Adenoviral Vector-Mediated Overexpression of IL-4 in the Joint of Mice with Collagen-Induced Prevents Destruction This information is current as of September 25, 2021. Erik Lubberts, Leo A. B. Joosten, Liduine van den Bersselaar, Monique M. A. Helsen, Andrew C. Bakker, Joyce B. J. van Meurs, Frank L. Graham, Carl D. Richards and Wim B. van den Berg

J Immunol 1999; 163:4546-4556; ; Downloaded from http://www.jimmunol.org/content/163/8/4546

References This article cites 53 articles, 9 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/163/8/4546.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 25, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Adenoviral Vector-Mediated Overexpression of IL-4 in the Knee Joint of Mice with Collagen-Induced Arthritis Prevents Cartilage Destruction1

Erik Lubberts,2* Leo A. B. Joosten,* Liduine van den Bersselaar,* Monique M. A. Helsen,* Andrew C. Bakker,* Joyce B. J. van Meurs,* Frank L. Graham,† Carl D. Richards,† and Wim B. van den Berg*

Rheumatoid arthritis is a chronic inflammatory joint disease, leading to cartilage and bone destruction. In this study, we inves- tigated the effects of local IL-4 application, introduced by a recombinant human type 5 adenovirus vector, in the knee joint of mice with collagen-induced arthritis. One intraarticular injection with an IL-4-expressing virus caused overexpression of IL-4 in the Downloaded from mouse knee joint. Enhanced onset and aggravation of the synovial inflammation were found in the IL-4 group. However, despite ongoing inflammation, histologic analysis showed impressive prevention of chondrocyte death and cartilage erosion. In line with this, chondrocyte proteoglycan synthesis was enhanced in the articular cartilage. This was quantified with ex vivo 35S-sulfate incorporation in patellar cartilage and confirmed by autoradiography on whole knee joint sections. Reduction of cartilage erosion was further substantiated by lack of expression of the stromelysin-dependent cartilage proteoglycan breakdown neoepitope VDIPEN in the Ad5E1 mIL-4-treated knee joint. Reduced metalloproteinase activity was also supported by markedly diminished http://www.jimmunol.org/ mRNA expression of stromelysin-3 in the synovial tissue. Histologic analysis revealed marked reduction of polymorphonuclear cells in the synovial joint space in the IL-4-treated joints. This was confirmed by immunolocalization studies on knee joint sections using NIMP-R14 staining and diminished mRNA expression of -inflammatory protein-2 in the synovium tissue. mRNA levels of TNF-␣ and IL-1␤ were suppressed as well, and IL-1␤ and production by arthritic synovial tissue were strongly reduced. Our data show an impressive cartilage-protective effect of local IL-4 and underline the feasibility of local gene therapy with this in arthritis. The Journal of Immunology, 1999, 163: 4546–4556.

heumatoid arthritis (RA)3 is a chronic inflammatory joint cartilage destruction, whereas TNF is a major cytokine in the in- disease, which is characterized by uncontrolled prolifer- flammatory process. Treatment with TNF Abs or soluble receptors by guest on September 25, 2021 R ation of synoviocytes and massive infiltration of leuko- markedly reduced the inflammatory signs in RA patients (4–6). cytes (1, 2). The perpetuation of the synovitis is accompanied by Apart from direct interference with TNF/IL-1, regulation can be erosion of bone and cartilage. Since a specific causative agent or exerted at the level of modulatory such as IL-10 and Ag has not been identified yet, bypassing the potential Ag and IL-4. IL-4 has the potential to effectively antagonize inflammatory targeting the cytokine disbalance might represent a solid way to and destructive mediators of the RA process (7). Induced synovial control this disease. production of IL-1 and TNF is strongly reduced; IL-1Ra is en- ␣ It is now generally accepted that the cytokines TNF- and IL-1 hanced; and spontaneous production of IL-6, TNF, leukemia-in- are principal mediators in RA (3, 4). These proinflammatory cy- hibitory factor, and PGE2 is also inhibited by IL-4 (8, 9). More- tokines are produced in increased quantities by RA synovium and over, IL-4 regulates expression and release of soluble receptors, detected in synovial fluid. IL-1, in particular, is a key cytokine in including IL-1RII and the TNF receptors (10, 11). Finally, in vitro studies demonstrated suppressive effects of IL-4 on Th1 activity (12, 13) and beneficial effects on cartilage matrix degradation (14– * Research Lab, Department of Rheumatology, University Hospital Nijmegen, Nijmegen, The Netherlands; and †Department of Pathology, McMaster 17). Of great importance, IL-4 could not be detected in synovial University, Hamilton, Ontario, Canada fluid, synovial supernatants, or synovium of RA patients (13). This Received for publication February 22, 1999. Accepted for publication August 2, 1999. lack of IL-4 is likely to contribute to the uneven Th1/Th2 balance The costs of publication of this article were defrayed in part by the payment of page and to the chronic nature of RA (13, 18). charges. This article must therefore be hereby marked advertisement in accordance In addition to the in vitro work, therapeutic studies have been with 18 U.S.C. Section 1734 solely to indicate this fact. done with IL-4 in animal models of arthritis. Due to the short t1/2, 1 This work was supported by grants from The Dutch League against Rheumatism, from the European Union (Biomed-2 Programm, Contract BMH4-CT96-1698), and treatment was systemic and injections were given daily. Suppres- partly supported by the Hamilton Health Sciences Corporation and St. Joseph’s Hos- sion of collagen arthritis and streptococcal cell wall arthritis was pital, Hamilton, ON, Canada. demonstrated, whereas synergy with IL-10 was evident (19, 20). 2 Address correspondence and reprint requests to Dr. Erik Lubberts, University Hos- However, a major risk of prolonged systemic treatment is the in- pital Nijmegen, Department of Rheumatology, Rheumatology Research Lab, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail address: E.Lubberts@ duction of generalized immunosuppression or allergic reactions at reuma.azn.nl multiple sites. Local gene transfer with a viral vector might cir- 3 Abbreviations used in this paper: RA, ; CIA, collagen-induced cumvent these problems. Evidence for the potential of successful arthritis; CII, collagen type II; H&E, hematoxylin and eosin; i.a., intraarticular; IL- 1Ra, IL-1 receptor antagonist; MMP, metalloproteinase; PMN, polymorphonuclear gene delivery of other cytokine-related constructs to synovial tis- cell; RT, room temperature; TIMP-1, tissue inhibitor of metalloproteinase. sue has already been obtained (21–25).

