The role of neutrophil elastase in aortic valve calcification

Normal human AV tissue was collected from a patient with aortic dissection, and calcified AV tissues were collected from three patients with calcific aortic valve disease undergoing aortic valve replacement. All the valve tissues were rapidly cut in two parts. One part was fixed in 4% paraformaldehyde, the other was frozen and stored in liquid nitrogen.

Serum was collected from 58 patients diagnosed with aortic valve stenosis. As controls, normal serum were obtained from 30 healthy people (Additional file 1: Table S1). Serum with hemolysis or blood lipid was excluded and stored at – 80 ℃ until assays were performed.

This study was approved by the Ethics Committee of Chongqing Medical University. Informed consent was obtained from all the patients and their family members.

All animal procedures were approved by the Ethics Committee of Chongqing Medical University. The male Apoe⁻/⁻ mice were obtained from the Animal Experiment Center of Chongqing Medical University. Six to 8 weeks mice were divided into four groups: normal diet (ND), normal diet with Alvelestat (ND + Alv), western diet (WD), western diet with Alvelestat (WD + Alv). WD group and WD + Alv group were fed western diet containing 40% fat for 16 weeks (XT108C, Jiangsu Xietong Pharmaceutical Bio-engineering Co., Ltd.). A dose of 3 mg/kg Alvelestat was used in this study to pharmacologically inhibit the NE activity in mice once a week based on the previous studies [14].

Echocardiography was conducted on mice using a Vinno 6 ultrasound system (VINNO, Suzhou, China) in this study. Mice were anesthetized by induction for 5 min with 5% isoflurane and maintained at 1.5% isoflurane with an oxygen flow rate of 2.0 L/min. The heart is imaged in a two-dimensional parasternal long axis view and an M-echocardiogram of the middle chamber is recorded at the myocardial level. Cusp separation and ejection fraction were obtained from the M-mode image. Doppler measures of blood velocity were all made with the intercept angle 60° between the targeted vessel and ultrasound beam.

The porcine aortic valve interstitial cells (pVICs) were isolated and cultured as previously reported [12]. In brief, the pVICs were removed in the sterile environment. In the biological safety cabinet, the surfaces of the valves were wiped repeatedly with sterile cotton swabs to remove the valve endothelial cells. Then valves were cut into small pieces of 2 × 2 mm and subjected to an 8-h collagenase digestion at 37 ℃. Then, pVICs were cultured in Media 199 (Hyclone, USA) supplemented with 10% fetal bovine serum (Cellmax, USA) and 1% penicillin/streptomycin (C0222, Beyotime Biotechnology). pVICs from passages 3–7 were used to following experiment.

To induce calcification, pVICs were cultured in osteogenic medium: M199 supplemented with 10% FBS, 1% penicillin/streptomycin, 10 mmol/L β-glycerin glyphosate (ST637, Beyotime Biotechnology), 50 µg/mL ascorbic acid (ST1434, Beyotime Biotechnology) and 100 nmol/L dexamethasone (ST1258, Beyotime Biotechnology). pVICs were treated with 1 µg/mL active human neutrophil elastase protein (ab91099, Abcam) for 12, 24, 48 h to assess its effect on pVICs apoptosis, phenotypic transition and inflammation. To further verify the effect of NE on pVICs calcification, pVICs were cultured in osteogenic medium with NE, 40 nmol/L Alvelestat (T3107, Topscience) and both. To further illustrate the different functions of PGRN and 45KD GRN on valve calcification, pVICs were added with 800 ng/mL recombinant PGRN protein or transfected with 45KD GRN virus or contrls virus in the presence of OM. In order to further verify the degradation effect of NE on PGRN, HEK293 cells were cultured in DMEM (Hyclone, USA) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Cells were added with 800 ng/mL recombinant PGRN protein with or without NE and Alvelestat. For the study of intracellular signaling pathways, pVICs were treated with HNE in the absence or presence of osteogenic medium for 24 h. The later signaling pathway inhibitor were added 1 h before stimulation with HNE: the NF-ĸB inhibitor BAY 11-7082 (Cat. No. 196870, Calbiochem), and the Akt inhibitor LY294002 (S1105, Selleckchem).

