Abstract
Resident intestinal bacteria likely play an important role in the pathogenesis of Crohn’s disease through their interaction with the gut immune system. SAMP1/YitFc mice spontaneously develop chronic, discontinuous, transmural ileitis with many features similar to Crohn’s disease. The aim of this study was to determine the effects and elucidate the mechanisms of action of antibiotic treatment in the SAMP1/YitFc mouse model of ileitis. Mice were treated orally with ciprofloxacin and metronidazole before the development of ileitis (prevention protocol) or after ileitis was fully established (treatment protocol). Terminal ilea were harvested for histological scoring, and lamina propria and mesenteric lymph node cells were isolated for analysis of activation markers and cytokine production. Antibiotic therapy significantly decreased the severity of ileitis both in the prevention (40% reduction, p < 0.05) and the treatment (25% reduction, p < 0.01) protocols, compared with untreated, control mice. These effects were associated with a decreased percentage of CD4+/CD45RBhigh lymphocytes in mesenteric lymph nodes of antibiotic-treated mice, as well as decreased production of IFN-γ (prevention: 0.53 ± 0.21 vs 1.84 ± 0.04 ng/ml, p < 0.05; treatment: 8.4 ± 0.4 vs 12.4 ± 0.7 ng/ml, p < 0.005) and TNF (prevention: 61.5 ± 13 vs 134 ± 19 pg/ml, p < 0.01; treatment: 333.5 ± 11 vs 496 ± 20 pg/ml, p < 0.001). The number of activated lamina propria lymphocytes was also reduced after antibiotic treatment. In conclusion, antibiotic therapy significantly ameliorates the severity of ileitis in SAMP1/YitFc mice by a mechanism involving down-regulation of activated gut lymphocytes and inhibition of intestinal Th1 cytokine production.
The etiology of Crohn’s disease (CD)3 remains largely unknown; however, both genetic and environmental factors seem to participate in the pathogenesis of the disease (1). Among the latter, increasing evidence supports a key role for the indigenous bacterial flora in both the initiation as well as the perpetuation of chronic intestinal inflammation characteristic of CD (2). Studies in animal models of colitis clearly show that development of intestinal inflammation is dependent upon the presence of commensal microorganisms, because animals raised in germ-free environment do not develop disease (3, 4, 5).
To further support this theory, manipulation of intestinal flora with antibiotics, especially those against anaerobes or Gram-negative microorganisms, has been proven effective in ameliorating inflammatory bowel disease (IBD) (6, 7, 8, 9). In one study, >60% of patients with active CD experienced a clinical response following a 10-wk course of ciprofloxacin/metronidazole, independent of concomitant corticosteroid therapy (6). The same combination of antibiotics resulted in clinical remission in 45% of patients with moderately active CD (9). In other studies, macrolides in combination with rifabutin (7) and clarithromycin (8) were effective in inducing remission in severe CD. Similar results were obtained from animal studies, which showed that administration of antimicrobial agents can prevent the development of intestinal inflammation or attenuate disease severity in experimental IBD (10, 11, 12, 13). Recently, the administration of probiotic compounds has been effectively used to increase the number of potentially protective microorganisms in the intestinal lumen and control intestinal inflammation in certain clinical settings, as well as in experimental IBD (14, 15).
The mechanism(s) by which indigenous bacteria mediate intestinal inflammation have not been fully elucidated. The most accepted theory is that of a direct interaction between bacteria and bacterial products and the gut immune system. Normally, the gut immune system develops tolerance against Ags from resident intestinal bacteria (16). However, this state of hyporesponsiveness is broken in clinical and experimental IBD (16, 17). Therefore, it is possible that, in a genetically predisposed individual, intestinal bacteria gain access to the lamina propria through a disrupted epithelial barrier, thus providing a constant antigenic stimulus that leads to dysregulated activation of the gut immune system. This, in turn, triggers and perpetuates chronic intestinal inflammatory responses (18).
