Nuclear receptors constitute a superfamily of ligand-dependent transcription factors, that include receptors for steroid hormones, retinoids, thyroid hormone, and eicosanoid metabolites, and orphan receptors for which the ligand has not yet been identified. The retinoid-related orphan receptors alpha, beta and gamma (RORα, β, and γ also referred to as NR1F1 to -3, respectively) comprise a distinct subfamily of nuclear receptors.1 RORs share a modular structure composed of an amino terminal domain, a DNA-binding domain, a hinge- and a ligand binding domain (LBD), that is typical of nuclear receptors. Through alternative splicing and promoter usage each ROR gene generates two or more isoforms (e.g. RORγ1 and γ2) that differ only in their amino terminus. The utilization of different promoters results in a cell type-specific expression of certain ROR isoforms. As a consequence these isoforms regulate different physiological processes and target genes. However, in cells where RORs are co-expressed, their functions may overlap. RORs regulate transcription by binding as a monomer to ROR-response elements (ROREs) consisting of the consensus sequence AGGTCA preceded by a 6 bp A/T rich region in the promoter regulatory region of target genes.2,3 Despite intensive research, to date few ROR target genes have been positively identified (Table 1).

Table 1 Characteristics of members of the ROR subfamily
Full size table

Transcriptional regulation by RORs is mediated via an interaction with co-repressor or co-activator complexes which through their histone deacetylase or acetylase activities, respectively, induce changes in chromatin structure and mediate the interaction of RORs with the basic transcriptional machinery.1,4,5 RORs interact with co-repressors, including nuclear co-repressor (N-CoR), and co-activators, such as steroid receptor co-activator 1 (SRC-1), suggesting that they can function as repressors and activators of transcription.1 Whether the activity of RORs is regulated by ligands or modulated by other signaling pathways is not yet clear. A recent study demonstrated that RORα- and γ-mediated transcriptional activation could be significantly enhanced by Ca2+-dependent calmodulin kinase IV (CaMKIV).6 These observations potentially link regulation of gene expression by RORs to a variety of signaling pathways that cause an increase in intracellular Ca2+. Although the mechanism of this stimulation of transactivation is far from understood, it does not appear to involve direct phosphorylation of RORs by CaMKIV. Alternatively, the increase in ROR-mediated transcriptional activation by CaMKIV may be mediated by phosphorylation and activation of a specific co-activator, inactivation of a co-repressor or increased synthesis of an endogenous ligand.1,6 Activation of the transcription factor MEF2 by CaMKIV has been reported to be linked to phosphorylation to histone deacetylases and their subsequent export to the cytoplasm.7 Such a mechanism may also account for the CaMKIV-enhanced transcriptional activation by RORs.

Recent studies of ROR-deficient mice have implicated RORs in the regulation of a number of biological processes and revealed potential roles for these proteins in several pathologies (Table 1).1 RORα−/− mice and staggerer (sg/sg) mice, a natural mutant strain disrupted in RORα expression, display an ataxic phenotype caused by severe cerebellar neurodegeneration.8,9,10 The latter involves a defective regulation of the differentiation and/or maintenance (possibly increased apoptosis) of cerebellar Purkinje cells. Interestingly, the cerebellar anomalies observed in hypothyroid animals are very similar to those displayed by RORα-deficient mice and involve extensive apoptosis in cerebellar granule cells.11 Whether the cerebellar neurodegeneration in RORα-deficient mice are mediated through a similar mechanism needs further study. RORα-deficient mice also exhibit abnormalities in bone formation and maintenance of bone tissue indicating that RORα functions as a positive regulator of osteogenesis.12 These studies suggest that changes in RORα signaling could play a role in osteoporosis. In addition, RORα−/− mice display an increased susceptibility to arteriosclerosis and show alterations in several immune responses.13,14 Overexpression of RORα has been reported to suppress TNFα-induced expression of COX-2, IL-6 and IL-8 in smooth muscle cells.15 This suppression appears to be linked to the transcriptional activation of the IκBα gene by RORα through an RORE present in the NF-κB promoter. This would subsequently lead to an increase in IκBα and inhibition of the NF-κB signaling pathway. These observations suggest that RORα may function as a negative regulator of inflammatory responses in vivo.

