Minimal tumor necrosis factor receptor binding protein: optimum biological activity of a truncated p55 soluble tumor necrosis factor receptor-IgG fusion protein.

ARTICLE

INTRODUCTION

The pleiotropic cytokine, tumour necrosis factor-alpha (TNF-alpha), produced primarily by activated macrophages, is intimately involved in many immune and inflammatory responses including septic shock, cachexia, and rheumatoid arthritis (RA) [1-5]. TNF interacts with two high affinity cell surface receptors, which also bind the structurally related cytokine, lymphotoxin (LT) [6]. Two human soluble TNF binding proteins, termed TBP I and TBP II, are homologous to the extracellular domains (ECDs) of the p55 and p75 TNF receptors respectively [7, 8], and are generated by proteolytic cleavage of the mature cell surface receptor [9]. Their levels are elevated in a number of disease states such as RA [10] and cancer [11], and thus they may play an important regulatory role in TNF-mediated inflammatory processes [10].

A number of studies have investigated the possibility of ameliorating TNF-related inflammatory diseases using anti-TNF neutralizing antibodies or soluble TNF receptors. Treatment with an anti-TNF antibody caused marked reduction in inflammation in collagen-induced arthritis in mice [12]. Four to twelve week, double-blind clinical trials of humanized anti-TNF antibodies similarly produced dramatic reductions in disease pathology in RA [13-15] and Crohn's disease [16, 17]. However, in longer term (up to 95 weeks) dose response trials, half of the patients produced human anti-chimeric antibody (HACA) responses [18], casting doubt on the suitability of these agents for long-term treatment of some patients.

Monomeric soluble receptors bind TNF in vitro, but only at very high molar ratios, doses which would not be feasible for therapeutic applications [19, 20]. We and others have found that dimers of both the p55 and p75 soluble TNF receptors fused to human IgG heavy chains markedly increased avidity for TNF in vitro compared with monomeric soluble receptors [19-22]. These fusion proteins have been shown to protect against lipopolysaccharide (LPS) and TNF-induced lethality in animal models [19, 21-23]. The much lower molar ratios required suggest that they may be as good as anti-TNF at combatting the harmful effects of TNF. Furthermore, they contain no murine variable regions and are thus likely to be less immunogenic than anti-TNF antibodies. This makes them an exciting prospect for clinical trials in diseases such as RA. Indeed, a recent three month, double blind trial of a human p75 soluble TNF-R-IgG fusion protein reported significant improvement in rheumatoid arthritis patients, with no side effects [24].

The TNF receptor proteins possess homologous ECDs, consisting of four repeating cysteine-rich motifs, conferring a high degree of structural homology between the receptors [25, 26]. The ECDs of three TNF receptors bind at the three corners of a TNF or LT trimer [27]. We have previously established that deletion of the fourth (membrane proximal) cysteine-rich repeat from the ECD of the p55 TNF receptor has no effect on TNF binding by either the soluble or membrane bound receptor [28]. Thus the fourth repeat is not involved in TNF binding. In contrast, each of the first three repeats is essential for optimal TNF binding [28]. These results concur with the reported crystallographic structure of a p55 soluble TNF receptor/LT complex, which predicts that the fourth repeat is spatially removed from the LT trimer [27].

The aim of this study was to construct dimers of the truncated p55 soluble receptor lacking the fourth cysteine rich repeat on IgG heavy chains and to compare the TNF binding of such chimeras with the full length soluble receptor chimera. By virtue of their smaller size, they may provide greater flexibility of interaction between the two arms of the IgG/TNF-R dimer molecule and thus with the ligand. We have previously found that a tetrameric Ig-p55 TNF-R fusion protein did not bind TNF significantly better than the dimeric form, possibly due to steric hindrance [22]. A truncated receptor may overcome this problem. Their smaller size may also render them less immunogenic, an important consideration in long term treatment of chronic inflammatory diseases such as RA.