In recent years, various animal models of intestinal inflammation have been described (19). Although these models provide valuable information in specific aspects of the pathogenesis underlying CD, they suffer from two important shortcomings. First, disease almost uniformly develops in the large intestine, leaving the site primarily involved in CD—the terminal ileum—unaffected. Second, the cause of inflammation is usually a clearly defined genetic, immunologic, or chemical intervention, highly unlikely to contribute to the human condition. The SAMP1/YitFc mouse is a new mouse model that overcomes these deficiencies, because it spontaneously develops chronic intestinal inflammation, characteristically located in the terminal ileum. The SAMP1/YitFc mouse represents a substrain developed at the University of Virginia by continuous brother-sister mating of the parental SAMP1/Yit strain, originally developed by S. Matsumoto (Yakult Central Institute for Microbiological Research, Tokyo, Japan) in Japan (20, 21). Two breeding pairs of the original SAMP1/Yit Japanese mice were kindly provided to our group in 1996. After >20 generations of brother-sister mating, mice in our colony developed several phenotypic changes that justified their description as a new murine substrain, which we have named SAMP1/YitFc.4 The striking similarities between disease in SAMP1/YitFc mice and CD include histological features of transmural, discontinuous infiltration of the bowel wall with both acute and chronic inflammatory cells, occasional formation of granulomas, and prominent hypertrophy of the muscularis propria. Recently, the occurrence of perianal disease was also described in our colony.4 Ileitis in SAMP1/YitFc mice is mediated by activated lymphocytes that heavily infiltrate the lamina propria and have the ability to adoptively transfer disease to MHC-matched SCID mice. (20). Moreover, mesenteric lymph node (MLN) cells from mice with established ileitis produce high levels of TNF and IFN-γ, consistent with Th1-mediated inflammation (20). Similar to human CD, intestinal inflammation in SAMP1/YitFc mice responds to corticosteroids and TNF blockade. The development of ileitis in SAMP1/YitFc mice may require the presence of intestinal bacteria, because the original SAMP1/Yit strain did not develop inflammation when raised in a germfree environment. Interestingly, ileitis was established when bacteria were introduced at a later age (21).
In the present study, to further investigate the role of indigenous bacterial flora in ileitis pathogenesis in SAMP1/YitFc mice, we tested whether 1) antibiotic treatment would be effective in preventing the development and/or attenuating the severity of established inflammation in SAMP1/YitFc mice and 2) the administration of antibiotics would lead to decreased activation of the gut immune system. In this paper, we report for the first time that antibiotic therapy in a spontaneous mouse model of human CD significantly reduces the severity of ileitis by a mechanism that involves down-regulation of activated gut lymphocytes and inhibition of intestinal Th1 cytokine production.
Materials and Methods
Animals and treatment protocols
SAMP1/YitFc mice were maintained under specific pathogen-free conditions in the animal facility of the University of Virginia. A combination of ciprofloxacin (50 mg/kg/day) and metronidazole (100 mg/kg/day) was administered to 40-wk-old SAMP1/YitFc mice (i.e., with established ileitis) for 4 wk by addition to drinking water (treatment protocol). In a similar fashion, 3-wk-old SAMP1/YitFc mice (i.e., before the development of ileitis) were treated for 6 wk with ciprofloxacin/metronidazole immediately after weaning (prevention protocol). In a separate experiment, to evaluate any long-term effects of antibiotic treatment, 40-wk-old mice were given a 4-wk course of ciprofloxacin/metronidazole and sacrificed 6 wk after the cessation of antibiotics. In all protocols, sex-matched littermates consuming regular drinking water served as controls. Mice were allowed ad libitum access to regular or medicated water. Water consumption and animal weight were monitored daily. There were no significant differences in water consumption between antibiotic-treated and control animals. All protocols were approved by the Animal Care and Use Committee of the University of Virginia. Fecal samples from SAMP1/YitFc mice were tested by PCR and consistently found to be negative for the presence of Helicobacter hepaticus, Helicobacter bilis, and other known murine Helicobacter species.
