Spontaneous autoimmune gastritis and hypochlorhydria are manifest in the ileitis-prone SAMP1/YitFcs mice

Abstract

SAMP1/YitFcs mice serve as a model of Crohn's disease, and we have used them to assess gastritis. Gastritis was compared in SAMP1/YitFcs, AKR, and C57BL/6 mice by histology, immunohistochemistry, and flow cytometry. Gastric acid secretion was measured in ligated stomachs, while anti-parietal cell antibodies were assayed by immunofluorescence and enzyme-linked immunosorbent spot assay. SAMP1/YitFcs mice display a corpus-dominant, chronic gastritis with multifocal aggregates of mononuclear cells consisting of T and B lymphocytes. Relatively few aggregates were observed elsewhere in the stomach. The infiltrates in the oxyntic mucosa were associated with the loss of parietal cell mass. AKR mice, the founder strain of the SAMP1/YitFcs, also have gastritis, although they do not develop ileitis. Genetic studies using SAMP1/YitFcs-C57BL/6 congenic mice showed that the genetic regions regulating ileitis had comparable effects on gastritis. The majority of the cells in the aggregates expressed the T cell marker CD3 or the B cell marker B220. Adoptive transfer of SAMP1/YitFcs CD4⁺ T helper cells, with or without B cells, into immunodeficient recipients induced a pangastritis and duodenitis. SAMP1/YitFcs and AKR mice manifest hypochlorhydria and anti-parietal cell antibodies. These data suggest that common genetic factors controlling gastroenteric disease in SAMP1/YitFcs mice regulate distinct pathogenic mechanisms causing inflammation in separate sites within the digestive tract.

the gastrointestinal tract faces the unique immunological challenge of coping with a dynamic and vast array of dietary and microbial antigens. Chronic inflammatory disease of the intestine, as well as the stomach, of adults and children (9, 10, 38) arises in genetically susceptible individuals that fail to properly regulate host responses to these luminal antigens. The interrelationship of inflammation in the stomach and intestine can be appreciated in studies that examine these tissues from subjects with inflammatory bowel disease (IBD). For example, sterile gastritis has been observed in subjects with Crohn's disease or ulcerative colitis (7, 31).

Several animal models have been used to study IBD; however, models of Crohn's disease are limited. SAMP1/Yit mice develop a spontaneous terminal ileitis with many features of Crohn's disease (17, 23). Ileitis in SAMP1/Yit mice is associated with proinflammatory cytokine production: T helper (Th) type 1 (Th1) cell responses in younger mice (17), but also Th2 cytokines (42) as they mature (4). It has been suggested that the lesions in SAMP1/YitFcs mice were limited to the terminal ileum (17). These mice have also been shown to have skin lesions as well as hepatitis (23). The SAMP1/Yit mice were originally derived from AKR mice with an unintentional backcross to an unknown strain (18). AKR mice have not been reported to develop spontaneous inflammation in their digestive tract.

Although many immunological responses are perturbed in SAMP1/Yit mice, some evidence suggests that the primary defect lies in the epithelial cell barrier. In the process of studying intestinal epithelial permeability (28), we showed a disrupted epithelial cell barrier function in the SAMP1/Yit stomach (34). While SAMP1/Yit mice develop a spontaneous terminal ileitis and have been used as a model of Crohn's disease, the health of their gastric mucosa has not been carefully examined. The purpose of this study was to more thoroughly investigate the gastric manifestations in SAMP1/Yit mice.

MATERIALS AND METHODS

Mice.

