Th1 cells have different capacities to develop into memory cells based on their production of IFN-γ. In this study, the mechanism by which a homogenous population of IFN-γ-producing CD4 T cells was eliminated in vivo was assessed. When such cells were transferred into naive mice and activated with Ag, a striking decrease in the frequency of cells in the spleen and lung was observed. However, administration of neutralizing anti-IFN-γ Ab at the time of Ag challenge largely prevented the elimination of such cells. To determine whether IFN-γ was mediating its effects directly and/or indirectly, the ability of IFN-γ to effectively signal in such cells was assessed in vitro. Indeed, there was reduced phosphorylation of STAT1 in response to IFN-γ as well as markedly reduced expression of the IFN-γR β-chain. Furthermore, transfer of such cells into IFN-γR-deficient mice limited their death following activation with Ag. Together, these data suggest that IFN-γ acts in a paracrine manner to mediate the death of activated IFN-γ-producing Th1 cells. In contrast to Ag stimulation, administration of CpG alone resulted in the elimination of Th1 cells in IFN-γR−/− mice. These results show that in response to Ag stimulation, the death of IFN-γ-producing effector Th1 cells is controlled in an IFN-γ-dependent manner, whereas in response to innate activation, the death of IFN-γ-producing Th1 cells can occur through an IFN-γ-independent pathway. Collectively, these data show the multiple mechanisms by which Th1 effector cells are efficiently eliminated in vivo.

The Th1 cells through the production of IFN-γ play a critical role in mediating the intracellular killing of a variety of infectious pathogens. In addition, Th1 responses are also associated with certain organ-specific autoimmune diseases. These observations suggest that tight regulatory control of Th1 responses must exist to allow for protection against infection but also limit reactivity against self or cross-reactive Ags. Thus, understanding the factors that regulate the generation and termination of such responses will be important in vaccine design for protection against intracellular pathogens.

Th1 cells generated in vitro or in vivo are heterogeneous with respect to production of IL-2 and IFN-γ (1, 2). We have previously shown that in vitro activated IFN-γ-producing cells have a limited capacity to survive following transfer into naive mice, whereas IFN-γ negative cells isolated from the same heterogeneous population of Th1 cells develop into long-term memory cells (3). Similarly, Th1 cells that are highly enriched for IFN-γ production are also rapidly eliminated following transfer and antigenic stimulation in vivo (4, 5). These observations led us to examine the mechanism by which IFN-γ-producing Th1 cells are killed in vivo.

One factor that may control Th1 cell death is IFN-γ itself. In experimental models of autoimmune and infectious diseases, mice deficient in IFN-γ (6, 7) or IFN-γ signaling (8) have increased CD4 T cell responses. Furthermore, in vitro generated malaria specific Th1 cells are eliminated in an IFN-γ-dependent manner following transfer into nude mice and subsequent infectious challenge (5). Also, IFN-γ signaling has been shown to directly mediate apoptosis of CD4 T cells in vitro (9). Collectively, these data provide strong evidence that IFN-γ has a critical role in regulating Th1 responses. It remains unclear, however, whether IFN-γ mediates its effects on Th1 cells in a direct or indirect manner in vivo.

In this report, we assessed the mechanism by which IFN-γ-producing Th1 cells are eliminated following antigenic stimulation in vivo. We show that neutralizing IFN-γ at the time of activation in vivo prevented the death of a population of Th1 cells that were homogeneous in their ability to produce IFN-γ. In terms of whether IFN-γ was mediating its effects directly on Th1 cells, we showed that IFN-γR expression and signaling was markedly diminished in such cells. In addition, the adoptive transfer of Th1 cells to IFN-γR-deficient mice largely prevented cell death following stimulation with Ag. These results establish that the IFN-γ released from Th1 effector cells mediates their death in vivo in a paracrine manner. These findings suggest a dual role for IFN-γ in regulating Th1 responses. Although IFN-γ is critical for establishing Th1 cells at the initiation of a response, it also plays an important regulatory role in the elimination of activated IFN-γ-producing effector cells to terminate the response.

DO11.10 or OT-II mice were obtained from Taconic Farms or The Jackson Laboratory, respectively. In some experiments, CD4 T cells from DO11.10 Rag2-deficient mice were used with similar results. STAT1 × DO11.10 mice were provided by K. Murphy (Washington University, St. Louis, MO). IFN-γR−/− mice were purchased from The Jackson Laboratory. All experiments were approved by the Vaccine Research Center Animal Care Use Committee.

Naive CD4 T cells from DO11.10 or OT-II mice were isolated from pooled lymph nodes and spleens by anti-CD4 microbeads and cultured in tissue culture flasks in the presence of irradiated splenocytes with 1 μg/ml OVA peptide (amino acids 323–339) under Th1 polarizing conditions (1 ng/ml IL-12 and 2 μg/ml anti-IL-4) or Th2 polarizing conditions (20 ng/ml IL-4 and 2 μg/ml anti-IL-12) for 4 days, followed by 3 days of rest in fresh medium containing low dose IL-2 (5 U/ml). IFN-γR-deficient CD4 cells were cultured in the presence of 3 μg/ml anti-CD3 rather than peptide. For Th1 wk 3 cells, cells were cultured under Th1 conditions as above for three cycles.

Two million naive CD4 T cells, Th1 wk 3 cells, or Th2 wk 3 cells were injected i.v. into naive BALB/c, C57BL/6, or IFN-γR-deficient mice. In some experiments, cells were labeled with CFSE (Invitrogen Life Technologies) before transfer. One day later, mice received 100 μg of OVA peptide, 50 μg of CpG (1826), or both i.v. In some experiments, mice received 1 mg of anti-IFN-γ Ab (XMG-1.2), 1 mg of an isotype control (GL113), 1 mg of anti-TNF-α (XT-22), or 10 mg of aminoguanidine (Sigma-Aldrich) i.p. on the day of cell transfer, and in some experiments, 0.5 mg of anti-IFN-γ 2 days after injection of OVA.

