Material Properties and Osteogenic Differentiation of Marrow Stromal Cells on Fiber-Reinforced Laminated Hydrogel Nanocomposites

Abstract

The fibrils in the bone matrix are glued together by ECM proteins to form laminated structures (osteons) to provide elasticity and a supportive substrate for osteogenesis. The objective of this work was to investigate material properties and osteogenic differentiation of bone marrow stromal (BMS) cells seeded on osteon-mimetic fiber-reinforced hydrogel/apatite composites. Layers of electrospun poly(L-lactide) (L-PLA) fiber mesh coated with a poly(lactide-co-ethylene oxide fumarate) (PLEOF) hydrogel precursor solution were stacked and pressed together, and crosslinked to produce a laminated fiber-reinforced composite. Hydroxyapatite (HA) nanocrystals were added to the precursor solution to produce an osteoconductive matrix for BMS cells. Acrylamide-terminated RGD peptide (Ac-GRGD) was conjugated to the PLEOF/HA hydrogel phase to promote focal point adhesion of BMS cells. Laminates were characterized with respect to Young’s modulus, degradation kinetics, and osteogenic differentiation of BMS cells. The moduli of the laminates under dry and wet conditions were significantly higher than those of the fiber mesh and PLEOF/HA hydrogel, and within the range of values reported for wet human cancellous bone. At days 14 and 21, ALPase activity of the laminates was significantly higher than those of the fiber mesh and hydrogel. Lamination significantly increased the extent of mineralization of BMS cells and laminates with HA and conjugated with RGD (Lam-RGD-HA) had 2.7-, 3.5-, and 2.8-fold higher calcium content (compared to laminates without HA or RGD) after 7, 14, and 21 days, respectively. The Lam-RGD-HA group had significantly higher expression of osteopontin (OP) and osteocalcin (OC) compared to the hydrogel or laminates without HA or RGD, consistent with the higher ALPase activity and calcium content of Lam-RGD-HA. Laminated osteon-mimetic structures have the potential to provide mechanical strength to the regenerating region as well as supporting the differentiation of progenitor cells to the osteogenic lineage.

1. Introduction

There are more than 0.5 million skeletal injuries in the United States annually that require bone graft procedures to ensure rapid skeletal repair [1]. These include bone loss after skeletal trauma, resection of tumors, voids following osteoporotic fractures, and maxillofacial defects [2, 3]. Novel biomaterials that can provide temporary structural support to the regenerating region, initiate the cascade of osteogenesis and mineralized matrix formation, and degrade concurrent with the production of ECM are urgently needed for limb, head, and face reconstruction of patients with multiple traumatic injuries [4]. For example, as a scaffold, the biomaterial could be used in patients who have undergone frontotemporal craniotomy, with severely resorbed maxilla, as resorbable trays in reconstruction of large mandibular defects, and in alveolar ridge augmentation [5].

The porous collagen sponge is the most widely used scaffold by orthopedic surgeons because it provides an osteoconductive matrix for migration of bone marrow stromal (BMS) cells and a template for mineralization. However, additional mechanical protection in the form of a metallic cage is required to prevent deformation of the scaffold due to soft tissue compression [6–8]. Bioactive calcium phosphate (CaP) ceramic scaffolds are used clinically in spine fusion but, due to low initial strength, their use is limited to defects that are subject to uniform loading [9–13]. To improve bending strength and compressive modulus, composites of CaP with poly(L-lactide) (L-PLA) [14], poly(D,L-lactide) [15, 16], and poly(lactide-co-glycolide) (PLGA) [17–19] have been developed. Although PLGA/HA composites provide some structural support during bone repair, the major drawback is the hydrophobicity of the PLGA polymer. Unlike the collagenous phase in the natural bone, PLGA can not support complex cell-matrix and cell-cell interactions, solubilization of proteins, and growth factor modulation required for osteogenesis and vasculogenesis which results in undesirably long implantation periods for regeneration [20]. Composites of bioactive ceramics with natural hydrogels like collagen type I promote bone formation and certain compositions harden in-situ but they are prone to fatigue fracture [21, 22]. Nutrients and oxygen diffuse readily through synthetic and natural hydrogels and their hydrophilic structure supports complex cell-cell and cell-matrix interactions [23, 24].

Bone is a composite matrix consisting of a mineral and a collagenous phase [25]. The mineral phase provides mechanical strength in compression and an osteoconductive substrate for bone formation [26, 27] while the collagenous phase plays a central role in differentiation and maturation of progenitor cells and the extent of mineral-collagen interaction [28–30]. Bone exhibits hierarchical levels of organization from macroscopic to microscopic to nanoscale [31]. On the nanoscale (1–10 nm), an amino acid sequence next to the N-terminal and a glutamic acid sequence in osteonectin protein bind to the collagen network and apatite nanocrystals, respectively, to produce a mechanically robust composite for transmission of compressive load between the collagenous and apatite phases [32–34]. On the microscale (1–100 μm), layers of fibrils are glued together by ECM proteins to form laminated structures, called osteons, that make bone elastic and allow diffusion of nutrients and oxygen to cells embedded in the bone matrix [35]. The apatite crystals provide mechanical support in compression, the laminated structure of the osteons confers elasticity, while the hydrophilic ECM proteins allow diffusion of nutrients and oxygen to cells in the bone matrix.

