Epithelial cells and lymphocytes, including γδ and αβ T cells, in the gastrointestinal tract epithelium represent a major host defense intranet that is incompletely understood. Cell-to-cell interactions between intraepithelial lymphocytes (IELs) and intestinal epithelial cells (IECs) comprise this intranet, and we have assessed the role of IECs in the regulation of γδ and αβ T cell responses. When highly purified CD3+ IEL T cells were stimulated via the TCR-CD3 complex, high proliferative responses and cytokine synthesis were induced. However, the addition of viable IECs or purified IEC membranes (mIEC) down-regulated T cell proliferative and cytokine responses. Further, the inhibitory effect of mIEC was not restored by antibodies to TGF-β, CD1d, E-cadherin, or MHC class I or II. This inhibitory effect was noted for both γδ and αβ T cell subsets from IELs, and mRNA levels were reduced for both Th1 (IL-2 and IFN-γ) and Th2 (IL-4 and IL-5) cytokines in γδ and αβ IELs. In contrast, a purified membrane fraction obtained from thymocytes did not inhibit IEL proliferative responses. Further, mIEC did not inhibit splenic αβ T cell proliferative responses. These findings show that cell-to-cell interactions between intraepithelial γδ and αβ T cells and IECs occur via cell surface molecules, suggesting an intranet to prevent potential inflammatory responses at the intestinal mucosal surface.
Mucosal surfaces such as those in the gastrointestinal (GI)3 tract provide a first line of defense from the plethora of potential pathogens and nonself Ags that perturb the host. This area is protected by the mucosal immune system, which includes both secretory IgA (S-IgA) and T cell-mediated immunity. It has been shown that orally encountered Ags are taken up by the M cells in a specialized epithelium overlying Peyer’s patches, which are sites for induction of protective mucosal immunity (1). In addition, a second major interface for absorption of soluble Ags from the lumen of the GI tract is a columnar intestinal epithelial cell layer that contains large numbers of T cells, commonly referred to as intraepithelial lymphocytes (IELs). The IELs have several unique features in their ontogeny and immunologic functions compared with their counterparts in lymphoid tissues. Most notably, the majority of IELs are T cells, with an approximately equal frequency of CD3+ IELs expressing either γδ or αβTCR heterodimeric chains (2, 3, 4, 5, 6).
Inasmuch as these IELs reside in the intestinal epithelium where major Ag trafficking from the digestive tissues into the host occurs, these IELs provide major effector functions for the mucosal immune system. For example, IELs exhibit both Ag-specific and polyclonal cytotoxic activities (7, 8). Further, both γδ and αβ T cells in the intestinal epithelium provide regulatory functions for IgA immune responses. It has been shown that γδ T cells can abrogate systemic unresponsiveness induced by orally administered Ags (9, 10). In addition, the loss of γδ T cells from the GI tract by specific δ gene deletion resulted in reduced IgA responses (11). The intestinal CD4+, CD8−, αβ T cells also provide typical helper functions in supporting Ag-specific IgA responses (10). These findings suggest that both γδ and αβ T cells in the intestinal epithelium can participate as regulatory T cells for the induction of IgA immune responses.
Our recent studies have shown that although whole IEL fractions respond poorly to mitogens or to TCR-CD3 complex-mediated stimulation signals, purified CD3+ IELs are capable of responding to activation signals provided via the TCR-CD3 complex (12). Thus, T cells from the intestinal epithelium respond to environmental Ags and mitogenic components that occur in the intestinal lumen. However, despite the fact that γδ and αβ T cells are continuously exposed to these stimuli, the majority of IEL T cells remain in the resting stage (13). These findings have also suggested possible cell-to-cell communication provided by the neighboring IECs that may arrest cell division of γδ and αβ T cells in the intestinal epithelium.
