Orphan Nuclear Receptor RORα Regulates Enzymatic Metabolism of Cerebral 24S-Hydroxycholesterol through CYP39A1 Intronic Response Element Activation

Oxysterols, important regulators of cholesterol homeostasis in the brain, are affected by neurodegenerative diseases. Early-onset Alzheimer’s disease is associated with higher levels of circulating brain-derived 24S-hydroxycholesterol (24S-OHC). Conversion of cholesterol to 24S-OHC is mediated by cholesterol 24S-hydroxylase in the brain, which is the major pathway for oxysterol elimination, followed by oxidation through hepatic first-pass metabolism by CYP39A1. Abnormal CYP39A1 expression results in accumulation of 24S-OHC, influencing neurodegenerative disease-related deterioration; thus, it is important to understand the normal elimination of 24S-OHC and the system regulating CYP39A1, a selective hepatic metabolic enzyme of 24S-OHC. We examined the role of transcriptional regulation by retinoic acid receptor-related orphan receptor α (RORα), a nuclear receptor that responds to oxysterol ligands. In humans, the promoter and first intronic regions of CYP39A1 contain two putative RORα response elements (ROREs). RORα binding and responses of these ROREs were assessed using electrophoretic mobility shift, chromatin immunoprecipitation, and luciferase reporter assays. CYP39A1 was upregulated by RORα overexpression in HEK293 cells, while RORα knockdown by siRNA significantly downregulated CYP39A1 expression in human hepatoma cells. Additionally, CYP39A1 was induced by RORα agonist treatment, suggesting that CYP39A1 expression is activated by RORα nuclear receptors. This may provide a way to increase CYP39A1 activity using RORα agonists, and help halt 24S-OHC accumulation in neurodegenerative illnesses.


Introduction
While the brain only accounts for 2% of the total body weight, it contains 25% of all the cholesterol in the body [1]. The balanced removal of oxysterol through cholesterol oxidation regulates cholesterol homeostasis in the brain. Removal of cholesterol from brain tissue is difficult, but it can cross the low-density lipoprotein receptor and 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase and synthase. CYP39A1 mRNA expression is upregulated upon dietary cholesterol intake in rats [28]. This response predicts the existence of regulatory elements in the promoter of the gene which respond in a positive fashion to cholesterol. Such elements may or may not differ from the sterol regulatory elements identified in genes coding for HMG-CoA synthase and others in the cholesterol synthesis pathways which respond negatively to cholesterol [29]. However, the transcriptional and inducible regulatory systems controlling the CYP39A1 gene are poorly understood. In this study, we aimed to elucidate the role of RORα and investigate the possibility of inducing CYP39A1 activity by RORα agonism.

RORα Bound to ROREs of CYP39A1 Promoter and Intronic Regions
Electrophoretic mobility shift assays (EMSAs) were used to assess RORα binding to DNA of two putative ROREs of the promoter and first intron of CYP39A1 (RORE1: −833/−822 as the upstream region; RORE2: +1082/+1093 as the downstream region; Figure 1A). In binding assays with RORE1 and RORE2 competing with IκB oligonucleotides, known ROREs that bind RORα, binding between IκB and RORα was inhibited. In contrast, RORE1 and RORE2 competition following a base substitution mutation did not inhibit binding of IκB DNA to RORα. Supershift experiments with the anti-RORα antibody showed specific binding of RORα to ROREs. Thus, RORα bound to CYP39A1 RORE1 and RORE2; of these, RORE2 and RORα had the highest binding affinity ( Figure 1B). Complexes containing RORα and RORE in the promoter and first intronic regions of the CYP39A1 gene were assessed using ChIP-PCR in HepG2 cells to identify RORE-RORα complexes that could be introduced into cells. Following ChIP using an RORα antibody, we performed PCR so that RORE1 (−995/−724, upstream region) and RORE2 (+887/+1264, downstream intron region) sequences could be included, resulting in RORα binding for both RORE1 and RORE2 ( Figure 1C). including the low-density lipoprotein receptor and 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase and synthase. CYP39A1 mRNA expression is upregulated upon dietary cholesterol intake in rats [28]. This response predicts the existence of regulatory elements in the promoter of the gene which respond in a positive fashion to cholesterol. Such elements may or may not differ from the sterol regulatory elements identified in genes coding for HMG-CoA synthase and others in the cholesterol synthesis pathways which respond negatively to cholesterol [29]. However, the transcriptional and inducible regulatory systems controlling the CYP39A1 gene are poorly understood. In this study, we aimed to elucidate the role of RORα and investigate the possibility of inducing CYP39A1 activity by RORα agonism.

