IFN-γ Promotes Generation of IL-10 Secreting CD4⁺ T Cells that Suppress Generation of CD8 Responses in an Antigen-Experienced Host

Mice

Four- to eight-week-old adult female C57BL/6 (H-2^b) mice were purchased specific pathogen free from the Animal Resource Centre (Perth, Australia). CD1d knockout (CD1d^−/−) mice, and IFN-γ receptor knockout mice (IFN-γR^−/− mice), each on at C57BL/6 background, were sourced from The Jackson Laboratory and bred in the biological resources facility at Princess Alexandra Hospital. Mice were kept under specific pathogen free conditions throughout, and all experiments were approved by and performed in compliance with the guidelines of the University of Queensland Animal Experimentation Ethics Committee.

Cell lines, peptides, and Abs

Spodoptera frugiperda (Sf-9) cells (Life Technique) were maintained in Sf-900 II medium with Sf-9 II Supplement (Life Technique) and 10% FBS (CSL) at 27°C. Anti-IL-10 receptor (anti-IL-10R) hybridoma (1B1.3a) (10) was provided by Dr. Warwick Britton of Centenary Institute, University of Sydney and was maintained in RPMI 1640 (Invitrogen) with 10% FBS.

Production of anti-IL-10 mAb has been described elsewhere (10). In brief, hybridoma cells were cultured in RPMI 1640 with 1% FBS for 72 h; supernatants were collected and passed through a Protein G column (Sigma-Aldrich) and eluted by running 100 mM glycine (Sigma-Aldrich) through the column. Eluted Abs were dialyzed extensively against PBS (0.15 M NaCl; 0.02 M PO₄ (pH 7.4)) and the concentration of Ab was measured and bioactivity was assessed by an in vitro suppression assay as previously described (9).

Anti-CD4 mAb (GK1.5) and anti-CD8 mAb (2.43) were produced from mouse ascites. Anti CD4-FITC mAb (RM4-4), anti-IL-10 mAb (JES5-16E3), anti-CD4-PE (RM 4-5), anti-GITR-PE (RAM34), anti-CCR5 (7A4), anti-CCR7 (4B12), were purchased from eBioscience. Anti-IFN-γ (R4-6A2), anti-IFN-γ α-chain (GR20) was purchased from BD Pharmingen. Anti-IFN-γ β-chain (MOB-47) and anti-CD3 (145-2C11) was purchased from Bio-Legend.

IFN-γ was purchased from eBioscience. OVA, zymosan, and LPS were purchased from Sigma-Aldrich.

The MHC class I (H-2 D^b) restricted HPV16 E7 peptide RAHYNIVTF (20) was synthesized and purified by Chiron Mimotopes.

Production of recombinant virus-like particles (VLPs)

VLP of various papillomaviruses were produced using recombinant baculoviruses encoding the L1 protein of bovine papillomavirus (L1 VLPs) or a synthetic fusion protein comprising the L1 protein of bovine papillomavirus and a cytotoxic T cell epitope of HPV16 E7 protein (RAHYNIVTF; single letter amino acid code) (L1E7 VLPs) as previously described (21). VLPs were purified from the nuclei of SF9 cells infected with recombinant baculovirus by CsCl gradient centrifugation as previously described (22). Samples were subjected to analysis by transmission electron microscope and immunoblotting to confirm the identity and integrity of the VLPs. For immunoblot analysis, protein samples were diluted in SDS-PAGE sample buffer, electrophoresed through a 10% SDS-PAGE gel, and transferred to nitrocellulose membrane. The membrane was probed with the anti-L1 mAb MC15. Bound Ab was detected by incubation of the membrane with HRP-conjugated sheep anti-mouse Ab (Silenus) and visualized using ECL (Amersham Biosciences). For electron microscopy, CsCl gradient-purified and dialysed VLP samples were mounted onto carbon-coated grids, stained with 2% ammonium molybdate (pH 6.2), and examined with a Hitachi H-800 electron microscope.

Immunization of mice

Groups of three to five mice were immunized as indicated with 30 or 50 μg of VLPs i.m. without any adjuvant; or were immunized with 100 μg of OVA and 30 μg of LPS i.p. Mice were lightly anesthetized with Isofluorane (Abbott).

