Regenerative Potential of Mandibular Condyle Cartilage and Bone Cells Compared to Costal Cartilage Cells When Seeded in Novel Gelatin Based Hydrogels

Abstract

The field of TMJ condyle regeneration is hampered by a limited understanding of the phenotype and regeneration potential of cells in mandibular condyle cartilage. It has been shown that chondrocytes derived from hyaline and costal cartilage exhibit a greater chondro-regenerative potential in-vitro than those from mandibular condylar cartilage. However, our recent in-vivo studies suggest that mandibular condyle cartilage cells do have the potential for cartilage regeneration in osteochondral defects, but that bone regeneration is inadequate. The objective of this study was to determine the regeneration potential of cartilage and bone cells from goat mandibular condyles in two different photocrosslinkable hydrogel systems, PGH and methacrylated gelatin, compared to the well-studied costal chondrocytes. PGH is composed of methacrylated poly(ethylene glycol), gelatin, and heparin. Histology, biochemistry and unconfined compression testing was performed after 4 weeks of culture. For bone derived cells, histology showed that PGH inhibited mineralization, while gelatin supported it. For chondrocytes, costal chondrocytes had robust glycosaminoglycan (GAG) deposition in both PGH and gelatin, and compression properties on par with native condylar cartilage in gelatin. However, they showed signs of hypertrophy in gelatin but not PGH. Conversely, mandibular condyle cartilage chondrocytes only had high GAG deposition in gelatin but not in PGH. These appeared to remain dormant in PGH. These results show that mandibular condyle cartilage cells do have innate regeneration potential but that they are more sensitive to hydrogel material than costal cartilage cells.

Introduction

It is estimated that 11% of individuals with temporomandibular joint (TMJ) disorders present with arthritic degeneration of the mandibular articular cartilage³⁵, a pathology which can lead to a cascade of problems resulting from functional and morphological changes in the joint⁴⁷. The arthritic process includes irreparable abrasion of articular cartilage and thickening and remodeling of underlying bone⁴⁷. There are currently no regenerative therapies to help the 1 million Americans with mandibular condylar cartilage pathology, only end-stage arthroplasty with a total joint prosthetic. While total joint replacements are an option for severe TMJ degeneration in certain patients^{18, 27}, regenerative therapies could offer an alternative of the relative young TMJ patient population, which are mainly women between puberty and menopause²⁸. Procedures used in repair of hyaline cartilage of other diarthroidal joints, such as microfracture⁹, mosaicplasty²⁴, ACI (autologous chondrocyte implantation)^{38, 41}, and MACI (matrix-induced ACI)¹⁶, are not accepted for regeneration of the TMJ condylar cartilage. The condylar cartilage anatomy is different than hyaline cartilage. For example, the articular surface of condylar cartilage is a thick fibrous layer compared to the thin superficial tangential layer of appendicular joint cartilage. A stable population of hypertrophic-like chondrocytes exists at the interface between cartilage and bone in condylar cartilage compared to the tide mark of other diarthrodial joints. Thus, a unique approach is needed for condylar cartilage regeneration. Furthermore, a reliable large-animal model that predicts clinical outcomes is needed to develop and translate therapies to the clinic.

We have developed the goat as a novel model to assess the efficacy of regenerative treatments for mandibular cartilage degeneration¹³. We have shown that the TMJ condylar cartilage has innate healing potential, with regeneration of both a superficial fibrous tissue and underlying cartilage. Full integration was observed between the regenerated and the native mandibular condylar cartilage, but the regenerated cartilaginous layer was not continuous across the entire defect. This suggests that cells from the surrounding tissues migrated into the scaffold to regenerate both fibrous and cartilaginous layers. However, no robust regeneration of the subchondral bone was observed, perhaps because osteoprogenitor migration and/or tissue formation is slower than that of fibrochondrocytes. Thus, the field of TMJ condyle regeneration is hampered by limited understanding of mandibular condyle cells that may have contributed to the regeneration we observed in our goat model¹³.

