A Requirement for the Vγ1⁺ Subset of Peripheral γδ T Cells in the Control of the Systemic Growth of Toxoplasma gondii and Infection-Induced Pathology¹

Although γδ T cells are relatively rare in lymphoid tissues, they are well represented in epithelia and in particular in the intestinal epithelium, where they can comprise the largest population of intraepithelial lymphocytes (IELs)⁵ (1, 2). The immunoregulatory properties of TCR-γδ⁺ IELs (3, 4) and their involvement in epithelial homeostasis (5, 6, 7) are consistent with their providing a first line of defense against infection (2, 8). Systemic populations of γδ T cells in contrast can contribute to both the early and the later stages of infection. The observation that pathogen infection of γδ T cell-deficient mice results in an exaggerated immune response, chronic inflammation, and tissue injury (9, 10, 11, 12, 13, 14, 15, 16, 17) suggests that systemic γδ T cells are of particular importance in controlling or limiting inflammatory responses and resolving pathogen-induced immune responses. The finding that γδ T cells play a role in macrophage homeostasis, especially in the elimination of activated pathogen-elicited macrophages (18, 19), and can influence αβ T cell responses (11, 20) is consistent with this interpretation. It is not clear, however, what the relationship is between γδ T cells that are active at different stages of pathogen-induced immune responses, whether different populations of γδ T cells have specific functions, and what factors determine the functional properties of individual populations. Two possibilities are that γδ T cell function is predetermined and that different populations have restricted or defined functions, or that γδ T cells are functionally heterogeneous with their response being shaped by environmental factors.

Based upon the functional analysis of individual subsets of γδ T cells in murine models of Coxsackievirus B3 infection and airway hypersensitivity, it has been proposed that γδ T cell function cosegregates with TCR variable (Vγ) gene usage (21, 22, 23, 24). In the Coxsackievirus infection model, the cytokines produced by different populations of γδ T cells have a significant influence on the CD4-Th cell phenotype with Vγ1⁺ (IL-4⁺) T cells being able to suppress and Vγ4⁺ (IFN-γ⁺) T cells promote susceptibility to the development of virus-induced myocarditis (21, 22). In airway hypersensitivity, in contrast, Vγ1⁺ T cells can promote airway inflammation, whereas Vγ4⁺ (IFN-γ⁺) pulmonary T cells can negatively regulate airway responsiveness (24). Our own studies of γδ T cell function in the murine model of listeriosis, however, have shown that certain populations of γδ T cell are functionally heterogeneous and that their response is not predetermined, but is instead influenced by environmental factors (25). Based upon cytokine profiling and TCR-CDR3 Vγ/Vδ sequence analyses of Listeria-elicited Vγ1⁺ T cells, it was shown that the populations responding during the early and late stages of the immune response are distinct and nonoverlapping. The question of how γδ T cell responses are determined is, therefore, unresolved. In considering how to redress these contrasting findings from the different experimental model systems, it is noteworthy that the evidence for functional plasticity of γδ (Vγ1⁺) T cells is based upon the analysis of a single infectious disease model. To directly address this issue and determine the universality of the proposed function of Vγ1⁺ T cells, we have used a murine model of peroral Toxoplasma gondii infection. This model of infection is particularly well suited to assessing the impact TCR-Vγ gene usage vs tissue environmental influences have on Vγ1⁺ T cell responses because the contribution Vγ1⁺ T cells in the intestine (IELs) vs peripheral tissues make to protecting the host from infection can be assessed.

