Preparation and evaluation of a chitosan salt–poloxamer 407 based matrix for buccal drug delivery

Journal of Controlled Release 102 (2005) 159 – 169 www.elsevier.com/locate/jconrel

Preparation and evaluation of a chitosan salt–poloxamer 407 based matrix for buccal drug delivery S. Cafaggi*, R. Leardi, B. Parodi, G. Caviglioli, E. Russo, G. Bignardi Universita` di Genova- Dipartimento di Chimica e Tecnologie farmaceutiche ed alimentari, Via Brigata Salerno 16147 Genova, Italy Received 1 July 2004; accepted 21 September 2004 Available online 6 November 2004

Abstract The aim of this work was to prepare and evaluate a matrix for buccal drug delivery composed of a chitosan salt and poloxamer 407. Different chitosan salts were formed by reacting chitosan with acetic, citric, and lactic acid. Various proportions of poloxamer 407 were added to the aqueous solution of chitosan salt, and the residue obtained by lyophilisation was compressed into discs, using a 30 kN compression force. An experimental design (32) was used to study the influence of the type of chitosan salt and of the relative amount of poloxamer on drug release capacity, swelling, erosion, and mucoadhesiveness of matrices. The results showed that matrix properties depended significantly on both relative amount of poloxamer and chitosan salt type. The rank orders of chitosan salts for the four processes evaluated were as follows: drug release: chitosan acetateNchitosan citrateNchitosan lactate; swelling: chitosan lactateNchitosan acetate=chitosan citrate; erosion: chitosan citrateNchitosan lactateNchitosan acetate; mucoadhesion: chitosan lactateNchitosan acetate=chitosan citrate. Mucoadhesion was particularly favoured when poloxamer 407 was present at about 30% (w/w). The matrix composed of chitosan lactate and poloxamer 407 showed the best characteristics for buccal administration. D 2004 Elsevier B.V. All rights reserved. Keywords: Buccal delivery; Chitosan; Poloxamer 407; Mucoadhesion; Controlled release

1. Introduction The buccal route presents several advantages compared to traditional methods of systemic drug administration [1,2]. The direct entry of the drug into * Corresponding author. Tel.: +39 10 353 2625; fax: +39 10 353 2684. E-mail address: [email protected] (S. Cafaggi). 0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2004.09.019

the systemic circulation obviates the first pass hepatic metabolism; in addition, the drug can be easily administered and, if necessary, removed from the site of application which is easily accessible for selfmedication. By contrast, some drawbacks must be taken into account when a dosage form is proposed for buccal administration. Among these is the need for the device to maintain its position for many hours against buccal motion and salivary flow, the latter also

160

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

being responsible for dissolving a possible relevant part of the drug, thus reducing the mucosal absorption. Consequently, the dosage form must have good adhesive properties and show an efficient control of drug delivery. This can be accomplished by using excipients with adequate characteristics. Numerous bioadhesive polymers have been investigated for purposes of buccal administration, namely, sodium carboxymethylcellulose, hydroxypropylcellulose, Carbopol, and polycarbophil [3,4]. Another interesting polymer, chitosan, mixed with sodium alginate, was studied as a vehicle in buccal tablets [5], while chitosan glutamate, interacted with polycarbophil and other anionic polymers, was proposed for bilaminated films and bilayered tablets. It has been shown that drug release is influenced by swelling and erosion of the matrix, whereas matrix adhesiveness can be modulated using different mixtures of polymers, both adhesive and not [6]. In this paper, we propose a matrix for buccal drug delivery, composed of a chitosan salt and poloxamer 407 (P407). Chitosan is the N-deacetylated product of chitin, a polysaccharide very abundant in nature. Chitosan is gaining increasing importance in the pharmaceutical field due to its favourable properties such as biocompatibility, nontoxicity, and biodegradability. It has been shown that this polymer has good mucoadhesiveness and a significant enhancing effect on the permeation of drugs across the buccal mucosa [7,8]. P407, also known as pluronic F127, is a polyoxyethylene–polyoxypropylene–polyoxyethylene type block copolymer consisting of 70% polyoxyethylene units. It has the ability to form a clear gel in aqueous media at a concentration of approximately 20% (w/w) or more and exhibits the unique property of reversible thermal gelation; this latter is achieved at a higher temperature (e.g., body temperature) and is reversible upon cooling (e.g., at refrigerator temperature) thereby yielding a low viscosity solution. In addition, P407 has low toxicity, high solubilizing capacity, and excellent drug-release characteristics, all of which have been exploited in the polymer’s use as a drug delivery vehicle for a variety of therapeutic agents [9–11]. The combination of chitosan acetate with P407 has been adopted for a mucosal vaccine delivery system, in which the two components showed a synergistic

