- Email: [email protected]
Controlled Drug Release from Gels Using Lipophilic Interactions of Charged Substances with Surfactants and Polymers
Journal of Colloid and Interface Science 248, 194–200 (2002) doi:10.1006/jcis.2001.8182, available online at http://www.idealibrary.com on
Controlled Drug Release from Gels Using Lipophilic Interactions of Charged Substances with Surfactants and Polymers Mattias Paulsson and Katarina Edsman1 Department of Pharmacy, Uppsala University, Uppsala Biomedical Centre, Box 580, SE-751 23 Uppsala, Sweden E-mail: [email protected] Received September 20, 2001; accepted December 10, 2001; published online February 21, 2002
The aim of this article was to study interactions between different gel forming polymers and amphiphilic drugs and surfactants with the intention of finding interactions that can be used for designing controlled release formulations. The release from gels was measured by detecting the UV-absorbance of drugs released from 6 mL gel into 250 mL release medium in a dissolution bath. The rheological behavior of gels was characterized using a controlled rate rheometer. The diffusion coefficient of alprenolol was 6.3× 10−6 cm2 /s when formulated in a 1% poly(acrylic acid) gel (PAA) and 2.8 × 10−6 cm2 /s in a lipophilically modified gel (LM-PAA). The addition of alprenolol to 1% LM-PAA increased the elasticity, G , from 123 to 182 Pa. Increased gel strength was also observed for a number of other amphiphilic drugs. The addition of 1% Brij 58 to LM-PAA decreased the diffusion coefficient of alprenolol to 2.3 × 10−6 cm2 /s. It was possible to sustain the release of charged drugs with high log P by adding surfactant micelles. However, the effect was small and only useful for drugs with adequate lipophilicity. The interaction between LM-PAA and amphiphilic drugs could be seen using rheology and was used for designing controlled release gel formulations. In this way surfactants can be avoided, thus decreasing toxicity problems. C 2002 Elsevier Science (USA) Key Words: gel; controlled release; surfactants; Carbopol; Gelrite; polymers; amphiphilic drugs.
1. INTRODUCTION
Gels are often used for mucosal drug delivery since their rheological properties give the formulation an extended residence time at the absorption site (1, 2). However, an extended residence time would only be advantageous if some of the drug remains in the formulation and is released during this time. Sustained release from gels can be achieved if the drug is suspended in the gel in the form of particles (3, 4), distributed to liposomes (5, 6), or if the drug interacts with the polymer (7–9). It has been shown that surfactants can be used to give a prolonged release from gels through the partition of unionized drugs to micelles (10) or by spontaneous vesicle formation when ion-
1
2. METHODS AND MATERIALS
2.1. Materials Alprenolol hydrochloride, atenolol, diphenhydramine hydrochloride, chlorpromazine hydrochloride, tetracaine, lidocaine, orphenadrine hydrochloride, amitriptyline hydrochloride, Brij 58 (polyoxyethylene 20 cetyl ether), and Brij 700 (polyoxyethelene 100 stearyl ether) were purchased from Sigma (St. Louis, MO). See Table 1 and Fig. 1 for drug characteristics. Pluronic F-127 was bought from BASF (Parsippany, NJ).
To whom correspondence should be addressed. Fax: +46 18 471 42 23.
0021-9797/02 $35.00
C 2002 Elsevier Science (USA)
All rights reserved.
ized drugs are mixed with oppositely charged surfactants (11). Since many drugs are lipophilic and carry a charge, they are amphiphilic and thus can have both coulombic and lipophilic interactions with surfactants and polymers. Interactions between polymers and surfactants have been studied extensively, see, e.g., the review by Goddard (12). Surfactants are reported to form mixed micelles with lipophilically modified polymers (13), and the elasticity of gels can be both increased or decreased by the addition of surfactants (14) or phase separation can occur (15). Using parts of the extensive literature on polymer–surfactant interactions as reference, we will investigate the interactions between pharmaceutically acceptable polymers and amphiphilic drugs or surfactants. The aim of this study was to investigate interactions between the gel matrix, amphiphilic drugs, and surfactants with the purpose of finding interactions that can be used for controlled release formulations. Three charged drugs differing in lipophilicity (as described by log P) were formulated in gels and the release was measured. The drugs were atenolol (a highly hydrophilic drug), alprenolol, and diphenhydramine, the latter being the most lipophilic. Four different types of gels were made: one physical gel, Gelrite, and three chemically cross-linked poly(acrylic acid) hydrogels, Carbopol 934 and Carbopol 940, which differ in the degree of cross-linking, and Carbopol 1342, which has a covalently bound, lipophilic modification. Gelrite is a cation-sensitive in situ gelling polysaccharide. The nonionic surfactants Brij 58, Brij 700, and poloxamer 407 were used to form micelles’ with various characteristics.
194
195
CONTROLLED DRUG RELEASE FROM GELS
TABLE 1 Characteristics of Drug Substances Used in the Release Experiments Drug
pK a
c Log P
Log D
Alprenolol Amitriptylin Atenolol Chlorpromazine Diphenhydramine Lidocaine Orphenadrine Tetracaine
9.6 9.4 9.6 9.4 9.0 7.9 8.4 8.5
2.65 4.85 0.16 5.20 3.54 3.25 4.04 3.02
0.45 2.55 −2.04 3.20 1.93 2.63 3.00 1.89
Note. c Log P is the calculated octanol/water partition coefficient. Log D is the distribution coefficient and the pH was 7.4.
Poly(acrylic acid) polymers cross-linked with allyl sucrose or allylpentaerythritol and with the proprietary names Carbopol 934P (C934), Carbopol 940NF (C940), and Carbopol 1342NF (C1342) were kind gifts from BF Goodrich (Brecksville, OH). Deacetylated gellan gum (Kelcogel F), also called Gelrite, was a kind gift from the Kelco division of the Monsanto Company
FIG. 2.
Structure of the polymers studied.
(San Diego, CA). See Fig. 2 for structures of the polymers. All other chemicals were from Sigma Chemical Co. and of analytical or “ultra” quality. Ultra-pure water, prepared using a MilliQ Water Purification System (Millipore, France), was used in all preparations. 2.2. Preparation of Samples
FIG. 1.
Structure of the model substances studied.
The composition of simulated tear fluid was adopted from a tear fluid analysis (16) using 8.3 g NaCl, 0.084 g CaCl2 · 2H2 O, and 1.4 g KCl in 1 liter of ultra-pure water. This is equivalent to 142 mM of Na+ , 19 mM K+ , and 0.6 mM of Ca2+ . The concentration of alprenolol, atenolol, and diphenhydramine was 18 mM in all gels except when otherwise stated. The concentration of surfactants was 1% (w/w) except when otherwise stated. 2.2.1. Preparation of Carbopol gels. The polymer powder was dispersed in simulated tear fluid containing the dissolved model drug with or without surfactants. The dispersions were then stirred using magnetic stirring bars for approximately 1 h at room temperature and eventually 1 or 2 M NaOH, depending on the polymer concentration, was added to neutralize the sample to pH 6.5–7. All gels were allowed to equilibrate for at least 16 h at room temperature. The pH of the gels was then adjusted to pH 7.4, simulated tear fluid was added to achieve the final volume, and the gels were left for at least 90 min before measurement commenced, to allow the gels to equilibrate. The polymer content of all Carbopol gels was 1% (w/w), except where otherwise stated. 2.2.2. Preparation of Gelrite gels. The polymer powder was dispersed in ultra-pure water, which contained dissolved surfactant for some systems. The dispersions were then stirred for 20 min at 100◦ C using a water bath. The model substances were
196
PAULSSON AND EDSMAN
added during the cooling to room temperature, then the solutions were allowed to equilibrate for at least 16 h. The polymer content of all Gelrite gels was 0.5% (w/w).
within the linear viscoelastic region at 35◦ C. Gelrite gels were studied using temperature sweeps over the range 90–5◦ C, with a cooling rate of 0.5◦ C/min as described elsewhere (19).
