Contact sensitivity (CS) responses to reactive hapten Ag, such as picryl chloride (PCl) or oxazolone (OX), are classical examples of T cell–mediated immune responses in vivo that are clearly subject to multifaceted regulation. There is abundant evidence that downregulation of CS may be mediated by T cells exposed to high doses of Ag. This is termed high dose Ag tolerance. To clarify the T cell types that effect CS responses and mediate their downregulation, we have undertaken studies of CS in mice congenitally deficient in specific subsets of lymphocytes. The first such studies, using αβ T cell–deficient (TCRα−/−) mice, are presented here. The results clearly show that TCRα−/− mice cannot mount CS, implicating αβ T cells as the critical CS-effector cells. However, TCRα−/− mice can, after high dose tolerance, downregulate α+/+ CS-effector T cells adoptively transferred into them. By mixing ex vivo and then adoptive cell transfers in vivo, the active downregulatory cells in tolerized α−/− mice are shown to include γδ TCR+ cells that also can downregulate interferon-γ production by the targeted CS-effector cells in vitro. Downregulation by γδ cells showed specificity for hapten, but was not restricted by the MHC. Together, these findings establish that γδ T cells cannot fulfill CS-effector functions performed by αβ T cells, but may fulfill an Ag-specific downregulatory role that may be directly comparable to reports of Ag-specific downregulation of IgE antibody responses by γδ T cells. Comparisons are likewise considered with downregulation by γδ T cells occurring in immune responses to pathogens, tumors, and allografts, and in systemic autoimmunity.
Cutaneous contact sensitivity (CS)1 responses to contact sensitizing hapten Ag such as picryl chloride (PCl) and oxazolone (OX) are classical manifestations of T cell–mediated immunity in vivo. In previously sensitized hosts, CS and the related delayed-type hypersensitivity (DTH) reaction, are manifest as macroscopically measurable inflammation (skin swelling) that peaks at 24–48 h after topical cutaneous Ag challenge (1).
It appears that early after challenge with hapten, local tissue mast cells (2, 3) and blood platelets (4) are induced to release the vasoactive amine serotonin (2–6) that facilitates local extravasation and subsequent migration and activation of Ag-specific CS-effector lymphocytes. Available data strongly suggest, but do not prove, that the responding lateacting lymphocytes are exclusively MHC class II–restricted, αβ TCR+, CS-effector T cells (7, 8), which customarily make Th1-type effector cytokines, such as IFN-γ (9–11).
By contrast to the conventional CS-effector response, which is induced by immunization via skin painting with reactive hapten Ag, mice respond to high intravenous doses of soluble hapten Ag by developing tolerance. This high dose Ag tolerance is of particular significance with regard to the capacity of individuals that become tolerant toward contact hapten allergens, and may also underlie aspects of tolerance to self Ag. Numerous studies have demonstrated that high dose tolerance can be associated with the development of regulatory T cells that limit the response of CSeffector T cells (12–17). In this regard, it is unclear whether or not such CS-downregulatory T cells are exclusively αβ T cells, or whether they may include γδ T cells.
An assay commonly used to demonstrate high dose downregulation of CS, is the mixing together in vitro of spleen cells from mice tolerized intravenously with high doses of Ag, together with CS-effector cells from contact sensitized mice. Thereafter, the mixed regulatory and CS-effector cells are adoptively transferred to naive recipients, and a subsequent measurement is made of the transferred CS-effector response in the recipients. The studies reported in this manuscript use this mixing assay and other established assays, along with use of TCR-deficient α−/− mice, to assess the cells responsible for CS elicitation and for CS downregulation, respectively. The studies show that γδ T cells cannot substitute for αβ T cells as effectors of CS elicitation, but that γδ T cells can medite downregulation of both αβ CS effectors in vivo and IFN-γ production by these CS effectors in vitro. The studies are particularly pertinent given the reported capacity of γδ T cells to negatively regulate αβ T cell–driven responses in allergic (18, 19) and other immune responses.
Materials And Methods
5–7-wk-old 129/J (H2b), BDF1 (H2b × d), C57Bl/6 (H2b), BALB/c (H2d), and CBA/J (H2k) mice were obtained from Jackson Labs (Bar Harbor, ME). TCRα−/− and TCRα+/− mice (20–22) with different but defined MHC backgrounds that included H2b, H2d, and H2b+d, were supplied locally. All mice were maintained in microisolator cages, and changed in a laminar flow hood. MHC haplotypes were determined by FACS® with antiMHC class Ib (PE-conjugated), and class Id (FITC-conjugated) mAb (PharMingen, San Diego, CA).
