It is well known that immune responses in the intestine remain in a state of controlled inflammation, suggesting that not only active suppression by regulatory T cells plays an important role in the normal intestinal homeostasis, but also its dysregulation leads to the development of inflammatory bowel disease. In this study, we demonstrate that the CD4+CD25bright T cells reside in the human intestinal lamina propria (LP) and functionally retain regulatory activities. All human LP CD4+ T cells regardless of CD25 expression constitutively expressed CTLA-4, glucocorticoid-induced TNFR family-related protein, and Foxp3 and proliferate poorly. Although LP CD4+CD25− T cells showed an activated and anergic/memory phenotype, they did not retain regulatory activity. In LP CD4+CD25+ T cells, however, cells expressing CD25 at high levels (CD4+CD25bright) suppressed the proliferation and various cytokine productions of CD4+CD25− T cells. LP CD4+CD25bright T cells by themselves produced fewer amounts of IL-2, IFN-γ, and IL-10. Interestingly, LP CD4+CD25bright T cells with regulatory T activity were significantly increased in patients with active inflammatory bowel disease. These results suggest that CD4+CD25bright T cells found in the normal and inflamed intestinal mucosa selectively inhibit the host immune response and therefore may contribute to the intestinal immune homeostasis.
The gastrointestinal tract is home to the largest number of leukocytes in the body as well as being the site where these cells encounter abundant exogenous stimuli. Despite this potential immune stimulus, it is well known that immune responses in the intestine remain in a state of controlled inflammation (1). Regulation of the immune response here is a balance between the need to mount protective immunity toward pathogens while not activating damaging inflammatory responses to the plethora of harmless Ags present, including those derived from resident bacteria (2, 3, 4). To maintain the intestinal homeostasis, including immunological tolerance, functionally distinct subsets have been clearly defined in T cells (5, 6). Among these subsets, the regulatory T (TR)3 cell subset down-regulates immune responses for both foreign and self-Ags and effectively participates in the suppression of autoimmune disorders (7, 8, 9). The importance of an intact immune system for the intestinal homeostasis is revealed by the fact that a number of immune manipulations, including deletion of cytokine genes and alterations in TR subsets, lead to the development of an animal model of inflammatory bowel disease (IBD) (10, 11, 12). Evidence emerging from these studies suggests that pathogenic responses in the intestine are derived by resident bacteria and are controlled by a functionally specialized population of TR cells in the GALT (13).
A variety of TR cells that display regulatory function in vitro or in vivo have been described. These can be subdivided into different subsets based on the expression of cell surface markers, production of cytokines, and mechanisms of action. Recent studies focused on CD25 as the best marker for CD4+ TR cells in mice and humans (14, 15, 16). CD4+CD25+ T cells, which constitute ∼10% of peripheral murine and human CD4+ T cells, express CTLA-4 (17, 18, 19, 20), glucocorticoid-induced TNFR family-related protein (GITR) (21, 22), and the forkhead/winged helix transcription factor Foxp3 (23, 24, 25). Although autoimmune diseases, including IBD, can be induced by reconstituting immunodeficient mice with peripheral CD4+ T cells, which have been depleted of CD4+CD25+ (26) or CD4+GITR+ T cells (27), not only the existence of human intestinal CD4+CD25+ TR cells in normal or diseased condition but also their role in the pathogenesis of IBD are largely unknown. Conversely, several findings, for example. the evidence of the accumulated CD25+, CD69+, and CD71+ (transferrin receptor) cells, indicate an increased state of activation of the T cell system in IBD (28). Although intestinal lamina propria (LP) T cells are already preactivated, the activation is further increased in both the circulation and mucosa from IBD patients as determined by several activation markers as compared with those from normal individuals (29). Furthermore, this activated phenotype is increased in involved compared with uninvolved areas or control tissue (29), being so far believed that these activated CD4+CD25+ cells should be pathogenic for the development of IBD. These conflicting findings prompted us to determine whether or not the LP CD4+CD25+ T cells in IBD patients were pathogenic (effector) or protective (TR) in this study.
Furthermore, there is much other evidence showing that 1) the peripheral CD4+CD25− T cell population also possesses some regulatory activity (26, 30, 31, 32, 33, 34), 2) anergic/memory T cells have regulatory properties (35, 36, 37, 38), and 3) LP CD4+ T lymphocytes, regardless of CD25 expression, are generally memory T cells, and hyporesponsive to TCR-mediated proliferative signals (39, 40), indicating these cells could be highly differentiated effector or effector memory T cells (41, 42) with a raised threshold of activation that prevents immune responses to harmless intraluminal Ags.
In the present study, we investigate the existence and the role of the human LP CD4+CD25bright as well as CD4+CD25+ and CD4+CD25− T cells. We demonstrate here that human LP CD4+ T cells from the normal intestine, regardless of CD25 expression, express other regulatory markers, CTLA-4, GITR, and Foxp3, and show anergic and memory features, but only CD4+CD25bright LP T cells retain TR function, which can mediate potent suppression of autologous T cell proliferation and various cytokine productions, such as IFN-γ and IL-2. Furthermore, we show here that these TR cells are significantly increased in patients with active IBD.
