Differences in regulatory pathways identify subgroups of T cell-derived Th2 cytokines

Abstract

We analysed regulatory mechanisms involved in the production of Th2 cytokines by freshly isolated human T cells. We used an in vitro culture system in which the primary signal was provided by a cross-linking anti-CD3 MoAb presented on the Fc receptors of P815 cells. Both CD80 and CD86, expressed on transfected P815 cells, were able to provide efficient costimulation for the production of IL-4, IL-5 and IL-13. IL-2 was also highly important for induction of all three Th2 cytokines. However, differences between IL-4 on the one hand and IL-5 and IL-13 on the other hand were observed when sensitivity to cyclosporin A (CsA) was studied. CsA (an inhibitor of calcineurin phosphatase activity) strongly inhibited IL-4 production, but it did either not affect or even increased IL-5 and IL-13 production. In accordance with this, CD80 and phorbol myristate acetate (PMA) (without anti-CD3 or calcium ionophore) were sufficient to induce production of IL-5 and IL-13, but not of IL-4. The subgrouping of Th2 cytokines was further confirmed at another level on the basis of differences in cell sources: IL-4 was predominantly produced by CD4⁺ T cells, while IL-5 and IL-13 were produced by both CD4⁺ and CD8⁺ T cells. Thus, differences in cell sources and in the requirement of the calcium/calcineurin-signalling pathway allowed us to identify two subgroups (IL-4 and IL-5/IL-13) among human Th2-type T cell cytokines.

INTRODUCTION

According to their cytokine secretions, murine memory and effector T helper cells have been functionally divided into distinct subsets: Th1 cells, which secrete IL-2, interferon-gamma (IFN-γ) and tumour necrosis factor-beta (TNF-β), which are all involved in cell-mediated immunity and delayed-type hypersensitivity; and Th2 cells, which produce a number of cytokines (IL-4, IL-5, IL-6, IL-10 and IL-13) involved in humoral immune responses and immediate-type hypersensitivity reactions (for review see [1–4]). These subsets differentiate from a common naive CD4⁺ T precursor helper cell [1–3]. The establishment of the Th1/Th2 balance is determined early during immune responses and depends on many factors including antigen structure, the functional status of antigen-presenting cells (APC), the strength of T cell activation, the presence of cytokines such as IL-12 and IL-4, costimulatory signals such as CD80 or CD86, and the microenvironment (for review see [1–3]). In humans, the distinction between these T helper subsets is less clear, but clinical conditions with a predominance of either Th1 or Th2 cytokine secretion have been identified [4], and T cell clones with cytokine secretion patterns corresponding to murine Th1 and Th2 cells have been generated [5]. Most studies of Th2 cytokine production have focused either on differentiation signals, or on production by differentiated T cell clones, and little is known about the regulation of production of these cytokines by primary human T cells. In this respect, some data point to the importance of costimulation by CD80 or CD86 [6].

The interaction between CD80 and CD86 on the APC with CD28 on T cells is considered a highly important costimulatory interaction for T cell activation in general (for review see [7,8]). Triggering of CD28 induces separate (but incompletely identified) intracellular signalling pathways, which are calcium-independent and which render T cell activation resistant to inhibition by cyclosporin A (CsA) [9–11]. The fact that IL-4, IL-5 and IL-13 are supposedly secreted by the same T cell subsets (Th2 cells) and are overproduced in similar clinical conditions (e.g. atopic allergy) suggests that their production is regulated by common mechanisms. This however, has not been proven, and there are indications that this might in fact not be the case [4]. In the present study we investigated the production of Th2 cytokines (IL-4, IL-5 and IL-13) by freshly isolated human T cells. We used an in vitro system of polyclonal T cell activation in which anti-CD3 MoAb was cross-linked on P815 mouse mastocytoma cells to provide the primary signal [10]. Alternatively, we used phorbol myristate acetate (PMA) and ionomycin as the primary signal. To provide costimulation, P815 cells transfected with CD80 or CD86 (the ligands for CD28) were used. We investigated IL-2 dependency and sensitivity to CsA, as well as the T cell subpopulations responsible for their production.

