European Journal of Pharmaceutical Sciences 20 (2003) 429–438

Interactions of ibuprofen with cationic polysaccharidesin aqueous dispersions and hydrogels

Rheological and diffusional implications

Rosalıa Rodrıguez, Carmen Alvarez-Lorenzo∗, Angel Concheiro

Departamento de Farmacia y Tecnolog´ıa Farmacéutica, Facultad de Farmacia, Universidad de Santiago de Compostela,Santiago de Compostela 15782, Spain

Received 8 July 2003; received in revised form 11 September 2003; accepted 11 September 2003

Abstract

Non-steroidal antiinflammatory drugs, such as ibuprofen, are amphiphilic substances capable of self-association in aqueous solutionsand able to be sorbed onto polymers through hydrophobic and electrostatic bonds. The aim of this work was to analyze the associationprocesses of sodium ibuprofen with cationic celluloses (Celquat® H-100 (PQ-4) and SC-230M (PQ-10)) and cationic guar gums (Ecopol®

261-S and 14-S) and their repercussions on the properties of the aqueous dispersions and cross-linked hydrogels. The interaction processwas studied in aqueous dispersions through transmittance, surface tension, fluorescence, conductivity, viscosity and oscillatory rheometrymeasurements. Belowcmc, the drug molecules weakly interact with the polymers through hydrophobic and ionic interactions. Around thecmc (4%), a notable decrease in the viscosity, and storage and loss moduli of the dispersions (even precipitation in PQ-10 systems) wasobserved. An additional increase in drug concentration induced the dispersions to recover their initial properties. Since ibuprofen/polymercationic groups ratio were in all cases above 1, these observations indicate that drug self-association induces the polymer to coil aroundthe micelles and, as the number of micelles increases (more drug concentration) the polymer chains interact with more of them, un-coiling again to some extent. Polymer (1%) dispersions containing 6% ibuprofen showed drug diffusion coefficients much lower than inwater. When a surfactant, sodium dodecylsulfate, was added to these systems the diffusion coefficients decreased even more, suggestingthe formation of new associative structures. Chemically cross-linked hydrogels made of these cationic polysaccharides absorb consid-erable amounts of ibuprofen (up to 15 g/g) and showed a pH-dependent release process. At acidic pH, drug–polymer affinity is main-tained, preventing drug release. In contrast, at pH 8 the interactions are broken and the release process is sustained for more than 4 h.In summary, ibuprofen interactions with cationic polysaccharides strongly determine the performance of their aqueous dispersions andhydrogels.© 2003 Elsevier B.V. All rights reserved.

Keywords:Amphiphilic drugs; Cationic cellulose; Cationic guar gum; Sodium dodecyl sulfate; Polymer–surfactant interactions

1. Introduction

Polymer association with complementary additives thatcan rapidly strengthen or induce connections between thepolymeric chains has been shown as a useful way to obtainconsiderable increases in the viscosity of the dispersions—avoiding the difficult handling of high concentrations orhigh molecular weight polymers—or even the creation ofresponsive systems whose properties are modulated bychanges of the affinity between the polymer and the additives

∗ Corresponding author. Fax:+34-981-547148.E-mail address:ffrusdog@usc.es (C. Alvarez-Lorenzo).

(Pouliquen et al., 2003). Most semi-synthetic polysaccha-rides have a particularly adequate structure to interact withamphiphilic molecules, which is interesting for this type ofapplications. Cellulose and guar gum derivatives present aglucosidic backbone that may establish hydrophobic inter-actions, while the presence of hydrophilic or charged groupsin their substituents provides the polymer with hydrogenbonding capacity or high affinity for oppositely chargedmolecules. In several papers, the use of surfactants to in-duce changes in the conformation of cationic polysaccha-rides and to promote the formation of aggregates has beenproposed as a way to obtain homogeneous aqueous disper-sions and to modulate their rheological behavior (Goddard

0928-0987/$ – see front matter © 2003 Elsevier B.V. All rights reserved.doi:10.1016/j.ejps.2003.09.004

430 R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438

and Hannan, 1977; Hoffman et al., 1996; Chronakis andAlexandridis, 2001). Additionally, the high bioadhesive ca-pacity (Gruber and Kreeger, 1996) and low toxicity (Annon,1988) of hydrophilic cationic polysaccharides make themparticularly useful for preparing cosmetic preparations anddrug delivery systems able to combine long residence timesat the application site with adequate mechanical properties.

As classical surfactants, many kinds of drugs—i.e., tri-cyclic antidepressants,�-blockers, phenotiazine tranquiliz-ers, antihistamines, non-steroidal antiinflammatory drugs(NSAIDs) and local anesthetics—are surface-active sub-stances capable of self-association in aqueous solutions(Florence and Attwood, 1998; Taboada et al., 2001) and ableto be sorbed onto polymers through hydrophobic and elec-trostatic bonds (Hugerth, 2001; Mura et al., 2003; Yomotaand Okada, 2003). The repercussions of these phenomenaon the properties of polymer gels depend on the tendencyof the drug to aggregate and on the strength of its interac-tions with the polymer (Alvarez-Lorenzo and Concheiro,2003). In Carbopol® 1342 gels, alprenolol forms, at lowdrug concentrations, micelle-like aggregates with the poly-mer lipophilic residues, increasing the elastic and viscousmoduli of the gels. However, as the drug concentration israised, the consistency decreases, the gel collapses and,when alprenolol amino groups neutralize the carboxylicacid groups of the acrylic polymer, precipitation occurs(Paulsson and Edsman, 2002). The changes observed inalprenolol diffusion rate are a consequence of both theinteractions with the polymer and the changes induced inthe viscosity of the systems. Interaction of phenothiazines(chlorpromazine, trifluoromazine, promazine, and promet-hazine) with hyaluronate causes the gels made of this an-ionic polymer to shrink; the minimum drug concentrationrequired for that being proportional to the critical micellarconcentration,cmc (Yomota and Okada, 2003). The po-tential as periodontal drug delivery systems of non-ioniccellulose ether–surface-active anesthetic drug gels was high-lighted byScherlund et al. (2000). Lidocaine and prilocainedo not interact with ethylhydroxyethyl cellulose (EHEC) orhydrophobically modified EHEC (HM-EHEC) but stronglyaffect their interactions with sodium dodecylsulfate andmyristoylcholine bromide. In consequence, the drug loadedsystems notably differed from polymer–surfactant disper-sions in the storage and loss moduli.