METHODS

Reagents

Oligonucleotides were purchased from National Biosciences (Plymouth, MN). DNA sequencing kits were from U.S. Biochemical Corporation (Cleveland, OH). Goat anti-human IgG Fc fragment-specific antibodies and alkaline phosphatase-conjugated goat anti-human IgG (heavy and light chains) were supplied by Jackson ImmunoResearch (West Grove, PA). Rabbit polyclonal anti-p55TNF-R antiserum was a kind gift from D. Wallach. Unlabelled recombinant human TNF (400 units/ng, > 97% pure) and LT (125 units/ng, > 97% pure) were obtained from R and D Systems (Minneapolis, MN) and ¹²⁵I-labelled human TNF from DuPont Company, NEN (Boston, MA). Human LT was radioiodinated with Na¹²⁵I (IMS30, Amersham) and Iodogen (Pierce), according to the manufacturers' instructions. CD1 mice (6-10 weeks) were obtained from Charles River Labs. (Raleigh, NC). E. coli lipopolysaccharide (0111:B4) was purchased from List Biological Labs. (Campbell, CA), and D-galactosamine from Sigma Chemical Co. (St. Louis, MO).

Construction of 4 TNF receptor-IgG chimeric vectors

Expression vectors: the heavy chain vectors used, designated sf2 (single fusion) have been described previously [22]. Briefly, the vectors were derived from plasmid pSV2-gpt, into which was inserted a cloned gene encoding the human Ig heavy chain promoter and leader peptide coding sequence, a cloning site, a coding sequence for eight amino acids of human J sequence, followed by a genomic sequence for the human IgG1 constant region of a chimeric mouse-human antibody, cM-T412 [29]. A modified vector, designated sf3, was constructed in which the IgG CH1 domain was deleted. An analogous vector was made from the cM-T412 light chain gene, which expressed a truncated light chain lacking a variable region [30].

TNF receptor constructs

The construction of a p55 soluble TNF-R/IgG chimera has been previously described [22]. To construct 4 derivatives, PCR fragments encoding amino acids 1-127 of the p55 TNF-R extracellular domain, excluding the signal peptide, and thus comprising the first three cysteine-rich repeats, were amplified from a full length cDNA clone [25] using the 3' oligo 5'-GGTGCACACGGTGTTCT-3' and various 5' oligos described below. The PCR fragments were blunt-end ligated into the sf2 and sf3 heavy chain vectors and the integrity of the sequence confirmed by sequencing. The three versions of 4 constructed in sf2 were: 4alt, which contains an Ig signal peptide and 2 altered amino acids at the N terminus, to give the consensus Ig recognition sequence; 4 N, which contains an Ig signal peptide and the TNF-R native N terminus; 4 N2, which contains the native TNF-R signal peptide and N terminus. The 5' oligos used were

5' CACAGGTGTGTCCCCAAGGAAAA 3',

5' CAGATAGTGTGTGTCCCCAAGGAAAA 3'

5' ATAAGAGGCCATAGCTGTCTGGC 3'

The 4alt version was also expressed in the sf3 vector.

A cDNA clone, encoding a monomeric non-fusion (nf) form of the p55 extracellular domain (p55-nf), including the leader peptide, has been described [25]. This cDNA was inserted into a dhfr mammalian expression vector under the transcriptional control of an SV40 early promoter. The various constructs are depicted in Figure 1.

Transfection and expression

Linearized plasmids encoding fusion proteins were transfected into the myeloma cell line X63-Ag8.653 by electroporation. Cell supernatants were assayed for IgG production by ELISA as described elsewhere [30]. Briefly, supernatants were incubated on plates coated with anti-human IgG Fc and bound protein was detected using alkaline phosphatase-conjugated anti-human IgG. The p55-nf construct was transfected into a dhfr^­ mutant CHO cell line by standard calcium phosphate precipitation techniques and amplified by selection in methotrexate.

