IDG-SW3 Osteocyte Differentiation and Bone Extracellular Matrix Deposition Are Enhanced in a 3D Matrix Metalloproteinase-Sensitive Hydrogel

Abstract

Osteocytes reside within a heavily mineralized matrix making them difficult to study in vivo and to extract for studies in vitro. IDG-SW3 cells are capable of producing mineralized collagen matrix and transitioning from osteoblasts to mature osteocytes, thus offering an alternative to study osteoblast to late osteocyte differentiation in vitro. The goal for this work was to develop a 3D degradable hydrogel to support IDG-SW3 differentiation and deposition of bone ECM. In 2D, the genes Mmp2 and Mmp13 increased during IDG-SW3 differentiation and were used as targets to create a MMP-sensitive poly(ethylene glycol) hydrogel containing the peptide crosslink GCGPLG-LWARCG and RGD to promote cell attachment. IDG-SW3 differentiation in the MMP-sensitive hydrogels improved over non-degradable hydrogels and standard 2D culture. Alkaline phosphatase activity at day 14 was higher, Dmp1 and Phex were 8.1-fold and 3.8-fold higher, respectively, and DMP1 protein expression was more pronounced in the MMP-sensitive hydrogels compared to non-degradable hydrogels. Cell-encapsulation density (cells/ml precursor) influenced formation of dendrite-like cellular process and mineral and collagen deposition with 80×10⁶ performing better than 2×10⁶ or 20×10⁶, while connexin 43 was not affected by cell density. The cell density effects were more pronounced in the MMP-sensitive hydrogels over non-degradable hydrogels. This study identified that high cell encapsulation density and a hydrogel susceptible to cell-mediated degradation enhanced mineralized collagen matrix and osteocyte differentiation. Overall, a promising hydrogel is presented that supports IDG-SW3 cell maturation from osteoblasts to osteocytes in 3D.

INTRODUCTION

In bone, osteocytes are the most abundant cells and are vital in regulating homeostasis and healing.^1,2 Osteocytes sense mechanical cues that transfer through the bone lacuno-canalicular network (LCN) during load-bearing activities, and in turn serve as orchestrators of bone remodeling by regulating bone formation and resorption in osteoblasts and osteoclasts.^3–5 Osteocytes arise from osteoblasts; when osteoblasts become encased in newly formed osteoid, the entrapped cells form dendrites that extend towards the mineralizing front or vascular space.⁶ The osteocyte dendritic processes facilitate cellular communication throughout bony tissue by creating a highly interconnected, three-dimensional (3D) cellular network.^6–8 While our understanding of osteocyte biology and function is improving, our understanding is far from complete. Novel in vivo studies have shed some light on osteocyte mechanisms (e.g.,^9–13), but such studies are difficult and costly. In vitro models that support the mature osteocyte phenotype and mimic in vivo conditions are thus needed to study osteocyte function in a controlled environment.¹⁴

The location of osteocytes within the mineralized matrix of bone makes primary osteocyte isolation and culture particularly challenging.¹⁵ As an alternative, osteocyte-related cell lines enable in vitro study of osteocytes, such as studying the mechanisms by which osteocytes mineralize bone¹⁶ or respond to fluid shear stress.¹⁷ MLO-Y4 cells exhibit osteocyte properties including a dendritic morphology concomitant with high expression of osteocalcin, low expression of the osteoblast marker, alkaline phosphatase, but limited ability to mineralize.¹⁸ MLO-A5 cells exhibit a post-osteoblast phenotype with high alkaline phosphatase and rapid mineralization. However, both these osteocyte-related cell lines lack markers of mature osteocytes, such as Dmp1 (encoding for dentine matrix protein 1) and Phex (encoding for phosphate-regulating neutral endopeptidase, X-linked).¹⁹

More recently, the murine IDG-SW3 cell line was developed to study the transition from osteoblasts to osteocytes.²⁰ IDG-SW3 cells cultured on collagen-coated plates showed up-regulation of mature osteocyte markers, including Dmp1 and Phex, and produced a mineralized matrix.²⁰ Recent 2D studies have investigated IDG-SW3 expression of mature osteocyte markers in response to various physical, biochemical, and mechanical cues (e.g.,^21–25). These types of studies were previously not possible due to the limitations of the other osteocyte-related cell lines. However, to date, IDG-SW3 osteocyte maturation has not yet been achieved and fully characterized in a 3D environment.

Three-dimensional, rather than two-dimensional, environments more accurately resemble the native tissue networking of cells.^26,27 Various studies have shown that osteocyte differentiation is significantly improved when cultured in 3D.^28–31 For example, primary human osteoblasts cultured on a biphasic calcium phosphate porous scaffold showed up-regulation of several osteocyte-related genes within two weeks of culture including Dmp1 and Phex.²⁸ The expression of these genes was either low or undetectable in 2D cultures. Pre-osteoblastic MC3T3-E1 cells were shown to differentiate towards an osteocyte phenotype in vitro, but only after migrating into a 3D collagen hydrogel.³¹ These studies indicate that osteocytes will maintain their in vivo phenotype, but only if cultured in 3D. To relate findings from in vitro studies to the in vivo environment, 3D culture systems are important. Some studies have established 3D cultures for osteoblasts and early osteocytes using scaffolds such as collagen type I hydrogels or microbeads (e.g.^13,32–34), but there are few 3D models that support mature osteocyte differentiation using late osteocyte cell lines, and, to date, there are no known 3D tunable hydrogel systems that support IDG-SW3 culture.

The goal of this study is to develop and characterize a hydrogel system that supports the transition from osteoblasts to osteocytes in a 3D environment. A poly(ethylene glycol) (PEG) hydrogel based on the thiol-norbornene photoclick chemistry was chosen for its cytocompatibility, tunability, and ease with which peptides are incorporated.³⁵ Specifically, cell adhesion peptides and matrix metalloproteinase (MMP)-sensitive crosslinks were introduced into the hydrogel. Given the tight polymer mesh of the hydrogel, cell-mediated degradation is important for the extension of cellular processes, such as dendrites, and formation of cell-cell contacts.³⁶ Moreover, studies have reported that incorporating MMP-sensitive crosslinks into a hydrogel enhanced osteogenesis,^37,38 supported ECM deposition,^39,40 and was biocompatible in a subcutaneous implant.⁴⁰ To develop this hydrogel system, this study focused on three aims. The first aim was to identify a MMP-sensitive hydrogel, which would support IDG-SW3 cell-mediated degradation. To this end, several MMPs were investigated for their known roles in bone development and homeostasis and include MMP 2,9,13 and 14.⁴¹ The second aim was to assess IDG-SW3 differentiation within the MMP-sensitive degradable hydrogel and compare to a non-degradable hydrogel and 2D culture on tissue culture polystyrene (TCPS). The final aim examined the effect of cell encapsulation density on cellular morphology and bone ECM deposition. In particular, cell density has been identified as a key regulator of osteoblast-to-osteocyte differentiation^30,42,43 and will influence cell-cell contacts as well as the amount of total ECM deposited. Overall, this study identified a promising hydrogel for IDG-SW3 culture in 3D that can be degraded by the encapsulated IDG-SW3 cells and which supports bone extracellular matrix deposition and mature osteocyte differentiation.

