International Journal of Pharmaceutics 215 (2001) 229240

The effects of terpene enhancers on the percutaneouspermeation of drugs with different lipophilicities

Ayman F. El-Kattan a,b, Charles S. Asbill a,c, Nanhye Kim a,d,Bozena B. Michniak a,e,*

a College of Pharmacy, Uni6ersity of South Carolina, Columbia, SC 29208, USAb Pfizer, Ann Arbor, MI, USA

c McWhorter School of Pharmacy, Samford Uni6ersity, Birmingham, AL, USAd Watson Pharmaceuticals, Salt Lake City, UT, 84108, USA

e New Jersey Center for Biomaterials, Laboratory for Drug Deli6ery Center, 111 Lock Street, Newark, NJ 07103, USA

Received 26 July 2000; received in revised form 7 December 2000; accepted 11 December 2000

Abstract

Four model drugs were selected based on their lipophilicity denoted as log P (nicardipine hydrochloride log P0.9990.1, hydrocortisone log P 1.4390.47, carbamazepine log P 2.6790.38, and tamoxifen log P 7.8790.75).The enhancing activities of four terpene enhancers (fenchone log P 2.1390.30, thymol log P 3.2890.20, D-limonenelog P 4.5890.23, and nerolidol log P 5.3690.38) were tested in vitro across full thickness hairless mouse skin witheach of the evaluated drugs formulated in hydroxypropyl cellulose gel formulations. The relationships betweenlipophilicity (log P) of the terpene enhancers and model drugs and efficacy (represented by the enhancement ratio offlux ERflux) of the drugs when coadministered with the enhancers were examined using linear regression. Terpeneenhancers had significant effect on the percutaneous permeation of the model drugs. Nerolidol (highest lipophilicity)provided the highest increase in the flux of the evaluated model drugs. The flux of nicardipine hydrochloride increasedby approximately 135-fold, hydrocortisone by 33-fold, carbamazepine 8-fold, and tamoxifen 2-fold. The lowestincrease in the flux was observed with fenchone. Linear relationships were generated between the ERflux of nicardipinehydrochloride, hydrocortisone, carbamazepine, and tamoxifen and the log P of the terpene enhancers [r0.951,(P0.049), r0.977, (P0.023), r0.942, (P0.057), and r0.874, (P0.126), respectively]. Furthermore, thefour terpene enhancers produced linear relationships, indicating that they were more effective at enhancing thepenetration of hydrophilic drugs rather than lipophilic drugs r 0.824 (P0.176) for fenchone, r 0.891(P0.109) for thymol, r 0.846 (P0.154) for limonene, and r 0.769 (P0.232) for nerolidol. 2001Published by Elsevier Science B.V.

www.elsevier.com:locate:ijpharm

1. Introduction

Skin, the largest human body organ, envelopsand protects the body from the external environ-

* Corresponding author. Tel.: 1-973-9729720; fax: 1-973-3535336.

E-mail address: bmichniak@earthlink.net (B.B. Michniak).

0378-5173:01:$ - see front matter 2001 Published by Elsevier Science B.V.

PII: S0 378 -5173 (00 )00699 -2

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240230

ment while helping maintain the integrity andappropriate function of the complex inner bodyorgans. The utilization of the transdermal drugdelivery is associated with various advantages thathave been extensively discussed in the literature.This route allows for controlled release of thedrug at rates approaching zero-order simulatingthose provided by IV infusion. It also delivers themedication in a continuous input, which is impor-tant for drugs that have short biological half-livesand low therapeutic indices. Once the drug isabsorbed through the skin, the first-pass intestinaland hepatic metabolism is bypassed. As a conse-quence of all these advantages, patient complianceis improved.

Despite the obvious advantages of transdermaldrug delivery, this route presents unique chal-lenges. The greatest obstacle is the stratumcorneum, the outermost layer of the skin, which isconsidered to form the primary rate-limiting bar-rier to the permeation of the drugs across the skin(Zettersten et al., 1997). It consists of dead, flat-tened cells, filled with keratin, that are embeddedin a lipid matrix (Bouwstra et al., 1999; Coderchet al., 1999). Lipids in the intercellular spaces aremade of fatty acids, ceramides, and cholesterol(esters) and are arranged in bilayer structures,which run mainly parallel to the skin surface(Elias, 1996a,b). The stratum corneum has beendescribed as hydrophilic protein bricks in a hy-drophobic mortar (Heisig et al., 1996). The searchfor solutions to this problem led investigators toemploy several enhancement techniques such asiontophoresis, electroporation (Banga et al.,1999), or the application of drug eutectic mixtures(Kaplun-Frischoff and Touitou, 1997; Stott et al.,1998). However, the most widely implementedtechnique is the use of chemical penetration en-hancers, which reversibly perturb the permeabilitybarrier of the stratum corneum. Examples includemany compounds such as Azone and its ana-logues (Michniak, 1993; Michniak et al., 1993,1994a,b), fatty acids (Aungst, 1989), alcohols(Takahashi et al., 1991), pyrrolidones (Southwelland Barry, 1983), sulfoxides (Choi et al., 1991),and terpenes (Williams and Barry, 1990; Cornwelland Barry, 1991; El-Kattan et al., 2000a,b). How-ever, the skin has a special role as a major barrier

in protecting a living body from cutaneous expo-sure to toxic chemicals, hence the safety of perme-ation enhancers is of primary consideration.Terpenes are of low cutaneous irritancy, possessgood toxicological profile, provide excellent en-hancement ability, and appear to be promisingcandidates for pharmaceutical formulations (Gaoand Singh, 1998). A variety of terpenes have beenshown to increase the percutaneous absorption ofboth hydrophilic (Zhao and Singh, 1999) andlipophilic drugs (Gao and Singh, 1998). Differen-tial scanning calorimetery (DSC) studies showedthat cyclic monoterpenes decreased the transitiontemperature associated with the stratum corneumlipids (Yamane et al., 1995). The DSC resultssupport the speculation that terpene enhancersmainly increase drug diffusivity in the skin bydisrupting the highly ordered intercellular lipidstructure of the stratum corneum. Furthermore,this interaction with the stratum corneum lipids isreversible.

