This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Giguère, V. Search for Related Content PubMed PubMed Citation Articles by Giguère, V.

Orphan Nuclear Receptors: From Gene to Function¹

Molecular Oncology Group, McGill University Health Centre and Departments of Biochemistry, Medicine, and Oncology, McGill University, Montréal, Québec, Canada H3A 1A1

	Abstract

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

 

I. Introduction

II. Nuclear Receptors: General Concepts

A. Anatomy of nuclear receptors

B. Mechanisms of action

III. Orphan Nuclear Receptors

A. Definition

B. Nomenclature

C. Structural and functional diversity

IV. Novel Hormone Response Systems: RXR and Its Heterodimeric Partners

A. RXR: rexinoids

B. PPAR: multiple ligands, multiple functions

C. PXR: pregnanes, xenobiotic compounds, and benzoate derivatives

D. CAR (constitutive androstane receptor): androstanes and phenobarbital

E. LXR: control of cholesterol metabolism by oxysterols

F. FXR: bile acids receptor

V. Orphans in Search of a Home

A. HNF4: diabetes and possible regulation by acyl-coenzyme A (CoA) thioesters

B. FTZ-F1: steroidogenesis and sexual development

C. Rev-Erb: singular members of the superfamily

D. ROR: neuron development and T cell selection

E. TR2: the testis receptors

F. TLX: forebrain development

G. COUP-TF: neurogenesis, angiogenesis, and heart development

H. ERR: placenta development and control of lipid metabolism

I. NGFI-B: hypothalamus-pituitary axis (HPA), T cells, and dopaminergic neurons

J. GCNF: one of a kind

K. DAX-1: adrenal development and sex determination

L. SHP: a promiscuous and inhibitory heterodimeric partner

VI. Concluding Remarks

	I. Introduction

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

 
THE RECOGNITION that the steroid, thyroid, and retinoid receptors constituted only a small subset of a large number of related gene products is around 10 yr old (1 ). The year 1988 witnessed the identification of the first cDNA clones encoding polypeptides with structural features suggestive of cryptic steroid hormone receptors (2 ), and at that time one could identify about a dozen or so distinct nuclear receptor-like proteins. There are now more than 50 identified in various species and, mainly through the various genome-sequencing projects, the number is now increasing at a rapid pace. Because the discovery of all these putative nuclear receptors had not been anticipated by previous physiological studies and therefore not linked with the biological effects of a particular hormone or ligand, these new gene products were referred to as orphan nuclear receptors. That the activity of orphan nuclear receptors could be potentially regulated by natural ligands led to the tantalizing suggestion that new hormone response systems remained to be discovered (2 ). Interest in orphan nuclear receptor research was also stimulated by the knowledge that classic members of the superfamily of nuclear receptors and their ligands play crucial roles in development, homeostasis, and disease. The possibility that nuclear receptors’ activity might be regulated by the direct action of natural and synthetic compounds makes orphan receptors good targets for drug discovery. Therefore, the existence of a large number of potential new receptors offers the exciting opportunity to develop novel therapeutic agents, even in the absence of known natural ligands. Recent advances in the field have shown that such drugs could be used to treat a variety of illnesses, including diabetes, lipid disorders, and cancer. For these reasons, the study of orphan nuclear functions has regrouped scientists from a wide variety of fields, and consequently the information being generated on this subject is vast, diversified, and often very confusing. In this review, I will attempt to regroup some of this information in a format accessible to nonspecialists and specialists alike. I will also try to demonstrate how the study of orphan nuclear receptors has revealed new modes of action for nuclear receptors that often challenged previous dogma and highlights the discovery of new hormone response systems as well as regulatory pathways controlling cell fate, organogenesis, and basic metabolic functions.

	II. Nuclear Receptors: General Concepts

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

 
Nuclear receptors provide multicellular organisms with a means to directly control gene expression in response to a wide range of developmental, physiological, and environmental cues. It is now recognized that nuclear receptor activity can be controlled by at least three distinct mechanisms: 1) binding of a small lipophilic ligand by the receptor or its partner in heterodimer complexes; 2) covalent modification, usually in the form of phosphorylation regulated by events at the cellular membrane or during the cell cycle; and 3) protein-protein interactions, generally through contacts with other transcription factors including nuclear receptors themselves. All three mechanisms can either work individually or in concert with each other to modulate a specific signal (Fig. 1). It should also be noted that some nuclear receptors mediate nongenomic effects that are too rapid to involve changes in gene transcription. This subject has been reviewed recently in this journal (3 ).

View larger version (61K): [in this window] [in a new window]   Figure 1. Signaling pathways regulating genomic actions of nuclear receptors. Nuclear receptor activity can be regulated by direct binding of small lipophilic ligands, protein-protein interactions with other transcription factors, or by covalent modification such as phosphorylation after stimulation of cell surface receptors or by cyclin-dependent kinases.

A. Anatomy of nuclear receptors
Nuclear receptors are composed of four independent but interacting functional modules (Fig. 2A). These are the modulator domain, the DNA-binding domain (DBD), the hinge region, and the ligand-binding domain (LBD). For some nuclear receptors, the sequence of the protein extends beyond the LBD at the carboxy-terminal end, but no specific role has been assigned to these additions when present.

View larger version (18K): [in this window] [in a new window]   Figure 2. Anatomy of nuclear receptors and typical gene structure. A, Nuclear receptors are composed of independent functional domains that include the DBD and LBD, the primary functions of which are to recognize specific DNA sequences and ligands, respectively. Nuclear receptors generally possess two transcription activation functions (AF-1 and -2) located at the amino and carboxy termini. The division of nuclear receptors into domains A–F is based on the degree of amino acid sequences conservation between the same receptor in different species. B, Schematic representation of the exon-intron organization of a typical nuclear receptor gene. The modulator domain is usually encoded by one or two exons. Distinct modulator domains can be generated by alternative promoter usage (arrows) and splicing (linked exons). The two zinc finger modules are generally encoded by distinct exons while the hinge and LBD are encoded by 6 to 10 exons. Additional alternative splicing may generate nuclear receptors with modified LBDs.

