Upregulation of p75 Tumor Necrosis Factor Alpha Receptor in ...

INFECTION AND IMMUNITY, Nov. 1999, p. 5762–5767 0019-9567/99/$04.00⫹0 Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Vol. 67, No. 11

Upregulation of p75 Tumor Necrosis Factor Alpha Receptor in Mycobacterium avium-Infected Mice: Evidence for a Functional Role ANGELO CORTI,1 LANFRANCO FATTORINI,2 OVE FREDRIK THORESEN,2 MARIA LUISA RICCI,2 ANNA GALLIZIA,1 MICAELA PELAGI,1 YONGJUN LI,2 AND GRAZIELLA OREFICI2* DIBIT, San Raffaele H Scientific Institute, Milan,1 and Laboratory of Bacteriology and Medical Mycology, Istituto Superiore di Sanita `, Rome,2 Italy Received 5 May 1999/Returned for modification 29 June 1999/Accepted 9 August 1999

The bacterial growth and the production of tumor necrosis factor alpha (TNF-␣) and TNF receptors (TNF-Rs) in the spleen and blood of BALB/c mice challenged with Mycobacterium avium complex (MAC) were monitored. Infection developed in two phases: the first, up to day 21, was associated with rapid MAC multiplication in the spleen and a drop in the mycobacteremia, and the second was associated with control of the infection in both compartments. In the spleen, TNF-␣ and TNF-RII mRNA levels peaked on day 21 and then slowly decreased; however, no increase in the level of TNF-RI mRNA was observed throughout these experiments. The level of circulating soluble TNF-RII (sTNF-RII) was transiently increased after day 21. In a model in which overproduction of bioactive TNF-␣ was triggered in response to a second infection with MAC, an increased production of sTNF-RII by cultured splenocytes was also observed. Administration of an antagonist anti-TNF-RII monoclonal antibody (MAb 6G1) to infected mice inhibited the bacterial growth in the spleen, suggesting that the TNF-RII and/or sTNF-RII was functionally involved in the mechanisms that control the infection. Overall, these observations suggest that upregulation of TNF-RII or sTNF-RII contributes to modulation of the TNF-␣ antibacterial activity in MAC infections. Organisms belonging to the Mycobacterium avium complex (MAC) are rarely pathogenic for healthy individuals but may become a major cause of disseminated bacterial infection in human immunodeficiency virus-infected patients (18). MAC can survive within macrophages (M␾) and affect various physiological functions, including the production of tumor necrosis factor alpha (TNF-␣), a cytokine of great importance for antiMAC resistance in ex vivo and in vivo models (3, 4, 13). It is generally accepted that the final outcome of TNF-␣ expression in different infection models may depend on its site of action, its local concentration, and the duration of exposure. An excess of TNF-␣ released into the blood can be detrimental for humans, as suggested by the observation that many symptoms of tuberculosis and chronic MAC infections related to TNF-␣, such as fever and weight loss, are improved by thalidomide, a drug that selectively destabilizes TNF-␣ mRNA (15, 28). TNF-␣ is involved in the development of granulomas, since administration of neutralizing anti-TNF-␣ antibodies leads to granuloma regression and mycobacterial dissemination (3, 17, 19). At the cellular level, TNF-␣ activity is downregulated by MAC in cultured human and mouse M␾ within the first 24 to 48 h of infection, thus preventing a local antimicrobial effect of this cytokine (9, 11, 14). The wide range of TNF-␣ activities can be in part explained by the presence on almost all nucleated cells of one or two distinct TNF-␣ receptors (TNF-Rs), namely, TNF-RI (p55) and TNF-RII (p75). Shedding of soluble forms from both receptors (sTNF-RI and sTNF-RII, respectively) can modulate the biological effects of TNF-␣ by inhibition of its bioactivity

(9, 29) or by stabilization of its quaternary structure (2). Most of the information on the role of TNF-␣ and TNF-Rs in murine mycobacterial infections has been obtained by using neutralizing antibodies (3, 12) and sTNF-Rs (1, 6, 25). Although these studies have provided evidence that TNF-␣ and TNF-RI play a role in mycobacterial infections in ex vivo and in vivo models, they gave little or no information on the site and temporal effects of TNF-␣ and TNF-R activation in MACinfected mice. An attempt to demonstrate a role for TNF-RI and TNF-RII in the control of MAC infection, based on double-TNF-RII-knockout mice, did not support a role for TNF-Rs in modulating the bacterial load but, rather, pointed to an important role in promoting chronic pathologic changes (8). To contribute to the dissection of these events and to further investigate the roles of TNF-Rs, we monitored the production of TNF-Rs in the spleen and blood of BALB/c mice throughout a 70-day period of infection. In addition, we used an antagonist anti-TNF-RII antibody to assess whether membrane TNF-RII or sTNF-RII is critically involved in the control of infection with MAC. MATERIALS AND METHODS Mice. Male BALB/c mice aged 6 to 7 weeks were obtained from Charles River (Calco, Como, Italy). They were bred and maintained under standard conditions, receiving sterilized chow and acidified water ad libitum. Organism and mouse infection. A clinical isolate of MAC 485 (11) was used throughout this study. Transparent colonies grown on Middlebrook 7H10 agar plates (Difco Laboratories, Detroit, Mich.) were suspended in phosphate-buffered saline (PBS) and sonicated for 10 s to disperse clumps. A suspension adjusted to an optical density of 0.2 at 500 nm (corresponding to approximately 6 ⫻ 108 CFU/ml) was prepared. Each mouse was intraperitoneally (i.p.) inoculated with 107 CFU in 0.2 ml of PBS. At different time points, mice were sacrificed and their spleens were aseptically collected and suspended in Middlebrook 7H9 medium (Difco), ground in homogenizers, and briefly sonicated. Each suspension was 10-fold diluted with distilled water and plated onto 7H10 agar medium; after incubation for 10 to 14 days at 37°C in a humidified 5% CO2