Materials and Methods
Patients and samples
Normal mucosal samples were obtained from macroscopically and microscopically unaffected areas of 54 colonic specimens from patients with colon cancer who underwent surgery and were obtained from the surgical resected samples of intestinal inflamed mucosa of 24 patients with colonic type Crohn’s disease (CD) and 25 patients with ulcerative colitis (UC). The mucosa was prepared immediately after stripping away the underlying submucosa by blunt dissection. All experiments were approved by the Committee on Human Subjects of both Tokyo Medical and Dental University Hospital (Tokyo, Japan) and Yokohama City Hospital (Yokohama, Japan). Informed consent was obtained from all patients before the study. Disease activity in each patient with CD was analyzed according to the Crohn’s Disease Activity Index and endoscopic and histopathological data. The disease had been present from 6 mo to 15 years. When the experimental study was performed in patients with CD, 4 patients were taking only steroids, 4 were taking both steroids and sulfasalazine, 3 were receiving both azathiopurine and sulfasalazine, 11 were taking only sulfasalazine, and 2 had been undergoing nonspecific therapy for the previous 3 mo. In the UC group, disease activity was defined by the Truelove-Witts criteria and endoscopic (Matts grade) and histopathological data. When the experimental study was performed, 1 patient was receiving only steroids, 14 were receiving both steroids and sulfasalazine, 6 were taking only sulfasalazine, 3 were receiving azathiopurine, steroids and sulfasalazine, and 1 patient had been undergoing nonspecific treatment for at least 3 mo.
Culture media, reagents, and Abs
In all in vitro assays, cells were cultured in RPMI 1640 medium supplemented with 2 mM l-glutamine, 100 IU/ml penicillin/100 μg/ml streptomycin (Sigma-Aldrich, St. Louis, MO), 2 × 10−5
Purified recombinant human IL-2 (rhIL-2; kindly provided by Shionogi Pharmaceutical, Osaka, Japan), purified anti-human CD3 mAb (HIT3a; BD Pharmingen), Con A (Sigma-Aldrich), PHA (Sigma-Aldrich), and mitomycin C (MMC, Sigma-Aldrich) were used for in vitro coculture assay.
Isolation of LP mononuclear cells (LPMCs) from intestinal mucosa
LPMCs were isolated from surgically resected intestinal specimens using enzymatic techniques as previously described (43). Briefly, the dissected mucosa was incubated in calcium and magnesium-free HBSS containing 2.5% FBS and 1 mM DTT (Sigma-Aldrich) to remove mucus. The mucosa was then incubated in medium containing 0.75 mM EDTA (Sigma-Aldrich) for 60 min at 37°C. During this treatment, intraepithelial lymphocytes and epithelial cells were released from the tissue and tissues containing LPMCs were collected and incubated in medium containing 0.02% collagenase (Worthington Biochemical, Freehold, NJ). The fraction was pelleted twice through a 40% isotonic percoll solution and the cells were then centrifuged over a Ficoll-Hypaque density gradient (40% and 60%). The purity of resulting LPMCs was analyzed by flow cytometry.
The CD4+ T cells from LPMCs and PBMCs were purified by positive selection with the CD4+ MultiSort kit (Miltenyi Biotec, Auburn, CA) (27). The MMC-treated CD4− fraction was used as APCs in the following experiments. The CD4+CD25−, CD4+CD25+, and CD4+CD25bright populations were isolated from LPMCs and PBMCs by sorting using a FACSVantage (BD Biosciences, San Jose, CA) in accordance with the method by Baecher-Allan et al. (44), who demonstrated that the regulatory CD4+ T cells in PBMCs of healthy human subjects preferentially reside within the CD4+CD25bright T cell population. Cells were incubated with FITC-conjugated anti-CD4 and PE-conjugated anti-CD25 mAbs. The analysis and sort gates were restricted to the population of lymphocytes by means of their forward and side scatter properties. Large, activated T cells were excluded. On reanalysis, the forward and side scatter properties of the CD4+CD25bright cells were not appreciably different from those of the CD4+CD25− population, indicating that these cell populations are similar in size. In another set of experiments, LPMCs were stained with CyChrome-anti-CD4, FITC-anti-CD25, and PE-anti-CD45RO or PE-anti-CD71 mAbs, and the various subsets of cells were purified by a FACSVantage. For polyclonal activation, cells were cultured with 5 μg/ml plate-bound anti-CD3 (UCHT1; BD Pharmingen) and 5 μg/ml soluble anti-CD28 (CD28.2; BD Pharmingen) for 48 h and used for FACS and RT-PCR analyses.
