Mesoporous bioactive glasses for controlled drug release

Mesoporous bioactive glasses for controlled drug release

Available online at Microporous and Mesoporous Materials 109 (2008) 210–215 Mesoporous bioac...

388KB Sizes 0 Downloads 20 Views

Available online at

Microporous and Mesoporous Materials 109 (2008) 210–215

Mesoporous bioactive glasses for controlled drug release Lingzhi Zhao, Xiaoxia Yan, Xufeng Zhou, Liang Zhou, Hongning Wang, Jiawei Tang, Chengzhong Yu * Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433, PR China Received 23 February 2007; received in revised form 19 April 2007; accepted 19 April 2007 Available online 3 May 2007

Abstract Mesoporous bioactive glasses (MBGs) with different compositions have been prepared and their drug release behaviors have been studied. Tetracycline (TC) is selected as the model drug. It is found that TC can be successfully loaded into the mesopores; and increasing the CaO content in MBG increases the loading ratio of TC/MBG. The drug release behaviors of MBGs are also strongly dependent on their composition. MBG 100S without CaO in the framework has a rapid drug release kinetics, while MBG 90S5C, 80S15C, and 70S25C show much slow release profiles and the release constant in the first 10 h decrease as the CaO content increases. However, the release constant of MBG 60S35C is abnormally larger compared to that of MBG 90S5C. The different drug release behaviors of MBGs have been explained by their difference in CaO composition.  2007 Elsevier Inc. All rights reserved. Keywords: Mesoporous bioactive glass; Controlled release; Drug delivery; SBF (simulated body fluid); Tetracycline

1. Introduction Bioactive glasses (BGs) have attracted much attention since the first report by Hench et al. in 1971 [1]. The inorganic composition of human bone is hydroxycarbonate apatite (HCA). When implanted in human body, a HCA layer is formed on the surface of BG, which chemically bonds with living bone [2,3]. In addition to the commercial applications in bone defect repair, this family of biomaterials could be used for bioactive coating of metallic implants [4], protein and/or cell activation [5] for tissue regeneration and tissue engineering [6]. In bone reconstruction surgeries, osteomyelitis caused by bacteria infection is the main complication. Conventional treatments include systemic antibiotic administration, surgical debridement, wound drainage and implant removal. These approaches, however, are rather inefficient and the patients would suffer for extra surgeries. A new


Corresponding author. Tel.: +86 21 55665103; fax: +86 21 65641740. E-mail address: czy[email protected] (C. Yu).

1387-1811/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2007.04.041

method to solve this problem is to introduce a local drug release system into the implant site. The advantages of this treatment include high delivery efficiency, continuous action, reduced toxicity and convenience to the patients. Ordered mesoporous materials such as SBA-15 and MCM-41 have been successfully used as drug delivery systems. Their characteristics such as tunable and uniform pore size, large surface area and high pore volume make it possible to adsorb drug molecules and release them from the mesostructured matrices by a controlled rate [7,8]. However, due to the low bioactivity, pure silica cannot be directly used as bone repair substitute. Very recently, mesoporous bioactive glasses (MBGs) have been successfully synthesized and this new family of biomaterials exhibits superior bone-forming bioactivities [9–11]. Compared to conventional BGs, MBGs show different structure/composition [12] and bioactivity/composition correlations [13]. Moreover, the high pore volume as well as the uniform mesopores of MBGs makes it possible for a direct and highly efficient immobilization of proteins and/or drug molecules within MBGs, which are practically important in future medical applications.

L. Zhao et al. / Microporous and Mesoporous Materials 109 (2008) 210–215





(S) (S) (Z) (S)







Fig. 1. The chemical structure of tetracycline molecule.

In this work the potential of MBGs as controlled drug release system has been studied. Tetracycline (TC), a well-known broad spectrum antibiotic [14], was selected as the model drug and introduced into MBGs with different compositions to investigate the influence of composition upon the release behavior. The chemical structure of TC is shown in Fig. 1.

