Chitosan Capsules for Colon-Specific Drug Delivery: Improvement of Insulin Absorption from the Rat Colon

HIDEYUKI TOZAKI*, JUNTA KOMOIKE*, CHIKA TADA*, TAKAKO MARUYAMA*, AKIRA TERABE, TSUTOMU SUZUKI, AKIRA YAMAMOTO*X, AND SHOZO MURANISHI*
Received January 17, 1997, from the *Department of Biopharmaceutics, Kyoto Pharmaceutical University, Misasagi, Yamashina-ku, Kyoto 607, Japan, and Aicello Chemical Company, Toyohashi 441-11, Japan. Final revised manuscript received June 2, 1997. Accepted for publication June 20, 1997X.
Abstract 0 The objective of this study was to estimate colon-specific insulin delivery with chitosan capsules. In vitro drug release experiments from chitosan capsules containing 5(6)-carboxyfluorescein (CF) were carried out by the Japan Pharmacopoeia (J. P.) rotating basket method with some slight modifications. The intestinal absorption of insulin was evaluated by measuring the plasma insulin levels and its hypoglycemic effects after oral administration of the chitosan capsules containing insulin and additives. Little release of CF from the capsules was observed in liquid 1, an artificial gastric juice (pH 1), or in liquid 2, an artificial intestinal juice (pH 7). However, the release of CF was markedly increased in the presence of rat cecal contents. A marked absorption of insulin and a corresponding decrease in plasma glucose levels was observed following the oral administration of these capsules that contain 20 IU of insulin and sodium glycocholate (PA% ) 3.49%), as compared with the capsules containing only lactose or only 20 IU of insulin (PA% ) 1.62%). The hypoglycemic effect started from 8 h after the administration of chitosan capsules when the capsules entered the colon, as evaluated by the transit time experiments with chitosan capsules. These findings suggest that chitosan capsules may be useful carriers for the colon-specific delivery of peptides including insulin.
terms of site-specific drug release to the colon. This poor reliability is due to large variations in transit times and pH depending on diet, food intake, intestinal motility, and disease states. Alternatively, Saffran and his co-workers developed a new approach to oral delivery of peptide and protein drugs.6 They coated peptide drugs with polymers cross-linked with azoaromatic groups to protect orally administered drugs from digestion in the stomach and small intestine. When the azopolymer-coated drugs reached the large intestine, the indigenous microflora reduced the azo bonds, broke the crosslinks, and degraded the polymer film, thereby releasing the drug into the lumen of the colon. These authors demonstrated the ability of the azopolymer coating to protect and deliver orally administered peptide drugs, such as insulin and vasopressin. However, they encountered some problems, such as variability in absorption rate, when administering coated drugs. Moreover, the safety of azopolymer, which is a synthetic polymer (not a natural product), has not been fully evaluated. Chitosan is a high molecular weight cationic polysaccharide derived from naturally occurring chitin in crab and shrimp shells by deacetylation. Chitosan has previously been employed as a pharmaceutical excipient in oral drug formulations to improve the dissolution of poorly soluble drugs or for the sustained release of drugs by a process of slow erosion from a hydrated compressed matrix.7 The compound is considered to be nontoxic, with an oral LD50 in mice of >16 g/kg. Recently, it was reported that this compound is also degraded by microflora, which are richly distributed in the colon. In this study, therefore, we prepared the capsules using chitosan and examined the colon-specific delivery of insulin with chitosan capsules in rats. We also examined the effect of various additives, such as absorption enhancers and protease inhibitors that are incorporated with insulin in these capsules, on insulin absorption from the colon with these capsules in rats.

Introduction
The intestinal absorption of peptide and protein drugs after oral administration is generally poor. This poor absorption has been attributed to the extensive hydrolysis of these drugs by the proteolytic enzymes in the gastrointestinal tract and/ or their poor membrane permeability characteristics.1 Therefore, various approaches, such as alternative routes, absorption enhancers, protease inhibitors, chemical modification, and dosage forms, have been examined to overcome the delivery problems of these peptides and proteins via the gastrointestinal tract.2,3 On the other hand, there has recently been increasing interest in targeting peptide and protein drugs to the colon because of the relatively low activity of proteolytic enzymes in the colon.4 Thus, these drugs, which are susceptible to proteolytic degradation and deactivation in the upper small intestine, may be more effectively absorbed from the colon. Based on this finding, many dosage forms, such as timecontrolled and pH-dependent release dosage forms have been examined for the specific delivery of these drugs to the colon.5 In the case of time-controlled release dosage forms, the drug is released after a specific time interval based on the expected transit time for the device to reach the colon. In the application of pH-dependent release dosage forms, it is assumed that the coating is stable in the low pH of the stomach and the neutral pH of the small intestine, and dissolves at the pH of the colon. These principles, however, are not very reliable in
X

