Assessment of skin absorption and irritation potential of arachidonic acid and glyceryl arachidonate using in vitro diffusion cell techniques.

Arachidonic acid (AA), a precursor of pro-inflammatory mediators, and its glycerin ester, glyceryl arachidonate (GA), are reportedly used in cosmetic products. In vitro skin penetration of AA and GA and GA's ester hydrolysis was determined in flow-through diffusion cells. AA penetration with human and rat skin was 19.5% and 52.3% of the applied dose respectively, a substantial amount of which remained in the skin at 24h. Similar penetration results were obtained with GA in human skin. However, GA penetration through cultured skin (EpiDerm) was 51% of the applied dose, almost all of which appeared in the receptor fluid. At least 27.8% of GA penetrating skin was hydrolyzed to AA. In vitro methods were used to assess skin irritation in diffusion cells. Skin irritation of AA, sodium lauryl sulfate (SLS), and Tween 80 was determined by changes in transepidermal water loss (TEWL), skin viability (3-(4,5-dimethylthiaxol-2-yl)-2,5-diphenyltetrazolium bromide, MTT, formation), and cytokine release (IL-1alpha). SLS irritation was much less pronounced in an emulsion versus an aqueous vehicle. No significant irritation was observed in vitro from AA in an emulsion. This work predicts that AA would penetrate human skin in vivo and that it could be formed in skin from topically applied GA.

In vitro human skin penetration of diethanolamine.

Concerns about the safety of diethanolamine (DEA) have been raised by the National Toxicology Program (NTP). Therefore, we measured the extent of DEA absorption in human skin relevant to exposures from shampoos, hair dyes and body lotions. Radiolabeled [14C]-DEA was added to two commercial products from each class and applied to excised viable and non-viable human skin in flow-through diffusion cells. The products remained on the skin for 5, 30 and 24 h for shampoos, hair dyes and body lotions, respectively. After 24 h, most of the absorbed dose was found in skin: 2.8% for shampoos, 2.9% for hair dyes and 10.0% for body lotions. Only small amounts were absorbed into the receptor fluid: 0.08%, 0.09% and 0.9% for shampoos, hair dyes and body lotions respectively. There was no significant difference in the absorption of DEA through viable and non-viable skin or from product application doses of 1, 2 or 3 mg lotion/cm2. In 72 h daily repeat dose studies with a lotion, DEA appeared to accumulate in the skin (29.2%) with little diffusing out into the receptor fluid. Therefore, skin levels of DEA should not be included in estimates of systemic absorption used in exposure assessments.

In vitro and in vivo percutaneous absorption of catechol.

The Cosmetic Ingredient Review Expert Panel found insufficient data to conclude that catechol could be used safely in permanent hair dye products. Information was lacking on the extent of oxidation and skin absorption of remaining catechol. In vitro percutaneous absorption studies were conducted in human and rat skin using a consumer permanent hair dye spiked with 0.6% catechol. A 30-min application demonstrated 0.4% of the applied dose was absorbed through human skin and 0.2% through rat skin. The minimal absorption observed was due to the short exposure time and to partial oxidation of catechol by the dye developer. The fate of catechol remaining in rat skin after exposure in vitro and in vivo was investigated with additional absorption studies using catechol in ethanol. At 72 h, 24-h application of 4% catechol resulted in skin absorption of 81% of the applied dose in vitro and 53% in vivo. Skin levels measured at 24 h remained unchanged after 72 h. Therefore the skin reservoir did not contribute to the estimated systemic absorption. A deconvolution technique employed to predict skin absorption using plasma levels from intravenous and dermal administration overestimated in vivo skin absorption due to volatility of catechol in an ethanolic vehicle.

In vitro percutaneous absorption models.

In vitro skin absorption studies are commonly used to estimate in vivo skin absorption in topical safety and efficacy evaluations. In vitro studies are more economical and result in minimization or elimination of the use of animals.

Percutaneous penetration and metabolism of 2-nitro-p-phenylenediamine in human and fuzzy rat skin.

