Alteration of Vβ Usage and Cytokine Production of CD4⁺ TCR ββ Homodimer T Cells by Elimination of Bacteroides vulgatus Prevents Colitis in TCR α-Chain-Deficient Mice¹

Recent advances in the understanding of the immunologic aspects of the gastrointestinal tract pointed to the disturbance of the intestinal mucosal immune system as a possible pathogenic factor in inflammatory bowel disease (IBD),³ e.g., Crohn’s disease, and ulcerative colitis (1). Impressive results in this area have been obtained by recent studies that treated animals suffering from colitis with chemical reagents such as haptens. For example, dextran sulfate sodium was orally administered (2), and trinitrobenzenesulfonic acid or oxazolone together with ethanol were rectally administered as a “barrier breaker” (3, 4, 5). In addition, the availability of specific gene-manipulated mice that spontaneously develop IBD-like chronic colitis as they age further enhanced our understanding of contribution of the mucosal immune system in IBD (6).

By 16–20 wk of age, TCR α-chain-deficient (TCR α^−/−) mice spontaneously develop chronic colitis, which shares some of the histologic and immunologic features of ulcerative colitis in humans (7, 8). For example, mice afflicted with this disease lack αβ T cells and develop a unique subset of T cells that express homodimeric forms of TCR β-chains without TCR α-chains (ββ T cells) and increased food Ag- and autoantigen-specific Ab responses (8, 9). Excessive development of Th2-type CD4⁺, ββ T cells accounted for the induction of IgG and IgA Ab-producing cells in the intestinal mucosa-associated tissues, i.e., colonic lamina propria (LP), Peyer’s patches (PP), and mesenteric lymph nodes (MLN). These CD4⁺, ββ T cells are thought to play a critical role for the development of IBD in TCR α^−/− mice (8, 9, 10). On the other hand, the role of excessive B cell responses in the mucosa-associated tissues of diseased mice is controversial. A recent study, which showed that double mutant mice (TCR α^−/− × Ig μ^−/−) lacking mature B cells developed more serious disease, concluded that B cells and Abs played a suppressive role in the development of IBD via the clearance of apoptotic bodies (11).

Our previous study demonstrated that diseased TCR α^−/− mice suffered from food sensitivities not seen in TCR α^−/− mice without colitis, presumably because of the higher incidence in their colonic LP of plasma cells producing IgG, IgE, and IgA Abs that are reactive with food proteins such as soybean and wheat (8). Further, immunopathologic CD4⁺, ββ T cells have been shown to react with some bacterial Ags (8, 9). Given these findings, we speculated that some luminal Ags, including foods and normal bacterial flora, might be involved in the development of IBD in TCR α^−/− mice. To examine this hypothesis, we fed TCR α^−/− mice an elemental diet (ED) to alter their intestinal microenvironment and then compared the histologic and immunologic characteristics of the ED-fed mice with those of diseased regular diet (RD)-fed mice.

TCR α^−/− mice with a background of 129 × C57BL/6 strain (H-2^b) were obtained from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 mice were purchased from Clea Japan (Tokyo, Japan). The mice were maintained under specific pathogen-free conditions in the Experimental Animal Facility at the Research Institute for Microbial Diseases, Osaka University (Osaka, Japan). All mice received autoclaved distilled water ad libitum, and the cages were changed every 5 days. When the mice reached 8 wk of age, they were fed ED (ELENTAL, Ajinomoto, Tokyo, Japan), which contains 17 kinds of chemically defined amino acids, dextrin, a small amount of soybean oil, trace elements, vitamins, and no protein. A control group continued to be fed RD (Certified diet MF, Oriental Yeast Co., Osaka, Japan). All mice fed ED or RD were sacrificed and examined at 16 wk of age, since our previous studies showed that most mice fed RD developed some signs of IBD (7, 8).

For histologic examination, the tissue samples obtained from the colon were fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin using a standard protocol. The diagnosis and grading of colitis were determined by histologic examination as described previously (7). Mucin-producing goblet cells were stained using the periodic acid-Schiff-Alcian blue procedure, as described previously (12).

