A novel approach to prepare a tissue engineering decellularized valve scaffold with poly(ethylene glycol)–poly(ε-caprolactone)

Abstract

The objective of this study was to explore the feasibility of preparing a decellularized valve scaffold with methoxy poly(ethylene glycol)–poly(ε-caprolactone) (MPEG–PCL). This is the first report of applying MPEG–PCL to decellularize porcine aortic valves (PAVs). We evaluated its decellularization activity versus two commonly used agents (Triton X-100 and sodium deoxycholate (SD)) in terms of histological morphology, amount of valve-related components, biocompatibility and mechanical properties. The results revealed that 1% MPEG–PCL fully decellularized the valve cells, and the valve fiber structure remained intact. Compared to untreated native valves, the amount of residual DNA in the decellularization groups treated with 1% MPEG–PCL, Triton X-100 and SD were significantly reduced. The water content and collagen content in none of the decellularization groups were significantly different from the native valves. However, the elastin content in the valves of all the decellularization groups was significantly lower than in the native valves. In all the decellularization groups, some degree of platelet adhesion was observed, but the hemolysis rates of all groups were significantly smaller versus the native group. Cytotoxicity testing showed that MPEG–PCL was non-cytotoxic. For every decellularization group, the mechanical properties of the valve scaffolds along circumferential and radial directions were not significantly different from that of the valves. Our study indicates that MPEG–PCL can be used to prepare decellularized PAV. MPEG–PCL is non-cytotoxic and can completely remove cells from the valve while maintaining an intact extracellular matrix ultrastructure. We provide a novel decellularization method for the construction of a tissue engineered heart valve.

---

1. Introduction

Tissue engineered heart valves (TEHVs) are a novel prosthetic valve material. Using tissue-engineering technology, seeded cells are grown on a scaffold so that after implantation to replace the diseased valve, the TEHV still retains the capacity for growth, reconstruction and repair. No postoperative anticoagulant treatment is needed, and the biological as well as mechanical properties of TEHV are comparable to those of native normal valves.^1,2 Current studies on TEHV have mainly focused on the following aspects: preparation and modification of the valve scaffold, selection and cultivation of seeded cells, and interaction between seeded cells and the valve scaffold material. Preparation of valve scaffolds is a prerequisite for the entire TEHV research. Thus, it is essential to prepare high performance valve scaffolds.

The decellularized extracellular matrix (ECM) has been widely used in tissue engineering due to its excellent biocompatibility and ability to provide a matrix microenvironment similar to that found in vivo for seed cell adhesion, proliferation and differentiation. Currently, the most commonly used decellularized TEHV scaffolds are porcine heart valves and bovine pericardium. These have good biological and mechanical properties, good availability, and low costs. Thus, they are used widely in TEHV research. The ideal decellularized scaffold would maintain the integrity of the valve ECM and show excellent biocompatibility and a three-dimensional ultra-fine porous structure to provide sufficient space for seeded cells and tissue growth. It is difficult for synthetic materials to imitate these features.^3,4 Valve decellularization is a crucial step in construction of TEHV using decellularized scaffolds. The goal of any decellularization method is to remove the cell components while keeping the ECM ultrastructure and mechanical properties intact.⁵

Depending on their nature, decellularization methods can generally be divided into physical, chemical, and enzymatic methods. Physical decellularization methods commonly used include rapid freezing and thawing, sonication, etc. The agents typically used in chemical methods include Triton X-100, sodium deoxycholate (SD), sodium lauryl sulfate (SDS), etc. Enzymes commonly used in enzymatic methods include low concentrations of trypsin. In practice, different decellularization methods or different agents can be used in combination to achieve the desired decellularization effect.⁶ The most current and widely applied agents to prepare decellularized biomaterials are surfactants—these are all toxic to a certain degree. Cationic surfactants are most toxic followed by anionic surfactants; nonionic surfactants have minimum toxicity. In terms decellularization capability, nonionic surfactants are the strongest followed by anionic surfactants.⁷

The technology to build TEHV decellularized biomaterial scaffolds is not yet mature, and all current decellularization methods will result in damage to valve structure and a potential loss of surface components. This accelerates the degeneration and calcification of the valve to a certain extent.⁸ In addition, the commonly used decellularization agents all have certain toxicity, and there will be some residue of these agents on the decellularized valves.^9,10 Decellularization requires a relatively long time—in most cases 48 hours or longer.¹¹ Using surfactant as the only decellularization agent for valves or similar tissues is often not sufficient to ensure complete removal of the cells. Also, it cannot maintain the integrity of the ECM or the mechanical properties of these tissues. Therefore, improved decellularization agents and methods are critical for constructing decellularized TEHV scaffolds.

Polyethylene glycol (PEG) has a basic structure of repeated glycols. It is highly hydrophilic, non-toxic, non-antigenic and non-immunogenic. Poly ε-caprolactone (PCL) is a liposoluble polymer with excellent drug permeability, biodegradability and biocompatibility. PEG and PCL play important roles in drug delivery systems, tissue engineering and other fields.^12–15 The U.S. Food and Drug Administration (FDA) has approved both PEG and PCL for use in man. PEG–PCL is an amphiphilic block copolymer. It has the structural features of a surfactant and is a non-toxic, non-ionic surfactant.

In this study, we show for the first time that methoxy poly(ethylene glycol)–poly(ε-caprolactone) (MPEG–PCL) decellularizes porcine aortic valves (PAV). We then compared its decellularization activity versus two other commonly used decellularization agents (Triton X-100 and SD) with respect to histology, content of valve-related components, biocompatibility and mechanical properties. The goal was to provide a novel decellularization method for TEHV construction.

