Mice

BALB/c and SCID mice of the same background were purchased from Japan Clea Co. (Tokyo, Japan). STAT4^−/− mice of the BALB/c background were obtained from the Jackson Laboratory (Bar Harbor, Maine, USA). All mice were housed in the experimental animal facility at the Research Institute for Microbial Diseases, Osaka University, and fed sterilised food (certified diet MF; Oriental Yeast Co., Osaka, Japan) and tap water ad libitum. All mice were aged 6–8 weeks at the beginning of the experiments.

Immunisation and adoptive transfer

BALB/c mice were primed by the subcutaneous route with 1 mg of OVA (fraction V; Sigma Chemical Co., St Louis, Missouri, USA) or 100 μg of keyhole limpet haemocyanin (KLH) (Sigma) in 100 μl of CFA (Difco Laboratories, Detroit, Michigan, USA), respectively. One week after systemic challenge, the spleen was removed aseptically and a single cell suspension was prepared by mechanical dissociation methods, as described previously.¹⁶ FACS separated CD4⁺ CD45RB⁺ T cells or CD4⁺ CD45RB^low T cells (4–5×10⁵) were resuspended in 200 μl of phosphate buffered saline (PBS) and adoptively transferred to SCID mice by tail vein injection. One week after the reconstitution, the recipient SCID mice were repeatedly challenged with 50 mg of OVA or 5 mg of KLH by gastric intubation three times a week. Body weight of naive and reconstituted SCID mice was measured every week. After 8–10 oral administrations, the spleen, and small and large intestines were removed aseptically.

Histological evaluation

For the histopathological study, the small and large intestines from recipient SCID mice were fixed in 4% paraformaldehyde and embedded in paraffin. Sections (5 μm) were stained with haematoxylin and eosin for evaluation of general histopathology changes and stained with periodic acid-Schiff for identification of goblet cells. The degree of inflammation in the microscopic sections of the colon was graded semiquantitatively in a blinded manner using a scoring system of 0 to 4 (0, no inflammation; 1, very mild inflammation; 2, mild inflammation; 3, moderate inflammation; 4, severe inflammation, with infiltration of mononuclear cells, crypt elongation, crypt abscesses, loss of goblet cells, and thickening of the colon wall).¹⁷

Neutralising monoclonal antibody treatment

SCID mice were injected intraperitoneally with 100 μg of antimouse interferon γ (IFN-γ) monoclonal antibody (mAb) (XMG1.2; Pharmingen, San Diego, California, USA), antimouse TNF-α mAb (MP6-XT3; Pharmingen), or isotype matched rat IgG2b mAb (R35-38; Pharmingen) once a week for the entire period of the experiments. The mAb treatment was started one week before the adoptive transfer.

Isolation of mononuclear cells

Intraepithelial lymphocytes and lamina propria lymphocytes from the small and large intestine were isolated as described previously.^{14, 18} In brief, intestinal tissues were digested using RPMI medium containing 1 mM EDTA and a collagenase (type IV 0.5 mg/ml of RPMI 1640; Sigma) after removal of Peyer's patches and colonic patches in the 37°C incubator. The single cell suspensions were pooled, washed, and placed on a discontinuous 40% and 70% Percoll gradient (Pharmacia, Uppsala, Sweden). After centrifugation for 20 minutes at 600 g, the cells were collected from the interface.

FACS analysis and cell sorting

Isolated mononuclear cells from the spleen, small intestine, and large intestine were preincubated with an Fc blocking mAb (2.4G2; Pharmingen) for 15 minutes on ice. The following Abs were purchased from Pharmingen and used in this study: FITC conjugated antimouse CD4 (L3T4), CD2 (RM2-5), CD11a (CL891F) mAbs, and PE conjugated antimouse CD48 (BCM1), CD69 (H1.2F3), CD44 (1M7), CD54 (3E2), and CD45RB (16A) mAbs. Two colour analysis was performed in flow cytometry analysis using a FACS Calibur (Becton Dickinson, San Jose, California, USA). In some experiments, the flow cytometry sorting separation was performed by FACS Vantage (Becton Dickinson) using FITC conjugated anti-CD4 and PE conjugated anti-CD45RB mAbs.

