Regulatory T cells (Tregs) play a fundamental role in regulating the immune system in health and disease. Considerable evidence exists demonstrating that transfer of Tregs can cure colitis and a variety of other inflammatory disorders. However, little is known about the effects of inflammation on resident Tregs. Mice (BALB/c or C57BL/6) treated with an intrarectal instillation of the haptenizing agent 2,4-dinitrobenzene sulfonic acid (DNBS) develop an acute inflammatory disease, the histopathology of which peaks at 3 days posttreatment and resolves spontaneously thereafter. In this study we demonstrate that DNBS (or oxazolone)-induced colitis causes a depletion of colonic Foxp3+ Tregs 8 days posttreatment, while the proportion of Foxp3+ cells in the ileum, mesenteric lymph nodes, and spleen remains unchanged. Replenishment of the colonic Treg population was associated with the reappearance of mucosal homing (α4β7+) CD4+Foxp3+ Tregs. Assessing the mechanism of local Treg depletion, we found no evidence to implicate cytokine-induced phenotypic switching in the Foxp3+ population or increased SMAD7 expression despite the essential role that TGF-β has in Foxp3+ Treg biology. Increased Fas ligand (FasL) expression was observed in the colon of colitic mice and in vitro stimulation with a Fas cross-linking Ab resulted in apoptosis of CD4+Foxp3+ but not CD4+Foxp3 cells. Furthermore, DNBS-induced colitis in Fas/FasL-deficient mice did not result in depletion of colonic Tregs. Finally, adoptively transferred synergic Fas−/− but not Fas+/+ Tregs were protected from depletion in the colon 8 days post-DNBS treatment, thus substantiating the hypothesis that inflammation-induced local depletion of Foxp3+ Tregs in the colon of mice occurs via Fas/FasL-mediated death.

The immune system is tasked with protecting the host from infection and disease and so must discriminate between self or otherwise harmless Ags and noxious, potentially dangerous Ags. In the intestine this represents a formidable challenge, because the large surface area required for nutrient digestion and absorption also provides ample opportunity for pathogen entry. Moreover, the mammalian intestine, particularly the lumen of the colon, is home to a vast microbiota that in this location is not only considered commensal but of benefit to the host (1, 2, 3). However, if the same organisms gain entry to the blood the outcome can be sepsis, multiorgan failure, and death. In light of this, it is not surprising that the gastrointestinal tract is the largest reservoir of T cells in the body and is often described as being in a continuous state of controlled inflammation, a state achieved through an exquisitely balanced interplay between proinflammatory, anti-inflammatory, and immunosuppressive/regulatory mechanisms.

Understanding the complexity of immune homeostasis and how this determines whether one recovers or succumbs to disease is of paramount importance, and a number of key soluble (e.g., TGF-β and IL-10) and cellular (e.g., regulatory T cells (Tregs)3) regulatory agents have been identified. Natural Tregs develop in the thymus where their transcriptional programming is imprinted and controlled by the transcription factor Foxp3 (4, 5, 6). In addition to natural Tregs and although the issue is a contentious one, there are data to support the possibility that peripheral T cells can be induced into a regulatory phenotype when stimulated appropriately to elicit Foxp3 expression (7). The site of Treg-controlled suppression is also a subject of debate; convincing data demonstrate that Treg suppression of mucosal inflammation is achieved during the priming of effector T cells in secondary lymphoid tissue (8), and there is also substantive data showing that immunosuppression at the effector site is necessary (9).

The importance of Tregs cells in the gastrointestinal tract has been demonstrated in rodent models of colitis, where spontaneous disease develops in their absence or upon the transfer of effector T cells devoid of Tregs into immunodeficient animals (10). Although there is little doubt that Tregs can suppress the development of inflammatory disease, there is a paucity of data on the reciprocal interaction, i.e., the effect(s) that inflammation has on Tregs resident in the colon. Thus, we asked the following three questions. First, does colitis alter the colonic Foxp3+ T cell population? Second, if distinct changes in the resident Treg population occur, are these events paralleled in secondary lymphoid tissues? Third, what is the mechanism underlying inflammation-induced changes in local, that is, colonic Tregs? Using established models of murine colitis, the data herein show for the first time that colitis results in a transient loss of Foxp3+ T cells that is restricted to the colon with no involvement of the small bowel, mesenteric lymph nodes, or spleen. Mechanistic investigations indicate that the depletion of Foxp3+ T cells in the colon is a consequence of Fas-Fas ligand (FasL)-induced death of this regulatory T cell population.

Six- to eight week old male BALB/c, C57BL/6, B6.Cg-IghaThy1aGpi1a/J (referred to as Thy1.1), and B6.MRL-Faslpr/J (referred to as B6-Fas−/−) mice were purchased from The Jackson Laboratory. BALB/c lpr/lpr gld/gld mice were provided by Dr. J. Erikson (Wistar Institute, Philadelphia PA). All mice were maintained in specific pathogen-free conditions at the University of Calgary (Calgary, Alberta, Canada) for 1 wk before use. Colitis was evoked by intrarectal (i.r.) instillation of 3 or 5 mg of 2,4-dinitrobenzene sulfonic acid (DNBS) in 100 μl of a 1:1 ethanol (EtOH): PBS solution via a 3-cm catheter in BALB/c and C57BL/6 mice, respectively. Mice receiving 50% EtOH alone or PBS only served as time-matched controls. Mice were sacrificed at 3, 8, 14, or 21 days posttreatment (EtOH controls or DNBS). In other studies, mice were treated with, oxazolone (5 mg i.r. in 50% EtOH), and assessed 8 days later. Colitis was assessed by clinical disease score (maximum = 5) based on presence (1) or absence (0) of the following: >10% weight loss, wet anus/soft stool/empty colon, shortening of the colon, anal bleeding/occult blood, macroscopic ulcers, and death. Histological damage scores were determined on H&E-stained sections (formalin-fixed, paraffin-embedded tissues; 3-μm sections) and are based on the following: loss of architecture, 0–3; inflammatory infiltrate, 0–3; goblet-cell depletion, 0 or 1; ulceration, 0 or 1; edema, 0 or 1; muscle thickening, 0–2; and presence of crypt abscesses, 0 or 1 (maximum score = 12) as published previously (11). All procedures were approved by the animal care committee at the University of Calgary and conformed to Canadian guidelines on animal care.

A 1-cm segment of a full thickness colon was excised from the mid-colon, snap frozen in liquid N2, processed using TRIzol according to manufacturer’s instructions (Invitrogen), and used for semiquantitative RT-PCR. Isolated RNA was reverse transcribed using an Iscript kit (Bio-Rad) and 1 μg of RNA was added along with the reverse transcriptase enzyme to the premixed buffer (dNTP, oligo(dT) primers, RNase H inhibitor). Synthesis of cDNA was performed in a Bio-Rad MyCycler, and 1 μg of this template cDNA was used in PCR with gene-specific primers (Table I) and the following parameters: initial denaturing (94°C for 5 min), 35 cycles (denaturing at 94°C for 30 s, annealing at 58°C for 5 s, extension at 72°C for 10 s), and a final extension at 72°C for 5 min. The PCR products were electrophoresed through 2% agarose gels containing 0.5 mg/ml ethidium bromide and visualized under UV light.

Table I.

Primer sequences used in this study

GeneForward Primer (5′-3′)Reverse Primer (5′-3′)
CD3 AAGGAGCAAAGGAGACG TGAGTCCTGCTGAGTTC 
GATA3 ATCAAGCCCAAGCGAAG GCTCTGCCTCTCTAACC 
β-Actin CCAGAGCAAGAGAGGTATCC CTGTGGTGGTGAAGCTGTAG 
Foxp3 ACTGCTGGCAAATGGAGTCT AAGTAGGCGAACATGCGAGT 
IL-4 TCAACCCCCAGCTAGTTGTC TGTTCTTCGCTGTGAGG 
IFN-γ ACTGGCAAAAGGATGGTGAC CGCAATCACAGTCTTGGCTA 
TNF-α GGCCTTTCCTACCTTCAGACC AGCAAAAGAGGAGGCAACAA 
Fas ATGCACACTCTGCGATGAAG TTCAGGGGTCATCCTGTCTCC 
FasL GCCCATGAATTACCCATGTC GTTCTGCCAGTTCCTTCTGC 
GeneForward Primer (5′-3′)Reverse Primer (5′-3′)
CD3 AAGGAGCAAAGGAGACG TGAGTCCTGCTGAGTTC 
GATA3 ATCAAGCCCAAGCGAAG GCTCTGCCTCTCTAACC 
β-Actin CCAGAGCAAGAGAGGTATCC CTGTGGTGGTGAAGCTGTAG 
Foxp3 ACTGCTGGCAAATGGAGTCT AAGTAGGCGAACATGCGAGT 
IL-4 TCAACCCCCAGCTAGTTGTC TGTTCTTCGCTGTGAGG 
IFN-γ ACTGGCAAAAGGATGGTGAC CGCAATCACAGTCTTGGCTA 
TNF-α GGCCTTTCCTACCTTCAGACC AGCAAAAGAGGAGGCAACAA 
Fas ATGCACACTCTGCGATGAAG TTCAGGGGTCATCCTGTCTCC 
FasL GCCCATGAATTACCCATGTC GTTCTGCCAGTTCCTTCTGC 

Splenocyte and mesenteric lymph node (MLN) cell isolation.

