The release of model macromolecules may be controlled by the hydrophobicity of palmitoyl glycol chitosan hydrogels

Journal of Controlled Release 80 (2002) 87–100www.elsevier.com/ locate / jconrel

The release of model macromolecules may be controlled by thehydrophobicity of palmitoyl glycol chitosan hydrogels

a a b cLee Martin , Clive G. Wilson , Fariba Koosha , Laurence Tetley ,a d a ,*Alexander I. Gray , Sevda Senel , Ijeoma F. Uchegbu

aDepartment of Pharmaceutical Sciences, Strathclyde Institute for Biological Sciences, University of Strathclyde, 27 Taylor St.,Glasgow G4 0NR, UK

bGlaxoSmithKline Pharmaceuticals, New Frontiers Science Park (South), Third Avenue, Harlow, Essex CM19 5AW, UKcElectron Microscopy Unit, IBLS, University of Glasgow, Glasgow G12 8QQ, UK

dFaculty of Pharmacy, Hacettepe University, 06100 Ankara, Turkey

Received 15 October 2001; accepted 20 December 2001

Abstract

A non-covalently cross-linked palmitoyl glycol chitosan (GCP) hydrogel has been evaluated as an erodible controlledrelease system for the delivery of hydrophilic macromolecules. Samples of GCP with hydrophobicity decreasing in the order

1GCP12.GCP11.GCP21 were synthesised and characterised by H NMR. Hydrogels were prepared by freeze-drying anaqueous dispersion of the polymer in the presence or absence of either a model macromolecule fluorescein isothiocyanate-dextran (FITC-dextran, MW 4400), and/or amphiphilic derivatives Gelucire 50/13 or vitamin E d-a-tocopherol poly-ethylene glycol succinate. Gels were analysed for aqueous hydration, FITC-dextran release, and bioadhesion, and imaged byscanning electron microscopy. The gels were highly porous and could be hydrated to up to 953 their original weight withoutan appreciable volume change and most gels eventually eroded. Hydration and erosion were governed by the hydrophobicityof the gel and the presence of the amphiphilic additives. GCP gels could be loaded with up to 27.5% (w/w) of FITC-dextranby freeze-drying a dispersion of GCP in a solution of FITC-dextran. The controlled release of FITC-dextran was governedby the hydrophobicity of the gel following the trend GCP21.GCP11.GCP12. GCP gels were bioadhesive but less so thanhydroxypropylmethylcellulose, Carbopol 974NF (7:3) tablets. 2002 Elsevier Science B.V. All rights reserved.

Keywords: Controlled release; Drug delivery; Hydrogel; Palmitoyl glycol chitosan; Macromolecules; Fluorescein isothiocyanate-dextran(FITC-dextran)

1. Introduction glutaraldehyde [1–4], glyoxal [5], 2,2-dimethoxy-2-phenylacetophenone [6], tripolyphosphate penta-

Chitosan based gels have been prepared by chemi- sodium salt [7], diacetaldehyde PEG followed bycal cross-linking of the polymer chains using either reduction with sodium cyanoborohydride [8], scleral-

dehyde (Schiff base formation) [9] or diethylsquarate[10]. Amide bond formation between acid moieties*Corresponding author. Tel.: 144-141-548-3895; fax: 144-on either poly(4-N-methacrylamidobenzoic acid)141-552-6443.

E-mail address: [email protected] (I.F. Uchegbu). [11] or poly(acrylic acid) [12] and the chitosan

0168-3659/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0168-3659( 02 )00005-6

88 L. Martin et al. / Journal of Controlled Release 80 (2002) 87 –100

amino groups is also a method of achieving chemi- Here we examine the effect of varying the hydro-cally cross-linked interpenetrating polymer network phobicity of the gel material on gel hydration, gel(IPN) gels. Additionally semi-IPN gels consisting of erosion, gel bioadhesion and the release of entrappedchemically cross-linked chitosan and polymers such macromolecules. The current communication showsas poly(oxypropylene glycol) [2], silk fibroin [4], that GCP gels slowly erode, may be loaded to a highpoly(acrylic acid) [13] and polyethylene glycol [14], concentration with the model macromolecule fluores-have been prepared. cein isothiocyanate dextran (FITC-dextran), control