The objective of this work was to develop a composite matrix to mimic the laminated structure of osteons in bone. Sheets of L-PLA nanofibers were fabricated by electrospinning. The sheets were coated with a hydrogel/hydroxyapatite (HA) precursor solution, stacked and pressed together, and allowed to crosslink by photopolymerization to form a fiber-reinforced hydrogel/HA laminated structure. The precursor solution was based on poly(lactide-co-ethylene oxide fumarate) (PLEOF) macromer that could be crosslinked in aqueous environment with redox or ultraviolet initiators to produce a biodegradable hydrogel [23, 36]. The crosslink density could be adjusted by the initiator concentration and number of fumarate groups on PLEOF macromer [37]. The degradability and water content of the hydrogel could be tailored to a particular application by varying the molecular weight of the lactide chains and the ratio of lactide (LA) to poly(ethylene glycol) (PEG) in PLEOF macromer [23, 37, 38]. PLEOF degradation could be modulated to the migration rate of progenitor cells by crosslinking PLEOF with a biologically degradable crosslinker [23]. The modulus of PLEOF/HA composite could be enhanced by treating the surface of the apatite crystals with an acrylate-functionalized glutamic acid sequence [33, 34]. Bioactive peptides could be bulk-conjugated or grafted to PLEOF to facilitate adhesion and differentiation of BMS cells [39, 40]. For example, integrin-binding RGD peptide was functionalized by reaction with acrylic acid to form an acrylamide-terminated RGD (Ac-GRGD) and conjugated to PLEOF hydrogel by the reaction between the acrylamide group of Ac-GRGD and PLEOF fumarate groups [39]. In this work, the effect of lamination on equilibrium water uptake, Young’s modulus, and degradation kinetics of the fiber-reinforced laminated composite was investigated. In addition, BMS cells seeded on RGD-conjugated laminated composites were evaluated with respect to osteogenic differentiation by measuring alkaline phosphatase (ALPase) activity, calcium content, and expression levels of osteogenic markers with incubation time.

2. Experimental

2.1. Materials

Triethylamine (TEA), tin (II) 2-ethylhexanoate (TOC), N-vinyl-2-pyrrolidone (NVP), and piperidine were purchased from Sigma-Aldrich (St. Louis, MO). Fumaryl chloride (FuCl; Sigma-Aldrich) was purified by distillation and PEG (Sigma-Aldrich; nominal molecular weight of 4.3 kDa) was dried by azeotropic distillation from toluene. Dichloromethane (DCM; Acros Organics) was dried by distillation over calcium hydride (Sigma-Aldrich). Diethylene glycol (DEG), N,N-dimethylformamide (DMF), trifluoroacetic acid (TFA), N,N-dimethylaminopyridine (DMAP), N,N′-diisopropylcarbodiimide (DIC), hydroxybenzotriazole (HOBt), acetonitrile (MeCN), triisopropylsilane (TIPS), N,N′-methylene bisacrylamide (BISAM), ammonium persulfate (APS) and N,N,N′,N′-tetramethylethylenediamine (TMEDA) were purchased from Acros Organics (Fisher; Pittsburg, PA). L-Lactide Monomer (LA; >99.5% Purity by GC) was purchased from Ortec (Easley, SC). High molecular weight L-PLA (0.9–1.2 dL/g intrinsic viscosity and 185 kDa weight average molecular weight) and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) were purchased from Durect Corp. (Pelham, AL) and VWR (West Chester, PA), respectively. The Rink Amide NovaGel^™ resin and all Fmoc-protected amino acids were purchased from Novabiochem (EMD Biosciences, San Diego, CA). Dulbecco’s phosphate-buffer saline (PBS) and Dulbecco’s Modified Eagle’s Medium (DMEM; 4.5 g/L glucose with L-glutamine and without sodium pyruvate) were purchased from Cellgro (Mediatech; Herndon, VA).

2.2. Synthesis and characterization of PLEOF macromer

Low molecular weight poly(L-lactide) (LMW-PLA) was synthesized by ring opening polymerization of LA monomer as described [36, 40]. DEG and TOC were used as the bifunctional initiator and polymerization catalyst, respectively. PLEOF macromer was synthesized by condensation polymerization of LMW-PLA and PEG with FuCl as described [23, 37, 38] (see Figure 1). The weight ratio of PEG to LMW-PLA was 70:30 to produce a hydrophilic water soluble terpolymer. After completion of the reaction, the product was dissolved in DCM, precipitated twice in ether, dried in vacuum (<5 mmHg) for at least 12 h and stored at −20°C until used. Chemical structure of the macromers was characterized by a Varian Mercury-300 ¹H NMR (Varian, Palo Alto, CA). The polymer sample was dissolved in CDCl₃ at a concentration of 50 mg/mL, and 1% v/v tetramethylsilane (TMS) was used as the internal standard. The molecular weight distribution of the macromers was measured by gel permeation chromatography (GPC) [23]. Measurements were carried out with a Waters 717 GPC system (Waters, Milford, MA) connected to a model 410 refractive index detector. 50 μL of the sample (2 mg/mL in THF) was eluted through a styragel HR 4E column (300 mm × 7.8 mm, Waters) with degassed THF at a flow rate of 1 mL/min. Monodisperse polystyrene standards (0.58–66.35 kDa and polydispersities of <1.1, Waters) were used to construct the calibration curve.

Figure 1.

Reaction scheme for the synthesis and crosslinking of poly(lactide-co-ethylene oxide fumarate) (PLEOF). PLEOF macromer was synthesized by condensation polymerization of low molecular weight poly(L-lactide) (LMW-PLA) and poly(ethylene glycol) (PEG) with fumaryl chloride (FuCl). Triethylamine (TEA) was used as a radical scavenger. Next, PLEOF macromer, N,N′-methylene bisacrylamide crosslinker (BISAM), and acrylamide terminated Ac-GRGD peptide were mixed in phosphate-buffer saline (PBS), equimolar amounts of ammonium persulfate (APS) polymerization initiator and N,N,N′,N′-tetramethylethylenediamine accelerator (TEMED) were added, and the polymerizing precursor solution was allowed to crosslink to form a hydrogel. Hydroxyapatite (HA) nanocrystals can be added to the polymerizing solution to produce a hydrogel/HA composite.

