Drug impregnation for laser sintered poly(methyl methacrylate) biocomposites using supercritical carbon dioxide

The Journal of Supercritical Fluids 136 (2018) 29–36

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids journal homepage: www.elsevier.com/locate/supflu

Drug impregnation for laser sintered poly(methyl methacrylate) biocomposites using supercritical carbon dioxide

T

⁎

Truc T. Ngoa, , Savanna Blaira, Katie Kuwaharab, Drew Christensena, Isabel Barreraa, Myles Domingoa, Sarat Singamnenic a

Shiley-Marcos School of Engineering, University of San Diego, 5998 Alcala Park, San Diego, CA, USA College of Arts and Sciences, University of San Diego, 5998 Alcala Park, San Diego, CA, USA c Auckland University of Technology, Auckland, New Zealand b

G RA P H I C A L AB S T R A C T

A R T I C L E I N F O

A B S T R A C T

Keywords: Supercritical carbon dioxide PMMA Selective laser sintering Drug impregnation Biocomposites

Multi-functional biocomposites were developed using a 3D-printing method coupled with supercritical carbon dioxide (scCO2) processing. Selective laser sintering was used to prepare poly(methyl methacrylate)/β-tricalcium phosphate biocomposites, which were subsequently treated with an anti-inflammatory drug (flurbiprofen) in scCO2. Drug impregnation behaviors were investigated under different temperatures and pressures (313 K and 323 K; 85–115 bar). Results show that drug molecules were able to diffuse into the biocomposite structure effectively, achieving > 26% by weight of drug uptake. Surface morphology of the materials remained unaffected upon pro-longed scCO2 processing. Drug release data was modeled using the Weibull function, indicating potential influences of scCO2 processing temperature on the PMMA surface and the interaction between flurbiprofen particles and PMMA matrix inside the composite structure.

1. Introduction In recent years, the rapid evolution of biomaterial processing has greatly expanded the functional diversity of composite products. A variety of polymeric materials have been developed, both biodegradable and biostable, for utilization as dental and orthopedic replacement devices, anti-microbial bone cements, as well as 3D scaffolding networks capable of promoting new tissue formation, amongst other

⁎

applications [1]. Commanding increased attention, especially amongst pharmaceutical companies, is the potential use of biocomposite materials as controlled drug release systems (CRDS) [2,3]. CRDS involve biomaterials internally and externally infused with various anti-biotic/ anti-inflammatory pharmaceuticals and have been tested to treat infections such as osteomyelitis and ophthalmic inflammation associated with surgical procedures [4,5]. Poly(methyl methacrylate) (or PMMA), an acrylic, hydrophobic,

Corresponding author. E-mail address: [email protected] (T.T. Ngo).

https://doi.org/10.1016/j.supflu.2018.01.030 Received 1 December 2017; Received in revised form 29 January 2018; Accepted 31 January 2018 Available online 02 February 2018 0896-8446/ © 2018 Elsevier B.V. All rights reserved.

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

together with the scCO2 processing for drug infusion and delivery. Flurbiprofen, an anti-inflamatory drug, was selected due to its favorable solubility in scCO2 without the use of additional co-solvents [16]. It is also compatible with PMMA-based copolymers, especially those prepared by traditional polymerization method [4]. The behavior of flurbiprofen in solutions during the impregnation process and the drug release process was characterized by UV/vis spectroscopy. Drug release profile was analyzed and surface morphology of the treated biocomposites was examined for any impact of scCO2 processing on the material.