Toxoplasma infection is perhaps the most prevalent parasitic infection of mankind in the world (26) and can be divided into two temporally distinct phases, acute and chronic (27, 28, 29). The acute phase is characterized by widespread dissemination of rapidly dividing tachyzoites that can invade virtually any cell type. Chronic infection is associated with differentiation of tachyzoites into bradyzoites that form quiescent cysts within the brain and skeletal muscle. The generation of long-lived protection to Toxoplasma infection is dependent upon the development of a strongly polarized, IL-12-dependent, Th1-type response that is characterized by the production of IFN-γ by NK and CD4⁺/CD8⁺ T cells (30, 31, 32, 33, 34). The failure to appropriately regulate the innate or adaptive response can lead to severe immunopathology, which is seen in susceptible inbred strains of mice that develop acute and lethal ileitis after oral infection (35, 36, 37). Circumstantial evidence suggests that γδ⁺ IELs may contribute to protection against Toxoplasma infection. The ability of adoptively transferred IELs to confer long-term protection against an otherwise lethal parasite challenge has been shown to be at least partially dependent upon the presence of TCR-γδ⁺ IELs in the host (38). In addition, γδ IELs may contribute to the local production of TGF-β (3, 4) that is required for regulating T. gondii-induced ileitis (39). However, it is not clear which population(s) of γδ IELs might contribute to host protection. The evidence for the involvement of systemic γδ T cells in the host response to Toxoplasma infection once it has invaded the epithelium is more compelling. γδ T cells expand in response to T. gondii infection in genetically dissimilar strains of mice (40, 41, 42), rats (43), and humans (44, 45, 46, 47), the timing of which in rodents is dependent upon the strain and dose of Toxoplasma used and the route of administration. The increased susceptibility and mortality of γδ T cell-depleted mice to infection (40, 41), and that the adoptive transfer of cytolytic and IFN-γ-producing γδ T cells prolongs survival against acute peroral parasite challenge (41) are consistent with a requirement for γδ T cells in immunity to Toxoplasma infection. Other studies, however, have found little or no evidence of a protective role for γδ T cells (38, 48). It is not clear, therefore, what role γδ T cells and the Vγ1 subset might play in the host response to Toxoplasma infection.

In this study, we show that whereas Vγ1⁺ IELs do not make a significant contribution to limiting parasite-induced inflammation and tissue injury in the intestine, systemic Vγ1⁺ T cells play a central role in limiting parasite growth and inflammation in peripheral tissues and, in particular, in the CNS, a property associated with the ability of Vγ1⁺ T cells to eliminate parasite-elicited and infected macrophages.

Six- to 8-wk-old male C57BL/6 mice were purchased from Harlan Laboratories. C57BL/6 TCR-δ^−/− and TCR-β^−/− mice were purchased from The Jackson Laboratory and bred in our facility. TCR-Vγ1^−/− mice were generated in our laboratory (25) and backcrossed onto the C57BL/6 background for at least six generations. All mice were housed in specific pathogen-free conditions at Leeds University. Tachyzoites of the type I RH strain of T. gondii-expressing yellow fluorescent protein (YFP) (49) were maintained by serial passage through confluent monolayers of 3T3 fibroblasts in DMEM supplemented with 2 mM l-glutamine, 10% heat-inactivated FBS, 5 U/ml penicillin, and 10 g/ml streptomycin and incubated at 37°C, 5% CO₂. All reagents were from Sigma-Aldrich. The type II ME-49 strain was maintained by passage in outbred, CD1 mice (Charles River Laboratories). Cysts were isolated from chronically infected brains by discontinuous Percoll (Amersham Biosciences) gradient centrifugation of brain homogenates. Bradyzoites were recovered from the cysts by digestion with an acid pepsin solution (50). Parasite numbers were determined by phase-contrast microscopy and diluted to a working concentration in PBS. Eight-week-old male mice were infected by oral gavage with 3000 ME-49 bradyzoites (equivalent of 6–10 cysts), or 10⁶ YFP-RH tachyzoites. All animal experiments were conducted at the University of Leeds and in full accordance with the Animals Scientific Procedures Act 1986 under Home Office approval.

