International Journal of Pharmaceutics, 74 (1991) 157-168
`© 199] Elsevier Science Publishers B.V. 0378-5173,/91/$03.50
`ADONIS 037851739 1(W12854
`
`LIP 02480
`
`157
`
`The enhancement index concept applied to terpene penetration
`enhancers for human skin and modellipophilic (oestradiol)
`and hydrophilic (5-fluorouracil) drugs
`
`A.C. Williams and B.W. Barry
`Posturaduaie Studies in Pharmaceutical Technology, The School of Pharmacy, University of Bradford, Bradford (UK)
`(Received 22 March 1991)
`(Accepted 12 April 1991)
`
`Key words: Percutaneous absorption; Skin penetration enhancer; Terpene; Oestradiol; 5-Fluorouracil:
`Enhancement index
`
`Summary
`
`Aseries of cyclic monoterpenes has heen assessed as skin penetration enhancers towards a model lipophilic drug. oestradiol. In
`vitro permeation experiments on human epidermal membranes showed that the terpenes varied im their activities; hydrocarbon
`{e.g.. limonene) and cyclic ether (e.g., 1.8-cineole) terpenes were effective accelerants providing approximately 4-fold increases in
`
`the permeability coefficient of aqueous oestradiol, whereas alcohols (e.g., carveol), ketones (e.g. menthone) and epoxides (e.g.,
`
`
`pinene oxide) were ineffective. The results of this study are compared with terpene activities towards a model hydrophilic drug,
`$-fluorouracil. A novel concept, the enhancement index (EJ), is introduced to compare differences in terpene activities towards the
`two permeants: El provides information as to the partition coefficient and maximum achievable permeation enhancement for a
`drug, together with a measure of a penetration enhancer’s activity towards that drug expressed as a percentage of the maximum
`effect. This approach permits useful comparisons between the activities of various enhancers towards different drugs.
`
`Introduction
`
`The rate determining step for transdermal de-
`livery of most drugs is provided by the stratum
`corneum (Scheuplein, 1965).
`Its structure has
`been depicted in the brick and mortar model
`(Michaels et al., 1975; Elias, 1981) in which anu-
`cleate keratinised cells are embedded in alipid
`mortar. The stratum corneum lipids are arranged
`
`Correspondence: B.W. Barry, Postgraduate Studies in Pharma-
`ceutical Technology. The School of Pharmacy, University of
`Bradford Bradford, BD7 [DP, U.K.
`
`in multiple bilayers providing alternate hydropho-
`bic and hydrophilic barriers. Drugs must diffuse
`through the intercellular lipid matrix, and to re-
`duce reversibly the resistance of
`this pathway
`researchers employ penetration enhancers (or ac-
`cclerants), These materials interact reversibly with
`stratum corneum constituents to disrupt the highly
`ordered structure and hence facilitate drug diffu-
`sion. Many established penetration enhancers are
`synthetic chemicals which are not yet approved
`by regulatory authorities for use with drugs. Re-
`cently, a novel series of penetration enhancers,
`classed as terpenes or terpenoids, has been de-
`scribed (Williams and Barry, 1989, 1990). These
`
`Noven Pharmaceuticals, Inc.
`EX2012
`Mylan Tech., Inc. v. Noven Pharma., Inc.
`IPR2018-00174
`
`0001
`
`