Non-steroidal antiinflammatory drugs (NSAIDs), such assodium ibuprofen, adsorb onto the non-ionic cellulose ethersnon-cooperatively up to thecmc and cooperatively abovethis concentration (Ridell et al., 1999). Nearcmc, ibupro-fen causes phase separation of EHEC dispersions, becauseof a strong hydrophobic interaction that shrinks the polymerchains but, abovecmc, micelles of ibuprofen solubilize thehydrophobic parts of the polymer. The presence of apolar re-gions and ammonium groups in the cationic polysaccharidescould enhance the intensity of interactions with the aromaticring and the hydrophilic carboxylic group of NSAIDs, mak-ing it possible to combine both hydrophobic association and

Coulombic forces. These interactions may induce changes inthe system which can be foreseen as useful to achieve an effi-cient control of the release process from aqueous dispersions(Paulsson and Edsman, 2001; Jimenez-Kairuz et al., 2002)or chemically cross-linked hydrogels (González-Rodrıguezet al., 2002; Rodrıguez et al., 2003a).

The aim of this work was to analyze the association pro-cesses of sodium ibuprofen with cationic celluloses andcationic guar gums and their repercussions on the propertiesof the aqueous dispersions and cross-linked hydrogels. Twovarieties of cationic hydroxyethyl celluloses and two vari-eties of cationic guar gums, differing in molecular weightand content and distribution of the cationic groups, wereused to evaluate the influence of polymer molecular struc-ture. Additionally, the effect of the concomitant presenceof an anionic surfactant, sodium dodecylsulfate, which maycompete with the drug for the binding to the polymer, wasconsidered. Since NSAIDs are suitable candidates for incor-poration into polymer dispersions (physical gels) for topicalapplication or into hydrogels that release them at the smallintestine, minimizing the adverse gastric drug effects afteroral administration (Dominkus et al., 1996), a preliminaryevaluation of the potential interest of the dispersions andhydrogels as drug delivery systems was also carried out.

2. Materials and methods

2.1. Materials

Polyquaternium-4 (PQ-4) (Celquat® H-100, batchFGS 1014) and Polyquaternium-10 (PQ-10) (Celquat®

SC-230 M, batch GFS 1139) were provided by Na-tional Starch and Chemical Ltd., UK. Ecopol® 261-S(E-261, batch L10159215) and Ecopol® 14-S (E-14, batchL10159214) were from Economy Polymers & Chemi-cals, USA. Sodium ibuprofen, sodium dodecylsulfate andethyleneglycol diglycidylether (EGDE, 50 vol.% in water)were from Sigma–Aldrich Chemical Co., USA. Purifiedwater by reverse osmosis (MilliQ®, Millipore Spain) withresistivity >1.82 M� cm was used. All other reactives wereof analytical grade.

2.2. Methods

2.2.1. Preparation of the dispersionsAqueous dispersions containing 1.0% (w/w) cationic cel-

lulose and a wide range of ibuprofen concentrations (0–8%,w/w) were prepared by dispersing the required amounts ofeach component in 100 ml of water under stirring. The sys-tems were left to stand for 24 h before characterization. Allstudies were carried out at 25◦C.

2.2.2. CloudinessThe cloudiness of dispersions was determined by mea-

suring transmittance at 800 nm (Shimadzu UV-240, Japan)against a blank of cationic polysaccharide dispersion.

R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438 431

2.2.3. pHThe measurements were made with a pH-meter Crison,

model GLP22 (Spain), equipped with a sensor (Ag/AgCl)for viscous media no. 52-21.

2.2.4. ConductivityConductance measurements were made in a rhadiometer

conductivity meter model CDM2e (Denmark) equipped witha Crison platinum sensor (Spain).

2.2.5. Surface tensionThe measurements were made using the platinum ring

method with a Lauda tensiometer TD1 (Germany) applyingthe necessary density corrections.

2.2.6. Steady-state fluorescence measurementsPyrene (Py) emission spectra (λ = 350–450 nm) were

recorded in a Perkin-Elmer LS50B fluorescence spectropho-tometer (UK), with the excitation wavelength set to 310 nmand slits set to 5 and 5, for excitation and emission, respec-tively. The samples were prepared by the addition of polymer(0.5% cationic celluloses or 0.1% cationic guar gums) andibuprofen (0.5–7%) to a pyrene aqueous solution (10−6 M)and stored at 25◦C for 24 h. The ratio of the intensity atthe first (I1 at 373 nm) and the third (I3 at 392 nm) vibronicpeaks was used as an index of the local hydrophobicity ofthe polymer–surfactant aggregates (Rodrıguez et al., 2003b).

2.2.7. ViscosityDeterminations of apparent and specific viscosity were

carried out in Cannon-Fenske capillary viscometers (Afora,Spain).