RORα bound to ROREs of CYP39A1 Promoter and Intronic Regions
Electrophoretic mobility shift assays (EMSAs) were used to assess RORα binding to DNA of two putative ROREs of the promoter and first intron of CYP39A1 (RORE1: −833/−822 as the upstream region; RORE2: +1082/+1093 as the downstream region; Figure 1A). In binding assays with RORE1 and RORE2 competing with IκB oligonucleotides, known ROREs that bind RORα, binding between IκB and RORα was inhibited. In contrast, RORE1 and RORE2 competition following a base substitution mutation did not inhibit binding of IκB DNA to RORα. Supershift experiments with the anti-RORα antibody showed specific binding of RORα to ROREs. Thus, RORα bound to CYP39A1 RORE1 and RORE2; of these, RORE2 and RORα had the highest binding affinity ( Figure 1B). Complexes containing RORα and RORE in the promoter and first intronic regions of the CYP39A1 gene were assessed using ChIP-PCR in HepG2 cells to identify RORE-RORα complexes that could be introduced into cells. Following ChIP using an RORα antibody, we performed PCR so that RORE1 (−995/−724, upstream region) and RORE2 (+887/+1264, downstream intron region) sequences could be included, resulting in RORα binding for both RORE1 and RORE2 ( Figure 1C). The promoter region of CYP39A1. Two predicted ROREs are indicated, along with the arrangement of RORE1 and RORE2 and their nucleotide sequences, which were used for mutation analyses. Mutated bases are indicated by lowercase characters. Upstream elements are indicated by a minus sign; downstream elements are indicated by a plus sign relative to the transcription start site (TSS), identified as +1. UTR; untranslated region, ORF; open reading frame. (B) Electrophoretic mobility shift assays (EMSAs) showing in vitro binding results of RORα interacting with ROREs of the CYP39A1 promoter. DNA oligonucleotides containing RORα binding sites were end-labeled with [γ-32 P] competitor DNA oligonucleotides of IκB, CYP39A1-RORE1 or -RORE2, and incubated with RORα extracts translated in vitro. For EMSAs, anti-RORα antibodies were added to each reaction; a negative control of EgrI, a transcription factor that does not bind to ROREs, was added for comparison with anti-EgrI antibodies. (C) Chromatin immunoprecipitation (ChIP) assays with the anti-RORα antibody showing in vivo binding results of RORα with the ROREs. PCR was performed using primers for two RORE-containing regions in the CYP39A1 promoter. Normal rabbit IgG was used as a negative control.

RORα and RORE Responses in CYP39A1 Promoter and Intronic Regions
Luciferase reporter assays were performed to estimate responses to RORα at the CYP39A1 ROREs. The reporter vector connected direct repeats of RORE1 or RORE2 to the upstream elements of the minimal SV40 promoter sequence. Responses were indicated by upregulation of RORE1 (1.6-fold) and RORE2 (11.8-fold). The reporter response disappeared following a base substitution ( Figure 2A). A vector that connected the CYP39A1 region and RORE1 (−1219/+86) or RORE2 (+887/+1264 linked to −235/+86 as the core promoter region) to the upstream luciferase reporter region was built, then cells were transfected with the RORα expression vector. When RORE1 or RORE2 was included, a RORα response was observed; this response disappeared when a base substitution was introduced into the reporter region ( Figure 2B).