2.3. Drug Release Measurements
2.6. Cryogenic Transmission Electron Microscopy (Cryo-TEM)
Drug release from the gels was measured by the USP paddle (XXI) method using gel containers with a fixed volume of 6 cm3 and a surface area of 21 cm2 , covered by a coarse mesh-size plastic net and a stainless steel net. The containers were immersed in 250 mL of simulated tear fluid maintained at 35◦ C and stirred at 20 rpm using a Pharma Test PTW II USP bath (Pharma Test Apparatebau, Germany). The stirring rate was chosen to give adequate convection but minimize surface erosion of the gels. On-line measurements of the concentration were performed by continuously pumping the dissolution media to a UV-vis spectrophotometer, Shimadzu UV-1601 (Shimadzu, Kyoto, Japan) using a peristaltic pump and ismaprene tubing (Ismatec SA, Z¨urich, Switzerland). The absorbance was measured every 150 s for the first 45 min, then after 65 min and eventually every 30 min, the last measurement being that made 6 h after the first one. The maximum absorbance wavelength was found to be 274 nm for atenolol, 271 nm for alprenolol, and 258 nm for diphenhydramine. The light scattering of polymer released from the gel at the end of the experiment (after approximately 4 h) was compensated for by subtracting the absorbance from a drug-free experiment. 2.4. Drug Release and Calculation of Diffusion Coefficient One-dimensional Fickian diffusion from a gel holder can be expressed by the following expression under sink conditions during the initial part of the release, Q = 2C0
Dt π
12
,
A small drop of the sample was deposited on a grid covered by a holey polymer film, and excess liquid was blotted with filter paper and thereafter vitrified in liquid ethane. The films were transferred to a Zeiss EM 902 transmission electron microscope and kept below 108 K throughout this part of the investigation. Details of the method are described elsewhere (20). 2.7. Statistical Analysis The 95% confidence interval was calculated for the slope of the√line when the fraction released was plotted as a function of t. 3. RESULTS AND DISCUSSION
3.1. Drug–Polymer Interactions. Effects on Gel Strength Interactions between polymers and surfactants have been studied extensively, see, e.g., the review by Goddard (12). Drug molecules are often amphiphilic, and in this part of the article, the interaction between drugs and polymers will be discussed with support from the abundant literature on polymer–surfactant interactions. In Fig. 3 the rheological behavior of C1342 gels formulated with some known amphiphilic drugs is shown. C1342 is a cross-linked poly(acrylic acid) with a lipophilic modification consisting of an alkyl acrylate chain of C10–C30. The drugs have lipophilicities with log P ranging from 2.65 to 5.20, as shown in Table 1. For alprenolol, diphenhydramine, and lidocaine the
[1]
where Q is the amount of drug released per unit area, C0 is the initial concentration of the drug in the gel, D is the diffusion coefficient of the drug in the gel, and t is the time that has elapsed since the release experiment started. The equation is valid for the first 60% of the fractional release (17, 18). In our laboratory setting the gel was placed in a confined space and was not allowed to swell during the study. Plots of the initial drug release versus the square root of time should give a straight line and the diffusion coefficient can be calculated from the slope of the line. 2.5. Rheological Measurements The rheological measurements were carried out using a Bohlin VOR Rheometer (Bohlin Reologi, Lund, Sweden), a controlled strain rate instrument of the couette type, in the dynamic oscillation mode. A concentric cylinder measuring system (C14) was used and silicone oil was applied to the surface of the sample to prevent evaporation. All measurements were performed
FIG. 3. The elastic modulus, G of 1% C1342 gels containing 18 mM of the drug. The reference line shows the G of a drug-free 1% C1342 gel.
CONTROLLED DRUG RELEASE FROM GELS
elastic modulus, G , observed is higher than that for the pure C1342 gel. The increase cannot be attributed to an increased ionic strength; on the contrary, the addition of inorganic salts has previously been shown to decrease the gel strength (21). When these drugs (alprenolol and diphenhydramine) are formulated in gels that lack the lipophilic modification (e.g., C934) no strengthening effect is seen (Table 2). It seems that the increased G , accompanied by an increased viscous modulus, G , is caused by interactions between the lipophilic modification of the polymer and the nonpolar part of the drug. For the more hydrophilic atenolol, which lacks amphiphilic properties, no increase in the gel strength is observed. Uncharged, lipophilic substances lacking amphiphilic properties (e.g., butylparaben) do not affect the gel strength (10). Surprisingly, the more lipophilic drug compounds in Fig. 3, such as amitriptylin, do not affect the gel strength when formulated at the concentration 18 mM. However, as can be seen in Fig. 4, when the concentration is lowered to 9 mM an interaction is seen as an increased elasticity. The balance of the polar and lipophilic character of the drug (seen as, e.g., CMC, log P) and hence the balance of electrostatic and lipophilic interactions between the polymer and the drug can explain (7, 22) the differences observed in the concentration dependence of the interaction. This is best exemplified by alprenolol in C1342
TABLE 2 Diffusion Coefficients (Mean ± 95% Confidence Interval) and Rheological Characteristics of Gel Formulations at 1 Hz Formulation Pure C934 (1%) Pure C1342 (1%) Alprenolol (18 mM), C934 (1%) Alprenolol (18 mM), C934 (2%) Alprenolol (18 mM), C940 (1%) Alprenolol (9 mM), C1342 (1%) Alprenolol (18 mM), C1342 (1%) Alprenolol (36 mM), C1342 (1%) Alprenolol (18 mM), C1342 (0.5%) Alprenolol (18 mM), C1342 (2%) Alprenolol (18 mM), C1342 (1%), B58 (1%) Alprenolol (18 mM), Gelrite (0.5%) Alprenolol (18 mM), Gelrite (0.5%), B58 (1%) Atenolol (18 mM), C934 (1%) Atenolol (18 mM), C940 (1%) Atenolol (18 mM), C1342 (1%) Atenolol (18 mM), C934 (1%), B58 (1%) Diphenhydramine (18 mM), C934 (1%) Diphenhydramine (18 mM), C1342 (1%) Diphenhydramine (18 mM), C1342 (1%), B58 (1%)
D (10−6 cm2 /s) G (Pa) G (Pa) — — 6.31 ± 0.23 3.48 ± 0.17 4.76 ± 0.30 2.77 ± 0.14 2.78 ± 0.10 2.61 ± 0.083 6.32 ± 0.069 1.44 ± 0.10 2.29 ± 0.13
65.4 123 68.9 493 206 207 182 79.3 15.4 853 157
6.32 ± 0.16 4.80 ± 0.33
0.02–104b 0.02–104b
6.02 ± 0.26 4.49 ± 0.24 3.50 ± 0.22 6.17 ± 0.25 6.51 ± 0.27 4.00 ± 0.18 3.32 ± 0.15
65.4 237 131 68.2 55.8a 178a 114a
5.98 7.32 6.55 41.4 18.6 24.5 42.1 32.4 4.4 202 47.2
4.46 7.8 11.3 4.85 5.32a 43.2a 21.4a
a Gels prepared according to the alternative method, mixing neutralized solution with gel in the ratio 1 : 1. b Gels formed in situ during the release experiment. The interval of the elastic modulus is shown.
