Interleukin (IL)-6 Directs the Differentiation of IL-4–producing CD4⁺ T Cells

Total CD4⁺ T cells were isolated from spleen and lymph nodes from either wild-type B10.BR, wild-type C57BL/6J, or C57BL/6J-backcrossed IL-6–deficient mice by negative selection as previously described (16). The naive population (CD4⁺CD45RB^highCD44^low) was purified from total CD4⁺ T cells obtained from cytochrome (Cyt)¹ c TCR transgenic mice by staining with a red⁶¹³-conjugated anti-CD4 mAb, a biotinylated anti-CD44 mAb, FITC-conjugated anti-CD45RB mAb, and PE-conjugated streptavidin (PharMingen, San Diego, CA) and cell sorting by using a FACS^PLUS®. Mitomycin C-treated (50 μg/ml) syngeneic splenocytes were used as source of APCs.

Total or naive CD4⁺ T cells were cultured at 10⁶ cells/ml in the presence of syngeneic APCs (5 × 10⁵ cells/ml) with Con A (Boehringer Mannheim GmbH, Mannheim, Germany) or synthetic moth Cyt c peptide (Yale University, New Haven, CT) plus IL-6 (R&D Sys., Inc., Minneapolis, MN), IL-12 (provided by Genetics Institute, Cambridge, MA), or IL-4 (provided by DNAX, Palo Alto, CA). After 4 d, a 40–60% increase in the number of cells was observed in the different conditions. Effector Th cells were exhaustively washed, counted, and restimulated at 10⁶ cells/ml with Con A (in the absence of APCs) or Cyt c peptide plus APCs (5 × 10⁵ cells/ml). Anti-CD3 mAb (2C11) was immobilized on plastic at 5 μg/ml and anti-CD28 (PharMingen) was used in a soluble form (1 μg/ml).

Expression of IL-6Rα was analyzed by FACS^® by double staining with a FITC-conjugated anti-CD4 mAb and a biotinylated anti–IL-6Rα mAb (PharMingen) in combination with PE-conjugated streptavidin.

Total RNA was extracted as described (17) from 2 × 10⁵ cells mixed with 10⁴ Raji cells which were added at harvest as an internal control, and the amount of human MHC class II HLA-DR cDNA from Raji cells was used as an internal standard. In brief, after dilution (1/2) of reverse transcription reaction mixture, 5 μl was assayed for levels of DR cDNA by PCR using DR-specific primers in the presence of DR competitor construct (50 pg) to confirm the efficiency of RNA extraction and reverse transcriptase (RT)-PCR procedure in each group. IL-4 transcript levels in the 5 μl of diluted (1/2) reverse transcription reaction mixture was semi-quantitated using the competitor (167 fg) in the presence of specific primers as described (18). DR primers used were (5′-CGAGTTCTATACTGAATCCTG, and 3′-GTTCTGCTGCATTGCTTTTGC). Competitor amounts shown (competition cDNA, fg or pg) are corrected to represent the amount of IL-4 or DR gene and not as the total plasmid amount. A multiple cytokine-containing competitor construct was a gift from R.M. Locksley (University of California, San Francisco, CA).

ELISA assays were performed using purified anti–IL-4, anti–IL-12, and anti–IL-6 mAbs (3 μg/ml) as primary antibodies, the corresponding biotinylated anti–IL-4, anti–IL-12 and anti–IL-6 mAbs (1 μg/ml; PharMingen), and horseradish peroxidase-conjugated avidin D (2.5 μg/ml) (Vector Labs., Inc., Burlingame, CA), the TMB microwell peroxidase substrate and stop solution (Kirkegaard & Perry Labs., Inc., Gaithersburg, MD), using the recommended protocol (PharMingen). Recombinant IL-4 (DNAX) and IFN-γ (GIBCO BRL, Gaithersburg, MD) were used as standards. The specific activity of the IL-4 and the IFN-γ that were used as standards for the ELISA assays were 10⁸ U/mg for IL-4 and 10⁷ U/mg for IFN-γ.