biostable polymer, has been identified in the literature as a highly suitable vessel for local drug delivery due to its excellent compatibility with human tissue and existing use for medical implants [6]. In addition, the fully amorphous structure and thermo-plasticity of PMMA allow for enhanced control over product dimensions and homogeneity of the overall structure [7,8]. The drug release profile of PMMA bone cements and other commercial PMMA-based devices depend on the porosity, surface extension, initial drug concentration, and thermostability of the compound [9]. Of elevated significance is the effect of inherent porosity of the polymer matrix; enhanced porosity is correlated with increased penetration of dissolution fluids into the matrix, thus resulting in higher quantities of drug elution from the device into the body [10,11]. To achieve higher internal porosity structure, PMMA-based composites enriched with β–tricalcium phosphate (β-TCP) are being investigated [11]. β-TCP is a biocompatible, osteo-conductive, porous ceramic commonly used as a bone substitute [6]. It invokes a unique biological response during bone remodeling in which the resorption of old bone minerals and formation of new bone occur simultaneously. During this process, the degraded calcium and phosphate ions are naturally metabolized, increasing the biodegradability of the ceramic [12]. In this way, PMMA/β-TCP biocomposites are able to maintain mechanical stability while simultaneously providing partial biodegradable character to mediate the release of bioactive drug molecules. In this study, PMMA/β-TCP biocomposites were prepared using the most appropriate and powder-based 3D-printing method, the selective laser sintering (SLS) technique, for potential use as a polymeric CRDS as previously developed by Velu and Singamneni [7]. SLS allows researchers and practitioners to have theoretically unlimited control and freedom over product customization and meso- or macrostructures, including interconnected porosity. In SLS, a CO2 laser is typically used to melt the polymer particles and sinter them into the desired 3D form from the bottom up, one layer at a time. To form the biocomposite, the ceramic powders are premixed with selected polymer particles before printing. The ceramic powders act as the functional, bioactive agent embedded within the polymer matrix. The ceramic phase also reinforces the polymer matrix and influences the overall strength of the biocomposite [13]. This novel biocomposite preparation method maintains structural integrity of the composite as well as homogeneity of the pores, which is essential for even drug distribution within the biomaterial. To further enhance the composite and give it the multi-functional property, supercritical carbon dioxide (scCO2) was used to impregnate an anti-inflammatory drug, in this case flurbiprofen, into the PMMA/βTCP composite structure. CO2 is a common greenhouse gas, with relatively low critical temperature and pressure (Tc = 304 K, Pc = 73.8 bar). In supercritical state, CO2 can act as a non-toxic solvent to chemical compounds that are normally insoluble in liquid CO2. For example, many anti-inflammatory drugs, such as triflusal, ketoprofen, piroxicam, nimesulide and flurbiprofen, have been shown to exhibit reasonable solubility in scCO2 within 303–323 K temperature range [14–16]. ScCO2 can also act as a carrier to transport material from one phase to another. It has been used widely as a benign medium to enhance material properties, especially in polymers and composites [17–19]. Utilization of scCO2 not only eliminates the need for toxic organic solvents but also has the potential to recycle the greenhouse gas back into the process as a closed-loop system. Moreover, this method lowers the risk of leaving residual toxic solvents inside the materials, and prevents thermal tension induced during the material formation process by conventional methods [20]. Evidently, 3D-printing methods allow building biocomposites of controlled porosity, while scCO2 processing is effective in terms of loading and dissipation of drugs. A suitable combination of the two techniques can be envisioned as an innovative and effective means of achieving enhanced drug delivery systems. This study aims at such a solution, coupling selective laser sintering of biocomposite powders

2. Materials and methods 2.1. Materials Poly(methyl methacrylate) (or PMMA) was purchased from Aldrich in powder form, with average particle size of 75 μm, melting and glass transition temperatures of 433 K and 379 K, respectively, and theoretical density of 1180 kg/m3. β-Tricalcium phosphate (or β-TCP, Ca3(PO4)2) powders were also purchased from Aldrich, with average particle diameter of 5 μm and molecular weight of 310.18 g/mol. Flurbiprofen (C15H13FO2, CAS 5104-49-4) was obtained from Sigma in white powder form, with melting point range of 383–385 K, and used as is without further purification. ACS reagent-grade liquid ethanol (Arcos Organics, CAS 64-17-5) with purity of 99.5% was used as the solvent for drug dissolution. USP medical grade carbon dioxide, purchased from Airgas with > 99.9% purity, was used in all scCO2 treatments. 2.2. Sample preparation The laser sintering experimental setup developed by Velu and Singamneni [7] was used to print the specimens. The test setup used a CO2 laser of 10.5-μm wavelength, which allowed higher absorptivity with most polymeric materials, including PMMA. Laser sintering results of PMMA and β-TCP previously reported by Velu and Singamneni [13] were used to identify the best composition and process parameter settings for the current experiments. Based on previous coalescence and porosity data, the 95% PMMA + 5% β-TCP composition was selected for the current work, with energy density of the laser set at 0.15 J/mm2. A laser power of 38 W and scan speed at 450 mm/s allowed to achieve this energy density and the consolidation of the PMMA/β-TCP composites to the best possible extent. The samples were sintered with five layers each, and approximately 10 mm × 20 mm in size. Some samples were cut in two halves using sharp scissors prior to scCO2 processing whereas some others were processed as a whole piece. The pre-sintered thickness of the spread powder was around 50–60 μm, which was finally reduced to approximately 30–40 μm due to further interaction with the laser energy input. The porosity of the sintered samples was estimated to be at around 50%, based on the evaluation of SEM images. 2.3. Supercritical carbon dioxide processing Supercritical CO2 (or scCO2) processing of PMMA/β-TCP biocomposite samples was performed in 316-stainless steel cells (volume of 53–60 mL), equipped with calcium fluoride windows for in situ UV/vis spectroscopic monitoring. The experimental setup is similar to the scCO2 processing setup previously reported by Ngo et al. [21]. Bicomposite samples were first loaded into the cell for each experiment. One experimental run (313 K and 85 bar) was performed on PMMA/βTCP sample with scCO2 only and in absence of flurbiprofen, to study the effect of scCO2 alone on the biocomposite surface morphology. The remaining experiments were performed in presence of flurbiprofen. The amount of flurbiprofen used for each experiment was calculated to be at least three times the solubility limits of the drug under each pressure and temperature condition. This was to ensure excess drug availability for absorption into PMMA/β-TCP biocomposite samples during 30

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

Fig. 1. Calibration curve for flurbiprofen solubility in ethanol.

