NKT Cells Are Critical for the Initiation of an Inflammatory Bowel Response against Toxoplasma gondii¹

Natural killer T cells represent a minor subset of T lymphocytes that share receptor structures with conventional T cells and NK cells (1, 2). Murine NKT cells express intermediate levels of a TCR using a semi-invariant Vα14-Jα281 TCR α-chain paired with Vβ8, -7, or -2 TCR β-chain together with NK cell receptors (NKR-P1, Ly-49, and NK1.1 in C57BL/6 mice) (3, 4). These cells are located mainly in the liver, spleen, thymus, and bone marrow and recognize Ag in the context of the monomorphic CD1d Ag-presenting molecule (5, 6). CD1d and the invariant TCR α-chain are essential for the normal development of NKT cells (7). CD1 molecules present hydrophobic lipid Ags (8), and the marine sponge derived glycolipid, commonly referred to as α-galactosylceramide (α-GalCer),⁴ was identified as a potent stimulatory factor for NKT cells (9).

A potential role of NKT cells in the regulation of immune responses has been hypothesized because of their capacity to rapidly release large amounts of IL-4 and IFN-γ upon activation (10). NKT cells play crucial roles in various immune responses, including antitumor, autoimmune, and antimicrobial immune responses (1, 11). Within hours of TCR engagement, CD1d-reactive T cells produce Th1 and/or Th2 cytokines (9, 11, 12) by a mechanism not yet identified that can influence other immune cells, such as conventional T (13, 14, 15), NK cells (16), and dendritic cells (DC) (17). NKT cell-derived Th1 cytokines (such as IFN-γ) are important in the initiation of the antitumor immune response, whereas NKT cell-derived Th2 cytokines (IL-4 and IL-10) are involved in down-regulation of the autoimmune response (18). When stimulated with α-GalCer, NKT cells exhibit the ability to proliferate and to produce both Th1 and Th2 cytokines (9, 19). However administration of α-GalCer at the time of priming of mice with Ag results in the generation of only Ag-specific Th2 cells. Thus, α-GalCer might be useful for modulating the immune response toward a Th2 phenotype (12).

Recent evidence suggests that NKT cells are important in the host/pathogen immune response, including cytotoxicity, Ab production, and regulation of Th1/Th2 differentiation. NKT cells have been shown to participate in the immune response to a range of different infectious agents, including Listeria, Mycobacteria, Salmonella, Plasmodium, viral hepatitis (20, 21), HIV (22), and even Toxoplasma gondii (23). T. gondii is an obligate intracellular parasite acquired by oral ingestion of tissue cysts containing either bradyzoites or sporozoites from contaminated soil. It has been observed that after oral infection with tissue cysts, the intestinal epithelial and lamina propria cells are invaded by the parasites. Parasite infection induces a strongly biased Th1 response in the gut that displays a dual effect. IFN-γ produced by the CD4 T cells from the lamina propria (24) limits parasite replication, conferring resistance in mice in certain inbred strains. However, in C57BL/6 (B6) mice, an overwhelming IFN-γ production leads to a lethal acute ileitis within 10 days after oral infection. This Toxoplasma-induced intestinal disease shares histological and immunological similarities with human inflammatory bowel disease, such as Crohn’s disease. The regulation of this inflammatory process requires a delicate homeostatic balance that is influenced by either a Th1 or Th2 response.

In this report the role of NKT cells in the initiation of the inflammatory process in response to oral infection with T. gondii was evaluated. Our findings suggest a potentially critical role for these early responder cells in the initiation and regulation of the lethal inflammatory process.

Female, 8- to 10-wk-old, inbred B6 mice and CBA were obtained from IFFA-Credo. Mice were housed under approved conditions of the Animal Research Facility at Institut Pasteur. IL-10^−/− mice were supplied by Dr. Bandeira (Institut Pasteur, Paris, France). We were provided with Jα281^−/− mice by Dr. M. Taniguchi (Riken Research Center for Allergy and Immunology, Yokohama, Japan) (9), Vα14Tg mice by Dr. A. Lehuen (Institut National de la Santé et de la Recherche Médicale, Paris, France) (25), actin-GFP mice by Dr. M. Okabe (Genome Information Research Center, Osaka University, Osaka, Japan) (26), and CD1^−/− mice by Dr. L. Van Kaer (Vanderbilt University School of Medicine, Nashville, TN) (7). All the genetically modified strains were on a B6 genetic background. 76K strain cysts isolated from the brains of chronically infected CBA mice were used for in vivo studies. Mice were infected orally by intragastric gavage of 35 cysts, a lethal condition for B6 wild-type mice as described previously (27). After infection, mortality was evaluated, and morbidity was estimated by the percentage of weight loss compared with the initial weight.

