The Combined Effects of Tryptophan Starvation and Tryptophan Catabolites Down-Regulate T Cell Receptor ζ-Chain and Induce a Regulatory Phenotype in Naive T Cells

Mice

Eight- to 10-wk-old DBA/2 (H-2^d) and 4-wk-old NOD (K^d, I-A^g7, D^b) mice were purchased from Charles River Breeding Laboratories. Gcn2^−/− mice, homozygous for the GCN2 targeted mutation, and wild-type Gcn2^+/+ controls were as described (26). NOD-SCID mice were bred in our animal facility under specific pathogen-free conditions to be used at ages 8–10 wk. Transgenic 8.3-NOD mice (27) and BDC2.5-NOD mice (28) were as described. Colorimetric strips were used to monitor glycosuria and hyperglycemia (>300 mg/dl) in NOD-SCID mice (4). Only female mice were used in this study, and all in vivo experiments were done in compliance with National and Perugia University Animal Care and Use Committee guidelines.

Reagents and treatments

The IDO inhibitor 1-methyl-dl-tryptophan (1-MT) was purchased from Sigma-Aldrich to be used in vitro at the concentration of 20 μM. To inhibit IDO activity in vivo, 1-MT or placebo pellets (150 mg/pellet, 7-day release at 0.9 mg/h; Innovative Research of America) were implanted as described (4). NRP-A7 (KYNKANAFL), a synthetic peptide mimotope recognized in the context of the MHC class I H-2K^d molecule by 8.3-CD8⁺ T cells (29), was as described (13). FITC-conjugated NRP-A7 pentamer and negative control peptide pentamer were prepared by ProImmune and validated by staining 8.3-NOD spleen cells according to manufacturer’s instructions. Pentamer-positive 8.3-CD8⁺ T cells were recovered by magnetic-activated sorting using anti-FITC MicroBeads (Miltenyi Biotec). Ab reagents for ELISAs of IL-2, IL-4, IL-10, and IFN-γ have been described (14). A specific TGF-β ELISA kit was purchased from Promega. Blockade in vitro of IL-10 and CTLA-4 was achieved by the respective use of SXC-1 (rat IgM) and 4F10 (hamster IgG) monoclonals (13). Purified anti-TGF-β 2G7 (mouse IgG2b) was a gift from M. Kopf (Swiss Federal Institute of Technology, Zurich, Switzerland). Neutralizing Abs were added at 10 (anti-IL-10) or 50 μg/ml (anti-CTLA-4, anti-TGF-β) and isotype controls were included in all assays. Standard IMDM (Invitrogen Life Technologies), containing 50 μM tryptophan, and low tryptophan (LT, 5 μM) medium (30), both with and without kynurenines (17), were used as culture media. Tryptophan and kynurenine concentrations were measured by HPLC (13).

Cell sorting and suppression assay

CD11c⁺CD8⁺ splenic DCs were purified by the sequential use of CD11c and CD8α MicroBeads (Miltenyi Biotec), as described (13), and were referred to as IDO⁻ DCs to distinguish them from IDO⁺ DCs, i.e., CD11c⁺CD8⁺ cells treated overnight with 200 U/ml IFN-γ so as to induce expression of IDO transcript, protein, and function (4). CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells were isolated from lymph nodes by magnetic-activated cell sorting as described (13, 14). The purity of either T cell fraction was >95%. For assay of regulatory activity, CD4⁺CD25⁻ cells were cocultured with irradiated T cell-depleted splenocytes and CD4⁺CD25⁺ cells for 3 days in the presence of soluble anti-CD3 (1 μg/ml) (14). Proliferation was measured by incorporation of [³H]thymidine according to standard procedures.