Picryl chloride TNP chloride (PCl) obtained from Chemica Alta (Edmonton, Alberta, Canada) was recrystallized from methanol/H2O, and protected from light and humidity. Trinitro-benzene sulfonic acid (TNBSA) (Eastman Fine Chemicals, Rochester, NY); oxazolone (4-ethoxymethylene-2-phenyloxazolone) (OX) (British Drug Houses via Gallard-Schlesinger Inds. Inc., Carle Place, NY); and anti–hamster IgG antibodycoated magnetic beads (1 μm iron magnetic particles; Advanced Magnetics, Inc., Cambridge, MA) were obtained from the manufacturers. Anti-TCR γδ mAb (clone UC7-135D5, hamster IgG), was obtained from Dr. J. Bluestone (University of Chicago, Chicago, IL; 23). Low-tox rabbit complement C was obtained from Pel-Freeze Biologicals (Brown Deer, WI).
Mice were actively sensitized by topical application of 0.15 ml of 5% PCl or 3% OX in 1:3 acetone–ethanol mixture, to the shaved abdomen, chest, and hind feet on day 0. 4 d later, zero time ear thickness was measured with an engineer's micrometer (Mitutoyo Mfg. Co., Tokyo, Japan) before local ear skin challenge via topical application to both sides of both ears of one drop of 0.8% PCl or OX solution in olive oil (for 129/J, BALB/c, TCRα+/− and α−/−), or 1% PCl to less reactive BDF1 or C57Bl/6. The subsequent optimal increase in ear thickness was measured at 24 h, and occasionally 48 h, after challenge as reported (1–8), and expressed in units of 0.01 mm ± SE. In all experiments, a group of nonimmune control animals was also challenged on the ears with 0.8% PCl or OX, and the resulting background increase in ear thickness (∼2 U at 24 h) was subtracted from responses of experimental groups to yield net ear swelling responses, shown in the figures.
Induction of High Ag Dose Tolerance.
For induction of TNP– specific high dose tolerance, mice received two intravenous injections of 3 mg TNBSA (1% TNBSA in distilled H20, readjusted to pH 7.2 with 1 M NaOH) in 0.3 ml on days 0 and 3. On day 7, spleens from these TNBSA-injected mice were harvested as a source of putative regulatory cells. Alternatively, mice were tolerized by intravenous injection of TNP-conjugated syngeneic spleen cells.
To induce OX-specific tolerance, mice were injected intravenously once with 0.3 ml of 10% OX-conjugated mouse red blood cells (MRBC), or with OX-conjugated spleen cells (5 × 107) on days 0 and 3. The MRBC or spleen cells were labeled with OX as described previously (12). Briefly, 1 ml of packed MRBC was resuspended in 10 ml of DPBS (calcium- and magnesium-free PBS) and mixed for 30 min with freshly prepared 20 ml aqueous OX solution (containing 20 mg OX). The OXMRBC were then washed twice with PBS containing 5% FCS, and before use, the peletted OX-MRBC were resuspended to 10% in PBS. Spleens from mice injected intravenously with OXMRBC, or with OX spleen cells, were harvested on d 7 as a source of OX-specific regulatory cells.
Adoptive Cell Transfer of CS Responses and Cell Mixing Assay.
Donors of CS-immune effector cells were contact sensitized with 5% PCl or 3% OX on days 0 and 3. Immune lymph node and spleen cells were harvested on day 7 and 7 × 107 were injected intravenously into normal syngeneic recipients. Immediately after transfer, recipients were challenged on the ears with 0.8% PCl or OX in olive oil. The increase in ear swelling was determined 24 h, and sometimes 48 h later. It was compared to actively sensitized positive controls and to negative controls that were simply challenged on the ears with 0.8% PCl or OX.
For the cell mixing assay, 5–7 × 107 CS-effector immune lymph node and spleen cells from PCl or OX contact sensitized α+/+ donors were incubated for 30 min at 37°C with 5 × 107 regulatory spleen cells from mice intravenously tolerized with TNBSA or with OX-MRBC. Regulatory spleen cells were from tolerized α−/− mice, or control α+/+ 129/J, C57Bl/6, BALB/c, or BDF1 mice. Positive controls were immune CS-effector cells incubated without regulatory cells. After incubation, washed mixed cells from each group were transferred intravenously, and recipients were tested for CS responses by ear challenge with topical application of 0.8% PCl or OX, within 30 min immediately after transfer.