α-GalCer was kept dissolved in PBS buffer containing 20% DMSO at 220 μg/ml as a stock solution. Mice received a single i.p. injection of 5 mg of α-GalCer the day before infection by T. gondii. Control mice received an i.p. injection of PBS/20% DMSO, which has no influence on the course of T. gondii infection.

Neutralization of IL-4 was conducted by injecting i.p. 1 mg of anti-IL-4 (11B11; provided by Dr. P. Launois, World Health Organization Immunology Research and Training Center, Institute of Biochemistry, Epalinges, Switzerland) mAb 24 h before α-GalCer treatment and 48 h before infection. Control mice were treated with rat IgG Abs (Sigma-Aldrich).

Mice were depleted of CD25⁺ cells by i.p. administration of 0.5 mg of anti-CD25 (PC61; provided by Dr. R. J. Noelle, Dartmouth Medical School, Lebanon, NH) mAb. Three days after the treatment, the efficiency of CD25⁺ cell depletion was controlled in peripheral blood by FACS analysis. The CD25⁺ cell depletion remained stable over 15 days. Control mice were treated with a mouse IgD1 isotype Ab (MOPC31C k; BD Pharmingen).

The method used to isolate intestinal lamina propria lymphocytes (LPLs) was modified as described previously (24). After dissection and removal of Peyer’s patches, the sectioned intestines were incubated in PBS-3 mM EDTA at 37°C and 5% CO₂ (four times, 20 min each time). Then intestinal pieces were incubated at 37°C in RPMI 1640-5% FCS with Liberase (0.14 Wunch units/ml; Roche) and DNase (10 U/ml; Sigma-Aldrich). After 45 min, the digested suspension containing LPLs was filtered on a cell strainer and washed twice, and the pellet was submitted to a Percoll gradient to isolate the lymphocytes. Total cells were resuspended in a 80% isotonic Percoll solution (Pharmacia Biotech) and overlaid with a 40% isotonic Percoll solution. Centrifugation for 30 min at 3000 rpm resulted in concentration of mononuclear cells at the 40–80% interface. The collected cells were washed once with PBS supplemented with 2% FCS. The purity of the LPL population was assessed by the relative percentage of B cells (>50%), CD4 T cells (∼20%), CD8 T (<3%) cells, and enterocytes (<5%).

IELs were isolated as previously described (28). Briefly, the small intestine was flushed with PBS and divided longitudinally after removal of Peyers’s patches. The mucosae were scraped, dissociated by mechanical disruption, in RPMI 1640 containing 4% FCS and 1 mmol/L DTT. After passage over a glass-wool column, the lymphocytes were separated by Percoll as described for LPLs. The purity of IEL population was assessed by the relative percentages of B cells (<2%), CD4 T cells (<10%), CD8 T cells (>80%), and enterocytes (<5%).

MLN and spleen were dissociated and freed of connective tissue by filtration (70 μm). Unless otherwise stated, each mouse was analyzed individually.

Single-cell suspensions were obtained from liver as described previously by us (29).

Single-cell suspensions were first incubated 10 min with an anti-FcγRII/III mAb (Fcblock, 2.4G2; BD Pharmingen), followed by a 1-h exposure to CD1d/α-GalCer tetramer-allophycocyanin under agitation at 4°C. CD1d/α-GalCer tetramers were prepared as described by Matsuda et al. (30). After two washes, other cell surface stainings were performed with the following Abs: anti-TCRβ (H57-597), anti-CD4 (RM4-5), anti-CD8 (53-6.7), anti-NK1.1 (PK136), anti-CD25 (C363 16A), anti-CD45RB (7D4), and anti-CD5 (BD Biosciences). PerCP-streptavidin and CyChrome-streptavidin were purchased from BD Biosciences. Cells were analyzed in PBS containing 2% FCS using a FACSCalibur flow cytometer and CellQuest software (BD Biosciences).

NKT cells stained with the tetramer were magnetically sorted. After tetramer CD1d/α-GalCer-allophycocyanin staining, cells suspensions were incubated for 15 min in PBS/2% FCS/2 mM EDTA at 4°C with anti-allophycocyanin beads as described by the provider (Miltenyi Biotec). After washing and filtration, samples were run on AutoMACS (Miltenyi Biotec). Purity was controlled by cytometric analysis, and the sorted cells were frozen until molecular biology analysis.

