Division of Biopharmaceutics and Pharmacokinetics Department of Pharmacy

University of Helsinki

Biopharmaceutical Evaluation of Microcrystalline Chitosan as Release-Rate-Controlling Hydrophilic Polymer in Granules for Gastro-Retentive Drug Delivery

Mia Säkkinen

Academic Dissertation

To be presented, with the permission of the Faculty of Science of the University of Helsinki,

for public criticism in Auditorium 1041 at Viikki Biocentre (Viikinkaari 5) on October 18th, 2003, at 12 noon

Helsinki 2003

Supervisor Professor Martti Marvola Division of Biopharmaceutics and Pharmacokinetics Department of Pharmacy University of Helsinki Finland

Reviewers Docent Jyrki Heinämäki Pharmaceutical Technology Division Department of Pharmacy University of Helsinki Finland

Docent Mika Vidgren Department of Pharmaceutics

Faculty of Pharmacy University of Kuopio Finland

Opponent Professor Jukka Mönkkönen Department of Pharmaceutics

Faculty of Pharmacy University of Kuopio Finland

© Mia Säkkinen 2003 ISBN 952-10-1049-5 (print) ISBN 952-10-1050-9 (pdf, http://ethesis.helsinki.fi/) ISSN 1239-9469

Yliopistopaino Helsinki 2003 Finland

i

Contents

Table of contents i

Abstract iv

List of original publications v

1. Introduction 1

1.1. Chitosan 3

1.1.1. Origin of chitosan 3

1.1.2. Physicochemical properties of chitosan 3

1.1.3. Chitosan in pharmaceutical applications 5

1.2. In vitro characterization of specific effects of chitosan 7

1.2.1. Effects of gel-forming chitosan on drug release 7

1.2.2. Mucoadhesive effects of chitosan 9

1.3. Mucoadhesive chitosan in slow-release formulations for gastro-retentive drug delivery 11

1.3.1. Special properties of chitosan 11

1.3.2. In vivo behaviour of chitosan formulations; unanswered questions 12

1.4. Safety of chitosan in oral use 15

2. Study strategy 16

2.1. Choice of microcrystalline chitosan and study variables 16

ii

2.2. Choice of study methods 17

2.2.1. In vitro studies 17

2.2.2. In vivo studies 18

3. Aims of study 21

4. Materials and methods 22

4.1. Chitosans 22

4.2. Model drugs 23

4.3. Study formulations 23

4.4. In vitro studies 24

4.4.1. Properties of chitosan gels (I) 24

4.4.2. Gel formation by chitosan in granules (I, IV) 24

4.4.3. Drug release from chitosan granules (I–IV) 24

4.4.4. In vitro mucoadhesion of chitosan (III) 25

4.5. In vivo studies 25

4.5.1. Bioavailability studies (II–III) 25

4.5.2. Gamma scintigraphic investigations (IV–V) 27

5. Results and discussion 29

5.1. In vitro characteristics of MCCh formulations 29

5.1.1. Gel-forming ability of MCCh 30

5.1.2. Effects of grade of MCCh on release and mucoadhesive characteristics of chitosan formulations 30

5.1.3. Effects of drug solubility and amount of MCCh on release characteristics of chitosan formulations 32

iii

5.2. Release characteristics of MCCh formulations in vivo 34

5.2.1. Bioavailabilities of drugs from MCCh granules; effects of amount and molecular weight of chitosan 34

5.2.2. Effects of other excipients in MCCh granules 36

5.3. Retention of MCCh formulations in the stomach 37

5.3.1. Information obtained via bioavailability studies 37

5.3.2. Gamma scintigraphic investigations 38

5.4. Additional information provided by gamma scintigraphy on the behaviour of MCCh formulations in vivo 41

6. Conclusions 43

Acknowledgements 45

References 47

Original publications I–V

iv

Abstract

Chitosan, a polymer derived from crustacean chitin, has a number of properties that make it potentially valuable as a pharmaceutical excipient. In the studies described, microcrystalline chitosan (MCCh), a highly crystalline grade of chitosan base, was evaluated for the first time as an excipient. The first objective was to determine its properties in vitro as a gel-forming and drug release rate controlling excipient in granules, and as a mucoadhesive polymer. The properties of MCCh were compared with those of conventional chitosan, which is predominantly amorphous. Attention was also given to the effects of the molecular weight (Mw) and degree of deacetylation (DD) of MCCh on its properties. The ultimate objective was to evaluate, in studies on human volunteers, whether MCCh granules could be useful as slow-release systems for gastro-retentive drug delivery.

Drug release in vitro could be controlled by altering the amount and grade of chitosan in the granules. Increasing the chitosan content decreased the release rate after the formation of chitosan gels in an acidic environment. MCCh had a more marked retardant effect than conventional chitosan, because of its superior gel formation. Retardation of drug release increased with the Mw of MCCh but DD had no marked effect. A slow release was achieved with model drugs that are slightly soluble in acidic environments (ibuprofen, furosemide). With a readily soluble drug (paracetamol) retardation of drug release was only slight. All the chitosans studied had fairly marked mucoadhesive tendencies in vitro (porcine oesophagus preparations). Adhesive strength increased with the Mw of chitosan, while other variables (DD and crystallinity) had no marked effects.

Bioavailability studies showed that use of MCCh granules in gelatine capsules is a simple way of preparing slow-release dosage forms of slightly soluble drugs, provided the drug concerned, like ibuprofen, is absorbed throughout the gastrointestinal tract (gi-tract). With furosemide, a drug that is specifically absorbed in the upper gi-tract, the amounts absorbed decreased with release rate in vitro. This indicates that the drug was being released beyond its “absorption window”. Gamma scintigraphic investigations revealed that MCCh granules were not retained in the stomach to sufficiently reproducible extents; adhesion of MCCh onto gastric mucosa took place only occasionally. Adherence of the formulation to the oesophagus in one volunteer suggests that problems could arise from the type of formulation studied. The results of in vivo studies demonstrated that MCCh can be used to prepare slow-release systems but not gastro-retentive formulations. The results of dissolution and mucoadhesion tests in vitro did not allow adequate prediction of the in vivo behaviour of MCCh formulations. This underscores the importance of conducting biopharmaceutical studies at an early stage during the evaluation of new excipients.

v

List of original publications

This dissertation is based on the following studies, which are referred to in the text by the Roman numerals I–V.

I Säkkinen, M., Seppälä, U., Heinänen, P., Marvola, M., 2002. In vitro evaluation of microcrystalline chitosan (MCCh) as gel-forming excipient in matrix granules. Eur. J. Pharm. Biopharm. 54, 33-40.

II Säkkinen, M., Linna, A., Ojala, S., Jürjenson, H., Veski, P., Marvola, M., 2003.

In vivo evaluation of matrix granules containing microcrystalline chitosan as a gel-forming excipient. Int. J. Pharm. 250, 227-237.

III Säkkinen, M., Tuononen, T., Jürjenson, H., Veski, P., Marvola, M., 2003.

Evaluation of microcrystalline chitosans for gastro-retentive drug delivery. Eur. J. Pharm. Sci. 19, 345-353.

IV Säkkinen, M., Marvola, J., Kanerva, H., Lindevall, K., Lipponen, M., Kekki, T.,

Ahonen, A., Marvola, M., 2003. Gamma scintigraphic evaluation of the fate of microcrystalline chitosan granules in human stomach. Eur. J. Pharm. Biopharm. In press.

V Säkkinen, M., Marvola, J., Kanerva, H., Lindevall, K., Ahonen, A., Marvola, M.,

2003. Scintigraphic verification of adherence of a chitosan formulation to the human oesophagus. Eur. J. Pharm. Biopharm. In press.

1

1. Introduction

Chitosan is a cationic polymer derived from the chitin of crustaceans. Its use in various applications has received considerable attention. One obvious advantage of this material is that it can be obtained from ecologically sound natural sources, namely crab- and shrimp-shell wastes. Chitosan has been widely studied in the biomedical field, and has been found to be highly biocompatible (Muzzarelli et al., 1988). To mention only some areas of investigation, chitosan has been studied as a wound dressing (Loke et al., 2000; Mi et al., 2001), as a haemostatic material (Rao and Sharma, 1997; Hoekstra et al., 1998) in the form of gels or sponges, and as sutures, in the form of fibres (Qin and Agboh, 1998). Some chitosan-based products, e.g. wound dressings (Illum, 1998), have reached the market. Chitosan has also been marketed throughout the world in non-medical products, as a fat-binder in cholesterol-lowering and slimming formulations (Shahidi et al., 1999). It has been claimed that chitosan entraps lipids in the intestine, because of its cationic nature (Kanauchi et al., 1995; Wuolijoki et al., 1999). Chitosan may also have technical applications, including, e.g., use as a seed coating and nitrogen source in agriculture, and as an adsorbent for water-purification (Li et al., 1992).

In recent decades there has also been considerable interest in the pharmaceutical field in using chitosan as an excipient, in various applications. In addition to the good biocompatibility of chitosan and the abundance of natural sources of the material, chitosan has a number of desirable properties that make study of it interesting. Because it is a polycationic polymer, chitosan forms gels most readily in acidic environments, such as that in the stomach. This makes chitosan interesting in relation to the development of slow-release dosage forms for oral administration. The mucoadhesive properties of chitosan, illustrated by its ability to adhere for instance to porcine gastric mucosa in vitro (Gåserød et al., 1998), could allow site-specific drug delivery. These desirable properties of chitosan have recently led to increasing interest in the development of slow-release formulations for gastro-retentive drug delivery. Chitosan-based dosage forms of this kind could be useful in relation, e.g., to the administration of antibiotics used for eradication of Helicobacter pylori from the stomach (Shah et al., 1999; Remuñan-López et al., 2000).

Chitosan is produced in grades that differ in their physicochemical properties, and as a base or a salt of a base. The existence of different grades could also be valuable. It seems that properties of chitosan-based dosage forms could be controlled by altering the grade of chitosan in formulations. The effects of chitosan on drug release rate (Goskonda and Upadrashta, 1993; Mi et al., 1997; Sabnis et al., 1997) and the capacity of chitosan for mucoadhesion (Lehr et al., 1992; He et al., 1998) depend on properties such as molecular

2

weight and degree of deacetylation of chitosan. Chitosan base seems likely to be more useful than chitosan salts in relation to development of slow-release formulations because it is in general less soluble.

Microcrystalline chitosan (MCCh) may be particularly valuable as an excipient. MCCh is a highly crystalline grade of chitosan base (Struszczyk, 1987), which can be prepared on a large scale using a method developed by the Finnish company Novasso (Finnish Pat. FI83426, 1991). MCCh has been studied in relation to various technical applications (Struszczyk and Kivekäs, 1992) and in cholesterol-lowering formulations (Wuolijoki et al., 1999) but not as a pharmaceutical excipient before the studies described here were carried out. Most drug formulation studies involving chitosans have used material produced commercially by conventional methods. Chitosans of this kind are fairly amorphous. However, Sabnis et al. (1997) reported that increasing the crystallinity of chitosan could offer advantages in relation to manufacturing process of pharmaceutical formulations, e.g. by making the chitosan more suitable for direct compression into tablets. Perhaps more significantly, the crystallinity of a chitosan could affect the behaviour of a formulation in which it was incorporated. Effects of the crystallinity of chitosan therefore required evaluation. One specific property of MCCh is its high capacity for retaining water (Struszczyk, 1987). This property could be advantageous in relation to the development of slow-release formulations because it might facilitate the formation of gels that would control drug release. The pronounced ability of MCCh to form hydrogen bonds (Struszczyk, 1987; Struszczyk and Kivekäs, 1992) could theoretically result in efficient mucoadhesion by MCCh. The properties of MCCh mentioned made it particularly interesting for study as a hydrophilic excipient controlling rate of drug release from formulations that were also intended to be mucoadhesive in the stomach.

It is obvious that no conclusions regarding the potential of MCCh, or any other grade of chitosan, as an excipient in drug delivery systems can be reached until biopharmaceutical studies have been conducted. A major defect in many previous studies is that only in vitro methods have been employed in determining the properties of chitosan-based formulations. Only in some studies in animals have attempts been made to demonstrate the value of chitosan as a gel-forming polymer that could control rate of drug release from oral dosage forms (Miyazaki et al., 1988a,b), and to evaluate whether chitosan-based formulations were gastro-retentive (Remuñan-López et al., 2000). It is, however, well known that results of in vitro studies, and even results of animal studies, often fail to predict fate of a formulation in man (Davis and Wilding, 2000). Biopharmaceutical studies in human volunteers are therefore important. Until the studies described here were carried out, no information was available about the adhesion of chitosan to the mucosa of the human stomach or whether chitosan-based formulations had potential as slow-release systems for gastro-retentive drug delivery in man.

3

1.1. Chitosan

1.1.1. Origin of chitosan

One reason why chitosan has become of interest is undoubtedly because it can be obtained from natural sources that are abundant and renewable. Chitosan is prepared from chitin, the polymer second most abundant in nature after cellulose (Roberts, 1992). Chitin is the primary structural component of the outer skeletons of crustaceans, and of many other species such as molluscs, insects and fungi. The role played by chitin is similar to the roles played by cellulose in plants and collagen in higher animals. It is a reinforcing material, which occurs in three polymorphic forms, α-, β- and γ-chitin. Where hardness in needed α-chitin is found, where flexibility is required β- and γ-chitin occur. Chitin is inert in aqueous environment. This property limits the use of chitin as such. Chitosan is prepared from chitin to obtain a more reactive polymer.

Chitosan is most commonly obtained from crustacean chitin, from crab- and shrimp-shell wastes (Roberts, 1992; Shepherd et al., 1997). Chitin accounts for approximately 70% of the organic compounds in such shells. In preparing chitosan, ground shells are treated with alkali and acid to remove proteins and minerals, respectively, after which the extracted chitin is deacetylated to chitosan by alkaline hydrolysis at high temperature. Preparation of chitosan from crustacean-shell wastes is economically feasible and ecologically desirable because large amounts of shell wastes are available as a by-product of the food industry. Production of chitosan from these is inexpensive and easy. It has also been suggested that other sources of chitin, e.g. chitin from squid pens, may be valuable in relation to the preparation of chitosan (Shepherd et al., 1997; Rhazi et al., 2000). Chitosan as such is rare in nature, except in certain fungi. In recent years, there has been increasing interest in the production of chitosan from fungi, using fermentation methods (Nwe et al., 2002).

1.1.2. Physicochemical properties of chitosan

Structure. Chitosan is a linear polysaccharide consisting of ß (1-4)-linked 2-amino-2-deoxy-D-glucose (D-glucosamine) and 2-acetamido-2-deoxy-D-glucose (N-acetyl-D-glucosamine) units (Fig. 1). The structure of chitosan is very similar to that of cellulose (made up of ß (1-4)-linked D-glucose units), in which there are hydroxyl groups at C2 positions of the glucose rings.

The term chitosan is used to describe a series of polymers of different degrees of deacetylation (DD), defined in terms of the percentage of primary amino groups in the polymer backbone, and average molecular weights (Mw) (Roberts, 1992). The DD of chitosan is usually between 70 and 95%, and the Mw between 10 and 1000 kDa. Changing

4

the reaction conditions during the manufacture of chitosan from chitin can alter the DD and Mw of chitosan. The chitosan polymer produced using conventional commercial manufacturing methods is fairly amorphous (Struszczyk, 1987; Genta et al., 1995; Portero et al., 1998; Mura et al., 2003). The crystallinity of the polymer can be increased using the method described by Struszczyk (1987), which results in microcrystalline chitosan (MCCh). The method is based on precipitation of MCCh from acidic solutions of conventional chitosan. The crystallinity of MCCh manufactured using this method is up to 30% greater than that of conventionally prepared chitosan, the crystallinity index of which is typically less than 60% (Struszczyk, 1987). The Finnish company Novasso has developed a continuous method (Finnish Pat. FI83426, 1991) for production of MCCh on a large scale. The DD of the MCCh produced typically ranges from 70 to 95%, and the Mw from 10 kDa to 300 kDa. MCCh is available as a base. Conventional chitosan is available both as a base and as a salt, e.g. the hydrochloride, malate or glutamate.

Figure 1. Structural units of chitosan and its parent substance chitin. Chitin consists mostly of N-acetyl-D-glucosamine –units (left). During the preparation of chitosan most units are deacetylated to D-glucosamine –units (right).

Behaviour in aqueous media. Chitosan, which contains a number of amino groups in its polymer backbone, is a polycationic base in acidic aqueous environments. A cationic character is rare in polysaccharides found in nature, which makes chitosan an interesting material for study. Several investigators (Cölfen et al., 2001; Berth and Dautzenberg, 2002) are currently characterizing chitosan in aqueous media in depth, from a macro-molecular point of view.

The D-glucosamine unit has a pKa value of 7.5 (Roberts, 1992). The basic nature of chitosan depends on its degree of deacetylation, and the pKa value for the polymer is around 6.5. Chitosan base dissolves slowly in acidic and slightly acidic aqueous solutions. If the pH of a solution is increased to close to neutral, the base precipitates from the solution. Low-molecular-weight chitosans (Mw approximately 10 kDa or below) and chitosan salts (e.g. hydrochlorides) may be more readily soluble. Chitosan is also hydrophilic, retaining water in its structure and forming gels spontaneously. Gel formation

5

takes place at acidic and slightly acidic pH values, because of the cationic nature of chitosan. Chitosan gels increase in viscosity as molecular weight or amount of polymer increase (Roberts, 1992; Kristl et al., 1993). A decrease in the pH increases viscosity, apparently because the conformation of chitosan is more extended (random coil) at low pH values, because of repulsive forces between positively charged amino groups (Tsaih and Chen, 1999). Viscosities also increase as the degree of deacetylation increases, because the polyelectrolytic characteristics of the chitosan become more marked. Chitosan gels exhibit pseudoplastic flow, which means that their viscosities decrease as shear force increases (Senel et al., 2000). Chitosan gels degrade gradually, as the chitosan dissolves.

1.1.3. Chitosan in pharmaceutical applications

Chitosan has received considerable attention as a possible pharmaceutical excipient in recent decades (Table 1). It has been evaluated in both conventional excipient applications, e.g. as a directly compressible diluent in tablets (Sawayanagi et al., 1982a,b), as a binder in wet granulation (Upadrashta et al., 1992), and in novel applications, e.g. as a carrier for mucosal delivery of antigens in connection with oral vaccination (Van der Lubben et al., 2001a,b; Chew et al., 2003). Chitosan has also recently been approved by the authorities, and a monograph relating to chitosan hydrochloride was included in the fourth edition of the European Pharmacopoeia (2002).