MATERIALS AND METHODS

Monoclonal antibodies and other reagents

Anti-CD3 MoAb UCHT1 and anti-CD45RO MoAb UCHL-1 were gifts from Dr P. Beverley (E. Jenner Institute for Vaccine Research, Compton, UK). Humanized anti-Tac MoAb (CD25, directed at the p55 chain of the human lymphocyte IL-2R) and humanized Mikβ1 (anti-CD122, directed at the p75 chain of the IL-2R) were gifts from Dr J. Hakimi (Hoffman-La Roche, Nutley, NJ). MoAb B-G5 (IgG1), a murine neutralizing anti-IL-2 MoAb, was purchased from Innotherapie (Besançon, France). Anti-IL-4R (anti-CDw124) MoAb was purchased from R&D Systems (Minneapolis, MN). PMA and ionomycin were purchased from Sigma Chemical Co. (St Louis, MO). CsA was from Sandoz Pharmaceuticals (Basel, Switzerland).

Cell lines

The P815 cell line (obtained from ATCC, Rockville, MD) is an NK-resistant DBA/2-derived murine mastocytoma cell line that expresses mouse Fcγ RII and Fcγ RIII [12]. The P815 cell lines, transfected with CD80 or CD86, were gifts from L. Lanier (DNAX Research Institute of Molecular and Cellular Biology, Palo Alto, CA).

Isolation of T cells

All subjects donating blood for these experiments were healthy volunteers of both sexes, aged 20–50 years. Peripheral blood mononuclear cells (PBMC) were isolated by centrifugation on Ficoll–Hypaque (density 1·077) gradients. Monocytes were removed by cold agglutination. The lymphocytes were further purified using complement-fixing anti-NK and anti-monocyte MoAbs and lympho-KWIK-T (One Lambda Inc., Los Angeles, CA) as previously reported [13]. The resulting cell preparations contained >98% CD3⁺ T cells.

CD4⁺ and CD8⁺ T lymphocytes were purified by magnetic immunoselection. The positive selection of CD4⁺ and CD8⁺ lymphocytes was performed using Dynabeads (Dynal, Oslo, Norway) coated with anti-CD4- and anti-CD8 MoAbs, respectively. In order to detach the beads, cells were incubated for 45 min at room temperature in complete medium with Detach-a-bead (Dynal). Detached beads were removed by using the magnetic particle concentrator. The CD4⁺ and CD8⁺ subpopulations were >97% pure on FACS analysis.

Naive CD45RO⁻ cells and memory CD45RA⁻ cells were purified by negative selection using MoAb UCHL-1 and MoAb 2H4 (Coulter Corp., Hialeah, FL), respectively. Goat anti-mouse IgG-coated magnetic beads (Dynal) were used at a ratio of 30 beads per target cell to remove the reciprocal subpopulation. The purity of the CD45RO⁻ and CD45RA⁻ subpopulations were 98% and 99%, respectively.

Cell culture for T cell activation

One million T cells in a final volume of 1 ml complete medium RPMI 1640 (Boehringer Ingelheim, Biowhittaker, Heidelberg, Germany) supplemented with 2 mm l-glutamine, penicillin 100 U/ml, streptomycin (100 μ g/ml) and 10% iron-supplemented BCS (Hyclone, Logan, UT) were cultured in flat-bottomed 24-well plates (Nunc, Roskilde, Denmark). P815, P815/CD80, P815/CD86 or P815/CD58 cells were pretreated with mitomycin C (50 μ g/ml), and used as accessory cells at a P815/T cell ratio of 1. Soluble anti-CD3 MoAb UCHT1 (2 μ g/ml) or PMA (1 ng/ml) and ionomycin (0·5 μ g/ml) were added as primary signals. All cultures were performed at 37°C in a 5% CO₂ atmosphere. After 1–4 days of culture, supernatants were taken and frozen until cytokine measurements were performed. The concentration of IL-4 in the T cell supernatants is influenced by autoconsumption [14]. For accurate detection, IL-4 was therefore measured either at day 1 of the culture (which is before significant consumption occurs), or at days 3–4 if anti-IL-4R MoAb had been added to block consumption [14].

