European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics

journal homepage: www.elsevier .com/locate /e jpb

Research paper

Effect of nanostructured lipid vehicles on percutaneous absorptionof curcumin

0939-6411/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.ejpb.2013.12.011

Abbreviations: MAD, monoolein aqueous dispersions; ORG, organogel; CUR,curcumin; x-gum, xanthan gum.⇑ Corresponding author. Dipartimento SVEB, Via Fossato di Mortara, 19, I-44121

Ferrara, Italy. Tel.: +39 0532 455259; fax: +39 0532 455953.E-mail address: ese@unife.it (E. Esposito).

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

Elisabetta Esposito a,⇑, Laura Ravani a, Paolo Mariani b, Nicolas Huang c,d, Paola Boldrini e,Markus Drechsler f, Giuseppe Valacchi a,g, Rita Cortesi a, Carmelo Puglia h

a Department of Life Sciences and Biotechnologies, University of Ferrara, Ferrara, Italyb Department of Life and Environmental Sciences and CNISM, Università Politecnica delle Marche, Ancona, Italyc Univ Paris-Sud, Faculté de Pharmacie, Châtenay-Malabry Cedex, Franced CNRS UMR 8612, Institut Galien Paris-Sud, Châtenay-Malabry Cedex, Francee Electron Microscopy Center, University of Ferrara, Ferrara, Italyf Macromolecular Chemistry II, University of Bayreuth, Germanyg Kyung Hee University, Dept. of Food and Nutrition, Seoul, South Koreah Department of Drug Science, University of Catania, Catania, Italy

a r t i c l e i n f o

Article history:Received 25 October 2013Accepted in revised form 15 December 2013Available online xxxx

Keywords:CubosomesOrganogelCurcuminReflectance spectroscopyTape-strippingControlled release

a b s t r a c t

The present study describes the production and characterization of monoolein aqueous dispersions(MAD) and lecithin organogels (ORG) as percutaneous delivery systems for curcumin (CUR).

In particular, MAD stabilized by sodium cholate/poloxamer and w0 3 ORG lipid carriers, both in thepresence and absence of CUR, have been considered: MAD morphology and dimensional distributionhave been investigated by Cryogenic Transmission Electron Microscopy (cryo-TEM) and Photon Correla-tion Spectroscopy (PCS), while the inner structure of MAD and ORG has been studied by X-ray scatteringtechniques. As a general result, CUR chemical stability has been found to be better controlled by MAD,probably because CUR is more protected in the case of CUR-MAD with respect to CUR-ORG.

To investigate the performance of differently composed lipid formulations as CUR delivery system,in vitro studies, based on Franz cell and stratum corneum–epidermis (SCE) membranes, and in vivo stud-ies, based on skin reflectance spectrophotometry and tape stripping, were then performed. The resultsindicated that ORG induces a rapid and intense initial penetration of CUR probably due to a strong inter-action between the peculiar supramolecular aggregation structure of phospholipids in the vehicle and thelipids present in the stratum corneum. Conversely, CUR incorporated into MAD can be released in a con-trolled fashion possibly because of the formation of a CUR depot in the stratum corneum. In this respectORG could be employed in pathologies requiring rapid CUR action, while MAD could be proposed forassuring a prolonged CUR activity.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

A wide variety of disorders and diseases affect the skin, e.g.acne, warts, multiple inflammatory dermatoses, skin cancers, auto-immune diseases, occupational dermatoses and contact dermatitis,requiring different investigations and therapies. Head and necksquamous cell carcinoma (HNSCC) is the sixth most common can-cer worldwide [1]. Treatment protocols include disfiguring surgery,platinum-based chemotherapy and radiation, all of which may

result in tremendous patient morbidity [2–4]. As a result, there issignificant interest in developing adjuvant chemotherapies toaugment currently available treatment protocols, which may allowdecreased side effects and toxicity without compromising thera-peutic efficacy. Curcumin (CUR) is one such potential candidate,in fact it has been recently demonstrated that CUR has a beneficialrole in skin diseases, in particular it can be employed as a singleagent in the treatment of HNSCC and used as an adjuvant agentin combination with standard platinum-based chemotherapy [5,6].

Despite the efficacy of CUR, its scarce water solubility may limitits administration. On the matter a number of research work hasbeen done in order to develop delivery strategies for CUR, e.g. lip-osomes, solid lipid nanoparticles, and cyclodextrins have beeninvestigated [7].

m. Bio-

2 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

Recently we have proposed the solubilization of CUR in monoo-lein aqueous dispersions (MAD) stabilized by different emulsifiers,resulting in the formation of complex lyotropic liquid crystallinephases, depending on the emulsifier used [8].

In order to treat cutaneous pathologies such as HNSCC, a semi-solid formulation constituted of biocompatible materials able toassure targeted delivery of actives, minimizing at the same timetoxic systemic effects appears the best strategy. A non-aqueousmicroemulsion system based on lecithin could be a good solutionat this aim [9,10].

It is well known that lecithin, thanks to its chemico-physicalproperties, is one of the most promising and useful agents ableto increase the skin permeation. In fact lecithin is a non-toxic, nat-urally occurring biocompatible surfactant able to form poly-molec-ular structures such as direct and inverted micelles or hexagonal,cubic and lamellar phases, offering the chance to produce innova-tive topical forms [10–13].

When a precise amount of water is added to an oil lecithinsolution, initially the lecithin molecules form spherical reversedmicelles, then the micellar aggregates entangle, forming athree-dimensional network in the bulk phase [14,15]. This pecu-liar w/o microemulsion is called lecithin organogel (ORG) since itconsists of a gel-like reverse micellar system in which theexternal phase is an organic solvent [14]. By adjusting the amountof added water it is possible to modulate the system viscosity[11,14,16].

ORG are particularly suitable as cutaneous delivery systems formany reasons. Indeed, they are thermodynamically stable, they areable to solubilize hydrophilic and lipophilic molecules and eventu-ally they can be produced with biocompatible solvents [17–21].

In the present study, we investigate the performances of differ-ent lipid based formulations for CUR cutaneous administration (i.e.MAD and ORG).

In particular, the first part of the present study describes theproduction and characterization of liquid MAD stabilized by polox-amer in mixture with sodium cholate and ORG. The MAD viscositywas then adjusted by the use of xanthan gum (x-gum).

The second part concerns an in vitro and in vivo investigationaimed to compare the release modalities of CUR after the adminis-tration of the semisolid lipid formulations on the skin.

CUR applied on the skin was studied determining its in vivotopical anti-inflammatory activity after cutaneous application ofMAD and ORG. The ultraviolet B (UVB)-induced erythema waschosen as inflammatory model on healthy human volunteers andwas monitored by reflectance visible spectrophotometry. More-over, tape-stripping experiments have been performed on skinafter topical administration of MAD and ORG to quantify CURpresence in the stratum corneum.

2. Experimental methods

2.1. Materials

Glyceryl monooleate RYLO MG 19 (monoolein) was a gift fromDanisco Cultor (Grindsted, Denmark). Pluronic F127 (Poloxamer407, poloxamer) (PEO98–POP67–PEO98) was obtained from BASF(Ludwigshafen, Germany). Curcumin (CUR), (1E,6E)-1,7-bis(4-Hy-droxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione, sodium cho-late (Na cholate) (3a,7a,12a-Trihydroxy-5b-cholan-24-oic acidsodium salt), Xanthan gum (x-gum) and isopropylpalmitate (IPP)were purchased from Sigma Chemical Company (St. Louis, MO,USA). The soybean lecithin (90% phosphatidyl choline) used forORG preparation was Epikuron 200 from Lucas Meyer, Hamburg,Germany. Solvents were of HPLC grade and all other chemicalswere of analytical grade.

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

2.2. MAD preparation

Production of dispersions was based on the emulsification ofmonoolein and emulsifier in water, as previously described [22].MAD composition was monolein 4.5% w/w, Na cholate 0.15% w/w, poloxamer 0.5% w/w and water.

After emulsification, the dispersions were subjected to homog-enization (15,000 rev min�1, Ultra Turrax, Janke & Kunkel, Ika-Werk, Sardo, Italy) at 60 �C for 1 min, then cooled and maintainedat room temperature in glass vials.

To produce CUR containing MAD (CUR-MAD), 7.5 mg of CUR(0.33% w/w with respect to the monoolein, 0.015% w/w withrespect to the dispersion) was added to the molten monoo-lein/emulsifier mixture and dissolved before addition to theaqueous solution. During production the vial was protectedfrom light with an aluminum foil to prevent photo-degradationof CUR.

