Ultrastructural Pathology of Atherosclerosis, Calcific Aortic Valve Disease, and Bioprosthetic Heart Valve Degeneration: Commonalities and Differences

Atherosclerosis, calcific aortic valve disease (CAVD), and bioprosthetic heart valve degeneration (alternatively termed structural valve deterioration, SVD) represent three diseases affecting distinct components of the circulatory system and their substitutes, yet sharing multiple risk factors and commonly leading to the extraskeletal calcification. Whereas the histopathology of the mentioned disorders is well-described, their ultrastructural pathology is largely obscure due to the lack of appropriate investigation techniques. Employing an original method for sample preparation and the electron microscopy visualisation of calcified cardiovascular tissues, here we revisited the ultrastructural features of lipid retention, macrophage infiltration, intraplaque/intraleaflet haemorrhage, and calcification which are common or unique for the indicated types of cardiovascular disease. Atherosclerotic plaques were notable for the massive accumulation of lipids in the extracellular matrix (ECM), abundant macrophage content, and pronounced neovascularisation associated with blood leakage and calcium deposition. In contrast, CAVD and SVD generally did not require vasculo- or angiogenesis to occur, instead relying on fatigue-induced ECM degradation and the concurrent migration of immune cells. Unlike native tissues, bioprosthetic heart valves contained numerous specialised macrophages and were not capable of the regeneration that underscores ECM integrity as a pivotal factor for SVD prevention. While atherosclerosis, CAVD, and SVD show similar pathogenesis patterns, these disorders demonstrate considerable ultrastructural differences.


Introduction
Atherosclerosis and calcific aortic valve disease (CAVD) represent chronic inflammatory disorders which are characterised by endothelial dysfunction, lipid deposition, macrophage infiltration, and the maladaptive remodelling of the extracellular matrix (ECM) in the arterial wall and valve leaflets [1][2][3][4][5]. The altered paracrine signalling of the dysfunctional endothelium and blood-derived immune cells results in the excessive release of extracellular proteases, as well as pro-inflammatory and pro-calcific cytokines and growth factors [1][2][3][4][5]. The pathological microenvironment also induces the proliferation, phenotypic switch, and osteochondrogenic differentiation of vascular smooth muscle cells (VSMCs) and valvular interstitial cells (VICs), which additionally disturbs the balance between ECM deposition

Results
We first performed a semi-quantitative analysis of all the samples (36 atherosclerotic plaques, 12 calcified native AVs, and 12 failed BHVs with SVD). As underlying health conditions could impact on the development of histopathological features, we have also analysed the prevalence of the patient comorbidities, but did not find any statistically significant differences between the groups (Table 1). Table 1. Clinicopathological features of the patients with atherosclerosis, CAVD, and BHV failure. Each of the disorders has been characterised by a specific histopathological pattern ( Table 2). Foam cells were detected in all plaques but were less frequent in diseased native AVs and BHVs. Canonical macrophages (i.e., mononuclear macrophages without any cytoplasmic inclusions) were encountered regardless of the tissue, whereas multinucleated giant cells were characteristic of failed BHVs. Neutrophil infiltration was often noted in plaques and BHVs, but not in the native AVs. Endothelialisation or pseudoendothelialisation (a layer of endothelial-like cells at the surface) were common for all the studied specimens. Neovascularisation was particularly notable in plaques. While haemorrhages were observed across all pathologies, they were caused by the leakage of neovessels in plaques and by red blood cell (RBC) penetration or delamination in native AVs and BHVs. Mineral deposits were detected in almost all the examined samples. We further carried out a detailed analysis of the indicated ultrastructural features. As atherosclerosis, CAVD, and SVD all involve macrophage infiltration and are largely driven by lipid accumulation, we primarily focused on