Osteoarthritis and Cartilage (2009) 17, 832e842

ª 2009 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.joca.2008.12.011

InternationalCartilageRepairSociety

Global analysis of nuclear receptor expression and dysregulation inhuman osteoarthritic articular cartilage

Reduced LXR signaling contributes to catabolic metabolismtypical of osteoarthritisL. A. Collins-Raciey, Z. Yangz, M. Araiy, N. Liz, M. K. Majumdarza, S. Nagpalzx, W. M. Mountsy,A. J. Dornery, E. Morrisz and E. R. LaValliey*yThe Departments of Biological Technologies, Cambridge, MA 02140, United StateszWomen’s Health and Musculoskeletal Biology, Wyeth Research, Cambridge, MA 02140, United StatesxWomen’s Health and Musculoskeletal Biology, Wyeth Research, Collegeville, PA 19426, United States

Summary

Objective: Compare the expression and regulation of nuclear receptors (NRs) in osteoarthritic and normal human articular cartilage.

Method: The transcriptional levels of 48 NRs and additional related proteins were measured in mRNA from human articular cartilage fromsubjects with osteoarthritis (OA) and compared to samples from subjects without OA, using microarrays, individual quantitative reversetranscriptase polymerase chain reaction assays, and a custom human NR TaqMan� Low Density Array (TLDA). The functional effect of liverX receptor (LXR) activity in cartilage was studied by measuring proteoglycan (PG) synthesis and degradation in articular cartilage explantcultures following treatment with the synthetic LXR agonist T0901317.

Results: Thirty-one of 48 NRs analyzed by TLDA were found to be measurably expressed in human articular cartilage; 23 of these 31 NRsshowed significantly altered expression in OA vs unaffected cartilage. Among these, LXRa and LXRb, and their heterodimeric partners retinoidX receptor (RXR)a and RXRb were all expressed at significantly lower levels in OA cartilage, as were LXR target genes ABCG1 and apoli-poproteins D and E. Addition of LXR agonist to human OA articular chondrocytes and to cartilage explant cultures resulted in activation ofLXR-mediated transcription and significant reduction of both basal and interleukin (IL)-1-mediated PG degradation.

Conclusions: Articular cartilage expresses a substantial number of NRs, and a large proportion of the expressed NRs are dysregulated in OA.In particular, LXR signaling in OA articular cartilage is impaired, and stimulation of LXR transcriptional activity can counteract the cataboliceffects of IL-1. We conclude that LXR agonism may be a possible therapeutic option for OA.ª 2009 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Key words: Osteoarthritis, Nuclear receptor, Liver X receptor, Retinoid X receptor, Gene expression, Proteoglycan.

Introduction

The nuclear receptor (NR) superfamily consists of a largegroup of ligand-inducible or ‘‘orphan’’ transcriptional regula-tor proteins. To date, 49 members have been identified inthe human genome1. NRs are a very ‘‘druggable’’ class ofproteins; synthetic ligand analogs have been produced fornearly all NRs with known natural ligands2. Since NRs mod-ulate diverse cellular functions associated with various dis-eases (cancer, osteoporosis, diabetes, etc.), they constitutean important group of candidate pharmacologicaltargets3e9.

Osteoarthritis (OA) is a highly prevalent, progressive, de-bilitating disease characterized by gradual destruction of ar-ticular cartilage. Chondrocytes constitute the primary cell

aPresent address: GlaxoSmithKline, 1250 S Collegeville Road,Collegeville, PA 19426, United States.

*Address correspondence and reprint requests to: Dr EdwardR. LaVallie, Department of Biological Technologies, WyethResearch, 35 CambridgePark Drive, Y1106 Cambridge, MA02140, United States. Tel: 1-617-665-7068; Fax: 1-617-665-7240;E-mail: elavallie@wyeth.com

Received 2 April 2008; revision accepted 20 December 2008.

832

type in articular cartilage and they are responsible for the syn-thesis and maintenance of the cartilage matrix that surroundsthe cells. This matrix provides cartilage with the biophysicalcharacteristics that are important for the tissue to act as aneffective articulating surface. Chondrocytes maintain properhomeostasis of cartilage tissue by balancing expression andresponse to anabolic and catabolic factors. In OA this bal-ance appears improperly skewed toward catabolism, result-ing in gradual loss of cartilage extracellular matrix. Since NRsare important transcriptional regulators within cells, dysregu-lation of NR expression in OA could provide important cluesinto the pathophysiology of the disease as well as providepossible points of therapeutic intervention.

Despite the recognized importance of NRs in the transcrip-tional regulation of various cellular pathways, to date onlya few NRs have been suggested to potentially play a rolein OA, including estrogen receptor (ER) alpha (ERa,NR3A1)10e19, vitamin D receptor (VDR, NR1I1)20e22, peroxi-some proliferator-activated receptor gamma (PPARg,NR1C3)23e26, glucocorticoid receptor (GR, NR3C1)27,28, andthe orphan NRs NURR1 (NR4A2)29 and Rev-ErbAa(NR1D1)30. Here we describe an evaluation of the expressionof 48 human NRs in articular cartilage from OA patients

833Osteoarthritis and Cartilage Vol. 17, No. 7

compared to unaffected individuals. These experiments re-vealed that articular cartilage expresses many NRs and severalof these are dysregulated in osteoarthritic cartilage compared tonormal cartilage, including most of those mentioned above withprior association with OA. Expression of liver X receptor (LXR)a(NR1H3) and LXRb (NR1H2) along with LXR heterodimericpartners retinoid X receptor (RXR)a (NR2B1) and RXRb(NR2B2) was found to be significantly reduced in human OAcartilage compared to normal. This deficit in LXR and RXR ex-pression in OA cartilage was accompanied by decreased ex-pression of LXR-responsive genes. Treatment of humanarticular cartilage explants in vitro with an LXR agonist in-creased LXR target gene expression and reduced cytokine-in-duced proteoglycan (PG) degradation. We conclude thatcorrection of an LXR signaling defect in OA cartilage by treat-ment with an LXR agonist may be a novel disease-modifyingtherapeutic option for treatment of OA.

Experimental procedures

MEDIA, CHEMICALS AND REAGENTS

Cell culture reagents were obtained from Gibco-BRL(Grand Island, NY, USA). Interleukin (IL)-1b, tumor necrosisfactor (TNF)a, and oncostatin M (OSM) were purchasedfrom R&D Systems (Minneapolis, MN, USA). LXRagonist T0901317 [N-(2,2,2-trifluoro-ethyl)-N-[4-(2,2,2-triflu-oro-1-hydroxy-1-trifluoromethyl-ethyl)-phenyl]-benzenesulf-onamide]31,32 was prepared following standard chemicalsyntheses from published literature. Human Universal Ref-erence Total RNA (catalog #636538) was purchased fromClontech (Mountain View, CA, USA). Dimethylmethyleneblue (DMMB) was purchased from Polysciences Inc. (War-rington, PA, USA). Shark and whale chondroitin sulfate wasobtained from Fluka Biochemika (Buchs, Switzerland). Pro-teinase K was from Roche Molecular Biochemicals(Indianapolis, IN, USA). Fresh human OA and normal artic-ular cartilage for cell culture experiments were obtainedfrom the National Disease Research Interchange (Philadel-phia, PA, USA).