Spleen, liver, and mesenteric lymph node (MLN) were homogenized; contaminating erythrocytes were lysed with 0.84% (w/v) ammonium chloride solution; and the cell suspension was passed through 0.7-μm nylon filter and washed before Ab staining. Small intestinal IELs were isolated from Peyer’s patch-excised small intestines, as described previously (51). Abs used for surface staining included: F(ab′)₂ of mAbs specific for TCR-Vγ1 (clone 2.11) (52) and TCR-δ (GL3). The 2.11 hybridoma cell line was provided by P. Pereira (Institut Pasteur, Paris, France). Commercial anti-mouse mAbs used were anti-CD3 (145 2C11), anti-TCR-γδ (GL3), anti-TCR-αβ (H57-597), anti-B220 (RA3-6B2), anti-Gr-1, and anti-F4/80 conjugated to FITC, PE, or PE-Cy5.5 purchased from Caltag-Medsystems or BD Pharmingen. Streptavidin conjugates of PE, FITC (Caltag Laboratories), or Alexa Fluor 633 (Molecular Probes) were used as secondary reagents. To block nonspecific Ab binding, cells were preincubated with anti-FcR Ab mixture, anti-CD16/32 (20μg/ml) (Caltag Laboratories). Isotype-matched Abs of irrelevant specificity were used to determine the level of nonspecific staining. Stained cells were analyzed on a FACSCalibur flow cytometer using CellQuest software (BD Pharmingen).

For histology, tissues were fixed overnight in phosphate-buffered 10% Formalin (Sigma-Aldrich) and embedded in paraffin. Five-micrometer sections were cut using a microtome, mounted on SuperFrost Plus microscope slides (BDH), and either stained directly with H&E or with biotinylated anti-active caspase 3 (BD Pharmingen), followed by avidin-biotin-HRP complexes (ARC amplification system; DakoCytomation) and diaminobenzidine substrate before counterstaining with hematoxylin. As an independent evaluation of liver injury serum, alanine aminotransferase (ALT) levels were measured on an ADVIA 2400 chemistry system analyzer (Bayer) following the manufacturer’s instructions.

Plastic-adherent peritoneal exudate cells, spleen, or liver mononuclear cells (5 × 10⁵) from day 10 ME-49-infected TCR-δ^−/− mice were incubated with 1 × 10⁶ nonadherent splenocytes from noninfected wild-type, TCR-δ^−/−, TCR-β^−/−, or TCR-Vγ1^−/− mice in 8-well chamber slides (Valeant Pharmaceuticals) and cultured at 37°C in RPMI 1640 (Sigma-Aldrich) with 10% FBS for 1 h, as previously described (53). For some experiments, γδ T cell-enriched preparations of splenocytes (>80% TCR-γδ⁺) prepared by immunomagnetic selection (25) from wild-type mice were used as effector cells. The extent of macrophage killing was determined by labeling target cells with the Live/Dead cell reagent (Molecular Probes) that contains fluorescent dyes that identify intracellular esterase activity of viable cells (3,3′-dioctadecyloxacarbocyanine) or that are incorporated in the nuclei of dead cells (propidium iodide). At least 100 live and/or dead cells were counted in four separate fields by UV microscopy.

Intracellular cytokines were detected by cytoplasmic staining of IELs and splenocytes recovered from parasite-infected wild-type mice after culture in vitro for 5 h with 10 mg/ml brefeldin A (Sigma-Aldrich). Cells were then surface stained with anti-CD3, anti-TCR-γδ, and anti-TCR-Vγ1, or anti-TCR-Vγ7 mAbs, fixed in 1% paraformaldehyde, and permeabilized with 0.5% of saponin (Sigma-Aldrich) before cytoplasmic staining with PE-conjugated anti-mouse cytokine mAbs to IL-2, IL-4, IL-5, IL-6, IL-10, IFN-γ, and TNF-α (Caltag-Medsystems or BD Pharmingen), or FITC-conjugated polyclonal Abs to MIP-1β and MCP-1 (Sigma-Aldrich). Anti-TGF-β and anti-latency-associated peptide (LAP) mAbs (R&D Systems) were conjugated to FITC and used for surface staining. PE- and FITC-conjugated isotype-matched mAbs of irrelevant specificity were used to determine levels of nonspecific staining of anti-cytokine mAbs. Serum cytokine levels were determined using flow cytometric cytokine bead arrays (BD Pharmingen), according to the manufacturer’s instructions.

All data were assessed for normal distribution using a Shapiro-Wilk test. For parametric data analysis was performed using Student t test, and for nonparametric data analysis was performed using Mann-Whitney U tests using the Statistical Package for the Social Sciences software, SPSS. Statistical significance was taken as p < 0.05.