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 4547

In the present study, we examined the effects of local IL-4 over- Scoring was done by two independent observers, without knowledge of the expression, through a recombinant human type 5 adenovirus vec- experimental groups. tor, in the knee joint of mice with collagen arthritis. We found that this treatment greatly protects against cartilage erosions, despite Determination of IL-4, IL-1␤, and IL-1Ra protein ongoing inflammation. The protective effect was associated with a To determine the levels of IL-4, IL-1␤, and IL-1Ra in patella washouts, reduction of PMNs in the synovial joint space, decreased NO syn- patellae were isolated in a standardized manner from knee joints, as pre- thesis, down-regulation of IL-1␤, and a reduction of the MMP-3/ viously described (28). Patellae were incubated in RPMI 1640 medium TIMP disbalance in the synovium. with 0.1% BSA, gentamicin (50 ␮g/ml), and L-glutamine (2 mM) (200 ␮l/patella) for1hatRT.After supernatant was harvested, the IL-4, IL-1␤, Materials and Methods and IL-1Ra levels were measured by ELISA. Briefly, ELISA plates were coated with the capture Ab (3 ␮g/ml) by overnight incubation at 4°C in Animals carbonate buffer (pH 9.6). Nonspecific binding sites were blocked by 1-h Male DBA-1/BOM mice were purchased from Bomholdga¨rd (Ry, Den- incubation at 37°C with 1% BSA in PBS/Tween. The supernatants from mark). The mice were housed in filter-top cages. The mice were immu- the patella cultures were tested by 3-h incubation at 37°C. The plates were nized between 10 and 12 wk of age. Water and food were provided ad then incubated for 1.5 h at 37°C with the biotinylated second Ab, followed libitum. by a 30-min incubation at 37°C with streptavidin-polyperoxidase conju- gate. Bound complexes were detected by reaction with orthophenylenedia- Adenoviral vectors mine and H2O2. Absorbance was measured at 492 nm using an ELISA plate reader (Titertek Multiscan MCC/340). The cytokine concentration in The recombinant replication-deficient adenovirus Ad5E1 mIL-4 was gen- the samples was calculated as pg/ml using recombinant murine IL-4, IL- erated by homologous recombination after cotransfecting 293 cells with 1␤, or IL-1Ra as a standard. The sensitivity of the IL-4 and IL-1␤ ELISA Downloaded from PACCMVmIL-4 and a virus-rescuing vector pAdBHG10, as described is 10 pg/ml, and of the IL-1Ra ELISA is 160 pg/ml. (26). The empty recombinant replication-deficient adenovirus Ad5de170-3 was used as a control vector throughout the study. High titers of recom- binant adenoviruses were amplified, purified, titered, and stored, as de- Isolation of RNA scribed (27). Mice were sacrificed by cervical dislocation, and the patellae and adjacent Materials synovium were immediately dissected (29). Synovium biopsy tissue was taken from 6 of 10 patella specimens. Two biopsy specimens with a di- Freund’s complete adjuvant and Mycobacterium (strain ameter of 3 mm were punched out, using a biopsy punch (Stifle, Wacht- http://www.jimmunol.org/ H37Ra) were obtained from Difco (Detroit, MI). Bovine type II collagen ersbach, Germany): one from the lateral side and one from the medial side. (CII) was prepared as described (20). RPMI 1640 was obtained from Life Three lateral and three medial biopsy samples were pooled to yield two Technologies (Breda, The Netherlands). ELISA plates (Maxisorb) were samples per group. The synovium samples were immediately frozen in purchased from Nunc (Copenhagen, Denmark). The following mAbs were liquid nitrogen. Ten patella specimens per experimental group were taken. used in the cytokine ELISAs: anti-murine IL-4 Abs (capture 18031D, de- Patellae were transferred to a 5% EDTA solution and kept on ice for 4 h. tection 18042D) were purchased from PharMingen (San Diego, CA). Anti- Thereafter, the cartilage layer was stripped, as previously described (20). ␤ murine IL-1 Abs (capture PM-425-B, detection MM-425-B) were ob- This procedure does not affect mRNA isolation or amplification efficiency. tained from Endogen (Cambridge, MA). Anti-murine IL-1Ra Abs (capture Total RNA from a pool of 10 cartilage samples from a particular group was MAB480, detection BAF480) were from R&D Systems (Minneapolis, extracted with 1 ml of Trizol reagent, an improved single-step RNA iso-

MN). Streptavidin-polyperoxidase conjugate was obtained from CLB (Am- lation method based on the method described by Chomczynski and Sacchi by guest on September 25, 2021 sterdam, The Netherlands). Murine rIL-4 was a kind gift of Dr. S. Smith (30). Synovium biopsy samples were ground to powder using a microdis- ␤ (Schering-Plough, Kenilworth, NJ). Murine rIL-1 was from R&D Sys- membrator II (B. Braun, Melsungen, Germany). Total RNA was extracted tems. NIMP-R14 (rat IgG2b anti-mouse mAb) was kindly provided by Dr. in 1 ml of Trizol reagent in a manner similar to that used for cartilage M. Strath (National Institute for Medical Research, London, U.K.). Anti- samples. VDIPEN IgG Abs were a generous gift of Drs. I. Singer and E. Bayne (Merck Research Laboratories, Rahway, NJ). PCR amplification Induction of CIA One microgram of synovial RNA and the total amount of cartilage RNA Bovine CII was diluted in 0.05 M acetic acid to a concentration of 2 mg/ml (pool of 10 cartilage layers) was used for RT-PCR. mRNA was reverse and was emulsified in equal volumes of Freund’s complete adjuvant (2 transcribed to cDNA using oligo(dT) primers, and one-twentieth of the mg/ml of M. tuberculosis). The mice were immunized intradermally at the cDNA was used in one PCR amplification. PCR was performed at a final ␮ ␮ base of the tail with 100 l of emulsion (100 g of collagen). On day 21, concentration of 200 ␮M dNTPs, 0.1 ␮M of each primer, and1UofTaq ␮ mice were given an i.p. booster injection of 100 g of CII diluted in PBS, As meant? Please verify. polymerase (Life Technologies) in standard PCR and normally arthritis onset occurs at about day 28. buffer. The mixture was overlaid with mineral oil and amplified in a ther- ␣ Study protocol mocycler (Omnigene, Hybaid, U.K.). Message for GAPDH, IL-1, TNF- , IL-1Ra, and inducible NO synthase was amplified using the primers de- CIA was induced in male DBA-1 mice, as described above. Just before scribed elsewhere (20). Primers for MMP-3, TIMP-1, macrophage-inflam- expected onset of CIA, mice were scored visually for the appearance of matory protein-2, and monocyte-chemotactic protein-1 were designed us- arthritis. Mice without macroscopic signs of arthritis in the paws were ing Oligo 4.0 and Primer Software. selected. Mice were anesthetized with ether, and a small aperture in the Samples (5 ␮l) were taken from the reaction tubes after a certain num- skin of the knee was performed for the intraarticular (i.a.) injection pro- ber of cycles. PCR products were separated on 1.6% agarose and stained cedure. When absence of arthritis was confirmed in the knee joint, i.a. with ethidium bromide. injections were performed with 106 or 107 PFU/6 ␮l of either an IL-4- expressing (Ad5E1 mIL-4) or an empty control (Ad5del70-3) recombinant human type 5 adenovirus vector or with saline. Seven days after the i.a. Histology injection of the viral vector, mice were sacrificed by cervical dislocation Mice were sacrificed by cervical dislocation. Thereafter, whole knee joints and the skin of the knee joint was removed. The appearance of arthritis in were removed and fixed for 4 days in 10% formalin. After decalcification the injected joints was assessed and severity score was recorded as previ- in 5% formic acid, the specimens were processed for paraffin embedding ously described (20). Thereafter, knee joints were isolated and processed (31). Tissue sections (7 ␮m) were stained with hematoxylin and eosin for light microscopy. (H&E) or Safranin O. Histopathological changes were scored using the Assessment of arthritis following parameters. Infiltration of cells was scored on a scale of 0–3, depending on the amount of inflammatory cells in the synovial cavity (ex- Mice were considered to have arthritis when significant changes in redness udate) and synovial tissue (infiltrate). Proteoglycan depletion was deter- and/or swelling were noted in the digits or in other parts of the paws. Knee mined using Safranin O staining. The loss of proteoglycans was scored on joint inflammation was scored visually after skin dissection, using a scale a scale of 0–3, ranging from fully stained cartilage to destained cartilage or of noninflamed (0), mild (1), marked (1.5), or severe (2) inflammation. complete loss of articular cartilage. A characteristic parameter in CIA is the 4548 CARTILAGE PROTECTION BY LOCAL IL-4 IN EXPERIMENTAL ARTHRITIS progressive loss of articular cartilage. This destruction was graded sepa- rately on a scale of 0–3, ranging from the appearance of dead chondrocyte (empty lacunae) to complete loss of the articular cartilage. Histopatholog- ical changes in the knee joints were scored in the patella/femur region on five semiserial sections of the joint, spaced 70 ␮m apart. Scoring was performed by two observers without knowledge of the experimental group, as described earlier (20). For autoradiographic analysis, radiolabeled sulfate (50 ␮Ci/mouse) was injected i.p. 2 h before dissection of the stifle joints. 35S-sulfate was in- corporated into the cartilage layer, but not the underlying subchondral bone, and its incorporation reflects newly synthesized proteoglycans (32). Sections (7 ␮m) were mounted on gelatin-coated slides, which were im- mersed in K5 emulsion (Ilford; Basildon, Essex, U.K.) and exposed for several weeks before being developed and stained with H&E. Assessment of chondrocyte proteoglycan synthesis Patellae (10 pieces), in a minimal amount of adjoining soft tissue, were placed in 2 ml RPMI 1640 medium with gentamicin (50 ␮g/ml), L-glu- ␮ 35 tamine (2 mM), and 20 Ci [ SO4]sulfate. At the end of the 3-h incubation period, patellae were washed in saline three times, fixed in 4% formalin, and subsequently decalcified in formic acid (5%) for 4 h. Patellae were punched out of the adjacent tissue, and dissolved in 0.5 ml Luma solve Downloaded from (Omnilabo, Breda, The Netherlands). The 35S-sulfate content of each pa- tella was measured by liquid scintillation counting and expressed as cpm. NIMP-R14 staining Influx of PMNs was analyzed on knee joint sections. Briefly, sections were deparaffinized, and preincubated for 15 min by RT with 20% normal rabbit serum. Thereafter, sections were incubated with NIMP-R14 (25–30 kDa http://www.jimmunol.org/ epitope mainly present on neutrophils) Ab for 1 h. After incubation with the second rabbit anti-rat peroxidase Ab for 30 min, sections were incu- bated with AEC substrate in the dark by 37°C for 10 min. Thereafter, sections were stained with hematoxylin for 30 s. As a negative control, sections were incubated with normal rat Ig instead of NIMP-R14 Abs. VDIPEN analysis FIGURE 1. Adenoviral vector-mediated IL-4 expression in the mouse VDIPEN expression was analyzed as described (33). Briefly, sections were knee joint of naive (A) and CII-immunized (B) mice. A total of 1.106 or deparaffinized, rehydrated, and digested with chondroitinase ABC (0.25 7 U/ml 0.1 M Tris-HCl, pH 8) for1hat37°C to remove chondroitin sulfate 1.10 PFU of Ad5E1 mIL-4 was injected before onset of CIA was noted. Seven days after the i.a. injection of the viral vector, patellae with adjacent by guest on September 25, 2021 from the proteoglycans. Sections were then treated with 1% H2O2 in meth- anol for 20 min, and subsequently with 0.1% Triton in PBS for 5 min. After synovium were isolated in a standardized manner from knee joints and incubation with 1.5% normal goat serum for 20 min, sections were incu- cultured for1hin200␮l RPMI medium at RT; thereafter, the culture bated overnight at 4°C with affinity-purified anti-VDIPEN IgG. This Ab supernatants were assayed for IL-4 by ELISA. The same doses of the recognizes the neoepitope specifically and does not recognize intact pro- Ad5del70-3 control vector i.a. injected in the left knee gave rise to unde- teoglycan (33). Subsequently, sections were incubated with biotinylated tectable levels of IL-4 (not graphed). No IL-4 levels were found in the goat anti-rabbit IgG and stained with avidin-peroxidase (Elite kit; Vector, circulation using 1.106 and 1.107 PFU doses of the control or IL-4-ex- Burlingame, CA). Development of the peroxidase product was done using pressing vector (data not shown). Results are the mean Ϯ SD of three mice nickel enhancement to increase sensitivity. Counterstaining was done using orange G (2%) for 5 min. As a negative control, sections were incubated per dose. The detection limit of the IL-4 ELISA is 5–10 pg/ml. with normal rabbit IgG instead of anti-VDIPEN Abs. Assessment of nitrite levels in 48-h patella washouts Ϫ Patellae with adjacent synovium were isolated in a standardized manner The concentration of NO2 (a stable breakdown product of NO) was de- from knee joints, as previously described (28), and incubated in RPMI termined by Griess reaction using NaNO2 standards (Merck, Darmstadt, 1640 medium with 0.1% BSA, gentamicin (50 ␮g/ml), and L-glutamine (2 Germany) in RPMI 1640 tissue culture medium with 0.1% BSA. Briefly, mM) (200 ␮l/patella) for 24 h at 37°C. Thereafter, supernatant was har- 100 ␮l of conditioned medium of 24 and 48 h was mixed with 100 ␮lof vested and new medium was added and incubated for another 24 h at 37°C. Griess reagent (0.1% naphthylethylenediamine dihydrochloride (Sigma, St.

