Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Tumor necrosis factor- (TNF-), originally identified by its antitumor properties and cytotoxicity to some transformed cell lines (Carswell et al., 1975), is now known to play a major role in promoting immunological, inflammatory, and pathobiological reactions (Tracey and Cerami 1994).

In malaria, TNF- plays a dual role. On the one hand, much of the pathology found during severe and cerebral malaria is considered to be mediated by high amounts of TNF- (Grau et al., 1989; Curfs et al., 1993; McGuire et al., 1994). On the other hand, low amounts of TNF- can mediate protection against malaria (Stevenson and Ghadirian 1989; Taverne et al., 1994; Kremsner et al., 1995; Jacobs et al., 1996). Recombinant human (rh) TNF--activated cells (e.g., macrophages and neutrophils) may play a role in parasite-killing (Kumaratilake et al., 1990; Taverne et al., 1994). In addition, TNF- inhibits erythropoiesis (Johnson et al., 1989; Moldawer et al., 1989), which may be involved in the suppression of parasitemia. It was recently demonstrated that low doses of rhTNF- protect against the development of experimental cerebral malaria (ECM) in Plasmodium berghei K173-infected mice and suppress parasitemia (our unpublished data). When administered by continuous infusion from i.p.-implanted osmotic pumps, about 10 times lower doses of rhTNF- were equally protective against ECM as a s.c.-administered injection of rhTNF- (unpublished data).

RhTNF- is cleared rapidly from the blood with a half-life of less than 20 min on i.v. injection into mice (Ferraiolo et al., 1988). Liver and kidney are major sites of accumulation of rhTNF- (Ferraiolo et al., 1988; Mori et al., 1996); in general, the liver plays an important role in protein clearance, although the kidney is an important site of catabolism for low molecular weight proteins (M_r < 50,000), including rhTNF- (Pessina et al., 1987; Franssen et al., 1992; McMartin, 1992). Therefore, carrier systems with a low affinity for the liver and kidney may provide an alternative approach to continuous infusion to obtain prolonged persistence of therapeutic concentrations of rhTNF-.

Liposomes improve the delivery of bioactive proteins and peptides by functioning as a circulating "microreservoir" for sustained release (Storm et al., 1991). Unfortunately, the residence time of liposomes in the bloodstream is limited because of the rapid recognition and clearance from the circulation by phagocytic cells of the mononuclear phagocyte system in liver and spleen (Senior, 1987). The development of novel formulations with, e.g., polyethylene glycol (PEG) lipid derivatives in the liposomal bilayer resulted in sterically stabilized liposomes with reduced mononuclear phagocyte system uptake and prolonged blood residence times (Blume and Cevc, 1990; Huang, 1992). Prolonged systemic delivery of the peptide drug vasopressin by long-circulating liposomes was demonstrated (Woodle et al., 1992). Thus, i.v. injection of rhTNF- encapsulated into sterically stabilized liposomes may provide an attractive alternative to continuous release from surgically implanted osmotic pumps to obtain protection against ECM.

Several other studies reported on the chemical modification of rhTNF- at the site of its primary amino groups to improve its pharmacokinetics directly (Tsutsumi et al., 1994) or after incorporation into carrier systems (Utsumi et al., 1991; Mori et al., 1996). A prolonged blood residence time of liposomal rhTNF- compared with free drug was demonstrated by Mori et al. (1996) after incorporation of a rhTNF--phospholipid conjugate into sterically stabilized liposomes. However, no therapeutic studies were performed. Coupling of rhTNF- to the outer surface of long-circulating liposomes offers another approach to increase the residence time of rhTNF- in the circulation.

Our purpose was to enhance the protective efficacy of rhTNF- against ECM by either coupling of rhTNF- to the outer surface of liposomes or by encapsulation of rhTNF- into liposomes. Two types of liposomes were used: conventional and sterically stabilized liposomes. Equal doses of bioactive rhTNF- either liposome-bound or -encapsulated rhTNF- were administered i.v. and the protection against ECM and the effect on parasitemia was compared with i.v. injection of free rhTNF-.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials

