Hypoxia-induced activation of the retinoic acid receptor-related orphan receptor alpha4 gene by an interaction between hypoxia-inducible factor-1 and Sp1.

Hypoxia plays a key role in the pathophysiology of many disease states, and expression of the retinoic acid receptor-related orphan receptor alpha (RORalpha) gene increases under hypoxia. We investigated the mechanism for this transient hypoxia-induced increase in RORalpha expression. Reverse transcription-coupled PCR analysis revealed that the steady-state level of mRNA for the RORalpha4 isoform, but not the RORalpha1 isoform, increased in HepG2 cells after 3 h of hypoxia. Transient transfection studies showed that the hypoxia-induced increase in RORalpha4 mRNA occurs at the transcriptional level and is dependent on a hypoxia-responsive element (HRE) located downstream of the promoter. A dominant-negative mutant of hypoxia-inducible factor-1alpha (HIF-1alpha) abrogates the transcription activated by hypoxia as well as the transcription activated by exogenously expressed HIF-1alpha, demonstrating the direct involvement of HIF-1alpha in the transcriptional activation. However, HIF-1 alone was not sufficient to activate transcription in hypoxic conditions but, rather, required Sp1/Sp3, which binds to a cluster of GC-rich sequences adjacent to the HRE. Deletion of one or more of these GC boxes reduced or eliminated the HIF-1-dependent transcription. Together, these results suggest that the hypoxia-responsive region of the RORalpha4 promoter is composed of the HRE and GC-rich sequences and that the transcriptional activation under hypoxia is conferred through the cooperation of HIF-1 with Sp1/Sp3.

includes nuclear receptors that recognize and bind to the AG-GTCA half-site motif as a monomer. Little is known about ligands for the members of this group.
The ROR 1 subfamily, which has three members (ROR␣, ROR␤, and ROR␥), belongs to the monomeric nuclear receptor subgroup (3). Recently, evidence has been accumulating that ROR␣ is involved in cell differentiation and proliferation. The homozygous mutant mouse staggerer (sg) was initially described as ataxic due to the presence of massive neurodegeneration in the cerebellum. The identification of the widely expressed ROR␣ gene as the site of mutation in the sg mouse has led in recent years to great progress in understanding the molecular basis of its phenotype (4). ROR␣ knock-out mice display the same cerebellar atrophic phenotype as the sg mouse (5,6). Detailed analyses of such ROR␣-deficient mice have revealed a number of additional phenotypes outside the nervous system. These include a greater susceptibility to atherosclerosis (7), immunodeficiencies linked to the overexpression of inflammatory cytokines (8,9), abnormalities in the formation and maintenance of bone tissue (10), changes in muscle differentiation (11), and an increase in ischemia-induced angiogenesis (12). ROR␣ has also been suggested to play a role in vascular function (13,14). It thus appears that ROR␣ has direct links to a number of age-related pathologies of great medical interest.
ROR␣ is composed of four isoforms that share common DNA and ligand binding domains but differ in their N-terminal regions (15)(16)(17)(18). They are produced from a single gene by way of different promoters. It has been suggested that the N-terminal region of ROR␣ affects the specificity of recognition by ROR␣-responsive elements (21). Although ROR␣ is expressed ubiquitously in vivo, each isoform shows a tissue-specific distribution (18 -20). Several tissues or cell types, including liver, brain, endothelial cells, and smooth muscle cells, express mainly ROR␣1 and ROR␣4, whereas ROR␣2 and ROR␣3 are expressed abundantly in testis. These observations suggest that each ROR␣ isoform plays a different role in a range of cellular processes. Although little is known about the regulation of ROR␣ expression, it has recently been demonstrated that expression of ROR␣ is transiently increased when ischemia is induced in the hindlimbs of mice (12) and in various cell types in response to oxygen deprivation (hypoxia) (20,22).
