|Cancer||Cell death||Cell cycle||Cytoskeleton||Exo/endocytosis||Differentiation||Division||Organelles||Signalling||Stem cells||Trafficking|
Polymethyl-methacrylate–sorbitol-based capsules as local drug delivery vehicles: a preliminary study
Dorottya Frank*, Gellért Cseh†, Tamás Nagy*, László Pótó‡, Béla Kocsis§ and Attila Miseta*1
*Department of Laboratory Medicine, Faculty of Medicine, University of Pcs, Ifjsg u. 13.,7624, Pcs, Hungary, †Department of Surgical Research and Techniques, Faculty of Medicine, University of Pcs, Ifjsg u. 13.,7624, Pcs, Hungary, ‡Institute of Bioanalysis, Faculty of Medicine, University of Pcs, Ifjsg u. 13.,7624, Pcs, Hungary, and §Department of Medical Microbiology and Immunology, Faculty of Medicine, University of Pcs, Ifjsg u. 13.,7624, Pcs, Hungary
Local delivery of antibiotics via PMMA (polymethyl-methacrylate) has been widely used in the treatment of chronic osteomyelitis for over 40 years. Unfortunately, PMMA is water insoluble, which seriously limits antibiotic delivery. In addition, the polymerization temperature of PMMA is high, and consequently, only heat-stable antibiotics can be used. Therefore our aim has been to develop an effective antibiotic delivery system, which can be loaded with a wide variety of drugs and deliver the molecules in a predictable manner. Capsules with wall thicknesses of 0.3–0.6 mm from PMMA mixtures containing 40–70 w/w% (weight percent) of sorbitol were prepared and their permeability tested with BPB (Bromophenol Blue). Sorbitol content and wall thickness correlated with the BPB release. SEM (scanning electron microscopy) showed that the canalization of capsules also was well correlated with both sorbitol content and wall thickness. The PMMA–sorbitol-based capsule can potentially be a versatile tool in assuring effective delivery of antibiotics and other substances.
Key words: antibiotic release, chronic osteomyelitis, infection, local therapy, polymethyl-methacrylate
Abbreviations: ANOVA, analysis of variance, BPB, Bromophenol Blue, LADV, local antibiotic delivery vehicle, PMMA, polymethyl-methacrylate, SEM, scanning electron microscopy, w/w%, weight percent
1To whom correspondence should be addressed (email email@example.com).
In the treatment of chronic osteomyelitis, LADV (local antibiotic delivery vehicles) have drawn an increasing interest, although the ideal LADV has yet to be developed (Nelson, 2004). Using LADV to treat local infections is preferable to systemic therapy, as the main benefit of these devices is to obtain locally a high antibiotic concentration at the site of the infection without increasing systemic toxicity. Hence, LADV can either replace or supplement the parenteral antibiotic administration, since the bioavailability of the systemic drugs is uncertain at the affected area due to poor vascular supply (Hanssen, 2005). In 1970, Buchholz and Engelbrecht (1970) reported their pioneering work on the use of antibiotic-impregnated PMMA (polymethyl-methacrylate). Since then, the antibiotic laden PMMA represents the standard in prevention and/or treatment of deep bone infections (Nelson, 2004).
However, the release properties of substances from acrylic bone cement are suboptimal (Belt et al., 2001; McLaren, 2004; Wu et al., 2006). Previous studies concluded that antibiotic-laden PMMA shows biphasic release kinetics, with rapid initial burst being followed by sustained low rate of release (Wu et al., 2006; Dunne et al., 2007; Moojen et al., 2008; Squire et al., 2008; Zilberman et al., 2008). Furthermore, due to the hydrophobicity of PMMA, <10% of the trapped drug is released from such composite materials (Cabanillas et al., 2000; Zilberman et al., 2008). To increase the accessible antibiotic content, PMMA has been complemented with various water-soluble fillers (sucrose, glycine, erythrol, xylitol, additional antibiotics) (McLaren et al., 2005, 2007; Dunne et al., 2007). Moreover, a number of commercially used antimicrobial drugs cannot withstand the high temperature (70–110°C) during polymerization (DiPisa et al., 1976). Another limitation is that antibiotics in aqueous solutions cannot be used due to mixing and hardening problems (Popham et al., 1991).
Therefore, our aim has been to develop a LADV that provides controlled local therapy for chronic osteomyelitis, capable of using a wide range of drugs with relatively steady time-dependent release kinetics over a protracted period. We have created a PMMA–filler (sorbitol)-based capsule system with different wall thicknesses and filler content. We assumed that the two properties are important determinants in the formation of elution channels. Theoretically, therefore, we should be able to control the capsule wall permeability by optimizing these properties. The permeability was preliminarily monitored by measuring the release of BPB (Bromophenol Blue) as a marker molecule throughout a clinically relevant period of time (42 days).