Purification and Western blotting of fusion proteins

Fusion protein cell supernatants were purified on a Protein A-Sepharose column as described [22]. The p55-nf was purified on a TNF affinity column as described [28]. Purified proteins were resolved by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) to estimate purity and correct size. The fusion proteins were Western blotted with rabbit polyclonal anti-TNF-R followed by alkaline phosphatase conjugated goat anti rabbit antibody.

WEHI cytotoxicity assay

The fusion proteins were assayed for their inhibition of the cytotoxic effect of TNF or LT on the TNF/LT sensitive cell line WEHI 164 clone 13 [31]. WEHI cells (50,000 cells per well) were adhered for 4 hours at 37° C and then incubated overnight at 37° C with 4 pM TNF or 3.6 pM LT, either alone or in combination with increasing concentrations of p55 fusion or non-fusion proteins. The cells were assayed as described elsewhere [31] except that they were lysed with dimethyl sulphoxide and analysed immediately.

Inhibition of binding of ¹²⁵I labelled TNF and LT to U937 cells

Increasing concentrations of p55 and 4 fusion proteins were preincubated with 1 nM ¹²⁵I labelled TNF or 2 nM LT for 1 hour at 37° C and then added in triplicate to U937 cells aliquoted at 2 x 10⁶ cells per well in 96 well plates, for 2 hours at 4° C. Non-specific binding was determined by preincubation of cells with 250 nM unlabelled cytokine, followed by ¹²⁵I cytokine. Cells were centrifuged through 20% sucrose in PBS to remove unbound cytokine and counted on an LKB 1260 Multigamma II gamma counter.

Binding assays and Scatchard analysis

96-well microtitre plates were coated with goat anti-human IgG Fc antibodies (10 µg/ml). Fusion proteins were captured at a concentration of 10ng/ml. Increasing concentrations of ¹²⁵I-labelled TNF and LT (4.9-20,000 pM) were added in PBS/1% BSA and bound for 2 hours at room temperature. Non-specific binding was determined with a 200-fold excess of unlabelled cytokine. Bound cpm were removed with 2N NaOH and counted.

In vivo mouse model of LPS induced lethality

Groups of fifteen CD1 mice were injected intravenously (i.v.) with either saline (control mice), or p55-sf2 or 4-sf2 in two quantities ­ 0.5 or 5 ug. One hour later, the mice were injected simultaneously i.v. with 50 µg E. coli LPS (0111:B4) and 15 mg D-galactosamine. Live mice were counted after 24 hours. Three separate experiments were performed.

RESULTS

The p55-sf2 and 4-sf2 constructs were expressed in association with the truncated light chain as this was necessary for secretion of the heavy chain fusion proteins from myeloma cells. The 4-sf3 construct, which lacks the IgG CH1 domain, was expressed in the absence of light chain. The purified fusion proteins were shown by SDS-PAGE to be > 95% pure and the appropriate size i.e. 168 kD for p55-sf2 [22], 162 kD for the 4 derivatives lacking the fourth cysteine-rich repeat, and 108 kD for 4-sf3 due to the additional deletion of the CH1 domain. Western blotting with anti-TNF-R confirmed the presence of TNF receptor derivatives fused to the Ig heavy chain (data not shown). The levels of expression of the 4 sf2 chimeras were very similar, as was their biological activity in the assays described below. Thus, provision of an Ig leader sequence or N-terminal recognition sequence is not necessary for stable expression of functional TNF-R/Ig chimeras. For brevity only the 4alt version is shown in the following experiments.

Inhibition of TNF or LT cytotoxicity in the WEHI assay

The ability of the soluble receptor fusion proteins to sequester TNF and LT, thus inhibiting their biological activities, was tested in a cytotoxicity assay using the TNF/LT-sensitive murine cell line, WEHI 164 clone 13, which expresses the p55 TNF-R. Figure 2A shows that p55 sf2, 4-sf2 and 4-sf3 exhibited identical dose-dependent inhibition of TNF cytotoxicity. All three fusion proteins effected 50% inhibition at a 2 fold molar excess over TNF. In contrast, the non-fusion soluble receptor was required in 500 fold excess to achieve 50% inhibition (Table 1). In the case of LT inhibition, 50% inhibition was achieved by p55-sf2 at an 11 fold molar excess (Figure 2B). However, the molar ratio of 4-sf2 and 4-sf3 required was significantly higher (18 fold and 21 fold respectively).