Hypoxia is an essential developmental and physiological stimulus that also plays a key role in the pathophysiology of cancer, heart attack, stroke, and other major causes of mortal-ity (for review, see Ref. 23 and references therein). Cellular adaptation to hypoxia is associated with induction of a number of genes, including vascular endothelial growth factor, erythropoietin, several glycolytic enzymes, and inducible nitric-oxide synthase (23). The transcriptional induction of these genes is mediated in large part by HIF-1, which is a heterodimeric transcription factor composed of HIF-1␣ and HIF1␤ (also known as ARNT) (24,25). HIF-1 binds directly to a hypoxiaresponsive element (HRE) and activates the transcription of genes to adapt to oxygen deprivation. Although HIF-1␤ is constitutively expressed and is stable under normoxia, expression of the HIF-1␣ protein is regulated by the level of oxygen (26 -28). During normoxia, HIF-1␣ protein undergoes rapid ubiquitination and proteasomal degradation. Upon hypoxia, the ubiquitination is blocked, which allows HIF-1␣ to dimerize with HIF-1␤ and to translocate it into the nucleus (29).
In the present study we investigated the mechanism by which the expression of ROR␣ gene is regulated during hypoxia. The results show that the hypoxia-induced increase of ROR␣ expression is specific for the ROR␣4 isoform and occurs at the transcriptional level. Promoter analysis revealed a novel structure for the hypoxia-responsive region, which consists of a putative HRE and a cluster of GC-rich sequences. The hypoxia-induced activation of the ROR␣4 promoter appears to be regulated through a functional interaction between HIF-1 and Sp1/Sp3.

EXPERIMENTAL PROCEDURES
Cell Culture and Induction of Hypoxia-HeLa cells and HepG2 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (Invitrogen). Caco2 cells were maintained in minimal essential medium supplemented with 20% fetal calf serum and nonessential amino acids. For exposure to hypoxia, cells were cultured in a humidified 1% O 2 , 5% CO 2 , 94% N 2 incubator for the indicated time. Alternatively, cells were cultured in medium containing 100 M cobalt chloride (Co 2ϩ ).
Plasmid DNA-A mouse genomic library was screened using the 5Ј-end of the ROR␣4 cDNA as a probe. About 6 kilobases of the XhoI fragment containing the ROR␣4 promoter region was subcloned into the XhoI site of the pGVB2 luciferase vector (Nippon Gene, Japan), in which the HindIII site has been disrupted (pROR␣4XXluc). 5Ј-Deletion mutants were constructed by exonuclease III/S1 nuclease digestion of pROR␣4XXluc at the HindIII site followed by digestion with NheI. After ligation with an EcoRI linker (5Ј-CCGAATTCGG-3Ј), the larger DNA fragment was isolated and self-ligated. For construction of 3Ј deletion mutants, pROR␣4HXLuc was constructed by removing the HindIII-NheI fragment from pROR␣4XXLuc, digestion by exonuclease III/S1 nuclease at the XhoI site, and ligation with the EcoRI linker. After digestion with SphI the larger DNA was isolated and subcloned into the SmaI (changed to an EcoRI site with the linker) and SphI sites of pGVB2. An internal deletion mutant was constructed by combination of the 5Ј and 3Ј deletion mutant. A HRE-mutated ROR␣4 reporter was constructed using the GeneEditor in vitro site-directed mutagenesis system (Promega). The primer used to introduce a mutation within the HRE was 5Ј-CGCGCCTGCCGGAGCACCGGTACAGCTTGAGGGG-3Ј. An expression vector for HIF-1␣ (pBoshHIF-1␣) was kindly provided by Dr. Y. Fujii. An expression vector producing a dominant negative HIF-1␣ (amino acids 1-363, pHIF-1␣DN) was constructed by partial digestion of pBoshHIF-1␣ with Age I.
Transfection of Cells-Transfection was performed using the Effectene TM reagent (Qiagen) according to the manufacturer's protocol. Briefly, cells were seeded in a 24-well microplate at a density of 4 -6 ϫ 10 4 /well 1 day before transfection. Fifty nanograms of reporter DNA, 50 ng of effector DNA, and 50 ng of pSG␤Gal were used for each well. The total amount of DNA was adjusted to 200 ng with pUC19 DNA. After transfection for 12-16 h, the cells were washed once with phosphatebuffered saline and then grown for 24 h in fresh medium. All transfections were carried out in duplicate, and each experiment was repeated at least three times. The relative luciferase activity was calculated by normalization to the ␤-galactosidase activity.