Table I. Biological activity of IL-4 by the Ad5E1mIL-4 vectora

Naive Mice Rabbit Igsb Naive Mice Anti-IL-4 Treatmentb Dose (pfu) Infiltratec Exudatec Infiltrate

Ad5del70-3 1.107 0 000 (n ϭ 8) (n ϭ 8) Ad5E1mIL-4 1.107 0.8 Ϯ 0.1 0.3 Ϯ 0.1 0.2 Ϯ 0.1 0.1 Ϯ 0.1 (n ϭ 8) (n ϭ 8)

a Ad5del70-3 or Ad5E1mIL-4 were given i.a. in the right knee joint of naive mice. Seven days later, mice were sacrificed by cervical dislocation and the knee joints were taken for histology. b Rabbit anti-mouse IL-4 Ab (0.5 ml) were given i.p. at days 0, 3, and 6 to verify biological activity of IL-4. As a control, the same dose of rabbit Ig was used. c Infiltration of cells was scored on a scale of 0–3, depending on the amount of inflammatory cells in the synovial cavity (exudate) and synovial tissue (infiltrate). Results are the mean Ϯ SD per group. The Journal of Immunology 4549 Downloaded from

FIGURE 4. Histological analysis of chondrocyte death (A) and cartilage erosion (B) after i.a. administration of 1.107 PFU of Ad5E1 mIL-4 during http://www.jimmunol.org/ FIGURE 2. Effects of Ad5E1 mIL-4 in the mouse knee joint with CII onset of CIA. For more detail, see Fig. 3. Chondrocyte death and cartilage p Ͻ 0.001 vs control group, by ,ء .arthritis. Immunized DBA-1 mice were i.a. injected in the right knee joint erosion were scored on a scale of 0–3 with 1.106 or 1.107 PFU of Ad5E1 mIL-4 or Ad5del70-3, before onset of Mann-Whitney rank sum test. CIA was noted. Seven days after i.a. injection of the viral vector, mice were sacrificed by cervical dislocation and the skin of the knee joint was re- moved. The appearance of arthritis in the injected joints was assessed (A) Statistical analysis and arthritis was scored for severity (B). Results are the mean Ϯ SD of two experiments; in the first experiment, both doses of 1.107 and 1.106 PFU Differences between experimental groups were tested using the Mann- Whitney rank sum test, unless stated otherwise. were tested with at least eight mice per group, and in a second experiment by guest on September 25, 2021 p ϭ ,ء .only the 1.107 PFU dose was given, with at least 12 mice per group 0.006 vs control group, by Mann-Whitney rank sum test. Results Adenoviral vector-mediated IL-4 expression in the knee joint Different doses of Ad5E1 mIL-4 were intraarticularly injected into the right knee joint of naive and CII-immunized DBA-1 mice, and IL-4 protein levels were measured at day 7 in washouts of joint Louis, MO), 1/1 diluted with 1% sulfanilamide (Sigma) in 5% H PO4) in 3 tissue. Considerable IL-4 levels were found after a single injection a flat-bottom microtiter plate (Costar, Cambridge, MA), and the OD at 545 7 nm was measured using an ELISA plate reader (Titertek Multiscan of 10 PFU (Fig. 1), whereas roughly a 10-fold lower IL-4 level MCC/340). was seen after injection of 106 PFU. No major differences were

FIGURE 3. Histological analysis of joint pathology after i.a. administration of Ad5E1 mIL-4 during onset of CIA. Immunized DBA-1 mice were i.a. injected in the right knee with 1.106 or 1.107 PFU of either Ad5E1 mIL-4 or Ad5del70-3 before onset of CIA was noted. Seven days after the i.a. injection of the viral vector, mice were sacrificed by cervical dislocation and the knee joints were taken for histology. Synovial infiltrates, (A), and p ϭ 0.002; #, p ϭ 0.012 vs control group, by Mann-Whitney rank ,ء .proteoglycan depletion (B) were scored on a scale of 0–3. For more detail, see Fig. 2 sum test. 4550 CARTILAGE PROTECTION BY LOCAL IL-4 IN EXPERIMENTAL ARTHRITIS Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 5. Protective effect of local IL-4 treatment on chondrocyte death and cartilage destruction in CIA. A, C, E, and G, Arthritic knee joint of a mouse 7 days after i.a. administration of 1.107 PFU of Ad5del70-3 control vector. Note the infiltrate and chondrocyte death (A ϩ C) (arrows), and severe cartilage surface disruption (E ϩ G) (arrows). B, D, F, and H, Knee joint of a mouse 7 days after i.a. injection of 1.107 PFU of Ad5E1 mIL-4. Note the enhanced infiltrate and almost complete prevention of chondrocyte death (B ϩ D) and cartilage damage (F ϩ H). P, patella; F, femur; S, synovium; JS, joint space; C, cartilage. A, B, E, and F, Original magnification ϫ100; C, D, G, and H, original magnification ϫ400. H&E was used in A–D, and Safranin O staining in E–H.