Cytokine production

Cytokine concentrations in culture supernatants were determined with a sandwich ELISA technique, using combinations of unlabelled and biotin-coupled MoAbs to different epitopes of each cytokine. MoAb pairs for IL-2, IL-5 and IL-13 were from PharMingen (San Diego, CA), and those for IL-4 were from BioSource Europe (Nivelle, Belgium). The MoAb clones were: 5344.11 and B33-2 for IL-2, 860 A4B3 and 860 F10H12 for IL-4, TRFK5 and JES1-SA10 for IL-5, and JES10-5A2 and B69-2 for IL-13.

RESULTS

Th2 cytokine production: importance of costimulatory signals and cell sources

In a first series of experiments, the capacity of CD80 and CD86 to provide costimulation for Th2 cytokine production was analysed. To this end, freshly isolated T cells were stimulated with anti-CD3 MoAb, cross-linked on the FcR of P815 cells. For costimulation, CD80- or CD86-transfected P815 cells were used. Figure 1 shows that T cells stimulated with anti-CD3 alone produced low amounts of cytokines. When costimulation was provided by CD80 or CD86, all three Th2 cytokines tested (IL-4, IL-5 and IL-13) were produced. Obviously, both CD80 and CD86 can provide costimulation for the production of Th2 cytokines, confirming previous data [6,11]. Because CD86 expression was lower than the CD80 expression on transfected P815 cells, differences in the amount of Th2 cytokines induced by both costimulatory signals can not be interpreted.

Fig. 1.

Effects of costimulation on Th2 cytokine production by human T cells. One million T cells were cultured with 10⁶ mitomycin C-treated wild-type P815 cells or P815 cells transfected with CD80 or CD86 in a final volume of 1 ml of complete medium in flat-bottomed 24-well culture plates. Anti-CD3 MoAb UCHT1 (2 μ g/ml) was used as the primary stimulus. After 1 day (IL-4) or 4 days (IL-5, IL-13), supernatants were collected and cytokines were measured by ELISA. The results are from three independent experiments.

The stimulatory conditions with P815/CD80 cells were then selected for identifying the cell sources of Th2 cytokines. We first compared the relative production of the different Th2 cytokines by naive versus memory T cells. Negatively selected CD45RO⁻ naive T cells and CD45RA⁻ memory/effector T cells were stimulated in parallel and in identical conditions with anti-CD3 cross-linked on P815/CD80 cells. As shown in Fig. 2, the memory T cell fraction is the main cell source for Th2 cytokines, while the naive fraction accounted for <15% of the production (Mann–Whitney test for IL-4, P = 0·0025; IL-5, P = 0·002; IL-13, P = 0·0032). All further results have to be interpreted as being relevant for memory and/or effector T cells.

Fig. 2.

Th2 cytokines are mainly produced by memory/effector T cells. One million CD45RO⁻ or CD45RA⁻ T cells were incubated with 10⁶ P815/CD80 cells in a final volume of 1 ml complete medium in flat-bottomed 24-well plates. UCHT1 (2 μ g/ml) was used as the primary stimulus. To block IL-4 consumption, anti-IL-4R MoAb was added at a concentration of 2·5 μ g/ml. Supernatants were collected after 3 days (IL-4) and 4 days (IL-5, IL-13) of culture. The concentrations of IL-4, IL-5 and IL-13 were determined by ELISA. The figure shows the results of five experiments.

