Biofunctionalization of decellularized porcine aortic valve with OPG-loaded PCL nanoparticles for anti-calcification

Decellularized valve stents are widely used in tissue-engineered heart valves because they maintain the morphological structure of natural valves, have good histocompatibility and low immunogenicity. However, the surface of the cell valve loses the original endothelial cell coverage, exposing collagen and causing calcification and decay of the valve in advance. In this study, poly ε-caprolactone (PCL) nanoparticles loaded with osteoprotegerin (OPG) were bridged to a decellularized valve using a nanoparticle drug delivery system and tissue engineering technology to construct a new anti-calcification composite valve with sustained release function. The PCL nanoparticles loaded with OPG were prepared via an emulsion solvent evaporation method, which had a particle size of 133 nm and zeta potential of −27.8 mV. Transmission electron microscopy demonstrated that the prepared nanoparticles were round in shape, regular in size, and uniformly distributed, with an encapsulation efficiency of 75%, slow release in vitro, no burst release, no cytotoxicity to BMSCs, and contained OPG nanoparticles in vitro. There was a delay in the differentiation of BMSCs into osteoblasts. The decellularized valve modified by nanoparticles remained intact and its collagen fibers were continuous. After 8 weeks of subcutaneous implantation in rats, the morphological structure of the valve was almost complete, and the composite valve showed anti-calcification ability to a certain extent.


Introduction
Heart valve disease is the leading cause of cardiovascular death, where calcied aortic valve disease (CAVD) remains dominant. 1 The incidence of CAVD has increased continuously with the increase in the aging population, which ranks second among the cardiovascular diseases. 2 To date, there is no effective drug for the prevention or treatment of CAVD, and the main clinical treatment for CAVD is valve replacement surgery. However, the existing mechanical valves require the long-term use of anticoagulant drugs. The same or different types of biological valves are calcied and have a high chance of decay, which affect the quality of life of patients aer surgery, resulting in great suffering. 3 Therefore, the development of new types of anti-calcication heart valves is necessary, which remains the goal of clinical researchers.
Tissue-engineered heart valves (TEHV) are an ideal valve replacement approach that overcomes the deciencies of existing mechanical and biological valves. They have broad application prospects clinically due to their characteristics of self-renewability and remodeling, no anticoagulation, low immunogenicity, durability, calcication, etc. 4,5 The scaffold materials used for TEHV mainly include polymer materials and natural decellularized materials. Decellularized valve stents are widely used in TEHV because they maintain the morphological structure of natural valves, have good histocompatibility and low immunogenicity. 6,7 However, the surface of the cell valve loses its original endothelial cell coverage, exposing collagen and causing calcication and decay of the valve in advance. 8 Therefore, it is of great value to biomodify decellularized valves for anti-calcication.
Nano drug-loading systems are a new type of drug carrier with broad application prospects for development. Nanoparticle drug-loading systems enhance the stability of protein drugs and have excellent sustained-release and controlledrelease properties, helping to achieve local therapeutic effects. 9 The drug-loading materials mainly include natural and synthetic polymers. Among them, nanoparticles prepared using synthetic polymer materials have a large range of applications due to their high bioavailability, good encapsulation, control release and non-toxic properties. 10,11 Polycaprolactone (PCL) is a hydrophobic biodegradable polymer with good biocompatibility due to its linear aliphatic polyester, which is obtained via ring-opening polymerization using 3-caprolactone (3-CL). However, PCL degrades rather slowly and has low biocompatibility with so tissues, which restrict its clinical application. 10 Polyethylene glycol (PEG) is a long-chain polymer with high hydrophilicity, non-toxicity and good histocompatibility. It has a wide range of applications in the medical and health elds. 12 The PEG-PCL copolymer is formed by binding PEG to PCL, which has greatly improved hydrophilicity, biodegradability and mechanical properties. Therefore, PEGylated PCL nanoparticles are considered to be more suitable for drug delivery than PCL nanoparticles alone. 13 Osteoprotegerin (OPG) is a soluble secretory glycoprotein belonging to the TNF receptor superfamily. Its N-terminal cysteine is required for the formation of disulde bonds in homodimers. 14 OPG competitively prevents RANK ligands from binding to the nuclear factor kB receptor activating factor and inhibits the abnormal osteogenesis of vascular and valvular stromal cells. Thus, the relative or absolute deciency of its secretion may be an important cause for valvular calcication. 15 Previous studies have shown that PEG cross-linked to decellularized heart valves improved the mechanical and biological properties of the valve stent. 16 In this study, the watersoluble protein OPG was encapsulated in MAL-PEG-PCL-modied PCL nanoparticles via a nano drug-loading system, which improved the bioavailability of OPG, and then the nanoparticles were terminated by Michael addition reaction. The combination of unsaturated carbon-carbon double bonds and thiolated decellularized valves allowed the nanoparticles to bridge the decellularized valve for the construction of a novel composite valve with sustained release function.