For the identification of anaerobic bacteria in the colon, direct smears of the cecal contents were examined by the RapID-ANA II System (Innovative Diagnostic Systems, Atlanta, GA) (13) at the bacteriologic laboratory of the Research Institute for Microbial Diseases, Osaka University.

The spleen and MLN were aseptically extirpated, and single-cell suspensions were prepared by a standard mechanical disruption procedure, as described previously (8, 14). Single-cell suspensions of PP lymphocytes and LP lymphocytes were prepared by an enzymatic dissociation method using type IV collagenase (Sigma, St. Louis, MO) as described previously (8, 14). Briefly, PP were carefully excised from the intestinal wall and then dissociated in Joklik’s modified medium (Life Technologies, Grand Island, NY) containing collagenase. After removal of PP and MLN, the intestine was opened longitudinally, washed thoroughly, and cut into small fragments. Epithelial cells and intraepithelial lymphocytes (IEL) were removed from intestinal tissue by incubating in RPMI 1640 (Sigma) containing 2% FCS, shaking vigorously, and filtering (14). The specimens were then minced and added to Joklik’s modified medium containing collagenase. Cells were dissociated by stirring at 37°C. Then, lymphocytes were isolated using the discontinuous density gradients procedure with Percoll (Amersham Pharmacia Biotech, Uppsala, Sweden) (14).

Total IgA, IgG, and IgM Ab-forming cells were evaluated in mononuclear cells isolated from mucosa-associated and peripheral lymphoid tissues by an ELISPOT assay as described previously (8). Briefly, nitrocellulose microtiter plates (Millipore, Bedford, MA) were coated with 100 μl of goat anti-mouse IgA, IgG, or IgM (Southern Biotechnology Associates, Birmingham, AL) at a concentration of 5 μg/ml in PBS. Then, different concentrations of mononuclear cells (10⁵–10³ cells in 100 μl of RPMI 1640 containing 10% FCS) were added to the plates and incubated in 5% CO₂ at 37°C for 4 h. The plates were washed with PBS, followed by PBS containing 0.05% Tween 20 (Wako, Osaka, Japan). For the detection of Ab-forming cells, alkaline phosphatase-conjugated goat anti-mouse IgA, IgG, or IgM (1 μg/ml; Southern Biotechnology Associates) was added and then visualized with the substrate, 5-bromo-4-chloro-3-indolyl phosphate (Wako)/nitroblue tetrazolium (Wako) in alkaline phosphatase buffer (100 mM Tris-HCl (pH 9.5) containing 100 mM NaCl and 5 mM MgCl₂). Spots were counted using a Leica MZ8 stereomicroscope (Leica, Heerbrugg, Switzerland).

FACS analysis was performed using a FACScan (Becton Dickinson, Mountain View, CA). Cells stained with a single-color reagent were used to set the appropriate compensation levels, and at least 10,000 events were analyzed. The following fluorescence-conjugated mAbs from PharMingen (San Diego, CA) were used: anti-CD4 (clone RM4-5), anti-TCRβ (H57-597), anti-TCR Vβ4 (KT4), anti-TCR Vβ5.1/5.2 (MR9-4), anti-TCR Vβ6 (RR4-7), anti-TCR Vβ7 (TR310), anti-TCR Vβ8.1/8.2 (MR5-2), anti-TCR Vβ8.3 (1B3.3), anti-TCR Vβ9 (MR10-2), anti-TCR Vβ10^b (B21.5), anti-TCR Vβ11 (RR3-15), anti-TCR Vβ12 (MR11-1), anti-TCR Vβ13 (MR12-3), and anti-TCR Vβ14 (14-2). For two-color flow cytometry, 10⁶ cells in 20 μl of PBS containing 2% FCS and 0.05% sodium azide were first incubated with anti-Fc receptor mAb (2.4G2) to prevent nonspecific staining and then stained with the appropriate FITC- and PE-conjugated mAbs (8). All mAbs were used at saturating concentrations. Data were analyzed using CellQuest software (Becton Dickinson) and were shown as log-log dot plots. For cell sorting, mononuclear cells were stained simultaneously for PE-conjugated anti-CD4 (RM4-5) and FITC-conjugated anti-TCRβ (H57-597) and were positively selected with FACSvantage (Becton Dickinson).