In this study, hydroxypropyl cellulose (HPC),which is a hydrophobic nonionic polymer, wasincorporated in the gel formulations. HPC hasseveral useful properties that render it widely usedin topical and transdermal formulations. It is usedas an emulsifier, suspending, and stabilizing agentfor topical preparations (Wu et al., 1998). Fur-thermore, it can inhibit the formation of sedi-ments (Wade and Weller, 1995).

The purpose of the present study was to investi-gate if a correlation exists between the efficacy ofthe terpene enhancers with different lipophilicity(denoted as log P) (fenchone log P 2.1390.30,thymol log P 3.2890.20, D-limonene log P4.5890.23, and nerolidol log P 5.3690.38) andthe log P of the model drugs (nicardipine hydro-chloride log P 0.9990.1, hydrocortisone log P1.4390.47, carbamazepine log P 2.6790.38, andtamoxifen log P 7.8790.75).

2. Methods

2.1. Materials

Fenchone, thymol, D-limonene, and nerolidolwere all purchased from Aldrich Chemical Co.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240 231

(Milwaukee, WI). Nicardipine hydrochloride, hy-drocortisone, carbamazepine, tamoxifen, potas-sium phosphate monobasic, and triethanolaminewere supplied by Sigma Chemical Co. (St. Louis,MO). Acetonitrile, methanol, and water used wereof HPLC grade and supplied by EM Science(New Briggs, NJ). Ethyl alcohol (USP) was ob-tained from Aaper Alcohol and Chemical Co.(Shelbyville, KY). HPC and glycerol were pro-vided by Spectrum Chemical ManufacturingCorp. (Gardena, CA). Female hairless mice strainSKH1 (h:h), 68 weeks old were obtained fromCharles River Lab., Inc., (Wilmington, MA). Allreagents were of analytical grades and used with-out further purification.

2.2. Preparation of gel formulations

HPC gels were prepared as described recentlyby El-Kattan et al. (El-Kattan et al., 2000a,b).The HPC gel composition is shown in Table 1.Briefly, a Sted Fast stirrer was used to mix etha-nol with glycerol and distilled water. Then HPCpowder was added to the obtained solution withcontinuous mixing until the gel was formed. Thetested drug followed by a terpene enhancer wereadded to the gel and mixed. The gels were leftovernight at ambient temperature. It is worthnoting that tamoxifen, carbamazepine, and hydro-cortisone were not completely soluble in the pre-pared gel formulations. Whereas, nicardipinehydrochloride was completely soluble in the gelformulations. Furthermore, all the evaluated ter-pene enhancers, but limonene were completelysoluble in the gel formulations.

2.3. In 6itro skin permeability studies

Franz diffusion cells (effective diffusion area0.64 cm2, receptor volume 5.1 ml) were used toevaluate the in vitro drug percutaneous perme-ation from HPC gel. The receptor compartmentwas filled with isotonic phosphate buffer (pH 7.2)and 0.1% formaldehyde (37%) as a preservative.The temperature was maintained at 3790.5Cand the receptor was constantly stirred at 600rpm. Excised female full-thickness hairless mouseskin (strain SKH1 h:h, 8 weeks old, Charles RiverLaboratories, Wilmington, MA) was mounted be-tween the donor and receptor compartments. Ap-proximately 300 mg of HPC gel was placed oneach skin. All cell donors were occluded with aParafilm. A 500-ml sample of the receptor wastaken at predetermined time intervals over 24 hand replenished immediately with equal volume ofthe diffusion buffer. Samples were frozen at 70C prior to HPLC analysis. After 24 h ofsampling, the skins were removed from the cellsand washed briefly in methanol (25 ml) for 15 s(Michniak et al., 1994a,b). Following drying atroom temperature for 10 min, each skin was cutup and then homogenized in 4-ml methanol. Thesamples were then centrifuged and the superna-tant layer was taken and stored at 70C beforeHPLC analysis.

Control gel formulations for each model drugwere developed. Control formulations consistedof all the formulation components except theterpene enhancer. All experiments were n5.

2.4. HPLC analysis of model drugs

All drugs were analyzed using a reverse phaseC18 column (MICROSORB-MV, 15 cm, 5 mm)and at a flow rate 1 ml:min. HPLC analysis wasperformed using Hewlett Packard 1100 with anautosampler equipped with a quaternary pump, avariable-wavelength detector. Nicardipine hydro-chloride was detected at 240 nm with a mobilephase composition of 60:40 acetonitrile:0.02 Mpotassium phosphate monobasic and injectionvolume of 20 ml (Kobayashi et al., 1993). Hydro-cortisone was detected at 242 nm with a mobilephase composition of 40:60 acetonitrile:water and

Table 1Composition (% w:w) of drug HPC gel formulations a

% w:wComponent

2Drug*2Terpene2HPC

Water 2810Glycerol

Ethanol 56

a * nicardipine hydrochloride, hydrocortisone, carba-mazepine, and tamoxifen.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240232

