Sustained IL-12 Signaling Is Required for Th1 Development ¹ Free

Protective immunity against intracellular pathogens requires the differentiation of naive Th cells into Th1 cells, characterized by the production of IFN-γ (reviewed in Refs. 1 and 2). This cytokine promotes the effective eradication of this class of pathogens through the induction of antimicrobial activity in macrophages. A key player in governing the development of Th1 responses is IL-12, a cytokine largely produced by APC, including dendritic cells (reviewed in Refs.3, 4, 5). IL-12 signals through STAT4, and the central role of this transcription factor in skewing naive Th cells toward the Th1 phenotype has been well documented (reviewed in Refs. 5 and 6).

Type I IFNs (IFN-αβ) have also been shown to activate STAT4 in both human T cells and NK cells (7, 8, 9). Phosphorylation of STAT4 occurs upon its recruitment to the IFN-αβ receptor complex through the C-terminal region of STAT2 (10, 11). Initially it was thought that utilization of the STAT4 pathway by IFN-αβ was restricted to human cells, as the mouse STAT2 gene harbors a minisatellite insertion affecting the C-terminal region, selectively disrupting the ability of STAT2 to recruit STAT4 (11). However, two recent reports demonstrated that STAT4 can be directly phosphorylated by IFN-αβ in murine T cells, indicating the existence of a STAT2-independent pathway for STAT4 activation (12, 13).

Because of their ability to activate STAT4, IFN-αβ have been proposed to direct the development of naive Th cells into Th1 cells (reviewed in Refs. 5 ,14 , and 15). Hence, IFN-α favored the development of Ag-specific human T cell clones with a Th1, rather than a Th2, cytokine profile (16). It has also been noted that IFN-α inhibits IL-4-induced gene expression (17, 18, 19) and suppresses IL-4-induced Th2 development (20), which would skew T cell differentiation along the Th1 lineage. Accordingly, it has been suggested that IFN-α may be an alternative to the IL-12 route to evoke Th1 responses (5, 14, 15). However, most studies do not directly compare the Th1-driving potential of IFN-α vs IL-12 during naive Th cell differentiation. In this context there are several observations that indicate that the Th1-polarizing activity of IFN-αβ is not equivalent to that of IL-12. For example, in humans, polarized Th2 responses can be reversed by IL-12 (21, 22), but not IFN-α (20) (H. H. Smits, unpublished observations). Furthermore, patients deficient in IL-12R-mediated signaling fail to induce protective Th1 responses and suffer from recurrent Mycobacteria spp. and Salmonella spp. infections despite having an intact IFN-αβ signaling pathway (23, 24). Thus, IFN-αβ cannot compensate for IL-12 in vivo under circumstances where optimal Th1 responses are required. The current dogma that IFN-α skews Th1 cell phenotype development may, therefore, need some revision.

In the present study an APC-free human in vitro Th1/Th2 differentiation model was employed to analyze the direct actions of IFN-α and IL-12 on naive Th cell phenotype development. It was found that IL-12 and IFN-α differed quite markedly in their potency in terms of driving Th1 cell differentiation; the IFN-α Th1-polarizing response was much weaker than that of IL-12. One explanation for this could be that activation of STAT4 by IL-12 is more long-lived than that in response to IFN-α, as reported previously (25, 26, 27, 28). In this study we further studied the differential kinetics of STAT4 activation by IFN-α and IL-12. We show that STAT4 activation in response to IL-12 can be sustained over a long time (>48 h), whereas the response to IFN-α is transient (4 h). Termination of STAT4 activation by IFN-α is a cis response and does not involve members of the suppressors of cytokine signaling (SOCS)⁵ family, but correlates with the IFN-α-induced down-regulation of the cell surface expression of IFN-αβ receptor subunit 1 (IFNAR1), rendering cells refractory to IFN-α. The kinetics of the IL-12/STAT4 response is very different from that induced by IFN-α. Thus, sustained STAT4 activation can be achieved by culturing T cells in the continuous presence of IL-12. Another important difference between IFN-α and IL-12 is that the T cell growth factor IL-2 prolongs the IL-12/STAT4 response, but not the IFN-α/STAT4 response. Importantly, the ability of IL-12 and IL-2 to synergize for sustained STAT4 activation is reflected in a requirement for IL-2 for optimal IL-12-induced Th1 cell differentiation of naive Th cells. Collectively these results show that the duration of responses to cytokines can determine the biological response.

