Background
Lesions in the knee joint surface are commonly treated with microfracture [
1], autologous cell implantation (ACI) [
2], or osteoarticular autograft transfer system (OATS) [
3]. Although patients have symptomatic relief, there is little convincing histological or biochemical data to support the contention that the new tissue that forms is characteristic of native hyaline cartilage found on the joint surface that is composed predominantly of type II collagen. ACI and microfracture most often result a fibrous cartilage repair that is high in type I collagen and not durable in weight bearing positions over the long term. Roberts et al. have reported that as many as 65 % of second look biopsies showed fibrocartilage [
4]. Recent publications including a meta-analysis suggest there are no differences among ACI, microfracture, and OATS [
5,
6]. Newer modifications of the ACI technique being tested in Europe involve an open weave or sponge matrix (MACI), frequently made from collagen or hyaluronic acid, where the cells are absorbed into the matrix before being secured in the lesion [
7]. These woven type scaffolds do not provide any immediate biomechanical integrity and can be crushed by the forces placed on the joint. Furthermore, recent results suggest that the new tissue formed is fibrocartilage and the fate of the cells is unknown [
8,
9]. Although MACI techniques also relieve pain, the long-term results using MACI are not yet available [
10‐
12]. Similarly, the long-term clinical results for cartilage regeneration employing particulated juvenile allogeneic cartilage (DeNovo NT, Zimmer, Warsaw, IN) and the cartilage autograft implantation system (CAIS, Johnson & Johnson, New Brunswick, NJ) are not fully known [
13‐
15]. In summary, the long-term outcomes of many cartilage restorative procedures are unsatisfactory and an improved method for joint surface repair is a clear unmet need in orthopaedic surgery.
Tissue engineering strategies combining chondrocytes or chondrocyte progenitor cells with biomimetic scaffolds made of natural or synthetic biomaterials could be a promising alternative for cartilage repair and regeneration. New cartilage matrix has been successfully produced in immunocompromised animals with cells placed on open fibrous scaffolds, such as collagen or polyesters, but these open lattice networks also permit invasion of inflammatory cells in immune competent animal models that can negatively affect matrix formation [
16]. Hydrogels have high water content and can be modified to have diverse physical properties for use in medical implants, biosensors, and drug-delivery devices [
17]. Prior to polymerization, hydrogels can be injected in a minimally invasive manner to fill defects of any size and shape [
18‐
20]. Chondrocytes or chondroprogenitors can be homogenously suspended in a three-dimensional hydrogel, where the encapsulated cells can retain a rounded morphology that may induce or sustain a chondrocytic phenotype. The porous nature of hydrogels permits the transport of nutrients and waste. They also permit the translation of mechanical loads to encapsulated cells, similar to normal physiological conditions [
21]. Although mechanically weak gels may be a drawback, the mechanical properties can be changed by crosslinking chemistry [
22]. Besides, many degradable hydrogels degrade too fast to permit sufficient ECM production for neocartilage formation. Nonetheless, various research groups including ours have successfully demonstrated neocartilage tissue formation using multiple kinds of hydrogels to encapsulate chondrocytes or mesenchymal stem cells [
23‐
28].