Experimental Section
MaterialssBovine insulin, sodium glycocholate (Na-GC), aprotinin, and 5(6)-carboxyfluorescein (CF) were purchased from Sigma Chemical Company (St. Louis, MO). Bacitracin, soybean trypsin inhibitor (STI), sodium oleate, n-dodecyl--D-maltopyranoside (LM), Glazyme insulin-EIA test, and glucose B test Wako were obtained from Wako Pure Chemical Industries Company (Osaka, Japan). Preparation of Suspension of Cecal ContentssFresh cecal contents from nonfasted rats were suspended in two volumes of bicarbonate buffer (NaHCO3, 9.240 g; Na2HPO412H2O, 7.125 g; NaCl, 0.470 g; KCl, 0.450 g; CaCl22H2O, 0.073 g; MgCl26H2O, 0.087 g/L). The pH of the buffer was adjusted to 7.0 by bubbling CO2 gas through it prior to use. The suspension was filtered through four layers of gauze.8,9 Preparation of Chitosan CapsulessThe chitosan capsules were obtained from Aicello Chemical Company Ltd. (Toyohashi, Japan). The chitosan capsules contained CF (1.0 mg) or insulin, with or without bacitracin (14 mg), aprotinin (4.0 mg), and STI (3.0 mg) as

Abstract published in Advance ACS Abstracts, August 1, 1997.

1016 / Journal of Pharmaceutical Sciences Vol. 86, No. 9, September 1997

S0022-3549(97)00018-X CCC: $14.00

1997, American Chemical Society and American Pharmaceutical Association

Figure 2sThe gastrointestinal transit of chitosan capsules. Key: (a) small intestine; (b) large intestine; (c) excretion.

Figure 1sCross section of a chitosan capsule.

protease inhibitors, or sodium glycocholate (9.8 mg), sodium oleate (6.0 mg), and LM (10 mg) as absorption enhancers. A cross section of a chitosan capsule is shown in Figure 1. The mean diameter and weight of these capsules were 3.5 1.6 mm and 1.2-1.5 mg, respectively. The surfaces of these capsules were coated with hydroxypropyl methylcellulose phthalate (HPMCP) as an enteric coating material. The HPMCP was applied to the capsule by the following procedure. First, HPMCP was dissolved in acetone/ethanol solvent (w/w ) 1.1), and then chitosan capsules were dipped in this HPMCP solvent to obtain chitosan capsules coated with HPMCP. In Vitro Drug Release Experiments from Chitosan CapsulessRelease studies of drugs from chitosan capsules were carried out by the Japan Pharmacopoeia (J. P.) rotating basket method with some slight modifications. CF, which was encapsulated in the chitosan capsules, was used as a model water-soluble compound. Liquid 1 (a model medium of an artificial gastric juice for the J. P. disintegration test), liquid 2 (a model medium of an artificial intestinal juice for the J. P. disintegration test), phosphate buffered saline (pH 6.0), and suspensions of the cecal contents were used as media in these experiments. The pH values of liquids 1 and 2 are 1 and 7, respectively. The rotation speed of the baskets was 100 rpm. Samples (100 L) were taken every 60 min, and the amount of CF released from the capsules was determined spectrofluorometrically at excitation and emission wavelengths of 490 and 520 nm, respectively. The electron microscopic photograph (40) of the chitosan capsule in liquid 2 at 5 h and in the suspension of cecal contents was scanned at 12 h after starting the in vitro drug release experiments. In Vivo Oral Administration of Chitosan CapsulessColonspecific delivery of insulin after oral administration was investigated with chitosan capsules. Male Wistar rats, 200-270 g, were fasted for 8 h before the experiments, and eight or ten capsules were orally administered to stomach via polyethylene tubing under light ether anesthesia. Two hundred microliters of distilled water were administered to rats every 3 h. The gastrointestinal transit of chitosan capsules was determined by counting the number of capsules in the gastrointestinal lumen by celiotomy at certain times after their oral administration. The colonic absorption of insulin was estimated by its hypoglycemic effects and plasma insulin concentrations. Blood samples were periodically collected and were centrifuged at 10 000 rpm for 5 min and plasma samples were obtained. The glucose concentrations in the plasma were determined by a glucose oxidase method. The decrement in plasma glucose level (D%) was calculated by a method similar to that described by Hirai et al.2 from the following equation:

Figure 3sRelease of CF from chitosan capsules, determined by the J. P. rotating basket method. Key: (b) liquid 1 (J. P.) f liquid 2 (J. P.) f 33% suspension of cecal contents; (O) phosphate buffered saline (pH 6.0). Results are expressed as the mean SE of two to four experiments. PA% ) D%p.o. Dosei.v. 100 D%i.v. Dosep.o. (2)

The insulin concentration in plasma was determined by an enzyme immunoassay kit (GLAZYME Insulin-EIA TEST). Statistical AnalysessResults are expressed as the mean ( SE, and statistical analyses were performed with the Students t test.

Results
In Vivo Gastrointestinal Transit ExperimentssThe gastrointestinal transit of chitosan capsules after oral administration is shown in Figure 2. After oral administration, the chitosan capsules were eliminated from the stomach in 2 h, moved into the small intestine after 2 to 6 h, reached the large intestine after 6 to 12 h, and then were excreted from the body. In Vitro Release of Drugs from the Chitosan CapsulessTo examine whether the chitosan capsules were degraded in the large intestine or not, the amount of CF released from the chitosan capsules was determined in the presence of rat cecal contents. In this study, rat cecal contents were used as a model of large intestinal contents because there was no significant difference in the release rate of CF from the capsules into cecal and colonic contents in our pilot studies. The release-time profile of CF from chitosan capsules, determined by the J. P. rotating basket method, is shown in Figure 3. Based on the results of the gastrointestinal transit time of chitosan capsules, drug release was investigated in liquid 1 (an artificial gastric juice, pH 1) between 0 and 2 h, in liquid 2 (an artificial intestinal juice, pH 7) between 2 and 6 h, and in a suspension of rat cecal contents between 6 and 12 h. A small amount of CF from the capsules was released in liquids 1 and 2. However, the release of CF was markedly increased in the presence of rat cecal contents. Scanning electron microscopic photographs of chitosan capsules in liquid

D% ) 1 -

AUC6 f 24 100% (24 - 6) h

100

(1)

In this work, the area above the 100% line was ruled out for calculating the AUC6 f 24. The D% value was 4.4% when the insulin solution (0.1 IU/rat) was intravenously administered by bolus injection. The pharmacological availability (PA%) was calculated from the following equation:

Journal of Pharmaceutical Sciences / 1017 Vol. 86, No. 9, September 1997

Figure 4sScanning electron microscopic photograph of chitosan capsules in (A) liquid 1 for 2 h and liquid 2 for 5 h and (B) in the suspension of rat cecal contents for 12 h after starting the in vitro drug release experiments.

1 for 2 h and then liquid 2 for 5 h, and in the suspension of cecal contents at 12 h after starting the in vitro drug release experiments are shown in Figures 4A and 4B, respectively. Disintegration of the chitosan capsule in the suspension of rat cecal contents is evident (Figure 4B). However, no disintegration of chitosan capsules in liquid 1 and 2 was evident by scanning electron microscopy (Figure 4A). In Vivo Drug Absorption Experiments with the Chitosan CapsulessThe plasma insulin concentration and the 1018 / Journal of Pharmaceutical Sciences Vol. 86, No. 9, September 1997

plasma glucose concentration-time profiles after oral administration of eight chitosan capsules are shown in Figure 5. No peak insulin or hypoglycemic effects were observed after the oral administration of insulin in solution. In addition, no hypoglycemic effects were observed after the oral administration of gelatin capsules containing insulin. There were no insulin peaks, however a few hypoglycemic effects were observed after the oral administration of chitosan capsules containing insulin. The hypoglycemic effects started from 6

