interaction of clonixin with epc liposomes used as membrane models

Interaction of Clonixin with EPC Liposomes Usedas Membrane Models

HELENA FERREIRA,1 MARLENE LUCIO,1 JOSE L.F.C. LIMA,1 CARLA MATOS,2 SALETTE REIS1

1REQUIMTE, Departamento de Quımica-Fısica, Faculdade de Farmacia, Universidade do Porto,Rua Anıbal Cunha, 164, 4050-047 Porto, Portugal

2REQUIMTE, Faculdade Ciencias da Saude, Universidade Fernando Pessoa, Rua Carlos da Maia,no. 296, 4200-150 Porto, Portugal

Received 15 December 2004; revised 15 February 2005; accepted 16 February 2005

Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20351

ABSTRACT: In this work, an overall analysis of clonixin interaction with liposomes wasachieved using different techniques, which allowed the evaluation of the change indifferent membrane’s characteristics as well as the possible location of the drug in themembrane. Clonixin acidity constants were obtained and the values are 5.5� 0.08 and2.2� 0.04. Clonixin partition coefficient (Kp) between liposomes and water was alsodetermined using derivative spectrophotometry, fluorescence quenching, and zeta-potential (z-potential). These three techniques yielded similar results. z-potentialmeasurements were performed and an increase of the membrane negative charge withan increase of drug concentration was observed. Drug location within the bilayer wasperformed by fluorescence quenching using a set of n-(9-anthroyloxy) fatty acid probes(n¼ 2, 6, 9, and 12). The fluorescence intensity of all probes was quenched by the drug.This effect is more noticeable for the outer located probe, indicating that the drug ispositioning in the external part of the membrane. These same probes were used forsteady-state anisotropy measurements to determine the perturbation in membranestructure induced by clonixin. Clonixin increased membrane fluidity in a concentrationdependent manner, with the highest perturbation occurring nearby the 2-AS probe,closely located to the bilayer surface. � 2005 Wiley-Liss, Inc. and the American Pharmacists

Association J Pharm Sci 94:1277–1287, 2005

Keywords: clonixin; liposomes; drug interaction; light scattering; UV/Vis spectro-scopy; fluorescence spectroscopy; partition coefficient; drug location; membrane fluidity

INTRODUCTION

Cyclooxygenase (COX), a membrane relatedenzyme, is the pharmacological target of non-steroidal anti-inflammatory drugs (NSAIDs),which are therefore commonly used in inflamma-tory diseases treatment. Although low doses ofNSAIDs inhibit prostaglandins biosynthesis, highconcentrations interfere with processes not de-pendent on these mediators. Membrane related

phenomena, such as neutrophil function inhi-bition, oxidative phosphorylation inhibition inmitochondria, signal transduction disruption,and the consequent interference with intracel-lular calcium mobilization and protein kinase Cactivity alteration, have all been reported byKlein et al.,1 as well as a membrane fluidityalteration, which has been mentioned by severalauthors.2–5 Furthermore, NSAIDs have beenshown to inhibit the cellular proliferation rate,to alter the cell cycle regulation, and to induceapoptosis in cancer cell lines, in a mechanism in-dependent from prostaglandin pathways.1

There is consensual evidence that the lipidaffinity of the NSAIDs is of major significance

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 6, JUNE 2005 1277

Correspondence to: Salette Reis (Telephone: þ351-222-078-966. Fax: þ351-222-004-427; E-mail: [email protected].)

Journal of Pharmaceutical Sciences, Vol. 94, 1277–1287 (2005)� 2005 Wiley-Liss, Inc. and the American Pharmacists Association

for their toxic and therapeutic actions. Indeed,depending on their hydrolipophilic character,NSAIDs can be distributed between the mem-brane and the aqueous phases. This distributiondetermines their concentration in each phase andthereby controls the extents of their penetrationinto the membrane and/or their interactions withphospholipids or other membrane components,such as COX enzymes, which are embedded inthe lipid bilayers.6 Thus, for the study of NSAIDs’action mechanisms and their side-effects, it is ofgreat importance to investigate the interactionsbetween these drugs and biomembranes. For thispurpose, this work was performed using lipo-somes of egg yolk phosphatidylcholine (EPC).Liposomes are generally accepted to be a suitablemodel for the study of membrane structure andproperties, because they are surrounded by a lipidbilayer structurally similar to the lipidic matrixof the cell membranes.6,7 Additionally, because ofbeing constituted by natural lipids, EPC liposomescan mimic the chemical and structural anisotropicenvironment of cell membranes. EPC liposomesalso appear to mimic the interfacial character aswell as the ionic, H-bond, dipole–dipole, andhydrophobic interactions, which may define parti-tioning in real biomembranes.6,7