In vitro OVA specific proliferative responses

Mononuclear cells isolated from the spleen and large intestine of the reconstituted SCID recipients were suspended in RPMI 1640 medium (Sigma) containing 10% heat inactivated fetal calf serum, HEPES buffer, l-glutamine, penicillin, and streptomycin. The cells were cultured in the presence of 1 mg/ml of OVA for four days, as described previously.¹⁶ To measure antigen specific proliferation, 0.5 μCi of [³H]thymidine (ICN, Costa Mesa, California, USA) was added for the final 18 hour incubation and the amount of [³H]thymidine incorporation was determined by scintillation counting.

Intracellular cytokine assay

For intracellular cytokine analysis, mononuclear cells isolated from the spleen and the small and large intestines were cultured with complete RPMI medium containing 10% fetal bovine serum, soluble anti-CD28 mAb (37.51, 2 μg/ml), and recombinant interleukin (IL)-2 in 24 well flat bottomed plates coated with anti-CD3ε mAb (145–2C11, 10 μg/ml) for 16 hours.^{11, 19} GolgiStop (2 μM/ml; Pharmingen) was added during the final four hours of incubation, and cytoplasmic staining was then performed using Cytofix/Cytoperm Kits (Pharmingen). FITC conjugated antimouse CD45RB (16A) and PE conjugated antimouse IFN-γ (XMG1.2), TNF-α (MP6-XT22), IL-4 (BUD4-1D11), IL-10 (JES5-16E3), and antirat IgG1 (R3-34) mAbs for isotype control were used. Labelled cells were analysed by flow cytometry analysis using FACS Calibur (Becton Dickinson). Unless otherwise indicated, mAbs were purchased from Pharmingen.

Quantitative RT-PCR method

A quantitative reverse transcription-polymerase chain reaction (RT-PCR) was employed for the assessment of cytokine specific mRNA expression by different subsets of CD4⁺ T cells according to a method described previously.¹¹ Total RNA was extracted using TRIzol reagent (Life Technologies, Gaithesburg, Maryland, USA), and 5 μg/ml of extracted RNA was subjected to RT reaction using Superscript II reverse transcriptase (Life Technologies). The cDNA from 10 ng of RNA was used for each cytokine specific PCR (for example, IFN-γ, TNF-α, IL-4, and IL-10). Rapid cycle DNA amplification was performed by a LightCycler (Boehringer Mannheim GmbH, Mannheim, Germany) with the double strand specific dye SYBER Green I.²⁰ The conditions of the PCR cycle used were as follows: initial denaturation at 95°C for two minutes, followed by 45 cycles of denaturation at 95°C for 0 seconds and combined annealing-extension at 55°C for five seconds and 72°C for 10 seconds. Cycle to cycle fluorescence emission readings were plotted on the computer screen for continuous monitoring of PCR product using LightCycler software.²⁰

Data analysis

Data are expressed as mean (SEM) and were evaluated by the Mann-Whitney U test for unpaired samples using a Statview II statistical program (SAS Inc., Cary, North Carolina, USA) designed for the Macintosh computer. p values less than 0.05 were assumed to be statistically significant.

Systemically primed splenic CD4⁺ T cells induce weight loss and provoke inflammation in the large intestine of the SCID mice after specific antigen feeding

