Vancomycin- and Poly(simvastatin)-Loaded Scaffolds with Time-Dependent Development of Porosity

Abstract

Biodegradable scaffolds are widely use in drug delivery and tissue engineering applications. The scaffolds can be modified to provide the necessary mechanical support for tissue formation and to deliver one or more drugs to stimulate tissue formation or for the treatment of a specific condition. In the current study, we developed biodegradable scaffolds that have the potential for dual drug delivery. The scaffolds consisted of simvastatin-containing prodrug, poly(simvastatin) entrapped in poly(β-amino ester) (PBAE) porogen particles and vancomycin encapsulated in poly(lactic-co-glycolic acid) (PLGA) microspheres, which were fused together around the PBAE porogens to create a slow-degrading matrix. Upon hydrolysis, poly(simvastatin) releases simvastatin acid, which has angiogenic and osteogenic properties, while the PLGA microspheres release vancomycin as an antibacterial agent. Degradation of PBAE porogens through hydrolysis of ester linkages led to the development of porosity in a controlled manner and led to water penetration that facilitated hydrolysis of PLGA. Higher porogen loading (~60% by weight) gave rise to ~70% interconnected porosity with pore spacing of ~180 μm. This open volume facilitated simvastatin acid release upon hydrolysis and entrapped vancomycin release via diffusion through and degradation of PLGA. During the study, ~162 μg of simvastatin acid and ~18 mg vancomycin were released from the highest porosity scaffolds. Bioactivity studies showed that released simvastatin acid stimulated preosteoblastic activity, indicating that scaffold fabrication did not damage the polymeric prodrug. Regarding mechanical properties, compressive modulus, failure strain, and failure stress decreased with increasing PBAE porogen content. These dual drug releasing scaffolds with controlled development of microarchitecture can be useful in bone tissue engineering applications.

1. Introduction

Simvastatin was originally developed to treat high cholesterol levels. This drug is administered in the form of an inactive lactone that is hydrolyzed to a pharmacologically active, open ring, hydroxyacid form.^1–2 Previous studies demonstrated that simvastatin can promote osteogenesis.^3–5 Our group developed and characterized the physical and chemical properties of a simvastatin-containing polymeric prodrug, termed poly(simvastatin).^4–6 Polymeric prodrugs have gained increased attention due to their higher drug loading, controlled release, enhanced stability, and easier processing.^7–9 We also demonstrated the bone-forming potential of poly(simvastatin) in vivo.¹⁰

Common scaffold fabrication processes employ organic solvents, higher temperatures, and acidic conditions, which can impede incorporation of bioactive molecules and drugs.^11–12 Paris et al. formed porous scaffolds of apatite/agarose composite for delivery systems releasing two therapeutic substances, zoledronic acid and ibuprofen, using freeze drying technique.¹³ Dual growth factor delivery was achieved by oligo(poly(ethylene glycol) fumarate) hydrogel scaffolds in presence of gelatin porogen microparticles.¹⁴ In a different approach, PLGA/β-tricalcium phosphate scaffolds used two different types of drug-loaded microspheres for bone regeneration by releasing dexamethasone and a model protein (bovine serum albumin).¹⁵ These materials, however, lack time-dependent development of porosity, which is vital for smooth transition from synthetic scaffold to regenerated tissue. Controlled porosity development allows cells to penetrate to the inner regions of the scaffold, maintaining structural integrity and allowing new tissue to take over from the degrading scaffold with sufficient vascularization for oxygen, nutrition, and waste transport.

The present study sought to use a different formulation strategy to deliver the simvastatin-containing polymeric prodrug in conjunction with an antibiotic. We designed scaffolds consisting of poly(simvastatin) entrapped in poly(β-amino ester) (PBAE) particles that were subsequently incorporated within a matrix of poly(lactic-co-glycolic acid) (PLGA) microspheres encapsulating vancomycin. For proof-of-concept, we were interested in two different actions: dual drug delivery and development of porosity. The entrapped vancomycin served as the antibacterial drug, and the poly(simvastatin) polymeric prodrug behaved as the osteogenic agent. This type of scaffold is especially beneficial in tissue engineering applications with incorporated active capable of stimulating tissue formation while also providing antibiotic to prevent infection. Degradation of the PBAE porogens also generates porosity in a time-dependent manner. To our knowledge, this is the first use of a polymeric prodrug in a hydrogel and PLGA matrix. By studying the full spectrum of drug release profile, in vitro bioactivity, degradation pattern, porosity development, and mechanical properties, future scaffolds can be designed to satisfy specific tissue engineering applications.