It is well established that intestinal epithelial cells (IECs) are capable of expressing MHC class II molecules and can act as effective APCs for the induction and activation of T cells in vitro (14, 15, 16). In this regard, several groups have suggested that IECs selectively activate T cells with suppressor function that are mediated by the CD1d molecule, which subsequently down-regulate immune responses (17, 18, 19). Further, it was shown that IECs can provide inhibitory signals to T cells via a PG cascade (20). In other studies, however, the IECs induced T cell proliferative responses in intestinal lamina propria without MHC class I or II restriction (21). Taken together, these results suggest that IECs can provide regulatory signals for T cells to maintain appropriate immunologic homeostasis in the GI tract; however, no studies have directly assessed cell-to-cell interactions between IECs and γδ and αβ IELs in the intestinal epithelium. This study has examined the effect of IECs on intestinal epithelial γδ and αβ T cells stimulated by signals through the TCR-CD3 complex.
Materials and Methods
C3H/HeN mice were obtained from Harlen Sprague-Dawley, Inc. (Indianapolis, IN). All mice were received at 5 to 6 wk of age, were maintained in horizontal lamina flow cabinets, and were provided with sterile food and water ad libitum. The mice used in this study were between 8 and 12 wk of age.
Isolation and purification of IELs, IECs, splenocytes, and thymocytes
Lymphocytes and epithelial cells were isolated by a modified protocol as described previously (13, 22). Briefly, dissected short segments of the small intestine from five mice were stirred at 37°C in prewarmed RPMI 1640 containing l-glutamine, penicillin, streptomycin, and gentamicin with 2% newborn calf serum for 10 min followed by vigorous shaking for 15 s. This process was performed twice, and the resulting supernatants were passed through a small cotton-glass wool column to remove cell debris and were then separated on a Percoll density gradient (Pharmacia Fine Chemicals, Pharmacia, Inc., Uppsala, Sweden). A discontinuous density gradient (25, 40, and 75%) was used. The cells that layered between the 40 and 75% fractions were collected as IELs, and the cells that layered between the 40 and 25% interface were collected as IECs, respectively. Highly purified IECs were then obtained by flow cytometry (FACStar Plus, Becton Dickinson Co., Sunnyvale, CA) according to cell size and granularity (Fig. 1,A). Following these procedures, cell yields of >96% viable cells that possessed alkaline phosphate activity were routinely obtained (Fig. 1,A). The purity of IECs was also assessed by flow cytometry staining with FITC-labeled anti-CD3 (145-2C11, PharMingen, San Diego, CA), FITC-anti-Ig (Southern Biotechnology Associates, Birmingham, AL), FITC-anti-Mac 1 (M3/84, PharMingen), and FITC-anti-H-2Kk (AF3-12.1, PharMingen; Fig. 1,A). Further, purified IECs were stained with FITC-anti-cytokeratin mAb (PCK-26, Sigma Chemical Co., St. Louis, MO) following cytospin, and the majority (>98%) of cells were cytokeratin positive (Fig. 1 B).
Splenocytes and thymocytes were prepared by a mechanical dissociation method using sterile stainless steel screens followed by density gradient purification (Lympholyte-M, Accurate Chemical and Scientific Co., Westbury, NY) (9). The viability of splenocytes and thymocytes was >98%. When the thymocytes were stained by FITC-anti-Thy1.2 (30-H12), >99% of cells were Thy 1.2+ cells. In some experiments, spleen cells were stained with FITC-anti-CD3, FITC-anti-γδTCR (GL3, PharMingen), biotinylated anti-αβTCR (H57-597), FITC-anti-CD4 (GK1.5), and biotin-anti-CD8 (53.6.72, PharMingen) followed by streptavidin-phycoerythrin (PharMingen) to obtain highly purified T cell subsets as described previously (9, 10, 11, 13). These cells were then sorted into specific T cell subsets according to the expression of γδTCR, αβTCR, CD4, and CD8 by flow cytometry. Our previous study showed that the highly purified T cell subsets from the intestinal epithelium are responsible for the activation signals provided via the TCR-CD3 complex (12).