Traditionally, the octanol–water partition coef-ficient (Kp) has been used to measure compounds’hydrophobicity, which is correlated to drug activ-ity. The octanol–water system is a good membranemodel when polar group interactions between thesolute and the phospholipid bilayer are minimal orabsent. However, since octanol can only modelnon-polar interactions,8 better systems are neededfor molecules which can establish electrostatic in-teractions with polar groups in the membrane.According to this, the study of clonixin’s partitionin a liposome/buffer system has been performed.There is a more satisfactory correlation betweenthis parameter and its pharmacological propertiessince clonixin has proved to be able to establishelectrostatic interactions with polar groups in thebiomembranes. The drug’s Kp was evaluated byderivative spectrophotometry, a technique thatcan be used when a solute’s spectral characteristicchanges between one media to another. Deriva-tive spectrophotometry eliminates the intensebackground signals that arise from light scatteredby lipid vesicles, and it also improves the resolu-tion of overlapping signals reported by severalauthors.9–12 Moreover, the liposome/water Kp wasalso determined by other experimental techni-ques: zeta-potential (z-potential) and fluorescence

quenching. Using the z-potential technique, it waspossible to evaluate the interaction of clonixinwith liposomes by measuring the membranepotential arising from the drug partitioning. Infact, biological membranes are charged, due toionized components (lipids, glycolipids, glycopro-teins), and the resulting surface potential plays acritical role in regulatory processes, membrane–membrane interactions, and in their binding capa-city to solutes in solution.13,14 Additionally, toelectrostatic effects, which can affect the confor-mation and activity of membrane and membrane-bound enzymes,15,16 several cell processes are alsorelated to electrostatic or polarization effects onthe cell membrane. In this context, the character-ization of the electrostatic membrane propertiesinduced by clonixin binding is a fundamentalparameter, and it also allows the quantificationof clonixin molecules in these membranes. Conse-quently, Kp values can be calculated.

Fluorescence quenching was also used to mea-sure clonixin’s coefficient partition. The fluores-cent n-(9-anthroyloxy)-stearic acids (n-AS, n¼ 2,6, 9, and 12) are the set of probes most widelyused for obtaining information on molecularaggregates, such as liposomes and natural mem-branes.17–22 For these probes, there is evidencethat the anthroyloxy fluorophores are located ata graded series of depths inside a membrane,depending on its substitution position (n) in thealiphatic chain.17 Therefore, these probes, due totheir exceptional environmental sensitivity, havebeen employed to monitor the microenvironmentof membranes. These appropriate measurementsallow information about the local membranestructure to be inferred. According to this, besidesthe determination of clonixin’s Kp, fluorescencequenching provides a mean to evaluate the posi-tion and orientation of the drug in the membraneby a comparative analysis of all probe’s quenching.Furthermore, the fluorescent probes are capable ofsensing a ‘‘fluidity’’ gradient through the bilayerleaflet and, therefore, they were used to assess theclonixin effect in the lipid membrane fluidity. Thiswas achieved using steady-state anisotropy mea-surements, since that anisotropy depends uponthe rotational motion of the fluorophore and it issensitive to hindrance forces imposed by themicroenvironment, property that has been widelyused to estimate membrane fluidity.18 Membranefluidity assessment gives useful physiologic infor-mation as biomembranes need to be in a fluid statein order to maintain complete biological function.Indeed, any alteration in membrane fluidity tends

1278 FERREIRA ET AL.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 94, NO. 6, JUNE 2005

to change movement and/or orientation of pro-teins floating within lipid bilayer, as reported byDave and Witorsch.23 Becceria et al.3 described adecrease in lymphocytes membrane fluidity inrheumatoid arthritis, and suggested membranefluidification as a good indicator of NSAID’s ther-apeutic effects. Another important aspect inphysiological activity, is the acid–base propertiesof drugs. Although little has been published aboutclonixin’s acid–base chemistry and lipophilicity,this characterization is necessary for thoroughunderstanding of its pharmacokinetic and phar-macodynamic parameters.

Finally, clonixin has been chosen for this studybecause, besides its anti-inflammatory effect, it isusually prescribed as an analgesic. Therefore, it isof great interest to investigate the behavior ofclonixin to compare with others classic NSAIDs,like nimesulide,24 once that therapeutic and toxicactivities can differ in structurally unrelated andhomologous NSAIDs.

EXPERIMENTAL SECTION

Reagents and Equipment

The anti-inflammatory drug clonixin was gener-ously supplied by its manufacturer Janssen-CilagPharmaceutica (Barcarena, Portugal) and wasused without further purification. The EPC and(� )-12-(9-anthroyloxy)-stearic acid (12-AS) werepurchased from Sigma Chemicals (St. Louis, MO);the other probes, (� )-2-(9-anthroyloxy)-stearicacid, (� )-6-(9-anthroyloxy)-stearic acid, and(� )-9-(9-anthroyloxy)-stearic acid (2-AS, 6-AS,and 9-AS) were purchased from Molecular Probes(Eugene, OR). All of these were used as supplied.All the other chemicals were from Merck (Darm-stadt, Germany) with pro analysi grade. Solutionswere prepared with double-deionized water (con-ductivity less than 0.1 ms/cm). The ionic strengthof all solutions was adjusted to 0.1M with NaCl.

The polycarbonate filters with a diameter poreof 100 nm (Nucleopore) were obtained from What-man (Maidstone, England).

Absorption spectra were recorded at 25.0�0.18C on a Perkin-Elmer Lambda 45 UV/VISspectrophotometer in the range 220–500 nm with1 nm intervals (a Hitachi U-2000 dual-beam spec-trophotometer, a 220–400 nm range, and 2 nmintervals for determination of acidity constants).In all the cases, quartz cells were used and thetemperature was kept constant by circulatingthermostated water in the cell holder.

