Common themes emerge in the transcriptional control of T helper and developmental cell fate decisions regulated by the T-box, GATA and ROR families

Abstract

Cellular differentiation requires the precise action of lineage-determining transcription factors. In the immune system, CD4⁺ T helper cells differentiate into at least three distinct effector lineages, T helper type 1 (Th1), Th2 and Th17, with the fate of the cell at least in part determined by the transcription factors T-box expressed in T cells (T-bet), GATA-3 and retinoid-related orphan receptor γt (RORγt), respectively. Importantly, these transcription factors are members of larger families that are required for numerous developmental transitions from early embryogenesis into adulthood. Mutations in members of these transcription factor families are associated with a number of human genetic diseases due to a failure in completing lineage-specification events when the factor is dysregulated. Mechanistically, there are both common and distinct functional activities that are utilized by T-box, GATA and ROR family members to globally alter the cellular gene expression profiles at specific cell fate decision checkpoints. Therefore, understanding the molecular events that contribute to the ability of T-bet, GATA-3 and RORγt to define T helper cell lineages can provide valuable information relevant to the establishment of other developmental systems and, conversely, information from diverse developmental systems may provide unexpected insights into the molecular mechanisms utilized in T helper cell differentiation.

Introduction

As cells differentiate from multipotent progenitors into defined lineages, specific gene networks must be expressed to ensure the proper functioning of the cell. Although the exact genes being activated or repressed differ in the development of each cell lineage, the global mechanisms utilized to control cell fate choices that begin in early embryogenesis and continue through the differentiation of adult tissues are likely to be shared. Key developmental transcription factors are responsible for regulating genetic programmes, but precise knowledge of the molecular mechanisms through which they achieve this is often lacking. It is striking that seemingly unrelated cell fate decisions are controlled by transcription factors representing a relatively small number of protein families. This may suggest that these families have evolutionarily conserved functional capabilities that are required for common processes in cellular differentiation. Importantly, disrupting the activities of these transcription factors can have drastic consequences for the organism, as highlighted by the association of human genetic diseases with their mutation.¹ The challenge is to precisely define the mechanisms through which lineage-determining transcription factors direct cell development and to determine whether the functional activities uncovered for one specific factor will be able to be generalized to other family members in diverse developmental systems.

There are currently three known types of effector CD4⁺ T helper cells, T helper type 1 (Th1), Th2 and Th17.^2–4 The differentiation of a naïve T helper cell into a specific lineage is directed by signals from antigen-presenting cells in response to the type of pathogen they have encountered. The series of molecular events induced by these signals includes the stable expression of lineage-determining transcription factors.^5–8 T-box expressed in T cells (T-bet), GATA-3 and retinoid-related orphan receptor γt (RORγt) are considered to be the lineage-specifying transcription factors for Th1, Th2 and Th17 cells, respectively, because when absent, the corresponding type of effector does not develop, regardless of the signals received (Fig. 1).^2–4 The requirement for T-bet, GATA-3 and RORγt is not limited to the development of T helper cells. These factors are expressed and play important roles both within the immune system, where T-bet is involved in natural killer T (NKT) cell development, and in other systems, such as GATA-3 activating targets in mammary cells.^10,11 Further, the dysregulation of both T-bet and GATA-3 has been linked to the development and/or progression of different cancers, and disruption of all three factors has been implicated in autoimmune disorders.^1,9,10,12,13 Significantly, these proteins also belong to larger transcription factor families that are important in many developmental systems (Fig. 1).^1,9,10 This suggests that research in the immune system may have larger implications for, and draw from, mechanistic studies in other fields.

Figure 1.

T-box, GATA and retinoid-related orphan receptor (ROR) factors play essential roles in the development of (a) T helper cells and (b) other tissues. (a) In this simplified representation, naïve CD4⁺ T cells differentiate into distinct T helper type 1 (Th1), Th2 and Th17 subsets based upon the expression and functional activities of T-bet, GATA-3 and RORγt, respectively. (b) Similarly, during embryogenesis, stem cells require the activities of T-box, GATA and ROR family members to differentiate into distinct lineages such as those that form the heart and pituitary. The asterisk indicates the possible role of RORα in the pituitary. In the development of both effector T helper cells and other tissues, there are many other proteins, including other members of the T-box, GATA and ROR families, that are required. Also, in the case of the heart and pituitary, there are many intermediate steps between the common stem cell and the differentiated tissue. However, it remains striking that members of the same transcription factor families are involved in such disparate developmental decisions as heart, pituitary and T helper cell differentiation.