2-Nitro-p-phenylenediamine (2NPPD) is a dye used in semipermanent and permanent (tinting color) hair dye formulations. National Toxicology Program toxicology and carcinogenesis testing of 2NPPD has raised concerns about its safety. Therefore, we initiated in vitro studies to measure absorption and metabolism of 2NPPD in human and fuzzy rat skin and rat jejunal tissue. Intestinal tissue metabolism of 2NPPD was compared to skin metabolism since toxicology data from oral 2NPPD studies will be used for future safety assessment purposes. Absorption was measured over 24 h by using flow-through diffusion cells with a receptor fluid consisting of Hepes-buffered Hank's balanced salt solution. Dosing vehicles were applied to skin and intestine in the diffusion cells for 30 min. 2NPPD metabolites were determined by high-performance liquid chromatography methodology. In human skin, the percentages of total applied dose absorbed (receptor fluid + skin) over 24 h were 9.2 +/- 5.7 (mean +/- SD) and 9.5 +/- 3.2 for the ethanol and semipermanent vehicles, respectively, with approximately 3% remaining in skin. In rat skin, the percentages of total applied dose absorbed over 24 h were 9.3 +/- 1.2 (mean +/- SE), 6.9 +/- 1.2, and 4.2 +/- 0.1 for the ethanol, semipermanent, and permanent formulation vehicles, respectively, with approximately 3% remaining in skin. In rat intestinal tissue, the percentage of total applied dose absorbed over 24 h was 10.9 +/- 1.2, with approximately 5% remaining in the tissue. In human and rat skin, 2NPPD was metabolized to triaminobenzene and N4-acetyl-2NPPD. 2NPPD was also metabolized to a sulfated 2NPPD metabolite in rat skin, but not in human skin. 2NPPD was extensively metabolized in both human and rat skin with ethanol application; metabolism was not as extensive with a semipermanent formulation application. In rat intestinal tissue, 62% of 2NPPD was metabolized upon absorption to triaminobenzene and N4-acetyl-2NPPD. Differences in the metabolic profiles (proportion of each metabolite formed) were found between the skin and intestinal tissue. These results suggest that 2NPPD is rapidly absorbed and extensively metabolized in both skin and intestinal tissue. The extent of metabolism and the metabolic profile were found to be species-, tissue-, and dosing vehicle-dependent. Metabolism information will be useful in predicting the extent of 2NPPD and/or 2NPPD metabolite systemic absorption relative to a dermal exposure, which will improve the health hazard assessment of 2NPPD.

The effects of an alpha hydroxy acid (glycolic acid) on hairless guinea pig skin permeability.

The barrier integrity of hairless guinea pig skin after treatment with an alpha hydroxy acid was assessed through in vivo topical application of an oil-in-water emulsion containing 5 or 10% glycolic acid at pH 3.0. The control was a commercial moisturizing lotion, pH 7.8. A dosing regimen for the glycolic acid formulations that was tolerated by the hairless guinea pigs and significantly decreased stratum corneum turnover time was determined using the dansyl chloride staining technique. Once-daily dosing of hairless guinea pig skin for 3 weeks with the glycolic acid formulations resulted in approximately a 36-39% decrease in stratum corneum turnover time compared with the control lotion. After this treatment, hairless guinea pigs were sacrificed for the in vitro measurement of the percutaneous absorption of [14C]hydroquinone and [14C]musk xylol. No significant differences in the 24-hour absorption of either test compound were found for skin treated with the control lotion or the glycolic acid formulations. There were also no significant differences found in the absorption of [3H]water through skin from the different treatment groups. Although no increase in skin penetration occurred after treatment with the glycolic acid formulations, histology revealed approximately a twofold increase in epidermal thickness. Also the number of nucleated cell layers nearly doubled in skin treated with 5% and 10% glycolic acid compared with the control lotion and untreated skin. These studies demonstrate that substantial changes in the structure of hairless guinea pig epidermis can occur without significant effect on skin permeability of two model compounds.

Percutaneous absorption of trinitrobenzene: animal models for human skin.

The percutaneous absorption of 1,3,5-trinitrobenzene (TNB) was studied in viable skin from hairless guinea pigs (HGP), Fischer 344 rats and humans. Skin was dermatomed and assembled in flow-through diffusion cells followed by TNB application in either an acetone or a water vehicle. Skin absorption was expressed as the percentage of applied dose absorbed into skin and receptor fluid within 24 h. Rapid absorption of TNB by rodent skin was obtained with both vehicles. For HGP skin, TNB absorption was 72.7+/-5.5% in the acetone vehicle and 82.3+/-4.5% in the water vehicle. For rat skin, TNB absorption was 61.0+/-4.1% (acetone) and 66.5+/-4.1% (water). Absorption of TNB from acetone was significantly reduced (38.0+/-11.0%, P = 0.0118) in human skin, but absorption from water remained high (75.5+/-10.8%). Little TNB remained in skin when a thin (200 microm) dermatome section was used (HGP and human skin). A thicker dermatome section was required (350 microm) with haired rat skin, and 13-21% of the absorbed radioactivity remained in the skin at 24 h. Rodent skin did not simulate satisfactorily the barrier properties of human skin when TNB absorption was reduced by application in a volatile solvent.

