PubMed2005
Occupational and environmental exposure to topical chemicals is usually in the form of complex chemical mixtures, yet risk assessment is based on experimentally derived data from individual chemical exposures from a single, usually aqueous vehicle, or from computed physiochemical properties. We present an approach using hybrid quantitative structure permeation relationships (QSPeR) models where absorption through porcine skin flow-through diffusion cells is well predicted using a QSPeR model describing the individual penetrants, coupled with a mixture factor (MF) that accounts for physicochemical properties of the vehicle/mixture components. The baseline equation is log k(p) = c + mMF + a sigma alpha2(H) + b sigma beta2(H) + s pi2(H) + rR2 + vV(x) where sigma alpha2(H) is the hydrogen-bond donor acidity, sigma beta2(H) is the hydrogen-bond acceptor basicity, pi2(H) is the dipolarity/polarizability, R2 represents the excess molar refractivity, and V(x) is the McGowan volume of the penetrants of interest; c, m, a, b, s, r, and v are strength coefficients coupling these descriptors to skin permeability (k(p)) of 12 penetrants (atrazine, chlorpyrifos, ethylparathion, fenthion, methylparathion, nonylphenol, rho-nitrophenol, pentachlorophenol, phenol, propazine, simazine, and triazine) in 24 mixtures. Mixtures consisted of full factorial combinations of vehicles (water, ethanol, propylene glycol) and additives (sodium lauryl sulfate, methyl nicotinate). An additional set of 4 penetrants (DEET, SDS, permethrin, ricinoleic acid) in different mixtures were included to assess applicability of this approach. This resulted in a dataset of 16 compounds administered in 344 treatment combinations. Across all exposures with no MF, R2 for absorption was 0.62. With the MF, correlations increased up to 0.78. Parameters correlated to the MF include refractive index, polarizability and log (1/Henry's Law Constant) of the mixture components. These factors should not be considered final as the focus of these studies was solely to determine if knowledge of the physical properties of a mixture would improve predicting skin permeability. Inclusion of multiple mixture factors should further improve predictability. The importance of these findings is that there is an approach whereby the effects of a mixture on dermal absorption of a penetrant of interest can be quantitated in a standard QSPeR model if physicochemical properties of the mixture are also incorporated.