2. Experimental

2.1 Materials and chemicals

Methoxy polyethylene glycol (MPEG, M_n 2000 D), ε-caprolactone, stannous octoate, tris(hydroxymethyl)aminomethane (TRIS), DNase I, and RNase A were purchased from Sigma-Aldrich (U.S.). An animal tissue genomic DNA kit was purchased from Beijing Zoman Biotechnology Co., Ltd. (China). A collagen enzyme-linked immunosorbent assay (ELISA) kit and elastin ELISA kit were purchased from RapidBio (U.S.). A hematoxylin–eosin (H&E) staining kit was purchased from Wuhan Boster Biological Technology. Masson's trichrome stain kit was purchased from Genmed Scientifics, Inc. (U.S.). Roswell Park Memorial Institute (RPMI) medium 1640 was purchased from Hyclone Corp. (U.S.). Fetal bovine serum was purchased from Biological Industries Israel Beit-Haemek Ltd. (Israel). TransDetect™ Cell Counting Kit and phosphate-buffered saline (PBS) were purchased from Beijing TransGen Biotech Co., Ltd. (China). Toluene, dichloromethane, ethyl ether and other reagents were purchase from Xilong Chemical Co., Ltd. (China). Human umbilical vein endothelial cells (HUVECs, CRL-1730™) were purchased from American Type Culture Collection (ATCC).

2.2 Synthesis and structural characterization of MPEG–PCL

Ring-opening polymerization was applied, and MPEG and ε-CL were polymerized to synthesize the MPEG–PCL copolymer. After drying, 2 g MPEG and 1.067 ml ε-CL were placed in a dry flask. Stannous octoate served as the catalyst, and toluene was the reaction solvent. The reaction took place under a nitrogen atmosphere. After magnetic stirring in a 90 °C oil bath for 24 h, the crude MPEG–PCL products were obtained. The remaining toluene was removed via rotary evaporation under reduced pressure. The products were cooled to room temperature and dichloromethane was added to fully dissolve the reaction products—these were then precipitated with ethyl ether and subjected to filtration under reduced pressure to obtain white precipitates. The precipitates were again dissolved in dichloromethane, precipitated with ethyl ether and subjected to filtration. This was repeated three times. The final product was sealed and preserved at −20 °C. Fourier transform infrared spectroscopy (FTIR, Nicolet 5700, Thermo Fisher Nicolet) and proton nuclear magnetic resonance (¹H-NMR, Bruker, Switzerland) spectroscopy were used to characterize the product.

2.3 Collection of PAV

Porcine hearts were freshly collected from a local slaughterhouse. The heart was rinsed with 4 °C saline to remove blood and expose the aortic root. Nearby heart muscles, tendons and other tissues were resected to remove the aortic root containing the leaflet—this was rinsed repeatedly with 4 °C saline before transport to the laboratory. In the laboratory, the aortic valve leaflet was cut in an ultraclean bench environment and was rinsed repeatedly with 4 °C saline.

2.4 Determination of the optimal MPEG–PCL concentration

To obtain the optimal MPEG–PCL for decellularization, PAVs were treated with 10 mM TRIS buffer (including 0.02% (w/v) EDTA and 1 mM PMSF) containing 0.5%, 1%, 2%, or 5% (w/v) MPEG–PCL on an incubator shaker (SHA-BA, Changzhou Long More Instruments Manufacturing Co. Ltd., China) at 37 °C and 75 rpm with continuous oscillation for 24 h. The PAVs were then treated with PBS containing nucleases (0.2 mg ml⁻¹ DNase I and 20 μg ml⁻¹ RNase A) on an incubator shaker at 37 °C and 75 rpm for 2 h. Finally, the decellularized valves were placed in PBS for 12 h. The prepared valves in each decellularized group was subjected to H&E, Masson's trichrome staining, and optical microscopy to observe cell removal and structural integrity of the valve to obtain the optimal MPEG–PCL concentration for decellularization.

2.5 Treatment steps of different groups

The PAVs were randomly divided into four groups. Decellularization of different groups was performed as follows. The PAV in Group A (MPEG–PCL group) was treated with 10 mM TRIS buffer containing 1% (w/v) MPEG–PCL. The PAV in Group B (Triton X-100 group) was treated with PBS containing 1% (v/v) Triton X-100. The PAV in Group C (SD group) was treated with 1% (w/v) SD water solution. For all three groups, the decellularization solution included 0.02% (w/v) EDTA and 1 mM PMSF. The PAVs and the corresponding decellularization solution were subjected to continuous oscillation at 37 °C, 75 rpm on an incubator shaker for 24 h. This was then treated with PBS containing nucleases (0.2 mg ml⁻¹ DNase I and 20 μg ml⁻¹ RNase A) and placed on an incubator shaker at 37 °C and 75 rpm for 2 h. Finally, the decellularized PAVs were placed in PBS for cleaning for 12 h. In Group D (native, untreated fresh valve), fresh PAVs were placed in PBS. This was subjected to continuous oscillation at 37 °C and 75 rpm on an incubator shaker for 24 h before being cleaned in fresh PBS for 12 h.

2.6 H&E and Masson's trichrome staining

For all groups, the valve was fixed in 4% paraformaldehyde and embedded in paraffin for sectioning. H&E staining and Masson's trichrome stain were performed following kit instructions. Under light microscopy, valve sections in each group were observed to examine if there were residual cells or cell fragments and whether the valve structure was intact and whether the fiber alignment was disordered.

2.7 Scanning electron microscopy (SEM)

The valves were fixed in 2.5% glutaraldehyde at 4 °C, and then subjected to dehydration with conventional gradient ethanol, CO₂ critical point drying, and ion sputtering for gold coating. The prepared samples in different groups were then observed using SEM (Quanta200F, FEI Company, U.S.) to observe the valve morphology and structural integrity.

2.8 Measurement of valve-related components

2.8.1 DNA content. The extraction of genomic DNA in all groups was performed in accordance with the instructions provided with the genomic DNA kit. Briefly, valve tissue was weighed, crushed, and lysed in a 56 °C water bath on a shaker for 3 h. Next, DNA was extracted using manufacturer's protocol. The DNA concentration was measured using a biological nucleic acid protein analyzer (BioPhotometer Plus, Eppendrof, Germany).