effect on the immune response [12]. P407 was also used in association with Carbopol to obtain mucoadhesive gels [13]. Taking into account the fact that chitosan salts have different physical properties and can have different effects on mucosa permeability [14–16], the aim of this work was to study the behaviour of a matrix composed of a chitosan salt and P407, investigating the effect of the type of chitosan salt and of the proportions of the components on matrix swelling, release capacity, and adhesion. It was expected that P407 could play an important role, based on its capacity of gelling during matrix hydration and the possibility of interaction with chitosan through hydrogen bonding. Propranolol hydrochloride, a water-soluble compound and a widely used h-blocker, was selected as a model drug, as it is among those drugs whose systemic bioavailability might be strongly improved by buccal delivery.

2. Material and methods 2.1. Materials P407 (Lutrol F 127R), was a kind gift from BASF, Milan, Italy. Medium molecular weight chitosan [molecular weight about 400 000, viscosity 286 mPa s at C=1% (w/w) in 1% (w/w) acetic acid, deacetylation grade 81%], lactic acid aqueous solution [85% (w/w)], and propranolol hydrochloride were all purchased from Sigma Aldrich (Milwaukee, USA). All the other chemicals were of reagent grade. Water was purified using the Milli-Q Plus system (Millipore, USA). 2.2. Methods 2.2.1. Experimental design In order to verify the influence of the type of chitosan salt (qualitative variable) and of the relative amount of poloxamer (quantitative variable) on the drug release capacity, swelling, erosion, and mucoadhesiveness of the compressed matrices under study, an experimental design was planned. A full factorial design at two factors and three levels (32), with a replicate at the central level of the quantitative variable for each value of the qualitative variable, was chosen.

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

Design of experiments and data analysis were performed using Modde software, Version 6.0 (Umetrics, Sweden). Results were at first processed using the multiple linear regression (MLR) method on the basis of a full quadratic model, represented by the following expression: Y ¼ b0 þ b1;a X1;a þ b1;b X1;b þ b1;c X1;c þ b2 X2 þ b1;a2 X1;a X2 þ b1;b2 X1;b X2 þ b1;c2 X1;c X2 þ b22 X22 þ e where b1;a þ b1;b þ b1;c ¼ 0 because X1;a þ X1;b þ X1;c ¼ 1 (presence–absence model) [17]. The full quadratic model was considered in order to take into account a possible curvature of the response ( Y) within the experimental domain in addition to the possible interaction between the chitosan salt type and the relative amount of poloxamer. X 1 was the qualitative variable representing the type of chitosan salt and could assume the values 0 and 1. It was associated to the qualitative factor corresponding to chitosan salt type, whose levels are indicated in Table 1 as a, b, c. X 2 was the quantitative variable, representing the relative amount of P407, which could assume the coded values reported in the same table. Nonsignificant terms were excluded from the model whenever it was useful to obtain a better prediction. Differences were considered to be significant at a level of Pb0.05. 2.2.2. Preparation of compressed matrices Chitosan 2% (w/w) and the minimum amount of acid (acetic, citric, lactic) necessary for its dissolution Table 1 Factors considered in the experimental design and their levels Factor

Level

Chitosan salt type a (acetate) b (citrate) c (lactate) P407 concentration 12 (coded 1) 31 (coded 0) 50 (coded+1) % (w/w)

