Effect of Hypoxia

8/13/2019 Effect of Hypoxia

1/12

ARTHRITIS & RHEUMATISM

Vol. 50, No. 11, November 2004, pp 35493560

DOI 10.1002/art.20596

2004, American College of Rheumatology

Effect of Hypoxia and Reoxygenation on Gene Expression and

Response to Interleukin-1 in Cultured Articular Chondrocytes

G. Martin,1 R. Andriamanalijaona,1 S. Grassel,2 R. Dreier,3 M. Mathy-Hartert,4

P. Bogdanowicz,1 K. Boumediene,1 Y. Henrotin,4 P. Bruckner,3 and J.-P. Pujol1

Objective.To determine the effects of hypoxia and

reoxygenation on the metabolism of chondrocytes and

their response to interleukin-1 (IL-1). The study

included activation of hypoxia-inducible factor 1 (HIF-

1), NF-B, and activator protein 1 (AP-1) transcription

factors, expression of matrix components and metallo-proteases and transforming growth factor (TGF)

and TGF receptors, and production of nitric oxide

(NO) and prostaglandin E2

(PGE2

).

Methods. Bovine articular chondrocytes (BACs)

were cultured to confluency in either 5% O2

(hypoxia) or

21% O2

(normoxia) in media supplemented with 10%

fetal calf serum (FCS). BACs were preincubated for 18

hours in media with 1% FCS only and then incubated

for 24 hours in the presence of IL-1. For reoxygenation

experiments, cells were treated in the same way in 5%

O2

, except that cultures were transferred to normal

atmospheric conditions and used after 4 hours for RNA

extraction or after 30 minutes for cytoplasmic or nu-clear protein extraction.

Results. In hypoxic and reoxygenated chondro-

cytes, we observed strong DNA binding of HIF-1. IL-1-

induced DNA binding of NF-B and AP-1 was signifi-

cantly higher in hypoxic and reoxygenated cultures than

in normoxia. Greater activation of the MAPKs was also

observed with IL-1 treatment in hypoxia compared

with normoxia. Steady-state levels of type II collagen

and aggrecan core protein messenger RNA (mRNA)

were decreased by IL-1 in all instances. Matrix metal-

loprotease 1 (MMP-1) and MMP-3 mRNA were in-

creased by IL-1 in normoxia and hypoxia, whereas only

MMP-3 mRNA was enhanced in reoxygenated cultures.

The MMP-2 mRNA level was not significantly affectedby IL-1 in normoxia or hypoxia, whereas it was en-

hanced in reoxygenated cultures. MMP-9 mRNA was

dramatically decreased by IL-1 only in low oxygen

tension. Tissue inhibitor of metalloproteinases 1

(TIMP-1) message was significantly enhanced by the

cytokine in most instances, whereas TIMP-2 message

was markedly decreased by IL-1 in reoxygenated cul-

tures. Stimulation of TGF1 expression by IL-1 was

observed only in normal atmospheric conditions. One of

the more striking findings of the study was the greater

stimulating effect of IL-1 on NO production observed

in hypoxia, which was much higher than in normoxia,

whereas the reverse was observed for IL-1induced

PGE2

production.

Conclusion. Oxygen level and reoxygenation

stress significantly modulate gene expression and the

response of articular chondrocytes to cytokines such as

IL-1. In hypoxic conditions, which mimic the in vivo

condition of cartilage, the effects of IL-1 on both

synthesis and degradative processes are significantly

different from those in normoxia, conditions that are

unlikely encountered by chondrocytes in a normal state.

In low oxygen tension, high IL-1induced NO produc-

tion is associated with a significant decrease in PGE2

synthesis. These data should influence our concept ofthe role of oxygen in the pathophysiology of joint disease

and may help define the best conditions in which to

develop bioartificial cartilage.

Studies to date on the role of hypoxia in theskeleton have involved tissues or cells of mesenchymalorigin. Reports are conflicting as to whether chondro-

Supported by an unrestricted educational grant from UCBPharma (Belgium) and by the Regional Council of Lower Normandy.

1G. Martin, PhD, R. Andriamanalijaona, PhD, P. Bogdanow-icz, PhD, K. Boumediene, PhD, J.-P. Pujol, PhD: Laboratory ofConnective Tissue Biochemistry, Caen, France; 2S. Grassel, PhD:Orthopedic University Clinic, Regensburg, Germany; 3R. Dreier, PhD,P. Bruckner, PhD: Institute for Physiological Chemistry and Pathobio-chemistry, Munster, Germany; 4M. Mathy-Hartert, PhD, Y. Henrotin,PhD: University of Liege, Liege, France.

Address correspondence and reprint requests to J.-P. Pujol,PhD, Laboratory of Connective Tissue Biochemistry, Faculty ofMedicine, 14032 Caen Cedex, France. E-mail: [email protected].

Submitted for publication November 12, 2003; accepted inrevised form July 28, 2004.

3549


2/12

cytes in growth plate cartilage, which is normally avas-cular, are hypoxic (1,2). In the case of adult cartilage,however, there is evidence that the oxygen supply toresident chondrocytes is rather limited (3,4). Oxygen, as

well as other nutrients, must diffuse into the tissue from

the synovial fluid. Microelectrode studies have shownthat an oxygen gradient exists in cartilage, with thesuperficial layers and deep zones having tension of6%O2 and 1% O2, respectively (4). The possibility alsoexists that changes in oxygen concentration may occurduring joint diseases, such as rheumatoid arthritis (RA)and osteoarthritis (OA). For example, it has been sug-gested that in these diseases decreased capillary densityof the synovium may contribute to reduced oxygentension in the joint (5,6). In contrast, it is likely that inadvanced fibrillated stages of OA cartilage, a higheroxygen supply can reach chondrocytes in altered zonesof the tissue.

Articular cartilage is not the only tissue living inlow oxygen tension. For example, measurements of bonemarrow aspirates from human donors yielded mean PO2

values of 6.4% (7,8). In inflamed tissue, infected tissue,tumors, wounds, and fracture sites, the PO2 level couldbe considerably lower (9). In the case of chondrocytes, ithas been suggested that the low oxygen tension prevail-ing in cartilage imposes energy limitations and regulatesmatrix synthesis (10). Nevertheless, chondrocytes havebeen shown to be well adapted to these conditions andcapable of maintaining the energy charge of the cell,principally based on the glycolytic pathway (11).

Thus, conventional cell culture using atmosphericair exposes chondrocytes to oxygen tensions that aremuch higher than pathophysiologic levels (12). As aconsequence, biochemical changes associated with vari-ations in oxygen delivery are still poorly understood.Results of a few studies have indicated that articularchondrocytes cultured in low oxygen tension show down-regulation of type II collagen, no change in aggrecanmessenger RNA (mRNA) levels, and up-regulation oftissue inhibitor of metalloproteinases 1 (TIMP-1) (1316). The oxygen level may also modulate the chondro-cyte response to the cytokines implicated in joint dis-eases, including interleukin-1 (IL-1), tumor necrosis

factor (TNF), and transforming growth factor (TGF). In this regard, it has been reported that hypoxiadecreases IL-1 and TNF-induced nitric oxide (NO)production in porcine cartilage explants (17).