found between IL-4 levels in naive and CII-immunized mice. No Local overexpression of IL-4 in the knee joint of naive mice detectable IL-4 was noted in washouts of normal or knee gave rise to local cell influx at day 7. No increase in cell mass was joints injected with a control vector (Ad5del70-3). noted using the same dose of Ad5del70-3. To verify biological The Journal of Immunology 4551

Table II. Classifications of individual mice, with respect to degree of chondrocyte death and cartilage destruction after Ad5E1mIL-4 treatmenta

Chondrocyte Deathb Cartilage Destructionb

AdIL-4 AdControl AdIL-4 AdControl AdIL-4 AdControl AdIL-4 AdControl (106 pfu) (106 pfu) (107 pfu) (107 pfu) (106 pfu) (106 pfu) (107 pfu) (107 pfu)

Arthritis 9/9 4/8 21/21 20/22 9/9 4/8 21/21 20/22 Non003000 70 Mild 3 0 10 0 3 0 14 0 Marked 6 1 8 8 6 1 0 15 Severe 0 3 0 12 0 3 0 5

a See Fig. 2 for more detail on the experimental protocol. b Arthritic mice were classified according to their degree of chondrocyte death and cartilage erosion, as scored on histologic sections. The ␹2 test showed significance ( p Ͻ 0.01) for chondrocyte and cartilage destruction between the 1.107 pfu of Ad5E1mIL-4 and Ad5del70-3 treatment comparing non/mild and marked/severe groups. AdIL-4 ϭ Ad5E1mIL-4; AdControl ϭ Ad5del70-3. activity of IL-4 expressed by Ad5E1 mIL-4, blocking experiments line injection (Fig. 2). Of note, a single injection of 107 adenovirus with anti-IL-4 were performed, resulting in prevention of cell in- did not induce joint inflammation in a knee joint of naive mice. flux (Table I). This indicates that adenovirus itself can accelerate expression of collagen arthritis, whereas this expression is further increased Downloaded from Local IL-4 aggravates expression of collagen arthritis when IL-4 is overexpressed. DBA-1 mice were immunized with CII, and shortly before ex- Apart from the incidence, we also analyzed the severity of the pected onset of collagen arthritis (day 25) a single injection of arthritis. As seen in Fig. 2B, severity of arthritis is also markedly Ad5E1 mIL-4, control vector, or saline was given in the right knee enhanced, even with the low dose of 106 PFU Ad5E1 mIL-4. Fur- joint. Seven days later, a 100% arthritis incidence was noted in the thermore, histologic analysis showed more inflammatory infiltrate in right knee joints of the Ad5E1 mIL-4 groups, both after injection the synovial tissue and exudate in the synovial cavity in the low dose http://www.jimmunol.org/ of 107 or 106 PFU. In contrast, 89% and 50% incidence was seen IL-4 group as compared with the control vector group (Fig. 3A). in joints injected with 107 or 106 PFU control vectors, respectively. The latter was not different from the incidence observed after sa- Local IL-4 overexpression prevents chondrocyte death and cartilage erosion Apart form the analysis of inflammatory aspects, the histologic knee joint sections were stained for proteoglycan content in the articular cartilage. Furthermore, semiserial sections were scored for the degree of chondrocyte death and cartilage surface erosions by guest on September 25, 2021 in the patella and femur region. Profound proteoglycan depletion was found in the control, arthritic group, and the depletion was not different in the IL-4 groups (Fig. 3B). However, despite the pronounced inflammation in the IL-4 groups, the degree of chon- drocyte death and cartilage erosion was highly reduced (Fig. 4). A second experiment is shown in Fig. 4, in which the effect on the chondrocyte death and cartilage damage of a single injection of 107 PFU is shown. As in the first experiment, 100% arthritis in- cidence was noted in the right knee joint of the IL-4 group with an arthritis score of 2 Ϯ 0.2 compared with 91% incidence in the adenoviral control group with an arthritis score of 1.8 Ϯ 0.6. Again a strong reduction of the degree of chondrocyte death and cartilage erosion was found in the IL-4 group. Mean values were 61% and 77% reduced for these parameters in the high dose IL-4 group as compared with its respective adenoviral control group (Figs. 4 and 5). In addition, the mice were ranked for individual scores of chon- drocyte death and cartilage erosion in four subclasses (Table II). Whereas the majority of mice in the high dose control group were ranked as having severe chondrocyte death, none of the mice of the IL-4 group was in this catagory, whereas 13 of 21 were ranked as having no or mild chondrocyte death. When cartilage erosion was scored, the differences between IL-4 and control were even more impressive. All mice in the control group were ranked as moderate FIGURE 6. Prevention of VDIPEN neoepitope expression by local IL-4 or severe, whereas all mice in the IL-4 group were nonerosive or in CIA. A, Arthritic knee joint of a mouse 7 days after i.a. injection of 1.107 mildly erosive. PFU of Ad5del70-3 control vector showing moderate erosion. Note the 6 expression of the VDIPEN neoepitope in the femur and patella. B, Knee Although the number of mice treated with the low dose (10 ϭ joint of a CII-immunized DBA-1 mouse 7 days after i.a. injection of 1.107 PFU) AdIL-4 is limited (n 9), a similar trend of protection PFU of Ad5E1 mIL-4. Note the reduction of VDIPEN neoepitope expres- against severe damage was noted. Of the four of eight mice show- sion in this group. P, patella; F, femur; JS, joint space. Original magnifi- ing arthritis in the control group, three were classified with severe cation ϫ100. chondrocyte death and erosion (Table II). None of the nine arthritic 4552 CARTILAGE PROTECTION BY LOCAL IL-4 IN EXPERIMENTAL ARTHRITIS

Table III. Synovial and cartilage mRNA levels after Ad5E1mIL-4 treatmenta

PCR Cyclesb

Synovium Cartilage

Ad5del70-3c Ad5E1mIL-4c Œd Ad5del70-3 Ad5E1mIL-4 Œ

IL-1␤ 26 31 Ϫ525 28 Ϫ3 TNF-␣ 28 32 Ϫ428 28 0 IL-1Ra 26 27 Ϫ126 26 0 iNOS 32 33 Ϫ134 35 Ϫ1 TIMP-1 21 21 0 24 22 ϩ2 MMP-3 21 26 Ϫ524 24 0 MCP-1 26 26 0 ND ND ND MIP-2 30 33 Ϫ3NDNDND

a Synovial and cartilage mRNA levels were determined by RT-PCR technology 7 days after the viral injection. Total RNA from a pool of 10 cartilage samples or 6 synovium biopsies from a particular group were pooled, as described in Materials and Methods. Values are the mean of two experiments. The PCR measurements of a particular mediator were routinely repeated three times. The variations in the two repeated experiments never exceeded more than two cycles. b Number of PCR cycles, corrected for GAPDH expression, in which gene product of interest was first detectable (ND ϭ not done). c Immunized DBA-1 mice were i.a. injected in the right knee joint with 1.107 pfu of Ad5E1mIL-4 or Ad5del70-3 before onset of CIA was noted. mRNA expression was determined 7 days after the viral injection. d Œ refers to the difference between the respective Ad5del70-3 and Ad5E1mIL-4 treated conditions. Downloaded from

mice of the low dose IL-4 group ended up in this category, again proving that IL-4 provides protection against progression to major erosion.

