UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Analysis of the mechanism of action of anti-human interleukin-6 and anti-human interleukin-6receptor-neutralising monoclonal antibodies

Kalai, M.; Montero-Julian, F.A.; Brakenhoff, J.; Fontaine, V.; de Wit, L.; Wollmer, A.; Brailly,H.; Content, J.; Grotzinger, J.Published in:European Journal of Biochemistry

DOI:10.1111/j.1432-1033.1997.t01-2-00690.x

Link to publication

Citation for published version (APA):Kalai, M., Montero-Julian, F. A., Brakenhoff, J., Fontaine, V., de Wit, L., Wollmer, A., ... Grotzinger, J. (1997).Analysis of the mechanism of action of anti-human interleukin-6 and anti-human interleukin-6 receptor-neutralising monoclonal antibodies. European Journal of Biochemistry, 249, 690-700.https://doi.org/10.1111/j.1432-1033.1997.t01-2-00690.x

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 17 Oct 2020

Eur. J . Biochem. 249, 690-700 (1997) 0 FEBS 1997

Analysis of the mechanism of action of anti-human interleukin-6 and anti-human interleukin-6 receptor-neutralising monoclonal antibodies Michael KALAI ', Felix A. MONTERO-JULIAN', Just P. J. BRAKENHOFF', VCronique FONTAINE', Lucas DE WIT', Axel WOLLMER4 HervC BRATLLY '. Jean CONTENT' and Joachim GROTZINGER" ' Institut Pasteur de Bruxelles, DCpartement de Virologie, Bruxelles, Belgium * Immunotech S. A,, Marseille, France

Department of Autoimmune Diseases, Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (CLB), Amsterdam, The Netherlands Institut fur Biochemie. Rheinisch-Westfalisch Technische Hochschule Aachen, Aachen, Germany

(Received 25 May/l9 August 1997) - EJB 97 0745/3

Anti-human interleukin-6 (human IL-6) and anti-human IL-6 receptor (L6R)-neutraIising monoclo- nal antibodies (mAbs) are among the most promising human IL-6-specific inhibitors and have been shown to exert short-term beneficial effects in clinical trials. Simultaneous treatment with different anti-human IL-6 or anti-human IL-6R mAbs was recently suggested to be a potent way to inhibit the action of the cytokine in vivo. Although some of these mAbs are already used, their mechanisms of action and the location of their epitopes on the surface of human IL-6 and human IL-6R are still unknown. Here, we analysed the capacity of several anti-human IL-6 and anti-human IL-6R mAbs to inhibit the interaction between human IL-6, human IL-6R, and human glycoprotein 130 (gp130). We mapped the epitopes of several of these mAbs by studying their binding to human IL-6 and human IL-6R mutant proteins. Our results show that several anti-human IL-6 and anti-human IL-6R-neutralising mAbs block the binding between human IL-6 and human IL-6R, whereas others block the binding to gpl30. We provide evidence that some of the latter mAbs inhibit interaction with gp130p1, whereas others interfere with the binding to gp130p2. Our results suggest that residues included in the C'D' loop of human IL-6R interact with gp 1 3Op2.

Keywords: epitope mapping ; interleukin-6 ; interleukin-6 receptor; glycoprotein 130 ; monoclonal anti- body.

Interleukin-6 (IL-6) is a pleiotropic cytokine mainly in- volved in the regulation of inflammation, immune responses, and hematopoi'esis. IL-6 binds first to a low-affinity a-subunit receptor called IL-6R. The IL-6LL-6R complex associates with two glycoprotein 130 (gp130) /$receptor chains. The two iden- tical molecules, which are designated here as gpl30pl and gp130p2, form a covalently linked homodimer that transduces the signal [ I , 21. The exact stoichiometry and composition of the high-affinity IL-6 receptor complex at the cell surface mem- brane are still under debate.

Overexpression of IL-6 has been suggested to participate in the pathogenesis of a variety of clinical disorders, including sur- gical trauma, rheumatoid arthritis, post-menopausal osteoporo- sis, chronic autoimmune diseases, Castleman's disease, multiple myeloma, renal carcinoma and AIDS [l]. To date, the rational design of efficient IL-6 inhibitors was only partially successful. Among the most promising human IL-6 (human IL-6)-specific inhibitors are anti-human IL-6 and anti-human IL-6R-neutralis-

Correspondence to J. Content, Institut Pasteur de Bruxelles, DCparte-

F a x : +322 373 3291. E-mail: jcontent @ ben.vub.ac.be URLr http ://hen.vuh.ac.be/-jdbraek/pasteur.htm Abbreviations. AP, alkaline phosphatase; gp130, glycoprotein 130;

Enzyme. Alkaline phosphatase (EC 3.1.3.1). Note. M. Kalai and E A. Montero-Julian contributed equally to this

ment de Virologie, 642 rue Engeland, B-1180 Bruxelles, Belgium

1L-6, interleukin-6; L 6 R , interleukin-6 receptor.

study.

ing antibodies. So far, these are the only ones tested in vivo. Anti-human IL-6 mAbs were shown to exert short-term benefi- cial effects both in animal models of some of the IL-6-related diseases, and in human clinical trials 13-61. In addition to anti- human IL-6 mAbs, other antibodies raised against human IL-6R were also shown to inhibit the activity of human IL-6 both in vitro and in vivo [7-111. A few of these anti-human IL-6R mAbs were humanized to prevent the development of human anti-mouse immunoglobulin response in trcatcd paticnts 112, 131. Inhibition of human IL-6 and human IL-6R activity by the humanized anti-human IL-6R antibody PM-1 has been shown to be effective for the treatment of patients with rheumatoid arthri- tis, Castleman's disease, and multiple myeloma [14].

A major drawback encountered during anti-human IL-6 ther- apy might prevent the clinical application of such antibodies in the near future. The problem observed in these trials was the accumulation of large amounts of circulating human IL-6 in im- mune complexes [15, 161. This stabilised human IL-6 is poten- tially biologically active and might be released from the immune complex by competing IL-6R [17]. The situation is aggravated in patients that produce large amounts of human IL-6 and solu- ble human IL-6R [18, 191. Pharmacokinetics studies in mice injected with radiolabelled human IL-6 have demonstrated that treatment with at least three different anti-human IL-6 mAbs resulted in rapid clearance of the mAblhuman IL-6 complexes, and antibody stabilisation of circulating cytokine was no longer observed [20, 211. It seems that simultaneous treatment with dif-

Kalai et al. (Eur: J . Biochem. 249) 69 1

ferent anti-human IL-6 mAbs may therefore be a potent way to inhibit the action of the cytokine in vivo.

These studies suggested that the mAbs included in the cock- tails should not compete with each other for the binding of the antigen. The mAbs should be high-affinity human IL-6 binders and should preferentially interfere with the binding of distinct receptor subunits to prevent as much as possible the formation of the signal transducing IL-6 receptor complex. Such a treat- ment may combine an efficient inhibition of human IL-6 with the long in vivo half-life of free antibodies and the rapid clear- ance of the polymeric immune complexes.

Use of combinations of anti-human IL-6R mAbs interfering efficiently with interactions between human IL-6, human IL-6R, and gp130 will inhibit human IL-6 action even better and may accelerate the removal of excess concentrations of soluble human IL-6R from circulation. Such a treatment may prove more beneficial for rheumatoid arthritis, Castleman's disease, and multiple myeloma patients than their injection with PM-1 alone.

Although several anti-human IL-6 and anti-human IL-6R mAbs are already used, their mechanism of action and the loca- tion of their epitopes on the structures of the corresponding antigens, are still unknown. In this study, we investigated the capacity of several anti-human IL-6 and anti-IL-6R-neutralising mAbs to inhibit the bioactivity of their antigens, studied the mechanism of action of these potent human IL-6 antagonists, identified residues included in their epitopes, and predicted their location on the human IL-6 and human IL-6R structure, respec- tively.

