ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1999, p. 1027–1033 0066-4804/99/$04.0010 Copyright © 1999, American Society for Microbiology. All Rights Reserved.
Vol. 43, No. 5
Thiolated Recombinant Human Tumor Necrosis Factor-Alpha Protects against Plasmodium berghei K173-Induced Experimental Cerebral Malaria in Mice NANCY S. POSTMA,1† ROB C. HERMSEN,2* DAAN J. A. CROMMELIN,1 WIJNAND M. C. ELING,2 1 AND JAN ZUIDEMA Department of Pharmaceutics, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, 3508 TB Utrecht,1 and Department of Medical Microbiology, University Hospital Nijmegen, 6500 HB Nijmegen,2 The Netherlands Received 14 October 1998/Returned for modification 30 December 1998/Accepted 25 February 1999
The introduction of reactive thiol groups in recombinant human tumor necrosis factor (TNF) alpha (rhTNF-a) by the reagent succinimidyl-S-acetylthioacetate resulted in the formation of a chemically stabilized rhTNF-a trimer (rhTNFa-AT; as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis). rhTNFa-AT showed a substantially enhanced protective efficacy against the development of experimental murine cerebral malaria (ECM) after intravenous injection compared to the protective efficacy of nonmodified rhTNF-a. Administration of thiolated rhTNF-a with protected thiol groups (rhTNFa-ATA; no stabilized trimers in vitro) exhibited the same protective efficacy against ECM, while in vitro bioactivity was reduced. Parasitemia was significantly suppressed in rhTNF-treated mice that were protected against ECM but not in treated mice that developed ECM. Protection against ECM was not related to increased concentrations in plasma of soluble TNF receptor 1 and 2 directly after injection or at the moment of development of ECM in nontreated mice. The results indicate that thiolation of rhTNF-a leads to the formation of stable trimers with increased potential in vivo. Tumor necrosis factor (TNF) alpha (TNF-a) is a pleiotropic cytokine that is primarily produced by stimulated macrophages and is a major mediator of the inflammatory and cellular immune responses. Although TNF-a was originally identified by its antitumor properties and toxicity for some transformed cell lines (8), it has also been implicated in the pathophysiology of septic shock, cachexia, certain inflammatory diseases (46), and infections, like malaria (9). Monomeric recombinant human TNF-a (rhTNF-a) is a nonglycosylated polypeptide of 157 amino acids with a molecular mass of 17 kDa (1). Analytical ultracentrifugation (3, 50, 51), cross-linking (21, 26, 41, 43), gel filtration chromatography (32, 35), and X-ray scattering (29) studies showed that rhTNF-a is present as a trimer in solution. X-ray crystallographic studies of rhTNF-a revealed that rhTNF-a forms a compact, bell-shaped homotrimer (14, 23). The effects of both rhTNF-a and recombinant human TNF beta (rhTNF-b) are mediated by two receptors with distinct molecular masses: TNF receptor 1 (TNFR1; p55) and TNF receptor 2 (TNFR2; p75) (5). Studies with TNFR1 provided evidence that each rhTNF-a trimer can interact with three receptor molecules (4, 30). Additional mutational analysis studies with rhTNF-a suggest that each trimer has three receptor interaction sites located in the intersubunit grooves near the base of the trimer and that the integrity of the trimeric structure is essential for rhTNF-a activity (48, 52). On the other hand, rhTNF-a oligomers (dimers and trimers) are unstable at low concentrations, resulting in the formation of inactive monomers and polymers (10, 35, 37, 41). This sug-
gests that injected rhTNF-a may be present only as bioactive rhTNF-a at or near the site of administration (37). Stabilization of rhTNF-a in the form of trimers by chemical crosslinking prevents the dissociation into inactive monomers and may therefore enhance the therapeutic potential of rhTNF-a due to higher and longer-circulating concentrations of bioactive rhTNF-a. Stable rhTNF-a trimers were prepared by reaction of the protein with the reagent succinimidyl-S-acetylthioacetate (SATA). Reactive thiol groups are introduced in the protein at the site of primary amino groups (13), which can form disulfide bridges between adjacent rhTNF-a monomers. TNF-a plays a dual role in experimental murine malaria. Whereas high doses of rhTNF-a induced cerebral pathology in mice with malaria (12), recent studies showed that treatment with low doses of rhTNF-a protect against experimental cerebral malaria (ECM) (39). Soluble TNF receptors derived from proteolytic cleavage of the cellular receptors TNFR1 and TNFR2 (15) may play a role in rhTNF-a-mediated protection against ECM. They competitively inhibit the interaction of TNF-a with membrane-bound receptors (27) and thus provide a possible mechanism for regulating the availability of TNF-a. The aim of the present study was to analyze the therapeutic potential of covalently stabilized rhTNF-a trimers in Plasmodium berghei K173-induced ECM in mice. The protective effect of rhTNF-a treatment on ECM was analyzed in relation to its effect on parasitemia and the production of soluble TNF receptors (sTNFR1 and sTNFR2).
* Corresponding author. Mailing address: Department of Medical Microbiology, University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Phone: 31-24-3614664. Fax: 31-243540216. E-mail:
[email protected]. † Present address: ACE Pharmaceuticals BV, 3890 BB Zeewolde, The Netherlands.
Materials. rhTNF-a (4 mg/ml in 10 mM sodium phosphate [pH 7], 200 mM NaCl) (derived from Escherichia coli) (36) (specific activity, 5 3 107 U/mg, as determined by bioassay with murine LM cells) was a gift from G. R. Adolf (Boehringer Ingelheim, Vienna, Austria). SATA was purchased from Pierce (Rockford, Ill.). All other reagents used were of analytical grade. Methods. (i) Thiolation of rhTNF-a by the SATA-reaction. Basically, the procedure for the thiolation of proteins with SATA described by Duncan et al.
MATERIALS AND METHODS
1027
1028
POSTMA ET AL.
