Fig 2: MMP21 plays a crucial role in basal extrusion of β-cat ΔN/RasV12 cells surrounded by β-cat ΔN cells.a, b Establishment of MDCK mCherry-β-cateninΔ131 and MDCK-pTR GFP-RasV12 mCherry-β-cateninΔ131 cells. a Constitutive expression of mCherry-β-cateninΔ131 and tetracycline-induced expression of RasV12. b TOP FLASH reporter assay. Indicated cell lines were co-transfected with TOP FLASH or FOP FLASH and Renilla expressing vector, and subjected to measurement of luciferase activity. Data are mean ± s.e.m. *P = 0.0015; **P = 0.0032, unpaired two-tailed t-test; n = three independent experiments. c Immunofluorescence images of β-cat ΔN/RasV12 cells cultured alone or surrounded by β-cat ΔN cells. d Quantification of cell fates of β-cat ΔN/RasV12 cells cultured alone or surrounded by β-cat ΔN cells. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0190 for basal extrusion between two conditions, unpaired two-tailed t-test. e Experimental design and heatmap for differentially expressed genes between β-cat ΔN/RasV12 cells cultured alone and β-cat ΔN/RasV12 cells co-cultured with β-cat ΔN cells. f Volcano plot showing the distribution of fold changes and corrected P-values calculated by unpaired two-tailed t-test in β-cat ΔN/RasV12 cells mixed with β-cat ΔN cells compared with a single culture. The red or blue dots represent upregulated or downregulated genes with statistical significance, respectively. The upregulated and downregulated genes are shown in Supplementary Data 1. g q-PCR analysis of MMP-related metalloproteases. β-cat ΔN/RasV12 cells were co-cultured with β-cat ΔN cells or cultured alone, and subjected to q-PCR analysis. Values are shown as fold change relative to β-cat ΔN/RasV12 cells cultured alone. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0221, unpaired two-tailed t-test. h Immunofluorescence images of MMP21. β-cat ΔN/RasV12 cells were cultured alone or co-cultured with β-cat ΔN cells in the presence of 2.5 μg ml−1 BFA. Cells were stained with Hoechst 33342 (blue) and anti-MMP21 antibody (white). i Quantification of fluorescence intensity of MMP21. Values are expressed as a ratio relative to β-cat ΔN/RasV12 cells cultured alone. Data are mean ± s.e.m. *P < 0.001, unpaired two-tailed t-test; n = 286 and 284 cells pooled from three independent experiments. j Establishment of MDCK-pTR GFP-RasV12 mCherry-β-cateninΔ131 cells stably expressing MMP21 shRNA1 or MMP21 shRNA2. Knockdown of MMP21 was confirmed by western blotting. k Immunofluorescence images of β-cat ΔN/RasV12 cells, β-cat ΔN/RasV12 cells expressing MMP21 shRNA1 or MMP21 shRNA2 surrounded by β-cat ΔN cells. Scale bars, 10 μm (c, h, k). l Quantification of basal extrusion. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0186; **P = 0.0001, unpaired two-tailed t-test.

Fig 3: MMP21 promotes the non-cell autonomous basal invasion of APCMin/RasV12-transformed cells in vivo.a, b The non-cell autonomous upregulation of MMP21 in APCMin/RasV12 cells. a Immunofluorescence images of MMP21 in intestinal villi. APCMin-Villin-GFP, Villin-RasV12, or APCMin-Villin-RasV12 mice were injected with tamoxifen and were sacrificed 3 days later. Paraffin-embedded sections were stained with DAPI (blue), anti-GFP antibody (green), and anti-MMP21 antibody (red). b Immunofluorescence images of MMP21 in intestinal organoids from APCMin-Villin-GFP, Villin-RasV12, or APCMin-Villin-RasV12 mice at 24 h after tamoxifen treatment. Arrowheads indicate APCMin/RasV12-transformed cells with elevated MMP21 expression (a, b). c–h Effect of MMP21-knockout on basal extrusion of APCMin/RasV12 cells. c Strategy for the CRISPR-mediated MMP21-targeted mice. Physical maps of the murine MMP21 gene locus and sgRNA sequences are shown. The red line indicates the targeted region. Protospacer adjacent motif (PAM) sequences are underlined. d Genomic sequence of MMP21 locus from F1 generation MMP21-knockout mice. Note that 5890 bp covering Exon 1–Exon 7 of the MMP21 gene is deleted. e Western blot analysis of protein lysates prepared from intestinal tissues of MMP21(+/+), (+/−), (−/−) genotypes. β-actin served as a loading control. f Immunofluorescence images of MMP21 in intestinal villi from APCMin-Villin-RasV12 or APCMin-Villin-RasV12-MMP21 KO mice. Paraffin-embedded sections were stained with DAPI (blue), anti-GFP antibody (green), and anti-MMP21 antibody (red). Arrowheads indicate APCMin/RasV12-transformed cells. g, h Frequency of the basal extrusion of APCMin/RasV12 cells. g Immunofluorescence images of intestinal villi from APCMin-Villin-RasV12 or APCMin-Villin-RasV12-MMP21 KO mice 3 days after tamoxifen injection. An arrow indicates a basally extruded cell. Ba and Ap stand for the basal and apical sides, respectively (b, f, g). Scale bars, 50 μm (a) and 10 μm (b, f, g). h Quantification of apical and basal extrusion of transformed cells. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0004, unpaired two-tailed t-test.

