Defective Nitric Oxide Effector Functions Lead to Extreme Susceptibility of Trypanosoma cruzi-Infected Mice Deficient in Gamma Interferon Receptor or Inducible Nitric Oxide Synthase

Abstract

Trypanosoma cruzi, the causative agent of Chagas’ disease, induces an innate and adaptive host immune response during the acute phase of infection. These responses were analyzed by comparing mouse lines deficient for the gamma interferon (IFN-γ) receptor (IFN-γR^−/−) or deficient for inducible nitric oxide synthase (iNOS^−/−). Both lines were highly susceptible, with similar and dramatically increased parasite burdens and severe histopathology and were incapable of surviving even very low doses, exhibiting similar mortality kinetics. This pathophysiological correlation has a common cause, since both mutant mouse strains were unable to respond to infection by producing nitric oxide (NO) with the consequence that mutant macrophages had impaired trypanocidal activities. These in vivo and subsequent in vitro studies further demonstrated that an IFN-γ-dependent pathway of iNOS induction is crucial for efficient NO production and mandatory for resisting acute infection with T. cruzi. Despite this defect, both mutant mouse strains had a rather normal proinflammatory cytokine response (interleukin-12 [IL-12], IFN-γ, IL-6), with the exception of an impaired tumor necrosis factor alpha and IL-1α response in IFN-γR^−/− mice, demonstrating that only the latter two cytokines are dependent on IFN-γ activation. Moreover, polarization of T cells in type 1 and type 2 T-helper (Th1/Th2) and cytotoxic T (Tc1/Tc2) cells as well as T. cruzi-specific antibody responses were normal in IFN-γR^−/− mice, demonstrating that IFN-γ is not necessary for the promotion of T-cell differentiation and T. cruzi-specific antibody responses.

Trypanosoma cruzi is an obligate intracellular protozoan parasite of mammals and the etiologic agent of Chagas’ disease. T. cruzi invades a variety of host cell types and replicates within the cytoplasm. In humans and in mice, infection with T. cruzi is followed by a severe immunosuppression mediated by T cells (27) and macrophages (9). Infected mice have a decreased ability to produce interleukin-2 (IL-2) (30) and to display IL-2 receptors (36) during the acute phase of infection. The acute phase is characterized by a large increase in parasite replication which can be monitored in the blood of the infected mice. However, immunocompetent mice are able to control the parasite load by an inflammatory innate and specific immune response but generally fail to completely eliminate the parasite. This may eventually lead to chronic chagasic pathology, in which autoimmune mechanisms also play a role (7).

Multiple components of both the innate and the adaptive immune system are simultaneously required for protection during the acute phase of infection, with gamma interferon (IFN-γ) being an important mediator of resistance to T. cruzi (29, 45). IFN-γ is believed to be produced by natural killer (NK) cells at the onset of infection (8) and later also by CD4⁺ (38) and CD8⁺ (43) T cells. Consequently, administration of recombinant IFN-γ increases resistance (33), whereas neutralization of endogenously produced IFN-γ increases susceptibility during the acute phase of infection (45). Moreover, IFN-γ-activated macrophages are a major source of protective inflammatory cytokines and induce trypanocidal activities (19). The latter can be blocked by l-arginine analogs that inhibit the induced nitric oxide synthase (iNOS) pathway (47). In addition, nitric oxide (NO) is released during the acute phase of T. cruzi infection in mice, and treatment of such mice with inhibitors of NO synthase exacerbates the infection (31, 47). While NO might be by itself cytotoxic, it also reacts with superoxide (O₂⁻) to yield peroxynitrite (ONOO⁻), a stronger cytotoxic molecule than its precursor (4, 32), which causes lipid and thiol oxidation and nitrosylation and nitrosylation of amino acids on target proteins and is highly toxic for T. cruzi (13). In this report we show the immunological consequence of T. cruzi infection in the absence of IFN-γ and iNOS by comparative in vivo studies using IFN-γ receptor (IFN-γR)- and iNOS-deficient (IFN-γR^−/− and iNOS^−/−, respectively) mice. Evidence is presented that both types of mutant mice are defective in NO production and trypanocidal activities, explaining their similar and extreme susceptibilities. These data demonstrate the crucial importance of IFN-γ-dependent, iNOS-mediated NO effector functions to resist acute T. cruzi infection. Despite an impaired tumor necrosis factor alpha (TNF-α) and IL-1α response, other proinflammatory cytokine responses (e.g., IL-12, IFN-γ, IL-6) were rather normal. Moreover, antibody production by B cells and isotype switching to immunoglobulin G2a (IgG2a) as well as T-cell differentiation were also independent of IFN-γ signalling.