greatest hypoglycemic effects were obtained 15 h after the oral administration of chitosan capsules containing insulin and sodium glycocholate (9.8 mg). The PA% of insulin after the oral administration of chitosan capsules containing 20 IU of insulin and various protease inhibitors or absorption enhancers is shown in Figure 7. Little PA% was obtained after the oral administration of an insulin solution. However, the PA% after the oral administration of chitosan capsules containing insulin slightly increased to 1.6%. Therefore, delivery of insulin to the colon with chitosan capsules did have some effects, however these effects were not sufficient. The PA% values were increased by coadministration of various additives. The protease inhibitors, such as bacitracin, aprotinin, and STI, increased the PA% values. There was a significant increase in PA% in the presence of aprotinin. The order for absorption enhancers was, sodium glycocholate > sodium oleate. However, LM with insulin did not enhance the PA%. The PA% after the oral administration of chitosan capsules containing insulin (20 IU) with LM (10 mg) was the same as the PA% after the oral administration of chitosan capsules containing only insulin (20 IU). The PA% was significantly increased by the administration of chitosan capsules containing insulin (20 IU) and sodium glycocholate (9.8 mg).

Discussion
In this study, we showed that chitosan capsules reached the large intestine 6 to 12 h after oral administration. This finding was almost consistent with the review by Mrsny10 who reported that transit through the proximal colon (cecum to mid-transverse colon), right colon (mid-transverse to sigmoid colon), and sigmoid colon requires 7-11 h, 9-11 h, and 12.5-18.5 h, respectively. In addition, we observed a decrement in plasma glucose levels 6 h after oral administration of chitosan capsules containing insulin and sodium glycocholate. Therefore, the onset of hypoglycemic effects in absorption studies was fairly well correlated with the results of gastrointestinal transit time of chitosan capsules to the colon. In the release studies, we measured the amount of CF released from the chitosan capsules in rat cecal contents. CF is a stable compound in both buffer solution and suspension of rat cecal contents and is easily assayed with a spectrofluorometer. Moreover, we previously reported that insulin was degraded by some proteolytic enzymes in rat cecal contents and the amount of insulin in rat cecal contents could not be measured exactly.8,9 Thus, we used this dye as a model compound. We found that the release of CF was markedly increased in the presence of rat cecal contents, suggesting that the chitosan capsules were degraded in the presence of rat cecal contents. The mechanism by which chitosan capsules were degraded in rat cecal contents is not fully understood. One possible mechanism is that chitosan may be degraded by Bacteroides Vulgatus, which is major component of cecal contents. Indeed, in our pilot studies, the swelling degree of chitosan film was increased in the presence of Bacteroides Vulgatus solution and its elongation strength was decreased in this solution. Thus, it may be possible that chitosan is degraded by some enzymes that are released from this bacteria. Alternatively, it was recently reported that the pH may actually decrease in the colon as compared with the pH of the small intestine.11 This low pH, which is caused by acidification of the colonic contents by the products of bacterial fermentation, may be related to the degradation of chitosan capsules in rat cecal contents because chitosan is unstable and is easily degraded in acidic condition. We found that the hypoglycemic effect was observed after oral administration of chitosan capsules containing insulin

Figure 5sPlasma insulin and glucose concentrations after the oral administration of chitosan capsules. Key: (4) solution (insulin, 20 IU); (0) gelatin capsules (insulin, 20 IU); (O) chitosan capsules (insulin, 20 IU); (2) chitosan capsules (insulin, 20 IU; sodium glycocholate, 9.8 mg). Results are expressed as the mean SE of five experiments.

h after the administration of the chitosan capsules, when the capsules containing insulin entered the colon. Conversely, sharp peaks in the plasma insulin concentration were observed after the oral administration of chitosan capsules containing insulin and sodium glycocholate. The plasma insulin concentration increased from 6 h after administration, a maximum plasma insulin concentration of 326 U/mL was observed 7 h after administration, and the bioavailability of insulin was 5.73%. Hypoglycemic effects accompanied the absorption of insulin. The hypoglycemic effects were observed between 6 and 24 h after the oral administration of the chitosan capsules containing insulin and sodium glycocholate. Hypoglycemic Effects and Pharmacological Availability (PA%) after the Oral Administration of Chitosan Capsules Containing Insulin and Various AdditivessThe hypoglycemic effects after the oral administration of chitosan capsules containing insulin and various protease inhibitors are shown in Figure 6A. The dose of each protease inhibitor was determined from our previous results of in situ insulin absorption experiments. We reported that the presence of these protease inhibitors at each dose enhanced the absorption of insulin from the large intestine by an in situ loop method.4 The hypoglycemic effects were increased by the coadministration of various protease inhibitors such as bacitracin (14 mg), aprotinin (4.0 mg), and STI (3.0 mg) compared with the administration of insulin alone. The hypoglycemic effects after the oral administration of chitosan capsules containing insulin and various absorption enhancers are shown in Figure 6B. The hypoglycemic effects were increased by coadministration of sodium glycocholate (9.8 mg) and sodium oleate (6.0 mg), however no absorption enhancing effect was obtained when LM (10 mg) was coadministered with insulin. The