(13) was used. rhTNF-a (4 mg/ml in 10 mM sodium phosphate [pH 7], 200 mM NaCl) was applied on a Sephadex G-25M column (PD-10 column; Pharmacia, Uppsala, Sweden) and was eluted with HEPES buffer (10 mM HEPES, 135 mM NaCl, 1 mM EDTA [pH 7.5]). The eluted rhTNF-a solution was concentrated with Microsep filters (molecular mass cutoff, 10 kDa; Pall Filtron, Breda, The Netherlands) at 4°C to approximately 2 mg/ml before reaction with SATA. SATA was dissolved in dimethylformamide, and the solution was mixed with the rhTNF-a solution in a volume ratio of dimethylformamide:buffer of 1:100 and a molar ratio of SATA:rhTNF-a of 8:1. The mixture was incubated at room temperature under continuous rotation for at least 20 min to allow the formation of rhTNF-a with protected thiol groups (rhTNFa-ATA). After incubation, the unreacted SATA was separated from rhTNFa-ATA on a PD-10 column. The protein fraction in the eluate was detected by monitoring the absorption at 280 nm (A280), collected, and stored at 220°C (for a maximum of 3 weeks). rhTNFaATA was deacetylated by adding a freshly prepared hydroxylamine-HCl solution (0.5 M hydroxylamine-HCl, 0.5 M HEPES, 25 mM EDTA [pH 7.5]). The mixture was incubated for 60 min at room temperature under continuous rotation at a volume ratio of rhTNFa-ATA:hydroxylamine of 10:1, yielding rhTNF-a with reactive thiol groups (rhTNFa-AT). This product was used in all studies. (ii) Protein determination. To determine the concentrations of rhTNF-a and rhTNFa-ATA, the method described by Wessel and Flugge (49) was used. However, the amount of rhTNFa-AT could not be determined by this method due to interference of hydroxylamine with the analysis. Therefore, the amount of protein in the rhTNFa-AT solutions was estimated from the amount of rhTNFaATA measured and was corrected for dilution upon addition of hydroxylamine, because it is not likely that the addition of hydroxylamine will change the amount of protein in any other way. (iii) GPC. The molecular masses of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT were determined by gel permeation chromatography (GPC) with a system consisting of a model 510 high-pressure liquid chromatography pump (Waters Associates, Milford, Mass.), a thermostated (35°C) precolumn (Biosep SEC S3000 PEEK; 50 by 7.5 mm, Phenomenex, Torrance, Calif.), and a TSK-GEL G3000SW column (75 by 300 mm; particle size, 10 mm; Tosohaas). The system was equilibrated with 100 mM sodium sulfate–100 mM sodium phosphate buffer (pH 6.7). The material was eluted with the same buffer at a flow rate of 1.0 ml/min, and the elution profile was monitored at 280 nm with a UV detector (model 486; Waters Associates). The system was calibrated with mixtures of standards of known molecular weight (dextran blue [2,000,000], ovalbumin [43,000], and RNase A [13,7000]; thyroglobulin [669,000], lactoglobulin [18,400], and insulin [6,000]; and immunoglobulin G [150,000], chymotrypsin [25,000], bovine serum albumin [67,000], and myoglobin [17,200]). (iv) SDS-PAGE. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) analysis of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT was performed in a system modified from that of Laemmli (25) by using a horizontal 8 to 18% gradient gel (ExelGel, Pharmacia), and the gels were silver stained. Gels were run with nonreduced and reduced (0.1 M dithiothreitol and 10% iodocetamide) samples. Molecular mass markers (Bio-Rad Laboratories, Veenendaal, The Netherlands) were used, with the molecular masses of the markers ranging from 14 to 200 kDa. (v) In vitro bioactivities. The biological activities of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT were determined by the WEHI cytotoxicity assay (16) by the method described previously (39). Briefly, 50 ml of 0.5 3 106 WEHI cells/ml of medium was pipetted into 96-well microtiter plates. Following incubation for 3 to 5 h at 37°C in 5% CO2, 50 ml of standard or test sample was added. A standard curve (10 to 20,000 pg/ml) was made from rhTNF-a (Boehringer, Ingelheim, Germany) with an originally defined biological activity of 6 3 107 U/mg in a murine LM bioassay according to World Health Organization standards. Serial dilutions of samples in medium were added directly to the cells. After incubation for 18 h at 37°C in 5% CO2, the viable cells were quantified by the XTT assay (40). Briefly, 50 ml of the XTT labelling mixture was added to the wells, and the plate was incubated for 90 min at 37°C in 5% CO2. The A490 was measured with a Bio-Rad Novapath microplate reader (Bio-Rad Laboratories, Veenendaal, The Netherlands) (reference wavelength, 655 nm). The lower detection limit of the WEHI cytotoxicity assay was approximately 80 pg of rhTNF-a/ml. In the WEHI cytotoxicity assay described above the specific biological activity of rhTNF-a (in units per milligram) used for the experiments was about two to four times higher in comparison with that of the standard rhTNF-a (Boehringer Ingelheim, Germany). (vi) Mouse TNFR1 and TNFR2 ELISAs. Mice treated with buffer, rhTNF-a, rhTNFa-ATA, or rhTNFa-AT at day 5 after infection were killed at predefined time points (4, 24, 72, or 96 h) after injection. Blood was collected in endotoxinfree tubes (4°C) which contained endotoxin-free (as determined by the Limulus amebocyte lysate assay) heparin (NPBI, Emmer Compascuum, The Netherlands). Plasma was collected by centrifugation within 2 h after bleeding and was stored at 220°C. Enzyme-linked immunosorbent assays (ELISAs) recognizing both free and complexed TNF receptors were carried out with the materials and by the methods described by Bemelmans et al. (6). Briefly, microtiter plates were coated with anti-TNFR1 or anti-TNFR2 monoclonal antibody. Nonspecific binding was blocked with 1% bovine serum albumin. After washing, mouse plasma samples were added. A standard curve was made with recombinant murine TNFR1 and TNFR2 proteins. Subsequently, the plates were washed and incubated with biotinylated anti-TNF-a receptor antibodies. After washing, the