FACS analysis of surface and intracellular Ags
For two-color flow cytometry analysis, freshly isolated LPMCs and PBMCs were stained with FITC-conjugated anti-CD4 and PE-conjugated anti-CD25 (27). For three-color flow cytometry analysis, the samples stained with CyChrome-conjugated anti-CD4 and FITC-conjugated anti-CD25 were also stained with control PE-IgG1, or PE-conjugated CD45RO, PE-conjugated anti-CD45RA, PE-conjugated anti-HLA-DR, or PE-conjugated anti-GITR mAb. Before staining with PE-conjugated anti-CTLA-4 mAb, the cells were fixed and permeabilized with Cytofix/Cytoperm (BD Pharmingen) at 4°C for 30 min. Staining and washing were performed in Perm/Wash Buffer (BD Pharmingen), and cells were washed once in PBS before analysis. For GITR staining, the sorted CD4+CD25− and CD4+CD25bright cells stimulated with anti-CD3/CD28 mAbs were also used.
Cell stimulation assays
The CD4+CD25− cells were plated at 1.0 × 104/well, while the CD4+CD25bright cells were plated at 0, 0.5 × 104, or 1.0 × 104/well (27). Thus, upon coculture, the cells were combined at various ratios of TR responder in 96-well round-bottom plates (Corning Costar, Cambridge, MA) in the presence or absence of Con A (5 μg/ml), PHA (5 μg/ml), or soluble anti-CD3 mAb (1 μg/ml). To evaluate the effect of IL-2 on the break of anergy, cells were cultured with soluble anti-CD3 (1 μg/ml) in the presence or absence of rhIL-2 (100 U/ml). To assess a role of suppressive activity via competition between TR cells and responder cells for IL-2, various concentrations of rhIL-2 (0, 2, 5, 10, or 100 U/ml) were added. All wells received 5 × 104 MMC-treated CD4− cells as APCs. To determine proliferation, 50 μl of the culture supernatant was removed from each before 1 μCi of [3H]thymidine (NEN, Boston, MA) was added on day 3, the final 9 h of culture before harvesting. Percent proliferation was determined as (cpm incorporated in the coculture)/(cpm of responder population alone) × 100%.
Intracytofluorimetric analysis of IL-2, IL-10, and IFN-γ synthesis at the single-cell level was performed. Briefly, 1 × 106 cells were stimulated with 5 ng/ml PMA plus 500 nM ionomycin for 6 h, the last 4 of which was in the presence of 5 μg/ml brefeldin A (GolgiPlug; BD Pharmingen). Cells were collected, washed, fixed, and saponin permeabilized (Cytofix/Cytoperm; BD Pharmingen) and stained with cytokine-specific mAbs or isotype. In another set of experiments, the supernatants, which were removed before addition of [3H]thymidine at proliferation assays, were diluted and analyzed by a cytometric bead array (Th1/Th2 Cytokine CBA 1; BD Pharmingen) according to the manufacturer’s instructions.
RT-PCR for the detection of Foxp3
Total cellular RNA was extracted from 5 × 105 freshly sorted CD4+CD25− and CD4+CD25bright45). Their sequences were as follows; Foxp3 primers: 5′-TTCATGCACCAGCTCTCAACGG-3′ and 5′-TCGTCCATCCTCCTTTCCTTGATC-3′ (Sawader, Tokyo, Japan). PCR cycling conditions consisted of 95°C for 6 min, followed by 45 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s. Cycle threshold (CT) values were compared against a standard curve to estimate starting amounts of mRNA, and the relative expression of Foxp3Foxp3 levels were measured by dual-labeled probe RT-PCR using Smart Cycler (Cepheid, Sunnyvale, CA). The PCR contained 0.3 mM of each primer, 0.2 mM probe, 3 mM MgCL2 and 0.75 U of Platinum Taq
The results were expressed as the mean ± SD. Groups of data were compared by Mann-Whitney U test. Differences were considered to be statistically significant when p < 0.05.
Phenotypical properties of human CD4+CD25+ and CD4+CD25− LP T cells
Paired samples of PBMCs and LPMCs from 15 individuals were analyzed by flow cytometry for the presence of the CD4+CD25+ T cells. Consistent with previous reports describing human naturally occurring CD4+CD25+ TR cells (44, 46-49), a total of 6.0 ± 0.5% of the peripheral blood (PB) CD4+ T cells were CD25+ (Fig. 1⇓, A and B). Similarly, 6.4 ± 0.6% of the LP CD4+ T cells was also CD25+ (Fig. 1⇓, A and B). In addition, the mean fluorescence intensity (MFI) of LP CD4+CD25+ T cells is similar to that of the PB CD4+CD25+ T cells (data not shown). Since it has been recently reported that both naturally occurring CD4+CD25+ TR cells (44, 50, 51) and anergic memory TR cells (52) preferentially resided in the CD4+CD25bright T cell population in humans, we subdivided the CD4+CD25+ T cells into CD4+CD25bright and CD4+CD25int T cells. Also, 1.06 ± 0.23% and 0.90 ± 0.16% of the PB and the LP CD4+ T cells, respectively, were CD4+CD25bright (Fig. 1⇓C), indicating that the LP CD4+ T cells also may contain TR cells as well.
