from two topical formulations, in vivo
Ingo Alberti1,2, Yogeshvar N. Kalia1,2, Aarti Naik1,2, Richard H. Guy1,2,3
1Centre Interuniversitaire de Recherche et d'Enseignement "Pharmapeptides", Campus universitaire, F-74166 Archamps (France); 2Laboratoire de Pharmacie Galénique, Section de Pharmacie, Université de Genève, Quai Ernest-Ansermet 30, CH-1211 Genève 4 (Switzerland); 3To whom corre-spondence should be addressed.
Abstract
Purpose: Skin occlusion by an impermeable dressing is a common method to increase the penetration and cutaneous bioavailability of a topical drug. It is generally believed that hydration is responsible for this effect. The aim of this study was to ascertain if normal "clinical" occlusion (i.e., during a reasonable time of 6 hours) was an adequate strategy to increase the cutaneous bioavailability of topically applied products.
Methods: Skin areas of 22.5 cm2 located on the ventral forearms of six volunteers were treated either with an emollient ointment or a gel in open (exposed to environmental conditions) or closed (covered by an impermeable dressing) condi-tions. Each formulation contained 1% (w/w) of a lipophilic model drug. After 6 hours, the formulation was removed and the penetrated drug was sampled by sequential tape stripping of the stratum corneum (SC). Extraction and HPLC analysis of the drug in the tapes permitted a concentration-depth profile to be obtained. Fitting of the experimental data to an adequate solution of Fick's second law of diffusion, as well as integration of the latter allowed estimation of the partitioning, diffusivity, and reservoir of the drug. Transepidermal water loss (TEWL) measurements also enabled as-sessment of the SC water diffusivity and apparent membrane thickness.
Results: A single dose treatment for 6 hours under occlusion did not unequivocally improve cutaneous bioavailability of the lipophilic model drug. Occlusion of the ointment did not improve any of the parameters mentioned above, whereas the occluded gel showed only an improved partitioning. However, an open application of the ointment resulted in a better partitioning and cutaneous bioavailability of the drug compared to the gel, suggesting a higher impact of the formulation than the occlusion.
Conclusions: Occlusion will probably exhibit its effect with longer exposures, but vehicle effects are more precocious.
Consequently, an appropriate formulation strategy is recommended as an alternative to occlusion, and is certainly more efficient and cosmetically appealing than an occlusive dressing. The methodology described offers a powerful tool to analyze the penetration behaviour of drugs from topical cutaneous formulations and their interaction with human hy-drated skin in vivo. It may help to provide some insight into the mode of action of delivery formulations, which is highly relevant for the design of an effective dermatological product.
Keywords: Cutaneous bioavailability; Drug concentration profile; Tape stripping; Occlusion
Introduction
he effect of hydration on stratum corneum (SC) permeability has been widely documented, and extensively reviewed by several authors [1-3].
The mechanism by which such an enhance-ment is induced is still unclear, although
at-tempts have been made to elucidate the possi-ble implication of the intercellular lipids [4]
and the structural keratins [5].
Occlusion of the treated skin is a common nonphysiological method to increase SC and, therefore, to enhance local drug cutaneous bioavailability. Generally, this may be
T
achieved either by applying an impermeable dressing or an emollient formulation (e.g., containing petrolatum) to the skin. In both cases, the effect is similar: the barrier reduces or cuts off transepidermal water loss (TEWL), thus leading to a build up of water in the SC.
The extent of hydration by occlusion de-pends mainly on the water vapour transmis-sion of the covering film (dressing or emol-lient) [6] and the time of application [7].
However, in vivo occlusion effects on human SC permeability are not spectacular, unless long-term occlusion is applied. In contrast, hairless mouse skin exhibits a greater perme-ability when occluded [5], so that those in vitro findings must be cautiously interpreted prior to extending them to in vivo human skin.
Occlusion by an impermeable or semi-permeable dressing is commonly used either alone or in combination with topical steroids for the treatment of local skin disorders such as psoriasis. In this condition, occlusion not only increases drug penetration, but also de-creases the mitotic rate of hyperproliferative skin [8]. However, for more extensive skin diseases like atopic or contact dermatitis, an occlusive dressing treatment is not conven-ient, and occlusion by an emollient is prefer-able.
