el formation in systems composed of drug containing catanionic vesiclesnd oppositely charged hydrophobically modified polymer
oel Dewa,∗, Katarina Edwardsb, Katarina Edsmana,1
Department of Pharmacy, Uppsala University, Box 580, SE-75123 Uppsala, SwedenDepartment of Physical and Analytical Chemistry, Uppsala University, Box 579, SE-751 23, Uppsala, Sweden
r t i c l e i n f o
rticle history:eceived 2 October 2008eceived in revised form 10 December 2008ccepted 14 December 2008vailable online 24 December 2008
eywords:
a b s t r a c t
The aim of this study was to explore if mixtures of drug containing catanionic vesicles and polymers giverise to gel formation, and if so, if drug release from these gels could be prolonged. Catanionic vesiclesformed from the drug substances alprenolol or tetracaine, and the oppositely charged surfactant sodiumdodecyl sulphate were mixed with polymers. Three polymers with different properties were employed:one bearing hydrophobic modifications, one positively charged and one positively charged polymer bear-ing hydrophobic modifications. The structure of the vesicles before and after addition of polymer was
atanionicesicleelolymerrolonged drug release
investigated by using cryo-TEM. Gel formation was confirmed by using rheological measurements. Drugrelease was studied using a modified USP paddle method. Gels were observed to form only in the casewhen catanionic vesicles, most likely with a net negative charge, were mixed with positively chargedpolymer bearing lipophilic modifications. The release of drug substance from these systems, where thevesicles are not trapped within the gel but constitute a founding part of it, could be significantly pro-longed. The drug release rate was found to depend on vesicle concentration to a higher extent than onpolymer concentration.
. Introduction
Gels have a variety of pharmaceutical uses; ocular, cutaneousnd nasal to name a few. Due to mucoadhesion [1] or because ofheir rheological properties [2], gels usually have a considerablyonger contact time with mucosa than aqueous solutions. The pro-onged contact time can only be advantageous if the drug substances released from the gel throughout the contact time. As the gel typi-ally contains at least 95% water, free drug molecules usually diffusehrough the gel as quickly as through pure water. Therefore, to takedvantage of a prolonged contact time the release rate of the drugubstance from the gel needs to be prolonged. Prolonged release for-ulations may increase the bioavailability of the drug compound
nd/or reduce dosing frequency and thereby increase patient com-liance. There has been many strategies to prolong the release fromel formulations; either by formulating the drug substance as a sus-
ension in the gel [3], distributing the substance to micelles [4] or
iposomes [5] in the gel, or letting the drug substance interact withhe gel forming polymer [6]. A more recent way of prolonging theelease of drug substances from gels is by incorporating so called
catanionic aggregates, formed by drug substance and an oppositelycharged surfactant in the gel [7–11].
Catanionic aggregates are formed spontaneously when solu-tions of a variety of oppositely charged surfactants are mixed[12,13]. The surfactant mixtures have a complex phase behaviorand formation of vesicles and different types of micelles have beenobserved.
Polymers may stabilize vesicles and function as release rate con-trollers in drug delivery [14]. There are several possible ways ofinteraction between the vesicles and the polymers; hydrophobicparts of the polymer can interact with hydrophobic parts of thevesicle [15–17], in charged systems there can be electrostatic inter-actions [18,19], or there may be a combination of these types ofinteractions [18,19]. Cross-links formed as a consequence of theinteractions may lead to gel formation.
Mixtures of polymers or polyelectrolytes and oppositelycharged vesicles have been studied in the past. Interactions inthese complex systems give rise to interesting phenomena that canbe utilized in pharmaceutics, paint and the cosmetics industry. Thepolymers used may be of different origin, bear different charges,
and have a large variety of modifications. Polymers investigated sofar have been of both synthetic and biological origin [16,20]. Someof the latter systems have been suggested as models for livingcells [21]. Previous studies have involved vesicles formed fromsurfactants of classical type; in this study we have explored the
Fig. 1. Molecular structures of (a) JR-400, positively charged polymer and (b)
88 N. Dew et al. / Colloids and Surfa
se of a conventional surfactant in combination with an oppositelyharged, surface active, drug substance.
In earlier studies by this research group the drug containingatanionic aggregates, have been incorporated within a conven-ional, preformed, gel. Usually, a covalently cross-linked gel haseen used, and the catanionic aggregates have then been enclosed
n these structures [7–11]. The conventional gels typically remainntact at the site of application also after the drug substance hasscaped the gel and been absorbed. In this study we explore the pos-ibilities of using the interactions between the catanionic vesiclesnd the polymers for the actual gel formation. Since the catanionicesicles most likely form the cross-links that give rise to the gel for-ation, the gel is expected to break down as the drug substance and
he oppositely charged surfactant diffuses from the vesicles. Theseystems are interesting from a pharmaceutical point of view, sincefter the active ingredient is removed from the vehicle there is notn “empty” gel left at the application site.
. Materials and methods
.1. Materials
Sodium chloride, sodium dodecyl sulfate (SDS), alprenololydrochloride and tetracaine hydrochloride was purchased fromigma Chemical Co. (St. Louis, MO, USA). All substances were of ana-ytic grade or “Ultra” quality. The cationic polymers UCARE JR-400nd SoftCAT SK-MH were kind gifts of The Dow Chemical Company.he uncharged polymer HM(C16-18)-PEG was a kind gift of Akzoobel Sweden. The polymers were used as received, without fur-
her purification. Carbopol polymers 940 and 1342 were kind giftsf Noveon, Inc. (Breeksville, OH, USA). Millipore water (Millipore,rance) was used in all experiments.
JR-400 and SK-MH are hydroxyethylcellulose derivates; JR-400s an N,N,N-trimenthylammonium derivative and SK-MH is an,N-dimethyl-N-dodecylammonium derivative, structures are dis-layed in Fig. 1a and b. JR-400 has a higher degree of cationicubstitution and charge density than SK-MH, JR-400 has a nitro-en content of 1.5–2.2% [22] and SK-MH has a nitrogen contentf 0.8–1.10% [23]. HM(C16-18)-PEG is an uncharged poly (ethylenelycol), modified with hydophobic hydrocarbon chains comprisedf C16–18, the hydrophilic part between these hydrocarbon chainsre composed of approximately 280 oxyethylene units, the struc-ure displayed in Fig. 1c. Carbopol 1342 has a lipophilic graft on itsackbone, a long chain alkyl acrylate [24].
