S100A4 contributes to colitis development by increasing the adherence of Citrobacter rodentium in intestinal epithelial cells



Last updated:

S100A4 has been implicated in cancer and several inflammatory diseases, but its role in inflammatory bowel disease has not been well investigated. Here, upon infection with Citrobacter rodentium, a model for enteropathogenic Escherichia coli infection in humans, induced the infiltration of a large number of S100A4+ cells into the colon in wild type (WT) mice. Deficiency of S100A4 reduced weight loss, bacterial colonization and colonic pathology. Furthermore, the expression of inflammatory cytokines and the recruitment of macrophages and neutrophils also decreased significantly in S100A4 knock out (S100A4−/−) mice. In vitro, soluble S100A4 directly up-regulated expression of integrin β−1 in intestinal epithelial cells and significantly increased the adherence of C. rodentium to intestinal epithelial cells. Additionally, the effects of S100A4 on the adherence of C. rodentium to epithelial cells could be abolished by a receptor for advanced glycation end products (RAGE)-specific inhibitor (FPS-ZM1). Therefore, these data indicate a novel mechanism for S100A4 that promotes colitis development by enhancing host adhesion and colonization of Citrobacter rodentium through the S100A4-mediated host inflammatory responses.

Enteropathogenic Escherichia coli (EPEC) and Enterohemorrhagic Escherichia coli (EHEC) are major causes of severe diarrhea and death worldwide1. They pose a significant public health risk, especially in developing countries where they contaminate food and water supplies2. EPEC causes infantile diarrhea, and the infection leads to dehydration and death3. EHEC is also a severe health threat, causing hemorrhagic colitis and hemolytic-uremic syndrome, a potentially fatal disease, and the infections happen particularly in young children, elderly people and immunocompromised individuals4,5,6. EPEC and EHEC, called attaching-and-effacing (A/E) pathogens, induce characteristic A/E lesions in the intestinal epithelium, which are important for establishing an infection in the host. They infect their hosts by intimately attaching to the surface of the intestinal epithelium and effacing the brush border microvilli7. Upon infection, A/E pathogens displace the commensal flora and cause intestinal inflammation characterized by crypt hyperplasia, goblet cell depletion, and damage to the epithelium. Additionally, infection with these pathogens induces infiltration of immune cells and edema within the lamina propria8,9. Both EPEC and EHEC are poorly pathogenic in mice, but C. rodentium is a Gram-negative A/E bacterium that specifically infects the mouse colon epithelial cells and causes damage to the epithelial layer10,11. Therefore, infection of mice with C. rodentium is an excellent in vivo model of colitis.

S100 proteins belong to a family of low molecular weight, EF-hand (E- helices and F-helices) calcium-binding proteins that regulate calcium-dependent and calcium-independent processes11. S100A4, also called fibroblast-specific protein 1, is a member of the S100 protein family. S100A4 was initially cloned in metastatic cells and fibroblasts, was identified as a metastasis promoter and has mainly been studied in relation to cancer12. It promotes motility and invasion of existing tumor cells, resulting in aggressive metastasis, and is expressed in various cell types, including fibroblasts, macrophages, and malignant cells13,14,15,16. Intracellularly, S100A4 binds to several targets regulating cytoskeletal dynamics and cell motility and proliferation17. Moreover, S100A4 is secreted from both tumor and non-malignant cells and exerts extracellular effects regulating cell mobility, invasion, and angiogenesis by interacting with annexin II, RAGE, and heparan sulfate proteoglycans18,19,20. In our previous study, we found an additional mechanism of action where S100A4+ cells promote 7, 12-dimethylbenz-(a)anthracene/ 12-O-tetradecanoylphorbol-13-acetate (DMBA/TPA) induced skin tumor development by promoting chronic inflammation21.

In fact, S100 proteins are associated with inflammatory responses. Expression of S100 proteins has been shown in arthritis and ulcerative colitis22. In addition, certain members, such as S100A7, S100A8 and S100A9, have the power to kill bacteria through modulating pH or inhibition of microbial growth23,24. It remains unclear whether S100A4 is involved in intestinal inflammation.

For this purpose, we used S100A4 knock out (S100A4−/−) and wild-type (WT) mice orally challenged with C. rodentium to establish a model for studying the role of S100A4 and its related molecular mechanism in colitis. In the present study, we found that S100A4 contributes to bacterial colonization at sites of infection. The expression of S100A4 is up-regulated in C. rodentium-infected mouse colons. Infection-induced inflammation and colonic pathology are attenuated in S100A4−/− mice. Further mechanistic studies found that S100A4 increased the bacteria adherence to intestinal epithelial cells by up-regulating adherence molecular-integrin β-1 and directly promoted colonization.