ANTIMICROB. AGENTS CHEMOTHER. plates were incubated with streptavidin peroxidase, followed by an enzyme reaction that provided a colored reaction product. The detection limit for soluble murine TNFR1 was 40 pg/ml, and that for soluble murine TNFR2 was 1,560 pg/ml. Mice. Female C57BL/6J mice (age, 6 to 10 weeks) were obtained from a specific-pathogen-free colony maintained at the Central Animal Facility of the University of Nijmegen. All mice were housed in plastic cages and received water and standard RMH food (Hope Farms, Woerden, The Netherlands) ad libitum. Parasite and development of cerebral malaria. P. berghei K173 (originally obtained from W. Kretschmar, Tu ¨bingen, Germany) has been maintained by weekly transfer of parasitized erythrocytes (PEs) from infected into naive mice for .30 years. Experimental C57BL/6J mice were injected intraperitoneally with 103 PEs from blood from infected donors. About 95% of the C57BL/6J mice injected intraperitoneally with 103 PEs die early in the second week after infection (days 9 to 12) due to the development of ECM (11). Approximately 1 day before death a progressive hypothermia develops, and this progressive hypothermia is strongly correlated with development of hemorrhage in the brain, as observed by histology (38). Mice that survive this critical period show only a limited, transient hypothermia, and the majority die in the third week or later after infection with severe anemia and a parasitemia of between 20 and 40%, but with no noticeable cerebral pathology, as determined by light microscopy. The day of death after infection was used as the parameter for ECM or protection against ECM in the experiments described in this paper. Experimental design. For each experiment mice received 103 PEs intraperitoneally on day 0. At day 5 mice were treated with a single intravenous injection of either rhTNF-a, rhTNFa-ATA, or rhTNFa-AT. Prior to injection, appropriate dilutions were prepared with HEPES buffer (10 mM HEPES, 135 mM NaCl, 1 mM EDTA [pH 7.5]) containing 1% mouse plasma. Placebo-treated mice received buffer only. Control mice did not receive any treatment (nontreated mice). The effect of treatment on parasitemia (percent infected erythrocytes) at day 8 after infection was studied from thin blood films made from tail blood and stained with May-Gru ¨nwald and Giemsa solutions. A considerable variation in parasitemia is observed among independent infected control groups at day 8 after infection (average 6 standard deviation parasitemia, 17% 6 8%; n 5 8 experiments). Therefore, within each experiment the average parasitemia for nontreated mice was determined and the parasitemia for each mouse from either a control, placebo, or a treatment group was divided by this average. These ratios were used to compare and summarize the results of independent repeat experiments. The effect of treatment on the development of ECM was studied by monitoring the survival of the mice. The death of mice within 2 weeks after infection was used to identify ECM-related death (11). Persistent, recurrent parasitemia and survival for more than 2 weeks after infection indicated protection against ECM. Statistical analysis. The effect of treatment on parasitemia was analyzed by one-way analysis of variance (ANOVA). The effect of treatment on survival was analyzed by the chi-square test. Differences were considered significant at a level of a equal to 0.05.
RESULTS Characterization of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT by GPC and SDS-PAGE. By GPC under nondenaturing conditions, rhTNF-a is eluted as a single peak corresponding to a molecular mass of approximately 42 kDa, which is indicative of a compact trimer. No other peaks were present (data not shown). Chromatography of both rhTNFa-ATA and rhTNFa-AT showed major peaks corresponding to molecular masses of 37 and 36 kDa, respectively (dimer or compact trimer). Small fractions were eluted in the void volume (molecular mass, .500 kDa) and an additional fraction had a molecular mass of approximately 80 kDa (multimers). When subjected to SDS-PAGE analysis, nonreduced samples of nonmodified rhTNF-a showed a strong monomeric band of 17 kDa (Fig. 1A, lane 1). rhTNFa-ATA exhibited two major bands of about 17 kDa (Fig. 1A, lane 2), whereas rhTNFa-AT showed a major band of 52 kDa (a band of this molecular mass is indicative of a trimer; Fig. 1A, lane 3). Reduced samples yielded a single band corresponding to the rhTNF-a monomer of about 17 kDa (Fig. 1B). No other products were observed. In vitro bioactivities. The bioactivities of nonmodified rhTNF-a, rhTNFa-ATA, and rhTNFa-AT were determined by a standard WEHI cytotoxicity assay. Equal concentrations of protein were added to the cells. rhTNF-a induces cell lysis, resulting in a lower A490 with increasing rhTNF-a concentra-
VOL. 43, 1999
THIOLATED rhTNF-a PROTECTS AGAINST ECM
1029
FIG. 1. Samples of rhTNF-a (lane 1), rhTNFa-ATA (lane 2), and rhTNFa-AT (lane 3) were analyzed by SDS-PAGE in the absence (A) or presence (B) of reducing conditions. Photometric scans of silver-stained SDS-polyacrylamide gels are presented. The major bands are indicated by arrows and their corresponding relative molecular masses (kD, kilodaltons).
tions. Figure 2 shows a typical example of the cytotoxicity of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT in the WEHI cytotoxicity assay. Almost 100% cytotoxicity was observed with protein concentrations of 20,000 to 40,000 pg/ml. With lower rhTNF-a-ATA concentrations, rhTNFa-ATA was less effective in causing cell lysis (higher A490 value) than equal concentrations of rhTNF-a (Fig. 2). This result was consistent in all experiments performed. The bioactivity of rhTNFa-ATA was about eight- to ninefold lower than that of nonmodified rhTNF-a (Fig. 2). The bioactivity was almost completely restored (70 to 90%) upon deacetylation of rhTNFa-ATA to rhTNFa-AT. Under the experimental conditions used, hydroxylamine did not interfere with the bioactivity of rhTNF-a (data not shown).
FIG. 2. Dose-response curves of the cytotoxic activities of rhTNF-a, rhTNFa-ATA, and rhTNFa-AT. An equal dose of rhTNF-a, rhTNFa-ATA, or rhTNFa-AT was added to TNF-a-sensitive WEHI cells. The A490 reflects the presence of viable cells and is plotted against the log10 of the protein concentration. rhTNF-a induces lysis of WEHI cells, and the level of lysis increases with increasing protein concentrations, as reflected by decreasing A490 values. Maximum cytotoxicity is reached at an A490 of 0.3 to 0.4. Samples were measured in triplicate, and averages 6 standard deviations are presented.
FIG. 3. Effect of treatment with rhTNF-a, rhTNFa-ATA, or rhTNFa-AT during infection on prevention of ECM. Mice received a single intravenous injection of rhTNF-a, rhTNFa-ATA, or rhTNFa-AT at day 5 after infection with 103 P. berghei parasites, and survival for more than 2 weeks after infection was used as a marker for protection against ECM. The percentage of mice protected against ECM is plotted against the protein dose. The numbers above each bar indicate numbers of mice protected/numbers of mice treated.