The pathogenesis of joint diseases is associatedwith cartilage breakdown due to the induction of severalmetalloproteases in both synoviocytes and chondrocytes,together with reduced synthesis of specific matrix com-

ponents including type II collagen and aggrecan. Thecytokines IL-1 and TNF are key factors in this meta-bolic dysfunction (18,19). They induce expression of theNO synthase 2 (NOS2) (2022) and cyclooxygenase 2(COX-2) (23,24) enzymes, resulting in increased pro-

duction of NO and prostaglandin E2(PGE2). It has beenshown that activation of NF-B and activator protein 1(AP-1) transcription factors by IL-1 is involved in theexpression of several proinflammatory genes, includingCOX-2 (25), NOS2 (2628), and matrix metallopro-teases (MMPs) (2931).

IL-1 is also capable of stimulating TGF1 ex-pression in articular chondrocytes by a mechanism thatis NO-dependent (Martin G, et al: unpublished obser-

vations). In turn, TGF1 can antagonize most of thedeleterious effects of IL-1and has been proposed as afactor promoting cartilage repair (32). It is likely that

this interplaying network of cytokines/growth factorscould be influenced by the oxygen tension present in themicroenvironment of chondrocytes. Furthermore, de-generation of the articular cartilage in OA causes fibril-lation in the tissue, and sometimes microfractures of thesubchondral bone may also occur. These processes prob-ably disrupt the gradient of oxygen in the cartilage andproduce reoxygenation of the chondrocytes. The pres-ence of higher oxygen tension may facilitate productionof reactive oxygen species, which are known to playsignaling roles (for review, see ref. 33) and could alsocontribute to the modulation of chondrocyte gene ex-pression.

In the present study, we examined the effect ofhypoxia (5% O2) and reoxygenation to atmospheric airon several functional parameters of articular chondro-cytes related to both anabolic and catabolic aspects oftheir activity. Cells were treated or not treated withIL-1to determine whether oxygen tension could mod-ulate the response of chondrocytes to this cytokine.Furthermore, the expression of TGFs and their recep-tors was studied as a critical aspect of cartilage repairpotential.

MATERIALS AND METHODS

Reagents. Mouse antiphospho-MAPK and rabbitantiMAPK-1/2 monoclonal antibodies were purchased from Up-state Biotechnology (Lake Placid, NY). P50, c-Rel, p65, c-Jun,c-Fos, and JunB polyclonal antibodies were obtained fromSanta Cruz Biotechnology (Santa Cruz, CA). Anti-rabbit oranti-mouse horseradish peroxidaselabeled secondary anti-bodies (Santa Cruz Biotechnology) were used for detectionwith an ECL Plus kit (Amersham Pharmacia Biotech, Orsay,France). The oligonucleotide probes were supplied by Invitro-

3550 MARTIN ET AL


3/12

gen (Cergy Pontoise, France) and were as follows: for HIF-1,5-TTGTGAGACGTGCGGCTTCCCA-3 ; fo r A P-1 , 5-CGCTTGATGAGTCAGCCCGGAA-3; for NF -B, 5-AGTTGAGGGGACTTTCCCAGGC-3. All other chemicalswere of the highest purity (Sigma-Aldrich, St. Quentin Falla-vier, France). Human recombinant IL-1 was a generous gift

from Dr. Soichiro Sato, Shizuoka, Japan.Culture and treatment of articular chondrocytes.

Chondrocytes were isolated from metacarpophalangeal jointsof freshly slaughtered calves by enzymatic treatment (12).Slices of cartilage were then digested in a special plasticchamber at low oxygen tension in order to prevent anyreoxygenation of the chondrocytes during the isolation process(see below). Sequential digestion with type XIV protease andtype I collagenase was performed, as previously described (12),in Dulbeccos modified Eagles medium (DMEM; Invitrogen)previously equilibrated at 5% O2. The cells were seeded inDMEM containing 10% heat-inactivated fetal calf serum(FCS; Invitrogen) and antibiotic/antimycotic (penicillin, eryth-romycin, and Fungizone). Three samples were cultured: two at5% O2 (hypoxia and reoxygenation) and one at 21% O2

(normoxia). A sealed chamber (model 818-GBB/22; BioblockScientific, Illkirch, France), equipped with gloves, was used. Asmall culture incubator was placed inside, and vessels wereintroduced by a special airlock. The chamber was previouslyequilibrated with 5% O2by flushing a gas mixture (N2 CO2)without oxygen (Bactal 2 Pro; Air Liquide, Mitry Mory,France). The oxygen tension was monitored with a specificprobe (model 74223; Bioblock Scientific). The oxygen concen-tration of the culture media was reduced from 20% to 5% O2by bubbling a gas mixture without oxygen (1.5 hours for a500-ml flask). The oxygen level was measured by inserting aprobe into the culture media (Oxi 330/set; Wissenschaftlich-Techniche Werkstatten, Weilheim, Germany). Phenol red,generally used as a pH indicator, was absent in the mediumbecause of its antioxidant properties.

To induce reoxygenation, some cell cultures weretransferred from hypoxic to atmospheric conditions and usedafter 30 minutes for electrophoretic mobility shift assay(EMSA) analysis and after 4 hours for Northern blottingexperiments. For the reoxygenation step, culture media werealso replaced with fresh media containing 21% O

2.Preparation of cytoplasmic and nuclear extracts.Cells

were rinsed twice with phosphate buffered saline (PBS) andscraped in a hypotonic buffer (34). After centrifugation at10,000gfor 10 minutes, the resulting supernatants were used ascytoplasmic extracts. The pellet was incubated in hypertonicbuffer for 4 hours at 4C and centrifuged at 10,000g for 30minutes to provide nuclear extracts (34). The protein amountwas determined by Bradfords colorimetric procedure (Bio-Rad SA, Ivry sur Seine, France).

Western blot analysis.Cellular lysates (15 g protein)were subjected to sodium dodecyl sulfatepolyacrylamide gelelectrophoresis (SDS-PAGE) under reducing conditions (35)and electrophoretically transferred to polyvinylidene difluo-ride membranes (NEN Life Science Products, Zawentem,Belgium). The membranes were then treated as previouslydescribed (36).