When a new formulation is developed for these inflammatory skin diseases, it would be valuable to predict the effect of occlusion on the cutaneous bioavailability of the drug,
since characterisation of the optimal formula-tion and applicaformula-tion condiformula-tions are essential starting points before commencing further clinical trials. Often, promising drugs are claimed insufficiently efficient because their skin penetration revealed to be too low de-spite excellent in vitro pharmacological activ-ity. Consequently, there is an obvious need for objective and quantitative assessment of their cutaneous bioavailability.
In vivo methods for measuring the effect of occlusion on cutaneous bioavailability are well described only for topical corticoids, since the latter exert a measurable, in vivo skin blanching effect due to vasoconstriction of the dermal capillaries [9]. Generally, a minimum occlusion time of 3 hours is re-quired for skin blanching, and a maximum is observed after a 10 hour occlusion period.
For topically applied drugs which do not exert a measurable in vivo effect, it is more appropriate to measure the drug concentration in the skin, and particularly in the SC. The latter can be the target site of the treatment (e.g., antifungals), but is also in direct contact with the underlying target sites (viable epi-dermis or epi-dermis) of drugs such as topical corticoids or anti-acneics.
Measurement of the drug concentration may even provide a suitable and more objec-tive alternaobjec-tive method to pharmacodynamic assessment, since recent evidence suggests that skin concentration is well correlated to a measure of drug effect in the skin [10,11].
In the present study, attention was turned to the investigation of the effect of occlusion through use of either a dressing or an emol-lient on the cutaneous bioavailability of a lipophilic model drug. Furthermore, we ad-dressed the issue whether "clinical" occlusion (i.e., during a reasonable time of 6 hours) was an adequate strategy to increase the cutaneous bioavailability of topically applied products.
Two formulations (an ointment and a gel) were investigated, each of which was applied for a period of 6 hours under either open (ex-posed to atmospheric conditions) or closed (covered with a dressing) conditions. The level of SC hydration was monitored by TEWL analysis, which provides a measure-ment of water diffusivity, and consequently offers an indirect means to evaluate water content within the membrane. The cutaneous bioavailability of the model drug was de-temined by assaying the drug content within the layers of SC removed. This was achieved by serial tape stripping of the SC and subse-quent extraction and chromatographic analy-sis of the permeant.
Materials and methods
Chemicals
The occlusive formulation was a petrola-tum based ointment, whereas the non-occlusive formulation was a silicium dioxyde based gel containing isopropyl myristate and
soya lecithin. The large (810 Da) and lipo-philic (log Ko/w ~2.8) drug was incorporated at 1% (w/w) in both topical formulations.
For the extraction of the model drug from the tapes and the HPLC analysis, all solvents were of HPLC grade: ultra-pure deionized water, acetonitrile (Sigma-Aldrich, Steinheim, Germany), and t-butylmethyl ether (Sigma-Aldrich, Steinheim, Germany). Phosphoric acid, incorporated in the buffer, was also pur-chased from Sigma-Aldrich.
Treatment protocol
Six healthy caucasian volunteers (5 fe-males, 1 male, 24-34 years, mean 29 years) with no history of dermatological disease were selected for this study. Written consent was obtained from all subjects, after approval by the local university hospital's ethical com-mitee. The treatment sites, covering individ-ual skin areas of 22.5 cm2 (9.0x2.5 cm), were located on the volar aspects of the forearms (which were non-hairy in all cases), 6 cm from the wrists.
Each formulation was applied for 6 hours in open (exposed to the atmospheric condi-tions but physically protected) or closed (cov-ered by an occlusive dressing) condition, in contralateral positions.
Two sets of experiments were carried out:
the first one concerned the ointment and in-cluded subjects A, B, C, D, and E (n=5), whereas in the second set the gel was tested, and included subjects A, C, D, E and F (n=5).
A single dose of 45 mg of formulation (corresponding to 2 mg/cm2) was applied to the skin. For the open application, the treated area was protected by a sheet of cellulose tissue, which was affixed by an adhesive elas-tic and porous dressing (Medipore®, 3M, St.
Louis, MN). For the closed situation, an inert occlusive polyester film (Scotchpak®, 3M, St.