2. Materials and methods 2.1. Preparation of materials Nonionic triblock copolymer EO20PO70EO20 [P123, BASF, where EO is poly (ethylene oxide) and PO is poly (propylene oxide)] was chosen as a structure directing agent. MBGs were synthesized according to the methods we reported before [9,12]. For the synthesis of MBGs with different compositions, the amounts of silicon and calcium precursors [tetraethyl orthosilicate (TEOS) and Ca(NO3)2 Æ 4H2O, respectively] were shown in Table 1, while the amounts of other reagents were the same. In the synthesis, 4.0 g of P123, 0.73 g of triethyl phosphate (TEP), 1.0 g of 0.5 M HCl, and the silicon and calcium precursors were dissolved in 60 g of ethanol and stirred at room temperature for 1 day. The molar ratio (percentage · 100) of SiO2 (S) and CaO (C) were used to denote MBG samples with different compositions and the P2O5 content is kept at 5% in all samples (see Table 1). The resulting sol was introduced into a Petri dish to undergo an evaporation-induced self-assembly (EISA) process. The dried gels were calcined at 973 K for 5 h to remove the template. Calcined MBGs with different compositions were ground and sieved, the fractions with sizes in the Table 1 The amounts of silicon and calcium precursors used for MBGs with different compositions


100S 90S5C 80S15C 70S25C 60S35C


Ca(NO3)2 Æ 4H2O/g

8.3 7.5 6.7 5.8 5.0

/ 0.47 1.4 2.4 3.3


range of 76–38 lm were selected and conformed in disks (0.1 g) by uniaxial pressure (3 MPa). The incorporation of the drug into MBGs is followed by the common method which is used to adsorb small molecules into mesopores [7,8,15]. The MBG disks were soaked in the saturated acetone solution of TC for 3 days to adsorb the drug. The disks loaded with drug were dried in air and the final product was designated as MBG–TC in text. 2.2. Characterization of samples The samples before and after drug adsorption were characterized by X-ray diffraction (XRD), thermogravimetric analysis (TGA) and N2 adsorption–desorption. The weight ratio of TC to MBG was determined from the TGA. The MBGs before and after drug adsorption were evaluated by Fourier-transform infrared (FTIR) analysis to detect the interaction between TC and MBGs. The XRD patterns were obtained with a Bruker D4 powder X-ray diffractometer using Cu Ka radiation (40 kV, 40 mA). The TGA profiles were carried out between 298 and 973 K under air flow using a Mettler Toledo TGA/SDTA851e instrument, the flow rate is 80 mL/min and the heating rate is 5 K/min. N2 adsorption–desorption isotherms were measured with a Tristar3000 analyzer at 77 K. The samples were conformed in disks by an uniaxial pressure of 3 MPa and degassed under vacuum at 453 K for 5 h before measurement. Absorbance FTIR spectra were collected with the use of a Nicolet FTIR 360 (KBr method). 2.3. In vitro drug release The release profile of TC was obtained by soaking MBG–TC in a simulated body fluid (SBF) [16] at 310 K with constant stirring. In spite of the different weight percentage of MBG–TC, the specific ratio of TC in starting MBG–TC to SBF is kept at 0.1 mg/ml. The concentration of TC released into SBF was evaluated by ultraviolet–visible (UV) spectroscopy at 267 nm using a Jasco V-550 spectrophotometer. 3. Results and discussion 3.1. Characterization of MBG and MBG–TC 3.1.1. XRD patterns XRD patterns of MBG 80S15C (S denotes silica, C represents calcium, and the numbers are the molar percentages of compositions in the MBG materials) and MBG 80S15C– TC are shown in Fig. 2. The pattern of MBG 80S15C exhibits three diffraction peaks at 1.21, 2.06, and 2.38, which can be assigned to the 10, 11, and 20 reflections of a 2D hexagonal space group (p6m) [12]. For MBG 80S15C–TC, the intensity of the 10 peak is dramatically decreased, while the 11 and 20 peaks cannot be observed


L. Zhao et al. / Microporous and Mesoporous Materials 109 (2008) 210–215

the pore sizes of both materials are similar (see Table 2). Therefore, the higher / in the latter material loaded with TC should be attributed to the difference in CaO contents. The above results indicate that there is a certain interaction between TC and calcium species and the introduction of CaO in MBG framework may finally influence the TC loading ratio in MBG materials.