Flow cytometry, pentamer staining, and immunoblot analysis

In all FACS analyses, cells were treated with rat anti-CD16/32 (2.4G2) for 30 min at 4°C for blockade of Fc receptors before assaying on an EPICS flow cytometer using EXPO 32 ADC software (Beckman Coulter). The following FITC-labeled Abs were used in cell surface staining (BD Pharmingen): anti-CD4 (GK1.5), anti-CD69 (H1.2F3), and anti-CD45RB (16A). FITC-labeled anti-Foxp3 (FJK-16s) was from eBioscience. PE-labeled anti-CD62L (MEL-14; BD Pharmingen), anti-glucocorticoid-inducible TNFR (GITR, 108619; R&D Systems), and anti-CD4 (GK1.5; BD Pharmingen) were also used. Detection of the surface molecule B and T cell lymphocyte attenuator (BTLA) was accomplished by the use of biotinylated anti-BTLA (a gift from K. Murphy, Washington University, St. Louis, MO) and Streptavidin-FITC. Total CTLA-4 expression was analyzed as described (13). For pentamer staining, spleen and lymph node cells were incubated for 30 min at room temperature with FITC-conjugated H-2K^d pentamers before FACS analysis. For intracellular ζ-chain staining, cells were washed and fixed for 5 min with 2% paraformaldehyde at room temperature, permeabilized and incubated for 30 min with PE-labeled anti-CD3 (6B10.2; Santa Cruz Biotechnology). All FACS analyses included the use of isotype control Abs. For CFSE or CellTrace Far Red DDAO-SE (Molecular Probes) labeling, the fluorescent dyes in the form of 5 mM stock solutions were added to 2 × 10⁷ CD4⁺ T cells/ml to a final concentration of 2 (CFSE) or 1 μM (DDAO-SE). The cells were incubated at room temperature for 10 min and washed twice with the same volume of FCS and then with culture medium. Immunoblot of TCR ζ-chain and ε-chain involved total T cells or CD4⁺ and CD8⁺ fractions purified by magnetic-activated sorting (Miltenyi Biotec), which were analyzed as described (18), using monoclonal anti-CD3ζ (6B10.2) and polyclonal anti-CD3ε (M20; both from Santa Cruz Biotechnology) in combination with the appropriate Abs conjugated to HRP, followed by ECL.

Analysis of lymphocyte function

For mitogen-induced cell proliferation and cytokine induction, CD4⁺ or CD8⁺ T cells (1 × 10⁶/ml) were cultured in the presence of 1 μg/ml plate-bound anti-CD3ε (145-2C11), 2.5 μg/ml Con A (Sigma-Aldrich), or 2 ng/ml PMA (Sigma-Aldrich) combined with 0.2 μM calcium ionophore (Sigma-Aldrich). In T cell-DC cocultures, T cells (1 × 10⁶) were stimulated with DCs (at a DC to T cell ratio of 1:5 and in the presence of 1 μg/ml soluble anti-CD3), and cultures were assayed for T cell proliferation by incorporation of [³H]thymidine or cell-free supernatants were analyzed for kynurenine or cytokine contents. The lytic activity of CD8⁺ T cells against ⁵¹Cr-labeled, NRP-A7-pulsed RMA-S/K^d target cells was measured as described (29).

Foxp3 expression

Real-time PCR were run on a Chromo4 Four-Color Real-Time Detector (Bio-Rad) using Foxp3-specific primers (5′-CCCAGGAAAGACAGCAACCTT-3′; 5′-TTCTCACAACCAGGCCACTTG-3′). Gapdh (5′-GCCTTCCGTGTTCCTACCC-3′; 5′-CAGTGGGCCCTCAGATGC-3′) was the normalizer gene. Conventional RT-PCR was performed using the following primers: Foxp3, 5′-CAGCTGCCTACAGTGCCCCTAG-3′ and 5′-CATTTGCCAGCAGTGGGTAG-3′; Gapdh, 5′-AGCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCTGTTGCTGTA-3′. Specific T cell populations were permeabilized and stained with FITC-labeled anti-Foxp3 according to the manufacturer’s instructions.

Statistical analysis

In all assays, the Student’s t test was used for analysis of in vitro and in vivo data.