The PB and the LP CD4+CD25−, CD4+CD25+, and CD4+CD25bright T cells were next analyzed for expression of surface Ags to gain insight into their mechanism of action and to more fully characterize the TR population in human intestinal LP. Consistent with previous reports (53), CD45RO, which can be associated with proliferative responses to recall Ags (memory T cells), was expressed at significantly higher levels by the PB CD4+CD25+ and CD4+CD25bright populations as compared with the CD4+CD25− (Fig. 2⇓). In contrast to the PB CD4+ T cells, all groups, the LP CD4+CD25+ and CD4+CD25bright, and CD4+CD25− T cells preferentially expressed CD45RO. In contrast, the expression of CD45RA, considered a marker for naive T cells, showed the opposite expression profile. Consistent with the evidence that activated human T cells express HLA-DR molecules on their surface (54), the PB and LP CD4+CD25+ and CD4+CD25bright T cells expressed significantly higher HLA-DR, as compared with the PB CD4+CD25− T cells. In addition, the LP CD4+CD25− T cells did express HLA-DR as compared with the PB CD4+CD25− T cells (Fig. 2⇓). Taken together, the LP CD4+ T cells, regardless of CD25 expression, showed the phenotype characterized as activated memory T cells.
LP CD4+CD25bright T cells express CTLA-4, GITR, and Foxp3
Since CD4+CD8−CD25+ T cells in the PB and thymus from normal individuals are TR cells (44, 46, 47, 48, 49), we assessed whether or not the LP CD4+CD25+ and CD4+CD25bright T cells also express well-known TR markers, such as CTLA-4 and GITR, although these are also known as T cell activation markers. Expectedly, both intracellular CTLA-4 and surface GITR were significantly increased in/on the LP CD4+CD25+ and CD4+CD25bright T cells as compared with the LP CD4+CD25− T cells (Fig. 3⇓A). Interestingly, these markers were significantly up-regulated in/on the CD4+CD25bright T cells as compared with the CD4+CD25+ T cells (Fig. 3⇓A). However, unexpectedly, the LP CD4+CD25− T cells did also express CTLA-4 intracellularly, albeit the PB CD4+CD25− T cells did not (Fig. 3⇓A). In turn, GITR on the LP CD4+CD25− T cells was also up-regulated as compared with that on the PB CD4+CD25− T cells (Fig. 3⇓A). These data were also confirmed by the differences of the MFI ± SD from each population (Fig. 3⇓A).
Recently, Foxp3, which encodes a forkhead/winged-helix transcription factor known as scurfin (45), was found to be expressed specifically by naturally occurring CD4+CD25+ TR cells, but not by previously activated CD4+CD25+ T cells in mice. Thus, we next questioned whether or not the LP CD4+CD25− and CD4+CD25bright T cells expressed Foxp3. Consistent with previous mice studies (23, 24, 25), the PB CD4+CD25bright T population, predominantly transcribed Foxp3, whereas the PB CD4+CD25− T cells did not (Fig. 3⇑B). Somewhat surprising, the LP CD4+CD25− T cells as well as the LP CD4+CD25bright T cells did transcribe the gene (Fig. 3⇑B), although the semiquantitative RT-PCR confirmed that Foxp3 in the LP CD4+CD25− T cells was clearly expressed, but to a significantly smaller extent than that in the LP CD4+CD25bright T cells (Fig. 3⇑C). These results indicate that TR cells may reside within the LP CD4+CD25− T cell population as well as the LP CD4+CD25bright T cell population, which expressed CTLA-4, GITR, and Foxp3.
LP CD4+ T cells regardless of CD25 expression are anergic
Since the CD4+CD25+ TR cells show a partially anergic phenotype, we then questioned whether the LP CD4+CD25bright/CD4+CD25− T cells proliferate with or without PHA, Con A, or soluble anti-CD3 mAb stimulation in the presence of MMC-treated APCs. In this proliferation assay, we first separated the CD4+CD25bright T cells and CD4+CD25− T cells from paired PBMCs and LPMCs, respectively, using a FACSVantage. In both the PB and LP CD4+ T cells, the CD4+CD25bright subsets were found to be hyporesponsive to PHA, Con A, or anti-CD3 mAb stimulation as compared with the CD4+CD25− T cells (Fig. 4⇓), indicating that the PB and the LP CD4+CD25bright were anergic. Similarly, the PB and the LP CD4+CD25+ subsets were also hyporesponsive to PHA, Con A, or anti-CD3 mAb stimulation as compared with the PB and the LP CD4+CD25− T cells, respectively (data not shown). Furthermore, the LP CD4+CD25− T cells did proliferate to a significantly smaller extent as compared with the PB CD4+CD25− T cells, indicating that the LP CD4+CD25− cells were also anergic as compared with the paired PB CD4+CD25− cells. However, in the presence of exogenously added rhIL-2 at the concentration of 100 U/ml, stimulation with anti-CD3 mAb elicited proliferation not only in the PB and the LP CD4+CD25bright T cells but also in the LP CD4+CD25− T cells (Fig. 4⇓).