not a limiting factor throughout the process. ScCO2 treatment under set conditions was maintained for 24 h. Although the system was able to reach equilibrium after approximately five hours, samples were left in scCO2 for 24 h for convenient monitoring and verification of sample stability under pro-longed treatment conditions. At the end of the experiment, the cell was first allowed to cool to room temperature. It was then slowly depressurized to prevent foaming of PMMA samples as previously observed by Lopez-Periago et al. [14]. Upon completion of cell depressurization, samples were removed from the cell. Loose drug particles were removed from the samples by applying low-pressure compressed air to all sample surfaces. Samples were weighed and compared with weights obtained prior to scCO2 treatment. Samples were then prepared for post-processing, including scanning electron microscopy (SEM) imaging or ethanol soak for drug loading quantification or drug release profile characterization.

treatment. For example, the initial amount of drug added to the cell ranged from 8.8 to 28.6 mg, depending on scCO2 processing conditions. Masses of PMMA/β-TCP samples varied between 38.4 and 84.3 mg, depending on the size of the sample. Flurbiprofen solubility data in scCO2 was based on Duarte et al. study [16] for pressure of at least 98 bar at 313 K and 323 K. For lower pressure range, solubility of flurbiprofen in scCO2 was performed using UV/vis spectroscopic experimental method developed by Ngo et al. [22]. After samples and drug were loaded, the cell was purged with CO2 gas for seven minutes to displace air from the cell interior. The cell was then closed immediately after purging was complete, and cell heating was started. Liquid CO2 was pumped into the cell using a SFT-10 CO2 pump (Supercritical Fluid Technologies, Inc.). Two temperatures were tested in the experiments: 313 K and 323 K, along with three pressure settings: 85 bar, 100 bar and 115 bar. Safety caution must be considered when working with high-pressure conditions as optical windows and other removable components may fail during the experiments. CO2 inlet flow was stopped once the cell reached both temperature and pressure respective set points. UV/vis spectroscopy (USB 2000+ Ocean Optics UV–vis spectrometer with UV-transparent optical fibers) was used to monitor the dissolution of flurbiprofen into scCO2 as the treatment progressed. UV absorbance of flurbiprofen was detected between 200 nm and 300 nm range. A shift in UV peak maximum of flurbiprofen in scCO2 solution was observed as fluid pressure changed. For example, at 323 K, UV peak maximum of flurbiprofen in scCO2 appeared at 247 nm at 100 bar, which was relatively close to peak detection of the drug when dissolved in ethanol. However, the peak maximum shifted to a shorter wavelength (blue shift) of 241 nm when the system was at 85 bar, and shifted to a longer wavelength (red shift), as high as 268 nm, when the system pressure was raised to 115 bar. Although some spectral shift due to changes in fluid density was expected [23], bond conjugations were most likely the dominant driving force for the significant red shift in UV absorption of flurbiprofen in scCO2. As pressure increased, flurbiprofen molecules were forced closer to one another in solution, increasing the chance for pi-bond conjugations and causing the absorption maximum shift to longer wavelengths [24]. UV peak absorbance was observed to increase initially as more flurbiprofen transitioned from solid phase into scCO2 fluid phase, but eventually leveled off when the system reached saturation limit. There was no observation of any decrease in UV absorbance of flurbiprofen in scCO2 fluid phase during the impregnation, confirming that the initially loaded drug amount in the cell was in excess and drug availability was

2.4. Material characterization Scanning electron microscopy (SEM) was used to characterize surface morphology of PMMA/β-TCP samples before and after scCO2 treatment. The samples were first coated with a 60/40 gold/palladium alloy for two minutes at 25-mA current setting in an argon-rich sputter chamber, before being scanned on the Hitachi S-3400 N scanning electron microscope. For more accurate quantification of drug absorption inside the biocomposites, scCO2-treated PMMA/β-TCP samples were soaked in 80 mL ethanol for at least 24 h at room temperature with constant stirring. The amount of extracted flurbiprofen in ethanol solution was then determined using UV/vis spectroscopy and Beer-Lambert law, averaged over 24 scans and 1-s integration time. To do this, a calibration curve was first established to correlate flurbiprofen concentration in ethanol solution with UV absorbance at λmax = 249 nm with three repeated measurements for each data point. Calibration result is shown in Fig. 1, with a linear best fit trendline. Drug release profile characterization was performed on selected samples, for different scCO2 processing conditions. To do this, scCO2treated samples were also soaked in 80-mL beaker of ethanol at room temperature. A small amount of solution was drawn from the beaker at approximately every 30 min, and measured on the UV/vis spectrometer. Three repeated measurements were performed each time for data validation. After each draw, the soak solution was replenished with pure ethanol to maintain a consistent 80 mL of total solution volume throughout the experiment. Further dilution was done to the 31

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

samples were free of foaming, there was no noticeable impact of scCO2 processing alone on sample surface, considering both PMMA and β-TCP phases. Considering the point-by-point consolidation mechanism, the laser sintered PMMA/β-TCP specimens are bound to be porous. The current specimens were estimated to be at approximately 50% porosity level. When flurbiprofen was added to scCO2 treatments, scCO2 dissolved flurbiprofen into solution, acting as a carrier to migrate the drug molecules inside the composite structure and layers. Concurrently, scCO2 was capable of swelling PMMA structure [19], providing pathways for more drug molecules to diffuse into the surface of PMMA particles. SEM analysis was performed on various PMMA/β-TCP samples treated in scCO2 and flurbiprofen under different temperature and pressure settings. Fig. 3 shows an example of SEM images for samples treated at 115 bar for both 313 K and 323 K. Result shows no apparent changes on surface integrity of the biocomposites under any of the tested conditions. These observations suggest that scCO2 treatment of PMMA/β-TCP biocomposites had insignificant impact on composite structure, and was able to preserve shape and size of the material.