Tryptophan catabolism down-regulates TCR ζ-chain in CD8⁺ T cells

Arginine consumption by macrophages has been reported to modulate the expression of TCR ζ-chain (CD3ζ) in T lymphocytes in the absence of apoptosis (31). We have previously shown that, contrary to Th1 cells, naive T cells do not undergo apoptosis when cultured with IDO⁺ DCs for 48 h (17). We assessed CD3ζ expression in CD4⁺ and CD8⁺ T lymphocytes from naive mice after culturing these cells with IDO⁺ DCs in the presence or absence of 1-MT, a specific enzyme inhibitor (Fig. 1⇓A). FACS analysis revealed a noticeable decrease of CD3ζ in CD8⁺ T cells cultured with DCs in the absence of 1-MT. Under these conditions, the initial 50 μM tryptophan concentration dropped to ∼5 μM by 24 h, a time when the concentrations of the major tryptophan catabolites l-kynurenine, 3-hydroxykynurenine, 3-hydroxyanthranilic acid (3-HAA), and quinolinic acid were in the low micromolar range (Fig. 1⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Culture of splenic T cells or purified CD8⁺ T cells with IDO⁺ DCs, or with LT and kynurenines, results in selective CD3ζ down-regulation that is concomitant with tryptophan consumption and kynurenine production and is mediated by GCN2. A, Total T cells and purified CD4⁺ or CD8⁺ fractions from DBA/2 mice were cultured for 48 h with soluble anti-CD3 (1 μg/ml) and IDO⁻ DCs (thick line histogram) or IDO⁺ DCs with (thin line histogram) or without (gray-filled histogram) 1-MT. Cells were labeled for surface CD3, CD4, or CD8 with the respective FITC-labeled reagents, and were then fixed, permeabilized and stained for total ζ-chain expression by means of PE-labeled anti-mouse CD3ζ. Filled histogram indicates isotype controls. Data are from one experiment representative of four. B, Concentrations of tryptophan (TRY) and different kynurenines (mean ± SD of three independent experiments) were measured in 24-h supernatants of total T cells and IDO⁺ DCs. The initial tryptophan concentration was 50 μM. In parallel cocultures established in the presence of 1-MT, the final concentration of tryptophan was ∼30 μM, whereas kynurenines were undetectable. C, Total T cells or purified CD4⁺ and CD8⁺ fractions were activated by plate-bound anti-CD3 (0.5 μg/ml) and anti-CD28 (0.5 μg/ml) in control medium (thick line histogram) or medium containing LT for 48 h, in the absence (thin histogram) or presence (gray-filled histogram) of kynurenines. The effect of addition of kynurenines to control medium was also investigated (dotted line histogram). Cells were fixed, permeabilized and stained for total ζ-chain by means of PE-labeled anti-CD3ζ. Filled histogram indicates isotype controls. D, CD8⁺ cells cultured as in C were assayed for ζ expression at 12, 24, 48, and 72 h. Results were expressed as fluorescence intensity values (mean ± SD) of three independent experiments. ∗, p < 0.01, LT-K vs control medium. E, CD4⁺ and CD8⁺ T cells cultured as in C in control medium (C) or in LT, kynurenines (K), or both (LT-K) were lysed and analyzed by immunoblot with anti-ζ or anti-ε. The effect of T cell coculture with IDO⁺ or IDO⁻ DCs was also investigated. Selected densitometry data (CD3ζ to CDε ratios) were as follows: CD4⁺ T cell in control medium, 1.1; CD4⁺ T cell in LT-K, 0.9; CD4⁺ T cell with IDO⁺, 1.0; CD8⁺ T cell in control medium, 1.0; CD8⁺ T cell in LT-K, 0.3; and CD8⁺ T cell with IDO⁺, 0.2. F, FACS analysis of CD3ζ expression in CD8⁺ cells cultured as in C with LT-K (gray-filled histogram) or control media using wild-type or Gcn2^−/− mice as donors. Filled histograms indicate isotype controls. All experiments were performed at least three times.

To investigate whether ζ-chain down-regulation occurred by the direct action of tryptophan catabolism and to identify the possible milieu responsible for this effect, we cultured CD4⁺ and CD8⁺ T cells in medium containing a LT (5 μM) concentration, a mixture of 3-hydroxykynurenine, 3-HAA, and quinolinic acid, each at 10 μM, or a combination of LT with kynurenines (LT-K). Plate-bound anti-CD3 and anti-CD28 were present in the cultures. FACS analysis revealed that the LT-K culture conditions, but not LT or kynurenines alone, decreased CD3ζ, and did so mostly in CD8⁺ T cells (Fig. 1⇑C). The LT-K conditions had no effect on the expression of other TCR components, including TCRαβ and TCRε, and did not affect costimulatory molecules, including CD28 (data not shown). The effect of LT-K was time-dependent (Fig. 1⇑D) and slowly reversible, as it required reculturing the cells for at least 72 h in control medium for 90% recovery of the original ζ expression (data not shown). The effect was also demonstrable by immunoblot analysis with anti-ζ but not anti-ε Abs, and was likewise observable using IDO⁺ DCs in place of LT-K (Fig. 1⇑E).