Electromobility Gel Shift Assay (EMSA) and DNase I Footprinting Analysis-Nuclear extracts from normoxic or hypoxic cells were prepared as described previously (30). The nuclear proteins was incubated in a 10-l reaction containing 25 mM HEPES, pH 7.9, 50 mM KCl, 10% glycerol, 1 mM dithiothreitol, 0.1% Nonidet P-40, and 50 g/ml of poly(dI-dC). After 10 min on ice, 32 P-labeled probe was added (10,000 cpm/reaction), and the reaction was further incubated on ice for 20 min. The DNA-protein complex was separated in a 5% polyacrylamide gel. For DNase I footprinting, the binding reaction was carried out in a total of 20 l under the same conditions as EMSA. After binding for 20 min, CaCl 2 , MgCl 2 , and DNase I were added to final concentrations of 0.25 mM, 5 mM, and 100 ng/ml, respectively. Digestion was performed for 1 min at room temperature and stopped by the addition of a final concentration of 10 mM EDTA. The DNA was then purified and analyzed using a 5% denaturing sequence gel. Oligonucleotides used in the binding experiments were as follows: HRE␣4, 5Ј-GCCGGAGCTCACGTAC-AGCTTGAG-3Ј; HRE␣4m, 5Ј-GCCGGAGCTCAAATACAGCTTGAG-3Ј; HREepo, 5Ј-CTAGCCCTACGTGCTGCCTCG-3Ј; GC-box, 5Ј-ATTCGA-TCGGGGCGGGGCGAGC-3Ј; GA-box, 5Ј-TCACTTGGGGGAGGATTG-CAT-3Ј; GRE, 5Ј-AGTTATGGTTACAAACTGTTCTTAAAACA-3Ј.

Induction of ROR␣4 Isoform Expression by Hypoxia-During
hypoxia, HepG2 cells have been shown to express increased levels of ROR␣ (20). To confirm this in our hands, we first used reverse transcription-coupled PCR to analyze the effects of hypoxia on the steady-state levels of ROR␣1 and ROR␣4 mRNA in HepG2 cells. To avoid artifacts due to fluctuations in the PCR reactions, amplification of both transcripts was carried out in the same reaction using two 5Ј primers specific for each isoform and a common 3Ј primer. Each 5Ј primer selectively amplified its target transcript (data not shown). The amount of ROR␣4 mRNA increased several-fold after the cells were exposed to hypoxia for 3 h, whereas that of ROR␣1 mRNA did not change (Fig. 1). In addition, the increased level of ROR␣4 mRNA decreased to the control level when the cells were reoxygenated (data not shown). Thus, hypoxia specifically increased expression of the ROR␣4 isoform in HepG2 cells.
Transcriptional Regulation of the ROR␣4 Gene Promoter by Hypoxia-Because each ROR␣ isoform differs from the other isoforms only in its N-terminal region, it is likely that the hypoxia-induced increase of ROR␣4 mRNA expression is regulated at the transcriptional level. To examine this, we analyzed the effect of hypoxia on ROR␣4 promoter activity in a transient transfection assay. We constructed a reporter DNA in which a fragment encompassing the region from Ϫ1015 to ϩ408 of the mouse ROR␣4 promoter was linked upstream to the firefly luciferase gene. As shown in Fig. 1C, transcription of the ROR␣4 promoter increased severalfold upon exposure of the transfected HepG2 cells to hypoxia. A similar increase in the promoter activity was also observed when cells were grown in medium containing 100 M CoCl 2 (Co 2ϩ ), which is known to mimic hypoxic conditions. Additionally, almost identical transcriptional responses to hypoxia were observed both in HeLa and Caco2 cells, each of which differently expresses ROR␣4 mRNA (data not shown). These results clearly indicate that the hypoxia-induced increase of ROR␣4 expression occurred at the transcriptional level. Because the promoter activity of ROR␣4 was highest in Caco2 cells (data not shown), we used these cells for the following transfection experiments.
To identify cis elements responsible for the hypoxia-induced activation of the ROR␣4 promoter, we constructed a series of mutant DNAs lacking the 5Ј-upstream or 3Ј-downstream region and analyzed their promoter activities under hypoxia (Fig.  2). Although deletion of the region upstream from Ϫ397 resulted in a decrease of the basal transcription to about 40% that of the control, any deletion of the 5Ј-flanking region did not affect the activation of transcription by hypoxia. Deletion of a region downstream from ϩ273 also did not affect the hypoxiainduced activation of transcription, but further deletion to ϩ241 resulted in a loss of responsiveness to hypoxia. An identical result was obtained in the presence of Co 2ϩ (data not shown). Inspection of the nucleotide sequence from ϩ240 to ϩ273 revealed the existence of a 5Ј-CACGTAC-3Ј sequence, which conforms well to the consensus HRE, 5Ј-(C/G/T)ACGT(G/ C)C(G/T) (Fig. 3A). To determine whether the putative HRE found within the ROR␣4 promoter is indeed responsible for the transcriptional activation under hypoxia, a specific mutation was introduced within the sequence, and its effect on transcription was analyzed. Fig. 3B shows that the hypoxia-induced activation of transcription was markedly abrogated by the mutation within the HRE, indicating its direct involvement in the activation of the ROR␣4 promoter under hypoxia. Similar results were obtained in the presence of Co 2ϩ (data not shown).