`

`[58
`
`chemicals may provide a series of safe, naturally
`occurring penetration enhancers whose toxicitics
`are well documented (e.g., Opdyke, 1974-1976),
`Several terpenes were shown to be effective ac-
`celerants for the hydrophilic cytotoxic drug, 5-flu-
`orouracil (Williams and Barry, 1991). The present
`study extends the investigation to the effects of
`some terpenes on transdermal permeation of a
`model lipophilic drug, oestradiol (ES).
`Topical oestrogens are employed when cn-
`dogenous hormonesare lacking, such as in post-
`menopausal women. A transdermal oestradiol
`patch, Estraderm TTS, has recently been devel-
`oped to treat menopausal symptoms.Clinical tri-
`als have indicated that transdermal delivery holds
`many advantages over oral oestrogen administra-
`tion,
`including reduced variation in serum hor-
`mone concentrations, a more normal oestronc:
`oestradiol ratio and minimal pharmacological ef-
`fects on hepatic proteins (Powers et al., 1985;
`Crust et al., 1989; Yum, 1989),
`Our report also compares the activities of ter-
`pene penetration enhancers towards the model
`hydrophobic drug (oestradiol) and the model hy-
`drophilic drug (5-fluorouracil). A novel concept,
`the enhancement index (EI), is used to compare
`accelerant actions for the two drugs; it
`is hoped
`that such an approach may be of value with a
`wide variety of drugs and enhancers.A significant
`advantage of the method is that
`it allows an
`assessment of the maximum benefit which can be
`expected in chemically enhancing the skin perme-
`ation of a particular drug.
`
`isolated from the oil (supplied by Field and Co.)
`by fractional distillation in vacuo (Pinder, 1960).
`The chemical
`formulae of these terpenes are
`given in Fig. 1.
`Anassessment of the terpene purities has been
`published (Williams and Barry, 1991); no single
`impurity was present in each terpene at greater
`than 2%, and such traces were considered to be
`at sufficiently low thermodynamic activities that
`their effects on skin permeability would be negli-
`gible compared with that of the main terpenc. As
`an initial assessment of accelerant activity. ail
`terpenes were employed as neat liquids.
`The model
`lipophilic permeant was [2,4.6,7-
`SH(N)Joestradiol (NEN Research Products),
`ra-
`diochemical purity 99%. Unlabelled oestradiol
`(Sigma Chemical Company) was used to prepare
`a saturated aqueous drug solution (0.003 mg/ml
`at 30°C, Michaels et al., 1975).
`
`® oo Y.
`
`ene
`
`Po
`
`d-Limonene
`
`3-Carene
`
`a -Terpineol
`
`OH
`
`0
`
`oO
`
`Terpiner-4-al
`
`™~
`
`Carveo}
`
`as
`
`Carvone
`
`ni
`Pulegone
`
`Materials and Methods
`
`°
`
`The terpenes used as received were a-pinenc,
`3-carene,
`terpinen-4-ol, carveol, carvone, pule-
`gone, menthone, a-pinene oxide, limonene oxide,
`cyclohexene oxide, cyclopentene oxide and 7-
`oxabicyclo[2.2.1]heptane
`supplied by Aldrich
`Chemical Company, d-limonene and 1,8-cineole
`provided by Sigma Chemical Company,
`a-
`terpineol obtained from BDH Chemicals Limited
`and piperitone from Field and Co. Ascaridole,
`the main constituent of oil of chenopodium was
`
`0002
`
`h
`
`0
`
`=
`
`Pipentone
`
`Menthone
`
`Cyclohexene
`oxide
`
`Limonene
`oxide
`
`0
`
`Sy
`
`Pinene
`Ose
`
`7
`
`oO
`
`neeanidole
`ASCAMNGCIE
`
`7 Oxabicycte
`{2 2 t]heptane
`
`~
`La crvaln
`
`Fig. 1. The structural formutae of terpenes used in this study.
`
`

`

`Preparation of human skin membranes
`Caucasian abdominal skin (male and female,
`age 17—89) was obtained postmortem. Excessfatty
`and connective tissues were removed and the
`samples stored at — 20°C (Harrison et al., 1984).
`Full thickness membranes.
`Skin samples were
`trimmed of fatty material
`to provide tissue ap-
`proximately 1 cm thick and essentially flat. The
`samples were clamped between stainless steel
`plates with a polythene sheet covering the stra-
`tum corneum and the membrane refrozen to ad-
`here the fatty layer to the metal. The upper plate
`and polythene sheet were removed andthe stra-
`tum corneum surface of the tissue was thawed
`slightly before a membrane, approximately 430
`jm, was cut using a Duplex Electro Dermatome
`7, providing a sample of full thickness skin com-
`prising stratum corneum, nucleate epidermis and
`some dermal tissue.
`Samples of
`Stripped full thickness membranes.
`full thickness skin with the stratum corneum re-
`moved were prepared by tape stripping (Clipper
`tape). Typically 25—30 strippings were required to
`remove the stratum corneum from full thickness
`skin membranes. Fully hydrated stratum corneum
`comprises approximately 30 um of the mem-
`brane, hence the resulting stripped full thickness
`tissue provided a sample approximately 400 um
`thick.
`Epidermal membranes. Epidermal mem-
`branes,
`incorporating the
`anucleate
`stratum
`corneum and nucleate epidermal
`tissue, were
`prepared by a heat separation technique (Klig-
`man and Christophers,
`1963). Skin samples
`trimmedoffatty tissue were immersed in water at
`60°C for 45 s, after which the epidermal mem-
`branes were teased off the underlying dermis.
`The membranes were floated on an aqueous solu-
`tion of 0.002%sodium azide for 36 h to ensure
`full hydration of the stratum corneum.
`Stratum
`Stratum corneum membranes.
`corneum samples were prepared from epidermal
`membranes (Kligman and Christophers, 1963):
`epidermal membranes were floated overnight on
`an aqueous solution of trypsin (0.0001% w/v)
`and sodium hydrogen carbonate (0.5% w/v) at
`37°C. The enzyme digests the nucleate epidermal
`tissue allowing the remnants to be removed by
`
`0003
`
`159
`
`swabbing. The stratum corneum membranes were
`floated on water before use to ensure full tissue
`hydration.
`A diagrammatical representation of the skin
`membranes used in this study is shownin Fig. 2.
`
`Permeation experiments
`Experiments at 32+ 1°C used an automated
`diffusion apparatus with 24 stainless-steel diffu-
`sion cells (diffusional area 0.126 cm*) and 0.002%
`aqueous sodium azide as flow-through receptor
`solution (Akhter et al., 1984).
`Fully hydrated epidermal membrane samples
`were mounted in the cells and treated with 150
`1 aliquots of saturated radiolabelled ES. To
`ensure saturation of the donor solution a crystal
`of ES, with the same radioactivity as the drug
`solution, was placed in each donor compartment,
`and the donor drug solution was replenished ev-
`ery 8 h. Under these conditions, ES ts at maximal
`
`sc
`
`FULL THICKNESS
`MEMBRANE
`
`STRIPPED FULL
`THICKNESS MEMBRANE
`
`MEMBRANE
`
`
`
`STRATUM CORNEUM
`
`EPIDERMAL
`MEMBRANE
`
`Fig. 2. A diagrammatical representation of skin membranes
`employed in this study (not to scale). SC. stratum corneum: E,
`nucleate epidermis; D, dermis.
`
`

