A MMP7-sensitive photoclickable biomimetic hydrogel for MSC encapsulation towards engineering human cartilage

Abstract

Cartilage tissue engineering strategies that use in situ forming degradable hydrogels for mesenchymal stem cell (MSC) delivery are promising for treating chondral defects. Hydrogels that recapitulate aspects of the native tissue have the potential to encourage chondrogenesis, permit cellular mediated degradation, and facilitate tissue growth. This study investigated photoclickable poly(ethylene glycol) (PEG) hydrogels, which were tailored to mimic the cartilage microenvironment by incorporating extracellular matrix analogs, chondroitin sulfate and RGD, and crosslinks sensitive to matrix metalloproteinase 7 (MMP7). Human MSCs were encapsulated in the hydrogel, cultured up to nine weeks, and assessed by mRNA expression, protein production and biochemical analysis. Chondrogenic genes, SOX9, ACAN, and COL2A1, significantly increased with culture time, and the ratios of COL2A1:COL10A1 and SOX9:RUNX2 reached values of ~20–100 by week six. The encapsulated MSCs degraded the hydrogel, which was nearly undetectable by week nine. There was substantial deposition of aggrecan and collagen II, which correlated with degradation of the hydrogel. Minimal collagen X was detectable, but collagen I was prevalent. After week one, extracellular matrix elaboration was accompanied by a ~two-fold increase in compressive modulus with culture time. The MMP7-sensitive cartilage mimetic hydrogel supported MSC chondrogenesis and promoted macroscopic neo-cartilaginous matrix elaboration representative of fibrocartilage.

Introduction

Cartilage repair is a significant clinical challenge due its limited self-healing capacity^(1),(2). Autologous chondrocyte implantation (ACI) is a clinically available cell-based therapy for the treatment of articular chondral defects. In ACI, autologous chondrocytes are harvested from a non-loading bearing region of the joint, expanded and injected into the defect covered by a membrane⁽³⁾. ACI has shown some success⁽⁴⁾; however long-term, randomized clinical trials indicate that ACI does not outperform microfracture⁽⁵⁾. Moreover, several limitations associated with ACI include donor site morbidity, limited number of cells that can be harvested, and de-differentiation of the chondrocytes during expansion^(6),(7). An alternative cell source for ACI is mesenchymal stem cells (MSCs), which eliminate donor site morbidity, have a higher proliferation capacity, and undergo chondrogenesis^(8)–(10). Although promising, MSCs introduce several challenges⁽¹¹⁾. Most notably, MSCs have an intrinsic differentiation profile to undergo hypertrophy during chondrogenic differentiation^(12)–(14), which is a precursor to endochondral ossification and osteoarthritis. In addition, robust cartilage regeneration by MSCs has not yet been demonstrated. Thus, a better understanding of the factors that control MSC chondrogenesis and cartilage regeneration is needed for their effective use in repairing focal chondral defects.

Delivering cells in situ within a three-dimensional (3D) matrix, such as a synthetic-based hydrogel, provides cells with structural support and as well creates the opportunity to introduce biochemical cues into the matrix to encourage differentiation. The extracellular matrix (ECM) of cartilage is comprised predominantly of collagen type II and aggrecan. Aggrecan is a proteoglycan made from sulfated glycosaminoglycans (GAGs) (e.g., chondroitin sulfate) and is linked by hyaluronic acid to create large aggrecan aggregates. On the contrary, hypertrophic cartilage is characterized by collagen type X and mineralization⁽¹⁵⁾. A number of studies have shown that incorporating different types of cartilage-derived ECM moieties into an otherwise synthetic hydrogel improves chondrogenesis of MSCs^{(10),(16),(17)}. Importantly, several of these studies have shown that creating a cartilage mimetic environment within the hydrogel suppresses the hypertrophic phenotype of MSCs ^(18)–(20).

Hydrogels comprised of cartilage’s GAGs, specifically hyaluronic acid and/or chondroitin sulfate, have been extensively studied as a vehicle to encapsulate MSCs and support chondrogenesis. Hyaluronic acid hydrogels^(21)–(26) and chondroitin sulfate hydrogels⁽²⁰⁾ support chondrogenesis with the former showing improvements over synthetic hydrogels, such as crosslinked poly(ethylene glycol) (PEG)⁽²⁷⁾. On the other hand, PEG hydrogels have served as a base chemistry to which cartilage ECM moieties are introduced. For example, the addition of covalently linked chondroitin sulfate into a PEG hydrogel led to enhanced chondrogenesis through cartilage-specific gene expression and matrix production when compared to PEG-only hydrogels^{(20),(28),(29)}. Interestingly, chondrogenesis was improved and hypertrophy was suppressed when sulfate groups were introduced into the backbone of hyaluronic acid,⁽³⁰⁾ suggesting an important role of sulfated GAGs.