Percutaneous absorption and metabolism of Coumarin in human and rat skin.

Coumarin is widely used as a fragrance in cosmetics, perfumes and soaps. The food and Drug Administration banned coumarin use in food because of reports that coumarin produced hepatotoxicity in rodents. Concerns about coumarin's safety have also been raised by toxicity testing conducted by the National Toxicology Program. Therefore, we initiated studies to measure the extent of coumarin absorption and metabolism in skin. [14C]Coumarin (ca. 0.5 microCi per cell) absorption in skin was measured by using two vehicles: ethanol (15 microliters cm-2) and an oil-in-water emulsion (3 mg cm-2). Absorption was determined for 24 h by using flow-through diffusion cells (0.64 cm2, exposed skin) with a receptor fluid consisting of HEPES-buffered Hank's balanced salt solution (pH 7.4). Coumarin metabolism was determined by high-performance liquid chromatography methodology. In rat skin (n = 3), the percentages of applied dose absorbed after 24 h were 54.9 +/- 0.63 (mean +/- SEM) and 86.8 +/- 5.4 for the ethanol and emulsion vehicles, respectively, with ca. 5% remaining in skin. In human skin (n = 2), the percentages of applied dose absorbed after 24 h were 64.4 +/- 0.29 and 98.0 +/- 5.3 for the ethanol and emulsion vehicles, respectively, with ca. 1% remaining in skin. The extent of skin absorption was greater from the emulsion vehicle than from the ethanol vehicle in both human and rat skin. Coumarin rapidly penetrated both rat and human skin with > 75% and > 95%, respectively, of the absorbed dose found in the receptor fluid within 6 h. No evidence of coumarin metabolism was found in either skin or receptor fluid fractions. These studies indicate that coumarin absorption is significant in skin. Systemic coumarin absorption must be expected after dermal contact with coumarin-containing products.

Use of in vitro skin penetration data and a physiologically based model to predict in vivo blood levels of benzoic acid.

A physiologically based pharmacokinetic (PB PK) model was developed to predict plasma levels of benzoic acid (BA) in the hairless guinea pig after topical exposure at three finite dose levels. The PB PK model consisted of four compartments: (1) rapidly perfused tissues; (2) slowly perfused tissues; (3) liver, representing the route of elimination of BA from the plasma after biotransformation to hippuric acid; and (4) plasma. The predictive capacity of the PB PK model was assessed by comparing plasma BA levels measured experimentally with those predicted by the model. The percutaneous absorption of finite doses of BA in the model was described by a transdermal input function, which was derived from in vitro percutaneous absorption studies in which viable hairless guinea pig skin in flow-through diffusion cells was exposed to BA. Physiological parameters used in the model were calculated from previously published values. Biochemical parameters, including partition coefficients and metabolic constants, were measured experimentally in vitro. The PB PK model predictions were generally in good agreement with measured plasma levels for each of the dose levels studied. The predicted plasma BA levels and the measured values were closer for the highest dose (120 microg/cm2) than for either of the other two doses used (12 and 40 microg/cm2). The effects of optimizing the metabolic constants and the transdermal input function parameters on the predicted curve shape and fit to that of the measured plasma BA levels were assessed. Varying the transdermal input parameters produced closer agreement between predicted and measured values.

In vitro percutaneous absorption of the fragrance ingredient musk xylol.