Louis, MN) was applied on the skin, and fixed by an adhesive elastic polyurethane film (Opsite®, Smith-Nephew, Hull, UK).
After 6 hours, the patch was removed and the excess formulation gently blotted with three dry cellulose swabs, without using any solvent (which could potentially affect the percutaneous penetration of the drug).
SC Sampling protocol
Drug cutaneous bioavailability was as-sessed by measuring the drug concentration in the SC, which was isolated by sequentially removing microscopic layers (0.5-1 µm) of tissue carrying the drug. This procedure is generally recognized as non-invasive and painless, and was easily carried out by simply applying a commercially available contact-sensitive adhesive tape (polypropylene back-ing and acrylic adhesive, Scotch® Book Tape, 3M, St. Louis, MN) on the treated site and subsequently removing it, after rubbing six times with a metallic spatula. In order to rig-orously remove the same constant area, the treated site was delimited by a template, ex-posing only a window of 9.0 x 2.5 cm of skin.
The first tape was not discarded, although some authors consider that it only contains non-penetrated drug. Up to 20 SC layers were removed.
Each tape was carefully weighed before and after the procedure with a 10-µg precision balance (Mettler AT 261, Greifensee, Swit-zerland), under controlled conditions of tem-perature (19±1°C) and relative humidity (34±2 %). The drug levels were normalized with respect to both the removed volume and the SC depth. The former was obtained as-suming that the SC average specific gravity was 1 g/cm3 [12], and the latter, from TEWL measurements, using a method described elsewhere [13].
Transepidermal water loss (TEWL) measurements
Water vapour flux through the skin, as-sessed as TEWL, is an important parameter when studying the effects of occlusion. In this study, it allowed us to determine the mean water diffusivity within the SC, as well as the apparent SC thickness. In addition, since wa-ter vapour flux is an indicator of barrier func-tion disrupfunc-tion by tape stripping [14], it per-mitted an assessment of barrier efficiency and hence, cessation of the stripping procedure at an appropriate level.
A TEWL reading was taken every 4 tape strips after a 1 minute stabilization, from the center of the treated site, by the means of an Evaporimeter (model EP1, Servomed,
Stock-holm, Sweden), according to the guidelines from the standardization group of contact dermatitis [15]. During each session, the sub-jects were placed in an air conditioned room at 21-22°C, and rested for 30 min, in order to avoid sweating, which obviously interferes with the TEWL [16]. Their skin temperature was recorded with a tele-thermometer (model 44, YSI, Yellow Springs, OH), and ranged between 29 and 31°C. Ambient relative hu-midity was also monitored but not controlled, and ranged between 24 and 32% (experiments took place in western Europe between the months of January and March).
Extraction and HPLC analysis of the drug from the tape strips
Absorbed drug was extracted from the in-dividual tape strips (paired above the 10th strip) by immersing in 7 mL of acetonitrile in polypropylene capped tubes, for a period of 1 hour under agitation. After filtration of the supernatant through a 0.45 µm nylon mem-brane (Nalgene, Nalge, Rochester, NY), the filtrate was submitted to HPLC analysis.
Validation of the extraction was carried out by spiking stripped blank tapes with 100 µL of a 0.1 mg/mL solution of neat drug, corsponding to 2.2% of the applied dose. A re-covery of 78.0±3.8% (n=5) was attained.
The HPLC system consisted of a Waters-Millipore (Milford,MA) pump 600, autosam-pler 717 Plus, and UV detector 486 set at 210 nm. A reversed phase C-18 AB,
125x4 mm (Macherey-Nagel, Düren, Ger-many), column mantained at 60°C was used for the analysis. The mobile phase comprised water, acetonitrile, t-butylmethyl ether, and phosphoric acid at volume fractions 600:300:70:0.2 (system A) or 200:660:70:0.2 (system B). The system solvents were mixed at ratios of 0% A + 100% B from 9 to 15 minutes run time, and 35% A + 65% B the remaining time. The flow rate was set at 1 mL/min, and a sample injection of 50 µL yielded a retention time of about 4 minutes.
Peak recording and data processing were performed with the built-in system manager (Millennium, version 2.0). The drug amounts were determined using the area-under-the-curve (AUC) method from calibration plots generated with the neat compound. The detec-tion limit was 25 ng on column.
