Unraveling the physiological roles of retinoic acid receptor-related orphan receptor α

Retinoic acid receptor-related orphan receptor-α (RORα) belongs to the orphan nuclear receptor (NR) superfamily. The RORα gene, which is located on human chromosome 15q22.2, generates by alternative splicing four human isoforms comprising a conserved DNA-binding domain (DBD), ligand-binding domain (LBD), hinge domain, and distinct N-terminal domain. RORα is expressed in the liver, skin, lung, adipose tissue, brain, and muscle. All four isoforms, RORα1–4, are expressed in humans, whereas only RORα1 and RORα4 are expressed in mice. Expression of these isoforms varies and the isoforms regulate specific functions in several physiological processes^1,2. The RORα recognizes a specific DNA sequence, the ROR response element (RORE), which comprises the AGGTCA consensus motif followed by an A/T-rich region, to regulate target gene transcription via interaction with coactivators and corepressors^3–5. For most NRs, ligand binding is a critical step that triggers a conformational change in the receptor and the possibility of being replaced by a coactivator. Although cognate endogenous ligands of RORα have not been clearly identified, cholesterol and cholesterol intermediates have been suggested. Crystal structure analysis has shown that these ligands reversibly bind to RORα, but whether cholesterol and cholesterol intermediates are relevant physiological ligands remains unclear. In addition to natural ligands, SR3335 and SR1078 constitute developed inverse agonists of RORα. Selective synthetic SR3335 inhibits expression of RORα target genes, such as the gluconeogenic enzymes G6pc and Pepck⁶, and effects of SR1078 on p53 regulation, adipose tissue inflammation, renal ischemia, and allergic asthma have been studied^7–10.

The naturally occurring mutant mouse strain “staggerer” (also known as Rorα^sg/sg)¹¹ was discovered by positional cloning¹²; it lacks up to 90% of the Purkinje cells (PCs) found in wild-type (WT) mice and has a short lifespan. The defective phenotype of staggerer mice is due to deletion of the LBD region of the RORα gene, which causes a shift in the reading frame and prevents adequate translation. These mice exhibit ataxia, motor deficiencies, irregular circadian rhythm, hyperinflammation, osteopenia, atherosclerosis, and muscle atrophy. Moreover, the similar phenotypes of RORα null-mutant (RORα^−/−)¹³ and Rorα^sg/sg indicate that RORα is an essential regulator of PC differentiation and cerebellar function. Furthermore, RORα is involved in various physiological processes, such as the circadian clock, inflammation, tumorigenesis, and metabolic diseases, and it is a tumor suppressor in many types of cancer. Its expression is downregulated in cancers of the skin, colon, prostate, liver, and breast, as well as in melanoma^14–19. Decreased RORα expression correlates with cancer aggressiveness, short overall survival, and poor prognosis. In addition, RORα modulates cyclic expression of BMAL1, a key transcription factor (TF) that regulates circadian rhythms. RORα contributes to the inflammatory response, M1/M2 polarization, and the nuclear factor-κB (NF-κB) pathway^20–22 and regulates lipid and glucose metabolic gene expression, implying its involvement in energy homeostasis.

The regulatory mechanisms of RORα in many physiological processes are bidirectional. Its cis-acting regulation is based on direct recruitment of RORα to the RORE of target gene promoters. However, RORα often affects transcriptional programs in transaction mode through the antagonistic mechanisms of other TFs along with altered coregulator recruitment^15,23. This review summarizes and discusses the current understanding of RORα function and molecular mechanisms in various pathophysiological processes, including cancer, inflammation, cerebellar development, circadian rhythm, and lipid homeostasis. This review also highlights potential targets for treating various diseases.

Oncogenes are abnormally activated during tumor progression, which might involve signal-dependent determination of target gene expression profiles^24–27. Cancer is associated with gene mutations, abnormal gene expression due to gene duplication, unbalanced numbers of chromosomes, or chromatin remodeling. These aberrant genetic changes can increase expression of oncogenes encoding tumor-promoting proteins or decrease that of genes encoding tumor-suppressor proteins^28–30. Gene mutations and altered gene expression are generally connected. For instance, a gene mutation that activates an oncogene functions as a negative regulator of tumor-suppressor gene expression, which leads to a decrease in tumor suppressors. Therefore, understanding gene mutations and changes in expression is important to fully comprehend cancer.