For the reconstitution experiment, NKT from the liver and the spleen of actin-GFP mice were sorted with both anti-CD5 biotin (53-7.3), and anti-NK1.1-PE (PK136) mAbs and streptavidin-allophycocyanin using a MoFlo (DakoCytomation). Purified NKT-GFP⁺ cells were collected in RPMI 1640 supplemented with 10% FCS. The purity of the sorted NKT-GFP cells was found to exceed 97%.

Highly purified NKT cells (1 × 10⁶) were injected i.v. into Jα281^−/− mice. At the same time these mice were treated with 5 μg of α-GalCer i.p. One day later, NKT cells were transferred, and α-GalCer-treated mice were infected.

Intestines were immediately fixed in buffered 10% formalin after dissection. Then they were embedded, sectioned, and stained with H&E for histological examination. Inflammation was scored by the ratio of the length/thickness of the villi (mean of 20 measures for a total of four different fields).

Intestinal and hepatic samples from NKT-GFP-transferred mice were microscopically examined. On day 7 after infection mice were scarified, and samples from intestines and livers were incubated for 24 h in paraformaldehyde (4%) and saccharose (30%). Then tissues were frozen in liquid nitrogen using OCT embedding compound (Sakura). Frozen sections (10 μm) were cut on a microtome HM 505 cryostat (Microcom Laboratory), fixed with PBS/paraformaldehyde (4%), permeabilized by PBS/Triton (0.1%), contrastained with rhodamine phallodin (Molecular Probes), and mounted with Vectashield (Vector Laboratories). Preparations were analyzed with fluorescent microscope Axioplan 2 imaging coupled with an ApoTome system (Zeiss). GFP-NKT cell trafficking was also assessed by FACS analysis performed on day 7 after infection with cell suspensions obtained from lamina propria and livers.

Recipient mice were lethally irradiated (900 rad) with a ¹³⁷Ce source. Then they received i.v. bone marrow cells (1 × 10⁷) recovered from femurs and tibias of donor mice. To generate mice with only NKT cells devoid of the IL-10 gene, a mix (50/50%) of bone marrow cells from Jα281^−/− mice and IL-10^−/− mice was prepared. Control mice received cells from B6, Jα281^−/− or IL-10^−/− mice alone. Six weeks after reconstitution, mice were bled, and the presence of CD4⁺, CD19⁺ (1D3), and CD11c⁺ (HL3) cells was monitored by flow cytometric analysis. Reconstitution with NKT was assessed (two mice per group) by staining the CD1d/α-GalCer-allophycocyanin tetramer cell suspensions obtained from the liver and lamina propria of the chimera. Chimeric mice were then infected. At different times after infection, LPLs and MLN cell suspensions were phenotyped by FACS analysis. Morbidity was evaluated daily by recording the weight loss, and mortality was also recorded.

Tissue samples from intestines and purified cells were kept frozen (−70°C) until mRNA extraction. Specimens were disrupted in a Polytron (Brinkmann Instruments) and homogenized in 350 ml of RLT buffer (Qiagen). RNA extraction and cDNA preparation were conducted following standard procedures using oligo(dT)₁₇ primers, and 10 U of avian myeloblastosis virus reverse transcriptase. Quantitative PCR was performed with the GeneAmp 7000 (Applied Biosystems) as indicated by the supplier. Primers and probes for the quantitative PCR assay of cytokines and actin were designed as previously described (31). Foxp3 mRNA were analyzed with applied assay on demand n°Mm00475156_m1 (Applied Biosystems).

DNA was extracted from the different organ samples using a DNAeasy kit (Qiagen). The Toxoplasma B1 gene was amplified by quantitative real-time PCR (32). Parasite titration by real-time PCR was performed with the GeneAmp 7000 (Applied Biosystems). The standard curve established from the serial 10-fold dilutions of T. gondii DNA of parasite concentrations ranging from 1 × 10⁶ to 10, showed linearity over a 6-log concentration range and was included in each amplification run. At different time points after infection, tissue samples were recovered, and their DNA were extracted with the DNeasy Tissue Kit (Qiagen). For each sample, parasite count was calculated by interpolation from the standard curve. The parasite burden was expressed as the number of parasites per milligram of samples. Cerebral parasite burden was evaluated by enumeration of the cysts on day 30 after infection.