SAMP1/YitFcs mice [the registered nomenclature for a strain that has been referred to previously as SAMP1/YitFc (35)] were derived from the SAMP1/Yit mice originally developed at the Yakult Central Institute for Microbiological Research (Tokyo, Japan). Mice were provided initially by Drs. Pizarro and Cominelli and subsequently by the University of Virginia. Congenic mouse strains generated at the University of Virginia (M.M.) included SAMP.B6-D9Mit297-D9Mit212 (SAMP.B6C9A), SAMP.B6-D9Mit297-D9Mit123 69BL (SAMP.B6C9BL), and SAMP.B6-DXMit166-DXMit156 (SAMP.B6CX). Congenic strains and (AKR × SAMP1/YitFcs) F1 mice were generated on site as described previously (18, 40). C57BL/6J, AKR/J, and C3Smn.CB17-Prkdc^scid/J (C3H.scid) mice were obtained from Jackson Laboratories (Bar Harbor, ME) or bred and housed in specific pathogen-free conditions at the University of Virginia. Male and female 8- to 58-wk-old mice were used in the studies. Mice were not infected with any known pathogenic Helicobacter spp. All mice were maintained specific pathogen-free with quarterly health surveillance monitoring, and protocols were performed with the approval of the University of Virginia Institutional Animal Care and Use Committee.

Characterization of gastritis.

Stomachs were collected from all mice, flushed with PBS, fixed in Bouin's fixative, oriented symmetrically (to allow uniform cross-sectional measurement of mucosa thickness), and prepared for staining with hematoxylin and eosin (1) or processed for immunohistochemistry to detect CD45R/B220, F4/80, and CD3. For B220 (catalog no. 553086, BD Bioscience, San Diego, CA) and F4/80 (catalog no. MCA497BB, Serotec, Raleigh, NC), tissue sections were pretreated by steaming with Diva solution (Biocare, Concord, CA) for 20 min and then incubated with the primary antibody at a concentration of 1:1,000 and 1:200, respectively. Detection was achieved using a polymer detection system (Rat-on-Mouse AP kit, Biocare) and Vulcan fast red (Biocare). CD3⁺ cells were detected using a monoclonal antibody (Ab 16669, Abcam, Cambridge, MA) at 1:1,000 dilution, with sequential application of a goat anti-rabbit antibody (BA-1000, Vector Laboratories, Burlingame, CA), Vector Elite ABC (Vector Laboratories), and NovaRed (Vector Laboratories).

All hematoxylin-eosin-stained slides were scanned and digitally stored using ScanScope (Aperio, Vista, CA) and read using ImageScope (Aperio) software to allow the use of a web-based file-sharing platform. Scoring was based on a modification of a comprehensive approach described previously (1, 2, 14) developed in consultation with a comparative pathologist (C.C.B.) and an investigator with extensive experience scoring autoimmune gastritis in mice (K.S.T.). Thickness of the corpus tissue was measured using the “ruler function” of ImageScope, which allows the start and end of the measurement to be precisely placed from the edge of the lumen to the outside of the serosa ∼5 mm from the junction of the forestomach and corpus. Multiple criteria for the three major regions of the murine stomach were used to grade the histopathology (Table 1). All photomicrographs were captured using ImageScope.

Table 1.

Summary of scoring criteria

	Score
Parameter	0	1	2	3
Thickness	<750 μm	>750 μm
Forestomach
Granulocytes	Absent	Present
Mononuclear cells	Absent	Focal, 1–5 cells in a ×40 field	Aggregates
Hyperkeratosis	<10 layers	>10 layers
Corpus
Granulocytes	Absent or rare	Focal, 2–5 cells in a ×40 field	>1 field with 5 cells	Luminal
Mononuclear cells	Rare	Focal	Diffuse, >30 cells in a ×40 field	Aggregates
Parietal cell loss	None	Irregular staining and morphology	Moderate loss of parietal cell mass	Marked loss of parietal cell mass
Antrum
Granulocytes	Absent or rare	Focal, 2–5 cells in a ×40 field	>1 field with 5 cells	Luminal
Mononuclear cells	Rare	Focal	Diffuse	Aggregates

Gastric mononuclear cells (MNC) from the corpus dissected free of the forestomach and antrum, gastric lymph nodes, or mesenenteric lymph nodes were prepared as reported previously (1, 2, 37), labeled for CD3, CD45, and B220 (reagents described in Table 2) immediately after isolation, and assayed by flow cytometry, as previously described (3, 25, 45). After gating on CD45⁺ cells excluding aggregates, samples were assayed using a flow cytometer (Fortessa, Becton Dickinson, San Jose, CA). Data were analyzed using Flowjo software (Treestar, Ashland, OR). Controls included beads coated with each antibody for compensation, and gates were set using gastric MNC labeled with combinations of all antibodies minus one (36), such that ≥97.5% of all nonspecific fluorescence was excluded.