FITC anti-CD25, allophycocyanin anti-CD62L, PerCP anti-CD4, allophycocyanin anti-IFN-γ, and FITC anti-IL-2 Abs were purchased from BD Biosciences. PE-KJ1–26 Ab (DO11.10 clonotypic Ab) was purchased from Caltag Laboratories. OT-II Th1 cells were identified following adoptive transfer by staining for PerCP anti-CD4, PE anti-Vα2, allophycocyanin anti-IFN-γ, and allophycocyanin anti-IL-2. C57BL/6 mice that did not receive transferred Th1 cells did not have any CD4+/Vα2+/IFN-γ+/IL-2+ cells following stimulation with OVA, CpG, or OVA plus CpG.

For direct ex vivo analysis of cytokine production, recipients were challenged i.v. with OVA peptide and sacrificed 2 h later. Single-cell suspensions were prepared in medium containing Brefeldin A (Sigma-Aldrich). Cells were then stained with anti-CD4 and KJ1–26 Mabs, fixed, permeabilized, and stained with anti-IFN-γ. For in vitro stimulation, single-cell suspensions were cultured in medium with OVA peptide (1 μg/ml) and anti-CD28 (1 μg/ml) or in some experiments, with PMA and ionomycin (Sigma-Aldrich). After 2 h, Brefeldin A was added, and the cells were incubated for a further 3 h. The cells were then fixed, permeabilized, and stained with anti-CD4, KJ1–26, and anti-IFN-γ. The percentage of CD4+/KJ1–26+, KJ1–26+/IFN-γ+, or CD4+/Vα2+/IFN-γ+/IL-2+ cells was assessed by FACS analysis.

mRNA was extracted (Oligotex kit; Qiagen) from resting or stimulated wk 1 or 3 Th1, Th2, or naive CD4 T cells. Stimulated cells were treated with anti-CD3 plus anti-CD28 for 3 h before extraction. A total of 50 ng/ml IFN-γ (R&D Systems) or 10 μg/ml anti-IFN-γ was added to some cells during the 3 h incubation. mRNA was treated with both DNA-free and Dnase solutions (Ambion) then added directly to one-step quantitative RT-PCR containing Superscript RT-Platinum Taq enzyme mix (Invitrogen Life Technologies). Probes were labeled with FAM reporter and QSY7 quencher (MegaBases). Reactions were run in duplicate and analyzed with an ABI 7700 real-time system (PE Applied Biosystems). Parallel reactions were run for β2-microglobulin to normalize input mRNA amount. Relative amounts of target mRNA were then calculated by the comparative (ΔΔ)Ct method. Primers and probes were designed against sequences for murine IFN-γR1, IFN-γR2, IFN-γ, IP10, and IL-4 using Primer Express software (PE Applied Biosystems).

Th1 cells (5 million cells/ml) were stimulated for 20 min at 37°C with 10,000 U/ml IFN-α or 50 ng/ml IFN-γ then lysed in 100 μl Triton lysis buffer (0.05 M Tris, 0.3 M NaCl, 0.5% Triton X-100, 2 mM EDTA, 0.4 mM sodium orthovanadate, 2.5 mM leupeptin, 2.5 mM aprotinin, and 2.5 mM NPGB). Protein concentrations of whole cell lysates were determined by bicinchoninic acid protein assay (Pierce). Forty-five μg total protein was separated by SDS PAGE, followed by transfer to nitrocellulose. Western blotting was performed using Abs against STAT1 (BD Biosciences), STAT4 (Santa Cruz Biotechnology), and phosphorylated STAT1 and STAT4 (Invitrogen Life Technologies).

Ten million CFSE-labeled Th1 wk 1 cells were adoptively transferred into naive BALB/c mice. One day later, the recipient mice received either OVA peptide or CpG i.v. Mice were sacrificed, and spleens were harvested at 24 and 72 h post-treatment. Frozen 5-μm-thick spleen sections were prepared, and apoptosis was detected based on labeling of DNA strand breaks using TUNEL technology (In Situ Cell Death Detection kit, TMR red, Roche). Visualization of GFP was conducted on a Leica SP5 confocal microscope (Leica Microsystems) using a 20× oil immersion objective numerical aperture 0.9. Fluorochromes were excited using a 405 nm diode laser for Hoechst and a 488 nm laser for GFP. To avoid possible cross-talk, the two wavelengths were collected separately and later merged. Images were processed using Leica TCS-SP software (version 2.1537) and Adobe Photoshop CS2 (Adobe systems).

The aim of this study was to assess the mechanism by which CD4 T cells that produce IFN-γ are eliminated following activation in vivo. To generate a population of Th1 cells in which all of the cells produce IFN-γ, naive CD4 T cells isolated from DO11.10 TCR-Tg mice were cultured under Th1 polarizing conditions for 3 wk. After each round of stimulation, cells were assessed for the expression of IL-2 and IFN-γ as well as cell surface markers of activation. As shown in Fig. 1, there was heterogeneous expression of IL-2, IFN-γ, CD25, and the lymph node homing receptor molecule CD62L from these cells after 1 wk of stimulation (Th1 wk 1). By contrast, after 3 wk of culture, all DO11.10 TCR-Tg cells were CD62L low and produced IFN-γ. For the remainder of this study for all adoptive transfers, we used Th1 cells that had been cultured for 3 wk (Th1 wk 3), a population of CD4 T cells in which all cells were capable of producing IFN-γ.

FIGURE 1.

Generation and characterization of Th1 cells. Naive DO11.10 TCR-transgenic (Tg)3CD4 T cells were cultured in Th1 polarization conditions for 7-day cycles (4 days of stimulation followed by 3 days of rest in IL-2) over the course of 3 wk. After each wk of culture, cells were assessed for the expression of intracellular IL-2 and IFN-γ and cell surface markers of activation.

FIGURE 1.

Generation and characterization of Th1 cells. Naive DO11.10 TCR-transgenic (Tg)3CD4 T cells were cultured in Th1 polarization conditions for 7-day cycles (4 days of stimulation followed by 3 days of rest in IL-2) over the course of 3 wk. After each wk of culture, cells were assessed for the expression of intracellular IL-2 and IFN-γ and cell surface markers of activation.