MATERIALS AND METHODS

Mice and parasites.

Young adult (7- to 8-week-old) IFN-γR^−/− mice (21), 129sv wild-type mice (IFN-γR^+/+), iNOS^−/− mice, and 129sv × C57BL/6 wild-type mice (iNOS^+/+) (28), maintained under specific-pathogen-free conditions, were used for the experiments. iNOS-deficient mice were generously provided by J. D. MacMicking, C. Nathan (Cornell University Medical College, New York, N.Y.), and J. S. Mudgett (Merck Research Laboratories, Rahway, N.J.).

A cloned population of the reticulotropic Trypanosoma cruzi strain Tulahuen (a kind gift from Simon Croft, London School of Hygiene and Tropical Medicine, London, Great Britain) was routinely maintained in mice. For experiments, groups of mice were intraperitoneally infected with trypomastigotes and the resulting parasitemia was monitored by hemacytometer counting of blood samples.

For preparation of inactivated T. cruzi (iTC), tissue culture trypomastigotes, and trypanocidal assays, monolayers of LLC-MK2 cells (American Type Culture Collection [ATCC] CCL7.1) were infected and cultured in complete ISCOVES medium (Gibco, Paisley, Great Britain) containing 10% heat-inactivated fetal calf serum (Gibco), 0.05 mM 2-mercaptoethanol (Roth, Karlsruhe, Germany), and penicillin and streptomycin (100 U/ml and 100 μg/ml, respectively) (Biochrom, Berlin, Germany). Inactivation of culture trypomastigotes was performed by 10 freeze-thaw cycles, as described previously (10).

Histopathological analyses.

Infected mice were killed by cervical dislocation after 17 days of infection. Tissue specimens were collected and fixed in paraformaldehyde (4% in phosphate-buffered saline) for further processing. Paraffin-embedded tissue sections were stained with hematoxylin-eosin and subjected to microscope analysis.

Trypanocidal assay.

T. cruzi trypomastigotes were harvested from infected LLC-MK2 cells and were incubated overnight before use in the trypanocidal assay (19). Amastigote contamination was <15% for all assays.

Bone marrow cells from IFN-γR^−/−, iNOS^−/−, and wild-type mice were flushed from mouse femora and cultivated at a concentration of 5 × 10⁵ cells per ml in hydrophobic Teflon film bags (Hereaus, Hanau, Germany) as previously described (15). The culture medium contained 70% high-glucose-formulation Dulbecco’s modified Eagle’s Medium (Gibco), supplemented with 2 mM l-glutamine, 0.01 mM sodium pyruvate, 5% heat-inactivated horse serum, 10% heat-inactivated fetal calf serum (Gibco), 0.05 mM 2-mercaptoethanol (Roth), penicillin and streptomycin (100 U/ml and 100 μg/ml, respectively) (Biochrom), and 30% L929 conditioned medium, as a source of macrophage colony-stimulating factor activity. L-cell conditioned medium was prepared as previously described (15). After 10 days of infection, a pure bone marrow macrophage (BMMφ) population developed and cells were used for the assays as previously described.