2.3. Characterization of MAD

2.3.1. Cryo-Transmission Electron Microscopy (cryo-TEM)Samples were vitrified as described in a previous study [22].The vitrified specimen was transferred to a Zeiss EM922Omega

(Carl Zeiss Microscopy, Oberkochen, Germany) transmission elec-tron microscope using a cryoholder (CT3500, Gatan, Munich, Ger-many). Sample temperature was kept below 100 K throughoutthe examination. Specimens were examined with reduced dosesof about 1000–2000 e/nm2 at 200 kV. Images were recorded bya CCD digital camera (Ultrascan 1000, Gatan, Munich, Germany)and analyzed using a GMS 1.8 software (Gatan, Munich,Germany).

2.3.2. Photon Correlation Spectroscopy (PCS)Submicron particle size analysis was performed using a Zetasiz-

er 3000 PCS (Malvern Instr., Malvern, England) equipped with a5 mW helium neon laser with a wavelength output of 633 nm.Glassware was cleaned of dust by washing with detergent andrinsing twice with sterile water. Measurements were made at25 �C at an angle of 90� with a run time of at least 180 s. Sampleswere diluted with bidistilled water in a 1:10 v:v ratio. Data wereanalyzed using the ‘‘CONTIN’’ method [23]. Measurements wereperformed on MAD after production and after 6 months fromproduction.

2.3.3. CUR content of MADThe encapsulation efficiency (EE) of CUR in the MAD was deter-

mined as described by Nayak and colleagues [24]. 100 ll aliquot ofMAD was loaded in a centrifugal filter (Microcon centrifugal filterunit YM-10 membrane, NMWCO 10 kDa, Sigma Aldrich, St. Louis,MO, USA) and centrifuged (Spectrafuge™ 24D Digital Microcentri-fuge, Woodbridge NJ, USA) at 8000 rpm for 20 min. The amount offree and entrapped CUR was determined by dissolving the super-natant with a known amount of ethanol (1:10, v/v). The amountof CUR in the supernatant was determined by high performance li-quid chromatography (HPLC) method, as below reported. The EEwas determined as follows:

EE ¼ TCUR � SCUR=TCUR � 100 ð1Þ

where TCUR stands for the total amount of CUR added to the formu-lation and SCUR for the amount of drug measured in the supernatant.

2.4. ORG preparation

ORG were prepared by dissolving lecithin (200 mM) in IPP.From these reverse micellar solutions, microemulsion gels were

prepared by adding water, under magnetic stirring, to obtain the

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx 3

gel. The amount of added water has been expressed as w0 that isthe molar water to lecithin ratio (w0 = [H2O]/[lecithin]) [11].w0max was determined, i.e. the highest amount of water that canbe incorporated into the lecithin solution with no phaseseparation.

To produce drug containing ORG (CUR-ORG), after 3 h of stir-ring, CUR was added (0.015% w/v) to the obtained formulations.ORG were then maintained under stirring for 12 h.

Afterward, the samples were maintained at 25 �C for 30 minand later examined under a microscope equipped with a polariza-tion device.

Solubility of CUR in ORG was determined by saturating the ORGwith an excess of CUR (7 mg/ml). The obtained mixture was main-tained under stirring at room temperature protected from expo-sure to light for 72 h. CUR content in the produced formulationwas evaluated by extraction of un-dissolved drug with centrifuga-tion cycles. In particular 1 ml of CUR-ORG was subjected to two cy-cles of centrifugation for 15 min at 6000 rpm (Spectrafuge™ 24DDigital Microcentrifuge, Woodbridge NJ, USA). 100 ll of superna-tant was then dissolved with ethanol (1:10, v/v), afterwards theamount of solubilized CUR was determined by HPLC, as belowreported.

2.5. X-ray scattering measurements

Low-angle and wide-angle X-ray diffraction experiments wereperformed using a laboratory 3.5 kW Philips PW 1830 X-ray gen-erator equipped with a Guinier-type focusing camera operatingwith a bent quartz crystal monochromator (k = 1.54 Å). Diffrac-tion patterns were recorded on GNR Analytical Instruments Imag-ing Plate system. Samples, held in a tight vacuum cylindrical cellprovided with thin Mylar windows, were analyzed at two differ-ent temperatures, 37 and 45 �C. In each experiment, a number ofBragg peaks were eventually detected in the low-angle X-ray dif-fraction region, and the peak indexing was performed consideringthe different symmetries commonly observed in lipid phases [25].From the averaged spacing of the observed peaks, the unit celldimension, a, was calculated using the Bragg law. The nature ofthe short-range sample conformation was derived analyzing thehigh-angle X-ray diffraction profiles [26].

ORG were also investigated by Small-Angle X-ray Scattering(SAXS) experiments, which were performed at the SAXS beamlineat the Elettra Synchrotron (Trieste, Italy). The wavelength of theincident beam was k = 1.54 Šand the explored Q-range extendedfrom 0.1 to 0.9 �1 (Q is the modulus of the scattering vector, de-fined as 4p sinh/k, where 2h is the scattering angle). At the SAXSbeamline, organogels were measured using 1 mm thick quartz cap-illaries at the 2 different temperatures of 37 and 45 �C. Particularattention was paid to checking for equilibrium conditions andmonitoring radiation damage. In a few tests, measurements wererepeated several times (up to 10) at the same temperature to ac-count for a constant scattering signal. Experimental intensitieswere corrected for background, solvent contributions, detectorinhomogeneities, and sample transmission, as usual [27]. Unfortu-nately, no absolute scale calibration of the experimental data wasavailable.

2.6. Prediction of long-term stability

The stability of CUR in MAD and in ORG was assessed in the ab-sence of light. CUR-MAD and CUR-ORG were stored in well-closedamber glass containers and kept at room temperature for6 months.

Chemical stability was evaluated on drug loaded formulations,determining CUR content by HPLC analyses. Shelf life values werecalculated [28].

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

Log (CUR residual content, %) was plotted against time and theslopes (m) were calculated by linear regression.

The slopes (m) were then substituted into the following equa-tion for the determination of k values:

k ¼ m� 2:303 ð2Þ

Shelf life values (the time for 10% loss, t90) and half-life (thetime for 50% loss, t1/2) were then calculated by the followingequations:

t90 ¼ 0:105=k ð3Þ

t1=2 ¼ 0:693=k ð4Þ

as reported by Wells [29].

2.7. Gel production

The MAD have been viscosized by adding x-gum (1% w/w) di-rectly into the dispersion and by slowly stirring for 1 h until com-plete dispersion of the gum. The obtained viscous MAD was namedCUR-MAD x-gum.

2.8. Rheological measurements

Rheological measurements were performed with an AR-G2rotational rheometer (TA Instruments). The geometry used wasan aluminum cone-plate with a diameter of 40 mm and an angleof 1�. Flow curve was obtained in increasing the shear rate from0.01 s�1 to 5000 s�1 with 5 points per decade, and each pointwas maintained for a duration of 180 s in order to perform mea-surements in the permanent regime. The temperature was set at25 �C or 35 �C and controlled with a Peltier plate. A solvent trapwas used to prevent water evaporation. Measurements were per-formed in triplicate for each sample, to ensure reproducibility.

2.9. In vitro skin permeation experiments

Samples of adult human skin (mean age 36 ± 8 years) were ob-tained from breast reduction operations. Subcutaneous fat wascarefully trimmed and the skin was immersed in distilled waterat 60 ± 1 �C for 2 min, after which SCE (stratum corneum/epider-mis) was removed from the dermis using a dull scalpel blade.SCE membranes were dried in a desiccator at �25% relative humid-ity. The dried samples were wrapped in aluminum foil and storedat 4 ± 1 �C until use.

SCE were then rehydrated by immersion in distilled water atroom temperature for 1 h before being mounted in Franz-type dif-fusion cells supplied by LGA (Berkeley, CA) as reported by Pugliaand colleagues [8].

The exposed skin surface area was 0.78 cm2 area (1 cm diame-ter orifice). The receptor compartment contained 5 ml of a mixtureof phosphate buffer 60 mM pH 7.4 (PBS) and ethanol (50:50, v/v) toallow the establishment of the sink conditions and to sustain per-meant solubilization [30]. This solution was stirred with the help ofa magnetic bar at 500 rpm and thermostated at 32 ± 1 �C during allthe experiments [31].

Approximately 500 ll of each formulation (CUR-MAD, CURMAD x-gum and CUR-ORG) was placed on the skin surface in thedonor compartment and the latter was sealed to avoid evaporation.At predetermined time intervals comprised between 1 and 24 h,samples (0.15 ml) of receptor phase solution were withdrawnand the CUR concentration in the receptor phase was measuredusing HPLC. Each removed sample was replaced with an equal vol-ume of simple receptor phase. The CUR concentrations were deter-mined six times in independent experiments and the meanvalues ± standard deviations were calculated. The mean values

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

4 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

were then plotted as a function of time. The diffusion coefficientswere computed from the linear portion of the accumulation curveand expressed both as experimentally observed fluxes (Fo) and asnormalized fluxes Fn

Fn ¼ Fo=C ð5Þ

where C is the CUR concentration in the analyzed form, expressed inmg/ml.