IL-6 is produced by a wide spectrum of cells including fibroblasts, endothelial cells, neuronal cells, macrophages, mast cells, tumor cell lines, and CD4⁺ Th2 cells, but from an immunological point of view, APCs represent the major source of IL-6 (14, 19). To determine the potential role of IL-6 in differentiation of naive CD4⁺ T cells into effector Th1 and Th2 cells, we first analyzed the effect of exogenous IL-6. Total mouse CD4⁺ T cells were isolated (16) and stimulated to differentiate with Con A with or without IL-4 (Th2) or IL-12 (Th1) in the presence or absence of exogenous IL-6. After 4 d, the effector Th1 or Th2 cells were exhaustively washed, counted, and equal number of cells were restimulated with Con A for 24 h before harvesting the supernatant which was then analyzed for cytokine production. Interestingly, even in the absence of any polarizing cytokine, IL-6 directed the differentiation of the CD4⁺ cells to a Th2 phenotype, since the cells produce high amounts of IL-4, but not IFN-γ (Fig. 1,A). IL-6 did not modify the differentiation of the Th2 cells directed by IL-4. However, differentiation into Th1 cells by IL-12, was impaired in the presence of IL-6. Thus, Th1 cells differentiated with Con A and IL-12 in the presence of IL-6, produced less IFN-γ, and more IL-4 (Fig. 1,A). IL-6, like IL-2, has been described to be a growth factor for a number of cells (14, 15). However, only IL-6, but not IL-2, was able to modify the polarization of the CD4⁺ cells to Th2 phenotype (Fig. 1,A), indicating that IL-6 is involved in differentiation rather than growth of T cells. To eliminate the possibility that IL-6 could favor the expansion or IL-4 secretion of the CD4⁺ memory subpopulation, which has been described to display a Th2 phenotype (20), we analyzed the role of IL-6 in the differentiation of purified naive CD4⁺ T cells. Thus, we used CD4⁺ T cells from TCR transgenic mice (21), which express the α and β chain of the TCR that recognizes a pigeon Cyt c peptide; before further purification, 90–95% of the CD4⁺ T cells express the naive phenotype. Total CD4⁺ T cells were stained with anti-CD44 and anti-CD45RB mAbs and the naive CD4⁺ CD44^lowCD45RB^high population was isolated by cell sorting and activated with the same polyclonal stimulus, Con A, in the presence or absence of different cytokines. We observed that IL-6, in the absence of any other cytokine, was able to promote the differentiation of naive CD4⁺ cells to IL-4–producing cells (Fig. 1,B). In addition, we also analyzed the effect of IL-6 in the differentiation of naive CD4⁺ T cells stimulated by specific antigen. The presence of IL-6 during the activation with Cyt c peptide drove differentiation of naive CD4⁺ T cells into IL-4–producing effector Th2 cells as well or better than IL-4 (Fig. 1 C). Therefore, the modulatory effect of IL-6 in T cell differentiation is not a consequence of an expansion of the memory subpopulation.

Other cytokines have also been described to indirectly modulate the polarization towards Th1 and Th2. IL-10, a Th2 cytokine, promotes Th2 and inhibits Th1 cells, their cytokines, and related immune phenomena. Although the mechanism is not clear yet, several lines of evidence indicate that IL-10 reduces the Th1 response by an inhibitory effect on the IL-12 expression by APCs such as macrophages (22–24). In contrast, the immunoregulatory role of IL-6 in the polarization of Th1 and Th2 is not an indirect effect on the APCs. Thus, IL-6 inhibited IFN-γ production and stimulated IL-4 synthesis by Th1 cells that have been differentiated in the presence of exogenous IL-12 (Fig. 1, A and B), eliminating the possibility of inhibiting the production of IL-12 by APCs. In addition, IL-6 did not affect the expression on the APCs of co-stimulatory molecules such as B7.1 and B7.1, which have also been involved in the differentiation of Th1 and Th2 cells (5, 6; data not shown). To further prove that the effect of IL-6 was on T cells directly, rather than APCs, we differentiated CD4⁺ T cells with immobilized anti-CD3 mAb plus anti-CD28 mAb in the complete absence of APCs, and then restimulated these cells with anti-CD3 mAb alone. The presence of IL-6 (or IL-4 as a control) during the first culture resulted in an increase of IL-4 production and a dramatic reduction of IFN-γ production (Fig. 1 D) by the cells that were elicited, showing that IL-6 directly favors the polarization of naive CD4⁺ T cells to Th2 cells via the T cell.