2. DBD. Nuclear receptors bind DNA as monomers, homodimers, and heterodimers (Fig. 3A) (reviewed in Ref. 22 ). While most heterodimeric complexes contain one of the retinoid X receptors (RXRs) as a common partner (23 ) (see below), alternative heterodimeric interactions between nuclear receptors have been reported and may be of physiological significance (24 25 26 27 28 29 30 31 32 ). Nuclear receptor DNA recognition sites, referred to as hormone response elements (HREs), contain one or two consensus core half-site sequences. For dimeric HREs, the half-sites can be configured as inverted, everted, or direct repeats. Steroid receptors recognize the half-site consensus sequence AGAACA while the estrogen receptors and other nuclear receptors bind to the half-site consensus sequence AGGTCA. For monomeric HREs, a single half-site is preceded by a 5'-flanking A/T-rich sequence. Half-site sequences can deviate quite considerably from the consensus sequences, especially for dimeric HREs in which a single conserved half-site is usually sufficient to confer high-affinity binding to the homo- or heterodimer complexes. Natural HREs rarely contain two perfect consensus half-sites.

View larger version (31K): [in this window] [in a new window]   Figure 3. DNA binding by nuclear receptors. A, Nuclear receptors can bind DNA as monomers, homodimers, and RXR heterodimers. Nuclear receptor binding sites are composed of one or two half-core motifs, generally AGGTCA or a close variant that could be preceded by a 5'-flanking A/T-rich sequence. The half-core motifs and 5'-flanking A/T-rich sequences are recognized by the first zinc module and the CTE, respectively. Half-core sequences in dimeric sites can be arranged as inverted, everted (not shown), or direct repeats. Intra- and intermolecular protein-protein interactions can influence DNA binding specificity and receptor dimerization. B, Schematic structure of a nuclear receptor DBD. The two zinc finger modules as well as the CTE are identified. Residues in black have been shown to make direct contacts with DNA. Closed and open symbols linked to certain residues represent residues that have been shown to mediate dimerization in distinct receptor complexes.

3. The hinge region. This region of the nuclear receptors is also highly variable in both length and primary sequence: as its name indicates, its main function is to serve as a hinge between the DBD and LBD. The hinge has to be very flexible to let the DBD rotate 180° to allow some receptors to bind as dimers to both direct and inverted HREs (22 ). Recent studies have also demonstrated that the hinge region may serve as a docking site for corepressor proteins (43 44 ).

4. LBD. The LBD is a multifunctional domain that mediates ligand binding, dimerization, interaction with heat shock proteins, nuclear localization, and transactivation functions. Although quite variable in primary sequence, nuclear receptor LBDs can be defined by a signature motif overlapping with helix 4 (45 ). In addition, ligand-dependent transactivation is dependent on the presence of a highly conserved motif, referred to as activation function-2 (AF-2), localized at the carboxy-terminal end of the LBD (Fig. 2A). X-ray crystallographic experiments suggest that LBDs have similar structures: they are formed by the folding of 11–13 -helices into three layers that bury the ligand-binding site within the core of the LBD (46 47 48 49 50 51 52 ). Ligand-dependent transactivation involves the recruitment of coactivators (see below), a process in which the AF-2 plays an obligatory role. Comparison of holo- and apo-LBD structures has led to the mouse trap model in which ligand binding induces a conformational change in the LBD allowing coactivators to bind (48 ). In this model, the AF-2 motif folds back against the core LBD upon ligand binding, closing the ligand-binding pocket and forming a novel interface involving residues from the AF-2 itself and at least three other helices (53 ). While transcriptionally competent interfaces are induced by receptor agonists, binding of antagonists to the LBD leads to the formation of a nonfunctional interface preventing interaction between the nuclear receptor and coactivator proteins (49 ).

B. Mechanisms of action
The textbook model of nuclear receptor action is often represented by an inactive cytoplasmic receptor in a complex with heat shock proteins which, upon ligand binding, translocates to the nucleus and activates gene expression. Although this model is valid for some steroid receptors (54 ), most nuclear receptors are constitutively nuclear and often bound to DNA in the absence of their ligand. It is also now widely recognized that in the absence of ligands, many nuclear receptors can act as a strong repressor of gene expression (43 44 55 56 57 58 ). To modulate transcription of their target genes, nuclear receptors interact with coregulatory proteins. Nuclear receptors have been shown to associate with various components of the general transcription machinery, corepressors, coactivators, and the cointegrator CBP (CREB-binding protein)/p300 (reviewed in Refs. 59 60 61 62 63 ). Corepressor proteins may function by recruiting histone deacetylases, an activity that keeps the chromatin in a repressive state (64 65 66 ). Upon ligand binding, the repressor complex dissociates from the receptor, which is then free to interact with the coactivator complex. The receptor-coactivator complex may contain one or more coactivators, including an RNA coactivator referred to as SRA (67 ), p/CAF (p 300/CBP-associated factor), CBP/p300, and other uncharacterized components. SRC-1, p/CAF, and CBP have been shown to possess intrinsic histone acetylase activity leading to a derepression of the chromatin structure (68 69 70 71 72 73 ). Taken together, these results delineate a new model for transcriptional regulation by nuclear receptors that includes three chromatin states: 1) normal chromatin in the absence of receptor that displays basal levels of histone acetylation and transcription; 2) repressive chromatin with deacetylated histones and no transcription in the presence of the unliganded receptor; and 3) active chromatin with high levels of histone acetylation and transcription in the presence of liganded receptor (reviewed in Ref. 74 ). However, chromatin disruption alone is not sufficient for transcriptional activation, indicating that additional interactions between nuclear receptors and the general transcription machinery are required to regulate gene expression (75 ).

Finally, nuclear receptors can also regulate gene transcription via direct interactions with other transcription factors, a process that does not depend on DNA binding by the nuclear receptor (reviewed in Refs. 76 77 ). In particular, the glucocorticoid receptor (GR) has been shown to antagonize AP-1 and nuclear factor-B activities via transcriptional interference (78 79 80 81 82 ). Recent in vivo experiments that used reverse genetics to engineer a mutant mouse carrying a DNA-binding deficient GR have demonstrated that development and survival of mice do not require HRE-mediated gene regulation (83 ). These observations emphasize the multifaceted control of nuclear receptor activities and the independence of each functional domain in carrying out physiological roles.