Poly(ethylene glycol) (PEG) is a linear carbon polymer that can be photochemically crosslinked into a stable hydrogel in which the biochemical and biophysical properties can be custom designed to obtain desirable properties that permit controlled cartilage matrix production. Our previous work has described a photocrosslinkable hydrogel for cell encapsulation using PEG polymer [
21,
29]. PEG has a long history in medical applications with desirable chemistry that allows easy chemical modification. For example, the PEG polymer chain can be methacrylated and referred to as PEG diacrylate (PEGDM). In the presence of a photoinitiator (Igracure 2959, BASF Corp., Florham Park, NJ) and ultraviolet light (365 nm), the PEGDM forms crosslinks between the linear chains and forms a hydrogel. The polymers and gelation process allow easy placement into joint lesions and provide mechanical and structural stability during the regeneration process with desirable transport properties [
30]. Previous work has shown that both the amount and distribution of extracellular matrix (ECM) by chondrocytes encapsulated in photopolymerized PEG hydrogels in vitro is directly correlated to the pore size, degradation rate, and swelling behavior of the gel networks [
30‐
33]. When the network mesh pore size is too small (low swelling), the distribution of large ECM molecules (e.g. glycsosaminoglycans) is confined to the pericellular region. By increasing the pore size, the GAG molecules can diffuse throughout the intercellular spaces [
32]. Additionally, the degradation properties of the PEG hydrogels (controlled by cross-linking density) can also affect the distribution of ECM. When chondrocytes are encapsulated in nondegradable PEG hydrogels, the collagen molecules are confined to the pericellular area, whereas the collagen is dispersed evenly into the void volume between the cells in gels that are designed to biodegrade [
33]. The degradation characteristics can be tailored to make gels that degrade rapidly or slowly over time depending on the amount of degradable lactide units grafted to the PEG. Creating gels that have bimodal degradation could allow the scaffolds to retain their architecture during the tissue formation and remodeling processes.
Based on our earlier work on the degradation kinetics of copolymer gels, several candidate mixtures were tested in preliminary studies to formulate favorable PEGDM copolymer hydrogels that supported chondrocyte survival and growth, and allowed new matrix production [
34]. We selected a degradable poly(ethylene glycol)-4,5 lactic acid dimethacrylate (PEG-LA-DM) and nondegradable PEGDM macromer for this study. Preliminary studies showed that PEG-LA-DM that was fully degradable dissolved too rapidly in vitro for the chondrocytes to make new ECM. Thus, the PEG-LA-DM was combined with nondegradable PEGDM in molar ratios that ranged from 90 % PEG-LA-DM/10 % PEGDM to a 50/50 ratio. The aims of our study were: 1) to demonstrate whether combining degradable and nondegradable PEGDM copolymer hydrogels would permit neocartilage formation by articular chondrocytes photoencapsulated into the gels; and 2) to investigate if the newly formed cartilage matrix could integrate with the native cartilage.
Methods
The Institutional Animal Care and Use Committee of the Massachusetts General Hospital approved all procedures with animal tissues and live animals.
Chondrocyte isolation
Swine articular cartilage was harvested from knee joints of three- to six- month-old female Yorkshire swine carcasses that were euthanized from other approved studies in our hospital (laboratory animal vendor: Tufts University, Grafton, MA). The cartilage was minced into 1 mm3 pieces and digested using 0.05 % collagenase type 2 (Worthington Biochemical Co., Freehold, NJ) solution for 16 h at 37 °C. After digestion the chondrocyte suspension was filtered through a 100 μm sterile cell strainer (Becton Dickinson, Franklin Lakes, NJ) to remove undigested debris. Cells were centrifuged and washed three times in phosphate-buffered saline. Using trypan blue and a hemacytometer, cell number and viability was assessed. Only cell isolations with viability above 90 % were used for the further experiments.
Polymer preparation
The synthesis of degradable PEG-LA-DM and nondegradable PEGDM used in this study have been previously described [
34]. Briefly, the macromer combinations were dissolved in sterile phosphate-buffered saline to a final concentration of 10 % (
w/w). Mixed molar ratios of 50/50, 60/40, 70/30, 80/20 and 90/10 (PEG-LA-DM:PEGDM) were prepared for the study. 2-hydroxy-1[4-(hydroxyethoxy)phenyl]-2-methyl-1- propanone (Igracure; I2959) was used as the photosensitive initiator added to the polymer solutions at a final concentration of 0.05 % (
w/w).
Chondrocyte photoencapsulation
Swine articular chondrocytes were added to the macromer/photoinitiator solutions at a concentration of 6 × 107 cells/mL. 100 μL of nonpolymerized gel containing the cells was placed in each cylindrical mould measuring 4.5 mm diameter × 6.5 mm in height. The gel-cell solutions were polymerized using ultraviolet (365 nm) irradiation at ~10 mW/cm2 for 10 min.
