SEARCH

SEARCH BY CITATION

Keywords:

  • axenic;
  • interleukin-10;
  • gene-deficient;
  • inflammatory bowel disease;
  • bacterial antigen

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References

Intestinal flora plays a critical role in the initiation and perpetuation of inflammatory bowel disease. This study examined whether live fecal bacteria were necessary for the initiation of this inflammatory response or whether sterile fecal material would provoke a similar response. Three preparations of fecal material were prepared: (1) a slurry of live fecal bacteria, (2) a sterile lysate of bacterial antigens, and (3) a sterile filtrate of fecal water. Each preparation was introduced via gastric gavage into the intestines of axenic interleukin-10 gene-deficient mice genetically predisposed to develop inflammatory bowel disease. Intestinal barrier integrity and degrees of mucosal and systemic inflammations were determined for each preparation group. Intestinal barrier integrity, as determined by mannitol transmural flux, was altered by both live fecal bacterial and sterile lysates of bacterial antigens, although it was not altered by sterile filtrates of fecal water. However, only live fecal bacteria initiated mucosal inflammation and injury and a systemic immune response. Fecal bacterial antigens in the presence of live bacteria and sterile fecal bacterial antigens have different effects on the initiation and perpetuation of intestinal inflammation.

Evidence from the last several decades has revealed that the body's indigenous flora plays an important role in the pathogenesis of inflammatory bowel diseases (IBDs).1–3 Although there is no specific disease-causing bacterium, monoassociation studies have demonstrated that certain individual bacteria can by themselves cause an inflammation in susceptible rodents, whereas other bacteria cannot.4–8 The disease pathology depends on the animal model used and the specific inherent or induced defect of the model. Furthermore, it has been shown that humoral and cellular responses to bacterial antigens drive the perpetuation of the inflammatory state.9 Feeding of fecal material, which includes the whole array of endogenous bacteria, induces a systemic response to various bacterial antigens.10 In the case of monoassociation, the response is directed against antigens from the inoculating bacteria.8

Recently, individual bacterial components have been identified as immunogens involved in intestinal inflammation. Among the bacterial antigens identified so far are bacterial flagellins, histones, superantigens, and peptidoglycans of bacterial cell-wall material.11–14 Colonic intramural or subserosal injection of bacterial cell-wall peptidoglycans can, for instance, cause chronic, spontaneously relapsing enterocolitis in rats.15,16 In addition, bacterial oligonucleotides have been reported to have modulating effects on intestinal inflammation.17

However, it is not clear whether these bacterial antigens can, in the absence of live bacteria, initiate a sustained inflammatory response. To investigate whether oral exposure of bacterial antigens alone may lead to a mucosal and/or systemic immune response resulting in a sustained inflammation, we gavaged axenic interleukin-10 (IL-10) gene-deficient mice with filter-sterilized fecal slurry containing all of the bacterial antigens and compared the results of this exposure to those of an exposure to the same live fecal bacteria.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References

Mice

Originally obtained from the Gnotobiotics facility at North Carolina State University, axenic IL-10 gene-deficient mice, generated on a 129/SvEv background, were maintained in a colony at the University of Alberta under germ-free conditions. The mice in this experiment were used at ≈20 weeks of age. Wild-type and conventionally housed IL-10 gene-deficient mice were maintained under specific-pathogen-free conditions in the animal facility.

Preparation of Fecal Material for Axenic Association

The fecal material for inoculation was prepared 3 different ways to provide preparations of a slurry of live fecal bacteria, a sterile lysate of bacterial antigens, and a sterile filtrate of fecal water.