2.2.8. Viscoelastic behaviorThe rheological behavior of the polymer dispersions with

and without ibuprofen was evaluated, in triplicate, in a Rheo-lyst AR-1000N rheometer (TA Instruments, UK) equippedwith an AR2500 data analyzer, fitted with a Peltier temper-ature control, and a 6 cm cone-plate measuring geometry,covered with a solvent trap. Oscillatory shear responses (G′or storage modulus, andG′′ or loss modulus) were deter-mined at 0.1 Pa over the frequency range of 0.01–50 rad s−1.The test conditions were inside the linearity range of theviscoelastic properties.

2.2.9. Ibuprofen diffusion assaysDrug release profiles from 1% polymer dispersions con-

taining 6% ibuprofen (and in some cases SDS at differentconcentrations) were obtained in triplicate in Franz-Chiendiffusion cells (Vidra-Foc, Spain), thermostated at 37◦C,and fitted with cellulose acetate filters (0.45�m pore size,Teknokroma, Spain) after previous storage at the same tem-perature for at least 1 h. The area available for diffusion was0.785 cm2. A sample of 2.5 ml was placed in the donor com-partment; while 6.0 ml of iso-osmotic NaCl aqueous solu-tion filled the recipient compartment and was stirred with a

magnetic rod. Samples (0.10 ml) were taken from the recip-ient compartment at intervals over a 10-h period, for deter-mination of ibuprofen on the basis of absorption at 273 nm(Shimadzu UV-240, Kyoto, Japan); in each case, recipientmedium volume was immediately made up with iso-osmoticsolution. Diffusion coefficients were estimated by non-linearregression applying theHiguchi (1962)equation

Q

A= 2C0

(Dt

π

)1/2

(1)

whereQ/A is the amount of ibuprofen released per unit areaat the timet, andC0 is the initial concentration of ibuprofenin the dispersion.

2.2.10. Hydrogel synthesisHydrogels were synthesized as described previously

(Rodrıguez et al., 2003a). Briefly, EGDE (2 ml) was addedto a 3% cationic cellulose dispersion in 0.10 M NaOHmedium (10 ml) or to a 2% cationic guar gum solution in0.05 M KOH medium (10 ml). After stirring for 5 min, themixture was transferred to a test tube of 10.5 mm i.d., andhermetically closed. The tubes were kept at 60◦C for 6 h,in the case of guar gums, and 24 h, in the case of cationiccelluloses. After cooling, the hydrogels were carefully re-moved and immersed in ultrapure water for 24 h to swell.The hydrogels were then transferred to HCl 0.1N solutionfor 12 h to neutralize the basic medium, and finally kept inultrapure water for 1 week, changing the medium every 12 hto allow a complete wash out of non-reacted substances.

2.2.11. Equilibrium swelling of hydrogelsAfter being immersed in ibuprofen solutions (as explained

below), equilibrium diametersd of pieces of cylindrical gelswere measured using a digital micrometer. The degree ofswelling was expressed as

Swelling ratio= V

V0=

(d

d0

)3

(2)

where d0 was the diameter of gel discs upon synthesis(10.5 mm).

The water content of the hydrogel was estimated by thedifference between the weight of fully swollen hydrogelsamples (W), after careful wiping of their surfaces with asoft tissue, and the weight of the samples after being dried(W0) for 1 week at 40◦C

Q = (W − W0)

W0(3)

2.2.12. Ibuprofen sodium loading and release fromcross-linked hydrogels

Pieces of the cylindrical gels (2–3 mm thickness) wereplaced in 10 ml aqueous solutions of 1–8% drug for 3 daysat room temperature. The amount loaded was determinedas the difference between the initial amount of drug andthe amount remaining in the outer solution after the 3 days,

432 R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438

which was measured spectrophotometrically (λ = 273 nm;Shimadzu UV-240, Japan). Then, each loaded hydrogel wassuperficially rinsed with water and directly immersed in50 ml water of pH 3 (HCl aq.), pH 6 (not adjusted), or pH 8(phosphate buffer), and kept at 37◦C. The release was alsoevaluated when the hydrogels where immersed first in pH3 medium (for 2 h) then, transfer to pH 6 medium (for 2 h),and subsequently to pH 8 solution. The experiments werecarried out in triplicate. The amount of ibuprofen releasedwas measured spectrophotometrically (λ = 273 nm) in pe-riodically taken samples and again placed in the same ves-sel so that the liquid volume was kept constant. The releaseprofiles between 10 and 80% were characterized by fittingthe following equation by non-linear regression (Korsmeyerand Peppas, 1981)

Mt

M∞= Ktn (4)

whereMt /M∞ is the fraction of drug released at timet.To detect the presence of ibuprofen remaining in the hy-

drogels after the release tests, IR spectroscopy was used.Samples of hydrogels used for the loading/release were driedin an oven at 40◦C until constant weight, and powderedin a mortar. IR spectra were recorded over the range of400–4000 cm−1, in a Bruker IFS 66 V FT-IR spectrometer(Germany), using the potassium bromide pellet technique.The characteristic peaks of ibuprofen appear at 2956 cm−1,benzene ring, and 1584 and 1409 cm−1, ionized carboxylicgroups (Higgins et al., 2001).

3. Results and discussion

The structure of the polymers is shown inFig. 1. PQ-10and PQ-4 are linear cationic hydroxyethyl celluloses inwhich the ammonium groups are bonded to the hydroxy-ethyl substituent (PQ-10) or directly grafted to the cellulosebackbone while the hydroxyethyl substituent remains un-altered (PQ-4). On the other hand, cationic guar gums areformed by mannose units linearly bonded that alternativelyhave a lateral ramification of galactopiranose. Some po-sition 6 hydroxyl groups of galactopiranose are boundedto a hydroxypropyltrimethylammonium group. In additionto these general differences, the polymers also differ inmolecular weight and amount of cationic groups as reportedpreviously (Table 1) (Rodrıguez et al., 2001).