drawn solution as necessary, depending on the concentration of the drawn solution at the time of measurement. UV absorbance for each measured solution was then converted to flurbiprofen concentration using the best fit linear equation shown in Fig. 1. Consequently, the amount of drug presence in the stock solution was calculated. Drug release profile was typically performed continuously for five to seven hours, with the last data point collected at the 24 h mark. 3. Results and discussion 3.1. Drug solubility in scCO2 In order to ensure proper treatment of PMMA/β-TCP biocomposites with flurbiprofen in scCO2, the drug needs to exhibit decent solubility behaviors in scCO2 solution at the tested temperature and pressure ranges. Duarte et al. reported flurbiprofen solubility in scCO2 at 303 K, 313 K and 323 K, for pressure range from 97 bar to 241 bar, measured using UV/vis spectroscopy [16]. To confirm the extrapolation validity of Duarte’s data to lower pressure ranges, a separate experiment was performed to check flurbiprofen solubility in scCO2 at 88.4 bar and 313 K, using the UV/vis spectroscopic experimental method outlined in Ngo et al.’s work [22]. UV absorbance for flurbiprofen was detected between 245 nm and 250 nm wavelengths, depending on the corresponding fluid density. Flurbiprofen solubility in scCO2 at 88.4 bar and 313 K was determined to be 0.022 g/L, which is less than 15% in deviation from the extrapolated value using Duarte’s data. Based on this result, it was concluded that Duarte’s solubilitycould be reliably extrapolated to determine the proper amount of drug to load inside the high-pressure cell for each scCO2 treatment of PMMA/β-TCP samples. However, Duarte’s solubility data extrapolation would result in a negative value for 323 K and 85 bar condition. Therefore, the solubility of flurbiprofen in scCO2 for 323 K and 85 bar setting was alternatively estimated using the proportional relationship between the measured UV peak absorbance of flurbiprofen in scCO2 and Duarte’s extrapolated solubility data at 313 K and 85 bar, assuming negligible change in the molar absorptivity of solution between the two processing conditions. Also, as previously stated in the Materials and Methods section, an excess of flurbiprofen was added to the high-pressure optical cell with at least three times the solubility limit for each of the PMMA/β-TCP treatment runs in scCO2. This ensured that drug loading in biocomposite samples was not limited by drug availability.

3.3. Temperature and pressure effects on drug loading ScCO2 treatment in presence of flurbiprofen was performed at three different pressures: 85 bar, 100 bar and 115 bar for both 313 K and 323 K. Three to five runs were carried out under each set of conditions to check for data repeatability. Treated samples were soaked in ethanol solution for at least 24 h and total drug amount extracted from the treated samples was determined using UV/vis spectroscopy, as outlined in the Experimental Methods section. Fig. 4 shows drug loading results for all tested processing conditions. Percent drug loading was calculated using Eq. (1): % Drug loading = Mass of drug uptake/Mass of sample prior to processing × 100% (1) Data shows that flurbiprofen uptake in SLS-prepared PMMA/β-TCP composites was comparable to other solute uptakes previously reported in literature using conventionally prepared PMMA under similar scCO2 processing conditions. For examples, Lopez-Periago et al. reported 15 wt.% of triflusal in PMMA [14]; Elvira et al. achieved up to 20 wt.% of cholesterol loading in PMMA spheres [27]; Üzer et al. showed that up to approximately 25 wt.% of naphthalene can be loaded in PMMA matrix at 150 bar and 313 K [28]. There was a strong correlation observed between drug loading of the biocomposites and scCO2 processing pressure, under both temperature settings. Single-factor ANOVA tests found that with 95% confidence, the differences in amount of drug loading at the three tested pressures were statistically significant, with F = 54 > Fcrit = 4 (p = 1 × 10−5 < .05) for 313 K runs, and F = 74 > Fcrit = 5 (p = 2 × 10−5 < .05) for 323 K runs. As pressure increased from 85 bar to 115 bar, PMMA/β-TCP biocomposite samples were able to uptake almost three times more drug amount at 313 K, and almost eight times more at 323 K. This trend is expected. First, PMMA swelling induced by scCO2 treatment has been shown by Üzer et al. to increase with pressure [28], allowing more available space for drug molecules to diffuse into the matrix. Secondly, as pressure increased at constant temperature, scCO2 fluid became denser, leading to improved intermolecular interactions between solvent and solute molecules, as observed in Duarte et al. study [16]. The increase in solubility of flurbiprofen in scCO2 increased the supply of drug molecules available for diffusing into PMMA/β-TCP composite structure. The drug was able to enter either through the porous composite structure or PMMA surface absorption facilitated by CO2 swelling. Because drug was supplied in excess, PMMA/β-TCP’s drug uptake was only dependent on drug solubility in scCO2, and the availability of receptor sites in the biocomposite structure. As previously noted, during the drug impregnation process