Spleens or MLNs were removed and placed into separate sterile tubes containing RPMI 10 medium supplemented with 5% FBS and 2% (v/v) penicillin/streptomycin (all from Sigma-Aldrich) and then mechanically disrupted by pressing through a 100-μm cell strainer. Pelleted cells were resuspended in 2 ml of ACK lysis buffer (0.15M NH4Cl, 10 mM KHCO3, and 0.1 MM EDTA; all from Sigma-Aldrich) for 2 min to remove erythrocytes. Following centrifugation, cells were resuspended at 5 × 106/ml for immediate use or purified into an enriched CD4+ T cell population (see below).

Colonic lamina propria lymphocyte (LPL) isolation.

LPLs were isolated according to a standard protocol (12) with minor modifications. Briefly, after mesentery and fat were removed, the excised colons were opened and cut into 0.5-mm pieces. Tissues were washed six times in Ca2+- and Mg2+-free (CMF) HBSS (Invitrogen). Tissues were incubated in tissue culture flasks on end with sterile stir bars in CMF HBSS with EDTA (8 mM) stirred at 300 rpm on stir plates at room temperature. After 15 min the supernatant containing epithelial cells and intraepithelial lymphocytes was discarded, fresh CMF HBSS with EDTA was added and the process was repeated an additional three times. Release of LPLs was accomplished by collagenase (150U/ml final, fraction V; Sigma-Aldrich) digestion of tissue, with agitation by stirring for 1 h at 37°C followed by the recovery of cells from the supernatant. To increase viability of the cells, 20 ml of RPMI 10 was added to the supernatants to dilute and neutralize the collagenase. Digestion with collagenase was repeated twice more, and the recovered cells were pooled, washed twice in CMF HBSS, resuspended at 106/ml, and kept on ice before use.

CD4+ T cell and Treg purification.

CD4+ T cells were purified from naive BALB/c spleens using MACS according to the manufacturer’s instructions (Miltenyi Biotec). Briefly, cells were resuspended in sterile, degassed 5% BSA/PBS (w/v) and incubated with a mixture of biotinylated Abs, leaving CD4+ T cells untouched, for 10 min at 4°C. Following the addition and binding of magnetic biotinylated microbeads, cells were added to a separation column held in a magnetic field that captured and retained the labeled cells while unlabeled CD4+ T cells are collected as flow through. Flow cytometry confirmed ≥95% CD4+ purity (data not shown). Treg cells were purified in a similar manner, with the additional step of positive selection for CD25+ cells in the CD4+ population. Purified gated lymphocytes were 91% CD4+CD25+, of which 87.5% were CD4+Foxp3+ as assessed by flow cytometry.

Purified Tregs were adoptively transferred to naive synergic recipients by i.p. injection of 107 cells in 100 μl of PBS (13, 14). Preliminary experiments were conducted with the transfer of CFSE-labeled BALB/c, or Thy1.2+ (C57BL/6) Tregs into synergic BALB/c or Thy1.1+ (C57BL/6) hosts to determine whether i.p. injection would allow for the homing and subsequent recovery of lamina propria Tregs 11 days after transfer. In other experiments, EtOH or 5 mg of DNBS (i.r.) was administered to recipient mice 3 days post-Treg transfer, and Treg recovery from the colon was assessed 8 days post-DNBS treatment. Adoptive transfer of Thy1.2+ cells into Thy1.1+ recipients allowed us to distinguish donor from recipient cells and to discriminate between simple loss of Foxp3 (as would be evidenced by increased CD4+Foxp3Thy1.2+ cells) or death (indicated by loss of CD4+Foxp3+Thy1.2+ without an increase in CD4+Foxp3Thy1.2+ cells).

Following a published protocol (14), 106 mixed splenocytes, MLN cells, or purified CD4+ T cells were cultured in the presence of plate-bound anti-CD3 Abs (1 μg/ml) with or without 2 μg/ml soluble anti-CD28 Abs (BD Biosciences) and 5 ng/ml TGF-β (R&D Systems). Five days later, Foxp3 protein expression was determined by immunoblotting. In addition, identical cell cultures also received mouse recombinant IFN-γ, IL-4, or TNF-α (all 20 ng/ml; Peprotech) simultaneously with TGF-β treatment or were pretreated with each cytokine for 3 days before treatment with the Foxp3 induction protocol.

Protein was extracted from isolated immune cells, full thickness colon, or mucosal scrapings. Laemmli sample buffer was added to 20 μg of protein and the preparation was boiled for 5 min and then subjected to discontinuous SDS-PAGE (4% stacking pH 6.8, 10% separating pH 8.8). Resolved proteins were transferred onto polyvinylidene difluoride membranes and blocked for 1 h in TBST (5% milk, Tris-buffered saline, and Tween 20). Membranes were washed three times (5 min each) and incubated for either 1 h (room temperature) or overnight (4°C) with primary Abs against one of the following Ags: Foxp3 (eBioscience), Zap70 mAb, cleaved caspase 3, total caspase 3, phospho-SMAD2, phospho-STAT1Y701 (Cell Signaling Technology), and SMAD7 (Abcam), STAT1 (Santa Cruz Biotechnology) (all of the Abs were used at 1/1000 and were rabbit anti-mouse, except for Foxp3, which was rat anti-mouse). Following washing, membranes were incubated with the appropriate secondary Ab (1/5000) in 5% TBST for 1 h at room temperature, washed extensively, and developed using ECL solution (15).

Flow cytometry was performed on freshly isolated splenocytes, MLN cells, and LPLs. In addition, blood obtained from cardiac puncture was collected into heparin sulfate-containing tubes and mixed, and 200 μl was transferred to a 50-ml conical tube along with 20 ml of PBS. After erythrocyte lysis (see above) the immune cells were resuspended, stained, and subjected to FACS. Cell staining was performed with anti-CD4, anti-Fas, anti-cleaved caspase 3 (BD Biosciences), anti-CD25, -anti-α4β7 (LPAM-1), anti-CD90.2 (Thy1.2), anti-FasL, and anti-Foxp3 (BioLegend) Abs. Briefly, 106 cells were transferred to flow cytometry tubes and washed once with staining buffer (BioLegend). Cells were resuspended in 100 μl of staining buffer and surface staining was performed for 20 min in the dark. Cells were washed twice in staining buffer followed by fixation and permeabilization in freshly prepared 1× Fix/Perm buffer (BioLegend). After extensive washing to remove residual fixative, cells were resuspended in permeabilization buffer and incubated with anti-mouse Foxp3-Alexa Fluor488. Following washing of the cells, data were acquired and analyzed on a FACScan using CellQuest Pro (BD Biosciences). Analysis was conducted on lymphocytes based on forward scatter and side scatter, with unstained, single stained, and isotype controls used to determine compensation and setting of gates. Absolute numbers of cells were calculated as described previously (16): absolute number = (percentage CD4+Foxp3+ among total cells) × (total cell count)/100.

Intracellular staining for SMAD7 and phosphorylated SMAD2 was performed as indicated above with Abs labeled with anti-rabbit Fab conjugated to either Alexa Fluor488 or PE (Zenon kit; Invitrogen).

Excised tissues (mid-colon, spleen, and MLN) were formalin fixed and paraffin embedded and 5-μm sections were collected onto coded slides. After deparaffinization, re-hydration, and epitope recovery (30 min in citrate buffer (citric acid and Triton X-100) in a food-steamer), endogenous peroxidases were blocked by incubation in 3% H2O2 (5 min). Sections were washed in TBS, blocked for 1 h with an IgG blocking reagent (mouse-on-mouse (MoM) kit, Vector Laboratories), washed in TBS (3 × 5 min), incubated at room temperature in MoM diluent for 5 min, and then treated with anti-Foxp3 (1/100; BioLegend) Ab for 30 min. Following three TBS washes, sections were incubated in biotinylated secondary Ab as indicated by the manufacturer’s instructions (Vector Laboratories) for 10 min. The biotin-streptavidin-based color development was performed using the Vectorstain Elite ABC kit with diaminobenzidine as a substrate that gives a positive brown reaction product. Slides were counterstained in hematoxylin, dehydrated and Foxp3+ cells enumerated in five random high power fields (×400 original magnification per tissue section). Staining for Fas ligand was performed on sections as described above with the following exceptions. Heat-induced epitope recovery was performed for 20 min followed by rapid cooling by placing the vessel containing the slides and buffer in ice-cold water for 10 min. Blocking was performed by incubation in 0.3% H2O2 for 30 min. Polyclonal Ab directed against FasL was added at 1/100 (Abcam) and incubated overnight at 4°C. Negative controls in which primary Ab was omitted and positive control tissues (i.e., spleen) were used to ensure specific staining.