	III. Orphan Nuclear Receptors

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

 
A. Definition
Classic members of the superfamily of nuclear receptors were originally cloned on the basis that known hormones were transducing their physiological functions through binding to proteins referred to as receptors (1 84 ). This is true for cortisol and aldosterone, estradiol, progesterone, and testosterone, vitamin D, T₃, and, to a certain extent, for atRA, for which the mechanism of action had been postulated to resemble that of steroid hormones (85 ), and for the insect hormone ecdysone (86 ). Therefore, the existence of these receptors was well recognized, and the combined efforts of biochemical and cloning experiments revealed their common structure and mode of action as defined above. Once investigators realized that nuclear receptors shared extensive homology at the amino acid and nucleotide sequence levels, a search for new members was undertaken using low-stringency screening of cDNA libraries with well conserved DBD fragments as probes. This cloning exercise led to two unexpected results. First, individual ligands, such as T₃ and atRA, were shown to regulate development and physiology through multiple receptors. This finding was first exemplified by the characterization of a second receptor for T₃ (87 ) and later by the cloning of three RAR genes, each encoding multiple isoforms (reviewed in Refs. 4 5 ). The recent identification of a second estrogen receptor in various species (11 88 89 ) demonstrates that even for classic ligands, the hunt for new receptors is still very active. Second, the search for new members of the superfamily led to the isolation of multiple cDNAs encoding proteins with structural features found in nuclear receptors. However, since ligands could not be linked to these putative receptors based on structural studies alone, these and other new members of the superfamily identified using various cloning strategies were referred to as "orphan nuclear receptors." For the purpose of this review, orphan nuclear receptors are defined as gene products that embody structural features of nuclear receptors that were identified without any prior knowledge of their association with a putative ligand. Using this definition, orphan nuclear receptors remain in this category even after the subsequent identification of specific ligands.

B. Nomenclature
The element of randomness associated with the cloning of orphan nuclear receptors led to a great diversity in the naming of these new genes. No common nomenclature or even a basic naming scheme was ever followed, and often the same receptors cloned in different species or by different groups were given unrelated names. Recently, a unified nomenclature system for the nuclear receptor superfamily has been adopted (90 ). The nomenclature is based on the well known system used for the cytochrome p450 superfamily. In this system, the gene subfamilies are designated by Arabic numerals, groups by capital letters, and individual genes by a second set of Arabic numerals. Receptor isoforms generated from the same gene by alternative promoter usage or differential splicing are designated by a lowercase letter at the end of the name. The introduction of this nomenclature system is not designed to replace the use of trivial names, but only to clearly identify which nuclear receptor was studied in a particular set of published experiments. The use of this system should be enormously helpful to both nuclear receptor aficionados and nonspecialists in attributing functional properties to each receptor.

A list of known vertebrate orphan nuclear receptors is presented in Table 1. In this table, orphan nuclear receptors are first classified in seven groups (0 to VI) according to the molecular phylogeny analysis performed by V. Laudet (91 ). Each group is divided into families referred to by their most commonly used trivial names, and each family member is identified by a Greek letter. Receptor isoforms generated from a single gene are identified by an Arabic numeral. Each receptor is then identified by its official name, and a list of other known trivial names is also provided. In this review, for simplicity and clarity, orphan nuclear receptors will be referred to by their family names (most commonly used trivial names), and subtypes will be referred to by a Greek letter. Readers are asked to use Table 1 as a guide to relate these names to other trivial names and to the official nomenclature.

View larger version (39K): [in this window] [in a new window]   Figure 4. Schematic representation of vertebrate orphan nuclear receptors. One member of each family is shown. The presence of a ligand-dependent AF-2 domain is indicated in black. Atypical sequences in DAX-1 and SHP are represented by a dotted domain.

All vertebrate orphan nuclear receptors possess a highly recognizable LBD (Fig. 5). The presence of a conserved LBD is often interpreted as a strong indication that all vertebrate orphan nuclear receptors possess the intrinsic ability to bind a specific ligand. However, since the LBD mediates multiple functions (such as dimerization and coactivator interaction), its presence may only be required for these activities, which could be regulated via covalent modifications or protein-protein interactions (see above). Moreover, widely phylogenetically divergent receptors can bind similar ligands, suggesting that the ligand-binding function of nuclear receptors has evolved independently several times during evolution (121 ). This hypothesis implies that a certain number of orphan nuclear receptors may never have acquired the ability to bind ligands. However, this hypothesis would also imply that the LBD possesses the intrinsic ability to bind ligands, and that only a few mutations would be necessary to modify an ordinary transcription factor into a ligand-modulated one. The reverse hypothesis seems more plausible, i.e., the ancestral nuclear receptor was a ligand-dependent transcription factor and that mutations during the course of evolution changed the ligand-binding specificity of novel nuclear receptors generated through gene duplication according to the increasing needs of more complex organisms. Some nuclear receptors may have lost their ligand-binding properties during evolution, but more drastic changes in their primary structures may have been expected, such as the loss of the AF-2 domain involved in ligand-dependent transactivation. While evolutionary studies are useful for stimulating speculative debates, well designed biochemical, molecular, and physiological experiments are more likely to provide answers on the roles and functions of orphan nuclear receptors and their putative ligands.

View larger version (104K): [in this window] [in a new window]   Figure 5. Sequence alignment of vertebrate orphan nuclear receptor LBDs. The putative secondary structure adopted by nuclear receptors is represented by -helices and ß-strands according to crystal structure of PPAR (52 ).

	IV. Novel Hormone Response Systems: RXR and Its Heterodimeric Partners

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

A. RXR: rexinoids
RXR was originally cloned as a result of its homology with the RAR DBD (125 ). Three RXR gene products referred to as RXR, -ß, and - were identified in mammals (125 126 127 128 ), as well as a Drosophila homolog, ultraspiracle (129 130 131 ). RXRs as a family are ubiquitously expressed, although individual RXR genes display unique but overlapping pattern of expression during development and in adult tissues (127 132 133 134 ).

RXR could be activated by supraphysiological doses of atRA, suggesting that the natural ligand for RXR might be a metabolite of atRA (125 ). Starting with this hypothesis, two groups independently identified 9-cis-retinoic acid (9cRA) as the RXR ligand (Fig. 6) (135 136 ). The identification of 9cRA as the natural RXR ligand was the first demonstration of reverse endocrinology, where the discovery of a receptor leads to the identification of a novel hormone. As 9cRA was also found to be a high-affinity ligand for RAR (136 ), several groups began a search for natural and synthetic RXR-specific ligands. Two noncyclic terpenoids, methoprene acid and phytanic acid, were found to bind and activate RXRs in a specific manner (137 138 139 ). Phytanic acid is a natural chlorophyll metabolite present in normal human diet, while methoprene acid is an environmental contaminant. While both ligands have the potential to regulate or disrupt natural RXR responses, the physiological significance of these findings remains to be proven, especially in view of the relative low affinities of these compounds for RXRs. Several synthetic compounds have now also been characterized that bear the characteristics of RXR-selective ligands, including both agonists and antagonists (140 141 142 143 144 ).