EXPERIMENTAL SECTION

2D Cell Culture.

IDG-SW3 cells (Kerafast, Inc., Boston, MA) are engineered with GFP expression under the control of the dentin matrix acidic phosphoprotein 1 (DMP1) promoter. The cells are also engineered to proliferate and remain immortal under interferon-gamma (INF-γ) at 33°C. When INF-γ is removed from the culture, and the cells are cultured in osteogenic media at 37°C, they revert to their in vivo phenotype of late osteoblasts and are capable of differentiating to osteocytes²⁰. IDG-SW3 cells were expanded in culture medium consisting of Modified Essential Medium (MEM) α containing L-glutamine and deoxyribonucleosides (Gibco) supplemented with 10% FBS (Atlanta Biologicals), 30 U/ml recombinant mouse interferon-gamma (INF-γ) (Peprotech), and penicillin/streptomycin/ amphotericin B (PSF, Invitrogen). The cells were seeded at a density of 3000 cells/cm² on to T-225 tissue culture polystyrene flasks that were coated with rat-tail collagen type-1 (Sigma-Aldrich). The cells were expanded in a regulated incubator at 33°C with 5% CO₂. Medium was replaced thrice weekly. Cells were treated with trypsin and passaged at ~80–90% confluency.

To induce osteogenesis, cells were plated on collagen type I-coated surfaces at 80,000 cells/cm² and cultured under osteogenic conditions and at 37°C. This process involved the removal of IFN-γ from the culture medium, and the introduction of 50 mg/mL of ascorbic acid and 4 mM β-glycerophosphate to the culture medium.

3D Hydrogel Formation and Culture.

An 8-arm PEG with terminal amines (20,000 g/mol; JenKem Technology USA, Plano, TX) was functionalized with norbornenes by reacting 5-norbornene-2-carboxylic acid (Sigma-Aldrich, St. Louis, MO) with 2-(1H-7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl uranium hexafluorophosphate methanaminium (Chem-Impex International, Inc., Wool Dale, IL), and N,N-diisopropylethylamine (Chem-Impex) in dimethylformamide (DMF)/ dichloromethane (DCM) (Sigma-Aldrich). The reaction was carried out overnight at room temperature under inert atmosphere. The product was precipitated in diethyl ether (Sigma-Aldrich), filtered, dialyzed, and lyophilized. The extent of conjugation of norbornene to each arm of the 8-arm PEG-amine was determined to be 92% using ¹H NMR by comparing the protons across the carbon-carbon double bond in the norbornene to the methylene protons in PEG. Two different crosslinkers were used: a non-degradable PEG-dithiol (1000 g/mol; Sigma-Aldrich) and an MMP-sensitive peptide with the amino acid sequence GCGPLG-LWARCG (GenScript) containing two cysteines. CRGDS (GenScript), which contains one cysteine for tethering, was also introduced as the cell adhesion peptide.

For cell encapsulation studies, MMP-sensitive, degradable hydrogels were formed from a precursor solution consisting of 6.5% (w/w) 8-arm PEG-norbornene, MMP-sensitive peptide at 0.65:1 thiol:ene ratio, 2 mM CRDGS, and 0.05% (w/w) photoinitiator, 1-(4-(2- Hydroxyethoxy)-phenyl)-2-hydroxy-2-methyl-1-propane-1-one (I2959; BASF, Tarrytown, NY), in phosphate-buffered saline (PBS). Non-degradable hydrogels were formed from a precursor solution consisting of 7% (w/w) 8-arm PEG-norbornene, PEG dithiol (1000 g/mol) at 0.5:1 thiol:ene, 2 mM CRGDS, and 0.05% (w/w) photoinitiator in PBS. The non-degradable hydrogel formulation was chosen to achieve a compressive modulus that was close to the MMP-sensitive hydrogel, with an acellular compressive modulus of 10 (1) kPa and 8 (1) kPa for the non-degradable and MMP-sensitive hydrogel formulations, respectively. Murine IDG-SW3 cells were encapsulated into the MMP-sensitive and non-degradable hydrogels at cell concentrations of (1) low, 2×10⁶ cells/mL, (2) medium, 20×10⁶ cells/mL, and (3) high, 80×10⁶ cells/mL. All hydrogels were photopolymerized for 10 minutes with 352 nm light at 5–10 mW/cm² in molds that were 3 mm in diameter and 3 mm in height and immediately cultured in osteogenic differentiation media.

Acellular Hydrogel Degradation.

Acellular, MMP-sensitive, degradable hydrogels were formed from a precursor solution consisting of 4.5% (w/w) 8-arm PEG-norbornene, MMP-sensitive peptide at 0.65:1 thiol:ene ratio, and 0.05% I2959 in PBS using the same molds and photopolymerization method as described above. A lower concentration of 8-arm PEG-norbornene was chosen to shorten the time of degradation. A solution of 10 nM MMP-13 or 11.1 nM MMP-2 (Calbiochem, EMD Millipore) was prepared in a buffer consisting of 50mM (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (Corning), 10mM CaCl₂ (Thermo Fisher Scientific), 20% glycerol (Mallinckrodt Chemicals), and 0.005% BRIJ-35 (Alfa Aesar). Hydrogels were placed initially in this buffer but without MMP and allowed to swell to equilibrium at 37°C for 48 hours. After equilibrium swelling, the buffer solution was replaced with new buffer solution containing either MMP-13 or MMP-2 and incubated at 37 °C. Every 24 hours for 3 days, the enzyme solution was refreshed, and samples were collected, weighed, tested for compressive modulus, and then lyophilized.

GFP Expression.

GFP-expressing DMP-1 was monitored over culture time up to 30 days by fluorescence microscopy (EVOS FL Imaging System; Life Technologies) and image acquisition using a camera (Sony ICX445 monochrome CCD camera) in the 2D cultures. At day 28, intact hydrogels were imaged by confocal microscopy (Zeiss LSM 5 Pascal system using a Zeiss Axiovert microscope). In the 2D culture experiment, the corresponding bright field images were also acquired.

Gene Expression.