RhTNF- (Escherichia coli derived) (Pennica et al., 1984) with a specific activity of 5 × 10⁷ U/mg (bioassay, murine LM cells) was a gift from Dr. G. R. Adolf (Boehringer Ingelheim, Vienna, Austria). RhTNF- with 5 to 10% (w/v) human serum albumin (HSA) (4 mg rhTNF-/ml, 20 mM sodium phosphate, pH 7, and 400 mM NaCl) was used to prepare liposome-encapsulated rhTNF-. rhTNF- without HSA (4 mg rhTNF-/ml, 10 mM sodium phosphate, pH 7, and 200 mM NaCl) was used to couple the protein to the liposomal bilayer after reaction with N-succinimidyl S-acetylthioacetate (SATA) (Pierce, Rockford, IL). Cholesterol (Chol) was obtained from Sigma Chemical Co. (St Louis, MO). Egg phosphatidylcholine (EPC) and egg phosphatidylglycerol were donated by Lipoid (Ludwigshafen, Germany). Maleimido-4-(p-phenylbutyrate)-phosphatidylethanolamine (MPB-PE) was synthesized from succinimidyl-4-(p-maleimidophenyl)-butyrate (Pierce) and EPC (Lipoid). Distearoylphosphatidylethanolamine-poly(ethylene glycol) 2000 (PEG-DSPE) was obtained from Avanti Polar Lipids (Alabaster, AL). All other reagents used were of analytical grade.

Thiolation of rhTNF-

Basically, the procedure to thiolate proteins with SATA, as described by Duncan et al. (1983), was used. The reaction scheme is shown in Fig. 1. rhTNF- (without HSA, 10 mM sodium phosphate, pH 7, and 200 mM NaCl) was applied on a Sephadex G-25 M column (PD-10 column; Pharmacia, Uppsala, Sweden) and eluted with HEPES buffer (10 mM HEPES, 135 mM NaCl, and 1 mM EDTA, pH 7.5). The eluted rhTNF- solution was concentrated using Microsep filters (Pall Filtron, Breda, The Netherlands) with a molecular mass cut-off of 10 kDa at 4°C to approximately 2 mg rhTNF-/ml before reaction with SATA. SATA was dissolved in dimethylformamide and mixed with the rhTNF- solution in a volume ratio of dimethylformamide/buffer of 1:100 and a molar ratio of SATA/rhTNF- of 8:1. The mixture was incubated at room temperature under continuous rotation for at least 20 min to allow formation of acetylthioacetyl-rhTNF- (rhTNF-ATA) (with protected thiol groups). After incubation, the unreacted SATA was separated from rhTNF-ATA on a PD-10 column. The protein fraction in the eluate was detected by monitoring the absorption at 280 nm, collected, and stored at 20°C (maximally 3 weeks).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Reaction scheme of the introduction of thiol groups into rhTNF- by the SATA reaction and coupling of thiolated protein to anchor liposomes. Thiol groups are introduced at the site of primary amines in the protein, i.e., the terminal amino group (as shown in the scheme) and lysines throughout the protein.

Liposomal rhTNF-

Liposomes composed of EPC/egg phosphatidylglycerol/Chol/MPB-PE at 2.4:0.25:1:0.094 and EPC/PEG-DSPE/Chol/MPB-PE at 1.78:0.15:1:0.075 (molar ratios) were prepared by the thin-film extrusion method (Olson et al., 1979). These liposomes are referred to as conventional and sterically stabilized liposomes, respectively.

Liposome-Bound rhTNF-. The lipid film was hydrated with HEPES buffer (10 mM HEPES, 135 mM NaCl, 1 mM EDTA, pH 7.5) and the resulting liposome dispersion was extruded through an appropriate combination of single or stacked 0.6-, 0.2-, 0.1-, and 0.05-µm (pore size) polycarbonate membrane filters (Nuclepore; Costar, Cambridge, MA) under nitrogen pressure. These liposomes are referred to as anchor liposomes. RhTNF-ATA was deacetylated before coupling to anchor liposomes by adding a freshly prepared hydroxylamine-HCl solution (0.5 M hydroxylamine-HCl, 0.5 M HEPES, and 25 mM EDTA, pH 7.5). The mixture was incubated for 60 min at room temperature under continuous rotation at a volume ratio of rhTNF-ATA/hydroxylamine solution of 10:1, yielding thiolated rhTNF- with reactive thiol groups (acetylthio-rhTNF-). Subsequently, acetylthio-rhTNF- was incubated with anchor liposomes (protein-to-lipid ratios varied from 15 to 40 µg protein/µmol lipid) at room temperature for 75 min under continuous rotation. The liposome-bound protein was separated from free protein by two ultracentrifugation wash steps (Beckmann Optima LE-80K; Beckman Instruments, Palo Alto, CA) at 275,000g for 60 min at 4°C. The liposome dispersion was resuspended with HEPES buffer (10 mM HEPES and 135 mM NaCl, pH 7) to a lipid concentration of approximately 40 mM.