Compared to RORα, which is widely expressed, the expression of RORβ is much more restricted.1 RORβ is expressed in regions of the central nervous system involved in processing sensory information, in the retina, and pineal gland. RORβ−/− mice manifest a duck-like gait and become blind due to a degenerative loss of the retina.16 Whether this degeneration is mediated through apoptosis has yet to be determined. The abnormalities in circadian rhythm displayed by RORβ−/− mice and the differential regulation of RORβ expression in the pineal gland as a function of circadian rhythm suggest a role for RORβ in the control of circadian rhythm.16,17

Recently, studies of mice deficient in the expression of RORγ unveiled several important physiological functions for this orphan receptor.1,18,19 RORγ−/− mice lack all lymph nodes and Peyer's patches suggesting that RORγ is essential for the development of secondary lymphoid organs. RORγ is highly expressed in CD3−CD4+CD45+IL-7Rα+ cells, early progenitors of lymphoid organ development. The lack of lymph node organogenesis in RORγ−/− mice is likely related to the loss of this population of precursor cells. Although the molecular mechanism by which RORγ regulates lymph node development has yet to be elucidated, RORγ appears to be required for differentiation or survival of these progenitor cells.

Another anomaly observed in RORγ−/− mice is associated with thymopoiesis. The thymus of 6–8-week-old RORγ−/− mice contains a greatly reduced number of thymocytes compared to thymus from wild-type mice.18,19 Homeostasis in the thymus is maintained by a delicate balance between proliferation, differentiation, selection, and apoptosis.20 Loss of RORγ expression enhances apoptosis as well as proliferation of distinct thymocyte populations. These observations indicate that loss of RORγ disturbs homeostasis in the thymus. In addition to increased apoptosis in vivo, the rate of apoptosis of cultured RORγ−/− thymocytes is greatly accelerated.18,19 Normal thymocytes undergo ‘spontaneous’ apoptosis over a period of 3–4 days of culture while accelerated apoptosis in RORγ−/− thymocytes occurs within several hours.

Thymopoiesis is a tightly regulated developmental program controlled by multiple signaling pathways involving cell:cell interactions, various cytokines and different transcription factors.20,21,22,23 The major stages in thymopoiesis are defined by the expression of the cell membrane co-receptors CD8 and CD4. At the first stage, CD4−CD8− (double negative, DN) thymocytes undergo a series of distinct changes, including TCRβ rearrangement and extensive proliferation, before they differentiate into CD4+CD8+ (double positive, DP) cells. After TCRα-recombination and expression, DP thymocytes containing TCRs with intermediate affinity for self-peptide-major histocompatibility complexes (MHC) are selected and continue maturation (positive selection) into single positive (SP) CD4+CD8− (T helper) or CD4−CD8+ cells. During this selection process more than 95% of the immature DP thymocytes, cells with low or high affinity for self-antigen-MHC complexes, are eliminated by apoptosis through processes referred to as death by neglect or negative selection, respectively. The precise mechanisms of these two processes of apoptosis are still poorly understood.24,25 In RORγ−/− mice the number of both DP and SP cells are greatly diminished due to increased apoptosis in DP cells.18,19 This enhanced apoptosis together with the fact that the expression of RORγ is largely restricted to DP thymocytes suggest that this population is a target for RORγ regulation. The enhanced cell death in DP thymocytes does not require TCR expression and is not mediated through the Fas-activated signaling pathway since DP thymocytes from RORγ−/− mice that were made defective in TCRα expression or Fas ligand (FasL) function did not differ in their phenotype from those of RORγ−/− mice.19 Moreover, expression of FasL was not altered in RORγ−/− thymocytes. These observations suggest that this increase in apoptosis does not occur through negative selection but appears to mimic death by neglect. Recent studies have shown that in addition to enhancing the survival of DP thymocytes, RORγ promotes the maturation of DN cells, and in particular that of the immature CD8 single positive (ISP) subpopulation,26 into DP thymocytes.27