Effect of treatment with rhTNF-a, rhTNFa-ATA, or rhTNFa-AT on ECM. Intravenous injection of increasing doses of rhTNF-a of 2 to 18 mg/mouse protected up to 30% of the mice against ECM (Fig. 3), and the protective effect did not increase when 80 mg of rhTNF-a was administered. No toxicity of this dose was observed (data not shown). Injection of buffer only did not protect against ECM (data not shown). Mean survival times for mice that were protected against ECM and mice that were not protected were 25 and 9 days after infection, respectively. Mice protected against ECM did not show any cerebral pathology in postmortem histological analysis, as was observed for mice that died from ECM (11). Mice protected against ECM eventually died from severe anemia. These observations are typical for all protected and nonprotected mice described in this paper. Injection of either rhTNFa-ATA or rhTNFa-AT (dose range, 1 to 18 mg of protein/mouse) significantly (P , 0.05; chi-square test) enhanced the protection against ECM compared to the protection offered by nonmodified rhTNF-a (Fig. 3). The protective efficacy increased up to 80% at doses of 9 and 18 mg of protein for both rhTNFa-ATA and rhTNFa-AT. The differences in the protective effects of rhTNFa-ATA and rhTNFa-AT were not significant (summarized data for doses of 1 to 18 mg of protein/mouse). In the WEHI cytotoxicity assay, equal protein concentrations were associated with an eight- to ninefold lower bioactivity of rhTNFa-ATA compared to the bioactivities of rhTNF-a and rhTNFa-AT (Fig. 2). A ninefold higher dose of rhTNFa-ATA (81 mg of protein corresponding to 9 mg of bioactive protein in the WEHI cytotoxicity assay) killed 13 of 15 mice within 24 h, which is in contrast to the observation that 80 mg of nonmodified rhTNF-a was not toxic to the infected
1030
POSTMA ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
TABLE 1. Effect of intravenous injection of rhTNF-a, rhTNFa-ATA, or rhTNFa-AT at day 5 after infection on parasitemia in relation to development of cerebral malaria in P. berghei-infected mice Ratio of parasitemia at day 8/mean parasitemia for nontreated micea
Treatment
No treatment rhTNF-a (2–18 mg) rhTNFa-ATA (1–18 mg) rhTNFa-AT (1–18 mg)
All mice
Mice with ECM
Mice without ECM
1 6 0.4 (n 5 27) 1 6 0.5 (n 5 33) 0.8 6 0.5 (n 5 52) 0.7b 6 0.4 (n 5 35)
1 6 0.4 (n 5 27) 1 6 0.5 (n 5 27) 1 6 0.6 (n 5 17) 0.7 6 0.2 (n 5 9)
(n 5 0) 0.6b,c 6 0.2 (n 5 6) 0.6b,c 6 0.2 (n 5 35) 0.6b 6 0.4 (n 5 26)
a The average parasitemia at day 8 after infection was determined for each control group, and the parasitemia for each mouse was divided by the average parasitemia for the corresponding control group. The ratios were used for further analysis. Values represent averages 6 standard deviations of the combined data for different doses (1 to 18 mg protein). b P , 0.05 (ANOVA) compared to the results for nontreated mice. c P , 0.05 (ANOVA) compared to the results for mice that in the same treatment group developed ECM.
mice. Comparable toxicity was observed after intravenous injection of 80 mg of rhTNFa-AT (data not shown). Effect of treatment with rhTNF-a, rhTNFa-ATA, or rhTNFa-AT on parasitemia. For the dose range studied (1 to 18 mg of protein/mouse), the effect of treatment on parasitemia was independent of dose for either rhTNF-a, rhTNFa-ATA, or rhTNFa-AT (ANOVA; data not shown). Therefore, the parasitemia data for all doses tested were summarized and are presented in Table 1. Parasitemia was significantly lower after treatment with rhTNFa-AT (P , 0.05 compared to the results for nontreated mice; ANOVA) but not after administration of rhTNF-a or rhTNFa-ATA (Table 1). When the effect of treatment on parasitemia was analyzed in relation to protection against ECM, parasitemia was significantly lower in mice protected against ECM but not in mice that developed ECM in the same treatment group (P , 0.05 compared to to the results for nontreated mice; ANOVA). It should be noted that significantly lower ratios of (0.6) for mice treated with rhTNF-a, rhTNFa-ATA, or rhTNFa-AT in relation to an average parasitemia of 17% for the controls at day 8 after infection (see Materials and Methods) indicate that at day 8 after infection treated mice still exhibit parasitemias that overlap those observed in the controls. Injection of buffer only did not affect parasitemia (data not shown). Effect of treatment with rhTNF-a, rhTNFa-ATA, and rhTNFa-AT on sTNFRs in plasma. The effect of treatment on changes in the concentrations of sTNFR1 and sTNFR2 in plasma was analyzed in comparison with the effects of buffer treatment in P. berghei-infected mice. Since rhTNFa-ATA exhibited an eight- to ninefold lower in vitro bioactivity than nonmodified rhTNF-a and rhTNFa-AT, the effects of both
treatments were analyzed with comparable amounts of total protein (9 mg) and bioactive rhTNFa-ATA (81 mg of protein). At all time points, the concentrations of sTNFR1 in plasma were below the detection limit (,40 pg/ml). The plasma sTNFR2 concentration at different time points after treatment is presented in Table 2. The plasma sTNFR2 concentration remained below the detection limit in buffer-treated mice until day 6 and increased thereafter. Treatment with 9 mg of rhTNF-a, rhTNFa-ATA, or rhTNFa-AT significantly (P , 0.05 compared to the results for buffer-treated mice; ANOVA) increased the concentration of sTNFR2 4 h after injection, and a further significant increase was noted when 81 mg of rhTNFa-ATA was injected. Analysis 24 h after injection showed that the plasma sTNFR2 concentrations had returned to the values for controls for mice treated with 9 mg of rhTNF-a, rhTNFa-ATA, or rhTNFa-AT, but the concentrations were still elevated in mice treated with 81 mg of rhTNFa-ATA. However, data for only two mice could be obtained because 13 of 15 mice died due to the toxicity of this dose of rhTNFa-ATA. On days 8 and 9 after infection, sTNFR2 concentrations in mice treated with either rhTNF-a, rhTNFa-ATA, or rhTNFa-AT increased again, as was observed in buffer-treated mice. Plasma sTNFR2 concentrations were significantly (P , 0.05 compared to the results for buffertreated mice; ANOVA) enhanced at day 8 after infection but did not reach significance at day 9 after infection compared with the concentrations in the plasma of buffer-treated mice. Treatment with rhTNFa-AT resulted in a twofold increase in plasma sTNFR2 concentrations at day 9 after infection compared to the plasma sTNFR2 concentrations obtained after treatment with rhTNF-a and rhTNFa-ATA and a threefold