2. Materials and methods
2.1. Preparation of PMMA–sorbitol capsules
Sixteen groups of capsules were prepared with different wall thickness and filler content. The chosen filler was sorbitol, an anticariogenic food sweetener that is inexpensive, water-soluble and is metabolized either slowly or not at all by bacteria (Burt, 2006).
We have prepared four types of Surgical Simplex P PMMA bone cement mixtures (Stryker-Howmedica-Osteonics) containing 40, 50, 60 and 70 w/w% (weight percent) of sorbitol powder (Sigma–Aldrich). First, the sorbitol was ground and passed through a filter with a pore size of 160 μm (Millipore), and then homogeneously dry-mixed with PMMA polymer powder. After adding the liquid monomer to the powder mix, cylindrical blocks were made. The technique used by Börzsei et al. (2006a, 2006b) was further improved by using a milling machine instead of a press to prepare the capsules. Therefore, capsules were milled out from these blocks with wall thickness of 0.3, 0.4, 0.5 and 0.6 mm. The plug air-tightly closes the capsule, as shown (Figure 1).
2.2. Determination of capsule wall permeability
BPB (Sigma–Aldrich) was used as a diffusion indicator preceding antibiotic elution studies. Accordingly, the following antibiotic elution studies could focus only on the relevant filler content and wall thickness. Therefore, the capsules were loaded with 40 μl of 1 mg/ml BPB in physiological saline solution before being carefully closed to avoid any undesirable leakage or BPB contamination of their outside surface. The capsules were submerged in 1 ml of 0.89% NaCl solution and left for 42 days at room temperature. The sample holders, each containing one capsule, were capped to prevent evaporation. Ten capsules of each type were used for parallel measurements. The absorption spectra of BPB showed a characteristic maximum at 592 nm. For the quantitative determination of the BPB released from the capsule into the submerging NaCl solution, we used a standard series of 0.89% NaCl solution complemented with 0, 0.1, 0.5, 1, 5, 10 and 50 μg/ml BPB. For the sample measurements, 800 μl aliquots were placed into spectrophotometer cuvettes, and absorption at 592 nm was recorded with a PerkinElmer Lambda 2 Spectrophotometer (PerkinElmer). Following measurements, these aliquots were returned into the sample holders. Similar readings were taken on days 1–6 and subsequently 9, 12, 15, 19, 23, 27, 32, 37 and 42 days. As a negative control, capsules were loaded from each group with saline solution and used as background.
2.3. Analysis of surface morphology by SEM (scanning electron microscopy)
The surface morphology of the capsules was examined and characterized using a JEOL JSM-6300 Scanning Electron Microscope (Akashima). The capsules used in the elution studies were collected on day 42 and coated with gold. Pictures were recorded at 60-fold magnification. The percentage ratios of the developed pores were calculated by using Image J image processing software (plugin: contour plotter analyser). Upon using the contour plotter, the pictures were first formatted into 8 bit, followed by the background intensity determination. Prior to determining the area of the voids, the limits of the contour intensities were adjusted to the backgrounds.
2.4. Statistical analysis
Data are presented as mean value±S.D. throughout. All statistical analyses were performed using SPSS 15.0 (SPSS Inc.). The effects of capsule wall thickness and sorbitol content for the release of BPB were analysed using the ANOVA (analysis of variance) test for each sampling time (1–42 days). Statistically significant differences were defined at P<0.05. ANOVA was followed by post hoc analyses of paired groups of capsules on days 6, 15 and 32 of incubation. Since 16 groups of capsules were compared at three different time-points, a total of 720 analyses were performed. We colour-coded identically the groups of capsules, which were similar with regards to their BPB release properties (P>0.05).
The relation between the release properties and the capsule surface morphologies was also assessed. Comparison of the surface area of pores, as determined by SEM, was plotted and analysed against the beginning of the initial release, the time needed to release half the total load, the daily maximum release of BPB and the time passed between 25 and 75% of the total load released. Within each experimental group, the data of SEM analysis (n = 3) were paired with the mean values obtained from BPB release (n = 10).