Target genes associated with several diseases, including cancer, are transcriptionally regulated by RORα. Transcriptional regulation via PTM mediated by RORα is affected by various cellular signaling pathways³³. SUMOylation is a PTM in which a small ubiquitin-like modifier (SUMO) is added to target proteins, to change their binding affinity³⁴. SUMO E3 ligase protein inhibitor of activated STAT SUMOylates RORα, whereas SUMO-specific peptidase 2 deSUMOylates it. SUMOylation upregulates RORα transcriptional activity and activates tumor-suppressive target genes³⁵.

The classical ability of RORα to regulate target gene expression is key for activation of tumor-suppressive target genes (Fig. 1)³⁶, although a transcription-independent function of RORα via a noncanonical pathway can potentiate tumor-suppressive roles. These functions involve direct interaction of RORα with molecules in the Wnt/β-catenin pathway, allowing RORα to transrepress Wnt pathways in colorectal cancer²². The Wnt signaling pathway can be classified as canonical Wnt/β-catenin or noncanonical Wnt/Ca²⁺ signaling^37,38. As a corepressor, RORα is responsible for downregulating expression of Wnt/β-catenin target genes. In the noncanonical Wnt pathway, Wnt5a activates protein kinase Cα (PKCα) via phosphorylation and phosphorylation of RORα by activated PKCα can serve as an intersection between the noncanonical and canonical Wnt pathways. Phosphorylated RORα transrepresses β-catenin in the canonical Wnt/β-catenin pathway and downregulates expression of cyclin D1, Axin, c-Jun, and c-Myc¹⁵. In addition, RORα contributes to the intracellular DNA damage response, which is a checkpoint that determines cell fate, progression to a tumor or cell cycle blockade, and apoptosis during the early stage of tumorigenesis³⁹. Under normal conditions, low levels of p53 are maintained by E3 ubiquitin ligases, such as Mdm2, but p53 expression is stabilized during DNA damage responses^40,41. Moreover, expression of RORα is increased by the DNA damage signal and it is then recruited to specific genomic sites by interacting with p53 as a coregulator. The target genes of p53 comprise those with HAUSP (herpes virus-associated ubiquitin-specific protease)-dependent (Hras, Casp1, and Siva) and -independent (Fas, Pmap1, and Mdm2) promoters, with RORα functioning as a coactivator for both types⁴².

The mechanisms through which RORα acts as a canonical TF via its DBD and coregulator differ considerably. The action of RORα can be determined by the types of PTMs involved, such as SUMOylation or phosphorylation. The fact that RORα regulates expression of the master tumor-suppressor p53 and increases p53 protein stability suggests that agonists enhance the effects of RORα and create synergistic effects in two layers of RORα downstream signaling.

Acute and chronic inflammation are major causes of cancer and understanding of their relationships is increasing⁴³. The fundamental mediators between inflammation and cancer include cytokines, noncoding RNAs, and NF-κB TFs⁴⁴, and the effects of these mediators principally cause anti- or protumorigenic inflammatory responses⁴⁵. NF-κB is a key mediator of inflammation and tumorigenesis, and its functional link is supported by experimental evidence explaining the mutual negative regulation of NF-κB and the tumor-suppressor RORα^22,42,46,47.

Metabolic changes have been observed in cancer cells and the tumor microenvironment (TME)⁴⁸, and the mechanisms underlying these changes related to RORα have recently been identified. For example, RORα responds to metabolic changes and influences cholesterol synthesis pathways in cytotoxic T cells by inhibiting NF-κB target genes. The contributions of these activities to cytotoxic CD8+ T cells have been verified using RORα agonists⁴⁹. Pharmacological stimulation of RORα using cholesterol sulfate and the synthetic agonist SR1078 activates downstream target genes of RORα, whereas the synthetic specific antagonist SR3335 functions as a selective inhibitor of RORα. Compared to SR1078 acting as a dual agonist of RORα and RORγ, cholesterol sulfate is a natural selective ligand of RORα. Given that directly targeting RORα via its LBD is a preferred approach for minimizing off-target effects, the selective agonist cholesterol sulfate has therapeutic anticancer effects. Cholesterol sulfate functions to regulate the effector responses of CD8+ T cells by reducing cholesterol esterification in an NF-κB suppression-dependent manner under RORα activation (Fig. 2). Immunosurveillance is an effective antitumor immune response achieved by innate and adaptive components of the immune system; however, cancer can evade immune cytotoxicity. Cytotoxic CD8+ T cells are considered the major effectors of antitumor immunity. The repressive abilities of RORα on NF-κB target genes, such as Acat1/2 and Abca1, induce CD8+ T cells to secrete cytotoxic cytokines, including tumor necrosis factor-α and interferon-γ. Anti-inflammatory drugs have considerable importance for potential use in cancer chemoprevention. Indeed, anti-inflammatory drugs modulate the cytotoxic effector functions of antitumor immunity, and the attenuated inflammation and increased cytotoxic effector function promoted by activated RORα are promising for the development of a versatile therapeutic approach. Furthermore, as a safer and more effective therapy than currently available classical anti-inflammatory agents, this approach may support antitumor immune-potentiating activity because of an increased selective agonistic effect on RORα signaling in the TME.