ELISA for IL-10 and IFN-γ cytokines from culture supernatants

ELISA for IL-10 and IFN-γ (R&D Systems) was performed as described previously, according to the manufacturer’s recommended procedures (9).

ELISPOT

ELISPOT was performed as described (23). In brief, single spleen cell or lymph node suspensions were added to membrane base 96-well plates (Millipore) coated with anti-IFN-γ (BD Harlingen). Peptide was added at various concentrations and cells held at 37°C with peptide for 18 h. Ag specific IFN-γ secreting cells were detected by sequential exposure of the plate to biotinylated anti-IFN-γ (BD Harlingen), avidin-HRP (Sigma-Aldrich) and diaminobenzidine (Sigma-Aldrich). The IFN-γ secreting cells were quantitated by ELISPOT reader system ELR02 (AID Autoimmun Diagnostica).

Positive selection of mouse spleen CD11c⁺ cells

Positive selection of mouse spleen CD11c⁺ cells has been described previously (10). In brief, C57 BL/6 or IFN-γR^−/− mouse spleens were held in 1 mg/ml Collagenase D (Roche Diagnostics) and 500 μl of Collagenase D was injected into each spleen. The spleens were then cut in smaller pieces, held in 5 ml of Collagenase D for 45–60 min at 37°C, and passed through a steel mesh. The cells were counted, washed in RPMI 1640 with 2% FBS, and resuspended in 400 μl RPMI 1640 with 2% FBS per 10⁸ total cells. One hundred microliters of MACS CD11c Microbeads (Miltenyi Biotec) were added, and held for 15 min at 6–12°C. After washing, cells were resuspended in 500 μl per 10⁸ cells. CD11c⁺ cells were positively selected with a LS column (Miltenyi Biotec) according to the manufacturer’s protocols. The purity of CD11c⁺ cells was around 80% as assessed by flow cytometry.

Sorting of CD4⁺ T cells

To generate a lymphocyte population enriched for Ag primed CD4⁺ T cells, mice were immunized with L1VLPs twice, and draining inguinal lymph nodes were removed 7 days after the second immunization. Cells were passed through a 70-μm nylon membrane (BD Harlingen) and resuspended in 1 ml RPMI 1640 plus 2% FBS. The cells were stained for CD4 and GITR expression. CD4⁺GITR⁺, or CD4⁺GITR⁻ cells from draining, nondraining lymph nodes of C57 BL/6 or IFN-γR^−/− mouse were sorted by using high-speed sorter MoFlo Flow (DakoCytomation). Sorted cells had >95% purity and the CD11c⁺ cells within the sorted cells were <2%.

In some experiments, CD4⁺ T cells were negatively enriched and then CD4⁺GITR⁺ T cells were positively selected by using MACS beads following the protocols from manufacturer. The purity of CD4⁺GITR⁺ T cells was around 90%.

Intracellular staining for Foxp3 and cytokines

Intracellular staining of lymphocytes from spleen and lymph nodes Foxp3 was performed using an Foxp3 intracellular staining kit from eBioscience following the instructions provided.

For the intracellular staining of cytokines, single cell suspensions were made from spleen or lymph nodes from immunized mice and stimulated with 25 ng/ml PMA and 1 μg/ml ionomycin for 3–5 h in the presence of monensin (BioLegend). For some experiments, the cells were cultured for 72 h in the presence of anti-CD3 and equal number of CD11c⁺ cells and then stimulated with same concentration of PMA and ionomycin for 5 h. After stimulation, intracellular staining for IL-10 and IFN-γ was performed using Per/Fix reagents (BD Pharmingen).

Statistical analysis

Statistical analysis was performed using the two tailed Student’s test, using Prism 4.0 (GraphPad Software).

