Society: AGA

Background and Aims: Ménétrier’s disease (MD) is a rare acquired protein-losing hypertrophic gastropathy characterized by giant gastric rugal folds, decreased acid secretion, increased gastric mucus production, and hypoalbuminemia due to protein loss in the gastric mucosa.Microscopically, it is characterized by massive foveolar hyperplasia, oxyntic gland atrophy (loss of parietal cell), and repatterning of cell specification. MD is caused by TGF-α overexpression in the stomach and it has been shown that the EGFR inhibiting antibody, cetuximab, can reverse the histopathologic phenotypes of MD, but failed to achieve long-term remission. In a proteomic analysis of gastric tissue of Ménétrier’s disease patients before and after cetuximab treatment, we observed decreased levels of the Notch ligand, Jagged1, after cetuximab treatment. We found that Notch signaling is activated in the stomach of MD exemplified by immunostaining for Hes1, a downstream target of Notch signaling. In addition, we observed that Jagged1 expression was localized in the chief cells in the MD mouse model stomach, suggesting EGFR signaling in the chief cell is important for the pathogenesis of MD. The aim of this study was to investigate the role of Egfr in the chief cell for the pathogenesis of MD and to optimize the therapeutics by combining inhibiting antibodies against Egfr, P1X/P2X, and a Notch inhibitor, DBZ. Methods: To examine the role of Egfr in the chief cell in MD, CMV-rtTA; tetO-TGFA mouse line was crossed with Mist1^CreERT2; Egfr^flox/flox mouse line. We knocked out Egfr in chief cells by tamoxifen treatment and then overexpressed TGF-α by treating mice with doxycycline. We counted the numbers of parietal, chief, neck, and foveolar cells by immunostaining them with their respective markers: HK-ATPase, GIF, GSII, and UEA1. We also tested the therapeutic effects of P1X/P2X, DBZ, and P1X/P2X in combination with DBZ in MT-TGF- α mice. Results: Egfr knock-out in chief cell rescued the phenotypes of MD. Parietal cells reappeared, and number of foveolar cells decreased. Also, combination therapy with P1X/P2X and DBZ was more effective in rescuing the phenotypes of MD compared to P1X/P2X or DBZ monotherapy. Conclusions: Our study shows that EGFR signaling in the chief cell is necessary for the pathogenesis of MD by inducing Jagged1 that activates Notch signaling. In addition, our data suggests that P1X/P2X and DBZ combination therapy is a promising treatment option for MD patients.

The transcription factor Sox2 and the adhesion molecule CD44 have been implicated in the development of inflammation-mediated gastric metaplasia and neoplasia. Lgr5 marks gastric progenitors that regulate epithelial homeostasis. We reported that inflammation and inhibition of Bone Morphogenetic Protein (BMP) signaling, achieved by expression of the BMP inhibitor noggin (Nog) in the corpus (H⁺/K⁺-Nog mice), and by crossing mice expressing Cre and GFP in Lgr5 cells to mice with floxed alleles of BMP receptor 1A (Lgr5-Cre;Bmpr1a^flox-flox mice), activate aberrant cells that give rise to SPEM, a pre-neoplastic metaplasia characterized by the expression of neck cell mucins, and CD44 at the base of glands of the corpus. We investigated the regulation, localization, and function of Sox2 in the context of inhibition of BMP signaling. Deletion of Sox2 and Cd44 in Lgr5 cells was achieved by crossing Lgr5-Cre mice to either Cd44^flox-flox mice or to Sox2^flox-flox mice to generate Lgr5-Cre;Cd44^flox-flox and Lgr5-Cre;Sox2^flox-flox mice. The role of Lgr5 cell-Cd44 and of Lgr5 cell-Sox2 in Nog-induced SPEM, was assessed by crossing Lgr5-Cre;Cd44^flox-flox and Lgr5-Cre;Sox2^flox-flox mice to H⁺/K⁺-Nog mice to generate Lgr5-Cre;Cd44^flox-flox;H⁺/K⁺-Nog and Lgr5-Cre;Sox2^flox-flox;H⁺/K⁺-Nog mice. Infection with H. felis (HF) was used to induce inflammation. Distribution of GFP-, Sox2-, and CD44-+ve cells was analyzed by immunostaining. Single cell RNA sequencing (scRNA-seq) was performed in both control and H⁺/K⁺-Nog mice. Cluster assignments relative to marker gene expression defined eighteen clusters. Transcriptomes of individual cells from these clusters were compared from these mice to identify differentially expressed genes by cluster. ChIP Enrichment Analysis (ChEA) was performed using the ChIP-X database to identify enriched transcription factors target genes. Sox2+ve staining was detected in the antrum but not in the corpus of wild type C57B/6 mice. In contrast, robust expression of Sox2 was observed in the presence of inhibition of BMP signaling in GFP+ve/Lgr5 cells located in the oxyntic mucosa of the lesser curvature of Lgr5-Cre;Bmr1a^flox-flox mice, an effect which was enhanced by HF. Similarly, Sox2 expression was seen in the corpus of H⁺/K⁺-Nog mice in CD44+ve cells. scRNA-seq analysis demonstrated an increase in the percentage of cells expressing Sox2 transcripts in the neck cell cluster of H⁺/K⁺-Nog mice. ChEA indicated over-representation of Sox2 target genes in the ChIP-X database of the mouse gastric mucosa highlighting the functional relevance of Sox2. Deletion of either Cd44 or Sox2 in Lgr5 cells attenuated the SPEM phenotype, underscoring the importance of these molecules in the development of metaplasia. Our findings demonstrate the significance of BMP signaling in the pathophysiology of metaplasia and in the regulation of Sox2 in the stomach.