RORα and RORE Responses in CYP39A1 Promoter and Intronic Regions
Luciferase reporter assays were performed to estimate responses to RORα at the CYP39A1 ROREs. The reporter vector connected direct repeats of RORE1 or RORE2 to the upstream elements of the minimal SV40 promoter sequence. Responses were indicated by upregulation of RORE1 (1.6-fold) and RORE2 (11.8-fold). The reporter response disappeared following a base substitution ( Figure 2A). A vector that connected the CYP39A1 region and RORE1 (−1219/+86) or RORE2 (+887/+1264 linked to −235/+86 as the core promoter region) to the upstream luciferase reporter region was built, then cells were transfected with the RORα expression vector. When RORE1 or RORE2 was included, a RORα response was observed; this response disappeared when a base substitution was introduced into the reporter region ( Figure 2B).

CYP39A1 Expression Increased in Cells Overexpressing RORα
To quantify changes in expressions of CYP39A1 mRNA and proteins in cells overexpressing RORα, qRT-PCR and immunoblotting analyses were performed, respectively, in HEK293 cells. Changes in RORα proteins were measured. RORα and CYP39A1 mRNA levels increased in cells overexpressing RORα compared with the pSG5 empty vector ( Figure 3A,B). Moreover, a 1.96-fold increase in CYP39A1 protein levels was observed, with a 3.95-fold increase in RORα protein levels ( Figure 3C,D).
of RORα on luciferase reporter gene activities of constructs containing direct triplet repeats of wildtype (Wt) or mutant-type (Mt) ROREs of the CYP39A1 gene. Data are pooled from three independent experiments. *, p < 0.05. (B) Luciferase assays showing effects of RORα on the reporter gene expression of constructs containing Wt or Mt RORE1 or RORE2 regions linked to the core promoter of the CYP39A1 gene. Data are presented as means ± standard error of the means (n = 3). *, p < 0.05.

CYP39A1 Expression Increased in Cells Overexpressing RORα
To quantify changes in expressions of CYP39A1 mRNA and proteins in cells overexpressing RORα, qRT-PCR and immunoblotting analyses were performed, respectively, in HEK293 cells. Changes in RORα proteins were measured. RORα and CYP39A1 mRNA levels increased in cells overexpressing RORα compared with the pSG5 empty vector ( Figure 3A,B). Moreover, a 1.96-fold increase in CYP39A1 protein levels was observed, with a 3.95-fold increase in RORα protein levels ( Figure 3C,D).

CYP39A1 Expression Decreased Following RORα Knockdown
Silencing of the RORα gene by siRNA was performed to determine the impact of decreased RORα levels on CYP39A1 mRNA and protein concentrations in HepG2 cells. A decrease in CYP39A1

CYP39A1 Expression Decreased Following RORα Knockdown
Silencing of the RORα gene by siRNA was performed to determine the impact of decreased RORα levels on CYP39A1 mRNA and protein concentrations in HepG2 cells. A decrease in CYP39A1 mRNA levels was observed following RORα knockdown ( Figure 4A), and this expression was not decreased by siGFP as a negative control. Lactate dehydrogenase levels were measured as indicators of cell toxicity in the siRNA-transfected cells. The proportion of LDH in the intracellular compartment of siROR-treated cells was similar to that in the siGFP-treated cells. No cell toxicity resulted from siRNA knockdown ( Figure 4B). A 0.5-fold decrease in RORα protein concentration resulted in decreased CYP39A1 protein concentration by 0.2-fold ( Figure 4C,D). mRNA levels was observed following RORα knockdown ( Figure 4A), and this expression was not decreased by siGFP as a negative control. Lactate dehydrogenase levels were measured as indicators of cell toxicity in the siRNA-transfected cells. The proportion of LDH in the intracellular compartment of siROR-treated cells was similar to that in the siGFP-treated cells. No cell toxicity resulted from siRNA knockdown ( Figure 4B). A 0.5-fold decrease in RORα protein concentration resulted in decreased CYP39A1 protein concentration by 0.2-fold ( Figure 4C,D).

CYP39A1 Expression Increased upon RORα Ligand Administration
To investigate whether the synthetic RORα agonist, SR1078, would induce CYP39A1 mRNA expression, CYP39A1 mRNA levels in HepG2 cells treated with or without SR1078 were analyzed using qRT-PCR. Robust induction of CYP39A1 mRNA expression was observed in HepG2 cells following SR1078 administration ( Figure 5). RORα expression was unchanged and BMAL1 expression, a positive control, was induced by SR1078.

