To evaluate the expression and function of the orphan nuclear receptor Rev-ErbAα in articular cartilage and to investigate its role in osteoarthritis (OA).
To evaluate the expression and function of the orphan nuclear receptor Rev-ErbAα in articular cartilage and to investigate its role in osteoarthritis (OA).
Expression of Rev-ErbAα was analyzed at both the messenger RNA and protein levels in human and bovine articular cartilage and chondrocytes by real-time polymerase chain reaction (TaqMan) and immunocytochemical techniques. The effects of cartilage catabolic and anabolic agents on the expression of Rev-ErbAα were evaluated by TaqMan analysis. Overexpression was achieved by either adenoviral transduction or treatment with a peroxisome proliferator–activated receptor α agonist, whereas expression was suppressed by antisense oligonucleotides.
Among the 48 known nuclear receptors, Rev-ErbAα was found to be the most highly expressed in OA cartilage. It is known to function as a transcription repressor. Treatment of articular chondrocytes with known catabolic agents resulted in the induction of Rev-ErbAα, whereas stimulation with anabolic agents led to a decrease in expression. Overexpression of the nuclear receptor was associated with an increase in the expression of matrix-degrading enzymes such as matrix metalloproteinase 13 and aggrecanase. In contrast, a decrease in Rev-ErbAα expression led to a concomitant reduction in the activity of matrix-degrading enzymes.
This study is the first to demonstrate that Rev-ErbAα is highly expressed in OA articular chondrocytes and that its expression is modulated by known cartilage catabolic and anabolic stimuli. We also demonstrated that modulation of Rev-ErbAα expression in chondrocytes may be a novel means of regulating the expression and production of multiple matrix-degrading enzymes. These observations suggest that Rev-ErbAα may be a novel therapeutic target for OA.
Osteoarthritis (OA) is the most common form of arthritis, and it is often associated with significant disability and an impaired quality of life. Age, sex, mechanical stress, injury, genetic susceptibility, and metabolic predispositions are among the known risk factors for this disease. A hallmark of the disease is the degradation of articular cartilage, which is composed of an extracellular matrix that is rich in proteoglycan and type II collagen. This degradation is associated with increased synthesis and release of several catabolic factors, such as proinflammatory cytokines, matrix metalloproteinases (MMPs), and other proteases, into the tissues (1–5). These catabolic factors have been found to be elevated in both synovial fluid and cartilage from OA joints (6, 7). These enzymes cleave the native collagen and other matrix proteins, such as aggrecan, and convert proMMP zymogens into active forms.
Chondrocytes are the only cells found in cartilage tissue and are believed to be responsible for the elaboration of catabolic cytokines, such as interleukin-1 (IL-1) and tumor necrosis factor α (TNFα), and the subsequent release of matrix-degrading proteases (8). To date, the accumulated findings show that selective inhibition of IL-1, MMPs, and aggrecanases could reduce the progression of structural changes in experimental OA (5, 9–13). Thus, the modulation of these catabolic factors may lead to the identification of new therapeutic targets for the treatment of OA.
Recent data suggest that modulation of the activity of nuclear receptors such as peroxisome proliferator–activated receptor γ (PPARγ) may result in down-regulation of genes that encode catabolic factors. PPARγ has been shown to be expressed in chondrocytes and synoviocytes (14, 15), and PPARγ agonists can suppress the expression of inducible nitric oxide synthase and MMP-13 in human chondrocytes, as well as the expression of MMP-1 in human synovial fibroblasts (16, 17). Consistent with these findings, it has recently been reported that a PPARγ agonist, pioglitazone, exhibited chondroprotective effects in an experimental model of OA in the guinea pig (8, 18).