Incorporating multiple ECM moieties into a hydrogel provides an opportunity to re-create the complexity of the native ECM. For example, when hyaluronic acid hydrogels were formed with either unmodified collagen type I, which introduces cell adhesion sites, or methacrylated chondroitin sulfate, cartilage matrix deposition was improved while hypertrophic-induced mineralization was reduced when compared to hyaluronic acid-alone hydrogels.^(18),(31) The use of small peptides over full proteins, such as collagen, enables a facile method to controllably incorporate cell adhesion functionalities. For example, PEG hydrogels that combine tethered chondroitin sulfate with the cell adhesion peptide, RGD, supported chondrogenesis of MSCs under culture conditions that did not readily induce chondrogenesis in PEG-only hydrogels⁽²⁹⁾. Alternatively, collagen mimetic hydrogels were created with peptides that bind cell-secreted GAGs^(32),(33), showing enhanced chondrogenic differentiation when compared to hydrogels without peptides^{(32),(34)–(36)}. Collectively, these and other studies support the idea that biomimetic hydrogels, which introduce cartilage ECM moieties and create environments that are more reminiscent of cartilage, improve MSC chondrogenesis.

Hydrogel degradation is essential to forming a macroscopic engineered tissue and ultimately regenerating cartilage. When cells are encapsulated in a hydrogel, the mesh size of the hydrogel dictates diffusion of newly secreted ECM molecules⁽³⁷⁾. One of the challenges is that the mesh size is smaller than that of most ECM molecules and notably that of collagen type II and the aggrecan aggregates⁽³⁸⁾. As a result, the hydrogel must reach its reverse gelation point prior to the formation of a macroscopic tissue^(38),(39). Moreover, when using MSCs, chondrogenesis must occur prior to substantial degradation and then the rate of degradation must reasonably match ECM synthesis. Hyaluronic acid hydrogels have been designed with hydrolytically labile linkers enabling degradation by water and enzymes^(40),(41). The incorporation of hydrolytically labile linkers led to improved collagen II and chondroitin sulfate deposition over enzyme-only degradable hydrogels, but collagen II remained limited to the pericellular space^(40),(41). Crosslinks sensitive to cell secreted matrix metalloproteinases (MMPs) have been introduced into synthetic-based hydrogels⁽⁴²⁾. In particular, MMP7 has been investigated because it is upregulated during chondrogenesis^(35),(43). When MSCs were encapsulated in a MMP7-sensitive PEG hydrogel without any additional ECM analogs, macroscopic deposition of collagen II was evident, but the presence of hypertrophic proteins was not evaluated⁽⁴³⁾. When a similar MMP7-sensitive crosslinker was incorporated into a collagen-based hydrogel that contained GAG binding peptides and the cell adhesion peptide RGD with MSCs, cartilaginous tissue deposition contained collagen II, but was limited to the pericellular space and was accompanied by collagen I and X, markers of fibrocartilage and hypertrophy, respectively^(35),(36). Although promising, macroscopic neocartilaginous tissue production by encapsulated MSCs has been limited and warrants further investigation.

The goal of this study was to develop an in situ forming enzyme-sensitive cartilage mimetic hydrogel for human MSC (hMSC) encapsulation and evaluate its potential for human cartilage tissue engineering. PEG hydrogels formed from the thiol:norbornene photoclick reaction were chosen as the base chemistry due to the mild photopolymerization conditions^(44),(45) and the ease with which thiolated ECM moieties and bis-cysteine crosslinks can be incorporated.^(46),(47) ECM moieties of chondroitin sulfate and RGD were chosen, as they have shown enhanced chondrogenesis of MSCs⁽²⁹⁾. A MMP7-sensitive peptide crosslinker was chosen given its promise in supporting macroscopic tissue deposition.⁽⁴³⁾ The hMSC-laden hydrogels were cultured for up to nine weeks and evaluated by mRNA expression, biochemical composition of cartilage ECM, hydrogel degradation, and mechanical properties. Overall, the results from this work indicate that a cartilage mimetic, MMP-7 sensitive PEG hydrogel formed from a thiol:norbornene photoclick reaction supports chondrogenesis of encapsulated MSCs, promotes formation of a macroscopic neo-cartilage tissue, and suppresses hypertrophy.