ISOLATION OF RNA FROM PRIMARY CARTILAGE TISSUE AND

FROM CHONDROCYTES IN CULTURE

RNA was isolated from human osteoarthritic articular carti-lage samples obtained from patients (n¼ 20, meanage¼ 66.2 years, range 49e84 years) undergoing totalknee replacement surgery (New England Baptist Hospital,Boston, MA, USA), or from non-osteoarthritic cartilage ob-tained from above-knee amputations (n¼ 10, meanage¼ 71.6 years, range 43e100) (Clinomics, Pittsfield, MA,USA). The OA cartilage samples were obtained as wholejoints within 2 h of surgery, and the articular cartilage wasshaved from the joint surfaces taking great care to avoidany pannus, fibrotic tissues, subchondral bone, and othernon-cartilaginous regions of the joint. Non-osteoarthritic(‘‘normal’’) cartilage samples were obtained from individualswithout a clinical diagnosis or symptoms of OA, and thespecimens were evaluated histologically to confirm the clas-sification prior to inclusion in this study. Cartilage pieceswere flash-frozen in liquid nitrogen and stored at �80�C untilprocessed for RNA isolation. RNA was prepared from the fro-zen cartilage as described previously33. Isolation of RNA fromchondrocytes in monolayer culture was performed using TRI-zol reagent (Invitrogen) according to the manufacturer’sprotocol.

CHONDROCYTE CELL CULTURE

Chondrocytes were isolated from fresh human articularcartilage using a standard method previously described33.Cells were cultured in 10% fetal bovine serum (FBS) con-taining Dulbecco’s modified Eagle medium (DMEM)/F12growth media for 2e3 days in 12 well culture plates at a den-sity of 1e2� 10E6 cells/well. Chondrocyte cultures werestimulated with cytokines (TNFa 10 ng/ml, IL-1b1 ng/ml) for 18 h or with LXR agonist (T0901317, 2 mM finalconcentration) for 26 h, in triplicate wells.

MEASUREMENT OF mRNA CHANGES IN OSTEOARTHRITIC

AND NORMAL CARTILAGE USING MICROARRAYS

Global gene expression changes in RNA from osteoar-thritic cartilage (n¼ 20) as well as from non-osteoarthriticcartilage (n¼ 10) samples described above were analyzedusing the Human Genome HG-U95Av2 GeneChip� Array(Affymetrix, Santa Clara, CA, USA), as describedpreviously33.

cDNA SYNTHESIS AND QUANTITATIVE REVERSE

TRANSCRIPTASE POLYMERASE CHAIN REACTION (QRT-PCR)

cDNA was prepared from purified RNA using the High-Capacity cDNA Archive Kit (Applied Biosystems, catalog#4322171) according to the manufacturer’s instructions.Quantitative real-time PCR utilized either human TaqMan�

Gene Expression Assays or TaqMan� Low Density Arrays(TLDAs) from Applied Biosystems. Thermal cycling wasperformed using either an ABI Prism 7900 Sequence De-tection System (for TLDA) or an ABI Prism 7700 SequenceDetection System (for TaqMan� Gene Expression Assays).RNA for TaqMan� analysis was purified from dissected andfrozen cartilage tissue as described above, followed by twomore rounds of phenol/chloroform extraction followed byRNeasy (Qiagen) column purification. RNA was treatedwith DNase (Qiagen) during RNeasy column purification,and again following the RNA purification using DNA-free(Ambion, Austin, TX, USA), following the manufacturer’s in-structions. Human Universal RNA (Clontech) was used togenerate standard curves for each assay. Pre-designedTaqMan� probe/primer assay sets (Gene Expression As-says, Applied Biosystems) for individual qRT-PCR assess-ments were obtained for the following LXR target genes:Abca1 (Hs01059122_m1), Abcg1 (Hs00245154_m1),Apoe (Hs00171168_m1), stearoyl-CoA desaturase (Scd)(Hs00748952_s1), sterol regulatory element binding tran-scription factor 1 (Srebf1) (Hs00231674_m1) along withthe endogenous control glucuronidase beta (Gusb)(4333767F).

A TLDA was custom designed and ordered from AppliedBiosystems, containing Gene Expression Assays capableof measuring mRNA for: 48 of the 49 known human NRs(NR2A3/Hnf4b was omitted due to lack of an available as-say); NR co-repressors (Ncor1 and Ncor2 known LXR tar-get genes [Abcg1, apolipoproteins D and E (Apod andApoe)]; endogenous control normalizer genes (18S rRNA,B2m, Gapdh, Gusb, Hprt1, Pgk1, Ppia, Rplp0, Tbp, Tfr);and a gene known to be overexpressed in OA articular car-tilage (Prss11/Htra1)34,35 as a positive control. The catalognumbers for these assays are listed in Table I. The assayprimer/probe sets were preloaded in triplicate into each ofthe 384 wells in a ‘‘Format 64’’ TLDA. Each port on theTLDA (four ports total per TLDA) received 100 ml TaqMan�

reaction consisting of cDNA corresponding to 75 ng of total

834 L. A. Collins-Racie et al.: NR expression in OA cartilage

RNA combined with 1� TaqMan� Universal Master Mix.The comparative CT method of relative quantification36 us-ing Gusb as normalizer compared to the CT value of the tar-get gene (DCT) was used. Relative quantification (RQ, orfold-change) between different sample groups (e.g., OAvs normal) was then determined according to the 2�DDCT

method36, where DDCT¼DCT treated sample�DCT con-trol sample(s). The mean of the expression values for allof the normal samples (n¼ 10) was used as the calibratorfor these calculations.

CARTILAGE PG ANALYSIS

Fresh cartilage explants (w15 pieces, a total ofw165 mg/well) from human OA donors were cultured for10 days in 1 ml of DMEM/F12 containing 1% Nutridoma�

(Roche Applied Science, Indianapolis, IN, USA). Duringthe 10 days, the explants were incubated with cytokines(1 ng/ml IL-1bþ 5 ng/ml OSM) with or without LXR agonistT0901317 (5, 10, 20 mM). Culture medium was replacedevery 2 days with fresh cytokines and LXR agonist. Releaseof PGs was measured in these cultures using a DMMBassay37. Possible cytotoxic effects of cellular treatmentswere assessed by measuring lactate levels in culture mediaas an indicator of cellular metabolism and viability usinga kit from Trinity Biotech (Berkeley Heights, NJ, USA).Explants were then digested with proteinase K and assayedfor total PG content.

STATISTICAL ANALYSIS

Data are expressed as mean� standard error of themean (S.E.M.). Statistical significance was determined bytwo-tailed Welch t test using either Expressionist Software(Genedata, Basel, Switzerland) or Microsoft Excel 2000;for either test, P< 0.05 was set as the significancethreshold.

Results

EXPRESSION AND DYSREGULATION OF NRs IN HUMAN OA

ARTICULAR CARTILAGE

A focused analysis to identify the spectrum of expres-sion of nuclear hormone receptors in global gene expres-sion data from OA vs normal articular cartilage33 wasundertaken. The mean transcript levels of NRs repre-sented on the gene chips and judged present in cartilageare listed in Table I. Although relatively few NRs were ex-pressed in cartilage at levels measurable on the genechips, these data support a previous report30 that Rev-Er-bAa (NR1D1) is one of the most abundant NRs in articularcartilage. In addition, LXRb and RXRa were also found tobe relatively highly expressed in articular cartilage, raisinginterest in a possible role for LXR signaling in cartilage.Preliminary qRT-PCR performed on cartilage RNA froma small subset of the donors profiled by gene chip con-firmed LXRa, LXRb, and RXRa expression, and sug-gested that these NRs may be downregulated in OAcartilage (data not shown).