To investigate whether antigen specific CD4⁺ T cells derived from the systemic immune compartment play a pathogenic role in the large intestine, splenic CD4⁺ T cells taken from mice systemically primed with OVA containing populations of both high and low expression levels of CD45RB molecule (CD45RB^high and CD45RB^low) were adoptively transferred to SCID mice which were then fed OVA for four weeks and sacrificed. We found that the mice had lost weight, had loose stools, and had gross colitis (fig 1A, B⇓). Histological evaluation of the spleens and small and large intestines revealed that lesions were restricted to the large intestine with the most severe lesions being located in the proximal colon. Prominent inflammatory cell infiltration, crypt elongation, and crypt abscesses were present (fig 1D⇓), and the number of goblet cells was decreased (fig 1F⇓) compared with the control group (fig 1C, E⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Induction of colitis in SCID mice by reconstitution with ovalbumin (OVA) primed splenic CD4⁺ CD45RB⁺ T cells. BALB/c mice were immunised by subcutaneous injection with OVA in complete Freund adjuvant. One week after systemic challenge, FACS separated CD4⁺ CD45RB⁺ T cells obtained from the spleens were adoptively transferred to SCID mice of the same background mice (5×10⁵ cells per mouse). Control SCID mice received splenic T cells from non-primed BALB/c mice. Beginning one week after the reconstitution, the recipient SCID mice were repeatedly challenged with 50 mg of OVA by gastric intubation. After four weeks of oral challenge, tissue sections were taken from the proximal, middle, and distal regions of the large intestine and stained with haematoxylin and eosin (C, D) or periodic acid Schiff staining (E, F). (A) Gross view of the large intestine of control SCID mice without colitis (upper) and reconstituted mouse with colitis (lower). (B) Change in body weight in control (open circle) and reconstituted (filled circle) SCID mice. (C, E) Histology of a control SCID mouse without colitis. (D, F) Histology of the proximal large intestine of a reconstituted SCID mouse with a severe pathological lesion. Periodic acid-Schiff staining shows decreased numbers of goblet cells in the proximal region of the large intestine from a diseased mouse (F) compared with similar tissue from a control mouse (E). Original magnification ×100–200.

The CD4⁺ CD45RB^low population was selectively expanded in the large but not the small intestine

To define the phenotype of mononuclear cells from the spleen, small intestine, and large intestine of the reconstituted SCID mice, flow cytometry analysis was performed using mAbs specific for various surface-membrane molecules (fig 2⇓). Higher numbers of CD4⁺ T cells were found in the large intestine than in the spleen and small intestine of the recipient SCID mice following repeated oral administration of OVA (for example, the frequency of CD4⁺ cell subset: spleen 16%, small intestine 25%, and large intestine 73%). In addition, the large intestinal CD4⁺ T cells contained high numbers of CD45RB^low CD44^high CD69^high T cells which are well known phenotypes of activated/memory type T cells (fig 2A⇓). Also, mononuclear cells isolated from the large intestine expressed high levels of LFA-1α, intercellular adhesion molecule 1 (ICAM-1), and LFA-2 compared with those of the small intestine (fig 2B⇓). Neither large intestinal nor small intestinal mononuclear cells expressed L-selectin, α₄β₇, or α_IELβ₇ (data not shown). The expression pattern of adhesion molecules in mononuclear cells isolated from the spleen was similar to that in cells from the large intestine (fig 2B⇓). As it is well established that increased expression of LFA-1, LFA-2, LFA-3, and ICAM-1 is a marker of activated/memory type cells,^{21, 22} our data demonstrate that activated/memory type CD4⁺ T cells preferentially migrated to and were expanded in the large intestine.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Expression of an activated/memory phenotype (A) and adhesion molecules (B) by the CD4⁺ T cells isolated from spleen, and small and large intestines of the reconstituted SCID mice after oral challenge with ovalbumin. The small boxes in the dot plot of the FACS data indicate the phenotype of the activated/memory type population (for example, CD45RB^low, CD44^high, and CD69^high). These data are representative of three independent experiments containing 3–5 mice per group. Detailed immunisation methods are described in the legend to fig 1⇑. ICAM-1, intercellular adhesion molecule 1; LFA, lymphocyte function associated antigen.