Detection of NE levels in serum was performed using ELISA according to the manufacturer’s standard protocol (DY008, R&D Systems).

NE activity in serum and supernatant were measured according to the procedure used by Virca, G.D. et al [15]. Briefly, the serum and supernatant were transferred into 96-well plates with 500 µM substrate solution N-methoxysuccinyl-Ala-Ala-pro-Val-p-nitroanilide(M4765, Sigma-Aldrich) in 20 mM Tris–HCl, pH 7.5, with the addition of 10 µg/mL human serum albumin (bs-0292P, Bioss) and incubated for 24 h at room temperature. The absorbance of N-methoxysuccinyl-Ala-Ala-pro-Val-p-nitroanilide cleaved by NE was measured at 405 nm.

Total RNA was extracted using RNA-quick purification kit (RN001, Esunbio) according to the manufacturer’ s recommendations. Reverse transcriptase reactions were performed using PrimeScriptTM RT Master Mix (RR036B, Takara). qRT-PCR was performed using 2X SYBR Green Fast qPCR Mix (RK02001, Biomarker) on a CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). All primer sequences are shown in Additional file 1: Table S2. Data were normalized by GAPDH expression and expressed as fold change relative to controls. All PCRs were performed at least in triplicate for each experimental condition.

All the AV tissues were dehydrated, embedded in paraffin, and cut in 5-µm-thick sections. The sections were deparaffinized with xylene and hydrated with gradient alcohol. For haematoxylin and eosin stain, sections were stained with hematoxylin (G1120, Solarbio) for 1 min and then with eosin for 1 min. For Alizarin red S staining, sections were stained with Alizarin Red S staining working solution (G3280, Solarbio) for 5 min and hematoxylin for 1 min. For immunohistochemical staining, after antigen retrieval with sodium citrate, sections were treated with 3% hydrogen peroxide for 10 min to block endogenous peroxidase and blocked by 5% normal goat serum (SP-9000, ZSGB-BIO) for 15 min, sections were incubated with primary antibodies against OPN and NE overnight at 4 ℃.Then sections were further incubated with biotin-labeled goat anti-rabbit IgG for 15 min and HRP-labeled Streptomyces ovoalbumin working solution for 15 min. The signal was revealed by using a DAB Substrate Kit (ZLI-9017, ZSGB-BIO).

Cultured pVICs were fixed in 4% paraformaldehyde and permeabilized by 0.5% Triton X-100(T8200, Solarbio). After blocked by 5% bovine serum albumin at room temperature for 30 min, cells were incubated with primary antibody overnight at 4 ℃ followed by AlexaFluor 647-conjugated anti-rabbit secondary antibodies (bs-0369 M, Bioss) for 2 h at room temperature in dark. Nuclei were stained with DAPI (C0065, Solarbio) for 10 min. Images were captured by Fluoview FV1000 laser scanning confocal system (Olympus, Tokyo, Japan).

A pool of three target-specific siRNAs was used to silence NE expression. The siRNAs targeting porcine NE transcript were designed and synthesized from RiboBio. A negative control (NC) siRNA was used as a scramble siRNA, targeting no known gene. Targeted sequences are as follows: siNE#1: CCGACGCTATGGTGATAAT; siNE#2: GCAGCATCATGCAGCATCT; siNE#3: GCAGCAGCTCAATGTGACT. pVICs reached at 70–80% consistency were transfected with 5 nM NE or NC siRNAs using Lipofectamine 2000 Transfection Reagent (Cat. No. 11668500; Invitrogen; Thermo Fisher Scientific, Inc.). Cell lysates were collected and examined for NE mRNA expression 24 h and protein expression 48 h after transfection.