For infection, macrophages were washed, counted, and adjusted to a concentration of 2.5 × 10⁶ cells/ml; 0.2 ml of BMMφ was added to each well of a Lab Tek eight-chamber slide (Nunc, Naperville, Ill.) and incubated for 2 h. After being washed three times with Hanks balanced salt solution (Gibco), BMMφ were infected with 50 μl of 2 × 10⁷ parasites/ml, with a final ratio of two parasites to one BMMφ, and incubated for 2 h. Extracellular parasites were removed by rinsing gently with Hanks balanced salt solution. Intracellular parasite elimination was determined after a 2-day incubation with complete media, supplemented with recombinant murine IFN-γ (100 or 10 U/ml), TNF-α (50 U/ml; Pharmingen, San Diego, Calif.) (the endotoxin level is <0.1 ng per mg), N^G-monomethyl-l-arginine (l-NMMA; 500 μM) (Calbiochem, San Diego, Calif.), and anti-IFN-γ (500 ng/ml) (R4-6A2; ATCC) or lipopolysaccharide (LPS) (10 μg/ml; Sigma, St. Louis, Mo.). Two hundred fifty cells per duplicate well were counted by light microscopy on Diff Quick (Baxter Scientific, Mundelein, Ill.)-stained slides. The percentage of cells infected with T. cruzi was recorded, and the percentage of parasite elimination was determined by the following calculation: 1 − [% infected cells_(stimulated)/% infected cells_(medium)] × 100%.

CD4⁺ and CD8⁺ T-cell enrichment.

Enrichment of the lymph node cell suspension for CD4⁺ cells was performed by positive selection with magnetic mouse CD4 Dynabeads and mouse CD4 DETACHaBEAD (Dynal; Robbins Scientific, Mountain View, Calif.). CD8⁺ cells were enriched by further incubation of the CD4⁺-depleted cell suspension with anti-B220-specific Dynabeads. Positive selected CD4⁺ cells from lymph nodes contained <5% CD8⁺, and negative enriched CD8⁺ cells contained <5% CD4⁺ cells, as determined by flow cytometry analysis.

Determination of cytokines, T. cruzi-specific antibodies, and nitrite.

Isolated cells were cultured in 48- or 96-well flat-bottom plates (Nunc) at 2 × 10⁶/ml. The cultures were stimulated either with iTC (4 × 10⁶/ml), LPS (5 μg/ml; Sigma), or plate-bound anti-CD3 (30 μg/ml) (145-2C-11; ATCC). Cell supernatants from the cultures were harvested after 24 or 48 h of culture.

Cytokine levels in the plasma and culture supernatants were detected by sandwich enzyme-linked immunosorbent assays as described previously (11). Plasma and culture supernatants and appropriate standards (Pharmingen) were used in threefold serial dilutions. The coating and biotinylated detection antibodies for IFN-γ, TNF-α, IL-4, IL-6, and IL-12 were purchased from Pharmingen. The unlabelled rabbit and hamster anti-IL-1α antibodies were from Genzyme (Cambridge, Mass.), and the biotinylated anti-rabbit antibody was from Southern Biotechnology (Birmingham, Ala.). Alkaline phosphatase coupled to streptavidin (Southern Biotechnology) was used to stain the detection antibodies. The cytokines and their detection levels were as follows: IFN-γ, 0.2 ng/ml; TNF-α, 0.05 ng/ml; IL-1α, 0.1 ng/ml; IL-4, 1 ng/ml; IL-6, 0.3 ng/ml; and IL-12, 0.3 ng/ml.

For antigen-specific antibody detection, plates were coated with 10⁴ inactivated trypomastigotes/well and incubated with the plasma of individual animals bled at day 0 and day 17 after infection, followed by incubation with alkaline phosphatase-conjugated isotype-specific antibodies for Ig, IgM, IgG1, IgG2a, IgG2b, and IgG3 (Southern Biotechnology). All assays were developed with p-nitrophenyl phosphate (Sigma).

The nitrite content in serial diluted triplicates was measured by adding 50 μl of freshly prepared Griess reagent to 50 μl of the samples in 96-well plates and reading the optical density at 550 nm after 10 min and subsequently comparing it with the optical density curves of serial dilutions of sodium nitrite in normal plasma or complete culture medium. For plasma nitrite quantification, the Griess reaction was used as described previously (35).

RESULTS

IFN-γR^−/− and iNOS^−/− mice are unable to survive acute T. cruzi infection with similar parasitemias and mortality kinetics.

