Region-specific Wnt signaling responses promote gastric polyp formation in patients with familial adenomatous polyposis

Abstract

Germline adenomatous polyposis coli (APC) mutation in patients with familial adenomatous polyposis (FAP) promotes gastrointestinal polyposis, including the formation of frequent gastric fundic gland polyps (FGPs). In this study, we investigated how dysregulated Wnt signaling promotes FGPs and why they localize to the corpus region of the stomach. We developed a biobank of FGP and surrounding nonpolyp corpus biopsies and organoids from patients with FAP for comparative studies. Polyp biopsies and polyp-derived organoids exhibited enhanced Wnt target gene expression. Polyp-derived organoids with intrinsically upregulated Wnt signaling showed poor tolerance to further induction, suggesting that high Wnt restricts growth. Targeted genomic sequencing revealed that most gastric polyps did not arise via APC loss of heterozygosity. Studies in genetic mouse models demonstrated that heterozygous Apc loss increased epithelial cell proliferation in the corpus but not the antrum, while homozygous Apc loss was not maintained in the corpus yet induced hyperproliferation in the antrum. Our findings suggest that heterozygous APC mutation in patients with FAP may be sufficient to drive polyp formation in the corpus region while subsequent loss of heterozygosity to further enhance Wnt signaling is not tolerated. This finding contextualizes the abundant yet benign nature of gastric polyps in FAP patient corpus compared with the rare, yet adenomatous polyps in the antrum.

Keywords: Adult stem cells; Gastric cancer; Gastroenterology; Tumor suppressors.

Figure 1. FGP biopsy samples from patients with FAP have increased Wnt target gene expression.

(A) Schematic of Wnt signaling in Wnt OFF, in Wnt ON, and in FAP patients with mutated APC (APC*) exhibiting nuclear localization of β-catenin in the absence of Wnt ligand. (B) Biopsies from paired fundic gland polyps (P) and surrounding nonpolyp (NP) corpus tissue were collected to establish an FAP patient biobank of RNA, DNA, and organoids. (C) Schematic of the APC protein labeled with specific familial mutation sites for the patients in our biobank. Patients with established polyp and nonpolyp organoids are designated. Only patients with known germline mutations are included on this schematic (see Table 1). (D) Relative mRNA abundance of select genes in FAP patient biopsies. qPCR analysis of Wnt target–, cell marker–, and inflammation-related transcripts, with HPRT used as an internal reference transcript. Data are displayed as log₂ fold-change (error bars minimum to maximum values, box length IQR, whiskers outliers) relative to patient-matched nonpolyp tissue (n_nonpolyp = 20–25 biopsies, n_polyp = 21–30 biopsies). Statistical analysis by unpaired parametric Student’s t test (*P < 0.05, **P < 0.005, ***P < 0.001; see Supplemental Figure 1).

Figure 2. Enhanced Wnt signaling in a subset of FAP gastric polyp-derived organoids.

(A) Wnt target gene expression was measured in patient-matched polyp (P)/nonpolyp (NP) organoid pairs following 6 days of growth in WR media or following 4 days of growth in WR and 2 days of growth in WR-free media. (B) LGR5 mRNA abundance in FAP organoids grown in WR or WR-free media. Data shown as mean ± SD mRNA abundance relative to the reference gene ACTB (n = 3 individual wells; *P < 0.05, **P < 0.005, ***P < 0.001 by 1-way ANOVA with Tukey’s multiple-comparison test). (C) Heatmap of relative Wnt target gene expression in polyp organoids. Data are shown as mean fold-change relative to the expression of each target gene in matched nonpolyp organoids grown in WR media for 6 days. (D) Heatmap of LGR5 mRNA expression in polyp organoids, shown as mean fold-change relative to patient-matched nonpolyp organoids and ordered from highest to lowest expression. High Wnt and low Wnt denote the classification of polyp-derived lines with increased or similar/decreased LGR5 gene expression. (E) Images of polyp organoids grown for 2 passages (12 days) in WR-free media, demonstrating Wnt-independent growth (scale bars = 100 μm). (F) Growth of organoids cultured for 12 days in 100% or 60% WR media was measured through ATP-dependent luminescence. (G) Images of H87 organoids at day 12 following growth in 100% or 60% WR (scale bars = 200 μm). (H) Relative growth of nonpolyp and polyp organoids clustered by Wnt target gene expression characteristics (high Wnt and low Wnt) in 60% compared with 100% WR media. Each point represents an individual organoid line, shown as the average of triplicate wells. Data shown as mean ± SD of the organoids in that group (n = 4–9 individual organoid lines; *P < 0.05, **P < 0.005 by 1-way ANOVA with Tukey’s multiple-comparison test).

Figure 3. Intrinsic Wnt tone patterns intrapatient FGP variability.

