Abstract
Commitment of Th lymphocytes to the Th1 phenotype, as characterized by the expression of the major proinflammatory cytokine IFN-γ, may be critically involved in the establishment of chronic inflammation and inflammatory autoimmune disease. To date, it has been shown that in IL-12-stimulated murine Th cell lines containing a major fraction of Th1 cells, Th2 cells can be induced by IL-4 until about 2 wk after initial activation, but not later. Here we analyze, based on the magnetic isolation of viable Th1 cells according to their specific expression of IFN-γ, the cytokine commitment of individual Th1 cells. After activation of naive Th cells with Ag and IL-12 for up to 5 wk, isolated IFN-γ-producing cells were restimulated with Ag and IL-4. Within the first 3 to 4 wk of IL-12 stimulation, some IFN-γ+ cells stopped expression of IFN-γ when restimulated with IL-4. However, within only 1 to 2 wk of IL-12 stimulation, few IFN-γ+ cells could be converted to produce IL-4. Others continued to express IFN-γ and thus were already committed to a proinflammatory, Th1-like phenotype. Surprisingly, within 3 wk of IL-12 stimulation, many of the IFN-γ-producing cells responded to IL-4 restimulation by expression of IL-10, but neither IFN-γ nor IL-4, i.e., by conversion to a suppressive, anti-inflammatory phenotype.
Thelper cells and their cytokines are key regulators of immune reactions. Differential expression of Th1 or Th2 cytokines, as characterized by the cytokines IFN-γ and IL-4, contributes to resistance or susceptibility to particular pathogens (1, 2, 3, 4, 5, 6). Expression of distinct Th1 or Th2 cytokines is also associated with and may contribute to pathologic immune reactions against allergens or autoantigens, i.e., allergy and autoimmunity (7, 8, 9). Understanding plasticity and stability of expression of cytokines in Th cells is thus of obvious importance for the development of vaccination strategies and therapeutic concepts as well as our basic understanding of chronic immune reactions.
Th differentiation can be efficiently polarized into a Th1- or a Th2-like phenotype in vitro (10, 11, 12, 13) as well as in vivo by IL-12 or IL-4, respectively (14, 15, 16). Treatment of Leishmania major-infected BALB/c Th2-type responder mice with anti-IL-4 mAb or IL-12 at the time of infection prevents the development of a fatal Th2 response and instead induces a protective Th1 reaction (15, 16, 17). However, already about 1 wk after infection, IL-12 or anti-IL-4 can no longer convert the Th2 into a protective Th1 response (16). After 1 wk of stimulation with IL-4, murine Th2 populations polarized in vitro can no longer be converted to the Th1 phenotype by restimulation with IL-12 (18, 19, 20), probably due to a defect in IL-12 signal transduction (18). IFN-γ seems to be able to restore IL-12 responsiveness under these conditions (21, 22, 23). Polarized murine Th1 populations can be converted to Th0/Th2-like populations by IL-4 after 1 wk of stimulation with IL-12 in vitro (18, 19, 20). After about 3 wk, however, Th1 populations become committed as well and no longer respond to IL-4 (20).
Until now, conversion of Th1 and Th2 responses has been analyzed on the level of polarized populations of Th cells. However, even highly polarized Th1 and Th2 populations are still heterogeneous with respect to cytokines they express, in that at least 10% and often >20% of the Th cells do not produce the cytokines in question (24, 25, 26). Therefore it has not been clear whether conversion of Th cell populations is the result of conversion of individual Th1 or Th2 cells or of selective outgrowth and differentiation of cells previously not expressing cytokines. Conversely, the stability of Th1 and Th2 populations observed after repeated polarization at later time points could be the result of selective survival of cells responding to the initial polarization.
Here, we have generated in vitro and isolated viable Th1 cells, producing IFN-γ, and analyzed the stability of their cytokine expression under converting conditions. Naive murine Th cells were stimulated for up to 5 wk with Ag and IL-12 in vitro. From such polarized populations, still containing about 30% of cells expressing neither IFN-γ nor IL-4; 70% producing IFN-γ, but not IL-4; and <1% expressing IL-10, IFN-γ-producing cells were isolated according to expression of surface IFN-γ (suIFN-γ), using magnetofluorescent liposomes for its detection (27, 28). The IFN-γ-producing cells were restimulated with Ag and IL-4 and then analyzed for expression of IFN-γ, IL-4, IL-5, and IL-10.
