Monocyte chemoattractant protein 1 (MCP 1) expression in human articular cartilage Induction by peptide regulatory factors and differential effects of dexamethasone and retinoic acid

Monocyte chemoattractant protein-1 (MCP-1)

expression in human articular cartilage.

Induction by peptide regulatory factors and

differential effects of dexamethasone and

retinoic acid.

P M Villiger, … , R Terkeltaub, M Lotz

J Clin Invest. 1992;90(2):488-496. https://doi.org/10.1172/JCI115885.

Monocyte influx and activation in synovial joints are important in the pathogenesis of both degenerative and inflammatory arthropathies. In this study, we demonstrate the potential of articular cartilage to directly modulate these events. IL-1-stimulated human articular

chondrocytes transcribed 0.7-kb monocyte chemoattractant protein-1 (MCP-1) mRNA. In situ hybridization of cartilage organ cultures revealed MCP-1 transcripts in chondrocytes in the superficial tangential zone within 2 h of stimulation with IL-1. Chondrocytes in deeper layers responded by 4 h and reached maximum MCP-1 mRNA levels by 8-12 h. IL-1-stimulated cartilage organ and chondrocyte monolayer cultures released functional

monocyte chemotactic activity. This was neutralized by a monoclonal antibody specific for MCP-1, and was associated with the synthesis and secretion of immunoreactive 13-kD and 15-kD isoforms of MCP-1. Regulators and signal transduction pathways involved with the expression of the MCP-1 gene in chondrocytes were analyzed. Steady-state mRNA levels were increased by the known chondrocyte activators IL-1, tumor necrosis factor alpha, LPS, platelet-derived growth factor, and transforming growth factor beta. In addition, leukemia inhibitory factor induced MCP-1 gene expression and protein synthesis, identifying this cytokine as a new regulator of chondrocyte function. Dexamethasone blunted the induction of MCP-1 gene expression by IL-1 and by activators of protein kinase A as well as protein kinase C signal transduction pathways. In contrast, retinoic acid strongly increased phorbol myristate […]

Research Article

Find the latest version:

Monocyte Chemoattractant Protein-1 (MCP-1) Expression

in

Human Articular Cartilage

Induction byPeptide Regulatory FactorsandDifferential Effects ofDexamethasone and Retinoic Acid

Peter M. Villiger,* RobertTerkeltaub,**and Martin Lotz*

SamandRoseStein Institutefor Researchon Aging, *Department ofMedicine, University ofCalifornia,San Diego, La Jolla,California 92093;and*SanDiegoVeteransAffairsMedical Center,

LaJolla, California 92093

Abstract Introduction

Monocyteinfluxandactivationinsynovialjointsareimportant inthepathogenesis of both degenerative andinflammatory

ar-thropathies.Inthis study,wedemonstrate the potential ofartic-ularcartilagetodirectlymodulate theseevents.

IL-1-stimulatedhumanarticular chondrocytes transcribed 0.7-kbmonocytechemoattractant protein-1(MCP-1)mRNA. Insitu hybridization of cartilageorganculturesrevealed

MCP-1transcriptsinchondrocytesinthesuperficial tangentialzone within 2 h ofstimulation with IL-1. Chondrocytes indeeper

layers respondedby4 h andreached maximum MCP-1mRNA

levels by 8-12h.

IL-i-stimulated

cartilageorgan and chondro-cyte monolayer cultures released functionalmonocyte chemo-tacticactivity. Thiswasneutralized byamonoclonalantibody specificfor MCP-1,andwasassociated withthesynthesisand

secretion of immunoreactive 13-kD and 15-kD isoforms of MCP-1.

Regulatorsandsignaltransductionpathwaysinvolved with

theexpression oftheMCP-1 geneinchondrocyteswere

ana-lyzed.Steady-statemRNAlevelswereincreased bythe known

chondrocyte activators IL-1,tumornecrosisfactoralpha,LPS,

platelet-derived growth factor,andtransforminggrowth factor

beta. Inaddition, leukemia inhibitoryfactor induced MCP-1

geneexpressionandproteinsynthesis,identifyingthiscytokine

as a new regulator ofchondrocyte function. Dexamethasone blunted the induction ofMCP-1geneexpression byIL-1andby activators of protein kinaseAaswellasproteinkinaseCsignal

transduction pathways. In contrast, retinoicacid strongly in-creasedphorbolmyristate acetate-induced MCP-1expression

andpotentiatedtheeffects ofIL-I andLPS.

Inconclusion,chondrocytesexpressMCP-1in response to

factors thatarepresent incartilageorsynovium.Thisprovides

amechanism by which cartilagecanplayanactiverole inthe

initiationand progression of arthritis. (J. Clin. Invest. 1992.

90:488496.)Key words: chondrocytes*cartilage. monocyte

chemoattractant protein-1-leukemia inhibitoryfactor * trans-forminggrowthfactorbeta *retinoic acid

AddresscorrespondencetoDr.Martin Lotz, Department of Medicine,

UniversityofCalifornia,SanDiego,9500Gilman Drive,LaJolla, CA 92093-0663.

Receivedforpublication16October1991 andin revisedform25 February1992.

TheJournalof Clinical Investigation, Inc.

Volume 90, August 1992, 488-496

Arthritis isgenerallyviewed as aprocessthatisbasedon the

infiltrationofsynovialtissueand/orfluid withinflammatory

cells that cause acute symptoms and leadto degradation of

cartilage.Inthisscenario, cartilage isseen asthepassive target

ofsynovial inflammation,anotion that has been revisedbythe

demonstration that chondrocytesaresecretorycellsthat

pro-duce notonly proteases (1)but alsomultiple regulatoryfactors that can act oncartilage, synovium,or onleukocytes

infiltrat-ingthejoint (2-6).

Synovial fluid monocytosis is a finding characteristic of osteoarthritis(7)adiseaseoriginatinginandprimarily

affect-ing cartilage. Furthermore, intraarticular monocyte infiltration

and activation arethoughtto beimportanteventsin the patho-genesisofcartilage degradationininflammatory and

degenera-tivearthropathies. Monocyte recruitment in arthritidesmay be

initiated byavarietyofchemotactic factors derived from articu-lartissues, including type II collagen peptides released by pro-teolysisofcartilage (8), andtransforming growth factorbeta

(TGF3)' andplatelet-derived growth factor (PDGF)secreted fromsynoviocytesandchondrocytes (5, 6). Inthis study,we examine theability of articular cartilageto expressmonocyte chemoattractant protein (MCP-l ),a morerecently identified cytokine (9-13) that isamember ofafamily of small, proin-flammatory polypeptides that includes platelet factor 4 and IL-8(14, 15). MCP- 1 is specifically recognized by monocytes, stimulatingnotonly chemotaxisbut alsothe release of

super-oxide andlysosomal proteases (16). MCP-1 is secreted as a

76-amino acid glycoprotein that migrates typically intwo

mo-lecular forms ( 13 and 15 kD)onSDS-PAGE under reducing conditions (12). The gene forMCP-1 has been cloned, and someof its regulatory elements have been characterized ( 17). Although multipleinducersand cellsourcesofMCP-1have been characterized, expression by joint tissue cells is largely

unknown. This study demonstrates thatchondrocytesexpress the MCP- 1 gene andthatMCP-l is the major monocyte che-motactantin IL-1-stimulatedcartilage organ culture. Regula-toryfactorsthat areknownto be presentin cartilageor syno-vium caninduceMCP- 1 geneexpressionbychondrocytes in monolayer and organ culture. Newagentsand pathways

in-1.Abbreviationsused in this paper: LIF, leukemia inhibitory factor; MCP, monocyte chemoattractant protein; PDGF, platelet-derived growthfactor;RA, retinoic acid;_TGF#,transforming growth factor beta; TNFa,tumornecrosisfactor alpha.

volved with the induction of MCP-1 in chondrocytes are iden-tified.