Buffy coats were purchased from the National Blood Service (South Thames Blood Bank, London, U.K.) or the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (Amsterdam, The Netherlands). Human PBMC were isolated by Lymphoprep (Nycomed, Torshov, Norway) density-gradient centrifugation. Subsequently, CD45RA⁺/CD45RO⁻ naive CD4⁺ T cells were isolated to high purity (>98%, as assessed by flow cytometry) through negative selection using the CD4⁺/CD45RO⁻ column kit (R&D Systems, Minneapolis, MN). Purified naive Th cells were stimulated in 96-well culture plates (1 × 10⁵ cells/well; Costar, Cambridge, MA) with immobilized anti-human CD3 mAb (CLB-T3/3; 1 μg/ml) and soluble anti-human CD28 mAb (CLB-CD28/1; 2 μg/ml, both obtained from CLB (Amsterdam, The Netherlands) in the presence or the absence of additional cytokines and/or neutralizing Abs to IL-2 (R&D Systems, Abingdon, U.K.) as indicated. All cultures were performed in IMDM (BioWhittaker, Walkersville, MD) supplemented with 10% FCS (HyClone Laboratories, Logan, UT) and gentamicin (80 μg/ml; Duchefa, Haarlem, The Netherlands). On day 14, quiescent Th cells were restimulated with PMA (10 ng/ml) and ionomycin (1 μg/ml; Sigma-Aldrich, St. Louis, MO) for 6 h, the last 5 h in the presence of brefeldin A (10 μg/ml; Sigma-Aldrich), to determine single-cell IL-4 and IFN-γ production by intracellular staining and flow cytometric analysis. Cells were fixed in 2% paraformaldehyde (Merck, Darmstadt, Germany), permeabilized with 0.5% saponin (ICN Biochemicals, Cleveland, OH), stained with anti-human IFN-γ-FITC and anti-human IL-4-PE (both from BD PharMingen, San Diego, CA), and analyzed by flow cytometry.

PBMC were activated in vitro with PHA (Murex Biotech, Dartford, U.K.) for 3 days and were maintained in RPMI 1640 medium, 10% heat-inactivated FCS, and 20 ng/ml human rIL-2 (Chiron, Emeryville, CA) for 1 wk using standard protocols (referred to as PBT cells). PBT cells were arrested in a quiescent state by washing twice in RPMI 1640 medium and replacing in human rIL-2-free RPMI 1640 medium and 10% heat-inactivated FCS for 48 h. Unless specified, stimulation of PBT cells was performed with 20 ng/ml human rIL-2, 20 ng/ml human rIL-12 (Hoffman-La Roche, Nutley, N.J.), or 10³ U/ml IFN-α (Wellferon; a mixture of natural IFN-α subtypes provided by Wellcome Research Laboratories (Beckenham, Kent, U.K.)).

Affinity precipitation of DNA binding proteins was performed with the optimal binding sequence for STAT4 (γ-activated sequence (GAS)-STAT4), GTGGCTTTCCGGGAATCCTTG (29, 30). Oligonucleotides of the 3′ and 5′ complementary strands of each sequence were synthesized at the Oligonucleotide Synthesis Service (Cancer Research UK). 5′ Strands contain a biotin moiety at their 5′ end. Annealed double-stranded oligonucleotides were used at 0.1 μg/μl in 10 mM Tris and 1mM EDTA, and 75 mM NaCl. Cells (2 × 10⁷/ml) were stimulated and lysed in 1% Nonidet P-40, 50 mM Tris (pH 7.9), 10 mM NaF, 100 mM EDTA, 10% glycerol, 150 mM NaCl, 1 mM PMSF, 1 mM Na₂VO₃, 1 mM DTT, and 1 μg/ml each pepstatin, chymostatin, leupeptin, and antipain. Cell nuclei and debris were eliminated by spinning for 5 min at 10,000 × g. Total protein content in the lysate was standardized using a protein assay (Bio-Rad, Hercules, CA), according to the manufacturer’s instructions, to 750-1000 μg/sample. Lysates were diluted with NaCl-free lysis buffer to 15 mM NaCl and precleared with 20 μl of streptavidin agarose beads in a wheel mixer for 15 min at 4°C. Precleared lysates were incubated with 1 μg of the corresponding double-stranded oligonucleotide and 30 μl of streptavidin agarose beads for 2 h at 4°C in a wheel mixer. Beads containing the oligonucleotide-bound proteins were washed three times with cold lysis buffer and resuspended in reducing sample buffer. All spins were performed at 5000 × g for 5 min. For the analysis of STAT4 protein levels, equal amounts (usually 100 μg) of total protein were acetone-precipitated from the total cell lysate and resuspended in reducing sample buffer.

