INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Mycobacteria are intracellular parasites which replicate within the shielded environment of monocyte-derived tissue macrophages. Activation of antibacterial killing mechanisms within these cells by cytokines is essential for the control of mycobacterial infections (24). Gamma interferon (IFN-) plays a central role in this since it is produced by a variety of lymphocytes responding to mycobacterial infections, including CD4⁺ and CD8⁺ T cells and T cells. Administration of recombinant IFN- protects mice against lethal Mycobacterium tuberculosis infection in some but not all experimental models (13, 20), whereas neutralization with anti-IFN- antibodies exacerbates the infection (13). The failure of mice deficient in IFN- or IFN- receptors to control M. tuberculosis infection confirms that this cytokine is essential for killing M. tuberculosis (12, 20). Studies with human and murine macrophages, however, have demonstrated that additional signals are required to fully activate mycobacterial killing (24). Potential activators include other cytokines, such as tumor necrosis factor (TNF), interleukin-4 (IL-4), IL-6, and granulocyte-macrophage colony-stimulating factor (15, 18, 19), and in humans 1,25-dihydroxy-vitamin D₃, the biologically active form of vitamin D₃ (14). TNF alone cannot activate macrophages sufficiently to kill mycobacteria, but it does synergize with IFN- to increase the antimycobacterial activity of infected macrophages in vitro (18). Administration of anti-TNF antibodies decreases the resistance of mice to infection with M. bovis bacillus Calmette-Guérin (BCG) (25) and M. tuberculosis (21). TNF is a necessary requirement for effective antimycobacterial immunity, since mice deficient in the 55-kDa TNF receptor I (TNFRI) develop progressive M. tuberculosis infection (21). The protective effects of TNF and the lethal consequences of anti-TNF antibodies have been observed in other models of intracellular bacterial infection, including infections by Listeria monocytogenes, Salmonella typhimurium, and Legionella spp. (37). Although in mycobacterial infections, such as leprosy, high levels of TNF have been associated with tissue damage and systemic toxicity, local TNF synthesis is essential for the control of mycobacterial infections (35).

Studies with neutralizing anti-human TNF monoclonal antibodies (MAb) demonstrated that the sequence from amino acids 65 to 85 of the TNF molecule was involved in binding to the TNF receptor (32). By use of truncated peptides, amino acids 70 to 80 were identified as essential for TNF activity (33). When this peptide sequence was modified by substitution of leucine-76 for isoleucine, the subsequent peptide TNF_(70-80) had increased stability in vitro in the presence of serum (32a) and possessed TNF mimetic properties both in vitro and in vivo (27). TNF_(70-80) stimulated a reactive oxygen burst in human and murine neutrophils (27) and activated human neutrophils to kill Plasmodium falciparum (27). In a murine model of Plasmodium chabaudi infection, systemic therapy with TNF_(70-80) increased the rate of recovery and clearance of parasites (27). More recently, TNF_(70-80) was found to reduce the weight loss and systemic effects in mice chronically infected with Pseudomonas aeruginosa (32a). The demonstrated properties of TNF_(70-80) and the known requirement of TNF for activating macrophages led us to examine whether this mimetic peptide would have antimycobacterial activity on a murine macrophage cell line. We now report that TNF_(70-80) synergizes with IFN- to activate murine macrophages to inhibit the growth of M. bovis BCG and that this property is dependent on its activation of inducible nitric oxide synthase (iNOS).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria and cell line. M. bovis BCG (CSL strain) was obtained from CSL (Melbourne, Australia) and grown in Middlebrook 7H9 broth supplemented with OACD (Difco, Detroit, Mich.) and 0.5% Tween 80 (Sigma, St. Louis, Mo.). The bacteria were stored in 30% glycerol at 70°C. After being thawed, the number of viable bacteria was determined by culture of serial dilutions on 7H11 agar containing OACD and glycerol for 3 weeks. Prior to use the BCG cells were washed and suspended in RPMI 1640 (Flow, Sydney, Australia) containing 10% fetal calf serum (CSL) and 2 mM L-glutamine (experimental medium) and sonicated briefly before infection of cells. The RAW264.7 cells (ATCC TIB 71) were kindly provided by G. Chaudhari (University of Sydney, Australia). These were maintained in Dulbecco modified Eagle medium (Gibco BRL, Gaithersburg, Md.) with 10% fetal calf serum, 2 mM L-glutamine, 100 U of penicillin (CSL) per ml, and 100 mg of streptomycin (CSL) per ml at 37°C in 5% CO₂. The cells were washed twice with RPMI before infection with M. bovis BCG.

