Introduction

The human enteric pathogens enteropathogenic Escherichia coli (EPEC) and enterohemorrhagic E. coli (EHEC) are major causes of food poisoning1. Infection by EPEC is associated with childhood mortality in developing countries, whereas infection by EHEC causes hemolytic uremic syndrome2,3. Attachment to intestinal epithelial cells by EPEC and EHEC induces distinctive pedestal-like structures on the host cell surface known as attaching and effacing (A/E) lesions. A related A/E-associated pathogen Citrobacter rodentium is used extensively to study the host-microbe relationship in mouse models4,5.

Mice infected with C. rodentium are susceptible to weight loss and develop soft stool and epithelial crypt hyperplasia6,7. Like EPEC and EHEC, the genome of C. rodentium contains a pathogenicity island known as the locus of enterocyte effacement (LEE)8. The LEE contains genes encoding a type III secretion system, a molecular syringe used by bacteria to inject virulence-associated proteins into the host cell in order to subvert its functions and to enhance the development of disease. The LEE-encoded proteins translocating intimin receptor (Tir) and the bacterial outer membrane adhesin intimin have roles in bacterial virulence and the formation of A/E lesions9. Tir is translocated into the host cell by the type III secretion system to serve as a receptor for intimin9,10,11,12. These proteins are necessary for inducing cytoskeletal rearrangements and actin-rich pedestal formation10,11.

Actin polymerization is an important innate immune mechanism which controls bacterial infection13. Rac-dependent actin polymerization is activated by the guanine nucleotide exchange factor Dedicator of cytokinesis 2 (DOCK2), a mammalian homolog of CED-5 from Caenorhabditis elegans and myoblast city (MBC) from Drosophila melanogaster14,15. The SH3 domain of DOCK2 associates with the C-terminal sequence of engulfment and cell motility (ELMO1)15,16,17. This interaction relieves autoinhibition of DOCK2, which is then allowed to fully activate Rac1. The importance of DOCK2 in the immune system is indicated by its expression in monocytes, macrophages, lymphocytes and other hematopoietic cells18,19,20. DOCK2 functions synergistically with DOCK5 to induce PMA-induced Rac activation, ROS production and formation of neutrophil extracellular traps in mouse neutrophils21. Importantly, dendritic cells defective in DOCK2 exhibit impaired endocytosis of soluble antigens and phagocytosis of insoluble antigens and larger particles22. In addition, DOCK2 contributes to T and B cell migration14,23. These findings highlight important roles of DOCK2 in the immune system. However, the function of DOCK2 in immunity to infectious diseases remains unknown.

Here, we showed that mice lacking DOCK2 were more susceptible to enteric C. rodentium infection. Mice lacking DOCK2 were prone to bacterial dissemination to the systemic organs, had an impaired ability to recruit immune cells and had a reduced capacity to prevent rapid bacterial attachment to the intestinal epithelium compared with wild-type mice. These findings identified DOCK2 as a critical regulator of gastrointestinal immunity to the enteric pathogen C. rodentium.

Results

DOCK2 provides host resistance to C. rodentium infection

We infected wild-type (WT) and Dock2−/− mice via oral gavage with 1 × 1010 CFU of C. rodentium and monitored their survival for 18 days. All WT mice controlled and survived the infection (Fig. 1A), consistent with the phenotype of self-limiting colitis induced by C. rodentium6. Compared with WT mice, Dock2−/− mice were significantly more susceptible, with 39% of the Dock2−/− mice succumbing to the infection by day 13 (P < 0.01; Fig. 1A). Dock2−/− mice lost body weight, especially 10 days after infection (Fig. 1B). Compared with infected WT mice, we found a significantly increased burden of C. rodentium bacteria in the stool of infected Dock2−/− mice on days 10 and 14, but not on days 4 and 7 (Fig. 1C). A significantly higher bacterial number was observed in the colon of Dock2−/− mice on day 14 compared with WT mice (Fig. 1D). Shortening of the cecum and colon – a hallmark of colitis – was more pronounced in infected Dock2−/− mice on day 14 compared with infected WT mice (Fig. 1E).

Figure 1: DOCK2 is required for host protection against C. rodentium infection.
figure 1

(A,B) Survival and body weight change of WT and Dock2−/− mice orally infected with 1 × 1010 CFU of C. rodentium. (C,D) C. rodentium CFU in fecal and colon samples. (E) Lengths of the colons on Days 7 and 14. (F) H&E staining of colon tissues and quantification of crypt length and intestinal damage. Each symbol represents an individual mouse. Data are representative of three independent experiments (mean and SEM). (A) Log-rank test. (B) Two-way ANOVA. (CF) Two-tailed t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not statistically significant.