To further probe the regenerative capability of cells from the mandibular condyle cartilage, we have developed the combined use of two photocrosslinkable hydrogel biomaterials: PGH and methacrylated gelatin (GEL-MA). PGH is composed of methacrylated poly(ethylene glycol), gelatin, and heparin¹⁰. PGH was designed to mimic components of cartilage, with the gelatin mimicking the collagenous component (providing adhesive cues and eliminating anoikis) and the heparin mimicking the glycosaminoglycans. Poly(ethylene glycol) was added as a bioinert material to control hydrogel swelling, mechanical stiffness, and contraction by cells. GEL is used as a control because is a popular hydrogel material for bone and cartilage engineering and commonly used bioink. In our hands, PGH and GEL-MA were used to differentially control bone marrow stem cell differentiation toward both cartilage and bone¹⁰.

The objective of this study was to determine the regeneration potential and the primary cell phenotype stability of cells from mandibular condylar cartilage and bone, seeded within PGH and GEL-MA. We also compared results to cells from costal cartilage as a reference, because they are a popular source used in TMJ tissue engineering studies^{20, 21,25,42}. We extracted, isolated, and seeded primary bone and cartilage cells from goat mandibular condyles and primary cartilage cells from goat ribs into PGH and GEL-MA hydrogels and cultured for 4 weeks in osteogenic and chondrogenic mediums. We assessed the phenotype stability of each cell type under the different hydrogel biomaterials. Histology, biochemistry, and unconfined compression testing were used to assess the performance of these cell sources in both PGH and GEL-MA after 4 weeks of culture.

Materials and Methods

Material Preparation

PEG (molecular weight: ~4000), type B gelatin taken from bovine skin (molecular weight: ~45000), dichloromethane, triethlyamine, diethyl ether, molecular sieves (4Ǻ), methacrylic anhydride, and dialysis tubing (4000 mw cutoff) were obtained from Sigma-Aldrich. Heparin sodium salt (molecular weight: 15000), sodium hydroxide, and phosphate buffered saline (PBS) were from ThermoFisher Scientific (Hampton, NH).

The 4kDa MW PEG was selected this molecular weight for two reasons. First, the versatility of alternate formulations of the PGH where PEG is employed as the crosslinker. Short crosslinkers are desirable to minimize its influence on hydrogel porosity and stiffness. The 4 kDa was the shortest with facile purification under dialysis. Second, the 4kDa MW was consistent with prior works studying chondrogenesis with PEG hydrogels^{29, 43}. However, it is important to note that low MW PEGs alone yield hydrogels with low mesh size (e.g. 22×10–10 m, ¹⁵), with limited diffusivity to large molecules such as albumin, antibodies, etc.

Hydrogels were formed according to our previously described protocols¹⁰. For all materials, 1H NMR was used to verify the methacrylation (Bruker Avance III 300MHz). Lithium phenyl-2,4,6-trimethylbenzoyl phosphinate (LAP) was synthesized as in Majima et al. ³² and was used as a cross liking agent that reacts to ultraviolet (UV) light. All levels of methacrylations for each polymer were as previously described ¹⁰, but briefly, 93% of the terminal hydroxyl groups of PEG are methacrylated, 100% of the lysine residues in GEL are methacrylated, and 10.4% of the available saccharide residues in HEP are methacrylated (10.4 methacrylate groups per 100 disaccharide units).

Briefly, PEG was dissolved in dry dichloromethane and reacted with methacrylic anhydride and trimethylamine in the presence of freshly activated molecular sieves (4Ǻ). The mixture was left to react for 4 days at room temperature in the dark. The modified PEG (PEGDA) was then precipitated in diethyl ether, filtered, dried, and dialyzed against distilled water for 48 hours. The solution was then retrieved, frozen, and lyophilized.

Raw gelatin was dissolved in PBS and reacted with methacrylic anhydride at 50 °C for 1 hour. The newly modified gelatin (GEL-MA) was then neutralized to pH 7.4 with dilute sodium hydroxide and dialyzed against PBS for 48 hours. The material was retrieved, frozen, and lyolphilized.