Materials and Methods
Mice
Mice homozygously transgenic for the DO11.10 αβ-TCR (OVA-TCRtg) (29) on BALB/c background were a gift from Dennis Y. Loh (Washington University School of Medicine, St. Louis, MO) and were bred under specific pathogen-free conditions in laminar flow incubators. Specific pathogen-free BALB/cJ mice (8–12 wk) were obtained from the Bomholtgard Breeding and Research Center (Ry, Denmark). Mice were killed by cervical dislocation for the isolation of spleen cells (SC).4
Isolation of CD62L+ CD4+ T cells by MACS MultiSort
OVA-TCRtg SC were stained with FITC-conjugated anti-CD4 mAb (GK1.5/4 (30)) and MultiSort anti-FITC microbeads (Miltenyi Biotec). CD4+ cells were isolated by positive selection on VS+ columns using the high gradient magnetic cell separation system MACS (Miltenyi Biotec), as described previously (31). Then, microbeads were cleaved off enzymatically and the CD4+ cells were stained for CD62L (L-selectin) with digoxigenized (DIG) anti-CD62L mAb (MEL-14 (32)) and anti-DIG MACS microbeads. CD62L+ CD4+ cells were then purified to 99% by positive MiniMACS selection on MS+ columns.
Stimulation of OVA-TCRtg Th cells in vitro
Cells were cultured at 1 to 2 × 106 cells/ml in complete RPMI 1640 (Life Technologies, Grand Island, NY) containing 100 U/ml penicillin, 0.1 mg/ml streptomycin, 0.3 mg/ml glutamine, 10 μM 2-ME, and 5% FCS (PAA, Linz, Austria).
The antigenic peptide OVA323–339 (Neosystem, Strasbourg, France) was used at a 1 μM concentration. OVA-TCRtg SC or congenic BALB/c SC were depleted of T cells with CD4-, CD8-, and Thy1.2- or CD5-microbeads by MACS (Miltenyi Biotec). Ninety-nine percent pure T cell-depleted SC (T−SC) were used as APC for OVA-TCRtg T cells at a 5:1 ratio. Recombinant murine IL-12 (gift from Maurice Gately, Hoffmann-La Roche, Nutley, NJ) was added at 100 U/ml, IL-4 (culture supernatant of NIH-3T3, transfected with murine IL-4 cDNA; gift from Werner Müller, Institute for Genetics, University of Cologne, Cologne, Germany) at 30% (v/v), and neutralizing anti-IL-4 mAb 11B11 (33) at 10 μg/ml as indicated. For restimulation, PMA and ionomycin (both from Sigma, St. Louis, MO) were used at 5 ng/ml and 1 μg/ml, respectively. Brefeldin A (Sigma) was added in some experiments at 5 μg/ml, 2 h before fixation.
Abs and flow cytometry
For intracellular staining, surface staining, or ELISA the following rat anti-mouse cytokine mAb were used: anti-IL-2 JES6-5H4 and JES6-1A12 (34, 35); anti-IL-4 11B11 (33), BVD4-1D11 and BVD6–24G2 (34, 35); anti-IL-5 TRFK4 and TRFK5 (36); and anti-IFN-γ AN18.17.24 (37), R4-6A2 (38), anti-IL-10 JES5-2A5 (34), and SXC1 and SXC2 (39).
FACScan and CellQuest research software (Becton Dickinson, Mountain View, CA) were used for flow cytometry. Dead cells were excluded according to light scatter and staining with propidium iodide (0.3 μg/ml).
Detection and isolation of IFN-γ-producing cells according to suIFN-γ
suIFN-γ, as specific marker for IFN-γ-expressing cells, was detected on activated Th cells using magnetofluorescent liposomes, as described previously (27, 28). Briefly, cells were incubated with purified rat IgG (0.1 mg/ml; Nordic, Tilburg, The Netherlands) in PBS/BSA/NaN3 at 4°C for 10 min to block unspecific binding of rat mAb. Then the cells were labeled with DIG-conjugated AN18.17.24 (5 μg/ml) in PBS/BSA/NaN3 at 4°C for 10 min. Cells were washed twice and stained with sheep anti-DIG Fab fragments conjugated to magnetofluorescent liposomes in PBS/BSA/NaN3 for 30 min at 4°C with gentle agitation. Cells were washed, resuspended, and analyzed by flow cytometry. suIFN-γ-positive cells were enriched to 99% purity on VS+ columns by MACS.