Fig 4: NF-κB signaling governs basal extrusion of transformed cells via MMP21 upregulation.a, b Enhanced NF-κB activity in β-cat ΔN/RasV12 cells co-cultured with β-cat ΔN cells. a GSEA plot showing a significant correlation between the transcriptional profile of β-cat ΔN/RasV12 cells co-cultured with β-cat ΔN cells and IL-6 or chemokine production gene signatures. The P-values were calculated based on a two-sided permutation test. b NF-κB reporter assay. β-cat ΔN/RasV12 cells were co-transfected with NF-κB luciferase reporter and Renilla expressing vector and were co-cultured with β-cat ΔN cells or cultured alone. Cells were then subjected to measurement of luciferase activity. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0498, unpaired two-tailed t-test. c–g Effect of BAY 11-7082 on MMP21 expression and cell fates of β-cat ΔN/RasV12 cells. c Immunofluorescence images of MMP21 in the absence or presence of BAY 11-7082. β-cat ΔN/RasV12 cells were mixed with β-cat ΔN cells in the presence of BFA and were treated with DMSO or BAY 11-7082. Cells were stained with Hoechst 33342 (blue) and anti-MMP21 antibody (white). d Quantification of fluorescence intensity of MMP21. Values are expressed as a ratio relative to DMSO treatment. Data are mean ± s.e.m. *P < 0.001, unpaired two-tailed t-test; n = 418 and 420 cells pooled from three independent experiments. e p65 ChIP assay. An illustration of the p65 binding site within dog MMP21 promoter region is shown in the upper picture. Effect of TNF-α on the recruitment of p65 NF-κB to MMP21 promoter region is shown (lower panel). Data are mean ± s.e.m. n = three independent experiments. *P = 0.0069; **P = 0.0046, unpaired two-tailed t-test. f Immunofluorescence images of β-cat ΔN/RasV12 cells surrounded by β-cat ΔN cells in the absence or presence of BAY 11-7082. g Quantification of apical and basal extrusion in the absence or presence of BAY 11-7082. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0018; **P = 0.0019; ***P = 0.0176; ****P = 0.0007, unpaired two-tailed t-test. h, i Activation of NF-κB signaling in APCMin/RasV12 cells. h Immunofluorescence images of p65 in intestinal villi. APCMin-Villin-GFP, Villin-RasV12, or APCMin-Villin-RasV12 mice were injected with tamoxifen and were sacrificed 3 days later. Paraffin-embedded sections were stained with DAPI (blue), anti-GFP antibody (green) and anti-p65 antibody (red). i Immunofluorescence images of p65 in intestinal organoids from indicated mice. Arrowheads indicate APCMin/RasV12-transformed cells with accumulated p65 in the nucleus (h, i). j–n Effect of NF-κB inhibition on MMP21 expression and cell fates. (j) Immunofluorescence images of MMP21 in intestinal organoids from APCMin-Villin-RasV12 mice in the absence or presence of BAY 11-7082. k Immunofluorescence images of intestinal organoids treated with 100 nM tamoxifen and cultured for 24 h in the absence or presence of BAY 11-7082. Ba and Ap stand for basal and apical sides, respectively (i–k). l Quantification of the effect of BAY 11-7082 on apical and basal extrusion ex vivo. Data are mean ± s.e.m. n = four independent experiments. *P = 0.0097 for basally extruded cells between DMSO- and BAY 11-7082-treated group, unpaired two-tailed t-test. m Immunofluorescence images of intestinal villi from APCMin-Villin-RasV12 mice administered with PBS or SN50. Magnified images depict basal (an arrow) or apical extrusion of transformed cells from PBS- or SN50-administered mice, respectively. Scale bars, 10 μm (c, f, h–k, m). n Quantification of extrusion of APCMin/RasV12-transformed cells in the absence of the presence of SN50. Data are mean ± s.e.m. n = three independent experiments. *P = 0.0486 for basally extruded cells between PBS- and SN50-administrated group, unpaired two-tailed t-test.