The role of Vγ1⁺ T cells resident in the small intestine and peripheral lymphoid tissues in the host response to T. gondii infection was investigated by comparing the outcome of infecting wild-type (C57BL/6) mice and strains genetically deficient of all γδ T cells (TCR-δ^−/−) or of the Vγ1⁺ T cell subset (Vγ1^−/−) (25) with virulent type I tachyzoites or avirulent type II bradyzoites of T. gondii. For the analysis of the involvement of γδ IELs in the acute stage of Toxoplasma infection, a recombinant form of the RH strain of tachyzoites that express a highly fluorescent reporter gene (tandem YFP) (49) was used. Because YFP-reporter gene expression was lost during tachyzoite to bradyzoite stage conversion and RH is a noncyst-forming strain, it was not possible to use YFP-RH to study the chronic stage of Toxoplasma infection. Instead, the avirulent and cyst-forming type II ME-49 strain was used to study the chronic stage (2–6 wk postinfection) of T. gondii infection. In a series of preliminary experiments, an infectious dose of 10⁶ tachyzoites and 3000 bradyzoites (equivalent to 5–10 cysts) was shown to produce a productive, but nonlethal acute and a chronic infection, respectively, in TCR-δ^−/− and Vγ1^−/− mice, and these doses were used in all subsequent experiments described in this study.

Both the onset and severity of T. gondii-induced ileitis were accelerated and more severe in TCR-δ^−/− mice infected with RH tachyzoites compared with wild-type mice (Table I and Fig. 1,A). The first signs of inflammation were evident in TCR-δ^−/− mice 4 days postinfection (Table I), some 3 days before the involvement of αβ T cells (35). Pathology was characterized by the presence of small distinct foci of infiltrating polymorphonuclear cells (PMNC) and villus edema and blunting. This was in contrast to the virtually normal histology of the intestine of wild-type and TCR-Vγ1^−/− mice (Table I). In contrast, TCR-β^−/− were more susceptible to RH infection, with the majority (8 of 10) developing a lethal ileitis within 7 days postinfection (Table I), consistent with the requirement for αβ T cells (Th1) for protection from Toxoplasma infection (30, 31, 32, 33, 34).

By 7 days postinfection, inflammation was more extensive and severe in TCR-δ^−/− mice (Fig. 1,A and Table I), with the almost complete loss of normal epithelial architecture; hemorrhaging in the lamina propria; massive accumulation of inflammatory cells comprising monocytes, macrophages, and lymphocytes throughout the ileum; and extensive necrosis. Pathology was also evident at this time in TCR-Vγ1^−/− mice, although it was less severe than that seen in TCR-δ^−/− mice and was more comparable to that of wild-type mice, with fewer foci of inflammation and necrosis throughout the ileum and less extensive changes in villous architecture (Fig. 1,A and Table I). The increased susceptibility of TCR-δ^−/− mice to Toxoplasma infection was not parasite strain or stage specific because a similar outcome was seen after infection with ME-49 bradyzoites (Table I) and with the less virulent type II PRU strain of tachyzoites (J. Dalton and S. Carding, unpublished observations). Although PRU- and ME-49-mediated inflammation in TCR-δ^−/− mice was as extensive as that of RH infection, it was less severe. Also, the onset of PRU- and ME-49-induced inflammation was delayed compared with RH infection and was first evident after 7–10 days in both wild-type and TCR-δ^−/− (Table I, and data not shown), which most likely reflects the decreased tissue transmigratory and long distance migratory capabilities of type II strains of Toxoplasma (54). Compared with TCR-δ^−/−-infected mice, TCR-Vγ1^−/− mice resembled the wild type in being more resistant to ME-49 infection, with the first signs of ME-49-induced ileitis apparent only after 2 wk postinfection (Table I). Collectively, these findings demonstrate that compared with TCR-δ^−/− mice, Vγ1^−/− mice are more resistant to parasite-induced inflammation and tissue injury.