197
FIG. 4. The elastic modulus, G of 1% C1342 gels with varying concentrations of alprenolol, amitriptylin, and diphenhydramine. For some gels precipitation occurred ( ) and rheological characterization was not possible ().
(Fig. 4) where rheological measurements show that G is larger than for the pure gel, not only for the 18 mM alprenolol concentration as discussed above, but also for the gels with concentrations of 4.5 and 9 mM. Owing to its amphiphilic properties, alprenolol can probably assist in the formation of micelle-like structures containing the lipophilic modifications of C1342 as has been shown with surfactants (23). These aggregates may explain the increased gel strength in the presence of alprenolol. The rheological behavior of hydrophobically modified polyacrylates and oppositely charged surfactants has been described by Senan et al. (24). The hydrophobic and electrostatic interactions between C934 polymers (lacking lipophilic modification) and oppositely charged surfactants has been discussed by BarreiroIglesias et al. (25). In the gel with 36 mM alprenolol, G is lower than that in pure C1342 gels, and with 72 mM alprenolol, the gel collapsed. At this alprenolol concentration there is significant shielding of the carboxylic acid groups on the polymer, neutralizing the polyelectrolyte and impeding the swelling of the hydrogel network. An approximate calculation of the number of ionic groups of C1342 can be made using the weak acid ionization theory (Henderson–Hasselbalch equation) with pH 7.4, pK a = 6.0, and a carboxylic acid content of 57% (the number of monomers carrying free carboxylic acid according to the manufacturer). At 1% C1342 the concentration of COO− is 76 mM, and comparing this with the alprenolol concentration of 72 mM, the stoichiometric ratio [alprenolol]/[COO− ] is close to unity. The degree of binding, β (which is the ratio of the number of bound amphiphilic molecules to the number of ionic groups on the polyelectrolyte) can, however, not be determined without some knowledge of the amount of drug bound to the polymer (e.g., through equilibrium dialysis measurements). It has previously
198
PAULSSON AND EDSMAN
been reported that phase separation occurs for completely ionized linear PAA when the surfactant dodecylpyridinium is added so that β is above 0.8 (15). The concentration of lipophilic sites (alkyl groups) on the C1342 polymer is about 5–10 mM for a 1% gel. This value was obtained by performing a rough calculation based on the assumption that the 5% differences in carboxylic acid content between C1342 and C934 (stated in the U.S. National Formulary, 18th ed.) is caused by the grafting of lipophilic modifications. Hydrophobically modified polymers in concentrations on the order of 1% have been reported to form hydrophobic microdomains that can solubilize hydrophobic molecules, see, for example, the review by Piculell et al. (26). The concentration of these domains can be calculated by dividing the concentration of alkyl groups by the aggregation number of lipophilic modifications in one domain. When the concentration of the amphiphilic drug compound alprenolol is kept constant (18 mM) but the polymer concentration is varied, it can be seen (Table 3) that the gel strength is increased for both 1 and 2% C1342 gels, but slightly decreased for 0.5 and 0.75% C1342 gels in comparison to pure C1342 gels. It has previously been observed that surfactants can interact with polymers that have lipophilic modifications (LM) and both increase and decrease the elasticity of the gel. High concentrations of surfactants give an increase in the elasticity of the gel through the solubilization of many LM in one micelle and thereby forming additional cross-links. This effect was seen neither at low nor at very high concentrations since there are either too few micelles to have an effect, or there are so many micelles that there will only be one LM per micelle and no cross-linking can occur (13, 14). 3.2. Drug–Polymer Interactions: Effects on Drug Release In this section the effect the interactions discussed above have on the diffusion of a drug in a gel will be examined. The emphasis will be on alprenolol formulations. It is widely recognized that the choice of polymer can be used to control the release from a gel (7, 8, 27). In Fig. 5 the release of alprenolol and atenolol from different gels is shown. The C934 gel has the lowest cross-linking density of the three (according to the manufacturer) and gives the fastest drug release. The lipophilically modified C1342 has the highest density and gives
TABLE 3 Rheological Properties of Carbopol 1342 Gels When 18 mM Alprenolol or 1% Brij 58 Was Added Concentration (%)
Pure gel G (Pa)
C1342 with alprenolol G (Pa)
C1342 with Brij 58 G (Pa)
0.5 0.75 1 2
20.8 72.7 127 350
15.4 64.4 208 853
15.7 53.3 88.4 181
FIG. 5. The release (mean ± standard deviation, n = 3) of atenolol (open symbols) and alprenolol (closed symbols) from gels of C934 (, ), C940 (, ), and C1342 (, ).
the slowest release, while C940 with intermediate cross-linking has an intermediate release rate. When comparing the release of atenolol and alprenolol from the three different gels it can be noted that the release of the two substances does not differ in C934 and C940. In C1342, however, alprenolol has a significantly slower release than atenolol. It is likely that the interactions detected in the rheological study are the reason for the slower diffusion, as in other systems which have been suggested for controlled release formulations (7, 9). In Table 2 it can be seen that diphenhydramine has a slower release when formulated with the C1342 polymer than when C934 is used. Surprisingly, although diphenhydramine has a higher log P, the diffusion coefficient is higher than that of both atenolol and alprenolol. Perhaps the slow release of alprenolol is caused by a different kind of interaction than the one that diphenhydramine has with the polymer. Cryo transmission electron microscopy (cryo-TEM) of the C1342 gel with alprenolol could not explain this difference (images not shown). Small micelle-like structures can, however, be difficult to see in a gel using this technique. As demonstrated by the diffusion coefficients in Table 2, the release of alprenolol from gels of C934 and Gelrite is the same despite the different structure of the covalently cross-linked poly(acrylic acid) C934 and Gelrite, cross-linked in situ by cations. The amount of lipophilic modification on the polymer compared with the amount of drug will influence the release rate from the formulation. In Table 2 the release of alprenolol from C1342 gels is shown with three different polymer concentrations. The diffusion coefficient decreases when the polymer concentration is increased from 0.5 to 2% because of the increased number of sites on the polymer with which the drug can interact lipophilically and/or electrostatically. Likewise, for C934 an increase
CONTROLLED DRUG RELEASE FROM GELS
FIG. 6. The release (mean ± standard deviation, n = 3) of diphenhydramine from gels of C934 (), C1342 (), and C1342 with 1% Brij 58 ().
in polymer concentration from 1 to 2% decreases the diffusion coefficient. This confirms that either electrostatic and/or lipophilic interactions are present in the C934 and the C1342 formulations. Further information regarding the C1342–alprenolol interaction can be obtained by varying the drug concentration and keeping the polymer concentration constant. When the polymer content was 1% C1342 and the drug concentration varied, no differences were seen in the release data between gels with 9, 18, and 36 mM alprenolol. Comparison with the concentration of lipophilic modifications (5–10 mM) calculated above shows that the concentration of drug is in excess of the lipophilic domains for all concentrations used since many LM are probably clustered into the lipophilic domains. As discussed in Section 3.1, the rheological behavior changes when the drug loading is varied.
