You are here

Article Text

Article menu

Download PDFPDF

Stomach
Original article
A signalling cascade of IL-33 to IL-13 regulates metaplasia in the mouse stomach
  1. Christine P Petersen1,2,
  2. Anne R Meyer1,2,
  3. Carlo De Salvo3,
  4. Eunyoung Choi2,4,5,
  5. Cameron Schlegel2,5,
  6. Alec Petersen2,
  7. Amy C Engevik2,5,
  8. Nripesh Prasad6,
  9. Shawn E Levy6,
  10. R Stokes Peebles7,
  11. Theresa T Pizarro3,
  12. James R Goldenring1,2,4,5
  1. 1Departments of Cell and Developmental Biology, Vanderbilt University, Nashville, Tennessee, USA
  2. 2Epithelial Biology Center, Vanderbilt University, Nashville, Tennessee, USA
  3. 3Department of Pathology, Case Western Reserve School of Medicine, Cleveland, Ohio, USA
  4. 4Nashville VA Medical Center, Vanderbilt University, Nashville, Tennessee, USA
  5. 5Department of Surgery, Vanderbilt University, Nashville, Tennessee, USA
  6. 6HudsonAlpha Institute for Biotechnology, Huntsville, Alabama, USA
  7. 7Department of Medicine, Vanderbilt University, Nashville, Tennessee, USA
  1. Correspondence to Dr James R Goldenring, Epithelial Biology Center and Section of Surgical Sciences, Vanderbilt University Medical Center, 10435-G MRB IV, 2213 Garland Avenue, Nashville, TN 37232, USA; jim.goldenring{at}vanderbilt.edu

Abstract

Objective Alternatively activated macrophages (M2) are associated with the progression of spasmolytic polypeptide-expressing metaplasia (SPEM) in the stomach. However, the precise mechanism(s) and critical mediators that induce SPEM are unknown.

Design To determine candidate genes important in these processes, macrophages from the stomach corpus of mice with SPEM (DMP-777-treated) or advanced SPEM (L635-treated) were isolated and RNA sequenced. Effects on metaplasia development after acute parietal cell loss induced by L635 were evaluated in interleukin (IL)-33, IL-33 receptor (ST2) and IL-13 knockout (KO) mice.

Results Profiling of metaplasia-associated macrophages in the stomach identified an M2a-polarised macrophage population. Expression of IL-33 was significantly upregulated in macrophages associated with advanced SPEM. L635 induced metaplasia in the stomachs of wild-type mice, but not in the stomachs of IL-33 and ST2 KO mice. While IL-5 and IL-9 were not required for metaplasia induction, IL-13 KO mice did not develop metaplasia in response to L635. Administration of IL-13 to ST2 KO mice re-established the induction of metaplasia following acute parietal cell loss.

Conclusions Metaplasia induction and macrophage polarisation after parietal cell loss is coordinated through a cytokine signalling network of IL-33 and IL-13, linking a combined response to injury by both intrinsic mucosal mechanisms and infiltrating M2 macrophages.

  • CELLULAR IMMUNOLOGY
  • CYTOKINES
  • GASTRIC PRE-CANCER
  • GASTRITIS
  • MACROPHAGES

https://doi.org/10.1136/gutjnl-2016-312779

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Significance of this study

    What is already known on this subject?

    • Gastric cancer evolves in a field of pre-cancerous metaplasia.

    • Spasmolytic polypeptide-expressing metaplasia (SPEM) and intestinal metaplasia in the human stomach are associated with chronic inflammation.

    • Mouse models of SPEM have shown that M2-macrophages promote the progression of metaplasia towards a more proliferative and advanced phenotype.

    What are the new findings?

    • Transcriptional profiling of isolated metaplasia-associated gastric mucosal macrophages in the stomach shows that they have predominantly M2a phenotype.

    • The primarily epithelial-associated cytokine interleukin (IL)-33 is also expressed in M2 macrophages.

    • Mice that are null for IL-33, the IL-33 receptor (ST2) or IL-13 do not develop metaplasia following acute parietal cell loss.

    • IL-13 treatment in ST2 null mice restored the development of SPEM following acute parietal cell loss, thus identifying a novel pathway required for metaplasia development.

    How might it impact on clinical practice in the foreseeable future?

    • The results suggest that intrinsic mucosal and macrophage cytokines are responsible for the induction of metaplasia in the atrophic stomach. Thus, antagonism against IL-33 and IL-13 may represent viable approaches for reversal of metaplastic changes in the stomach.

    Introduction

    Gastric cancer is a leading cause of cancer-related deaths worldwide.1 Intestinal-type gastric cancer, the most abundant type of gastric adenocarcinoma, is frequently detected at later stages due to poor early clinical indicators. Helicobacter pylori infection is the most common risk factor for developing gastric adenocarcinoma by inducing a chronic inflammatory environment. Helicobacter infection results in the loss of acid-secreting parietal cells (oxyntic atrophy) within the corpus of the stomach, leading to the evolution of metaplasia in the gastric epithelium.2 Two different types of metaplasia are typically observed in the atrophic human stomach: spasmolytic polypeptide-expressing metaplasia (SPEM) and intestinal metaplasia.3 SPEM arises from mature chief cells at the base of the gland that transdifferentiate into mucus-expressing and TFF2-expressing cell lineages.3 Under chronic inflammatory conditions, increasing evidence suggests that SPEM gives rise to intestinal metaplasia.4–6

    Little is known about the molecular pathways that induce or drive the progression of metaplasia towards gastric cancer. It is difficult to study the normal pathophysiological process of gastric cancer development because mice do not typically develop goblet cell intestinal metaplasia or gastric cancer after extended Helicobacter infection.7 Instead, metaplasia advances towards an intestinalised phenotype with increased proliferation and elevated levels of intestinal gene expression, which we have defined as advanced SPEM.8 Two drugs, DMP-777 and L635, have previously been used in our lab to induce acute oxyntic atrophy and study the progression of metaplasia in mice.9 ,10 DMP-777 is a parietal cell protonophore and neutrophil elastase inhibitor that causes oxyntic atrophy and metaplasia with minimal inflammation in the stomach after 10 days of treatment.9 L635, an analogue of DMP-777, causes oxyntic atrophy without inhibiting the inflammatory response.3 Mice treated with L635 for 3 days develop a prominent inflammatory infiltrate that results in proliferative SPEM with increased expression of intestinal transcripts (advanced SPEM) that closely resembles lineage changes observed after 6–12 months of Helicobacter felis infection.8

    We recently reported that macrophages are necessary for the progression of metaplasia to a more advanced SPEM phenotype.11 Macrophage-depleted L635-treated mice developed fewer proliferative SPEM cells and less intestinal transcripts compared with control L635-treated mice. Thus macrophages that infiltrate into the stomach after parietal cell loss promote the advancement of metaplasia.11 In addition, macrophages infiltrating into the mucosa after L635 treatment displayed M2-like markers.11 M2 macrophages are characterised as anti-inflammatory, tumour-associated cells driven by Th2 cytokines.12 However, macrophages also specialise in response to local environmental stimuli, and therefore have diverse gene expression and function within different organ systems.13 In addition, the generalised M1 and M2 classification, which relies on a limited number of markers, is overly simplified.14 Thus an in-depth characterisation of metaplasia-associated macrophages is necessary to identify key inflammatory mediators that promote the induction and progression of metaplasia in the stomach.

    We have now performed RNA sequencing of macrophages associated with SPEM (DMP-777-treated mice) or advanced SPEM (L635-treated mice) from the stomach corpus to generate a profile of metaplasia-associated macrophages. We found that IL-33 is significantly upregulated in M2 macrophages from advanced metaplasia. Using IL-33 and IL-33 receptor (ST2) knockout mice, we determined that an IL-33 signalling pathway is required for the development of metaplasia after parietal cell loss. Downstream pathway analysis confirmed that IL-33 is necessary for Th2 cytokine induction in the stomach. Investigations to determine the role of IL-5, IL-9 and IL-13 on metaplasia development revealed that IL-13 is necessary and sufficient for chief cell transdifferentiation into metaplasia following parietal cell loss. These findings reveal a multicellular signalling pathway necessary to induce metaplasia in the stomach after parietal cell loss and mucosal injury.