Table 2Effect of terpene enhancers on the percutaneous parameters of nicardipine hydrochloride formulated in HPC gels a

ERflux Q24 (mg:cm2)Terpene SC (mg:g)Flux (mg:cm2 per h)

1.0Control 45952.591.0 43059400Fenchone 44.895.0 17.9 793919 13 35991023Thymol 45.596.2 18.2 811962 98039845

60.0 33049350150.0920.0 21 48394532Limonene337.2931.2Nerolidol 134.8 61059590 18 04693425

a Mean9S.D., n5. ERflux, Enhancement ratio of nicardipine HCl flux; Q24, cumulative amount of nicardipine HCl in thereceptor after 24 h; SC, skin content of nicardipine HCl after 24 h.

injection volume of 40 ml (Michniak, 1993). Car-bamazepine was detected at 210 nm with a mobilephase composition of acetonitrile:water (30:70)and injection volume of 20 ml (Psillakis et al.,1999). Tamoxifen was detected at 275 nm with amobile phase composition of wa-ter:triethanolamine:methanol and injection vol-ume of 100 ml (70:1.8:928) (Herrlinger et al.,1992).

2.5. Data analysis

The in vitro skin permeation data obtained wasgraphically plotted as the cumulative correctedamount of drug penetrated into the receptorphase as a function of time. The permeationprofiles provided the following parameters: Theslope of the straight line portions of this plot (atsteady state) yielded the values of flux (mg:cm2 perh) and the cumulative corrected receptor concen-trations at 24 h Q24 (mg:cm2). Skin content valueswere expressed as mg of drug per gram of hy-drated full-thickness mouse skin.

Log P values of terpenes enhancers and drugswere determined using ACD software program(Advanced Chemistry Incorporated, Ontario,Canada).

Enhancement ratios for flux (ERflux) were calcu-lated using the following equation

ERflux

Model drug flux with terpene in gelModel drug flux without terpene in gel (control)

(1)

Controls were assigned the value of 1.00 andthe data was presented as mean9S.D. of five

experiments. Statistical analyses were performedusing one-way analysis of variance (one wayANOVA). Correlation analyses were performedby the least squares linear regression method.Correlation coefficients were examined for signifi-cance (PB0.05) by Students t test.

3. Results

3.1. Percutaneous absorption of nicardipinehydrochloride in 6itro

The effects of terpene enhancers on the percuta-neous permeation parameters of nicardipine hy-drochloride (flux, ERflux, cumulative amount ofnicardipine hydrochloride after 24 h (Q24), andskin content) from HPC gel formulations areshown in Table 2. Controls consisted of gel for-mulations with no terpene enhancers. The valuesfor the percutaneous permeation parameters ofnicardipine hydrochloride control gels are as fol-lows 2.591.0 mg:cm2 per h for flux, 4595 mg:cm2 for Q24, and 43059400 mg:g for nicardipinehydrochloride skin content. All the evaluated ter-pene enhancers had significant effects on the per-cutaneous permeation of nicardipinehydrochloride relative to the control. Nerolidolprovided the best enhancement activity fornicardipine hydrochloride. It increased nicardip-ine hydrochloride flux 134.8-fold relative to thecontrol (P0.0000042) followed by limonenewith 60.0-fold (P0.00021), thymol 18.2 fold(P0.000244), and fenchone 17.9-fold (P0.00013). The highest increase in the Q24 was alsoprovided by nerolidol (61059590 mg:cm2) fol-

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240 233

lowed by limonene (33049350 mg:cm2), thymol(811962 mg:cm2), and fenchone (793919 mg:cm2). It should be emphasized that the increase inthe Q24 provided by nerolidol, limonene, thymol,and fenchone was significantly higher than thatrecorded with the control (P4106, 4106, 0.00011, and 0.000063, respectively). Thehighest nicardipine hydrochloride skin contentwas observed with limonene (21 48394532 mg:g)followed by nerolidol (18 04693425 mg:g), fen-chone (13 35991023 mg:g), and thymol (98039845 mg:g), which were significantly higher thanthat provided by the control (P0.0027, 0.0025,0.0014, and 0.00090, respectively).

3.2. Percutaneous absorption of hydrocortisone in6itro

The effects of terpene enhancers on the percuta-neous permeation parameters of hydrocortisone(flux, ERflux, cumulative amount of hydrocorti-sone after 24 h (Q24), and skin content) from HPCgel formulations are presented in Table 3. Controlvalues for hydrocortisone were determined to be6.091.5 mg:cm2 per h for flux, 14596 mg:cm2 forQ24, and 80529290 mg:g for hydrocortisone skincontent. All tested terpene enhancers had signifi-cant effects on both hydrocortisone flux and Q24relative to the control. The highest increase in theflux was observed with nerolidol, which increasedthe flux by 32.7-fold relative to the control (P0.0011) followed by limonene with 28.0-fold (P0.00071), thymol 10.5-fold (P0.00088), andfenchone with 7.8 fold (P0.000169). Similar tothe ERflux results, nerolidol provided the highestincrease in the Q24 (18379106 mg:cm2) that was

significantly higher than that observed with thecontrol (P8106) followed by limonene(15699232 mg:cm2) (P0.00095) and thymol(11589168 mg:cm2) (P0.00059). The lowest in-crease in the Q24 was recorded with fenchone(10179148 mg:cm2) yet it was significantly higherthan that observed with the control (P0.00094).Terpene enhancers did not increase the hydrocor-tisone skin content values significantly relative tothe control. All the terpene enhancers, with theexception of fenchone (10 61193139 mg:g) hadlower skin content values relative to the control.