Proteins were separated by SDS-PAGE (Protogel; National Diagnostics, Atlanta, GA) using 7.5% acrylamide/0.03% bis. Proteins were transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore, Bedford, MA) and blocked in 5% dried milk and 0.05% Tween 20 (BDH, Poole, U.K.)/PBS. The following Abs were used: anti-STAT4 (C20; Santa Cruz Biotechnology, Santa Cruz, CA) and anti-STAT1 pan mAb, anti-STAT3 pan mAb, anti-STAT5 pan mAb (all from BD Transduction Laboratories, Lexington, KY). HRP-conjugated anti-mouse or anti-rabbit (Amersham Pharmacia Biotech, Arlington Heights, IL) Abs were respectively used. Proteins were visualized using the ECL detection system (Amersham Pharmacia Biotech). For sequential detection, membranes were stripped in 100 mM 2-ME, 2% SDS, and 62.5 mM Tris (pH 6.8) for 45 min at 50°C.

RNase protection assays were conducted as described previously (31). The human SOCS multiprobe template set was purchased from BD PharMingen. Briefly, probes were synthesized from T7 transcription vectors and labeled with [³²P]UTP to a sp. act. of 2–5 × 10⁸ cpm/μg of input DNA. Aliquots equivalent to 1–3 × 10⁵ cpm of each probe/assay and 5–10 μg of total RNA/assay were used.

Abs were obtained conjugated to PE, and nonconjugated Abs were revealed using biotinylated anti-mouse, anti-rabbit, or anti-rat Abs, respectively, followed by incubation with streptavidin-Tricolor or streptavidin-PE. For staining, 2 × 10⁶ cells/sample were incubated with saturating concentrations of corresponding Abs at 4°C for 20 min. All incubations were performed in a 96-well, V-bottomed microtiter plate in 100 μl of 1% BSA/PBS. Cells were washed three times with 1% BSA/PBS between incubations and before analysis. The following Abs were used: anti-human CD25-PE (rabbit; BD PharMingen), anti-human IFNAR1 (4B1; mouse) and IFNAR2 (R2.2; rabbit polyclonal). The anti-human IL-12Rβ1 (2B10; rat IgG2a) and IL-12Rβ2 (2B6; rat IgG2a) were provided by Dr. F. Sinigaglia (Roche, Milan, Italy). The following Abs of unknown specificity were used as isotype-matched controls: anti-rat IgG2aκ (for IL-12Rβ1/β2), whole rabbit IgG (for IFNAR2; Jackson ImmunoResearch Laboratories, West Grove, PA), and whole mouse IgG (for IFNAR1; Sigma-Aldrich). Stained cells were analyzed on a FACSCalibur (BD Biosciences, Mountain View, CA). Events were collected and stored ungated in list-mode using CellQuest (BD Biosciences) software. Live cells were gated according to their forward/side scatter profiles and by eliminating dead cells that positively stained with the DNA binding dye TO-PRO3 (Molecular Probes). Data were analyzed using CellQuest software.

Binding of IL-12 to its receptor results in the masking of the IL-12Rβ2 epitope that is recognized by the anti-IL-12Rβ2 Ab (2B6). To analyze IL-12Rβ2 cell surface expression on IL-12-stimulated cells, acid wash was used to remove IL-12. Stimulated cells were washed twice in ice-cold PBS before and after being resuspended in ice-cold RPMI 1640 medium (pH 4.0) for 1 min. All spins were performed for 20 s at 10,000 × g at 4°C. The structure and Ab recognition of both chains of the IL-12R were not altered by this procedure. IL-12 removal was confirmed by comparison with a PBS-only-treated aliquot of each sample using anti-IL-12Rβ2 staining and flow cytometric analysis.