* Corresponding author. Mailing address: Laboratory of Bacteriology and Medical Mycology, Istituto Superiore di Sanita`, Viale Regina Elena 299, 00161 Rome, Italy. Phone: 39 06 49902333. Fax: 39 06 49387112. E-mail: [email protected]. 5762

VOL. 67, 1999

UPREGULATION OF TNF-RII IN M. AVIUM-INFECTED MICE

atmosphere, the colonies were counted. Blood was drawn from the retro-orbital sinus, diluted with distilled water, and analyzed for bacterial load as above. Serum samples were stored in aliquots at ⫺80°C for determination of TNF-␣ and sTNF-R levels. Measurement of TNF-␣ and TNF-R mRNA expression. MAC-infected and PBS-injected (control) mice were sacrificed at the indicated times; their spleens were removed and immediately frozen in liquid nitrogen. (i) RNA extraction. Frozen tissues were ground in a mortar, and guanidine thiocyanate solution (4 M guanidinium thiocyanate, 0.025 M sodium citrate [pH 7], 0.5% Sarkosyl) was added. The homogenates were mixed sequentially with sodium acetate, water-saturated phenol, and chloroform-isoamyl alcohol (49:1), vortexed thoroughly, and incubated on ice for 15 min. After centrifugation at 12,900 ⫻ g for 20 min at 4°C, the RNA in the upper (aqueous) phase was precipitated with isopropanol. Each sample was then incubated for 30 min at ⫺20°C and further centrifuged for 10 min. The RNA pellet was resuspended in cold 75% ethanol, centrifuged, air dried, dissolved in distilled water, and stored frozen (7). RNA gel electrophoresis was performed to confirm that the RNA was intact and that the concentration had been correctly determined. (ii) Reverse transcription (RT) of RNA. Eppendorf tubes containing 40 ␮l of total RNA (0.1 ␮g/␮l) and 4 ␮l of oligo(dT) (27-7858; Pharmacia Biotech, Uppsala, Sweden) were incubated for 10 min at 65°C and chilled on ice (7). Each tube was incubated for 60 min at 37°C after addition of 36 ␮l of a solution containing 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, a 0.5 mM concentration of each deoxynucleoside triphosphate (Pharmacia), 10 mM dithiothreitol, and 10 U of Moloney murine leukemia virus reverse transcriptase (Life Technologies, Grand Island, N.Y.) per ␮l. The reaction mixture was heated at 95°C for 5 min to denature the reverse transcriptase and cooled on ice. cDNA was diluted 1:10 in distilled water and stored frozen at ⫺20°C until use. (iii) Semiquantitative PCR. Synthesized cDNAs were amplified by PCR with primers for murine TNF-␣ and TNF-Rs designed to span introns and to enable differential detection of cDNA from genomic DNA. The following sequences were used: 5⬘-GATCTCAAAGACAACCAACTAGTG-3⬘ (5⬘ primer for TNF-␣) (33), 5⬘-CTCCAGCTGGAAGACTCCTCCCAG-3⬘ (3⬘ primer for TNF-␣) (33), 5⬘-CTGGCCTGCTGCTGTCACTG-3⬘ (5⬘ primer for TNF-RI), 5⬘-GTCAGCT TGGCAAGGAGAGATC-3⬘ (3⬘ primer for TNF-RI), 5⬘-GTCGCGCTGGTCT TCGAACTG-3⬘ (5⬘ primer for TNF-RII), and 5⬘-GGTATACATGCTTGCCT CACAGTC-3⬘ (3⬘ primer for TNF-RII). Briefly, PCRs were carried out with 20-␮l final volumes containing 50 mM KCl, 10 mM Tris-HCl, 0.1% gelatin, 1.5 mM MgCl2, a 0.2 mM concentration of each deoxynucleoside triphosphate, a 0.5 ␮M concentration of each primer, 0.5 U of Taq polymerase (Perkin-Elmer Cetus, Foster City, Calif.), and 2 ␮l of cDNA and incubated in a DNA thermal cycler (Perkin-Elmer Cetus). After an initial incubation at 94°C for 5 min, temperature cycling was initiated with each cycle as follows: 94°C for 1 min, the appropriate annealing temperature for 1 min, and 72°C for 2 min. The PCR products and the ␾X174 DNA HaeIII fragments (Life Technologies), used as the molecular weight markers, were analyzed by agarose gel electrophoresis (1% agarose), stained with ethidium bromide, and photographed under UV light. The intensity of the bands on the negatives was quantitated by scanning the gels on a laser densitometer. To standardize the PCR procedure, preliminary experiments were performed to establish optimal annealing temperatures and cycle numbers. To verify that comparable amounts of cDNAs were added in each PCR within an experiment, samples were normalized to contain equal amounts of cDNA of the housekeeping ␤-actin gene. The amount of each PCR product was determined by comparing the signal density with that of a standard curve generated by simultaneous PCR on the cDNAs of spleen cells stimulated overnight with 2 ␮g of concanavalin A per ml. The results are expressed as values relative to the means of the PBS controls (day 1), which were arbitrarily given a value of 1. Production of recombinant sTNF-RI and sTNF-RII. The cDNAs coding for the extracellular domains of TNF-RI and TNF-RII were amplified by PCR on plasmids containing the entire cDNA sequences of murine receptors (kindly provided by R. Goodwin, Immunex, Seattle, Wash.), using the following primers: 5⬘-CATGGAAACATACATCCATCAGGGGTCACTGGA-3⬘ (5⬘ primer for sTNF-RI); 5⬘-ATGGGATCCTTACGCAGTACCTGAGTCCTGGGG-3⬘ (3⬘ primer for sTNF-RI); 5⬘-CATGGATCCGTGCCCGCCCAGGTTGTC-3⬘ (5⬘ primer for sTNF-RII); and 5⬘-ATGGGATCCTTACCTGGTACTTTGTTCAA TAATGGG-3⬘ (3⬘ primer for sTNF-RII). sTNF-RI and sTNF-RII were produced by cloning the fragments in the pET-11a system (Novagen, Madison, Wis.) and expressing them in the BL21(DE3) Escherichia coli strain (Novagen). Both receptors were expressed as insoluble forms (40 mg of sTNF-RI and 80 mg of sTNF-RII per liter of culture). Solubilization, denaturation, and partial refolding were performed essentially as described for human sTNF-␣ receptors (20). Soluble receptors were purified by reverse-phase chromatography with a Source 15 RPC column (Pharmacia), with a final yield of 5 to 10%. Recombinant sTNF-RI and sTNF-RII were quantified by commercial enzyme-linked immunosorbent assay (ELISA) kits (HyCult Biotechnologies). Sodium dodecyl sulfatepolyacrylamide gel electrophoresis revealed bands corresponding to the expected molecular mass for monomeric sTNF-RI and sTNF-RII (24 and 29 kDa, respectively). Production of polyclonal and monoclonal anti-TNF-R antibodies. sTNF-RI and sTNF-RII were used to raise polyclonal anti-sTNF-RI and anti-sTNF-RII in rabbits by standard techniques; the immunoglobulin G (IgG) fraction of each