Fluorescence studies were carried out on aPerkin-Elmer LS 50B steady-state fluorescencespectrometer equipped with a constant-tempera-ture cell holder. All data were recorded at25.0� 0.18C in 1-cm cuvettes with excitation andemission slits between 4.0 and 9.0 nm. Excitationwavelength was set to 384 nm and emissionwavelength to 446 nm for 12-AS and 9-AS probes,and to 451 nm for 6-AS and 2-AS ones. Fluo-rescence intensity data were corrected for absor-bance of the quencher (clonixin) at the excitationwavelength.25

z-potential values and size distribution of ex-truded EPC liposomes, with and without incorpo-rated drug, were determined at pH 7.4 (HEPESbuffer), at 25.0� 0.18C, by quasi-elastic lightscattering analysis using a ZET 5104 cell in aMalvern ZetaSizer 5000, with a 908scatteringangle.

Spectrophotometric Determinationof the Acidity Constants

Clonixin acidity constants were obtained fromUV/Vis data in aqueous solution with ionicstrength adjusted to 0.1M by addition of NaCl.The �log [Hþ] value was consecutively changedby potentiometric titration of 25.00 mL of anacidified aqueous solution (with hydrochloric acid30%) of the drug (approximately 30 mM) withNaOH (0.02M), under a nitrogen stream. Thepotentiometric system calibration was performedby the Gran method,26 in terms of hydrogen ionconcentration, by titrating solutions of strong acid(10�3M HCl) with strong base (0.02M NaOH).Absorption spectra were recorded in the systemdescribed in ‘‘Reagents and Equipment.’’ Calcula-tions were performed with data obtained from, atleast, four independent experiments using theprogram pHab.27

Liposome Preparation and Drug Incorporation

Liposomes were prepared by the thin film hydra-tion method. According to this method, a knownamount of EPC was dissolved in chloroform/methanol (9:1). The organic solvent was evapo-rated under a nitrogen stream and the residualtraces of solvent were removed by a further eva-poration for, at least, 3 h under the same stream.The resulting dried lipid film was dispersed by theaddition of the buffers with different pH values(pH 3.0: 34.6 mM hydrochloric acid, 50 mM glycin;pH 7.4: 10 mM HEPES; pH 10.3: 43.2 mM sodium

INTERACTION OF CLONIXIN WITH EPC LIPOSOMES 1279


hydroxide, 50 mM boric acid; all buffers withI¼ 0.1M and for pH 3.0 and 10.3 buffers the finalpH was adjusted by addition of strong acid orbase). These mixtures were then vortexed abovetheir phase transition temperature (room tem-perature) to produce multilamellar liposomes(MLVs). MLVs suspensions were then equili-brated for 30 min and extruded ten times throughpolycarbonate filters with a pore diameter of100 nm (Nucleopore)6 at room temperature toproduce large unilamellar vesicles (LUVs). EPCconcentration in vesicle suspensions was deter-mined by phosphate analysis using the phospho-molybdate method.28

In fluorescence studies, the fluorescence probesdissolved in ethanol were gently added to aliposome suspension to achieve a final probe tolipid ratio smaller than 1:100, to prevent changesin membranes structure. To ensure completeincorporation of the probe in the lipid bilayer, thesuspensions were left to stand in the dark for30 min.29 Subsequent assays of liposome charac-terization by z-size measurements were made toconfirm that no changes occurred in the fluores-cent marked vesicles.

After the liposome preparation and/or labeling,the drug samples were prepared by mixing aknown volume of the drug to a suitable aliquot ofvesicle suspension in buffer. The correspondentreference solutions were identically prepared, inthe absence of drug. All suspensions were thenvortexed for 5 min and incubated for 30 min atroom temperature.

Kp Determination byDerivative Spectrophotometry

Clonixin Kp was determined in LUVs suspensionsat pH 3.0, 7.4, and 10.3. In the derivativespectrophotometry studies, a series of bufferedsuspensions containing a fixed concentration ofdrug (30 mM) and increasing concentrations ofEPC (in the range 70–1200 mM) were prepared.The correspondent reference solutions were iden-tically prepared in the absence of drug. Allsuspensions were then vortexed and incubatedin the dark for 30 min at room temperature andthe absorption spectra were recorded.

z-Potential and Size Determinations

z-potential values and size distribution of ex-truded EPC liposomes, with and without incorpo-rated drug, were determined at pH 7.4 (HEPES

buffer), at 25.0� 0.18C. Lipid concentrationwas kept constant at approximately 400 mM.Clonixin concentrations ranged between 0 and500 mM. The values for viscosity and refrac-tive index were taken as 0.890 cP and 1.330,respectively.30

Particle size of the pure vesicles was found tobe 133� 5 nm (average and standard deviation ofthe measurements of six independently preparedsuspensions).

Kp Determination by Fluorescence Quenching

In fluorescence quenching studies, liposomes pre-pared with the 2-AS probe were added to HEPES(pH 7.4, 10 mM, I¼ 0.1M) buffered solutions ofclonixin as described above. The EPC concentra-tion of the liposomes ranged between approxi-mately 80 and 900 mM and drug concentrationbetween 0 and 500 mM. The resulting suspensionswere incubated in the dark for 1 h at roomtemperature.

Membrane Fluidity and Drug Location Studies

Membrane fluidity was estimated by fluorescenceanisotropy while drug location was determined byfluorescence quenching; in both cases all n-ASfluorescence probes in LUV suspensions at pH 7.4were used. The EPC concentration was approxi-mately 500 mM. HEPES (pH 7.4, 10 mM, I¼ 0.1M)buffered solutions of clonixin were added to theliposomes prepared with the n-AS probes as pre-viously described. Drug concentrations were inthe range of 0 to 500 mM.

