Behavioral characterization of MDCK cells on liquid–liquid interfaces
Time-lapse observation was performed for MDCK cells cultured on fibronectin-coated or non-coated liquid–liquid or solid-liquid interfaces (Movie S1). After seeding, cells cultured on a fibronectin-coated liquid–liquid interface (Movie S1D) and those on fibronectin-coated or non-coated solid-liquid interfaces (Movie S1A and B), exhibited elongated shapes and active migration. As the culture time progressed, cells cultured on both fibronectin-coated liquid–liquid (Movie S1D) and solid-liquid interfaces (Movie S1A and B), reached a confluent state through continued active migration and division. In contrast, cells cultured on the non-coated liquid–liquid interface organized into island-like cell populations or cell-free-areas in the culture vessel, with cells in island-like regions showing reduced migration (Movie S1C). As the culture time progressed further, the island-like regions expanded through cell growth (Movie S1C). These results suggest that cell behaviors, such as migration and growth, serves as important cues for the maturation of epithelial cells on liquid–liquid interfaces.
To understand the temporal profile of cell growth during maturation on various culture surfaces, the density of cells cultured on liquid–liquid and solid-liquid interfaces in the culture vessel were monitored until t = 96 h (Fig. 1). At t = 24 h, the seeding density of cells cultured on non-coated liquid–liquid interfaces initially decreased, but these values increased over time. By t = 72 h, the number of cells across all culture conditions had increased to 2.6×105–3.2×105 cells/cm2, resembling values as at t = 96 h and indicating contact inhibition. These results indicate that the growth of cells cultured on liquid–liquid interfaces is comparable to that of cells cultured on solid-liquid interfaces.
Figure 1
To characterize MDCK cells during maturation, immunostaining for nuclei, F-actin, and ZO-2 of MDCK cells at t = 24 h and 72 h was performed, and immunostaining images were randomly captured in the culture vessel (Fig. 2). Additionally, the entire images of the culture vessel were captured to understand the spatial distribution of MDCK cells on the culture surfaces in each culture condition. Figures 3A-D show the immunostaining images of cell nuclei in a quarter of the culture vessel at t = 24 h. Notably, when cultured at the solid-liquid interface (Fig. 2A, B, E, F), the cells did not form island-like populations, and the cell densities were low in both conditions (Fig. 2a-2 and b-2). Actin stress fibers were abundant in cells cultured on non-coated solid-liquid interfaces (Fig. 2a-3), and these cells did not express ZO-2 (Fig. 2a-4). When cultured on solid-liquid interfaces coated with fibronectin, stress fiber formation was inhibited (Fig. 2b-3), and ZO-2 localization was observed in a portion of the cellular membrane (Fig. 2b-4). In contrast, cells cultured on fibronectin-coated liquid–liquid interfaces formed an island-like population (Fig. 2D), exhibiting a high-density state (Fig. 2d-2). The size of the island-like population on the fibronectin-coated liquid–liquid interfaces was larger than that of the island-like population on the non-coated liquid–liquid interfaces (Fig. 2G and H). Cells cultured on liquid–liquid interfaces with high population density and reduced stress fiber formation showed ZO-2 localization in the cytoplasm (Fig. 2c-2, c-3, c-4, d-2, d-3, d-4). Ultimately, at t= 72 h, the polygonal cells in the high-density region had a well-defined ZO-1 line surrounding the cell (Fig. 2L and P). Furthermore, the colocalization of ZO-2 and F-actin (Fig. 2L) indicated the stabilization of cell-cell adhesion. These results suggest that the density and distribution of cells, as a consequence of cell-cell communication through actin network, are related to the maturation of epithelial cells in confluent states.
