Bilayered, Peptide-Biofunctionalized Hydrogels for In Vivo Osteochondral Tissue Repair

Abstract

Osteochondral defects present a unique clinical challenge due to their combination of phenotypically distinct cartilage and bone, which require specific, stratified biochemical cues for tissue regeneration. Furthermore, the articular cartilage exhibits significantly worse regeneration than bone due to its largely acellular and avascular nature, prompting significant demand for regenerative therapies. To address these clinical challenges, we have developed a bilayered, modular hydrogel system that enables the click functionalization of cartilage- and bone-specific biochemical cues to each layer. In this system, the crosslinker poly(glycolic acid)-poly(ethylene glycol)-poly(glycolic acid)-di(but-2-yne-1,4-dithiol) (PdBT) was click conjugated with either a cartilage- or bone-specific peptide sequence of interest, and then mixed with a suspension of thermoresponsive polymer and mesenchymal stem cells (MSCs) to generate tissue-specific, cell-encapsulated hydrogel layers targeting the cartilage or bone. We implanted bilayered hydrogels in rabbit femoral condyle defects and investigated the effects of tissue-specific peptide presentation and cell encapsulation on osteochondral tissue repair. After 12 weeks implantation, hydrogels with a chondrogenic peptide sequence produced higher histological measures of overall defect filling, cartilage surface regularity, glycosaminoglycan (GAG)/cell content of neocartilage and adjacent cartilage, and bone filling and bonding compared to non-chondrogenic hydrogels. Furthermore, MSC encapsulation promoted greater histological measures of overall defect filling, cartilage thickness, GAG/cell content of neocartilage, and bone filling. Our results establish the utility of this click functionalized hydrogel system for in vivo repair of the osteochondral unit.

Graphical Abstract

1. Introduction

Osteochondral defects represent a highly prevalent clinical problem across the US and the world, with an estimated 900,000 individuals affected each year in the US alone [1]. Articular cartilage repair in particular shows poor clinical outcomes and high patient morbidity due to the tissue’s lack of inherent regenerative capacity, presenting an ongoing clinical demand [1,2]. Current surgical techniques for the repair of cartilage also tend to generate poorly functional fibrocartilage rather than the hyaline cartilage found in healthy osteochondral tissue [3]. Furthermore, osteochondral defects typically involve not only significant articular cartilage damage, but also changes to the underlying subchondral bone structure, presenting a unique challenge for clinicians and tissue engineers to simultaneously repair these phenotypically distinct tissues [4,5].

The differences between bone and cartilage tissues are numerous and include distinct cell phenotypes, extracellular matrix (ECM) composition and organization, and the presence or absence of mineralization [4]. To better mimic these differential factors, tissue engineers have developed bilayered constructs with gross compositional differences in mineral content, proteins, mechanical properties, or cell types [6–9]. However, many strategies have yet to recapitulate the specific biochemical cues responsible for divergence of bone and cartilage differentiation from mesenchymal stem cells (MSCs) during native tissue developmental processes [10]. To this end, we previously developed a modular hydrogel system that enables the click conjugation of tissue-specific biomolecular cues for bone and cartilage development, ranging from small molecules such as bioactive peptide sequences to larger macromolecules such as glycosaminoglycans [11,12]. These bone- and cartilage-specific hydrogels are fabricated by conjugating the polymeric hydrogel crosslinker, poly(glycolic acid)-poly(ethylene glycol)-poly(glycolic acid)-di(but-2-yne-1,4-dithiol) (PdBT), with tissue-specific moieties via a mild, aqueous, and Cp*Ru(cod)Cl-catalyzed click reaction, followed by mixing the biofunctionalized crosslinker with thermoresponsive poly(N-isopropylacrylamide-co-glycidyl methacrylate) (P(NIPAAm-co-GMA)) and MSCs at physiological temperature [11]. We further demonstrated the ability of these tissue-specific hydrogels to promote distinct bone and cartilage development from MSCs in vitro, with outcomes dependent on the identity and concentration of biomolecular cues conjugated to PdBT [12]. While the individual bioactivity of PdBT macromers has been established for osteogenesis and chondrogenesis, the ability to produce clinically relevant osteochondral tissue formation in a physiological environment was yet unknown. In the present study, we thus sought to develop bilayered hydrogels for the in vivo repair of full thickness osteochondral defects, by implanting cartilage- and bone-specific hydrogel layers together in a rabbit femoral defect model (Figure 1).

Figure 1:

In vivo assessment of osteochondral tissue regeneration by implantation of bilayered, tissue-specific hydrogels. Cartilage-specific layers were generated by the click conjugation of a chondrogenic N-cadherin peptide sequence to the hydrogel crosslinker, while bone-specific layers were generated by the click conjugation of an osteogenic glycine-histidine-lysine peptide sequence.