EVALUATION OF NR REGULATION IN OA CARTILAGE USING

A CUSTOM TLDA

The preliminary analysis of NR expression providedimpetus to expand the study beyond the use of gene chips

(limited by sensitivity) and individual qRT-PCR assays (lim-ited by the amount of cartilage RNA from OA and normalhuman donors) to a methodology that would allow compre-hensive measurement of all human NRs with limitedamounts of RNA and with high sensitivity and precision.To accomplish this, a custom human NR TLDA (hNR-TLDA) was created containing assays for 48 of the 49known human NRs (Table I; no assay was available forNR2A3/HNF4b), two NR co-repressors (NCOR1 andNCOR2), 10 endogenous control ‘‘normalizer’’ genes,and a gene (PRSS11) known to be robustly expressedand upregulated in OA cartilage from the gene chip data[Fig. 1(A)]. In addition, assays were included to measureexpression of three known LXR target genes (ABCG1,APOD, and APOE)38e41 to determine whether the putativedownregulation of LXR expression in OA cartilage wouldbe reflected in the expression of genes directly regulatedby LXR. All of the NR assays were capable of detectingexpression in amplification reactions containing 75 ng ofuniversal RNA with CT values< 34, with the exception ofthe photoreceptor-specific nuclear receptor (PNR) assay(presumably due to under-representation of the highly tis-sue-specific PNR in the universal RNA42). The design ofthe hNR-TLDA incorporated triplicate assays for eachgene. Intra- and inter-assay reproducibility was also testedusing the universal RNA control and found to be acceptablewhen CT values were less than 34 (data not shown).

RNA from normal (n¼ 10) and OA cartilage (n¼ 20) sam-ples was assayed using this hNR-TLDA (75 ng RNA perport). The choice of GUSB as the endogenous control nor-malizer gene was based upon a post hoc analysis of vari-ability in the expression of the 10 endogenous controlgenes across all cartilage samples and because its expres-sion level (CT value) was similar to most of the NRs understudy (Table I).

PRSS11 and APOD were the only two genes repre-sented on the TLDA that were expressed highly enoughin both OA and normal cartilage to test the concordanceof hNR-TLDA data to microarray data in our sample set.Average expression of PRSS11 in each cohort measuredby either the HG-U95Av2 GeneChip [Fig. 1(A)] or thehNR-TLDA [Fig. 1(B)] provided very similar results thatconfirm upregulation of this gene in OA articular carti-lage34,35. For APOD, the microarray data [Fig. 1(C)] andthe TaqMan data [Fig. 1(D)] both indicated a significant re-duction in APOD expression in OA cartilage. However, thefold-differences in APOD gene expression between OAand normal cartilage did not closely agree between thetwo technologies, presumably because the average ex-pression value for APOD in OA cartilage by microarraywas at the lower limit of reliable detection [w10 parts permillion (ppm)] while TaqMan expression values forAPOD in OA cartilage were well within the range of quan-titation (average CT w 25).

Thirty-one of the 48 NRs assayed were found to be mea-surably expressed by human articular cartilage using thehNR-TLDA (Table I). Genes for which the particular assaysdid not reach a CT value< 34 when data from all subjects(OA and normal) were averaged were considered not to bereliably expressed and were excluded from further analysis.In addition, data from the assay for NR1I1 (VDR) were highlyvariable between subjects and were also excluded. Themean expression values for all normal articular cartilagesamples were used as the calibrator and GUSB was usedas the normalizer for this analysis. A list of the NRs and theirexpression values in normal articular cartilage (expressed asaverage CT) are shown in Table I. Of the 31 NRs reliably

Table IGene Expression Assay content and cartilage expression data for genes on Affymetrix GeneChips and the human custom NR-TLDA

Gene symbol Gene name Alias GeneChip� averagesignal (ppm)

TaqMan�

assayTaqMan�

average CT

Welch t test

NRs OA vs normal

NR0B1 NR subfamily 0, group B,member 1

DAX1 ND Hs00230864_m1 37.56* No value

NR0B2 NR subfamily 0, group B,member 2

SHP1 ND Hs00222677_m1 40.00* No value

NR1A1 Thyroid hormone receptor,alpha

TRa ND Hs00268470_m1 26.99 0.012

NR1A2 Thyroid hormone receptor,beta

TRb ND Hs00230861_m1 27.25 0.529

NR1B1 Retinoic acid receptor, alpha RARa ND Hs00230907_m1 27.22 1.2E L 07NR1B2 Retinoic acid receptor, beta RARb ND Hs00233407_m1 28.05 0.019NR1B3 Retinoic acid receptor, gamma RARg ND Hs00171273_m1 26.85 0.049NR1C1 Peroxisome proliferator-activated

receptor alphaPPARa ND Hs00231882_m1 28.97 0.033

NR1C2 Peroxisome proliferator-activatedreceptor delta

PPARd ND Hs00606407_m1 28.32 2.2E L 06

NR1C3 Peroxisome proliferator-activatedreceptor gamma

PPARg ND Hs00234592_m1 31.24 0.003

NR1D1 NR subfamily 1, group D,member 1

Rev-Erba 20.6 Hs00253876_m1 24.27 0.010

NR1D2 NR subfamily 1, group D,member 2

Rev-Erbb ND Hs00233309_m1 27.25 0.137

NR1F1 RAR-related orphan receptor A RORa ND Hs00536545_m1 27.64 0.288NR1F2 RAR-related orphan receptor B RORb ND Hs00199445_m1 33.27* No valueNR1F3 RAR-related orphan receptor C RORg ND Hs00172860_m1 28.45 0.308NR1H2 NR subfamily 1, group H,

member 2LXRb 9.3 Hs00173195_m1 28.37 0.001

NR1H3 NR subfamily 1, group H,member 3

LXRa ND Hs00172885_m1 29.40 0.032

NR1H4 NR subfamily 1, group H,member 4

FXR ND Hs00231968_m1 39.50* No value

NR1I1 Vitamin D (1,25-dihydroxyvitaminD3) receptor

VDR ND Hs00172113_m1 32.42y No value

NR1I2 NR subfamily 1, group I,member 2

PXR ND Hs00243666_m1 31.95 0.010

NR1I3 NR subfamily 1, group I,member 3

CAR ND Hs00231959_m1 37.21* No value

NR2A1 Hepatocyte nuclear factor 4,alpha

HNF4A ND Hs00230853_m1 35.17* No value

NR2A2 Hepatocyte nuclear factor 4,gamma

HNF4G ND Hs01071345_m1 34.86* No value

NR2B1 Retinoid X receptor, alpha RXRa 24.3 Hs00172565_m1 26.37 4.7E L 04NR2B2 Retinoid X receptor, beta RXRb ND Hs00232774_m1 25.90 0.008NR2B3 Retinoid X receptor, gamma RXRg ND Hs00199455_m1 35.19* No valueNR2C1 NR subfamily 2, group C,