IFN-γR^−/− mice on a 129sv background were infected regularly with sublethal doses of 100 trypomastigotes of T. cruzi, and the course of infection was compared with that of similarly infected 129sv wild-type mice (50% lethal dose = 250 trypomastigotes). During the acute phase of infection, mutant mice exhibited an earlier onset and an increased level (15-fold) of parasitemia compared to infected controls (Fig. 1a). Moreover, all mutant mice died between days 14 and 21 (Fig. 1b) whereas most wild-type mice survived acute infection. By reducing the infection dose to 15 or 50 parasites, mortality was delayed. Nevertheless, all mutant mice died within 28 and 33 days postinfection. These results demonstrate the extreme susceptibility of IFN-γR^−/− mice and their inability to survive T. cruzi infection. Similar experiments were performed with iNOS^−/− mice and their wild-type controls (129sv × C57BL/6). When sublethal doses of 15 trypomastigotes were used, all mutant mice showed earlier and fivefold-exacerbated parasitemia (Fig. 1c) and died between day 21 and 26 (Fig. 1d), whereas most wild-type mice survived acute infection. These results demonstrate the extreme susceptibility of iNOS^−/− mice and their inability to survive T. cruzi infection. In conclusion, both IFN-γ and iNOS seem to have key functions which are crucial in surviving the acute phase of T. cruzi infection.

FIG. 1.

Parasitemia and survival of mutant mice infected with T. cruzi. IFN-γR^−/− (open symbols) and wild-type (closed symbols) mice were each infected with 100 blood trypomastigotes, and the subsequent parasitemia (trypomastigotes per microliter) (a) was observed as described in Materials and Methods. Mice were each infected with 100 (○, •), 50 (◊), or 15 (▵) blood trypomastigotes, and survival (b) was monitored. iNOS^−/− (open symbols) and wild-type (closed symbols) mice were each infected with 15 blood trypomastigotes, and parasitemia (trypomastigotes per microliter) (c) and survival (d) were monitored. Results are expressed as the means ± standard deviations (error bars) of five mice/group.

Extensive necrotic lesions and parasite dissemination in infected IFN-γR^−/− and iNOS^−/− mice.

On day 17 postinfection, the liver (Fig. 2a and b), heart, and spleen (not shown) of infected wild-type mouse strains showed comparably low numbers of parasite-infected cells, with inflammatory mononuclear cell infiltration in the liver and spleen but without histopathological lesions in these organs. In contrast, the infected organs of IFN-γR^−/− and iNOS^−/− mice revealed an increased parasite burden accompanied by large necrotic lesions at the sites of infection in the liver (Fig. 2c and d) and spleen (not shown). Furthermore, amastigote nests were found in the cytoplasm of infected cells and disseminated parasites were found in the necrotic areas (Fig. 2e and f). Infected organs of IFN-γR^−/− mice showed more and larger parasite nests than those of iNOS^−/− mice, in accordance with the observed higher parasitemia of the former (Fig. 1). In summary, these results strongly suggest that both mutant mouse strains succumb to the pathophysiology of the infection, which is typical for a severe acute phase of experimental Chagas’ disease. Moreover, the correlation of comparable early parasitemias and mortality kinetics when taken together with the similar pathophysiologies observed in IFN-γR^−/− and iNOS^−/− mice may indicate related immunological mechanisms. These possibilities were further investigated.

FIG. 2.

Histopathology of heart and liver tissues from T. cruzi-infected IFN-γR^−/− and iNOS^−/− mice. IFN-γR^−/− and control mice were each infected with 100 trypomastigotes, iNOS^−/− and control mice were each infected with 50 trypomastigotes, and hematoxylin-eosin-stained sections were prepared at 17 days postinfection. Liver tissue from IFN-γR^+/+ (a) and iNOS^+/+ (b) control mice showed comparable multiple small foci of mononuclear cell infiltration with minimal tissue damage. In contrast, liver tissue from IFN-γR^−/− (c and e) and iNOS^−/− (d and f) mice showed severe destruction (>60%) of the liver parenchyma with confluent necrosis. Mononuclear cell infiltration was comparable in both groups of deficient mice, but parasite numbers and amastigote nest sizes in the liver were greater in IFN-γR^−/− mice (e) than in iNOS^−/− mice (f). IFN-γR^−/− (g) and iNOS^−/− (h) heart sections showed many amastigote nests but poor inflammatory cell infiltrates. Shown are representative sections from five individual analyzed mice/group. Bar = 0.03 mm.