Figure 2
Relationship between maturation extent and cell distribution in the culture vessel
To quantitatively investigate the relationship between the extent of TJ formation and local cell density, we determined the frequencies of local squares, FX, against nucleus density on a local square, XLN, of MDCK cells cultured on fibronectin-coated or non-coated liquid–liquid or solid-liquid interfaces at t = 24 h and 72 h (Fig. 3). Data were obtained from MDCK cells in triplicate experiments for each culture condition. The average values of cells cultured on non-coated and fibronectin-coated solid-liquid interfaces were \(\stackrel{-}{{X}_{\text{LN}}}\) = 1.0×105 \(\text{±}\) 0.22×105 nuclei/cm2 and \(\stackrel{-}{{X}_{\text{LN}}}\) = 1.2×105 \(\text{±}\) 0.38×105 nuclei/cm2, respectively (Fig. 3a and b). With elapsed culture time, the XLN values increased in both culture conditions, ranging from XLN = 2.5×105 nuclei/cm2 to XLN = 5.5×105 nuclei/cm2 at t = 72 h (Fig. 3e and f). This indicates the presence of uneven cell distribution in the culture vessel. The FZ values, reflecting broad distributions of XLN values in both non-coated and fibronectin-coated culture conditions, were FZ = 0.13 \(\text{±}\) 0.12 and FZ = 0.18 \(\text{±}\) 0.01, respectively (Fig. 3e and f). In contrast, when cultured on fibronectin-coated liquid–liquid interfaces, a broad distribution of XLN was observed, ranging from XLN = 0 to XLN = 3.8×105 nuclei/cm2 (Fig. 3d). The frequency of XLN = 0 on non-coated liquid–liquid interfaces was the highest (Fig. 3c), indicating numerous areas without cells. However, in areas where cells adhered to the liquid interface, the cell density was high, ranging from XLN = 2.0×105 nuclei/cm2 to XLN = 3.0×105 nuclei/cm2 at t = 24 h (Fig. 3c). At t = 72 h, the XLN distribution on fibronectin-coated liquid–liquid interfaces became narrower than that on non-coated liquid–liquid interfaces, and the FZ values reached their peak (FZ = 0.80 \(\text{±}\) 0.02). Collectively, these findings suggest that both high density and even distribution of cells in the culture vessel promote TJ formation, subsequently facilitating the maturation process.
Figure 3
Facilitation of uniform maturation through cell behaviors on liquid–liquid interfaces
Y27632, a specific inhibitor for Rho-dependent protein kinase (ROCK), induces cell survival and suppression of stress fiber formation [21, 22]. Therefore, in this study, we investigated the effect of the behavior of epithelial cells on their maturation on liquid–liquid or solid-liquid interfaces when exposed to media with and without 10 µg/mL Y27632 at t = 0–48 h (Fig. 4). At t = 24 h, exposing cells cultured on solid-liquid interfaces to Y27632 induced the suppression of stress fiber formation (Fig. 4B), leading to TJ formation (Fig. 4F). However, at t = 72 h, cells cultured on solid-liquid interfaces in both conditions caused F-actin disassembly (Fig. 4I and J), resulting in TJ disruption (Fig. 4M and N). In contrast, cells cultured on liquid–liquid interfaces in both culture conditions became highly dense (Fig. 4C and D), with ZO-2 localization in the cytoplasm (Fig. 4G and H). Similar to solid-liquid interface culture, when not exposed to Y27632, cells cultured on liquid–liquid interfaces at t = 72 h exhibited F-actin disassembly and TJ disruption (Fig. 4K and O). However, when exposed to Y27632, cells cultured on liquid–liquid interfaces at t = 72 h formed the actin belt (Fig. 4L), and ZO-2 colocalized with F-actin, indicating the stabilization of TJ formation at the plasma membrane (Fig. 4L and P). Interestingly, the staining image trend of F-actin and ZO-2, when exposed to Y27632, resembled that seen in cultures on fibronectin-coated surfaces (Figs. 2 and 4).
Figure 4
To further elucidate the correlation between TJ formation and even cell distribution in the culture vessel, the analysis was performed according to the procedure outlined in Fig. S2C. Specifically, we examined the correlation between the HLN and FZ values of MDCK cells cultured on fibronectin-coated or non-coated liquid–liquid or solid-liquid interfaces at t = 72 h to achieve high cell density. The homogeneity index, HLN, serves as an indicator of the even distribution of cells on the surfaces in each culture vessel. Notably, both the HLN and FZ values of cells cultured on the fibronectin-coated liquid–liquid interface, as well as of Y27632-exposed cells, were significantly higher than those cultured on the non-coated liquid–liquid interfaces (Fig. 5). Across all culture conditions, the FZ and HLN values were closely correlated (Fig. 5) (R2 = 0.93), suggesting that achieving an even distribution of cell nuclei via cell-cell interaction is crucial for maintaining TJ formation during the late phase of maturation.
Figure 5