Understanding local respiratory toxicity and bioavailability of inhaled pesticides in an occupational exposure setting

Background: Occupational exposure to xenobiotic aerosols may occur within different settings, with a common example being the unintentional inhalation of pesticide aerosols sprayed to protect various crops. Current regulatory guidelines do not exploit emerging in vitro toxicity models and make the unvalidated assumption that 100% of the inhaled pesticide is absorbed into systemic circulation, in stark contrast to oral or dermal routes of exposure, which use established evidence-based in vitro and in silico methods, in addition to in vivo models. The assumption of 100% respiratory bioavailability overlooks several established clearance mechanisms, such as mucociliary clearance and lung metabolism. Data-driven approaches for predicting toxicity and estimating respiratory bioavailability based on experimental evidence are therefore urgently needed and have important implications for future risk assessments. This body of work is focused on (i) the suitability of in vitro respiratory models, (ii) the use of in vitro and in silico data for toxicokinetic predictions and (iii) NMR metabolomics as a sensitive measure of changes to lung cell phenotype following pesticide exposure at subtoxic concentrations. Methods: To address these knowledge gaps regarding regional respiratory toxicity throughout the airway and bioavailability of xenobiotics in an occupational exposure setting, several experimental methods were employed alongside in silico modelling approaches. Epithelial cell models (RPMI-2650, Calu-3, 16HBE14o-, BEAS-2B, TT1 and A549), reflecting different regions of the airway were utilised to investigate the toxicity, permeability and metabolism of a range of pesticides. The results from these respiratory models were compared against the gastrointestinal epithelial cell line Caco-2, which has been more widely employed in the field. Techniques used included various cytotoxicity assays, Phase I and II metabolism studies to highlight enzyme activity in the different models, NMR metabolomics to assess the changes to lung cell phenotype at subtoxic xenobiotic concentrations, and permeability studies using various Transwell models (Calu-3, Caco-2, PAMPA and lung lipid extract), alongside in vitro protein binding assays. Additionally, in vitro data was paired with in silico predictions for lung deposition and pharmacokinetics to predict the bioavailability of inhaled xenobiotics in an occupational exposure scenario. Results: Significant differences were found in the suitability of the different biological models for assessing toxicity, metabolism and permeability within the lungs. The effect of the physicochemical properties of different xenobiotics on the predicted bioavailability following inhalation was also shown to be substantial. Finally, in addition to cytotoxicity, several changes to cell phenotype were observed following realistic (and non-cytotoxic) pesticide exposure concentrations of 1-10 µM. Cytotoxicity, mitochondrial inhibition, CYP1A1 induction and increased intracellular glutathione were observed at within this concentration range for various pesticides, whilst chlorothalonil 20 µM was found to be capable of causing epithelial damage to Calu-3 cells cultured at air-liquid interface. Conclusion: This study confirmed that the assumption of 100% bioavailability for inhaled pesticides is often an overestimate and regional airway models may better predict the actual dose fraction reaching the systemic circulation. These in vitro data combined with in silico approaches highlight the potential to improve predictions of respiratory toxicity and bioavailability and improve future occupational risk assessments for respiratory exposure to xenobiotics.