Several properties of chitosan make it potentially valuable as a pharmaceutical excipient. Good biocompatibility and low toxicity of chitosan (Arai et al., 1968; Muzzarelli et al., 1988; Knapczyk et al., 1989; Kim et al., 2001), and the fact that sources of chitosan are abundant, are properties that any new excipient material should have. One property that makes chitosan particularly interesting for study as an excipient is its ability to become hydrated and form gels in acidic aqueous environments. Because of its gel-forming ability, a major area of interest since studies began has been use of chitosan to prepare slow-release drug delivery systems. Chitosan has been evaluated in vitro as a drug carrier in hydrocolloids and gels (Knapczyk, 1993b; Kristl et al., 1993), and as a hydrophilic matrix retarding drug release in tablets (Kawashima et al., 1985; Acartürk, 1989), granules (Hou et al., 1985; Goskonda and Upadrashta, 1993) and microparticles (Thanoo et al., 1992; Chandy and Sharma, 1993; Kas, 1997). The hydrophilic nature of chitosan has also aroused interest in its use in immediate-release formulations, e.g. as a disintegrant in small amounts in tablets, where it has been found to have effects similar to or better than those of microcrystalline cellulose (Sawayanagi et al., 1982a,b; Ritthidej et al., 1994), and as an excipient to increase the rate of dissolution of poorly soluble drug substances (Shiraishi et al., 1990; Acartürk et al., 1993; Portero et al., 1998).

6

One area in which interest is growing is use of chitosan as a bioadhesive material. Many commercially available chitosans exhibit fairly good mucoadhesive properties in vitro (Lehr et al., 1992), and interest in use of chitosan to prepare mucoadhesive systems has therefore been aroused. It has been suggested that times of residence of formulations at sites of drug action or absorption could be prolonged through use of chitosan. It has also been suggested that chitosan might be valuable for delivery of drugs to specific regions of the gastrointestinal tract, e.g. the stomach (Gåserød et al., 1998; Remuñan-López et al., 2000), small intestine (Lehr et al., 1992; He et al., 1998; Shimoda et al., 2001), and buccal mucosa (Miyazaki et al., 1995; Remuñan-López et al., 1998). Delivery to other mucosal surfaces, e.g. delivery of peptide drugs on to the nasal epithelia has also been studied (Illum et al., 1994; 2002; Aspden et al., 1997; Tengamnuay et al., 2000).

Table 1. Pharmaceutical applications of chitosan investigated.

Application Reference • Diluent in direct compression of tablets Sawayanagi et al., 1982a,b; Knapczyk,

1993a • Binder in wet granulation Upadrashta et al., 1992; Tapia et al., 1993;

Henriksen et al., 1993 • Slow-release of drugs from tablets and granules Kawashima et al., 1985; Hou et al., 1985;

Acartürk, 1989 • Drug carrier in microparticle systems Thanoo et al., 1992; Chandy and Sharma,

1993; Okhamafe et al., 1996 • Films controlling drug release Remuñan-López and Bodmeier, 1997;

Senel et al., 2000 • Preparation of hydrogels, agent for increasing

viscosity in solutions Knapczyk, 1993b; Kristl et al., 1993; Khalid et al., 1999

• Wetting agent, and improvement of dissolution of poorly soluble drug substances

Shiraishi et al., 1990; Acartürk et al., 1993; Genta et al., 1995

• Disintegrant Sawayanagi et al., 1982a,b; Ritthidej et al., 1994

• Bioadhesive polymer Lehr et al., 1992; Miyazaki et al., 1995; Bernkop-Schnürch et al., 1998

• Site-specific drug delivery (e.g. to the stomach or colon)

Tozaki et al., 1997; 1999; 2002; Shah et al., 1999; Remuñan-López et al., 2000

• Absorption enhancer (e.g. for nasal or oral drug delivery)

Illum et al., 1994; Schipper et al., 1996; 1997; 1999; Kotzé et al., 1997

• Biodegradable polymer (implants, microparticles) Song et al., 1996; Jameela et al., 1998 • Carrier in relation to vaccine delivery or gene

therapy Lee et al., 1998; Aral et al., 2000; Van der Lubben et al., 2001a,b

7

The potential value of chitosan as a novel excipient which could find extensive application in pharmaceutical products has been highlighted in several reports, and in numerous review articles relating to the field (Li et al., 1992; Illum, 1998; Dodane and Vilivalam, 1998; Felt et al., 1998; Paul and Sharma, 2000; Singla and Chawla, 2001). Despite the substantial research relating to chitosan that has been carried out in recent decades, many questions remain unanswered. More information is needed about the effect of chitosan grade on the properties of pharmaceutical formulations. The fact that commercially available chitosan is not always well characterized has limited use of chitosan (Felt et al., 1998). Comparison of results obtained by different research groups has been difficult because the properties of the chitosans studied, e.g. degree of deacetylation and/or molecular weight, have not been specified. However, in recent years attention has been paid producing chitosans with particular physicochemical properties, and desired combinations of DD and Mw. The progress made has allowed studies in which the effect of altering chitosan grade on the properties of pharmaceutical formulations could be determined. Another important issue is the in vivo behaviour of chitosan-based formulations. A major deficiency of studies in this field is that many have determined the properties of the chitosan-based formulations solely by means of in vitro methods. Information on the in vivo behaviour of the formulations, especially in human beings, has been lacking. Results of in vivo studies would reveal the value or otherwise of chitosan as a pharmaceutical excipient, and allow products containing chitosan to be marketed.

1.2. In vitro characterization of specific effects of chitosan

1.2.1. Effects of gel-forming chitosan on drug release

Mechanism of action. Chitosan formulations can easily be prepared using conventional tableting or granulating methods. The simplest way of achieving slow release of drugs by means of chitosan is to employ it as a gel-forming excipient in matrix-type formulations. Retardant effects of chitosan on drug release were noticed as early as the 1980s. Chitosan was found to decrease rates of release of drugs from tablets during dissolution tests at acidic and slightly acidic pH levels (Kawashima et al., 1985; Acartürk, 1989). However, Akbuğa (1993a), who carried out studies at higher pH levels (pH 7.4), reported that chitosan exhibited no slow-release properties. This finding indicated that the effects of chitosan depend on the pH level.

Mi et al. (1997) studied the mechanism of action of chitosan in slow-release matrix tablets in detail, and confirmed earlier findings. The formation of swellable and erodible chitosan gels in tablets resulted in a release mechanism typical of hydrogel-based slow-release systems, namely non-Fickian diffusion (Ritger and Peppas, 1987b). Release

8

kinetics were almost zero-order. The results also showed that hydration of and gel formation by chitosan in formulations takes place more readily at acidic pH levels (pH 1.2) than at pH levels close to neutral (7.2). This could be explained by the cationic nature of chitosan. The retardant effect of chitosan on drug release was therefore found to be most marked in an acidic environment (pH 1.2). Gel-forming chitosan also retards drug release from matrix granules (Hou et al., 1985; Miyazaki et al., 1988a,b) and the effects are most marked in acidic environments (Goskonda and Upadrashta, 1993).

Variables affecting release. The rate at which a drug is released from a hydrophilic chitosan matrix depends on the amount of chitosan involved, and on the nature of the drug. Increasing the amount of chitosan in tablets (Kawashima et al., 1985; Acartürk, 1989; Mi et al., 1997; Kristmundsdottir et al., 1995; Inoyatov et al., 1998) and granules (Miyazaki et al., 1988a,b) decreases release rates. Slow-release has been achieved with slightly soluble drug substances (Kawashima et al., 1985; Miyazaki et al., 1988a,b; Acartürk, 1989; Tapia et al., 1993; 2002). Formulations of readily soluble drugs have, in most cases, also contained excipients other than chitosan that affect release rates (Nigalaye et al., 1990; Adusumilli and Bolton, 1991; Kristmundsdóttir et al., 1995). Akbuğa et al. (1993b) and Henriksen et al. (1993) studied in detail the effects of the physicochemical properties of drug substances in formulations containing a chitosan salt (malate). They found that drugs with low solubilities in water and/or high molecular weights were released most slowly.

The grade of chitosan has also been found to affect drug release but the effects of physicochemical properties of chitosan have not so far been systematically evaluated. The effects of altering molecular weight or degree of deacetylation have been discussed in some publications but no reports of studies on the effects of altering the crystallinity of chitosan on its ability to retard drug release are available. The effects of molecular weight were first evaluated using chitosan hydrocolloids. Drug release was found to be most effectively retarded by chitosan of high molecular weight (Kristl et al., 1993). Drug release rates from granules (Goskonda and Upadrashta, 1993) and tablets (Mi et al., 1997) were also found to decline as molecular weight increased. The effects of altering the molecular weight of chitosan were assessed over wide ranges (from approximately 70 kDa to 2000 kDa) but the possible effects of degree of deacetylation were not discussed (DD was not specified). The effect of the increase in molecular weight was considered to be based on increase in viscosity of the gels formed by chitosan.

Sabnis et al. (1997) determined the effects of altering the degree of deacetylation of chitosan (DD 74, 87 and 92%). They found that drug release rates from tablets decreased as degree of deacetylation increased, in dissolution tests at pH 1.2. It was concluded that the rate of uptake of solvent by the polymer matrix increased as the number of ionizable amino groups in the chitosan increased (i.e. as DD increased), resulting in increasing formation of a chitosan gel barrier. Sabnis et al. (1997) also suggested that release might

9

also have been retarded by the formation of complexes between positively charged chitosan and the negatively charged drug studied, diclofenac. The formulations contained depolymerized chitosans of low molecular weight (approximately 20 kDa). Further studies by the same group indicated that chitosans of low molecular weight (Mw << 100 kDa) are particularly liable to interact, because of easy accessibility to amino groups (Rege et al., 1999; Sabnis and Block, 2000).

Cross-linking treatment. Drug release from chitosan-based formulations can often be modified by cross-linking treatment of chitosan. Drug release from microparticles (Thanoo et al., 1992; Orienti et al., 1996), and various in vitro properties of chitosan hydrogels, e.g. swelling of the gels and gel viscosity (Guan et al., 1996; Yao et al., 1998; Khalid et al., 1999), can be controlled by cross-linking of chitosan. During preparation of such systems the chitosan is treated with reagents that cause formation of networks involving covalent or ionic bonds. Systems of these kinds are swellable in aqueous environment and have properties that depend on the density of cross-linking and the reagent used. One advantage of this method is that it enhances the retardant effects of chitosan (Ganza-González et al., 1999). However, the method necessitates use of reagents that can have harmful effects in vivo. Toxic effects have, e.g., been noted with glutaraldehyde, an agent commonly used to induce cross-linking (Carreño-Gómez and Duncan, 1997).

1.2.2. Mucoadhesive effects of chitosan

Mechanism of action. Many hydrophilic polymers adhere to mucosal surfaces as they attract water from the mucus gel layer adherent to the epithelial surface. This is the simplest mechanism of adhesion and has been defined as “adhesion by hydration” (Chen and Cyr, 1970). Various kinds of adhesive force, e.g. hydrogen bonding between the adherent polymer and the substrate, i.e. mucus, are involved in mucoadhesion at the molecular level (Peppas and Buri, 1985).

Mucoadhesion of chitosan was first studied by Lehr et al. (1992), who reported that many commercially available chitosans adhere fairly strongly in vitro. In the studies of Lehr et al., forces required to detach chitosan films were measured in isolated porcine intestinal preparations. One important finding during the studies was that the adhesive properties of chitosans persisted well during repeated contacts of chitosan and the substrate, with the chitosan in a swollen state. It was suggested that not only adhesion by hydration was involved but also additional mechanisms, such as hydrogen bonding and ionic interactions. Important mechanism of action was suggested to be ionic interactions between positively charged amino groups in chitosan and the negatively charged mucus gel layer. In further studies, interactions between chitosan and mucus were demonstrated (Fiebrig et al., 1995; He et al., 1998), and the primary mechanism of action at the

10

molecular level was found to be electrostatic (Deacon et al., 2000). It has also been shown that adsorption of hydrophilic chitosan on mucosal surfaces and mucosal dehydration are involved in the adhesion of chitosan (Patel et al., 1999; 2000).

Variables affecting mucoadhesion. Interactions of chitosan with mucus and its mucoadhesive properties are affected by both physiological factors and the properties of chitosan. The principal component of mucus is mucous glycoprotein, or mucin, which contains negatively charged residues, e.g. sialic acid in gastric mucin (Allen, 1978). Chitosan adsorbs mucin. The extent of adsorption increases with the amount of sialic acid in the mucin (He et al., 1998). Deacon et al. (1999) showed that interactions are more marked with mucins from the cardiac region of the porcine stomach than from the corpus or antrum. This finding could be explained by the fact that amounts of sialic acid in secretions from the cardiac region are high (Nordman et al., 1997). Because amounts of sialic acid in mucosal secretions vary, the strength with which chitosan adheres to mucus may differ between tissues. Since the mechanism of adhesion also involves electrostatic interactions pH is also important. Interactions are strong at acidic and slightly acidic pH levels, at which the charge density of chitosan is high (He et al., 1998). Accordingly, microspheres coated with chitosan adhered to isolated porcine gastric mucosa in simulated gastric fluid (pH ~1.2), but adhesion to oesophagus moistened with artificial saliva (pH ~7) was less strong (Gåserød et al., 1998).

The molecular weight and charge of chitosan are important properties in relation to its capacity for mucoadhesion. Increase in molecular weight of chitosan results in stronger adhesion (Lehr et al., 1992). This would be expected, because penetration of polymer chains into the mucus layer should become more efficient as the chain length of the polymer increases (Peppas and Buri, 1985). He et al. (1998) reported that an increase in the charge density of chitosan results in adhesive properties becoming more marked. The charge density of a chitosan in a formulation is affected by its degree of deacetylation and any cross-linking treatment. Using rat-gut-loop preparations, He’s research group showed that amounts of chitosan microspheres adhering to the intestine were greatest when the density of cross-linking of chitosan was least, i.e. when the number of free amino groups in chitosan was greatest. This finding suggests that the adhesive properties of chitosan should become more marked as degree of deacetylation increases. It is also obvious from the results that cross-linking procedures should be avoided when formulations for mucoadhesive drug delivery are being developed. Results of other studies have also indicated that cross-linking reduces mucoadhesive effects of chitosan (Bernkop-Schnürch et al., 1998; Genta et al., 1999).

Like other hydrophilic mucoadhesive polymers, chitosan adheres most readily when the degree of hydration is optimal. In the study of Gåserød et al. (1998) adhesive tendencies of chitosan formulations decreased markedly when they were hydrated in

11

simulated gastric fluid with mucin for longer than 15 minutes before tissue contact (isolated porcine gastric mucosa).

Novel effects of mucoadhesives. There is currently a search for mucoadhesive polymers that are multifunctional, e.g. polymers that can also temporarily increase epithelial permeability (Lehr, 1996; Lee et al., 2000). Such polymers are intended for particular purposes, e.g. enhancement of the absorption of peptide drugs. The low solubilities of chitosan bases and also many chitosan salts make them unsuitable for use in such applications but soluble chitosan derivatives (e.g. N-trimethylchitosan chloride) have been found to show promise as absorption enhancers. They dissolve readily over a wide pH range and then increase epithelial permeability in vitro by opening the intercellular junctions of epithelial cells (Kotzé et al., 1997; 1998; 1999; Thanou et al., 1999). Studies characterizing the in vitro mucoadhesive properties of such chitosan derivatives are in progress (Snyman et al., 2003).

1.3. Mucoadhesive chitosan in slow-release formulations for gastro-retentive drug delivery

1.3.1. Special properties of chitosan

Even in early publications it was suggested that chitosan could be of particular value for controlling drug release from oral dosage forms (Hou et al., 1985). More recently, it has been suggested that chitosan formulations could also be used for gastro-retentive drug delivery (Shah et al., 1999; Remuñan-López et al., 2000; Hejazi and Amiji, 2002). Two properties in particular make chitosan potentially valuable in such dosage forms, namely its ability to form gels that control drug release at acidic pH levels, and its mucoadhesive properties. Both properties are closely related to the cationic nature of chitosan.

Chitosan could be ideal for use in formulations intended to release drugs slowly in the stomach, since the gel formation by cationic chitosan that is pronounced at acidic pH levels results in marked retardant effects on drug release (Section 1.2.1.). Orally administered formulations are initially exposed to the acidic milieu of the stomach, especially if they have been administered to subjects in fasted states, in whom gastric pH is likely to range from approximately 1 to 2. Polymers commonly used in preparing slow-release formulations do not form gels very efficiently in such environments. For example, with hydroxypropylmethylcellulose (HPMC), which is non-ionic, stable gels are formed within the pH range 3–11 (Alderman, 1984). Anionic polymers may even require pH values to be close to neutral before they will form gels. Gel formation by chitosan, in contrast, takes place readily at acidic pH levels, because of its cationic character.

12

Mucoadhesive effects of chitosan (Section 1.2.2.) are interesting. Hydrated chitosan has a positive charge and can adhere to negatively charged mucus gel layers, as results of studies e.g. on isolated porcine gastric mucosa preparations have shown (Gåserød et al., 1998). This ability could result in formulations containing chitosan being retained in the stomach. Adhesion would be expected to be particularly marked under the acidic conditions in the stomach, where cationic chitosan would be highly charged. Other mucoadhesive polymers are most commonly anionic (e.g. acrylic acid co-polymers) or non-ionic (e.g. cellulose derivatives) (Lee et al., 2000). With such polymers adhesion could take place via hydrogen bonding between the polymer and substrate. Although this mechanism could also be involved in relation to chitosan, the positive charge of the latter could result in there being additional forces of attraction (Lehr et al., 1992).

