Acta Biomaterialia 10 (2014) 4983–4995

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

An anti-inflammatory cell-free collagen/resveratrol scaffold for repairing osteochondral defects in rabbits Wei Wang, Liang Sun, Pengfei Zhang, Junfei Song, Wenguang Liu ⇑ School of Materials Science and Engineering, Tianjin Key Laboratory of Composite and Functional Materials, Tianjin University, Tianjin 300072, PR China

a r t i c l e

i n f o

Article history: Received 12 May 2014 Received in revised form 14 August 2014 Accepted 18 August 2014 Available online 27 August 2014 Keywords: Osteochondral damage Acellular Resveratrol Collagen Anti-inflammation

a b s t r a c t Inflammatory factor overexpression is the major cause of cartilage and osteochondral damage. Resveratrol (Res) is known for its anti-inflammatory, antioxidant and immunmodulatory properties. However, these effects are hampered by its water insolubility and rapid metabolism in vivo. To optimize its therapeutic efficacy in this study, Res was grafted to polyacrylic acid (PAA, 1000 Da) to obtain a macromolecular drug, PAA-Res, which was then incorporated into atelocollagen (Coll) hydrogels to fabricate anti-inflammatory cell-free (Coll/Res) scaffolds with improved mechanical strengths. The Coll/Res scaffolds demonstrated the ability to capture diphenylpicrylhydrazyl free radicals. Both pure Coll and Coll/ Res scaffolds could maintain their original shape for 6 weeks in phosphate buffered saline. The scaffolds were degraded by collagenase over several days, and the degradation rate was slowed down by Res loading. The Coll and Coll/Res scaffolds with excellent cytocompatibility were shown to promote the proliferation and maintain the normal phenotype of the seeded chondrocytes and bone marrow stromal stem cells (BMSCs). In addition, the Coll/Res scaffold exhibited the capacity to protect the chondrocytes and BMSCs against reactive oxygen species. The acellular Coll/Res scaffolds were transplanted into the rabbit osteochondral defects. After implantation for 2, 4 and 6 weeks, the samples were retrieved for quantitative real-time polymerase chain reaction, and the inflammatory related genes interleukin-1b, matrix metalloproteinases-13, COX-2 and bone and cartilage related genes SOX-9, aggrecan, Coll II and Coll I were determined. Compared with the untreated defects, the inflammatory related genes were downregulated and those bone and cartilage related genes were up-regulated by filling the defect with an anti-inflammatory scaffold. After 12 weeks, the osteochondral defects were completely repaired by the Coll/Res scaffold, and the neo-cartilage integrated well with its surrounding tissue and subchondral bone. Immunohistochemical and glycosaminoglycan staining confirmed the distribution of Coll II and glycosaminoglycans in the regenerated cartilage. The anti-inflammatory acellular Coll/Res scaffolds are convenient to administer in vivo, holding a greater potential for future clinical applications. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Osteoarthritis (OA) involving full-thickness damage to the hyaline cartilage and underlying bone, is encountered frequently in the clinical setting. The therapy of OA is seriously hampered by the limited self-repair potential of hyaline cartilage due to the absence of vasculature and lymphatics [1–4]. Surgical treatments, such as microfracture, osteochondral grafting and autologous chondrocyte implantation (ACI), are employed to restore the cartilage and repair osteochondral damage. All of these treatments have shown success in relieving pain, but suffer from severe limitations

⇑ Corresponding author. E-mail address: [email protected] (W. Liu). http://dx.doi.org/10.1016/j.actbio.2014.08.022 1742-7061/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

[2–7]. Recent efforts have focused on exploring tissue engineering and regenerative medicine to repair osteochondral damage. Bioabsorbable tissue engineering scaffolds have been used to repair osteochondral lesions. The optimal scaffold for articular cartilage repair should be biocompatible, sufficiently robust in initial strength, degradable and absorbable without being toxic and causing inflammation [3,8,9]. Naturally derived materials such as collagen, alginate, agarose, fibrin, hyaluronic acid (HA), gelatin, chitosan and chondroitin sulfate have been shown to be preferable materials for cartilage tissue engineering scaffolds because of their ability to elicit biological repair actions and stimulate an immune-mediated response. Collagen, the most important constituent of cartilage and bone, has already been used in the ACI procedure for years [10]. However, for clinical applications, safety and non-immunogenicity of the scaffold is essential. Atelocollagen as a scaffold material

4984

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

is considered to be safer than telocollagen because atelocollagen is devoid of terminal telopeptides (the antigenic determinants on the peptide chain of collagen type I). In a previous study, an atelocollagen hydrogel film was successfully used to regenerate a rabbit cornea without problems of immunogenicity or inflammation [11]. On the other hand, compared with telocollagen, a much higher concentration can be acquired by atelocollagen, allowing much easier fabrication of high-strength hydrogel for soft tissue engineering [11]. Pro-inflammatory cytokines, such as cytokine interleukin-1b (IL-1b) and tumor necrosis factor-a (TNF-a) mediate the catabolic degradation of extracellular matrix (ECM) in articular cartilage and play a key role in rheumatoid arthritis (RA) and OA pathogenesis. These cytokines induce chondrocytes and synoviocytes to synthesize proteolytic enzymes, such as matrix metalloproteinases (MMP) and inflammatory mediators. Overexpression of MMPs is implicated in cartilage loss. Additionally, these pro-inflammatory cytokines induce chondrocyte apoptosis [12–14]. Thus, it is critical to inhibit inflammation in the treatment of RA and OA. The clinical results showed that the inhibition of IL-1b could reduce pain and swelling of patients with RA [15]. Recently, resveratrol (trans3,5,40 -trihydroxystilbene, Res), a natural polyphenolic compound, has been shown to have anti-inflammatory, antioxidant, antiaging, anticarcinogenic and immunmodulatory properties [16]. Res has demonstrated anti-inflammatory effects through inhibition of TNF-a, IL-1b and COX-2 by blocking NF-jB activation [17–19]; it can inhibit the IL-1b-induced degradation of mitochondria and apoptosis in chondrocytes and bone marrow stromal stem cell (BMSC)-derived chondrocytes [20,21]. Res can also promote osteogenic differentiation of BMSCs and adipose-derived stem cells (ADSCs), enhance DNA synthesis and alkaline phosphatase (ALP) activity in osteoblasts and prevent femoral bone loss in rats [22,23]. After intra-articular injection, Res can significantly decrease the expression of inflammatory factors and prevent the degradation of glycosaminoglycans (GAGs), and relieve the symptoms of arthritis remarkably [13,24]. Despite these extensive biological effects, Res is known to be quickly metabolized in the human body, resulting in an extremely short plasma half-life. The results of pharmacokinetic tests showed that the half-life of Res in the rabbit and rat blood was 14.4 and 10.3 min, respectively. Moreover, the high hydrophobicity of Res and its ability to be oxidized easily may cause serious problems during use [18,22]. Therefore, preventing the rapid metabolism of Res is essential for its application. Sheu et al. fabricated oxidized hyaluronic acid (OxiHA) hydrogel with chemically bonded Res, which was shown to protect the chondrocytes against lipopolysaccharide (LPS) [16]. Li et al. used Res to modify the surface of porous polycaprolactone (PCL). The Res-PCL led to a significant increase in osteogenesis [22]. Recently, many attempts to utilize the anti-inflammatory property of Res have been reported, but most of these research projects were carried out in vitro only [25–27]. To the best of our knowledge, there has been no report on a Res-incorporated scaffold to restore osteochondral defects in vivo. In this study, Res was grafted to polyacrylic acid (PAA, 1000 Da) to achieve a macromolecular drug (PAA-Res) with much improved water solubility compared to Res alone (Fig. 1A). Then PAA-Res was incorporated into atelocollagen to construct an anti-inflammatory scaffold (Coll/Res) (Fig. 1B). The degradation property of this scaffold and the release behavior of Res were characterized in vitro. The ability of Res to capture diphenylpicrylhydrazyl (DPPH) free radicals and the protective effect of Coll/Res scaffold on the chondrocytes and BMSCs in the reactive oxygen species (ROS) environment caused by H2O2 were assessed. Finally, Coll/Res scaffolds were transplanted into osteochondral defects in rabbits (Fig. 1C). The inflammatory and chondrogenic gene expressions were evaluated by a quantitative real-time polymerase chain reaction (qRT-PCR) and the neo-tissue in the defect zone was examined