CYP39A1 Expression Increased upon RORα Ligand Administration
To investigate whether the synthetic RORα agonist, SR1078, would induce CYP39A1 mRNA expression, CYP39A1 mRNA levels in HepG2 cells treated with or without SR1078 were analyzed using qRT-PCR. Robust induction of CYP39A1 mRNA expression was observed in HepG2 cells following SR1078 administration ( Figure 5). RORα expression was unchanged and BMAL1 expression, a positive control, was induced by SR1078.

Discussion
Oxysterol accumulation in the brain causes neuronal death in AD [11]. Almost all brain oxysterol is in the form of 24S-OHC, which can pass through the BBB for elimination through the circulatory system [4], involving metabolism by the hepatic CYP39A1 oxidizing pathway [5,30]. The 24S-OHC accumulation is prevented by the normal expression and function of CYP39A1; however, CYP39A1 regulation remains understudied, including its inducible nuclear receptor ligands. We showed, for the first time, that the nuclear receptor RORα regulates CYP39A1 expression levels in human hepatoma cells, which can be induced by SR1078, a RORα agonist. SR1078 also acts as an agonist of RORγ [31]. Therefore, CYP39A1 may be regulated by activation of RORα and RORγ.
RORα regulates transcription by binding its ROREs as monomers upon oxysterol ligand binding. In silico modeling revealed the promoter and first intronic regions of the human CYP39A1 gene contained two putative ROREs ( Figure 1A). RORα bound and responded to the CYP39A1 ROREs, as determined by in vitro EMSAs ( Figure 1B), and in cellulo ChIP ( Figure 1C) and luciferase reporter assays (Figure 2). The first intronic region of the CYP39A1 gene contains RORE enhancers as transcription activation sites. Some introns contain enhancer elements or alternative promoters, while others elevate mRNA expression by an alternative intron-mediated enhancement process [32]. Here, RORE2 in the first intronic region, as well as RORE1 in the promoter region, upregulated CYP39A1 transcription. Moreover, CYP39A1 was upregulated through RORα overexpression in HEK293 cells (Figure 3), while RORα knockdown by siRNA significantly downregulated CYP39A1 expression in human hepatoma cells (Figure 4), suggesting that CYP39A1 expression is activated via RORα nuclear receptors. It has been reported that RORγ, involved in the transcriptional regulation of lipid metabolic genes, also acts as a transcription factor in the ROREs [33][34][35]. Therefore, RORγ may also act synergistically on the regulation of RORα. Whether CYP39A1 is involved in the regulatory mechanism of these two transcription factors remains to be elucidated.
The 24S-OHC molecule, a CYP39A1 substrate, acts on RORα and liver X receptors (LXRs) involved in oxysterol synthesis. Oxysterols, including 24S-OHC, 25-OHC, and 27-OHC are endogenous ligands for both RORα and LXRs [23]. RORα-and LXR-deficient mouse studies have revealed a potentially important functional crosstalk between RORα and LXRs [25]. LXRs possibly act as the major target of oxysterols, especially in the regulation of cholesterol metabolism, by binding to their respective response elements (LXREs, which contain two direct repeats of the consensus AGGTCA sequence, separated by four nucleotides) via the formation of heterodimers with an

