3.3. LPS-induced changes in gut barrier and BBB
To simulate the inflammatory response of gut-brain axis, our GBA chip was treated with LPS. The TEER values of the gut and the BBB were decreased in response to LPS treatment, with increased permeability of both barriers (Fig. 6A~D). LPS can be found on the cell walls of Gram-negative bacteria (Schumann, 1992), which is known to disrupt gut integrity and can penetrate systemic circulation across the gut barrier (Ghosh, Wang, Yannie, & Ghosh, 2020; Hirotani et al., 2008). This endotoxin can also impair BBB integrity and increased circulating LPS is associated with brain-related issues (Banks & Erickson, 2010; Qin et al., 2007; Zhu et al., 2017). Although the gut epithelium and BBB showed a similar trend in overall, the extent of response to LPS was different depending on the culture condition. For example, both cells seemed less sensitive to LPS in transwell condition than chip condition. In chip condition, physiological properties of Caco-2 such as barrier function, absorptive property and enzyme activity can be altered (Choe et al., 2017), which suggests that the response of Caco-2 cells to LPS may be affected by the flow.
It is also notable that the changes in absorption permeability was more significant than the changes in TEER in response to LPS. In the gut chip developed by Chi et al., Caco-2 cells featured 0.9 fold change in TEER and 2 fold change in permeability when exposed to flow (Chi et al., 2015). In the study conducted by Hirotani et al., a similar trend was observed, where the absorption permeability changed more dramatically than the values of TEER (Hirotani et al., 2008). The TEER value reflects the ionic conductance of the paracellular pathway in the cell monolayer while the flux of non-electrolyte tracers represents the paracellular water flow and the pore size of the tight junctions (Srinivasan et al., 2015). As TEER and flux of non-electrolyte tracers indicate different entities, the change in TEER value does not always correspond to the change in permeability of paracellular pathway (Zucco et al., 2005).
A proinflammatory cytokine, IL-8 is thought to engage in several inflammatory processes (Ehrlich et al., 1998), and BBB could be a crucial source of them, which could influence brain microenvironment (Banks & Erickson, 2010; Chen et al., 2001; Verma, Nakaoke, Dohgu, & Banks, 2006). The bEnd.3 cell line used in our study is a murine cell line, which shows inflammatory response different from human cells (Asfaha et al., 2013). Therefore, we used the primary human brain microvascular endothelial cell, hBMECs, to study inflammatory response in our GBA chip (Fig 6E~G). The IL-8 assay was conducted to confirm if proper alteration in the cytokine secretion can be observed after the endotoxin stimulation in the gut. Caco-2 cells in the well did not show significant changes. In contrast, hBMECs in the well showed significant increase of the cytokine after the stimulation. This result implies that the hBMECs were more sensitive to LPS treatment and the induced inflammatory response observed in our GBA chip originate mostly from brain endothelial cells.