Intracellular staining of cytokines
Cells were fixed, permeabilized, and stained for intracellular cytokines and surface markers as described previously (24). For detection of IL-4, IFN-γ, and IL-10, the mAb 11B11, R4-6A2, AN18.17.24, and JES5-2A5 were used as DIG- and/or nitrophenylacetyl-conjugated Ab with FITC-conjugated sheep anti-DIG Fab (Boehringer Mannheim, Mannheim, Germany) and/or phycoerythrin-conjugated anti-nitrophenylacetyl mAb S43-10 (40) as secondary reagents. Stained cells were analyzed by flow cytometry.
ELISA for murine IL-2, IL-4, IL-5, IFN-γ, and IL-10
Concentrations of IL-2, IL-4, IL-5, IFN-γ, and IL-10 in the culture supernatants were determined by sandwich ELISA as described previously (41, 42), using mAb JES6-5H4, BVD4-1D11, TRFK5, R4-6A2, and SXC2 as coating Ab and biotinylated mAb JES6-1A12, BVD6-24G2, TRFK4, AN18.17.24, and SXC1 as developing Ab, respectively. The detection limits of the ELISA were 0.1 ng/ml for IL-2, 0.2 to 0.4 ng/ml for IL-4, 0.6 to 3 ng/ml for IL-5, 1.6 to 2 ng/ml for IL-10, and 0.1 to 0.2 ng/ml for IFN-γ, respectively.
Results
Isolation of naive CD62L+ CD4+ T cells by MACS MultiSort
Naive Th cells were isolated from DO11.10 TCR transgenic (OVA-TCRtg) mice by positive selection of CD62L+ CD4+ splenic lymphocytes by multiparameter magnetic cell sorting with MACS MultiSort. First, CD4+ T cells were specifically labeled with superparamagnetic MultiSort microbeads and enriched up to 99% purity (Fig. 1⇓). The microbeads were then cleaved off from the cell surface enzymatically. From the purified CD4+ cells, CD62L+ cells were isolated in a second magnetic separation step (Fig. 1⇓). Thus, CD62L+ CD4+ T cells were enriched to a purity of 99%. CD62L+ CD4+ cells expressed IL-2, but no detectable IL-4, IL-5, IL-10, or IFN-γ, upon stimulation with PMA and ionomycin for 4 h, as analyzed by intracellular cytokine staining (data not shown).
Isolation of naive CD62L+ CD4+ T cells by MACS MultiSort. OVA-TCRtg SC were stained with FITC-conjugated anti-CD4 mAb and MultiSort anti-FITC-microbeads. CD4+ cells were isolated by MACS. Then microbeads were cleaved off enzymatically, and the CD4+ cells were stained for CD62L (MEL-14) with specific MACS microbeads. CD62L+ CD4+ cells were enriched by a second, positive MACS selection.
Polarization of naive Th cells into IFN-γ-expressing (Th1) cells
CD62L+ CD4+ cells from OVA-TCRtg mice were stimulated for 1 wk with Ag and APC (T cell-depleted SC) in the presence of IL-12 to induce polarized Th1 populations. In one experiment (Expt. 3), anti-IL-4 mAb was also added to neutralize any endogenous IL-4. After 1 wk, all viable T cells expressed CD4 and the transgenic TCR, detectable with the clonotype-specific mAb KJ1-26.1 (43) (data not shown). As has been described previously (20, 26), no IL-4-producing cells (<0.1%) were detectable by intracellular immunofluorescence (Fig. 2⇓) and ELISA (data not shown) upon restimulation with PMA and ionomycin. IL-10-positive cells were detectable at very low frequencies (<1%; data not shown), even in the presence of anti-IL-4 mAb. While secreting high amounts of IFN-γ into the culture supernatant (11, 44, 45) (data not shown), such polarized Th1 populations are still heterogeneous with respect to expression of IFN-γ (20, 24, 26) (Fig. 2⇓). At most, 75% of the cells expressed IFN-γ as detectable by intracellular staining (Fig. 2⇓). Twenty to thirty percent of the Th cells did not produce IFN-γ. We have shown previously that only those cells staining for IFN-γ intracellularly also secrete IFN-γ, by correlation of intracellular staining with suIFN-γ expression, isolation of suIFN-γ+ cells, and analysis of secretion (28).