Materials and Methods

Macromer Synthesis

An 8-arm PEG amine (10kDa) reactant was used to synthesize the ‘ene’ monomer, 8-arm PEG norbornene. The PEG amine was dissolved in dimethylformamide (DMF) and reacted with 8x molar excess of 5-norbornene-2-carboxylic acid in the presence of 4 molar excess n,n-diisopropylethylamaine (DIEA) and 1-[Bis(dimethylamaino0methylene]-1H-1,2,3-triazolog[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) overnight at room temperature under argon. The 8-arm PEG norbornene product was recovered by precipitation in ethyl ether, purified via dialysis for 2–3 days, filtered (0.2μm), and lyophilized. Conjugation of norbornene to each arm of the 8-arm PEG was determined to be ~100% via ¹H NMR by comparing the area under the peak for the allylic hydrogen closest to the norbornene hydrocarbon group (δ=3.1–3.2ppm) to the peak of the PEG backbone methyl groups (δ=3.4–3.85ppm).

Thiolated chondroitin sulfate (ChS-SH) was synthesized as described by Shu et al. via a carbodiimide chemistry with thioacid dihydrazide⁽⁴⁸⁾. ChS (Chondroitin sulfate A, Sigma Aldrich) was dissolved in water and reacted with 2x molar excess dithiobis(propanoic dihydrazide) (DTP) and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDCI) overnight at an adjusted pH of 4.75 using 1.0M HCl. To stop the reaction, the pH was raised to 7 with the addition of 1.0M NaOH. A 6.5 molar excess of dithiothreitol (DTT) was added and reacted overnight at a pH of 8.5 to reduce the thiol groups of the DTP. The thiolated chondroitin sulfate product (ChS-SH) was purified and recovered by dialysis against 0.3mM HCl, centrifuged to remove any particulates, and the supernatant lyophilized. Conjugation of the thiol groups to the ChS was found to be ~15% (~7 thiol groups per molecule of ChS) via ¹HNMR by comparing the area under the peaks for the methylene groups of DTP (δ=2.5–2.6 and 2.6–2.8ppm) to the area under the peak of the methyl protons of the acetyl amine side chain of the chondroitin sulfate backbone (δ=1.8–2.0 ppm).

Human MSC (hMSC) Culture

Human mesenchymal stem cells (26 year old female) were purchased from Texas A&M and expanded in MSC expansion media consisting of 20% fetal bovine serum (FBS, Atlanta Biologicals), 50 U ml⁻¹ penicillin, 50 mg ml⁻¹ streptomycin, 20 mg ml⁻¹ gentamicin, and 5 ng ml⁻¹ basic fibroblast growth factor (bFGF) (Invitrogen) in low glucose Dulbecco’s modified Eagle media (DMEM, Invitrogen). The hMSCs were expanded under standard cell conditions (37°C, 5%C O₂) to 80% confluency and passaged at 3000 cells cm⁻². Passage 3 was used.

Cell-laden Hydrogel Preparation

Cartilage biomimetic degradable hydrogels were fabricated via photopolymerization of 9wt% PEG-norbornene (8-arm, 10kDa), 1wt% ChS-SH, 0.1mM CRGDS (Genscript), and 2.5wt% MMP7 sensitive peptide (CRDPLE-LRADRC)⁽⁴³⁾ (Genscript) in the presence of 0.05wt% photoinitiator Igracure 2959 (I2959) in phosphate buffer saline (PBS) under 352nm light at 5 mW cm⁻² for 8 minutes. The hMSCs were encapsulated at 50 million cells ml⁻¹ of filter-sterilized (0.2 um filter) monomer precursor solution and photopolymerized. This cell density was chosen based on prior studies using encapsulated chondrocytes⁽⁴⁷⁾ and encapsulated MSCs⁽⁴⁹⁾, which supported macroscopic tissue elaboration and deposition.

Cell-laden hydrogels (5mm diameter × 2.5mm height) were placed in 24-well tissue culture plates in 2 milliliters of chondrogenic differentiation media, adapted from Texas A&M stem cell distribution center^(50),(51), but with a lower concentration of TGF-β3 (1% ITS+ Premix, 100 nM dexamethasone, 2.5 ng ml⁻¹ TGF-β3, 50 μg ml⁻¹ l-ascorbic acid 2-phosphate, 50 U ml⁻¹ penicillin, 50 mg ml⁻¹ streptomycin, 250 μg ml⁻¹ Fungizone, and 10 mg ml⁻¹ gentamicin in high glucose Dulbecco’s modified Eagle media) which was replaced every other day. A lower concentration of TGF-β3 was used to assess the chondrogenic differentiation capabilities of the hydrogel environment ⁽²⁹⁾. The hMSC-laden hydrogels were cultured under standard cell conditions of 37°C with 5% CO₂ up to 9 weeks.