Liposome-Encapsulated rhTNF-. For the preparation of liposome-encapsulated rhTNF-, the lipid film was hydrated with 1.5 mg rhTNF-/ml (5-10% HSA, 20 mM sodium phosphate, pH 7, and 400 mM NaCl, diluted in bi-distilled water) to a lipid concentration of 40 mM. The liposomes were extruded through 0.6- and 0.2-µm polycarbonate membrane filters (Nuclepore) under nitrogen pressure. Nonencapsulated rhTNF- was removed by two ultracentrifugation wash steps (Beckmann Optima LE-80K) at 275,000g for 45 min at 4°C. The liposome dispersion was resuspended with HEPES buffer (10 mM HEPES and 135 mM NaCl, pH 7) to a lipid concentration of approximately 40 mM.

Liposome Characterization. Lipid phosphate was measured by the method of Rouser et al. (1970). Mean particle size was determined by dynamic light scattering at 25°C with a Malvern 4700 system using a 25 mW He-Ne laser (NEC, Tokyo, Japan) and the automeasure version 3.2 software (Malvern Ltd., Malvern, UK). For viscosity and refractive index, the values of pure water were used as the standard. As a measure of particle size distribution, the system reports a polydispersity index. This index ranges from 0.0 for a monodisperse dispersion and up to 1.0 for an entirely polydisperse dispersion.

Bioactivity of rhTNF- In Vitro

The bioactivity of rhTNF- either free, liposome-bound, or liposome-encapsulated (expressed as dose of free rhTNF-) was determined in the WEHI cytotoxicity assay according to the method described previously by Espevik and Nissen-Meyer (1986). WEHI 164 clone 13 mouse fibrosarcoma cells were cultured in Iscove's modified Dulbecco's medium with glutamine (GIBCO Life Technologies, Breda, The Netherlands) supplemented with 10% fetal bovine serum (Integro, Zaandam, The Netherlands), penicillin (100 IU/ml), streptomycin (100 µg/ml), and amphotericin B (0.25 µg/ml) at 37°C in a humidified atmosphere containing 5% CO₂. One day before testing, WEHI cells were transferred to new flasks and grown overnight. The next day cells were collected to a final concentration of 0.5 × 10⁶ cells/ml medium of which 50 µl were pipetted into 96-well microtiter plates (Falcon, Lelystad, The Netherlands). After incubation for 3 to 5 h at 37°C and 5% CO₂, 50 µl of a standard or test sample were added. A standard curve (up to 20,000 pg/ml) was made from rhTNF- (Boehringer Ingelheim) with an originally defined biological activity of 6 × 10⁷ U/mg in a murine LM bioassay according to WHO standards. Serial dilutions of nonliposomal rhTNF- in medium were added directly to the cells. In the case of liposomal rhTNF-, liposomes were solubilized first with 10% Triton X-100 in HEPES buffer (20 mM HEPES, 149 mM NaCl, and 0.5% bovine serum albumin, pH 7.4) in a volume ratio of 1:1. No effect of Triton X-100 was observed in the bioactivity assay of rhTNF- with a sample dilution of 30,000×. After incubation for 18 h at 37°C and 5% CO₂, the viable cells were quantitated using the sodium 3-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate assay (Scudiero et al., 1988). Briefly, 100 µl of N-methyldibenzopyrazine methyl sulfate (0.4 mg/ml in phosphate-buffered saline) from Sigma (Bornem, Belgium) were added to 5 ml of sodium 3-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate (Sigma) with a concentration of 1 mg/ml RPMI 1640 (GIBCO Life Technologies). A total of 50 µl of sodium 3-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate labeling mixture were added to the wells and incubated for 90 min at 37°C and 5% CO₂. The absorbance was measured using a Bio-Rad Novapath microplate reader (Bio-Rad Laboratories, Veenendaal, The Netherlands) at a wavelength of 490 nm (reference wavelength: 655 nm). The lower detection limit was approximately 80 pg rhTNF-/ml. In the above WEHI cytotoxicity assay, the specific biological activity of rhTNF- (U/mg) used for the experiments was about two to four times higher than for the standard rhTNF- (Boehringer Ingelheim). The two batches rhTNF-, with or without HSA, showed a similar specific biological activity (U/mg).