IECs were resuspended in RPMI 1640 supplemented with 5 μg/ml of transferrin, 5 μg/ml of insulin, 5 ng/ml of sodium selenite, and 20 ng/ml of epidermal growth factor (23, 24). Cells were then cultured in rat collagen type I coated 24-well plates (Collaborative Biomedical Products, Becton Dickinson Labware, Bedford, MA) for 48 h at 37°C in a moist atmosphere of 10% CO2. During this incubation period, the cell viability was maintained at >60% for 24 h and approximately 40% after 48 h.
T cell proliferation
Purified T cells (2 × 106 cells/ml) were added to 96 wells precoated with 100 μl of anti-CD3 mAb (10 μg/ml; 145-2C11; PharMingen) for 48 h at 37°C in a moist atmosphere of 10% CO2. During the last 15 h of incubation, 0.5 μCi of tritiated thymidine ([3H]TdR)/well was added, the cells were harvested, and the amount of [3H]TdR incorporation was determined by scintillation counting.
Isolation of IEC membranes
A membrane preparation of IEC (mIEC) was obtained using the hypotonic lysis method as described previously (25). Briefly, cells were washed with cold PBS, and the cell pellet was mixed with lysis buffer containing mannitol (50 mM) and HEPES buffer (pH 7.4; 5 mM). Cells were then homogenized by aspiration through a 21-gauge needle and syringe. CaCl2 was added to a 10 mM final concentration followed by additional homogenization with a 25-gauge needle. The homogenized material was placed on ice for 10 min and subjected to 3,000 × g centrifugation for 15 min to remove insoluble material. The mIEC was sedimented by centrifugation at 100,000 × g for 12 min at 4°C. Purified mIEC were then resuspended in RPMI 1640 and were dialyzed against RPMI 1640. The amount of protein in mIEC was determined by bicinchoninic acid (BCA) protein assay reagent (Pierce Chemical Co., Rockford, IL).
Ab inhibition studies
The anti-CD1d (1B1), anti-H-2Kk (AF3-12.1), anti-H-2Dk (15-5-5), and anti-I-Ak (10-3.6) mAbs were purchased from PharMingen. Anti-E-cadherin was purchased from Zymed Laboratories, Inc. (South San Francisco, CA). Anti-human TGF-β that recognizes murine TGF-β (26) was purchased from Genzyme (Cambridge, MA). Ten micrograms of mIEC was preincubated with 100 μg of mAbs for 30 min at 4°C. The membranes were washed with PBS to remove unbound mAbs. The mAb-treated or untreated mIEC was then incubated with 2 × 106 cells/ml of intraepithelial T cells in the wells precoated with anti-CD3 mAb.
Cytokine-specific reverse transcriptase-PCR (RT-PCR)
The Th1 and Th2 cytokine-specific RT-PCR was performed as previously described (12, 13, 27). Briefly, RNA were extracted by the acid guanidinium thiocyanate phenol chloroform procedure, and RNA preparations from IEL T cells were subjected to the IFN-γ-, IL-2-, IL-4-, and IL-5-specific RT-PCR. Total RNA from T cells was added to a reaction mixture containing MgCl2 (Perkin-Elmer/Cetus, Norwalk, CT), PCR buffer II (Perkin-Elmer Cetus), dNTPs (Perkin-Elmer/Cetus), RNase inhibitor (Promega, Madison, WI), Moloney murine leukemia virus RT (Life Technologies, Grand Island, NY), and oligo(dT)16 (Perkin-Elmer). Samples were reverse transcribed, and the RT products were added to each tube with AmpliTaq DNA polymerase, 5′ primer, 3′ primer, MgCl2, and PCR buffer II (Perkin-Elmer/Cetus) and were amplified for 35 cycles. After RT-PCR, the products were stored at 4°C until analyzed.
