Collagen-Induced Arthritis and Comparison and IL-1 Blockade in Α

Home , Marc Feldmann

Evaluation of TNF-α and IL-1 Blockade in Collagen-Induced Arthritis and Comparison with Combined Anti-TNF-α/Anti-CD4 Therapy This information is current as of September 28, 2021. Richard O. Williams, Lilia Marinova-Mutafchieva, Marc Feldmann and Ravinder N. Maini J Immunol 2000; 165:7240-7245; ; doi: 10.4049/jimmunol.165.12.7240 http://www.jimmunol.org/content/165/12/7240 Downloaded from

References This article cites 35 articles, 7 of which you can access for free at:

http://www.jimmunol.org/content/165/12/7240.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 28, 2021

*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 © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Evaluation of TNF-␣ and IL-1 Blockade in Collagen-Induced Arthritis and Comparison with Combined Anti-TNF-␣/ Anti-CD4 Therapy1

Richard O. Williams,2 Lilia Marinova-Mutafchieva, Marc Feldmann, and Ravinder N. Maini

We have evaluated the effects of anti-TNF-␣, anti-IL-1, and combined anti-TNF-␣/anti-CD4 therapy in collagen-induced arthritis. Blockade of TNF-␣ or IL-1 before disease onset delayed, but did not prevent, the induction of arthritis. When treatment was initiated after onset of arthritis, anti-TNF-␣, anti-IL-1␤, and anti-IL-1R (which blocks IL-1␣ and IL-1␤) were all found to be effective in reducing the severity of arthritis, with anti-IL-1R and anti-IL-1␤ showing greater efﬁcacy than anti-TNF-␣. Anti-IL-1␤ was equally as effective as anti-IL-1R, indicating that IL-1␤ plays a more prominent role than IL-1␣ in collagen-induced arthritis. An additive effect was observed between anti-TNF-␣ and anti-IL-1R in the prevention of joint erosion and in normalization of the Downloaded from levels of serum amyloid P. Combined anti-TNF-␣/anti-CD4 therapy also caused normalization of serum amyloid P levels. The therapeutic effect of anti-TNF-␣ plus anti-CD4 was comparable to that of anti-TNF-␣ plus anti-IL-1R, suggesting that combined anti-TNF-␣/anti-CD4 therapy prevents both TNF-␣- and IL-1-mediated pathology. Anti-TNF-␣ treatment reduced IL-1␤ expression in the joint and, conversely, anti-IL-1␤ treatment reduced TNF-␣ expression. Combined anti-TNF-␣/anti-CD4 treatment almost completely blocked the expression of IL-1␤, thereby conﬁrming the ability of this form of combination therapy to prevent ␤

IL-1 -mediated pathology. The Journal of Immunology, 2000, 165: 7240–7245. http://www.jimmunol.org/

lacebo-controlled clinical trials have now firmly estab- Studies of cytokine blockade also point to an important patho- lished that anti-TNF-␣ therapy is effective in reducing dis- logical role for IL-1 in CIA (14–17). However, some doubt as to P ease activity in rheumatoid arthritis (RA)3 (1–5). The de- the pathological significance of IL-1 remains as continuous infu- cision to test anti-TNF-␣ therapy in RA was based on a variety of sion with IL-1R antagonist (IL-1Ra) was found to be relatively diverse findings that, taken together, provided compelling evi- ineffective in adjuvant arthritis (17) and neutralization of IL-1 in dence that TNF-␣ was playing a dominant pathological role in the streptococcal cell wall-induced arthritis failed to affect the clinical disease. For example, it was shown that blockade of TNF-␣ in RA severity of disease, although a reduction in cartilage proteoglycan synovial cell cultures led to reduced production of another impor- depletion was observed microscopically (18). In human RA, clin- by guest on September 28, 2021 tant pro-inflammatory cytokine, IL-1, suggesting that the produc- ical trials with IL-1Ra have shown only modest reductions in dis- tion of IL-1 in RA is driven by TNF-␣ (6). Subsequently, the ease activity (19) although in view of the poor pharmacokinetics, beneficial effects of anti-TNF-␣ therapy were demonstrated by a this may be due to incomplete neutralization of IL-1 (17). number of groups in collagen-induced arthritis (CIA), an animal We report here on a comparative study to evaluate the effect of model of RA (7–11). A further piece of evidence that confirmed neutralizing both IL-1␣ and IL-1␤ (using a mAb that blocks sig- the arthritogenic capacity of TNF-␣ was the observation that TNF- naling via the type I IL-1R), IL-1␤ (using anti-IL-1␤ mAb), or ␣-transgenic mice, which express human TNF-␣ in a disregulated TNF-␣ (using anti-TNF-␣ mAb) in CIA. In addition, we compare fashion, spontaneously develop severe erosive arthritis that can be the effects of two different forms of combination therapy, anti- prevented by anti-human TNF-␣ mAb (12). However, more re- TNF-␣ plus anti-IL-1R and TNF-␣ plus anti-CD4. In this study we cently it was shown that arthritis could also be prevented in TNF- are able to confirm and extend previous findings suggesting that ␣-transgenic mice by the administration of a blocking anti-IL-1R IL-1, like TNF-␣, represents a promising therapeutic target for RA. (type I) mAb, indicating that the induction of arthritis by TNF-␣ in In addition, we provide data indicating that IL-1␤, rather than IL- this model is dependent on IL-1 (13). 1␣, is the major contributor to joint pathology in CIA in DBA/1 mice. Finally, we demonstrate that combined anti-TNF-␣/anti- CD4 therapy leads to effective suppression of both TNF-␣ and ␤ Kennedy Institute of Rheumatology Division, Imperial College School of Medicine, IL-1 expression. London, United Kingdom Received for publication May 19, 2000. Accepted for publication September 11, 2000. Materials and Methods Purification of and immunization with type II collagen The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance Bovine type II collagen was purified from articular cartilage as described with 18 U.S.C. Section 1734 solely to indicate this fact. (20). Male DBA/1 mice (8–12 wk of age) were immunized intradermally 1 This work was supported by the Wellcome Trust, the Arthritis Research Campaign, with type II collagen (200 ␮g/mouse), emulsified in CFA. and Centocor, Inc. 2 Address correspondence and reprint requests to Dr. Richard Williams, Kennedy mAb treatment Institute of Rheumatology Division, Imperial College School of Medicine, 1 Aspenlea Road, London W6 8LH, U.K. E-mail address: [email protected] All mAbs were administered by i.p. injection. Isotype controls for all of the 3 Abbreviations used in this paper: RA, rheumatoid arthritis; CIA, collagen-induced mAbs used in this project have been shown to have no discernable effect arthritis; IL-1Ra, IL-1R antagonist; PIP, proximal interphalangeal; SAP, serum amy- on the progression of arthritis (8, 14, 21, 22). The mAbs that were used are loid P; AUC, area under the curve. described below.