macroscopically, histologically and immunohistologically after 12 weeks. 2. Materials and methods 2.1. Materials Type I porcine atelocollagen was supplied by Nippon Meat Packers Inc. (Tokyo, Japan). Res was obtained by the Tianjin Jianfeng Natural Product R&D Co. Ltd (Tianjin, China). N,N0 -Carbonyldiimidazole (CDI) and DPPH were obtained from the Shanghai Pepmed Co. Ltd (Shanghai, China). 1-Ethyl-3-(3-dimethyl aminopropyl) carbodiimide (EDC), 3-(4,5-dimethyl-2-thiazoyl)2,5-diphenyl tetrazolium bromide (MTT, 98%) and N-hydroxysuccinimide (NHS) were supplied by Fluka (Buchs, Switzerland). Collagenase type I and type II were purchased from Sigma-Aldrich. All the other reagents were of analytical grade and used without further purification. 2.2. Preparation and characterization of PAA-Res Freeze-dried PAA was dissolved in dimethyl sulfoxide (DMSO), followed by the addition of CDI (10% w/v, in DMSO) with a CDI:COOH molar ratio = 3:10. This reaction was kept at 60 °C for 3 h. After cooling to room temperature, a 3-fold excess Res (molar ratio) was added to the resultant solution, which was kept in the dark at room temperature for 12 h. The product was dialyzed against pure water through a dialysis membrane (molecular weight cutoff = 1000) in the dark for 3 days. Finally, the solution was lyophilized to collect dry PAA-Res. The Res, PAA and PAARes were characterized by proton nuclear magnetic resonance (1H-NMR, Varian INOVA 500MHZ). The formula below was used to calculate the Res grafting ratio:

Res% ¼

7:5 X I1 6

9

,

2:8 X I2 1

3

I1 is the sum of the integral area of those peaks in 6–7.5 which are attributed to a-CH of the phenolic hydroxyl groups in Res, and I2 is the sum of the integral area of those peaks in 1–2.8 which are attributed to ACH2ACHA in PAA. The water solubility of PAA-Res was evaluated by the absorption at 305 nm [28] on a TU-1810 UV-Vis spectrophotometer (Pgeneral, China). DPPH, a stable and commercially available free radical, has been extensively applied in the study of antioxidant activity of natural compounds. The scavenging effect on DPPH radicals was determined in terms of the method described in the work of others [26,29]. Briefly, DPPH was dissolved in methanol to make a 25 lM stock solution. Then, 98 ll of DPPH solution was mixed with Res solution or PAA-Res solution. A control sample was prepared by mixing 98 ll of DPPH solution with 2 ll of Tris buffer. All the samples were stored at room temperature for 30 min in the dark. After that, the absorbance was measured at 517 nm by a spectrophotometer. The free radical scavenging effect was calculated according to the following formula:

Scavenging effect ð%Þ ¼

ðADPPH  Asample Þ  100 ADPPH

ADPPH and Asample are the absorbance of DPPH and the PAA-Res samples at 517 nm, respectively. 2.3. Preparation and characterization of Coll/Res scaffold The Coll scaffold was fabricated according to a previously reported method [11]. Briefly, 0.3 g of Coll solution (13.7% w/w)

4985

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

Fig. 1. Schematic illustration depicting the procedures of scaffold construction and implantation: (A) synthesis of PAA-Res, (B) fabrication of Coll/PAA-Res, (C) transplantation into the articular defect in rabbits.

was thoroughly mixed with 0.7 ml of water containing no or different concentrations of PAA-Res in a syringe-mixing system under an ice-water bath. The final concentration of Coll and Res was 4% and the final concentration of Res was controlled as 0.1%, 0.3% and 0.5%. This mixture was injected with EDC/NHS solution (12 mg EDC and 6 mg NHS are used to cross-link 1 g collagen (13.7%)). After adjusting the pH to 5.5 (a NaOH solution of 1 mol l1 was employed to adjust the pH), the mixture was cast into glass molds (diameter 4 mm; height 6 mm) or between two glass plates (both length and width = 10 cm, thickness = 0.25 cm) separated by a spacer frame with a thickness of 500 lm. The molds were left at room temperature with 100% humidity for 16 h, and then transferred into an incubator at 37 °C for 5 h. After incubation, the molds were immersed in phosphate buffered saline (PBS) for 30 min, followed by cautious removal of the hydrogels from the molds. The resulting hydrogel cylinders or sheets were eluted in PBS, and PBS was replaced at 8 h intervals for 3 days and then stored at 4 °C in the dark. The equilibrium water content was measured according to weight and dry mass of the gels. The Fourier transform infrared (FTIR) spectra of the Res, PAA-Res, Coll and Coll/PAA-Res film were measured using a Thermo Scientific Nicolet 380 FTIR Spectrometer (Thermo electron cooperation, USA). The compression strength of the hydrogels was measured using a WDW-05 electromechanical tester (Time Group Inc., China). Crosshead speed was set at 10 mm min1, and the breaking load was used to calculate the compression strength (the average strain was 35%). Columnar samples (4.0 mm in diameter and 4.0 mm in height) were employed, and three specimens were tested for each group. The scavenging effects on DPPH radicals by Coll and Coll/ PAA-Res scaffold were determined by the above method. The degradation of the Coll/PAA-Res was evaluated in PBS and collagenase solution. For PBS, the Coll and Coll/PAA-Res scaffold (4.0 mm in diameter and 6.0 mm in height) were separately immersed in sterile PBS at 37 °C for 6 weeks, and the samples were taken out at different time intervals to measure the mass. For the collagenase solution, the Coll or Coll/PAA-Res scaffolds were placed in 20 ml collagenase (5 U) solution. The percentage residual mass of hydrogels was calculated according to the following equation:

Residual mass ration ð%Þ ¼

Wt  100% W0

where W0 is the initial weight of the hydrogel and Wt is the weight of the hydrogel at each time point. During the process of degradation, 1 ml solution was taken out and then 1 ml fresh collagenase solution was added at different time intervals. The absorbance at 305 nm was measured to quantify the amount of Res released [30]. 2.4. Chondrocytes seeded on the Coll/PAA-Res scaffolds Chondrocytes were isolated from New Zealand rabbits by digesting the pieces of articular cartilage with 0.2% collagenase II and culturing in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS). The cell viability was evaluated by the MTT assay. The chondrocytes were seeded on the scaffolds; 24 h later, MTT (5 mg ml1) solution was added. After being continuously cultured for another 4 h, DMSO was added to dissolve the formazan pigment reduced by viable cells. The mixture was centrifuged at 3000g for 10 min to ensure the complete separation of the pigment. The absorbance at 570 nm was measured on a micro-plate reader (Bio-Rad 550). Cell proliferation was assayed using the same method during the in vitro culture for 7 days. To observe the cell distribution, the cell cultured scaffolds were cultured for 2 days and rinsed with PBS gently. Then, 0.5 mg ml1 fluorescein diacetate (FDA)/PBS solution was added. After incubation for 10 min and rinsing with PBS, the fluorescence was observed using fluorescence microscope (Olympus, CKX41). To generate a ROS environment, the cell-hydrogel constructs after 24 h of cultivation were treated with 300 lM H2O2 for 2 h [31]. After another 24 h, the cell viability was evaluated by MTT. 2.5. BMSCs seeded on the Coll/PAA-Res scaffolds Rabbit BMSCs were isolated by short-term adherence to plastic, according to the reported protocol [32,33]. Briefly, after anesthesia the rabbit tibia was punctured with a 16-gauge needle. 4–5 ml of bone marrow was aspirated through a sterile tube into a 10 ml syringe containing 5000 U of heparin. The bone marrow was centrifuged to remove fat and heparin. The precipitated cells were cultured in a complete medium containing DMEM/10% v/v FBS. After a confluent cell layer had formed, the cells were detached

4986

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

using 0.25% trypsin and routinely passaged. The cell viability, proliferation and staining procedures were carried out with the same methods as those of chondrocytes.

3. Results

2.6. Surgical grafting procedure and tissue retrieval

Fig. 2A and B shows the synthesis process of PAA-Res and the H-NMR results, respectively. In the NMR spectra, the peaks at 9.3 and 9.6 ppm belonged to phenolic hydroxyl groups in Res. The peaks between 6 and 7.5 ppm were assigned to @CAH, which appeared in both PAA-Res and Res. The spectra indicated that PAARes was successfully formed. From the NMR spectra, it can be found that only the 40 phenolic hydroxyl group was reacted with the PAA, while the 3 and 5 phenolic hydroxyl groups remained unreacted, which may be attributed to the different steric effect of the phenolic hydroxyl group. The results showed that 14% COOH was substituted with Res. Actually, the degree of substitution can be maximized to 40% by adjusting the molar ratio of COOH:CDI:Res. However, more Res grafting led to insolubility in water. In this research, a substitution degree of 14% Res was used for the fabrication of scaffolds and subsequent characterization. At this degree of substitution, the solubility of PAA-Res in water was 11.8 g l1, equivalent to 2.23 g l1 soluble Res, which was determined by UV spectroscopy, whereas only 0.03 g l1 pure Res was dissolved in water [28]. Obviously, modification with hydrophilic polymer can dramatically enhance the solubility of Res in water. The scavenging effect on free radicals by Res and PAA-Res was determined by the DPPH method. The free radical scavenging effect of Res and PAA-Res was 90.4% ± 1.1% and 58.4% ± 1.5%, which matched well with the amount of phenolic hydroxyl group in the two compounds. Because one of the three phenolic hydroxyl groups in Res was consumed in the PAA grafting reaction, the free radical scavenging effect of PAA-Res was equivalent to two-thirds that of Res. This result also indicated that the antioxidant activity of Res was well preserved in PAA-Res. The columnar scaffolds (Fig. 3A) changed from transparent to opaque with an increment in Res content. At a higher content, hydrophobic Res cannot be well compatible with collagen. These scaffolds had more than 95% equilibrium water contents. The FTIR spectra (Fig. S.2) suggested that the Res was incorporated into Coll. Compared with the pure Coll scaffold, the peaks at 1147 and 950 cm1 separately attributable to the CAO stretch of phenol and the CAH stretch of the benzene ring in Res [16,25] appear in Coll/PAA-Res. As shown in Fig. 3B, the pure Coll scaffold had a compressive strength of 0.52 MPa, whereas the compressive strength of Coll/ PAA-Res hydrogel increased to 0.59 MPa (0.1% PAA-Res, P = 0.067), 0.71 MPa (0.3% PAA-Res, P = 0.034) and 0.82 MPa (0.5% PAA-Res, P = 0.0077) (the P value is calculated by comparison with the pure Coll scaffold). The enhancement in the strength of Coll/PAA-Res may be due to the denser cross-linking of collagen caused by the introduction of more COOH groups in PAA-Res. Cross-linking more of PAA-Res with the collagen chain could contribute to the higher strength of the Coll/PAA-Res scaffold. Moreover, this could also be verified by measuring the compressive strength of Coll/PAA. The compression strength of Coll/PAA with a similar composition to that of the Coll/PAA-Res (0.3%) was 0.67 ± 0.10 MPa, which is higher than that of Coll (0.52 ± 0.05 MPa). Furthermore, the hydrophobic interaction and the hydrogen bond may be increased with the PAA-Res incorporation into collagen, which also contributes to better mechanical properties of Coll/PAA-Res. Fig. 3C exhibits the free radical capture abilities in different scaffolds. The scavenging effect on DPPH radicals was increased with the content of Res, and that Coll/PAA-Res (0.5%) scaffolds was 4-fold greater than that of pure Coll scaffold. Fig. 4A shows images of Coll and Coll/PAA-Res scaffolds after immersing in PBS at 37 °C for 6 weeks. All the scaffolds remained