Local IL-4 overexpression prevents VDIPEN neoepitope http://www.jimmunol.org/ expression VDIPEN neoepitope is a marker of metalloproteinase (MMP)-me- diated cleavage of aggrecan, the major proteoglycan of articular cartilage, and earlier studies revealed that expression mainly oc- curred at sites and stages of advanced damage (33–35). To further demonstrate the protective effect of local IL-4, sections were stained for this neoepitope. As a typical example, VDIPEN was

highly expressed at the edges of the pattela and throughout the by guest on September 25, 2021 femoral cartilage in sections of control, arthritic mice with mod- erate erosions (Fig. 6). In contrast, VDIPEN staining was markedly reduced in the IL-4 group and only present at the margins of the cartilage. In line with the reduced VDIPEN expression, we analyzed the mRNA expression of stromelysin, a major MMP able to induce the neoepitope cleavage site. RT-PCR measurements revealed a strongly reduced level of stromelysin in the synovial tissue (Table III). TIMP-1 was equally expressed in the synovium of control and

FIGURE 7. Effects of Ad5E1 mIL-4 in the mouse knee joint with CII arthritis on the chondrocyte proteoglycan synthesis. A separate set of im- FIGURE 8. Effects of local IL-4 overexpression on the metabolic ac- munized DBA-1 mice was i.a. injected in the right and left knee joint on tivity of the chondrocytes. Autoradiography of the right (injected with day 26 with 1.107 PFU of Ad5E1 mIL-4 or Ad5del70-3. Seven days later, 1.107 PFU of Ad5del70-3) (A) and left (injected with 1.107 PFU of Ad5E1 mice were sacrificed by cervical dislocation. 35S-sulfate was given 2 h mIL-4) (B) knee joint of the same mouse with CIA, 7 days after i.a. in- before sacrifice. The appearance of arthritis in the injected joints was as- jection of the viral vector. 35S-sulfate labeling for2hinvivo (see Fig. 7). sessed and arthritis score was given (A). In addition, proteoglycan synthesis Black spots represent 35S-sulfate incorporation. P, patella; F, femur; JS, of chondrocytes in patellar cartilage was measured by 35S-sulfate incorpo- joint space. Original magnification ϫ100. Note the absence of grains in ,patella and femur, in particular in the Ad5del70-3 control vector group ,ء .ration ex vivo (B). Results are the mean Ϯ SD of 10 patellae per group p ϭ 0.006 vs control group, by Mann-Whitney rank sum test. reflecting strong inhibition of chondrocyte proteoglycan synthesis. The Journal of Immunology 4553 Downloaded from http://www.jimmunol.org/

FIGURE 9. Local IL-4 prevents influx of granulocytes in the synovial joint space. A and C, Arthritic knee joint of a mouse 7 days after i.a. administration of 1.107 PFU of Ad5del70-3 control vector. Note the large numbers of granulocytes in the exudate and adhered to the cartilage layer (arrows). B and D, Knee joint of a mouse 7 days after i.a. injection of 1.107 PFU of Ad5E1 mIL-4. Note the marked prevention of granulocytes in the synovial joint space. F, femur; JS, joint space; C, cartilage. A and B, original magnification ϫ200; C and D, original magnification ϫ1000. H&E was used in A and B, and immunolocalization of NIMP-R14 in C and D.

IL-4 group, implying down-regulation of the MMP-3/TIMP-1 bal- despite the fact that joint inflammation in these mice was undi- by guest on September 25, 2021 ance in the synovium of the IL-4-treated mice. Up-regulation of minished. Because we noted above that local erosions and chon- TIMP-1 was noted in the cartilage of the IL-4 group. drocyte death were a major phenomenon in the control group, we also performed autoradiography on whole joint sections, after 35S- Effects of local IL-4 overexpression on chondrocyte metabolic sulfate labeling in vivo. This showed variable but low metabolic function activity of the chondrocytes in the control group. The IL-4 treat- The above studies revealed that IL-4 did prevent chondrocyte ment induced a general increase in chondrocyte proteoglycan syn- death. To investigate whether these chondrocytes are still meta- thesis in chondrocytes throughout the whole cartilage (Fig. 8). bolically active and produce proteoglycans, we measured 35S-sul- fate incorporation in the whole patellar cartilage, ex vivo. As shown in Fig. 7, the 35S-proteoglycan synthesis is markedly en- Local IL-4 overexpression reduces influx of granulocytes in the hanced in the IL-4 group as compared with the arthritic control, synovial joint space Histologic analysis revealed differences in nature of the synovial Table IV. Adenoviral vector-mediated IL-4 expression in the mouse infiltrate between the IL-4 group compared with the control vector knee joint suppresses local IL-1␤ and NO levelsa group. In the control arthritic joint, numerous granulocytes ad- hered to the cartilage, whereas this was absent in the IL-4-treated Arthritis Score IL-1␤ (pg/ml)b NO (␮M)c joint (Fig. 9, A and B). Moreover, local IL-4 overexpression mark- Ad5del70-3 2.0 Ϯ 0.06 309 Ϯ 42 18.0 Ϯ 3.8 edly diminished the influx of PMNs in the synovial joint space. Ad5E1mIL-4 2.0 Ϯ 0.05 75 Ϯ 21* 9.7 Ϯ 3.7† This was confirmed by immunolocalization studies on knee joint

a Immunized DBA-1 mice with a booster injection on day 22 were i.a. injected in sections using NIMP-R14 staining (Fig. 9, C and D) and dimin- the right and left knee joint on day 26 with 1.107 pfu of Ad5E1mIL-4 or Ad5del70-3 ished mRNA expression of macrophage-inflammatory protein-2 in before onset of CIA was noted. Seven days after the i.a. injection of the viral vector, the synovial tissue (Table III). patella washouts were taken and assayed for IL-1␤ and nitric oxide levels. Note the marked suppression of IL-1␤ and NO, despite the severe inflammation. b Patellae with adjacent synovium were isolated in a standardized manner from knee joints and cultured for1hin200␮l RPMI 1640 medium at room temperature; Down-regulation of IL-1 and NO synthesis by local IL-4 thereafter, IL-1␤ levels were measured in these culture supernatants using an IL-1␤ ELISA. Results are the mean Ϯ SD of six patella washouts per group. Previously, we have shown the pivotal role of IL-1 in cartilage c Patellae with adjacent synovium were isolated in a standardized manner from knee joints and cultured for 24 h in 200 ␮l RPMI 1640 medium at 37°C; thereafter, destruction in CIA (36–38). In addition, IL-1 is a key mediator in medium was changed and the samples were incubated for another 24 h. NO levels the regulation of the expression of MMP-induced neoepitope (33). were measured using Griess reagents. The results are the mean Ϯ SD of the total nitric oxide levels of the first and the second incubation period. Furthermore, it has been shown that IL-1 together with NO played *, p ϭ 0.002; †, p ϭ 0.005 vs control group, by Mann-Whitney rank sum test. an important role in the inhibition of the chondrocyte metabolic 4554 CARTILAGE PROTECTION BY LOCAL IL-4 IN EXPERIMENTAL ARTHRITIS activity (39, 40). Furthermore, it has been shown that IL-4 up- expression was mainly found at particular sites in the articular regulates IL-1Ra, the receptor antagonist of IL-1 (9, 19). We there- cartilage, associated with severe proteoglycan depletion and oc- fore examined the effects of Ad5E1 mIL-4 on the local IL-1 ex- currence of chondrocyte death (J. B. J. Van Meurs, manuscript in pression in the synovium and cartilage and the local NO synthesis preparation). Moreover, although inhibition of proteoglycan deple- production. In addition, we analyzed the effects of local IL-4 on tion was not always achieved with blocking of IL-1 in the various IL-1Ra and the balance IL-Ra/IL-1␤. RT-PCR revealed major arthritis models, IL-1 blocking always fully prevented VDIPEN down-regulation of mRNA expression of IL-1␤ in the synovium expression and late erosions. The present study further supports and also in articular cartilage (Table III). In line with this, pro- this link between erosions and VDIPEN expression. Although IL-4 nounced suppression of IL-1␤ protein levels (76%) in the syno- treatment did not diminish proteoglycan depletion, it did prevent vium was found (Table IV). Although no up-regulation of IL-1Ra VDIPEN expression and surface erosions. It is known that IL-1 is mRNA levels (Table III) or protein levels (AdControl, 3831 Ϯ a potent inducer of stromelysin (MMP-3) and that MMP-3 can 1329 pg/ml vs AdIL-4, 1959 Ϯ 131 pg/ml) was found, the balance induce the VDIPEN epitopes. The present study further revealed between IL-1Ra/IL-1␤ protein levels was slightly enhanced by lo- that MMP-3 mRNA levels were highly reduced by IL-4, compat- cal IL-4 treatment (IL-1Ra/IL-1␤ ratio for AdControl, 12 vs ible with a role of MMP-3 in erosion. It is not yet clear whether the AdIL-4, 26). NO levels in the synovium were also decreased by reduction of IL-1 was indirectly responsible for the lowered Ad5E1 mIL-4 (Table IV). This indicates that the cartilage-protec- MMP-3 levels or whether the IL-4 directly reduced the enzyme. In tive effect of local overexpression of IL-4 in the synovium of mice vitro studies have suggested that IL-4 is capable of direct reduction with collagen arthritis might be due to suppression of local IL-1 of these metalloproteinases (46, 47). production and subsequent reduction of MMP activity. Apart from major reduction in cartilage erosions, we also noted Downloaded from an amelioration of the inhibition of chondrocyte proteoglycan syn- thesis, by direct measurement and autoradiography. Decreased Discussion 35S-proteoglycan synthesis is a hallmark of events in the articular This is the first study to demonstrate the major protection against cartilage during joint inflammation, and IL-1 as well as NO, as a cartilage erosion by local overexpression of IL-4 in the inflamed secundary mediator, were shown by us to be pivotal mediators in