The formation of hydrogels from polymers using the release of the model macromolecule and arenon-covalent means is a useful method of preparing bioadhesive. Hydrophobicity has been found tohydrogels for drug delivery since these gels are influence gel hydration /erosion and the release of alikely to be more biocompatible as gel formation model hydrophilic macromolecule. Since this gel isdoes not require the use of organic solvents or designed for controlled buccal delivery, the desirablechemical reactions which may be potentially deleteri- attributes will be a high drug loading capacity, theous to the drug load. The need to deal with residual ability to control the release of drugs, the ability tolevels of organic solvent within the gels is thus adhere to the buccal mucosa and the eventual erosionremoved. Such physically cross-linked chitosan that will avoid the need to remove the drug deliverybased gels are formed by exploiting either hydrogen device after the dose has been delivered.bonding or hydrophobic attractions. As such, The buccal route is a promising route for thechitosan based gels have been prepared by grafting delivery of drugs which either undergo extensivelow molecular weight moieties such as acetyl [15], hepatic first pass metabolism or which are degradedethylenediamine tetracetic acid [16] or palmitoyl [17] in the gastrointestinal tract such as peptides [26].units or the use of graft copolymers of chitosan This route is well vascularised with venous bloodbearing groups such as methoxypolyethylene glycol draining the buccal mucosa reaching the heart direct-[8], poly(lactic acid) [18] and poly(N-iso- ly via the internal jugular vein. Since polycationsproylacrylamide) [19]. These latter high molecular such as chitosan have been reported to increaseweight grafts destroy the crystallinity of chitosan and epithelial permeability by enhancing paracellularlead to an irregular packing of the polymer chains transport [27,28] and chitosan in particular has been[18,19] in an amorphous network. Physical gels held used to enhance the absorption of insulin via thetogether by non-covalent forces may also be pre- nasal route [29], chitosan derivatives are ideal candi-pared by exploiting polyelectrolyte complexation as date materials for use in the fabrication of buccalin the case of complexes formed between the amino drug delivery dosage forms.groups of chitosan and the carboxylic acid groups ofcarboxymethylcellulose [20]. Additionally chitosanblended with poly(vinylalcohol) [1,21], xylan [22] or 2. Materials and methodspoly(N-isopropylacrylamide) [19] also results in theformation of physically cross-linked hydrogels and 2.1. Materialsonce again hydrogen bonding is the main contributorto gelation. Physically cross-linked gels have also All materials were used as received. Glycolbeen prepared with other polymers, for example the chitosan (GC; MW 164 kDa), palmitic acid N-hy-combination of a dextran-L-lactic acid and a dextran- droxysuccinimide (PNS), FITC-dextran (MW 4.4D-lactic acid conjugate gives rise to a gel termed a kDa) and phosphate buffered saline (PBS) tabletsstereocomplex [23]. were all purchased from Sigma Aldrich (UK).

Our work has focused on the use of pendant Sodium bicarbonate was obtained from Fluka (UK).hydrophobic groups to achieve non-covalent cross- Carbopol 974NF (CP) was obtained from BF Good-linking [17]. This communication deals with further rich (USA). Gelucire 50/13 (a mixture of mono-, di-developments of this type of gel, which is formed and triglycerides and mono- and di-fatty C16 andfrom a variant of the amphiphilic vesicle forming C18 acid esters of polyethylene glycol, MW 1500),polymer palmitoyl glycol chitosan (GCP) [24,25]. vitamin E d-a-tocopherol polyethylene glycol succi-

L. Martin et al. / Journal of Controlled Release 80 (2002) 87 –100 89

Scheme 1. Synthesis of palmitoyl glycol chitosan.

nate (TPGS) and hydroxypropylmethyl cellulose the reaction mixture and the resulting solution evapo-(HPMC) were all kind gifts from Gattefosse rated under reduced pressure at 45 8C to remove(France), Eastman (UK) and Dow Chemical Com- organic solvents. The resulting mixture was extractedpany (USA), respectively. Fresh excised porcine with 33 diethyl ether (100 ml) and subjected tobuccal mucosa was obtained from the local abattoir. exhaustive dialysis against water (5 l) for 24 h withAbsolute ethanol, acetone and diethyl ether were all six changes. The polymer dispersion was sub-purchased from Merck (UK). sequently freeze-dried (see Section 2.2.2 below).

1Analysis of the polymer was carried out by H2.2. Methods NMR [17] with the gel dissolved in deuterated

1dimethylsulphoxide and the H NMR of the parent2.2.1. Synthesis and characterisation of GCP GC molecule (dissolved in deuterated water) used as

Synthesis of the polymer was carried out as shown a standard. Briefly the degree of hydrophobic modi-in Scheme 1. Table 1 lists the ratios of reactants used fication was calculated by comparing the level offor each of the GCP samples. The method given palmitoyl methyl protons (0.86 ppm) in the GCPbelow is given for GCP11. GC (250 mg) was spectrum with the level of sugar protons (3.28–4.15dissolved in water (40 ml) to which was added ppm) in the GC spectrum using the acetyl protonssodium bicarbonate (190 mg) and absolute ethanol (1.83–2.00 ppm) in both spectra for standardization.(25 ml). A solution of PNS (396 mg) in absoluteethanol (300 ml) was added drop-wise to the alkaline 2.2.2. Preparation of GCP gelssolution of GC over a 1-h period. After 72-h stirring,protected from light, acetone (50 ml) was added to 2.2.2.1. Plain GCP gels. The purified polymer dis-

Table 11Level of hydrophobic modification in various GCP samples as determined by H NMR

GCP gel Glycol Palmitic acid Level of palmitoylationformulation chitosan (g) N-hydroxysuccinimide in mol.% of

(g) sugar monomers

GCP21 0.500 0.396 4.2062.62GCP11 0.250 0.396 10.0167.32GCP12 0.250 0.792 20.3162.22


persion obtained from above was frozen at 220 8C, each data point was generated by a different pre-freeze-dried in either 96-well plates to yield dried weighed gel sample.non-covalently cross-linked cylindrical gels (9 mmin length and 6 mm in diameter) or freeze-dried in 2.2.3.2. FITC-dextran loading. The amount of FIT-7-ml bijou bottles to yield discs which were manual- C-dextran loaded onto the gels was determined byly compressed to 15 mm in diameter and 2 mm probe sonicating a weighed piece of FITC-dextranthick. Disc shaped gels were used for the bioadhe- loaded gel in the presence of PBS (pH 7.4, 5 ml).sion experiments while cylindrical gels were used for This dispersion was then diluted with PBS (pH 7.4,all other tests. 15 ml). A 1-ml sample of this diluted dispersion was

then further diluted with PBS (pH 7.4, 4 ml) and2.2.2.2. Gelucire 50/13 or TPGS containing gels. centrifuged (30003g for 20 min). The supernatantDried GCP12 gels (68.4 mg) from above and either (3 ml) was separated, diluted with PBS (pH 7.4, 5TPGS or Gelucire 50/13 (3.6 mg) were probe ml) and analysed fluorimetrically (exc. 495 nm, em.sonicated (Lucas Davis Ultrasonics, UK) in water (2 525 nm, Perkin Elmer LS-50B). The concentrationml) and freeze-dried as described above to give of FITC-dextran in these samples was determinedGCP12 gels containing either Gelucire 50/13 or with respect to an FITC-dextran standard curveTPGS at a level of 5% (w/w). prepared within the concentration range of 0.001–

