|Home | About | Journals | Submit | Contact Us | Français|
Helicobacter pylori-induced gastritis is the strongest singular risk factor for gastric adenocarcinoma. Matrix metalloproteinase-7, MMP-7, is a proteolytic enzyme that can modify the intestinal microbial replicative niche as well as affect tumorigenesis and H. pylori stimulates expression of MMP-7 in gastric epithelial cells in vitro. Utilizing a transgenic murine model of H. pylori-mediated injury, our experiments now demonstrate that gastric inflammation is increased within the context of MMP-7 deficiency, which involves both Th1- and Th17-mediated pathways. Enhanced gastritis in H. pylori-infected mmp-7−/− mice is strongly linked to accelerated epithelial cellular turnover. However, more severe inflammation and heightened proliferation and apoptosis are not dependent on MMP-7-mediated bacterial eradication. Collectively, these studies indicate that H. pylori-mediated induction of MMP-7 may serve to protect the gastric mucosa from pathophysiological processes that promote carcinogenesis.
Helicobacter pylori induces a gastric inflammatory response that persists for decades, which increases the risk for gastric adenocarcinoma (1, 2). A host molecule that may influence disease outcome in conjunction with H. pylori is matrix metalloproteinase-7 (MMP-7), a member of a family of proteolytic enzymes that play important roles in tissue destruction and remodeling (3). MMP-7 is one of a small number of MMPs expressed in polarized glandular epithelium (3), and in the intestinal tract, MMP-7 mediates the production of defensins which regulate microbial colonization (4). However, MMP-7 also influences cellular proliferation and apoptosis and is over-expressed in gastric malignancies (5–8). Further, our group and others have demonstrated that H. pylori increases expression of MMP-7 in vitro (9–11). Therefore, we sought to more clearly define the role of MMP-7 within the context of H. pylori-induced injury in vivo by utilizing a transgenic murine model of MMP-7 deficiency.
H. pylori strain SS1 was grown in Brucella broth with 5% FBS (Gibco) for 18 hours (11). C57BL/6 mmp-7−/− mice were generated by backcrossing mmp-7+/− heterozygous 129 mice (12) to C57BL/6 mice beyond the N12 generation. Six to 8-week old wild-type and mmp-7−/− male C57BL/6 mice were used. Experiments were approved by the Vanderbilt IACUC Committee. Brucella broth containing 5 × 109 H. pylori or broth alone was delivered to mice (13).
Mice were euthanized 12 or 36 weeks post-challenge since infection of INS-GAS mice with H. felis increases MMP-7 expression by 3 months, which persists to 9 months post-infection (14). For one-half of the stomach, linear strips from the squamocolumnar junction through the duodenum were paraffin-embedded (13); one-fourth of the stomach was stored at −80°C for RNA extraction, while the other one-fourth was cultured. H. pylori colonization density was assessed by immunohistochemistry. Inflammation was graded on a 0–3 ordinal scale based on the Sydney System as follows: acute inflammation; Grade 0-no polymorphonuclear (PMN) cells present, Grade 1-focal mild neutrophil infiltration (<10 PMN/high-powered field (HPF)), Grade 2-focal dense neutrophil infiltration (10 to 20 PMN/HPF), Grade 3-diffuse and dense PMN infiltration (>20 PMN/HPF); chronic inflammation (mononuclear cell infiltration independent of lymphoid follicles); Grade 0-no inflammation, Grade 1-mild inflammation (slight increase in mononuclear cells), Grade 2-moderate inflammation (dense but focal mononuclear inflammatory cells), Grade 3-severe inflammation (dense and diffuse mononuclear inflammatory cells).
Seventy-five mice were used: uninfected wild-type (n=5, n=4 at 12 and 36 weeks, respectively); uninfected mmp-7−/− (n=5, n=5 at 12 and 36 weeks, respectively); infected wild-type (n=19, n=7 at 12 and 36 weeks, respectively); and infected mmp-7−/− (n=22, n=8 at 12 and 36 weeks, respectively).
RNA was extracted from mouse gastric tissues using the RNeasy Mini Kit (Qiagen) and cDNA prepared using the iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR for cytokines was performed using the iQ SYBR Green Supermix (Bio-Rad) and the primer sequences: IFN-γ (F: 5′-ACTGGCAAAAGGATGGTGAC-3′ and R: 5′-TGAGCTCATTGAATGCTTGG-3′), IL-17 (F: 5′-GCTCCAGAAGGCCCTCAGA-3′ and R: 5′-CTTTCCCTCCGCATTGACA-3′), and IL-10 (F: 5′-CCAAGCCTTATCGGAAATGA-3′ and R: 5′-TCACTCTTCACCTGCTCCAC-3′) with data standardized to β-actin (F: 5′-CCAGAGCAAGAGAGGTATCC-3′ and R: 5′-CTGTGGTGGTGAAGCTGTAG-3′).