TABLE 2. Effect of treatment with nonmodified rhTNF-a, rhTNFa-ATA, or rhTNFa-AT on mouse sTNFR2 in plasma of P. berghei-infected micea Time after infection, time after injection
Day Day Day Day Day
5, 5, 6, 8, 9,
0 4 24 72 96
Mouse sTNFR2 concn (ng/ml) Buffer
rhTNF-a (9 mg)
rhTNFa-ATA (9 mg)
rhTNFa-ATA (81 mg)
rhTNFa-AT (9 mg)
,1.6 (n 5 3) ,1.6 (n 5 3) ,1.6 (n 5 3) 22 6 10 (n 5 5) 28 6 7 (n 5 3)
62b 6 6 (n 5 4) ,1.6 (n 5 4) 38b 6 8 (n 5 4) 45 6 23 (n 5 4)
85b 6 25 (n 5 7) ,1.6 (n 5 7) 36b 6 11 (n 5 7) 52 6 18 (n 5 7)
96b,c 6 4 (n 5 4) 136, 129d
84b 6 17 (n 5 3) ,1.6 (n 5 3) 33b 6 7 (n 5 3) 103 6 28 (n 5 3)
a Mice received an intravenous injection of buffer, rhTNF-a, rhTNFa-ATA, or rhTNFa-AT at day 5 after infection with 103 parasites. At predefined time points after injection, the mice were killed and blood samples were taken. The plasma was stored at 220°C until analysis of sTNFR2. Values represent averages 6 standard deviations. b P , 0.05 (ANOVA) compared to the results for buffer-treated mice. c P , 0.05 (ANOVA) compared to the results for mice treated with rhTNF-a d Thirteen of 15 mice died; values represent data for the surviving mice (n 5 2).
VOL. 43, 1999
THIOLATED rhTNF-a PROTECTS AGAINST ECM
increase compared to the plasma sTNFR2 concentration obtained after treatment with buffer. DISCUSSION The present study demonstrates that the introduction of reactive thiol groups in rhTNF-a by the SATA reaction results in the formation of a covalently stabilized trimer (rhTNFaAT) with a significantly enhanced protective efficacy against P. berghei-induced ECM compared to the efficacy of nonmodified rhTNF-a. Administration of thiolated rhTNF-a with protected thiol groups (rhTNFa-ATA; no stable trimers in vitro) exhibited protective efficacy similar to that of rhTNFa-AT against ECM, despite a ninefold reduced in vitro bioactivity. Characterization of thiolated rhTNF-a. SDS-PAGE analysis under nonreducing conditions confirmed that stable rhTNF-a trimers were formed upon thiolation of rhTNF-a (rhTNFaAT) with SATA (Fig. 1). Stable trimers of thiolated rhTNF-a probably result from the formation of disulfide bridges between monomers already locked into a trimeric position. Monomeric rhTNF-a contains one terminal valine and six lysyl residues. The total number of thiolated residues in rhTNF-a was not determined in the present study, but acetylation of up to 12 residues/trimer was shown before (47). The heterogeneity of bands on the SDS-gel probably reflects differences in disulfide bridging, resulting in different molecular complexes (21). The absence of other products besides the trimer suggests that the trimer is the actual conformational state of rhTNF-a in solution and is not merely a cross-linking phenomenon of rhTNF-a monomers. GPC analysis of rhTNF-a under nondenaturing conditions revealed a single peak with a molecular mass of 42 kDa, which is indicative of a dimer or a compact trimer, which is in accordance with data in the literature (26, 37, 51). Under denaturing conditions (SDS-PAGE) only the monomeric form of rhTNF-a was present, demonstrating that the rhTNF-a oligomer is held together by noncovalent interactions. The presence of bioactive oligomers and inactive monomers and polymers of human TNF-a has also been demonstrated in vivo and may represent a “natural” mechanism in the regulation of cytokine activity (34). Taken together the present data suggest that the enhanced therapeutic potential of rhTNFa-AT results from an increased availability of bioactive rhTNF-a (i.e., rhTNF-a trimers) in vivo. In vitro cytotoxicity and in vivo effects on ECM and parasitemia. Interestingly, administration of rhTNFa-ATA showed the same protective efficacy against ECM as that of rhTNFaAT, while the majority of the protein was present in the injected solution as noncovalently bound oligomers (GPC data; Fig. 1), and in vitro bioactivity was eight- to ninefold lower compared with the bioactivity of the same amount of rhTNF-a and rhTNFa-AT (Fig. 2). In addition, treatment with 80 mg of rhTNFa-ATA was toxic in vivo, similar to treatment with rhTNFa-AT, while 80 mg of rhTNF-a was not toxic in vivo. These data indicate that rhTNFa-ATA is transformed into a bioactive compound in vivo. A possible explanation is that rhTNFa-ATA is deacetylated in vivo and forms stable, bioactive trimers in vivo. This remains to be investigated. The mechanism by which treatment with rhTNF-a, rhTNFaATA, or rhTNFa-AT protects against the development of ECM remains to be elucidated. Our data suggest that inhibition of parasitemia might be a factor that plays a role, particularly because the inhibition was specifically observed in mice protected against ECM but not in treated mice that developed ECM (Table 1). On the one hand, inhibition of parasitemia by rhTNF-a has been described before in other models (42, 45)
1031