Inhibition of binding of ¹²⁵I TNF or LT to U937 cells

The U937 human monocytic cell line expresses both p55 and p75 TNF-R [32]. As seen in Figure 3A, the 4-sf2 and 4sf3 fusion proteins effectively inhibited binding of TNF to this cell line, with IC_50s very close to that of the wild type fusion protein (Table 1). The molar ratios required (approx. 7:1) were higher than for the WEHI assay. Inhibition of LT binding by p55-sf2 was similar to the profile seen with TNF, as shown in Figure 3B. However, a relatively higher molar ratio of the 4 derivatives (16:1 and 15:1 for 4-sf2 and 4-sf3 respectively) was required to achieve IC₅₀ (Table 1). As before, much higher molar ratios of the monomeric receptor were required to reach IC₅₀-200:1 with TNF and 300:1 with LT.

Affinity studies

The binding affinity of the fusion proteins for TNF and LT was determined by incubation with increasing concentrations of ¹²⁵I labelled ligand. The binding curves in Figure 4A and 5A demonstrate specific saturable binding of TNF and LT respectively, by p55-sf2, 4-sf2 and 4-sf3. Scatchard analysis reveals a single set of high affinity binding sites for each. p55-sf2 (Figure 4B) has a Kd for TNF of 29 pM similar to that previously described [22]. The affinities of both 4-sf2 (s) and 4-sf3 (Figure 4D) were consistently slightly higher in each individual experiment performed (n = 5), showing mean Kds of 26 and 25 pM respectively, although this difference was not statistically significant. The normal circula-ting levels of TNF are 10 ^{­ 18} M, increasing to up to 10 ^{­ 9} M in disease states. Thus the soluble receptor fusion proteins would be able to sequester increasing levels of TNF. The affinities of the fusion proteins for LT were much lower at 117, 240 and 170 pM respectively for p55-sf2 (Figure 5B), 4-sf2 (Figure 5C) and 4-sf3 (Figure 5D). The affinity of 4-sf2 for LT was considerably lower than that of the wild type receptor and a paired Student's T test revealed this difference to be statistically significant (p < 0.005). The affinity of 4-sf3 for LT was intermediate and not statistically different from p55-sf2.

In vivo model

The D-galactosamine (GaLN) sensitization model of LPS induced cytotoxicity [33] has been used previously to analyse TNF-R-Ig chimeras and was used to analyse the in vivo biological activity of the 4-sf2 chimera. GaLN increases sensitivity to LPS or TNF toxicity by up to 100,000 fold. TNF has been well established as the key inflammatory mediator of LPS toxicity, and has identical lethality effects in this model. As shown in Figure 6, for both the p55sf2 and 4 sf2 fusion proteins, the lower dose of 0.5 mg afforded some protection against LPS challenge, in comparison with the control mice which exhibited 100% mortality after 24 hours. Increasing the dose of fusion protein 10 fold led to almost complete inhibition of LPS lethality in both cases. Overall, the ability of the two TNF-R fusion proteins to sequester TNF and thus protect against LPS cytotoxicity was not significantly different in repeated experiments.

DISCUSSION

A wealth of evidence demonstrates the central role of TNF in inflammatory processes. In murine collagen-induced arthritis, treatment with anti-TNF antibodies ameliorates the disease [12]. In recent clinical trials (open and placebo-controlled), patients with active rheumatoid arthritis and Crohn's disease treated with anti-TNF monoclonal antibodies showed marked clinical improvement, indicating that blockage of this cytokine is of therapeutic benefit [13, 14, 17].