In a previous study, we demonstrated that anti-CD3 cross-linked on P815 cells transfected with CD80 was effective in activating highly purified CD8⁺ T cells in the absence of helper cells [15]. As shown in Fig. 3, we found that IL-4 was produced by the CD4⁺ subpopulation. In contrast, IL-5 was produced more abundantly by CD8⁺ T cells compared with CD4⁺ T cells (Wilcoxon test for IL-5: P = 0·0195). CD4⁺ and CD8⁺ cells produced similar amounts of IL-13 (P = 0·0742).

Fig. 3.

Th2 cytokine production by CD4⁺ and CD8⁺ T cells. One million CD4⁺ or CD8⁺ T cells isolated from blood of five (a) or 10 (b,c) different donors were cultured with 10⁶ mitomycin C-treated P815/CD80 cells in a final volume of 1 ml complete medium in flat-bottomed 24-well plates. Anti-CD3 MoAb UCHT1 (2 μ g/ml) was used as the primary stimulus. For the cultures of (a), anti-IL-4R MoAb (2·5 μ g/ml) was added to block IL-4 consumption. Supernatants were collected after 4 days of culture. The concentrations of IL-4, IL-5 and IL-13 were determined by ELISA.

CsA differentially affects IL-4 and IL-5/IL-13

CsA binds to cyclophilin, and the CsA–cyclophilin complex acts by binding to calcineurin and by blocking its phosphatase activity. CsA therefore can be used as a tool to study the importance of calcineurin for the activation of cytokine genes. In previous studies, we and others have demonstrated that costimulation through CD28 results in CsA-resistant T cell activation, at least at the level of T cell proliferation, generation of cytotoxic T cell activity and IL-2 production [9–11]. We now investigated the effect of CsA on IL-4, IL-5 and IL-13 production by T cells in the experimental conditions described above. Figure 4 shows that the effect of CsA on CD28-costimulated T cells was divergent, depending on the cytokine studied. The production of IL-4 was strongly inhibited (residual levels on average 10% of the controls, Wilcoxon test for IL-4: P = 0·0156). In contrast, the production of IL-5 and IL-13 was enhanced (on average 232% and 169% of the controls, respectively) by CsA (Wilcoxon test for IL-5, P = 0·0055; IL-13, P = 0·0026). In accordance with earlier findings [9–11], IL-2 production in this series of experiments was only slightly inhibited by CsA (residual levels on average 79% of the controls; not shown). Addition of CsA resulted in similar changes in Th2 cytokine production in cultures of highly purified CD45RA⁻ T cells as in total T cells (not shown). Thus, these data suggest that calcineurin signalling is required for IL-4 production, while the production of IL-5 and IL-13 is apparently calcineurin-independent, at least when the T cells are costimulated through CD28.

Fig. 4.

Cyclosporin A (CsA) differentially affects the production of Th2 cytokines. One million T cells were cultured with 10⁶ P815/CD80 cells in 1 ml medium. UCHT1 MoAb was added at a concentration of 2 μ g/ml. CsA was added to parallel cultures at a final concentration of 400 ng/ml. After 1 (IL-4) or 4 (IL-5, IL-13) days of culture, supernatants were collected and IL-4, IL-5 and IL-13 concentrations were determined by ELISA. The figure shows the results of seven (a) and 16 (b,c) independent experiments.

Calcium fluxes and Th2 cytokine production

We then tried to confirm differences in the role of calcium signalling pathways for the production of Th2 cytokines using a different approach. We compared the effect of PMA, ionomycin and PMA + ionomycin (indicated as PMA/iono) as primary stimulating agents for T cells, with or without a CD80 costimulatory signal. In Fig. 5 the results are presented for IL-2, IL-4, IL-5 and IL-13 production.

Fig. 5.

Effects of calcium rise on Th2 cytokine production. One million T lymphocytes were incubated with 10⁶ P815 or P815/CD80 cells in a final volume of 1 ml complete medium in flat-bottomed 24-well plates. Phorbol myristate acetate (PMA; 1 ng/ml) and/or ionomycin (0·5 μ g/ml) were added as primary signals. Supernatants were taken at day 1 (IL-2, IL-4) or at day 4 (IL-5, IL-13). Cytokine levels were determined by ELISA. The figure shows the results of independent experiments on T cells from eight to 19 donors.