2. Experimental

Materials

All glassware was oven-dried before use and cooled under a constant stream of dry nitrogen. Simvastatin was purchased from Haorui Pharma-Chem (Edison, NJ). PLGA (50:50, IV: 0.55–0.75 dl g⁻¹, acid-terminated; was purchased from Durect Corporation, Pelham, AL). Vancomycin was purchased from GoldBiotechnology, St. Louis, MO. 2,2-dimethoxy-2-phenylacetophenone (DMPA), Triazabicyclodecene (TBD), 5 kDa monomethyl ether poly (ethylene glycol) (mPEG), anhydrous diethyl ether, Iso-butyl amine and dichloromethane (DCM) were obtained from Sigma-Aldrich (St. Louis, MO). Diethylene glycol diacrylate (DEGDA) and poly(ethylene glycol) 400 diacrylate (PEGDA) were bought from Polysciences, Inc. (Warrington, PA). Tetrahydrofuran (THF) stabilized with 3,5-di-tertbutyl-4-hydroxytoluene (BHT) was purchased from Fisher Scientific (Pittsburgh, PA). All materials were used as received. Polystyrene Medium EasiVials (2 mL) purchased from Agilent (Santa Clara, CA) were used as the GPC standards. MC3T3-E1 preosteoblastic cells (CRL-2593) were obtained from ATCC (Manassas, VA). Alpha-minimum essential medium (α-MEM) without ascorbic acid was purchased from Gibco. Premium grade fetal bovine serum (product no: 1500-500/lot: 350K15) was purchased from Seradigm. For the alkaline phosphatase assay, the phosphatase substrate was purchased from Sigma-Aldrich. For the MTT assay, thiazolyl blue tetrazolium bromide (MTT reagent) was purchased from Alfa Aesar. For the DNA quantification assay, calf thymus deoxyribonucleic acid sodium salt and bisbenzimide (H33258) fluorochrome were purchased from EMD Millipore Corp USA.

Poly(simvastatin) synthesis and preparation of poly(simvastatin)-containing PBAE porogens

Poly(simvastatin) was prepared using ring-opening polymerization in a dry nitrogen environment following the previously published procedure.⁶ Hydrogel particles composed of AH6 2:1 PBAE macromer were used in this study. Macromer was named according to the classification system from Hawkins et al¹⁶, which was adapted from that of Anderson et al.¹⁷ . In this case, the molar ratio DEGDA (A) to PEGDA (H) was 2:1, and the molar ratio of total acrylate to amine reactive group was 1.2:1. The diacrylate and amine were weighed out to meet the desired molar ratios and reacted in a sealed conical flask maintained at 85°C for 48 hours^{16, 18}

PBAE-entrapped poly(simvastatin)-containing hydrogel was formed by mixing 10% (w/w) poly(simvastatin) into macromer dissolved in 80% (w/w) DCM, followed by addition of 1% (w/w) DMPA photoinitiator. All weight percentages were based on the initial macromer mass. After adding 2 mL of the macromer-poly(simvastatin) mixture to a transparent Pyrex petri dish (internal diameter 5.1 cm), hydrogels were formed by UV photopolymerization via exposure to a UV flood source for 5 minutes on each side. The resulting hydrogels were soaked in ethanol for 5 minutes to remove unreacted components and then ground using a mortar and pestle to obtain particles that were sieved (<500 μm) for use as the porogen.

A Shimadzu Prominence LC-20 AB HPLC system connected to a Waters 2410 refractive index detector was used to measure the number average (M_n) and weight average (M_w) molecular weights of polymeric prodrug and the macromer. Two Resipore columns in series (300 × 7.5 mm, 3 μm particle size; Agilent Technologies) were used for separation. Samples were dissolved in THF at 1 mg/mL and injected using THF as the mobile phase at a 1.0 mL/min flow rate at 40 °C. The polystyrene standards used to calculate molecular weight ranged from 162 Da to 364 kDa. Molecular weights were M_n 14 kDa and M_w 24 kDa for poly(simvastatin) and M_n 3 kDa and M_w 5 kDa for the macromer.

Vancomycin-containing PLGA microspheres

Microspheres were prepared following the previously reported water/oil/water (W₁/O/W₂) double emulsion procedure.¹⁹ Briefly, vancomycin (0.6 g) was dissolved in 4 mL phosphate-buffered saline (PBS), pH 7.4, as the W₁ phase. PLGA (2 g) was dissolved in 15 mL of DCM as the organic or oil phase. The W₂ phase consisted of 3.5 g of methylcellulose surfactant and 50 g of NaCl dissolved in 600 mL of deionized water. The W₁ phase was added to the oil phase and sonicated three times for 20 seconds (amplitude 100%) with intervals to prevent excessive heating. The W₁/O mixture was poured into the W₂ phase while being homogenized at 4000 RPM for 5 minutes. The suspension was then stirred overnight to allow the DCM to evaporate. After centrifugation, the collected microspheres were lyophilized for 48 hours and then sieved to <250 μm before use.

