Mechanisms and Models of Skin Penetration

Mindmap of the world of Skin Penetration

Measurement of chemical penetration through the skin is important in both pharmacology and toxicology.  In the pharmaceutical industry, transdermal drug delivery devices and topical pharmaceuticals are a very important part of the market, and the focus in this industry is to enhance the penetration of drugs into and through the skin. In contrast, from a toxicological perspective, penetration of chemicals through the skin is normally a bad thing and the focus of toxicology is to assess and predict the hazards from cutaneous penetration of chemicals in the workplace, home and environment. The “science” of Skin Penetration has been developed to meet these needs. The mindmap shown above summarizes many of the important aspects related to skin penetration. It is the goal of our laboratory to improve the quality and use of skin penetration measurements.

(link back to main page)

In vitro methods:

One of the easiest ways to estimate penetration of drugs or chemicals through the skin is to take excised human (frequently leftover from surgeries) or laboratory animal (rat, guinea pig or pig) skin and place it between two chambers with the compound of interest in the donor chamber.  In these in vitro studies, steady-state flux is determined by measuring the appearance of the compound in the receptor solution in the other chamber.  Radioactive compounds are often used to make analysis of the receptor solution easier. Steady-state flux (mass/time or mass/area/time) is determined from the slope of the linear portion of the cumulative chemical absorbed vs time plot. Comparing these fluxes across a range of pharmaceutical products with exactly the same method allows acceptable rank-ordering of compounds for their ability to penetrate skin, which can be useful for selection of compounds or development of formulations.  Unfortunately, this type of study does not provide information that is accurate enough for predicting the rate of penetration of a specific chemical in humans for a variety of reasons.  Briefly, this is because of the dissimilarity between the environment of the skin in the diffusion cell and the skin on the human body (flux doesn’t extrapolate very well).  That said, careful experimentation with diffusion cells can provide fluxes that are “in the ballpark” and the studies are cheaper and easier than whole animal pharmacokinetic studies. 

We have used static diffusion cells to measure penetration (across the skin) and absorption (into the skin) of a saturated solution of organic chemical in water for short periods of time (5-20 minutes) (McDougal and Jurgens-Whitehead 2001).  The purpose of this study was to investigate the kinetics of this organic chemical in the skin at times before steady state was reached as would occur with showering or bathing.  We showed that absorption into the skin was rapid and linear from the beginning, but that the chemical did not out come through the skin for about 20 minutes.  This demonstrated that flux into the skin is constant from time zero, and the lag time is an artifact of the way we measure penetration. We concluded that the amount of chemical in the skin was approximately 8 times the amount of chemical that penetrated the skin and therefore the amount of chemical in the skin is important for body burden calculations from showering and bathing scenarios.  For more information about showering and bathing scenarios and absorption through the skin see (Bunge and McDougal 1998).

We used diffusion cells to estimate dermal penetration of Break-Free CLP® (a petroleum-based liquid solution for cleaning, lubricating and protecting firearms) in mouse, rat and pig skin (Arfsten et al. 2006).  Breakfree is a mixture of polyalphaolefin oil (65%), synthetic oils, esters and synthetic proprietary ingredients (27%), isoparaffinic hydrocarbons (5%), and dibasic ester (3%).  We analyzed the gas chromatographic peaks from seven major components to estimate the total absorption and penetration of the mixture.  In our study, mouse skin was 37 times more permeable to Break-Free CLP than pig skin and 6 times more permeable than rat skin.  Flux measurements from static diffusion cells showed an inverse correlation with mouse, rat, and pig skin thickness, as expected.  The concentration of Break-Free in mouse skin was 4.5 times higher than the amount found in rat skin and about 17 times higher than the amount absorbed by pig skin.  This paper showed a viable approach to determining the skin permeability of a mixture by analyzing the penetration of individual components.