The GCN2 stress-response kinase is activated by starvation for amino acids and functions to control transcriptional and translational programs that couple cell growth to amino acid availability (26). Recent evidence indicates that GCN2 activation could be an upstream mechanism in the immunoregulatory effects of IDO (32). To ascertain whether down-regulation of CD3ζ by tryptophan catabolism requires GCN2, we used Gcn2^−/− mice as a source of CD8⁺ T cells that were exposed to LT-K medium (Fig. 1⇑F). In contrast to the Gcn2^+/+ mice, we observed no changes in CD3ζ expression in the T cells lacking GCN2.

TCR ζ-chain down-regulation is associated with impaired function

We investigated whether the diminished expression of CD3ζ induced by LT and kynurenines in CD8⁺ T cells correlates with an impaired immune function in vitro. Purified CD8⁺ and CD4⁺ T cells were exposed to LT-K for 72 h, washed and recultured for 48 h in the presence of anti-CD3, Con A, or a combination of PMA and calcium ionophore (Fig. 2⇓A). Proliferation was significantly reduced by LT-K conditioning upon activation of CD8⁺ cells, but not CD4⁺ cells, with anti-CD3 or Con A. No effect was found on activation with PMA and calcium ionophore in either cell type, demonstrating a selective impairment of TCR-mediated signaling events in the CD8⁺ cells conditioned by LT-K. Under the same conditions, the production of IFN-γ and IL-2 by CD8⁺ T cells in response to anti-CD3 was severely impaired (Fig. 2⇓B). The cytokine response of CD4⁺ T cells was apparently unaffected (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

TCR ζ-chain down-regulation by tryptophan catabolism correlates with impaired T cell function in vitro and occurs in vivo. A, Purified CD8⁺ or CD4⁺ T cells were precultured for 72 h in control or LT-K medium in the presence of plate-bound anti-CD3 and anti-CD28, and were then exposed for 48 h to plate-bound anti-CD3, Con A, or PMA and calcium ionophore, to be assayed for cell proliferation as measured by [³H]thymidine incorporation (mean ± SD of triplicate samples). ∗, p < 0.001, LT-K vs control in one experiment representative of three. B, Cultures of CD8⁺ T cells established as in A and recultured with anti-CD3 were assayed for IFN-γ and IL-2 contents in 48-h supernatants (mean ± SD of triplicate samples). ∗, p < 0.005, LT-K vs control in one experiment representative of three. C, IDO⁻ and IDO⁺ DCs pulsed with NRP-A7 were injected s.c. into NOD-SCID mice (1 × 10⁶), which also received NRP-A7 pentamer-positive 8.3-CD8⁺ T cells (5 × 10⁶) i.v. The injection of IDO⁻ or IDO⁺ NRP-A7-pulsed DCs was repeated after 1 wk and, after 48 h, CD8⁺ cells were recovered from the spleen and lymph nodes; of these cells, 95.4 and 97.3% were pentamer-positive cells in the spleen and lymph nodes, respectively. The CD8⁺ T cells were pooled and assayed for CD3ζ expression by FACS analysis as in Fig. 1⇑C. Isotype control treatment (open histogram) is indicated. The mean channel fluorescence intensity is indicated for one experiment representative of three. The mean fluorescence value ± SD for the three experiments were 22.3 ± 4.1 and 9.5 ± 2.0 for IDO⁻ and IDO⁺ DC treatments, respectively (p < 0.01). D, Portions of the CD8⁺ cells recovered as in C were assayed for lytic activity to NRP-A7-pulsed RMA-S/K^d target cells at different E:T ratios (mean ± SD of replicate samples). ∗, p < 0.01, IDO⁻ vs IDO⁺ DC treatment of the prospective donors of CD8⁺ cytotoxic effectors. One experiment representative of three is shown.