Cytokine production by purified CD4⁺, ββ T cells from colonic LP was analyzed by modified cytokine-specific RT-PCR as previously described (15, 16). The mRNA was isolated from FACS-purified CD4⁺, ββ T cells by using TRIzol reagent (Life Technologies), treated with DNase I (Life Technologies), and reverse transcribed into cDNA using PCR buffer (Life Technologies), RNase inhibitor (Toyobo, Tokyo, Japan), oligo(dT)₁₆ (Life Technologies), Superscript II reverse transcriptase (Life Technologies), and dNTPs (Amersham Pharmacia Biotech). The mixture was incubated at 42°C for 120 min and then heated to 90°C for 5 min. After treatment with RNase H (Toyobo), the synthesized cDNA and a series of diluted standard oligonucleotides were quantified with a spectrofluorometer using an OliGreen ssDNA Quantification Kit (Molecular Probes, Eugene, OR). PCR amplification from 10 ng of cDNA for each sample was performed with GeneAmp PCR System 9700 (Perkin-Elmer/Cetus, Branchburg, NJ). Cytokine-specific primers and amplification protocols were described previously (16). RT-PCR was also performed using the same protocol to detect TCR Vβ-specific mRNA expression by Bacteroides-stimulated splenic ββT cells isolated from ED-fed TCR α^−/− mice. The sequences of Vβ and Cβ primers were described previously (8). The amplified products were separated by electrophoresis in 1.8% agarose gel and were visualized with ethidium bromide (1 μg/ml)

The strains of Bacteroides spp. isolated from the cecal contents were maintained on GAM broth (Nissui, Tokyo, Japan)-based agar supplemented with hemin (5 mg/L; Wako), menadione (10 mg/L; Wako), and 5% defibrinated rabbit blood (Nippon Bio-Test Laboratories, Tokyo, Japan) at 37°C under anaerobic conditions (17). The colonies of B. vulgatus or B. distasonis grown in the medium were harvested by centrifugation, freeze-dried, and stored at 4°C (18) for later use as bacterial Ags in experiments. The commensal aerobic bacterial Ags were prepared under similar aerobic conditions and used as controls.

For in vitro experiments, bacterial cells were disrupted by ultrasonic disruptor UD-21 (TOMY, Tokyo, Japan) for 5 min intermittently on ice. After unlysed cells were removed by centrifuging at 5000 × g for 5 min, the lysates were filter-sterilized using a 0.2-μm syringe filter and quantified by Coomassie Protein Assay Reagent (Pierce, Rockford, IL).

To examine the effect of B. vulgatus on CD4⁺, ββ T cells, spleens were aseptically removed from ED-fed TCR α^−/− mice without IBD, and single-cell suspensions were prepared in complete RPMI 1640 consisting of sodium bicarbonate, l-glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 U/ml), and 10% FCS. Then, whole cells (10 μg/ml) of lyophilized B. vulgatus, B. distasonis, or commensal aerobic bacteria were added to individual splenic mononuclear cell cultures (10⁶ cells/ml). Staphylococcal enterotoxin B (SEB; Sigma; 10 μg/ml) was used as a positive control (9). These cultures were incubated for 72 h at 37°C in a humidified incubator at 5% CO₂. Then, the TCR Vβ mRNA expression of these in vitro-stimulated CD4⁺, ββT cells was examined by RT-PCR. Further, the cytokine production of these CD4⁺, ββ T cells was quantified by cytokine-specific ELISA.