Anti-TNF-␣. TN3-19.12, a neutralizing hamster IgG1 anti-TNF-␣/␤ images (magniﬁcation, ϫ200) per specimen were recorded and quantitative mAb (23) was a gift from Robert Schreiber (Washington University Med- analysis was performed according to color separation. The data are pre- ical School, St. Louis, MO), in conjunction with Celltech (Slough, U.K.). sented as the area within a region of interest covered by positively stained The concentration of TN3-19.12 required to cause 50% inhibition of killing cells, expressed as a percentage of the total area covered by cells. of WEHI 164 cells by 15 pg/ml TNF-␣ is 62.0 ng/ml (24). In vivo administration of TN3-19.12, at doses of 300 ␮g/mouse given once every 3 days, Measurement of anti-type II collagen IgG was earlier shown by us to be effective in reducing the severity of CIA (8). Circulating levels of anti-collagen IgG were measured by solid phase Anti-IL-1R. 35F5 is a rat IgG1 mAb that blocks the binding of IL-1␣ ELISA, as described (8). and IL-1␤ to the mouse type I IL-1R and hence neutralizes IL-1 bioactivity. A concentration of 100 ng/ml 35F5 caused a 50% inhibition of prolifera- Measurement of serum amyloid P (SAP) tion of the D19.G4.1 cell line, stimulated with 12 ng/ml IL-1␣ (25). Pre- treatment with 35F5 at a dose of 200 ␮g completely prevented weight loss Serum levels of SAP were measured by a solid phase ELISA that uses the ϩ and attenuated the acute phase response in mice following challenge with Ca2 -dependent binding of SAP to trinitrophenylated keyhole limpet he- turpentine (26). 35F5 was generated and supplied by Richard Chizzonite mocynanin (31). (Hoffmann-LaRoche, Nutley, NJ). Anti-IL-1␤. 1400.24.17 is a mouse IgG1 mAb that neutralizes IL-1␤ but Results not IL-1␣. The concentration of 1400.24.17 required to induce a 50% in- TNF-␣ and IL-1 blockade before the onset of CIA hibition of proliferation of the D10S cell line, stimulated with 100 pg/ml ␣ IL-1␤, was 50 ng/ml (14). Treatment of DBA/1 mice with 1400.24.17, at To determine the effect of blocking TNF- and/or IL-1 during the doses of 100 ␮g/mouse (three times per week) from the time of collagen induction of arthritis, but before the onset of clinical disease, mice immunization, was earlier shown to produce a marked reduction in the were given two injections on days 14 and 16 after collagen im- severity of CIA (14). 1400.24.17 was generated and supplied by Harry munization with 300 ␮g anti-TNF-␣ alone, 200 ␮g anti-IL-1R Downloaded from Towbin (Novartis Pharmaceuticals, Basel, Switzerland). alone, or 300 ␮g anti-TNF-␣ plus 200 ␮g anti-IL-1R. Controls Anti-CD4 received PBS. Mice started to develop arthritis 17 days after im- Lytic anti-CD4 mAb (rat IgG2b) consisted of a mixture (1:1) of YTS munization in the PBS-treated group, compared with 22 days in the 191.1.2 and YTA 3.1.2 (27–29). YTA 3.1.2 was a gift from Herman Wald- group treated with anti-TNF-␣ and 24 days in the groups treated mann (then at the University of Cambridge, U.K.) and YTS191.1.2 was ob- with anti-IL-1R or anti-IL-1R/anti-TNF-␣ (Fig. 1). This suggests tained from the European Collection of Animal Cell Cultures (Salis- that either TNF-␣-blockade or IL-1-blockade during this prearth- http://www.jimmunol.org/ bury, U.K.). ritic period is capable of protecting mice from disease for a short Clinical assessment of arthritis period after injection. However, this protection was short-lived as arthritis developed very rapidly in the groups of mice treated with Following immunization with type II collagen, mice were monitored for ␣ the ﬁrst signs of arthritis (redness and/or swelling in one or more paws, or anti-TNF- and/or anti-IL-1R mAbs, such that 100% incidence of limping). To compare the severity of arthritis in the different treatment arthritis was reached in all groups (Fig. 1). This clearly indicates groups, clinical score and paw-swelling were monitored over a 10 day that neither TNF-␣ blockade, IL-1 blockade, or combined TNF- treatment period. A scoring system was used with the following scale: 0, ␣/IL-1 blockade had a long-lasting protective effect, although all normal; 1, slight swelling and/or erythema; and 2, pronounced edematous three treatments delayed the start of arthritis.