210.05 mg ml . Each determination was performed a2.2.2.3. FITC-dextran loaded gels. Dried GCP12 minimum of three times.gels (40 mg) obtained from above were dispersed inFITC-dextran (4 ml) solutions at FITC-dextran con- 2.2.3.3. Release of FITC-dextran. Pre-weighed FIT-

21centration levels of 0.2, 0.4, 2.0 and 4.0 mg ml . C-dextran loaded gels were placed in PBS (pH 7.4,After probe sonication, the resulting solutions were 20 ml) in stoppered bottles that were in turn placedfreeze-dried as described above to give GCP12 gels in a shaking water bath with the temperature set atloaded with varying concentrations of FITC-dextran. 35 8C. At pre-determined time points, aliquots of theAll other formulations were loaded with FITC-dex- release medium (1 ml) were collected and replaced

21tran at a concentration of 2.0 mg ml using the by the same volume of fresh PBS (pH 7.4). Thesesame procedure as described above. 1-ml samples were then diluted to 5 ml with PBS

(pH 7.4) and centrifuged (30003g for 20 min). Then3 ml of the resulting supernatant was removed and

2.2.3. Characterisation of the GCP gels diluted to 8 ml with PBS (pH 7.4) and the resultingsolutions assayed fluorimetrically as describedabove.

2.2.3.1. Hydration and erosion studies. Dried cylin-drical gels (9 mm in length36 mm in diameter) were 2.2.3.4. Bioadhesion studies. Disc shaped gels, 15weighed and immersed in water (25 ml) contained mm in diameter and 2 mm in height, were manuallywithin stoppered bottles that were in turn placed in a attached to the texture analyser probe (Stable Microshaking water bath set at 35 8C. At various time Systems TA-XT2 Texture Analyser). Control tabletspoints gels were removed, drained and their weights consisting of HPMC, CP (7:3) were prepared. Ex-

Wt]recorded. Swelling ratio was recorded as Q 5 , cipients were weighed to give the desired com-W0

where W 5the weight of the dried hydrogel and position and the powders blended for 30 min in a0

W 5the weight of the hydrated hydrogel at time t. Turbula Mixer prior to compression. After blending,t

Experiments were performed a minimum of three the resulting powder was directly compressed at atimes. weight of 100 mg and at a diameter of 7 mm, using a

The gradual erosion of GCP gels was also studied single-punch tablet press (Model E2, Manesty).by observing the time at which W began to decrease. Control tablets were fully characterised for weight,t

Experiments were performed a minimum of three diameter, thickness and hardness (Erweka TGB30times and to minimise the effects of gel handling, tablet tester). Tablets (n55) had an average weight


of 95 mg and were 15 mm in diameter, 3 mm thickand with a mean hardness of 6 kp.

Freshly excised porcine buccal mucosa was ob-tained from male / female pigs from the local slaug-hterhouse. Pigs had an average weight of 65 kg andhad a mean age6S.D. of 6.560.5 months at the timeof death. The animals were killed by electroshockfollowed by exsanguination. The tissue was thenremoved and placed in PBS solution (pH 7.4, 300ml) and used within 2 h of the animal’s death.

The force of detachment of the gels and tabletsfrom a piece of porcine mucosa was recorded bylowering tablets or gels (affixed to the instrument’s

21Perspex support) at a speed of 10 mm s untilcontact with the tissue was achieved. A predeter-mined contact force of 0.02 N and contact time of 60s was used. The probe was then raised (in an attemptto detach the gel / tablet from the mucosa) to adistance of 15 mm and the force of detachment

Fig. 1. GCP12, TPGS (95:5) gel, bar represents 6 mm.recorded as a measure of the bioadhesiveness of theformulations.

shaped gels, 15 mm in diameter and 2 mm thick, was2.2.3.5. Scanning electron microscopy. Gel formula- 15 mg, showing that the material is very light.tions were fast frozen in liquid propane and freeze- Analysis revealed a diminishing hydrophobicity fromdried on cover slips at 280 8C overnight under a GCP12.GCP11.GCP21 (Table 1). The large stan-

26vacuum of 10 Torr. Dried specimens were gold- dard deviation observed in the GCP11 batches didcoated on a Peltier-cold stage sputter-coater and not prevent the observation of hydrophobicity depen-examined using a Phillips 500 Scanning Electron dent effects in the rest of this study and reflectMicroscope at 3 or 6 kV accelerating voltage. 1 K31 variations between polymer batches. Materials withK images of the surface and cross-sectional morphol- distinctly different levels of palmitoylation were usedogy of the various freeze-dried GCP gel formulations for the comparative tests on GCP21, GCP11 andwere captured using a digital interface, and recorded GCP12.onto CD.