Molecular mechanisms of cellular differentiation

To establish T helper cell differentiation programmes, lineage-specifying transcription factors must be able to both activate and repress gene networks. Determining how this is achieved at the molecular level will allow us to truly understand what occurs in disease states and potentially provide the opportunity to induce cells to fight infections. The immune system provides a tractable model system where primary cells can be expanded and differentiated both in vivo and in vitro to address these questions in detail. These types of studies may provide testable hypotheses for developmental systems where such manipulations are more difficult. In this review, we will address three areas of research that are required to determine the mechanisms by which gene expression networks are established. We will examine them from the perspective of three key transcription factors involved in T helper cell differentiation as well as the counterpart family members found in other developmental systems. The first essential area of research is to define which genes are regulated, both positively and negatively, by each factor and whether that regulation is direct or indirect. Secondly, examining the protein–protein interactions for the lineage-defining transcription factors can shed light on the types of functional activities that are being regulated. For example, if a transcription factor physically interacts with a histone-modifying enzyme, this could indicate that epigenetic changes play a role at target genes. Finally, it is important to understand the post-translational modifications for each protein. Not only do they modulate protein–protein interactions, but they also indicate the types of signalling pathways that are utilized to alter the activity of the factor. Although this is by no means a comprehensive list of ongoing questions, it does provide a starting point for examining differentiation of both T helper cells and additional cell types that may broaden our mechanistic understanding of developmental transitions.

T-bet and the T-box family

Overview

In the immune system, the T-box factor T-bet is required for Th1 cell differentiation.^3,14 Other members of the T-box transcription factor family are expressed in a wide variety of tissues and are required for diverse cell fate decisions.¹ The critical role of the T-box family in cellular differentiation is best demonstrated by the number of human genetic diseases that are associated with mutations in T-box factors (Table 1). Not surprisingly, the disease phenotypes correspond to the tissue where each T-box protein exerts its control on cell fate decisions. For instance, mutations in Tbx5, which normally influences cardiac differentiation, result in congenital heart defects. In addition, mutations in other T-box factors are found in disease states such as Tbx19-mediated adrenocorticotropin (ACTH) deficiency, Tbx3-mediated ulnar-mammary syndrome and Tbx22-mediated X-linked cleft palate.^23–29 Taken together, genetic studies have provided strong evidence that T-box factors regulate developmental transitions.

Table 1.

Indicated are the expression patterns, human disease associations and phenotypes for the murine homozygous deletion of the T-box, GATA and retinoid-related orphan receptor (ROR) family members

