Biomaterials 21 (2000) 193 } 198

Silica xerogel as an implantable carrier for controlled drug delivery*evaluation of drug distribution and tissue e!ects after implantation
Pirjo Kortesuo *, Manja Ahola , Stefan Karlsson , Ilkka Kangasniemi , Antti Yli-Urpo , Juha Kiesvaara
Orion Corporation, Orion Pharma, Pharmaceutical Development Department, P.O. Box 425, FIN-20101, Turku, Finland Institute of Dentistry, University of Turku, Lemminka ( isenkatu 2, FIN-20520, Turku, Finland Received 30 September 1998; accepted 15 July 1999

Abstract The purpose of the present study was to examine controlled delivery of toremifene citrate from subcutaneously implanted silica xerogel carrier and to evaluate silica xerogel related tissue e!ects after implantation. Toremifene citrate was incorporated into hydrolyzed silica sol in a room temperature process. Toremifene citrate treated silica xerogel implants were tested both in vitro and in vivo using healthy mice. Silica xerogel with tritium-labelled toremifene was implanted subcutaneously in mice for 42 d. To determine the amount of tritiated toremifene remaining in the silica discs at the implantation site, the discs were excised periodically and radioactivity measured. The amount of tritiated toremifene in the implant after 42 d was still about 16% and the amount of silica xerogel about 25%. In a histopathological study silica xerogel did not show any tissue irritation at the site of the implantation. A "brotic capsule was formed around the implant. No silica xerogel related histological changes in liver, kidney, lymph nodes and uterus were observed during the implantation period. The silica xerogel discs showed a sustained release of toremifene citrate over 42 d. Histologically, toremifene-related changes in the uterus were also detectable at all studied time points. These "ndings suggest that silica xerogel is a promising carrier material for implantable controlled drug delivery system. 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Sol}gel processed silica xerogel; Implantation study; Drug delivery; Toremifene citrate; Tissue distribution

1. Introduction Sol}gel processed porous silica glasses have a considerable potential as carriers for controlled drug release [1}4]. In recent years, processing factors controlling the degradation rate of the matrix in vitro have been extensively studied [1,3]. However, in vivo degradation and drug delivery of low-temperature processed silica xerogel are not well documented. Parenterally implanted sintered sol}gel produced glasses containing 85% silica and also pure silica glasses are biocompatible [5,6]. Bioactive glasses in soft tissue or

* Corresponding author. Tel.: #358-2-2727690; fax: #358-22727273. E-mail address: pirjo.kortesuo@orion." (P. Kortesuo) Presented in part at the 23rd Annual Meeting of the Society for Biomaterials, New Orleans, 1997.

bone enhance "broblast or osteoblast formation resulting in increased collagen formation, which in turn bonds bioactive glass implants to bone or soft tissue [5,7]. However, crystalline silicon dioxide in the form of aerosol is known to cause a rapid in#ux of in#ammatory cells and increased collagen deposition in lungs and histological changes in the pulmonary lymph nodes [8]. Toremifene citrate, a non-steroidal antiestrogenic compound, which has antitumor activity in breast and endometrial cancer, was used as model drug in this study [9]. Local therapy of antiestrogens after hormone-dependent breast cancer surgery could give a targeted and long-lasting disease control while minimizing systemic distribution. In this study toremifene citrate/H-toremifene was incorporated into the silica gel matrix during a room-temperature gelation process and silica gel was dried to a constant weight at 403C without sintering at a higher temperature. The objectives of this work were: (i) to

0142-9612/00/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 9 9 ) 0 0 1 4 8 - 9

194

P. Kortesuo et al. / Biomaterials 21 (2000) 193 } 198

evaluate the capability of silica xerogel to deliver toremifene in a sustained manner, (ii) to study tissue distribution of toremifene and (iii) to evaluate histological e!ects after implantation of the silica xerogel itself on the uterus, liver, kidneys, lymph nodes and at the implantation site.