Animal Model of Malaria

Mice. Female C57Bl/6J mice, 6 to 10 weeks old, 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 rat-mouse-hamster food (Hope Farms, Woerden, The Netherlands) ad libitum.

Parasitemia and Development of Experimental Cerebral Malaria. Mice were infected with the murine parasite P. berghei K173. The parasite was maintained by weekly transfer of parasitized erythrocytes from infected into naive mice. About 95% of C57Bl/6J mice infected i.p. with 10³ parasitized erythrocytes die in the second week after infection (days 9-12) because of the development of ECM (Curfs et al., 1992). One day before death, a progressive hypothermia develops that is strongly correlated with development of hemorrhages in the brain as observed by histology (Polder et al., 1992). Mice that survive this critical period only show a limited, transient hypothermia and die in the third week or later after infection with severe anemia and a parasitemia of 20 to 40% but without any noticeable cerebral pathology. Time of death after infection was used as the parameter for ECM or protection against ECM in the experiments described herein.

Experimental Design

For each experiment mice received 10³ parasitized erythrocytes i.p. on day 0. At day 5 after infection mice received a single i.v. injection of either free rhTNF- or liposome-bound or -encapsulated rhTNF-. Before injection, appropriate dilutions were prepared using HEPES buffer (10 mM HEPES, 135 mM NaCl, and 1 mM EDTA, pH 7.5) containing 1% mouse plasma. Placebo-treated mice received either anchor liposomes or HEPES buffer only. Control mice did not receive any treatment (nontreated mice). The effect of treatment on parasitemia (percentage of infected erythrocytes) at day 8 after infection was studied from thin blood films made from tail-blood and stained with May-Grünwald and Giemsa's solutions. A considerable variation in parasitemia is observed among infected control groups of independent repeat experiments (average parasitemia ± S.D.: 17 ± 8%; n = 8 experiments). Therefore, within each experiment the average parasitemia in nontreated mice was determined and the parasitemia of each individual mouse from either a control 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 survival of mice. Death of mice within 2 weeks after infection was used to identify ECM death (Curfs et al., 1992). 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. The effect of treatment on protection against ECM was analyzed by the ² test. Differences were considered significant at a level  = .05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of Liposome-Bound rhTNF-

Thiolated rhTNF- was incubated with either conventional or sterically stabilized anchor liposomes at different protein-to-lipid ratios, varying from 15 to 40 µg protein/µmol lipid. The bioactivity of rhTNF- in the final liposome dispersion was determined by the WEHI cytotoxicity assay and expressed as dose of free rhTNF-. Typical binding was about 0.9 µg bioactive protein/µmol lipid for both types of liposomes, with a range of 0.4 to 1.4 in 8 experiments. The protein-to-lipid ratios in the final product were independent of the protein-to-lipid ratios during incubation. The influence of protein-to-lipid ratio during incubation on liposome size and polydispersity is shown in Table 1. The average liposomal size and the polydispersity index of the liposomes increased with increasing protein-to-lipid ratios used during incubation. This effect was more pronounced for conventional liposomes than for sterically stabilized liposomes. Large aggregates were formed at protein-to-lipid ratios of 35 to 40 µg protein/µmol lipid.

View this table: [in this window] [in a new window]   TABLE 1 Effect of protein/lipid ratio during incubation with thiolated rhTNF- on liposome size Preformed, sized anchor liposomes were incubated with varying amounts of thiolated rhTNF- for 60 min at room temperature. Directly after coupling of the protein, the size of the liposomes was measured by dynamic light scattering, yielding an average particle size and polydispersity index.

Characterization of Liposome-Encapsulated rhTNF-

Nonmodified rhTNF- was encapsulated into conventional or sterically stabilized liposomes by the film method and the ratio of bioactive protein-to-lipid (in micrograms per micromoles) was 2.3 and 2.8 for conventional and sterically stabilized liposomes, respectively. The encapsulation efficiency was approximately 6%. The average size of both liposomal dispersions was 0.2 µm (polydispersity index <0.1).