EMSA. Nuclear extracts (10 g protein) were incu-bated in binding buffer for 30 minutes at 25C with comple-mentary DNA (cDNA) probes, previously labeled with

[-32P]ATP (25 fmoles) using T4 polynucleotide kinase (LifeTechnologies). The final binding reaction for NF-B wasperformed in 20 mM HEPES (pH 7.5), 50 mM KCl, 4 mMMgCl2, 0.2 mM EDTA, 0.5 mMdithiothreitol (DTT), 0.05%Nonidet P40, 20% glycerol, 1 mg/ml bovine serum albumin,and 0.025 mMpoly(dI-dC). The binding reaction mixture for

HIF-1 and AP-1 was composed of 20 mMHEPES, 2 mMTrisHCl (pH 8), 1 mMDTT, 80 mMNaCl, 10% glycerol (volume/volume), 0.2 mM phenylmethylsulfonyl fluoride, 0.4 mMEDTA, and 0.2 mMEGTA. The samples were then submittedto 8% PAGE in 0.5 TBE (45 mMTris [pH 7.8], 45 mMboricacid, and 1 mM EDTA) and visualized by autoradiography(37).

Determination of TGF1 and TGF2 proteins byenzyme-linked immunosorbent assay (ELISA). TGF1 andTGF2 proteins were assayed in culture media, using anELISA kit (R&D Systems, Minneapolis, MN).

RNA extraction and Northern blot analysis. Prepara-tion of probes. Probes were obtained by reverse transcriptionpolymerase chain reaction (RT-PCR) of total RNA extractedfrom 5 106 chondrocytes, using the RNeasy Mini protocol

(Qiagen, Courtaboeuf, France) or the guanidium isothiocya-nate phenol chloroform procedure (38). Samples were treatedwith DNase during the RNA isolation procedure according tothe manufacturers protocol (Invitrogen). The RT reaction andPCR amplification were performed as previously described(39). Cycling parameters included an initial denaturation stepfor 5 minutes at 95C, followed by 35 or 40 cycles of 30 secondsat 95C, 1 minute at various annealing temperatures accordingto the temperature of the primers (55C for AC; 60C for D,E, H, I, and J; 56.5C for F and G), and 1 minute at 72C.

For quantification of MMP and TIMP mRNAs, PCRproducts were purified using a PCR purification kit (Qiagen),subcloned into the pCR II-TOPO vector (Invitrogen) andsubsequently sequenced for identification (ABI Prism geneticanalyzer, model 310; Amersham Pharmacia Biotech). The

GAPDH probe was a gift from Dr. H. Kresses group (Mun-ster, Germany).Primer sequences. The primer sequences were as fol-

l ow s: f or t yp e I I c ol la ge n ( 64 8 b p) , s en se 5-GAC-CCCATGCAGTACATG-3, antisense 5-GACCGTCTTGC-CCCACTT-3; for aggrecan core protein (200 bp), sense5-CCCTGGACTTTGACAGGGC-3, antisense 5-AG-GAAACTCGTCCTTGTCTCC-3; for 18S RNA (100 bp),sense 5-CTTTTCAGAGGGACAAGTGG-3, antisense 5-CCTACGGAAACCTTGTTACG-3; for MMP-1 (442 bp),sense 5-GGGAAATCCTGTTGGGAGAACA-3, antisense5-GATGGCCTGGATCCCATCAA-3; for MMP-2 (334 bp),sense 5-CGCTCGTGCCTTCCAAGTCT-3 , antisense 5-TGGTGGAACACCAAAGGAAGC-3; for MMP-3 (429 bp),sense 5-TGGCTCATGCCTACCCACCT-3 , antisense 5-AAGAGATCAAATGAAATTCAGGTTC-3; for membranetype 1 (MT1)MMP (625 bp), sense 5-TCAAGGCCA-ATGTTCGAAGGA-3, antisense 5-GGCCACGGTGTCA-AA GTTCC-3; f or T IM P- 1 ( 46 9 b p) , s en se 5-GC-CTCTGGCATCCTGTTGCT-3, antisense 5-GGTCCGTC-CACAAGCAGTGA-3; for TIMP-2 (500 bp), sense 5-GATCAGGGCCAAAGCAGTCAA-3, antisense 5- CGAT-GTCCAGAAACTCCTGCTTG-3; for MMP-9 (495 bp),sense 5- GAACCACGAACCAACCTCAC-3, antisense 5-CATCTCCCTGAATGCCTTG-3.

RESPONSE OF CHONDROCYTES TO IL-1 AND DEPENDENCE ON OXYGEN TENSION 3551


4/12

Labeling of the probes.Type II collagen, aggrecan, and18S probes were 32P-radiolabeled, using a random priming kit(Invitrogen). The other probes were labeled with the Prime-A-Gene kit (Promega, Charbonnieres, France) according tothe manufacturers protocol.

Northern blotting. Type II collagen and aggrecan

Northern blots were prehybridized (1 hour at 42C) andhybridized (18 hours at 4860C) in Na2HPO4 (pH 7.27.4)/SDS 20% (2:1 [v/v]). The blots were washed several times in2 standard saline citrate (SSC) and 0.1% SDS at 42C. Finalwashes were in 0.1 SSC plus 0.1% SDS at 42C. The32P-radiolabeled cDNAmRNA hybrids were visualized byautoradiography. The signals were then measured by densito-metric scans of the x-ray film (Kodak X-Omat; Kodak, Roch-ester, NY) and quantified by the ImageQuant program (Mo-lecular Dynamics, Sunnyvale, CA).

For the other blots, filters were hybridized eitherovernight (MMPs and TIMPs) or for 1 hour (GAPDH) at60C, after prehybridization with 5 ml of QuikHyb hybridiza-tion solution (Stratagene, Europe, Amsterdam, The Nether-lands) for 1 hour at 60C. Filters were subsequently washed,

once or twice each, in washing buffer containing 2 SSC plus0.1% SDS, 1 SSC plus 0.1% SDS, 0.2 SSC plus 0.1% SDS,and 0.1 SSC plus 0.1% SDS for each, 10 minutes at 60C.Filters were exposed to Storm phosphoimager screens (Amer-sham Pharmacia Biotech) either overnight (MMPs andTIMPs) or for 45 hours (GAPDH), scanned with a Stormphosphoimager (Amersham Pharmacia Biotech), and sub-jected to a densitometric evaluation with ImageQuant analysissoftware (Amersham Pharmacia Biotech).

Filters were repeatedly hybridized with up to 5 differ-ent probes and between hybridizations, filters were strippedwith 10 mMEDTA, 10 mMTris, and 0.1% SDS for 10 minutesat 8595C. Subsequently, filters were equilibrated twice in 2SSC at room temperature.