Table 2.

Reagents used for flow cytometry

Description	Fluorochrome	Channel	Source	Stock No.	Isotype
Anti-CD3	FITC	FL1	eBioscience	11003182	Hamster IgG
Anti-B220	PE-TX red	D yellow/green	BD	551489	Rat IgG2a, IgG2k
Anti-CD45	PerCP	B blue	BD	557235	Rat IgG2b, IgG2k

PE, phycoerythrin; TX red, Texas red; PerCP, peridinin-chlorophyll-protein complex.

Adoptive transfer studies.

To investigate the pathogenicity of lymphocytes in the gastritis, adoptive transfer studies were performed as previously described for studying the role of these cells in ileitis (27). Briefly, mesenteric lymph node cells from SAMP1/YitFcs or C3H mice were prepared using standard techniques (15) and used as a source of lymphocytes that were enriched into CD4⁺ or B220⁺ subsets by magnetic beads; then 5 × 10⁵ T cells, B cells, or both were adoptively transferred into the major histocompatibility complex-compatible C3H.scid mice by intraperitoneal injection. After 6 wk, tissues were collected and scored for histology.

Evaluation of gastric acid secretion.

Mice were fasted in metabolic cages overnight, weighed, and anesthetized. The pylorus was isolated, and a snug ligature was applied at the junction of the antrum and pylorus. After 3 h of recovery, the mice were anesthetized, the proximal end of the stomach was ligated at the junction of the esophagus, and the entire stomach was removed. The gastric contents were collected, weighed to estimate the volume, and assayed for pH. In some cases, pentagastrin (50 μg/kg ip) was administered in a bolus, but no differences were observed between this treatment and the 3-h unstimulated gastric acid secretion (data not shown).

Autoantibody detection.

Sera were screened for autoantibodies by immunofluorescence. Briefly, frozen sections of normal mouse stomach were incubated with a 1:50 dilution of the serum to be screened from SAMP1/YitFcs, AKR, or C57BL/6 mice and then with a FITC-conjugated anti-murine antibody (IgM, IgG1, IgG2a, IgG2b, or IgA) and detected by fluorescent microscopy (43). The intensity of fluorescence was graded (assigned a score of 0–3) against negative and positive control sections. In addition, gastric MNC isolated as described above were compared with MNC isolated from bone marrow and spleen (25, 45) by enzyme-linked immunosorbent spot (ELISpot) assay for total IgM-, IgG-, and IgA-producing cells (16), as well as antigen-specific spots, using plates coated with 50 μg/ml of a lysate of enriched parietal cells isolated from C57BL/6 mouse stomachs prepared as described previously (29) or porcine H⁺-K⁺-ATPase (Arotec Diagnostics). Plates coated with AGS cell lysates served as a negative control.

Statistical analysis.

The inflammation scores were compared by Student's t-test with Bonferroni's correction for multiple comparisons when indicated. Multivariate analysis was used to examine the effects of age and sex on inflammation. The correlation of corpus inflammation and anti-parietal cell antibody titers was also assessed. The pH of gastric secretions was compared using Wilcoxon's rank sum test. All analyses were done using the statistical package in Sigmaplot (GraphPad Software, San Diego, CA).

RESULTS

SAMP1/YitFcs mice develop a corpus-dominant gastritis.

Pathological changes in gastric tissue were assessed in the forestomach, corpus, and antrum. Figure 1 shows no significant evidence of gastritis in C57BL/6 mice. In contrast, tissue from the SAMP1/YitFcs mice was thickened (mean of 10 animals: 905.3 and 479 μm as measured at the junction of the forestomach and corpus for SAMP1/YitFcs and C57BL/6 respectively, P < 0.05) and had a marked chronic, mononuclear cell infiltrate, including defined multifocal aggregates (Fig. 1A vs. Fig. 1, B and C) and occasional accumulations in the lumen of the gastric glands that appeared to include dead mononuclear cells but rarely identifiable polymorphonuclear cells, in the corpus. Infiltration with polymorphonuclear cells was remarkably modest. Epithelial glands in the corpus were often distended or tortuous, with evidence of epithelial cell proliferation. In areas of inflammation, there was loss of parietal cell mass. Inflammation was more modest in the antrum, and when inflammatory cells were increased, they were focal and located at the base of the glands close to or on either side of the muscularis mucosa. Histological scoring is summarized in Fig. 1D.