For each preparation, 20 random fresh fecal pellets were collected from conventionally housed 129/SvEv mice and placed in a sterile 15-mL conical tube containing 3 mL sterile distilled water. All of the equipment used in the following processes was sterile. To construct the first preparation, a slurry of live fecal bacteria, the fresh fecal pellets were broken up with a pestle and vortexed into a thick solution. To construct the subsequent preparations, the slurry of live fecal bacteria was spun down at 4000 rpm for 15 min (5810R Eppendorf centrifuge) to yield a supernatant and pellet. From the pellet, the second preparation of a sterile lysate of bacterial antigens was produced by lysing the pellet in a miniblender at 5000 rpm with glass beads, according to the method described by Dieleman et al,18 modified from a method described by Cong et al.19 This yielded a sterile lysate of fecal bacterial antigens consisting of cell-wall proteins and bacterial genetic material (data not shown). From the supernatant, the third preparation of a sterile filtrate of fecal water was produced by filter sterilizing (Millex-GV 0.22-μm syringe-driven filter) the supernatant into sterile 1.5-mL microfuge tubes. This yielded a sterile filtrate of fecal water devoid of bacterial cell-wall material and protein (data not shown). The sterility of both preparations was confirmed using standard aerobic and anaerobic bacterial culture methods.

Axenic IL-10 gene-deficient mice received 100 μL of the slurry of live fecal bacteria, sterile lysate of fecal bacterial antigens, or sterile filtrate of fecal water by oral gavage for 4 consecutive days. Mice receiving the slurry of live fecal bacteria were subsequently housed in conventional facilities; mice receiving sterile preparations were maintained in axenic facilities until day 7, when all experimentation was performed.

Microscopic Injury Score

Experimental and control mice were killed by cervical dislocation. The axenic status of the experimental mice was tested by analytical culture of the stool. The colon and cecum were harvested from the treated and control mice and cut longitudinally. Approximately half of the colonic and cecal tissue was embedded in paraffin in toto, sectioned at 4 μm, and stained with hematoxylin and eosin for light-microscopic examination. The slides were reviewed in a blinded fashion and were assigned a histological score for intestinal inflammation using a modification of a scheme described by Saverymuttu et al.20 Briefly, histological grades were used to represent the numerical sum of 4 scoring criteria: mucosal ulceration, epithelial hyperplasia, lamina propria mononuclear infiltration, and lamina propria neutrophilic infiltration.

In Vitro Cytokine Secretion by Intestinal Tissue

The other approximately half of the colon, cecum, and ileum was individually rinsed with phosphate-buffered saline containing 50 μg/mL gentamicin (Life Technologies, Burlington, Ontario, Canada) and cut into small fragments. The fragments were blotted on sterile gauze to remove excess liquid and placed in 1-mL complete RPMI consisting of RPMI 1640 (Roswell Park Memorial Institute) medium supplemented with nonessential amino acids, 100 U/mL penicillin, 100 U/mL streptomycin, 2 mmol/L l-glutamine (all from Life Technologies), 2 × 105 mmol/L β-mercaptoethanol (Sigma, St Louis, MO), 50 μg/mL gentamicin, and 5% fetal calf serum (Life Technologies). After a 24-hour incubation period at 37°C under 5% CO2, the contents of the cultured wells were harvested into sterile vials. The vials were centrifuged at 10,000 rpm for 10 min, and the supernatants were collected and assayed to determine the levels of cytokine release.

In Vitro Systemic T Cell Responses

Mice spleens were removed and minced between the frosted ends of slides to prepare a single-cell suspension in complete RPMI with 10% fetal calf serum. Red blood cells were lysed by osmotic shock, and lymphocytes were placed into the wells of 96-well plates at a concentration of 2 × 105/well. Bacterial sonicates were added at a concentration of 50 μg/mL, and the T cell mitogen ConA (Sigma) was added at a concentration of 1 μg/mL. We have described our method for preparation of bacterial sonicates in detail elsewhere.10 Control stimuli included plate-bound anti-CD3η clone 145-2C11 (PharMingen Canada, Mississauga, Ontario) and medium alone. After 48 h of incubation at 37°C in a humidified incubator at 5% CO2, the plates were centrifuged, and the amounts of interferon-γ (IFN-γ) and IL-4 in the supernatants were quantified with standard enzyme-linked immunosorbent assay (ELISA) techniques.