the formation of foam cells. Depending on the dehydration efficiency, they appeared as large round, oval, or irregularly shaped cells full of black or white globules ( Figure 1A). Among all the examined cardiovascular tissues, foam cells were the most abundant in the atherosclerotic plaques, where they generally located as multiple aggregations randomly distributed across the neointima ( Figure 1A). Native AVs and BHVs contained a few lipid-laden cells ( Figure 1A), although fatty streaks were also prominent at the inflow and outflow surfaces ( Figure 1A,B). randomly distributed across the neointima ( Figure 1A). Native AVs and BHVs contained a few lipidladen cells ( Figure 1A), although fatty streaks were also prominent at the inflow and outflow surfaces ( Figure 1A,B).  We next investigated the histopathological diversity of the macrophages in the affected vascular and valvular tissues. The majority of the macrophages had elliptic or round large nuclei and did not have any specific inclusions in the cytosol, suggesting their moderate activity in the ECM remodelling ( Figure 2A). However, some of the valvular macrophages, typically located within the degraded ECM, contained numerous electron-dense granules and had an intermediate phenotype between canonical macrophages and multinucleated giant cells ( Figure 2B). This macrophage subtype was especially prominent in BHVs ( Figure 2B). Multinucleated giant cells with multiple cytoplasmic inclusions were detectable in the atherosclerotic plaques and BHVs ( Figure 2C). Besides the macrophages, the plaques and BHVs were infiltrated by neutrophils, suggestive of active inflammation ( Figure 2D). We next investigated the histopathological diversity of the macrophages in the affected vascular and valvular tissues. The majority of the macrophages had elliptic or round large nuclei and did not have any specific inclusions in the cytosol, suggesting their moderate activity in the ECM remodelling ( Figure 2A). However, some of the valvular macrophages, typically located within the degraded ECM, contained numerous electron-dense granules and had an intermediate phenotype between canonical macrophages and multinucleated giant cells ( Figure 2B). This macrophage subtype was especially prominent in BHVs ( Figure 2B). Multinucleated giant cells with multiple cytoplasmic inclusions were detectable in the atherosclerotic plaques and BHVs ( Figure 2C). Besides the macrophages, the plaques and BHVs were infiltrated by neutrophils, suggestive of active inflammation ( Figure 2D).  Monocyte infiltration followed by a macrophage-driven ECM disintegration represents a multi-step process. Electron microscopy snapshots enabled to illustrate its events, including rolling ( Figure 3A), adhesion ( Figure 3B), the invasion of the cells into the ECM ( Figure 3C,D), the degradation of the surrounding ECM and engulfment of its components ( Figure 3E), and ECM delamination ( Figure 3F). Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 6 of 18 red arrows) and macrophage (indicated by blue arrows) infiltration, magnification x1000. Note the absence of neutrophils in the native AVs.
Monocyte infiltration followed by a macrophage-driven ECM disintegration represents a multistep process. Electron microscopy snapshots enabled to illustrate its events, including rolling ( Figure  3A), adhesion ( Figure 3B), the invasion of the cells into the ECM ( Figure 3C,D), the degradation of the surrounding ECM and engulfment of its components ( Figure 3E), and ECM delamination ( Figure  3F).  Leukocyte rolling and attachment are induced by endothelial activation and the loss of endothelial integrity, although endothelial dysfunction does not necessarily mean endothelial injury, and re-endothelialisation is a common regenerative phenomenon. Atherosclerotic plaques and native AVs were covered by a layer of endothelial cells elongated along the direction of flow, while BHVs had single endothelial-like cells attached to the prosthetic surface ( Figure 4A). Mesenchymal cells in the examined cardiovascular tissues mostly included vascular/valvular SMCs or VICs ( Figure 4B). Fibroblasts were rarely encountered ( Figure 4C). Leukocyte rolling and attachment are induced by endothelial activation and the loss of endothelial integrity, although endothelial dysfunction does not necessarily mean endothelial injury, and re-endothelialisation is a common regenerative phenomenon. Atherosclerotic plaques and native AVs were covered by a layer of endothelial cells elongated along the direction of flow, while BHVs had single endothelial-like cells attached to the prosthetic surface ( Figure 4A). Mesenchymal cells in the examined cardiovascular tissues mostly included vascular/valvular SMCs or VICs ( Figure 4B). Fibroblasts were rarely encountered ( Figure 4C).  