Taking into account that a higher content of Res may result in a better anti-inflammatory effect, the Coll/PAA-Res (0.5%) scaffolds were employed as in vivo implants. The Committee on Animal Experimentation of the Chinese Academy of Medical Sciences and Peking Union Medical College Institute of Biomedical Engineering approved all the animal experiments. A total of 24 New Zealand white rabbits weighing 2.0–2.5 kg were obtained at 3 months old and acclimated for 1 week before use. The rabbits were anesthetized by injection of ketamine hydrochloride (3.5 mg kg1 of body weight) into the peritoneal cavity. The rear knee joints were exposed through an incision from the medial edge of the patella. With the knees flexed, a corneal trephine was used to create an osteochondral defect on each articular cartilage (4 mm in diameter and 4 mm in depth, Fig. 1C). These osteochondral defects were randomly filled by the Coll or Coll/PAA-Res scaffold, or left untreated. 14 samples were assigned to each group. There were four legs without defects. After operation, the muscle and the skin were closed with suture. After transplantation, the rabbits were allowed to move freely through their housing. At 2, 4 and 6 weeks, the rabbits were sacrificed by injection of excess ketamine hydrochloride. The tissue in the defects was retrieved by corneal trephine (diameter = 4 mm) for qRT-PCR assay. Then, total RNA was extracted using the RNeasy Mini Kit (Qiagen) and 1 lg of total RNA was used for reverse-transcription into complementary DNA (cDNA) with the Omniscript RT Kit (Qiagen). The primer sequences specific to the target gene and the internal control gene (glyceraldehyde-3-phosphate dehydrogenase (GAPDH)) used for qRT-PCR are listed in Table S.1. The RT-PCR was performed in an ABI 7300 RT-PCR system (Applied Biosystems) with SYBR Green PCR master mix (Applied Biosystems) under the condition of 15 s at 90 °C and 60 s at 60 °C, and the fluorescence intensity was recorded for 40 cycles. The amount of mRNA was normalized by the normal cartilage, which was set as 100%. For histological and immunochemical evaluation, the rabbits were sacrificed at 12 weeks. Samples for histological evaluation were fixed in neutral buffered formalin for 7 days. Each specimen was decalcified with 0.25 mol l1 ethylenediaminetetraacetic acid in PBS, dehydrated in graded alcohol and embedded in paraffin wax and sectioned in cross-sections through the center of the implants (5 lm thick sections). The cross-sections were routinely stained with hematoxylin and eosin (H&E) and periodic acid-schiff (PAS) staining. The sections were also used for the immunohistochemical staining of collagen type II by collagen type II antibody (PV-9000, Poly-HRP anti-Mouse/Rabbit IgG Detection system, Beijing Zhongshan Biotechnology). The samples were immersed in PBS containing 3% H2O2 at 25 °C for 10 min to block nonspecific reactions. Subsequently, the sections were incubated in anticollagen type II antibody working dilution at 4 °C overnight. After washing with PBS three times, the samples were incubated in anti-mouse antibody (DAKO, Carpinteria, CA) working dilution at 37 °C for 20–30 min. Finally, the samples were incubated in 3,30 -diaminobenzidine tetrahydrochloride (DAB) solution. Collagen type II was stained brown. 2.7. Statistical analysis Data were expressed as mean ± standard deviation (SD). Statistical analysis was performed using the two-population Student’s t-test. The significant level was set as P < 0.05.

3.1. Properties of the scaffolds

1

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

4987

Fig. 2. Synthesis (A) and 1H NMR characterization (B) of the water soluble macromolecular drug PAA-Res.

Fig. 3. Properties of the Coll/Res scaffold. (A) Gross morphology of the scaffolds; A1 to A4 denote Coll, Coll/PAA-Res (0.1%), Coll/PAA-Res (0.3%) and Coll/PAA-Res (0.5%) scaffold, respectively. (B) Compressive strength of the Coll and Coll/PAA-Res scaffolds. (C) Ability of the scaffolds to capture DPPH free radicals. All values are mean ± SD of three experiments. * denotes P < 0.05.

intact, indicating a very slow degradation rate in PBS. In fact, the mass losses of Coll/PAA-Res and Coll scaffolds were 10–30% after 6 weeks. It was found that the scaffolds with more Res showed a darker color after immersion in PBS, which was caused by the oxidation of Res. The color of scaffold can be maintained for 6 months without any sign of yellowing if the scaffolds are shielded from light exposure at 4 °C. Fig. 4B shows the degradation of the Coll and Coll/PAA-Res scaffolds in collagenase solution. All of these four scaffolds can be degraded by collagenase in 3 days, and more Res content resulted in a slower degradation rate. The pure Coll scaffold was completely degraded in 26 h, while the complete

degradation times for Coll/Res (0.3%) and Coll/Res (0.5%) were 38 and 62 h, respectively. The inset table shows the time for 50% mass loss of these four scaffolds. The degradation rate of Coll/PAA-Res (0.5%) was 1.2-fold slower than that of pure Coll. Fig. 4C shows Res release behavior in collagenase solution. It can be seen that the scaffold with higher Res contents exhibited slower release rates. All the three scaffolds showed a burst release in the first 12 h, and then slowed down. The complete release times for Coll/ PAA-Res (0.1%), Coll/PAA-Res (0.3%) and Coll/PAA-Res (0.5%) were 26, 50 and 74 h, respectively, which coincided with the degradation behavior. Actually, Res can be released in PBS, but only 30%

4988

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

Fig. 4. Degradation properties of the Coll and Coll/PAA-Res scaffolds and Res releasing behavior. (A) Scaffolds in PBS at 37 °C for 6 weeks. A1 to A4 denote Coll, Coll/PAA-Res (0.1%), Coll/PAA-Res (0.3%) and Coll/PAA-Res (0.5%) scaffold, respectively. (B) Degradation properties of the Coll/PAA-Res scaffolds in the collagenase solution (5 U ml1) at 37 °C. (C) Res releasing behavior.

Res came out even after 6 weeks. The results indicated that the breakdown of collagen backbone by collagenase could facilitate the release of Res. There are two factors that contributed to the degradation behavior. Firstly, the density of cross-linking is enhanced by adding the PAA because its COOH can be cross-linked with collagen chains. Secondly, the hydrophobic interaction and the hydrogen bond may be increased with the PAA-Res incorporation into collagen. Both these two factors hinder the degradation of

the Coll/PAA-Res, so the degradation rate of the Coll/PAA-Res showed a dependence on PAA-Res content. 3.2. Chondrocytes and BMSC culture with the Coll/PAA-Res scaffolds Fig. 5A shows the cytotoxicity of the Coll and Coll/PAA-Res scaffolds. As seen in the figure, more chondrocytes adhered on the film of Coll and Coll/PAA-Res than on TCPS, suggesting excellent

Fig. 5. Chondrocyte behavior on the Coll or Coll/PAA-Res scaffolds. (A) Cell viability after 24 h measured by MTT. (B) Florescent images of the chondrocytes on the Coll/Res scaffolds, stained by FDA. B1–B4 denote Coll, Coll/PAA-Res (0.1%), Coll/PAA-Res (0.3%) and Coll/PAA-Res (0.5%) scaffold, respectively. (C) Proliferation behavior of the chondrocytes on the Coll/PAA-Res scaffolds. (D) Relative cell viability after being treated by 300 lmol l1 H2O2 for 2 h. All values are mean ± SD of three experiments. * denotes P < 0.05.

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

4989

Fig. 6. BMSC behavior on the Coll or Coll/PAA-Res scaffolds. (A) Cell viability after 24 h measured by MTT. (B) Florescent images of the BMSCs on the Coll/PAA-Res scaffolds, stained by FDA. B1–B4 denote Coll, Coll/PAA-Res (0.1%), Coll/PAA-Res (0.3%) and Coll/PAA-Res (0.5%) scaffolds, respectively. (C) Proliferation behavior of the BMSCs on the Coll/ PAA-Res scaffolds. (D) Relative cell viability after being treated by the 300 lmol l1 H2O2 for 2 h. All values are mean ± SD of three experiments. * denotes P < 0.05.