Capillary electrophoresis analysis of RT-PCR products
The capillary electrophoresis with the laser-induced fluorescence detection system (LIF-P/ACE, Beckman Instruments, Fullerton, CA) was applied to RT-PCR products to quantitate the relative changes in the levels of cytokine-specific mRNA as described previously (28, 29, 30). Briefly, the analysis of cytokine-specific RT-PCR products was conducted using coated capillary tubes (Beckman Instruments) in Tris-borate EDTA-containing replaceable linear polyacrylamide gel and the fluorescent intercalator. RT-PCR products were run for 25 min at 200 V/cm. The level of cytokine-specific mRNA was normalized to the corresponding β-actin signal (30).
Results are expressed as the mean ± 1 SD of the mean, and statistical significance (p < 0.05) was analyzed by the Mann-Whitney U test.
Effects of IECs on IEL T cell proliferation
Although freshly isolated IELs respond poorly to stimulation signals provided by T cell mitogens or TCR-CD3 complex-specific mAbs (4, 31, 32), our previous studies have shown that highly purified CD3+ IEL T cells respond to stimulation via the TCR-CD3 complex (12). Since most procedures for isolation of IELs from the murine GI tract generally result in fractions containing approximately 10% enterocytes (5, 13), we considered the possibility that enterocytes actually induce hyporesponsiveness of intraepithelial T cells stimulated with anti-CD3, anti-γδ, or anti-αβ in vitro, effects that could mimic the in vivo situation. As expected, highly purified CD3+ T cells isolated from murine intestinal epithelium responded well to activation via solid phase anti-CD3, anti-γδ, or anti-αβ mAbs (Fig. 2). In contrast, when different ratios of IECs were added to CD3+ IEL T cell cultures (1:100, 1:50, 1:10, and 1:5), ratios of 1:5 and 1:10 of IEC:T cells inhibited proliferative responses (Fig. 2), suggesting that IECs dispatch inhibitory signals to neighboring γδ and αβ T cells in the intestinal epithelium.
Since freshly isolated IECs die rapidly in vitro, it was possible that denatured cellular proteins and enzymes from lysed IECs caused inhibition of T cell proliferative responses. To test this possibility, IECs were incubated for periods of up to 48 h, and supernatants were harvested from IEC cultures and added to IEL T cell cultures. IEC culture supernatants failed to inhibit intraepithelial T cell proliferation induced by anti-CD3 mAb (Fig. 3). On the contrary, when IEC culture supernatants were added to the CD3+ IEL cultures without other stimulation, low, but significant, proliferation was induced (data not shown). Taken together, these findings suggested that IECs did not spontaneously produce inhibitory cytokines that down-regulated IEL T cell responses.
The mIEC provide inhibitory signals to IEL T cells
Thus far, our findings suggested that direct cell-to-cell interactions between IECs and IEL T cells down-regulate T cell activation. To directly address this issue, mIEC from intestinal epithelium were added to anti-CD3 mAb-stimulated IEL T cell cultures (Fig. 4). The addition of mIEC to anti-CD3-stimulated IEL T cell cultures resulted in inhibition of T cell proliferation (Fig. 4). As little as 5 μg of mIEC inhibited intraepithelial T cell proliferative responses in vitro. Further, complete inhibition of IEL T cell proliferative responses was seen with 10 μg of mIEC, whereas membrane fractions from thymocytes did not affect the activation of intraepithelial T cells (Fig. 4). These results showed that the mIEC could inhibit neighboring intraepithelial CD3+ T cells in response to stimulation signals provided through the TCR-CD3 complex. We also examined the possibility that mIEC competitively bound to the TCR-CD3 complex with subsequent blocking of T cell activation. To this end, CD3+ T cells were preincubated with immobilized anti-CD3 mAb, and after 2 h, the mIEC was then added to T cell cultures. The mIEC inhibited IEL T cell proliferation (Fig. 5), indicating that the inhibitory effect was not due to blocking of the TCR-CD3 complex on intraepithelial T cells by mIEC.