Cytokine-specific ELISA

To assess cytokine production by intestinal fragments, a quantitative, cytokine-specific sandwich ELISA was performed on 96-well, high-protein-binding ELISA plates (Costar, Corning Inc, Corning, NY). The following antibodies were used: anti-IFN-γ and anti-IL-4 (as capture antibodies) and biotinylated anti-IFN-γ and biotinylated anti-IL-4 (as detection antibodies). All antibodies and recombinant cytokine standards were purchased from PharMingen and used at pretitered concentrations to provide optimal results. The color was developed using streptavidin peroxidase (Jackson ImmunoResearch Laboratories, West Grove, PA) and 3,3,5,5-tetramethylbenzidine (Sigma). The reaction was stopped with 1 mmol/L H2SO4. The plates were read in an ELISA reader (SLT Instruments, Salzburg, Austria), and the amounts of cytokine in individual samples were assessed against the standard curve.

Epithelial Barrier Function

To study epithelial transport function, a segment of the colons of monoassociated mice and control mice was assayed in Lucite chambers in the manner described by Madsen et al.21 Briefly, in the chambers, the mucosal and serosal surfaces were exposed to oxygenated Krebs buffer maintained at 37°C. Fructose at a concentration of 10 mmol/L was added to either side. To measure the basal mannitol flux, 10 μCi H3-labeled mannitol was added to the mucosal side. The spontaneous transepithelial potential difference was determined, and the tissue was clamped at zero voltage by continuously introducing an appropriate short-circuit current using an automatic voltage clamp (DVC 1000, World Precision Instruments, New Haven, CT), except for 5 to 10 s every 5 min when potential difference was measured by removing the voltage clamp. Tissue ion conductance was calculated from the potential difference and short-circuit current according to Ohm's law. Potential difference was expressed as millivolts, short-circuit current as microamperes per square centimeter, and tissue ion conductance as millisiemens per square centimeter. Baseline short-circuit current and tissue ion conductance were measured after a 20-min equilibration period. Increases in short-circuit current were induced by addition of the adenylate cyclase-activating agent forskolin (10−5 mmol/L) to the serosal surface. Epithelial responsiveness was defined as the maximal increase in short-circuit current to occur within 5 min of exposure to the secretagogue.

Statistical Analysis

Data are expressed as mean ± SEM. Differences between mean values were evaluated with analysis of variance or paired t tests when appropriate (SigmaStat, Jandel Corp, San Rafael, CA).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References

Both Live Bacteria and Sterile Bacterial Antigens Induce a Defect in Intestinal Barrier Integrity

It is well established that conventionally housed IL-10 gene-deficient mice develop a defect in intestinal barrier integrity that is measurable before the onset of the disease.22 It seems likely that this defect is critical to the initiation and perpetuation of the disease. Heightened permeability may facilitate the passage of endogenous bacteria and/or bacterial antigens to the underlying tissue, where they trigger an immunological response. In the absence of endogenous luminal flora, axenic IL-10 gene-deficient mice do not develop the defect in intestinal barrier integrity (Fig. 1). However, after inoculation with a slurry of live fecal bacteria, a defect in barrier integrity, demonstrated by an increase in mannitol flux (Fig. 1), occurs. Furthermore, although inoculation of axenic IL-10 gene-deficient mice with a sterile filtrate of fecal water had no effect on intestinal barrier integrity, inoculation with a sterile lysate of fecal bacterial antigens caused a defect in intestinal barrier integrity, at the same time point (7 days after feeding) and to a similar magnitude as did association with the slurry of live fecal bacteria.

thumbnail image

Figure FIGURE 1. Colonic epithelial barrier integrity. Colonic epithelial barrier integrity, as measured by transmural mannitol flux in Ussing chambers, is significantly altered in axenic IL-10 gene-deficient mice inoculated with a slurry of live fecal bacteria and sterile lysate of fecal bacterial antigens, but not in mice inoculated with sterile filtrate of fecal water devoid of bacterial antigens. Values represent mean ± SEM from 4 to 8 mice per group. *P < 0.05, compared with untreated axenic group.