swelling. Each limb was graded, giving a maximum score of 8 per mouse. by guest on September 28, 2021 Paw thickness in affected hind-paws was measured with calipers throughout the treatment period and a value for total paw-swelling over this period Dose-dependent therapeutic effect of anti-IL-1R mAb in was obtained by determining the area under the curve, using the trapezoidal established CIA method. We had earlier demonstrated a dose-dependent beneficial influence Histological assessment of arthritis of anti-TNF-␣ mAb after the onset of clinical arthritis, a finding At the end of the treatment period the mice were killed, bled, and their that helped to establish the validity of the therapeutic effect of joints were processed for histology. The first limb to show clinical evidence TNF-blockade (8). Consequently, we set out to demonstrate a of arthritis was removed, fixed, decalcified, and embedded before section- comparable effect for IL-1 blockade in established CIA. Mice were ing and staining with hemotoxylin and eosin. Saggital sections of the prox- immunized with type II collagen and monitored for the first signs imal interphalangeal (PIP) joint of the middle digit were examined by of arthritis. On days 1, 4, and 7 of arthritis, the mice were treated microscopy in a blinded fashion for the presence or absence of erosions, as ␮ defined (21). Thus, comparisons were made of the same joints and the with different doses of anti-IL-1R mAb (20, 100, or 200 g/ arthritis was of identical duration in each case. The PIP joint was chosen because in previous studies, erosions have been found to be present in this joint in 90–100% of untreated arthritic mice on day 10 of arthritis. A significant reduction in this figure is taken to represent a beneficial therapeutic effect on joint erosion. Immunohistochemistry Joints from treated arthritic mice were analyzed for TNF-␣ and IL-1␤ expression using a previously described procedure (30). In brief, joints were embedded in OCT embedding matrix, snap-frozen, and stored at Ϫ70°C until use. Sagittal sections (6 ␮M thick) were air-dried then fixed in 4% paraformaldehyde. After blocking endogenous peroxidase activity the slides were incubated with anti-TNF-␣ mAb (MP6-XT22; PharMingen, San Diego, CA) or anti-IL-1␤ mAb (B122; Genzyme, Cambridge, MA). Sections were then washed and incubated with biotinylated secondary Ab (Vector Laboratories, Burlingame, CA). Ab-biotin conjugates were detected with an avidin-biotin-HRP complex (Vectastain Elite ABC; Vector Laboratories) and developed with diaminobenzidine.

Quantiﬁcation of cytokine immunostaining FIGURE 1. Effect of TNF-␣-blockade and/or IL-1-blockade during the Sections stained for TNF-␣ or IL-1␤ were examined using an Olympus induction phase of CIA. Mice were treated with anti-TNF-␣ mAb or anti- BH-2 microscope (New Hyde Park, NY) and analyzed by computer image IL-1R mAb on days 14 and 16 postimmunization. There were 12 analysis (AnalySIS; Soft Imaging System, Munster, Germany). Six digital mice/group. 7242 TNF-␣ AND IL-1 BLOCKADE IN CIA mouse). The experiment was terminated on day 10. Anti-IL-1R treatment had a dose-dependent therapeutic effect, as judged by clinical score, paw-swelling and joint erosion (Table I). Thus, a dose of 20 ␮g/mouse of anti-IL-1R failed to modulate any of the disease parameters, whereas doses of 100 ␮g and 200 ␮g reduced clinical score, paw-swelling, and joint erosion, although the small number of mice used in the study (4 mice/group) precluded sta- tistical analysis of the data. For all further studies, doses of 200 ␮g/mouse were used.