The gels or other images shown are always representative of at least three separate experiments. Graphical data are mean ± SE and n values refer to the number of mice used in two or more separate experiments. Statistical analysis was performed for multiple groups using one-way ANOVA followed by the appropriate pair wise post hoc statistical test. Student’s t test was used for two-group comparisons. A level of statistically significant difference was accepted at p < 0.05.

Consistent with previous observations (11), DNBS treatment of BALB/c mice resulted in a transient colitis characterized by weight loss, diarrhea, tissue edema, ulceration, and loss of tissue architecture. As assessed by clinical disease and histological damage scores, the severity of the DNBS-induced colitis peaked at 3 days posttreatment and resolved thereafter; by day 14 post-DNBS treatment, the mice appeared ostensibly normal although, upon histological examination, a small amount of inflammatory cell infiltrate was still apparent in the colon (Fig. 1, open bars). A similar progression of clinical disease occurred in DNBS-treated C57BL/6 mice (data not shown), an observation consistent with other publications (12). In addition, the inflammation elicited by the haptenizing agent oxazolone was maximal 3 days posttreatment (data not shown) with substantial improvement noted by 8 days posttreatment (Fig. 1, closed bars).

FIGURE 1.

Intrarectal instillation of 3 mg of DNBS (open bars), or 5 mg of oxazolone (closed bars) into BALB/c mice induces a transient colitis as assessed by clinical disease scores (A) and histological damage scores (B) assessed by examination of H&E-stained sections of colonic tissue. ∗, p < 0.05 compared with EtOH (vehicle for DNBS delivery) and control (DNBS); #, p < 0.05 compared with control (oxazolone); n = 14–20, mean ± SEM. EtOH-treated mice were autopsied as time-matched controls with the DNBS time course and because the data were not appreciably different, the data were pooled and are shown as a single control EtOH group.

FIGURE 1.

Intrarectal instillation of 3 mg of DNBS (open bars), or 5 mg of oxazolone (closed bars) into BALB/c mice induces a transient colitis as assessed by clinical disease scores (A) and histological damage scores (B) assessed by examination of H&E-stained sections of colonic tissue. ∗, p < 0.05 compared with EtOH (vehicle for DNBS delivery) and control (DNBS); #, p < 0.05 compared with control (oxazolone); n = 14–20, mean ± SEM. EtOH-treated mice were autopsied as time-matched controls with the DNBS time course and because the data were not appreciably different, the data were pooled and are shown as a single control EtOH group.

Close modal

The effect of inflammation on colonic Foxp3+ Tregs was assessed during overt inflammation and restitution by several complimentary techniques. RNA isolated from the colons of mice 8 days post-DNBS-induced colitis revealed complete abolition of detectable Foxp3 mRNA during restitution of inflammation as shown by RT-PCR (Fig. 2,A). Significant reductions in the absolute number of Foxp3+ cells and the proportion of CD4+Foxp3+ cells in gated LPLs at 8 days post-DNBS treatment were also observed by immunohistochemistry (Fig. 2,B) and flow cytometry (Fig. 2,C). In addition, as determined by flow cytometry, the absolute number of CD4+Foxp3+ T cells recovered from the colon 8 days post-DNBS treatment was significantly reduced (Fig. 2,D). Depletion of Foxp3 protein and mRNA was not unique to BALB/c mice, as colonic Treg depletion was also observed in gated LPLs from C57BL/6 mice as assessed by flow cytometry for mean percentage (Fig. 2,E; 1.09 ± 0.03% vs 0.49 ± 0.03%) and absolute numbers of CD4+Foxp3+ cells (Fig. 2,F). This phenomenon was not restricted to DNBS colitis, as reduced Foxp3 mRNA was detected in BALB/c mice 8 days posttreatment with oxazolone (Fig. 2 G).

FIGURE 2.

Enteric inflammation reduces colonic Foxp3+ TregsA, RT-PCR performed on full thickness colonic tissues from BALB/c mice reveals reduced Foxp3 mRNA 8 days post-DNBS treatment that returned to normal by day 14 posttreatment (each lane represents a colon from an individual mouse). con, Control. B–D, This reduction in Foxp3 was also observed at the protein level as shown by reduced Foxp3+ cells detected by immunohistochemistry on colonic tissue sections (mean ± SEM, n = 6–8; H.P.F., high power microscopical field) (B), by flow cytometry conducted on colonic lamina propria lymphocytes from (n = 4–8 mice) (C), and by absolute numbers of recovered CD4+Foxp3+ cells (D). E and F, C57BL/6 mice (n = 4–6) also display reduced CD4+Foxp3+ cells in their colon 8 days post-DNBS treatment (presented as a percentage of gated lymphocytes) (E), and this depletion was also observed when absolute numbers of recovered CD4+Foxp3+ cells were calculated (F). G, Reduced Foxp3 mRNA was observed in colonic tissue 8 days after oxazolone-induced colitis in BALB/c mice (each lane represents tissue from an individual mouse). ∗, p < 0.05 compared with control.

FIGURE 2.

Enteric inflammation reduces colonic Foxp3+ TregsA, RT-PCR performed on full thickness colonic tissues from BALB/c mice reveals reduced Foxp3 mRNA 8 days post-DNBS treatment that returned to normal by day 14 posttreatment (each lane represents a colon from an individual mouse). con, Control. B–D, This reduction in Foxp3 was also observed at the protein level as shown by reduced Foxp3+ cells detected by immunohistochemistry on colonic tissue sections (mean ± SEM, n = 6–8; H.P.F., high power microscopical field) (B), by flow cytometry conducted on colonic lamina propria lymphocytes from (n = 4–8 mice) (C), and by absolute numbers of recovered CD4+Foxp3+ cells (D). E and F, C57BL/6 mice (n = 4–6) also display reduced CD4+Foxp3+ cells in their colon 8 days post-DNBS treatment (presented as a percentage of gated lymphocytes) (E), and this depletion was also observed when absolute numbers of recovered CD4+Foxp3+ cells were calculated (F). G, Reduced Foxp3 mRNA was observed in colonic tissue 8 days after oxazolone-induced colitis in BALB/c mice (each lane represents tissue from an individual mouse). ∗, p < 0.05 compared with control.

Close modal

The reduced number of Foxp3+ cells was not due to dilution resulting from CD4+Foxp3 recruitment, as colonic (Fig. 2,C) and peripheral blood CD4+ T cells were not increased until day 14 post-DNBS treatment (Fig. 3, A and B). Thus, the absence of CD4+Foxp3+ cells represents a real reduction and is not simply a consequence of increased infiltration of CD4+Foxp3 cells into the colon.

FIGURE 3.

Reappearance of colonic CD4+Foxp3+ T cells in the colon is associated with increased peripheral blood α4β7+ CD4+Foxp3+ cells. A, DNBS-induced colitis significantly reduces CD4+ Foxp3 and CD4+Foxp3+ cells in the blood 3 days post-DNBS treatment. Despite the reappearance of Tregs in the blood by day 8 post-DNBS treatment (when colonic Treg depletion occurs; see Fig. 2 C), these cells do not bear the mucosal homing integrin α4β7+. B, Replenishment of colonic Tregs at day 14 post-DNBS treatment is associated with the reappearance of α4β7-expressing Tregs in the blood. C, CD4+Foxp3+ cells as a percentage of gated lymphocytes were decreased in the spleens of mice treated 14 days previously with DNBS. ∗, p < 0.05; ∗∗, p < 0.01; n = 3–5 mice; although a similar event was not observed in mesenteric lymph node cells (data not shown).

FIGURE 3.

Reappearance of colonic CD4+Foxp3+ T cells in the colon is associated with increased peripheral blood α4β7+ CD4+Foxp3+ cells. A, DNBS-induced colitis significantly reduces CD4+ Foxp3 and CD4+Foxp3+ cells in the blood 3 days post-DNBS treatment. Despite the reappearance of Tregs in the blood by day 8 post-DNBS treatment (when colonic Treg depletion occurs; see Fig. 2 C), these cells do not bear the mucosal homing integrin α4β7+. B, Replenishment of colonic Tregs at day 14 post-DNBS treatment is associated with the reappearance of α4β7-expressing Tregs in the blood. C, CD4+Foxp3+ cells as a percentage of gated lymphocytes were decreased in the spleens of mice treated 14 days previously with DNBS. ∗, p < 0.05; ∗∗, p < 0.01; n = 3–5 mice; although a similar event was not observed in mesenteric lymph node cells (data not shown).

