffness is particularly notable as a myofibroblast-inducing factor, as VICs cultured on traditional 2D tissue culture plastic will spontaneously differentiate into aVICs. ⁶ Characteristics of aVICs include increased αSMA expression and contractility, along with increased expression of multiple ECM molecules associated with fibrosis, most notably type I collagen, fibronectin, and PGs. ⁵⁵ It is presumed that VICs are the primary source of the myriad other upregulated ECM components found in stenotic valves, ²⁴ although the cellular source for many of these molecules has not yet been confirmed, as the gene expression analyses of these stenotic valves have generally been performed on the bulk sample, without mapping to specific cell types within the valve. Production of matrix remodeling enzymes (eg, MMPs), which are positively associated with AV severity, is also elevated in aVICs. ⁵⁵

VICs may also differentiate into an osteoblastic phenotype (obVICs).^56,57 Osteoblastic VIC differentiation can be stimulated by the mechanical environment, inflammatory factors, and growth factors such as TGF-β1 and bone morphogenetic proteins (BMPs).^16,58–60 The differentiation of VICs into obVICs is marked by expression of RUNX2, osteopontin, osteocalcin, alkaline phosphatase, bone sialoprotein, and BMP −2 and −4.^4,56 However, while VICs can exhibit some osteoblastic characteristics, their osteogenic activity is still far less robust than true osteoblasts, so they are typically referred to as ‘osteoblast-like’ cells.^56,61 Although aVICs are likely responsible for the majority of ECM remodeling during CAS, obVICs also participate in this process, as they demonstrate increased expression of elastin and multiple collagen-related genes. ⁶²

Crosstalk between VECs and VICs also influences fibrosis and osteogenesis, although there are conflicting findings on the nature of this relationship. VECs have been shown to mitigate fibrosis by helping VICs maintain a more quiescent phenotype, as evidenced by decreased expression of multiple markers of VIC pathology, including αSMA.^63,64 However, some recent findings suggest that VECs actively contribute to VIC-driven pathogenesis via paracrine signaling.^2,65 These conflicting results may indicate that the influence of VECs on VIC function is dependent upon the composition and mechanics of the culture environment, as the aforementioned studies were performed across 2D and 3D platforms with varied ECM composition.

Inflammatory activity

Inflammatory cytokines and cells are not typically found in healthy valves; however, a combination of numerous other pathological events, including VEC and VIC dysfunction, lipid deposition, and upregulation of reactive oxygen species all combine to create an environment that recruits and sustains inflammatory activity. Lipid accumulation was one of the original CAS histological hallmarks described by Otto et al in 1994. ²⁶ More recent work has demonstrated that the presence of oxidized lipoproteins in the valve fibrosa plays a central role in triggering CAS inflammatory activity and initiating osteogenic reprograming of VICs. Altered shear stresses contribute to dysregulation of VEC barrier functions, which can allow lipoproteins (eg, LDL, ApoB, ApoE) to enter the valve fibrosa.^26,66,67 Once these lipid species enter the valve fibrosa, they rapidly become oxidized by reactive oxygen species, which are also elevated in stenotic valves. ⁶⁸ These reactive oxygen species are created by oxidative stress in the aortic valve, a result of dysregulation of the endothelial nitric oxide synthase (eNOS) pathway. ⁶⁹ This oxidative environment transforms deposited lipoproteins into Ox-PLs, thereby initiating chronic, low-grade inflammatory activity within the aortic valve.

Higher levels of Ox-PLs have been associated with more severe progression of CAS and more extensive fibrocalcific remodeling.^70,71 Further, in vitro studies have shown that oxidized LDL (oxLDL) and other Ox-PLs promote aortic valve calcification.^72,73 The means by which these oxidized species contribute to valvular fibrosis ⁷⁰ and calcification ⁷¹ involve both direct and indirect mechanisms. For example, oxLDL can directly stimulate the production of sulfated GAGs by VICs, ²⁵ which are not only the major contributor to early-stage valvular thickening, but also serve to further trap more Ox-PLs. Sulfated GAGs can profoundly affect VIC behavior, as chondroitin sulfate increases both myofibroblastic and osteoblastic activity in VICs, in addition to stimulating collagen I production. ⁶⁴ OxLDL also stimulates the production of biglycan by VICs, ²³ where biglycan is correlated with CAS severity and is capable of stimulating osteogenic gene expression in VICs. ⁷⁴ Finally, although commonly thought to be a “good” lipoprotein, high density lipoproteins (HDL) have also been implicated in CAS. OxHDL is elevated in patients with CAS, and treatment with oxHDL in vitro led to increased osteoblastic VIC activity. ⁷⁵