199
surfactant in the same poly(acrylic acid) gel (10). The higher log P for diphenhydramine than for alprenolol suggests a greater proportion of lipophilic interactions with Brij 58 in the formation of mixed micelles. Both drugs are, however, affected by the presence of Brij 58 to the same extent. Being highly hydrophilic, the diffusion of atenolol is not affected by the presence of Brij 58. Carbopol formulations were also made with other nonionic surfactants such as 1% Brij 700 and 1% poloxamer and the diffusion coefficients calculated. These formulations showed only slightly—and not statistically significant—lower diffusion coefficients so the data are not shown. In Fig. 6 the release of diphenhydramine from the C1342 gel with Brij 58 is slower than from the C1342 gel without surfactant, but after 180 min the curves intersect and the Brij 58 formulation has a faster release rate. A plausible explanation could be that the fraction of drug that interacts with the surfactants also is released together with the surfactants, whereas the fraction of drug that interacts with the LM-PAA polymer matrix has a longer distance over which to diffuse since the matrix is stationary. The fraction of the drug that interacts with the LM is probably smaller when Brij 58 micelles are present. As the micelles are released from the gel this difference is shown by the crossover of the curves after 180 min. The addition of nonionic surfactants to the C1342 formulations in this study only brought about minor differences. In Gelrite gels the addition of 1% Brij 58 decreased the diffusion coefficient from 6.3 × 10−6 to 4.8 × 10−6 cm2 /s. In this gel there are no lipophilic modifications that can participate in lipophilic interactions with the drug, and the addition of micelles gives a greater reduction than it did in the C1342 formulations. With these formulations, for which the interaction with the polymer or the formation of mixed micelles is used to give controlled release, charged surfactants are avoided, and thereby the
3.3. Addition of Surfactants By including micelle-forming surfactants in the formulation, the diffusion of drug molecules can be further reduced since lipophilic interactions can take place between the drug and both the polymer and micelle. Brij 58 is a nonionic surfactant with a large polyoxyethylene shell. The CMC is 0.007 mM (according to the manufacturer), and in the presence of C934 the CMC is lowered to CAC, which is likely half of the CMC (based on data for nonionic surfactants in C934 (25)). Compared to pure C1342 gels, both diphenhydramine (Fig. 6) and alprenolol have significantly slower release rates when Brij 58 micelles are present (Table 2). The amphiphilic drugs can probably interact with and form mixed micelles with Brij 58, resulting in a slower diffusion. The solubilization of uncharged, lipophilic model substances in 1% Brij 58 has previously been reported to give a diffusion coefficient 20 times lower than when they are formulated without
FIG. 7. The effect of Brij 58 on the elastic modulus, G , () and the phase angle, δ, () of 1% C1342 gels.
200
PAULSSON AND EDSMAN
toxicity is possibly reduced. However, these formulations have a smaller effect on the release than was seen when oppositely charged surfactants were used as described by Paulsson and Edsman (11). This reference describes a totally different way to control the release based on the formation of vesicles. The addition of surfactants often affects the gel strength, and in Fig. 7 and Table 3 it can be seen that when Brij 58 is added, the gel strength is decreased for all polymer concentrations. The presence of these surfactants seems to disturb the structure of the gel matrix, thereby decreasing the elasticity (G ) of the gel. The Brij 58 concentration is well above the CMC (low CMC ≈ 0.007 mM, according to the manufacturer) and thus does not contribute to the cross-linking of the gel as described by J¨onsson et al. (13) and discussed above. 4. SUMMARY
The interaction between lipophilically modified PAA and drug compounds can be measured using rheology and can be predicted to some degree by considering the lipophilicity of the drug. However, the ratio of lipophilic modification of polymer to drug concentration also seems to play a major role. The interaction between lipophilically modified PAA and amphiphilic drugs can be used for constructing controlled release gel formulations. It is also possible to sustain the release of charged drugs with adequate lipophilicity by adding nonionic surfactants with which the drugs can form mixed micelles; however, these effects are only minor in lipophilically modified gels. From a toxicological point-of-view, formulations without surfactants are often preferable. ACKNOWLEDGMENTS The excellent experimental skills of Mr. Stefano Bonfante and Mr. G¨oran Karlsson (who produced the cryo-TEM images) are gratefully acknowledged.
REFERENCES 1. Saettone, M. F., Giannaccini, B., Ravecca, S., LaMarca, F., and Tota, G., Int. J. Pharm. 20, 187 (1984). 2. Zhou, M., and Donovan, M. D., Int. J. Pharm. 135, 115 (1996). 3. Desai, S. D., and Blanchard, J., Drug. Deliv. 7, 201 (2000). 4. Di Colo, G., Carelli, V., Giannaccini, B., Serafini, M. F., and Bottari, F., J. Pharm. Sci. 69, 387 (1980). 5. Paavola, A., Kilpel¨ainen, I., Yliruusi, J., and Rosenberg, P., Int. J. Pharm. 199, 85 (2000). 6. Bochot, A., Fattal, E., Grossiord, J. L., Puisieux, F., and Couvreur, P., Int. J. Pharm. 162, 119 (1998). 7. Sj¨oberg, H., Persson, S., and Caram-Lelham, N., J. Control. Release 59, 391 (1999). 8. am Ende, M. T., and Peppas, N. A., J. Control. Release 48, 47 (1997). 9. Inoue, T., Chen, G., Nakamae, K., and Hoffman, A. S., J. Control. Release 49, 167 (1997). 10. Paulsson, M., and Edsman, K., J. Pharm. Sci. 90, 1216 (2001). 11. Paulsson, M., and Edsman, K., Pharm. Res. 18, 1586 (2001). 12. Goddard, E. D., J. Soc. Cosmet. Chem. 41, 23 (1990). 13. J¨onsson, B., Lindman, B., Holmberg, K., and Kronberg, B., “Surfactants and Polymers in Aqueous Solution.” Wiley, Chichester, England, 1998. 14. Magny, B., Iliopoulos, I., Zana, R., and Audebert, R., Langmuir 10, 3180 (1994). 15. Shimizu, T., Colloid Surf. A-Physicochem. Eng. Asp. 94, 115 (1995). 16. Stjernschantz, J., and Astin, M., in “Biopharmaceutics of Ocular Drug Delivery” (P. Edman, Ed.), p. 1. CRC Press, Boca Raton, FL, 1993. 17. Higuchi, W. I., J. Pharm. Sci. 51, 802 (1962). 18. Ritger, P. L., and Peppas, N. A., J. Control. Release 5, 23 (1987). 19. Paulsson, M., H¨agerstr¨om, H., and Edsman, K., Eur. J. Pharm. Sci. 9, 99 (1999). 20. Almgren, M., Edwards, K., and Karlsson, G., Colloid. Surf. A 174, 3 (2000). 21. Kriwet, B., and Kissel, T., Int. J. Pharm. 127, 135 (1996). 22. Lowe, T. L., Virtanen, J., and Tenhu, H., Polymer 40, 2595 (1999). 23. Piculell, L., Thuresson, K., and Ericsson, O., Faraday Disc. 101, 307 (1995). 24. Senan, C., Meadows, J., Shone, P. T., and Williams, P. A., Langmuir 10, 2471 (1994). 25. Barreiro-Iglesias, R., Alvarez-Lorenzo, C., and Concheiro, A., J. Control. Release 77, 59 (2001). 26. Piculell, L., Guillemet, F., Thuresson, K., Shubin, V., and Ericsson, O., Adv. Colloid Interface Sci. 63, 1 (1996). 27. Zentner, G. M., Rathi, R., Shih, C., McRea, J. C., Seo, M. H., Oh, H., Rhee, B. G., Mestecky, J., Moldoveanu, Z., Morgan, M., and Weitman, S., J. Control. Release 72, 203 (2001).