It is rare for polymers used as pharmaceutical excipients to have cationic nature. The effects of chitosan are therefore quite unique. It is, consequently, not surprising that there has been increasing interest in use of chitosan to prepare slow-release formulations for gastro-retentive drug delivery. It has been suggested that such dosage forms could be valuable, e.g., in relation to administration of antibiotics intended to eradicate Helicobacter pylori (Shah et al., 1999; Remuñan-López et al., 2000). They could also be useful for administration of drug substances that are site-specifically absorbed from the upper gastrointestinal tract or degraded in the intestine. A slow-release formulation that was retained for a time in the stomach could increase the bioavailability of such substances. There are already patents relating to various kinds of chitosan systems intended to release drugs slowly in the stomach, e.g. microspheres (Japanese Pat. JP1224311, 1989; Japanese Pat. JP5017371, 1993), and cross-linked gels (United States Pat. US5620706, 1997; United States Pat. US5904927, 1999). Most such patents are based on results of in vitro studies. However, de la Mata et al. (European Pat. Application EP1238663 A2, 2002) have shown that floating chitosan gel mixtures are being retained in the stomachs of pigs and are therefore potentially valuable as gastro-retentive drug delivery systems.

1.3.2. In vivo behaviour of chitosan formulations; unanswered questions

So far, surprisingly little information has been published on the in vivo behaviour of chitosan formulations, especially in human beings. Information is needed on both the slow-release characteristics of chitosan formulations and their mucoadhesive properties.

Evidence relating to slow release in vivo. Attempts have been made to confirm the value of chitosan systems in animals. Miyazaki et al. (1988a,b) were among the first to carry out experiments in vivo on slow-release characteristics of chitosan formulations. Using a small-scale extrusion method, they prepared chitosan matrix granules, release of drug from which was thought to be controlled by formation of chitosan gels. Release of

13

indomethacin from the granules in vitro was slow, and the granules also seemed to behave as slow-release formulations in rabbits (Miyazaki et al., 1988a) and beagle dogs (Miyazaki et al., 1988b). However, there were only three animals in each study, and the results obtained varied considerably. Miyazaki et al. were unable to demonstrate statistically significant differences between rates of absorption (evaluated in terms of tmax) from a conventional capsule formulation and from chitosan granules. Inoye et al. (1988), who tested chitosan tablets containing prednisolone in three beagle dogs, obtained similar results. Larger groups would have been needed to determine whether the formulations had potential as slow-release dosage forms. The information obtained relating to slow-release properties of chitosan formulations in vivo was limited because only a few variables were studied. Miyazaki et al. (1988a,b) reported that high amounts of chitosan in granules retarded drug absorption but did not discuss effects of altering chitosan grade (neither the Mw or DD of the chitosan used was stated). In the study of Inoye et al. (1988) tablets contained chitosans with different degrees of deacetylation but a possible effect of molecular weight was not taken into account (DD was 93% and 50–60%, Mw was not stated). The tablets contained also different amounts of other excipients, making comparisons between the effects of the chitosans studied difficult. Bioavailability studies by Shiraishi et al. (1993) in beagle dogs and by Imai et al. (2000) in human volunteers have shown that slow-release formulations of indomethacin can be prepared using cross-linked chitosan in granules.

Evidence relating to mucoadhesion in vivo. Adhesion of chitosan formulations to the mucosa of the gastrointestinal tract has been assessed in only a few studies. Miyazaki et al. (1988a) may have been the first to discover evidence that gastric residence times of formulations could be prolonged through use of chitosan. They found that chitosan granules were retained in the rabbit stomach for at least three hours (results recorded for only one rabbit). They suggested that retention in the stomach could result from flotation of the granules on gastric contents, but did not raise the possibility that chitosan might have mucoadhesive properties. Following further studies, it has been suggested that adherence of chitosan to the mucosa may be the important mechanism in relation to prolonged residence times of chitosan-based formulations in regions of the gastrointestinal tract.

The mucoadhesive properties of chitosan were first assessed in vitro, in isolated mucosal preparations. In a study by Gåserød et al. (1998) microspheres coated with chitosan were found to adhere to isolated porcine gastric mucosa. Chitosan has also been found to adhere to isolated porcine intestinal mucosa in studies in which forces required to detach chitosan films were measured (Lehr et al., 1992), and to rat intestine in wash-off studies in which percentages of chitosan microspheres that adhered to an isolated rat gut loop were taken as measures of the mucoadhesion of chitosan (Takeuchi et al., 1996; He et al., 1998; Yamamoto et al., 2000).

14

Results relating to the possible adherence of chitosan formulations in vivo have been published only recently. There are reports of studies of adherence of chitosan microspheres to murine gastric mucosa (Remuñan-López et al., 2000) and rat intestinal mucosa (Shimoda et al., 2001). In the studies mentioned, the stomach and small intestine of an anesthetized animal were excised a predetermined time after administration of formulations containing chitosan and a fluorecent label. Fluorecence in various regions was measured and taken as a measure of adherence of the formulations. In the study of Remuñan-López et al. (2000), chitosan microspheres were even found in the murine stomach four days after administration. The authors did not, however, provide information on the kinetics of adhesion over time, or record amounts of microspheres remaining. Chitosan microspheres were also found to adhere to intestinal mucosa in studies by Shimoda et al. (2001). Different amounts of microspheres were retained in the rat intestine over a period of eight hours. Although the chitosan formulations studied exhibited mucoadhesive properties in rodents, the authors did not discuss whether their results might be reproducible in human beings. It is, however, well known that extrapolation to man of results obtained in animal studies with formulations intended to retain in the stomach is subject to substantial limitations. For example, Harris et al. (1990a,b), who studied commonly used mucoadhesive polyacrylic-acid-based polymers, found a poor correlation between results in the rat and in man. Although the polymers exhibited good adhesive properties in the rat, adhesion was rare in human volunteers. No significant differences were found between gastric residence times of the polymers studied and residence time of a reference formulation (lactose), in gamma scintigraphic investigations in man. It has been concluded that the poor correlation between results obtained in studies in rodents and results obtained in studies in higher species (the dog, man) can be explained for the most part on the basis of the physiological differences between the species (Jiménez-Castellanos et al., 1993). The discrepancies between results obtained in animal studies and in studies in human volunteers show the importance of studies in man as well as in animals.

Study results make it obvious that more information on the in vivo behaviour of systems containing chitosan is needed, and that there should be particular focus on study of chitosan formulations in human beings. Only after in vivo studies in human volunteers can conclusions be reached regarding the value of chitosan formulations as slow-release systems for gastro-retentive drug delivery. To allow optimization of the characteristics of chitosan-based drug delivery systems in vivo, the effects of different variables, e.g. chitosan content and the grade of chitosan used, on drug absorption and/or gastric residence times of formulations needed to be studied. The effects of the physicochemical properties of chitosan also required systematic evaluation, in vitro and in vivo.

15

1.4. Safety of chitosan in oral use

The biocompatibility of chitosan has been evaluated in several studies (Arai et al., 1968; Muzzarelli et al., 1988; Knapczyk et al., 1989; Kim et al., 2001). Chitosan is currently regarded as relatively safe following oral administration. It is not specifically degraded in the human intestine, because chitinase and chitosanase enzymes are not present (Shahidi et al., 1999). It is therefore not absorbed from the intestine in significant amounts. Limited digestion of chitosan by enzymes produced by bacteria in the intestine may occur (Okamoto et al., 2001; Zhang and Neau, 2002). Results of toxicity tests in rodents have shown that chitosan is relatively harmless following oral administration. Administration at a dose of 15 g/kg in rats (Wistar), in solution or as a powder for 14 days, resulted in no acute toxic effects or pathological changes (Knapczyk et al., 1989). The oral LD50% of chitosan in mice has been reported to be over 16 g/kg (Arai et al., 1968).

Side effects of chitosan following oral administration relate to its effects within the gastrointestinal tract. Consumption of several grams of chitosan daily by humans can result in constipation or diarrhoea, because chitosan entraps water and lipids in the intestine (Koide, 1998; Pittler et al., 1999). In studies in rats, long-term ingestion of high doses of chitosan has also led to reductions in absorption of minerals and fat-soluble vitamins, because of chitosan’s ability to adsorb lipids (Deuchi et al., 1995). In man, in trials lasting for up to 12 weeks, no clinically significant adverse effects or changes in laboratory values relating to safety have been noted. Chitosan treatment had no effects on serum electrolyte levels (Jing et al., 1997) or fat-soluble vitamin levels (Wuolijoki et al., 1999, Pittler et al., 1999). The amounts of chitosan needed in pharmaceutical formulations are fairly low (less than 1000 mg/dose), and risks of side effects in the gastrointestinal tract are therefore also low. Although chitosan is clinically well tolerated it has been suggested that it might not be desirable for administration to individuals allergic to crustaceans, in whom consumption of crustaceans frequently results in allergic reactions, and serious adverse reactions are possible (Ylitalo et al., 2002). On the other hand, chitosan has been widely used in non-medical natural products. No information about allergic reactions relating to these products has been available.

16

2. Study strategy

In the studies described a highly crystalline grade of chitosan base, microcrystalline chitosan (MCCh) was evaluated for the first time as a pharmaceutical excipient. Studies intended to allow determination of properties of MCCh as excipient in granules from which drug release would be controlled through gel formation by the MCCh, and which were also expected to be mucoadhesive. A final aim was to evaluate whether MCCh formulations would allow slow-release of drugs in the human stomach. In addition to the general advantages which it has been suggested that chitosan exhibits in formulations administered orally (see Section 1.3.1.), MCCh could offer particular advantages, such as efficient gel formation in formulations and marked retarding effects on drug release.

2.1. Choice of microcrystalline chitosan and study variables

In previous studies of pharmaceutical formulations little attention has been paid to the effects of the crystallinity of chitosan. Most studies were carried out using chitosans produced commercially, using conventional methods. Chitosans of this kind are fairly amorphous, as indicated by powder X-ray diffraction patterns (Struszczyk, 1987; Genta et al., 1995; Portero et al., 1998; Mura et al., 2003). MCCh differs from conventional chitosan in respect of greater crystallinity, energy of hydrogen bonds, and water retention (Struszczyk, 1987). Both high energy of hydrogen bonds and high water retention are properties reflecting the increase in crystallinity and the substantial surface area of MCCh. The ability of MCCh to retain high amounts of water is a property which could be of particular value in relation to slow-release formulations. MCCh can retain three to four times as much water as the parent chitosan (Struszczyk, 1987). This might result in MCCh having a greater capacity than conventional chitosan to form gels in formulations, and result in marked retardant effects on drug release.

Mucoadhesive tendency of chitosan might also depend on its crystallinity. Efficient gel formation by MCCh could result in substantial mucoadhesion, at least as far as “adhesion by hydration” is concerned. Results of studies relating to technical applications of chitosan have indicated that the reactivity of MCCh is greater than that of conventional chitosan, because of the greater ability of MCCh to form hydrogen bonds (Struszczyk and Kivekäs, 1992). Because adhesion of chitosan to mucosa takes primarily through hydrogen bonding and electrostatic interactions, differences in ability to form hydrogen bonds might be reflected in differences in capacity to adhere to mucosa.

17

All of the properties of MCCh mentioned suggest that it might be used to prepare slow-release and mucoadhesive formulations better than those that can be made using conventional chitosan. It was obvious that the effects of use of MCCh needed to be determined and compared with those seen when conventional chitosan was used. Results of previous in vitro studies relating to conventional chitosan suggested that both drug release from (see Section 1.2.1.) and mucoadhesive properties of (see Section 1.2.2.) chitosan formulations could be affected by altering the molecular weight and degree of deacetylation of chitosan. In the studies described it therefore seemed reasonable in the first place to determine the effects of physicochemical properties of chitosan, including the effects of differences in crystallinity (MCCh versus conventional chitosan), and the effects of differences in molecular weight (Mw) and degree of deacetylation (DD) of MCCh. Other formulation variables that could affect drug release (see Section 1.2.1.) and thus also require evaluation in MCCh formulations were the amount of MCCh used, and the drug substance concerned.

2.2. Choice of study methods

2.2.1. In vitro studies

Studies in vitro in standardized environments are fundamental for characterization of effects of formulation-related factors on the properties of a drug delivery system. The basic idea behind the studies described here was to develop MCCh granules from which drug release could be controlled through formation of gels, and which were also expected to be mucoadhesive. To allow the in vivo behaviour of such formulations to be understood and controlled, the effects of the study variables mentioned in section 2.1. needed to be evaluated first in vitro, concentrating on gel formation by MCCh in granules, drug release from granules, and mucoadhesive tendency.

Gel formation studies. Studies relating to gel formation were needed in the first place to determine whether there were differences in relation to gel formation between MCCh and conventional chitosan. Conduct of such studies was justified particularly by the fact that the ability of a hydrophilic polymer to form gels is usually related to its retardant effects on drug release (Alderman, 1984). Simple methods of determining capacity for gel formation are measurement of increase in weight of a formulation during hydration (gravimetry) and observation of swelling of a formulation (swelling method). Granules can be studied using the latter method.

Dissolution tests. The potential of MCCh as an excipient to modify drug release rates can be assessed by means of dissolution tests on formulations containing different amounts of MCCh and incorporation of model drugs that differ in their solubilities in water at

18

physiological pH levels. Paracetamol was chosen as a representative of drug substances readily soluble throughout the physiological pH range (a Class-I drug in the Biopharmaceutics Classification System (BCS)) (Amidon et al., 1995). Ibuprofen (a BSC Class-II drug) and furosemide (a BCS Class-IV drug) were, in contrast, representative of drug substances slightly soluble in acidic environment. Ibuprofen and furosemide differ in their in vivo absorption characteristics. These differences are discussed in detail in Section 2.2.2. (In vivo studies).

Mucoadhesion tests. In vitro tests were also needed to determine mucoadhesive tendency of MCCh. Several methods currently exist for determining the mucoadhesive tendencies of pharmaceutical excipients in vitro (Peppas et al., 2000). One common method is measurement of detachment force. The potentially mucoadhesive polymer is brought in contact with an isolated mucosal preparation for a predetermined time, and the force required to separate the material from the tissue is then measured. This result can be used as a measure of adhesion. This method has been very popular for decades, perhaps because it is simple and quick. An example of this type of method is the isolated porcine oesophagus preparation developed by Marvola et al. (1982; 1983). This method has turned out to be useful for evaluating the adhesive tendencies of different kinds of pharmaceutical materials, and it was felt that it would also be useful in studies of chitosan. In particular, it was anticipated that results obtained using the method would not represent overestimates of the likely value of the chitosan grades studied in relation to preparation of dosage forms for gastro-retentive drug delivery. Adhesion of chitosan to gastric mucosa in vivo may be greater than adhesion to the oesophageal tissue, because in the stomach pH levels are highly acidic and the mucus gel layer is thick, with a high charge density (Gåserød et al., 1998). The oesophagus contains rather small amount of mucus, pH of which is about 5 to 6. Because the aim of the studies described was to develop MCCh formulations that would be gastro-retentive, it was felt to be desirable to immerse formulations in simulated gastric fluid (pH ~1.2) before adhesion testing, to enhance gel formation by chitosan.

2.2.2. In vivo studies

Conditions in all commonly used in vitro methods differ markedly from conditions in vivo. Data from in vitro studies, and even data from animal studies, often fail to predict the fate of a formulation in man (Davis and Wilding, 2000). This has led to the conclusion being drawn that in many cases “the best model for man is man” (Newman et al., 2003). Since there was little evidence that chitosan formulations behave as intended in vivo (see Section 1.3.2.), the focus of studies described here was on evaluation of the characteristics of MCCh formulations in human volunteers. In vivo study methods that could demonstrate whether MCCh granules had potential as slow-release formulations for gastro-retentive

19

drug delivery were needed. Bioavailability studies and gamma scintigraphic investigations were felt to be appropriate for these purposes. To overcome the limitations that very commonly apply when data from animal studies is extrapolated to human situations, studies were carried out in volunteers.

Bioavailability studies. The slow-release characteristics of MCCh granules can be evaluated by means of bioavailability studies. Such studies were needed to determine whether granules from which drug release was slow could be formulated using MCCh, and whether the slow-release characteristics of such granules in vivo could be controlled by altering variables such as the grade or amount of MCCh. The choice of model drug to be incorporated in the formulations investigated was important. Initially it was felt appropriate to carry out the studies using ibuprofen, for two reasons. Firstly, ibuprofen is readily absorbed throughout the gastrointestinal tract (Wilson et al., 1989). Secondly, it has a short elimination half-life (t½ ~2 h). It was assumed that the short elimination half-life would facilitate investigation of the effects of altering formulation variables on absorption rate, since a decrease in drug release rate would rapidly be reflected in the length of the elimination phase. Subsequently, furosemide was also used. The bioavailability studies with furosemide formulations were intended in particular to provide indirect information on the gastro-retentive properties of MCCh formulations. Absorption of furosemide is strongly site-specific, and takes place in the stomach and upper parts of the small intestine (Staib et al., 1989). If a slow-release MCCh formulation is gastro-retentive, the tmax value for furosemide should be higher and the Cmax value lower than with a conventional furosemide formulation, while the AUC0-∞ value should be similar to or higher than that of a conventional formulation. If, in contrast, a slow-release formulation is not retained in the stomach, both AUC0-∞ and Cmax values should decrease, because most of the furosemide passes the sites of absorption in the upper gastrointestinal tract before being released from the formulation.

Gamma scintigraphy. Results of gamma scintigraphic studies can be used to draw conclusions on the value of MCCh in preparing gastro-retentive formulations. Gamma scintigraphy is currently regarded as the best imaging technique for obtaining data on the fate of a formulation in the gastrointestinal tract by non-invasive means (Wilding et al., 2001; Newman et al., 2003). The greatest advantage of the method over radiological studies that have been widely used is that it allows visualization over time of the entire course of transit of a formulation through the gastrointestinal tract, with reasonably low exposure of subjects to radiation. In X-ray studies the data obtained are inevitably limited because to conduct serial X-rays would expose subjects, e.g. healthy volunteers, to high doses of radiation.

Gamma scintigraphy based on neutron activation has proved to be a particularly good method for study of modified-release drug delivery systems (Wilson, 1998; Wilding

20

et al., 2001). It was felt it would be appropriate for determining whether MCCh formulations were gastro-retentive. The method involves labelling the formulation studied with a nuclide emitting gamma radiation, allowing subsequent imaging of the formulation in the gastrointestinal tract by means of an external gamma camera. A stable isotope (in this case 152samarium) is incorporated in the formulation during manufacture, and activated in a neutron flux to a radioactive isotope (in this case 153samarium, t½ 46.3 hours) before the in vivo studies. There are several advantages of the neutron activation method over other radio-labelling techniques used in gamma scintigraphy. In particular, handling of radioactive isotopes during manufacture is avoided (Digenis and Sandefer, 1991). Other gamma scintigraphic techniques most commonly involve use of 99mtechnetium, which is radioactive during manufacture of the drug products concerned. A short half-life of the label (for 99mTc it is 6 hours) could also result in problems.