The percutaneous absorption of the fragrance fixative musk xylol was measured in vitro in human and hairless guinea pig skin. For comparison, musk xylol was applied to skin in an oil-in-water emulsion or the volatile solvent, methanol. After 24 hr, total absorption of musk xylol in hairless guinea pig skin was 55% from the emulsion vehicle and 45% from the methanol vehicle. With human skin, permeation of musk xylol from both vehicles decreased to 22% of the applied dose. When human studies were continued for an additional 6 days after skin surface washing, only 6% of the applied dose remained in skin. The data suggest that most of the absorbed musk xylol in skin at 24 hr will be systemically absorbed in vivo within 1 wk. Throughout the 24-hr absorption study, absorbed musk xylol was not metabolized. A permeability constant for musk xylol permeation through hairless guinea pig skin was determined by a modified procedure for the lipophilic compound. At each time point, some diffusion cells were terminated so that skin and receptor fluid levels could be determined. Under steady-state absorption the permeability constant was 6.86 x 10(-5) cm/hr. The amount of musk xylol penetrating skin from three types of cosmetic products was also calculated on the basis of actual conditions of use. Products that are applied to large areas of the body and remain on the skin for long periods will result in the greatest absorption of musk xylol.

Metabolism of benzocaine during percutaneous absorption in the hairless guinea pig: acetylbenzocaine formation and activity.

The effect of dose and enzymatic inhibition on the percutaneous absorption and metabolism of benzocaine was studied in vitro in the hairless guinea pig. At the dose level of 2 micrograms/cm2, benzocaine was rapidly absorbed and extensively metabolized (80%) by acetyltransferase. As the applied dose of benzocaine was increased to 40 and 200 micrograms/cm2, N-acetylation of benzocaine decreased to 44 and 34%, respectively, suggesting saturation of the acetyltransferase system. Total 14C absorption after benzocaine application was not significantly different between control and enzyme-inhibited skin and therefore does not appear to be affected by the extent of benzocaine metabolism during percutaneous penetration. Skin provides a significant first-pass metabolic effect for therapeutic doses of percutaneously absorbed benzocaine, and the primary metabolite formed, acetylbenzocaine, is biologically active.

Growing a stratified, cornified primary culture of rat keratinocytes with epidermis-like water permeation barrier function.

The culture of cutaneous keratinocytes grown on a Puropore nylon microporous membrane at the air-liquid interface has been shown to be similar to the epidermis in a number of molecular and morphologic characteristics but to exhibit a significantly greater degree of tritiated water permeation. Various culture conditions have been altered in an effort to improve the water barrier properties. A Kp value in the range of 5.5 +/- 1.6 x 10(3) has been obtained for 79% of the cultures a) by plating 0.9 x 10(6) viable basal cells on a piece (13-mm diameter) of membrane for 7 days of submerged growth, b) by placing two membranes on two stacked glass fiber filters (47-mm extra-thick) in a culture dish (60 mm) for 14 days of growth at the air-liquid interface, c) by replacing the growth medium, i.e., 1 ml of complete minimum essential medium (CMEM) every 24 h after lifting, d) by using 10% fetal bovine serum (FBS) in the CMEM during the submerged culture period and 15% FBS in the CMEM during the lifted culture period, and e) by adding a dialysis membrane on top and a Puropore nylon membrane below the culture when the cultures were inserted in the permeation cell for testing. The percentage of cultures with this value for Kp can be increased to 90% if only cultures with yellow, smooth, and shiny surfaces are tested. This system should be useful as a replacement for skin in testing the cutaneous permeation of some chemicals.

Influence of metabolism in skin on dosimetry after topical exposure.

Metabolism of chemicals occurs in skin and therefore should be taken into account when one determines topical exposure dose. Skin metabolism is difficult to measure in vivo because biological specimens may also contain metabolites from other tissues. Metabolism in skin during percutaneous absorption can be studied with viable skin in flow-through diffusion cells. Several compounds metabolized by microsomal enzymes in skin (benzo[a]pyrene and 7-ethoxycoumarin) penetrated human and hairless guinea pig skin predominantly unmetabolized. However, compounds containing a primary amino group (p-aminobenzoic acid, benzocaine, and azo color reduction products) were substrates for acetyltransferase activity in skin and were substantially metabolized during absorption. A physiologically based pharmacokinetic model has been developed with an input equation, allowing modeling after topical exposure. Plasma concentrations in the hairless guinea pig were accurately predicted for the model compound, benzoic acid, from in vitro absorption, metabolism, and other pharmacokinetic parameters.

Characterization of esterase and alcohol dehydrogenase activity in skin. Metabolism of retinyl palmitate to retinol (vitamin A) during percutaneous absorption.