For MMP-7, sections were incubated with a rat anti-MMP-7 antibody (9, 15). For anti-H. pylori, anti-CD3, anti-CD20, anti-Ki67, and anti-caspase-3 staining, sections were rehydrated and placed in heated Target Retrieval Solution (Labvision, Fremont, CA). Endogenous peroxidase was neutralized with 0.03% hydrogen peroxide containing sodium azide, followed by a casein-based protein block (Dako, Carpinteria, CA). Sections were then incubated with either rabbit polyclonal anti-H. pylori (B0471; Dako; 1:100), rabbit polyclonal anti-CD3 (A0452; Dako; 1:50), mouse monoclonal anti-CD20 (M0755; Dako; 1:200) for 30 minutes, or rabbit polyclonal anti-Ki-67 (VP-K451; Vector; 1:2000) or rabbit anti-human cleaved caspase-3 (Promega, Madison, WI; 1:600) for 60 minutes, subjected to the Dako Envision+ HRP/DAB System (Dako), and counterstained with hematoxylin. The number of positively stained cells was quantified per high-powered field examined.
Intergroup comparisons of histology scores were performed by two-tailed Fisher’s exact test and linear mixed-effects models. Comparisons of immunostaining data and cytokine levels were performed by unpaired Student’s t test and ANOVA with Student Newman-Keuls test, respectively. Significance was defined as P≤0.05.
All mice challenged with H. pylori were successfully infected. Twelve weeks post-challenge, there was no evidence of MMP-7 expression in uninfected mice. In contrast, MMP-7 was detected in the stomachs of wild-type mice challenged with H. pylori. Immunolabeling localized exclusively to gastric epithelial cells (Figure 1), which mirrors the pattern of MMP-7 expression previously observed in H. pylori-infected humans (9).
We next infected wild-type and mmp-7−/− mice and surprisingly, H. pylori-colonized mmp-7−/− mice developed more severe gastritis than infected wild-type mice (Figure 1, Figure 2). Detailed analysis of inflammation revealed significantly increased levels of acute inflammation in the antrum (P=0.008) and corpus (P=0.03) 12 weeks post-infection and in the antrum (P=0.03) 36 weeks post-infection (Figure 1, Figure 2). Chronic inflammatory scores were significantly higher 36 weeks post-infection in infected mmp-7−/− mice in the antrum (P=0.007) and corpus (P=0.05), although there was a trend toward higher levels of chronic inflammation in the corpus (P=0.08) 12 weeks post-infection (Figure 1, Figure 2).
We next immunophenotyped inflammatory infiltrates within infected gastric mucosa harvested from wild-type or mmp-7−/− mice at the 36-week time-point. Similar to the pattern for overall inflammation, gastric mucosa from infected mmp-7−/− mice contained significantly more T cells and B cells compared to mucosa from infected wild-type mice (Figure 3A). To delineate potential differences in specific T cell responses, we quantified expression of Th1 (IFN-γ), Th2 (IL-10), and Th17 (IL-17) cytokines within the same samples of colonized mucosa. While levels of these cytokines were not altered in uninfected mmp-7−/− mice (data not shown), levels of IFN-γ and IL-17 were significantly higher and IL-10 levels were significantly lower in infected mmp-7−/− mice compared to infected wild-type mice (Figure 3B).
To determine if differences in inflammation were due to altered levels of colonization, we compared colonization density in wild-type and mmp-7−/− mice infected for 36 weeks. There were no significant differences in levels of colonization (Figure 3C). To ascertain whether altered levels of colonization may precede this time-point, we examined colonization density in a subset of samples obtained from mice infected for 12 weeks and, similar to 36 weeks, there were no differences between infected wild-type or mmp-7−/− mice (Figure 3C). Collectively, these results indicate that MMP-7 deficiency enhances the intensity of Th1- and Th17-mediated responses, independent of colonization density.
Since loss of MMP-7 augmented inflammation, we determined whether epithelial proliferation and apoptosis were differentially affected in a subset of colonized wild-type or mmp-7−/− mice. A trend (P=0.1) toward higher proliferation levels was present in infected versus uninfected wild-type mice at 36, but not 12 weeks post-challenge (Figures 4A, 4B). Infected mmp-7−/− mice exhibited significantly higher proliferation levels than either uninfected mmp-7−/− mice or infected wild-type mice at 12 (P<0.001) and 36 (P<0.001) weeks post-infection (Figures 4A, 4B). In wild-type mice, Ki67+ epithelial cells were tightly clustered within the neck region of the gastric glands; however, staining in mmp-7−/− mice extended bidirectionally from the isthmus (Figure 4A).
