polydimethylsiloxane xerogel and in vitro release of drug

Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 193–207 www.elsevier.com/locate/actabiomat

Correlation between physicochemical properties of doxorubicin-loaded silica/polydimethylsiloxane xerogel and in vitro release of drug Magdalena Prokopowicz * Medical Academy of Gdan´sk, Division of Physical Chemistry, Hallera 107, 80-416 Gdan´sk, Poland Received 19 February 2008; received in revised form 6 June 2008; accepted 24 July 2008 Available online 6 August 2008

Abstract The aim of this study was to prepare organically modified silica xerogels by a two-step acid/base catalyzed sol–gel process that would provide a slow release of an anticancer drug – doxorubicin hydrochloride (DOX). The influence of different amounts of PDMS added on the properties of xerogels intended for the release of the drug and the dissolution of xerogels was investigated. SEM, BET, IR and nitrogen gas adsorption/desorption measurements were performed to characterize the microstructures and chemical properties of xerogels. It is shown that an increase in the proportion of PDMS in the silica network is associated with a decrease in bulk density, specific surface area, total volume of pores and proportion of silanol groups (higher hydrophobicity). These results also revealed the influence of PDMS on the release of doxorubicin hydrochloride and the dissolution behavior of xerogels. An increase in PDMS content results in a decrease in both the rate of drug release and dissolution of xerogels. After the release study, the changes of microstructures and physicochemical properties of these xerogels were also examined. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Doxorubicin hydrochloride; Sol–gel methods; Silica/polydimethylsiloxane xerogels; Drug release

1. Introduction Silica materials obtained by the sol–gel process are amorphous, inorganic and porous, and have found applications as a coating on implants and medical products [1,2], biocatalyst [2], biosensor [2,3] and matrix for a controlled release of drugs [4–15]. A growing field of interest in silica materials was found in the application for a local drug release at the implanted site. This treatment appears to be a much more interesting alternative, where the goals of drug-releasing systems are to maintain the drug in the desired therapeutic range with just a single dose, localize delivery of the drug to a particular part of the body, reduce the need for follow-up care, preserve drugs that are rapidly destroyed by the body, and to increase patient comfort and/or improve compliance. *

Fax: +48 58-3493206. E-mail address: [email protected]

The principal advantage of utilized silica materials as an implantable delivery vehicle is that the sol–gel method is non-toxic, uncomplicated, inexpensive, takes place at room temperature (important for thermosensitive drugs), and it does not require the use of pharmaceutically unacceptable solvents. These materials are biocompatible in vivo, cause no adverse tissue reactions and degrade in the body to silicic acid, i.e., Si(OH)4, which is eliminated through the kidneys [8]. Another advantage, namely bioactivity (e.g., materials function like the tissue in which they are implanted) of the sol–gel-derived materials has been widely studied [16–21]. These materials show apatite-forming ability in a simulated body fluid (SBF) and may lead to the bone bonding in vivo circumstances. Therefore, it seems to be a very attractive idea to look for materials that could release a drug in a local and controlled way while showing bioactive properties. The sol–gel process involves the formation of a colloidal suspension of appropriate compounds (sol) and its gelation

1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2008.07.027

194

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

through hydrolysis followed by polycondensation until a solid network (gel) is formed [22,23]. Bioactive agents can be added during the sol–gel manufacturing process or introduced into gel by adsorbing a drug onto the surface of the gel [4–15]. In the former case, during aggregation of the formed colloidal particles, the drug is incorporated in the lattice of silica gel polymer. Subsequent to further polymerization, aging and drying under atmospheric conditions, the solid sol–gel material called xerogel with encapsulated drug is obtained. In the latter case, a gel of desired porosity is immersed in solution of a drug to be encapsulated [13,14]. This approach is important for labile agents (e.g., biomolecules) to avoid their decomposition under the conditions used for gel production. The release mechanisms of drugs from sol–gel-derived silica materials have not been extensively studied. Some literature reports indicate that the drug release is controlled by simple diffusion through solvent-filled capillary channels while other papers concluded that the release of drug occurs according to a combined process of the diffusion and the erosion process of the matrix [4–15]. Diffusion-controlled release was found for the release of nifedipine, trypsin inhibitor and lidocaine from silica xerogels [4,10,15]. However, the release of toremifene was found diffusioncontrolled from crushed particles but conformed to zeroorder kinetics from monoliths [5,6]. The zero-order release was found also for drugs with high molecular weight, such as heparin (MW 1600) [7]. A diffusion-controlled release of drug seems to be achieved if the drug is released prior to erosion of the silica materials. The diffusion of drug was faster with decreasing device diameter and surface area to volume ratio of the material or with increasing size of open pores in silica materials. In addition to this, the water-soluble drugs and drugs with smaller molecular weight were diffused easier than non-soluble or high-molecular-weight ones. On the other hand, zero-order release of the drug is attained if the drug was released simultaneously with polymers eroding. The erosion rate of silica materials seems to be dependent on the total amount of material and generally decreases as the material is depleted [8]. The rate of silica xerogel erosion in vitro depends on their pore characteristics and on their surface properties. Specifically, the presence of an apatite surface film significantly reduced xerogel erosion resulting also in decrease of drug release [16]. In addition to these, the active agent may also interact with the silica backbone and hence the rate of releasing is controlled by the interactions between drug and polymer. Altogether these studies document that in the design of a more or less diffusion-controlled release system, all aspects, such as chemical properties and molecular size of a drug, the possibilities of interaction between the drug and polymer material, and the size of matrix, have to be taken into consideration. In addition to these, the conditions of gel synthesis, especially pH value and a water/alkoxysilane molar ratio, which affect hydrolysis and condensation reactions, and therefore the texture of silica gel [4–15], also influence the design of less or more diffuse systems. Longer

gelation times produce linear silica aggregates and a more condensed microstructure, resulting in the less diffuse system [6]. However, faster gelation leads to the branched structure with a broader distribution of larger pores in silica gels and hence resulting in the opposite effect. The gelation time is longest near the isoelectric point of silica at pH 2, and decreases with increasing pH of the sol [6]. In addition to pH, an increase in the water/silicon alkoxide molar ratio decreases time of gelation and also correlates with increasing specific surface area, which also partly controls the release of drug. An addition to colloidal tetraethoxysilane-based sols of organic polymers (e.g., polyethylene glycol [4]) or organosilanes (e.g., methyltriethoxysilane, propyltriethoxysilane, [9,15]), most often slows down the rate of drug release compared to non-modified silica gels, which makes them attractive materials for long-term delivery systems. Organic/ inorganic gels made by the sol–gel method combine the advantages of both organic and inorganic materials and are expected to possess new properties that individual organic or inorganic materials could not achieve. The inorganic precursors can directly bond to the organic polymers and covalent bonds are formed between them and/or weaktype interactions such as hydrogen bonding or interpenetrating networks may be formed between the organic and inorganic phases. The objective of previous work was to prepare a doxorubicin hydrochloride-loaded non-modified silica xerogel by the sol–gel method and to examine its stability and drug release [11]. Doxorubicin hydrochloride is a cytotoxic, anthracycline antibiotic used in antimitotic chemotherapy. It was chosen as an active agent due to its common use in the treatment of bone cancer [24]. Presently, the treatment is systematic and, due to the narrow therapeutic index of doxorubicin, a substantial increase in systematic dose to achieve high concentration of the drug at the bone sarcomas is not possible. Currently, doxorubicin is incorporated into physically self-assembled structures (liposomes, micelles) and polymer–drug conjugates [25–28]. Despite enhanced therapeutic effects, low encapsulation efficiency, drug leakiness and low stability of doxorubicin-loaded liposomes or micelles pose serious problems. In the author’s previous work it was found that sol–gels synthesized at room temperature can be used successfully for a complete loading with doxorubicin without losses or leaching out and a prolonged drug release was achieved [11]. Furthermore, drug encapsulation significantly delayed doxorubicin degradation kinetics. The aim of this work was to prepare silica/polydimethylsiloxane xerogels with encapsulated doxorubicin hydrochloride that would provide a slower drug release compared to non-modified silica. Tetraethoxysilane was used as a SiO2 source. Silanol-terminated PDMS was chosen as an organic, oligomeric component for the synthesis of matrices due to its capability to change hydrophilic properties of the matrix surface, mechanical properties of the matrices and their microstructure [29,30]. In general,

