Main

Allergic symptoms provoked by ingested foods include not only eczema or respiratory distress, but also diarrhea, especially in infants and young children. The allergic response is for simplicity distinguished into two types, the anaphylactic type and the cell-mediated type. The anaphylactic type can be seen within 30 min to 9 h after antigen ingestion, whereas the cell-mediated type occurs later, persists for a longer duration, and sometimes causes protracted diarrhea and dehydration. When intestinal biopsy specimens are examined, mucosal edema can be observed in the anaphylactic type, whereas patients suffering from the cell-mediated type display mucosal damage such as villous atrophy, crypt hyperplasia, and increased numbers of IEL(1). The precise mechanisms leading to these morphologic changes have not been defined.

CY is known to inhibit suppressor T and B cells(2) and to reduce total secretion of IgA(3), allowing orally presented food antigens to pass easily through the mucosal barrier and, in this manner, induces hypersensitive reactions against food antigens. Mowat and Ferguson(4) first reported this phenomenon in mice. When OVA was administered after treatment with CY, they observed the induction of mucosal damage and an elevation of the crypt cell reproduction rate(4). Further immunologic studies have not been reported.

In this study, IEL and LPL were isolated before and after the antigen challenge in this OVA-sensitive mouse model, and we investigated the subpopulations of lymphocytes and their proliferative response to OVA, to confirm that antigen-specific reactions cause mucosal damage in food-sensitive enteropathy.

METHODS

Animal treatment. Female BALB/c mice at 8 wk of age were used. Four groups, consisting of 10 mice each, were prepared. Group A was the nontreated control group. Groups B and C were treated with OVA or CY alone, respectively. Group D represented the OVA-sensitive mouse model prepared by treatment with both CY and OVA as previously described(4). CY was injected intraperitoneally (1 mg; 40 mg/kg) and 2 d later, 2 mg of OVA were given orally to sensitize the mice to this antigen. All of the mice were fed commercial pellets lacking OVA, or any other protein derived from egg, during the experiments. After 30 d, 0.1 mg of OVA dissolved in drinking water was given for 10 d to produce enteropathy (see Table 1). Mice were killed under anesthesia with ether and intestinal samples were removed for further examination.

Table 1 Protocol for producing a food-hypersensitive mouse model and its OVA sensitization
Full size table

Morphologic evaluations. More than 10 jejunum specimens, each about 1 cm long, were removed from every mouse and washed gently with PBS. Each sample was fixed in Clarke's solution and stained with Schiff's solution(4). More than 10 pairs of villous heights and crypt depths were measured in each sample to confirm the severity of mucosal damage. To estimate the extent of mucosal damage, the ratio between villous height and crypt depth was calculated for each sample. This ratio decreases according to the severity of mucosal damage.

The number of IEL per 100 epithelial cells was also counted for each specimen. Each sample was removed, fixed with 10% formalin, and stained with hematoxylin and eosin, and the average numbers of IEL per 100 epithelial cells were determined.

Isolation of IEL and LPL. IEL and LPL were isolated from the OVA-sensitive mice before and after the antigen challenge by modified procedures as described previously(5). Each intestinal sample was removed and flushed with Ca2+,Mg2+-free Hanks' balanced salt solution, and all Peyer's patches were removed. The intestines were then opened longitudinally and cut laterally into small pieces. Each segment was incubated in Ca2+,Mg2+-free Hanks' balanced salt solution containing 0.1 mM EDTA and stirred for 30 min twice to remove epithelial cells. Cell suspensions were filtered through nylon mesh and then centrifuged. The cell pellet, consisting of IEL and epithelial cells, was suspended in RPMI 1640 medium containing 10% FCS, 10 mM L-glutamine, 0.05 mM 2-mercaptoethanol, 100 U/mL penicillin, and 100 μg/mL streptomycin. The remaining fragments were then transferred to flasks containing RPMI 1640 with 90 U/mL of collagenase and stirred gently for 60 min in a 37°C water bath. Cell suspensions, containing LPL, were filtered through nylon mesh and then centrifuged. IEL and LPL were purified using a 45-70% discontinuous Percoll gradient. After centrifugation at 600 × g for 20 min at 4°C, cells at the interface were collected, washed, and resuspended in RPMI.

