Fig 2: PEAK1 does not promote CRC progression in vivo. (a) Timeline of the short-term VilCreERT2; Apcfl/fl mouse models to study hyperproliferation. (b) BrdU incorporation in the small intestine of VilCreERT2; Apcfl/fl (WT) and VilCreERT2; Apcfl/fl; Peak1−/− (Peak1−/−) mice, visualized by immunohistochemistry. Scale bar, 200 μm. (c). The average percentages of BrdU-positive cells per crypt are shown. Each point represents the average percentage per mouse (n = 6 mice per group; at least 25 crypts per sample were scored). (d, h, l) Timeline of the long-term mouse models to study tumorigenesis. Mice were injected with tamoxifen and sacrificed when they showed symptoms of intestinal tumors (Endpoint). (e, i, m) Kaplan-Meier survival plots comparing WT versus Peak1−/− mice after Cre-Recombinase induction with tamoxifen. Around day 540 all remaining VilCreERT2; Apcfl/+ (WT) and VilCreERT2; Apcfl/+; Peak1−/− (Peak1−/−) animals were sacrificed. Censored animals are indicated on the Kaplan-Meier curve as tick marks. Survival data and information for censoring are provided in Supplementary Table S1. (e). P values were calculated using the log-rank (Mantel-Cox) test (f, j, n). The number of tumors in the small intestine was scored macroscopically at the endpoint of the experiment. (g, k, o). The tumor burden was defined as the sum of the area (πr2) of all tumors. n = 14 mice per group, except for VilCreERT2; Apcfl/+ (WT) and VilCreERT2; Apcfl/+; Ptenfl/fl (WT) n = 13. Mann-Whitney U test was performed to determine statistical significance. ns, not significant.

Fig 3: Pathological analysis of intestinal lesions.The entire intestinal tract, including duodenum/jejunum, ilium, cecum, proximal colon, and distal colon/rectum,was examined in both WT and Peak1−/− mice for specific lesions, such as aberrant crypt foci (ACF), gastrointestinal intraepithelial neoplasia (GIN), adenoma, and adenocarcinoma. (a) Histograms illustrate the percentage of each lesion type present in the intestines of mice with different genotypes. Lesions were scored as detailed in the Materials and Methods section. Error bars represent the standard error of the mean; ns denotes not significant (Two-Way ANOVA with Šídák’s multiple comparison test between groups); For each group, n = 5, representing the 5 sections of the GI tract that were analyzed. (b) Microphotographs of H&E-stained sections illustrate each type of lesion as described. Scale bars: Upper panel: 500 μm (left), 200 μm (middle), 50 μm (right.). Middle panel: 100 μm (left and middle), 20 μm (right). Lower panel: 50 μm (left and right).

Fig 4: Decreased size of PEAK1-deficient Caco-2 and SW480 spheroids. (a,b) Caco-2 (a) or SW480 (b) cells were seeded in Matrigel and imaged on day 1, 4 and 7. Data were obtained from three independent experiments. In total, 300–330 spheroids were analyzed per condition. (c) Confocal microscopy z-stack images of spheroids isolated from the Matrigel on day 7, showing E-cadherin (Caco-2) or β-catenin (SW480), actin, and the cell nuclei (stained with DAPI). Scale bar, 20 μm. (d) Caco-2 cells were seeded in Matrigel and imaged on day 7 using bright field microscopy (left panels). Scale bar, 200 μm. Subsequently, spheroids were isolated from the gels, stained for E-cadherin, PAR3, and the cell nuclei using DAPI, and imaged using confocal microscopy (right panels). Scale bar, 20 μm. (e) Quantifications of the lumen size as a percentage of the total spheroid size. Data were obtained from at least 3 independent experiments. Total number of spheroids analyzed: 38 (WT); 50 (KO2); 50 (KO3); 42 (WT + EGF); 38 (KO2 + EGF); 53 (KO3 + EGF). (f) Representative western blots showing the expression of Scribble, PAR3, and E-cadherin in PEAK1 wild-type and deficient Caco-2 cells. GAPDH