3.2. Physical cross-linking2.2.3.6. Statistics. ANOVA tests were performed ondata sets using the Mintab 7.1 software package. In the current study a physically cross-linkedStatistical significance was set at P,0.05. hydrogel [17] has been prepared by freeze-drying a

dispersion of the amphiphilic polymer palmitoylglycol chitosan [24,25]. Cross-linking is achieved by

3. Results and discussion the hydrophobic interactions between pendant pal-mitoyl groups (Fig. 2) which together with the

3.1. GCP synthesis polymer molecular weight (164 kDa) allow thepolymer to hydrate without solubilising. Dehydration

The freeze-drying of palmitoyl glycol chitosan (freeze-drying) of the polymer dispersion is thesamples results in the formation of a spongy material initiator for cross-linking. It is likely that inter-chain(Fig. 1). The average weight of the cylindrical gels, hydrogen bonding between sugar units may also6 mm in diameter and 9 mm in length, shown in Fig. contribute to the gel structure (Fig. 2). This method1, was 2 mg and the average weight of the disk of preparing hydrogels from a hydrophilic com-


3.4. Gel hydration and erosion

This porous structure leads to a high hydrationcapacity with the gels swelling to up to 95 timestheir original weight (Fig. 4). Porous highly swellingstructures are also formed from freeze-driedchitosan-xylan gels (which are cross-linked by anassociation of the glucoronic acid (xylan) and amino(chitosan) functionalities) [22] and freeze-driedchitosan-poly(ethylene glycol) semi-IPN gels [5].Freeze-dried gels are known to hydrate to a higherextent than air-dried gels prepared from the samepolymeric compounds [5] as the high free volume inFig. 2. Schematic representation of polymer chain cross-linking tothese gels allows for the accumulation of largeproduce the erodible hydrogel.volumes of water.

The gels transform from an opaque solid to aponent and a water insoluble component has been translucent mass on hydration and do not suffer anused previously in a number of studies (see review appreciable change in volume or shape in the earlyby Kamath and Park [30]). The water insolubility is hydration phase. On hydration, all GCP gels gradual-usually achieved by covalent cross-linking [1–6,8– ly erode with time due to the hydration of the12,31] which ultimately increases molecular weight. hydrophilic parts of the gel and the hydrostaticPhysical cross-linking as is used with the current pressure causing the hydrophobic contacts to dimin-hydrogels allows the hydrogel to gradually erode ish in strength. The time of discernable erosion onsetpreferably after delivery of its payload. This latter of the hydrated gels increases as the level ofproperty would negate the need for removal of the hydrophobicity of GCP increases following the trenddelivery system after use. GCP21,GCP11,GCP12 (Table 2), evidence that

an increase in hydrophobic cross-links increases the3.3. Gel structure resistance of the gel to erosion. It is likely that

erosion begins before the times shown in Table 2 butThe macroscopic (Fig. 1) and microscopic (Fig. 3) cannot be ascertained by the methods used here

appearance of the gel is of a porous structure by because both erosion and hydration are dynamic andvirtue of the freeze-drying step [17] with the pores competing processes with opposing effects on thebeing the result of ice crystal formation. GCP gel weight of the gel.microstructure is influenced by gel hydrophobicity The erosion of the gel in turn impacts on the gelwith the average pore size decreasing as gel hydro- swelling data with the fast eroding more hydrophilicphobicity increases (Fig. 3a–c). The more hydrogel GCP21 revealing a diminished swelling due tophilic gels trap larger ice crystals as expected since a the influence of erosion which continues during thehigher level of hydrophilicity would result in more hydration experiments but is only discernable underwater being bound to the polymer prior to freeze- our own experimental conditions when there is a lossdrying. This increased association with water with in gel weight. This interplay between erosion (beingincreasing hydrophilicity (decreasing hydrophobicity highest for GCP21) and gel hydration (being lowestand hydrophobic cross-links) would increase pore for GCP12) explains the hydration trend observed,size. Similarly cross-linking chitosan with glutaral- i.e. GCP11.GCP12.GCP21 (Fig. 4a), and GCP11dehyde led to a decrease in pore size of chitosan gels is seen to swell to 95% of its dry weight.prepared by the solvent evaporation technique [32]. As the current drug delivery system is beingThe incorporation of both Gelucire 50/13 and TPGS developed for buccal use, amphiphilic compoundsincreased the pore size of GCP12 gels and abolished were added in order to enhance drug penetrationthe mesh-like structure (Fig. 3d). through the squamous epithelium. Soluble detergents


Fig. 3. Scanning electron micrographs of GCP gels, bar represents 200 mm: (a) 100% GCP21; (b) 100% GCP11; (c) 100% GCP12; (d)GCP12, Gelucire 50/13 (95:5); (e) GCP12, FITC-dextran (86:14).