In the 2D experiment for the first study of the paper (Figure 1), samples were collected at days 1, 4, 14, 21 and 35, and the cells were lysed directly in the well plates with TRK lysis buffer (Omega) and stored at −80°C. In the 2D experiment for the second study of the paper (Figure 3), samples were collected at days 1, 7, 14, and 28. In the 3D culture experiments, samples were collected at days 1, 7, 14, and 28, placed in TRK lysis buffer (Omega) and stored at −80°C. Samples were disrupted using a tissue lyser (Qiagen), and RNA was isolated using E.Z.N.A. microelute kit (Omega) per the manufacturer instructions. The amount of pure RNA was quantified using a Nanodrop instrument (ND-1000, Thermo Scientific) with A260/280 greater than 1.90. Purified RNA was reverse transcribed into cDNA using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) per the manufacturer instructions. Samples were analyzed by qPCR with Fast SYBR Green Master Mix (Applied Biosystems) on a 7500 Fast system (Applied Biosystems). Custom primers were designed using Primer Express 3.0 software (Applied Biosystems) and were evaluated for efficiency. Primer sequences, efficiencies, and accession numbers are listed in Table 1. Data are presented as relative expression (RE) to the housekeeping gene L32 given by

$$\text{Relative\hspace{0.17em}expression\hspace{0.17em}}\left(\text{RE}\right)=E_{HKG}^{C_{t\left(HKG\right)}}/E_{HKG}^{C_{t\left(GOI\right)}}$$

where E is the true primer efficiency, HKG is the housekeeping gene, GOI is the gene of interest, and C_t is the cycle number where the sample crosses the threshold. Data are also presented as normalized expression (NE) given by

$$NE={\left(E_{GOI}\right)}^{\Delta C_{t,GOI}\left(\mathit{\text{control}}-\mathit{\text{sample}}\right)}/{\left(E_{REF}\right)}^{\Delta C_{t,HKG}\left(\mathit{\text{control}}-\mathit{\text{sample}}\right)}$$

where the gene expression of the sample is normalized to a control, as described in the text.

Figure 1.

A) Schematic of 2D experiment. B) Brightfield fluorescence images and fluorescent microscopy images depicting the expression of DMP-1 (green) in IDG-SW3 cells cultured on collagen type-1 coated tissue culture polystyrene on days 7, 10, 14, 21, and 30 of osteogenic differentiation. Scale bar is 150 μm. C) Relative and D) Normalized gene expression of Mmp2, Mmp9, Mmp13, and Mmp14 genes expressed in IDG-SW3 cells cultured on collagen type-1 coated tissue culture polystyrene on days 1, 4, 14, 21, and 35 of osteogenic differentiation. For RE, letter symbols denote significance between MMPs at each time point. For normalized expression, symbols represent significance from day 1 for a given MMP and are shown in the legend for each MMP.

Figure 3.

A) Schematic showing study outline. The high encapsulation density (80×10⁶ cells/ml) was used in this study. Abbreviations are as follows: ALP = Alkaline Phosphatase Assay, qPCR = quantitative Polymerase Chain Reaction, GFP = DMP-1 Green Fluorescent Protein positive imaging. B) ALP Activity as a function of culture condition. * is difference from day 1, # is difference from day 14. one symbol is p<0.05, two symbols is p<0.01, three symbols is p<0.001. C&D) Normalized gene expression for (C) Dmp1 and (D) Phex expressed in IDG-SW3 cells at days 1, 7, 14 and 28 in osteogenic differentiation media and encapsulated at high cell seeding density in MMP-deg (dotted) hydrogels, Non-deg (dashed) hydrogels and collagen type-1 coated tissue-culture polystyrene (solid). Normalized expression is the relative expression that is normalized to the Day 0 trypsin-treated cells prior to adding in differentiation media and encapsulating in hydrogels. Symbols represent significance from day 1 for a given culture condition; one symbol p<0.05; two symbols p<0.01; three symbols p<0.001; symbols are shown in legend for each condition. Letter symbols denote significance between culture conditions at Day 28 (a’s are statistically different from b’s; b’s are not statistically different from each other). E) Confocal images depicting live IDG-SW3 cells at day 28 expressing DMP-1 GFP (green) cultured in osteogenic differentiation media and encapsulated at high cell seeding density in non-degradable PEG (left) and MMP-sensitive (right) hydrogels. Scale bar is 150μm.

Table 1.

Primer sequences, accession numbers, and efficiencies for each gene used in this study

Gene	Forward Primer	Reverse Primer	Accession #	Efficiency
L32	CCATCTGTTTTACGGCATCATG	TGAACTTCTTGGTCCTCTTTTTGA	NM_172086	1.83
Dmp1	GCTTCTCTGAGATCCCTCTTCG	GCGATTCCTCTACCCTCTCT	NM_016779.2	1.97
Phex	CAACGTTCCGCGGTCAATAC	GTGTTGCTTGGTCCAGCTTC	NM_011077.2	1.88
Mmp2	AACGGTCGGGAATACAGCAG	GTAAACAAGGCTTCATGGGGG	NM_008610.3	1.91
Mmp9	GCCGACTTTTGTGGTCTTCC	TACAAGTATGCCTCTGCCAGC	NM_013599.4	1.96
Mmp13	GGAGCCCTGATGTTTCCCAT	GTCTTCATCGCCTGGACCATA	NM_008607.2	1.88
Mmp14	GCCCTCTGTCCCAGATAAGC	ACCATCGCTCCTTGAAGACA	NM_008608.4	2.00

Alkaline Phosphatase Activity.

Hydrogel specimens were removed from culture at 1, 14 and 28 days and rinsed in PBS for 1 hour, lysed in deionized water (diH₂O), frozen in liquid nitrogen, and stored at −80°C. Samples were disrupted using a tissue lyser (Qiagen), then subjected to freeze-thaw-sonicate cycles to lyse the cells. DNA content was measured using a Quant-iT PicoGreen dsDNA Assay Kit (Thermo Fisher Scientific) by fluorescence with an excitation at 485 nm and emission at 520 nm according manufacturer specifications. Alkaline phosphatase activity was determined by measuring the number of moles of p-nitrophenol phosphate catalyzed to p-nitrophenol, which was measured by absorbance at 450 nm using a spectrophotometer.

Cell Viability and Morphology in 3D Hydrogels.

The viability and morphology of encapsulated IDG-SW3 cells were assessed using a live/dead assay based on Calcein AM (Corning), which stains the cytosol of live cells, and ethidium homodimer (Corning), which enters the nucleus of compromised cells and stains DNA. Hydrogels were incubated with 4 μM Calcein AM and 2 μM ethidium homodimer for 10 minutes. The GFP expression could not be distinguished from the cytosolic stain of Calcein, but the latter will stain all live cells. Qualitative assessment of dead cells was not affected by GFP expression. Intact hydrogels were imaged by confocal microscopy (Zeiss LSM 5 Pascal system using a Zeiss Axiovert microscope).

Immunohistochemistry and Histology.

Hydrogel specimen were removed from culture at day 28, fixed immediately in 4% paraformaldehyde at 4°C for 24 hours, dehydrated, paraffin embedded, and sectioned to 10 μm thickness. For connexin 43 staining, sections were treated with an antigen retrieval (Retrivagen A, BD Biosciences), blocked for 2 hours at room temperature with 10% normal goat serum, 2% bovine serum albumin, and 0.25% triton X-100 in PBS. Sections were then treated with the primary antibody to connexin 43 (ab11370, Abcam) at 1:1000 overnight at 4°C. An Alexa Fluor 546 goat anti-rabbit secondary (4 μg/mL; Life Technologies) was applied for 1 hour at room temperature. Sections were also stained with von Kossa and counterstained with nuclear red, according to standard protocols and imaged using light microscopy (Zeiss Axioskop 40) with either a 20x or 40x objective and a digital camera (Diagnostic Instruments, MN 14.2 Color Mosaic) using SPOT Software v. 4.6. For collagen type I, sections were enzyme treated with pepsin (280 kU), protease (400 U) and 0.25% trypsin and EDTA for 1 hour at 37°C. Sections were treated for antigen retrieval as described above and then blocked with 1% BSA and permeabilized with 1% BSA 0.25% Triton-X-100. Sections were treated with primary anti-collagen type I (Abcam ab34710) at 1:50 overnight at 4°C followed by treatment with Alexa Fluor 546 goat anti-rabbit secondary (4 μg/mL; Life Technologies) for 1 hour at room temperature. Both connexin 43 and collagen type I stained sections were counterstained with 4,6-diamidino-2-phenylindole (DAPI) for nucleus detection and then imaged by confocal microscopy (Nikon A1R Confocal System) with a 40x or 60x objective and NIS-Elements Confocal software.