Methods

Chondrocyte isolation and culture. Cartilagewasobtainedat autopsy from donors without known history of joint disease. For all experi-mentsreportedhere, cartilage fromthefemoral condylesand tibial plateausof the knee joints was used.Cartilageslices were prepared and washed with DMEM (Whittaker M.A. Bioproducts, Walkersville, MD). Theywereminced withascalpel and treatedwithtrypsin(10% vol/vol)for 15 minin a370Cwaterbath. The samples weretransferred

toDMEMcontaining5%fetalbovineserum(FBS), penicillin-strepto-mycin-fungizone,and 2mg/ml clostridial collagenase type IV (Sigma ChemicalCo., St.Louis, MO)anddigested for3 honagyratory shaker until thetissuefragmentsweredissolved.The cellswerewashed three timesand culturedasprimary chondrocytes. Forcartilageorgan

cul-tureexperiments,slices ofkneecartilageof0.5X0.5X0.3-cm size were prepared.Care was taken not todisturbthesurfaceand toobtainfull thicknesscartilage.Sliceswereweighedand cultured in 12-well plates in DMEM 1% FBS. For studiesongeneexpression chondrocyteswere

usedasprimaryorsubcultured cellsasindicated for each experiment. Chemotaxis assay. Enriched preparations ofhuman peripheral blood monocytes wereisolated fromtheheparin-anticoagulated

ve-nousbloodof normal volunteers.First,mononuclearleukocyteswere

separated bycentrifugationat400 gfor30 minoverFicoll-Hypaque, washedthreetimes, and resuspended in RPMI 1640 that contained 10% FBS. 3-ml aliquots of mononuclear leukocytes were carefully layered overadiscontinuousPercoll(PharmaciaFineChemicals, Pis-cataway,NJ)densitygradient(5.0 ml 30%Percoll;5.0 ml 40% Percoll; 15.0 ml 50.5% Percoll; 7.5 ml 75% Percoll) in 50-mlcentrifugetubes. After centrifugation at 400 gfor 30 min at 220C, the monocyte-enriched fraction wasremovedfromthe40-50.5%interface ofeach gradient. Monocyte enrichment wastypically>85% under these con-ditions,asassessedbystainingwith the mature monocyte marker Mo2 (CD14).

Themonocyte-enriched fractionwaswashed threetimes in RPMI 1640, and resuspended at 106cells/mlin RPMI 1640, supplemented with 10 mM HEPES,1%penicillin-streptomycin,and1%BSA. Mono-cytechemotaxis assays wereperformedintriplicateina48-well modi-fied Boyden chamber apparatus (Neuro Probe, Cabin John, MD) ( 18). Monocytes were separatedfrom purifiedagonists (2 nM human recombinant MCP-1, purchased fromBiosource,Camarillo, CA) or 10 nMFMLP, purchasedfrom SigmaChemical Co. (St. Louis,MO)or

from chondrocyte-conditioned media byapolycarbonate polyvinyl-pyrrolidone-free filterwithaporesize of5um(Neuroprobe).Before testing,chemoattractants(0.1 ml)wereallowedtoequilibrate (37°C for 30_min,4°Cfor 1 h) with(0.01 ml) ofbuffer,or38 nM

nonim-mune mouseIgG,or apreviously characterized, neutralizing murine

monoclonalantibodytoMCP-I (antibodyEl1 _[191,agenerousgift from Drs. Ed Leonard andTeizoYoshimura, National Cancer Insti-tute,Fredrick, MD).Cells(106/mlin 0.05ml)wereincubated with chemoattractants(0.025 ml)for 90 minat37°C, followed by Wright-Giemsastaining.Cellmigrationtothe underside ofthefilterwas quan-tified byreadingfivehighpower fields for each well.

Metabolic labelingandimmunoprecipitation. MCP-1 protein syn-thesiswasanalyzed by metaboliclabelingandimmunoprecipitation. Primarychondrocyteswerewashed in PBS andincubatedin cysteine-andmethionine-free RPMI (ICNBiomedicals,Inc., Costa Mesa,CA) supplemented with L-Glu, L-Met, 1% FBS, [35S]cysteine,and [35S]_

methionine (ICNBiomedicals,Inc.; 100 Ci/ml)for 48 h. The

super-natants werecollected and precleared byincubatingwithproteinG

sepharose (ZymedLab, Inc., South SanFrancisco, CA)for 1 h. The sampleswereincubatedovernightin the presence of the monoclonal murineantibodyE11 (19)toMCP-1 (2Mg/sample)on arotor, and

proteinGsepharosewasadded foroneadditional hour. The

precipi-tates werewashed four times with PBS containing 0.05%Tween-80,

0.1% Triton X-100, and I mM PMSF followed by two washes in PBS. The beadswereboiled for 5 min in Laemmli sample buffer and the proteinswereseparatedon 15% polyacrylamide gels underreducing conditions. After staining with Coomassie blue the gels were treated withAmplify (Amersham Corp., Arlington Heights, IL)forIh,dried,

andexposedtoKodak XAR film(Eastman Kodak, Rochester, NY)at

-70'Cfor 2 d.

RNA probe preparation.The cDNA for MCP-Iinsertedin EcoRI sites ofpBS-SK(-) was kindly provided by Drs. Ed Leonard and Teizo Yoshimura, National Cancer Institute, Fredrick, MD. For develop-ment of a probe for fl-actin, two 15-bp oligonucleotides (5'-CGTCGTCGACAACGG-3' and5'-GACCGTAGCACTACC-3') de-fining a 216-bp fragment of the cDNA were designed. The restriction sitesEcoRI and HindIlwereaddedattheir 5' ends,respectively.After amplification by polymerase chain reaction the fragment was inserted intopGEM-4z (Promega Corp., Madison, WI).

Therecombinant plasmids were linearized and transcribed with the T7 or T3 RNA polymerase to obtain antisense and sense probes, respec-tively. For Northern blotting, the probes were labeled with [32P]UTP (Amersham) and separated from unincorporated nucleotides bygel filtration(Centri-sep columns; Princeton Separations,Inc., Adelphia, NJ). The probes for in situ hybridization were generated in the pres-enceof35S-UTP. Aftertranscription,thetemplatewasdigestedwith DNase and the products were recovered by phenol extraction and eth-anolprecipitation. The specific activity was- _{2 x 108dpm/,ug}

tem-plate.