member 1TR2 ND Hs00172676_m1 28.37 4.9E L 04

NR2C2 NR subfamily 2, group C,member 2

TR4 ND Hs00231489_m1 28.17 0.004

NR2E1 NR subfamily 2, group E,member 1

TLX ND Hs00172664_m1 40.00* No value

NR2E3 NR subfamily 2, group E,member 3

PNR ND Hs00183917_m1 38.69* No value

NR2F1 NR subfamily 2, group F,member 1

COUP-TF1 ND Hs00818842_m1 32.10 0.082

NR2F2 NR subfamily 2, group F,member 2

COUP-TF2 ND Hs00819630_m1 29.23 0.002

NR2F6 NR subfamily 2, group F,member 6

EAR2 ND Hs00172870_m1 26.95 0.036

NR3A1 ER 1 ERa ND Hs00174860_m1 27.52 0.030NR3A2 ER 2 ERb ND Hs00230957_m1 33.17* No valueNR3B1 Estrogen-related receptor

alphaERRa ND Hs00607062_gH 27.96 0.008

NR3B2 Estrogen-related receptorbeta

ERRb ND Hs00374442_m1 39.28* No value

NR3B3 Estrogen-related receptorgamma

ERRg ND Hs00155006_m1 38.99* No value

NR3C1 Glucocorticoid receptor GR 8.4 Hs00230818_m1 25.22 0.002

(continued on next page)

835Osteoarthritis and Cartilage Vol. 17, No. 7

Table I (continued )

Gene symbol Gene name Alias GeneChip� averagesignal (ppm)

TaqMan�

assayTaqMan�

average CT

Welch t testNRs OA vs normal

NR3C2 NR subfamily 3, group C,member 2

MR ND Hs00230906_m1 30.45 0.022

NR3C3 Progesterone receptor PR ND Hs00172183_m1 32.18 0.001NR3C4 Androgen receptor AR ND Hs00171172_m1 28.21 0.034NR4A1 NR subfamily 4, group A,

member 1NUR77 ND Hs00374225_m1 28.99 0.335

NR4A2 NR subfamily 4, group A,member 2

NURR1 ND Hs00428691_m1 29.68 0.656

NR4A3 NR subfamily 4, group A,member 3

NOR1 ND Hs00175077_m1 31.37 4.8E L 04

NR5A1 NR subfamily 5, group A,member 1

SF1 ND Hs00610436_m1 40.00* No value

NR5A2 NR subfamily 5, group A,member 2

LRH-1 ND Hs00187067_m1 31.22* No value

NR6A1 NR subfamily 6, group A,member 1

GCNF ND Hs00364256_m1 33.29* No value

Endogenous controlsGUSB Glucuronidase beta 15.10 Hs99999908_m1 27.14 0.863TFRC Transferrin receptor (p90) CD71 6.40 Hs99999911_m1 27.62 0.863

LXR targetsABCG1 ATP-binding cassette, sub-family G

(WHITE), member 1ND Hs00245154_m1 30.82 5.9E L 05

APOD Apolipoprotein D 106 Hs00155794_m1 21.09 8.9E L 05APOE Apolipoprotein E ND Hs00171168_m1 26.72 2.4E L 04

Positive controlPRSS11 HTRA serine peptidase 1 HTRA 100.8 Hs00170197_m1 22.46 0.002

Co-repressorsNCOR1 NR co-repressor 1 13.2 Hs00196920_m1 26.38 4.7E L 05NCOR2 NR co-repressor 2 SMRT 18 Hs00196955_m1 25.73 2.0E L 06

A listing of the genes represented by assays on the hNR-TLDA included 48 of the 49 known human NRs (all except NR2A3), two NR co-

regulators, a positive control representing a transcript previously shown to be well-expressed in cartilage and upregulated in OA cartilage,

three known LXR target genes, and 10 endogenous control ‘‘normalizer’’ genes. All assays tiled on the custom hNR-TLDA are listed in alpha-

betical order, grouped by function, with columns representing the Gene Expression Assay (Applied Biosystems) number, average signal value

(in ppm) for that particular gene within all normal samples on GeneChip, average CT values for each assay across all normal samples, and the

P values for the comparison of the TLDA results for the OA vs normal comparisons.

*Eliminated from further analysis because combined average CT> 34 across normal and OA cohorts.

yEliminated from further analysis because of erratic assay performance.

836 L. A. Collins-Racie et al.: NR expression in OA cartilage

expressed in articular cartilage, 23 were expressed at signif-icantly different levels (P< 0.05 by Welch t test) in OA artic-ular cartilage compared to normal (Table I). The differencesin expression of each NR in all OA cartilage samples com-pared to normal, expressed as fold-change, are shown inFig. 2. In addition to the 48 NRs, the transcript levels forco-repressors NCOR1 and NCOR2 were also assayed inall cartilage samples. Compared to normal cartilage,NCOR1 expression was decreased 1.6-fold (P¼ 4.4E e 5)in OA cartilage and NCOR2 expression was 2.3-fold lowerin OA cartilage (P¼ 2.0E e 6).

EVALUATION OF LXR TARGET GENE EXPRESSION IN HUMAN

OA CARTILAGE COMPARED TO UNAFFECTED CARTILAGE

LXRs and associated co-receptors were among the 23NRs dysregulated in OA cartilage, confirming earlier pre-liminary data. LXRa (NR1H3), LXRb (NR1H2), RXRa(NR2B1) and RXRb (NR2B2) were all found to be ex-pressed at significantly lower levels in OA cartilage com-pared to normal (Fig. 2). Analysis of qRT-PCR data wasperformed to determine whether decreased expressionof LXRs and RXRs resulted in significant reduction in ex-pression of LXR target genes in these same OA and

normal cartilage samples. As shown in Fig. 3, ABCG1,APOD, and APOE all showed significantly decreased ex-pression in OA cartilage compared to normal, reflectinga deficit in LXR activity in OA cartilage.

OA CHONDROCYTES INDUCE LXR TARGET GENES FOLLOWING

TREATMENT WITH A SYNTHETIC LXR AGONIST

Given reduced LXR and RXR expression in OA chon-drocytes, experiments were performed to determinewhether these cells were responsive to a synthetic LXRagonist. Primary articular chondrocytes isolated from hu-man OA donors (n¼ 3) were grown in monolayer cultureswith the LXR agonist T0901317 (2 mM) or vehicle in trip-licate cultures per treatment. Following 16 h of incubation,the effect of T0901317 on the expression of LXR targetgenes was assayed by qRT-PCR. As shown in Fig. 4, ex-pression of ABCA1 and ABCG1 as well as LXRa, APOE,SREBF1, and SCD, were all significantly increased inLXR agonist-treated cultures compared to vehicle con-trols; the ABC cassette genes ABCA1 and ABCG1showed the most dramatic induction by T0901317 treat-ment of all of the genes studied (>20- and 80-fold,respectively).