161

were added to an aqueous solution of propranolol hydrochloride, whose concentration ranged from 0.2% to 0.7% (w/w). Excess of acid was avoided by controlling solution pH during chitosan dissolution under magnetic stirring. Thereafter, P407 in the concentration range 0.5–5% (w/w) was added, and the resulting suspension was stirred at room temperature until complete dissolution of the polymer. The final solution was then transferred into a glass capsule and subjected to a lyophilisation process. This was carried out using a Model 79480 freeze dry system (Laboconco, USA), maintaining samples in the drying chamber at T=30 8C for 18 h, then heating from 30 to 20 8C within an hour and maintaining at this temperature for 3 h to complete drying. The lyophilised product was then cut into small pieces and compressed in discs (13-mm diameter, 1-mm thickness, mass of about 140 mg) containing about 10 mg of drug. Compression was achieved by applying a force of 30 kN for 1 min by means of a hydraulic press equipped with a 13-mm die (Perkin Elmer, England). Table 2 reports the experimental data regarding matrix preparation. The slight difference between the fixed levels of the quantitative variable and the values obtained experimentally depended on the preparation method, in particular on the extent of drying of matrices and acid volatility. 2.2.3. Release study in vitro A Premiere 5100 dissolution system (Distek, USA) with automatic sampling was used for release studies in vitro (USP paddle apparatus). After fixing to a stainless steel support with a cyanoacrylate adhesive, discs were placed at the bottom of vessels containing phosphate buffer solution (C=0.07 mol/L, V=900 mL, pH=6.4). Working temperature was (37.0F0.5) 8C, and rotation speed was set at 50 rpm. At prefixed time intervals, aliquots were automatically withdrawn, and the drug was assayed using a spectrophotometer (HP 8453, Hewlett Packard, USA) equipped with a multicell system. Absorbance was measured at k=290 nm, the wavelength at which propranolol gave maximum absorption. Results are reported as the mean of two determinations. The calculation of the amount of drug released was made by taking into account the weight of each disc and the relative content of the drug in the matrix (Table 2).

162

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

Table 2 Experimental data regarding matrix preparation Acid type Matrix code Weight (g)

Acetic

Citric

Lactic

A B C B replicate D E F E replicate G H I H replicate

Drug concentration % (w/w )

P407 concentration % (w/w)

Chitosan Water P407

Acid

Drug

Total

Solution Freeze Expected subjected dried to freeze product drying

Obtained

Expected

Obtained

0.812 0.811 0.823 0.805 0.817 0.806 0.820 0.813 0.804 0.801 0.801 0.803

0.312 0.310 0.304 0.313 1.453 1.444 1.423 1.437 0.616 0.660 0.624 0.591

0.097 0.129 0.185 0.130 0.175 0.250 0.329 0.253 0.116 0.153 0.216 0.149

41.390 41.828 42.729 41.949 42.385 43.814 45.189 43.829 42.335 43.045 43.134 42.417

33.40 33.97 36.29 35.72 37.44 40.52 40.35 39.31 35.60 37.61 37.11 35.93

7.6 6.9 7.5 7.8 6.6 7.3 6.8 7.3 7.0 6.9 7.0 7.2

12.1 31.6 49.5 31.4 12.9 30.5 49.1 30.1 12.1 30.7 49.8 31.3

13.2 30.7 51.8 34.5 12.8 30.1 48.7 29.4 12.2 30.4 49.5 31.9

40.00 40.00 40.13 40.13 39.60 40.28 40.27 40.31 40.60 40.76 39.96 40.21

0.169 0.578 1.287 0.571 0.340 1.034 2.347 1.016 0.199 0.671 1.533 0.664

2.2.4. Swelling and erosion study Discs were weighed at dry state and fixed to a stainless steel support with a cyanoacrylate adhesive. They were then placed at the bottom of vessels containing phosphate buffer solution, using the same experimental conditions of the release study. At prefixed time intervals, discs were withdrawn from vessels, carefully blotted with filter paper to remove excess liquid, and immediately weighed. Water uptake was determined by gravimetry on the basis of the formula Water uptake ð%Þ ¼ ðW2  W 1Þ=W1  100 where W 2 was the weight of hydrated disc and W 1 was its initial weight. Erosion was evaluated at 7 h from starting experiments. At this time, discs from swelling study were withdrawn from vessels, put in an oven at (37.0F0.5) 8C overnight, and then weighed after cooling at room temperature in a dessicator. Erosion was calculated according to the following formula: Matrix erosion ð%Þ ¼ ðW1  W3 Þ=W1  100 where W 3 was the weight of dried discs and W 1 was the initial weight. Results are reported as the mean of two determinations.