2.8.2 Water content. In each group, valve pieces (six pieces from each group) were dried as much as possible using filter paper, and accurately weighed (weight denoted as m₀) before placed in a lyophilized tube. After subjected to prefreezing in a freeze dryer (FD-1A-50, Beijing Boyikang Laboratory Instruments Co., Ltd., China) at −56 °C for 3 h, the samples were dried under vacuum for 20 h before accurately weighed again (weight denoted as m₁). The water content of the valve in each group was calculated.

Water content of each group (%) = (m₀ − m₁)/m₀ × 100%

2.8.3 Collagen and elastin contents. The valve collagen content was measured in accordance with the instructions in a porcine collagen ELISA kit. Briefly, valve tissue was weighed and immediately frozen with liquid nitrogen. After thawing on ice, PBS was added for full homogenization, and the sample was centrifuged at 4 °C and 2500 rpm for 20 min. The supernatant was collected and the collagen content was measured following the kit instructions. The valve elastin content was measured with the porcine elastin ELISA kit similar to the collagen measurements.

2.9 Platelet adhesion test

An appropriate amount of fresh blood was collected from healthy rabbits, and 3.8% sodium citrate was used as the anticoagulant. The final volume ratio of blood to sodium citrate solution was 9 : 1. The collected blood was then subjected to centrifugation at 8 °C and 700 rpm for 10 min to obtain platelet rich plasma (PRP). This was diluted with PBS at a ratio of 1 : 1. For each group, 0.1 ml diluted PRP was dropped on a 1 cm × 1 cm piece of decellularized valve. This was incubated at 37 °C for 1 h. Next each valve piece was rinsed with PBS three times to wash away non-adhered or weakly adhered platelets. Finally, the valve piece was fixed at 4 °C in 2.5% glutaraldehyde and subjected to dehydration with conventional gradient ethanol with CO₂ critical point drying and ion sputtering for gold coating. The platelet distribution on the decellularized valve in each group was observed by SEM.

2.10 Hemolysis test

Using a method described by Li,¹⁶ a 1 cm × 1 cm piece of valve from each group was placed in a centrifuge tube containing 10 ml PBS at 37 °C for 1 h. Then 0.2 ml diluted rabbit blood (8 ml blood diluted with 10 ml PBS) was added and incubated with the valve piece at 37 °C for another 1 h. Triple-distilled water served as a positive control, and PBS was the negative control. The tube was centrifuged at 1500 rpm for 10 min, and the supernatant was collected for absorbance measurements at 545 nm. The hemolysis rate (HR) was calculated using the following formula:

HR = (AS − AN)/(AP − AN)

where AS, AN, and AP represent the absorbance values of the supernatant in the valve groups, the negative control group, and the positive control group, respectively.

2.11 Biomechanical characterization

The biomechanical properties of the valves in all groups were characterized using the uniaxial tensile loading test. Considering the anisotropy of the valves, indicators of valve-related biomechanical properties were measured along the circumferential and radial directions with a Computer Servo Control Tensile Tester (HD-B609BS, Haida International Instrument Co., Ltd., China, Model: BAB-5MT, CAP: 5 kg). For each group, the valves were cut into a 15 mm × 5 mm strip along the circumferential (Fig. 1A), and a 10 mm × 5 mm strip along the radial direction (Fig. 1B). Both were fixed between the clamps of the tester (Fig. 1D). A piece of native normal valve is shown in Fig. 1C. The samples were fixed and then loaded to failure at a constant speed of 10 mm min⁻¹. The stress–strain curves were produced using the load–displacement curves from uniaxial tensile loading tests. Four parameters were calculated—ultimate tensile strength, fracture strength, fracture strain, and elastic modulus. The length, width, and thickness of these valves were measured using digital calipers. All testing was conducted at room temperature.

Fig. 1 The strips of valve cut along the circumferential and radial directions for each group are shown (A and B), respectively; normal valve is shown in (C); the valve strips are fixed between the clamps of the testers as shown in (D).

2.12 Cytotoxicity assay

The effect of the valves on cell viability was investigated using a Cell Counting Kit (CCK) assay according to the standard protocol stated in ISO 10993-5 for evaluation of in vitro cytotoxicity of medical devices. RPMI 1640 cell culture medium containing 10% FBS was used to culture HUVECs in an incubator (HERACELL 150i, Thermo Fisher Scientific Company, U.S.) at 37 °C and 5% CO₂. To prepare valve conditioned medium from each group, the valve in a 1 cm × 1 cm piece was placed in a test tube containing 10 ml cell culture medium for incubation at 37 °C under sterile conditions for 24 h. The resulting supernatants were then sterilized using 0.22 μm filters for later use. HUVECs in the logarithmic growth phase were prepared into cell suspensions of 4 × 10³ ml⁻¹ and seeded at 100 μl per well into a 96-well plate. For each experimental condition, six replicate wells were used. The plate was placed in a culture incubator. After the cells were adherent, we added 10 μl of the corresponding valve conditioned medium to each well in the experimental group. The control group received 10 μl of cell culture medium. The cells were cultured for another 24 h or 48 h, when 11 μl CCK solution was carefully added to each well, and the plate was again incubated for 2 h. Finally, the absorbance value of each well at 450 nm was measured using a microplate reader (VARIOSKAN, Thermo Fisher Scientific Company, U.S.). The relative growth rate (RGR) of the cells was calculated as follows: RGR (%) = mean absorbance of the experimental group/mean absorbance of the control group. To directly observe cell growth under the microscope, cells cultured for 48 h were imaged via H&E staining before observation.