Evaluation of mRNA by qPCR

Prior to encapsulation and at prescribed culture times, hMSC-laden hydrogels (n=3/time point) were removed from culture, homogenized (TissueLyzer II, Qiagen) at 30Hz for 10 minutes in RNA lysis buffer, and RNA was extracted using a MicroElute Total RNA Kit (Omega) per manufacturer instruction. RNA was transcribed to cDNA using a high capacity reverse transcription kit (Applied Biosystems) per manufacturer instruction. Quantitative PCR (qPCR) of chondrogenic genes, SOX9, ACAN, and COL2A1 and hypertrophic genes, RUNX2 and COL10A1 was performed. Primers for each gene are given in Table 1 along with primer efficiency. The qPCR was performed using Fast SYBR Green Master Mix (Applied Biosystems) and a 7500 Fast Real-time PCR machine (Applied Biosystems). Relative gene expression were calculated using true efficiencies (based on the Pfaffle method⁽⁵²⁾) for the gene of interest (GOI) and the housekeeping gene (HKG) L30. Normalized gene expression (NE) was calculated following the Pfaffl method⁽⁵²⁾ using the true efficiencies from Table 1 by the following:

$$NE=\frac{{(E_{\mathit{GOI}})}^{\mathrm{\Delta }Ct(\mathit{calibrator}-\mathit{sample})}}{{(E_{\mathit{HKG}})}^{\mathrm{\Delta }Ct(\mathit{calibrator}-\mathit{sample})}}$$

where E is the PCR efficiency. ΔCt is the difference in Ct values between the calibrator (pre-encapsulated MSCs) and the sample (encapsulated at day 21).

Table 1.

Primer Sequences and Efficiency for qPCR Analysis

Gene	Forward Sequence	Reverse Sequence	Efficiency
L30	5′-TTAGCGGCTGCTGTTGGTT-3′	5′-TCCAGCGACTTTTTCGTCTTC-3′	94%
SOX9	5′-TGACCTATCCAAGCGCATTACCA-3′	5′-ATCATCCTCCACGCTTGCTCTGAA-3′	95%
ACAN	5′-AGTATCATCGTCCCAGAATCTAGCA-3′	5′-AATGCAGAGGTGGTTTCACTCA-3′	88%
COL2A1	5′-CAACACTGCCAACGTCCAGAT-3′	5′-TCTTGCAGTGGTAGGTGATGTTCT-3′	102%
RUNX2	5′-TTGGCCTGGTGGTGTCATTA-3′	5′-GAGTCCTTCTGTGGCATGCA-3′	98%
COL10A1	5′-TTTTGCTGCTAGTATCCTTGAACT-3′	5′-ACCTCTAGGGCCAGAAGGAC-3′	87%

Histological and immunohistochemical analysis

At prescribed culture times, hMSC-laden hydrogels (n=3/time point) were fixed in 4% paraformaldehyde, dehydrated, and paraffin embedded following a protocol using gradual concentration of ethanol to Neoclear to paraffin. Paraffin embedded hydrogels were then sectioned to 10μm using a microtome. Sections were stained with Safranin-O/Fast Green to visualize sulfated glycosaminoglycans (sGAGs) using light microscopy (Ziess Pascal, Olympus DP70). Immunohistochemistry was performed as follows. Sections were deparaffinized, rehydrated and pre-treated with appropriate enzyme treatments (hyaluronidase 200U/ml for aggrecan and collagen II, chondroitinase ABC (10mU) and keratinase I (4mU) for aggrecan, pepsin (280kU), protease (400U) and 0.25 %trypsin and EDTA for collagen X) for 1hr at 37 °C as well as antigen retrieval (collagen I, aggrecan, PEG). Sections were blocked with 1% BSA in PBS, washed twice, and permeabilized with 1% BSA 0.25% Triton-X-100 for all stains but PEG. Sections were treated with primary antibodies against aggrecan (1:5, US Biological A1059-53F), collagen type II (1:50, US Biological C7510-21C), collagen type X (1:50, Abcam ab49945), collagen type I (1:50, Abcam ab34710) and PEG (1:50, Anti-PEG 6.3 courtesy of Dr. Steve Roffler), followed by secondary antibodies with conjugated AlexaFluor 488 or 546 probes and counterstained with DAPI for nucleus detection. Positive controls using bovine meniscus and articular cartilage were used for primary antibodies for collagen I and collagen II. Previous studies have confirmed the use of the collagen X antibody^(53)–(57) and the authors have reported on the antibody previously⁽²⁹⁾. A laser scanning confocal microscope (Ziess LSM 5 Pascal) was used to acquire images at 400x magnification. Semiquantitative analysis of representative confocal images (n=4 images per hydrogel, n=3 hydrogels) was performed. Sections were stained simultaneously and the gain adjustment was set and maintained for all images to minimize variations in the intensity of the stain between images. The total intensity of the positively stained protein or PEG was normalized to the number cells in each image.