The immunohistochemistry microscopy images for connexin 43 and collagen type I were quantitated using Image J. For each image (n = 3 per group for connexin 43, n = 9 per group for collagen I), the blue (DAPI) channel and red (connexin 43 or collagen I) channel were converted to binary using the same threshold value for each respective channel. The number of nuclei per image was counted using Analyze Particles with the size range of 10 – infinity μm². The percent total area of red connexin 43 or collagen I staining were calculated using Analyze Particles with the size range of 0 – infinity μm².

Biochemical Assays.

Hydrogel specimens were removed from culture at 28 days and subsequently rinsed in PBS for 1 hour, lysed in diH₂O, frozen in liquid nitrogen, and stored at −80°C. Samples were disrupted using a tissue lyser (Qiagen) and subjected to freeze-thaw-sonicate cycles. A known amount of sample was measured for calcium content using the Calcium (CPC) Liquicolor® Assay (Stanbio) according to manufacturer specifications, with absorbance measured at 540 nm. Hydroxyproline, an amino acid in high abundance in collagen, was measured. A known amount of sample was hydrolyzed in 6 M hydrochloric acid for 3 hours at 120°C, prior to being reacted with 4-(dimethylamino)benzaldehyde in a chloramine-T/oxidation buffer for 90 min. Absorbance was measured at 570 nm using a spectrophotometer and normalized to a baseline at 620 nm.

Mechanical Testing of 3D Hydrogel Constructs.

Mechanical testing was conducted on a Mechanical Testing System Insight II (MTS; Eden Prairie, MN) in unconfined compression acellular hydrogels were tested in their swollen state with a 2 N load cell. Testing was performed with the top platen out-of-contact with the hydrogel and then a constant displacement rate of 1.2 mm/min until an approximate 50% strain (based on the initial height of the hydrogel) was reached. The point of contact was determined using MATLAB. The compressive modulus was calculated from the slope of the linear region of the engineering stress-strain curve between 10 and 15% strain.

For mechanical testing of cellular hydrogels, the specimens were removed from culture at days 14, 21 and 28 for mechanical assessment. A 2 mN pre-load was used to establish consistent contact with each sample. Hydrogels were then compressed at a constant displacement rate of 0.5 mm/min to 15% strain. The compressive modulus was calculated as described above between 10 and 15% strain.

Statistical Analysis.

Statistical analysis was performed using Real Statistics add-in for Excel. ANOVAs were performed with α = 0.05 where factors included time, culture condition, and/or cell encapsulation density as described in the results. If significant interactions between factors were observed, follow-up one-way ANOVAs were performed holding each factor constant. Post-hoc analysis was performed using Tukey’s HSD and α = 0.05. In comparisons that were limited to two groups, a Student’s t-test was performed assuming independent samples and equal variances. P-values from the one-way ANOVAs are provided to indicate the level of significance with p < 0.05 being considered statistically significant. Data were confirmed to follow a normal distribution and exhibit homogeneous variance. All numerical results are presented as mean with standard deviation listed parenthetically in the text. Graphical results are presented as mean with standard deviation as error bars. The sample size was n = 3 unless otherwise noted.

RESULTS

IDG-SW3 Differentiation In 2D Culture and MMP Gene Expression Profiles.

IDG-SW3 cells were cultured in 2D on collagen-coated TCPS and analyzed over the course of a 35-day study using the study design in Figure 1A. Differentiation was confirmed by GFP expression indicating activity by the DMP-1 promoter. Brightfield images and corresponding fluorescent microscopy images are shown over time (Figure 1B). The cells were confluent by day 7 and remained confluent through the experiment. There was minimal expression of GFP detected on days 7 and 10. Qualitatively by day 14, GFP was noticeable in a small number of cells. By day 21, GFP was more prevalent and remained consistent at day 30.

Relative expression for genes encoding matrix degradative enzymes Mmp2, Mmp9, Mmp13, and Mmp14 was assessed as a function of culture time (Figure 1C). Comparing MMP type at each time-point revealed a dynamic pattern of Mmp expression. At day 1, relative expression for Mmp2 was highest (p < 0.001) when compared to the other three MMPs, which continued through day 21. Mmp9 and Mmp13 relative expression levels were not different from each other and were the lowest of the MMPs at day 1. By day 35, Mmp2 and Mmp13 relative expressions were not significantly different from each other, but both were higher (p < 0.001) than Mmp9 and Mmp14.

Relative expression for each MMP type was normalized to its expression level prior to the initiation of differentiation (i.e., day 1) and shown as a function of culture time (Figure 1D). Normalized expression levels increased from day 1 to 4 by 1.7-fold (p = 0.005) for Mmp2, 1.9-fold (p < 0.001) for Mmp9, and 1.8-fold (p = 0.025) for Mmp14. By day 35, Mmp2 and Mmp14 normalized expression levels returned to levels that were not significantly different from their day 1 values. Mmp9 levels remained elevated, but no further increase was observed after day 4. On the contrary, Mmp13 normalized expression exhibited a distinctly different expression profile than the other MMPs. Mmp13 expression significantly increased at each time point resulting in a 200-fold increase (p < 0.001) from day 1 to day 35. Collectively, these results demonstrate that Mmp2 expression is high in the IDG-SW3 cells prior to osteocyte differentiation and is maintained during differentiation. In contrast, Mmp13 expression is low prior to differentiation and increases with osteocyte differentiation.

Acellular Characterization of an MMP-Sensitive PEG Hydrogel.

MMP-sensitive PEG hydrogels were formed via the photoclick reaction between thiols and norbornene functionalized monomers (Figure 2A, B). We chose the peptide sequence GPLG-LWAR for the crosslinker because it was previously identified for its specificity for MMP-13.⁴⁴ Degradation occurs via an enzyme-catalyzed hydrolysis of the peptide bond. Acellular hydrogels exposed to exogenous MMP-2 and MMP-13 was assessed by the compressive modulus (Figure 2C). The compressive modulus decreased (p < 0.001) in the first day for both MMP-2 and MMP-13 exogenous treatment. The hydrogels exposed to MMP-2 degraded more rapidly by day 2, as shown by the lower (p = 0.025) modulus as compared to MMP-13. However, by day 3, hydrogels exposed to MMP-2 and MMP-13 were not significantly different.