Methods
Preparation of nanoparticles loaded with OPG. MAL-PEG-PCL-modied PCL nanoparticles were prepared via an emulsion solvent evaporation method. The schematic diagram of the preparation of nanoparticles loaded with OPG is shown in Fig.  1. Briey, OPG (1 mg mL À1 ) was dissolved in sterile trihydrated water and the concentration was adjusted to 1 mg mL À1 . Then 10 mg soybean phospholipid was accurately weighed and dissolved in 1 mL of t-butanol, and then in 1 mL of 1 mg mL À1 OPG aqueous solution was added to 1 mL of 10 mg mL À1 soybean phospholipid/tert-butanol. The solution was thoroughly blown with a pipette, pre-frozen at À60 C for 2 h in a freeze dryer, and then dried for 22 h to obtain an OPG-phospholipid complex. The blank phospholipid complex was prepared as described before. As shown in Fig. 1, the O/W emulsion was rst prepared, and then 4 mg of MAL-PEG-PCL, 12 mg of PCL and 10 mg of OPG-phospholipid complex were dissolved in 1.5 mL of methylene chloride as the oil phase (O), and 6 mL 2% (w/v) PVA aqueous solution as the aqueous phase (W). The oil phase was slowly added to the aqueous phase and magnetically stirred for 1 min. The stirred mixed solution was ultrasonicated in an ice bath (SCIENTZ-II D, Ningbo Xinzhi Bio Technology Co., Ltd.), ultrasonicated at a power of 60 W for 2 min, and then stirred at room temperature for 5 h to fully volatilize the methylene chloride and nally obtain the MAL-PEG-PCL-modied OPG nanoparticles (OPG-NPs). Non-loaded OPG nanoparticles (NL-NPs) were prepared similarly, except that a blank phospholipid complex was added.
Characterization of nanoparticles loaded with OPG Particle size and zeta potential. The particle size and zeta potential distribution of the nanoparticles were measured via laser light scattering (LLS, Zetasizer Nano, Malvern, UK). 1 mL of the prepared nanoparticle suspension was diluted 4 times with triply distilled water, thoroughly mixed, and then the particle size and zeta potential distribution were determined using a measuring instrument.
Transmission electron microscopy. The morphology and particle diameter of the nanoparticles were observed via transmission electron microscopy (TEM, JEM-2100, JEOL Ltd.). The prepared nanoparticle suspension was diluted 100 times with triply distilled water, and thoroughly mixed. Then 10 mL of the diluted nanoparticle suspension was dropped on a copper mesh covered with a support lm, which was placed in a ventilated area to dry naturally and then dropped in 2% phosphorus. The tungstic acid solution was dried for 2 min. Aer drying, the copper mesh was placed under TEM to observe the morphology of the nanoparticles.
Encapsulation efficiency. The content of encapsulated OPG in the nanoparticles was detected using an enzyme-linked immunosorbent assay (ELISA). Briey, the prepared nanoparticle suspension was diluted to 10 mL with PBS and mixed thoroughly. Subsequently, 6 mL of nanoparticle suspension was aspirated by ultracentrifugation (32 000 rpm, 20 min, 4 C), and then the supernatant was collected. 1 mL of nanoparticle suspension was added to an EP tube and 20 mL of dimethyl sulfoxide (DMSO) was added, and then the mixture was thoroughly mixed.
