doi:10.1006/cyto.2001.0939, available online at http://www.idealibrary.com on

PATHWAYS BY WHICH INTERLEUKIN 17INDUCES ARTICULAR CARTILAGE BREAKDOWN

IN VITRO AND IN VIVO

Liping Cai,1 JianPing Yin,3 Melissa A. Starovasnik,3 Deborah A. Hogue,1

Kenneth J. Hillan,2 John S. Mort,4 Ellen H. Filvaroff

Overexpression of interleukin (IL-)17 has recently been shown to be associated with a numberof pathological conditions. Because IL-17 is found at high levels in the synovial fluid surroundingcartilage in patients with inflammatory arthritis, the present study determined the direct effect ofIL-17 on articular cartilage. As shown herein, IL-17 was a direct and potent inducer of matrixbreakdown and an inhibitor of matrix synthesis in articular cartilage explants. These effectswere mediated in part by leukemia inhibitory factor (LIF), but did not depend on interleukin-1activity. The mechanism whereby IL-17 induced matrix breakdown in cartilage tissue appearedto be due to stimulation of activity of aggrecanase(s), not matrix metalloproteinase(s). However,IL-17 upregulated expression of matrix metalloproteinase(s) in chondrocytes cultured inmonolayer. In vivo, IL-17 induced a phenotype similar to inflammatory arthritis when injectedinto the intra-articular space of mouse knee joints. Furthermore, a related protein, IL-17E, wasfound to have catabolic activity on human articular cartilage. This study characterizes themechanism whereby IL-17 acts directly on cartilage matrix turnover. Such findings haveimportant implications for the treatment of degenerative joint diseases such as arthritis.

� 2001 Academic Press

From the Department of 1Molecular Oncology, 2Pathology, 3ProteinEngineering, Genentech Inc., South San Francisco, CA 94080,USA; 4Joint Diseases Laboratory, Shriners Hospital forChildren, Montreal, Quebec, Canada H3G1A6

Correspondence to: Dr Ellen H. Filvaroff, Department of MolecularOncology, Genentech Inc., MS 37, 1 DNA Way, South SanFrancisco, CA 94080-4990, USA. Tel: 650-225-1159; Fax: 650-225-6497; E-mail: filvarof@gene.com

Received 1 April 2001; received in revised form 30 June 2001;accepted for publication 16 July 2001

� 2001 Academic Press1043–4666/01/190010+12 $35.00/0

KEY WORDS: cartilage/cytokines/interleukins/rheumatoid arthritis/T lymphocytes

Cytokines are involved in the inflammation andcartilage destruction characteristic of arthritic dis-orders. Interleukin (IL-)1 and tumor necrosis factor-�tumor necrosis factor-(TNF-)�, which are present athigh levels in diseased joints,1–3 induce cartilage matrixbreakdown and expression of other pro-inflammatorymolecules, and inhibit synthesis of cartilage matrixproteins. Neutralizing IL-1 or TNF-� activity in vivosuppresses inflammation and/or protects skeletaltissues in arthritic animal models.1,2,4 In order tosuppress inflammatory disease as well as skeletaldestruction in rheumatoid arthritis, a combination oftherapies, i.e. TNF-� inhibitors3 along with IL-1ra,5

may prove to be particularly effective.

10

In humans, as well as in animal models, rheuma-toid arthritis (RA) is characterized by leukocyte infil-tration, synovitis, and pannus formation.1 Activated Tcells are sufficient6 and perhaps necessary7 for induc-tion of cartilage and bone loss in animal models of RA.In humans, activation of T cells is likely requiredespecially in the early stages of RA.8,9 Cytokinesreleased by activated T cells can stimulate macro-phages, fibroblasts and other T cells, thus enhancingthe local immune response.9

IL-17, which is produced by activated T cells, is alikely contributor to the pathogenesis of arthritis.10–12

IL-17 is present at high levels in the synovium andsynovial fluid of arthritic patients10–12 and stimulatesproduction of nitric oxide by human articular carti-lage13 and isolated chondrocytes.14 IL-17 induces pro-duction of pro-inflammatory cytokines by stromalcells, synoviocytes, chondrocytes, and macrophages,some of which can enhance the effects of IL-17.14–17

For example, IL-17 stimulates production of, andcan synergize with, IL-1 and TNF-�.10,15–17 IL-17 alsostimulates production of IL-12, leukemia inhibitoryfactor (LIF), IL-6, IL-8, granulocyte/macrophage-colony stimulating factor (G-CSF) and inflammationmediators, such as prostaglandins (PGE2) and cyclo-oxygenase-2 (COX-2).14–18 Accordingly, treatment of

CYTOKINE, Vol. 16, No. 1 (7 October), 2001: pp 10–21

Catabolic pathways in articular cartilage / 11

RA synovial cultures with antibodies which neutralizeIL-17, causes a substantial reduction in LIF15 andIL-610 production, collagenase activity,19 and osteo-clast formation.11 IL-17 may also be at least part of themechanism whereby other cytokines, such as IL-15,exert proinflammatory activity in vivo.12 Thus, IL-17appears to play a role in both induction and expansionof the proinflammatory cytokine cascade, and, assuch, IL-17 may initiate as well as amplify the jointdestruction which occurs in patients with rheumatoidarthritis.9

Because of the role IL-17 appears to play inregulation of the immune system and its possiblecontribution to clinical disorders, identification andcharacterization of related molecules has been of par-ticular interest.20,21 To this end, a novel member of thisfamily, IL-17E, and its receptor, IL-17Rh1, haverecently been identified.21 IL-17E, which is 16–20%identical to IL-17, IL-17B, and IL-17C at the aminoacid level, is expressed at very low levels in a number oftissues, including brain, kidney, lung, prostate, testis,spinal cord, adrenal gland, and trachea.21 AlthoughIL-17E does not bind to the IL-17 receptor, IL-17Einduces NF-�B activity and IL-8 production infibroblasts in vitro in a manner similar to IL-17.21

While the role of IL-17 in inflammation has beenstudied by a number of investigators, the direct effectsof IL-17 on articular cartilage have not been well-characterized. In the present report, we tested whether

IL-17 has direct effects on cartilage in the absence ofinflammatory cells. To this end, we have (1) measuredthe direct effect of IL-17 on isolated articular cartilage;(2) identified factors which may mediate these effects ofIL-17; (3) determined which enzymes are activated incartilage by IL-17 treatment; (4) characterized in vivoeffects of locally administered IL-17 within the jointand (5) measured the effect of IL-17E on humanarticular cartilage.