Semiquantitative RT-PCR.Three micrograms of RNA

was transcribed in a 50-l volume containing 10 l of 5

FirstStrand buffer, 20 M oligo(dT)16, 1 l of recombinant ribo-nuclease inhibitor (RNAseOUT; Invitrogen) (40 units/l), 3 lof dNTPs (10 mMeach), and 1 l of Moloney murine leukemiavirus (200 units/l) for 15 minutes at 42C and 5 minutes at99C. PCR was performed with 3 l of cDNA in 40l of PCRreaction sample with 0.2 l ofTaq polymerase (15 units/l), 1l of dNTPs (10 mMeach), 1 l of 50 mMMgCl2and 20 pMof each primer. Amplification of the cDNA sequence wasperformed by PCR, using the following primer sequences:for TGF1, sense GCCCTGGACACCAACTATTGC, anti-sense GCTGCACTTGCAGGAGCGCAC; for TGF2, senseGCTTTGGATGCGGCCTATTGC, antisense GCTGCATT-TGCAAGACTTTAC; for TGF receptor type I (TGFRI),sense ATTGCTGGACCAGTGTGCTTCGTC, antisenseTAAGTCTGCAATACAGCAAGTTCCATTCTT; forTGFRII, sense CGCTTTGCTGAGGTCTATAAGGCC, an-tisense GATATTGGAGCTCTTGAGGTCCCT. The follow-ing amplification protocol was used: 1 minute at 95C, 1 minuteat 55C, and 1 minute at 72C. PCR products were analyzed in2% agarose gels stained with ethidium bromide. The banddensities were quantified using ImageQuant. The relativeexpression amount of messengers was calculated as the ratio tothe expression of 18S mRNA, as a housekeeping gene refer-ence. The procedure was repeated in 3 different experiments.

Measurement of NO and PGE2 production. NO pro-duction was assessed by measuring the concentration ofnitrate/nitrite in media, using the Griess reaction (40). PGE2was assayed in the culture media according to a previouslydescribed radioimmunoassay (41).

Gelatin zymography. Fifty-microliter aliquots of me-

dium were mixed with equal volumes of 2-fold concentratedsample loading buffer (2 mMEDTA, 2% SDS, 20% glycerol,0.02% bromphenol blue, 20 mM Tris HCl [pH 8.0]) andsubjected to electrophoresis on a 1% gelatincontaining 4.515% gradient SDS polyacrylamide gel. The gels were washedtwice for 30 minutes in 2.5% Triton X-100, rinsed in distilledwater, and developed for 16 hours at 37C in 50 mMTris HCl(pH 8.5) containing 5 mM CaCl2. Finally, they were stainedwith Coomassie brilliant blue R250 (Serva, Heidelberg, Ger-many) to visualize protease activity and photographed.

Statistical analysis. Three different experiments wereperformed. Data are presented as the mean SEM.Pvalueswere calculated by Students t-test.

RESULTS

Strong HIF-1 DNA binding induced by hypoxia

and reoxygenation.We first examined the binding activ-ity of the redox-sensitive transcription factor HIF-1,

which is known to regulate the expression of glycolyticenzymes in most cell types. In normoxia, HIF-1 bindingactivity was very low or undetectable, but it stronglyincreased in hypoxia (Figure 1A). When cells werereoxygenated after culturing in low oxygen tension,HIF-1 binding was maintained at the same level. Expo-sure to IL-1 did not induce a significant change inHIF-1 binding activity in both situations. Competition

analysis, using a wild-type probe, revealed the specificityof this complex (Figure 1A).Enhancement of IL-1induced NF-B and AP-1

DNA binding by hypoxia and reoxygenation. To investi-gate the influence of normoxia, hypoxia, and reoxygen-ation on the response of articular chondrocytes, weperformed a series of EMSA analyses, using consensusDNA binding sequences for NF-B and AP-1 transcrip-tion factors. Nuclear extracts were obtained from BACscultured in either normoxia (21% O2) or hypoxia (5%O2) and treated with IL-1 (10 ng/ml) for several timeperiods. Part of the cells cultured in hypoxia weresubmitted to a 30-minute period of reoxygenation to

produce an oxidative stress. As expected, the bindingactivities of NF-B and AP-1 complexes were stronglyinduced by the exposure of chondrocytes to IL-1 innormoxia (Figures 1B and C). One-hour treatment withIL-1was found to be sufficient to produce the bindingof 3 NF-B complexes (Figure 1B), whereas 6-hourtreatment was required to clearly stimulate the bindingof AP-1 (Figure 2A). IL-1induced binding of NF-B

3552 MARTIN ET AL


5/12

and AP-1 was greater in hypoxia compared with nor-moxia (Figures 1C and 2B). Moreover, hypoxic condi-tions were found to increase the basal binding activity of

AP-1. Interestingly, it must be noted that an additionalNF-B protein complex, with lower molecular weight,

appeared in IL-1treated hypoxic cultures comparedwith those treated in normoxic conditions (Figure 1C).

Competition analysis, using wild-type or mutatedprobes, revealed the specificity of these complexes (Fig-

Figure 1. Hypoxia-inducible factor 1 (HIF-1) and NF-B DNA bind-

ing activity under interleukin-1 (IL-1) treatment: effect of oxygen

tension and reoxygenation. Bovine articular chondrocytes were cul-tured in normoxia (21% O2) or hypoxia (5% O2) and were treated or

not treated with IL-1(10 ng/ml) for several time periods (1, 2, 6, and

24 hours). Part of the cells cultured in 5% O2 were transferred fromhypoxic to atmospheric conditions in order to induce an oxidative

stress and used after 30 minutes. At this stage, nuclear extracts were

prepared and used in electrophoretic mobility shift assay. The bindingof HIF-1 was studied using a specific cDNA probe. A, The binding

specificity of complexes induced by hypoxia was assessed with unla-

beled HIF-1 wild-type (Wt) duplexes as competitors. B and C, Thebinding of NF-B was studied using a specific cDNA probe. D, The

binding specificity of complexes induced by IL-1 was characterized

using unlabeled NF-B wild-type or mutant (Mt) duplexes as compet-

itors. E, Inducible complexes were identified using antic-Rel, anti-p50, and anti-p65 antibodies. C control; IC inducible complex;

SS supershift. a, b, and c denote binding complexes. Results shown

are representative of 3 independent experiments.

Figure 2. Effect of oxygen tension and reoxygenation on activator

protein 1 (AP-1) DNA binding and ERK1/2 MAPK activity underinterleukin-1 (IL-1) treatment. Bovine articular chondrocytes

(BACs) were cultured as described in Figure 1, and nuclear extracts

were used in electrophoretic mobility shift assay (A and B). Thebinding specificity of complexes induced by IL-1 was characterized

using unlabeled NF-B duplexes, wild-type (Wt), or mutant (Mt)

duplexes as competitors (C). Finally, inducible complexes were iden-tified using antic-fos and anti-JunB antibodies (D). BACs were

cultured either in normoxia (21% O2) or hypoxia (5% O2) and treated

with IL-1 (10 ng/ml) for 30 minutes. Part of the cells cultured in

hypoxia were transferred to a normoxic environment for 15 minutesbefore the end of IL-1 treatment in order to generate an oxidative stress.