The identification of soluble TNF binding proteins in normal human serum and urine, and increased levels of these proteins in inflammatory disease suggest that they may play an important protective role against the harmful effects of excess TNF production in inflammatory diseases [10, 11]. However, we and others have shown that these monomeric soluble receptors fully inhibit TNF cytotoxicity in vitro only at extremely high molar ratios, which are not found in vivo. Reported ratios vary from 2,000:1 to 30,000:1 [19, 20, 22]. Such a high dosage requirement excludes the possibility of administering recombinant monomeric soluble receptors, which are cleared quickly via the kidney, as an anti-inflammatory strategy. Avidity for TNF is increased dramatically by the dimerization of soluble receptors on IgG backbones, presumably due to the fact that the stoichiometry of TNF binding to these receptors is 3 receptors to 1 TNF trimer. Location of two receptor-IgG molecules side by side along the boundaries of the TNF subunits thus leads to a rapid stable binding equilibrium. We have previously shown that a number of p55-IgG fusion proteins bind TNF with high affinity (Kd = 22pM) compared with the monomeric soluble receptor (Kd = 1,900 pM) [22]. Functionally, the molar ratio of fusion protein to TNF required to inhibit TNF cytotoxicity by 50% was 2, which was 1,000 fold less than for the non-fusion protein. Thus, these soluble receptor Ig chimeras may provide a further advance on the use of anti-TNF antibodies, particularly as they are entirely humanized, i.e. have no murine variable regions and thus are likely to be less immunogenic. Indeed, a p75TNF-R-IgG chimera has recently been shown to be beneficial in the treatment of RA [24]. In this study we have compared several versions of a truncated soluble TNF-R fusion protein with the full length version, and have found that their biological efficacy in sequestering TNF is identical to or slightly better than that of the full length chimera, both in in vitro assays and in an in vivo model of LPS-induced lethality. Thus, these fusion proteins are likely to be as effective as the wild type receptor for future therapeutic applications.

A major finding from these studies is that while the 4 chimeras have an identical affinity for TNF compared with the wild type, 4-sf2 has a lower affinity for LT (Kd of 240 vs 117 for p55-sf2). These Scatchard data were reflected by the greater molar ratio of 4-sf2 required to achieve IC₅₀ with LT in the WEHI and U937 assays. Structurally, this is probably due to the different structure of the membrane proximal apices of TNF and LT, which lie close to the deleted fourth cysteine-rich repeat of the receptor. The TNF apex forms a tight cone whereas the apical structure of LT is more open [34, 35]. This decreased affinity of 4-sf2 may reduce interference with the action of LT concomitant with sequestration of TNF. As LT is produced at much lower levels than TNF in a TNF-mediated inflammatory response, the full length p55-sf2 may sequester it efficiently even at levels too low to fully sequester TNF. This effect might be deleterious because, while TNF and LT have many physiological effects in common, LT has a number of important unique effects, in particular regulation of lymph node and spleen development and maintenance of normal peripheral B cell populations [36, 37]. Furthermore, it is likely that LT may bind a separate, as yet unidentifed receptor which does not bind TNF and may thus have other distinct functions [37]. Therefore, use of 4-sf2 rather than p55-sf2 to sequester TNF may be advantageous as it is likely to interfere less with LT functions.

Our investigation of the biological properties of the 4-sf3 variant also resolve, a discrepancy concerning the comparative biological activity of 4 and the full length soluble receptor. Marsters et al. reported that a 4 version of the TNF-R fused to an Ig backbone lacking the CH1 domain bound TNF with 10 fold lower affinity than the wild type soluble receptor fused to the same backbone [38]. In contrast, our previous work with monomeric soluble receptors indicates that the biological activity of 4 is equivalent to that of the wild type receptor [28]. To resolve this anomaly, we tested Marsters' observation by fusing the 4 to sf3, which encodes a similar Ig backbone, lacking the CH1 domain. In our hands, this fusion protein had very similar biological activity to the 4 TNF receptor fused to a full length Ig backbone. As discussed above, both proteins exhibited similar TNF binding characteristics to the wild type TNF-R fusion protein. Thus, our results in two separate studies are in contrast to those of the previous Marsters' study and concur with crystallographic data which depicts the fourth cysteine-rich repeat as spatially removed from the LT trimer [27].