TNF-␣ and IL-1 blockade in established CIA One of the principle objectives of this study was to compare the effects TNF-␣ blockade, IL-1 blockade and combined TNF-␣/IL-1 blockade in established arthritis. Therefore, mice were treated with anti-TNF-␣ alone, anti-IL-1R alone or anti-TNF-␣ plus anti-IL- 1R. Additional groups were given anti-IL-1␤ alone, anti-IL-1␤ FIGURE 2. Effect of anti-TNF-␣, anti-IL-1R, and combined anti-TNF- ␣ plus anti-TNF-␣, or anti-TNF-␣ plus anti-CD4. Controls were /anti-IL-1R treatment on paw-swelling established CIA. Day 1 corre- given PBS. The first injection was given on day 1 of arthritis (the sponds to the first day that clinical arthritis was observed and mAbs were administratered on days 1, 4, and 7. Data are presented as means Ϯ SE. first day that clinical arthritis was detected) and repeated on days Downloaded from ␣ Each group contains 14 or more mice. Additional parameters of disease 4 and 7. The experiment was terminated on day 10. Anti-TNF- severity are shown in Table II. treatment alone caused a significant reduction in paw-swelling although anti-IL-1R treatment had a much more pronounced therapeutic effect (Fig. 2 and Table II). Combined treatment with anti- indicates that IL1␤ plays a more important role in the pathogenesis TNF-␣ and anti-IL-1R was more effective than anti-IL-1R alone, of CIA than anti-IL-1␣. although the differences between the two groups were not statis- This study has established that most of the pathology in CIA can http://www.jimmunol.org/ tically significant. Clinical scores were significantly reduced in the be attributed to either TNF-␣ or IL-1, and in an earlier study we mice treated with anti-TNF-␣ and even more so in the mice given had reported that combined treatment with anti-TNF and anti-CD4 anti-IL-1R (Table II). Finally, the extent of joint erosion was com- prevented mice from much of the pathology associated with CIA. pared by histological evaluation of the erosive changes in the PIP These two findings suggest that combined anti-TNF/anti-CD4 joints. Of the control arthritic mice, 86% of the joints showed treatment results in the suppression of both TNF-␣ and IL-1 ac- erosive changes, compared with 52% for the mice given anti- tivity. To further support this hypothesis, a direct comparison was TNF-␣ alone, 36% for the mice given anti-IL-1R alone, and only made of combined treatment with anti-TNF-␣ plus anti-CD4 and 8% in the group given anti-TNF-␣ plus anti-IL-1R (Table II). anti-TNF-␣ plus anti-IL-1R/anti-IL-1␤. As judged by clinical It was concluded that anti-IL-1R therapy was more effective in score, paw-swelling, and joint erosion, the magnitude of the ther- by guest on September 28, 2021 this model of arthritis than anti-TNF-␣ therapy, and a combination apeutic effect of combined anti-TNF-␣/anti-CD4 treatment was of anti-IL-1R and anti-TNF-␣ showed a trend toward even greater very similar to, or greater than, that of anti-IL-1R/anti-IL-1␤ alone suppression of disease. A similar result was obtained when TNF-␣ or anti-TNF-␣ plus anti-IL-1R/anti-IL-1␤ (Table II). This suggests blockade was compared with IL-1 blockade using anti-IL-1␤ mAb. that treatment with anti-TNF-␣ plus anti-CD4 leads to the sup- Thus, anti-IL-1␤ was more effective than anti-TNF-␣ in reducing pression of IL-1 activity. paw-swelling, clinical score, and joint erosion (Table II). How- ever, combined treatment with anti-TNF-␣ plus anti-IL-1␤ re- Effect of therapy on the acute phase response sulted in even further reductions in paw-swelling and joint erosion It is reported that the pro-inflammatory cytokines, TNF-␣, IL-1, than anti-IL-1␤ alone (Table II). and IL-6, are major mediators of the acute phase response (32). The anti-IL-1R mAb used in these experiments neutralizes the Furthermore, serum levels of at least one acute phase protein, SAP, activity of both IL-1␣ and ␤ through its interaction with the type have been found to be elevated in CIA (33). Therefore, an addi- I IL-1R. In view of the potent ameliorative effect of anti-IL-1R tional study was conducted to compare levels of SAP in mice mAb in this study, the question is raised of which cytokine, IL-1␣ treated with anti-TNF-␣ and/or anti-IL-1R or anti-TNF-␣ plus anti- or IL-1␤, plays the dominant pathological role in CIA. In this CD4. Neither anti-TNF-␣ nor anti-IL-1R treatment alone had any study, very similar levels of suppression of clinical score, paw- significant effect on levels of SAP although there was a trend to- swelling, and joint erosion were observed in the groups treated ward reduced levels in the anti-IL-1R-treated group (Table III). with anti-IL-1R mAb or anti-IL-1␤ mAb (Table II). This finding However, there was a significant reduction in SAP levels in the

Table I. Dose-dependent therapeutic effect of anti-IL-1R mAb in established CIAa

Paw Swelling Proportion of PIP Treatment No. Clinical Score (mean Ϯ SE) (AUC; mean Ϯ SE) Joints With Erosions

Controls 5 3.2 Ϯ 0.37 4.3 Ϯ 1.12 6/6 Anti-IL-1R (20 ␮g) 4 3.0 Ϯ 0.58 5.30 Ϯ 0.53 4/4 Anti-IL-1R (100 ␮g) 4 1.7 Ϯ 0.48 1.77 Ϯ 0.58 2/4 Anti-IL-1R (200 ␮g) 4 1.0 Ϯ 0.41 1.36 Ϯ 0.12 1/4

a Mice were treated over a 10-day period, starting on the day of onset of arthritis. Clinical scores were compared on day 10 of arthritis. Paw swelling was monitored throughout the treatment period and a value for total paw swelling was obtained by determining the area under the curve (AUC). For the histology, saggital sections of the PIP joints were examined by light microscopy for the presence or absence of erosions. The Journal of Immunology 7243

Table II. Therapeutic effects of TNF-␣/IL-1 blockade in CIAa

Paw Swelling Over the Treatment Period (AUC) Histology Clinical Score Treatment No. (mean Ϯ SE) Paw Swelling (mean Ϯ SE) Relative Reduction PIP Joints With Erosions