RESULTS AND DISCUSSION

Spectrophotometric Determinationof Acidity Constants

Clonixin contains two proton-binding sites (carbo-xylate group and aminopyridinyl moiety) and canexist in four protonated forms in solution. Proto-nation equilibria suggested are shown in Figure 1,in a similar fashion to what has been suggest-ed for a structurally similar molecule, niflumicacid.31 Neutral species are presumed to exist in anegligible percentage, in detriment of the zwit-terionic species, and from this assumption, onlythree protonated states are considered: positive,zwitterionic, and negative species.

Due to the low solubility of clonixin, acidityconstants were determined by spectrophotometry.



This method can be used to determine the clonixinpKa values, since the protonation of one or otherfunctional group by changing the solution pH,results in spectral changes. The pKa1 and pKa2

values of clonixin obtained were 5.5� 0.08 and2.2� 0.04, respectively. These results resemblethe ones obtained by Takacs-Novak et al. withniflumic acid31 and can be used to calculate the pH-dependent percentages of the species. Therefore,at blood physiological pH (pH¼ 7.4), clonixinexists predominantly (98.7%) in the negative form,increasing this percentage with the pH value.

Kp Determination byDerivative Spectrophotometry

Kp of a compound between vesicle suspensionsand the aqueous solution is defined as:30

Kp ¼AT� �

m

AT½ �wð1Þ

where [AT]m is the local drug concentration of A inthe lipid phase and [AT]w in the water phase basedon lipid and water solution volumes, Vm and Vw.

The concentration of drug partitioned in eachphase can be determined by UV spectrophotome-try, by non-linear regression of the data collectedin different EPC concentrations, providing thatthe solute’s spectral characteristics (e and/or lmax)change when it permeates from the aqueous to thelipid phases. However, it is not usually possible toobtain Kp using direct spectroscopic methods,owing to intense background signals arising fromlipid vesicles light scattered. Derivative spectro-photometry eliminates the effect of backgroundsignals and it improves the resolution of over-lapping bands.9–12

The derivative UV spectra of clonixin (Figure 2)show a decrease in absorption intensities inthe presence of increasing amounts of EPC. Fur-thermore, the absorption spectra of clonixin inEPC exhibit isosbestic points and a bathochromicshift in lmax with increasing lipid concentration(Figure 2). These observations suggest that thedrug exists in two forms: drug in polar bulk waterand in EPC bilayers.9,11

Derivative intensities can be related to thepartition coefficient by the following expression:

AbsT ¼ Absw þ Absm � Abswð ÞKp L½ �Vf

1 þ Kp L½ �Vfð2Þ

where AbsT, Absm, and Absw are the total, lipid,and aqueous absorbances of the drug, respec-tively, Kp is the partition coefficient, [L] is thelipid concentration, and Vf the lipid molarvolume. For EPC, the mean molecular weightwas considered to be 700 g/mol and Vf to be 0.688L/mol.32 This equation was fitted to the experi-mental second and third derivative spectrophoto-metric data (DT versus [L]) using a non-linearleast-squares regression method (Figure 3), atwavelengths where the scattering is eliminat-ed. The value of Kp obtained in HEPES buffer(pH¼ 7.4) was 660� 100. Kp were also determin-ed also at pH 3.0 and 10.3, and the resultsobtained were 530� 100 and 870� 50, respec-tively (mean and standard deviation of at leasttwo independent assays). Results show that,

Figure 1. Protonation equilibria of clonixin.

Figure 2. Second and third derivative spectra ofclonixin at different concentrations of egg yolk phos-phatidylcholine (EPC): (1) 0; (2) 86; (3) 181; (4) 272; (5)352; (6) 441; (7) 525; (8) 629; (9) 764; (10) 851; (11) 872;(12) 958; (13) 1127 mM.



despite what was observed for other NSAIDs inprevious articles,24,30,33,34 the Kp values do notdramatically change with the change in mediumpH. Since clonixin shows electrical chargedgroups at any pH value its hydrophilicity is main-tained and its affinity to the lipid bilayer does notsignificantly change.

z-Potential Determinations

Since clonixin is predominantly in its negativestate at pH 7.4, its partition in the liposomes leadsto the formation of a charged membrane surface.z-potential values, z, determined in different drugconcentrations, can be used to calculate surfacepotential values (C0); and these allow the deter-mination of surface charge density on the mem-brane, expressed in number of charged moleculesper area unit, s*.30

The determination ofKp from z-potential resultscan be achieved using Langmuir isotherms:30

s� ¼ K�ps

�max A�

i

� �ð3Þ

where, A�i

� �the concentration at the interface can

be obtained using the Boltzmann equation.35 Theapparent partition coefficient K�

p can be trans-formed in Kp by:30

K�p ¼ Kp �

½L�aA þ �nn½L�aL

½Am�aA þ ½L�aLð4Þ

where aA and aL represent the molecular area forthe drug and lipid, respectively, and �nn stands for

the maximum mole lipid:drug ratio that is, themembrane’s loading capacity. The values of s�

max

can be obtained from the plot of s� versus theconcentration of total drug added to the system,[AT], fitting the binding isotherm:36