Close modal

Previously, we reported that a homogeneous population of activated, IFN-γ-producing Th1 cells that was transferred into naive mice was quickly eliminated without any further stimulation (3). However, by culturing the Th1 cells in a low dose of IL-2 for several days before adoptive transfer, these cells were able to survive following transfer into naive mice (Fig. 2,A). This enabled us to examine the fate of such cells after challenge with specific Ag in vivo. Two hours after injection with OVA, the frequency of adoptively transferred Th1 cells that produced IFN-γ was assessed immediately ex vivo by intracellular cytokine staining (10). As shown in Fig. 2 B, ∼60–70% of the Th1 cells produced IFN-γ in the spleens or lungs of recipient mice following challenge with Ag. These results demonstrate that the in vivo activation of Th1 cells with OVA was rapid and efficient for inducing production of IFN-γ.

FIGURE 2.

The fate of Th1 cells following adoptive transfer and activation in vivo. A, Th1 wk 3 cells were cultured in low dose IL-2 for 3 days. Such cells (2 × 106) were labeled with CFSE and transferred into naive BALB/c mice (n = 3 per group). As controls, similar numbers of IL-2 cultured Th1 wk 3 cells, that were activated for 3 h in vitro with immobilized anti-CD3 mAb and soluble anti-CD28 before transfer, were labeled with CFSE and transferred. Three days post-transfer, KJ1–26+CFSE + cells were identified in the spleen. B–D, IL-2 cultured wk 3 cells were labeled with CFSE, transferred (2 × 106) into BALB/c mice, and challenged with 100 μg of OVA peptide, 50 μg of CpG, or both i.v. Two hours later (B), mice were sacrificed, and the frequency of IFN-γ-producing KJ1–26+ CD4 T cells was assessed immediately ex vivo by intracellular cytokine staining. Two weeks later (C and D), the frequency of KJ1–26+ CD4 T cells was assessed in the spleen and lung, and extent of division was determined by loss of CFSE intensity in the lung. ∗∗, p < 0.01, significantly different by t test compared with no treatment (Tx) group in the same organ.

FIGURE 2.

The fate of Th1 cells following adoptive transfer and activation in vivo. A, Th1 wk 3 cells were cultured in low dose IL-2 for 3 days. Such cells (2 × 106) were labeled with CFSE and transferred into naive BALB/c mice (n = 3 per group). As controls, similar numbers of IL-2 cultured Th1 wk 3 cells, that were activated for 3 h in vitro with immobilized anti-CD3 mAb and soluble anti-CD28 before transfer, were labeled with CFSE and transferred. Three days post-transfer, KJ1–26+CFSE + cells were identified in the spleen. B–D, IL-2 cultured wk 3 cells were labeled with CFSE, transferred (2 × 106) into BALB/c mice, and challenged with 100 μg of OVA peptide, 50 μg of CpG, or both i.v. Two hours later (B), mice were sacrificed, and the frequency of IFN-γ-producing KJ1–26+ CD4 T cells was assessed immediately ex vivo by intracellular cytokine staining. Two weeks later (C and D), the frequency of KJ1–26+ CD4 T cells was assessed in the spleen and lung, and extent of division was determined by loss of CFSE intensity in the lung. ∗∗, p < 0.01, significantly different by t test compared with no treatment (Tx) group in the same organ.

Close modal

To extend these findings, mice that received transferred Th1 wk 3 cells were then challenged with OVA, CpG, or OVA plus CpG. The use of CpG alone enabled the determination of the role that innate immunity alone, or in combination with Ag, has on the survival of Th1 cells in vivo. The frequency of transferred Th1 cells was markedly diminished in the spleens and lungs of recipient mice treated with OVA, CpG, or OVA plus CpG compared with unchallenged mice (Fig. 2,C). Of note, there were essentially no cells detected in spleens compared with lungs in response to any of the stimuli. Mechanisms to account for these findings are discussed below. Finally, Th1 wk 3 cells in the lung that were activated with OVA or OVA plus CpG underwent several divisions as indicated by loss of CFSE intensity, whereas those that were activated with CpG alone did not appear to divide more than untreated cells (Fig. 2 D). Combined, these data show that following activation with Ag, there is rapid induction of IFN-γ followed by elimination.

The mechanism by which the IFN-γ-producing Th1 cells were eliminated in vivo was explored. As IFN-γ was so rapidly induced from the transferred cells following OVA stimulation (Fig. 2,B) and has been shown to mediate the death of CD4 T cells (9), its role in the elimination of the Th1 cells in vivo was assessed. Naive mice that received transferred naive DO11.10 CD4 T, Th1 wk 3, or Th2 wk 3 cells were stimulated with OVA or OVA plus CpG. A group of mice was treated with anti-IFN-γ at the same time they were stimulated with OVA. At 9 days poststimulation, the Th1 cells were present at a high frequency in the spleens and lungs of recipient mice without OVA injection (Fig. 3, A and B). However, there were few detectable cells in the spleens or lungs of recipient mice that received OVA or OVA plus CpG. By contrast, treatment with anti-IFN-γ at the time of OVA injection prevented the elimination of such cells in the lungs and, to a lesser extent, in the spleens of recipient mice. The frequency of naive CD4 T cells, which produce little IFN-γ immediately upon activation, was not decreased following injection of OVA and was increased following injection of OVA plus CpG. However, the frequency of Th2 wk 3 cells, which also produce little IFN-γ upon activation, was decreased following OVA and OVA plus CpG injections in both the spleens and lungs. Yet, unlike the Th1 wk 3 cells, the elimination of the Th2 wk 3 cells was not prevented by the administration of anti-IFN-γ. These data clearly establish a role for IFN-γ in the specific elimination of Th1 cells in vivo.

FIGURE 3.