Cytokines, peptide, and antibodies. The sequence of the TNF_(70-80) peptide is H-Pro-Ser-Thr-His-Val-Leu-Ile-Thr-His-Thr-Ile-OH. The peptide was synthesized by the F-moc-polyamine method (4) of solid-phase peptide synthesis with the PepSyn KA solid resin (33) and purified by high-pressure liquid chromatography. Murine IFN- (10⁶ U/ml) and TNF (2 × 10⁷ U/ml) were purchased from Genzyme (Cambridge, Mass.). Recombinant human TNF (1.7 × 10⁷ U/ml) was provided by Peptech Ltd. (Sydney, Australia). Fresh dilutions of the peptides and cytokines were prepared daily in experimental medium. The relative potencies of human TNF and TNF_(70-80) were compared for the stimulation of nitric oxide (NO) release, and 5.0 µg of TNF_(70-80) per ml was found to have activity corresponding to 1,000 U of human TNF per ml. Hamster anti-IFN- MAb antiserum was purchased from Genzyme. Anti-human TNF MAb 054 (immunoglobulin G1 [IgG1]) (32) was used as an ammonium sulfate precipitate of ascites fluid. Isotype-matched control MAb L5 (IgG1) binds the M. leprae 18-kDa protein (8). A soluble form of the human TNFRI receptor composed of the recombinant 55-kDa TNFRI fused to the human IgG heavy-chain domain (TNFRI-IgG, designated p55-sf2) (9) was kindly provided by B. Scallon (Centocor, Malvern, Pa.) along with control immunoglobulin.

M. bovis BCG culture in macrophage cell line. RAW264.7 cells were washed and adjusted to 10⁶/ml in experimental medium without antibiotics. Portions (100 µl; 10⁵ cells) were distributed into microtiter wells (Nunc, Roskilde, Denmark) and incubated for 6 h in 5% CO₂ at 37°C. The cells were then incubated with various concentrations of cytokines or the peptide for 16 h. After the cells were washed in warm RPMI, 100 µl of experimental medium was added to them. They were then infected by adding 100 µl of RPMI containing 10⁵, 10⁶, or 10⁷ BCG organisms and cultured for 4 days. A multiplicity of infection of 10:1 resulted in optimal BCG growth and was used in subsequent experiments. The culture supernatants were collected from the cytokine-stimulated cells at various time points and tested immediately or stored at 70°C for NO₂ measurement.

Assay for BCG growth. After 3 days of culture, culture supernatants were gently aspirated, and the infected RAW264.7 cells were washed twice with warm medium to remove any extracellular mycobacteria. The cells were lysed by adding 100 µl of 0.1% saponin (Sigma) in experimental medium, followed by the addition of 100 µl of ³H-uracil (Amersham, Amersham, United Kingdom) diluted to 10 µCi/ml in RPMI. The plates were incubated for a further 24 h at 37°C in 5% CO₂ to allow incorporation of ³H-uracil into the RNA of viable mycobacteria. Mycobacteria were then harvested onto glass microfiber filters (Whatman, Maidstone, United Kingdom), and ³H-uracil incorporation was determined by liquid scintillation spectroscopy. The percent inhibition of ³H-uracil incorporation into M. bovis BCG was calculated as follows: (mean of triplicate cultures without cytokines  mean of triplicate cultures with cytokines)/(mean of triplicate cultures without cytokines).