Northernblot hybridization. Primaryorsubcultured(P1 to P8) chondrocytesweremaintained in T175 flasks andstimulatedas indi-catedfor each experiment. Total RNA was extracted by the single step guanidiniumthiocyanate-phenol-chloroform method(20). 10-30 Mg of total RNA was separated on 1% formaldehyde gels,blottedonto

nylon filters,andcross-linkedwith UV light for 5 min each side. The blots wereprehybridized in 50% formamide, 6x standard saline citrate (SSC), 0.5% SDS, 0.1% Tween 20, 100Mg yeast RNA/ml for 15 minat

65°C. Theprehybridization mixture was replaced with fresh solution containing 106 cpm/ml of probe. Hybridizationwasperformed

over-nightat65°Candfollowedbywashes inlXSSC, 0.1% SDS at room temperature(2 x 30min)and 0.1X SSC, 0.1% SDSat65°C(2X30 min).The damp filterswereexposedtoKodak XAR film at -70°C for 2-24 h. To confirm equal RNA load and complete transfer the 18S and 28S bandswerevisualized with ethidiumbromide on all filters. In addition,RNA loadwasexamined byprobingfor,B-actinmRNA. For

mostaspects ofMCP-1mRNAexpressionatleast three separate exper-imentswereperformed.

Insitu hybridization. Cartilage organ cultureswerepreparedas

outlined above. Afterappropriateincubation time the sliceswerecut perpendiculartothesurfaceintotwopiecesof equalsize. Theywere

snapfrozen in dryice/2-methyl-butaneand storedat-70°Cuntiluse.

Frozen sections (5 Mm) were cut and mountedontoprecleaned slides (FisherScientific,SanFrancisco, CA). Theywerefixed in4% parafor-maldehyde,washed in l xPBS,andtransferredto2x SSC.

Hybridizationwasperformedasdescribedelsewhere(21 ).In brief, thespecimenswere acetylatedin triethanolaminewith acetic anhy-dride andincubated inglycinebuffer.Hybridizationwascarried out withI0 cpm/Alofprobe in 50%formamide,3xSSC,500Mug/mlyeast tRNA, 1 mg/mlsingle-strandedcalfthymus DNA,2mg/ml BSA, 10 mMDTT,and 1%polyethylene glycol overnightat50°C. Posthybridi-zationtreatmentconsisted of washes with 2xSSC,50%formamideat

50°C,incubation with RNase solution and further washes in 2X SSC at 55°Cand thereafter in2xSSCat roomtemperature.Specimenswere

dehydrated, air-dried, and covered with Kodak NTB nuclear track emulsion(Eastman Kodak)forautoradiography. After exposure for 3-6 dat4°C,the slidesweredeveloped, fixed,and counterstained with Giemsa. Cellswereconsideredpositive,when therewere>20 granules

Figure 1.

IL-l-stimu-so latedcartilage secretes

biologically

active

64

1 L ^ ^ MCP-I. Cartilage frag-0 48 4 4 4 4 ments (from a 37-yr-old

donor)wereincubated

o 32 o o o o inDMEM+

I%

FBS

o o o o o for 24h with or without

E _{IL-I# (10 ng/ml).}

Conditioned media

0 wereanalyzedfor

monocytechemotactic " _°''

_t<'~

#._O0

_{activity usinga}

multi-wellBoyden chamber assay as described in Methods.Conditionedmediaand rMCP-I(0.1 ml)were pretreated with0.01mlbuffer, or murineanti-MCP-l Ab, or nonimmune mouse IgG at a final concentration of 3.8 x 108 M. Human rMCP- 1 was used as achemoattractantata final concentra-tion of 2x 10-' M.Theresults.represent mean values of triplicate determinations.

probe or with an unrelated antisense probetoserveasnegative controls foreach condition.

Reagents. The following reagentswere used: recombinanthuman

TGF,3l,PDGF,IL-I#,tumor necrosisfactoralpha(TNFa), leukemia inhibitory factor (LIF), andIL-6(R&DSystems,Minneapolis,MN); retinoic acid, dexamethasone, phorbol myristate acetate (PMA),

cAMP, and LPS from SalmonellaMinnesota(SigmaChemicalCo.).

Results

IL-I-stimulated cartilagesecretesbiologically active MCP-J.

To assesswhether chondrocytes have thecapacityto release

biologically active MCP-1, conditioned media ofcartilage in organ culturesweretested inahuman monocyte chemotaxis assay. Theconditionedmedia of IL- 1-stimulated,butnot

un-stimulated cartilage,contained monocyte chemoattractant

ac-tivity comparabletothebioactivity elicitedby 2 nM human

recombinant MCP-1 (Fig. 1). Conditioned media from 24-h

culturesweretested since this isanoptimaltimepointfor the

induction ofcytokine release by IL-1 in chondrocytes. The

chemotacticactivityof both rMCP-1 and theconditioned

me-dia ofIL-1-stimulatedcartilage,werenearly completely

abro-gatedbypretreatment with a monoclonal, neutralizing

anti-bodyto MCP-1, but notwith nonimmune murineIgG (Fig.

1).Specificityof thisassay systemfor MCP-1-induced mono-cytechemotaxiswasvalidated by: (a)establishingthat the

hu-man neutrophil and T lymphocyte chemotactant, IL-8 (10 nM) didnotinduceachemotacticresponse(not shown); and

(b)the lackofaneutralizingeffect of identicalpretreatmentof

10 nM FMLP with monoclonalanti-MCP-1 antibodyonthe

elicitedmonocytechemotaxisresponse(90±3monocytes/five high power fields without anti-MCP-1, 110±18 monocytes/

fivehighpowerfields with anti-MCP-1, n =_{3). Furthermore,}

theaddition ofIL-1indoses up to 10 ng/ml to culture media

did notinduce detectable monocytechemotaxis. Thus,

acti-vatedchondrocytes inorganculture secretedbiologicallyactive MCP- 1, which was the predominant chemoattractant for

monocytes intheir conditioned media.