cytocompatibility. The incorporation of Res had no effect on the biocompatibility of Coll. The chondrocytes on the Coll and Coll/ PAA-Res films assumed a spherical shape, indicating the normal phenotype in this system (Fig. 5B). The proliferation behavior of the chondrocytes on the Coll and Coll/PAA-Res film was evaluated in 7 days (Fig. 5C). The chondrocytes on both the Coll and Coll/PAARes films showed a rapid cell proliferation behavior. To estimate the anti-inflammation of Coll/PAA-Res, H2O2 was used to create a ROS environment. As shown in Fig. 5D, the cell viability was increased from 26% to 48% with an increase in the content of Res. An increment of 1.8-fold in cell viability at a ROS environment was achieved by the incorporation of 0.5% Res. Fig. 6 shows the BMSC growth behavior on the Coll and Coll/ PAA-Res scaffolds. The BMSCs appeared to have a spindle shape. The variation trend of cell viability and proliferation of BMSCs with Res was similar to that of chondrocytes. As shown in Fig. 6D, BMSC viability was increased from 40% to 75% with an increase of Res at a ROS environment, i.e. a 1.9-fold increase in cell viability by the incorporation of 0.5% Res. The results of chondrocytes and BMSCs confirmed the anti-inflammation functionality of Res-loaded scaffolds, and this functionality was enhanced with increasing Res. 3.3. Anti-inflammatory functionality in vivo From the results in vitro, it was clear that a higher Res content resulted in a better anti-inflammatory functionality, and hence Coll/PAA-Res (0.5%) scaffolds were selected for transplantation into the osteochondral defects in rabbits. The anti-inflammatory effects

were evaluated by mRNA expression of IL-1b, MMP-13 and COX-2. As shown in Fig. 7A, the mRNA-IL-1b in all these three groups was up-regulated significantly after the operation and decreased with time. By transplanting a Coll and Coll/PAA-Res plug, the amount of mRNA-IL-1b decreased sharply. In the 6 week experiment period, the amount of expressed mRNA-IL-1b decreased in the order: untreated > Coll > Coll/PAA-Res. The mRNA-IL-1b in the untreated defects was 7-fold higher than that in normal cartilage in the first 2 weeks. 4 weeks after operation, the mRNA-IL-1b decreased in the untreated defects, but it was still much higher than that of Coll and Coll/PAA-Res scaffolds. 6 weeks after the operation, the amount of mRNA-IL-1b in these three groups decreased, but still exceeded the normal cartilage. The mRNA-MMP-13 (Fig. 7B) showed the same tendency as that of IL-1b. However, the expression of mRNAMMP-13 was much higher than that of IL-1b after the injury. As for COX-2 (Fig. 7C), there was some difference. In the untreated defects, mRNA-COX-2 peaked at 4 weeks after operation. Except for that, the variation trend of mRNA-COX-2 is similar to that of IL-1b and MMP-13. Generally, all these three genes were up-regulated after the acute injury, which could be alleviated by transplanting the Coll or Coll/PAA-Res scaffolds. By transplanting the pure Coll scaffold, the inflammation related genes were reduced by half, while the expression of these genes decreased to one-third of that in untreated defects after the Coll/PAA-Res implantation. After 12 weeks’ implantation, an inflammatory symptom in the histological section appeared (Fig. S.3). Only small inflammatory areas were found in the Coll/PAA-Res group (Fig. S.3A), and no other inflammatory areas can be observed. There were some sporadic

4990

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

Fig. 7. qRT-PCR data of the anti-inflammatory properties of Coll/PAA-Res (0.5%) scaffold in the rabbit osteochondral defect. (A–C) The relative gene expression of IL-1b, MMP-13 and COX-2, respectively. All values are mean ± SD of three experiments. * denotes P < 0.05.

inflammatory signs in the Coll group (Fig. S.3B), while some sheet zones showed inflammatory signs in the untreated group (Fig. S.3C). The results were in good agreement with the PCR data, verifying the anti-inflammatory feature of the Coll/PAA-Res scaffold.

3.4. Cartilage and bone related gene expression in vivo Fig. 8 shows the cartilage- and bone-related gene expression 6 weeks after operation. The gene expression level of SOX-9 was up-regulated after the injury, and peaked at 4 weeks, and decreased at 6 weeks (Fig. 8A). By transplantation of Coll or Coll/PAA-Res scaffold, the gene expression level increased significantly. In this 6 week time period, the gene expression level in the Coll/PAA-Res group was higher than that of the Coll group. 6 weeks after implantation, the gene expression level of SOX-9 was back to a normal level in the untreated defects, while the expression in Coll and Coll/PAA-Res group was 3-fold higher than that in normal cartilage. As for aggrecan shown in Fig. 8B, the gene expression level also peaked at 4 weeks. However, the untreated defect group showed a higher gene level than that transplanted with a scaffold at 2 weeks. Fig. 8C displays the gene expression level of Coll II. The gene expression level of Coll II was increased by transplanting a scaffold into the defects, especially when the Coll/PAA-Res scaffold was used to fill the defects. At 2 weeks, the gene expression level in the three groups was lower than that in normal cartilage, especially in the untreated group. At 4 weeks, the expression level of Coll II increased significantly, and the untreated group showed a level comparable to the normal cartilage, while the Coll or Coll/PAA-Res groups demonstrated a >2-fold gene expression compared to that of normal cartilage. At 6 weeks, the gene expression level of Coll II was inferior to

the normal cartilage in the untreated group; however, the gene expression level of Coll II in the Coll or Coll/PAA-Res groups was still superior to normal cartilage. The gene expression of Coll I was also assessed (Fig. 8D). Overall, the expression of Coll I exhibited a similar change tendency to that of Coll II. However, at 2 weeks, the gene expression level in the three groups was slightly higher than that in normal cartilage. At 4 weeks, the gene expression of Coll I increased sharply, and the untreated group showed a level of 5.5-fold that in normal cartilage. In contrast, the Coll or Coll/PAA-Res groups achieved 5-fold greater gene expression than normal cartilage. At 6 weeks, the gene expression level of Coll I was superior to the normal cartilage in all of these three groups. As a whole, SOX-9 and aggrecan have shown a much higher expression level than that of Coll II and Coll I. The transplantation of a scaffold in the defects can up-regulate these four genes compared to the untreated defects, while much higher expression level can be obtained by Coll/PAA-Res scaffold.

3.5. Histological examination After implantation for 12 weeks, in gross view, no obvious changes in walking pattern were observed for all the rabbits regardless, of the groups. The joints did not show contracture or synovitis in the Coll and Coll/PAA-Res group, which indicated good biocompatibility. However, synovitis was found in the untreated defects. Fig. 9A exhibits the macroscopic observation of the neo-tissue. Clearly, the defect in the Coll/PAA-Res group was covered with a smooth cartilage-like tissue, presenting an overall smooth joint surface and a normal color and texture as that of the surrounding

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

4991

Fig. 8. qRT-PCR data of the cartilage and bone related genes. (A–D) The relative gene expression of SOX-9, aggrecan, Coll II, and Coll I, respectively. All values are mean ± SD of three experiments. * denotes P < 0.05.