Controls 21 3.4 Ϯ 0.2 4.67 Ϯ 0.38 – 18/21 (86%) Anti-TNF-␣ (300 ␮g) 23 2.2 Ϯ 0.2** 3.28 Ϯ 0.36* 30% 12/23* (52%) Anti-IL-1R (200 ␮g) 14 1.4 Ϯ 0.2** 1.40 Ϯ 0.23** 70% 5/14** (36%) Anti-TNF-␣ (300 ␮g) plus anti-IL-1R (200 ␮g) 14 1.4 Ϯ 0.2** 1.03 Ϯ 0.23** 78% 1/12** (8%) Anti-IL-1␤ (300 ␮g) 6 1.2 Ϯ 0.2** 1.4 Ϯ 0.17** 70% 1/6* (17%) Anti-TNF-␣ (300 ␮g) plus anti-IL-1␤ (300 ␮g) 6 1.2 Ϯ 0.2** 0.85 Ϯ 0.23** 82% 0/6* (0%) Anti-TNF-␣ (300 ␮g) plus anti-CD4 (200 ␮g) 7 1.4 Ϯ 0.2** 0.72 Ϯ 0.17** 84% 1/7* (14%)

a Mice were treated over a 10-day period, starting on day 1 of arthritis. Clinical scores were compared on day 10 of arthritis. Paw swelling was monitored throughout the treatment period and a value for total paw swelling was obtained by determining the AUC. For the histology, sections were examined by microscopy for the presence or absence of erosions. Data was pooled from two experiments. .p Ͻ 0.01; p values refer to differences between treated mice and controls ,ءء ;p Ͻ 0.05 ,ء

group given anti-TNF-␣ plus anti-IL-1R and an even more pro- Discussion ␣ nounced reduction in the group given anti-TNF- plus anti-CD4. In this study we have evaluated the effects of TNF-␣ blockade and Thus, SAP levels found in anti-TNF-␣/anti-CD4-treated mice were IL-1 blockade in CIA. In the first experiment, mAbs to TNF-␣ or Downloaded from similar to or less than those found in a group of age- and sex- to the type I IL-1R were administered during the induction phase matched, unimmunized DBA/1 mice. These data suggest that of CIA and it was found that both forms of treatment caused a ␣ TNF- and IL-1 are not direct inducers of SAP, but act through delay of 5–7 days in the manifestation of clinical arthritis (Fig. 1). another mediator, such as IL-6 (34), and this further supports the After this delay, arthritis developed rapidly in treated mice, pre- ␣ concept that combined anti-TNF- /anti-CD4 treatment results in sumably as a consequence of the clearance of mAbs from the cir- suppression of other pro-inflammatory cytokines in addition to ␣ culation. From this it was concluded that both TNF- and IL-1 are http://www.jimmunol.org/ ␣ TNF- . required for the induction of CIA but a brief pulse of anti-TNF-␣ Immunohistological evaluation of TNF-␣ and IL-1␤ expression or anti-IL-1R treatment during this prearthritic phase fails to mod- following therapy ulate the ongoing autoimmune response as all mice eventually developed arthritis. We next set out to address the question of the extent to which A second set of experiments was then performed in which the ␣ anti-IL-1 therapy affects the level of TNF- expression and, con- effect of TNF-␣ blockade and/or IL-1 blockade were assessed in ␣ versely, the extent to which anti-TNF- , or combined anti-TNF- established arthritis. Two different IL-1 blocking mAbs were used, ␣ /anti-CD4 therapy, affects the level of IL-1 expression. Mice with anti-IL-1R and anti-IL-1␤. Anti-TNF-␣ treatment alone was found established CIA were treated on days 1 and 4 of arthritis with to reduce the clinical and histological severity of arthritis, a finding by guest on September 28, 2021 ␣ ␣ anti-IL-1R, anti-TNF- , or anti-TNF- plus anti-CD4. The mice that is consistent with previously published reports (7–11). How- were killed on day 6 of arthritis and undecalcified joints were ever, anti-IL-1 treatment, using either anti-IL-1R mAb or anti- ␣ ␤ cryosectioned and stained for TNF- /IL-1 expression. The pro- IL-1␤ mAb, resulted in significantly greater suppression of dis- portion of positively stained cells was quantified using image anal- ease, both clinically and histologically than anti-TNF-␣ (Table II). ␣ ysis. Both anti-IL-1R and anti-TNF- gave significant reductions The findings presented here are consistent with previous findings ␣ ␤ in the proportions of TNF- -positive cells and IL-1 -positive cells by Van den Berg and colleagues (15, 16), which reported marked ␣ (Table IV). Combined anti-TNF- /anti-CD4 therapy caused even amelioration of established CIA following treatment with poly- ␣ ␤ greater suppression of TNF- , and particularly IL-1 expression. clonal Abs against IL-1␣ and IL-1␤, anti-IL-1␤ alone or high ␣ This confirms the ability of combined anti-TNF- /anti-CD4 treat- doses of IL-1Ra. However, the results of comparisons between ␤ ment to profoundly suppress the level of IL-1 expression at the different mAbs should be interpreted with caution, as the mAbs site of disease activity. may differ appreciably in their ability to neutralize their target cytokine or to penetrate the site of disease activity. Nevertheless, Effect of therapy on the anti-collagen Ab response these findings may suggest that IL-1, like TNF-␣, represents a None of the different treatments tested had any significant effect on potential therapeutic target for RA. circulating anti-type II collagen IgG levels, assayed at day 10 (data A question that we have addressed in this report is whether not shown). combined treatment of CIA with a TNF-␣ neutralizing mAb plus

Table III. Levels of SAP in arthritic mice following treatment with anti- Table IV. Immunohistochemical analysis of TNF-␣ and IL-1␤ ␣ ␣ a TNF- and/or anti-IL-1R or combined anti-TNF- /anti-CD4 treatment expression in arthritic joints following therapya