The expression of S100A4 is up-regulated in C. rodentium-infected mouse colons

It is still not clear whether S100A4 is expressed in the mouse colon during C. rodentium infection. To investigate the kinetics of S100A4+ cells during this process, C57BL/6 mice were orally inoculated with 2 × 109 CFU C. rodentium in 200 μl PBS. Various tissues were collected both prior to and during C. rodentium infection. The mRNA expression of S100A4 was examined by real time quantitative PCR. As shown in Fig. 1A, similar low levels of S100A4 mRNA in different tissues of uninfected WT mice were observed. However, high S100A4 mRNA levels were detected in colons after C. rodentium infection on day 7 (P < 0.01). The expression of S100A4 mRNA of mice colons on day 0, day7, day 14 and day 21 p.i. were further detected (Fig. 1B). The expression of S100A4 mRNA in mice colons was up-regulated on day 7 and day 14 p.i. (P < 0.01) and down-regulated on day 21 p.i. S100A4+ cells in colons at different time points were then stained by immunohistochemistry (IHC) (Fig. 1C). There were very few S100A4+ cells in the untreated colon. However, C. rodentium infection led to a rapid increase of these cells. There were more S100A4+ cells in the C. rodentium affected colon on day 7 than in untreated colon, which increased more on day 14. The number of S100A4+ cells decreased on day 21. Most S100A4+ cells accumulate in the lamina propria and lymphoid follicle in the colons (Figs. 1C,D) (P < 0.05, P < 0.01). Those consistent results clearly demonstrate that the S100A4 expression in the colons of WT mice was up-regulated during C. rodentium infection.

Figure 1
S100A4, S100A4 contributes to colitis development by increasing the adherence of Citrobacter rodentium in intestinal epithelial cells

The expression of S100A4 is up-regulated in C. rodentium-infected mouse colons. C57BL/6 mice were infected orally with C. rodentium as described in the Materials and Methods. Colons and other tissues were collected at different time points. (A) Real-time quantitative PCR analysis of S100A4 mRNA in various tissues from infected C57BL/6 mice on day 0 and day 7 after infection. GAPDH was used as the reference control, (n = 4); *P < 0.05, **P < 0.01. The mRNA level of liver of non-infected mice is set as 1.00 to calibrate the relative levels in other tissues. (B) Real-time quantitative PCR analysis of S100A4 mRNA in colons from infected WT mice on day 0, 7, 14, 21 after infection. GAPDH was used as the reference control, (n = 4); **P < 0.01. The mRNA level of non-infected mice (day 0) is set as 1.00 to calibrate the relative levels in other days. (C) Immunohistochemical staining for S100A4 (brown) in colon sections from C57BL/6 mice on day 0 (uninfected) and day 7, 14, and 21 after infection (n = 4). (D) The numbers of S100A4+ cells per HPF ( × 200), (n = 4); *P < 0.05, **P < 0.01. The experiment was performed by an observer blinded to the experimental condition. (E) Flow cytometry analysis of the phenotypes of S100A4+ cells in the colons of S100A4+/+.GFP mice after C. rodentium infection 7 days by staining GFP+ cells with CD45, F4/80, CD11c, CD4, CD8 and CD19 antibodies.

Full size image

To identify the cellular source of S100A4 in the colon, S100A4+/+.GFP transgenic mice expressing green fluorescent protein (GFP) under the control of the S100A4 promoter were treated with DSS, and then cells were isolated from colon tissues, were co-stained with cellular marker antibodies for various cell types and were analyzed by flow cytometry. As shown in Fig. 1E, among the S100A4-GFP+ cells, approximately 96.5% were CD45+, mainly S100A4-GFP+ cells expressing myeloid cell markers, 48.4% were F4/80+ and 23.6% were CD11c+. In addition, a small number of the S100A4-GFP+ cells expressed markers of T cells, B cells and granulocytes (Fig. 1E). Double immunofluorescence staining of S100A4 and different cellular markers in the colon tissues showed similar results (Supplementary Fig. 1). In addition, S100A4+ cells seldom expressed α-SMA detected by double staining, showing that they were not fibroblasts (Supplementary Fig. 1).

Infection-induced weight loss and C. rodentium colonization are attenuated in S100A4-deficient mice

To determine the role of S100A4 in the host response to an A/E pathogen in vivo, both S100A4−/− mice and WT mice were orally infected with 2 × 109 CFU C. rodentium in 200 μl PBS. Infection-induced weight loss in mice was monitored for 20 days. As shown in Fig. 2A, weight loss in S100A4−/− mice was minimal and significantly less than that seen in WT mice on days 2, 4, 6, 8 (P < 0.01) and 10 (P < 0.05). Furthermore, C. rodentium colonization of colons in S100A4−/− mice was significantly less than those in WT mice after infection on day 7 (P < 0.05) (Fig. 2B). In addition, S100A4−/− mice had significantly lower fecal counts of C. rodentium on day 7 p.i. (P < 0.05) (Fig. 2C) and lower systemic infection in the mesenteric lymph nodes (MLN) (P < 0.05), spleen (P < 0.01) and liver (P < 0.01) on day 7 p.i. (Fig. 2D–F). Collectively, our data suggested that the counts of C. rodentium were reduced in S100A4−/− mice on day 7 p.i. compared with…