Fig. 4. (a) The swelling ratio (mean6S.D.) of GCP gels: m5100% GCP11; j5100% GCP12; n5100% GCP21; *significant difference[(P,0.05) between GCP11 gel and both GCP12 and GCP21 gels which were not significantly different from each other; significant

difference ( p,0.05) between GCP21 gel and both GCP11 and GCP21 gels, n5a minimum of 3 observations. (b) The swelling ratio(mean6S.D.) of GCP12 gels: j5100% GCP12; d55% (w/w) Gelucire 50/13; s55% (w/w) Vitamin E TPGS; *significant difference(P,0.05) between each of the gel formulations; n5a minimum of 3 observations.

such as glycodeoxycholate have been shown to phiphilic additives into GCP 12 gels results in anenhance the buccal bioavailability of FITC-dextran overall increase in average pore size (Fig. 5d) and(MW 4.4 kDa) in the pig model [33]. When additives the disappearance of the mesh-like structure. Itare introduced into the gel, the rate of gel swelling appears that despite the influence of Gelucire 50/13on hydration follows the trend GCP12, TPGS. on pore size, the hydration of Gelucire 50/13GCP12.GCP12, Gelucire 50/13 (Fig. 4b). The high containing gels is kinetically limited when comparedspeed of hydration shown on incorporation of TPGS to TPGS containing gels (Fig. 4b). Evidence for thisinto GCP12 gels must result from the good aqueous is provided by the diminished erosion and slowsolubility of this detergent. The opposite is true for swelling of Gelucire 50/13 containing gels (Fig. 4b,Gelucire 50/13. The incorporation of the am- Table 2) hence Gelucire 50/13 makes the gels more


Table 2Erosion of hydrated GCP gels (n53) and bioadhesion of GCP12 gels (n53)

Gel formulation Time of erosion Mean detachment force22onset (h) (N cm )6S.D.

[100% GCP21 4 0.06860.009100% GCP11 6 0.09960.015

†100% GCP12 48 0.07860.020GCP12, TPGS (95:5) 24 0.10360.033GCP12, Gelucire 50/13 .76 0.15360.037(95:5)GCP12, FITC-dextran .76 ND(86:14)GCP12, FITC-dextran .76 NDTPGS (85:10:5)GCP12, FITC-dextran, .76 NDGelucire 50/13 (80:15:5)HPMC, CP (70:30) ND 0.45260.112*

*Significant difference between HPMC, CP tablets and all other gels.†Significant difference between GCP12 and GCP12, Gelucire 50/13.[Significant difference between GCP21 and GCP11 (P,0.05).

Fig. 5. (a) The release of FITC-dextran (mean6S.D.) from GCP12 hydrogels: s51.9% (w/w) FITC-dextran; d54.4% (w/w) FITC-dextran; h513% (w/w) FITC-dextran; j527.5% (w/w) FITC-dextran; *significant difference (P,0.05) between 13% (w/w) and 1.9%

[(w/w) samples only; significant difference (P,0.05) between 27.5% (w/w) and 13% (w/w) samples only; n5a minimum of 3observations. (b) The release of FITC-dextran (mean6S.D.) from GCP gels: n5GCP2119.0% (w/w) FITC-dextran; m5GCP11110.0%(w/w) FITC-dextran; d5GCP12114.0% (w/w) FITC-dextran; *significant differences (P,0.05) in release rates between each of the gelformulations; n5a minimum of 3 observations. (c) The release of FITC-dextran (mean6S.D.) from GCP12 gels: j514.0% (w/w)FITC-dextran; s514.9% (w/w) FITC-dextran15% (w/w) Gelucire; h510.5% (w/w) FITC-dextran15% (w/w) TPGS; *significantdifference (P,0.05) between Gelucire 50/13 and TPGS loaded gels only. There was no significant difference between 100% GCP12 gelsand both Gelucire 50/13 and TPGS loaded gels; n5a minimum of 3 observations.


Fig. 5. (continued)

hydrophobic. The increase in pore size conferred on 3.5. FITC-dextran GCP gels: loading and releasethe gel by Gelucire 50/13 thus cannot be explainedby simply considering the effect of the hydrophil- The efficiency of loading of FITC-dextran on toicity /hydrophobicity of the amphiphilic mixture GCP12 gels was proportional to the concentration ofconstituting Gelucire 50/13. A further more complex FITC-dextran in the loading solution prior to freeze-

21interaction between Gelucire 50/13 and GCP12 must drying and a concentration range of 0.2–4.0 mg mlbe responsible for the enlarged pores seen in the in the loading solution yielded gels with a loading ofGCP12/Gelucire 50/13 gels 1.9227.5% (w/w) with a good correlation (r5

The inclusion of the model hydrophilic macro- 0.998) between the FITC-dextran concentration inmolecule FITC-dextran increased the erosion onset the loading solution and the final FITC-dextranvalue for the plain and TPGS containing GCP12 gels solute load on the gel. It is worth highlighting that(Table 2), thus this model macromolecule makes a GCP12 gels could be loaded with up to 27.5% (w/w)contribution to gel structure. of FITC-dextran because the loading of solutes on to


chitosan gels usually results in low solute levels, after 4 h of hydration (Table 2), released virtually alltypically between 0.1 [31] and 3% (w/w) [34]. The of its contents (9664.9%) within the first 30 minloading of chitosan hydrogels is usually accom- (Fig. 5b) with release preceding discernible erosion.plished by mixing the gel forming mixture with the Hydrogels act as drug delivery systems by phys-solute load prior to gelation, followed by the gelation ically entrapping bioactive molecules, which are thenstep [5,15,34] and gelation is usually achieved by slowly released. When the hydrogel comes intophysical [15] or chemical [5,31] cross-linking. In the contact with an aqueous solution, it begins to imbibepresent case gelation is achieved on freeze-drying a water. This water uptake leads to considerabledispersion of the polymer and this is done in the swelling of the polymer, dissolution of the drug, andpresence of a loading solute yielding a surprisingly the resultant diffusion of drug out of the polymer.efficient yet gentle method of loading. This form of Thus, drug release depends upon three rate pro-gel loading may be applicable for the loading of cesses; water diffusion into the polymer, chainlabile materials such as hydrophilic peptides and relaxation and drug dissolution. An increase in theproteins. hydrophobicity of the gel polymer will limit the first