Journal of Pharmaceutical Sciences / 1019 Vol. 86, No. 9, September 1997

Figure 6sConcentrationtime profile of glucose in the plasma after the oral administration of chitosan capsules containing insulin with (A) various protease inhibitors and (B) various absorption enhancers. Key: (A) (O) insulin, 20 IU; (b) insulin, 20 IU + bacitracin, 14 mg; (9) insulin, 20 IU + soybean trypsin inhibitor, 3.0 mg; (4) insulin, 20 IU + aprotinin, 4.0 mg, (B) (O) insulin, 20 IU; (b) insulin, 20 IU + sodium oleate, 6.1 mg; (9) insulin, 20 IU + n-dodecyl--D-maltopyranoside, 10 mg; (2) insulin, 20 IU + sodium glycocholate, 9.8 mg. Results are expressed as the mean SE of five experiments.

Figure 7sPharmacological availability (PA%) of insulin administered orally with various additives using chitosan capsules. Key: (a) sodium glycocholate; (b) sodium oleate; (c) n-dodecyl--D-maltopyranoside; (d) soybean trypsin inhibitor; (*) p < 0.05, compared with the administration of chitosan capsules (insulin, 20 IU). The results are expressed as the mean SE of five experiments.

(20 IU), although no significant hypoglycemic effect was noted after large intestinal administration of insulin solution (20 IU). This result may be attributed to the improved stability of insulin in the gastrointestinal tract within chitosan capsules, because insulin is degraded by various proteolytic enzymes in the gut4,12 and the chitosan capsules may protect insulin from this degradation. Moreover, it has recently been reported by Illum et al.13 that chitosan considerably enhanced the absorption of peptides, such as insulin and calcitonin, across the nasal epithelium. More recently, Artursson et al.14,15 reported that chitosan increased the permeability of mannitol across Caco-2 monolayers in a concentration- and pH-dependent way. These findings may be involved in the enhanced mechanisms of insulin using chitosan capsules in the colon. The hypoglycemic effects were further increased by the coadministration of insulin with a variety of protease inhibitors and absorption enhancers. The greatest hypoglycemic effect was observed when sodium glycocholate was coadministered with insulin. These results reflect the results of in 1020 / Journal of Pharmaceutical Sciences Vol. 86, No. 9, September 1997

situ closed loop experiments.4 We already reported that sodium glycocholate had no effect on the stability of insulin in large intestinal mucosal homogenates or suspension of rat cecal contents.4,8 Therefore, we consider that the absorptionenhancing action of sodium glycocholate was more predominant for improving insulin absorption from the colon than its protease inhibitory activities in this study.16 Bacitracin and STI are known to be leucine aminopeptidase and trypsin and chymotrypsin inhibitors, respectively.17-19 In this study, the hypoglycemic effects observed after oral administration of insulin were increased by these protease inhibitors. Therefore, these inhibitors may reduce the activities of various proteases, thereby inhibiting the degradation of insulin in the large intestinal fluid and large intestinal mucosae. Moreover, we recently found that bacitracin increased both small and large intestinal absorption of phenol red and fluorescein isothiocyanate dextran, with an average molecular weight of 4000 (FD-4),20 using in situ closed loop experiments and in vitro Ussing chamber experiments. Thus, we demonstrated that bacitracin has not only a proteolytic inhibitory action but also an absorption-enhancing capability, and the hypoglycemic effects of bacitracin reflect both of these actions. The alkylsaccharide LM is known to have absorptionenhancing activities. Murakami et al.21 reported that insulin absorption was increased by the coadministration of LM in the rectum. However, no significant increase in the PA% was observed when LM was coadministered with insulin. We consider that this negative result of LM can be explained by the low solubility of LM and the different release rates between insulin and LM from the capsules. Oleic acid, an unsaturated fatty acid, is also known to be a typical absorption enhancer.22 Therefore, the positive effect of oleic acid may be attributed to its absorption-enhancing action. In conclusion, we demonstrated that chitosan capsules are stable in the stomach and the small intestine. However, they were specifically degraded by microorganisms in the cecal contents of the rats when they reached the colon. Furthermore, we showed that insulin absorption from the large

intestine was improved by the coadministration of a variety of additives. Thus, these capsules may be useful carriers for the colon-specific delivery of peptides including insulin.