LP CD4+CD25bright T cells are regulatory
We next investigated the regulatory properties of the LP CD4+CD25bright/CD4+CD25− T cells by testing their ability to suppress the proliferative responses of the PB and LP CD4+CD25− T cells obtained from the same individuals. In a series of experiments, we used the CD4+CD25bright T cells for assessing their regulatory activity, because our preliminary data showed that the PB and LP CD4+CD25+ T cells, including CD25bright and CD25int subpopulations, did not suppress the proliferation of the PB CD4+CD25− responder cells at a high ratio of 1 TR:1 responder (data not shown). In contrast, the PB CD4+CD25bright T cells were able to suppress the proliferation of the PB CD4+CD25− cells when cocultured at a ratio of 0.5 TR:1 responder or 1 TR:1 responder in the presence of MMC-treated APCs and PHA (percent proliferation compared with that at culturing with responder alone, 0.5 TR:1 responder, 19.5 ± 4.4% (p < 0.05); 1 TR:1 responder; 12.6 ± 2.6% (p < 0.05), n = 7) (Fig. 5⇓A), indicating that the PB CD4+CD25bright T cells were regulatory. As a control, it was shown that titration of the same dose of the PB CD4+CD25− cells into the cultures did not affect the degree of proliferation, thereby excluding the possibility that an increase in total responder cell number was responsible for the suppressive effect (Fig. 5⇓A). However, the LP CD4+CD25bright T cells could not suppress the proliferation of the LP CD4+CD25− responder cells even at a high ratio of 1 TR:1 responder (percent proliferation, 88.5 ± 4.6% (p = 0.32), n = 7; Fig. 5⇓B), probably affecting the original feature of the anergic LP CD4+CD25− cells as shown in Fig. 4⇑. To re-estimate their regulatory properties, we next used the PB CD4+CD25− T cells from the same individuals as responders for the proliferation assay. Expectedly, the LP CD4+CD25bright T cells were able to suppress the proliferation of the PB CD4+CD25− cells when cocultured at a ratio of both 0.5 TR:1 responder and 1 TR:1 responder (percent proliferation, 0.5 TR:1 responder, 22.7 ± 4.4% (p < 0.05); 1 TR:1 responder, 12.8 ± 2.2% (p < 0.05), n = 7; Fig. 5⇓C), indicating that the LP CD4+CD25bright T cells were also regulatory. In contrast, the LP CD4+CD25− T cells could not suppress the proliferation of the PB CD4+CD25− cells (Fig. 5⇓C), although these cells expressed CTLA-4, GITR, and Foxp3 and were anergic (Figs. 3⇑ and 4⇑).
LP CD4+CD25bright T cells suppress cytokine productions from the CD4+CD25− T cells
The cytokine profile of the LP CD4+CD25bright T cells and the effect of these cells on cytokines produced by the CD4+CD25− cells were examined. After stimulation with PMA and calcium ionophore, no detectable levels of IL-2, IFN- γ, and IL-10 could be measured in both the PB and LP CD4+CD25bright T cells (Fig. 6⇓A). In contrast, the PB and LP CD4+CD25− T cells produced IL-2 and IFN-γ, but not IL-10 (Fig. 6⇓A). The striking difference between the LP CD4+CD25bright and CD4+CD25− T cell populations was that CD25bright cells failed to secrete IL-2, indicating that these cells have a specific defect in production of IL-2 (Fig. 6⇓A). To next determine whether the LP CD4+CD25bright T cells suppress cytokine secretion by CD4+CD25− responder T cells, the supernatants were collected at 63 h before addition of [3H]thymidine at proliferation assays and analyzed by a cytometric bead array, which allows multiparameter analysis in a single sample. As shown in Fig. 6⇓B, although the PB CD4+CD25− T cells (1 × 104 and 2 × 104 cells) in the absence of the LP CD4+CD25bright T cells produced a large amount of IL-2, IL-5, TNF-α, and IFN-γ, the LP CD4+CD25bright, but not the LP CD4+CD25− T cells, did clearly suppress these cytokine productions, indicating that the LP CD4+CD25bright T cell did affect not only the proliferation, but also the cytokine productions of the surrounding responder cells.
GITR/Foxp3 expressions in LP T cells after TCR stimulation
We next assessed the expression of GITR and FoxP3 on/in LP T cells in association with activation and memory markers, because it has been reported that some anergic activated/memory T cells have regulatory properties, indicating the peripheral development of TR cells. To do so, we first divided the LP CD4+ cells into CD71+ (activated) and CD71− or CD45RO+ (memory) and CD45RO− cells, respectively, thereafter we examined the expression of CD25/GITR and Foxp3 using FACS and RT-PCR analysis. As shown in Fig. 7⇓, A and B, both CD25/GITR and Foxp3 were preferentially expressed on/in LP CD71+ and CD45RO+ T cells as compared with LP CD71− and CD45RO− T cells, respectively. Furthermore, the semiquantitative RT-PCR confirmed that the Foxp3 expression in the LP CD71+ and CD45RO+ cells was significantly higher than that in the LP CD71− and CD45RO− cells, respectively (Fig. 7⇓C).