The Th1-driving ability of IFN-α vs IL-12 was assessed in an in vitro APC-free human Th1/Th2 differentiation model (32, 33). Activation of human naive Th cells in the absence of polarizing factors (neutral) yielded a mixed population of IFN-γ and/or IL-4 producers (Fig. 1). In this model, clear-cut development of strongly polarized Th1 and Th2 cells could be achieved by the cytokine IL-12 or IL-4, respectively. IFN-α, however, only marginally affected naive Th cell development. Even at the highest concentration of IFN-α (1000 U/ml, optimal for STAT4 activation; data not shown), the proportion of IFN-γ-producing cells only doubled compared with cells cultured under neutral conditions. Moreover, the percentage of IL-4 producers was not significantly reduced by IFN-α. In contrast, IL-12 yielded >90% of IFN-γ producers and strongly impaired the development of IL-4-producing cells. Thus, in comparison with IL-12, IFN-α only poorly skews the development of naive Th cells toward the Th1 phenotype.

Differences in the regulation of STAT4 by IL-12 and IFN-α may underlie the poor ability of IFN-α to induce Th1 cell responses. However, the initial kinetics and magnitude of the IFN-α/STAT4 and IL-12/STAT4 responses are indistinguishable (7, 9, 20). Nevertheless, Th cell differentiation is a process that takes several days, and it is possible that differences in the duration of signaling by cytokines could explain differences in their biological actions. It has been reported that induction of STAT4 tyrosine phosphorylation in response to IL-12 is more sustained than that in response to IFN-α (25, 26, 28). However, more recently it was proposed by Wang et al. (27) that IL-12 signaling in T cells is terminated within 24 h through the specific degradation of phosphorylated STAT4. Accordingly, we compared the long term kinetics of IL-12 and IFN-α signal transduction through their ability to activate STAT4, a key transcription factor for Th1 cell differentiation. IFN-αβ receptors are expressed constitutively on naive T cells, whereas IL-12Rs are restricted to activated cells. Therefore, a comparison of the kinetics of IFN-α- and IL-12-STAT4 responses was performed in T lymphoblasts (PBT cells; see Materials and Methods) that express both IL-12 and IFN-α receptors. Affinity precipitation using biotinylated, double-stranded oligonucleotides bearing the optimal STAT4 DNA binding site GAS-STAT4 (GTGGCTTTCCGGGAATCCTTG) coupled to streptavidin agarose beads was used to study cytokine-induced STAT4 activation through DNA binding. PBT cells were rested for 48 h and then stimulated with IL-12 or IFN-α over a 9-h period. Cell extracts were used for affinity precipitation, and the presence of STAT4 was monitored by Western blot analysis using specific anti-STAT4 Abs. Two forms of activated STAT4 with distinct electrophoretic mobility can be observed up to 1 h of cytokine treatment: a more rapidly migrating STAT4p1, phosphorylated on tyrosine, and a more slowly migrating STAT4p2, phosphorylated on both tyrosine and serine (7, 9). At later time points (>1 h) only the STAT4p2 form can be observed (7, 9). The data show that no STAT4 protein was detected in GAS-STAT4 complexes isolated from quiescent T cells, but for both IL-12 and IFN-α, high and comparable levels of activated STAT4 could be observed after 3 h of cytokine treatment (Fig. 2,A). Assessing STAT4 activation at later time points, however, revealed major differences in the long term kinetics of IL-12 vs IFN-α responses; IL-12-induced STAT4 activation was maintained for at least 9 h, whereas the IFN-α/STAT4 response was terminated after 3 h. In additional experiments it was seen that IL-12 activation of STAT4 was also detectable 24 and 48 h after IL-12 stimulation, although at these time points the response had declined to ∼20–30% of maximal (see below). In contrast, transient activation of STAT4 by IFN-α, lasting for only 3–4 h, was consistently observed (Fig. 2 A).

Levels of activated STAT4 are determined by the rate of production of active STAT4 vs its dephosphorylation/degradation. We therefore examined whether the prolonged activation of STAT4 by IL-12 could be induced by a pulse of IL-12 or was dependent on the continuous presence of the cytokine. T cells were pulsed for 1 h with IL-12, washed to remove the cytokine, and then cultured for various times in the absence or the presence of IL-12. Upon withdrawal of IL-12, the levels of activated STAT4 decreased and were completely lost after 4 h (Fig. 2 B). These results show that the half-life for active STAT4 is relatively short (1–2 h), and continued exposure to IL-12 is required to maintain STAT4 activation in T cells.