The role of IFN-γ in mediating the elimination of Th1 cells following activation in vivo. Naive DO11.10 CD4 T, Th1 wk 3, and Th2 wk 3 cells were transferred into naive BALB/c mice. At day 1 post-transfer, mice were injected i.v. with 100 μg of OVA peptide or 100 μg of OVA peptide together with 50 μg of CpG; some mice received anti-IFN-γ i.p. As an additional control, a group of mice were not injected with OVA (No Tx). At day 9 postpeptide injection, the frequency of transferred cells was measured in the spleen (A) and lung (B) as the percent of CD4 T cells that were KJ1–26+. ∗, p < 0.05, significantly different by t test compared with treatment with OVA.

FIGURE 3.

The role of IFN-γ in mediating the elimination of Th1 cells following activation in vivo. Naive DO11.10 CD4 T, Th1 wk 3, and Th2 wk 3 cells were transferred into naive BALB/c mice. At day 1 post-transfer, mice were injected i.v. with 100 μg of OVA peptide or 100 μg of OVA peptide together with 50 μg of CpG; some mice received anti-IFN-γ i.p. As an additional control, a group of mice were not injected with OVA (No Tx). At day 9 postpeptide injection, the frequency of transferred cells was measured in the spleen (A) and lung (B) as the percent of CD4 T cells that were KJ1–26+. ∗, p < 0.05, significantly different by t test compared with treatment with OVA.

Close modal

IFN-γ could mediate the death of Th1 cells in vivo directly by acting on the transferred cells in an autocrine manner, indirectly by the induction of factors from host cells, or both. To address these mechanisms, the expression of the IFN-γR in activated Th1 wk 3 cells by real-time PCR was compared with activated naive or Th2 cells. The IFN-γ membrane receptor is comprised of two chains, IFN-γR1, the ligand-binding chain, and IFN-γR2, the signaling chain (11). As shown in Fig. 4, mRNA expression for IFN-γR1 was similar for Th1 wk 3 and naive cells and increased in Th2 cells compared with naive cells (Fig. 4,A and Table I). In contrast, the expression of IFN-γR2 was demonstrably lower in Th1 wk 3 cells compared with the other cell types. As IFN-γ itself can decrease IFN-γR2 receptor expression (12), a similar analysis was done on activated naive, Th1 wk 3, and Th2 cells in which anti-IFN-γ was added to the cultures (Fig. 4,B and Table II). Similar to the results above, there was minimal mRNA expression for IFN-γR2 from activated Th1 wk 3 cells cultured for 2 h in the presence of anti-IFN-γ compared with naive or Th2 cells. IFN-γ responsiveness was also determined by assessing whether the presence of exogenous IFN-γ could increase the expression of the IFN-γ-inducible gene IP-10. Although there was a statistically significant increase in mRNA expression of IP-10 in activated naive cells and in Th2 cells, there was little change in expression from Th1 wk 3 cells (Table II). Similar results were observed in resting cells. These data are consistent with a previous study that IFN-γR2 expression is greatly reduced in Th1 wk 3 cells and that these cells are thus unresponsive to IFN-γ signaling (13).

FIGURE 4.

IFN-γR expression in Th1 cells. A, IFN-γR1 and IFN-γR2 expression in activated naive DO11.10 CD4 T, Th1 wk 3, and Th2 cells was determined using RT-PCR on triplicate cultures. In B, cells were activated in the presence of anti-IFN-γ or IFN-γ. Samples were normalized to one naive sample (A) and one naive anti-IFN-γ sample (B). Error bars represent mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01; and ∗∗∗, p < 0.001, significantly different by t test compared with naive with same treatment.

FIGURE 4.

IFN-γR expression in Th1 cells. A, IFN-γR1 and IFN-γR2 expression in activated naive DO11.10 CD4 T, Th1 wk 3, and Th2 cells was determined using RT-PCR on triplicate cultures. In B, cells were activated in the presence of anti-IFN-γ or IFN-γ. Samples were normalized to one naive sample (A) and one naive anti-IFN-γ sample (B). Error bars represent mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01; and ∗∗∗, p < 0.001, significantly different by t test compared with naive with same treatment.

Close modal
Table I.

mRNA expression of the IFN-γR components and IFN-γ in naive CD4 T, Th1 wk 3, and Th2 cellsa

Cells TestedIFN-γR1IFN-γR2IFN-γ
Naive CD4 T cells 1.41 ± 0.21 1.55 ± 0.32 1.27 ± 0.18 
Th1 wk 3 cells 0.81 ± 0.25 0.14 ± 0.03b 993.20 ± 197.40c 
Th2 cells 2.09 ± 0.32 2.17 ± 0.20 9.83 ± 2.01b 
Cells TestedIFN-γR1IFN-γR2IFN-γ
Naive CD4 T cells 1.41 ± 0.21 1.55 ± 0.32 1.27 ± 0.18 
Th1 wk 3 cells 0.81 ± 0.25 0.14 ± 0.03b 993.20 ± 197.40c 
Th2 cells 2.09 ± 0.32 2.17 ± 0.20 9.83 ± 2.01b 
a

Mean ± SEM. n = 3. The t test measures significant differences between naive CD4 T and Th1 week 3 cells or Th2 cells.

b

p < 0.05 and

c

p < 0.01.

Table II.

mRNA expression for IFN-γR2, IP10, IFN-γ, and IL-4 from activated naive CD4 T, Th1 wk 3, and Th2 cellsa