After Franz cell experiments, skin samples were treated withreducing agents and fixed with glutaraldehyde 2.5% (w/v) in phos-phate buffer 0.1 M pH 7.4, post-osmicated with osmium tetroxide2% in the same buffer, dehydrated with increasing quantities ofacetone and included in Araldite Durcupan ACM (Fluka). Semifinesections were made on ultramicrotome Reichert Ultracut S, stainedwith an aqueous solution of Toluidin blue 1% (w/v) and studied bylight microscope Nikon Eclipse E800.

2.10. In vivo studies

2.10.1. Volunteers recruitmentIn vivo experiments were performed on two groups of ten

volunteers: group A enrolled for the first in vivo experimentation(evaluation of anti-inflammatory activity) and group B enrolledfor the second one (tape-stripping).

Experiments were conducted in accordance with The Code ofEthics of the World Medical Association (Declaration of Helsinki)for experiments involving humans.

The volunteers were of both sexes in the age range 25–55 yearsand they were recruited after medical screening including the fill-ing of a health questionnaire followed by physical examination ofthe application sites. After they were fully informed on the natureof the study and on the procedures involved, they gave their writ-ten consent. The participants did not suffer from any ailment andwere not on any medication at the time of the study. They wererested for 15 min prior to the experiments and room conditionswere set at 22 ± 2 �C and 40–50% relative humidity.

2.10.2. In vivo anti-inflammatory activityUVB-induced skin erythema was monitored by using a reflec-

tance visible spectrophotometer X-Rite model 968 (XRite Inc.,Grandville, MI, USA), calibrated and controlled as previously re-ported [32,33]. Reflectance spectra were obtained over the wave-length range 400–700 nm using illuminant C and 2� standardobserver. From the spectral data obtained, the erythema index(EI) was calculated using Eq. (6) [34]:

EI ¼ 100 log1

R560þ 1:5 log

1R540

þ log1

R580

� �� 2 log

1R510

þ log1

R610

� �� �

ð6Þ

where 1/R is the inverse reflectance at a specific wavelength (560,540, 580, 510 and 610). The skin erythema was induced by UVBirradiation using a UVM-57 ultraviolet lamp (UVP, San Gabriel,CA, USA) whose specific parameters are reported elsewhere [33].The MED was preliminarily determined, and an irradiation dosecorresponding to twice the value of MED was used throughoutthe study. For each subject (group A), seven sites on the ventral sur-face of each forearm were defined using a circular template (1 cm2)and demarcated with permanent ink. One of the seven sites of eachforearm was used as control, three sites were treated with 300 mgof CUR-MAD and the remaining three with 300 mg of CUR-ORG. Thepreparations were spread uniformly by means of a solid glass rodand then the sites were occluded for 6 h using Hill Top Chambers(Hill Top Research, Cincinnati, OH). After the occlusion period, thechambers were removed and the skin surfaces were gently washedto remove the gel and allowed to dry for 15 min. Each pretreated

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

site was exposed to UV-B irradiation 1, 3 and 6 h (t = 1, t = 3 andt = 6, respectively) after CUR-MAD and CUR-ORG removal and theinduced erythema was monitored for 52 h. EI baseline values weretaken at each designated site before application of gel formulationand they were subtracted from the EI values obtained after UV-Birradiation at each time point to obtain DEI values. For each site,the AUC was computed using the trapezoidal rule. To better outlinethe results obtained, the PIE was calculated from the AUC valuesusing following equation:

Inhibition ð%Þ ¼ AUCðCÞ � AUCðTÞAUCðCÞ

� 100 ð7Þ

where AUC(C) is the area under the response/time curve of the vehi-cle-treated site (control) and AUC(T) is the area under the response/time curve of the drug-treated site. Statistical differences of in vivodata were determined using repeated measure analysis of variance(ANOVA) followed by the Bonferroni-Dunn post hoc pair-wise com-parison procedure. A probability, P, of less than 0.05 was consideredsignificant in this study.

2.10.3. Tape strippingThe first steps of the experimental protocol previously de-

scribed were also employed in this second in vivo study; in fact,for each subject of group B, seven sites (2 cm2) were defined onthe ventral surface of each forearm, and 200 mg of each formula-tions (CUR-MAD and CUR-ORG) was applied on these cutaneoussites (three sites of application for each formulation in duplicate).The preparations were spread uniformly on the site by means ofa solid glass rod and were then occluded for 6 h. After the occlusionperiod, the residual formulations were removed by gently wipingwith cotton balls (different for each pretreated site). Ten individual2 cm2 squares of adhesive tape (Scotch Book Tape 845, 3M) wereutilized to sequentially tape-strip the stratum corneum on theapplication sites. To obtain a realistic comparison between theresults of this experimentation and the ones obtained in theprevious in vivo study, the removal of stratum corneum in eachpretreated site was effected at 1 h (t = 1), 3 h (t = 3) and 6 h(t = 6) after gel removal [32].

Each adhesive square, before and after skin tape stripping, wasweighed on a Sartorius balance (model ME415S, sensitivity 1 mg)to quantify the weight of stratum corneum removed. After eachstripping, the tapes were put in the same vial containing 2 ml ofthe HPLC mobile phase (methanol, 2% acetic acid and acetonitrile,5:30:65 v/v) and subjected to vortical stirring over 30 s. Theextracted CUR was then quantified by HPLC. The recovery of CURwas validated by spiking tape-stripped samples of untreated stra-tum corneum with 2 ml of a mobile phase containing CUR 10 mg/ml. The extraction efficiency of CUR was 96.8 ± 0.9% (n = 3).

2.11. Statistical analysis

Statistical differences of in vivo data were determined usingrepeated-measures analysis of variance (ANOVA) followed by theBonferroni–Dunn post hoc pairwise comparison procedure.The employed software was Prism 5.0, Graph Pad Software Inc.(La Jolla, CA, USA). A probability of less than 0.05 is consideredsignificant in this study.

2.12. HPLC procedure

HPLC determinations were performed using a two-plungersalternative pump (Jasco, Japan), an UV-detector operating at425 nm, and a 7125 Rheodyne injection valve with a 50 ll loop.Samples were loaded on a stainless steel C-18 reverse-phase

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx 5

column (15 � 0.46 cm) packed with 5 lm particles (Grace� – Allti-ma, Alltech, USA).

Elution was performed with a mobile phase containing metha-nol, 2% acetic acid and acetonitrile 5:30:65 v/v at a flow rate of0.5 ml/min. Retention time of CUR was 7.0 min.

Fig. 1. Cryo-transmission electron microscopy images (cryo-TEM) of MAD in theabsence (panel A) and in the presence (panel B) of CUR.

3. Results

3.1. Preparation of MAD

It is well known that monoolein is able to form cubosome dis-persions when it is emulsified in water in the presence of a surfac-tant [35–37]. In particular poloxamer 407 is extensively employed[38–40].

Recently we found that stable monoolein dispersions can be ob-tained by the use of Na cholate in mixture of poloxamer as emul-sifiers [41].

In this study MAD and CUR-MAD were produced by the mixtureof Na cholate and poloxamer, resulting in translucent dispersions,yellowish in the case of CUR-MAD.

Cryo-electron transmission microscope analyses were con-ducted in order to investigate the internal structures of MAD andto compare the influence of the emulsifier on the nanostructureof the dispersed phase. Fig. 1 reports cryo-TEM images of MAD (pa-nel A) and CUR-MAD (panel B). One can observe the presence ofunilamellar vesicles in both panels. The black points are ascribedto ice crystals contamination due to sample preparation. The pres-ence of CUR does not seem to affect MAD aspect.

3.1.1. MAD dimensional analysesPCS studies were conducted to determine the dimensional dis-

tribution of MAD dispersions, in the absence and in the presence ofCUR. Table 1 summarizes the obtained results. One can observedthat the mean diameter of the MAD disperse phase after produc-tion is around 150 nm. The presence of CUR leads to a slight de-crease in mean diameters and polydispersity index.

After 6 months from production, mean diameters are almostunvaried in the absence of MAD (2% increase with respect to theinitial mean diameters), while they undergo a 20% increase in thepresence of CUR.

