Control of T Helper Cell Differentiation--in Search of Master Genes

See allHide authors and affiliations

Science's STKE  12 Sep 2000:
Vol. 2000, Issue 49, pp. pe1
DOI: 10.1126/stke.2000.49.pe1


Naïve T helper (TH0) cells can differentiate into one of two distinct populations: TH1 and TH2. Each population is characterized by the expression of specific cytokines and their ability to participate in cell-mediated or humoral immune responses. Recent efforts at identifying the molecular mechanisms through which TH0 cells become TH1 or TH2 cells have been promising. A number of transcription factors, including GATA-3 and T-bet, have been identified that promote the differentiation of TH0 cells and the maintenance of the differentiated cell phenotype. Dong and Flavell review recent findings on proteins that control the fate of TH0 differentiation, whether by promotion or inhibition, and discuss the role of epigenesis in the differentiation process.

Naïve T helper (TH0) cells exist in our peripheral lymphoid organs as immunocompetent precursors that, upon activation, can differentiate into one of two functional subsets that are distinguished by their cytokine production profiles and their roles in immune regulation (Fig. 1). TH1 cells produce proinflammatory cytokines such as interferon-γ (IFN-γ) and lymphotoxin-α; effector TH2 cells, in contrast, produce cytokines [interleukin-4 (IL-4), IL-5, IL-9, IL-10, IL-13, and so forth] that together instruct antiparasite and, to some extent, humoral immune responses. T helper (TH) cell differentiation can be observed not only during in vivo immune reactions, but also can be recapitulated in vitro, when TH cells are activated through their T cell receptors (TCRs) and costimulatory receptors (i.e., provided by antigen-presenting cells (APCs) stimulation or by anti-CD28 cross-linking). Under these conditions, addition of IL-12, normally produced by activated APCs in vivo, can strongly induce TH1-polarized differentiation, whereas addition of IL-4 skews TH cells to become TH2 effectors. Facilitated by this culture system and powerful molecular cloning and mouse genetics techniques, several key transcription factors and signaling molecules that govern the TH differentiation program have been identified over the past years.

Fig. 1.

Transcription factors and TH cell differentiation. Upon activation by APC, naïve TH cells produce IL-2, undergo clonal expansion, and become TH0 cells, which can then differentiate into TH1 or TH2 effector cells. TH1 and TH2 cells differ in the cytokines that they produced and in the immune regulation performed. TH1 and TH2 differentiation is determined by several key cytokines: IL-12, produced by APC, potentiates TH1 differentiation; IL-4, on the other hand, drives TH2 development. Polarized TH cell differentiation is also characterized and mediated by transcription factors expressed or by function in lineage-specific fashion. These factors include STAT4 and T-bet for TH1 and STAT6, GATA-3, c-Maf, JunB, and NFATc for TH2 differentiation.

IL-4 Transcription Regulation

IL-4 is produced by differentiating TH2 cells and by fully differentiated TH2 cells and is also the key factor that drives uncommitted, bipotential TH0 cells into the TH2 pathway. IL-4 gene transcription has been extensively studied by a number of groups. The 800-base pair (bp) proximal promoter of the IL-4 gene was shown to confer TH2 specificity in transgenic mice despite a substantially lower expression level than the endogenous gene (1). Furthermore, reporter vectors containing an IL-2 or IL-4 minimal promoter into which the NFAT-AP1 composite DNA element from the 800-bp proximal promoter has been inserted (1-3) were expressed in a TH2-specific fashion, indicating that the NFAT and AP1 transcription factor families can regulate IL-4 gene expression. NFATc and NFATp bind to the NFAT-AP1 composite DNA site in the IL-4 promoter and activate IL-4 transcription in vitro (4, 5). However, analysis of mice deficient in NFATc and NFATp distinguished the roles of these genes in IL-4 regulation: Mice deficient for NFATp or for both NFATp and NFAT4 have an exaggerated TH2 response (6-8), whereas NFATc-deficient mice displayed impaired TH2 differentiation (9, 10). Thus, NFATp and NFATc seem to play opposite regulatory roles in IL-4 regulation. Precisely why this is so remains a puzzle.