Articular cartilage repair was enabled by the bioconjugation of an N-cadherin (NC) peptide sequence, mimicking a protein that has been implicated in formation of cell-cell contacts during MSC condensation, a critical early-stage chondrogenic process in utero [13]. Subchondral bone repair was supported by the bioconjugation of a glycine-histidine-lysine (GHK) peptide fragment from osteonectin, an osteogenic glycoprotein prominently involved in mid- to late-stage bone mineralization [14]. These peptides were selected on the basis of their in vitro bioactivity when conjugated to PdBT [12]. Given the strong clinical demand for articular cartilage repair, we focused on investigating the effects of NC peptide bioconjugation on cartilage formation. Furthermore, given the involvement of MSCs in osteochondral tissue development and their prior utilization in vitro [3,12], we assessed if the delivery of encapsulated MSCs supported tissue development within bilayered hydrogels. Thus, we hypothesized that articular cartilage repair would be enhanced by both the click conjugation of the NC peptide to the top hydrogel layer and by the encapsulation of MSCs within the entire bilayered construct. To investigate these hypotheses, we implanted three experimental groups in full thickness rabbit osteochondral defects (Figure 2) for a duration of 12 weeks, based on prior studies using this animal model [8,9], and characterized the degree of osteochondral tissue regeneration through histological analysis. Ultimately, we sought to develop a stratified, tissue-specific hydrogel system for the simultaneous repair of cartilage and bone tissue within the osteochondral unit.

Figure 2:

Bilayered hydrogel formulations for implantation in rabbit medial femoral condyles.

2. Materials and Methods

2.1. Materials

Fetal bovine serum (FBS) was purchased from Gemini Bio-Products (Sacramento, CA). Low glucose Dulbecco’s modified eagle medium (DMEM) and antibiotic-antimycotic were purchased from ThermoFisher Scientific (Waltham, MA). Phosphate buffered saline (PBS) and 10% neutral buffered formalin were purchased from MilliporeSigma (St. Louis, MO). New Zealand White (NZW) rabbits were purchased from Charles River Laboratories (Wilmington, MA). Lactated Ringer’s Solution (LRS), ethylenediaminetetraacetic acid (EDTA) solution, and all other described surgical supplies and substances were purchased from Patterson Veterinary Supply (Devens, MA). Ultrapure water was obtained from a Millipore Super-Q water system (Billerica, MA).

2.2. MSC Harvest and Culture

MSCs were harvested by aspiration of bone marrow from the tibia of six anesthetized 6 month old male NZW rabbits, in accordance with protocols approved by the Rice Institutional Animal Care and Use Committee and in agreement with the animal care and use guidelines set forth by the National Institutes of Health [11]. Cells from all rabbits were pooled to minimize the effect of interanimal variability on the cell population. Adherent cells from the bone marrow were cultured in DMEM supplemented with 10% v/v FBS and 1% v/v antibiotic-antimycotic inside a humidified incubator at 37 °C and 5% CO₂, and cryopreserved until point of usage. All MSCs used for encapsulation were utilized at passage 3.

2.3. Fabrication of Tissue-Specific Hydrogels

The osteogenic GHK (“GGGGHKSP”) and chondrogenic NC (“GGGHAVDI”) peptide sequences with N-terminal azides were synthesized by solid phase peptide synthesis as described in detail elsewhere [11,15]. The PdBT crosslinker, peptide functionalized crosslinkers GHK/PdBT and NC/PdBT, and thermoresponsive P(NIPAAm-co-GMA) were synthesized as described previously, with functionalization of GHK/PdBT and NC/PdBT at a 2:1 molar ratio of peptide:PdBT [11,12]. All polymers were pre-sterilized by UV exposure for 3 h, while Teflon molds and other materials were sterilized by autoclaving or ethylene oxide exposure.

Tissue-specific hydrogel layers of 3 mm diameter and 1.5 mm height were fabricated for bilayered implantation in defects of 3 mm diameter and 3 mm height. Hydrogel layers were fabricated by mixing P(NIPAAm-co-GMA); either MSCs or no cells; and either PdBT (non-bioactive), GHK/PdBT (osteogenic), or NC/PdBT (chondrogenic) in PBS on ice to generate homogeneous pre-hydrogel suspensions, which were then injected into cylindrical Teflon molds that were stored inside a humidified incubator at 37 °C and 5% CO₂. After 1.5 h of gelation and crosslinking, hydrogel layers were sterilely transferred to 1 mL PBS with 1% v/v antibiotic-antimycotic and used within 16 h. Hydrogel formulations contained 10% w/v P(NIPAAm-co-GMA), either 0 or 5,000,000 MSCs/mL, and 25 mM of the selected crosslinker [12].