Ox-PLs also promote ECM remodeling and fibrosis in an indirect manner by increasing the production of multiple pro-inflammatory cytokines and surface receptors by VICs, which, in turn, increases the recruitment and invasion of immune cells that regulate ECM remodeling.^76,77 Macrophages, mast cells, and T cells comprise the inflammatory infiltrate found in stenotic valves. ⁷⁸ These immune cells secrete myriad pro-inflammatory cytokines and growth factors (eg, IL-1β, IL-6, IL-8, TNF-α, IGF-1, and TGF-β1) that serve as powerful regulators of ECM production and degradation. Perhaps the most potent pro-fibrotic stimulus in the valve is TGF-β1. The TGF-β1 in the diseased aortic valve comes from both the inflammatory cells and the VICs themselves. Throughout the body, TGF-β1 is known for its ability to stimulate ECM production and significantly increase deposition of fibrillar collagen, ⁷⁹ both of which also hold true in the case of VICs. ⁸⁰ In addition to increasing collagen production, TGF-β1 promotes the elongation of valve PGs, ⁸¹ which serve a key role in sequestering Ox-PLs.

While the association between TGF-β1 and valvular fibrosis is fairly well established, the role of other pro-inflammatory molecules in CAS progression can be a bit more nuanced. For example, TNF-α activates the NF-κB pathway in VICs, which promotes upregulation of calcification genes RUNX2 and BMP2 and the formation of calcium deposits. ⁸² TNF-α treatment of VICs also induces production of IL6, which further promotes aortic valve inflammatory activity ⁸² and obVIC differentiation. ¹⁶ However, the net role of TNF-α in CAS is unclear, as it may also decrease myofibroblastic differentiation and collagen production by VICs.^83,84 In other cell types (including mitral valve VICs), TNF-α was able to downregulate the production of TGF-β1. ⁸³

In addition to cytokines, macrophages secrete numerous proteolytic enzymes (eg, MMPs and cathepsins), that contribute to the ECM remodeling that typifies CAS.^16,85 Moreover, macrophages and VICs participate in a reciprocal relationship with respect to signaling ECM remodeling; while we most often think of macrophages releasing molecules that affect VIC phenotype, this paracrine signaling functions in both directions. For example, VICs can stimulate macrophages to significantly decrease their MMP-9 production, ⁸⁶ thereby leading to excessive ECM accumulation. Finally, immune cells undergo apoptosis as a consequence of oxidative stress in the diseased valve, creating nucleation sites for calcium deposition.

Hypoxia

The healthy aortic valve measures a mere 300 microns in thickness. ³⁸ However, the extensive fibrocalcific remodeling of CAS transforms thin aortic valve tissue to thickened fibrotic tissue, with extensively diseased valves being as thick as 2.2 mm. ³⁸ In this thickened environment, the VICs can no longer obtain oxygen through simple diffusion, and they experience an increase in hypoxia relative to healthy valves. Evidence of increased hypoxic conditions can be found in diseased valves, where expression of hypoxia inducible factor-1α (HIF-1α) and pro-angiogenic vascular endothelial growth factor (VEGF) in diseased valves co-localizes with calcific nodules. ⁸⁷ Hypoxia has also been linked to altered production of multiple ECM components by VICs, including collagen and GAGs.^88,89 Further, diseased valves show evidence of neo-vascularization, another consequence of hypoxia.^87,90,91

Although increased relative hypoxia has been identified in CAS progression, there is still much to learn about the role that hypoxia plays in this disease and its accompanying ECM remodeling and fibrosis. HIF-1α is a transcription factor known as the master regulator of the hypoxia response. HIF-1α has been shown to upregulate calcification, fibrosis, and inflammatory pathways, all of which are key steps in CAS progression.^88,92 In other tissues in the body, HIF1α expression typically leads to significant upregulation of collagen I production, hydroxylation, and crosslinking,^88,93,94 although initial studies suggest that the effects of HIF-1α on valvular ECM may be different. ⁸⁸ There is also evidence that hypoxia may stimulate MMP-9 activity in CAS. ⁹⁵ Finally, HIF-1α activation may increase aortic valve calcification by aiding in the osteogenic differentiation of VICs. ⁹⁶

Sex differences in CAS and associated ECM remodeling

[Disclaimer: It should be noted that most of the studies that have interrogated sex differences in CAS have been published in association with clinical reports where the method used to determine sex is not explicitly described, but is assumed to be based upon sex assigned at birth.]