2.3. Electrospinning and characterization of L-PLA fiber mesh

The L-PLA fiber mesh was prepared by electrospinning from HFIP solvent (9 wt%) as described [41]. The use of HFIP as the solvent resulted in the production of bead-free fibers at the lowest L-PLA concentration of 9 wt%. The optimum conditions of 1.0 mL/h injection rate, 25 kV electric potential, and 7.0 cm needle-to-collector distance were used [41]. A programmable KDS100 syringe pump (KD Scientific, Holliston, MA) was used to transfer the polymer solution from a 1 ml syringe (Norm-Ject, Henke Sass Wolf GmbH, Germany) to the end of a 21 gauge needle (GTW-PrecisionGlide, 0.7 mm I.D., Becton-Dickinson, Franklin, NJ). A positively charged Pt electrode of a high voltage supply (ES40P-5W/DAM, Gamma High Voltage Research) was connected to the end of the needle. An aluminum plate connected to the ground electrode was used as the collector. To determine morphology and size, the fiber mesh was attached to a SEM stub, coated with gold using an Ion Sputter Coater (JEOL, JFC-1100) at 20 mA current for 1 min, and imaged with a JSM-5400 scanning electron microscope (SEM; JOEL, Japan) at an accelerating voltage of 10 kV. SEM images were analyzed to determine the average fiber size using the ImageJ software (National Institutes of Health, Bethesda, MD).

2.4. Synthesis and characterization of acrylamide-terminated GRGD peptide

The chemical structure of Ac-GRGD is shown in Figure 1. GRGD peptide was synthesized manually on Rink Amide NovaGel^™ resin in the solid phase using a previously described procedure [42, 43]. Briefly, the Fmoc-protected amino acid (6 eq.), DIC (6.6 eq.), and HOBt (12 eq.) were added to 100 mg resin and swelled in DMF (3 mL). Next, 0.2 mL of 0.05 M DMAP was added to the mixture and the coupling reaction was allowed to proceed for 4–6 hr at 30°C with orbital shaking. If the Kaiser test for the presence of unreacted amines was negative [42, 44], the resin was treated with 20% piperidine in DMF and the next Fmoc-protected amino acid was coupled using the same procedure. To improve reaction yield [45], the peptide was functionalized directly on the peptidyl resin by coupling acrylic acid to the N-terminal amine group under conditions used for the amino acid coupling reaction. The acrylamide-terminated peptide was cleaved from the resin by treating with 95% TFA/2.5% TIPS/2.5% water and precipitated in cold ether. The product was purified by preparative HPLC on a 250×10 mm, 10 μm Xterra® Prep RP18 column (Waters, Milford, MA), flow rate of 2 mL/min using a gradient 5–95% MeCN in 0.1% aqueous TFA with a photodiode array detector (model 996, Waters) at 214 nm wavelength [39]. The retention time of the peptide was 8.04 min. The purified peptide was characterized with a Finnigan 4500 Electro Spray Ionization (ESI) spectrometer (Thermo Electron, Waltham, MA). In the ESI-MS spectrum, mass numbers (m/z) 457 and 479 corresponded to monovalent hydrogen cation [(M+H)+] and monovalent sodium cation [(M+Na)+] of the purified Ac-GRGD peptide, respectively.

2.5. Fabrication of fiber-reinforced laminated composites

Schematic diagram of the lamination process is shown in Figure 2. Fiber layers were coated with the hydrogel precursor solution, coated layers were stacked together, compressed, and the multi-layer assembly was crosslinked to form a laminated nanocomposite. The hydrogel/HA precursor solution was prepared by first dispersing HA nanowhiskers (principle axis=80 nm and aspect ratio=4; Berkeley Advanced Biomaterials, Berkeley, CA) in water. HA nanocrystals (10% by weight of PLEOF, 32 mg HA) were added to 0.825 mL PBS and the resulting dispersion was sonicated for 5 min with a 3mm probe sonicator (Cole-Parmer Instruments, Vernon Hills, IL) to break down the aggregates and produce a homogeneous dispersion [33, 34]. Next, 30 mg BISAM and 3.2 mg Ac-GRGD (1% by weight of PLEOF) peptide were dissolved in the HA dispersion, followed by the addition of 315 mg PLEOF macromer, and the mixture was heated to 50°C and vortexed to aid dissolution. Then, 105 μL 0.2M APS and 105 μL 0.2M TMEDA were added to initiate polymerization. The L-PLA nanofiber mesh was placed on a Teflon plate, the hydrogel precursor solution was brushed over the fibers, followed by placing another nanofiber layer on top of the precursor solution. This process was repeated to produce a four-layer laminated composite. The precursor solution acted as a “glue’ to hold together the fibrous layers. Another Teflon plate was placed on top of the laminate, a pressure of 4.7 kPa was applied to the assembly to squeeze out the extra solution, and allowed to crosslink for 30 min at 50°C. The final thickness of the laminate was 85±15 μm.

Figure 2.

Schematic diagram for the fabrication of fiber-reinforced hydrogel/HA laminated composite. The electrospun L-PLA nanofiber mesh (a) was placed on a Teflon plate, the hydrogel precursor solution (b) was brushed over the fibers (c), followed by placing another fiber layer on top of the precursor solution. This process was repeated to produce a multi-layer laminated composite (d). The precursor solution acted as a “glue’ to hold together the fiber layers. Next, another Teflon plate was placed on top of the laminate (e), a pressure of 4.7 kPa was applied to the assembly to squeeze out the extra solution, and allowed to crosslink (e). The hydrogel/HA precursor solution was prepared by mixing HA nanocrystals, BISAM crosslinker, Ac-GRGD cell adhesion peptide, APS initiator, and TMEDA accelerator in PBS.