Pattern of chondrocyte MCP-J mRNA expression in

re-sponsetoIL-I incartilageorganculture. Manyparameters of

theresponsivenessofchondrocyteswithincartilageto

stimula-tion with IL- I havenotpreviouslybeencharacterized. To

ex-amine kinetics,the distributionand number of cells that

ex-pressMCP-1 mRNA, cartilage fragments of identical size and shape containingalllayersofthetissue,were cutand cultured in the presence orabsenceof IL-1 (10 ng/ml). Fig.2 shows that unstimulated cells (Co) remained negative for MCP-l

mRNA as assayed by in situhybridization.By 2 h after

stimula-tion with IL-1, themajority of the cells in thesuperficial

tan-gentialzoneshowedsomepositivityforMCP-l mRNA(white

granulesarecoveringcells indark-fieldpicture). However,by

4h, superficial cellswereagain negative,while cells indeeper

location (round cells in chondrons) showed an increase inthe numberofpositive cells andintheintensityofthe

hybridiza-tionsignalper cells. After 6 h itwaspredominantly chondro-cytes indeepercell layersthat were strongly positive(bottom panel). Additional experiments showed that cells in deeper

location remained positive for at least 18 h(not shown). Stimulation with LPS (1

gg/ml),

whichreadily activates MCP- 1 expression in chondrocyte monolayer cultures(see re-sultsbelow)resulted only in a weak response ofcells in

superfi-cial location. Only in the presence of high concentrations of LPS(>10Ig/ml) some cells in deeper layers responded.

Collec-tively, these results demonstrate that chondrocytes when surroundedbyextracellular matrix incartilageorgan culture respond tostimulation with IL-l with the expression ofMCP- 1 mRNA. The response is rapid, and the IL-I effect ispropagated through cartilage matrixto chondrocytesindeeperlocations. In contrast, LPS penetration through thematrixappears tobe

poor, and wasassociated largely with activation of cells in the superficial zone.

Chondrocytes synthesize and release the13-and15-kD

iso-forms

ofMCP-J. IL-1-activated cartilage expressed MCP-l

mRNAand secretedbiologically activeMCP-1. We next ana-lyzed whetheractivated chondrocytes synthesize thiscytokine and characterized the MCP- 1 proteins produced. Primary

chondrocytes were cultured in the presence of [35S]methionine

and[35S]cysteine for48 h, andstimulated withIL-1 (10 ng/

ml)or LIF(10 ng/ml). Conditioned mediawere immunopre-cipitated, and analyzed by a 15%SDS-PAGEunderreducing

conditions (Fig. 3).Chondrocytessynthesized two isoformsof

MCP- 1 in similar quantity, and ofthecharacteristic13-kD and 15-kDsizes ( 12).Stimulationwith not onlyIL-1,butalso with LIFinduced a markedincreasein the intensity of the bands,

demonstrating upregulation ofproteinsynthesisinresponse to these cytokines.

Regulation ofMCP-J geneexpressioninchondrocytes.We

further evaluated regulation of MCP-1 expression using cul-tured chondrocytes. Northern blotanalysis showed that

pri-Figure 2.Time-andsite-specific expression of MCP-1 mRNA by chondrocytes in cartilage organ culture. Cartilage fragments were cultured in DMEM+ 1%FBS in the absence (Co) or presence of IL-I(10ng/ ml)for 2 h, 4 h, and 6 h. MCP-I mRNAexpression was analyzed by in situ hybridization (positive cells are covered with white granules in dark-field pictures). Unstimulated chondrocytes were negative (Co). After 2-h stimulation positive cells were mainly located in the superficial tangential zone. After 4 h and 6 h superficial cells became negative, but abundant MCP-ImRNA was then found in cells in the deeper layers. Each section was photographed in bright field ( left panels) or dark field ( right panels).