(A) Polyp (P) and nonpolyp (NP) biopsies were obtained to establish multiple polyp-derived organoids from each patient with FAP. (B) Organoids were grown in CN media with varying concentrations of CHIR, and growth was measured through ATP-dependent luminescence on day 5. (C) Representative images of H102NPα and H102Pγ following culture in 0–3 μM CHIR (scale bars = 100 μm). (D) Relative growth of H102 organoid lines as a function of CHIR concentration. Growth was normalized to the maximal growth observed for each line, with data shown as mean ± SD of triplicate wells. (E–G) Relative growth of organoid lines from 12 patients with FAP plotted as a function of CHIR concentration. Individual organoid lines are shown in gray, and the average of all lines is shown in color. (E) Nonpolyp organoid lines (blue, n = 23 lines). (F) Polyp organoid lines demonstrating normal-like growth (pink, n = 14 lines). (G) Polyp organoid lines demonstrating enhanced Wnt-independent growth (purple, n = 21 lines). (H) Relative growth of H102 organoid lines after 5 days of growth in 0 μM CHIR (WR-free media) versus LGR5 mRNA abundance after 6 days of growth in 60% WR. The trend line was calculated via linear regression analysis. (I) Relative mRNA abundance of Wnt target genes in FAP organoids after 6 days’ growth in 60% WR. Data shown as mean ± SD fold-change relative to patient-matched nonpolyp. NP: n = 19; P-N: n = 11; P-E: n = 13 (*P < 0.05, **P < 0.005 by 1-way ANOVA with Tukey’s multiple-comparison test). (J) Average NP, P-N, or P-E organoid thickness after 6 days’ growth in 60% WR (***P < 0.001). (K) Representative images of nonpolyp, normal-like polyp, and enhanced polyp organoids after 6 days’ growth in 60% WR (scale bars = 50 μm).

Figure 4. Infrequent somatic APC mutations in FAP patient FGPs.

(A) Targeted sequencing of 19 FAP patient polyp and nonpolyp samples using a QIAGEN Human Comprehensive Cancer Panel detected 1 patient (H85) polyp with a novel somatic APC loss-of-function mutation. (B) A total of 6 patients, including H85, demonstrated APC-specific copy number variation (duplication or deletion) in their sequenced polyp DNA. (C) Copy number variations detected through sequencing across the panel of 283 target genes by patient: none = 0 CNVs (4/19), few = 1–4 (9/19), moderate = 5–25 (3/19), extensive = >26 (4/19). (D) APC mRNA sequence of normal and mutant alleles in patients H72 and H73 at the familial mutation site (c.5826_5829). The highlighted CAGA sequence is deleted in the germline mutant APC allele. (E) Chromatogram of sequenced APC cDNA harvested from H72 polyp organoids grown in 100% WR media, H72 polyp organoids grown for 2+ passages in WR-free media, and H73 polyp organoids grown for 3+ passages in WR-free media, depicting 7 bp 5′ of the mutation site and 6 bp 3′ of the mutation site. The blue underlined sequence aligned with the wild-type sequence. The red underlined sequence aligned with the mutant sequence. (F) APC mRNA sequence of the wild-type and mutant alleles in patient H87 at the familial mutation site (c.4782_4785). The highlighted AGCC sequence is deleted in the germline mutant APC allele. (G) Chromatogram of sequenced APC cDNA harvested from H87 polyp organoids grown in WR media or for 2+ passages in WR-free media. The blue underlined sequence aligned with the wild-type sequence. The red underlined sequence aligned with the mutant sequence.

Figure 5. Gastric region–specific proliferation in FAP mouse model.

(A) Adult Sox2-CreER^T2 Apc^fl/+ (heterozygous), Sox2-CreER^T2 Apc^fl/fl (homozygous), and control (Apc^fl/+ or Apc^fl/fl) mice were treated with tamoxifen (TX) to delete Apc exon 14, and tissue was harvested 1 month later. (B) Agarose gel showing PCR products amplified from genomic DNA with primers flanking a loxP site in the Apc gene to identify either the un-recombined allele (fl, 314 bp) or recombined allele (Δ, 258 bp). DNA was isolated from full-thickness corpus (C) or antral (A) tissue from control (Apc^fl/fl) mice 1 month post-TX or from homozygous Sox2CreER^T2 Apc^fl/fl mice either 48 hours or 1 month post-TX. (C and E) Representative images of corpus (C) and antral (E) tissue from control, Sox2-CreER^T2 Apc^fl/+, and Sox2-CreER^T2 Apc^fl/fl mice 1 month post-TX stained for 5-ethynyl-2′-deoxyuridine (EdU) (green) to mark proliferating cells (scale bar = 100 μm). (D and F) Morphometric quantification of proliferating EdU⁺ cells/μm of corpus (D) or antral (F) tissue in control, heterozygous (fl/+), and homozygous (fl/fl) mice 1 month post-TX. Data are presented as mean ± SEM (n = 3–9 mice per group, ***P < 0.001 by 1-way ANOVA with Tukey’s multiple-comparison test). (G and H) Size and representative images of corpus (G) or antral (H) organoids derived from control (Sox2-CreER^T2), heterozygous (fl/+), and homozygous (fl/fl) mice. Data are presented as fold-change organoid area relative to control (error bars minimum to maximum values, box length IQR, whisker outliers) (**P < 0.05, ***P < 0.005 by 1-way ANOVA with Tukey’s multiple-comparison test). Sox2-CreER^T2 (n = 3 mice; 208 corpus organoids; 217 antrum organoids), Sox2-CreERT2 Apc^fl/+ (n = 3 mice; 171 corpus organoids; 256 antrum organoids), and Sox2-CreERT2 Apc^fl/fl (n = 3 mice; 218 corpus organoids; 278 antrum organoids).