GATA-3 was identified as a lineage-specific factor selectively expressed in the TH2 pathway (11, 12). GATA-3 is present at low levels in naïve and newly activated CD4+ T cells. However, in differentiated TH2 cells, GATA-3 levels are elevated (11, 12), whereas TH1 cells have undetectable levels of GATA-3 mRNA and IL-4 gene expression. Ouyang et al. (13) found that the reduced GATA-3 expression during TH1 differentiation is dependent on IL-12. Within the past 3 years, several studies have demonstrated that GATA-3 is the critical regulatory transcription factor involved in TH2 differentiation. T cells isolated from transgenic mice overexpressing GATA-3 produced IL-4, IL-5, IL-6, and IL-10 mRNA when cultured under TH1 conditions in the complete absence of exogenous IL-4 (12). Zhang et al. also found that GATA-3 could activate the IL-5 promoter in vitro (11). Moreover, GATA-3 was also reported to inhibit IL-12-mediated IFN-γ production in T cells (13). On the other hand, reduction of GATA-3 function by antisense expression in cloned TH2 cells led to the extinction of IL-4, IL-5, IL-6, IL-10, and IL-13 TH2 cytokine gene mRNA expression and protein secretion (12). This result was recently confirmed by expression of a dominant-negative GATA-3 mutant in murine T cells, which resulted in reduced expression of the TH2 cytokines IL-4, IL-5, and IL-13 (14). Arai and colleagues showed recently that retrovirally expressed GATA-3 could cause differentiated cloned TH1 cells to produce IL-4 and IL-5 (15), and Ouyang et al. (16), using a similar approach, found that GATA-3 overexpression in signal transducer and activator of transcription protein 6 (STAT6)-deficient cells resulted in TH2 cytokine expression and the establishment of TH2-specific deoxyribonuclease I (DNase I)-hypersensitive sites in the IL-4 gene locus. Together, these experiments show that GATA-3 is the master TH2 regulatory factor both necessary and sufficient to generate TH2 responses. However, how GATA-3 works is still not understood. The proximal promoter of the IL-4 gene lacks a strongly functional GATA-3 binding site; therefore, its major role is likely to be at a TH2-specific enhancer(s) or even perhaps a locus control region outside of the minimal IL-4 gene promoter.

In addition to GATA-3, two bZIP transcription factors, c-Maf and JunB, are also expressed in a TH2-specific manner and bind to the proximal IL-4 promoter (2, 17, 18). Overexpression of c-Maf strongly induced IL-4 expression in non-T cells (18). Transgenic mice in which c-Maf overexpression was driven by the CD4 promoter exhibited enhanced TH2 responses dependent on IL-4 (19). The importance of c-Maf in controlling IL-4 expression is further strengthened by a report that c-maf-/- TH and natural killer T cells exhibit impaired IL-4 production (20). However, TH2 differentiation and cytokine production appear normal in the absence of c-Maf, probably because of compensation by normal levels of IL-13. On the other hand, overexpression of the AP1 family member JunB in transgenic mice caused developing TH1 cells to overproduce IL-4 and IL-5 and, to a lesser extent, IL-6 and IL-10 (all TH2 products) through a synergistic mechanism involving c-Maf (17). This synergy was specific for JunB because neither c-Jun nor JunD is capable of synergizing with c-Maf.

The IL-4 gene is located on chromosome 11 in a locus containing the other TH2 cytokine genes, IL-5 and IL-13, which suggests that the expression of these genes is coordinated. Two groups have identified several DNase I-hypersensitive sites in the IL-4 and IL-13 gene locus that are selectively associated with TH2 differentiation (21, 22), indicating that chromatin may be remodeled during TH2 commitment, making the locus more accessible for transcription. In addition, remodeling of the IL-4 gene locus is accompanied by DNA demethylation and was shown to require both TCR and IL-4 receptor signaling (21). Loots et al. examined 1 megabase of human and murine DNA containing TH2-regulated genes and found several homologous noncoding sequences shared between them. Some of these sequences are conserved across several mammalian species (23). These regions could serve as candidate enhancer/locus control sequences for TH2 genes. Indeed, these investigators generated a transgenic mouse strain carrying a yeast artificial chromosome encoding the human IL-4 locus and showed that the human TH2 cytokine genes could be expressed properly in these mice. Deletion of one of the conserved noncoding regions greatly reduced the efficiency but not the TH2 specificity of gene expression, indicating that the DNA region between the IL-4 and IL-13 genes could act as an important enhancer sequence that regulates TH2 cytokine expression. Agarwal et al. identified another putative enhancer: a DNase I-hypersensitive site located downstream of the IL-4 gene (24). In this study, they found that GATA-3 but not c-Maf bound to this enhancer; in response to TCR signaling, NFATp also bound to the site, in a cyclosporin A-inhibitable manner. Although this element is likely to support GATA-3-dependent IL-4 gene transcription, it is possible that additional GATA-3 targets exist in the locus.