Total protein were prepared by lysing the cells in RIPA buffer supplemented with protease inhibitors and phosSTOP phosphatase inhibitors (Cat. No. 56-25-7, Roche). The protein samples (20 µg) were separated by electrophoresis on a 10% SDS–polyacrylamide gel and transferred to a polyvinylidene fluoride membrane (C3117, Millipore). After blocking with 5% skimmed milk powder at 37 ℃ for 2 h, use the primary antibody overnight at 4 ℃ to detect the protein of interest (Additional file 1: Table S3). After washed three times in TBST, the membranes were incubated with corresponding secondary antibodies conjugated with horseradish peroxidase (ZB-2305, ZSGB-BIO) at dilution of 1:5000 for 1 h at room temperature. Finally, the membranes were visualized by enhanced chemiluminescence (4AW011-200B, 4A BIOTECH) and calculated the gray value of each band by using Image Lab software. The internal control GAPDH was used as control, and the ratios of the gray value of the target protein bands to the gray value of the corresponding internal control bands were used as the expression level of the target protein.

pVICs were seeded into the sterile coverslip (BS-14-RC, Biosharp) at a density of 1 × 10³ cells/well. After the cells are completely adherent, 1 µg/mL NE was added into the cells for 24 h. Then, apoptotic cells were detected using TUNEL assay kit (E-CK-A321, Elabscience) according to the manufacturer's standard protocol.

pVICs in the logarithmic growth phase were seeded into 96-well plates at a density of 5 × 10³ cells/well with three replicate wells. After 8 h incubation, cells were treated with 1 µg/mL NE for 24 h. The medium was removed and 10% MTT solution (5 g/L, M8180, Solarbio) was added into the plates. After 4 h incubation, the supernatant was discarded and 100ul DMSO (D8371, Solarbio) was added to each well. The absorbance at 490 nm was valued by the microplate reader.

pVICs were fixed in 4% paraformaldehyde for 20 min and washed in PBS for 5 min. Then, cells were incubated with 200ul alkaline phosphatase solution (C3206, Beyotime Biotechnology) or alizarin red solution (130-22-3, Solarbio) for 30 min. Cells were washed in distilled water for 5 min and then observed using a Leica DM IL LED microscope.

To investigate whether NE plays an important role in valve calcification, we detected the levels and activity of NE in patients with CAVD. As shown in Fig. 1A and B, the level and activity of NE in patients with CAVD were significantly higher than in non-CAVD group. Moreover, immunohistochemical analysis and Western blot results further showed that NE expression was increased in calcified valve tissues contrast with normal valve tissue (Fig. 1C, D). Besides, after pVICs were treated with osteogenic medium for 24 h, the expression and activity of NE were significantly increased (Fig. 1E–G).

We then further established the role of NE on apoptosis and inflammation in pVICs. Immunofluorescence results presented that fluorescence intensity was enhanced after HNE treatment, suggesting that NE can enter into cells to function (Additional file 1: Fig. S2). The result of TUNEL assay showed NE treatment induced more apoptosis of pVICs (Fig. 2A) while no obvious effect on cell proliferation (Fig. 2B). Meanwhile, NE promoted the increase of bax and the decrease of bcl-2 in pVICs (Fig. 2C–E). We then examined the role of NE in phenotypic transformation and inflammation in pVICs at different time points. NE increased α-SMA and Collagen I mRNA expression at 12 h and elevated RUNX2 mRNA expression from 24 to 48 h of treatment (Fig. 2F). In addition, NE increased the expression of inflammatory factors TNF-α, IL-6 and IL-1β (Fig. 2G).

As shown in Fig. 3A–H, NE induced a significant increase of RUNX2 and OPN at both mRNA and protein levels under the induction of osteogenic medium. Besides, NE also stimulated the expression of TNF-α and bax, although no elevation of bcl-2. For confirmation, pVICs were treated with NE in combination with its inhibitor Alvelestat for 24 h in osteogenic medium. The mRNA and protein results indicated that Alvelestat prevented NE from promoting the increase of RUNX2, OPN, TNF-α and bax. ALP staining and Alizarin Red Staining results showed that the number of ALP positive cells and calcium deposition were markedly increased after 7 or 14 days of incubation in osteogenic medium in the presence of NE, compared with OM group. However, Alvelestat still reversed this effect (Fig. 3I, J).