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

PDMS was found to change high stiffness and brittleness of non-modified silica and produce materials with more toughness and flexibility. Furthermore, because PDMS surfaces have good hemo- and bio-compatibility, it is generally believed that surface biocompatibility may be improved by modifying biomaterial surfaces with organic silicone-like properties [31]. The preparation of silica/PDMS xerogels was carried out through a two-step acid/base catalyzed sol–gel process as described previously [11]. In the case of the two-step process, during an acidic step hydrolysis proceeds much faster than condensation and most of alkoxysilanes are hydrolyzed to the silanol group [22,23]:

The addition of a base catalyst in the second step accelerates condensation:

195

properties such as pore size, pore volume and specific surface area were studied by the adsorption–desorption of N2. 2. Materials and methods 2.1. Reagents All reagents and the drug were obtained from Sigma Chemical Company (St. Louis, MO, USA) and used without further purification: doxorubicin hydrochloride (referred to as DOX) (M = 580.0 g mol1; purity 99%); silicon standard solution (acidic) (1000 mg l1); tetraethoxysilane (TEOS, silanol-terminated poly(dimM = 208.33 g mol1); ethylsiloxane) (PDMS, 150 cSt fluid); ethanol (98%); sodium lauryl sulfate (1%); hydrochloric acid (0.01 M) and ammonia (0.2%). The release medium was simulated body fluid (SBF, pH 7.4) with ion concentrations nearly equal to those of human blood plasma [33]. SBF had the following composition (in mM): NaCl (136.8), NaHCO3 (4.2), KCl (3.0), K2HPO4H2O (1.0), MgCl26H2O (1.5), CaCl22H2O (2.5) and Na2SO4 (0.5). It was buffered at pH 7.4 with a tris[hydroxymethyl]aminomethane solution (50 mM) and hydrochloric acid (1 M). A Millipore Milli Q-plus system was used to purify water for the preparation of buffers and reagent solutions (resistance of water 18 MX, pH 7.0). 2.2. Preparation of silica/PDMS xerogels by the sol–gel method

Silica/PDMS xerogels were obtained by an addition of silanol-terminated polydimethylsiloxane to the acid-catalyzed colloidal suspension which can form interpenetrating networks between the organic and inorganic phases or undergo a co-condensation reaction with unreacted silanol groups of alkoxysilanes, as proposed by Wilkes et al. [32]:

Preparation of the DOX-loaded silica/PDMS xerogels by a two-step sol–gel process was made in polypropylene containers at room temperature, under atmospheric pressure and kept out of the light (DOX has been found stable in the dark and to not adsorb onto polypropylene [34]). Hydrochloric acid and ammonia were used as catalysts with a molar ratio TEOS/H2O/C2H5OH/HCl/NH4OH 1/6/8/8  105/5  105 as previously described [11]. The

Doxorubicin-loaded silica/PDMS xerogels were obtained by an addition of the drug in the second step of preparation. The freeze-drying technique was chosen to dry DOX-loaded xerogels because traditional drying at a higher temperature effects the decomposition of DOX. The objective of this research was to study the effect of modifier content on the release behavior of freeze-dried doxorubicin-loaded silica/PDMS xerogels and dissolution of xerogels under soaking in simulated body fluid. The molecular surface structure of the xerogels was investigated by IR and SEM microscopy. Some of the key physical

solutions were prepared as follows: TEOS (1.73 g) and ethanol (3.07 g) were slowly stirred magnetically for 15 min. Next, de-ionized water (1.8 g) with CaCl2 (0.05 g), and a catalyst – HCl (60 lL, 0.01 M, pH 3) – were added. The solutions were stirred for 1 h to obtain uniform sols. After that, various amounts of PDMS as an organic modifier and sodium lauryl sulfate (SLS) (1%) as an anionic surfactant were added dropwise to the solutions. Table 1 shows the weight ratios of these precursors in the three samples prepared (Si–8% PDMS, Si–15% PDMS and Si–21% PDMS). The mixtures were tightly closed and sonicated in a cold

196

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

Table 1 Weight ratios of synthesis precursors and final appearance of silica/ polydimethylsiloxane xerogel Sample

Experimental wt. ratio PDMS/ TEOS

PDMS Experimental Appearance* theoretical wt. ratio wt.% in final SLS/PDMS samples

Si–8% PDMS Si–15% PDMS Si–21% PDMS

1/35 1/17 1/12

8 15 21

*

1/8 1/8 1/8

T o O

T, transparent; o, slightly opaque; O, opaque.

water bath by applying ultrasonic radiation of 20 kHz until they were clear (8–10 min). The sono-sols were stirred in a covered flask for one day at a temperature of 50 °C and then a base catalyst – ammonia (40 ll, 0.2%, pH 10.5) was added. The pH of mixtures was raised to about 5.5–6.0 (DOX is stable in the pH range 4.5–6.5 [34,35]). After 4 h of stirring, 1 ml of a water-based solution containing 2 mg of doxorubicin hydrochloride was added. Mixtures with the drug were cast into disk-shaped polypropylene moulds and allowed to stand undisturbed at +4 °C (DOX is thermosensitive [11,34]) for 7 days to continue polycondensation and gelation. Next, the diskshaped, solid gels were placed in the drying chamber of an Alpha 1–2 LD Freeze-Dryer (Christ, Germany), and cooled to 55 °C. Freeze-drying was performed at a pressure of 2 Pa for 48 h as described previously [11]. Samples of modified xerogels (without drug) were also prepared in an identical manner. 2.3. Physicochemical characteristics of silica/PDMS xerogels The bulk density of the disk-shaped, drug-loaded silica/ PDMS xerogels was calculated from sample weight and volume. The other studies were performed with a crushed form of samples. The structural variables used to characterize the pores in these samples, such as the pore volume, volume fraction porosity, specific surface area, pore distribution and average pore diameter, were obtained from the nitrogen gas adsorption/desorption at 77 K (Micromeritics ASAP 2405 N). The specific surface areas and average pore diameters were determined using the Brunauer–Emmet– Teller (BET) method. Pore distributions were determined from the desorption isotherms using the Barret–Joyner– Halenda (BJH) method and pore shapes were calculated from the BJH desorption and adsorption isotherms. The total pore volume of pores was obtained at a single point measurement p/po = 0.9845. FTIR spectroscopic studies of the bulk structure of samples were carried out in the operating range of 4000– 400 cm1, using a Jasco model 410 FTIR and KBr pellet technique. An FTIR spectrometer (Jasco model 410, 4 cm1 resolution) equipped with the Horizontal Attenuated Total Reflectance Accessory (45° incident angle, zinc selenide ATR prism) was used to record ATR/FTIR spec-