Close modal

Moreover, Treg depletion was restricted to the colon (i.e., the site of inflammation), as DNBS-colitis did not reduce CD4+Foxp3+ mRNA in terminal ileum with or without Peyer’s patches or MLNs at any time point examined (data not shown). The proportions of splenic or MLN CD4+Foxp3+ gated lymphocytes were not altered 3 and 8 days post-DNBS treatment but were significantly decreased in the spleen at 14 days post-DNBS treatment (2.48 ± 0.06% to 1.59 ± 0.1%) (Fig. 3,C), suggesting an egress of these cells from the spleen and mucosal homing. Indeed, while mucosal homing α4β7+CD4+Foxp3+ cells are reduced in the blood 8 days post-DNBS treatment, Treg re-population of the colon occurred only when CD4+Foxp3+α4β7+ cells were found at normal (i.e., pretreatment) levels in the blood at 14 days post-DNBS treatment (compare Figs. 2,B and 3 B).

DNBS-induced colitis was associated with increased TNF-α, IFN-γ, and IL-4 mRNA as detected by RT-PCR. To test the hypothesis that these proinflammatory mediators might underlay the loss of Foxp3 in the colon, purified CD4+ T cells from the spleen were stimulated with anti-CD3+TGF-β+ and anti-CD28 (known to induce Foxp3) with or without IFN-γ, TNF-α, or IL-4 (all 10 ng/ml). Pretreatment or cotreatment with IFN-γ, TNF-α, or IL-4 failed to diminish the Foxp3 signal (data not shown). Similar changes in cytokine mRNA were noted in segments of terminal ileum (with or without Peyer’s patches) from mice treated with DNBS, but this region of intestine showed neither signs of inflammation nor depletion of CD4+Foxp3+ cells. These findings indicate that exposure to proinflammatory cytokines does not account for Treg depletion in the colon of colitic mice.

Having identified a tissue-specific depletion of Foxp3+ Tregs, we considered whether an inflammation-induced TGF-β signaling dysfunction was responsible for the depletion of colonic Tregs, as this signaling cascade is critical for the induction and maintenance of Tregs (7, 17). Immunoblotting of protein extracts from colonic mucosal scrapings revealed an increase in the inhibitory SMAD7 protein at 8 days post-DNBS treatment, correlating with the depletion of colonic Foxp3+ Tregs (Fig. 4,A, lower panel). In accordance with our analysis of the mouse colon, increased SMAD7 expression in tissues from patients with inflammatory bowel disease (IBD) has been presented where increased IFN-γ activity was postulated as being responsible for the SMAD7 signal (16). Despite DNBS-induced increases in IFN-γ mRNA (data not shown), IFN-γ-elicited signaling as gauged by STAT1 phosphorylation was not detected in the colonic mucosa (Fig. 4,A, upper panel). Inflammation-induced SMAD7 production was not restricted to the colonic mucosa, as expression was increased in Western blotting conducted on protein extracts from splenocytes and MLN cells 8 days post-DNBS treatment where Foxp3+ cell numbers were unaltered (data not shown). However, this alteration to SMAD signaling did not effect CD4+ T cells (and therefore not Foxp3+ cells), as purified CD4+ T cells from MLNs (data not shown) or flow cytometry of CD4+ gated MLN and LPL cells revealed no increase in SMAD7 post-DNBS treatment (Fig. 4 B).

FIGURE 4.

SMAD signaling dysfunction occurs as a response to DNBS-induced colitis but does not account for the colonic Treg depletion. A, Phosphorylation of STAT1 (pSTAT1; Tyr701) and SMAD7 expression were assessed in extracts of colonic mucosa from BALB/c mice 3 or 8 days post-DNBS treatment. STAT1 is present in the mucosa, but STAT1 phosphorylation was not detected. Mucosal SMAD7 expression is increased at day 8 post-DNBS treatment (each lane represents a different mouse; X, empty lane; + CON, positive control whole cell lysate from IFN-γ-stimulated (20 ng/ml for 30 min) splenocytes; a representative experiment is shown of n = 8–10 mice/group). B, LPLs were isolated from mice 8 days post-DNBS treatment and controls and stimulated in vitro with TGF-β (10 ng/ml for 30 min). Cells were stained for CD4, SMAD7, and phospho-SMAD2 and analyzed by flow cytometry of gated CD4+ lymphocytes. SMAD7 expression in CD4+ LPLs from control and colitic (8 days post-DNBS treatment) mice was not significantly different (left FACS plot and left graph). TGF-β stimulation increased the percentage of phospho-SMAD2+ in CD4+ gated LPLs isolated from control but not DNBS mice (middle and right FACS plot, right graph). Data are mean ± SEM,;∗, p < 0.01 compared with control (unstimulated) cells from control (ethanol-treated) mice; n = 4–6).

FIGURE 4.

SMAD signaling dysfunction occurs as a response to DNBS-induced colitis but does not account for the colonic Treg depletion. A, Phosphorylation of STAT1 (pSTAT1; Tyr701) and SMAD7 expression were assessed in extracts of colonic mucosa from BALB/c mice 3 or 8 days post-DNBS treatment. STAT1 is present in the mucosa, but STAT1 phosphorylation was not detected. Mucosal SMAD7 expression is increased at day 8 post-DNBS treatment (each lane represents a different mouse; X, empty lane; + CON, positive control whole cell lysate from IFN-γ-stimulated (20 ng/ml for 30 min) splenocytes; a representative experiment is shown of n = 8–10 mice/group). B, LPLs were isolated from mice 8 days post-DNBS treatment and controls and stimulated in vitro with TGF-β (10 ng/ml for 30 min). Cells were stained for CD4, SMAD7, and phospho-SMAD2 and analyzed by flow cytometry of gated CD4+ lymphocytes. SMAD7 expression in CD4+ LPLs from control and colitic (8 days post-DNBS treatment) mice was not significantly different (left FACS plot and left graph). TGF-β stimulation increased the percentage of phospho-SMAD2+ in CD4+ gated LPLs isolated from control but not DNBS mice (middle and right FACS plot, right graph). Data are mean ± SEM,;∗, p < 0.01 compared with control (unstimulated) cells from control (ethanol-treated) mice; n = 4–6).

Close modal

Inflammation-induced alterations to SMAD signaling were not restricted to increased SMAD7 expression. Western blotting and flow cytometry conducted on isolated MLN cells and splenocytes demonstrated that cells isolated from DNBS-treated mice, but not controls, exhibited increased basal SMAD2 phosphorylation and were refractive to TGF-β stimulation in vitro (data not shown). Phosphorylation of SMAD2 in CD4+ gated lymphocytes increased following TGF-β stimulation of LPLs from control mice from 7.91 ± 0.05% to 69.8 ± 1.8%. In contrast, 79.3 ± 3.5% LPLs isolated from colon 8 days post-DNBS treatment had phosphorylated SMAD, and this was not significantly increased by TGF-β treatment (77.0 ± 3.9%) (Fig. 5 B). SMAD2 phosphorylation in CD4+ gated MLN lymphocytes increased after TGF-β stimulation from 4.00 ± 0.25% to 81.5 ± 2.8% and 62.7 ± 10.9% to 83.9 ± 1.9% in control and DNBS-treated mice, respectively. Thus, SMAD signaling dysfunction in CD4+ T cells elicited by DNBS-induced colitis is restricted to saturation of the SMAD signaling cascade. Additionally, as altered SMAD signaling occurs in tissues without subsequent decreases in Tregs, SMAD signaling dysfunction cannot be the primary event responsible for the depletion of colonic Tregs.

FIGURE 5.

Enteric inflammation induced by DNBS increases FasL expression. A, Increased expression of FasL mRNA was detected by RT-PCR in full thickness colonic tissue excised from BALB/c mice treated 3, 8 or 14 days previously with DNBS (each lane represents an individual mouse, n = 6–8 mice total) and had returned to control levels by 21 days post-DNBS treatment. B, Immunohistochemical staining for FasL confirmed increased FasL protein following the induction of colitis. Expression of FasL at 3 days post-DNBS treatment was localized to epithelial cells, intraepithelial leukocytes, and mononuclear cells in the lamina propria, while at 8 days post-DNBS treatment FasL expression was increased within the smooth muscle. FasL positive cells from five randomly chosen high power fields (H.P.F.; original magnification, ×200). Data are mean ± SEM; ∗, p < 0.05; n = 4–5 mice/group.

FIGURE 5.

Enteric inflammation induced by DNBS increases FasL expression. A, Increased expression of FasL mRNA was detected by RT-PCR in full thickness colonic tissue excised from BALB/c mice treated 3, 8 or 14 days previously with DNBS (each lane represents an individual mouse, n = 6–8 mice total) and had returned to control levels by 21 days post-DNBS treatment. B, Immunohistochemical staining for FasL confirmed increased FasL protein following the induction of colitis. Expression of FasL at 3 days post-DNBS treatment was localized to epithelial cells, intraepithelial leukocytes, and mononuclear cells in the lamina propria, while at 8 days post-DNBS treatment FasL expression was increased within the smooth muscle. FasL positive cells from five randomly chosen high power fields (H.P.F.; original magnification, ×200). Data are mean ± SEM; ∗, p < 0.05; n = 4–5 mice/group.