3.2. Preparation of ORG

Solutions of 50–250 mM phosphatidylcholine in organic solventcan be transformed into transparent gels by addition of critical andwell defined amounts of water [11]. The addition of water leads toa strong change in viscosity until a transparent gel is obtained [13].It should be stressed that the transparency of the system has con-siderable advantages with respect to topical and transdermalapplications [21]. In lecithin gels in fact, complete solubilizationof solid compounds can be easily evaluated [14,15].

It is well known that the viscosity of the gel is a function of theadded water, expressed as w0 [11]. ORG and CUR-ORG with w0

values 1, 2 and 3 were studied, the determined w0max was 3. Forin vitro and in vivo studies the w0 3 CUR-ORG was chosen, thus pro-viding suitable viscosity for topical application. The producedorganogels were transparent, yellow, macroscopically monophasicand isotropic under polarized light. In the case of CUR-ORG theyellow color was more intense.

3.3. X-ray diffraction analyses

X-ray diffraction was used to investigate the inner structuralorganization of MAD and ORG. Experiments were performed as afunction of temperature, both in the presence and in the absence

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

of CUR. A first example of diffraction results is shown in Fig. 2. Con-sidering the expected Bragg peak sequence, the X-ray diffractionprofiles indicate the presence of a Pn3m inverted bicontinuous cu-bic phase both in the absence and in the presence of CUR [42].However, two results should be underlined: first, as no changesare detected in diffraction profiles, the inner structure of MADand CUR-MAD samples is the same, but a larger unit cell dimensionis observed in the presence of CUR: at 37 �C, the unit cell parameterchanges from 95.5 to 102.4 Å. Second, no phase transitions are ob-served on heating. In both cases, heating only determines a smallreduction of the unit cell size (when temperature increases from37 to 45 �C, the unit cell reduces from 95.5 to 83.8 Å and from102.4 to 98.7 Å for MAD and CUR-MAD samples, respectively), asexpected considering the temperature-induced decrease in thehydrocarbon chain order parameter. Finally, no changes were de-tected in wide-angle (WAXS) profiles (data not shown), indicatingthat the fluid nature of the short-range lipid conformation is notmodified by the presence of CUR.

In the case of ORG samples, synchrotron SAXS experimentswere performed in addition (Fig. 3A), as no low-angle scatteringwas observed using a standard X-ray diffractometer. While in theabsence of CUR a typical SAXS profile for a disordered micellarsystem is detected, in the CUR-ORG sample two very low and large

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

Table 1Dimensional parameters of MAD and CUR-MAD after production (0 d) and after 6 months (180 d) from production.

Formulation Mean diameter (nm) ± SD Polydispersity index

Average Analysis by intensity Analysis by volume Analysis by number

0 (d) 180 (d) 0 (d) 180 (d) 0 (d) 180 (d) 0 (d) 180 (d) 0 (d) 180 (d)

MAD 158.9 ± 2.0 161.9 ± 4.2 164.3 ± 5.1 166.7 ± 5.3 137.3 ± 3.3 141.2 ± 6.3 98.2 ± 4.1 99.1 ± 3.8 0.43 ± 0.07 0.30 ± 0.05CUR-MAD 147.6 ± 3.1 178.7 ± 2.8 154.5 ± 6.2 157.0 ± 3.2 131.5 ± 2.9 132.4 ± 6.3 95.7 ± 3.3 96.5 ± 2.5 0.34 ± 0.01 0.15 ± 0.04

Data are the mean of 5 independent determinations of different MAD batches.

Fig. 2. X-ray scattering results from MAD samples, both in the presence andabsence of CUR. Scattering profiles have been measured at 37 �C (continuous line)and 45 �C (dotted lines) and curves have been displaced along the intensity axis forclarity. The vertical lines indicate the expected positions of the Bragg peaks in thecase of a cubic Pn3m structure [42].

6 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

Bragg peaks were observed. As the position of the two peaks scalesas 1:

p3, a local, short-range 2D hexagonal order between the lec-

ithin cylindrical reverse micelles can be suggested [14]. Noticeableis the fact that the unit cell parameter of the Hexagonal (H) phase,a, (which corresponds to the lateral distance between two rod-shaped micelles) is practically independent on temperature andequal to about 170 Å. Wide-angle diffraction profiles (WAXS)shown in Fig. 3B evidence a broad band at Q = 1.35 Å�1, which cor-responds to a repeat distance of 4.65 Å and which suggests a fluidconformation of the hydrocarbon chains.

In order to derive structural information, the SAXS curves havebeen analyzed by a very simple and phenomenological model al-ready used to describe the clustering effects in macromolecularsystems [43–46] (Supplementary Data, SD). From such a model,the clustering strength and the correlation length for the micellarcylinders, e.g. the so-called ‘‘blob’’ size, can be derived.

3.4. Efficiency of CUR encapsulation

In the case of CUR-ORG, the preparation method prevents druglosses, allowing to obtain 100% recovery of drug into the formula-tion. Conversely, in the case of CUR-MAD the production is based

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

on the use of a stirrer and a homogenizer, so the drug can be lostpartly on the paddle employed for MAD production.

For these reason, to determine CUR recovery in MAD, the Ultra-centrifugation method (UC) was employed [8,24].

Because the method allows to separate CUR associated with thedisperse phase from CUR free in the dispersion, it enables to obtainboth recovery and encapsulation data. The latter are given indi-rectly by difference and refer to free CUR found in the supernatantafter UC of CUR-MAD. CUR EE was almost quantitative (99.19%with respect to the CUR amount used for the preparation). CURrecovery data determined by MAD dilution in ethanol and stirringwere superposable (data not shown). The data indicate that thewhole CUR employed for MAD production was encapsulated intothe MAD disperse phase and that no drug has been lost duringproduction.

It should be emphasized the solubility power of ORG with re-spect to CUR. It was found that ORG is able to solubilize 5.5 mg/ml of CUR. This value is 5.5-fold higher than the CUR solubility inethanol (1 mg/ml). In the case of MAD, the solubility power forCUR was 350 lg/ml that is 15-fold lower than its solubility inORG, but dramatically higher than its solubility in water (11 ng/ml).

3.5. CUR stability in MAD and in ORG

CUR content in the different formulations was calculated as afunction of time and expressed as percentage of the total amountused for the preparation.

Fig. 4 reports the first order kinetics obtained plotting Log (CURresidual content, % with respect to drug content at time 0) againsttime. From the slopes (m), determined by linear regression, shelflife (t90) and half life (t1/2) values were calculated and reported inTable 2.

All data were statistically significant (p < 0.0001).It was found that CUR-MAD was decidedly more efficacious

than CUR-ORG in controlling CUR stability. In fact MAD couldmaintain 90% of CUR stability for almost 23 years, while in the caseof ORG, t90 is around one month. The t1/2 values reach 154 years forCUR-MAD and 7 months for CUR-ORG.

Anyway the efficacy in controlling CUR stability is startling forboth formulations, in fact it should be underlined that CUR in phos-phate buffer 0.1 M is rapidly decomposed (t1/2 9.4 min) [47].

The macroscopic aspect of both CUR-ORG and CUR-MAD did notchange by time, in fact they did not show phase separation phe-nomena, maintaining the almost absence of aggregates also aftersix months from production.

3.6. Production of viscous MAD

In order to obtain vehicles suitable for possible administrationon the skin and to compare ORG with MAD for in vitro andin vivo studies, the low viscosity of the MAD was adequately im-proved by the use of xanthan gum x-gum. The plain x-gum (1%w/w) in water has a viscosity of 0.4 Pa s.

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

A

B

Fig. 3. X-ray scattering results from ORG samples, both in the presence and absenceof CUR. SAXS profiles (panel A) have been measured at 37 �C (continuous line) and45 �C (dotted lines), while WAXS results (panel B) refer to 37 �C. Curves have beendisplaced along the intensity axis for clarity. Note the different Q-ranges. Thevertical lines indicate the expected positions of the Bragg peaks in the case of a 2D Hphase.

1.8

1.84

1.88

1.92

1.96

2

0 20 40 60 80 100

Log

(% C

UR

resi

dual

)

time (days)

Fig. 4. Variation of CUR residual content in CUR-MAD (s) and CUR-ORG (h) as afunction of time. Data are the means of 5 analyses on different batches of the sametype of dispersions.

Table 2Curcumin recovery in MAD and ORG by time.

Formulation Shelf life values

K t90 (d)a t1/2 (d)b

CUR-MAD 0.000012 254,010 56,341CUR-ORG 0.00325 968.4 213.23

The reported results represent the average of four independent experiment, SDwere within 5%.

a Time at which the drug concentration has lost 10% with respect to drugrecovery at 0 day.

b Time at which the drug concentration has lost 50% with respect to drugrecovery at 0 day.