3.3. Percutaneous absorption of carbamazepine in6itro

The effects of terpene enhancers on the percuta-neous permeation parameters of carbamazepine(flux, ERflux, cumulative amount of carba-mazepine after 24 h (Q24), and skin content) fromHPC gel formulations are presented in Table 4.The control percutaneous permeation parametersfor carbamazepine were 24.694.5 mg:cm2 per hfor flux, 6809123 mg:cm2 for Q24, and 38739644 mg:g for carbamazepine skin content. It isinteresting to note that all the terpene enhancers,with the exception of fenchone had major effectson the carbamazepine flux and Q24 relative to thecontrol. Nerolidol provided the highest increase inthe carbamazepine flux relative to the control. Itincreased the flux by 7.5-fold (P0.00041) fol-lowed by limonene with 6.6-fold (P0.00011),thymol 4.2-fold (P0.00077), and fenchone with1.5-fold (P0.049). Similar to the ERflux results,nerolidol provided the highest increase in the Q24(12839251 mg:cm2) relative to the control (P

Table 3Effect of terpene enhancers on the percutaneous parameters of hydrocortisone formulated in HPC gels a

Flux (mg:cm2 per h) ERflux Q24 (mg:cm2) SC (mg:g)Terpene

14598Control 1.0 805292906.091.510 61193139101791487.8Fenchone 47.294.9

10.563.4910.6 664092015Thymol 1158916815699232168.9929.3Limonene 28.0 834291250

32.7Nerolidol 46039142118379106196.5939.0

a Mean9S.D., n5. ERflux, enhancement ratio of hydrocortisone flux; Q24, cumulative amount of hydrocortisone in the receptorafter 24 h; SC, skin content of hydrocortisone after 24 h.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240234

Table 4Effect of terpene enhancers on the percutaneous parameters of carbamazepine formulated in HPC gels a

ERflux Q24 (mg:cm2)Terpene SC (mg:g)Flux (mg:cm2 per h)

1.0Control 680912324.694.5 38739644Fenchone 37.896.4 1.5 7739123 16869525

103.9913.2Thymol 4.2 994979 7840920126.6 1121948162.5916.97 16 74291531Limonene

Nerolidol 7.5185.1925.0 12839251 34 91099924

a Mean9S.D., n5. ERflux, enhancement ratio of carbamazepine flux; Q24, cumulative amount of carbamazepine in the receptorafter 24 h; SC, skin content of carbamazepine after 24 h.

0.020) followed by limonene (112148 mg:cm2)(P0.0041) and thymol (994979 mg:cm2) (P0.020). The lowest increase in the Q24 was pro-vided by fenchone (7739123 mg:cm2), which wasnot significantly different from the control (P0.41). It is interesting to note that nerolidol,limonene, and thymol significantly increased thecarbamazepine skin contents relative to the con-trol (P0.0058, 0.00016, and 0.032, respectively).However, the carbamazepine skin contentrecorded with fenchone (16869525 mg:g) wassignificantly lower than that provided by the con-trol (P0.0081).

3.4. Percutaneous absorption of tamoxifen in 6itro

The effects of terpene enhancers on the percuta-neous permeation parameters of tamoxifen (flux,ERflux, cumulative amount of tamoxifen after 24 h(Q24), and skin content) from HPC gel formula-tions are shown in Table 5. Controls consisted ofgel formulations with no terpene enhancer and thevalues for the percutaneous permeation of tamox-ifen control gel were found to be 2.891.0 mg:cm2per h for flux, 52925 mg:cm2 for Q24, and358491642 mg:g for tamoxifen skin content. Un-like the results found with other model drugs,terpene enhancers did not have major effects onthe percutaneous permeation parameters of ta-moxifen relative to the control. However,nerolidol provided the best enhancement activityfor tamoxifen. It increased the flux 1.7-fold rela-tive to the control followed by limonene with1.6-fold, and thymol 1.4-fold. Unlike other terpe-nes, fenchone decreased the flux 0.6-fold relativeto the control. Nerolidol also provided the highest

increase in the Q24. It increased Q24 to 82913mg:cm2 followed by limonene (81925 mg:cm2),thymol (74955 mg:cm2), and fenchone (57915mg:cm2). It should be emphasized that the ob-tained Q24 results with nerolidol, limonene, thy-mol, and fenchone were not significantly higherthan that provided by the control (P0.12, 0.21,0.51, and 0.78, respectively). The highest tamox-ifen skin content was observed with limonene(714091932 mg:g) followed by thymol (711593600 mg:g), fenchone (577391000 mg:g), andnerolidol (553291431 mg:g). It should be empha-sized that limonene, thymol, fenchone, andnerolidol did not increase tamoxifen skin contentssignificantly relative to the control (P0.065,0.196, 0.117, and 0.208, respectively).

3.5. Correlation of terpene efficacy with log P6alues of model drugs

In this study, the effect of terpene enhancers onthe percutaneous permeation of model drugs withdifferent lipophilicity as denoted by their log Pvalues were investigated. Correlation coefficientswere established between the lipophilicties of themodel drugs and the ERflux and between the ter-pene enhancers lipophilicities and the ERflux.