    Results

    Analysis of the macrophage transcriptome from acute SPEM models

    We have previously determined that M2-like macrophages are present in the corpus in association with advanced SPEM that demonstrated elevated intestinal gene expression, and increased proliferation, determined by increased nuclear Ki-67 staining.11 Therefore, we sought to determine the characteristics of this macrophage population to identify factors that induce or promote the progression of metaplasia. F4/80-positive cells from the corpus of L635-treated mice and DMP-777-treated mice were isolated and RNA sequenced for transcriptome analysis (see online supplementary figure S1). The purity of the macrophage-enriched cell populations was established using the normalised expression values from the RNA sequencing (figure 1A). Macrophage-specific genes (Cd68, Cd11b, Cd11c) were enriched as compared with other immune (Cd3e, Ly6c, Ly6g, Prg2/Mbp, Cma1) or gastric epithelial cell populations (Dclk1, Chga, Cckbr, Wfdc2, Mist1). We assessed macrophage polarisation markers and determined that macrophages associated with advanced SPEM demonstrated more prominent upregulation of M2 polarisation-associated markers (Fizz1, Ym1, Klf4, Alox15, Il33, Mrc1, Ccl2) as opposed to canonical M1-associated markers (Il6, Il12, Nos2) (figure 1B). However, there was still prominent expression of some M1-associated genes (eg, Klf6, Irf5 and Il18). Previous studies relied upon only a few key markers to determine macrophage polarisation. Our sequencing data confirms known polarisation markers and provides a more detailed analysis that illustrates that macrophages associated with advanced metaplasia may represent a subclass of M2a polarised macrophages (Ccl17). Our analysis concentrated primarily on secreted factors or receptors, as macrophages impact the local environment predominantly through paracrine and juxtacrine signalling.15 ,16 Genes fitting this description that were significantly upregulated in L635-associated macrophages were organised into groups: macrophage genes (Csf1, Syk), cytokines (Il33, ll4), chemokines (Ccl24, Cxcl3), matrix remodelling enzymes (Mmp25, Mmp9), growth factors (Hgf, Areg, VegfA) and cell adhesion molecules (Itgb2, Itga4) (figure 1C).

    Figure 1
    Figure 1

    Transcriptome analysis of an enriched macrophage population associated with metaplasia in the stomach. Gene expression profile of macrophages from metaplasia was determined for three replicates/each sample (A) Expression values of macrophage genes (Cd68, Cd11b, Fn1), other immune cell types (Cd3e: T-cells, Ly6c: neutrophils, Prg2: eosinophils) and gastric epithelial cell lineages (Dclk1: tuft cells, Chga: endocrine cells, Cckbr: parietal cells, Mist1: chief cells) demonstrate a population enriched for macrophages. (B) Expression levels of M1 (blue) and M2 (red)-associated genes, in macrophages isolated from advanced spasmolytic polypeptide-expressing metaplasia (SPEM) (L635-treated mice) plotted with SEM. Macrophages associated with advanced SPEM express increased M2-like transcripts (Fizz1, Klf4, Il33) compared with M1 genes (Nos2, Ifng). (C) Factors significantly upregulated in macrophages from advanced SPEM (L635-treated mice) (p<0.01; fold change ≥2 compared with macrophages isolated from DMP-777-treated mice). Representative genes selected were grouped based on function.

    IL-33 and the IL-33 receptor (ST2) are necessary for SPEM induction

    Based on our RNA sequencing results, the IL-1 family member IL-33 was the most significantly upregulated cytokine in macrophages associated with an advanced SPEM from L635-treated mice. IL-33 expression is localised to surface mucus cells in the normal stomach.21 However after parietal cell loss, macrophages expressing IL-33 are present within the corpus mucosa (see online supplementary figure S2). No IL-33 staining was observed in either mucosal cells or macrophages in IL-33-deficient mice (see online supplementary figure S2). Western blot analysis showed elevations in IL-33 expression in gastric mucosa from mice treated with L635 and mice infected with H. felis for 6 months (see online supplementary figure S2C). As observed in L635-treated mice, H. felis-infected mice showed expression of IL-33 in foveolar cells as well as in macrophages (see online supplementary figure 2D).

    To investigate the role of IL-33 in the initiation and progression of SPEM, we treated IL-33 knockout (IL33KO) mice with L635 to induce advanced metaplasia (figure 2⇓). Consistent with previous studies, acute parietal cell loss in wild-type mice resulted in the loss of the mature chief cell transcription factor, Mist1, in chief cells located at the base of the glands.17 While Mist1-positive chief cells also significantly decreased after L635 treatment in IL33KO mice, there were significantly more Mist1-positive chief cells remaining at the base of glands compared with wild-type L635-treated mice (figure 2C). SPEM is characterised by lineages co-expressing Muc6 (GSII-lectin), CD44 variant 9 (Cd44v9) and gastric intrinsic factor (GIF).18 The number of GIF-only positive cells (chief cells) and cells co-positive for GIF, GSII and Cd44v9 (SPEM cells) was quantified to determine the presence and amount of SPEM development. In wild-type mice, L635-treatment induced GIF-staining cells co-expressing GSII and CD44v9, indicative of SPEM. L635-treated IL33KO mice, showed significantly fewer SPEM cells as compared with wild-type L635-treated mice (figure 2D). Further, a hallmark of inflammation-associated advanced metaplasia in the stomach is the development of a highly proliferative SPEM, visualised by cells triple-positive for GIF, Ki67 and GSII-lectin. In L635-treated wild-type mice, approximately 45% of SPEM cells were proliferative, indicative of advanced SPEM.8 However, IL33KO L635-treated mice had almost no proliferative metaplastic cells (figure 2E).

    Figure 2
    Figure 2

    Interleukin (IL)-33 is necessary for induction of metaplasia and polarisation of M2 macrophages. (A) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker Griffonia simplicifolia lectin II (GSII-lectin), chief cell granule marker gastric intrinsic factor (GIF), spasmolytic polypeptide-expressing metaplasia (SPEM) marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker diamidino-2-phenylindol (DAPI) in untreated wild-type and L635-treated wild-type mice. (B) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker GSII-lectin chief cell granule marker gastric intrinsic factor (GIF), SPEM marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker DAPI in untreated IL-33 knockout (IL33KO) and L635-treated IL33KO mice. (C) Quantification of Mist1-positive nuclei per 20× field. L635-treated IL33KO mice had significantly fewer Mist1-positive cells compared with untreated IL33KO mice, and significantly more Mist1-positive cells compared with wild-type L635-treated mice. (D) Quantification of SPEM cells determined by per cent of GIF-positive cells co-positive for CD44v9 and GSII-lectin. L635-treated IL33KO mice have significantly fewer SPEM cells compared with wild-type L635-treated mice. (E) Quantification of proliferative SPEM cells determined by the per cent of SPEM cells positive for Ki67. L635-treated IL33KO mice have significantly fewer proliferative SPEM cells compared with wild-type L635-treated mice. (F) Quantification of F4/80-positive macrophages (red) and F4/80 and CD163 co-positive M2 macrophages (green) per 20× field. L635-treated IL33KO mice had similar F4/80-expressing macrophage infiltration as wild-type L635-treated mice, but significantly fewer F4/80 macrophages were co-positive for CD163. *Denotes significant difference (p<0.05) compared with untreated mice, + denotes significant difference (p<0.05) compared with L635-treated wild-type mice. White dashed boxes are magnified in the insets. Scale bars=100 μm.

    Figure 2
    Figure 2

    Continued

    After acute parietal cell loss in wild-type L635-treated mice, F4/80-positive macrophages infiltrate into the corpus mucosa.11 To determine if IL-33 is required for macrophage recruitment after parietal cell loss, F4/80-positive cells were counted in IL33KO L635-treated mice. No significant difference in macrophage infiltration was observed in IL33KO L635-treated mice compared with wild-type L635-treated mice (figure 2F). Our RNA sequencing data showed macrophages associated with advanced metaplasia have a M2 polarised phenotype (figure 1B). Using co-labelling of macrophage marker F4/80 with the M2 macrophage marker CD163, we found that macrophages in L635-treated IL33KO mice had a significant reduction in M2 macrophages compared with L635-treated wild-type mice (figure 2F). We confirmed a significant loss of M2 macrophage polarisation in L635-treated IL33KO mice using quantitative real-time PCR for the M2 marker, Ym1 (see online supplementary figure S3).