3.1. Aqueous dispersions

3.1.1. Study of the interactions in aqueous dispersionsAqueous dispersions of cationic celluloses and cationic

guar gums prepared with several ibuprofen concentrationswere, in general, homogeneous. Transmittance was in somecases even greater than that of the polymer alone dispersions.Precipitation was only observed for the PQ-10 dispersionscontaining 4–5% ibuprofen (Fig. 2). The pH of the systems

O

H

HO

H

HO

H

H

OHHO

OH

O

H

H

HO

H

H

OHHO

O (CH2CH2O)n CH2CHOHCH2N+(CH3)3

Cl-

O

H

HOHO

H

H

OHHO

O

O

H

H

HO

H

H

OHHO

O (CH2CH2O)nH

Cl-

(CH2CH2O)nH

N+

N+

Cl-

O

H

H

HO

OH

H

HH

O

OH

O

H

H

HO

OH

H

HHO

O

O

OH

H

H

HO

H

OHHH

O R

z

3

PQ-10

PQ-4

E-14 and E-261

O

ONa

Ibuprofen

R = H or CH2CHOHCH2N+(CH3)3Cl-

Fig. 1. Molecular structures of the cationic celluloses (polyquaternium-4and polyquaternium–10), the cationic guar gums (Ecopol® 261-S and14-S differ in the content in the cationic groups), and ibuprofen.

increased linearly from around 7 (for PQ-10 and PQ-4) or6 (E-14 and E-261) to 8.5 when ibuprofen concentrationchanged from 0 to 8%; the ibuprofen–cationic polysaccha-ride systems showing the same pH values as the ibuprofensodium solutions. The conductivity of the systems also in-creased from 650–800 to 10,000–12,000�S/cm with drugconcentration, not showing any inflection points.

The surface activity of ibuprofen was clearly seen whenthe evolution of pyrene fluorescence and surface tension

Table 1Molecular weight (MW), intrinsic viscosity in water at 25◦C ([η]), andcharacteristics of substitution of the cationic celluloses and cationic guargums (Rodrıguez et al., 2001)

Polymer MWa [η] (dl/g) MSHECb %N Ammonium

group/sugarrepeating unit

PQ-4 1400000 14.59 2.71 1.27 1:5.3PQ-10 1700000 70.09 1.28 1.95 1:2.9E-14 500000 46.78 – 1.16 1:6.5E-261 200000 45.85 – 0.91 1:8.5

a Data provided by the supplier.b MSHEC = average number of hydroxyethyl substituents bonded to

each glucopiranose unit.

R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438 433

0 1 2 3 4 5 6 7 8

Tra

nsm

itta

nce

(%)

60

90

120

150

180

210

0 1 2 3 4 5 6 7 8

Surf

ace

tens

ion

(mN

/m)

30

40

50

60

70

Ibuprofen (%)0 1 2 3 4 5 6 7 8

Vis

cosi

ty (

mP

a·s)

0

1000

2000

3000

4000

5000

6000

Tra

nsm

itta

nce

(%)

20

40

60

80

100

120

Surf

ace

tens

ion

(mN

/m)

30

40

50

60

70

Vis

cosi

ty (

mP

a·s)

0

200

400

600

800

1000

E-14 E-261

P Q-4P Q-10

Ibuprofen only

P

P

Fig. 2. Transmittance, viscosity and surface tension of aqueous dispersions of cationic polysaccharides (1%) containing different proportions ofibuprofen(P: precipitation of PQ-10 dispersions).

measurements as a function of drug concentration wererecorded. With both techniques, thecmc obtained foribuprofen in water, 4%, was similar to the value found byRidell et al. (1999). The minimum observed can be at-tributed to some surface-active impurities in the drug. Themicropolarity of the medium was measured by the changesin the I1/I3 ratio in the emission spectra of pyrene. TheI1/I3 ratio is higher in a polar medium (1.87 in water) thanin an apolar medium (0.66 in cyclohexane) (Anthony et al.,1998). Without ibuprofen, cationic polysaccharides showeda I1/I3 ratio around 1.2. In the presence of 5% ibuprofen,this value decreased to 0.9, in PQ-4 dispersions, and to0.70–0.75, in the other cationic polysaccharide systems.The greaterI1/I3 ratio of PQ-4/ibuprofen dispersions is ex-plained by the more hydrophilic character of this polymer,which have more hydroxyethyl groups. The surface tensionof the ibuprofen–cationic polysaccharide dispersions de-creased, from 50–60 mN/m, in the presence of the drug toreach a minimum, 30 mN/m, around 4% ibuprofen (Fig. 2).No significant differences were observed in the surface ten-sion patterns with respect to that obtained in the absenceof polymer, which indicates that none of these cationicpolysaccharides significantly interferes in the micellizationprocess of the drug and that the main interactions may hap-pen for drug concentrations around or abovecmc. Theseobservations were confirmed by the changes that occurredin the transmittance and rheological properties of the dis-persions. The viscosity of the systems containing ibuprofenat a concentration aroundcmcwas considerably lower than

that of the polymer dispersions in the absence of the drug(Fig. 2). The effect was less pronounced for PQ-4 systemsthat showed, as the E-14 dispersions, a significant increasein viscosity for lower drug concentrations. In general, thevalues ofG′ andG′′ followed a tendency similar to that ofviscosity and continuously decreased until the drug con-centration reached thecmc (4%). An additional increasein drug concentration induced the dispersions to recovertheir initial properties, showingG′ and G′′ values similarto or even greater than those of polymer alone dispersions(Figs. 3 and 4).