Fig. 1 presents the relationships between thelipophilicities of the model drugs (nicardipine hy-drochloride, hydrocortisone, carbamazepine, andtamoxifen) and their ERflux with the terpene en-hancers. The findings suggest that the increase inthe lipophilicities of the model drug is associatedwith a decrease in the percutaneous permeation ofthe drug. The relationships for the four terpeneenhancers were found to be linear [r 0.824,

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240 235

(P0.176) for fenchone, r 0.891, (P0.109)for thymol, r 0.846, (P0.154) for limonene,and r 0.769, (P0.232) for nerolidol].

Fig. 2 represents the relationships between thelipophilicities of the terpene enhancers (fenchone,thymol, limonene, and nerolidol) and the ERflux ofthe tested model drugs. It can be suggested thatthe increase in the lipophilicities of the terpeneenhancers is associated with an increase in thepercutaneous permeation of the drug. The rela-tionships for the five model drugs were found tobe linear [r0.951, (P0.049) for nicardipinehydrochloride, r0.977, (P0.023) for hydro-cortisone, r0.942, (P0.057) for carba-mazepine, and r0.874, (P0.126) fortamoxifen].

4. Discussion

Terpenes were reported to be effective penetra-tion enhancers for both hydrophilic and lipophilicdrugs (Cornwell and Barry, 1991; Hori et al.,1991; Gao and Singh, 1998; Zhao and Singh,1998). In this study, the tested terpene enhancersprovided significant enhancements for the fluxand cumulative amounts of the evaluated modeldrugs at 24 h (Q24). Nerolidol in particular wasthe most effective terpene enhancer in promotingthe permeation of all the model drugs followed bylimonene and thymol. Fenchone provided thelowest increase in the flux for all the tested modeldrugs. The obtained results are similar to thosereported with other research groups. In 1991,Cornwell et al. evaluated the effect of terpeneenhancers on the percutaneous permeation of 5-

fluorouracil across the skin. They reported thatnerolidol provided the highest enhancement forthe permeation of 5-fluorouracil (Cornwell andBarry, 1991). Furthermore, the high enhancementactivity of nerolidol was reported by Arellano etal., who found that nerolidol was an effectiveenhancer for the permeation of diclofenac sodiumacross the rat skin (Arellano et al., 1996). Theeffective promoting activity of nerolidol was at-tributed to its amphiphilic structure that was suit-able for alignment within the lipid lamellae of thestratum corneum and disrupting its highly orga-nized packing (Cornwell and Barry, 1994).

It is worth noting that limonene providedhigher enhancing activity for the permeation ofnicardipine hydrochloride (hydrophilic calciumchannel blocker log P 0.9990.1) and hydro-cortisone (a polar steroid with a log P of 1.4390.25) relative to fenchone and thymol. Thesefindings conflict with the results observed by otherauthors. It has been recognized that hydrophilicterpenes capable of hydrogen binding (such asfenchone and thymol) are more active towardspromoting the permeation of hydrophilic drugs,whereas, hydrocarbon terpenes (such as limonene)provide higher enhancing activity for lipophilicdrugs (Hori et al., 1991; Moghimi et al., 1997). In1992, Katayama et al. studied the effect of L-men-thol on the skin permeability of mannitol, corti-sone or indomethacin using hairless mouse skin invitro (Katayama et al., 1992). They observed thatwhen the donor solution contained L-menthol inan aqueous ethanol vehicle at pH 7.4 the perme-ability coefficients of mannitol and indomethacinwere increased by 100-fold relative to thatrecorded with the control (an aqueous vehicle)

Table 5Effect of terpene enhancers on the percutaneous parameters of tamoxifen formulated in HPC gels a

Flux (mg:cm2 per h) ERflux Q24 (mg:cm2) SC (mg:g)Terpene

35849164252925Control 2.891.0 1.057915 5773910000.6Fenchone 1.890.8749551.43.991.5Thymol 711593600

7140919351.64.591.3Limonene 819251.7Nerolidol 553291431829134.890.7

a Mean9S.D., n5. ERflux, enhancement ratio of tamoxifen flux; Q24, cumulative amount of tamoxifen in the receptor after 24h; SC, skin content of tamoxifen after 24 h.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240236

Fig. 1. The relationship between the log P of model drugs and the ERflux of the evaluated model drugs (nicardipine hydrochloride,hydrocortisone, carbamazepine, and tamoxifen), using fenchone, thymol, limonene, and nerolidol as penetration enhancers.

and increased those of cortisone by about 10-fold.However, L-menthol did not increase the perme-ation of indomethacin at pH 3.0, since the major-ity of the species would be in the unionized form.

The effects of terpenes on the permeation ofpropranolol hydrochloride (hydrophilic drug) anddiazepam (lipophilic drug) were evaluated by Horiet al. The purely hydrocarbon terpenes promotedboth propranolol and diazepam permeation (Horiet al., 1991). Whereas, the terpenes with hydro-gen-bonding ability only enhanced the flux ofpropranolol.

The data in our study can be explained by thefollowing information. The higher enhancementactivity of limonene relative to thymol and fen-chone can be attributed to its higher thermody-namic activity in the gel since limonene was notcompletely soluble in the gel at 2% concentration.On the contrary, thymol and fenchone were foundto be completely soluble in the gel at the evalu-ated concentration. Similar results were also re-ported by Obata et al., who attributed the higherenhancing activity of the permeation of diclofenacby limonene to its higher thermodynamic activity

in the 40% ethanol-buffer solution at the evalu-ated concentration relative to L-menthol (Obata etal., 1993).