IEL responses have been associated with epithelial cell-directed cytotoxicity, the release of immunoregulatory cytokines/chemokines, and up-regulation of Th1 cytokines (reviewed in Ref.2). We, therefore, investigated whether these responses involved or required Vγ1⁺ IELs during the initial response to T. gondii infection in vivo. Staining of sections of ileum with Abs specific for active caspase 3, an early indicator and mediator of apoptotic cell death (55), identified very few, but equivalent numbers of apoptotic cells in the epithelial mucosa of wild-type, γδ-, or Vγ1⁺-deficient mice at 4 days postinfection, a time point coincident with the onset of ileitis in TCR-δ^−/− mice (Fig. 1,B). Analysis of cytokine production by γδ IELs during the development of ileitis (Fig. 1,C) showed that the majority (>90%) did not synthesize any of the cytokines analyzed, including examples of those known to be constitutively expressed (e.g., TGF-β) or up-regulated by IELs upon activation (e.g., IFN-γ, TNF-α, and TGF-β) (2). Among the small number of cytokine-producing Vγ1⁺ IELs, synthesis of IL-4, IL-6, and IL-10 was detected. Among Vγ7⁺ IELs, cytokine (primarily IL-6) production was restricted to a small subset (∼10%) of cells (Fig. 1 C).

By 1 wk postinfection, the parasite burden in the majority of infected tissues, including the liver, peritoneum, and lungs of TCR-δ^−/− and TCR-Vγ1^−/− mice, was increased compared with wild-type mice (Fig. 2,A, and data not shown). Within the spleen, the majority of parasites were intracellular within F4/80⁺ macrophages (Fig. 2,A). This increase in parasite burden in TCR-δ^−/− and TCR-Vγ1^−/− mice occurred in parallel with the development of the peripheral γδ and Vγ1⁺ T cell response to Toxoplasma infection in wild-type mice (Fig. 2,B). The γδ T cell response was dominated by Vγ1⁺ T cells that comprised 60–90% of all γδ T cells in the MLN, spleen, and liver at the time points analyzed. This pattern of the γδ and Vγ1⁺ T cell response was seen in both tachyzoite- and bradyzoite-induced infections (data not shown). γδ T cell expansion was first evident 7 days postinfection in both the lymphoid tissues draining the initial site of infection and in primary lymphoid organs, reaching maximum numbers at ∼10 days postinfection, after which they declined (Fig. 2 B).

The effects of the absence of γδ and Vγ1⁺ T cells during the acute stages of Toxoplasma infection were also evident in the primary sites of parasite invasion, including the liver, lungs, and CNS after infection with either RH tachyzoites (Table I) or ME-49 bradyzoites (Fig. 3 and Table I). The onset and extent of inflammation were more rapid and severe after infection with virulent RH tachyzoites compared with avirulent ME-49 bradyzoites. By 4 days postinfection, liver pathology was first apparent in RH-infected TCR-δ^−/− mice and was characterized by portal tract inflammation and the presence of PMNC and lymphocytes, and by 7 days postinfection had developed into severe inflammation with extensive necrosis (Table I). Liver pathology in infected TCR-β^−/− mice resembled that of TCR-δ^−/− with evidence of inflammatory foci and lesions, although they were fewer in number and smaller compared with those seen in TCR-δ^−/− mice (data not shown). Severe liver pathology and injury in TCR-δ^−/− mice were also reflected in changes in serum levels of the liver enzyme, ALT, which at 7 days postinfection was significantly (p < 0.05) higher in TCR-δ^−/− (333 ± 31 IU/L) compared with wild-type (174 ± 23 IU/L) and TCR-β^−/− (199 ± 14 IU/L) mice. The extent of liver pathology in T. gondii-infected TCR-Vγ1^−/− mice was similar to that seen in TCR-δ^−/− mice, although the onset of inflammation and necrosis was delayed. Up to 10 days postinfection, the liver histology in TCR-Vγ1^−/− mice resembled that of wild-type mice, whereas by 14 days it was comparable to that of TCR-δ^−/− mice (Fig. 3). Changes in serum ALT levels in TCR-Vγ1^−/− mice (241 ± 34 IU/L at day 7, 284 ± 28 at day 10, and 379 ± 32 at day 14) also reflected the delay in the development of liver pathology. A similar pattern of pathology was seen in the lungs of T. gondii-infected mice, with the almost complete ablation of normal tissue architecture due to a massive infiltrate of inflammatory cells (primarily PMNC and lymphocytes) apparent in TCR-δ^−/− mice by 7 days postinfection and by 10–14 days postinfection in TCR-Vγ1^−/− mice (data not shown).