5763

serum sample (named RAM-RI and RAM-RII, respectively) was purified by affinity chromatography on protein A-Sepharose (Pharmacia). Monoclonal antibody (MAb) 6G1 (rat anti-murine TNF-RII IgG2a) was prepared as follows. One rat (Wistar) was immunized by injecting 20 ␮g of sTNF-RII in 250 ␮l of a 1:1 emulsion with complete Freund’s adjuvant into the base of the tail and into the hind footpad. The animal was boosted with the same amount of sTNF-RII in incomplete Freund’s adjuvant and four times with sTNF-RII in PBS, subcutaneously, at the base of the tail, all at 15-day intervals. Three days after the last boost, the rat was killed. The spleen cells were isolated and fused with P3-X63 Ag8-NS1 myeloma cells by standard procedures to generate hybridomas. Hybridomas secreting anti-TNF-RII antibodies were screened by ELISA, using microplates coated with sTNF-RII. MAb 6G1 was purified from cell supernatants by ammonium sulfate precipitation, ion-exchange chromatography (three cycles), and ultrafiltration through a YM 10 membrane (Amicon, Danvers, Mass.). The final product was ⬎90% pure by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. MAb 19E12 (IgG2a), used as irrelevant antibody in in vivo studies, was purified by ammonium sulfate precipitation and affinity chromatography on protein A-Sepharose (21). This antibody is directed to Thy.1.1 and does not cross-react with the Thy.1.2 antigen expressed by BALB/c mice. Thymocyte proliferation assay. Thymocytes from C57BL/6 mice (Charles River) were obtained by thymus excision and teasing. The thymocytes were washed twice with RPMI 1640 medium and plated in 96-well flat-bottom plates (Costar, Cambridge, Mass.) (106 cells/well) in a final volume of 100 ␮l of RPMI 1640 (Life Technologies) supplemented with 20% heat-inactivated (56°C for 30 min) fetal calf serum, 1% glutamine, 100 U of penicillin per ml, and 100 ␮g of streptomycin per ml in the presence of 50 U of interleukin-2 per ml. After 48 h, the cells were pulsed with 1 ␮Ci of [3H]thymidine (Amersham, Milan, Italy) per well for 20 h, harvested onto a fiberglass filter, and lysed with distilled water. The radioactivity bound to the filter was measured with a liquid scintillation counter. Measurement of TNF-␣ bioactivity and sTNF-Rs. (i) Determination of TNF-␣ bioactivity levels by a cytolytic assay. TNF-␣ bioactivity was measured by a cytolytic assay based on mouse WEHI 164 clone 13 cells (10, 21); the detection limit of this assay was 0.21 pg/ml. (ii) Determination of sTNF-RI and sTNF-RII levels by ELISA. sTNF-RI and sTNF-RII were quantified by sandwich ELISA with rabbit anti-sTNF-RI or anti-sTNF-RII polyclonal IgG, prepared as described above, in the capturing step and biotinylated anti-sTNF-RI or anti-sTNF-RII MAbs (HyCult Biotechnologies, Uden, The Netherlands) as the second antibody. Streptavidin-peroxidase dissolved in PBS containing 0.5% bovine serum albumin (BSA) and 1% normal goat serum was used as the detection reagent (24). The detection limits of the assays were 0.32 ng/ml for sTNF-RI and 0.1 ng/ml for sTNF-RII. Preparation and stimulation of splenocytes from mice infected once or twice with MAC. Mice were infected intravenously by injecting 5 ⫻ 103 CFU of MAC into the tail vein. After 60 days, some mice were sacrificed and their spleens were aseptically collected and processed for CFU counts as described above. Infected mice were reinfected i.p. with 105 CFU of MAC on day 60; uninfected mice of the same age were infected in parallel and used as once-infected control mice. All the mice were sacrificed after 3 months and processed for determination of CFU, bioactive TNF-␣, sTNF-RI, and sTNF-RII in spleen cell supernatants. Spleen cells collected from individual mice of each group were incubated (in 24-well plates at 3 ⫻ 106 cells/well) in culture medium (RPMI 1640 medium containing 10% heat-inactivated fetal calf serum, 25 mM HEPES, and 2 mM glutamine) with or without 10 ␮g of MAC 485 sonicate per ml. After 48 h, the levels of bioactive TNF-␣, sTNF-RI, and sTNF-RII in the cell supernatants were measured as described above. The sonicate was prepared by lysing MAC cells at 4°C with an ultrasonic disintegrator (VCX 400 W; Sonics & Materials Inc., Danbury, Conn.) for 20 min (1-min cycles with 1-min cooling intervals); after centrifugation, the protein content in the supernatant was determined by the method of Bradford (5).