JP-8 jet fuel has also been studied in our diffusion cell apparatus (McDougal et al. 2000).  We used static diffusion cells to measure both the flux of JP-8 and components across the skin and the kinetics of absorption into the skin.  Total flux of the hydrocarbon components was 20.3 micrograms/cm2/h.  Thirteen individual components of JP-8 penetrated into the receptor solution.  The fluxes ranged from a high of 51.5 micrograms/cm2/h (an additive, diethylene glycol monomethyl ether) to a low of 0.334 micrograms/cm2/h (tridecane).  Aromatic components penetrated most rapidly.  Six components (all aliphatic) were identified in the skin.  Concentrations absorbed into the skin at 3.5 h ranged from 0.055 micrograms per gram skin (tetradecane) to 0.266 micrograms per gram skin (undecane).  These results suggest: (1) that JP-8 penetration will not cause systemic toxicity because of low fluxes of all the components; and (2) the absorption of aliphatic components into the skin may be a cause of skin irritation. 

Because excised skin has no blood flow it is necessary when using diffusion cells to remove part of the dermis below the capillaries with a dermatome to mimic the distance chemicals must diffuse before being picked up by the blood.  We have estimated the level of the capillaries (and several other parameters) in several laboratory animal species (Grabau et al. 1995; Mattie et al. 1994). 

We also believe that these studies provide much better information when the amount of the chemical in the skin (absorption) is characterized in addition to the flux (penetration), because the amount in the skin may be subsequently absorbed. Once the exposure is over, if the skin is not occluded, the rate of evaporation (if there is one) and the rate of penetration from the skin depot will ultimately determine the amount absorbed. With volatile chemicals it is also very important to immediately flash-freeze the skin with liquid nitrogen before analyzing the amount of chemical in the skin to avoid loss due to volatilization. 

(link back to main page)

In vivo Methods:

Studies of skin penetration in the whole animal have the potential to provide much better information because blood flow, metabolism as well as nervous and hormonal responses of the skin are intact!  However, they are much more expensive and time consuming.  In vivo studies require particularly careful control of the dose applied to the skin.  The surface area exposed and the concentration (or mass) on the skin are the driving forces for penetration and unlike diffusion cell studies (where the area is carefully controlled and excess chemical is usually applied) these are harder to control in whole animal exposures.  See Poet and McDougal (2002) for some details.

One fascinating problem is how to expose just the skin of whole animals to chemical vapors.  Because the respiratory tract is so efficient at absorbing vapors and the skin is not, any of the chemical vapor that can get to the lungs will overwhelm (and mask) chemical that might come through the skin.  We developed and validated a dermal vapor exposure chamber to be able to determine permeability of the skin to organic chemical vapors.  The concept was to provide a mask for the rat to breathe fresh air while the rest of the body was exposed to vapor in a chamber.  We basically built a gas mask for a rat (thanks to the dental clinic and the custom mask shop on WPAFB).  The rats were trained to wear the masks over a 3-day period and then we carefully clipped their fur, hooked them up in the chamber, provided them with positive pressure clean air to breathe and drew blood samples during the exposure (with an indwelling jugular cannula) to determine the amount of chemical absorbed.  Using a PB-PK model validated for inhalation exposures, we could shut off the inhalation input and adjust the skin permeability (Kp) to match the blood concentrations (McDougal et al. 1985; McDougal et al. 1990).  This system received a US patent (4,582,055) (McDougal et al. 1986a) and as far as we know, nobody has gone to the trouble to duplicate the system.

Chamber and gas mask for rats

We used this chamber to measure the vapor permeability coefficient estimate the percent of vapor which might be taken up through the skin during a whole body exposure for eight volatile organic chemicals (McDougal et al. 1990).  The permeability coefficients measured in our study were generally 2-4 times greater than the permeability coefficients we estimated from studies done by other authors, primarily in Europe.

These studies suggest (if we are comfortable extrapolating them to humans) that in most cases absorption of organic chemical vapors through the skin will not contribute significantly to the hazard unless the individual is wearing a respirator.  When wearing a respirator, the dermal route may provide the largest route for absorption but still be insignificant, toxicologically.  In our minds, at least, this put to rest a lot of concern about whether vapor penetration through the skin might be hazardous.