Effect of rhTNF- Treatment on Development of ECM

Intravenous injection of an increasing dose of free rhTNF- (2-9 µg bioactive protein) at day 5 after infection protected up to 15% of the mice against the development of ECM (Fig. 2). Intravenous injection of rhTNF- (1-9 µg), either encapsulated into conventional or sterically stabilized liposomes, protected up to 10% and 17% of the mice, respectively. The protective efficacy of rhTNF- was significantly (p < .0001, summarized data of all doses) enhanced by treatment with rhTNF- coupled to either conventional (1-4 µg) or sterically stabilized liposomes (1-9 µg) (Fig. 2). Approximately 55 to 100% of the mice were protected against ECM at doses of 2 to 9 µg protein. Treatment with 9 µg rhTNF- coupled to conventional liposomes caused death, probably as a result of large aggregates in the liposomal dispersion. Placebo-treated and nontreated mice were not protected against ECM (data not shown).

View larger version (34K): [in this window] [in a new window]   Fig. 2.   The effect of treatment with rhTNF- during infection on protection against ECM. At day 5 after infection with 103 P. berghei parasites, mice received a single i.v. injection of free rhTNF- (2-9 µg), liposome-encapsulated rhTNF- (1-9 µg), or liposome-bound rhTNF- (1-9 µg). 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 dose of rhTNF-. The number of mice in each treatment group varied from 5 to 16 animals. died because of toxicity.

Effect of rhTNF- Treatment on Parasitemia

The effect of an i.v. injection of free or liposome-bound or -encapsulated rhTNF- at day 5 after infection on parasitemia at day 8 was determined. The data are expressed as the ratio of the parasitemia of a certain mouse divided by the mean parasitemia of the infected but otherwise nontreated mice in the same experiment. Calculation of this ratio permitted us to summarize the results of independent repeat experiments. In the experimental protocols used and in the studied dose range, the effect of treatment on parasitemia was independent of the dose of rhTNF-, as is shown for liposome-bound rhTNF- in Fig. 3, A and B. No differences were observed between treatment with conventional or sterically stabilized liposomes. Similar data were obtained after treatment with free or liposome-encapsulated rhTNF-. The summarized data of different doses of rhTNF- for the various rhTNF- formulations are shown in Table 2. The overall data show that parasitemia was significantly suppressed (p < .05) by treatment with liposome-bound rhTNF- but not by treatment with either free or liposome-encapsulated rhTNF- (Table 2). When the effect of treatment on parasitemia was studied in relation to protection against ECM, parasitemia was significantly (p < 0.05) suppressed in ECM-protected mice but not in mice developing ECM after treatment with either free rhTNF- or liposome-bound rhTNF- (Table 2). Injection of HEPES buffer only or of anchor liposomes did not affect parasitemia (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   A and B, The effect of treatment with rhTNF- coupled to liposomes during infection on parasitemia. Mice received a bolus i.v. injection of rhTNF- coupled to conventional liposomes (A, 1-4 µg rhTNF-) or sterically stabilized liposomes (B, 1-9 µg rhTNF-) at day 5 after infection with 103 P. berghei parasites. The average parasitemia at day 8 after infection was determined in each control group and the parasitemia of each individual mouse was divided by the average of the corresponding control group. The ratio is plotted against the dose of rhTNF- for each mouse. When treatment has no effect on parasitemia the ratio scatters around 1, although ratios <1 indicate that parasitemia is suppressed.