Ag-experienced CD4⁺GITR⁺ T cells secrete IL-10 upon antigenic stimulation

Animals primed to Ag by immunization are inhibited from producing IFN-γ-secreting CD8 T cells in response to further immunization with Ag, and inhibition requires primed CD4 T cells and IL-10. To identify which population of Ag experienced CD4⁺ T cells produce IL-10 and suppress induction of effector cells in Ag primed mice, we immunized mice with bovine papillomavirus 1 L1 virus like particles (L1 VLPs) and compared CD4⁺ T cells from lymph nodes draining the site of immunization with cells from non draining lymph nodes for GITR and Foxp3 expression. The percentage of CD4 cells expressing GITR was similar between draining and nondraining lymph nodes, while Foxp3 expression was slightly lower in draining lymph nodes (Fig. 1⇓A). However, while equal numbers of CD4⁺GITR⁺ T cells expressed or did not express FoxP3 in the draining lymph nodes, in non draining lymph nodes the majority of CD4⁺GITR⁺ T cells were also FoxP3 positive (Fig. 1⇓B). These results suggested that although the number of CD4⁺GITR⁺ T cells was similar between draining and nondraining lymph node, their function might be different.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Ag-experienced CD4⁺GITR⁺ T cells express more FoxP3 and secrete higher levels of IL-10 than naive CD4⁺GITR⁺ T cells upon antigenic stimulation. C57BL/6J mice were immunized with bovine PV L1 virus like particles on days 0 and 14, and draining and nondraining lymph nodes harvested on day 20. A, Expression of Foxp3 and GITR on CD4 cells from draining and nondraining nodes. B, Expression of FoxP3 on CD4⁺ GITR⁺ cells by triple staining flow cytometry (left) on draining and nondraining nodes (right). C, FACS sorted CD4⁺GITR⁺ T cells from draining and nondraining nodes, and from unimmunized animals, were assessed for IL-10 production after culture with or without membrane bound anti-CD3 Ab. D, CD11c⁺ DCs (0.5 × 10⁵), previously exposed to 40 μg/ml bovine PV L1 virus-like particles overnight or left unexposed, were cocultured overnight with 0.5 × 10⁵ of FACS sorted CD4⁺GITR⁺ or CD4⁺GITR⁻ T cells from draining and nondraining lymph nodes of bovine PV L1 virus like particle immunized mice. IL-10 level from supernatants were measured by ELISA. Results throughout are the mean and 1 SEM of triplicates and represent one of two independent experiments. NT in C and D means not tested.

To study the capacity of Ag-experienced and naive CD4⁺GITR⁺ T cells to secrete IL-10 in response to stimulation, CD4⁺GITR⁺ cells from unimmunized mice, and from draining and nondraining lymph nodes from VLP immunized mice, were cultured in vitro with or without anti-CD3 Ab, and IL-10 produced in culture supernatants was measured by ELISA. After stimulation by anti-CD3, CD4⁺GITR⁺ T cells from draining lymph nodes secreted higher levels of IL-10 (Fig. 1⇑C) than CD4⁺GITR⁺ T cells from naive or nondraining lymph nodes. To investigate whether Ag-specific stimulation generated similar results, CD11c⁺ cells were pulsed with L1VLPs and then cocultured with GITR⁺CD4⁺ T cells from draining or nondraining lymph nodes, or GITR-CD4⁺ T cells from draining lymph nodes. CD4⁺GITR⁺ T cells responded to cognate Ag stimulation similarly to anti-CD3 stimulation: GITR⁺ cells from draining lymph nodes produced high level of IL-10, unlike GITR⁺ cells from nondraining lymph nodes, and GITR⁻ cells from draining lymph nodes (Fig. 1⇑D). These data demonstrate that CD4⁺GITR⁺ T cells induced by immunization and responsive to cognate Ag are the main producers of IL-10 upon repriming, indicating these cells might be the cells that mediate in vivo suppression of induction of IFN-γ secreting CD8 T cells following reimmunisation.

Ag-experienced CD4⁺GITR⁺ T cells secrete high levels of IFN-γ, and IFN-γ signaling is required for inhibition of induction of IFN-γ secreting CD8 T cells