3.1. Wall permeability of capsules
Panels show our results (Figures 2A–2D). The mean values±S.D. (y-axis) of BPB release at 0.3, 0.4, 0.5 and 0.6 mm wall thickness as a function of incubation time (x-axis) are indicated, respectively (Figures 2A, 2B, 2C and 2D). Each panel contains four curves for sorbitol content of 40, 50, 60 and 70 w/w%, respectively. It is apparent that BPB release slowed down significantly with increasing wall thickness. The curves representing 40–70 w/w% sorbitol concentrations gradually flatten with increasing wall thickness. At the 70 w/w% sorbitol level, full equilibration was seen by day 6 with the 0.3 mm wall thickness. Full equilibration was postponed in time when the capsule walls were thicker: 19, 23 and 42 days, respectively. As expected, the same tendencies were seen in capsules with lower sorbitol contents. Comparison of the effect of sorbitol content on BPB release at a given capsule wall thickness showed that less and less BPB is measured in the incubation medium when the sorbitol content decreases. Consequently, only about half the total BPB load is released in the thinnest, 0.3 mm/40 w/w% sorbitol capsules by the end of experiment.
ANOVA analysis of the released BPB as the function of wall thickness or filler content gave highly significant differences at any time-point (P<0.001). As expected, the combined effect of these two factors had a significant cumulative impact on the release profiles (P<0.001). Pairwise, post hoc analyses were also performed on days 6, 15 and 32 (Table 1). We considered the release characteristics significantly different at P<0.05. Capsules with similar release characteristics were labelled with the same shading (Table 1). Considering only the wall thickness, the thinner is the capsule wall at given filler concentration, the faster is the release group to which it belonged. Likewise, the higher the sorbitol content of the capsule at any given wall thickness, the faster the release of BPB.
Table 1 Results of post hoc analyses of paired groups of capsules
Capsules with significantly different release characteristics are labelled with different shading (P<0.05). Capsules labelled with * were used as references in the same colour-coded group.
3.2. Surface morphology of capsules by SEM
In SEM pictures recorded at day 42, circular dark voids were present (Figures 3A–3D). The calculated area of the pores is shown (Figure 4). We found that the larger the area of surface pores, the sooner the initial burst occurs (Figure 5A). An exponential curve provided the best fit to our data (R2 = 0.868, P<0.001). The relationship between the surface porosity and the time needed to release half of the total load gave a negative exponential correlation (R2 = 0.945, P<0.001) (Figure 5B). The relation between the surface porosity and the daily maximum release of BPB can be seen (Figure 5C). As expected, the data gave a strong exponential positive correlation (R2 = 0.939, P<0.001). The surface porosity compared with time passed between the 25 and 75% total BPB released plot confirmed strong negative exponential relationship between the variables (R2 = 0.945, P<0.001) (Figure 5D).
Antibiotic-laden PMMA represents the current standard for local drug delivery in orthopaedic surgery (Nelson, 2004). PMMA delivers antibiotics to the tissues by elution, as the body fluid surrounding the bone cement dissolves the contained drug. An impediment to this mechanism is the zero water permeability of the bone cement (McLaren et al., 2005). Therefore, following the fast release of surface-accessible antibiotics, sustained suboptimal release occurs due to slow canalization of PMMA. Therefore, various bioinert water-soluble fillers have been added to PMMA to facilitate delivery of an antibiotic by causing interconnecting voids in the polymer (McLaren et al., 2005, 2007; Dunne et al., 2007). However, the bone cement inevitably encapsulates some of the complementing antibiotics, and totally, delivery cannot be achieved. In addition, the polymerization process occurs at high temperature, which prevents the use of heat-labile antibiotics. Therefore, an alternative LADV, offering more complete and efficient release, would be more appealing. Hence, our aim was to develop a PMMA–water-soluble filler-based LADV, with the intention of creating a capsule which could be loaded with virtually any type of antimicrobial agents, and that effectively releases its content in a time-dependent manner.
We have therefore assessed the permeability of 16 types of capsules to select the capsules with the best release modality. Our hypothesis was that if PMMA–sorbitol capsules are loaded and placed in an aqueous environment, interconnecting pores (elution channels) will be formed due to the solubilization of the filler. We assumed that the development of the elution channels, and therefore the release of BPB, would correlate positively with sorbitol content and negatively with wall thickness. Since the release of BPB was presumed to be associated with the canalization of the capsule wall, we also assumed that the surface area of elution channels would similarly relate to the release characteristics of BPB.
The data obtained demonstrate that LADV with various controlled release kinetics can be prepared by setting capsule wall thickness and filler content. A rapid high initial release was associated with high filler content and thin capsule wall. If the filler content was low and the capsule wall thick, relatively even but slow release kinetics were observed. However, the correct combination of filler content and capsule wall thickness resulted in relatively steady release kinetics with almost complete discharge of BPB by day 42. Consequently, our results compare favourably with those of others (Cabanillas et al., 2000; Dunne et al., 2007; Moojen et al., 2008). In the case of the sorbitol-filled capsules, formation of elution channels starts immediately, but effective canalization is prolonged where the capsule wall is thick and/or the filler concentration is low. The more elution channels that are open, the quicker is the equilibration process, but the acceleration is dampened by the fact that in the meantime, the capsule BPB concentration decreases. This allows the creation of a LADV with relatively constant and efficient discharge of its load.