Figure 2.

A) Schematic of hydrogel formation and cell encapsulation by a photoclickable thiol:norbornene reaction. Non-degradable hydrogels were formed by reacting monomers of eight-arm PEG-norbornene (PEG-NB) with crosslinkers of PEG-dithiol. MMP-sensitive hydrogels were formed by reacting PEG-NB with a bis-cysteine peptide sequence shown. Both non-degradable and MMP-sensitive hydrogels had the same molar concentration of RGDS and weight percent of I2959 Photoinitiator. B) The PEG-NB weight percent and thiol:ene molar ratio used to form the hydrogels. C) Compressive modulus of acellular MMP-sensitive hydrogels when exposed to exogenous MMP treatment for 3 days. Symbols represent significance from day 0 and are shown in legend for each MMP treatment; one symbol p<0.05; two symbols p<0.01; three symbols p<0.001. Letter symbols denote significance between MMP treatment each day.

IDG-SW3 Differentiation in 3D MMP-Sensitive Hydrogels.

IDG-SW3 cells were cultured under three conditions: (1) 2D collagen-coated TCPS, (2) 3D non-degradable hydrogels, and (3) 3D MMP-sensitive hydrogels. The study design is shown in Figure 3A. The high cell encapsulation density (80×10⁶ cells/mL) was used for this study. The effect of culture condition on osteocyte differentiation over time was assessed by alkaline phosphatase activity, expression of the osteocytic genes, Dmp1 and Phex, and GFP expression associated with DMP1 (Figure 3B–E).

Alkaline phosphatase (ALP) activity was assessed as a function of time and culture condition (Figure 3B). Considering time as the only factor, ALP activity in 2D culture increased (p < 0.001) from day 1 to 14 and then decreased (p = 0.025) by day 28 where mean ALP levels were greater (p = 0.06) than levels from day 1. Similar findings were observed for ALP activity from the 3D culture in non-degradable hydrogels, but day 28 levels were not different from day 1. In MMP-sensitive hydrogels, ALP activity also reached its highest level at day 14 when compared to day 1 (p < 0.001) and day 28 (p < 0.009) with levels remaining higher (p = 0.03) than day 1 at day 28. At day 14, ALP activity was 2.4- (p = 0.01) and 3.7- (p = 0.003) fold higher in the MMP-sensitive hydrogels compared to the 2D culture and 3D culture in non-degradable hydrogels, respectively. There were no differences between the 2D culture and the 3D culture in non-degradable hydrogels at any time point. Although the magnitude of ALP activity was lower at day 28, similar trends were observed between culture conditions as in day 14.

Dmp1 expression is normalized to that of the pre-encapsulated cells (Figure 3C). Considering time as the only factor, Dmp1 levels in 2D culture increased (p = 0.001) with time resulting in a 2200-fold change from day 1 to 28. In 3D culture non-degradable hydrogels, Dmp1 levels increased (p < 0.001) from day 1 to 14 and then decreased (p < 0.001) from day 14 to 28 to levels that were not significantly different from day 1. For the MMP-sensitive hydrogels, Dmp1 levels increased (p < 0.001) by 2500-fold from day 1 to 28. Considering culture condition as the only factor, Dmp1 normalized levels at day 28 were 2.9- (p = 0.016) and 8.1- (p = 0.004) fold higher in the MMP-sensitive hydrogel compared to the 2D culture and 3D culture non- degradable hydrogels, respectively.

Phex expression is also shown as normalized to that of the pre-encapsulated cells (Figure 3D). Considering time as the only factor, Phex levels in 2D culture increased (p = 0.015) from day 1 to 14, but by day 28 were not significantly different from day 1. In 3D culture non-degradable hydrogels, Phex levels increased (p = 0.005) from day 1 to 14 and then decreased (p = 0.049) from day 14 to 28 to levels that were not significantly different from day 1. For the MMP-sensitive hydrogels, Phex levels increased (p < 0.001) by 140-fold from day 1 to 28. Considering culture condition as the only factor, Phex levels at day 28 were 3.2- (p = 0.005) and 3.8- (p = 0.004) fold higher in the MMP-sensitive hydrogel compared to the 2D culture and 3D culture non-degradable hydrogels, respectively.

Activity of the DMP-1 promoter was confirmed by the presence of GFP in the cells in 3D culture in the non-degradable and the MMP-sensitive hydrogels (Figure 3E). At day 28, a few cells expressed GFP in the non-degradable hydrogel, while many cells expressed GFP in the MMP-sensitive hydrogel. Qualitatively, there appeared to be more GFP+ cells in the MMP-sensitive hydrogel, but quantification was not performed. Collectively, these results along with gene expression and ALP activity demonstrate that osteocyte differentiation is improved in the MMP-sensitive hydrogels over 3D culture in non-degradable hydrogels and the 2D cultures.

The Effect of IDG-SW3 Cell Encapsulation Density and Hydrogel Type on Osteocyte Morphology and Bone ECM Deposition

After confirming that IDG-SW3 cells underwent osteocyte differentiation in the MMP-sensitive hydrogels, we next investigated if cell density within the hydrogels influenced differentiation and bone ECM deposition (study design in Figure 4). The effect of cell encapsulation density and hydrogel type (MMP-sensitive vs non-degradable) was assessed for cellular morphology that included formation of dendrite-like processes and connexin 43 staining, matrix mineralization, collagen deposition, and compressive modulus.

Figure 4.

Schematic showing study outline. Abbreviations are as follows: qPCR = quantitative Polymerase Chain Reaction, MTS = Mechanical Testing System compressive modulus, IHC = immunohistochemistry, Biochemical assays = calcium and hydroxyproline, Histology = von Kossa.

Cellular morphology was evaluated as a function of time (day 1, 14, and 28) and cell encapsulation density (low, medium and high) in non-degradable and MMP-sensitive hydrogels (Figure 5). At day 1, cells retained a round morphology after encapsulation with increased cell numbers evident at increased cell encapsulation densities. With time, cells retained the round morphology at the low cell encapsulation density in the non-degradable hydrogels. On the contrary, in the medium and high cell encapsulation density, there was evidence of short dendritic-like cellular processes developing by day 14 in the non-degradable hydrogels. In the MMP-sensitive hydrogels, dendritic-like cellular processes extending into the hydrogel were observable across all cell densities by day 14. There were few dead cells observed in the confocal microscopy images for the low and medium cell encapsulation density. However, there were dead cells present in the high cell encapsulation at day 1 after encapsulation (see also Figure S1, which shows the dead cells only). Thus, the apparent decrease in live cells for the high cell encapsulation density could be attributed to the initial cell death after encapsulation. It is worth noting that there was no apparent overlap between green cells and those staining as dead cells. Thus, any GFP+ cells that overlap with the calcein stained cells are expected to be limited to live cells.

Figure 5.

Confocal images depicting live, stained by Calcein AM (green), and dead, stained by ethidium homodimer (red), IDG-SW3 cells at days 1, 14 and 28 in osteogenic differentiation media encapsulated at low, medium, and high cell seeding densities in MMP-sensitive and non-degradable PEG hydrogels. Note that any GFP staining would overlap. Scale bar is 150 μm.