Recently, IFN-γ and IL-10 secreting CD4 T cells have been show to have regulatory function in parasitic infection (24, 25). We therefore investigated whether IFN-γ signaling was required for inhibition of induction of IFN-γ secreting CD8 T cells in an Ag primed animal. We first determined whether Ag experienced CD4⁺GITR⁺ T cells produce IFN-γ upon TCR stimulation. Ag-experienced and naive CD4⁺GITR⁺ T cells were stimulated with anti-CD3 Ab, and supernatants were assessed for IFN-γ, TNF-α, IL-6, and MCP-1. Neither naive nor Ag-experienced CD4⁺GITR⁺ T cells produced IL-6, while only Ag-experienced CD4⁺GITR⁺ T cells produced IFN-γ and TNF-α. (Fig. 2⇓A). Intracellular staining results show that there is a fraction of CD4⁺GITR⁺ T cells secreting both IL-10 and IFN-γ (Fig. 2⇓A, FACS data). To investigate whether IFN-γ contributes to inhibition of induction of IFN-γ secreting CD8 T cells in response to further Ag exposure in an Ag experienced animal, we tested whether such inhibition could be observed in IFN-γR^−/− mice. Naive IFN-γR^−/− mice generate slightly higher E7 specific IFN-γ CD8 T cell numbers in response to immunization than naive wild-type C57BL/6 mice (Fig. 2⇓B) after L1E7VLPs immunization. After priming with L1VLPs, and subsequent immunization with L1E7 VLPs, IFN-γR^−/− mice mounted significantly greater E7 specific IFN-γ secreting CD8 T cell responses than IFN-γR^+/+ mice similarly immunized, and these were similar in number to the CD8 response to L1E7 VLPs in naive animals (Fig. 2⇓B). Thus, IFN-γ signaling is inhibitory to the optimal induction of specific CD8 responses in Ag-primed mice and both IL-10 and IFN-γ signaling contribute to inhibition.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Ag-experienced CD4⁺GITR⁺ T cells secrete IFN-γ which can inhibit generation of specific CTL. A, left, CD4⁺GITR⁺ T cells from draining and nondraining lymph nodes of naive or bovine BPVL1 VLPs immunized mice on days 0 and 14 were FACS sorted and cultured at 0.5 × 10⁵ per well with or without membrane bound anti-CD3 Ab overnight, and supernatants assessed for cytokine secretion by multi cytokine array methods described in Materials and Methods. A, right, FACS results: C57BL/6 mice were immunized with virus like particles on days 0 and 14. Two days after final immunization, single cells from draining lymph nodes were directly stimulated with PMA and ionomycin for 5 h, stained for CD4 GITR, and then intracellularly stained for IL-10 and IFN-γ. CD4⁺GITR⁺ T cells were gated and intracellular IFN-γ and IL-10 shown. B, C57BL/6 and C57BL/6 IFN-γR^−/− mice were immunized with bovine BPVL1 virus like particles on days 0 and 14, or left unimmunized. On days 21 and 35, all mice were immunized with chimeric L1E7VLPs. Six days after final immunization, the number of E7 specific IFN-γ secreting CD8 cells per 10⁶ splenocytes was assessed by ELISPOT for all mice. Results are shown for each mouse individually and pooled from two independent experiments.

CD11c⁺ cells from IFN-γ receptor knockout but not wild-type mice secrete more IL-10 and less IL-12 upon stimulation with VLPs

We next wished to establish a mechanism by which IFN-γ might inhibit induction of CD8 T cells. Professional Ag presentation cells play a critical role in controlling the adaptive immune response, and can respond to IFN-γ. However, IFN-γ negatively regulates IL-10 secretion by macrophages and dendritic cells (DCs) (26, 27). To establish whether IFN-γ-mediated inhibition of induction of CD8 T cells might be mediated by a phenotypic change in DC, CD11c⁺ cells isolated from mouse spleen were exposed either to L1VLPs or to Zymosan, a TLR3 agonist, as a positive control, in the presence or absence of IFN-γ. CD11c⁺ cells from IFN-γ R^−/− mice, whether stimulated with VLPs or Zymosan, produced less IL-12 and more IL-10 than those from C57BL/6 mice (Fig. 3⇓A). To examine the contribution of IFN-γ to the response, CD11c⁺ cells were then stimulated with VLPs with or without IFN-γ. IL12p40 levels were higher if DCs were stimulated with VLP in the presence of IFN-γ (Fig. 3⇓B). DCs stimulated with Zymosan with or without IFN-γ showed similar results (data not shown). These results suggested that IFN-γ signaling through DCs would promote Th1 responses, and was therefore unlikely to be responsible for the observed IFN-γ mediated inhibition of induction of IFN-γ-secreting CD8 T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

CD11c⁺ cells from IFN-γR^−/− mice secrete more IL-10 and less IL-12 upon antigenic stimulation. A, CD11c⁺ DCs (10⁵) from C57BL/6 or IFN-γR^−/− mice were left exposed to 40 μg/ml bovine PV virus-like particles (VLP) or 50 μg/ml Zymosan for 18 h and supernatants were assayed for IL-10 and IL-12 by ELISA. Results are the means and 1 SEM of triplicates and represent one of two independent experiments. B, CD11c⁺ DCs (10⁵) from C57BL/6 mice were exposed to L1VLPs (40 μg/ml) and IFN-γ at differing concentrations as shown, or were left unexposed to VLPs (unpulsed). Culture supernatants were assayed for IL-12 by ELISA. Results are the means and 1 SEM of triplicates.