Degradation of silica xerogel matrix was determined by measuring dissolved Si(OH) spectrophotometrically  as a molybden blue complex at 820 nm [11]. 2.3. Animal experiments The animals were housed according to the recommendations in DHEW pub. (NIH) 85-23 entitled &Guide for the care and use of laboratory animals'. Thirty female mice (C57Bl, Denmark) with an average weight of about 19.6 g (SD 1.2 g) received implants subcutaneously on the back on either side of the backbone with tritium-labelled toremifene loaded silica implants (a total of 2 implants were placed in each mouse) and 30 mice with untreated silica implants. The H-toremifene dose was about 80 Ci/kg (0.8 Ci/implant), toremifene citrate 350 mg/ kg (approx. 3.4 mg/implant) and silica xerogel about 1.53 g/kg body weight. 2.4. Histology The animals were killed at 7, 14, 21, 28, 35 and 42 d (5 mice per time period) after implantation. The silica xerogel implants on the left side of the backbone were removed together with the surrounding tissue and "xed in 70% ethanol and embedded in PMMA (Technovit). Sections of 20 m were stained with toluidine blue. Samples of liver, kidney, mesenterial lymph node and uterus were "xed in bu!ered formaldehyde and embedded in para$n. Sections of 6 m were stained with hematoxylin}eosin. Samples were evaluated using light microscopy. The samples of the implantation site were studied from all interim kills but liver, kidney, lymph node and uterus only from the animals treated 7, 28 and 42 d. 2.5. Drug release and silica xerogel erosion studies The mice were killed, whereupon the silica xerogel implants on the right side of the backbone were cut o! from the surrounding "brous capsule and dried at room temperature in a desiccator for 24 h. The percentage of implant remaining at each time point was calculated. To determine the amount of toremifene remaining in the implants, the dried implants were dissolved in 0.1 N NaOH solution and the activity was measured in liquid scintillation counter. 2.6. Organ radioactivity counting After killing the mice, the tissue samples taken from liver, kidney, uterus, lung, blood and application area were burned in an oxidizer (Junitek, Kaarina, Finland) and the radioactivity was measured by scintillation counting.

2. Materials and methods 2.1. Preparation of toremifene citrate incorporated silica gels Silica gel was prepared by the hydrolysis and polycondensation of tetraethoxysilane (TEOS) with polyethylene glycol (PEG, M 4600), water and acetic acid in a  mole ratio of TEOS : PEG : H O : CH COOH"   1.0 : 0.0012 : 14.2 : 0.5 at room temperature. Toremifene citrate (Orion Corporation, Turku, Finland) was dissolved into the silica solution after 1 h hydrolysis at the concentration of 33 mg/g. Tritiated toremifene was synthetized as described earlier [10]. The speci"c activity of the material was 16 Ci/mmol. It was added to the silica sol to produce a "nal radioactivity of 16 Ci/g. Solution was pipetted into round moulds (100 l/mould) for polycondensation and aging at 403C for 18 h. The polymerized silica gels were soaked in water to leach out residual organic matter within the gel and dried at 403C to a constant weight. The silica xerogel implants were round with the diameter approximately 4.7 mm and height 0.9 mm weighing about 15 mg. 2.2. In vitro dissolution test The dissolution pro"les (each data point is the mean of 2 values) of H-toremifene and silica from test specimens were studied using the USP XXIII dissolution apparatus II (paddle method, Sotax AT6, Basel, Switzerland) at a constant temperature (373C). Simulated body #uid (SBF, pH 7.4) containing 0.5% (m/v) sodiumdodecylsulphate was used as dissolution medium. SBF was prepared by dissolving reagent-grade NaCl (136.8 mM), NaHCO (4.2 mM), KCl (3.0 mM), K HPO ;3H O     (1.0 mM), MgCl ;6H O (1.5 mM), CaCl ;2H O     (2.5 mM) and Na SO (0.5 mM) in distilled water. The   solution was bu!ered at pH 7.4 with tris(hydroxymethyl)aminomethane (50 mM) and hydrochloric acid. The volume of dissolution medium was 250 ml and the rotation speed 50 rpm. At each sample interval a 2 ml sample was withdrawn and replaced with fresh medium. To study the amount of tritiated toremifene released from the matrix, samples of 0.5 ml were mixed with 10 ml of scintillation liquid (Ready Safe2+, Beckmann, CA, USA) and measured by scintillation counting (Model 81 000, LKB-Wallac, Turku, Finland).