Results are expressed as the mean ± SD. Statistical differences between groups were analyzed using Student’s t test. A value of p < 0.05 was considered significant.

The presence of the NKT lymphocyte subpopulation within the gut was demonstrated by FACS analysis using the CD1d/α-GalCer tetramers. In the lamina propria of naive B6 mice, 2% of the mononuclear cells (LPLs) were detected (Fig. 1,A). NKT cells were not detected in cell suspensions from the IEL compartment (Fig. 1,A). Seventy to 80% of the tetramer-positive cells were CD4⁺; the remainder were CD4⁻CD8⁻ double negative. During the days following infection, a decrease in the number of tetramer-positive cells was observed (Fig. 1,B) that could be due to TCR down-regulation. Serial time point phenotyping after infection demonstrated that all NKT cells were CD25⁻. To assess NKT cell trafficking into the intestine after infection, Jα281^−/− mice were transferred with NKT-GFP⁺ cells (1 × 10⁶) highly purified from the livers of GFP transgenic mice on the basis of CD5 and NK1.1 expression (Fig. 1,D, a). At 7 days after infection, GFP⁺ cells were found in cell suspension obtained from the liver (Fig. 1,D, b) and lamina propria (Fig. 1,D, c) of the transferred mice. Histological examination by confocal microscopy revealed that within the liver, NKT-GFP⁺ cells were distributed among hepatocytes near the sinusoids (Fig. 1,E). Within the gut, NKT-GFP⁺ cells were always localized in the lamina propria and were never associated with the IEL compartment (Fig. 1 E). These data indicated that NKT cells traffic to the intestine, where they localize within the lamina propria.

The involvement of NKT cells in the initiation of the intestinal inflammation after oral infection with T. gondii was investigated by comparing the outcome of the infection in wild-type B6 mice and mice genetically deficient in NKT cells (Jα281^−/− mice). As expected, all control B6 mice died within 7–10 days of severe ileitis after oral challenge with 35 cysts (Fig. 2,A). The intestinal inflammation and subsequent morphological changes were characterized by cellular infiltration within the lamina propria; short, thickened villi; and patchy transluminal necrosis. In contrast, Jα281^−/− mice developed a less severe disease (Fig. 2,B) associated with 1) a decrease in the length/thickness ratio of the villi compared with B6 infected mice (Fig. 2,C), 2) a significantly delayed time of death, and 3) a decrease in the mortality rate compared with B6 mice (Fig. 2,A). This outcome was not parasite dose dependent, as determined using a lower infectious dose of cysts (10 cysts/mouse) in which all the Jα281^−/− mice survived, whereas 25% of the B6 died (Fig. 2,D). These results indicate that the absence of NKT cells correlates with a more resistant phenotype. However, CD1d^−/− mice were even more susceptible than B6 mice (Fig. 2 A). In addition to NKT depletion, regulatory cells, such as IEL and B cells, are also reduced in CD1d^−/− mice (33, 34).

To further explore the potential role of NKT cells in the inflammatory process, mice that overexpressed NKT cells (Vα14Tg mice) were infected. Both B6 and Vα14Tg mice died within 7–10 days when infected with 35 cysts (Fig. 2,A). However, in the experiment using a lower dose of cysts (10 cysts/mouse), all the Vα14Tg mice died, whereas only 25% of the B6 mice died (Fig. 2 D). These data confirm that NKT cells are important in the innate host response to oral parasite infection and are involved in disease susceptibility.

IFN-γ is an important cytokine in mediating host defense against T. gondii infection. It limits parasite replication, but, at the same time, if overproduced, it leads to the development of overwhelming intestinal inflammation. Therefore, because NKT cell-deficient mice (Jα281^−/−) were more resistant to the development of lethal ileitis after T. gondii infection, the expression of IFN-γ in their intestines was measured at different times after oral challenge with the cysts. Between days 2 and 3 after infection, IFN-γ mRNA expression peaked in the intestine of B6 mice, and there was a significant difference in IFN-γ mRNA expression between B6 mice and Jα281^−/− mice. By quantitative RT-PCR, the level of mRNA expression in B6 mice was 9–10 times higher than that in Jα281^−/− mice (Fig. 2 E). Over time, inflammatory cytokine production in Jα281^−/− mice may increase, contributing in the delayed time to death due to lethal intestinal inflammation. The lack of early production of IFN-γ might also explain the 2-fold increase in parasite burden in Jα281^−/− mice on day 8 after infection. These findings strongly suggest that NKT cell activation after oral infection with T. gondii is associated with early initiation of the Th1 process observed in the intestines of B6 mice.