and is probably mediated through products released from rhTNF-a-activated cells (e.g., macrophages and neutrophils) (24, 45). Inhibition of erythropoiesis by rhTNF-a treatment, however, may also play a role (22, 33). On the other hand, it should be noted that almost all mice infected with inocula of between 10 and 10,000 parasites/mouse, which results in widely varying parasitemias after infection, develop ECM (11). Moreover, after infection with the same number of parasites, the parasitemias vary extensively among mice in the same group as well as among mice in independent experiments, resulting in a wide range of parasitemias (average 6 standard deviation, 17% 6 8%) at day 8 after infection, shortly before they die from ECM. The last observations tend to suggest that the actual parasitemia is not a critical factor in the development of cerebral malaria in P. berghei-infected mice. Thus, the rhTNFa-mediated protection is not dependent on the magnitude of the inhibition of parasitemia, which is in line with the absence of an effect of the rhTNF-a dose (1 to 18 mg of protein/mouse) on parasite inhibition. Possibly, rhTNF-a treatment affects bone marrow function, and the inhibition of erythropoiesis, which inhibits parasitemia, is only one of the outcomes. The effects of the derivatives rhTNFa-ATA and rhTNFa-AT on parasitemia and protection against ECM may be obtained through the same interactions through which the effects of rhTNF-a are obtained. Protection against cerebral malaria and sTNFRs. Since in another murine model of ECM increased levels of bioactive rhTNF-a in plasma were shown to be responsible for the development of ECM (19), protection against ECM could depend on the timely production of sufficient sTNFRs. sTNFRs competitively inhibit the interaction of TNF-a with membranebound receptor (27), and transgenic mice expressing high levels of soluble TNFR1 were protected against ECM (18). On the other hand, an increase in sTNFR levels results in a decreased amount of receptors on the cell membrane, which could reduce the sensitivity of cells to TNF-a. Previous studies have shown that TNFR2 knock-out mice are protected against ECM in a murine model (31). No increase in sTNFR1 levels was observed in our studies, possibly because peak levels are reached earlier (,2 h after injection) (7). On the other hand, treatment with rhTNF-a induced the production of sTNFR2 (4 h after injection; Table 2), whereas rhTNF-a binds only to mouse TNFR1 (28), indicating that triggering of TNFR1 is sufficient to cause shedding of both TNFRs (7). The increased concentrations of sTNFR2 24 h after injection of 81 mg of rhTNFa-ATA compared with those obtained after treatment with 9 mg of protein may reflect a prolonged time of circulation of bioactive protein and a strong compensating response of the body to an overdose of bioactive rhTNFa-ATA. The enhanced protection against cerebral malaria by treatment with rhTNFa-ATA and rhTNFa-AT compared to the protection offered by rhTNF-a is not reflected by a difference in sTNFR2 levels at day 8 after infection (Fig. 3 and Table 2). Mice with nontreated infections develop ECM, despite the presence of increasing concentrations of sTNFRs (20). The observation that protection against ECM is not correlated with sTNFR concentrations is in agreement with the observations that no bioactive mTNF-a was present in the plasma of nontreated infected mice and that mice could not be protected against ECM by treatment with anti-TNF antibody in this model (20). The enhanced plasma sTNFR2 concentrations at day 9 after injection of rhTNFa-AT were not related to an enhanced suppression of parasitemia (Table 1) or to an enhanced protective efficacy against ECM compared with the suppression of parasitemia or the protective efficacy of
1032
POSTMA ET AL.
rhTNFa-ATA (Fig. 3). However, parasitemia was determined at day 8 after infection, which was 1 day before elevated plasma sTNFR2 concentrations were observed. The mechanism behind the increase in the sTNFR2 concentration and its potential therapeutic implication remains to be elucidated. Still another mechanism that may be involved in the protection against ECM by rhTNF-a and thiolated rhTNF-a is the development of refractoriness or tolerance. Several studies demonstrated that pretreatment with TNF-a induces tolerance to the (toxic) effects of subsequent stimuli that provoke TNF-a release or to the effects of TNF-a itself (2, 17). Tolerance persists for several days and might be another “natural” mechanism (in addition to sTNFRs) that controls excessive stimulation by TNF-a. Tolerance was not based on differences in the levels of sTNFRs between tolerant and nontolerant mice (44), which is in agreement with our observations that rhTNF-amediated protection against ECM is not related to elevated concentrations of sTNFRs. The results obtained in the present study do not permit determination of whether treatment with rhTNF-a and thiolated rhTNF-a induces tolerance and thus protection against ECM. In summary, intravenous injection of covalently stabilized trimers of rhTNF-a (rhTNFa-AT) strongly enhances the bioactivity of rhTNF-a in vivo. This is reflected in the protection against ECM provided by low doses (2 to 18 mg of protein). Moreover, the data suggest that rhTNFa-ATA is transformed into a bioactive compound in vivo. Finally, the effect of parasitemia on rhTNF-a-, rhTNFaATA-, and rhTNFa-AT-mediated protection against ECM remains to be clarified, but the effects on sTNFRs are probably not involved. ACKNOWLEDGMENTS We gratefully thank G. R. Adolf (Boehringer Ingelheim, Vienna, Austria) for the generous gift of rhTNF-a. We also thank A. Mekking and J. Smal (Octoplus, Leiden, The Netherlands) for performing the SDS-PAGE analysis. We thank W. Buurman (Department of General Surgery, University Limburg, Maastricht, The Netherlands) for supplying us with the materials used for the sTNFR ELISAs and J. van de Meer (University Hospital Nijmegen, Nijmegen, The Netherlands) for the gift of rhTNF-a used as a standard in the WEHI cytotoxicity assay. We thank M. J. van Steenbergen (Department of Pharmaceutics, Utrecht University, Utrecht, The Netherlands) for performing the GPC analysis, and we thank G. Poelen, T. van de Ing, and Y. Brom for skilled biotechnical assistance. 1.
2.
3. 4.
5. 6.
7. 8.