View larger version (17K): [in this window] [in a new window]   Figure 6. Representative ligands and activators shown to modulate the activity of RXR and its heterodimeric partners.

Because RXRs play a dual role in nuclear receptor signaling (as receptor for 9cRA and as a heterodimer partner for several nuclear receptors), it has been difficult to assess the precise contribution of RXR-selective pathways in developmental and physiological processes. Two lines of evidence point to the fact that RXR could act as a specific receptor in its own right. First, 9cRA binding to RXRs promotes the formation of RXR homodimers (159 ), demonstrating that RXRs can function independently of other signaling pathways when bound to a DR-1 HRE (160 ). Second, transgenic mice expressing a chimeric RXR protein in which the RXR LBD was fused to the DBD of the yeast activator Gal4, together with a ß-galactosidase reporter gene driven by Gal4 upstream-activated sequences, showed a receptor-specific activation pattern in the developing spinal cord consistent with a role for endogenous RXR ligands in vivo (161 ). This technique offers the potential to investigate the activation pattern of any nuclear receptors and should be particularly useful in studying orphan nuclear receptor functions.

Genetic ablation experiments have revealed that RXR plays a primary role in placenta, heart, and eye morphogenesis (162 163 164 165 166 167 168 169 170 ). The putative functions of RXR in adult animals are unknown due to the embryonic lethal phenotype. RXRß mutant mice have abnormal spermatogenesis (171 ), while RXR null mice are apparently normal (172 ). While generation of RXR and RAR compound mutants have clearly demonstrated that the RXR/RAR heterodimer complex transduces the retinoid signal for a number of RA-dependent processes during development (162 173 174 ), surprisingly, there is no genetic evidence yet available indicating that RXRs actively participate in other hormone response pathways in vivo.

B. PPAR: multiple ligands, multiple functions
Three PPAR genes generating a number of isoforms have been identified in mammals: PPAR, -ß, and - (Table 1) (reviewed in Ref. 175 ). The members of the PPAR family have been cloned by techniques including screening of a cDNA library with a mixture of oligonucleotides directed against a conserved region of the DBD, low-stringency screening, expression library screening using radiolabeled HREs, and DNA affinity purification and microsequencing of nuclear proteins. The primary sequences of PPAR, -ß, and - are more divergent than members of other families, reflecting a rapid evolution from the ancestral PPAR gene. Each PPAR gene displays a distinct expression pattern during development and in adult animals. PPAR is highly expressed in heart, liver, kidney, intestine, and brown fat, tissues that demonstrate high rates of fatty acid ß oxidation (176 177 ). Hepatic PPAR expression levels have been observed to vary widely in individual animals (177 ), possibly due to hormonal modulation of PPAR expression by glucocorticoids (178 ), physical stress (179 ), or changes in serum insulin levels (180 ). PPARß is more widely expressed in adult tissues: high levels of PPARß transcripts are detected in the brain, kidney, small intestine, and Sertoli cells (177 181 ). Interestingly, PPAR isoforms are expressed in a tissue-specific fashion: PPAR1 transcripts are abundantly expressed in the spleen, intestine, and white adipose tissue (177 ), while the PPAR2 isoform is preferentially expressed in white and brown fat.

PPARs bind to DR1 HREs as a heterodimer with RXR (147 182 183 ). However, analysis of natural PPREs has clearly shown that nucleotides that are present 5' of the two consensus half-sites regulate the efficiency with which specific PPAR isoforms recognize PPREs (120 184 185 ). PPREs have been identified in genes controlling all aspects of carbohydrate and lipid metabolism (147 182 183 184 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 ). In addition to directly binding to their cognate response element, it has been suggested that PPARs may regulate gene expression by forming heterodimers with other nuclear receptors such as T₃Rß and liver X receptor (LXR) (27 212 213 ).

The initial characterization of PPAR led to the observation that peroxisome proliferators, a group of structurally unrelated compounds that cause proliferation of hepatic peroxisomes, liver hyperplasia, and hepatic malignancies in rodents (reviewed in Refs. 175 203 ), could stimulate its activity when employed at pharmacological doses (176 ). A search for more potent and natural PPAR activators first demonstrated that fatty acids could activate PPAR (214 ). PPAR activity was subsequently shown to be induced by eicosanoids (215 ), carbaprostacyclin (216 ), nonsteroidal anti-inflammatory drugs (NSAIDs) (217 ), and leukotriene ß₄ (LTB₄) (218 ). Although PPARß and - can be activated by common PPAR ligands such as docosahexenoic acid and certain prostaglandins (215 ), PPAR was shown to specifically bind to thiazolidinediones (TZDs), a class of antidiabetic drugs (219 220 ). Other PPAR ligands include the natural prostaglandin metabolite 15-deoxy-^12,14-prostaglandin J₂ (PGJ2) (219 221 ), polyunsaturated fatty acids (222 ), and NSAIDs such as ibuprofren (223 ). Specific synthetic PPARß ligands have been identified by screening biased chemical libraries (224 ); however, a natural high-affinity PPARß ligand has yet to be characterized. Representative PPAR ligands and activators are shown in Fig. 6. In addition to direct activation by PPAR ligands, PPAR-RXR heterodimers can be activated by RXR-specific ligands. Transient transfection studies showed that maximal activation of a PPRE-containing gene promoter was achieved by simultaneous treatment with the synthetic PPAR ligand WY14,643 and 9cRA (225 ). As mentioned above, this synergism is also observed in vivo, where the efficacy of TZDs in reducing fasting hyperglycemia and hypertriglyceridemia in db/db obese mice is potentiated by dual treatment with rexinoids (157 ). Finally, PPAR isoforms display isoform-specific transactivation potential due to their distinct ligand-independent AF-1 domain (226 227 ). In particular, the PPAR2 AF-1 domain contains a consensus MAPK site, and phosphorylation of that site after stimulation of the MAPK pathway by epidermal growth factor, and insulin inhibits the adipogenic potential of liganded PPAR2 (226 228 ), presumably because phosphorylation of this site reduces ligand-binding affinity via an intramolecular communication between the AF-1 and the LBD (229 ). The PPAR2 AF-1 domain has also been shown to interact with a coactivator protein termed PGC-2 (230 ). Ectopic expression of PGC-2 in preadipocytes leads to an increase in fat cell differentiation, suggesting that PGC-2 may be a limiting cofactor for the adipogenic action of PPAR2.

