Chemometrical assessment of adverse effects in lung cells induced by vehicle engine emissions.

Abstract

Vehicle engine exhausts contain complex mixtures of gaseous and particulate pollutants, which are known to affect lung functions adversely. Many in vitro studies have shown that exposure to engine exhaust can induce oxidative stress in lung cells, leading to cellular inflammation and cytotoxicity. However, it remains challenging to identify key harmful components and their specific adverse effects via traditional toxicological assessments. Machine learning (ML) methods offer new ways of analyzing such complex datasets and have gained attention in predicting toxicity outcomes and identifying key pollutants in mixtures responsible for adverse effects in a non-biased way. This study aims to understand the contribution of exhaust components to lung cell toxicity using ML techniques. Data were reanalyzed from previous studies (2015-2018), where a 3D human epithelial airway tissue model was exposed to gasoline and diesel engine exhausts under air-liquid interface (ALI) conditions with different fuels and exhaust after-treatment systems. This dataset included exhaust characteristics (particle number (PN), carbon monoxide (CO), total gaseous hydrocarbons (THC), and nitrogen oxides (NOx) levels) and corresponding biological responses (cytotoxicity, oxidative stress, and inflammatory responses). The relationships between pollutants and biological responses were explored using ML techniques, including hierarchical and nonhierarchical clustering and principal component analysis. The findings reveal both gaseous (CO, THC, and NOx) and particulate pollutants contribute to oxidative stress, inflammation, and cytotoxicity in lung cells, highlighting the significant role of each gaseous component. In addition, unmeasured factors beyond CO, THC, NOx, and PN likely contribute to biological effects, indicating the need for a more detailed characterization of exhaust parameters in ML analysis. By successfully integrating ML techniques, this study shows the potential of ML in identifying pollutant-specific contributions to cell toxicity. These insights can guide the analysis of complex exposure scenarios and inform regulatory measures and technical developments in emission control.