i t

_{~~~~~~~~~~evl}

if

~~~~~~~~~~~.Y _istilers'

''':

A_{d~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~p}

*.*

, ..., ''...:.. :*

0::$f,

-'

:

...

:,.

:'

...

....:'

.

,

!.i:.

yA;}

q...'..

,'

a,^..

...',, .,_.'.

k

Vorf'*'

492 P.M. Villiger, R. Terkeltaub,andM. Lotz

C

IL-i

LIF

Figure 3. Chondrocytes

secretede novo synthe-sizedMCP-l in re-sponse to stimulation

21.0 kD-

withIL-l

orLIF.Pri-marychondrocyteswere

20.1 kD

-

cultured for 48

h

in the

presence of [35S]-methionine and

[35SJ-cysteine. Conditioned

14.2

kD-

s

mediawereimmuno-precipitatedand

ana-lyzed bySDS-PAGE underreducing condi-tions.Lane I shows

un-* . stimulated.... cells,lane 2stimulation withIL-I,

lane3 stimulation with LIF(bothstimuliat aconcentrationof 10 ng/ml). MCP-I proteins appearasthecharacteristicdoubletat13 kD and 15kD.

marychondrocytes cultured inthe presenceofserum consis-tently expressed 0.7-kb MCP-l mRNA (see results below). This decreased tolow levelsafter overnight serum-starvation (Fig. 4, lane 1) and remained lowover aperiod ofatleast48h

(not shown).IL-Ifl,TNFa, andLPS(lanes 2, 5, and 6,

respec-tively) induced a marked increase in MCP-l mRNA levels while IL-6, an important regulator ofchondrocyte function (22, 23) had no detectable effect (lane 3). Surprisingly, LIF

causedonlyaslightincrease inMCP-l mRNAlevelat5 h(lane

4). Thiswasnotexpectedbecauseofthe markedeffect ofLIF onMCP-1 synthesis(Fig. 3) and suggested that the kinetics of

MCP-lexpression inresponse to LIFdiffered from the kinetics inresponse to IL- 1.

Wethus compared the kinetics ofMCP-l mRNA

expres-sion inresponse toIL-1, LIF, andTGF,3(Fig.5), and observed that LIF induced a very rapid increase in MCP-1 mRNA, which reached its maximal levels by 2 h.By 5 h, themRNA

levels hadalready returned tobaseline (Fig. 5 B). As we

re-cently showed thatLIFincreases theexpressionofLIF mRNA

(24) thesameblot that had beenprobed for

MCP-l

was hybrid-izedto alabeledLIF RNAprobe. This showed that incontrast toMCP-1,LIF mRNAlevels remain elevated foratleast9 h. Theveryrapid and transient induction ofhigh MCP-l mRNA

levelsbyLIF wasconfirmed intwoadditionalexperiments.

TGF,3wasidentifiedas anadditionalnewinducer of

MCP-C IL-I IL-6 LIF TNF LPS Figure 4. Chondrocytes expressMCP-l mRNA. Primary chondrocytes

(froma31-yr-old do-nor)werestimulated for MCP-1- L 5h withIL-1(10ng/

ml;lane 2),IL-6(10 ng/ml;lane 3),LIF(10 ng/ml; lane 4), TNFa, (I 0 U/ml;lane5), LPS (I1,g/ml, lane 6);

a-actln-

unstimulatedcells in

lane 1. Total RNA was extracted andsubjected

toNorthernblotting.Afterhybridizationwithan

MCP-l

RNA-probe thefilterwashybridizedto a#-actinprobetoassesstheamountof

RNAper lane.

1 mRNA expression. Its effects occurred more slowly with a

maximumafter 5-8 h(Fig.5 C). MCP-l mRNA levels often

remained elevated formorethan 11 h in response toTGF,3and wereonlyslightlylower thanthoseseenafter stimulationwith IL-1. PDGF induced MCP-l gene expression with kinetics

comparabletothose observed with IL- I (not shown). Collec-tively, these experiments identify LIF andTGF# as two new

inducers of MCP-1. The distinct patterns of MCP-l mRNA

inductionsuggestthat IL-1,LIF,andTGF,3may use different

mechanismstoregulate MCP-I mRNAlevels.

Dexamethasone andretinoicacid

differentially

affect MCP-I gene expression.Glucocorticoidssuchasdexamethasone in-hibitIL- 1 responses in a varietyof cells and suppress the

ex-pressionof IL-l-inducible cytokines (25 ). Fig. 6Ashows that chondrocyte culture in 1%FBS is sufficienttoinduce

detect-able levels of MCP-l mRNA (lane 1). Dexamethasone (lane 3) abrogated theserum-induced MCP-l geneexpressionand

markedly reduced theresponse toIL-l (lane 4). Inaddition, dexamethasone inhibited MCP- 1 expression in response to

cAMP andto PMA(Fig.6B). Thus, dexamethasone isa gen-eralinhibitor of MCP-Iexpressioninchondrocytes inhibiting serum-induced(basal) and induciblegeneexpression.

Retinoic acid (RA)has potenteffectsondifferentiation of mesoderm-derived tissues (26) and inhibits IL-l-induced

ex-pression ofcollagenase and stromelysin and production of proinflammatory cytokines (27-29). Fig.7 Ashows that stimu-lation of chondrocytes with RA(1

gM)

in serum containing medium for 72hincreased MCP-l mRNAlevels(lane2).This RA concentration of1 ,uM was optimal for the induction of

theseeffects. Dexamethasone,asdocumentedabove, reduced serum-induced MCP-1 expression (lane 3), and this was

re-A IL-1 2h 2h 5h 8h 11h

IMICPd1

-I

p-actin-B LIF 2h 2h 5h 9h I8h

3~~~~~~~~~~~~~~~~i

-

,

p3-actin-C

_TGFIPi

2h 2h 5h 8h ilh

MCP-1

-

f3-actln-Figure 5. IL-1, LIF, and

TGF#induce MCP-I expressionwith

differ-enttimekinetics. Sub-culturedchondrocytes

wereserum-starvedfor 24 h and then stimu-latedwithIL- I(A),LIF

(B),or

TGF/B

(C) (all stimuliat10ng/ml)for timesindicated above the lanes.First lane in allpanels represent

un-stimulated cells.Total

RNA wasextracted,

subjectedtoNorthern blotting,andanalyzed withanMCP-l RNA

probe. Toassessthe

amountofRNAper

lanethe filterwas sub-sequentlyhybridizedto

Dex Figure6.(A)

Dexa-+ _{methasone reduces the}

IL-1 IL-IeffectonMCP-I

expression. Articular chondrocytes (P4)were

_ cultured in DMEM

+ 1%FBSfor24hand then stimulatedwith IL-i (10 ng/ml; lane 2), dexamethasone

(IO-7M; lane3),or

IL-1and dexamethasone

(lane 4)for 5 h. Lane1

shows RNA from

un-PMA stimulatedcells. RNA

wasextracted and

ana-lyzedbyNorthern

blot-ting.Toassessthe

amountofRNAper

lanethe filterwas

+ stained with ethidium

bromide. (B)

Dexa-methasonereduces

MCP-1 expressionin

tweensynovium and cartilage. Through the release of MCP- 1,

mononuclearphagocytesareattractedtocartilage and thismay

beonefactor that directs theproliferationof rheumatoid pan-nusontothe articular surface.Conversely,theactivated prolif-erating synovium isa source of factors suchasIL-1, TNFa,

LIF, TGF(B, and PDGF thatstimulateorperpetuate chondro-cyte MCP-1 expression and this can create a circuit that is

likelytoresult in cartilagedestruction.

Havingestablishedthat chondrocytescanproduce MCP-1, this study definedcharacteristicsof expression of thisgenethat

areimportantoruniquetocartilage. The first pointaddressed

the role of extracellular matrix in the induction andsecretion

ofMCP-1. We havedemonstratedthat cells in cartilageorgan