In this study, we identified another member of the nuclear receptor family, Rev-ErbAα, that is highly expressed in chondrocytes. It is known to be expressed in several tissues and cell types (19). Unlike other liganded nuclear receptors, Rev-ErbAα belongs to a subfamily of orphan receptors that are repressors of target gene transcription. Rev-ErbAα binds either as a monomer to AGGTCA response elements or as a homodimer to a direct repeat of this core motif separated by 2 nucleotides (Rev-DR2) and preceded at the 5′-end by an A/T-rich sequence (20, 21). It lacks the activation function 2 domain that is present at the carboxy-terminal of the ligand-binding domain of most nuclear receptors and acts as a transcription repressor through direct interaction with the nuclear receptor corepressor family of corepressor proteins (22). Its expression is significantly increased during adipogenesis and is down-regulated during muscle differentiation (23, 24). Furthermore, overexpression of Rev-ErbAα in myoblasts was shown to completely abolish differentiation, indicating a role of Rev-ErbAα in metabolic control and energy homeostasis (25).
Because of the high levels of expression of Rev-ErbAα in chondrocytes, we were interested in understanding its role in chondrocyte biology and in OA. We report herein the characterization of Rev-ErbAα expression and function in chondrocyte biology and its potential role in OA.
Adult human articular cartilage from patients undergoing knee replacement surgery for OA was obtained from the National Disease Research Interchange (Philadelphia, PA). Patients gave their informed consent before tissues were obtained, and the Institutional Review Board approved the study. Cartilage from which chondrocytes were isolated was primarily derived from the total knee, which included both the load-bearing and non–load-bearing regions. Bovine cartilage was obtained from the entire front knee of calves.
Human and bovine chondrocytes were isolated according to a method described previously (26). On average, the yield of human articular chondrocytes ranged from 6 × 106 to 6.5 × 106 cells/gm of tissue. The yield of bovine articular chondrocytes was better, ranging from 6 × 107 to 7 × 107 cells/gm of tissue. Human articular chondrocytes were plated in complete medium containing high-glucose Dulbecco's modified Eagle's medium (DMEM) and 10% heat-inactivated fetal bovine serum (FBS) on the same day of isolation. Bovine articular chondrocytes were plated in complete medium containing Ham's F-12 with 10% heat-inactivated FBS on the same day of isolation. Cells were cultured for 3–4 days and then lysed for the cell lysate preparations or were used in the cell-based assay without passaging.
For evaluating the effects of catabolic and anabolic factors on the expression of Rev-ErbAα, the cells were treated for 18–24 hours. The following catabolic factors were tested: IL-1, TNFα, and retinoic acid. The following anabolic factors were tested: insulin-like growth factor 1 (IGF-1) (27, 28), transforming growth factor β (TGFβ), bone morphogenetic protein 2 (BMP-2) and BMP-4 (29–32), hyaluronic acid (33), and dexamethasone.
Rabbit polyclonal antibodies were generated against a peptide of human Rev-ErbAα, RALRALVLKNRPLETSR. The peptide is located within the ligand-binding domain and was chosen based on solvent accessibilities, antigenicities, and uniqueness. For affinity purification of the Rev-ErbAα antibody, the peptide was immobilized on an active support. Antiserum was passed through the serum column, which was then washed. Specific antibodies were eluted via a pH gradient into neutralizing buffer, collected, and stored in borate buffer.
Total RNA was extracted from the treated cells using TRIzol reagent (Invitrogen, San Diego, CA) and additional purification was done using an RNeasy Mini kit (Qiagen, Chatsworth, CA). To remove any residual genomic DNA, each sample was treated with 10 units of DNase I (Ambion, Austin, TX). The efficiency of the DNase I procedure was validated in a standard TaqMan assay using RNA samples that had not been subjected to a reverse transcription step and a GAPDH primer set. The DNA-free total RNA samples were quantified using RiboGreen RNA quantitation reagent (Molecular Probes, Eugene, OR). First-strand complementary DNA (cDNA) synthesis was performed using a High Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA), and 300 ng of total RNA was converted for each sample.
An equivalent of 10 ng of total RNA was loaded into each well on the 384-well optical microplate, with technical replicates. Gene-specific primers were designed using Primer Express software (Applied Biosystems) and confirmed by BLAST searches using public and propriety sequence databases. The concentration of the forward and reverse primers and the probe for each assay was 900 nM, 900 nM, and 100 nM, respectively. Quantitative PCR was performed using a 7900HT Sequence Detector System (Applied Biosystems) in a 10-μl reaction volume. TaqMan Universal PCR Master Mix (2× concentrated; Applied Biosystems) and universal PCR conditions recommended by the vendor were used. Relative abundances were calculated, and expression levels of each gene were normalized using control genes GAPDH and cyclophilin.
