Retinoid‐related orphan receptor α (RORα) (NR1F1) is a member of the nuclear receptor superfamily whose biological functions are largely unknown. Since staggerer mice, which carry a deletion in the RORα gene, suffer from immune abnormalities, we generated an adenovirus encoding RORα1 to investigate its potential role in control of the inflammatory response. We demonstrated that RORα is expressed in human primary smooth‐muscle cells and that ectopic expression of RORα1 inhibits TNFα‐induced IL‐6, IL‐8 and COX‐2 expression in these cells. RORα1 negatively interferes with the NF‐κB signalling pathway by reducing p65 translocation as demonstrated by western blotting, immunostaining and electrophoretic mobility shift assays. This action of RORα1 on NF‐κB is associated with the induction of IκBα, the major inhibitory protein of the NF‐κB signalling pathway, whose expression was found to be transcriptionally upregulated by RORα1 via a ROR response element in the IκBα promoter. Taken together, these data identify RORα1 as a potential target in the treatment of chronic inflammatory diseases, including atherosclerosis and rheumatoid arthritis.
Retinoids, vitamin D, fatty acid derivatives, thyroid and steroid hormones regulate developmental and physiological processes in vertebrates by binding to specific transcription factors belonging to the nuclear receptor superfamily (Mangelsdorf et al., 1995). In addition to these ligand‐activated receptors, other members of the family are structurally related proteins for which no ligand has yet been identified, and therefore are referred to as ‘orphan’ receptors. The retinoid‐related orphan receptor (ROR) subfamily contains three members, RORα (NR1F1), RORβ (NR1F2) and RORγ (NR1F3). The RORα gene generates four isoforms that share common DNA‐binding domains (DBD) and putative ligand‐binding domains (LBD) but are distinguished by different N‐terminal domains (Carlberg et al., 1994; Giguère et al., 1995). RORα isoforms bind, as monomers, to ROR response elements (RORE) composed of a 6 bp A/T rich region immediately preceding a half core AGGTCA motif (Giguère et al., 1995; McBroom et al., 1995; Ma et al., 1998). In addition, as similarly reported for Rev‐erbα, another orphan receptor RORα is also able to bind DR‐2 elements as a homodimer (Moraitis and Giguère, 1999).
Numerous binding sites for RORα have been found within gene promoters, but the transcriptional regulation of these genes by RORα has only been functionally demonstrated for γF‐crystallin (Tini et al., 1995), N‐myc (Dussault and Giguère, 1997), laminin B1 (Matsui, 1996), rat apolipoprotein AI (Vu‐Dac et al., 1997), Purkinje cell protein‐2 (Matsui, 1997) and prosaposin (Jin et al., 1998). Since RORα regulates a wide spectrum of genes expressed in various tissues, it is difficult to assign a precise role to this orphan receptor. However, analysis of RORα‐deficient mice revealed that RORα plays a crucial role in the development of the central nervous system (Hamilton et al., 1996; Dussault et al., 1998; Steinmayr et al., 1998). RORα‐deficient mice display the cerebellar defects of the ataxic staggerer mice: they suffer from impaired motor coordination, hanging and equilibrium deficits. Homozygous staggerer mice carry a deletion within the RORα gene that prevents the translation of the ligand‐binding domain (Hamilton et al., 1996). Interestingly, the premature death of the staggerer mice does not correlate with the cerebellar defects, suggesting that RORα may play additional roles in development and physiology.
Since it has been demonstrated that staggerer mice fed a high‐fat diet develop severe atherosclerosis (Mamontova et al., 1998), and since these mice display immune abnormalities associated with increased mRNA levels of IL‐1β, IL‐6 and TNFα (Kopmels et al., 1992), we studied RORα expression in primary smooth‐muscle cells (SMC) and examined whether RORα regulates the inflammatory response in these cells. Using an adenovirus encoding for RORα1, we demonstrate that RORα1 negatively regulates the inflammatory response by interfering with the NF‐κB signalling pathway. This action of RORα1 on NF‐κB is likely to be due to the induction of IκBα, the major inhibitory protein of the NF‐κB signalling pathway, whose expression was found to be transcriptionally upregulated by RORα1 via a RORE in the IκBα promoter.