2.4. Study Design and Animal Surgery

Three different experimental groups (Figure 2) were utilized for implantation in twenty 6 month old male NZW rabbits, in accordance with protocols approved by the Rice Institutional Animal Care and Use Committee and in agreement with the animal care and use guidelines set forth by the National Institutes of Health. Experimental groups were designed to enable independent analysis of the effects of 1. chondrogenic peptide presentation (Cell+/Chondro− vs. Cell+/Chondro+) and 2. MSC delivery (Cell−/Chondro+ vs. Cell+/Chondro+) on the primary clinical target of cartilage development. Each formulation was implanted in 12–14 femoral condyle defects as determined using power analysis and prior studies utilizing the same animal model [8,9]. Treatments were performed bilaterally, with randomized assignment of hydrogel formulations to osteochondral defects.

Surgeries were performed according to previously established methods with the same animal model [8,9,16]. All rabbits were sedated by subcutaneous injection of ketamine (40 mg/kg) and acepromazine (0.5 mg/kg), followed by intubation and inhalation general anesthesia under isoflurane and oxygen. Buprenorphine (0.5 mg/kg) and bupivacaine, 0.25% (<1 mL/kg) were also administered by subcutaneous injection preoperatively, with the latter administered at the site of incision for local analgesia. Enrofloxacin (10 mg/kg) and carprofen (4 mg/kg) were also administered by subcutaneous injection perioperatively and twice postoperatively at days 1 and 2 to minimize infection risk and postoperative discomfort. Incisions were generated longitudinally and medial to the patellar tendon, followed by lateral luxation of the patella and exposure of the medial femoral condyle. Full thickness, cylindrical (3 mm diameter and 3 mm depth) osteochondral defects were generated in the medial femoral condyle of each knee using a dental drill with 2 mm, 2.75 mm, and 3 mm drill bits and 3 mm depth stops under constant irrigation with LRS. 3 mm diameter and depth have been established in the literature as a critical size defect for this medial femoral condyle model, in which the osteochondral defect does not undergo spontaneous healing in the absence of treatment [8,16,17]. After cleaning of defects with LRS and subsequent hydrogel implantation, the patella was repositioned and the joint capsule, fascia, and skin were sutured using 3–0 Vicryl, 3–0 Monocryl, and 4–0 Vicryl, respectively. Rabbits were returned to cages upon recovery from anesthesia with allowance of unrestricted movement and weight-bearing activity, followed by postoperative assessment of appropriate activity and behavior.

2.5. Tissue Processing and Histology

Rabbits were euthanized 12 weeks postoperatively by intravenous administration of Beuthanasia (0.22 mL/kg) and bilateral thoracotomy as confirmation of euthanasia, under ketamine/acepromazine sedation and isoflurane anesthesia as described above. Tissue processing was performed by previously established methods [8,9]. The medial femoral condyles were isolated, examined, fixed in 10% neutral buffered formalin for 48h at 37 °C, and stored in 70% ethanol. Tissue samples were then decalcified in EDTA solution, followed by dehydration through a graded series of ethanol to paraffin embedding at 58 °C. Longitudinal sections of 5 μm thickness were acquired from each tissue sample using a microtome and stained with H&E, Safranin O/Fast Green, van Gieson’s Picrofuchsin, and Masson’s trichrome. Sections were acquired by trimming down to the defect region, followed by acquisition of section bands spaced at 100 μm apart, with 2–3 sections per stain per implant as technical replicates. All sections were imaged at 10x using a Nikon (Tokyo, Japan) Eclipse Ti2-E inverted microscope system with an attached DS-Fi3 color camera, with automated stitching to generate a single image of each section.

Histological sections were scored blindly and independently by three evaluators (Y.S.K., G.L.K., and J.L.) using a modified O’Driscoll histological scoring system that includes 11 parameters for overall defect filling, subchondral bone formation, and articular cartilage formation (Table S1) [8,9,18].

2.6. Quantification of Safranin O Staining Intensity

To analyze the intensity of Safranin O staining in surface tissue, a consistent 1 mm × 3 mm region of interest was placed at the edge of the surface tissue and within the defect margins of each Safranin O/Fast Green image. The region of interest was subtracted of background, deconvoluted to its red channel to isolate Safranin O staining, converted to binary, and analyzed for stain intensity using ImageJ [19,20].

2.7. Statistical Analysis

Histological scores within each of the 11 scoring categories were analyzed using the non-parametric Wilcoxon Ranked Sum Test (α=0.05) [8,21]. Safranin O staining intensity was analyzed using a parametric one-way analysis of variance, followed by post-hoc testing using Tukey’s honestly significant difference (α=0.05) [20].

3. Results

3.1. Overall Observations and Histological Imaging

All rabbits resumed normal activity and behavior within 7 days postoperatively, and exhibited negligible signs of infection or injury at the incision site. After tissue retrieval at 12 weeks postoperatively, the isolated femoral condyles showed no gross signs of infection, inflammation, or swelling (Figure S1). Medial femoral condyles were sectioned at 5 μm thickness and stained with H&E, Safranin O/Fast Green, and van Gieson’s Picrofuchsin.