RESULTS

0

4000B

cpm

/mg

tiss

ue

350030002500200015001000500

IL-170 1

**

0.2

*

0.11

**

0.2

**

0.1

*

IL-1

0

1.4A

PG

rel

ease

(µg

/mg

tiss

ue)

1.2

1.0

0.8

0.6

0.4

0.2

IL-170 1

**

0.2

**

0.1

*

1

**

0.2

**

0.1IL-1

0

4500D

cpm

/mg

tiss

ue 3500

30002500200015001000500

0 1

**

0.1

**

0

**

1

**

0.1

*

IL-1

0

1.2C

PG

rel

ease

(µg

/mg

tiss

ue)

1.0

0.8

0.6

0.4

0.2

0 1

**

0.1

**

0

**

1

**

0.1IL-1

IL-17

IL-170 11100

0 11100

4000

**

Figure 1. Induction of cartilage matrix breakdown and inhibition of synthesis by IL-17.

Porcine articular cartilage explants were treated with IL-17 or IL-1� at various concentrations (0.1, 0.2 or 1 ng/ml), and proteoglycan release (A,C) and synthesis (B, D) were measured. Data represent the average of five independent samples�SEM. Results presented in A, B are from aseparate experiment from those shown in C, D. *=P<0.05 and **=P<0.005 vs control, by Student’s t-test.

Effect of interleukin-17 on cartilage matrixturnover

To determine whether IL-17 affects cartilagematrix metabolism, porcine articular cartilage explantswere treated with a range of IL-17 concentrations, andproteoglycan synthesis and release were measured. Atconcentrations as low as 0.1 ng/ml, IL-17 inducedsignificant cartilage matrix release (Fig. 1A) and inhib-ited new matrix synthesis (Fig. 1B), with comparablepotency to IL-1�. More specifically, IL-17 at 1 ng/mlincreased proteoglycan release by 60% and decreasedproteoglycan synthesis by 55% relative to untreatedcontrols. When IL-1� (1 ng/ml) and IL-17 (1 ng/ml)were combined, an enhancing, apparently additive,effect was observed on both matrix release (increasedby 130%) (Fig. 1C) and synthesis (decreased by 73%)

12 / Cai et al. CYTOKINE, Vol. 16, No. 1 (7 October, 2001: 10–21)

(Fig. 1D) relative to untreated controls. Unlike pre-vious studies in other cell systems,15 no synergismbetween IL-1� and IL-17 was observed.

To test for species-related effects, the ability ofIL-17 to alter matrix metabolism in bovine articularcartilage explants was measured. While both IL-17 andIL-1� increased proteoglycan breakdown and inhibitedmatrix synthesis in a dose-dependent manner, bovinearticular cartilage was less responsive than porcinetissue to IL-17 as evidenced by the difference in theirresponse to low concentrations (�1 ng/ml) of IL-17(data not shown). At higher concentrations (>1 ng/ml),IL-17 induced matrix breakdown and inhibited matrixsynthesis in bovine articular cartilage explants (datanot shown). This decrease in sensitivity to IL-17 in thebovine tissue could be due to the older age (18–24months) of these animals, since adult bovine cartilagehave been shown to be less growth factor responsivethan cartilage from young or adolescent animals.22

Because diseased human cartilage likely hasunique biological properties and may be relativelyinsensitive to in vitro treatment with specific cytokinessuch as IL-1�, human articular cartilage from individ-uals with or without reported joint degeneration wastested. Two tissue samples from individuals (66 and 92years old) without reported joint degeneration andtwenty-four tissue samples from individuals (ages from39–92 years old) with severe joint degeneration weretested. In both types of cartilage, as well as in samplesfrom human menisci (from individuals ranging inage from 62–70 years old), IL-17 (10 nM) increasedcartilage matrix release and inhibited synthesis ofproteoglycans (data not shown).

Thus, for all species tested, and for both hyalinecartilage (articular cartilage) and fibrocartilage (menis-cus), IL-17 was found to be a potent catabolic cytokine.

IL-17 induction of catabolic proteinsTo determine the role of IL-1 in IL-17-induced

matrix degradation and synthesis, explants weretreated with IL-17 plus IL-1 receptor antagonist (IL-1ra). Although IL-1ra inhibited the effects of IL-1� onarticular cartilage explants, IL-1ra did not block IL-17-induced matrix release (Fig. 2A) nor prevent inhibitionof matrix synthesis by IL-17 (Fig. 2B). Thus, the effectsof IL-17 on matrix degradation and synthesis were notdependent on IL-1 production by chondrocytes.

To determine the role of LIF in IL-17 activity,articular cartilage explants were treated with anti-bodies to LIF (anti-LIF) alone, or in combination withIL-17 or IL-1�. Inhibition of LIF significantlydecreased IL-17 and IL-1� induced matrix breakdown(Fig. 2A) and partially overcame the inhibitory effectsof IL-17 or IL-1� on matrix synthesis (Fig. 2B).Co-treatment with anti-LIF decreased induction of

proteoglycan release by �20% and overcame suppres-sion of proteoglycan synthesis by �40% relative toIL-17 (1 ng/ml) or IL-1� (1 ng/ml) alone. Thus, IL-17-and IL-1�-induced changes in cartilage matrix turn-over appear to be mediated, at least in part, by LIF.Similarly, human articular cartilage explants frompatients with degenerative joint disease treated withanti-LIF antibodies showed significantly higher levelsof proteoglycan synthesis than untreated tissue (datanot shown). The ability of anti-LIF to protect cartilagein the absence of additional cytokines, suggests thatarticular cartilage explants from both young pigs andolder humans constitutively synthesize active LIFunder serum-free conditions in culture (Fig. 2).

To determine whether IL-17 is initiating a cascadeof cytokine expression, human articular cartilageexplants were treated with IL-17, and levels of pro-inflammatory and anti-inflammatory cytokines weremeasured. No change in the levels of IL-2, IL-4, IL-5or IFN-� was found after IL-17 treatment. However,treatment with IL-17 or TNF-� caused an increase inIL-10 protein levels in explants of human articularcartilage or human menisci (Table 1).

0

3500B

cpm

/mg

tiss

ue 3000

2500200015001000500

- 1 ra

††

αLIF

††

-

**

1raαLIF

##

-

**

1raαLIFIL-1IL-17

0

1.2A

PG

rel

ease

(µg

/mg

tiss

ue)

1.0

0.8

0.6

0.4

0.2

- 1 ra

††

αLIF

††

-

**

1raαLIF

##

-

**

1raαLIF

*

IL-1IL-17

Figure 2. Effect of anti-LIF antibodies or IL1ra on PG metabolism.

Porcine articular cartilage explants were treated with IL-17 (1 ng/ml)or IL-1� (1 ng/ml) alone, or in the presence of antibodies to�LIF (1.1 �g/ml) or an interleukin 1 receptor antagonist (IL-1ra)(0.2 �g/ml), and matrix breakdown (A) or synthesis (B) was measured.Data represent the average of five independent samples�SEM.*=P<0.05 and **=P<0.005 vs control, ##=P<0.005 vs IL-17alone, ††=P<0.005 vs IL-1 alone, by Student’s t-test.