Fig. 2. XRD patterns of calcined MBG 80S15C and MBG 80S15C–TC.

in the XRD pattern. The above observation indicates that the drug molecules have been adsorbed into the channels of MBG, similar to the XRD pattern evolution before and after loading ibuprofen in mesoporous silicas [17]. 3.1.2. TGA results and N2 sorption analysis The physicochemical parameters measured by N2 sorption analysis of both MBGs and MBG–TC materials as well as the drug loading percentage in MBG–TC with different compositions are shown in Table 2. Comparing MBG–TC to MBG with the same composition, it can be seen that the surface area, pore volume and pore size are decreased, suggesting that TC has been loaded into the mesopores of MBG materials. The drug loading weight percentage (/) in MBG–TC materials with different compositions show a strong dependence on the composition of MBG. Increasing the CaO content from 0% (MBG 100) to 5% (MBG 90S5C) leads to a dramatic increase in / from 10.5% to 15.5%. Further increasing the CaO content from 5% to 35% (MBG 60S35C) results in a relatively small augment in / from 15.5% to 18.3%. Moreover, it should be noted that the pore volume and surface area of MBG 90S5C are larger than those of MBG 60S35C, and Table 2 Physicochemical properties of MBGs and MBG–TC samples

3.1.3. FTIR The FTIR spectra of TC, MBG 100S and MBG 100S– TC are shown in Fig. 3a. Three peaks located at 1520, 1456 and 1400 cm1 can be found in the spectrum of TC, and these peaks generally can be found in the same wavenumber in the spectrum of MBG 100–TC. In contrast, for MBG 80S15C–TC, the FTIR spectrum shows peaks at 1497, 1448, and 1414 cm1 in the same region, and the peaks are broadened (Fig. 3b). The above observations strongly indicate that the immobilized TC in the mesopores may chemically bond with calcium species in the framework of MBG materials. This interaction is possibly attributed to the chelation of TC and the calcium species in the pore walls [18]. 3.2. In vitro drug release The TC release profiles from five MBG–TC materials in SBF as a function of time are shown in Fig. 4. Generally the release curves can be separated to two parts: an initial fast release followed by a slow release pattern. The fast release is mainly caused by the dissolution of the drug which is physically adsorbed in MBGs, and the slow release may be attributed to the chemically adsorbed drugs. To make a semi-quantitative comparison, we introduce the following equation to predict the release rate of the fast release part of the profiles according to the theoretical analysis by Higuchi [19], which is used to dealing with the release from a planar system with porous matrix. a ¼ kt1=2


where a is the release amount after time t and k is the release constant. When the drug is dispersed in a matrix and the diffusion occurs through solvent-filled pores, the formulation of the constant k is given by






k ¼ f ðD; e; s; C; AÞ


269 215 338 287 229 194 147 128 159 146

0.45 0.35 0.46 0.28 0.31 0.19 0.24 0.15 0.27 0.15

6.1 5.5 5.5 5.3 5.2 5.0 5.2 5.1 4.6 4.5

/ 10.5 / 15.5 / 15.8 / 17.4 / 18.3

where D is the diffusivity of the drug in the solvent, s is the tortuosity factor of the system, e is the porosity of the matrix, A is the total amount of the drug present in the matrix and C is the solubility of the drug in the solvent used. Table 3 shows that the release constant k of each MBG in the first 10 h. We can see that during the fast release, the k of MBG 100S is 20.7, which is much larger than the other materials, the MBG90S5C has a release constant of 4.56, a little larger than that of MBG 80S15C and 70S25C (k is 3.23 and 3.05, respectively). However, it is noted that MBG 60S35C with the largest CaO content presents a

100S 100S–TC 90S5C 90S5C–TC 80S15C 80S15C–TC 70S25C 70S25C–TC 60S35C 60S35C–TC

Note: S is BET surface area, V is pore volume and D is BJH pore size of MBGs and MBG–TC materials. / is the TC loading weight percentage in MBG–TC materials.


L. Zhao et al. / Microporous and Mesoporous Materials 109 (2008) 210–215


Fig. 4. The release profiles of tetracycline as a function of time from MBGs with different compositions.

Table 3 The release constant of MBGs materials in the first 10 h, R is the liner regression coefficient and k is the release constant

Fig. 3. The FTIR spectra of: (a) TC, MBG 100S and MBG 100S–TC and (b) TC, MBG 80S15C, MBG 80S15C–TC.

quite abnormal larger k of 5.01 even compared to that of MBG 90S5C.