REFERENCES Aggarwal, B. B., W. J. Kohr, P. E. Hass, B. Moffat, S. A. Spencer, W. J. Henzel, T. S. Bringham, G. E. Nedwin, and D. V. Goeddel. 1985. Human tumor necrosis factor, production, purification and characterisation. J. Biol. Chem. 260:2345–2354. Alexander, H. R., B. C. Sheppard, J. C. Jensen, H. N. Langstein, C. M. Buresh, D. Venzon, E. C. Walker, D. L. Fraker, M. C. Stovroff, and J. A. Norton. 1991. Treatment with recombinant human tumor necrosis factoralpha protects rats against the lethality, hypotension, and hypothermia of gram-negative sepsis. J. Clin. Invest. 88:34–39. Arakawa, T., and D. A. Yphantis. 1987. Molecular weight of recombinant human tumor necrosis factor-a. J. Biol. Chem. 262:7484–7485. Banner, D., A. D’Arcy, W. Janes, R. Gentz, H. J. Schoenfeld, C. Broger, H. Loetscher, and W. Lesslauer. 1993. Crystal structure of the soluble human 55 kd TNF receptor-human TNF beta complex: implications for TNF receptor activation. Cell 73:431–445. Bazzoni, F., and B. Beutler. 1995. How do tumor necrosis factor receptors work? J. Inflamm. 45:221–228. Bemelmans, M. H., D. Abramowicz, D. J. Gouma, M. Goldman, and W. A. Buurman. 1994. In vivo T cell activation by anti-CD3 monoclonal antibody induces soluble TNF receptor release in mice. Effects of pentoxifylline, methylprednisolone, anti-TNF and anti-IFN-gamma antibodies. J. Immunol. 153:499–506. Bemelmans, M. H. A., D. J. Gouma, and W. A. Buurman. 1993. LPS-induced sTNF-receptor release in vivo in a murine model. J. Immunol. 151:5554– 5562. Carswell, E. A., L. J. Old, R. L. Kassel, S. Greene, N. Fiore, and B. Wil-
ANTIMICROB. AGENTS CHEMOTHER.
9. 10. 11. 12.
13.
14. 15.
16. 17. 18.
19. 20.
21. 22. 23. 24. 25. 26.
27.
28.
29.
30.
31. 32.
liamson. 1975. An endotoxin-induced serum factor that causes necrosis of tumors. Proc. Natl. Acad. Sci. USA 72:3666–3670. Clark, I. A. 1978. Does endotoxin cause both the disease and parasite death in acute malaria and babesiosis? Lancet ii:75–77. Corti, A., G. Fassina, F. Marcucci, E. Barbanti, and G. Cassani. 1992. Oligomeric tumor necrosis factor a slowly converts into inactive forms at bioactive levels. Biochem. J. 284:905–910. Curfs, J. H. A. J., C. C. Hermsen, J. H. E. T. Meuwissen, and W. M. C. Eling. 1992. Immunization against cerebral pathology in Plasmodium berghei infected mice. Parasitology 105:7–14. Curfs, J. H. A. J., C. C. Hermsen, P. Kremsner, S. Neifer, J. H. E. T. Meuwissen, N. Van Rooyen, and W. M. C. Eling. 1993. Tumor necrosis factor-a and macrophages in Plasmodium berghei-induced cerebral malaria. Parasitology 107:125–134. Duncan, R. J. S., P. D. Weston, and R. Wrigglesworth. 1983. A new reagent which may be used to introduce sulfhydryl groups into proteins, and its use in the preparation of conjugates for immunoassay. Anal. Biochem. 132:68– 73. Eck, M. J., and S. R. Sprang. 1989. The structure of tumor necrosis factor-a at 2.6 Å resolution. J. Biol. Chem. 264:17595–17605. Engelmann, H., D. Novick, and D. Wallach. 1990. Two tumor necrosis factor-binding proteins purified from human urine. Evidence for immunological cross-reactivity with the cell surface tumor necrosis factor receptors. J. Biol. Chem. 265:1531–1536. Espevik, T., and J. Nissen-Meyer. 1986. A highly sensitive cell line, WEHI 164 clone 13, for measuring cytotoxic factor/tumor necrosis factor from human monocytes. J. Immunol. Methods 95:99–105. Fraker, D. L., M. C. Stovroff, M. J. Merino, and J. A. Norton. 1988. Tolerance to tumor necrosis factor in rats and the relationship to endotoxin tolerance and toxicity. J. Exp. Med. 168:95–105. Garcia, I., Y. Miyazaki, K. Araki, R. Lucas, G. E. Grau, G. Milon, Y. Belkaid, C. Montixi, W. Lesslaeuer, et al. 1995. Transgenic mice expressing high levels of soluble TNF-R1 fusion protein are protected from lethal septic shock and cerebral malaria, and are highly sensitive to Listeria monocytogenes and Leishmania major infections. Eur. J. Immunol. 25:2401–2407. Grau, G. E., P.-F. Piguet, P. Vassalli, and P.-H. Lambert. 1989. Tumor necrosis factor and other cytokines in cerebral malaria: experimental and clinical data. Immunol. Rev. 112:49–70. Hermsen, C. C., J. Van De Crommert, H. Frederix, R. W. Sauerwein, and W. M. C. Eling. 1997. Circulating tumor necrosis factor a is not involved in the development of cerebral malaria in Plasmodium berghei-infected C57BI mice. Parasite Immunol. 19:571–577. Hlodan, R., and R. H. Pain. 1995. The folding and assembly pathway of tumor necrosis factor TNFa, a globular trimeric protein. Eur. J. Biochem. 231:381–387. Johnson, R. A., T. A. Waddelow, J. Caro, A. Oliff, and G. D. Roodman. 1989. Chronic exposure to tumor necrosis factor in vivo preferentially inhibits erythropoiesis in nude mice. Blood 74:130–138. Jones, E. Y., D. I. Stuart, and N. P. C. Walker. 1989. Structure of tumor necrosis factor. Nature 338:225–228. Kumaratilake, L. M., A. Ferrante, and C. M. Rzepczyk. 1990. Tumor necrosis factor enhances neutrophil-mediated killing of Plasmodium falciparum. Infect. Immun. 58:788–793. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. Lam, K. S., P. Scuderi, and S. E. Salmon. 1988. Analysis of the molecular organization of recombinant human tumor necrosis factor (rTNF) in solution using ethylene glycolbis(succinimidylsuccinate) as the cross-linking reagent. J. Biol. Response Modif. 7:267–275. Lesslauer, W., H. Tabuchi, R. Gentz, M. Brockhaus, E. J. Schlaeger, G. Grau, P. F. Piguet, P. Pointaire, P. Vassalli, and H. Loetscher. 1991. Recombinant soluble tumor necrosis factor receptor proteins protect mice from lipopolysaccharide-induced lethality. Eur. J. Immunol. 21:2883–2886. Lewis, M., L. A. Tartaglia, G. L. Lee, G. C. Rice, G. H. W. Wong, and E. Y. Chen. 1991. Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific. Proc. Natl. Acad. Sci. USA 88:2830–2834. Lewit-Bentley, A., R. Fourme, R. Kahn, T. Prange´, P. Vachette, J. Tavernier, G. Hauquier, and W. Fiers. 1988. Structure of tumor necrosis factor by X-ray solution scattering and preliminary studies by single crystal X-ray diffraction. J. Mol. Biol. 199:389–392. Loetscher, H., R. Gentz, M. Zulauf, A. Lustig, H. Tabuchi, E. J. Schlaeger, M. Brocjhaus, H. Gallati, M. Manneberg, and W. Lesslauer. 1991. Recombinant 55-kDa tumor necrosis factor (TNF) receptor. Stoichiometry of binding to TNFa and TNFb and inhibition of TNF activity. J. Biol. Chem. 266:18324–18329. Lucas, R., J.-N. Lou, P. Juillard, M. Moore, H. Bluethmann, and G. E. Grau. 1997. Respective role of TNF receptors in the development of experimental cerebral malaria. J. Immunol. 72:143–148. Luksa, J., V. Menart, S. Milicic, and B. Kus. 1994. Purification of human tumor necrosis factor by membrane chromatography. J. Chromatogr. A 661:161–168.