Biochemical Analysis

At prescribed culture times, hMSC-laden hydrogel constructs (n=3/time point) were flash frozen in liquid nitrogen and stored at -80 °C. Hydrogels were lyophilized, homogenized (TissueLyzer II, Qiagen) at 30Hz for 10 minutes, and digested with papain for 16 hours at 60°C. DNA content in the hydrogel constructs was measured using Hoechst 33258 (n=3). Sulfated glycosaminoglycan (sGAG) content was assessed using dimethyl methylene blue (DMMB) assay (n=3)⁽⁵⁸⁾.

Hydrogel Characterization

The compressive modulus of the cell-laden hydrogels was evaluated at prescribed culture times (n=3/time point). Hydrogels were compressed to 15% strain at a strain rate of 0.1mm min⁻¹ to obtain stress strain curves (MTS Synergie 100, 10N). The compressive modulus was determined by the slope tangential to the linear region of the stress-strain curves from 10 to 15% strain.

Statistical Analysis

Data are represented as the mean with standard deviation. A one-way analysis of variance (ANOVA) was performed with time as the factor followed by Tukey’s post-hoc analysis. P values are reported to indicate the level of significance with p<0.05 considered to be statistically significant.

Results

A photoclickable MMP7-sensitive cartilage mimetic PEG hydrogel was developed to encapsulate hMSCs (Figure 1). Chondrogenesis of hMSCs was evaluated by mRNA expression over the course of nine weeks in culture (Figure 2). Chondrogenic genes SOX9, a transcription factor, and ACAN and COL2A1, the main ECM molecules in cartilage, were evaluated. Time was a factor for ACAN (p<0.0001) and COL2A1 (p=0.0018), but was not for SOX9 expression. ACAN levels were maintained from week one to three and then increased (p<0.0001) by 500-fold at week six. From week six to week nine, ACAN levels decreased (p<0.0001) by 30-fold. COL2A1 levels exhibited a similar trend to that of ACAN with a 10,000-fold increase (p=0.0002) from week three to six followed by a 200-fold decrease from week six to nine.

Figure 1.

A schematic of the hydrogel precursors and the encapsulation of human mesenchymal stem cells (hMSCs) in a MMP7-sensitive cartilage mimetic hydrogel. Hydrogel precursors included 8-arm PEG functionalized with norbornene, MMP7-sensitive peptide flanked with cysteines on each end, thiolated chondroitin sulfate, and cysteine containing RGD sequence.

Figure 2.

Gene expression of MSCs encapsulated in an MMP7 degradable hydrogel normalized to gene expression of pre-encapsulated MSCs. The chondrogenic genes SOX9, ACAN, and COL2A1, and the hypertrophic genes RUNX2 and COL10A1 were evaluated with culture time. Data are represented as the mean with error bars as standard deviation (n=3).

Hypertrophic markers were evaluated by RUNX2, a transcription factor, and COL10A1. Time was a factor for RUNX2 (p=0.0045) and COL10A1 (p=0.043). Both genes were significantly up-regulated at week one when compared to the pre-encapsulated MSCs. RUNX2 levels decreased from week one to three and by week nine remained low (p=0.005) when compared to week one. COL10A1 levels increased (p=0.04) by ~100-fold at week six, but by week nine had returned to levels similar to that of week one. To further probe the phenotype of the differentiating MSCs, the ratios of COL2A1 to COL10A1 gene expression and the ratio of SOX9 to RUNX2 gene expression were evaluated. By week six, the COL2A1 to COL10A1 ratio increased (p=0.03) by ~100-fold when compared to week one or week three. SOX9 to RUNX2 ratio increased (p=0.00003) by ~100-fold from week one and increased (p=0.0004) by 10-fold from week three. By week nine, the ratio was lower (p=0.0008) than week six and not significantly different from week one.