    We also examined the effects of IL-33 receptor (ST2) deletion (ST2KO) on the evolution of SPEM induced by L635 treatment. Figure 3 demonstrates that ST2KO mice showed a similar pattern to IL33KO mice. L635-treated ST2KO mice have decreased Mist1 expression (figure 3B), loss of SPEM (figure 3C) and reduced proliferative SPEM (figure 3D). In addition, macrophages infiltrating into the mucosa of L635-treated ST2KO mice did not express the M2 marker CD163 (figure 3E). L635-treated ST2KO mice also had reduced Ym1 expression (see online supplementary figure S3). These data suggest a role for IL-33 signalling in the initiation of metaplasia and M2 macrophage polarisation.

    Figure 3
    Figure 3

    Interleukin (IL)-33 signalling drives spasmolytic polypeptide-expressing metaplasia (SPEM) development and M2 macrophage polarisation. (A) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker Griffonia simplicifolia lectin II (GSII-lectin), chief cell granule marker GIF, SPEM marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker diamidino-2-phenylindol (DAPI) in untreated ST2 (IL33 receptor) knockout mice and L635-treated ST2 knockout mice. (B) Quantification of Mist1-positive nuclei per 20× field. L635-treated ST2 knock out (ST2KO) mice had significantly fewer Mist1-positive cells compared with untreated ST2KO mice, although some Mist1-positive cells remained. (C) Quantification of SPEM cells determined by per cent of GIF-positive cells co-positive for CD44v9 and GSII-lectin. L635-treated ST2KO mice did not have significant changes in SPEM cell number compared with untreated ST2KO mice. (D) Quantification of proliferative SPEM cells determined by the per cent of SPEM cells positive for Ki67. Although L635-treated ST2KO had very few SPEM cells, some of the SPEM cells were proliferating so there is a significantly higher per cent of proliferative SPEM cells in L635-treated ST2KO mice compared with untreated ST2KO mice. (E) Quantification of F4/80-positive macrophages (red) and F4/80 and CD163 co-positive M2 macrophages (green) per 20× field. L635-treated ST2KO mice have significant F4/80-positive macrophage infiltration compared with untreated ST2KO mice. L635-treated ST2KO mice also have significantly more F4/80 and CD163 co-positive cells compared with untreated ST2KO mice; however, the per cent of F4/80 macrophages that are co-positive for CD163 is around 10% in both mouse models. *denotes significant difference (p<0.05) compared with untreated mice. White dashed boxes are magnified in the insets. Scale bars=100 μm.

    Loss of Th2 cytokines correlates with the absence of M2a polarised macrophages in L635-treated IL33KO mice

    Using a mouse-specific cytokine and chemokine array (Qiagen, PAMM-150Z), we examined the impact that IL-33 loss has on downstream cytokines and chemokines after parietal cell loss. Whole stomach corpus from wild-type L635-treated mice and IL33KO L635-treated mice was assessed for 84 different chemokine and cytokine transcripts. Seven genes had significant differential expression (p<0.05) between wild-type L635-treated mice and IL33KO L635-treated mice: Ccl11, Ccl24, Ccl17, Il4, Il5, Il9 and Il13 (figure 4A, B). The genes Ccl11, Ccl24 and Ccl17 are characteristic of an M2a macrophage polarised environment, while Il4, Il5, Il9 and Il13 are Th2-related cytokines.19 Thus, L635-treated IL33KO mice had a significant reduction in M2a macrophage polarising chemokines and Th2-related cytokines compared with L635-treated wild-type mice.

    Figure 4
    Figure 4

    Altered cytokine response in interleukin (IL)-33 knockout (KO) mice. Analysis of 84 mouse chemokines and cytokines in whole stomach corpus RNA from L635-treated wild-type mice and L635-treated IL-33 knockout (IL33KO) mice (n=3). (A) Alternatively activated macrophage-related chemokines (Ccl11, Ccl24 and Ccl17) were significantly decreased in L635-treated IL33KO mice compared with L635-treated wild-type mice. (B) Th2-related M2 macrophage polarising cytokines were significantly decreased in L635-treated IL33KO mice compared with L635-treated wild-type mice. *p<0.05.

    IL-5 and IL-9 do not impact metaplasia development or macrophage polarisation

    IL-33 is known to induce the expression of Th2 cytokines in the stomach, lungs and intestine, and drive alternative activation of macrophages (M2).20–23 However, little is understood about the specific influence of Th2 cytokines on metaplasia induction or progression in the stomach. IL-5 represents a major regulator of eosinophil activation. Eosinophils are present in the overall inflammatory milieu in the stomach mucosa during Helicobacter infection and are also implicated in GI eosinophilia.24 ,25 While the F4/80 antibody is predominately used as a monocyte/macrophage and dendritic cell marker, it is also expressed at low levels on eosinophil granules.26 Major basic protein (Prg2), an eosinophil granule protein,27 and Ccr3, a chemokine receptor highly expressed on eosinophils,19 ,28 were upregulated over sixfold in the F4/80+ cells isolated from L635-treated mice (see online supplementary table S1). DMP-777 and L635-treated mice have a similar robust eosinophil infiltration into the mucosa during SPEM induction (see online supplementary figure S4A). We have previously identified macrophages as necessary for the progression to advanced SPEM,11 and re-examination of L635-treated mice also treated with clodronate to deplete macrophages revealed fewer eosinophils compared with control L635-treated mice (see online supplementary figure S4B). Thus, macrophages impact eosinophil recruitment to the stomach after parietal cell loss.

    supplementary table

    Alterations in mRNA expression in F4/80 positive macrophages isolated from DMP-777 and L635-treated mouse stomachs.

    Since IL-5 and IL-9 were significantly decreased in L635-treated IL33KO mice, we investigated the role of these Th2 cytokines in metaplasia induction in the stomach corpus. To this end, IL-5 or IL-9 was depleted from mice followed by L635 administration. We evaluated IL-5 and IL-9-depleted mice as performed for IL33KO and ST2KO mice (figure 5). Immunofluorescence staining for Mist1 expression was conducted in the following L635-treated groups: non-specific IgG-treated, anti-IL-5-treated and anti-IL-9-treated wild-type mice. We observed the loss of Mist1-expressing cells at the bases of the glands in non-specific IgG L635-treated wild-type mice. The loss of Mist1 was also observed in L635-treated anti-IL-5 and L635-treated anti-IL-9 mice (figure 5B). The presence of SPEM was evaluated by quantitation of the number of cells triple-positive for GSII-lectin, CD44v9 and GIF. We observed that L635-treated mice depleted of IL-5 or IL-9 developed SPEM similar to non-specific IgG L635-treated wild-type mice (figure 5C). To determine whether SPEM was proliferative, we evaluated the number of Ki67 positive SPEM cells in each mouse model. L635-treated anti-IL-5 and anti-IL-9-treated mice did not have a significant difference in proliferative SPEM cells compared with non-specific IgG L635-treated wild-type mice (figure 5D). These results demonstrate that neither IL-5 nor IL-9 is directly necessary for the development of SPEM. Furthermore, we examined the presence of eosinophils in L635-treated IL-5-depleted mice (see online supplementary figure S4C) and found an absence of eosinophils in the mucosa. Thus, eosinophils are not required for the induction of SPEM as IL-5-depleted L635-treated mice develop SPEM normally.

