End stage renal disease‐induced hypercalcemia may promote aortic valve calcification via Annexin VI enrichment of valve interstitial cell derived‐matrix vesicles

Patients with end‐stage renal disease (ESRD) have elevated circulating calcium (Ca) and phosphate (Pi), and exhibit accelerated progression of calcific aortic valve disease (CAVD). We hypothesized that matrix vesicles (MVs) initiate the calcification process in CAVD. Ca induced rat valve interstitial cells (VICs) calcification at 4.5 mM (16.4‐fold; p < 0.05) whereas Pi treatment alone had no effect. Ca (2.7 mM) and Pi (2.5 mM) synergistically induced calcium deposition (10.8‐fold; p < 0.001) in VICs. Ca treatment increased the mRNA of the osteogenic markers Msx2, Runx2, and Alpl (p < 0.01). MVs were harvested by ultracentrifugation from VICs cultured with control or calcification media (containing 2.7 mM Ca and 2.5 mM Pi) for 16 hr. Proteomics analysis revealed the marked enrichment of exosomal proteins, including CD9, CD63, LAMP‐1, and LAMP‐2 and a concomitant up‐regulation of the Annexin family of calcium‐binding proteins. Of particular note Annexin VI was shown to be enriched in calcifying VIC‐derived MVs (51.9‐fold; p < 0.05). Through bioinformatic analysis using Ingenuity Pathway Analysis (IPA), the up‐regulation of canonical signaling pathways relevant to cardiovascular function were identified in calcifying VIC‐derived MVs, including aldosterone, Rho kinase, and metal binding. Further studies using human calcified valve tissue revealed the co‐localization of Annexin VI with areas of MVs in the extracellular matrix by transmission electron microscopy (TEM). Together these findings highlight a critical role for VIC‐derived MVs in CAVD. Furthermore, we identify calcium as a key driver of aortic valve calcification, which may directly underpin the increased susceptibility of ESRD patients to accelerated development of CAVD.

The pathophysiology of CAVD is complex, but shares similar mechanisms to physiological bone mineralization (Mohler et al., 2001).
Valve interstitial cells (VICs), the most abundant cell type in the aortic heart valve, play a key role in CAVD progression. Numerous studies have demonstrated the ability of VICs to undergo osteogenic transdifferentiation and calcification (Monzack & Masters, 2011;Osman, Yacoub, Latif, Amrani, & Chester, 2006). While the mechanisms underpinning this process have yet to be fully elucidated, it is highly plausible that parallels with chondrocyte and vascular calcification exist, whereby matrix vesicles (MVs) initiate the calcification process through interactions with collagen within the extracellular matrix (ECM) (Chen, O'Neill, Chen, & Moe, 2008;Kapustin & Shanahan, 2012;New et al., 2013).
However, the role of MVs in CAVD has yet to be fully determined.
This study aimed to characterize the role of MVs during aortic valve calcification. We have therefore, undertaken analysis of clinical CAVD tissues in conjunction with in vitro calcification studies in ratderived VICs to address the hypothesis that Ca and Pi induce aortic valve calcification through a MV-mediated mechanism.
These findings yield novel insights into the mechanisms of CAVD and highlight a critical role for the involvement of VIC-derived MVs.
Furthermore, we identify Ca as a key driver of aortic valve calcification, which may directly underpin the increased susceptibility of ESRD patients to accelerated development of CAVD. Life Technologies) and 1% gentamicin (Life Technologies) and cultured at 37°C, in the presence of 5% CO 2 . Before experimentation, isolated VICs were expanded in growth media for 2-4 passages, and cells used for experiments were between passage 4-6.

| Rat VIC line
The RVIC Sv40T (rat VIC-derived cell line) was generated by transducing primary rat VICs with recombinant lentivirus expressing Simian virus (SV) 40 large T antigen. Cell immortalization is achieved by silencing the expression of the tumor suppressors such as p53 and retinoblastoma protein (Rb), through a siRNA expressed by the lentivirus (Capital Biosciences, Gaithersburg, MD).

