Figure 6. Expression of ESRP1 and ESRP1-dependent hGPR137 isoforms is down-regulated in the diseased intestine and predicts CRC patient survival.

(A) ESRP1 transcript levels were measured in inflamed versus matched, non-inflamed intestinal biopsies from Crohn’s disease (CD) patients and normalized to EPCAM expression. Normalized ESRP1 transcript levels in the non-inflamed biopsy were set to one for each patient and fold induction was calculated for the corresponding inflamed biopsy (n = 15 samples per group). Data represent means ± standard error of the mean. (B) Immunohistochemistry was performed on intestinal tissue of CD patients to detect nuclear ESRP1, which was measured using automated quantification and normalized. Representative pictures are shown from a patient during remission and active disease, respectively (n = 31, 32, 7 and 14 biopsies per indicated group of cases). Data represent means ± standard error of the mean. Scale bars: 50 µm. (C) Kaplan-Meier survival curves of CRC patients with high (n = 77) or low (n = 88) expression of ESRP1 in tumor tissues. Representative IHC showing ESRP1-high and -low intestinal tumors. Scale bars: 100 µm. (D) Correlation between ESRP1 and hGPR137_Short expression in normal tissue of the large intestine (n = 51). (E) hGPR137_Short and hGPR137_Long isoform expression in tumor (n = 647) versus normal (n = 51) tissue of the large intestine. Data represent means ± standard error of the mean. (F) Kaplan-Meier survival curves of CRC patients with a high (n = 142) or low (n = 261) ratio of hGPR137_Short to hGPR137_Long transcripts in tumor tissues. Statistics: (A) Wilcoxon signed-rank test, (B) One-way ANOVA with Bonferroni post-test, (C) and (F) Log-rank test, (D) Spearman correlation, (E) Kruskal-Wallis with Dunn’s post-test. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Figure 6—figure supplement 1. Expression of GPR137 isoforms in the healthy and diseased intestine.

(A) Transcript levels of ESRP1 and hGPR137_Short were measured in normal human colon tissue (n = 10) and normalized to GAPDH expression. Expression of hGPR137_Short was assessed using primers specific for hGPR137 exon 4 and 5. (B) Transcript levels of Gpr137_Long were measured in normal and tumor intestinal tissue of AOM/DSS-treated WT mice. Data represent means ± standard error of the mean. n = 6 mice per group. (C) Kaplan-Meier survival curves of CRC patients with high (n = 163) or low (n = 241) hGPR137_Short expression in tumor tissues. Statistics: (A) Spearman correlation, (B) Student's t test and (C) Log-rank test.

Figure 6—figure supplement 2. ESRP1-dependent alternative mRNA splicing is required for epithelial integrity and intestinal homeostasis.

(A) (1) The Esrp1^Triaka allele generates altered mRNA splicing events. (2) Perturbation in the distribution of specific isoforms or generation of aberrant protein isoforms in epithelial cells reduce the integrity of the intestinal barrier. (3) This results in a facilitated penetration of microbes or microbial products into the intestinal mucosa. (4) A systemic antibody response is raised against translocated intestinal microbes. (5) Altered splicing through Esrp1^Triaka (6) decreases the proliferative capacity of intestinal epithelial cells. (7) This affects intestinal wound-healing, inflammation, and tumor development. (B) Mechanistically, Esrp1^Triaka leads to an altered ratio of Gpr137 transcript isoforms, among others. Preferential generation of Gpr137 Long in Esrp1^Triaka epithelial cells is associated with reduced Wnt/β-catenin signaling, and consequently diminished intestinal homeostasis and function. (C) ESRP1 protein is gradually downregulated during the adenoma to carcinoma sequence in human intestinal tumors. This correlates with a reduced expression of hGPR137_Short and an increase in hGPR137_Long isoforms, thereby possibly supporting tumor progression.