As shown by histological imaging, almost all samples demonstrated some degree of new tissue development, particularly in the subchondral bone layer, as well as fragmentation of the implanted hydrogels (Figure 3). Residual hydrogel was visible as translucent, unstained material that was mostly localized to the subchondral layer (Figures 3, S2). Pockets of implant degradation with incomplete tissue filling were also observed as void space near the center of many samples (Figure 3). Overall, Cell+/Chondro+ implants (Figures 3D–3F), which originally contained MSCs and a chondrogenic peptide sequence, appeared to contain grossly more tissue within the defect site than Cell+/Chondro− implants (Figures 3A–3C), which had MSCs but no chondrogenic peptide sequence, and Cell−/Chondro+ implants (Figures 3G–3I), which were acellular but had a chondrogenic peptide sequence. Cell−/Chondro+ samples mostly demonstrated defect filling with residual hydrogel fragments (Figures 3G–3I), and this leftover implant material was not considered as neotissue in any quantitative scoring.

Figure 3:

Example histological images of the overall defect for (A-C) Cell+/Chondro−, (D-F) Cell+/Chondro+, and (G-I) Cell−/Chondro+ implants at 12 weeks postoperatively. From left to right: H&E, Safranin O/Fast Green, and van Gieson’s Picrofuchsin. Dashed boxes indicate the approximate defect region, with articular cartilage in the upper third of the defect. Numbers at the bottom left and bottom right of each H&E image indicate histological scores for overall defect filling and implant degradation, respectively, that were given to that particular sample. All scale bars are 500 μm.

In terms of subchondral bone repair, all hydrogel formulations produced histological sections with mixtures of primarily compact bone and fibrous tissue in the subchondral region, with relatively small amounts of trabecular bone (Figure 4). Trabecular bone was visible as highly porous and pink-stained tissue on H&E sections, while compact bone was visible as densely packed, non-porous and pink-stained tissue [22]. Fibrous tissue, on the other hand, was represented by pockets of purple-stained tissue on H&E, with co-localized red staining of collagen on van Gieson’s Picrofuchsin [22] (Figures 4, S2). Overall, the subchondral regions of Cell+/Chondro+ implants (Figures 4D–4F) appeared to be more consistently filled with new tissue than Cell+/Chondro− implants (Figures 4A–4C) and Cell−/Chondro+ implants (Figures 4G–4I). Interestingly, 6 out of 12 (50%) implants in the Cell−/Chondro+ group showed a combination of hydrogel fragmentation and non-bone matrix development at the bottom of the subchondral region, as indicated by sparse and isolated staining of GAGs and collagen (Figures 4H, 4I). This phenomenon was also seen in 4 out of 13 (30.8%) implants of the Cell+/Chondro− group and 2 out of 14 (14.3%) implants of the Cell+/Chondro+ group.

Figure 4:

Example histological images of the subchondral bone for (A-C) Cell+/Chondro−, (D-F) Cell+/Chondro+, and (G-I) Cell−/Chondro+ implants at 12 weeks postoperatively. From left to right: H&E, Safranin O/Fast Green, and van Gieson’s Picrofuchsin. Numbers at the bottom left and bottom right of each H&E image indicate histological scores for bone filling and bone bonding, respectively, that were given to that particular sample. All scale bars are 500 μm.

The articular cartilage regenerated to varying degrees between experimental groups and individual implants, albeit with many samples exhibiting surface fissures (Figure 5). For the Cell+/Chondro− implants, the staining of GAGs by Safranin O (red), which represents more hyaline-like matrix composition, was visibly sparse in the surface tissue of most samples (Figure 5B). The Cell+/Chondro+ formulation, on the other hand, generally presented more robust GAG staining of both the neocartilage and adjacent cartilage (Figures 5E). Interestingly, 4 out of 14 (28.6%) implants in the Cell+/Chondro+ group also appeared to be developing a discernable osteochondral interface – with chondrocytes, GAGs, and collagen bridging the articular cartilage and subchondral bone (Figure S3) – compared to 1 out of 13 (7.7%) implants in the Cell+/Chondro− group and 0 out of 12 (0%) implants in the Cell−/Chondro+ group. Additional staining by Masson’s trichrome (Figure S4) revealed highly characteristic and dense blue/green collagen staining of articular cartilage within Cell+/Chondro+ implants, with a gradated transition to conventional bone staining patterns of red background and cytoplasm mixed with blue collagen in the subchondral layer (Figure S4B). Cell+/Chondro− implants, on the other hand, showed thin and uneven blue/green collagen staining that was interspersed with non-cartilage-specific staining in the articular cartilage layer (Figure S4A). Cell−/Chondro+ implants, similarly, showed highly disorganized collagen staining with significant pockets of partially degraded hydrogel and void space (Figure S4C).