A drawback of neutron activation is that it can alter the in vitro and in vivo properties of a formulation. Irradiation can result in degradation of polymeric excipients, e.g. of hydroxypropylmethylcelluloses (HPMC), resulting in decreases in polymer gel viscosities and consequent faster drug release from formulations containing the polymer (Ahrabi et al., 1999; 2000). Before gamma scintigraphic studies effects of neutron activation on MCCh therefore needed to be determined. Because it was anticipated that any changes in polymer structure would be reflected in the properties of gels formed by the polymer, as had been shown in previous studies, in vitro quality control studies, in particular swelling and dissolution studies, were needed to demonstrate that neutron activation process had had no effect on the MCCh. Only MCCh grades the properties of which do not change markedly during irradiation can be used in the in vivo tests intended to reveal whether formulations containing MCCh are gastro-retentive.

21

3. Aims of study

The first objective of this study was to determine the properties of MCCh, a highly crystalline grade of chitosan base, as an excipient that would form gels and control the release rate of drugs, and which was also expected to be mucoadhesive. The properties of MCCh were compared with those of conventional chitosan, which is predominantly amorphous. The effects of altering the molecular weight and degree of deacetylation of MCCh were also studied.

The ultimate objective of the study was to evaluate whether MCCh granules had potential as slow-release systems for gastro-retentive drug delivery. The basic idea was that a gel that controlled drug release would form readily in the stomach from the cationic chitosan used to construct a hydrophilic matrix in the granules. Because hydrated chitosan has been found in several in vitro studies to be mucoadhesive, it was expected that it would adhere to the gastric mucosa and thus make the formulation gastro-retentive. In detail, the aims of the study were:

1. To determine in vitro whether gel formation by chitosan in granules is affected by the crystallinity of chitosan, by comparing formulations made using MCCh with formulations made using conventional chitosan, and to determine the effects of altering the crystallinity, molecular weight and degree of deacetylation on drug release from and the mucoadhesive characteristics of chitosan formulations.

2. To determine in vitro how model drugs (paracetamol, ibuprofen and furosemide) with different aqueous solubilities at physiological pH levels behaved in MCCh granules, and to evaluate the potential of granules containing different amounts of MCCh as slow-release systems for these drugs.

3. To determine, by means of bioavailability tests in human volunteers, how drug absorption is controlled by MCCh in granule formulations, by using in the formulations a drug substance that is readily absorbed throughout the gastro-intestinal tract (ibuprofen) and one absorbed only in the upper regions of the gastrointestinal tract (furosemide).

4. To determine, by means of gamma scintigraphic investigations, whether MCCh formulations have gastro-retentive properties in human volunteers.

22

4. Materials and methods

4.1. Chitosans

The characteristics of microcrystalline chitosan (MCCh) as polymer capable of forming gels that might control drug release, and likely to possess mucoadhesive properties were evaluated. MCCh (Novasso Ltd, Finland) differing in extents of deacetylation (DD) and molecular weights (Mw), and conventional chitosan (Ch) (Primex Ingredients ASA, Norway) were employed. The chitosans studied and their properties, as reported by the supplier, are shown in Table 2.

Table 2. Chitosans studied. MCCh, microcrystalline chitosan; Ch, conventional chitosan; DD, approximate degree of deacetylation; Mw, approximate molecular weight a.

Quality DD (%) Mw (kDa)

MCCh A 75 25

MCCh B 75 150

MCCh C 75 240

MCCh D 90 120

Ch E 90 160 a Range of variation in Mw ± 20 – 30 kDa

MCCh was manufactured from conventional chitosan in accordance with specifications (Struszczyk, 1987) using a continuous method (Finnish Pat. FI83426, 1991). The crystallinity of MCCh manufactured using this method is up to 30% higher than that of conventional chitosan. Chitosan of grade A was representative of a low-Mw chitosan, and grades B, C, D and E of high-Mw chitosans (Table 2). Grades A, B and C had low DD,grades D and E high DD. All of the chitosans studied were in the form of the base. Studies were initially carried out using chitosan of grades A, B, D and E to prepare the formulations concerned. Following a request by us, the MCCh supplier provided MCCh of grade C from study III on.

23

4.2. Model drugs

Model drugs (paracetamol, ibuprofen and furosemide (Ph.Eur.)) that differ in their aqueous solubilities and in their absorption characteristics were incorporated in the formulations.

In the in vitro studies the weak acids paracetamol (pKa 9.5, Mw 151.2 g mol-1) (I), which is readily soluble in water throughout the physiological pH range, and ibuprofen (pKa 5.3, Mw 206.3 g mol-1) (I and II) and furosemide (pKa 3.9, Mw 330.7 g mol-1) (II–IV), which are sparingly soluble in acidic aqueous solutions, were used.

In the bioavailability studies ibuprofen (II), which is readily absorbed throughout the gastrointestinal tract (Wilson et al., 1989), and furosemide (II and III), which is site-specifically absorbed only in the upper regions of the gastrointestinal tract (Staib et al., 1989), were used. Both drugs have short elimination half-lives (t½). Values reported are 2 ± 0.5 hours for ibuprofen and 0.5 to 2 hours for furosemide (Handbook of Clinical Drug Data, 1993).

4.3. Study formulations

The studies were carried out using chitosan matrix granules (I–V), except in the case of the in vitro mucoadhesion studies, in which the chitosans were compressed into tablets (III). In preliminary tests (I) dilute chitosan gels were studied. Detailed descriptions of the compositions and preparation of the formulations are given in the original publications (I–V). Conventional granulating and tableting methods were used.

Granules. Percentages of chitosan in granules varied from 5 to 95 (I–V). In study II, enteric polymers insoluble at gastric pH levels were incorporated in the granule matrices (0, 2.5, 5 and 10%, Eudragit S100®, Röhm Pharma, Germany) or used as coatings (0, 5, 10 and 20%, Aqoat AS-HF®, Shin-Etsu Chemical Co., Japan) to determine whether they could be used to reinforce the gels formed by MCCh, and control drug release at acidic pH levels. In study III an organic acid was added to the formulation (0, 2.5, 5 and 10% of tartaric or citric acid (Ph.Eur.)), to determine whether doing so enhanced gel formation by MCCh, and prolonged drug release. In the gamma scintigraphic studies (IV and V) natural-abundance samarium oxide (Sm2O3) (Aldrich, USA) was used as radio-label in the granules. Lactose (Pharmatose DCL 21, DMV International, Netherlands) granules which contained polyvinylpyrrolidone (PVP K25, Fluka Chemie, Switzerland) as binder were used as reference formulation.

Tablets. The chitosan tablets used in the in vitro adhesion studies (III) contained only chitosan and 1% of magnesium stearate (Ph.Eur.). The reference tablets were enteric-coated (Eudragit S100®, Röhm Pharma, Germany) tablets of microcrystalline cellulose (Emcocel LP 200, Mendell, USA).

24

4.4. In vitro studies

4.4.1. Properties of chitosan gels (I)

The effect of grade of chitosan on the properties of gels formed was first studied, at various chitosan concentrations (1, 2 and 4%). The gels were prepared in slightly and markedly acidic aqueous solutions. Gel pH levels (InLab412 combination electrode, Mettler Toledo, Switzerland) and apparent single viscosity values (DV-II Brookfield digital viscometer, Brookfield Engineering Laboratories, USA) were determined at room temperature.

4.4.2. Gel formation by chitosan in granules (I, IV)

The effect of grade of chitosan on gel formation by chitosan in granules (I) was studied by means of light microscopy (Leica DMLB 020-519.511, Leica Mikroskopie und Systeme, Germany). Increases in granule diameters were measured over time, under markedly (pH 1.2 and 2.2) and slightly (pH 5.8) acidic pH conditions. Gel formation was also studied (at pH 1.2) when determining the effects of neutron activation on MCCh (IV).

4.4.3. Drug release from chitosan granules (I–IV)

The effects on drug release of grade of chitosan, amount of chitosan used, and other excipients in the granules were studied by means of dissolution tests (Apparatus Distek Premiere 5100, Distek, USA), using the basket method (USP 24). The dissolution media used were phosphate buffer pH 5.8 (1000 ml, 37 ± 0.5oC) for granules containing the slightly soluble acidic drugs ibuprofen and furosemide, and hydrochloric acid buffer pH 1.2 and phosphate buffer pH 5.8 (500 ml, 37 ± 0.5oC) for granules containing paracetamol, which is readily soluble. In studies III and IV, with furosemide as model drug, granules were pre-treated at pH 1.2 (in 7.5 ml of buffer) for one hour, to simulate pH conditions in the stomach, before dissolution tests were conducted at pH 5.8. Speeds of rotation in tests were 100 min-1 (ibuprofen, furosemide) and 50 min-1 (paracetamol). Amounts of drug released were determined spectrophotometrically (Ultrospec 4000, Pharmacia Biotech, UK). Dissolution tests were also carried out when determining the effect of neutron activation on MCCh (IV).

Drug-release kinetics (I) were evaluated using the empirical equation Mt/M∞ = ktn

describing fractional release from swellable devices (Ritger and Peppas, 1987a,b). In this equation, Mt/M∞ is the fractional amount of drug released at time t, k is the kinetic constant and n is an exponent indicative of the diffusional mechanism of release. The exponent nhas limit values that depend on the geometry of the device, which indicate whether

25

diffusion-controlled Fickian release is taking place. Greater values of n indicate non-Fickian release. In such a case drug release depends on the ratio between polymer relaxation rate and rate of diffusion of drug in gel.

4.4.4. In vitro mucoadhesion of chitosan (III)

The effect of grade of chitosan on its mucoadhesive tendency was studied using the isolated porcine oesophagus preparation described by Marvola et al. (1982; 1983). Tablets containing various grades of chitosan were first moistened in simulated gastric fluid (pH ~1.2) without enzymes (USP 24) and then placed on an oesophageal preparation. The shear force needed to detach the tablet from the preparation was taken as a measure of the mucoadhesive tendency of the grade of chitosan concerned. Enteric-coated (Eudragit S100®, Röhm Pharma, Germany) tablets were used as reference formulations in the studies because the enteric coating was expected to exhibit only a slight tendency to adhere to the isolated porcine oesophagus.

4.5. In vivo studies

4.5.1. Bioavailability studies (II–III)

Six groups of eight or nine healthy volunteers participated in a series of randomized, crossover single-dose studies. Each volunteer had given informed consent in writing to participation in the studies. The investigations were carried out in accordance with the recommendations of the Declaration of Helsinki (World Medical Assembly, 1964) as revised in Edinburgh in 2000. The study protocol had been approved by the Ethics Committee of the University Hospital of Tartu.

Procedure. Compositions of the formulations studied are described in Table 3. The amounts of ibuprofen and furosemide administered in the study were 300 mg (II), and 40 mg (II and III), respectively. Granules dispensed in size-0 gelatine capsules (Ph.Eur.) were administered to each subject with 200 ml of water, after the subject had fasted overnight for at least 10 hours. Lunch was provided four hours after drug administration. Venous blood samples were collected at intervals. The washout period between administration of different formulations was at least one week.

Assay methods. Drug concentrations in plasma were determined by means of high-performance liquid chromatography (HPLC). Methods described by Avgerinos and Hutt (1986) and Beermann (1982) were used, with slight modifications, for ibuprofen and furosemide determinations, respectively. Accuracy and precision of the methods were

26

investigated as recommended by Shah et al. (2000). Validation criteria were met by both methods (II, Section 2.6.; III, Section 2.7.).

Data analysis. Maximum concentration (Cmax) and time to peak concentration (tmax)were obtained directly from individual time-versus-plasma-concentration curves. Pharmacokinetic parameters assessed (Siphar®, Simed, France) were absorption rate constant (ka), area under the concentration-time curve (AUC0-∞), apparent elimination half-life (t½) and mean residence time (MRT). AUC0-∞ values were calculated using the trapezoidal method. The method of Wagner and Nelson was used to estimate the apparent absorption rate constant (ka). Rate of absorption was also evaluated by means of the ratio Cmax/AUC. Statistical analyses were carried out using Student’s paired t-test or Wilcoxon’s non-parametric test (for tmax values). When results in different groups of volunteers were compared, Student’s t-test relating to independent groups or the Mann-Whitney non-parametric test (for tmax values) was used.

Table 3. Compositions of granules in bioavailability studies. For groups 1 to 3 the drug was ibuprofen (300 mg), for groups 4 to 6 furosemide (40 mg).

Study II Group 1 Group 2 Group 3 Group 4

Drug 60% 60% 60% 60% 60 b 60 b 60 b 60% 60%

MCCh Mw 150 kDa 40% 30% 35% 37.5% 40 b 40 b 40 b 40% 30%

Enteric polymer a - 10% 5% 2.5% - - - - 10%

Coating aa - - - - 5% 10% 20% - -

Study III Group 5 Group 6

Drug 20% 5% 5% 5 bb

MCCh Mw 150 kDa 80% 95% - -

MCCh Mw 240 kDa - - 95% 95 bb

Tartaric acid - - - 5%

a Enteric polymer in granules matrices; aa Enteric polymer as granule coating, b Polymer to drug ratio 40:60 in granule matrices; bb Polymer to drug ratio 95:5 in granule matrices, granules prepared from masses containing 95% of MCCh and 5% of furosemide by replacing 5% of the mass with tartaric acid.

27

4.5.2. Gamma scintigraphic investigations (IV–V)

Three groups of five healthy male volunteers participated in the gamma scintigraphic studies. Their weights varied from 62 to 97 kg and their body mass indices (BMI) from 19 to 27 kg m-2. Informed consent to participation in the studies had been obtained in writing. The investigations were carried out in accordance with International Conference on Harmonization (ICH) Good Clinical Practice Guidelines and the Declaration of Helsinki (World Medical Assembly, 1964) and subsequent amendments. The study protocol had been approved by the Ethics Committee of Helsinki University Hospital (HUS) and the Finnish National Agency for Medicines. The studies were carried out in HUS Nuclear Medicine Division, which has a radiation safety licence issued by STUK (Radiation and Nuclear Safety Authority of Finland). Safety requirements were set in accordance with the guidelines established by STUK. The ALARA (as-low-as-reasonably-achievable) principle was observed, and exposure to radiation was minimised in every situation.

Procedure. Compositions of study formulations are described in Table 4. Granules were dispensed in size-0 gelatine capsules (Ph.Eur.) by volume. The masses of the contents of individual capsules were 260 mg (Formulation 1), 290 mg (Formulation 2) and 370 mg (Formulation 3). Each capsule contained 4 mg of 152Sm2O3 in the granules. The 152Sm was activated in a thermal neutron flux to the gamma emitting nuclide 153Sm (t½ 46.3 hours), using a 250kW TRIGA Mark II nuclear research reactor (General Atomics, USA) at the VTT Technical Research Centre of Finland. Gamma scintigraphic studies were carried out 48 hours after neutron activation. This time period allowed decay of unwanted radioisotopes; primarily 24Na. Gamma spectra and radioactivity of 153Sm were measured to determine the safety of the formulations used in the in vivo studies in healthy volunteers. Safety requirements were met for every formulation (IV, Section 3.2.).

Table 4. Compositions of study formulations in gamma scintigraphic studies (IV and V). The granules were dispensed into size-0 gelatine capsules, each containing 4 mg of samarium oxide.

Ingredient Formulation 1 Formulation 2 Formulation 3 MCCh Mw 150 kDa 95% 40% -

Lactose 3.4% 58.6% 96.4%

Sm2O3 1.6% 1.4% 1.1%

PVP - - 2.5%

28

A capsule containing the granules was administered to each volunteer in a sitting position, with 180 ml of water, after the subject had fasted overnight for at least 12 hours. The volunteers were not allowed to eat or drink during the imaging period. Anterior and posterior images, each of one-minute duration, were recorded continuously for the first 30 minutes, by means of a dual-head gamma camera (ADAC Forte, ADAC Laboratories, USA), after which six images, each of one-minute duration, were recorded every 15 minutes for the next three to four hours. During imaging each subject lay supine beneath the gamma camera. At other times they could move freely.

Data analysis. Regions of interest (ROI) relating to the stomach (IV and V) or oesophagus (V) were drawn manually on gamma images, and gamma counts relating to ROIs were calculated using Hermes software (version 3.7, Nuclear Diagnostics, Sweden). Geometric means of counts in paired anterior and posterior images were calculated. Gastric emptying of the formulations was expressed in terms of remaining relative counts (RELcounts) in each ROI as a function of time. Lag times before onset of gastric emptying were obtained directly from individual gastric emptying curves. RELcounts between 0.90 and 0.10 were used to determine the gastric-emptying rate constant (k) by means of linear regression analysis. Times at which half of the granules had left the stomach (t50%) were used in evaluating gastric residence times. Statistical analyses were carried out using non-parametric Kruskal-Wallis analysis of variance.

29

5. Results and discussion

5.1. In vitro characteristics of MCCh formulations

The various grades of MCCh and conventional chitosan retarded release of the slightly soluble drug substances ibuprofen and furosemide from granules, and exhibited fairly marked tendencies to adhere to oesophageal mucosa in vitro, suggesting that the chitosans studied could be used to prepare slow-release formulations that would also be mucoadhesive. Gel formation by chitosan and its retardant effects on drug release, and its mucoadhesive capacity could be controlled by altering the grade of chitosan used. Other formulation-related variables that affected drug release were the nature of the drug, the amount of chitosan used, and the natures of other excipients in granules. The pH of the environment influenced the effects of altering these variables. The main results from the in vitro studies are shown in Table 5.

Table 5. Effects of altering formulation variables and environmental factors on the in vitro characteristics of formulations containing different grades of chitosan.