Transfer into Tolerized Mice to Confirm Active Tolerance and Agspecificity.
To test tolerance and Ag specificity of downregulatory cells in another system, α−/− mice were tolerized with a high dose of TNBSA or with OX-MRBC intravenously on days 0 and 3 before serving as cell transfer recipients. Then on day 7, nontolerized control or TNP- or OX- tolerized α−/− mice received 7 × 107 CS-effector cells from normal TCRα+/+ mice (BALB/c) that were Ag homologous, or were of unrelated Ag specificity (TNP versus OX). α−/− recipient mice were then challenged immediately on the ears with homologous hapten (TNP or OX), and ear tested for 24 h ear swelling.
Immuno-magnetic Bead Cell Fractionation.
Spleen cells (1.5 × 108) from four TNBSA-tolerized α+/+ 129/J, or α−/− donor mice were incubated in 20 ml UC7 hamster anti-TCR γδ mAb (12 μg/ml) hybridoma supernatant for 60 min on ice. After mAb coating, cells were washed three times with PBS + 2% FCS, and resuspended in 40 ml PBS + 2% FCS containing goat anti–hamster Ig–coated paramagnetic beads at 5–10 beads per target cell. The cells were then incubated in a flat vertical 50-ml flask for 30 min on ice, after which a magnet (Advanced Magnetics Inc., Cambridge, MA) was applied to one side of the flask. 10 min later, magnetic bead nonadherent cells were recovered, and then the magnetic bead adherent cells were recovered. Nonseparated regulatory cells were used in experiments shown in Figs. 3, 5, and 6 b.
Recoveries of γδ cells were as follows. For 129/J TNBSA-tolerized donors, we harvested 15 × 107 separated spleen cells from three donors, obtaining ∼106 anti-γδ magnetic bead positive cells that were 60% γδ+ by FACS®. Thus, there were ∼1.6 × 105 γδ cells transferred to each of four recipients, while there were 10 × 105 γδ cells in the starting spleen of one donor mouse. For α−/− TNBSA-tolerized donors, we harvested 20 × 107 spleen cells from three donors, obtaining 3.4 × 106 anti-γδ bead positive total cells that were on average 80% γδ+ by FACS®, thus allowing for 7 × 105 γδ-enriched cells transferred to each of four recipients. FACS® analysis of the anti-γδ immunobead separated subpopulations was presented elsewhere (24), in which we reported that the percentage of γδ cells adhering to the anti-γδ beads can vary between 55–95% of the total adherent cells; the rest are not specifically bound, and on assay did not have biologic activity. Also, we noted that unavoidable splenic macrophage phagocytosis of some mAb-coated γδ cells that were attached to beads, probably contributed to variability in γδ T cell recoveries.
Mixing of CS-effector Cells and Magnetic Bead-enriched Regulatory Cells.
Immune CS-effector cells (7 × 107) were incubated with 5 × 107 unseparated regulatory T cells, with an equivalent number of separated magnetic bead positive γδ cells, or with an equivalent number of magnetic bead nonadherent γδ− cells (i.e., αβ TCR+-remaining cells) originating from either tolerized 129/J α+/+ mice or TCRα−/− mice for 30 min at 37°C. The washed effector and regulatory cell mixture was then injected intraperitoneally into naive 129/J recipients that were tested for CS responses by challenging ears with 0.8% PCl or OX in olive oil the next day. Recipients were then measured for ear swelling 24 h after Ag challenge to the ears.
In Vitro Culture of CS-effector Cells and Hapten-conjugated APC for Elaboration of IFN-γ, and Regulation by Tolerized Cells.