We conclude that a truncated version of a human soluble p55 TNF-R/IgG chimera containing the three membrane distal cysteine-rich repeats has identical TNF binding affinity to the wild type receptor. In contrast, it binds LT with significantly lower affinity. In future therapeutic applications, 4-sf2 is likely to be a potent inhibitor of TNF-mediated inflammatory responses and may be more selective than the wild type receptor in its cytokine-blocking effects. Its efficacy, in dimeric form, sets the stage for construction of tetrameric fusion proteins, which may bind TNF with even greater avidity. Additionally, although no immunogenic effects of humanized Ig-soluble TNF-R therapy have yet been reported [24], these may arise in longer term therapy and 4-sf2 may be less immunogenic than the full length soluble receptor, by virtue of its smaller size.

REFERENCES

1. Beutler B, Cerami A. 1988. Tumor necrosis, cachexia, shock, and inflammation: a common mediator. Annu. Rev. Biochem. 57: 505.

2. Larrick J W, Wright S C. 1990. Cytotoxic mechanism of tumor necrosis factor-alpha. (Review) FASEB J 4: 3215.

3. Brennan F M, Chantry D, Jackson A, Maini R, Feldmann M. 1989. Inhibitory effects of TNF-alpha antibodies on synovial cell interleukin-1 production in rheumatoid arthritis. Lancet ii: 244.

4. Feldmann M, Brennan F M, Chantry D, Haworth C, Turner M, Abney E, Buchan G, Barrett K, Barkley D, Chu A, Field M, Maini R N. 1990. Cytokine production in the rheumatoid joint: implications for treatment. Ann. Rheum. Dis. 49: 480.

5. Keffer J, Probert L, Cazlaris H, Georgopoulos S, Kaslaris E, Kioussis D, Kollias G. 1991. Transgenic mice expressing human tumour necrosis factor: a predictive genetic model of arthritis. EMBO J. 10: 4025.

6. Stauber G B, Aggarwal B B. 1989. Characterization and affinity cross-linking of of receptors for human recombinant lymphotoxin (Tumour necrosis factor-ß) on a human histiocytic lymphoma cell line, U-937. J. Biol. Chem. 264: 3573.

7. Seckinger P, Isaaz S, Dayer J M. 1989. Purification and biologic characterization of a specific tumor necrosis factor-alpha inhibitor. J. Biol. Chem. 264: 11966.

8. Engelmann H, Novick D, Wallach D. 1990. Two tumor necrosis factor-binding proteins purified from human urine. Evidence for immunological cross-reactivity with cell surface tumor necrosis factor receptors. J. Biol. Chem. 265: 1531.

9. Porteu F, Brockhaus M, Wallach D, Engelmann H, Nathan C F. 1991. Human neutrophil elastase releases a ligand-binding fragment from the 75 kDa tumor necrosis factor (TNF) receptor. Comparison with the proteolytic activity responsible for shedding of TNF receptors from stimulated neutrophils. J. Biol. Chem. 266: 18846.

10. Cope A, Aderka D, Doherty M, Engelmann H, Gibbons D, Jones A C, Brennan F M, Maini R N, Wallach D, Feldmann M. 1992. Increased levels of soluble tumour necrosis factor receptors in the sera and synovial fluids of patients with rheumatic diseases. Arthritis Rheum. 35: 1160.

11. Aderka D, Englemann H, Hornik V, Skornick Y, Levo Y, Wallach D, Kushtai G. 1991. Increased serum levels of soluble receptors for tumor necrosis factor in cancer patients. Cancer Res. 51: 5602.