According to the manufacturer's instructions for the OPG ELISA kit, 100 mL of standard, supernatant and DMSO and nanoparticle suspension were added to the corresponding wells of the plate, incubated at 37 C for 1 h, and then the liquid in the well was discarded. Subsequently, 100 mL of test solution A was added to each well, incubated at 37 C for 1 h, and then the liquid in the well discarded. 350 mL washing solution was added to each well, which was washed 3 times. Then the enzyme plate was dried and 100 mL of test solution B added to each well and incubated at 37 C for 30 min. The liquid in the well was discarded and the well was washed with washing solution 5 times, and then 90 mL of TMB substrate solution was added to each well, and color development was observed at 37 C for 20 min in the dark. Then 50 mL of reaction stop solution was added to each well, and nally the ELISA plate was placed on the multifunctional enzyme. The absorbance of each well was measured at 450 nm by placing it on a standard instrument (VARIOSKAN, Thermo Fisher Scientic, USA). The supernatant OPG content (M 0 ) and the total OPG content of the nanoparticle suspension (M 1 ) were calculated separately. The nanoparticle encapsulation efficiency (EE) was calculated as follows: EE (%) OPG cumulative release of nanoparticles. PBS (pH 7.4) was used as the release medium for the release of the nanoparticles in vitro, and the released OPG content was detected via ELISA. The prepared nanoparticle suspension was diluted to 10 mL with PBS and mixed thoroughly. 6 mL of nanoparticle suspension was placed in an ultracentrifuge tube and centrifuged at 32 000 rpm for 20 min at 4 C, the supernatant discarded, and the nanoparticles resuspended in 5 mL PBS. The pellet was placed in a constant temperature water bath oscillator (SHA-BA, Changzhou Langyue Instrument Manufacturing Co., Ltd.) and continuously shaken at 37 C and 60 rpm. Aer intervals of 6 h, 12 h, 24 h, 2 d, 3 d, 5 d, and 7 d, the tubes were taken out of the shaker bath, centrifuged at 4 C and 32 000 rpm for 20 min, and the supernatant collected. The nanoparticle pellet was then resuspended in 5 mL PBS, placed in a constant temperature water bath shaker, and then the supernatant was collected again at the next time point.
Isolation and culture of rat bone marrow mesenchymal stem cells (rBMSCs). According to the method described by Li et al., 17 3 week-old male Sprague-Dawley rats were sacriced by cervical dislocation. The rats were immersed in 75% ethanol for 15 min. The femur and tibia of the rats were isolated under aseptic conditions. The bone marrow cavity was washed with a medium containing 10% FBS, 100 U mL À1 penicillin, and 100 mg mL À1 streptomycin using a 5 mL syringe. The cell suspension was centrifuged at 1000 rpm for 5 min, the supernatant was discarded, the cells were re-suspended in complete medium, thoroughly mixed and then placed in a new culture ask at 37 C in a 5% CO 2 incubator (HERAcell 150i, American Thermo Scientic).
Flow cytometry. Second generation BMSCs were collected, trypsinized and then centrifuged (1000 rpm, 5 min). The supernatant was discarded, the cells were suspended in sterile PBS and the cell density was adjusted to 1 Â 10 6 /mL, and 500 mL of cell suspension was pipetted into a ow tube. 10 mL of FITClabeled CD29, CD90, and CD45 and 10 mL of PE-labeled CD34 antibodies were added to each ow tube. The isotype IgG antibody was used as an isotype control, incubated at 4 C for 30 min in the dark, and then placed in a ow cytometer (FACS Calibur, USA, BD) to detect the expression of each antibody. Cytotoxicity of nanoparticles. The effect of the OPG nanoparticles on the proliferation of BMSCs was detected via the CCK-8 method. The prepared nanoparticle suspension was centrifuged at a low temperature (32 000 rpm, 20 min, 4 C), the supernatant was discarded, resuspended in 2 mL of sterile PBS, and then ltered through a 0.22 mm sterile membrane for use. The experiment involved the blank nanoparticle group (NL-NPs), OPG-loaded nanoparticle group (OPG-NPs), PBS and blank control group. Third generation BMSCs were inoculated on a 96-well culture plate at a density of 1 Â 10 4 /mL, each well was inoculated with 100 mL of cell suspension, placed at 37 C, incubated in a 5% CO 2 incubator for 24 h, and then the corresponding 10 mL NL-NPs, OPG-NPs, and PBS were added to the wells. Only 100 mL PBS was added to the blank wells. Aer 12 h, 24 h, and 48 h, the cell culture was replaced with new culture medium, followed by the addition of 10 mL CCK-8 test solution. The cell culture was le to incubate for 2 h and then placed on an enzyme label instrument. The absorbance values of the respective wells were measured at 450 nm on a detector.