0

50

100

150

200

250

Normal OA Normal OA

Normal OA Normal OA

2.0 X

**

PR

SS

11 T

ran

scrip

t

L

evels (p

pm

)A

0

0.5

1

1.5

2

2.5

3

3.5

4

PR

SS

11 A

ve

rage

Fo

ld C

hange

2.9 X

**

B

0

20

40

60

80

100

120

140

Ap

oD

T

ran

scrip

t L

ev

els (p

pm

) 9.8 X

**

C

0

0.5

1

1.5

2

2.5

3

Ap

oD

A

verag

e F

old

C

han

ge

18.4 X

**

D

Fig. 1. Comparison of qRT-PCR TLDA assay data to GeneChip probe set data. Two genes represented by assays on the TLDA were ex-pressed highly enough in both normal and OA cartilage to allow measurement and comparison of expression differences on gene chipsand TLDA. Expression differences between OA and normal cartilage were determined for PRSS11 (A) and APOD (C) on gene chips, andcompared to TLDA data for PRSS11 (B) and APOD (D). Bars represent the average expression of PRSS11 or APOD in normal cartilage(n¼ 10; light gray bars) or OA cartilage samples (n¼ 20, average of 10 non-lesional OA and 10 lesional OA cartilage; black bars) measuredeither by Affymetrix U95Av2 gene chips (expressed in transcript ppm) or by the custom hNR-TLDA (expressed in average fold-change percohort, using GUSB as normalizer and the mean of all normal cartilage samples as the calibrator). Values represent the mean� S.E.M. for

each group. **P< 0.01 by Welch t test.

0.1

1

10 Normal CartilageOA Cartilage

Averag

e F

old

C

han

ge vs. A

ll N

orm

als

TRα*

TRβ

RAR

α**

RAR

β*

RAR

γ*

PPAR

α*

PPAR

δ**

PPAR

γ** Rev

-Erb

α** Rev

-Erb

β

RO

Rα

RO

Rγ

LXR

β**

LXR

α*PX

R**

RXR

α**

RXR

β**

TR2*

*

TR4*

*

CO

UP-

TF1

CO

UP-

TF2*

*

EAR

2*

ERα*

ERR

α**

GR

**

MR

*PR

**

AR*

NU

R77

NU

RR

1

NO

R1*

*

Fig. 2. Comparison of NR expression in human osteoarthritic articular cartilage compared to normal cartilage. NR expression data from normalcartilage (n¼ 10) and OA cartilage (n¼ 20) RNA samples using the hNR-TLDA, expressed as mean RQ (fold-change)� S.E.M. for that cohortcompared to the mean of all normal samples following normalization to the GUSB endogenous control. *P< 0.05; **P< 0.01 by Welch t test.

837Osteoarthritis and Cartilage Vol. 17, No. 7

Fig. 3. Target genes for LXR are expressed at significantly lowerlevels in osteoarthritic articular cartilage compared to normals.Comparison of the average fold-change (RQ) for LXR target genesABCG1, APOD and APOE measured on the hNR-TLDA. Expres-sion values (CT) for each sample were normalized to the averageexpression (CT average) for all endogenous controls, and com-pared to all normal samples. Bars represent the mean RQ� S.E.M.for each cohort (normal, n¼ 10; OA, n¼ 20). **P< 0.01 by Welch t

test.

838 L. A. Collins-Racie et al.: NR expression in OA cartilage

TRANSCRIPTIONAL REGULATION OF LXR AND RXR BY

INFLAMMATORY CYTOKINES

The reduction in LXR and RXR expression in OA chon-drocytes suggested that they might be under the transcrip-tional control of signaling pathways that are altered in OA. Itis well established that IL-1 and TNF are important media-tors of OA43e47, so experiments were performed to deter-mine whether these cytokines can regulate LXR and RXRexpression in articular chondrocytes. Primary articularchondrocytes were isolated from human donors (n¼ 7;n¼ 4 normal, n¼ 3 OA) and grown in monolayer cultureswith IL-1b, TNFa, or vehicle, and the expression of LXRand RXR isoforms was measured by qRT-PCR. Theseexperiments showed that RXRa expression was signifi-cantly reduced by both IL-1 and TNF treatment in articularchondrocytes from both normal and OA donors [Fig. 5(C)].Interestingly, LXRa expression was significantly reducedby both IL-1 and TNF treatment only in chondrocytes fromnormal donors [Fig. 5(A)]. A trend toward reduced LXRaexpression was also seen in IL-1 and TNF-treated

Fig. 4. Primary chondrocytes from human OA articular cartilage respondArticular chondrocytes isolated from late-stage OA knee cartilage from hwith or without LXR agonist T0901317 (2 mM) in triplicate cultures per treawas assayed by qRT-PCR to measure the effect of T0901317 on the expreas the average fold-change for each gene in treated cultures compared

ABCG1. Panel B, NR1H3 (LXRa), APOE, S

chondrocytes from OA donors, but the effect did not reachsignificance in this sample set. In contrast, LXRb expres-sion was largely unchanged by cytokine treatment of pri-mary chondrocytes [Fig. 5(B)]; only IL-1 treatment ofnormal chondrocytes showed a small but significant reduc-tion in LXRb expression. RXRb showed an expression pat-tern similar to LXRb whereby significant downregulationwas seen only for IL-1 treated normal chondrocytes[Fig. 5(D)]. RXRg was expressed at levels too low for quan-titative assessment of gene regulation by cytokine treat-ment (data not shown).

LXR AGONIST TREATMENT MITIGATES CYTOKINE-INDUCED

PG DEGRADATION IN HUMAN OA ARTICULAR CARTILAGE

Experiments were performed to evaluate the effect ofLXR agonist treatment of human articular cartilage explantson PG synthesis and degradation following treatment withcatabolic cytokines. Figure 6 shows the results from a repre-sentative experiment in which cartilage explants fromhuman OA cartilage were treated with or without IL-1b/OSM to stimulate PG degradation following addition of in-creasing doses of the LXR agonist compound T0901317.After 10 days in culture, the highest dose of T0901317(20 mM) significantly decreased endogenous PG release(no cytokine treatment) from the OA cartilage explantsinto the conditioned medium [20 mM T0901317 vs dimethylsulfoxide (DMSO) control], and this activity was alsoreflected in significantly increased retention of PG in theexplants [20 mM T0901317 vs DMSO control]. Addition ofIL-1b/OSM to the explant cultures resulted in a significantand substantial increase in PG release from the explantsand a concomitant decrease in PG remaining in theexplants, while co-treatment with 10 mM or 20 mMT0901317 resulted in a significant reduction in cytokine-mediated PG release to the conditioned media and in-creased PG retention in the explants.

Discussion

An analysis of microarray data on human cartilage sam-ples from OA patients and unaffected controls33 for NR ex-pression and regulation in OA cartilage suggesteddysregulation of some NRs in OA, but the inherently lowlevels of NR expression limited the analysis because mostNRs were below the microarray lower limit of quantitation.

to LXR agonist treatment by upregulating known LXR target genes.uman donors (n¼ 3) were grown in monolayer culture and treatedtment group for 26 h. RNA prepared from the cells following culturession of LXR target genes in OA chondrocytes. Data are presentedto untreated, �S.E.M. Panel A, ABC cassette genes ABCA1 andREBF1, SCD. *P< 0.05, **P< 0.01.