The constructs were also evaluated by their DNA content and total sulfated GAG (sGAG) content as a function of culture time (Figure 3). For total DNA content per construct (Figure 3A), time was not a factor indicating that cell number remained constant in the hydrogels for the duration of the study. Time was a factor (p=0.0005) in the amount of sGAGs per construct (Figure 3B). The sGAG content was relatively constant from week one to six, with a slight mean decrease (p=0.06) by ~20%. By week nine, the sGAGs per construct were the lowest, decreasing (p=0.005) by ~40% from week one. The spatial distribution of sGAGs in the hydrogels was also evaluated (Figure 3C). At all time points, positive sGAG staining was present throughout the construct. However, it was not possible to differentiate between the chondroitin sulfate that was incorporated into the hyrogel and the newly synthesized sGAG.

Figure 3.

A. Total DNA content per construct is shown as a function of culture time. B. sGAGs per construct are shown as a function of culture time. Data are represented as the mean with standard deviation shown parenthetically or as error bars (n=3). C. Representative microscopy images of histological assessment by Safranin O/Fast Green, which stains sulfated glycosaminoglycans (sGAGs) red, scale bar is 50 μm.

The spatial distribution of cartilage-related proteins and the PEG polymer associated with the hydrogel were assessed by immunohistochemistry (Figure 4). The proteins included aggrecan and collagen II for hyaline cartilage, collagen X for hypertrophic cartilage, and collagen I for fibrocartilage. There was minimal aggrecan deposition detected at week 1, but its presence appeared to increase throughout the duration of the study and by week nine was present throughout the construct. There was some detectable staining for collagen II at week one. Similar to aggrecan, collagen II presence appeared to increase with culture time and was prevalent throughout the constructs by week nine. There was minimal collagen I detected at week one, but it also appeared to increase with culture time and was present throughout the construct by week nine. There was minimal collagen X detected at weeks one and three, but its presence became apparent by weeks six and nine. Its deposition, however, appeared to be localized pericellularly and not all of the cells stained positive. The spatial presence and disappearance of PEG was also evaluated with culture time. Positive staining for PEG was evident throughout the construct at week one, but its staining diminished over time with minimal staining by week nine.

Figure 4.

Representative immunohistochemical images for aggrecan (red), collagen II (green), collagen I (green), collagen X (green), and PEG (green) as a function of culture time. Images were acquired by confocal microscopy. Nuclei are stained blue. Scale bars are 20 μm.

The deposition of collagen II and correspondingly PEG disappearance was quantified from the immunohistochemistry images by measuring the intensity of positively stained collagen II and the intensity of positively stained PEG, each normalized to cell nuclei (Figure 5A and 5B). The mean intensity of collagen II increased from week one to three, although not significantly, and was maintained at week six. By week nine, collagen II intensity was the highest (p=0.001–0.032). The opposite trend was observed for PEG intensity. From weeks one to three, there was a decrease in mean PEG intensity per nuclei, although not significant, which continued to further decrease (p=0.01) by week six and remained similarly low at week nine. Additionally, PEG intensity per nuclei was plotted against collagen II intensity per nuclei (Figure 5C). A linear relationship was observed in the data from week one to six resulting in a Pearson correlation coefficient of -0.87. Interestingly, collagen intensity increased although the PEG intensity was already at its lowest, suggesting that the cells may continue to build their surrounding ECM even after the hydrogel has degraded.

Figure 5.

Semi-quantitative analysis of immunohistochemical images of (A) collagen II intensity per nuclei and (B) PEG intensity per nuclei as a function of culture time. (C) A scatter plot of PEG intensity per nuclei plotted against collagen II intensity per nuclei. A linear correlation between PEG and collagen II intensity is shown with a linear Pearson correlation coefficient of -0.87. The data points above a value of four for collagen II intensity per nuclei were not included in the linear correlation.

The construct modulus under compression was measured as a function of culture time (Figure 6). The initial modulus, which was measured after one day, was 18 (2) kPa. The modulus dropped (p<0.0001) by ~70% after one week of culture. By week six, the modulus increased (p=0.047) from week 1 to 10 (2) kPa and was maintained at week nine. Although the construct modulus increased from weeks one to nine, the final modulus was lower (p=0.006) than the initial modulus.

Figure 6.

Compressive modulus measurements initially (day 1) and with culture time of the cell-laden MMP7 degradable hydrogels. Data represent mean with standard deviation as error bars (n=3).

Discussion

In this study, we present a MMP7-sensitive cartilage mimetic hydrogel that supports MSC chondrogenesis and promotes neo-cartilaginous matrix production. The MMP7 crosslinker facilitated cell-mediated hydrogel degradation, which is necessary for macroscopic tissue elaboration. In accordance with hydrogel degradation, there was an increase in ECM deposition and a concomitant rise in compressive modulus. The neo-cartilage tissue that was formed by the encapsulated hMSCs was comprised of aggrecan and collagen type II, the main ECM molecules that make up cartilage, with minimal hypertrophy. However, the presence of collagen type I indicates fibrocartilage formation.