References and Notes

1. Lee, V. H. L.; Yamamoto, A. Adv. Drug Del. Rev. 1990, 4, 171207. 2. Hirai, S.; Yashiki, T.; Mima, H. Int. J. Pharm. 1981, 9, 165172. 3. Tenma, T.; Yodoya, E.; Tashima, S.; Fujita, T.; Murakami, M.; Yamamoto, A.; Muranishi, S. Pharm. Res. 1993, 10, 1488-1492. 4. Yamamoto, A.; Taniguchi, T.; Rikyuu, K.; Tsuji, T.; Fujita, T.; Murakami, M.; Muranishi, S. Pharm. Res. 1994, 11, 1496-1500. 5. Ashford, M.; Fell, J. T.; Attwood, D.; Woodhead, P. J. Int. J. Pharm. 1993, 91, 241-245. 6. Saffran, M.; Kumar, G. S.; Savariar, C.; Burnham, J. C.; Williams, F.; Neckers, D. C. Science 1986, 233, 1081-1084. 7. Kristl, J.; Smid-Korbar, J.; Struc, E.; Schara, M.; Rupprecht, H. Int. J. Pharm. 1993, 99, 13-19. 8. Tozaki, H.; Emi, Y.; Horisaka, E.; Fujita, T.; Yamamoto A.; Muranishi, S. Biol. Pharm. Bull. 1995, 18, 929-931. 9. Tozaki, H.; Emi, Y.; Horisaka, E.; Fujita, T.; Yamamoto, A.; Muranishi, S. J. Pharm. Pharmacol. 1997, 49, 164-168. 10. Mrsny, R. J. J. Controlled Release 1992, 22, 15-34. 11. Pye, G.; Evans, D. F.; Ledingham, S.; Hardcastle, J. D. Gut 1990, 31, 1355-1357. 12. Tozaki, H.; Taniguchi, T.; Yamamoto, A.; Muranishi, S. Pharm. Sci. 1996, 2, 365-368. 13. Illum, L.; Farraj, N. F.; Davis, S. S. Pharm. Res. 1994, 11, 11861189.

14. Artursson, P.; Lindmark, T.; Davis, S. S.; Illum, L. Pharm. Res. 1994, 9, 1358-1361. 15. Schipper, N. G. M.; Varum, K. M.; Artursson, P. Pharm. Res. 1996, 11, 1686-1692. 16. Hersey, S. J.; Jackson, R. T. J. Pharm. Sci. 1987, 76, 876-879. 17. Roth, R. A. Biochem. Biophys. Res. Commun. 1981, 98, 431438. 18. Raehs, S. C.; Sandow, J.; Wirth, K.; Merkle, H. P. Pharm. Res. 1988, 5, 689-693. 19. Morimoto, K.; Yamaguchi, H.; Iwakura, Y.; Miyazaki, M.; Nakatani, E.; Iwamoto, T.; Ohashi, Y.; Nakai, Y. Pharm. Res. 1991, 8, 1175-1179. 20. Gotoh, S.; Nakamura, R.; Nishiyama, M.; Quan, Y. S.; Fujita, T.; Yamamoto, A.; Muranishi, S. J. Pharm. Sci. 1996, 8, 858862. 21. Murakami, M.; Kusanoi, Y.; Takada, K.; Muranishi, S. Int. J. Pharm. 1992, 79, 159-169. 22. Fukui, H.; Murakami, M.; Yoshikawa, H.; Takada, K.; Muranishi, S. J. Pharmacobio-Dyn. 1987, 10, 236-242.

Acknowledgments
The authors thank T. Matsumoto, Y. Hagino, and T. Sugiyama (Research and Development, Aicello Chemical Company, Ltd.) for their careful preparation of chitosan capsules. We also thank M. Murakami and T. Fujita for many helpful discussions during the course of the work, and T. Taniguchi, Y. Emi, E. Horisaka, T. Odoriba, K. Watanabe, and Y. Tada for their technical assistance.

JS970018G

Journal of Pharmaceutical Sciences / 1021 Vol. 86, No. 9, September 1997