Encapsulation efficiency and loading

The encapsulation efficiency of vancomycin in PLGA microspheres was measured by dissolving 30 mg of PLGA-entrapped vancomycin microspheres in 3 mL of DCM for 2 hours. After complete dissolution, 10 mL of PBS were added, vortexed for ~1 min, and then centrifuged at 2500 RPM for 5 minutes. Vancomycin in the PBS layer was quantified by measuring the absorbance at 280 nm using a PowerWave HT microplate spectrophotometer with Gen5 analysis software. The amount of vancomycin entrapped in PLGA microspheres was determined by comparison with a vancomycin standard curve. The percentage drug loading and encapsulation efficiency were calculated using as follows.

$$\text{\%}\text{Drug}\text{Loading}\text{=}\left(\text{Mass}\text{of}\text{drug}\text{in}\text{microspheres}\text{/}\text{Mass}\text{of}\text{microspheres}\right)\text{x}\text{100}$$

$$\text{\%}\text{Encapsulation}\text{Efficiency}\text{=}\left(\text{Mass}\text{of}\text{drug}\text{contained}\text{in}\text{known}\text{quantity}\text{of}\text{microspheres}\text{/}\text{Theoretical}\text{amount}\text{of}\text{drug}\text{in}\text{that}\text{quantity}\text{of}\text{microspheres}\right)\text{x}\text{100}$$

To determine the amount of poly(simvastatin) entrapped in hydrogel particles, a known quantity of sieved PBAE particles were immersed in PBS for 7 days to ensure full degradation of PBAE. The amount of simvastatin acid was measured using a Hitachi Primeaide HPLC system. One Luna C18 column (150×4.60 mm, 5 μm particle size) was used with an isocratic mobile phase of acetonitrile and water containing 0.1% trifluoroacetic acid (70:30 v/v) with a flow rate of 1.0 mL/min. Absorbance was measured at 238 nm. Simvastatin acid standards were prepared by dissolving simvastatin acid in PBS. Samples were filtered (0.45 μm) before injecting to the HPLC.

Scaffold Fabrication

A home-built, spring-loaded mold was used to prepare the scaffolds. Average dimensions of the scaffolds were ~6.22 mm diameter and ~5.33 mm height. Vancomycin-containing PLGA microspheres and poly(simvastatin)-containing PBAE particles were mixed at weight ratios of 60:40, 50:50 and 40:60 PLGA:PBAE. The mixture was gently ground to a more homogeneous distribution using a mortar and pestle and loaded into the mold wells at 200 mg each. The mold was then incubated for 48 hours at 50 °C to allow PLGA microspheres to fuse around the PBAE particles.

In vitro degradation and drug release

Scaffolds were immersed in 2 mL of PBS under continuous shaking at 37 °C, and the supernatant was collected and completely replaced with fresh PBS at predetermined time points. Simvastatin acid release was then measured by HPLC as described previously. Vancomycin release was quantified by determining the absorbance at 280 nm using a PowerWave HT microplate spectrophotometer with Gen5 analysis software as mentioned previously.

Swelling and mass loss

Swelling was determined by immersing the corresponding scaffolds in 2 mL PBS and weighing at pre-determined time points. The samples were quickly removed from the PBS, patted dry with a Kim-wipe, and weighed to determine the mass gained from the swelling in PBS. Mass loss was calculated by measuring the weights of lyophilized scaffolds at predetermined time points.

Microarchitecture

Microcomputerized tomography (microCT) was used to evaluate the morphology and microarchitecture of the lyophilized scaffolds. Samples were analyzed using a Scanco μCT 40 (SCANCO Medical, Switzerland) operating at high resolution (6 μm voxel size) and parameters of 55 kV, 145 μA, and 8W. Three-D images were reconstructed and microarchitectural parameters were determined using the manufacturer’s morphometry software (version 6). Scaffolds were evaluated in their dry state before degradation as well as after 3, 6, 9, 12, 15, and 18 days of incubation in PBS.

Bioactivity

MTT cell viability assay.

Cytotoxicity of the collected supernatants was assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Supernatants collected from days 3, 6, 12, and 18 were used for this study without any further dilution. Briefly, mouse preosteoblasts (MC3T3-E1, clone 4; passage number 7) cultured in α-MEM medium (without ascorbic acid) with 10% fetal bovine serum were seeded in a 96-well plate at 10,000 cells per well in 100 μL of medium and allowed to adhere overnight in an incubator (37 °C, 5% CO₂). Ten μL of supernatant filtered through 0.2 μm sterile filter was added to each well. Sterile PBS was used as the control. After 24 hours of incubation, 25 μL of freshly prepared MTT solution (5 mg/mL in PBS) was added to each well and incubated another 2 hours. After that, 100 μL of lysing buffer (20% sodium dodecyl sulfate in 50% n,n-dimethyl formamide, pH 4.7) was added to each well and incubated for overnight. The optical density of formazan was read at its maximum absorbance wavelength of 570 nm using a PowerWave HT microplate spectrophotometer with Gen5 analysis software. In addition, a gradient of simvastatin acid reference standards was prepared and used for a parallel MTT assay.