RESULTS Kinetics of MAC infection in mice. MAC-infected BALB/c mice developed a chronic infection which persisted for more than 70 days (Fig. 1). In the spleen, after a 1.3 log10 unit increase in the number of CFU from day 2 to day 21, the bacterial burden remained fairly constant throughout the remaining period. About 102 CFU/ml was found in the blood between days 2 and 14, and then mycobacteremia dropped to undetectable levels. No animal deaths were recorded during the whole experimental period. Kinetics of TNF-␣ and TNF-R production. The production of TNF-␣, TNF-RI, and TNF-RII mRNAs in the spleens of infected mice was monitored by reverse transcription-PCR throughout the 70-day experimentation period (Fig. 2). The TNF-␣ mRNA level increased fourfold (P ⫽ 0.002) on day 21 in comparison with the level in control mice and then slowly decreased (Fig. 2A). The level of TNF-RI mRNA did not

5764

CORTI ET AL.

INFECT. IMMUN.

FIG. 1. Time course of MAC infection in the spleens and blood of BALB/c mice inoculated i.p. with 107 CFU. Data are shown as CFU per spleen and CFU per milliliter of blood. Each point represents the mean value for five mice and the standard deviation.

significantly increase at any time (Fig. 2B); in contrast, TNFRII mRNA was significantly overexpressed in the spleen on day 21 (15-fold increase [P ⫽ 0.009]) (Fig. 2C). Of note, the increased production of TNF-␣ mRNA was not accompanied by an increase in the level of circulating TNF-␣, since no cytotoxic activity was detected in the blood of infected animals even when a highly sensitive bioassay was used (detection limit, 0.21 ␮g/ml). However, the production of TNFRII mRNA in the spleen was concomitant with a significant increase in the level of sTNF-RII antigen in the blood (Fig. 3). Production and characterization of an antagonist anti-TNFRII MAb. To investigate the role of TNF-RII in MAC infection, we have generated a new rat anti-murine TNF-RII MAb (MAb 6G1). Polyclonal anti-TNF-RI and anti-TNF-RII IgGs (termed RAM-RI and RAM-RII, respectively) were also prepared. No cross-reactivity of polyclonal anti-TNF-R antibodies or MAbs with irrelevant sTNF-RI or sTNF-RII was observed in the ELISA (results not shown). The capability of MAb 6G1 and polyclonal antibodies to mimic (agonist activity) or inhibit (antagonist activity) TNF-␣induced effects was investigated by a thymocyte proliferation assay (Fig. 4, left). Since thymocyte proliferation can be triggered by selective stimulation of TNF-RII (27), this assay is a good model for investigating the TNF-RII agonist or antagonist properties of each antibody. Polyclonal RAM-RII alone, but not RAM-RI, induced thymocyte proliferation, indicating that RAM-RII behaves as an agonist anti-TNF-RII antibody. In contrast, MAb 6G1 alone was unable to induce proliferative effects, whereas it inhibited the TNF-␣-induced proliferation of thymocytes. Thus, MAb 6G1 behaves as an antagonist anti-TNF-RII antibody. Moreover, MAb 6G1 inhibited the binding of murine TNF-␣ to sTNF-RII in an ELISA system (Fig. 4, right), indicating that this antibody was directed to an epitope sterically overlapping the TNF-␣-binding site on the receptor. Altogether, these results indicate that MAb 6G1 can prevent the binding of TNF-␣ to both sTNF-RII and membrane-bound TNF-RII.