Controlled Drug Release from Gels Using Lipophilic Interactions of Charged Substances with Surfactants and Polymers Mattias Paulsson and Katarina Edsman1 Department of Pharmacy, Uppsala University, Uppsala Biomedical Centre, Box 580, SE-751 23 Uppsala, Sweden E-mail: [email protected] Received September 20, 2001; accepted December 10, 2001; published online February 21, 2002
The aim of this article was to study interactions between different gel forming polymers and amphiphilic drugs and surfactants with the intention of finding interactions that can be used for designing controlled release formulations. The release from gels was measured by detecting the UV-absorbance of drugs released from 6 mL gel into 250 mL release medium in a dissolution bath. The rheological behavior of gels was characterized using a controlled rate rheometer. The diffusion coefficient of alprenolol was 6.3× 10−6 cm2 /s when formulated in a 1% poly(acrylic acid) gel (PAA) and 2.8 × 10−6 cm2 /s in a lipophilically modified gel (LM-PAA). The addition of alprenolol to 1% LM-PAA increased the elasticity, G , from 123 to 182 Pa. Increased gel strength was also observed for a number of other amphiphilic drugs. The addition of 1% Brij 58 to LM-PAA decreased the diffusion coefficient of alprenolol to 2.3 × 10−6 cm2 /s. It was possible to sustain the release of charged drugs with high log P by adding surfactant micelles. However, the effect was small and only useful for drugs with adequate lipophilicity. The interaction between LM-PAA and amphiphilic drugs could be seen using rheology and was used for designing controlled release gel formulations. In this way surfactants can be avoided, thus decreasing toxicity problems. C 2002 Elsevier Science (USA) Key Words: gel; controlled release; surfactants; Carbopol; Gelrite; polymers; amphiphilic drugs.
1. INTRODUCTION
Gels are often used for mucosal drug delivery since their rheological properties give the formulation an extended residence time at the absorption site (1, 2). However, an extended residence time would only be advantageous if some of the drug remains in the formulation and is released during this time. Sustained release from gels can be achieved if the drug is suspended in the gel in the form of particles (3, 4), distributed to liposomes (5, 6), or if the drug interacts with the polymer (7–9). It has been shown that surfactants can be used to give a prolonged release from gels through the partition of unionized drugs to micelles (10) or by spontaneous vesicle formation when ion-
1
2. METHODS AND MATERIALS
2.1. Materials Alprenolol hydrochloride, atenolol, diphenhydramine hydrochloride, chlorpromazine hydrochloride, tetracaine, lidocaine, orphenadrine hydrochloride, amitriptyline hydrochloride, Brij 58 (polyoxyethylene 20 cetyl ether), and Brij 700 (polyoxyethelene 100 stearyl ether) were purchased from Sigma (St. Louis, MO). See Table 1 and Fig. 1 for drug characteristics. Pluronic F-127 was bought from BASF (Parsippany, NJ).
To whom correspondence should be addressed. Fax: +46 18 471 42 23.
0021-9797/02 $35.00
C 2002 Elsevier Science (USA)
All rights reserved.
ized drugs are mixed with oppositely charged surfactants (11). Since many drugs are lipophilic and carry a charge, they are amphiphilic and thus can have both coulombic and lipophilic interactions with surfactants and polymers. Interactions between polymers and surfactants have been studied extensively, see, e.g., the review by Goddard (12). Surfactants are reported to form mixed micelles with lipophilically modified polymers (13), and the elasticity of gels can be both increased or decreased by the addition of surfactants (14) or phase separation can occur (15). Using parts of the extensive literature on polymer–surfactant interactions as reference, we will investigate the interactions between pharmaceutically acceptable polymers and amphiphilic drugs or surfactants. The aim of this study was to investigate interactions between the gel matrix, amphiphilic drugs, and surfactants with the purpose of finding interactions that can be used for controlled release formulations. Three charged drugs differing in lipophilicity (as described by log P) were formulated in gels and the release was measured. The drugs were atenolol (a highly hydrophilic drug), alprenolol, and diphenhydramine, the latter being the most lipophilic. Four different types of gels were made: one physical gel, Gelrite, and three chemically cross-linked poly(acrylic acid) hydrogels, Carbopol 934 and Carbopol 940, which differ in the degree of cross-linking, and Carbopol 1342, which has a covalently bound, lipophilic modification. Gelrite is a cation-sensitive in situ gelling polysaccharide. The nonionic surfactants Brij 58, Brij 700, and poloxamer 407 were used to form micelles’ with various characteristics.
194
195
CONTROLLED DRUG RELEASE FROM GELS
TABLE 1 Characteristics of Drug Substances Used in the Release Experiments Drug
pK a
c Log P
Log D
Alprenolol Amitriptylin Atenolol Chlorpromazine Diphenhydramine Lidocaine Orphenadrine Tetracaine
9.6 9.4 9.6 9.4 9.0 7.9 8.4 8.5
2.65 4.85 0.16 5.20 3.54 3.25 4.04 3.02
0.45 2.55 −2.04 3.20 1.93 2.63 3.00 1.89
Note. c Log P is the calculated octanol/water partition coefficient. Log D is the distribution coefficient and the pH was 7.4.
Poly(acrylic acid) polymers cross-linked with allyl sucrose or allylpentaerythritol and with the proprietary names Carbopol 934P (C934), Carbopol 940NF (C940), and Carbopol 1342NF (C1342) were kind gifts from BF Goodrich (Brecksville, OH). Deacetylated gellan gum (Kelcogel F), also called Gelrite, was a kind gift from the Kelco division of the Monsanto Company
FIG. 2.
Structure of the polymers studied.
(San Diego, CA). See Fig. 2 for structures of the polymers. All other chemicals were from Sigma Chemical Co. and of analytical or “ultra” quality. Ultra-pure water, prepared using a MilliQ Water Purification System (Millipore, France), was used in all preparations. 2.2. Preparation of Samples
FIG. 1.
Structure of the model substances studied.
The composition of simulated tear fluid was adopted from a tear fluid analysis (16) using 8.3 g NaCl, 0.084 g CaCl2 · 2H2 O, and 1.4 g KCl in 1 liter of ultra-pure water. This is equivalent to 142 mM of Na+ , 19 mM K+ , and 0.6 mM of Ca2+ . The concentration of alprenolol, atenolol, and diphenhydramine was 18 mM in all gels except when otherwise stated. The concentration of surfactants was 1% (w/w) except when otherwise stated. 2.2.1. Preparation of Carbopol gels. The polymer powder was dispersed in simulated tear fluid containing the dissolved model drug with or without surfactants. The dispersions were then stirred using magnetic stirring bars for approximately 1 h at room temperature and eventually 1 or 2 M NaOH, depending on the polymer concentration, was added to neutralize the sample to pH 6.5–7. All gels were allowed to equilibrate for at least 16 h at room temperature. The pH of the gels was then adjusted to pH 7.4, simulated tear fluid was added to achieve the final volume, and the gels were left for at least 90 min before measurement commenced, to allow the gels to equilibrate. The polymer content of all Carbopol gels was 1% (w/w), except where otherwise stated. 2.2.2. Preparation of Gelrite gels. The polymer powder was dispersed in ultra-pure water, which contained dissolved surfactant for some systems. The dispersions were then stirred for 20 min at 100◦ C using a water bath. The model substances were
196
PAULSSON AND EDSMAN
added during the cooling to room temperature, then the solutions were allowed to equilibrate for at least 16 h. The polymer content of all Gelrite gels was 0.5% (w/w).
within the linear viscoelastic region at 35◦ C. Gelrite gels were studied using temperature sweeps over the range 90–5◦ C, with a cooling rate of 0.5◦ C/min as described elsewhere (19).