The possibility that transient IFN-α/STAT responses were due to the apoptotic and/or antiproliferative properties described for IFNs in nonlymphoid cells (34) was excluded by flow cytometric analysis on propidium-stained cells and by cell counting. Rather than inducing apoptosis, IFN-α treatment for both short (1 to 6 h) and long (24 to 72 h) periods of time improved the survival of PBT cells considerably (data not shown). Hence, IFN-α promotes the survival of lymphoid cells, as described previously (35, 36).

The fact that sustained activation of STAT4 by IL-12 required the constant presence of the cytokine made us question whether the transient STAT4 activation response seen in IFN-α-treated T cells reflected that this cytokine became depleted from cell cultures. To examine this possibility, T cells were preincubated for 4 h in the absence or the presence of IFN-α. At this time point the primary response to IFN-α was extinguished. The consequences of further addition of IFN-α were then examined. In this experiment the activity of other IFN-α-activated STATs, notably STAT1, 3, and 5, was also monitored by their ability to bind to the same GAS-STAT4 oligonucleotide (Fig. 3,A). If cells had not previously seen IFN-α they gave the expected high level of STAT activation (Fig. 3,A, third lane). However, if they had been preincubated with IFN-α for 4 h, they were unable to reinduce STAT activation in response to a second stimulation with this cytokine (Fig. 3, last lane). Thus, transient IFN-α responses are not due to extinction of IFN-α from the medium, and IFN-α pretreatment renders cells refractory to further IFN-α stimulation.

An interesting question is whether IFN-α-treated T cells are refractory to STAT4 activation because they induce a negative regulator of STATs. One possibility for such negative regulators is the SOCS family of proteins that is recognized as one of the main negative regulatory mechanisms in cytokine signaling (reviewed in Ref.37). SOCS expression is induced upon cytokine stimulation, and SOCS1, -3, and -5 have recently been implicated in the regulation of Th1/Th2 development (38, 39, 40). The impact of IFN-α on the induction of SOCS proteins in T cells has not been examined previously. Therefore, we analyzed the expression of SOCS family members in T cells activated by IFN-α. As a positive control in these experiments the effect of the cytokine IL-2 was also examined, as this is a known inducer of the SOCS family in T cells. The results of RNase protection assays show that both IL-2 and IFN-α induced the expression of SOCS1, SOCS3, and cytokine-inducible SH2-containing protein (CIS) (Fig. 4), whereas SOCS5, -6, and -7 were not induced in these T cells at the time points analyzed. SOCS2 was only marginally induced by IL-2 and IFN-α. SOCS1, SOCS3, and CIS were also induced by IL-12, but to a much lesser extent than observed in response to IFN-α and IL-2.

The induction of SOCS1, SOCS3, and CIS by IFN-α could explain the loss of STAT4 activation in IFN-α-treated cells. However, these molecules are able to inhibit in trans (38, 41). Hence, if up-regulation of SOCS proteins were the explanation for transient activation of STAT4, then it would be predicted that IFN-α pretreatment of cells would prevent subsequent activation of STAT4 by IL-12. To examine this possibility, T cells were preincubated for 4 h in the absence or the presence of IFN-α and then rechallenged with either IL-12 or IFN-α. As shown above (see Fig. 3,A), IFN-α could only induce STAT4 activation in cells when it was given as a primary stimulus, but it had no effect in cells that had been preincubated with IFN-α (Fig. 3,B, third and fourth lanes). In contrast, IL-12 strongly activated STAT4 regardless of whether cells had been preincubated with IFN-α (Fig. 3 B, last two lanes). These data indicate that loss of STAT4 responses in IFN-α-treated cells is a cis response, and IFN-α does not trans-inhibit the IL-12/STAT4 response in T cells.