`

`160
`
`thermodynamic activity throughout the diffusion
`membranes wasas described previously (Williams
`and Barry, 1991). Fully hydrated stratum corneum
`experiment with negligible donor depletion. Sam-
`ples of the receptor solution (2 ml) were collected
`samples were equilibrated in a terpene for }2 h.
`The tissues were blotted dry and placed in a
`every hour for 24 h,
`to which 5 ml OptiPhase
`saturated radiolabelled aqueous solution of ES
`HiSafe
`II
`scintillation fluid (Pharmacia) was
`for 4h. The samples were blotted dry, solubilised
`added, and the radiolabelled drug determined by
`and the drug determined byliquid scintillation
`liquid scintillation counting (Packard Tri-Carb
`460). The permeant solution was washed from the
`counting. Triplicate partition coefficients (stra-
`tum corneum/water) were determined using tis-
`membrane with 0.002% aqueous sodium azide
`and replaced with 150 ul of a terpene. After 12 h
`suc samples from three different human sources.
`treatment,
`the terpene was washed from the
`Controls were stratum corneum samples un-
`membrane and ES permeation was redetermined
`treated with terpene. The results were expressed
`as a partition ratio Pp where:
`as above. Linear regression analysis of the pseudo
`steady-state diffusion results after the lag time
`allows evaluation of the permeability coefficient
`(Kp) of the drug in the membrane before and
`after terpene treatment. As a measure of the
`penetration enhancing activity of the terpenes,
`the enhancement ratio (ER) was calculated as
`(Goodman and Barry, 1988):
`
`partition coefficient after terpene treatment
`Rk
`aoe
`re
`a
`~
`-
`~
`partition coefficient with untreated membrane
`
`ER
`
`Kp after terpene treatment
`Kp before terpene treatment
`
`(1)
`
`Values reported are meanratios from 4—13 repli-
`cates.
`Permeation across the epidermis and dermis
`and clearance into the aqueous receptor fluid
`may provide the rate determining step in the
`permeation of very highly lipophilic drugs. The
`barrier to ES diffusion was therefore investigated
`using skin membranes composed of a varicty of
`layers: stratum corneum alone, epidermis (includ-
`ing stratum corneum), full thickness skin (stratum
`corneum, nucleate epidermis and some dermal
`tissue) and tape stripped full thickness skin (stra-
`tum corneum removed). Oestradiol is not a very
`highly lipophilic drug (log P octanol / water =
`2.29) and has a small but significant aqueous
`solubility (0.003 mg/ml at 30°C). Thus, removal
`of the drug into the flow through receptorfluid is
`unlikely to provide a significant resistance to drug
`permeation, but passage
`across
`epidermal /
`dermal tissue may.
`
`Partitioning experiments
`The method used to assess the effects of ter-
`penes on ESpartitioning into stratum corneum
`
`0004
`
`Solubility studies
`The solubility of ES in the terpenes was deter-
`mined by the method of Williams and Barry
`(199]). Terpenes were saturated with radiola-
`belled crystals of ES and the saturated drug con-
`centration determined in triplicate by liquid scin-
`tillation counting.
`
`Results and Discussion
`
`The experimental design for permeation stud-
`ies of determining Kp,
`treating the membranes
`with a terpene and then redetermining Kp, allows
`each piece of tissue to act as its own control,
`thereby reducing errors due to the biological vari-
`ability of human skin. The conditions for drug
`delivery were optimised by the use of saturated
`drug solutions,
`replenished every 8&8 h, which
`maintains the permeant at near maximal thermo-
`dynamic activity. Typical permeation profiles un-
`der
`these conditions are in Fig. 3. From the
`diffusion experiments, the mean permeability co-
`efficient of aqueous ES through normal
`(un-
`treated) human epidermal membranesat 32°Cis
`3.68 + 0.36 x 107° em/h (n = 144). This result
`shows good agreement with literature values of
`3.2 x 1074 em/h (Goodman and Barry, 1988) and
`5.2 10°* cem/h (Michaels et al., 1975; Flynn
`and Stewart, 1988).
`
`