We found intraplaque haemorrhage as a frequent event in the atherosclerotic plaques because of leaky neovessels ( Figure 4D). In contrast to the vasa plaquorum, microvessels within the native AVs and capillary-like tubes in the BHVs were impermeable for red blood cells (RBCs) ( Figure 4D). Yet, both types of valve were also penetrated by RBCs, which was accompanied by impaired ECM integrity ( Figure 4E) as well as degradation ( Figure 5A) and the disorientation of the collagen fibers ( Figure 5B). The disintegration of the ECM was associated with haemorrhagic infiltration (Figure 5C), the delamination of the tissue (Figure 5D), and the migration of RBCs into the hollows ( Figure 5E). Mineral deposits were characterised by a remarkable heterogeneity. Both plaques and native/bioprosthetic valves contained macrocalcifications between the inflow and outflow surfaces ( Figure 6A), which could be surrounded (plaques and native AVs) or covered (BHVs) by a connective Mineral deposits were characterised by a remarkable heterogeneity. Both plaques and native/bioprosthetic valves contained macrocalcifications between the inflow and outflow surfaces ( Figure 6A), which could be surrounded (plaques and native AVs) or covered (BHVs) by a connective tissue ( Figure 6B). Some of the calcifications had uneven or sharp edges, suggestive of its invasive pattern ( Figure 6C). Multiple microcalcifications of varying diameter were often detected in all the examined tissues ( Figure 6D). Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 10 of 18 tissue ( Figure 6B). Some of the calcifications had uneven or sharp edges, suggestive of its invasive pattern ( Figure 6C). Multiple microcalcifications of varying diameter were often detected in all the examined tissues ( Figure 6D).

Discussion
The concept that the development of atherosclerosis and heart valve disease is interrelated was a matter of debate during the last decade [24][25][26][27]. The arguments include common genetic susceptibility loci [28][29][30] and a number of identical circulating biomarkers [31][32][33]; yet, the similarities and differences in the pathophysiology of atherosclerosis, CAVD, and structural degeneration of BHVs have not been systematically compared. In particular, there are a lack of ultrastructural data connecting the pathophysiology of these disorders with their clinical manifestations. The sample preparation of extraskeletal calcified tissues such as plaques or dysfunctional heart valves is complicated, since it is incompatible with histological sectioning. Having applied our original protocol for the staining, embedding, and visualisation of the

Discussion
The concept that the development of atherosclerosis and heart valve disease is interrelated was a matter of debate during the last decade [24][25][26][27]. The arguments include common genetic susceptibility loci [28][29][30] and a number of identical circulating biomarkers [31][32][33]; yet, the similarities and differences in the pathophysiology of atherosclerosis, CAVD, and structural degeneration of BHVs have not been systematically compared. In particular, there are a lack of ultrastructural data connecting the pathophysiology of these disorders with their clinical manifestations. The sample preparation of extraskeletal calcified tissues such as plaques or dysfunctional heart valves is complicated, since it is incompatible with histological sectioning. Having applied our original protocol for the staining, embedding, and visualisation of the mineralised cardiovascular specimens, here we carried out an electron microscopy investigation of the main histopathological features of atherosclerosis, CAVD, and SVD-i.e., macrophage phenotypic switch and infiltration, intraplaque/intraleaflet haemorrhage, ECM degradation, and calcification.