E7-specific IFN-γ inhibition can be observed in CD1d knockout mice

NKT cells are the main producer of IFN-γ at the early stage of an immune response. We therefore wished to establish whether IFN-γ secreted from NKT cells might contribute to the inhibition of induction of IFN-γ-secreting CD8 T cells observed in Ag-primed mice. We primed CD1d knockout mice, which lack CD1d restricted NKT cells, with L1VLPs and then further immunized these animals with L1E7VLPs. We measured E7 peptide-specific IFN-γ secretion by CD8 T cells. Naive CD1d^−/− mice developed E7-specific IFN-γ-secreting CD8 cells in response to immunization, and showed impaired responses when previously primed with L1VLPs in the same manner that CD1d replete animals did (Fig. 4⇓), demonstrating that CD1d restricted NKT cells and therefore IFN γ secretion by NKT cells are not necessary for inhibition of induction of CD8 T cells following priming.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

CD1d⁺ T cells are not required for inhibition of CD8 T cell responses following Ag exposure. Groups of C57BL/6 mice and C57 BL/6 CD1d^−/− mice were either left unimmunized or immunized with bovine PV virus-like particles on days 0 and 14. On days 21 and 35, mice were immunized with chimeric L1E7 bovine PV virus-like particles. Six days after the final immunization, the frequency of E7 specific IFN-γ secreting T cells in spleen was measured by ELISPOT assay. Results shown were from one of two independent experiments.

Viral capsid protein-experienced CD4⁺GITR⁺ T cells have higher IFN-γ receptor expression and optimal IL-10 secretion by these cells requires IFN-γ signaling