Figure 5:

Example histological images of the articular cartilage for (A-C) Cell+/Chondro−, (D-F) Cell+/Chondro+, and (G-I) Cell−/Chondro+ implants at 12 weeks postoperatively. From left to right: H&E, Safranin O/Fast Green, and van Gieson’s Picrofuchsin. Numbers at the bottom left and bottom right of each Safranin O/Fast Green image indicate histological scores for the GAG/cell content of neocartilage and adjacent cartilage, respectively, that were given to that particular sample. All scale bars are 500 μm.

3.2. Histological Scoring and Quantification

All histological sections were quantitatively and independently scored using a well-established, modified O’Driscoll scoring system that has been previously utilized by our laboratory and collaborators (Table S1) [8,9]. After scoring, the Cell+/Chondro+ group demonstrated statistically higher overall defect filling compared to both the Cell+/Chondro− and Cell−/Chondro+ groups (Figure 6A). For categories of subchondral bone repair, the Cell+/Chondro+ formulation produced statistically greater bone filling compared to Cell+/Chondro− and Cell−/Chondro+, as well as a higher degree of bone bonding with the surrounding tissue compared to Cell+/Chondro− (Figure 6B). In terms of articular cartilage repair, the Cell+/Chondro+ group scored higher on cartilage thickness and GAG/cell content compared to the acellular control of Cell−/Chondro+ (Figure 6C). Importantly, the Cell+/Chondro+ group displayed statistically greater surface regularity of neocartilage, GAG/cell content of neocartilage, and GAG/cell content of adjacent cartilage compared to the non-chondrogenic control of Cell+/Chondro− (Figure 6C). The Cell+/Chondro− and Cell−/Chondro+ groups did not demonstrate statistical significance from each other in any scoring category.

Figure 6:

Histological scores for (A) overall defect evaluation, (B) subchondral bone evaluation, and (C) articular cartilage evaluation. All data are reported as means ± standard deviation. * indicates statistical significance compared to the Cell+/Chondro− group (chondrogenic peptide-free control), # indicates significance compared to the Cell−/Chondro+ group (acellular control) (α=0.05).

The histological categories that contained statistically significant differences were then further examined for their score distributions within each hydrogel group (Figure 7).

Figure 7:

Histological score distributions for statistically significant categories of (A) overall defect repair, (B-C) subchondral bone repair, and (D-G) articular cartilage repair. Distributions are shown as the percentage of sections in each hydrogel group that received a specific numerical score (0–3).

For the parameters of overall defect filling and bone filling, a greater proportion of the Cell+/Chondro+ group received scores of 2 or 3, representing 50–100% filling of the overall defect and subchondral layer, respectively, with newly formed tissue (Figure 7A–B). The Cell+/Chondro− and Cell−/Chondro+ groups, on the other hand, primarily received scores of 1, which indicates <50% filling by volume. In terms of bone bonding, the Cell+/Chondro+ group also received the highest number of score 3’s, indicating continuous bonding of newly formed bone tissue in the subchondral layer (Figure 7C).

In the categories of articular cartilage repair, the Cell+/Chondro+ formulation produced a similar distribution of cartilage thickness scores compared to Cell+/Chondro−, albeit with fewer scores of 0, which are defined as a complete lack of cartilage formation (Figure 7D). Surface regularity, while statistically higher on average for Cell+/Chondro+ compared to Cell+/Chondro−, was predominantly scored at 0 or 1 across all groups, indicating the presence of surface fissures of 25–100% thickness in the majority of samples (Figure 7E). The Cell+/Chondro+ formulation, however, produced a much greater number of score 2’s for the GAG/cell content of neocartilage relative to both other formulations, indicating normal cellularity and moderate GAG staining (Figure 7F). Assessment of the adjacent cartilage revealed a large number of score 1’s for all groups, which represents fewer cells and poor GAG staining in the regions adjacent to the defect site (Figure 7G). To further support the histological scoring of neocartilage GAG/cell content, we used image analysis to calculate the intensity of Safranin O staining. For each Safranin O/Fast Green image, a 1 mm × 3 mm region of interest denoting surface tissue within the defect margins was extracted and quantified for the intensity of red Safranin O staining (Figure S5) [19,20]. Image analysis revealed that the Cell+/Chondro+ group produced statistically greater intensities of Safranin O staining in newly formed surface tissue compared to the Cell+/Chondro− and Cell−/Chondro+ groups (Figure S5), corroborating the results of histological scoring (Figures 6C, 7F).