MATERIALS AND METHODS

Reagents. Unless specified otherwise, the buffer used for dilution in all of the immunoassays was 0.1 M Hepes, pH 7.2, 0.15 M NaCI, 1 mM MgCl,, 0.1 mM ZnCI,, 2 mM NaN,, and 20 g/l BSA. The wash buffer was 137 mM NaCl, 2.7 mM KCI, 0.9 mM CaCl,, 0.5 mM MgCl,, 1.5 mM KH,PO,, 8 mM Na,HPO,, pH 7.2 (NaCl/P,), supplemented with 0.01 % NdN,, 1 % BSA.

Antibodies. Anti-human IL-6 mAbs AH-64, AH-65, and anti-human IL-6R mAb M-91 were obtained from Immunotech and prepared as described 122, 231. Anti-human IL-6 mAbs B-E4 and B-E8 were kindly provided by Dr J. Wijdenes (Dia- clone, BesanGon, France) [21, 241. CLB. IL-6 mAbs nos 8, 12, 14, and 16 and LN1-75-8 (B2) were prepared as described [25,26].

Anti-human IL-6R mAbs M-37, M-91, M-113, M-164, M182, and M-195 were kindly provided by Dr J. Brochier (Institut National de la Santt et de la Rechereche MCdicale U291, Montpellier, France) [7]. Anti-human IL-6R mAbs 17.6, 22.1, 32.3, 34.4, 38.8, 48.8, and 50.6 were kindly provided by Dr D. Novick (The Weizmann Institute of Science, Rehovot, Israel) [27]. Anti-human 1L-6 mAb PM-1 and anti human gp130 mAbs AM-64 and AM-66 were a gift from Dr K. Yasukawa (Tosoh Corperation, Kanagawa, Japan) 1281.

Wild-type and mutant human IL-6 and soluble human IL8R proteins. Wild-type human IL-6 and soluble human IL-6R were obtained from Immunotech [22, 231 or prepared as described in [29, 301. Human IL-6 mutants I32 M, F74E, F78E, R168 M, Q175L, R179E, R179W, M184K, double mutants F74E/F78E, K171N/Q175L, and R179E/R182E and triple mutant R30T/Y31H/I32V were produced as described in [29]. Human IL-6 mutants 2a, double mutant Q159E/T162P, triple mutant 2a/Q159E/T162P, deletion mutant 426, and quadruple mutant DFRD were produced as described in 125, 26, 31, 321.

Production of the soluble human IL-6R mutants was carried out as described in 129, 301.

Radiolabelling of human IL-6 and soluble human IL-6R. Metabolic "S-labelling of wild-type human IL-6 and soluble hu- man IL-6R was carried out as described in [30]. Human IL-6 was labelled with "'I using the chloramine-T method as de- scribed previously [21].

Determination of the kinetic parameters of human IL-6 binding by anti-human 1L-6 mAbs in equilibrium. Associa- tion kinetics. Different concentrations of 1 ml biotinylated mAb diluted in NaCUP, supplemented with 0.2% BSA and 5 mM NaN, were added to avidin-coated Eppendorf tubes (Immuno- tech). Following overnight incubation at 4"C, liquid was dis- carded and tubes were washed. To each tube was added 0.5 ml '2'I-human IL-6 at 1.5X10' cpdml . After 1, 5, 10, 20, 30, 40, 60, 120, 180, 240, 300, 360, 420, and 1000 min of incubation, liquid was removed, the tubes were washed, and the bound ra- dioactivity was measured using a y counter.

KO, was calculated using the formula

KO, = -In( 1 - B/B,,,)/Ab,, X t,

where Ab,,+B,,,, Ab, is the initial mAb concentration (equal to the free mAb concentration), B is the bound human IL-6 concen- tration, B,,, is the bound human IL-6 maximal concentration (equal to the initial human IL-6 concentration), and t is time.

Affinity constants. Different concentrations of biotinylated anti-human IL-6 mAb were mixed with an equal amount of human IL-6. Reactions were carried out in a total volume of 1 ml NaCl/P, supplemented with 0.2% BSA and 5 mM NaN, in Eppendorf tubes. Following overnight incubation at 4°C or 37"C, depending on the experiment, 100 pI Sephacryl avidin (Immunotech) was added to each of the tube. Mixtures were incubated for several h at the corresponding initial temperature. Following centrifugation at 2000 rpm in an Eppendorf centri- fuge, the concentration of free human IL-6 in each supernatant was measured by the human IL-6 enzyme immunoassay (EIA- 1120) kit (Immunotech) [33]. Affinity constants were calculated using Scatchard plots and the following formula:

Kd = Ab,,xF/(T-F),

where Ab&T Ab, is the initial mAb concentration (equal to the free mAb concentration), F is the free human IL-6 concentra- tion, and T is the initial human TL-6 concentration (equal to the total human IL-6 concentration).

IL-6 bioassay. The effect of different concentrations of anti- human IL-6 mAbs on human IL-6 bioactivity was measured using the IL-6-dependent mouse hybridoma cell line B9 as de- scribed in [25, 341. Briefly, cells were first washed and incu- bated for 12 h in the absence of IL-6. Cells were then cultured for 48 h in flat bottom microtiter plates at a concentration of 104 cells/ml in a total volume of 200 pl/well RPMI 1640 medium supplemented with 5% fetal calf serum, human IL-6, and dif- ferent concentrations of anti-human IL-6 mAbs. Following incu- bation, 0.25 pCi/well [3H]thymidine (Amersham) was added, and cells were incubated for additional 8 h. Cells were then har- vested on a Whatman filter and mixed with 2 rnl scintillation liquid. The radioactivity incorporated in the cell pellet was deter- mined with a p-counter. A standard range of serial dilutions of human IL-6 was included in every experiment. The specific ac- tivity, which is defined as the human IL-6 dose that gave half of the maximal B9 proliferation response, was 3-5 pg/ml.

Immunoprecipitation of 35S-labelled human IL-6 and sol- uble human ILdR. Immunoprecipitations were carried out in total volumes of 100pl of each mAb at a final concentration of 3 pg/ml in Eppendorf tubes and 20% (by vol.) protein-A- Sepharose (Pharmacia). 5 p1 Y-labelled human IL-6 or soluble

692 Kalai et al. ( E m J. Biochem. 249)

human IL-6R was mixed with the mAbs in the presence or ab- sence of about tenfold (220 ng/ml) excess of unlabelled human IL-6 or soluble human IL-6R depending on the experiment as indicated in Figs 2 and 6. Following overnight incubation at 4°C on a turning wheel, the Sepharose was washed three times. Im- munocomplexes were dissociated in Laemmli buffer, separated by SDSFAGE, and visualised by autoradiography.

Antibody cross-reactivity assays. Competition sandwich ELISA (enzyme-linked immunosorbent assay) was used for the detection of cross-reactivity between the anti-human IL-6 mAbs. mAbs AH-65, AH-64, B-E4, and B-E8 were biotinylated using a protein biotinylation kit according to the manufacturer's in- structions (Boehringer-Mannheim). Distinct microtiter plates were coated overnight with 100 pl/well of a 5 pg/ml solution of each of the anti-human IL-6 mAbs diluted in NaCVP,. Following incubation, plates were washed and incubated with 50 pl/well serial dilutions of human IL-6 and 50 pl/well 1 pg/ml biotin- ylated anti-human IL-6 mAb used as the tracer. Following 6 h of incubation at 4°C on an orbital shaker, plates were washed three times and bound biotinylated mAb was detected with alka- line-phosphatase-conjugated Extravidin (Sigma). Following 1 h incubation, plates were washed again and phosphatase activity was measured by adding 150 pl/well p-nitrophenyl phosphate at 1 mg/ml in a buffer containing 1 M diethanolamine, 1 mM MgCl,, adjusted to pH 8.9 with HC1. The color reaction was allowed to develop and the absorbance was measured at 405 nm by an ELISA reader. Assays were carried out in duplicate.