Because α-GalCer can influence the nature of the cytokines produced by NKT cells and consequently the orientation of the adaptative Th response, mice were treated with α-GalCer the day before infection. Up to 30 days after infection, this treatment prevented death in both B6 (100%) and V14αTg mice overexpressing NKT cells (80%; Fig. 3,A). Histological examination performed on day 7 after infection revealed that treatment with α-GalCer interfered with the development of ileitis (Fig. 3, B and C). In addition, B6 mice treated with α-GalCer exhibited less weight loss compared with untreated infected controls (Fig. 3,D). To assess the cell population targeted by α-GalCer treatment, NKT-deficient mice (Jα281^−/−) were treated with α-GalCer the day before infection. This treatment had no effect on the infection outcome in Jα281^−/− mice (Fig. 3,A), as attested by the early time of death and the histological damages observed in treated mice (Fig. 3 B). These observations strongly suggest that α-GalCer modulates the functional abilities of NKT cells. Treatment with α-GalCer was not directly toxic to the parasite, because there was no difference in parasite burden in Jα281^−/− mice treated or not treated on day 30 after infection (data not shown). Treatment with α-GalCer 2 days after infection failed to impact the development of the hyperinflammatory response in small intestine.

One of the consequences of α-GalCer treatment was the increase in the number of NKT cells in the lamina propria of infected mice (Fig. 1,C). The production of selected cytokines in the whole intestine of α-GalCer-treated mice was monitored by quantitative RT-PCR. A significant increase in IL-10 (180-fold) and IL-4 (80-fold) mRNA expression was observed in the intestines of α-GalCer-treated mice on days 3 and 5, respectively, after infection. In contrast, no increase in IL-13 mRNA expression in the whole intestine of α-GalCer-treated mice was measured at serial time points after infection. mRNA for IFN-γ was also significantly decreased (10-fold) in α-GalCer-treated mice (data not shown). This result demonstrated a shift in cytokine production toward a Th2-like profile after treatment with α-GalCer and infection and a decline in the Th1-like immune response. To better assess the contribution of intestinal NKT cells in this shift, α-GalCer-treated mice and untreated control mice were killed on day 8 after infection, and NKT cells were purified from the lamina propria (Fig. 4,A). As shown in Fig. 4,A, the purity of the sorted population was >90% in both α-GalCer-treated (Fig. 4,A, a) and untreated (Fig. 4,A, b) animals. The mRNA production of different cytokines by the purified NKT cell population was measured by RT-PCR. The results are expressed as the relative increase or decrease in mRNA expression for different cytokines in NKT cells isolated from α-GalCer-treated mice compared with control infected, but untreated, mice. Compared with controls, IL-10, IL-4, and IL-13 mRNA expressions were increased in the NKT cell population isolated on day 8 from mice treated with α-GalCer and infected (Fig. 4 B). These data indicate that treatment with α-GalCer shifts the NKT cell cytokine pattern to a Th2-like profile.

The production of IL-10 and IL-4 by NKT cells stimulated with α-GalCer was increased in the intestines of treated mice. In contrast, IL-13 production by NKT cells after treatment with α-GalCer did not lead to an increase in this cytokine in the whole intestine throughout the serial time points after infection.

The contribution of IL-4 production associated with α-GalCer treatment to interference with the induction of T. gondii-induced death was evaluated by a series of experiments using blocking Ab. Blocking of IL-4 the day before α-GalCer treatment partially reversed its beneficial effect, as shown by a 50% survival rate compared with 100% survival of mice in the α-GalCer alone-treated group (Fig. 5 A). These observations suggest a partial role for IL-4 in the protection induced by α-GalCer in this model.

The contribution of IL-10 production associated with α-GalCer treatment in interfering with the induction of T. gondii-induced death was evaluated using genetically deficient and chimeric mice. Strikingly α-GalCer treatment had no beneficial effect on protection in IL-10^−/− mice (Fig. 5 B). These observations suggest a pivotal role for IL-10.