To verify whether inhibition of NE activity affects pVICs calcification, pVICs were treated with 40 nmol/L Alvelestat to observe its effect on calcification, inflammation and apoptosis related indicators. As shown in Fig. 4A, Alvelestat reduced the increase of NE activity induced by OM. Also, Alvelestat inhibited early fibrosis index α-SMA and Collagen I at 24 h (Fig. 4B–D) and inhibited calcification indicators RUNX2 and OPN, inflammatory factor TNF-α, apoptosis protein bax as well as NE at 48 h (Fig. 4E–K). Furthermore, Alvelestat reduced ALP staining and calcium deposition (Fig. 4L, M).

We have previously demonstrated that NE expression was elevated during valve calcification. Next, we transfected Ne-targeted siRNA into pVICs to verify the effect of NE knockdown on valve calcification. The NE knockdown efficiency was validated by markedly diminished NE mRNA level in pVICs (Fig. 5A). With NE knockdown in pAVICs, we found reduced mRNA and protein levels of RUNX2 and OPN. In line with this, NE silencing also inhibited TNF-α and bax (Fig. 5B–I). In addition, both Alizarin red and ALP staining results displayed that NE deficiency decreased the osteogenic differentiation and calcium deposition (Fig. 5J, K).

We previously reported that 45KD GRN, the degradation fragment of PGRN is markedly increased in the calcified AV tissues and we further studied the effect of full-length PGRN and 45KD GRN on valve calcification [4]. As shown in Fig. 6A, PGRN inhibited the expression of RUNX2 and OPN induced by OM, while 45KD GRN had the opposite biological function. PGRN can be cleaved into short peptides by NE [5]. Therefore, we hypothesized that the increase of 45KD GRN in calcified valves was due to elevated NE levels. We examined the effect of NE treatment on 45KD GRN in the presence or absence of calcification induction in pVICs. To evaluate whether NE can regulate PGRN at the transcriptional level, mRNA levels of PGRN were assessed by quantitative real-time PCR after treatment with NE in the presence or absence of OM. The Q-PCR results showed that there was no significant difference in mRNA level of PGRN after NE treatment (Fig. 6D, F). However, Western blot results showed that 45KD GRN, the degradation fragment of PGRN was notably increased after NE treatment (Fig. 6E, G). Similarly, in 293 cells, NE treatment also reduced the full-length PGRN and promoted the production of 45KD GRN (Fig. 6I).

The possible intracellular mechanisms by which NE exerts its procalcification effects in pVICs were analyzed. We examined the effects of NE on classical osteogenic signaling pathway Smad1/5/8, non-classical osteogenic related signaling pathway ERK, AKT and NF-κB. As shown in Fig. 7A, NE induced phosphorylation of AKT and NF-κB while had no effect on phosphorylation of smad 1/5/8 and ERK. Furthermore, NE promoted OM-induced phosphorylation of AKT and NF-κB (Fig. 7F). LY294002, the inhibitor of AKT pathway and BAY11-7082, the inhibitor of NF-κB pathway, were used to evaluate whether they can reverse the pro-calcification effect of NE. LY294002 and BAY11-7082 were significantly blocked NE-induced upregulation of osteogenic markers OPN expression (Fig. 7K, L).

To further explore the translational value of NE inhibition in prevention of valve calcification, Alvelestat was orally administered to APOE⁻/⁻ mice. As shown in Fig. 8A and B, APOE⁻/⁻ mice fed the western diet showed significant valve thickening and increased NE and fibrotic markers α-SMA. While Alvelestat significantly reduced the valve thickness and NE and α-SMA expression. We then used echocardiography to assess left heart function in APOE⁻/⁻ mice. WD group exhibited increased transaortic peak velocity compared with ND + Alv group, while other indicators, such as EF and cusp separation, were not statistically different (Fig. 8C–E). Meanwhile, increased NE activity was observed in WD group compared with ND group. Significantly lower levels of NE activity were observed in mice administered Alvelestat (Fig. 8F).