Family	Transcription factor	Cell type expression	Human disease association	Mouse knockout phenotype
T-box	Brachyury	Embryo, mesoderm	Weak association with spina bifida15	Embryonic lethal
	Tpit	Pituitary	ACTH deficiency	ACTH deficiency and pigment defects
	Tbx20	Heart	Congenital heart defects (CHD)16	Embryonic lethal
	Tbx10	Palate	Unknown	Cleft lip/palate
	Tbx1	Craniofacial tissue, heart, mesoderm	Digeorge syndrome; heart, glandular and vascular defects	Neonatal lethal
	Tbx22	Palate	X-linked cleft palate	Unknown
	Tbx18	Heart, vertebrae	Unknown	Perinatal lethal
	Tbx15	Limb, craniofacial	Unknown	Skeletal limb defects
	Tbx2	Heart, limb, embryo	Duplicated in breast cancer17	Embryonic lethal
	Tbx3	Heart, limb	Ulnar-mammary syndrome	Embryonic lethal
	Tbx4	Limb, embryo	Small patella syndrome18	Embryonic lethal
	Tbx5	Heart, limb	Holt–Oram syndrome, CHD	Embryonic lethal
	Tbx6	Vertebrae, mesoderm	Unknown	Embryonic lethal
	T-bet	Immune cells, including Th1, NKT, NK, CD8+ T cells and dendritic cells	Asthma with decreased T-bet	Immune defects including Th1 and NKT cell deficiency, asthma, increased tumour metastisis,69 resistance to EAE,70 colitis71
	Eomesodermin	Brain, mesoderm, immune cells	Microcephaly, chronic infections, motor delays19	Embryonic lethal
	Tbr1	Brain	Unknown	Perinatal lethal20
GATA	GATA-1	Erythroid cells, testis	Dyserythropoietic anaemia, thrombocytopenia	Embryonic lethal
	GATA-2	Pituitary, testis, adipocytes	Coronary artery disease, CML21	Embryonic lethal22
	GATA-3	Immune cells, including Th2 cells, mammary cells	HDR syndrome	Embryonic lethal
	GATA-4	Pancreas, testis, heart, neurons	CHD	Embryonic lethal
	GATA-5	Intestine, heart, pancreas	Unknown	Genitourinary defects in females
	GATA-6	Heart, lung, liver, pancreas, testis	Unknown	Embryonic lethal
ROR	RORα	1: Cerebellum	Unknown	Slow hair growth, ataxia, cerebellar atrophy
		2: Cerebellum and widely expressed
	RORβ	1: Brain, retina	Unknown	Duck-like gait, blindness, retinal degeneration
		2: Pineal gland, retina
	RORγ	1: Widely expressed	Unknown	No Peyer’s patches or lymph nodes
		2(γt): Immune cells, including Th17

The cell type expression pattern listed for each family member represents the best characterized cell types that have high levels of expression, but this should not be considered an exhaustive list. Similarly, the human diseases associated with each factor in the table are primarily attributable to polymorphisms and there are probably more pathogenic conditions related to the dysregulation of each factor. In the ROR cell type expression section, the numbers refer to the two characterized isoforms for each factor. Unless indicated, the information in this table has been expertly reviewed by Naiche et al.¹ for the T-box family, Viger et al.¹⁰ for the GATA family, and Jetten & Joo⁹ for the ROR family.

NK, natural killer; Th, T helper; ACTH, adrenocorticotropin; EAE, experimental autoimmune encephalomyelitis; CML, chronic myeloid leukemia; HDR, hypoparathyroidism, deafness, and renal dysplasia.

T-bet and other members of the T-box family are defined by the highly conserved T-box DNA-binding domain.¹ This domain is large, consisting of approximately 180 amino acids, and is located in the centre portion of each T-box protein. The high conservation observed in this region explains the ability of T-box factors to bind a consensus sequence in the DNA, termed the T-site.¹ In addition, there are both conserved and divergent sequences not required for DNA contact within the T-box DNA-binding domain.³⁰ The N- and C-termini of the T-box proteins also are divergent in length and amino acid sequence between family members. The conserved nature of regions within the T-box, together with the divergence of the N- and C-termini, suggests that T-box factors will have both common and factor-specific functions. Collectively, the data generated to date compellingly demonstrate the significance of this family in development, but the molecular mechanisms through which individual members regulate cellular differentiation are not yet clear.

Identification of target genes

The majority of studies examining T-box protein target gene regulation have focused on prototypic, tissue-specific genes. Perhaps the best characterized studies involve the regulation of Anf by Tbx5 in the heart or the ability of T-bet to activate Ifng in the immune system.^{1,3,14,25,26,30–38} Although these hypothesis-driven studies have been very informative, they cannot provide information about the complex range of gene expression networks regulated by T-box factors. Significantly, it is unlikely that such drastic changes in the function of a cell lineage will be induced by the activation of a small subset of genes. To date, the major difficulty in using unbiased approaches to study the target genes of the T-box family has been that the deletion of most T-box factors in mice causes an embryonic lethal phenotype.¹ This makes comparative microarray strategies difficult at best. Despite these challenges, some global target gene approaches have been used successfully. cDNA microarray studies examining the role of Tbx1 and Tbx5, as well as a ChIP-on-chip study of T-bet, have provided a number of potential targets.^31,39,40 Based upon the published data, it appears that Tbx1 and Tbx5 up-regulate some of the same genes. Among the common targets, a few, such as the transcription factor Wilms’ tumour 1 (Wt-1), may be needed in many developmental transitions. However, other potentially common T-box targets do not have an obvious connection to developmental programmes. In order to manipulate these gene expression systems in a meaningful and specific way in either helper T cells or other cell types, we need to understand fully the nature of target activation and the potential overlap between family members.