tra of the surface structure of samples in the range of 1300– 700 cm1. The thickness of the gel layer was about 1 mm, i.e., much greater than the effective penetration depth of the IR light passing the ATR crystal, making ATR an excellent technique for surface studies [36]. For a better comparison, all the IR spectra were normalized to maximum absorption of dominant peak at 1074 cm1 attributed to the asymmetric stretching of siloxane bands. In order to determine the degree of fractions of Si–OH and Si–O groups in the sample surface network, the ATR spectra were deconvoluted between 1300 and 900 cm1. The fitting was performed by a non-linear least squares method, using Gaussian and Lorentzian functions. The proportion of Si–OH and Si–O groups was calculated by the ratio: (area of Si–OH and Si–O bands)/(total area of the mas Si–O–Si band) plus (area of SiOH and Si–O bands)) according to Fidalgo et al. [37]. The morphology of the samples was observed by scanning electron microscopy (SEM, Philips XL 30) at an acceleration voltage of 15 kV or 30 kV, respectively. Samples for SEM were fixed on carbon tape and fine gold sputtering was applied for 2 min. 2.4. Release test of DOX in vitro The dissolution tests of DOX from different silica/ PDMS xerogels were studied to determine the rate of drug release for these systems under physiological conditions. Prior to the studies, the resulting monolithic DOXloaded silica/PDMS xerogels were crushed and sieved to a desired diameter of granules 1–2 mm (±0.1 mm). Weighed xerogels with various amounts of PDMS (500 mg) containing 2 mg of DOX were examined. The studies were performed twice for an initial time up to 24 h and for an extended time up to 250 h. In the former case, drug-loaded xerogels were subjected to dissolution testing at 37 °C in 50 ml of simulated body fluid solution (SBF), pH 7.4, under sink condition of drug (the theoretical highest concentration of the drug in the medium was much below the 15% drug aqueous solubility, which is a requirement for fulfillment of ‘‘sink” conditions). The release medium in a polypropylene flask was shaken in a thermostatted shaking water bath (75 shakes per minute) (Julaba, Germany (±0.02 °C)) and kept out of the light. At various time intervals, 2 ml aliquots of the solution were filtered through a membrane filter (cellulose acetate, 0.22 lm) and the spectrophotometric assay of DOX was performed (UV/VIS spectrometer Jasco model V-500 in the classical mode for quantitative determinations of DOX) at k = 480 nm). The release medium was replaced by a fresh 2 ml aliquot of SBF buffer to maintain constant volume. For the extended-time study, the medium was regularly replaced, following the measurement of drug release, every 24 h during the time of release (this replacement was necessary because of rapid decomposition of dissolved form of DOX [11]). The volume of SBF medium was 10 ml.

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

The rest of the experimental conditions were the same as for the initial release studies. Quantitative determinations of the amount of DOX released were based on pre-calibration of the spectrometer using standard solutions of the drug in simulated body fluid. To understand the release mechanisms of doxorubicin hydrochloride, release data were fitted using a semi-empirical power law equation: Mt ¼ k  tn M1

ð1Þ

where Mt is the amount of the drug released at time t; M1 is the total amount of loaded drug; k is a characteristic constant of the silica/PDMS xerogels and n is the release exponent indicating the type of drug release mechanism. As the k value becomes higher, the release occurs faster. The n value approaching 1 corresponds to zero-order release kinetics, 0.5 < n < 1 implies a non-Fickian release model and n approaching 0.5 indicates Fickian diffusion for non-swelling polymers (Higuchi model) [38]. From the plot of  log

Mt M1

¼ f ðlog tÞ; kinetic parameters, n and k, were

calculated. After the release test, samples were removed from the SBF buffer, rinsed with de-ionized water, freeze-dried under the same conditions as mentioned above, and weighed to determine the weight loss. An analytical balance (RADWAG, Poland) was used to determine the weight of the samples before and after release test. Furthermore, the measurements of physicochemical properties of samples were repeated. 2.5. Dissolution test of silica/PDMS xerogels The granular form (1–2 mm) of silica/PDMS xerogels (without drug) weight-to-solution-volume ratio (500 mg/ 50 ml) was used in order to determine the dissolution rate and the saturation level of Si. Seven samples from each type of xerogel were immersed in the SBF medium for a specific time of 0.5, 2, 6, 10, 12, 16 and 24 h, respectively. The dissolution medium in a polypropylene flask was shaken in a thermostatted shaking water bath (Julaba, Germany (75 shakes per minute) at temperature 37 ± 0.02 °C. The dissolution of silica/PDMS xerogels at the specific time of immersion was determined by measuring dissolved silicic acid spectrophotometrically (UV/vis spectrometer Jasco model V-500) using the molybdenum blue method (k = 820 nm) [39]. The silicic acid analysis was based on the reduction with 1-amino-2-naphthol-4sulfonic acid. Quantitative determinations of silicic acid were based on pre-calibration of the spectrometer using standard solutions of silicon (IV) in simulated body fluid. The experiments were terminated when the equilibrium (saturation with Si) was reached. The degree of dissolution of samples was expressed as the percentage

197

of mean mass of silicic acid released from the weighed xerogel (wt.%). 2.6. Statistical analysis The release, weight loss and dissolution tests were repeated five times and the results are given as mean ± standard deviation (SD). The structural experiments were repeated three times and the results are presented as medians. The Kruskal–Wallis nonparametric analysis of variance was used with Dunn’s post hoc procedure to analyse differences between more than two groups. The differences were considered significant if P < 0.05. 3. Results and discussion Non-modified silica xerogels have a long-range amorphous structure, resulting from a random network of elementary SiO4 units arranged in a variety of siloxane structures, both linear and cyclic species (mainly fourmembered siloxane rings and six-membered rings), with hydrophilic hydroxyls on the surface [40,41]. Silanol-terminated PDMS is one of the organically modified precursors employed for the introduction of organic groups, in this case methyl groups, that act as sol–gel network modifiers. According to the literature [29,30], pH of the sol–gel reaction has a main influence on co-condensation of TEOS with PDMS and with an increase in pH of mixtures, the efficiency of co-condensation increases. In this work, modifications of the silica network occur at pH values not higher than 5.5–6.0, determined by the above-mentioned doxorubicin stability. Fig. 1A–C shows scanning electron micrographs of freeze-dried silica/PDMS xerogels with varying PDMS contents resulting from different pH values during the sol–gel reaction. With an increase in the amount of PDMS, the surface of samples changed. For Si–8% PDMS, the surface was homogenous with one visible phase, whereas for Si–15% PDMS and Si–21% PDMS, two distinguishable phases were observed with a good overall dispersion of PDMS within silica on the micro scale. The smallest, spherical microphase (diameter of about 0.5 lm) was obtained for the smallest amount of PDMS, whereas the largest sphere (diameter of about 1 lm) was obtained for the largest amounts of PDMS. These results correlated with the appearance of obtained samples presented in Table 1, which was transparent up to 8 wt.% of PDMS, and slightly opaque and opaque for 15 and 21 wt.% of PDMS, respectively. The opacity phenomenon depends both on the material porosity (largest pores, i.e., macropores, resulted in increased light scattering, and thus higher opacity) and on the length of oligomer chains as well as on their concentration [29]. Because no macroscopic phase separation occurs in these materials, the latter cause seems to be more important. For long PDMS chains or for higher concentrations, opacity occurred because of the chance of forming PDMS chains

198

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

Fig. 1. Morphology (A–C) before and (D–F) after release test of Si–8% PDMS, Si–15% PDMS and Si–21% PDMS, respectively, monitored by scanning electron microscope.

by self-condensation of silanol-terminated PDMS. The self-condensation of PDMS is a competitive reaction, which can occur when the hydrolysis of the alkoxide is not sufficiently rapid or pH of condensation is not sufficiently high such as in this work. This suggests that under desired conditions of gelation, chemical modification by covalent bonding of the gels, took place to a small extent only and the structure is mostly composite perhaps with the presence of weak-type hydrogen bonding interactions. Since the backbone of PDMS and the products formed by hydrolysis and condensation of TEOS have similar nature of chemical bonds (Si–O–Si), the organic and inorganic components may be compatible. This is also the reason why a dispersion of PDMS in the final network structure occurs, although a certain degree of localized microphase separation may exist [32].