VOL. 43, 1999
THIOLATED rhTNF-a PROTECTS AGAINST ECM
33. Moldawer, L. L., M. A. Marano, H. Wei, Y. Fong, M. L. Silen, G. Kuo, K. R. Manogue, H. Vlassara, H. Cohen, A. Cerami, et al. 1989. Cachectin/tumor necrosis factor-a alters red blood cell kinetics and induces anaemia in vivo. FASEB J. 3:1637–1643. 34. Møller, B., S. C. Mogensen, P. Wendelboe, K. Bendtzen, and C. M. Petersen. 1991. Bioactive and inactive forms of tumor necrosis factor-a in spinal fluid from patients with meningitis. J. Infect. Dis. 163:886–889. 35. Narhi, L. O., and T. Arakawa. 1987. Dissociation of recombinant tumor necrosis factor-a studied by gel permeation chromatography. Biochem. Biophys. Res. Commun. 147:740–746. 36. Pennica, D., G. E. Nedwin, J. S. Hayflick, P. H. Seeburg, R. Derynck, M. A. Palladino, W. J. Kohr, B. B. Aggarwal, and D. V. Goeddel. 1984. Human tumor necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 312:724–729. 37. Petersen, C. M., A. Nykjær, B. S. Christiansen, L. Heickendorff, S. C. Mogensen, and B. Møller. 1989. Bioactive human recombinant tumor necrosis factor-a: an unstable dimer? Eur. J. Immunol. 19:1887–1894. 38. Polder, T. W., W. M. C. Eling, J. H. A. J. Curfs, C. R. Jerusalem, and M. Wijers-Rouw. 1992. Ultrastructural changes in the blood-brain barrier of mice infected with Plasmodium berghei. Acta Leiden 60:31–46. 39. Postma, N. S., D. J. A. Crommelin, W. M. C. Eling, and J. Zuidema. Treatment with liposome-bound recombinant human tumor necrosis factor-a suppresses parasitemia and protects against P. berghei K173 induced experimental cerebral malaria in mice. J. Pharmacol. Exp. Ther, in press. 40. Scudiero, D. A., R. H. Shoemaker, K. D. Paull, A. Monks, S. Tierney, T. H. Nofziger, M. J. Currens, D. Seniff, and M. R. Boyd. 1988. Evaluation of a soluble tetrazolium/formazan assay for cell growth and drug sensitivity in culture using human and other tumor cell lines. Cancer Res. 48:4827–4833. 41. Smith, R. A., and C. Baglioni. 1987. The active form of tumor necrosis factor is a trimer. J. Biol. Chem. 262:6951–6954. 42. Stevenson, M. M., and E. Ghadirian. 1989. Human recombinant tumor
necrosis factor alpha protects susceptible A/J mice against lethal Plasmodium chabaudi AS infection. Infect. Immun. 57:3936–3939. Tang, P., M.-C. Hung, and J. Klostergaard. 1996. Human pro-tumor necrosis factor is a homotrimer. Biochemistry 35:8216–8225. Takahashi, N., P. Brouckaert, M. H. Bemelmans, W. A. Buurmans, and W. Fiers. 1994. Mechanism of induction of tolerance to tumor necrosis factor (TNF): no involvement of modulators of TNF bioavailability or receptor binding. Cytokine 6:235–242. Taverne, J., N. Sheikh, J. B. De Souza, J. H. L. Playfair, L. Probert, and G. Kollias. 1994. Anaemia and resistance to malaria in transgenic mice expressing human tumor necrosis factor. Immunology 82:397–403. Tracey, K. J., and A. Cerami. 1994. Tumor necrosis factor: a pleiotropic cytokine and therapeutic target. Ann. Rev. Med. 45:491–503. Utsumi, T., M.-C. Hung, and J. Klostergaard. 1992. The role of amino functions in recombinant human tumor necrosis factor in expression of biological activity. Mol. Immunol. 29:77–81. Van Ostade, X., J. Tavernier, and W. Fiers. 1994. Structure-activity studies of human tumor necrosis factors. Protein Eng. 7:5–22. Wessel, D., and U. I. Flugge. 1984. A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Anal. Biochem. 138:141–143. Wingfield, P., R. H. Pain, and S. Craig. 1987. Tumor necrosis factor is a compact trimer. FEBS Lett. 211:179–184. Yoshimura, T., S. Sone, S. Maezawa, K. Kameyama, and T. Takagi. 1988. Molecular weight of tumor necrosis factor determined by gel permeation chromatography alone or in combination with low-angle laser light scattering. Biochem. Int. 17:1157–1163. Zhang, X.-M., I. Weber, and M.-J. Chen. 1992. Site-directed mutational analysis of human tumor necrosis factor-a receptor binding site and structure-functional relationship. J. Biol. Chem. 267:24069–24075.
43. 44.
45. 46. 47. 48. 49. 50. 51.
52.
1033