Osteogenic differentiation. Complete medium for osteogenic differentiation containing DMEM high glucose medium, 10% FBS, 100 U mL À1 penicillin, 100 mg mL À1 streptomycin, 10 mmol L À1 b-glycerophosphate, 50 mg mL À1 ascorbic acid, and 100 nM dexamethasone, hereinaer referred to as DAG, induces the differentiation of BMSCs into osteoblasts. 17 An appropriate amount of prepared OPG nanoparticle (OPG-NPs) and blank nanoparticle (NL-NPs) suspension was taken, centrifuged at low temperature (32 000 rpm, 20 min, 4 C), the supernatant discarded, and then re-suspended in 2 mL sterile PBS. Subsequently, it was ltered through a 0.22 mm sterile lter for further use. The third generation BMSCs were seeded in a 24-well culture plate at a density of 2 Â 10 4 /mL, and cultured at 37 C in a 5% CO 2 incubator. The DAG, OPG-NPs, and NL-NPs groups included 40 mL PBS, OPG-NPs, NL-NPs and 500 mL DAG mixed solution and 540 mL complete culture only was added in the normal control group. Six replicate wells for each experimental group were prepared. When the cell fusion reached 60-70%, the medium in the wells of the culture plate was carefully discarded, thoroughly rinsed with sterile PBS, and then the respective mixture was added to each well. The medium was changed once a week. Calcication was evaluated on days 7, 14, and 21 aer osteogenic differentiation.
Alizarin red staining. Alizarin red forms a red complex with calcium salts, and the formation of calcium nodules was observed by alizarin red staining. At day 21, the plate was washed with PBS, 500 mL of 4% paraformaldehyde was added to each well, xed at room temperature for 30 min, and rinsed thoroughly with PBS. Subsequently, 500 mL of alizarin red dye solution was added to each well, stained for 5 min, rinsed with PBS and then observed under a light microscope.
Calcium deposition test. The calcium deposition in each group was quantitatively detected using the calcium ion content in the calcium nodules. Firstly, 500 mL of 0.6 N HCL was added to each well, and then the mixture was incubated overnight at 37 C. The extracts of each group were collected the next day, centrifuged at 2500 rpm for 5 min, and then the supernatant was collected.
According to the manufacturer's instructions of the calcium quantitative test kit, 5 mL of supernatant was added to the bottom of a 96-well culture plate, and then 200 mL of the test solution was added to each well. Aer standing at room temperature for 5 min, the absorbance values were measured at 630 nm.
Preparation of decellularized valve modied with OPG nanoparticles. Fresh porcine hearts were obtained from the local slaughterhouse under clean conditions, and then the porcine aortic valve (PAV) was cut in a sterile environment. The aortic valve was then preserved in PBS containing antibiotics (0.1 mg mL À1 streptomycin, 100 U mL À1 penicillin, and 0.25 mg mL À1 amphotericin B). The obtained aortic valve was placed in PBS containing 0.05% (w/v) trypsin and 0.02% (w/v) EDTA at 37 C for 12 h. It was further placed in PBS containing 1% (v/v) Triton X-100 at 4 C for 48 h, washed with PBS, and then treated with PBS containing nuclease (DNase I 0.2 mg mL À1 and RNase A 20 mg mL À1 ) at 37 C for 1 h. Aer depletion with PBS, a decellularized porcine aortic valve (DPAV) was obtained. According to Chen, 18 the prepared decellularized valve was reacted with thiolation reagent SATA at 37 C for 2 h, and then the reaction was terminated by elution with PBS. Subsequently, it was reacted with hydroxylamine hydrochloride at 37 C for 2 h to protect the acetylated thiol group, and washed with PBS for thiolation.