Polarization of naive Th cells into Th1- or Th2-like cells. Naive CD62L+ CD4+ OVA-TCRtg cells were stimulated in vitro for 7 days with OVA323–339 and T cell-depleted SC in the presence of IL-12 or IL-4. The cells were then restimulated for 4 h with PMA and ionomycin, fixed, permeabilized, and stained intracellularly for IFN-γ and IL-4.
This heterogeneity of the polarized populations poses a fundamental problem for the analysis of cytokine commitment, since within expanding populations, the initial phenotype of committed or converted cells is not clear. Therefore, we isolated live IFN-γ-expressing cells from a polarized Th1 population and analyzed the stability of cytokine expression (Fig. 3⇓).
Isolation and conversion of suIFN-γ-positive cells. Naive CD62L+ CD4+ cells from OVA-TCRtg mice were stimulated with OVA323–339 and T cell-depleted SC in the presence of IL-12 (with or without anti-IL-4 mAb) for 1 wk. Then cells were restimulated for 4 to 5 h with PMA and ionomycin and stained for suIFN-γ using magnetofluorescent liposomes (27, 28). By MACS, suIFN-γ+ cells were enriched from 47% before MACS to 99% brightly stained cells after MACS. The IFN-γ+ cells were then cultured for 6 days with IL-2 in the presence of IL-4. Then the cells were restimulated with the antigenic peptide OVA323–339 and T cell-depleted SC as APC in the presence of IL-4 for 4 days. Finally, cells were restimulated with PMA and ionomycin. After 4 h (with brefeldin for the last 2 h), cells were fixed, permeabilized, and stained for intracellular IL-4 and IFN-γ or for IL-4 and IL-10.
Isolation and restimulation of live IFN-γ-expressing cells
CD62L+ CD4+ T cells from OVA-TCRtg mice were stimulated with the antigenic peptide OVA323–339 and APC in the presence of IL-12 (and anti-IL-4 mAb in Expt. 3) for 1 wk, as described above. The cells were then restimulated for 4 to 5 h with PMA and ionomycin, and stained for suIFN-γ with digoxigenized AN18.17.24 mAb and anti-DIG magnetofluorescent liposomes (27, 28). By high gradient magnetic cell sorting with MACS, suIFN-γ+ cells were enriched from 37 to 60% to 99% (Fig. 3⇑ and Table I⇓). The separation was controlled for enrichment of IFN-γ-expressing cells by staining of intracellular IFN-γ (inIFN-γ). Among the 99% pure suIFN-γ+ cells, >96% also stained for inIFN-γ when fixed and analyzed directly after the sort (Table I⇓). Isolated suIFN-γ+ cells, when stimulated for 2 h more with PMA and ionomycin in the presence of the secretion inhibitor brefeldin A, stained to 99.5% for inIFN-γ (Table I⇓).
Isolation of viable IFN-γ-expressing cellsa
After MACS, unseparated Th cells and purified IFN-γ-expressing Th cells were cultured for 5 to 6 days with IL-2 in the presence or the absence of IL-4, IL-12, or IL-12 plus anti-IL-4. Unseparated and isolated IFN-γ+ Th cells were then restimulated again with the antigenic peptide OVA323–339 and T cell-depleted SC as APC in the presence or the absence of IL-4, IL-12, or IL-12 plus anti-IL-4, i.e., under neutral (no external cytokine), converting (IL-4), or preserving (IL-12 with or without anti-IL-4) conditions for 4 days (Expt. 1 and 3) or 7 days (Expt. 2).
The expression of IL-4, IL-10, and IFN-γ in OVA-TCRtg Th cells was analyzed by intracellular immunofluorescence after restimulation with PMA and ionomycin for 4 h, with brefeldin A added for the last 2 h before fixation (Figs. 3⇑ and 4⇓A, and Table II⇓). After 24 h of PMA and ionomycin stimulation, the concentrations of IL-2, IL-4, IL-5, IL-10, and IFN-γ in the culture supernatant were analyzed by ELISA (Fig. 4⇓B).