100S–TC 90S5C–TC 80S15C–TC 70S25C–TC 60S35C–TC

20.7 4.56 3.23 3.05 5.01

0.976 0.992 0.992 0.992 0.991

The TC release profile of MBG 100S shows a quite different behavior compared to those of other MBGs in the entire time range (0–120 h) under study. For MBG 100S, the time at which half of TC is released is 1.9 h, and the release profile reaches equilibrium at 24 h and the TC released content is higher than 90%. For the other four MBGs, the TC concentration released in SBF reaches equilibrium after 72 h, and the TC released content is 98%, 56%, 25%, 25%, 38% for MBG 100S, 90S5C, 80S15C, 70S25C, 60S35C, respectively. Again, the release profile and the equilibrium released concentration of TC for MBG 60S35C is abnormal compared to other MBGs. The different drug release behavior of MBGs can be explained by their difference in composition. In our studies, all the MBGs possess the same structure, similar pore size, pore volume and surface area, and the release experiment is carried out all in SBF, therefore the parameters of D, e, s and C should be similar for all MBGs, the difference in the release kinetics (k) in the first 10 h should be mainly attributed to A. It is proposed that there are two types of TC molecules adsorbed in the mesopores of MBG (Fig. 5): the physically adsorbed TC (type I) and chemically adsorbed TC (type II) through chelation with calcium species on the pore wall of MBGs. In the initial release period (<10 h) of MBG–TC in SBF, the fast release kinetics may associate predominantly with type I TC. Although MBG 100S–TC without CaO content has the smallest / (Table 2), the majority of the


L. Zhao et al. / Microporous and Mesoporous Materials 109 (2008) 210–215




Ca2+ TC Ca2+ TC —Ca TC —Ca TC —Ca TC



TC Ca —

TC Ca —

—Ca TC

TC Ca —



pore wall

pore wall

TC Ca —



TC Ca — TC Ca —


—Ca TC —Ca TC



TC TC Ca —

Fig. 6. The FTIR spectra of unreacted MBG 80S15C–TC and 80S15C– TC reacted in SBF for 24 h. Fig. 5. Proposed scheme of different types of TC molecules adsorbed in MBG and during the release process: (I) physically adsorbed TC, (II) TC chelated with calcium on the pore wall, and (III) TC chelated with calcium ions in SBF solution.

TC (>72 h) in MBG 60S35C is larger than that in MBG 80S15C and 70S25C. 3.3. In vitro bioactivity of MBGs 80S15C–TC

drug adsorbed in the material is type I TC according to our IR results. Compared MBG 90S5C to MBG 100S, it is proposed that both types I and II TC molecules are present in the mesopores and the fraction of type I TC is dramatically decreased due to the presence of calcium species in the former material. Increasing the content of CaO in MBGs may further decrease the fraction of type I TC and in turn, A, which should represent the initial concentration of type I TC in MBG, decreases as the CaO content increases because the initial ratio of TC/SBF is a constant. As a consequence, the release kinetics of MBG–TC is greatly influenced by their composition (more specifically, the CaO content), similar to the influence of composition upon A we discussed above. Moreover, the change in equilibrium released concentration of TC (>72 h) in different MBG– TC materials can be understood following our proposed mechanism. Increasing the CaO content may increase the fraction of type II TC in MBG–TC materials, which are much more difficult to release compared to type I TC molecules. In the above discussion, the abnormal release behavior of MBG 60S35C is not considered. In fact, during the TC release process in SBF, the calcium ions in solutions may chelate with TC (Fig. 5), this type III TC should be taken into account. The chelation between TC and calcium in solution (type III TC) may compete with the chelation between TC and calcium in the solid framework (type II TC) and finally influence the release behavior. In our previous study, it has been observed that the maximum calcium concentration released by MBG 60S35C is about two times higher than that of other MBGs [13]. Therefore, the higher concentration of calcium ions in solution favors the formation of type III TC and fastens the release kinetics. As a result, the release rate of MBG 60S35C is eventually larger and the equilibrium released concentration of