IL-4 is the most effective differentiation factor for Th2 cells; it acts by promoting the secretion of more IL-4 and inhibiting the production of IFN-γ by T cells (9, 10). It was therefore possible that IL-6 could directly upregulate the synthesis of IL-4 by T cells and, consequently, that the IL-6 effect was mediated through IL-4, which would be responsible for the suppression of Th1 differentiation, while favoring a Th2 response. To address this hypothesis, CD4⁺ T cells were differentiated with Con A and IL-6 in the presence or absence of neutralizing anti–IL-4 mAb, and after 4 d, the cells were washed and restimulated with Con A. The ability of IL-6 to polarize CD4⁺ T cells toward the Th2 phenotype was blocked by anti–IL-4 mAbs, since the cells were unable to produce IL-4 upon restimulation (Fig. 1 E). These results indicated that the differentiation of Th2 cells by IL-6 is dependent on the endogenous production of IL-4. It was therefore possible that IL-6 may trigger the Th2 pathway by inducing small amounts of IL-4, which in turn would act as an autocrine differentiation factor for the Th2 cells.

As mentioned above, APCs represent the major source of IL-6 early in the immune response. To examine the physiological role of IL-6 in T cell differentiation, we first measured the production of IL-6 during the differentiation of Th1 and Th2 cells. CD4⁺ T cells were stimulated with Con A and IL-4 or IL-12 in the presence of APCs, and supernatants were harvested after different periods of time to measure IL-6 secretion. After 2 d of culture, identical levels of IL-6 were detected in the IL-4 and IL-12 cultures (Fig. 2,A). However, while the IL-6 level in the presence of IL-12 was sustained during the course of T cell differentiation, it decayed dramatically during the differentiation of Th2 cells in the presence of IL-4. No IL-6 was detected after restimulation of either Th1 or Th2 cells with Con A in the absence of APCs (data not shown), suggesting that the IL-6 that we detected during the first stimulation was secreted mainly by the APCs. The difference in the kinetics of IL-6 synthesis during the differentiation of Th1 and Th2 could, therefore, be due either to upregulation of IL-6 on the APCs in the presence of IL-12, or greater IL-6 consumption by Th2 cells than by Th1 cells. To test this, we examined IL-6 production during the differentiation of Th1 and Th2 in the presence of an anti–IL-6 receptor (IL-6R) mAb, to block IL-6 consumption. In the presence of anti– IL-6R mAb IL-6 did not diminish during the differentiation of Th2 cells (Fig. 2,B). These data indicate that the IL-6 produced by APCs was consumed during the differentiation of Th2 cells, but not during differentiation of Th1 cells, supporting the idea that IL-6 plays a role during Th2 polarization by acting directly on T cells. IL-6R is a heterodimer of the signal transducer gp130 (which is also a component of the IL-11, ciliary neurotrophic factor, leukemia inhibitory factor, and onconstatin M receptors) and the specific IL-6Rα chain (25–30). IL-6R has been found in both unstimulated CD4⁺ and CD8⁺ T cell subsets and its expression is downregulated upon activation (31). Differential expression of the IL-6Rα chain during Th1 or Th2 differentiation could explain the higher IL-6 consumption by the Th2 cells. We also analyzed, therefore, the expression of IL-6Rα during the differentiation of CD4⁺ T cells in effector Th1 or Th2 cells. However, the expression of cell surface IL-6Rα was regulated similarly during the differentiation of Th1 or Th2 cells in the presence of either IL-4 or -12 (Fig. 2 C). Low levels of IL-6Rα were present on unstimulated CD4⁺ T cells, and downmodulation occurred after the first day of stimulation, remaining at almost undetectable levels during the differentiation of both Th1 and Th2 cells. These results indicated that the effects of IL-6 in the differentiation of Th2 cells were not due to a differential distribution of the IL-6Rα.