Cardiovascular disease is an arena where there is clear evidence of sexual dimorphism.^97,98 Some features of cardiovascular disease where sexual dimorphism is especially prominent are cardiac fibrosis, inflammation and calcification. While the 2016 National Institutes of Health policy on sex as a biological variable (SABV) has improved the inclusion of females in studies involving vertebrate animals or humans, this policy has not translated to including SABV in cellular-scale research. Thus, the origin of sexual dimorphic behavior in cardiovascular disease is not well understood, and the relative contributions of chromosomal sex versus sex hormones are only beginning to be examined.^98–100

Although sex differences in CAS incidence have long been known, ¹⁰¹ it is only within the last 12 years that sexual dimorphism in CAS pathophysiology has been described. In 2012, the existence of cellular-scale sex differences in VICs was demonstrated. ¹⁰² Then, in 2017, a landmark study ¹⁰³ revealed new information about sexual dimorphism in clinical CAS presentation: for the same extent of stenosis, women presented with a more fibrotic outcome compared with males, who tended towards development of calcific nodules. Furthermore, along with a lower risk of aortic calcification, women also experienced more fibrotic remodeling than men, irrespective of age and valve phenotype. ¹⁰⁴ Thus, changes to the valve ECM during CAS seem to be strongly dependent upon sex. At a cellular level, we see further evidence of sex-specific differences in ECM remodeling in CAS.^105,106 Specifically, in cell cultures separated by chromosomal sex, male VICs display a greater tendency to assume a myofibroblastic phenotype,^102,105 in addition to greater secretion of matrix-degrading enzymes compared to female VICs. ¹⁰⁵ Meanwhile, female VICs exhibited greater production of fibrillar collagen. ¹⁰⁵

Studies of sexual dimorphism in CAS have largely focused on genetic sex, while the role of sex hormones in CAS progression largely remains unaddressed. Like many other cardiovascular diseases, incidence of CAS is markedly increased when women enter post-menopause, suggesting that estrogen may mediate protection against CAS initiation and/or progression earlier in life. The ‘cardioprotective’ label for E2 was first coined due to the observation that women's risk of atherosclerosis experiences a rapid increase at the time of menopause, when estrogen levels have declined. ¹⁰⁷ While little is known about estradiol's influence in the specific context of CAS, estradiol (E2) signaling through its primary receptors (ER-α, ER-β, and GPER) has been shown to influence the pathophysiology of several other cardiovascular diseases that share pathological features with AS. These cardioprotective actions may be related to E2 inhibiting proliferation of vascular smooth muscle cells and decreasing oxidative stress. ¹⁰⁷ Estradiol has also been shown to alter collagen production and remodeling. For example, E2 decreases the production of TGF-β1^108,109 and fibrillar collagens^110,111 in multiple other tissue types.

ECM-inspired scaffolds to study CAS

Historically, the mechanisms underlying CAS onset and progression have been investigated using either animal models (both diet-induced and genetically modified) or two-dimensional in vitro cell cultures. These models have been summarized in recent reviews. ⁸³ Although animals provide complex, multi-organ systems in which to study disease, they are often costly, not always representative of human physiology, and do not permit precise control over variables that may be needed to probe mechanistic questions. Thus, despite their reductionist nature, there is still a need for in vitro platforms. Despite serving as the standard in vitro platform for decades, 2D cultures have proven particularly problematic in the context of VICs, whose myofibroblastic differentiation is stimulated by the high stiffness of tissue culture plastic. ⁶ Moreover, typical 2D cultures cannot mimic the valvular ECM, which is thought to be highly influential in regulating CAS. In the following sections, we revisit the CAS hallmarks described earlier in this review and briefly discuss efforts and unrealized opportunities to create 3D, ECM-inspired culture platforms that examine the role of these hallmarks in the progression of CAS.