Variable Effect Chitosan • Crystallinity MCCh formed gels most efficiently and had a more marked

retardant effect on drug release than conventional chitosan (I), no effect on mucoadhesive capacity of chitosan (III)

• Mw Increase in the Mw of MCCh decreased drug release rate (I, III) and increased mucoadhesive capacity of chitosan (III)

• DD DD of MCCh had no significant effects on drug release rate (I) or mucoadhesive capacity of chitosan (III)

• Amount Increasing the amount of MCCh decreased drug release rate (I, III). With slightly soluble acidic drug substances the effect could be opposite at pH levels that were only slightly acidic (I)

Model drug Slow-release granules were obtained only when using slightly soluble drug substances (I–III)

Environmental pH The lower the pH the more marked were gel formation by chitosan and retardant effects on drug release (I, III)

Other excipients The slow-release characteristics of chitosan granules could be modified by use of enteric polymers in granule matrices or as coatings (II) and by use of organic acids in granule matrices (III)

30

5.1.1. Gel-forming ability of MCCh

Hydration of chitosans in granules and consequent gel formation by the chitosans resulted in swelling of the granules, which made it possible to determine gel-forming abilities of the chitosans via measurements of granule diameters (I). Granules containing MCCh increased more in diameter than corresponding granules made from conventional chitosan (I, Fig. 3), indicating that hydration and subsequent gel formation takes place more efficiently with MCCh than with conventional chitosan. Gel formation was greatest when pH levels were markedly acidic. On the other hand, the less efficient gel formation by conventional chitosan resulted in more rapid granule degradation.

Gel formation by chitosan in granules (Orienti et al., 1996) and tablets (Mi et al., 1997) is known to depend on pH because of the cationic character of the polymer. The results of our in vitro studies indicate that the crystallinity of chitosan could also affect its gel-forming properties. A particular property of MCCh is its ability to take up substantial amounts of water on hydration (Struszczyk, 1987). Our findings support the idea that this property could be reflected in the efficacy of gel formation by MCCh in granule formulations. It is well known that pronounced gel formation by a hydrophilic polymer is beneficial when slow-release, matrix-type dosage forms are being developed. This has been seen, e.g., in studies with HPMC (Alderman, 1984). The property is usually related to marked retardant effects on drug release. Our findings therefore suggested that MCCh might be better than conventional chitosan as a gel-forming excipient in matrix-type dosage forms. The fact that MCCh forms gels most readily at acidic pH levels make it of potential value in formulations intended to release drugs slowly in the stomach.

5.1.2. Effects of grade of MCCh on release and mucoadhesive characteristics of chitosan formulations

Results of further studies indicated that drug release from chitosan granules could be controlled by altering the crystallinity of the chitosan used, e.g. by using MCCh grade or the more amorphous conventional chitosan (I). Altering the properties of MCCh used affected both the slow-release and mucoadhesive characteristics of formulations (I, III). In this respect the important property of MCCh grade was its molecular weight.

MCCh versus conventional chitosan. The mechanism of drug release from chitosan granules was regarded as non-Fickian diffusion, on the basis of exponent (n) values (I, Tables 3 and 4). Such a mechanism can be explained on the basis of control of drug release through gel formation by chitosan, and thereafter both by diffusion of drug through the gel and gel erosion. On the basis of the value for kinetic constant (k), drug release was most retarded with MCCh in the granules (I, Tables 3 and 4; MCCh grades A and C versus

31

conventional chitosan grade D), as expected from the results of studies of gel formation in vitro. The greatest retardant effect on drug release was achieved with the MCCh grade which had also a pronounced ability for gel formation. The granules concerned contained 40% of MCCh with a degree of deacetylation of 75% and molecular weight of 150 kDa (I, grade A), and ibuprofen. These in vitro results indicate that increasing the crystallinity of chitosan could make its retardant effects on drug release from matrix-type formulations more marked, as a result of increasingly efficient gel formation by the chitosan. Better slow-release dosage forms might be obtained through use of MCCh rather than conventional chitosan. On the basis of these promising results an application has been made for a patent relating to slow-release granule formulations containing MCCh (Finnish Pat. Application. FI20000780, 2000).

In studies of chitosans as mucoadhesive excipients, the various grades of MCCh and conventional chitosan exhibited fairly marked mucoadhesive tendencies (III). Forces of detachment of chitosan tablets from isolated oesophagus preparations were always statistically significantly greater (about 2.5- to 6-fold, p < 0.01) than those for the enteric-coated reference tablet. MCCh and conventional chitosan had equally strong capacities to adhere to mucosal tissue (III, Fig. 1; grades II and III versus grade I). It is known that the properties of a polymer relating to hydration and gel-forming can affect its mucoadhesive capacity (Junginger, 1991). The finding that MCCh had a particularly pronounced ability to form gels suggested that MCCh and conventional chitosan might differ in relation to their mucoadhesive properties. Results of studies relating to technical applications of MCCh have also indicated that the reactivity of MCCh is greater than that of conventional chitosan (Struszczyk and Kivekäs, 1992). This property could theoretically result in MCCh adhering to mucus to a greater extent than conventional chitosan. However, neither of these two properties of MCCh were reflected in terms of mucoadhesive strength of chitosan in our studies using isolated oesophagus preparations. This finding suggests that chitosan crystallinity had no effect on the mucoadhesive properties of the chitosans studied. MCCh and conventional chitosan could both be valuable excipients in the development of mucoadhesive formulations.

Effects of the Mw and DD of MCCh. The properties of MCCh formulations could be controlled by using MCCh of different molecular weight but not by using MCCh of different degree of deacetylation. The retardant effects of MCCh on drug release from granules (I, Fig. 4; III, Fig. 2) and the mucoadhesive capacity (III, Fig. 1) were most marked with MCCh of highest molecular weights (Mw 150 kDa and 240 kDa). These findings are similar to those in previous studies, in which decreases in rates of release of drug from granules (Goskonda and Upadrashta, 1993) and increases in the mucoadhesive capacities of chitosan (Lehr et al, 1992) as chitosan molecular weight increased were reported. The findings are understandable, because increases in the molecular weight of

32

MCCh markedly increase viscosities of gels formed (I, Fig. 1). Increases in the viscosity of chitosan gels result in increasingly retarded drug release (Kristl et al., 1993), and could enhance the mucoadhesive properties of the polymer gel (Junginger, 1991). As the chain length of the polymer increases penetration of polymer chains into the mucus layer could also increase (Peppas and Buri, 1985). In our studies using isolated oesophagus preparations, forces of detachment of chitosan tablets increased 2.5-fold as the molecular weight of the MCCh increased from 150 kDa to 240 kDa (III, Fig. 1; 0.75 ± 0.29 N (grade II); 1.85 ± 0.60 N (grade IV), p < 0.01).

In contrast, degree of deacetylation of MCCh had no significant effects on drug release (I, Fig. 4) or the mucoadhesive properties of formulations (III, Fig. 1). However, MCCh of low molecular weight had unexpectedly good mucoadhesive properties (III, Fig. 1). This might be explained on the basis that the reactivity of amino groups is high with chitosans of low molecular weight because the amino groups are easily accessible in the structures of such polymers (Sabnis and Block, 2000). If so, effects of degree of deacetylation on the mucoadhesive capacity of chitosan could be particularly evident with chitosans of low molecular weight. In a recent study by Sabnis et al. (1997) it was found that drug release could also be controlled by changing the degree of deacetylation of chitosan. However, Sabnis et al. used chitosans of very low molecular weight (20 kDa), which might explain why the findings differ from those in our study. In the studies reported here, effects of altering degree of deacetylation on drug release were studied using only MCCh grades of high molecular weight (I; grade A versus grade C).

It was concluded from the results of our in vitro studies that MCCh grades of high molecular weights were likely to behave best in formulations intended to release drugs slowly in the stomach, and such grades were therefore chosen for studies in vivo. Such grades of chitosan had marked retardant effects on drug release and strong mucoadhesive capacities in vitro. For the first bioavailability studies (II) MCCh of molecular weight of 150 kDa was available. Later, MCCh of 240 kDa was also used.

5.1.3. Effects of drug solubility and amount of MCCh on release characteristics of chitosan formulations

Effect of drug solubility. Drug solubility was found to affect the slow-release characteristics of MCCh granules substantially (I and II). Model drugs which are slightly soluble in acidic environments (I and II; ibuprofen and furosemide) were relatively slowly released but retardation of release of a readily soluble drug (I; paracetamol) was minimal. This finding is similar to findings in previous studies of granules containing conventional chitosan. Slow release from formulations containing slightly soluble drug substances, e.g. indomethacin and diclofenac has been reported (Miyazaki et al., 1988a,b; Tapia et al.,

33

1993) but preparation of slow-release formulations of a readily soluble drug substance (theophylline) has been found difficult (Henriksen et al., 1993).

Release of ibuprofen at pH 5.8 was slower than release of furosemide from corresponding MCCh formulations (II, Figs. 1 and 3; 40% MCCh of Mw 150 kDa). At pH 5.8 furosemide, which is a stronger acid (pKa 3.9) than ibuprofen (pKa 5.3) dissolves more easily than ibuprofen. Retardant effects of MCCh on drug release were most marked when gel formation by the MCCh took place at markedly acidic pH levels, when the granules had been hydrated at pH 1.2 prior being subjected to dissolution testing at pH 5.8 (III, Fig. 3). This result was expected, because gel formation by MCCh in granules had been found to be most marked at acidic pH levels (I, Fig. 3), and pH level affected the viscosity of MCCh gels (I, Table 2). Gels prepared in a buffer at pH 1.2 were found to be markedly more viscous than gels prepared in a buffer at pH 5.8. Our results are similar to findings in studies relating to conventional chitosan (Goskonda and Upadrashta, 1993; Mi et al., 1997). The ability of MCCh to retard drug release most efficiently at acidic pH levels could be valuable in relation to development of formulations intended to release drugs slowly in the stomach.

Effect of amount of MCCh. Rate of drug release could be controlled by varying the amount of MCCh in granules. The extent of the effect depended on the model drug concerned, and pH. Results relating to paracetamol, which dissolves readily regardless of pH throughout the physiological pH range, differed from those with ibuprofen and furosemide, which are slightly soluble at acidic pH levels. Increasing the amount of MCCh decreased rate of paracetamol release from granules (I, Fig. 8). This finding is understandable because increasing the amount of MCCh markedly increases the viscosities of gels formed by MCCh (I, Fig. 1). However, even incorporation of a substantial percentage of MCCh in the granules (60%) did not result in a satisfactory slow-release formulation for paracetamol. This finding was similar to that of Henriksen et al. (1993), who studied a readily soluble drug (theophylline) in granules containing conventional chitosan (high Mw, DD 77%). Henriksen et al. concluded that gel formation by chitosan would have to be more pronounced for retardant effects on drug release to be marked. However, the results of our study show that even use of a grade of chitosan that has good gel-forming properties did not result in slow release of a readily soluble drug substance.

In contrast, formulations containing ibuprofen and furosemide, which are slightly soluble at acidic pH levels, exhibited satisfactory slow-release properties in vitro. Increasing the percentage of MCCh in the granules decreased rate of drug release, but only when gel formation by MCCh took place at a markedly acidic pH level, when the granules had been hydrated at pH 1.2 prior being subjected to dissolution testing at pH 5.8 (III, furosemide). When dissolution testing took place at pH 5.8 with no pre-treatment, high percentages of MCCh were associated with highest release rates (I, Fig. 7; ibuprofen; III,

34

Fig. 3; furosemide). One mechanism of action could be that high amounts of MCCh in granules increase pH levels of the gels formed (I, Fig. 2), with consequential increases in rates of dissolution of the acidic drugs studied. This mechanism could come into play in slightly acidic environments, in which gel formation by MCCh is less efficient, with drugs that are weak acids.

The conclusion from the results of the in vitro studies was that slow-release granules are likely to be achievable only if the drugs concerned are slightly soluble in acidic environments, like ibuprofen and furosemide. The granules studied might behave best as slow-release formulations if they were administered to subjects in a fasted state. The granules would then be exposed to an acidic gastric milieu, and gel formation by MCCh would be expected to be most efficient and its retardant effects on drug release most marked. However, administration of granules to individuals in a fasted state would be likely to restrict the extent to which the granules could be used. In an attempt to overcome this disadvantage small amounts of acidic excipients (2.5 to 10% of tartaric and citric acids) were incorporated in granules, in the hope of enhancing gel formation by MCCh and ensuring slow release of acidic drugs even at pH levels close to neutral. Addition of these acids did not, however, have the desired effect. A marked retardant effect on drug release occurred only if MCCh had formed a gel in a markedly acidic environment (III, Fig. 4).

5.2. Release characteristics of MCCh formulations in vivo

5.2.1. Bioavailabilities of drugs from MCCh granules; effects of amount and molecular weight of chitosan

MCCh granules were evaluated as slow-release dosage forms in human subjects in a fasted state. The granules were administered in gelatine capsules. A formulation containing 40% of MCCh of molecular weight 150 kDa was a satisfactory slow-release dosage form of ibuprofen, as determined from tmax value. The mean tmax was 2.9 ± 1.1 hours (II, Table 1). With conventional dosage forms, maximum ibuprofen concentration in plasma occurs after only about one hour (Davies, 1998). Results of studies with furosemide showed that the slow-release properties of the formulations studied depend significantly on the drug concerned. Furosemide absorption from granules similar to those that had released ibuprofen slowly was fairly rapid (II, Fig. 4 versus Fig. 6). The mean tmax for furosemide from granules containing 40% of MCCh was only slightly higher (not statistically significant) than the mean tmax for a conventional furosemide tablet (II, Table 3). Slight increase in tmax could be explained by the fact that drug absorption from MCCh granules was subject to a lag time (II, Fig. 6). It was obvious that results of in vitro studies could not allow adequate prediction of how formulations would behave in human subjects.

35

Attempts were made (III) to retard rate of release of furosemide from granules by incorporating higher percentages of MCCh (80 to 95%) and by using MCCh of higher molecular weight (240 kDa). Although decreases in rates of release of drug had been noted in the in vitro studies (III, Fig. 2), it was found that tmax did not change significantly (III, Table 2). However, amounts of furosemide absorbed, as determined from AUC0-∞ and Cmax values, were lower as in vitro release rates decreased. An obvious explanation for these findings is that as rate of release of furosemide from granules decreased, most of the drug in the granules had moved beyond the “absorption window” for furosemide in the upper gastrointestinal tract. There was no change in tmax, because furosemide is poorly absorbed once it has passed its “absorption window”.

Little information has previously been available on the behaviour in vivo of slow-release chitosan formulations, especially with chitosan as the only excipient. The results of one of the first studies to be carried out in animals indicate that slow-release formulations might be obtainable by using chitosan in granule matrices (Miyazaki et al., 1988a). Increasing the percentage of chitosan (Mw or DD not stated) in granules from 33 to 66% increased tmax for indomethacin, which is slightly soluble and readily absorbed. However, the effect was not statistically significant in a small group of animals (see Section 1.3.2. for details). The results of bioavailability studies with MCCh formulations also indicate that use of chitosan could allow simple preparation of slow-release dosage forms of slightly soluble drugs. In our study in human volunteers, absorption of ibuprofen was markedly retarded by MCCh. The results of our studies with furosemide, which is absorbed only in the upper gastrointestinal tract, suggest, however, that MCCh can be utilised in preparation of slow-release formulations only if the drug concerned is absorbed throughout the gastrointestinal tract.

Slow release of a drug from a formulation is beneficial if, e.g., there is a need to prevent high peak concentrations of the drug in plasma and it is desirable instead for drug concentrations in plasma to be approximately constant over a long period. A satisfactory slow-release dosage form of ibuprofen was obtained when granules containing 40% of MCCh (Mw 150 kDa) were administered in gelatine capsules to subjects in a fasted state. The formulation did not, however, meet the criterion for a prolonged-release formulation, because the apparent elimination half-life (t½) for ibuprofen did not change markedly. It was 2.4 ± 0.5 hours (II, Table 1), which is fairly similar to the half-life of the drug substance itself (~ 2 hours). Incorporation of higher amounts of MCCh or MCCh of higher molecular weight in the granules might result in greater retardation of drug release. This is suggested by the results of our studies relating to furosemide.

36

5.2.2. Effects of other excipients in MCCh granules

Attempts were made to make MCCh granules that not only released drugs slowly but also acted as prolonged-release formulations. It could be desirable for drug release from a dosage form in vivo to be so slow that it is the rate-limiting step in drug kinetics, i.e. very slow absorption would restrict the rate of elimination of the drug. In such a case, the increase in the apparent half-life of the drug substance would allow a longer interval between doses. This could be beneficial in particular in relation to drug substances the elimination half-lives of which are short. If the aim were to decrease rate of release of drug in vivo, incorporation of higher percentages of MCCh would undoubtedly be one of the first steps to be tried. However, incorporation of high percentages of excipients in formulations is impractical if a high dose of drug is also required. An alternative means of preparing prolonged-release formulation could be incorporation of a chitosan of high molecular weight. MCCh grades of molecular weights greater than 150 kDa were not available during the early stages of our studies in vivo (II). We therefore tried other approaches foreseen when the study strategy was planned.

The effects of incorporating enteric polymers (2.5 to 10%) in MCCh matrices or employing them as granule coatings (5 to 20%) were studied with granules containing ibuprofen (II). It was aimed that the polymers, which are insoluble at low pH levels, would reinforce the gels formed by the MCCh and control release of drug in the stomach. In other experiments, tartaric acid (5%) was added to the matrices of granules containing furosemide (III). The aim in this case was to retard release of drug further by enhancing gel formation by MCCh, and to ensure that the drug was released slowly even at pH close to neutral.

Only incorporation of enteric polymer in MCCh granule matrices was found to have any effect in vivo. Absorption of ibuprofen could easily be retarded by including enteric polymer in granules (II, Fig. 4, Table 1). There were statistically significant differences (p < 0.05) in relation to the rate parameters ka, tmax, MRT and Cmax/AUC between granules containing only MCCh (40%, Mw 150 kDa) and granules in which 10% of the MCCh in the matrices had been replaced with enteric polymer. The marked increase in apparent t½

for ibuprofen indicated that absorption had become the rate limiting step in drug kinetics, and that a prolonged-release formulation had been achieved. Apparent t½ values were 2.4 ± 0.5 hours for granules containing only MCCh and 4.9 ± 2.0 hours for granules in which 10% of the MCCh had been replaced by enteric polymer (p < 0.01). A two-fold increase in t½ allows the interval between doses to be doubled. The results demonstrate that incorporation of small amounts of enteric polymers in granule matrices makes it possible to prepare prolonged-release MCCh formulations simply. However, the need for different kinds of chitosan grades to be on the market is obvious.

37

5.3. Retention of MCCh formulations in the stomach

5.3.1. Information obtained via bioavailability studies

Because furosemide is absorbed only in the upper parts of the gastrointestinal tract (Staib et al., 1989), the results of the bioavailability studies with MCCh formulations containing furosemide indicated whether the formulations were gastro-retentive (II, III). If the drug was released in the stomach, amounts of furosemide absorbed would be similar to or greater than those after administration of a conventional formulation.