Discussion
Oxysterol accumulation in the brain causes neuronal death in AD [11]. Almost all brain oxysterol is in the form of 24S-OHC, which can pass through the BBB for elimination through the circulatory system [4], involving metabolism by the hepatic CYP39A1 oxidizing pathway [5,30]. The 24S-OHC accumulation is prevented by the normal expression and function of CYP39A1; however, CYP39A1 regulation remains understudied, including its inducible nuclear receptor ligands. We showed, for the first time, that the nuclear receptor RORα regulates CYP39A1 expression levels in human hepatoma cells, which can be induced by SR1078, a RORα agonist. SR1078 also acts as an agonist of RORγ [31]. Therefore, CYP39A1 may be regulated by activation of RORα and RORγ.
RORα regulates transcription by binding its ROREs as monomers upon oxysterol ligand binding. In silico modeling revealed the promoter and first intronic regions of the human CYP39A1 gene contained two putative ROREs ( Figure 1A). RORα bound and responded to the CYP39A1 ROREs, as determined by in vitro EMSAs ( Figure 1B), and in cellulo ChIP ( Figure 1C) and luciferase reporter assays ( Figure 2). The first intronic region of the CYP39A1 gene contains RORE enhancers as transcription activation sites. Some introns contain enhancer elements or alternative promoters, while others elevate mRNA expression by an alternative intron-mediated enhancement process [32]. Here, RORE2 in the first intronic region, as well as RORE1 in the promoter region, upregulated CYP39A1 transcription. Moreover, CYP39A1 was upregulated through RORα overexpression in HEK293 cells (Figure 3), while RORα knockdown by siRNA significantly downregulated CYP39A1 expression in human hepatoma cells (Figure 4), suggesting that CYP39A1 expression is activated via RORα nuclear receptors. It has been reported that RORγ, involved in the transcriptional regulation of lipid metabolic genes, also acts as a transcription factor in the ROREs [33][34][35]. Therefore, RORγ may also act synergistically on the regulation of RORα. Whether CYP39A1 is involved in the regulatory mechanism of these two transcription factors remains to be elucidated.
The 24S-OHC molecule, a CYP39A1 substrate, acts on RORα and liver X receptors (LXRs) involved in oxysterol synthesis. Oxysterols, including 24S-OHC, 25-OHC, and 27-OHC are endogenous ligands for both RORα and LXRs [23]. RORα-and LXR-deficient mouse studies have revealed a potentially important functional crosstalk between RORα and LXRs [25]. LXRs possibly act as the major target of oxysterols, especially in the regulation of cholesterol metabolism, by binding to their respective response elements (LXREs, which contain two direct repeats of the consensus AGGTCA sequence, separated by four nucleotides) via the formation of heterodimers with an obligate partner, the retinoid X receptor [36]. Both nuclear receptors recognize response elements, including a commonly conserved half-site, but target genes of transcriptional regulation are not perfectly homologous [37]. Multiple RORα and LXR isoforms may exist or CYP39A1 may be regulated by response elements further upstream.
RORα and its inverse processes, including REV-ERB (reverse orientation the c-erbA-1 gene) nuclear receptors, may be targeted using synthetic ligands for treating several diseases [27], including atherosclerosis [38,39], nonalcoholic steatohepatitis (NASH) [40], and autism [41]. We showed that CYP39A1 was induced following RORα agonist treatment ( Figure 5), suggesting CYP39A1 expression can be upregulated, increasing elimination and inhibiting accumulation of 24S-OHC, thus reducing neuronal death. Another possibility is that 24S-OHC acts as an endogenous inverse agonist at RORα and RORγ [24] and an agonist at LXR [18]. Others have reported that LXR does not act as an activator in the brain or liver in vivo, even if 24S-OHC production is increased in a transgenic mouse model with constitutive cholesterol 24-hydroxylase expression [42]. Therefore, 24S-OHC may concentration-dependently influence ROR-and LXR-mediated gene expression, although the inducible mechanisms through which it may act as an endogenous CYP39A1 substrate are unknown. Although the in vivo action of 24S-OHC is unclear, hepatic CYP39A1 expression is induced by steroid analogs via nuclear receptor activation and may have neuroprotective effects.
Treatment with the thiazolidinedione class of peroxisome proliferator-activated receptor gamma (PPARγ) agonists, widely prescribed for treating type II diabetes mellitus, have been shown to significantly improve memory and cognition in patients with AD [43]. The PPARγ transcriptional network is regulated by RORα in hepatic lipid homeostasis [44], but the relationship between RORα activity and AD remains unclear at the clinical, cellular, and molecular levels.
Ultimately, we showed the RORα nuclear receptor directly bound to two ROREs located upstream of the promoter and downstream of the first intronic region of the CYP39A1 transcription start site, resulting in CYP39A1 expression regulation. Moreover, RORα agonist administration induced CYP39A1 expression, suggesting RORα agonists could be used to suppress neuronal death.

Cell Culture
HepG2 and HEK293 cells obtained from the American Type Culture Collection (Manassas, VA, USA) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 • C and 5% CO 2 .