These observations may be explained taking into accountthe electrostatic and hydrophobic interactions that may beestablished between the drug and the polymers. Even forthe lowest drug concentration studied (0.5%), the molar ra-tio of carboxylic drug groups/ ammonium polymer groupswas greater than 1; this means that in all systems thereare enough ibuprofen molecules to ionically neutralize allcationic groups of the polymer. Therefore, although be-low cmc the interaction seems to be scarcely significant,some drug molecules may act as bridges between differentpolymer chains, favoring the formation of a three dimen-sional network of slightly greater consistency. Consideringthe viscosity and transmittance data, the critical aggrega-tion concentration,cac, may be established to be around2% ibuprofen. When thecmcis reached, the drug moleculesself-associate causing the polymer chains to coil to wrapthe micelles. This folding favors intramacromolecular bondsand, in consequence, produces a decrease in the consistency

434 R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438

Angular frequency (rad/s)

0.01 0.1 1 10 100

G´

(Pa)

10-5

10-4

10-3

10-2

10-1

100

101

G´´

(P

a)

10-5

10-4

10-3

10-2

10-1

100

101

0 %1 % 2 %3 %4 %5 %6 %7 %8 %

Fig. 3. Effect of ibuprofen concentration on the viscoelastic behavior ofPQ-10 dispersions.

Ibuprofen (%)0 1 2 3 4 5 6 7 8

G´

and

G´´

(P

a)

0.1

1

10

100

G´

and

G´´

(P

a)

0.001

0.01

0.1

1 P Q - 4

P Q - 1 0

E-14

E-261

Fig. 4. Storage (G′, solid symbols) and loss (G′′, open symbols) moduli ofcationic polysaccharide dispersions (1%) containing different proportionsof ibuprofen (angular frequency 0.9 rad/s).

Fig. 5. Schematic drawing of the effect of ibuprofen on the conformationof cationic polysaccharides, as the drug concentration increases. In allsystems studied, the molar ratio between anionic groups of ibuprofen andcationic groups of the polymer was above 1. When thecmcof ibuprofenis reached, the polymer chains coil around the micelles owing to ionic andhydrophobic interactions. This causes a strong decrease in the consistencyof the system. Increasing the number of micelles, the polymer chainsuncoil and the system recovers its initial properties.

of the systems and, in the case of PQ-10, even precipitation(Fig. 5). A similar effect was described byHugerth (2001)indispersions of dextran sulfate or carrageenan with amitripty-line. Abovecmc, the presence of more micelles makes theaggregates and the hydrophobic regions of the polymersmore soluble. The mixed micelles become the connectingpoints for transient networks (Hoffman et al., 1996) and,therefore, the viscosity and transparency of the dispersionsrise again.

Compared to common anionic surfactants, such as SDS,in which thecacwith cationic polysaccharides is one to twoorders of magnitude lower than thecmc (Anthony et al.,1998; Rodrıguez et al., 2003b), ibuprofen sorption is lessfavored and begins nearer thecmc. The binding of ibupro-fen molecules among themselves is preferred to the bind-ing to the polymer. Similar results were reported byRidellet al. (1999)when the association process of this drug withtwo non-ionic cellulose ethers was analyzed. The differentaffinity and effects that SDS and ibuprofen induce in theproperties of the cationic polysaccharides may be related toimportant structural differences between both amphiphilicmolecules. SDS has a greater surface activity and, therefore,a higher tendency to hydrophobically sorb onto the poly-mer is expected. Additionally, its sorption is facilitated by

R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438 435

the relatively flexible apolar chain, compared to the morerigid benzene ring of ibuprofen (Bustamante et al., 2000).Finally, the sulfate group has a more acidic character proneto interact ionically. Therefore, in the case of SDS, sorp-tion begins at concentrations much lower thancmc form-ing hemimicellar structures with the polymer. In contrast,ibuprofen first forms micelles into which or around whichthe polymer chains distribute owing to electrostatic and hy-drophobic interactions (Fig. 5). Formation of aggregates be-tween cationic polysaccharides and ibuprofen micelles ismore favored as the polymer has more hydrophobic residuesand more cationic groups; i.e., cationic cellulose PQ-10 andcationic guar gum E-14. This explains that the least relevantchanges in viscosity were observed for PQ-4 dispersions,which is a polymer with the greatest content in hydrophilichydroxyethyl groups. The fact that the hydroxyethyl and theammonium groups are separated or bound together justifiesthe different behavior shown by the systems based on thetwo cationic celluloses when they interact with ibuprofen orSDS. The presence of the free hydroxyethyl groups in PQ-4provides enough hydrophilicity to avoid the precipitationof the aggregates, even when PQ-4 charges were neutral-ized and the coiled of the polymer occurred. In PQ-10, theammonium groups are directly bonded to the hydroxyethylsubstituents, which may be included in the aggregates dur-ing association with the amphiphilic molecules. This leavesthe more hydrophobic unsubstituted cellulose chain exposedto the water interface, promoting precipitation (Rodrıguezet al., 2003a).