P. Kortesuo et al. / Biomaterials 21 (2000) 193 } 198

195

3. Results 3.1. In vitro and in vivo release of tritium-labelled toremifene and erosion of silica xerogel matrix About 90% of the loaded tritium-labelled toremifene was released after 99 h in vitro. The amount of silica released in 99 h is equivalent to about 65% weight loss of silica xerogel matrix (Fig. 1). Weight loss of silica gel matrix in vivo was about 75 wt% during 42 d. The erosion rate was fast during the "rst 28 d and then decreased (Fig. 2). By an estimate from the dissolution pro"le, the silica xerogel implant weighing about 15 mg is totally absorbed after 60}70 d. The silica xerogel discs showed sustained release of toremifene during the test period. About 40% of loaded toremifene eluted during the "rst week. After that the release of toremifene slowed down and 50% was released in 3 weeks. The amount of H-toremifene remaining in the implant after 42 d was still about 16% (Fig. 2). The radioactivity of tritiated toremifene was detected upto 42 d in the liver, kidneys, and lung. The radioactivity reached a maximum value at 14 d and then remained relatively constant until the end of the implantation test. Radioactivity was not found in blood or uterus (Fig. 3). Toremifene-induced uterine glandular hyperplasia, the characteristic hormonal change seen in mice treated with this drug, was observed at all studied time points. 3.2. Ewects after implantation The local e!ect of the silica xerogel implant was restricted to the close vicinity of the implant. A well-

Fig. 2. Percentage of remaining silica xerogel implant and H-toremifene radioactivity at di!erent time points in vivo.

Fig. 3. Biodistribution of H-toremifene radioactivity at di!erent time points in some of the organs in mice.

Fig. 1. In vitro release of H-toremifene and silica from silica xerogel matrix.

organized "brous capsule was formed around the implant from 14 d time point onward (Fig. 4A). The implant did not cause necrosis and the in#ammatory response observed was mainly due to a minor in"ltration of macrophages at time points 14 d and later. Acute in#ammation was observed only at the 7-d point. Phagocytic macrophages were present close to the implant at 14 d or later (Fig. 4B). No accumulation of dissolved material occurred. Toremifene released from the implant caused tissue irritation, necrosis and an unfavorable environment for phagocytic macrophages causing an accumulation of

196

P. Kortesuo et al. / Biomaterials 21 (2000) 193 } 198

Fig. 4. (A) An untreated implant (S) in a mouse 35 d after implantation. The implant is surrounded by a "brotic capsule (arrow, original magni"cation 16;). (B) The surface (S) of an untreated implant 21 d after implantation. The "brous capsule (arrow) is well formed and phagocytic macrophages (open arrowheads) are abundant close to the surface (original magni"cation 100;).

dissolved silica inside the "brotic capsule. These "ndings were present from the 14 d point onward. No extensive silica xerogel related histological changes could be observed in the liver, lymph nodes or kidney at the 1.5 g/kg dose.