The encapsulated hMSCs readily degraded the MMP7-sensitive crosslinks within the hydrogel as indicated by the loss of PEG as a function of time. MSCs are known to secrete MMPs and as they differentiate, the types of MMPs released can change⁽⁵⁹⁾. In particular, MMP7 is highly expressed during early stages of chondrogenic differentiation, but is not known to be secreted by fully differentiated chondrocytes^(34),(43). In this study, a significant drop in compressive modulus was observed during the first week of culture. Bahney et al⁽⁴³⁾ reported that the MMP7 gene was undetectable in MSCs, but its expression spiked during the first week of chondrogenesis and then rose slowly and eventually leveled off by three weeks. Since the initial stage of chondrogenesis was not accompanied by significant ECM synthesis and deposition as shown by minimal staining for aggrecan and collagens, it is not surprising that the modulus drop was observed in the first week. This observation is consistent with other studies, which have reported a decrease in compressive modulus due to the unmatched rate of hydrogel degradation to ECM production^{(43),(60),(61)}. It is important to mention that other enzymes, including aggrecanase and MMPs 1, 2, and 13 have been reported to degrade this particular MMP7 sensitive peptide sequence^{(34),(62)–(64)}. Thus, it is possible that the hydrogel, especially during the first week, may have been degraded by enzymes (e.g., MMP2) that are known to be secreted by MSCs^(35),(43).

Over the nine weeks of culture, the MMP7-sensitive cartilage mimetic hydrogel supported MSC chondrogenesis and importantly the formation of a macroscopic neo-cartilaginous tissue that was composed of sGAGs, aggrecan and collagen II. Although it was not possible to distinguish between the chondroitin sulfate incorporated into the hydrogel and that which is deposited by the encapsulated cells, conclusions can be inferred from the histology, immunohistochemistry and biochemical analysis results. The dark red positive staining at week 1 is most likely from the chondroitin sulfate incorporated into the hydrogel, as there is little to no observable staining for aggrecan and this positive staining is characteristic of the hydrogel⁽⁶⁵⁾. Safranin O stained sections at week 3 and 6 represent the transitional phase between hydrogel and macroscopic tissue deposition and elaboration. As the hydrogel degrades and is replaced by macroscopic tissue, evident by the deposition and elaboration of aggrecan and collagen II, there is a change in the morphology. At week 9, regions of positively stained neo-tissue with nuclei are surrounded by weakly stained regions that are likely the remnants of the hydrogel and are consistent with the faint presence of PEG at this time point. The growth of the neo-tissue was accompanied by an increase in the modulus.

While other studies have encapsulated MSCs in MMP-sensitive hydrogels to investigate in vitro chondrogenesis and cartilage formation, the elaborated ECM is often limited with deposition primarily restricted to the pericellular space and little to no interconnectivity in ECM^{(35),(43),(66),(67)}. A few studies, however, have also reported an interconnected ECM within regions of the hydrogel similar to that reported in this study, however, at higher concentrations of the chondrogenic growth factor TGFβ⁽⁴³⁾. The neo-cartilaginous tissue that was formed in this study is attributed to a combination of the biochemical cues, chondroitin sulfate and RGD, and the degradable crosslinks^(29),(68). The incorporation of chondroitin sulfate can enhance ECM synthesis by the introduction of fixed negative charges, which, similar to cartilage, elevates the local osmolarity^(69),(70), and can bind and retain chondrogenic growth factors (e.g., TGFβ) within the hydrogel⁽⁷¹⁾. RGD, which is found in fibronectin, is known to enhance chondrogenesis of MSCs during early stages of differentiation and when incorporated at low concentrations or through degradable tethers enhances chondrogenesis^{(36),(72)–(74)}. It is worth noting that MMP7 may also contribute to ECM synthesis. MMP7 has been shown to support cartilage development by facilitating collagen II production through mobilization and release of bound growth factors (e.g., TGFβ) from the negative charged sGAGs^{(63),(75)–(78)}. In addition, sGAGs have been shown to bind MMP7 and promote its activation,⁽⁶³⁾ which can lead to localization of MMP7 activity in the pericellular region⁽⁶³⁾. While MMP activity is critical to hydrogel degradation, additional studies are needed to determine whether MMP7 contributes positively to ECM synthesis. Moreover, as noted above, other enzymes that are expressed in differentiated chondrocytes (e.g., ADAMTS4), but which are inactive during chondrogenesis may also have contributed to the continued degradation of the hydrogel long-term to help facilitate neo-tissue growth^(79),(80). Taken together, our results show that MSC chondrogenesis and cartilaginous matrix elaboration is supported by the MMP7-sensitive cartilage mimetic hydrogel.