CD45RB^low subsets of the large intestine produced predominant Th1 type cytokines

To determine the profile of cytokine synthesis of CD4⁺ CD45RB⁺ T cells isolated from the spleen, small intestine, and large intestine of the recipient SCID mice, we performed intracellular cytokine staining and cytokine specific RT-PCR assays. Analysis of cytokine production at the single cell level revealed that a significant number of large intestinal CD45RB^low subsets expressed IFN-γ compared with those from the small intestine (fig 3⇓). Furthermore, large numbers of TNF-α secreting cells were also detected in the CD45RB^low population of the large intestine and spleen (fig 3⇓). In contrast, no IL-4 or IL-10 secreting cells were detected in the spleen, small intestine, or large intestine of recipient SCID mice (fig 3⇓). When the cytokine specific RT-PCR assays were performed, colonic CD4⁺ T cells were found to possess high levels of mRNA for IFN-γ and TNF-α; however, no mRNA for IL-4 and IL-10 was detected (table 1⇓). Taken together, these findings suggest that the development of colitis in reconstituted SCID mice is caused by the spleen derived colonic CD4⁺ CD45RB^low T cells producing IFN-γ and TNF-α in the colon.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Analysis of cytokine production by ovalbumin primed large intestinal CD4⁺ and CD45RB^low T cells. Details are described in the methods section. The main quadrant line represents staining with isotype control monoclonal antibody. These data are representative of two independent experiments containing 3–5 mice per group. TNF-α, tumour necrosis factor α; IFN-γ, interferon γ; IL, interleukin.

Administration of neutralising Ab to TNF-α but not to IFN-γ protected the development of colitis in SCID mice reconstituted with systemically primed splenic CD4⁺ CD45RB^low T cells

To directly examine the role of Th1 type cytokines in the development of colitis, reconstituted SCID mice were treated with neutralising Abs to IFN-γ or TNF-α. Neutralising Abs to TNF-α but not to IFN-γ completely prevented clinical and histological evidence of colitis. Histologically, the proximal large intestine of anti-IFN-γ Ab treated SCID mice had prominent crypt elongation, crypt distortion, reduced numbers of goblet cells, and inflammatory cell infiltration (fig 4B⇓, F) whereas these signs of disease were absent in mice treated with anti-TNF-α Ab (fig 4C⇓). Also, mice treated with anti-IFN-γ lost weight but mice treated with anti-TNF-α did not (fig 4E⇓). As expected, colitis was normally developed in the mock Ab treated recipient SCID mice (fig 4A, F⇓). As neutralisation with anti-IFN-γ Ab did not protect against the development of colitis, we next tested whether the STAT4 signalling cascade is involved in the development of colitis. Donor CD4⁺ T cells were obtained from the spleens of STAT4^−/− mice following subcutaneous priming with OVA in CFA. Reconstitution of SCID mice with the splenic CD4⁺ T cells resulted in colitis (fig 4D, F⇓), and the level of disease severity was similar to that seen in the SCID mice reconstituted with CD4⁺ T cells from the primed wild type mice (fig 1D⇑). Thus the lack of IFN-γ inducing pathway did not prevent the development of colitis by CD4⁺ CD45RB^low T cells. Taken together, our data suggest that TNF-α rather than IFN-γ is crucial for the development of the colitis studied here.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Effects of administering mock antibody (Ab) (A) or neutralising Abs to interferon γ (IFN-γ) (B) or tumour necrosis factor α (TNF-α) (C) to SCID mice reconstituted with splenic CD4⁺ CD45RB⁺ T cells of systemically primed BALB/c mice. Histological sections of the proximal large intestine are shown. (D) Large intestine of SCID mice reconstituted with splenic CD4⁺ CD45RB⁺ T cells of systemically primed STAT4^−/− mice. (E) Percentage starting body weight in each group corresponding to those of (A–D). (F) Histological score graded semiquantitatively from 0 to 4. Details of histological score and immunisation schedules are described in the methods and in the legend to fig 1⇑, respectively. Original magnification ×100.