joint. Histologic analysis revealed almost complete protection this process, in vivo (39, 40). The lower IL-1 and NO levels in the http://www.jimmunol.org/ against chondrocyte death and surface erosions of the articular IL-4-treated mice are in line with these previous findings. Despite cartilage, despite sustained, full-blown inflammation. This protec- the recovery of synthetic activity, pronounced net proteoglycan tive effect of local IL-4 was associated with down-regulation of loss was still observed in the cartilage of these mice, suggesting IL-1, TNF, and NO in the synovium, whereas the MMP-3/TIMP-1 that there is overruling breakdown due to the inflammatory pro- imbalance was improved. cess, independent of IL-1 and NO. Of high interest, we found a Previously, we showed the pivotal role of IL-1 in cartilage de- similar uncoupling of NO dependence of proteoglycan synthesis struction in murine collagen arthritis (36). Both IL-1␣ and IL-1␤ and breakdown in recent studies in inducible NO synthase-defi- appear to be involved in cartilage destruction in Ag-induced ar- cient mice (40). thritis, whereas dominant involvement of IL-1␤ was observed in The prevention of cartilage damage in CIA was IL-4 dose de- by guest on September 25, 2021 collagen arthritis (37). This autoimmune model is driven by the pendent. Both the high and low dose of IL-4 vector significantly combination of cellular and humoral immunity against cartilage enhanced the expression of collagen arthritis, in line with potent CII and is characterized by rapid and severe erosions of cartilage proinflammatory potential of IL-4 (7). The best protection against and bone (41–44). The onset of CIA is dependent on IL-12, TNF, cartilage erosion was obtained with the high dose, suggesting that and IL-1, and under the stringent control of endogenous IL-10 (45, substantial levels of IL-4 are needed to control erosion. Of interest, 20). The fact that the local IL-4 treatment strongly reduced IL-1 net proteoglycan depletion was not different between the two IL-4 makes it likely that the protection against erosions is at least partly dose groups, underlining uncoupling between depletion and ero- achieved through this pathway. sion and suggesting that different mediators are involved in these It has been shown that IL-4 up-regulates IL-1Ra (9, 19). In the processes. present study, no direct up-regulation of IL-1Ra by local IL-4 Enhancement of synovial inflammatory mass by local IL-4 ex- overexpression was found. One explanation could be that IL-1Ra pression is a significant but not unexpected side effect. IL-4 has levels are high in arthritic tissue, as a feedback reaction to IL-1. been shown to be a chemoattractant for and fibro- When prolonged treatment with IL-4 highly reduces the IL-1 lev- blast and inducer of fibroblast proliferation (48–50). Increased cell els, this will result in significant down-regulation and this will influx and tissue proliferation have been noted after local IL-4 outbalance the mere up-regulation by IL-4. Interestingly, because overexpression in the pancreas, trachea, and liver (51–53). Re- of the marked suppression of IL-1␤, we found that the IL-1Ra/ cently, up-regulation of ␤ integrin, VCAM-1, IL-6, and monocyte- IL-1␤ balance was slightly up-regulated by local IL-4. Another chemotactic protein-1 was reported after exposure of lung fibro- explanation that could be involved is the fact that local IL-4 mark- blast to IL-4 (54). edly prevents cell influx of granulocytes into the joint space. It is The inflammation of collagen arthritis is characterized by a known that PMNs are important producers of IL-1Ra. The strong florid exudate in the joint space, containing numerous amounts of reduction of granulocytes in the exudate by local IL-4 may play an granulocytes, and a progressive destruction of the articular carti- important role in the explanation that IL-1Ra is not up-regulated. lage. Apart from the high numbers of granulocytes in the joint Protection against erosion was also substantiated by lack of cavity, the synovial tissue contains large numbers of macrophages VDIPEN neoepitope expression in the cartilage of the IL-4-treated and lymphocytes, but also in this compartment granulocytes are mice. Breakdown of aggrecan, the major proteoglycan of articular prominent in the first 2 wk after onset. The most characteristic cartilage, predominantly occurs through two enzymatic pathways: feature of collagen arthritis is the aggressive attack of the inflam- aggrecanase attack, resulting in NITEGE neoepitopes, and MMP- matory process at the articular cartilage. In this model, heavy stick- mediated degradation, resulting in VDIPEN neoepitopes. Previous ing of granulocytes at the cartilage surface is a common finding. work from our group revealed that NITEGE expression is an early Granulocytes may play an active role in the cartilage destruction, phenomenon in proteoglycan depletion and that VDIPEN epitope linked to sticking to anti-CII immune complexes in the surface The Journal of Immunology 4555 layers. Igs adherent to cartilage surfaces have been identified in Randomized double-blind comparison of chimeric monoclonal antibody to tumor ␣ rheumatoid joints (55, 56), which can provide an anchorage and necrosis factor (cA2) versus placebo in rheumatoid arthritis. Lancet 344:1105. 6. Rankin, E. C. C., E. H. S. Choy, D. Kassimos, G. H. Kingsley, A. M. Sopwith, trigger for granulocyte activation (57). Reactive oxygen species D. A. Isenberg, and G. S. Panayi. 1995. The therapeutic effects of an engineered and proteolytic enzymes present in the PMNs can be released di- human anti- ␣ antibody (CDP571) in rheumatoid arthritis. rectly into the surface of the cartilage, thereby escaping inhibitors Br. J. Rheumatol. 34:334. 7. Chomarat, P., and J. Banchereau. 1997. An update on interleukin-4 and its re- present in the synovial fluid (58). PMNs need this close contact to ceptor. Eur. Cytokine Netw. 8:333. cartilage to inflict cartilage damage (59–61). In the present study, 8. Miossec, P., J. Briolay, J. Dechanet, J. Wijdenes, H. Martinez-Valdez, and we found numerous granulocytes adhered to the cartilage layer in J. Banchereau. 1992. Inhibition of the production of proinflammatory cytokines and immunoglobulins by interleukin-4 in an ex vivo model of rheumatoid syno- the control arthritis joint, whereas this was absent in the IL-4- vitis. Arthritis Rheum. 35:874. treated joint. This indicates that the marked reduction of granulo- 9. Chomarat, P., E. Vannier, J. Dechanet, M. C. Rissoan, J. Banchereau, cytes (PMNs) in the synovial joint space by local IL-4 may play an C. A. Dinarello, and P. Miossec. 1995. Balance of IL-1 receptor antagonist/IL-1␤ in rheumatoid synovium and its regulation by IL-4 and IL-10. J. Immunol. 154: important role in the prevention of cartilage destruction. 1432. The above paragraphs discussed various pathways of cartilage 10. Colotta, F., F. Re, M. Muzio, R. Bertini, N. Polentarutti, M. Sironi, S. K. Dower, destruction in collagen arthritis. Apart from IL-1-mediated regu- J. E. Sims, and A. Mantovani. 1993. Interleukin-1 type II receptor: a decoy target for IL-1 that is regulated by IL-4. Science 261:472. lation of granulocyte influx, IL-1 is crucial in induction of latent 11. Cope, A. P., D. L. Gibbons, and D. Aderka. 1993. Differential regulation of tumor enzymes in the articular cartilage, still needing further activation necrosis factor receptors (TNF-R) by IL-4: up-regulation of p55 and p75 TNF-R by other enzymes. Granulocytes may provide elastase, which is a on synovial joint mononuclear cells. Cytokine 3:205. 12. O’Garra, A. 1998. Cytokines induce the development of functionally heteroge- potent activator of latent stromelysin. IL-4 can interfere with car- nous T helper cell subsets. Immunity 8:275. tilage destruction at multiple levels. It reduces IL-1, it reduces 13. Miossec, P., and W. B. van den Berg. 1997. Th1/Th2 cytokine balance in arthritis. Downloaded from stromelysin, and it prevents activation of latent stromelysin, Arthritis Rheum. 40:2105. 14. Yeh, L.-A., A. J. Augustine, P. Lee, L. S. Riviere, and A. Sheldon. 1995. Inter- through reduction of granulocyte influx and adherence to the car- leukin-4, an inhibitor of cartilage breakdown in bovine articular cartilage ex- tilage surface. Although IL-4 enhances the inflammatory mass in plants. J. Rheumatol. 22:1740. the joint, cells do not show aggressive behavior. It is intriguing to 15. Shingu, M., S. Miyauchi, Y. Nagai, C. Yasutake, and K. Horie. 1995. The role of IL-4 and IL-6 in IL-1-dependent cartilage matrix degradation. Br. J. Rheumatol. note that joint inflammation in various patient groups can show a 34:101. destructive and nondestructive phenotype. It is tempting to spec- 16. Van Roon, J. A. G., J. L. A. M. van Roy, F. H. J. Gmelig-Meyling, http://www.jimmunol.org/ ulate that relative levels of IL-4 might make a difference. F. P. J. G. Lafeber, and J. W. J. Bijlsma. 1996. Prevention and reversal of car- tilage degradation in rheumatoid arthritis by interleukin-10 and interleukin-4. It is known that IL-4 has a suppressive effect on Th1 activity and Arthritis Rheum. 39:829. is a crucial factor in differentiation of naive T cells to the Th2 17. Cawston, T. E., A. J. Ellis, H. Bigg, V. Curry, E. Lean, and D. Ward. 1996. phenotype (12). In the present study, no substantial IL-10 expres- Interleukin-4 blocks the release of collagen fragments from bovine nasal cartilage treated with cytokines. Biochim. Biophys. Acta 1314:226. sion was found in tissue washouts of the IL-4-treated group (data 18. Odeh, M. 1997. New insights into the pathogenesis and treatment of rheumatoid not shown). In contrast, preliminary studies on local IL-12 mRNA arthritis. Clin. Immunol. Immunopathol. 83:103. expression in the synovial tissue suggested marked reduction of 19. Allen, J. B., H. L. Wong, G. L. Costa, M. J. Bienkowski, and S. M. Wahl. 1993. Suppression of monocyte function and differential regulation of IL-1 and IL-1ra this cytokine, implying that lower Th1 rather than enhanced Th2 by IL-4 contribute to resolution of experimental arthritis. J. Immunol. 151:4344. activity contributes to the net protection. Changes in activity of 20. Joosten, L. A. B., E. Lubberts, P. Durez, M. M. A. Helsen, M. J. M. Jacobs, by guest on September 25, 2021 CII-specific T cells in such an infiltrate of IL-4-treated mice are at M. Goldman, and W. B. van den Berg. 1997. Role of interleukin-4 and interleu- kin-10 in murine collagen-induced arthritis: protective effect of interleukin-4 and present under investigation. interleukin-10 treatment on cartilage destruction. Arthritis Rheum. 40:249. In conclusion, this study represents the first demonstration of 21. Roessler, B. J., E. D. Allen, J. M. Wilson, J. W. Hartman, and B. L. Davidson. cartilage-protective effects of local IL-4 gene therapy in experi- 1993. Adenoviral-mediated gene transfer to rabbit synovium in vivo. J. Clin. Invest. 92:1085. mental arthritis, despite ongoing inflammation. It furthermore il- 22. Roessler, B. J., J. W. Hartman, D. K. Vallance, J. M. Latta, S. L. Janich, and lustrates the feasibility of gene transfer. Apart from overexpression B. L. Davidson. 1995. Inhibition of interleukin-1 induced effects in synoviocytes of inhibitors, probably asking for high and prolonged expression transduced with the human IL-1 receptor antagonist cDNA using an adenoviral vector. Hum. Gene Ther. 6:307. levels (38), the introduction of modulators such as IL-4 may pro- 23. Nita, I., S. C. Ghivizzani, J. Galea-Lauri, G. Bandara, H. I. Georgescu, vide a promising alternative to treat destructive arthritis. Our find- P. D. Robbins, and C. H. Evans. 1996. Direct gene delivery to synovium: an ings also underline the often suggested, but hardly proven concept evaluation of potential vectors in vitro and in vivo. Arthritis Rheum. 39:820. 24. Sawchuk, S. J., G. P. Boivin, L. E. Duwel, W. Ball, K. Bove, B. Trapnell, and that the balance of destructive and protective mediators determines R. Hirsch. 1996. Anti-T cell receptor monoclonal antibody prolongs transgene the relative erosive nature of a given arthritis, rather than the bulk expression following adenovirus-mediated in vivo gene transfer to mouse syno- of the inflammatory mass. The main clinical problem of chronic vium. Hum. Gene Ther. 7:499. 25. Ghivizzani, S. C., E. R. Lechman, R. Kang, C. Tio, J. Kolls, C. H. Evans, and arthritides is the destruction of cartilage and bone. Our data make P. D. Robbins. 1998. Direct adenovirus-mediated gene transfer of interleukin 1 it clear that IL-4 is a promising regulator of these elements. and tumor necrosis factor ␣ soluble receptors to rabbit knees with experimental arthritis has local and distal anti-arthritic effects. Proc. Natl. Acad. Sci. USA 95:4613. Acknowledgments 26. Hogaboam, C. M., B. A. Vallance, A. Kumar, C. L. Addison, F. L. Graham, J. Gauldie, and S. M. Collins. 1997. Therapeutic effects of interleukin-4 gene The Central Animal Laboratory, Faculty of Medicine, University of transfer in experimental inflammatory bowel disease. J. Clin. Invest. 100:2766. Nijmegen is acknowledged for the animal care. 27. Xing, Z., Y. Ohkawara, M. Jordana, F. L. Graham, and J. Gauldie. 1996. Transfer of GM-CSF gene to rat lung induces eosinophilia, monocytosis and fibrotic re- References actions. J. Clin. Invest. 97:1102. 28. Lubberts, E., L. A. B. Joosten, M. M. A. Helsen, and W. B. van den Berg. 1998. 1. Firestein, G. S., J. M. Alvaro-Garcia, and R. Maki. 1990. Quantitative analysis of Regulatory role of interleukin-10 in joint inflammation and cartilage destruction cytokine gene expression in rheumatoid arthritis. J. Immunol. 144:3347. in murine streptococcal cell wall (SCW) arthritis: more therapeutic benefit with 2. Chen, E., E. C. Keystone, and E. N. Fish. 1993. Restricted cytokine expression IL-4/IL-10 combination therapy than with IL-10 treatment alone. Cytokine 10: in rheumatoid arthritis. Arthritis Rheum. 36:901. 361. 3. Arend, W. P., and J.-M. Dayer. 1995. Inhibition of the production and effects of 29. Van Meurs, J. B. J., P. L. E. M. van Lent, L. A. B. Joosten, P. M. van der Kraan, interleukin-1 and tumor necrosis factor ␣ in rheumatoid arthritis. Arthritis Rheum. and W. B. van den Berg. 1997. Quantification of mRNA levels in joint capsule 38:151. and articular cartilage of the murine knee joint by RT-PCR. Rheumatol. Int. 4. Feldmann, M., M. J. Elliot, J. N. Woody, and R. N. Maini. 1997. Anti-tumor 16:197. necrosis factor-␣ therapy of rheumatoid arthritis. Adv. Immunol. 64:283. 30. Chomczynski, P., and N. Sacchi. 1987. Single step method of RNA isolation by 5. Elliot, M. J., R. N. Maini, M. Feldmann, J. R. Kalden, C. Antoni, J. Smolen, acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162: B. Leeb, F. C. Breedveld, J. D. Macfarlane, H. Bijl, and J. N. Woody. 1994. 156. 4556 CARTILAGE PROTECTION BY LOCAL IL-4 IN EXPERIMENTAL ARTHRITIS