| Induction of calcification
Primary VICs were seeded in growth media at a density of 1.67 × 10 4 /cm 2 in multi-well plates. Calcification was induced as previously described in VSMCs (Reynolds et al., 2004). In brief, cells were grown to confluence (Day 0) before treatment with control (1.05 mM Ca/0.95 mM Pi) or test media; high calcium (Ca media; 2.7, 3.6, 4.5, or 5.4 mM Ca), high phosphate (Pi media; 2, 2.5, 3, 4, or 5.0 mM Pi), or both (CaPi media; 1.5-2.7 mM Ca/1.5-2.5 mM Pi). The standard CaPi media used was 2.7 mM Ca and 2.5 mM Pi. VICs were incubated for up to 5 days in 5% CO 2 at 37°C, and the medium was changed every 2nd/3rd day.

| Determination of calcification
Calcium deposition was quantified based on a method previously described (Zhu et

| MV isolation
MVs were obtained by differential centrifugation using a modified MV isolation protocol (Reynolds et al., 2004). Primary VICs or RVIC Sv40T cells were cultured in control (serum free growth media) or in serum free standard CaPi media, for 16 hr at 37°C. The media was subsequently aspirated and centrifuged at 3000 g for 20 min to pellet cell debris. The supernatant was then transferred to Beckman Coulter Ultra Clear™ ultracentrifuge tubes (VWR International Ltd, Lutterworth, UK) and centrifuged at 270,000 g for 1 hr at 4°C, using a Beckman Optima XL-90 ultracentrifuge (Beckman Coulter, Buckinghamshire, UK). The supernatant was discarded and the pellets were re-suspended in PBS. The concentration of pelleted MVs was measured using DC assay (Bio-Rad). Keynes, UK). Control sections were incubated with non-immune mouse or/and rabbit IgG (Sigma) (2 μg IgG/ml) in place of the primary antibody.

| Transmission electron microscopy
Human valve tissue samples were cut into 1-2 mm 3 pieces and fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4).
Samples were then processed with an automated routine tissue processor, Leica EMTP (Leica Biosystems) through a series of dehydration steps and penetration using TAAB 812 Epoxy resin (TAAB, Aldermaston, UK). The samples were then polymerized in TAAB embedding capsules (TAAB), sectioned at 0.5 μm, stained with toluidine blue, and scanned using a Zeiss Axioscan Z1 slide scanner (Carl Zeiss, Jena, Germany).
Ultrathin sections (thickness 90 nm) were prepared on a Leica UC6 (Leica Biosystems) and contrasted with 5% uranyl acetate for 15 min, and lead citrate for 5 min, on a Leica AC20 (Leica Biosystems). For immunogold labeling, the samples were initially fixed in 0.5% glutaradehyde in 4% PFA.
Following the same routine electron microscopy routine (using LR White Resin; TAAB), after ultrathin sectioning to 60 nm thick, the sections were collected on formvar-carbon coated nickel grids, then heated in citrate buffer (pH 6) before letting the samples cool down. The slides were blocked with 5% bovine serum albumin (BSA; Sigma) for 1 hr before incubating with anti-annexin VI antibody (1:1000; Santa Cruz Biotechnology, Dallas, TX), at 4°C, overnight. On the following day, they were incubated with rabbit anti-goat Gold antibody (1:300, Sigma) at room temperature, for 2 hr. Finally, they were contrasted with 5% uranyl acetate for 15 min, and lead citrate for 5 min as mentioned previously.