Antigen specific reactions accelerate the development of colitis

To clarify if antigen specific responses are involved in the development of colitis, SCID mice were reconstituted with splenic CD4⁺ CD45RB⁺ T cells isolated from donor mice systemically primed with OVA or KLH in CFA. After four weeks of oral administration with the relevant or irrelevant protein, the severity of colitis was compared. Severe colitis developed in SCID mice reconstituted with OVA primed T cells following oral challenge with OVA (fig 5A⇓). In contrast, mild colitis developed in the reconstituted SCID mice following oral challenge with irrelevant protein or PBS alone; their histological scores were lower than those of SCID mice orally challenged with relevant protein (fig 5A⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Role of ovalbumin (OVA) specific CD4⁺ T cell responses in the development of colitis. (A) Histopathological score of the proximal large intestine of the reconstituted SCID mice. Donor splenic CD4⁺ CD45RB⁺ T cells were isolated from BALB/c mice systemically primed with OVA or keyhole limpet haemocyanin (KLH) in complete freund adjuvant (CFA) and then adoptively transferred to each group of SCID recipient via tail vein injection. After one week of the reconstitution, SCID recipients were fed relevant (OVA) or irrelevant (KLH) proteins. Details of histological scoring are described in the methods section. (B) Assessment of OVA specific T cell proliferation by mononuclear cells isolated from SCID mice reconstituted with naive splenic CD4⁺ CD45RB⁺ T cells or OVA primed splenic CD4⁺ CD45RB⁺ T cells. Stimulation index is expressed as [³H]thymidine incorporation by cells incubated with OVA divided by [³H]thymidine incorporation by cells in controls, incubated without OVA. Levels of [³H]thymidine incorporation of control wells were 500–1000 cpm. *p<0.05 versus group of SCID mice reconstituted with naive CD4⁺ CD45RB⁺ T cells.

To further investigate involvement of antigen specific T cell responses in the development of colitis in SCID recipients, mononuclear cells were isolated from the spleens and large intestines of SCID mice, and OVA specific T cell proliferative responses were then examined. High levels of OVA specific T cell proliferation were seen in the spleen and large intestine of the SCID recipient reconstituted with the OVA primed CD4⁺ CD45RB⁺ T cells following repeated oral administration of relevant antigen—that is, OVA (fig 5B⇑). On the other hand, mice reconstituted with naive T cells had low levels of T cell proliferation. These results further suggest that antigen specific reactions could accelerate the development of colitis.

Reconstitution of OVA primed CD45RB^low CD4⁺ T cells leads to colitis in SCID recipients

To directly ascertain a pathogenic role for CD4⁺ CD45RB^low T cells in the development of colitis, SCID mice were reconstituted with splenic CD4⁺ CD45RB^low T cells isolated from OVA primed mice. Following repeated oral administration of OVA, histological score of the large intestine and intracellular cytokine expression by large intestinal mononuclear cells were examined. The SCID mice adoptively transferred with OVA primed splenic CD4⁺ CD45RB^low T cells developed severe colitis characterised by infiltration of inflammatory cells, crypt elongation, loss of goblet cells, and thickened large intestinal wall (fig 6⇓). Most interestingly, increased numbers of TNF-α secreting cells were detected in SCID recipients reconstituted with OVA primed splenic CD4⁺ CD45RB^low T cells compared with those of SCID mice adoptively transferred with non-primed CD4⁺ CD45RB^low T cells (fig 6⇓). In contrast, no significant difference was seen in IFN-γ secreting cells between the two groups (fig 6⇓). These results further implicate TNF-α in the development of colitis in this model. In summary, TNF-α producing antigen specific activated/memory CD4⁺ CD45RB^low T cells play a critical role in the development of colitis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Pathological role of ovalbumin (OVA) primed CD4⁺ CD45RB^low T cells in the development of colitis. Histopathological analysis of the proximal large intestine was performed by haematoxylin and eosin staining, and Th1 type cytokine expression by the large intestinal mononuclear cells was determined by intracellular cytokine staining. SCID mice were reconstituted with unprimed or OVA primed splenic CD4⁺ CD45RB⁺ T cells. Beginning one week after reconstitution, SCID mice were repeatedly challenged by gastric intubation with 50 mg of OVA, Original magnification ×100. TNF-α, tumour necrosis factor α; IFN-γ, interferon γ; CFA, complete Freund adjuvant.