2.6. Characterization of fiber-reinforced laminated composites

The laminated composites were imaged with an SEM with accelerating voltage of 10 kV as described in section 2.3. For determination of water uptake, crosslinked laminated disks (8 mm diameter) were soaked in PBS for 24 h at 37°C with a change of media every 6 h, and swollen weight, W_s, was measured. Next, the disks were placed in distilled deionized (DDI) water for at least 12 h to remove excess electrolytes. Then, samples were dried at ambient conditions for 12 h followed by drying in vacuum at 40°C for 1 h and the dry sample weight, W_d, was recorded [23]. The equilibrium water uptake of the laminates was determined by Q = (Ws − W_d)/W_d. Degradation was measured as a function of time in 5 mL primary media [DMEM supplemented with 10% FBS, 100 units/mL penicillin (PEN), 100 μg/mL streptomycin (SP), 50 μg/mL gentamicin sulfate (GS) and 250 ng/mL fungizone (FZ)], without fetal bovine serum (FBS) at 37°C under mild agitation. At each time point, samples were removed from the media, washed with DDI water, and dried in vacuum. The dry sample weight was recorded and compared with the initial dry weight to determine fractional mass remaining. A Rheometrics RSA III dynamic mechanical analyzer (DMA; TA Instruments, New Castle, DE) was used to measure Young’s modulus of the laminated nanocomposites at 37°C. The DMA was used in the static mode at a strain rate of 0.002/s. Rectangular films 20 mm in length, 4 mm in width, and 85 μm in thickness were used to generate the strain-stress curves.

2.7. BMS cell isolation and seeding

BMS cells were isolated from the bone marrow of young adult male Wistar rats [23, 39, 40]. Cell isolations were performed under a protocol approved by the Institutional Animal Care and Use Committee of the University of South Carolina. After aseptically removing the femurs and tibias, the marrow was flushed out with 20 mL of cell isolation media (DMEM supplemented with 100 units/mL penicillin (PEN), 100 μg/mL SP, 20 μg/mL FZ and 50 μg/mL GS). The cell suspension was centrifuged at 200×g for 5 min, cell pellets were re-suspended in 12 mL of primary media, aliquoted into T-75 flasks, and maintained in a humidified 5% CO₂ incubator at 37°C. Cultures were replaced with fresh media at 3 and 7 days to remove haematopoietic and other unattached cells. After 10 days, cells were detached from the flasks with 0.05% trypsin-0.53 mM EDTA and seeded on laminated composites.

For cell seeding, the sample (1 cm in diameter) was placed on a sterile round glass coverslip and the edge was coated with a silicone sealant (uncatalyzed peroxide-initiated Class VI medical grade liquid silicone rubber; Dow Corning; Midland, MI). The sealant was allowed to harden for 12 h under sterile conditions to prevent separation of the sample from the coverslip in culture media. Next, the construct was sterilized by ultraviolet radiation for 1 h with a BLAK-RAY 100 W mercury long wavelength (365 nm) UV lamp (UVP; Upland, CA) followed by 75% ethanol, and washed with PBS prior to cell seeding. 250 μl of the BMS cell suspension at a density of 4×10⁵ cells/ml in primary media was placed on each sample and incubated for 24 h. After cell attachment, media was replaced with complete osteogenic media (primary media supplemented with 100 nM dexamethasone, 50μg/ml ascorbic acid, and 10mM β-glycerophosphate), and cultured for up to 21 days. For imaging with SEM, cell-seeded samples were fixed with 4% paraformaldehyde (Sigma-Aldrich) in PBS for 40 min at ambient conditions. Next, samples were stained with osmium tetroxide (Sigma-Aldrich), dehydrated in sequential ethanol solutions, and dried by critical point drying. The dried specimen was mounted on a stub, coated with gold (Polaron sputter coater, Quorum Technologies, New Haven, UK) and observed with a JEOL SEM at an accelerating voltage of 10 kV (JSM-6300; Tokyo, Japan).

2.8. Osteogenic differentiation of BMS cells on laminated composites

At each time point (7, 14, 21 days), cell-seeded laminates were washed with serum-free DMEM for 8 h to remove serum components, washed with PBS, lysed, and used for measurement of DNA content, ALPase activity and calcium content. The double stranded DNA content of the samples was determined using a Quant-it PicoGreen assay (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions. An aliquot (100μL) of the working solution was added to 100μL of the cell lysate and incubated for 4 min at ambient conditions. The fluorescence was measured with a Synergy HT plate reader (Bio-Tek; Winooski, VT) at emission and excitation wavelength of 485 and 528 nm, respectively. Measured fluorescent intensities were correlated to cell numbers using a calibration curve constructed with BMS cells of known concentration ranging from zero to 4×10⁴ cells/mL. ALPase activity was assessed using QuantiChrom ALPase Assay Kit (BioAssay Systems, Hayward, CA) according to manufacture’s instructions. A 10 μL aliquot of the sonicated cell lysate was added to 190 μL of the reagent solution containing 10 mM p-nitrophenyl phosphate and 5 mM magnesium acetate and absorbance was recorded at time zero and again after 4 min. ALPase activity was calculated using the equation [(A_t=4 − A_t=0)/(A_calibrator − A_ddH2O)×808] expressed as IU/L. The absorbance was measured on a Synergy HT plate reader at 405 nm. The calcium content was measured using a QuantiChrom Calcium Assay Kit (BioAssay Systems) according to manufacturer’s instructions. 0.2 mL of 2M HCl was added to 0.2 mL aliquot of the sonicated cell lysate to dissolve the calcium content of the mineralized matrix. Next, a 5 μL aliquot of the supernatant was added to 200μL of the working solution. After incubation for 3 min, the absorbance was measured with a plate reader at 612 nm. Measured intensities were correlated to equivalent amounts of Ca²⁺ using a calibration curve constructed with calcium chloride solutions of known concentration ranging from zero to 200 μg/mL. For HA containing samples, the measured intensities at day 4 (negligible mineralization after 4 days) were used as the baseline, and subtracted from the measured intensities at days 7–21.