Total RNA was extracted from the treated cells using TRIzol reagent, and additional purification was done using an RNeasy Mini kit. To remove any residual genomic DNA, each sample was treated with 10 units of DNase I. A total of 500 ng of RNA per sample was reverse transcribed using a random hexamer kit from Clontech (Palo Alto, CA). The transcribed product was amplified by PCR using gene-specific primers.
For detection of Rev-ErbAα, Western blot assays were performed. Total cell lysates were prepared by lysing cells in buffer containing 20 mM Tris HCl (pH 8.0), 0.1% sodium dodecyl sulfate (SDS), 100 mM NaCl, 1 mM Na3VO4, aprotinin (2 μg/ml), leupeptin (2 μg/ml), pepstatin (2 μg/ml), and 1 mM phenylmethylsulfonyl fluoride for 30 minutes at 4°C. Protein concentrations of chondrocyte lysates were determined by bicinchoninic acid protein assay (Bio-Rad, Richmond, CA), and lysates containing equal amounts of protein (20 μg) were separated by SDS–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The membranes were blocked in 3% milk, probed with anti–Rev-ErbAα, followed by a horseradish peroxidase (HRP)–conjugated anti-rabbit secondary antibody (Promega, Madison, WI). Immunoreactive proteins were detected by enhanced chemiluminescence (Amersham, Arlington Heights, IL).
Three different antisense oligonucleotides and 1 sense oligonucleotide were designed. The antisense oligonucleotides were as follows: first, 5′-TGGAGCAGGTACCATGTGATCC-3′; second, 5P-CGTTCCCTCGGCAGTAATATTTCAC-3′, and third, 5′-GACTGGGATTTGTAGTC-3′. The sense oligonucleotide was 5′-GGATCACATGGTACCTGCTCCA-3′. The oligonucleotides at different concentrations were transfected into cells in a 6-well plate using FuGene 6 transfection reagent (Roche Diagnostics, Indianapolis, IN). Cells were transfected at ∼75% confluency for 72 hours and were then lysed for RNA or protein preparations or were treated with 10 ng/ml of IL-1β for an additional 24 hours. All 3 antisense oligonucleotides suppressed the expression of Rev-ErbAα in human chondrocytes. However, maximum suppression was achieved by the second antisense oligonucleotide. Therefore, this antisense oligonucleotide was used to suppress the expression of Rev-ErbAα in all later experiments. The sense oligonucleotide had no effect on suppression.
Bovine articular chondrocytes in suspension were infected overnight with purified adenovirus containing the Rev-ErbAα gene construct at concentrations of 50 focus-forming units per cell. The infected cells were then incubated at 37°C for an additional 72 hours and were then lysed for RNA preparation.
The chondrocyte-based aggrecanase assay was run essentially as described elsewhere (34). Bovine chondrocytes in Ham's F-12 medium containing 10% heat-inactivated fetal calf serum (FCS) were grown on streptavidin-coated plates (NEN, Boston, MA) containing a C-terminal–biotinylated peptide substrate consisting of the aggrecanase-sensitive sequence NITEGE–ARGS. After 5 days, the cells were washed, and the medium was replaced with DMEM containing 0.5% FCS and the appropriate treatments (e.g., IL-1 or PPARα). After 20–24 hours, the cells were washed, lysed in PBS/0.05% Tween 20 (wash buffer), and the immunoreactive product ARGS generated by aggrecanase-mediated cleavage was detected by enzyme-linked immunosorbent assay (ELISA). The OA-1 monoclonal antibody (anti-ARGSVIL; 1 μg/ml) in 1% BSA/PBS/Tween 20 (0.05%) was added to each well and incubated for 1 hour at 22°C. The solution was removed, and the plate was washed 6 times with wash buffer. Anti-ARGSVIL antibody that adhered to the immobilized biotinylated product was detected using HRP-conjugated goat anti-mouse IgG (Calbiochem, La Jolla, CA) in 1% BSA/PBS/Tween 20 (0.05%). The levels of HRP were determined using tetramethylbenzidine (DakoCytomation, Carpinteria, CA), and absorbance was determined at 450 nm following acidification with 1N H2SO4.