Recently, two reports demonstrated that IL-4 is expressed in a monoallelic fashion. Bix and Locksley established TH2 cell clones from BALB/c x CaST/Ei F1 mice and analyzed the IL-4 genes for evidence of allelic expression (25). They found that most clones expressed IL-4 in a monoallelic fashion and that the allelic pattern was transmitted as a stable epigenetic trait. In another study, Riviere et al. generated a mouse strain in which one allele of the IL-4 gene was replaced by the human CD2 gene (26). Using this model, they demonstrated that most TH2 cells expressed only the functional IL-4 allele or only the targeted allele. The frequency of monoallelic versus biallelic expression in this model suggests that allelic activation is a stochastic process whereby the chance of activating each cytokine gene allele can result in the generation of diverse profiles of cytokine expression and subsequent effector cell populations. In addition, Reiner and colleagues also found that IL-4 expression in T cells requires passage through more cell cycles than are required for IFN-γ expression, suggesting that rounds of cell division in concert with changing profiles of polarizing cytokines may relieve epigenetic repression (27). In contrast to differentiating TH cells, fully differentiated TH effector cells exhibit stable epigenetic modification and cell cycle-independent gene expression.

The above studies point out an interesting new direction for understanding TH2 cytokine gene regulation. TCR signaling, possibly through NFATc, initiates the clonal expansion and low-level IL-4 expression by TH cells. At this stage, IL-4, produced by other regulatory cell types during immune reactions or by proliferating TH0 cells through a STAT6-dependent mechanism, increases and sustains GATA-3 expression. It is possible that GATA-3, with other lineage nonspecific nuclear factors, mediates chromatin remodeling to render the TH2 locus on one chromosome accessible to other lineage- and gene-specific factors, such as c-Maf and JunB. Increased production of IL-4 would enhance T cell proliferation and, on the other hand, might provide stronger signals to open up the additional allele.

Regulation of IFN-γ Expression

The prototypic cytokine for TH1 cells is IFN-γ. However, in comparison to the IL-4 gene, less is known about how IFN-γ gene transcription is regulated. Agarwal et al. showed that, during TH1 differentiation, chromatin remodeling of the IFN-γ locus occurs (21). Penix et al. identified c-Jun/ATF2 sites and a series of other ATF binding sites within two functionally active elements called the proximal and distal IFN-γ elements (28, 29). Transgenic mice were generated whereby these two elements were linked to luciferase reporter constructs, demonstrating that the proximal element showed TH1 specificity, whereas the distal element did not (30, 31). In addition, Young and colleagues found an intronic enhancer in the IFN-γ gene (32). Interestingly, c-Rel was found to bind to this enhancer but not to a similar sequence from the IL-2 promoter.

A TH1-restricted transcription factor has recently been described (33). T-bet is a T-box protein that is expressed in TH1 clones and differentiating TH1 cells. Overexpression of T-bet activates IFN-γ expression in cell lines or primary activated T cells. Even more striking, T-bet expression by retroviral transduction induced IFN-γ expression and suppressed IL-2, IL-4, and IL-5 production in differentiating and differentiated TH2 cells. Thus, T-bet appears to be a candidate TH1 master gene. The mechanism of its action and regulation remains to be solved and will no doubt provide a better understanding of TH differentiation regulation.