Heparin was dissolved in distilled water and reacted with methacrylic anhydride at pH 8.5. Following the initial reaction, the mixture was allowed to react for 12 hours at 4°C, upon which the pH was brought back up to 8.4 due. Modified heparin (HEP-MA) was precipitated in ethanol, dialyzed against distilled water for 48 hours, frozen, and lyophilized.

Cell Isolation

Mandibular condylar chondrocytes (MCCs) were isolated from six different goat heads that were acquired from a local abattoir, from young goats at slaughter age (<1 year); all animals were slaughtered less than 24 hours before cell isolation. Cells were not pooled and all six different donors were used for each group. Mandibular condyles with the temporomandibular joint disc attached were extracted from the head using sterile tools. The mandibular condylar cartilage was dissected from the surface of the condyle, from both the fibrous and hyaline layers, using a sterile scalpel. The cartilage was minced, placed in sterile glass flasks with high glucose Dulbecco’s modified Eagle medium (DMEM) (ThermoFisher) supplemented with 2 mg/ml type II collagenase (Worthington), and digested overnight at 37 °C under mechanical agitation on an orbital shaker. Viable cells were plated on tissue culture plastic (TCP, ThermoFisher) and cultured for 3 days at 37 °C in high glucose DMEM with 10% (v/v) fetal bovine serum (FBS, Atlanta Biologicals / R&D Systems, MN) 1X non-essential amino acids (ThermoFisher), and 1X penicillin-streptomycin (ThermoFisher). Cells were subsequently trypsinized (0.25% w/v trypsin, ThermoFisher) and frozen in 10% (v/v) dimethyl sulfoxide (Sigma) in FBS, at which stage the cells were designated at P0. No substantial heterogeneity in cell fate and biosynthesis was observed from culturing cells from both the fibrous and cartilage layer, but more detail experiments need to be performed to determine actual differences.

Mandibular condylar bone derived cells (BDCs) were isolated for the condyles after cartilage removal, per established procedures in the literature to isolate osteoblasts via outgrowth of explants ³⁴. First, the trabecular bone of the mandibular condyle was drilled out and washed extensively and repeatedly with phosphate buffered saline (PBS) to remove marrow. The bone was minced using a sterile pin and ligature cutter, and then these bone chips were cultured on TCP with alpha MEM (αMEM) with GlutaMAX™ (ThermoFisher) and 2 mg/mL collagenase (Worthington) to allow for cells to migrate out of the bone. After 7 days of culture, the migrated and attached BDCs were trypsinized and frozen at P0.

Costal chondrocytes (CCs) were isolated from the ribs of young (<1 year) goats, which were obtained from a local abattoir within 4 hours of slaughter. The fibrocartilage was dissected from retrieved ribs, minced, and digested in overnight in DMEM with 2 mg/mL type II collagenase (Worthington) at 37 °C under mechanical agitation. Viable CCs were then frozen at P0 as done for MCCs, with plating on TCP, culture for 3 days, and trypsinization.

Scaffold Culture

All primary cells were passaged until P3 in their respective plating mediums (For MCCs and CCs, supplemented DMEM as above; For DCBs, αMEM with GlutaMAX™, 10% (v/v) FBS and 1X penicillin-streptomycin) and then frozen again prior to seeding. Before seeding, the cells were plated one more time for three days to recover in their respective mediums, trypsinized, and then seeded into the scaffolds.