In summary, we have developed and carried out preliminary tests on PMMA-based capsules consisting of graded amounts of water-soluble filler with different wall thickness as a promising vehicle for local drug delivery. In all capsules, BPB release occurred in accordance with the capsule properties (sorbitol content and wall thickness), as higher sorbitol content provided better permeability, and increased capsule wall resulted in protracted release. These findings suggest that the PMMA–sorbitol-based capsules may be potentially useful LADVs in orthopaedic surgery (and other applications) by providing more controlled and effective release kinetics. Verification of the significance of these in vitro observations will require further in vitro antibiotic elution studies.
All the authors participated in this research. Gellért Cseh and Attila Miseta raised the research question. Dorottya Frank, Béla Kocsis and Attila Miseta designed the study. Dorottya Frank and Tamás Nagy carried out the experiments and gathered the data. Dorottya Frank and László Pótó analysed the data. Dorottya Frank, Tamás Nagy and Attila Miseta wrote the initial draft. Béla Kocsis, Gellért Cseh and Attila Miseta assessed the accuracy of the data and analysis. Dorottya Frank and Gellért Cseh contributed the literature research.
We thank Ferenc Szvacsek for the preparation of capsules and Béla Dolgos and László Seress for the valuable technical and professional assistance in SEM.
This work was supported by the
Belt, H, Neut, D, Schenk, W, Horn, JR, Mei, HC and Busscher, HJ (2001) Infection of orthopedic implants and the use of antibiotic-loaded bone cements. A review. Acta Orthop Scand 72, 557-71
Börzsei, L, Mintál, T, Horváth, A, Koós, Z, Kocsis, B and Nyárády, J (2006a) Comparative study of antibiotic-containing polymethylmethacrylate capsules and beads. Chemotherapy 52, 1-8
Börzsei, L, Mintál, T, Koós, Z, Kocsis, B, Helyes, Zs, Kereskai, L and Nyárády, J (2006b) Examination of a novel, specified local antibiotic therapy through polymethylmethacrylate capsules in a rabbit osteomyelitis model. Chemotherapy 52, 73-9
Cabanillas, PF, Pena, ED, Barrales-Rienda, JM and Frutos, G (2000) Validation and in vitro characterization of antibiotic-loaded bone cement release. Int J Pharm 209, 15-26
Dunne, N, Hill, J, McAfee, P, Todd, K, Kirkpatrick, R, Tunney, M and Patrik, S (2007) In vitro study of the efficacy of acrylic bone cement loaded with supplementary amounts of gentamicin. Acta Orthop 78, 774-85
McLaren, AC, Nelson, CL, McLaren, SG and Wassel, DL (2005) Phenolphthalein used to assess permeability of antibiotic-laden polymethylmethacrylate: a pilot study. Clin Orthop Relat Res 439, 48-51
Moojen, DJ, Hentenaar, B, Vogeley, C, Verbout, AJ, Castelein, RM and Dhert, WJ (2008) In vitro release of antibiotics from commercial PMMA beads and articulating hip spacers. J Arthroplasty 23, 1152-6
Squire, MW, Ludwig, BJ, Thompson, JR, Jagodzinski, J, Hall, D and Andes, D (2008) Premixed antibiotic bone cement: an in vitro comparison of antimicrobial efficacy. J Arthroplasty 23 ((6 Suppl 1)), 110-4
Received 7 October 2010; accepted 11 November 2010
Published as Cell Biology International Immediate Publication 11 November 2010, doi:10.1042/CBI20100712
© The Author(s) Journal compilation © 2011 Portland Press Limited
ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)
Figure 2 Cumulative release of BPB from PMMA capsules with wall thickness of 0.3 (A), 0.4 (B), 0.5 (C) or 0.6 mm (D) containing various amounts of sorbitol: 40 (⧫), 50 (▾), 60 (•), 70 w/w% (▪)
Figure 3 Representative SEM pictures of the surfaces of PMMA capsules with wall thickness of 0.6 mm containing 40 (A), 50 (B), 60 (C) and 70 w/w % (D) of sorbitol at the same magnification (×60)
Figure 4 The percentage of porous surface compared with various wall thicknesses and sorbitol contents: 40 (black columns), 50 (light grey columns), 60 (white columns), 70 w/w% (dark grey columns)