Download figure to PowerPoint

Live Bacteria, but Not Sterile Bacterial Antigens, Induce Intestinal Mucosal Injury and Inflammation

Inoculation of axenic IL-10 gene-deficient mice with a slurry of live fecal bacteria induced a marked histopathological injury (Fig. 2) and mucosal inflammatory response (Fig. 3) in both colon and cecum, similar in magnitude to that seen in IL-10 gene-deficient mice housed in conventional facilities from birth. In contrast, the sterile lysate of fecal bacterial antigens that caused a defect in intestinal barrier integrity (Fig. 1) induced neither an intestinal mucosal injury (Fig. 2) nor an intestinal mucosal inflammation (Fig. 3).

thumbnail image

Figure FIGURE 2. Histopathological injury score. Histopathological injury score from colon (A) and cecum (B) is significantly increased in axenic IL-10 gene-deficient mice inoculated with a slurry of live fecal bacteria, but not in mice inoculated with a sterile lysate of bacterial antigens or a sterile filtrate of fecal water devoid of bacterial antigens. Values represent mean ± SEM from 4 to 8 mice per group. *P < 0.05, compared with axenic group.

Download figure to PowerPoint

thumbnail image

Figure FIGURE 3. Mucosal inflammatory response. Intestinal inflammation in colon (A) and cecum (B), as determined by spontaneous mucosal release of IFN-γ, is significantly increased in 24-hour intestinal explant cultures from axenic IL-10 gene-deficient mice inoculated with a slurry of live fecal bacteria, but not from mice inoculated with a sterile lysate of fecal bacterial antigens or a sterile filtrate of fecal water devoid of bacterial antigens. Values represent mean ± SEM from 4 to 8 mice per group. *P < 0.05, compared with axenic group.

Download figure to PowerPoint

Live Bacteria, but Not Sterile Bacterial Antigens, Induce a Systemic Inflammatory Response

Concomitant with intestinal inflammation, IL-10 gene-deficient mice demonstrate a lack of tolerance to exposure to bacterial antigens. Spleen cells from conventionally housed IL-10 gene-deficient mice release high amounts of IFN-γ when stimulated in vitro with antigens derived from pure strains of various endogenous bacteria.8 Similarly, when spleen cells are isolated from IL-10 gene-deficient mice after their inoculation with fecal slurry that contains live bacteria, bacterial stimulation leads to elevated IFN-γ levels compared with spleen cells derived from axenic mice.10 To investigate whether inoculation of axenic mice with sterile fecal material would trigger a systemic response, we stimulated in vitro spleen cells from each group with bacterial sonicates prepared from pure cultures of Lactobacillus reuteri, Clostridium sordellii, Bacteriodes vulgatus, and Enterobacter cloacae.

Only those axenic mice inoculated with a slurry of live fecal bacteria demonstrated a systemic inflammatory response to these pure cultures; those that had been inoculated with a sterile filtrate of fecal water or a sterile lysate of bacterial antigens did not show the same response (Fig. 4). This pattern of response was similar for all the bacterial antigens used for spleen-cell stimulation (Fig. 4).

thumbnail image

Figure FIGURE 4. Systemic inflammatory response. Release of IFN-γ in spleen cell cultures after stimulation with sonicates of pure cultures derived from L reuteri, C sordellii, B vulgatus, and E cloacae. Of the axenic IL-10 gene-deficient mice, only those inoculated with a slurry of live fecal bacteria showed enhanced IFN-γ release in antigen-stimulated spleen-cell cultures similar to the levels measured in cultures from conventionally raised IL-10 gene-deficient mice. Gavage with sterile fecal water or bacterial antigens did not stimulate enhanced IFN-γ release in splenocytes. Shown is the mean value of data derived from 4 to 10 mice per group.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References