.2. Sample preparation and phase studies
.2.1. Catanionic phase studyThe first part of the phase studies consisted of exploring the
hase behavior in mixtures of oppositely charged surfactants,here one of the surfactants is a drug substance. Catanionic aggre-
ate mixtures were prepared by mixing solutions of drug substancend oppositely charged surfactant. The proportions of drug sub-tance and surfactant were varied, keeping the total concentrationonstant; then the total concentrations were varied between 5nd 160 mM. The total surfactant concentration will henceforth beeferred to as the vesicle concentration. All mixtures were madeith 0.9% sodium chloride solution. The mixtures were allowed to
quilibrate for at least 48 h before visual inspection, allowing phaseeparated mixtures to set. The physical long-term stability wasontrolled after 5 months. To confirm findings in the visual study
olutions containing vesicles were investigated using cryogenicransmission electron microscopy (cryo-TEM). The phase behaviorf tetracaine and SDS has previously been studied thoroughly byhis research group [9] and those results were used to find suit-ble ratios and concentrations where vesicle content was observed
molecular structure of SK-MH, positively charged polymer bearing lipophilic mod-ifications. Both redrawn from DOW product information; (c) HM(C16-18)-PEG. nis approximately 280, R is 16–18 and m is not disclosed, according to Akzo NobelSweden.
in non-phase separated samples. A more detailed presentation ofthose results can be found elsewhere [9]. The alprenolol and SDSphase table explored and used in this study is presented in thiswork.
2.2.2. Vesicle–polymer gel formation studyTo study if gel formation could occur, catanionic vesicle contain-
ing solutions and polymer solutions were mixed. Three differentpolymers were used; one bearing positive charges, one unchargedpolymer bearing lipophilic modifications, and one polymer bear-ing lipophilic modifications and positive charges. These variationswere included to show which type, or types, of properties thatare important for gel formation. In the vesicle–polymer study thevesicle containing solutions were mixed with polymer solutionsof different concentrations, both of double the intended final con-centrations, so that when equal volumes were mixed the desiredconcentrations were obtained. These samples were mixed withmagnetic stirrers for several days before evaluation as the polymersolutions were very viscous. Samples containing large amounts oftrapped air were centrifuged before evaluation. Mixtures appear-ing to have resulted in gel formation were examined further usingrheological measurements and cryo-TEM.
2.3. Cryogenic transmission electron spectroscopy
Cryo-TEM was used to confirm the results of the visual phasestudies. A drop of the sample was deposited on a holey poly-mer film covered grid. Excess sample was blotted away with filter
paper and the remaining liquid was vitrified by plunging the gridinto liquid ethane held at a temperature of −170 ◦C. After trans-fer to a Zeiss EM 902 transmission electron microscope the filmswere kept below −165 ◦C for the viewing process. All observationswere made in the zero-loss bright-field mode at an accelerating
ces B:
v[
2
Vrcmsmtvo(m
2
moamacTPTtsctrtwcmtuemleabf
Q
wtcstpr
tcc9mdo
N. Dew et al. / Colloids and Surfa
oltage of 80 kV. Details of this method can be found elsewhere25].
.4. Rheology
Viscoelastic measurements were carried out using a BohlinOR rheometer (Bohlin reologi, Lund, Sweden), a controlled rateheometer of couette type. The measuring system used was a con-entric cylinder (C14). All samples were centrifuged before theeasurements to avoid trapped air. Silicon oil was applied to the
ample surface which prevented evaporation of fluid during theeasurements. Measurements were carried out at 20 ◦C and up
o 37 ◦C. Oscillation measurements were carried out in the lineariscoelastic regions, which were found by strain sweeps. The rhe-logical properties are presented as the elastic (G’) and viscousG”) modulus, the elastic module is also referred to as the storage
odule and the viscous module as the loss module.
.5. Release study
The release of drug substances from gels was studied using aodified USP paddle method. Three measurements were carried
ut on each sample. Gel containers with a volume of 6 cm3 andsurface area of 21 cm2 were filled with gel and covered with aesh size plastic net and a coarse plastic net; graphical details
bout this equipment can be found elsewhere [10]. The filled gelontainers were immersed in 200–500 mL of 0.9% NaCl solution.hese were stirred at 20 rpm and maintained at 37 ± 0.5 ◦C using aharma Test PTW II USP bath (Pharma Test Apparatebau, Germany).he volume of receiving medium was chosen so that sink condi-ions were kept throughout the entire experiment and so that thepectrophotometrical measurements could be done within suitableoncentration intervals. By peristaltic pumping through ismapremeubing (Ismatech, SA, Zurich, Switzerland) the saline solution on theeceiving side was continuously pumped through the UV–vis spec-rophotometer (Shimadzu UV-1601, Shimadzu, Kyoto, Japan). Theavelengths used were 271 nm for alprenolol and 310 nm for tetra-
aine. The release of drug substance was studied for 13.5 h withore frequent measurements during the first 40 min, where less
han 60% of the drug hade been released, and these values weresed to calculate the apparent diffusion coefficients, D. The appar-nt diffusion coefficient measures the rate at which a substanceoves randomly from a region of high concentration to a region of
ower concentration. These calculated diffusivities depend on sev-ral factors such as vesicle deterioration, drug loading in aggregatesnd the actual diffusivities. The apparent diffusion coefficient cane calculated using Eq. (1) which describes the Fickian diffusionrom a gel during sink conditions:
= 2C0
(Dt
�
)1/2(1)
here Q is the amount of drug released per unit area, C0 the ini-ial concentration of the drug in the gel, D the apparent diffusionoefficient of the drug in the gel and t is the time elapsed since thetart of the experiment. This equation is valid for the first 60% ofhe release [26,27]. For each sample a control for a linear fit waserformed when plotting the released amount versus the squareoot of time.