Identification of the genes that are modulated in response to the presence of a T-box factor must be followed by determination of whether that gene is a direct or indirect target. Here, studies of the role of T-bet in Th1 development have been very informative and may provide unexpected insight into the direct target genes for the greater T-box family during cellular differentiation. It is interesting to note that many of the genes identified in our ChIP-on-chip experiments were bound by T-bet, but, to date, we have not been able to detect any functional regulation in the context of Th1 or other immune cell types.^31,34 It could be that some of these are targets of other T-box factors. In support of this possibility, another T-box factor, Brachyury, can bind to T-bet targets, including Ifng, but it is unable to induce transcription of prototypic T-bet targets.^30,34 Thus, we can speculate that embedded within the ChIP-on-chip target gene data is useful information pertaining to T-box family member target genes. Current data suggest that there also may be certain classes of genes that are regulated by all T-box factors, although this has not been formally demonstrated. If this is indeed the case, it will provide insight into common events that take place during cellular differentiation and offer a note of caution for future therapeutic manipulation.

Protein/protein interactions

To date, the two major classes of proteins with which T-bet has been shown to interact are other transcription factors and proteins with chromatin-remodelling activity. In some cases, a physical association has been demonstrated, while other data have classified functional interactions. The association between T-bet and other transcription factors can either enhance or inhibit the activity of T-bet, dependent upon the characteristics of the partner protein and the context of the interaction. One study has shown that Hlx, a homeobox transcription factor, both is regulated by T-bet and works co-operatively with it, to activate other target genes.⁴¹ Conversely, T-bet has been shown to bind to GATA-3, and this represses the activity of both proteins.⁴² Interestingly, the protein interaction data generated thus far with T-bet have remarkable similarities to themes in other developmental systems. Significantly, T-box, homeobox and GATA factors work in concert or opposition at a number of developmental transitions. In the heart, for example, it has been shown that Tbx5, Nkx2-5 and GATA-4 all work together to activate genes important in cardiomyocyte development.^{1,10,25,26,36} Perhaps similar to the mammalian immune system, T-box and GATA homologues in Caenorhabditis elegans activate opposing developmental gene profiles early in the differentiation of the endoderm and mesoderm.⁴³ It therefore seems likely that the ability of these protein families to functionally co-operate and antagonize each other has been conserved, but that the interactions themselves may have evolved specific consequences in each developmental system.

T-bet has also been shown to interact with chromatin-remodelling factors, specifically histone-modifying enzymes. Recently, it has been shown that T-bet interacts with at least two types of histone-modifying enzymes: an H3K27me2/3-demethylase and an H3K4me2-methyltransferase.^30,34 Surprisingly, these interactions localized to two physically separable domains within the larger T-box DNA-binding domain that are highly conserved between T-box factors. Consistent with the conservation in the interaction interface, the ability to associate with these epigenetic modifying activities is a common function for the T-box family.³⁰ Importantly, several mutations associated with human genetic diseases localize to one of these domains, suggesting the essential nature of this activity in T-box-mediated developmental transcription.^{23–26,29,30} Although the recruitment of H3K27-demethylase and H3K4-methyltransferase activity is required for optimal transcriptional potential at a subset of T-bet target genes, there appears to be a differential requirement for these activities at specific target promoters. It will be interesting to determine the logic behind the differential utilization of these functions and whether there is a group of promoters where the histone-modifying activities are not required in any circumstances.