Mapping the epitopes of anti-human IL-6 mAbs by sand- wich ELISA and the use of human IL-6 mutants. The degree of recognition of human IL-6 mutants 426, Q159E/T162P, 2a, 2a/Q159E/T162P, and DFRD, by the different anti-human IL-6 mAbs was measured by ELISA as described in [25, 26, 351. Briefly, microtiter plates were coated with 1 pg/ml each of the anti-human IL-6 mAbs in NaCUP,. After washing, the plates were incubated with different concentrations of wild-type or variant human IL-6. Bound human IL-6 was revealed using bio- tinylated affinity purified goat anti-human IL-6 and horseradish peroxidase streptavidin (Amersham). Bound peroxidase was de- tected with 3,3'5,5'-tetramethylbenzidine/H,02 (Merck). The re- action was stopped by the addition of 2 M H,SO, and the absor- bance was read at 450nm by a Titerteck multiskan ELISA reader. Assays were carried out in duplicate.

The recognition of site-1 human IL-6 mutants, triple mutant R30TN31H/I32V, I32 M, F74E, F78E, double mutant F74E/ F78E, R168 M, double mutant K171N/Q175L, Q175L, R179E, R179W, double mutant R179E/R182E, and M184K by the dif- ferent anti-human IL-6 mAbs was tested by competition sand- wich ELISA [29].

Soluble human gpl30-binding assay. The effect of increas- ing concentrations of anti-human IL-6 and anti-human IL-6R mAbs on the association of human IL-6 and soluble human IL-6R with soluble human gp130 was measured as described in [30]. Briefly, avidin precoated microtiter plates (Immunotech) were coated with biotinylated anti-human gp130 monoclonal an- tibody AM-64 (kindly provided by Dr K. Yasukawa) [28, 361.

The each AM-64-coated well was added 50 pl soluble human gp130 (500 ng/ml), 25 pl soluble human IL-6R (50 ng/ ml), 25 p1 human IL-6 (1 pg/ml), and 10 pl M-91-AP as tracer, at 10 U enzyme/ml. Following 6 h incubation at 4°C on an or- bital shaker, the wells were washed three times and phosphatase activity was measured as above. Assays were carried out in du- plicate.

Mapping the epitopes of anti-human IL-6R mAbs. Identi- fication of amino acid residues included in the epitopes of anti- human IL-6R mAbs was carried out using competition sandwich ELISA as described in [27]. Microtiter plates coated with the

Table 1. Characteristics of anti-human IL-6 mAbs. ID,, is the mAb concentration required for half-maximal inhibition of 2 U/ml human IL-6 in the B-9 assay.

mAb k,., K,, (4 "C) Kd (37 "C) ID,,

M-1 s-l M

AH-64 7.1 X lo5 3.7 X 1.7 X 2.3 X lo-'' B-E8 2 . 9 ~ 105 5.9 x 10-12 2 . 2 ~ 1.1 x B-E4 3.6 x los 1.36 X 10-12 6.7 X 1.2 X lo-'' AH-65 5.9 x 10' 4.8 X 10-l2 4.8 X 10-l2 3.3 X lo-''

following: 100 pl/well5 pg/ml of each of the anti-human IL-6R mAbs, which was used as a phase to capture the soluble human IL-6R variants, and the alkaline-phosphatase-conjugated M-91 anti-human IL-6R mAb (M-91-AP, Immunotech) as a tracer to detect bound soluble human IL-6R molecules.

Zn vitro human IL-6 and soluble human IL-6R-binding assays. The potential of the anti-human IL-6 and anti-human IL-6R mAbs to interfere with the binding between human IL-6 and soluble human IL-6R in solution, or the recognition of their complex by AH-65, B-E4, or M-91, was tested as follows. Increasing concentrations of the assayed mAb were added to mixtures containing 85.7 ng/ml human IL-6 and 7.14 ng/ml sol- uble human IL-6R. Reactions were performed in microtiter plates coated with 5 pg/ml anti-human IL-6 mAb AH-65 [22] (Immunotech) or B-E4 to capture cytokine and with 15 U enzyme/ml anti-human IL-6R mAb M-91-AP (Immunotech) to detect human IL-6/soluble human IL-6R complexes, as de- scribed previously [23, 29, 301. Assays were camed out in dupli- cate.

RESULTS

Analysis of anti-human IL-6 mAbs. We tested the cross-reac- tivity between eight anti-human IL-6 mAbs, using competition sandwich ELISA experiments. The results demonstrated that the mAbs can be divided into at least three distinct epitope groups. Group I includes mAbs AH-64, B-E8, and 8, group I1 includes mAbs B-E4, 12, and B2, and group I11 includes mAbs AH-65 and 16. Some competition was also observed between mAbs 12 or B2 from group I1 and mAb 8 from group I.

We measured the kinetic parameters and the dissociation constants at 4°C and 37°C of several of these mAbs (i.e. AH-64, AH-65, B-E4, and B-E8) and analysed their capacity to inhibit the human IL-6-dependent proliferation of B9 cells (Table 1). The results demonstrated that the mAbs are high-affinity human IL-6 binders and potent inhibitors of its bioactivity. Moreover, the findings suggested that mAbs AH-65 and AH-64 were more potent inhibitors than B-E8, which was used in several prelimi- nary clinical trials [4-61. B-E4 seemed to be the less potent inhibitor out of the four (Table 1).

Since anti-human IL-6-neutralising antibodies can interfere with the binding of human IL-6 to IL-6R (site 1) or to gp130 (sites 2 and 3) (Fig. l), we studied the mechanism of inhibition of several of the anti-human IL-6 mAbs in a stepwise approach that followed the different stages of the human IL-6/human IL-6R/human gp130 complex assembly in v i m .

Identification of mAbs that block the binding of human IL-6 to human IL-6R. Assuming that antibodies which interfere with the binding of human IL-6 to human IL-6R shall not be able to immunoprecipitate the intact human IL-6/soluble human IL-6R complex, we identified first those mAbs that can block the for-

Kalai et al. (Eul: J. Biochern. 249) 693

Fig. 1. A schematic representation of the three-dimensional model of human IL-6. Receptor binding sites are colored as follows: yellow (human ILdR, site l), red (gpl3Op1, site 2); green (gp130p2, site 3). The locations of a few selected residues or regions are indicated.

kDa 200 -

97.4 - 69 - 46 - 30 -

shIL-6R

14.5 - z + - + - + - + - hIL-6

+ + + + + + + + shILdR

Anti-hIL-6 mAbs Fig. 2. Identification of anti-human IL-6 mAbs that block the forma- tion of the human IL-6/soluble human ILdR complex. Immunopreci- pitation of '3-labelled soluble human IL-6R by anti-human IL-6 mAbs in the presence or absence of unlabelled human IL-6. An autoradiograph of an SDSPAGE analysis of one out of three repeated experiments, which gave very similar results, is presented.

mation of this complex. As expected, all of the mAbs immuno- precipitated 35S-radiolabelled human IL-6 (data not shown) and none were able to precipitate the 3sS-radiolabelled soluble hu- man IL-6R in the absence of human IL-6 (Fig. 2). mAbs AH-65 and B-E4 precipitated 3sS-soluble human IL-6R in the presence of human IL-6, whereas AH-64 and B-E8 could not (Fig. 2) . These results suggest that AH-64 and B-E8 may recognise resi- dues directly involved in the binding to human IL-6R, whereas AH-65 and B-E4 are most probably not interacting with this site.