The circadian clock is a periodic biological system that interconnects the diurnal sleep–wake cycle, metabolism, body temperature control, and immune response. Therefore, circadian rhythm dysfunction correlates with the onset of metabolic diseases, cancer, and neuropsychiatric disorders. The master axis of the circadian rhythm is generated by a specific feedback loop involving BMAL1 expression and CLOCK activation of cryptochrome (CRY1 and CRY2) and period (PER1, PER2, and PER3) genes. BMAL1 and CLOCK TF heterodimer formation is critical for controlled expression as a negative regulatory mechanism. RORα participates in regulating the core loop by transcriptional regulation of the BMAL1 gene with REV-ERBs^54–56 and the dynamic interplay between RORα and REV-ERBs modulates BMAL1 expression and provides a feedback loop for timely expression. RORα works competitively with REV-ERBs by sharing the same DNA response element in the BMAL1 promoter (Fig. 3); expression of BMAL1 is silenced and increased by REV-ERBs and RORα, respectively. Oscillation by these two NRs plays a crucial role in enabling elaborate regulation of BMAL1 expression. In particular, PGC-1α is a coactivator of RORα in BMAL1 regulation; it interacts with RORα via the LXXLL motif to recruit RORE in the proximal BMAL1 promoter along with p300 and GCN5. The function of PGC-1α depends on RORE and the ability of RORα to induce BMAL1 expression is abolished in PGC-1α-null hepatocytes⁵⁷. The CLOCK target gene depends on both RORα and REV-ERBs. Chromatin immunoprecipitation assays have revealed that RORα and REV-ERBs bind to two functional ROREs in the promoter of CLOCK and control BMAL1 expression⁵⁸. Receptor-interacting protein 140 (RIP140) supports RORα transcriptional activity as a positive activator of the BMAL1 gene, whereas REV-ERBs repress RIP140 expression and directly interact with RORα. As RIP140 exerts no effects in the absence of RORα, RIP140 might ROR-dependently participate in the feedback mechanism of the circadian clock and BMAL1 expression⁵⁹.

Overall, circadian clock genes might be associated with the onset and risk of breast, prostate, and colorectal cancers and lung carcinoma^60–62. Cancer predisposition statistically correlates with genetic variations in the circadian pathway via adaptive rank truncated product (ARTP)-based gene expression and pathways⁶⁰. RORα ranks top among genes associated with the risk of breast and lung cancer. Moreover, RORα plays dual roles and has a functional link between circadian rhythms and cancer pathogenesis.

A coordinated transcriptional network regulates expression of lipid and glucose metabolic target genes for energy homeostasis. In particular, changes in clock genes result in abnormal metabolic phenotypes in models in vivo. The roles of RORα in metabolism have been characterized in RORα^−/− mice. Triglyceride (TG) and total plasma cholesterol levels, as well as ApoCIII, APOA1, and APOA2 expression, are lower in Rorα^sg/sg mice than in WT mice. In addition, expression of the cholesterol transporters Abca1 and Abca8/G1 is attenuated in Rorα^sg/sg mice, suggesting that the phenotype of impaired high-density lipoprotein biosynthesis is caused by decreased expression of ABCA1⁶³. Reduced TGs in Rorα^sg/sg mice are due to downregulation of transcriptional regulators responsible for fatty acid biosynthesis and lipogenesis, including SREBP1c, FAS, and ABCG1. Indeed, Rorα^sg/sg mice are resistant to standard high-fat diet (HFD)-induced obesity and weight gain. Although heart and liver tissue weights do not significantly differ between WT and Rorα^sg/sg mice, epididymal and inguinal fat depots are distinct. Moreover, PCK1 expression is reduced in the livers of Rorα^sg/sg mice after a 6 h fast. The gluconeogenic enzyme PCK1 is a rate-determining factor in the adaptive response of the fasting period. Hepatic gluconeogenesis and adipose tissue glyceroneogenesis are positively controlled by RORα and its regulation is mediated by PCK1⁶⁴. Expression of thermogenic genes, such as UCP1, is increased in brown and subcutaneous inguinal adipose compartments of Rorα^sg/sg mice, suggesting increased thermal control and cold tolerance in these mice⁶⁵. This process is associated with increased oxygen consumption and energy expenditure.