Alkaline phosphatase (ALP) activity assay.

The ability of released simvastatin acid to stimulate osteoblastic differentiation was assessed by measuring cellular ALP activity. The supernatants collected from day 3, 6, 12, and 18 were used for this study. All the supernatants were diluted using PBS until the final concentration of simvastatin acid was 10⁻⁷ M and filtered through 0.2 μm sterile filter. From a pilot study, and past literature,²⁰ we found that maximal ALP activity resulted when the simvastatin acid concentration was 10⁻⁷ M.

MC3T3-E1 cells (passage number 7) cultured in α-MEM medium (without ascorbic acid) with 10% fetal bovine serum were seeded into a 48-well plate at 40,000 cells per well in 500 μL of medium and allowed to adhere overnight in an incubator. The medium was then replaced with 500 μL α-MEM containing 1% FBS, 10 mM β-glycerophosphate, and 50 μg/mL of ascorbic acid. To this were added 50 μL of the corresponding diluted supernatants. Sterile PBS was used as the control. Medium with treatments was replaced every 2 days. After 7 days, the wells were washed twice with warm PBS and the cells lysed by adding 150 μL of high salt solution (0.05 M NaH₂PO₄, 2M NaCl, 2 mM EDTA, pH 7.4) to each well followed by sonication (20 seconds, amplitude 20%, pulse 10). An aliquot of 20 μL of lysate was mixed with 50 μL of ALP substrate solution (4 mM MgCl₂ and 10 mM p-nitrophenyl phosphate in 0.6 M 2-amino-2-methyl-1-propanol buffer, pH 10), and the absorbance at 400 nm was measured as a function of time. ALP activity was normalized by DNA content determined using the Hoechst DNA assay to eliminate any discrepancy arising from different numbers of cells present in the wells. After lysing the cells, bisbenzimide (H33258) was added at 0.5 μg/mL in the high salt buffer, and the fluorescence measured (λ_ex 356 nm, λ_em 458 nm). A standard curve consisting of calf thymus deoxyribonucleic acid sodium salt was prepared.

Mechanical testing

Compression testing was performed using a Bose Electroforce 3300 mechanical testing system. Lyophilized samples were used for the analysis and generally had dimensions of 7.95 mm diameter of 8.06 mm height. Uniaxial compression tests were run at 0.25 mm/s to 50% strain or failure, whichever occurred first. Stress-strain plots were constructed from the load-displacement data and physical dimensions of the samples. The modulus was determined from the slope of the initial linear elastic region of each curve.

Statistical Analysis

Analysis of variance (ANOVA) followed by post-hoc pairwise comparisons (with Bonferroni adjustment) was performed to test the significant differences in the treatment means. Values of p < 0.05 were deemed statistically significant.

Results

Drug loading in PLGA microspheres and PBAE hydrogels

The PLGA microspheres had vancomycin loading of 17.4% and encapsulation efficiency of 75%. The resulting scaffolds consisting of 60:40, 50:50, and 40:60 PLGA to PBAE thereby contained ~21, ~17, and ~14 mg of vancomycin, respectively.

According to the HPLC analysis, 10 mg of hydrogel particles released 30±0.57 μg of simvastatin acid after complete degradation. The scaffolds composed of 60:40, 50:50, and 40:60 PLGA to PBAE should therefore release ~240, ~300, and ~360 μg of simvastatin acid, respectively.

Microarchitecture analysis

Figure 1 shows cross-sections of the three scaffold types at increasing times of degradation. All scaffold types were initially solid masses with no distinct differentiation or porosity between them. The day 3 samples also mainly appeared as a solid mass, but the hydrogel particles were clearly visible with their boundaries indicating the initiation of degradation. In day 3 PLGA:PBAE 60:40 and 50:50 scaffolds, the porogen boundaries were more visible closer to the edges, whereas they were clearly visible throughout the scaffold cross-section of PLGA:PBAE 40:60 scaffolds.

Figure 1:

Cutplane images of microCT reconstructions of 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds during degradation. Scale bar is 1 mm.

