Fig. 6: Treatment with LPS produces similar effects as faecal microbiota from non-responsive patients on the gut-on-a-chip.
a, Heat map of top R- and NR-associated taxa based on Wilcoxon test unadjusted P < 0.05. Bars on the right indicate log2 fold change of mean relative abundance (R/NR). b, Relative abundances (in a log10 base) of detected UniRef90 genes involved in LPS synthesis in patients’ faeces (nNR = 5, nR = 4, two-tailed Wilcoxon rank-sum test P = 4.1 × 10−7). c, Fluorescence microscopy images (n > 2 in 2 independent experiments) of dextran permeability assay (TRITC-dextran 4.4 kDa) on epithelial cells at day 8 (left) and day 9 (right) in control chips (top) and LPS-treated (24 h, 10 μg ml−1) chips (bottom). d, Normalized fluorescence intensity profile of dextran permeability assay (TRITC-dextran 4.4 kDa) of control and LPS-treated chips at day 8 and day 9. Data were normalized on maximum value. e, Apparent permeability (Papp) quantification of epithelial barrier in control and LPS-treated chips at day 8 and day 9 (P = 0.0072 for LPS day 9 versus CTR day 9, P = 0.0041 for LPS day 9 versus LPS day 8). f, Fluorescence microscopy images of ZO-1 (green) and E-cadherin (ECad, white) on epithelial cells in control and LPS-treated chips. g, Fluorescence microscopy images of ICAM-1 (red) on endothelial cells in control chips and after LPS treatment of epithelial tubule. h, Mean fluorescence intensity quantification of ZO-1 (P = 0.0182), E-cadherin (ECad, P = 0.0041) and ICAM-1 (P = 0.0022). i, Fluorescence microscopy images and quantification of adherent T cells (green, P = 0.0157) on endothelial cells cultured in endothelial medium or pre-treated with LPS. Bars represent mean ± s.d., ordinary one-way ANOVA (e), two-tailed Mann–Whitney test (h left and right), two-tailed unpaired t-test (h centre), two-tailed paired t-test (i) and Wilcoxon test (b). *P ≤ 0.05; **P ≤ 0.01.