Cytoplasmic extracts were prepared and used for Western blot analysis,

using antiphospho-ERK1/2 (P-Erk1 and P-Erk2, respectively) (E). Theproteinantibody complexes were visualized by chemiluminescence, using

horseradish peroxidaseconjugated secondary antibody. Results shown

are representative of 3 independent experiments. C control; IC inducible complex; SS supershift; NIC noninducible complex. d and

e denote binding complexes.



6/12

ures 1D and 2C). In an attempt to identify the NF-Bcomponents induced by IL-1, we performed supershiftassays with specific antibodies to NF-B subunits. Antibod-ies to p65 and to p50 subunits caused a disappearance ofthe complex, suggesting that IL-1induced the binding of

a p65 homodimer and a heterodimer that could be formedby association of p65 and p50 subunits (Figure 1E). Asimilar approach demonstrated that IL-1 specifically ac-tivated the c-Fos and JunB subunits of AP-1 (Figure 2D).Reoxygenation of hypoxic cultures was found to increasethe IL-1induced binding of NF-B and AP-1, ascompared with cultures maintained in hypoxia.

Greater IL-1 stimulation of MAPKs in low

oxygen tension. To further study the effect of oxygentension on the signaling pathways upstream of NF-Band AP-1 DNA binding, we performed Western blotanalysis of MAPKs. A 30-minute treatment with IL-1induced phosphorylation of ERK-1 and ERK-2 in thechondrocytes, reflecting activation of the kinases (Figure2E). In addition, the bands were always of greaterintensity for chondrocytes cultured in low oxygen ten-sion, indicating that IL-1 stimulation was more effec-tive in hypoxia than in normoxia. In contrast, reoxygen-ation of hypoxic cultures caused a decrease of IL-1effect, as compared with hypoxic controls.

Modulation of IL-1 effect on aggrecan core

protein and type II collagen mRNA steady-state levels

by oxygen tension. Northern blot analysis revealed thatthe mRNA levels of both aggrecan core protein and typeII collagen were lower in hypoxic and reoxygenation

conditions, as compared with normoxia (Figure 3A andB). However, the decrease was not statistically signifi-cant between normoxia and hypoxia. As expected, IL-1treatment (24 hours) caused a reduction of these mRNAamounts in the 3 experimental conditions. This down-regulation was more marked for collagen in hypoxia andreoxygenation conditions than in normoxia (Figure 3B).

Influence of oxygen tension on IL-1 regulation

of MMP expression. As studied by Northern blotting,the response of MMP genes to oxygen tension and IL-1

was dependent on the MMP t ype. As expected, MMP-1and MMP-3 mRNAs were increased by treatment ofchondrocytes with IL-1 in both normoxic and hypoxic

conditions, although the difference was not statisticallysignificant (Figures 4A and C). However, in reoxygen-ated cultures, only MMP-3 expression was enhanced byIL-1, with no change of the MMP-1 mRNA levelobserved in that situation. MMP-2 and MMP-9 genes

were found to respond in a manner different from thatof MMP-1 and MMP-3. In normoxic cultures of chon-drocytes, MMP-2 and MMP-9 mRNA levels were notsignificantly affected by IL-1 (Figures 4B and D). In

Figure 3. Modulation of the effect of interleukin-1 (IL-1) on

aggrecan core protein and type II collagen expression by oxygentension and oxidative stress. Bovine articular chondrocytes were

cultured as described in Figure 1 and treated with IL-1(10 ng/ml) for

24 hours. Part of the cells cultured in 5% O2were submitted to 4-hourreoxygenation. Total RNA was isolated and used in Northern blotting

analysis, with 2 labeled probes corresponding to aggrecan core protein

(A)and COL2A1 (B) cDNA. Hybridization with a probe correspond-ing to 18S RNA is shown as an internal control of RNA loading.

Histograms show the mean and SEM values obtained by densitometric

scanning of membranes prepared with RNA from 3 different experi-ments. The normoxic control (C) was arbitrarily set at 100%. NS not

significant. P 0.05; P 0.01.

3554 MARTIN ET AL


7/12

low oxygen tension, MMP-2 mRNA was reduced by halfas compared with normoxic conditions, but IL-1 didnot affect this level (Figure 4B). In contrast, MMP-9mRNA was significantly decreased by IL-1 in hypoxia(Figure 4D). Finally, in reoxygenated cultures, theMMP-2 mRNA level was enhanced by IL-1, whileMMP-9 mRNA was not affected. The expression of amembrane-type metalloprotease, MT1-MMP, was alsoinvestigated. First, we observed that the mRNA steady-

state level of this gene was dramatically increased inhypoxia and in reoxygenated cultures, as compared withnormal atmosphere (Figure 4E). IL-1 treatmentcaused an increase of MT1-MMP mRNA in normoxiaand reoxygenation conditions, whereas no effect was

observed in the setting of low oxygen tension.In order to obtain more information on the

enzymatic activity produced by chondrocytes in theseexperimental conditions, we also performed gelatin zy-mography analysis, using the conditioned culture media.We observed a band corresponding to a protein with anapparent molecular mass 66 kd in all of the samples,regardless of the treatment. The enzyme presumablycausing this band is proMMP-2, because it comigrated

with a gelatinolytic protein in media of human OAchondrocytes known to contain latent and activeMMP-2. Supporting this hypothesis, an immunoblottingexperiment with an antibody to MMP-2 (catalog no.IM33T; Oncogene Research Products, Cambridge, MA)revealed a band 66 kd under reduced conditions(results not shown). Culture at 5% O2led to productionof an additional gelatinolytic molecule with an electro-phoretic mobility slightly less than 97 kd. This proteo-lytic activity was not detectable in conditioned media ofchondrocytes cultured at 21% O2or reoxygenated aftera hypoxic period. It was not related to proMMP-9 oractive MMP-9, because no reactivity was detected inimmunoblots with an antibody to MMP-9 (Biomol,Plymouth Meeting, PA; catalog no. SA-106) (data notshown).