0.10.20.30.40.50.60.70.80.91.01.11.2 RXRβ

Averag

e F

old

C

han

ge vs. U

ntreated

D

0.10.20.30.40.50.60.70.80.91.01.11.2 LXRα

Averag

e F

old

C

han

ge vs. U

ntreated

A

∗∗

∗∗

0.10.20.30.40.50.60.70.80.91.01.11.2 LXRβ

Averag

e F

old

C

han

ge vs. U

ntreated

B

∗

0.10.20.30.40.50.60.70.80.91.01.11.2 RXRα

Averag

e F

old

C

han

ge vs. U

ntreated

∗∗∗∗

∗

C

∗∗

∗∗

IL-1 treated Normal Chondrocytes

TNF treated Normal Chondrocytes TNF treated OA Chondrocytes

IL-1 treated OA Chondrocytes

Fig. 5. Primary OA chondrocytes downregulate LXR and RXR in response to treatment with IL-1b or TNFa. Primary chondrocytes isolatedfrom human donors (OA n¼ 3, normal n¼ 4) were treated in monolayer culture with either 1 ng/ml IL-1b or 10 ng/ml TNFa for 18 h in triplicatecultures per treatment. RNA prepared from the cells following culture was assayed by qRT-PCR to measure the effect of cytokine treatment onthe expression of (A) LXRa, (B) LXRb, (C) RXRa, and (D) RXRb. Bars represent the individual average fold-change in expression values forthe cytokine-treated cultures for each donor compared to untreated cultures from the same donor, �S.E.M. *P< 0.05, **P< 0.01 vs control.

839Osteoarthritis and Cartilage Vol. 17, No. 7

Since NR dysregulation in OA was largely unknown, the im-portant implications of these data with regard to increasingunderstanding of gene regulation in OA and discoveringnovel drug targets led us to consider utilizing a recently-de-veloped, moderate-throughput technology e TLDAs e tomonitor all human NRs in OA vs normal cartilage. This ap-proach allowed simultaneous measurement of all NR tran-scripts with much greater sensitivity than the microarrays,which was necessary to measure the relatively low-leveltranscription of many of the NRs. Moreover, the TLDA plat-form uses far less RNA than ‘‘traditional’’ TaqMan� individ-ual assays for the number of genes studied here, which wasan important consideration since the OA and normal carti-lage RNA samples were precious and finite.

We found that almost two-thirds of all known NRs wereexpressed in human articular cartilage. These findingswere consistent with a previous study by Chaturvediet al.30 who found evidence for expression of about 30NRs in human OA cartilage and particularly abundantexpression of Rev-Erba, but no comparison of NR expres-sion in OA vs normal cartilage was reported. In the presentstudy, a surprisingly high proportion (approximately 75%) ofexpressed NRs showed altered expression in OA

compared to normal cartilage. Although this seems inordi-nately high, this is one of the first reported in-depth studiesof global NR regulation in a disease tissue and future inves-tigations on affected tissues from other disease states willdetermine whether or not our findings in OA cartilage areunusual. The changes in NR expression that we observedin OA cartilage are consistent with most previous reportslinking alterations in activity or expression of some NRswith OA. For instance, an association of ER and OA hasbeen found by haplotype analyses18,19, and reduced estro-gen signaling has been postulated to contribute to an in-creased incidence of OA in post-menopausal women.There is evidence that estrogen replacement therapy mayprovide a protective effect in OA48, and 17-b-estradiol hasbeen shown to suppress matrix metalloproteinase (MMP)expression in synovial cells49 and in chondrocytes50. Con-sistent with these findings, we found that ERa expressionin human OA cartilage is almost 10-fold lower than normal.Similarly, reduced expression of GR27 and PPARg51 in OAcartilage has been previously reported, consistent with ourobservations, and the use of agonists to these NRs hasbeen shown to diminish some catabolic activities in diseasetissue25,51.

Fig. 6. LXR agonist treatment of human OA cartilage explants mitigates cytokine-induced PG degradation. Fresh human OA cartilage explantswere cultured for 10 days with 1 ng/ml IL-1b plus 5 ng/ml OSM with or without LXR agonist T0901317 (5, 10, 20 mM). Percent of total PGsreleased into the culture medium� S.E.M. (A) and the percent of total PGs retained in the explants� S.E.M. (B) are shown. #P< 0.05,##P< 0.01 in comparison to DMSO control; *P< 0.05, **P< 0.01 in comparison to DMSOþ IL-1/OSM control. Cartilage explants from sixhuman donors were evaluated in six separate experiments of the same design. The results of a representative experiment on one of the donor

samples are shown.

840 L. A. Collins-Racie et al.: NR expression in OA cartilage

Our findings of dysregulation of several other NRs in OAcartilage have not been reported previously. In particular,our finding that LXRa, LXRb, RXRa, and RXRb areexpressed at significantly lower levels in OA cartilage sug-gested a potential LXR signaling defect in OA, which wassubstantiated by significantly decreased expression ofgenes in OA cartilage known to be directly regulated byLXR. Previous literature describing the expression or func-tion of LXR and/or RXR in cartilage is sparse. LXR and RXRexpression was reported in chicken hypertrophic chondro-cytes52 and responded to synthetic LXR agonists byincreasing LXR target genes involved in cholesterol efflux.Treatment of bovine chondrocytes with hyaluronan hexa-saccharides was shown to increase expression of retinoicacid receptor (RAR) and RXR53. The impact of changesin RXR expression is difficult to predict since RXR can existas a homodimer, or can heterodimerize with several otherNRs besides LXRs (reviewed in Ref. 54). RXR biology isfurther complicated by the fact that some heterodimericreceptor complexes (e.g., LXRs, farnesoid X receptor(FXR), and PPARs) can be independently activated byeither the RXR’s ligand, the RXR partner’s ligand, or byboth; alternatively, other RXR heterodimeric receptor com-plexes require the partner’s ligand for activation (e.g.,VDR and thyroid hormone receptor (TR))54. Burrageet al.55 reported suppression of IL-1b-mediated inductionof MMP-1 and MMP-13 expression in the SW1353 humanchondrosarcoma cell line by treatment with the RXR-spe-cific agonist LG100268. In contrast, Castrillo et al.56 re-ported that LXR agonist treatment of murine macrophages

did not suppress IL-1b-mediated MMP-13 expression. Wefind that LXR agonist treatment of human cartilage explantsdoes not suppress IL-1b-mediated MMP-1 or MMP-13 ex-pression (Li et al., submitted), suggesting that either theRXR agonist effect on MMP expression is not mediatedthrough an LXR/RXR heterodimer or that SW1353 cellsare not accurately reflecting the response of primary humancartilage. Studies comparing LXR and RXR agonists ongene expression and matrix degradation in explant culturesshould address this question.

A protective role for LXR signaling in OA cartilage is sug-gested by the ability of a synthetic LXR agonist to block thecytokine-mediated degradation of PG in human cartilageexplants. The anti-inflammatory potential of LXR activationhas been documented previously in other cell types, includ-ing macrophage57 where LXR agonist treatment sup-pressed nuclear factor (NF)-kB signaling elicited byexposure to lipopolysaccharide (LPS) and inhibited numer-ous inflammatory downstream targets of NF-kB such as IL-6, cyclo-oxygenase-2 (COX-2) and inducible nitric oxidesynthase (iNOS). These inflammatory mediators are knownto regulate enzymes responsible for cartilage matrix de-struction56,58. Since chondrocytes and macrophages aresimilar in their production of, and response to, inflammatorycytokines59, it is possible that the mechanism of LXR anti-inflammatory effects in chondrocytes and macrophagemay also be similar. In chondrocytes, we find that IL-1band TNFa treatment results in significantly decreasedexpression of LXRs and RXRs, especially the alpha iso-forms. This may accentuate the catabolic effects of NF-kB

841Osteoarthritis and Cartilage Vol. 17, No. 7

stimulation by reducing an important negative feedback.Studies to further elucidate the mechanism(s) of LXR sup-pression of cytokine-induced cartilage matrix degradationin OA are currently in progress.