IL-17 induction of matrix-degrading enzymesIn order to determine which enzymes are involved

in cytokine-induced matrix breakdown, explants weretreated with actinonin, at a concentration which can

Catabolic pathways in articular cartilage / 13

inhibit ‘‘aggrecanase’’ activity but not matrix metallo-proteinase (MMP) activity.23 In this system, actinonindecreased basal and IL-1�- or IL-17- induced matrixcatabolism (Fig. 3A). Actinonin at 10 µM decreasedproteoglycan release by �20% relative to control andsuppressed IL-17 (1 ng/ml) and IL-1� (1 ng/ml) activityby �40%. These effects of actinonin were not due tocytotoxicity, as actinonin had no untoward effects oneither proteoglycan synthesis in articular cartilageexplants (data not shown) or on survival of isolatedprimary chondrocytes in vitro (data not shown). Theseresults suggest that both IL-1� and IL-17-inducedmatrix breakdown is mediated by aggrecanase(s), andnot by MMPs.

Additional support for this hypothesis was pro-vided by experiments showing no change in the level ofexpression or activation of members of the MMPfamily of enzymes in explants treated with IL-1� orIL-17 (Fig. 3B). Rather suprisingly, when chondro-cytes were grown on tissue culture plates as a mono-layer, both IL-1� and IL-17 induced expressionof MMP-2 and stromelysin (Fig. 3B). However,these enzymes were present predominantly in theirpro-enzyme (i.e. inactive) form.

Analysis of the major proteoglycan degradationproducts released from cells can also be used to deter-mine which enzyme(s) have been activated. One of themajor catabolic sites (between Asn341 and Phe342)found within the interglobular domain (IGD) of aggre-can is believed to be due to MMP activity, and theother (between Glu373 and Ala374) is likely due to‘‘aggrecanase’’ activity. Western blot analysis of aggre-

can fragments released into the media from tissuestreated with IL-1 or IL-17 or control media wasperformed using antibodies which recognized either thenew NH2 terminal ARGSVG generated by aggreca-nase (antibody no. 71), or the new NH2 terminalFFGVG generated by MMP activity (antibody no.247).24 As expected, MMP generated neo-epitopeswere found in explants treated with the MMP activatorp-aminophenylmercuric acetate (APMA; Fig. 4). Incontrast, no such fragments were found in explantstreated with IL-1 or IL-17 (Fig. 4 left panel). Aggreca-nase generated fragments were found in both the IL-1and IL-17 treated samples, and the pattern of thesefragments was similar under the two different condi-tions (Fig. 4 right panel). Namely, bands of reactivitywere detected at �230 kDa, �200 kDa, �150 kDa,�110 and �64 kDa. The high molecular mass band at230 kDa most likely represents the C-terminal aggre-can fragment formed by initial cleavage at the Glu373–Ala374 bond within the IGD. Additional cleavage inthe C terminus at other sites, likely accounts for thesmaller products. Thus, both IL-1 and IL-17 generateproteoglycan fragments indicative of aggrecanase, butnot MMP, activity in articular cartilage explants.

TABLE 1. Induction of IL-10 by IL-17 or TNF-�

Sample Treatment

Cytokine measured (pg/ml)

IL-5 IL-4 IL-2 IL-10

1 Control 1.5 2.2 0 2.6IL-17 1.4 2.9 1.7 5.4

2 Control 0 2 2.2 64.4IL-17 1.7 1.8 2.7 112.0

3 Control 2.1 4.5 0 4.8IL-17 1.9 3.1 0 28.3

4 Control 1.6 1.6 0 6.5IL-17 0 1.1 1.6 25.1

TNF-� 2.6 5.6 2.1 30.45 Control 1.6 2.2 1.7 54.9

IL-17 0 2.2 0 154.6TNF-� 1.4 1.7 1 86.7

M-5 Control 1.5 0 1.6 85.9IL-17 1.5 0 1.7 185.8

TNF-� 0 0 0 107.8

Articular cartilage from clinically normal (1,2) or diseased (3,4,5) joints, and human meniscus from a diseased joint (M-5), wereuntreated (Control) or treated with IL-17 (10 nM) or TNF-� (5 ng/mL). Conditioned media was collected at 48hr and theconcentration of various cytokines (TNF-�, IL-5, IL-4, IL-2, or IL-10) was measured using the Cytometric Bead Array (CBA) kit(Becton Dickonson). Source of human specimens: sample 1 is from a 66 year old Caucasian male without a history of joint disease;sample 2 is from a 92 year old black male without a history of joint disease; sample 3 is from a 65 year old Caucasian female withosteoarthritis (OA); sample 4 is from a 49 year old Caucasian female with OA; and sample 5 and M-5 is from a 69 year old femalewith OA. ‘0’ indicates levels below detection.

In vivo effects of IL-17In the explant system, articular cartilage is

dissected away from surrounding tissues. In order totest the effects of IL-17 in another physiologicallyrelevant system, we treated intact skeletal elements,i.e. whole patellae, with IL-1� or IL-17 in vitro.Under these conditions, IL-17 decreased matrix

14 / Cai et al. CYTOKINE, Vol. 16, No. 1 (7 October, 2001: 10–21)

�

�����

����

���

�� �

� �

�����

��

���

���

���

���

���

���

���

���

�

��

�

��

��

�

��

��

!"� !"��

#���

���

��

$#�

#�

��

� �� %� # � �� �� �

�

�� �� �

��

# � �� �� �

�

�� �� �

��&��

#

���

���

$#�

� �� %�

#�

$#�

#�

'�������($)��

%

*

Figure 3. Effect of an aggrecanase inhibitor on matrix metabolism.

Porcine articular cartilage explants were treated with IL-17 or IL-1 (1 ng/ml) alone or in combination with actinonin (a) (10 �M), an inhibitorof aggrecanase activity. (A) Levels of proteoglycan release were determined. Data represent the average of five independent samples�SEM.**=P<0.005 vs control. ##=P<0.005 vs IL-17 alone. ††=P<0.005 vs IL-1 alone, by Student’s t-test. (B) Detection of MMPs by zymographyin supernatants from chondrocytes either in explants (Tissue explants) or isolated chondrocytes cultured as monolayers (Cells). Explants weretreated with IL-1� (�) at 50 ng/ml, IL-17 (17) at 50 ng/ml, or p-aminophenylmercuric acetate (APMA), a known MMP activator (A) (1 mM).Cells were treated with IL-1� (�) or IL-17 (17) at 1 ng/ml (1), 10 ng/ml (10), or 50 ng/ml (50). The upper panels are gelatin zymograms, and lowerpanels are casein zymograms. pM9, pro-MMP-9; M9, active MMP-9; pM2, pro-MMP-2; M2, active MMP-2; pM3, pro-MMP-3; M3, active,MMP-3.