Nitric oxide measurements. Nitrite levels were measured by using the Greiss reagent (22). Briefly, Greiss reagent was freshly prepared by mixing 3.0% phosphoric acid, 1.0% sulfanilamide, and 0.1% n-(1-naphthyl)ethylenediamine (Sigma) in distilled water. Then 100-µl portions of the culture supernatants were incubated with Greiss reagent in microtiter trays in triplicate for 10 min at room temperature and the optical density at 540 nm was measured. The nitrite levels were determined from a standard curve prepared by using serial dilutions of sodium nitrite (Sigma) in water. In some experiments competitive inhibitors of iNOS, n^G-monomethyl-L-arginine monoacetate (n-MMA) (Calbiochem-Behring, San Diego, Calif.) or aminoguanidine bicarbonate (ICN, Cosa Mesa, Calif.), were added to the culture medium at concentrations of 0.01 to 10 mM from the time of incubation of the RAW264.7 cells with the cytokines.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Synergistic effect of TNF and IFN- on growth of BCG. The macrophage cell line RAW264.7 supported the growth of BCG in vitro with an optimal multiplicity of infection of 10:1 and 10⁵ cells/well. After 3 days the adherent macrophages were lysed, and the replication of BCG was confirmed by determining the uptake of ³H-uracil into viable mycobacteria. Maximal uptake of 50,000 to 60,000 cpm occurred after 16 h. When the RAW264.7 cells were pretreated with IFN- alone, there was partial reduction in ³H-uracil uptake, resulting in 57% inhibition of BCG replication at an IFN- concentration of 64 U/ml (Fig. 1A). When human TNF was added at increasing doses, there was a further reduction in ³H-uracil uptake so that at 64 U of IFN- per ml and at 1,000 U of TNF per ml there was complete inhibition of BCG growth (Fig. 1A). This was associated with the dose-dependent production of NO by the macrophages as measured by the release of nitrite into the culture supernatant (Fig. 1B). After stimulation with IFN- and TNF, maximum nitrite release by the RAW264.7 cells occurred after 3 days of culture (data not shown). When TNF alone was used, the maximum nitrite release was 1.4 µg/ml at a TNF concentration of 1,000 U/ml with 28% inhibition of BCG growth. Murine TNF had an effect similar to that of human TNF in synergizing with IFN- to stimulate nitric oxide release and inhibit BCG replication (data not shown).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Activation of the macrophage cell line RAW264.7 by IFN- and human TNF (hTNF). Activation was determined by measuring the inhibition of BCG growth as reflected by 3H-uracil incorporation (A) and by nitrite release at 3 days postinfection (B). The cells were incubated with IFN- alone () or IFN- with increasing concentrations of hTNF (, 64 U/ml; , 250 U/ml; , 1,000 U/ml) for 24 h prior to BCG infection. The maximum uptake of 3H-uracil into replicating M. bovis BCG in the absence of cytokines ± the standard deviation was 54,780 ± 3,062 cpm.

TNF_(70-80) peptide and IFN- inhibit BCG growth. TNF_(70-80) peptide is based on the sequence from amino acids 70 to 80 of the human TNF protein. When the TNF_(70-80) peptide was added with IFN-, a similar response was observed with dose-dependent NO production and BCG growth inhibition (Fig. 2). At concentrations of 1.25 µg of TNF_(70-80)/ml and of 64 U of IFN-/ml there was almost complete inhibition of ³H-uracil uptake. TNF_(70-80) peptide alone (at 5.0 µg/ml) stimulated low levels of nitrite release (maximum, 1.5 µg/ml), resulting in a maximum growth inhibition of 26%, but synergized with IFN- to stimulate nitrite levels of 17 µg/ml. Control peptide from other regions of human TNF [TNF_(6-18)] had no synergistic effect with IFN- to activate RAW264.7 cells (data not shown).

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Activation of the macrophage cell line RAW264.7 by IFN- and peptide TNF(70-80). Activation was determined by measuring the inhibition of BCG growth as reflected by 3H-uracil incorporation (A) and by nitrite release at 3 days postinfection (B). The cells were incubated with IFN- alone () or IFN- with increasing concentrations of peptide TNF(70-80) (, 0.31 µg/ml; , 1.25 µg/ml; , 5.0 µg/ml) for 24 h prior to BCG infection. The maximum uptake of 3H-uracil into replicating M. bovis BCG in the absence of cytokines ± the standard deviation was 56,204 ± 3,465 cpm.

Effect of anti-cytokine antibodies on TNF_(70-80) peptide activity. The inhibitory effects of TNF_(70-80) on BCG growth were dependent on coactivation with IFN-. When antibodies that neutralize IFN- were added during the activation phase, there was reduced nitrite release and inhibition of BCG growth (Table 1). Neutralizing MAb to human TNF blocked the effect of both human TNF and TNF_(70-80) on BCG-infected RAW264.7 cells (Fig. 3). This anti-TNF MAb neutralizes the activity of human, but not murine, TNF (32) and had no effect on the activation of macrophages by murine TNF and IFN- (data not shown).

View this table: [in this window] [in a new window]   TABLE 1.   Effect of anti-murine IFN- antibodies on the activation of antimycobacterial activity of RAW246.7 cells by the combination of IFN- with TNF or peptide TNF(70-80) or by IFN- alonea

View larger version (25K): [in this window] [in a new window]   FIG. 3.   Inhibitory effect of anti-human TNF antibodies on the activation of the macrophage cell line RAW264.7 by the combination of IFN- and either hTNF or peptide TNF(70-80). The inhibitory effect was determined by measuring the suppression of antimycobacterial activity as reflected by 3H-uracil incorporation into BCG (A) and by nitrite release at 3 days postinfection (B). The cells were incubated with IFN- (100 U/ml) and hTNF (250 U/ml) (open columns) or with IFN- (100 U/ml) and peptide TNF(70-80) (1.25 µg/ml) (shaded columns) with increasing concentrations of anti-hTNF antibodies for 24 h prior to BCG infection.