Histological assessment of ileitis
Upon completion of prevention and treatment protocols, mice were anesthetized and sacrificed by cervical dislocation. Terminal ilea were removed, rinsed with cold PBS, rolled over a plastic pipette (Swiss-roll technique), and fixed in Bouin’s solution (LabChem, Pittsburgh, PA). H&E-stained sections were evaluated histologically by a single pathologist (C.A.M.) in a blinded fashion. A validated semiquantitative scoring system was used, as has been reported previously (22). Briefly, scores ranging from 0 (normal histology) to 3 (maximum severity of histologic changes) were used to evaluate histologic indices for 1) active inflammation (neutrophil infiltration), 2) chronic inflammation (lymphocytes, plasma cells and macrophages in the mucosa and submucosa), and 3) villous distortion (flattening and/or widening of normal villous architecture). The total inflammatory index represents the sum of the three individual indices.
Isolation of MLN and lamina propria mononuclear (LPMN) cells
MLN cells were aseptically removed at the time of sacrifice and passed through a 100-μm-cell strainer to obtain single-cell suspensions. LPMN cell isolation was performed as previously reported (20). Briefly, freshly resected terminal ilea were washed with cold HBSS and cut into 2- to 5-mm pieces. Samples were then sequentially incubated in HBSS solution containing dithioerythritol (0.145 mg/ml) and EDTA (0.37 mg/ml) at 37°C for 15 min to remove debris and epithelial cells, and further digested twice in RPMI 1640 containing 100–400 U/ml collagenase VIII (Sigma-Aldrich, St. Louis, MO) at 37°C for 60 min. Lymphocyte-enriched populations were isolated at the 40%/100% interface of a discontinuous Percoll gradient (Amersham Pharmacia Biotech, Uppsala, Sweden).
Cell culture
MLN cells were cultured either unstimulated or under stimulation with immobilized anti-CD3. To stimulate T cells, 24-well plates were incubated with 10 μg/ml of anti-CD3 Abs (BD PharMingen, San Diego, CA) for 90 min at 37°C and washed three times with PBS. MLN cells (1 × 106 cells in 1 ml medium per well) were cultured for 24 or 48 h in complete medium (RPMI 1640 with 10% FBS, 2 mM l-glutamine, and 1% penicillin/streptomycin), and supernatants were harvested for analyses.
Cytokine measurement
Flow cytometry
For the determination of cell surface activation markers, cells in suspension (106 cells per condition) were labeled by 30-min incubation with the appropriate fluorochrome-tagged Abs. Abs against CD4, CD8, CD69, CD44, CD45RB and (BD PharMingen, San Diego, CA) were used. Cells were washed with cold PBS to remove Ab excess and fixed in 1% paraformaldehyde. After gating for lymphocytes by forward and side scatter, the percentage of cells expressing surface markers and the intensity of expression were determined by single- and two-color analyses, using a FACSCalibur (BD Biosciences, San Jose, CA) and CellQuest software (BD Immunocytometry Systems, San Jose, CA).
Statistical analysis
Mean histological scores, mean number of cells expressing markers of activation, and mean cytokine levels between treated and control animals were compared using a two-sided Student t test. Data are expressed as mean ± SEM. An alpha level of 0.05 was considered significant (p < 0.05).
Results
Ciprofloxacin/metronidazole administration prevents the development of ileitis
At the time of weaning (week 3), there was no histological evidence of ileitis in SAMP1/YitFc mice (data not shown). In contrast, by week 9, ileitis was established in control animals (Fig. 1⇓, A and B). SAMP1/YitFc mice, treated with ciprofloxacin/metronidazole for 6 wk, starting from wk 3, demonstrated significantly lower total inflammatory indices by wk 9 (6.2 ± 1.2 vs 10.3 ± 1.2 in untreated controls, p < 0.05) (Fig. 2⇓). Antibiotic-treated mice had significantly lower inflammatory indices for villous distortion (2.6 ± 0.5 vs 4.3 ± 0.5 for controls, p < 0.05), active inflammation (2.4 ± 0.5 vs 3.9 ± 0.5, p < 0.05), and chronic inflammation (1.3 ± 0.2, vs 2.1 ± 0.3, p < 0.05) (Fig. 2⇓). In addition, 6 of 15 animals in the antibiotic group had very low total inflammatory indices (i.e., ≤ 2.5), indicating almost complete prevention of ileitis in 40% of mice (Fig. 1⇓, C and D).