E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx 7

The viscosity value of MAD x-gum and CUR-MAD x-gum mea-sured at 25 �C were 1.290 and 1.649 Pa s respectively, at a shearrate value of 10 s�1.

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

The ORG and CUR-ORG viscosity at a shear rate value of 10 s�1

were 8.082 and 8.216 Pa s respectively. Thus the presence of CURinduces a light increase in MAD x-gum and ORG viscosity.

3.7. Rheological analyses

Fig. 5 shows flow curves for CUR-MAD x-gum (panels A and B)and CUR-ORG (panels C and D), at 25 �C and 35 �C. In particular theviscosity vs. shear rate (panels A and C) and vs. shear stress (panelsB and D) has been plotted.

The points on the curves are the means of three experiments,and error bars represent standard deviation.

In Fig. 5A, CUR-MAD x-gum shows a marked non-Newtonianshear thinning behavior: the steady shear viscosity sharply de-creased as an increase in shear rate while the first and second pla-teaux (Newtonian viscosity regions) are not observed at low andhigh shear rates respectively [48]. The temperature does not influ-ence the behavior, in fact the curves at 25 �C or 35 �C aresuperposable.

A similar trend is reported in the flow curve of Fig. 5B, referringto CUR-MAD x-gum vs. shear stress.

In Fig. 5C, CUR-ORG, at 25 �C or 35 �C, exhibits a Newtonian pla-teau at low shear rate, and is shear thinning at higher shear rate[49]. More precisely, at 25 �C, the Newtonian plateau extends from0.01 s�1 to 1 s�1, and the viscosity decreases significantly above10 s�1. For this temperature, the viscosity curve as a function ofstress (Fig. 5D) also shows a sharp decrease in the viscosity above

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

A B

C D

Fig. 5. Flow curves for CUR-MAD x-gum (panels A and B) and CUR-ORG (panels C and D) determined at 25 and 35 �C. Viscosity vs. shear rate (panels A and C) and vs. shearstress (panels B and D) has been plotted.

1

y = 0.020028 + 0.12301x R= 0.99417

y = 0.0034273 + 0.013116x R= 0.99135

y = 0.011447 + 0.019436x R= 0.99148

8 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

a stress of about 100 Pa (yield stress). At 35 �C, the same shearthinning behavior is observed one decade higher in shear rate(but no yield stress is observed): the Newtonian plateau extendsfrom 0.01 s�1 to 10 s�1, and the viscosity decreases significantlyabove 100 s�1. For shear rates corresponding to the Newtonian pla-teau, viscosity at 35 �C (� 0.8 Pa s) is about one decade lower thatviscosity at 25 �C (� 11 Pa s). At high shear rate (>400 s�1), viscos-ities at 25 �C and 35 �C are identical: the shear is high enough tobreak significantly the structure of the gel (shear thinning behav-ior) and to conceal the effect of temperature. It should be notedthat the second Newtonian plateau, normally observed at very highshear rate, was not reached at 10,000 s�1. However, for the CUR-ORG (Fig. 5C) at 25 �C, the slight decrease in the viscosity slopeat 10,000 s�1 may be the sign of the very beginning of this plateau.

Flow curves of MAD x-gum and ORG are not shown since theyare superposable to the reported curves, thus viscosity is not af-fected by the presence of CUR.

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10

CU

R (u

g/cm

2 )

time (hours)

Fig. 6. In vitro permeation profiles of CUR from CUR-MAD (s), CUR MAD x-gum (d)and CUR-ORG (j). Data represent the mean of six independent experiments ± SD.

3.8. In vitro CUR diffusion kinetics

The in vitro CUR diffusion was studied by Franz cell in order toinvestigate the efficiency of CUR-MAD and CUR-ORG designed astopical vehicles.

In particular CUR diffusion from liquid CUR-MAD, viscous MAD(CUR-MAD x-gum) and CUR-ORG were compared.

A non-physiological receptor phase with 50% v/v of ethanol wasused due to CUR scarce solubility in aqueous media, in fact it wasfound previously that the use of physiological media as receptorphases led to negligible diffusion kinetics [31].

In Fig. 6 the cumulative plots of the amount of CUR permeatedthrough SCE membranes as a function of time are reported. Fromthe obtained equation, the CUR steady state flux values (Fn) forthe different vehicles were calculated.

The Fn of CUR formulated in ORG resulted 0.82 lg/cm2, whilethe Fn was 0.129 lg/cm2 and 0.087 lg/cm2 for CUR formulated inCUR-MAD and in CUR-MAD x-gum respectively.

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

Statistical analysis revealed significant differences (p < 0.01) be-tween the Fn values obtained for all formulations.

An important question preliminary to pharmaceutical applica-tions is whether, and to what extent, CUR-MAD and CUR-ORGare harmful to human skin [21]. To try to clarify this point, we car-ried out a light microscopic investigation of human skin afterin vitro diffusion experiments. In particular, samples of skin treatedfor 8 h with either PBS (taken as control), CUR-MAD or CUR-ORGwere analyzed (Fig. 7). No significant alterations of the skin wereapparent for all the treated samples. In particular, the stratumcorneum was still intact after treatment. There was no difference

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx 9

between the control samples treated with PBS (Fig. 7A) and skinsamples treated with CUR-MAD and CUR-ORG.

3.9. In vivo anti-inflammatory activity

CUR is able to inhibit skin erythema by cutaneous applicationdue to its anti-inflammatory activity [50,51].

Formulation CUR-ORG and CUR-MAD x-gum were further stud-ied in vivo to determine their ability to inhibit the UVB-inducedskin erythema on healthy human volunteers. Skin reflectance spec-trophotometry was used to determine the extent of the erythemaand to assess the inhibition capacity of the formulations after theirpreventive application onto the skin [32]. The AUC was determinedfor each subject plotting DEI values vs. time. Table 3 reports theobtained AUC0–52 values. An inverse relationship was found be-tween the AUC and the inhibition of UVB-induced erythema.Fig. 8A reports the PIE values.

Fig. 7. Light micrographs of SCE membranes treated for 8 h with PBS (panel A)taken as control or with CUR-MAD (panel B) and CUR-ORG (panel C). (Forinterpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

CUR-ORG showed to be more effective than CUR-MAD x-gum ininhibiting the induced erythema 1 h after its removal (P < 0.05),while at 6 h, the formulation CUR-MAD x-gum showed the bestinhibitory ability (P < 0.05; Table 3, Fig. 8A). No difference betweenthe formulations was found, in terms of anti-inflammatory activity,after 3 h after their removal (p > 0.05).

3.10. Tape-stripping technique

The technique was useful for quantifying drug depletion in theviable epidermis and the amount of CUR responsible for the anti-inflammatory effect [32]. From the analysis of the data, a decreasein the amount of CUR in the stratum corneum was detected forboth CUR-MAD x-gum and CUR-ORG (Fig. 8B). Nonetheless theamount of CUR recovered in the stratum corneum after removalof CUR-ORG was significantly different with respect to the onefrom the CUR-MAD x-gum.

In fact in the case of CUR-ORG the amount decreased from37 ng/cm2 after 1 h to 20 ng/cm2 after 6 h, while in the case ofCUR-MAD x-gum the CUR amount passed from 192 ng/cm2 after1 h to 69 ng/cm2 after 6 h (Fig. 8B).

4. Discussion

Lipid based formulations have a potential as matrixes able todissolve and deliver active molecules in a controlled fashion, there-by improving their bioavailability and reducing side-effects[37,52,53]. Their topical application allows the treatment of cuta-neous pathologies improving local delivery of the incorporateddrugs or modulating drug diffusion, as a function of the lipidcomponents.

In the present study we have produced lipid based topical formswith the aim to find new vehicles for cutaneous administration ofthe lipophilic CUR molecule. For this purpose we focused our ef-forts toward two different colloidal systems: the first based on ao/w monoolein nanodispersion (MAD) mainly constituted of vesi-cles and cubosomes and the second based on a w/o lecithin micro-emulsion (ORG) constituted of an entanglement of elongatedreverse micelles.

With regard to MAD, from cryo-TEM images it can be assertedthat the use of Na cholate as emulsifier in mixture with poloxamer,does not allow the formation of cubic or sponge-like structures butonly unilamellar vesicles (Fig. 1). Indeed by the presence of Pn3minverted bicontinuous cubic phases was detected both in the ab-sence and in the presence of CUR (Fig. 2). Cryo-TEM and X-ray dataapparently disagree but the latter are in accordance with theobservations by Lindstrom and colleagues which found thatmonoolein in dilute micellar bile salt solutions forms vesiclesand different liquid crystalline phases such as the cubic one [54].In this regard it should be underlined that in few cases inner struc-tured cubosomes or hexasomes could not be trapped during freez-ing processes, mostly it should be observed that only a very smallvolume of the sample is examined by TEM. By contrast, a very largesample volume is analyzed by X-ray diffraction, and even smallamounts of ordered phases can be easily detected. Therefore, thewhole results for plain CUR-MAD suggest the coexistence ofvesicles and inner structured particles.