Conflict of interest

The work described in this manuscript was funded byWyeth Pharmaceuticals. The authors are current or pastemployees of Wyeth Research and received monetarycompensation from Wyeth.

Acknowledgments

The authors wish to thank Maryann Whitley, Andrew Hill,and Roger Ford for their help with microarray and TLDA bi-oinformatics analysis, Dina Anderson and Tad Kornaga fortechnical contributions, and Dawn DeThomas for assis-tance with graphics.

References

1. Robinson-Rechavi M, Carpentier AS, Duffraisse M, Laudet V. Howmany nuclear hormone receptors are there in the human genome?Trends Genet 2001;17:554e6.

2. Pearce KH, Iannone MA, Simmons CA, Gray JG. Discovery of novelnuclear receptor modulating ligands: an integral role for peptide inter-action profiling. Drug Discov Today 2004;9:741e51.

3. Tsuchida A, Yamauchi T, Kadowaki T. Nuclear receptors as targets fordrug development: molecular mechanisms for regulation of obesityand insulin resistance by peroxisome proliferator-activated receptorgamma, CREB-binding protein, and adiponectin. J Pharmacol Sci2005;97:164e70.

4. Tobin JF, Freedman LP. Nuclear receptors as drug targets in metabolicdiseases: new approaches to therapy. Trends Endocrinol Metab2006;17:284e90.

5. Margolis R. Nuclear receptors and bone. Ann N Y Acad Sci 2007.6. Heinlein CA, Chang C. Androgen receptor in prostate cancer. Endocr

Rev 2004;25:276e308.7. Vrbanec D, Petricevic B. Estrogen and progesterone receptor status in

primary breast cancer e a study of 11,273 patients from the year 1990to 2002. Coll Antropol 2007;31:535e40.

8. Mullican SE, Zhang S, Konopleva M, Ruvolo V, Andreeff M, Milbrandt J,et al. Abrogation of nuclear receptors Nr4a3 and Nr4a1 leads to devel-opment of acute myeloid leukemia. Nat Med 2007;13:730e5.

9. Saez E, Tontonoz P, Nelson MC, Alvarez JG, Ming UT, Baird SM, et al.Activators of the nuclear receptor PPARgamma enhance colon polypformation. Nat Med 1998;4:1058e61.

10. Hart DJ, Doyle DV, Spector TD. Incidence and risk factors for radio-graphic knee osteoarthritis in middle-aged women: the ChingfordStudy. Arthritis Rheum 1999;42:17e24.

11. Spector TD, Nandra D, Hart DJ, Doyle DV. Is hormone replacementtherapy protective for hand and knee osteoarthritis in women? TheChingford Study. Ann Rheum Dis 1997;56:432e4.

12. Sowers MF, Hochberg M, Crabbe JP, Muhich A, Crutchfield M,Updike S. Association of bone mineral density and sex hormonelevels with osteoarthritis of the hand and knee in premenopausalwomen. Am J Epidemiol 1996;143:38e47.

13. Samanta A, Jones A, Regan M, Wilson S, Doherty M. Is osteoarthritis inwomen affected by hormonal changes or smoking? Br J Rheumatol1993;32:366e70.

14. Hannan MT, Felson DT, Anderson JJ, Naimark A, Kannel WB. Estrogenuse and radiographic osteoarthritis of the knee in women. The Fra-mingham Osteoarthritis Study. Arthritis Rheum 1990;33:525e32.

15. Nevitt MC, Cummings SR, Lane NE, Hochberg MC, Scott JC,Pressman AR, et al. Association of estrogen replacement therapywith the risk of osteoarthritis of the hip in elderly white women. Studyof Osteoporotic Fractures Research Group. Arch Intern Med 1996;156:2073e80.

16. Zhang Y, McAlindon TE, Hannan MT, Chaisson CE, Klein R,Wilson PW, et al. Estrogen replacement therapy and worsening of ra-diographic knee osteoarthritis: the Framingham Study. ArthritisRheum 1998;41:1867e73.

17. Oliveria SA, Felson DT, Klein RA, Reed JI, Walker AM. Estrogenreplacement therapy and the development of osteoarthritis. Epidemi-ology 1996;7:415e9.

18. Bergink AP, van Meurs JB, Loughlin J, Arp PP, Fang Y, Hofman A, et al.Estrogen receptor alpha gene haplotype is associated with radio-graphic osteoarthritis of the knee in elderly men and women. ArthritisRheum 2003;48:1913e22.

19. Ushiyama T, Ueyama H, Inoue K, Nishioka J, Ohkubo I, Hukuda S.Estrogen receptor gene polymorphism and generalized osteoarthritis.J Rheumatol 1998;25:134e7.

20. Tetlow LC, Woolley DE. Expression of vitamin D receptors and matrixmetalloproteinases in osteoarthritic cartilage and human articularchondrocytes in vitro. Osteoarthritis Cartilage 2001;9:423e31.

21. Solovieva S, Hirvonen A, Siivola P, Vehmas T, Luoma K,Riihimaki H, et al. Vitamin D receptor gene polymorphisms andsusceptibility of hand osteoarthritis in Finnish women. ArthritisRes Ther 2006;8:R20.

22. Jordan KM, Syddall H, Dennison EM, Cooper C, Arden NK. Birthweight,vitamin D receptor gene polymorphism, and risk of lumbar spine oste-oarthritis. J Rheumatol 2005;32:678e83.

23. Fahmi H, Di Battista JA, Pelletier JP, Mineau F, Ranger P, Martel-Pelletier J. Peroxisome proliferator-activated receptor gamma activa-tors inhibit interleukin-1beta-induced nitric oxide and matrix metallo-proteinase 13 production in human chondrocytes. Arthritis Rheum2001;44:595e607.

24. Fahmi H, Pelletier JP, Di Battista JA, Cheung HS, Fernandes JC, Martel-Pelletier J. Peroxisome proliferator-activated receptor gamma activa-tors inhibit MMP-1 production in human synovial fibroblasts likely byreducing the binding of the activator protein 1. Osteoarthritis Cartilage2002;10:100e8.

25. Kobayashi T, Notoya K, Naito T, Unno S, Nakamura A, Martel-Pelletier J, et al. Pioglitazone, a peroxisome proliferator-activatedreceptor gamma agonist, reduces the progression of experimental os-teoarthritis in guinea pigs. Arthritis Rheum 2005;52:479e87.

26. Boileau C, Martel-Pelletier J, Fahmi H, Mineau F, Boily M, Pelletier JP.The peroxisome proliferator-activated receptor gamma agonist piogli-tazone reduces the development of cartilage lesions in an experimen-tal dog model of osteoarthritis: in vivo protective effects mediatedthrough the inhibition of key signaling and catabolic pathways. ArthritisRheum 2007;56:2288e98.

27. DiBattista JA, Martel-Pelletier J, Antakly T, Tardif G, Cloutier JM,Pelletier JP. Reduced expression of glucocorticoid receptor levelsin human osteoarthritic chondrocytes. Role in the suppression ofmetalloprotease synthesis. J Clin Endocrinol Metab 1993;76:1128e34.

28. Stove J, Schoniger R, Huch K, Brenner R, Gunther KP, Puhl W, et al.Effects of dexamethasone on proteoglycan content and gene expres-sion of IL-1beta-stimulated osteoarthrotic chondrocytes in vitro. ActaOrthop Scand 2002;73:562e7.