Catabolic pathways in articular cartilage / 15

337DIPENFFGVG . . .

MMP

247

368NITEGEARGSV. . . . . . . . . . . .

Aggrecanase

71

216

132

78

45

50A 1 5010

β IL-17

247

216

132

78

45

50Con 10 501

βIL-17

71

Figure 4. Analysis of aggrecan fragments released from articular cartilage explants.

Bovine articular cartilage explants were treated with IL-17 (1, 10, or 50 ng/ml), IL-1� (�) (50 ng/ml) or APMA (1 mM) (A), and the media wasanalyzed by Western blotting using specific antibodies recognizing neoepitopes of aggrecan which are exposed upon cleavage by MMPs (antibody247, left panel) or by aggrecanase (antibody 71, right panel). The relatively high basal aggrecanase activity in the control may be due to the factthat explants were cultured in serum-free media. The pattern of bands in IL-1� treated samples (data not shown) was identical to those for IL-1�treated samples. Specific cleavage sites for MMPs or aggrecanase are shown (bottom, in bold) including epitopes recognized by antibody no. 71or antibody no. 247.

synthesis, but to a lesser extent than did IL-1� (datanot shown)25.

In order to test in vivo effects, either IL-17 inbuffer [phosphate-buffered saline (PBS)+0.1% bovineserum albumin (BSA)] or buffer alone was injected intothe intra-articular space of the knee joints of mice.Patellae were harvested 24 h after the second of twodaily injections, and proteoglycan synthesis wasmeasured. Proteoglycan synthesis decreased signifi-cantly at high (80 ng) (Fig. 5) but not low (12 ng) (datanot shown), doses of IL-17. However, the extent ofdecrease with IL-17 at 80ng (decreased by 30%) was

significantly less than that seen with IL-1� (decreasedby 70%), even when much lower doses of IL-1� (12 ng)were used (Fig. 5).

In order to better understand the in vivo effects ofIL-17, joints injected with IL-17 (80 ng) were processedfor histological examination. Joints injected with IL-1�(1 ng) were used as a positive control. Joints from PBSinjected animals were normal or showed, at most, mildperi-articular reactive inflammation around the injec-tion site (Fig. 6A, D and G). Joints from all cytokine-treated animals examined at day 3 showed evidence ofarthritis (Fig. 6B, C, E, F, H and I), characterized by

16 / Cai et al. CYTOKINE, Vol. 16, No. 1 (7 October, 2001: 10–21)

moderate to severe inflammation of peri-articular tis-sues, reactive synovitis, and intra-articular leukocyteinfiltration. Leukocytes were often adherent to thesurface of the articular cartilage, and irregularity ofthe normally smooth articular surface was found. Theinflammatory infiltrate contained both neutrophils andmononuclear cells. The morphologic features of theinfiltrate were indistinguishable between the IL-1� andIL-17 treated animals. The intensity of safranin Ostaining (Fig. 6G, H and I), which highlights thecontent of glycosaminoglycans in articular cartilage,was reduced in the most severely inflamed joints, whencompared with controls.

Effects of a IL-17 homologue, IL-17EBecause of its homology to IL-17, we tested the

activity of a novel cytokine, IL-17E, on human articu-lar cartilage. IL-17E inhibited matrix synthesis andstimulated nitric oxide release and IL-6 production inhuman articular cartilage (Fig. 7). IL-17E at 10 nMdecreased proteoglycan synthesis by 20%, increasednitric oxide release by 400% and increased IL-6production by 330%.

In conclusion, we have shown that IL-17 hasdirect catabolic effects on articular cartilage, which aremediated in part by LIF, but not IL-1. The enzyme(s)responsible for IL-17 induced cartilage matrix break-down appear to be aggrecanase(s), not MMPs.Histological analysis showed that IL-17 inducesinflammation and leukocyte infiltration in vivo in amanner which resembles joints with inflammatoryarthritis. Finally, a protein homologous to IL-17,IL-17E, which does not bind to the IL-17 receptor,

0

3500

IL-1α

cpm

3000

2500

2000

1500

1000

500

– +

**

– +

*

IL-17

Figure 5. In vivo effect of IL-17 on matrix synthesis.

Proteoglycan synthesis was measured in patellae isolated from miceinjected with IL-1 (12 ng) into the right knee (+), and buffer into theleft (�), or from mice injected with IL-17 (80 ng) into the right knee(+) and buffer into the left (�). Shown is the mean value�thestandard error for individual patellae isolated from 5 month oldmice with 5 mice/group. *=P<0.05 and **=P<0.005 vs control, byStudent’s t-test.

induces IL-6 and nitric oxide production, and inhibitsproteoglycan synthesis by human articular cartilage.

DISCUSSION

Loss of cartilage tissue in arthritic patients resultsfrom an imbalance between matrix breakdown andsynthesis. Release of proteoglycans from articular car-tilage leads to impaired chondrocyte function andcartilage biomechanics, and may contribute to loss ofother matrix molecules such as collagens. This is thefirst report to show that IL-17 can act directly onarticular cartilage to stimulate loss of proteoglycansthrough induction of activity of aggrecanase(s), notmatrix metalloproteinases. The severe inflammation ofperi-articular tissues, reactive synovitis, intra-articularleukocyte infiltration, and inhibition of proteoglycansynthesis, found in mouse joints injected with IL-17further support the hypothesis that IL-17 can inducean inflammatory arthritic phenotype in vivo. While thismanuscript was in preparation, it was reported that theeffects of IL-17 on knee joints were similar to those ofIL-1.26 In contrast to our findings, no change inproteoglycan synthesis was detected, perhaps due tothe different dose of IL-17 utilized.26 Unlike other Tcell-derived cytokines, which are difficult to detect inRA synovium, IL-17 appears to be produced at highlevels in RA joints.10–12,27 Thus, production of IL-17may be responsible, at least in part, for the compro-mised articular cartilage volume and integrity whichis one of the major clinical problems of arthriticpatients.

Soluble factors made by T cells, monocytes andsynovial fibroblasts may act in concert as these celltypes are found in close proximity in RA synovium.The fact that IL-17 induces expression of othercytokines, such as TNF-� and IL-1�,15,17 which arefound at high levels in diseased joints,1,2 raises theintriguing possibility that IL-17 is involved in theinitiation of the inflammatory cascade in arthritis.However, the effects of IL-17 were not mediated byproduction of IL-1, suggesting that clinical use ofIL-1ra is not likely to abrogate IL-17 induced tissuebreakdown in arthritic joints. As described herein,IL-17 disrupted cartilage matrix homeostasis and aug-mented the detrimental effects of IL-1� on articularcartilage matrix turnover. Thus, in the absence of highlevels of inhibitory factors such as IL-4,25 IL-17 in adiseased joint could amplify the inflammatory cascade,and exacerbate skeletal tissue breakdown in humanjoints.