Close modal

Given that Fas ligation induces apoptosis in human Tregs (18) and the effect of Fas stimulation on mouse Tregs had not been reported, we assessed the contribution of Fas/FasL interaction to the depletion of colonic Tregs. Colonic FasL mRNA expression in DNBS-treated mice followed a pattern of expression that was opposite to that of Foxp3, increasing 3 and 8 days post-DNBS treatment (Fig. 5,A). Fas mRNA expression was constitutive and unaltered by inflammation (Fig. 5,A). Immunohistochemistry for FasL confirmed significantly increased numbers of FasL+ lamina propria mononuclear cells and epithelial cells 3 days post-DNBS treatment (Fig. 5 B). Expression of FasL 8 days post-DNBS treatment was apparent on smooth muscle cells, lamina propria mononuclear cells, and epithelial cells from the base to the tip of the crypts, with only a few epithelial cells at the luminal surface positive for FasL by 14 days post-DNBS-induced colitis. In contrast to colonic tissue, FasL expression was not increased in the spleen or MNL nodes of colitic mice. For example, flow cytometry revealed that 5.39 ± 0.45% of spleen cells from control mice were FasL+ and 5.87 ± 0.86 (n = 3) of splenocytes excised from mice 8 days post-DNBS treatment were FasL+.

Isolated LPLs from BALB/c mice were positive for surface Fas expression as determined by flow cytometry. Expression of Fas on colonic LPL CD4+Foxp3+ cells was constitutive and remained unchanged 3 days post-DNBS treatment, a time when inflammation is ongoing and Tregs are still present (data not shown). We assessed the susceptibility of Tregs to Fas/FasL-induced apoptosis by stimulating naive BALB/c splenocytes with the Fas-activating mAb Jo-2 with or without anti-CD3 in vitro, monitoring for caspase 3 cleavage by flow cytometry. Cleaved caspase 3 was significantly increased in CD4+Foxp3+- but not CD4+Foxp3-gated T cells after incubation with Jo-2 with or without anti-CD3 (Fig. 6), demonstrating that mouse Foxp3+ Tregs are sensitive to Fas-induced apoptosis.

FIGURE 6.

CD4+Foxp3+ Tregs but not CD4+Foxp3 cells are rapidly killed in vitro in response to stimulation through Fas. Isolated splenocytes from BALB/c mice were treated with staurosporine (STS; 2.5 μM) or the Fas cross-linking (activating) Ab (Jo-2, 10 μg/ml; anti-Fas (αFas) mAb) with or without the plate-bound anti-CD3 (αCD3) Ab (1 μg/well) for 8 h and apoptosis was assessed by caspase-3 cleavage as detected by flow cytometry. A and B, CD4+Foxp3+ cells (A) show a significant increase in basal expression of cleaved caspase 3 and, unlike CD4+Foxp3 cells (B), are highly susceptible to anti-Fas and anti-CD3-induced death. C and D, The graphs show the FACS data normalized to control (i.e., unstimulated) and as the percentage of cells expressing cleaved caspase-3 in CD4+Foxp3+- (C) and CD4+Foxp3-gated (D) cells, respectively. ∗ p < 0.05; ∗∗, p < 0.01; n = 3 mice.

FIGURE 6.

CD4+Foxp3+ Tregs but not CD4+Foxp3 cells are rapidly killed in vitro in response to stimulation through Fas. Isolated splenocytes from BALB/c mice were treated with staurosporine (STS; 2.5 μM) or the Fas cross-linking (activating) Ab (Jo-2, 10 μg/ml; anti-Fas (αFas) mAb) with or without the plate-bound anti-CD3 (αCD3) Ab (1 μg/well) for 8 h and apoptosis was assessed by caspase-3 cleavage as detected by flow cytometry. A and B, CD4+Foxp3+ cells (A) show a significant increase in basal expression of cleaved caspase 3 and, unlike CD4+Foxp3 cells (B), are highly susceptible to anti-Fas and anti-CD3-induced death. C and D, The graphs show the FACS data normalized to control (i.e., unstimulated) and as the percentage of cells expressing cleaved caspase-3 in CD4+Foxp3+- (C) and CD4+Foxp3-gated (D) cells, respectively. ∗ p < 0.05; ∗∗, p < 0.01; n = 3 mice.

Close modal

To determine whether colonic Treg depletion was due to increased Fas/FasL interactions, we evaluated this T cell population following DNBS-induced colitis in mice deficient in both Fas (lpr/lpr) and FasL (gld/gld). Mice were given DNBS i.r. at 5 wk of age, a time when there was no evidence of autoimmunity or lymphoadenopathy (19). Despite the deficiency in both Fas and FasL, DNBS treatment of these mice resulted in disease as assessed by clinical disease scores (Fig. 7,A). Additionally, colonic Treg depletion was not observed, with CD4+Foxp3+ cells comprising 2.7 ± 0.23% and 2.23 ± 0.17% of gated LPLs 8 days post-EtOH and post-DNBS treatment, respectively (Fig. 7; n = 3 mice/group).

FIGURE 7.

Transient depletion of CD4+Foxp3+ cells in the colon does not occur in mice lacking Fas and FasL. A, Wild-type and Fas/FasL−/− mice develop colitis in response to 3 mg of DNBS (i.r.) as gauged by disease clinical scores. B and C, Despite the colitis there is no depletion of CD4+Foxp3+ cells as assessed by the percentage of gated lamina propria lymphocytes (B) or total cells (C) as determined by flow cytometry, respectively (n = 3 mice/group).

FIGURE 7.

Transient depletion of CD4+Foxp3+ cells in the colon does not occur in mice lacking Fas and FasL. A, Wild-type and Fas/FasL−/− mice develop colitis in response to 3 mg of DNBS (i.r.) as gauged by disease clinical scores. B and C, Despite the colitis there is no depletion of CD4+Foxp3+ cells as assessed by the percentage of gated lamina propria lymphocytes (B) or total cells (C) as determined by flow cytometry, respectively (n = 3 mice/group).

Close modal

To decisively assess the contribution of Fas/FasL-induced apoptosis to the reduction of Foxp3+ cells, Thy1.2+B6-Fas+/+ or Thy1.2+B6-Fas−/− Tregs were adoptively transferred into synergic Thy1.1+ recipients 3 days before DNBS treatment (Fig. 8, A and B). Although the severity of DNBS-colitis was not reduced in Thy1.2+B6-Fas−/− or Thy1.2+B6-Fas+/+Tregs recipients, Fas−/− Tregs but not Fas+/+ Tregs were protected from depletion in the colon. Although the proportion of CD4+Foxp3+ gated LPLs was unchanged in Thy1.2+B6-Fas−/−Tregs recipients (1.35 ± 0.29% compared with 1.71 ± 0.07%), these cells were significantly reduced in Thy1.2+B6-Fas+/+Treg recipients (1.17 ± 0.04% compared with 0.32 ± 0.04%) in DNBS-treated mice (Fig. 8). This event was substantiated by calculation of the number of CD4+Foxp3+ cells in the various animal groups, in which only the transfer of Fas+/+ cells and not Fas−/− cells resulted in reduced numbers of CD4+Foxp3+ T cells retrievable from the colon 8 days post-DNBS treatment (Fig. 8 D). Conversion of CD4+Foxp3+ cells into other T cell types was assessed by determining the proportion of Thy1.2+CD4+Foxp3 (donor Tregs) following DNBS-induced colitis. Although a small proportion of CD4+Foxp3 cells were Thy1.2+, this cell population was not significantly increased following DNBS-induced colitis as compared with recipients of Thy1.2+ Tregs treated with EtOH.

FIGURE 8.

Fas-bearing Foxp3+ Tregs are subject to depletion following DNBS colitis. A and B, In pilot studies, purification protocols for Tregs revealed 87.5% purity as defined by Foxp3 staining (A) and i.p. administration of 107 CSFE-labeled Thy 1.2+Foxp3+ Tregs resulted in significant recovery of these cells from the colon of mice 11 days later (B). C, Purified Th1.2+Foxp3+-expressing cells with (+/+) or without (−/−) Fas (107) were adoptively transferred into Thy1.1+ synergic recipient C57BL/6 mice by i.p. injection and 3 days later (a time determined experimentally to allow for T cell homing; see B) mice received either EtOH or 5 mg of DNBS i.r. as described in Materials and Methods. Mice were sacrificed 8 days post-DNBS treatment and LPLs were isolated and analyzed by flow cytometry for CD4, Foxp3, and Thy1.2 (the later indicating donor cells). B6-Fas−/− Thy1.2+ but not wild-type Thy1.2+ Fas+/+ Tregs transferred into Thy1.1 synergic recipients were protected from depletion in the lamina propria. D, The effect of the adoptive transfer in terms of the absolute number of cells recovered from the lamina propria of the treated mice (n = 4 mice/group).

FIGURE 8.