For decades, CAVD was considerd as a passive degenerative disease. But recent studies have shown that it is an active and dynamic disease which is mediated by endothelial dysfunction, inflammation and lipid deposition [16]. In terms of pathogenic factors, CAVD shares many pathophysiological features with atherosclerosis. Aortic valve tissue is divided into three layers: fibrosa, spongiosa and ventricularis [17]. During the disease progression, the valve leaflets gradually become thicker and stiff, with structural and collagen disorders [18]. In line with this, we found that compared with normal valve tissue, calcified valves were significantly thickened with increased elastin fiber fragments and calcium nodules. The valve interstitial cells (VICs) are the main cell type of valve. VICs exhibits mesenchymal phenotype characterised by high expression of α-SMA (alpha-smooth muscle actin) and vimentin instead of CD31 [19]. Therefore, we used these mesenchymal markers to phenotypic identify isolated porcine valve interstitial cells prior to the initiation of in vitro experiments (Additional file 1: Fig. S1).

Inflammation is essential for initiating valve calcification and accompanying the entire process of calcification [20]. Meanwhile, NE is closely associated with some inflammatory diseases such as arthritis [21] and acute pancreatitis [22]. NE levels are rapidly upregulated during inflammatory responses. The first finding of this study was that the expression and activity of NE of CAVD-patients was higher than that of control group. Besides, In vitro osteogenic medium also promoted the expression and activity of NE in pVICs, suggesting that it may play an important role in valve calcification. This experiment confirmed for the first time that valve interstitial cells could also secrete NE. In fact, NE was not only produced by neutrophils, but also secreted by macrophages and endothelial cells [23]. Since previous studies have reported that there are infiltration of macrophages, monocytes and other inflammatory cells in the pathological valve tissue [24], we consider that these inflammatory cells may be one of the sources of NE in calcified valve tissues. Our team will further investigate the mechanism of NE elevation in calcified valves.

Apoptosis is a crucial regulator of initiation and progression of CAVD [25]. Meanwhile, it is reported that NE could induce endothelial cells apoptosis [26]. Here, we found that NE promoted bax protein level with or without osteogenic medium induction. NE had no significant effect on the expression of bcl-2 under the conditions of OM induction. We speculate that it may be due to the negative feedback regulation mechanism of pVICs activated by OM. In general, NE exhibited the effect of promoting pVICs apoptosis, which suggests that NE may partly influence the osteogenic differentiation of pVICs through apoptosis.

In healthy valves, pVICs remain in a quiescent state [27]. With the change of valve homeostasis, pVICs are activated and further differentiated into myofibroblast-VIC characteristiced by high expression of α-SMA and osteogenic-VIC characteristiced by high expression of RUNX2 [28]. Therefore, we examined the effect of NE treatment on VIC phenotypic transformation at different time points. Our results demonstrated that NE promoted the activation of pVICs in the early stage, and then promoted the osteogenic differentiation of pVICs accompanied by an increase in inflammatory factors.