Since stimulation of peripheral CD4+CD25− T cells through TCR and CD28 leads to several outcomes, including the proliferation and induction of CD25 expression, it is possible that some or all of the LP CD4+CD25bright TR cells are the result of recent activation in the periphery. As shown in Fig. 7⇑D, both CD25 and GITR were up-regulated after the stimulation by anti-CD3/CD28 mAbs, although their expression was relatively low when compared with that of CD4+CD25bright cells. To next determine the nature of the Foxp3 expression in previous activation of LP T cells, we demonstrated Foxp3 expression in both sorted CD4+CD25− and CD4+CD25bright cells stimulated with anti-CD3/CD28 mAbs. As shown in Fig. 7⇑, E and F, anti-CD3/CD28-stimulated LP CD4+CD25− cells expressed the Foxp3 transcripts to an extent similar to that in anti-CD3/CD28-stimulated LP CD4+CD25bright cells. However, the expression of Foxp3 tended to be increased after the stimulation as compared with that in the starting freshly isolated population of CD4+CD25− T cells, but was not significant.
Origin of LP CD4+CD25bright T cells
Although human naturally arising CD4+CD25+ TR cells are thought to be mainly derived from thymus (46), a recent study by Walker et al. (55) suggests that Foxp3+CD4+CD25+ TR can be also developed from peripheral Foxp3−CD4+CD25− T cells by in vitro anti-CD3/CD28 mAb stimulation. Furthermore, it is well known that anergic human CD4+ T cells that are accompanied by an increased level of CD25 expression possess TR activity (52). Thus, it was important to determine whether our human LP CD4+CD25bright cells were naturally arising TR cells from the thymus or anergic cells that were developed in the periphery. As one of the differences between naturally arising CD4+CD25+ TR cells and anergic cells, it is well known that in vitro suppressive activity of naturally arising PB CD4+CD25+ TR cells usually can occur even if the TR:responder ratio is >0.1 TR:1 responder (49), whereas that of anergic cells can occur only if the ratio is >1 TR:1 responder (52). Furthermore, since it has been known that both the natural-occurring TR and the anergic cells require stimulation through the TCR to activate its TR program, we used soluble anti-CD3 mAb in place of PHA in this setting. Although the PB CD4+CD25bright T cells were able to significantly suppress the proliferation of the PB CD4+CD25− cells when cocultured even at a low ratio of 0.06 TR:1 responder (Fig. 8⇓A), the LP CD4+CD25bright T cells could suppress only at a relatively higher ratio of >0.25 TR:1 responder (Fig. 8⇓A). We next conducted an IL-2 consumption assay because it has been previously shown that low amounts of IL-2 abrogate the inhibition of the responder cells by anergic cells (52), but not by natural-occurring CD4+CD25+ TR cells (56). As shown in Fig. 8⇓B, the LP CD4+CD25bright cells as well as the PB CD4+CD25bright cells could inhibit the proliferative responses of the PB CD4+CD25− cells in the presence of low concentrations of IL-2 (2, 5, and 10 U/ml), but not a high concentration of IL-2 (100 U/ml), in combination with soluble anti-CD3 mAb (HIT3a, 1 μg/ml) and MMC-treated APCs. Importantly, there were no differences between the LP CD4+CD25bright cells and the PB CD4+CD25bright cells regarding regulatory activity at a 1:1 responder:TR ratio and at any concentrations of exogenously added IL-2.
LP CD4+CD25bright T cells from patients with active IBD also suppress the PB CD4+CD25− T cell proliferation
Finally, we assessed the role of CD4+CD25+ and CD4+CD25bright T cells in the LP from patients with active IBD to determine whether the LP CD4+CD25+/CD4+CD25bright T cells in IBD patients are activated pathogenic T cells or TR cells, which should be cells opposite of each other in terms of their pathogenesis. As shown in Fig. 9⇓, A and B, both the LP CD4+CD25+ and CD4+CD25bright T cells were significantly increased in patients with active CD or active UC as compared with normal individuals. Of note, the percentages of the LP CD4+CD25bright T cells per total LP CD4+CD25+ T cells were also significantly increased in IBD patients compared with those of normal individuals (Fig. 9⇓C). Although Barrat et al (57) recently reported a combination of immunosuppressive drugs, vitamin D3 and dexamethasone, induced human and mouse naive CD4+ T cells to differentiate in vitro into IL-10-producing regulatory T cells (named Tr1), there were no correlations between the medical treatments and the CD25 expression in IBD patients (data not shown). Similarly to the normal LP CD4+ T cells, the LP CD4+CD25− T cells as well as the LP CD4+CD25+ and CD4+CD25bright T cells from IBD patients did express intracellular CTLA-4 and surface GITR, although their expression was apt to be less than in/on the LP CD4+CD25bright/CD4+CD25+ T cells (Fig. 9⇓D). Furthermore, the LP CD4+CD25bright, but not CD4+CD25− T cells from active IBD patients, did suppress the proliferation of the paired PB CD4+CD25− T cells to an extent similar to that of the normal LP CD4+CD25bright T cells (Fig. 9⇓E), indicating that the LP CD4+CD25bright T cells from IBD patients contain protective CD4+CD25bright TR cells.