Cell viability was assessed by trypan blue dye exclusion. IEL and LPL were greater than 85% viable and the stained preparations of IEL and LPL almost exclusively displayed the characteristic morphology of lymphocytes with minimal epithelial cell contamination.

Lymphocyte subpopulations. Immunofluorescence staining was performed on IEL and LPL with monoclonal anti-mouse L3T4 (CD4) (rat IgG2b), LFA-1 (rat IgG2a) (Becton Dickinson, San Jose, CA), Ly-2(CD8) (rat IgG2a), Thy-1.2 (rat IgG2b), TCR-α/β(hamster IgG), and TCR-γ/δ (hamster IgG) (Pharmingen, San Diego, CA) antibodies. Positively stained cells were detected with FITC-conjugated anti-rat/hamster IgG. These antibodies were used at recommended dilutions according to the manufacturer's instructions. Two-color flow cytometric analysis were performed using a FACScan (Becton Dickinson), and dead cells were excluded from analysis on the basis of performed iodide dye exclusion. The percentage of positive cells and the mean intensity of fluorescence was calculated in comparison with a negative control consisting of cells stained with FITC-conjugated anti-rat/hamster IgG alone. Total cell numbers of each subpopulation were calculated as: Equation

Lymphocyte proliferation assays. To confirm the antigenspecific response of mucosal lymphocytes to the food antigen, a proliferation assay was performed as previously described(6, 7). IEL and LPL (2 × 105 cells; 100 μL/well) were cultured with irradiated(2000 rad) nontreated BALB/c mice spleen cells (1 × 105 cells; 100 μL/well), as antigen-presenting cells, and incubated at 37°C in 95% air-5% CO2 in 96-well flat bottom plates. Lymphocytes were stimulated with OVA at concentrations of 0, 0.4, 4, 40, and 400 μg/mL in RPMI medium for 72 h. Lymphocyte proliferation was determined by [3H]thymidine incorporation during the last 16 h by pulsing with 1 μCi of[3H]thymidine. Radioactivity was then counted using aβ-scintillation counter (Wallac 1450 MicroBeta Plus, Wallac Oy, Turku, Finland). Nontreated mice spleen cells were also present in the unstimulated controls as antigen-presenting cells. Each assay was done in triplicate. the stimulation index (S.I.) was calculated as: Equation

Results are expressed as the mean ± SD. Differences between means were determined by the Mann-Whitney U test.

RESULTS

Mucosal damage, including villous atrophy and crypt hyperplasia, was confirmed in the OVA-sensitive mouse and significantly differed from all control groups (p < 0.01) (Fig. 1). Figure 2 shows the number of IEL per 100 epithelial cells. The number of IEL was significantly increased in the OVA-sensitive mouse compared with all control groups (p < 0.01). These findings suggested that mucosal damage was introduced by orally administrated antigen in the OVA-sensitive mice.

Figure 1
figure 1

The ratio between villous height and crypt depth in jejunum samples was determined for the nontreated control, controls treated with OVA or CY alone, and the OVA-sensitive mice after the antigen challenge. Results are presented as mean ± SD of data from 10 animals(*p < 0.01).

Full size image
Figure 2
figure 2

The numbers of IEL per 100 epithelial cells were estimated in the nontreated control, controls treated with OVA or CY alone, and the OVA-sensitive mice after the antigen challenge. Results are presented as mean ± SD of data from 10 animals (*p < 0.01).

Full size image

To precisely confirm that lymphocyte infiltration occurs in food-sensitive enteropathy, IEL and LPL were isolated from the intestinal mucosa of OVA-sensitive mice before and after the antigen challenge and their total numbers were estimated, as shown in Table 2. The numbers of IEL and LPL were significantly increased in the OVA-sensitive mice after the antigen challenge (p < 0.01).