    Figure 5
    Figure 5

    Interleukin (IL)-5 and IL-9 are not required for spasmolytic polypeptide-expressing metaplasia (SPEM) development and M2 macrophage polarisation in the atrophic stomach. (A) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker Griffonia simplicifolia lectin II (GSII-lectin), chief cell granule marker gastric intrinsic factor (GIF), SPEM marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker diamidino-2-phenylindol (DAPI) in wild-type mice treated with non-specific IgG or the following L635 treated mice: non-specific IgG, anti-IL-5 and anti-IL-9. (B) Quantification of Mist1-positive nuclei per 20× field. IL-5 or IL-9-depleted L635-treated mice had significant loss of Mis1-positive cells similar to non-specific IgG L635-treated mice. (C) Quantification of SPEM cells determined by per cent of GIF-positive cells co-positive for CD44v9 and GSII-lectin. IL-5 or IL-9-depleted L635-treated mice have a significant per cent of SPEM cells similar to non-specific IgG L635-treated mice. (D) Quantification of proliferative SPEM cells determined by the per cent of SPEM cells positive for Ki67. IL-5 or IL-9-depleted mice have a significant per cent of proliferating SPEM cells similar to non-specific IgG L635-treated mice. (E) Quantification of F4/80-positive macrophages (red) and F4/80 and CD163 co-positive M2 macrophages (green) per 20× field. IL-5 or IL-9-depleted mice had significant infiltration of F4/80-positive macrophages, most of which were co-positive for CD163 similar to non-specific IgG L635-treated mice. *denotes significant difference (p<0.05) compared with non-specific IgG-treated mice without L635-treatment. White dashed boxes are magnified in the insets. Scale bars=100 μm.

    Th2 cytokines promote the alternative activation of M2-polarised macrophages.29 To determine the impact that loss of specific Th2 cytokines had on the polarisation of infiltrating macrophages, we assessed the macrophage populations in non-specific IgG, anti-IL-5 and anti-IL-9 L635-treated mouse models. The number of F4/80-positive infiltrating macrophages was quantitated and we did not detect a significant difference in the number of F4/80-positive macrophages present after L635 treatment between either of the Th2 cytokine depleted mouse models (figure 5E). M2 macrophage polarisation was determined through dual immunolabelling for F4/80 and CD163 from whole stomach corpus. We observed that almost all macrophages in non-specific IgG L635-treated control mice were co-positive for macrophage marker F4/80 and M2 macrophage marker CD163. Furthermore, both L635-treated anti-IL-5 and anti-IL-9-treated mice had similar high numbers of CD163-positive macrophages, and M2-macrophage polarisation after L635 treatment was confirmed by high expression of Ym1 (see figure 5E and online supplementary figure S3).

    IL-13 is necessary for M2 macrophage polarisation and SPEM development in the stomach

    We next evaluated the role of IL-13 as a putative Th2 cytokine influencing the polarisation of M2 macrophages and SPEM development in the stomach corpus. We examined the ability of L635 to induce SPEM and macrophage polarisation in IL-13 deficient (IL13KO) mice (figure 6). Untreated IL13KO mice demonstrated a normal compendium of chief cells. L635 induced parietal cell loss similar to that observed in wild-type mice (data not shown). Following L635-treatment, there was a reduction in Mist1-positive chief cells in IL13KO mice; however, IL13KO mice maintained some Mist1-positive chief cells at the gland bases (figure 6B). In addition, L635-treated IL13KO mice exhibited few triple-positive GIF+GSII-lectin+CD44v9+ SPEM cells (figure 6C) and very few proliferative SPEM cells (figure 6D). L635-treated IL13KO mice also showed few CD163 and F4/80 co-positive macrophages consistent with low expression of Ym1 (see figure 6E and online supplementary figure S3). Together, these data implicate a role for IL-13 in the polarisation of macrophages and the progression of metaplasia following parietal cell loss.

    Figure 6
    Figure 6

    Interleukin (IL)-13 signalling downstream of IL-33/ST2 drives metaplasia induction and M2 macrophage polarisation. (A) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker Griffonia simplicifolia lectin II (GSII-lectin), chief cell granule marker gastric intrinsic factor (GIF), spasmolytic polypeptide-expressing metaplasia (SPEM) marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker diamidino-2-phenylindol (DAPI) in untreated IL13KO mice and L635-treated IL13KO mice. (B). Quantification of Mist1 positive nuclei per 20x field. L635-treated IL13KO mice had significantly fewer Mist1-positive cells compared to untreated IL13KO mice, although some Mist1-positive cells remained. (C) Quantification of SPEM cells determined by per cent of GIF-positive cells co-positive for CD44v9 and GSII-lectin. L635-treated IL13KO mice have a significantly higher per cent of SPEM cells; however, this number is still relatively low. (D) Quantification of proliferative SPEM cells determined by the per cent of SPEM cells positive for Ki67. L635-treated IL13KO had a few SPEM cells and some of the SPEM cells were proliferating so there is a significantly higher per cent of proliferative SPEM cells in L635-treated IL13KO mice compared with untreated IL13KO mice. (E) Quantification of F4/80-positive macrophages (red) and F4/80 and CD163 co-positive M2 macrophages (green) per 20× field. L635-treated IL13KO mice have significant F4/80-positive macrophage infiltration compared with untreated IL13KO mice. L635-treated IL13KO mice also have significantly more F4/80 and CD163 co-positive cells compared with untreated IL13KO mice; however, the per cent of F4/80 macrophages that are co-positive for CD163 is around 15% in both mouse models. *denotes significant difference (p<0.05) compared with untreated mice. White dashed boxes are magnified in the insets. Scale bars=100 μm.

    IL-13 is sufficient to induce metaplasia following parietal cell loss in the absence of IL33 signalling

    IL-13 binds the heterodimer of IL4r1 receptor subunits to activate downstream signalling.30 Two published microarray data sets from microdissected chief cells indicate that IL13rα1 is expressed on both chief cells and SPEM cells in the murine stomach.8 ,31 We have further evaluated the presence of IL13rα1 protein in the gastric oxyntic mucosa. Immunostaining demonstrated strong staining for IL13rα1 in chief cells in the gastric corpus (see online supplementary figure S5). We therefore sought to ascertain if IL-13 was sufficient to restore SPEM induction after parietal cell loss in the absence of IL-33 signalling. ST2KO mice were treated with 10 µg of recombinant mouse IL-13 during L635-treatment (ST2KOL635-IL13). IL-13-treated ST2KO mice (ST2KOIL13) and non-specific IgG L635-treated ST2KO mice (ST2KOL635) were also assessed for Mist1 expression, SPEM development, proliferative SPEM and M2 macrophage polarisation (figure 7). We did not observe any phenotypic changes in ST2KOIL13 mice compared with non-specific IgG-treated ST2KO mice (figure 7A–E). In agreement with our previous findings, ST2KOL635 mice did not develop SPEM or proliferative SPEM in response to parietal cell loss (figure 7C, D). ST2KOL635-IL13 mice did not have a significant change in the number of Mist1-positive cells compared with ST2KOL635 mice (figure 7B). However, using CD44v9 co-labelling with GSII and GIF to identify SPEM cells, we found that ST2KOL635-IL13 mice had a significantly increased (p<0.0001) number of SPEM cells compared with ST2KOL635 mice (figure 7C). To determine if the metaplasia present in ST2KOL635-IL13 mice was proliferative, we immunolabelled with Ki67 and SPEM markers and found that ST2KOL635-IL13 mice did not develop proliferative SPEM (figure 7D). Macrophage infiltration and M2 macrophage polarisation were also assessed in ST2KOL635-IL13 mice. There was significant F4/80-expressing macrophage infiltration, of which approximately half were co-positive for the M2 marker, CD163 (figure 7E). These findings suggest that IL-13 is sufficient for the induction of SPEM, but progression to proliferative SPEM as well as complete M2 macrophage polarisation may accrue from another IL-33-dependent effector.

    Figure 7
    Figure 7

    Interleukin (IL)-13 is sufficient to induce metaplasia after parietal cell loss in the absence of IL-33/ST2 signalling. (A) Immunofluorescence staining for chief cell transcription factor Mist1, mucus marker Griffonia simplicifolia lectin II (GSII-lectin), chief cell granule marker gastric intrinsic factor (GIF), spasmolytic polypeptide-expressing metaplasia (SPEM) marker CD44 variant isoform 9 (CD44v9), proliferation marker Ki67, macrophage marker F4/80, M2 macrophage marker CD163 and nuclei marker diamidino-2-phenylindol (DAPI) in ST2 knock out (ST2KO) mice treated with IL-13 (ST2KOIL13), ST2KO mice treated with a non-specific IgG and L635 (ST2KOL635), or ST2KO mice treated with IL-13 and L635 (ST2KOL635-IL13). (B) Quantification of Mist1-positive nuclei per 20× field. ST2KOL635-IL13 mice had significant loss of Mist1-positive cells similar to ST2KOL635. (C) Quantification of SPEM cells determined by per cent of GIF-positive cells co-positive for CD44v9 and GSII-lectin. ST2KOL635-IL13 mice have a significantly higher per cent of SPEM cells compared with ST2KOIl13 and ST2KOL635 mice. (D) Quantification of proliferative SPEM cells determined by the per cent of SPEM cells positive for Ki67. Although there is more SPEM in ST2KOL635-IL13 mice, there is no significant difference in proliferative SPEM. (E) Quantification of F4/80-positive macrophages (red) and F4/80 and CD163 co-positive M2 macrophages (green) per 20× field. ST2KOL635-IL13 mice exhibit similar levels of F4/80-positive macrophage infiltration to ST2KOL635 mice, but there are significantly more F4/80 and C163 co-positive M2 macrophages in ST2KOL635-IL13 mice. *denotes significant difference (p<0.05) compared with untreated mice, + denotes significant difference (p<0.05) compared with ST2KO non-specific IgG-treated L635-treated mice. White dashed boxes are magnified in the insets. Scale bars=100 μm.