FIG. 2. Semiquantitative analysis of TNF-␣, TNF-RI, and sTNF-RII gene expression by RT-PCR in the spleens of MAC-infected and PBS-treated BALB/c mice. Data are presented as arbitrary units; each point represents the mean value for three mice and the standard deviation. At each time point, a statistically significant increase in expression compared to PBS-injected mice is labelled with an asterisk, indicating P ⬍ 0.05 (Student’s t test).

Effect of anti-TNF-RII antibodies on MAC growth in mice. To assess whether TNF-RII and/or its soluble form is involved in the regulation of TNF-␣ activity during MAC infection, we carried out an experiment in which MAb 6G1 was administered to animals. This treatment significantly inhibited MAC growth in the spleens of mice after both 21 days (P ⫽ 0.0002)

FIG. 3. Kinetics of release of sTNF-RII in the blood of MAC-infected and PBS-treated BALB/c mice. Each point represents the mean value and the standard deviation for three mice. At each time point, a statistically significant increase compared to PBS-treated mice is labelled with an asterisk indicating P ⬍ 0.05.

VOL. 67, 1999


5765

FIG. 4. Effect of MAb 6G1 on the TNF-␣-induced proliferation of mouse thymocytes (left) and on the binding of TNF-␣ to sTNF-RII (right). The thymocyte proliferation assay was carried out in the absence and in the presence of murine TNF-␣, without or with 2 ␮g of the indicated antibody per ml (left). The effect of MAb 6G1 on the binding of 3 ␮g of TNF-␣ per ml to sTNF-RII (right) was evaluated by a competitive ELISA with microplates coated with sTNF-RII (10 ␮g/ml in PBS, 1 h at room temperature) and blocked with 3% BSA for 2 h. After a 1-h incubation with biotin–TNF-␣ in PBS containing 1% BSA and MAb 6G1 at various doses, the plates were washed with PBS and further incubated with the detecting reagents (anti-TNF-␣ rabbit IgGs, followed by goat anti-rabbit IgG–peroxidase conjugate and chromogenic substrate) by standard techniques.

and 42 days (P ⫽ 0.0005) of infection (Fig. 5). Treatment with an irrelevant antibody (MAb 19E12) of the same isotype (IgG2a) did not cause significant changes in the number of CFU in comparison with the control (saline-treated) animals. This observation indicates that TNF-RII and/or its soluble form is critically involved in the control of MAC infection in mice. Production of TNF-␣ and sTNF-Rs in mice infected once or twice with MAC. After 60 days from the initial infection, the number of CFU per spleen in untreated mice was (2.6 ⫾ 0.5) ⫻ 106. To explore the response to a further mycobacterial stimulation, infected mice were reinfected with MAC on day 60. After 3 months, CFU, TNF-␣, and sTNF-R levels were deter-

mined and compared with those for the once-infected control mice. No significant difference in viable counts between twiceand once-infected mice was observed; the CFU were (0.71 ⫾ 0.46) ⫻ 106 and (1 ⫾ 0.46) ⫻ 106, respectively. Conversely, the first infection induced a stronger response to the second in terms of bioactive TNF-␣ production, as suggested by the observation that spleen cells of mice infected twice secreted at least 11 times as much cytokine as did those infected once (Fig. 6A). No detectable secretion of sTNF-RI was observed in supernatants from both groups of mice (Fig. 6B). In contrast, upregulation of TNF-␣ secretion was associated with a strong increase in sTNF-RII shedding from twiceinfected mouse splenocytes in comparison with control mouse cells (Fig. 6C). Stimulation of the spleen cells of once- and twice-infected mice with MAC antigens caused a 15-fold and 2-fold increase of bioactive TNF-␣ production, respectively, in comparison with unstimulated cells (Fig. 6A). No increase in the level of sTNF-RI above the assay detection limit was induced by MAC antigens in spleen cell supernatants of both groups of mice (Fig. 6B), whereas a consistent upregulation of sTNF-RII shedding was observed in the supernatants of once- and twiceinfected mice in comparison with unstimulated cells (Fig. 6C). DISCUSSION

FIG. 5. Effect of anti-TNF-RII treatment on the time course of MAC growth in the spleens of BALB/c mice infected with 107 CFU i.p. The mice were treated i.p. with 100 ␮g of anti-TNF-RII antibody (MAb 6G1) or an irrelevant antibody (MAb 19E12) of the same isotype (IgG2a) as MAb 6G1, administered 1 day before infection and every 5 days up to day 34. Control (infected) mice were treated with saline. Each point represents the mean value and the standard deviation of two independent experiments, each with four mice per group. At each time point, a statistically significant decrease compared to control mice is labelled with two asterisks, indicating P ⬍ 0.005.