FITC-dextran loading led to an increase in overall two processes.gel pore size (Fig. 3e) and diminished erosion (Table GCP12 was selected for further investigation due2). FITC-dextran is likely to be situated within to its ability to retard the release of FITC-dextran tohydrophilic pockets within the GCP gel and will the greatest extent over a 5-h time period. Thedecrease phase separation between the solid dis- addition of both the amphiphiles TPGS and Gelucirepersants and water, thereby increasing the level of 50/13 had no significant effect on the release ofice crystal formation during freezing, hence the FITC-dextran from GCP12 gels (Fig. 5c).increased pore size. The reduced level of erosion It must be noted that the complete release ofpoints to the fact that dextran although primarily FITC-dextran took much longer than the 5-h timeemployed as a model solute also gives mechanical period shown in Fig. 5 and release was initially faststrength to the gel. This fact is obvious from the followed by a slower release phase. After 24 h fortexture of the gel although the exact mechanism of example only 66.466.7% of FITC-dextran had beenincreasing gel mechanical strength is unclear. The released from the GCP11 gel.formulation thus behaves as a polymer blend. Chitosan is slowly degraded to glucosamine andChitosan based polymer blends have given rise to N-acetyl glucosamine in simulated gastric fluid USPhydrogels when chitosan was blended with a number XXII and simulated intestinal fluid USP XXII [36]of polymers such as polyvinyl alcohol [21,35] and and as such this agent could be applied as anpoly(N-isopropylacrylamide) [19]. erodible controlled drug delivery system via the

In order to examine whether the level of macro- buccal route. This is especially true as a highmolecule load influenced the release rate, GCP12 percentage of the drug is released from the deliverygels were loaded with various levels of FITC-dextran system during the 5-h period due to the porousand it was found that the level of macromolecule nature of the freeze-dried gel. Prolonged releaseload affected the release rate although no clear trend kinetics, which would necessitate prolonged resi-emerged and only GCP12 gels loaded with 13% dence times in the mouth for the buccal delivery(w/w) of FITC-dextran showed a different release devices, may not appeal to a large majority ofrate from the other FITC-dextran loaded gels (Fig. patients. Freeze-dried chitosan-polyethylene glycol5a). [5] and chitosan-PVP semi-IPNs [37], which are

The most notable observation of this study, when porous, also tend to release their contents faster thanthe FITC-dextran load was maintained within a air-dried gels.narrow range (9–14%, w/w), was that the release ofFITC-dextran by GCP gels was influenced by the 3.6. Gel bioadhesionhydrophobicity of the gels and followed the trendGCP21.GCP11.GCP12 (Fig. 5b). The fastest In order to evaluate the adhesion of the dry geleroding gel GCP21, which shows visible erosion formulations a CP, HPMC formulation was used as a


control. The GCP formulations are found to be polymers [39,41] and chitosan [39,41]. Both of thesebioadhesive but less bioadhesive than the CP, HPMC agents are bioadhesive although the Carbopol poly-control with the more hydrophobic gels (GCP11 and mers are more bioadhesive as also shown here.GCP12) being more bioadhesive than the less hydro-phobic gel (GCP21) and significantly so in the caseof GCP11 (Table 2). Formulation hydrophobicity has 4. Conclusionsbeen reported to improve bioadhesion [38]. Theinclusion of Gelucire 50/13 in GCP12 gels improved In conclusion, a physically cross-linked erodiblebioadhesion (Table 2). The force of attachment of and porous hydrogel has been prepared from GCP, athe dry gel to the surface of the mucosa will depend chitosan based amphiphile without the need foron the level of hydration of the formulation [39] (as organic solvents and chemical reactions potentiallydehydration of the mucosal surface drives mucoadhe- deleterious to the drug load. This hydrogel ission [38]), the possibility of hydrophobic interactions bioadhesive and may be loaded with high levels ofbetween the gel surface and the mucosal surface and the model macromolecule FITC-dextran and solutethe possibility of hydrogen bonding between the gel release controlled by the hydrophobicity of the gelformulation and the mucosal surface. When consider- polymer.ing the dehydration of the mucosal surface the factthat the GCP11 gel is a fast hydrating and minimallyeroding gel may account for its superior mucoadhe-

Acknowledgementssion when compared to GCP21, which is a fasthydrating and fast eroding gel. GCP11 gels would

L.M. Martin is in receipt of a BBSRC-GSK CASEthus dehydrate the mucous but retain their structureAward.while the dehydration of the mucous by GCP21 gels

would be accompanied by a loss of gel structure.This may explain why the trend observed in thebioadhesion data (Table 2) exactly resembles the Referencestrend observed with the hydration data (Fig. 2a). Anexplanation for the increased bioadhesiveness of [1] S. Nakatsuka, A.L. Andrady, Permeability of vitamin B12 in

chitosan membranes—effect of cross-linking with poly(vinylGCP12, Gelucire 50/13 gels when compared toalcohol) on permeability, J. Appl. Polym. Sci. 44 (1992)GCP12 gels is more difficult. However the erosion17–28.