Table 2 Total numbers of IEL and LPL isolated before and after the antigen challenge in the OVA-sensitive mouse model
Full size table

To examine their characteristics, subpopulations of CD4+, CD8+, and Thy-1+ cells were analyzed by FACScan (Fig. 3). Before the antigen challenge, about 70% of the IEL were CD8+ cells, whereas the ratio of CD4+ to CD8+ cells among the LPL was almost equal, and about 45% of both IEL and LPL were Thy-1+ cells. Hence, after the antigen challenge, the number of CD8+ cells was significantly increased among the IEL (p < 0.01), and, in the case of LPL, the numbers of both CD4+ and CD8+ cells were increased (p < 0.01) to an equal extent. The numbers of Thy-1+ cells were also increased among both the IEL and LPL after the antigen challenge (p < 0.01).

Figure 3
figure 3

Fluorescence flow cytometry was performed to determine the numbers of CD4, CD8, Thy-1, TCR-α/β, and TCR-γ/δ positive cells among IEL and LPL of OVA-sensitive mice before (open column) and after (hatched column) the antigen challenge. Positively stained cells were analyzed by FACScan. Results are presented as mean ± SD of data from 10 animals (*p < 0.01).

Full size image

Subpopulations of TCR-α/β+ and TCR-γ/δ+ cells are also shown in Figure 3. Before the antigen challenge, 20% of the IEL were TCR-γ/δ+, whereas almost none of these cells was detected among the LPL. After the antigen challenge, only TCR-α/β+ cells were increased significantly (p < 0.01) in the case of both IEL and LPL.

Our findings concerning LFA-1 expression on IEL and LPL are shown in Figure 4. The curve shifted toward the right after antigen challenge, indicating that the expression of LFA-1 was enhanced in both IEL and LPL after the OVA challenge, and this shift suggested that these lymphocytes were activated.

Figure 4
figure 4

LFA-1 expression on IEL and LPL isolated from OVA-sensitive mice before (dashed line) and after (solid line) the antigen challenge was analyzed by fluorescence flow cytometry after reaction with anti-LFA-1 MAb and detection with FITC-conjugated goat anti-rat IgG.

Full size image

To determine whether IEL and LPL recognize and react to food antigen specifically, a proliferation assay was performed in the presence of serially diluted OVA (Fig. 5). The stimulation index was calculated and it was found that the proliferative response of both IEL and LPL isolated from the OVA-sensitive mice increased in a dose-dependent manner after the antigen challenge. The stimulation index of the IEL increased to a level five times greater than the unstimulated level after the antigen challenge (p < 0.01).

Figure 5
figure 5

Proliferation activity of IEL and LPL isolated from OVA-sensitive mice before (open circles) and after (closed circles) the antigen challenge was determined by [3H]thymidine incorporation. Each assay was done in triplicate. Stimulation index values are presented as mean ± SD of data from 10 animals (*p < 0.01).

Full size image

DISCUSSION

Mucosal damage, such as villous atrophy and crypt hyperplasia, was confirmed in the antigen challenged OVA-sensitive mouse model (Fig. 1), and the numbers of both IEL and LPL increased to levels twice as high as observed in the control groups (Fig. 2 and Table 2). Because those cells that migrated into the epithelial layer were mostly CD8+ cells (Fig. 3), it appears that the cytotoxic activities of these cells may directly damage epithelial cells and promote crypt cell reproduction as seen in damaged mucosa. In addition, the numbers of both CD4+ and CD8+ T cells were increased in the lamina propria (Fig. 3), suggesting that production of cytokines and eicosanoids by these lymphocytes may play an important role in mucosal damage. This process may be influenced not only by the interaction between CD4+ T cells and CD8+ T cells in the lamina propria, but also by IEL, because IEL have been found to regulate antibody and cytokine production(8, 9) and to suppress the proliferation of LPL(10). Recently, inflammatory bowel disease-like enteropathies have been observed in IL-2 or IL-10 deficient mice and also in TCR mutant mice(11–13). These studies demonstrated that lymphocytes and their products are critically involved in the development of mucosal damage in enteropathy.