Effect of oxygen tension and IL-1 on TIMP-1and TIMP-2 expression. Because TIMPs are key ele-ments in the regulation of MMP activity in articularcartilage, we wanted to determine whether oxygen ten-sion could modulate this expression. Northern blotanalysis revealed that the variations of TIMP-1 andTIMP-2 mRNA in our experimental conditions wereclearly different. The amount of TIMP-1 mRNA was notsignificantly altered by oxygen tension (Figure 4G),

whereas it was increased by IL-1, especially in reoxy-genated cultures (P 0.01). In contrast, the basal levelof TIMP-2 mRNA was significantly augmented in reoxy-genated cultures and dramatically reduced by IL-1 in

these conditions (Figure 4H).Effect of oxygen tension and IL-1 on the expres-

sion of TGF1, TGF2, and TGFRs. TGFs areimplicated in cartilage homeostasis, exerting oppositeeffects against IL-1 on extracellular matrix metabolism.Therefore, it was interesting to study the expression ofTGF1 and TGF2, together with that of TGFRI andTGFRII, in our system. Real-time PCR analysis wasperformed on total RNA, and ELISA was carried out to

Figure 4. Effect of hypoxia and reoxygenation on matrix metallopro-

tease (MMP) and tissue inhibitor of metalloproteinases (TIMP)

expression by chondrocytes treated or not treated with interleukin-1

(IL-1). Total RNA from chondrocyte cultures treated as described in

Figure 1 was used in Northern blot analysis to quantify the mRNA

steady-state levels of MMP-1 (A), MMP-2 (B), MMP-3 (C), MMP-9

(D), membrane type 1 (MT1)MMP (E), TIMP-1 (G), and TIMP-2

(H) genes. Data are the mean and SEM of 3 experiments. Thegelatinolytic activities produced by chondrocyte cultures was analyzedby zymography (F). Culture media from cultures treated as described

in Figure 1 were subjected to electrophoresis on a 1% gelatin

containing 4.515% gradient sodium dodecyl sulfatepolyacrylamidegel (see Materials and Methods). Protease activity was visualized with

Coomassie brilliant blue R250 staining. The apparent lower amount in

reoxygenation samples is attributable to the fact that the culture

medium was washed out and replaced with fresh fully oxygenatedmedium to ascertain rapid reoxygenation. As a consequence, the

values are likely to represent production over the last 4 hours. C control. P 0.05; P 0.01. ??? unidentified molecule.



8/12

assay TGF1 and TGF2 proteins in culture media. Inlow oxygen tension and after reoxygenation, the chon-drocytes expressed more TGF1 mRNA and proteinthan in normal atmosphere (Figures 5A and C). Thisexpression was significantly increased by IL-1, at both

the mRNA and protein levels (P 0.01) in normoxia,whereas it was rather reduced in low oxygen tension andreoxygenation conditions. The production of TGF2protein was found to be much lower in hypoxic condi-tions and in reoxygenated cultures than in normoxia

(Figure 5D). However, there was no correlation with themRNA steady-state levels (Figure 5B), in contrast withTGF1 mRNA. In normoxic conditions, IL-1 treat-ment induced a decrease in TGF2 protein (P 0.001),

which was not reflected at the mRNA level (Figures 5Band D). The cytokine did not affect the protein level inthe hypoxia and reoxygenation protocols, whereas themRNA levels were decreased significantly in hypoxia(Figures 5B and D).

The mRNA level of TGFRI was greater in

chondrocytes cultured in low oxygen or reoxygenatedafter a hypoxic period than in normal atmosphere (Fig-ure 5E). IL-1 did not affect this level in normoxia andhypoxia but caused a significant decrease in reoxygen-ated cultures. Contrasting with the TGFRI mRNAlevel, the TGFRII mRNA level did not significantlychange, regardless of the conditions (Figure 5F). How-ever, IL-1 reduced by 2-fold the value of TGFRII

Figure 5. Effect of hypoxia and reoxygenation on transforminggrowth factor 1 (TGF1), TGF2, and TGF receptor (TR)

mRNA expression by chondrocytes treated or not treated with

interleukin-1(IL-1).A, B, E,and F, Total RNA from chondrocytes

treated as described in Figure 1 was used in reverse transcriptionpolymerase chain reaction analysis to quantify the mRNA steady-state

levels of TGF1, TGF2, TRI, and TRII genes (see Materials and

Methods). C and D, Media were collected to assay the levels ofcorresponding TGF1 and TGF2 proteins, using an enzyme-linked

immunosorbent assay kit. Data are the mean and SEM of 3 experi-

ments. The statistical significance of the differences between control(C) and IL-1treated samples was evaluated by Studentst-test. P 0.05; P 0.01; P 0.001.

Figure 6. Production of nitric oxide (NO) and prostaglandin E2(PGE2) by normoxic, hypoxic, and reoxygenated chondrocyte cultures.Media from cultures treated as described in Figure 1 were collected

and used to determine NO production by the Griess reaction(A) and

PGE2 levels by radioimmunoassay (B). The statistical significance of

differences between control (C) and interleukin-1 (IL-1)treatedcultures was evaluated by Students t-test. Data are the mean and

SEM. P 0.05; P 0.01; P 0.001.

3556 MARTIN ET AL


9/12

mRNA in normoxia (P 0.01) and by one-third thevalue in low oxygen tension (P 0.05).

Differential effect of oxygen tension on IL-1

induced NO and PGE2

production. IL-1induced NOproduction by chondrocytes was significantly greater

(8-fold) in hypoxia, compared with that in normal atmo-sphere (Figure 6A). Similar values were also observedfor cultures that were reoxygenated after the low-oxygenperiod. Interestingly, the production of PGE2displays anopposite profile, being 3-fold lower in the setting of lowoxygen and after reoxygenation of cells (Figure 6B). Thedata clearly indicate that inducible NOS (iNOS) andCOX are differently regulated by oxygen tension.

DISCUSSION

Chondrocytes can survive in oxygen tension aslow as 0.1% (14), indicating that these cells are welladapted to hypoxia (11). As a consequence, they displayan unusual response to changes in oxygen tension. Forexample, in contrast to several cell types that increasetheir glycolytic enzyme expression and their anaerobicrespiration (42), articular chondrocytes produce lesslactic acid in low oxygen tension and are less metaboli-cally active than in normoxia (11). Despite this existingknowledge, the response of articular chondrocytes tocytokines and growth factors in hypoxic conditions is stillpoorly understood.

Here, we found significantly greater HIF-1 DNAbinding in hypoxic and reoxygenated chondrocyte cul-

tures, compared with normoxic controls. These findingsare consistent with those of a recent study showing thattargeted deletion of HIF-1 in growth-plate chondrocytesrenders them unable to maintain ATP levels and inducesreduced synthesis of aggrecan and t ype II collagenmRNA in a hypoxic microenvironment (43). Takentogether, these results clearly indicate the critical role ofHIF-1 in the maintenance of anaerobic glycolysis andextracellular matrix synthesis of chondrocytes submittedto low oxygen tension.

Interestingly, our results revealed that the IL-1induced DNA binding of NF-B and AP-1 was signifi-cantly enhanced in hypoxic and reoxygenated chondro-

cyte cultures. Furthermore, the NF-B complexesformed in conditions of low oxygen tension and duringreoxygenation appeared slightly different from thoseobserved in normoxic control cultures. The intensity ofthe chondrocyte response to IL-1 is therefore greaterin an environment of low oxygen tension, such as thatfound in living cartilage, suggesting that previous dataobtained in normoxic conditions may have been under-

estimated. This can be extended to MAPKs, which werealso activated by IL-1 to a greater extent in 5% O2concentration, a finding that confirms results of a pre-

vious study on human cartilage explants (44). In ourstudy, we confirmed that type II collagen and aggrecan

core protein mRNA levels were reduced in hypoxia(14,15), and we found that the IL-1inhibitory effect ontype II collagen expression was somewhat greater inhypoxia than in normoxia. The IL-1induced NF-Bactivation can provide an explanation for the down-regulation of type II collagen expression, because it hasrecently been demonstrated that inhibition of COL2A1transcription by IL-1 was mediated by the 57/125-bp promoter region and was associated with in-duction of Egr-1, NF-B, and ESE-1 binding activities innuclear extracts (45).