y ¼ Kb½AT�1 þ Kb½AT� ð5Þ

in which y, the degree of saturation, is given byy ¼ s�=s�

max. Knowledge of s�max allows the calcu-

lation of �nn: 30

s�max ¼ 1

ðaA þ �nn aLÞð6Þ

Results for potential measurements obtained inHEPES buffer (pH 7.4) were obtained at increas-ing clonixin concentrations, and used to calculateKp by the mathematical formula described before.The values of s�

max were determined by fittingthe Equation 6 the plot of s� versus [AT], as canbe observed in Figure 4, which shows an examplefor this determination at several concentrationsof clonixin. The value of s�

max obtained was 1.20�0.20�10�3 molecules/A2 (average and standarddeviations of three independent experiments).The value of was calculated by Equation 6, as-suming aL as 60 A2,37 and as aA (molecular surfacearea of clonixin) was not determined, one hasassumed the same value (60 A2), because, asdescribed previously, this parameter does not

Figure 3. Second-derivative spectrophotometric dataat l¼ 316 nm for clonixin at different EPC concentra-tions (0; 86; 181; 272; 352; 441; 525; 629; 764; 851; 872;958; and 1127 mM). The curve represents the best fit toEquation 2.

Figure 4. Dependence of number of charged mole-cules per area unit (s*) with concentration of clonixin atpH 7.4, in the presence of 400 mM of EPC. The curverepresents the best fit to Equation 5.



significantly affect the results obtained.30 Thevalue calculated for �nn was 14.5� 2.0.

The results obtained show a decrease in Kp

values with the increase in drug concentration(from 580 for 20 mM to 240 for 500 mM), indicatingthat the influence of electrostatic effects on Kp

values cannot be neglected for higher drug con-centrations. The value presented in Table 1 is theaverage of the values obtained for clonixin con-centrations lower than 50 mM (the order ofmagnitude of the drug concentrations used in theother presented techniques).

Kp Determination by Fluorescence Quenching

Quenching properties provide a means to deter-mine quantitatively the Kp of the drugs betweenlipid and aqueous phases. For this purpose, thecapacity of clonixin to quench the fluorescence ofthe probe 2-AS was evaluated by determination ofthe apparent Stern–Volmer constant, Kapp

sv , atdifferent lipid concentrations. The Kapp

sv valuesdepend not only on the quencher efficiency butalso on its Kp between the aqueous and the lipidphase, since only the quencher molecules in themembrane are responsible for quenching. Thisdependence can be described by the equation:38

Kappsv ¼ Ksv

Kp

KpVm þ 1ð7Þ

The knowledge of Kappsv for several lipid concentra-

tions (Figure 5) allows the determination ofclonixin Kp and Stern–Volmer constant (KSV),by fitting Equation 7 using a non-linear regres-sion method. The Kp value obtained is included inTable 1.

Drug Location

Measurements of the clonixin quenching efficien-cies, by addition of increasing amounts of drug to

liposome suspensions incorporating the fluores-cent n-AS probes, in which an ester linkageattaches the fluorescent anthracene at differentpositions along the fatty acid chain, is a sensitivemethod of determining the relative position ofquenching molecules in the lipid bilayer. When n-AS probes are included in the lipid bilayer, thecarboxyl terminal group is located at the inter-facial region of the membrane and the anthracenegroup is located in precise and known positionsalong the membrane depth plane.34

The fluorescence intensity of n-AS probesdecreases with an increase of drug concentration.Figure 6 shows this behavior for the 12-AS probe inEPC unilamellar liposomes.

Table 1. Partition Coefficients (Kp) (Adimensional)for Clonixin in EPC Unilamellar Liposomes (LUV) atpH 7.4, by Derivative Spectrophotometry, FluorescenceQuenching, and z-Potential*

Method Kp

Derivative spectrophotometry 660� 100Fluorescence quenching 670� 130z-potential 580� 120

*The reported values are the mean of at least twoindependent measurements; the error that affects each valueis the standard deviation.

Figure 5. Apparent Stern–Volmer constants ðKappSV Þ

of 2-AS for clonixin in EPC unilamellar liposomes (LUV)obtained for different lipid concentrations ([L]).

Figure 6. Fluorescence quenching of 12-AS probe inEPC unilamellar liposomes (500 mM, pH 7.4) by clonixin.



The existence of fluidity and polarity gradientsthrough the plane of the liposomes membrane asmentioned by Thulborn et al.18,39 is reflected indifferent t0 values found for each probe. However,the effect of the microenvironment differencessurrounding the probes is eliminated in bimole-cular constant, kq,38 and, therefore, the determi-nation of these values provides an useful tool topredict drug location. As the fluorescence lifetime(t0) for all n-(9-anthroyloxy) stearic acid probesat pH 7.4 have been reported by Vincent et al.,40

and from the knowledge of the determined Kappsv

values, it is possible to determine an apparent kq

that is kapp:

Kappsv ¼ kapp t0 ð8Þ

The comparison of the kapp values obtained foreach probe as a tool to access drug location ispreferable to the comparison of Kapp

sv values, sincein the first the effect of lifetime variation withinthe probe series is eliminated. The location of theanti-inflammatory drug at pH 7.4 was thereforeachieved using kapp (Table 2). The observation ofthese values leads to the conclusion that all n-ASprobes were quenched by clonixin and the relativequenching efficiencies are in the order 2-AS> 6-AS> 9-AS> 12-AS. This suggests that the anti-inflammatory drug is not deeply buried inside thelipid bilayer, but is preferably located near thephospholipid headgroups, probably with an elec-trostatic binding between the negative drug andthe positive pole of the zwitterionic phophatidyl-choline headgroup. This proposed location corro-borates the observation of clonixin’s structure(Figure 1), where one can see a strong quenchinggroup (chloride atom), which is responsible for thequenching of the probes. In addition, clonixinshows electrical charged groups and that can ex-plain the noticeable preference for the probes thatlie near the surface of the phospholipids acyl chain.