The high levels of LIF in synovial fluid of arthriticpatients may be due to IL-17 expression by the syno-vivum.15 As with IL-17, LIF stimulates production ofTNF-� and IL-1�,28 and conversely these cytokines

Catabolic pathways in articular cartilage / 17

Figure 6. In vivo effect of IL-17.

Following intra-articular injection of either IL-17 or IL-1, joints were harvested for analysis. Representative images of knee joints from animalsafter injection with buffer (PBS with 0.1% BSA) (A, D and G), IL-1� (1 ng) (B, E and H) or IL-17 (80 ng) (C, F and I). Panels A–F are stainedwith H&E and G–I with Safranin O. Joints from PBS injected animals were essentially normal. In animals treated with IL-1� and IL-17, thejoints showed a moderate to severe peri-articular mixed inflammatory cell infiltrate (arrows, B and C), reactive synovitis (arrowheads, B, C, Eand F), and arthritis (B, C, E and F) with adherence of intra-articular leukocytes to the articular surface (arrow, E). The articular cartilagesurface in cytokine-treated animals showed mild irregularity (E, F, H and I). The intensity of Safranin O staining of articular cartilage wasreduced in severely inflamed joints (H and I) when compared with controls (G). Scale bar shown in A represents 100 �m in panels A–C. Scalebar in D represents 100 �m in panels D–I.

induce expression of LIF.29,30 In addition, the effects ofLIF on nitric oxide (NO) production are additive withthose of IL-17.18 As shown herein, endogenous LIFproduction appears to mediate, at least in part, theeffects of inflammatory cytokines such as IL-1� orIL-17 on articular cartilage. Thus, inhibition of LIF,for example through the use of antibodies, may proveto be a useful therapy for arthritis, either alone or incombination with other treatments.

Proteases of the MMP and aggrecanase familiesare believed to be responsible at least in part for thecartilage matrix degradation which occurs during jointdestruction (reviewed in 31). Synovial fluid (SF) orarticular cartilage from arthritic patients containproteoglycan fragments containing both MMP- andaggrecanase generated termini.32–34 While IL-17

induces MMP-3 mRNA in isolated human chondro-cytes14 and MMP-9 in macrophages,35 no evidence forincreased MMP activity in IL-17 or IL-1 treated tissueexplants was found. Rather, as shown by examinationof aggrecan fragments in the media and by inhibitionof breakdown with actinonin, IL-17 or IL-1�-inducedrelease of matrix fragments from cartilage explantsappears to be due to stimulation of aggrecanaseactivity. Similarly, increased matrix catabolism inhuman OA cartilage correlates with aggrecanase, notMMP, activity.36 Our findings that chondrocytesplated as monolayer show responses to IL-1� andIL-17 which are distinct from those of chondrocytescultured as explants suggest that the response ofchondrocytes to cytokines can be affected by theirculture conditions.

18 / Cai et al. CYTOKINE, Vol. 16, No. 1 (7 October, 2001: 10–21)

MATERIALS AND METHODS

ReagentsInterleukins (IL-1�, IL-1�, and IL-17), purchased from

R&D systems (Minneapolis, MN, USA), were resuspendedin buffer (PBS+0.1% BSA) prior to use. APMA was pur-chased from Aldrich Chemical Company (Milwaukee, WI,USA). Cetylpyridinium chloride, iodoacetamide, phenyl-methanesulphonyl fluoride (PMSF), pepstatinA, KeratanaseI and chondroitinase ABC were purchased from Sigma.35S-sulfate was purchased from ICN Radiochemicals (Irvine,CA, USA). Anti-rabbit Ig, horse radish peroxidase-linkedwhole antibody (from donkey) were purchased fromAmersham Life Science (Arlington Heights, IL, USA). Kera-tanase II was purchased from Seikagaku (Falmouth, MA,USA). Anti-LIF antibodies were prepared and characterizedat Genentech (San Francisco, CA, USA) as described.37

0

14 000C

IL-6

(pg

/ml)

12 000

10 000

8000

6000

4000

2000

– IL-17E0.1 nM

IL-17E1 nM

IL-17E10 nM

0

8 B

NO

(µM

) 6

4

2

– IL-17E0.1 nM

**

IL-17E1 nM

**

IL-17E10 nM

**

0

500A

cpm

/mg

tiss

ue 400

300

200

100

– IL-17E0.1 nM

IL-17E1 nM

*

IL-17E10 nM

**

Figure 7. Effect of IL-17E on human articular cartilage.

Human cartilage was cultured with IL-17E (at various concen-trations) to determine effects on (A) PG synthesis. (B) NO produc-tion. (C) IL-6 production. *=P<0.05 and **=P<0.005 vs control,by Student’s t-test.

Articular cartilage explantsThe metacarpo-phalangeal joint of 4–6 month old

female pigs or 18–24 month old cows was aseptically opened,and articular cartilage was dissected free of the underlyingbone. The cartilage was pooled, minced, washed and culturedin bulk for at least 24 h in a humidified atmosphere of 95%air and 5% CO2 in serum-free low glucose 50:50 Dulbecco’sModified Eagle Medium (DMEM):F12 media with 0.1%BSA, 100 U/ml penicillin/streptomycin (Gibco, Rockville,MD, USA), 2 mM L-glutamine, 1� GHT, 0.1 mM MEMSodium Pyruvate (Gibco), 20 �g/ml Gentamicin (Gibco),1.25 mg/l Amphotericin B, 10 �g/ml transferrin and 5 �g/mlVitamin E. Approximately 50 mg of articular cartilage wasaliquoted into Micronics tubes and incubated for at least 24 hin above media before changing to media without transferrinand Vitamin E. Test proteins were then added. Media washarvested and changed at various time points (0, 24, 48,72 h). Human knee articular cartilage, received from theNational Disease Research Interchange (Philadelphia, PA,USA), was cultured and treated in explants as above forporcine cartilage. Two tissue samples from individuals (66

We have found that another member of the IL-17 family, IL-17E,21 can induce cartilage matrixcatabolism, production of other cytokines, such asIL-6, as well as potential mediators of inflammationsuch as nitric oxide. Thus, IL-17E appears to haveactivities similar to that of IL-17, despite previousfindings that IL-17E binds to a distinct receptor,namely IL-17Rh1, but not to the IL-17 receptoritself.21 Our results suggest that signaling throughIL-17Rh1 may result in downstream activities similarto that of stimulation of the IL-17 receptor. Similarly,IL-17E has been shown to be expressed in tracheae,which contains significant amounts of cartilage.21

Further analysis of IL-17E expression in normal anddiseased human tissues will enable us to determine therole of this protein in skeletal disorders.