Circadian and metabolic disruptions are associated with an increased incidence of specific types of cancer and inflammatory diseases; however, the relationships between causes and mediators remain unclear^71,72. As RORα regulates cancer, inflammation, circadian rhythm, and metabolism, it not only functions as a transcriptional effector but also as a sensitive sensor that captures and reconstructs distinct changes in the cellular environment into transcriptional responses⁷³. Elucidating the signaling pathways of second messengers, hormones, and chromatin configurations, which modulate interactions between RORα and its binding partners, is important. Such knowledge is relevant to the identification and development of novel compounds that increase or decrease the binding affinity of RORα and its complexes, and direct the activity of cofactors toward disease crosstalk pathways.

Overall, targeting individual pathways using a direct inhibitor of RORα as a TF by suppressing the DBD or LBD only inhibits one or two pathways. Disease progression can be decelerated, but other pathways in which RORα functions as a coregulator and transregulator of target genes may be upregulated as compensation, thus generating resistance to therapy. If a defined role of RORα as a coregulator is targeted and its function prevented, combined pathways will be concurrently inhibited. Such a strategy blocks the alternate pathway of compensatory upregulation, decreasing the onset of acquired drug resistance. In this scheme, a small-molecule drug designed to block a specific action of RORα, such as a PKCα agonist acting as a coregulator, can lead to simultaneous inhibition of an array of Wnt/β-catenin downstream target genes. As PTMs designating the action of RORα as a TF or a coregulator are a result of prior specific enzymatic reactions, drugs can inhibit or activate modifying enzymes. These approaches are important for understanding the regulation of specific noncanonical pathways in which drugs do not modulate RORα via LBDs, suggesting that a mechanism-based drug administration strategy may be useful for specifically inhibiting cancer activity.

Future investigation in these areas will be critical for defining the sensory and physiological codes of the dual RORα roles as a TF and coregulator, and recognizing the potential of dual concepts as therapeutic targets. Only a few signaling pathways or PTMs that regulate the activities of specific recruitment of RORα as a direct TF or coregulator have been elucidated thus far. The specific signaling pathways and enzymes responsible for conducting the signaling cascade in terms of TFs or coregulators and the inhibitors that would be applicable for clinical applications await further investigation in the future.

Many efforts to find ligands of RORα, such as the discovery of endogenous ligands and the chemical synthesis for synthetic ligand design, have been applied to verify its importance α in various biological systems and disease models. Although T0901317, first known as an inverse agonist of RORα, was revealed to be a nonspecific ligand for several NRs, it contributed to the development of a new synthetic ligand. In addition, SR1078 showed efficacy in p53 stabilization and subsequent apoptosis, and may be utilized in cancer treatment in the future. In addition, the inverse agonist SR1001, which interferes with the interaction between receptor and coactivator, has been proven to be effective for treating autoimmune disease. Many researchers have attempted to discover a potent and selective synthetic ligand of RORα and verify its efficacy in disease models. Although RORα is a very promising target that can be applied to circadian rhythm, cancer, and metabolic and immune diseases, there are no drugs precisely targeting RORα on the market. Several pharmaceutical companies are dedicated to the discovery and development of using natural products as selective agonists of RORα. For example, GENFIT and BICOLL identified the first potent bioavailable natural product neoruscogenin as a specific agonist of RORα, and they are developing neoruscogenin as a therapeutic drug for autoimmune disorders in the preclinical stage⁷⁶. Therefore, the development of drugs targeting RORα in the future will be a significant and valuable achievement for treating related diseases.