was used as loading control. Quantifications of Scribble signal intensities normalized to GAPDH levels are shown (n = 3 for WT and KO2 samples; n = 4 for WT and KO3 samples; bars show mean with s.d.). Original blots are presented in Supplementary Fig. 4. (g) Immunofluorescence analysis of cell-cell junctions in Caco-2 cells seeded on coverslips and fixed after 3 days. Scale bar, 20 μm. Violin and box plots range from the smallest to largest value; lines indicate the median and 25th to 75th percentiles. Mann-Whitney U test was performed to determine statistical significance. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Fig 5: PEAK1 deletion triggers ZO-1 degradation by exposing an LC3-interacting region on ZO-1.a–b Representative western blot analysis of wild-type (WT) and PEAK1 knockout (KO) Caco-2 cells treated with 100 nM Bafilomycin A1 (BafA1, autophagy inhibitor) (a) or 100 μM Rapamycin (Rapa, autophagy activator) (b) for 12 h. c Co-immunoprecipitation of Caco-2 cell lysates using anti-ZO-1 antibodies, showing LC-3B binding to ZO-1. d Immunofluorescence images showing colocalization of ZO-1 and LC-3B in autophagosomes of Caco-2 cells treated with 100 nM BafA1. Scale bars, 10 µm. e In vitro pull-down assays using different FC-tagged ZO-1 mutants and GFP-tagged PEAK1. FC-tagged ZO-1 deletions are visualized by Ponceau staining. f Predicted interaction motifs between ZO-1 and LC-3B, identified using an online tool (https://ilir.warwick.ac.uk/). PSSM: Position Specific Scoring Matrix (PSSM). High PSSM scores are assigned to the most frequent residues. g, h Co-immunoprecipitation analysis of HEK293T cells transfected with GFP-tagged ZO-1 (WT) or GFP-tagged mutant ZO-1 with a 1212-1217 amino acid deletion (Δ1212-1217), confirming this amino acid region regulates ZO-1 binding to LC-3B (g) and PEAK1 (h). (i) Co-immunoprecipitation using an anti-ZO-1 antibody in wild-type (WT) and PEAK1 knockout (KO) Caco-2 cells treated with 100 nM BafA1 for 12 h. j Representative Immunofluorescence images (left panel) and quantification of the ratio of ZO-1 and LC-3B co-localization to total ZO-1 puncta per cell (right panel) in WT and PEAK1 KO Caco-2 cells treated with 100 nM BafA1. Scale bars, 10 µm. KO = PEAK1 KO. Data are presented as mean ± SD. 79 cells per group were analyzed. Unpaired two-tailed Student’s t-test. k Co-immunoprecipitation using anti-ZO1 antibodies in Caco-2 cells treated with 100 nM BafA1 for 12 h, followed by an additional treatment with 10 uM PP2 inhibitor or vehicle (DMSO) for 2 h. (l) Left panel: immunofluorescence microscopy images detecting ZO-1 (red) and LC-3B (green) in Caco-2 cells treated with 100 nM BafA1 for 12 h, followed by treatment with 10 uM PP2 inhibitor or vehicle (DMSO) for 2 h. Scale bars, 10 µm. Right panel: quantification of the ratios of ZO-1 and LC-3B co-localization to total ZO-1 puncta per cell. Data are presented as mean ± SD. WT, n = 79 cells per group. Unpaired two-tailed Student’s t-test. m Co-immunoprecipitation analysis in the protein lysates from WT and Y724F mutant Caco-2 cells expressing CSK (sgNC controls) or lacking CSK (CSK KO) and treated with 100 nM BafA1 for 12 h. n Representative confocal images showing ZO-1 (red) and LC-3B (green) in WT and Y724F mutant Caco-2 cells expressing CSK (sgNC controls) or lacking CSK (CSK KO) and treated with 100 nM BafA1 for 12 h. Scale bars, 10 µm. Right panel: quantification of the ratios of ZO-1 and LC-3B co-localization to total ZO-1 puncta per cell. Data are presented as mean ± SD. n = 79 cells per group. Kruskal-Wallis test, followed by Dunn’s multiple comparisons test. The quantified relative expression levels shown in western blots are indicated in red, and represents the results of at least three repeated experiments.