Post-translational modifications

Few studies have focused on identifying the covalent modifications of T-bet or other T-box factors. However, there are at least two phosphorylation events that can modulate the ability of T-bet to interact with partner proteins.^42,44 For instance, the interaction between T-bet and GATA-3, which represses the activity of both proteins, requires the phosphorylation of residues in the C-terminus of T-bet.^42,45 It will be informative to determine if similar modification requirements exist in other T-box factors that are known to interact with GATA family members. It is also possible that the modification status will predict whether the partner proteins will work in a co-operative or antagonistic fashion in a particular cellular setting.

It is interesting to note that T-bet contains several predicted sites for modifications including phosphorylation, methylation, sumoylation and sulphanation. As these types of modifications regulate the activity of many other proteins, it seems likely that, if present in T-box proteins, they could be used in similar ways. It is also possible that the modification state of the partner protein itself may be a point of regulation. In particular, within regions of the T-box that are required for the interaction with the H3K4-methyltransferase and H3K27-demethylase activities, there are two conserved sumoylation interaction motifs.³⁰ This suggests the possibility that the sumoylation state of the histone-modifying complexes may influence their ability to interact with T-bet or other T-box proteins. Utilization of specific post-translational modifications to modulate key protein–protein interactions may provide a means to finely control T-box factor functional activity in response to cellular signalling pathways in Th1 cells as well as other developmental transitions.

GATA-3 and the GATA family

Overview

In mammals, the GATA family of zinc finger transcription factors can be subdivided into two groups. GATA-1, -2 and -3 all play a role in the haematopoietic system, in addition to their function in other tissues.¹⁰ In contrast, GATA-4, -5 and -6 are not involved in blood cell development but instead are required in multiple mesoderm- and endoderm-derived tissues, including the heart and liver.¹⁰ Although the members of each subgroup are more closely related to each other, all GATA proteins are defined by two highly conserved zinc fingers that are involved in DNA binding. The zinc finger DNA-binding domain allows for the interaction with a conserved site in the DNA, WGATAR, from which the family name is derived.¹⁰ Outside of the DNA-binding domain, there is little similarity between factors, although there are some key residues surrounding the zinc fingers that have been conserved through evolution.⁴⁶ Therefore, like the T-box family, it is likely that there are both common and distinct functions for GATA factors outside of DNA binding. This must be the case because they are non-redundant in cellular differentiation, but are expressed in partially overlapping patterns during development.

Like the T-box proteins, the association between GATA mutants and developmental pathologies illustrates the essential nature of GATA factor activity (Table 1). For example, germline deletion of GATA-3 results in early embryonic death.⁴⁷ Additionally, if GATA-3 is conditionally deleted only in the immune system, Th2 cells fail to develop.^4,48 GATA-4-deficient mice die in utero when the heart fails to develop properly and human mutations in GATA-4 have been associated with defects in atrial and ventricular septation.^36,49 GATA-1 is absolutely required for erythroid cell development and mice deficient in this factor also die early in embryogenesis.⁵⁰ Taking these findings together, it is clear that GATA factors are important in many developmental systems, including Th2 development, but the precise mechanisms utilized to regulate cellular programmes are not yet fully understood.

Identification of target genes

Similar to the T-box family, the majority of known GATA target genes have been identified in hypothesis-driven experiments. As a result of these studies, GATA-3 is known to regulate genes encoding the signature cytokines of Th2 cells, interleukin-4 (IL-4), IL-5 and IL-13.⁴ In addition, the cyclin D2 gene and Cdk4 were identified as direct GATA-4 target genes in a recent study using cells from a conditional GATA-4 knockout.⁴⁹ Given the conserved nature of GATA-binding sites, it is possible that other family members also will have the ability to regulate genes which are important for the G1/S transition. Indeed, GATA-1 has been shown to be directly required for the expression of related cell cycle genes.⁵¹ Regulating genes involved in the cell cycle makes particular sense when thinking about Th2 development. As Th2 cells differentiate, they must also expand to eliminate the invading pathogen. Controlling the genes that are important for cell cycle progression with the same transcription factor that defines the effector characteristics of the cell would be an efficient way to achieve this requirement.