AH-65 and B-E4 inhibit the binding of human IL-6 to gp130. Both AH-65 and B-E4 were able to inhibit the biological activity

h w

U AH-65

+ AH-65-F(ab)',

-1- AH-65-Fab

2 100 Y

$ CI 75

50 5 F z B-E4 ' 25

0

B 1 a 0

4 M

2 8 - - e 8 0 mAb (Wml)

Fig. 3. Dose-dependent inhibition of the binding of the human IL-6/ soluble human IL-6R complex to soluble human gp130 by anti-hu- man IL-6 mAbs B-E4, and AH-65 and the F(ab); and Fab prepara- tions of AH-65.

of human IL-6 (Table 1). however they did not block its binding to human IL-6R (Fig. 2). Therefore, we investigated whether they can interfere with the binding to soluble human gp130. Analysis of the effect of increasing concentrations of B-E4 and AH-65 on the binding of human IL-6 to soluble human gp130 in the presence of soluble human IL-6R has demonstrated that both of the mAbs can inhibit the binding to human gp130 (Fig. 3). However, as seen with the biological assay (Table I), mAb AH-65 is about ten times more efficient than mAb B-E4. F(ab')2 and Fab preparations of mAb AH-65 gave very similar results suggesting that the effect of the antibody may be direct and not merely sterical (Fig. 3).

Identification of residues included in the epitopes of several anti-human IL-6 mAbs. The identification of residues included in the epitopes of neutralising antibodies may provide informa- tion not only on their mechanism of action, but also yield details concerning the structure and function of the antigen [25, 26, 30, 371. We used a panel of human IL-6 variants that were produced previously for the identification of residues included in the epi- topes of the anti-human IL-6 studied here. In this method, resi- dues included in the epitope of a given mAb are delineated by its failure to recognise the corresponding site-specific single amino acid substitution variants of the antigenic protein.

mAb AH-65 and mAb 16 interact with site 3, whereas B-E4, B2, and 12 interact with site 2. The human IL-6 variant Q159El T162P and the human/murine-IL-6 chimera-2a were the first to indicate the location of a possible gpl30-binding site on the structure of human IL-6 [26, 381. This combined gpl30-binding epitope is located at the beginning of the AB loop and the D helix, and is often referred to as site 3 (Fig. 1).

We compared the degree of recognition of increasing con- centrations of these site 3 mutants and wild-type human IL-6 by mAbs AH-65, B-E4, and 16. mAb AH-65 bound to the double mutant Q159E/T162P only at very high concentrations (Fig. 4A) and showed a decrease in the recognition of mutant 2a (Fig. 4B). However, mAb B-E4 showed no difficulties in recognising the double mutant Q159E/T162P and the triple mutant 2a/Q159E/ T162P. As shown before, mAb 16 hardly reacted with these mu- tants (Fig. 4A and B) [26, 391.

The results indicated that the epitope recognised by AH-65 is shared with mAb 16 and that both may block the binding of gp130p2 to site 3. They also suggested that B-E4 might interfere with the binding of gpl30pl to site 2. This gpl30-binding epi- tope may be disrupted by the substitution of Tyr31 and Val121

694 Kalai et al. (Eur: J . Biochem. 249)

A + w - QI59E.Tl62P

IL-6 (ndmlj IL-6 (ng/ml) IL-6 (ng/ml)

B - w T -+- 2a -c-- 2a,Q159E,T162P

+ WT - DFRD

Fig.4. Differential recognition of site 3 and site 2 human IL-6 mutants by several anti-human IL-6 m.- .s determined by sandv .-.. ELISA. (A) Reactivity of mAbs AH-65, B-E4, and 16 with site 3 human IL-6 mutant Q159E/T162P. (B) Reactivity of mAbs AH-65, B-E4, and 16 with site 3 human/murine-IL-6-chimera mutant 2a, and triple mutant 2dQ159E/T162P. (Cj Reactivity of mAbs B-E4, B2, and 12 with site 2 human IL-6 mutant DFRD.

by aspartic acid, Gly35 by phenylalanine, and Ser l l8 by argi- nine, resulting in an antagonistic human IL-6 variant called DFRD (Fig. 1) [31].

We produced the DFRD mutant and tested its recognition by B-E4, and several other mAbs. mAbs B-E8, 8, and 16, which were tested as controls, recognised the mutant (data not shown). B-E4, and mAbs included in its epitope group, i.e. B2 and 12 (Table 1), hardly recognised the DFRD mutant (Fig. 4C). Of all of the inAbs studied here, these three mAbs were the only ones to show a significantly decreased binding of the human 1L-6 deletion variant 426, in which the first 26 residues starting with the N-terminal of the mature form of the protein were removed (data not shown). These results indicate that B-E4 interferes most probably with gpl30-binding by blocking site 2 and sug- gest that B2 and 12 may act in a similar way. Residues included in their epitopes are most probably located in the A helix and may also be in the C helix.

Anti-human IL-6 mAbs AH-64, B-E8, a d 8 recognise amino acid residues crucial for the binding to human IL-6R. mAbs B-E8, AH-64, and 8 can inhibit the binding between human IL- 6 and human IL-6R (Fig. 2 B [26]). This suggested that they may recognise residues included in site 1 or located in its proximity (Fig. 1). Therefore, we tested the binding of these mAbs to a panel of human IL-6 mutants in which residues predicted to re- side i n the human IL-6R-binding epitope were altered [29].

B-E8 did not bind to any of the Arg179 mutants although it recognised all of the other mutants (Fig. 5) . AH-64 failed to recognise mutants of Arg179 and showed a significant decrease

in the recognition of mutants of Arg168, Phe74, and Phe78 (Fig. 5). inAb 8 failed to recognise mutant F78E, double mutants F74E/F78E and R179E/R182E, and showed a significant decrease in the recognition of mutants of Arg179, indicating that its epitope covers the ends of both the AB loop and the D helix (Fig. 5). These results correlate well with the cross-reactivity ob- served between these three mAbs and suggest that they have overlapping but distinct epitopes (Table 1, Fig. 1).

Our results suggest that the identification of the epitopes of these mAbs by other techniques such as a peptide scan would have been difficult since they seem to be discontinuous and some of the residues they include are very distant in the primary structure.

Analysis of anti-human IL8R mAbs. The 14 anti-human IL-6R mAbs studied here are directed against the extracellular domain of human IL-6R and recognise the soluble form of the receptor subunit. All of the antibodies except M-91 and 22.1 are capable of inhibiting the biological activity of human IL-6R in cell culture assays [7,9,40-421. M-91 was suggested to interact with the Ig-like domain of the receptor. Competition was ob- served between mAbs M-37, M-113, M-164, 32.3, 34.4, 38.8, 48.8, and 50.6, and also between mAbs M-195 and PM-1. No competition was observed between mAbs 17.6, 22.1, and M-182 and any other of the anti-human IL-6R mAbs 17, 401.

Structure/function studies of human IL-6R have suggested that several distinct regions in the molecule are interacting with other molecules in the high-affinity human 1L-6 receptor com-

I Helix A I I

Kalai et al. (Eur J. Biochem. 249)

AB loop I Helix D 1 Phase I Tracer

100

50

0 150

n

F b 50

O O &? 150

looj 0 .. 1 01 . . . . 50 .:.: .:.: .. ,. . . . . .. . . ,. . . . . .. . . .. . . . . ,. . . . . . . . . . 0 " " s Z i 3 z n

5 % B p?