The increased susceptibility of Dock2−/− mice to C. rodentium infection was validated by histological analysis. Increased crypt lengths and levels of transmissible murine crypt hyperplasia owing to thickening of the mucosa were found in infected Dock2−/− mice on day 14 (Fig. 1F)7. Histological analyses also revealed that Dock2−/− mice developed more severe lesions than WT mice on both days 7 and 14 after C. rodentium infection (Fig. 1F). These results collectively suggested that DOCK2 contributed to the host protection against C. rodentium infection.

DOCK2 mediates resistance to C. rodentium dissemination but is dispensable for the production of cytokines or anti-microbial peptides

A consequence of certain enteric bacterial infection is a breach of the intestinal barrier, causing bacterial dissemination from the gut to the systemic organs of a host. The increased fecal and colon C. rodentium burden in Dock2−/− mice led us to hypothesize that bacteria are readily disseminated to the systemic organs of these mice, which could be responsible for the increased susceptibility and rate of mortality (Fig. 1). We infected WT and Dock2−/− mice via oral gavage with 1 × 1010 CFU C. rodentium per mouse, harvested the spleen, liver and mesenteric lymph nodes (MLNs) 14 days post-infection and analyzed the presence of viable bacteria. We observed significantly more bacteria in the liver and MLNs of infected Dock2−/− mice compared with infected WT mice (Fig. 2A,B). No viable bacteria was detected in the spleen of WT mice, whereas 104 CFU of C. rodentium were found in the spleen of Dock2−/− mice on day 14 (Fig. 2C). Accordingly, infected Dock2−/− mice had enlarged spleen and MLNs compared with infected WT mice (Fig. 2D,E).

Figure 2: DOCK2 mediates resistance to Citrobacter dissemination into systemic organs.
figure 2

(AC) WT and Dock2−/− mice were orally infected with 1 × 1010 CFU of C. rodentium. The bacterial load was determined in the liver, MLNs and spleen 14 days post-infection. (D,E) The weight of the MLNs and spleen on Days 0 and 14 post-infection. Each symbol represents an individual mouse. Data are representative of two independent experiments (mean and SEM). Two-tailed t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not statistically significant.

The production of protective cytokines and anti-microbial peptides are hallmarks of immune responses being mounted towards the infection. We found significantly elevated levels of the pro-inflammatory cytokines, IL-6 and KC (also known as CXCL1) in the colon tissues of infected Dock2−/− mice on days 7 and 14 compared with infected WT mice (Fig. 3A,B). We confirmed these results and found elevated circulating IL-6 and KC in the sera of Dock2−/− mice (Fig. 3A,B). No difference in the levels of these cytokines between uninfected WT and Dock2−/− mice and between infected WT and Dock2−/− mice 4 days post-infection was observed (Fig. 3A,B). Moreover, we measured the levels of additional cytokines that have been shown to orchestrate immunological functions against C. rodentium infection, including IL-17, IFN-γ and TNF7. We found similar levels of IL-17 and IFN-γ in the colon tissues of infected WT mice and Dock2−/− mice and elevated levels of TNF in the colon tissues of infected Dock2−/− mice compared with WT mice 14 days post-infection (Supplementary Fig. S1A).

Figure 3: Production of pro-inflammatory cytokines and anti-microbial peptides in WT and Dock2−/− mice.
figure 3

(A,B) The levels of IL-6 (A) and KC (B) proteins in the colon and serum in uninfected WT and Dock2−/− mice or mice that had been infected with C. rodentium. (CG) Real time qRT-PCR analysis of the expression of the gene encoding RegIIIβ (C), RegIIIγ (D), LCN2 (E), S100A8 (F), and S100A9 (G) in colon tissues of WT and Dock2−/− mice infected with C. rodentium. Data are representative of two independent experiments (mean and SEM). Two-tailed t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not statistically significant.