31. Van den Berg, W. B., L. A. B. Joosten, and L. B. A. van de Putte. 1984. Electrical 45. Joosten, L. A. B., E. Lubberts, M. M. A. Helsen, and W. B. van den Berg. 1997. charge of the antigen determines intraarticular handling and chronicity of arthritis Dual role of IL-12 in early and late stages of murine collagen type II arthritis. in mice. J. Clin. Invest. 74:1850. J. Immunol. 159:4094. 32. Van den Berg, W. B., H. J. van Beusekom, L. B. A. van de Putte, W. A. Zwarts, 46. Lacraz, S., L. Nicod, B. Galve-de Rochemonteix, C. Baumberger, J.-M. Dayer, and M. van der Sluis. 1982. Antigen handling in antigen-induced arthritis in mice: and H. G. Welgus. 1992. Suppression of metalloproteinase biosynthesis in human an autoradiographic and immunofluorescence study using whole joint sections. alveolar macrophages by interleukin-4. J. Clin. Invest. 90:382. Am. J. Pathol. 108:9. 47. Prontera, C., G. Crescenzi, and D. Rotilio. 1996. Inhibition by interleukin-4 of 33. Van Meurs, J. B. J., P. L. E. M. van Lent, I. I. Singer, E. K. Bayne, stromelysin expression in human skin fibroblasts: role of PKC. Exp. Cell Res. F. A. J. van de Loo, and W. B. van den Berg. 1998. Interleukin-1 receptor 224:183. antagonist prevents expression of the metalloproteinase-generated neoepitope 48. Hiester, A. A., D. R. Metcalf, and P. A. Campbell. 1992. Interleukin-4 is che- VDIPEN in antigen-induced arthritis. Arthritis Rheum. 41:647. motactic for mouse macrophages. Cell. Immunol. 139:72. 34. Van Meurs, J. B. J., P. L. E. M. van Lent, A. E. M. Holthuysen, R. Stoop, 49. Postlethwaite, A. E., and J. M. Seyer. 1991. Fibroblast induction by I. Singer, E. K. Bayne, D. Visco, J. S. Mudgett, and W. B. van den Berg. 1998. human recombinant interleukin-4. J. Clin. Invest. 87:2147. Expression of the MMP-induced neoepitope FVDIPEN is linked to severe car- 50. Monroe, J. G., S. Haldar, M. B. Prystowsky, and P. Lammie. 1988. Lymphokine tilage damage: an essential role for stromelysin in antigen-induced arthritis. Trans regulation of inflammatory processes: interleukin-4 stimulates fibroblast prolif- 44th Orthop. Res. Soc. 23:856 (Abstr.). eration. Clin. Immunol. Immunopathol. 49:292. 35. Singer, I. I., D. W. Kawka, E. K. Bayne, A. A. Donatelli, J. R. Weidner, 51. Mueller, R., T. Krahl, and N. Sarvetnick. 1997. Tissue-specific expression of H. R. Williams, J. M. Ayala, R. A. Mumford, M. W. Lark, T. T. Glant, et al. 1995. interleukin-4 induces extracellular matrix accumulation and extravasation of B VDIPEN, a metalloproteinase-generated neoepitope, is induced and immunolo- cells. Lab. Invest. 76:117. calized in articular cartilage during inflammatory arthritis. J. Clin. Invest. 95: 52. Lukacs, N. W., C. L. Addison, J. Gauldie, F. Graham, K. Simpson, R. M. Strieter, 2178. K. Warmington, S. W. Chensue, and S. L. Kunkel. 1997. Transgene-induced 36. Van den Berg, W. B., L. A. B. Joosten, M. M. A. Helsen, and A. J. J. van de Loo. production of IL-4 alters the development and collagen expression of T helper 1994. Amelioration of established murine collagen-induced arthritis with anti- cell 1-type pulmonary . J. Immunol. 158:4478. IL-1 treatment. Clin. Exp. Immunol. 11:237. 53. David, A., J. Chetritt, H. Coupel-Clauce, L. Tesson, A. Cassard, G. Blancho, 37. Joosten, L. A. B., M. M. A. Helsen, F. A. J. Van de Loo, and W. B. Van den Berg. B. Charreau, J. Sigalla, F. Buzelin, B. Le Mauff, et al. 1997. Adenovirus-medi- 1996. Anticytokine treatment of established type II collagen-induced arthritis in ated gene transfer in rat liver of interleukin 4 but not interleukin 10 produces Downloaded from DBA/1 mice: a comparative study using anti-TNF␣, anti-IL-1␣/␤, and IL-1Ra. severe acute hepatitis. Cytokine 9:818. Arthritis Rheum. 39:797. 54. Doucet, C., D. Brouty-Boye, C. Pottin-Clemenceau, G. W. Canonica, C. Jasmin, 38. Bakker, A. C., L. A. B. Joosten, O. J. Arntz, M. M. A. Helsen, A. M. Bendele, and B. Azzarone. 1998. Interleukin (IL) 4 and IL-13 act on human lung fibro- F. A. J. Van de Loo, and W. B. Van den Berg. 1997. Prevention of murine blasts: implication in asthma. J. Clin. Invest. 101:2129. collagen-induced arthritis in the knee and ipsilateral paw by local expression of 55. Jasin, H. E. 1985. Autoantibody specificities of immune complexes sequestered human interleukin-1 receptor antagonist protein in the knee. Arthritis Rheum. in articular cartilage of patients with rheumatoid arthritis and . Ar- 40:893. thritis Rheum. 28:241.