Gastric carcinogenesis is correlated with the presence of metaplasia and its progression. Therefore, understanding the events that lead to metaplasia progression remains a priority. Pyloric metaplasia characterized by the presence of spasmolytic polypeptide-expressing metaplasia (SPEM) cells is the earliest type of metaplasia seen in the stomach after parietal cell loss. Using transgenic Mist1-Kras mice, we have established murine SPEM gastroid lines (Meta1) from metaplastic glands isolated at 1 month after induction of active KRas (G12D). Meta1 gastroid lines maintain their SPEM cell features across multiple passages in vitro. Moreover, single cell RNA-sequencing data revealed that Meta1 have a distinctive transcriptional profile in comparison with mouse dysplastic gastroids (Meta4), established previously. Recent investigations from our laboratory have shown an increased number of intrinsic immune cells, type 2 innate lymphoid cells (ILC2s), in pyloric metaplasia tissues of both human and mouse. ILC2s are known to release cytokines including IL-13 and we demonstrated that IL-13 could promote SPEM lineage cell development in mice. We therefore studied the effects of IL-13 in normal gastroids by treatment with either recombinant IL-13 or 0.1% BSA. IL-13-treated gastroids showed an increase in the proportion of MUC6-, CD44v9- or AQP5-positive SPEM cells (Fig 1A). Immunoblotting showed up-regulation of AQP5 and STAT6 phosphorylation (pSTAT6) in normal gastroids after the IL-13 treatment. We also observed that co-culture of Meta1 with isolated mouse ILC2s promoted gastroid growth by increasing the number of TFF2- and AQP5-co-expressing SPEM cells and proliferative Ki67-positive cells. Immunoblotting also showed upregulation of STAT6 phosphorylation in the co-culture condition. The IL-13 also increased in organoid size of Meta1 gastroids. MUC6-positive SPEM cells and proliferative Ki-67-positive cells were also increased in IL-13-treated Meta1 cells. Immunoblotting showed pSTAT6 in Meta1 gastroids treated with IL-13 in a dose-dependent manner; and RT-qPCR revealed a significant increase in the expression of SPEM-related genes. We additionally observed that IL-13 receptor a-1 (IL-13Ra-1) is expressed in the basolateral membranes of epithelial cells from the stomach corpus at 1 month after induction of active KRas (G12D) (Fig 1B). The expression is consistently observed in the stomachs with intestinal metaplasia (IM) and dysplasia at 3 and 4 months after induction, respectively. When IM and dysplastic gastroids (Meta3 and Meta4, respectively) were treated with either IL-13 or BSA, we observed that IL-13 treatment caused the phosphorylation of STAT6 (Fig 1C). These results suggest that IL-13 could play important roles in different stages of the stomach carcinogenesis.

<b>Figure 1. IL-13 effects in mouse gastroids. A)</b> Left panel, Immunofluorescence staining for SPEM markers (Muc6 and AQP5) in normal gastroids, showing an increase of SPEM cell marker in IL-13-treated gastroids. Right panel, Immunoblotting showing an up-regulation of AQP5 and phosphorylated STAT6 after treatment with IL-13. <b>B)</b> Immunofluorescence staining for IL-13Ra1 and P120 (epithelial cell marker) in Mist1-Kras mouse stomach tissues after 1, 3 or 4 months of induction. <b>C)</b> Immunoblotting showing an up-regulation of phosphorylated STAT6 after treatment with IL-13 in Meta3 (IM) and Meta4 (dysplasia) gastroids.

Figure 1. IL-13 effects in mouse gastroids. A) Left panel, Immunofluorescence staining for SPEM markers (Muc6 and AQP5) in normal gastroids, showing an increase of SPEM cell marker in IL-13-treated gastroids. Right panel, Immunoblotting showing an up-regulation of AQP5 and phosphorylated STAT6 after treatment with IL-13. B) Immunofluorescence staining for IL-13Ra1 and P120 (epithelial cell marker) in Mist1-Kras mouse stomach tissues after 1, 3 or 4 months of induction. C) Immunoblotting showing an up-regulation of phosphorylated STAT6 after treatment with IL-13 in Meta3 (IM) and Meta4 (dysplasia) gastroids.