Fig. 1.

SAMP1/YitFcs mice develop a corpus-dominant gastritis. Gastric tissue was collected from control C57BL/6 mice (A) and SAMP1/YitFcs mice (B), and inflammation was assessed by histology. A: normal, uninflamed gastric mucosa. Higher power (A2) shows a typical uniform population of parietal cells in the corpus; representative parietal cells are indicated by yellow arrows in A3. Gastric mucosa of SAMP1/YitFcs mice is typically thickened (B; scale bars, ∼500 μm in A and B). At higher magnification, mononuclear cell aggregates are apparent in the corpus (green arrows, B2–B4); dilated glands containing accumulations of dead cells are indicated by white arrows (B2–B4). Images captured at a higher power show intact parietal cells in control mouse (A4) and inflammation with associated loss of parietal cell mass in SAMP1/YitFcs mouse (B4). Aggregates (encircled) were observed in the forestomach of some mice (C; scale bar, ∼100 μm). D: scoring for forestomach (FS), corpus, and antrum from 5 SAMP1/YitFcs (SAMP1) mice and 6 C57BL/6 (BL/6) age-matched mice, representing a sample of >60 mice. Values are means ± SE. *P < 0.05. As several genetic elements regulating the onset of ileitis in SAMP1/YitFcs mice have been identified, gastritis was assessed in these crosses. E: scoring for the forestomach, corpus, and antrum from 10 SAMP1/YitFcs, 9 AKR, and 9 (AKR × SAMP1/YitFcs) F1 mice. F: gastritis scores for 10 C57BL/6 mice and 13 SAMP1/YitFcs mice compared with scores for SAMP1/Yit/Fcs-C57BL/6 congenic mice, including 8 SAMP.B6C9A (B6C9A), 3 SAMP.B6C9BL (B6C9BL), and 3 SAMP.B6CX (B6CX) mice. Values are means ± SE. *P < 0.05 vs. SAMP1.

Ileitis and gastritis in SAMP1/YitFcs mice share some of the same genetic control.

The SAMP1/YitFcs mice were originally derived from AKR mice (40), so gastritis was assessed in this strain as well. AKR mice displayed a corpus-dominant, spontaneous, chronic gastritis. As neither strain had been reported previously to have gastritis, these observations represent two new murine models of gastritis.

When SAMP1/YitFcs mice are crossed to AKR mice, the resultant F1 mice lack ileitis (40). When F1 mice from this cross were examined, the gastritis was markedly attenuated (Fig. 1E), approaching the degree of inflammation in C57BL/6 mice. In other reports (18), congenic mice generated by backcrosses onto C57BL/6 mice yielded offspring with decreased inflammation in the terminal ileum. The congenic strains with attenuated ileitis (SAMP.B6C9A and SAMP.B6CX) also had decreased gastritis, while the SAMP.B6C9/BL congenic strain had no effect on gastritis (Fig. 1F), which, again, mimicked the effect of this cross on ileitis (18).

A mixed T and B cell infiltrate constitutes the majority of the chronic gastritis.

The mononuclear cells within the aggregates expressed B220 (Fig. 2, A and C), as well as CD3 (Fig. 2, B and D). Both of these markers identified the majority of the mononuclear cells in the aggregates in the corpus, as well as those that accumulated diffusely throughout the tissue. Cells expressing F4/80 were not detected in the aggregates and were sparse throughout the mucosa (data not shown).

Fig. 2.