Cell laden hydrogel scaffolds were manufactured using each cell types in both GEL-MA and PGH hydrogels as per Chen et al¹⁰. The two hydrogel formulations were tested: 8% (w/v) GEL-MA and 8% (w/v) PGH (2.4% PEGDA, 3.2% GEL-MA, and 2.4% HEP-MA). Cells were seeded at 30 million cells/mL in cylindrical scaffolds of 5 mm diameter and 2 mm height. Each scaffold was placed in a well of a 12-well plate with 2.5 mL media in each well and cultured for four weeks under gentle agitation on a mechanical rocker in the incubator at 37 °C and 5% CO₂. Media was replenished every other day. Hydrogels containing osteoblasts (OB) were cultured in osteogenic media consisting of αMEM with GlutaMAX™, 10% (v/v) FBS, 1X penicillin-streptomycin, 10 nM dexamethasone, 5 mM β-glycerolphosphate, and 50 µM L-ascorbic acid 2-phosphate (latter supplements from Sigma). Hydrogels containing costal chondrocytes (CC) and mandibular condylar chondrocytes (MCC) were cultured in chemically defined chondrogenic media consisting of high glucose DMEM supplemented with 1X ITS-G (ThermoFisher), 100 μM L-ascorbic acid 2-phosphate, 100 nM dexamethasone, 40 μg/mL L-proline, 10 ng/mL TGFβ-3, and 1X non-essential amino acids.

Histology

For each group, n = 3 scaffolds were processed for cyrosectioning after four weeks of culture as per Chen et al¹⁰. Briefly, scaffolds were fixed in neutral buffered formalin overnight, and then infiltrated with sucrose and O.C.T Compound (ThermoFisher) over three days, followed by snap freezing and embedding in O.C.T. The frozen sections were stained with hematoxylin & eosin to visualize cell content and distribution, Safranin O & Fast Green to visualize glycosaminoglycan (GAG) deposition, and silver nitrate via the Von Kossa procedure for mineral content.

Biochemistry

For each group, n = 3 scaffolds as fabricated (time-point 0 no culture controls) and n = 6 scaffolds after four weeks culture were used for biochemistry assays according to our established protocols ^{2, 4,1,3,5,7,8,21,20,31}. The samples were first digested in 0.1M papain at 65 °C overnight. Following digestion, the hydrogels were mechanically agitated to break up the scaffolds. Samples were then stored at −20 °C prior to the biochemical assays. DNA quantification was done using a picogreen assay (Molecular Probes / ThermoFisher). The total amount of glycosaminoglycans (GAG) content was measured using the Blyscan™ 1,9-dimethylmethlyene blue colorimetric assay with chondroitin-4-sulfate as the standard (Biocolor Ltd.). GAG content was divided by the DNA content to derive a normalized measure based on cell density.

Unconfined Compression

For each group, n = 3 scaffolds as fabricated (time-point 0 no culture controls) and n = 6 scaffolds were immediately removed from culture after 4 weeks culture and stored at 4 °C in 24 well plates filled with PBS. Scaffolds were tested within 24 hours. The diameter and thickness of the scaffolds were measured using calipers. An MTS Insight 1 kN testing apparatus with a 10 N load cell (MTS) was used to perform unconfined compression of the hydrogels. Scaffolds were placed in a bath with PBS. preloaded to 0.005 N, and height recorded. Then the scaffolds were subjected to 10 cycles of preconditioning at 9% strain/mm to 10% strain. After preconditioning, force versus displacement measures were made to a maximum of 10% applied strain (same rate) and over a 30 minute relaxation period. The peak stress is then reported as the force recorded at the end of the ramp to 10% strain, divided by the cross-sectional area. The dynamic modulus is calculated by the slope of the linear region of the stress-strain curve from 8%–10% strain during ramp. The modulus is termed dynamic since it is dependent on strain rate due to contributions from both the solid bulk material and fluid flow. The percent relaxation is the decrease in stress from peak to 30 minutes of relaxation (1-(Stress Relaxation/Stress Peak)).

Statistical Analysis

All data is represented as the average and standard deviation across sample replicates. For biochemical data (DNA and GAG) a general linear model (GLM) was used to detect statistical differences for scaffold type (GEL-MA, PGH), cell source (MCC, CC, OB), and time (0 Weeks, 4 Weeks). For compression properties (Peak Stress, Dynamic Modulus, Relaxation), since only one timepoint was measured, a GLM was performed for scaffold type (GEL-MA, PGH), cell source (MCC, CC, OB). The GLMs were followed by a Tukey’s post-hoc for multiple comparisons of each means to detect significant differences at an alpha of p = 0.05. Only statistical differences between GEL-MA and PGH for each cell type are reported, as determination of biomaterial effects on each cell population was the main objective of this study.