3.1.2. Ibuprofen diffusion studiesTo carry out the diffusion studies, the dispersions con-

taining 6% ibuprofen were selected since they combined atherapeutically convenient drug dose (Higgins et al., 2001)with adequate rheological properties. Suitable proportionsof SDS were added to some dispersions in order to lead tothe maximum interaction with the cationic polysaccharides(i.e., those that caused the greatest increases in viscosityin the polymer alone dispersions) (Rodrıguez et al., 2003a)with the aim of establishing the possibilities that the addi-tion of this surfactant offers to modulate the properties ofthe systems. All formulations showed sustained release formore than 10 h and the Higuchi equation fitted the diffusionprofiles well (r2 > 0.98); the diffusion coefficients beingparticularly low in cationic guar gums dispersions owingto their thickening capability (Table 2). It is interesting tonote that, in the presence of ibuprofen, SDS did not causea significant increase in the viscosity of the polymer disper-sions, but a decrease. Despite this effect, the diffusion coef-ficients of ibuprofen were significantly lower, especially incationic cellulose systems, in the presence of SDS. There-fore, ibuprofen may remove SDS from the binding sites ofthe polymer leading to the formation of new ionic interac-tions with the polymer, hydrophobic interactions with SDS,or mixed micelles of ibuprofen/SDS/polymer, making thediffusion through the dispersion more difficult (Paulsson and

Table 2Apparent viscosities and ibuprofen diffusion coefficients obtained forcationic polysaccharide (1%)/SDS dispersions

Polymer SDS (%) Viscosity (mPa s) D (cm2/min)

PQ-4 0 41.01 (0.26) 34.1× 10−4 (4.5 × 10−4)0.25 36.81 (0.21) 24.9× 10−4 (3.5 × 10−4)0.30 44.01 (0.66) 19.5× 10−4 (1.2 × 10−4)

PQ-10 0 316.2 (0.91) 11.9× 10−4 (2.9 × 10−4)0.05 156.3 (0.42) 4.8× 10−4 (0.6 × 10−4)0.10 147.7 (1.89) 6.3× 10−4 (0.5 × 10−4)

E-14 0 589.4 (26.2) 5.9× 10−4 (0.5 × 10−4)0.01 407.6 (2.47) 5.2× 10−4 (0.3 × 10−4)0.05 285.9 (3.69) 6.0× 10−4 (0.8 × 10−4)

E-261 0 2129.9 (3.3) 5.3× 10−4 (0.7 × 10−4)0.01 1029.3 (5.1) 5.2× 10−4 (0.4 × 10−4)0.05 955.3 (11.5) 6.5× 10−4 (0.1 × 10−4)

Means of three replicates (S.D.). Ibuprofen diffusion coefficient in waterwas 180× 10−4 (2.2 × 10−4) cm2/min.

Edsman, 2001). In the case of cationic guar gums, the pro-portions of SDS that can be added without causing phaseseparation are relatively small, and, in consequence, the ef-fect on ibuprofen diffusion is scarce.

3.2. Chemically cross-linked hydrogels

3.2.1. Drug loadingFor some applications, it may be interesting to keep

the viscosity and integrity of the polymer-based systemconstant, using the amphiphilic drug–polymer interactionsto regulate the release process. In that case, chemicallycross-linked hydrogels obtained by reacting the cationicpolysaccharides with ethyleneglycol diglycidylether maybe useful (Rodrıguez et al., 2003b). These hydrogels aretranslucent, with a homogeneous surface and high consis-tency and elasticity. The volume of the hydrogels whenfully swollen in water was several times their volume asfreshly prepared: 7.4 for PQ-4, 9.9 for PQ-10, 3.9 for E-261,and 3.4 for E-14. These values corresponded to a hydra-tion level of 70, 250, 114, and 66 g water/g dry hydrogel,respectively. This high water affinity makes the hydrogelspotentially useful as main components of biocompatiblesystems with a high permeability to water and small solutes(Capitan et al., 2000).

Discs of hydrogels immersed in sodium ibuprofen solu-tions of several concentrations (1–8%) showed a decreasein their volume of 40%, which is related to neutralization ofsome ionic charges and to an increase in the ionic strength ofthe medium. The amount of drug sorbed linearly increasedwith the concentration of ibuprofen in the outer solution(Fig. 6). The isotherm patterns were similar to the G-typemodel (Duro et al., 1999). This means that the interactionsbetween the polymer and the drug are relatively weak. Nev-ertheless, the hydrogels were able to take up a considerablyhigh amount of drug, up to 15 mg/mg dry hydrogel, except

436 R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438

Ibuprofen (%)0 1 2 3 4 5 6 7 8

0

2

4

6

8

10

12

14

16

Ibup

rofe

n lo

aded

(g

/g d

ry g

el)

0

2

4

6

8

10

12

14

16PQ-10PQ-4

E-261E-14

Fig. 6. Ibuprofen sorption isotherms on chemically cross-linked cationiccelluloses and cationic guar gums.

for E-14 hydrogels in which drug sorption was limited to8 mg/mg.

3.2.2. Drug releaseIbuprofen release profiles in water from the hydrogels

loaded in 6, 7, or 8% ibuprofen solutions are shown inFig. 7A. The release process reached equilibrium after 2 h,when a 40–50% of the amount of ibuprofen loaded (60, 72,or 80 mg per disc, respectively) still remained in the hydro-gels, which did not experiment significant changes in volumeduring the process. The release occurred by Fickian diffu-sion (Eq. (4), n = 0.46±0.5, r2 > 0.98). The permanence ofthe drug was confirmed by IR spectroscopy (Fig. 7B). After8 h in water, the hydrogels were dried and the characteris-tic bands of ibuprofen (2956 cm−1 benzene ring; 1584 and1409 cm−1 ionized carboxylic groups) were still clearly vis-ible. Also, the bands corresponding to the carboxylic groupsof ibuprofen were slightly moved to lower wave numbers,suggesting ionic interactions with the polymer. Therefore,only the unbound or weakly bound drug molecules are re-leased from the hydrogel. The remaining drug molecules areionically trapped by the ammonium groups of the polymernetwork, similarly to that observed byJimenez-Kairuz et al.(2002) for lidocaine in carbopol gels. Note that ibuprofenconcentration is, in all cases, clearly below its solubility co-efficient in water (Cs = 20%;Bustamante et al., 2000). Anincrease in pH and the presence of more ions in the mediumfacilitate the diffusion of all ibuprofen, by acting as counter-