As shown in Figure 2a, at day 6, the PLGA:PBAE 40:60 scaffolds showed rapid porosity development compared to the other two compositions. From day 3 to day 6, these scaffolds gained ~21% average porosity. The average porosity increment was 9% and 14% for the 60:40 and 50:50 PLGA: PBAE scaffolds, respectively. Overall porosity was 37±2%, 46±3%, and 55±2% for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds, respectively, on day 6. A majority of the PBAE porogens were degraded within the first 6 days in all three types of scaffolds. A less rapid but gradual linear increment in porosity was observed for 50:50 and 40:60 PLGA:PBAE scaffolds from day 6 onward. For 60:40 PLGA:PBAE scaffolds, the porosity increased even less rapidly until day 12 and then plateaued. From day 6 to day 18, the porosity increased by 5%, 17%, and 20% in 60:40, 50:50, and 40: 60 PLGA:PBAE scaffolds, respectively. During our 18 day study, the 60:40, 50:50, and 40:60 scaffolds showed final porosities of 43±2%, 63±4%, and 74±2%. respectively. There was a significant difference in porosity between each pair of scaffolds at days 6, 9, 12, and 18 (p<0.05). At day 15, there was a significant difference in porosity between the following pairs: 50:50 versus 60:40 PLGA:PBAE, and 60:40 versus 40:60 PLGA:PBAE (p<0.05).

Figure 2:

a) Percentage of porosity, b) pore size, and c) interpore thickness of scaffolds at different time points during degradation in PBS. Data are shown as mean ± standard error (n=3).

For all three scaffold types, the pore size increased and interpore thickness decreased during degradation (Figure 2b,c). At the end of the study, the scaffolds of 60:40, 50:50, and 40:60 PLGA:PBAE showed pore size of 135±19, 173±76, and 186±15 μm, respectively. There was a significant difference in pore size between 40:60 and 60:40 scaffolds at days 15 and 18. Similarly, pore sizes in 50:50 and 60:40 sizes were significantly different (p<0.05) from each other at days 15 and 18.

There was a rapid reduction in interpore thickness from day 3 to day 6; it decreased approximately 45, 30, and 37% relative to day 3 for 60:40, 50:50, and 40:60 PLGA:PBAE, respectively. After day 6, further decreases in interpore thickness were minimal. After 18 days, the 60:40, 50:50, and 40:60 scaffolds had interpore thickness of 98±24, 60±4 μm, and 34±3 μm, respectively. There were significant differences in interpore thickness between the pairs of 50:50 versus 60:40 PLGA:PBAE, and 60:40 versus 40:60 PLGA:PBAE at days 9, 12, and 15 (p<0.05). The interpore thicknesses in 60:40 versus 40:60 PLGA:PBAE were significantly different from each other at day 18 (p<0.05).

Dry and wet mass changes

The mass loss profiles of the scaffolds are depicted in Figure 3a. At day 3, all three scaffold types showed similar mass loss. At day 6, there was a rapid mass loss of ~26, 36, and 52% for the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds, respectively. The dry mass loss from day 3 to day 6 was ~14, 17, and 35% for the scaffolds 60:40, 50:50, and 40:60 PLGA:PBAE, respectively. After day 6, the percentage mass loss was ~18, 21, and 13% for the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds, respectively. From day 6 onward, the mass loss slowed and decreased in a relatively linear pattern for PLGA:PBAE 50:50 and 40:60 scaffolds. For the PLGA:PBAE 60:40 scaffolds, the mass remaining plateaued after day 12. Overall, remaining mass for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds was 56±2, 45±1, and 35±2% respectively, at day 18, and there was a significant difference in overall remaining mass between each pair of scaffolds.

Figure 3:

a) Degradation and b) swelling profiles for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds. Data are shown as mean ± standard error (n=3).

Figure 3b shows the swelling profiles. The scaffolds exhibited maximum swelling ratios of 128±3, 134±4, and 140±4% for 60:40, 50:50, and 40:60 PLGA:PBAE, respectively, on day 6, after which there was a gradual decrease of swelling. There were no statistically significant differences in swelling between the three types of scaffolds tested.

Drug Release

All three scaffold types showed two different kinetic behaviors for release of vancomycin without initial burst (Figure 4a). During the first 9 days, the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds released vancomycin at rates of 0.8, 0.9, and 1.0 mg/d, respectively. From day 9 to day 18, vancomycin was released at much lower rates of 0.16, 0.14, and 0.12 mg/d for 60:40, 50:50, and 40:60 scaffolds, respectively. Although the 40:60 PLGA:PBAE composition contained the smallest amount of vancomycin-loaded microspheres, the scaffolds released the largest amount of vancomycin, followed by 50:50 and 60:40 scaffolds. The 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds released ~41%, ~54%, and ~75%, respectively, of the total encapsulated vancomycin during the study. There was a significant difference in overall cumulative vancomycin release between 60:40 and 40:60 PLGA:PBAE for the 18 day period.

Figure 4.

Profiles for cumulative release of a) vancomycin and b) simvastatin acid from 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds. Data are shown as mean ± standard error (n=3).