    Discussion

    The present study identifies that the IL-33/IL-13 cytokine-signalling network is required and sufficient for the induction of metaplasia following parietal cell loss. Using in vivo mouse models, we established that IL-33 signalling via IL-13 drives chief cell transdifferentiation into a mucus-producing metaplastic lineage. In the absence of IL-33, metaplasia does not develop, which is a necessary process for gastric epithelial repair.32 In addition to being necessary for the induction of metaplasia, IL-33 signalling drives M2 macrophage polarisation that is associated with the progression towards a more advanced metaplasia.11

    Previous studies in the stomach have attempted to ascertain the mechanism of metaplasia induction and progression. We recently identified infiltrating macrophages as necessary for the proliferation and intestinalisation of metaplasia in mice.11 Macrophage depletion in L635-treated mice resulted in a less advanced metaplasia. Our model established that macrophages promoted the progression of metaplasia, but is not required for metaplasia induction. Efforts to circumvent metaplasia development through manipulation of epidermal growth factor (EGF) ligands/receptors (Areg,33 EGFR33 ,34), endocrine pathways (gastrin,10 ,35 histamine36) and cytokines (interferon-γ ,37–39 IL-1140) have yet to identify factor(s) necessary for the induction of metaplasia in response to parietal cell loss in either an acute drug-induced system or chronic Helicobacter infection. Previous studies found that administration of IL-33 is sufficient to induce hypertrophy and mucous metaplasia in the airway, stomach, and intestine through the induction Th2 cytokines and infiltration of myeloid cells and eosinophils.20 ,21 ,41–44 Furthermore, it is well established that IL-33 signalling results in upregulation of the expression of IL-13, which is required for the development of mucous metaplasia in the airway allergic response. Our findings demonstrate an analogous role for IL-33 as an inducer of IL-13 in the promotion of metaplasia in the stomach following parietal cell loss. Recent investigations have suggested that SPEM is involved in the process of repair after acute mucosal injury.32 ,45 It therefore seems likely that an IL-33/IL-13 pathway leading to metaplasia induction is important not only for gastric pre-neoplasia, but also for gastric epithelial repair. These findings are consistent with previous studies that have shown that decreased IL-33 in TFF2-deficient mice may contribute to defective metaplasia development that delays gastric epithelial repair.21 ,32 ,45

    We found that IL-33 signalling was necessary for metaplasia induction. In IL33KO and ST2KO mice, chief cells were able to downregulate Mist1 but were unable to transdifferentiate into mucus-producing SPEM cells. Interestingly, the loss of IL-33 or ST2 receptor did not impact macrophage recruitment, although we detected decreased expression of M2a-associated cytokines (CCL17, CCL22, CCL24) and confirmed a loss of M2-polarised macrophages in both L635-treated IL-33 and ST2 knockout mice. Thus, IL-33 was not necessary for macrophage recruitment, nor was macrophage recruitment alone sufficient to induce transdifferentiation of chief cells in SPEM. IL-33 induces Th2 cytokine expression from ILC2 cells and T helper cells.20 ,46–48 A recent investigation has implicated ILC2 cells in the promotion of neoplasia in a model of diffuse gastric carcinogenesis.49 Indeed, IL33KO mice treated with L635 have reduced Th2 cytokine expression, further illustrating the relationship between IL-33 and Th2 cytokines induction.

    We investigated the role of Th2 cytokines downstream of IL-33 and found that IL-5 and IL-9-depleted mice treated with L635 developed metaplasia normally after parietal cell loss. However, IL-13 knockout mice treated with L635 did not develop metaplasia, indicating that IL-13, downstream of IL-33, induces metaplasia after acute injury and parietal cell loss. It seems likely that IL-13 can act directly on chief cells to promote the transdifferentiation of metaplasia due to the localisation of the IL-13 receptor on chief cells. Thus, we present a novel pathway in the stomach required for metaplasia induction where IL-33 stimulates IL-13 release, likely from ILC2 cells, which then acts as a key initiator of metaplasia development. ST2KO mice treated with IL-13 concurrently with L635 were able to develop metaplasia in response to parietal cell loss, further supporting this finding. Therefore, IL-13 is sufficient to induce metaplasia after parietal cell loss in the absence of IL-33 signalling. Nevertheless, since treatment of ST2KO mice with IL-13 did not induce proliferation in SPEM after L635-treatment, a factor other than IL-13 must be required for IL-33-dependent induction of proliferative SPEM.

    In summary, our investigations of the induction of metaplasia after acute parietal cell loss revealed the requirement of IL-33 in the development of SPEM. We further demonstrated that IL-33 is necessary to promote a Th2 inflammatory response and M2a polarisation of recruited macrophages, which drives advancement of SPEM.11 The identification of IL-13 downstream of IL-33 signalling as a critical regulator of metaplasia development supports a central role for cytokine and immune regulation of the epithelial response to acute injury. It is important to note that while IL33KO, ST2KO and IL13KO mice all show profound parietal cell loss after L635 treatment and reductions in Mist1 expression, the chief cells fail to complete transdifferentiation. Thus, parietal cell loss alone is not sufficient to induce chief cell transdifferentiation. Rather, cytokine responses after parietal cell loss are also required to promote transdifferentiation as part of repair and regeneration of the gastric epithelium. Therapeutic modalities currently used to modulate IL-13 could potentially promote epithelial regeneration. Similarly, inhibiting IL-33 or IL-13 cytokine pathways can modulate M2a macrophage polarisation, a critical step in the progression of metaplasia towards gastric neoplasia.

    Methods

    Mice

    C57BL/6J mice were purchased from Jackson Labs (Bar Harbor, Maine, USA). IL33KO mice were generated as previously described and maintained on the C57BL/6J background.50 ST2KO mice were developed as previously described51 and backcrossed to a Balb/C background.52 IL13KO mice used are previously described53 and backcrossed to a Balb/C background.54 Regular mouse chow and water ad libitum was provided during experiments in a temperature-controlled room with 12-hour light-dark cycles. All treatment maintenance and care of animals in these studies followed protocols approved by the Institutional Animal Care and Use Committees of Vanderbilt University and Case Western Reserve University.

    Drug treatment

    Three to six mice were used per group. L635 and DMP-777 treatment and dosage was conducted as previously described.11 Clodronate depletion of macrophages was performed as previously described.11 Anti-IL-5 (1 mg/kg, R&D Systems, clone TRFK5) and anti-IL-9 (1 mg/kg, R&D Systems, clone #222622) was administered via intraperitoneal injection to deplete IL-5 and IL-9, respectively. Depletion protocols were determined based upon previous studies.55 ,56 Anti-IL-5 was given daily for two days prior to L635-treatment, and once on day 2 of L635 treatment protocol. Anti-IL-9 antibody was given to mice on days 1 and 2 of L635 treatment. Recombinant mouse IL-13 dosage (5 ) was determined based upon prior studies and administered once daily for two days with sacrifice occurring 2 hours after final dose.56 ,57 IL-13 administration was given on days 2 and 3 of L635 treatment. Vehicle control non-specific IgG antibodies (Purified Rat IgG2a, Leaf; BioLegend, San Diego, California, USA) were administered to control mice at a similar dosage and timing as for experimental mice.