CellsTreatmentIFN-γR2eIP10fIFN-γeIL-4e
Naive CD4 T cells anti-IFN-γ 1.19 ± 0.15 1.01 ± 0.07 0.96 ± 0.11 0.76 ± 0.12 
 IFN-γ 0.57 ± 0.02 4.50 ± 0.85b 0.92 ± 0.04 0.59 ± 0.05 
Th1 wk 3 cells anti-IFN-γ 0.05 ± 0.01c 0.83 ± 0.21 456.90 ± 71.68c 0.10 ± 0.06c 
 IFN-γ 0.07 ± 0.00d 1.32 ± 0.46 769.30 ± 76.51d 0.14 ± 0.06c 
Th2 cells anti-IFN-γ 2.21 ± 0.49 8.27 ± 0.69 12.03 ± 3.90b 33.32 ± 6.58c 
 IFN-γ 0.92 ± 0.17 15.23 ± 0.23d 7.00 ± 3.26 42.90 ± 8.77c 
CellsTreatmentIFN-γR2eIP10fIFN-γeIL-4e
Naive CD4 T cells anti-IFN-γ 1.19 ± 0.15 1.01 ± 0.07 0.96 ± 0.11 0.76 ± 0.12 
 IFN-γ 0.57 ± 0.02 4.50 ± 0.85b 0.92 ± 0.04 0.59 ± 0.05 
Th1 wk 3 cells anti-IFN-γ 0.05 ± 0.01c 0.83 ± 0.21 456.90 ± 71.68c 0.10 ± 0.06c 
 IFN-γ 0.07 ± 0.00d 1.32 ± 0.46 769.30 ± 76.51d 0.14 ± 0.06c 
Th2 cells anti-IFN-γ 2.21 ± 0.49 8.27 ± 0.69 12.03 ± 3.90b 33.32 ± 6.58c 
 IFN-γ 0.92 ± 0.17 15.23 ± 0.23d 7.00 ± 3.26 42.90 ± 8.77c 
a

Mean ± SEM. n = 3.

b

p < 0.05;

c

p < 0.01; and

d

p < 0.001. Values are normalized to one naive α-IFN-γ sample.

e

For IFN-γR2, IFN-γ, and IL-4 expression, t test measures significant differences between naive with same treatment and Th1 or Th2 cells.

f

For IP10, t test measures significant differences between anti-IFN-γ and IFN-γ for each cell type.

To provide further functional evidence that IFN-γR2 expression was necessary to mediate signaling by IFN-γ, the phosphorylation of STAT1 protein in naive, Th1 wk 1, and Th1 wk 3 cells in response to exogenous IFN-γ was assessed. As a positive control, IFN-α, which can also phosphorylate STAT1, was used. As shown in Fig. 5, A and B, the IFN-γR is required for phosphorylation of STAT1 in response to IFN-γ since Th1 wk 3 cells generated from IFN-γR-deficient mice do not phosphorylate STAT1. Similarly, Th1 wk 3 cells generated from wild-type DO11.10 mice, which have very low expression of IFN-γR2, were unable to phosphorylate STAT1 in response to IFN-γ; however, STAT1 was phosphorylated in the presence of IL-12 or IFN-α in both IFN-γR-deficient and wild-type mice. Finally, STAT4 was phosphorylated to a similar degree in all three cell types in the presence of IL-12, indicating that other signaling pathways were fully functional (Fig. 5, A and C). In addition, naive CD4 T cells, but not Th1 wk 1 or Th1 wk 3 cells, were able to phosphorylate STAT1 in response to IFN-γ (Fig. 5, D and E). Taken together, these data show that functional IFN-γ signaling is markedly diminished in differentiated Th1 cells. Therefore, it is unlikely that IFN-γ secreted from Th1 cells acts directly on such cells to initiate a signaling cascade that leads to their death.

FIGURE 5.

IFN-γ signaling in Th1 cells. The capacity of IFN-γ to phosphorylate STAT1 protein in Th1 cells was measured by Western blot in Th1 wk 3 cells generated from IFN-γR−/−, DO11.10 STAT1−/−, and DO11.0 STAT1+/+ mice (A) and in naive CD4 T, Th1 wk 1, and Th1 wk 3 cells (D). STAT phosphorylation was measured by normalizing pSTAT1 to total STAT1 (B and E) and pSTAT4 to STAT4 (C) using densitometry.

FIGURE 5.

IFN-γ signaling in Th1 cells. The capacity of IFN-γ to phosphorylate STAT1 protein in Th1 cells was measured by Western blot in Th1 wk 3 cells generated from IFN-γR−/−, DO11.10 STAT1−/−, and DO11.0 STAT1+/+ mice (A) and in naive CD4 T, Th1 wk 1, and Th1 wk 3 cells (D). STAT phosphorylation was measured by normalizing pSTAT1 to total STAT1 (B and E) and pSTAT4 to STAT4 (C) using densitometry.

Close modal

Since Th1 cells appear unable to signal in response to IFN-γ, we hypothesized that IFN-γ was acting indirectly on the cells of the adoptive host rather than in an autocrine manner on the transferred Th1 cells to mediate their death. Thus, Th1 cells were transferred into IFN-γR-deficient mice to determine whether the Th1 cells would be preserved when the host cells were incapable of responding to IFN-γ. Recipient wild-type and IFN-γR-deficient mice were then stimulated with OVA, OVA plus CpG, or CpG alone. Twenty days after transfer and in vivo stimulation, there were few Th1 wk 3 cells in the spleens of either strain of mice. However, there were substantially more Th1 wk 3 cells present in the lungs of IFN-γR-deficient mice stimulated with OVA or OVA plus CpG than in the lungs of wild-type mice (Fig. 6,A). These data suggest that IFN-γ acts on the cells of the recipient mice to mediate the death of the transferred Th1 cells in the lung. Importantly, injection of CpG alone nearly eliminated the Th1 cells from the spleens and lungs of both wild-type and IFN-γR-deficient mice. Taken together, these data show that Ag alone or in combination with innate stimuli mediates the death of Th1 cells in the lung through an IFN-γ-dependent paracrine mechanism while innate activation alone mediates death via an IFN-γ-independent mechanism. Remarkably, the death caused by administering CpG or OVA alone is overcome by antigenic stimulation with OVA plus CpG. This is likely the result of increased Th1 cell proliferation and/or survival due to the additional stimulation received by costimulatory molecules that are up-regulated by CpG (14, 15). Indeed, in a separate experiment, more CFSE negative Th1 cells were detected in the OVA plus CpG group than the OVA or CPG groups in the IFN-γR-deficient mice, suggesting that these cells had divided more (Fig. 6 B).

FIGURE 6.