Expression of IL-2, IL-4, IL-5, IL-10, and IFN-γ in restimulated IFN-γ+ Th cells. Naive CD62L+ CD4+ OVA-TCRtg T cells were stimulated in vitro for 6 to 7 days with OVA323–339 and T cell-depleted SC (T−SC) in the presence of IL-12. Then cells were restimulated for 4 to 5 h with PMA and ionomycin, stained for cell surface-associated IFN-γ, and sorted by MACS. After MACS, unseparated and IFN-γ+ Th cells were cultured for 5 to 6 days with IL-2 in the absence (OVA) or the presence of IL-4 (OVA and IL-4). Then unseparated and IFN-γ+ Th cells were restimulated with the antigenic peptide OVA323–339 and T cell-depleted SC as APC in the absence (OVA) or the presence of IL-4 (OVA and IL-4) for 7 days. Finally, cells were restimulated with PMA and ionomycin. A, After 4 h (with brefeldin for the last 2 h), cells were fixed, permeabilized, and stained for intracellular IL-4, IL-10, and IFN-γ vs OVA-TCR. The frequency of cytokine-producing cells was determined by flow cytometry (Fig. 2⇑). B, After 24 h, the concentrations of IL-4, IL-5, IL-10, IL-2, and IFN-γ in the culture supernatant were determined by ELISA (< indicates that the value was below the indicated detection limit).
Expression of IL-4, IL-10, and IFN-γ in restimulated IFN-γ+ Th1 cellsa
Frequencies of cytokine-expressing cells, after 4 h of restimulation with PMA and ionomycin, were generally higher in Expt. 2, i.e., after 7 days of antigenic stimulation, than in Expt. 1 and 3 (Fig. 4⇑, Table II⇑), i.e., after 4 days of stimulation with Ag, indicating that some cells might have been refractory to restimulation early after activation (46).
In addition to cytokine expression, expansion and proliferation of sorted and unsorted OVA-TCR+ Th cells, as detected with the clonotype-specific mAb KJ1-26.1 (43), were analyzed (data not shown). On the average, the number of OVA-TCR+ Th cells increased about threefold after restimulation with Ag. In Expt. 1 and 3 no significant difference between cultures with or without exogenous IL-4 added was found, while in Expt. 2 OVA-TCR+ Th cells expanded better in the presence of IL-4 than in its absence (sixfold compared with threefold). The average number of cell divisions corresponded directly to the expansion of OVA-TCR+ Th cells when proliferation was analyzed using PKH26 membrane labeling (Sigma; data not shown). No indication of massive cell death was found.
Stability of IFN-γ expression upon neutral or preserving restimulation
When unseparated cells and isolated IFN-γ-expressing cells were restimulated under neutral (medium alone) or preserving (IL-12 with or without anti-IL-4) conditions, most of them (60–98%) continued to produce IFN-γ (Fig. 4⇑ and Table II⇑). In all experiments, the frequency of IFN-γ+ OVA-TCRtg cells was higher among the isolated IFN-γ-expressing cells than among the unseparated cells.
In the first experiment, no IL-4-expressing cells were detectable in neutrally or preservingly restimulated, unseparated or isolated IFN-γ-producing cells (Table II⇑). In Expt. 2, we could detect a low number of IL-4-expressing cells (0.6%) in the unseparated population when restimulated with IL-12 (Table II⇑). To neutralize endogenous IL-4, which might have been produced by T cell-depleted SC used as APC, in the third experiment we added anti-IL-4 mAb during the initial Th1 stimulation as well as to the preserving restimulation. No IL-4-expressing cells were detectable then (Table II⇑).
In contrast to IL-4, low numbers of IL-10-expressing cells (1.5–4%) were consistently detectable in preserving restimulations (Table II⇑). In the primary Th1 population after 1 wk of antigenic stimulation with IL-12, in the presence of anti-IL-4 mAb, IL-10-positive cells were detectable, although they were few in number (<1%; data not shown).
Conversion of individual Th1 cells into IL-4- or IL-10-expressing cells
When unseparated IL-12-polarized Th cells and purified IFN-γ-producing Th cells from such cultures were restimulated under converting conditions, i.e., in the presence of exogenous IL-4, IL-4-expressing cells were readily detectable in both cultures (up to 25 and 12%, respectively; Figs. 3⇑ and 4⇑, and Table II⇑), demonstrating that individual IFN-γ+ Th1 cells had been converted to IL-4 production. In the second experiment, the frequency of IL-4-positive cells induced in the unseparated cells was twice as high as that in the isolated IFN-γ-expressing cells (Table II⇑). In the other two experiments, the frequencies of IL-4-positive cells induced in the unseparated and in the isolated IFN-γ+ Th1 cells were similar (Table II⇑).