To examine the effect of B. vulgatus on mucosal ββ T cells, colonic LP CD4⁺, ββ T cells isolated from ED-fed TCR α^−/− mice without IBD were cocultured with bacterial Ag-prepulsed APCs as described previously (19). For the preparation of Ag-prepulsed APCs, splenic mononuclear cells from C57BL/6 mice were prepared and cocultured with the lysates of B. vulgatus (100 μg/ml), B. distasonis (100 μg/ml), commensal bacteria (100 μg/ml), and SEB (10 μg/ml) at 2 × 10⁷ cells/5 ml for 48 h at 37°C. After washing twice, the cells were resuspended at 4 × 10⁶ cells/ml in complete RPMI 1640 as described above. These APCs were then irradiated with 3000 rad before coculture with colonic CD4⁺, ββ T cells. Purified colonic CD4⁺, ββ T cells (4 × 10⁵ cells/well) isolated from ED-fed TCR α^−/− mice without IBD were incubated in the presence of 4 × 10⁵ Ag-prepulsed APCs/well in wells of a 24-well flat-bottom tissue culture plate (Costar 3526, Corning, NY) for 72 h at 37°C in a humidified incubator at 5% CO₂. Then, the cytokine production by these CD4⁺ ββ T cells was quantified by cytokine-specific ELISA of culture supernatants.

For assessment of cytokine production by colonic or splenic CD4⁺, ββ T cells cocultured with B. vulgatus-, B. distasonis-, or commensal aerobic bacteria-prepulsed APCs or with these bacterial Ags in the presence of APCs, respectively, a quantitative cytokine-specific sandwich ELISA was performed using the Biotrak IFN-γ, IL-4, and IL-6 ELISA systems (Amersham Life Science, Aylesbury, U.K.).

To examine the ability of B. vulgatus to induce colitis in this model, disease-free ED-fed TCR α^−/− mice were rectally administered with B. vulgatus (10 μg/ml in PBS, 300 μl/body). Rectal administration was performed using a 1-ml syringe fitted with a gastric incubation needle (Fuchigami-Kikai, Kyoto, Japan) under ketamine anesthesia. Mice were treated at weekly intervals for 4 wk. The same dose of B. distasonis suspension or PBS was administered rectally to disease-free ED-fed mice as a control. Four weeks after the final administration, mice were sacrificed, and histologic examination of the colon was performed by hematoxylin and eosin staining. Th1- and Th2-type cytokine production of infiltrated ββT cells in colonic mucosa was examined by RT-PCR analysis as described above.

Significant differences between mean values were determined by Student’s t test. p < 0.05 was considered statistically significant.

To investigate the role of intestinal microenvironments in the development of IBD, TCR α^−/− mice were fed ED, consisting of chemically defined amino acids without antigenic proteins. TCR α^−/− mice fed RD containing proteins from soybean, wheat, and fish were used as a control group. At 16 wk of age, >80% of the RD-fed TCR α^−/− mice developed colitis, and some of those mice suffered from extracolic lesions, i.e., anorectal prolapse, enlargement of MLN, and splenomegaly (Table I) (7). Histologic examination of the colon of mice afflicted with colitis showed marked hyperplasia of the LP, infiltration of inflammatory cells into the LP, elongation of crypts, decrease of goblet cells, and the presence of crypt abscesses, as described previously (Fig. 1) (7, 8). On the other hand, ED-fed TCR α^−/− mice revealed neither the macroscopic inflammatory change in the colon nor the extracolic lesions (Table I). They also showed no alteration of their histologic features. The numbers of their goblet cells were comparable to those seen in normal (data not shown) and TCR α^−/− mice without IBD (Fig. 1).

