Self-organized intestinal epithelial monolayers in crypt and villus-like domains show effective barrier function

Abstract

Intestinal organoids have emerged as a powerful in vitro tool for studying intestinal biology due to their resemblance to in vivo tissue at the structural and functional levels. However, their sphere-like geometry prevents access to the apical side of the epithelium, making them unsuitable for standard functional assays designed for flat cell monolayers. Here, we describe a simple method for the formation of epithelial monolayers that recapitulates the in vivo-like cell type composition and organization and that is suitable for functional tissue barrier assays. In our approach, epithelial monolayer spreading is driven by the substrate stiffness, while tissue barrier function is achieved by the basolateral delivery of medium enriched with stem cell niche and myofibroblast-derived factors. These monolayers contain major intestinal epithelial cell types organized into proliferating crypt-like domains and differentiated villus-like regions, closely resembling the in vivo cell distribution. As a unique characteristic, these epithelial monolayers form functional epithelial barriers with an accessible apical surface and physiologically relevant transepithelial electrical resistance values. Our technology offers an up-to-date and novel culture method for intestinal epithelium, providing an in vivo-like cell composition and distribution in a tissue culture format compatible with high-throughput drug absorption or microbe-epithelium interaction studies.

Figure 1

Epithelial monolayers with in vivo-like cellular organization are formed on Matrigel-coated hard substrates. (A) Scheme depicting the different experimental set-ups employed. Either organoids-derived crypts or single cells where seeded on top of thin or thick layers of Matrigel-coated polystyrene plates. Epithelial structures obtained under each of these conditions are shown. (B) Representative images corresponding to co-immunofluorescence of filamentous actin (F-actin) and Ki67 from organoid-derived crypts and single cells cultured either on a thin film of Matrigel, referred as “Hard” substrate, or on a thick layer of Matrigel, referred as “Soft” substrate after 5 days of culture. Scale bars: 100 µm. (C) Co-immunofluorescence images for GFP (Lgr5-GFP⁺) and Ki67 (left panel), Lysozyme (Lyz) and GFP (middle panel), and Ki67 and cytokeratin 20 (CK20) (right panel) of epithelial monolayers derived from crypts (upper row) or single cells (lower row) grown on Matrigel-coated hard substrates for 7 days. Scale bars: 50 µm. (D) Image of Periodic acid-Schiff base (PAS) staining of epithelial monolayers grown on Matrigel-coated hard substrates for 7 days. Scale bar: 100 µm. (E) Immunostaining for zonula occludens 1 (ZO-1) and F-actin of epithelial monolayers grown on Matrigel-coated hard substrates for 7 days. Upper and lower panels show the top view and the orthogonal cross-sections of the monolayers, respectively. Scale bars: 50 µm (upper panel); 10 µm (lower panel). In all images Dapi was used to stain the nuclei.

Figure 2

Stem cells self-assemble in crypt-like domains and originate self-sustainable epithelial monolayers. (A) Live-imaging sequence of overlapped bright field and GFP signal corresponding to 60 hours after seeding organoid-derived crypts and (B) 96 hours after seeding organoid-derived single cells on a thin film of Matrigel. The corresponding time for each snapshot is shown in each panel. White arrow heads indicate Lgr5-GFP⁺ cells. Scale bars: 50 µm (A, upper row) and 100 µm (A, lower row, and B). (C) Immunofluorescence of EdU, Ki67, and ZO-1 of an EdU pulse-chase experiment. Representative images of 0 h, 24 h, 48 h, 72 h, and 96 h after EdU chase of either crypt-pieces (upper panels) or single cell (lower panels) derived cultures. Scale bars: 50 µm. (D) Immunofluorescence for Ki67 and F-actin of epithelial monolayers from passage 0 to passage 3 (P0, P1, P2, P3). Epithelial monolayers were passed by enzymatical digestion after every 6 to 8 days in culture. Scale bars: 100 µm.