For FuGene 6–mediated transfection experiments, the cells were cultured for 4 days. On the fourth day, cells were transiently transfected with FuGene 6 in serum-free medium. Five hours later, serum and IL-1 were added. The rest of the assay was performed as described above.
For in situ hybridization, cryosections of nonfixed, nondecalcified adult human articular knee cartilage from 10 OA patient samples were hybridized with 500 bp of riboprobes for Rev-ErbAα that had been labeled with digoxigenin-UTP by in vitro transcription. Antisense strands provided the positive probe, and sense strands were used as a negative control. Hybridization was detected with alkaline phosphatase–conjugated antidigoxigenin antibody. Antibody was detected with nitroblue tetrazolium/BCIP, resulting in a blue precipitate over positive cells. Cryosections were counterstained with nuclear fast red.
For immunocytochemical analyses, human articular chondrocytes were analyzed for Rev-ErbAα protein expression using the antibody described above. An alkaline phosphatase–based immunoenzymatic method using HRP/diaminobenzidine chromogen was performed according to the manufacturer's instructions (LSAB kit; DakoCytomation). Antibody binding to human articular chondrocytes was indicated by a brown precipitate. Blocking of immunizing peptide was used as a negative control.
Rev-ErbAα was previously identified as one of the most abundant genes in an analysis of cDNA libraries prepared from RNA isolated from OA cartilage (35). Additionally, among the different tissues in which Rev-ErbAα is expressed, the expression levels are highest in cartilage. We sought first to confirm the expression of Rev-ErbAα in cartilage. In our analysis of all 48 known nuclear receptors in human OA cartilage, the levels of expression of Rev-ErbAα were the highest (Figure 1A). Among the remaining nuclear receptors, the levels of Rev-Erbβ, retinoid X receptor α (RXRα), RXRβ, retinoic acid receptor γ, and glucocorticoid receptor expression were also notable. The high level of Rev-ErbAα expression was also confirmed in human OA articular chondrocytes (Figure 1B).
Although Rev-ErbAα was expressed in bovine articular chondrocytes, the level of expression was much lower than that in human OA articular chondrocytes (Figure 1B). Bovine articular chondrocytes were derived from young calves, whereas the human articular chondrocytes were derived from mature human donors with OA; thus, the higher expression in human chondrocytes could be attributable to the age and physiologic/pathologic condition of the chondrocytes.
The expression of Rev-ErbAα messenger RNA (mRNA) and protein was confirmed by immunochemical techniques in cartilage derived from human donors. The expression of Rev-ErbAα was detected in chondrocytes of the middle zone of OA cartilage (Figure 1C) as compared with the deep and superficial zones (results not shown). Human OA cartilage chondrocytes also expressed Rev-ErbAα protein, as demonstrated by immunohistochemistry using anti–Rev-ErbAα antibody. The specificity of the findings was demonstrated by the absence of staining in the presence of immunizing peptide (Figure 1D).
Chondrocytes produce mediators associated with inflammation and degradation, including cytokines such as IL-1 and proteolytic enzymes such as aggrecanases and MMPs that can cause degradation of cartilage matrix (4, 6). Therefore, we examined the effect of several catabolic agents on Rev-ErbAα expression in primary bovine articular chondrocytes. The expression of Rev-ErbAα was up-regulated in bovine chondrocytes by the 3 catabolic stimuli we tested (Figure 2A). Since IL-1 is one of the key catabolic factors of cartilage, we confirmed the induction of Rev-ErbAα by IL-1 at the protein level by Western blot analysis (Figure 2B). The results demonstrated that the known catabolic factors of cartilage increased the expression of Rev-ErbAα in chondrocytes. A rather high basal expression of Rev-ErbAα in human articular chondrocytes precluded a similar analysis in the OA articular chondrocytes.