There have also been microarray experiments performed in GATA-1 null cell lines that have identified target genes on a global level during erythroid development.⁵⁰ Of the genes activated by GATA-1, one group does not require new protein synthesis, whereas a delayed subset, including the β-globin gene, might require the production of cofactors. Unfortunately, the requirement for GATA factors in early embryonic development has made it difficult to perform these unbiased microarray-based experiments to look for target genes of the GATA family by comparing wild-type versus deficient cells. Here, studies of conditional knockouts have allowed for the development of the organism so that later tissues, including Th2 cells in the immune system, can be examined to identify the genes that are regulated in a specific cellular context. The information gained in conditional knockout studies is useful to form hypotheses for putative target genes in other organs as well.

Protein–protein interactions

The GATA factors have a number of well-characterized protein-binding partners. Perhaps the best studied are the physical and functional interactions with the friend of GATA (FOG) proteins. These coactivators were first identified in a yeast two-hybrid screen for GATA-1 interacting partners.²² Their role is complex because they have been shown to either enhance or inhibit GATA activity depending on the developmental context. For example, FOG-1 enhances GATA-1 activity in erythrocyte development, whereas overexpression of FOG-1 inhibits GATA-1 in eosinophils.²² In this case, it appears that a cofactor provides cell type-specific control of transcriptional activity. Interestingly, GATA factors also have been shown to interact with each other to regulate gene expression. In cardiomyocyte differentiation, GATA-4 and GATA-6 bind to each other through their zinc fingers and C-termini.⁴⁶ This interaction is important for the activation of cardiac target genes such as Anf and ET-1. Significantly, T-box factors also activate Anf, suggesting that T-box and GATA proteins may coregulate a subset of genes in cardiomyocytes. As noted in the T-box section, T-bet and GATA-3 physically associate. This, together with the fact that some target promoters have both T-box and GATA binding sites, suggests that these two families coregulate gene activity in several systems, including T helper cell differentiation.

Like the T-box family, GATA proteins also interact with enzymes that have histone-modifying activity. Several GATA factors interact with the acetyltransferase p300, which modifies lysine residues in histones as well as other proteins. In fact, p300 acetylates GATA-1, -2, -3 and -4 at conserved residues just outside the zinc fingers.^46,52 Although not currently tested, it is likely that GATA-5 and -6 also will be modified by this enzyme due to amino acid conservation at the site of acetylation. It is not yet clear whether subsets of target genes require the acetylation of their promoter or if the acetylation of the GATA factor itself is more important for gene activation. As with all proteins that catalyse post-translational modifications, if the substrate is present, it will probably be modified. The currently unanswered question is: which is the rate-limiting modification the enzyme was recruited to perform? It is possible that the acetylation of histones will be needed at some promoters, but not at others. This is an important point, considering the wealth of data generated in genome-wide histone modification analyses. This information needs to be sifted through to determine which of the characterized marks are functionally relevant versus those that are coincident with the recruitment of the enzymatic machinery for a different purpose.

Post-translational modifications

As previously mentioned, several GATA factors are acetylated at conserved lysine residues by the acetyltransferase p300. To date, all GATA factors examined need to be modified for optimal activity. Importantly for Th2 differentiation, GATA-3 cannot efficiently activate target genes unless it is acetylated, with the loss of acetylation resulting in decreased T-cell survival and abnormal homing.⁵³ Mechanisticly, GATA-1 must be acetylated in order to occupy its DNA-binding site within a chromatin template.⁵⁴ A GATA-4 mutant protein that cannot be acetylated acts as a dominant negative, suggesting the requirement for acetylation in its normal function.⁴⁶ Collectively, the data obtained in studies examining several GATA factors emphasize the importance of acetylation in regulating their function, but it is currently unclear whether acetylation causes differential effects on family member activity or whether some of the perceived mechanistic differences are attributable to the types of assays used to characterize the mutant proteins in each study.