In this study, silica/PDMS xerogels have been investigated as a potential implantable delivery material. Therefore, the examination of possible impurities in the matrix is important. These organic impurities can remain in the xerogels because of the processing conditions or incomplete chemical reactions. A comparison of FTIR spectra of samples is shown in Fig. 2. The absence of a strong ethanol band (at 880 cm1, attributed to the methyl and methylene rocking modes) indicates that very little or no ethanol remained in the xerogels. The appearance of characteristic CH2/CH3, C–O and C–C bands in the fingerprint region: 800–1500 cm1, and C–H stretching vibrations near 2940 cm1 most probably indicates the presence of unreacted or partially reacted TEOS [41]. However, these peaks overlapped with those of PDMS, thus making a clear observation of TEOS remaining in the samples quite diffi-

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

199

Fig. 2. FTIR spectra of silica/polydimethylsiloxane xerogels normalized to mas(Si–O–Si) at 1074 cm1. Dotted line: Si–8% PDMS, dashed-dotted line: Si–15% PDMS and solid line: Si–21% PDMS.

cult. More precise experiments may be achieved by examining the hydrolysis and condensation of TEOS using, e.g., in situ ATR/FTIR or 29Si NMR. In the spectra of samples, there is a gradual decrease of the bands at 3400 cm1, 1650 cm1, 950 cm1 and 560 cm1 with an increase in PDMS content. The evaluation of the 950 cm1 component quantifies the population of unreacted Si–OH and Si–O groups [37,40]. A relative intensity of this band decreases with the PDMS content and leads to increases of hydrophobicity of samples. These results also confirm that the polymer does not hinder the silica condensation process, leading to a final xerogel with less unreacted silanol groups. The increasing hydrophobicity of samples correlated well with a decrease in both bands attributed to water deformation band at  1650 cm1 and to all silanol groups with adsorbed water at 3400 cm1. Furthermore, in samples with an increased content of PDMS, the band centered at 3400 cm1 experienced a small upshift to about 3425 cm1 (Fig. 2), suggesting different contributions of silanol types and of the PDMS–OH stretching mode. These changes correlated with an increase in the bands at 2940 cm1 and 1269 cm1 corresponding to the C–H vibrations of Si–CH3 groups, coming from PDMS, confirming the hybrid or composite character of these samples. Additionally, an increase in the band at 800 cm1 was also observed, which was attributed to an increase in Si–O–Si groups in the silicate–PDMS network. The broad dominant band at 1074 cm1 with the shoulder on the high-frequency side of this peak related to the Si–O–Si asymmetric stretching mode of the silica network (transverse optic (TO) and longitudinal optic (LO) components [37], respectively) was also changed. The frequency of the maximum (TO) peak slightly upshifted from 1074 cm1 (pure silica-spectra not shown) to 1080 cm1 for Si–21%

PDMS. In agreement with this, the peak became narrower with an increase in PDMS, suggesting the change in structure porosities [37]. According to the central force network model [41], the TO antisymmetric stretching frequency is related to the Si–O–Si intertetrahedral bond angle, and to the Si–O stretching force constant. An increase in the Si– O–Si angle increases the frequency of vibration; on the other hand, an increase in the S–O bond length leads to a decrease in the stretching force constant and a shift in the opposite direction. Thus, the upshift of this mode with an increase in PDMS content can also be related to a decrease in the Si–O bond length and/or to a bridging angle increase. Additionally, IR information showing the downshift of a band from 460 cm1 to 450 cm1 corresponds to a decrease in the Si–O–Si angles [40,41]. Therefore, the formation of stronger and shorter Si–O bonds with an increase in PDMS has the predominant effect on the upshift of TO antisymmetric stretching. The observed slight decrease of the 560 cm1 band, attributed by Fidalgo et al. to four-membered siloxane rings [40], indicated a small decrease in the proportion of four-membered siloxane rings instead of six-membered rings with an increase of PDMS. Fidalgo et al. [40] deduced that the presence of strained structural rings, such as fourmembered siloxane rings, correlated with a more porous structure because of the existence of non-bridging oxygens (Si–O). Therefore, a decrease in this kind of structural units with an increase in PDMS content suggests the formation of a final silica/PDMS structure more linear than usual for pure silica, and responsible for the lowest porosity. All these observations suggest that an increase in the amount of PDMS results in the occurrence of heterogeneity regions with polymer predominance in the silica/PDMS xerogels network. These heterogeneous regions, formed as

200

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

a result of PDMS addition, may induce local network deformations responsible for both an increase in hydrophobicity and a decrease in the Si–O–Si bond lengths and angles and also porosity of the final sample. These IR observations are well supported by the results obtained by N2-adsorption/desorption experiments. N2adsorption/desorption isotherms of these samples are shown in Fig. 3A and the pore distribution determined by the BJH method is depicted in Fig. 4A. The following physical properties of xerogels containing PDMS up to 21 wt.% are summarized in Table 2: bulk density, pore size, specific surface area, micropore area, micro- and mesopore volume and total pore volume. As shown in Fig. 3A, all xerogels reveal a type I isotherm with a sharp initial increase of the adsorbed gas volume at a low relative pressure p/po, indicating that xerogels have a significant contribution of micropores (pore diameter <2 nm), according to the IUPAC recommendation [42]. Nevertheless, these isotherms showed hysteresis loops – type H1, characteristic of capillary condensation in mesopores (pore diameter >2 nm) [42] with the presence of cylindrical shape of the pore system. This shape of pores is also confirmed by adsorption/desorption BJH data in this study. As presented in Table 2, an increase in PDMS content in silica decreased density, pore volume and surface area without affecting the diameter of pores. The differences between bulk densities, pore volumes, and surface area values were found highly significant (P = 0.01) whereas no significant differences between pore diameters were found (P > 0.05). The pore distribution determined by the BJH method rapidly decreases with an increasing pore diameter up to 10 nm, and gives an average pore diameter of 2 nm for all silica/PDMS xerogels (Fig. 4A). This average pore diameter is also confirmed by the BET method (Table 2). The same pore distribution for all xerogels confirms that PDMS does not hinder hydrolysis and condensation reac-

tions of silica. A large decrease in the specific surface area (from 650 to 318 m2 g1) and total pore volume (from 0.369 to 0.172 cm3 g1) with an increase in PDMS content (up to 21 wt.%) indicates decreased porosities of the samples. This effect is associated with the above-mentioned formation of interpenetrating networks between the PDMS and silica phases. The decrease of densities of samples from 1.52 g cm3 for non-modified silica (characteristic of a dense xerogel) to 1.22 g cm3 for SiO2–21wt.% PDMS resulted from lower capillary tensions, since the increasing of hydrophobic properties reduced the wall interactions between pores and a solvent and hence caused their shrinkage [29,30]. The observed differences in both chemical properties and morphology of samples with an increase in PDMS content may result in a different behavior of drug-loaded silica/PDMS xerogels during a release study. The aim of this study was to develop a doxorubicin delivery system as a potentially implantable, bioactive matrix for the treatment of bone cancer. The comparison between in vitro and in vivo degradation and release study of the same xerogels is not documented as well. Radin et al. [16] reported that the total degradation of silica granules of the size similar to this work occurred after 1 week of in vitro immersion in physiological solution with solution exchange. In contrast, only a decrease in size, not a total degradation was observed for these granules after bone implantation up to 4 weeks. Such behavior of silica correlated well with the release study. Kortesuo et al. [14] also showed that in vitro about 70% of the loaded amount of toremifene citrate was released after 24 h, whereas in vivo a sustained release could be observed over 40 days. The results show that the in vitro release profile does not correlate with the in vivo profile on a real-time basis. The reason for the difference in behavior between the in vitro and in vivo data was probably due to the difference in the ratio of surface area of the implant to the volume of the solution at implantation

Fig. 3. Adsorption–desorption isotherms of N2 at 77 K of silica/polydimethylsiloxane xerogels: (A) before and (B) after release test.