A single cell suspension of just CS-effector lymph node cells was obtained aseptically from PCl contact sensitized mice. Suspensions of regulatory cells were from spleen cells of mice that were injected with TNBSA or OX-MRBC intravenously on days 0 and 3, or from control nonimmune mice. In some experiments, spleen cells from TNBSA-tolerized mice, or nonimmune controls, were incubated with 1 μg biotinylated anti-γδ TCR (GL3; PharMingen) per 106 cells at 2 × 107 cells/ml for 20 min on ice. Then, 10 μl of streptavidin-conjugated magnetic microbeads (Miltenyi, Sunnyvale, CA and BioTek Instrs. Inc., Sunnyvale, CA) per 107 cells were added and incubated for 15 min on ice. After washing, this cell and bead suspension was applied to a MiniMACS magnetic separation column (Miltenyi). After recovering of nonadherent γδ-depleted cells and washing the column, γδ-enriched adherent cells were eluted. Equivalent numbers of γδ-enriched, or γδdepleted cells were added in vitro as regulatory cells, to the CSeffector cells (1:1; 2 × 105/well), and then were added together for 72 h to mitomycin C–treated TNP- or OX-conjugated normal spleen cells as APC (2 × 105) in 0.2 ml of RPMI 1640 containing 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 25 mM Hepes, 5 × 10−5 M/2-mercaptoethanol, and 10% fetal calf serum. At 48 h, culture supernatants were collected for IFN-γ quantitation.
ELISA Quantitation of IFN-γ in Culture Supernatants.
Quantitative sandwich ELISA for IFN-γ was performed with two different mAb specific for mouse IFN-γ (clones R4-6A2 and XMG1.2; Pharmingen). 1 μg/ml of capture mAb to IFN-γ in 0.1 M Na2CO3, pH 8.3, was briefly coated onto 96-well microtrays (Corning Glass Incorporated, Corning, NY) overnight at 4°C. After blocking with 3% dry milk in PBS, samples and standard recombinant mouse IFN-γ (Genzyme Corp., Cambridge, MA) were applied to the wells and incubated overnight at 4°C. Then, 0.5 μg/ml of biotinylated separate mAb to IFN-γ was applied for 45 min at 25°C, followed by incubation with 1:3,000 dilution of horseradish peroxidase–conjugated streptavidin (Vector Labs., Inc., Burlingame, CA), for 30 min at 25°C. TMB (tetramethylbenzidine) microwell peroxidase substrate, and TMB one component stop solution (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, MD) were used for reaction development and were read at OD 405.
Double-tailed Student's t test was used to assess the significance of differences between groups, with P <0.05 taken as a minimum level of significance.
TCRα−/− Mice Have No 24 h CS Responses.
CS responses in contact sensitized TCRα−/− mice and control 129/J mice were compared by sensitizing (via abdominal skin painting) with concentrated PCl (5%), followed 4 d later by topical challenge with dilute PCl (0.8%) applied to the ears, and subsequent measurement of ear swelling 24 h later. It was found that αβ T cell–bearing 129/J mice developed a classical 24 h ear swelling response (Fig. 1, group A), whereas TCRα−/− mice did not (Fig 1, group B).
When α−/− mice were used as recipients of αβ+ cells from MHC-matched, immunized αβ T cell–sufficient BDF1 mice, they could mount CS upon ear challenge with hapten (Fig. 1, group D), similar to normal mice (Fig. 1, group C). This demonstrated that the vascular and cellular environment for eliciting CS was normal in TCRα−/− knockout mice. The finding that the 24 h component of CS cannot be elicited in actively sensitized TCRα−/− mice is consistent with other studies indicating that the active CSeffector cell type is TCRαβ+ (7, 8, 24–26).
High Dose Ag Tolerance of TCRα−/− mice Induces Regulatory Spleen Cells that Downregulate CS.
Because TCRα−/− mice do not mount CS responses, it is impossible to determine directly whether high dose Ag tolerization of such mice induces downregulatory cells. Instead, CS-effector cells (7 × 107) from contact-sensitized 129/J mice were mixed and incubated for 30 min at 37°C with 5 × 107 spleen cells from α−/− mice, that were harvested on d 7 after two intravenous injections of tolerogenic TNBSA. Analogous TNBSA-tolerized cells were derived from TNBSA-treated MHC haplotype-matched TCRα+/− mice, or TCRα+/+ (129/J) mice. The cells were then transferred back to haplotype-matched recipient mice that were challenged with hapten.
Fig. 2 shows that upon transfer back to recipient mice, CS-effector cells elicit a 24 h response that can be significantly reduced by preincubation of the effector cells with cells from tolerized TCRα+/+ mice (Fig. 2, group B), TCRα+/− mice (Fig. 2, group C), or TCRα−/− mice (Fig. 2, group D). Essentially the same results were obtained when effector cells from sensitized BDF1 mice were mixed with cells from either TNBSA-tolerized BDF1 mice or TNBSA-tolerized TCRα−/− mice (Fig. 2, groups E–G).