A link can be established between IL-1 stimula-tion of MAPKs, AP-1, and NF-B transcription factors,and the expression of several genes bearing AP-1 andNF-Bresponsive elements in their promoters, includ-ing MMPs, iNOS, and COX. For instance, it has previ-ously been shown that IL-1 activation of MAPKs, NF-B, and AP-1 was associated with up-regulation of someMMP gene expression, including MMP-3 and MMP-13,in BACs (46). The promoters of several MMPs areknown to contain AP-1responsive elements, and therole of AP-1 in their transcription has been largelydocumented (47,48). Implication of NF-B, whose sitesare not systematically found in all MMP promoters, hasalso been reported. For example, NF-B activation is

involved in the regulation of MMP-1 and MMP-13 genesin chondrosarcoma cells (49). In the present study, nosystematic relationship between IL-1induced activa-tion of AP-1 and NF-B and the transcriptional activityof the MMP genes could be found, because the responseto cytokine and oxygen tension was largely dependent onthe MMP type.

Collagenase 1 (MMP-1) and stromelysin(MMP-3) are known to be expressed in a generallycoordinated manner. Their promoters contain commonregulatory sequences. In the present study, MMP-1 andMMP-3 mRNA levels were similarly increased by IL-1in hypoxia and normoxia. However, we did not observe

greater stimulation of these genes, despite the higherAP-1 and NF-B binding in low oxygen tension. It islikely that other factors may contribute to regulateMMP-1 and MMP-3 gene transcription in hypoxic con-ditions. It is tempting to speculate that production ofreactive oxygen species, which is likely to be greater innormoxia than in hypoxia, may synergize with IL-1,

whereas the higher level of NO elicited by IL-1 in



10/12

hypoxia could reduce the cytokine effect on MMP-1 andMMP-3 (33). Interestingly, we observed that the MMP-3mRNA level was significantly increased by IL-1 inreoxygenated cultures, while the amount of MMP-1 wasnot altered, indicating that there may be a differential

regulation of these 2 MMP genes during oxidative stress.Gelatinase A (MMP-2) and gelatinase B

(MMP-9) have the same substrates, but they are con-trolled by different mechanisms in most cell types.

Analysis of the 5-flanking region of the MMP-2 genehas revealed striking differences with several MMPpromoters, including the absence of TATA box andTRE sequence. Furthermore, the presence of a bindingsite for AP-2, several Sp1-binding elements, and 2silencer sequences in the MMP-2 promoter suggests thatMMP-2 could be regulated by tissue-specific repressorsrather than activators. In the present study, these 2MMPs were not stimulated by IL-1 in either normoxicor hypoxic cultures, in contrast to MMP-1 and MMP-3.Moreover, MMP-9 mRNA levels were strongly de-creased by IL-1 in hypoxia. Interestingly, IL-1 wascapable of stimulating MMP-2 expression during theoxidative stress induced by reoxygenation, suggestingthat the reactive oxygen species produced in theseconditions may up-regulate this MMP.

MT1-MMP is a membrane-associated MMPfound in human cartilage (50). It possesses properties ofcollagenase and is able to activate the gelatinase MMP-2and collagenase 3 (MMP-13) (51). Here, baseline ex-pression of MT1-MMP mRNA was significantly in-

creased in chondrocytes cultured in low oxygen tension(12-fold) or reoxygenated (10-fold), compared withnormoxic controls, suggesting that this gene could beHIF-1 dependent. Interestingly, MT1-MMP mRNA lev-els were further increased by exposure to IL-1 inreoxygenated cultures.

This study is the first to show that a gelatinolyticactivity, which remains to be characterized, is expressedby chondrocytes in low oxygen tension. This activity wasnot apparently regulated by IL-1. Here, too, it is tempt-ing to speculate that the expression of this enzyme couldbe linked to HIF-1 activation, as it is for MT1-MMP.TIMPs can control the biologic activity of MMP.

TIMP-1 can bind activated forms of MMP-1 and MMP-3and both latent and active MMP-9. In contrast, TIMP-2is mainly associated with MMP-2. In agreement with thisapparent functional difference, the promoters of theseTIMP genes show variation in their regulatory elements:they both contain several Sp1 sites and 1 TRE sequence,but the TRE site of TIMP-2 is part of an inhibitoryelement. Thus, IL-1, TNF, and TPA generally up-

regulate TIMP-1, as well as MMP-1 and MMP-3,whereas TIMP-2 and MMP-2 are weakly influenced bythese factors. This was the case in our study, becauseTIMP-1 mRNA levels were increased by IL-1 in nor-moxic, hypoxic, and reoxygenated cultures, while those

of TIMP-2 were not significantly altered in normoxiaand hypoxia and were IL-1inhibited during oxidativestress.

We previously observed that TGF1 expressionby chondrocytes was enhanced by IL-1 (34), and thatblocking NO production led to further stimulation bythe cytokine (Martin G, et al: unpublished observations).This indicates that NO exerts an inhibitory effect onTGF1 transcription. Here, we confirmed the IL-1up-regulation of TGF1 expression in normoxia, but wefound an inhibition by the cytokine in conditions of lowoxygen tension and during oxidative stress. This findingis probably attributable to the higher level of NOproduced by IL-1action in these conditions, leading toan overall negative effect of the cytokine. The TGF2gene promoter does not contain the same regulatoryelements as TGF1 and, therefore, is not regulated inthe same way by IL-1 (52). TGF2 protein was reducedby IL-1 only in normoxia, whereas no effect wasobserved in hypoxic and reoxygenated cultures.

TGFRII plays a crucial role in the homeostasisof cartilage (53,54). Our present data clearly show thatIL-1 down-regulates the expression of TGFRII inboth normoxia and hypoxia, suggesting that the cytokinecould be responsible for the dramatic decrease in this

receptor in OA cartilage (53). The enhanced expressionof TGFRI in hypoxia may potentially alter the ratiobetween the 2 receptors and also contribute to modula-tion of the TGF effect.

Finally, our findings indicate that oxygen tensioncan modulate the relative production of NO and PGE2induced by IL-1 in chondrocytes. NO production wassignificantly increased in hypoxic and reoxygenated cul-tures compared with normoxic controls, whereas thesituation for PGE2synthesis was exactly the opposite. Asan explanation for our results, it is likely that high levelsof NO may exert an inhibitory effect on COX expressionand activity, as previously reported (55,56). Our data are

not in agreement with those of a recent study showingthat porcine cartilage explants in hypoxia produced lessNO under the effect of IL-1 than in normoxia (17).However, the culture system used by those authors hadtechnical limitations and did not allow the experimentsto be performed in conditions of continuous low oxygentension, because culture media could not be replenishedin hypoxic conditions.