I I

hIL-6 mutant

695

Fig.5. Differential recognition of site 1 human IL-6 mutants by anti-human IL-6 mAbs B-E8, 8, and AH-64 determine by sandwich ELIS(A). The concentrations of equal amounts of each human IL-6 variant were determined using solid phase coated with mAbs AH-65, B-E8, 8, or B-E4, and biotinylated anti-human IL-6 mAb B-E4 or AH-64 as tracers as indicated. Results are presented as percentages of the values obvained for wild-type (WT) human IL-6. Values are the mean? SD of at least two independent experiments carried out in duplicate. (ND, not determined).

A 200 - 97.4 -

69 - 46 - 30 -

Anti-hIL-6R mAbs 6

200 - 97.4 -

69 - 46 - 30 -

> hIL-6

> hIL-6

14.5 - shIL-6R + - + - + - + - + - + - + - - - r! m ""$?$Z?? V Z $ d 9 2 x x m m b * m I o

Anti-hILdR mAbs Fig. 6. Identification of anti-human IL-6R mAbs that block the formation of the human IL-6/soluble human IL-6R complex. Immunoprecipita- tion of '5S-labelled human IL-6 by anti-human IL-6R mAbs in the presence or absence of unlabelled soluble human IL-6R. An autoradiograph of an SDSPAGE analysis of one out of three repeated experiments that gave very similar results is presented.

696 Kalai et al. (ELM J. Biochern. 249)

Fig. 7. A schematic representation of the three-dimensional model of human IL-6R. The locations of a few selected residues and regions are indicated. Residues included in the human IL-6 and human gp130p1 binding sites are colored yellow and red, respectively. The green circle indicates the predicted location of the C'D' loop in domain 11.

plex. Some are involved in human IL-6 binding [29], and others interact with gp130 [43, 441.

We decided to first identify mAbs that block binding to hu- man IL-6. Therefore, we tested their ability to immunoprecipi- tate the intact human IL-6/soluble human IL-6R complex.

mAbs M-37, M-113, M-164, M-195, PM-I, and 48.8 can block the binding of human IL-6 to human IL-6R. All of the mAbs immunoprecipitated the "S-soluble human IL-6R (data not shown) and none were able to precipitate the 35S-human IL-6 in the absence of soluble human IL-6R (Fig. 6A and B). About eight of the anti-human IL-6R mAbs were able to co-immuno- precipitate the human 35S-IL-6 in the presence of soluble human IL-6R (Fig. 6A and B). Two of these, i.e. 32.3 and 50.6 (Fig. 6B), seemed to interfere at least partially with the binding of human 1L-6 to human IL-6R. The most potent inhibitors were mAbs M-37, M-113, M-164, M-195, PM-1, and 48.8 (Fig. 6A and B).

Identification of residues included in the epitopes of several anti-human IL-6R mAbs. Preliminary attempts to map the epi- topes of anti-human IL-6R mAbs 17.6, 22.1, 32.3, 34.4, 38.8, 48.8, 50.6 using linear synthetic peptides were only partially successful and suggested that mAbs 34.4 and 50.6 may bind to peptides corresponding to residues 257 -266 and 25 1 -260 of human L 6 R , respectively (Fig. 7) [45]. Attempts to map the epitopes of mAbs M-37, M-91, M-113, M-164, M-182, and M-195 using a similar approach were unsuccessful [41].

Here, we used 15 different soluble human IL-6R mutants representing ten distinct amino acid residues predicted to reside in the human IL-6-binding epitope of the receptor. The recogni- tion of each of the mutants by each of the anti-human IL-6R

mAbs was compared with the recognition of the wild-type solu- ble human IL-6R (Fig. 8). mAbs 17.6 and 22.1 recognised all of the mutants and the wild-type. The recognition of mutant V190G by most of the mAbs was poor suggesting that Val190 is impor- tant for the conservation of the structural integrity of the mole- cule at least locally (Fig. 8).

mAbs M-37, M-113, and M-164 hardly recognised mutants of Ser186, Tyr188, and Val190 (Fig. 8A). The recognition of these mutants by mAbs 50.6, 38.8, 34.4, and 32.3 was less af- fected (Fig. 8B). Among the latter, the effect on mAb 50.6 was the strongest (Fig. 8B). These results suggest that Ser186 and Tyr188 and perhaps Val190 are included or are in the vicinity of the epitopes interacting with mAbs M-37, M-113, M-164, and probably also 50.6 (Fig. 7). mAbs 32.3, 34.4, 38.8, and 48.8 were much less affected by the substitution of these residues and were also less efficient in inhibiting the binding between human IL-6 and soluble human IL-6R (Figs 8B and 6B). Since some competition exists between the two groups, we suggest that they recognise two distinct but close epitopes.

M-195 and PM-1 showed a slight decrease in the recognition of mutants of Tyr249 and a strong decrease in the recognition of mutants of glutamic acid at positions 296 and 297 (Fig. 8A).

mAbs M-182, 34.4, and 38.8 are predicted to interfere with the binding of the human IL-6/human IL-6R complex to gp130p2. Some of the anti-human IL-6R mAbs did not block the binding between human IL-6 and human IL-6R (Fig. 6). Nevertheless, they might inhibit the binding to gp130. This was shown recently to be the case of mAb M-182 [41]. It is not known whether this neutralising mAb interferes with the binding to gpl30pl or gp130p2. We searched for a rapid way to distin- guish between the two possibilities.

Two different studies have demonstrated that direct interac- tions occur between human IL-6R and gpl30pl [43,44]. So far, no evidence indicating that such interactions also occur between human IL-6R and gp130p2 has been provided. Determining if such contacts exist and the identification of their location may be important for understanding how the high-affinity IL-6 receptor complex is assembled. Since crystal and NMR structures of this complex are unavailable, any additional information concerning contact regions with gp130p2 is of extreme importance for the construction of the corresponding three-dimensional model. Such a model may be useful for the study of the high-affinity receptor complexes of other IL-6-type cytokines.

Discrimination between interference with the binding to gp130p1 or to gp130p2 is not a simple task. As demonstrated above, the anti-human IL-6 mAbs B-E4, and AH-65 interact with residues included in site 2 and site 3 on human IL-6, and interfere with the binding to gpl30pl and gp130p2, respectively. Therefore, we used the anti-site 2 mAb B-E4 to play the role of gpl30pl and the anti-site 3 mAb AH-65 as a substitute for gpl30p2. Using these mAbs, we searched for anti-human IL-6R mAbs that would possibly interfere with the binding of the hu- man IL-6/soluble human IL-6R complex to one of the two gp130 subunits.

The formation of human IL-6/soluble human IL-6R com- plexes and their capture by a solid phase coated with either anti- human IL-6 mAb AH-65 or mAb B-E4, in the presence and absence of various concentrations of each of the different anti- human IL-6R mAbs, was monitored by an alkaline-phosphatase (AP)-conjugated anti-human IL-6R mAb M-91.

In such an experimental system, mAbs that interfere with the binding between human IL-6 and human IL-6R will decrease the detection of their complex in both assays in a similar manner. Similar results may be observed if the additional mAb interferes with the binding of the complex by the tracer mAb M-91-AP. A decrease in the binding of the human IL-6/human IL-6R com-

Kalai et al. (Eul: J. Biochem. 249) 697

1504 22.1 I 100

50

0 I I

1504 M-37 - I 100

50

0

1504 I 100

50

0

1501 M-164 I

1504 17.6

50

0 I

1504 2 2 . 3 I 50

0 I I I h 1504 34.4

k 100

50 50 0 0

1501 M-182 I 100

50

0 I 1

IW

50

0

shIL-6R

I v M 1504 38.8 fi a .: 50

.I

I

1504 48.8 I 1W

50

0

1504 50.6 I

shIL-6R Fig. 8. Identification of residues included in the epitopes of anti-human ILdR mAbs by sandwich ELISA and soluble human ILBR mutants. Amino acid substitutions are indicated by the single-letter code. Results are presented as percentages of the values obtained with the wild-type (WT) soluble human IL-6R. Values are the mean of 2-4 independent experiments carried out in duplicate.

plex by one of the phase mAbs and not the other would suggest that the tested mAb interferes also with the binding to the corre- sponding gp130 (i.e. interference with B-E4 is predicted to decrease the binding to gpl30P1, whereas, interference with AH-65 is predicted to inhibit the interaction with gp130p2).