ECM-inspired scaffolds to study VIC and VEC differentiation

ECM-inspired scaffold systems provide several different mechanisms by which VIC phenotype can be controlled, such as stiffness, composition, and addition of exogenous stimuli. Multiple types of scaffold-based cell culture platforms have been developed to control VIC phenotype, and are particularly well-suited for studying stiffness-induced VIC differentiation.^112–114 For example, photodegradable polyethylene glycol (PEG)-based hydrogels have been synthesized to examine the effects of stiffness on VIC myofibroblastic differentiation. ¹¹⁵ In these studies, 70% of VICs on stiff (15 kPa) substrates underwent myofibroblastic activation, compared to 20% of VICs on soft (3 kPa) substrates. Intriguingly, the authors showed that VIC activation could be reversed by dynamically decreasing the stiffness during VIC culture. ¹¹⁵

ECM composition has also been found to be a powerful regulator of VIC phenotype in biomimetic scaffolds. Early studies of VICs on different 2D ECM coatings revealed that VICs are highly sensitive to the identity of the underlying ECM, with some ECM components (eg, fibronectin, HA) supporting more quiescent VIC function, and other components (eg, fibrin, laminin) allowing more myofibroblastic activity and dystrophic nodule formation.^116,117 Across numerous studies, hyaluronic acid (HA) has emerged as an ECM molecule that appears to consistently support a physiologically healthy VIC phenotype.^25,118–121 While VICs spontaneously underwent myofibroblastic differentiation in methacrylated gelatin (GelMA) hydrogels, the inclusion of methacrylated HA (HAMA) in these materials prevented this spontaneous differentiation. These scaffolds permitted induction of an aVIC phenotype via treatment with exogenous TGF-β1, ¹¹⁹ or differentiation to an obVIC phenotype upon treatment with osteogenic media and TNF-α. ¹²⁰ Inclusion of HA in collagen-based hydrogels yielded similar findings. ¹²¹ Specifically, hydrogels were synthesized using varying ratios of collagen to HA while maintaining similar mechanics, and VICs cultured in constructs with lower HA content exhibited more aVIC and obVIC differentiation. In fact, when VIC interactions with HA are blocked in ex vivo leaflet organ cultures, this is sufficient to spur myofibroblastic differentiation of the VICs. ¹²² 2D studies of VICs on HA coatings suggest that VICs may also be sensitive to HA molecular weight, ¹²² although this has not been confirmed in other culture systems. VIC phenotype within 3D hydrogels can also be controlled via addition of exogenous factors, such as bFGF to maintain a qVIC phenotype, ¹²³ or TGF-β1 and osteogenic media to induce aVIC and obVIC phenotypes, respectively.

To date, the use of 3D culture platforms specifically targeted at studying EndMT in VECs remains relatively rare. Most prior work with VECs has focused on studying the suitability of different biomaterial environments for supporting proper VEC function, rather than biomaterial-based models to recapitulate VEC dysfunction in CAS. That said, a handful of ECM-inspired culture platforms have included co-culture of VICs and VECs ¹²⁴ and subsequently analyzed pathological differentiation. ⁶⁵ The role of different GAGs (HA, CS, and dermatan sulfate (DS)) in modulating calcification was examined in a collagen I-based hydrogel platform containing both VICs and VECs. ⁶⁴ The authors found that high stiffness and the presence of CS promoted calcification; however, in DS- or HA-enriched scaffolds, VECs attenuated the pathological differentiation of VICs.

ECM-inspired scaffolds to study fibrosis and calcification

Historically, 3D scaffolds for valvular cell culture have been pursued for the purpose of creating tissue-engineered heart valves as options for valve replacements. Healthy valve tissue engineering is extensively reviewed elsewhere. ¹²⁵ However, it is only within the last decade that this work has been extended to develop scaffolds that mimic features of CAS progression.

As discussed earlier, GAG enrichment is one of the earliest CAS hallmarks. GelMA-GAG hydrogel systems have been used to explore how this GAG enrichment influences CAS progression. In one instance, GelMA was combined with methacrylated HA or CS in amounts meant to represent the healthy valve environment versus the GAG-enriched environment found in early CAS. ²⁵ This work found that enrichment in CS did not directly induce pathological behaviors in VICs, but rather influenced VICs in an indirect manner. Specifically, CS was able to trap oxLDL which, in turn, promoted the production of multiple inflammatory cytokines by VICs, as well as expression of pro-fibrotic molecules and production of more CS, thereby creating a positive feedback loop for disease progression. Meanwhile, enrichment in HA was not associated with significant lipid entrapment, but did stimulate production of pro-angiogenic factors by VICs. ²⁵

Others have found that CS can indeed directly stimulate pathological VIC behaviors,^64,126 a finding which may differ from those described above due to the scaffold platform mimicking a later stage of CAS. To represent the dense fibrillar features and mechanics of valves that have progressed to a more fibrotic state, CS was incorporated into collagen I hydrogels of varied stiffness. ⁶⁴ In this system, stiffness alone was the most influential variable in guiding VIC differentiation, but CS itself was able to stimulate significant myofibroblastic and osteoblastic activity by the encapsulated VICs.