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

201

Fig. 4. Pore size distribution obtained from adsorption data using the BJH method. Silica/polydimethylsiloxane xerogels (A) before and (B) after release test.

Table 2 Comparison of some physical properties of of silica/polydimethylsiloxane xerogel and non-modified silica xerogels Sample Bulk density (g cm3) Pore diameter (4 V/A by BET) (nm) BET surface area (m2 g1) Micropore area (m2 g1) Total pore volume (cm3 g1) Micropore volume (cm3 g1) Mesopore volume (cm3 g1)c

Si–0% PDMS 1.52 2.22 650.81 154.91 0.362 0.066 0.296

Si–8% PDMS

Si–15% PDMS

Si–21% PDMS

1.42 2.32a 5.90b 535.7a 279.9b 115.2a 6.2b 0.311a 0.413b 0.049a 0.0018b 0.262a 0.411b

1.38 2.45a 5.39b 383.2a 248.6b 35.7a –19.2b 0.235 a 0.353b 0.014a –0.012b 0.221a 0.353b

1.22 2.25a 5.33b 318.2a 262.7b 29.9a –7.8b 0.179a 0.350b 0.012a –0.005b 0.167a 0.350b

a

Before release study. After release study. c Calculated by the difference between the total pore volume and micropore volume. b

site and in dissolution study. In addition, the study in vitro of silica xerogels aimed at their bioactivity and potential application in bone implantology [1,20] revealed that the silica dissolution/re-precipitation appears to be related to apatite formation and this formation is related to a limited dissolution of silica. Altogether, the results suggest that in vivo resorption of xerogels implanted into bone was slower compared to degradation in vitro. Perhaps the reason was due to the limited blood flow to the skeletal tissue [34] and therefore the saturation level of silica ion was achieved faster. Hence, the rate of resorption may be controlled in vivo by the rates of dissolution/re-precipitation of silica and diffusion of silica ion into the local tissue around the implant. Therefore, in this release study a large material weight-to-

solution-volume ratio (500 mg/50 ml) was used in order to reach solution saturation with Si, and prevent further sol–gel degradation during immersion, such as in paper of Santos [10]. The solution was replaced every 24 h as a result of decomposition of doxorubicin under static conditions and to simulate dissolution/re-precipitation cycles of silica. This proposed route of release can simulate the in vivo release model in which the dissolution/re-precipitation of silica may be the rate-limiting mechanism of DOX release at bone implantation site. The effect of PDMS content on the release of the drug was studied for samples in the granular formulation, containing 2 mg of doxorubicin as a model formulation and shown in Fig. 5. The kinetic data are listed in Table 3. Fig. 5B shows the cumulative DOX release for Si–8% PDMS, Si–15% PDMS and Si–21% PDMS samples as a function of immersion time up to 250 h while Fig. 5A presents the release of doxorubicin over the initial time of release up to 24 h. The differences between samples were highly significant (P = 0.01). The data suggest that there were two stages of release for all silica/PDMS xerogels and that there was no burst effect – defined as an initial large bolus of the drug released before the release rate reaches a stable profile [43]. When fitted to the zero-order release (r > 0.98), the drug release was linear up to 24 h, and the obtained release exponent, n, was close to 1 for all the profiles (Table 3), suggesting that the release rate of drug was independent of the initial time. The release percent of DOX was 51 ± 1.4 wt.%, 33 ± 0.9 wt.% and 10.3 ± 0.7 wt.% for Si–8% PDMS, Si–15% PDMS and Si–21% PDMS, respectively in the first 24 h. After that, the release rate slowed down for all the silica/PDMS xerogels and a decrease in rates (smallest k in Table 3) was

202

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

Fig. 5. Doxorubicin release profiles from silica/polydimethylsiloxane xerogels as a function of (A) initial time, (B) extended time and (C) square root of time release. Dashed lines are linear fits of calculated values. Best linear fits do not include the initial release (up to 24 h).

Table 3 Release kinetics parameters Sample (%)

Release exponent na

Kinetic constant (wt.%/hn)a

Release exponent nb

Kinetic constant (wt.%/hn)b

Si–8 PDMS Si–15 PDMS Si–21 PDMS

0.89 ± 0.05

4.23 ± 0.12

0.50 ± 0.02

6.7 ± 0.21

1.00 ± 0.02

0.86 ± 0.03

0.56 ± 0.03

2.1 ± 0.06

0.92 ± 0.03

0.83 ± 0.04

0.58 ± 0.04

1.2 ± 0.04

a b

Initial time of release up to 24 h. Extended-time release up to 250 h.

observed with an increase in PDMS content. The total cumulative DOX release was 80 ± 2.3 wt.% for Si–8% PDMS, while 50 ± 2.1 wt.% and 33 ± 1.3 wt.% releases were found at the end of the study for Si–15% PDMS

and Si–21% PDMS, respectively. The release exponent in the range 0.5–0.6 for all the silica/PDMS xerogels suggested diffusion-controlled release of DOX at the extended time. In case of diffusional release, plotting the cumulative percentage of drug release versus square root of time should yield a linear correlation according to the Higushi model [44]. Excluding the initial release by omitting the early time data points (up to 24 h), good fits were obtained (r > 0.98), as shown in Fig. 5C. However, the initial (up to 24 h) relatively rapid release of the drug (zero-order release, Fig. 5A) is likely to be the consequence of a mechanical erosion and/or dissolution process after immersion in the SBF medium. To confirm these results, the dissolution behavior of the samples during the initial time (up to 24 h) of immersion in SBF was studied. In fact, as seen in Fig. 6, the initial dissolution process occurred for all samples; however, the saturation of Si was not rapidly

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

Fig. 6. Dissolution profiles of silica/polydimethylsiloxane xerogels in SBF as a function of immersion time.

achieved. A closer inspection of these profiles revealed that the dissolution rate (before the saturation level of Si) decreased as the amount of PDMS increased and the differences between xerogels were highly significant (P = 0.01). The rates of dissolution calculated from the linear portion of the curves before the saturation with Si was reached were 0.263 ± 0.03 wt.% h1, 0.163 ± 0.02 wt.% h1, and 0.073 ± 0.01 wt.% h1 for Si–8% PDMS, Si–15% PDMS and Si–21% PDMS, respectively. The tested solutions showed silica saturation level close to 0.06 mg ml1 within 6, 10 and 16 h for Si–8% PDMS, Si–15% PDMS and Si– 21% PDMS, respectively. The literature solubility of the amorphous, non-modified sol–gel-derived silica is in the range 130–150 ppm [22], whereas the level obtained in this