Results

Histology

At four weeks culture in osteogenic medium, mandibular BDCs seemed to produce mineralized matrix only in the GEL-MA hydrogel (Figure 1). Mineralized nodules seemed to be present throughout the GEL-MA hydrogel (Figure 1B), but none were evident in the PGH hydrogel (Figure 1D). The cellular distribution also appeared different between hydrogels, with no BDCs present at the periphery of GEL-MA hydrogels (Figure 1A), but BDCs uniformly distributed throughout the PGH hydrogel (Figure 1C).

Figure 1.

Unlike in gelatin hydrogels (top row), mandibular condyle osteoblasts in PGH hydrogels (bottom row) do not deposit mineralized matrix during 4 weeks culture in osteogenic media. Hematoxylin & eosin (H&E, left) and Von Kossa stains (right). N=3/group, Scale bar = 500 μm, and 125 μm for inset. (A) H&E of gelatin hydrogel. (B) Von Kossa of gelatin hydrogel with pink stain for the background and black for phosphates and other polyanions. (C) H&E of PGH hydrogel but low cellularity. (D) Von Kossa of PGH hydrogel with no observable mineral deposition.

At four weeks culture in chondrogenic medium, CCs seemed to have deposited GAGs in both GEL-MA and PGH hydrogels, but the GAG distribution appeared different between both hydrogels (Figure 2). In GEL-MA, GAG staining occurred throughout the scaffold around the periphery of the CCs (Figure 2B). Near the scaffold center, CC morphology was enlarged as evidenced by a large lacunae (Figure 2A). In addition, CCs deposited pockets of fibrous tissue as evidenced by eosin stain (Figure 2A). In PGH, GAG stain was more diffuse around the cells and inhomogeneous throughout the scaffold (Figure 2D). CCs in PGH did not seem to show enlarged lacunae and fibrous tissue staining (Figure 2C) as in GEL-MA.

Figure 2.

Costal chondrocytes in gelatin hydrogels (top row) and PGH hydrogel (bottom row) deposit GAGs during 4 weeks culture in chondrogenic media. Hematoxylin & eosin (H&E, left) and Safranin-O stains (right). N=3/group, Scale bar = 500 μm, and 125 μm for inset. (A) H&E of gelatin hydrogel with nuclear staining, and enlarged chondrocytes (indicating hypertrophy) in the center compared to the periphery. (B) Safranin-O of gelatin hydrogel with red indicating GAG deposition. (C) H&E of PGH hydrogel with cellular nuclear staining and no enlarged chondrocytes in the center. (D) Safranin-O of PGH hydrogel with GAG deposition.

At four weeks culture in chondrogenic medium, only MCCs in GEL-MA seemed to show evidence of GAG deposition. The GAG staining seemed to be more diffuse (less focal staining around lacunae, Figure 3B) compared to CC in GEL-MA (Figure 2B). Significant fibrous tissue staining seemed evident for MCCs in GEL-MA around the periphery where staining was low (Figure 3A). Few enlarged lacunae were evident compared to CCs in GEL-MA. In PGH, the MCCs did not seemed to deposit GAGs (Figure 3D) and appeared more fibroblast-like in morphology, but without staining for fibrous tissue (Figure 3C).

Figure 3.

Mandibular condyle chondrocytes deposit GAGs in gelatin hydrogels (top row) but not in PGH hydrogels (bottom row) during 4 weeks culture in chondrogenic media. Hematoxylin & eosin (H&E, left) and Safranin O stains (right). N=3/group, Scale bar = 500 μm, and 125 μm for inset. (A) H&E of gelatin hydrogel with nuclear staining, and no enlarged chondrocytes. (B) Safranin-O of gelatin hydrogel with GAG staining. (C) H&E of PGH hydrogel with nuclear staining, but with apparently lower cellularity. (D) Safranin-O of PGH hydrogel indicating no GAG deposition.