Retinyl palmitate, a widely used ingredient in cosmetic products, is promoted for its beneficial effects on the appearance of skin. Previous studies suggest that enzymes are available in skin to metabolize this ingredient during skin absorption. Esterase activity hydrolyzes retinyl palmitate to retinol (vitamin A), which is oxidized in many tissues to retinoic acid primarily by alcohol dehydrogenase. The activities of esterase and alcohol dehydrogenase were characterized in hairless guinea pig skin by using flow-through diffusion cells and radiolabeled model compounds (methyl salicylate and benzyl alcohol) previously shown to be metabolized by these enzymes. Methyl salicylate was hydrolyzed by esterase to a greater extent in viable skin than in nonviable skin. Glycine conjugation of salicylic acid and benzoic acid occurred only in viable skin. The metabolism of methyl salicylate and benzyl alcohol occurred to a greater extent in male guinea pig skin than in female guinea pig skin. The percutaneous absorption of both radiolabeled compounds was similar in viable and nonviable skin. About 30 and 18% of topically applied retinyl palmitate were absorbed from an acetone vehicle by hairless guinea pig skin and human skin, respectively. Less than 1% of the applied dose of this lipophilic compound diffused from skin into the receptor fluid. Retinol was the only detectable metabolite of retinyl palmitate in both hairless guinea pig and human skin. In human skin, 44% of the absorbed retinyl palmitate was hydrolyzed to retinol. The use of retinyl palmitate in cosmetic formulations may result in significant delivery of retinol into the skin.

Dose-response relationship in skin sensitization.

The dose-response relationship (challenge phase) of the skin sensitization response was investigated in previously sensitized Hartley guinea pigs. Larger numbers of animals were used per group at the lower doses so that statistically significant observations could be made. Model compounds known to be skin sensitizers were used: a strong sensitizer, dinitrochlorobenzene (DNCB), and a weaker sensitizer, p-phenylenediamine (PPDA). A gradation in response to changing DNCB doses was easily observed by using either the open epicutaneous test (OET) or the Buehler occlusive patch test. The Buehler test was used to study the dose-response relationship of DNCB sensitization. The sensitivity of the OET and Buehler test was judged not adequate to measure the dose response for PPDA, because at high doses a high incidence of responders was not obtained. Therefore, the maximization test was used to evaluate PPDA. Similar, non-linear dose-response curves were obtained with each compound. The higher doses produced a somewhat linear relationship, but at lower doses the curves flattened out and more slowly approached a zero response. Thus, for potent sensitizers, concentrations found in exposure situations might be in the linear portion of the dose-response curve. For weak responders, use concentrations might be in the shallow portion of the curve, where reactions would be underestimated if a linear dose-response curve were assumed.

Reduction of azo dyes during in vitro percutaneous absorption.

The azoreduction of phenylazo-2-naphthol (Sudan I), 5-(phenylazo)-6-hydroxynaphthalene-2-sulfonic acid [aniline subsidiary color of FD&C Yellow No. 6 (ANSC)], and phenylazophenol [Solvent Yellow 7 (SY7)] in skin during percutaneous absorption was measured and the contributions of cytosolic and microsomal reductions were characterized. By using a series of azo dyes with a common U-14C-labeled phenylazo moiety, percutaneous absorption and metabolism were measured in vitro in flow-through diffusion cells with Sencar mouse, hairless guinea pig, and human skin. Azoreductase assays using subcellular fractions from skin of all species were used to examine intracellular rates and distribution for the series of dyes. The absorption of the lipophilic dyes Sudan I and SY7 from a finite surface dose of 5 micrograms/cm2 was extensive in all species. In mouse, 32.8 +/- 2.8% of the applied Sudan I dose and 64.1 +/- 3.3% of the applied SY7 dose were absorbed in 24 hr. In the hairless guinea pig, 57.6 +/- 5.9% of the applied Sudan I dose and 67.8 +/- 4.6% of the applied SY7 dose were absorbed in 24 hr. Human skin was least permeable, with 26.4 +/- 6.7% of the applied Sudan I dose and 36.1 +/- 4.5% of the applied SY7 dose absorbed in 24 hr. Sudan I and SY7 were extensively reduced in skin of all species during percutaneous absorption (29.5 and 26.5%, respectively, of the absorbed dose in human skin and greater than 50% of the applied dose in other species). ANSC was the least absorbed, with 5% or less penetrating. SY7 was preferentially reduced in the skin cytosol of all species, whereas Sudan I was equally reduced in the skin cytosol and microsomal fractions. The site of azoreduction in the cell may affect the metabolic fate of the liberated arylamine. The extensive azoreduction observed during percutaneous absorption may modulate the toxicities of these compounds and must be considered when effective doses are determined for quantitative risk assessments from dermal exposures.