Toxoplasmic encephalitis was more severe in TCR-Vγ1^−/− mice infected with ME-49 bradyzoites compared with wild-type mice (Fig. 3). By 3 wk postinfection, both TCR-δ^−/− and TCR-Vγ1^−/− mice developed severe encephalitis characterized by vascular cuffing and an extensive infiltrate of inflammatory cells comprising of macrophages, granulocytes, and lymphocytes in the meninges and in brain tissue itself, where it encircled cysts. Inflammation persisted in the brain of TCR-Vγ1^−/− mice for >6 wk after the initial infection (Fig. 3, and data not shown), consistent with the requirement for Vγ1⁺ T cells in controlling inflammation.

In view of the chronic inflammation seen in Toxoplasma-infected TCR-Vγ1^−/− mice and the known involvement of Vγ1⁺ T cells in macrophage homeostasis (25, 53), we sought evidence for the disruption of macrophage homeostasis in T. gondii-infected TCR-Vγ1^−/− mice and the involvement of macrophage-reactive Vγ1⁺ T cells in controlling toxoplasmosis. The chronic inflammatory reaction in tissues of Toxoplasma-infected TCR-δ^−/− and TCR-Vγ1^−/− mice was characterized by the accumulation of activated macrophages and granulocytes (Fig. 4), which persisted for up to 4 wk after initial infection. This chronic macrophage and granulocyte response was seen in all major sites of Toxoplasma infection in TCR-Vγ1^−/− mice, including the liver, spleen, lungs, peritoneal cavity (Fig. 4, and data not shown), and the CNS, where it was still evident 6 wk after initial infection (Fig. 3). In wild-type mice, the kinetics of the macrophage response was mirrored by the Vγ1⁺ T cell response to Toxoplasma infection (Fig. 2,B). Further evidence of an uncontrolled and ineffective inflammatory anti-parasite response in TCR-δ^−/− and TCR-Vγ1^−/− mice was provided by the analysis of serum levels of the proinflammatory cytokines TNF-α and IFN-γ (Fig. 4) and CNS parasite burden (Fig. 5,A). Significantly higher levels of both cytokines were present in the serum of TCR-δ^−/− and TCR-Vγ1^−/− mice throughout the first 2 wk of infection, with the highest levels seen between 7 and 10 days postinfection. Significantly higher numbers of parasite cysts were recovered from the brain of TCR-δ^−/− and TCR-Vγ1^−/− mice (Fig. 5 A) with occasional histological evidence of cyst rupture in the brain of TCR-δ^−/− and TCR-Vγ1^−/− mice, suggestive of reactivation that was not seen in wild-type mice (data not shown) and of a role for Vγ1⁺ T cells in containing parasite growth in the CNS during the chronic phase of T. gondii infection.

Because macrophages are a major target of Toxoplasma infection (Fig. 2,A), we determined whether the killing of parasite-elicited and activated macrophages might be a mechanism by which macrophage cytocidal Vγ1⁺ T cells can help restrict parasite dissemination and growth. The killing of T. gondii-elicited, activated macrophages was assessed using a fluorescence-based cytotoxicity assay that we have previously used to demonstrate the killing by Vγ1⁺ T cells of Listeria-elicited macrophages (53), which is Fas-Fas ligand mediated (56). Peritoneal macrophages from day 10 ME-49-infected TCR-δ^−/− mice were used as target cells, and bulk or γδ T cell-enriched splenocytes (>80% γδ T cells) from noninfected wild-type, TCR-δ^−/−, TCR-Vγ1^−/−, or TCR-β^−/− mice were used as effector cells (25). As shown in Fig. 5,B, T. gondii-elicited macrophages were efficiently killed (>80%) by γδ T cells from wild-type mice and TCR-β^−/− mice, but not from TCR-Vγ1^−/− mice or by splenocytes from TCR-δ^−/− mice. Due to the difficulty of identifying parasites within the harvested macrophages, it was not possible to determine the extent of killing of infected vs noninfected, but activated macrophages by Vγ1⁺ T cells. Assuming that at least some of the target macrophages were infected, it was not possible to detect the release of any viable parasites from lysed macrophages either microscopically or in in vitro fibroblast reinfection assays (data not shown). Comparable levels of killing were also seen using the spleen or liver as a source of target macrophages (data not shown). Evidence that this killing was directly attributable to Vγ1⁺ T cells was obtained from experiments showing that in using γδ T cells from Vγ1^−/− mice (Fig. 5,B) the level of macrophage killing was not significantly different from the level of spontaneous macrophage death in cultures of macrophages in the absence of T cells. As observed previously (53), the susceptibility of macrophages to lysis by γδ T cells was dependent upon their prior activation, as γδ T cells did not kill macrophages from noninfected wild-type mice (data not shown). Activation was not, however, a requirement for γδ T cell macrophage cytocidal activity (Fig. 5 B).

