Systems toxicology assessment of a representative e-liquid formulation using human primary bronchial epithelial cells

      Marescotti, D.; Mathis, C.; Belcastro, V.; Leroy, P.; Acali, S.; Martin, F.; Dulize, R.; Bornand, D.; Baumer, K.; Peric, D.; Guedj, E.; Ortega Torres, L.; Biasioli, M.; Fuhrimann, M.; Fernandes, E.; Frauendorfer, F.; Luettich, K.; GonzalezSuarez, I.; Sciuscio, D.; Ivanov, N. V.; Peitsch, M. C.; Hoeng, J.
      Conference date
      Mar 12, 2019
      Conference name
      Society of Toxicology (SOT) 2019

      Cigarette smoke (CS) is a risk factor for respiratory and systemic diseases, smoking cessation remains the most effective approach for risk reduction. Innovative products with the potential to reduce the risk of smoking related diseases are being developed with the aim to provide a better alternative to smokers who would otherwise continue to smoke. Literature suggests that e-vapor products likely have a significantly lower risk profile than cigarettes. This is based on the notion that chemical composition of e-vapor aerosols is much simpler than that of CS, and ingredients are generally regarded as safe given their use in the food industry. Yet, data available for safety assessment in the context of inhalation toxicology are limited. A list of flavors used in e-liquids was grouped based on common physicochemical properties and available toxicological data, as defined by the European Food Safety Authority. Within each group, at least one representative flavor was selected to create a mixture, with 28 representative flavors dissolved in a matrix containing 41% propylene glycol, 38% vegetable glycerin, and 0.6% nicotine. We evaluated the effects of exposing normal human bronchial epithelial cells in submerged condition to 1) the e-liquid/flavor mix and 2) the 28 individual flavors. Finally, new mixtures were generated by selectively removing those flavors exhibiting the largest cytotoxic effects. Each e-liquid solution was tested over a 24-hour period using a real-time, impedance-based assay. Phenotypic effects were further evaluated using a battery of high-content screening endpoints. For additional mechanistic insights, effects of the flavor mixture and its corresponding matrix were investigated using gene expression analysis combined with a computational approach leveraging mechanistic network models to identify and quantify perturbed molecular pathways. The 28-flavor mixture (flavor content ~5%) was more cytotoxic than the matrix following a 24-hour exposure. Assessing each flavor individually, D-L-citronellol (0.48%) and alpha-pinene (0.048%) were found to exhibit the highest cytotoxicity. A decrease in cytotoxicity was observed only when D-L-citronellol was removed, suggesting that D-L-citronellol was accountable for most of the mixture’s overall cytotoxicity.