2.13 Statistical analysis

The PAVs were randomly assigned to different groups. All experiments were repeated in triplicate. Data are expressed as mean ± standard deviation. For comparison between each decellularization group and the native group, we performed a one-way analysis of variance (ANOVA). Any reference to a difference in the results implies statistical significance when P < 0.05.

3. Results and discussion

Decellularized ECM is widely used in tissue engineering because it offers excellent biocompatibility and the ability to provide a matrix microenvironment similar to in vivo conditions for seed cell adhesion, proliferation and differentiation. The effect of decellularization agents depends on the type of tissue, the type of the decellularization agent, and interaction between the two. A number of decellularization agents are used for tissues. Currently, the most widely applied type is a surfactant such as Triton X-100, SD and SDS. Triton X-100 is a non-ionic surfactant, and is the most frequently used decellularization agent. And the decellularization effect of Triton X-100 relies on its concentration and the type of the tissue.^10,17 SD and SDS are ionic surfactants, and are more toxic than Triton X-100. In the decellularization process, SD and SDS will reduce the contents of certain components of the tissue such as elastin and glycosaminoglycan. It can even destroy the structures of the ECM.^18,19

MPEG–PCL is a non-ionic surfactant. Its synthesis is simple, and the product is non-toxic. MPEG–PCL possesses the corresponding functions of surfactants, and can disrupt lipid–lipid and lipid–protein interactions, but leave protein–protein interactions intact.^9,20 Here, we evaluated the effect of MPEG–PCL on PAV decellularization and compared it with that of two commonly used decellularization agents (Triton X-100 and SD) to provide a new method for preparing decellularized TEHV scaffold. We used the MPEG–PCL that we synthesized ourselves as a decellularization agent and assessed its decellularization effects in multiple aspects including histology and morphology, the contents of valve-related components, biocompatibility and mechanical properties.

3.1 Structural characterization of MPEG–PCL

Ring-opening polymerization was used to synthesize MPEG–PCL copolymer. The structure and molecular weight of which were then characterized with FTIR and ¹H-NMR. The synthetic copolymer was a white powder after drying, and the synthetic yield was about 90%. The chemical structure of MPEG–PCL was . Fig. 2A shows the infrared spectrum of MPEG–PCL. The peak at 3443.05 cm⁻¹ is the stretching vibration peak of –OH at the end of the PCL block: 1728.66 cm⁻¹ is the stretching vibration peak of C=O in PCL block; 1109.81 cm⁻¹ is the stretching vibration peak of C–O–C in the MPEG block, and 2800–3000 cm⁻¹ is the stretching vibration peak of C–H stretching vibration of methylene group. These peaks demonstrate that the synthesized copolymer consists of a MPEG chain block and PCL chain block. This was consistent with the results of other studies.^21,22

Fig. 2 The FTIR spectrum (A) and ¹H NMR spectrum (B) of MPEG–PCL copolymer.

Fig. 2B shows the ¹H-NMR spectrum of the copolymer MPEG–PCL. The following peaks can be observed: the methylene proton peak in the PCL block (δ = 1.38, 1.65, 2.31 and 4.06 ppm), the methylene proton peak in the MPEG block (primarily δ = 3.64 ppm), and a relatively weak peak at 4.23 ppm. This is related to –OCH₂CH₂O– at the junction between MPEG and PCL. These data suggest that the synthesized product is a MPEG–PCL copolymer, which is consistent with the results reported in other studies.^21,22 The mean molecular weight of the PCL block can be estimated by the δ 2.31 ppm proton peak in the PCL block and the δ 3.64 ppm proton peak in the MPEG block.²³ In this way, the mean molecular weight of the synthesized MPEG–PCL copolymer was estimated to be 4000 D.

3.2 Determination of optimal MPEG–PCL concentration

MPEG–PCL is an amphiphilic block copolymer. It belongs to non-ionic agents and possesses the structural characteristics of surfactants. As a decellularization agent, MPEG–PCL can disrupt lipid–lipid and lipid–protein interactions. However, at different concentrations, the decellularization effects of MPEG–PCL vary widely. In this study, PAVs were treated with MPEG–PCL from four different concentrations for decellularization treatment. After H&E staining and Masson's trichrome stain, the cell residues and structural integrity of valves in different concentration groups were observed and compared to determine the optimal MPEG–PCL concentration for decellularization. The H&E staining and Masson's trichrome stain results on different concentration groups are shown in Fig. 3.

Fig. 3 H&E staining results (A–D) and Masson's trichrome stain results (E–H) on decellularized valves treated with MPEG–PCL of different concentrations. A–D (E–H) represent the 0.5%, 1%, 2% and 5% concentration groups, respectively.

In the 0.5% MPEG–PCL group, valve decellularization was not complete; there were outlines of the cells, but the structure of the valve remained intact. In the 1% MPEG–PCL group, valve decellularization was complete. There were no cell outlines or cell debris. The valve structure was intact, and the fibers took a wavy shape showing clear layers. In the 2% MPEG–PCL group, the valve decellularization was complete. The valve structure was relatively intact. The fibers were wavy but did show some fracture. In the 5% MPEG–PCL group, the valve structure was incomplete although no cell residues were observed. The majority of the fibers were broken showing disordered shapes. By comparison, it can be seen that the overall decellularization effect of MPEG–PCL is optimal when its concentration is 1%—not only is the decellularization complete, but the ultrastructure of the valve remains intact.

For a certain agent, the decellularization effect does not just depend on its concentration. The type of biological tissues to be decellularized and the decellularization time are also rather important.²⁴ Here, we only considered the concentration and determined the optimal decellularization concentration of the MPEG–PCL. The MPEG–PCL was a two-block copolymer. If the molecular weight of each block changed, the decellularization effect of the copolymer at the same concentration might be different.