Another major area of concern is how to do good experiments to skin penetration rates from liquid exposures to the skin, in vivo.  A very common method of trying to quantify absorption through the skin is to apply chemical to the surface of the skin and determine the percentage of the dose absorbed based on what is left on the surface (easiest way) or analyzing the animal and its sources of loss, i.e. breath, urine & feces (much, much harder).  Percentage absorbed is not an accurate means of predicting absorption.  Briefly (and hopefully avoiding a tirade), the percentage of the dose applied to the skin that is absorbed is not constant for other applied doses or other exposure times.  Here are two illustrations:

1) It is possible to apply more chemical to the skin surface than is available for absorption through the skin. If this is done experimentally, then the excess chemical is included in the dose and underestimates the dose that might be absorbed if there were not excess chemical. In an excess chemical situation, if you double the dose (with the same surface area) you will halve the percentage absorbed!

2) Percentage absorbed may depend on the time that you measure the amount absorbed, if you do not wait until absorption is complete. Consider a chemical placed on the skin that penetrates the skin at a specific rate.  If there is no way to lose the chemical from the surface (no evaporation or other loss from the surface), the ultimate percentage absorbed will be 100%!  You could experimentally get any percentage absorbed between zero and 100, depending on where you measured the amount absorbed.

Percentage of chemical absorbed is only useful if you can mimic a known human exposure scenario, i.e. a specific dose, area and time, in the experiment. These percentage numbers do not extrapolate accurately to other exposure scenarios. There are ways to get in vivo skin penetration information that can be predictive and will extrapolate to other concentrations, areas and times.

The approach we took was to control the surface area exposed and evaporation by supergluing a glass chamber to the back of a rat and inserting a jugular cannula to serially draw blood before, during and after the exposure.  Blood concentrations can be related to total chemical absorbed using a PB-PK model (see Models below for more information).

Drawing of a rat showing jugular cannula which is covered with a veterinary wrap and the glass cell for administration of chemical

Using this system, we studied the penetration of 14 volatile organic chemicals (1,2-dichloroethane, bromochloromethane, chloroform, benzene, tetrachloroethylene, dibromomethane, trichloroethylene, toluene, xylene, hexane, ethylbenzene, styrene, carbon tetrachloride and 1,2,2-trichloroethane) as 1/3, 2/3 and saturated aqueous solutions, as well as pure chemical (Morgan et al. 1991).  Blood levels of 1,2-dichloroethane and benzene continued to increase during the 24-hr exposure to neat chemical, while blood levels of the other neat VOCs peaked within 4 hr and then either decreased or remained about the same for the duration of the exposure.  Absorption of VOCs from one-third, two-thirds, or saturated aqueous solutions was rapid, and resulted in depletion of the chemical from the solution although only a small amount of water was absorbed.  Blood levels of each VOC were directly related to the exposure concentrations.  The rapid appearance of VOCs in the blood from aqueous solutions demonstrates that detectable amounts of VOCs were absorbed during exposure of only about 1% of the skin surface area of the rat.  Unfortunately, we never used PB-PK models to estimate flux and permeability coefficients for these experiments.

We did, however, completely analyze and model the non-steady state absorption of aqueous solutions of dibromomethane and bromochloromethane over 24 hours using our in vivo approach (Jepson and McDougal 1997).  The blood concentrations peaked at about 1-2 hr and diminished to nearly nothing at 24 hr.  Physiologically based models were used to estimate permeability coefficients for each of the exposures, although none of the exposures reached steady state due to the decreasing concentration of chemical on the surface of the skin, a constant permeability coefficient adequately described the blood concentrations during the prolonged exposure.  This demonstrated the utility of PB-PK models for skin exposures where the concentration on the skin was not constant.