2.9. mRNA analysis of BMS cells on laminated composites

At each time point, total cellular RNA was isolated using TRIzol (Invitrogen, Carlsbad, CA) plus RNeasy Mini-Kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. The qualitative and quantitative analysis of the RNA samples was performed with NanoDrop 2000 (Thermo Scientific, Waltham, MA). The obtained RNA histograms and gel images were analyzed for the intact 28S and 18S ribosomal RNA. 1μg of the extracted total RNA was subjected to cDNA conversion using the Reverse Transcription System (Promega, Madison, WI). The obtained cDNA was subjected to classical polymerase chain reaction (PCR) amplification with appropriate gene specific primers and the control primer ARBP (acidic ribosomal phosphoproteins) for 35 cycles. Primers for real-time PCR analysis were designed and selected using the Primer3 web-based software as described [43]. The PCR products were analyzed by agarose gel electrophoresis with 2% ethidium bromide staining (Sigma-Aldrich). The annealing temperatures and other parameters for amplification were optimized by classical PCR and agarose gel electrophoresis as described [43]. Real-time PCR (RT-qPCR) was performed to analyze the differential expression of osteopontin (OP), osteocalcin (OC), and osteonectin (ON) genes with SYBR green RealMasterMix (Eppendorf, Hamburg, Germany) using Bio-Rad iCycler machine (Bio-Rad, Hercules, CA) and iCycler optical interface version 2.3 software. The following forward and reverse primers were synthesized by Integrated DNA technologies (Coralville, IA): Osteonectin: forward 5′-ACA AGC TCC ACC TGG ACT ACA and reverse 5′-TCT TCT TCA CAC GCA GTT T; Osteopontin: forward 5′-GAC GGC CGA GGT GAT AGC TT and reverse 5′-CAT GGC TGG TCT TCC CGT TGC; Osteocalcin: forward 5′-AAA GCC CAG CGA CTC T and reverse 5′-CTA AAC GGT GGT GCC ATA GAT; and ARBP: forward 5′-CGA CCT GGA AGT CCA ACT AC and reverse 5′-ATC TGC TGC ATC TGC TTG. Quantification of gene expression was based on the crossing-point threshold value (CT; number of cycles required for the RT-qPCR fluorescent signal to cross the threshold) for each sample. This was evaluated by the Relative Expression Software Tool (RESTTM; [46]) as the average of three replicate measurements. The expression of the ARBP house-keeping gene was used as the reference and the fold difference in gene expression was normalized to that at time zero. The model of Pfaffl, which includes an RT-qPCR efficiency correction factor of the individual transcripts, was used to determine the expression ratio of the gene [46]. The CT values were processed and analyzed for standard error and significant difference between the groups with the Q-gene software (www.biotechnique.com/softlib/qgene.html) [47].

2.10. Statistical analysis

Data are expressed as means ± standard deviation. Significant differences between two groups were evaluated using a two-tailed student t-test. A value of p<0.05 was considered statistically significant.

3. Results and discussion

3.1. Polymer characterization

Chemical structure of the synthesized macromers was characterized by ¹H-NMR. The ratio of the chemical shift with peak position at 3.6 ppm to that at 5.1 ppm was directly related to $\overline{M_n}$ of LMW-PLA, which was consistent with $\overline{M_n}$ values obtained from GPC. The $\overline{{\text{M}}_{\text{n}}}$ and polydispersity index (PDI) of LMW-PLA were 1.2 kDa and 1.4, and those of 30/70 PLEOF were 8.0 kDa and 1.6., as measured by GPC. The SEM images of the fiber mesh, one-layer laminate, and 4-layer laminate are shown in Figures 3a, 3b, and 3c, respectively. The average diameter of the fibers, produced from 9 wt% L-PLA in HFIP solvent, was 610±320 nm. Arrows in Figure 3c show each layer in the multi-layer laminate.

Figure 3.

SEM images of the fiber mesh (a), one-layer laminate (b), and 4-layer laminate (c). Arrows in Figure 2c show each layer in the multi-layer laminate.

3.2. Water uptake and modulus of laminated composites

Figure 4a shows water uptake (weight of water/dry sample weight) of the samples. The gray area in Figure 4a is the range for water content of the collagenous phase of the natural bone excluding the mineral component (bone water content is 15–20% by weight [48] corresponding to 43–50% (0.75–1.0 g/g dry) without the mineral component). The water uptake of the hydrogel and fiber mesh was 17.5±3.4 and 4.6±1.3, respectively. The porosity of the fiber mesh contributed to the relatively high water uptake of the hydrophobic L-PLA fibers. The addition of 10% HA to the hydrogel reduced water content to 14.2±7.0. Lamination of the fiber mesh with the hydrogel drastically reduced water content from 17.5±3.4 (95% water) to 1.6±0.3 (61%). For the laminates, the hydrogel swelling pressure was counterbalanced by the elastic force of the fiber mesh, resulting in a significant reduction in water uptake. With the addition of HA, RGD, and RGD-HA to the hydrogel, water uptake reduced from 1.6±0.3 (no RGD and HA) to 1.4±0.4, 1.4±0.5, and 1.1±0.3, respectively. The higher water uptake of the ionic Ac-GRGD, conjugated to the hydrogel, was offset by the higher extent of crosslinking (via the unsaturated acrylate group of Ac-GRGD). According to Figure 4a, lamination of the fiber mesh with the hydrogel produced composites with water content similar to that of the collagenous phase of the bone matrix.

Figure 4.