Fifteen to fifty percent of the 6- to 7-day IL-12-stimulated IFN-γ-producing Th cells were induced to express IL-10 after restimulation with IL-4 (Figs. 3⇑ and 4⇑, and Table II⇑). Again, in two experiments the frequencies of IL-10-positive cells generated in the unseparated and sorted IFN-γ+ populations were very similar, while in the third experiment 40% more IL-10 producers were generated in the unsorted population (Table II⇑).
In the presence, but not in the absence, of exogenous IL-4, we could detect IL-5 in the culture supernatant of both IFN-γ+ and unsorted cells (Fig. 4⇑B). Due to limitations in cell numbers, the production of IL-5 was analyzed only by ELISA.
The presence of exogenous IL-4 further resulted in a strong reduction of the expression of IFN-γ (Figs. 3⇑ and 4⇑, and Table II⇑). Only 5 to 20% of the unsorted cells and 6 to 23% of the sorted cells still produced IFN-γ after converting restimulation, compared with 60 to 98% under preserving conditions.
By intracellular costaining of IL-4, IL-10, and IFN-γ in isolated IFN-γ+ cells after converting restimulation, we analyzed whether suppression of IFN-γ production and induction of IL-4 and IL-10 expression were independent effects of IL-4 on individual Th1 cells. The analysis shows that IL-4- and IL-10-positive cells were induced with similar frequencies among cells still expressing IFN-γ, namely 13% of all IFN-γ-producing cells coexpressed IL-4, and 42% coexpressed IL-10; those cells that stopped to express IFN-γ, namely 12% of those produced IL-4, and 36% of those produced IL-10 (Table III⇓). Thus, induction of IL-4 and IL-10 production is apparently independent of suppression of IFN-γ expression in individual Th1 cells.
Coexpression of IL-4, IL-10, and IFN-γ in isolated IFN-γ+ cells after restimulation in the presence of IL-4a
Differential induction of IL-4 and IL-10 in individual IFN-γ+ cells
Differential inducibility of IL-4 and IL-10 expression was analyzed further in isolated IFN-γ+ cells polarized for various times by repeated restimulations in the presence of IL-12. To this end, IFN-γ+ cells were isolated by MACS after 1 wk of primary activation with Ag and polarization with IL-12 and anti-IL-4. After purification, IFN-γ+ Th cells were cultured for 7 more days with IL-2 in the presence of IL-12 and anti-IL-4. IFN-γ+ Th cells were then restimulated three times with the antigenic peptide OVA323–339 and T cell-depleted SC as APC in the presence of IL-12 plus anti-IL-4 or in the presence of IL-4 for 4 to 6 days each time. The expressions of IL-4, IL-10, and IFN-γ in OVA-TCRtg Th cells were analyzed by intracellular immunofluorescence after recall stimulation with PMA and ionomycin for 4 h, and brefeldin A was added for the last 2 h (Fig. 5⇓A). After 48 h of PMA and ionomycin stimulation, the concentrations of IL-2, IL-4, IL-5, IL-10, and IFN-γ in the culture supernatant were analyzed by ELISA (Fig. 5⇓B).
Inducibility of IL-4 is lost earlier than inducibility of IL-10 in IFN-γ+ Th1 cells. Naive CD62L+ CD4+ OVA-TCRtg T cells were stimulated in vitro for 7 days with OVA323–339 and T cell-depleted SC in the presence of IL-12 and anti-IL-4 mAb. Then cells were restimulated for 4 to 5 h with PMA and ionomycin, stained for cell surface-associated IFN-γ, and sorted by MACS. After MACS, IFN-γ+ Th cells were cultured for 6 days with IL-2 in the presence of IL-12 and anti-IL-4 mAb. Then IFN-γ+ Th cells were restimulated three times with the antigenic peptide OVA323–339 and T cell-depleted SC (T−SC) as APC in the presence of IL-12 and anti-IL-4 mAb or in the presence of IL-4 for 4 to 6 days, each time changing a cell sample from preserving (IL-12 and anti-IL-4) to converting (IL-4) restimulations after about 2, 3, or 4 wk (2w., 3w., or 4w. IL-12), or without changing the IL-12 treatment (5w. IL-12). Finally, cells were restimulated with PMA and ionomycin. A, After 4 h (with brefeldin for the last 2 h), cells were fixed, permeabilized, and stained for intracellular IL-4, IL-10, and IFN-γ vs OVA-TCR. The frequency of cytokine-producing cells was determined by flow cytometry (Fig. 2⇑). B, After 48 h, the concentrations of IL-4, IL-5, IL-10, IL-2, and IFN-γ in the culture supernatant were determined by ELISA (< indicates that the value was below the indicated detection limit).