1.035 1.530 2.111 1.410 2.354 3.178 4.304 3.097 1.379 1.929 2.664 1.764

7.0 7.1 7.1 7.1 6.6 7.4 6.9 7.5 7.0 7.0 7.0 7.0

2.2.5. Mucoadhesion study Mucoadhesion was evaluated through the measure of maximum force and calculation of work necessary for the detachment of the disc from the surface of a sample of porcine buccal mucosa, using a modified testing machine (model LRX, Lloyd Instruments, Fareham, England) equipped with a 10 N loadcell. Experimental data were processed by a dedicated software (Nexygen, Lloyd Instruments). Samples of buccal mucosa obtained from pigs immediately after slaughtering (CLAI, Imola, Italy) were frozen at T=20 8C and thawed just before use by immersion in physiological solution at room temperature. Defrosted mucosa was assumed to be comparable to the fresh one, regarding adhesion, as found by Park and Munday for bovine buccal mucosa [18]. Buccal mucosa, separated from its connective tissue using fine-point forceps and surgical scissors, was then fixed using a cyanoacrylate adhesive to the upper support of the instrument, whereas the disc was attached with the same adhesive to the lower one. Just before making contact between the two surfaces, mucosa was moistened with 15 AL of phosphate buffer solution (C=0.07 mol/L), at pH 6.4, made isotonic with NaCl. A contact force of 0.5 N (contact area 1.3 cm2) was applied for 10 min. The

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

mucosa was then moved upwards from the disc at a constant speed of 15 mm/min, and the detachment force was measured. The peak detachment force per unit area (N/cm2) was calculated as the quotient of the recorded maximum force (N) and the contact area (cm 2 ). The work of adhesion (mJ) was determined from the area under the force/time curve. Results are reported as the mean of at least three determinations.

3. Results and discussion 3.1. Matrix preparation and set-up of experimental design Matrices obtained as lyophilised residues were initially compressed at a lower force (0.7 kN) in order to obtain a relatively porous disc that could facilitate matrix rehydration. However, a preliminary investigation of their mucoadhesion capacity revealed that adhesion prevailed over cohesion, thus rendering impossible such type of measures. Inasmuch as matrices compressed at a much higher force gave a similar drug release and did not present that drawback, the subsequent work focused

163

on discs obtained by matrices compressed at 30 kN. Assuming that excess acid was present in negligible amount, it could be considered that matrices consisted of three components: the drug, the chitosan salt, and the poloxamer. The study was carried out taking into account two factors, namely, the type of chitosan salt and the relative amount of poloxamer, maintaining propranolol hydrochloride as a constant. Having fixed the amount of drug, this corresponded, from a quantitative point of view, to study the effect of the component proportions. Acids (acetic, lactic, and citric) chosen for chitosan solubilization presented different levels of either hydroxylation or carboxylation, thus making it possible to verify the influence of these functional groups on the interaction between the polymers through hydrogen bond formation and, more generally, on the structure of the obtained matrices. On the basis of preliminary data, the statistical experimental design described above was planned to rationalise the subsequent work. Factors involved in the plan and their levels are reported in Table 1. Five responses were chosen: the relative amount of drug released expressed as dissolution efficiency [19], the relative weight gain due to water uptake (swelling), the relative weight loss due to erosion, and finally, the

Table 3 Experimental plan and results of the experiments Exp. No

Run order

Matrix code

Chitosan salt type

P407 concentration % (w/w )

Dissolution efficiencya,b (%)

Maximum water uptakec (%)

Matrix erosiond (%)

Work of adhesione (mJ)

Detachment forcef (N/cm2)

1 2 3 4 5 6 7 8 9 10 11 12

8 4 5 12 1 10 9 2 3 6 11 7

A D G B B replicate E E replicate H H replicate C F I

Acetate Citrate Lactate Acetate Acetate Citrate Citrate Lactate Lactate Acetate Citrate Lactate

13 13 12 31 34 30 29 30 32 52 49 50

66.3 56.7 50.6 72.4 64.7 57.2 56.5 55.8 53.7 73.7 71.2 67.0

342 332 604 250 230 227 250 499 522 159 126 281

5.1 45.2 30.1 25.2 29.1 66.3 69.2 27.2 26.1 46.2 73.7 57.1

1.90 1.46 3.12 1.91 2.39 2.08 1.70 4.51 3.86 1.40 1.95 3.67

0.89 0.69 1.28 1.49 1.58 0.85 0.79 2.07 2.20 0.90 1.27 1.63

a

Defined as the area under the dissolution curve up 7 h, expressed as a percentage of the area of the rectangle described by 100% dissolution in the same time. b n=2, pooled S.D.=5.1 (%). c n=2, pooled S.D.=7.5 (%). d n=2, pooled S.D.=1.8 (%). e 3VnV5, pooled S.D.=0.38 (mJ). f 3VnV5, pooled S.D.=0.66 (N/cm2).