3.2. Surface morphology Surface morphology of the laser sintered PMMA/β-TCP samples before and after scCO2 treatments was characterized by SEM. Prior to processing, the laser sintered PMMA/β-TCP samples constitute of layers made of partially or fully coalesced round PMMA particles of various sizes, with the irregularly shaped β-TCP particles scattered within PMMA matrix, as described by Velu and Singamneni in 2014 [13]. To check for the effect of CO2 alone on the surface integrity of the composite, a PMMA/β-TCP sample was processed with pure scCO2 at 313 K and 85 bar. Fig. 2 compares composite surface before and after pure scCO2 treatment, under different cell depressurization rate at the end of the experiments. Result shows that when CO2 depressurization was carefully controlled at the completion of each experimental run, no foaming was observed in PMMA particles. However, when depressurization was done at faster rate, CO2 molecules, which previously diffused into PMMA to induce swelling of the polymer structure, escaped the polymer too rapidly, causing foaming on the outside surface of PMMA particles. This observation is consistent with other studies reported in literature for PMMA and other polymers, and was often used to its advantage to form various polymer foam structures [14,25,26]. For the purpose of this work, foaming of PMMA is undesirable and must be avoided. Therefore, CO2 depressurization was controlled carefully, with CO2 escape rate of no more than 0.1 bar/min. When processed 32

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

Fig. 2. SEM of PMMA/β-TCP sample with pure scCO2 treatment at 313 K and 85 bar: (a) pre-processing, (b) post-processing with rapid depressurization, (c) post-processing with controlled/slow depressurization.

Fig. 3. SEM of PMMA/β-TCP samples treated with flurbiprofen in scCO2 at 115 bar.

the UV absorbance of flurbiprofen in scCO2 increased initially and leveled off eventually. No reduction in peak intensity was observed throughout the experiment. These observations indicate that phase partitioning plays an important role in the drug impregnation, similar to findings reported in Shen et al.’s study in 2008 [29]. Upon initial pressurization, scCO2 swells up PMMA matrix, creating separate phases which the drug can diffuse to and partition in. As drug molecules start to dissolve into scCO2 fluid phase, drug partitioning becomes active between scCO2 fluid phase and PMMA/β-TCP composite phase. This process continues until both composite matrix and scCO2 phase are saturated with drug molecules and the system equilibrium is reached. Considering temperature effect on drug uptake, lower temperature seemed to have resulted in higher drug uptake by the biocomposites at

lower CO2 pressures (e.g. 85 bar). However, the trend reversed at higher pressures (e.g. 115 bar). Two-tailed t-tests were performed to compare drug uptake levels at three pressure settings (85 bar, 100 bar, 115 bar) under constant temperature (313 K or 323 K). For 85 bar, drug uptake by the biocomposite samples at 313 K was significantly higher than the uptake level at 323 K (p = .003 < .05). On the other hand, for 115 bar, drug uptake by the composite samples at 313 K was lower than the uptake level at 323 K (p = .036 < .05). These differences were statistically significant with 95% confidence. For 100 bar, the difference in drug uptake levels at 313 K and 323 K was essentially insignificant, with two-tailed t-test showing p-value p-value of .071 with 95% confidence. The observed effect of temperature on drug uptake by the biocomposite samples under constant pressure might seem random;

Fig. 4. Drug loading for PMMA/β-TCP samples under various scCO2 processing conditions.

33

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

the composition of PMMA and β-TCP mixture, power setting and laser scan speed during the SLS 3D-printing process [7]. The results obtained in this study present high potentials for drug loading optimization in PMMA/β-TCP biocomposites for specific biomedical application needs. This will be the focus of follow-up investigations with a wider range of processing parameters and varied porosity levels inside the composite materials.

however, this trend was indeed consistent with the solubility behavior of flurbiprofen in scCO2. According to Duarte et al., flurbiprofen was more soluble in scCO2 at lower temperatures for pressures less than approximately 110 bar [16]. However, when pressure rises above about 110 bar, flurbiprofen becomes more soluble in scCO2 at higher temperatures. This behavior can be explained by several competing factors that affect the ability of flurbiprofen molecules to dissolve in scCO2 solution: solute vapor pressure, solvent density and intermolecular interactions in the fluid phase [16]. For treatment at lower pressures (e.g. 85 bar), increasing temperature resulted in lowering scCO2 fluid density, leading to weaker intermolecular interactions between flurbiprofen and CO2 molecules in the fluid phase. Consequently, fewer drug molecules were able to dissolve into the fluid phase, thus less available for diffusion into the biocomposite structure. However, at higher pressure range (e.g. 115 bar), increasing temperature would have a more significant effect on increasing solute vapor pressure, surpassing any negative impact on drug solubility due to solvent density. As a result, drug solubility increases as temperature increases, allowing more drug molecules to be transported inside the biocomposite structure by scCO2 at higher pressures. Although drug uptake at 100 bar pressure appeared to be higher at 313 K than at 323 K, which still follows the drug solubility influence theory, the difference was statistically insignificant. Drug loading in PMMA/β-TCP composites exhibits a strong correlation with drug solubility in scCO2 for both 313 K and 323 K settings, as shown in Fig. 5. Note that solubility data included in Fig. 5 was derived from data extrapolation of Duarte et al.’s work [16], with some additional calculation based on the measured UV peak absorbance of flurbiprofen in scCO2 at 85 bar settings as part of this study. The relationship appears to be linear, with a crossover between the two temperature settings at a lower solubility point as explained previously. This linear correlation suggests that drug partitioning between the scCO2 fluid phase and the PMMA/β-TCP composite phase played a key role in the impregnation process, as similarly found by Shen et al. and Diankov et al. [29,30]. Under the tested experimental conditions, the highest drug uptake achieved was 26.1% ± 4.4% at 115 bar and 323 K. Based on the observed trends in this study, along with drug solubility data reported by Duarte et al. [16], drug uptake level can be increased further by increasing scCO2 pressure, processing temperature, and the porosity level inside the biocomposite structure. According to Velu and Singamneni, porosity of the biocomposites can modulated by changing