Analysis of Bcl-2 family members revealed that the expression of Bcl-xL mRNA and protein was almost totally repressed in RORγ−/− thymocytes (Figure 1).18,19 Little change was observed in the expression of proapoptotic Bax and Bak mRNAs. However, Bax protein which in normal thymocytes is largely localized to the cytoplasm, is translocated to the mitochondria in RORγ−/− thymocytes. Since Bax can form an oligomer with Bcl-xL, the repression of Bcl-xL may be directly responsible for the translocation of Bax. Thymocytes from RORγ−/− mice expressing Bcl-xL under the control of the proximal promoter of the lck gene, which directs expression to DP thymocytes, exhibit a lifespan and rate of apoptosis similar to those of wild-type thymocytes. These findings suggest that suppression of Bcl-xL is an essential and early event in the enhanced apoptosis in RORγ−/− thymocytes. This conclusion is in agreement with studies showing that loss of Bcl-xL comprises the survival of DP thymocytes in Bcl-xL−/− mice.28 Interestingly, during thymopoiesis the expression of Bcl-xL, as that of RORγ2, is restricted to DP thymocytes.28,29 The fact that Bcl-xL expression is repressed in RORγ−/− thymocytes, coupled with the observation that RORγ and Bcl-xL are co-expressed in wild-type DP thymocytes suggests that RORγ positively controls Bcl-xL transcription. However, little is known about the transcriptional regulation of the Bcl-xL gene in thymocytes and the mechanism by which RORγ controls Bcl-xL expression. RORγ may regulate Bcl-xL expression directly by interacting with an RORE in the promoter regulatory region of Bcl-xL. Although no RORE-like sequence was identified in the 700 bp promoter flanking region of Bcl-xL, ROREs important for RORγ-mediated regulation could be located further up- or downstream of this region. Alternatively, RORγ may regulate Bcl-xL expression indirectly by regulating the expression of other transcription factors or signaling pathways. Rel/NF-κB transcription factors have been implicated in the up-regulation of Bcl-xL expression and shown to promote survival of lymphocytes.30,31 It has to be established whether changes in the Rel/NF-κB signaling pathway are involved in the regulation of Bcl-xL and survival of DP thymocytes in RORγ−/− mice.

Figure 1
figure 1

RORγ regulates the survival of DP thymocytes. Loss of RORγ expression results in repression of Bcl-xL expression which in turn leads to degradation of p27Kip1 protein by the proteasome and increased CDK2 activity. The repression of Bcl-xL further induces the translocation of Bax to the mitochondria, dissipation of ΔΨm and subsequently the activation of caspases and cell death. CDK2 activity plays an active role in accelerated apoptosis since the CDK2 inhibitor roscovitine inhibits the collapse of ΔΨm and steps further downstream in apoptosis. The apoptosis promoting function of CDK2 may be related to changes in the specificity of its kinase activity due to either its association with a protein other than cyclin A or E (a yet unidentified protein X) or to alterations in its subcellular localization. Ectopic expression of Bcl-xL restores survival of DP thymocytes. Whether accelerated apoptosis in RORγ−/− thymocytes involves release of cytochrome c or is mediated by a cytochrome c-independent pathway has yet to be established. The pancaspase inhibitor ZVAD-fmk has little effect on the dissipation of ΔΨm but inhibits steps further downstream in the execution phase of apoptosis

Full size image

The activity of cyclin-dependent kinase 2 (CDK2) was found to be dramatically enhanced in thymocytes from RORγ-deficient mice.19 Roscovitine, an inhibitor of CDK2 activity, greatly inhibits accelerated apoptosis in RORγ-deficient thymocytes suggesting that CDK2 plays an active role in this induction of apoptosis. Although roscovitine inhibits apoptosis, it does not totally prevent apoptosis suggesting that alternative apoptotic pathways may operate in RORγ−/− thymocytes. The observed reduction in the level of the CDK-inhibitor p27Kip1 in RORγ−/− thymocytes appears to be at least partly related to the increase in CDK2 activity (Figure 1). Exogenous expression of Bcl-xL blocks accelerated apoptosis in RORγ−/− thymocytes and prevents the increase in CDK2 activity and the reduction in p27Kip1 suggesting that Bcl-xL acts upstream of these proteins.19 A similar relationship between Bcl-2/Bcl-xL expression, CDK2 activity and the level of p27Kip1 protein has been observed in thymocytes during dexamethasone-mediated apoptosis.32,33 The regulation of CDK2 activity by Bcl-xL and Bcl-2 in thymocytes appears to be linked at least in part to a delay in the degradation of p27Kip1 by the proteasome.34 Although increased CDK2 activity plays an important role in the induction of apoptosis in thymocytes, how it affects apoptosis is not understood. The apoptosis-promoting role of CDK2 contrasts with the well-established function that nuclear CDK2-cyclin A or -cyclin E complexes have in the regulation of cell cycle progression. Interestingly, in its apoptosis-promoting role in dexamethasone-induced apoptosis, CDK2 was not found in complex with cyclin A or E34 and it has been suggested that CDK2 may be associated with another protein that may change the target specificity of the CDK2 kinase. Alternatively, the apoptotic fate of the cell may depend on the subcellular localization of CDK2 as has recently been demonstrated in ultraviolet irradiation-induced apoptosis in mesangial cells.35 The apoptosis enhancing activity of CDK2 may involve phosphorylation of (a) specific target protein(s) that promote(s) cell death. Since roscovitine inhibits the loss of mitochondrial membrane potential (ΔΨm), elevation of CDK2 activity may influence the function of proteins involved in controlling ΔΨm (Figure 1).