4. Discussion

Osteochondral defect repair remains a significant clinical and scientific challenge, with difficulties arising from the heterogeneity of tissue phenotypes in the osteochondral unit as well as the poor self-regenerative capacity of articular cartilage [10]. We evaluated the ability of bilayered, click biofunctionalized hydrogels to promote in vivo osteochondral repair after implantation in the medial femoral condyles of rabbits. This animal model, while less clinically representative than larger ovine and porcine models, offers greater compositional similarity to human osteochondral tissue than common murine models and has furthermore been validated by guidelines in ASTM F2451-05: Standard Guide for In Vivo Assessment of Implantable Devices Intended to Repair or Regenerate Articular Cartilage [16,23]. Importantly, full thickness osteochondral defects can be generated in this animal model to assess the degree of bone and cartilage tissue regeneration by implanted materials and tissue engineering constructs [8,9,16,17]. As prior studies have demonstrated highly immature tissue formation in the medial femoral condyle defect model after 6 weeks but quantifiable osteochondral regeneration after 12 weeks with successful treatment [8,9], we chose to investigate tissue regeneration at the 12 week timepoint. Nevertheless, further studies at earlier timepoints may help elucidate the developmental mechanisms promoted by click biofunctionalized biochemical cues.

Bilayered constructs that present distinct biochemical properties for the bone and cartilage have been well-established as superior methods for induction of osteochondral repair compared to monolayer scaffolds that present bone-specific, cartilage-specific, or mixed biochemical cues [3,7,24]. Traditionally, cartilage and bone tissue engineering constructs have utilized controlled release for the delivery of two or more tissue-specific growth factors to the osteochondral defect site [3,9]. However, controlled release often results in the uncontrolled diffusion of biomolecules, which reduces bioactivity and creates off-target effects that compromise the spatial localization of tissue development for heterogeneous tissues such as the osteochondral unit [24]. By covalently tethering our tissue-specific peptides to distinct hydrogel layers in a new application of Cp*Ru(cod)Cl-catalyzed click chemistry, we provided stratified biochemical cues for cartilage and bone repair. To our knowledge, this represents the first application of bioconjugation chemistry for the simultaneous repair of bone and cartilage tissue.

Furthermore, while prior strategies for osteochondral repair have involved bilayered constructs with gross compositional differences between each layer in mineral content, mechanical properties, proteins, or cell types, few constructs have recapitulated the specific biochemical cues responsible for the differential development of cartilage and bone [7–9]. In this study, osteochondral repair was achieved through the functionalization of a modular P(NIPAAm)-based construct with developmentally inspired peptide sequences – an N-cadherin mimicking sequence from early chondrogenic development and an osteonectin-derived sequence from mid-late stage osteogenic development [13,14]. While in vivo mineralization and bone development have been well-characterized in P(NIPAAm)-based constructs through microcomputed tomography, mechanical testing, and other techniques, the development of histologically well-scored articular cartilage remains an outstanding challenge and was thus the focus of our study [9,25,26]. Furthermore, given that P(NIPAAm), poly(ethylene glycol), and other synthetic polymer hydrogel systems are capable of supporting significant bone development even in the absence of cells or biochemical cues [27–29], we focused instead on evaluating articular cartilage repair using the O’Driscoll scoring system that predominantly evaluates parameters of cartilage development [8,9]. As evidenced by the higher histological scores produced by the Cell+/Chondro+ group relative to the Cell+/Chondro− group, NC peptide presentation promoted greater cellularity and GAG production within the neocartilage and adjacent cartilage and also promoted greater surface regularity of the neocartilage. The histological scoring of neocartilage GAG content was further corroborated by image analysis of Safranin O/Fast Green sections, which showed more intense Safranin O staining in the neocartilage of Cell+/Chondro+ samples. While Safranin O can also stain non-cartilage components such as GAGs within the secretory granules of mast cells [30], the intensity of Safranin O staining within articular cartilage, in particular, has been correlated with the quantity of cartilage-specific proteoglycans and GAGs [20,31]. The stronger GAG staining and scoring of Cell+/Chondro+ implants thus suggests more hyaline-like composition of the neocartilage ECM [2]. This combination of more hyaline-like matrix and greater surface regularity in Cell+/Chondro+ implants may contribute to more functional tissue, as articular cartilage function relies on tissue hydration by GAGs and the provision of a uniform surface for articulation [32]. Additionally, the Safranin O staining of tissue samples corroborates prior in vitro studies that demonstrate production of GAGs and expression of chondrogenic markers CDH2 and SOX9 by rabbit MSCs encapsulated within NC peptide-functionalized hydrogels [12]. While the tissue processing method utilized for our specimens is not compatible with RNA extraction or direct assessment of gene expression [8,33], more fundamental biological studies can help elucidate the specific genes that may be upregulated by the N-cadherin peptide sequence in vivo. Cell+/Chondro− and Cell−/Chondro+ groups, on the other hand, appear to have produced surface tissue with more fibrocartilage-like matrix composition, as exemplified by relatively lower GAG content scored in the Safranin O/Fast Green sections. Cell+/Chondro+ implants also exhibited much denser and well-organized collagen matrix in the articular cartilage when stained by Masson’s trichrome, suggesting higher quality tissue development, while Cell+/Chondro− and Cell−/Chondro+ implants showed sparse and non-uniform distributions of collagen. A major clinical challenge is the successful production of hyaline cartilage rather than poorly functional fibrocartilage [3], and our results support mechanistic studies which show that the mimicry of early developmental cues, such as the NC peptide sequence, promotes formation of hyaline-like cartilage [34,35]. While prior bilayered constructs have utilized controlled release to deliver cartilage- and bone-specific growth factors such as insulin-like growth factor-1 (IGF-1) and bone morphogenetic protein-2 (BMP-2), these strategies have produced minimal effects on the histological quality of cartilage formation under the O’Driscoll scoring system, with improvements seen mostly in chondrocyte clustering [9]. The absence of bioactive effects in prior cases has been attributed to the diffusion of solubly released growth factors [9,24], and our bioconjugation approach addresses these pitfalls by covalently tethering peptide-based cues to the construct, producing statistically significant improvements to multiple histological measures of cartilage repair that were not observed in prior studies. Overall, our click conjugation of the cartilage-specific NC peptide shows promise for improving the quality of newly regenerated cartilage.