3.4. Drug release profile The drug-loaded PMMA/β-TCP biocomposite samples were soaked in ethanol solution at room temperature and the drug release profile was characterized. Percent drug release was calculated using Eq. (2):

% Drug release = Mass of drug released / Mass of sample prior to proces sin g × 100%

(2)

Fig. 6 shows different drug release profiles for PMMA/β-TCP samples treated under different scCO2 conditions. Drug release profile seems to correspond directly to the drug uptake level of the biocomposites. The initial rate of drug being released from the samples with the most drug uptakes was significantly higher than that from the samples with lower drug load. About 50% of the drug uptake was able to release within four hours of soaking in ethanol. This is expected due to porous structure of the biocomposites and high solubility of flurbiprofen in ethanol solution. Like the drug uptake, the level of porosity inside the biocomposite structure can modulate the drug release profile. Flurbiprofen release profile appears similar to that of cholesterol-impregnated PMMA samples (with up to 20 wt.% loading) previously reported by Elvira et al., showing almost all of the solute was released into ethanol solution within the first 24 h of sample soaking [27]. The experimental drug release data was fitted to the Weibull function, which is often used to empirically model the drug release from delivery systems [31]:

Mt / M∞ = 1 − exp(−αt b)

(3)

Eq. (3) can also be written in an alternative form as shown in Eq. (4) [32]:

Mt / M∞ = 1 − exp[−(t / τ )b]

(4)

Mt and M∞ are the masses of drug released at times t and infinity, respectively; α and b are constants; and τ is a modified constant of α

Fig. 5. Correlation between drug loading in PMMA/β-TCP composite and drug solubility in scCO2.

34

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

Fig. 6. Drug release profile of scCO2 treated PMMA/β-TCP biocomposites in ethanol solution (Note: y-axis values are calculated based on Eq. (2)).

behavior of flurbiprofen out of the composite structure and into the solution. More in-depth investigations on the structural changes of PMMA matrix between processed temperature settings will need to be performed in future work in order to confirm this hypothesis. Using the derived constants τ and b from the model, the drug release amounts were calculated based on Eq. (4) and compared against the actual drug release amounts obtained from the experimental data. Fig. 7 indicates a linear fit between the predicted data and the experimental data. This linear relationship validates the suitability of modeling the release of flurbiprofen from laser sintered and scCO2processed PMMA/β-TCP biocomposites into the solution using the Weibull function.

bearing the same time unit as t. Data from Fig. 6 was fitted to Eq. (4), using the amount of drug released into the solution at the end of each drug release experiment as M∞, the instantaneous amounts of drug released as Mt, and time t in minutes. Table 1 shows the values of constants τ and b calculated from the Weibull empirical model, along with R2-values representing the goodness of fit for each of the data curves. The average values for constants τ and b for composite samples processed in scCO2 at 313 K were calculated to be 242.4 ± 4.9 and 1.782 ± 0.091, respectively, and for composite samples processed in scCO2 at 323 K were 263.5 ± 2.9 and 2.007 ± 0.154, respectively. The fitted b-values appear higher than typical b-values found in other controlled drug release scenarios from simpler polymer matrices. Higher b-values in Weibull distributions typically indicate the absence of plasticizer and coating around the drug particles [31,32]. It could also be due to the mixed presence of PMMA spheres and β-TCP particles inside the host composite matrix from which flurbiprofen molecules were released. Moreover, drug diffusion originated from not only the impregnated PMMA spherical matrices but also the voids present inside the whole porous PMMA/β-TCP composite structure, thus complicating the overall drug release mechanism. According to Kosmidis et al., constant τ strongly depends on the surface of the host matrix in the composite, whereas b is influenced by the interaction between drug particles and the host matrix inside the composite [31]. T-tests show that with 95% confidence, there were statistically significant differences in the average constant values obtained between 313 K and 323 K processed samples, with pvalue = .03 < .05 in both τ and b cases. This result indicates that scCO2 processing temperature had some potential influences not only on the surface of PMMA but also on the interaction between flurbiprofen particles and the PMMA matrix, thus affecting the diffusion