Bacteroides spp. recovered from the cecal contents of RD- and ED-fed TCR α^−/− mice were compared (Table II) (13). Colonies of B. vulgatus were detected in 80% of the overall and in 88% of the diseased RD-fed TCR α^−/− mouse population, but were not found in ED-fed TCR α^−/− mice. Furthermore, B. vulgatus-specific Abs were detected in the serum of all diseased RD-fed mice (data not shown). Interestingly, colonies of B. distasonis were detected in 83% of ED-fed mice, but were not detected in RD-fed TCR α^−/− mice. These findings were not observed in RD- and ED-fed TCR α^+/+ mice (data not shown). RD- and ED-fed TCR α^−/− mice did not significantly differ in their detection of other bacterial species of the flora, such as Escherichia spp., Enterococcus spp., Enterobacter spp., Klebsiella spp., and Proteus spp. (data not shown).

Increased mucosal B cell responses are one of the unique immunologic features of TCR α^−/− mice with IBD (8). To examine the effect of ED on the aberrant B cell responses, isotype-specific ELISPOT assays were performed on mononuclear cells isolated from the spleen, MLN, and colonic LP of ED- and RD-fed TCR α^−/− mice. In both peripheral and mucosa-associated tissues, the frequencies of IgA and IgG Ab-forming cells in ED-fed TCR α^−/− mice without IBD were lower than those in RD-fed TCR α^−/− mice with IBD (Fig. 2). For example, the colonic LP lymphocytes of ED-fed TCR α^−/− mice exhibited only 20% of the number of IgA Ab-forming cells detected in that of RD-fed TCR α^−/− mice with IBD. These marked reductions of Ab-forming cells in colonic LP were not found in ED-fed TCR α^+/+ mice (data not shown). These findings suggest that the ED-induced alteration of intestinal microenvironments affected the development of the aberrant colonic B cell responses found in diseased TCR α^−/− mice, the role of which in the development of disease is, however, still controversial (8, 11).

The increase in a unique population of CD4⁺, ββ T cells in mucosa-associated lymphoid tissues was considered to play a critical role in the development of IBD (8). To assess how ED-modified intestinal microenvironments might affect the development of CD4⁺, ββ T cells, the frequencies of these T cells in the mucosa-associated and peripheral lymphoid tissues of ED-fed TCR α^−/− mice without IBD or diseased RD-fed TCR α^−/− mice were examined and compared by flow cytometric analysis (Fig. 3,A). Interestingly, similar frequencies of CD4⁺, ββ T cells were noted in the mucosal and peripheral tissues of both groups. The absolute numbers of CD4⁺, ββ T cells infiltrated in the colonic LP were also similar (Fig. 3 B). These findings indicate that the alteration of intestinal microenvironments induced by feeding with ED did not affect the development of CD4⁺, ββ T cells.

Having assessed the qualitative alterations in CD4⁺, ββ T cells induced by feeding with ED, we next analyzed Th1- and Th2-type cytokine production by CD4⁺, ββ T cells isolated from the colonic LP of ED- or RD-fed TCR α^−/− mice. The CD4⁺, ββ T cells were purified from colonic LP mononuclear cells by FACS, and the profile of Th1- and Th2-type cytokine mRNA expression was determined by cytokine-specific RT-PCR. The CD4⁺, ββ T cells isolated from the colonic LP of diseased RD-fed TCR α^−/− mice expressed mRNA specific for IFN-γ, IL-4, IL-6, and IL-10. In contrast, IL-4, IL-6, and IL-10 were not detected in the CD4⁺, ββ T cells isolated from the colonic LP of ED-fed TCR α^−/− mice without colitis (Fig. 4). These findings demonstrate that the modification of intestinal microenvironments by ED feeding results in the removal of Th2-type responses.

To further analyze the ED-induced qualitative alterations in CD4⁺, ββ T cells, flow cytometric analysis of Vβ repertoires was performed. The major Vβ repertoires used by CD4⁺, ββ T cells isolated from ED-fed TCR α^−/− mice without IBD were Vβ6 and Vβ14, followed by Vβ11, Vβ8, and Vβ4 (Fig. 5). In contrast, CD4⁺, ββT cells isolated from RD-fed TCR α^−/− mice with IBD showed a predominance of Vβ 8 (8, 9), although they also expressed, to a lesser degree, some of the other Vβ subfamilies seen in ED-fed TCR α^−/− mice without colitis. These findings suggest that ED-induced alterations in the intestinal microenvironments change the dominant expression of Vβ in CD4⁺, ββT cells from Vβ 8 to Vβ 6, Vβ11, and Vβ14.