For all cytokines, the combination of PMA/iono +CD80 was the most efficient one. However, most informative is the effect of CD80 combined with PMA. This combination induced production of IL-2, IL-5 and IL-13. IL-4 was not induced (except for low IL-4 production in one experiment). Neither PMA nor CD80 induced calcium release. The conclusion based on these results is that IL-5 and IL-13 genes can be induced in a calcium-independent manner, which is in accordance with the data in the previous paragraph on CsA resistance of IL-5 and IL-13 production. T cells stimulated with CD80 alone or T cells without any stimulation produced no Th2 cytokines (not shown).

Contribution of endogenous IL-4 production to IL-5 and IL-13 production

It is well established that IL-4 has an essential role in the differentiation of precursor T cells towards Th2 cells. It is less clear whether IL-4 has an effect on Th2 cytokine secretion by freshly isolated memory/effector T cells. To study this issue, T cells were stimulated with P815/CD80 cells and anti-CD3 in the absence or presence of an anti-IL-4R MoAb which completely blocks the IL-4R [14]. Figure 6 shows that when CD45RA⁻ T cells were stimulated by anti-CD3 and CD80, IL-13 production was enhanced by IL-4, as indicated by the significant decrease in IL-13 production after IL-4 receptors had been blocked (on average, 37% of control; Wilcoxon test: P = 0·0078). However, blocking IL-4 receptors did not influence IL-4 and IL-5 production (on average 96% and 98% of controls, respectively; Wilcoxon test for IL-4, P = 0·9375; IL-5, P = 0·4375). Th2 cytokine production by CD28-costimulated memory/effector T cells is therefore largely IL-4-independent.

Fig. 6.

Effects of blocking IL-4 bioactivity on Th2 cytokine production by memory/effector T cells. CD45RA⁻ memory T cells (5 × 10⁵/ml) were cultured with P815/CD80 cells (5 × 10⁵/ml) and UCHT1 MoAb at a concentration of 2 μ g/ml, in the presence or absence of a blocking anti-IL-4R MoAb (2·5 μ g/ml). Supernatants were collected after 72 h and the IL-4, IL-5 and IL-13 concentrations in the supernatants were determined by ELISA. The results of six to eight independent experiments with different donors are shown.

Contribution of endogenous IL-2 production to Th2 cytokine production

We analysed whether or not the different Th2 cytokines were dependent on autocrine or paracrine IL-2 production. In order to block completely all IL-2 activity, we used a panel of two anti-IL-2R MoAbs, directed at the α- (CD25) and β- (CD122) chains of the IL-2 receptor in combination with a neutralizing anti-IL-2 MoAb [11]. This mixture is further indicated as anti-IL-2(R). When T cells were stimulated with anti-CD3 only, Th2 cytokine production was very low and entirely dependent on IL-2 signalling (not shown). Anti-IL-2(R) MoAb reduced IL-4, IL-5 and IL-13 production (residual production on average 23%, 47%, 39% of the controls, respectively) when T cells were costimulated via CD80 (Fig. 7a). Although we can not exclude the possibility of incomplete neutralization of IL-2 activity by anti-IL-2(R), a direct IL-2-independent signalling effect of CD28 costimulation on Th2 cytokine genes was strongly suggested by these data.

Fig. 7.

Effects of blocking IL-2 bioactivity on the production of Th2 cytokines after CD80 costimulation. One million T cells were incubated with 10⁶ P815/CD80 cells. UCHT1 MoAb at 2 μ g/ml, phorbol myristate acetate (PMA; 1 ng/ml) and ionomycin (0·5 μ g/ml) were added as the primary stimulus. For inhibition of IL-2 activity, a combination of anti-Tac (anti-IL-2Rα) MoAb at 5 μ g/ml, Mikβ1 (anti-IL-2Rβ) and B-G5 (anti-IL-2) MoAbs both at 2 μ g/ml was used. To block IL-4 consumption, anti-IL-4R MoAb was added at a concentration of 2·5 μ g/ml. All cytokine concentrations were determined by ELISA after 4 days of culture.