IL-22 and IL23 have emerged as key players in the host protection against C. rodentium infection24,25,26,27. Of particular importance is that IL-23 drives IL-22-mediated production of antimicrobial peptides within the Reg family, RegIIIβ and RegIIIγ, which critically provides early defense against C. rodentium infection26. We measured the expression of the genes encoding IL-22, IL-23p19, and the anti-microbial peptides RegIIIβ and RegIIIγ in the colon tissues of WT and Dock2−/− mice. We found similar expression levels of the gene encoding IL-23p19 in the colon tissues between WT and Dock2−/− mice on both days 4 and 14 (Supplementary Fig. S1B). The expression of the genes encoding IL-22, RegIIIβ and RegIIIγ were even elevated in infected Dock2−/− mice on both days 4 and 14 (Supplementary Fig. S1C; Fig. 3C,D).

C. rodentium infection also induces production of the anti-microbial peptides LCN2, S100A8 and S100A928. However, we found similar levels of these mediators in the colon tissues of WT and Dock2−/− mice 4 days after infection and increased expression of these mediators after 14 days of infection (Fig. 3E–G). Overall, DOCK2 deficiency did not impair the production of protective pro-inflammatory cytokines and anti-microbial peptides at earlier time points during the infection. The levels of certain pro-inflammatory cytokines and anti-microbial peptides were elevated in infected Dock2−/− mice predominately at later time points, which could be a consequence of the increased number of bacteria in their colon tissues. Taken together, these results suggested that the susceptibility of Dock2−/− mice to C. rodentium infection was largely not owing to the inability to the host to produce pro-inflammatory cytokines and antimicrobial peptides.

DOCK2 is required for infiltration of additional immune cells at later stages of infection

To investigate the role of DOCK2 in mucosal immunity against C. rodentium infection, we used immunohistochemistry techniques to localize macrophages and neutrophils in colon tissues of WT and Dock2−/− mice infected with C. rodentium. Both macrophages and neutrophils were found in the colon of WT and Dock2−/− mice 7 days after infection (Fig. 4A). We counted the number of macrophages and neutrophils localized to the mucosa versus the number localized to the submucosa to determine the relative distribution of these cells at these sites. We observed that the relative distribution of both immune cell types in the mucosa and submucosa was similar in WT and Dock2−/− mice 7 days post-infection (Fig. 4A). However, we found a reduced proportion of the total macrophages and neutrophils infiltrating the mucosa in infected Dock2−/− mice compared with infected WT mice 14 days after infection (Fig. 4B). Abundant numbers of macrophages and neutrophils were found in the submucosa of Dock2−/− mice which failed to infiltrate the lamina propria (Fig. 4B). These data suggested that DOCK2 mediated infiltration of macrophages and neutrophils into the mucosa at later stages of C. rodentium infection.

Figure 4: DOCK2 is required for infiltration of immune cells at late stages of infection with C. rodentium.
figure 4

(A,B) WT and Dock2−/− mice were orally infected with 1 × 1010 CFU of C. rodentium. Immunohistochemistry staining of macrophages and neutrophils from colon tissue sections on Day 7 (A) and 14 (B) post-infection. Relative percentages of macrophages or neutrophils in the mucosa verses their relative percentage in the submucosa per field in WT mice (Day 7, n = 6; Day 14, n = 6) and Dock2−/− mice (Day 7, n = 4; Day 14, n = 6) were determined. At least five different fields for each mouse were quantified. Each bar presents an individual mouse.

DOCK2 impairs bacterial attachment and formation of A/E lesions

The differential ability of macrophages and neutrophils to infiltrate the mucosa in WT and Dock2−/− mice at the later stages of infection could be influenced, in part, by earlier events that establish the infection. C. rodentium employs an attaching and effacing mechanism to attach to enterocytes. This host-pathogen interaction is characterized by effacement of the brush border microvilli and the formation of pedestal-like structures on the host enterocyte7. We hypothesized that DOCK2 is able to prevent bacterial attachment to enterocytes at the earlier stages of C. rodentium infection, which is responsible for the differential production cytokines and antimicrobial peptides, infiltration of immune cells and bacterial dissemination.