Signaling Regulation and Cross-Regulation

TH cell differentiation is largely determined by the cytokine environment. Cytokine-induced receptor dimerization activates the Janus family of tyrosine kinases, which then phosphorylate and initiate the nuclear transport of STAT transcription factors. STAT6, which acts downstream of IL-4, is required for TH2 differentiation (34-36). Bcl-6 was reported to compete with STAT6 for the same binding site and thus was considered to be a negative regulator of TH2 responses (37). This conclusion is supported by studies in Bcl-6-deficient mice that exhibited TH2-specific autoimmune diseases (37). Although the function of STAT6 in TH differentiation has been well established, the mechanism of its action still remains unclear. For instance, the proximal IL-4 promoter lacks a STAT binding site. One study reported that STAT6 may bind to a site in an intronic repressor and release the inhibition (38); however, whether this is the physiological function for STAT6 remains to be determined. Alternatively, STAT6 may function to regulate expression of other transcription factors involved in TH2 differentiation by binding to sites in genes such as GATA-3 and c-maf and by functioning to regulate their expression. Indeed, Kurata and colleagues found that retroviral expression of STAT6 can result in GATA-3 and c-Maf expression and TH2 cytokine production in differentiated TH1 cells (39). On the other hand, STAT4 is activated by IL-12, and STAT4-deficient mice are profoundly deficient in TH1 cells (40, 41). How STAT4 regulates TH1 differentiation is very unclear. STAT4 may induce T-bet expression, bind to a TH1-specific enhancer to mediate chromatin remodeling, or simply antagonize STAT6 function. IFN-γ induces the expression of the major histocompatibility complex (MHC) class II transactivator (CIITA) (42). Recently, CIITA-deficient T cells were found to have increased TH2 cytokine production and exhibited differentiation skewed toward the TH2 phenotype (43), suggesting that CIITA serves as a negative regulator of the TH2 program. Thus, CIITA may provide another mechanism for TH1/2 cross-inhibition.

There are three major mitogen-activated protein kinase (MAPK) pathways in mammals: the extracellularly regulated kinase (ERK), p38, and Jun NH2-terminal kinase (JNK) pathways. The Ras-ERK pathway has an important role in early thymocyte differentiation and selection; however, its role in peripheral TH cell differentiation was not studied until recently. Yamashita et al. used dominant-negative Ha-ras transgenic mice (in which ERK activation by the TCR was severely compromised) and showed that this pathway is required for TH2 differentiation (44). Similar results were found using wild-type cells treated with inhibitors against MEKs, the upstream kinases that activate ERKs. Yamashita et al. also demonstrated that the function of the ERK pathway was to enhance IL-4-induced phosphorylation of STAT6 and IL-4R, suggesting a mechanism of IL-4 signaling regulation by ERK (44).

The p38 MAPK pathway is selectively activated in mouse TH1 effector cells. Studies using imidazole inhibitors of the p38 kinases, transgenic mice expressing a dominant-negative p38α molecule (45), or mice deficient in the p38 upstream kinase MAPK kinase kinase 3 (MKK3) (46) all demonstrated that IFN-γ production by TH1 cells required p38 signaling. Consistent with these results, primary T cells expressing constitutively active MKK6, one of the upstream MEKs that activates p38 kinase, had higher levels of IFN-γ than control cells (45). Candidate downstream targets of p38 regulation are likely to include transcription factors of the ATF family.

JNK phosphorylates c-Jun and increases AP1 transcriptional activity (47). JNK activation in Jurkat tumor cells has been shown to require both TCR and CD28 ligation, and it has been suggested that JNK integrates signals from both sets of receptors to induce IL-2 expression (48). Unlike Jurkat cells, however, naïve CD4+ T cells express only very low levels of JNK proteins and mRNA (49). Induction of JNK follows activation of primary mouse T cells, but this process is slow and peaks at ~2 or 3 days after T cell activation; over this period, there is substantial de novo synthesis of JNK protein. Thus, JNK activity correlates well with the expression of effector cytokines in both the TH1 and TH2 pathways but correlates poorly with the expression of the growth factor IL-2. Indeed, mice deficient for either Jnk1 (50) or Jnk2 (51) expression, Jnk1-/-Jnk2-/- mice (52), and mice lacking the MKK7 gene (whose protein product activates JNKs) (52) all exhibited defective phenotypes associated with TH differentiation but not with TH activation. JNK2-deficient mice showed impaired TH1 immune responses (51), and reduced levels of IFN-γ in Jnk2-/- mice appear to derive from a reduced production of the IL-12 receptor β chain (IL-12Rβ2). On the other hand, mice deficient for JNK1, both JNK1 and JNK2, or MKK7 showed an exaggerated TH2 response (50, 52). Consistent with this exaggerated TH2 response, infection of JNK1-deficient mice with Leishmania leads to greatly exacerbated disease with failure to heal skin lesions (53); the disease advances to ulceration in a manner similarly observed in BALB/c mice, which have a genetically determined strong TH2 response. This indicates that JNK is required in vivo and in vitro to restrict TH2 cytokine expression and to maintain polarized TH1 responses. Examination of expression of TH2 transcription factors at this stage shows that JNK1-deficient mice have elevated NFATc in the nucleus (50). The mechanism whereby NFATc accumulates has now been elucidated (54). Specifically, JNK serves as an NFATc kinase, phosphorylating NFATc on two serine residues, one of which lies in the calcineurin binding site. JNK inhibits the targeting of the phosphatase calcineurin to NFATc and thus opposes NFATc nuclear translocation. In the absence of JNK signaling, NFATc is inefficiently removed from the nucleus, which results in nuclear accumulation of NFATc and excessive production of TH2 cytokines.