12. Williams R O, Feldmann M, Maini R N. 1992. Anti-tumour necrosis factor-ameliorates joint disease in murine collagen-induced arthritis. Proc. Natl. Acad. Sci. USA 89: 9784.

13. Elliott M J, Maini R N, Feldmann M, Kalden J R, Antoni C, Smolen J S, Leeb B, Breedveld F C, Macfarlane J D, Bijl H, Woody J N. 1994. Randomized double blind comparison of a chimeric monoclonal antibody to tumour necrosis factor-alpha (cA2) versus placebo in rheumatoid arthritis. Lancet 344: 1105.

14. Rankin E C, Choy E H S, Kassimos D, Kingsley G H, Sopwith A M, Isenberg D A, Panayi G S. 1995. The therapeutic effects of an engineered human anti-tumour necrosis factor-alpha antibody (CDP571) in rheumatoid arthritis. Br. J. Rheumatol. 34: 334.

15. Lorenz H M, Antoni C, Valerius T, Repp R, Grunke M, Schwerdtner N, Nusslein H, Woody J, Kalden J R, Manger B. 1996. In vivo blockade of TNF-alpha by intravenous infusion of a chimeric monoclonal TNF-alpha antibody in patients with rheumatoid arthritis. Short term cellular and molecular effects. J. Immunol. 156: 1646.

16. Targan S R, Hanauer S B, van Deventer S J, Mayer L, Present D H, Braakman T, DeWoody K L, Schaible T F, Rutgeerts P J. 1997 A short-term study of chimeric monoclonal antibody cA2 to tumor necrosis factor-alpha for Crohn's disease. Crohn's disease cA2 study group. N. Engl. J. Med. 337: 1029

17. Stack W A, Mann S D, Roy A J, Heath P, Sopwith M, Freeman J, Holmes G, Long R, Forbes A, Kamm M A. 1997. Randomised controlled trial of CDP571 antibody to tumour necrosis factor-alpha in Crohn's disease. Lancet 349: 521.

18. Elliott M J, Maini R N, Feldmann M, Long-Fox A, Charles P, Bijl H, Woody J N. 1994. Repeated therapy with monoclonal antibody to tumour necrosis factor-alpha (cA2) in patients with rheumatoid arthritis. Lancet 344: 1125.

19. Ashkenazi A, Marsters S A, Capon D J, Chamow S M, Figari I S, Pennica D, Goeddel D V, Palladino M A, Smith D H. 1991. Protection against endotoxic shock by a tumor necrosis factor receptor immunoadhesin. Proc. Natl. Acad. Sci. USA 88: 10535.

20. Peppel K, Crawford D, Beutler B. 1991. A tumour necrosis factor (TNF) receptor-IgG heavy chain chimeric protein as a bivalent antagonist of TNF activity. J. Exp. Med. 174: 1483.

21. Lesslauer W, Tabuchi H, Gentz R, Brockhaus M, Schlaeger E J, Grau G, Piguet P F, Pointaire P, Vassalli P, Loetscher H. 1991. Recombinant soluble tumour necrosis factor receptor proteins protect mice from lipopolysaccharide-induced lethality. Eur. J. Immunol. 21: 2883.

22. Scallon B J, Trinh H, Nedelman M, Brennan F M, Feldmann M, Ghrayeb J. 1995. Functional comparisons of different Tumor Necrosis Factor Receptor/IgG fusion proteins. Cytokine 7: 759.

23. Bertini R, Delgado R, Faggioni R, Gascon M P, Ythier A, Ghezzi P. 1993. Urinary TNF-binding protein (TNF soluble receptor) protects mice against the lethal effect of TNF and endotoxic shock. Eur. Cytokine Netw. 4: 39.

24. Moreland L W, Baumgartner S W, Schiff M H, Tindall E A, Fleischmann R M, Weaver A L, Ettlinger R E, Cohen S, Koopman W J, Mohler K, Widmer M B, Blosch C M. 1997. Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Gc fusion protein. N. Engl. J. Med. 337: 141.