4. Discussion The silica xerogel matrix gave sustained release of toremifene citrate for more than six weeks. This was

con"rmed by erosion/release studies, activity measurements of tritiated toremifene in tissue samples and by histological studies. Histologically, toremifene-related changes in the uterus were detectable at all studied time points. About 40% of loaded toremifene was eluted from silica xerogel matrix during the "rst week. After that the release rate slowed down and about 16% of the loaded amount was still remaining after 42 d. Therefore, radioactivity in tissue samples reached a maximum value at 14 d and remained relatively constant in the liver, lung and kidneys throughout the test period. It was

P. Kortesuo et al. / Biomaterials 21 (2000) 193 } 198

197

demonstrated earlier that as a lipophilic agent, tritiumlabelled toremifene was distributed throughout the body with highest radioactivity in the liver and lung [10]. Toremifene citrate is metabolized without "rst-pass metabolism. After enterohepatic circulation, the metabolites are excreted mainly in feces. In tritiated toremifene the radioactivity is not eliminated within the metabolic pathways [10]. Compared to sintered silica xerogels (7003C) where a silica xerogel matrix was impregnated with toremifene citrate, incorporation of drug molecules into the non-sintered silica xerogel matrix retarded the dissolution of toremifene citrate both in vitro and in vivo. However, a non-sintered silica xerogel matrix degraded faster in vivo than sintered silica gel in an earlier in vivo study [12]. It is known from literature that sintering decreases the dissolution rate of silica [13]. The in vitro toremifene citrate release was faster than in tissue. About 90% of the loaded amount was released after 99 h and the amount of remaining silica was about 35% at the same time point. Weight loss of silica xerogel matrix in vivo after 42 d was about 75%. The results show that the in vitro release pro"le does not correlate to the in vivo pro"le on a real time basis. The origin of the di!erence of the behavior between in vitro and in vivo data is probably due to the di!erence in the surface area of the implant to the volume ratio of the solution at subcutaneous implantation site and in dissolution study. However, the observed release pro"les suggest that erosion of silica xerogel controls the release of toremifene citrate from the matrix. The untreated silica xerogel implant did not cause irritation of the surrounding tissue. Local e!ects at the implantation site are considered as a normal woundhealing process. Acute in#ammation was observed only at the 7-d point probably as a postoperative response. Implantation of foreign material caused acute in#ammation of relatively short duration. The "nal stage of healing response was the development of a "brous capsule at the 14-d point around the implant, which typically has been considered a sign of biocompatibility of the material [14,15]. Sol}gel produced glasses also appeared nontoxic in in vitro experiments showing normal growth of "broblasts [6]. A "brous capsule may retard the release of drug from implanted material and distribution to other organs [14]. However, in this study toremifenerelated changes in the uterus were observed at all time points. Toremifene-treated implants caused irritation at the implantation site throughout the study period. The difference in tissue reactions between untreated silica xerogel and with H-toremifene treated implant may be due to irritating e!ect of toremifene, which causes an unfavorable environment for phagocytic macrophages. In earlier studies toremifene has proved to have an irritating e!ect following intraperitoneal administration (unpublished result). As an antiestrogenic agent, tor-

emifene may also stimulate "broblast growth by TGFor radioactivity may produce tissue in#ammation [16]. Silica xerogel dose of 1.5 g/kg did not produce any histological changes in liver, kidney, lymph nodes and uterus during the test period (42 d). However, earlier studies claim that silica has several biological e!ects, which a!ect the immune system particularly [17]. Especially crystalline quartz (silicon oxide) is known to be cytotoxic in macrophages and following chronic inhalation, quartz dust causes increased granulation tissue in the pulmonary or tracheobronchial lymph nodes of rats [17}19]. The destructive e!ect of crystalline quartz is probably due to the property of the crystalline form to disrupt various biological membranes. However, these "ndings were not observed with amorphous sol}gel produced silica xerogel. In conclusion, silica xerogel gave sustained release of tritium-labelled toremifene for six weeks. The release of toremifene was controlled by the bioerosion of the silica xerogel matrix. No silica xerogel related changes in liver, kidney, uterus or lymph nodes were observed even though the applied amount of silica xerogel was high (1.5 g/kg body weight). Furthermore, silica xerogel was biocompatible with the surrounding tissue. Thus silica xerogel seems to be promising carrier material for implantable controlled drug delivery system.