Administration of antibiotics prevents the initial development of ileitis in SAMP1/YitFc mice. Histologic assessment of the ileum from a 9-wk-old SAMP1/YitFc mouse shows discontinuous inflammation, characterized by distorted villous architecture, expansion of lamina propria, and thickening of muscularis propria (A, ×40 original magnification). The inflammatory infiltrate in the lamina propria is composed of both chronic (lymphocytes and macrophages) as well as acute (neutrophils) inflammatory cells (B, ×200 original magnification). Nine-week-old SAMP1/YitFc mouse treated with ciprofloxacin/metronidazole for 6 wk shows complete prevention of ileitis, with normal-appearing villi, no thickening of the muscularis propria (C, ×40 original magnification), and absence of inflammatory infiltration in the lamina propria (D, ×200 original magnification).
Blinded histologic scores of the ilea from SAMP1/YitFc mice receiving regular drinking water or water containing ciprofloxacin/metronidazole for 6 wk, starting before the onset of ileitis (3 wk of age, prevention protocol). Villous distortion, active inflammation, and chronic inflammation were independently scored by a semiquantitative scoring system and added to give a total inflammatory index. n = 16 mice per group. Data are presented as mean ± SEM.
Ciprofloxacin/metronidazole administration ameliorates established ileitis
We next tested whether the combination of ciprofloxacin/metronidazole could be efficient in reversing established, long-standing chronic ileitis characteristic of 40-wk-old SAMP1/YitFc mice (Fig. 3⇓, A and B). Four weeks of antibiotic treatment resulted in significant amelioration of intestinal inflammation (total inflammatory index: 7.7 ± 0.6 vs 10.1 ± 0.6 in controls, p < 0.01) (Fig. 3⇓, C and D, and Fig. 4⇓). The most significant decrease was observed in the villous distortion index (3.3 ± 0.2 in the antibiotic-treated group vs 4.3 ± 0.3 in controls, p < 0.01). However, both chronic and active indices were also significantly decreased in the antibiotic-treated group (2.6 ± 0.3 vs 3.4 ± 0.3, p < 0.05, for active inflammation, and 1.8 ± 0.2 vs 2.5 ± 0.2, p < 0.05, for chronic inflammation) (Fig. 4⇓). Overall, ciprofloxacin/metronidazole administration resulted in a less pronounced amelioration of established ileitis (25% reduction of total inflammatory score) compared with the effects of antibiotics in preventing the initial development of intestinal inflammation (40% reduction of total inflammatory score). No differences were observed in the severity of ileitis between control and ciprofloxacin/metronidazole-treated mice, when animals were evaluated histologically 6 wk after the cessation of antibiotics (data not shown).
The severity of established ileitis in SAMP1/YitFc mice is ameliorated after treatment with antibiotics. Histologic assessment of the ileum from a 44-wk-old SAMP1/YitFc mouse shows severe chronic inflammation with marked villous blunting and thickening of the muscularis propria (A, ×40 original magnification). Lamina propria is expanded with heavy inflammatory infiltrate consisting mainly of chronic inflammatory cells (lymphocytes and macrophages) along with aggregates of neutrophils (B, ×200 original magnification). Forty-four-week-old SAMP1/YitFc mouse treated with ciprofloxacin/metronidazole for 4 wk shows marked histological improvement with minimal thickening of muscularis propria and remodeling of villous architecture (C, ×40 original magnification). High power magnification of the lamina propria shows minimal infiltration with predominantly chronic inflammatory cells and few scattered neutrophils (D, ×200 original magnification).
Blinded histologic scores of the ilea from SAMP1/YitFc mice receiving regular drinking water or water containing ciprofloxacin/metronidazole for 4 wk, starting after the establishment of ileitis (40 wk of age, treatment protocol). Villous distortion, active inflammation, and chronic inflammation were independently scored by a semiquantitative scoring system and added to give a total inflammatory index. n = 25 mice per group. Data are presented as mean ± SEM.