Anti-MAC resistance in BALB/c mice developed in two phases: the first, up to day 21, was associated with bacterial multiplication in the spleen, a drop in mycobacteremia, and transient activation of the TNF-␣/TNF-RII system; the second was associated with containment of infection in the spleen and blood. The marked increase of TNF-␣ and TNF-RII mRNA levels on day 21 coincided with or immediately preceded the highest protective activity against infection and indicated that the control of MAC growth in the spleen closely correlates with the local activation of the TNF-␣/TNF-RII system. This is in keeping with results of previous studies (3) in which early neutralization of TNF-␣ with antibodies led to an enhanced bacterial load in tissues. TNF-RII appeared to be regulated not only at the transcriptional level but also posttranscriptionally, as

5766

CORTI ET AL.

FIG. 6. Release of bioactive TNF-␣ (A), sTNF-RI (B), and sTNF-RII (C) in the supernatants of spleen cells of once- or twice-infected mice, either unstimulated (solid bars) or stimulated with 10 ␮g of MAC sonicate per ml (open bars). Each point represents the mean value and the standard deviation for six mice. Detection limits are indicated by dotted lines.

shown by a concomitant increase in the level of sTNF-RII in the blood. The correlation between activation of TNF-RII in the spleen and blood and the capability of animals to efficiently control MAC growth strongly suggests that this receptor and its soluble form play a role in modulating TNF-␣ activity locally and systemically during the immune response to MAC. Interestingly, in contrast to TNF-RII, transcription of TNF-RI mRNA in the spleen did not change appreciably during infection. Previous reports showed that TNF-RI is essential for protection against Mycobacterium tuberculosis (12) and My-

INFECT. IMMUN.

cobacterium bovis BCG (25) infections in mice. Although TNF-RI may also be involved in signalling in a constitutive mode, our data suggest that the fine modulation of TNF-␣ activity in MAC infection during the early phase is mainly due to TNF-RII. It has been shown that TNF-RII can play an accessory role by increasing the local concentration of TNF-␣ and by rapid “ligand passing” to TNF-RI for signalling (26), while TNF-RI can mediate most of the effects of TNF-␣ (29, 31). Moreover, TNF-RII can also enhance TNF-RI-mediated activities by cooperative signalling (32). Besides these supportive or modulating effects, TNF-RII can also directly contribute to local restriction of the responses induced by the membranebound precursor of TNF-␣ (16) and to the mediation of other TNF-␣ activities, such as thymocyte and T-cell proliferation and granulocyte-macrophage colony-stimulating factor production (27, 30). Thus, the modulation of TNF-RII during MAC infection not only may regulate the signals triggered by soluble TNF-␣ through TNF-RI but also could markedly affect the overall activity of soluble and membrane-bound TNF-␣. Shedding of sTNF-RII can additionally contribute to the downregulation of the receptors and to the release of inhibitors that locally or systemically impair interaction with membrane receptors. Recent ex vivo studies have shown that the supernatant of MAC-infected M␾ contains a significant amount of TNF-␣ antigen, mostly inactive (9); shedding of soluble TNF-RII has also been suggested as one of the mechanisms of TNF-␣ inhibition in cell culture supernatants. These studies suggest that upregulation of soluble TNF-␣ receptors may play a role in the ability of some virulent MAC strains to survive in M␾ (9). Our findings that shedding of TNF-RII indeed occurs in vivo during infection support the hypothesis of a functional importance of TNF-RII and its soluble form in the regulation of TNF-␣ activity during MAC infection. This hypothesis is strengthened by the finding that the antagonist anti-TNF-RII MAb 6G1 can affect bacterial growth in the spleens of infected mice. Interestingly, this antibody was able to significantly reduce bacterial growth after 21 or 42 days of infection, suggesting that the antigen accessible to this antibody acts as a negative modulator of the antibacterial activity exerted by TNF-␣ in the early phase of infection (3). Circulating sTNF-RII is probably accessible to systemically administered antibodies. Since MAb 6G1 can prevent the binding of TNF-␣ to sTNF-RII, one possibility is that the MAb 6G1 blocked the inhibitory activity of this soluble receptor, leaving TNF-␣ free to interact with membrane receptors (at least TNF-RI) and to trigger antibacterial effects. However, since MAb 6G1 can also prevent the binding of TNF-␣ to membrane-bound TNF-RII, the inhibition of MAC growth may be related to inhibition of TNF-␣ binding to both forms of this receptor. Although it is difficult to draw conclusions on this point, these data suggest that the TNF-RII system is indeed functionally involved in the modulation of TNF-␣ during infection. The above results indicate that upon a single challenge with MAC the TNF-␣/TNF-R system is transiently activated during the early phase of infection. Since mycobacterial lesions are acutely sensitive to further stimulation with mycobacterial products or whole organisms (22, 23), we explored whether restimulation of infected mice with live MAC could induce a chronic activation of the TNF-␣/TNF-R system in the spleens of twice-infected or once-infected mice. Although CFU numbers in twice-infected and once-infected mice were not significantly different 3 months after reinfection, the TNF-␣/TNFRII system was markedly activated in the spleens of twiceinfected mice, as suggested by a consistent upregulation of the release of bioactive TNF-␣ and sTNF-RII in spleen cell super-