A prediction that follows from the data shown in Fig. 2 (groups A–G) is that TCRα−/− mice will not support the actions of adoptively transferred TCRα+/+ CS-effector cells if the recipient α−/− mice are subject to high dose Ag tolerance beforehand. The data in Fig. 2, groups H and I confirm this prediction. This result is important in that it rules out that the detection of TCRα−/− downregulatory cells in adoptive transfer experiments (Fig. 2, groups D and G) was artefactual, due to manipulation of the cells in vitro and their subsequent transfer. Furthermore, induction of high dose tolerance by TNBSA in TCRα−/− mice was reproducibly demonstrated using transfers from individual TCRα− mice (data not shown), indicating the generality of the findings.
Expression of γδ TCR on CS-regulatory Cells from Tolerized TCRα−/− Mice.
To determine the phenotype of CS-regulatory cells induced by high Ag dose tolerance, splenocytes from TNBSA-tolerized mice were fractionated by use of anti-TCR γδ mAb-coated magnetic beads. γδ+, γδ−, or unfractionated tolerized cells were then mixed for 30 min at 37°C with lymph node and spleen CS–effector cells (7 × 107) from contact immunized, haplotype–matched mice, and the mixture transferred to recipients that were subsequently assayed for CS responses by ear challenge. In these experiments, anti-γδ bead adherent cells ranged from 55 to 95% γδ TCR+ when reanalyzed by FACS® (see Materials and Methods). Thus, we estimate that ⩾106 γδ cells were adoptively transferred into each recipient when unseparated TCRα−/− cells were used, and about ⩾7 × 105 γδ cells were transferred when magnetic bead γδ–enriched cells were used.
The data in experiments 1 and 2 of Fig. 3 show that while the cells from TNBSA-treated α+/+ 129/J mice that are responsible for downregulating CS effectors are exclusively TCR γδ− (group D), the most potent regulatory cell fraction from tolerized TCRα−/− mice scored as TCR γδ+ (Fig. 3, group F). Similar results were obtained in experiment three, in which cells from H2d TCRα−/− mice were compared with cells from H2d BALB/c mice for their response to high dose Ag tolerance. As before, when measuring the capacity of tolerized cells to downregulate a CSeffector response to PCl, the TCR γδ+ fraction from TCRα−/− mice, but not from α+/+ BALB/c mice, was again active (experiment 3, group C). Conversely, the active fraction from normal α−/− mice was TCR αδ+ (group D), and the TCR γδ− fraction from TCRα−/− mice (group E) contained negligible residual activity.
γδ T Cell–mediated Regulatory Activity Shows Antigen Specificity.
The discovery that γδ T cells in TCRα−/− mice can mediate downregulation of CS effector cells prompted an investigation of Ag specificity, since γδ T cells and αβ cells appear to have a different set of Ag specificities. To test this, cells from TCRα−/− mice tolerized with high doses of OX or TNP, were mixed with CS-effector cells from TCRα+/+ mice sensitized with PCl or OX, and compared in their capacities to downregulate the 24 h CS response after transfer back to recipient mice. Data presented in Fig. 4,a show no capacity of cells from OX-tolerized TCRα−/− mice to influence the CS response to TNP (group B), whereas TNP responses were inhibited significantly after mixing effector cells with cells from TNP-tolerized TCRα−/− mice (Fig. 4,a, group C). Data in Fig. 4 a, groups D–F demonstrate that reciprocal Ag specificity was likewise apparent (i.e., cells from OX-tolerized TCRα−/− mice downregulated effector cells from OX-sensitized mice, but not α+/+ TNP–immune cells) (group E versus F).
To confirm the above findings we tested Ag-specificity in tolerized α−/− mice by attempting to transfer α+/+ CS-effector cells into these recipients. α−/− mice were tolerized with a high dose of TNBSA or OX-MRBC intravenously before cell transfer. Then, on day 7, control (nontolerized) or TNP or OX-tolerized α−/− mice received 7 × 107 TNP or OX immune cells from normal α+/+ (BALB/c) mice. The data presented in Fig. 4,b show there was no ability to adoptively transfer immunity into TNP tolerized α−/− mice with TNP-specific CS-effector cells (group B), while TNP-immune CS-effector cells transferred normal CS responses into OX-tolerized α−/− recipients (group C). Similar data were obtained when OX-specific immune cells were transferred (group D versus F). Thus, total inhibition of adoptive cell transfer was obtained when OXspecific immune CS-effector cells were transferred to OXtolerized α−/− mice (Fig. 4,b, group E), whereas OX CS was successfully transferred to TNP-tolerized α−/− mice (Fig. 4 b, group F). Thus, this system that did not involve cell harvest, cell mixing, nor cell transfer, confirmed the presence of active Ag-specific downregulation in high dose–tolerized α−/− mice.