and hydration data seem to suggest that Gelucire [2] K.D. Yao, T. Peng, M.F.A. Goosen, J.M. Min, Y.Y. He, pH50/13 (a mixture of polyethylene glycol triglyceride sensitivity of hydrogels based on complex forming chitosan-amphiphiles) diminishes gel erosion and hydration :polyether interpenetrating network, J. Appl. Polym. Sci. 48

(1993) 343–354.and makes the gel more hydrophobic (Table 2, Fig.[3] T. Peng, K.D. Yao, C. Yuan, M.F.A. Goosen, Structural-4b). It is possible that hydrophobic interactions

changes of pH-sensitive chitosan polyether hydrogels inbetween the gel and the mucosal surface are further different pH solution, J. Polym. Sci. A Polym. Chem. 32enhanced with the amphiphilic mixture constituting (1994) 591–596.Gelucire 50/13 gels. [4] X. Chen, W.J. Li, W. Zhong, Y.H. Lu, T.Y. Yu, pH sensitivity

and ion sensitivity of hydrogels based on complex formingFor buccal drug delivery, adhesion to the oralchitosan /silk fibroin interpenetrating polymer network, J.mucosa permits not only the intimacy of contact andAppl. Polym. Sci. 65 (1997) 2257–2262.

the possibility of improved drug absorption, but also [5] V.R. Patel, M.M. Amiji, Preparation and characterization ofthe ability to achieve an optimum residence time at freeze-dried chitosan-poly(ethylene oxide) hydrogels for site-the site of administration [40], thus ensuring suffi- specific antibiotic delivery in the stomach, Pharm. Res. 13

(1996) 588–593.cient time for the controlled release of the drug into[6] Y.M. Lee, S.S. Kim, S.H. Kim, Synthesis and properties ofthe systemic circulation via the buccal mucosa.

poly(ethylene glycol) macromer /beta-chitosan hydrogels, J.Diverse classes of polymers have been investigated Mater. Sci. Mater. Med. 8 (1997) 537–541.for their potential use as bioadhesive drug delivery [7] S. Senel, G. Ikinci, S. Kas, A. Yousefi-Rad, M.F. Sargon,systems including the Carbopol (polyacrylic acid) A.A. Hincal, Chitosan films and hydrogels of chlorhexidine


gluconate for oral mucosal delivery, Int. J. Pharm. 193 [23] S.J. de Jong, B. van Eerdenbrugh, C.F. van Nostrum, J.J.(2000) 197–203. Kettenes-van de Bosch, W.E. Hennink, Physically cross-

[8] M.D. Bentley, M.J. Roberts, J.M. Harris, Reductive amina- linked dextran hydrogels by stereocomplex formation oftion using poly(ethylene glycol) acetaldehyde hydrate gener- lactic acid oligomers: degradation and protein release be-ated in situ: applications to chitosan and lysozyme, J. Pharm. havior, J. Controlled Release 71 (2001) 261–275.

¨Sci. 87 (1998) 1446–1449. [24] I.F. Uchegbu, A.G. Schatzlein, L. Tetley, A.I. Gray, J.[9] B. Guo, A. Elgsaeter, B.E. Christensen, B.T. Stokke, Sludden, S. Siddique, E. Mosha, Polymeric chitosan-based

Sclerox-chitosan co-gels: effects of charge density on swell- vesicles for drug delivery, J. Pharm. Pharmacol. 50 (1998)ing of gels in ionic aqueous solution and in poor solvents, 453–458.and on the rehydration of dried gels, Polym. Gels Netw. 6 [25] W. Wang, A.M. McConaghy, L. Tetley, I.F. Uchegbu,(1998) 471–492. Controls on polymer molecular weight may be used to

[10] A.A. De-Angelis, D. Capitani, V. Crescenzi, Synthesis and control the size of palmitoyl glycol chitosan polymeric13C CP-MAS NMR characterization of a new chitosan-based vesicles, Langmuir 17 (2001) 631–636.polymeric network, Macromolecules 31 (1998) 1595–1601. [26] R.B. Gandhi, J.R. Robinson, Oral cavity as a site for

[11] C. Peniche, C. Elvira, J.S. Roman, Interpolymer complexes bioadhesive drug-delivery, Adv. Drug Deliv. Rev. 13 (1994)of chitosan and polymethacrylic acid derivatives of salicylic 43–74.acid: preparation, characterization and modification by ther- [27] G.T.A. McEwan, M.A. Jepson, B.H. Hirst, N.L. Simmons,mal treatment, Polymer 39 (1998) 6549–6554. Polycation induced enhancement of epithelial paracellulal

[12] J.W. Lee, S.Y. Kim, S.S. Kim, Y.M. Lee, K.H. Lee, S.J. Kim, permeability is independent of tight junctional characteris-Synthesis and characteristics of interpenetrating polymer tics, Biochim. Biophys. Acta 1148 (1993) 51–60.network hydrogel composed of chitosan and poly(acrylic [28] P. Artusson, T. Lindmark, S.S. Davis, L. Illum, Effects ofacid), J. Appl. Polym. Sci. 73 (1999) 113–120. chitosan on the permeability of monolayers of intestinal

[13] H.F. Wang, W.J. Li, Y.H. Lu, Z.L. Wang, Studies on chitosan epithelial cells (Caco-2), Pharm. Res. 11 (1994) 1358–1361.and poly(acrylic acid) interpolymer complex, 1. Preparation, [29] T.J. Aspden, L. Illum, O. Skaugrad, Chitosan as a nasalstructure, pH sensitivity and salt sensitivity of complex- delivery system: evaluation of insulin absorption enhance-forming poly(acrylic acid):chitosan semi-interpenetrating ment and effect of nasal membrane integrity using ratpolymer network, J. Appl. Polym. Sci. 65 (1997) 1445– models, Eur. J. Pharm. Biopharm. 4 (1996) 23–31.1450. [30] K.R. Kamath, K. Park, Biodegradable hydrogels in drug