`

`161
`
`(a)
`
`(b)
`
`5.0
`4.5
`4.0
`3.5
`3.0
`2.5
`2.0
`1.5
`1.0
`0.5
`
`
`
`
`
`CUMULATIVECPM/cm?(x10*)
`
`
`
`
`
`CUMULATIVECPM/cm?(x10*)
`
`00> 4
`
`6
`
`B
`
`employedin diffusion experiments. Relevant data
`are summarised in Table 1.
`These results illustrate no significant differ-
`ence in ES permeation through stratum corneum,
`epidermal or full
`thickness membrane with an
`aqueous receptor fluid (P=0.05). This clearly
`illustrates that intact stratum corneum is the main
`barrier to oestradiol permeation. The resistance
`to drug permeation provided by various skin lay-
`ers may be calculated as the reciprocal of the
`
`16
`#18 20 27
`24
`o
`2
`4
`6
`8
`10
`12
`14
`drug permeability coefficient. It should be noted
`TIME=(h)
`that
`the resistance of each membrane includes
`those contributions arising from donor and recep-
`tor stationary layers. Due to the natural variabil-
`ity of human skin, no significant difference exists
`between the resistances of stratum corneum, epi-
`dermal and full thickness membranes (P = 0.05).
`However,
`the stratum corneum provides signifi-
`cantly greater resistance to drug permeation than
`stripped full
`thickness skin (E/D in Table 1;
`P=(.05). After removal of the stratum corneum
`the permeability coefficient of ES increases by a
`factor of 5 compared with full thickness skin with
`the stratum corneum intact. This result
`implies
`that
`the maximum enhancement effect
`that an
`accelerant such as a terpene may induce is a
`5-fold increase in permeability corresponding to
`full removal of the barrier resistance of the stra-
`tum corneum. Clearly, if the barrier nature of the
`stratum corneum is diminished by the use of
`penetration enhancers,
`then the resistance pro-
`vided by epidermal/dermal
`tissue will exert a
`proportionally greater
`influence on oestradiol
`permeation.
`Table 2 shows data on penetration enhancing
`activities of the terpenes. These results demon-
`strate that the hydrocarbon terpenes are acceler-
`
`12
`
`16
`
`18
`
`20
`
`22 24
`
`14
`10
`(h)
`TIME
`Fig. 3. Typical permeation profiles of aqueous oestradiol
`through human epidermal membranes showing the effects of
`donor replenishment. a. Donor not
`replenished. b. Donor
`replenished. Every second experimental value plotted.
`
`For highly lipophilic drugs, the rate determin-
`ing step in transdermal permeation may reside in
`the partitioning process into the essentially aque-
`ous nucleate epidermis and dermis, and/or in
`clearance from the skin into the systemic circula-
`tion. To investigate the rate limiting step for ES
`permeation, a variety of skin membranes were
`
`TABLE|
`
`The permeability coefficient (Kp) of aqueous oestradiol permeating through various layers of human skin, with SE
`
`Membrane
`Description
`Kp (em/h, x 10%)
`Tissue resistance (h/cm)
`n
`
`3
`SC
`Stratum corneum
`447+0.77
`224 +39
`144
`SC/E
`Epidermal
`3.68 + 0.36
`272 +27
`6
`SC/E/D
`Full thickness
`2.45 + 0.42
`408 +89
`
`Stripped full thickness 13.0 + 1.23 769+ 7.8E/D 8
`
`
`
`
`-
`
`SC, stratum corneum; E, epidermis; D, dermis; m, number ofreplicates.
`
`0005
`
`