TNF_(70-80) peptide acts through the TNFRI. The fusion protein, p55-sf2, inhibits the binding of human and murine TNF to the 55-kDa TNFRI (9). When the RAW264.7 cells were incubated with TNF, IFN-, and increasing concentrations of TNFRI-IgG, macrophage activation was blocked, with reduction in NO release and increased ³H-uracil uptake by the BCG (Fig. 4). TNFRI-IgG also blocked the activation of RAW264.7 cells by TNF_(70-80) (Fig. 4), indicating that the peptide signaling is dependent on the TNFRI.

View larger version (29K): [in this window] [in a new window]   FIG. 4.   Inhibitory effect of a soluble TNFRI-IgG fusion protein on the activation of the macrophage cell line RAW264.7 by the combination of IFN- and either hTNF or peptide TNF(70-80). The inhibitory effect was determined by measuring the suppression of antimycobacterial activity as reflected by 3H-uracil incorporation into BCG (A) and by nitrite release at 3 days postinfection (B). The cells were incubated with IFN- (100 U/ml) and hTNF (250 U/ml) (open columns) or IFN- (100 U/ml) and peptide TNF(70-80) (1.25 µg/ml) (shaded columns) with increasing concentrations of soluble TNFRI-IgG fusion protein for 24 h prior to BCG infection.

Activity of TNF_(70-80) peptide is dependent on NO. The competitive inhibitor n-MMA blocks the production of NO by inducible NO synthase. n-MMA at 1 mM blocked the induction of NO release by the combination of IFN- and either TNF or TNF_(70-80) (Fig. 5B) and the accompanying inhibition of BCG growth (Fig. 5A). Aminoguanidine also inhibits inducible NO synthase. At concentrations of 1 to 10 mM, aminoguanidine blocked the effects of both TNF and TNF_(70-80) (Table 2). Therefore, the inhibition of BCG growth by TNF_(70-80) is dependent on NO production.

View larger version (13K): [in this window] [in a new window]   FIG. 5.   Inhibitory effect of n-MMA on the activation of the macrophage cell line RAW264.7 by the combination of IFN- with hTNF, murine TNF, or peptide TNF(70-80). The inhibitory effect was determined by measuring the suppression of antimycobacterial activity as reflected by 3H-uracil incorporation into BCG (A) and by nitrite release at 3 days postinfection (B). The cells were incubated with IFN- alone (100 U/ml) (), IFN- (100 U/ml) and hTNF (250 U/ml) (), IFN- (100 U/ml) and murine TNF (250 U/ml) (), or IFN- (100 U/ml) and peptide TNF(70-80) (1.25 µg/ml) () and increasing concentrations of n-MMA for 24 h prior to BCG infection.

View this table: [in this window] [in a new window]   TABLE 2.   Inhibitory effect of aminoguanidine on the activation of antimycobacterial activity of RAW264.7 cells by the combination of IFN- and either TNF or peptide TNF(70-80)a

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The peptide TNF_(70-80) demonstrated properties similar to those of soluble human and murine TNF in synergizing with IFN- to activate the murine macrophage cell line and to inhibit the replication of intracellular mycobacteria. This peptide sequence is close to the region of human TNF and its homolog, alpha lymphotoxin, which is considered to interact with the TNF receptor (5, 17). The MAb 054, which neutralizes the action of human TNF on tumor cells (32), blocked the activation of murine macrophages by both human TNF and TNF_(70-80). TNF engages two distinct receptors, the 55-kDa TNFRI and the 75-kDa TNFRII, on the surfaces of leukocytes, with the resultant signaling inducing a wide range of biological effects including apoptosis, macrophage activation, and cellular proliferation (6). A soluble form of the TNFRI receptor has been demonstrated to protect mice against endotoxic shock (3) and to block TNF-induced pathology in experimental allergic encephalitis (26). The TNFRI-IgG fusion protein abrogated the activity of TNF and TNF_(70-80) on macrophage activation (Fig. 4), a finding consistent with the signaling induced by TNF_(70-80) occurring via the TNFRI.