In this study, we have shown for the first time that inoculation of axenic IL-10 gene-deficient mice with either sterile fecal water or sterile fecal bacterial antigens does not initiate an intestinal injury or inflammation. This is the case despite the fact that a sterile lysate of fecal bacterial antigens, but not the sterile filtrate of fecal water, was able to cause a defect in intestinal epithelial barrier integrity. In contrast, inoculation of axenic IL-10 gene-deficient mice with a slurry of live fecal bacteria will rapidly induce both a defect in barrier integrity and subsequent intestinal inflammation and injury.

IL-10 gene-deficient mice are genetically susceptible to develop enterocolitis; however, the initiation of the disease appears to depend on the presence of the “right” luminal bacteria. Enterocolitis can be initiated rapidly within 3 days, reaching a maximum intensity in 7 days, in axenic mice gavaged with a slurry of live fecal bacteria that contains the array of endogenous bacteria normally seen in fecal samples.10 Nevertheless, trying to identify which single or group of bacteria from this fecal sample initiates the inflammation has been difficult. Monoassociation studies on axenic IL-10 gene-deficient mice have demonstrated that a limited number of bacterial strains are able to initiate any type of intestinal inflammation when present as the single bacterial species in the gut. Currently, only 3 bacteria have been shown to cause enterocolitis in this genetically modified strain of mice: Enterococcus faecalis, Escherichia coli, and Enterobacter cloacae.4,5 Other tested strains that do not cause disease in a time frame of up to 30 weeks include Viridans group streptococcus, C sordellii, B vulgatus, Helicobacter hepaticus, Pseudomonas fluorescens, Candida albicans, and various Lactobacillus species.4,5,7,8

The reason why some bacteria cause intestinal inflammation, whereas others do not, is largely unknown. The difference may be rooted in the varying effects that individual bacterial species have on the immune system. For example, Gram-positive strains belonging to the genera Lactobacillus, Streptococcus, and Eubacterium have little immunogenic effect in stimulating an immunoglobulin A response or release of IFN-γ in monoassociated mice, whereas Gram-negative strains such as Bacteroides and Escherichia are highly stimulatory.8,23 In addition, Umesaki et al24 have shown that bacteria-induced changes differ between the small and large intestines of mice monoassociated with segmented, filamentes bacteria and/or Clostridium bacteria.

It is well established that loss of tolerance to antigens of the endogenous flora plays a major role in the development of intestinal inflammation. Loss of tolerance occurs as a result of an imbalance in the homeostatic environment and leads to overstimulation of inflammatory cytokines.25 Chandran et al26 hypothesize that quorum-sensing molecules, when secreted in high concentration by bacteria, may alter the state of mucosal tolerance through modulation of the immune system. Moreover, studies on the sera of Crohn's disease patients have revealed that loss of tolerance to the various microbial antigens appears to be patient specific.27 Thus, responsiveness to individual bacterial antigens may in part be dependent on immunological genetic makeup.

Recently, much effort has been concentrated on identifying bacterial components that are recognized by pattern recognition molecules. Bacterial compounds such as bacterial cell-wall material, nucleic acid, flagellin, and superantigens such as lipopolysaccharide have been studied in detail. Their effects on the immune system through interaction with toll-like receptors (TLRs) found on intestinal epithelial cells can greatly influence the immunological homeostasis of the intestinal environment, making them prime candidates for involvement in IBD pathogenesis. Bacterial flagellins have been demonstrated to induce epithelial proinflammatory gene expression after binding to TLR5,28 whereas bacterial DNA has been shown to exert anti-inflammatory effects through interaction with TLR9.29 The gene products of the NOD2/CARD15 gene also are involved in bacterial antigen recognition. The protein product of NOD2 is thought to be a cytoplasmic receptor for muramyl dipeptide, which is produced after the lysozyme-mediated breakdown of the polysaccharide component of bacterial cell-wall peptidoglycans.30,31NOD2/CARD15 is believed to act through the NF-κB pathway to regulate production of tumor necrosis factor and other proinflammatory cytokines.32