Antibiotic treatment decreases activation of lamina propria and MLN lymphocytes
Previous studies in SAMP1/Yit mice have shown that the establishment of ileitis correlates with heavy infiltration of small intestine lamina propria with CD4+ and CD8+ lymphocytes that are overexpressing surface markers of activation (20). Therefore, we addressed the hypothesis that the beneficial effect of antibiotic administration is mediated by down-regulation of the activation status of intestinal lymphocytes. The lamina propria lymphocyte population of ilea from 9-wk-old SAMP1/YitFc mice receiving regular or medicated water was analyzed by flow cytometry. No difference was observed in the number of CD8+ cells, whereas amelioration of intestinal inflammation following the administration of ciprofloxacin/metronidazole was associated with a marked decrease in the number of CD4+ cells infiltrating the lamina propria (Fig. 5⇓). Accordingly, the number of CD4+ lymphocytes expressing the surface activation markers CD44 and CD69 was also decreased in the antibiotic-treated group (Fig. 5⇓). These data indicate that, following antibiotic administration, the number of effector CD4+ lymphocytes that infiltrate the lamina propria and display an immunologically activated phenotype was decreased. In addition, a lower number of lamina propria lymphocytes expressing the CD4+/CD45RBhigh (naive) phenotype was present in SAMP1/YitFc mice treated with antibiotics, as compared with control animals (Fig. 5⇓).
Flow cytometric analysis of lamina propria lymphocytes from SAMP1/YitFc mice receiving regular water or ciprofloxacin/metronidazole for 6 wk (prevention protocol). LPMN cells were double labeled with CD4-FITC- and CD8-, CD44-, CD69-, or CD45RB-PE-conjugated mAbs. Lymphocytes were gated by forward and side scatter and analyzed by flow cytometry. Results are from a representative experiment using pooled LPMN cells from eight mice per group.
In previous studies, we have shown that MLN lymphocytes from >30-wk-old SAMP1/Yit mice have an activated phenotype with the ability to transfer disease to MHC-matched SCID mice (20). Therefore, we subsequently studied the effects of antibiotic treatment on the MLN compartment in SAMP1/YitFc mice. No differences were observed between control and antibiotic-treated mice in our prevention protocol (data not shown). In contrast, antibiotic administration to SAMP1/YitFc mice with established ileitis resulted in changes in MLN cellularity and composition. The total number of cells present in the MLNs of antibiotic-treated mice was decreased (32.1 × 106 vs 45.7 × 106 in controls). In addition, flow-cytometric analysis of MLN cells from antibiotic-treated and control animals showed that, although no differences in the relative proportions of CD4+ and CD8+ lymphocytes were observed (data not shown), the population of lymphocytes with the CD4+/CD45RBhigh (naive) phenotype was decreased after antibiotic administration (Fig. 6⇓).
Decreased numbers of CD4+/CD45RBhigh lymphocytes in MLNs from antibiotic-treated SAMP1/YitFc mice. Mice were treated with ciprofloxacin/metronidazole for 4 wk (treatment protocol). MLN cells were double labeled with CD4-PE- and CD45RB-FITC-conjugated mAbs and analyzed by flow cytometry. Dot-plot histograms shown are representative of three independent experiments.
Antibiotic treatment inhibits intestinal Th1 cytokine production
MLN lymphocytes from inflamed SAMP1/Yit mice have been shown to express Th1 cytokine polarization with production of high levels of IFN-γ and TNF upon stimulation (20). Therefore, we addressed the question of whether amelioration of ileitis following the administration of antibiotics was associated with a reduced production of proinflammatory cytokines in MLN cells. Production of IFN-γ and TNF from unfractionated MLN cells was elevated at wk 9 and markedly increased by wk 44 (Fig. 7⇓). Preventive administration of ciprofloxacin/metronidazole resulted in a significant decrease of both TNF (61.5 ± 13 pg/ml vs 134 ± 19 pg/ml for the untreated group, p < 0.01) and IFN-γ production (0.53 ± 0.2 ng/ml vs 1.84 ± 0.4 ng/ml for the untreated group, p < 0.05) (Fig. 7⇓, upper panels). Similarly, MLN cells from mice with established ileitis treated with ciprofloxacin/metronidazole produced significantly lower amounts of TNF (333 ± 11 pg/ml vs 496 ± 20 pg/ml for the untreated group, p < 0.001) and IFN-γ (8.4 ± 0.4 ng/ml vs 12.4 ± 0.7 ng/ml for the untreated group, p < 0.005) (Fig. 7⇓, upper panels). In contrast, no significant differences were observed in the production of IL-10 and IL-4 between antibiotic-treated and control animals, in either the prevention or treatment protocols (Fig. 7⇓, lower panels).