native cartilage (Fig. 9A1). The border between the neo-cartilage and the surrounding normal tissue could not be distinguished, manifesting a uniform integration of the neo-cartilage. The defect in the Coll group was fully covered with white tissue, which appeared to be a slightly different color to that of the surrounding normal tissue. Moreover, the regenerated cartilage was irregular with a rougher surface, and the parts adjacent to the surrounding tissues can be distinguished (Fig. 9B1). The defect seemed to be not repaired at all. A vacancy in the defect was observed, and the bottom was filled with a gel-like substance (Fig. 9C1). The neo-tissue was then analyzed by H&E, PAS and immunohistochemical staining of collagen type II. As shown in Fig. 9A2, the defect was fully filled by the neo-tissues in the Coll/PAA-Res group. The surface of the regenerated tissue was smooth, and both sides of the neo-tissue integrated very well with adjacent tissue. The upper layer of the neo-tissue was stained a dark purple-magenta color by PAS and brown by Coll II staining (Fig. 10A). It indicated that the tissue was composed of GAGs and Coll II. The surface layer exhibited the same thickness with its surrounding cartilage and the cell density in this layer was somewhat thinner than that of the native cartilage, and lacunas and cell clusters were apparent (Fig. 9A). All the histological appearances confirmed that the upper layer of the neo-tissue was cartilage. This neo-tissue integrated very well with its subchondral bone, which was also well organized. Since the depth of the initial defects was up to subchondral bone, the observed subchondral bone should be newly regenerated. There was no observable residual scaffold in the neo-tissue. As for the Coll group, the defects were also fully filled with repaired tissue. However, the surface and subchondral zone of the neo-tissue were highly irregular (Figs. 9B and 10B). At the

surface of the neo-tissue, there were vacancies left penetrating into the subchondral bone in the center, as observed in Fig. 9B2. However, the neo-tissue integrated well with adjacent tissue. The upper layer of neo-tissue can be stained a purple-magenta color by PAS and brown by Coll II, which were weaker than those of normal cartilage, indicative of fewer GAGs and Coll II in the neo-tissue. In the neo-tissue near the native cartilage, the cells in partial region has exhibited a regular column-like alignment and basically were directed to the center of the defects (Fig. 9B3). The neo-tissue in the defect center displayed the subchondral bone regeneration, while the surface layer exhibited the less cartilage-like feature. The whole surface layer penetrated irregularly to the subchondral bone, suggesting a partial restoration by implantation of pure Coll scaffold. The untreated defect was filled with fibrous tissue only (Figs. 9C and 10C). Only some of the cartilage matrix near the host tissue was stained by PAS and Coll II. Chondrocyte-like cells were observed only at the connected region. Subchondral bone was formed incompletely and irregularly.

4. Discussion Recently, cell therapy has been clinically used in the treatment of cartilage and osteochondral damage, but the effectiveness of any cellular repair approach depends on the retention of cell viability after implantation. Oshima et al. prepared polylactic acid (PLA) scaffold loaded with 1  106 BMSCs. The transplanted BMSCs numbers decreased from 7.8  105 at 1 week, to 3.5  105 at 4 weeks, to 3.8  104 at 12 weeks, and were not detectable at 24 weeks. The

4992

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

Fig. 9. Gross view (A1–C1) and histological images (A2–C2) of the neo-cartilage after transplantation for 12 weeks in rabbit knees. A3–C3, adjacent regions of the defect and host tissue, which correspond to the same side of A2–C2. A4–C4, center zone of the neo-cartilage with higher magnification (marked with box zones of A2–C2). (A–C) Coll/PAARes (0.5%) scaffold, pure Coll scaffold and untreated defect, respectively. The arrows in A2–C2 indicate the boundaries between the defect and the host tissue.

results indicate that endogenous cells participated in the reparative process, and the cells in the neo-tissue were the host cells instead of the transplanted cells [34]. In addition, cell delivery has encountered crucial barriers in therapeutic translation, including immune rejection, pathogen transmission, potential tumorigenesis, issues with packaging, storage and shipping and difficulties in clinical adoption and regulatory approval [35]. Delivery of cell-free materials may avoid the high cost and regulatory barriers involved in the cellular systems [36]. However, the implantation of biomaterials by themselves seemed to yield inferior results to cell therapy. In the previous study, PLGA/fibrin gel scaffolds transplanted into the osteochondral defects in rabbits showed fibrous tissue in the defects. The chondrocyte-like cells were hardly observed in the defect area, and little cartilage matrix was stained by GAGs and Coll II [32,33]. Also, there were some reports that fibrocartilaginous repair tissue was found after cellfree pure alginate treatment [37], and sporadically hyaline-like islands were observed by the cell-free alginate-gelatin hydrogels

[9,38]. The difference in scaffolds may significantly influence the cell behavior, and thereby the repair results. Numerous scaffolds have been used in the bone and cartilage regeneration. However, there are several criteria to be considered for designing the scaffold for treating osteochondral damage: (1) maintaining the morphology and phenotype of chondrocytes and BMSCs; (2) degradation by-products without toxicity or inflammation, and being cleared from the body easily; (3) matching biodegradation and tissue growth rate; (4) matching scaffold and native cartilage mechanical properties [3]. The atelocollagen scaffold fabricated in this study can meet the above four criteria. Firstly, in vitro results showed that the Coll scaffolds could support the growth and maintain the morphology and phenotype of chondrocytes and BMSCs (Figs. 5 and 6). Secondly, the degradation byproducts of Coll are nutrient substance for regulating and directing the cell function. The atelocollagen was used for cartilage regeneration, without problems of immunogenicity or inflammation [39,40]. Third, the Coll scaffold in this research showed slow

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

4993

Fig. 10. Cartilage ECM in the neo-cartilage. GAGs straining by PAS (A1–C1) and the immunohistochemical staining of Coll II (A2–C2).

degradation rate in PBS and can be degraded in collagenase solution, and the degradation rate could be controlled by the content of Res. Coll/Res scaffold showed a much slower degradation rate than pure Coll scaffold in collagenase solution. When the neo-tissue in the defects was retrieved at 2, 4 and 6 weeks for PCR, the degradation behavior of the transplanted scaffolds could be observed from the visual inspection. The profile of the Coll and Coll/PAA-Res scaffolds can be distinguished after 2 weeks’ implantation, while only scattered fragments were observed after 4 weeks, and the scaffolds disappeared after 6 weeks. However, by naked-eye observation only, it was difficult to distinguish the different degradation rate of Coll and Coll/PAA-Res in vivo. From the results of PCR, it was interesting to find that all of the bone and cartilage related genes peaked at 4 weeks in the untreated defects, which indicated a maximum endogenous regeneration capacity at this time. It is reasonable to suggest that the scaffold in osteochondral defects should be degraded slowly in the first 2 weeks due to the necessity for mechanical support, and the scaffold should be degraded quickly at 4 weeks to make room for the new tissue. Taken in this sense, the biodegradation of Coll and Coll/PAA-Res scaffolds matched the tissue growth rates. In addition, our PCR results showed that the inherent regeneration rates changed with time, which may provide a clue for how to design the biodegradable scaffold to match the tissue growth rates. Fourth, 4% Coll was employed in this study, and a scaffold with the compressive strength of 0.5 MPa was obtained. The strengths can be further enhanced by incorporating Res. The compressive strength of our scaffolds was comparable with the normal cartilage (the equilibrium confined compression modulus of cartilage ranges from 0.08 to 2.1 MPa from superficial to deep layers of adult bovine cartilage [41]). Typically, a low Coll solution (0.5%) was commonly used to fabricate a scaffold, which led to a very poor strength and shrinking sharply by the cells and also resorption quickly in vivo. In this research, the high Coll solution provided the scaffold with a strong mechanical strength, which may promote BMSCs migration and differentiation into the osteoblast [42]. Pro-inflammatory cytokines have been identified to initiate enhanced apoptosis of chondrocytes [12,13]. For these reasons, it is of clinical significance to find natural remedies to inhibit the release of these pro-inflammatory cytokines. In this research, antiinflammatory scaffolds (Coll/PAA-Res) showed an enhanced ability to capture DPPH free radicals, and eventually offered a protective effect on the chondrocytes and BMSCs in a ROS environment, as