3.3 Histological evaluation of each experimental group

Based on the above result, 1% MPEG–PCL was used as a decellularization agent. Its decellularization effect was assessed from the perspectives of histology, the contents of valve-related components, biocompatibility and mechanical properties. Such an effect was compared with that of commonly used decellularization agents including Triton X-100 and SD. This was in order to assess the potential of MPEG–PCL as a novel decellularization agent.

The decellularization effects regarding histological morphology of different valve groups were examined primarily via H&E staining and Masson's trichrome stain. In each group we determined whether there were residual cells or cell debris, whether the valve structure was intact, and whether the shape of the fibers was disordered. The H&E staining and Masson's trichrome stain results are shown in Fig. 4. Native normal valve tissues contained large numbers of cells, and the valve structure was intact. The fibers showed a wavy shape and were arranged in a continuous, neat fashion. In the MPEG–PCL group, decellularization was complete and no residual endothelial cells, valvular interstitial cells, or cell debris were observed. The valve fiber structure remained intact and showed a loose and wavy shape with clear layers. In the Triton X-100 group and SD group, the valve structure remained intact. There were no residual cells, but some cell outlines were still visible. The fibers were intact and wavy and lined up in an orderly manner. Thus, H&E staining and Masson's trichrome stain results reveal that the decellularization effect of 1% MPEG–PCL is good. Cell removal is complete while the valve fiber structure remains intact.

Fig. 4 H&E staining results and Masson's trichrome stain results of different experimental groups ((A and E) MPEG–PCL group; (B and F) Triton X-100 group; (C and G) SD group; and (D and H) native valve).

3.4 SEM

The SEM results of different valve groups were similar to histology staining results. Fig. 5 shows the SEM images of valves from different groups. The surface of the normal valve was covered with a layer of cells, and no fibers were exposed. In the MPEG–PCL group after decellularization, no cells or cell debris were observed. The fibers were arranged like wavy strips with no obvious damage. These were arranged in neat rows. In the Triton X-100 group, there were no residual cells, but the fibers appeared rough—some showed fractures. In the SD group, some regions had residual cell debris. The fibers did not show a clear wavy shape and were clustered in some areas.

Fig. 5 SEM images of different experimental groups ((A) MPEG–PCL group; (B) Triton X-100 group; (C) SD group; and (D) native valve).

3.5 Measurement of valve-related components

The native valve has excellent biological and mechanical properties, which depend on its composition and structural integrity. As a candidate TEHV scaffold material, the contents of the relevant components in the decellularized valve should be optimal. However, in the actual decellularization process, a loss of some components is inevitable. An excellent decellularization agent should not cause a loss of valve-related components, but rather should protect these constituents.

3.5.1 DNA content. The decellularization effect was further verified using the residue DNA content in each decellularized valve group. Lower amounts of residual DNA suggest more complete cell removal. Fig. 6A shows the amounts of residual DNA detected from different groups. Versus the normal valve group, the DNA content of each decellularization group was significantly reduced (P < 0.05). In particular, the amount of residual DNA in the MPEG–PCL group was minimal. Antigenicity is an important indicator for assessing the decellularization effect. This is because certain residual antigens in the decellularized valve such as DNA, α-Gal, and MHC-I are important antigens that can cause a variety of rejection reactions.^25–27 Hence, before the bioprosthetic valve is implanted in the human body, its immunogenicity must be minimized.²⁸ Currently, there are no universally accepted international standards. Crapo et al.²⁹ proposed that dsDNA per mg ECM dry weight < 50 ng or DNA fragment length < 200 bp could be used as a minimum safety standard for DNA residue. In the current study, the amount of residual DNA in the MPEG–PCL decellularization group was 4.99 ± 0.11 ng mg⁻¹ (wet weight) in line with the above minimum safety standard.

Fig. 6 DNA contents (A) and water contents (B) of different experimental groups; *significant differences (P < 0.05).

3.5.2 Water content. Water makes up most of the weight of fresh valves. Different decellularization methods may lead to different water contents in the decellularized valves. Damage to collagen fiber structure and loss of elastin may result in increased water content.³⁰ Some decellularization agents may cause collagen fibers to be cross-linked and tightly packed leading to a reduction in the water content. Fig. 6B compares the valve water contents between different groups. Versus the native valve, none of the decellularized valve groups showed significantly different water contents (P > 0.05).

3.5.3 Collagen and elastin contents. Collagen and elastin are important components of valve ECM—changes in their contents directly affect the biological and mechanical properties of the valve.³¹ The content, diameter and cross-linking degree of collagen vary between different species and different ages.³² Fig. 7A shows the collagen contents of different valve groups. Versus the normal valve, none of the decellularized valve groups show significantly different collagen contents (P > 0.05).

Fig. 7 Collagen contents (A) and elastin contents (B) of different experimental groups; *significant differences (P < 0.05).

The presence of elastin enhances the elastic properties of the valve. Fig. 7B shows the elastin contents of different valve groups. Versus the normal valve group, the elastin contents of all decellularized valve groups were decreased significantly (P < 0.05). This suggests that after treatment with the three decellularization agents, valves will lose elastin. Meyer et al.³³ applied Triton X-100 to decellularize rat heart valves and reported decreased elastin content. Triton X-100 may even destroy the ultrastructure of collagen fibers.³⁴ Zhou et al.⁸ used 1% SD for decellularization of bovine pericardium and found that the main components and structure of ECM remained intact. However, Zang et al.¹⁸ applied 4% SD to decellularize murine trachea. The results showed that the elastic modulus and glycosaminoglycan (GAG) content were decreased whereas other mechanical properties were unaffected.

3.6 Platelet adhesion test

Exposure of valve ECM will cause platelet adhesion and aggregation.³⁵ The number of platelets adhered to the valve reflects the anti-platelet adhesion ability of the valve. Fig. 8 shows the SEM results of the platelet adhesion test for different groups. In the native valve group, the platelets were not activated because the fibers were covered, and no platelet adhesion was observed. In contrast, the fibers were exposed on the valve surface because they were not covered by endothelial cells in each decellularization group. Thus, a certain degree of platelet adhesion was observed. Different decellularization groups did not differ notably from each other. Further investigation of the anti-platelet adhesion properties of decellularized valves is needed to reduce platelet adhesion.