| Histology and immunohistochemistry
Tissues were fixed in 10% neutral buffered formalin (NBF) for 24 hr before being dehydrated, embedded in paraffin wax, and sectioned (4 μm) using standard procedures as previously described (Zhu et al., 2016). Sections were de-waxed in xylene and stained with Von Kossa or Alizarin red (Sigma) to visualize Ca deposition, and haematoxylin and eosin (H&E) to assess cell architecture. Immunohistochemistry was 2.10 | Analysis of gene expression RNA was extracted from VICs using RNeasy minikit (Qiagen, West Sussex, UK), according to the manufacturer's instructions. RNA was reverse transcribed and specific cDNAs were quantified by real-time PCR using the SYBR green detection method as previously reported CUI ET AL. | 2987 (Mackenzie et al., 2011;Staines, Zhu, Farquharson, & MacRae, 2014).
Primers were obtained from Qiagen and Primer Design (Primer Design, Southampton, UK).

| iTRAQ-based quantitative proteomics analysis
VICs were incubated in control and standard CaPi media for 16 hr, and MVs were isolated by differential centrifugation from the supernatants to produce two biological replicates for both control and CaPi media. MVs (25 μg

| iTRAQ data analysis
The raw mass spectrometric data files obtained were collated into a single data set using Proteome Discover version 2.0 (ThermoFisher Scientific) and the Mascot search engine version 2.4 (www.

| Western analysis
Cultured cells were lysed in RIPA buffer (ThermoFisher Scientific).

| Statistical analysis
Two-sample Student's t-test was used to analyze the significance between two sets of data. For more than two groups, one-way analysis of variance (ANOVA) using the general linear model (GLM) incorporating pairwise comparisons was performed. Data are presented as mean ± standard error of the mean (S.E.M). All statistical analysis was performed using Minitab 17 (Minitab Inc., Coventry, UK). p < 0.05 was considered to be significant, and p-values are represented as: *p < 0.05; **p < 0.01; ***p < 0.001.

| Ca regulates VIC calcification
Initial studies confirmed that rat primary VICs isolated in the present investigation were free from endothelial contamination. Cells were negative for the endothelial cell marker, CD31 (Figure 3a). In addition, cells showed positive staining for both α-SMA (green; Figure 3b) and vimentin (red; Figure 3c), in agreement with previous reports (Latif et al., 2015;Liu, Joag, & Gotlieb, 2007).
In ESRD, systemic Ca and Pi concentrations typically exceed 2.4 mM and 2.0 mM, respectively (Reynolds et al., 2004). To understand the accelerated progression of CAVD in patients with ESRD, it is essential to appreciate the calcification potential of VICs

| Proteomics analysis of MVs from calcified VICs reveals calcification regulators and exosome markers
To elucidate the specific structural and functional features of MVs derived from calcified VICs, we analyzed their protein composition using iTRAQ-based quantitative mass spectrometry analysis (Supplementary Table). Remarkably, a number of established exosomal proteins were enriched in MVs isolated from VICS cultured with standard CaPi calcification media, including CD9, CD63, LAMP-1, and LAMP-2 (Table 1). Of further interest was the concomitant upregulation of the calcium-binding Annexins (I, II, III, IV, V, VI, VII, and XI) (Table 2), which have been previously shown to specifically accumulate in chondrocyte and VSMC-derived mineralization competent MVs in a calcium-dependent manner (Kapustin et al., 2011;Wang, Xu, & Kirsch, 2003).

| Co-localization of Annexin VI with MVs in human CAVD tissue
Following recent reports establishing that Annexin VI is required for MV mediated VSMC calcification (Kapustin et al., 2011), we further

| Identification of novel putative MV pathways
With a view to highlighting novel biological processes underpinning MV-mediated calcification in VICs, we used IPA to identify canonical pathways, biological functions, and networks of interacting proteins.
Key pathways crucial to cardiovascular function were identified, including aldosterone, P2Y purinergic receptor signaling and thrombin signaling pathways (Figure 7a