According to Figure 4b, all three types of scaffold did not exhibit burst release of simvastatin acid and had similar cumulative release (~12 μg) during the first 3 days. After day 3, two different kinetic behaviors were observed. From day 3 to day 9, 60:40, 50:50, and 40:60 scaffolds released simvastatin acid at rates of 9.9, 12.4, and 20 μg/d, respectively. From day 9 to day 18, the scaffolds released simvastatin acid at much lower rates of 2.4, 2.7, and 8.5 μg/d for 60:40, 50:50, and 40:60 PLGA:PBAE, respectively, without plateauing. Release was significantly different between 50:50 vs. 40:60, and 60:40 vs. 40:60 scaffolds for each day starting from day 6 (p<0.05). Overall, the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds released 90, 113, and 163 μg simvastatin acid, respectively.

Cytotoxicity

Cytotoxicity followed a similar pattern for all three scaffold types (Figure 5), with higher cytotoxicity observed at days 3 and 6. Absorbance of formazan formed from release supernatants was significantly different from that for simvastatin acid standards and that of control (PBS) at the 3 and 6 d time points. The 12- and 18-day supernatants, however, had a negligible cytotoxic effect. The equivalent simvastatin acid concentrations were not cytotoxic at any of the four time points. Literature indicates that even high levels of vancomycin did not affect bone cell viability.²¹

Figure 5.

Corresponding cytotoxicity for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds. Data are shown as mean ± standard error (n=3).

For the ALP activity study, all supernatants were diluted to have a final simvastatin acid concentration of 10⁻⁷ M. Supernatants from all the three types of scaffold exhibited ALP activity (Figure 6). Compared to 3 and 18 day samples, greater ALP activity was observed for 6 and 12 day samples. At day 12, there was a statistically significant difference between 40:60 and 60:40 PLGA:PBAE scaffolds (p<0.05). Significant differences in overall ALP activity between scaffolds, however, were otherwise not observed for the 18 day study.

Figure 6.

Corresponding ALP activity for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds. Data are shown as mean ± standard error (n=3).

Mechanical Properties

For all three scaffold types, the modulus, failure strain, and failure stress decreased with increasing porosity (Figure 7). The zero day, undegraded samples showed a higher modulus compared to all the other time points. The majority of degraded samples displayed a compressive modulus between 2 and 10 MPa, but the undegraded samples displayed moduli ranging between 15 and 74 MPa, with high standard deviation within the series. Failure strain was highest for day 3 samples, but was otherwise relatively stable at approximately 5% for 60:40 and 50:50 scaffolds, and 2% for 40:60 scaffolds. Likewise, failure stress was higher for day zero samples, and gradually decreased to stable values of 430, 240, and 60 kPa for 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds, respectively.

Figure 7.

a) Compressive modulus, b) failure strain, and c) failure stress of 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds in their dry state. Data are shown as mean ± standard error (n=3).

Discussion

Dual drug delivery scaffolds were prepared by compressing different ratios of PLGA microspheres and PBAE hydrogel particles, wherein the PBAE particles behaved as the porogen in addition to releasing drug. The physical and mechanical properties of the scaffolds were evaluated to understand the effects of different amounts of PBAE porogens on the scaffolds. Ideally, scaffolds will provide mechanical support and degrade in concert with neotissue infiltrating the porous network.

Development of porosity

The PBAE hydrogel particles were prepared using AH6 2:1 macromer, which degrades over approximately 2 weeks. These PBAE particles behaved as porogens, with their degradation giving rise to development of porosity, which is important for the transport of nutrients, oxygen, and waste. The porous nature of the scaffold increases the surface area for protein adsorption and facilitates the interaction between scaffold and cells. Also, the capillary forces created by the porosity have been reported to improve bone cell attachment on the scaffold surface and facilitate penetration into the scaffold interior.^22–23 Initially, all three scaffold types were solid masses with negligible porosity, but degradation of PBAE particles resulted in increased porosity. The variation of porosity was due to the different PBAE porogen content in each scaffold.

There was a systematic increment of pore size and systematic reduction of interpore thickness, indicating well-controlled porous microstructure development occurring in the scaffolds’ interior. The scaffolds ultimately showed pore sizes in the range of 135-186 μm, well below to the PBAE porogen particle size (<500 μm). Porogen sieving was conducted using ethanol-soaked PBAE hydrogels followed by lyophilization, which explains why the embedded porogens and resulting pores had smaller sizes than sieved particles. Pore sizes above 100 μm and porosity higher than 75% are necessary for appropriate bone infiltration.²⁴

Drug release and mass loss

Along with the development of porosity, degradation of the scaffolds released entrapped drug molecules. The polymeric prodrug poly(simvastatin) hydrolyzed to release simvastatin acid, and the PLGA microspheres released entrapped vancomycin.