PPARα agonists have previously been shown to up-regulate Rev-ErbAα expression in human hepatocytes (36). To assess whether PPARα plays a role in regulating Rev-ErbAα in chondrocytes, we evaluated the effect of a small-molecule PPARα agonist (SB692926, a proprietary compound of GlaxoSmithKline) on the expression of Rev-ErbAα in human chondrocytes. The 50% effective concentration (EC50) of this agonist is ∼300 nM for PPARα. Similar to the other catabolic agents, the PPARα agonist caused a modest increase in Rev-ErbAα expression (Figure 2C).
In our experience, bovine chondrocytes derived from young calves consistently exhibit low basal levels of Rev-ErbAα that precluded measurement of any further decrease. In contrast, human chondrocytes derived from aged donors consistently exhibit a much higher level of basal expression, which allows us to test any inhibitory effect on Rev-ErbAα expression. Thus, we tested the effect of known anabolic growth factors of cartilage on human OA articular chondrocytes.
The expression of Rev-ErbAα exhibited a trend toward reduction with all of the factors we tested (Figure 3). The reduced expression was more prominent in samples treated with anabolic growth factors, such as IGF-1 and TGFβ. These anabolic factors had minimal effects on the levels of expression of Rev-ErbAα in bovine articular chondrocytes (data not shown). Thus, our findings showed that Rev-ErbAα expression is modulated by anabolic and catabolic factors of cartilage.
The expression of certain MMPs and other inflammatory molecules is elevated in diseased cartilage (4, 6) and, in general, the high expression correlates with their increased activity in cartilage matrix degradation. Since the expression of Rev-ErbAα was observed to be elevated by catabolic stimuli, we examined whether overexpression of this protein would result in increased expression of MMPs and aggrecanases. Overexpression of Rev-ErbAα was achieved by transducing bovine articular chondrocytes with purified adenovirus particles engineered to contain the gene encoding Rev-ErbAα. As shown in Figure 4A, an ∼3-fold increase in Rev-ErbAα mRNA expression led to a significant increase in MMP-13 (∼150-fold), aggrecanase 1 (∼200-fold), and aggrecanase 2 (∼5-fold) expression.
Using a chondrocyte-based enzymatic assay (34), we next evaluated whether increasing the expression levels of Rev-ErbAα would lead to a concomitant increase in aggrecanase activity in chondrocytes. This assay is based on an immobilized aggrecanase-specific peptide substrate that is cleaved by aggrecanase that is produced by the chondrocyte, resulting in the production of a neoepitope that is recognized by a specific antibody in an ELISA. Consistent with the RNA data shown in Figure 2, treatment of bovine articular chondrocytes with the PPARα agonist caused an increase in the expression of Rev-ErbAα at the protein level (Figure 4B).
We also tested the effect of overexpression of Rev-ErbAα mediated by PPARα agonists on aggrecanase activity. Similar to the increase in the aggrecanase activity achieved by IL-1 treatment, PPARα agonist–mediated overexpression of Rev-ErbAα also resulted in an increase in aggrecanase activity in bovine articular chondrocytes (Figure 4C). Furthermore, treatment with the PPARα agonist in combination with IL-1 resulted in a greater increase in aggrecanase activity over that induced by IL-1 alone. These findings suggest that increasing the expression of Rev-ErbAα leads to a corresponding increase in the expression of mRNA and the activity of proteases, such as aggrecanases, that are involved in cartilage degradation.
The preceding experiment indicated that overexpression of Rev-ErbAα leads to an increase in the expression of matrix-degrading proteases. To determine the effects of decreased Rev-ErbAα expression, we used antisense technology to suppress Rev-ErbAα expression and examined its effect on the mRNA expression of MMP-3, MMP-13, aggrecanase 1, and aggrecanase 2. Treatment with antisense oligonucleotide directed against Rev-ErbAα led to a ∼3.5-fold reduction in Rev-ErbAα expression (Figure 5A). A corresponding reduction in the levels of expression of MMPs and aggrecanases was observed (range 4–15-fold).