To be able to compare the release of drug substances fromhe catanionic vesicle–polymer gels obtained in the phase study,arbopol reference gels were used. The effect of lipohpilic modifi-
ations was evaluated by using two different carbopols: carbopol40, which is unmodified and carbopol 1342, which bears lipophilicodifications. Carbopol gels containing catanionic vesicles or pure
rug substance were prepared by mixing stock solutions and gelsf double the intended final concentration. The 2% Carbopol 940
Biointerfaces 70 (2009) 187–197 189
and 1342 gels were prepared with 0.9% NaCl-solution and set atpH = 7.4 ± 0.1 before mixed (1:1) with the vesicle solution, render-ing a 1% gel. The pH was then controlled again and adjusted to7.4 ± 0.1, when needed.
To assure that the SK-MH polymer did not have any UV-absorbance an absorbance spectrum was obtained for a polymersolution. The polymer concentration of this solution was the high-est possible in the receiving medium, should all polymer enclosedin the gel container escape. The polymer did not interfere with thedrug concentration measurements at the wavelengths used in thisstudy.
2.6. X-ray measurements
Selected gel samples were investigated with X-ray measure-ments to elucidate whether or not cubic or highly ordered micellarstructures were present in the samples. Details about the methodcan be found elsewhere [28].
2.7. Statistical analysis
The apparent diffusion coefficients were analyzed statisticallyusing ANOVA and as a post hoc test Bonferroni’s multiple com-parisons with a significance level (p < 0.05) was considered asstatistically significant. Prism 4 for Windows, by GraphPad SoftwareInc. (San Diego, CA, USA) was the software used. The confidenceinterval (CI) for each group of samples is given.
3. Results
3.1. Phase behavior and gel formation
3.1.1. Catanionic phase behaviorSpontaneous vesicle formation was observed in many of the
mixtures of alprenolol and SDS, at high SDS ratios, as presented inFig. 2. Vesicles were noted at several different ratios of alprenololand SDS, for all concentrations examined. In the higher concentra-tion ranges several viscous solutions were observed, indicating thepresence of elongated micelles. Mixtures resulting in vesicle for-mation were used in the vesicle–polymer gel formation study. Themixtures were kept at ambient room temperature for 5 months andthe samples were then checked visually. The long-term stability wasgenerally good; the vesicle containing samples had not separated.
3.1.2. Polymer–surfactant interactionThe interactions between each of the surface active substances
and the SK-MH polymer were studied. Only the SK-MH polymerwas used, as gel formation was not observed with either of theother polymers, see Section 3.1.3. When a 20 mM solution of SDSwas added to a 2% SK-MH solution a very viscous, and slightly white,phase in equilibrium with a non-viscous phase resulted. The viscousphase is most likely constituted by polymers and strongly associ-ated SDS while the non-viscous phase is saline solution with a smallamount of surfactant. Polymer solution mixed with alprenolol ortetracaine at 20 mM did not result in phase separation. The rhe-ological investigation, see Section 3.3, showed, however, that theelastic and viscous properties of the polymer were affected uponaddition of the drug substances.
3.1.3. Vesicle–polymer gel formation
Catanionic vesicles were mixed with polymer solutions in order
to explore the potential for gel formation. Three different polymerswere used in this study: one uncharged hydophobically modifiedHM(C16-18)-PEG polymer, one positively charged JR-400 polymerand one positively charged SK-MH polymer bearing hydrophobic
190 N. Dew et al. / Colloids and Surfaces B: Biointerfaces 70 (2009) 187–197
F tionsV mpler
mbfcwilsTbSc
ga7St
3
a
Fas
ig. 2. Phase behavior of alprenolol and SDS mixtures. Unmarked areas signify soluesicles are noted in text. Viscous samples are marked �. The symbol � signifies saise to a low degree of opaqueness.
odifications. Mixtures of the JR-400 polymer, lacking hydropho-ic modifications, and catanionic vesicles did not result in gelormation. The uncharged HM(C16-18)-PEG polymer produced vis-ous samples when mixed with catanionic vesicles, but noneere gel-like. Gels were formed, however, upon mixing the catan-
onic vesicles with the positively charged SK-MH polymer bearingipophilic grafts. Based on the above-described result subsequenttudies were focused on mixtures containing the SK-MH polymer.he results displayed in Fig. 3 show that a critical concentration ofoth polymer and vesicles are required for gel formation to occur.elected samples of pure and mixed samples of tetracaine/SDS vesi-les and SK-MH polymer are displayed in Fig. 4.
The pH of the tetracaine/SDS catanionic vesicle–SK-MH polymerels were in the range of 6.8–7.0. The alprenolol containing gels hadpH ranging from 7.0 to 7.3. Adjustments to a physiological pH of
.4 ± 0.1 did not affect phase behavior or rheology in the systems.amples were kept at ambient room temperature for 5 months, andhe long-term stability of the gels was good, as examined visually.
.2. Cryo-TEM
Solutions containing alprenolol or tetracaine, and SDS, as wells gels containing drug substance, SDS and the polymer SK-MH
ig. 3. Phase behavior of (a) alprenolol/SDS vesicles (in a 30:70 ratio) mixed with SK-MHnd (c) tetracaine/SDS vesicles (in a 35:65 ratio) mixed with SK-MH polymer. Unmarked aeparated samples. Gel samples are noted in text. The symbol – indicates an unstudied m
, gray areas signify phase separated samples and black areas signify precipitations.s where a visual inspection indicated vesicle content, but low concentrations gave
were all characterized using cryo-TEM. The presence of vesicles wasconfirmed both before and after addition of the SK-MH polymers.Alprenolol/SDS systems displayed spherical bi- and multi-lamellarvesicles. Elongated and invaginated bilayer structures were alsopresent. The polymer free systems are displayed in Fig. 5a and c.As SK-MH polymer was added there was a reduction in size ofthe lamellar structures. Further, the fraction of elongated bilayerstructures increased. Typical structures are displayed in Fig. 5b andd.
The micrograph displayed in Fig. 6a shows uni-, bi- and multi-lamellar vesicles present in the tetracaine/SDS systems. Frequently,these vesicles displayed large pores, or holes, in the membrane. InFig. 6b the SK-MH polymer containing system is displayed. It con-tained uni-lamellar vesicles and large lamellar sheets. It should benoted that the polymer, due to poor contrast, cannot be visualizedby the cryo-TEM technique employed in the present study.