Fas-bearing Foxp3+ Tregs are subject to depletion following DNBS colitis. A and B, In pilot studies, purification protocols for Tregs revealed 87.5% purity as defined by Foxp3 staining (A) and i.p. administration of 107 CSFE-labeled Thy 1.2+Foxp3+ Tregs resulted in significant recovery of these cells from the colon of mice 11 days later (B). C, Purified Th1.2+Foxp3+-expressing cells with (+/+) or without (−/−) Fas (107) were adoptively transferred into Thy1.1+ synergic recipient C57BL/6 mice by i.p. injection and 3 days later (a time determined experimentally to allow for T cell homing; see B) mice received either EtOH or 5 mg of DNBS i.r. as described in Materials and Methods. Mice were sacrificed 8 days post-DNBS treatment and LPLs were isolated and analyzed by flow cytometry for CD4, Foxp3, and Thy1.2 (the later indicating donor cells). B6-Fas−/− Thy1.2+ but not wild-type Thy1.2+ Fas+/+ Tregs transferred into Thy1.1 synergic recipients were protected from depletion in the lamina propria. D, The effect of the adoptive transfer in terms of the absolute number of cells recovered from the lamina propria of the treated mice (n = 4 mice/group).

Close modal

Although T and B cells are important components for maintaining an individual’s overall health status, they also participate in the development of autoimmune disease and immunopathology. It has become apparent that Foxp3+ T cells are critically involved in the suppression of immune responses, as both mice and humans lacking Foxp3 develop severe multiorgan pathology and a potentially fatal autoimmune syndrome (20). Experimental depletion of Tregs allows for activation of self-reactive effector T cells and the development of autoimmune disease (21). Thus, Foxp3+ T cells are considered beneficial and numerous studies have revealed their ability to reduce inflammation in the gut (10, 22) and joints (23). However, although seminal observations have been made on the factors/conditions that control Foxp3+ Treg development (17) and function in suppressing immune responses (7, 22, 24), there is a lack of data on the fate of Tregs during inflammation. In light of studies demonstrating that adoptive transfer of Tregs reduces the severity of colitis in mice, our goal was to assess the fate of colonic Tregs in mice with hapten-induced colitis.

Intrarectal administration of DNBS results in a self-limiting colitis with tissue damage peaking ∼3 days posttreatment and resolving thereafter (11, 25). BALB/c and C57BL/6 mice treated with DNBS (or oxazolone) show transient depletion of colonic Foxp3+ Tregs 8 days posttreatment, a time point when tissue restitution from the inflammatory insult has begun. Highlighting the fact that this reduction was not simply a consequence of increased inflammatory cell infiltration, reduced absolute numbers of Foxp3+ cells were observed by immunohistochemistry and flow cytometry on gated LPLs. The immunological ramifications of this transient loss of enteric Foxp3+ Tregs are not clear. Transient loss of Foxp3+ Tregs following epithelial restitution would remove the suppressive influence over local effector cells, allowing them to mount a response to any residual bacteria in the tissue (26). It is also feasible that the absence of Tregs could leave the host vulnerable to immunopathological events should there be a concomitant bacterial insult (27). For instance, CD4+CD25high blood cells and LPLs from Helicobacter pylori infected-patients have significantly increased Foxp3 expression (28), limiting the development of gastritis while preventing bacterial clearance (27). Similarly, infection with Leishmania major is accompanied by increased Tregs (29), with enhanced parasite clearance occurring after Treg depletion. In addition, this co-opting of regulatory cells by pathogens not only suppresses the adaptive immune system but also the components of innate immunity (30, 31).

At least three possible explanations exist for the depletion of Treg from the colon of mice with colitis: 1) inflammatory cytokines affect TGF-β signaling required for Foxp3+ Tregs; 2) reduced availability of systemic Tregs; and, 3) enhanced apoptosis of colonic Tregs. We addressed each of these possibilities.

Following DNBS treatment, a rise in local and systemic levels of proinflammatory cytokines occurs, including those of IFN-γ and others (11, 32). However, and consistent with other reports, exposure to IFN-γ, TNF-α, or IL-4 did not reduce the expression of Foxp3 in anti-CD3 (with or without anti-CD28) plus TGF-β-treated CD4+ cells in vitro. Given the ability of IFN-γ to inhibit TGF-β-elicited signal transduction (33) and the critical importance of TGF-β to the induction and survival of Tregs (7, 9, 17), we reasoned that a local inflammatory response may decrease peripheral Treg survival. Additionally, as colonic LPLs from IBD patients have increased expression of SMAD7 (34), an inhibitor of TGF-β-SMAD signaling, the contribution of SMAD7 to Treg depletion was assessed. DNBS colitis enhanced SMAD7 expression in the colon, MLN, and spleen but, critically, not in CD4+ T cells. This finding demonstrates that inflammation-induced SMAD7 expression cannot account for the loss of Tregs, because the number of Foxp3+ cells in the MLN and spleen 8 days post-DNBS treatment were similar to that of controls. Furthermore, flow cytometry revealed increased basal SMAD2 phosphorylation in CD4+ LPLs and MLNs of DNBS-treated mice. Both the increase in SMAD7 in non-CD4+ T cells and phospho-SMAD2 in T cells are suggestive of increased bioactive TGF-β during the restitution process. As evidence of this, it has been shown that patients with IBD have paradoxically increased TGF-β expression (35) and increased SMAD7 expression (36). Although increased SMAD7 expression occurs in humans, this increase is not restricted to T cells in mice (37). Highlighting this, mice with disruption of TGF-β signaling in epithelial cells have an increased susceptibility to mucosal damage and delayed repair (38). Together, these and our data suggest that although altered TGF-β signaling is a consequence of enteric inflammation, it is unlikely to play a substantial role in the depletion of colonic Foxp3+ Tregs.

Natural turnover of Tregs in conjunction with reduced recruitment from the circulation could contribute to a local loss of Tregs. As reduced CD4+Foxp3+ cells were not detected in the spleen and MLN in DNBS-treated mice, peripheral pools of Tregs existed to replenish the colon. However, peripheral blood CD4+Foxp3+ cells were significantly reduced before colonic Treg depletion, adding credence to a reduced recruitment hypothesis. Colonic Tregs did not recover until CD4+Foxp3+α4β7+ T cells were detected at normal levels in the blood 14 days post-DNBS treatment. Peripheral blood lymphocyte depletion has been described in colitis, which may be a result of anemia due to bleeding and/or damage to the thymus by recirculating T cells (39). These data support the contention that reduced Treg recruitment to the colon by the mucosal homing integrin α4β7 could contribute to reduced colonic Treg numbers; however, they do not satisfactorily explain the Treg depletion at 8 days post-DNBS treatment.

Regulatory T cell homeostasis is mediated by FasL, as deficiency in FasL increases the number of Foxp3+ Tregs (40). In addition, human peripheral blood Tregs are sensitive to FasL-mediated killing (18), although similar findings have not been reported for mice. T cell homeostasis following inflammation can occur through the ligation of Fas on the T cell surface with FasL expressed by activated T cells (41), or other cells, including dendritic cells (42), intraepithelial leukocytes (43), and neurones (44), which can also increase their FasL expression. DNBS-induced colitis results in significantly increased expression of FasL on lamina propria mononuclear cells, the colonic epithelium, and smooth muscle cells. However, and notably, FasL was not increased in the ileum, spleen, or MLNs of DNBS-treated mice, and these tissues had normal numbers of CD4+Foxp3+ cells. The correlation between increased FasL expression and loss of Foxp3+ Tregs suggested that apoptosis could account for Treg depletion, a hypothesis supported by in vitro studies demonstrating that murine CD4+Foxp3+ but not CD4+Foxp3 cells are sensitive to Fas-mediated cell death. The higher basal rate of apoptosis and the increased susceptibility of murine CD4+Foxp3+ cells to apoptotic stimuli is intriguing and is in stark contrast to CD4+Foxp3 cells from the same animals. Although this would be expected to impinge on immunoregulation, the reason for this is unclear, but we speculate that it may be due to decreased anti-apoptotic gene expression in Tregs as has been reported in human Foxp3+ cells (45, 46).

A role for Fas/FasL in Treg depletion following DNBS-induced colitis was further confirmed in vivo through use of mice deficient for both Fas and FasL (lpr/lpr gld/gld). Complementing previous reports (40), these mice possessed increased CD4+Foxp3+ Tregs compared with wild-type mice. In support of our hypothesis, we found no depletion of Foxp3+ Tregs in the colonic LPL population in Fas/FasL-deficient mice 8 days post-DNBS treatment. Although these data are highly suggestive of the possibility that Fas/FasL regulates the number of Tregs in the inflamed gut, it is important to note that soluble FasL is a chemoattractant for polymorphonuclear cells (47, 48), primary components of the inflammatory infiltrate in DNBS colitis and, hence, the disease in the Fas/FasL−/− mice would be qualitatively different from that in normal mice. In light of this, we performed adoptive transfers of Thy1.2+B6-Fas+/+ or Thy1.2+B6-Fas−/− Tregs into Thy1.1+ recipients to determine whether lack of Treg Fas expression prevented the depletion of CD4+Foxp3+ Tregs from the colon. Survival of donor Thy1.2+B6-Fas−/− but not Thy1.2+Fas+/+ Tregs in Thy1.1+ recipients demonstrated that the depletion observed in DNBS colitis is due to Fas/FasL-mediated cell death.