The p38 and JNK MAPKs are both selectively activated in TH1 cells; however, the mechanism that underlies this selectivity was not understood until recently. Li et al. found that the expression of the small guanosine triphosphatase (GTPase) Rac2 is enhanced in TH1 cells (55). Because MAPK pathways can be activated by small GTPases, we tested the hypothesis that elevated Rac2 was responsible for mediating this selective activation. Indeed, Rac2 activation increases IFN-γ transcription in vitro and in vivo in Rac2 transgenic mice through several signaling pathways, including p38. Blocking the Rac2 pathway led to the inhibition of IFN-γ production and gene transcription. Rac2-deficient T cells also had reduced levels of IFN-γ.

Rac2 also mediates IFN-γ transcription by activating the inhibitor of κB (IκB) kinase-nuclear factor κB (NF-κB) pathway (55). IκB, a suppressor of NF-κB nuclear translocation, blocks Rac2-activated IFN-γ transcription in vitro, consistent with an earlier study on the human IFN-γ promoter (32). In support of this mechanism, Aronica et al. used transgenic mice expressing a similar IκB mutant that cannot be phosphorylated and recently demonstrated that the NF-κB pathway is required for TH1-dependent delayed-type hypersensitivity responses but not for the TH2-characterized airway hypersensitivity response (56).

Future Perspective--Is Everything Known Yet?

Much knowledge and insight has been gained from the past 10 years of intensive study. It appears that after T cell activation, there is chromosomal remodeling of cytokine gene loci that is likely mediated by transcription factors, including the NFAT and AP1 family, which results in low levels of cytokine expression and leads to the so-called TH0 stage. At this stage, different environmental factors and cytokines activate signaling pathways culminating in the activation or suppression of transcription factors such as T-bet and GATA-3 that regulate the development of TH1 or TH2 cells, respectively (summarzied in Fig. 2). In this way, gene transcription patterns are established and might be transmitted epigenetically. Cells are programmed to become effector cells that use lineage-restricted transcription factors (c-Maf and JunB) and specific signaling pathways (p38, JNK, and NF-κB) and produce high levels of cytokines to orchestrate immune responses.

Fig. 2.[AU1]

Signaling integration and cross-regulation in TH cell differentiation. Transcription factors mediating cytokine production in TH1 or TH2 cells are regulated by signal transduction pathways initiated from TCR or cytokine (IL-4 or IL-12) receptors. Some pathways cross-potentiate or inhibit other distinct pathways. In the diagram, positive regulation is indicated as an arrow, and negative regulation is indicated as a block arrow.

Real-life immune responses and mechanisms of TH cell differentiation, however, are probably much more complicated than this. In vitro culture systems have allowed the rapid identification and characterization of factors involved in TH differentiation, many of which still await "trials" of real-life immune challenges. Genetic manipulations in mice will greatly facilitate these in vivo experiments, which are compounded by numerous factors, including the nature of antigen presentation, the presence of different costimulatory factors, the physical sites of immune reactions, the production of inflammation cytokines and chemokines, feedbacks from effector cells, hormonal regulation, and so forth. Some of the in vitro findings will hold as time passes, and some will not. These types of studies will likely reveal still greater complexities of immune regulation and will ultimately provide useful therapeutics for treating immune diseases.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
View Abstract

Navigate This Article