2.3. Drug Release Measurements
2.6. Cryogenic Transmission Electron Microscopy (Cryo-TEM)
Drug release from the gels was measured by the USP paddle (XXI) method using gel containers with a fixed volume of 6 cm3 and a surface area of 21 cm2 , covered by a coarse mesh-size plastic net and a stainless steel net. The containers were immersed in 250 mL of simulated tear fluid maintained at 35◦ C and stirred at 20 rpm using a Pharma Test PTW II USP bath (Pharma Test Apparatebau, Germany). The stirring rate was chosen to give adequate convection but minimize surface erosion of the gels. On-line measurements of the concentration were performed by continuously pumping the dissolution media to a UV-vis spectrophotometer, Shimadzu UV-1601 (Shimadzu, Kyoto, Japan) using a peristaltic pump and ismaprene tubing (Ismatec SA, Z¨urich, Switzerland). The absorbance was measured every 150 s for the first 45 min, then after 65 min and eventually every 30 min, the last measurement being that made 6 h after the first one. The maximum absorbance wavelength was found to be 274 nm for atenolol, 271 nm for alprenolol, and 258 nm for diphenhydramine. The light scattering of polymer released from the gel at the end of the experiment (after approximately 4 h) was compensated for by subtracting the absorbance from a drug-free experiment. 2.4. Drug Release and Calculation of Diffusion Coefficient One-dimensional Fickian diffusion from a gel holder can be expressed by the following expression under sink conditions during the initial part of the release, Q = 2C0
Dt π
12
,
A small drop of the sample was deposited on a grid covered by a holey polymer film, and excess liquid was blotted with filter paper and thereafter vitrified in liquid ethane. The films were transferred to a Zeiss EM 902 transmission electron microscope and kept below 108 K throughout this part of the investigation. Details of the method are described elsewhere (20). 2.7. Statistical Analysis The 95% confidence interval was calculated for the slope of the√line when the fraction released was plotted as a function of t. 3. RESULTS AND DISCUSSION
3.1. Drug–Polymer Interactions. Effects on Gel Strength Interactions between polymers and surfactants have been studied extensively, see, e.g., the review by Goddard (12). Drug molecules are often amphiphilic, and in this part of the article, the interaction between drugs and polymers will be discussed with support from the abundant literature on polymer–surfactant interactions. In Fig. 3 the rheological behavior of C1342 gels formulated with some known amphiphilic drugs is shown. C1342 is a cross-linked poly(acrylic acid) with a lipophilic modification consisting of an alkyl acrylate chain of C10–C30. The drugs have lipophilicities with log P ranging from 2.65 to 5.20, as shown in Table 1. For alprenolol, diphenhydramine, and lidocaine the
[1]
where Q is the amount of drug released per unit area, C0 is the initial concentration of the drug in the gel, D is the diffusion coefficient of the drug in the gel, and t is the time that has elapsed since the release experiment started. The equation is valid for the first 60% of the fractional release (17, 18). In our laboratory setting the gel was placed in a confined space and was not allowed to swell during the study. Plots of the initial drug release versus the square root of time should give a straight line and the diffusion coefficient can be calculated from the slope of the line. 2.5. Rheological Measurements The rheological measurements were carried out using a Bohlin VOR Rheometer (Bohlin Reologi, Lund, Sweden), a controlled strain rate instrument of the couette type, in the dynamic oscillation mode. A concentric cylinder measuring system (C14) was used and silicone oil was applied to the surface of the sample to prevent evaporation. All measurements were performed
FIG. 3. The elastic modulus, G of 1% C1342 gels containing 18 mM of the drug. The reference line shows the G of a drug-free 1% C1342 gel.
CONTROLLED DRUG RELEASE FROM GELS
elastic modulus, G , observed is higher than that for the pure C1342 gel. The increase cannot be attributed to an increased ionic strength; on the contrary, the addition of inorganic salts has previously been shown to decrease the gel strength (21). When these drugs (alprenolol and diphenhydramine) are formulated in gels that lack the lipophilic modification (e.g., C934) no strengthening effect is seen (Table 2). It seems that the increased G , accompanied by an increased viscous modulus, G , is caused by interactions between the lipophilic modification of the polymer and the nonpolar part of the drug. For the more hydrophilic atenolol, which lacks amphiphilic properties, no increase in the gel strength is observed. Uncharged, lipophilic substances lacking amphiphilic properties (e.g., butylparaben) do not affect the gel strength (10). Surprisingly, the more lipophilic drug compounds in Fig. 3, such as amitriptylin, do not affect the gel strength when formulated at the concentration 18 mM. However, as can be seen in Fig. 4, when the concentration is lowered to 9 mM an interaction is seen as an increased elasticity. The balance of the polar and lipophilic character of the drug (seen as, e.g., CMC, log P) and hence the balance of electrostatic and lipophilic interactions between the polymer and the drug can explain (7, 22) the differences observed in the concentration dependence of the interaction. This is best exemplified by alprenolol in C1342
TABLE 2 Diffusion Coefficients (Mean ± 95% Confidence Interval) and Rheological Characteristics of Gel Formulations at 1 Hz Formulation Pure C934 (1%) Pure C1342 (1%) Alprenolol (18 mM), C934 (1%) Alprenolol (18 mM), C934 (2%) Alprenolol (18 mM), C940 (1%) Alprenolol (9 mM), C1342 (1%) Alprenolol (18 mM), C1342 (1%) Alprenolol (36 mM), C1342 (1%) Alprenolol (18 mM), C1342 (0.5%) Alprenolol (18 mM), C1342 (2%) Alprenolol (18 mM), C1342 (1%), B58 (1%) Alprenolol (18 mM), Gelrite (0.5%) Alprenolol (18 mM), Gelrite (0.5%), B58 (1%) Atenolol (18 mM), C934 (1%) Atenolol (18 mM), C940 (1%) Atenolol (18 mM), C1342 (1%) Atenolol (18 mM), C934 (1%), B58 (1%) Diphenhydramine (18 mM), C934 (1%) Diphenhydramine (18 mM), C1342 (1%) Diphenhydramine (18 mM), C1342 (1%), B58 (1%)
D (10−6 cm2 /s) G (Pa) G (Pa) — — 6.31 ± 0.23 3.48 ± 0.17 4.76 ± 0.30 2.77 ± 0.14 2.78 ± 0.10 2.61 ± 0.083 6.32 ± 0.069 1.44 ± 0.10 2.29 ± 0.13
65.4 123 68.9 493 206 207 182 79.3 15.4 853 157
6.32 ± 0.16 4.80 ± 0.33
0.02–104b 0.02–104b
6.02 ± 0.26 4.49 ± 0.24 3.50 ± 0.22 6.17 ± 0.25 6.51 ± 0.27 4.00 ± 0.18 3.32 ± 0.15
65.4 237 131 68.2 55.8a 178a 114a
5.98 7.32 6.55 41.4 18.6 24.5 42.1 32.4 4.4 202 47.2
4.46 7.8 11.3 4.85 5.32a 43.2a 21.4a
a Gels prepared according to the alternative method, mixing neutralized solution with gel in the ratio 1 : 1. b Gels formed in situ during the release experiment. The interval of the elastic modulus is shown.