Fig. 8 SEM images from the platelet adhesion test for different experimental groups ((A) MPEG–PCL group; (B) Triton X-100 group; (C) SD group; and (D) native valve).

3.7 Hemolysis test

Hemolysis is another indicator for assessing the blood compatibility of biological materials. Fig. 9 shows the hemolysis test results on different valve groups. All decellularization groups differed significantly from the native normal valve group (P < 0.05). There were no significant differences between the decellularization groups, but the HR of the MPEG–PCL was the lowest. This suggests that of the three decellularization agents, MPEG–PCL has the best blood compatibility. According to F 756-00 standard issued by the American Society for Testing and Materials (ASTM),³⁶ a hemolysis degree of 2% is considered nonhemolytic for biomaterials. The valves in all groups fit this criterion, and the HRs of all decellularization groups was lower than that of the native normal valve group. This indicates that in these experiments, valve decellularization does not increase hemolysis. In fact, it even improves the blood compatibility of the valve to some extent.

Fig. 9 Hemolysis rates of different experimental groups; *significant differences (P < 0.05).

3.8 Biomechanical characterization

A good decellularization method should not only completely remove the cells, but also maintain the biomechanical properties of the original valve. The valve ECM is layered, and rich in collagen, elastin and proteoglycans.³⁷ The collagen and elastin play important roles in maintaining the mechanical properties of the ECM.³⁸ In valves, the main valve collagen fibers are type I (74%), III (24%) and V (2%).³⁹ Type I collagen runs along the circumferential direction. Thus, in this direction, the mechanical properties of the valve are superior to that in the radial direction.^40,41 In view of the anisotropy of the valve, we compared the mechanical properties of the valves in each decellularization group with that of native valves in circumferential and radial directions to evaluate the decellularization effect of each group. Fig. 10 shows the ultimate tensile strength, fracture strength, fracture strain and elastic modulus of the valves in different groups along the circumferential and radial direction obtained from the uniaxial tensile loading test.

Fig. 10 The ultimate tensile strength (A), fracture strength (B), fracture strain (C), and elastic modulus (D) of valves in different experimental groups along the circumferential and radial directions. *Significant differences (P < 0.05).

The ultimate tensile strength did not differ significantly in either circumferential or radial directions between the different groups (Fig. 10A) (P > 0.05). This suggests that all the decellularization methods have little effect on the valve ultimate tensile strength. It further indicates that none of the decellularization agents affect the collagen structure notably. This is similar to the results on collagen content measurement. Fig. 10B shows the fracture strengths of the valves in different groups along the circumferential and radial directions. In the circumferential direction, the SD group and the normal group showed a statistically significant difference (P < 0.05), whereas the MPEG–PCL group and Triton X-100 group did not (P > 0.05). In the radial direction, there was no significant difference among all groups (P > 0.05). As illustrated in Fig. 10C, the fracture strain did not differ significantly among all groups (P > 0.05). The elastic modulus is an important parameter for assessing the ability of the material to resist elastic deformation. As shown in Fig. 10D, the elastic modulus did not show a significant difference in either the circumferential or the radial direction among the different groups (P > 0.05). However, the elastic modulus of all decellularization groups was lower than that of the normal valve group in the circumferential direction. This indicates that all decellularization agents have some influence on the elastic modulus of the valve.

Different decellularization agents have different impacts on the mechanical properties of the valve. However, given the type and alignment direction of collagen in the valves, the mechanical properties in the circumferential direction are superior to that in the radial direction. Thus, a number of studies have obtained similar results.^42–44 Yang et al.¹⁹ reported that use of Triton X-100 and SD could deteriorate the mechanical properties of bovine pericardium and destroy the ECM structure. Zou et al.⁴⁵ used 1% Triton X-100 for decellularization of porcine descending thoracic aortas. The results showed that the decellularized ECM retained the original critical mechanical and structural properties. The MPEG–PCL used here is a non-ionic surfactant. We showed that it fully removes cells from the valve, while maintaining the mechanical properties of normal valves. As shown by the results on the above four indicators, the mechanical properties of the valves treated with MPEG–PCL are superior to that treated with the other two decellularization agents.

3.9 Cytotoxicity assay

In theory, the ECM can provide a microenvironment for seed cell growth and is not toxic. However, various decellularized scaffolds may have some toxicity due to residual decellularization agents.²⁹ When implanted, such scaffolds will exert toxic effects on host cells, and may induce hypersensitivity.⁴⁶ Therefore, the residual decellularization agents should be removed as much as possible.

The cytotoxicity of the extracts from each valve group can be reflected by the degree of cell proliferation—this can be measured via the CCK method. Fig. 11 shows the RGRs of HUVECs at 24 h and 48 h in different groups. The RGR of the MPEG–PCL group was comparable to that of the normal group (P > 0.05), whereas the RGRs of both the Triton X-100 group and SD group were lower than that of the normal group (P < 0.05). The SD group exhibited the lowest RGR. Fig. 12 illustrates the H&E staining results on HUVECs cultured for 48 h in the valve conditioned medium of different groups. In all groups, the HUVECs showed normal morphology yet the cell density differed to a certain extent among the groups. The cell density of the MPEG–PCL group was comparable to that of the normal whereas the cell densities of the Triton X-100 group and the SD group were slightly lower. The H&E staining results were consistent with the RGR results.

Fig. 11 The relative growth rates of HUVECs in different experimental groups at 24 h and 48 h; *significant differences (P < 0.05).

