A three-compartment epithelial structure formed by a 3.0-µm-pored membrane

(a) Schematic diagram of experiments using normal human epidermal keratinocytes cultured on a microporous membrane and exposed to air for 14 days to produce 3D epidermis. (b) Hematoxylin and eosin (HE) staining and vimentin (Vim: brown) labeling of 3D tissue derived from each microporous membrane (Mem). Scale bar: 50 μm. (c) HE staining of 3D tissue formed by a 3.0-µm-pored membrane. Scale bar: 50 μm. (d) E-cadherin (E-cad)/N-cadherin (N-cad) (upper images), Slug (middle images), and Vim (lower images) staining of 3D tissue formed by a 0.4-µm-(left) or 3.0-µm-pored (right) membrane. Scale bar: 50 μm (low-magnified images) and 20 μm (high-magnified image). (e) Vim staining of 3D tissue formed by a 3.0-µm-pored membrane. Scale bar: 20 μm. (f) Quantitative RT-PCR (qRT-PCR) of CDH1 and CDH2. (g) qRT-PCR of FN1, VIM, SNAI1, and SNAI2. N = 12 biological replicates per group. Data were analyzed with two-tailed Mann–Whitney U tests. * p < 0.05; *** p < 0.001; **** p < 0.0001. (h) Schematic diagram of the three-compartment epithelial structure. Representative images from three or more replicates per group are shown.

Oscillatory movement of keratinocytes on a 3.0-µm-pored membrane

(a) Schematic diagram of LifeAct-KCs cultured on a microporous membrane. Observation at the level of cell exit from the micropores. (b) Time-lapse images of LifeAct-KCs at the level of cell exit from 3.0-µm micropores. Representative images are shown at 30-minute time intervals. Time stamp shows hour:minute. Arrows indicate the cells that exit from the pores and expand under the membrane. Arrowheads indicate the cells that show oscillatory movement. Scale bar: 20 µm. (c) Schematic diagram of LifeAct-KCs cultured on a microporous membrane. Observation at the level of cell entry into the micropores. (d) Time-lapse images of LifeAct-KCs at the level of cell entry into 3.0-µm micropores. Representative images are shown at 5-minute intervals. Time stamp shows hour:minute. Arrowheads indicate the cells that show oscillatory movement. Scale bar: 20 µm. (e, f) Time-lapse images and schematic diagrams of LifeAct-KCs at the level of cell entry into 3.0-µm (e) or 8.0-µm (f) micropores. Representative images are shown at 2-minute intervals. Time stamp shows hour:minute. Scale bar: 10 µm. (g, h) Quantification of oscillatory movement of LifeAct-KCs on 3.0-µm or 8.0-µm micropores. (g) LifeAct signals (white) are shown as a spectrogram, where the X-axis represents each pore, and the Y-axis represents time. Data are summarized from three independent experiments. (h) Graphs comparing the number of pores presenting the LifeAct signals within 3.0-µm or 8.0-µm micropores per time. The left panel shows oscillation time as the X-axis and the number of pores as the Y-axis. The right panel shows unidirectional movement of the cells (no oscillation within 30 minutes) within each micropore. Data are summarized from three independent experiments. (i, j) E-cad and Vim labeling of LifeAct-KCs at the level of cells above (i) or within (j) 3.0-µm micropores. Autofluorescence of micropores is shown in blue (j). Scale bar: 10 µm. Representative images from three or more replicates per group are shown.

Chemical treatment on KC entry into and exit from 3.0-µm micropores

(a–c) Quantification of LifeAct-KCs present within 3.0-µm micropores 6 hours after cell seeding. Cells were treated with TGF-β ligand (a), TGF-β receptor inhibitors, SB431542, and SB525334 (b), or cytochalasin D (c). (d–f) Quantification of LifeAct-KCs present below 3.0-µm micropores 24 hours after cell seeding. Cells were treated with TGF-β ligand (d), TGF-β receptor inhibitors (e), or cytochalasin D (f). Each dotted line represents an experimental batch, and the red line indicates the overall mean. N = 7 biological replicates per group. Data were analyzed using a gamma generalized linear model with a log link function, adjusting for experimental batch as a categorical fixed effect. * p < 0.05; **** p < 0.0001; ns, not significant.

Regulatory role of Piezo1 and KRT6 on KC entry into and exit from 3.0-µm micropores

(a, b) Quantification of ruthenium red–treated LifeAct-KCs present within 3.0 µm micropores at 6 hours (a) and the cells present below 3.0-µm micropores 24 hours after cell seeding (b). (c, d) Quantification of PIEZO1 KO LifeAct-KCs present within 3.0 µm micropores at 6 hours (c) and the cells present below 3.0-µm micropores 24 hours after cell seeding (d). (e, f) Quantification of Yoda1-treated LifeAct-KCs present within 3.0-µm micropores at 6 hours (e) and the cells present below 3.0-µm micropores 24 hours after cell seeding (f). Each dotted line represents an experimental batch, and the red line indicates the overall mean. N = 7 (a, b) or 9 (c–f) biological replicates per group. Data were analyzed using a gamma generalized linear model with a log link function, adjusting for experimental batch as a categorical fixed effect. * p < 0.05; *** p < 0.001; **** p < 0.0001; ns, not significant. (g) Schematic diagram of bulk RNA-seq comparing HaCaT KCs cultured on a 0.4-µm- or 3.0-µm-pored membrane. The magenta dashed rectangles indicate the cells compared in the bulk RNA-seq. (h) Heatmap of differentially expressed genes identified by the bulk RNA-seq. The full gene lists are provided in Figure 4—source data 1. (i) Gene ontology biological process terms enriched in upregulated and downregulated genes. (j) Heatmap of differentially expressed keratin genes in HaCaT KCs cultured on a 0.4-µm- or 3.0 µm-pored membrane. (k) Keratin 6 (KRT6) labeling of normal human epidermal KCs cultured on a 0.4-µm- or 3.0-µm-pored membrane for 14 days. Scale bar: 20 µm. Representative images from three replicates per group are shown. (l, m) Quantification of pan-KRT6 KO LifeAct-KCs present within 3.0-µm micropores at 6 hours (l) and the cells present below 3.0-µm micropores 24 hours after cell seeding (m). Each dotted line represents an experimental batch. N = 9 biological replicates per group. The parental data are shared between panels c, d and l, m, because the parental, PIEZO1 KO, and pan-KRT6 KO cell lines were assayed in parallel in the same experiments. Data were analyzed using a gamma generalized linear model with a log link function, adjusting for experimental batch as a categorical fixed effect. *** p < 0.001.

Primers used in qRT-PCR