The identification of ligands and target genes linked to lipid metabolism greatly facilitated the analysis of PPAR functions. Numerous in vitro studies have provided strong support for a crucial role for PPAR in adipogenesis (reviewed in Refs. 203 231 232 ). In particular, differentiation of fibroblasts into adipocytes is accompanied by increased expression of PPAR (233 ), and overexpression of PPAR2 is sufficient to induce fibroblasts to undergo adipocyte differentiation in the presence of ligands (234 ). It has been shown that activation of PPAR induces cell growth arrest in fibroblast cell lines, which suggests that PPAR may play an important role in cell cycle withdrawal during adipogenesis in vivo (235 ). Although PPAR was first thought to be an adipocyte-specific modulator (236 ), PPAR transcripts have now been detected in many tissues including normal mammary epithelia and breast adenocarcinomas (237 ), colon (238 239 240 241 ), and macrophages (242 243 ). Remarkably, treatment of human breast cancer cells with TZDs led to a series of molecular and morphological changes that are associated with a more differentiated state and to the induction of apoptotic pathways, suggesting that PPAR-induced cell differentiation may offer a novel therapeutic approach to breast tumors (237 244 ). A small clinical trial involving three patients with advanced liposarcoma also suggests that troglitazone (Parke-Davis, Ann Arbor, MI) could be effective at promoting the differentiation of this type of solid tumor (245 ). PPAR ligands have also been demonstrated to slow the growth and induce the differentiation of human colon cancer cells in culture or implanted tumors (246 ). Somatic PPAR mutations that impaired the function of the protein were also found in sporadic colon cancers, suggesting that loss of function of PPAR may contribute to the etiology of human colon cancer (247 ). However, mice genetically susceptible to develop polyps in the colon treated with PPAR ligands show an increased frequency of colon adenocarcinomas (240 241 ). These apparently contradicting results may reflect a distinct role for PPAR in the context of normal colon epithelium (proliferation) and tumor cells (growth arrest) (248 ). These observations raise a warning flag to long-term use of TZDs and beg for additional investigation of the role of PPAR in normal and abnormal colon physiology. PPAR has also recently been indirectly implicated in the regulation of monocyte functions (reviewed in Ref. 249 ). At relatively high doses, PPAR ligands were shown to be effective in reducing the levels of inflammatory cytokines and production of nitric oxide by isolated monocytic cells (242 250 ). In addition, PPAR appears to be involved in the maturation of monocytes along the macrophage lineage, more specifically in the conversion of monocytes to foam cells, which can be induced upon exposure of monocytes to oxidized LDL. It has been proposed that two oxidized derivatives of linoleic acid present in LDL, 9- and 13-hydroxyoctadecadienoic acid (9- and 13-HODE), can act as PPAR ligands once internalized into foam cells by oxidized LDL receptor-mediated endocytosis (243 251 ). PPAR activation by 9- and 13-HODE then leads to foam cell maturation and directly enhances the expression of the CD36 lipoprotein scavenger receptor gene promoter, resulting in increased macrophage LDL and oxidized LDL uptake, which promotes cholesterol deposition in atherosclerotic plaques. Again, both findings have direct implications for use of TZDs as therapeutic agents. On one hand, TZDs, which are well tolerated in patients with noninsulin-dependent diabetes mellitus, could replace the less well tolerated NSAIDs to treat inflammatory diseases. In contrast, prolonged use of TZDs could accelerate the formation of atherosclerotic plaques and increase cardiovascular diseases.

The study of PPAR null mice have revealed three important functions for PPAR in vivo (198 ). First, it has been demonstrated that PPAR is an essential mediator of the hepatic response to peroxisomal proliferators such as Wy-14,643, clofibrate, and DHEA-S (198 252 ). However, PPAR is not essential for peroxisome biogenesis as normal numbers of hepatic peroxisomes are present in PPAR^-/- mice. Second, PPAR appears to orchestrate the expression of genes encoding mitochondrial, peroxisomal, and cytochrome P450 enzymes involved in cellular fatty acid utilization in response to changes in intracellular levels of fatty acid metabolites (209 253 254 ). Third, PPAR null mice show a prolonged inflammatory response when challenged by its natural ligand, LTB₄ (218 ). It has been proposed that the prolonged response to LTB₄ is due to the disruption of the normal feedback mechanism controlling the degradation of this chemotactic inflammatory agent, implying that liganded PPAR regulates transcription of genes involved in this catabolic pathway. Finally, PPAR ligands have been shown to inhibit the inflammatory response of aortic smooth muscle cells in vivo, which participate in plaque formation and postangioplasty restenosis (255 ). Thus, in contrast to PPAR ligands, activators of PPAR may have beneficial vascular effects in atherosclerosis.

While the role of the widely expressed PPARß remains elusive, studies using cyclooxygenase-2 (COX2) null mice suggest that the essential role played by the COX2-derived prostacyclin PGI₂ in implantation and decidualization is transduced by this receptor in the uterus (256 ). PPARß, RXR, COX2, and PGI synthase are coexpressed in stromal cells surrounding the implanting blastocysts, and PPARß agonists (PGI₂, carbaprotacyclin, and L, 165,041) that can specifically promote PPARß/RXR heterodimerization and transactivate the receptor complex in vitro restore implantation and decidualization in COX2 null mice. Once again, the potential role of PPARß ligands in implantation suggests that the development of PPAR-directed drugs should be vigilantly monitored to avoid unwanted side effects.