The mean AUC0-∞ for furosemide from granules containing 40% of MCCh (Mw

150 kDa) (3450 ± 1180 µg l-1 h) was only slightly smaller (not statistically significant) than that from a conventional tablet (3800 ± 941 µg l-1 h) (II, Table 3). The bioavailability of furosemide was high, despite a lag time in relation to absorption (~ 30 minutes). These findings suggested that the MCCh granules had been retained in the stomach, and led to extension of the patent application relating to MCCh formulations (Finnish Pat. Application FI20000780, 2000) for gastro-retentive drug delivery (International Pat. Application WO01/76562 A1, 2001). However, the results of the following studies (III) were less promising. Amounts of furosemide absorbed decreased as in vitro release rates decreased, when a relatively high amount of MCCh (80 and 95%) or MCCh of relatively high molecular weight (240 kDa) was incorporated in granules (III, Figs. 2 and 5, Table 2). When the amount of MCCh (Mw 150 kDa) in the granules was 95%, the mean AUC0-∞

(2700 ± 1070 µg l-1 h) for furosemide was 29% lower than the AUC0-∞ for a conventional tablet that had been determined in study II. Clear et al. (2001) reported similar decreases in amounts of furosemide absorbed when drug release took place in the proximal intestine instead of the stomach (comparing events after conventional tablets had been released from an Intelisite® capsule in the duodenum and after oral administration of tablets). The findings relating to the granules that contained 95% of MCCh (Mw 150 kDa) obviously do not suggest substantial retention in the stomach. With a formulation containing the same percentage of MCCh of higher molecular weight (240 kDa), the decrease in AUC0-∞ was more marked (50%), indicating that much of the furosemide had passed the “absorption window” before it had been released from the formulation. It is obvious that the granules had remained in the stomach for too short a time in relation to the release rate of the drug.

The bioavailability studies were useful in determining the properties of MCCh formulations and indirectly helped evaluation of their potential as gastro-retentive systems. Although the MCCh grades studied exhibited fairly marked mucoadhesive capacities in vitro, no similar effect was evident from the results of the bioavailability studies. The correlation between the results of in vitro and in vivo studies was especially poor in relation to MCCh of molecular weight 240 kDa. Although this grade was found to adhere most strongly in the in vitro studies (III, Fig. 1; grade IV), the bioavailability of furosemide

38

was lowest from formulations that contained MCCh of Mw 240 kDa (III, Table 2). This finding suggests that no substantial adhesion of MCCh to gastric mucosa had taken place. Further studies that would provide direct information of the fate of MCCh formulations in the human gastrointestinal tract (such as gamma scintigraphic investigations) were needed.

5.3.2. Gamma scintigraphic investigations

Preliminary studies. Only MCCh grades the properties of which did not change markedly during neutron activation were usable in formulations for gamma scintigraphic investigation. The effects of irradiation on MCCh (Mw 150 kDa and 240 kDa) were therefore studied in vitro (IV). It was found that irradiation did not affect the properties of MCCh of molecular weight 150 kDa (IV, Figs. 1a and 2a). However, the MCCh grade with the highest molecular weight (240 kDa) was particularly sensitive to irradiation. Gels were formed less readily by this grade of MCCh following irradiation, resulting in more rapid drug release (IV, Figs. 1b and 2b). The effect become more marked as the irradiation dose increased. These findings are in accordance with findings in previous studies indicating that irradiation can result in degradation of polymers with glycoside linkages (e.g. ethyl cellulose (Waaler et al., 1997), HPMC and pectin (Ahrabi et al., 1999) and MCCh (us in study IV)).

Gamma scintigraphic investigations. MCCh of molecular weight 150 kDa was selected for use (IV). Because we had found that the amount of MCCh in granules affected the bioavailability of furosemide (see Section 5.3.1.), the same percentages (40 and 95%) of MCCh were used in formulations in gamma scintigraphic investigations.

The reference formulation (lactose granules in gelatine capsules) passed fairly rapidly through the stomach (IV, Fig. 5, Table 4). After a short lag time before onset of gastric emptying (IV, Table 3) it took from 15 minutes to about one hour for the granules to be cleared from the stomachs of five volunteers, on the basis of the rate constant (k) (IV, Table 4). In most cases gastric emptying took place in small and frequent boluses. In contrast, passage of MCCh granules from the stomach was prolonged in three occasions out of nine (IV, Figs. 3 and 4), in one subject of the four who received a formulation containing 95% of MCCh, and in two subjects of the five who received a formulation containing 40% of MCCh. Maximum individual t50% values (times at which half of the granules had left the stomach) were 2.1 hours (95% of MCCh, Subject 1) and 2.3 hours (40% of MCCh, Subject 5) (IV, Table 4). For the reference formulation, t50% varied from 0.2 hours to 1 hour (mean 0.5 ± 0.3 hours). From the gamma images it was concluded that the time MCCh granules remained in the stomach was prolonged because the formulation had adhered to the gastric mucosa. This is shown in Figure 2 relating to Subject 5, who had

39

Figure 2. Gamma scintigraphic images showing the fates of granules containing 40% of MCCh of Mw 150 kDa in the stomach of Subject 5. (a) Gelatine capsule containing the granules has not disintegrated (4 minutes). (b) Capsule has disintegrated and granules have become attached to gastric mucosa (12 minutes). (c) Granules still attached to the gastric mucosa but some have passed through the pylorus and reached the small intestine (2.5 hours). (d) Most granules have left the stomach (4 hours).

been given granules containing 40% of MCCh. The granules were released from the gelatine capsule in approximately 5 minutes, after which about two thirds of the granules (IV, Fig. 4; Subject 5), became attached to the gastric mucosa (corpus region), and remained attached two hours after administration. However, in the other six volunteers who had received MCCh granules (containing 40 or 95% of MCCh), radioactivity in the stomach region declined fairly rapidly. The curves relating to passage of MCCh granules from the stomach were rather similar to those for the reference formulation (IV, Figs. 3 and 4 versus Fig. 5). It was obvious that in most cases the MCCh granules did not adhere to the gastric mucosa or that adhesion was not strong.

The results of the gamma scintigraphic investigations are in good accordance with the results of the bioavailability studies. In the latter, amounts of furosemide absorbed decreased as rate of release of drug from the granules declined, indicating that the granules

40

did not remain long enough in the stomach in relation to release rate of drug. The results of the gamma scintigraphic investigations showed that MCCh formulations rarely adhere to gastric mucosa. The results also showed that there were no significant differences in the times for which formulations containing different amounts of MCCh (40 and 95%) remained in the stomach. The finding in bioavailability studies that the AUC0-∞ for furosemide was higher for granules containing the lowest amount of MCCh can be explained by more rapid release of the drug from the formulation concerned. This would have allowed furosemide absorption to take place mostly in the upper gastrointestinal tract, in the “absorption window”, irrespective of whether the granules had adhered to gastric mucosa, and resulted in the higher bioavailability observed.

It is evident from the results of the gamma scintigraphic investigations that the behaviour of MCCh formulations studied as gastro-retentive systems is not sufficiently reproducible. Although the results of several in vitro studies (Lehr et al., 1992; He et al., 1998; Gåserød et al., 1998; and us in study III), and results of a few studies in animals (Remuñan-López et al., 2000; Shimoda et al., 2001) suggest that chitosans could be of value as excipients that adhere to the mucosa of the gastrointestinal tract, we found no evidence from studies in human subjects that chitosan formulations could be used for gastro-retentive drug delivery. Prolonged gastric residence times of formulations were rare, even though the study conditions had been designed to eliminate factors that could hinder gastro-retention (e.g., the study was carried out in fasting volunteers).

Various physiological and formulation-related factors could explain why the systems did not work as expected. One formulation-related factor is that the granules were administered in gelatine capsules. It has been suggested that administration of formulations containing mucoadhesive polymers in gelatine capsules makes the gastro-retentive properties of the formulations less marked (Harris et al., 1990b). The gelatine in the capsule shell could interact with the mucoadhesive polymer during dissolution, and diminish the adhesive properties of the formulation. The consequences of administering granules in gelatine capsules require study. It is nevertheless obvious that administration of granules in capsules is more convenient than their administration loose. Another factor that could affect in vivo behaviour is granule size. In a recent study in mice, Remuñan-López et al. (2000) suggested that chitosan granules might be gastro-retentive if they were very small, i.e. if microparticles that could be taken into the mucosa were prepared.

However, physiological factors might be most significant in making the in vivo behaviour of formulations erratic. There are substantial differences between conditions in in vivo and in vitro studies. It is therefore not surprising that many polymers known to possess mucoadhesive properties in vitro (e.g. the polyacrylic-acid-based polymers investigated by Khosla and Davis (1987) and Harris et al. (1990a) and the MCCh studied by us) fail to exhibit gastro-retentive properties in studies in human beings. The

41

physiological conditions in the stomach may be such that it is difficult to make formulations containing mucoadhesive polymers remain there for long. The layer of mucus to which adhesion could take place is continually eroding, and this could hinder persistence of adhesive systems in the stomach (Allen et al., 1993). Exposure of chitosan to soluble mucins in the stomach lumen before it has a chance to interact with mucus gel layer could also adversely affect the adhesive properties of formulations (Gåserød et al., 1998). Although results of studies in animals (e.g. the rat, dog) relating to mucoadhesion of polymers have been promising, little tendency to adhesion has been found in studies in man (Harris et al., 1990a,b; Davis et al., 1993). Because there are always limitations on extrapolation of results of animal studies to human beings, studies in human subjects are needed. In the studies described here, gamma scintigraphy allowed information about the fate of formulations in the gastrointestinal tracts of the human subjects to be obtained directly, and made it possible to reach conclusions about whether the MCCh formulations studied had potential as gastro-retentive dosage forms.

5.4. Additional information provided by gamma scintigraphy on the behaviour of MCCh formulations in vivo

Gamma scintigraphy provided additional information about the behaviour of the MCCh formulations studied in the gastrointestinal tract. Adherence of the formulation to the oesophagus was noted in one volunteer. It is well known that adherence of a dosage form to the oesophagus can result in oesophageal injury, which can be severe. More than 70 drugs in common use have been reported to be associated with oesophageal disorders (Jaspersen, 2000). Failure of formulations to move rapidly through the oesophagus is a problem associated not only with pharmaceutical products but also with non-medical natural products, in which the presence of substantial amounts of gel-forming polymers can result in oesophageal obstruction (Opper et al., 1990). Our finding of adherence of an MCCh formulation to the oesophagus indicates problems that might be associated with use of formulations of the kinds studied. A detailed case report has therefore been published (V). Results relating to the passage of the formulation from the stomach of this volunteer were excluded from analysis (IV) because the situation was clearly so different from situations in the other subjects.

The formulation concerned consisted of granules containing 95% of MCCh of molecular weight of 150 kDa, in a gelatine capsule. The capsule adhered initially to the distal oesophagus (V, Fig. 1). The gelatine shell started to disintegrate within five minutes of administration, after which radioactivity in the oesophageal area started to decline as the contents of the capsule moved towards the stomach (V, Fig. 2). However, about two thirds of the radioactivity remained detectable in the oesophageal region for 1.75 hours. This

42

could be explained on the basis that not only the gelatine shell but also MCCh granules were adhering to the oesophageal mucosa after disintegration of the capsule shell had commenced. This explanation is consistent with the results of in vitro studies (III) in which MCCh exhibited a fairly marked tendency to adhere to the isolated porcine oesophagus (III, Fig. 1).

Normally, oesophageal transit times of dosage forms are short, only some 10 to 20 seconds (Channer and Virjee, 1985; Wilson et al., 1988; Burton et al., 1995). It is, however, well known that adhesion of dosage forms to the oesophagus can occur in up to 20% of cases, the risk of adhesion being high, e.g., with gelatine capsules that become sticky as they hydrate. It is interesting that the results suggest that gel-forming chitosan also adhered to the oesophageal mucosa. This finding suggests that use of chitosan may need to be avoided in formulations containing drug substances associated with oesophageal injury, particularly if gelatine capsules are also used. Caution might also be needed in relation to non-medical natural products containing chitosan. Amounts of chitosan in such products are often high, daily intakes usually amounting to several grams. The substantial amounts of gel-forming chitosan and the large size of the products could make them likely to stagnate in the oesophagus, and increase risk of oesophageal obstruction.

43

6. Conclusions

In vitro and in vivo studies were undertaken on the properties of MCCh, a highly crystalline grade of chitosan base, as an excipient. The most significant information to emerge from this study related to the in vivo behaviour of MCCh formulations in humans. One major defect in previous studies is that most have relied exclusively on in vitro methods to determine the properties of chitosan-based formulations. In the study described here the results of dissolution and mucoadhesion tests in vitro often did not allow adequate prediction of how MCCh formulations would behave in human subjects. This underscores the importance of conducting biopharmaceutical studies in human volunteers at an early stage during the evaluation of new excipients. The following conclusions can be drawn from the results of this study:

1. Various grades of MCCh and more amorphous conventional chitosan were useful in preparing granules from which drug release in vitro was controlled by formation of gels in acidic and slightly acidic environments. The release mechanism was drug diffusion through the gel and gel erosion. All the chitosans studied had fairly marked mucoadhesive capacities in vitro.

2. Gel formation by chitosan in vitro, retardant effects of chitosan on drug release and mucoadhesive capacities of chitosan depended on its physicochemical properties. MCCh was superior in terms of gel formation and accordingly had a more marked retardant effect on drug release than conventional chitosan. The molecular weight of MCCh was important: the greater the Mw the slower the release and the greater the strength of adhesion to mucus. Degree of deacetylation of MCCh had no significant effect. All these findings indicate that MCCh of high Mw should be valuable as an excipient in preparing slow-release dosage forms that are also mucoadhesive.

3. Drug release in vitro could also be controlled by altering the amount of MCCh in the formulation. Increasing the MCCh content decreased the release rate after gel formation at acidic pH. The solubility of the drug concerned substantially affected the slow-release characteristics of the granules. Slow release took place only with drugs having low solubility at acidic pH levels (ibuprofen, furosemide). Even incorporation of high amounts of MCCh did not result in slow release of the readily soluble drug (paracetamol).

44

4. The results of bioavailability studies in human volunteers showed that MCCh granules can be used in gelatine capsules to prepare slow-release dosage forms of slightly soluble drugs, provided these are absorbed throughout the gastrointestinal tract (such as ibuprofen). With furosemide, which is absorbed only in the upper gastrointestinal tract, there were no significant differences in tmax between formulations, even though differences in release rates had been seen in vitro. The amounts of furosemide absorbed were affected, indicating that the granules did not remain long enough in the stomach in relation to the release rate of the drug. These findings demonstrate the importance of carrying out studies in vivo.

5. Incorporation of small amounts of an enteric polymer into MCCh granule matrices considerably prolonged ibuprofen absorption. This finding suggest that a combination of chitosan, which forms gels, and a polymer insoluble at acidic pH levels could be valuable in the development of prolonged-release formulations.

6. The results of gamma scintigraphic investigations showed that the behaviour of MCCh formulations as gastro-retentive systems is insufficiently reproducible. Although MCCh exhibited a fairly marked mucoadhesive capacity in vitro, adhesion of MCCh onto gastric mucosa in vivo was rare. These results also demonstrate the importance of biopharmaceutical studies. Gamma scintigraphy was valuable in studying the fate of formulations in the human stomach.

7. The results of gamma scintigraphy revealed that a formulation containing MCCh had adhered for a considerable time to the oesophagus in one volunteer. This finding indicates that there could be problems associated with the use of MCCh granules, particularly if the granules are administered in gelatine capsules. It also serves as a reminder that drugs known to be capable of damaging the oesophagus should not be used in formulations where there is the risk of oesophageal adherence.

45

Acknowledgements

This study was carried out at the Division of Biopharmaceutics and Pharmacokinetics, Department of Pharmacy, University of Helsinki. Part of the work was performed at Tartu University Hospital, at Helsinki University Hospital (HUS), Division of Nuclear Medicine, and at the Technical Research Centre of Finland (VTT).

I wish to express my sincere gratitude to my supervisor Professor Martti Marvola, Head of the Division of Biopharmaceutics and Pharmacokinetics, for suggesting the topic of this study and for providing excellent working facilities. I am most grateful for his scientific guidance as well as for his constant enthusiasm and encouragement, all of which made the completion of this study possible.

I thank most sincerely the reviewers of this thesis, Docent Jyrki Heinämäki and Docent Mika Vidgren, for their constructive criticism and for giving me valuable suggestions for its improvement. I am also grateful to Edmund Gordon (Lääketieteellinen käännöstoimisto Oy - Medical translations) for revising the language of the manuscript.

I also express my sincere gratitude to Professor Peep Veski, Head of Institute of Pharmacy at the University of Tartu, Estonia, for his valuable support in the bioavailability studies. My thanks also go to Erja Piitulainen for her expert technical assistance during the HPLC analysis of the biological samples back here at the Division of Biopharmaceutics and Pharmacokinetics.

I want to thank my co-investigators, Janne Marvola and Hanna Kanerva, for their excellent collaboration during the gamma scintigraphic studies carried out at HUS. I also express my gratitude to Docent Aapo Ahonen from the Division of Nuclear Medicine, and to the personnel of the isotope laboratory. Our associates at VTT FiR-1 research laboratory, Seppo Salmenhaara and Juhani Aittomäki, who were responsible for neutron activation of the study formulations, and Maija Lipponen and Tommi Kekki, who were responsible for the gamma spectra measurements, made valuable contributions.

My sincere thanks are due to Niina Tapanainen, Ulrika Seppälä, Paula Heinänen, Sanna Ojala, Anu Linna, Tiina Tuononen and Jarno Kaarlas for their participation as M.Sc. students in this study.

I am grateful to all my colleagues at the Division of Biopharmaceutics and Pharmacokinetics for creating a friendly atmosphere in which to work, and for sharing my moments of success and despair. I particularly wish to thank Jaana Hietala for her great company and friendship during all the years of research and my Ph.D. studies in biopharmaceutics, and Pirjo Nykänen for sharing the excitement towards the end.

46

Finally, I owe my warmest thanks to dear Mika for his patience and loving support throughout the years of research relating to chitosans and especially during the final hectic months when the results were incorporated into the thesis. I also thank my parents, Mirja-Liisa and Lasse, for their everlasting support during the course of my education, and especially my late father Lasse, from whom I have inherited my sense of ambition.