To visualize attachment of bacteria to the intestinal epithelium, we performed immunostaining of the C. rodentium virulence factor Tir on colon sections from WT and Dock2−/− mice. Tir is expressed during infection and is translocated into host cells via the Type III secretion system to mediate actin rearrangements and pedestal formation10. We observed an increased frequency of Tir staining covering the intestinal epithelial surface of the distal colon in Dock2−/− mice compared with WT mice 4 days post-infection (Fig. 5A). To investigate whether C. rodentium is directly attaching to the intestinal epithelial cells and whether DOCK2 might be influencing this process, we performed transmission electron microscopy to visualize the bacteria and the presence of associated A/E lesions. We found an intact intestinal epithelium in both uninfected WT and Dock2−/− mice (day 0, Fig. 5B). Remarkably, we frequently identified attachment of C. rodentium to microvilli and induction of pedestal-like structures on enterocytes along the colon of Dock2−/− mice 4 and 7 days after infection (Fig. 5B). In contrast, we did not readily observe these events in WT mice on days 4 and 7. However, we frequently observed attachment of C. rodentium and destruction of microvilli in both WT and Dock2−/− mice on day 10 (Fig. 5B). C. rodentium bacteria were also found throughout the mucosa and in the submucosa of Dock2−/− mice. The increased ability of C. rodentium to attach to enterocytes of Dock2−/− mice was not due to differential expression of genes encoding mucin (Supplementary Fig. S2). Indeed, the expression of genes encoding MUC1, MUC2, MUC3 and MUC4 was similar in the colon tissues of WT and Dock2−/− mice (Supplementary Fig. S2). These results showed that C. rodentium attachment to enterocytes and microvilli destruction occurred earlier in mice lacking DOCK2 compared with WT mice.

Figure 5: DOCK2 impairs bacterial attachment and formation of A/E lesions by C. rodentium.
figure 5

(A) WT and Dock2−/− mice were orally infected with 1 × 1010 CFU of C. rodentium. Intestinal tissue sections were stained with anti-C. rodentium Tir antibody. Percentages of the area of the distal colon surface positive for Tir staining per field in infected WT mice (n = 5) and infected Dock2−/− mice (n = 6). Six different fields for each mouse were quantified. (B) Transmission electron microscopy images of the colonic intestinal epithelium of WT and Dock2−/− mice on Day 0 (uninfected), Day 4, Day 7 and Day 10 post-infection with C. rodentium. Arrowheads indicate bacterial attachment and formation of A/E lesions. Data are representative of two independent experiments (mean and SEM). Two-tailed t-test. ***P < 0.001.

Discussion

Genetic studies revealed that mutations in the gene encoding DOCK2 are associated with colorectal cancer and esophageal adenocarcinoma29,30, raising the possibility that DOCK2 might play an important role in maintaining homeostasis of mucosal surfaces. However, the physiological function of DOCK2 in infectious diseases has remained undefined. A recent study identified in five children biallelic mutations in the gene encoding DOCK2 which largely impair expression of the DOCK2 protein31. These children all suffered multiple immunological dysfunctions and were susceptible to early invasive bacterial and viral infection31. Additional clues for a role of DOCK2 in immunity to infectious disease were highlighted in a study showing that a virulence factor known as Nef encoded by HIV-1 associates with DOCK2 to inhibit chemotaxis of T cells32. This finding suggests that HIV-1 has evolved strategies to specifically prevent DOCK2 functions to overcome the immune system.

Our study identified an important role for DOCK2 in the host defense against the pathogenic bacterium C. rodentium in mice. Infiltration of immune cells to the site of infection is crucial for host defense against pathogens and for mediating their clearance7. We found that DOCK2 is required for recruitment of macrophages and neutrophils to the mucosa where attachment and colonization of C. rodentium occurs. The impaired ability of macrophages and neutrophils to migrate to the mucosa may explain the defective bacterial clearance at the later stages of C. rodentium infection in mice lacking DOCK2. The gene encoding Dock2 is predominately expressed in hematopoietic cells14,18,19,20. Therefore, it is likely that Dock2-deficient macrophages and neutrophils have an intrinsic defect in their ability to infiltrate the mucosa during C. rodentium infection. Given the established role for DOCK2 in orchestrating actin reorganization, it is possible that macrophages lacking DOCK2 may have an impaired ability to induce bacterial uptake even if they encounter the pathogen. Actin polymerization and cell-autonomous immunity are inextricably linked, which allows immune cells such as macrophages to phagocytose and control the number of bacteria per cell13,33. Indeed, we previously demonstrated that dendritic cells defective in DOCK2 exhibit impaired endocytosis of soluble antigens or phagocytosis of insoluble antigens and larger particles22, suggesting that DOCK2 is necessary for the engulfment process.

To identify earlier events leading to bacterial dissemination in vivo, we investigated the ability of C. rodentium to attach to the intestinal epithelium. Mice lacking DOCK2 failed to resist early colonization on enterocytes by C. rodentium via a yet-undefined mechanism. Although the expression of many genes or proteins encoding mucin, pro-inflammatory cytokines and other anti-microbial peptides were not impaired in Dock2−/− mice, it is possible that other anti-microbial factors produced by intestinal epithelial cells or immune cells may account for the differential susceptibility to enterocyte attachment by C. rodentium in WT and Dock2−/− mice.