To determine whether IL-10 produced by NKT cells was sufficient to suppress lethal intestinal inflammatory lesions, double-chimeric mice were generated. B6 mice were irradiated and reconstituted by a 50/50% mix of bone marrow cells from Jα281^−/− (NKT cell-deficient) and IL-10^−/− mice. After reconstitution, the double-chimeric mice expressed a normal immunological phenotype, except for the NKT cells that were IL-10^−/− (NKT IL-10^−/−). These NKT IL-10^−/− chimeric mice and their appropriate controls (B6 mice, Jα281^−/− and IL-10^−/− mice) were treated with α-GalCer the day before infection. NKT IL-10^−/− chimeric mice treated with α-GalCer rapidly lost more weight than α-GalCer-treated B6 mice (Fig. 5 C), indicating that the lack of IL-10 production by the NKT cells alone conferred greater susceptibility to the infection.

However, in contrast to what was expected, the decreased protective effect of α-GalCer treatment in NKT IL-10^−/− chimeric mice did not lead to a significant increase in the mortality rate (80% survival; Fig. 5,D). These results, demonstrating the complete lack of effect of α-GalCer treatment in IL-10^−/− mice (Fig. 5,B) and a reduced effect of this treatment in NKT IL-10^−/− chimeric mice (Fig. 5, C and D), suggested that other cell types might be the source of the IL-10 that is critical for protection. T regulatory cells (CD4⁺CD25⁺) that express the transcription factor FoxP3 and are known as important IL-10 producers were assessed after treatment with α-GalCer and infection. Interestingly, the number of CD4⁺CD25⁺ cells from intestines and MLNs were increased on days 6 and 9, respectively (data not shown), after infection, and this correlates with an increased expression of FoxP3 in the intestine on day 6 and in MLNs on day 9 from B6 mice, but not from Jα281^−/− mice (Fig. 6,A). The sorted CD4⁺CD25⁺ cell subpopulation exhibited IL-10 mRNA expression (data not shown). Whatever the time after infection and the treatment with or without α-GalCer, the sorted NKT cell population failed to express either FoxP3 or CD25. To better characterize the implication of these T regulatory cell subpopulations to the protective process induced by α-GalCer, the effect of this treatment in mice also treated with blocking anti-CD25 Abs was studied. Treatment with anti-CD25 abrogated the protection (Fig. 6 B), indicating the crucial role of these cells in the anti-inflammatory process induced by treatment with α-GalCer.

In contrast to B6 mice that develop acute lethal ileitis after oral infection with T. gondii, mice deficient in NKT cells, although permissive to parasite replication, are more resistant to this severe immunopathological manifestation, suggesting a critical role of these cells in the intestinal inflammation. NKT cells, present in the intestine at early stages after infection, can secrete IFN-γ that will initiate a Th1-like immune response mediating the lethal ileitis. The critical role of IFN-γ was confirmed by studies showing that mice deficient in IFN-γ production do not develop ileitis (27).

Results from this study show that the harmful effect of NKT cells can be neutralized by treatment with a single injection of α-GalCer. When intestinal NKT cells were stimulated by α-GalCer the day before infection, minor intestinal lesions developed, and the mice survived the infection. The beneficial effect of α-GalCer was accompanied by a shift in cytokine production by the intestinal NKT cells toward a Th2 profile (IL-4 and IL-10) and a dramatic increase in CD4⁺CD25⁺Foxp3⁺ cells in MLNs. Depletion of regulatory T cells abrogated the protective effect of treatment with α-GalCer before the infection. This observation indicates that activation of NKT cells by α-GalCer triggers a regulatory T cell response that helps control the inflammatory intestinal disease observed after T. gondii infection.

We showed for the first time that conventional CD1d-restricted NKT cells are present in the small intestine of T. gondii-infected mice; more precisely, they are located within the lamina propria compartment. They are not associated with IELs in this model, contrary to what was described in previous studies that have identified NK-like T cells within the intraepithelial compartment of the mouse small intestine (35). The presence of unconventional NKT cells, non-CD1d-restricted cells, was also described in the large intestine (36). In this study it was observed that the purified NKT cells were mainly of the CD4⁺ phenotype, with double-negative CD4⁻8⁻ cells making up the difference.