Characterization of mononuclear cells within the aggregates. Photomicrographs represent 2 magnifications of sections stained for B220 (A and C) and CD3 (B and D). Number of B220⁺ cells within these aggregates exceeded the number of CD3⁺ cells. Images are from 1 mouse and are representative of 5 separate SAMP1/YitFcs mice. E: flow cytometric analysis of mononuclear cells isolated from the gastric corpus lamina propria and stained for B220 and CD3. Results from representative experiment show that almost all the freshly isolated gastric mononuclear cells expressed CD3 or B220, while a significant subset expressed both markers. PE, phycoerythrin.

Gastric mononuclear cells were isolated from the corpus or from the draining gastric lymph nodes. These cells expressed CD3 (60%) and B220 (30%) in preparations from the mucosa (Fig. 2E), as well as the draining gastric lymph nodes (data not shown), approximating the results observed by immunohistochemistry. Interestingly, a small percentage (4–7%) of CD3⁺ cells expressed B220, which would be consistent with T cells being highly activated (44), as recognized previously in the liver (26) and gut (6, 12).

Adoptive transfer of CD4⁺ Th cells induces a pangastritis and duodenitis distinct from the spontaneous disease in SAMP1/YitFcs mice.

As the ileitis in SAMP1/YitFcs mice can be adoptively transferred into immunodeficient mice (27), we determined if the same was true for the gastritis. CD4⁺ Th cells purified from the mesenteric lymph nodes of SAMP1/YitFcs mice were transferred into severe combined immunodeficiency (SCID) recipient mice (Fig. 3). While gastritis emerged, it shifted from a corpus-predominant gastritis found in the SAMP1/YitFcs donors to a pangastritis and lacked the characteristic mononuclear cell aggregates observed in the donor mice. Cotransfer of B cells from SAMP1/YitFcs donors did not significantly alter the lesions that developed in the recipients after 6 wk. Moreover, a severe duodenitis with total ablation of the villi was detected in virtually all recipients given T cells alone or in combination with B cells (Fig. 3B). The latter was surprising, as duodenitis was a relatively unusual finding in SAMP1/YitFcs mice and has not been investigated previously in the transfer model. These findings suggest that the adoptive transfer of Th cells was proinflammatory but failed to directly mimic the characteristic gastritis that develops spontaneously in these mice.

Fig. 3.

Adoptive transfer of CD4⁺ T helper (Th) cells induces a pangastritis and duodenitis distinct from the spontaneous disease in SAMP1/YitFcs mice. To investigate the role of Th cells in the pathogenesis of gastritis, severe combined immunodeficiency (SCID) recipient mice were given 5 × 10⁵ CD4⁺ Th cells with or without an equal number of B cells, both prepared from the mesenteric lymph node of SAMP1/YitFcs mice. After 6 wk, tissues were collected. A: control SCID tissue had little inflammation and a completely healthy duodenum (black arrow). SCID recipients given Th cells and B cells from SAMP1/YitFcs mice (same magnification as SCID control) had a slightly thickened corpus and evidence of a generalized pangastritis, with more obvious lesions in the antrum in the recipients than the donor SAMP1/YitFcs mice. In addition, their duodenal architecture was virtually obliterated (red arrow and higher power in B; scale bar, ∼100 μm). Parietal cells are obvious and intact in these recipients. C: histological scoring. Data are from 2 SCID and 4 reconstituted mice per group.

Gastritis in SAMP1/YitFcs mice is associated with an autoimmune gastritis and hypochlorhydria.