Impaired TNF-α and IL-1α responses in IFN-γR^−/− mice but normal inflammatory cytokine responses in iNOS^−/− mice after infection with T. cruzi.

The activation of macrophages by IFN-γ is one major function of a protective inflammatory cytokine response. To determine if these responses were different in the various mutant mouse strains, we measured the in vivo production of IFN-γ, TNF-α, and IL-1α in the blood of infected mice 7 and 10 days postinfection. IFN-γR^−/− mice showed elevated levels of IFN-γ but reduced TNF-α and IL-α levels compared to those of wild-type controls (Fig. 3a and c). In contrast, these inflammatory cytokine responses were normal in iNOS^−/− mice (Fig. 3b and d). To determine the contribution of T cells and macrophages to this inflammatory response, spleen cells from infected IFN-γR^−/−, iNOS^−/−, and wild-type mice were isolated at day 10 postinfection and restimulated with iTC, LPS, or anti-CD3 (Fig. 4). Only cells from infected mice were able to induce a cytokine response to iTC (data not shown). After restimulation with iTC and LPS, IFN-γR^−/− mouse-derived spleen cells showed a striking reduction of IL-1α production and an impaired TNF-α secretion after restimulation with LPS (Fig. 4). After iTC restimulation, TNF-α levels were also reduced but varied in different experiments. In contrast, production of IL-6 and IL-12 was not affected in response to iTC and LPS (Fig. 4). Similar results were found with iTC-restimulated peritoneal exudate cells (data not shown). The overall anti-CD3 induced cytokine responses were rather low (Fig. 4), with the exception of IFN-γ, indicating that T cells were not major contributors to the measured IL-1α, TNF-α, and IL-6 inflammatory cytokine responses at this point of the infection course. LPS and iTC restimulation induced similar cytokine levels, suggesting that macrophages are the major contributors to the IL-12, IL-1α, TNF-α, and IL-6 responses. Therefore, the impaired TNF-α and IL-1α production is probably caused by defective IFN-γR^−/− macrophages. However, neither T. cruzi-induced production of IFN-γ itself nor that of IL-12 or IL-6 was impaired in these mice, suggesting that cell activation by IFN-γ is not mandatory for sufficient production of these cytokines in this particular infection model.

FIG. 3.

Levels of IFN-γ, TNF-α, and IL-1α in plasma. Groups of five IFN-γR^−/− (a and c) or iNOS^−/− (b and d) mice (open bars) and their wild-type controls (hatched bars) were each infected with 1,000 or 500 blood trypomastigotes, and the cytokine content of the plasma was determined 7 (a and b) and 10 (c and d) days postinfection as described in Materials and Methods. Asterisks indicate statistically significant differences (P < 0.05) from values of wild-type controls as calculated by Student’s t test. Error bars, standard deviations.

FIG. 4.

Cytokine synthesis from restimulated IFN-γR^−/− spleen cells. Mice were each infected with 1,000 blood trypomastigotes of the Tulahuen strain of T. cruzi. At day 10, spleen cells of wild-type (hatched bars) and IFN-γR^−/− (open bars) mice were isolated and cultured for 48 h in the presence of either anti-CD3, iTC, LPS, or medium alone, and the cytokine levels of the cell culture supernatants were determined as described in Materials and Methods. For the experiment whose results are shown, the pooled cells of five mice per group were used. Results are expressed as the means + standard deviations (error bars) of triplicate wells. Asterisks indicate statistically significant differences (P < 0.05) from values of wild-type controls as calculated by Student’s t test and are shown only for cases in which statistically significant differences were found in two independent experiments.