Cytokine production from antibiotic-treated and control SAMP1/YitFc mice. MLN cells were cultured for 48 h with immobilized anti-CD3, and secreted IFN-γ, TNF, IL-10, and IL-4 proteins were measured in the culture supernatants by ELISA. Antibiotic-treated SAMP1/YitFc mice showed significantly decreased production of IFN-γ and TNF (upper panels), but no significant differences were detected in the production of IL-10 or IL-4 (lower panels). Data are presented as mean ± SEM.
Discussion
In the present study, we report that administration of broad-spectrum antibiotics both prevents and treats ileitis in SAMP1/YitFc mice, indicating a central role for the indigenous microflora in initiating and perpetuating chronic intestinal inflammation. In addition, we demonstrate that attenuation of ileitis following antibiotic administration is associated with down-regulation of activated lamina propria and MLN lymphocytes in the ileum, as well as decreased Th1-cytokine production. These results provide evidence consistent with the hypothesis that ileitis in SAMP1/YitFc mice is related to a dysregulated immune response toward resident intestinal bacteria.
Our data show that ciprofloxacin/metronidazole were more effective as preventive therapy, with complete absence of ileitis development in a large percentage of mice. In contrast, long-standing ileitis was more resistant to antibiotic therapy. These data are consistent with recent studies reporting greater success in prevention rather than treatment of intestinal inflammation with antibiotics. In HLA-B27 transgenic rats, neither metronidazole nor ciprofloxacin significantly ameliorated established colitis, whereas both were effective in preventing its development (13). Ciprofloxacin or the combination of metronidazole/neomycin prevented the development of colitis in IL-10 gene-deficient mice; however, after disease was established, it was more resistant to antibiotic treatment (11). Recently, it was reported that ciprofloxacin/metronidazole ameliorated disease in acute dextran-sulfate-sodium-colitis but had no effect once the chronic phase was established (10). In our study, various explanations may account for the resistance of ileitis to treatment, as compared with prevention. First, our preventive protocol was started immediately at the time that young mice were weaned. Weaning represents a critical time for the development of gut flora in mice. Studies in BALB/c (23) and IL-10 gene-deficient mice (11) have clearly shown that, during weaning, the introduction of new dietary elements leads to rapid changes in gut microflora, especially in the small intestine (23), including the appearance of new bacterial species (11). In SAMP1/YitFc mice, no indication of ileitis is observed at wk 3, but intestinal inflammation develops during the first 10 wk of age. It is possible that alterations in bacterial flora also take place in SAMP1/YitFc mice after weaning, and new microorganisms provide the antigenic material responsible for initiation of the inflammatory response. Prevention of ileitis by early antibiotic administration can be attributed to qualitative and/or quantitative changes in the composition of this pathogenic indigenous flora. The possibility that food-derived Ags trigger gut immune responses after weaning is not supported by the absence of ileitis in germ-free SAMP1/Yit mice (21). Second, the reduced effects of antibiotic treatment in established ileitis may be related to the augmented Th1 response in older animals, as indicated by the dramatic increase in IFN-γ and TNF production from MLN cells. Therefore, it is unlikely that antibacterial therapy alone is sufficient to down-regulate the activation of intestinal lymphocytes in established ileitis. Combining antimicrobial therapy with anti-Th1 cytokine blockade may prove more effective in treating established ileitis in CD. Finally, ileitis per se seems to be more resistant to antibiotic therapy, in comparison to colitis. This is supported by reports that patients with CD-colitis benefit more from antibiotics than patients with small bowel involvement (6, 24). In addition, mice with experimental colitis show a better response to antibiotics than the response seen in the present study. Because bacterial counts are higher in colon than ileum and the distribution of microorganisms is similar, this resistance of ileitis, as compared with colitis, cannot be easily attributed to differences in the bacterial flora. Most likely it relates to the fact that the gut immune system is more affluent in the distal part of the small intestine than in the colon. Therefore, it is possible that immunologic activation in the terminal ileum can be sustained with lower levels of bacterial load.