Concerning X-ray diffraction results, noticeable is the relativelylarge unit cell increase induced by CUR (Fig. 2): this fact clearlysuggests that CUR, according to its amphiphilic nature, promotesan increase in the negative curvature of the polar/apolar interfacewith a subsequent increased hydration.

Rheology studies indicated for MAD x-gum, in the presence andin the absence of CUR, a steady shear flow behavior typical ofxanthan gum dispersions [48]. The large molecules of xanthan

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

Table 3AUC0–52 values obtained pre-treating skin sites with formulation CUR-ORG and CUR-MAD x-gum and applying UVB radiations after 1 h (t = 1), 3 h (t = 3) or 6 h (t = 6) from theirremoval.

Subjects AUC0–52

t = 1 t = 3 t = 6 Control

CUR-ORG CUR-MAD x-gum CUR-ORG CUR-MAD x-gum CUR-ORG CUR-MAD x-gum

A 590.2 1321.4 921.8 980.2 1164.3 831.2 1625.9B 621.4 1286.3 1000.2 975.1 1120.2 823.6 1568.3C 670.5 1279.1 896.4 880.2 1100.3 770.2 1489.4D 700.2 1325.2 912.6 926.4 1200.8 790.4 1410.3E 632.5 1295.5 1058.3 958.1 1190.6 754.3 1184.9F 690.4 1288.6 954.2 944.3 1160.4 780.2 1320.1

Mean 650.9 1299.4 957.3 944.1 1156.1 791.7 1433.2

0

0.05

0.1

0.15

0.2

0.25

CUR-ORG CUR-MAD x-gum

Cur

cum

in u

g/cm

2

A

B

0

10

20

30

40

50

60

CUR-ORG CUR-MAD x-gum

P.I.E

.

Fig. 8. In vivo comparative study of CUR anti-inflammatory effect (panel A) and CURamount in the stratum corneum (panel B) after topical application and removal offormulation CUR-ORG and CUR-MAD x-gum. PIE: percentage of erythema index;t = 1, 3, and 6: hours from the removal of formulations. Data represent the mean forten subjects.

10 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

gum form aggregates through hydrogen bonding and polymerentanglement, resulting in a high shear viscosity at low shear ratesor at rest. As shear rate increases, the steady shear viscosity de-creases due to the disentanglement of the polymer network andto the alignment of individual macromolecules in the direction ofthe shear flow, this results in a low shear viscosity at high shearrate region. The network structure reforms rapidly upon removing

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

of shear, hence the viscosity is recovered almost instantaneously[55]. The presence of monooleine–poloxamer–sodium cholatebased nanosystems does not seem to affect the rheology behaviorof MAD x-gum.

The lecithin organogel is characterized by a typical supramolec-ular organization described by other authors as a three-dimen-sional network of entangled interpenetrating polymer-like orworm-like micelles [14,56]. In the micellar evolution there arefew stages upon successive addition of water, resulting in abranching of aggregates from the initial linear growth [13]. Thenthe micelles disintegrate into a mixture consisting of shorter andlonger ones that persist in constituting a three-dimensional net-work stabilized by hydrogen bonds between lecithin and polar sol-vent molecules. The ORG were prepared using of a lecithin solutionin IPP, a biocompatible oil which has been demonstrated to be par-ticularly suitable for transdermal delivery [52,57]. Lecithin concen-tration was fixed at 200 mM, since this concentration was found tobe effective for penetration enhancement [12].

In the case of ORG, the SAXS profile suggests that the presenceof CUR induces the formation of a short-range, local 2D hexagonalorganization of the lecithin cylindrical reverse micelles, while thebroad band in WAXS profile confirms the fluid conformation ofthe hydrocarbon chains [26], independently on the presence ofCUR. It is also interesting to observe that the results obtained fit-ting the SAXS curves by equation (S) in the SD confirms the in-creased ordering of the organogel after the addition of CUR (e.g.,the entanglement length, which is a measure of the distance alongthe micelle separating neighboring entanglements, increases fromabout 420 to 620 Å). Therefore, the presence of CUR induces an in-creased ordering of the system, possibly due to an increase in theintermolecular hydrogen bonding in the CUR-ORG [58,59].

From a rheological point of view, CUR-ORG exhibited the typicalbehavior of lecithin organogels both in the absence and in the pres-ence of CUR. As found by other authors, a Newtonian flow wasfound for over the low shear rate range, as indicated by a near con-stant viscosity over shear rate [49,60]. This suggests that the net-work structure formed by the worm-like micelles does not breakin this range. At higher shear rates, CUR-ORG displayed non-New-tonian flow, giving a decrease in viscosity with increasing shearrate. This phenomenon may be a result of the collapse of the net-work structure due to disentanglement and/or breakage of theworm-like micelles as the shear rate increases [49,60].

It was found that CUR solubility in ORG is 5.5-fold higher thanthe CUR solubility in ethanol (1 mg/ml). Within the organogel, CURis probably localized partially between the acyl chains of the leci-thin molecules by hydrophobic interactions [61], as previouslyfound for other lipophilic molecules [21,62] and partially dissolvedin the external oil phase.

Conversely in the MAD, that is constituted of an external aque-ous phase and an inner nanostructured lipophilic phase, CUR

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx 11

would be encapsulated into the inner phase of the dispersion [8],its solubility in this case is only 350 lg/ml. The difference in CURlocation within the two formulations should be responsible ofthe different CUR stability. It can be hypothesized that CUR is moreexposed to degradation in the case of CUR-ORG and more pro-tected in the case of CUR-MAD.

In vitro CUR diffusion studies demonstrated that CUR Fn valuesfrom CUR-ORG were 6-fold higher with respect to CUR-MAD x-gum and 9-fold higher than CUR-MAD. The differences in Fn ofCUR-MAD and CUR-MAD x-gum may be ascribed to the vehicle vis-cosity, indeed the presence of x-gum represents an obstacle to CURdiffusion through the vehicle.

Conversely the difference in CUR fluxes from CUR-ORG andCUR-MAD has to be related to the technological aspect of the vehi-cle. In fact, as above reported, in CUR-ORG the molecule is partiallysolubilized in the external phase constituted of IPP, so that it ispromptly available for diffusion through the SCE membrane [21].Whereas in CUR-MAD the drug is associated with the nanostruc-ture present in the disperse phase while the external phase is con-stituted of water. Thus the lower CUR fluxes from CUR-MAD andCUR-MAD x-gum could be justified by the presence of differentstructural systems (i.e. cubosomes and vesicles) which could slowCUR diffusion.

In vivo experiments gave precious information about CUR cuta-neous biodistribution after topical application with MAD or ORG.

The low anti-inflammatory effect exerted by CUR-MAD x-gumafter 1 h from the removal of the formulation followed by a pro-gressive increase after 3 and 6 h, could be tentatively justified bythe formation of a CUR depot in the stratum corneum induced byCUR-MAD presence in this layer, as indicated by experimental evi-dences reported by other authors [32,52]. Since Algrem and col-leagues have found a similarity between the cubic phasestructure and the structure of the stratum corneum [63], it is rea-sonable to suppose the formation of a mix of cubosomal monooleinwith stratum corneum lipids. According to this hypothesis, fromthe stratum corneum depot a sustained release of CUR towardthe deeper skin layers could take place. Thus the stratum corneumlipids might play a role in modulating CUR pharmacodynamic re-sponse [32].

On the other hand, the high initial anti-inflammatory effect in-duced by CUR-ORG followed by a rapid decrease could suggest arapid and intense initial penetration of CUR. This phenomenoncould be justified by a strong interaction with the stratum corneumlipids and results in a high concentration of CUR in the vascularizedsection of the skin, from which CUR is rapidly removed by theblood stream [52]. This strong interaction could be due to the pe-culiar supramolecular aggregation structure of phospholipids inthe CUR-ORG promoting a CUR penetration enhancer effect [52].

Tape stripping experiments helped to better elucidate the‘‘in vivo’’ behavior of CUR applied on the skin [64]. In the case ofCUR-MAD x-gum, 1 h after the occlusion, the CUR amount in thestratum corneum was 5-fold higher than that one found for CUR-ORG. Afterward CUR contents decreased time dependently in bothforms with similar trend and CUR-ORG constantly lower in CURconcentration. This profile justifies the above reported hypothesisof the formation of a cubosome depot in the stratum corneum fromwhich CUR can be released in a controlled fashion [32]. Therefore,the decrease in CUR has to be ascribed to a drug slow release fromthis reservoir. The same amounts were probably responsible forthe CUR-MAD x-gum anti-inflammatory profile recorded.