In addition to acetylation, phosphorylation also has been shown to regulate the activity of at least one GATA factor. GATA-1 is activated when it is phosphorylated by AKT in erythroid precursors.⁵² Mutation of the phosphoacceptor site results in a dominant negative protein that blocks cellular maturation in the fetal liver. The AKT recognition site, located directly downstream of the C-terminal zinc finger, is conserved in GATA-2 and -3, but not GATA-4, -5 and -6.⁵² This suggests that signalling through AKT will probably regulate GATA factors involved in haematopoietic development, but not the other subfamily members. It will be important to examine these possibilities, especially when the goal is to discover ways to pharmacologically alter the activity of one factor in a disease state while leaving the activity of the other family members intact.

RORγt and the ROR family

Overview

There are three members of the mammalian ROR family, RORα, RORβ and RORγ, which belong to the steroid hormone receptor superfamily.⁹ Each member of the ROR family produces at least two isoforms through alternative promoter usage and/or exon splicing. These variants differ from each other in tissue expression pattern as well as the genes they activate. For instance, the shorter RORγt isoform is expressed predominantly in developing thymocytes, Th17 cells, and lymphoid tissue inducer cells.⁹ In contrast, RORγ1 is more ubiquitously expressed. The DNA-binding domain for all ROR family members consists of two conserved zinc fingers that recognize ROR response elements with the consensus motif AGGTCA.⁹ Although other nuclear receptors typically bind to DNA as a dimer, the ROR family contains a structural kink that is hypothesized to preclude them from forming such complexes, therefore suggesting a monomeric binding element. The ligand-binding domain for the ROR family is moderately conserved, but it appears that the secondary structure of this domain, which contains 12 alpha helices, is more important than the primary amino acid sequence.⁹

The importance of the ROR family in development has predominantly been studied in mouse models (Table 1). As with the T-box and GATA factors, the absence of individual ROR proteins results in the failure of specific cell lineages to develop normally. RORα deficiency or mutation results in mice that display ataxia and cerebellar atrophy, emphasizing the need for this protein in the development of the cerebellum.⁵⁵ RORγt^−/− mice fail to develop lymph nodes or Peyer’s patches while RORβ^−/− mice are blind as a result of retinal degeneration a few weeks after birth.^56–58 These data indicate that RORγt and RORβ are required for the development of peripheral lymphoid organs and the retina, respectively. RORγt is also required for the development of Th17 cells, which are sometimes called inflammatory T helper cells.² Th17 cells play a role in the clearance of extracellular bacteria and they have been implicated in autoimmune disease models including colitis and experimental autoimmune encephalomyelities (EAE).^59,60 Of all the transcription factors governing effector T helper cell differentiation, the least is known about RORγt. This is not surprising because Th17 cells were only recently identified.⁶¹ The relative lack of knowledge concerning the role of RORγt in T helper cells makes it all the more important to gain insight from the studies performed on its family members in other developmental systems.

Identification of target genes

There have been some studies undertaken to identify ROR target genes, but it is not always clear whether these targets are direct or indirect. Still, the ROR family has been shown to regulate some tissue-specific target genes. In Th17 cells, RORγt is required to activate IL17.² Additionally, Th17 cells produce a related cytokine, IL-17F, as well as IL-22, possibly as a result of the direct or indirect activity of RORγt.⁶² Outside the immune system, the current data suggest that RORα directly activates genes involved in the glutamatergic pathway in the nervous system.^9,63 In addition to tissue-specific gene regulation, ROR factors play a common functional role in the regulation of the circadian rhythm through the activation of a transcription factor, Bmal1, which is important in the molecular feedback loops that control this process.⁹ The individual loss of a given ROR does not completely disrupt target gene oscillation, even in tissues where only one family member is expressed. This suggests that ROR proteins may be co-operating with other transcription factor families. Perhaps most exciting to work in the immune system, RORα directly regulates genes involved in signal-dependent calcium release.⁶³ The common regulation of these target genes could have significant consequences given the importance of calcium signalling in T-cell activation. However, at present, the functional activities that ROR factors utilize to regulate common and tissue-specific targets for the development of individual cell types are not known. In order to reasonably manipulate ROR activity in disease settings, this area of research must become a priority.