In vitro release of nitroglycerin from topical products by use of artificial membranes.

An in vitro testing method for measuring the release of nitroglycerin from topical drug products was evaluated. The method involved measuring the amount of nitroglycerin that diffused from ointments and patches through various synthetic membranes into receptor fluid contained in a modified Franz diffusion cell. Plots of the amount of nitroglycerin released against the square root of time for all 10 synthetic membranes were linear. Five membranes (group I) were found to release nitroglycerin at similar rates (difference not significant; p > 0.05). Use of the other five membranes (group II) in diffusion cells resulted in release rates significantly slower than those of group I membranes (p < 0.05). Rapid permeation of nitroglycerin from a solution through a polysulfone membrane demonstrated a lack of significant diffusion barrier properties of this membrane. Rates of release of nitroglycerin from commercially available ointments were found to be similar. A comparison of three commercial nitroglycerin patches revealed that these products released nitroglycerin at different rates in vitro.

Percutaneous absorption and metabolism of pyrene, benzo[a]pyrene, and di(2-ethylhexyl) phthalate: comparison of in vitro and in vivo results in the hairless guinea pig.

The in vitro and in vivo absorption and metabolism of pyrene, benzo[a]pyrene, and di(2-ethylhexyl) phthalate (DEHP) were investigated in the hairless guinea pig. The in vitro method, which involved the use of flow-through diffusion cells and Hepes-buffered Hanks' balanced salt solution containing 4% bovine serum albumin as perfusate, was demonstrated to be a suitable system for predicting in vivo absorption of the above lipophilic compounds. The successful application of the in vitro technique for these compounds is significant because no satisfactory in vitro method has hitherto been developed to predict in vivo absorption of highly lipophilic chemicals. Quantification of parent compounds and metabolites that permeated into perfusates and those that remained in skin discs provided insight into the process by which the chemicals penetrated through the skin. Pyrene was absorbed primarily by a passive diffusion process, although a small fraction of the administered dose was biotransformed into metabolites in the skin and partitioned into the receptor fluid. Absorption of benzo[a]pyrene was mediated by biotransformation processes. A metabolite derived from the ultimate carcinogen of this compound, benzo[a]pyrene r-7, t-8,9,10-tetrahydrotetrol, was identified in the receptor fluid. Most of the administered DEHP remained in the skin and only a very small fraction of the dose partitioned into the receptor fluid in either viable or nonviable skin. Data from the present study led to the conclusion that the in vitro method can be utilized to predict in vivo absorption for compounds of high lipophilicity and that dermal metabolism facilitates partitioning of metabolites into the receptor fluid and hence may affect the biological activities of dermally applied compounds.

Percutaneous absorption/metabolism of phenanthrene in the hairless guinea pig: comparison of in vitro and in vivo results.

The in vitro and in vivo percutaneous absorption/metabolism of phenanthrene was investigated in hairless guinea pigs. Flow-through diffusion cells and Hepes-buffered Hanks' balanced salt solution (HHBSS) as receptor fluid were used in the in vitro system. When phenanthrene was applied to excised guinea pig skin mounted on the cells at dose levels of 6.6 and 15.2 micrograms/cm2, 89.7 and 79.1% of the administered doses were respectively absorbed into the skin and receptor fluids during a 24-hr perfusion period. These results are consistent with the in vivo data which showed approximately 80% absorption over the same period of time. Phenanthrene was metabolized in vitro into phenanthrene 9,10-dihydrodiol, 3,4-dihydrodiol, 1,2-dihydrodiol, and traces of hydroxy phenanthrenes. Of the materials absorbed in vitro, 92% was the parent compound and 7% the dihydrodiol metabolites. When a nonviable in vitro system was used, 68% of the applied 15.2 micrograms/cm2 dose was absorbed. Data from the present study demonstrate that the in vitro system is a good model for predicting in vivo percutaneous absorption of phenanthrene, and that penetration of phenanthrene through the skin is controlled more by the passive rate of diffusion than by metabolism.