25. Gray P W, Barrett K, Chantry D, Turner M, Feldmann M. 1990. Cloning of human tumor necrosis factor (TNF) receptor cDNA and expression of recombinant soluble TNF binding protein. Proc. Natl. Acad. Sci. USA 87: 7380.

26. Smith C A, Davis T, Anderson D, Solam L, Beckmann M P, Jerzy R, Dower S K, Cosman D, Goodwin R G. 1990. A receptor for tumor necrosis factor defines an unusual family of cellular and viral proteins. Science 248: 1019.

27. Banner D W, D'Arcy A, Janes W, Gentz R, Schoenfeld H J, Broger C, Loetscher H, Lesslauer W. 1993. Crystal structure of the soluble human 55 kd TNF receptor-human TNF-ß complex: implications for TNF receptor activation. Cell 73: 431.

28. Corcoran A E, Barrett K, Turner M, Brown A, Kissonerghis A M, Gadnell M, Gray P W, Chernajovsky Y, Feldmann M. 1994. Characterization of ligand binding by the human p55 tumour-necrosis-factor receptor. Involvement of individual cysteine-rich repeats. Eur. J. Biochem. 223: 831.

29. Looney J E, Knight D M, Arevalo-Moore M, Trinh H, Pak K Y, Dalesandro M R, Rieber E P, Riethmuller G, Daddona P E, Ghrayeb J. 1992. High-level expression and characterization of a mouse-human chimeric CD4 antibody with therapeutic potential. Hum. Antibodies Hybridomas 3: 191.

30. Knight D M, Trinh H, Le J, Siegel S, Shealy D, McDonough M, Scallon B J, Moore M A, Vilcek J, Daddona P E, Ghrayeb J. 1993. Construction and initial characterization of a mouse/human chimeric anti-TNF antibody. Mol. Immunol. 30: 1443.

31. Espevik T, Nissen-Meyer J. 1986. A highly sensitive cell line, WEHI 164, clone 13, for measuring cytotoxic factor/tumour necrosis factor from human monocytes. J. Immunol. Methods 95: 99.

32. Brockhaus M, Schoenfeld H J, Schlaeger E J, Hunziker W, Lesslauer W, Loetscher H. 1990. Identification of two types of tumor necrosis factor receptors on human cell lines by monoclonal antibodies. Proc. Natl. Acad. Sci. USA 87: 3127.

33. Lehmann V, Freundenberg M M, Galanos C. 1987. Lethal toxicity of lipopolysaccharide and tumor necrosis factor in normal and D-galactosamine treated mice. J. Exp. Med. 165: 657.

34. Jones E Y, Stuart D I, Walker N P. 1989. Structure of tumour necrosis factor. Nature 338: 225.

35. Eck M J, Ultsch M, Rinderknecht E, de Vos A M, Sprang S R. 1992. The structure of human lymphotoxin (tumor necrosis factor-ß) at 1.9-A resolution. J. Biol. Chem. 267: 2119.

36. De Togni P, Goellner J, Ruddle N H, Streeter P R, Fick A, Mariathasan S, Smith S C, Carlson R, Shornick L P, Strauss-Schoenberger J, Russell J H, Karr R, Chaplin D D. 1994. Abnormal development of peripheral lymphoid organs in mice deficient in lymphotoxin. Science 264: 703.

37. Koni P A, Sacca R, Lawton P, Browning J L, Ruddle N H, Flavell R A. 1997. Distinct roles in lymphoid organogenesis for lymphotoxins alpha and ß revealed in lymphotoxin ß-deficient mice. Immunity 6: 491.

38. Marsters S A, Frutkin A D, Simpson N J, Fendly B M, Ashkenazi A. 1992. Identification of cysteine-rich domains of the type 1 tumor necrosis factor receptor involved in ligand binding. J. Biol. Chem. 267: 5747.