39. Van de Loo, F. A. J., L. A. B. Joosten, P. L. E. M. van Lent, O. J. Arntz, and 56. Veto, A. A., M. Mannila, E. Zatavain-Rios, and M. H. Weber. 1990. Immune http://www.jimmunol.org/ W. B. van den Berg. 1995. Role of interleukin-1, tumor necrosis factor ␣, and deposits in articular cartilage of patients with rheumatoid arthritis have a granular interleukin-6 in cartilage proteoglycan metabolism and destruction: effect of in pattern not seen in osteoarthritis. Rheumatol. Int. 10:13. situ blocking in murine antigen- and zymosan-induced arthritis. Arthritis Rheum. 57. Ugai, K., H. Ishikawa, K. Hirohata, and H. Shirane. 1983. Interaction of poly- 38:164. morphonuclear leukocytes with immune complexes trapped in rheumatoid artic- 40. Van de Loo, F. A. J., O. J. Arntz, F. H. J. van Enckevort, P. L. E. M. van Lent, ular cartilage. Arthritis Rheum. 26:1434. and W. B. van den Berg. 1998. Reduced cartilage proteoglycan loss during zy- 58. Schalkwijk, J., W. B. Van den Berg, L. A. B. Joosten, and L. B. A. Van de Putte. mosan-induced gonarthritis in NOS2-deficient mice and in anti- 1987. Elastase secreted by activated polymorphonuclear leukocytes causes chon- interleukin-1-treated wild-type mice with unabated joint inflammation. Arthritis drocyte damage and matrix degradation in intact articular cartilage: escape from Rheum. 41:634. inactivation by ␣-1-proteinase inhibitor. Br. J. Exp. Pathol. 68:81. 41. Trentham, D. E., A. S. Townes, and A. H. Kang. 1977. Autoimmunity to type II 59. Campbell, E. J., E. K. Silverman, and M. A. Campbell. 1989. Elastase and ca- collagen: an experimental model of arthritis. J. Exp. Med. 146:857. thepsin G of human monocytes: quantification of cellular content, release in re-

42. Courtenay, J. S., M. J. Dallman, A. B. Dayan, A. Martin, and B. Mosedale. 1980. sponse to stimuli, and heterogeneity in elastase-mediated proteolytic activity. by guest on September 25, 2021 Immunization against heterologous type II collagen-induced arthritis in mice. J. Immunol. 143:2961. Nature 2843:666. 60. Kowanko, I. C., and A. Ferrante. 1996. Adhesion and TNF priming in neutrophil- 43. Stuart, J. M., A. S. Townes, and A. H. Kang. 1982. Nature and specificity of mediated cartilage damage. Clin. Immunol. Immunopathol. 79:36. immune responses to collagen in type II collagen-induced arthritis in mice. 61. Van Lent, P. L. E. M., L. Van den Bersselaar, L. B. A. Van de Putte, and J. Clin. Invest. 69:673. W. B. Van den Berg. 1990. Immobilization aggravates cartilage damage during 44. Myers, L. K., E. F. Rosloniec, M. A. Cremer, and A. H. Kang. 1997. Collagen- antigen-induced arthritis in mice: attachment of polymorphonuclear leukocytes to induced arthritis, an animal model of autoimmunity. Life Sci. 61:1861. articular cartilage. Am. J. Pathol. 136:1407.