3.3. Rheology
The rheological properties of the viscous, high concentration,SK-MH polymer solutions were investigated. The solutions dis-played rheological profiles typical for entangled polymer solutions.At 1% of SK-MH polymer (in 0.9% NaCl solution) and 1 Hz G′ was
polymer, (b) alprenolol/SDS vesicles (in a 40:60 ratio) mixed with SK-MH polymerreas signify solutions, black areas signify precipitations and gray areas signify phaseixture.
N. Dew et al. / Colloids and Surfaces B:
Fig. 4. Photographs of samples from the phase table. The bottom row displays thesamples inverted. (a) SK-MH polymer solution (2%); (b) tetracaine/SDS vesicles (ata 35:65 ratio) at 20 mM and SK-MH polymer at 0.1%; (c) tetracaine/SDS vesicles (ata 35:65 ratio) at 20 mM and SK-MH polymer at 0.2%; (d) tetracaine/SDS vesicles(at a 35:65 ratio) at 20 mM and SK-MH polymer at 1%; (e) tetracaine/SDS vesicles(at a 35:65 ratio) at 20 mM and SK-MH polymer at 2%; (f) tetracaine/SDS vesicles(sa
1tpGc1t
cAqsfinsmwtlF
bpvvmmset
3.5.1.2. Alprenolol systems. The release rate of drug substance wassignificantly prolonged from all alprenolol vesicle containing gels,
at a 35:65 ratio) at 20 mM without polymer. Samples b and c contain precipitates;amples d and e are gels. This photo was obtained shortly after mixture of vesiclend polymer solutions, and homogeneity of the gel samples increase with time.
10 Pa and G′ ′ was 76 Pa. The rheological effects upon addition ofhe model drug substances were also investigated. At a 1% SK-MHolymer concentration and with an addition of 20 mM of alprenolol′ was lowered to 71 Pa and G′ ′ to 57 Pa, at 1 Hz. However, as tetra-aine was added G′ was increased to 358 Pa and G′ ′ to 271 Pa, atHz. The change in rheology upon addition of catanionic vesicles
o a polymer solution is illustrated in Fig. 7.Rheological measurements were carried out on all gel-like
atanionic vesicle–SK-MH polymer samples from the phase study.weak gel structure was confirmed, i.e. G′ � G′ ′ with a slight fre-
uency dependence. Both the elastic and viscous modules increaselightly with frequency. An example of a typical rheological pro-le is displayed in Fig. 8. Macroscopically, the gel structures wereot weak; the samples displayed high rigidity and would stay foreveral hours in the top part of an overturned vial. The rheologicaleasurements were carried out at 20 ◦C and when the temperatureas raised in steps to 37 ◦C the samples still displayed gel proper-
ies. As the temperature was risen the G′ and G′ ′ values were slightlyowered, as could be expected. An example of this is displayed inig. 9.
In the alprenolol containing systems, gel strength increases withoth polymer concentration and total concentration of vesicles. Theolymer concentration has a greater impact on the increase than theesicle concentration. The rheological behavior of alprenolol/SDSesicle–SK-MH polymer gels is illustrated by showing the elasticodulus G′ at 1 Hz, see Fig. 10a and b. In most investigated poly-
er concentrations there seems to be a maximum in elasticity; the
amples with the highest vesicle concentrations typically had lowerlasticities than the ones of half the alprenolol/SDS vesicle concen-ration. In the alprenolol system there was also a curiosity; when
Biointerfaces 70 (2009) 187–197 191
changing the ratio of alprenolol in the catanionic aggregates from40% to 30% the elasticity increased, in some cases.
In the mixtures containing tetracaine the gel strength alsoincreased with both polymer concentration and total concentra-tion of vesicles, see Fig. 10c. The effect on the elasticity dependedto a higher extent on the polymer concentration than on the vesi-cle concentration. There seemed to be a maximum in the elasticity,as the elasticity decreased in the samples with the highest vesi-cle concentration. However, a larger number of samples need to beanalyzed in order to confirm this behavior.
3.4. X-ray measurements
Other research groups have studied systems of polyelectrolyteand catanionic surfactant where a highly ordered structure ofmicelles appear [29]. Several of the tetracaine mixtures appeared tohave a very high viscosity, a clear transparent appearance indicat-ing absence of vesicles, and possibly a characteristic ringing sound,indicating a cubic phase. However, investigations using small angleX-ray spectroscopy, SAXS, did not indicate cubic structures in anyof the investigated vesicle–SK-MH polymer mixtures (results notshown).
3.5. Drug release study
A selection of vesicle–SK-MH polymer gels was used in therelease study. The choice of samples made comparisons in releaserate from samples differing in catanionic vesicle concentrations andpolymer concentrations possible. To compare the release rate fromconventional gels, with and without lipophilic modifications, refer-ence samples with drug substance only or in vesicles with SDS wereprepared using carbopol 940 or 1342 gels. The apparent diffusioncoefficients from all investigated systems are displayed in Table 1.Typical release profiles are displayed in Fig. 11, where the release ofalprenolol from vesicle–SK-MH gels and carbopol gels containingalprenolol with and without SDS are compared.
3.5.1. Drug release from vesicle–SK-MH gels3.5.1.1. Tetracaine systems. The release of tetracaine was prolongedfrom all gel systems containing catanionic vesicles. In the catanionicvesicle–SK-MH polymer gel with the lowest vesicle concentrationsthe interactions are only strong enough to form a very weak gel; therelease from this gel was therefore the least prolonged. To inves-tigate the effect of changes in polymer and vesicle concentrationa series of measurements were carried out, in which the releaseof tetracaine was studied as either the polymer or vesicle concen-tration was kept constant while the other was varied. The releaseof drug substance was prolonged as the vesicle concentration wasincreased from 20 mM to 80 mM, which is illustrated in Fig. 12a.There was no significant difference in release rates between the gelscontaining 40 mM and 80 mM of catanionic vesicles. The releasefrom the gel containing 20 mM catanionic vesicles, however, wassignificantly different from that of the gels with higher vesicle con-tent (p < 0.001). As illustrated in Fig. 12b, the drug release rate wassomewhat prolonged when increasing the SK-MH polymer concen-tration from 1% to 2%. The changes in apparent diffusion coefficientsare illustrated in Fig. 13.
see Table 1. The release rate of alprenolol from the catanionicvesicle–SK-MH polymer gels was not significantly different fromthe reference carbopol gels with catanionic vesicles at any of theinvestigated concentrations (p < 0.001).