Biochemical Assay

Proliferation of cells depended on the hydrogel material type. After 4 weeks in culture, the DNA ratio from 4 weeks to 0 weeks for MCCs in GEL-MA was higher than in PGH (p<0.05) (Figure 4A). Conversely, the DNA ratio from 4 weeks to 0 weeks for CCs was higher in PGH than GEL-MA(p<0.05), but both were above 1. The DNA ratio from 4 weeks to 0 weeks for BDC on both hydrogels was well below 1, showing a decrease in DNA content from 0 to 4 weeks.

Figure 4.

CC proliferate and deposit GAGs in gelatin and PGH during 4-weeks of culture, but MCC do not proliferate or deposit GAGs in PGH. BDC viability decreased during hydrogel culture. (A) Ratio of DNA content at 4 weeks to 0 weeks for mandibular condylar chondrocytes (MCC), costal chondrocytes (CC), and mandibular condyle bone derived cells (DBC) in both gelatin and PGH hydrogels. Data is presented as average standard deviation. N=6 per group at each time-point. (B) Ratio of GAG content to total DNA of the hydrogel scaffolds for mandibular condylar chondrocytes (MCC), and costal chondrocytes (CC). N=6 per group at each time-point. A “*” represents a significant (p<0.05) difference between GEL and PGH for each cell type. Other significant differences were detected, but are not reported since no new ideas are elucidated.

GAG deposition ratio to DNA also depended on the hydrogel material type (Figure 4B). After 4 weeks in culture, the MCCs deposited significant GAG matrix in GEL-MA (p<0.05), but no detectable amounts were found in PGH. Conversely, CCs deposited significant GAG matrix in both GEL-MA and PGH (p<0.05), with no detectable statistical difference between hydrogel types. In terms of total GAG, CCs deposited about 650 µg of GAG per scaffold in both GEL-MA and PGH at 4 weeks, but PGH had a high background reading of about 70 µg of GAG per scaffold at 0 weeks, while GEL-MA had a background reading of about 8 µg of GAG per scaffold at 0 weeks.

Unconfined Compression

After 4 weeks in culture, the peak stress was higher in GEL-MA than in PGH for both MCC (16±6 kPa and 1.7±1.7 kPa, respectively) and CC (44±27 kPa and 3.6±1.3 kPa, respectively) (p<0.05) (Figure 5 A). The dynamic modulus was higher in GEL-MA than in PGH for both MCC (269±114 kPa and 24.6±11 kPa, respectively) and CC (775±426 kPa and 53±22 kPa, respectively) (p<0.05) (Figure 5 B). For the BDCs, no statistical differences in peak stress and dynamic modulus were detected between GEL-MA and PGH. In terms of relaxation behavior (Figure 5C), there were no detectable differences between hydrogel materials for all cell types at about 70% relaxation. It is important to note that the loads measured for the BDCs in both hydrogels were within the noise range of the load cell, thus the apparent reduced relaxation for the BDCs in both hydrogels is underestimated. Furthermore, the loads measured for all groups at 0 weeks was below the noise range of the load cells, and thus no comparisons were made for time.

Figure 5.

Measured mechanical properties of GEL and PGH hydrogels seeded with 30×10⁶/ml mandibular condylar chondrocytes (MCC), costal chondrocytes (CC), mandibular condyle bone derived cells (DBC) after 4 weeks of culture. N=6. Data is presented as average standard deviation. A “*” represents a significant difference (p < 0.05) between GEL and PGH for each cell type. (A) Peak stress and (B) Dynamic Modulus at 10% strain (0.09/s rate), and (C) Percent Relaxation after 30 minutes at 10% strain. Other significant differences were detected, but are not reported since no new ideas are elucidated.