Direct Involvement of HIF-1 in the Regulation of the ROR␣4 Promoter-Activation of hypoxia-inducible genes is mediated in large part by the binding of HIF-1 to the HRE. HIF-1␣ contains two transcriptional activation domains in its C-terminal half (31). We constructed a mutant form of HIF-1␣ lacking the Cterminal region from amino acid residue 393, which we used as a dominant negative form (HIF-1␣DN). The hypoxia-induced activation of the ROR␣4 promoter was almost completely inhibited by HIF-1␣DN (Fig. 3C). The transcription activated by the wildtype HIF-1␣ was also repressed efficiently in the cells cotransfected with HIF-1␣DN under normoxia. Fig. 4A shows proteins binding to the region containing the HRE. Several nuclear proteins from normoxic and hypoxic cells showed identical patterns of DNA binding. Cold probe DNA competed completely with the binding of all of the proteins, but a specific oligonucleotide containing the HRE failed to compete with the binding of any protein. However, an EMSA experiment carried out using the HRE oligonucleotide showed that proteins bound specifically to the HRE are induced under hypoxia. An excess amount of the cold wild-type HRE successfully competed with the binding, but a mutated HRE did not, demonstrating the specificity of the binding. Additionally, the HREbinding protein induced by hypoxia disappeared quickly after reoxygenation (data not shown). Another protein, which is present in both normoxic and hypoxic cells, was also found to bind specifically with the HRE. It is not clear at present whether this constitutive protein might be activating transcription factor or CREB (cAMP-response element-binding protein), as has been demonstrated previously (32).
To further confirm that HIF-1 binds to the ROR␣4 HRE, we performed a DNase I footprinting analysis using a probe from ϩ40 to ϩ341 with nuclear proteins from normoxic or hypoxic cells (Fig. 4C). Consistent with the result obtained in the EMSA experiments, no protein binding to the HRE region was detected with low amounts of nuclear proteins from either normoxic or hypoxic cells. However, the HRE did bind to protein when an increasing amount of the nuclear proteins was used, and the pattern differed between the hypoxic nuclear proteins and the normoxic nuclear proteins in that the HRE region protected by the normoxic nuclear proteins was relatively broad. Together with the results obtained in the transfection experiments, these observations indicate that the HIF-1 induced by hypoxia binds to the HRE and activates the transcription of the ROR␣4 promoter.
HIF-1-induced Activation Requires Three GC-rich Elements-Hypoxia-inducible genes often require cis elements in addition to the HRE for their efficient activation by hypoxia (33)(34)(35)(36). In addition to the HRE, three other regions (Box1, Box2, and  5, 11, and 15), a mutated ROR␣4 HRE oligonucleotide (lanes 12 and 16), or an Epo HRE oligonucleotide (lanes 13 and 17). Each competitor was used at a 30 M excess. Lanes 1, 6, or 9 contain the probe alone. Arrows indicate specific binding to the HRE probe. C, DNase I footprinting analysis of protein binding to the downstream region of the ROR␣4 promoter. Box3) bound with protein(s) that are abundant and expressed constitutively in both normoxic and hypoxic cells (Fig. 4, A and  C). The region between ϩ160 and ϩ250 is rich in GC and contains several GC-box-like sequences (Fig. 5), including a typical GC-box at ϩ210, two GGGAGG (GA-box) sequences at ϩ216 and ϩ234, and three GGGTGG sequences at ϩ171, ϩ191, and ϩ239. To address whether these three Boxes are involved in the hypoxia-induced activation of the ROR␣4 promoter, we analyzed the effects of their deletion on the HIF-1-dependent transcription. As shown in Fig. 6, deletion of Box3 (from ϩ223 to ϩ243) reduced the HIF-1-dependent transcription to about 50% of that seen with the wild-type DNA. Deletion of the region from ϩ202 to ϩ243 (deletion of the Box2 and Box3) resulted in a marked reduction of the transcription to about 20% of the control. A similar reduction in transcription was observed in a mutant lacking the region between ϩ155 and ϩ227 (deletion of the Box1 and Box2). In contrast, only a 50% reduction of the HIF-1-dependent transcription was detected on a mutant lacking the region from ϩ155 to ϩ215 (deletion of Box1 and Box2). Although the deletion from ϩ155 to ϩ215 removed both the GC-box and GA-box within the Box2, the ligation of an EcoRI linker created a new GGGAGG sequence in this reporter construct. The Box1 appeared to be dispensable, because deletion of the region between ϩ155 and ϩ178 did not significantly affect the HIF-1-dependent transcription.