Evidence against a role for TGF-β, CD1d, E-cadherin, and MHC class I or II inhibitory molecules
In the next series of studies, we tested whether the inhibitory signals from IECs to IEL T cells were provided via the known cell surface molecules, since several molecules and cytokines have been shown to be involved in cell-to-cell interactions between IECs and T cells. Thus, mIEC were incubated with anti-TGF-β anti-CD1d, anti-E-cadherin, anti-H-2Kk, anti-H-2Dk, or anti-I-Ak mAbs before cultivation with anti-CD3 stimulated IEL T cells. None of these Abs blocked inhibitory signals of mIEC for IEL T cells (Fig. 6). Further, the addition of these mAbs to individual wells containing anti-CD3-stimulated IEL T cells and mIEC did not block the inhibitory effect of mIEC (data not shown).
The mIEC inhibits intraepithelial γδ and αβ T cells, but not splenic T cells
Inasmuch as the study described above demonstrated that the mIEC inhibited the activation of CD3+ IEL T cells, it was important to examine whether the mIEC specifically inhibited a subset of T cells, since IEL T cells contain multiple phenotypes based on the expression of CD4, CD8, γδTCR, and αβTCR (5). Initially, CD4+ and CD8+ T cells isolated from IELs or spleen were stimulated via the TCR-CD3 complex in either the presence or the absence of mIEC. Both purified IEL CD4+ and CD8+ T cell subsets resulted in elevated T cell proliferation in response to activation signals provided by anti-CD3 mAb. The proliferative response in both IEL T cell subsets was inhibited by mIEC (Fig. 7). In contrast, mAb anti-CD3-activated CD4+ and CD8+ T cells isolated from spleen were not inhibited by mIEC (Fig. 7). These findings suggested that IECs can deliver inhibitory signals to both CD4+ and CD8+ T cells in the intestinal epithelium, but not to splenic T cells, via specific cell-to-cell interactions.
In the next series, IEL T cells were separated into γδ and αβ T cell subsets, since a unique feature of IEL CD3+ T cells is an enrichment (up to 50%) of CD3+ T cells that express γδ chains of TCR (2, 3, 4, 5, 6). The stimulation of both intraepithelial γδ and αβ T cells via the TCR-CD3 complex resulted in high proliferative responses (Fig. 8), as our previous studies have shown (12). The addition of mIEC markedly inhibited anti-CD3-stimulated γδ and αβ T cell proliferation; however, the proliferative responses of splenic αβ T cells remained normal, even in the presence of higher concentrations of mIEC (Fig. 8). These findings further suggest that the IECs can specifically down-regulate both intraepithelial γδ and αβ T cells, but not splenic T cells, via their specific cell-to-cell interactions.
The mIEC regulate IEL cytokine synthesis
Our previous studies have shown that freshly isolated IELs spontaneously produce IFN-γ and IL-5 in both γδ (Fig. 9, A and B) and αβ (Fig. 9, C and D) T cell subsets (5, 13, 33). Further, when IEL T cells were stimulated via the TCR-CD3 complex in vitro, these T cells also produced IL-2 and IL-4 (Fig. 9, A–D) (33). Therefore, we next examined the effect of mIEC on subsequent cytokine synthesis by anti-CD3 mAb-stimulated intraepithelial γδ and αβ T cells. Interestingly, mIEC inhibited de novo production of IL-2- and IL-4-specific mRNA; however, mIEC was less effective on existing mRNA levels of IL-5 and IFN-γ (Fig. 9, A–D). When 10 μg/well of the mIEC fraction was added to anti-CD3-stimulated intraepithelial γδ and αβ T cell cultures, mRNA levels for both Th1 (IL-2 and IFN-γ) and Th2 (IL-4 and IL-5) cytokines were reduced in these T cell cultures. It should be pointed out that mRNA for IL-5 and IFN-γ were still detected in anti-CD3 mAb-stimulated γδ and αβ T cells even in the presence of mIEC. On the other hand, IL-2- and IL-4-specific mRNA was not detected in these γδ and αβ T cell cultures treated with mIEC (Fig. 9, A–D). These results showed that mIEC down-regulated not only cell proliferation but also the expression of Th1 (IL-2) and Th2 (IL-4) cytokine-specific mRNA, which most likely serve as T cell growth factors within the epithelium.
