Macroscopic images and histological analysis of stomachs from H. felis-infected INS-GAS, GAS-KO, and B6 wild-type (wt) mice at 12 MPI. H. felis-infected INS-GAS mice showed accelerated gastric corpus metaplasia and dysplasia (indicated by yellow arrows) at 12 MPI, as compared with infected B6 wild-type and GAS-KO mice. A: Macroscopic images of the stomachs at 12 MPI, left: H. felis-infected INS-GAS mice, middle: infected GAS-KO mice, right: infected B6 wild-type mice. B: H&E images of the gastric corpus. Left: H. felis-infected INS-GAS mice; middle: infected GAS-KO mice; right: infected B6 wild-type mice (magnification = original ×60; scale bar = 200 μm). C: H&E images of the gastric antrum. Left: H. felis-infected INS-GAS mice; middle: infected GAS-KO mice; right: infected B6 wild-type mice (magnification = original ×150; scale bar = 100 μm). D: The histological scores for gastric corpus of H. felis-infected INS-GAS, GAS-KO, and B6 wild-type mice. *P < 0.05. **P < 0.01. (n = 6 for each group).

Serum gastrin levels in H. felis-infected INS-GAS and B6 mice gradually increased in time course, while those of infected B6 wild-type (wt) mice were stable. Total amidated gastrin as well as human-specific gastrin-17 (G-17) in serum of infected INS-GAS mice gradually increased at 6, 12, and 18 MPI compared with uninfected mice. In contrast, serum amidated gastrin levels in H. felis-infected B6 wild-type mice did not show significant changes in time course. A: Total (= mouse plus human) amidated gastrin in serum of INS-GAS mice with or without H. felis infection. B: Human-specific G-17 in serum of INS-GAS mice with or without H. felis infection. C: Total amidated gastrin in serum of B6 wild-type mice with or without H. felis infection (y axis unit: pM; n = 8 for each group).

Immunohistochemical studies with anti-Ki-67 and anti-TFF2 antibodies for H. felis-infected INS-GAS, GAS-KO, and B6 wild-type (wt) mice at 18 MPI. The numbers of Ki-67 positive cells in corpus were significantly higher in H. felis-infected INS-GAS mice than other two groups (Figure 5A), while those numbers in antrum was almost unchanged among three groups (Figure 5B) at 18 MPI. With respect to TFF2 positive cells, which represent spasmolytic peptide expressing metaplasia, we also observed widely and strongly positive areas in the corpus of H. felis-infected INS-GAS mice, while minimal spasmolytic peptide expressing metaplasia areas were detected in infected GAS-KO mice. (A) (corpus) and (B) (antrum): immunohistochemistry with anti-Ki-67 antibody. Left: The gastric corpus or antrum of H. felis-infected INS-GAS mouse, middle: infected GAS-KO mouse, right: infected B6 wild-type mouse. (18 MPI, magnification = original ×150; scale bar = 100 μm) C: immunohistochemistry with anti-TFF2 antibody. Left: The gastric corpus of H. felis-infected INS-GAS; middle: infected GAS-KO; right: infected B6 wild-type. (18 MPI, magnification = original ×100; scale bar = 200 μm).

Gastric corpus histology of INS-GAS, GAS-KO, and B6 wild-type (wt) mice without H. felis infection at 24 months of age. INS-GAS mice without H. felis infection spontaneously developed mild gastric corpus dysplasia (indicated by yellow arrow) at the older age, while no changes were detected in uninfected B6 wild-type and GAS-KO mice. A and B: H&E images of the gastric corpus of INS-GAS mice (A) and B6 wild-type mice (B) without H. felis infection of 24 month age. C and D: H&E images of the gastric corpus (C) and the gastric antrum (D) of GAS-KO mice without H. felis infection of the same age as A and B. Magnification= original ×100 (A, B, C); and ×200 (D). Scale bar = 100 μm).