(a) Water uptake (weight of water/dry weight) and (b) Young’s moduli under dry and wet conditions of the laminates. Experimental groups include 30/70 PLEOF hydrogel (Gel), PLEOF hydrogel with 10% HA (Gel-HA), L-PLA electrospun fiber mesh (Fiber), L-PLA fiber mesh/PLEOF hydrogel laminate (Lam), Laminate with 10% HA in PLEOF (Lam-HA), laminate conjugated with 1% Ac-GRGD in PLEOF (Lam-RGD), and laminate with 1% Ac-GRGD and 10% HA (Lam-RGD-HA). For wet conditions in (b), samples were soaked in primary media without FBS for 24 h prior to DMA measurements. The gray area in (a) is the range for water content of the natural bone excluding the mineral component. The gray area in (b) is the reported range for Young’s modulus of wet human cancellous bone [26]. Error bars correspond to means ± 1 SD for n = 3.

The reduced water content of the composites with lamination affected their moduli under dry and wet conditions, as shown in Figure 4b. The gray area in Figure 4b is the reported range for Young’s modulus of wet human cancellous bone [26]. The modulus of the fiber mesh under dry and wet conditions was 140±3 MPa, but modulus of the hydrogel under wet condition was two orders of magnitude lower than that under dry condition (0.50±0.07 versus 139±23 MPa). Addition of HA to the hydrogel did not significantly improve modulus under wet condition. Lamination of the fiber mesh with PLEOF hydrogel increased modulus dramatically (compared to the fiber mesh or hydrogel) to 570±130 MPa. More importantly, Lam had similar modulus under dry and wet conditions (570±130 for dry and 575±14 MPa for wet). It is interesting to note that all laminated composites exhibited modulus values within the range for that of wet human cancellous bone (within the gray area) while modulus of the fiber mesh or hydrogel was significantly less than that of cancellous bone. The fiber mesh reinforced and reduced water uptake of the matrix, resulting in a laminated structure with robust mechanical properties under wet and dry conditions. Due to low interfacial interaction, addition of 10% HA to the hydrogel reduced modulus of the laminate under dry and wet conditions. We and others have demonstrated that energetic interactions at the interface of the hydrogel and filler nanoparticles significantly affect viscoelastic properties of the composite [33, 34, 49–53]. For example, we have shown previously that the addition of an acrylamide-terminated glutamic acid sequence to apatite nanocrystals improves shear modulus of the composite by an order of magnitude [33, 34]. Addition of Ac-GRGD to the hydrogel phase of the laminate substantially increased modulus of the dry laminate (590±230 MPa; due to higher degree of crosslinking) but reduced its modulus under wet condition (300±70 MPa, due to higher water uptake). The modulus of Lam-RGD-HA samples under dry and wet conditions was 470±25 and 380±47 MPa, respectively, which was between those of Lam-HA and Lam-RGD. These results demonstrate that fiber-reinforced laminated composites have the potential to provide structural support to the regenerating region in load-bearing orthopedic applications. The mechanical characteristics of the laminates can be further improved by the extent of compression in the lamination step and crosslinking, fiber to hydrogel ratio, HA content and surface treatment, water uptake of the hydrogel, fiber orientation, and the number of fiber layers in the laminate.

3.3. Degradation Characteristics of laminated composites

Mass loss of the L-PLA fiber mesh (Fiber), RGD-conjugated PLEOF hydrogel (Gel-RGD), laminate of L-PLA with PLEOF hydrogel (Lam), and laminate with 10% (by PLEOF weight) HA (Lam-HA) with incubation time is shown in Figure 5. Gel-RGD sample completely degraded in 3 weeks while the fiber mesh lost <10% of its mass in 4 weeks. For a given degradation time, mass loss of the laminate was less than that of the hydrogel but greater than the fiber mesh. The laminates (without HA) lost 18, 29, 34, and 44% of its mass after 1, 2, 3, and 4 weeks, respectively. The laminates with and without 10% HA showed a similar trend in mass loss with time. Hydrogel mass loss can be tailored to a particular application by varying the ratio of lactide to PEG in the macromer and extent of crosslinking. For example, UV-crosslinked PLEOF hydrogels with lactide fractions of 10, 20, 30, and 40% had 19, 23, 70, and 86% mass loss after 4 weeks in primary media at 37°C. Higher lactide fractions in PLEOF increase the density of degradable ester bonds which increase hydrogel degradation rate [37]. The results in Figure 5 demonstrate that composites with bimodal degradation profile can be produced by varying the ratio of L-PLA fibers to PLEOF hydrogel in the laminate: a slow degradation rate by the fibers for structural stability and a fast degradation rate by the hydrogel to increase volume for cell migration and ECM production.

Figure 5.

Degradation characteristics with incubation time for L-PLA fiber mesh (Fiber), 30/70 PLEOF hydrogel conjugated with 1% RGD (Gel-RGD), laminate of L-PLA with 30/70 PLEOF hydrogel (Lam), and laminate with 10% HA (Lam-HA).

3.4. Adhesion of BMS cells to laminated composites

BMS cells were seeded on laminated composites and cultured in complete osteogenic media for 21 days. Experimental groups included 30/70 PLEOF hydrogel with 1% RGD (Gel-RGD), L-PLA fiber mesh (Fiber), 4-layer L-PLA nanofiber/PLEOF hydrogel laminated composite (Lam), laminate with 10% HA (Lam-HA), laminate conjugated with 1% RGD (Lam-RGD), and laminate with 1% RGD and 10% HA (Lam-RGD-HA). A typical surface coverage and morphology of the BMS cells on a laminated surface (Lam-HA) is shown in Figure 6. It should be noted that the observed morphology in Figure 6 could be due to shrinkage of the cell aggregates after freeze-drying from confluent cell layers. Nevertheless, the SEM image clearly shows the cuboidal to polygonal morphology of BMS cells on the laminates. It is well-established that mechanical factors such as surface hardness and modulus as well as biochemical factors affect cell adhesion and spreading [54, 55]. For example, Kim and Park reported that fibroblast cells had a rounded morphology on untreated PLGA fiber mesh while the cells spread on RGD-grafted fibers [56]. The fiber mesh had relatively higher cell coverage but fewer focal point adhesions due to lack of specific cell-substrate interactions. This was consistent with Mo and collaborators observation that untreated electrospun poly(L-lactid-co-ε-caprolactone) (PLLA-CL) nanofibers support docking/attachment of endothelial and smooth muscle cells [57]. The laminates without Ac-GRGD, due to non-adherent nature of the PLEOF hydrogel, had lower cell coverage but coverage improved by reinforcement with HA, conjugation with Ac-GRGD, or modification with HA and Ac-GRGD.