Signaling via the TNFRI is essential for controlling intracellular bacterial infections. Mice deficient in the 55-kDa TNFRI are highly susceptible to infection with L. monocytogenes (31) or with M. tuberculosis, the latter type of infection resulting in uncontrolled replication of the organisms and the failure to develop epithelioid cells within granulomas (21). This is consistent with the observations that the treatment of mice with recombinant TNF enhanced resistance to a lethal challenge with L. monocytogenes (16) and that endogenous TNF was essential for the enhanced resistance conferred by therapy with exogenous recombinant IFN- (16, 28). TNF is required both to control acute mycobacterial infections and to prevent reactivation in the chronic stage of infection. Treatment of mice with chronic tuberculosis infection with an adenovirus expressing the gene for the p55 TNFRI resulted in exacerbation of the disease and the destruction of granulomas (1). Recently, a child with disseminated M. avium infection was found to have defective TNF production, confirming the importance of TNF in human mycobacterial infections (36).

The inhibitory effect of TNF_(70-80) on BCG replication was dependent on the synthesis of reactive nitrogen metabolites. Both human TNF and TNF_(70-80) synergized with IFN- to activate iNOS, with the maximum release of nitrite occurring at 3 days. When the competitive inhibitors of iNOS function, n-MMA and aminoguanidine, were added to the BCG-infected macrophages, the mycobacterial inhibitory effects of both TNF_(70-80) and TNF were lost. NO is also essential for the killing of M. tuberculosis in vitro by murine macrophages (11), although there is variability in the sensitivity to macrophage-generated NO among different virulent strains of M. tuberculosis (34). This requirement for reactive nitrogen intermediates was confirmed by the observation that in vivo treatment with iNOS inhibitors prevented mice from controlling the replication of M. tuberculosis (10). Mice genetically deficient in iNOS are unable to restrain M. tuberculosis (30) and other intracellular pathogens such as L. monocytogenes (29), emphasizing the central role of reactive nitrogen intermediates in the bactericidal activity of murine macrophages.

The TNF mimetic properties of TNF_(70-80) have been demonstrated with other infectious agents. TNF_(70-80) stimulated human polymorphonuclear neutrophils to undergo a respiratory burst and to release their granular contents, leading to enhanced killing of P. falciparum in vitro (27). This activity was not due to contaminating lipopolysaccharide (LPS), since the activity of TNF_(70-80) was destroyed by boiling, against which LPS is resistant (27). Further, the addition of polymyxin B, which binds and inactivates LPS, to the TNF_(70-80) had no effect on its activity. Treatment of mice infected with P. chabaudi with the peptide significantly reduced the parasitemia, with the effect observable within 7 h, whereas a control peptide had no effect. The immunostimulatory properties of TNF_(70-80) were also evident in chronic P. aeruginosa infection of mice, where peptide therapy resulted in reduced weight loss and enhanced clearance of the organisms (32a).

The activities of other cytokines have been mimicked by short peptides. A nonpeptide sequence derived from amino acids 163 to 171 of IL-1, which corresponds to an amphipathic region crucial for binding to the IL-1 receptor (7), has immunostimulatory activity without the pyrogenic and inflammatory effects of IL-1 (2). This peptide enhanced the antitumor efficacy of anti-idiotype vaccines directed against a mouse B-cell lymphoma when delivered as a fusion protein or as a DNA vaccine encoding the fusion protein (23). Small peptides generated from a random phage display library bound and activated the erythropoietin receptor (38). The peptides mimicked the action of the hormone erythropoietin by stimulating erythropoiesis in mice, even though the amino acid sequences of the peptides were not present in the primary sequence of erythropoietin (38). Both TNF and the related cytokine alpha lymphotoxin are considered to bind to the TNF receptors as trimers, although the exact nature of interaction is unresolved (6). The mechanism by which the TNF mimetic peptide, TNF_(70-80), binds and signals via the TNFRI is currently under investigation.

In related studies, the peptide TNF_(70-80) has been found to be free of toxicity when used in mice in concentrations of up to 400 mg/kg (27). The peptide has neutrophil and macrophage stimulatory properties without the adverse effects associated with TNF administration. This suggests the peptide could be used to modify the response to infections with intracellular pathogens. Treatment of BCG-infected mice with TNF_(70-80) modulated the pathological response to the organism (34a). It has been more difficult to demonstrate cytokine modulation of antimycobacterial activity in human macrophages (24). Further studies are required to confirm whether the TNF_(70-80) peptide has similar effects on mycobacterium-infected human macrophages and to determine whether it has the potential to synergize with chemotherapy to increase the rate of clearance of organisms.