Survival of Th1 cells in IFN-γR-deficient recipient mice and in situ detection of cell death in spleens. A and B, A total of 2 × 106 Th1 wk 3 cells, generated from OT-II TCR-Tg mice, were labeled with CFSE, transferred into C57BL/6 mice (n = 2–3 per group), and injected with OVA, CpG, or OVA plus CpG the following day. The frequency of transferred cells was assessed in spleen and lung 20 days postchallenge. ∗, p < 0.05, significantly different compared with OVA plus CpG in wild-type mice by t test. C, A total of 10 × 106 CFSE-labeled Th1 wk 3 cells, generated from DO11.10 TCR-Tg mice, were adoptively transferred into naive mice. Recipient mice were then injected with OVA or CpG, and spleens were harvested 24 or 72 h post-treatment. Frozen sections were stained for apoptotic cells using TUNEL technology. Slides were examined using confocal microscopy. Images show Th1 cells in green clusters in T cell regions and apoptotic cells in red. True colocalization was not detected as CFSE stains the cytoplasm while TUNEL stains the nucleus.

FIGURE 6.

Survival of Th1 cells in IFN-γR-deficient recipient mice and in situ detection of cell death in spleens. A and B, A total of 2 × 106 Th1 wk 3 cells, generated from OT-II TCR-Tg mice, were labeled with CFSE, transferred into C57BL/6 mice (n = 2–3 per group), and injected with OVA, CpG, or OVA plus CpG the following day. The frequency of transferred cells was assessed in spleen and lung 20 days postchallenge. ∗, p < 0.05, significantly different compared with OVA plus CpG in wild-type mice by t test. C, A total of 10 × 106 CFSE-labeled Th1 wk 3 cells, generated from DO11.10 TCR-Tg mice, were adoptively transferred into naive mice. Recipient mice were then injected with OVA or CpG, and spleens were harvested 24 or 72 h post-treatment. Frozen sections were stained for apoptotic cells using TUNEL technology. Slides were examined using confocal microscopy. Images show Th1 cells in green clusters in T cell regions and apoptotic cells in red. True colocalization was not detected as CFSE stains the cytoplasm while TUNEL stains the nucleus.

Close modal

Finally, it is notable that in all experiments there are far less cells detected in the spleen compared with the lung following activation in vivo. These results suggest that there may be differential regulation of Th1 cell death in lymphoid and nonlymphoid tissues in response to antigenic or innate activation. Alternatively, the failure to detect cells in the spleen may be due to migration of the Th1 cells from lymphoid to nonlymphoid organs following activation. To address this, confocal microscopy was used to identify apoptotic cells in the spleen following transfer of CFSE-labeled Th1 cells and activation with OVA or CpG alone. As shown in Fig. 6 C, there were very few apoptotic cells in the T cell regions of untreated mice. Twenty-four hours after stimulation with OVA, the Th1 cells had begun to divide as indicated by loss of CFSE intensity, and there were a few apoptotic cells in the T cells regions of the spleen. However, 24 h after treatment with CpG, a large number of cells in the T cell regions had died while very few Th1 cells appeared to have divided. By 72 h poststimulation, very few CFSE bright cells were detected in the spleens of mice treated with OVA, whereas some still existed in those treated with CpG. In both groups, a number of apoptotic cells remained. These results suggest that CpG mediates the death of Th1 cells in the spleen, whereas antigenic stimulation leads to Th1 cell division and migration to the periphery.

The data presented above suggest that IFN-γ mediates the death of Th1 wk 3 cells by acting on other cells to induce factors that can then lead to their demise. Thus, we explored the role of two IFN-γ-inducible factors that are associated with cell death, TNF-α and NO, in Th1 cell death. Th1 wk 3 cells were transferred into naive BALB/c mice and stimulated with OVA peptide one day later. At the time of stimulation, mice received either control Ab, anti-IFN-γ, anti-TNF-α, or aminoguanidine to inhibit NO production. Eleven days later, the frequency of KJ1–26+ CD4 T cells in the spleens and lungs of recipient mice was assessed (Fig. 7). As shown previously, injection of OVA peptide resulted in a considerable decrease of Th1 cells in both the spleens (Fig. 7,A) and lungs (Fig. 7 B) of recipient mice, which could be largely prevented by the addition of anti-IFN-γ at the time of injection. Treatment of mice with anti-TNF-α Ab did not prevent the disappearance of Th1 cells in the spleen; however, most of the Th1 cells in the lung were maintained. Similar results were obtained when aminoguanidine was administered to inhibit NO formation. Thus, while neither anti-TNF-α or aminoguanidine were as effective as anti-IFN-γ in preventing Th1 cell death, both had a considerable effect in the lung. These results show that both TNF-α and NO are capable of mediating Th1 cell death and may play a role in IFN-γ mediated Th1 cell death.

FIGURE 7.

Role of IFN-γ-inducible mediators on Th1 cell survival. IL-2 cultured Th1 wk 3 cells were transferred (2 × 106) into BALB/c mice. One day later, mice were challenged with 100 μg of OVA peptide i.v., together with either 1 mg control Ab (GL113), 1 mg anti-IFN-γ, 1 mg anti-TNF-α, or 10 mg aminoguanidine i.p. Eleven days later, the frequency of KJ1–26+ CD4 T cells was assessed in the spleen (A) and lung (B). ∗∗, Significantly different to OVA plus GL113, t test (p < 0.01).

FIGURE 7.

Role of IFN-γ-inducible mediators on Th1 cell survival. IL-2 cultured Th1 wk 3 cells were transferred (2 × 106) into BALB/c mice. One day later, mice were challenged with 100 μg of OVA peptide i.v., together with either 1 mg control Ab (GL113), 1 mg anti-IFN-γ, 1 mg anti-TNF-α, or 10 mg aminoguanidine i.p. Eleven days later, the frequency of KJ1–26+ CD4 T cells was assessed in the spleen (A) and lung (B). ∗∗, Significantly different to OVA plus GL113, t test (p < 0.01).