Results and Discussion
Since RORα may play a role in inflammation (Kopmels et al., 1992) and atherosclerosis (Mamontova et al., 1998), we first examined RORα expression in human primary SMC by RT–PCR analysis. RORα expression was readily detected in the different vascular SMC types analysed (Figure 1A). To investigate a potential role of RORα in the control of the inflammatory response, we generated by homologous recombination an adenovirus encoding the RORα1 isoform (Ad‐RORα1), which is the predominantly expressed transcript among the four isoforms (Forman et al., 1994). Infection of primary SMC with Ad‐RORα1 led to a strong expression of RORα mRNA levels whereas infection with a control Ad‐GFP did not affect endogenous levels (Figure 1C). Immunocytochemistry experiments using a specific polyclonal antibody raised against RORα (aa 163–225) confirmed these results at the protein expression level and further demonstrated that both endogenous and exogenous RORα are principally localized in the nucleus (Figure 1D).
Next, the influence of RORα1 on different markers of the inflammatory response to TNFα was tested. In absence of stimulation, both IL‐6 and IL‐8 were detected at low levels in the culture medium of Ad‐GFP‐infected cells. This low cytokine production was significantly repressed in RORα1‐infected cells (Figure 2A and B). TNFα treatment of Ad‐GFP‐infected cells for 24 h led to a sharp induction of both cytokines, which was significantly repressed by RORα1 infection (Figure 2A and B). COX‐2 protein levels were monitored by western blot analysis in the same experiments. TNFα strongly induced COX‐2 expression, which was completely blunted in Ad‐RORα1‐infected cells (Figure 2C). Finally, in Ad‐GFP‐infected cells, we failed to detect endogenous RORα protein, which is likely to be due to the low affinity of our antibody. However, in Ad‐RORα1 cells, high amounts of RORα protein were detected and its levels were not affected by TNFα stimulation (Figure 2D). These data indicate that RORα1 negatively regulates the TNFα‐induced inflammatory response in primary aortic SMC.
Since the IL‐6, IL‐8 and COX‐2 genes have been reported to be NF‐κB‐driven genes (Thurberg and Collins, 1998), we hypothesized that RORα1 may negatively interfere with the NF‐κB signalling pathway. Because of the inability to transfect primary cultured human aortic SMC, rat SMC PAC1A cells were transiently transfected with an NF‐κB‐driven promoter construct in the presence or absence of RORα1. Since these cells were unresponsive to TNFα stimulation, lipopolysaccharide (LPS) was used as an NF‐κB signalling inducer. LPS treatment significantly (>3‐fold) increased NF‐κB‐dependent promoter activity. This induction was abolished by RORα1 cotransfection whereas basal promoter activity was not significantly affected (Figure 3A). These results suggest that RORα1 may exert its anti‐inflammatory activities by inhibiting NF‐κB transcriptional activity.
To obtain further insight into the molecular mechanisms of the anti‐inflammatory activities of RORα1, we next tested, by western blot analysis and by immunocytochemistry experiments, the influence of RORα1 overexpression on p65 nuclear translocation, a critical event in NF‐κB signaling pathway activation. As expected, TNFα treatment for 30 min led to p65 nuclear translocation in Ad‐GFP‐infected cells (Figure 3B and C). This translocation was less pronounced in Ad‐RORα1‐infected cells (Figure 3B and C), suggesting that RORα1 may control p65 translocation without regulating total p65 protein expression (Figure 3B, lower panel). To explore the consequences of this reduced translocation, electrophoretic mobility shift assays (EMSAs) were performed. Twenty‐four hours after cell infections, TNFα was added for 30 min, and nuclear extracts were prepared and subjected to EMSA using an NF‐κB radiolabelled consensus probe. Ad‐GFP‐infected cells expressed a basal NF‐κB DNA‐binding activity, which was significantly lowered in Ad‐RORα1‐infected cells, consistent with our previous results (Figure 2A and B). As expected, TNFα stimulation led to a strong increase in NF‐κB DNA binding, which was again sharply reduced in Ad‐RORα1 cells (Figure 4A). The presence of p65 protein in the NF‐κB complex was verified by shifting the complex using an anti‐p65 antibody (data not shown). As a control, neither TNFα nor RORα1 overexpression affected Sp1 DNA binding (Figure 4A), indicating that RORα1 specifically interferes with NF‐κB DNA‐binding activities. Since many nuclear receptors have been reported to inhibit NF‐κB‐dependent transcription by physical interaction with the p65 subunit (Göttlicher et al., 1998), we next explored the existence of such a mechanism. The influence of RORα1 co‐transfection on transactivation of an NF‐κB response element‐driven reporter vector by transfected p65 was tested in PAC1A cells. As expected, p65 transfection resulted in a strong induction of reporter activity (Figure 4B). However, this induction was not affected by RORα1 co‐transfection, indicating that the negative interference of RORα1 with NF‐κB signalling is independent of direct association with p65 and thus occurs upstream of p65. Similar results were obtained with p50 co‐transfection (data not shown).