Since the adaptive transfer of Th cells failed to recreate the corpus-dominant gastritis in SAMP1/YitFcs mice, the role of other cells was explored. Approximately 50% of the SAMP1/YitFcs mice displayed hyperkeratosis at the junction of the forestomach and corpus (Fig. 4, A–C), often associated with an overgrowth of rod-shaped bacteria (Fig. 4D). As the number of bacteria on the surface of the forestomach epithelium appeared increased, gastric acid secretion was assessed. Measurement of the 3-h unstimulated gastric acid secretion yielded an average of 0.765 ml of gastric juice from all mice sampled and showed that approximately half of the SAMP1/YitFcs and AKR mice displayed hypo- or achlorhydria (Fig. 4E). As parietal cell damage and loss of acid production are usually associated with autoimmune gastritis, the serum was assayed for anti-parietal cell antibodies. As shown in Fig. 5, serum from SAMP1/YitFcs mice had significant titers of autoantibodies that recognized parietal cells specifically, but not adjacent cells. These antibodies were not detected in the C57BL/6 (Fig. 5B) and (AKR × SAMP1/YitFcs) F1 mice (data not shown) that lacked gastritis. Moreover, the titer of the anti-parietal cell antibody correlated to the intensity of the corpus gastritis (Fig. 5C). Examination of the isotype of these anti-parietal cell antibodies revealed multiple isotypes, including IgG1, IgG2a, IgG2b, and IgA (data not shown). ELISpot assays revealed that antibodies recognizing parietal cell lysates and H⁺-K⁺-ATPase were produced predominantly by IgA-secreting B cells isolated from the gastric corpus (Fig. 5D) compared with plates coated with irrelevant proteins, including IgG or IgM or cell lysates from AGS cells.

Fig. 4.

SAMP1/YitFcs mice have evidence of bacterial overgrowth and achlorhydria. Hyperkeratosis was observed in the forestomach in ∼50% of the SAMP1/YitFcs mice (B and C) but was not evident in the C57BL/6 mice (A). In some cases, a marked accumulation of rod-shaped bacteria was evident (D). These findings suggest that there may be bacterial overgrowth due to altered gastric acid production. Examination of the gastric pH over 3 h in unstimulated, ligated stomachs revealed abnormal gastric acid production that reached pH 7 in some cases in SAMP1/YitFcs and AKR mice, suggestive of achlorhydria (E). (AKR × SAMP1/YitFcs) F1 mice, which lacked gastritis, had essentially normal gastric acid production. Data are from 6–9 mice per group. Values are median scores. *P < 0.05 vs. C57BL/6 mice (Wilcoxon's rank sum test).

Fig. 5.

SAMP1/YitFcs and AKR mice have an autoimmune gastritis and produce anti-parietal cell antibodies. Given the effect on gastric pH, SAMP1/YitFcs (n = 6), AKR (n = 3), and (AKR × SAMP1/YitFcs) F1 mice (n = 5) were assayed for anti-parietal cell antibodies and compared with C57BL/6 mice (n = 6). A: representative immunofluorescence image in which serum from SAMP1/YitFcs mice (left) binds selectively to the parietal cells, in contrast to the sample from C57BL/6 mice (right). B: percentage of each strain with titer scores of 0 (negative), 1 (+), 2 (++), or 3 (+++). C: analysis of the inflammation score in the corpus of all 20 samples and the titers of anti-parietal cell antibodies showed a significant correlation between the 2 variables (r = 0.9274, P < 0.05). D: number of nonspecific (IgM, IgG, or IgA) and antigen-specific antibody-producing cells in gastric mononuclear cells (gMNC), spleen (SPL), and bone marrow (BM) isolated from SAMP1/YitFcs mice detected by enzyme-linked immunosorbent spot assay. Values are means ± SE for 3 observations.

DISCUSSION

This report describes two new models of autoimmune gastritis in the SAMP1/YitFcs and AKR mice. As SAMP1/YitFcs mice are derived from AKR mice (40), the presence of gastritis in both strains may not be unexpected, although only the former manifest terminal ileitis (17). F1 mice derived from a cross of SAMP1/YitFcs and AKR lose their ileitis (40). Surprisingly, they also had markedly attenuated gastritis, which may reflect an interaction between distinct alleles of a given gene or chromosomal segment that produces a phenotype distinct from either parental phenotype. Examination of congenic strains derived from crosses of C57BL/6 with SAMP1/YitFcs mice (18) showed that ileitis and gastritis fell under the same genetic control. Although the pathogenesis of the two lesions may differ, these studies suggest that control of disease in the ileum and stomach of the SAMP1/YitFcs mice shares some of the same genetic elements.