Since feeding with ED was found to eliminate B. vulgatus and so inhibit disease development, it was important to determine the effect of B. vulgatus on the pathogenesis of CD4⁺, ββ T cells. The ability of B. vulgatus to direct CD4⁺, ββ T cells to produce Th1- and/or Th2-type cytokines was examined in vitro (Fig. 6). CD4⁺, ββ T cells isolated from colonic LP of ED-fed TCR α^−/− mice without IBD were cocultured with APCs prepulsed with lysates of B. vulgatus, B. distasonis, or commensal aerobic bacteria. Although some spontaneous Th1- and Th2-type cytokine productions were noted in colonic ββ T cells cocultured with nonprepulsed APCs, B. vulgatus enhanced the secretion of Th2-type cytokines (e.g., IL-4 and IL-6), while B. distasonis instead enhanced the secretion of Th1-type cytokines (IFN-γ; Fig. 6,A). B. vulgatus was also found to enhance the secretion of Th2-type cytokines in splenic ββ T cells, while B. distasonis instead enhanced the secretion of Th1-type cytokines (Fig. 6 B). These results suggest an interesting possibility that a selective intestinal bacterium (e.g., B. vulgatus) directs preferential induction of Th2-type cytokines by the pathogenic CD4⁺, ββ T cells.

Next, we used Vβ-specific RT-PCR analysis to examine the effect of B. vulgatus and B. distasonis on the Vβ expression of splenic ββ T cells (Fig. 7). ββ T cells from B. vulgatus-treated mononuclear cell cultures expressed predominantly Vβ8-specific mRNA, while ββ T cells cultured with B. distasonis expressed a diversified repertoire of Vβ (i.e., Vβ6, Vβ8, Vβ11, and Vβ14; Fig. 7). These findings suggest that the presence of B. vulgatus in the gut lumen of TCR α^−/− mice may lead to the induction of Th2-type ββ T cells with a limited TCR Vβ repertoire (e.g., Vβ8) for the development of IBD.

To examine the direct effect of B. vulgatus on the onset of colitis in TCR α^−/− mice, B. vulgatus was administered rectally to disease-free ED-fed TCR α^−/− mice. Histologic examination of B. vulgatus-treated TCR α^−/− mice showed marked inflammatory changes in colon and rectum, such as hyperplasia of the LP region, elongation of crypts and microvilli, infiltration of inflammatory cells into the LP region, and decrease in the number of goblet cells (Fig. 8). In contrast, PBS-treated mice showed no sign of inflammation, and B. distasonis-treated mice revealed only minor inflammatory changes (Fig. 8). To analyze the cytokine profile of infiltrated ββT cells in these B. vulgatus-treated and diseased mice, Th1-/Th2-type cytokine-specific mRNA expression was examined in FACS-purified CD4⁺, ββT cells by RT-PCR analysis. The CD4⁺, ββ T cells isolated from the colonic LP of B. vulgatus-treated mice expressed mRNA specific for IFN-γ, IL-4, IL-5, IL-6, and IL-10. This cytokine profile was similar to those found in the infiltrated colonic ββ T cells isolated from RD-fed diseased mice. In contrast, a lack of Th2-type cytokines was noted in the colonic ββ T cells isolated from PBS-treated ED-fed mice (Fig. 9). These findings suggest that the exposure of colonic mucosa to B. vulgatus induces colitis with infiltration of Th2-type ββT cells.