reported previously [16,20]. Then, an acute trauma defect in the rabbit joint was created to evaluate the anti-inflammatory effect of the Coll/PAA-Res scaffold. The acute trauma and foreign body implants may cause inflammation of the host in the defect region to elicit the gathering of polymorphonuclear cells, catabolic enzymes and IL-1b and TNF-a [43]. After in vivo implantation into the trauma defect, both Coll and Coll/Res scaffold can down-regulate inflammatory factors, such as the IL-1b, MMP-13 and COX-2, as reported by others [12,13,18]. It was interesting to find that transplantation Coll and Coll/PAA-Res scaffold seemed to be no foreign body reaction. Filling the defect with a scaffold capable of suppressing inflammation could be in favor of the migration of BMSCs to the defect zone. It was noted that the acute foreign body reaction disappeared after 2 weeks’ implantation. Due to the best anti-inflammatory effect in vitro, the Coll/PAARes (0.5%) scaffolds were implanted into the osteochondral defects. About 400 lg Res was loaded in the one scaffold. Some researchers demonstrated that Res was effective at concentrations ranging from 5 to 100 lmol l1 in protecting chondrocytes in vitro [12,17]. In this study, although there was only 400 lg Res in each scaffold, Res/Coll scaffolds exhibited a notable protective effect on the chondrocytes and BMSCs in the H2O2 environment, thus inhibiting the expression of IL-1b, MMP-13 and COX-2 in vivo. After implantation into an osteochondral defect, the diffusion of released Res was restricted to one direction (articular surface), because the other sides of scaffold were encompassed by the host tissue. In view of steric effect, there may be a high concentration of Res in the osteochondral defects for a long period. In this study, an osteochondral defect was made in the rabbit joint, and the defects were filled with bone marrow immediately after the damage, so the BMSCs from the bone marrow can migrate readily to the defects, just as those in a classical microfracture procedure. The BMSCs may be the major cells to restore the osteochondral defects. However, as shown in Fig. 9A2, B2 and C2, it was observed that both the cells in the cartilage layer and in the subchondral bone adjacent to the host tissue are orderly arranged in some extent. Those cells were directed to the center defects from the host tissue, implying that the cells migrated from the host tissue to the defects. That is, the neo-cartilage can be restored from the sides of the defect border other than the bottom of the defects and a cell flow was involved in the reparative process [32,33,40]. In our work, Coll/PAA-Res scaffold with mechanical properties comparable to normal tissue and excellent tissue biocompatibility

4994

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995

may facilitate the migrating, adhering and proliferation of the stem cells. Furthermore, the inflammatory factors were significantly inhibited by Res loaded in the scaffold; thus the microenvironment became more conducive to stem cell growth and differentiation to targeted chondrocytes and osteoblasts. On the other hand, growth factors have been implicated in stimulating cellular behavior and matrix production [3,33,35,40,44]. The implementation of growth factors in cartilage repair faces severe regulatory hurdles. In this work, we fabricated a resveratrol-loaded scaffold which showed high efficiency to restore the osteochondral defect in rabbit joint. Being absent of cells and growth factors, our scaffolds are much easier to be approved for straightforward surgery. In addition, these scaffolds can be transplanted under arthroscopy, which will provide one step and a minimally invasive option for the repair of osteochondral defects. 5. Conclusion In this study, naturally occurring Res was grafted to PAA to achieve a macromolecular drug with much improved water solubility compared to Res alone. PAA-Res was then incorporated with atelocollagen to construct an anti-inflammatory scaffold (Coll/ PAA-Res). The Coll/PAA-Res scaffold possessed the compressive strength comparable to normal cartilage, and demonstrated excellent cytocompatibility. The Coll/Res scaffolds were capable of scavenging DPPH free radicals and thus protected the chondrocytes and BMSCs against damaging ROS. After implantation into the osteochondral defect of rabbits, the Coll/PAA-Res scaffold exhibited anti-inflammatory activity by down-regulating the gene levels of IL-1, MMP13 and COX-2, and meanwhile up-regulating SOX-9, aggrecan, Coll II and Coll I. The Coll/PAA-Res scaffold was highly efficiently in restoring the osteochondral defects in rabbit joint after 12 weeks’ implantation. This study represents a potentially significant advancement in clinical options for osteochondral damage repair. Acknowledgements The authors gratefully acknowledge the support for this work from the National Natural Science Foundation of China (Grants 51103096, 51173129) and National Natural Science Funds for Distinguished Young Scholar (No. 51325305). Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1, 2, 4, 5, 6, 9, and 10, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/ 10.1016/j.actbio.2014.08.022. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2014.08. 022. References [1] Zhen GH, Wen CY, Jia XF, Li Y, Crane JL, Mears SC, et al. Inhibition of tgf-beta signaling in mesenchymal stem cells of subchondral bone attenuates osteoarthritis. Nat Med 2013;19:704–12. [2] Emans PJ, van Rhijn LW, Welting TJM, Cremers A, Wijnands N, Spaapen F, et al. Autologous engineering of cartilage. Proc Natl Acad Sci USA 2010;107:3418–23. [3] Huey DJ, Hu JC, Athanasiou KA. Unlike bone, cartilage regeneration remains elusive. Science 2012;338:917–21. [4] Chung C, Burdick JA. Engineering cartilage tissue. Adv Drug Deliver Rev 2008;60:243–62.