As a result of the implementation of biomaterial platforms that are intended to maintain healthy VIC phenotypes^25,118–121 or to mimic elements of CAS fibrosis,^25,64,126 it is becoming evident that GAGs play a critical role in regulating CAS progression, and that the molecules within this ECM class perform distinct, independent functions in this regard. Multiple types of tissue-engineered constructs have implicated CS as a driver of pathological VIC behavior and other hallmarks of CAS, with both direct and indirect influences on VIC differentiation.^25,64,126 Similarly, several disparate pieces of evidence are converging to suggest that HA may be important in supporting non-pathological VIC function.^25,118–121

The biomaterial platforms discussed above employ changes in ECM composition and stiffness to mimic fibrosis, but spatial ECM organization is also a feature that changes with valve fibrosis. This aspect of fibrosis was mimicked using photo-tunable PEG-based hydrogels to create organized versus disorganized patterns of elastic moduli. ¹²⁷ In short, photomasks and UV light were applied to PEG hydrogels (10 kPa) to selectively soften the scaffolds (to 4 kPa) in regular or random patterns (ie, causing interspersed soft and stiff sections of the material), thereby representing the heterogeneous nature of collagen distribution and focal areas of stiffness in fibrotic valves. Regardless of whether the patterns of soft versus stiff areas were random or organized, VICs responded to an increased overall percentage of stiff areas with increased myofibroblastic differentiation, which is consistent with many previous findings regarding stiffness and VIC differentiation. However, even when two different substrates had the same number of soft versus stiff sections, the nature of the spatial patterning alone was sufficient to control VIC phenotype; namely, when the soft versus stiff segments were present in organized patterns, there was greater αSMA expression by VICs, but VIC proliferation was greater on random patterns. ¹²⁷

In the latest stages of disease, fibrosis progresses to calcification. To study VIC behavior within a calcified environment in vitro, the traditional approach has been treatment with well-established mineralization media cocktails, rather than modifying the scaffold composition itself to represent a mineralized matrix. Some modifications to scaffold composition, such as inclusion of collagen I, ¹²⁸ can facilitate mineralization upon treatment with osteogenic medium, since collagen I can promote hydroxyapatite deposition. However, collagen I alone is typically not sufficient to initiate this process. Thus, recent innovative approaches have incorporated hydroxyapatite into 3D scaffolds to produce systems that resemble the calcification present in CAS and thus gain insight into events driving cellular behaviors in the later stages of disease. For example, 3D collagen hydrogels have been combined with hydroxyapatite microcrystals of varying crystallinity. ¹²⁹ After 7 days of VIC and VEC culture in these scaffolds, calcium deposition was achieved in the absence of mineralization medium. Additionally, the extent of crystallinity of the hydroxyapatite particles was found to be highly influential in guiding VIC and VEC differentiation, with less crystalline particles driving more pathological outcomes. ¹²⁹

Future opportunities to understand and mimic fibrosis in CAS

The development of new bioengineering and -omics tools has enabled advancements in our understanding of ECM dynamics in CAS. There are many opportunities yet to be realized in this realm, both in terms of describing the valve ECM and in mimicking it. Recent proteomic analyses of valves have opened up new information about the valve ECM that has yet to be explored in prospective experiments.^19–23 Characterization of ECM changes with normal ageing, and the roles of biological sex and sex hormones in this process, also remains incomplete.

As discussed in this review, CAS features such as lipid infiltration and oxidation, inflammatory activity, and increased hypoxia all intersect with ECM changes in CAS and promote pro-fibrotic and osteogenic actions. However, it is less common for these elements to be directly incorporated into CAS-mimicking scaffold platforms. Inflammatory stimuli have typically been delivered to VIC-containing scaffolds via inclusion of biomolecules (eg, TNF−α) to the culture media ¹³⁰ or use of macrophage-conditioned media. ¹⁶ Thus, there is an opportunity to incorporate immune cell co-cultures, as well as oxidized lipid species, into ECM-inspired valve scaffolds in order to understand the interplay between these pathological cues and improve our ability to mimic CAS progression. Moreover, the culture of these platforms under dynamic conditions is likely to offer further insight into disease pathology, as ECM changes during CAS alter hemodynamic forces, which, in turn, cause further aberrant ECM remodeling. Overall, given the critical role of the ECM in both regulating and driving CAS pathogenesis, ECM-inspired scaffold models hold promise in deciphering the mechanisms of CAS progression and in developing therapeutic strategies.