Implantation and harvest
Experimental constructs were placed into subcutaneous pockets in 5-week-old athymic male mice (nu/nu) (MGH, Boston, MA) weighing 20–25 g and housed in a pathogen free facility. The animals were exposed to a 12 h light–dark cycle and allowed free access to sterile water and standard mouse chow. In a sterile biosafety cabinet the animals were anesthetized with an intraperitoneal injection of Avertin (250 mg/kg). A one centimeter incision was made through the skin on the dorsum of the mice and lateral subcutaneous pockets were made with blunt scissor dissection. The cell-scaffold implants were inserted into the pockets. The wound was infiltrated with 0.5 mL of sterile 0.5 % Bupivicaine for analgesia. At the time of specimen harvest the animals were euthanized by exposure to carbon dioxide from a gas source according to the recommendations of the American Veterinary Medical Association and death was confirmed by lack of breathing and heart beat.
A pilot study was performed in which implants from the five different polymer ratios (4/test polymer ratio) were made and implanted for 6 weeks. The results from the pilot study permitted us to focus only on the polymer ratios in which the matrix was contiguous and the construct volume was highest upon harvest. The full study focused only on ratios of 60/40 and 70/30. The sample size was 32 specimens for each polymer mixture. At 6, 12, and 18 weeks after implantation, eight specimens from each group were harvested. Specimens were evaluated for macroscopic appearance, weighed, and randomly selected for either histological examination or biochemical analyses (DNA; glycosaminoglycan (GAG); hydroxyproline). An additional 8 samples were prepared for each polymer mixture, but were not implanted for use as controls for assays at time zero.
Ring model preparation
A previously published model using cartilage rings was used to study the integration of the engineered cartilage with existing cartilage matrix [
35]. Using an 8 mm punch biopsy, disks of swine articular cartilage were made measuring 8 mm × 2 mm. Using a 5 mm biopsy punch, the center of the cartilage disk was removed leaving a ring of native cartilage matrix. To eliminate any influence on healing and integration by the chondrocytes in the rings, they were devitalized using five freeze–thaw cycles. Gel solution containing the chondrocytes cells was placed into the central cavities of the disks and photopolymerized. The ring model constructs were examined grossly and processed for histology to evaluate the interface between the native cartilage ring and the new cartilage formed in the copolymer gel.
Specimen evaluation
Specimens in the pilot study were harvested at 6 weeks to assess the volume and cartilage forming ability among the various ratios of the copolymers. The dimensions and wet weight of the specimens were recorded. These specimens from the pilot study were only evaluated by histology. Samples were fixed in 10 % buffered formalin for a minimum of 24 h and embedded in paraffin. Serial 5 mm sections were made, deparaffinized, and stained with hematoxylin and eosin (H&E).
Specimens in the full study comparing 60/40 and 70/30 polymer ratios were collected after 6, 12, and 18 weeks in vivo. The specimens were cut in half with one-half used for biochemical assays and the remaining half for histological processing. In addition to H&E staining to evaluate overall tissue morphology, histological sections were stained with safranin O to assess proteoglycan content of the neotissue. Sections also were immunostained with antibodies against collagen types I and II (Chondrex, Redmond WA). Briefly, slides were treated with 2 % bovine testicular hyaluronidase at room temperature for 30 min. A blocking reagent consisting of 0.3 % hydrogen peroxide in methanol was added for another 30 min followed by 10 % goat serum for 30 min. On separate test slides, antibodies against collagen type I and type II were applied for 1 h. N-Universal Negative Control was applied for the negative control, and the secondary horseradish peroxidase (HRP) labelled antibody was added for 20 min. 3,3-diaminobenzidine was applied to each slide to react with the HRP, and the sections were counterstained with hematoxylin.