Using the same methods, we investigated the effect of the vehicle containing the chemical on the rate of penetration through the skin (Jepson and McDougal 1999).  We studied the penetration of dibromomethane and bromochloromethane, in vivo, with 12 combinations of these chemicals and corn oil, mineral oil and water for 24 hours.  We looked at 25, 50 & 75% mixtures with these organic chemicals in the oils and 25, 50, 75 and 100% saturated in water.  We used a PB-PK model to estimate permeability coefficients (cm/hr) based on the blood concentrations over the 24 hours.  The permeability coefficients from the aqueous solutions were greater by a factor of 73 (dibromomethane) and 40 (bromochloromethane) even though the actual exposure concentrations in the water were more than 2 orders of magnitude less than the oils.  This was a dramatic demonstration of the ability of a water vehicle to “push” and organic chemical into the skin!  Because the partition coefficient between the chemical of interest and the vehicle determines the thermodynamics of the exposure, we showed that the specific chemical partition coefficient between the vehicle and the skin could be used to predict the permeability in other vehicles.  We predicted the rate of penetration of dibromomethane and bromochloromethane from a different oil – peanut oil and then did the experiments and observed permeability coefficients in agreement with the predictions.

(link back to main page)

Mechanisms:

The ability to extrapolate skin penetration depends on how well we can understand the factors that affect the process of absorption into and penetration through the skin.  Chemical into and through the skin is a passive process that is influenced by the characteristics of the skin and the chemical (including any vehicle).  We frequently treat the skin as a simple membrane, but it contains many layers, about 30 cell types and many appendages (hair follicles, sebaceous glands etc.).  The epidermis or thin outer layer of the skin is continually being replenished as cells in the basal layer differentiate and move toward the surface where they form a tight lipid/protein matrix for a couple of weeks before being replaced by newly differentiated cells from below.  This active layer (epidermis) has no blood flow and must receive nourishment (and eliminate waste) by diffusion from the capillaries in the dermis below.  The dermis provides support and elasticity with collagen and other proteins.  See Poet and McDougal (Poet and McDougal 2002) for details about structure.

It is the upper part of the epidermis, the stratum corneum, that provides the barrier function of the skin.  The stratum corneum is optimized to prevent the absorption of water in wet environments and the loss of water in dry environments.  This is a good thing – otherwise we would be big like a grape or small like a raisin depending on the environment we were in!  We showed using thermal gravimetric analysis that the diffusivity of water through porcine stratum corneum was about 2 orders of magnitude slower than diffusivity through the dermis (Liron et al. 1994).

If we compare the correlation approach for penetration of chemicals from an aqueous solution to penetration of chemicals from a lipid-like solution (JP-8) we see that the slope is negative in fuel and positive in water.  This difference in slope clearly illustrates that the predictive equations that have been developed for the correlation approach will not work when the vehicle is different from water.

Understanding the “first principle” mechanisms of skin penetration is necessary to accurately predict absorption and penetration of chemicals when they come into contact with the skin.  A very useful approach that incorporates this understanding is biologically based models.

(link back to main page)

Models:

It is not always possible to duplicate expected human exposures in order to estimate penetration of chemicals through the skin. The amount of chemical that penetrates depends on surface area exposed and surface concentration (or mass/area). Fortunately, modification of Fick’s law gives us a reasonable means to try to extrapolate between differences in exposure parameters such as area and concentration:

In words: “the amount of chemical that penetrates the skin barrier is directly proportional to the exposure concentration, the surface area exposed, the permeability of the chemical and the exposure time”.  Exposure concentration, area exposed and exposure time are parameters that can be carefully controlled in the laboratory, but are estimated or variable in human exposure scenarios.  Permeability is estimated from experiments, but sometimes (as a last resort), is estimated from correlations between known permeabilities and molecular weights and octanol/water partition coefficients.  Permeability is frequently expressed as distance/time, i.e. cm/hr and is often symbolized Kp.  Permeability can be estimated from steady-state flux by dividing the flux by the exposure concentration (see Bunge and McDougal (1998) for lots more detail).

Biologically-based pharmacokinetic models (also called, physiologically-based pharmacokinetic models, PB-PK) are the best way to predict the effects (therapeutic or toxic) of skin exposures.  As it sounds, biologically-based models are constrained by the biology!  Unlike compartmental models, which are used to fit a decrease in blood concentrations over time (for example), the compartments in biologically-based models are “real” compartments in the body (skin, liver, fat, brain etc.) and lumped compartments (rapidly perfused, slowly perfused) that are connected by “real” blood flows and have “real” affinities for the chemical being modeled.  The advantage of this type of model is the ability to extrapolate to other exposure concentrations, routes of exposure or other species (McDougal 1998; McDougal 2004). 