The presence of connexin 43 was investigated by immunohistochemistry as a function of cell encapsulation density in the non-degradable and MMP-sensitive hydrogels (Figure 6). Positive staining for connexin 43 was evident by punctate staining near the encapsulated cells (Figure 6A). There was little to minimal staining in the non-degradable hydrogels at low encapsulation cell density. With medium and high cell encapsulation density there was some evidence of connexin 43 in the non-degradable hydrogels. On the contrary, there was greater positive staining for connexin 43 in the MMP-sensitive hydrogels for all three cell encapsulation densities. The images were quantified by amount of connexin 43 per cell as measured by positive staining area (%) per nuclei for each condition (Figure 6B). Connexin 43 on a per cell basis was affected by hydrogel type (p = 0.0076), indicating an overall higher amount of connexin 43 in the MMP-sensitive hydrogels over the non-degradable gels; although pair-wise comparisons were not statistically significant. Cell encapsulation density did not affect the amount of connexin 43 staining on a per cell basis.

Figure 6.

A) Representative microscopy images of IDG-SW3 cells at day 28, for connexin 43 (white, indicated by arrows) and nuclei counterstained with DAPI (blue) cultured in osteogenic differentiation media and encapsulated at low, medium, and high cell seeding densities in non-degradable PEG (top) and MMP-sensitive (bottom) hydrogels. Scale bar is 10 μm. B) Semi-quantitative analysis of positive staining area (%) per nuclei for connexin 43.

Mineralization was quantified by calcium content (Figure 7A) and its spatial distribution assessed by von Kossa staining (Figure 7B) at day 28 of differentiation. Calcium content was affected by hydrogel type (p < 0.001) and cell encapsulation density (p < 0.001) with no significant interaction between the two factors. Calcium content increased (p < 0.001) with increasing cell encapsulation density and was consistently higher (p < 0.001) in the MMP-sensitive hydrogel compared to the non-degradable hydrogel. Spatially, there was minimal staining in the low cell encapsulation density in the non-degradable hydrogels. There was mineralization present in all other experimental groups with mineralization predominantly being localized to regions near the cells.

Figure 7.

Assessment of mineral content in IDG-SW3 cells encapsulated MMP-degradable and non-degradable PEG hydrogels at low, medium, and high cell seeding densities at day 28 of osteogenic differentiation. A) Calcium content; p-values are shown for pairwise comparisons between hydrogel and from low (Indicated by *) and from medium (indicated by #) for the same hydrogel. One symbol is p < 0.05, two symbols is p < 0.01, and three symbols is p < 0.001. B) Representative microscopy images of IDG-SW3 cells for mineralization by von Kossa staining (black) with nuclei counterstained with methyl red (red). Scale bar is 100 μm.

IDG-SW3 cells have been shown to produce a mineralized collagen matrix,²⁰ therefore total collagen content was quantified by hydroxyproline, an amino acid in high abundance in collagen (Figure 8A) and the spatial distribution of collagen type 1 (the major collagen type found in bone)⁴⁵ assessed by immunohistochemistry (Figure 8B) at day 28 of differentiation. The immunohistochemistry images were analyzed to quantify the amount of collagen I per cell as measured by positive staining area per nuclei (Figure 8C). Considering the non-degradable hydrogel only, total collagen content (i.e., Figure 8A) was not different between the low and medium but was higher (p < 0.01) in the high cell encapsulation density condition. Similarly, considering only the MMP-sensitive hydrogel, total collagen content was not different between the low and medium but was higher (p < 0.01) in the high cell encapsulation density condition. Comparing the two hydrogels, total collagen content was 3.2- (p = 0.004) and 2.9-fold (p = 0.006) higher in the MMP-sensitive hydrogels compared to the non-degradable hydrogels for the medium and high cell encapsulation density conditions, respectively. Collagen type I was spatially restricted to regions in and near the encapsulated cells in all conditions (Figure 8B). However, there were differences in the amount of collagen I staining per cell (i.e., Figure 8C). Considering only non-degradable hydrogels, the high encapsulation density showed greater collagen I per cell when compared to either low (p = 0.005) or medium (p = 0.007) cell encapsulation densities. Considering only the MMP-sensitive hydrogels, the amount of collagen I per cell was not affected by cell encapsulation density. This result suggests that the increased total collagen content in Figure 8A is due to a higher cell number. Comparing the two hydrogel types, the amount of collagen I per cell was 4.1-fold higher (p = 0.0114) for low and 4.2-fold higher (p = 0.0026) for medium cell encapsulation densities in the MMP-sensitive hydrogels. At high cell encapsulation density, there was no difference in the amount of collagen I per cell between the hydrogel types. However, there was more total collagen deposited in the MMP-sensitive hydrogels, which could be due to an increase in cell proliferation, although this was not measured.

Figure 8.

Assessment of collagen content in IDG-SW3 cells encapsulated MMP-degradable and non-degradable PEG hydrogels at low, medium, and high cell seeding densities at day 28 of osteogenic differentiation. A) Hydroxyproline content. B) Representative microscopy images of IDG-SW3 cells for collagen type I (red) with nuclei counterstained with DAPI. Scale bar length is 20 μm. C) Quantitative analysis of positive staining area (%) per nuclei for collagen I in the immunohistochemistry images. P-values are shown for pairwise comparisons between hydrogel and from low (indicated by *) and from medium (indicated by #) for the same hydrogel. One symbol is p < 0.05, two symbols is p < 0.01, and three symbols is p < 0.001.

The compressive modulus of the hydrogels was measured over time as a function of hydrogel type and cell encapsulation density (Figure 9). The non-degradable hydrogels maintained their compressive modulus with an average equilibrium modulus of 10.9 kPa over the duration of the 28 days for all three cell encapsulation densities. In the MMP-sensitive hydrogels, the modulus remained low for the low cell encapsulation density and did not change with time. For the medium cell encapsulation density, the modulus of the MMP-sensitive hydrogels decreased (p = 0.006) with time resulting in a modulus of 1.3 kPa by day 28. On the contrary, the modulus for the high cell encapsulation density in the MMP-sensitive hydrogels increased (p < 0.001) with time resulting in a modulus of 4 kPa at day 14 and increasing (p < 0.001) to 14 kPa by day 28.

Figure 9.

Compressive modulus measurements of IDG-SW3 encapsulated MMP-sensitive and non-degradable PEG hydrogels at low, medium and high cell seeding densities at days 14, 21 and 28 of osteogenic differentiation. P-values for MMP-13 sensitive hydrogels denote significance from day 1.

Collectively, these results demonstrate that the MMP-sensitive hydrogel led to enhanced dendritic-like cellular processes and increased positive staining for connexin 43. Further, the high cell encapsulation density in the MMP-sensitive hydrogels improved collagenous matrix mineralization when compared to the lower MMP-sensitive encapsulation densities and all encapsulation densities in the non-degradable hydrogels.