Model of concentration-depth profile
The concentration-depth profile was fitted to a model (equation 1) obtained from a solu-tion of Fick's second law of diffusion [17], assuming that (i) dermal capillaries act as a sink, (ii) binding or metabolism of the drug are negligible, (iii) the drug concentration in the vehicle at the surface remains constant throughout the application time, and (iv) the vehicle does not modify the membrane's per-meability.
This expression predicts the instantaneous membrane concentration (Cx) of a drug dis-solved in a vehicle at a given concentration (Cv), at a given time (t) and at all relative po-sitions (x/L) within the SC. The parameters of interest are the SC-vehicle partitioning cient (K) and the apparent diffusion coeffi-cient of the drug (D/L2).
Data obtained from HPLC analysis were fitted to equation (1) by the means of a non-linear fitting software package (Grafit 3.03, Erithacus).
SC cutaneous bioavailability
This parameter was assessed as SC drug reservoir, namely the area under the concen-tration-depth profile of the drug. The latter was obtained by solution of the integrated form of equation (1) from the relative value of 0 to 1 (i.e., over the entire SC thickness), yielding equation (2):
0 1
x v 2 2
C d(x
L)=K C (1 2 - 4
(e xp(-D
L ) e xp(-9D
L )))
π2 π2 1 π2
t+9 t
Equation (2)
This approach allows extrapolation of drug levels in the lower layers of SC, even where stripping was not performed.
Results and discussion
Effect of occlusion on SC hydration
Data from TEWL measurements during sequential tape stripping allowed evaluation of the water diffusivity (Dw) and the apparent thickness of the SC (L), reported in Table I.
Water diffusivity is an indirect marker of the hydration (water content) of SC, since it has been shown that diffusing water encoun-ters lower resistance as the water content in-creases [18]. Furthermore, it is apparent from Fick's law of diffusion that increased water flux (TEWL) is provoked by an increased water concentration gradient. However, in the experiment described occlusion did not ap-pear to affect water diffusivity, since in all situations (occlusion by a dressing or by the emollient ointment formulation) there was no statistically significant difference in the Dw
values (Table I). This outcome suggests that 6 hours of occlusion are not enough to ensure a sufficient accumulation of water in the SC.
Although longer periods of occlusion may be expected to elevate SC hydration, it is unlikely to enhance permeation as a result of depletion of the applied dose. That is, a multi-dose regimen is more likely (than extended occlusion) to improve cutaneous bioavailabil-ity.
The second marker of hydration, the SC thickness, also gives an idea of the effect of occlusion by the swelling of the SC struc-tures. This effect was observed and modeled, among others, by Blank and co-workers [18].
In our study (Table I), the biomembrane thickness assessed by TEWL varies between 8.7 and 14.0 µm, which agrees well with the literature [19]. However, no significant (p>0.05) swelling was observed, either due to occlusion or to formulation constituents, as compared by the Student t-test on paired val-ues.
Effect of occlusion on drug penetration and cutaneous bioavailability
Concentration-depth data obtained from HPLC analysis were fitted to the model de-scribed in equation 1. The curves of best fit to the experimental data are presented in Figure 1 (ointment data) and Figure 2 (gel data). The areas under these curves were calculated
us-ing equation 2, thus yieldus-ing an estimate of the local cutaneous bioavailability of the drug.
All parameters obtained from the curves (K = SC/vehicle partitioning, D/L2 = drug diffusiv-ity, AUC = cutaneous bioavailabildiffusiv-ity, and P = SC permeability coefficient of the drug) are reported in Table I. Statistical analysis was performed by the means of a paired Student t-test, with significance level fixed at p<0.05.
Table I: Data obtained from TEWL measurements (SC thickness, L; water diffusivity, Dw), and from HPLC analysis for the drug (K = SC/vehicle partitioning, D/L2 = diffusivity in SC, AUC = SC reservoir). SC perme-ability coefficient for the drug (P) was obtained from Fick's first law of diffusion: P = K*D/L = K*D/L2*L (note that D/L2 and L are obtained from HPLC and from TEWL data, respectively). Each individual value indicated the mean±SD (n=5, except for the gel, where n=4 for drug diffusivity, SC reservoir and permeabil-ity). Statistical analysis was performed using the paired Student t-test (Sigmastat version 1.0 software Jan-del), and level of significance was fixed at p<0.05.