197
FIG. 4. The elastic modulus, G of 1% C1342 gels with varying concentrations of alprenolol, amitriptylin, and diphenhydramine. For some gels precipitation occurred ( ) and rheological characterization was not possible ().
(Fig. 4) where rheological measurements show that G is larger than for the pure gel, not only for the 18 mM alprenolol concentration as discussed above, but also for the gels with concentrations of 4.5 and 9 mM. Owing to its amphiphilic properties, alprenolol can probably assist in the formation of micelle-like structures containing the lipophilic modifications of C1342 as has been shown with surfactants (23). These aggregates may explain the increased gel strength in the presence of alprenolol. The rheological behavior of hydrophobically modified polyacrylates and oppositely charged surfactants has been described by Senan et al. (24). The hydrophobic and electrostatic interactions between C934 polymers (lacking lipophilic modification) and oppositely charged surfactants has been discussed by BarreiroIglesias et al. (25). In the gel with 36 mM alprenolol, G is lower than that in pure C1342 gels, and with 72 mM alprenolol, the gel collapsed. At this alprenolol concentration there is significant shielding of the carboxylic acid groups on the polymer, neutralizing the polyelectrolyte and impeding the swelling of the hydrogel network. An approximate calculation of the number of ionic groups of C1342 can be made using the weak acid ionization theory (Henderson–Hasselbalch equation) with pH 7.4, pK a = 6.0, and a carboxylic acid content of 57% (the number of monomers carrying free carboxylic acid according to the manufacturer). At 1% C1342 the concentration of COO− is 76 mM, and comparing this with the alprenolol concentration of 72 mM, the stoichiometric ratio [alprenolol]/[COO− ] is close to unity. The degree of binding, β (which is the ratio of the number of bound amphiphilic molecules to the number of ionic groups on the polyelectrolyte) can, however, not be determined without some knowledge of the amount of drug bound to the polymer (e.g., through equilibrium dialysis measurements). It has previously
198
PAULSSON AND EDSMAN
been reported that phase separation occurs for completely ionized linear PAA when the surfactant dodecylpyridinium is added so that β is above 0.8 (15). The concentration of lipophilic sites (alkyl groups) on the C1342 polymer is about 5–10 mM for a 1% gel. This value was obtained by performing a rough calculation based on the assumption that the 5% differences in carboxylic acid content between C1342 and C934 (stated in the U.S. National Formulary, 18th ed.) is caused by the grafting of lipophilic modifications. Hydrophobically modified polymers in concentrations on the order of 1% have been reported to form hydrophobic microdomains that can solubilize hydrophobic molecules, see, for example, the review by Piculell et al. (26). The concentration of these domains can be calculated by dividing the concentration of alkyl groups by the aggregation number of lipophilic modifications in one domain. When the concentration of the amphiphilic drug compound alprenolol is kept constant (18 mM) but the polymer concentration is varied, it can be seen (Table 3) that the gel strength is increased for both 1 and 2% C1342 gels, but slightly decreased for 0.5 and 0.75% C1342 gels in comparison to pure C1342 gels. It has previously been observed that surfactants can interact with polymers that have lipophilic modifications (LM) and both increase and decrease the elasticity of the gel. High concentrations of surfactants give an increase in the elasticity of the gel through the solubilization of many LM in one micelle and thereby forming additional cross-links. This effect was seen neither at low nor at very high concentrations since there are either too few micelles to have an effect, or there are so many micelles that there will only be one LM per micelle and no cross-linking can occur (13, 14). 3.2. Drug–Polymer Interactions: Effects on Drug Release In this section the effect the interactions discussed above have on the diffusion of a drug in a gel will be examined. The emphasis will be on alprenolol formulations. It is widely recognized that the choice of polymer can be used to control the release from a gel (7, 8, 27). In Fig. 5 the release of alprenolol and atenolol from different gels is shown. The C934 gel has the lowest cross-linking density of the three (according to the manufacturer) and gives the fastest drug release. The lipophilically modified C1342 has the highest density and gives
TABLE 3 Rheological Properties of Carbopol 1342 Gels When 18 mM Alprenolol or 1% Brij 58 Was Added Concentration (%)
Pure gel G (Pa)
C1342 with alprenolol G (Pa)
C1342 with Brij 58 G (Pa)
0.5 0.75 1 2
20.8 72.7 127 350
15.4 64.4 208 853
15.7 53.3 88.4 181
FIG. 5. The release (mean ± standard deviation, n = 3) of atenolol (open symbols) and alprenolol (closed symbols) from gels of C934 (, ), C940 (, ), and C1342 (, ).
the slowest release, while C940 with intermediate cross-linking has an intermediate release rate. When comparing the release of atenolol and alprenolol from the three different gels it can be noted that the release of the two substances does not differ in C934 and C940. In C1342, however, alprenolol has a significantly slower release than atenolol. It is likely that the interactions detected in the rheological study are the reason for the slower diffusion, as in other systems which have been suggested for controlled release formulations (7, 9). In Table 2 it can be seen that diphenhydramine has a slower release when formulated with the C1342 polymer than when C934 is used. Surprisingly, although diphenhydramine has a higher log P, the diffusion coefficient is higher than that of both atenolol and alprenolol. Perhaps the slow release of alprenolol is caused by a different kind of interaction than the one that diphenhydramine has with the polymer. Cryo transmission electron microscopy (cryo-TEM) of the C1342 gel with alprenolol could not explain this difference (images not shown). Small micelle-like structures can, however, be difficult to see in a gel using this technique. As demonstrated by the diffusion coefficients in Table 2, the release of alprenolol from gels of C934 and Gelrite is the same despite the different structure of the covalently cross-linked poly(acrylic acid) C934 and Gelrite, cross-linked in situ by cations. The amount of lipophilic modification on the polymer compared with the amount of drug will influence the release rate from the formulation. In Table 2 the release of alprenolol from C1342 gels is shown with three different polymer concentrations. The diffusion coefficient decreases when the polymer concentration is increased from 0.5 to 2% because of the increased number of sites on the polymer with which the drug can interact lipophilically and/or electrostatically. Likewise, for C934 an increase
CONTROLLED DRUG RELEASE FROM GELS
FIG. 6. The release (mean ± standard deviation, n = 3) of diphenhydramine from gels of C934 (), C1342 (), and C1342 with 1% Brij 58 ().
in polymer concentration from 1 to 2% decreases the diffusion coefficient. This confirms that either electrostatic and/or lipophilic interactions are present in the C934 and the C1342 formulations. Further information regarding the C1342–alprenolol interaction can be obtained by varying the drug concentration and keeping the polymer concentration constant. When the polymer content was 1% C1342 and the drug concentration varied, no differences were seen in the release data between gels with 9, 18, and 36 mM alprenolol. Comparison with the concentration of lipophilic modifications (5–10 mM) calculated above shows that the concentration of drug is in excess of the lipophilic domains for all concentrations used since many LM are probably clustered into the lipophilic domains. As discussed in Section 3.1, the rheological behavior changes when the drug loading is varied.