The major histological changes reflected a chronic inflammation with multifocal aggregates of mononuclear cells composed predominantly of B220⁺ B cells and CD3⁺ T cells. These aggregates were primarily in the oxyntic mucosa, with occasional lesions in the forestomach but substantially fewer inflammatory cells infiltrating the antral mucosa. The presence of T cells was associated with mRNA for IFN-γ and TNF-α in the gastric mucosa from the AKR and SAMP1/YitFcs mice, with little evidence of IL-4 or IL-17A (data not shown). To address the pathogenesis of the gastritis more completely, we assessed the role of B cells in the production of autoantibodies. The morphology of the oxyntic mucosa in SAMP1/YitFcs mice was often very perturbed, including loss of parietal cell mass. In approximately half of the SAMP1/YitFcs and AKR mice, pH was significantly increased. Parietal cell-specific antibody-producing cells were detected in the mucosa of the SAMP1/YitFcs mice, while anti-parietal cell antibodies were detected in the majority of the SAMP1/YitFcs mice and ∼50% of the AKR mice. Interestingly, IgA-producing cells were the predominant isotype in the stomach, which is unlike the pattern of isotype expression observed in the MLN of the SAMP1/YitFcs mice (27) or in the mucosa of human subjects with IBD (20) or gastritis (8, 10). Direct examination of the antisera for their ability to react to host cells in the ileum failed to identify an obvious autoantibody reaction (data not shown).

The adoptive transfer of ileitis by CD4⁺ Th cells is exacerbated by the coadministration of B cells (11, 27). It has been suggested that B cells from SAMP1/YitFcs mice express a glucocorticoid-induced TNF receptor-related protein (GITR) ligand, which, when engaged into its receptor GITR, decreases the function of regulatory Th cells (27). B cells in SAMP1/YitFcs mice also produce less IL-10 and transforming growth factor-β (24). Thus it is possible that the B cell defect(s) in SAMP1/YitFcs mice affect(s) immune regulatory function, which is one mechanism known to control gastritis (2, 13, 32, 37). However, reconstitution with T and B cells isolated from SAMP1/YitFcs mice failed to recapitulate the corpus-predominant gastritis observed in SAMP1/YitFcs mice. These observations support the notion that the B cells are producing an autoantibody that is necessary for the corpus-dominant gastritis to emerge and that the accumulation of this antibody and its damage takes >6 wk. The role of B cells in this process is being addressed through long-term reconstitutions, which may provide the time required for the development of autoantibodies and parietal cell damage.

Reconstitution of the recipients with donor cells from SAMP1/YitFcs mice induced a novel duodenitis. The presence of gastritis in the recipients is in agreement with the recent findings of Reuter et al. (33); however, because they did not report examining the duodenum, a direct comparison of their study with the present study is difficult. Both studies used the same H2^k-matched recipients, as described previously (27), which should prevent a major contribution of graft vs. host disease. However, the presence of duodenitis may reflect a minor mismatch in the major histocompatibility complex or the tendency of donor lymphocytes from these mice to promote disease easily in multiple tissues of recipients.

It has been reported that, in addition to ileitis, SAMP1/YitFcs mice develop skin lesions and display hepatic mononuclear cell aggregates (23). However, in another report, “no significant extraintestinal inflammation” was detected in the stomach, liver, spleen, thymus, and various lymph nodes (17). There are several possible explanations, besides reader error, for the varying ability to detect gastritis. 1) The morphology of the SAMP1/YitFcs mice was compared with that of the founder AKR strain (33), which, based on the evidence in the present study, also develops autoimmune gastritis. Such a comparison lacking a negative control would complicate an interpretation, while contrasting the morphology to a strain without gastritis, such as the C57BL/6 strain, reveals the lesion easily. 2) There could also be genetic drift between lines. While possible, this seems unlikely, since by genetic analysis at our institution, founding and current strains appear identical (Drs. Jesus Rivera-Nieves and M. McDuffie, personal communication). Moreover, SAMP1/YitFcs mice were originally derived from AKR mice (41); thus the tendency to develop gastritis may lie in the AKR background, which is embedded in the SAMP1/YitFcs strain. 3) Epigenetic factors may also affect the outcome, as could age and sex. However, lesions are present in males and females and in some mice as young as 3–4 wk old (unpublished observations; 33). 4) The inability to detect gastritis could be due to differences in the microbiome in the mice. This is quite possible, as bacteria, while not necessary, exacerbate the lesions in the SAMP1/Yit mice (5, 33). In other cases, the presence of certain bacterial species attenuates ileitis (21, 22, 30). Together, these observations lead one to conclude that the immunological abnormalities in SAMP1/YitFcs mice facilitate heightened responses in multiple tissues that are modulated by microbial communities.