The thiolated DPAV was immersed in a diluted OPG-NPs solution, and then placed on a constant temperature shaker at 37 C, 60 rpm in the dark for 8 h. The nanoparticles that were not bound to the valve were washed away with PBS to nally obtain the OPG-loaded PCL nanoparticle-modied composite valve (OPG-NPs-DPAV). The nanoparticle-modied DPAV without OPG was also prepared using this method (NL-NPs-DPAV).
Characterization of the complexed DPAV HE and MASSON staining. The valves in each group (PAV, DPAV, OPG-NPs-DPAV, and NL-NPs-DPAV) were xed with 4% paraformaldehyde for 24 h, embedded in paraffin, sectioned, and stained with reference to HE and MASSON staining kit instructions. The decellularization effect and the valvular collagen bers were then observed under a light microscope.
Scanning electron microscopy. DPAV and OPG-NPs-DPAV were xed in 2.5% glutaraldehyde for 24 h, dehydrated using an ethanol gradient, dried at the CO 2 critical point, sprayed with gold by ion sputtering, and then placed under a scanning electron microscope (FEI Quanta 200F, American EI) to observe the combination of granules and decellularized valves.
Attenuated total reection Fourier infrared spectroscopy (ATR-FITR). OPG-NPs-DPAV, NL-NPs-DPAV and DPAV were plated on a watch glass, frozen in a freeze dryer at À56 C for 3 h, and then vacuum dried for 20 h. The dried groups of the valves were ground into a powder and thoroughly mixed with potassium bromide. An appropriate amount of each sample was utilized to form a tablet, and then its infrared absorption spectrum was measured by Fourier transform infrared spectroscopy (FTIR, Nicolet 5700, Nico, USA).
Anti-calcication of composite valve in vivo. The rat subcutaneous implantation model is used as a routine method for the detection of the anti-calcication biological properties of biological valves. 19 To simulate the calcication of valves in each group (PAV, DPAV, OPG-NPs-DPAV, and NL-NPs-DPAV) in vivo, 4 week-old male SD rats were divided into 4 groups, with 9 in each group, and then anesthetized by intraperitoneal injection of 1% sodium pentobarbital. The subcutaneous tissue was blunt free. The valves of each group (1 cm Â 1 cm) were sutured and xed on the fascia, with one valve in each incision. Two valves were placed per rat. The back incision was sutured and then the back suture was removed aer 7 days. The animals were sacriced by CO 2 asphyxiation aer 2, 4 and 8 weeks and the samples were retrieved. Aer xing in 4% paraformaldehyde, the samples were embedded in paraffin and then sliced. HE staining and osteocalcin (OCN) immunohistochemical staining were performed. The inltration of valve cells, collagen ber shape and OCN expression were observed under a light microscope.
Statistical analysis. Data are expressed as mean AE standard deviation. Statistical differences in the measured properties between groups were determined using one-way ANOVA with Student-Newman-Keuls. P values < 0.05 were considered to be statistically signicant.

Characterization of OPG nanoparticles
The OPG nanoparticles were fabricated via emulsion solvent volatilization. 20 As shown in Fig. 2A, the OPG nanoparticle suspension prepared in this study was semi-transparent and pale blue opalescence, with no precipitation observed. Its particle size as measured by the LLS method was 133 nm, the dispersion index (PDI) was 0.131, and the zeta potential was À27.8 mV (Fig. 2B). TEM demonstrated that the prepared nanoparticles were round, regular in size, uniform in distribution, and exhibited no adhesions to each other (Fig. 2C). However, the particle size of the nanoparticles as measured by TEM was smaller than that by laser light scattering, which is due to the shrinkage of the nanoparticles aer drying. 21 Nanoparticles prepared via the emulsifying solvent volatilization method have a lower encapsulation efficiency for water-soluble drugs and higher encapsulation efficiency for hydrophobic drugs. 14 In this study, water-soluble OPG and phospholipids were freeze-dried to form OPG-phospholipid complexes, which encapsulated a layer of hydrophobic phospholipids on the surface of OPG, improving the encapsulation efficiency of nanoparticles and avoiding direct contact between OPG and organic solvents. 22 We calculated the nanoparticle encapsulation efficiency to be as high as 75% using the content of OPG wrapped in nanoparticles and the total OPG content.