In the present study, we have provided new evidence that IECs inhibit the activation of intraepithelial γδ and αβ T cells. Further, this inhibitory signal was delivered from IECs to T cells via cell surface molecules, since mIEC, but not culture supernatants, inhibited anti-CD3 mAb-stimulated IEL γδ and αβ T cells. To place these findings in perspective, other studies should also be discussed. For example, it has been shown that whole IEL preparations obtained from the murine small intestine respond poorly to T cell-specific activation signals (4, 12, 31, 32). However, it should be noted that most IEL preparations obtained by published procedures, which consist of mechanical dissociation followed by discontinuous Percoll gradient centrifugation, normally contain approximately 10% enterocytes (4, 5, 31). On the other hand, it was shown that highly purified CD3+ IEL T cells respond to stimulation signals transduced via the TCR-CD3 complex (12). Taken together, these findings suggested the possibility that even small numbers of IEC could transduce inhibitory signals to adjacent intraepithelial γδ and αβ T cells, which led to the present investigation.
To address the possible effects of IECs, our initial experiments were aimed at testing whether IEC could down-regulate the activation of intraepithelial CD3+ T cells. When IEC were added to anti-CD3 mAb-stimulated IEL CD3+ T cell cultures at ratios of 1:5 or 1:10, the reduction of CD3+ IEL proliferative responses to mAb specific for CD3, γδ, and αβTCR were observed (Fig. 2). In this regard, IECs have been shown to inhibit proliferative responses of rat lymph node cells stimulated by either Ag and/or mitogens, an effect attributable to PG secretion (20). An additional study has shown that rat IEC culture supernatants contain an inhibitory factor that associated with a protein with an approximate molecular mass of 32 kDa and down-regulated proliferative responses of lymphocytes to Con A, IL-2, or Ag (34). In human studies, it has been shown that pretreatment of peripheral T cells with intestinal mucosa supernatants results in the reduction of proliferative responses to protein kinase C activators (e.g., phorbol 12,13-dibutyrate and ionomycin) (35). In addition, several studies have reported that IECs were capable of producing inhibitory cytokines such as TGF-β (36, 37, 38, 39). In the present study, however, the addition of IEC culture supernatants did not inhibit T cell proliferative responses in mAb anti-CD3-stimulated IEL T cells (Fig. 3). Furthermore, treatment of mIEC with anti-TGF-β mAb did not block inhibitory signals from IECs (Fig. 6). These results indicated that IEC did not spontaneously produce inhibitory molecules such as TGF-β and/or the membrane form of TGF-β for CD3+ IEL proliferative responses in vitro. However, after appropriate stimulation of IECs, these cells may produce inhibitory cytokines. Instead, it was interesting to note that addition of IEC culture supernatants induced CD3+ IEL proliferative responses in the absence of any other stimulus (data not shown). In this regard, our separate study has shown that IECs express mRNA for IL-7, and this cytokine has been shown to induce the proliferation of CD3+ IEL T cells (22).