192 N. Dew et al. / Colloids and Surfaces B: Biointerfaces 70 (2009) 187–197
Fig. 5. cryo-TEM micrographs of alprenolol/SDS (in a 40:60 ratio) vesicles: (a) with total concentration of 80 mM in saline solution and (b) the same conditions but in a 2%SK-MH polymer solution, (c) with total concentration of 20 mM in saline solution and (d) shows the same conditions but in a 1% SK-MH polymer solution. Examples of icecrystals deposited on the surface of the vitrified sample indicated with arrows.
Fig. 6. cryo-TEM micrographs of tetracaine/SDS (in a 35:65 ratio) vesicles: (a) with a total concentration of 80 mM in saline solution and (b) shows the same conditions butin a 2% SK-MH polymer solution. Some large openings in the vesicles are indicated with arrows.
N. Dew et al. / Colloids and Surfaces B: Biointerfaces 70 (2009) 187–197 193
Fig. 7. Rheological profiles of a 1% SK-MH polymer solution, marked with solid lines,and a 1% SK-MH polymer solution with alprenolol/SDS vesicles (at a 10 mM concen-tration and a 30:70 ratio) added, marked with hatched lines. Circular shapes signifythe elastic module G′ and triangular shapes signify the viscous module G′ ′ .
Fig. 8. A typical rheological profile of a catanionic vesicle–SK-MH polymer gel sam-ple. This sample contains 20 mM of alprenolol/SDS (in a 40:60 ratio) vesicles and1% of SK-MH polymer. (�) Signifies the elastic modulus G′ , (�) signifies the viscousmodulus G′ ′ and (�) signifies the phase angle.
Table 1Apparent diffusion coefficient of drug substance from carbopol 940, carbopol 1342 and ca
a The sample had a p < 0.001 when compared to the carbopol 940 reference sample and* p < 0.05 (statistical difference from the carbopol reference samples).
** p < 0.01 (statistical difference from the carbopol reference samples).*** p < 0.001 (statistical difference from the carbopol reference samples).
Fig. 9. Rheological properties of a tetracaine/SDS vesicle (20 mM concentration anda 35:65 ratio)–SK-MH polymer gel at the frequency 1 Hz at different temperatures.(�) Signifies the elastic modulus G′ , (�) signifies the viscous modulus G′ ′ and (�)signifies the phase angle.
3.5.2. Comparisons with carbopol gels3.5.2.1. Tetracaine systems. In systems containing 5 mM of tetra-caine and SDS, the release of tetracaine from the catanionicvesicle–SK-MH polymer gel was statistically different from therelease of tetracaine from catanionic vesicles in carbopol 940 butnot from catanionic vesicles in carbopol 1342, see Table 1. Atthe 20 mM and 80 mM tetracaine/SDS concentration levels thedifferences were not significant. The fraction of released drug sub-stance at the end of the experiment is somewhat higher from thevesicle–SK-MH gels than from the carbopol gels. At all concentra-tion levels the release of tetracaine from the reference samples, i.e.
samples lacking SDS, was significantly faster than from gels contain-ing catanionic vesicles (p < 0.001 except one case where p < 0.05).
3.5.2.2. Alprenolol systems. In the case of alprenolol, the apparentdiffusion coefficients of systems with the same vesicle concentra-
tanionic vesicle–SK-MH polymer gels.
Gel Polymer conc. D (cm2/s) CI
C940 1% 5.57 × 106 2.60 × 10−7
C1342 1% 3.34 × 106 1.23 × 10−7
SKMH 1% 2.77 × 10−7*** 8.14 × 10−8
C940 1% 2.08 × 10−7*** 6.45 × 10−9
C1342 1% 2.85 × 10−7*** 7.64 × 10−8
SKMH 1.5% 2.32 × 10−7*** 4.37 × 10−8
C940 1% 2.27 × 10−7*** 7.06 × 10−8
C1342 1% 2.40 × 10−7*** 6.65 × 10−8
SKMH 2% 4.46 × 10−8*** 5.26 × 10−9
C940 1% 7.97 × 10−8*** 3.48 × 10−8
C1342 1% 4.18 × 10−8*** 1.04 × 10−8
C940 1% 4.02 × 10−6 5.49 × 10−7
C1342 1% 2.38 × 10−6 1.04 × 10−7
SKMH 1% 1.20 × 10−6***/** ,a 2.39 × 10−7
C940 1% 2.78 × 10−8*** 6.20 × 10−8
C1342 1% 6.11 × 10−7*** 5.73 × 10−8
SKMH 1% 2.11 × 10−7*** 3.49 × 10−8
SKMH 1.5% 1.84 × 10−7*** 1.93 × 10−8
C940 1% 6.65 × 10−8*** 9.21 × 10−9
C1342 1% 1.53 × 10−7*** 1.66 × 10−8
SKMH 2% 1.60 × 10−7*** 1.79 × 10−8
SKMH 2% 4.33 × 10−8*** 7.05 × 10−9
SKMH 2% 3.06 × 10−8*** 8.76 × 10−9
C940 1% 9.91 × 10−9*** 1.25 × 10−9
C1342 1% 1.38 × 10−8*** 1.22 × 10−9
a p < 0.01 when compared to the carbopol 1342 reference sample.
194 N. Dew et al. / Colloids and Surfaces B: Biointerfaces 70 (2009) 187–197
Fig. 10. Elastic modulus, G′ , at 1 Hz in samples of (a) alprenolol/SDS (in a 40:60ratio) and SK-MH polymer gels. Note that at 80 mM there is no gel phase present at1% SK-MH; (b) alprenolol/SDS (in a 30:70 ratio) and SK-MH polymer gels. Note thatat 20 and 40 mM there are no gel phase present at 1% SK-MH; (c) tetracaine/SDS (ina 35:65 ratio) and SK-MH polymer gels. Note that at 40 and 80 mM there are no gelphases present at 1% SK-MH, and not at 80 mM at 1.5% SK-MH.