In vitro cumulative release of OPG
We used PBS as the release medium to detect the release of OPG for 7 days. From the release curve (Fig. 2D), it can be seen that OPG was released from the nanoparticles at a faster rate, but with no bursts during the early stage of release. The 24 hour OPG cumulative release rate reached 52%, and then the release rate remained at. By day 7, the OPG cumulative release rate reached 89%. This experiment used phospholipid molecules to form a protective "sheath" on the surface of the OPG, slowing the movement of OPG outwards, and preventing it from being released quickly.

Isolation and culture results of BMSCs
The newly inoculated cells were round and suspended in the medium (Fig. 3A). On day 3 of culture, some of the cells began to adhere to the wall and were grown in a long spindle shape (Fig. 3B). On day 5, most of the cells were already adhered to form cell colonies and gradually extended outwards (Fig. 3C). On day 7, the cells were merged by more than 70%, and then the cells were arranged in a certain direction, showing a spiral shape (Fig. 3D). Aer passage, the cells grew vigorously with a uniform morphology and long spindle-shaped adherent growth (Fig. 3E and F). Flow cytometry showed that the BMSCs highly expressed CD29 and CD90, but did not express or showed low expression of CD34 and CD45 (Fig. 3G). Effect of OPG nanoparticles on the proliferation of BMSCs As shown in Fig. 4, aer the addition of nanoparticles to the culture for 12 h, 24 h, and 48 h, the growth of the nanoparticles in the NL-NPs and OPG-NPs groups was not affected by the nanoparticles, showing no statistical difference between the groups (P > 0.05) when compared with the control group (Fig. 4). With time, the absorbance of each group also increased, indicating that the nanoparticles showed no effect on the proliferation of BMSCs and the prepared OPG nanoparticles were nontoxic to BMSCs.

Effects of OPG nanoparticles on osteogenic differentiation of BMSCs
The staining of the calcium nodules is shown in Fig. 5. Aer 21 days, the calcium deposits were stained with alizarin red to orange-red, and the DAG and NL-NPs groups showed a large number of calcium nodules clustered together, while the OPG-NPs group showed scattered calcium nodules. The number of calcium nodules was small, and no calcium nodules were found in the normal control group, indicating that OPG-NPs inhibited the differentiation of BMSCs into osteoblasts to some extent.
The calcium deposition of each group was represented by the extracted calcium ion concentration (Fig. 6) on day 7 of osteoinduction. The calcium concentration in the DAG group was signicantly different from that in the OPG-NPs group (P < 0.05). On day 14, the calcium concentration in the DAG group and the NL-OPG-NPs group increased signicantly, showing a statistically signicant difference compared with the OPG-NPs group (P < 0.001).
Up to day 21, the calcium ion concentration of each group increased continuously, with a signicant increase in the DAG group, NL-OPG-NPs group and OPG-NPs group. However, the calcium ion concentration of the OPG-NPs group was still signicantly lower than that of the DAG group and NL-OPG-NPs group (P < 0.001). There was no signicant difference between the OPG-NPs group and the normal control group from 1 to 14 days (P > 0.05). However, on day 21, there was a signicant difference in calcium ion concentration between the OPG-NPs group and normal control group (P < 0.001), but the OPG-NPs group showed a stronger anti-calcication effect than the DAG group and the NL-OPG-NPs group.

Characterization of decellularized valve modied with OPG nanoparticles
The morphology of the valve tissue of each group was observed by HE and MASSON staining. In the PAV group, a large number of cells was observed, the valve structure was intact, and the bers were in shape and continuous in a certain direction ( Fig. 7A and E, respectively). No cell residue was observed in the DPAV group, and the bers were continuously shaped in a certain direction ( Fig. 7B and F, respectively). No cell residue was observed in the NPs-DPAV group, and the valve ber structure remained intact, wavy, and the layers were clear ( Fig. 7C and G, respectively). The NL-NPs-DPAV group was   observed to be similar to the OPG-NPs-DPAV group ( Fig. 7D and H, respectively), where the structure of the decellularized valve aer nanoparticle modication remained intact.