Results from several studies strongly suggest that immunoregulatory cells, particularly CD4⁺ T cells in intestinal mucosa-associated tissues, play an important role in the pathogenesis of Crohn’s disease or ulcerative colitis (20, 21, 22). However, the pathogenic Ags in human IBD and the characteristics and behavior of the disease-specific subsets of CD4⁺ T cells remain unknown. Using new and unique IBD models, researchers have determined that specific subsets of CD4⁺ T cells, including CD4⁺, CD45RB⁺ T cells, Tr1 cells, and Th1-type cells, play a critical role in disease development (23, 24, 25, 26). Other investigators found that TCR α^−/− mice with IBD showed aberrant humoral responses, including the production of Abs against autoantigens or food Ags (8, 27). The development of these aberrant B cell responses was influenced by a unique subset of CD4⁺, ββ T cells with Th2-type cytokine synthesis (8, 9, 28, 29).

Along with such immunologic alterations, an intestinal microenvironment consisting of foods and bacterial flora is also considered to be a causative factor for IBD (30). The findings of the present study suggest a crucial role for luminal environmental Ags in the development of IBD in TCR α^−/− mice. ED decreases intestinal physiologic functions such as peristalsis and peptic secretion as well as immunologic functions. Further, it is likely that ED can be fully absorbed in the upper jejunum, with the colon remaining in a quiescent mode. ED feeding resulted in the alteration of the enteric microflora in a quantitative as well as qualitative manner (31, 32). Thus, the ED feeding of TCR α^−/− mice may lead to qualitative and quantitative alteration of commensal anaerobic bacteria, such as the reduction or loss of B. vulgatus, and so prevent the oligoclonal expansions of Th2-type CD4⁺, ββ T cells, thus protecting against the development of IBD. Our results provide new evidence that the alteration of enteric flora by feeding with ED could prevent the development of IBD.

A selective population of patients with IBD has been treated with nutritional therapy using ED (33). The ED treatment sustained remission, especially in patients with Crohn’s disease (34, 35, 36), probably by reducing digestive processes and so allowing the bowel to rest (37). However, recent studies of IBD in rats indicate the suppressive effect of ED on the mucosal immune system (38, 39). Rats fed ED were found to have lower numbers of CD4⁺ T cells and IgA-containing cells than rats fed ED containing oleic acid (38). Moreover, ED reduced the progression of peptidoglycan-polysaccharide-induced enteritis by modulating the activation of T cells, the production of NO, and the generation of oxygen free radicals (39). Although the present study did not find ED feeding to significantly alter the number of CD4⁺, ββT cells in either IBD-afflicted or healthy TCR α^−/− mice, it did show that the quality of those cells was drastically changed by feeding with ED (e.g., reduction of Th2 cytokine synthesis). This qualitative alteration of CD4⁺, ββT cells resulted in the inhibition of aberrant B cell responses in the ED-fed TCR α^−/− mice. Further, our present study suggests the interesting possibility that the qualitative changes in these T cells were due to the alteration of luminal microenvironments, including enteric flora, by feeding with ED. Supportive of such a conclusion is the finding that B. vulgatus, dominant in the colonic microflora of RD-fed diseased mice, was absent from mice fed ED (Table II). Further, incubation of colonic or splenic ββT cells isolated from ED-fed mice without colitis in the presence of B. vulgatus-stimulation resulted in the generation of Th2-type ββT cells that predominantly expressed Vβ 8 in vitro (Figs. 6 and 7).

Previous findings suggested that IBD could be initiated in TCR α^−/− mice by a specific organism or group of organisms normally present in the enteric microflora (10, 40, 41). TCR α^−/− mice maintained under germfree or limited bacterial flora conditions exhibited no colonic inflammation (40). Disease activity was influenced by the development of CD4⁺, ββ T cells, which are associated with an altered immune response (from polyclonal to oligoclonal) to cecal aerobic bacterial Ags (9, 10). Examination using the appendectomy model for TCR α^−/− mice suggested that the priming of immunopathologic lymphocytes with these microflora Ags occurred in the appendix lymphoid follicles at a young age (41). In a study using C3H/HeJBir mice, a Th1-driven colitis model, CD4⁺ T cells reactive with conventional Ags of the enteric bacterial flora were also found to mediate IBD (19). Although the genus was not identified, these results strongly suggest the important role played by some micro-organisms normally residing in the colon in initial inflammation. Our findings provide new evidence that one such micro-organism, B. vulgatus, may be associated with the development of IBD in RD-fed TCR α^−/− mice, since it was found to be capable of inducing a selected subpopulation of aberrant Th2-type CD4⁺, ββ T cells (Table II and Figs. 6 and 7). In addition, our hypothesis was further supported by the finding that rectal administration of this micro-organism to disease-free ED-fed mice resulted in the development of colitis with mucosal infiltration of Th2-type CD4⁺, ββ T cells (Figs. 8 and 9).