Our data show that IL-17 is active on intactarticular cartilage in the absence of other inflammatorycells or synoviocytes. The ability of IL-17 to increasecartilage matrix breakdown appears to be mediated bymembers of the aggrecanase family, while IL-17 stimu-lates expression of MMP in human RA synoviocytes.19

The effects of IL-17 on cartilage may result, at least inpart, from induction of other catabolic proteins, suchas leukemia inhibitory factor. Local, in vivo treatmentof joint tissues with IL-17 induces a phenotype similarto inflammatory arthritis. Finally, a recently clonedmember of the IL-17 family, IL-17E, has cataboliceffects on human articular cartilage.

Blockade of IL-17, through use of anti-inflammatory or anabolic cytokines, or through neu-tralizing antibodies or small molecule inhibitors, mayprove to be an effective treatment for disorders ofcartilage and bone such as arthritis. In vitro inhibitionof IL-17 production by RA synovium has been shownto inhibit production of catabolic cytokines11 andenzymes.19 Of particular interest will be whether inhi-bition of IL-17 in vivo will suppress inflammatorydisease activity as well as protect against cartilage andbone destruction.

Catabolic pathways in articular cartilage / 19

and 92 years old) without reported joint degeneration andtwenty-four tissue samples from individuals (ages from 39–92years old) undergoing joint replacements, as well as foursamples from human menisci (from individuals ranging inage from 62–70 years old) were tested.

Measurement of proteoglycan breakdown inarticular cartilage explants

Media harvested at various time points were assayedfor amount of proteoglycans (PG) using the 1,9-dimethylmethylene blue (DMB) colorimetric assay.38 Themeasure of glycosaminoglycan in the media is a good indi-cator of proteoglycan degradation.39 Values shown representthe amount of PG released per wet weight of articularcartilage tissue. Chondroitin sulfate (Sigma, St. Louis, MO,USA) ranging from 0.0–5.0 �g was used to make the stan-dard curve.

Measurement of proteoglycan synthesis inarticular cartilage explants

After 2 days treatment with factors, 35S-sulfate (to afinal concentration of 10 �Ci/ml) was added to the cartilageexplants, and tissue was incubated for 17 h at 37�C40. Mediawas then saved for measurements of nitric oxide or pro-teoglycan content, and cartilage pieces were washed twotimes using explant media. Digestion buffer containing10 mM EDTA (pH 8.0), 0.1 M Sodium phosphate (pH 6.5)and 1 mg/ml proteinase K (Gibco BRL) was added to eachtube and incubated overnight in a 50�C water bath. Thedigest supernatant was mixed with an equal amount of 10%W/V cetylpyridinium chloride (Sigma). Samples were spun at1000�g for 15 min. The supernatant was removed, and500 �l formic acid (Sigma) was added to the samples todissolve the precipitate. Solubilized pellets were transferredto scintillation vials containing 10 ml scintillation fluid(ICN), and samples were read in a scintillation counter.

Measurement of NO10 �l of 0.05 mg/ml 2,3-diaminonapthalene (DAN) in

0.62 M HCl was added to 100 �l media from explants.Samples were mixed and incubated at room temperaturefor 10–20 min. Reaction was terminated with 5 �l of2.8 M NaOH. The fluorescent product, 2,3-diamino-naphthotriazole, was measured using a Cytoflor fluorescentplate reader with excitation at 360 nm and emission read at409 nm.

Western blot analysisCetylpyridinium chloride (CPC) was added to culture

media from explants treated for 3 days to a final concen-tration of 1% (W/V). Precipitated proteoglycans and pro-teoglycan fragments were collected by centrifugation. Thepellet was washed with 1% (W/V) CPC then dissolved inisopropanol/water (3:2, V/V). Two volumes of ethanol satu-rated with potassium acetate were added at 4�C, and theproteoglycan samples (now as their potassium salts) werecollected by centrifugation. The pellet was then washed twicewith ethanol, then with ether, and air dried.

Proteoglycan samples were dissolved at 10 mg/ml in0.1 M Tris–acetate, pH 7.0, containing 10 mM EDTA,10 mM iodoacetamide, 5 mM phenylmethanesulphonylfluoride, 0.36 mM pepstatin A, 0.24 unit/ml keratanase I(endo �-galactosidase, Sigma), 0.1 unit/40 mg Keratanase IIand 0.12 unit/ml chondroitinase ABC and incubated at 37�Covernight. Digestion was terminated by addition of sodiumdodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and incubation for 3 min in a boilingwater bath. Samples were analysed on 4–12% Tris–glycineSDS-PAGE gradient gels followed by electroblotting tonitrocellulose membranes (Novex, San Diego, CA, USA),which were probed with 1:1000 dilution of the monoclonalantibody raised in rabbit recognizing the aggrecanase gener-ated N-terminal neopeptide ARGSV . . . (antibody no. 71)and the MMP-generated N-terminal IGD neoepitopeFFGVG . . . (antibody no. 247).24 Subsequently, membraneswere incubated with sheep anti-rabbit Ig horseradish peroxi-dase conjugate (Amersham Life Science) and aggrecan cata-bolites visualized by incubation with the SuperSignal WestPico Chemiluminescent Substrate (Pierce, Rockford, IL,USA) for 5 min and then exposed.

In vivo experimentsRecombinant murine IL-1� or IL-17 (R&D Systems) in

a volume of 3 �l in buffer [PBS with 0.1% BSA (Sigma)] wasinjected through the intrapatellar ligaments into the jointspace of C57Bl6 mice. Buffer alone (PBS with 0.1% BSA) wasused as a control. Mice were killed 24 h after the secondinjection (of two daily injections), and patellae were eitherharvested for measurements of PG synthesis, or included inthe joint tissues fixed for histological analysis.

Histological analysisFollowing sacrifice of animals, knees were fixed in 4%

buffered formalin, followed by decalcification in Formical for4–8 h. Samples were then processed for paraffin embeddingand for histological assessment. Three micron thick stepsections were cut in the coronal plane and stained withhematoxylin and eosin (H&E) or safranin O.

Measurement of proteoglycan synthesis in isolatedpatellae

Cartilage PG synthesis was measured by sulfate incor-poration into patellae from 5 month old mice ex vivo. Briefly,patellae were dissected away from the patellar tendon andother soft tissues, labeled with 35S sodium sulfate (30 �Ci/ml)for 4 h, washed and fixed in 4% buffered formalin overnight.Patellae were then decalcified in 5% formic acid for 4 h, andcartilage was dissected away from underlying bone. Thepatellar cartilage was then transferred to scintillation vialscontaining 500 �l of the solubilizer Solvable (PackardBioscience, Meriden, CT, USA) and incubated at 60�C for1.5 h. 10 ml of scintillation fluid (HIONIC-fluor, PackardBioscience, Meriden, CT, USA) was added, and sampleswere counted.