Analysis of cytokine production by peripheral, liver and splenic, Vγ1⁺ T cells responding to parasite infection showed that <6% of T cells produced any of the cytokines tested with IL-4, IL-10, and LAP being synthesized by a small subset of liver- and/or spleen-derived Vγ1⁺ T cells. The poor cytokine response of peripheral Vγ1⁺ T cells resembled that of Vγ1⁺ IELs (Fig. 1 C).

Although γδ T cells are found in a number of different anatomical sites, with localization often associated with the expression of distinct TCR-Vγ chains, it is not clear what the relative contribution of the different populations is to protecting the host from pathogens, and whether their functional phenotype is predetermined and cosegregates with TCR-Vγ chain usage or is influenced by environmental factors. Using the protozoan parasite T. gondii as a model enteric pathogen to investigate the contribution that Vγ1⁺ IELs vs Vγ1⁺ peripheral T cells make to infection, we have shown that whereas Vγ1⁺ IELs do not appear to make a significant contribution to the acute phase of infection or to preventing or limiting parasite-induced ileitis, systemic Vγ1⁺ T cells play an essential role in limiting parasite growth and inflammation in peripheral tissues and in the CNS. These findings also demonstrate that the macrophage cytocidal activity of Vγ1⁺ T cells contributes to antiparasitic as well as antibacterial responses and may therefore be a universal feature of pathogen-induced immune responses.

More extensive accumulation and rapid dissemination of parasites and accelerated onset of ileitis in mice deficient of γδ IELs are consistent with these cells playing an important role in resisting pathogen invasion in the small intestine. Although we have been unable to establish the identity of the protective γδ IEL populations, our data suggest that whereas Vγ1⁺ IELs do not contribute to epithelial defense, the increased resistance of Vγ1^−/− mice to parasite invasion in which ∼90% of γδ IELs are Vγ7⁺ (25) suggests that Vγ7⁺ IELs have a direct role in limiting pathogen invasion and the development of intestinal inflammation. The nature of this protective function of Vγ1⁻ γδ IELs is not known, although it appears to be independent of the cytolytic Th1-skewed immunoregulatory properties ascribed to IELs (2). These findings do not, however, discount the importance of γδ IEL cytotoxicity and cytokine production that may be of more relevance to the host response later in the course of infection once parasites have migrated from the intestinal mucosa. Indeed, as ileitis develops ∼1 wk after infection, there is evidence of increased IEL activity and cytokine production (42), which precedes lamina propria or Peyer’s patch T cell involvement (35, 57). IEL-mediated killing of parasite-infected cells is only apparent after 1 wk (∼9 days) postinfection (58). In addition, the ability of adoptively transferred bulk IELs, which include αβ and γδ T cells to confer protection against severe and lethal ileitis in otherwise susceptible mice (38, 51, 59), may be attributable to IELs killing parasite-infected cells (58) and secreting anti-inflammatory cytokines such as TGF-β and IL-10 (39, 60), or by influencing the local activity or homeostatic regulation of other immune cells (53, 61, 62), including pathogenic αβ (CD4⁺) T cells (35, 36, 57).