C. PXR: pregnanes, xenobiotic compounds, and benzoate derivatives
The characterization of PXR may well represent the first identification of a steroid hormone-based response system in many decades. Murine PXR was identified through a computer search of expressed sequence tags (ESTs) derived from a liver cDNA library (257 ). PXR is predominantly expressed in the liver and intestine of embryos and adult animals and is most closely related to the VDR at the structural and amino acid sequence levels. Using a Gal4-PXR chimeric protein to perform an initial search for PXR activators, Kliewer et al. (257 ) found that synthetic pregnanes (C21 steroids) and both glucocorticoid agonists and antagonists were potent inducers of PXR activity. Interestingly, some of these compounds, in particular dexamethasone and pregnenolone 16-carbonitrile (PCN), were well known to induce the expression of the cytochrome p450 CYP3A gene in rodent liver and intestine and cultured hepatocytes (see Refs. 257 258 ). CYP3A is involved in the hydroxylation of steroid hormones as well as various toxic xenobiotics. Induction of CYP3A and other members of the rodent p450 3A family is believed to confer protection against drugs and toxic xenobiotics by increasing their catabolism (259 ). Like its close relative VDR, PXR was found to recognize DR-3 HREs as a heterodimer with RXR, and a DR-3 motif previously demonstrated to be responsible for the activation of the CYP3A promoter by dexamethasone and PCN (260 261 262 ) was also found to mediate PXR activation of the CYP3A promoter by these compounds (257 ). Interestingly, mouse and human PXR display important differences in their activation profile by certain drugs (258 263 ). While PCN acts as a potent inducer of mouse PXR, it has only a weak inductive effect on human PXR. Conversely, rifampicin is a strong activator of human PXR but has very little activity on mouse PXR. These species-specific activation profiles were not entirely unexpected as marked interspecies differences had been observed in the induction of CYP3A genes. The comparative analysis of human and mouse PXR function provide an elegant molecular explanation for these species-specific responses. Taken together, these results provide convincing evidence that PXR is responsible for the induction of CYP3A genes in response to treatments with PCN, dexamethasone, and various xenobiotic agents. However, these are synthetic compounds that must mimic the action of endogenous steroids whose normal physiological role could be to regulate steroid and sterol metabolism in liver and intestine. Based on the finding that PXR is best activated by pregnenolone and its derivatives, Kliewer et al. (257 ) proposed that the natural ligand for PXR is likely to be a pregnane; hence, the name pregnane X receptor (PXR). The broader activation profile and low affinities for human PXR activators has also been interpreted to mean that PXR could function as a steroid and xenobiotic sensor that directly regulates the activity of catabolic p450 enzymes in response to the presence of their substrates (123 263 ). If this hypothesis is correct, a high-affinity ligand for PXR may not be required. Regardless of which hypothesis turns out to be correct, the first direct application to spring from these discoveries is likely to be the development of more rapid and accurate PXR-based assays to screen for the ability of drugs to induce CYP3A genes, an important component of the drug development process (258 ).

The Xenopus PXR ortholog, referred to as BXR (264 ) and xONR1 (265 ), appears to play a completely different role in that organism. BXR is expressed early during Xenopus development, and biochemical purification of transcriptionally competent embryonic extracts tested in a BXR-dependent activation assay led to characterization of endogenous benzoate metabolites as BXR ligands (Fig. 6) (264 ). While the exact role of these compounds in vertebrate development is unknown, their identification as orphan nuclear receptor ligands suggests that this class of molecules may participate in previously unrecognized morphogenetic signaling pathways.

D. CAR (constitutive androstane receptor): androstanes and phenobarbital
Study of CAR function has recently introduced another new concept in nuclear receptor action. CAR was originally identified through screening of a cDNA library with a degenerate oligonucleotide based on a conserved region of the nuclear receptor DBDs (266 ). CAR was found to bind DR-5 HREs as a heterodimer with RXR, sites previously shown to be regulated by RAR-RXR complexes in the presence of retinoids or rexinoids. However, CAR was found to activate reporter genes driven by promoters containing DR-5 HREs (266 267 ) or a complex HRE present in the CYP2B gene (268 ) in the absence of retinoids, rexinoids, or any other exogenously added ligands. Thus, the name CAR was referring to constitutively active receptor. Despite its constitutive behavior, the likelihood that CAR activity could be regulated by a ligand was considered high, mainly on the basis of its association with RXR. The search for a ligand revealed that CAR is, in fact, a steroid receptor for androstenol and androstanol (Fig. 6), but contrary to previous dogma concerning steroid receptor action, these ligands switch the activity of CAR off instead of on (269 ). A biochemical analysis of CAR function suggests that the two steroids act as inverse agonists, not antagonists, since addition of these ligands induces the dissociation between CAR and a coactivator protein in vitro. A legitimate question to ask is whether androstenol and androstanol are the best ligands for CAR. The affinities of both compounds for CAR (>400 nM) are well below normal physiological levels found in the plasma. Nonetheless, this discovery should stimulate studies on this previously unrecognized androstane-related signaling pathway and provide a new model to investigate the molecular and structural mechanisms underlying nuclear receptor activation and repression.

While the constitutive activity of CAR can be suppressed by androstanes, various phenobarbital-type inducers have been shown to reverse the negative effect of androstanes on the human cytochrome P450 (CYP) 2B6 promoter (270 ). Phenobarbital is the prototype for xenochemicals that induce CYP2B genes. This observation suggests that CAR, together with PXR and PPAR, may participate in a nuclear receptor-based regulatory pathway controlling the expression of CYP genes in response to exogenous xenochemicals and endogenous compounds such as steroids and lipids.

E. LXR: control of cholesterol metabolism by oxysterols
LXR was so named based on its initial isolation from a human liver cDNA library and the observation that its expression is enriched in that tissue (148 ). LXR is also expressed at significant levels in other organs such as the intestine, kidney, and spleen (148 271 ). A second member of the family, LXRß, is ubiquitously expressed (272 273 274 275 ). Both LXR and -ß recognize DR-4 HREs as heterodimers with RXR (148 271 272 274 ). LXR binding sites, referred to as LXREs, preferentially contain nonconsensus half-sites (AGTTCA). LXR has also been found to heterodimerize with PPAR: this interaction does not lead to the formation of a transcriptionally active complex and has not yet been shown to be physiologically relevant (213 ). The LXR-RXR complex belongs to the class of permissive heterodimers as the RXR ligand 9cRA was found to be a potent activator of LXR transcriptional activity (148 ). Interestingly, RXR was shown to occupy the 5'-half-site on the LXRE, a position that does not allow ligand binding in other heterodimeric complexes. This suggests that RXR-ligand activation potential is not exclusively dictated by receptor binding polarity on DNA (42 ) but is rather dependent on the RXR partner and primary sequence of the HRE (150 ). Furthermore, activation of the LXR-RXR complex by 9cRA requires the LXR but not the RXR AF-2 domain, suggesting that ligand binding by one receptor induces conformational changes in its partner that lead to transcriptional activation (150 276 ). Similar conclusions were reached in studies using the synthetic rexinoid LG100754, which can activate the nonpermissive RAR-RXR heterodimer via the unliganded RAR (277 ). This phenomenon was referred to as the phantom ligand effect.