`

`ho2
`
`TABLE 2
`
`3.72 + 0.61
`2BY + 1.47
`1.57 + 0.32
`
`The mean permeability coefficients (Kp), with standard error of the mean, of aqueous oestradiol through epidermal membranes before
`and after treatment with a terpene, with mean (and SE) enhancement ratios (ER) and the number ofreplicates (n)
`Terpene
`Kp (em/h, x 104)
`ER
`"
`
`Initial (contro)
`Treated
`
`Hydrocarbons
`a-Pinene
`d-Limonene
`3-Carene
`
`+093
`8.40 + O41
`6.57 + 0.69
`
`3.09 + (L89
`3.75 + 1.32
`4.364 1.02
`
`12
`12
`4
`
`Alcohols
`c-Terpineol
`Terpinen-4-ol
`Carveoal
`
`Ketones
`
`Carvone
`Pulegone
`Piperitone
`Menthone
`
`Oxides
`
`Cyclohexene oxide
`«-Pinene oxide
`Limonene oxide
`Ascaridole
`7-Oxubicyelo-
`{2.2.1 Jheptane
`1.8-Cineole
`
`5.98 + 0.73
`2.52 + 0.49
`7.36 + 0.49
`
`4.09 + 1,06
`3.01 + 1.11
`3.54 + 0.89
`3,22 + 0.7]
`
`3.33 + 0.59
`2.68 + 0.31
`2.29 + 0.64
`3.39 + O87
`
`2.94 + 0,99
`3.26 + 1.24
`
`2.34 + 0.38
`1.63 + 0.68
`1.064 O31
`
`0.40 + 0417
`0.98 + 0.26
`01.63 + 0.18
`Lol + 0.35
`
`4.68 + O84
`S.l44 1.29
`3.46 + 1.00
`15.8 4+ 4.42
`
`12.7 +343
`+ 3.78
`
`(1.33 + 0.10
`O45 +010
`42 +010
`
`0.104 0.04
`0.34 + 0.08
`O.17 + 0.4)2
`0.364 0.17
`
`1.424 0.75
`1.90 + O87
`i.61 + 0.76
`4.75 + 1.34
`
`4.93 + 1.50
`4.40 + 0.57
`
`6
`5
`u
`
`4
`3
`5
`7
`
`f
`5
`5
`s
`
`4
`7
`
`ants for the lipophilic drug, providing enhance-
`ment ratios of between 3 and 4. The alcohol and
`ketone terpenes did not enhance oestradiol per-
`meation and may in fact hinder passage of the
`lipophilic drug. The oxide terpenes show varied
`accelerant activities; cyclohexene oxide, limonene
`oxide and a@-pinene oxide provide no significant
`increase
`in oestradiol permeation (P = 0.05)
`whereas
`ascaridole,
`7-oxabicyclo{2.2.1}heptane
`and 1,8-cineole all induce a 4—5-fold increase in
`diffusion of the lipophilic drug. It is interesting to
`note that the ineffective oxide terpenes all con-
`tain a 1,2-oxygen linkage and may broadly be
`classified as epoxides. The effective oxide ter-
`penes contain 1.4-(ascaridole. 7-oxabicyclo[2.2.1}+
`heptane) or 1,8-(1.8-cineole) oxygen linkages and
`may be classed as cyclic ethers. It has been sug-
`gested that structural conformations may be a
`
`factor in determining penctration enhancer activ-
`ities of terpenes (Williams and Barry, 1991).
`The mechanisms underlying terpene penetra-~
`tion enhancement were investigated by partition-
`ing and solubility studies. Terpene effects on
`oestradiol! partitioning into isolated fully hydrated
`stratum corneum membranes were assessed (Ta-
`ble 3). The control (non-terpene treated) parti-
`tion coefficient (P) was 14.5, providing a log P
`(stratum corneum/ water) of 1.16. Literature P
`values include 46 for human stratum corneum
`(Scheuplein et al., 1969), 58.9 for male abdominal
`hairless mouse skin (Valia and Chien, 1984) and
`80 for rat stratum corneum (Wepierre et al.,
`1990). The partition coefficient (octanol / water),
`which gives a rank order approximationto parti-
`tioning into the stratum corneum (Barry, 1983),
`for oestradiol is 195. Thus, the results from this
`
`0006
`
`