We then evaluated the role of NE on VIC during OM induction. As expected, NE aggravated the osteogenic differentiation, apoptosis and infalmmation of pVICs induced by OM, together with increased ALP and calcium deposits. However, these results were reversed by NE inhibitor Alvelestat. Alvelestat (AZD9668) is a novel, orally bioavailable NE inhibitor [29]. Early NE inhibitors, such as Sivelestat, bind irreversibly to NE, which causes serious side effects [30]. Compared to Sivelestat, the interaction between Alvelestat and NE is rapidly reversible. Moreover, clinical studies were conducted on pharmacology, tolerability and safety of Alvelestat. The results showed that Alvelestat could be quickly absorbed after oral administration, eliminating half-life between 5 and 15 h, and most of Alvelestat could be eliminated by the kidneys [31]. Therefore, it is less toxic than other early NE inhibtors. At present, the application of Alvelestat in the treatment of COPD and bronchiectasis has entered the clinical phase II study, suggesting its better clinical application prospect [32‐34]. Alvelestat also has been reported to be protective in some inflammatory diseases, such as abdominal aortic aneurysms [35] and cystic fibrosis [36]. However, its role in cardiovascular diseases, especially CAVD, has not been studied. Hence, we further studied the role of Alvelestat in OM-induced osteogenic differentiation of pVICs. We chose a concentration of 40 nmol/L Alvelestat for this experiment, because previous studies has reported that the pIC50 values of Alvelestat for the whole-blood and cell-associated assays was about 40 nmol/L [33]. Our analysis indicated that Alvelestat could effectively inhibit the increase in NE activity and expression induced by OM. Meanwhile, Alvelestat also had a certain inhibitory effect on apoptosis and inflammation. These results suggest the potential of Alvelestat to inhibit osteogenic differentiation of pVICs. To achieve satisfactory inhibition of NE, we next used siRNA against NE to knockdown the NE mRNA expression during OM induction. Similarly, NE silencing also reduced OM-induced apoptosis, inflammation and osteogenic differentiation of pVICs. In addition, in order to further increase the clinical value of Alvelestat, we then explored the effect of Alvelestat on APOE⁻/⁻ mice fed with western diet. Our results showed that Alvelestat markly remitted the thickening and fibrosis of the valves. The results of echocardiography were further revealed that Alvelestat could decreased the transaortic peak velocity to alleviate cardiac function in APOE⁻/⁻ mice fed with western diet. Although there was no statistical difference in trends, mice in the WD group showed decline in cusp separation and ejection fraction. The reason for this phenomenon may be due to the small number of mice in the WD group. Therefore, the number of mice in WD group needs to be expanded to obtain more stable results.

Our group previously investigated the role of the inflammatory regulator PGRN in valve calcification and we found for the first time that 45KD GRN, a degraded fragment of PGRN, had a characteristic increase in calcified valve tissues. We demonstrated that full-length PGRN antagonized TNF-α and inhibited valve calcification while 45KD GRN played the opposite biological function of pro-inflammatory and pro-calcification [12]. So why does the degradation of PGRN to 45KD GRN increase when the valve is calcified? We hypothesized that this phenomenon may be caused by the elevation of a certain enzyme related to PGRN degradation. It is well known that PGRN can be cleaved by neutrophil elastase and degraded into short peptides called GRN [13]. Our experimental results showed that NE levels and activity were significantly increased during valve calcification. In vitro experiments showed that NE promoted the production of 45KD GRN in the presence or absence of OM. Overall, NE plays its role in promoting valve calcification at least partially by degrading PGRN to 45KD GRN.

We then explored the signaling pathways involved in NE promoting osteogenic differentiation in pVICs. The effect of NE seems to be mediated by several pathways. It has been reported that NE increased the nuclear translocation of NF-κB p-P65 and thus relieved LPS-induced lung injury of rats [37]. Another research reported that NE activated ERK and induced IL-8 production [38]. In addition, NE also induced tumor cell survival and migration via Src/PI3K-dependent activation of AKT signaling [39]. However, the effect of NE on Smad 1/5/8 signaling pathway is unclear. In our current study, we found that NE did not affect the classical osteogenesis Smad 1/5/8 signaling pathway and the non-classical osteogenesis-related ERK signaling pathway, but activated the NF-κB and AKT signaling pathway in the presence or absence of OM. Further verification was carried out by using BAY11-7082, the inhibitor of NF-KB, and LY294002, the inhibitor of AKT pathway.

We have discovered that NE accelerates osteogenic differentiation of pVICs through promoting the production of 45KD GRN and activating the NF-KB and AKT pathways, thus enhancing valve calcification progression. In addition, we found that inhibition of NE by Alvelestat significantly attenuates valve calcification induced by OM. These findings open the possibility that NE may be a potential therapeutic targets for CAVD.