As mentioned above, RORγ−/− DP thymocytes undergo accelerated apoptosis when placed in culture.18,19 This apoptosis is associated with loss of ΔΨm, an increase in caspase activity and annexin V binding, and ultimately internucleosomal degradation. Release of cytochrome c from mitochondria into the cytoplasm is often observed during early stages of apoptosis. Cytochrome c has been demonstrated to form a complex with APAF-1 and pro-caspase 9, referred to as apoptosome, resulting in the activation of caspase 9.36 However, preliminary results appear to suggest that apoptosis in RORγ−/− thymocytes may not involve mitochondrial release of cytochrome c. Future studies have to determine the role of cytochrome c-dependent and -independent mechanisms play in this apoptosis. The broad-spectrum caspase inhibitor ZVAD-FMK inhibited but did not block apoptosis and had little effect on the collapse of ΔΨm suggesting that caspase activation may not be involved in early steps of accelerated apoptosis in RORγ−/− thymocytes. The relationship between the repression of Bcl-xL, Bax translocation, increased CDK2 activity, collapse of ΔΨm, and apoptosis is still controversial and not yet completely understood.37,38,39,40,41 Although the molecular mechanism of accelerated apoptosis in cultured RORγ−/− thymocytes is not yet precisely known, clearly the down-regulation of Bcl-xL and increased CDK2 activity are critical events. However, other signals, e.g. reactive oxygen species or loss of cell–cell interactions, may act synergistically in accelerating apoptosis in cultured RORγ−/− thymocytes.

Although RORγ−/− mice are initially healthy, they are highly susceptible to early onset of thymic lymphoma.42 At 4 months of age more than 50% of the mice have died as a result of T cell lymphoma formation. These tumor cells frequently metastasize to liver and spleen and occasionally to kidney. No other tumor types have been detected in RORγ−/− mice. Although analysis of the immunophenotype of the lymphoblastic cells showed some heterogeneity in CD4/CD8 subpopulations among the different RORγ−/− lymphomas often an increase in the number of (immature?) SP CD8+ cells was observed. All T cell lymphomas stain highly for PCNA and TUNEL indicating a high rate of both proliferation and apoptosis within the tumors. RORγ−/− lymphoma cells also undergo accelerated apoptosis when placed in culture. This apoptosis is, in contrast to that of RORγ−/− thymocytes, resistant to inhibition by roscovitine suggesting differences in the regulation of apoptosis between lymphoblastic cells and RORγ−/− thymocytes. Lymphoma formation does not appear to involve alterations in p53 expression. Although the mechanism that leads to lymphoma development has yet to be determined, it likely relates to the changes in control proliferation and survival of thymocytes in RORγ−/− mice. Increased proliferation of RORγ−/− thymocytes may promote the probability of genetic alterations in tumor suppressor or proto-oncogenes and as a consequence lead to development of lymphoma.42 Future identification of target genes may help in elucidating the molecular mechanisms by which RORγ controls thymic homeostasis and lymphoma development.

Recently, the crystal structure of the LBD of RORβ in complex with stearate, a fortuitous ligand that co-purifies with the protein, and a NR-box peptide of the coactivator SRC-1 has been determined.43 These studies revealed that RORβ and γ contain a relatively large ligand binding pocket of 7663 and 705,3 respectively, while that of RORα is smaller (568).3 Stearate occupies only a small fraction of the ligand binding pocket and does not induce RORβ-mediated transactivation suggesting that it does not function as a true ligand. Based on their studies, Stehlin et al.43 concluded that the activity of RORs is regulated by a ligand-mediated mechanism. However, based on homology modeling of RORα, Harris et al.44 concluded that the RORα is a constitutively active receptor. Although the question of whether the activity of RORs is regulated by endogenous ligands is still controversial, insight into the structure of the binding pockets of RORs may enhance the design of synthetic agonists and antagonists that regulate the activity of RORs. Such synthetic compounds will provide powerful tools in the study of ROR signaling pathways and in the identification of ROR target genes. Although the precise roles for RORs in human disease still have to be determined, recent studies have revealed potential roles for RORs in arteriosclerosis, osteoporosis, cancer, and anomalies in circadian behavior.12,13,14,15,16,42 Discovery of ROR ligands may lead to the development of novel therapeutic strategies for these diseases.