Samples of native articular swine cartilage and portions of the experimental samples were analyzed biochemically. The specimens were weighed to obtain the wet weight of the samples then lyophilized. After lyophilization, the dry weight of the specimens was recorded, and the difference between the wet and dry weights was the water content of the neotissue. Papain type III solution (Sigma-Aldrich, St. Louis, MO) at 125 mg/mL was added to the lyophilized specimens to digested the cartilage matrix for 16– 24 h at 60 °C. A PicoGreen dsDNA Quantitation Assay Kit (Molecular Probes, Eugene, Oregon) was used to determine the amount of DNA or cells in the specimens. A standard curve of Lambda DNA from the kit was generated. The amount of DNA from the test specimens was determined and used as an indicator of the proliferative potential of the photoencapsulated chondrocytes. A previously published dimethylmethylene blue dye method was employed to measure the amount of GAG in the specimens using chondroitin sulfate B (Sigma-Aldrich) as a standard in the interpretation of the data [
36]. The amount of hydroxyproline in the tissues served as a surrogate marker for total collagen. Hydroxyproline was quantified using a previously published method using L-4- Hydroxyproline (Fluka Biochemika, Steinleim, Switzerland) as the standard [
37]. All the DNA, GAG, and hydroxyproline data were normalized by wet tissue weight.
Statistical analysis
Quantitative data of the study was statistically analyzed using a Student t-test (Sigmastat 2.0, SPSS Science, Chicago, IL) and ANOVA. All the values are reported as the mean ± standard deviation and the level of statistical significance was set at p < 0.05.
Discussion
Development of effective and minimally invasive tissue engineering techniques using injectable hydrogel materials to support chondrocyte growth and ECM production could be an ideal therapeutic strategy for cartilage repair and regeneration. Numerous hydrogels have been formulated for cartilage generation including natural materials such as fibrin, chitosan, collagen, hyaluronic acid (HA), and self-assembling peptides to mention a few [
19,
27,
38‐
40]. Synthetic materials also have been developed into hydrogels that permit encapsulation of chondrocytes and new cartilage matrix formation, such as poly(vinyl alcohol) and poly(ethylene glycol) [
18,
31]. Extracellular matrix components, such as GAG or HA, or growth factors have been added to these gels to promote cartilage formation in some cases [
41]. Most studies, however, have explored only single, base polymer formulations and not combinations thereof.
Gels can be formulated that are non-degradable and resistant to breakdown allowing maintenance of three-dimensional architecture. However, these resilient gels inhibit cell-to-cell contact and interfere with intracellular ECM production. Data from previous studies show that nondegradable PEGDM gels permit extracellular matrix formation in the pericellular region, but the nondegradable polymer interferes with generating contiguous cartilage matrix [
32]. Synthesizing degradable PEG gels that can break down within predictable periods of time can be achieved by adding in water soluble crosslinks to the macromer backbone. It is challenging, however, to delicately time the degradation process while maintaining volume and three-dimensional shape in vivo. Photochemical crosslinking to polymerize gels allows for molding them into predetermined shapes permitting cartilage formation in the desired form. Previously published in vitro data showed that fully degradable photocrosslinked PEG-LA-DM degraded too rapidly for cartilage matrix to form [
34]. Pilot studies in vivo with this degradable formulation confirmed this as well. By combining nondegradable with degradable PEG, we hoped to tailor the gels to form articular cartilage yet maintain structural integrity during tissue formation. Thus, the goal of this study was to evaluate neocartilage formation employing copolymers composed of a combination of degradable and nondegradable PEG macromers.
Modifications of the polymerization process can affect the mechanical and biological behaviors of the final hydrogel. The amount of crosslinking has a direct impact on the cartilage formation capacity of the chondrocytes [
21], although the crosslinking process does not negatively affect cell viability throughout the gels [
30]. Also, Omobono et al. showed that secondary treatment of photochemically crosslinked collagen gels with chemical crosslinking using N-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) was able to slow the enzymatic degradation without compromising cell viability [
42]. In the current study, we chose two polymer formulations that could be photopolymerized simultaneously with the same photoinitiator agent and the same wavelength of light. This unique strategy permitted us to blend these two different PEG formulations with the cells and polymerize them simultaneously into the final hydrogel promoting cartilage matrix formation. The pilot study confirmed that gels high in degradable PEG (80/20 and 90/10) content dissolved too rapidly for cohesive neocartilage formation. Formulations where the nondegradable PEG content was 50 % inhibited contiguous cartilage matrix formation due to the resistance of the polymer to biodegradation. These results narrowed the focus of the larger study on the photochemically crosslinked co-polymer formulations of 60/40 and 70/30.