Membrane Fluidity Studies

Fluorescence anisotropy variation is a parameterthat has frequently been used to examine the

effect of molecules on the physical properties ofphospholipid bilayers. There is considerable evi-dence indicating that membrane fluidity can beunderstood in terms of the rate and amplitude ofthe anisotropic motion of the phospholipid acylchains. Fluorescent probes are used to report suchmotion.40 The so-called ‘‘fluiditity gradient’’ canbe investigated using the previously mentionedseries of n-AS probes. As reported by New,29 thisprovides labeling at a graded series of depths inthe bilayer. Therefore, fluorescence anisotropystudies with these probes constitute a relativelyeasy means of establishing the depth-dependenceof fluidity.

The method described by Lakowicz38 was usedto measure the steady-state anisotropies (rss), bythe equation:

rss ¼Ivv �GIvhIvv þ 2GIvh

ð9Þ

where G is an instrumental correction factor, Ivvand Ivh are the emission intensities polarizedvertically and horizontally to the direction of thepolarized light.

For a fluorescent molecule dissolved in anisotropic solvent, the anisotropy can be describedby the Perrin equation,38 which expresses theeffect of either lifetime or microviscosity variationson anisotropy. Under certain specific conditions, inwhich the ‘‘out-of-plane’’ motion of the n-AS probesin membranes is totally unhindered, the inten-sity decays exponentially to zero and the Perrinequation can be used.38,40,41 These statements aremade on the assumption that clonixin does notchange the intensity decay of the probe to multi-exponential.

The Perrin equation can be rearranged to yield:

r0 ¼ yþ t0

yþ t0� rss ð10Þ

where t0 and t0 are the fluorescence lifetime of thefluorophore in the absence and presence of thedrug, respectively, and y is the rotational correla-tion time (directly related to local microviscosity).

Table 2. Values of Apparent Stern–Volmer Constants ðKappSV Þ and Apparent

Bimolecular Rate Constants kapp Obtained for Clonixin in Unilamellar Liposomes*

2-AS 6-AS 9-AS 12-AS

KappSV =Mð Þ 8080� 300 7970� 300 8300� 200 8060� 300

kapp � 109ð=MsÞ 2.59� 0.77 2.35� 0.56 2.11� 0.13 1.56� 0.12

*The reported values are the means of two independent measurements; the error is the standarddeviation.



The values of t0 can be easily calculated using thefollowing equation:38

I0

I¼ t0

t0ð11Þ

where I0 and I are the corrected fluorescenceintensity of the fluorophores (n-AS probes) in theabsence and presence of the drug, respectively.The values of y and t0 have been published for then-AS probes used.40

Considering, in Equation 10, rss as the valuemeasured for t0 (the value in the absence ofquencher) and the published values of y and t0

for n-AS probes,40 a curve is generated with thevalues of the corrected anisotropy, r0. These r0

values are the variation of anisotropy that shouldbe obtained due to the lifetime changes of thefluorophore (gives a correction for the influence ofthe probe itself in membrane microviscosity) andare then compared with the experimental rssvalues. From the difference between rss and r0

one can assess the real variation of anisotropycaused by the drug without the illusory effect of theintrinsic variation due to the decrease of probefluorescence lifetime. Once that rss�r0 decreaseswith increasing clonixin concentration, it is con-cluded that a membrane fluidization happenedfor the NSAID studied. The fluidizing effect ratiocan be calculated by: % fluidizing effect¼ [(r0�rss)/r00]� 100.

As clonixin partitions into the membrane andsome fraction of the quencher remains in theaqueous phase, it is necessary to correct the drugconcentration, by the following equation:38

Qm½ � ¼ Kp QT½ �Kpam þ 1 � amð Þ ð12Þ

where [Qm] is the quencher concentration inmembrane and [QT] is the total concentration ofquencher added; Kp is the lipid–water Kp and amis the volume fraction of the membrane phase(am¼Vm/VT; Vm and VT represent the volumes ofthe membrane and water phases, respectively).

Figure 7 shows a plot of % fluidizing effectversus effective drug concentration for each probestudied. The graph shows that the fluidizing effectis practically the same for probes 9-AS, 6-AS, and12-AS, and it is more effective for 2-AS.

As clonixin, from the quenching studies pre-sented, seems to be preferably located near themembrane surface, this ‘‘fluidizing’’ effect near themembrane surface can be explained by a freevolume effect once that clonixin binds to the

headgroup region, increasing the spacing betweenlipid molecules.