The lower CUR amount found in the stratum corneum in thecase of CUR-ORG could be related to a different kind of interactionbetween the formulation and the skin.

Considering both lecithin and CUR-ORG abilities to promoteCUR absorption through the skin [65–67], it appears reasonableto suppose that already after 1 h of occlusion, CUR has penetrated

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

through the upper compartments of the epidermis and then canreach the dermal zone, where it displays an anti-inflammatoryactivity.

Afterward in the time points corresponding to 3 and 6 h fromthe occlusion period the amounts of CUR undergo a furtherdecrease.

5. Conclusions

This study has highlighted the CUR performances in two differ-ent nanostructured lipid vehicles. Both are able to solubilize thislipophilic molecule and to protect it from degradation.

In vivo data have demonstrated that CUR-ORG could be a prom-ising strategy to treat skin diseases (e.g. scleroderma, psoriasis andskin cancer) because the vehicle promotes CUR absorption throughskin, while CUR-MAD x-gum could be employed to treat the dis-ease by time in a controlled fashion, since CUR delivery can bemodulated and its action prolonged. To verify this hypothesis ani-mal studies should be performed.

Acknowledgements

The authors are grateful to Sarah Mosbah from Institut GalienParis-Sud, Châtenay-Malabry, France for rheological characteriza-tion and Clelia Miracco from Department of Medicine and Surgery,University of Siena, Italy for her expert advice about microscopydata of skin samples.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ejpb.2013.12.011.

References

[1] D.L. Miller, M.A. Weinstock, Nonmelanoma skin cancer in the United States:incidence, J. Am. Acad. Dermatol. 30 (1994) 774–778.

[2] F. Farshadpour, H. Kranenborg, E.V. Calkoen, Survival analysis of head and necksquamous cell carcinoma: influence of smoking and drinking, Head Neck 33(2011) 817–823.

[3] S. Marur, A.A. Forastiere, Head and neck cancer: changing epidemiology,diagnosis, and treatment, Mayo Clin. Proc. 83 (2008) 489–501.

[4] V. Bouvard, R. Baan, K. Straif, Y. Grosse, B. Secretan, F. El Ghissassi, A review ofhuman carcinogens – Part B: Biological agents, Lancet Oncol. 10 (2009) 321–322.

[5] M.M. LoTempio, M.S. Veena, H.L. Steele, B. Ramamurthy, T.S. Ramalingam, A.N.Cohen, R. Chakrabarti, E.S. Srivatsan, M.B. Wang, Curcumin suppresses growthof head and neck squamous cell carcinoma, Clin. Cancer Res. 11 (2005) 6994–7002.

[6] R. Wilken, M.S. Veena, M.B. Wang, E.S. Srivatsan, Curcumin: a review of anti-cancer properties and therapeutic activity in head and neck squamous cellcarcinoma, Mol. Cancer 10 (2011) 1–19.

[7] R.K. Bhawana, H.S. Basniwal, V.K. Buttar, N. Jain, Curcumin nanoparticles:preparation, characterization, and antimicrobial study, J. Agric. Food Chem. 59(2011) 2056–2061.

[8] C. Puglia, V. Cardile, A.M. Panico, L. Crascì, A. Offerta, S. Caggia, M. Drechsler, P.Mariani, R. Cortesi, E. Esposito, Evaluation of monooleine aqueous dispersionsas tools for topical administration of curcumin: characterization, in vitro andex-vivo studies, J. Pharm. Sci. 102 (2013) 2349–2361.

[9] E. Esposito, E. Menegatti, R. Cortesi, Design and characterization of fenretinidecontaining organogels, Mater. Sci. Eng. C 33 (2013) 383–389.

[10] F. Dreher, P. Walde, P.L. Luisi, P. Elsner, Human skin irritation studies of alecithin microemulsion gel and of lecithin liposomes, Skin Pharmacol. 9 (1996)124–129.

[11] R. Scartazzini, P.L. Luisi, Organogels from lecithins, J. Phys. Chem. 92 (1988)829–833.

[12] M.B. Fawzi, U.R. Iyer, M. Mahjour, Use of Commercial Lecithin as SkinPenetration Enhancer, US Patent 4, 783, 450, 1988.

[13] P.L. Luisi, R. Scartazzini, G. Haering, P. Schurtenberger, Organogels from water-in-oil microemulsions, Colloid Polym. Sci. 268 (1990) 356–374.

[14] Y.A. Shchipunov, A micellar system with unique properties, Colloids Surf. A185 (2001) 541–554.

[15] P. Schurtenberger, R. Scartazzini, L.J. Magid, M.E. Leser, P.L. Luisi, Structuraland dynamic properties of polymer like reverse micelles, J. Phys. Chem. 94(1990) 3695–3701.

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-

12 E. Esposito et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx

[16] Y.A. Shchipunov, E.V. Shumilina, Lecithin-bridging by hydrogen bonds in theorganogel, Mater. Sci. Eng. C 3 (1995) 43–50.

[17] K.D. Patil, S.R. Bakliwal, S.P. Pawar, Organogel: topical and transdermal drugdelivery system, Int. J. Pharm. Res. Dev. 3 (2011) 58–66.

[18] S. Gupta, R.P. Singh, A. Sarkar, H. Panchal, D. Pandey, Organogel: a viablealternative for existing carrier system, Int. J. Compr. Pharm. 2 (2011) 1–5.

[19] H.L. Willimann, P.L. Luisi, Lecithin organogels as matrix for the transdermaltransport of drugs, Biochem. Biophys. Res. Commun. 177 (1991) 897–900.

[20] P. Walde, H. Willimann, C. Nastruzzi, R. Scartazzini, P. Schurtenberger, P.L.Luisi, Lecithin microemulsion gels as matrix for the transdermal delivery ofdrugs, in: Proceed. Intern. Symp. Control. Rel. Bioact. Mater., vol. 17, 1990, pp.421–422.

[21] H. Willimann, P. Walde, P.L. Lusi, F. Stroppolo, Lecithin organogel as matrix fortransdermal transport of drugs, J. Pharm. Sci. 81 (1992) 871.

[22] E. Esposito, P. Mariani, L. Ravani, C. Contado, M. Volta, S. Bido, M. Drechsler, S.Mazzoni, E. Menegatti, M. Morari, R. Cortesi, Nanoparticulate lipid dispersionsfor bromocriptine delivery: characterization and in vivo study, Eur. J. Pharm.Biopharm. 80 (2012) 306–314.

[23] R. Pecora, Dynamic light scattering measurement of nanometer particles inliquids, J. Nanoparticle Res. 2 (2000) 123–131.

[24] A.P. Nayak, W. Tiyaboonchai, S. Patankar, B. Madhusudhan, E.B. Souto,Curcuminoids-loaded lipid nanoparticles: novel approach towards malariatreatment, Colloids Surf. B 81 (2010) 263–273.

[25] M. Pisani, S. Bernstorff, C. Ferrero, P. Mariani, Pressure induced cubic-to-cubicphase transition in monoolein hydrated system, J. Phys. Chem. B 80 (2001)3109–3119.

[26] G. Zolese, M. Wozniak, P. Mariani, L. Saturni, E. Bertoli, A. Ambrosini, Differentmodulation of phospholipase A2 activity by saturated and monounsaturatedN-acylethanolamines, J. Lipid Res. 44 (2003) 742–753.

[27] P. Andreozzi, S.S. Funari, C. La Mesa, P. Mariani, M.G. Ortore, R. Sinibaldi, F.Spinozzi, Multi- to unilamellar transitions in catanionic vesicles, J. Phys. Chem.B 114 (2010) 8056–8060.

[28] E. Esposito, L. Ravani, C. Contado, A. Costenaro, M. Drechsler, D. Rossi, E.Menegatti, A. Grandini, R. Cortesi, Clotrimazole nanoparticle gel for mucosaladministration, Mater. Sci. Eng. C 33 (2013) 411–418.

[29] J.I. Wells, Pharmaceutical Preformulation: The Physicochemical Properties ofDrug Substances Ellis Hortwood, Chichester, England, 1988.

[30] E. Toitou, B. Fabin, Altered skin permeation of highly lipophilic molecule:Tetrahydrocannabinol, Int. J. Pharm. 43 (1988) 17–22.