164

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

work of adhesion and the detachment force, i.e., the two parameters generally employed to assess adhesiveness of a material. Table 3 reports the experimental plan and the results of the experiments performed. The pooled standard deviations reported in Table 3 were obtained by extracting the square root of the pooled variance of each set of replicate measurements calculated by the formula s2p

¼ ððN1 

1Þs21

þ ðN2 

1Þs22

þ N ::

þ ðN12  1Þs212 Þ=ðN1 þ N2 þ N þ N12  12Þ where s p2 was the pooled variance, N was the sample size of the set, and 12 was the number of the mean values in each column. 3.2. Kinetics and mechanism of drug release Results of the drug release study are reported in terms of dissolution efficiency, calculated considering the entire time interval 0–7 h, over which the release curves that are depicted in Fig. 1 were registered. Kinetics and release mechanisms can be evaluated on the basis of the equation of Ritger and Peppas [20] Mt =Ml ¼ k t n where M t /M 8 is the drug fraction released at time t, k is a proportionality constant which takes into account the matrix characteristics, and n is an exponent whose

Table 4 Values of k and n (F95% confidence intervals) obtained by log plot of drug release curves from discs of different composition. R 2 values are also reported Matrix code

Chitosan salt type

k(hn )

n

R2

A B C D E F G H I

Acetate

34.6F1.1 35.4F0.9 40.3F0.2 30.4F0.3 33.9F0.8 35.2F0.6 25.2F0.2 25.9F0.3 31.8F0.3

0.65F0.05 0.65F0.04 0.62F0.01 0.55F0.01 0.52F0.02 0.63F0.02 0.60F0.01 0.62F0.01 0.67F0.01

0.994 0.997 0.999 0.999 0.996 0.998 0.999 0.999 0.999

Citrate

Lactate

value is indicative of the drug release mechanism. The equation is valid only if data points covering up to 70% of the released drug are considered. The parameters k and n can be obtained from the initial experimental data through a log transformation and by applying the least square regression method to the resulting linear curves. This gives log k and n as intercept and slope, respectively. The values of k and n obtained from the curves of the different matrices are reported in Table 4. It can be seen that n values varied from 0.52 to 0.67. Taking into account the criteria assumed to be valid in the treatment of drug release data from swelling systems [20], these values suggest that the release mechanism could be considered, in general, nonFickian, although they were very close to the Fickian limit of n=0.5, especially in the case of chitosan citrate.

Fig. 1. Drug release curves from matrices compressed at 30 kN, containing different chitosan salts and various amounts of P407. Standard deviation bars are omitted for clarity. Chitosan acetate—A, B, C; chitosan citrate—D, E, F; chitosan lactate—G, H, I. Concentration of P407, expressed as % (w/w), is indicated in parentheses.

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

This means that drug diffused partially through the swollen matrix and partially through the hydrated expanding matrix, with an increasing diffusional path. As for k values, which were generally high, these indicated the presence of a burst effect for all the matrices. 3.3. Experimental design: analysis of results 3.3.1. Drug release A graphical representation of the values of the coefficients of the model is depicted in Fig. 2, where the values of the parameters indicating the model adequacy, i.e., R 2 and Q 2, are given in the caption. They represent the fraction of variation of the response explained by the model and the fraction of variation of the response which can be predicted by the model, respectively. Values close to 1 for both R 2 and Q 2 indicate a very good model with excellent predictive power. Because the coefficients of the equation are directly proportional to the effects of the variables, the desired information can be achieved by evaluating their significance. It must be taken into account that when making considerations about the physical significance of the coefficients, the reference term, when the presence–absence model is adopted, is represented by a mean value of the response which is only a hypothetical value corresponding to a hypothetical reference state. This does not prevent the comparison of the coefficients to each other on