| DISCUSSION
Given the significant contribution of increased circulating Ca to the arterial medial calcification associated with ESRD, in conjunction with reports showing the concentration of Ca in calcified aortic valves to exceed 13.5 mM/g tissue (Dahm et al., 2000), a more complete understanding of the role of calcium in aortic valve calcification is essential.
We report for the first time that VIC calcification in vitro can be driven by elevated Ca levels, with a concomitant increase in the expression of osteogenic markers Msx2, Runx2, and Alpl. Pi treatment alone was surprisingly ineffective at calcification induction. Intriguingly, a notable synergistic effect of Ca and Pi in combination on VIC calcification was observed, corroborating earlier reports demonstrating that elevated Ca induced human VSMC calcification in vitro with a synergistic effect of Ca and Pi (Reynolds et al., 2004). Abnormalities in Ca and Pi metabolism may therefore, directly underpin the increased susceptibility of ESRD patients to accelerated progression of CAVD.
This study offers new insight into the role of MVs in cardiovascular disease and provides direct evidence to suggest that MVs contribute to the pathological process of aortic valve calcification. Our experiments have established that MVs are released by viable VICs, particularly in the presence of elevated levels of extracellular Ca and Pi. These data support recent mechanistic studies of VSMCderived MVs (Kapustin et al., 2011), which proposed that Ca-loaded vesicles are released from cells to protect against the cytotoxic effects of intracellular Ca overload (Fleckenstein-Grün, Thimm, Czirfuzs, Matyas, & Frey, 1994;Hsu & Camacho, 1999). Furthermore, our ultrastructural analyses have identified the presence of MVs in calcified human aortic valves. These data extend previous reports demonstrating the existence of MVs in medial arterial calcification (Kim, 1976), and atherosclerotic intimal plaques (New et al., 2013).
Recent data have challenged current views on the plasma membrane origin of MVs, identifying VSMC-derived MVs as exosomes emanating from intracellular MV bodies (Kapustin et al., 2015). In the present study, iTRAQ-based quantitative mass spectrometry analysis revealed that a number of established exosomal proteins were enriched in MVs isolated from calcified VICs. It is therefore, conceivable that unlike bone-derived MVs, which are understood to be released through polarized budding (Cui et al., 2016), MVs implicated in cardiovascular calcification are of exosomal origin.
Further studies are therefore, required to more fully characterize the regulation of MV biogenesis in aortic valve calcification.

MVs produced by calcifying chondrocytes and VSMCs have been
shown to contain Annexin II, V, and VI, the membrane-associated proteins known to mediate Ca influx into MVs (Balcerzak et al., 2008;Kapustin et al., 2011;Xiao et al., 2007). Here we demonstrate the up-regulation of Annexins I, II, III, IV, V, VI, VII, and XI in calcifying VIC-derived MVs, underscoring the significance of the Annexin family in the process of cardiovascular calcification. We reveal for the first time not only the presence of Annexin VI expression in calcified human aortic valve tissue, but also the co-localization of Annexin VI expression within calcifying MVs. These data suggest a novel role for Annexin VI in the regulation of calcium homeostasis and MV release in CAVD, building on previous reports demonstrating a specificity of Annexin V for macrophage-derived MVs (New et al.,  (Balcerzak et al., 2008;Kapustin et al., 2011;Xiao et al., 2007).
Numerous proteins associated with aldosterone and thrombin canonical pathways were identified in our analysis, which have both been previously linked to accelerated vascular calcification (Borissoff et al., 2012;de Rita, Hackam, & Spence, 2012;Hillaert et al., 2013). P2Y purinergic receptor signaling was also up-regulated, which has been recently linked to aortic medial calcification in chronic kidney disease (CKD) patients through the activation by uridine adenosine tetraphosphate (Up4A) (Schuchardt et al., 2012). Additionally, we observed increased expression of several members of the Rho signaling pathway, which has been previously shown to be associated with VIC calcification in vitro (Gu & Masters, 2011). We also identified novel pathways associated with PARK7, which is involved in copper-dependent signaling and functions as a redox-sensitive chaperone and as a sensor for oxidative stress (Shendelman et al., 2004;Zhou et al., 2006).

DISCLOSURE
Authors wish to declare that there are no conflicts of interest.