VOL. 67, 1999

natants. These data suggest that under conditions in which reinfection was efficiently controlled, a stronger chronic activation of the TNF-␣/TNF-RII system than the one observed in mice infected with a single MAC dose occurred. The observation that MAC antigens can induce a vigorous increase (⬎15fold) in the production of bioactive TNF-␣ by in vitro stimulation of spleen cells from once-infected mice and only a 2-fold increase in spleen cells from twice-infected mice can be explained in part by an increased production of sTNF-RII. Overall, our data indicate that TNF-RII can be activated to modulate TNF-␣ activity in infected mice either transiently, in the first month of infection, or chronically, in reinfected mice. The time-related modulation of TNF-␣ activity by TNF-RII may be an important mechanism by which immunocompetent hosts efficiently control natural infections by this environmental organism. ACKNOWLEDGMENTS We thank Antonio Cassone, Istituto Superiore di Sanita`, Rome, Italy, for help in reviewing the manuscript. We thank Elisabetta Iona and Yuming Fan, Istituto Superiore di Sanita`, and Barbara Colombo, DIBIT San Raffaele Scientific Institute, for valuable technical assistance. This work was supported in part by the Italian Tuberculosis Project, Istituto Superiore di Sanita`, Ministero della Sanita` (grant 28), and by the Italian AIDS Project, Istituto Superiore di Sanita` (grant 50 B/E). REFERENCES 1. Adams, L. B., C. M. Mason, J. K. Kolls, D. Scollard, J. L. Krahenbuhl, and S. Nelson. 1995. Exacerbation of acute and chronic murine tuberculosis by administration of a tumor necrosis factor alpha receptor-expressing adenovirus. J. Infect. Dis. 171:400–405. 2. Aderka, D., H. Engelmann, Y. Maor, C. Brakebusch, and D. Wallach. 1992. Stabilization of the bioactivity of tumor necrosis factor by its soluble receptors. J. Exp. Med. 175:323–329. 3. Appelberg, R., A. G. Castro, J. Pedrosa, R. A. Silva, I. M. Orme, and P. Minoprio. 1994. Role of gamma interferon and tumor necrosis factor alpha during T-cell-independent and -dependent phases of Mycobacterium avium infection. Infect. Immun. 62:3962–3971. 4. Bermudez, L. E. M., P. Stevens, P. Kolonosky, M. Wu, and L. S. Young. 1989. Treatment of experimental disseminated Mycobacterium avium complex infection in mice with recombinant IL-2 and tumor necrosis factor. J. Immunol. 143:2996–3000. 5. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–254. 6. Champsi, J., L. S. Young, and L. E. M. Bermudez. 1995. Production of TNF-␣, IL-6 and TGF-␤, and expression of receptors for TNF-␣ and IL-6, during murine Mycobacterium avium infection. Immunology 84:549–554. 7. Coligan, J. E., A. M. Kruisbeek, D. H. Margulies, E. M. Shevach, and W. Strober. 1992. Molecular biology, p. 10.11.5–10.11.14. In R. Coico (ed.), Current protocols in immunology, unit 10, vol. 2. John Wiley & Sons, Inc., New York, N.Y. 8. Doherty, T. M., and A. Sher. 1997. Defects in cell-mediated activity affect chronic, but not innate, resistance to Mycobacterium avium infection. J. Immunol. 158:4822–4831. 9. Eriks, I. S., and C. L. Emerson. 1997. Temporal effect of tumor necrosis factor alpha on murine macrophages infected with Mycobacterium avium. Infect. Immun. 65:2100–2106. 10. 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. 11. Fattorini, L., Y. Xiao, C. M. Ausiello, F. Urbani, A. LaSala, M. Mattei, and G. Orefici. 1996. Late acquisition of hyporesponsiveness to Mycobacterium avium-infected human macrophages in producing tumor necrosis factor-␣ but not interleukin-1␤ and -6. J. Infect. Dis. 173:1030–1033. 12. Flynn, J. L., M. M. Goldstein, J. Chan, K. J. Triebold, K. Pfeffer, C. J. Lowenstein, R. Schreiber, T. W. Mak, and B. R. Bloom. 1995. Tumor necrosis factor-␣ is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2:561–572.