culturesarenotonly inducedtoproduce MCP-1, but that bio-logically active MCP-1 issecretedfrom the tissue into the

cul-ture media. Although MCP-l is a heparin-binding protein,

substantialamountsofthiscytokinewerereleased from

carti-lage, which contains a proteoglycan-rich matrix. The

func-A

responsetocAMP and PMA. Articular chondrocytes(P3)were

cul-tured in DMEM+1% FBSfor 24 h and thenstimulated withcAMP

(250,g/ml)orPMA(20 ng/ml)inthepresence(+)orabsence(-)

ofdexamethasone ( 10-' M)for5 h. LaneIshows RNA from

unsti-mulated cells, lane2showsthe effect.

RA

C RA Dex Dex

so 'fM> -.

MCP-1

-versed by RA (lane 4). Furthermore,asshown inFig.7B,RA

not only elevated serum-induced MCP-1 expression but also

markedly upregulatedtheMCP-l responsetoIL-1,LPS,and PMA (Fig. 7 B). To test whether this effect was limited to

MCP-1 geneexpression or was a more general phenomenon

thesameNorthernblotwassequentially probedforIL-6, IL-8, and LIF mRNAs. With all stimuli that induceexpression of thesecytokines amarked upregulationby RA wasobserved. Thus, incontrasttoitseffectoncytokine expressionin

mono-cytes(29),RAisapotentstimulusof MCP-1geneexpression

in chondrocytes and can upregulate the effect of different

agents that utilize distinct mechanisms in the induction of MCP-1.

Discussion

Thisstudy shows theexpression ofbiologicallyactive MCP-l by chondrocytesinhumanarticularcartilage.We have charac-terizedthe kinetics and distribution of cellsexpressingMCP-1

in cartilage organ culture by in situ hybridization and have

carriedoutadetailedanalysisofinducers, inhibitors,and

sig-nal transduction pathways, identifying LIF, TGF,3, and RAas newactivators of MCP-1 expression.

MCP-1isamember ofarecently identified family of proin-flammatory cytokines (14, 15). Its principal biological func-tionsaremonocytechemotaxisand activation.Thefindings of this study suggest that, through the secretion of biologically active MCP- 1,chondrocytes have thepotentialto attract

mono-cytesintosynovial fluid and membrane,toactivate oxidative metabolism andproteaserelease and thustoactivelypromote

cartilagedegradation.Theproductionof MCP- I by chondro-cytes also representsanimportantpathogeneticinteraction

be-

Et-bromide-IL-1 LPS PMA

lng Ong I O1ng

MCPl--RA - ₊ - +

Et-bromide-- +

-Figure 7.Dexamethasoneandretinoic aciddifferentiallyaffect

MCP-I expression. (A)Articular chondrocytes(P3)wereculturedin

DMEM+ 1%FBS in the absence(lanes I and 2)orpresence(lanes

3 and4)of retinoic acid(RA)for3 d. Dexamethasone(10-'M;lanes

2and4)wasadded for the last 5 h. RNAwasextractedandanalyzed

byNorthernblotting. Toassesstheamountof RNAperlanethe

filterswerestained with ethidium bromide.(B)Retinoic acid

in-creasesthe induction of MCP-l mRNA byIL-1,LPS, andPMA.

Chondrocytes(P3)wereculturedinDMEM+10%FBS in the

pres-ence(+)orabsence(-)of retinoic acid(10-6M)for 3 d.Thecells

werethenstimulatedwith IL-I (lanes 1-4),LPS(lanes 5-8),orPMA

(lanes 9-12)atconcentrationsasindicated above the lanesfor 4 h.

RNAwasextractedandanalyzedasabove.

494 P.M. Villiger, R. Terkeltaub,and M. Lotz

A C IL-1 De:

MCP-1-MCP>1 $

Et-bromide-Co cAMP

B

MCP-1

Dex - + +

tionalsignificance of the heparin binding of MCP-l hasnotyet beendetermined. It could provide a basis for storage ofthis

cytokine incartilage extracellular matrix. Apotential roleof heparinbindinginmodulatingMCP- 1bindingtoits receptors

asithas beenshown for bFGF (30), alsoneeds to beevaluated.

Insituhybridization experiments providednewinsight into the kinetics and distribution of IL-l effects on chondrocyte

geneexpression in cartilageorgancultures. Cells in the superfi-cialtangentialzoneexpressed MCP-l already 2 h after

stimula-tion with IL-1. Thisresponse wastransientand mRNAlevels had alreadydecreasedafter4h. Interestingly, the response of the cells in the deeper layers ofcartilagewas notonly delayedin onset but of much longer duration. Since the effect of IL-l aloneonMCP- 1 is usually fast and brief, the prolonged expres-sionbychondrocytes in deeper layersmaybe duetosecondary

events.Oneinterestinghypothesis is that thismaybe dueto the

expression ofTGF,3which can cause a similarlylong lasting increase in MCP-1 mRNAlevels. Inprevious studieswe

dem-onstrated that IL-l induces a late increase in

TGF#lI

and a

rapidincrease in TGF,33 mRNAlevels in chondrocytes (5). Thesedifferences in the kinetics ofMCP- Iexpressionmay also

beindicativeof functionalheterogeneity ofchondrocytesin the different layers of cartilage.

Thecharacterizationofinducersof MCP-I expression that

are presentincartilageorsynovium resulted in the identifica-tion oftwopeptideregulatoryfactorsas newactivatorsof this

gene. Inadditionto IL- 1,TNFa,LPS,PDGF, and serum, previ-ously knowninducers of MCP-1in other celltypes, we

demon-stratethatTGF3 andLIFstimulatethis cytokine in

chondro-cytes.Except for the induction of boneresorptioninvivo and invitro (31, 32), LIFhasnot yetbeencharacterized with re-spect toits effectsonconnective tissue metabolism. The find-ing thatLIFinducesMCP-l expressionrepresentsthefirst doc-umentation that LIF canregulate thesynthesis of other

cyto-kines and that it affects chondrocyte function. We recently showed thatchondrocytesandsynoviocytesproduceLIFand thatsynovialfluids containLIFactivity (24).Basedonits pres-ence in thejoint, LIF is probably an important inducer of

chondrocyte MCP-l productioninvivo.

TGFB is abundant ininflammatory jointfluids(33,34).It

isapotentregulatorof chondrocytefunctionand also known

to profoundly affect specific as well as nonspecific immune reactions (reviewed in 35).IncartilageTGF#hasmainly

ana-bolicfunctions.It promoteschondrocyteproliferation (5, 23),

thesynthesis of extracellular matrix (36)andit reduces IL-l induction of collagenase (37) and theexpression ofIL-1 recep-tors onchondrocytes (38). Amongits effects onimmune

re-sponses,TGFBhas beenshowntoinhibitTandBlymphocyte proliferation, to modulate immunoglobulin synthesis and to

deactivate monocytes. Incontrast to these antiinflammatory actionsarethepotenteffects of TGFBas aneutrophil (39)and monocyte chemoattractant in vitro and in vivo (40, 41). In

addition,weshowed recentlythat

_TGF,#

induces production of

the neutrophil chemoattractant, IL-8, by chondrocytes (4). This study has demonstrated that

TGF3

induces MCP- I mRNAlevels comparableto thelevels seenafter stimulation withIL- 1. Thus, TGF, candirectlyandindirectly contribute