Despite lipid deposition being prominent regardless of the tissue type, it was most pronounced in atherosclerotic plaques where the foam cells formed multiple conglomerates in contrast to native and prosthetic valves. However, the macrophage diversity in plaques (and also in calcified native AVs) was limited, whereas BHVs contained a number of macrophage subtypes (foam cells, canonical macrophages, mononucleated macrophages with multiple cytoplasmic inclusions, and multinucleated giant cells). The blood vessels in plaques and native AVs had distinct structural properties. Plaque neovessels frequently leaked and were responsible for intraplaque haemorrhages. While native AVs also relied on the microvasculature to meet the nutritional needs, it generally did not leak and intraleaflet haemorrhages have been associated with stress-driven ECM degradation and delamination. Substantial delamination and ECM disintegration were characteristic of BHVs which were incapable of regeneration. The calcification patterns in the arteries and heart valves were similar and included both large mineral deposits and microcalcifications. The schematic representation of the study findings is depicted in Figure 7.
Macrophage infiltration is evident in both atherosclerotic and valvular disease, being responsible for the neointimal and dystrophic remodeling, respectively [34][35][36][37]. The phenotyping of macrophages is mostly based on immunodetection methods such as immunostaining or flow cytometry upon fluorescence-or magnetic-activated cell sorting. Briefly, all macrophages are differentiated into pro-inflammatory (M1) and anti-inflammatory (M2), although this classification is oversimplified (at least for in vivo scenarios), and macrophages demonstrate a remarkable plasticity depending on the microenvironmental cues [38][39][40][41]. Although being widely established, this approach should ideally be combined with the ultrastructural interrogation of macrophages which also permits to identify their location and co-localisation with other plaque or valvular compartments. The current macrophage hierarchy does not consider their ECM remodelling activity, which is of crucial importance in the development of atherosclerosis, CAVD, and SVD [1][2][3]5,42], while backscattered electron microscopy permits the high-quality visualisation of cellular content and allows us to evaluate the integrity of the surrounding ECM.
Recent single-cell sequencing studies have indicated the leading role of non-lipid-laden macrophages in promoting the neointimal inflammation [43][44][45]. In our study, non-lipid-laden macrophages with numerous electron-dense granules and multinucleated giant cells were located in the areas with degraded ECM that might suggest their pro-inflammatory role. The appreciable diversity of ultrastructural macrophage phenotypes in BHVs in comparison with native cardiovascular tissues is probably explained by the amount of the heterologous biomaterial possessing residual immunogenicity, which is caused by galactose-α-1,3-galactose and N-glycolylneuraminic acid, essential components of bovine and porcine ECM that have been lost by humans during evolution [46]. These ultrastructural findings support the current vision on immune rejection as one the main factors contributing to the BHV failure [42]. In turn, the excess the lipid-laden macrophages in the plaques reflects the paramount role of lipid retention in atherosclerosis, whereas in dysfunctional heart valves it merely complements the mechanical stress and immune rejection mechanisms, though still being among the major contributors.
properties. Plaque neovessels frequently leaked and were responsible for intraplaque haemorrhages. While native AVs also relied on the microvasculature to meet the nutritional needs, it generally did not leak and intraleaflet haemorrhages have been associated with stress-driven ECM degradation and delamination. Substantial delamination and ECM disintegration were characteristic of BHVs which were incapable of regeneration. The calcification patterns in the arteries and heart valves were similar and included both large mineral deposits and microcalcifications. The schematic representation of the study findings is depicted in Figure 7. Common and unique ultrastructural features of atherosclerotic plaques, calcified native AVs, and failed BHVs. Atherosclerosis is mainly driven by a lipid retention accompanying by a foam cell formation. Subsequent migration and phenotypic switch of the vascular smooth muscle cells, active neovascularisation, and mineralisation further promote the development of atherosclerosis. Foam cells are also observed in the native AVs and BHVs, yet in considerably lower amounts, in particular when compared with non-lipid-laden macrophages. The highest macrophage diversity is observed in BHVs which additionally contain ECM-degrading/ECM-scavenging macrophages and multinucleated giant cells. In BHVs, mesenchymal cells form large outgrowths of a connective tissue which is termed pannus. Both heart valve types are characterised by protease-and fatigue-induced ECM degradation. Calcification and haemorrhages are common for all studied disorders, albeit intraplaque haemorrhages are caused by a neovessel leakage, while intravalvular bleedings are triggered by a loss of the ECM integrity.