Cytometric bead assayNormal human cartilage, diseased human cartilage and

diseased human meniscus were cultured in the same way as

20 / Cai et al. CYTOKINE, Vol. 16, No. 1 (7 October, 2001: 10–21)

porcine explant as described above. Conditioned media at48 h was harvested and straight media was assayed forTh1/Th2 cytokines using the Cytometric Bead Array (CBA)kit (Becton Dickonson, San Diego, CA, USA).

ZymographyConditioned media from cartilage explant culture or

primary chondrocyte culture were mixed with NOVEX Tris–Glycine SDS Sample Buffer (2�) and let stand 10 min atroom temperature. Samples were applied to 10% Zymogram(gelatin) gels or 12% Zymogram (casein) gels and run forabout 90 min at 125 V. After running, gels were first incu-bated in 1�NOVEX Zymogram Renaturing Buffer withgentle agitation for 30 min at room temperature, then equi-librated in 1�Zymogram Developing Buffer for 30 min atroom temperature with gentle agitation. Fresh 1� develop-ing buffer was then added and gels were incubated overnightat 37�C for maximum sensitivity. GELCODE Blue StainReagent from Pierce was used to stain the gels. Areas ofprotease activity will show up as clear bands.

IL-17E expression and purificationDNA (DNA147531) encoding IL-17E21 was first ampli-

fied by polymerase chain reaction (PCR), then subcloned intopET15b (Novagen, Madison, WI, USA) in order to intro-duce an N-terminal His-tag and thrombin cleavage site. Afteranother PCR step, the coding region was subcloned into thebaculovirus transfer vector pAcGP67B (PharMingen, SanDiego, CA, USA). The transfer vector was co-transfectedwith BaculoGold DNA (PharMingen, San Diego, CA, USA)into Sf9 cells, and recombinant virus was isolated andamplified to 2�108 pfu/ml. For protein production, Hi5 cellswere infected with amplified virus. After 3 days culture at27�C, the medium was harvested by centrifugation, andrecombinant IL-17E was purified on a Ni-NTA (Qiagen)column. The majority of the IL-17E was present as adisulfide-bonded dimer, although a portion ran as a mono-mer on non-reducing SDS PAGE. IL-17E fractions werepooled and dialyzed into PBS, pH 6.5, together with 1unit/mg thrombin (Calbiochem, San Diego, CA, USA) over-night at 4�C. The sample was then concentrated, and thethrombin and His-tag were removed by purification over aSuperdex-75 column in 50 mM sodium phosphate, pH 6.0,500 mM NaCl; both forms of IL-17E migrated as a dimer.IL-17E fractions were pooled and dialyzed into 50 mMsodium phosphate, pH 6.0, 150 mM NaCl. N-terminalsequencing confirmed the identity of the purified protein asIL-17E having an additional four amino acids (GSHM) atthe N-terminus introduced by the vector.

Statistical analysisValues are expressed as the mean�SEM. Differences

between experimental groups were tested using Student’st-test.

Acknowledgements

The authors are grateful for the support andadvice of Dr T. A. Stewart. In addition, we thank

Noelyn Desjardin and Bruce Hultgren for their tech-nical assistance, and Zhonghua Lin, Pam Suboc andDavid Fei for their help in initiating parts of thisproject.

REFERENCES

1. Arend WP, Dayer JM (1990) Cytokines and cytokineinhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum33:305–315.

2. Arend WP, Dayer JM (1995) Inhibition of the productionand effects of interleukin-1 and tumor necrosis factor-� inrheumatoid arthritis. Arthritis Rheum 38:151–160.

3. Feldmann M, Elliott MJ, Woody JN, Maini RN (1997)Anti-tumor necrosis factor-alpha therapy of rheumatoid arthritis.Adv Immunol 64:283–350.

4. Joosten LA, Helsen MM, Saxne T, van de Loo FA,Heinegard D, van den Berg WB (1999) IL-1�� blockade preventscartilage and bone destruction in murine type II collagen-induced arthritis, whereas TNF-� blockade only ameliorates jointinflammation. J Immunol 163:5049–5055.

5. Bresnihan B, Alvaro-Gracia JM, Cobby M, Doherty M,Domljan Z, Emery P, Nuki G, Pavelka K, Rau R, Rozman B, WattI, Williams B, Aitchison R, McCabe D, Musikic P (1998) Treatmentof rheumatoid arthritis with recombinant human interleukin-1receptor antagonist. Arthritis Rheum 41:2196–2204.

6. Kong YY, Feige U, Sarosi I, Bolon B, Tafuri A, Morony S,Capparelli C, Li J, Elliott R, McCabe S, Wong T, Campagnuolo G,Moran E, Bogoch ER, Van G, Nguyen LT, Ohasi PS, Lacey DL,Fish E, Boyle WJ, Penninger JM (1999) Activated T cells regulatebone loss and joint destruction in adjuvant arthritis throughosteoprotegerin ligand. Nature 402:304–308.

7. Panayi GS, Lanchbury JS, Kingsley GH (1992) The impor-tance of the T cell in initiating and maintaining the chronic synovitisof rheumatoid arthritis. Arthritis Rheum 35:729–735.

8. Kingsley GH, Panayi GS (1997) Joint destruction inrheumatoid arthritis: biological bases. Clin Exp Rheumatol 15:S3–S14.

9. Miossec P (2000) Are T cells in rheumatoid synoviumaggressors or bystanders? Curr Opin Rheumatol 12:181–185.

10. Chabaud M, Durand JM, Buchs N, Fossiez F, Page G,Frappart L, Miossec P (1999) Human interleukin-17: a T-cell derivedproinflammatory cytokine produced by the rheumatoid synovium.Arthritis Rheum 42:963–970.

11. Kotake S, Udagawa N, Takahashi N, Matsuzaki K, Itoh K,Ishiyama S, Saito S, Inoue K, Kamatani N, Gillespie MT, MartinTJ, Suda T (1999) IL-17 in synovial fluids from patients withrheumatoid arthritis is a potent stimulator of osteoclastogenesis. JClinical Invest 103:1345–1352.

12. Ziolkowska M, Koc A, Luszczykiewicz G, Ksiezopolska-Pietrzak K, Klimczak E, Chwalinska-Sadowska H, Maslinski W(2000) High levels of IL-17 in rheumatoid arthritis patients: IL-15triggers in vitro IL-17 production via cyclosporin A-sensitivemechanism. J Immunol 164:2832–2838.

13. Attur MG, Patel RN, Abramson SB, Amin AR (1997)Interleukin-17 up-regulation of nitric oxide production in humanosteoarthritis cartilage. Arthritis Rheum 40:1050–1053.

14. Shalom-Barak T, Quach J, Lotz M (1998) Interleukin-17induced gene expression in articular chondrocytes is associated withactivation of mitogen-activated protein kinases and NF-�B. J BiolChem 273:27 467–27 473.