199
surfactant in the same poly(acrylic acid) gel (10). The higher log P for diphenhydramine than for alprenolol suggests a greater proportion of lipophilic interactions with Brij 58 in the formation of mixed micelles. Both drugs are, however, affected by the presence of Brij 58 to the same extent. Being highly hydrophilic, the diffusion of atenolol is not affected by the presence of Brij 58. Carbopol formulations were also made with other nonionic surfactants such as 1% Brij 700 and 1% poloxamer and the diffusion coefficients calculated. These formulations showed only slightly—and not statistically significant—lower diffusion coefficients so the data are not shown. In Fig. 6 the release of diphenhydramine from the C1342 gel with Brij 58 is slower than from the C1342 gel without surfactant, but after 180 min the curves intersect and the Brij 58 formulation has a faster release rate. A plausible explanation could be that the fraction of drug that interacts with the surfactants also is released together with the surfactants, whereas the fraction of drug that interacts with the LM-PAA polymer matrix has a longer distance over which to diffuse since the matrix is stationary. The fraction of the drug that interacts with the LM is probably smaller when Brij 58 micelles are present. As the micelles are released from the gel this difference is shown by the crossover of the curves after 180 min. The addition of nonionic surfactants to the C1342 formulations in this study only brought about minor differences. In Gelrite gels the addition of 1% Brij 58 decreased the diffusion coefficient from 6.3 × 10−6 to 4.8 × 10−6 cm2 /s. In this gel there are no lipophilic modifications that can participate in lipophilic interactions with the drug, and the addition of micelles gives a greater reduction than it did in the C1342 formulations. With these formulations, for which the interaction with the polymer or the formation of mixed micelles is used to give controlled release, charged surfactants are avoided, and thereby the
3.3. Addition of Surfactants By including micelle-forming surfactants in the formulation, the diffusion of drug molecules can be further reduced since lipophilic interactions can take place between the drug and both the polymer and micelle. Brij 58 is a nonionic surfactant with a large polyoxyethylene shell. The CMC is 0.007 mM (according to the manufacturer), and in the presence of C934 the CMC is lowered to CAC, which is likely half of the CMC (based on data for nonionic surfactants in C934 (25)). Compared to pure C1342 gels, both diphenhydramine (Fig. 6) and alprenolol have significantly slower release rates when Brij 58 micelles are present (Table 2). The amphiphilic drugs can probably interact with and form mixed micelles with Brij 58, resulting in a slower diffusion. The solubilization of uncharged, lipophilic model substances in 1% Brij 58 has previously been reported to give a diffusion coefficient 20 times lower than when they are formulated without
FIG. 7. The effect of Brij 58 on the elastic modulus, G , () and the phase angle, δ, () of 1% C1342 gels.
200
PAULSSON AND EDSMAN
toxicity is possibly reduced. However, these formulations have a smaller effect on the release than was seen when oppositely charged surfactants were used as described by Paulsson and Edsman (11). This reference describes a totally different way to control the release based on the formation of vesicles. The addition of surfactants often affects the gel strength, and in Fig. 7 and Table 3 it can be seen that when Brij 58 is added, the gel strength is decreased for all polymer concentrations. The presence of these surfactants seems to disturb the structure of the gel matrix, thereby decreasing the elasticity (G ) of the gel. The Brij 58 concentration is well above the CMC (low CMC ≈ 0.007 mM, according to the manufacturer) and thus does not contribute to the cross-linking of the gel as described by J¨onsson et al. (13) and discussed above. 4. SUMMARY
The interaction between lipophilically modified PAA and drug compounds can be measured using rheology and can be predicted to some degree by considering the lipophilicity of the drug. However, the ratio of lipophilic modification of polymer to drug concentration also seems to play a major role. The interaction between lipophilically modified PAA and amphiphilic drugs can be used for constructing controlled release gel formulations. It is also possible to sustain the release of charged drugs with adequate lipophilicity by adding nonionic surfactants with which the drugs can form mixed micelles; however, these effects are only minor in lipophilically modified gels. From a toxicological point-of-view, formulations without surfactants are often preferable. ACKNOWLEDGMENTS The excellent experimental skills of Mr. Stefano Bonfante and Mr. G¨oran Karlsson (who produced the cryo-TEM images) are gratefully acknowledged.
REFERENCES 1. Saettone, M. F., Giannaccini, B., Ravecca, S., LaMarca, F., and Tota, G., Int. J. Pharm. 20, 187 (1984). 2. Zhou, M., and Donovan, M. D., Int. J. Pharm. 135, 115 (1996). 3. Desai, S. D., and Blanchard, J., Drug. Deliv. 7, 201 (2000). 4. Di Colo, G., Carelli, V., Giannaccini, B., Serafini, M. F., and Bottari, F., J. Pharm. Sci. 69, 387 (1980). 5. Paavola, A., Kilpel¨ainen, I., Yliruusi, J., and Rosenberg, P., Int. J. Pharm. 199, 85 (2000). 6. Bochot, A., Fattal, E., Grossiord, J. L., Puisieux, F., and Couvreur, P., Int. J. Pharm. 162, 119 (1998). 7. Sj¨oberg, H., Persson, S., and Caram-Lelham, N., J. Control. Release 59, 391 (1999). 8. am Ende, M. T., and Peppas, N. A., J. Control. Release 48, 47 (1997). 9. Inoue, T., Chen, G., Nakamae, K., and Hoffman, A. S., J. Control. Release 49, 167 (1997). 10. Paulsson, M., and Edsman, K., J. Pharm. Sci. 90, 1216 (2001). 11. Paulsson, M., and Edsman, K., Pharm. Res. 18, 1586 (2001). 12. Goddard, E. D., J. Soc. Cosmet. Chem. 41, 23 (1990). 13. J¨onsson, B., Lindman, B., Holmberg, K., and Kronberg, B., “Surfactants and Polymers in Aqueous Solution.” Wiley, Chichester, England, 1998. 14. Magny, B., Iliopoulos, I., Zana, R., and Audebert, R., Langmuir 10, 3180 (1994). 15. Shimizu, T., Colloid Surf. A-Physicochem. Eng. Asp. 94, 115 (1995). 16. Stjernschantz, J., and Astin, M., in “Biopharmaceutics of Ocular Drug Delivery” (P. Edman, Ed.), p. 1. CRC Press, Boca Raton, FL, 1993. 17. Higuchi, W. I., J. Pharm. Sci. 51, 802 (1962). 18. Ritger, P. L., and Peppas, N. A., J. Control. Release 5, 23 (1987). 19. Paulsson, M., H¨agerstr¨om, H., and Edsman, K., Eur. J. Pharm. Sci. 9, 99 (1999). 20. Almgren, M., Edwards, K., and Karlsson, G., Colloid. Surf. A 174, 3 (2000). 21. Kriwet, B., and Kissel, T., Int. J. Pharm. 127, 135 (1996). 22. Lowe, T. L., Virtanen, J., and Tenhu, H., Polymer 40, 2595 (1999). 23. Piculell, L., Thuresson, K., and Ericsson, O., Faraday Disc. 101, 307 (1995). 24. Senan, C., Meadows, J., Shone, P. T., and Williams, P. A., Langmuir 10, 2471 (1994). 25. Barreiro-Iglesias, R., Alvarez-Lorenzo, C., and Concheiro, A., J. Control. Release 77, 59 (2001). 26. Piculell, L., Guillemet, F., Thuresson, K., Shubin, V., and Ericsson, O., Adv. Colloid Interface Sci. 63, 1 (1996). 27. Zentner, G. M., Rathi, R., Shih, C., McRea, J. C., Seo, M. H., Oh, H., Rhee, B. G., Mestecky, J., Moldoveanu, Z., Morgan, M., and Weitman, S., J. Control. Release 72, 203 (2001).