Fig. 11. Comparison of release rates of 20 mM alprenolol from gels: (�) 20 mM from1% carbopol 940 gel and (×) 40 mM from carbopol 1342 gel. Release of alprenololfrom 80 mM alprenolol/SDS catanionic vesicles (at a 40:60 ratio): (�) from a 2% SK-MH polymer gel; (�) from a 1% carbopol 940 gel; (�) from a 1% 1342 carbopol gel.Bars indicate standard deviation.
Fig. 12. Effects on release rates of tetracaine from catanionic tetracaine/SDS (at a35:65 ratio) vesicle–SK-MH polymer gels: (a) SK-MH polymer concentration 2% and
vesicle concentrations: (�) 20 mM, (�) 40 mM and (�) 80 mM. (b) Vesicle concen-tration 20 mM and polymer concentrations: (�) 1% (�) 1.5% and (�) 2%. Bars indicatestandard deviation.
tions were not significantly different. However, as in the tetracainesystems, the fraction of released drug substance at the end ofthe experiments was generally somewhat lower than from thevesicle–SK-MH polymer gels (data not shown). The release ofalprenolol from reference gels without SDS, i.e. without catan-
ionic vesicles, was more prolonged when using carbopol 1342 thanwhen using carbopol 940 (p < 0.001). At all concentration levels therelease of alprenolol from the catanionic vesicle containing sam-ples, i.e. samples with SDS, was significantly slower than from gelscontaining only drug substance (p < 0.001).
Fig. 13. Comparison of apparent diffusion coefficients of tetracaine, from catan-ionic tetracaine/SDS (in a 35:65 ratio) vesicle–SK-MH polymer gels with a variety ofpolymer and vesicle concentrations.
ces B: Biointerfaces 70 (2009) 187–197 195
4
4
vfa[fScvc[
csvfoagum[avtp
ortcf1sucfati1tac
pial[httat
cbtcfc
Fig. 14. Schematic illustrations of (a) interactions between the catanionic vesiclesand the SK-MH polymer. The electrostatic and lipophilic interactions between poly-mer grafts and catanionic vesicles give rise to cross-links that result in gel formation.One vesicle has been cross-cut to show lipophilic interactions. The polymer back-bone is light gray and the grafts are dark gray. In the vesicles the light color represents
N. Dew et al. / Colloids and Surfa
. Discussion
.1. Phase behavior and gel formation
In previous studies the formation of catanionic micelles andesicles has been noted for mixtures of oppositely charged sur-actants, where one of the surfactants is a drug substance. Theggregates have also been used for prolonged drug release purposes7–11]. Our investigations of the alprenolol/SDS mixtures show thatormation of vesicles only occur in the samples belonging to theDS rich part of the phase table, see Fig. 2. As a result, the vesi-les formed most likely bear a net negative charge. The increasediscosity of the samples formed at the high concentrations indi-ate elongated micelles, as has been observed in previous work9].
The phenomenon of gel formation due to interactions in systemsontaining catanionic vesicles and polymers has been observed byeveral research groups [15–17,19]. Marques et al. mixed catanionicesicles with oppositely charged JR-400 polymers and found gelormation in polymer-rich mixtures [19]. Faceted vesicles, disks andther bilayer structures were imaged using cryo-TEM. Medronho etl. used an uncharged polymer bearing lipophilic grafts, and foundel formation when mixed with catanionic vesicles [17]. Whensing a hydrophobically modified chitosan, Lee et al. found gel for-ation upon mixing with oppositely charged catanionic vesicles
16]. Asbaugh et al. used hydrophobically modified sodium poly-crylate polymer, and found gel formation when the catanionicesicles had an anionic net charge [15]. They found the gel forma-ion to largely depend on the hydrophobic interactions betweenolymer and vesicles.
The mixing of our drug containing catanionic vesicles andppositely charged polymer, bearing hydrophobic modifications,esulted in gel formation. In accordance with earlier sugges-ions [17], our investigations show that a critical amount of bothatanionic vesicles and oppositely charged polymer are neededor gel formation to occur. When using the uncharged HM(C16-8)-PEG polymer gel formation was not achieved. This was nourprise. Earlier studies have shown that although mixing ofncharged lipophilically modified polymer and catanionic vesi-les may result in samples with increased viscosity, gels are notormed [15,29]. In the vesicle–HM(C16-18)-PEG mixtures the inter-ctions are too weak to result in durable cross-links. Similar tohe observations in this study, Nilsson et al. reported an increasen viscosity upon addition of catanionic vesicles to the HM(C16-8)-PEG polymer [29]. This increase in viscosity was derivedo an organized mixed-micelle structure, consisting of cat- andnionic surfactants and the hydrophobic grafts on the polymerhains.
Antunes et al. showed that the positively charged JR-400olymer interacted strongly with SDS/DDAB vesicles [18]. These
nteractions led to the formation of faceted vesicles, vesicle rupturend even caused formation of disk-like aggregates. The polygonal-ike shape was suggested to be the result of chain crystallization30]. It is plausible that the high charge density and the lack ofydrophobic grafts on the JR-400 polymer cause the vesicles inves-igated in this study to disintegrate. The cross-links needed forhe formation of a gel network are thus not able to form; insteadn associative phase separation occurs due to surfactant adsorp-ion.
When using the SK-MH polymer, which bears both positiveharges and hydrophobic grafts, both electrostatic and hydropho-
ic interactions are made possible. The resulting interactions seemo be durable but not too strong to cause vesicle rupture or asso-iative phase separation. Instead cross-links are formed and gelormation results. A schematic illustration of how catanionic vesi-les may interact with the SK-MH polymer, and how this differs
a positively charged substance and the darker a negatively charged one (in both aand b). (b) Catanionic vesicles entrapped within a carbopol gel network, shown inlight gray. In this case the vesicle is not a founding part of the gel structure, as thepolymer is covalently cross-linked. The illustrations are not to scale.
from how the vesicles are located in the carbopol gels, is shown inFig. 14.