Formulation Application L
(µm) Dw *109
(cm2/s)
K (dim.less)
D/L2 *106 (s-1)
P*109 (cm/s)
AUC*103 (M) Ointment Occluded 10.12±2.69 3.24±1.50 1.67±0.43b 3.29±1.85 6.01±4.91 6.05±2.90 Ointment Open 8.69±3.53 2.17±0.71 2.34±0.59c 3.75±2.88 7.25±6.84 7.55±3.17d Gel Occluded 14.03±4.32 3.84±1.52 1.04±0.30a,b 1.18±0.59 1.67±0.87 2.38±0.98 Gel Open 12.15±3.02 2.52±1.41 0.42±0.26a,c 1.79±2.48 1.12±1.39 1.07±1.01d
a,b,c,d
Sudent t-test indicates that those paired values are significantly different (p<0.05)
Figure 1: Fitted concentration-depth profiles of the model drug in human SC in vivo, released from the ointment after 6 hours treatment under occlusive dressing (A) or exposed to atmospheric conditions (B). Individual data points are not shown, for sake of clarity.
1 0.8
0.6 0.4
0.2 0.05
0.04
0.03
0.02
0.01
0
Relative SC depth
[Drug] in SC (M)
Subject A Subject B Subject C Subject D Subject E
Ointment - occluded (A)
1 0.8
0.6 0.4
0.2 0
0.05
0.04
0.03
0.02
0.01
0
Relative SC depth
[Drug] in SC (M)
Subject A Subject B Subject C
Subject D Subject E
Ointment - open (B)
Figure 2: Fitted concentration-depth profiles of the model drug in human SC in vivo, released from the gel after 6 hours treatment under occlusive dressing (A) or exposed to atmospheric conditions (B). Individual data points are not shown, for sake of clarity.
1 0.8
0.6 0.4
0.2 0
0.05
0.04
0.03
0.02
0.01
0
Relative SC depth
[Drug] in SC (M)
Subject A Subject C Subject D Subject E Subject F
Gel - occluded (A)
1 0.8
0.6 0.4
0.2 0
0.05
0.04
0.03
0.02
0.01
0
Relative SC depth
[Drug] in SC (M)
Subject A Subject C Subject D Subject E Subject F
Gel - open (B)
In the case of the ointment, an open appli-cation results in an apparently better penetra-tion of the drug for all subjects (Figure 1B) compared to the closed protocol (Figure 1A).
However, values of diffusivity of the drug from Table I show no significant difference.
This indicates that the SC barrier efficacy was not reduced by occlusion, as additionally confirmed by the unltered water diffusivity.
On the other side, Figure 1B suggests that the drug exhibits a better solubility in the SC for most subjects (namely, A, B, and C) in the open situation, since partitioning is slightly higher than in the closed application (Table I), suggesting a better partitioning potential for this formulation in normal than overhydrated SC [9].
On the whole, permeability of SC for the drug released from the ointment is not in-creased after 6 hours occlusion by a dressing, and even in the open application of the oint-ment the occlusivity of petrolatum is not suf-ficient to increase water diffusivity in com-parison to the open gel (Table I), but the 5-fold better partitioning of the ointment (2.34 versus 0.42) suggests that the 7-fold increase in the cutaneous bioavailability (7.55 versus 1.07) is to be attributed to the release proper-ties of the formulation. This is an important information for the rational design of the for-mulation, since modification of the physico-chemical properties of a formulation can lead to an efficacy of the product better than that obtained by an occlusive dressing.
Effect of occlusion on the skin penetration of the drug from the gel is quite different than from the ointment. Actually, closed applica-tion (figure 2A) of the gel generated slightly higher drug levels than during the open one (figure 2B). This effect is not due to the in-creased hydration, since water diffusivity is
Effect of occlusion on the skin penetration of the drug from the gel is quite different than from the ointment. Actually, closed applica-tion (figure 2A) of the gel generated slightly higher drug levels than during the open one (figure 2B). This effect is not due to the in-creased hydration, since water diffusivity is