In resting cells, IκBα sequesters p50–p65 heterodimers in a cytoplasmic inactive complex. Since RORα1 inhibits p65 translocation and subsequent NF‐κB DNA‐binding activity, IκBα expression levels were analysed in both Ad‐GFP‐ and Ad‐RORα1‐infected cells. Ad‐GFP‐infected cells express low IκBα mRNA levels (Figure 5A). Interestingly, IκBα transcript levels were drastically increased in Ad‐RORα1‐infected cells (Figure 5A), suggesting that RORα1 inhibits NF‐κB signalling by inducing IκBα gene expression. In order to provide genetic evidence for a role of RORα in the control of IκBα gene expression, IκBα mRNA levels were analysed in aortas from wild‐type and homozygous staggerer mice by dot‐blot analysis (Figure 5B). Aortas from staggerer mice display significantly lower basal levels of IκBα mRNA than in vivo wild‐type mice, indicating that RORα regulates IκBα gene expression in the vascular wall (Figure 5B). Furthermore, staggerer mice displayed an exacerbated inflammatory response, as demonstrated by hyperproduction of IL‐6 after phorbolester treatment of splenocytes isolated from staggerer mice, compared with those from wild‐type mice (see Supplementary data, available at EMBO reports Online). To further study IκBα gene regulation by RORα1, a (−929 +22) promoter fragment was PCR amplified and inserted upstream of the luciferase gene (Ito et al., 1994). Co‐transfection of RORα1 induced 3‐fold the (−929 +22) but not the (−385 +22) promoter construct (Le Bail et al., 1993), indicating that the regulatory region locates between −929 and −385 (Figure 5C). Sequence analysis revealed the presence of a putative RORE located between −911 and −895 (Figure 5C). Mutation of this putative RORE site confered unresponsiveness to RORα1 (Figure 5C), indicating that this site mediates IκBα transcriptional induction by RORα1.
Since the RORα gene generates several isoforms (Giguère et al., 1994), we next evaluated the influence of RORα1, RORα2 and RORα3 on a (IκBα‐RORE)2‐driven promoter construct. As expected, RORα1 strongly induced the reporter construct activity (almost 10‐fold) (Figure 6A), showing that the IκBα‐RORE can function in a promoter‐independent manner. By contrast, RORα2 and RORα3 failed to induce this latter construct (Figure 6A), suggesting that IκBα transcriptional induction is specific to the RORα1 isoform. A dominant‐negative RORα1Δ235 lacking the LBD was found to inhibit RORα1 transcriptional activity by competing for the same DNA‐binding site (McBroom et al., 1995). The effect of this mutant was next tested on RORα1‐induced (IκBα‐RORE)2‐driven promoter activity. As expected, co‐transfection of RORα1 led to a strong induction of the promoter activity, which was completely abolished by RORα1Δ235 cotransfection (Figure 6B), suggesting that RORα1 binding to the promoter is essential for the transcriptional induction. Finally, RORα1 binding to the identified RORE site was verified by EMSA using in vitro translated RORα1 proteins (Figure 6C). Competition experiments using wild‐type and mutated cold oligonucleotides, as well as supershift experiments, demonstrate that RORα1 binds strongly to the wild‐type but not to the mutated RORE site.