Although the loss of Foxp3 expression and conversion into another T cell type was conceivable, the adoptive transfer experiments argue against this possibility. Although the majority of reports describe the generation of Th17 cells from naive T cells in the presence of TGF-β and IL-6 (49), the conversion of Tregs in the presence of IL-6 has been recently described in vitro (50). Despite increased IL-6 expression occurring during hapten induced colitis (51), depletion of colonic Tregs is not likely due to conversion. Adoptive transfer of Thy1.2+ Tregs into Thy1.1+ recipients allows one to distinguish host from recipient cells. If depletion of colonic Tregs was due to loss of Foxp3 expression and/or conversion of the Thy1.2+CD4+Foxp3+ into a Th17 phenotype, this should have been accompanied by an increase in Thy1.2+CD4+Foxp3 cells. The fact that Thy1.2+CD4+Foxp3 cells were not increased in recipients of Thy1.2+ Tregs (either Fas−/− or Fas+/+) following colitis indicates that that conversion of Thy1.2+CD4+Foxp3+ to another phenotype is not responsible for the depletion of Tregs from the colon.

This study demonstrates that colonic inflammation in mice elicits a profound but localized and temporary depletion of CD4+Foxp3+ Tregs. We demonstrate for the first time that the loss of this immunoregulatory population occurs by a Fas-dependent mechanism. Repopulation of the colon with CD4+Foxp3+ Tregs occurs after the FasL signal subsides, by mechanism(s) that may involve α4β7-based homing and recruitment to the gut. Clearly, this system is dynamic and, as with all immune responses, the key to intestinal health and homeostasis is balance. Temporary removal of Tregs may allow robust immune responses to occur following localized insult and injury. However, prolonged depletion could leave the host vulnerable to unchecked aggressive immune responses.

The authors have no financial conflict of interest.

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 work was supported Natural Sciences and Engineering Research Council of Canada Grant 341924. D.M.M. is supported by a Canada Research Chair (Tier 1) and an Alberta Heritage Foundation for Medical Research Scientist Award. C.R. was a recipient of a Natural Sciences and Engineering Research Council of Canada studentship.

3

Abbreviations used in this paper: Treg, regulatory T cell; CMF, Ca2+- and Mg2+-free; DNBS, 2,4-dinitrobenzene sulfonic acid; EtOH, ethanol; FasL, Fas ligand; IBD, inflammatory bowel disease; i.r., intrarectal; LPL, lamina propria lymphocyte; MLN, mesenteric lymph node.