When considering vancomycin release, we did not observe the typical triphasic release profile expected with delivery from PLGA.²⁵ Vancomycin was used to provide the potential to prevent infection following implantation of the scaffolds. Compared to simvastatin acid, vancomycin was released rapidly and in higher amounts initially and then slowed, resulting in a biphasic release kinetics. One reason we did not observe the initial burst may due to the lower loading of the vancomycin or that vancomycin was fully entrapped inside the PLGA microspheres. As such, drug release happened only after diffusion of water molecules into the inner core of the microsphere. In addition, the numerous washing steps to remove the surfactants from the microspheres would have removed unencapsulated and weakly bound drug. A previous study involving PLGA (50:50) / risperidone implants also did not show the typical triphasic release profile of the drug molecules.²⁶ Higher initial burst release is a significant problem for many scaffold-based drug delivery systems. As an example, 20-50% of drugs have been reported to be released within 24 h in vitro, ^27–28 and in some cases, even up to 60-70%.^{2, 29} Due to this, controlled-release drug delivery systems have become increasingly important in modern medicine to achieve an optimal therapeutic level and reduce side effects. In this study, only 4-9% of entrapped vancomycin was released in the first 24 hours.

Relatively more vancomycin was released by the PLGA:PBAE 40:60 scaffolds compared to PLGA:PBAE 60:40. In PLGA:PBAE 40:60 scaffolds, due to the higher interconnected porosity, the water molecules could easily penetrate to the inner core of the scaffold and release more vancomycin from the PLGA microspheres through diffusion, although the scaffolds had lower initial vancomycin content. Later vancomycin release occurred following degradation of the PLGA matrix, which is a much slower process compared with diffusion. The amount of vancomycin release was correlated with the dry mass loss, assuming all the PBAE particles were fully degraded during this time period. This is a viable assumption because we did not observe PBAE porogens in the 18-day samples according to microCT analysis. Also, the mass loss was well correlated with the final porosities of the scaffolds. If we consider the final porosities of the scaffolds, then the estimated average mass loss of the scaffolds should be 86, 126, and 150 mg for the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds, respectively. These values closely resemble the total weight loss of the final scaffolds.

The initial simvastatin acid release likely came from the hydrogel porogens on the periphery of the scaffold. The bulk of the PBAE particles were fully encapsulated by smaller PLGA microspheres. Thus, the PLGA microspheres served as a barrier for water penetration, lowering swelling and slowing the subsequent degradation of PBAE. This lowered simvastatin acid release during the first 3 days. But once water diffused into the scaffold interior, PBAE swelling and degradation increased, with higher PBAE contents giving rise to greater porosity and hydrolysis of entrapped poly(simvastatin). Hydrolysis of the scaffolds generated an acidic environment due to the PBAE and PLGA degradation byproducts, which catalyzed ester hydrolysis of the poly(simvastatin) prodrug. From day 3 onward, the scaffolds showed, two different kinetic behaviors of simvastatin acid release. We observed rapid porosity and swelling increment after day 3, which peaked at day 6 with higher simvastatin acid release. Due to water penetration followed by degradation, acidic byproducts accumulated initially inside the scaffolds, further catalyzing poly(simvastatin) hydrolysis and facilitating the corresponding release of simvastatin acid. As porosity increased at later time points, generated protons could easily diffuse from the scaffold interior to the bulk solution, decreasing internal acidity and decreasing the rate of simvastatin release. Assuming complete hydrolysis of the poly(simvastatin) polymer backbone, the 60:40, 50:50, and 40:60 PLGA:PBAE scaffolds should theoretically release 240, 300, and 360 μg of simvastatin acid, respectively. Experimentally, however, the scaffolds released 90, 114, and 163 μg of simvastatin acid, respectively, indicating that the polymer chains were not fully degraded within the time scale of our experiment. Poly(simvastatin) ester hydrolysis is a relatively slow process according to our previous reports.^4–6 The drug release profiles can be tuned by altering the amount of drug molecules entrapped in the polymer matrix or by changing the composition of the drug-loaded microspheres or porogens in the scaffolds.

