The human respiratory system is continuously subject to exposure to a wide range of potentially hazardous compounds in both indoor and outdoor air, including particulate matter (PM), volatile organic and inorganic compounds, and biological pollutants (Manisalidis et al., 2020; Gonzalez-Martin et al., 2021). Air pollution is a leading environmental risk factor for premature mortality and loss of healthy life years worldwide, with the largest contribution from fine ambient PM (Landrigan et al., 2018). Given the diverse range of potential health hazards, robust and predictive experimental models are needed for hazard identification and risk assessment. There is currently an ongoing effort to develop new approach methodologies (NAMs) to reduce the use of animals in research and to improve the physiological relevance for human exposure. Recent advances in inhalation toxicology include more complex three-dimensional (3D) cell culture models exposed at the air-liquid interface (ALI). In this approach, epithelial lung cells are cultivated on porous membranes in direct contact with the air on one side and cell culture media on the other side, which provides a more physiologically relevant model of the human respiratory system (Hiemstra et al., 2018; Yaqub et al., 2022).

The airway epithelium constitutes the first line of defence against inhaled toxicants. It forms a structural barrier that separates the external milieu from the underlying tissue and plays an important role in the innate immunity of the lung (Whitsett and Alenghat, 2015; Leiva-Juarez et al., 2018). The tracheobronchial airways are lined with a pseudostratified epithelium dominated by ciliated cells. However, a range of other cell types, such as basal progenitor cells, mucus-secreting goblet cells, club cells, serous cells and neuroendocrine cells, are also present at the mucosal surface and in submucosal glands, and are involved in the secretion of fluids, mucins, and host-defence proteins (Gao et al., 2015; Whitsett and Alenghat, 2015). In contrast, the alveolar epithelium is lined with only two types of epithelial cells, the thin squamous type I epithelial cells and the cuboidal type II epithelial cells. The type I cells cover most of the alveolar surface and are involved in gas exchange and alveolar fluid homeostasis, while the type II cells are secretory cells mainly responsible for synthesizing the pulmonary surfactant and constitute the main progenitor cell in the alveoli (Johnson et al., 2002; Whitsett and Alenghat, 2015; Lopez-Rodriguez et al., 2017). Expression of xenobiotic metabolizing enzymes are widely distributed in the human lung. However, CYP1A1 and CYP1B1, which are central in metabolism of polycyclic aromatic hydrocarbons (PAH), are most strongly expressed and induced in ciliated bronchial epithelial cells and club (Clara) cells, followed by alveolar type II cells, and alveolar macrophages (Hukkanen et al., 2002; Oesch et al., 2019).

As recently reviewed, several approaches are currently in use for representing the human lung epithelium, including various immortalized cell lines, primary lung epithelial cells, and induced pluripotent stem cell models (Hiemstra et al., 2018; Yaqub et al., 2022). Moreover, the epithelial cells can also be cultured together with other cell types, like macrophages, dendritic cells, mast cells, fibroblasts, mesenchymal cells, and endothelial cells (Klein et al., 2013; Chortarea et al., 2015; Marrazzo et al., 2016; Gras et al., 2017). Although cell lines do not fully recapitulate the physiology of the lung epithelium, they offer several advantages over more complex models in terms of increased availability, lack of inter-donor variability, ease of use, lower costs, and suitability for high-throughput screening purposes. Several epithelial lung cell lines are currently used, such as BEAS-2B, 16HBE14o-, Calu-3, A549, and NCI-H441, with A549 being the most common. These cells differ regarding the origin of the cells, whether they can recapitulate important processes like the production of mucus and lung surfactant, and the ability to form tight epithelial barriers (Hiemstra et al., 2018; Yaqub et al., 2022). By now, several studies have reported differences in responses between epithelial models based on primary cells and cell lines exposed to different stimuli (Kooter et al., 2013; Zavala et al., 2016; Rossner et al., 2021). However, comparative studies evaluating the responses of different 3D cell culture models based on commonly used cell lines exposed to the same stimuli are still scarce.

The purpose of the present study was to compare the responses of different advanced cell culture models of the human respiratory system when exposed to the same stimuli under the same experimental conditions. Three different epithelial cell lines were cultivated on porous membrane inserts in combination with differentiated THP-1 macrophage-like cells (Daigneault et al., 2010; Lund et al., 2016), and the endothelial cell line EA.hy926 (Edgell et al., 1983). Calu-3 and HBEC3-KT are human bronchial epithelial cell lines derived from lung adenocarcinoma and through immortalization of central bronchiole primary cells, respectively (Ramirez et al., 2004; Delgado et al., 2011), while A549 is a human lung carcinoma cell line with characteristics of alveolar type II epithelial cells (Lieber et al., 1976; Foster et al., 1998; Ramirez et al., 2004; Delgado et al., 2011; Bessa et al., 2023). Among these, only the Calu-3 cells are known to form a functional epithelial barrier (Wan et al., 2000; Kreft et al., 2015; Ren et al., 2016; Bessa et al., 2023). We first characterized the 3D cell culture models based on Calu-3, A549 and HBEC3-KT cells in terms of cellular composition and the permeability of the epithelial barrier capacity. Next, the responsiveness of the different 3D cell models was assessed by exposure to a reference sample of fine ambient air pollution PM under ALI conditions. Toxicological endpoints included cytotoxicity, release of pro-inflammatory cytokines, and expression of genes linked to inflammation, redox responses, and xenobiotic metabolism. The results show that SRM 2786 induced pro-inflammatory responses in all three cell models, which were strongly influenced by the presence of THP-1-derived macrophages. Moreover, the responses in the different compartments of the models appear to depend on the ability of the epithelial cells to form a tight epithelial barrier.