We have used these models several ways:

1. As a “tool” to predict body burden from cutaneous exposures based on serial blood concentrations after dermal vapor exposures (McDougal et al. 1986b; McDougal et al. 1990) and aqueous solutions (Jepson and McDougal 1997).

2. As a means of predicting the effects of vehicles (aqueous and oily) on the permeability of volatile organic chemicals through the skin (Jepson and McDougal 1999).

3. To compare the PB-PK approach to predicting absorption (into the skin) and penetration (through the skin) with other approaches in showering and bathing scenarios (McDougal and Jurgens-Whitehead 2001).

4. To explore the utility of serial and parallel skin subcompartments as predictors of penetration (Bookout, Jr. et al. 1996; Bookout, Jr. et al. 1997).

5. To model other routes of absorption for methylchloroform in three species after four routes of absorption (Reitz et al. 1988).

6. To model HCFC-123 blood concentration and metabolism (Vinegar et al. 1994) and extrapolation to humans (Williams et al. 1996) after inhalation exposures.

Although experimentally determining the parameters required for biologically-based models (partition coefficients, metabolic rates etc.) are relatively expensive the payoff frequently makes it worthwhile.  Biologically-based mathematical modeling of mixtures and mixed exposures is an extremely important area that is being developed.  Most toxicity studies are done with single chemicals even though there are almost no single chemical exposures in any workplace or home environment.  Biologically-based models that are developed to describe metabolic, pharmaco­dynamic and toxicodynamic interactions will allow prediction of effects in realistic environments.

(link back to main page)

Reference List

Arfsten, D. P., Garrett, C. M., Jederberg, W. W., Wilfong, E. R., and McDougal, J. N. (2006). Characterization of the skin penetration of a hydrocarbon-based weapons maintenance oil. J. Occup. Environ. Hyg. 3(9), 457-464. PM:16801258

Bookout, R. L., Jr., McDaniel, C. R., Quinn, D. W., and McDougal, J. N. (1996). Multilayered dermal subcompartments for modeling chemical absorption. SAR QSAR. Environ. Res. 5(3), 133-150. PM:9114511

Bookout, R. L., Jr., Quinn, D. W., and McDougal, J. N. (1997). Parallel dermal subcompartments for modeling chemical absorption. SAR QSAR. Environ. Res. 7(1-4), 259-279. PM:9501509

Bunge, A. L., and McDougal, J. N. (1998). Dermal Uptake. In Exposure to Contaminants in Drinking Water (S.S.Olin, Ed.), pp. 137-181. ILSI Press, Washington DC.

Grabau, J. H., Dong, L., Mattie, D. R., Jepson, G. W., and McDougal, J. N. Comparison of anatomical characteristics of the skin for several laboratory animals. AL/OE-TR-1995-0066, 1-27. 1995. Wright-Patterson AFB, OH, Air Force Research Laboratory. (link to Graubau tr 1995)

Jepson, G. W., and McDougal, J. N. (1997). Physiologically based modeling of nonsteady state dermal absorption of halogenated methanes from an aqueous solution. Toxicol. Appl. Pharmacol. 144(2), 315-324. PM:9194415

Jepson, G. W., and McDougal, J. N. (1999). Predicting vehicle effects on the dermal absorption of halogenated methanes using physiologically based modeling. Toxicol. Sci. 48(2), 180-188. PM:10353309

Liron, Z., Wright, R. L., and McDougal, J. N. (1994). Water diffusivity in porcine stratum corneum measured by a thermal gravimetric analysis technique. J. Pharm. Sci. 83(4), 457-462. PM:8046596

Mattie, D. R., Grabau, J. H., and McDougal, J. N. (1994). Significance of the dermal route of exposure to risk assessment. Risk Anal. 14(3), 277-284. PM:8029499

McDougal, J. N. (1998). Prediction - Physiological models. In Dermal Absorpion and Toxicity Assessment (M.Roberts and K.Walters, Eds.), pp. 189-202. Marcel Dekker, Inc., New York.