Similar to proliferation, apoptosis levels were higher (P=0.05) in infected versus uninfected wild-type mice at 36, but not 12 weeks post-challenge (Figures 4C, 4D). Infected mmp-7−/− mice exhibited significantly higher apoptosis levels than either uninfected mmp-7−/− mice or infected wild-type mice at 12 (P<0.001) and 36 (P<0.001) weeks (Figure 4D). In contrast to the topography of proliferation, apoptotic epithelial cells primarily localized to the upper one-third of the gastric foveolae (Figure 4C). These results suggest that increased levels of cellular turnover in H. pylori-infected MMP-7-deficient mice may be influenced by paracrine signaling from infiltrating immune cells (14, 16, 17).
Persistent inflammation induced by H. pylori likely promotes the development of gastric adenocarcinoma; therefore, our results were somewhat unexpected. However, a potential mechanism through which augmented inflammation may occur in mmp-7−/− mice is via dysregulation of epithelial-derived chemoattractant production. Gastric epithelial cells secrete the chemokine IL-8 in response to H. pylori (18), which establishes a haptotactic gradient towards the epithelial surface. In mouse models of acute lung injury, MMP-7 promotes the formation of a transepithelial gradient of KC, the murine functional homolog of human IL-8, via ectodomain shedding of syndecan-1, a heparan sulfate proteoglycan found on epithelial cell surfaces (19). KC associates with shed syndecan-1, establishing a gradient that drives immune cells toward the alveolar lumen. In mmp-7−/− mice, the ability of inflammatory cells to migrate from the interstitium into the alveolar compartment is impaired due to a lack of this gradient, effectively trapping immunocytes at the epithelial-matrix interface (19). H. pylori-infected mmp-7−/− mice may similarly lack the ability to establish proper chemoattractant gradients within the stomach, thus preventing transepithelial migration of immune cells to sites of microbial replication, manifesting as a phenotype of increased inflammation.
H. pylori persistence is likely due to inadequate adaptive immune responses characterized by insufficient Th1 and Th17 responses and inappropriate regulatory T cell activation (18). We found that gastric mucosa from mmp-7−/− mice contained increased numbers of T cells, and a >2–fold increase in expression of the prototype Th1 and Th17 cytokines, IFN-γ and IL-17, respectively, in conjunction with a concomitant decrease in the Th2/Treg cytokine IL-10. While these effects would be expected to enhance T cell function, there was no reduction in bacterial colonization, indicating that other immune components in mmp-7−/− mice may also be dysregulated. Consistent with this, we demonstrated increased B cell infiltration of the stomach, which has also been implicated in H. pylori persistence (18).
Functions that have been ascribed to MMP-7 include stimulation of apoptosis and proliferation (7, 8). Our results, however, indicate that both of these responses are paradoxically augmented in infected mice lacking MMP-7, similar to heightened levels of inflammation. H. pylori induces an influx of a myriad of immune cells that can stimulate apoptosis via increases in oxidative stress (18). In turn, cell loss reciprocally stimulates the proliferative response of epithelial cell precursors, to compensate for reduced cell mass. We speculate that, in mmp-7−/− mice, increased numbers of infiltrating immune cells further augment this process, accounting for differences in cell turnover.
We did not observe dysplasia or cancer in infected wild-type or MMP-7-deficient mice. However, previous data (20) have indicated that H. pylori-infected mice on a BL/6 background rarely develop gastric cancer prior to 15 months post-challenge. In contrast, H. pylori infection of hypergastrinemic INS-GAS mice (on a FVB/N background) leads to the development of premalignant lesions by 6 weeks and gastric cancer by 24 weeks (13). We are currently generating mmp-7−/− INS-GAS mice to more precisely define the role of MMP-7 in carcinogenesis, which will provide a framework for future studies.
In conclusion, MMP-7 expression is increased by H. pylori and negatively regulates inflammation and epithelial turnover. These findings may not only improve our understanding of H. pylori-induced carcinogenesis, but may also provide mechanistic insights into other malignancies that arise within the context of inflammatory states.
Grant support: NIH CA-116087, DK-58587, and DK-77955 (RMP), DK-075161 and DK-053620 (KTW), and F13GM-083500 (NDL), and the Office of Medical Research, Department of Veterans Affairs (KTW).