totherecruitment ofneutrophilsand monocytesviathe induc-tion ofIL-8 and MCP- 1.

Twoantiinflammatory steroids, dexamethasone and RA, which inhibit cytokine synthesis in other cell systems, were shown todifferentially affect MCP-1 geneexpression in

chon-drocytes.As has been shown for othercytokinesand for MCP-1 in other cell types (25), dexamethasone inhibited MCP-1 mRNA expression. Glucocorticoids bind to intracellular

re-ceptors that recognize glucocorticoid-responsive elements in thepromoterregionsofgenes.Inaddition, glucocorticoidscan modulateAP-l DNAbinding activityandtherebyinhibitgene expression(42). Dexamethasone abrogated MCP-l mRNA

in-duction after stimulation not only with IL-l but also in

re-sponsetodirect activators ofprotein kinaseA(cAMP)and the

proteinkinase C(PMA) signal transductionpathways.

Incontrast,RA wasidentifiedas a new andpotent inducer

of MCP- 1 geneexpressioninchondrocytes.Theeffectsof RA ongeneexpression dependonitsbindingtointracellularRA

receptors. TheRA/receptor complexes interact with retinoic

acidresponse elements (RARE)inthepromoter region of sev-eralgenes(43, 44).As forglucocorticoids,aninteractionwith

AP-l binding proteinshasbeendescribed (45, 46). Inthe case ofRA, the formation of unproductive complexes with c-Jun is thoughtto lead to adecreasein AP-l activity (47). The

pro-moterregion of MCP-1 has recently been sequenced,but de-tailedanalyses concerningfunctionallyimportant regulatory elements are lacking. A phorbol ester responsive element (TRE) confers PMAresponsivenessandthis has been shown

tobefunctionally important in fibroblasts ( 17). Interestingly, PMA caused only a minimal increase in MCP-l mRNA in chondrocytes,butafter preincubationofthe cellsfor3d with

RA,PMAinducedaverystrongincrease in mRNAlevels.In

additiontothe PMA response, MCP-l induction by IL-l and

LPSwasalsoupregulated byRA. Thesefindingsareincontrast to recentresults obtained with human monocytes, whereRA

inhibited IL- 1-induced synthesis of IL-6 (29). This suggested that either the regulation of IL-6 induction by IL-Idiffers from the regulation of MCP-1 by IL-1, orthat there are cell-type specific factors which account for these contrasting results.

Analysisofgeneexpressionofothercytokines(IL-6, IL-8, LIF) revealed that the stimulated expressionof all these cytokines

wasmarkedlyupregulated inchondrocytesby pretreatment of the cells with RA. Furthermore, AP-I DNA-bindingactivityas

analyzed by mobility shiftassaydocumented that in

chondro-cytesRAtreatment for 72 h resulted inan increase in AP- 1

activity(unpublished results). Thus, the profound stimulatory effect of RAonexpression of MCP-1 inchondrocytes,

appar-entlyrepresents atissuespecificdifference in the effect ofthis agent on geneexpression,andmaybe relatedto anincreasein

AP-1 activity.

Inconclusion,through the production ofbiologicallyactive MCP-1inresponse toinducers thatare presentin thearticular

tissues of patients with inflammatory anddegenerativejoint

diseases, cartilagecandirectlycontributetosynovial inflamma-tion and destrucinflamma-tion of its extracellular matrix. Furthermore,

understandingoftheregulationofcytokineandgrowth factor-induced MCP-l expression incartilageshould be valuable in the understanding andcontrol of the clinical manifestations andsequelaeof arthritides.

Acknowledgments

Wegratefully acknowledgeDenise Santoro andJacquelineQuachfor experttechnical assistance. Theencouragement,andgenerous

provi-sion ofimportant reagents by Drs. EdLeonard,TeizoYoshimura, and TomLane,isappreciated.

Naval Research (ONR N00014-90-J-1637) andby aMerit Review

Award from the VeteransAffairs Research Service. We thank Drs. C.

Colwell,K.Benirschke, R. D.Coutts,and Glenn Greenleaf from the

UniversityofCalifornia,SanDiego,tissue bank forprovidingthe

carti-lagesamples used in this study.

References

1.Pujol, J. P., andG.Loyau. 1987.Interleukin-Iandosteoarthritis.LifeSci.

41:1187-1198.

2. Towle,C. A., M. E.Trice,F.Ollivierre,B. J.Awbrey, and B. V. Treadwell. 1987.Regulation of cartilage remodelingby IL-1:evidence for autocrine synthe-sisofIL-I bychondrocytes.J.Rheumatol.14:11-13.

3. Guerne, P. A., D.A.Carson,and M. Lotz.1990. IL-6production byhuman

articularchondrocytes.Modulation of itssynthesis by cytokines, growthfactors,

and hormones invitro.J.Immunol. 144:499-505.

4.Lotz, M., R.Terkeltaub,and P. M.Villiger.1992.Chondrocytesandjoint

inflammation: expressionofIL-8inresponse topeptide regulatory factorsand

proinflammatoryagents. J. Immunol. 148:466-473.

5.Villiger, P. M., and M.Lotz.1992.DifferentialexpressionofTGFjB

iso-formsby humanarticular chondrocytesin response togrowth factors.J.Cell.

Physiol.151:318-325.

6.Peracchia,F., G.Ferrari,A.Poggi,and D.Rotilio.1991.IL-1beta-induced

expression ofPDGF-AAisoforminrabbitarticular chondrocytesis modulatedby

TGF-beta 1. Exp.CellRes. 193:208-212.

7.Louthrenoo,W., M.Sieck, G. Clayburne,S.Rothfuss,and H. R.

Schu-macher, Jr. 1991.Supravitalstaining ofcells innoninflammatory synovial fluids:

analysis oftheeffect of crystalsoncellpopulations.J.Rheumatol. 18:409-413.

8.Postlethwaite,A., andA.H.Kang.1976.Collagenandcollagen

peptide-in-ducedchemotaxis ofhumanbloodmonocytes. J.Exp. Med.143:1299-1307.

9.Matsushima,K.,C.G.Larsen,G. C.DuBois,and J. J.Oppenheim. 1989.

Purificationandcharacterizationofanovel monocytechemotacticand

activat-ing factor produced by a human myelomonocytic cell line. J. Exp. Med.

169:1485-1490.

10.Robinson,E.A., T.Yoshimura,E.J.Leonard,S.Tanaka,P.R.Griffin,J.

Shabanowitz,D. F.Hunt,and E.Appella. 1989.Complete aminoacid sequence

of a human monocyte chemoattractant,aputativemediatorofcellularimmune

reactions.Proc.Nati. Acad.Sci.USA. 86:1850-1854.

11.Furutani,Y., H. Nomura, M.Notake,Y.Oyamada,T.Fukui,M.

Ya-mada,C. G. Larsen, J. J.Oppenheim,and K.Matsushima.1989.Cloningand

sequencing ofthe cDNAforhuman monocytechemotacticandactivatingfactor

(MCAF).Biochem.Biophys.Res.Commun. 159:249-255.

12.Yoshimura,T., E. A.Robinson,S.Tanaka,E.Appella,J.Kuratsu,and

E. J.Leonard.1989. Purificationandamino acidanalysis oftwohuman

glioma-derived monocyte chemoattractants. J.Exp. Med. 169:1449-1459.

13.Yoshimura, T., N. Yuhki, S. K. Moore, E. Appella, M. I. Lerman, and E. J. Leonard. 1989. Human monocyte chemoattractantprotein-l (MCP-1).

Full-length cDNAcloning, expressioninmitogen-stimulatedblood mononuclear leukocytes, and sequencesimilarityto mousecompetence gene JE. FEBS(Fed.