165

the basis of their uncertainty. The proposed model allowed evidencing of the absence of interactions between the type of chitosan salt and the relative amount of poloxamer. This means that the effects of the factors were only additive, without any synergistic action. Data were then refitted by eliminating interaction terms, thus obtaining the results depicted in Fig. 2. It can be seen that drug release was markedly higher when matrices contained chitosan acetate, whereas it was lower when chitosan citrate was present and showed a further decreasing with chitosan lactate (rank order: chitosan acetateNchitosan citrateNchitosan lactate). Drug release also increased linearly with the amount of poloxamer, as indicated by the significance of the corresponding term, the quadratic one being not significant. 3.3.2. Swelling and erosion study Water uptake by matrices and their erosion can be evaluated from the curves depicted in Fig. 3 and from gravimetric data. Data from the swelling and erosion study are reported in Table 3. Swelling extent is represented by the maximum water uptake, as revealed from the corresponding curves. Water uptake and erosion data were fitted well by the proposed quadratic model but without interaction terms, yielding the results depicted in Fig. 4. It can be seen that, when the relative amount of P407 increased, swelling decreased and erosion increased linearly. Among chitosan salts studied, chitosan lactate

Fig. 2. Graphical representation of the coefficient values of the quadratic model, without interaction terms, to describe factor effects on drug release (R 2=0.867, Q 2=0.572). Bars represent 95% confidence intervals.

166

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

Fig. 3. Plot of water uptake (swelling) by matrices of different composition. Chitosan acetate—A, B, C; chitosan citrate—D, E, F; chitosan lactate—G, H, I. Concentration of P407, expressed as % (w/w), is indicated in parentheses. Bars represent standard deviations.

produced swelling to a much higher extent than acetate and citrate, which gave reduced swelling in a similar manner (rank order: chitosan lactateNchitosan acetate=chitosan citrate). Regarding erosion, chitosan citrate showed it to the highest level, while chitosan acetate and lactate behaved quite similarly, both exerting the phenomenon less markedly. Erosion due to a chitosan salt is obviously related to its solubility in the medium. The rank order found here (chitosan citrateNchitosan lactateNchitosan acetate) is substantially in agreement with solubility data on chitosan salt films presented by other authors [21]. Taken as a whole, these results are in keeping with those yielded by the dissolution study. Indeed, matrices containing chitosan lactate gave a more sustained drug release, a finding consistent with the fact that swelling, favoured by the presence of a hydroxyl group in the molecule of lactic acid, increased the length of the diffusional path of the drug and erosion played a limited role in this case. Erosion was still less important in the case of chitosan acetate, but swelling was reduced, likely due to the lack of other hydrophilic groups in the acid molecule, thus determining a higher drug release than chitosan lactate. In the case of citrate, which swelled like acetate, erosion was higher and should have contributed significantly to the release of drug. The fact that drug release was, on the contrary, markedly lower than acetate could be explained by the compactness of the matrix, whose structure could impair drug diffusion and swelling, as it

probably gave rise to a more rigid hydrogel during hydration. It must be noted, in fact, that citric acid contains three carboxylic groups in addition to one hydroxyl group, thus determining, besides a greater solubility of the matrix, the possibility of the formation of an electrostatically cross-linked polymer network when reacting with chitosan to yield the corresponding salt and/or during matrix rehydration with buffer phosphate at pH 6.4 [22]. 3.3.3. Mucoadhesion study Results of the mucoadhesion study are reported in Table 3. As is known, both work of adhesion and detachment force can be used to assess mucoadhesion, although some authors claim that the former is the best parameter [23,24]. Both parameters were used here, but work of adhesion presented more reproducible data. Processing data by using the full quadratic model revealed that no significant interactions occurred between the chitosan salt type and the relative amount of P407. Consequently, the initial model was refined by eliminating interaction terms, thereby obtaining a better prediction. The results are depicted in Fig. 5. As can be seen from Fig. 5, chitosan lactate yielded by far the highest response, corresponding to both higher mucoadhesion work and detachment force. Chitosan acetate and citrate gave significantly lower values, showing a similar behaviour in the case of mucoadhesion work (rank order: chitosan

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

167

Fig. 4. Graphical representation of the coefficient values of the quadratic model, without interaction terms, to describe factor effects on water uptake (R 2=0.958, Q 2=0.846) and erosion (R 2=0.920, Q 2=0.746). Bars represent 95% confidence intervals.