Editor: J. R. McGhee

5767

13. Furney, S. K., P. S. Skinner, A. D. Roberts, R. Appelberg, and I. M. Orme. 1992. Capacity of Mycobacterium avium isolates to grow well or poorly in murine macrophages resides in their ability to induce secretion of tumor necrosis factor. Infect. Immun. 60:4410–4413. 14. Gan, H., G. W. Newman, and H. G. Remold. 1995. Human macrophages acquire a hyporesponsive state of tumor necrosis factor alpha production in response to successive Mycobacterium avium serovar 4 stimulation. Infect. Immun. 63:1921–1926. 15. Gori, A., F. Franzetti, G. Marchetti, C. Rossi, M. L. Fusi, S. Ruzzante, and M. Clerici. 1998. Clinical and immunological improvement in a patient who received thalidomide treatment for refractory Mycobacterium avium complex infection. Clin. Infect. Dis. 26:184–185. 16. Grell, M., E. Douni, H. Wayant, M. Lohden, M. Clauss, B. Maxeiner, S. Georgopoulos, W. Lesslauer, G. Kollias, K. Pfizenmaier, and P. Scheurich. 1995. The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 Kda tumor necrosis factor receptor. Cell 83:793–802. 17. Hansch, H. C. R., D. A. Smith, M. E. A. Mielke, H. Hahn, G. J. Bancroft, and S. Ehlers. 1996. Mechanisms of granuloma formation in murine Mycobacterium avium infection: the contribution of CD4⫹ T cells. Int. Immunol. 8:1299–1310. 18. Inderlied, C. B., C. A. Kemper, and L. E. M. Bermudez. 1993. The Mycobacterium avium complex. Clin. Microbiol. Rev. 6:266–310. 19. Kindler, V., A.-P. Sappino, G. E. Grau, P.-E. Piguet, and P. Vassalli. 1989. The inducing role of tumor necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 56:731–740. 20. Merli, S., A. Corti, and G. Cassani. 1995. Production of sTNF-R1 in Escherichia coli: optimization of the refolding yields by a microtiter dilution assay. Anal. Biochem. 230:85–91. 21. Moro, M., M. Pelagi, G. Fulci, G. Paganelli, P. Dellabona, G. Casorati, A. G. Siccardi, and A. Corti. 1997. Tumor cell targeting with antibody-avidin complexes and biotinylated tumor necrosis factor-␣. Cancer Res. 57:1922– 1928. 22. Nagao, S., and A. Tanaka. 1985. Necrotic inflammatory reaction induced by muramyl dipeptide in guinea pigs sensitized by tubercle bacilli. J. Exp. Med. 162:401–403. 23. Rook, G. A. W., and R. Hernandez-Pando. 1996. The pathogenesis of tuberculosis. Annu. Rev. Microbiol. 50:259–284. 24. Schmid, J. A., and A. Billich. 1997. Simple method for high sensitivity chemiluminescence ELISA using conventional laboratory enrichment. BioTechniques 22:278–280. 25. Senaldi, G., Y. Songmei, C. L. Shaklee, P. F. Piguet, T. W. Mak, and T. R. Ulich. 1996. Corynebacterium parvum- and Mycobacterium bovis bacillus Calmette-Guerin-induced granuloma formation is inhibited in TNF receptor I (TNF-RI) knockout mice and by treatment with soluble TNF-RI. J. Immunol. 157:5022–5026. 26. Tartaglia, L. A., D. Pennica, and D. V. Goeddel. 1993. Ligand passing: the 75-kDa tumor necrosis factor (TNF) receptor recruits TNF for signalling by the 55-kDa TNF receptor. J. Biol. Chem. 268:18542–18548. 27. Tartaglia, L. A., D. V. Goeddel, C. Reynolds, I. S. Figari, R. F. Weber, B. M. Fendly, and M. A. Palladino, Jr. 1993. Stimulation of human T-cell proliferation by specific activation of the 75-KDa tumor necrosis factor receptor. J. Immunol. 151:4637–4641. 28. Tramontana, J. M., U. Utaipat, A. Molloy, P. Akarasewi, M. Burroughs, S. Masonkawkeyoon, B. Johnson, J. D. Klausner, W. Rim, and G. Kaplan. 1995. Thalidomide treatment reduces tumor necrosis factor ␣ production and enhances weight gain in patients with pulmonary tuberculosis. Mol. Med. 1:384–397. 29. Vandenabeele, P., W. Declercq, R. Beyaert, and W. Fiers. 1995. Two tumor necrosis factor receptors: structure and function. Trends Cell Biol. 5:392– 399. 30. Vandenabeele, P., W. Declercq, D. Vercammen, M. Van de Craen, J. Grooten, H. Loetscher, M. Brockhaus, W. Lesslauer, and W. Fiers. 1992. Functional characterization of the human tumor necrosis factor receptor p75 in a transfected rat/mouse T cell hybridoma. J. Exp. Med. 176:1015–1024. 31. Van Ostade, X., P. Vandenabeele, J. Tavernier, and W. Fiers. 1994. Human tumor necrosis factor mutants with preferential binding to and activity on the p55 receptor- and p75 receptor-selective TNF-alpha mutants. EMBO J. 13:843–850. 32. Weiss, T., M. Grell, B. Hessabi, S. Bourteele, G. Muller, P. Scheurich, and H. Wajant. 1997. Enhancement of TNF-receptor p60-mediated cytotoxicity by TNF receptor p80: requirement of the TNF-receptor associated factor-2 binding site. J. Immunol. 158:2398–2404. 33. Wynn, T. A., E. Eltoum, A. W. Cheever, F. A. Lewis, W. C. Gause, and A. Sher. 1993. Analysis of cytokine mRNA expression during primary granuloma formation induced by eggs of Schistosoma mansonii. J. Immunol. 151: 1430–1440.