4. Conclusion Selective laser sintered biocomposites of PMMA/β-TCP were treated with flurbiprofen drug using supercritical carbon dioxide. ScCO2 acted as the dissolution medium for the drug from solid powder phase into fluid phase. It also served as the transporting medium for the drug molecules to migrate from fluid phase into the PMMA/β-TCP biocomposite through the porous structure and polymer surface absorption. Drug loading ability of the biocomposites directly correlates to the solubility of flurbiprofen in scCO2. As the pressure increased at constant temperature so did the level of drug uptake. When comparing drug loading at different temperatures, drug loading increased at higher temperatures for the upper tested pressure range (above approximately 110 bar); however, drug loading decreased at higher temperatures for the lower pressure range (below 110 bar). The drug release profile was also characterized, showing 50% of the loaded drug was able to escape the biocomposite structure within four hours of soaking in ethanol solution. The Weibull modeling of drug release data indicates potential influences of scCO2 processing temperature on the PMMA surface and the interaction between drug particles and the host PMMA matrix inside the composite structure. Results from this study prove the feasibility of drug impregnation inside laser sintered biocomposite structures. Because the biocomposite structures can be conveniently engineered by changing laser sintering parameters, the level of drug uptake and drug release can be directly controlled and designed with temperature and scCO2 settings to meet specific application needs.

Table 1 Calculated values for τ and b based on the Weibull function. ScCO2 processing condition

Weibull function parameters

T (K)

P (bar)

τ (minutes)

b

313 313 313 323 323 323

85 100 115 85 100 115

241.0 247.8 238.3 266.8 261.4 262.2

1.885 1.714 1.748 2.182 1.892 1.947

Goodness of fit, R2

0.99 0.98 0.98 0.92 0.97 0.89

Acknowledgement This research was made possible by financial support from the 35

The Journal of Supercritical Fluids 136 (2018) 29–36

T.T. Ngo et al.

Fig. 7. Comparison between Weibull model-predicted drug release and actual drug release.

National Science Foundation Research Experience for Undergraduates program (Grant ID: CHE#1460645), the ALSAM Foundation, the Institute of International Education, the University of San Diego Associated Students and Engineering Research Funds. The authors would also like to thank Rachel Lloyd and Diego Giordani for their assistance with some of the experimental setup.

[15]

[16] [17]

References [18] [1] J.F. Burke, P. Didisheim, D. Goupil, et al., Application of materials in medicine and dentistry, in: B.D. Ratner, A.S. Hoffman, F.J. Schoen, J.E. Lemons (Eds.), Biomaterials Science: An Introduction to Materials in Medicine, Academic Press, San Diego, California, 1996, pp. 283–375. [2] S.G. Kazarian, Supercritical fluid impregnation of polymers for drug delivery, in: P. York, U.B. Kompella, B.Y. Shekunov (Eds.), Supercritical Technology for Drug Product Development, Informa Healthcare, United Kingdom, 2004, pp. 322–342. [3] I. Kikic, P. Sist, Applications of supercritical fluids to pharmaceuticals: controlled drug release systems, in: E. Kiran, B.G. Debenedetti, C.J. Peters (Eds.), Supercritical Fluids, Kluwer Academic PublishersDordrecht, The Netherlands, 2000, pp. 291–306. [4] A.R.C. Duarte, A.L. Simplicio, A. Vega-González, P. Subra-Paternault, P. Coimbra, M.H. Gil, H.C. de Sousa, C.M.M. Duarte, Supercritical fluid impregnation of a biocompatible polymer for ophthalmic drug delivery, J. Supercrit. Fluids 42 (2007) 373–377. [5] G. Giavaresi, E.B. Minelli, M. Sartori, A. Benini, T.D. Bora, V. Sambri, P. Gaibani, V. Borsari, F. Salamanna, L. Martini, N.N. Aldini, M. Fini, Microbiological and pharmacological tests on new antibiotic-loaded PMMA-based composites for the treatment of osteomyelitis, J. Orthop. Res. 30 (2011) 348–355. [6] G. Lewis, Properties of antibiotic-loaded acrylic bone cements for use in cemented arthroplasties: a state-of-the-art review, J. Biomed. Mater. Res 89B (2009) 558–574. [7] R. Velu, S.B. Singamneni, Evaluation of the influences of process parameters while selective laser sintering PMMA powders, Proc. Inst. Mech. Eng. Part C: J. Mech. Eng. Sci. 229 (2014) 603–613. [8] T. Jaeblon, Polymethylmethacrylate: properties and contemporary uses in orthopaedics, J. Am. Acad. Orthop. Surg. 18 (2010) 297–305. [9] E.B. Minelli, A. Benini, PMMA as drug delivery system and in vivo release from spacers, in: E. Meani, C. Romano, L. Crosby, G. Hofmann, G. Calonego (Eds.), Infection and Local Treatment in Orthopedic Surgery, Springer, Cham, Switzerland, 2007, pp. 79–91. [10] H. van de Belt, D. Neut, D.R.A. Uges, W. Schenk, J.R. van Horn, H.C. van der Mei, H.J. Busscher, Surface roughness porosity and wettability of gentamicin-loaded bone cements and their antibiotic release, Biomaterials 21 (2000) 1981–1987. [11] G. Giavaresi, E.B. Minelli, M. Sartori, A. Benini, A. Parrilli, M.C. Maltarello, F. Salamanna, P. Torricelli, R. Giardino, M. Fini, New PMMA-based composites for preparing spacer devices in prosthetic infections, J. Mater. Sci.: Mater. Med. 23 (2012) 1247–1257. [12] F. Barrère, C.A. van Blitterswijk, K. de Groot, Bone regeneration: molecular and cellular interactions with calcium phosphate ceramics, Int. J. Nanomed. 1 (2006) 317–332. [13] R. Velu, S. Singamneni, Selective laser sintering of polymer biocomposites based on polymethyl methacrylate, J. Mater. Res. 29 (2014) 1883–1892. [14] A.M. López-Periago, A. Vega, P. Subra, A. Argemí, J. Saurina, C.A. García-González,