The results obtained with mAbs M-37, M-164, M-195, PM-1, 32.3, and 48.8 correlated well with expectations based on the immunoprecipitation experiments (Fig. 6). All of these mAbs inhibited the formation of the human IL-6/human IL-6R complex in a concentration-dependent manner in both of the ex- perimental systems. The results obtained for M-195 are shown as an example (Fig. 9).

Anti-human gp130 mAb AM-66 which was used as a nega- tive control did not affect the recognition or the formation of the complex (data not shown). Several other mAbs, which were able to immunoprecipitate efficiently radiolabelled human IL-6 in the presence of soluble human TL-6R (Fig. 6) nevertheless affected significantly the recognition of the human IL-6/soluble human IL-6R complex in one or both of the experimental systems. mAb 17.6 showed inhibitory effects even at low concentrations (data not shown). This interference was independent of the anti-hu- man IL-6 used in the assay and was dose dependent only at

low concentrations. These results might imply that mAb 17.6 interferes with the binding of human IL-6/soluble human IL-6R complex to the tracer M-91-AP. The effect of mAb 22.1 was minor in both cases though somewhat stronger when the phase mAb was AH-65 (data not shown).

The results obtained for mAbs M-182, 34.4, and 38.8 corre- lated better to those obtained previously (Fig. 6) when the solid phase was coated with mAb B-E4 (Fig. 9). When the solid phase was coated with AH-65 these monoclonal antibodies were inhib- itory. The degree of inhibition varied between high for mAbs 38.8 and 34.4, and low for mAb M-182 (Fig. 9). It seems there- fore, that the binding of these anti-human IL-6R mAbs to the human IL-6/soluble IL-6R complex interferes with its capture by the anti-human IL-6 mAb AH-65 probably because of steric hindrance.

These results suggest that mAbs 34.4,38.8, M-182, and 22.1 may also interfere with the binding of the human IL-6/soluble human IL-6R complex to gpl30m. Analysis of their effect on the binding to sgpl30 indicated that mAbs 34.4, and 38.8 are more efficient inhibitors of the binding between the human IL-6/ soluble human IL-6R complex and soluble gp130 than mAb M-182 (Fig. 10).

698 Kalai et al. ( E m J. Biochem. 249)

M-195 125{

34.4 38.8 125

mAb (ng/ml)

Fig. 9. Intereference of anti-human IL-6R mAbs with the formation of the human IL-6/soluble human IL-6R complex or with its detec- tion by anti-human IL-6 mAbs AH-65 (B) or B-E4 (0) and anti- human IL-6R mAb M-91.

- 2 V . V

* M-182

mAb (Fg/ml)

Fig. 10. Dose-dependent inhibition of the binding of the human IL-6/ soluble human IL-6R complex to soluble human gp130 by anti-hu- man ILdR mAbs 34.4, 38.8, and M-182.

DISCUSSION

Anti-human IL-6 mAbs. The present study and previous results show that anti-human IL-6 mAbs AH-64, B-E8, 8, 12, B2, B-E4, AH-65, and 16 are all potent inhibitors of human IL-6 bioactiv- ity. These antibodies can be divided into at least three distinct epitope groups.

mAbs AH-64, B-E8, and 8 interact with residues included in site 1 (Fig. 1) and block the binding of human IL-6 to human IL-6R. mAbs 12, B2, and B-E4 interact with residues included in site 2 (Fig. 1) and probably block the binding to gpl30pl. mAbs AH-65, and 16 interact with residues included in site 3 (Fig. 1) and block the binding to gpl3Op2. These results support all previous findings concerning the location of the different

receptor binding sites on the human IL-6 structure and the residues they include [26, 29, 38, 39, 461.

The data we present may help in choosing which mAbs would be most beneficial in the treatment of human IL-6-related disease and in which combinations. It is suggested that the anti- bodies used in such a treatment will include representatives of each of the three epitope groups. To inhibit efficiently the bioac- tivity of human IL-6 such mAb cocktails could include the following: mAbs AH-64 or 8 to block site 1, mAbs B-E4 or B2 to block site 2 and mAbs AH-65 or 16 to block site 3.

Anti-human IL-6R mAbs. We compared many anti-human IL-6R mAbs to facilitate the choice of mAb combinations to use in anti-human IL-6R clinical treatment. Our findings and previ- ous results demonstrate that the antibodies can be divided into at least six distinct epitope groups.

Our results show that the anti-human IL-6R mAbs M-37, M-113, M164, M-195, and PM-1 are strong inhibitors of the binding of human IL-6 by human IL-6R, whereas mAbs 32.3, 48.8, and 50.6 affect the reaction to a lesser degree. They sug- gest that residues Ser186, Tyr188, and Val190 are located in the epitope recognised by anti-human IL-6R mAbs M-37, M-113, M-164, and 50.6 or in its vicinity, and that residues Glu296, Glu297, and Tyr249 may interact with mAbs PM-1, and M-195 (Fig. 7). The results in this study support further our previous suggestion that these residues are exposed on the surface of the human IL-6R molecule and that Tyr188, Tyr249, Glu296, Glu297, and maybe also Val190 are directly interacting with hu- man IL-6 [29, 301. Moreover they suggest that the EF, B’C’, and F’G‘ loops of human IL-6R are immunogenic.

The results concerning mAbs M-182, 17.6, 22.1, 32.3, 34.4, 38.8, and 48 are less direct. We have shown here that mAbs M-182, 17.6, 22.1, 34.4, and 38.8 do not inhibit the binding between human IL-6 and human IL-6R. mAb 17.6 showed a slight competition with M-91 for the binding of the human IL-6/ human IL-6R complex. This effect was not observed with human IL-6R alone. It is therefore suggested that mAb 17.6 rec- ognises an epitope distinct from the one interacting with mAb M-91, although it is probably close enough to allow for steric hindrance.

We showed that mAbs 34.4, 38.8, and M-182 inhibit the binding of the human IL-Ghurnan IL-6R complex to gp130. Since they interfered also with the binding of this complex to anti human IL-6 mAb AH-65, which was shown here to recog- nise residues included in the human IL-6 site 3 (Fig. 4), we sug- gest that like AH-65, these mAbs interfere with the binding to gp130p2.

Interestingly, the region recognised by mAb 34.4 was iden- tified recently by a peptide scan. The sequence common to the four peptides that it interacts with corresponds to residues Tyr257-Thr266 in the C’D’ loop of domain I1 of human IL-6R (Fig. 7, green circle). Moreover, the peptide that reacts best with mAb 34.4 was shown to inhibit the activity of human IL-6 in different biological assays and was suggested to interfere with the binding to gp130 [45].

In our three-dimensional model of the human IL-6/human IL-6Wuman gpl3Opl complex [47], the corresponding region is exposed on the surface of the human IL-6R with an orienta- tion similar to that of site 3 in human IL-6 (Figs l and 7). Inter- estingly, the alteration of several residues included in the C’D’ loop was shown to reduce significantly the bioactivity of the corresponding soluble human IL-6R variants (i.e.double mutant E260A/R261G and triple mutant R261G/S262A/K263G), but did not affect their human IL-6-binding capacity [43] Taken to- gether, these results suggest that residues included in the C’D’ loop of human IL-6R are interacting with gp130p. Further,

Kalai et al. (Eui: J. Biochem. 249) 699

human IL-6R mutagenesis studies could verify this prediction. This information may help the construction of a three-dimen- sional model of the human IL-6 high-affinity receptor complex.