␮ Treatment No. SAP Level ( g/ml) Treatment (no.) TNF-␣ Expression (%) IL-1␤ Expression (%) Ϯ Controls 16 847 157 Controls (4) 15.0 Ϯ 1.3 6.8 Ϯ 0.4 ␣ Ϯ Anti-TNF- alone 15 1075 317 (NS) Anti-TNF-␣ (4) 8.6 Ϯ 1.0 ( p Ͻ 0.001) 4.0 Ϯ 0.8 ( p Ͻ 0.01) Ϯ Anti-IL-1R alone 13 502 137 (NS) Anti-IL-1R (4) 8.4 Ϯ 0.5 ( p Ͻ 0.001) 2.5 Ϯ 0.2 ( p Ͻ 0.001) ␣ Ϯ Ͻ Anti-TNF- plus anti-IL-1R 14 334 74 ( p 0.05) Anti-TNF-␣/anti-CD4 (4) 6.0 Ϯ 0.3 ( p Ͻ 0.001) 1.8 Ϯ 0.2 ( p Ͻ 0.001) Anti-TNF-␣ plus anti-CD4 7 179 Ϯ 64 ( p Ͻ 0.05) Nonarthritic age-matched 6 303 Ϯ 45 a Mice with established CIA were treated on days 1 and 4 of arthritis with anti- DBA/1 mice IL-1R, anti-TNF-␣, or anti-TNF-␣ plus anti-CD4. The mice were killed on day 6 of arthritis, and joints were stained for TNF-␣ and IL-1␤ expression. Image analysis was a Serum samples were taken on day 10 of arthritis and SAP levels were measured used to quantitate the level of TNF-␣ or IL-1␤ expression. The data are presented as by ELISA. Values of p refer to differences between mAb-treated groups and the the area covered by positively stained cells, expressed as a percentage of the total area controls. covered by cells. Values of p refer to differences between treated groups and controls. 7244 TNF-␣ AND IL-1 BLOCKADE IN CIA an IL-1 neutralizing mAb provides greater therapeutic effect than in the case of anti-TNF-␣. In CIA, previous findings have shown either mAb alone. Anti-TNF-␣ treatment combined with anti- that IL-1 plays an important role in cartilage degradation (15, 16, IL-1R or anti-IL-1␤ was found to provide increased suppression of 35) and our findings indicate that IL-1 blockade is highly effective arthritis than anti-IL-1R alone or anti-IL-1␤ alone, though the dif- in limiting joint erosion. These findings strengthen the case for ferences were not statistically significant. Histologically, an addi- further evaluation and optimization of anti-IL-1 therapy in the tive effect was indeed observed between anti-TNF-␣ and anti- clinic. IL-1R or anti-IL-1␤ in the protection against joint erosion (Table II). In addition, there was clear evidence of an additive therapeutic References ␣ effect between anti-TNF- and anti-IL-1R in suppression of the 1. Elliott, M. J., R. N. Maini, M. Feldmann, A. Long-Fox, P. Charles, P. Katsikis, acute phase response, as judged by circulating levels of SAP F. M. Brennan, J. Walker, H. Bijl, J. Ghrayeb, and J. N. Woody. 1993. Treatment (Table III). of rheumatoid arthritis with chimeric monoclonal antibodies to tumour necrosis factor ␣. Arthritis Rheum. 36:1681. In previous studies we demonstrated synergy between anti- 2. Elliott, M. J., R. N. Maini, M. Feldmann, A. Long-Fox, P. Charles, H. Bijl, and TNF-␣ and anti-CD4 in the amelioration of established CIA and it J. N. Woody. 1994. Repeated therapy with a monoclonal antibody to tumour was postulated that an important mechanism underlying this syn- necrosis factor ␣ in patients with rheumatoid arthritis. Lancet 344:1125. ␣ 3. Elliott, M. J., R. N. Maini, M. Feldmann, J. R. Kalden, C. Antoni, J. S. Smollen, ergy was the suppression, not just of TNF- , but also of other B. Leeb, F. C. Breedfeld, J. D. Macfarlane, H. Bijl, and J. N. Woody. 1994. pro-inflammatory cytokines, including IL-1 (21). In the present Treatment with a chimaeric monoclonal antibody to tumour necrosis factor ␣ study we have demonstrated, first, that most of the pathology as- suppresses disease activity in rheumatoid arthritis: results of a multi-centre, ran- domised, double blind trial. Lancet 344:1105. sociated with CIA can be attributed to either IL-1 or TNF-␣ and 4. Rankin, E. C. C., E. H. S. Choy, D. Kassimos, G. H. Kingsley, A. M. Sopwith, second, that the magnitude of the therapeutic effect of combined D. A. Isenberg, and G. S. Panayi. 1995. The therapeutic effects of an engineered Downloaded from ␣ human anti-tumour necrosis factor ␣ antibody (CDP571) in rheumatoid arthritis. anti-TNF- /anti-CD4 treatment was comparable to combined Br. J. Rheumatol. 34:334. treatment with anti-TNF-␣ plus anti-IL-1R/anti-IL-1␤. These two 5. Moreland, L. W., S. W. Baumgartner, M. H. Schiff, E. A. Tindall, findings provide evidence to support the concept that combined R. M. Fleischmann, A. L. Weaver, R. E. Ettlinger, S. Cohen, W. J. Koopman, ␣ K. Mohler, et al. 1997. Treatment of rheumatoid arthritis with a recombinant anti-TNF- /anti-CD4 treatment results in the effective elimination human tumor necrosis factor receptor (p75)-Fc fusion protein. N. Engl. J. Med. of the pathology attributable to both TNF-␣ and IL-1. Further ev- 337:141. 6. Brennan, F. M., D. Chantry, A. Jackson, R. Maini, and M. Feldmann. 1989. idence in support of this concept is provided by the finding that http://www.jimmunol.org/ ␣ Inhibitory effect of TNF␣ antibodies on synovial cell interleukin-1 production in combined anti-TNF- /anti-CD4 treatment led to normalization of rheumatoid arthritis. Lancet 2:244. SAP levels as elevated levels of SAP are reported to be induced by 7. Thorbecke, G. J., R. Shah, C. H. Leu, A. P. Kuruvilla, A. M. Hardison, and a number of pro-inflammatory cytokines, including TNF-␣, IL-1 M. A. Palladino. 1992. Involvement of endogenous tumor necrosis factor ␣ and transforming growth factor ␤ during induction of collagen type II arthritis in and most importantly, IL-6 (34). The fact that combined anti-TNF- mice. Proc. Natl. Acad. Sci. USA 89:7375. ␣/anti-CD4 treatment caused a greater reduction in SAP levels 8. Williams, R. O., M. Feldmann, and R. N. Maini. 