The search for LXR ligands led to the discovery that endogenous oxysterols are potent and selective LXR activators (149 ). The most active compounds identified were 22(R)- and 24(S)-hydroxycholesterol, 24(S),25-epoxycholesterol, and 7-hydroxycholesterol (Fig. 6) (149 223 278 ). Oxysterols are oxidized derivatives of cholesterol that serve as intermediary substrates in the rate-limiting steps of steroid hormone and bile acid synthesis (279 ). One prediction from these findings is that liganded LXR could act as a sensor of cholesterol and regulate its metabolism. The phenotype of mice carrying a targeted null mutation in the LXR gene confirmed this hypothesis (280 ). The absence of LXR results in a block of cholesterol catabolism, leading to accumulation of hepatic cholesterol in mice fed a high cholesterol diet (2%) associated with mysregulation of CYP7A, which encodes a key regulatory step involved in bile acid synthesis. LXR mutant mice also display abnormal fatty acid synthesis. A role for LXR as a cholesterol sensor is strongly supported by the observation that LXR knock-out mice do not up-regulate bile acid synthesis or diminish their cholesterol uptake in response to high cholesterol levels. However, since the phenotype is observed only when the mice are exposed to very high levels of cholesterol in their diet, the role of LXR under normal physiological conditions remains to be elucidated. Nonetheless, these discoveries will certainly lead to the search for possible mutations in the LXR gene in patients with defects in cholesterol metabolism as well as the development of therapeutic agents targeting the oxysterol response pathway.

F. FXR: bile acids receptor
FXR is most closely related to the Drosophila EcR and binds to EcRE (IR-1) and DR-4 HREs in a complex with RXR (274 281 ). Transactivation assays have shown that rat FXR can be activated by high concentrations of farnesol (281 ). Farnesol is an isoprene intermediate in the mevalonate biosynthetic pathway and most likely activates FXR via its conversion into a higher affinity derivative (Fig. 6). However, the activity of the highly homologous mouse FXR (RIP14) is not induced by farnesol but rather can be stimulated by atRA and the synthetic retinoid TTNPB (282 ). As observed for the action of farnesol on rat FXR, high concentrations of atRA are required to activate mouse FXR, and no evidence of direct binding by atRA and TTNPB was obtained, again suggesting that FXR could serve as a receptor for an unknown metabolite of these activators. Because farnesol and retinoids share common metabolic pathways, it is expected that the activating compound would be a retinoid metabolite.

More recently, however, FXR was shown to be a receptor for bile acids (283 284 285 ). The combined results of three groups demonstrate that several bile acids are potent inducers of FXR transcriptional activity and promote FXR interaction with a coactivator at physiological concentrations. Bile acid activated FXR was also shown to enhance transcription from the intestinal bile acid binding protein (IBABP) promoter and inhibit transcription from the CYP7A promoter, most likely via antagonism of LXR action. Taken together, these results suggest that FXR is a general regulator of bile acid metabolism, acting both at the level of the liver through suppression of CYP7A to reduce synthesis and at the level of the intestine through activation of IBABP to increase recycling of bile acids (reviewed in Ref. 286 ).

	V. Orphans in Search of a Home

Top Abstract I. Introduction II. Nuclear Receptors: General... III. Orphan Nuclear Receptors IV. Novel Hormone Response... V. Orphans in Search... VI. Concluding Remarks References

 
While functional studies of nuclear receptors in the absence of ligands are intrinsically more limited in scope and pharmacological relevance, extensive usage of modern molecular, biochemical, and genetic tools have allowed investigators to obtain a first glance of the biological functions associated with many unliganded orphan nuclear receptors. Below is a brief summary of current knowledge on the molecular mechanisms and functions associated with these putative receptors.

A. HNF4: diabetes and possible regulation by acyl-coenzyme A (CoA) thioesters
HNF4 (hepatocyte nuclear factor 4) represents one of two orphan nuclear receptors not associated with RXR for which an association with a putative ligand has been proposed. HNF4 was initially identified as a transcription factor required for liver-specific gene expression (287 ). To date, three genes encoding HNF4 subtypes have been identified in vertebrates: two in human and rodents (HNF4 and -) (287 288 289 ) and one in Xenopus (HNF4ß) (290 ). Each gene product differs significantly in its expression pattern and transactivation potential (288 289 291 ). HNF4 is expressed at high levels in liver, kidney, intestine, and pancreas and at low levels in the testis (287 291 292 ). HNF4 transcripts are not expressed in liver but can be found at low levels in the kidney, intestine, and pancreas (291 ). During development, HNF4 is expressed in primary endoderm at 4.5 days postcoitus (d.p.c.) and in visceral endoderm between 5.5 d.p.c. and 8.5 d.p.c. (293 ). Hepatic expression of HNF4 is detected in the liver primordia by 8.5 d.p.c. and during all subsequent stages of development. HNF4 expression in other tissues begins at 9.5 d.p.c. in the gut and at 10.5 d.p.c. in the developing pancreas and the mesonephric tubules (293 294 ). HNF4 subtypes bind as homodimers to DR-1 HREs (295 ) and regulate the expression of genes involved in cholesterol, and xenobiotic and amino acid metabolism, as well as all aspects of carbohydrate and lipid metabolism and various liver specific-genes (201 292 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 ).