203

work (about 60 ppm for all silicate–PDMS) was lower. These differences are related to different ways of preparing the silica phase, resulting in different structure of silica. Peltola et al. [19] compared the saturation level of Si for sol–gel-derived silica aged for different times. The obtained results confirmed that the differences in structural properties of xerogels affect the obtained level of saturation. Additional processes used in this work, such as sono-catalysis and freeze-drying resulted in more stable structure compared to the typical sol–gel-derived amorphous xerogel [11,41]. However, the reason for the differences in the dissolution rates observed among these xerogels seems to relate to an addition of PDMS to silica network. Iler [23], Brinker [22], and Santos et al. [10] studied the dissolution process of pure silica. According to Brinker et al., silanol groups on the silica surface upon immersion into water solution take up large quantities of water and degrade by hydrolysis of the siloxane bonds through the entire network. Based on these reports, two parameters affect the rate of dissolution of pure silica. The first one is associated with pH of solution: with an increase in pH the silica materials dissolve faster (at pH >10 silica solubility rapidly increases). In this work, the pH of the dissolution medium equal to 7.4 was kept constant for all samples. The second parameter was attributed to the bulk microstructure of silica: less porous structures with a smaller fraction of silanol groups resulted in a slower hydrolysis of the siloxane bonds. The content of silanol groups is very important for the dissolution behavior, because these groups increase the surface hydrophilicity and thus wettability in aqueous solutions [22,23]. These conclusions are in good agreement with the results obtained in this work: the bulk structure of xerogels with the highest PDMS content is least porous with the smallest number of silanol groups and thus these xerogels dissolve more slowly com-

Fig. 7. Example of ATR/FTIR spectra of silica/polydimethylsiloxane xerogels normalized to mas(Si–O–Si) at 1074 cm1. Solid line: before and dotted line: after release test.

204

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

pared to others. In addition, after the release test of DOX, the smallest weight loss close to 6.0 ± 0.2% of the initial mass of matrix was found for the sample with the largest amount of PDMS. The weight losses of the remaining samples were 7.5 ± 0.5 wt.% and 10 ± 0.6 wt.% for the Si–15% PDMS and Si–8% PDMS, respectively. The differences between the silica/PDMS xerogels were significant (P < 0.05). The observed weight loss well of the xerogels results from the regular 24 h replacement of SBF-medium in the time study of DOX release. In addition to weight loss, the granules gradually decreased in size, but not in shape. This size reduction was smallest for the highest PDMS content in silica network. The obtained results are in line with the observed changes both in chemical properties and in particle morphology of silica/PDMS xerogels after the release test. A comparison of the ATR/IR spectra of surface structure of the samples before and after the release test is shown in Fig. 7. It reveals clearly visible changes in the structures. The residue of silanol groups on the surface of xerogels decreased about twofold for all the samples and also the upshift of the dominant band at 1074 cm1 (related to Si–O–Si) was observed. These results also suggest that upon immersion in SBF the silica/PDMS xerogels network was well established by consuming the free Si–OH groups resulting in the restructuring of samples. N2 adsorption/ desorption isotherms of these samples are presented in Fig. 3B. The shape of adsorption/desorption isotherms for all the samples changed to shape IV after the release test, indicating a mesoporous solid, defined as a material with pores in the range 2.0–50 nm, according to the IUPAC recommendation [42]. In addition, less gas is adsorbed at low relative pressures and a sudden increase in adsorption occurs at high pressures, indicating the predominance of mesopores. A characteristic feature of a type IV isotherm is the hysteresis that is normally attributed to the presence of pore cavities larger in diameter than the openings (necks) leading into them, forming the so-called bottleneck or ink-bottle character of pore system [41]. This shape of pores is also confirmed by adsorption/desorption BJH data in this study. The pore distribution of these samples shifted to higher pore sizes, and gives an average pore diameter of 5 nm for all samples (Fig. 4B). The relevant data are shown in Table 2. For all the silica/PDMS xerogels, the mesopore volume and diameter of pores increased dramatically with a decrease in surface area. However, for Si–8% PDMS the specific surface area decreased by a factor of two, thus confirming the largest changes in the structure of the sample with the lowest PDMS content compared to others. Changes in grain morphology after the release study can also be seen in SEM photographs in Fig. 1D–F: the greatest disturbance of grains is observed for the silica containing the smallest amount of PDMS and releasing the largest quantity of the drug. However, the observed increase in pore size, mesoporous volume, and also the restructuring of xerogels were not associated with the DOX release, since it was also observed for xero-

gels without the drug. It was reported that during immersion of silica gels, silica acids were released continually to reach solution saturation with Si, leading to the weight loss of silica and also an increase in their pore size systems [10]. Santos et al. [10] suggested that the immersion-induced restructuring of silica is probably associated with the well-established phenomenon of silica sol–gel aging by dissolution/reprecipitation reactions in aqueous salt solutions. Fig. 8 shows an example of SEM micrographs at high-magnification of surface of silica/PDMS xerogels before (a) and after immersion in SBF solutions (b). The micrograph of surface before immersion (Fig. 8a) reveals that the solid skeletal phase has a globular morphology with primary particles, that are joined together to form agglomerates, 20–40 nm in diameter. In contrast, the SEM micrograph (Fig. 8b) of the surface after immersion indicates a hierarchical morphology with the clusters of these agglomerates organized into larger clusters approaching 100 nm in diameter. Therefore, the observed initial dissolution of the xerogels and change in the physicochemical properties of xerogels upon immersion in the SBF medium indicated that the process of dissolution/re-precipitation occurred

Fig. 8. Example of morphology (a) before and (b) after release test of silica/polydimethylsiloxane xerogels monitored at high-magnification of scanning electron microscope.

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

during the release test. The formation of pores of largest size and volume seems to result from this repetitive dissolution/re-precipitation of silica. In summary, this study was designed in such a way that the large weight material to volume medium was used and hence doxorubicin release would not follow from matrix dissolution, but would rather be the result of a diffusioncontrolled mechanism. According to the literature, the release kinetics of water-soluble drugs such as doxorubicin hydrochloride from a porous, non-swelling matrices such as silica/ PDMS xerogels (results on swelling not shown) depend on: the water (SBF) concentration gradients at the matrix–water interface; the physicochemical properties of the matrices; the size of drug, loading degree, and the uniformity of drug distribution in matrices; and the possibility of interactions between the drug and the network of silica/PDMS. The best model for the release of doxorubicin under the conditions used was expressed by the Higuchi model. This model relates the release rate (Q) to the parameters potentially of interest in determining the DOX release kinetics [44]: Q ¼ f ðD; e; s; A; C i ; tÞ

ð2Þ

where D is the diffusivity of the drug in the release medium, e is the porosity, s describes tortuosity (the complexity of pore geometry), A is the surface area available for outward diffusion, Ci is the initial loading degree, and t is the time of immersion. This study was designed in such a way that the size of matrices and the loading degree of DOX were the same for all hybrids. The uniformity of drug distributions obtained by sol–gel loading was studied in another author’s study [45], in which it is reported that in the sol–gel procedure of loading DOX in the soluble form, the drug is uniformly distributed in the network of matrices. It is known that the size and morphology of drug carriers can be crucial for their release behavior and the influence of pore size on release kinetics is limited as long as it remains much larger than the drug dimensions. The surface area and volume of DOX molecule in its protonated form are 7.37 nm2 and 1.34 nm3, respectively [46], indicating a larger size of molecules than pore diameters, corresponding to 2 nm for all silica/PDMS xerogels before the release test. This effect is also responsible for no burst release of doxorubicin from all the xerogels and confirms that the initial dissolution/ erosion process of the xerogels is necessary to initiate spontaneous diffusion of the drug from the pores to the surrounding medium. These results are in agreement with other studies on sol–gel-derived non-modified silica xerogels as drug release agents for doxorubicin [11,13]. The release profiles described in the literature for silica xerogels were also fitted to the Higushi model [4,10]. In these works the release of the drug was diffusion-controlled, because the solubility of silica under given conditions is much lower than that of the encapsulated drug [15]. Upon immersion into release medium, the studied drugs were released out