NOD mice expressing the 8.3-TCRαβ transgene (8.3-NOD) (27) possess a prevalent population of autoreactive CD8⁺ T cells that recognize a β cell autoantigen and are reactive to the islet-related NRP-A7 peptide (29, 33, 34). To ascertain whether ζ-chain down-regulation occurs in vivo, we used pentamer technology to transfer NOD-SCID mice with a combination of NRP-A7-pulsed, IDO⁺ DCs and NRP-A7 pentamer-positive 8.3-CD8⁺ T cells. The CD8⁺ cells were harvested from the spleen and lymph nodes, and were analyzed for CD3ζ expression (Fig. 2⇑C) and for ability to lyse peptide-pulsed target cells (Fig. 2⇑D). Noticeable down-regulation of the ζ-chain and impaired lytic activity were observed in the CD8⁺ T cells recovered from NOD-SCID mice immunized with IDO⁺ DCs at the time of T cell transfer.

Tryptophan catabolism converts CD25⁻Foxp3⁻ into CD25⁺Foxp3⁺ cells

Despite the lack of LT-K effects on proliferation and cytokine production by CD4⁺ T cells following a 48-h exposure, we asked whether a prolonged exposure to LT and kynurenines might act on naive CD4⁺ T cells to change their functional properties. When CD4⁺ T cells exposed to LT-K for 7 d were recultured for 24 h with anti-CD3 in standard medium, they produced little or no IFN-γ and IL-4; yet, they produced noticeable amounts of IL-10 and TGF-β (Fig. 3⇓). A similar pattern of cytokine production was observed on assaying CD4⁺ T cells exposed to IDO⁺ DCs in place of LT-K (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Tryptophan catabolism and tryptophan catabolites favor the emergence of cells producing IL-10 and TGF-β. CD4⁺ T cells were cultured for 7 days in control or LT-K medium in the presence of plate-bound anti-CD3 and anti-CD28 and were then activated for 24 h with anti-CD3 to be assayed for the production of IL-10, TGF-β, IFN-γ, and IL-4. Results are means ± SD of replicate samples. One experiment representative of three is shown.

Accumulating evidence suggests that Tregs, including those characterized by IL-10 and TGF-β secretion, represent a dedicated T cell lineage, and that the forkhead family transcription factor Foxp3 functions as the Treg cell lineage specification factor (35). We used real-time PCR analysis for measuring Foxp3 transcript expression in unfractionated CD4⁺ T cells exposed to IDO⁻ or IDO⁺ DCs or to LT or LT-K for 1, 3, or 6 days (Fig. 4⇓A). The results demonstrated an 8- to 40-fold increase in Foxp3 transcription at 3–6 days of cell culture with IDO⁺ DCs or in the LT-K medium. The effect of culturing CD4⁺ T cells with IDO⁺ DCs on Foxp3 transcript expression was negated by the presence of 1-MT.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Induction of Foxp3 transcripts in CD4⁺ T cells cultured with IDO⁺ DCs or in LT-K medium for different times. A, Sorted CD4⁺ lymphocytes (3 × 10⁵) were exposed to IDO⁻ or IDO⁺ DCs (1 × 10⁵) and soluble anti-CD3, or were activated by plate-bound anti-CD3 and anti-CD28 in LT or LT-K media for 1, 3, or 6 days. A group of CD4⁺ T cells was cultured with IDO⁺ DCs in the presence of 1-MT. Foxp3 mRNA levels were quantified by real-time PCR using Gapdh normalization. Data (mean ± SD of triplicate samples in one of three experiments) are presented as normalized Foxp3 transcript expression in the samples relative to normalized Foxp3 transcript expression in the respective control cultures, i.e., cells exposed to IDO⁻ DCs or maintained in standard medium (fold change = 1; dashed line). B, CD4⁺ T cells were fractionated and CD25⁺ and CD25⁻ fractions were cultured as in A in control (C) or LT-K medium for 1, 3, or 6 days. Foxp3 message expression was analyzed by RT-PCR. On examining data as normalized Foxp3 transcript expression in CD4⁺CD25⁻ cell samples relative to normalized Foxp3 transcript expression in control cells (i.e., maintained in standard medium; fold change = 1), the 6-day fold change was 8.3. One experiment representative of three is shown. C, FACS analysis of Foxp3 expression in permeabilized CD4⁺CD25⁻ or CD4⁺CD25⁺ cells cultured in control medium or exposed to LT-K or IDO⁺ DCs for 1, 3, or 6 days as in A. Isotype controls were included in all assays and are shown for day 6. The dot plots were gated on CD4⁺ cells. Number (upper right quadrant) indicates percentage of double-positive cells. Data are representative of three experiments. D, Cytofluorometric analysis of naive CD4⁺ T cells cultured in LT and kynurenines. Sorted CD4⁺CD25⁻ cells were activated by plate-bound anti-CD3 and anti-CD28 in LT-K medium for 7 days (bottom), to be stained with Abs to CD25 (dot plot) or CD69, CD45RB, CD62L, CTLA-4, BTLA, and GITR (histograms). Also depicted is the profile of the original CD4⁺CD25⁻ population (top). Number indicated is percentage of cells in each quadrant. Control cultures were stained with isotype-matched Ab (open histogram).