Rac1 and Rac2 are essential components mediating the function of DOCK2 34. It is possible that DOCK2 signals via Rac1 and/or Rac2 in the host defense against C. rodentium infection. However, unlike mice lacking DOCK2, mice deficient in Rac2 are susceptible to C. rodentium infection only after 12 days of infection35. Further studies directly comparing Dock2−/−, Rac1−/−, and Rac2−/− mice will help clarify the signaling components governing DOCK2-mediated defense during C. rodentium infection.

The functional role of DOCK2 is likely to extend beyond the gastrointestinal tract. Mouse microglial cells lacking DOCK2 have an impaired ability to produce pro-inflammatory cytokines in response to LPS36, indicating a potential immunological role for DOCK2 in the central nervous system. Further studies are required to elucidate the biological functions of DOCK2 in response to different pathogens. In conclusion, our study identified DOCK2 as an important regulator of host defense against C. rodentium infection.

Methods

Mice

Dock2−/− mice have been described previously14. C57BL/6 mice were used as WT controls. Mice were housed in a pathogen-free facility. Animal procedures were approved by, and performed in accordance with the relevant guidelines from the St. Jude Children’s Research Hospital Committee on Use and Care of Animals.

Infection

Citrobacter rodentium (ATCC #51459) was grown in pre-warmed LB broth for 9 h at 37 °C with shaking. Mice were fasted for 4 h prior to infection with 1 × 1010 CFU per mouse by oral gavage. Spleen, liver, colon, MLN and fecal pellets were harvested as described previously37. To determine bacterial counts, serial dilutions of homogenized tissues and fecal pellets were plated on MacConkey agar plates and incubated for 24 h at 37 °C.

Histology and immunohistochemistry

Colons were fixed in 10% formalin, embedded in paraffin, sectioned and stained with H&E as described previously38. C. rodentium Tir was immunostained with C. rodentium-specific Tir antibody (gift from W. Deng and B.B. Finlay, University of British Columbia, Canada). Macrophages and neutrophils were stained with anti-F4/80 (1:500 dilution, MS48000, Caltag) and anti-neutrophils 7/4 (1:2,500 dilution, RM6500, Caltag) antibodies, respectively. Histological findings, including inflammation, edema, hyperplasia, the extent of colonic damage and crypt length, were evaluated at St. Jude Children’s Research Hospital by a pathologist in a blinded fashion.

Electronic Microscopy

Intestinal tissue samples were fixed in 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) and post-fixed for 1.5 h in 2% osmium tetroxide in 0.1 M sodium cacodylate buffer supplemented with 0.3% potassium ferrocyanide. After rinsing in the post-fixed buffer, samples were dehydrated through a series of graded ethanol to propylene oxide buffers, and infiltrated and embedded in epoxy resin, followed by polymerization at 70 °C overnight. Semi-thin sections of 0.5 micron thickness were prepared and stained with toluidine blue for light microscope examination. Ultrathin sections of 80 nm thickness were sectioned and imaged using a FEI Tecnai F 20 TEM FEG Electron Microscope (FEI, Hillsboro) equipped with an ATM XR41 camera.

Cytokine analysis

Colon tissues were homogenized in RIPA buffer supplemented with protease and phosphatase inhibitors (Roche). Levels of cytokines and chemokines in colon homogenates and sera were determined by multiplex ELISA according to the manufacturer’s instructions (Millipore).

Quantitative RT-PCR

RNA was isolated using Trizol, followed by conversion to cDNA as described previously38. Real-time quantitative PCR was performed on an ABI 7500 real-time PCR instrument with 2 × SYBR Green kit (Applied Biosystems) and the appropriate primers (sequences are found in Supplementary Table S1).

Statistical analysis

GraphPad Prism 6.0 software was used for data analysis. Data are shown as mean ± SEM. Statistical significance was determined by t tests (two-tailed) for two groups. Body weight change was compared using two-way ANOVA. Survival curves were compared using the log-rank test. P < 0.05 was considered statistically significant.

Additional Information

How to cite this article: Liu, Z. et al. DOCK2 confers immunity and intestinal colonization resistance to Citrobacter rodentium infection. Sci. Rep. 6, 27814; doi: 10.1038/srep27814 (2016).