[14] M.N. Khalid, L. Ho, F. Agnely, J.L. Grossiord, G. Couar- delivery, Adv. Drug Deliv. Rev. 11 (1993) 59–84.raze, Swelling properties and mechanical characterization of [31] S. Senel, G. Ikinci, S. Kas, A. Yousefi-Rad, M.F. Sargon,a semi-interpenetrating chitosan/polyethylene oxide net- A.A. Hincal, Chitosan films and hydrogels of chlorhexidinework—comparison with a chitosan reference gel, Stp Pharma gluconate for oral mucosal delivery, Int. J. Pharm. 193Sci. 9 (1999) 359–364. (2000) 197–203.

[15] J. Kristl, J. Smidkorbar, E. Struc, M. Schara, H. Rupprecht, [32] B. Krajewska, Diffusional properties of chitosan hydrogelHydrocolloids and gels of chitosan as drug carriers, Int. J. membranes, J. Chem. Technol. Biotechnol. 76 (2001) 636–Pharm. 99 (1993) 13–19. 642.

[16] C. Valenta, B. Christen, A. Bernkop-Schnurch, Chitosan– [33] A.J. Hoogstraate, J.C. Verhoef, B. Tuk, A. Pijpers,EDTA conjugate: a novel polymer for topical gels, J. Pharm. L.A.M.G.v. Leengoed, J.H.M. Verheijden, H.E. Bodde, Buc-Pharmacol. 50 (1998) 445–452. cal delivery of fluorescein isothiocyanate-dextran 4400 and

[17] L. Noble, A.I. Gray, L. Sadiq, I.F. Uchegbu, A non-co- the peptide drug buserelin with glycodeoxycholate as anvalently cross-linked chitosan based hydrogel, Int. J. Pharm. absorption enhancer in pigs, J. Controlled Release 41 (1996)192 (1999) 173–182. 77–84.

[18] X. Qu, A. Wirsen, A.C. Albertsson, Synthesis and characteri- [34] E. Ruel-Gariepy, A. Chenite, C. Chaput, S. Guirguis, J.C.zation of pH-sensitive hydrogels based on chitosan and Leroux, Characterization of thermosensitive chitosan gels forD,L-lactic acid, J. Appl. Polym. Sci. 74 (1999) 3193–3202. the sustained delivery of drugs, Int. J. Pharm. 203 (2000)

[19] S.Y. Kim, S.M. Cho, Y.M. Lee, S.J. Kim, Thermo- and 89–98.pH-responsive behaviours of graft copolymer and blend [35] T. Koyano, N. Koshizaki, H. Umehara, M. Nagura, N.based on chitosan and N-isopropyl acrylamide, J. Appl. Minoura, Surface states of PVA/chitosan blended hydrogels,Polym. Sci. 78 (2000) 1381–1391. Polymer 41 (2000) 4461–4465.

[20] D.D. Long, D. van Luyen, Chitosan-carboxymethylcellulose [36] F. Chellat, M. Tabrizian, S. Dumitriu, E. Chornet, C.H.hydrogels as supports for cell immobilization, J. Macromol. Rivard, L. Yahia, Study of biodegradation behavior ofSci. Pure Appl. Chem. A33 (1996) 1875–1884. chitosan-xanthan microspheres in simulated physiological

[21] M.G. Cascone, S. Maltinti, N. Barbani, M. Laus, Effect of media, J. Biomed. Mater. Res. 53 (2000) 592–599.chitosan and dextran on the properties of poly(vinyl alcohol) [37] M.V. Risbud, A.A. Hardikar, S.V. Bhat, R.R. Bhonde, pH-hydrogels, J. Mater. Sci. Mater. Med. 10 (1999) 431–435. sensitive freeze-dried chitosan-polyvinyl pyrrolidone hydro-

[22] I. Gabrielii, P. Gatenholm, Preparation and properties of gels as controlled release system for antibiotic delivery, J.hydrogels based on hemicellulose, J. Appl. Polym. Sci. 69 Controlled Release 68 (2000) 23–30.(1998) 1661–1667. [38] Y. Jacques, T. NguyenXuan, E. Ionescu, G.P. Ravelli, P.


Buri, P. Baehni, R. Gurny, In vivo evaluation of hydrophilic [40] K. Knuth, M. Amiji, J.R. Robinson, Hydrogel deliveryand hydrophobic mucoadhesive semi-solid formulations con- systems for vaginal and oral applications, Adv. Drug Deliv.taining sucralfate and lidocaine for intraoral use, Eur. J. Rev. 11 (1993) 137–167.Pharm. Biopharm. 43 (1997) 59–63. [41] I.G. Needleman, F.C. Smales, In-vitro assessment of

[39] C.M. Lehr, J.A. Bouwstra, E.H. Schacht, H.E. Junginger, In bioadhesion for periodontal and buccal drug-delivery,vitro evaluation of mucoadhesive properties of chitosan and Biomaterials 16 (1995) 617–624.some other natural polymers, Int. J. Pharm. 78 (1992) 43–48.

Documents

The release of model macromolecules may be controlled by the hydrophobicity of palmitoyl glycol chitosan hydrogels