Eur.Biochem.Soc.)Lett.244:487-493.

14.Oppenheim, J. J., C.0.C.Zachariae, N. Mukaida, and K.Matsushima.

1991.Properties ofthe novelproinflammatorysupergene"intercrine" cytokine

family.Annu.Rev.Immunol.9:617-648.

15.Leonard, E. J., and T. Yoshimura. 1990. Human monocyte

chemoattrac-tantprotein-I(MCP-I).Immunol. Today.11:97-101.

16.Zachariae,C. O., A.0.Anderson, H. L. Thompson, E. Appella, A.

Man-tovani,J.J.Oppenheim, and K. Matsushima. 1990.Propertiesofmonocyte

chemotactic andactivating factor (MCAF) purified fromahumanfibrosarcoma

cellline.J. Exp. Med. 171:2177-2182.

17.Shyy,Y.J., Y. S.Li,and P. E.Kolattukudy. 1990. Structureofhuman

monocyte chemotactic protein gene and its regulation by TPA. Biochem.

Biophys.Res. Commun. 169:346-351.

18. Falk, W., R. H. Goodwin, Jr., and E. G. Leonard. 1980. A 48-well

micro-chemotaxisassemblyfor rapid and accurate measurement of leukocyte

migra-tion. J. Immunol.Methods.33:239-249.

19.Yoshimura,T., M. Takeya, K.Takahashi, J.-I. Kuratsu, and E. J.

Leon-ard. 1991. Production and characterization of mouse monoclonal antibodies

against human monocytechemoattractant protein- 1 (MCP-1). J. Immunol.

147:2229.

20. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA

isola-tionby acid guanidiniumthiocyanate-phenol-chloroform extraction. Anal.

Bio-chem. 162:156-159.

21.Villiger,P.M., M. T. Cronin, T.Amenomori,W. Wachsman, and M.

Lotz. 1991. IL-6 production by human T lymphocytes. Expression in HTLV-

1-infectedbut not in normal T cells. J.Immunol.146:550-559.

22.Lotz,M., and P.-A. Guerne. 1991. Interleukin-6 induces the synthesis of

tissueinhibitorofmetalloproteinases-l/erythroid potentiating activity

(TIMP-1/EPA).J.Biol.Chem.266:2017-2020.

23.Guerne, P.-A.,and M. Lotz.1991. Interleukin-6 andtransforminggrowth

factorP synergisticallystimulatechondrosarcoma cellproliferation.J.Cell.

Phys-iol. 149:117-124.

24.Lotz, M.,T.Moats,and P.Villiger.1992.Leukemiainhibitory factor is

expressedincartilageandsynovium.Potential contributionstothepathogenesis

of arthritis.J.Clin.Invest. In press.

25.Mukaida, N.,C.C.Zachariae,G. L.Gusella,and K.Matsushima.1991.

Dexamethasoneinhibitstheinduction ofmonocytechemotactic-activating

fac-torproductionby IL-lor tumornecrosis factor. J.Immunol. 146:1212-1215.

26.Giguere, V.,E.S.Ong,R. M.Evans,andC. J.Tabin.1989.Spatialand

temporal expression oftheretinoic acidreceptor in theregeneratingamphibian

limb. Nature(Lond.).337:566-569.

27.Lafyatis,R., S. J.Kim,P.Angel,A.B.Roberts,M. B.Sporn,M.Karin,

and R. L. Wilder. 1990.Interleukin-1 stimulates and all-trans-retinoic acid in-hibitscollagenasegeneexpression throughits 5'activatorprotein-1-bindingsite.

Mol.Endocrinol.4:973-980.

28.Nicholson,R.C.,S.Mader,S.Nagpal,M.Leid,C.Rochette-Egly,and P.

Chambon. 1990.Negative regulationoftheratstromelysingenepromoterby

retinoic acid ismediatedbyan APIbindingsite. EMBO(Eur.Mol.Biol.Organ)

J.9:4443-4454.

29.Gross, V.,P. M.Villiger, andM.Lotz. 1991.Retinoic acidregulates

monocytefunctionandinhibitsIL-I-induced geneexpression.ArthritisRheum.

34:S144.(Abstr.)

30.Yayon, A.,M.Klagsbrun,J. D.Esko,P.Leder,and D. M.Ornitz.1991.

Cellsurface, heparin-likemoleculesarerequiredforbindingof basicfibroblast

growth factortoitshigh affinityreceptor.Cell.64:841-848.

31.Reid,L.R.,C.Lowe,J.Cornish,S. J.Skinner,D. J.Hilton,T.A.Willson,

D. P.Gearing,and T. J.Martin.1990.Leukemiainhibitoryfactor:anovel

bone-activecytokine.Endocrinology. 126:1416-1420.

32.Hilton,D. J.,and N. M. Gough. 1991.Leukemiainhibitoryfactor:a

biological perspective.J.Cell. Biochem.46:21-26.

33.Fava, R.,N.Olsen,J.Keski-Oja,H.Moses,and T.Pincus. 1989. Active

and latentforms oftransforming growthfactorbetaactivityinsynovialeffusions.

J.Exp. Med.169:291-296.

.34. Lotz, M.,J.Kekow,and D.A.Carson. 1990.Transforming growth

factor-betaand cellular immune responses insynovialfluids. J. Immunol. 144:4189-4194.

35.Palladino,M. A., R. E.Morris,H. F.Starnes,andA.D.Levinson. 1990.

Thetransforming growth factor-betas.A newfamilyofimmunoregulatory

mole-cules.Ann.NYAcad.Sci. 593:181-187.

36.TenDijke,P.,and K. K. Iwata. 1989. Growthfactors forwoundhealing. Biotechnology.7:793-797.

37.Chandrasekhar,S.,andA.K.Harvey.1988.Transforming growth

factor-beta isapotent inhibitorofIL-1inducedproteaseactivityandcartilage

proteogly-candegradation.Biochem.Biophys.Res.Commun. 157:1352-1359.

38.Harvey,A.K.,P.S.Hrubey,andS.Chandrasekhar. 1991.Transforming

growth factor-beta inhibition of interleukin-1activityinvolvesdown-regulation

ofinterleukin-1 receptorsonchondrocytes. Exp.CellRes. 195:376-385.

39.Brandes,M.E.,U.E. H.Mai,K.Ohura,and S. M.Wahl.1991.TypeI

transforming growth factorbeta receptorsonneutrophilsmediatechemotaxisto

transforminggrowth factorbeta. J.Immunol.147:1600-1606.

40.Wahl,S.M.,D.A.Hunt,L. M.Wakefield,N.McCartney-Francis,L. M.

Wahl,A.B.Roberts,and M. B.Sporn. 1987.Transforming growth factortype

beta induces monocytechemotaxisandgrowthfactorproduction.Proc.Natl.

Acad.Sci. USA. 84:5788-5792.

41.Allen,J.B.,C. L.Manthey,A.R.Hand,K.Ohura,L.Ellingsworth,and

S. M. Wahl. 1990.Rapidonsetsynovialinflammationandhyperplasiainduced

bytransforminggrowthfactor beta. J.Exp. Med. 171:231-247.

42.Jonat,C.,H.J.Rahmsdorf,K. K.Park,A.C.Cato,S.Gebel,H.Ponta,

and P.Herrlich. 1990.Antitumorpromotionandantiinflammation:

down-modu-lationof AP-1 (Fos/Jun)activitybyglucocorticoidhormone.Cell.

62:1189-1204.

43.Petkovich, M.,N. J.Brand,A.Krust,andP.Chambon. 1987.Ahuman

retinoic acid receptor whichbelongstothefamilyof nuclear receptors. Nature

(Lond.).330:444-450.

44.Mangelsdorf,D.J.,E. S.Ong,J.A.Dyck,and R. M. Evans. 1990. Nuclear

receptor thatidentifiesanovelretinoic acid responsepathway.Nature(Lond.).

345:224-229.

45.Schule, R.,K.Umesono,D.J.Mangelsdorf,J.Bolado,J. W.Pike,and

R.M.Evans. 1990. Jun-Fos and receptorsforvitaminsAandDrecognizea commonresponse element in the humanosteocalcingene.Cell. 61:497-504.

46.Schule,R.,P.Rangarajan,N.Yang,S.Kliewer,L. J.Ransone,J.Bolado,

1.M.Verma,and R. M. Evans. 1991.Retinoic acid isanegative regulatorof

AP-l-responsivegenes. Proc.Natl.Acad.Sci. USA. 88:6092-6096.

47.Schule,R.,P.Rangarajan,S.Kliewer,L. J.Ransone,J.Bolado,N.Yang,

I.M.Verma,andR. M. Evans. 1990.Functionalantagonismbetween

oncopro-tein c-Jun and theglucocorticoidreceptor. Cell.62:1217-1226.