The IR spectra of MBG 80S15C–TC and MBG 80S15C–TC after soaking in SBF for 24 h are shown in Fig. 6. A new peak at 958 cm1 can be observed, which corresponds to the P–O band [20]. The IR observation suggests that an apatite layer can be formed on the surface of MBG–TC after immersing in SBF for 24 h, in accordance with our previous results [13]. 4. Conclusions In this work, we studied the in vitro drug release behaviors of MBGs with different compositions for the antiinflammatory TC. The results show that the CaO contents of MBGs significantly influence the release kinetic because of the chelation between the drug and the calcium species. The release rate of MBG 100S without CaO in the framework is much faster than the other MBGs. MBG 90S5C, 80S15C, and 70S25C show slow release profiles and the release rate decreases as the CaO content increases. However, in the case of MBG 60S35C, as the CaO content is further increased, the release of the drug in the first 10 h is actually accelerated even compared to MBG 90S5C. Our results clearly depict the composition/drug release behavior in MBG biomaterials. By understanding the composition/drug release and composition/bioactivity correlations of MBGs, it is expected that MBG-drug may be used as bone repair material with build-in anti-inflammation function in bone reconstruction applications. Acknowledgments This work is supported by the NSF of China (20573021), the Ministry of Education of China (20060246010, FANEDD 200423), the State Key Research

L. Zhao et al. / Microporous and Mesoporous Materials 109 (2008) 210–215

Program (2006CB0N0302), Shanghai Science Committee (05SU07098, 05DJ14005), and Shanghai HuaYi Group. References [1] L.L. Hench, R. Splinter, W. Allen, T. Greenlee, J. Biomed. Mater. Res. 2 (1971) 117. [2] L.L. Hench, J.M. Polak, Science 295 (2002) 1014. [3] M. Vallet-Regi, J. Chem. Soc., Dalton Trans. (2001) 97. [4] G. Goller, F.N. Oktar, L.S. Ozyegin, E.S. Kayali, E. Demirkesen, Mater. Lett. 58 (2004) 2599. [5] P. Sepulveda, J.R. Jones, L.L. Hench, J. Biomed. Mater. Res. 59 (2002) 340. [6] M. Vallet-Regi, C.V. Ragel, A.J. Salinas, Eur. J. Inorg. Chem. (2003) 1029. [7] P. Horcajada, A. Ramila, J. Perez-Pariente, M. Vallet-Regi, Micropor. Mesopor. Mater. 68 (2004) 105. [8] A.L. Doadrio, E.M.B. Sousa, J.C. Doadrio, J.P. Pariente, I. Izquierdo-Barba, M. Vallet-Regi, J. Control. Release 97 (2004) 125. [9] X.X. Yan, C.Z. Yu, X.F. Zhou, J.W. Tang, D.Y. Zhao, Angew. Chem., Int. Ed. 43 (2004) 5980.


[10] Q.H. Shi, J.F. Wang, J.P. Zhang, J. Fan, G.D. Stucky, Adv. Mater. 18 (2006) 1038. [11] A. Lopez-Noriega, D. Arcos, I. Izquierdo-Barba, Y. Sakamoto, O. Terasaki, M. Vallet-Regi, Chem. Mater. 18 (2006) 3137. [12] X.X. Yan, H.X. Deng, X.H. Huang, G.Q. Lu, S.Z. Qiao, D.Y. Zhao, C.Z. Yu, J. Non-Cryst. Solids 351 (2005) 3209. [13] X.X. Yan, X.H. Huang, C.Z. Yu, H.X. Deng, Y. Wang, Z.D. Zhang, S.Z. Qiaoc, G.Q. Lu, D.Y. Zhao, Biomaterials 27 (2006) 3396. [14] M. Wikesjo¨, J. Baker, L. Christersson, J. Genco, M. Lyall, S. Hic, M. Diflorio, P. Terranova, J. Periodont. Res. 21 (1986) 322. [15] I. Izquierdo-Barba, A. Martinez, A.L. Doadrio, J. Perez-Pariente, M. Vallet-Regi, Eur. J. Pharm. Sci. 26 (2005) 365. [16] T. Kokubo, H. Kushitani, S. Sakka, T. Kitsugi, T. Yamamuro, J. Biomed. Mater. Res. 24 (1990) 721. [17] F.Y. Qu, G.S. Zhu, H.M. Lin, W.W. Zhang, J.Y. Sun, S.G. Li, S.L. Qiu, J. Solid State. Chem. 179 (2006) 2027. [18] E.C. Newman, C.W. Frank, J. Pharm. Sci. 65 (1976) 1728. [19] T. Higuchi, J. Pharm. Sci. 52 (1963) 1145. [20] M. Vallet-Regi, A.M. Romero, C.V. Ragel, R.Z. LeGeros, J. Biomed. Mater. Res. 44 (1999) 416.