Introduction
Gastric cancer develops through a cascade from precancerous metaplasia to dysplasia and adenocarcinoma. We have previously developed the Mist1-Kras mouse model recapitulating the carcinogenic cascade and reported that pyrvinium pamoate (pyrvinium) is highly effective to induce cell death in dysplastic cells. Here, we aimed to define specific targets of pyrvinium in dysplastic cells and further explore the effects in gastric pre-cancer organoid lines.
Method
Gastric organoid lines established from the Mist1-Kras were utilized to evaluate effects of Pyrvinium as well as MEK inhibitor and STAT3 inhibitors. For in vivo assay, pyrvinium was administered to the Mist1-Kras mice 3 months after tamoxifen injection. The mice were sacrificed at 1 and 2 weeks after the drug treatment and immunostaining in stomach tissue sections for various cell lineage markers were performed. Twenty human pre-cancer organoid lines were established from the surgically resected gastric tissues and characterized based on morphology and metaplastic or dysplastic marker expression.
Result
Pyrvinium induced cell death in dysplastic organoids by inhibiting both STAT3 and ERK phosphorylation, whereas it showed no cytotoxicity to metaplastic organoids and only suppressed organoid growth. Trametinib, a MEK inhibitor which downregulates ERK phosphorylation, only led to a slight decrease in proliferation in both metaplastic and dysplastic organoids without enhancing cell death. Western blot analysis demonstrated that the basal phospho-STAT3 level was much higher in dysplastic cells than in metaplastic ones and the cytotoxic effect of STAT3 inhibitors exerted a similar effect on dysplastic organoids compared to pyrvinium. These findings imply that pyrvinium's effect on dysplastic organoids is mediated by targeting abnormally enhanced STAT3 activity. Furthermore, the Mist1-Kras mice treated with pyrvinium showed a significantly reduced number of dysplastic glands along with increased normal gastric cells and a reduced proliferation index, indicating that pyrvinium has an ability to prevent metaplasia to dysplasia progression and reverse the metaplasia to normal glands in the stomach. We also found that responses of 20 human gastric pre-cancer organoid lines to pyrvinium were significantly correlated with dysplastic morphology and characteristics, including high TROP2, and low CD44v9 expression.
Conclusion
Pyrvinium induces cell death in dysplastic cells by blocking STAT3 signaling pathway and controls growth in both dysplastic and metaplastic organoids by inhibiting the MAPK pathway. Therefore, our study suggests that Pyrvinium has dual effects on metaplastic and dysplastic cells by blocking MAPK and STAT3 signaling pathways and is a promising candidate drug for not only eliminating dysplastic cells but also preventing metaplasia to dysplasia progression in the stomach.

Gastric cancer develops through a series of glandular pre-cancerous lesions in the stomach mucosa. The first stage is pyloric metaplasia which is developed by the transdifferentiation of gastric chief cells into spasmolytic polypeptide-expressing metaplastic (SPEM) cells. These SPEM cells are then seen throughout the final two pre-cancerous stages, intestinal metaplasia and dysplasia. However, it is currently unclear what roles they play in the progression of metaplasia during carcinogenesis. We have previously observed that SPEM cells display strong expression of the transcription factor Sox9 in both human and murine tissues of gastric pre-cancer. Sox9 is expressed in adult intestinal stem cells and is believed to control stem cell dynamics. We therefore hypothesize that Sox9 may act to control proliferation and differentiation in SPEM cells and regulate metaplasia progression.

To investigate this question, we utilized a mouse model (GCK) expressing activated Kras under the chief cell-specific driver (GIF) with a doxycycline-inducible rtTA-Cre system, which drives gastric carcinogenesis. Using the GCK mouse allele, we generated a new mouse allele which deletes the Sox9 gene only in gastric chief cells (GCK-Sox9KO) and performed immunostaining for cell lineage markers to examine cellular and morphological changes in stomach mucosa.

At three weeks after Kras-induction, the GCK mice displayed pyloric metaplasia in nearly 100% of their glands which can be identified by a population of CD44v9+/AQP5+/Sox9+ SPEM cells in the base of the gland, loss of parietal cells in the neck, and hyperplasia of surface cells. We observed that glands contained CD44v9+ SPEM cells without Sox9 expression in the GCK-Sox9KO mouse stomachs, suggesting that the Sox9 upregulation might not be necessary for the development of SPEM cells. At six weeks after the induction, the GCK mice show intestinal metaplasia featuring TFF3-expressing intestinal metaplastic (IM) cells in the neck of the gland, and SPEM cells at the base of the gland. However, GCK-Sox9KO mice displayed a significant reduction in intestinal metaplasia and the appearance of a previously unobserved gland type that has parietal cells re-populating the base of the gland instead of SPEM cells. By ten weeks, the majority of glands in the GCK-Sox9KO mouse stomachs recovered to normal types of gastric glands containing both chief cells and parietal cells, while the glands in the GCK models progressed to dysplasia with no chief cells or parietal cells.

These results show a significant disruption in the progression of metaplasia by the loss of Sox9 in SPEM cells and a possible recovery phenotype after failure of pyloric metaplasia to progress to more advanced stages. Therefore, this study suggests that Sox9 in SPEM cells may control the cell lineage progression in gastric carcinogenesis.

Alexis Guenther

Appearances