1
Rakoff-Nahoum, S., J. Paglino, F. Eslami-Varzaneh, S. Edberg, R. Medzhitov.
2004
. Recognition of commensal microflora by Toll-like receptors is required for intestinal homeostasis.
Cell
118
:
229
-241.
2
Stappenbeck, T. S., L. V. Hooper, J. I. Gordon.
2002
. Developmental regulation of intestinal angiogenesis by indigenous microbes via Paneth cells.
Proc. Natl. Acad. Sci USA
99
:
15451
-15455.
3
Rhee, K. J., P. Sethupathi, A. Driks, D. K. Lanning, K. L. Knight.
2004
. Role of commensal bacteria in development of gut-associated lymphoid tissues and pre-immune antibody repertoire.
J. Immunol.
172
:
1118
-1124.
4
Fontenot, J. D., M. A. Gavin, A. Y. Rudensky.
2003
. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells.
Nat. Immunol.
4
:
330
-336.
5
Hori, S., T. Nomura, S. Sakaguchi.
2003
. Control of regulatory T cell development by the transcription factor Foxp3.
Science
299
:
1057
-1061.
6
Watanabe, N., Y. H. Wang, H. K. Lee, T. Ito, Y. H. Wang, W. Cao, Y. J. Liu.
2005
. Hassall’s corpuscles instruct dendritic cells to induce CD4+CD25+ regulatory T cells in human thymus.
Nature
436
:
1181
-1185.
7
Fantini, M. C., C. Becker, G. Monteleone, F. Pallone, P. R. Galle, M. F. Neurath.
2004
. Cutting edge: TGF-β induces a regulatory phenotype in CD4+CD25 T cells through Foxp3 induction and down-regulation of Smad7.
J. Immunol.
172
:
5149
-5153.
8
Denning, T. L., G. Kim, M. Kronenberg.
2005
. Cutting edge: CD4+CD25+ regulatory T cells impaired for intestinal homing can prevent colitis.
J. Immunol.
174
:
7487
-7491.
9
Peng, Y., Y. Laouar, M. O. Li, E. A. Green, R. A. Flavell.
2004
. TGF-β regulates in vivo expansion of Foxp3-expressing CD4+CD25+ regulatory T cells responsible for protection against diabetes.
Proc. Natl. Acad. Sci. USA
101
:
4572
-4577.
10
Uhlig, H. H., J. Coombes, C. Mottet, A. Izcue, C. Thompson, A. Fanger, A. Tannapfel, J. D. Fontenot, F. Ramsdell, F. Powrie.
2006
. Characterization of Foxp3+CD4+CD25+ and IL-10-secreting CD4+CD25+ T cells during cure of colitis.
J. Immunol.
177
:
5852
-5860.
11
Hunter, M. M., A. Wang, C. L. Hirota, D. M. McKay.
2005
. Neutralizing anti-IL-10 antibody blocks the protective effect of tapeworm infection in a murine model of chemically-induced colitis.
J. Immunol.
174
:
7368
-7375.
12
Lefrançois, L., N. Lycke.
2003
. Basic protocol 3: isolation of lymphocytes from intestinal LP.
Current Protocols in Immunology
3.19.6
-3.19.15. John Wiley & Sons, Inc., Hoboken, NJ.
13
Wise, J. T., T. J. Baginski, J. L. Mobley.
1999
. An adoptive transfer model of allergic lung inflammation in mice is mediated by CD4+CD62LlowCD25+ T cells.
J. Immunol.
162
:
5592
-5600.
14
Szanya, V., J. Ermann, C. Taylor, C. Holness, C. G. Fathman.
2002
. The subpopulation of CD4+CD25+ splenocytes that delays adoptive transfer of diabetes expresses L-selectin and high levels of CCR7.
J. Immunol.
169
:
2461
-2465.
15
Reardon, C., D. M. McKay.
2007
. TGF-β suppresses IFN-γ-STAT1-dependent gene transcription by enhancing STAT1-PIAS1 interactions in epithelia but not monocytes/macrophages.
J. Immunol.
178
:
4284
-4295.
16
Poccia, F., B. Cipriani, S. Vendetti, V. Colizzi, Y. Poquet, L. Battistini, M. Lopez-Botet, J.-J. Fournie, M.-L. Cougeonn.
1997
. CD94/NKC2 inhibitory receptor complex modulates both anti-viral and anti-tumoral responses of polyclonal phosphoantigen-reactive Vy9VS2 T lymphocytes.
J. Immunol.
159
:
6009
-6017.
17
Marie, J. C., J. J. Letterio, M. Gavin, A. Y. Rudensky.
2005
. TGF-β1 maintains suppressor function and Foxp3 expression in CD4+CD25+ regulatory T cells.
J. Exp. Med.
201
:
1061
-1067.
18
Fritzsching, B., N. Oberle, N. Eberhardt, S. Quick, J. Haas, B. Wildemann, P. H. Krammer, E. Suri-Payer.
2005
. In contrast to effector T cells, CD4+CD25+FoxP3+ regulatory T cells are highly susceptible to CD95 ligand- but not to TCR-mediated cell death.
J. Immunol.
175
:
32
-36.
19
Cohen, P. L., R. A. Eisenberg.
1991
. Lpr and gld: single gene models of systemic autoimmunity and lymphoproliferative disease.
Annu. Rev. Immunol.
9
:
243
-269.
20
Schubert, L. A., E. Jeffery, Y. Zhang, F. Ramsdell, S. F. Ziegler.
2001
. Scurfin (FOXP3) acts as a repressor of transcription and regulates T cell activation.
J. Biol. Chem.
276
:
37672
-37679.
21
Kim, J. M., J. P. Rasmussen, A. Y. Rudensky.
2007
. Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice.
Nat. Immunol.
8
:
191
-197.
22
Fantini, M. C., C. Becker, I. Tubbe, A. Nikolaev, H. A. Lehr, P. Galle, M. F. Neurath.
2006
. Transforming growth factor-β induced FoxP3+ regulatory T cells suppress Th1 mediated experimental colitis.
Gut
55
:
671
-680.
23
Nadkarni, S., C. Mauri, M. R. Ehrenstein.
2007
. Anti-TNF-α therapy induces a distinct regulatory T cell population in patients with rheumatoid arthritis via TGF-β.
J. Exp. Med.
204
:
33
-39.
24
Williams, L. M., A. Y. Rudensky.
2007
. Maintenance of the Foxp3-dependent developmental program in mature regulatory T cells requires continued expression of Foxp3.
Nat. Immunol.
8
:
277
-284.
25
Sanovic, S., D. P. Lamb, M. G. Blennerhassett.
1999
. Damage to the enteric nervous system in experimental colitis.
Am. J. Pathol.
155
:
1051
-1057.
26
Chen, X., M. Baumel, D. N. Mannel, O. M. Z. Howard, J. J. Oppenheim.
2007
. Interaction of TNF with TNF receptor type 2 promotes expansion and function of mouse CD4+CD25+ T regulatory cells.
J. Immunol.
179
:
154
-161.
27
Rad, R., L. Brenner, S. Bauer, S. Schwendy, L. Layland, C. P. da Costa, W. Reindl, A. Dossumbekova, M. Friedrich, D. Saur, et al
2006
. CD25+Foxp3+ T cells regulate gastric inflammation and Helicobacter pylori colonization in vivo.
Gastroenterology
131
:
525
-537.
28
Lundgren, A., E. Stromberg, A. Sjoling, C. Lindholm, K. Enarsson, A. Edebo, E. Johnsson, E. Suri-Payer, P. Larsson, A. Rudin, et al
2005
. Mucosal FOXP3-expressing CD4+CD25high regulatory T cells in Helicobacter pylori-infected patients.
Infect. Immun.
73
:
523
-531.
29
Suffia, I. J., S. K. Reckling, C. A. Piccirillo, R. S. Goldszmid, Y. Belkaid.
2006
. Infected site-restricted Foxp3+ natural regulatory T cells are specific for microbial antigens.
J. Exp. Med.
203
:
777
-788.
30
Maloy, K. J., L. Salaun, R. Cahill, G. Dougan, N. J. Saunders, F. Powrie.
2003
. CD4+CD25+ T(R) cells suppress innate immune pathology through cytokine-dependent mechanisms.
J. Exp. Med.
197
:
111
-119.
31
De Luca, A., C. Montagnoli, T. Zelante, P. Bonifazi, S. Bozza, S. Moretti, C. D'Angelo, C. Vacca, L. Boon, F. Bistoni, et al
2007
. Functional yet balanced reactivity to Candida albicans requires TRIF, MyD88, and IDO-dependent inhibition of Rorc.
J. Immunol.
179
:
5999
-6008.
32
Khan, W. I., P. A. Blennerhasset, A. K. Varghese, S. K. Chowdhury, P. Omsted, Y. Deng, S. M. Collins.
2002
. Intestinal nematode infection ameliorates experimental colitis in mice.
Infect. Immun.
70
:
5931
-5937.
33
Ulloa, L., J. Doody, J. Massague.
1999
. Inhibition of transforming growth factor-β/SMAD signalling by the interferon-γ/STAT pathway.
Nature
397
:
710
-713.
34
Monteleone, G., A. Kumberova, N. M. Croft, C. McKenzie, H. W. Steer, T. T. MacDonald.
2001
. Blocking Smad7 restores TGF-β1 signaling in chronic inflammatory bowel disease.
J. Clin. Invest.
108
:
601
-609.
35
Babyatsky, M. W., G. Rossiter, D. K. Podolsky.
1996
. Expression of transforming growth factors α and β in colonic mucosa in inflammatory bowel disease.
Gastroenterology
110
:
975
-984.
36
Monteleone, G., F. Pallone, T. T. MacDonald.
2004
. Smad7 in TGF-β-mediated negative regulation of gut inflammation.
Trends Immunol.
25
:
513
-517.
37
Boirivant, M., F. Pallone, C. Di Giacinto, D. Fina, I. Monteleone, M. Marinaro, R. Caruso, A. Colantoni, G. Palmieri, M. Sanchez, et al
2006
. Inhibition of Smad7 with a specific antisense oligonucleotide facilitates TGF-β1-mediated suppression of colitis.
Gastroenterology
131
:
1786
-1798.
38
Beck, P. L., I. M. Rosenberg, R. J. Xavier, T. Koh, J. F. Wong, D. K. Podolsky.
2003
. Transforming growth factor-β mediates intestinal healing and susceptibility to injury in vitro and in vivo through epithelial cells.
Am. J. Pathol.
162
:
597
-608.
39
Faubion, W. A., Y. P. De Jong, A. A. Molina, H. Ji, K. Clarke, B. Wang, E. Mizoguchi, S. J. Simpson, A. K. Bhan, C. Terhorst.
2004
. Colitis is associated with thymic destruction attenuating CD4+25+ regulatory T cells in the periphery.
Gastroenterology
126
:
1759
-1770.
40
Mohamood, A. S., C. J. Trujillo, D. Zheng, C. Jie, F. M. Murillo, J. P. Schneck, A. R. Hamad.
2006
. Gld mutation of Fas ligand increases the frequency and up-regulates cell survival genes in CD25+CD4+ TR cells.
Int. Immunol.
18
:
1265
-1277.
41
Ju, S. T., D. J. Panka, H. Cui, R. Ettinger, M. EI-Khatib, D. H. Sherr, B. Z. Stanger, A. Marshak-Rothstein.
1995
. Fas(CD95)/FasL interactions required for programmed cell death after T-cell activation.
Nature
373
:
444
-448.
42
Zhang, H. G., M. Fleck, E. R. Kern, D. Liu, Y. Wang, H. C. Hsu, P. Yang, Z. Wang, D. T. Curiel, T. Zhou, J. D. Mountz.
2000
. Antigen presenting cells expressing Fas ligand down-modulate chronic inflammatory disease in Fas ligand-deficient mice.
J. Clin. Invest.
105
:
813
-821.
43
Lin, T., T. Brunner, B. Tietz, J. Madsen, E. Bonfoco, M. Reaves, M. Huflejt, D. R. Green.
1998
. Fas ligand- mediated killing by intestinal intraepithelial lymphocytes: participation in intestinal graft-versus-host disease.
J. Clin. Invest.
101
:
570
-577.
44
Sayani, F. A., C. M. Keenan, M. D. Van Sickle, K. R. Amundson, E. J. Parr, R. D. Mathison, W. K. MacNaughton, J. E. Braun, K. A. Sharkey.
2004
. The expression and role of Fas ligand in intestinal inflammation.
Neurogastroenterol. Motil.
16
:
61
-74.
45
Taams, L. S., J. Smith, M. H. Rustin, M. Salmon, L. W. Poulter, A. N. Akbar.
2001
. Human anergic/suppressive CD4+CD25+ T cells: a highly differentiated and apoptosis-prone population.
Eur. J. Immunol.
31
:
1122
-1131.
46
Vukmanovic-Stejic, M., Y. Zhang, J. E. Cook, J. M. Fletcher, A. McQuaid, J. E. Masters, M. H. Rustin, L. S. Taams, P. C. Beverley, D. C. Macallan, A. N. Akbar.
2006
. Human CD4+ CD25hi Foxp3+ regulatory T cells are derived by rapid turnover of memory populations in vivo.
J. Clin. Invest.
116
:
2423
-2433.
47
Seino, K. I., K. Iwabuchi, N. Kayagaki, R. Miyata, I. Nagaoka, A. Matsuzawa, K. Fukao, H. Yagita, K. Okumura.
1998
. Cutting edge: chemotactic activity of soluble Fas ligand against phagocytes.
J. Immunol.
161
:
4484
-4488.
48
Ottonello, L., G. Tortolina, M. Amelotti, F. Dallegri.
1999
. Soluble Fas ligand is chemotactic for human neutrophilic polymorphonuclear leukocytes.
J. Immunol.
162
:
3601
-3606.
49
Bettelli, E., Y. Carrier, W. Gao, T. Korn, T. B. Strom, M. Oukka, H. L. Weiner, V. K. Kuchroo.
2006
. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells.
Nature
441
:
235
-238.
50
Xu, L., A. Kitani, I. Fuss, W. Strober.
2007
. Cutting edge: regulatory T cells induce CD4+CD25Foxp3 T cells or are self-induced to become Th17 cells in the absence of exogenous TGF-β.
J. Immunol.
178
:
6725
-6729.
51
Atreya, R., J. Mudter, S. Finotto, J. Mullberg, T. Jostock, S. Wirtz, M. Schutz, B. Bartsch, M. Holtmann, C. Becker, et al
2000
. Blockade of interleukin 6 trans signaling suppresses T-cell resistance against apoptosis in chronic intestinal inflammation: evidence in Crohn’s disease and experimental colitis in vivo.
Nat. Med.
6
:
583
-588.