Fig. 12 H&E staining results on HUVECs cultured for 48 h in valve conditioned medium from different groups ((A) MPEG–PCL group; (B) Triton X-100 group; (C) SD group; and (D) native valve).

The cytotoxicity assay results largely reflect the toxicity of the residual decellularization agent in each group. The MPEG–PCL and Triton X-100 both belong to nonionic surfactants, but MPEG–PCL has a better biocompatibility than Triton X-100. Its toxicity is lower or even zero thereby meeting the requirement of TEHV construction. SD belongs to ionic surfactants, and its toxicity is greater than nonionic surfactants. Cebotari et al.¹⁰ applied 1% SD and 1% SDS to decellularize porcine pulmonary valves and used human endothelial cells to measure the toxicity of the rinse fluid. The results showed that the first batch of rinse fluid was highly toxic to human endothelial cells, yet as the number of rinses increased, the toxicity decreased. Residual decellularization agents will affect the adhesion and growth of seeded cells. To increase the number of rinse or use a non-toxic decellularization agent makes the decellularized scaffold more conducive to seed cell growth.

This study was the first to apply MPEG–PCL—a non-toxic non-ionic surfactant as a decellularization agent to decellularize PAV. We successfully prepared biomimetic material for aortic valve scaffold with excellent biological and mechanical properties. This result provides a new method for construction of the TEHV scaffold. Compared to commonly used decellularization agents of the same concentration, the decellularization effect of MPEG–PCL was found to be better. As a non-ionic surfactant, MPEG–PCL can be used for decellularization of valve tissues. MPEG–PCL is a high molecular weight polymer, and as the lengths of the MPEG and PCL blocks change, its decellularization effect may also vary. In this study, we did not explore the decellularization effect of MPEG–PCL of other molecular weights, nor did we examine the relationship between the ratio of MPEG to PCL and the effect of decellularization.

Residual decellularization agents affect the application of decellularized valves. This is especially true for decellularization agents with certain cytotoxicity. Their residual influences the seed cell adhesion, proliferation, and differentiation. Thus, to seek environment-friendly, nontoxic, and biocompatible decellularization agents is one of the future directions in the field of tissue engineering.

In this study, the object of decellularization was PAV. Martin et al.⁴⁷ pointed out that some animal valve models could not fully simulate human valves, and valves from different animals at different ages also present different mechanical properties. Therefore, the decellularization effect of MPEG–PCL on the valves of other species remains to be explored.

Because the surface of the native valve is completely covered by endothelial cells, it will not activate platelets. On the contrary, on the surface of decellularized valves, the fibers are exposed. This may activate platelets in vivo resulting in platelet adhesion and aggregation. This in turn may result in thrombosis.³⁵ The anti-platelet adhesion ability of the decellularized valves prepared in the present study still needs further improvement. Modification of decellularized valves also merits further exploration. For example, it is possible to apply relevant methods or techniques to immobilize certain biologically active factors to the surface of the decellularized valve. Thus, the decellularized valve can possess certain features such as ECM protection, anti platelet adhesion, and an increase in endothelialization of the scaffold in vivo.

This study only examined the decellularization effect of MPEG–PCL in vitro. It did not involve in vivo experiments. Thus, whether the prepared decellularized valves manifest excellent biological and mechanical properties in vivo still needs further investigation.

4. Conclusion

As a non-toxic surfactant, the MPEG–PCL can be used to prepare decellularized PAV. The results of this study demonstrate that the decellularization effect of 1% MPEG–PCL is better than that of Triton X-100 and SD at the same concentration. MPEG–PCL can completely remove cells from the valve, retain the ultrastructure of ECM, and have no cytotoxicity. After decellularization with MPEG–PCL, the valve shows good biocompatibility, and its mechanical properties are comparable to native normal valves. In sum, MPEG–PCL can be used as a novel type of decellularization agent for preparation of decellularized TEHV scaffolds.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This study was funded by a grant from the National High-tech Research and Development Program (863 Program) of China (No. 2014AA020539), the National Natural Science Foundation of China (No. 81260047, 81260054).