Inhibition of induction of IFN-γ-secreting CD8 T cells following priming requires Ag experienced CD4 T cells, IL-10, and IFN-γ signaling. We hypothesized that IFN-γ signaling to CD4⁺GITR⁺ T cells might be required for their acquisition of a suppressor phenotype, and that Ag experienced FoxP3⁺CD4⁺GITR⁺ T cells might express more IFN-γR than their naive FoxP3- counterparts. We first investigated the IFN-γ receptor α- and β-chain expression on CD4⁺GITR⁺ cells. The percentage of cells expressing IFN-γ α- and β-chain (Fig. 5⇓A) was higher in CD4⁺GITR⁺ cells in both draining and nondraining lymph nodes compared with CD4⁺GITR⁻ cells. The results were consistent from three independent experiments. The mean fluorescence intensity of IFN-γ α- or β-chain on CD4⁺GITR⁺ T cells was higher than on CD4⁺GITR⁻ T cells, whether from draining or nondraining lymph nodes (Fig. 5⇓A, right). Next, we identified Foxp3, IFN-γR β-chain and CD4 expression on cells from draining or nondraining lymph nodes of immunized mice. The percentage of CD4⁺Foxp3⁺ cells expressing IFNgR β-chain was similar in draining and nondraining lymph nodes (Fig. 5⇓B). We therefore investigated whether IFN-γ signaling to Ag experienced CD4⁺GITR⁺ T cells promoted IL-10 secretion by these cells. CD4⁺GITR⁺ T cells from C57BL/6 or IFN-γR^−/− mice were stimulated with anti-CD3 Ab; supernatants were measured for IL-10 level. As shown in Fig. 5⇓C, IL-10 secretion by CD4⁺GITR⁺ T cells from C57BL/6 mice was twice that seen from the same cell population from IFN-γR^−/− mice. Intracellular staining of cells for IL-10 confirmed these findings (Fig. 5⇓C, FACS results), and neutralizing IFN-γ reduced IL-10 secretion from GITR⁺ cells (Fig. 5⇓C). We confirmed these findings using Ag-experienced CD4⁺GITR⁺ T cells from OVA/LPS immunized C57BL/6 mice, and these also produced more IL-10 than cells from similarly immunized IFN-γR^−/− mice (214 ± 7pg/ml vs 137 ± 1pg/ml). Taken together, these results revealed that IFN-γ signaling through Ag-experienced GITR⁺CD4⁺ T cells, unlike IFN-γ signaling through DCs, regulates secretion of IL-10, the key cytokine for induction of the suppressor phenotype.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Ag-experienced CD4⁺GITR⁺ T cells have higher IFN-γ receptor expression and optimal IL-10 secretion by these cells requires IFN-γ signaling. C57BL/6J mice were immunized with bovine PV virus-like particles on days 0 and 14 or left unimmunised. Draining and nondraining lymph nodes cells were assessed (A) for CD4, GITR, and IFN-γ receptor α- and β-chain expression on day 20. Left, FACS profiles; Right, Mean fluorescence intensity for IFN-γ receptor α- or β-chain and (B) for CD4, Foxp3, and for IFN-γR β-chain on day 17. C, left, FACS-sorted Ag experienced CD4⁺GITR⁺ T cells were stimulated with anti-CD3 in the presence or absence of anti-IFN-γ Ab overnight, and supernatants were assayed for IL-10. Results are representing of two independent experiments. Right: Group of three C57BL/6 or IFN-γR^−/− mice were immunized with BPVL1 twice as indicated in the Materials and Methods, Ag-experienced CD4⁺GITR⁺ T cells were isolated by MACs beads and stimulated with membrane bound anti-CD3 for 18 h, IL-10 level from supernatants were measured by ELISA. Right, FACS profile, C57BL/6 or IFN-γR^−/− mice were immunized with VLPs twice as indicated in the Materials and Methods, Ag-experienced CD4⁺GITR⁺ T cells were isolated by MACs beads and stimulated with 1 μg/ml anti-CD3 and equal number of CD11c⁺ cells for 48 h, then stimulated with PMA and ionomycin for 5 h before performing intracellular staining for IL-10 and IFN-γ as described in Materials and Methods.

OVA/LPS immunization generate IL-10 secreting CD4⁺ T cells and neutralizing IL-10 overcome CTL inhibition in OVA/LPS-primed host

To investigate whether the CTL inhibition observed in VLP-primed mice could be demonstrated for a soluble protein Ag as well as a particulate viral Ag, mice were primed with OVA/LPS, and then immunized with OVA/LPS alone, or coadministered with anti-IL-10R Ab or normal rat serum. OVA CD8-specific IFN-γ responses were determined by ELISPOT assay. OVA-specific CD8 T effector cells were more frequent in OVA experience mice reimmunized with OVA/LPS, and also administered anti-IL-10R Ab, when compared with animals similarly immunized but not administered anti-IL-10R (Fig. 6⇓). As observed with BPV VLP Ag, OVA/LPS primed CD4⁺ T cells, especially CD4⁺GITR⁺ T cells, produced high amount of IL-10. Thus, whether examined for OVA or for bovine papillomavirus L1 protein, prior immunity to Ag inhibits generation of new CTL responses following further immunization, and neutralizing IL-10 at the time of Ag administration in an already Ag primed animal can overcome this inhibition.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Neutralizing IL-10 increases OVA specific CTL responses to immunization in OVA/LPS primed animals. A, Group of four C57BL/6J mice were primed with 100 μg of OVA with 30 μg of LPS on day 0, and then boosted with same amount of OVA/LPS on day 14 in the presence of 0.5 mg/ml anti-IL-10R Ab or rat serum i.p.; 6 days after final immunization, spleen from immunized mice were collected and ELISPOT assay performed. Results were pooled from two independent experiments. B, Two C57BL/6 mice were immunized i.p. with OVA/LPS as indicated in Materials and Methods; Ag-experienced CD4, CD4⁺GITR⁺, and CD4⁺GITR⁻ T cells were isolated by MACs beads and stimulated with anti-CD3 and equal number of spleen CD11c⁺ cells for 48 h before IL-10 level of the supernatants were measured by ELISA. Results throughout are the mean and 1 SEM of triplicates and represent one of two independent experiments.