Discussion

These results show that mandibular condyle chondrocytes (MCCs) do have innate regeneration potential. MCCs seeded in GEL-MA hydrogels, reached near native levels of compressive properties to mandibular condylar cartilage¹². Specifically, the peak stress and moduli of the native porcine mandibular condylar cartilage and that of the MCC in GEL-MA were both around 20 kPa and 250 kPa, respectively. Furthermore, the chondrocytes respond to the hydrogel material type, with MCCs more sensitive than costal chondrocytes (CCs). The potential mechanisms underlying this difference are not fully explored by this study, but some inferences can be drawn.

It is apparent that the hydrogel type impacts the phenotype stability of chondrocytes. At 4 weeks, CCs showed initial morphologic signs of hypertrophy only in GEL-MA while GAG measures were similar in both hydrogel types. This suggest that the PGH hydrogel enhances maintenance of chondrocyte phenotype, and is consistent with our prior work showing that PGH promotes chondrogenic differentiation of bone marrow derive stem cells (MSCs) while inhibiting their progression to hypertrophy¹⁰.

In addition, the hydrogel type impacts chondrogenesis by chondrocytes. It is likely that the MCCs respond uniquely to the PGH hydrogel beyond effects on chondrocyte redifferentiation. Serial expansion of chondrocytes in 2D planar culture conditions is known to yield dedifferentiation, which occurs faster in sub-confluent cultures^{23, 48}. 3D culture in hydrogels, commonly alginate beads, is employed to redifferentiate these but the transcriptome is not fully restored with high expansion^{11, 30}. To control for this, we expanded the cells at high density, but without supplement with basic FGF or TGF-1 to further enhance maintenance¹⁷. It is possible that the PGH hydrogel is less conducive to redifferentiation than the GEL-MA hydrogel, consistent with the lower mechanical measures for CCs at 4 weeks in PGH versus GEL-MA hydrogels. The MCCs may have undergone greater dedifferentiation than the CCs, yielding lower GAG and mechanical outcomes compared to CCs for each hydrogel type. However, the absence of MCC proliferation in PGH is not explained by dedifferentiation alone, as dedifferentiated chondrocyte proliferate. Furthermore, lower redifferentiation potential of PGH is inconsistent with our prior work showing that PGH promotes chondrogenic differentiation of MSCs under the same chondrogenic medium used here¹⁰. Instead, the MCCs appeared dormant in PGH hydrogel similar to an anergic state, decreasing both anabolic and catabolic responses¹⁴. While MCCs appear dormant in PGH, the lack of hypertrophy will likely lead to prevention of mineralization of the cartilage layer, which is the current caveat of cartilage regenerative therapies.

Prior works by others have shown that chondrocytes derived from hyaline and costal cartilage exhibit a greater chondro-regenerative potential in-vitro than those from mandibular condylar cartilage⁴⁵. However, our goat study has shown that condylar cartilage of the TMJ possesses innate regenerative potential in-vivo, unlike the hyaline cartilage of other diarthroidal joints¹³. However, MCCs are more sensitive to the biomaterial than CCs. This helps explain why CCs have not performed well in previous studies, since the scaffold material was not optimized or too many passages were used⁴⁵. Indeed, the results from this study suggest that native MCCs do have an innate regeneration potential, and they can be a part of regenerative therapy strategies. All these findings, when taken together, show that MCCs do have a regeneration potential, on par with costal cartilage, given a suitable hydrogel biomaterial for matrix deposition. The MCCs are more sensitive to these hydrogel materials than the CCs.

For BDCs, histology supported the theme of PGH inhibiting mineralization, as seen in our previous study¹⁰. However, osteoblasts obtained from the condyle did not survive prolonged in-vitro culture in both hydrogels as seen by the DNA quantification and low cellularity from H&E. Mineralization in vivo and in vitro is associated with high cell density and necrosis/apoptosis for both osteoblasts and chondrocytes^{26, 49}. The 5 mM concentration of -glycerophosphate used here does not appear to kill cells³³. However, the hydrogels may accumulate higher local concentration of inorganic phosphate that decrease viability, as cellular alkaline phosphatase acts on the -glycerophosphate, consistent with GEL-MA results. We have shown that PGH inhibits mineralization¹⁰, which is why no mineral was observed in PGH after 4 weeks of culture. However, this does not explain the decrease in cellularity of BDCs in PGH over time. Interestingly, BDCs in GEL-MA had similar mechanical properties to PGH, even though there were pockets of mineralization in GEL-MA. We hypothesize that the level mineralization was not sufficient matrix to impact the low strain (10%) unconfined compression properties, and thus other testing modalities will need to be employed to detect mechanical differences. Future in-vivo therapies of osteochondral regeneration for the condyle could use GEL-MA hydrogels, along with growth factors, to promote osteogenesis or bone regeneration. Furthermore, PGH hydrogels could be used in sites where mineralization is not desired.