The mIEC obtained from intestinal epithelium showed inhibitory effects on intraepithelial T cell proliferative responses induced via the TCR-CD3 complex. Further, different T cell subsets of activated IEL T cells, including both γδ and αβ T cells as well as CD4+ and CD8+ T cell subsets, were inhibited by mIEC. However, this effect was not seen with splenic T cells. These findings suggest two related possibilities: 1) that specific cell-to-cell interactions occur between IECs and IELs to provide down-regulatory signals from the former to the latter cell population; and 2) that the inhibitory signal was provided by novel and undefined cell surface molecules, including adhesion and/or ligand molecules. Since we did not examine the effect of cell membrane molecules from other epithelial cells or cell lines, it will be important to further examine cell membrane fractions from epithelial cells of different tissues to elucidate the latter possibility. A recent study has provided supportive evidence that specific IEC and IEL cross-talk occurs in the intestinal epithelium, where the removal of δTCR gene (e.g., lack of γδ IEL) results in a reduction in the generation and differentiation of IECs (40). Together with the results of our current study, it is now clear that a intranet between IEL and IEC is an important cellular and molecular cross-talk for the maintenance of an appropriate immnologic homeostasis in the GI tract. An interesting aspect of a present study is the possibility of a novel and undefined cell surface molecule(s) expressed on IEC that may contribute to this mucosal intranet.
Others have reported that the inhibitory effect of IECs on the lymphocyte proliferative response is related to the activation of CD8+ T cells (18, 41). In the rat system, these IEC-activated CD8+ T cells from lymph nodes have been shown to suppress Ag-specific proliferative responses (41). In contrast, human IEC-stimulated peripheral blood CD8+ T cells were found to down-regulate proliferative responses in a polyclonal fashion (18). Thus, based on previous studies the inhibitory mechanisms provided by IECs could be explained by the induction of CD8+ T cells with suppressor functions. For cellular interactions between IECs and CD8+ T cells, it has been shown that CD1d, a class I-like molecule expressed on murine and human epithelial cells (42, 43), may play an important role, since Abs to CD1d were capable of inhibiting proliferation of CD8+ T cells in these cocultures (19). However, a recent study demonstrated that CD1d transfectants did not activate CD8-associated p56lck, a src-like tyrosine kinase that is necessary for activation of CD8+ T cells (44). Further, anti-CD1d mAb did not block the inhibitory effects of mIEC. In this regard, Abs to MHC class I or II also failed to block inhibitory signals from IECs (Fig. 6). αEβ7 integrin has been shown to be expressed on IEL T cells, but not on peripheral T lymphocytes (45), and mediates specific adhesive interactions between IELs and IECs (46, 47). Further, it has been reported that heterotypic adhesion between IECs and IELs is mediated by E-cadherin and αEβ7 integrin (48, 49). Thus, an alternative possibility would be that signals through E-cadherin may inhibit the activation of T cells expressing αEβ7 integrin. However, anti-E-cadherin-treated mIEC still inhibited IEL T cell proliferative responses. Taken together, these results suggest that novel and yet undefined molecule(s) expressed on IEC may provide inhibitory effects for intraepithelial T cells, and this inhibition is not due to CD8+ T cells. However, our results have not excluded the possibility that other known cell surface molecules may indeed contribute to the inhibition of intraepithelial T cell responses.
From an anatomical and histologic standpoint, the intestinal epithelium contains a large number of CD3+ T cells that are located adjacent to IECs (7, 8). Interestingly, approximately 80% of these CD3+ IELs are CD8+ (2, 3, 4, 5, 6, 7, 8). Therefore, it was natural to assume that IECs may provide activation signals to intraepithelial CD8+ T cells that lead to active suppression. However, our results showed that IECs dispatch inhibitory signals to neighboring intraepithelial CD4+ T cells, CD8+ T cells, γδ T cells, and αβ T cells via direct cell-to-cell contact. Possible explanations for these findings could be that compared with the peripheral T cells, which are anatomically and completely isolated from the direct and continuous exposure to environmental Ags, IELs are chronically exposed to microenvironmental Ags and mitogens from the intestinal epithelium. If IELs chronically responded to these Ags, it would create a hyper-T cell-reactive region in the intestinal epithelium that would lead to severe inflammation. Therefore, specific inhibitory signals generated by IECs to neighboring intraepithelial γδ and αβ T cells would be an important mechanism for the maintenance of immunohomeostasis in the intestinal environment. To support this view, our previous results have shown that although freshly isolated IELs contain some CD3+ T cells that spontaneously secrete Th1- and Th2-type cytokines, >90% of T cells were in the inactive cell cycle stage of G0 to G1 (13).