The MMP7-sensitive cartilage mimetic hydrogel was able to reduce hypertrophy, evident by high expression of cartilage genes (SOX9 and COL2A1) relative to hypertrophic genes (RUNX2 and COL10A1), which was accompanied by minimal staining for collagen X relative to aggrecan and collagen II. RUNX2 expression is important during mesenchymal condensation ⁽⁸¹⁾ and has been reported to increase in studies inducing chondrogenesis in vitro with the addition of TGF-β3⁽⁸²⁾. Although there was an initial increase at week 1 in this study, the dominance of SOX9 to RUNX2 is believed to play a dominant role in the fate of MSCs⁽⁸³⁾. The discrepancy between COL10A1 and collagen X protein expression has been previously observed⁽⁸⁴⁾ and may in part be due to the differences in transcriptional and translational control of differentiating MSCs⁽⁸⁵⁾. Additionally, COL10A1 has been expressed by undifferentiated MSCs and chondrogenically differentiating MSCs when cultured with TGFβ^(86),(87). Although we do not know the exact mechanisms that may limit the hypertrophic phenotype in these hydrogels we hypothesize that the degradable nature of the hydrogel may play an important role. Our previous studies have reported that collagen X is prevalent in non-degradable PEG hydrogels containing no ECM analogs, as well as RGD and chondroitin sulfate at the same concentrations presented in this study ⁽²⁹⁾. The results presented herein are consistent with others which have reported reduced hypertrophy in MMP-sensitive hydrogels when compared to non-degradable hydrogels⁽⁶⁸⁾ suggesting that the elaboration of a cartilage ECM due to hydrogel degradation may help to minimize hypertrophy.

However, the presence of aggrecan and collagen II was accompanied by collagen I indicating the formation of fibrocartilage. Our results are consistent with previous studies which have reported similar findings in degradable hydrogels, where chondrogenically differentiating MSCs produce collagen I alongside collagen II⁽¹⁷⁾. Engineering fibrocartilage is important for regenerating certain cartilage tissues in the body, most notably the intervertebral disc. However, fibrocartilage is undesirable for treating cartilage defects in articulating joints. It is important to note that the hydrogels in this study were cultured in the absence of any mechanical stimulation, which has been suggested as an important factor in mediating fibrocartilage formation⁽⁸⁸⁾. Our previous work has shown that mechanical stimulation can inhibit collagen I while maintaining aggrecan and collagen II in degradable hydrogels containing fully differentiated chondrocytes⁽⁴⁴⁾. Thus, for articular cartilage tissue engineering, future work will need to investigate the effects of mechanical stimulation in the ability to control fibrocartilage development.

There are several limitations of this study. The activity of MMPs was not determined and thus we cannot confirm if MMP7 or other MMPs were critical to the positive outcome in ECM growth observed in this study. The compressive modulus of the neo-cartilaginous tissue was much lower than native cartilage tissue. Scaffold-less approaches based on condensed mesenchymal cell bodies have resulted in moduli similar to that of native cartilage.⁽⁸⁹⁾ However, the ability to deliver cells in an injectable hydrogel has benefits for clinical translation and compatibility with arthroscopic assisted surgery^(90),(91). The presence of mechanical stimulation⁽⁹²⁾ and confinement within a cartilage defect⁽⁹³⁾ may enhance the mechanical properties of the engineered cartilage within this hydrogel. Additionally, the study was limited to one donor and future work will need to evaluate more donors. Moreover, the initial hydrogel properties (e.g., crosslinking), which influences the degradation kinetics, may require optimization for different donors depending on the relative rates of MMP synthesis rates and ECM synthesis rates⁽³⁸⁾.

Conclusion

In this study, we developed a photoclickable cartilage mimetic PEG hydrogel with MMP7-senstive crosslinks that supported hMSC chondrogenesis and promoted macroscopic formation of human neo-cartilaginous tissue comprised of aggrecan and collagen II. Human MSCs were capable of degrading the MMP7-senstive peptide crosslinks resulting in a loss of PEG within the construct, which correlated closely with ECM growth and an increase in compressive modulus. Notably, the hydrogel inhibited hypertrophy, but led to collagen I deposition, indicating fibrocartilage. Overall, this hydrogel holds promise for cartilage tissue engineering using MSCs and warrants further investigation into improving the tissue mechanical properties and promoting articular cartilage.