29. Mix KS, Attur MG, Al-Mussawir H, Abramson SB, Brinckerhoff CE,Murphy EP. Transcriptional repression of matrix metalloproteinasegene expression by the orphan nuclear receptor NURR1 in cartilage.J Biol Chem 2007;282:9492e504.

30. Chaturvedi P, Pratta M, Steplewski K, Connor J, Kumar S. Functionalcharacterization of an orphan nuclear receptor, Rev-ErbAalpha, inchondrocytes and its potential role in osteoarthritis. Arthritis Rheum2006;54:3513e22.

31. Repa JJ, Turley SD, Lobaccaro JA, Medina J, Li L, Lustig K, et al. Reg-ulation of absorption and ABC1-mediated efflux of cholesterol by RXRheterodimers. Science 2000;289:1524e9.

32. Schultz JR, Tu H, Luk A, Repa JJ, Medina JC, Li L, et al. Role of LXRs incontrol of lipogenesis. Genes Dev 2000;14:2831e8.

33. LaVallie ER, Chockalingam PS, Collins-Racie LA, Freeman BA,Keohan CC, Leitges M, et al. Protein kinase Czeta is up-regulatedin osteoarthritic cartilage and is required for activation of NF-kappaBby tumor necrosis factor and interleukin-1 in articular chondrocytes.J Biol Chem 2006;281:24124e37.

34. Grau S, Richards PJ, Kerr B, Hughes C, Caterson B, Williams AS, et al.The role of human HtrA1 in arthritic disease. J Biol Chem 2006;281:6124e9.

35. Wu J, Liu W, Bemis A, Wang E, Qiu Y, Morris EA, et al. Comparative pro-teomic characterization of articular cartilage tissue from normal donorsand patients with osteoarthritis. Arthritis Rheum 2007;56:3675e84.

36. Livak KJ, Schmittgen TD. Analysis of relative gene expression datausing real-time quantitative PCR and the 2(-Delta Delta C(T)) method.Methods 2001;25:402e8.

37. Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrim-ination of sulphated glycosaminoglycans by use of dimethylmethyleneblue. Biochim Biophys Acta 1986;883:173e7.

38. Schwartz K, Lawn RM, Wade DP. ABC1 gene expression and ApoA-I-mediated cholesterol efflux are regulated by LXR. Biochem BiophysRes Commun 2000;274:794e802.

39. Laffitte BA, Repa JJ, Joseph SB, Wilpitz DC, Kast HR, Mangelsdorf DJ,et al. LXRs control lipid-inducible expression of the apolipoprotein Egene in macrophags and adipocytes. Proc Natl Acad Sci U S A2001;98:507e12.

842 L. A. Collins-Racie et al.: NR expression in OA cartilage

40. Venkateswaran A, Repa JJ, Lobaccaro JM, Bronson A, Mangelsdorf DJ,Edwards PA. Human white/murine ABC8 mRNA levels are highly in-duced in lipid-loaded macrophages. A transcriptional role for specificoxysterols. J Biol Chem 2000;275:14700e7.

41. Hummasti S, Laffitte BA, Watson MA, Galardi C, Chao LC,Ramamurthy L, et al. Liver X receptors are regulators of adipocytegene expression but not differentiation: identification of apoD as a di-rect target. J Lipid Res 2004;45:616e25.

42. Kobayashi M, Takezawa S, Hara K, Yu RT, Umesono Y, Agata K, et al.Identification of a photoreceptor cell-specific nuclear receptor. ProcNatl Acad Sci U S A 1999;96:4814e9.

43. Hedbom E, Hauselmann HJ. Molecular aspects of pathogenesis in oste-oarthritis: the role of inflammation. Cell Mol Life Sci 2002;59:45e53.

44. Aigner T, Kurz B, Fukui N, Sandell L. Roles of chondrocytes in the path-ogenesis of osteoarthritis. Curr Opin Rheumatol 2002;14:578e84.

45. Goldring MB. The role of the chondrocyte in osteoarthritis. ArthritisRheum 2000;43:1916e26.

46. Goldring MB. Osteoarthritis and cartilage: the role of cytokines. CurrRheumatol Rep 2000;2:459e65.

47. Fernandes JC, Martel-Pelletier J, Pelletier JP. The role of cytokines inosteoarthritis pathophysiology. Biorheology 2002;39:237e46.

48. Hanna FS, Wluka AE, Bell RJ, Davis SR, Cicuttini FM. Osteoarthritisand the postmenopausal woman: epidemiological, magnetic reso-nance imaging, and radiological findings. Semin Arthritis Rheum2004;34:631e6.

49. Lu T, Achari Y, Sciore P, Hart DA. Estrogen receptor alpha regulatesmatrix metalloproteinase-13 promoter activity primarily through theAP-1 transcriptional regulatory site. Biochim Biophys Acta 2006;1762:719e31.

50. Richette P, Dumontier MF, Francois M, Tsagris L, Korwin-Zmijowska C,Rannou F, et al. Dual effects of 17beta-oestradiol on interleukin 1beta-induced proteoglycan degradation in chondrocytes. Ann Rheum Dis2004;63:191e9.

51. Afif H, Benderdour M, Mfuna-Endam L, Martel-Pelletier J, Pelletier JP,Duval N, et al. Peroxisome proliferator-activated receptor gamma1expression is diminished in human osteoarthritic cartilage and isdownregulated by interleukin-1beta in articular chondrocytes. ArthritisRes Ther 2007;9:R31.

52. Gentili C, Tutolo G, Pianezzi A, Cancedda R, Descalzi Cancedda F.Cholesterol secretion and homeostasis in chondrocytes: a liver X re-ceptor and retinoid X receptor heterodimer mediates apolipoproteinA1 expression. Matrix Biol 2005;24:35e44.

53. Ohno S, Im HJ, Knudson CB, Knudson W. Hyaluronan oligosaccharide-induced activation of transcription factors in bovine articular chondro-cytes. Arthritis Rheum 2005;52:800e9.

54. Germain P, Chambon P, Eichele G, Evans RM, Lazar MA, Leid M, et al.International Union of Pharmacology. LXIII. Retinoid X receptors.Pharmacol Rev 2006;58:760e72.

55. Burrage PS, Huntington JT, Sporn MB, Brinckerhoff CE. Regulation ofmatrix metalloproteinase gene expression by a retinoid X receptor-specific ligand. Arthritis Rheum 2007;56:892e904.

56. Castrillo A, Joseph SB, Marathe C, Mangelsdorf DJ, Tontonoz P.Liver X receptor-dependent repression of matrix metalloprotei-nase-9 expression in macrophages. J Biol Chem 2003;278:10443e9.

57. Joseph SB, Castrillo A, Laffitte BA, Mangelsdorf DJ, Tontonoz P. Recip-rocal regulation of inflammation and lipid metabolism by liver X recep-tors. Nat Med 2003;9:213e9.

58. Tsutsumi R, Ito H, Hiramitsu T, Nishitani K, Akiyoshi M, Kitaori T, et al.Celecoxib inhibits production of MMP and NO via down-regulation ofNF-kappaB and JNK in a PGE2 independent manner in human artic-ular chondrocytes. Rheumatol Int 2007.

59. Attur MG, Dave M, Akamatsu M, Katoh M, Amin AR. Osteoarthritis or os-teoarthrosis: the definition of inflammation becomes a semantic issuein the genomic era of molecular medicine. Osteoarthritis Cartilage2002;10:1e4.