Previous reports indicated that Helicobacter hepaticus, newly recognized as murine Helicobacter, was detected with high frequency in several lines of mutant mice that spontaneously developed colitis (42). IL-10-deficient (IL-10^−/−) mice, which develop human IBD-like enterocolitis spontaneously under conventional conditions, show only minimal inflammatory change under specific pathogen-free conditions (42). However, H. hepaticus infection induced colitis in SPF-reared IL-10^−/− mice accompanied by a Th1-type cytokine response (43). We also examined whether this micro-organism can be detected in the diseased TCR α^−/− mice by PCR analysis using H. hepaticus-specific primers (44). Although multiple tissue samples (i.e., ileum, cecum, colon, rectum, and liver) from the diseased TCR α^−/− mice were examined, no H. hepaticus-specific gene was detected (data not shown). This finding suggests that H. hepaticus is not involved in the development of disease in TCR α^−/− mice.

It was also shown that oral administration of selected antibiotics (e.g., metronidazol) effectively remitted human IBD (45). Metronidazol has also been reported to reduce the incidence of Bacteroides spp. in the enteric microflora of patients with Crohn’s disease (46). Another study suggested a pathogenic role for B. vulgatus in ulcerative colitis (47). Taken together, these findings suggest that luminal micro-organisms exert an influence on the development of IBD in some patients, a conclusion further supported by the findings of our own present study. However, contradictory evidence was provided by a study that, using a semiquantitative bacteriologic analysis method, found that Bacteroides spp. was not significantly altered in patients with ulcerative colitis compared with that in healthy patients or those suffering from Crohn’s disease (48). The genus Bacteroides represents about one-third of the isolates from human fecal samples, and most healthy human adults harbor higher levels of B. vulgatus than of B. distasonis (49). Thus, more investigation is needed to determine whether B. vulgatus plays a pathogenic role in initiating IBD in clinical cases. However, it is interesting to hypothesize that B. vulgatus may become a pathogenic micro-organism for immunocompromised subjects (e.g., those suffering from an immunodeficiency such as a lack of αβT cells) in whom the immunologic homeostasis between the mucosal immune system and the intestinal microflora has been drastically disturbed. The immunopathologic stimulation provided by B. vulgatus would then be more likely to lead to the development of aberrant Th2-type CD4⁺, ββ T cells in immunocompromised TCR α^−/− than in TCR α^+/+ mice.

The TCR repertoire of CD4⁺, ββ T cells from ED- and RD-fed TCR α^−/− mice were found to be quite distinctive. Feeding with ED resulted in a high frequency of Vβ6, Vβ11, and Vβ14 positive cells, while feeding with RD led to a predominantly Vβ8 positive T cell population (Fig. 5). These findings further suggest that selected microflora (e.g., B. vulgatus) may cause disease development by inducing a specific population of CD4⁺ T cells with a selected Vβ usage (e.g., Vβ8). However, it is also possible that the disease in TCRα^−/− mice may not be caused by only one anaerobic micro-organism. Although our current data provide supportive evidence for the pathologic role of B. vulgatus in the development of Th2-type ββT cells in TCR α^−/− mice, one must realize that there is some anaerobic and/or aerobic species in commensal bacteria that may have ability for the induction of pathogenic Th cells in other immunocompromised conditions.

We thank Dr. Kimbery K. McGhee (University of Alabama, Tuscaloosa, AL) for editorial assistance.