McDougal, J. N. (2004). Physiologically based pharmacokinetic modeling. In Dermatotoxicology (H.Zhai and H.Maibach, Eds.), Sixth ed., pp. 590-619. CRC Press, Boca Raton.

McDougal, J. N., Gargas, M. L., Strohaver, R. A., Jepson, G. W., Thimling, K. R., and Williams, M. A. In vivo Dermal Absorption method and system for Laboratory Animals. The United States of America as represented by the Secretary of the Air Force. 06/612,776(4,582,055). 4-15-1986a.  Washington, DC. 5-22-1984a.

McDougal, J. N., Jepson, G. W., Clewell, H. J., III, and Andersen, M. E. (1985). Dermal absorption of dihalomethane vapors. Toxicol. Appl. Pharmacol. 79(1), 150-158. PM:4049402

McDougal, J. N., Jepson, G. W., Clewell, H. J., III, Gargas, M. L., and Andersen, M. E. (1990). Dermal absorption of organic chemical vapors in rats and humans. Fundam. Appl. Toxicol. 14(2), 299-308. PM:2318354

McDougal, J. N., Jepson, G. W., Clewell, H. J., III, MacNaughton, M. G., and Andersen, M. E. (1986). A physiological pharmacokinetic model for dermal absorption of vapors in the rat. Toxicol. Appl. Pharmacol. 85(2), 286-294. PM:3764915

McDougal, J. N., and Jurgens-Whitehead, J. L. (2001). Short-Term dermal absorption and penetration of chemicals from aqueous solutions: theory and experiment. Risk Anal. 21(4), 719-726. PM:11726022

McDougal, J. N., Pollard, D. L., Weisman, W., Garrett, C. M., and Miller, T. E. (2000). Assessment of skin absorption and penetration of JP-8 jet fuel and its components. Toxicol. Sci. 55(2), 247-255. PM:10828255

McDougal, J. N., and Robinson, P. J. (2002). Assessment of dermal absorption and penetration of components of a fuel mixture (JP-8). Sci. Total Environ. 288(1-2), 23-30. PM:12013544

Morgan, D. L., Cooper, S. W., Carlock, D. L., Sykora, J. J., Sutton, B., Mattie, D. R., and McDougal, J. N. (1991). Dermal absorption of neat and aqueous volatile organic chemicals in the Fischer 344 rat. Environ. Res. 55(1), 51-63. PM:1855490

Poet, T. S., and McDougal, J. N. (2002). Skin absorption and human risk assessment. Chem. Biol. Interact. 140(1), 19-34. PM:12044558

Reitz, R. H., McDougal, J. N., Himmelstein, M. W., Nolan, R. J., and Schumann, A. M. (1988). Physiologically based pharmacokinetic modeling with methylchloroform: implications for interspecies, high dose/low dose, and dose route extrapolations. Toxicol. Appl. Pharmacol. 95(2), 185-199. PM:3420611

Vinegar, A., Williams, R. J., Fisher, J. W., and McDougal, J. N. (1994). Dose-dependent metabolism of 2,2-dichloro-1,1,1-trifluoroethane: a physiologically based pharmacokinetic model in the male Fischer 344 rat. Toxicol. Appl. Pharmacol. 129(1), 103-113. PM:7974482

Williams, R. J., Vinegar, A., McDougal, J. N., Jarabek, A. M., and Fisher, J. W. (1996). Rat to human extrapolation of HCFC-123 kinetics deduced from halothane kinetics: a corollary approach to physiologically based pharmacokinetic modeling. Fundam. Appl. Toxicol. 30(1), 55-66. PM:8812223

James N. McDougal
Office and Lab: 216 Health Sciences
Phone: (937) 775-3697
Fax: (937) 775-7221
E-Mail: james.mcdougal@wright.edu