Upon polyclonal or Ag-specific stimulation through the TCR, CD1d-restricted NKT cells have the capacity to produce IL-4 and IFN-γ (11). In this model of pathogen-driven ileitis, we observed that intestinal CD1d-restricted NKT cells promote an IFN-γ response, as reflected by the marked reduction of IFN-γ mRNA expression at serial time points after infection in Jα281^−/− mice devoid of NKT cells compared with wild-type control mice. This early IFN-γ production by intestinal NKT cells may influence the Th1/Th2 balance and thus favor the switch toward a local inflammatory Th1 immune response. Secretion of IFN-γ by intestinal NKT cells may induce DC to secrete IL-12, resulting in an increased production of IFN-γ and TNF-α by lamina propria CD4⁺ T cells that are important effector cells in the hyperinflammatory process associated with oral T. gondii infection. IFN-γ produced by NKT may activate other cell types, such as macrophages and neutrophils (37), that will act on NK cells and CD8 T cells to enhance their IFN-γ production. Our data confirmed the findings of previous studies in which NK1.1⁺ cells were identified as a source of IFN-γ that is essential to limit parasite replication (32, 46) and also point out their role in triggering an exacerbated IFN-γ response leading to immunopathology.

NKT cells are certainly not the only source of IFN-γ. In Jα281^−/− mice, characterized by the absence of NKT cells, a limited amount of IFN-γ was secreted after infection, followed by a significant increase in cytokine production with time (day 8). This late IFN-γ production indicates that other cells within the responding immune population (e.g., CD4⁺ T cells from the lamina propria) are specifically activated and probably are responsible for the death of 75% of the Jα281^−/− mice and the mild inflammation observed in the intestines of surviving mice.

NKT cells can be activated through different pathways. Activation through TCR ligation by CD1d-associated glycolipid is one possibility. Alternatively, IL-12 might activate NKT cells directly, in the absence of TCR engagement (38, 39), or might synergize its effect to that of TCR engagement (40). The activation pathway responsible for NKT cells activation after T. gondii infection remains unclear. It is indeed unknown whether TCR engagement by Toxoplasma Ag or through recognition of self Ag is required. Recently, Brigl et al. (40) have described a model in which NKT cells in the presence of IL-12 were activated after recognition of self Ags presented by CD1d. IL-12 was first made by DCs in response to microbial products, and this cytokine, in turn, activated NKT cells to up-regulate CD69 expression and IFN-γ production. One of the potential Toxoplasma Ag responsible directly or indirectly for NKT activation is the surface Ag-1 (SAG1) protein, the major surface protein of the parasite. The SAG1 molecule induces the dominant Ab response during infection (41) and a strong Th1 immune response characterized by high levels of IFN-γ production by CD4 T cell from the lamina propria and CD8 T lymphocytes (42, 43). SAG1 is a GPI-anchored protein and could be a potential ligand for CD1d molecule.

The hypothesis of the activation of NKT cells through TCR recognition of CD1d-presented Ag is attractive in our model. However, after oral infection with T. gondii, CD1^−/− (B6 background) mice developed an acute and lethal ileitis within 7 days despite the absence of NKT cells. This suggests that CD1d may act via several alternative pathways. Besides its activity on NKT cell activation, CD1d is important for the activation of IELs (33) that down-regulate the intestinal inflammation after T. gondii infection. Indeed, upon Ag activation these IEL secrete copious amounts of TGF-β that participate in the maintenance of gut homeostasis (28). The lack of CD1 expression leads to the absence of protective IELs, and the absence of regulatory mechanisms overcome the absence of inflammatory NKT cells. In addition, the CD1d molecule is expressed on both the apical and the basolateral membranes of intestinal epithelial cells (44), and its ligation induces IL-10 secretion by these cells (45). Thus, the regulation of CD1 expression and its recognition by the TCR could play important roles in the regulation of intestinal inflammatory processes.

In this model of pathogen-driven inflammatory disease, NKT cells are important for the initiation of the robust Th1 inflammatory immune response in the intestine after oral parasite infection. Alternatively, α-GalCer and related glycolipids can modulate NKT cell responses toward a Th2-like profile (11, 12, 46). Our observations demonstrate that α-GalCer treatment has an impact on the intestinal immune response by shifting the cytokine profile production by NKT cells toward a Th2 phenotype, resulting in orientation of the lamina propria CD4 response. A single dose of α-GalCer prevented the development of lethal ileitis after infection with T. gondii. This treatment resulted in a Th2 immune response characterized by the production of IL-4, IL-10, and IL-13 by intestinal NKT cells. The major cytokine implicated in this protection is IL-10, because the beneficial effect of α-GalCer treatment was completely abrogated in IL-10-deficient mice.