Tshibassu and Desmet(14) reported an increase of CD8+ cells among IEL but a decrease of CD4+ cells among LPL in patients with food-sensitive enteropathy. They estimated the subpopulation of lymphocytes by immunohistochemical staining. This method does not indicate whether the number of CD4+ cells decreased in the whole intestine. Although the ratio of CD4+ to CD8+ cells is decreased, the absolute number of CD4+ cells might possibly increase in the lamina propria, and these cells could play an important role in immunologic reactions when large numbers of lymphocytes have infiltrated the intestine causing inflammation as seen in our mouse model (Fig. 3). Isolation of lymphocytes from the whole intestine is required to know the absolute number of CD4+ cells involved in enteropathy.

IEL mature extrathymically and show low expression of Thy-1(15). About 20 to 40% of IEL are TCR-γ/δ+ cells in normal mice(16). Hence, our findings of increased numbers of Thy-1+ cells, as well as TCR-α/β+ cells (Fig. 3), suggest that the IEL isolated from the intestine of antigen-challenged, OVA-sensitive mice are characteristically different from resident cells in the epithelial layer. In fact, they are very similar to those of peripheral blood. Kutlu et al.(17) also reported that the number of TCR-α/β+ cells, but not TCR-γ/δ+ cells, correlated with the grade of villous atrophy in coeliac patients on a long-term diet. We have concluded that some possibility exists that it is not resident but migrated lymphocytes, such as Thy-1+ and TCR-α/β+ cells, that may be directly involved in the pathogenesis of mucosal damage in food-sensitive enteropathy.

LFA-1 expression was enhanced on IEL and LPL after the antigen challenge in the OVA-sensitive mice (Fig. 4). LFA-1 is an adhesion molecule of which expression is increased when lymphocytes are activated(18). Increased expression of LFA-1 suggests that the IEL and LPL were activated by the orally administered antigen. The proliferative responses of both IEL and LPL were greatly augmented after the OVA challenge (Fig. 5), which also suggested that these lymphocytes were specifically activated by orally administered OVA. It is interesting that IEL show this food antigen-specific proliferative response, because IEL generally display low proliferative responses to various T cell stimuli, such as phytohemagglutinin, concanavalin A, or anti-CD3 antibody, as compared with peripheral blood lymphocytes(6).

There are several articles which suggest that cell-mediated immunologic reactions play a role in the pathogenesis of mucosal damage in enteropathy. Manluenda et al.(19) first reported that lymphocyte infiltration occurred after antigen challenge in cow's milk sensitive enteropathy. We have confirmed that the number of DR+CD4+ cells increase in the lamina propria after antigen challenge in the intestine of individuals with food-sensitive enteropathy, and their numbers return to the prechallenged level after treatment(20). Infiltration by these lymphocytes has been observed also in damaged intestinal mucosa in graft versus host reactions(21). Natural killer cell activity of IEL is thought to be one of the factors responsible for mucosal damage in graftversus host reactions. Mowat and Felstein(22) demonstrated that prevention of this mucosal damage can be achieved by injecting anti-asialo-GM1 antibody to deplete natural killer cells in graftversus host reactions. Intestinal damage has been observed also in cultured tissue from human small intestine after stimulation with pokeweed mitogen or anti-CD3 antibody which activates T cells in the intestine. In this instance, mucosal damage was prevented by adding cyclosporin A to the culture medium to inhibit T cell activation(23). However, to clarify the extent of participation of lymphocytes in the process of mucosal damage in food-sensitive enteropathy, it is necessary to prove that the migrated lymphocytes are activated by food antigen stimuli. In this study, we have demonstrated that specific food antigen actually stimulated the lymphocytes in the intestine and induced mucosal damage in the OVA-sensitive mouse model. Hence, further studies about cytokines and other chemical mediators produced by activated lymphocytes are necessary to explain how activated lymphocytes are involved and cause mucosal damage in food-sensitive enteropathy. We expect that such studies will clarify the mechanism of mucosal damage in food-sensitive enteropathy and provide a new strategy for clinical treatment.