In the present study, we demonstrated that the LP CD4+CD25bright T cells isolated from normal human individuals possessed regulatory activity. Of particular importance, although many mucosal immunologists believed so far that the activated LP CD25+ T cells that were accumulated in the inflamed mucosa of IBD patients should be pathogenic, we also demonstrated that not only these cells and particularly their subpopulation of CD4+CD25bright T cells from IBD patients were significantly increased as compared with those from normal individuals, but also the LP CD4+CD25bright T cells from IBD patients retained regulatory activity to an extent similar to that of normal individuals. These results indicate that the LP CD4+CD25bright TR cells should be involved at least in part in the local T cell homeostasis both in normal and inflammatory conditions.
Although accumulating evidence shows that human PB and thymic CD4+CD8−CD25+ T cells possess regulatory activity (46, 47, 48, 49) with the same characteristics in mice, Baecher-Allan et al. (44) recently demonstrated that only CD4+CD25bright, but not the whole CD4+CD25+ T cells, have regulatory activity. This indicates that the whole CD4+CD25+ T cells contain a relatively high proportion of previously activated T cells rather than naturally occurring CD4+CD25bright TR cells, thereby the freshly isolated whole CD4+CD25+ T cells cannot suppress the proliferation of autologous CD4+CD25− T cells in coculture. Consistent with their report, we also detected TR activity in the fraction of the PB and LP CD4+CD25bright, but not in the whole CD4+CD25+ T cells. Furthermore, the evidence that the LP CD4+CD25bright T cell fraction rather than whole CD4+CD25+ T cells in patients with IBD was significantly increased as compared with normal individuals may indicate that the LP CD4+CD25bright T cells function to control or down-modulate the excessive immune responses in inflamed mucosa of IBD patients. In support of this, it has recently been shown that CD4+CD25+ TR cells can reverse the established intestinal inflammation in CD4+CD45RBhigh T cell-transferred colitic mice (58).
However, the question arises as to why the patients with active IBD in this study were obligated to receive the surgical operation because of the limitation by the conventional drug therapy to control the severe inflammation, although the increased LPCD4+CD25bright TR cells resided in the inflamed mucosa. We believe that the answer lies in the fact that the high levels of various cytokines (e.g., IL-2, IL-6, and IL-15) that are produced in severely inflamed mucosa from active IBD patients (40) and the up-regulated expression of costimulatory molecules, such as CD80 and CD86, and class II MHC molecules (40), thereby these would abolish the regulatory activity of the LP CD4+CD25bright TR cells and enhance the LP CD4+ T cell activation in situ. In line of this hypothesis, first, two reports (59, 60) revealed that not only a high dose of IL-2 along with TCR stimulation triggers their proliferation and neutralizes the suppressive activity during their proliferation, but also upon removal of IL-2, the suppression-broken CD4+CD25+ T cells revert to their original suppressive state. Second, IL-6 produced by activated dendritic cells in response to TLR ligand during chronic inflammation is critical for T cell activation and efficient blockade of TR activity (61). Thus, the possibility cannot be excluded that the increased LP CD4+CD25bright TR cells do not function well at the site of severe inflammation, thereby our patients could not overcome their intestinal inflammation by the conventional drug therapy, such as corticosteroids and immunosuppressants. We are now investigating whether additional IL-2 or IL-6 affects the suppressive activity of the LP CD4+ T cells in an in vitro assay and the correlation between the number of the LP CD4+CD25bright TR cells and clinical course, especially the onset of IBD.
Furthermore, it was very important to determine whether or not the LP CD4+CD25bright TR cells were identical to the PB CD4+CD25bright TR cells originally described by others (46, 47, 48, 49) in terms of their origin. Unlike the PB CD4+CD25bright TR cells that could suppress responder cells even at a TR:responder ratio of 0.06 TR:1 responder in the presence of soluble anti-CD3 mAb, the LP CD4+CD25bright TR cells needed a relatively higher responder:TR ratio of 0.25 TR:1 responder to suppress the responders in the same setting. Like the PB CD4+CD25bright TR cells, however, the suppression by the LP CD4+CD25bright TR cells could not be reversed by addition of lower amounts of IL-2 (2–10 U/ml). Although it might well be that in this special situation, mucosal TR activity is relatively mild as compared with the PBCD4+CD25bright TR cells, it is likely that the LP CD4+CD25bright TR cells are derived mainly from natural-occurring TR cells because the suppression by the LP TR cells was observed at <1 TR:1 responder (0.5 TR:1 responder and 0.25 TR:1 responder) and was not caused by IL-2 consumption in our analysis.
Besides the natural-occurring CD4+CD25bright TR cells and the anergic cells, the existence of other distinct groups of TR cells, such as Tr1 cells and Th3 cells and their role in controlling intestinal inflammation along with suppressor cytokines IL-10 and TGF-β, should also be considered (4). As a possibility, two groups have recently reported data suggesting that contact with CD4+CD25+ TR cells causes CD25− responder T cells to become suppressive themselves, the mechanism of which is referred to as “infectious tolerance,” although its cytokine dependency is still controversial (37, 62). This mechanism of infectious tolerance could explain not only how such a small LP CD4+CD25bright population (1–2%) of cells can regulate a much larger population of responder cells in vivo, but also how the LP CD4+CD25− and CD4+CD25int T cells could express CTLA-4, GITR, and Foxp3.