γδ T Cell–mediated Regulatory Activity is MHC Unrestricted.
To determine possible MHC restriction of γδ downregulatory T cells, magnetic bead-enriched γδ TCR+ cells from TNBSA-tolerized α−/− mice were incubated with 7 × 107 PCl-immune CS-effector cells from MHC-compatible (129/J, H2b), or -incompatible (CBA/J, H2k) mice, for 30 min at 37°C, followed by subsequent transfer to recipients syngeneic with the CS-effector cells. The recipient mice were challenged immediately (before any significant host response to the allogeneic γδ graft could develop), and ear swelling responses determined. The results in Fig. 5 show that γδ+ cells from TNBSA-tolerized TCRα−/− mice downregulated CS-effector responses of allogeneic CBA/J H2k CS-effector cells (group D), as effectively as the regulation of H2-matched CS-effector cells (group B).
Phenotype and dose–response Activity of CS-regulatory γδ T Cells.
To establish the phenotype of γδ T cells capable of inhibiting CS-effector responses, spleen cells from TNBSA-tolerized α−/− mice were treated with anti-CD4 or anti-CD8 mAb or PBS, and then rabbit complement 1:75. The three resultant populations of TCRα−/− cells were then individually incubated with 129/J immune CS-effector cells for 30 min at 37°C, and each mixture of cells was subsequently assayed for CS-effector function after transfer to naive recipients. The results in Fig. 6 a demonstrate once more that TNBSA-treated TCRα−/− mice are a reliable source of downregulation of CS-effectors (group B), and that this activity is unaffected by treatment with either antiCD4 (group C), or anti-CD8 (group D). Results are consistent with the active regulators being double negative (DN) γδ T cells.
DN γδ T cells have been reported in the lymphoid system of essentially all vertebrates examined, but they occur in mice in very low numbers relative to αβ T cells. Therefore, to assess the potency of DN γδ T cells to downregulate CS, a dose–response experiment was performed. 10fold dilutions (2.5 × 105, 2.5 × 104, or 2.5 × 103) of 60% γδ TCR+ cells (purified from TNBSA-tolerized α−/− mice by use of magnetic beads) were incubated, each with 7 × 107 PCl-immune CS-effector 129/J cells for 30 min at 37°C, and then were harvested, washed, and transferred. Recipient ear swelling results (Fig. 6 b) showed a strong dose response of the assay to the inocula (groups B–D). Even the dose of 2.5 × 103 γδ T cells had activity in the assay (group D), albeit weaker than the other doses. The capacity of very small numbers of downregulatory γδ T cells to score in this bio-assay is discussed below.
γδ Cells From TNBSA-tolerized TCRα−/− Mice Inhibit In Vitro Production of IFN-γ by TNP-specific CS–effector T Cells.
Since IFN-γ is a central cytokine in Th-1–mediated responses, such as DTH and CS, we determined whether IFN-γ production by CS-effector TNP-immune cells was affected by regulatory cells induced by high dose intravenous TNBSA or OX tolerogenesis. Using two different anti–IFN-γ antibodies in an ELISA assay, it was first established that 2 × 105 lymph node cells from contact immunized mice, stimulated in vitro by TNP-conjugated APC, secreted >3 ng/ml IFN-γ over a 48 h period. Compared to production of IFN-γ by CS-effector cells mixed with nonimmune spleen cells (Fig. 7,a, groups A and D), IFN-γ production by CS-effector cells and TNP-APC was substantially inhibited by mixing with spleen cells from intravenous TNBSA tolerized normal α+/+ mice (Fig. 7,a, group B) and also with spleen cells from intravenous TNBSA tolerized α−/− mice (Fig. 7 a, group E). Also, as controls, spleen cells from nonimmune and nontolerized animals incubated alone or in the presence of TNP or OX-APC did not secrete detectable levels (0.1 ng) of IFN-γ (data not shown).