Changing from preserving (IL-12 and anti-IL-4) to converting (IL-4) restimulations at different time points of Th1 cell differentiation, i.e., after 2, 3, or 4 wk, showed that the inducibility of IL-4 and IL-10 expression as well as the ability to suppress IFN-γ production by IL-4 were largely lost upon repeated conserving stimulations, although with different kinetics (Fig. 5⇑). Inducibility of IL-4 expression by exogenous IL-4 was nearly completely lost after 3 wk of IL-12 polarization, while inducibility of IL-10, although reduced to about 40%, as well as inducibility of IL-5 were still clearly detectable then. At this time point, suppression of IFN-γ and IL-2 expression in Th1 cells was still comparable to IFN-γ and IL-2 suppression after 2 wk of IL-12 polarization. However, 1 wk later, after about 4 wk of IL-12, only expression of IL-5, but not that of IL-4 or IL-10, was induced by IL-4. At this point, production of IFN-γ and IL-2 could only be partially suppressed in IFN-γ+ (Th1) cells upon restimulation with IL-4.
Discussion
Several studies have shown that within polarized Th1 populations, within 1 wk after primary activation, some cells can be induced to express IL-4 by restimulation in the presence of exogenous IL-4 (18, 19, 20). However, it has not been clear whether these IL-4-producing cells had been derived from IFN-γ-expressing polarized Th1 cells or from Th cells not responding to IL-12 with expression of IFN-γ. Such cells are present in polarized Th1 populations at frequencies of 20 to 30%. Here we directly address the fundamental question of plasticity of cytokine expression in Th1 cells by isolation and functional analysis of live IFN-γ-producing cells from polarized Th1 populations. IFN-γ-expressing Th cells can be isolated according to low abundant suIFN-γ, detectable with magnetofluorescent liposomes. Expression of suIFN-γ is precisely linked to secretion of IFN-γ (28). By sorting of cells according to the expression of suIFN-γ, we could also demonstrate that inIFN-γ immunofluorescence detects all cells secreting IFN-γ, as detectable by ELISA (28). Apart from establishing the validity of intracellular immunofluorescence for detection of cytokine-expressing cells, this result confirms the apparent heterogeneity of polarized Th1 populations with respect to expression of IFN-γ.
To assess stability and plasticity of cytokine expression in IFN-γ+ Th1 cells, purified IFN-γ+ cells were restimulated under neutral, i.e., first IL-2-, then Ag-, preserving (i.e., same plus IL-12) or converting (i.e., same plus IL-4) conditions. While under neutral and preserving conditions, most IFN-γ+ cells retain their phenotype, restimulation with IL-4 converts a considerable fraction of cells to expression of IL-4 (5–12% of IFN-γ+ and 5–25% of unseparated Th cells), IL-5, and/or IL-10 (15–50% of IFN-γ+ and 19–56% of unseparated Th cells). It can be excluded that the converted cells are derived from contaminating IFN-γ− cells contained in the population of sorted IFN-γ+ cells. First, IFN-γ+ cells had been enriched to 99.5%. If the remaining 0.5% IFN-γ− cells had selectively survived and/or expanded within the 1 to 2 wk of culture and restimulation after sorting, either 90% of the IFN-γ+ cells would have had to die or the IFN-γ− cells would have had to divide three or four times more frequently than IFN-γ+ cells to increase their frequency from 0.5 to 5%. However, we did not observe massive death of IFN-γ+ cells, and proliferation rates were similar in both sorted (99.5% pure IFN-γ+, 0.5% IFN-γ− cells) and unsorted (50–70% pure IFN-γ+, 30–50% IFN-γ− cells) populations. The frequencies of IL-4- and IL-10-expressing cells as well as the amount of secreted IL-5 induced by IL-4 were at most twofold different in the separated and unseparated cell populations. Our results show that IL-4 induces the expression of IL-4, IL-5, or IL-10 in individual IFN-γ+ Th1 cells activated for 1 wk in the presence of IL-12 with about the same efficiency as in IFN-γ− cells of that same population.