15. Chabaud M, Fossiez F, Taupin JL, Miossec P (1998)Enhancing effect of IL-17 on IL-1-induced IL-6 and leukemiainhibitory factor production by rheumatoid arthritis synoviocytesand its regulation by Th2 cytokines. J Immunol 161:409–414.

16. Fossiez F, Djossou O, Chomarat P, Flores-Romo L,Ait-Yahia S, Maat C, Pin JJ, Garrone P, Garcia E, Saeland S,Blanchard D, Gaillard C, Das Mahapatra B, Rouvier E, Golstein P,Banchereau J, Lebecque S (1996) T-cell interleukin-17 induces

Catabolic pathways in articular cartilage / 21

stromal cells to produce proinflammatory and hematopoieticcytokines. J Exp Med 183:2593–2603.

17. Jovanovic DV, DiBattista JA, Martel-Pelletier J, JolicoeurFC, He Y, Zhang M, Mineau F, Pelletier JP (1998) IL-17 stimulatesthe production and expression of proinflammatory cytokines,IL-beta and TNF-alpha, by human macrophages. J Immunol160:3513–3521.

18. Martel-Pelletier J, Mineau F, Jovanovic D, DiBattista JA,Pelletier JP (1999) Mitogen-activated protein kinase and nuclearfactor �B together regulate interleukin-17-induced nitric oxideproduction in human osteoarthritic chondrocytes. Arthritis Rheum42:2399–2409.

19. Chabaud M, Garnero P, Dayer JM, Guerne PA, Fossiez F,Miossec P (2000) Contribution of interleukin 17 to synovium matrixdestruction in rheumatoid arthritis. Cytokine 12:1092–1099.

20. Li H, Chen J, Huang A, Stinson J, Heldens S, Foster J,Dowd P, Gurney AL, Wood WI (2000) Cloning and characterizationof IL-17B and IL-17C, two new members of the IL-17 cytokinefamily. Proc Natl Acad Sci 97:773–778.

21. Lee J, Ho WH, Maruoka M, Corpuz RT, Baldwin DT,Foster JS, Goddard AD, Yansura DG, Vandlen RL, Wood WI,Gurney AL (2001) IL-17E, a novel proinflammatory ligandfor the IL-17 receptor homolog IL-17Rh1. J Biol Chem276:1660–1664.

22. Luyten FP, Yu YM, Yanagishita M, Vukicevic S,Hammonds RG, Reddi AH (1992) Natural bovine osteogenin andrecombinant human bone morphogenetic protein-2B are equipotentin the maintenance of proteoglycans in bovine articular cartilageexplant cultures. J Biol Chem 267:3691–3695.

23. Hughes CE, Little CB, Buttner FH, Bartnik E, Caterson, B(1998) Differential expression of aggrecanase and matrix metallopro-teinase activity in chondrocytes isolated from bovine and porcinearticular cartilage. J Biol Chem 273:30 576–30 582.

24. Sztrolovics R, White RJ, Roughley PJ, Mort JS (2001) Themechanism of aggrecan release from cartilage differs with tissueorigin and the agent used to stimulate catabolism. Biochem J (inpress).

25. Lubberts E, Joosten LA, van de Loo FA, van denBersselaar LA, van den Berg WB (2000) Reduction of interleukin-17-induced inhibition of chondrocyte proteoglycan synthesis inintact murine articular cartilage by interleukin 4. Arthritis andRheum 43:1300–1306.

26. Dudler J, Renggli-Zulliger N, Busso N, Lotz M, So A(2000) Effect of interleukin 17 on proteoglycan degradation inmurine knee joints. Ann Rheum Dis 59:529–532.

27. Aarvak T, Chabaud M, Miossec P, Natvig JB (1999) IL-17is produced by some proinflammatory Th1/Th0 cells but not by Th2cells. J Immunology 162:1246–1251.

28. Villiger PM, Geng Y, Lotz M (1993) Induction ofcytokine expression by leukemia inhibitory factor. J Clin Invest91:1575–1581.

29. Lotz M, Moats T, Villiger PM (1992) Leukemia inhibitoryfactor is expressed in cartilage and synovium and can contribute tothe pathogenesis of arthritis. J Clin Invest 90:888–896.

30. Campbell IK, Waring P, Novak U, Hamilton JA (1993)Production of leukemia inhibitory factor by human articularchondrocytes and cartilage in response to interleukin-1 and tumornecrosis factor alpha. Arthritis Rheum 36:790–794.

31. Smith RL (1999) Degradative enzymes in osteoarthritis.Front Biosci 4:D704–D712.

32. Sandy JD, Flannery CR, Neame PJ, Lohmander LS (1992)The structure of aggrecan fragments in human synovial fluid.Evidence for the involvement in osteoarthritis of a novel proteinasewhich cleaves the Glu373-Ala 374 bond of the interglobular domain.J Clin Invest 89:1512–1516.

33. Lohmander LS, Neame PJ, Sandy JD (1993) The structureof aggrecan fragments in human synovial fluid. Evidence thataggrecanase mediates cartilage degradation in inflammatory jointdisease, joint injury, and osteoarthritis. Arthritis Rheum 36:1214–1222.

34. Lark MW, Bayne EK, Flanagan J, Harper CF, HoernerLA, Hutchinson NI, Singer II, Donatelli SA, Weidner JR, WilliamsHR, Mumford RA, Lohmander LS (1997) Aggrecan degradation inhuman cartilage. Evidence for both matrix metalloproteinase andaggrecanase activity in normal, osteoarthritic, and rheumatoidjoints. J Clin Invest 100:93–106.

35. Jovanovic DV, Martel-Pelletier J, Di Battista JA, MineauF, Jolicoeur FC, Benderdour M, Pelletier JP (2000) Stimulation of92-kd gelatinase (matrix metalloproteinase 9) production byinterleukin-17 in human monocyte/macrophages. Arthritis Rheum43:1134–1144.

36. Little CB, Flannery CR, Hughes CE, Mort JS, RoughleyPJ, Dent C, Caterson B (1999) Aggrecanase versus matrix metallo-proteinases in the catabolism of the interglobular domain ofaggrecan in vitro. Biochem J 344:61–68.

37. Kim KJ, Alphonso M, Schmelzer CH, Lowe D (1992)Detection of human leukemia inhibitory factor by monoclonalantibody based ELISA. J Immunol Methods 156:9–17.

38. Farndale RW, Buttle DJ, Barrett AJ (1986) Improvedquantitation and discrimination of sulphated glycosaminoglycans byuse of dimethylmethylene blue. Biochem Biophys Acta 883:173–177.

39. Handley CJ, Buttle DJ (1995) Assay of proteoglycandegradation. Methods Enzymology 248:47–58.

40. Von den Hoff JW, Van Kampen GPJ, Van de Stadt RJ,Van der Korst JK (1993) Kinetics of proteoglycan turnover in bovinearticular cartilage explants. Matrix 13:195–201.