Time (min)0 60 120 180 240

Ibup

rofe

n re

leas

ed (

%)

0

10

20

30

40

50

60

Wavenumbers (cm-1)

60012001800240030003600

(A)

(B)

Fig. 7. (A) Ibuprofen release profiles in water from PQ-10 hydrogelsloaded in 6% (�), 7% (�), and 8% () ibuprofen solutions; and (B) fromtop to the bottom, FT-IR spectra of freshly prepared PQ-10 hydrogel,of PQ-10 hydrogel after ibuprofen loading in 6% solution and release inwater, and of pure ibuprofen sodium.

ions. This explains that when the hydrogels were transferredto pH 8 phosphate buffer, all ibuprofen was released. Thisallowed to confirm the total amount of ibuprofen initiallyloaded by the hydrogels.

Fig. 8 shows the release profiles of hydrogels subse-quently immersed in media of increasing pH, simulating thein vivo situation after oral administration. At pH 3 (dilutedHCl aqueous solution), the hydrogels just released a low

Time (min)0 60 120 180 240 300 360 420 480 540 600

Ibup

rofe

n re

leas

ed (

%)

0

20

40

60

80

100pH 3 pH 6 pH 8

Fig. 8. Effect of pH on ibuprofen release rate from PQ-10 (�), PQ-4(�), E-14 (�), and E-261 (�).

R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438 437

proportion of ibuprofen. The lower solubility of the drugin this media may contribute to hinder the release processdirectly and, indirectly, by promoting hydrophobic interac-tions with the polymer network. When they were transferredto water, there was no release at all; the hydrogels behaveddifferently from when they were directly immersed in wa-ter. Finally, when the pH increased to 8, the release processbegan again and finished after 6 h. The release exponents,n, in pH 8 phosphate buffer were significantly lower than inwater, especially for hydrogels based on the polymers witha greater content in cationic groups (0.45 for PQ-4, 0.27 forPQ-10, 0.29 for E-14, 0.40 for E-261). This finding sug-gests that, at least phenomenologically, the release processis different in both media. Similar results were observedfor another NSAID molecule, diclofenac (Rodrıguez et al.,2003b). At pH below the pKa of ibuprofen (4.5;Bustamanteet al., 2000), in addition to some ionic interactions, hy-drophobic interactions are strongly promoted. This strongerassociation is maintained in water and broken when the pHand ion concentration rise.

4. Conclusions

Cationic cellulose ethers and cationic guar gums interacthydrophobically and ionically with ibuprofen. The associa-tion phenomena are directly related to the self-aggregationof the drug molecules; dramatically increasing the intensityof the interactions above thecmc. This behavior notably dif-fers from that previously reported for the association pro-cesses with common anionic surfactants, in which the mainchanges in the properties of cationic polysaccharide systemsoccur when the charges of the polymer are neutralized bythe surfactant. Ibuprofen-polymer interactions cause impor-tant modifications in the rheological properties of the poly-mer dispersions and considerably delay the drug diffusionrate. Moreover, the addition to the gel of small amounts ofSDS, which causes a slight decrease in viscosity, provokesa reduction in the values of the diffusion coefficients of al-most 50%. In the presence of the surfactant, new bondsare expected between the drug and the surfactant and be-tween the mixed micelles of polymer/drug/surfactant. Theinformation provided by these studies was particularly use-ful in the design of chemically cross-linked hydrogels ableto control the release process of ibuprofen as a function ofthe pH of the medium. At pH 3, hydrophobic and ionicbonds are strongly promoted and the drug is not released.The association was partially broken in water and totallyvanished in pH 8 phosphate buffer. In summary, the inter-action of cationic polysaccharides with anionic amphiphilicmolecules has considerable consequences on the consistencyand diffusional properties of their aqueous dispersions andchemically cross-linked hydrogels, and highlights the needfor a profound characterization of the associative phenom-ena as a first step in the development of drug delivery sys-tems with these components.

Acknowledgements

This work was financed by the Xunta de Galicia (PGIDT00PX120303PR) and the Ministerio de Ciencia y Tec-nologıa, Spain (RYC2001-8). The authors also express theirgratitude to the Xunta de Galicia for an equipment grant(DOG 04/06/97) and to National Starch and Chemical Ltd.for providing free samples of cationic celluloses.

References

Alvarez-Lorenzo, C., Concheiro, A., 2003. Effects of surfactants on gelbehavior. Design implications for drug delivery systems. Am. J. DrugDeliv. 1, 77–101.

Annon, 1988. Final report on the safety assessment of polyquaternium-10.J. Am. Coll. Toxicol. 7, 335–351.

Anthony, O., Marques, C.M., Richetti, P., 1998. Bulk and surface behav-ior of cationic guars in solutions of oppositely charged surfactants.Langmuir 14, 6086–6095.

Bustamante, P., Peña, M.A., Barra, J., 2000. The modified extendedHansen meted to determine partial solubility parameters of drugs con-taining a single hydrogen bonding group and their sodium derivatives:benzoic acid/Na and ibuprofen/Na. Int. J. Pharm. 194, 117–124.