One explanation for the loss of IFN-α signaling could be down-regulation of IFN-αβ receptor expression, as it is well established that IFN-αβ occupancy of its receptor induces receptor internalization and degradation (42). This prompted the question of whether transient vs sustained signaling by IFN-α and IL-12 could be explained by differences in the metabolism of these receptors. The data in Fig. 5,A show what happened to IFN-αβ and IL-12R expression when cells were maintained in the presence of their respective cytokine. The IFN-αβ receptor is a dimer comprised of IFNAR1 and IFNAR2 subunits. Surface levels of IFNAR2 stayed constant when cells were exposed to IFN-α, but surface levels of the IFNAR1 subunit were down-regulated at 4 h, and this down-regulation was maintained at 24 h. The IL-12R has two subunits, β1 and β2; the expression of the β2 subunit is rate-limiting for IL-12 responses (8, 43). Cell surface levels of IL-12Rβ2 were reduced after 18 min of exposure to IL-12 and decayed even further by 50–60% after 4 h of IL-12 stimulation. However, this down-regulation was transient, and IL-12Rβ2 cell surface expression recovered and was then increased at 24 h (∼130%) and 48 h (∼110%) in the presence of IL-12 (Fig. 5 B, right panels). The IFN-α- and IL-12-induced down-regulation of their receptors was ligand specific, as these cytokines did not reduce cell surface expression of CD25 (data not shown).

The cytokine IL-2 has been shown to synergize with IL-12 to regulate both NK cell and cytotoxic T cell differentiation and cytolytic activity. Notably, IL-2 has been previously reported to increase IL-12Rs and STAT4 levels in NK cells (26). The experiment presented in Fig. 6,A explores the effect of IL-2 on IL-12 activation of STAT4. PBT cells were left unstimulated or were stimulated with IL-12, IL-2, or their combination for different periods of time. The data show that IL-2 did not influence the immediate IL-12/STAT4 response, but strongly potentiated the long term IL-12/STAT4 response. This ability of IL-2 to synergize with IL-12 for STAT4 activation was consistently observed. We next examined whether IL-2 could synergize with IFN-α to activate the IFN-α/STAT response. PBT cells were left unstimulated or were stimulated with IFN-α in the presence or the absence of IL-2 for different periods of time. The results in Fig. 6 B show that IL-2 did not enhance or prolong the kinetics of the IFN-α/STAT4 response or prolong IFN-α-induced phosphorylation of STAT3 and STAT1.

One way that IL-2 could potentiate STAT4 activation is by regulation of expression of this transcription factor. In this respect, Wang et al. (27) have proposed that IL-12 signaling in T cells is terminated within 24 h through the specific degradation of phosphorylated STAT4. Our Western blot analysis of titrated cell lysates with STAT4 antisera show that in T cells cultured with IL-12 alone the loss of cellular levels of STAT4 was only marginal (Fig. 6 C). However, IL-2 quite clearly up-regulated STAT4 expression in either the presence or the absence of IL-12.

The ability of IL-2 to up-regulate cellular levels of STAT4 would contribute to the synergy seen between IL-2 and IL-12 for STAT4 activation, but cannot be the sole explanation for this phenomenon because of the selectivity of IL-2 for IL-12, but not IFN-α, responses. IL-2 has been shown to up-regulate the expression of IL-12Rs in NK cells (26). The results in Fig. 7,A confirm that IL-2 up-regulated IL-12Rβ1 and -β2 levels in T cells. In contrast, IL-2 did not up-regulate the expression of IFN-αβ receptor subunits (Fig. 7 B) and was not able to rescue IFN-α-induced down-regulation of IFNAR1 levels (data not shown). The differential actions of IL-2 on IL-12 vs IFN-αβ receptor expression thus correlates well with the biological synergy seen with IL-2 and IL-12, but not IL-2 and IFN-α.

The results presented above show that IFN-α and IL-12 differ in their kinetics of signaling, correlating with their Th1-polarizing activities: IL-12 induces a sustained STAT4 response and strongly drives Th1 cell development, whereas IFN-α responses are transient and only lead to poor Th1 cell polarization. The results furthermore show that IL-2 can prolong IL-12 signaling. This predicts that IL-2 will synergize with IL-12 for optimal Th1 development. The experiment in Fig. 8 explores the importance of IL-2 for IL-12-induced IFN-γ production in naive Th cells and for IL-12-induced Th1 differentiation. Naive Th cells were activated in the absence or the presence of IL-12, IL-2, and neutralizing anti-IL-2 Abs for 72 h. Supernatants were harvested to measure IFN-γ levels in these primary cultures. As shown in Fig. 8,A, high levels of IFN-γ were only observed when naive Th cells had been activated in the presence of both IL-12 and IL-2. Neutralization of endogenous IL-2 completely abolished IL-12-induced IFN-γ production. Furthermore, naive Th cells activated in the presence of IL-2 alone produce low levels of IFN-γ. Subsequently, all experimental groups were cultured for another 11 days in the presence of exogenous IL-2 to allow for their expansion and survival. All T cell cultures expanded to a similar extent, as assessed by cell counting (data not shown). Resting cells were restimulated, and their IL-4/IFN-γ cytokine production profile was assessed (Fig. 8 B). Optimal Th1 cell development was only achieved when naive Th cells had been activated in the presence of both IL-12 and IL-2.