There are several limitations to this work. First, several scaffolds were tested for mechanical properties after sitting in PBS for over 24 hours, which could impact the values measured due to swelling of the hydrogels. Second, the load cells used were of insufficient sensitivity to accurately measure the mechanical properties of the OB containing hydrogels (force measures were within the noise range of the load cells). Third, the PGH hydrogel has heparin, thus there is a GAG background reading that must be accounted for time 0, or low GAG content producing cells. Fourth, earlier time-points are required to determine the temporal profile of chondrogenesis and probe potential dedifferentiation of the cells, including characterization of the cells immediately prior to hydrogel encapsulation. As well, longer time-points are needed to determine if MCC laden hydrogels achieve similar properties as CC hydrogels. Fifth, empty scaffolds were not tested after 4 weeks of culture, but empty scaffolds were tested at 0 weeks with load levels below the accuracy of the mechanical testing load cell, which we inferred would not change over time, thus differences between materials and potential mineral contribution cannot be detected. Furthermore, to fully understand material effects on chondrogenic redifferentiation, chondrogenesis, and chondrocyte phenotype stability, additional anabolic (e.g. collagen type II), hypertrophy (e.g. collagen type X), and catabolic (e.g. matrix metalloproteinase 1, 13) markers are needed. Another limitation is that that differences in cell fate and matrix deposition within the heterogeneous MCC population were not examined. Future work will seek to determine if the material effects on MCCs were due to promotion of the growth of different subpopulations between the two hydrogel materials. It is possible that the MCCs dedifferentiate more than the CCs during serial expansion in monolayer culture, and therefore show less GAG in PGH. However, we discounted this possibility because the data suggest MCCs remain dormant in PGH. Specifically, the eosin staining for MCCs also shows a similar lower level in PGH than GEL. Further work will investigate the potential for dormancy, e.g. methyl green and pyronin Y staining, and stemness markers of MSCs in our hydrogels (e.g. Sox2, Oct4, Nanog). Future studies will explore the roles of permeability, stiffness, cell-material interactions, and sequestering of growth factors on hydrogel effects.

While there are tissue engineering studies that have been pivotal in getting closer to a TMJ condyle^{6, 19, 22, 36, 37, 39, 40, 44–46}, it is important to note that the main in-vivo focus has been the bone. It is imperative to regrow the cartilage layer, or mandibular condylar cartilage, to restore full function and long-term longevity of the implant. Furthermore, the mandibular condylar cartilage needs to be restored with low friction to ensure proper articulation with the TMJ disc. These two hydrogel materials may be used in composite or biphasic scaffolds to control cell phenotype and regenerate osteochondral tissue. PGH can be use used with stem cells that are differentiated towards a cartilage phenotype. This is based on the finding that as opposed to the GEL-MA groups, the costal chondrocytes seeded in PGH did not appear to hypertrophy, and mandibular condyle osteoblast and MSCs differentiated towards osteoblasts do not mineralize. We envision the use of PGH and GEL-MA manufactured as a bilayer scaffold to control the regeneration of both bone and the cartilage top. We also believe that the TMJ provides a unique environment for this technology, where infiltration of fibrocartilage progenitors can be manipulated towards a cartilage phenotype, and the bone marrow progenitors can be differentiated towards a bone phenotype. This is different than hyaline knee cartilage, where there is no source of cartilage progenitors, and bone marrow progenitors on their own only make fibrous tissue.