The functions of mucosal T cells are largely uncertain, but cells with a “memory” phenotype predominate in both the epithelium and the LP, indicating that they have been exposed to Ags (40). Thus, LP CD4+ T cells may be of particular importance to local immune regulation. They are generally anergic to TCR-mediated proliferative signals in normal mucosal conditions (28, 40, 63). The role of CTLA-4 that is expressed in activated/memory, but not naive T cells, in maintaining peripheral tolerance is well established (64, 65). Nevertheless, to our best knowledge, our observation that CTLA-4 is constitutively expressed in the LP CD4+CD25− as well as CD4+CD25bright T cells, but not in the PB CD4+CD25− T cells is the first report and may be in line with the general view on the acquisition and maintenance of mucosal tolerance. Further study will be needed to determine whether or not blockade of CTLA-4 abolishes the anergic condition of the LP CD4+CD25− T cells.
Based on the evidence that the LP CD4+CD25− T cells not only expressed CTLA-4, GITR, and Foxp3, but also showed anergic cells as compared with the paired PB CD4+CD25− T cells, we questioned whether the LP CD4+CD25− T cells are also TR cells as well as the LP CD4+CD25bright cells. Unexpectedly, although we could not detect regulatory activity in the LP CD4+CD25− subpopulation in our in vitro system at a ratio of 1 TR:1 responder, it is possible that only a part of the LP CD4+CD25− subpopulation expressed CTLA-4, GITR, and Foxp3 and retains TR activity, thereby none of the LP CD4+CD25− cells showed regulatory function in vitro by the indicated ratio. However, it should be noted that anti-CD3/CD28-stimulated LP CD4+CD25− T cellsexpressed Foxp3 to an extent similar to the stimulated LPCD4+CD25bright TR cells. Consistent with this, Walker et al. (55) have very recently demonstrated that anti-CD3/CD28-stimulated CD4+CD25− peripheral T cells express Foxp3 and that these activated cells are equally as suppressive as natural-occurring peripheral CD4+CD25+ TR cells. In addition, increasing evidence suggests that regulatory CD4+ T cells exist in the CD25− compartment in mice as well (26, 30, 31, 32, 33, 34), although it is widely accepted that CD4+CD25+ T cells are the major population that contains TR cells. In particular, three independent groups, including us, recently demonstrated that several fractions of CD4+CD25− T cells, such as αE+CD4+CD25− (33), LAP+CD4+CD25− (66), and GITR+CD4+CD25− (27), could suppress the development of murine CD4+CD45RBhigh T cell-transferred colitis. Further studies will be needed to address the correlation between the anergic memory LP CD4+CD25− T cells and TR function by sorting CD4+CD25− T cell fractions, which are enriched TR cells, such as LP αE+CD4+CD25−, LAP+CD4+CD25−, and GITR+CD4+CD25− T cells.
In conclusion, the LP CD4+CD25bright T cells from both normal individuals and IBD patients were demonstrated to be mucosal TR cells, indicating that these cells might not only serve as mucosal immune controllers to maintain T cell homeostasis against many stimuli, such as intestinal bacteria, but also as ideal cellular immunosuppressive cells as beneficial therapy for IBD patients in place of chemical immunosuppressants, which retain many side effects.
We express special thanks to Drs. Ryuichi Okamoto, Eriko Okada, Kiichiro Tsuchiya, Takahiro Kawamura, Shin Namiki, and Tomoko Matsumoto for the assessment of patients’ records and to Drs. Claudio Fiocchi, Warren Strober, and Barfor R. Sartor for helpful discussion and encouragement.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This study was supported in part by Grants-in-Aid for Scientific Research, Scientific Research on Priority Areas, Exploratory Research and Creative Scientific Research from the Japanese Ministry of Education, Culture, Sports, Science and Technology; the Japanese Ministry of Health, Labor and Welfare; the Japan Medical Association; Foundation for Advancement of International Science; Terumo Life Science Foundation; Ohyama Health Foundation; Yakult Bio-Science Foundation; and Research Fund of Mitsukoshi Health and Welfare Foundation.
↵2 Address correspondence and reprint requests to Dr. Takanori Kanai, Department of Gastroenterology and Hepatology, Graduate School of Medicine, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. E-mail address:
3 Abbreviations used in this paper: TR, regulatory T; IBD, inflammatory bowel disease; GITR, glucocorticoid-induced TNFR family-related protein; LP, lamina propria; LPMC, lamina propria mononuclear cell; MMC, mitomycin C; CD, Crohn’s disease; UC, ulcerative colitis; rh, recombinant human; CT, critical threshold; PB, peripheral blood; MFI, mean fluorescence intensity.
- Received December 10, 2003.
- Accepted June 23, 2004.
- Copyright © 2004 by The American Association of Immunologists