The mechanism(s) by which indigenous microflora trigger clinical and experimental IBD are not fully understood. Recent studies in patients with CD and animal models of colitis show a direct association between bacterial Ags and activation of gut lymphocytes. Mucosal Abs against intestinal bacteria have been described in patients with IBD (25). Duchmann et al. (16) reported that tolerance normally exists toward autologous intestinal flora but is abrogated in active CD. Similarly, in the trinitrobenzene sulfonic acid model of colitis, induction of intestinal inflammation is associated with loss of tolerance against normal intestinal bacteria (26). Finally, CD4+ MLN lymphocytes from spontaneously colitic C3H/HeJBir mice proliferate and secrete Th1 cytokines upon stimulation with cecal bacterial sonicates (27). Several findings from our study indicate that, in SAMP1/YitFc mice, suppression of bacterial flora with antibiotics down-regulates the gut immune system. First, in the lamina propria compartment, we observed a decrease in the number of CD4+ lymphocytes and, accordingly, in the number of CD4+ lymphocytes expressing the activation markers CD44 and CD69. These results are consistent with a reduction in the number of effector cells after antimicrobial treatment. Second, the decreased cellularity in the MLN supports the concept of decreased recruitment and/or proliferation of gut immune cells following suppression of intestinal bacterial load. Third, we observed a significant reduction in IFN-γ and TNF production by MLN cells following ciprofloxacin/metronidazole treatment. We have previously shown that ileitis in SAMP1/YitFc mice is characterized by increased production of IFN-γ and TNF from MLN cells (20). The suppression of Th1 cytokines by antimicrobial treatment raisesthe possibility that antigenic stimuli originating from intestinal bacteria are responsible for intestinal Th1 polarization in SAMP1/YitFc mice.
An interesting immunologic finding in our study is the decrease in the number of CD4+/CD45RBhigh lymphocytes after antibiotic treatment, both in the lamina propria and MLNs. Previous studies have shown that CD4+/CD45RBhigh cells produce high levels of IFN-γ upon stimulation and induce an IBD-like colitis when they are adoptively transferred into SCID mice (28). Of note, in the CD4+/CD45RBhigh transfer model of colitis, CD4+ lymphocytes from SCID mice with intestinal inflammation proliferate and produce Th1 cytokines upon stimulation with bacterial sonicates (29). In addition, intestinal disease is ameliorated after antibiotic administration (30). The observed decrease in CD4+/CD45RBhigh lymphocytes in both the lamina propria and MLNs of antibiotic-treated mice indicates that this lymphocyte subpopulation may also play an important role in the SAMP1/YitFc mouse model of ileitis.
Recent studies have proposed specific associations between defined bacterial species and intestinal inflammation in animal models of IBD. Rath et al. (31) have clearly shown that Bacteroides vulgatus is capable of inducing colitis in HLA-B27 transgenic rats. However, a recent publication from the same group reported that ciprofloxacin prevented the development of colitis in this model without affecting the number of colonic Bacteroides species and that the broad-spectrum vancomycin/imipenem treatment was more effective than metronidazole alone in disease attenuation (13). Similarly, in IL-10-deficient mice, Enterococcus faecalis has been proposed as an important pathogen for the development of colitis (32). However, treatment with metronidazole/neomycin resulted in prevention and treatment of colitis despite the fact that the number of adherent or translocated Enterococcus species was actually increased (11). These results indicate that, although the role of well-defined microorganisms may be important in certain animal models of IBD, the primary mechanism of action of antibiotics appears to be the ability to suppress the overall intestinal bacterial load. This can also explain results from human studies where antibiotics with diverse antimicrobial spectra have proven to be effective in ameliorating intestinal inflammation in CD (7, 8, 33). The combination of ciprofloxacin/metronidazole, used in our study, provides a wide antimicrobial spectrum that covers the majority of microorganisms in the bowel lumen, namely anaerobes and Gram-negative bacteria. Our results cannot discriminate whether the effects of antibiotics were mediated by a decrease in total bacterial counts or by suppression of specific bacterial species. Preliminary data from our group support the former, because administration of either ciprofloxacin or metronidazole as single therapies was capable of attenuating the severity of ileitis in SAMP1/YitFc mice (our unpublished data). The lack of significant histological differences between antibiotic-treated and control animals 6 wk after cessation of treatment indicates that constant suppression of the intestinal bacterial load is needed to maintain the anti-inflammatory effect. Recently, immunomodulating properties have been proposed for both metronidazole (34) and ciprofloxacin (35). These effects cannot be excluded in our study. However, studies in other models have shown that, in doses equal to or higher than the ones used in our protocols, ciprofloxacin and metronidazole exert their actions mainly through their antibacterial properties (11).