[31] M. Siewert, J. Dressman, C.K. Brown, V.P. Shah, FIP/AAPS guidelines todissolution/in vitro release testing of novel/special dosage forms, AAPSPharmSciTech 4 (2003) E7.

[32] E. Esposito, R. Cortesi, M. Drechsler, L. Paccamiccio, P. Mariani, C. Contado, E.Stellin, E. Menegatti, F. Bonina, C. Puglia, Cubosome dispersions as deliverysystems for percutaneous administration of indomethacin, Pharm. Res. 22(2005) 2163–2173.

[33] M. Ricci, C. Puglia, F. Bonina, C. Di Giovanni, S. Giovagnoli, C. Rossi, Evaluationof indomethacin percutaneous absorption from nanostructured lipid carriers(NLC): in vitro and in vivo studies, J. Pharm. Sci. 94 (2005) 1149–1159.

[34] J.B. Dawson, D.J. Barker, D.J. Ellis, E. Grassam, J.A. Catterill, G.W. Fischer, J.W.Feather, A theoretical and experimental study of light absorption andscattering by in vivo skin, Phys. Med. Biol. 25 (1980) 696–709.

[35] K. Larsson, Aqueous dispersion of cubic lipid/water phases, Curr. Opin. ColloidInterf. Sci. 5 (2000) 64–69.

[36] G. Garg, S. Saraf, S. Saraf, Cubosomes: an overview, Biol. Pharm. Bull. 30 (2007)350–353.

[37] B. Siekmann, H. Bunjes, M.H.J. Koch, K. Westesen, Preparation and structuralinvestigations of colloidal dispersions prepared from cubic monoglyceride/water phases, Int. J. Pharm. 244 (2002) 33–43.

[38] G. Worle, M. Drechsler, M.H.J. Koch, B. Siekmann, K. Westesen, H. Bunjes,Influence of composition and preparation parameters on the properties ofaqueous monoolein dispersions, Int. J. Pharm. 329 (2007) 150–157.

[39] J. Gustafsson, H. Ljusberg-Wharen, M. Almgrem, K. Larsson, Cubic lipid/waterphase dispersed into submicron particles, Langmuir 12 (1996) 4611–4613.

[40] E. Esposito, L. Ravani, P. Mariani, C. Contado, M. Drechsler, C. Puglia, R. Cortesi,Curcumin containing monoolein aqueous dispersions: a preformulative study,Mater. Sci. Eng. C 33 (2013) 4923–4934.

[41] J. Gustafsson, H. Ljusberg-Wharen, M. Almgrem, K. Larsson, Submicronparticles of reversed lipid phases in water stabilized by a nonionicamphiphilic polymer, Langmuir 13 (1997) 6964–6971.

[42] P. Mariani, V. Luzzati, H. Delacroix, Cubic phases of lipid containingsystems. Structure analysis and biological implications, J. Mol. Biol. 204(1988) 165–189.

Please cite this article in press as: E. Esposito et al., Effect of nanostructured lpharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.12.011

[43] B. Hammouda, D.L. Ho, S. Kline, Insight into clustering in poly(ethylene oxide)solutions, Macromolecules 37 (2004) 6932–6937.

[44] J.E. Curtis, A. McAuley, H. Nanda, S. Krueger, Protein structure and interactionsin the solid state studied by small-angle neutron scattering, Faraday Discuss.158 (2012) 285–299.

[45] R.G. Larson, The lengths of thread-like micelles inferred from rheology, J.Rheol. 56 (2012) 1363–1374.

[46] R. Granek, M.E. Cates, Stress relaxation in living polymers: results from aPoisson renewal model, J. Chem. Phys. 96 (1992) 4758–4767.

[47] Y.J. Wang, M.-H. Pan, A.-L. Cheng, L.-I. Lin, Y.-S. Ho, C.-Y. Hsieh, J.-K. Lin,Stability of curcumin in buffer solutions and characterization of itsdegradation products, J. Pharm. Biomed. Anal. 15 (1997) 1867–1876.

[48] K.-W. Song, Y.-S. Kim, G.-S. Chang, Rheology of concentrated xanthan gumsolutions: Steady shear flow behaviour, Fiber Polym. 7 (2006) 129–138.

[49] Y.A. Schipunov, H. Hoffmann, Thinning and thickening effects induced byshearing in lecithin solutions of polymer-like micelles, Rheol. Acta 39 (2000)542–553.

[50] B.B. Aggarwal, K.B. Harikumar, Potential therapeutic effects of curcumin, theanti-inflammatory agent, against neurodegenerative, cardiovascular,pulmonary, metabolic, autoimmune and neoplastic diseases, Int. J. Biochem.Cell Biol. 41 (2009) 40–59.

[51] J.S. Jurenka, Anti-inflammatory properties of curcumin, a major constituent ofCurcuma longa: a review of preclinical and clinical research, Altern. Med. Rev.14 (2009) 141–153.

[52] F.P. Bonina, L. Montenegro, N. Scrofani, E. Esposito, R. Cortesi, E. Menegatti, C.Nastruzzi, Effects of phospholipid based formulations on in vitro and in vivopercutaneous absorption of methyl nicotinate, J. Control. Release 34 (1995)53–63.

[53] A. Yaghmur, O. Glatter, Characterization and potential applications ofnanostructured aqueous dispersions, Adv. Colloid Interf. Sci. 147 (2009)333–342.

[54] M. Lindstrom, H. Ljusberg-Wahren, K. Larsson, B. Borgstrom, Aqueous lipidphases of relevance to intestinal fat digestion and absorption, Lipids 16 (1981)749–754.

[55] M. Marcotte, R. Taherian Hoshahili, H.S. Ramaswamy, Rheological properties ofselected hydrocol as a function of concentration and temperature, Food Res.Int. 34 (2001) 695–703.

[56] A. Vintiloiu, J.-C. Leroux, Organogels and their use in drug delivery: a review, J.Control. Release 125 (2008) 179–192.

[57] M. Changez, J. Chander, A.K. Dinda, Transdermal permeation of tetracainehydrochloride by lecithin microemulsion: in vivo, Colloids Surf. B: Biointerf. 48(2006) 58–66.

[58] Y. Jeong, A. Friggeri, I. Akiba, H. Masunaga, K. Sakurai, S. Sakurai, S. Okamoto, K.Inoue, S. Shinkai, Small-angle X-ray scattering from a dual-componentorganogel to exhibit a charge transfer interaction, J. Colloid Interf. Sci. 283(2005) 113–122.

[59] D. Satapathy, D. Biswas, B. Behera, S.S. Sagiri, K. Pal, K. Pramanik, Sunflower-oil-based lecithin organogels as matrices for controlled drug delivery, J. Appl.Polym. Sci. 129 (2013) 585–594.

[60] K. Hashizaki, N. Tamaki, H. Taguki, Y. Saito, K. Tsuchiya, H. Sakai, M. Abe,Rheological behavior of worm-like micelles in a mixed nonionic surfactantsystem of a polyoxyethylene phytosterol and a glycerin fatty acid monoester,Chem. Pharm. Bull. 56 (2008) 1682–1686.

[61] G. Began, E. Sudharshan, K. Udaya Sankar, A.G. Appu Rao, Interaction ofcurcumin with phosphatidylcholine: a spectrofluorimetric study, Agric. FoodChem. 47 (1999) 4992–4997.

[62] N. Duffy, H.C.G. Blonk, C.M. Beindorff, M. Cazade, A. Bot, G.S.M.J.E. Duchateau,Organogel-based emulsion systems, micro-structural features and impact onin vitro digestion, J. Am. Oil Chem. Soc. 86 (2009) 733–774.

[63] M. Almgrem, K. Edwards, G. Karlsson, Cryo transmission electron microscopyof liposomes and related structures, Colloids Surf. A 174 (2000) 3–21.

[64] J. Lademann, U. Jacobi, C. Surber, H.-J. Weigmann, J.W. Fluhr, The tape strippingprocedure-evaluation of some critical parameters, Eur. J. Pharm. Biopharm. 72(2009) 317–332.

[65] R. Kumar, O.P. Katare, Lecithin organogels as a potential phospholipid-structured system for topical drug delivery: a review, AAPS PharmSciTech 6(2005) E298–E310.

[66] F. Dreher, P. Walde, P. Walther, E. Wehrli, Interaction of a lecithinmicroemulsion gel with human stratum corneum and its effect ontransdermal transport, J. Control. Release 45 (1997) 131–140.

[67] M. Changez, J. Chander, A.K. Dinda, Transdermal permeation of tetracainehydrochloride by lecithin microemulsion: in vivo, Colloids Surf. B: Biointerf. 48(2006) 58–66.

ipid vehicles on percutaneous absorption of curcumin, Eur. J. Pharm. Bio-