In analogy with the anti-human IL-6 mAbs, combining anti- bodies such as M-37 or M-164 with M-195 or PM-1, and 34.4, 38.8, or M-182 in the treatment of IL-6-related disease may also result in an efficient inhibition of the human IL-6 activity and rapid clearance of soluble human IL-6R.

We thank Dr D. Novick, Dr J. Brochier, Dr J. Wijdenes, and Dr K. Yasukawa for anti-human IL-6R, anti-human IL-6, and anti-human gp130 mAbs which were used in this study. We thank Dr G. Muller- Newen for critical reading. This work was supported by grants from the CGER (Caisse GCnCrale d'Epargne et de Retraite) Fonds de Recherche Cance'rologique and the Fund for Medical Scientific Research (Belgium).

REFERENCES 1. Akira, S., Taga, T. & Kishimoto, T. (1993) Interleukin-6 in biology

and medicine, Adv. Immunol. 54, 1-78. 2. Murakami, M., Hibi, M., Nakagawa, N., Nakagawa, T., Yasukawa,

K., Yamanishi, K., Taga, T. & Kishimoto, T. (1993) IL-6-induced homodimerization of gp-I30 and associated activation of a tyro- sine kinase, Science 260, 1808-1810.

3. Klein, B., Wijdenes, J., Zhang, X.-G., Jourdan, M., Boiron, J.-M., Brochier, J., Mannoni, P., Sany, J. & Bataille, R. (1991) Murine anti-interleukin-6 monoclonal antibody therapy for a patient with plasma cell leukemia, Blood 78, 1198-1204.

4. Wijdenes, J., Racadot, E. & Wendling, D. (1994) interleukin-6 anti- bodies in rheumatoid arthritis, J. Int. Res. 14, 297-298.

5. Brochier, J., Legouffe, E., Liautard, J., Gaillard, J. P., Mao, L. Q,, Bataille, R., Rossi, J. F. & Klein, B. (1995) Immunomodulating IL-6 activity by murine monoclonal antibodies, Int. J. Immuno- pharmacol. 17,41-48.

6. Emilie, D., Wijdenes, J., Gisselbrecht, C., Jarrousse, B., Billaud, E., Blay, J.-Y., Gabarre, J., Gaillard, J.-P., Brochier, J., Raphael, M., Boue, F. & Galanaud, P. (1994) Administration of an anti-in- terleukin-6 monoclonal antibody to patients with acquired immu- nodeficiency syndrome and lymphoma: effect on lymphoma growth and on B clinical symptoms, Blood 84, 2472-2479.

7. Liautard, J., Gaillard, J. P., Mani, J. C., Montero-Julian, F., Duper- ray, C., Lu, Z . y., Jourdan, M., Klein, B., Brailly, H. & Brochier, J. (1994) Epitope analysis of human IL-6 receptor gp80 molecule with monoclonal antibodies, Eur: Cytokine Nem. 5, 293-300.

8. Daveau, M., Liautard, J., Gaillard, J. P., Hiron, M., Brochier, J. & Lebreton, J. P. (1994) IL-6-induced changes in synthesis of a 1-acid glycoprotein in human hepatoma Hep3B cells are distinc- tively regulated by monoclonal antibodies directed against dif- ferent epitopes of IL-6 receptor (gp80), Eui: Cytokine Netw 5, 601-608.

9. Novick, D., Shulman, L. M., Chen, L. & Revel, M. (1992) Enhance- ment of interleukin-6 cytostatic effect on human breast carcinoma cells by soluble IL-6 receptor from urine and reversion by mo- noclonal antibody, Cytokine 4, 6- l l .

10. Suzuki, H., Yasukawa, K., Saito, T., Goitsuka, A,, Hasegawa, A., Ohsugi, Y., Taga, T. & Kishimoto, T. (1992) Anti-human inter- leukin-6 receptor antibody inibits human myeloma growth in vivo, Eur: J. Immunol. 22, 1989-1993.

11. Taga, T., Narazaki, M., Yasukawa, K., Saito, T., Miki, D., Ham- guchi, M., Davis, s., Shoyad, M., Yancopoulos, G. D. & Kishi- moto, T. (1992) Functional inhibition of hematopoietic and neuro- trophic cytokines by blocking the interleukin 6 signal transducer gp130, Proc. Natl Acad. Sci. USA 89, 10998-11001.

12. Sato, K., Tsuchiya, M., Saldanha, J., Koishihara, Y., Ohsugi, Y., Kishimoto, T. & Bendig, M. M. (1993) Reshaping a human anti- body to inhibit the interleukin-6-dependent tumor cell growth, Cancer Res. 53, 851-856.

13. Sato, K., Tsuchiya, M., Saldanha, J., Koishihara, Y., Ohsugi, Y., Kishimoto, T. & Bendig, M. M. (1994) Humanization of a mouse anti-human interleukin-6 receptor antibody comparing two meth-

ods for selecting human framework regions, Mol. Immunol. 31,

14. Kishimoto, T. (1996) IL-6 family of cytokines and gpl30-biological and clinical implications, EUK Cytokine Netw. 7, 429, abstract no. 8.

15. Lu, Z. Y., Brochier, J., Wijdenes, J., Brailly, H., Bataille, R. & Klein, B. (1992) High amounts of circulating interleukin (1L)-6 in the form of monomeric immune complexes during anti-IL-6 therapy. Towards a new methodology for measuring overall cytokine pro- duction in human in vivo, Eui: J. Immunol. 22, 2819-2824.

16. Lu, Z. Y., Brailly, H., Wijdenes, J., Bataille, R., Rossi, J.-F. & Klein, B. (1995) Measurment of whole body interleukin-6 (IL-6) produc- tion: prediction of the efficacy of anti-IL-6 treatments, Blood 86,

17. Klein, B. & Brailly, H. (1995) Cytokine-binding proteins: stimulat- ing antagonists, Immunol. Today 16, 216-220.

18. Lavabre-Bertrand, T., Exbrayat, C., Liautard, J., Gailard, J.-P., Baskevitch, P.-P., Poujol, N., Duperray, C., Bourquard, P. & Brochier, J. (1995) Detection of membrane and soluble interleu- kin-6 receptor in lymphoid malignancies, Brit. J . Haematol. 91,

19. De Benedetti, F., Massa, M., Pignatti, P., Albani, S., Novick, D. & Martini, A. (1994) Serum soluble interleukin 6 (IL-6) receptor and IL-6/soluble IL-6 receptor complex in systemic juvenile rheu- matoid arthritis, J. Clin. Invest. 93, 2114-2119.

20. Montero-Julian, F. A,, Gautherot, E., Wijdenes, J., Klein, B. & Brailly, H. (1994) Pharmacokinetics of interleukin-6 during ther- apy with anti-interleukin-6 monoclonal antibodies : enhanced clearence of interleukin-6 by combination of anti-interleukin-6 antibodies, J. Int. Res. 14, 301-302.

21. Montero-Julian, F. A., Klein, B., Gautherot, E. & Brailly, H. (1995) Pharmacokinetic study of anti-interleukin-6 (IL-6) therapy with monoclonal antibodies: enhancment of IL-6 clearance by cock- tails of anti-IL-6 antibodies, Blood 85, 917-924.

22. Brailly, H., Montero-Julian, F. A,, Zuber, C. E., Flavetta, S., Grassi, J., Houssiau, F. & Van Snick, J. (1994) Total interleukin-6 in plasma measured by immunoassay, Clin. Chem. 40, 116-123.