The results of our study could have important implications for the management of CD. Contrary to all other models of IBD that uniformly develop colitis, the SAMP1/YitFc mouse spontaneously develops intestinal inflammation that, as in the majority of CD cases, is localized to the terminal ileum. Thus, this model most closely resembles the human condition and offers the opportunity to study the effects and mechanisms of action of antibiotics on small intestinal inflammation. The intriguing possibility exists that common bacterial-derived Ags trigger the dysregulated immune response seen in both SAMP1/YitFc mice and CD. To our knowledge, this is the first animal study describing the ability of antibiotics to ameliorate small intestinal inflammation and down-regulate the activation of intestinal lymphocytes, as well the production of proinflammatory Th1 cytokines, such as IFN-γ and TNF. Interestingly, the first gene shown to be clearly associated with a genetic predisposition to CD (i.e., NOD2) is a member of a family of genes implicated in sensing bacteria in the gastrointestinal lumen (36, 37). With an increased understanding of genetic predisposition to IBD, genetically susceptible individuals could be protected from disease development with the appropriate manipulation of bacterial flora at an early time point. Of note, metronidazole was effective in preventing the postoperative relapse of CD one year after ileal resection (38). Moreover, antibiotics can be used as first-line therapy with the goal of decreasing the antigenic load and down-regulating the gut immune system, potentially augmenting the efficacy of other immunomodulatory therapies for CD. The recent unsatisfactory results of IL-10 treatment in CD (39) suggest that such combination therapies may offer a therapeutic advantage as compared with single-agent treatments.
In conclusion, ileitis in SAMP1/YitFc mice can be prevented and treated with the administration of broad-spectrum antibiotics. Antibiotic administration down-regulates the activity of intestinal immune cells, suggesting that intestinal inflammation in the SAMP1/YitFc mice is mediated by a dysregulated immune response against resident intestinal bacteria.
Acknowledgments
We thank Sharon Hoang and Ming Lei for technical assistance and Dr. Satoshi Matsumoto for kindly providing the original SAMP1/Yit breeding pairs.
Footnotes
-
↵1 This work was supported by U.S. Public Health Service/National Institutes of Health Grants DK-57880, DK-42191, and DK-55812.
-
↵2 Address correspondence and reprint requests to Dr. Fabio Cominelli, Digestive Health Center of Excellence, University of Virginia Health System, P.O. Box 800708, Charlottesville, VA 22908. E-mail address: fc4q{at}virginia.edu
-
↵3 Abbreviations used in this paper: CD, Crohn’s disease; IBD, inflammatory bowel disease; LPMN, lamina propria mononuclear cells; MLN, mesenteric lymph node.
-
↵4 J. Rivera-Nieves, G. Bamias, A. Vidrich, M. Marini, T. T. Pizarro, M. J. McDuffie, S. Cohn, C. A. Moskaluk, and F. Cominelli. Emergence of perianal fistulizing disease in the SAMP1/YitFc mouse, a spontaneous model of chronic ileitis. Submitted for publication.
- Received July 3, 2002.
- Accepted August 27, 2002.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