To address this point further, the effect of IL-2 neutralization on Th2 cytokine production was also studied when T cells were stimulated with PMA and ionomycin. Importantly, PMA and ionomycin provide a potent primary signal for IL-2 production, which is not significantly enhanced by CD80 costimulation (Fig. 5). IL-4, and the small levels of IL-5 and IL-13 produced after stimulation with PMA and ionomycin alone, were all strongly reduced in the presence of anti-IL-2(R): residual levels were on average 10%, 10% and 21% of the controls, respectively (not shown). When T cells were costimulated with CD80, the production of the Th2 cytokines was less reduced by anti-IL-2(R) MoAb (Fig. 7b; residual levels in the presence of anti-IL-2(R), on average 39%, 31% and 42% of the controls, respectively). Since PMA +ionomycin with or without CD28 costimulation induced a similar amount of IL-2, while anti-IL-2(R) had a lower inhibitory effect on the production of IL-4, IL-5 and IL-13 in case of CD28 costimulation compared with PMA +ionomycin alone, these data thus point to an IL-2-independent direct effect of the CD28 signalling pathway on the IL-4, IL-5 and IL-13 genes.

DISCUSSION

In this study we have examined regulatory pathways for production of Th2 cytokines by freshly isolated human memory/effector T cells. The B7 molecules CD80 and CD86 provided efficient costimulation for the production of all Th2 cytokines, as already shown by others [6]. Most importantly, costimulation with CD80 enabled us to identify a differential requirement for calcineurin activity between two subgroups of Th2 cytokines. CD80 costimulation of T cells induces CsA resistance for IL-5 and IL-13 production, but not for IL-4 production. This difference in calcineurin dependency was also supported by the findings that PMA and CD80 were sufficient to induce IL-5 and IL-13, but not IL-4. Thus, a calcium-dependent pathway and calcineurin activity are essential for induction of IL-4 but not for IL-5 and IL-13. The concept of two subgroups was further reinforced by differences in cell sources: IL-4 is produced by CD4⁺ T cells, whereas IL-5 and IL-13 are produced by isolated CD4⁺ and even more so by CD8⁺ T cells.

It has to be emphasized that the Th2 cytokines in our cultures were produced by memory/effector T cells, and that we did not address the problem of differentiation of T precursor cells towards Th2 cells. We obtained similar results on Th2 cytokine production and calcineurin dependency in cultures of total T cells and in cultures of CD45RA⁻ memory/effector T cells. Except for a low production of IL-13, CD45RO⁻ naive T cells were unable to produce Th2 cytokines. This is consistent with recent data that commitment of naive T cells to IL-4 production requires several cell cycles. Once committed, cell division is however not essential for Th2 cytokine production [16]. In the short-term cultures performed in this study, cell division was not a prerequisite, because cytokines were already detected after 24 h. The stimulatory (anti-CD3) and costimulatory (CD80/CD86) signals used here thus activated precommitted T cells. It is thereby interesting to note that these committed cells are still largely dependent on costimulation for optimal cytokine production.

Our data on the calcium and calcineurin dependence of IL-4 production are in agreement with data published by others. The IL-4 gene is known to be controlled by the transcription factor NFAT [17–19]. The NFAT binding site on the IL-4 promoter is adjacent to the AP-1 binding site [19]. Both NFAT and AP-1 transcription activities are induced by intracellular release of calcium and activation of protein kinase C [20]. The IL-2 and IL-4 NFAT/AP-1 composite sites synergize with GATA3 to induce IL-4 gene expression in Th2 T cells [21].