The lipophilicity of the permeant, as well as theenhancer molecule is thought to play an impor-tant role in determining the enhancers promotingactivity on the permeation of the drug across theskin (Okabe et al., 1990; Hori et al., 1991; Lee etal., 1993, 1994; Phillips and Michniak, 1995; Bor-ras-Blasco et al., 1997; Godwin and Michniak,1999; El-Kattan et al., 2000a,b; Sung et al., 2000).

In the present study, terpene enhancers in com-bination with the gel solvent mixtures exhibitedhigh enhancement for flux and Q24 of the testedmodel drugs. The enhancers produced the highestactivity for the most hydrophilic drug i.e.nicardipine hydrochloride followed by hydrocorti-sone, and carbamazepine. The lowest activity ofthe enhancers was recorded with the mostlipophilic compound i.e. tamoxifen. Terpene en-hancers did not increase tamoxifen flux and Q24significantly relative to the control (Table 5). Theresults were anticipated, since other studies havedemonstrated the significant effect of model drug

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240 237

lipophilicity on the penetration enhancing proper-ties of terpenes (Kitagawa et al., 1998). Guy et al.reported that compounds with log P of more than2 may have potential problems in achieving steadyplasma concentrations in reasonable time spansdue to the fact that the drug is delayed in thestratum corneum where a reservoir may be estab-lished (Guy and Hadgraft, 1989). Kitagawa et al.evaluated the effect of L-menthol in 15% ethanolon the permeation of paraben analogues throughexcised guinea pig dorsal skin. They found thatthe permeability coefficients of the parabens cor-related with their lipophilicity (log P). Addition of1% L-menthol in 15% ethanol increased the per-meability coefficient of methyl paraben 16-fold,whereas this enhancer decreased that of butylparaben to about one fifth of the control value(Kitagawa et al., 1997).

In our study, significant linear correlations wereestablished between the ERflux and the log P ofthe tested model drugs for the four tested terpeneenhancers [r 0.824, (P0.176) for fenchone,r 0.891, (P0.109) for thymol, r 0.846,(P0.154) for limonene, and r 0.769, (P0.232) for nerolidol (Fig. 1)].

In the present investigation, the lipophilicitydenoted as log P of the terpene enhancers greatlyinfluenced the enhancement ratios of the evalu-ated model drugs at log P ranges of 2.135.36and strong correlation coefficients were estab-lished between the ERflux of nicardipine hydro-chloride, hydrocortisone, carbamazepine, andtamoxifen and the log P of the terpene enhancers[r0.951, (P0.049), r0.977, (P0.023), r0.942, (P0.057), and r0.874, (P0.126), re-spectively, (Fig. 2)]. These results are in a goodagreement with earlier findings reported byWilliams et al., who found linear relationshipsbetween the lipophilicity of the epoxide and ke-tone terpenes and the enhancement ratio for 5-fluorouracil permeation (Williams and Barry,1991). The mechanism of permeation enhance-ment of the terpene enhancers has been evaluatedusing differential scanning calorimetery, Fouriertransform infrared, and X-ray diffraction. Thesestudies suggested that terpenes enhance the per-meation of the drug across the skin mainly bydisrupting the highly ordered intercellular packingof the stratum corneum lipids and increasing drugdiffusivity (Williams and Barry, 1991; Cornwell et

Fig. 2. The relationship between the log P of the terpene enhancers and the ERflux of the evaluated model drugs (nicardipinehydrochloride, hydrocortisone, carbamazepine, and tamoxifen), using female hairless mouse skin.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240238

al., 1994). Taking these finding into consideration,it can be speculated that the increase in thelipophilicity of the terpenes may increase theirability to disrupt the highly hydrophobic lipidlamella of the stratum corneum.

5. Conclusions

In conclusion, terpene enhancer lipophilicityhad a significant effect on the percutaneous per-meation of the model drugs tested. Increasing thelog P of the terpene enhancers was associatedwith an increase in drug permeation. Nerolidolwas the most effective terpene enhancer for thepermeation of the drugs followed by limoneneand thymol. Fenchone was the least effective ter-pene enhancer. However, the higher enhancingactivity of limonene relative to fenchone and thy-mol was unexpected and attributed to its higherthermodynamic activity in the gel. Unlike fen-chone and thymol, limonene was not completelysoluble in the gel at 2% concentration. This maysuggest the importance of the gel formulationvehicle compositions on the enhancing activity ofthe terpene enhancers. Furthermore, model druglipophilicity had a significant impact on the ter-pene enhancers promoting activity. Increasing thelipophilicity of the model, drug was associatedwith a decrease in the enhancement activity of theterpenes. Terpene enhancers provided the highestenhancement activity for nicardipine hydrochlo-ride, which was the most hydrophilic compound.The lowest enhancement activity was provided fortamoxifen, which was the most lipophilic com-pound. It should be stressed that the obtainedresults were in agreement with earlier findings ofKitagawa et al., who reported that the addition of1% L-menthol in 15% ethanol increased the per-meability coefficient of methyl paraben 16-fold,whereas this enhancer decreased that of butylparaben to about one fifth of the control value(Kitagawa et al., 1997).

Further work needs to be conducted using ad-ditional drugs and other series of enhancercompounds.

References

Arellano, A., Santoyo, S., Martin, C., Ygartua, P., 1996.Enhancing effect of terpenes on the in vitro percutaneousabsorption of diclofenac sodium. Int. J. Pharm. 130, 141145.

Aungst, B.J., 1989. Structure:effect studies of fatty acid iso-mers as skin penetration enhancers and skin irritants. J.Pharm. Sci. 6, 244.

Banga, A.K., Bose, S., Ghosh, T.K., 1999. Iontophoresis andelectroporation: comparisons and contrasts. Int. J. Pharm.179, 119.