In this work, besides the quantification of theinteraction of clonixin with EPC liposomes, onewas able to evaluate the effects of this interactionby studying the changes of membrane potentialand fluidity. The data obtained for clonixins’location and partition are consistent with the z-potential studies, all pointing to a preferableinteraction of this drug with the membrane sur-face. In fact, analyzing the dependence of theclonixin concentration on z-potential values, it isevident that clonixin provokes an appreciabledecrease of the z-potential of the bilayer. Thisconfirms that this compound is negatively chargedat the pH of the studies. Similarly, in previousworks, indomethacin and acemetacin proved alsoto be predominantly in the negative form at the pHstudied and, consequently, they were able todecrease the z-potential of the membrane asclonixin did. In spite of this, indomethacin andacemetacin can reach the inner part of the bilayerand this contrasting location may be explained bythe presence of longer hydrocarbonated chains,which put the strong quenching group (chlorideatom) closely positioned to 12-AS probe.30,34 Onthe other hand, another previously studied NSAID(nimesulide), does not affect the z-potential of themembrane, since it is almost in a neutral form atthe same pH. Therefore, nimesulide can reach theinner part of the bilayer, presenting an invertedorder of probes’ quenching (12-AS> 9-AS> 6-

Figure 7. Dependence of the fluidizing effect ratio (%)for each n-AS probe in EPC unilamellar liposomes(500 mM, pH 7.4) caused by increasing effective concen-trations of clonixin.



AS> 2-AS) when compared to the results obtainedfor clonixin (2-AS> 6-AS> 9-AS> 12-AS).

Both electrostatic and fluidity properties arefundamental for regulatory mechanisms, such asmembrane enzyme activity and electric stimulaeconduction. These in turn, can be directly relatedto inflammatory and analgesic actions of clonixin.In fact, the results show that clonixin can build aconcentration dependent negative charge on themembrane’s surface and it can increase bilayerfluidity substantially, especially on the surfacewhere the drug is preferentially located. Thisagrees with previous studies performed for otherNSAID drugs,24,33,34 where a parallelism betweenlocation and membrane fluidification is also found.However, for the others, NSAIDs studied theirpreferential location and higher membrane fluidi-fication effects are observed in the core of thebilayer, whereas clonixins’ effects are mainlynoticed at the surface.

From the overall studies performed one canconclude that anti-inflammatory activity can berelated to membrane effects as described forclonixin.

ACKNOWLEDGMENTS

The authors thank FCT and FEDER for financialsupport through the contract POCTI/FCB/47186/2002. Some of us, H. F. and M. L., thank FCT forthe fellowships (BD 6829/01) and (BD 21667/99),respectively.

REFERENCES

1. Klein T, Ullrich V, Pfeilschifter J, Nusing R. 1998.On the induction of Cyclooxygenase-2, induciblenitric oxid synthase and soluble phospholipase A2in rat mesangial cells by a nonsteroidal anti-inflammatory drug: The role of cyclic AMP. MolPharmacol 53:385–391.

2. Knazek RA, Liu SC, Dave JR, Keller RJ. 1981.Indomethacin causes a simultaneous decrease ofboth prolactin binding and fluidity of mouse livermembranes. Prostag Med 6:403–411.

3. Beccerica E, Piergiacomi G, Cutarola G, Ferretti G.1989. Effect of antirheumatic drugs on lymphocytemembrane fluidity in rheumatoid arthritis: Afluorescence polarization study. Pharmacology 38:16–22.

4. Beccerica E, Ferretti G, Cutarola G, Cervini C.1990. Diacethylrhein and rhein: In vivo and in vitro

effect on lymphocyte membrane fluidity. Pharma-col Res 22:277–285.

5. Simonetti O, Ferretti G, Offidani AM, Gervasi P,Curantola G, Bossi G. 1996. Plasma membranefluidity of keratinocytes of normal and psoriaticskin: A study using fluorescence anisotropy oftrimethylammonium diphenylhexatriene (TMA-DPH). Arch Dermatol Res 288:51–54.

6. Lasic DD. 1993. Liposomes—From Physics to appli-cations. New York: Elsevier.

7. Mason RP, Rhodes DG, Herbette LG. 1991. Reeval-uating equilibrium and kinetic binding parametersfor lipophilic drugs based on a structural model fordrug interaction with biological membrane. J MedChem 34:869–877.

8. Betageri GV, Rogers JA. 1988. The liposome as adistribution model in QSAR studies. Int J of Pharm46:95–102.

9. Welti R, Mullikin LJ, Yoshimura T, HelmkampGM. 1984. Partition of amphiphilic moleculesinto phospholipid vesicles and human erythrocyteghosts: Measurements by ultraviolet differencespectroscopy. J Biochem 23:6086–6091.

10. Kitamura K, Imayoshi N. 1992. Second-derivativespectrophotometric determination of the bindingconstant between chlorpromazine and cyclodextrinin aqueous solutions. Ana Sci 8:497–501.

11. Kitamura K, Imayoshi N, Goto T, Shiro H, Mano T,Nakai Y. 1995. Second derivative spectropho-tometric determination of partition coefficients ofchlorpromazine and promazine between lecithinbilayer vesicles and water. Anal Chim Acta 304:101–106.

12. Gursoy A, Senyucel B. 1997. Characterization ofciprofloxacin liposomes: Derivative ultraviolet spec-trophotometric determinations. J Microencapsul14:769–776.

13. McLaughlin S. 1989. The electrostatic propertiesof membranes. Annu Rev Biophys Biophys Chem18:113–136.

14. Cevc G. 1990. Membrane electrostatics. BiochimBiophys Acta 1131:311–382.

15. Russel AJ, Fersht AR. 1987. Rational modificationsof enzyme catalysis by engineering surface charge.Nature 328:496–500.