To determine the phenotype of regulatory cells responsible for the in vitro suppression of IFN-γ production, γδ TCR+ or γδ TCR− cells in the spleens were separated, using anti-γδ antibody coated magnetic beads that were used to separate spleen cells from nontolerized, TNBSA-tolerized, or OX-tolerized TCRα+/+ or TCRα−/− mice, and then equivalent numbers of each cell subpopulation were added to TNP-immune CS–effector cells in vitro, and cultured together with TNP-hapten–conjugated APC. There was a striking suppression of IFN-γ production by γδ+ cells from TNBSA-tolerized α−/− mice (Fig. 7,b, group E, left) that did not occur with γδ+ cells from TNBSA-tolerized α+/+ mice (Fig. 7,b, group B, left). In contrast, the inhibitory activity of TNBSA-tolerized α+/+ mice was shown to be due to γδ− cells (Fig. 7,b, group B, right), and thus was probably due to αβ T cells. In contrast to the downregulatory activity of γδ+ cells from TNBSA-tolerized α−/− mice, the γδ− cells of these mice were totally without inhibitory activity (Fig. 7,b, group E, right), as were OX-tolerized γδ+ and γδ− cells (Fig. 7 b, group C and F, left), again varifying Ag-specificity of γδ+ cells mediating tolerance.
These in vitro findings were consistent with the in vivo findings. Thus, we concluded that suppressive cells in Agtolerized α−/− mice were potent Ag-specific γδ downregulatory T cells, both in vivo with contact sensitivity, and in vitro with IFN-γ production, while T cells that inhibited CS in vivo and IFN-γ in vitro that were obtained from normal α+/+ mice, were not γδ T cells, and probably were αβ T cells. The failure of most cell mixtures tested in Fig. 7 b to reduce IFN-γ production demonstrated that selected reductions in IFN-γ production were not readily attributable to nonspecific effects of extra cells in the assay, such as binding of IFN-γ to IFN-γ receptors, or to degradation of secreted IFN-γ.
αβ T cell–deficient mice do not elicit the classical 24 h component of CS after active contact sensitization and subsequent skin challenge. This result extends previous findings implicating αβ T cells as principal effectors in CS responses (7, 8, 25, 26): i.e., αβ T cells and γδ T cells are not redundant in their functional capacity as effector cells in CS. By contrast, TCRα−/− mice can respond to high dose Ag tolerance by the induction of cells that can downregulate CS-effector activity from α+/+ immunized donors. This was shown in two ways: splenocytes from tolerized TCRα−/− mice downregulate CS-effector cells with which they are mixed in vitro, before cotransfer and elicitation of CS, and high dose Ag-treated tolerized TCRα−/− mice inhibit the transfer of CS-effectors that function normally in untreated TCRα−/− mice.
The experiments presented here demonstrate that γδ cell– enriched fractions from TCRα−/− mice are better mediators of high dose Ag tolerance than are γδ cell–depleted fractions, suggesting that γδ T cells are the active regulatory cells, as opposed to TCRβ+α− cells that are a potential feature of TCRα−/− mice (27–28). The lack of conventional MHC restriction shown by the TCRα−/− regulatory cells, and their CD4−CD8− phenotype are both consistent with most assays of γδ function (29). The apparent hapten specificity of the regulatory γδ cells demonstrated herein for the first time may be considered consistent with the hypothesis that the potential diversity (30) of the γδ TCR is greater than that of any other Ag receptor (31). Thus, overall, downregulation can be mediated by Ag-specific γδ T cells, and need not be solely mediated by specific αβ T cells. However, in this study we have provided no evidence that the γδ cells directly recognized either PCl/TNP or OX, nor whether the γδ cells recognized either conventional APCs, or the effector T cells with which they were mixed, or some other cells. Conceivably all of these cells may have been haptenated.
Commonly, it has been argued that TCRα−/− mice are deficient in immune effector responses because αβ cells are uniquely capable of specifically responding to the challenging antigen. However, the apparent capacity of γδ cells in TCRα−/− mice to distinguish between CS-effector cells that are responsive to different hapten Ag raises questions about the completeness and/or validity of this explanation. Possibly, potential γδ effector cells responsive to challenge by specific Ag may be unable to mount strong immune effector responses because they cannot interact efficiently with the professional antigen presenting cell system. This may be for physiologic reasons; e.g., γδ cells may occur in an inappropriate anatomical location, in insufficient numbers, or produce inappropriate effector molecules. However, equally attractive is the idea that γδ cells, like B cells, only rarely recognize molecular complexes of antigenic peptides and MHC. In such a case, γδ cells would not “learn” that the body was infected from APCs and could not be expected to mount a conventional immune response of the kind that we associate with APC-αβ T cell interactions (31).