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29] [30]

[31]

[32]

36

C. Domingo, Supercritical CO2 processing of polymers for the production of materials with applications in tissue engineering and drug delivery, J. Mater. Sci. 43 (2008) 1939–1947. S.J. Macnaughton, I. Kikic, N.R. Foster, P. Alessi, A. Cortesi, I. Colombo, Solubility of anti-inflammatory drugs in supercritical carbon dioxide, J. Chem. Eng. Data 41 (1996) 1083–1086. A.R.C. Duarte, P. Coimbra, H.C. de Sousa, C.M.M. Duarte, Solubility of flurbiprofen in supercritical carbon dioxide, J. Chem. Eng. Data 49 (2004) 449–452. A.R. Berens, G.S. Huvard, R.W. Korsmeyer, F.W. Kunig, Application of compressed carbon dioxide in the incorporation of additives into polymers, J. Appl. Polym. Sci. 46 (1992) 231–242. T.T. Ngo, J. McCarney, J.S. Brown, M.J. Lazzaroni, K. Counts, C.L. Liotta, C.A. Eckert, Surface modification of polybutadiene facilitated by supercritical carbon dioxide, J. Appl. Polym. Sci. 88 (2003) 522–530. T.T. Ngo, C.L. Liotta, C.A. Eckert, S.G. Kazarian, Supercritical fluid impregnation of different azo-dyes into polymer: in situ UV/Vis spectroscopic study, J. Supercrit. Fluids 27 (2003) 215–221. P.I. Sealy, C. Nguyen, M. Tucci, H. Benghuzzi, J.D. Cleary, Delivery of antifungal agents using bioactive and nonbioactive bone cements, Ann. Pharmacother. 43 (2009) 1606–1615. T.T. Ngo, C. Lambert, B. Dorren, B. Gee, S. Go, R.D. George, Effects of deposition and supercritical CO2 treatment parameters on physical and electrical properties of pentacene thin films, Synthetic Met. 220 (2016) 384–393. T.T. Ngo, J.E. Keegan, R.D. George, Processing behaviors of thin-film pentacene and benzene-1,4-diboronic acid in supercritical carbon dioxide, Thin Solid Films 520 (2011) 1022–1026. J.K. Rice, E.D. Niemeyer, F.V. Bright, Evidence for density-dependent changes in solute molar absorptivities in supercritical CO2: impact on solubility determination practices, Anal. Chem. 67 (1995) 4354–4357. Y.R. Sharma, Elementary Organic Spectroscopy, S. Chand Publishing, India, 2007, p. 34. M. Li, P. Cheng, R. Zhang, G. Luo, Q. Shen, L. Zhang, Preparation of PMMA/graphene oxide microcellular foams using supercritical carbon dioxide, Mater. Sci. Eng. 87 (2015) 012042. J. Martin-de Leon, V. Bernardo, M.A. Rodriguez-Perez, Low density nanocellular polymers based on PMMA produced by gas dissolution foaming: fabrication and cellular structure characterization, Polymers 8 (2016) 265–281. C. Elvira, A. Fanovich, M. Fernandez, J. Fraile, J.S. Roman, C. Domingo, Evaluation of drug delivery characteristics of microspheres of PMMA-PCL-cholesterol obtained by supercritical-CO2 impregnation and by dissolution-evaporation techniques, J. Control. Release 99 (2004) 231–240. S. Üzer, U. Akman, Ö. Hortaçsu, Polymer swelling and impregnation using supercritical CO2: a model-component study towards producing controlled-release drugs, J. Supercrit. Fluids 38 (2006) 119–128. Z. Shen, G.H. Huvard, C.S. Warriner, M. McHugh, J.L. Banyasz, M.K. Mishra, CO2assisted fiber impregnation, Polymer 49 (2008) 1579–1586. S. Diankov, D. Barth, A. Vega-Gonzalez, I. Pentchev, P. Subra-Paternault, Impregnation isotherms of hydroxybenzoic acid on PMMA in supercritical carbon dioxide, J. Supercrit. Fluids 41 (2007) 164–172. K. Kosmidis, P. Argyrakis, P. Macheras, A reappraisal of drug release laws using Monte Carlo simulations: the prevalence of the Weibull function, Pharm. Res. 20 (2003) 988–995. A. Hadjitheodorou, G. Kalosakas, Analytical and numerical study of diffusion-controlled drug release from composite spherical matrices, Mater. Sci. Eng. C 42 (2014) 681–690.