Capitan, D., Nobile, M.A., Mensitieri, G., Sannino, A., Segre, A.L.,2000. 13C solid state NMR determination of cross-linking degree insuperabsorbing cellulose-based networks. Macromolecules 33, 430–437.

Chronakis, I.S., Alexandridis, P., 2001. Rheological properties of op-positely charged polyelectrolyte–surfactant mixtures. Effect of poly-mer molecular weight and surfactant architecture. Macromolecules 34,5005–5018.

Dominkus, M., Nicolakis, M., Kotz, R., Wilkinson, F.E., Kaiser, R.R.,Chlud, K., 1996. Comparison of tissue and plasma levels of ibuprofenafter oral and topical administration. Arzeneimittel.-Forschung./DrugRes. 46, 1138–1143.

Duro, R., Souto, C., Gómez-Amoza, J.L., Martınez-Pacheco, R.,Concheiro, A., 1999. Interfacial adsorption of polymers and surfac-tants: implications for the properties of disperse systems of pharma-ceutical interest. Drug Dev. Ind. Pharm. 25, 817–829.

Florence, A.T., Attwood, D., 1998. Physicochemical Principles of Phar-macy, third ed. Macmillan, London, pp. 199–251.

González-Rodrıguez, M.L., Holgado, M.A., Sanchez-Lafuente, C.,Rabasco, A.M., Fini, A., 2002. Alginate/chitosan particulate systemsfor sodium diclofenac release. Int. J. Pharm. 232, 225–234.

Goddard, E.D., Hannan, R.B., 1977. Polymer/surfactant interactions. J.Am. Oil Chem. Soc. 54, 561–566.

Gruber, J., Kreeger, R., 1996. Cellulose ethers, cationic. In: PolymericMaterials Encyclopedia, vol. 2. CRC Press, New York, pp. 1113–1118

Higgins, J.D., Gilmor, T.P., Martelucci, S.A., Bruce, R.D., 2001. Ibuprofen.In: Analytical Profiles of Drug Substances and Excipients, vol. 27.Academic Press, San Diego, pp. 265–300

Higuchi, W.I., 1962. Analysis of data on the medicaments release fromointments. J. Pharm. Sci. 51, 802–804.

Hoffman, H., Kästner, U., Donges, R., Ehrler, R., 1996. Gels from mod-ified hydroxyethylcellulose and ionic surfactants. Polym. Gels Netw.4, 509–526.

Hugerth, A.M., 2001. Micropolarity and microviscosity of amitripty-line and dextran sulfate/carrageenan-amitriptyline systems: thenature of polyelectrolyte–drug complexes. J. Pharm. Sci. 90, 1665–1677.

Jimenez-Kairuz, A.F., Allemandi, D.A., Manzo, R.H., 2002. Mechanismof lidocaine release from carbomer-lidocaine hydrogels. J. Pharm.Sci. 91, 267–272.

438 R. Rodr´ıguez et al. / European Journal of Pharmaceutical Sciences 20 (2003) 429–438

Korsmeyer, R.W., Peppas, N.A., 1981. Effect of the morphology ofhydrophilic polymeric matrices on the diffusion and release of watersoluble drugs. J. Membr. Sci. 9, 211–227.

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

Paulsson, M., Edsman, K., 2001. Controlled drug release from gels usingsurfactant aggregates. II. Vesicles formed from mixtures of amphiphilicdrugs and oppositely charged surfactants. Pharm. Res. 18, 1586–1592.

Paulsson, M., Edsman, K., 2002. Controlled drug release from gels usinglipophilic interactions of charged substances with surfactants andpolymers. J. Colloid Interface Sci. 248, 194–200.

Pouliquen, G., Porcar, I., Tribet, C., Amiel, C., 2003. Photoresponsivethickening in polyamphiphile-based physical gels: the examples ofmicelle, protein, and cyclodextrin cross-linkers. ACS Symp. Ser. 833,262–288.

Ridell, A., Evertsson, H., Nilsson, S., Sundelöf, L.O., 1999. Amphiphilicassociation of ibuprofen and two nonionic cellulose derivatives inaqueous solution. J. Pharm. Sci. 88, 1175–1181.

Rodrıguez, R., Alvarez-Lorenzo, C., Concheiro, A., 2001. Rheologi-cal evaluation of the interactions between cationic celluloses andCarbopol 974P in water. Biomacromolecules 2, 886–893.

Rodrıguez, R., Alvarez-Lorenzo, C., Concheiro, A., 2003a. Influence ofcationic cellulose structure on its interactions with sodium dodecyl-sulfate: implications on the properties of the aqueous dispersions andhydrogels. Eur. J. Pharm. Biopharm. 56, 133–142.

Rodrıguez, R., Alvarez-Lorenzo, C., Concheiro, A., 2003b. Cationiccellulose hydrogels: kinetics of the cross-linking process and char-acterization as pH-/ion-sensitive drug delivery systems. J. Control.Release 86, 253–265.

Scherlund, M., Brodin, A., Malmstem, M.J., 2000. Nonionic celluloseethers as potential drug delivery systems for periodontal anesthesia.Colloid Interface Sci. 229, 365–374.

Taboada, P., Ruso, J.M., Garcıa, M., Mosquera, V., 2001. Comparisonof the thermodynamic properties of structurally related amphiphilicantidepressants in aqueous solution. Colloid Polym. Sci. 279, 716–720.

Yomota, C., Okada, S., 2003. Various interactions of drugs withcross-linked hyaluronate gel. ACS Symp. Ser. 83, 326–338.