`

`TABLE3
`
`Theeffects of terpenes on the partitioning of oestradiol from aqueoussolution into fully hydrated stratum corneum membranes, and the
`solubility of oestradiol in the terpenes
`
`Terpene
`P+SE
`Ppt+SE
`Solubility
`
`(mg/ml)
`
`163
`
`0.003
`1.00
`1.79
`14.5 +4
`Control (water)
`0.011
`0.99 + 0.15
`13.64 0.19
`a-Pinene
`0.025
`0.96 + 0.08
`13.54 0.77
`d-Limonene
`0.046
`1.93 + 0.17
`32.1 + 3.02
`3-Carene
`5.00
`3.29 + 1.21
`54.6 + 19.9
`a-Terpineol
`6.96
`2.04 + 0.06
`29.54 3.56
`Terpinen-4-ol
`5.27
`2.10 + 0.46
`28.3 + 3.95
`Carveol
`12.80
`3.31 + 0.25
`47.44 5.94
`Carvone
`12.33
`4.56 + 0.81
`61.74 2.06
`Pulegone
`13.06
`3.94 + 0.27
`56.04 5.51
`Piperitone
`5.60
`7.26+ 1.15
`100) +141
`Menthone
`755
`1.68 + 0.18
`23.5+ 1.76
`Cyclohexene oxide
`6.83
`2.34 + 0.60
`30.6 + 3.25
`a-Pinene oxide
`7.36
`2.65 + 0.30
`37.7 + 5.71
`Limonene oxide
`0.291
`4.30 + 0.86
`718+ 14.9
`Ascaridole
`64,57
`1.57 + 0.27
`21.34 0.66
`7-Oxabicyclo[2.2.] Jheptane
`
`
`
`30.0 + 2.37 2.11+0.121,8-Cineole 8.02
`
`P, partition coefficient, stratum corneum/water, mean of 3 replicates, with SE. Pp = partition ratio = P/control = P/14.5, mean of
`3 individual values, with SE.
`
`study and the literature confirm that the stratum
`corneum provides a more polar environment than
`octanol.
`Terpene treatment generally increases parti-
`tioning of the lipophilic drug into the stratum
`corneum, asillustrated by the partition ratio, Pp,
`where:
`
`partition coefficient after terpene treatment
`R™ partition coefficient with untreated membrane
`
`(3)
`
`However, it is evident from Table 3 that treat-
`ment with d-limonene and a-pinene does not
`improve such partitioning whereas the other hy-
`drocarbon tested, 3-carene, provides a near dou-
`bling in partitioning. Improved partitioning would
`be expected considering the solubility of ES in
`the terpenes; the drug is more soluble in all the
`terpenes than in water, and is more soluble in
`oxygen-containing terpenes than in the hydrocar-
`bons (Table 3). No mathematical relationship ex-
`ists between the partition ratios and drug solubili-
`ties, although a trend is apparent with the hydro-
`carbons havinglittle or no effect on partitioning
`and providing a drug solubility of 0.011—0.046
`
`mg/ml compared with the ketones which im-
`prove partitioning approximately 4-fold and pro-
`vide a greater drug solubility of 5.6-13 mg/ml.
`Terpene effects on the apparent lag time to
`pseudo steady-state diffusion have been used to
`gain an insight into the accelerant effects towards
`5-fluorouracil permeation (Williams and Barry,
`1991). However, oestradiol permeation through
`untreated epidermal membranes providesa rela-
`tively short lag time of 4.42 + 0.36 h (SE; n = 93).
`Consequently, variations in the apparent lag times
`are not sufficient
`for a critical analysis of the
`modes of action of
`terpene penetration en-
`hancers. The inconsistent nature of the apparent
`lag time, due to the biological variability of hu-
`man skin,
`is well illustrated by the above value:
`the 93 replicate values provide a standard devia-
`tion of 79%. Additionally, the lag time aboveis
`considerably lower than literature values of 19 h
`(Mollgaard and Hoelgaard,
`1983) and 103
`h
`(Scheuplein et al., 1969).
`The oestradiol work has shown that hydrocar-
`bon and cyclic ether terpenes are effective accel-
`erants whereas alcohols, ketones and epoxides
`
`0007
`
`

`

`164
`
`including
`not. Hydrocarbon terpenes,
`are
`limonene,
`increased percutaneous absorption of
`indomethacin in rats (Nagai et al., 1989; Okabe et
`al., 1989), by direct effects on the barrier nature
`of skin. This implies that the hydrocarbonsact by
`increasing diffusivity of the drug in the stratum
`corneum.
`Indomethacin (log P octanol / water
`4.27),
`like oestradiol (log P 2.29),
`is a lipophilic
`drug and the authors reported that oxygen-con-
`taining terpenes, such as carvone, a-terpincol
`and 1,8-cineole were ineffective. Our results par-
`tially support
`this view; hydrocarbon terpenes
`promoted ES permeation whereas alcohol and
`ketone terpenes did not. However, this studyalso
`demonstrated that
`the
`cyclic cther
`terpenes
`(ascaridole, 7-oxabicyclo[2.2.1]heptane and 1.8-
`cineole) promoted ES permeation across human
`epidermal membranes, to a similar extent as the
`hydrocarbons. Another study employing lipophilic
`drugs and terpene penetration enhancers found
`that terpenes containing polar groups did not act
`as enhancers (Hori et al., 1989).
`A comparison of the terpene activities towards
`the polar drug 5-fluorouracil (Williams and Barry,
`1991) and oestradiol
`is given in Fig. 4a. The
`terpenes apparently are more active towards 5-
`fluorouracil
`(5-FU) than towards ES. The en-
`hancement ratios of |,8-cineole, an effective ac-
`celerant
`for both drugs, are 94.5 and 4.40 for
`5-FU and ES respectively. However, a compari-
`son between these results may be misleading,
`particularly in respect of the apparent low value
`for ES, as the scope for enhancement of these
`drugs varies considerably. The enhancement ratio
`for 5-FU provided by tape stripping the stratum
`corneum from full thickness human skin is 1045;
`the corresponding value for ES is only 5.3. We
`can deal with this consideration by treating the
`skin as a laminate with barriers in series.
`The resistance (AR) of
`intact,
`full
`thickness
`human skin to drug diffusion may be described
`mathematically by (Barry, 1983):
`
`
`
`pe ly
`Dsc Kc
`
`te to
`Dy Kp
`DyK,,
`
`where the subscripts SC, E and D refer to the
`stratum corneum, nucleate epidermis and dermis
`
`0008
`
`100
`
`90
`°
`80
`&
`70
`c
`60
`5
`50
`Ww
`a8
`40
`30
`<
`20
`z
`“10
`
`0 234567 a°pias
`
`C 5-Fu
`
`10
`
`8
`6
`
`igwo
`z
`
`4
`
`2
`
`0
`
`b
`a=
`6z
`<=
`+ 2lw
`-4
`
`12345 6
`
`7
`
`8 9 1011121314151617
`TERPENE
`CJ) 5-Fu @ €s
`
`Fig. 4. a. The mean enhancement ratios of terpenes towards
`5-fluorouracil and oestradiol. b. The enhancement
`indices
`calculated from Eqn 7. 1. a-pinene; 2. d-limonene: 3, 3-carene:
`=
`4. a-terpineol: 5,
`terpinen-4-ol; 4, carveol:
`7. carvone: &,
`pulegone; 9, piperitone; 10, menthone: | 1, cyclohexene oxide:
`12.
`limonene oxide; 13,
`a@-pinene oxide:
`14. cyclopentene
`oxide: 15, ascaridole; 16. 7-oxabicyclo[2.2. | heptane: 17. 1.8-
`cineole.
`
`respectively, #, D and K refer to the thickness,
`diffusion coefficient and partition coefficient of
`the various
`layers, and stationary layers and
`clearance are neglected.
`Assuming that the nucleate epidermis and der-
`mis provide similar barrier properties per unit
`thickness, then:
`
`h
`Ne
`R= +
`Dgyc-Kg-
`DK
`
`(5)
`
`where h=hy +hy, D=D,.=D, and K=K, =
`K,,. A problem with tape stripping studies arises
`from the mechanical weakness of epidermal
`
`