3558 MARTIN ET AL


11/12


12/12

species in homeostasis and degradation of cartilage. OsteoarthritisCartilage 2003;11:74755.

34. Andriamanalijaona R, Felisaz N, Kim SJ, King-Jones K, LehmannM, Pujol JP, et al. Mediation of interleukin-1induced transform-ing growth factor 1 expression by activator protein 4 transcriptionfactor in primary cultures of bovine articular chondrocytes: possi-ble cooperation with activator protein 1. Arthritis Rheum 2003;48:156981.

35. Laemmli UK. Cleavage of structural proteins during the assemblyof the head of bacteriophage T4. Nature 1970;227:6805.

36. Legendre F, Dudhia J, Pujol JP, Bogdanowicz P. JAK/STAT butnot ERK1/ERK2 pathway mediates interleukin (IL)-6/solubleIL-6R down-regulation of type II collagen, aggrecan core, and linkprotein transcription in articular chondrocytes: association with adown-regulation of SOX 9 expression. J Biol Chem 2003;278:290312.

37. Bellido T, Borba VZ, Roberson P, Manolagas SC. Activation ofthe Janus kinase/STAT (signal transducers and activator of tran-scription) signal transduction pathway by interleukin-6-type cyto-kines promotes osteoblast differentiation. Endocrinology 1997;138:366676.

38. Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidium thiocyanate-phenol-chloroform extraction. AnalBiochem 1987;162:1569.

39. Chadjichristos C, Ghayor C, Herrouin JF, Ala Kokko L, Suske G,Pujol JP, et al. Down-regulation of human type II collagen geneexpression by transforming growth factor-1 (TGF-1) in articularchondrocytes involves SP1/SP3 ratio. J Biol Chem 2002;277:4390317.

40. Green L, Wagner D, Glogowski J, Skipper PL, Wishnok JS,Tannenbaum SR. Analysis of nitrate, nitrite and [15N]nitrate inbiological fluids. Anal Biochem 1982;126:1318.

41. Serteyn D, Deby-Dupont G, Pincemail J, Mottart E, Philippart C,Lamy M. Equine postanaesthetic myositis: thromboxanes, prosta-cyclin and prostaglandin E2 production. Vet Res Commun 1988;12:21926.

42. Graven KK, McDonald RJ, Farber HW. Hypoxic regulation ofendothelial glyceraldehyde-3-phosphate dehydrogenase. Am JPhysiol 1998;274:C34755.

43. Pfander D, Cramer T, Swoboda B, Johnson RS. HIF-1 plays acritical role in growth-plate chondrocytes during low oxygentension. Osteoarthritis Cartilage 2003;11 Suppl A:S5S6.

44. Dudhia J, Wheeler-Jones C, Bayliss MT. Transient activation ofp42/p44 MAPK by IL-1 is enhanced by a low oxygen tension inhuman articular cartilage. Proceedings of the 46th Annual Meet-ing, Orthopaedic Research Society; 2000 Mar 12-15; Orlando, FL.

45. Goldring MB, Tan L, Choy BK, Osaki M. Multiple interactingIL-1-induced transcription factors converge to suppress constitu-tive COL2A1 promoter activity [abstract]. Osteoarthritis Cartilage2002;10 Suppl A:S12.

46. Liacini A, Sylvester J, Li WQ, Zafarullah M. Inhibition of inter-leukin-1-stimulated MAP kinases, activating protein-1 (AP-1) andnuclear factor B (NF-B) transcription factors down-regulatesmatrix metalloproteinase gene expression in articular chondro-cytes. Matrix Biol 2002;21:25162.

47. Buttice G, Quinones S, Kurkinen M. The AP-1 site is required for

basal expression but is not necessary for TPA-response of thehuman stromelysin gene. Nucleic Acids Res 1991;19:372331.48. Pendas AM, Balbin M, Llano E, Jimenez MG, Lopez-Otin C.

Structural analysis and promoter characterization of the humancollagenase-3 gene (MMP13). Genomics 1997;40:22233.

49. Mengshol JA, Vincenti MP, Coon CI, Barchowsky A, BrinckerhoffCE. Interleukin-1 induction of collagenase 3 (matrix metallopro-teinase 13) gene expression in chondrocytes requires p38, c-JunN-terminal kinase, and nuclear factor B: differential regulation ofcollagenase 1 and collagenase 3. Arthritis Rheum 2000;43:80111.

50. Buttner FH, Chubinskaya S, Margerie D, Huch K, Flechtenma-cher J, Cole AA, et al. Expression of membrane type 1 matrixmetalloproteinase in human articular cartilage. Arthritis Rheum1997;40:7049.

51. Imai K, Ohuchi E, Aoki T, Nomura H, Fujii Y, Sato H, et al.Membrane-type matrix metalloproteinase 1 is a gelatinolytic en-

zyme and is secreted in a complex with tissue inhibitor ofmetalloproteinases 2. Cancer Res 1996;56:270710.52. Noma T, Glick AB, Geiser AG, OReilly MA, Miller J, Roberts

AB, et al. Molecular cloning and structure of the human trans-forming growth factor- 2 gene promoter. Growth Factors 1991;4:24755.

53. Boumediene K, Conrozier T, Mathieu P, Richard M, Marcelli C,Vignon E, et al. Decrease of cartilage transforming growth fac-tor- receptor II expression in the rabbit experimental osteoar-thritis: potential role in cartilage breakdown. Osteoarthritis Car-tilage 1998;6:1469.

54. Serra R, Johnson M, Filvaroff EH, LaBorde J, Sheehan DM,Derynck R, et al. Expression of a truncated, kinase-defectiveTGF- type II receptor in mouse skeletal tissue promotes terminalchondrocyte differentiation and osteoarthritis. J Cell Biol 1997;139:54152.

55. Amin AR, Attur M, Patel RN, Thakker GD, Marshall PJ, Rediske

J, et al. Superinduction of cyclooxygenase-2 activity in humanosteoarthritis-affected cartilage: influence of nitric oxide. J ClinInvest 1997;99:12317.

56. Fermor B, Weinberg JB, Pisetsky DS, Misukonis MA, Fink C,Guilak F. Induction of cyclooxygenase-2 by mechanical stressthrough a nitric oxide-regulated pathway. Osteoarthritis Cartilage2002;10:7928.

57. Clancy RM, Amin AR, Abramson SB. The role of nitric oxide ininflammation and immunity. Arthritis Rheum 1998;41:114151.

3560 MARTIN ET AL

Effect of Hypoxia

Documents

Transcript of Effect of Hypoxia