Figure 3

Epithelial monolayers maintain in vivo-like cellular organization when grown on Matrigel-coated Transwell insert membranes. (A) Scheme depicting the experimental setup. Organoid-derived digested crypts were seeded on top of Matrigel-coated polycarbonate porous membranes and cultured under asymmetric administration of ENR_CV-medium. ENR_CV-medium: EGF, Noggin, R-Spondin 1, CHIR99021 and valproic acid-medium. Representative immunofluorescence images of (B) F-actin and Ki67, (C) GFP, and (D) EPHB2 expressed on epithelial monolayers grown on Matrigel-coated polycarbonate membranes and cultured for 10 days. White arrowheads indicate crypt-like regions. The inset panels in B, C, and D correspond to zoomed images of the crypt-like regions. Scale bars: 100 µm. In all images Dapi was used to stain the nuclei.

Figure 4

Biochemical factors produced by intestinal subepithelial myofibroblasts and Wnt3a boost epithelial growth. (A) Representative bright field images of epithelial monolayers grown on Matrigel-coated polystyrene plates and cultured with ENR_CV-medium, ISEMF_CM supplemented with ENR_CV, and ISEMF_CM supplemented with ENR_CV and Wnt3a after 5 days of culture. Black dashed lines mark the epithelial borders. Scale bars: 100 µm. (B) Graph showing the surface coverage percentage (with respect to the total substrate area) for each culture condition. Quantitative data was evaluated from 12 randomly selected regions from bright field images obtained from 3 different samples for each condition (n = 3). The data is represented as mean ± standard deviation. Statistics were performed with Student’s t-test setting *p < 0.05. (C) Representative fluorescence images of F-actin and cell nuclei for monolayers cultured in Transwell inserts under the asymmetric administration of ENR_CV-medium supplemented with ISEMF_CM and Wnt3a, at day 10 (upper row) and 20 (lower row) of culture. Left panels show the tile-scan images of the entire Transwell membrane surface (0.33 cm²), while right panels show higher magnification images for each condition. Scale bars: 1 mm (left panel) and 100 µm (right panel). (D) Graph plotting the surface coverage percentages with respect to the total substrate area as a function of the cell culture time. Six samples were analysed for day 10 (n = 6) and four samples for day 20 (n = 4). ENR_CV-medium: EGF, Noggin, R-Spondin 1, CHIR99021 and valproic acid containing -medium and ISEMF_CM: Intestinal subepithelial myofibroblast-conditioned medium. The data were represented as mean ± standard deviation. Statistics were performed through a Student’s t-test setting *p < 0.05.

Figure 5

Epithelial monolayers self-organized in crypt and villus-like domains show effective barrier function. (A) Representative immunofluorescence images of epithelial monolayers stained for GFP, Ki67, and CK20 after 10 and 20 days in culture on Matrigel-coated Transwell inserts. Scale bars: 100 µm. (B) Graph plotting the percentage of positive cells from the total cell number counted for each marker at day 10 (n = 3) and at day 20 (n = 2) of culture. The data were represented as mean ± standard deviation (statistics performed through pairwise Student’s t-test, significance level set at *p < 0.05). (C) Representative orthogonal cross-sections of randomly selected flat epithelial monolayers stained for F-actin and nuclei at day 10 and 20 of culture. Scale bars: 10 µm. The average cell heights, represented in the figure, were computed as 7.7 ± 1.3 µm and 10.3 ± 1.7 µm for day 10 (h₁₀) and day 20 (h₂₀), respectively. For the measurements ten randomly selected cells per sample were measured and a total of three independent samples (n = 3) for each condition were analysed. The data were represented as mean ± standard deviation. (D) Plot of the transepithelial electrical resistance (TEER) values measured for the epithelial monolayers as a function of the cell culture time. Reported values are corrected from their corresponding background and account for the total area of the samples. Values are reported as the mean ± standard deviation and correspond to three independent experiments with 5 technical replicas (n = 5). (E) Plot of the amount of permeated FITC-dextran 4.4 kDa (FD4) through the epithelial monolayers grown on Transwell inserts. As controls, the flux through the porous Transwell insert membranes coated with a thin layer of Matrigel and through the porous membranes alone was measured. Values are reported as the mean ± standard deviation and correspond to three technical replicas (n = 3).