Our data are in full agreement with previous work reporting the high susceptibility of IL-10-deficient mice to the development of lethal ileitis after oral T. gondii infection (47). This susceptibility is associated with the defect of T cells to produce IL-10, because mice with an inactivation of the IL-10 gene restricted to T cells generated by Cre/loxP-mediated targeting of the IL-10 gene succumb to severe immunopathology upon infection with T. gondii (48).

IL-10 secreted by NKT cells also participated in the protective effect of α-GalCer treatment, because double-chimeric mice in which NKT cells alone were impaired in IL-10 secretion were more susceptible to the development of ileitis than controls after α-GalCer injection. However, other IL-10-producing cells are also implicated, because treatment with α-GalCer reduced the mortality of these double-chimeric mice. Regulatory CD25⁺ T cells are the likely candidates, because they are present in the intestine, and the anti-CD25 treatment blocked the protective effect of α-GalCer injection.

IL-10 produced by NKT cells has been shown to exert an important regulatory function in experimental models of different pathologies, such as diabetes (49) and allergic encephalomyelitis (50). The link between the shift in the cytokine profile produced by NKT cells toward a Th2 profile and the activation of regulatory CD4⁺ T cells is as yet unknown. IL-10-producing CD4⁺ NKT cells are involved in the generation of regulatory CD8⁺ T cells after Ag exposure in the anterior chamber of the eye (51). Several reports indicate that NKT cells may contribute to immunoregulation via DC maturation (52). DC maturation in the presence of IL-10 may equally induce T regulatory 1 or Th3 regulatory T cells (53). Secretion of IL-4 and IL-10 by intestinal NKT cells after α-GalCer treatment may act directly on local DCs during induction of the polarization of the immune response and promote a Th2 profile. There is evidence that DCs that mature in the presence of NKT cells produce greater amounts of IL-10 and lose the ability to secrete IL-12, a phenotype consistent with a tolerogenic function (17).

The participation of IL-4 in this process cannot be ruled out. The role of IL-4 seems to be complex in toxoplasmosis. Our data indicate that neutralization of IL-4 cannot render α-GalCer-treated mice as susceptible as wild-type, infected, untreated mice, indicating the participation of other cytokine, such as IL-10. In addition, these experiments might indicate, as suggested by Nickdel et al. (54), that IL4-deficient mice are more resistant than wild-type mice to the development of ileitis. However, our data for IL-4 corroborate previous findings reporting that treatment with α-GalCer or OCH (a synthetic glycolipid that has shorter hydrophobic chain) improves mucosal Th1/Th2 cytokine balance by increasing IL-10 and IL-4 production and prevents experimental colitis in mice (55).

The important role played by NKT cells in the regulation of the intestinal immune response has also been previously suggested in a colitis model induce by chemical agents such as dextran sodium sulfate (56) or oxazolone (57). The pathogenic pathway leading to tissue injury in dextran sodium sulfate-induced colitis and, by extension, in Crohn’s disease was attributed to production of Th1 cytokines such as IFN-γ and to the presence of NK1.1⁺ T cells (56). However, the pathogenic pathway leading to tissue injury in oxazolone colitis was also associated with NKT cells secreting IL-13 (57).

The presence of IL-10-secreting T regulatory lymphocytes has been associated with regulation of intestinal inflammation (33), and in our model these cells may be ultimately responsible for the protective effect seen after treatment with α-GalCer. These data illustrate the dual potential of NKT cells in orienting distinct (i.e., Th1 or Th2) immune responses depending on the stimuli used.

After activation with T. gondii, NKT cells are important mediators of the immune response via a robust IFN-γ-mediated effect that limits parasite replication and allows for parasite clearance. However, this early and influential response is not without drawbacks and can be detrimental to the host. This response, when uncontrolled, leads to the development of an acute inflammatory process and death within 7 days of infection in this experimental model of pathogen-driven ileitis. Our data highlight the crucial role of NKT cells derived from the gut in the modulation of intestinal homeostasis.

We thank Emanuelle Perret for technical advice on confocal microscopy, and the Anatomopathology Unit from Department of Pathology, Institut Pasteur, for histological preparations and valuable advice. We are grateful to Anne Louise for the FACS cell sorting. We are especially indebted to Pharmaceutical Research Laboratory, Kirin Brewery Co., Ltd. (Gunma, Japan), for providing α-GalCer, and to M. Kronenberg and P. Van Erdert for providing plasmid containing CD1d and β₂-microglobulin genes and helping us to prepare CD1d/α-GalCer-tetramer, respectively.

The authors have no financial conflict of interest.