References
[1] Bo K ttcher H, Kallies K-H, Haufe H. Model investigations of controlled release of bioactive compounds from thin metal oxide layers. J Sol}Gel Sci Technol 1997;8:651}4. [2] Nicoll SB, Radin S, Santos EM, Tuan RS, Ducheyne P. In vitro release kinetics of biologically active transforming growth factor1 from a novel porous glass carrier. Biomaterials 1997;18:853}9. [3] Tan BH, Santos EM, Ducheyne P. Ultramicroscopic pore size and porosity of xerogels for controlled release of biological molecules. Fifth World Biomaterials Congress, vol. 161, 1996. [4] Unger K, Rupprecht H, Valentin B, Kircher W. The use of porous and surface modi"ed silicas as drug delivery and stabilizing agents. Drug Dev Ind Pharm 1983;9:69}91. [5] Wilson J, Pigott GH, Schoen FJ, Hench LL. Toxicology and biocompatibility of bioglasses. J Biomed Mater Res 1981; 15:805}17. [6] Palumbo G, Avigliano L, Strukul G, Pinna F, del Principe D, Angelo ID, Annicchiarico-Petruzzelli M, Locardi B, Rosato N. Fibroblast growth and polymorphonuclear granulocyte activation in the presence of a new biologically active sol}gel glass. J Mater Sci Mater Med 1997;8:417}21. [7] Klein CPAT, Li P, de Blieck-Hogervorst JMA, de Groot K. E!ect of sintering temperature on silica gels and their bone bonding ability. Biomaterials 1995;16:715}9. [8] Absher MP, Trombley L, Hemenway DR, Mickey RM, Leslie KO. Biphasic cellular and tissue response of rat lungs after eightday aerosol exposure to the silicon dioxide cristobalite. Am J Pathol 1989;134:1243}51. [9] Kallio S, Kangas L, Blanco G, Johansson R, Karjalainen A, Perila K M, Piippo I, Sundquist H, So K dervall M, Toivola R. A new triphenylethylene compound, Fc-1157a. I. Hormonal e!ects. Cancer Chemother Pharmacol 1986;17:103}8.

198

P. Kortesuo et al. / Biomaterials 21 (2000) 193 } 198 [15] Anderson JM. In vivo biocompatibility of implantable delivery systems and biomaterials. Eur J Pharm Biopharm 1994; 40:1}8. [16] Border WA, Noble NA. Targeting of TGF- for treatment of disease. Nature Med 1995;1:1000}1. [17] Burns CA, Zarkower A, Ferguson FG. Murine immunological and histological changes in response to chronic silica exposure. Environ Res 1980;21:298}307. [18] Allison AC, Harrington JS, Bibeck M. An examination of the cytotoxic e!ects of silica on macrophages. J Exp Med 1966;124: 141}61. [19] Gopinath C, Prentice DE, Lewis DJ. Atlas of experimental toxicological pathology. Norwell, MA, USA: MTP press, 1987.

[10] Kangas L, Haaparanta M, Paul R, Dirk R, Sipila K H. Biodistribution and scintigraphy of C-toremifene in rats bearing DMBAinduced mammary carcinoma. Pharmacol Toxicol 1989;64:373}7. [11] Koch OG, Koch-Dedic GA, Handbuch der Spurenanalyse. Berlin: Springer, 1974. p. 1105. [12] Kortesuo P, Ahola M, Karlsson S, Kangasniemi I, Kiesvaara J, Yli-Urpo A. Sol}gel-processed sintered silica xerogel as a carrier in controlled drug delivery. J Biomed Mater Res 1999;44:162}7. [13] Li P, Ohtsuki C, Kokubo T, Nakanishi K, Soga N. Apatite formation induced by silica gel in a simulated body #uid. J Am Ceram Soc 1992;75:2094}7. [14] Anderson JM, Niven H, Pelagalli J, Olano! LS, Jones RD. The role of the "brous capsule in the function of implanted drug-polymer sustained release systems. J Biomed Mater Res 1981;15: 889}902.