Gastric histology and histopathological scores for H. felis-infected INS-GAS, GAS-KO, and B6 wild-type (wt) mice at 18 MPI. H. felis-infected INS-GAS and B6 wild-type mice showed severe gastric corpus metaplasia and dysplasia (indicated by yellow arrows) at 18 MPI, while GAS-KO mice only showed severe gastritis. In contrast, GAS-KO and B6 wild-type mice showed moderate antral hyperplasia with mild dysplastic change (indicated by blue arrows), while infected INS-GAS mice only showed mild antral hyperplasia. A: H&E images of the gastric corpus. Left: H. felis-infected INS-GAS mice; middle: infected GAS-KO mice; right: infected B6 wild-type mice (magnification = original ×60; scale bar = 200 μm). B: H&E images of the gastric antrum. Left: H. felis-infected INS-GAS mice; middle: infected GAS-KO mice; right: infected B6 wild-type mice (magnification = original ×100; scale bar = 200 μm). C, D: The histological scores for gastric corpus (C) and antrum (D) of H. felis-infected INS-GAS, GAS-KO, and B6 wild-type mice. *P < 0.05. **P < 0.01. (n = 6 for each group).

Gastrin overexpression or deficiency does not alter H. felis colonization or Th1-Th2 immunobalance. DNA copy numbers of H. felis normalized to GAPDH in the corpus were almost the same among three groups, while in the antrum, DNA copy numbers of H. felis normalized to GAPDH were the highest in infected GAS-KO and the lowest in infected INS-GAS mice (Figure 4A). H. felis DNA copy numbers in the antrum were always higher than those in the corpus for each group (Figure 4A). H. felis-specific antibody titers did not show significant differences among three groups, which confirmed gastrin-overexpression or gastrin–deficiency did not affect H. felis colonization (Figure 4B). Cytokine expression levels such as TNF-α and IL-4 in the stomachs did not show significant differences among three groups (Figure 4C). Serum fractions of IgG1 (representing Th2) and IgG2c (representing Th1) for H. felis-specific antibody titers did not show significant differences among three groups (Figure 4D). IgG2c titer was higher than IgG1 titers for all of the mice in each group (except in serum of H. felis-infected INS-GAS mice at 12 MPI), which also indicated Th1 dominancy in H. felis-infected mice for all three groups (Figure 4D). A: Quantitative real-time PCR analysis of genomic DNAs obtained from corpus and antrum of H. felis-infected mice stomachs among three groups at 18 MPI (n = 4 for each group). B: H. felis-specific antibody titers in serum of the mice among three groups at 6, 12, and 18 MPI (n = 6 for each group). C: Quantitative real-time RT-PCR analysis of cytokine expressions such as TNF-α and IL-4 in the mice stomachs among three groups at 12 MPI (n = 3 for each group). D: serum IgG1 and IgG2c fractions for H. felis-specific antibody titers of the mice among three groups at 6, 12, and 18 MPI (n = 6 for each group). *P < 0.05 and **P < 0.01.

The expression of growth factors and MMPs in H. felis-infected INS-GAS, GAS-KO, and B6 wild-type mice. Quantitative real-time RT-PCR analysis revealed growth factors such as Reg I and Amphiregulin, and MMPs such as MMP-9 and -13, were significantly up-regulated in H. felis infected INS-GAS mice compared with infected GAS-KO and B6 wild-type mice (A–B). The up-regulation of MMP-13 was confirmed by immunohistochemistry with anti-MMP-13 antibody (C). A: Real-time RT-PCR analysis of the expression of Reg I and Amphiregulin in the stomachs of H. felis infected INS-GAS, GAS-KO and B6 wild-type mice at 18 MPI (n = 3 for each group). B: Real-time RT-PCR analysis of the expression of MMP-9 and -13 in the stomachs of H. felis-infected INS-GAS, GAS-KO, and B6 wild-type mice at 18 MPI (n = 3 for each group). C: Immunohistochemistry with anti-MMP-13 antibody. Left: The gastric corpus of H. felis-infected INS-GAS, right: H. felis-infected GAS-KO. (18 MPI, magnification = original ×100, scale bar = 200 μm).