Cytotoxicity and bioactivity

The highest cytotoxicity was observed in day 3 supernatants, independent of the scaffold type. The initial accumulation of unreacted starting materials, degradation byproducts, and other impurities, such as remaining surfactants, may be the reason for this observed trend. At the later time points, however, most of the degradation products and entrapped drugs were washed out from the scaffolds, so we did not observe any cytotoxicity for the day 12 and 18 samples. To determine whether any cytotoxicity occurred because of simvastatin acid alone or because of vancomycin plus degradation by products, we compared the effects of scaffold supernatants with reference samples containing only the same concentration of simvastatin acid. (Figure 5). Because the simvastatin acid alone at similar concentrations to those in supernatants were not cytotoxic, we can conclude that the cytotoxicity is related to release of vancomycin or scaffold degradation byproducts. Because the osteogenesic potential of the polymeric prodrug was of primary interest, we focused on cytotoxicity of the released simvastatin acid, although the scaffolds release both vancomycin and simvastatin acid. Preosteoblastic MC3T3-E1 cells exhibited a reduction in cell migration in hydroxyapatite scaffolds for vancomycin concentrations higher than 1 mg/mL, but there was no difference at lower concentrations.^30–31

Alkaline phosphatase was measured to determine the ability of released simvastatin acid to stimulate osteoblastic differentiation in preosteoblastic MC3T3-E1 cells. All three scaffold types stimulated ALP activity at the corresponding time points following dilution to 10⁻⁷ M, a concentration based on the literature²⁰ and a pilot study. Basically, bioactivity of simvastatin acid was not lost during incorporation of poly(simvastatin) into and release from the PBAE porogens and PLGA microsphere matrices. ALP activity is an early marker of osteoblastic differentiation, and, additional testing should include expression of osteocalcin and bone sialoproteins as well as calcium deposition to fully evaluate osteoblastic differentiation.³²

Mechanical properties

As-prepared PLGA:PBAE scaffolds displayed high mechanical properties prior to PBS exposure, but after immersion these mechanical properties rapidly decayed and plateaued after approximately 6 days of immersion in PBS. Once these properties stabilized, the relative mechanical properties were strongly correlated with initial PLGA content.

The markedly high mechanical properties displayed by the undegraded samples may be attributable to the fabrication procedures used for the hybrid scaffolds: the dry, heated polymer fusion process produced highly compact cylinders with smooth and continuous outer surfaces. This intact outer layer likely contributed significant mechanical strength to the samples by acting as a unified bounding surface that equally distributed load through the samples. Once this layer was compromised, imperfections in the sample could more readily act as stress concentrators and cause the sample to yield more readily; the high variation between zero-day samples may be explained by partial damage to the sample sides during extraction and handling.

Scaffold integrity is likely disrupted after liquid immersion and PBAE swelling and degradation. Significant differences in the behavior of PLGA and PBAE materials used in this study in response to liquid exposure are likely causal factors in ultimate scaffold behavior. PLGA does not significantly swell or degrade over the timescale observed, while PBAE particles both begin measurable degradation and swell significantly within a fraction of the observed period. PBAE particles have been previously observed to swell upwards of 50% in every dimension, producing over a 125% in volume in unrestricted settings. Thus, in this entrapped particle setting, PBAE swelling likely produces internal pressures that expand the PLGA lattice, introducing void space, decreasing effective sample density, and disrupting finer PLGA network branches.

Subsequent drying processes reduced swollen particles to original or lesser, degraded sizes, leaving significant void space within samples in which PBAE particles move independently and consequently fail to mechanically contribute prior to resisting significant deformation of the PLGA network and potential failure. This effect appears to become fully pronounced by the 6-day mark, after which sample mechanical properties plateaued. The anomalous failure strains displayed by 3-day samples are likely due to partial pore expansion or particle degradation insufficient to prevent PBAE particles from contributing to stress distribution. At this point, the PLGA network may be able to come into contact with the entrapped PBAE particles and distribute load continuously through the sample, resisting sample failure. The ultimate mechanical properties of PLGA:PBAE scaffolds appear to be largely dictated by their PLGA content, network density, and base mechanical properties, and are only significantly affected by PBAE content within the first 6 days.

Trabecular bone shows a compression modulus in the range of 50-78 MPa and articular cartilage has a modulus of 0.079-2.1 MPa.^33–34 Our undegraded scaffolds had moduli comparable to that of trabecular bone, and the compressive modulus of degrading scaffolds exceeded the modulus for articular cartilage. The resulting lower modulus for degrading scaffolds having higher PLGA microsphere content may be due to the lack of PLGA microsphere fusion. This can be solved by increasing the sintered sintering temperature further above the glass transition temperature of the PLGA or by using a more efficient heating and cooling setup for sintering the scaffolds.

Conclusion

The present study demonstrated how different contents of drug-loaded porogen can affect drug release and control microarchitecture development in PLGA matrix scaffolds. Here, we used the polymeric prodrug poly(simvastatin) as the drug molecule entrapped in PBAE porogen as a novel formulation strategy. Compared to traditional drug molecules, polymeric prodrugs have advantages related to loading/active content, release kinetics, stability, and formulation. Localized delivery of vancomycin provides the peri-scaffold environment with antibacterial activity while simvastatin acid enhances osteogenesis. The scaffolds were able to control release the drugs without the initial burst release and maintained appropriate mechanical properties and microarchitecture for bone tissue engineering applications.