The financial support of the Graduate School in Pharmaceutical Research (Ministry of Education, Finland), the National Technology Agency of Finland (TEKES), the Finnish Cultural Foundation, and the Pharmacal Research Foundation (Finland) is gratefully acknowledged.

Helsinki, August 2003

Mia Säkkinen

47

References

Acartürk, F., 1989. Preparation of a prolonged-release tablet formulation of diclofenac sodium. Pharmazie 44, 547-549.

Acartürk, F., Sencan, A., Celebi, C., 1993. Evaluation of the effect of low-molecular weight chitosan on the solubility and dissolution characteristics of spironolactone. Pharmazie 48, 605-607.

Adusumilli, P.S., Bolton, S.M., 1991. Evaluation of chitosan citrate complexes as matrices for controlled release formulations using a 32 full factorial design. Drug Dev. Ind. Pharm. 17, 1931-1945.

Ahrabi, S.F., Sande, S.A., Waaler, T., Graffner, C., 1999. Effects of thermal neutron irradiation on some potential excipients for colonic delivery systems. Drug Dev. Ind. Pharm. 25, 453-462.

Ahrabi, S.F., Heinämäki, J., Sande, S.A., Graffner, C., 2000. Influence of neutron activation factors on matrix tablets for site specific delivery to the colon. Eur. J. Pharm. Sci. 10, 225-235.

Akbuğa, J., 1993a. Use of chitosonium malate as a matrix in sustained-release tablets. Int. J. Pharm. 89, 19-24.

Akbuğa, J., 1993b. The effect of the physicochemical properties of a drug on its release from chitosonium malate matrix tablets. Int. J. Pharm. 100, 257-261.

Alderman, D.A., 1984. A review of cellulose ethers in hydrophilic matrices for oral controlled-release dosage forms. Int. J. Pharm. Technol. Prod. Manuf. 5, 1-9.

Allen, A., 1978. Structure of gastrointestinal mucus glycoproteins and the viscous and gel-forming properties of mucus. Br. Med. Bull. 34, 28-33.

Allen, A., Flemström, G., Garner, A., Kivilaakso, E., 1993. Gastroduodenal mucosal protection. Physiol. Rev. 73, 823-857.

Amidon, G.L., Lennernäs, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm. Res. 12, 413-420.

Arai, K., Kinumaki, T., Fujita, T., 1968. Toxicity of chitosan. Bull. Tokai Reg. Fish. Res. Lab. 43, 89-94.

Aral, C., Özbas-Turan, S., Kabaskal, L., Keyer-Uysal, M., Akbuğa, J., 2000. Studies of effective factors of plasmid DNA-loaded chitosan microspheres. I. Plasmid size, chitosan concentration and plasmid addition techniques. S.T.P. Pharma. Sci. 10, 83-88.

Aspden, T.J., Mason, J.D.T., Jones, N.S., Lowe, J., Skaugrud, O., Illum, L., 1997. Chitosan as a nasal delivery system: the effect of chitosan solutions on in vitro and in vivo mucociliary transport rates in human turbinates and volunteers. J. Pharm. Sci. 86, 509-513.

Avgerinos, A., Hutt, A.J., 1986. High performance liquid chromatographic determination of ibuprofen in human plasma and urine by direct injection. J. Chromatogr. 380, 468-471.

Beermann, B., 1982. Kinetics and dynamics of furosemide and slow-acting furosemide. Clin. Pharmacol. Ther. 32, 584-591.

48

Bernkop-Schnürch, A., Humenberger, C., Valenta, C., 1998. Basic studies on bioadhesive delivery systems for peptide and protein drugs. Int. J. Pharm. 165, 217-225.

Berth, G., Dautzenberg, H., 2002. The degree of acetylation of chitosan and its effect on the chain conformation in aqueous solution. Carbohydr. Polym. 47, 39-51.

Burton, S., Washinton, N., Steele, R.J.C., Musson, R., Feely, L., 1995. Intragastric distribution of ion-exchange resins: A drug delivery system for the topical treatment of the gastric mucosa. J. Pharm. Pharmacol. 47, 901-906.

Carreño-Gómez, B., Duncan, R., 1997. Evaluation of the biological properties of soluble chitosan and chitosan microspheres. Int. J. Pharm. 148, 231-240.

Chandy, T., Sharma, C.P., 1993. Chitosan matrix for oral sustained delivery of ampicillin. Biomaterials 14, 939-944.

Channer, K.S., Virjee, J.P., 1985. The effect of formulation on oesophageal transit. J. Pharm. Pharmacol. 37, 126-129.

Chen, J.L., Cyr, G.N., 1970. Compositions producing adhesion through hydration. In: Manly, R.S. (Ed.), Adhesion in biological systems, Academic Press, London, 172.

Chew, J.L., Wolfowicz, C.B., Mao, H.-Q., Leong, K.W., Chua, K.Y., 2003. Chitosan nanoparticles containing plasmid DNA encoding house dust mite allergen, Der p 1 for oral vaccination in mice. Vaccine 21, 2720-2729.

Clear, N.J., Milton, A., Humphrey, M., Henry, B.T., Wulff, M., Nichols, D.J., Anziano, R.J., Wilding, I., 2001. Evaluation of the Intelisite capsule to deliver theophylline and frusemide tablets to the small intestine and colon. Eur. J. Pharm. Sci. 13, 375-384.

Cölfen, H., Berth, G., Dautzenberg, H., 2001. Hydrodynamic studies on chitosan in aqueous solution. Carbohydr. Polym. 45, 373-383.

Davies, N.M., 1998. Clinical pharmacokinetics of ibuprofen. The first 30 years. Clin. Pharmacokinet. 34, 101-154.

Davis, S.S., Wilding, E.A., Wilding, I.R., 1993. Gastrointestinal transit of a matrix tablet formulation: comparison of canine and human data. Int. J. Pharm. 94, 235-238.

Davis, S.S., Wilding, I.R., 2000. Oral drug absorption studies: the best model for man is man. Drug Discov. Today 6, 127-128.

Deacon, M.P., Davis, S.S., White, R.J., Nordman, H., Carlstedt, I., Errington, N., Rowe, A.J., Harding, S.E., 1999. Are chitosan-mucin interactions specific to different regions of the stomach? Velocity ultracentrifugation offers a clue. Carbohydr. Polym. 38, 235-238.

Deacon, M.P., McGurk, S., Roberts, C.J., Williams, P.M., Tendler, S.J., Davies, M.C., Davis, S.S., Harding, S.E., 2000. Atomic force microscopy of gastric mucin and chitosan mucoadhesive systems. Biochem. J. 348, 557-563.

Deuchi, K., Kanauchi, O., Shizukuishi, M., Kobayashi, E., 1995. Continuous and massive intake of chitosan affects mineral and fat-soluble vitamin status in rats fed on a high-fat diet. Biosci. Biotechnol. Biochem. 59, 1211-1216.

Digenis, G.A., Sandefer, E. P., 1991. Gamma scintigraphy and neutron activation techniques in the in vivo assessment of orally administered dosage forms. Crit. Rev. Ther. Drug Carrier Syst. 7, 309-345.

49

Dodane, V., Vilivalam, V.D., 1998. Pharmaceutical applications of chitosan. Pharm. Sci. Technol. Today 1, 246-253.

European Pat. Application EP1238663 A2, 2002. Mixed hydrophilic gels as controlled release vehicles of floating dispersants. de la Mata, M.B., Escribano, A.M., Matji, T.J.A., Dominiguez, V.H., Gallo, M.R., Garcia, M.F., Gerrero, G.-P., Pivel, R.J.P.

Felt, O., Buri, P., Gurny, R., 1998. Chitosan; a unique polysaccharide for drug delivery. Drug Dev. Ind. Pharm. 24, 979-993.

Fiebrig, I., Harding, S.E., Rowe, A.J., Hyman, S.C., Davis, S.S., 1995. Transmission electron microscopy on pig gastric mucin and its interactions with chitosan. Carbohydr. Polym. 28, 239-244.

Finnish Pat. FI83426, 1991. Method for continuous manufacture of microcrystalline chitosan. Novasso Ltd.

Finnish Pat. Application FI20000780, 2000. Peroraalinen lääkemuoto lääkeaineen hallituksi vapauttamiseksi. Säkkinen, M., Marvola, M.

Ganza-Gonzáles, A., Anguiano-Igea, S., Otero-Espinair, F.J., Méndez, B.J., 1999. Chitosan and chondroitin microspheres for oral-administration controlled release of metoclopramide. Eur. J. Pharm. Biopharm. 48, 149-155.

Genta, I., Pavanetto, F., Conti, B., Giunchedi, P., Conte, U., 1995. Improvement of dexamethasone dissolution rate from spray-dried chitosan microspheres. S.T.P. Pharma. Sci. 5, 202-207.

Genta, I., Perugini, P., Pavanetto, F., Modena, T., Conti, B., Muzzarelli, R.A., 1999. Microparticulate drug delivery systems. EXS 87, 305-313.

Goskonda, S., Upadrashta, S., 1993. Avicel RC591/chitosan beads by extrusion-spheronization technology. Drug. Dev. Ind. Pharm. 19, 915-927.

Guan, Y.L., Shao, L., Yao, K.D., 1996. A study on correlation between water state and swelling kinetics of chitosan based hydrogels. J. Appl. Polym. Sci. 61, 2325-2335.

Gåserød, O., Jolliffe, A.G., Hampson, F.C., Dettmar, P.W., Skjåk-Bræk, G., 1998. The enhancement of the bioadhesive properties of calcium alginate gel beads by coating with chitosan. Int. J. Pharm. 175, 237-246.

Handbook of Clinical Drug Data, 1993. Knoben, J.E., Anderson, P.O., (Ed.), 7th ed., Drug Intelligence Publications, Hamilton, 585, 669.

Harris, D., Fell, J.T., Sharma, H.L., Taylor, D.C., 1990a. Gastrointestinal transit of potential bioadhesive formulations in man: a scintigraphic study. J. Controlled Release 12, 45-53.

Harris, D., Fell, J.T, Taylor, D.C., Lynch, J., Sharma, H.L., 1990b. GI transit of potential bioadhesive systems in the rat. J. Controlled Release 12, 55-65.

He, P., Davis, S.S., Illum, L., 1998. In vitro evaluation of the mucoadhesive properties of chitosan microspheres. Int. J. Pharm. 166, 75-88.

Hejazi, R., Amiji, M., 2002. Stomach-specific anti-H. pylori therapy. I: Preparation and characterization of tetracycline-loaded chitosan microspheres. Int. J. Pharm. 235, 87-94.

Henriksen, I., Skaugrud, O., Karlsen, J., 1993. Use of chitosan malate as an excipient in wet granulation of three water soluble drugs. Int. J. Pharm. 98, 181-188.

50

Hoekstra, A., Struszczyk, H., Kivekäs, O., 1998. Percutaneous microcrystalline chitosan application for sealing arterial puncture sites. Biomaterials 19, 1467-1471.

Hou, W.-M., Miyazaki, S., Takada, M., Komai, T., 1985. Sustained release of indomethacin from chitosan granules. Chem. Pharm. Bull. 33, 3986-3992.

Illum, L., Farraj, N.F., Davis, S.S., 1994. Chitosan as a novel nasal delivery system for peptide drugs. Pharm. Res. 11, 1186-1189.

Illum, L., 1998. Chitosan and its use as a pharmaceutical excipient. Pharm. Res. 15, 1326-1331.

Illum, L., Watts, P., Fisher, A.N., Hinchcliffe, M., Norbury, H., Jabbal-Gil, I., Nankervis, R., Davis, S.S., 2002. Intranasal delivery of morphine. J. Pharmacol. Exp. Ther. 301, 391-400.

Imai, T., Shiraishi, S., Otagiri, M., 2000. A strategy for the immediate-release of indomethacin from a sustained-release preparation using chitosan hydrolysate. S.T.P. Pharma. Sci. 10, 57-62.

Inouye, K., Machida, Y., Sannan, T., Nagai, T., 1988. Buoyant sustained release tablets based on chitosan. Drug Des. Delivery 2, 165-175.

Inoyatov, N., Celebi, N., Acarturk, F., 1998. Preparation and evaluation of a prolonged release pentoxyphylline tablet with chitosan. Pharm. Ind. 60, 472-475.

International Pat. Application WO01/76562 A1 (PCT/FI01/00322), 2001. A peroral pharmaceutical formulation for controlled release of a drug. Säkkinen, M., Marvola, M.

Jameela, S.R., Kumary, T.V., Lal, A.V., Jayakrishnan, A., 1998. Progesterone-loaded chitosan microspheres: long acting biodegradable controlled delivery system. J. Controlled Release 52, 17-24.

Japanese Pat. JP1224311, 1989. Production of floating granule retaining in stomach. Nagai, T.

Japanese Pat. JP5017371, 1993. Drug composition. Otagiri., M.

Jaspersen, D., 2000. Drug-induced oesophageal disorders. Drug Saf. 22, 237-249.

Jiménez-Castellanos, N.R., Zia, H., Rhodes, C.T., 1993. Mucoadhesive drug delivery systems. Drug Dev. Ind. Pharm. 19, 143-194.

Jing, S.B., Li, L., Ji, D., Takiguchi, Y., Yamaguchi, T., 1997. Effect of chitosan on renal function in patients with chronic renal failure. J. Pharm. Pharmacol. 49, 721-723.

Junginger, H.E., 1991. Mucoadhesive hydrogels. Pharm. Ind. 53, 1056-1065.

Kanauchi, O., Deuchi, K., Imasato, Y., Shizukuishi, M., Kobayashi, E., 1995. Mechanism for the inhibition of fat digestion by chitosan and for the synergistic effect of ascorbate. Biosci. Biotechnol. Biochem. 59, 786-790.

Kas, S., 1997. Chitosan: properties, preparations and application to microparticulate systems. J. Microencapsulation. 14, 689-711.

Kawashima, Y., Lin, S.Y., Kasai, A., Handa, T., Takenaka, H., 1985. Preparation of a prolonged release tablet of aspirin with chitosan. Chem. Pharm. Bull. 33, 2107-2113.

Khalid, M.N., Ho, L., Agnely, F., Grossiord, J.L., Couarraze, G., 1999. Swelling properties and mechanical characterization of a semi-interpenetrating chitosan/polyethylene oxide network. Comparison with a chitosan reference gel. S.T.P. Pharma. Sci. 9, 359-364.

Khosla, R., Davis, S.S., 1987. The effect of polycarbophil on the gastric emptying of pellets. J. Pharm. Pharmacol. 39, 47-49

51

Kim, S.-K., Park, P.-J., Yang, H.-P., Han, S.-S., 2001. Subacute toxicity of chitosan oligosaccharide in Sprague-Dawley rats. Arzneim.–Forcsh. 51, 769-774.

Knapczyk, J., Krówczynski, L., Pawlik, B., Liber, Z., 1989. Pharmaceutical dosage forms with chitosan. In: Skjåk-Bræek, G., Anthonsen, T., Sandford, P. (Eds.), Chitin and Chitosan: Sources, Chemistry, Biochemistry, Physical Properties and Applications, Elsevier Applied Science, London, 665-670.

Knapczyk, J., 1993a. Excipient ability of chitosan for direct tableting. Int. J. Pharm. 89, 1-7.

Knapczyk, J., 1993b. Chitosan hydrogel as a base for semisolid drug forms. Int. J. Pharm. 93, 233-237.

Koide, S.S., 1998. Chitin-chitosan – properties, benefits and risks. Nutr. Res. 18, 1091-1101.

Kotzé, A.F., de Leeuw, B.J., Luessen, H.L., de Boer, A.G., Verhoef, J.C., Junginger, H.E., 1997. Chitosans for enhanced delivery of therapeutic peptides across intestinal epithelia: in vitro evaluation in Caco-2 cell monolayers. Int. J. Pharm. 159, 243-253.

Kotzé, A.F., Luessen, H.L., de Leeuw, B.J., de Boer, A.G., Verhoef, J.C., Junginger, H.E., 1998. Comparison of the effect of different chitosan salts and N-trimethylchitosan chloride on the permeability of intestinal epithelial cells (Caco-2). J. Controlled Release 51, 35-46.

Kotzé, A.F., Luessen, H.L., de Boer, A.G., Verhoef, J.C., Junginger, H.E., 1999. Chitosan for enhanced intestinal permeability: Prospects for derivatives soluble in neutral and basic environments. Eur. J. Pharm. Sci. 7, 145-151.

Kristl, J., Smid-Korbar, J., Struc, E., Schara, M., Rupprecht, H., 1993. Hydrocolloids and gels of chitosan as drug carriers. Int. J. Pharm. 99, 13-19.

Kristmundsdottir, T., Ingvarsdottir, K., Saemundsdottir, G., 1995. Chitosan matrix tablets: the influence of excipients on drug release. Drug Dev. Ind. Pharm. 21, 1591-1598.

Lee, K.Y., Kwon, I.C., Kim, Y.-H., Jo, W.H., Jeong, S.Y., 1998. Preparation of chitosan self-aggregates as a gene delivery system. J. Controlled Release 51, 213-220.

Lee, J.W., Park, J.H., Robinson, J.R., 2000. Bioadhesive based dosage forms: The next generation. J. Pharm. Sci., 89, 850-866.

Lehr, C.M., Bouwstra, J.A., Schacht, E.H., Junginger, H.E., 1992. In vitro evaluation of mucoadhesive properties of chitosan and some other natural polymers. Int. J. Pharm. 78, 43-48.

Lehr, C.M., 1996. From sticky stuff to sweet receptors – achievements, limits and novel approaches to bioadhesion. Eur. J. Drug Metab. Pharmacokinet. 21, 139-148.

Li, Q., Dunn., E.T., Grandmaison, E.W., Goosen, M.F.A., 1992. Applications and properties of chitosan. J. Bioact. Compat. Polym. 7, 370-397.

Loke, W.K., Lau, S.K., Yong, L.L., Khor, E., Sum., C.K., 2000. Wound dressing with sustained anti-microbial capability. J. Biomed. Mater. Res. 53, 8-17.

Marvola, M., Vahervuo, K., Sothman, A., Marttila, E., Rajaniemi, M., 1982. Development of a method for study of the tendency of drug products to adhere to the oesophagus. J. Pharm. Sci. 71, 975-977.