Although there is no obvious requirement for TCR-Vγ1⁺ IELs in limiting parasite invasion and the development of intestinal inflammation, systemic anti-parasite responses are severely compromised in the absence of Vγ1⁺ T cells. Macrophage cytocidal Vγ1⁺ T cells dominate the systemic γδ T cell response to T. gondii as well as a number of other pathogens (63). In comparison with other experimental infection models, however, the Toxoplasma-elicited Vγ1⁺ T cell response is more potent, resulting in greater numbers of Vγ1⁺ T cells and elimination of the vast majority (>80%) of parasite-elicited, activated macrophages (Fig. 5 B). Assuming that some of these target macrophages are infected with parasites, Vγ1⁺ cells are likely to play a central and important role in Toxoplasma containment because macrophages are major cellular targets for Toxoplasma infection and dissemination in peripheral tissues (this study) (64) and their elimination would be an effective means of controlling parasite infection. In particular, the unresolved inflammation and increased parasite cyst burden in the brain of infected TCR-δ^−/− and Vγ1^−/− mice suggest that Vγ1⁺ T cells can contribute to resolving inflammation in the brain of T. gondii-infected mice by killing activated macrophages that make up a large part of the CNS infiltrate (this study) (65) and can contribute to the development of inflammatory lesions (66). In particular, they may help prevent or limit the extent of CNS inflammation by eliminating macrophages that during the acute phase are infected with cyst-forming bradyzoites (65) and can serve as a kind of “Trojan horse” and transport mechanism for entry of parasite into the brain (64).

The significance of the production of immunoregulatory cytokines, including IL-10, by a small subset of systemic Vγ1⁺ T cells is difficult to interpret due to their small numbers and the fact that these cytokines can be produced by more than one cell type in both the liver and spleen. The production of IL-10 and LAP by Vγ1⁺ T cells may, however, be significant in view of their potent immunosuppressive properties and role in down-modulating pathogenic CD4 T cells and limiting infection-induced pathology (36, 37). Similarly, the significance of the production of IL-6 by a small subset of γδ (Vγ7⁺) IELs is also uncertain, although it might play a role in limiting parasite invasion and the development of intestinal inflammation (67). It is also worth noting that because our analysis of cytokine production by γδ IELs was not exhaustive and did not include all of the cytokines and chemokines known to be produced by these cells (3, 4), it is possible that Vγ1⁺ as well as other γδ IELs populations may contribute to epithelial defense by the production of cytokines other than those analyzed in this study.

Conflicting evidence for both detrimental (61) and beneficial (16) roles of γδ T cells in influencing CNS inflammation has been reported. However, because the identity of the γδ T cells involved was not established, the contribution made by Vγ1⁺ T cells to pathogenic or protective γδ T cell responses in these studies needs to be determined. If Vγ1⁺ T cells are involved, then the timing and nature of their response will be an important factor in determining what the nature of their involvement is. The Vγ1⁺ T cell response to infection is functionally heterogeneous and biphasic with early responding cells of a proinflammatory phenotype and late responding cells with anti-inflammatory properties (25), the timing and magnitude of both being strongly influenced by the nature of the pathogen (virulence), the route, dose, and site of infection, as well as host factors (68). The apparently conflicting roles for γδ, and possibly Vγ1⁺, T cells in the pathogenesis of CNS disease may therefore be explained by differences in the nature of the pathogen and conditions under which γδ T cells are recruited and whether the prevailing microenvironmental conditions favor the involvement of pro- or anti-inflammatory populations.

In demonstrating differences in the involvement of populations of mucosal and systemic γδ T cells that use the same TCR-Vγ chain in regulating host responses to infection, this study increases our understanding of the factors that influence γδ T cell function during both acute and chronic stages of infections. Although the clonal relationship of peripheral vs intestinal Vγ1⁺ T cells has not been established, their contrasting level of involvement and contribution to the host response to Toxoplasma infection suggests that they are functionally distinct, reflecting microenvironmental influences unique to the different anatomical sites in which they reside and operate.

We thank the personnel of the pathology laboratory of the Leeds Teaching Hospital Trust for their help in developing immunohistochemical protocols, and Dr. Antonio Barragan (Karolinska Institut, Stockholm, Sweden) and Professor Jim Alexander and Dr. Craig Roberts (Strathclyde University, Glasgow, Scotland) for helpful comments and discussions during the preparation of the manuscript.

The authors have no financial conflict of interest.