In IBD pathogenesis, the intestinal inflammatory response is frequently accompanied by a loss of epithelial integrity. Although it is uncertain whether this defect results from or is the cause of the inflammatory response, it is believed that this defect allows easier passage of bacteria and their antigens; thus, it perpetuates a bacteria-stimulated response in the underlying tissue.

Metabolic stress of the intestinal epithelia in the form of chemicals such as trinitrobenzene sulfonic acid and dextran sodium sulfate can cause immunological imbalances that lead to intestinal inflammatory response.33–35 This response also is characterized by a loss of tolerance for the endogenous flora.36 In the IL-10 gene-deficient mouse, an epithelial defect preexists the onset of the disease.22 This defect appears to be initiated shortly after weaning, when the mice acquire their intestinal flora. In the SCID transfer model, the epithelial defect is detected after the transfer of colitogenic T cells.37 Thus, the defect may require the stimulation of lymphocytes by bacterial antigens.

Our results demonstrate that loss of epithelial integrity can occur in the absence of live bacteria. Stimulation with bacterial components alone, in the absence of any added chemicals, was sufficient to affect the epithelial integrity. However, only the sterile lysate of fecal bacterial antigens that contained sufficient cell-wall material had an effect on epithelial permeability. Although we cannot exclude the possibility that there also were higher concentrations of other bacterial products such as proteins and/or nucleic material in the sterile lysate of fecal bacterial antigens compared with the sterile filtrate of fecal water preparation, these results suggest that loss of epithelial integrity may be dependent on the inclusion of cell-wall material.

Despite the influence that bacterial antigens can exert on the mucosal immune response, it appears that bacterial antigens alone are not sufficient to initiate intestinal inflammation. There are several possible reasons why oral bacterial antigens do not induce inflammation. These reasons include the inaccessibility of the gut-associated intestinal tissue to bacterial antigens, the lack of stimulation of anti-inflammatory responses, and generation of oral tolerance. It is conceivable that in healthy individuals, bacterial antigens may be denied access to the underlying tissue as a result of the glycocalyx coating of the cells. In the absence of live bacteria, bacterial antigens are not newly generated and therefore may have a limited lifetime in the intestinal environment and may not be circulated in the bloodstream in a sufficient amount, consistent with our finding that no systemic response was generated. Alternatively, bacterial antigens may stimulate anti-inflammatory responses that suppress any proinflammatory responses. Haller et al38 have demonstrated that nonpathogenic bacteria can induce anti-inflammatory effects in epithelial cells through the inhibition of NF-κB activation and the induction of transforming growth factor-β production. In this context, it is interesting to note that IL-4 is produced in consistently higher amounts in spleen cells from IL-10 gene-deficient mice inoculated with a sterile lysate of fecal bacterial antigens than in spleen cells from axenic mice (B.C.S., unpublished observation). Finally, feeding of antigens in high concentrations has been shown to induce peripheral deletion of antigen-specific T cells after an initial increase in reactive T cells.39 A large amount of bacterial antigens given orally in the absence of live bacteria may thus stimulate a form of oral tolerance rather than stimulate an inflammatory response.

In summary, we have shown that sterile lysates of fecal bacterial antigens can cause a defect in intestinal barrier integrity but that they cannot initiate intestinal inflammation or injury in mice genetically predisposed to developing IBD. This is in contrast to live fecal bacteria, which both cause a defect in intestinal barrier integrity and lead to the development of intestinal inflammation and injury.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References

The authors wish to acknowledge the University of Alberta Health Sciences Laboratory Animal Services for care of the axenic mouse colony and Valerie Cooper for daily gavage of axenic mice.

References

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. References