16. Volwerk JJ, Jost PC, Haas GHDe, Griffith OH.1986. Activation of porcine pancreatic phospholi-pase A2 by the presence of negative charges atthe lipid–water interface. Biochemistry 25:1726–1733.

17. Blatt E, Sawyer WH. 1985. Depth-dependent fluo-rescent quenching in micelles and membranes.Biochim Biophys Acta 822:43–62.

18. Thulborn KR, Tilley LM, Sawyer WH, Treolar FE.1979. The use of n-(9-anthroyloxy) fatty acids todetermine fluidity and polarity gradients in phos-pholipid bilayers. Biochim Biophys Acta 558:166–178.



19. Melo ECCA, Costa SMB. 1987. A critical evaluationof partition constants in nonionic micelles. J PhysChem 91:5635–5640.

20. Katsikas H, Quinn PJ. 1983. Fluorescence probestudies of the distribution of ubiquinone homolo-gues in bilayers of dipalmitoylglycerophosphocho-line. Eur J Biochem 131:607–612.

21. Fato R, Battino M, Esposti MD, Castelli GP, LenazG. 1986. Determination of partition and lateraldiffusion coefficients of ubiquinones by fluorescencequenching of n-(9-anthroyloxy)stearic acids inphospholipid vesicles and mitochondrial mem-branes. Biochemistry 25:3378–3390.

22. Stillwell W, Ehringer W, Wassall SR. 1992. Inter-action of a-tocopherol with fatty acids in mem-branes and ethanol. Biochim Biophys Acta 1105:237–244.

23. Dave JR, Witorsch RJ. 1983. Indomethacindecreases both luteinizing hormone binding andfluidity of testicular microsomal membranes in rat.Prostag Leukotr Med 12:371–380.

24. Ferreira H, Lucio M, de Castro B, Gameiro P, LimaJL, Reis S. 2003. Partition and location of nimesu-lide in EPC liposomes: A spectrophotometric andfluorescence study. Anal Bioanal Chem 377:293–298.

25. Coutinho A, Prieto M. 1993. Ribonuclease T1 andalcohol dehydrogenase fluorescence quenching byacrylamide: A laboratory experiment for under-graduate students. J Chem Educ 70:425–428.

26. Gran G. 1952. Determination of the equivalencepoint in potentiometric titrations. Part II. Analyst77:661–671.

27. Gans P, Sabatini A, Vacca A. 1999. Determinationof equilibrium constants from spectrophotometricdata obtained from solutions of known pH: Theprogram pHab. Ann Chim 89:45–49.

28. McClare CWF. 1971. An accurate and convenientorganic phosphorus assay. Anal Biochem 39:527–530.

29. New RRC. 1990. Liposomes—A practical approach.New York: Oxford University Press.

30. Matos C, Castro B, Gameiro P, Lima JLFC, Reis S.2004. Zeta-potential measurements as a tool toquantify the effect of charged drugs on the surfacepotential of egg phosphatidylcholine liposomes.Langmuir 20:369–377.

31. Takacs-Novak K, Avdeef A, Box KJ, Podanyi B,Szasz G. 1994. Determination of protonationmacro- and microconstants and octanol/water par-tition coefficient of the anti-inflammatory drugniflumic acid. J Pharm & Biomed Anal 11:1369–1377.

32. White SH, Jacobs RE, King GI. 1987. Partialspecific volumes of lipid and water in mixtures ofegg lecithin and water. J Biophys 52:663–665.

33. Castro B, Gameiro P, Lima JLFC, Matos C, Reis S.2001. A fast and reliable spectroscopic methodfor the determination of membrane-water parti-tion coefficients of organic compounds. Lipids 36:89–96.

34. Castro B, Gameiro P, Lima JL, Matos C, Reis S.2001. Location and partition coefficients of antiin-flammatory drugs in EPC liposomes. A fluorescencequenching study using n-(9-antroyloxy)-stearicprobes. Colloid Surface A 190:205–212.

35. McLaughlin S, Harary H. 1976. The hydrophobicadsorption of charged molecules to bilayer mem-branes: A test of the applicability of the Sternequation. Biochemistry 15:1941–1948.

36. Connors KA. 1987. Binding constants. The mea-surement of molecular complex stability. NewYork: John Wiley & Sons, pp. 59–61.

37. Rooney EK, East JM, Jones OT, McWhieter J,Simmonds AC, Lee AG. 1983. Interaction of fattyacids with lipid bllayers. Biochim Biophys Acta728:159–170.

38. Lacowicz JR. 1999. Principles of fluorescence spec-troscopy, 2nd ed. New York: Kluwer Academic/Plenum Press.

39. Thulborn KR, Sawyer WH. 1978. Properties andthe locations of a set of fluorescent probes sensitiveto the fluidity gradient of the lipid bilayer. BiochimBiophys Acta 511:125–140.

40. Vincent M, de Foresta B, Gallay J, Alfsen A. 1982.Nanosecond fluorescence anisotropy decays of n-(9-anthroyloxy) fatty acids in dipalmitoylphosphati-dylcholine vesicles with regard to isotropic sol-vents. Biochemistry 21:708–716.

41. Lucio M, Ferreira H, Lima JLFC, Matos C, CastroB, Reis S. 2004. Influence of some anti-inflam-matory drugs in membrane fluidity studied byfluorescence anisotropy measurements. Phys ChemChem Phys 6:1493–1498.



interaction of clonixin with epc liposomes used as membrane models

Documents