References

1. N. Thierfelder, F. Koenig, T. Hollweck, C. Fano, M. Dauner, C. Hagl and B. Akra, J. Tissue Eng. Regener. Med., 2014, 8, 127.
2. Z. H. Syedain, A. R. Bradee, S. Kren, D. A. Taylor and R. T. Tranquillo, Tissue Eng., Part A, 2013, 19, 759–769.
3. S. Jana, B. J. Tefft, D. B. Spoon and R. D. Simari, Acta Biomater., 2014, 10, 2877–2893.
4. R. A. Rippel, H. Ghanbari and A. M. Seifalian, World J. Surg., 2012, 36, 1581–1591.
5. P. Somers, F. De Somer, M. Cornelissen, H. Thierens and G. Van Nooten, Artif. Cells, Blood Substitutes, Biotechnol., 2012, 40, 151–162.
6. A. Azhim, N. Syazwani, Y. Morimoto, K. S. Furukawa and T. Ushida, J. Biomater. Appl., 2014, 29, 130–141.
7. Y. Xue, S. Zhang, M. Tang, T. Zhang, Y. Wang, Y. Hieda and H. Takeshita, J. Appl. Toxicol., 2012, 32, 480–487.
8. Z. Jianye, O. Fritze, M. Schleicher, H. P. Wendel, K. Schenke-Layland, C. Harasztosi, H. Shengshou and U. A. Stock, Biomaterials, 2010, 31, 2549–2554.
9. T. W. Gilbert, T. L. Sellaro and S. F. Badylak, Biomaterials, 2006, 27, 3675–3683.
10. S. Cebotari, I. Tudorache, T. Jaekel, A. Hilfiker, S. Dorfman, W. Ternes, A. Haverich and A. Lichtenberg, Artif. Organs, 2010, 34, 206–209.
11. S. Eichhorn, D. Baier, D. Horst, U. Schreiber, H. Lahm, R. Lange and M. Krane, Int. J. Biomater., 2013, 2013, 693793.
12. M. J. Joralemon, S. McRae and T. Emrick, Chem. Commun., 2010, 46, 1377–1393.
13. T. K. Dash and V. B. Konkimalla, Mol. Pharm., 2012, 9, 2365–2379.
14. J. Zhou, S. Hu, J. Ding, J. Xu, J. Shi and N. Dong, BioMedical Engineering OnLine, 2013, 12, 1–12.
15. T. K. Dash and V. B. Konkimalla, J. Controlled Release, 2012, 158, 15–33.
16. M. Li, K. G. Neoh, L. Q. Xu, R. Wang, E.-T. Kang, T. Lau, D. P. Olszyna and E. Chiong, Langmuir, 2012, 28, 16408–16422.
17. P. Vavken, S. Joshi and M. M. Murray, J. Orthop. Res., 2009, 27, 1612–1618.
18. M. Zang, Q. Zhang, E. I. Chang, A. B. Mathur and P. Yu, Plast. Reconstr. Surg., 2012, 130, 532–540.
19. M. Yang, C.-Z. Chen, X.-N. Wang, Y.-B. Zhu and Y. J. Gu, J. Biomed. Mater. Res., Part B, 2009, 91, 354–361.
20. A. M. Seddon, P. Curnow and P. J. Booth, Biochim. Biophys. Acta, 2004, 1666, 105–117.
21. H. Danafar, S. Davaran, K. Rostamizadeh, H. Valizadeh and M. Hamidi, Adv. Pharm. Bull., 2014, 4, 501–510.
22. S. S. Payyappilly, S. Dhara and S. Chattopadhyay, Soft Matter, 2014, 10, 2150–2159.
23. N. Nasongkla, X. Shuai, H. Ai, B. D. Weinberg, J. Pink, D. A. Boothman and J. M. Gao, Angew. Chem., Int. Ed., 2004, 43, 6323–6327.
24. J. Dong, Y. Li and X. Mo, J. Surg. Res., 2013, 183, 56–67.
25. F. Naso, A. Gandaglia, L. Iop, M. Spina and G. Gerosa, Xenotransplantation, 2012, 19, 215–220.
26. M. L. Wong and L. G. Griffiths, Acta Biomater., 2014, 10, 1806–1816.
27. T. J. Keane, R. Londono, N. J. Turner and S. F. Badylak, Biomaterials, 2012, 33, 1771–1781.
28. D. D. Cissell, J. C. Hu, L. G. Griffiths and K. A. Athanasiou, J. Biomech., 2014, 47, 1987–1996.
29. P. M. Crapo, T. W. Gilbert and S. F. Badylak, Biomaterials, 2011, 32, 3233–3243.
30. M. L. Wong, J. L. Wong, K. A. Athanasiou and L. G. Griffiths, Acta Biomater., 2013, 9, 6492–6501.
31. Q. Chen, A. Bruyneel, C. Carr and J. Czernuszka, 3rd International Conference on Tissue Engineering (Icte2013), 2013, vol. 59, pp. 247–254.
32. V. Kasyanov, R. A. Moreno-Rodriguez, M. Kalejs, I. Ozolanta, P. Stradins, X. Wen, H. Yao and V. Mironov, Connect. Tissue Res., 2013, 54, 394–402.
33. S. R. Meyer, B. Chiu, T. A. Churchill, L. Zhu, J. R. T. Lakey and D. B. Ross, J. Biomed. Mater. Res., Part A, 2006, 79, 254–262.
34. C. R. Deeken, A. K. White, S. L. Bachman, B. J. Ramshaw, D. S. Cleveland, T. S. Loy and S. A. Grant, J. Biomed. Mater. Res., Part B, 2011, 96, 199–206.
35. M. T. Kasimir, G. Weigel, J. Sharma, E. Rieder, G. Seebacher, E. Wolner and P. Simon, Thromb. Haemostasis, 2005, 94, 562–567.
36. M. Bauer, C. Lautenschlaeger, K. Kempe, L. Tauhardt, U. S. Schubert and D. Fischer, Macromol. Biosci., 2012, 12, 986–998.
37. R. B. Hinton and K. E. Yutzey, in Annual Review of Physiology, ed. D. Julius and D. E. Clapham, 2011, vol. 73, pp. 29–46.
38. F. Auricchio, A. Ferrara and S. Morganti, Ann. Solid Struct. Mech., 2012, 4, 1–14.
39. E. Fallahiarezoudar, M. Ahmadipourroudposht, A. Idris and N. M. Yusof, Mater. Sci. Eng., C, 2015, 48, 556–565.
40. L. Jun, E. M. Joyce and M. S. Sacks, Biomaterials, 2008, 29, 1065–1074.
41. C. E. Eckert, R. Fan, B. Mikulis, M. Barron, C. A. Carruthers, V. M. Friebe, N. R. Vyavahare and M. S. Sacks, Acta Biomater., 2013, 9, 4653–4660.
42. S. A. Korossis, C. Booth, H. E. Wilcox, K. G. Watterson, J. N. Kearney, J. Fisher and E. Ingham, J. Heart Valve Dis., 2002, 11, 463–471.
43. Z. H. Syedain, L. A. Meier, J. M. Reimer and R. T. Tranquillo, Ann. Biomed. Eng., 2013, 41, 2645–2654.
44. Z. H. Syedain and R. T. Tranquillo, Biomaterials, 2009, 30, 4078–4084.
45. Y. Zou and Y. Zhang, J. Surg. Res., 2012, 175, 359–368.
46. P. Barkvoll, Den Norske Tannlaegeforenings Tidende, 1989, 99, 82–84.
47. C. Martin and W. Sun, J. Biomed. Mater. Res., Part A, 2012, 100, 1591–1599.

---

Footnote

† These authors contributed equally to this work.

---