Binding of Sp1 Family to the GC-rich Regions Adjacent to the HRE-Because the proteins bound to the three GC-rich boxes were constitutively expressed and appeared to be abundant (Fig. 4C), it is highly likely that they are Sp1 and/or its related proteins. To examine this, we performed competitive protein binding experiments with the same DNA probe used in the EMSA (shown in Fig. 4A). As shown in Fig. 7, among several protein-DNA complexes, at least three complexes disappeared specifically in the presence of an oligonucleotide containing either a GC-box or a GA-box. Moreover, these three complexes were recognized specifically by antibodies for either Sp1 or Sp3. The identical protein-DNA complexes were also detected with the hypoxic nuclear proteins (Fig. 4A and  data not shown).
The competitive binding of proteins was also analyzed using the hypoxic nuclear proteins in a DNase I footprinting experiment (Fig. 7C). Protein binding at all four sites was competed out by the addition of cold probe DNA, whereas the GC-box oligonucleotide competed only for the protein binding at Box2 and Box3 but not at Box1 and the HRE. The same results were observed with competition experiments using the GA-box oligonucleotide (data not shown). In contrast, the HRE oligonucleotide specifically inhibited HIF-1 binding at the HRE site. DISCUSSION In the present study we showed that the promoter of the ROR␣4 isoform is activated in response to hypoxia. We identified a hypoxia-responsive region downstream of the transcription initiation site that consists of an HRE and several GC-rich sequences. The HRE alone was not sufficient for the efficient activation of transcription of the ROR␣4 promoter by hypoxia; the activation was also dependent on at least two GC-rich sequences within Box2 and Box3 to which Sp1/Sp3 bind. The typical GC-box within Box2 appears to be the most important for the transcriptional activation. In contrast, the GC-box oligonucleotide did not compete for the protein binding to Box1, and no protein binding was observed at the GGGTGG sequence at ϩ191, indicating that Sp1/Sp3 does not bind to Box1. The characteristic features of the hypoxia-inducible region within the ROR␣4 promoter shares similarities with another hypoxiainducible gene, carbohydrase IX, in which the Sp1 family in concert with HIF-1 has been demonstrated to fully activate transcription (37). It is known that a functional and/or physical interaction between HIF-1 and another transcription factor is crucial for the transcriptional activation of the hypoxia-inducible genes. Several transcription factors are known to interact with HIF-1, including activating transcription factor/CREB-1 (cAMP-response element-binding protein) in the lactate dehydrogenase A gene (33,34), AP-1 in the VEGF gene (35), and HNF-4 in the erythropoietin gene (36). Together with the results obtained in the transfection experiments, these observations strongly suggest that the binding of Sp1 and/or Sp3 to at least two GC-rich sequences within Box2 and Box3 is essential for the HIF-1 dependent transcription of the ROR␣4 promoter. Our results, therefore, add members of the Sp1 family as another set of transcription factors cooperating with HIF-1 to achieve a high level of hypoxia-induced gene transcription.
The precise molecular mechanism of the interplay between HIF-1 and the Sp1 family is not known at present. The Sp1 family contains ubiquitously expressed transcription factors that frequently work in concert with other sequence-specific transactivators to control inducible promoters (38 -41). Functional and physical interactions between Sp1 and such factors often result in the synergistic activation of specific target promoters. Mutation of the HRE did not affect the basal transcription level of the ROR␣4 promoter under normoxia (data not shown), suggesting that the constitutive protein that specifically binds with the HRE (see Fig. 4B) might not be involved in regulating its transcription. Additionally, because deletion of any of the GC-rich sequences also did not result in a decrease in basal transcription (see Fig. 6), Sp1/Sp3 alone might not be involved in the transcriptional activation of this promoter under normoxia. Therefore, Sp1/Sp3 may help recruit HIF-1 to the promoter and/or may serve to better link the action of HIF-1 to the basal transcription machinery.