Next it was determined whether sustained IL-12 signaling is required for optimal Th1 cell differentiation. Naive Th cells were activated and cultured constantly in the presence of IL-2, but were exposed to IL-12 for 16, 24, 48, and 72 h. Sixteen hours was chosen as the earliest time point because both functional IL-12Rs and STAT4 are absent in naive Th cells and are induced upon TCR triggering. Efficient STAT4 activation in response to IL-12 can be observed 12 h after T cell activation (44) (data not shown). The data in Fig. 9 show that the time of exposure of naive Th cells to IL-12 markedly influences the ability of T cells to differentiate toward the Th1 phenotype. A brief 16-h exposure to IL-12 did not markedly synergize with IL-2 to induce Th1 cell development. Rather, prolonged exposure (72 h) to IL-12 was necessary to promote optimal Th1 cell differentiation.

A key step in Th1 cell differentiation is activation of the transcription factor STAT4. In human T cells, both IFN-α and IL-12 activate STAT4, yet IFN-α cannot substitute for the loss of IL-12 function in vivo, as judged by the fact that loss of IL-12R expression in humans results in a failure to induce protective Th1-mediated immune responses (23, 24). The present study yields some insight as to why IFN-α cannot substitute for IL-12 in vivo. Thus, the present data demonstrate that in comparison with IL-12, the capacity of IFN-α to directly induce Th1 phenotype commitment in naive Th cells is only poor. The differential action of these cytokines may be explained by kinetic differences in signal transduction. Hence, IL-12 induces a sustained signaling through STAT4 in human T cells, whereas the IFN-α response is transient. A second difference between IFN-α and IL-12 is that the kinetics of the IL-12/STAT4 response can be prolonged by the cytokine IL-2, whereas IL-2 cannot augment or prolong IFN-α-induced STAT4 activation.

The differential kinetics of STAT4 activation by IFN-α and IL-12 have been described previously (25, 26, 27, 28), but whether the observed difference explains the different abilities of IFN-α and IL-12 to polarize naive Th cell differentiation was not addressed. In the present study this issue was examined experimentally by determining whether transient exposure of naive Th cells to IL-12 was required for optimal Th1 differentiation. The data show that optimal Th1 differentiation requires prolonged exposure to IL-12 and cannot be induced by a transient pulse of IL-12. Further evidence that sustained IL-12 signaling is necessary for Th1 differentiation is provided by the observation that IL-2, which is shown in this study to prolong IL-12 activation of STAT4, synergizes with IL-12 to drive naive Th cell differentiation to the Th1 lineage. Thus, the duration of STAT4 activation in response to IL-12 correlates with the ability of this cytokine to drive Th1 cell development.

An interesting question is the mechanism that underlies the different kinetics of IFN-α vs IL-12 signal transduction. The present study has excluded that the early termination of STAT4 activation is due to depletion of cytokine from the media. Wang et al. (27) described that phosphorylated STAT4 is targeted for degradation by proteasomes, and we considered that IFN-α might exhaust cytosolic pools of STAT4. However, this does not seem to occur, as the IL-12/STAT4 response is not altered/diminished in cells that are refractory to IFN-α activation of STAT4 following prolonged exposure to IFN-α. This latter observation also excludes that IFN-α/STAT4 responses terminate because of the expression of negative regulators of STAT activation, such as SOCS1, SOCS3, and CIS. IFN-α does induce the expression of these proteins, but not at a level that blocks IL-12-induced activation of STAT4. Moreover, if induction of CIS, SOCS1, and SOCS3 were able to negatively regulate STAT4 activation, then it would have been predicted that IL-2, which potently induces these proteins, would antagonize STAT4 induction. However, IL-2 has no impact on the IFN-α/STAT4 response (Fig. 6 B and data not shown) and even potentiates activation of STAT4 in response to IL-12.