Close modal

Although it has been previously shown that IFN-γ controls the expansion and death of CD4 (6, 7) and CD8 T cell responses (16, 17), the data presented in this study provide a mechanism for how IFN-γ mediates the elimination of a functionally defined population of Th1 effector cells in vivo. We show that in response to activation with Ag, IFN-γ mediates the death of Th1 cells indirectly. Numerous factors associated with cell death, such as TNF-αRs (18), NO (19), IDO (20, 21), TRAIL (22), galectin-9 (23), and Fas/Fas ligand (24), have been identified as IFN-γ-inducible molecules. It is likely that any or all of these factors from the host cells act on Th1 cells to initiate their death. Indeed, we observed that neutralization of TNF-α or inhibition of NO saved a substantial number of Th1 cells from death. These results suggest that IFN-γ may be controlling death of Th1 cells by inducing multiple death pathways. In contrast to the results seen with antigenic stimulation, it was striking that CpG alone efficiently killed Th1 cells in an IFN-γ-independent manner in both spleens and lungs of IFN-γR-deficient mice. The mechanism for CpG mediated death could be a direct effect of IDO released from dendritic cells (25, 26) in an IFN-γ-independent manner. Alternatively, induction of Type I IFN induced by CpG may also mediate the death of such cells. Indeed, bystander cell death has been reported for CD8 T cells through Type I IFN (27). Finally, we would note that CpG is a potent inducer of IL-12, which indirectly can induce IFN-γ. Thus, CpG would also be able to mediate death of Th1 effector cells through an IFN-γ-dependent pathway.

Our results extend previous studies and show that IFN-γ has a central role for controlling all stages of Th1 differentiation and memory formation. During the generation of a primary Th1 response, IFN-γ acts as a positive regulator by selectively inducing Th1 differentiation through increased transcription of T-bet, which results in enhanced IL-12 responsiveness and suppressed Th2 lineage commitment (28, 29, 30). In addition, IFN-γ enhances IL-12 transcription, (31) which further optimizes the generation of Th1 cells. In contrast, in the early stages of CD4 T cell activation, IFN-γ can also negatively control T cell expansion by direct killing of such cells through caspase 8 (9). As Th1 cells develop, however, the expression of the IFN-γR diminishes, and they lose their capacity to be directly effected. At this point, the regulatory aspects of IFN-γ are indirect. Collectively, these data reveal a dynamic role for IFN-γ, showing its ability to prime a Th1 response, mediate effector function, and then terminate the response.

The biological relevance of IFN-γ mediated control of Th1 effector responses is that a careful balance is established between maintaining a sufficient frequency of effector/memory Th1 responses to mediate protection against infectious pathogens and limiting the excessive inflammation caused by Th1 cells specific for cross-reactive or self-Ags. This is illustrated in experimental mouse models of organ-specific autoimmune diseases in which IFN-γ (6) or IFN-γR-deficient mice (32) have enhanced susceptibility and disease progression due to an increased number of pathogenic T cells. Thus, the distinct mechanisms by which Ag and innate stimulation are able to mediate the death of IFN-γ-producing Th1 cells ensure the elimination of such cells to prevent potential self-reactivity in both lymphoid and nonlymphoid organs. In addition, this regulation has important implications for vaccine design for infections requiring Th1 immunity. The failure of IFN-γ-producing effector cells to survive following activation with Ag and their elimination by innate stimuli alone puts potential constraints on sustaining enough Th1 effector/memory cells over a prolonged period of time that will be sufficient to mediate protection. In this regard, we postulate that persistent boosting throughout the lifetime of the host will be important to ensure protective immunity with nonlive vaccines for infections requiring Th1 responses.

We thank Ken Murphy and Lisa Berenson for providing reagents and critical advice, and Fred Finkelman for providing neutralizing Ab against IFN-γ and TNF-α.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work is supported in part by grants from the National Key Basic Research Program of China (2007CB512404 to C.Y.W.) and the National Natural Science Foundation of China (30321004 to C.Y.W.).

3

Abbreviations used in this paper: Tg, transgenic; Tx; treatment.