Scanning electron microscopy demonstrated that the decellularized bers were visible in the decellularized valve, and no cell adhesion was observed (Fig. 8A). However, the decellularized valve modied with OPG nanoparticles exhibited a layer of nanoparticles attached to the surface of the ber, and its arrangement appeared more regular (Fig. 8B). The infrared spectroscopy results showed that most of the stretching vibration peaks of DPAV, OPG-NPs-DPAV and NL-NPs-DPAV were consistent, but OPG-NPs-DPAV and NL-NPs-DPAV showed a peak at 1744.30 cm À1 , which is ascribed to the C]O stretching vibration peak in the nanoparticle composition, but not in DPAV (Fig. 8C). The material on the surface of the composite valve was connected with the nanoparticles, indirectly indicating that the nanoparticles were attached to the decellularized valve.

Anti-calcication of composite valve in vivo
As shown in Fig. 9, at week 2, the valve bers were loosely arranged in each group. There were more inammatory cells inltrated in the supercial layer of the PAV group and DPAV group, and small amounts of inammatory cells inltrated in the deep layer of the DPAV group. A small amount of inammatory cell inltration was observed in the supercial layer of the OPG-NPs-DPAV group and NL-NPs-DPAV group. At week 4, more inammatory cells inltrated in the deep layer of the DPAV group, while in the other groups, small amounts of inammatory cells were observed in the deep layer. At week 8, the inammatory cell inltration was similar to that at week 4, but new capillary formation was observed in each group, with more new capillaries in the DPAV group. During the 8 weeks of implantation, the valves of each group remained in a complete shape and the bers appeared continuous. At week 8, a small amount of degradation occurred in the valves of each group. Osteocalcin (OCN) immuno-histochemical staining (OCN was stained brown) showed low expression of OCN in the DPAV and NL-NPs-DPAV groups at week 2. At week 4, OCN expression in the OPG-NPs-DPAV group was signicantly lower compared to the other groups. At week 8, a small amount of OCN expression was observed in the OPG-NPs-DPAV group, which may due to the complete release of OPG from the nanoparticles, but was still lower than the other groups. This demonstrated that the released OPG from DPAV could inhibit the expression of OCN.
OCN is a late marker of calcication in osteoblastogenesis, which appears in the later stages of skeletal bone formation. 23 Aortic valve calcication is similar to skeletal bone formation and the process is mediated by an osteoblast-like phenotype. 24 OPG could inhibit bone-like synthesis in the valve to prevent valve calcication. OPG strongly suppresses the levels of OCN, a protein that is specic to bone-like tissue, in an approximate proportion to suppress the valve calcication. 25 The slow release of OPG in the composite valve prepared in this study inhibited the expression of OCN to a certain extent, indicating that the composite valve can delay the calcication of the valve.

Discussion
This study aimed to construct a new type of tissue-engineered anti-calcication composite valve to provide a new strategy for the modication of tissue-engineered heart valves. Firstly, we prepared nanoparticles with anti-calcication factor OPG using the emulsion solvent evaporation method. The nanoparticles were uniform in size, with a encapsulation efficiency, slow in vitro release, non-cytotoxic to BMSCs, and delayed osteogenesis of BMSCs during osteogenesis conditions. Then, we used the unsaturated carbon-carbon double bond of the nano-particle PEG-terminal maleimide and the thiolated decellularized valve to form a Michael addition reaction to introduce the OPGloaded nanoparticles into the decellularized valve. The decellularized valve modied by OPG nanoparticles and the bers remained intact. Scanning electron microscopy and Fourier transform infrared spectroscopy showed that the nanoparticles were successfully bound to the decellularized valve. The in vivo study demonstrated that the decellularized valve modied by OPG had a certain anti-calcication ability.