    Macrophage isolation and fluorescence-activated cell sorting

    Three mice were pooled for each macrophage preparation, and three preparations were used for RNA sequencing. L635 and DMP-777-treated mice were sacrificed on the final day of treatment. The stomach was removed and opened along the greater curvature. Stomach contents were rinsed with phosphate buffered saline (PBS). The antrum and forestomach were removed with a razor blade and discarded, and the corpus was diced into pieces roughly 1 mm3 in size, and washed with 0.07% dithiothreitol in PBS three times for 5 min. Stomach corpus pieces were then enzymatically digested using 1 U/mL of Dispase in DMEM/F-12 (Stem Cell Technologies, Vancouver, British Columbia, Canada) for 7 min at 37°C and subsequently diluted using an equal volume of macrophage growth media, Iscove's Modified Dulbecco's Medium (ThermoFisher), with 10% fetal bovine serum (FBS). Pieces were sieved through a 100 µm filter (Corning) and crushed with a rubber syringe plunger into a single-cell suspension over a 50 mL conical tube. Cells were spun down twice at 500 g for 10 min at 4°C and resuspended in PBS with 10% FBS to remove excess cell debris and media. Cells were blocked using Mouse BD Fc Block (1:100, BD Pharmingen) for 15 min on ice and incubated with rat anti-mouse F4/80 conjugated to PE (1:25, BD Bioscience) for 30 min on ice. Prior to cell sorting, cells were washed twice in PBS with 10% FBS and incubated with 4′,6-diamidino-2-phenylindol (DAPI) (1:10 000). Cells were sorted using a BD fluorescence-activated cell sorting (FACS) Aria III (BD Biosciences, San Jose, California) and initially segregated from debris using forward scatter (FSC) and side scatter (SSC) properties of the 488 nm laser. Single cells were selected using the voltage pulse geometries of the FSC diode and SSC photomultiplier tube detectors. Dead cells were excluded based on their DAPI staining. Macrophages expressing F4/80 were sorted directly into Trizol (Invitrogen) using a 100 µm nozzle. RNA was extracted from Trizol and DNAse-treated (Promega) using the manufacturer's protocol. RNA sequencing was performed as detailed in online supplementary methods and results were deposited in the GEO database (GSE77195).

    Quantitative real-time PCR analysis

    The stomach corpus was isolated using a razor blade and placed into RNAlater or frozen on dry ice. Trizol was used to extract RNA from the tissue. One microgram of DNAse-treated (Promega) RNA was transcribed into complementary DNA using Superscript III Reverse Transcriptase (Invitrogen). Quantitative real-time PCR was conducted and analysed as previously described using SYBR Green Supermix.11 Ym1 primers detecting M2 macrophage transcripts were obtained from Qiagen (QuantiTect) and followed manufacturer protocol. Results were normalised to GAPDH and displayed as expression level (2−ΔCt).11

    Immunocytochemistry

    Mouse stomachs were fixed in 4% paraformaldehyde overnight and transferred into 70% ethanol for paraffin embedding. The protocol for subsequent processing and sectioning was previously reported.11 Tissue sections were blocked in Protein Block Serum-Free (DakoCytomation) for 1 hour at room temperature. The following primary antibodies were incubated overnight at 4°C in Antibody Diluent with Background Reducing Components (DakoCytomation): goat anti-intrinsic factor (1:1000, a gift from Dr. David Alpers, Washington University, St. Louis, Missouri, USA), rat anti-Ki67 (1:50, BioLegend clone 16A8), rabbit anti-Ki67 (1:500, Cell Signaling), rat anti-F4/80 (1:500; Invitrogen, Grand Island, New York, USA), rabbit anti-Mist1 (1:1000, a gift from Jason Mills, Washington University, St. Louis, Missouri, USA), rabbit anti-IL-13rα (1:250, Abcam, 1:500), rat anti-Cd44v9 (1:25 000; Cosmo Bio, Japan), mouse anti-CD163 (1:200, NeoMarkers, Fremont, California, USA) and goat anti-IL-33 (1:200; R&D Systems). Fluorescent secondary antibodies (1:500) and Alexa 488 conjugated Griffonia simplicifolia lectin II (GSII-lectin) (1:1000; Molecular Probes, Eugene, Oregon, USA), and were incubated at room temperature for 1 hour. Zeiss Axio Imager M2 microscope with Axiovision digital imaging system (Zeiss, Jena GmBH, Germany) or an Ariol SL-200 automated slide scanner (Leica Biosystems, Buffalo Grove, Illinois, USA) in the Vanderbilt Digital Histology Shared Resource was used to analyse sections.

    Image quantitation

    Images were analysed using CellProfiler or manually counted.58 Experimental groups contained three or six mice, with three to five representative images taken from each mouse at 20×. Mist1 or GIF-positive cells denoted chief cells, while SPEM lineages were identified as GIF, GSII-lectin and CD44v9 triple-positive cells. Proliferative SPEM was identified as GIF, GSII lectin and Ki67 triple-positive cells. Macrophages were counted using the marker F4/80. M2 macrophages were identified and counted as CD163 and F4/80 dual-positive cells. A one-way analysis of variance with post hoc examination of significant means with Bonferroni's test was used to determine statistical significance among multiple groups and Mann-Whitney test was employed for comparison between two group paradigms.

    Acknowledgments

    The authors thank David Flaherty in the Flow Sorting Shared Resource Core for technical assistance with these studies.