If IL-6 was required for Th2 cell differentiation in vitro, it would follow that T cells from IL-6^−/− mice should be incapable of developing Th2 effector cells. The experiments described below show that this appears to be true. In correlation with the previous characterization of IL-6^−/− mice (32, 33), analysis of cellular populations in the spleen and thymus did not show any differences between wildtype (WT) and IL-6^−/− mice (data not shown), and IL-6^−/− mice developed normally. We therefore used splenocytes from IL-6^−/− mice as APCs to further establish that the production of endogenous IL-6 by the APC plays a critical role in the polarization of the CD4⁺ T cells to effector Th2 cells. We purified CD4⁺ T cells from WT mice and stimulated them for 4 d with Con A in the presence of APCs from WT mice or IL-6^−/− mice in the absence of any added exogenous cytokines. Restimulation with Con A resulted in significantly less IL-4 production in cells differentiated in the presence of IL-6^−/− than in WT APCs (Fig. 3 A). Although the production of IFN-γ under these conditions (absence of exogenous IL-12) was low, the levels of IFN-γ produced by CD4 cells differentiated in the presence of IL-6^−/− APCs were higher than those produced by cells stimulated in the presence of WT APCs (data not shown).

In vivo experiments have showed an impaired immune response in IL-6^−/− mice. These mice fail to control infection with Listeria monocytogenes and have a diminished T cell– dependent body response against vesicular stomatitis virus (32–34). In addition, CD4⁺ cells from WT mice infected with Candida albicans express IL-4, but reduced IgE and IL-4 mRNA was observed in IL-6^−/− mice (35). Nevertheless, due to the numerous effects of IL-6 in vivo, it is difficult to study by which mechanism IL-6 affects the immunopathology of infection and whether this is mediated by a modification in T cell differentiation. We examined, therefore, the in vitro differentiation of CD4⁺ T cells from IL-6–deficient mice. IL-6^−/− CD4⁺ T cells stimulated with Con A in the presence of WT APCs, which can provide the IL-6 required, produced IL-4, but this IL-4 production was impaired when the IL-6 pathway was blocked by the addition of anti–IL-6 mAbs (Fig. 3,B). Most significantly, however, was that in the complete absence of IL-6, when IL-6^−/− CD4⁺ T cells were stimulated in the presence of IL-6^−/− APCs, no in vitro Th2 response was obtained (Fig. 3,B). The difference between complete absence of IL-4 production when IL-6^−/− APCs were used with IL-6^−/− T cells compared with IL-6^−/− APCs and WT T cells may have been the result of either IL-6 production from contaminating WT APCs in the T cells or a contribution of IL-6 from the WT T cells. Moreover, high IL-4 production was restored in cells that were differentiated in the presence of IL-6^−/− APCs together with an exogenous source of IL-6 (Fig. 3 B). Similar levels of IL-4 were detected when CD4⁺ T cells were differentiated in the presence of IL-4, and no significant additional increase was observed when both IL-4 and IL-6 were present during the differentiation, indicating again that both IL-4 and IL-6 were using the same pathway.