`

`to tape-strip the
`is not practical
`it
`membranes;
`stratum corneum from epidermal membranes.
`Thus, the maximum enhancement ratios for the
`drugs obtained with the barrier layer removed
`were determined using stripped full thickness skin
`samples, approximately 400 um thick. These ER
`values are then compared with data obtained
`using epidermal membranes, approximately 80
`um thick of which approximately 30 um is fully
`hydrated stratum corneum and 50 um nucleate
`epidermis (Marks, 1981; Sato et al., 1991). Since
`the nucleate epidermal tissue represents approxi-
`mately one eighth of the stripped full
`thickness
`skin, then for epidermal membranesusedin dif-
`fusion studies Eqn 5 may be rewritten as:
`
`hee
`R=—— +
`
`h
`
`
`(6)
`
`With the stratum corneum intact, drug perme-
`ation across the bulk of the skin (nucleate epider-
`mis and dermis)
`is
`rapid for hydrophilic and
`moderately lipophilic molecules compared with
`passage across the stratum corneum. However,
`when the barrier layer is removed, permeation
`across the nucleate epidermal and dermal tissue
`becomesrate limiting. Assuming that the epider-
`mis and dermis provide similar
`resistances to
`drug permeation per unit thickness, a correction
`may be made for the differences in skin sample
`thickness between the two experimental proto-
`cols;
`the permeability coefficients of the drugs
`will be approximately 8 times greater through 50
`wm of nucleate epidermis than through 400 4m
`epidermal/ dermal
`tissue. Thus,
`the maximum
`achievable enhancement ratios through nucleate
`epidermal membranes can be corrected to ap-
`proximately 8400 for 5-FU and 42 for ES. Clearly
`these two values are approximations, but they are
`more appropriate than the maximum enhance-
`ment ratios taken simply from the stripped full
`thickness tissue studies. Literature reports of
`penetration enhancers seldom provide data for
`drug permeation through epidermal
`tissue with
`the stratum corneum removed. In such cases an
`estimate of drug permeability coefficients may be
`calculated from the drug diffusivity in aqueous
`
`0009
`
`165
`
`systems, assuming the epidermis to be a porous
`hydrogel.
`To describe the activities of penetration en-
`hancers in a more informative way, we propose
`that an enhancement index (EI) may be useful.
`The maximum achievable enhancement is depen-
`dent,
`in part, on the partition coefficient of the
`model permeant. The log P (octanol/ water) of
`the drug is thus provided as a superscript to the
`enhancement index, and the maximum enhance-
`ment
`ratio provided by stripping the stratum
`corneum from nucleate epidermal membranes,
`corrected as discussed above,
`is given in a sub-
`script. These two values provide information
`which places into context
`the enhancement ef-
`fect. The enhancement index is then calculated as
`the fraction of the maximum achievable enhance-
`ment ratio induced by accelerant treatment, ex-
`pressed as a percentage. However, our definition
`of the enhancement ratio, which is the ratio of
`drug Kp after to that before accelerant
`treat-
`ment, dictates that an enhancement ratio of 1.0
`indicates that an accelerant has no penetration
`enhancing activity. To correct
`for this,
`the en-
`hancement index maybe defined as:
`
`-y permeantlog P
`Elmum ER =
`
`(Enhancement ratio after
`
`accelerant treatment) — |
`r
`(Maximum enhancement ratio,
`stratum corneum removed) — 1
`
`* 100
`
`(7)
`
`The enhancement indices of the terpenes as
`calculated towards the model drugsare listed in
`Table 4. These values show that,
`for example,
`while the enhancement ratio for 1,8-cineole to-
`wards 5-FU is large,
`in terms of the maximum
`achievable enhancement
`the terpene activity is
`low, Elfigsoo = 1.1%. Contrariwise,
`for oestra-
`diol, although the ER is apparently low, E1P22? =
`8.3%, i.e., the terpene shows 8 times more activ-
`ity towards the lipophilic drug as compared with
`the polar drug based on assessments of the maxi-
`mum achievable effect. Table 4 illustrates that
`the hydrocarbon and oxide terpenes are more
`active,
`in terms of the maximum possible en-
`hancement,
`towards ES than towards 5-FU,
`whereas the alcohol and ketone terpenes are
`
`