The CD25 molecule, which is expressed by 5–10% of CD4⁺ T cells in naive mice, is able to operationally if not specifically differentiate naturally present autoimmune-preventive Tregs from other T cells (22). Using RT-PCR, we examined Foxp3 transcripts in CD25⁺ and CD25⁻CD4⁺ fractions after cell culture in LT-K for 1, 3, or 6 days (Fig. 4⇑B). We obtained evidence for the appearance of Foxp3 transcripts in the CD25⁻ fraction at 3–6 days of culture. Foxp3 expression in the CD25⁺ cells remained unchanged. The presence of Foxp3 protein in the CD25⁻ fraction after LT-K exposure was confirmed by flow cytometry (Fig. 4⇑C).

We used FACS analysis to investigate the possible conversion of CD25⁻ cells into CD25⁺ cells in concomitance with Foxp3 appearance, and to evaluate the expression of a series of phenotypic and functional markers including CD69, CD45RB, CD62L, CTLA-4, BTLA, and GITR. At 1 wk of LT-K exposure, >90% of the original CD25⁻ cells appeared to express the CD25 molecule. The converted cells were CD69⁻, CD45RB^low, CD62L⁺, CTLA-4⁺, BTLA^low, and GITR⁺, thus expressing markers associated with the function of CD4⁺CD25⁺ Tregs and subsets thereof in different experimental settings (28, 36, 37, 38, 39) (Fig. 4⇑D).

Next we investigated whether the conversion of CD4⁺CD25⁻ T cells into CD4⁺CD25⁺ Tregs by LT-K could be the result of expansion of residual CD4⁺CD25⁺ cells or selective CD4⁺CD25⁺ cell survival. Using individual populations or combinations of CD4⁺CD25⁻ and CD4⁺CD25⁺ T cells we examined proliferation, cell division, and viable cell recovery data at different times of LT-K exposure. Purified CD4⁺CD25⁻ T cells but not CD4⁺CD25⁺ proliferated in LT-K medium, by both [³H]thymidine uptake and trypan blue dye exclusion analysis (Fig. 5⇓A). Coculturing CFSE-labeled CD4⁺CD25⁻ cells with 10% DDAO-SE-labeled CD4⁺CD25⁺ cells in LT-K revealed selective proliferation of CD4⁺CD25⁻ cells (Fig. 5⇓B). Thus, neither expansion of residual CD4⁺CD25⁺ cells nor selective CD4⁺CD25⁺ cell survival contributed significantly to the conversion of CD4⁺CD25⁻ T cells into CD4⁺CD25⁺ Tregs by LT-K.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Proliferation and viable cell recovery of T cell subsets cultured in LT and kynurenines. A, Purified CD4⁺CD25⁻ T cells (5 × 10⁴), but neither CD4⁺CD25⁺ nor CD8⁺ T cells, proliferate in LT-K medium, by both [³H]thymidine uptake and trypan blue dye exclusion analysis. B, Coculturing CFSE-labeled CD4⁺CD25⁻ cells (6 × 10⁵) with 10% DDAO-SE-labeled CD4⁺CD25⁺ cells for different times reveals selective proliferation of CD4⁺CD25⁻ cells, as determined by flow cytometry. All culture conditions included the use of plate-bound anti-CD3 and anti-CD28.