    References

    1. ↵
      1. de Martel C,
      2. Ferlay J,
      3. Franceschi S, et al
      . Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol 2012;13:607–15. doi:10.1016/S1470-2045(12)70137-7
      OpenUrlCrossRefPubMed
    2. ↵
      1. Blaser MJ,
      2. Parsonnet J
      . Parasitism by the “slow” bacterium Helicobacter pylori leads to altered gastric homeostasis and neoplasia. J Clin Invest 1994;94:4–8. doi:10.1172/JCI117336
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Nam KT,
      2. Lee HJ,
      3. Sousa JF, et al
      . Mature chief cells are cryptic progenitors for metaplasia in the stomach. Gastroenterology 2010;139:2028–37.e9. doi:10.1053/j.gastro.2010.09.005
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Tatematsu M,
      2. Tsukamoto T,
      3. Mizoshita T
      . Role of Helicobacter pylori in gastric carcinogenesis: the origin of gastric cancers and heterotopic proliferative glands in Mongolian gerbils. Helicobacter 2005;10:97–106. doi:10.1111/j.1523-5378.2005.00305.x
      OpenUrlCrossRefPubMed
    5. ↵
      1. Yoshizawa N,
      2. Takenaka Y,
      3. Yamaguchi H, et al
      . Emergence of spasmolytic polypeptide-expressing metaplasia in Mongolian gerbils infected with Helicobacter pylori. Lab Invest 2007;87:1265–76. doi:10.1038/labinvest.3700682
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Choi E,
      2. Hendley AM,
      3. Bailey JM, et al
      . Expression of activated ras in gastric chief cells of mice leads to the full spectrum of metaplastic lineage transitions. Gastroenterology 2016;150:918–30. doi:10.1053/j.gastro.2015.11.049
      OpenUrl
    7. ↵
      1. Varon C,
      2. Dubus P,
      3. Mazurier F, et al
      . Helicobacter pylori infection recruits bone marrow-derived cells that participate in gastric preneoplasia in mice. Gastroenterology 2012;142:281–91. doi:10.1053/j.gastro.2011.10.036
      OpenUrlCrossRefPubMed
    8. ↵
      1. Weis VG,
      2. Sousa JF,
      3. LaFleur BJ, et al
      . Heterogeneity in mouse spasmolytic polypeptide-expressing metaplasia lineages identifies markers of metaplastic progression. Gut 2013;62:1270–9. doi:10.1136/gutjnl-2012-302401
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Goldenring JR,
      2. Ray GS,
      3. Coffey RJ, et al
      . Reversible drug-induced oxyntic atrophy in rats. Gastroenterology 2000;118:1080–93. doi:10.1016/S0016-5085(00)70361-1
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Nomura S,
      2. Yamaguchi H,
      3. Ogawa M, et al
      . Alterations in gastric mucosal lineages induced by acute oxyntic atrophy in wild-type and gastrin-deficient mice. Am J Physiol Gastrointest Liver Physiol 2005;288:G362–75. doi:10.1152/ajpgi.00160.2004
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Petersen CP,
      2. Weis VG,
      3. Nam KT, et al
      . Macrophages promote progression of spasmolytic polypeptide-expressing metaplasia after acute loss of parietal cells. Gastroenterology 2014;146:1727–38.e8. doi:10.1053/j.gastro.2014.02.007
      OpenUrlCrossRefPubMed
    12. ↵
      1. Mills CD,
      2. Kincaid K,
      3. Alt JM, et al
      . M-1/M-2 macrophages and the Th1/Th2 paradigm. J Immunol 2000;164:6166–73. doi:10.4049/jimmunol.164.12.6166
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Gautier EL,
      2. Shay T,
      3. Miller J, et al
      . Gene-expression profiles and transcriptional regulatory pathways that underlie the identity and diversity of mouse tissue macrophages. Nat Immunol 2012;13:1118–28. doi:10.1038/ni.2419
      OpenUrlCrossRefPubMed
    14. ↵
      1. Martinez FO,
      2. Gordon S
      . The M1 and M2 paradigm of macrophage activation: time for reassessment. F1000Prime Rep 2014;6:13. doi:10.12703/P6-13
      OpenUrlCrossRef
    15. ↵
      1. Suganami T,
      2. Nishida J,
      3. Ogawa Y
      . A paracrine loop between adipocytes and macrophages aggravates inflammatory changes: role of free fatty acids and tumor necrosis factor alpha. Arterioscler Thromb Vasc Biol 2005;25:2062–8. doi:10.1161/01.ATV.0000183883.72263.13
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Lu H,
      2. Clauser KR,
      3. Tam WL, et al
      . A breast cancer stem cell niche supported by juxtacrine signalling from monocytes and macrophages. Nat Cell Biol 2014;16:1105–17. doi:10.1038/ncb3041
      OpenUrlCrossRefPubMed
    17. ↵
      1. Nozaki K,
      2. Ogawa M,
      3. Williams JA, et al
      . A molecular signature of gastric metaplasia arising in response to acute parietal cell loss. Gastroenterology 2008;134:511–22. doi:10.1053/j.gastro.2007.11.058
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Wada T,
      2. Ishimoto T,
      3. Seishima R, et al
      . Functional role of CD44v-xCT system in the development of spasmolytic polypeptide-expressing metaplasia. Cancer Sci 2013;104:1323–9. doi:10.1111/cas.12236
      OpenUrlCrossRefPubMed
    19. ↵
      1. Mantovani A,
      2. Sica A,
      3. Sozzani S, et al
      . The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol 2004;25:677–86. doi:10.1016/j.it.2004.09.015
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Schmitz J,
      2. Owyang A,
      3. Oldham E, et al
      . IL-33, an interleukin-1-like cytokine that signals via the IL-1 receptor-related protein ST2 and induces T helper type 2-associated cytokines. Immunity 2005;23:479–90. doi:10.1016/j.immuni.2005.09.015
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Buzzelli JN,
      2. Chalinor HV,
      3. Pavlic DI, et al
      . IL33 is a Stomach Alarmin that initiates a Skewed Th2 response to Injury and Infection. CMGH 2015;1:203–221.e3. doi:10.1016/j.jcmgh.2014.12.003
      OpenUrl
    22. ↵
      1. Carriere V,
      2. Roussel L,
      3. Ortega N, et al
      . IL-33, the IL-1-like cytokine ligand for ST2 receptor, is a chromatin-associated nuclear factor in vivo. Proc Natl Acad Sci USA 2007;104:282–7. doi:10.1073/pnas.0606854104
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Kurowska-Stolarska M,
      2. Stolarski B,
      3. Kewin P, et al
      . IL-33 amplifies the polarization of alternatively activated macrophages that contribute to airway inflammation. J Immunol 2009;183:6469–77. doi:10.4049/jimmunol.0901575
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Lee A,
      2. Fox JG,
      3. Otto G, et al
      . A small animal model of human Helicobacter pylori active chronic gastritis. Gastroenterology 1990;99:1315–23.
      OpenUrlPubMedWeb of Science
    25. ↵
      1. Zuo L,
      2. Rothenberg ME
      . Gastrointestinal eosinophilia. Immunol Allergy Clin North Am 2007;27:443–55. doi:10.1016/j.iac.2007.06.002
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. McGarry MP,
      2. Stewart CC
      . Murine eosinophil granulocytes bind the murine macrophage-monocyte specific monoclonal antibody F4/80. J Leukoc Biol 1991;50:471–8.
      OpenUrlAbstract
    27. ↵
      1. Frigas E,
      2. Loegering DA,
      3. Solley GO, et al
      . Elevated levels of the eosinophil granule major basic protein in the sputum of patients with bronchial asthma. Mayo Clin Proc 1981;56:345–53.
      OpenUrlPubMedWeb of Science
    28. ↵
      1. Ponath PD,
      2. Qin S,
      3. Ringler DJ, et al
      . Cloning of the human eosinophil chemoattractant, eotaxin. Expression, receptor binding, and functional properties suggest a mechanism for the selective recruitment of eosinophils. J Clin Invest 1996;97:604–12. doi:10.1172/JCI118456
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Lawrence T,
      2. Natoli G
      . Transcriptional regulation of macrophage polarization: enabling diversity with identity. Nat Rev Immunol 2011;11:750–61. doi:10.1038/nri3088
      OpenUrlCrossRefPubMed
    30. ↵
      1. Zurawski SM,
      2. Vega F Jr.,
      3. Huyghe B, et al
      . Receptors for interleukin-13 and interleukin-4 are complex and share a novel component that functions in signal transduction. EMBO J 1993;12:2663–70.
      OpenUrlPubMedWeb of Science
    31. ↵
      1. Weis VG,
      2. Petersen CP,
      3. Mills JC, et al
      . Establishment of novel in vitro mouse chief cell and SPEM cultures identifies MAL2 as a marker of metaplasia in the stomach. Am J Physiol Gastrointest Liver Physiol 2014;307:G777–92. doi:10.1152/ajpgi.00169.2014
      OpenUrlCrossRefPubMed
    32. ↵
      1. Engevik AC,
      2. Feng R,
      3. Choi E, et al
      . The Development of Spasmolytic Polypeptide/TFF2-Expressing Metaplasia (SPEM) During Gastric Repair is Absent in the Aged Stomach. Cell Mol Gastroenterol Hepatol 2016;2:605–24.
      OpenUrl
    33. ↵
      1. Nam KT,
      2. Varro A,
      3. Coffey RJ, et al