DISCUSSION

This study identified an MMP-sensitive hydrogel that is susceptible to degradation by MMP-2, MMP-13 and IDG-SW3 cells and supports osteoblast to osteocyte differentiation of IDG-SW3 cells. As compared to the non-degradable hydrogel, the MMP-sensitive hydrogel led to increased osteocyte differentiation and bone extracellular matrix deposition. Overall, a high cell encapsulation density combined with the MMP-sensitive hydrogel is a promising 3D culture system that supports the osteoblast to osteocyte transition of IDG-SW3 cells.

In this work, we confirmed that throughout osteocyte differentiation IDG-SW3 cells express MMPs of type 2, 9, 13, and 14 at the gene level. These four MMPs have been identified as playing an important role in skeletal development.⁴¹ Herein, Mmp2 expression maintained consistently higher expression throughout differentiation. Mmp14 and Mmp9 mirrored the temporal expression pattern of Mmp2, but at lower levels, and with Mmp9 at the lowest expression. Studies have reported that mice lacking MMP-14⁴⁶ or MMP-2⁴⁷ had aberrant bone formation. MMP-9 has been shown to be involved in apoptosis and angiogenesis⁴⁸ and important in endochondral ossification.⁴⁹ Thus, the IDG-SW3 osteoblast to osteocyte transition may require MMP-2 and MMP-14, but not MMP-9.

The temporal expression pattern of Mmp13 was uniquely different from the other three MMP types during IDG-SW3 osteoblast to osteocyte transition. While its expression was initially low compared to Mmp2 and Mmp14, a 200-fold increase occurred during differentiation from day 1 to day 35. Studies have reported that mice lacking MMP-13 displayed defects in remodeling and ECM organization in the perilacunar space in bone⁵⁰ In a separate in vitro study, upregulation of Mmp13 correlated to DMP-1 expression during migration of differentiating osteocytes in a dense collagen matrix.⁵¹ Further, MLO-A5 cells, which do not express the osteocyte marker DMP-1, expressed low Mmp13 levels during their differentiation period.⁵² Thus, it is possible that MMP-13 plays a critical role in the formation of the LCN and late stage osteocyte differentiation.

Based on the Mmp2 and Mmp13 expression profiles of IDG-SW3 cells during osteocyte differentiation in the established 2D culture, we chose an MMP-sensitive peptide crosslinker that was originally identified for its susceptibility to MMP-13,⁴⁴ but which we confirmed was also susceptible to MMP-2. IDG-SW3 cell-mediated hydrogel degradation was also confirmed by evidence of cell spreading. Extension of cellular processes was more pronounced within the MMP-sensitive hydrogel than in the non-degradable PEG hydrogel. This result is consistent with previous findings whereby a MMP inhibitor prevented cell spreading within similar MMP-sensitive PEG based hydrogels.⁴⁰ Bulk hydrogel degradation in the MMP-sensitive hydrogels was also evident, which indicates that cell-secreted enzymes diffused through the hydrogel and subsequently cleaved crosslinks in the bulk. This was particularly evident in the medium cell density condition where the hydrogel construct modulus decreased from days 14 to 28, which indicates bulk degradation. Although the exact MMP type or types used by IDG-SW3 cells during differentiation to degrade the MMP-sensitive hydrogel was not identified, the MMP-sensitive hydrogel formed with the peptide crosslinker, GCGPLG-LWARCG, supported the formation of dendritic-like cellular processes, which are important for osteocyte differentiation and a critical first step to forming the LCN of mature bone.

IDG-SW3 cells were derived from long bone chips and are capable of expressing osteoblast markers, producing a mineralized collagen matrix, as well as expressing early to late osteocyte markers.²⁰ To investigate the osteoblast to osteocyte transition of IDG-SW3 cells in this 3D hydrogel, this study assessed both osteoblast and osteocyte characteristics with the goal of identifying culture conditions that enhance osteocyte differentiation while supporting a bone ECM deposition. Osteocyte differentiation of the IDG-SW3 cells was most pronounced in the MMP-sensitive hydrogel. ALP is an osteoblast marker that is downregulated as osteoblasts differentiate towards osteocytes.^53,54 Regardless of the culture environment, ALP increased in the first 14 days of culture, and by day 28 ALP decreased to levels that were either similar to day 1 or lower than day 14. These results indicated a shift towards an osteocyte phenotype. IDG-SW3 cells in the MMP-sensitive hydrogels showed the highest expression of the osteocyte genes, Dmp1 and Phex¹⁹ and showed greater number of DMP1-GFP-positive cells by day 28. In vitro and in vivo, DMP-1 controls phosphate metabolism during osteocyte differentiation and is important for the organization of the osteocyte LCN.^55,56 PHEX, which is a membrane-bound endoprotease, binds DMP-1 to control phosphate homeostasis and ultimately mineralization.⁵⁷ The temporal profiles of Dmp1 and Phex expressions increased over the 28 days in the MMP-sensitive hydrogels but declined between 14 and 28 days in the non-degradable hydrogel.

Another hallmark of the osteoblast-to-osteocyte transition is dendrite formation.⁷ IDG-SW3 cells showed thin cellular processes that extend from the cell body wall, which may be an early form of osteocyte dendrite formation. There was evidence of short cellular processes in the non-degradable hydrogels at medium and high cell encapsulation densities by day 28. However, the cellular processes were longer and more pronounced in the MMP-sensitive hydrogels at all cell densities, likely due to cell-mediated hydrogel degradation. While dendrites that formed were less numerous than osteocyte LCN in bone tissue, the dendrites appeared to extend towards neighboring cells. In the non-degradable hydrogels, extension of these processes is only possible if they can extend through the mesh of the hydrogel. Our observations suggest that the hydrogel mesh restricted the development of the dendrite-like processes in the non-degradable hydrogels, which may have limited osteocyte differentiation in the non-degradable hydrogels. Yet, the evidence of cell-mediated degradation in the MMP-degradable hydrogels suggests that the MMP-degradation created space for more abundant dendrite formation and therefore enhanced osteocyte differentiation.

Connexin 43 staining was observed in both hydrogel types at day 28, and was greater in the MMP-sensitive hydrogels. In bone, connexin 43 is expressed by both osteoblasts and osteocytes and acts not only as a gap junction between cells, but also functions as an unopposed hemichannel. Hemichannels are located on the cell membrane, and unlike gap junctions, they do not require cell-to-cell connections.⁵⁸ Punctate staining for connexin 43 was observed and localized to cells, which is consistent with previous connexin 43 immunohistochemical staining.^59,60 Further studies are needed to confirm whether the observed connexin 43 is associated with hemichannels or gap junctions. Given that connexin 43 staining was similar regardless of cell density, the former role may dominate in this system. Connexin 43 is necessary for osteoblast and osteocyte survival and function⁶¹ and thus we postulate that it is likely present and playing a role in IDG-SW3 differentiation. The greater amount of connexin 43 staining in the MMP-sensitive hydrogels correlates to the observed greater prevalence of cellular processes and higher expression levels for osteocyte genes Dmp1 and Phex. Taken together, we hypothesize that the continued ability to extend cellular processes in the MMP-sensitive hydrogels by day 28 improved osteocyte differentiation, but that differentiation was impaired when growth of cellular processes was restricted in the non-degradable hydrogels. This observation suggests that the formation of cellular processes was critical to osteocyte differentiation and that degradable hydrogels facilitated formation of osteocyte networks in vitro. However, additional studies are needed to confirm this hypothesis.