205

of the pore structure of the room-temperature synthesized sol–gels (xerogels) in a time-dependent manner [10]. An addition of PDMS to the silica structure resulted in a slower release of the drug compared to pure silica synthesized under the same conditions [11]. An increase in PDMS content correlated with higher hydrophobicity, decrease in specific surface area, pore volume and also porosities of the bulk xerogels, and consequently resulted in a decrease in drug diffusion from less porous material. The author presumes that the release mechanism of water-soluble DOX from these silica/PDMS materials is as follows. As already mentioned, the structure of silica/PDMS xerogels containing 8–21 wt.% PDMS consists predominantly of inorganic, relatively hydrophilic silica phase, in which well-dispersed, hydrophobic PDMS is present. Dissolution of these local siloxane structures devoid of the silanol groups is hindered due to impeded penetration of water into the pores of the material. Consequently, during the first 24 h doxorubicin is released rapidly, according to zero-order kinetics, from the hydrophilic part of silica gel, which at the same time undergoes dissolution until saturation with Si is reached. As a result of repeated cycles of dissolution/re-precipitation of silica gel during drug release, pores with larger volumes and diameters (mesopores >2 nm) are formed and the drug present inside the matrix is gradually released. The release rates continuously diminish with time due to the increasing diffusional distance that hinders drug diffusion from the center of the device. Another important aspect affecting the rate of release of DOX is a possible interaction of the drug with the silica/ polydimethylsiloxane. Doxorubicin is an amphipathic weak base consisting of an anthraquinone moiety and an amino sugar [46]. The doxorubicin molecule is typical of so-called ‘‘chameleon” molecules, i.e., molecules undergoing a hydrophobic collapse when transferred from a polar to an apolar medium [47]. Therefore, according to the literature [9,15] the following three types of interactions between a drug such as DOX and the hybrids can exist: (1) electrostatic interactions, (2) hydrogen bonding and (3) hydrophobic interactions. The DOX release is significantly suppressed by an addition of PDMS to silica network, which can be partly explained by the hydrophobic interactions between DOX and the organic groups. There have been several reports that support the role of the drug–silica matrix hydrophobic interactions on the drug release. According to Kortesuo et al. [9], increasing the number or length of the organic groups attached to silicon reduced the release of dexmedetomidine from an alkylsubstituted silica xerogel. The same conclusions are also reported in the study by Wu et al. [15]. The aim of this study was to design a doxorubicin delivery system as a potential implantable matrix for the treatment of bone cancer. At present, surgery is the best treatment for bone cancer, with the following chemotherapy as a treatment of choice. After removing the tumor, the empty space is usually filled with a bone cement plug. Despite this aggressive approach, surgery does not pre-

206

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207

clude the possibility of tumor regrowth because the therapeutic efficiency of administrated drugs is strongly restricted due to the limited blood flow to the skeletal tissue. It has been reported that administration of a 30 mg m2 of doxorubicin as an intravenous bolus dose only resulted in a marrow drug concentration of 0.52 lg g1, 2.5 h after administration [34]. As a consequence, the development of more efficient therapy becomes very important. According to the literature, the potential use of the sol–gel silica carriers in medical research involves both the possibility of releasing a drug in a local and controlled way, and forming a bioactive hydroxyapatite-like surface upon immersion under physiological conditions. Therefore, the author assumes that by using an implant with doxorubicin at the post-surgery site it will be possible to accomplish drug targeting and a prolonged release of the drug without the conventional systemic administration. Upon releasing of the drug, these materials are slowly resorbed (supposedly, the bulk structure of matrix starts to resorb because silica surrounds the PDMS-domains) and may form bonds with the surrounding tissue – bone (through the formation of apatite on the material surface) and even promote the growth of new tissue. Hence, the removal of the implant seems to be unnecessary. Important papers of Hench, Peltola and Gallardo [18–20] summarize the variables affecting the nucleation and crystallization of hydroxyapatite on porous silica matrices. Hench et al. suggest that the texture is one of the critical variables with the rate of hydroxyapatite formation increasing as pore size and pore volume increase, with a pore size >2 nm required to achieve rapid kinetics of fully crystallized hydroxyapatite. Peltola et al. suggest that the surface properties such as topography and local physicochemical properties, e.g., the hydroxyl groups, are important for the hydroxyapatite formation and with an increase in the number of OH groups the bioactivity in vitro increases. However, no relationship between the porosity characteristics, Si–OH contents or synthesis conditions and in vitro ability to form apatite was found in the paper of Gallardo [20]. Gallardo et al. suggest that the silica dissolution and re-precipitation observed in vitro related to the apatite formation. Therefore, a promising way to continue this experiment would be to study the in vitro bioactivity of the same samples upon immersion in simulated body fluid for a prolonged time. 4. Conclusions The present study showed that doxorubicin could be successfully encapsulated in the sol–gel-produced micro/mesoporous silica/polydimethylsiloxane xerogel and released over a prolonged time period. The rate of release of the drug under the in vitro conditions used is affected by PDMS content, which depends on both the microstructure and chemical properties of the materials. An increase in PDMS content results in an increase in the hydrophobicity and the decrease in porosity of materials, leading to the decrease

in drug release. The kinetics of drug release from all matrices was biphasic: zero-order release in the initial time up to 24 h and smallest release characterized by a linear relationship against the square root of time in the extended time. In summary, silica/polydimethylsiloxane xerogels loaded with doxorubicin seem to be a promising carrier material used as an implantable drug delivery system for long-term disease control (e.g., bone disease). Acknowledgements The author thanks Prof. Jerzy Łukasiak from the Medical Academy of Gdan´sk, and Dr. Andrzej Przyjazny from Kettering University, Flint, MI, for theirs constructive input. References [1] Gallardo J, Galliano P, Moreno R, Duran A. Bioactive sol–gel coatings for orthopaedic prosthesis. J Sol–Gel Sci Tech 2000;19:107–11. [2] Livage J. Sol–gel processes. Curr Opin Solid State Mater Sci 1997;2:132–8. [3] Li J, Tan SN, Ge H. Silica sol–gel immobilized amperometric biosensor for hydrogen peroxide. Anal Chim Acta 1996;335:137–45. [4] Bo¨ttcher H, Slowik P, Su¨b W. Sol–gel carrier systems for controlled drug delivery. J Sol–Gel Sci Tech 1998;13:277–81. [5] Ahola M, Kortesuo P, Kangasniemi I, Kiesvaara J, Yli-Urpo A. In vitro release behaviour of toremifene citrate from sol–gel processed sintered silica xerogels. Drug Dev Ind Pharm 1999;25:955–9. [6] Ahola M, Kortesuo P, Kangasniemi I, Kiesvaara J, Yli-Urpo A. Silica xerogel carrier material for controlled release of toremifene citrate. Int J Pharm 2000;195:219–27. [7] Ahola M, Sa¨ilynoja E, Raitavuo M, Vaahtio M, Salonen J, Yli-Urpo A. In vitro release of heparin from silica xerogels. Biomaterials 2001;22:2163–70. [8] Kortesuo P, Ahola M, Kangas M, Yli-Urpo A, Kiesvaara J, Marvola M. In vitro release of dexmedetomidine from silica xerogel monoliths: effect of sol–gel synthesis parameters. Int J Pharm 2001;221:107–14. [9] Kortesuo P, Ahola M, Kangas M, Leino T, Laakso S, Vuorilehto L, et al. Alkyl-substituted silica gel as a carrier in the controlled release of dexmedetomidine. J Control Release 2001;76:227–38. [10] Santos EM, Radin S, Ducheyne P. Sol–gel derived carrier for the controlled release of proteins. Biomaterials 1999;20:1695–700. [11] Prokopowicz M. In vitro controlled release of doxorubicin from silica xerogels. J Pharm Pharmacol 2007;59:1365–73. [12] Prokopowicz M. Silica–polyethylene glycol matrix synthesis by sol– gel method and evaluation for diclofenac diethylammonium release. Drug Deliv 2007;14:119–27. [13] Prokopowicz M, Przyjazny A. Synthesis of sol–gel mesoporous silica materials providing a slow release of doxorubicin. J Microencapsul 2007;24(7):694–713. [14] Kortesuo P, Ahola M, Karlsson S, Kangasniemi I, Kiesvaara J, YliUrpo A. Sol–gel-processed sintered silica xerogel as cartier in controlled drug delivery. J Biomed Mater Res 1999;44:162–7. [15] Wu Z, Joo H, Lee TG, Lee K. Controlled release of lidocaine hydrochloride from the surfactant-doped hybrid xerogels. J Control Release 2005;104:497–505. [16] Radin S, El-Bassyouni G, Vresilovic J, Schepers E, Ducheyne P. In vivo tissue response to resorbable silica xerogels as controlled-release materials. Biomaterials 2005;26:1043–52. [17] Chen Q, Kamitakahara M, Miyata N, Kokubo T. Preparation of bioactive PDMS-modified CaO–SiO2–TiO2 hybrids by the sol–gel method. J Sol–Gel Sci Tech 2000;19:101–5.