The presence of gastritis in association with defects in epithelial permeability has been described in SAMP1/YitFcs mice (34). Gastritis in SAMP1/YitFcs mice has been proposed as a model of Crohn's gastritis (33), although these investigators did not report any investigations of autoimmune gastritis and, hence, could not discount it. As the data in the present study have shown unequivocally that B cells are present in the corpus and include autoreactive B cells likely recognizing H⁺-K⁺-ATPase from parietal cells, it is our conclusion that gastritis in these mice is more evident in the corpus in association with the antigen that drives the B cell expansion. It is possible that the epithelial permeability defect favors the translocation of microbes that modifies the host responses in these mice.

Gastritis in humans has many causes (39) and includes ulcerative colitis and Crohn's disease (31). While the SAMP1/YitFcs mice have been used as a model of Crohn's disease, it is not clear that this gastritis resembles the “sterile” Helicobacter-free gastritis associated with Crohn's disease or ulcerative colitis (19). Since mice are coprophagic, it is likely that gastric microbiota play a significant role in modifying the inflammation, making this model very distinct from sterile gastritis associated with Crohn's disease. Crohn's gastritis is characterized as “focally enhanced gastritis” (31), usually limited to superficial regions in the antrum (7). However, in SAMP1/YitFcs mice, lesions included multifocal mononuclear cell infiltrates predominantly in the corpus, usually adjacent to the muscularis mucosae, which is more typical of autoimmune gastritis. While autoimmune gastritis in Crohn's disease has been described in a case report (19), it is not generally recognized as being typical.

In summary, the SAMP1/YitFcs mouse continues to provide an interesting model for studying the pathogenesis of inflammation. However, disease in these mice is not limited to the ileum but is also manifest in the skin and liver and, in particular, the stomach. How these lesions relate to each other remains unclear, although they share genetic control of the inflammation in the stomach and ileum. It is likely that they differ significantly in their tissue-specific effector mechanisms that escape regulation due to one or more abnormalities in the SAMP1/YitFcs mice.

GRANTS

This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-051677 and DK-84063 (to P. B. Ernst), a Research Fellowship from the Canadian Association of Gastroenterology (to K. G. Scott), National Institute of Allergy and Infectious Diseases AI-41236 and AI-51420 (to K. S. Tung), and DK 57880 (M. McDuffie).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.B.E., L.D.E., K.G.S., C.C.B., M.S.A., M.M., J.L.W., R.H.H., and K.S.T. are responsible for conception and design of the research; P.B.E., W.M.L., K.G.S., E.B.W., F.J.T.-V., M.S.A., S.G.B., M.M., S.H.F., G.W.M., I.T.P., R.H.H., and K.S.T. performed the experiments; P.B.E., L.D.E., K.G.S., C.C.B., F.J.T.-V., M.M., J.L.W., G.W.M., and K.S.T. analyzed the data; P.B.E., L.D.E., W.M.L., K.G.S., C.C.B., F.J.T.-V., M.S.A., M.M., J.L.W., and K.S.T. interpreted the results of the experiments; P.B.E., E.B.W., and K.S.T. prepared the figures; P.B.E. and M.M. drafted the manuscript; P.B.E., L.D.E., E.B.W., C.C.B., F.J.T.-V., M.S.A., M.M., J.L.W., R.H.H., and K.S.T. edited and revised the manuscript; P.B.E., L.D.E., W.M.L., K.G.S., E.B.W., C.C.B., F.J.T.-V., M.S.A., S.G.B., M.M., S.H.F., J.L.W., G.W.M., I.T.P., R.H.H., and K.S.T. approved the final version of the manuscript.