While IL-4 induces up to 12% of IL-4-expressing cells and 50% of IL-10-expressing cells in isolated IFN-γ+ Th1 cells activated for 1 wk with Ag plus IL-12, stimulation of naive Th cells for 1 wk with Ag plus IL-4 results in up to 35% IL-4-producing cells and 70% IL-10-producing cells (Fig. 2⇑; data not shown). The difference of about 20% more IL-4- and/or IL-10-expressing naive vs converted IFN-γ+ Th1 cells corresponds roughly to the 20% of Th1 cells in the converted population that still produce IFN-γ, suggesting that after exposure to IL-12 for 1 wk, about one-third of IFN-γ+ cells is stably committed to expression of IFN-γ, while two-thirds have retained the plasticity of naive Th cells. The molecular basis for this dichotomy is not clear, nor is that for the loss of plasticity.
Nevertheless, it is surprising that, after 1 wk of priming with IL-12, most individual IFN-γ+ cells are not stably committed to expression of IFN-γ, unlike at later time points, and stop to express IFN-γ upon converting restimulation with IL-4. Costaining of these cells for IL-4, IL-10, and IFN-γ reveals that expressions of IL-4 and IL-10 are induced with similar patterns and efficiencies in early Th1 cells, which either still do or no longer express IFN-γ, after converting restimulation with IL-4. This suggests that suppression of IFN-γ production and induction of IL-4 and IL-10 expression are independent effects of IL-4 on individual Th1-like cells. Our results show that the shift from stimulation with IL-12 to stimulation with IL-4 after 1 wk of antigenic stimulation has a drastic effect on IFN-γ+ cells, most of which at that stage still respond to IL-4 by stopping to express IFN-γ, upon antigenic challenge. The induction of IL-10 in quite a few cells should indirectly further support this suppression of IFN-γ expression (47, 48). The shift from Th1- to Th2-inducing conditions thus seems a reasonable strategy to at least suppress if not convert immune reactions in the early phase of immune responses. The in vitro experiments described here suggest that IFN-γ expression can be switched off efficiently by IL-4 within the first 3 wk of primary activation in the presence of IL-12.
In IL-12-mediated differentiation of IFN-γ+ Th1 cells, inducibility of IL-4 expression by IL-4 is lost about 1 wk earlier (after about 3 wk with IL-12) than inducibility of IL-10 expression and suppression of IFN-γ, while IL-5 could still be induced by IL-4 even after about 4 wk of IL-12 polarization. For both effects of IL-4, suppression of IFN-γ and induction of IL-4 expression, STAT-6 activation seems to be required (49, 50, 51), which is detectable in early Th1 cultures at the population level (52) and in all cells analyzed here by fluorescence microscopy and flow cytometry (T. Stamm, M. Löhning, and A. Radbruch, unpublished observation). Whether in individual cells the production of IFN-γ is suppressed and/or the expression of IL-4, IL-5, or IL-10 is induced, has to be regulated by additional transcription factors. For induction of IL-4 expression, specific transcription factors, such as c-maf (53) and GATA-3 (54), were shown to be required. Interestingly, inhibition of GATA-3 seems to have a strong effect on IL-4, a less strong effect on IL-10, and only a weak effect on IL-5 expression, reflecting different thresholds (or alternative pathways) for induction of IL-4 vs IL-10 vs IL-5 (54), which could explain the different kinetics of inducibility of IL-4, IL-10, and IL-5 expression observed here in the course of Th1 cell differentiation.
Acknowledgments
We thank Dr. D. Y. Loh for providing the OVA-TCR transgenic DO11.10 mice. We are grateful to Drs. A. O’Garra, E. Schmitt, M. Gately, and W. Müller for generous gifts of mAb and reagents. We thank S. Irlenbusch for expert technical assistance, and R. Christine, R. Manz, A. Thiel, and K. Petry for helpful discussion and comments on the manuscript.
Footnotes
-
↵1 This work was supported by the Bundesministerium für Bildung und Forschung.
-
↵2 Both authors contributed equally to this work.
-
↵3 Address correspondence and reprint requests to Dr. Andreas Radbruch, Deutsches Rheuma-Forschungszentrum Berlin, Hannoversche Str. 27, D-10115 Berlin, Germany. E-mail address: radbruch{at}drfz.de
-
↵4 Abbreviations used in this paper: SC, spleen cells; CD62L, CD62 ligand; MACS, magnetic-activated cell sorting; DIG, digoxigen(ized); suIFN-γ, surface IFN-γ; inIFN-γ, intracellular IFN-γ.
- Received December 19, 1997.
- Accepted May 21, 1998.
- Copyright © 1998 by The American Association of Immunologists