M. Prokopowicz / Acta Biomaterialia 5 (2009) 193–207 [18] Hench LL. Bioactive ceramics theory and clinical applications. Bioceramics 1994;7:3–14. [19] Peltola T, Jokinen M, Veittola S, Rahiala H, Yli-Urpo A. Influence of sol and stage of spinnability on in vitro bioactivity and dissolution of sol–gel derived SiO2 fibers. Biomaterials 2001;22:589–98. [20] Gallardo J, Galliano PG, Porto Lopez JM. Preparation and in vitro evaluation of porous silica gels. Biomaterials 2002;23:4277–84. [21] Tsuru K, Aburatani Y, Yabuta T, Hayakawa S, Ohtsuki Ch, Osaka A. Synthesis and in vitro behaviour of organically modified silicate containing Ca ions. J Sol–Gel Sci Tech 2001;21:89–96. [22] Brinker CJ, Scherer GW. Sol–gel science: the physics and chemistry of sol–gel processing. San Diego, CA: Academic Press; 1990. [23] Iler RK. The chemistry of silica – solubility, polymerization, colloid and surface chemistry and biochemistry. New York: Wiley; 1979. [24] Itokazu M, Kumazawa S, Wada E, Wenyi Y. Sustained release of adriamycin from implanted hydroxyapatite blocks for the treatment of experimental osteogenic sarcoma in mice. Cancer Lett 1996;107:11–8. [25] Gabizon A, Shmeeda H, Barenholz Y. Pharmacokinetics of pegylated liposomal doxorubicin – review of animal and human studies. Clin Pharmacokinet 2003;42:419–36. [26] Nakanishi T, Fukushima S, Okamoto K, Suzuki M, Matsumura Y, Yokoyama M, et al. Development of the polymer micelle carrier system for doxorubicin. J Control Release 2001;74:295–302. [27] Yoo HS, Lee KH, Oh JE, Park TG. In vitro and in vivo anti-tumour activities of nanoparticles based on doxorubicin-PLGA conjugates. J Control Release 2000;68:419–31. [28] Nori A, Jensen KD, Tijerina M, Kopeckova P, Kopecek J. Tatconjugated synthetic macromolecules facilitate cytoplasmic drug delivery to human ovarian carcinoma cells. Bioconjugate Chem. 2003;14:44–50. [29] Foussaier O, Menetrier M, Videau J, Duguet E. Polydimethylsiloxane-based ORMOSIL microstructure: correlation with compressive behaviour. Mater Lett 2000;42:305–10. [30] Guo L, Hyeon-Lee J, Beaucage G. Structural analysis of poly(dimethylsiloxane) modified silica xerogels. J Non-Cryst Solids 1998;243:61–9. [31] Brash JL. Hydrophobic polymer surfaces and their interactions with blood. Ann NY Acad Sci 1977;283:356–71. [32] Wilkes GL, Orler B, Huang HH. Ceramers: hybrid materials incorporating polymeric/oligomeric species into inorganic glasses utilizing a sol–gel approach. Polym Prep 1985;26:300–2.

207

[33] Kokubo T, Kim HM, Kawashita M. Novel bioactive materials with different mechanical properties. Biomaterials 2003;24:2161–75. [34] Fan H, Dash A. Effect of cross-linking on the in vitro release kinetics of doxorubicin from gelatine implants. Int J Pharm 2001;213:103–16. [35] Janssen MJH, Crommelin DJA, Storm G, Hulshoff A. Doxorubicin decomposition on storage, Effect of pH, type of buffer and liposome encapsulation. Int J Pharm 1985;23:1–11. [36] Planinsˇek O, Planinsˇek D, Zega A, Breznik M, Srcˇicˇ S. Surface analysis of powder binary mixtures with ATR FTIR spectroscopy. Int J Pharm 2006;319:13–9. [37] Fidalgo A, Ilharco LM. Correlation between physical properties and structure of silica xerogels. J Non-Cryst Solids 2004;347:128–37. [38] Ritger PL, Peppas N. A simple equation for description of solute release.I. Fickian and non-Fickian release from non-swellable devices in the form of slabs, spheres, cylinders or discs.. J Control Release 1987;5:23–6. [39] Koch OG, Koch-Dedic GA. Silconmolybdanblau-Verfahren. In: Koch OG, Koch-Dedic GA, editors. Handbuch der Spurenanalyse. Berlin: Springer; 1974. p. 1105. [40] Fidalgo A, Ilharco LM. The defect structure of sol–gel derived silica/ polytetrahydrofuran hybrid films by FTIR. J Non-Cryst Solids 2001;283:144–54. [41] Jung HY, Gupta RK, Oh EO, Kim YH, Whang CM. Vibrational spectroscopic studies of sol–gel derived physical and chemical bonded Oromisils. J Non-Cryst Solids 2005;351:372–9. [42] Rouquerol F, Rouquerol J, Sing K. Adsorption by powders and porous solids. Principles, methodology and applications. Academic Press; 1999. [43] Huang X, Brazel CS. On the importance and mechanisms of burst release in matrix-controlled drug delivery systems. J Control Release 2001;73:121–36. [44] Higuchi T. Mechanism of sustained-action medication: theoretical analysis of solid drugs dispersed in solid matrices. J Pharm Sci 1963;52(12):1145–9. [45] Prokopowicz M, Łukasiak J, Przyjazny A. Synthesis and application of doxorubicin-loaded silica gels as solid materials for spectral analysis. Talanta 2005;65:663–71. [46] T} urker L. Quantum chemical studies on certain anthracycline antibiotics. Theochem 2002;583:81–7. [47] Poupaert JH, Couvreur P. A computationally derived structural model of doxorubicin interacting with oligomeric polyalkylcyanoacrylate in nanoparticles. J Control Release 2003;92:19–26.