lactateNchitosan acetate=chitosan citrate). For detachment force, the rank order was slightly different (chitosan lactateNchitosan acetateNchitosan citrate ). Mucoadhesion work and detachment force increased, but not linearly, with the relative amount of P407, as indicated by the significance of the quadratic term of the model. In other words, intermediate values of P407 concentration caused a significant increase of mucoadhesion, whereas extreme values did not give a significant variation. These results indicate that both chitosan lactate and P407 played an important role in the matrix mucoadhesion. It is known that this phenomenon is a consequence of several steps [25,26] such as polymer hydration, mucosa wetting, and interpenetration of the mucoadhesive polymer

with the mucus gel. Due to its swelling properties, chitosan lactate provided a large adhesive surface for maximum contact with mucosa, absorbing moisture from the mucosal surface and thus permitting interpenetration of the polymer into the mucosal layer and subsequent hydrogen bonding. When present at about 30% (w/w), P407 significantly favoured mucoadhesion for all samples. This result suggests that such a poloxamer concentration was optimal, probably by influencing the first two steps mentioned above, for polymer mobility and flexibility, thus promoting the mucoadhesive polymer chain/ mucin interpenetration. Preliminary experiments to examine the adhesion of discs made of such a matrix to human cheek

168

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

Fig. 5. Graphical representation of the coefficient values of the quadratic model, without interaction terms, to describe factor effects on matrix mucoadhesion (work of adhesion, R 2=0.932, Q 2=0.794; detachment force, R 2=0.841, Q 2=0.508). Bars represent 95% confidence intervals.

mucosa showed that they adhere within a few seconds and remain in place for at least 2 h. 3.3.4. Morphology of the swollen matrices A last point to emphasise is that a visual inspection of discs during both the release and the swelling study showed that discs made from matrices containing a low relative amount of P407 tended to swell, giving in general an irregular surface, whereas an intermediate proportion of poloxamer permitted a better control of

the swelling process from a morphological point of view. This was true particularly in the case of matrices containing chitosan citrate, as can be seen from the photos shown in Fig. 6.

4. Conclusions This study has demonstrated that the behaviour of a matrix composed of poloxamer 407 and a chitosan salt can differ based on the salt type and proportion of polymer. Drug release, matrix swelling, and erosion as well as mucoadhesion, depended significantly both on chitosan salt type and relative amount of poloxamer. Independently of chitosan salt type, mucoadhesion was significantly favoured when the P407 concentration in the matrix was about 30% (w/w). Chitosan lactate gave good sustained release, controlled swelling, and higher mucoadhesion when combined with P407 present in the matrix at the above concentration. These results indicate that such a matrix could find useful application in buccal drug delivery systems. Acknowledgements

Fig. 6. Pictures of swollen matrices, containing different chitosan salt type (1—chitosan acetate; 2—chitosan citrate; 3—chitosan lactate) and different amounts of P407 [A—at about 50% (w/w); B—at about 30% (w/w); C—at about 12% (w/w)]. The images were taken 4 h after starting the swelling study.

This work was supported by a grant from MIUR (Ministero dell’Universita` e della Ricerca, Rome, Italy).

S. Cafaggi et al. / Journal of Controlled Release 102 (2005) 159–169

References [1] T. Nagai, Y. Machida, Buccal delivery systems using hydrogels, Adv. Drug Deliv. Rev. 11 (1993) 179 – 191. [2] M.J. Rathbone, J. Hadgraft, Absorption of drugs from the human oral cavity, Int. J. Pharm. 74 (1991) 9 – 24. [3] H.P. Merkle, G.J.M. Wolany, in: D. Duchene (Ed.), Buccal and Nasal Administration as an Alternative to Parenteral Administration, Editions de Sante´, Paris, 1996, pp. 110 – 124. [4] B. Taylan, Y. Capan, O. Guven, S. Kes, A.A. Hincal, Design and evaluation of sustained release and buccal adhesive propranolol hydrochloride tablets, J. Control. Release 38 (1996) 11 – 20. [5] S. Miyazaki, A. Nakayama, M. Oda, M. Takada, D. Attwood, Drug release from oral mucosal tablets of chitosan and sodium alginate, Int. J. Pharm. 118 (1995) 257 – 263. [6] C. Remunan-Lopez, A. Portero, J.L. Vila-Jato, M.J. Alonso, The design and evaluation of chitosan: ethylcellulose mucoadhesive bilayered devices for buccal drug delivery, J. Control. Release 55 (1998) 143 – 152. [7] S. Senel, M.J. Kremer, H.S. Kas, P.W. Wertz, A.A. Hincal, C.A. Squier, in: M.G. Peter, R.A.A. Muzzarelli, A. Domard (Eds.), Advances in Chitin Science, vol. 4, University of Potsdam, 2000, pp. 254 – 258. [8] M.J. Kremer, S. Senel, S.H. Kas, P.W. Wertz, A.A. Hincal, C.A. Squier, Oral mucosal drug delivery: chitosan as vehicle and permeabilizer, J. Dent. Res. 77 (1999) 718. [9] K. Morikawa, F. Okada, M. Hosokawa, H. Kobayashi, Enhancement of therapeutic effects of recombinant interleukin 2 on a transplantable rat fibrosarcoma by the use of a sustained release vehicle, Pluronic gel, Cancer Res. 47 (1987) 37 – 41. [10] R. Bhardwaj, J. Blanchard, Controlled-release delivery system for the alpha-MSH analog melanotan-I using poloxamer 407, J. Pharm. Sci. 85 (1996) 915 – 919. [11] S.D. Desai, J. Blanchard, In vitro evaluation of pluronic F127based controlled-release ocular delivery systems for pilocarpine, J. Pharm. Sci. 87 (1998) 226 – 230. [12] M.A.J. Westerink, S.L. Smithson, N. Srivastava, J. Blonder, C. Coeshott, G.J. Rosenthal, ProJuvantTM (Pluronic F127R/ chitosan) enhances the immune response to intranasally administered tetanus toxoid, Vaccine 20 (2002) 711 – 723. [13] S.-C. Shin, J.-Y. Kim, I.-J. Oh, Mucoadhesive and physicochemical characterization of carbopol–poloxamer gels containing triamcinolone acetonide, Drug Dev. Ind. Pharm. 26 (2000) 307 – 312.