Our data showed that the average weight and volume of the neocartilage samples were statistically higher than the initial values at the time of implantation for both groups (60/40 and 70/30) at 18 weeks. The neocartilage samples were increasingly opaque over time indicating more ECM deposition. The changes in the biochemistry of the ECM paralleled the changes in morphology over time. There was a significant increase in GAG, hydroxyproline, and total DNA content during the 18 weeks of study, and there were no statistical differences between the 60/40 and 70/30 groups at each time point. Normally, articular chondrocytes in cartilage are nurished by diffusion from the surrounding synovial fluid facilitated by the loading patterns of the joint. Ectopic implantation of these cells into subcutaneous space alters dramatically their biological environment, while the nutritive elements received from the interstitial fluid and the biomechanical stimuli of the joint are quite different.
The strength of the interface of the engineered cartilage with existing adjacent native cartilage is of great importance in articular cartilage repair because weakness at this interface could cause collapse of the new cartilage tissue and jeopardize the cartilage restoration. Promoting integration of the new and existing matrix is correlated with collagen synthesis and deposition by viable cells at the interface [
43‐
46]. The results from this study showed that the engineered cartilage integrated with existing native cartilage using these copolymer hydrogel formulations seeded with cells. Partial integration of the neocartilage with the opposing surface of the native cartilage was noted by 6 weeks (data not shown), but by week 12 and 18, the integration interface showed repopulation, deposition of matrix macromolecules, and tight adhesion between the the engineered and native cartilage. The improved integration and the tight interface junction at the later times may be related to the breakdown of the polymer scaffold over time with the simultaneous growth of new cartilage matrix.
Most studies evaluating polymer scaffolds seeded with cells have been initially tested in vitro prior to pilot testing by implantation into nude mice. Nude mice lack a repertoire of mature T-cells and cannot mount a cellular rejection response to allogeneic or xenogeneic cells. In many ways they can be considered a “living” petri dish for studying cartilage (and bone) matrix formation from cell-seeded scaffolds. We chose a mouse model with subcutaneous implantation to generate neocartilage because this permits better cartilage formation than in vitro systems. The biomechanical microenvironment in the joint cannot be replicated in the subcutaneous pockets on the back of nude mice, however. We recognize that the model may have some deficiencies, but the nude mouse is a suitable small animal model to pilot screen and test new scaffolds and tissue formation. Further modifications in hydrogel chemistry or photopolymerization mechanism can improve the scaffold properties of PEG derivatives for tissue engineering of articular cartilage, particularly in terms of the intrinsic adhesive properties. The application of these copolymers in immunocompetent animal models will be the focus of subsequent studies.
Conclusions
Our findings demonstrated that copolymers that are composed of degradable and nondegradable PEGDM macromers are favorable scaffolds for engineering articular cartilage. No obvious differences were observed among the studied copolymers at different ratios (60/40 and 70/30) in the cell proliferation, biochemical performance and integrative properties between the engineered tissue and adjacent native tissue. Potentially, a liquid macromer/chondrocyte suspension like PEGDM copolymer could be injected into the cartilaginous defect and photopolymerized to provide a minimally invasive technique to promote or enhance articular cartilage repair. Further studies are required to investigate the potential of this tissue engineering approach for chondrogenesis involving large animal models and focusing on the biomechanical properties of the engineered articular cartilage.
Abbreviations
PEG, poly (ethylene glycol); PEGDM, poly (ethylene glycol) dimethacrylate; ECM, extracellular matrix; ACI, autologous cell implantation; OATS, osteoarticular autograft transfer system; H&E, hematoxylin and eosin; GAG, glycosaminoglycan; HYP, hydroxyproline; ANOVA, Analysis of variance
Acknowledgements
The authors would like to thank Mitun Ranka for supporting in preparation and implantation of the constructs and Arthur Foubert for his great assistance during harvesting the donor animal tissues. This work was supported by grants from the National Institutes of Health (R01DE12998) and Plastic Surgery Educational Foundation.