Converted CD4⁺CD25⁺ cells mediate CTLA-4 and IL-10-dependent suppression

To investigate whether the CD4⁺CD25⁺ Foxp3-expressing cells induced by LT-K display regulatory function in vitro, we cocultured various numbers of those cells with responder CD4⁺CD25⁻ T lymphocytes from naive mice in the presence of anti-CD3 and T cell-depleted splenocyte samples as APCs (Fig. 6⇓A). CD4⁺CD25⁻ cells were exposed to LT-K or cultured with IDO⁺ DCs for 7 days (a time when no viable CD8⁺ T cells would be recoverable from parallel cultures; Fig. 5⇑A), and cells were then assayed for suppressive activity after separation into CD25⁺ and CD25⁻ fractions. The CD4⁺CD25⁺ T cells cultured in LT-K or with IDO⁺ DCs for 7 days greatly inhibited the proliferation of responder T cells in contrast to the minority fraction that had remained CD25⁻ after LT-K or IDO⁺ DC treatment. (These cells did not express Foxp3 transcripts, data not shown.) Suppression was comparable to that of natural Tregs, i.e., CD4⁺CD25⁺ Foxp3-expressing T cells isolated from lymph nodes of naive mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The CD4⁺CD25⁺ cells induced by LT-K have regulatory function in vitro that is dependent on CTLA-4 expression and IL-10 production, and require GCN2 and autocrine TGF-β for the induction of Foxp3. A, Sorted CD4⁺CD25⁻ cells were activated by plate-bound anti-CD3 and anti-CD28 in LT-K media or cocultured with IDO⁺ DCs in the presence of soluble anti-CD3 for 7 days. Cells were then recovered and fractionated, and various numbers of CD25⁺ or CD25⁻ cells were cocultured with 1 × 10⁵ naive CD4⁺CD25⁻ cells and anti-CD3 and accessory cells at different Treg to responder cell ratios. Natural Tregs were assayed for comparison. Proliferation was measured as [³H]thymidine incorporation (mean ± SD) at 48 h. B, CD4⁺CD25⁻ cells were exposed to LT-K to generate a CD25⁺ fraction that was purified and assayed for regulatory activity as described in A, in the presence of Abs to IL-10, CTLA-4, or TGF-β or of isotype controls. Control, CD4⁺CD25⁻ cells cultured in the absence of a CD25⁺ fraction. Significantly different (∗, p < 0.01) from control. In both A and B, one experiment is reported of three performed with similar results. C, CD4⁺CD25⁻ cells were obtained from Gcn2^−/− or wild-type mice to be activated with anti-CD3 and anti-CD28 in control or LT-K media, with or without anti-TGF-β. At 7 days, cells were recovered and Foxp3 mRNA levels were quantified by RT-PCR using Gapdh normalization. D, Portions of the cells cultured in LT-K as described in C were admixed with CD4⁺CD25⁻ cells (1 × 10⁵) and accessory cells in the presence of anti-CD3 at different Treg to responder cell ratios. After 48 h, proliferation was assessed by [³H]thymidine incorporation (mean ± SD). One experiment representative of three.

We investigated whether the expression of CTLA-4, TGF-β, or IL-10 has any role in mediating the regulatory function of the CD25⁺ cells induced by LT-K. Abs to CTLA-4, TGF-β, or IL-10 were added, either singly or in combination, to the cocultures of Treg and responder CD25⁻ T cells in the proliferation assay (Fig. 6⇑B). The Abs had limited (anti-IL-10, anti-CTLA-4) or no (anti-TGF-β) effect when used singly, but the combination of anti-IL-10 and anti-CTLA-4 resulted in ablation of the regulatory properties of the Tregs generated by LT-K exposure.

GCN2 and TGF-β are required for generation of Foxp3⁺ Tregs by LT-K

Naive CD4⁺CD25⁻ T cells can be converted to a CD25⁺ Treg phenotype by TGF-β, which is thought to act through the induction of Foxp3 (37). Although the production of this cytokine seemed to have little role in the in vitro downstream effects of the T cells converted by LT-K, we asked whether autocrine TGF-β is involved in Foxp3 induction. We also examined whether activation of the GCN2 kinase is a prerequisite for the induction of Foxp3 and Treg function by LT-K. Gcn2^−/− and wild-type mice served as a source of CD4⁺CD25⁻ cells that were used for generating Tregs by 7-day exposure to LT-K in the presence of anti-CD3 and anti-CD28, with or without anti-TGF-β. Foxp3 transcript expression (Fig. 6⇑C) and ability to inhibit the proliferation of responder T cells (Fig. 6⇑D) were examined in the recovered cells according to the conditions described above. The GCN2 kinase and TGF-β appeared to be required for the expression of Foxp3 and regulatory properties in CD4⁺ T cells exposed to LT-K. Although it is possible that the anti-TGF-β would also neutralize any TGF-β present in the bovine sera added to the media, it should be noticed that the DCs did not produce significant amounts of TGF-β in vitro in response to LT-K (data not shown).