Marvola, M., Rajaniemi, M., Marttila, E., Vahervuo, K., Sothman, A., 1983. Effect of dosage form and formulation factors on the adherence of drugs to the oesophagus. J. Pharm. Sci. 72, 1034-1036.

52

Mi, F.L., Her, N.L., Kuan, C.Y., Wong, T.B., Shyu, S.S., 1997. Chitosan tablets for controlled release of theophylline – effect of polymer-drug wet or dry blending and anionic-cationic interpolymer complex. J. Appl. Polym. Sci. 66, 2495-2505.

Mi, F.L., Shyu, S.S., Wu, Y.B., Lee, S.T., Shyong, J.Y., Hunag, R.N., 2001. Fabrication and characterization of a sponge-like asymmetric chitosan membrane as a wound dressing. Biomaterials 22, 165-173.

Miyazaki, S., Yamaguchi, H., Yokouchi, C., Takada, M., Hou, W.-M., 1988a. Sustained release and intragastric-floating granules of indomethacin using chitosan in rabbits. Chem. Pharm. Bull. 36, 4033-4038.

Miyazaki, S., Yamaguchi, H., Yokouchi, C., Takada, M., Hou, W.-M., 1988b. Sustained release of indomethacin from chitosan granules in beagle dogs. J. Pharm. Pharmacol. 40, 642-643.

Miyazaki, S., Nakayama, A., Oda, M., Takada, M., Attwood, D., 1995. Drug release from oral mucosal adhesive tablets of chitosan and sodium alginate. Int. J. Pharm. 118, 257-263.

Mura, P., Zerrouk, N., Mennini, N., Maestrelli, F., Chemtob, C., 2003. Development and characterization of naproxen-chitosan solid systems with improved drug dissolution properties. Eur. J. Pharm. Sci. 19, 67-75.

Muzzarelli, R., Baldassare, V., Conti, F., Ferrara, P., Biagini, G., Gazzanelli, G., Vasi, V., 1988. Biological activity of chitosan: ultrastructure study. Biomaterials 9, 247-252.

Newman, S.P., Hirst, P.H., Wildng, I.R., 2003. New developments in radionuclide imaging for assessing drug delivery in man. Eur. J. Pharm. Sci. 18, 19-22.

Nigalaye, A.G., Adusumilli, P., Bolton, S., 1990. Investigation of prolonged drug release from matrix formulations of chitosan. Drug Dev. Ind. Pharm. 16, 449-467.

Nordman, H., Davies, J.R., Herrman, A., Karlsson, N.G., Hanson, G.C., Carlstedt, I., 1997. Mucus glycoproteins from pig gastric mucosa: identification of different mucin populations from the surface epithelium. Biochem. J. 326, 903-910.

Nwe, N., Chandrkrachang, S., Stevens, W., Maw, T., Tan, T., Khor, E., Wong, S., 2002. Production of fungal chitosan by solid state and submerged fermentation. Carbohydr. Polym. 49, 235-237.

Okamoto, Y., Nose, M., Miyatake, K., Sekine, J., Oura, R., Shigemasa, Y., Minami, S., 2001. Physical changes of chitin and chitosan in canine gastrointestinal tract. Carbohydr. Polym. 44, 211-215.

Okhamafe, A.O., Amsden, B., Chu, W., Goosen, M.F., 1996. Modulation of protein release from chitosan-alginate microcapsules using pH-sensitive polymer hydroxypropyl methylcellulose acetate succinate. J. Microencapsulation. 13, 497-508.

Opper, F.H., Isaacs, K.L., Warshauer, D.M., 1990. Oesophageal obstruction with a dietary fibre product designed for weight reduction. J. Clin. Gastroenterol. 12, 667-669.

Orienti, I., Aiedeh, K., Gianasi, E., Bertasi, V., Zecchi, V., 1996. Indomethacin loaded chitosan microspheres. Correlation between the erosion process and release kinetics. J. Microencapsulation 13, 463-472.

Patel, D., Smith, A.W., Grist, N., Barnett, P., Smart, J.D., 1999. An in vitro mucosal model predictive of bioadhesive agents onto the oral cavity. J. Controlled Release 61, 175-183.

53

Patel, D., Smith, J.R., Smith, A.W., Grist, N., Barnett, P., Smart, J.D., 2000. An atomic force microscopy investigation of bioadhesive polymer adsorption on to human buccal cells. Int. J. Pharm. 200, 271-277.

Paul, W., Sharma, C.P., 2000. Chitosan, a drug carrier for the 21st century: a review. S.T.P. Pharma. Sci. 10, 5-22.

Peppas, N.A., Buri, P.A., 1985. Surface, interfacial and molecular aspects of polymer bioadhesion on soft tissues. J. Controlled Release 2, 257-275.

Peppas, N.A., Little, M.D., Huang, Y., 2000. Bioadhesive controlled release systems. In: Wise, D.L. (Ed.), Handbook of Pharmaceutical Controlled Release Technology, Marcel Dekker, New York, 263-266.

Pittler, M.H., Abbot, N.C., Harkness, E.F., Ernst, E., 1999. Randomized, double-blind trial of chitosan for body weight reduction. Eur. J. Clin. Nutr. 53, 379-381.

Portero, A., Remuñan-López, C., Vila-Jato, J.L., 1998. Effect of chitosan and chitosan glutamate enhancing the dissolution properties of the poorly water soluble drug nifedipine. Int. J. Pharm. 175, 75-84.

Qin, Y., Agboh, O.C., 1998. Chitin and chitosan fibres. Med. Device Technol. 11, 24-28.

Rao, S.P., Sharma, C.P., 1997. Use of chitosan as biomaterial. Studies on its safety and haemostatic potential. J. Biomed. Mater. Res. 34, 21-28.

Rege, P.R., Shukla, D.J., Block, L.H., 1999. Chitinosans as tabletting excipients for modified release delivery systems. Int. J. Pharm. 181, 49-60.

Remuñan-López, C., Bodmeier, R., 1997. Mechanical, water uptake and permeability properties of crosslinked chitosan glutamate and alginate films. J. Controlled Release 44, 215-225.

Remuñan-López, C., Portero, A., Vila-Jato, J.L., Alonso, M.J., 1998. Design and evaluation of chitosan/ethylcellulose mucoadhesive bilayered devices for buccal drug delivery. J. Controlled Release 55, 143-152.

Remuñan-López, C., Portero, A., Lemos, M., Vila-Jato, J.L., Nuñez, M.J., Riveiro, P., López, J.M., Piso, M., Alonso, M.J., 2000. Chitosan microspheres for the specific delivery of amoxycillin to the gastric cavity. S.T.P. Pharma. Sci. 10, 69-76.

Rhazi, M., Desbriéres, J., Tolamaite, A., Alagui, A., Vottero, P., 2000. Investigation of different natural sources of chitin: influence of the source and deacetylation process on the physicochemical characteristics of chitosan. Polym. Int. 49, 337-344.

Ritger, P.L., Peppas, N.A., 1987a. 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. Controlled Release 5, 23-36.

Ritger, P.L., Peppas, N.A., 1987b. A simple equation for description of solute release. II: Fickian and anomalous release from swellable devices. J. Controlled Release 5, 37-42.

Ritthidej, G., Chomto, P., Pummangura, S., Menasveta, P., 1994. Chitin and chitosan as disintegrants in paracetamol tablets. Drug Dev. Ind. Pharm. 20, 2109-2134.

Roberts, G.A.F, 1992. Chitin Chemistry, The Mac Millan Press, London, 1-110, 274-315.

Sabnis, S., Rege, P., Block, L.H., 1997. Use of chitosan in compressed tablets of diclofenac sodium: inhibition of drug release in an acidic environment. Pharm. Dev. Technol. 2, 243-255.

54

Sabnis, S., Block, L.H., 2000. Chitosan as an enabling excipient for drug delivery systems. I. Molecular modifications. Int. J. Biol. Macromol. 27, 181-186.

Sawayanagi, Y., Nambu, N., Nagai, T., 1982a. Directly compressed tablets containing chitin or chitosan in addition to lactose or potato starch. Chem. Pharm. Bull. 30, 2935-2940.

Sawayanagi, Y., Nambu, N., Nagai, T., 1982b. Directly compressed tablets containing chitin or chitosan in addition to mannitol. Chem. Pharm. Bull. 30, 4216-4218.

Senel, S., Ikinci, G., Kas, S., Yousefi-Rad, A., Sargon, M.F., Hincal, A.A., 2000. Chitosan films and hydrogels of chlorhexidine gluconate for oral mucosal delivery. Int. J. Pharm. 193, 197-203.

Shah, S., Qaqishi, R., Patel, V., Amiji, M., 1999. Evaluation of the factors influencing stomach-specific delivery of antibacterial agents for Helicobacter pylori infection. J. Pharm. Pharmacol. 51, 667-672.

Shah, V.P., Midha, K.K., Findlay, J.W.A., Hill, H.M., Hulse, J.D., McGilveray, I.J., McKay, G., Miller, K.J., Patnaik, R.N., Powell, M.L., Tonelli, A., Viswanathan, C.T., Yacobi, A., 2000. Bioanalytical method validation – a revisit with a decade of progress. Pharm. Res. 17, 1551-1557.

Shahidi, F., Arachchi, J.K.V., Jeon, Y.J., 1999. Food applications of chitin and chitosans. Trends Food Sci. Technol. 10, 37-51.

Shepherd, R., Reader, S., Falshaw, A., 1997. Chitosan functional properties. Glycoconjugate J. 14, 535-542.

Shimoda, J., Onishi, H., Machida, Y., 2001. Bioadhesive characteristics of chitosan microspheres to the mucosa of rat small intestine. Drug Dev. Ind. Pharm. 27, 567-576.

Shipper, N.G.M., Vårum, K.M., Artursson, P., 1996. Chitosans as absorption enhancers for poorly absorbable drugs 1: Influence of molecular weight and degree of acetylation on drug transport across human intestinal epithelial (Caco-2) cells. Pharm. Res. 13, 1686-1692.

Shipper, N.G.M., Olsson, S., Hoogstraate, J.A., Deboer, A.G., Vårum, K.M., Artursson, P., 1997. Chitosans as absorption enhancers for poorly absorbable drugs 2: Mechanism of absorption enhancement. Pharm. Res. 14, 923-929.

Shipper, N.G.M., Vårum, K.M., Stenberg, P., Ocklind, G., Lennernäs, H., Artursson, P., 1999. Chitosans as absorption enhancers for poorly absorbable drugs 3: Influence of mucus on absorption enhancement. Eur. J. Pharm. Sci. 8, 335-343.

Shiraishi, S., Arahira, M., Imai, T., Otagri, M., 1990. Enhancement of dissolution rates of several drugs by low-molecular chitosan and alginate. Chem. Pharm. Bull. 38, 185-187.

Shiraishi, S., Imai, T., Otagiri, M., 1993. Controlled release of indomethacin by chitosan-polyelectrolyte complex: optimization and in vivo/in vitro evaluation. J. Controlled Release 25, 217-225.

Singla, A.K., Chawla, M., 2001. Chitosan: some pharmaceutical and biological aspects – an update. J. Pharm. Pharmacol. 53, 1047-1067.

Snyman, D., Hamman, J.H., Kotzé, A.F., 2003. Evaluation of the mucoadhesive properties of N-trimethylchitosan chloride. Drug Dev. Ind. Pharm. 29, 61-69.

Song, Y., Onishi, H., Machida, Y., Nagai, T., 1996. Drug release and antitumor characteristics of N-succinyl-chitosan-mitomycin C as an implant. J. Controlled Release 42, 93-100.

55

Staib, A.H., Woodcock, B.G., Loew, D., 1989. Remote control of gastrointestinal drug delivery in man. In: Prescott, L.F., Nimmo, W.S. (Eds.), Novel Drug Delivery and Its Therapeutic Application, Wiley, London, 79-88.

Struszczyk, H., 1987. Microcrystalline chitosan. I: Preparation and properties of microcrystalline chitosan. J. Appl. Polym. Sci. 33, 177-189.

Struszczyk, H., Kivekäs, O., 1992. Recent developments in microcrystalline chitosan applications. In: Brine, C.J., Sandford, P.A., Zikakis J.P., (Eds.), Advances in Chitin and Chitosan, Elsevier Applied Science, London, 549-555.

Takeuchi, H., Yamamoto, H., Niwa, T., Hino, T., Kawashima, Y., 1996. Enteral absorption of insulin in rats from mucoadhesive chitosan-coated liposomes. Pharm. Res. 13, 896-901.

Tapia, C., Buckton, G., Newton, J.M., 1993. Factors influencing the mechanism of release from sustained release matrix pellets, produced by extrusion/spheronisation. Int. J. Pharm. 92, 211-218.

Tapia, C., Costa, E., Moris, M., Sapag-Hagar, J., Valenzuela, F., Basualto, C., 2002. Study of the influence of the pH media dissolution, degree of polymerization, and degree of swelling of the polymers on the mechanism of release of diltiazem from matrices based on mixtures of chitosan/alginate. Drug Dev. Ind. Pharm. 28, 217-224.

Tengamnuay, P., Sahamethpat, A., Sailasuta, A., Mitra, A.K., 2000. Chitosans as nasal absorption enhancers of peptides: comparison between free amine chitosans and soluble salts. Int. J. Pharm. 197, 53-67.

Thanoo, C., Sunny, M.C., Jayakrishnan, A., 1992. Cross-linked chitosan microspheres: preparation and evaluation as a matrix for the controlled release of pharmaceuticals. J. Pharm. Pharmacol. 44, 283-286.

Thanou, M.M., Verhoef, J.C., Romeijn, S.G., Nagelkerke, J.F., Merkus, F.W., Junginger, H.E., 1999. Effects of N-trimethyl chitosan chloride, a novel absorption enhancer, on Caco-2 intestinal epithelia and the ciliary beat frequency of chicken embryo trachea. Int. J. Pharm. 185, 73-82.

Tozaki, H., Komoike, J., Tada, C., Maruyama, T., Terabe, A., Suzuki., T., Yamamoto, A., Muranishi, S., 1997. Chitosan capsules for colon-specific drug delivery: Improvement of insulin absorption from the rat colon. J. Pharm. Sci. 86, 1016-1021.

Tozaki, H., Fujita, T., Odoriba, T., Terabe, A., Suzuki, T., Tanaka, C., Okabe, S., Muranishi, S., Yamamoto, A., 1999. Colon-specific delivery of R68070, a new thromboxane synthase inhibitor, using chitosan capsules: therapeutic effects against 2,4,6-trinitrobenzenesulphonic acid-induced colitis in rats. Life Sci. 64, 1155-1162.

Tozaki, H., Odoriba, T., Okada, N., Fujita, T., Terabe, A., Suzuki, T., Okabe, S., Muranishi, S., Yamamoto, A., 2002. Chitosan capsules for colon-specific drug delivery: enhanced localization of 5-aminosalisylic acid in the large intestine accelerates healing of TNBS-induced colitis in rats. J. Controlled Release 82, 51-61.

Tsaih, M., Chen, R., 1999. Effects of ionic strength and pH on the diffusion coefficients and conformation of chitosans in molecule solution. J. Appl. Polym. Sci. 73, 2041-2050.

United States Pat. US5620706, 1997. Polyionic insoluble hydrogels comprising xanthan and chitosan. Dumitriu, S., Chornet, E., Vidal, P.

United States Pat. US5904927, 1999. Drug delivery using pH-sensitive semi-interpenetrating network hydrogels. Amiji, M.

56

Upadrashta, S.M., Katikaneni, P.R., Nuessle, N.O., 1992. Chitosan as a tablet binder. Drug Dev. Ind. Pharm. 18, 1701-1708.

Van der Lubben, I.M., Verhoef, J.C., Van Aelst, A., Brochard, G., Junginger, H.E., 2001a. Chitosan microparticles for oral vaccination: preparation, characterization and preliminary in vivo uptake studies in murine Peyer’s patches. Biomaterials 22, 687-694.

Van der Lubben, I.M., Konings, F.A.J., Brochard, G., Verhoef, J.C., Junginger, H.E., 2001b. In vivo uptake of chitosan microparticles by murine Peyer’s patches: visualization studies using confocal laser scanning microscopy and immunohistochemistry. J. Drug Targeting 939-947.

Waaler, T., Sande, S.A., Müller, B.W., Schüller-Lisether, G., 1997. The influence of thermal neutron irradiation on the in vitro characteristics of ASA oral dosage forms. Validation of neutron activation. Eur. J. Pharm. Biopharm. 43, 159-164.

Wilding, I.R., Coupe, A.J., Davis, S.S., 2001. The role of γ-scintigraphy in oral drug delivery. Adv. Drug Delivery Rev. 46, 103-124.

Wilson, C.G., Washington, N., Norman, S., Greaves, J.L, Peach, J.M., Pugh, K., 1988. A gamma scintigraphic study to compare oesophageal clearance of “Expidet” formulations, tablets and capsules in supine volunteers. Int. J. Pharm. 46, 241-246.

Wilson, C.G., Washington, N., Greaves, J.L., Kamali, F., Rees, J.A., Sempik, A.K., Lambard, J.F., 1989. Bimodal release of ibuprofen in a sustained-release formulation: a scintigraphic and pharmacokinetic open study in healthy volunteers under different conditions of food intake. Int. J. Pharm. 50, 155-161.

Wilson, C.G., 1998. In vivo monitoring of dosage forms. J. Pharm. Pharmacol. 50, 383-386.

Wuolijoki, E., Hirvelä, T., Ylitalo, P., 1999. Decrease in serum LDL cholesterol with microcrystalline chitosan. Methods Find. Exp. Clin. Pharmacol. 21, 357-361.

Yamamoto, H., Takeuchi, H., Hino, T., Kawashima, Y., 2000. Mucoadhesive liposomes: physicochemical properties and release behaviour of water-soluble drugs from chitosan coated liposomes. S.T.P. Pharma. Sci. 10, 63-68.

Yao, K-D., Liu, J., Cheng, G.X., Zhao, R.Z., Wang, W., Wei, L., 1998. The dynamic swelling behaviour of chitosan based hydrogels. Polym. Int. 45, 191-194.

Ylitalo, R., Lehtinen, S., Wuolijoki, E., Ylitalo, P., Lehtimäki, T., 2002. Cholesterol-lowering properties and safety of chitosan. Arzneim.–Forsch. 52, 1-7.

Zhang, H., Neau, S.H., 2002. In vitro degradation of chitosan by bacterial enzymes from rat cecal and colonic contents. Biomaterials 23, 2761-2766.