      . Potentiation of oxyntic atrophy-induced gastric metaplasia in amphiregulin-deficient mice. Gastroenterology 2007;132:1804–19. doi:10.1053/j.gastro.2007.03.040
      OpenUrlCrossRefPubMed
    34. ↵
      1. Ogawa M,
      2. Nomura S,
      3. Varro A, et al
      . Altered metaplastic response of waved-2 EGF receptor mutant mice to acute oxyntic atrophy. Am J Physiol Gastrointest Liver Physiol 2006;290:G793–804. doi:10.1152/ajpgi.00309.2005
      OpenUrlCrossRefPubMedWeb of Science
    35. ↵
      1. Aikou S,
      2. Fukushima Y,
      3. Ogawa M, et al
      . Alterations in gastric mucosal lineages before or after acute oxyntic atrophy in gastrin receptor and H2 histamine receptor-deficient mice. Dig Dis Sci 2009;54:1625–35. doi:10.1007/s10620-009-0832-2
      OpenUrlCrossRefPubMed
    36. ↵
      1. Nozaki K,
      2. Weis V,
      3. Wang TC, et al
      . Altered gastric chief cell lineage differentiation in histamine-deficient mice. Am J Physiol Gastrointest Liver Physiol 2009;296:G1211–20. doi:10.1152/ajpgi.90643.2008
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Liu Z,
      2. Demitrack ES,
      3. Keeley TM, et al
      . IFNgamma contributes to the development of gastric epithelial cell metaplasia in Huntingtin interacting protein 1 related (Hip1r)-deficient mice. Lab Invest 2012;92:1045–57. doi:10.1038/labinvest.2012.73
      OpenUrlCrossRefPubMed
    38. ↵
      1. Syu LJ,
      2. El-Zaatari M,
      3. Eaton KA, et al
      . Transgenic expression of interferon-γ in mouse stomach leads to inflammation, metaplasia, and dysplasia. Am J Pathol 2012;181:2114–25. doi:10.1016/j.ajpath.2012.08.017
      OpenUrlCrossRefPubMedWeb of Science
    39. ↵
      1. Tu SP,
      2. Quante M,
      3. Bhagat G, et al
      . IFN-γ inhibits gastric carcinogenesis by inducing epithelial cell autophagy and T-cell apoptosis. Cancer Res 2011;71:4247–59. doi:10.1158/0008-5472.CAN-10-4009
      OpenUrlAbstract/FREE Full Text
    40. ↵
      1. Howlett M,
      2. Chalinor HV,
      3. Buzzelli JN, et al
      . IL-11 is a parietal cell cytokine that induces atrophic gastritis. Gut 2012;61:1398–409. doi:10.1136/gutjnl-2011-300539
      OpenUrlAbstract/FREE Full Text
    41. ↵
      1. Kurowska-Stolarska M,
      2. Kewin P,
      3. Murphy G, et al
      . IL-33 induces antigen-specific IL-5+ T cells and promotes allergic-induced airway inflammation independent of IL-4. J Immunol 2008;181:4780–90. doi:10.4049/jimmunol.181.7.4780
      OpenUrlAbstract/FREE Full Text
    42. ↵
      1. Stolarski B,
      2. Kurowska-Stolarska M,
      3. Kewin P, et al
      . IL-33 exacerbates eosinophil-mediated airway inflammation. J Immunol 2010;185:3472–80. doi:10.4049/jimmunol.1000730
      OpenUrlAbstract/FREE Full Text
    43. ↵
      1. Préfontaine D,
      2. Nadigel J,
      3. Chouiali F, et al
      . Increased IL-33 expression by epithelial cells in bronchial asthma. J Allergy Clin Immunol 2010;125:752–4. doi:10.1016/j.jaci.2009.12.935
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. De Salvo C,
      2. Wang XM,
      3. Pastorelli L, et al
      . IL-33 drives eosinophil infiltration and pathogenic Type 2 Helper T-cell immune responses leading to chronic experimental ileitis. Am J Pathol 2016;186:885–98. doi:10.1016/j.ajpath.2015.11.028
      OpenUrlCrossRefPubMed
    45. ↵
      1. Farrell JJ,
      2. Taupin D,
      3. Koh TJ, et al
      . TFF2/SP-deficient mice show decreased gastric proliferation, increased acid secretion, and increased susceptibility to NSAID injury. J Clin Invest 2002;109:193–204. doi:10.1172/JCI12529
      OpenUrlCrossRefPubMedWeb of Science
    46. ↵
      1. Wang YH,
      2. Angkasekwinai P,
      3. Lu N, et al
      . IL-25 augments type 2 immune responses by enhancing the expansion and functions of TSLP-DC-activated Th2 memory cells. J Exp Med 2007;204:1837–47. doi:10.1084/jem.20070406
      OpenUrlAbstract/FREE Full Text
    47. ↵
      1. Fort MM,
      2. Cheung J,
      3. Yen D, et al
      . IL-25 induces IL-4, IL-5, and IL-13 and Th2-associated pathologies in vivo. Immunity 2001;15:985–95. doi:10.1016/S1074-7613(01)00243-6
      OpenUrlCrossRefPubMedWeb of Science
    48. ↵
      1. von Moltke J,
      2. Ji M,
      3. Liang HE, et al
      . Tuft-cell-derived IL-25 regulates an intestinal ILC2-epithelial response circuit. Nature 2015;529:221–5. doi:10.1038/nature16161
      OpenUrl
    49. ↵
      1. Hayakawa Y,
      2. Ariyama H,
      3. Stancikova J, et al
      . Mist1 expressing gastric stem cells maintain the normal and neoplastic gastric epithelium and are supported by a perivascular stem cell niche. Cancer Cell 2015;28:800–14. doi:10.1016/j.ccell.2015.10.003
      OpenUrlCrossRefPubMed
    50. ↵
      1. Maywald RL,
      2. Doerner SK,
      3. Pastorelli L, et al
      . IL-33 activates tumor stroma to promote intestinal polyposis. Proc Natl Acad Sci USA 2015;112:E2487–96. doi:10.1073/pnas.1422445112
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Townsend MJ,
      2. Fallon PG,
      3. Matthews DJ, et al
      . T1/ST2-deficient mice demonstrate the importance of T1/ST2 in developing primary T helper cell type 2 responses. J Exp Med 2000;191:1069–76. doi:10.1084/jem.191.6.1069
      OpenUrlAbstract/FREE Full Text
    52. ↵
      1. Mangan NE,
      2. Dasvarma A,
      3. McKenzie AN, et al
      . T1/ST2 expression on Th2 cells negatively regulates allergic pulmonary inflammation. Eur J Immunol 2007;37:1302–12. doi:10.1002/eji.200636520
      OpenUrlCrossRefPubMedWeb of Science
    53. ↵
      1. McKenzie GJ,
      2. Bancroft A,
      3. Grencis RK, et al
      . A distinct role for interleukin-13 in Th2-cell-mediated immune responses. Curr Biol 1998;8:339–42. doi:10.1016/S0960-9822(98)70134-4
      OpenUrlCrossRefPubMedWeb of Science
    54. ↵
      1. Müller U,
      2. Stenzel W,
      3. Köhler G, et al
      . IL-13 induces disease-promoting type 2 cytokines, alternatively activated macrophages and allergic inflammation during pulmonary infection of mice with Cryptococcus neoformans. J Immunol 2007;179:5367–77. doi:10.4049/jimmunol.179.8.5367
      OpenUrlAbstract/FREE Full Text
    55. ↵
      1. Herndon FJ,
      2. Kayes SG
      . Depletion of eosinophils by anti-IL-5 monoclonal antibody treatment of mice infected with Trichinella spiralis does not alter parasite burden or immunologic resistance to reinfection. J Immunol 1992;149:3642–7.
      OpenUrlAbstract
    56. ↵
      1. Steenwinckel V,
      2. Louahed J,
      3. Lemaire MM, et al
      . IL-9 promotes IL-13-dependent paneth cell hyperplasia and up-regulation of innate immunity mediators in intestinal mucosa. J Immunol 2009;182:4737–43. doi:10.4049/jimmunol.0801941
      OpenUrlAbstract/FREE Full Text
    57. ↵
      1. Manocha M,
      2. Shajib MS,
      3. Rahman MM, et al
      . IL-13-mediated immunological control of enterochromaffin cell hyperplasia and serotonin production in the gut. Mucosal Immunol 2013;6:146–55. doi:10.1038/mi.2012.58
      OpenUrlCrossRefPubMed
    58. ↵
      1. Jones TR,
      2. Kang IH,
      3. Wheeler DB, et al
      . CellProfiler Analyst: data exploration and analysis software for complex image-based screens. BMC Bioinformatics 2008;9:482. doi:10.1186/1471-2105-9-482
      OpenUrlCrossRefPubMed

    Footnotes

    • Contributors CPP: designed and performed experiments, analysed data and drafted manuscript. ARM, CS, AP, AE and NP: performed experiments, analysed data and revised manuscript. CDS and EC: designed and performed experiments, analysed data and revised manuscript. SEL and RSP analysed data and revised manuscript. TTP and JRG: designed experiments, analysed data and revised manuscript. All authors had access to the study data and reviewed and approved the final manuscript.

    • Funding These studies were supported by grants from a Department of Veterans Affairs Merit Review Award (I01BX000930) and NIH RO1 DK071590, as well as a grant from the Martell Foundation (to JRG), and funding from the DeGregorio Family Foundation and a Pilot & Feasibility Award from P50 CA150964 (to TTP). CPP was supported by an NIH NRSA Predoctoral Fellowship (F31 DK104600). ARM was supported by NIH T32 GM008554. This work was supported by core resources of the Vanderbilt Digestive Disease Center, (P30 DK058404) the Vanderbilt-Ingram Cancer Center (P30 CA68485, Chemical Synthesis Core), and imaging supported by both the Vanderbilt Combined Imaging Shared Resource and the Vanderbilt Digital Histology Shared Resource.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement The RNA sequencing results are included in online supplementary table S1.