Figure 6.

A typical SEM image of the BMS cells on the laminate with 10% HA (Lam-HA).

3.5. Osteogenic differentiation of BMS cells on laminated composites

The seeded BMS cells were analyzed for DNA content, ALPase activity (early marker for osteogenesis) and calcium content (late marker for osteogenesis) with incubation time in osteogenic media. Figure 7a shows DNA content of the samples as a function of time. DNA content of the BMS cells on L-PLA fiber mesh, due to higher surface area, was significantly higher that that of the laminate groups, as shown in Figure 7a. However, it should be noted that osteogenic differentiation of BMS cells, which is strongly dependent on cell-cell distance on the substrate, on Gel-RGD and all laminate groups could be directly compared because they had similar surface area for cell seeding and similar seeding density (see Figure 7a). Among the laminate groups, Lam-RGD-HA had the highest DNA content compared to Lam, Lam-RGD, and Lam-HA for all time points. For example, DNA content of Lam group increased from 170±10 to 366±32 and 432±16 ng/ml after 7, 14, and 21 days, respectively, while that of Lam-RGD-HA increased from 267±24 to 395±43 and 620±10 ng/ml. Shofer and collaborators observed that the normalized density of human mesenchymal stem cells (hMSC) on L-PLA/collagen I blend nanofibers increased from 42±8 to 60±6 and 75±5 when L-PLA to collagen ratio was increased from 1:2 to 2:1 and 4:1 [58]. The Shofer results demonstrate that substrate modulus (L-PLA fraction) plays a significant role in cell attachment. Patel and collaborators demonstrated that the adhesion of preadipocytes seeded on YIGSR-conjugated PEG hydrogels increased from a low of 0.2 μg/gel for the untreated gel to a high of 0.45 μg/gel for YIGSR-conjugated hydrogel [59]. Similarly, the density of fetal human osteoblasts, seeded on poly(hydroxyethyl methacrylate) (pHEMA) hydrogels increased from a low of 420±50 cells/cm² for untreated pHEMA to 1050±100 cells/cm² for poly(L-lysine) coated pHEMA [60].

Figure 7.

DNA content (a), ALPase activity (b), and calcium content (c) of the BMS cell seeded on different samples with incubation time. Groups include Gel-RGD, Fiber, Lam, Lam-HA, Lam-RGD, and Lam-RGD-HA. One star indicates statistically significant difference between the test group and Fiber for the same incubation time. Two stars indicate statistically significant difference between the test group (Lam-HA, Lam-RGD, or Lam-RGD-HA) and lam group for the same time. Error bars correspond to means ± 1 SD for n = 3.

Figure 7b shows ALPase activity of the BMS cells seeded on different samples as a function of time. The L-PLA fiber mesh showed the lowest ALPase activity for any time point, despite having the highest DNA content (see Figure 7a). The hydrophobic nature of the L-PLA fibers and lack of specific cell-substrate interaction did not support osteogenic differentiation of BMS cells. In contrast, the RGD-conjugated PLEOF hydrogel (Gel-RGD) and L-PLA fibers laminated with RGD-conjugated hydrogel, with relatively lower cell density compared to L-PLA fibers (see Figure 7a), had significantly higher ALPase activity. For example, ALPase activity of BMS cells on Gel-RGD was 1.0±0.08 and 1.37±0.09 IU/μg DNA after 14 and 21 days, respectively, while that of L-PLA fiber was 0.44±0.04 and 0.90±0.08 IU/μg. ALPase activity of the fiber-reinforced hydrogel laminate (Lam) and Lam-RGD after 21 days was 1.75±0.01 and 1.80±0.01 IU/μg, respectively, which was significantly higher than that of the fiber mesh (0.88±0.08 and 1.37±0.09). Lamination of the fiber mesh with PLEOF hydrogel provided a less deformable substrate for cell docking and attachment while the RGD-conjugated hydrogel facilitated cell-laminate interaction, consistent with previous results on RGD modified substrates [39, 61, 62]. Furthermore, the addition of HA to the hydrogel phase of the laminate dramatically increased ALPase activity of BMS cells. For example, ALPase activity of Lam-RGD-HA and Lam-HA at day 21 were 2.92±0.16 and 2.83±0.43 IU/μg DNA, respectively, while that of Lam-RGD and Lam were 1.82±0.01 and 1.75±0.01 IU/μg. This is consistent with previous results showing ALPase activity of MC3T3-E1 preosteoblast cells on poly(ε-caprolactone) (PCL)/HA was higher than that of PCL alone [63]. Overall, ALPase activity of Lam-RGD-HA was higher than the other groups for any time point.

Figure 7c shows calcium content of the BMS cells seeded on different samples as a function of time. L-PLA fiber mesh, Gel-RGD, Lam, and Lam-RGD groups had similar calcium contents ranging between 0.003–0.008 and 0.021–0.026 mg/cm² after 14 and 21 days, respectively. Addition of HA to the laminates (Lam-HA) increased calcium content by 2.4-fold compared to Lam to 0.062±0.002 mg/cm² after 21 days. RGD Conjugation and HA reinforcement (Lam-RGD-HA) increased calcium content of the laminates by 2.7-, 3.5-, and 2.8-fold (compared the lam) after 7, 14, and 21 days, respectively.