1
Foulds, K. E., C. Y. Wu, R. A. Seder.
2006
. Th1 memory: implications for vaccine development.
Immunol. Rev.
211
:
58
-66.
2
Roman, E., E. Miller, A. Harmsen, J. Wiley, U. H. Von Andrian, G. Huston, S. L. Swain.
2002
. CD4 effector T cell subsets in the response to influenza: heterogeneity, migration, and function.
J. Exp. Med.
196
:
957
-968.
3
Wu, C. Y., J. R. Kirman, M. J. Rotte, D. F. Davey, S. P. Perfetto, E. G. Rhee, B. L. Freidag, B. J. Hill, D. C. Douek, R. A. Seder.
2002
. Distinct lineages of T(H)1 cells have differential capacities for memory cell generation in vivo.
Nat. Immunol.
3
:
852
-858.
4
Hayashi, N., D. Liu, B. Min, S. Z. Ben-Sasson, W. E. Paul.
2002
. Antigen challenge leads to in vivo activation and elimination of highly polarized TH1 memory T cells.
Proc. Natl. Acad. Sci. USA
99
:
6187
-6191.
5
Xu, H., J. Wipasa, H. Yan, M. Zeng, M. O. Makobongo, F. D. Finkelman, A. Kelso, M. F. Good.
2002
. The mechanism and significance of deletion of parasite-specific CD4+ T cells in malaria infection.
J. Exp. Med.
195
:
881
-892.
6
Chu, C. Q., S. Wittmer, D. K. Dalton.
2000
. Failure to suppress the expansion of the activated CD4 T cell population in interferon γ-deficient mice leads to exacerbation of experimental autoimmune encephalomyelitis.
J. Exp. Med.
192
:
123
-128.
7
Dalton, D. K., L. Haynes, C. Q. Chu, S. L. Swain, S. Wittmer.
2000
. Interferon γ eliminates responding CD4 T cells during mycobacterial infection by inducing apoptosis of activated CD4 T cells.
J. Exp. Med.
192
:
117
-122.
8
Feuerer, M., K. Eulenburg, C. Loddenkemper, A. Hamann, J. Huehn.
2006
. Self-limitation of Th1-mediated inflammation by IFN-γ.
J. Immunol.
176
:
2857
-2863.
9
Refaeli, Y., L. Van Parijs, S. I. Alexander, A. K. Abbas.
2002
. Interferon γ is required for activation-induced death of T lymphocytes.
J. Exp. Med.
196
:
999
-1005.
10
Reinhardt, R. L., A. Khoruts, R. Merica, T. Zell, M. K. Jenkins.
2001
. Visualizing the generation of memory CD4 T cells in the whole body.
Nature
410
:
101
-105.
11
Bach, E. A., M. Aguet, R. D. Schreiber.
1997
. The IFN γ receptor: a paradigm for cytokine receptor signaling.
Annu. Rev. Immunol.
15
:
563
-591.
12
Bach, E. A., S. J. Szabo, A. S. Dighe, A. Ashkenazi, M. Aguet, K. M. Murphy, R. D. Schreiber.
1995
. Ligand-induced autoregulation of IFN-γ receptor β chain expression in T helper cell subsets.
Science
270
:
1215
-1218.
13
Pernis, A., S. Gupta, K. J. Gollob, E. Garfein, R. L. Coffman, C. Schindler, P. Rothman.
1995
. Lack of interferon γ receptor β chain and the prevention of interferon γ signaling in TH1 cells.
Science
269
:
245
-247.
14
Gett, A. V., F. Sallusto, A. Lanzavecchia, J. Geginat.
2003
. T cell fitness determined by signal strength.
Nat. Immunol.
4
:
355
-360.
15
Boise, L. H., C. B. Thompson.
1996
. Hierarchical control of lymphocyte survival.
Science
274
:
67
-68.
16
Lohman, B. L., R. M. Welsh.
1998
. Apoptotic regulation of T cells and absence of immune deficiency in virus-infected γ interferon receptor knockout mice.
J. Virol.
72
:
7815
-7821.
17
Badovinac, V. P., A. R. Tvinnereim, J. T. Harty.
2000
. Regulation of antigen-specific CD8+ T cell homeostasis by perforin and interferon-γ.
Science
290
:
1354
-1358.
18
Tsujimoto, M., Y. K. Yip, J. Vilcek.
1986
. Interferon-γ enhances expression of cellular receptors for tumor necrosis factor.
J. Immunol.
136
:
2441
-2444.
19
Bogdan, C..
2001
. Nitric oxide and the immune response.
Nat. Immunol.
2
:
907
-916.
20
Fallarino, F., U. Grohmann, C. Vacca, R. Bianchi, C. Orabona, A. Spreca, M. C. Fioretti, P. Puccetti.
2002
. T cell apoptosis by tryptophan catabolism.
Cell Death Differ.
9
:
1069
-1077.
21
Mellor, A. L., D. H. Munn.
2004
. IDO expression by dendritic cells: tolerance and tryptophan catabolism.
Nat. Rev. Immunol.
4
:
762
-774.
22
Sedger, L. M., D. M. Shows, R. A. Blanton, J. J. Peschon, R. G. Goodwin, D. Cosman, S. R. Wiley.
1999
. IFN-γ mediates a novel antiviral activity through dynamic modulation of TRAIL and TRAIL receptor expression.
J. Immunol.
163
:
920
-926.
23
Zhu, C., A. C. Anderson, A. Schubart, H. Xiong, J. Imitola, S. J. Khoury, X. X. Zheng, T. B. Strom, V. K. Kuchroo.
2005
. The Tim-3 ligand galectin-9 negatively regulates T helper type 1 immunity.
Nat. Immunol.
6
:
1245
-1252.
24
Xu, X., X. Y. Fu, J. Plate, A. S. Chong.
1998
. IFN-γ induces cell growth inhibition by Fas-mediated apoptosis: requirement of STAT1 protein for up-regulation of Fas and FasL expression.
Cancer Res.
58
:
2832
-2837.
25
Hayashi, T., L. Beck, C. Rossetto, X. Gong, O. Takikawa, K. Takabayashi, D. H. Broide, D. A. Carson, E. Raz.
2004
. Inhibition of experimental asthma by indoleamine 2,3-dioxygenase.
J. Clin. Invest.
114
:
270
-279.
26
Munn, D. H., M. D. Sharma, J. R. Lee, K. G. Jhaver, T. S. Johnson, D. B. Keskin, B. Marshall, P. Chandler, S. J. Antonia, R. Burgess, et al
2002
. Potential regulatory function of human dendritic cells expressing indoleamine 2,3-dioxygenase.
Science
297
:
1867
-1870.
27
McNally, J. M., C. C. Zarozinski, M. Y. Lin, M. A. Brehm, H. D. Chen, R. M. Welsh.
2001
. Attrition of bystander CD8 T cells during virus-induced T-cell and interferon responses.
J. Virol.
75
:
5965
-5976.
28
Maldonado, R. A., D. J. Irvine, R. Schreiber, L. H. Glimcher.
2004
. A role for the immunological synapse in lineage commitment of CD4 lymphocytes.
Nature
431
:
527
-532.
29
Mullen, A. C., F. A. High, A. S. Hutchins, H. W. Lee, A. V. Villarino, D. M. Livingston, A. L. Kung, N. Cereb, T. P. Yao, S. Y. Yang, S. L. Reiner.
2001
. Role of T-bet in commitment of TH1 cells before IL-12-dependent selection.
Science
292
:
1907
-1910.
30
Whitmire, J. K., N. Benning, J. L. Whitton.
2005
. Cutting edge: early IFN-γ signaling directly enhances primary antiviral CD4+ T cell responses.
J. Immunol.
175
:
5624
-5628.
31
Ma, X., J. M. Chow, G. Gri, G. Carra, F. Gerosa, S. F. Wolf, R. Dzialo, G. Trinchieri.
1996
. The interleukin 12 p40 gene promoter is primed by interferon γ in monocytic cells.
J. Exp. Med.
183
:
147
-157.
32
Manoury-Schwartz, B., G. Chiocchia, N. Bessis, O. Abehsira-Amar, F. Batteux, S. Muller, S. Huang, M. C. Boissier, C. Fournier.
1997
. High susceptibility to collagen-induced arthritis in mice lacking IFN-γ receptors.
J. Immunol.
158
:
5501
-5506.