Freshly isolated intraepithelial γδ and αβ T cells expressed IL-5 and IFN-γ mRNA, but not IL-2 and IL-4. When these CD3+ IELs were stimulated by mAb anti-CD3, IL-2- and IL-4-specific mRNA were induced in addition to the enhanced IL-5- and IFN-γ-specific mRNA expression (Fig. 9, A-D), in accord with our earlier findings (5, 13, 33). It was important to note that mIEC inhibited the expression of these Th1 and Th2 cytokine-specific mRNA in both γδ and αβ T cells. These observations have further shown that mIEC provide inhibitory signals that down-regulate the expression of Th1 and Th2 cytokine-specific mRNA in intraepithelial γδ and αβ T cells. It is well known that Th1 and Th2 cells down-regulate each other via specific cytokines (e.g., IFN-γ and IL-10, respectively) (50). However, mIEC inhibited both Th1 and Th2 cytokines simultaneously. This interesting finding provided two possibilities. The first is that new or known inhibitory substances (e.g., TGF-β) that down-regulate both Th1 and Th2 responses are produced by IELs following the interaction with cell surface molecules on IECs. Alternatively, a unique and new signal through a surface molecule on IEC may directly inhibit both Th1 and Th2 cytokine activation pathways. These issues are currently under investigation by our group.
It is interesting to note that although the levels of elevated IFN-γ and IL-5 mRNA were decreased following treatment of activated γδ and αβ T cells with mIEC-treated IELs, these cytokine-specific mRNA were still detected. On the other hand, the expression of IL-2- and IL-4-specific mRNA was completely inhibited by the presence of mIEC. Thus, the profile of Th1 and Th2 cytokines seen in the mIEC-treated intraepithelial γδ and αβ T cells (e.g., IFN-γ and IL-5, but not IL-2 and IL-4) that had been activated via the TCR-CD3 complex is similar to that seen in the freshly isolated γδ and αβ IELs (Fig. 9, A-D). Thus, it appeared that expression of cytokine-specific mRNA by IELs is down-regulated by IECs in vivo. Since intraepithelial γδ and αβ T cells are continuously exposed to a wide variety of intestinal Ags and stimulatory molecules, down-regulatory signals provided by IECs would be necessary for the maintenance of an appropriate cytokine balance in the GI tract environment. Thus, these IECs down-regulate quiescent IEL T cells that are otherwise responsive to exogenous and pathogenic agents, including bacteria, virus, and parasites.
In summary, our study is the first to show that IECs directly inhibit proliferative responses as well as cytokine synthesis in neighboring, activated intraepithelial γδ and αβ T cells induced by stimulation signals through the TCR-CD3 complex. However, IECs did not produce inhibitory molecules spontaneously. Further, the inhibitory signals from IECs were not blocked by Abs to TGF-β, CD1d, E-cadherin, or MHC class I or II, suggesting that specific molecules expressed by IECs may provide an inhibitory signal(s) to intraepithelial γδ and αβ T cells. Our current effort is focused on the isolation and characterization of this unknown molecule for the understanding of the precise mechanisms for inhibitory signal transduction from IECs to IELs.
The authors thank Dr. Charles O. Elson for his critical review, and Ms. Sheila D. Turner for the preparation of this manuscript.
This work was supported by U.S. Public Health Service Grants AI35932, DE09837, AI35544, AI18958, DE08228, DE04217, DE12242, and DK44240 as well as grants from the Ministry of Education, Science, Sports, and Culture and the Ministry of Health and Welfare in Japan.
Abbreviations used in this paper: GI, gastrointestinal; S-IgA, secretory immunoglobulin A; IELs, intraepithelial lymphocytes; IECs, intestinal epithelial cells; mIEC, intestinal epithelial cell membranes; RT-PCR, reverse transcriptase-polymerase chain reaction.