Electrophoretic Mobility Shift Assay (EMSA)
EMSAs were performed as described previously [45]. RORα was synthesized in vitro using the TNT Quick Coupled Transcription/Translation System (Promega, Madison, WI, USA) and incubated with reaction mixture on ice in the absence or presence of a double-stranded RORE oligonucleotide. After 20 min, 0.02 pmoles of a 32 P-labeled double-stranded RORE probe of the NF-κB inhibitor (IκB) gene, a known RORα target [46], was added, and incubated at 30 • C. For competition assays, 5-, 10-, or 20-fold excess concentrations of unlabeled competitor oligonucleotides (two ROREs of CYP39A1-wild-type [wt] or mutant type [mt]) were added before the labeled oligonucleotides. For supershift assays, specific antibodies for anti-RORα (sc-28612; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or anti-EgrI (as negative controls; sc-110; Santa Cruz Biotechnology) were added prior to the radiolabeled probe. Mixtures were subjected to 4% nondenaturing polyacrylamide gel electrophoresis in 0.5× TBE (tris/borate/ethylenediaminetetraacetic acid) buffer, visualized by autoradiographic imaging on a Typhoon 9400 (GE Healthcare, Little Chalfont, UK). Table S1 shows the probe sequences.

Luciferase Reporter Assays
Sixteen hours before transfection, HepG2 cells were seeded in 24-well plates (0.5 × 10 5 cells/well). Luciferase reporter plasmids, pRORE1-wt × 3 or pRORE2-wt × 3 carrying three RORE1 or RORE2 regions of CYP39A1 connected to the minimal SV40 promoter region in a PGV-P2 vector (Toyo Ink, Tokyo, Japan), or pCYP39A1-RORE1-wt (from −838 to +86 relative to the transcription start site [TSS]) or pCYP39A1-RORE2-wt (from +1077 to +1264 connected to the core promoter region, from −235 to +86 relative to the TSS) carrying the human CYP39A1 promoter and first intronic regions in a PGV-B2 vector (Toyo Ink) were cotransfected with expression vectors for RORα using an empty pSG5 vector (Agilent Technologies, Santa Clara, CA, USA), along with an internal β-galactosidase standard (each with 100 ng added/well), using Lipofectamine 2000 (Life Technologies, Gaithersburg, MD, USA) [45]. Reporter construct primers are shown in Table S1. Luciferase constructs, including RORE1 or RORE2, were mutated by a base substitution and compared to wt constructs. Cells were incubated for 32 h, then subjected to luciferase reporter assays, normalized relative to β-galactosidase activities, using PicaGene Luminescence Kits (Toyo Inc, Tokyo, Japan).

Quantitative Reverse Transcription-PCR (qRT-PCR)
Total RNA was extracted using ISOGEN kits (Nippon Gene, Tokyo, Japan). cDNA synthesis and qRT-PCR were performed as previously described [45]. Briefly, 300 ng of total RNA was reverse transcribed using random hexamers (Takara Bio) and Moloney murine leukemia virus reverse-transcriptase (Thermo Fisher Scientific). Of each cDNA, 10 ng was added to the SYBR Green Realtime PCR Master Mix (Toyobo, Osaka, Japan) or LightCycler 480 SYBR Green I Master (Roche, Basel, Switzerland) with 1 µM of each primer (see Table S1). Real-time fluorescence monitoring was performed using LightCycler 2.0 or LightCycler 480 II (Roche). Values were normalized to 18S rRNA, expressed relative to controls (treated empty vector (pSG5), small interfering RNA (siRNA) targeting green fluorescent protein (siGFP), or vehicle). The BMAL1 gene, a known RORα target, was used as a positive control.

Overexpression Analysis
HEK293 cells (1 × 10 5 cells/well) were seeded in 12-well plates 16 h before transfection. RORα expression plasmids with an empty pSG5 vector (0.3 µg/well for 12-well plates or 0.6 µg/well for 6-well plates) were transfected using Lipofectamine 2000 (Life Technologies); cells were harvested 48 or 72 h later. Total RNA and protein were extracted as described above.

Statistical Analyses
Data are shown as means ± standard errors of the mean. Student's t-tests were used to compare group means using Microsoft Excel software (version no. 1908; Microsoft corporation, Tokyo, Japan). p < 0.05 was considered statistically significant.