`

`166
`
`TABLE 4
`
`(E1) of some
`(ER) and indices
`ratios
`The enhancement
`monoterpenes,
`towards a model polar penetrant, 5-fluorouracil
`(log P = — 0.89) and a model
`lipophilic penetrant, oestradiol
`(log P = 2.3), El values quoted to 2 significant figures
`
`
`Terpene
`
`Oestradiol
`5-Fluorouracil
`ER Elai
`ER
`Elias
`(%)
`(%)
`
`
`94
`14
`20.0
`12.2
`
`21.2
`27.7
`37.9
`2.4
`13.7)
`1.0
`30.9
`82.5
`
`0.10
`(h11
`0.23
`.13
`
`0.24
`0.32
`(44
`0.017
`O15
`012
`0.35
`0.97
`
`0.33 *
`0.45 *
`0.42 *
`O.10 4
`
`0.34 *
`O.17 “
`(1.36 *
`1.42
`1.90
`1.61
`-
`4,75
`
`—1.6
`—13
`— td
`-2.2
`
`—1.6
`2.0
`— 1.6
`1.0
`2.2
`1.5
`-
`9.1
`
`a-Terpineol
`Terpinen-4-ol
`Carveol
`Carvone
`
`Pulegone
`Piperitone
`Menthone
`Cyclohexene oxide
`«-Pinene oxide
`Limonene oxide
`Cyclopentene oxide
`Ascaridole
`7-Oxabicyclo-
`[2.2.1 Jheptane
`1.8-Cineole
`
`molecule may achieve through skin after all the
`stratum corneum resistance is removed. This is
`illustrated in Fig. 4 which compares the enhance-
`ment ratios and enhancement indices ofthe ter-
`penes towards the two drugs. An apparently ef-
`fective penetration enhancer, increasing the drug
`permeability coefficient markedly, may actually
`be poor
`in
`relation to the maximum values
`achievable, although the marked increase in drug
`
`a-Pinene 12©0.0024 3.09 5.1
`
`
`flux may still prove to be clinically acceptable.
`Limonene
`2.1
`0.013
`3.75
`6.7
`Thus,
`the ER provides a practical measure of
`3-Carene
`3.2
`0,026
`4.36
`8.2
`enhancement while the EI informs the investiga-
`tor how close he has approached the limit of
`complete
`chemical
`removal of
`the
`stratum
`corneum barrier.
`Another way of looking at the phenomenon is
`simply that hydrophilic drugs have great potential
`for enhancement because their permeability coef-
`ficients are low: lipophilic drugs have much less
`potential because their unmodified coefficients
`already approach a maximum value (Flynn and
`Stewart, 1988).
`The expression of penetration enhancement by
`the enhancement index implies the assumptions:
`(1)
`that
`the accelerant acts only to alter the
`barrier nature of the stratum corneum; (2) that
`the stratum corneum is the rate limiting barrier
`to drug diffusion; and (3) that the vehicle does
`not alter the solvent nature of nucleate epider-
`mal / dermal tissue in tape stripped skin.
`This third assumption merits further consider-
`ation. When aqueous vehicles are employed in
`permeation studies, the maximum drug flux may
`be obtained through stripped full
`thickness skin
`(stratum corneum removed) as described in this
`report; assuming that the nucleate epidermis and
`dermis are porous hydrogels, then no appreciable
`changes in the solvent nature of the skin samples
`should occur on removal of the stratum corneum
`and on direct contact with aqueous donor. Simi-
`larly, for non-aqueous vehicles which are essen-
`tially immiscible with dermal tissue (and water),
`for example liquid paraffin, permeant diffusion
`through skin with the stratum corneum removed
`should also be capable of experimental determi-
`nation. Howeve