Mineralization and collagen were also present in both hydrogel types at day 28, but to a greater amount in the MMP-sensitive hydrogels. ALP activity, which plays a role in bone mineralization, was also highest in the MMP-sensitive hydrogels,⁵⁴ and collagen can serve as a nucleation site for mineralization.⁶² Collagen I is the most abundant collagen type found in bone, and contributes approximately 20–25% of the composition of bone.⁴⁵ Osteocytes are known for orchestrating osteoblast bone-forming activity,⁶³ but osteocytes themselves can also deposit new bone matrix directly around osteocyte lacunae.⁶⁴ The exact characteristics that define osteocytes, as compared to their parent osteoblast population, are still being elucidated.^65,66 Due to the restricted transport of large ECM molecules, such as collagen, through the crosslinks of the hydrogel,⁶⁷ deposition was limited to the pericellular region in this study, which also closely resembles osteocyte perilacunar modeling. In the MMP-sensitive hydrogel, this region can grow providing increasingly more space for more ECM to deposit, which is supported at the low and medium cell densities by the higher amount of collagen type I positive staining per nuclei. Longer studies are needed to determine if this deposition of bone ECM would continue to replace hydrogel beyond the perilacunar space. In this study, osteocyte differentiation and bone ECM deposition were both enhanced in the MMP-sensitive hydrogel. Similarly, the pre-osteoblastic cell line MC3T3-E1 has shown higher amounts of Dmp1 and Phex expression concomitant with mineralization in a collagen gel.⁶⁸ Thus, it is possible that the ability to mineralize is a quality important to both osteoblasts and osteocytes in vitro, and that the presence of bone ECM may be necessary for osteocyte differentiation.

The cell encapsulation density was varied to investigate its impact on cellular morphology and bone ECM deposition. For the MMP-sensitive hydrogels, the formation of cellular processes and the positive staining of connexin 43 per cell were similar with different cell concentrations, indicating no effect on these osteocyte characteristics. Similarly, the spatial deposition of mineral around the cells and the collagen I staining on a per cell basis were similar across all cell densities in the MMP-sensitive hydrogels. On the other hand, the total amount of calcium and hydroxyproline content increased with cell density and, therefore, a higher number of cells.⁶⁷ Moreover, from day 14 to day 28 the compressive modulus tripled for the MMP-sensitive hydrogels with high cell encapsulation density which is attributed to a greater amount of ECM in the hydrogel. In the non-degradable hydrogels, cell encapsulation density had a greater effect on osteocyte characteristics. The hydrogels with low cell density performed poorly with minimal signs of cellular processes or connexin 43 and a lack of mineral deposits. On the contrary, the hydrogels with medium and high cell densities supported osteocyte characteristics albeit to a lesser degree than the MMP-sensitive hydrogels, which was supported by the quantitative analysis of the connexin 43 and collagen type I staining per cell. Overall, these results indicate that for the MMP-sensitive hydrogel, cell density does not have a significant impact on the formation of dendritic-like processes, expression of connexin 43, and bone ECM deposition. Rather, higher cell concentrations produce overall more mineralized collagen matrix, which leads to increases in the construct compressive modulus over time. The formation of a mineralized matrix, or osteoid, may be necessary to continue IDG-SW3 osteocyte differentiation into mature osteocytes, but cell density may not be a critical factor. Additional studies are needed to determine the long-term effect of differences in cell density.

There are several limitations of this study. Notably, there were several differences in hydrogel formulation. First, a PEG-dithiol crosslinker was chosen for the non-degradable hydrogel to ensure stability of the hydrogel. We have observed (unpublished observations) that scrambled peptide sequences can be broadly susceptible to degradation by collagenases. Second, the formulation of the hydrogel was varied due to differences in crosslinker reactivity, which led to differences in the PEG-NB concentration and thus slight differences in the crosslinked structure. Third, there were small differences in the modulus between the non-degradable and MMP-sensitive hydrogel (8 vs 10 kPa). It is possible that differences in the crosslink chemistry, formulation and/or modulus could have impacted the cells, but which was not tested in this study. We assessed ALP activity at discrete time points and therefore may have missed the timing of optimal activity, which typically peaks between 14 and 21 days. We also assessed gene expression at discrete time points and thus may have not captured the true trends in differentiation. Although, the IDG-SW3 cells were confirmed to differentiate in 2D and 3D in differentiation medium in the first part of this study, we did not investigate differentiation markers in the study with cell encapsulation density. Thus, it is possible the cell encapsulation density may have altered the timing of differentiation. This study focused on IDG-SW3 cells given their known ability to differentiate into mature osteocytes. Future studies should investigate whether this promising MMP-sensitive hydrogel could support osteocyte differentiation of other cell types, which do not readily differentiate into mature osteocytes, such as MLO-Y4, MLO-A5, and MC3T3-E1.

CONCLUSIONS

This study provides insight into the importance of 3D culture conditions on IDG-SW3 osteocyte differentiation. We show that an MMP-sensitive hydrogel promotes osteoblast to osteocyte differentiation by the enhanced expression of mature osteocytic genes, formation of dendritic-like cellular processes, positive staining for connexin 43, and a mineralized collagen matrix. An IDG-SW3 cell-laden hydrogel promotes bone ECM deposition around the cells as they differentiate from osteoblasts into osteocytes, thus mimicking aspects of in vivo bone formation, but in a highly controlled environment. The combination of IDG-SW3 cells with an MMP-sensitive hydrogel offers a more realistic in vitro culture system when compared to the traditional 2D culture on TCPS for the study of osteoblast to osteocyte differentiation. Moreover, once the cells have differentiated into osteocytes, formed dendrite-like processes, and are embedded in a mineralized collagen matrix, this platform can be used to investigate for example, the effects of drug treatment, hormone levels, or mechanical loading on osteocytes.

ABBREVIATIONS

MMP

matrix metalloproteinase

2D

two-dimensional

3D

three-dimensional

LCN

lacuno-canalicular network

TCPS

tissue culture polystyrene

PEG

poly(ethylene glycol)

ECM

extracellular matrix

INF-γ

interferon-gamma

MEM

Modified Essential Medium

FBS

fetal bovine serum

PSF

penicillin/streptomycin/ amphotericin B

PBS

phosphate buffered saline

GFP

green fluorescent protein

DMP1

dentin matrix acidic phosphoprotein 1

qPCR

quantitative Polymerase Chain Reaction

RE

relative expression

NE

normalized expression

HKG

housekeeping gene

GOI

gene of interest

IHC

immunohistochemistry

DAPI

4,6-diamidino-2-phenylindole ALP, alkaline phosphatase

MTS

Mechanical Testing System

ANOVA

analysis of variance