To demonstrate whether the upregulation of IL-4 production by IL-6 was due to an induction of IL-4 gene expression rather than a posttranslational mechanism, e.g., increasing IL-4 secretion, we examined the levels of IL-4 mRNA. Naive CD4⁺ T cells were isolated from Cyt c TCR transgenic mice (see above) and stimulated with specific Cyt c peptide in the absence or presence of IL-6 or IL-4. After 4 d, cells were washed and restimulated with Cyt c peptide, and total RNA was isolated 20 h later. The expression of the IL-4 gene was determined by competitive RT-PCR of IL-4–specific transcripts (18). The presence of exogenous IL-6 or IL-4 during the differentiation (Th2 cells) resulted in the induction of high levels of IL-4 mRNA upon antigen restimulation (Fig. 3 C). Quantitation of IL-4 transcripts in cells differentiated in the presence of IL-6 or IL-4 were 100-fold higher than in the absence of cytokines (data not shown). Together, these data indicate that IL-6 causes the differentiation of Th2 cells by upregulating IL-4 gene expression. The transcriptional regulation of the IL-4 gene has not yet been characterized as extensively as the regulation of the IL-2 gene. Several positive and negative regulatory elements in the 5′ flanking region of the IL-4 gene have been described (36–39). Among five different nuclear factor of activated T cells (NFAT) binding elements, one located between −90 and −60 is important for the activation of the IL-4 gene expression. Like the distal NFAT site in the IL-2 promoter (40–42), this element binds NFAT (either NFATp or NFATc) in association with AP-1 components (37, 43). In addition, recently three functional binding sites for the C/EBP family of transcriptional factors have been identified in the IL-4 promoter (44), which might be regulated by IL-6 since members of this family (NF-IL6) have already been described as having been activated by IL-6 in other systems (45–50). Whether IL-6 acts through these sites will be the subject of future studies.

Both in vivo and in vitro experiments support the idea that, although other factors can play immunomodulatory roles, IL-4 and IL-12 are both necessary and sufficient for the polarization of Th1 and Th2 during the response to different pathogens. Considerable interest in the initial source of the IL-4 that drives Th2 cell differentiation has been expressed and recent attention has focused on mast cells or NK1.1⁺ CD4 T cells. We propose here that IL-6 is one of the stimuli in the initial production of IL-4 which provides a new mechanism to initiate Th2 CD4⁺ T cell differentiation. The levels of IL-6 that we used in these experiments are relatively high, just as the levels of IL-4 that also must be used to direct Th2 differentiation in vitro are (500– 1,000 U/ml IL-4). Presumably, for both IL-4 and IL-6 produced endogenously, lower levels produced in a local microenvironment suffice. Consistent with this idea, although the IL-6 effect is mediated through IL-4 (since it is blocked by anti–IL-4), the levels of IL-4 that can be detected are very low, presumably because of local consumption, but are sufficient to direct Th2 cell differentiation. Our data suggest a model (Fig. 4) for Th2 differentiation that is strikingly similar to the mechanism used for Th1 differentiation. In both cases, the source of the polarizing cytokine is the APC, IL-12 in the case of Th1 cells and IL-6 in the case of Th2 cells. Although APCs are the major source of IL-6 in vitro, other cell types that produce IL-6, such as fibroblasts, endothelial cells, keratinocytes or others, could also contribute to the overall in vivo IL-6 production. In our in vitro model, activation of the APC is proposed to result in the secretion of IL-6 which, in combination with the antigen, will participate in the induction of IL-4 gene expression in naive CD4⁺ T cells. In the second step, the low levels of IL-4 secreted by the T cells is sufficient to induce upregulation of the IL-4 and IL-4 receptor genes in an autocrine manner, while it inhibits the expression of the IFN-γ gene. Cells are therefore subsequently polarized to a Th2 phenotype using their own source of IL-4.

We thank S. Samanta for technical assistance and F. Manzo for assistance with manuscript preparation. We also thank Genetics Institute for providing the IL-12 and DNAX Institute for a gift of IL-4.