Provocatively, the data presented here bear striking similarities with the role of γδ cells reported in another system, namely the IgE allergic responsiveness of mice to protein antigen aerosols (18, 19). Thus γδ T cells reduced ovalbumin-specific IgE responses in rats and mice that were tolerized with ovalbumin administered via repetitive respiratory aerosol administration (18, 19). Not only were γδ T cells shown to be active in both cases, but strikingly low numbers of γδ T cells per animal (∼2 × 103) were effective, as also was the case here (Fig. 6 b). Likewise, systemic oral tolerance to ovalbumin was suggested to be due to γδ T cells (32). Recently it was also shown that mice deficient in γδ T cells, and responding to infection with a natural intestinal pathogen, Eimeria, had exaggerated gut immunopathology attributable to unrestrained αβ T cell function, or the effects thereof, suggesting a downregulatory function of γδ T cells (33). In other studies, γδ T cells from T. cruzi infected mice were shown to be suppressive in vivo and in vitro, correlating with autoimmunity (34). Finally, hepatic γδ T cells could adoptively transfer tolerance to allogeneic skin grafts, and also reduced lymphokine production (35). γδ T cells in tumor bearing mice were reported to decrease anti-tumor cytotoxic T cell activity (36). Thus, the data presented here, in the classical CS system, provide a clear demonstration of a fundamental functional phenotype of γδ cells that is operative in numerous systems, including allergy to aerosolized antigen, oral tolerance, autoimmunity, transplantation, tumor immunity, and pathogenesis of infectious diseases. As such, the downregulation of αβ T cells by the effects of γδ T cells is emerging as a major and probably general functional capacity.
The downregulatory effect of γδ T cells that we describe may in part be effected by a direct or indirect inhibition of IFNγ production by the CS-effector cells, as our results suggest. The mechanism for this inhibition has yet to be resolved. An obvious possibility is the production by γδ T cells of cytokines that inhibit T cell production of IFN-γ, such as TGFβ, IL-4, or IL-10. The production by γδ T cells of IL-4 would be consistent with the capacity of γδ T cells from TCRα−/− mice to support copious production of antibodies (22). Thus, regulation of CS by γδ T cells may be another instance of Th2 regulation or diversion of Th1 cells (37). Alternatively, γδ T cells may regulate CS-effector αβ T cells by their being eliminated via injury, apoptosis, or cytotoxic lysis. There are many reports that γδ T cells possess cytotoxic activity (38, 39). However, it is noteworthy that we have recent evidence that in vitro tolerance for IFN-γ production by splenocytes of TNBSA-tolerized mice can be reversed by IL-12 (Ushio, H., and P.W. Askenase, unpublished results), suggesting that cytotoxicity may not be the mechanism for γδ T cell-mediated TNBSA tolerance.
In summary, DTH and CS are examples of important in vivo Th1 responses that can play a crucial role in defense against diverse microbial pathogens and tumor cells, and as effectors of some allergic and autoimmune diseases. This may be the reason for such tight regulation of these responses. It is possible that changes in immunoregulation could be a cause for decreased immune resistance in serious infections, or in cancers, or perhaps for the abnormally increased immune responses that occur in allergies and autoimmunity. The findings of the current study clearly establish that γδ T cells can function as a newly recognized component of immune downregulation, and may be able to do so in an Ag-specific fashion, the underlying mechanism of which needs now to be defined.
The authors are especially grateful to Scott Roberts (University of Connecticut, Storrs, CT), Topher Dudley (National Institutes of Health), and Adrian Smith (Yale University, New Haven, CT) for review of the manuscript, and to Marilyn Avallone for her excellent secretarial skills.
This work was supported by National Institutes of Health grants AI-12211, AI-26639, and AI-02174 to P.W. Askenase, and AI-27855 to A.C. Hayday, the Polish Committee of Scientific Research, Maria Sklodowska-Curie Fund II (Polish American Agreement) to W. Ptak, and the Markey Foundation to L.R. Anderson.
1Abbreviations used in this paper: DN, double negative T cells (CD4−, CD8−); CS, contact sensitivity; DTH, delayed-type hypersensitivity; MRBC, mouse red blood cells; OX, oxazolone; PCl, picryl chloride; TMB, tetramethylbenzidine; TNBSA, Trinitro-benzene sulfonic acid.
Address correspondence to Philip W. Askenase, Section of Allergy and Clinical Immunology, Department of Internal Medicine, Yale University School of Medicine, 333 Cedal Street, New Haven, CT 06520-8013.