View this table: [in this window] [in a new window]   TABLE 2 Effect of rhTNF- treatment on parasitemia in relation to development of experimental cerebral malaria in P. berghei-infected mice Mice received a single i.v. injection of either free rhTNF-, liposome-encapsulated rhTNF- (conventional or sterically stabilized liposomes), or liposome-bound rhTNF- (conventional or sterically stabilized liposomes) at day 5 after infection with 103 parasites. Doses of bioactive protein ranged from 1 to 9 µg. The average parasitemia at day 8 after infection was determined in each control group and the parasitemia of each individual mouse was divided by the average of the corresponding control group. The ratios were used for further analysis. Values represent averages ± S.D. of the summarized data of the different doses of rhTNF-. a p < 0.05 compared with nontreated mice, b p < 0.05 compared with mice developing ECM in the same treatment group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Our principle finding is that treatment with liposome-bound rhTNF- shows enhanced protection against ECM and suppression of parasitemia as compared with treatment with free rhTNF-. The efficacy was the same for i.v.-injected rhTNF- coupled to either conventional or sterically stabilized liposomes. The presence of PEG-DSPE in the bilayer reduced to some extent the formation of large aggregates at high protein-to-lipid ratios by steric hindrance (Table 1). Encapsulation of rhTNF- into liposomes did not improve the protective efficacy of i.v.-injected rhTNF- against ECM. A possible explanation is that liposome-encapsulated rhTNF- was cleared rapidly from the circulation and that no sufficient sustained release was obtained. The encapsulation efficiency of rhTNF- into liposomes as determined by the WEHI bioassay was approximately 6%. This is in agreement with encapsulation efficiencies reported in the literature by radiolabeled rhTNF- (2-12%) (Debs et al., 1989; Mori et al., 1996). The relatively low encapsulation efficiency is caused by the hydrophilic nature of the nonmodified rhTNF-.

The mechanism by which rhTNF- protects against the development of ECM remains to be elucidated. Our data suggest that reduction of parasitemia plays a role, because the reduction was observed in ECM-protected mice but not in rhTNF--treated mice developing ECM (Table 1). It should be noted that the parasitemia in controls varies extensively before they die of ECM, indicating that the number of circulating parasites on a given day is not critical for development of ECM (Curfs et al., 1992). Thus, the suppression of parasitemia by treatment with rhTNF- may either interfere with the development of ECM or is an associated but not causally related factor in protection against ECM. Another possible mechanism that may be involved may be the development of refractoriness or tolerance after treatment with rhTNF- or stimuli that provoke TNF- release (Alexander et al., 1991; Takahashi et al., 1995). Refractoriness or tolerance persists for several days and might be one of the "natural" regulatory mechanisms to control excessive stimulation by this cytokine (Takahashi et al., 1993).

The enhanced protective efficacy of liposome-bound rhTNF- compared with free rhTNF- may be the result of an increased residence time of rhTNF- in the bloodstream, as was demonstrated by Mori et al. (1996) after incorporation of a lipid-rhTNF- conjugate into sterically stabilized liposomes. Surprisingly, similar results were obtained after treatment with rhTNF- coupled to either conventional or sterically stabilized liposomes (Table 2 and Fig. 2). This may either argue against a difference in circulation time of the two formulations or indicate that activity is not directly related to differences in the blood residence time of the two types of liposomes.

Another explanation for the enhanced efficacy of liposome-bound rhTNF- may be the stabilization of the bioactive configuration of rhTNF-. rhTNF- exists as a noncovalent trimer in solution, and this trimeric state of rhTNF- is considered to be the biologically active form (Smith and Baglioni 1987; Van Ostade et al., 1994). rhTNF- trimers have been shown to be unstable at low concentrations, resulting in the formation of inactive monomers and aggregates (Narhi and Arakawa 1987; Corti et al., 1992). Each monomer contains several primary amino groups (one terminal valine and six lysyl residues; Van Ostade et al., 1994) that may react with SATA. It is hypothesized that the introduction of several reactive thiol groups in rhTNF- by the SATA reaction leads to the formation of disulfide bridges between monomers in the trimeric configuration. The disulfide-stabilized trimer of rhTNF- is subsequently coupled to the liposome surface. This may prevent or delay the dissociation of rhTNF- into inactive monomers upon injection in vivo, resulting in prolonged concentrations of bioactive trimeric rhTNF- and enhanced access to deeper body compartments. This may also explain why encapsulation of rhTNF- into liposomes could not improve the protective efficacy of rhTNF-.

In summary, i.v. injection of liposome-bound rhTNF- significantly enhanced protection against P. berghei-induced ECM in mice as compared with i.v. injection of free rhTNF-. Protection against ECM was associated with suppression of parasitemia after treatment with any of the rhTNF- formulations. It is hypothesized that thiolation of rhTNF- and coupling to the liposomal bilayer stabilizes the bioactive form of rhTNF- and therefore leads to prolonged concentrations of bioactive rhTNF-.