(Neo)vascularisation provides a route for the migration of macrophages and other immune cells to the degraded ECM and altered microenvironment, thereby promoting inflammation and calcification in plaques and AVs [47][48][49][50]. While neovessels were detected in both plaques and AVs, plaque vessels were immature and leaky and often resulted in intraplaque haemorrhage, while valvular neovessels retained the integrity of the endothelial barrier. Nevertheless, haemorrhages were also common for the native AVs and BHVs because of ECM disintegration and delamination upon the critical valve fatigue [51]. Our results contradict some previously published studies which reported that microvascular leakage is a frequent phenomenon in CAVD [52,53], and this discrepancy warrants further investigation. The disorganisation of the ECM meshwork was notable in diseased native AVs and especially in the BHVs. Combined with intraleaflet haemorrhage, this provoked tissue swelling and delamination/pseudoaneurysm, further disrupting the distribution of the mechanical load and impairing the biomechanical properties of the valve. In addition, local haemorrhages may enhance the development of the valvular and vascular inflammation through iron oxidation and the subsequent generation of reactive oxygen [54] and nitrogen [55] species.
The ultimate outcome of inflammation, mechanical stress, and intraplaque or intraleaflet haemorrhage is calcification [56][57][58], which has a variety of patterns considerably affecting the disease course [59][60][61][62][63]. The dystrophic calcification of degraded ECM components and biomineralisation mediated by the osteochondrogenic differentiation of mesenchymal cells are two different modalities of valve and plaque ossification [1,[56][57][58]. Besides its trigger, extraskeletal calcification pattern is largely defined by the ambient conditions. Whether the mineral deposit is smooth or sharp, large or small, amorphous or crystalline strongly depends on microenvironmental pH as well as amount and proportions of available mineral ions [1]. Currently, microcalcifications < 5 µm diameter are not treated as hazardous [60], while calcifications between 5 and 100 µm diameter contribute to the plaque rupture because of uneven stress distribution across the delaminated tissue [61][62][63] and those > 100 µm diameter stabilise the plaque [61,63]. Despite the controversies on the role of different calcification modalities and patterns, it is clear that all of them are encountered in plaques, stenotic valves, and BHVs.
Differences in the prevalence and localisation of different ultrastructural features in distinct cardiovascular tissues may be linked to the site-or stage-specific alterations. Seemingly, the deposition of lipids in the ECM and their internalisation by macrophages play a key role in both the initiation and progression of atherosclerosis, as foam cells are abundant across the entire neointima. However, the localisation of lipid-laden macrophages in dysfunctional native or bioprosthetic valves is typically restricted to their inflow or outflow surfaces, suggesting that, in these settings, lipid retention and foam cells are responsible mainly for the initiation and do not have a pivotal significance further.
The main shortcoming of this study is that is has been limited to the observational approach and did not include any interventions. Nevertheless, here we focused on drawing parallels between the main inflammation-, remodeling-, and calcification-related disorders of the circulatory system components that did not require interventional study. To conclude, atherosclerosis, CAVD, and BHV failure share general pathogenetic mechanisms, while almost every of them has specific nuances and ultrastructural features which deserve further investigation.

Materials and Methods
Atherosclerotic plaques (n = 36) were excised during carotid endarterectomy conducted because of chronic brain ischemia. Calcified native AVs (n = 12) and BHVs (PeriCor or UniLine, NeoCor, Kemerovo, Russian Federation, n = 12) were obtained from the patients who underwent primary heart valve replacement because of CAVD or repeated heart valve replacement due to SVD, respectively. In-stent restenosis was the exclusion criterion for the patients with carotid atherosclerosis, whereas bicuspid AV, rheumatic heart disease, and infective endocarditis were exclusion criteria for those with CAVD. Median of BHV functioning was 5

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.