23. Montero-Julian, F. A,, Liautard, J., Flavetta, S., Romagne, F., Gail- lard, J. P., Brochier, J., Klein., B. & Brailly, H. (1994) Immuno- assay for functional human soluble interleukin-6 receptor in plasma based on ligandlreceptor interactions, J , Immunol. Meth-

24. Wijdenes, J., Clement, C., Klein, B., More-fourier, B., Vita, N., Fenara, P. & Peters, A. (1991) Human recombinant dimeric IL-6 binds to its receptor as detected by anti-IL-6 monoclonal anti- bodies, Mol. Immunol. 28, 1183-1192.

25. Brakenhoff, J. P. J., Hart, M., De Groot, E. R., Di Padova, F. & Aarden, L. A. (1990) Structure-function analysis of human IL-6. Epitope mapping of neutralizing monoclonal antibodies with amino- and carboxyl-terminal deletion mutants, J. Immunol. 145,

26. Brakenhoff, J., de Hon, F. D., Fontaine, V., Tenboekel, E., School- tink, H., Rose-John, s., Heinrich, P. C., Content, J. & Aarden, L. A. (1994) Development of a human interleukin-6 receptor antago- nist, J. Biol. Chem. 269, 86-93.

27. Novick, D., Engelmann, H., Revel, M., Leitner, 0. & Rubinstein, M. (1991) Monoclonal antibodies to the soluble human IL-6 re- ceptor: affinity purification, ELISA, and inhibition of ligand bind- ing, Hybridoma 10, 137-146.

28. Yasukawa, K., Futatsugi, K., Saito, T., Yawata, H., Narazaki, M., Suzuki, H., Taga, T. & Kishimoto, T. (1992) Association of re- combinant soluble IL-6-signal transducer, gp130, with a complex of IL-6-receptor and soluble IL-6- receptor, and establishment of an ELISA for soluble gp130, Immunol. Lett. 31, 123-130.

29. Kalai, M., Montero-Julian, F. A,, Grotzinger, J., Fontaine, V., Paul, V., Deschuyteneer, R., Wollmer, A,, Brailly, H. & Content, J. (1997) Analysis of the hIL-6/hIL-6R-binding interface at the amino acid level : proposed mechanisms of interaction, Blood 89,

30. Kalai, M., Montero-Julian, F. A,, Grotzinger, J . , Wollmer, A,, Mo- relle, D., Brochier, J., Rose-John, S., Heinrich, P. C., Brailly, H. & Content, J. (1 996) Participation of two Ser-Ser-Phe-Tyr repeats in IL-6-binding sites of human interleukin-6 receptor, Eui: J . Bio- chem. 238, 714-723.

371 -381.

3123- 3131.

871 - 877.

ods 169, 111-121.

561 -568.

1319-1333.

700 Kalai et al. (Eur. J. Biochem. 249)

31. Savino, R., Ciapponi, L., Lahm, A,, Demartis, A,, Cabibbo, A., Toniatti, C., Delmastro, P., Altamura, S. & Ciliberto, G . (1994) Rational design of a receptor super-antagonist of human interleu- kin-6, EMBO J. 13, 5863-5870.

32. de Hon, F. D., Ehlers, M., Rose-John, S., Ebeling, S. B., Klaasse Bos, H., Aarden, L. A. & Brakenhoff, J. P. J. (1994) Development of an interleukin (IL) 6 receptor antagonist that inhibits IL-6- dependent growth of human myeloma cells, J. Exp. Med. 180,

33. Brailly, H., Yasukawa, K., Montero-Julian, F. A,, Nakano, K. & Klein. B. (1994) Soluble IL-6RP is present at high concentration in normal serum as a dimer that binds efficiently to IL-6AL-6Ra complexes, Leucocyte Typing V; 13 - 17.

34. Content, J., De Wit, L., Pierard, D., Derynck, R., De Clercq, E. & Fiers, W. (1982) Secretory proteins induced in human fibroblasts under conditions used for the production of interferon @, Proc. Natl Acad. Sci. USA 79, 2768-2772.

35. Brakenhoff, J., De Hon, F. D., Fontaine, V., Hart, M., De Groot, E. R., Content, J. & Aarden, A. (1992) in IL-6 physiopathology and clinical potentials. Serono symposium on 1L-6 (Revel, M., ed.) vol. 88, pp. 33-41, Raven Press, NY.

36. Hibi, M., Murakami, M., Saito, M., Hirano, T., Taga, T. & Kishimoto, T. (1990) Molecular cloning and expression of an IL-6 signal transducer, gp130, Cell 63, 1149-1157.

37. Fontaine, V., Savino, R., Arcone, R., De Wit, L., Brakenhoff, J. P. J., Content, J. & Ciliberto, G. (1993) Involvement of the Arg179 in the active site of human IL-6, Eul: J. Biochem. 211, 749-755.

38. Ehlers, M., Grotzinger, J., de Hon, F. D., Mullberg, J., Brakenhoff, J., Liu, J. W., Wollmer, A. & Rose-John, S. (1994) Identification of two novel regions of human IL-6 responsible for receptor bind- ing and signal transduction, J. Imrnunol. 153, 1744- 1753.

39. de Hon, F. D., ten Boekel, E., Herrman, J., Clement, C., Ehlers, M., Taga, T., Yasukawa, K., Ohsugi, Y., Kishimoto, T., Rose-John, S., Wijdenes, J., Aarden, L. A. & Brakenhoff, J. P. J. (1995) Func-

2395 -2400.

tional distinction of two regions of human interleukin 6 important for signal transduction, Cytokine 7, 398-407.

40. Gaillard, J. P., Liautard, J., Duperray, C. & Brochier, J. (1993) Analysis of monoclonal antibodies (mAb) against human gp80 interleukin-6 receptor (IL-6R), Tissue Antigens 42, 330.

41. Gaillard, J.-I?, Liautard, J., Mani, J. C., Fernandez Suarez, J. M., Klein, B. & Brochier, J. (1996) Identification of a novel antigenic structure of the human receptor for interleukin-6 involved in the interaction with the glycoprotein 130 chain, Immunology 89,

42. Hirata, Y., Taga, T., Hibi, M., Nakano, N., Hirano, T. & Kishimoto, T. (1989) Characterization of IL-6 receptor expression by mo- noclonal and polyclonal antibodies, J. Immunol. 143,2900-2906.

43. Yawata, H., Yasukawa, K., Natsuka, S., Murakami, M., Yamasaki, K., Hibi, M., Taga, T. & Kishimoto, T. (1993) Structure-function analysis of human IL-6 receptor - dissociation of amino acid residues required for IL-6- binding and for IL-6 signal transduc- tion through gp130, EMBO J. 12, 1705-1712.

44. Salvati, A. L., Lahm, A., Paonessa, G., Ciliberto, G . & Toniatti, C. (1995) interleukin-6 (IL-6) antagonism by soluble IL-6 receptor a mutated in the predicted gp130-binding interface, J. Biol. Chem.

45. Halimi, H., Eisenstein, M., Oh, J.-W., Revel, M. & Chebath, J. (1995) Epitope peptides from interleukin-6 receptor which inhibit the growth of human myeloma cells, Eur. Cytokine Netw. 6,135 - 143.

46. Ciapponi, L., Graziani, R., Paonessa, G., Lahm, A,, Ciliberto, G. & Savino, R. (1995) Definition of a composite binding site for gp130 in human interleukin-6, J. Biol. Chem. 270, 31249- 31 254.

47. Grotzinger, J., Kurapkat, G., Wollmer, A., Kalai, M. & Rose-John, S. (1997) The family of the IL-6-type cytokines: specificity and promiscuity of the receptor complexes, Profeins Struct. Funct. Genet. 27, 96-109.

135- 141.

270, 12 242- 12 249.