HNF4 is a constitutive inducer of gene expression that interacts with steroid hormone coactivators and p300 (323 ), suggesting that its activity may be regulated by an endogenous ligand present in most cell types (287 ). Recently, long-chain fatty acyl-CoA thioesters have been shown to modulate gene activation by binding directly to HNF4 (324 ). Long-chain fatty acyl-CoA thioesters are amphiphilic molecules that play important roles in the regulation of energy metabolism through direct interaction with a variety of cellular enzymes (reviewed in Ref. 325 ). High intracellular fatty acyl-CoA concentrations, which sometime result from prolonged fasting or diabetes mellitus, have been shown to inhibit the activity of glycolytic enzymes resulting in fatty acid oxidation replacing glycolysis as the primary cellular energy source (325 326 ). The observation that long-chain fatty acyl-CoA thioesters could also regulate the transcription of genes implicated in these metabolic pathways is of considerable interest. However, long-chain fatty acyl-CoA thioesters displayed marked differences in their ability to regulate HNF4 transcriptional activity: poly- and monounsaturated acyl-CoAs inhibited the constitutive activity of HNF4, while different saturated acyl-CoAs activated HNF4 over its normal activity (palmitoyl-CoA) or inhibited (stearoyl-CoA) it. In addition, treatments with long-chain fatty acyl-CoA thioesters never resulted in more than a 2-fold change in HNF4 activity measured in transcriptional assay (324 ). Since a mixture of long-chain fatty acyl-CoA thioesters may have mutually antagonistic effects on HNF4 function, it may be difficult to demonstrate the importance of these ligands as modulators of HNF4 in vivo. Another caveat to this observation is the previous finding that long-chain fatty acyl-CoA thioesters may also regulate gene expression by interfering with T₃R signaling (327 ). Nonetheless, if confirmed by additional physiological studies, the identification of long-chain fatty acyl-CoA thioesters as natural HNF4 ligands could lead to the development of more specific synthetic HNF4 ligands, which could be used to differentiate the effects of these compounds mediated by HNF4 from those mediated by direct enzyme inhibition.

Insights into HNF4 function in vivo have come from both population and reverse genetics. Recently, a locus linked to maturity-onset diabetes of the young (MODY1) has been associated with mutations in the human HNF4 gene (328 329 330 331 ). The association between MODY1 and HNF4 is probably specific to this form of diabetes, as HNF4 mutations have not been identified so far in patients with other forms of noninsulin-dependent diabetes mellitus (332 ). Furthermore, disruption of the HNF4 binding site in the HNF1 promoter has been identified in an Italian family with MODY, providing an unusual example of patients whose disease state likely results from a combined impairment of HNF4 and HNF1 function (333 ). Of particular interest, HNF4 mutations identified in MODY1 patients can alter the cellular localization of HNF4 or reduce its activity in transcriptional assays, providing strong support for a direct link between reduced HNF4 function and the MODY phenotype (331 334 335 ). On the other hand, gene targeting experiments have not been informative with regard to possible HNF4 functions in liver development or metabolic control in adult animals. Ablation of the Hnf4 gene results in apoptosis of embryonic ectoderm at 6.5 d.p.c, followed by abnormal mesoderm differentiation and embryonic death (336 ). However, Hnf4 ablation in either ES cells or 8.5 d.p.c. embryos is associated with significantly reduced expression of glycolytic enzymes as well as glucose and fatty acid transport proteins (334 ).

B. FTZ-F1 (fushi tarazu-factor 1): steroidogenesis and sexual development
FTZ-F1 was initially characterized as an adrenal gland-specific factor (SF-1) able to bind to conserved AGGTCA consensus motif in the proximal promoter regions of steroid hydroxylases CYP11A, CYP11B2, and CYP21 genes, suggesting that this factor was a member of the nuclear receptor superfamily (337 ). FTZ-F1 was eventually cloned from an adrenal gland cDNA library based on its homology to the RXRß DBD (338 ). The FTZ-F1 gene generates several distinct isoforms through alternative splicing and promoter usage (339 340 341 ). A second closely related gene, FTZ-Fß, has also been identified and may be an important regulator of the -fetoprotein locus and the CYP7A gene (342 343 344 ). During development, FTZ-F1 expression is first detected at 9.0 d.p.c. in the urogenital ridge (345 ). At 10–10.5 dpc, FTZ-F1 expression is associated with the precursors of adrenal steroidogenic tissue and gonadal steroid-producing cells. FTZ-F1 expression is also detected in the ventromedial hypothalamic nucleus (VMH) after 11.5 d.p.c. and in the pituitary gland after 13.5 d.p.c. Pituitary FTZ-F1 expression precedes the onset of FSH expression in gonadotropes, suggesting that FTZ-F1 might either directly regulate FSH gene transcription or regulate gonadotrope differentiation (346 ). In adult mice, FTZ-F1 expression is highest in steroid-secreting cells of the adrenal gland and gonads; lower level expression is present in the spleen and pituitary gonadotropes (340 ). FTZ-F1 is a monomeric receptor that binds to HREs with the consensus sequence TCAAGGTCA (see Ref. 108 ). FTZ-F1 target genes include steroidogenic enzymes (reviewed in Ref. 337 ), Müllerian inhibiting substance (MIS) and its receptor (347 348 ), the pituitary glycoprotein -subunit (349 ), the LH ß-subunit (350 ), the ACTH receptor (351 ), the steroidogenic acute regulatory protein (StAR) (352 353 354 ), oxytocin (355 ), and the orphan nuclear receptor DAX-1 (356 357 ), all supporting an important role for FTZ-F1 in steroid metabolism and sexual differentiation.

FTZ-F1 usually constitutively activates gene expression, and its activity is regulated by phosphorylation: in vitro, protein kinase A-induced phosphorylation of FTZ-F1 reduces the receptor’s DNA-binding affinity, while in vivo, FTZ-F1 phosphorylation may regulate cAMP-dependent gene induction (358 359 360 ). In addition, phosphorylation of the AF-1 domain (located in the hinge region) leads to increased SF-1 transcriptional activity via direct recruitment of the coactivator GRIP-1 (361 ). These observations suggest a way by which peptide hormones such as ACTH could regulate steroid synthesis via ligand-independent activation of SF-1. However, recent studies have shown that certain oxysterols, distinct from those regulating the activity of LXR, increase FTZ-F1 transcriptional activity (362 ). The oxysterol 25-hydroxy-cholesterol is the most efficacious FTZ-F1 inducer (EC₅₀ 5 µM), while 26-hydroxy-cholesterol (EC₅₀ 5 µM), 27-hydroxy-cholesterol (EC₅₀ 5 µM), and 21-hydroxy-pregnenolone (EC₅₀ 11 µM) are less efficient FTZ-F1 activators, and the potent LXR activator 22(R)-hydroxycholesterol does not alter FTZ-F1 activity. Although