Protein–protein interactions

The best characterized protein interaction for the ROR family is with the forkhead transcription factor Foxp3. The physical association between Foxp3 and RORγt inhibits the functional activity of RORγt in in vitro stimulated CD4⁺ T cells.⁶⁴ It has been shown that the second exon of Foxp3 is required for this interaction and the functional repression of RORγt activity.⁶⁴ Although the domain of RORγt required for its association with Foxp3 has not been identified, it is possible that helix 12 of the ligand binding domain will be needed, similar to its requirement for the interaction between RORα and Foxp3.⁶⁵ At present, the direct transcriptional consequence of the association between Foxp3 and RORα is not clear. In luciferase assays, Foxp3 inhibits the ability of RORα to activate a transiently transfected synthetic promoter-reporter construct.⁶⁵ However, in vivo, RORα is required for the repression of the inflammatory genes tumour necrosis factor α (Tnf-α) and Il-6 in macrophages and mast cells.^9,66 Similarly, RORα^−/− CD8⁺ T cells show increased production of Ifn-γ, suggesting that RORα is suppressing effector function or, alternatively, altering a differentiation programme.⁶⁶ It would be unlikely that Foxp3, the key transcription factor for regulatory T cells, would inhibit the activity of RORα if it indeed plays a direct role in the repression of inflammatory target genes. Importantly, RORα has been shown to synergize with RORγt in IL-17 transactivation and the development of Th17 cells.⁶² Therefore, it seems more likely that the interaction between Foxp3 and RORα is playing an antagonistic role in CD4⁺ T-cell fate choices. A better understanding of the direct target genes of the ROR family will help to address the mechanisms that account for these observations.

There is some evidence that ROR family members functionally interact with different proteins at specific target promoters. Interestingly, RORα is required for the recruitment of β-catenin and p300 to the Shh promoter, whereas it recruits Tat interacting protein 60 (Tip60) in addition to β-catenin and p300 to the Pcp4 promoter.⁶³ Currently, the mechanisms by which this occurs and whether the proteins are brought to the promoters through a direct interaction with RORα are unclear. If there is a physical interaction, it will be interesting to determine whether the other ROR family members also interact with these cofactors and whether they are commonly recruited to target genes in the immune system.

Post-translational modifications

In addition to their interaction with coactivators and repressors, the ROR proteins bind several members of the ubiquitin/proteasome machinery.^9,67 The ability of the nuclear receptor superfamily to remodel chromatin and transactivate target genes is precisely controlled by ubiquitination and proteosomal degradation.⁶⁸ For example, when the proteasome is inhibited, there is an accumulation of poly-ubiquitinated RORα, preventing the transactivation of its target genes.⁹ It is also interesting to note that catalyses sumoylation, ubiquitin conjugating enzyme 9 (UBC9), interacts with ROR family members, suggesting that they might be sumoylated.⁹ However, it is not currently known what, if any, functional consequence this association has for ROR activity. Like the T-box proteins, the ROR family contains several predicted modification sites, but to date there have been no studies examining their functional significance. In the future, it will be important to learn the extent to which post-translational modifications modulate ROR activity.

Conclusion

The process by which a cell differentiates from a multipotent progenitor into a defined and committed end-point lineage requires the precise co-ordination of gene expression networks at specific developmental time-points. To understand lineage commitment, we must examine the activity of key developmental transcription factors. Interestingly, the transcription factors that are needed for diverse cell fate decisions belong to a surprisingly small number of protein families. Included in this list are the key transcription factors required for the development of effector T helper cells in the immune system, namely T-bet, GATA-3 and RORγt. The immune system provides a powerful tool for studying developmental transitions in a well-characterized and controlled environment, in this case, the ex vivo differentiation of T helper cells. The conservation among family members, especially in their DNA-binding domains, makes it possible to broaden important discoveries made in immune cells to other developmental transitions. We can also utilize information gained in studies of other systems to understand immune differentiation. The challenge is to integrate the data from diverse developmental systems, looking for commonalities and differences between protein family members, to ultimately be able to rationally manipulate the events that go awry in pathogenic conditions for the benefit of human health.

Disclosure

The authors declare no conflicts of interest.