[5] Kandel RA, Grynpas M, Pilliar R, Lee J, Wang J, Waldman S, et al. Repair of osteochondral defects with biphasic cartilage-calcium polyphosphate constructs in a sheep model. Biomaterials 2006;27:4120–31. [6] Kock NB, Hannink G, van Kampen A, Verdonschot N, van Susante JLC, Buma P. Evaluation of subsidence, chondrocyte survival and graft incorporation following autologous osteochondral transplantation. Knee Surg Sport Tr A 2011;19:1962–70. [7] Lai JH, Kajiyama G, Smith RL, Maloney W, Yang F. Stem cells catalyze cartilage formation by neonatal articular chondrocytes in 3d biomimetic hydrogels. Sci Rep 2013;3:3553. [8] Asheesh Bedi RJWI. Cartilage resurfacing using synthetic composite plugs. Tech Knee Surg 2009;8:29–36. [9] Schagemann JC, Erggelet C, Chung HW, Lahm A, Kurz H, Mrosek EH. Cell-laden and cell-free biopolymer hydrogel for the treatment of osteochondral defects in a sheep model. Tissue Eng Part A 2009;15:75–82. [10] Burdick JA, Mauck RL, Gorman JH, Gorman RC. Acellular biomaterials: an evolving alternative to cell-based therapies. Sci Transl Med 2013;5:176ps4. [11] Liu WG, Deng C, McLaughlin CR, Fagerholm P, Lagali NS, Heyne B, et al. Collagen-phosphorylcholine interpenetrating network hydrogels as corneal substitutes. Biomaterials 2009;30:1551–9. [12] Shakibaei M, John T, Seifarth C, Mobasheri A. Resveratrol inhibits il-1 betainduced stimulation of caspase-3 and cleavage of parp in human articular chondrocytes in vitro. Ann N Y Acad Sci 2007;1095:554–63. [13] Elmali N, Baysal O, Harma A, Esenkaya I, Mizrak B. Effects of resveratrol in inflammatory arthritis. Inflammation 2007;30:1–6. [14] Wang J, Gao JS, Chen JW, Li F, Tian J. Effect of resveratrol on cartilage protection and apoptosis inhibition in experimental osteoarthritis of rabbit. Rheumatol Int 2012;32:1541–8. [15] Wehling P, Reinecke J, Baltzer AA, Granrath M, Schulitz KP, Schultz C, et al. Clinical responses to gene therapy in joints of two subjects with rheumatoid arthritis. Human Gene Ther 2009;20:97–101. [16] Sheu SY, Chen WS, Sun JS, Lin FH, Wu T. Biological characterization of oxidized hyaluronic acid/resveratrol hydrogel for cartilage tissue engineering. J Biomed Mater Res A 2013;101:3457–66. [17] Dave M, Attur M, Palmer G, Al-Mussawir HE, Kennish L, Patel J, et al. The antioxidant resveratrol protects against chondrocyte apoptosis via effects on mitochondrial polarization and atp production. Arthritis Rheum 2008;58:2786–97. [18] Baur JA, Sinclair DA. Therapeutic potential of resveratrol: the in vivo evidence. Nat Rev Drug Discov 2006;5:493–506. [19] Liu FC, Hung LF, Wu WL, Chang DM, Huang CY, Lai JH, et al. Chondroprotective effects and mechanisms of resveratrol in advanced glycation end productsstimulated chondrocytes. Arthritis Res Ther 2010;12. [20] Csaki C, Keshishzadeh N, Fischer K, Shakibaei M. Regulation of inflammation signalling by resveratrol in human chondrocytes in vitro. Biochem Pharmacol 2008;75:677–87. [21] Lei M, Liu SQ, Liu YL. Resveratrol protects bone marrow mesenchymal stem cell derived chondrocytes cultured on chitosan-gelatin scaffolds from the inhibitory effect of interleukin-1 beta. Acta Pharmacol Sin 2008;29:1350–6. [22] Li Y, Danmark S, Edlund U, Finne-Wistrand A, He X, Norgard M, et al. Resveratrol-conjugated poly-epsilon-caprolactone facilitates in vitro mineralization and in vivo bone regeneration. Acta Biomater 2011;7:751–8. [23] Dosier CR, Erdman CP, Park JH, Schwartz Z, Boyan BD, Guldberg RE. Resveratrol effect on osteogenic differentiation of rat and human adipose derived stem cells in a 3-d culture environment. J Mech Behav Biomed 2012;11:112–22. [24] Henrotin Y, Lambert C, Couchourel D, Ripoll C, Chiotelli E. Nutraceuticals: do they represent a new era in the management of osteoarthritis? – a narrative review from the lessons taken with five products. Osteoarthr Cartilage 2011;19:1–21. [25] Ansari KA, Vavia PR, Trotta F, Cavalli R. Cyclodextrin-based nanosponges for delivery of resveratrol: in vitro characterisation, stability, cytotoxicity and permeation study. AAPS Pharmscitech 2011;12:279–86. [26] Caddeo C, Manconi M, Fadda AM, Lai F, Lampis S, Diez-Sales O, et al. Nanocarriers for antioxidant resveratrol: formulation approach, vesicle selfassembly and stability evaluation. Colloid Surf B 2013;111:327–32. [27] Lucas-Abellan C, Fortea I, Lopez-Nicolas JM, Nunez-Delicado E. Cyclodextrins as resveratrol carrier system. Food Chem 2007;104:39–44. [28] Mont Kumpugdee-Vollrath YI. Increasing solubility of poorly water soluble drug resveratrol by surfactants and cyclodextrins. Adv Mater Res 2012;418– 420:2231–4. [29] Lu Z, Cheng B, Hu YL, Zhang YH, Zou GL. Complexation of resveratrol with cyclodextrins: solubility and antioxidant activity. Food Chem 2009;113: 17–20. [30] Walle T, Hsieh F, DeLegge MH, Oatis JE, Walle UK. High absorption but very low bioavailability of oral resveratrol in humans. Drug Metab Dispos 2004;32:1377–82. [31] Li JJ, Shu Y, Hao T, Wang Y, Qian YF, Duan CM, et al. A chitosan-glutathione based injectable hydrogel for suppression of oxidative stress damage in cardiomyocytes. Biomaterials 2013;34:9071–81. [32] Wang W, Li B, Li Y, Jiang Y, Ouyang H, Gao C. In vivo restoration of fullthickness cartilage defects by poly(lactide-co-glycolide) sponges filled with fibrin gel, bone marrow mesenchymal stem cells and DNA complexes. Biomaterials 2010;31:5953–65. [33] Wang W, Li B, Yang JZ, Xin L, Li YL, Yin HP, et al. The restoration of fullthickness cartilage defects with bmscs and tgf-beta 1 loaded plga/fibrin gel constructs. Biomaterials 2010;31:8964–73.

W. Wang et al. / Acta Biomaterialia 10 (2014) 4983–4995 [34] Oshima Y, Harwood FL, Coutts RD, Kubo T, Amiel D. Variation of mesenchymal cells in polylactic acid scaffold in an osteochondral repair model. Tissue Eng Part C 2009;15:595–604. [35] Lee CH, Cook JL, Mendelson A, Moioli EK, Yao H, Mao JJ. Regeneration of the articular surface of the rabbit synovial joint by cell homing: a proof of concept study. Lancet 2010;376:440–8. [36] Ma T. Acellular biomaterials in mesenchymal stem cell-mediated endogenous tissue regeneration. J Mater Chem B 2014;2:31–5. [37] Fragonas E, Valente M, Pozzi-Mucelli M, Toffanin R, Rizzo R, Silvestri F, et al. Articular cartilage repair in rabbits by using suspensions of allogenic chondrocytes in alginate. Biomaterials 2000;21:795–801. [38] Diduch DR, Jordan LCM, Mierisch CM, Balian G. Marrow stromal cells embedded in alginate for repair of osteochondral defects. Arthroscopy 2000;16:571–7. [39] Ito Y, Ochi M, Adachi N, Sugawara K, Yanada S, Ikada Y, et al. Repair of osteochondral defect with tissue-engineered chondral plug in a rabbit model. Arthroscopy 2005;21:1155–63.

4995

[40] Zhang W, Chen JL, Tao JD, Jiang YZ, Hu CC, Huang L, et al. The use of type 1 collagen scaffold containing stromal cell-derived factor-1 to create a matrix environment conducive to partial-thickness cartilage defects repair. Biomaterials 2013;34:713–23. [41] Schinagl RM, Gurskis D, Chen AC, Sah RL. Depth-dependent confined compression modulus of full-thickness bovine articular cartilage. J Orthop Res 1997;15:499–506. [42] Engler AJ, Sen S, Sweeney HL, Discher DE. Matrix elasticity directs stem cell lineage specification. Cell 2006;126:677–89. [43] Hollander AP, Dickinson SC, Sims TJ, Brun P, Cortivo R, Kon E, et al. Maturation of tissue engineered cartilage implanted in injured and osteoarthritic human knees. Tissue Eng 2006;12:1787–98. [44] Diao H, Wang J, Shen C, Xia S, Guo T, Dong L, et al. Improved cartilage regeneration utilizing mesenchymal stem cells in tgf-beta1 gene-activated scaffolds. Tissue Eng Part A 2009;15:2687–98.