4.2. Cryo-TEM
Cryo-TEM has proven to be a useful tool to characterize struc-tures formed in catanionic systems [9] and also to study viscousgel samples [7,9–11,19]. Earlier studies have shown that polymersmay affect vesicle size and integrity [20,30]. In this study the posi-tively charged lipophilically modified polymer SK-MH was foundto induce a decrease in the alprenolol/SDS vesicle size. It alsodecreased the presence of bi-lamellar vesicles. The elongated struc-tures displayed in Fig. 5 are, to a large extent, probably an effect ofthe blotting of the viscous samples.
Due to the samples’ sensitivity to radiation the tetracaine sys-tems were not as easily characterized as the alprenolol systems. Ourcryo-TEM investigation revealed, however, that the tetracaine/SDSvesicle–SK-MH samples contained lamellar sheets and vesicles with
openings. Vesicles displaying holes in the membrane were observedalso in the absence of polymer, and it is likely that the lamellarsheets form as a consequence of bilayer–polymer interactions. Poly-mer induced disintegration of vesicles has been noted previously[19,30].
1 ces B:
sl
4
wdoh[t
chvl[sphtaEsstted
simaicwIsvhTrbdicuphto
4
ftsabWbt
96 N. Dew et al. / Colloids and Surfa
Cryo-TEM micrographs obtained in this study did not reveal anyigns suggesting that the hydrophobic grafts cause chain crystal-ization and formation of faceted vesicles.
.3. Rheology
All samples that in the visual study were classified as gel-likeere investigated rheologically. The investigated systems typicallyisplayed a maximum in the elasticity, G′, see Fig. 10. Similar rhe-logical profiles have been observed when studying mixtures ofydrophobically modified polyelectrolytes and cationic surfactants31]. Magny et al. observed that a viscosity maximum correspondso the number of hydrophobic grafts in each mixed micelle [31].
The SK-MH–tetracaine mixture had increased elastic and vis-ous modules compared to the pure SK-MH solution. On the otherand, the SK-MH–alprenolol mixture had decreased elasticity andiscosity compared to the pure SK-MH solution. Tetracaine has aower critical micelle concentration (CMC) than alprenolol, 75 mM9] and 100 mM [11], respectively. The concentrations of drug sub-tances used in the present study are in the range of 1–40 mM. It isossible that only tetracaine has formed mixed micelles with theydrophobic modifications on the SK-MH polymer. These interac-ions and the formation of mixed micelle aggregates could possiblyffect the release rate of the drug from the gels, as Paulsson anddsman showed using alprenolol and carbopol 1342 [4,11]. At lowurfactant concentrations there are not enough micelles to form aignificant amount of cross-links or the critical aggregation concen-ration (CAC) has not yet been reached. At very high concentrationshe number of hydrophobic grafts per micelle is not enough tonable crosslink formation. Thereby the amount of cross-links isecreased and the strengthening effect is lost [31,32].
Because of the size of the vesicle aggregates used in this studyeveral hydrophobic grafts can connect to the same vesicle. Thesenteractions could result in a higher elasticity than when mixed
icelles are formed. As Medronho et al. showed in 2006, the stor-ge and loss modulus dimensions depend on vesicle size, and anncreased amount of hydrophobic modifications give rise to moreross-links [17]. Similar results were obtained by Asbaugh et al.hen studying a variety of vesicle and polymer compositions [15].
t was also suggested that the results could be generalized to otherimilarly charged mixtures. In the alprenolol containing gels theesicles with 30% of drug substance seem to give rise to a slightlyigher elasticity than the ones containing 40% of drug substance.his is valid at the lower vesicle concentrations, see Fig. 10. Theeason for this can only be speculated upon, since the mechanismehind the gel formation is not fully understood. In order to gain aeeper understanding of the gel formation process systematic stud-
es concerning the influence of vesicle properties, such as size, netharge and charge density are needed. Further, the amount of free,naggregated, surfactant and drug need to be determined. Someolymer properties are also unclear; e.g. the distance between theydrophobic grafts on the polymer backbone is not known. Inves-igations designed in order to clarify these matters lie, however,utside the scope of this study.
.4. Drug release
The catanionic vesicle–SK-MH polymer gels are most likelyormed by cross-links. These form as the hydrophobic grafts ofhe polymer are included in the vesicle bilayer and as the oppo-ite charges of the polymer and vesicles attract. As drug substance
nd oppositely charged surfactant is released the vesicles are slowlyroken up, and monomers and/or small aggregates are released.hen the number of cross-links decreases, the gel itself is slowly
roken up. This makes these gels a kind of “self-destructing” sys-ems, contrary to the conventional gels that are slowly emptied of
Biointerfaces 70 (2009) 187–197
drug substance but still remain rather intact even after a substan-tial amount of time. A gel that breaks down as the active ingredientleaves it might more easily be eliminated from its site of actionthan a conventional gel. This could potentially be an advantage ina clinical situation.
Interactions between tetracaine and alprenolol and the SK-MHpolymer discussed in the rheology section could have effects onthe release rate of drug substance from the gels. It has been shownthat alprenolol interacts with the lipophilic modifications of car-bopol 1342 [4,11]. This is true also for tetracaine, and the reportedCAC of tetracaine is lower than that of alprenolol. As gels are onlyformed when mixing catanionic aggregates with SK-MH polymer,drug release from polymer mixtures without SDS could not bestudied. The effect on drug release caused by drug–polymer inter-actions is thus not definable. The apparent diffusion coefficientsof alprenolol and tetracaine from the reference carbopol gels, notcontaining catanionic aggregates, are in the same range. However,the drug release from carbopol 1342 is more prolonged than fromcarbopol 940, due to the interactions with the lipophilic modifica-tions.
When comparing the effect of alterations of vesicle or polymerconcentrations on the release rate of tetracaine from the vesicle–SK-MH polymer gel systems, the vesicle concentration was found tohave a larger influence, at investigated intervals. A higher vesicleconcentration resulted in a slower release rate of drug substancebut the effect of an increased polymer concentration was not asdistinct, see Fig. 12. There are limitations to how much the vesi-cle concentrations can be raised, as lamellar and phase separatedregions exist at higher concentrations. Raised polymer concentra-tions would also cause difficulties in the gel preparation process.Altering the amount or type of hydrophobic graft on the polymercould also affect the CACs and other system properties.