169

[14] A.F. Kotze, H.L. Luessen, B.J. De Leeuw, A.G. De Boer, J.C. Verhoef, H.E. Junginger, Comparison of the effect of different chitosan-salts and N-trimethyl-chitosan chloride on the permeability of intestinal epithelial cells (Caco-2), J. Control. Release 51 (1998) 35 – 46. [15] G.C. Ritthidej, T. Phaechamud, T. Koizumi, Moist heat treatment on physicochemical change of chitosan salt films, Int. J. Pharm. 232 (2002) 11 – 22. [16] A.F. Kotze, B.J. De Leeuw, H.L. Luehen, A.G. De Boer, J.C. Verhoef, H.E. Junginger, Chitosans for enhanced delivery of therapeutic peptides across intestinal epithelia: in vitro evaluation in Caco-2 cell monolayers, Int. J. Pharm. 159 (1997) 243 – 253. [17] G.A. Lewis, D. Mathieu, R. Phan-Tan-Luu, Pharmaceutical Experimental Design, Marcel Dekker, New York, 1999, p. 28. [18] C.R. Park, D.L. Munday, Development and evaluation of a biphasic buccal adhesive tablet for nicotine replacement therapy, Int. J. Pharm. 237 (2002) 215 – 226. [19] K.A. Khan, The concept of dissolution efficiency, J. Pharm. Pharmacol. 27 (1975) 48 – 49. [20] P.L. Ritger, N.A. Peppas, A simple equation for description of solute release: II. Fickian and anomalous release from swellable devices, J. Control. Release 5 (1987) 37 – 42. [21] G.C. Ritthidej, T. Phaechamud, T. Koizumi, Moist heat treatment on physicochemical change of chitosan salt films, Int. J. Pharm. 232 (2002) 11 – 22. [22] X.Z. Shu, K.J. Zhu, W. Song, Novel pH-sensitive citrate crosslinked chitosan film for drug controlled release, Int. J. Pharm. 212 (2001) 19 – 28. [23] G. Ponchel, F. Touchard, D. Duchene, N.A. Peppas, Bioadhesive analysis of controlled release systems: I. Fracture and interpenetration analysis in poly(acrylic acid) containing systems, J. Control. Release 5 (1987) 129 – 141. [24] F. Lejoyeux, G. Ponchel, D. Wouessidjewe, N.A. Peppas, D. Duchene, Bioadhesive tablets influence of the testing medium composition on bioadhesion, Drug Dev. Ind. Pharm. 15 (1989) 2037 – 2048. [25] J. Schulz, M. Nardin, in: A. Pizzi, K.L. Mittal (Eds.), Handbook of Adhesive Technology, Marcel Dekker, New York, 1994, pp. 19 – 34. [26] H.E. Junginger, Mucoadhesive hydrogels, Pharm. Ind. 53 (1991) 1056 – 1065.