Currently, all of the decellularization methods can damage the valve structure and cause potential loss of surface components, leading to a decrease in its mechanical properties. However, aer biofunctionalizing DPAV with nanoparticles, the ultimate tensile strength and the fracture strength of the hybrid valve had no difference compared with PAV, indicating that the nanoparticle modication improved some of the mechanical properties of DPAV. 26 The biocompatibility of nanomaterials decides their application prospect. In this study, the MAL-PEG-PCL modied PCL nanoparticles were prepared via the emulsion solvent evaporation method. PEG-PCL and PCL composed nanoparticles are nontoxic and degradable, and their degradation products are also nontoxic and environmentally friendly. 27,28 In addition, our in vitro study demonstrated that the OPG-nanoparticles showed no effect on the proliferation of BMSCs. Further, our previous study also demonstrated that the decellularized valve scaffold possesses good biocompatibility. 29 In this study, OPG and phospholipids were rst freeze-dried to form OPG-phospholipid complexes, forming a protective "sheath" surrounded by phospholipid molecules around OPG, and avoiding direct contact between OPG and organic solvents. Li et al. prepared an insulin-phospholipid complex via the same method, which demonstrated good physicochemical stability in an oil solvent, allowing insulin to exert a better biological effect. In addition, the low encapsulation efficiency of hydrophilic drugs has always been one of the problems plaguing researchers. 22 Several studies have shown that the doubleemulsion solvent evaporation method improves the encapsulation efficiency of water-soluble drugs, but allows hydrophilic drugs to directly contact with organic solvents; however, it remained difficult to determine whether the activity of the drug is affected. 30,31 The OPG-phospholipid complex prepared in this study has hydrophobic properties, which not only improved the encapsulation efficiency of hydrophilic drugs, but also protected the biological activity of OPG. In this study, osteogenic induction medium was used to induce BMSCs to differentiate into osteoblasts and simultaneously added to the medium containing OPG nanoparticles. The results showed that OPG released from the nanoparticles inhibited the calcication of BMSCs to a certain extent, but the intervention mechanism of OPG preventing the differentiation of BMSCs into osteoblasts needs further investigation.
PEG and PCL are commonly used carrier materials for the preparation of nanoparticles. Both PEG and PCL are non-toxic and have good biocompatibility, and they are approved by the FDA for human application. This study also conrmed that the nanoparticles prepared with PEG and PCL are non-cytotoxic to BMSCs. In addition, PEG is also a commonly used cross-linker for tissue-engineered heart valves. 32 In this study, the anti-calcication biological factor OPG was introduced into the decellularized valve by using the Michael addition reaction of the unsaturated carbon-carbon double bond of the PEGterminal maleimide of the nanoparticle carrier material and the thiolated decellularized valve, and the preparation was This journal is © The Royal Society of Chemistry 2019 successfully carried out, yielding a composite valve that is resistant to calcication.
In this study, a composite valve with certain anti-calcication ability was prepared via tissue engineering technology and nano drug-loading technology. Subcutaneous implantation in rats showed that the composite valve began to calcify at week 8, probably because the nanoparticles were degraded in the body and completely released OPG. Therefore, the degradation of the nanoparticles and the drug loading of nanoparticles require further exploration and optimization. On the other hand, aer 8 weeks of subcutaneous implantation in rats, the morphological structure of the valve was almost complete, which indicated that the composite valve has good biocompatibility. However, we did not test the blood compatibility and mechanical properties of the composite valve; thus, whether the composite valve can maintain normal function and biological performance in a complex blood ow environment still needs further research.

Conclusions
In this study, a novel composite valve with controlled release OPG was prepared via tissue engineering technology and a nano drug-loading system to introduce anti-calcication biological factor OPG on a decellularized valve. In vitro experiments showed that the OPG nanoparticles delayed the calcication of BMSCs. BMSCs were differentiated into osteoblasts, and subcutaneous implantation experiments in rats indicated that the composite valve has anti-calcication ability to a certain extent. Further optimization and investigation into this treatment may prove benecial for long-term anti-calcication.

Ethical statement
All experiments were performed in compliance with the Institutional Animal Care and Use Committee (IACUC) and all experiments followed institutional guidelines. All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of the Second Affiliated Hospital of Nanchang University. All research protocols were approved by the Institutional Animal Care and Use Committee of the Second Affiliated Hospital of Nanchang University.

Conflicts of interest
There are no conicts to declare.