The regulatory pathways for IL-5 and IL-13 are apparently completely distinct from those of IL-4. PMA and CD28 signalling is sufficient for IL-5 and IL-13 production, and thus IL-5 and IL-13 do not require a calcium signal. This finding can also explain why the production of IL-5 and IL-13 was resistant to addition of CsA, which targets the calcineurin-signalling pathway. Recently, Umland et al. [22] demonstrated that IL-5 mRNA stability was prolonged in the presence of CsA. Both Schandenéet al. [23] and van der Pouw Kraan et al. [24] demonstrated a negative effect of calcium signalling on IL-5 and IL-13 production, respectively. We also found that IL-5 and IL-13 production can be enhanced in the presence of CsA, but have no explanation yet for this.

Besides important differences in the regulation of the three Th2 cytokines studied, there are also a number of similarities, especially with respect to their dependence on costimulatory signals (CD80 or CD86) and their partial IL-2 dependency. In two recent papers, rIL-2 was used as the single stimulatory signal for IL-5 production in PBMC [25] or T cell clones [26]. Both studies decided on the existence of a CsA- and rapamycin-insensitive [25], FK506-sensitive [26] unique IL-2 signalling pathway which stimulates IL-5 production, which is distinct from the IL-2R-mediated signalling pathways for proliferation, and which is distinct from the TCR-derived signalling pathways. Houssiau et al. [27] and Holter et al. [28] previously reported that IL-4 production by human T cells is dependent on IL-2 production. A costimulatory effect by the CD80–CD28 interaction on IL-4 production has already been suggested previously [29], although it was not excluded that the effect of CD28 costimulation was indirect, resulting from enhanced IL-2 production [29]. Importantly, we have arguments that Th2 cytokine production can indeed proceed through an IL-2-independent mechanism. The IL-2 dependence of Th2 cytokine production was studied by blocking IL-2 bioactivity in T cell cultures with MoAb to IL-2 and to the IL-2R. We have previously shown the efficacy of this combination in blocking IL-2 activity, even when there is a high production in the culture [11]. We measured IL-4 production after neutralization of its autoconsumption [14] and found that the production induced by PMA and ionomycin in combination with CD28 signalling was at least partially IL-2-independent. Indeed, neutralization of IL-2 activity had a stronger effect on Th2 cytokine production when T cells were stimulated with PMA +ionomycin alone than when stimulated with PMA +inomycin +CD80, although IL-2 production in both conditions was similar. To our knowledge, no such strong evidence has previously been provided that CD28 directly costimulates the production of IL-4 by human T cells. Similarly, both IL-5 and IL-13 were only partially dependent on autocrine or paracrine IL-2 production, again suggesting a direct signalling effect through CD28 for IL-5 and IL-13 production. It has also been demonstrated by other groups that CD28 signalling can augment T cell gene expression of IL-5 [23] and IL-13 [30].

The totally distinct regulatory pathways for IL-4 and IL-13 production are remarkable, as IL-13 is a functionally and evolutionarily close relative of IL-4 [31]. Earlier studies have already focused on the different kinetics of the production of IL-4 and IL-13 [32,33]. Moreover, by using T cell clones different requirements of calcium- and protein kinase C (PKC)-mediated signalling for IL-4 and IL-13 production were found [32]. The notion of subgroups among the Th2 cytokines has already been proposed by Palmer et al. [34]. Focusing on the role of cytokines and accessory cell-dependent costimulatory signals in differentiation of precursor T cells into Th2 cells, they found that the stimulatory requirements for production of IL-4 by these Th2 cells were distinct from the requirements for IL-5 and IL-13 production, the former being PMA +ionomycin and the latter being the accessory cells and a cytokine mixture (IL-1, IL-4 and IL-6).

The differential effects of CsA on these different subgroups of Th2 cytokine responses might be relevant when CsA is scheduled as a treatment modality for autoimmune disorders, allergic diseases and asthma. Whereas CsA markedly improves atopic dermatitis [35], which is considered an IL-4-mediated disease [4], its efficacy on atopic asthma, in which IL-4 but also IL-5 and IL-13 play an essential role [36,37], is less clear [38]. Our findings on different regulatory pathways open perspectives for differential modulation of individual Th2 cytokines.