Interestingly, inclusion of the NC peptide sequence enhanced histological measures of bone repair – specifically, the degree of bone filling and bonding with adjacent tissue. Prior studies have demonstrated that the stratified delivery of IGF-1 and BMP-2 can synergistically enhance subchondral bone development [9]. Our findings appear to demonstrate a similar phenomenon involving the GHK and NC peptides. Since these peptides are covalently tethered, it is possible that the release of secondary paracrine factors, the diffusion of the hydrogel sol fraction, or the degradative release of covalently tethered peptides mediates these synergistic effects. Multilayered P(NIPAAm)-based hydrogels have been shown to form continuous interfaces through the same hydrophobic interactions that govern thermal gelation [8,9,36], and these interfaces allow for the diffusion of paracrine signals and other small molecules between layers [9]. In considering the osteogenic layer, rabbit MSCs have been shown to produce bone-like matrix deposition as well as expression of the osteogenic markers Runx2 and OPN in response to GHK/PdBT when cultured in vitro in P(NIPAAm) hydrogels [12]. In literature, the selected NC peptide sequence (“HAVDI”) has also been posited to mimic homotypic cell-cell contacts that occur during MSC chondrogenesis [13] and alternatively, to promote heterotypic cell-cell contacts that occur during osteogenesis if the peptide is presented in tandem with osteogenic cues or cell-adhesive sequences such as “RGD” [37]. Further investigations may thus elucidate the in vivo effects of NC peptide presentation in the presence of additional chondrogenic and osteogenic biochemical cues. Since we observed more prominent formation of osteochondral interfaces within Cell+/Chondro+ implants, our results suggest that the localized co-presentation of GHK and NC peptide sequences at the interface between hydrogel layers may in fact help promote development of transitional tissue phenotypes The presentation of chondrogenic cues to superficial regions of the subchondral bone has also been shown to promote a smoother bone-to-cartilage transition within the osteochondral interface [38,39], and in our construct, the 1.5 mm height of the cartilage-specific layer may produce similar effects through the extension of approximately 0.5 mm of NC-functionalized hydrogel into the subchondral region. While histological staining by H&E, Safranin O/Fast Green, and van Gieson’s Picrofuchsin can demonstrably indicate formation of bone, cartilage, and fibrous tissue when scored using the O’Driscoll system [8,16], further staining of collagen subtypes I, II, and X will be necessary in future studies to characterize more complex tissue phenotypes at the osteochondral interface [39]. Aside from this, it should be noted that small amounts of fibrous tissue were present in the subchondral region of almost every implant, as shown by bone morphology scores for fibrous tissue that were statistically similar across all experimental groups. Furthermore, all implants produced primarily compact bone rather than native-like trabecular bone, which could produce anisotropic mechanical properties in the subchondral bone plate and impact its function [40]. Thus, further studies can optimize the subchondral hydrogel formulation through osteogenic peptide selection or the incorporation of bone-specific additives such as calcium phosphates. Furthermore, the optimization of hydrogel mechanical properties could provide a more physically mimetic environment for bone development, given the established importance of substrate stiffness to the osteogenic differentiation of MSCs [41]. While P(NIPAAm)-based hydrogels demonstrate a Young’s modulus within the range of 1–100 kPa, subchondral bone typically exhibits a much higher modulus around 100–300 MPa, and this gap could be addressed through the incorporation of high stiffness additives such as calcium phosphate cements or graphene oxide [42–44].