In this study, we report that the overexpression of RORα1 in human aortic SMC prevents TNFα‐induced IL‐6, IL‐8 and COX‐2 expression, three important markers of the inflammatory response. RORα1 negatively regulates the cytokine‐induced inflammatory response by upregulating IκBα, the major inhibitor of the NF‐κB signalling pathway, at the transcriptional level, thereby reducing p65 nuclear translocation. Previous studies reported that staggerer mice display immune abnormalities such as IL‐1β hyperproduction in macrophages (Kopmels et al., 1990, 1992; Vernet‐der Garabedian et al., 1998). Interestingly, we found that staggerer mice express lower levels of IκBα transcript in the vascular wall compared with wild‐type mice (Figure 5B). In addition, cotransfection of a dominant‐negative form of RORα abolished RORα1‐induced IκBα transcription (Figure 6B). These results may, at least in part, explain the inflammatory phenotype of the staggerer mice. Interestingly, 5‐lipoxygenase, an important enzyme involved in the control of allergic and inflammatory reactions, has been reported to be a RORα target gene (Steinhilber et al., 1995). 5‐lipoxygenase was shown to be downregulated by RORα and RZRα in human B lymphocytes. RORα1, but not RORα2 nor RORα3, efficiently binds to RORE site in its promoter. However, this study was based on the use of melatonin as a specific RORα activator (Becker‐André et al., 1994), data which are controversial (Becker‐André et al., 1997). Altogether, the results of the present study provide a novel function for RORα1 as a negative regulator of the vascular inflammatory response, as well as a potential molecular basis for its anti‐inflammatory activity. The discovery of synthetic RORα ligands may lead to the development of compounds useful in the treatment of inflammatory disorders.
Primary human aortic (HA) and coronary artery (CA) SMC (PromoCell, Heidelberg, Germany) and primary SMC from saphenous veins (VSMC: a kind gift of Dr Walsh, Boston, MA) were cultured under standard conditions. Cells from passages 5–8 were used for the experiments.
RNA preparation and northern blot hybridizations were performed as previously described (Staels et al., 1992) using IκBα and GAPDH cDNA fragments as probes. For RT–PCR analysis of RORα expression, total RNA was reverse transcribed and subsequently amplified by PCR using the following primers: for RORα, 5′‐GTCAGCAGCTTCTACCTGGAC‐3′ and 5′‐GTGTTGTTCTGAGAGTGAAAGGCACG‐3′ (fragment size: 482 bp); for GAPDH, 5′‐ATGCAGCCCCGAATGCTCCTCATCGTGGCC‐3′ and 5′‐TTCTTGGAGGCCATGTGG GCCAT‐3′ (fragment size: 239 bp).
The recombinant adenovirus (Ad‐GFP and Ad‐RORα1) was obtained by homologous recombination in Escherichia coli (Chartier et al., 1996) after insertion of the cDNAs into pAdCMV2 vector. Viral stocks were then created as previously described (Sardet et al., 1995). Viral titres were determined by a plaque assay on 293 cells and defined as pfu/ml. Cells were infected at an input multiplicity (MOI) of 100 virus particles per cell, by adding virus stocks directly to the SMC culture medium.
Plasmids and transient transfections.
The expression vectors pSG5‐RORα1, ‐RORα2, ‐RORα3 and ‐RORα1Δ235 were derived from the previously reported pCMX vectors (McBroom et al., 1995). The p65 and p50 expression plasmids were as previously described (Staels et al., 1998). A −929+22 IκBα promoter fragment was amplified using human genomic DNA and inserted into pGL2 basic vector (Promega) yielding p(−929+22) IκBα‐Luc. A mutation in the IκBα promoter ROR site was introduced using the site‐directed mutagenesis kit (Stratagene), leading to the mutation of the ROR site AGGTCA to ACCTCA. The p(−385+22) IκBα‐Luc reporter vector was generously provided by Dr Israël (Institut Pasteur, Paris, France). The (IκBα‐RORE)2‐TK‐Luc reporter vector was generated by inserting two copies of the IκBα ROR site in front of the minimal TK promoter. The (NF‐κB)3‐Luc reporter construct was provided by Stratagene. PAC1A cells (a rat aortic SMC line) were transfected using a lipid‐cationic technique as previously described (Delerive et al., 2000). Phosphoglycerate kinase (PGK) β‐galactosidase expression plasmid (50 ng) was cotransfected as a control for transfection efficiency. All transfection experiments were repeated at least three times.
Immunostaining was performed as previously described (Chinetti et al., 1998), using specific antibodies raised against RORα and p65 (Santa Cruz, sc‐109). Proteins were visualized using secondary rhodamine‐conjugated anti‐rabbit antibodies with a LEITZ DMR fluorescence microscope.
Supplementary data for this paper are available at EMBO reports Online.
The authors are grateful to Dr Habib for providing COX‐2 antibodies. This work was supported by grants of the Institut Pasteur de Lille, INSERM and the Région Nord‐Pas‐de‐Calais/Feder. P.D. is supported by a grant of the Région Nord‐Pas‐de‐Calais.
- Copyright © 2001 European Molecular Biology Organization