1992. Anti-tumor necrosis factor than combined anti-TNF/anti-IL-1R treatment raises the possibil- ameliorates joint disease in murine collagen-induced arthritis. Proc. Natl. Acad. ␣ Sci. USA 89:9784. ity that treatment with anti-TNF- plus anti-CD4 resulted also in 9. Piguet, P. F., G. E. Grau, C. Vesin, H. Loetscher, R. Gentz, and W. Lesslauer. the suppression of additional cytokines, such as IL-6. Finally, we 1992. Evolution of collagen arthritis in mice is arrested by treatment with anti- were able to confirm using immunohistochemistry that combined tumour necrosis factor (TNF) antibody or a recombinant soluble TNF receptor. by guest on September 28, 2021 Immunology 77:510. anti-TNF-␣/anti-CD4 therapy leads to marked suppression in the 10. Wooley, P. H., J. Dutcher, M. B. Widmer, and S. Gillis. 1993. Influence of a level of expression of both TNF-␣ and IL-1␤ at the site of disease recombinant human soluble tumour necrosis factor receptor Fc fusion protein on activity (Table IV). type II collagen-induced arthritis in mice. J. Immunol. 151:6602. 11. Williams, R. O., J. Ghrayeb, M. Feldmann, and R. N. Maini. 1995. Successful The mechanism by which combined anti-TNF-␣/anti-CD4 treat- therapy of collagen-induced arthritis with TNF receptor-IgG fusion protein and ment could conceivably give rise to the suppression of IL-1 activ- combination with anti-CD4. Immunology 84:433. 12. Keffer, J., L. Probert, H. Cazlaris, S. Georgopoulos, E. Kaslaris, D. Kioussis, and ity is a matter for conjecture. One possibility is that there are two G. Kollias. 1991. Transgenic mice expressing human tumour necrosis factor: a major pathways involved in the induction of TNF-␣ and IL-1 in predictive genetic model of arthritis. EMBO J. 10:4025. CIA. The first pathway may involve the production of pro-inflam- 13. Probert, L., D. Plows, G. Kontogeorgos, and G. Kollias. 1995. The type I inter- matory cytokines by cells in the synovium, following their activa- leukin-1 receptor acts in series with tumor necrosis factor (TNF) to induce ar- ϩ thritis in TNF-transgenic mice. Eur. J. Immunol. 25:1794. tion by CD4 T cells. The second pathway may function indepen- 14. Geiger, T., H. Towbin, A. Cosenti-Vargas, O. Zingel, J. Arnold, C. Rordorf, and dently of T cells and may involve the induction of TNF-␣ and IL-1 K. Vosbeck. 1993. Neutralization of interleukin-1␤ activity in vivo with a monoclonal antibody alleviates collagen-induced arthritis in DBA/1 mice and prevents production by TNF itself in an autocrine/paracrine fashion. This is the associated acute-phase response. Clin. Exp. Rheumatol. 11:515. supported by the observation that the spontaneous production of 15. Van den Berg, W. B., L. A. Joosten, M. Helsen, and, F. A. van de Loo. 1994. IL-1 by human synovial cells in vitro is effectively blocked by the Amelioration of established murine collagen-induced arthritis with anti-IL-1 ␣ treatment. Clin. Exp. Immunol. 95:237. addition of anti-TNF- Abs (6). This would account for the fact 16. Joosten, L. A. B., M. M. A. Helen, F. A. J. van de Loo, and W. B. Van den Berg. that anti-CD4 treatment alone is relatively ineffective in amelio- 1996. Anticytokine treatment of established type II collagen-induced arthritis in rating already established arthritis. In contrast, combined anti- DBA/1 mice: a comparative study using anti-TNF␣, anti-IL-1␣/␤, and IL-1Ra. ␣ Arthritis Rheum. 39:797. TNF- /anti-CD4 treatment would modulate cytokine production 17. Bendele, A., T. McAbee, G. Sennello, J. Frazier, E. Chlipala, and D. McCabe. elicited by both pathways. 1999. Efficacy of sustained blood levels of interleukin-1 receptor antagonist in The impressive therapeutic effect achieved by IL-1-blockade in animal models of arthritis: comparison of efficacy in animal models with human clinical data. Arthritis Rheum. 42:498. this study raises the possibility that this form of treatment is im- 18. Kuiper, S., L. A. Joosten, A. M. Bendele, C. K. Edwards, 3rd, O. J. Arntz, munosuppressive as well as anti-inflammatory. However, circulat- M. M. Helsen, F. A. Van de Loo, and W. B. Van den Berg. 1998 . Different roles of tumour necrosis factor ␣ and interleukin 1 in murine streptococcal cell wall ing anti-type II collagen IgG levels were not significantly altered arthritis. Cytokine 10:690. by anti-IL-1R/anti-IL-1␤ therapy (data not shown). This is consis- 19. Bresnihan, B., J. M. Alvaro-Gracia, M. Cobby, M. Doherty, Z. Domljan, tent with the results of a previous study in which anti-IL-1␤ mAb, P. Emery, G. Nuki, K. Pavelka, R. Rau, B. Rozman, et al. 1998. Treatment of rheumatoid arthritis with recombinant human interleukin-1 receptor antagonist. administered continuously from the time of collagen immuniza- Arthritis Rheum. 41:2196. tion, had no effect on anti-collagen IgG levels (14). 20. Miller, E. J. 1972. Structural studies on cartilage collagen employing limited In conclusion, we have demonstrated that IL-1, in addition to cleavage and solubilization with pepsin. Biochemistry 11:4903. ␣ 21. Williams, R. O., L. J. Mason, M. Feldmann, and R. N. Maini. 1994. Synergy TNF- , is a potential target for therapeutic intervention in RA. between anti-CD4 and anti-TNF in the amelioration of established collagen-in- Increasingly, the goal in RA therapy is to prevent joint damage, as duced arthritis. Proc. Natl. Acad. Sci. USA 91:2762. The Journal of Immunology 7245