The expression level of osteogenic markers OP, OC, and ON as a function of time is shown in Figures 8a, 8b, and 8c, respectively. In Figure 8, one star indicates statistically significant difference between the laminate groups and Gel-RGD while two stars indicate a significant difference between the test group (Lam-HA, Lam-RGD, or Lam-RGD-HA) and Lam. Three stars indicate a significant difference between the Lam-RGD-HA group and Lam-HA or Lam-RGD. OP and OC expression levels increased while that of ON decreased with incubation time. At 14 and 21 days, Lam-RGD-HA group had significantly higher OP expression compared to Gel-RGD (one star), Lam (two star), and Lam-HA or lam-RGD (three stars), consistent with the ALPase activities and calcium contents shown in Figures 7b and 7c, respectively. After 7 and 14 days, Lam-RGD-HA group had significantly higher OC expression compared to Gel-RGD (one star), Lam (two star), and Lam-HA or lam-RGD (three stars), consistent with the ALPase activities and calcium contents shown in Figure 7b and 7c, respectively. After 21 days, the fold difference in OP expression of Gel-RGD, Lam, Lam-HA, Lam-RGD, and Lam-RGD-HA groups was 12, 10, 12, 18, and 21, respectively, and that of OC was 297, 262, 288, 346, and 371. After 21 days of incubation, the OP expression of Lam-RGD-HA group was higher than Gel-RGD, Lam, and Lam-HA while the OC expression of lam-RGD-HA was higher than Gel-RGD and Lam. The fold difference in ON expression decreased with time as shown in Figure 8c, and Lam-RGD had the lowest ON expression (0.16 fold) after 21 days. The changes in the expression level of osteogenic markers with time, shown in Figure 8, are consistent with the reported expression levels for BMS cells on porous collagen-glycosaminoglycan scaffolds cultured in osteogenic media [64]. These results demonstrate that conjugation of RGD modulated osteogenic activity of the laminated composites by enhancing cell adhesion and spreading.

Figure 8.

mRNA expression levels (as fold difference) of osteopontin (a), osteocalcin (b), and osteonectin (c) genes of the BMS cells, seeded on different samples and cultured in osteogenic media, as a function of incubation time. Groups include Gel-RGD, Lam, Lam-HA, Lam-RGD, and Lam-RGD-HA. One star indicates statistically significant difference between the test group and Gel-RGD group for the same incubation time. Two stars indicate statistically significant difference between the test group (Lam-HA, Lam-RGD, or Lam-RGD-HA) and Lam group for the same time. Three stars indicate statistically significant difference between the Lam-RGD-HA group and Lam-HA or Lam-RGD group for the same time. Error bars correspond to means ± 1 SD for n = 3.

Engler and collaborators investigated the differentiation of BMS cells seeded on collagen I coated polyacrylamide gels with tunable degree of elasticity [65]. On soft gels with elastic modulus <1 kPa, BMS cells exhibited a branched filopodia-rich morphology while on stiff gels with modulus >25 kPa that mimicked the crosslinked collagen of osteoids, BMS cells exhibited a polygonal morphology similar to that of osteoblasts [65]. Furthermore, the transcription factor CBFα1, an early marker for osteogenesis, of the BMS cells was upregulated on stiffer gels compared to softer gels. They also showed that nonmuscle myosin II was implicated in elasticity directed differentiation of BMS cells to different lineages. Our results support the findings of Engler and collaborators as the ALPase activity (Figure 7b), calcium content (Figure 7c), and OP (Figure 8a) and OC (Figure 8b) expression of BMS cells were significantly higher on stiffer fiber-reinforced RGD-conjugated laminates compared to Gel-RGD (not reinforced with L-PLA fiber mesh).

The physical and mechanical properties of the laminates and the differentiation of seeded BMS cells can be further improved by varying the fiber fraction and orientation, hydrophilicity of the hydrogel, amount, size, and surface treatment of apatite nanocrystals, number of layers in the laminate, and conjugation of bioactive agents. Laminated composites with macropores can be produced by the addition of a porogen (e.g., sucrose crystals, gelatin microspheres, or sodium chloride crystals) [43, 66] to the hydrogel precursor solution followed by laminating, crosslinking, and leaching the porogen from the laminate. The macroporous composites could be attractive as a scaffold in a variety of applications in regenerative medicine to provide structural support and pore volume for integration with the surrounding tissue. These include osteochondral grafts for cartilage lesions [67], grafts for torn tendons [68], cardiac patches for regeneration of myocardium after myocardial infarction [69], and small diameter vascular grafts [70].

4. Conclusions

Lamination of L-PLA fiber mesh with PLEOF hydrogel/HA precursor solution produced a multi-functional substrate for bone formation. The fiber layers can provide structural support and dimensional stability to prevent the regenerating region from soft tissue collapse. For example, lamination increased modulus under wet condition dramatically (compared to the fiber mesh or hydrogel) to 570±130 MPa, within the range for that of wet human cancellous bone, while modulus of the fiber mesh (140±3 MPa) or hydrogel (0.50±0.07 MPa) was significantly less than that of cancellous bone. The hydrogel phase could support diffusion of oxygen and nutrients, solubilization of growth factors, and conjugation of bioactive peptides, as well as providing a matrix with bimodal degradation profile: a slow degradation phase for structural stability and a fast degradation phase to increase volume for cell migration and ECM production. For example, the laminate lost 44% of its mass after 4 weeks while the hydrogel completely degraded and the fiber mesh lost <10% of its mass after the same duration of time. The HA nanocrystals, suspended in the hydrogel phase, can provide an osteoconductive substrate for differentiation and mineralization of BMS cells. For example, BSM cells seeded on the laminates with 10% HA nanocrystals and conjugated with 1% RGD (Lam-RGD-HA group) has significantly higher ALPase activity after 14 days incubation in osteogenic media, higher calcium content after 21 days, and higher expression of osteogenic markers osteopontin and osteocalcin after 21 days, compared to other laminate groups (without RGD or HA), fiber mesh, or PLEOF hydrogel.