Borras-Blasco, J., Lopez, A., Morant, M., Diez-Sales, O.,Herraez-Dominguez, M., 1997. Influence of sodium laurylsulfate on in vitro percutaneous absorption of compoundswith different lipophilicity. Eur. J. Pharm. Sci. 5, 1522.

Bouwstra, J.A., Dubbelaar, F.E., Gooris, G.S., Weerheim,A.M., Ponec, M., 1999. The role of ceramide compositionin the lipid organisation of the skin barrier. Biochim.Biophys. Acta 1419, 127136.

Choi, H.K., Amidon, G.L., Flynn, G.L., 1991. Some generalinfluences of n-decylmethyl sulfoxide on the permeation ofdrugs across hairless mouse skin. J. Invest. Dermatol. 96,822826.

Coderch, L., de Pera, M., Perez-Cullell, N., Estelrich, J., de laMaza, A., Parra, J.L., 1999. The effect of liposomes onskin barrier structure. Skin Pharmacol. Appl. Skin Physiol.12, 235246.

Cornwell, P., Barry, B., 1991. Sesquiterpene components ofvolatile oils as skin penetration enhancers for the hy-drophilic permeant 5-flurouracil. J. Pharm. Pharmacol. 46,261269.

Cornwell, P.A., Barry, B.W., 1994. Sesquiterpene componentsof volatile oils as skin penetration enhancers for the hy-drophilic permeant 5-fluorouracil. J. Pharm. Pharmacol.46, 261269.

Cornwell, P.A., Barry, B.W., Stoddart, C.P., Bouwstra, J.A.,1994. Wide-angle X-ray diffraction of human stratumcorneum: effects of hydration and terpene enhancer treat-ment. J. Pharm. Pharmacol. 46, 938950.

Elias, P.M., 1996a. Stratum corneum architecture, metabolicactivity and interactivity with subjacent cell layers. Exp.Dermatol. 5, 191201.

Elias, P.M., 1996b. The stratum corneum revisited. J. Derma-tol. 23, 756758.

El-Kattan, A.F., Asbill, C.S., Michniak, B.B., 2000a. Theeffect of terpene enhancer lipophilicity on the percutaneouspermeation of hydrocortisone formulated in HPMC gelsystems. Int. J. Pharm. 198, 179189.

El-Kattan, A.F., Asbill, C.S., Michniak, B.B., 2000. The effectof terpene enhancer lipophilicity on the percutaneous per-meation of hydrocortisone formulated in HPMC gel sys-tems. Inter. J. Pharm., in press.

Gao, S., Singh, J., 1998. In vitro percutaneous absorptionenhancement of a lipophilic drug tamoxifen by terpenes. J.Contr. Rel. 51, 193199.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240 239

Godwin, D.A., Michniak, B.B., 1999. Influence of druglipophilicity on terpenes as transdermal penetration en-hancers. Drug Dev. Ind. Pharm. 25, 905915.

Guy, R., Hadgraft, J., 1989. Selection of candidates fortransdermal drug delivery. In: Hadgraft, J., Guy, R.(Eds.), Transdermal Drug Delivery. Marcel Dekker, NewYork, pp. 5982.

Heisig, M., Lieckfeldt, R., Wittum, G., Mazurkevich, G.,Lee, G., 1996. Non steady-state descriptions of drug per-meation through stratum corneum. I. The biphasic brick-and-mortar model. Pharm. Res. 13, 421426.

Herrlinger, C., Braunfels, M., Fink, E., Kinzig, M., Metz,R., Sorgel, F., Vergin, H., 1992. Pharmacokinetics andbioavailability of tamoxifen in healthy volunteers. Int. J.Clin. Pharmacol. Ther. Toxicol. 30, 487489.

Hori, M., Satoh, S., Maibach, H.I., Guy, R.H., 1991. En-hancement of propranolol hydrochloride and diazepamskin absorption in vitro: effect of enhancer lipophilicity.J. Pharm. Sci. 80, 3235.

Kaplun-Frischoff, Y., Touitou, E., 1997. Testosterone skinpermeation enhancement by menthol through formationof eutectic with drug and interaction with skin lipids. J.Pharm. Sci. 86, 13941399.

Katayama, K., Takahashi, O., Matsui, R., Morigaki, S.,Aiba, T., Kakemi, M., Koizumi, T., 1992. Effect of L-menthol on the permeation of indomethacin, mannitoland cortisone through excised hairless mouse skin.Chem. Pharm. Bull. (Tokyo) 40, 30973099.

Kitagawa, S., Hosokai, A., Aseda, Y., Yamamoto, N.,Aneko, Y., Matsuoka, E., 1998. Permeability of benzoicacid derivatives in excised guinea pig dorsal skin and theeffect of L-menthol. Int. J. Pharm. 161, 115122.

Kitagawa, S., Li, H., Sato, S., 1997. Skin permeation ofparabens in excised guinea pig dorsal skin, its modifica-tion by penetration enhancers and their relationship withn-octanol:water partition coefficients. Chem. Pharm.Bull. (Tokyo) 45, 13541357.

Kobayashi, D., Matsuzawa, T., Sugibayashi, K., Morimoto,Y., Kobayashi, M., Kimura, M., 1993. Feasibility of useof several cardiovascular agents in transdermal therapeu-tic systems with L-menthol-ethanol system on hairless ratand human skin. Biol. Pharm. Bull. 16, 254258.