Initiation of transcription in eukaryotic cells is a complicated multistep process involving a large number of cofactors that act in the remodeling of chromatin and/or the recruitment of RNA polymerase II to the promoters of target genes (42,43). Of a dozen cofactors, CBP/p300 is a multidomain transcriptional cofactor that acts in conjunction with other factors to regulate transcription. Several lines of evidence show that HIF-1␣ recruits and interacts physically with CBP/p300 at its C-terminal transactivation domain (31, 44 -49). In contrast, it has been demonstrated that Sp1 does not physically interact with CBP/ p300 (50, 51), although they co-precipitate in the same complexes (52). Requirement of CBP/p300 for transcriptional synergism between Sp1 and other transcription activators has been also demonstrated (40). Therefore, it is possible that HIF-1 and Sp1/Sp3 might coexist in the same cofactor complex to synergistically activate transcription of the ROR␣4 promoter. Alternatively, Sp1/Sp3 might interact directly with HIF-1. Although in the present study we detected no direct, physical interaction between HIF-1 and Sp1, it is possible that some other factor stabilizing their interaction might be involved in the hypoxia-induced activation of the ROR␣4 promoter. It has recently been shown that a direct interaction between these two transactivators could be further stabilized by Smad3, which interacts both with Sp1 and HIF-1 (53). The existence of the GGGAGG sequence that is the constituent of the hypoxia-inducible region of the ROR␣4 promoter is also of interest. This sequence element has previously been identified as a binding site for ZBP-89, a Krü ppel-type zinc finger transcription factor. ZBP-89 binds efficiently to this element and activates or represses transcription, depending on the promoters of the target genes (50, 54 -57). Moreover, ZBP-89 has been shown to interact both with Sp1 and CBP/p300 (50). Although the present study showed that the major nuclear proteins bound with the GGGAGG sequence are Sp1/Sp3, a few protein complexes were observed that bound to the GA-box oligonucleotide but were not recognized by the antibodies for Sp1/Sp3 (see Fig. 4B). ZBP-89 might play a role in the collaboration of HIF-1 with Sp1 to activate the ROR␣4 promoter under hypoxia. Whether this is indeed the case requires further study.
To understand the possible roles of ROR␣ in cellular adaptation to hypoxia, it is essential to identify its target genes. Although there are several ROR␣-responsive element-containing genes that are expressed in the liver (7, 58 -60), little is known about their association with hypoxia. As described above, several lines of evidence show that ROR␣ has an antiinflammatory effect. Lipopolysaccharide-activated macrophages in the sg mice overproduce interleukin-1␤ and tumor necrosis factor ␣ (9), suggesting that ROR␣ negatively regulates expression of such proinflammatory cytokines to set in motion a mechanism to limit inflammation. Interestingly, ROR␣ has been suggested to interfere with the NF-B signaling cascade by reducing its nuclear translocation (61). NF-B plays a central role in a variety of cellular processes by regulating the transcription of a number of genes, including interleukin 1␤, IL-6, tumor necrosis factor ␣, and cyclooxygenase-2 (62,63). Recently, several stimuli other than oxygen tension have been shown to modulate HIF-1 expression and its consequent function. For example, interleukin 1␤ and tumor necrosis factor ␣ have been demonstrated to stimulate the DNA binding activity of HIF-1 and have also been demonstrated to increase expression of ROR␣ mRNA in endothelium and vascular muscle (19). The results obtained in the present study might support the possibility that the interleukin 1␤/ tumor necrosis factor ␣-induced increase of ROR␣ expression results from the activation of HIF-1. If this is the case, ROR␣ would be involved in a regulatory loop of proinflammatory cytokine production and would thereby play a protective role against inflammation and tissue injury associated with hypoxia. More studies are required to identify whether and how ROR␣ modulates the NF-B signaling cascade and consequently regulates transcription of the proinflammatory cytokine genes. Identification of genes whose expressions correlate with the increased expression of ROR␣ under hypoxia will help to define its function in the liver and in other tissues as well. Experiments along these lines are in progress.