One explanation for the different signaling kinetics for IFN-α vs IL-12 almost certainly lies in differences in the regulation of IFN-α and IL-12 receptors by their cognate cytokine. The present data show that IFN-α down-regulates the expression of its receptor. Indeed, IFN-α-induced down-regulation of its receptor by induction of receptor internalization and degradation is a long-established phenomenon (42). That this receptor loss terminates IFN-α signaling has been revealed in experiments in which removal of the IFNR tyrosine activation motif domain in the intracellular region of the IFNAR1 receptor subunit was shown to prevent receptor internalization and prolong activation of STATs (45). The action of IL-12 on its receptor is quite different; initially ligand occupancy of the IL-12R does cause receptor down-regulation, but this is only transient. It is transient because of a well-documented positive feedback loop by which IL-12 induces up-regulation of the expression of the IL-12Rβ2 subunit (8, 22). In a similar context, the differential effect of IL-2 on IL-12- vs IFN-α-induced STAT4 activation is probably also explained by differential receptor regulation. Thus, IL-2 promotes the expression of IL-12β1 and β2 subunits, but has no impact on IFN-αβ receptor levels. IL-2 is an important cytokine for naive Th cell development. It is produced by T cells early upon TCR triggering in a cell cycle-independent fashion (46), and it may also be provided by dendritic cells (47). IL-2 is required for the polarization of both Th1 and Th2 cells. It drives cell cycle progression in activated T cells, which is required for the stable induction of both IFN-γ and IL-4 (46, 48). In addition, as shown in the current report, it supports IL-12-induced Th1 development through enhancing IL-12R expression and STAT4 levels. A similar role for IL-2 in supporting IL-12-induced IFN-γ production and cytolytic activity has been proposed for NK cells (26). The up-regulatory effect of IL-2 on the β1 chain of the IL-12R is probably pivotal for facilitating sustained IL-12 signaling; this subunit is down-regulated upon IL-12 treatment, and unlike IL-12Rβ2, its expression is not induced by IL-12 itself.

An interesting question is whether the short-lived IFN-α/STAT4 response has any physiological relevance. In this context, a pivotal role for IFN-αβ-induced STAT4 phosphorylation was recently demonstrated in an in vivo virus infection model (13). Induction of protective high levels of IFN-γ by CD8⁺ T cells in response to infection with lymphocytic choriomeningitis virus was dependent on STAT4 activation by IFN-αβ. With respect to Th1/Th2 differentiation, the importance of the transient IFN-α response may reside in the fact that naive Th cells express IFN-αβ receptors, whereas IL-12R expression is induced by engagement of Ag and costimulatory receptors. IFN-α could thus be important in initiating Th cell polarization toward the Th1 lineage in naive Th cells. However, in the absence of IL-12, this response would not be sustained. IFN-α and IL-12 can thus deliver temporally overlapping signals for Th1 cell development; the transient nature of the IFN-α response is a mechanism to ensure that Th1 responses are determined by the long term availability of IL-12 or other cytokines, such as IL-18, which act in concert with IFN-α to induce IFN-γ production. During viral infection IL-18 is a likely candidate to play a critical role, as it is produced upon viral infection and strongly synergizes with IFN-α in inducing IFN-γ production by T and NK cells (12, 28, 49).

In summary, the present data underline the importance of the duration of cytokine signaling for biological responses. This was first established for the cytokine IL-2, when it was shown that the duration of IL-2 exposure was a critical determinant for T cell mitosis (50). It is also known that the duration of Ag receptor engagement is important for immune outcome; transient exposure to Ag is insufficient to stimulate a full effector/memory T cell immune response (51). The present results reveal that sustained IL-12 signaling is necessary for Th1 cell differentiation. This has important implications because IL-12 is not made by Th cells, but is produced by APC, notably dendritic cells. This would mean that Th cells would need to remain in close proximity to IL-12-producing APC to ensure differentiation to the Th1 lineage. Any signal that potentiates or inhibits the ability of Th cells to stay in proximity to cells producing IL-12 would thus have a big impact on the ability of Th cells to differentiate toward the Th1 phenotype.

We thank Caetano Reis e Sousa and Ian Kerr for critically reading the manuscript.