Lee, C.K., Uchida, T., Kitagawa, K., Yagi, A., Kim, N.S.,Goto, S., 1993. Effect of hydrophilic and lipophilic vehi-cles on skin permeation of tegafur, alclofenac andibuprofen with or without permeation enhancers. Biol.Pharm. Bull. 16, 12641269.

Lee, C.K., Uchida, T., Kitagawa, K., Yagi, A., Kim, N.S.,Goto, S., 1994. Skin permeability of various drugs withdifferent lipophilicity. J. Pharm. Sci. 83, 562565.

Michniak, B., Player, M., Chapman, J., Sowell, J., 1993. Invitro evaluation of a series of azone analogs as dermalpenetration enhancers. Int. J. Pharm. 91, 8593.

Michniak, B., Player, M., Fuhrman, L., Christensen, C.,Chapman, J., Sowell, J., 1993. In vitro evaluation of aseries of azone analogs as dermal penetration enhancers.II. (Thio) amides. Int. J. Pharm. 94, 203210.

Michniak, B., Player, M.R., Chapman, J.M., Sowell, J.W.,1994a. Azone analogues as penetration enhancers: effectof different vehicles on hydrocortisone acetate skin per-meation and retention. J. Contr. Rel. 32, 147154.

Michniak, B.B., Player, M.R., Chapman, J.M., Sowell, J.W.,1993b. Optimization of dermal delivery using azoneanalogs as enhancers. Proc. 12th Pharm. Technol. Conf.2, 158173.

Moghimi, H., Williams, A., Barry, B., 1997. A lamellar ma-trix model for stratum corneum intercellular lipids. V.Effects of terpene penetration enhancers on the structureand thermal behavior of the matrix. Int. J. Pharm. 146,4154.

Obata, Y., Takayama, K., Maitani, Y., Machida, Y., Nagai,T., 1993. Effect of pretreatment of skin with cyclicmonoterpenes on permeation of diclofenac in hairless rat.Biol. Pharm. Bull. 16, 312314.

Okabe, H., Obata, Y., Takayama, K., Nagai, T., 1990. Per-cutaneous absorption enhancing effect and skin irritationof monocyclic monoterpenes. Drug Design Del. 6, 229.

Phillips, C.A., Michniak, B.B., 1995. Transdermal delivery ofdrugs with differing lipophilicities using azone analogs asdermal penetration enhancers. J. Pharm. Sci. 84, 14271433.

Psillakis, T., Tsatsakis, A.M., Christodoulou, P.,Michalodimitrakis, M., Paritsis, N., Helidonis, E., 1999.Carbamazepine levels in head hair of patients underlong-term treatment: a method to evaluate the history ofdrug use. J. Clin. Pharmacol. 39, 5567.

Southwell, D., Barry, B., 1983. Penetration enhancers forhuman skin: mode of action of 2-pyrrolidone anddimethylformamide on partition and diffusion of modelcompounds water, alcohols, and caffeine. J. Invest. Der-matol. 66, 243252.

Stott, P.W., Williams, A.C., Barry, B.W., 1998. Transdermaldelivery from eutectic systems: enhanced permeation of amodel drug, ibuprofen. J. Contr. Rel. 50, 297308.

Sung, K.C., Fang, J., Yoa-Pu Hu, O., 2000. Delivery ofnalbuphine and its prodrugs across skin by passive diffu-sion and iontophoresis. J. Contr. Rel. 67, 18.

Takahashi, K., Tamagawa, S., Katagi, T., Yoshitomi, H.,Kamada, A., Rytting, J.H., Nishihata, T., Mizuno, N.,1991. In vitro transport of sodium diclofenac across ratabdominal skin: effect of selection of oleaginous compo-nent and the addition of alcohols to the vehicle. Chem.Pharm. Bull. (Tokyo) 39, 154158.

Wade, A., Weller, P., 1995. Handbook of PharmaceuticalExcipients, Second edn. American Pharmaceutical Associ-ation, Washington DC.

Williams, A., Barry, B., 1990. Differential scanning calorime-tery does not predict the activity of terpene penetrationenhancers in human skin. J. Pharm. Pharmacol. 42,156P.

Williams, A.C., Barry, B.W., 1991. Terpenes and the lipid-protein-partitioning theory of skin penetration enhance-ment. Pharm. Res. 8, 1724.

A.F. El-Kattan et al. : International Journal of Pharmaceutics 215 (2001) 229240240

Wu, P., Huang, Y., Fang, J., Tsai, Y., 1998. Percutaneousabsorption of captopril from hydrophilic cellulose deriva-tives through excised rabbit skin and human skin. DrugDev. Indust. Pharm. 24, 179182.

Yamane, M.A., Williams, A.C., Barry, B.W., 1995. Terpenepenetration enhancers in propylene glycol:water co-sol-vent systems: effectiveness and mechanism of action. J.Pharm. Pharmacol. 47, 978989.

Zettersten, E.M., Ghadially, K.R., Feingold, D., Crumrine,A., Elias, P.M., 1997. Optimal ratios of topical stratum

corneum lipids improve barrier recovery in chrono-logically aged skin. J. Am. Acad. Dermatol. 37, 403408.

Zhao, K., Singh, J., 1998. Mechanisms of percutaneous ab-sorption of tamoxifen by terpenes: eugenol, D-limoneneand menthone. J. Contr. Rel. 55, 253260.

Zhao, K., Singh, J., 1999. In vitro percutaneous absorptionenhancement of propranolol hydrochloride throughporcine epidermis by terpenes:ethanol. J. Contr. Rel. 62,359366.

.