There is reason to believe that the release mechanism of drugsubstances from conventional carbopol gels differ from that of thevesicle–polymer gels. The carbopol gel networks are intact through-out the release study and most likely only monomeric or smallaggregates can diffuse through the gel network and be released,as indicated by Brohede et al. [33]. However, the movements ofmonomeric surfactant molecules within the vesicle–polymer gelsremain to be studied. The release rate of drug substance from thecatanionic vesicle–SK-MH polymer gels was in this study comparedto drug release from conventional carbopol gels. Even though thevesicle–SK-MH gels are broken down as drug substance is released,the apparent diffusion coefficients of both alprenolol and tetracainefrom these gels are generally not significantly different from theones of the carbopol gels. There is no doubt that the catanionicvesicle–polymer systems are an interesting future research sub-ject. Further studies are needed to elude how the release of drugsubstance can be even further sustained, and to understand themechanisms behind the drug release from these systems.
5. Conclusions
Mixture of drug containing net negatively charged vesicles andoppositely charged polymers bearing lipophilic grafts can result ingel formation. Data of the present study show that there needsto be a well balanced amount of polymer and vesicles to achievestrong enough interactions to form a gel structure. Gel structureswere confirmed with rheological measurements and the presenceof vesicles in the gels was visualized using cryo-TEM. Variationsin vesicle concentration affected the release rate of the drug sub-
stance from the gels to a greater extent than did variations inpolymer concentration, at investigated intervals. Even though thevesicles are a founding part of the gel structure in these systems,the apparent diffusion coefficients of the drug substances, whenusing vesicle–polymer gels, were not significantly different from
ces B:
tpt
A
f
R
[
[[
[[
[
[[
[
[
[
[
[[[
[[[[
[[
N. Dew et al. / Colloids and Surfa
hose obtained using covalently cross-linked, preformed, gels. Theromising results obtained in this study merit further investiga-ions and development of these systems for drug delivery purposes.
cknowledgement
Financial support from the Swedish Research Council is grate-ully acknowledged.
eferences
[1] K. Edsman, H. Hagerstrom, Journal of Pharmacy and Pharmacology 57 (2005)3.
[2] J. Carlfors, K. Edsman, R. Petersson, K. Jornving, European Journal of Pharma-ceutical Sciences 6 (1998) 113.
[3] S.D. Desai, J. Blanchard, Drug Delivery 7 (2000) 201.[4] M. Paulsson, K. Edsman, Journal of Colloid and Interface Science 248 (2002) 194.[5] A. Paavola, I. Kilpelainen, J. Yliruusi, P. Rosenberg, International Journal of Phar-
maceutics 199 (2000) 85.[6] H. Sjoberg, S. Persson, N. Caram-Lelham, Journal of Controlled Release 59 (1999)
391.[7] T. Bramer, N. Dew, K. Edsman, Journal of Pharmaceutical Sciences 95 (2006)
769.[8] T. Bramer, G. Karlsson, K. Edwards, K. Edsman, Journal of drug delivery and
science technology 17 (2007) 285.[9] T. Bramer, M. Paulsson, K. Edwards, K. Edsman, Pharmaceutical Research 20
(2003) 1661.10] N. Dew, T. Bramer, K. Edsman, Journal of Colloid and Interface Science 323 (2008)
386.11] M. Paulsson, K. Edsman, Pharmaceutical Research 18 (2001) 1586.12] T. Bramer, N. Dew, K. Edsman, Journal of Pharmacy and Pharmacology 59 (2007)
1319.
[[
[
Biointerfaces 70 (2009) 187–197 197
13] C. Tondre, C. Caillet, Advances in Colloid and Interface Science 93 (2001) 115.14] M. Kasbauer, D.D. Lasic, M. Winterhalter, Chemistry and Physics of Lipids 86
(1997) 153.15] H.S. Ashbaugh, K. Boon, R.K. Prud’homme, Colloid and Polymer Science 280
(2002) 783.16] J.H. Lee, J.P. Gustin, T.H. Chen, G.F. Payne, S.R. Raghavan, Langmuir 21 (2005) 26.17] B. Medronho, F.E. Antunes, B. Lindman, M.G. Miguel, Journal of Dispersion Sci-
ence and Technology 27 (2006) 83.18] F.E. Antunes, E.F. Marques, R. Gomes, K. Thuresson, B. Lindman, M.G. Miguel,
Langmuir 20 (2004) 4647.19] E.F. Marques, O. Regev, A. Khan, M.D. Miguel, B. Lindman, Macromolecules 32
(1999) 6626.20] J.H. Lee, V. Agarwal, A. Bose, G.F. Payne, S.R. Raghavan, Physical Review Letters
96 (2006).21] H. Ringsdorf, E. Sackmann, J. Simon, F.M. Winnik, Biochimica Et Biophysica Acta
2008.25] K. Edwards, M. Silvander, G. Karlsson, Langmuir 11 (1995) 2429.26] W.I. Higuchi, Journal of Pharmaceutical Sciences 51 (1962) 802.27] P.L. Ritger, N.A. Peppas, Journal of Controlled Release 5 (1987) 23.28] P. Nilsson, Interaction between Crosslinked Polyelectrolyte Gels and Oppositely
Charged Surfactants, Uppsala University, Medicinska vetenskapsområdet, Fac-ulty of Pharmacy, Department of Pharmacy, Uppsala universitet, Uppsala, 2007,p. 68.
29] S. Nilsson, M. Goldraich, B. Lindman, Y. Talmon, Langmuir 16 (2000) 6825.30] F.E. Antunes, R.O. Brito, E.F. Marques, B. Lindman, M. Miguel, Journal of Physical
Chemistry B 111 (2007) 116.31] B. Magny, I. Iliopoulos, R. Zana, R. Audebert, Langmuir 10 (1994) 3180.32] K. Holmberg, B. Jönsson, B. Kronberg, B. Lindman, Surfactants and Polymers in
Aqueous Solution, 2nd ed., John Wiley & Sons Ltd., Chichester, 2003.33] U. Brohede, T. Bramer, K. Edsman, M. Stromme, Journal of Physical Chemistry B