Regarding the effects of cell encapsulation, Cell+/Chondro+ hydrogels outperformed the acellular control group, Cell−/Chondro+, on various measures of overall defect healing, subchondral bone repair, and articular cartilage repair. However, Cell−/Chondro+ hydrogels achieved similar in vivo outcomes to Cell+/Chondro− hydrogels on every scoring parameter, indicating that the biological effects of the NC peptide on native, infiltrating cells may be sufficiently potent to overcome the absence of hydrogel-encapsulated cells. It should be noted, however, that the majority of Cell−/Chondro+ implants displayed a mixture of hydrogel fragmentation and development of non-bone ECM in the subchondral layer, as shown by staining of GAGs and collagen in these regions. This may indicate a more immature developmental stage of the subchondral bone [3], and it can be inferred the absence of encapsulated cells and their additional differential and chemotactic cues can effectively delay bone development [45]. Interestingly, the cellularity of neocartilage scored comparably between the Cell−/Chondro+ and Cell+/Chondro− groups, suggesting significant chemotaxis of endogenous cells in the presence of the NC peptide. Cellularity in the subchondral layer, while not scored in the O’Driscoll system, appeared comparable between all three groups, with mixtures of bone-associated and fibrous tissue-associated cells in almost all samples. While synthetic hydrogels often experience loss of encapsulated cells both in vitro and in vivo due to the absence of cell-adhesive biochemical signals [28,46], our prior investigations have shown that bioconjugation of the NC peptide effectively resists these effects by mimicking cell-cell contacts within the hydrogel [12,13]. Overall, our results show that the inclusion of encapsulated MSCs enhances osteochondral repair.

While Cell+/Chondro+ hydrogels produced improved histological scores for osteochondral tissue repair compared to Cell+/Chondro− and Cell−/Chondro+ hydrogels, it is worth noting that nearly all implants experienced partial implant degradation by 12 weeks, with average histological scores corresponding to less than 50% degradation for all groups. While partially degraded hydrogel was mostly localized to the bottom of subchondral regions, prior studies have indicated that, to the contrary, deeper areas of a defect typically experience faster implant degradation due to a greater degree of tissue contact [47]. Along these lines, the collapse of residual hydrogel material during implant degradation or histological processing could explain the localization of undegraded matter to the lower regions of defects. Additionally, the biological degradation of this PdBT-crosslinked hydrogel system depends on hydrolysis of PdBT, followed by resorption of uncrosslinked P(NIPAAm-co-GMA). However, P(NIPAAm-co-GMA), which has a molecular weight of approximately 10 kDa [48], may be poorly resorbable and effectively non-biodegradable in vivo due to its state of thermal gelation at physiological temperature [48,49]. Osteochondral tissue repair may thus be enhanced by improving the biodegradation of the hydrogel system, which may be achieved by incorporating co-monomers that modulate the thermal gelation state of P(NIPAAm-co-GMA) [25] to enable solubilization or by utilizing alternative polymers with hydrolytically or enzymatically degradable components. Future studies may also elucidate how more complex spatial patterning of GHK, NC, and other osteochondral peptides can be used to influence the development of articular cartilage, subchondral bone, and transitional tissue phenotypes within the osteochondral unit.

5. Conclusions

In this study, bilayered, tissue-specific hydrogels for osteochondral repair were fabricated by the click conjugation of developmentally inspired peptides – specifically, a chondrogenic NC peptide and an osteogenic GHK peptide – to stratified hydrogel layers. We found that NC peptide presentation and MSC encapsulation promoted greater histological scores of overall defect filling with newly formed, non-implant tissue and also enhanced several categories of cartilage repair including the GAG and cell content of articular cartilage, surface regularity, and cartilage thickness. Furthermore, presentation of the NC peptide to the articular cartilage and incorporation of cells throughout the entire construct enhanced histological scores for subchondral bone filling and the degree of bone bonding with adjacent tissue. Overall, our results establish the utility of this bioconjugated, stratified hydrogel system for repair of the osteochondral unit.

Statement of Significance.

Osteochondral repair requires mimicry of both cartilage- and bone-specific biochemical cues, which are highly distinct. While traditional constructs for osteochondral repair have mimicked gross compositional differences between the cartilage and bone in mineral content, mechanical properties, proteins, or cell types, few constructs have recapitulated the specific biochemical cues responsible for the differential development of cartilage and bone.

In this study, click biofunctionalized, bilayered hydrogels produced stratified presentation of developmentally inspired peptide sequences for chondrogenesis and osteogenesis. This work represents, to the authors’ knowledge, the first application of bioconjugation chemistry for the simultaneous repair of bone and cartilage tissue. The conjugation of tissue-specific peptide sequences successfully promoted development of both cartilage and bone tissues in vivo.