As shown in Figure 2, IL-1 induced the expression of Rev-ErbAα in bovine articular chondrocytes. Therefore, we tested the effect of antisense oligonucleotide on the expression Rev-ErbAα induced by IL-1 in bovine articular chondrocytes. IL-1 induction of Rev-ErbAα at the protein level was suppressed by treatment with antisense oligonucleotide (Figure 5B). We next tested whether a decrease in Rev-ErbAα expression produced by the antisense oligonucleotide would also lead to a concomitant reduction in aggrecanase activity. A reduction in the levels of expression of Rev-ErbAα achieved by antisense oligonucleotide led to a significant reduction in IL-1–induced aggrecanase activity. As expected, this effect was not observed with the sense oligonucleotide (Figure 5C). These findings indicate that modulating Rev-ErbAα expression affects chondrocyte biology by altering the expression and activity of genes responsible for cartilage catabolism.
In the present study, we characterized Rev-ErbAα and its role in the biology of cartilage. While Rev-ErbAα has been known for many years, this is the first study to demonstrate its potential role in OA. Unlike several other nuclear receptors, Rev-ErbAα belongs to a subfamily of orphan receptors that are repressors of target gene transcription (21). Among the 48 known nuclear receptors, the expression of Rev-ErbAα was the highest in human cartilage. This unique expression pattern prompted us to further examine its biology and association with diseased cartilage.
Our studies suggest that the constitutive expression of Rev-ErbAα is higher in OA chondrocytes and cartilage isolated from mature adults undergoing knee joint replacement, as compared with normal chondrocytes derived from immature bovine calves. The expression of Rev-ErbAα in cartilage was down-regulated by anabolic stimuli and is up-regulated by catabolic stimuli. Furthermore, overexpression or down-regulation of this nuclear receptor resulted in changes in the expression of the matrix-degrading proteases and that, in turn, affected their activity. These findings suggest that modulation of the expression, and perhaps the activity, of Rev-ErbAα may play an important role in pathogenesis of OA (Figure 6).
However, it is still unclear whether the effect of these catabolic or anabolic agents on Rev-ErbAα expression is direct. Rev-ErbAα expression has been reported to be regulated by a variety of stimuli, and the molecular mechanisms that regulate its expression have been partly defined. Rev-ErbAα represses its own expression by binding to the Rev-DR2 element on its promoter (20). In contrast, in human hepatocytes, hyperlipidemia drugs of the fibrate class have been shown to up-regulate its expression (36). This induction of Rev-ErbAα gene expression occurs at the level of transcription in hepatocytes and is mediated by PPARα. PPARα induces Rev-ErbAα expression by interfering with the negative autoregulatory loop of Rev-ErbAα expression via the Rev-DR2 site. We have shown herein that a PPARα agonist also increases the expression of Rev-ErbAα in chondrocytes. Since Rev-ErbAα is a repressor of gene transcription, it may regulate the expression of MMPs and aggrecanases indirectly. It could enhance the expression of aggrecanases and MMPs by interfering negatively with the expression or activity of a transcription factor that represses MMP and aggrecanase expression.
The homology modeling studies published for Rev-ErbAα indicate that the putative ligand-binding cavity is occupied by side chains, suggesting that this receptor may not have any endogenous ligands (37). However, recent findings by Reinking et al (38) demonstrate that the ligand-binding pocket of the Drosophila ortholog of Rev-ErbAα, E75, contains a heme prosthetic group that is bound to heme. Furthermore, the oxidation state of the heme iron determined the interaction of E75 with the corepressor DHR3, the Drosophila ortholog of retinoic acid–related orphan receptor α, and this interaction was shown to be regulated by the binding of nitric oxide to the heme center. Since nitric oxide has been implicated in the pathogenesis of OA (39), this finding is very intriguing to us and raises the possibility that heme could be a natural ligand for mammalian Rev-ErbAα.
In summary, we have demonstrated that modulation of the expression of Rev-ErbAα in chondrocytes may be a novel means by which to regulate the expression and production of multiple matrix-degrading enzymes. The exact molecular mechanism by which Rev-ErbAα regulates the expression and activity of these proteases awaits further studies.
The authors wish to acknowledge the expert technical assistance of Hu Do, Feilan Wang, and Tonie Newman-Tarr. Our thanks go to Dr. Ganesh Sathe for designing the antisense oligonucleotides. Thanks to Joan Stuart for the real-time PCR experiments. Critical reading of the manuscript by Dr. John Emery and Bart Votta was much appreciated.