Tregs induced by LT-K protect hosts from diabetes transfer

Clonotypic T cell NOD mice expressing the BDC2.5-TCRαβ transgene (BDC2.5/NOD) are characterized by a prevalent population of CD4⁺ T cells that recognize a mimotope homologous to a peptide of glutamic acid decarboxylase 65 (40, 41, 42). Most BDC2.5/NOD mice do not develop spontaneous diabetes, but do develop insulitis accompanied by the appearance of activated islet-specific T cells. Evidence suggests that these cells are actively prevented from causing disease by a CTLA-4-dependent immunoregulatory mechanism and by the action of CD4⁺CD25⁺CD69⁻ Tregs (43). At variance with BDC2.5/NOD mice, BDC2.5 TCR transgenic mice backcrossed into the NOD-SCID background show fulminant diabetes, strongly arguing for the regulatory role of endogenously rearranged T cells lacking the transgenic TCR (44). Rapid diabetes also develops in NOD-SCID mice adoptively transferred with spleen cells or purified CD4⁺ T cells from BDC2.5/NOD donors (45). In this setting, however, onset of disease is prevented by the cotransfer of CD4⁺CD25⁺CD62L⁺ cells from prediabetic NOD mice (45). Using the cotransfer model, we investigated whether the Tregs generated by LT-K in vitro are capable of protecting NOD-SCID mice from the induction of diabetes by BDC2.5 transgenic T cells. We tested CD4⁺CD25⁻ cells cultured for 7 d in LT-K medium in comparison with purified CD4⁺CD25⁺CD62L⁺ cells from NOD mice (Fig. 7⇓A). The results showed that the Tregs generated by LT-K prevented transfer of diabetes in a durable and dose-dependent fashion and to an extent comparable to the control Treg population. However, when the recipients of the cotransfer received 1-MT pellets at the time of cell transfer, neither the CD4⁺CD25⁺ Treg population generated by LT-K nor the control Treg population was capable of protecting the hosts from diabetes transfer, and the proportions of disease-free animals at 30 days were 0% and 25% for the two treatments, respectively (Fig. 7⇓B). This result demonstrates that the Tregs induced by LT-K protect NOD-SCID mice from diabetes transfer via IDO-dependent mechanisms, and so does the CD25⁺CD62L⁺ subset of Tregs from early prediabetic NOD mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Tregs generated in vitro by LT and kynurenines protect NOD-SCID mice from developing fulminant diabetes. A, Eight- to 12-wk-old NOD/BDC2.5 splenocytes (1 × 10⁶) were injected i.v. either alone or in combination with two different doses (5 × 10⁵ or 3 × 10⁶) of unfractionated cells derived from an originally CD4⁺CD25⁻ population exposed to LT-K for 7 days in the presence of anti-CD3 and anti-CD28. CD4⁺CD25⁻ cells cultured in standard medium (control T cells) and CD62L⁺ T cells isolated from 4-wk-old prediabetic NOD mice were also used for comparison. The incidence of diabetes is indicated over time for the different groups (n ≥ 6), as revealed by glycosuria and hyperglycemia. One representative experiment of three conducted is shown. Breaks in lines and axis indicate scale interruption. B, Naive CD4⁺CD25⁻ cells were cultured with LT-K for 7 days as indicated in A. Sorted CD25⁺ cells or CD62L⁺ T cells from prediabetic NOD mice (2 × 10⁶) were cotransferred with diabetogenic NOD/BDC2.5 splenocytes (1 × 10⁶) into NOD-SCID mice that had been implanted with placebo or 1-MT pellets. The animals were then monitored for diabetes incidence. One of three experiments with similar results.