22. Butler, D. M., A.-M. Malfait, R. N. Maini, F. M. Brennan, and M. Feldmann. 29. Qin, S., S. Cobbold, H. Tighe, R. Benjamin, and H. Waldmann. 1987. CD4 1999. Anti-IL-12 and anti-TNF antibodies synergistically suppress the progres- monoclonal antibody pairs for immunosuppression and tolerance induction. Eur. sion of murine collagen induced arthritis. Eur. J. Immunol. 7:2205. J. Immunol. 17:1159. 23. Sheehan, K. C., N. H. Ruddle, and R. D. Schreiber. 1989. Generation and char- 30. Marinova-Mutafchieva, L., R. O. Williams, L. J. Mason, C. Mauri, M. Feldmann, acterization of hamster monoclonal antibodies that neutralize murine tumor ne- and R. N. Maini. 1997. Dynamics of proinflammatory cytokine expression in the crosis factors. J. Immunol. 142:3884. joints of mice with collagen-induced arthritis (CIA). Clin. Exp. Immunol. 107: 24. Baker, D., D. Butler, B. J. Scallon, J. K. O’Neill, J. L. Turk, and M. Feldmann. 507. 1994. Control of established experimental allergic encephalomyelitis by inhibition of tumour necrosis factor (TNF) activity within the central nervous system 31. Serban, D., and C. Rordorf-Adam. 1986. Quantitation of serum amyloid P com- using monoclonal antibodies and TNF receptor-immunoglobulin fusion proteins. ponent by an enzyme-linked immunoassay. J. Immunol. Methods 90:159. Eur. J. Immunol. 24:2040. 32. Baumann, H., and J. Gauldie. 1994. The acute phase response. Immunol. Today 25. McIntyre, K. W., G. J. Stepan, K. D. Kolinsky, W. R. Benjamin, J. M. Plocinski, 15:74. C. A. Campen, R. A. Chizzonite, and P. L. Kilian. 1991. Inhibition of interleukin 33. Bliven, M. L., P. H. Wooley, M. B. Pepys, and I. G. Otterness. 1986. Murine type 1 (IL-1) binding and bioactivity in vitro and modulation of acute inflammation in II collagen arthritis: association of an acute-phase response with clinical course. vivo by IL-1 receptor antagonist and anti-IL-1 receptor monoclonal antibody. Arthritis Rheum. 29:1131. J. Exp. Med. 173:931. 26. Gershenwald, J. E., Y. M. Fong, T. J. Fahey, S. E. Calvano, R. Chizzonite, 34. Steel, D. M., and A. S. Whitehead. 1994. The major acute phase reactants: C- S. F. Lowry, and L. L. Moldawer. 1990. Interleukin 1 receptor blockade atten- reactive protein, serum amyloid P component and serum amyloid A protein. uates the host inflammatory response. Proc. Natl. Acad. Sci. USA 87:4966. Immunol. Today 15:81. 27. Galfre, G., C. Milstein, and B. Wright. 1979. Rat ϫ rat hybrid myelomas and a 35. van de Loo, F. A., L. A. Joosten, P. L. van Lent, O. J. Arntz, and monoclonal anti-Fd portion of mouse IgG. Nature 277:131. W. B. Van den Berg. 1995. Role of interleukin-1, tumor necrosis factor ␣, and 28. Cobbold, S. P., A. Jayasuriya, A. Nash, T. D. Prospero, and H. Waldmann. 1984. interleukin-6 in cartilage proteoglycan metabolism and destruction: effect of in Therapy with monoclonal antibodies by elimination of T-cell subsets in vivo. situ blocking in murine antigen- and zymosan-induced arthritis. Arthritis Rheum. Nature 312:548. 38:164. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021