Spleen cells from iNOS^−/− mice showed comparable IFN-γ, IL-1α, TNF-α, and IL-6 levels after restimulation with iTC, LPS, and anti-CD3 (data not shown). In conclusion, the absence of iNOS appears not to affect a protective inflammatory cytokine response to T. cruzi.

Normal T. cruzi-specific antibody response and T-cell polarization in IFN-γR^−/− mice.

To analyze specific immune responses in IFN-γR^−/− mice, T. cruzi-specific antibody titers were determined 14 days postinfection with 100 trypomastigotes (Fig. 5). T-cell polarization into Th1/Th2 and Tc1/Tc2 effector cells was determined following restimulation with immobilized anti-CD3 of enriched CD4⁺ and CD8⁺ lymph node-derived T cells (Fig. 6). Slightly increased T. cruzi-specific antibody titers of all measured isotypes, including antigen-specific IgG2a, were observed in IFN-γR^−/− mice. Moreover, normal IFN-γ and IL-4 levels were found in the supernatants of the T-cell subsets of mutant mice, suggesting an unaltered Th1/Th2 and Tc1/Tc2 polarization. In summary, these data suggest normal T-cell polarization and normal B-cell antibody responses in T. cruzi-infected IFN-γR^−/− mice.

FIG. 5.

Immunoglobulin isotype distribution in plasma of IFN-γR^−/− mice each infected with 100 blood trypomastigotes of the Tulahuen strain of T. cruzi. Blood was collected from wild-type (closed symbols) and IFN-γR^−/− (open symbols) mice at day 17 postinfection. Ig isotypes of trypomastigote-specific antibodies in plasma were determined by an antigen-specific enzyme-linked immunosorbent assay as described in Materials and Methods. Antibody titers of individual mice are shown. Horizontal bars indicate mean values of five mice/group. Comparable results were obtained in another independent experiment.

FIG. 6.

IFN-γ and IL-4 synthesis of CD4⁺ and CD8⁺ cells from IFN-γR^−/− mice each infected with 1,000 blood trypomastigotes. At day 14 postinfection, lymph node cells and sorted CD4⁺ and CD8⁺ cells of wild-type (hatched bars) and IFN-γR^−/− (open bars) mice were cultured for 48 h in the presence of immobilized anti-CD3 and the cytokine secretion in the supernatants was determined as described in Materials and Methods. In the experiment shown, pooled cells of five mice per group were used. Results are expressed as the means + standard deviations (error bars) of triplicate wells. Comparable results were obtained in another independent experiment.

Defective trypanocidal and NO activity of IFN-γR^−/− and iNOS^−/− macrophages.

After infection with T. cruzi, activated macrophages and other cells produce NO, an effector molecule able to kill T. cruzi (19). In T. cruzi-infected wild-type mice, substantial production of NO was measured in the blood (>40 mM) and supernatants of iTC- or LPS-restimulated peritoneal exudate and spleen cells (>30 mM) (Table 1), demonstrating an effective NO response in these mice. In contrast, iNOS^−/− mice showed only low-level NO production in vivo (<1.5 mM) and after antigen or LPS restimulation in vitro (<1.5 mM) (Table 1). Moreover, infected IFN-γR^−/− mice were also strikingly impaired in their ability to produce NO in vivo (<1.5 mM) and in vitro (1.7 to 6.3 mM) (Table 1), suggesting an IFN-γ dependency for effective NO production. This suggestion was confirmed by using mutant and wild-type BMMφ. T. cruzi-infected mutant macrophages were unable to induce a NO response and a reduction of intracellular amastigotes (Fig. 7). For NO production and mediation of their following function, T. cruzi-infected wild-type macrophages required incubation with IFN-γ. Both NO production and trypanocidal activities could be inhibited by coincubation with the iNOS inhibitor l-NMMA or by a neutralizing IFN-γ antibody (Fig. 7). This demonstrates and confirms an IFN-γ dependency of iNOS-mediated NO production as a major defense effector molecule for T. cruzi. In conclusion, these in vivo and in vitro results strongly suggest that the NO defect in infected IFN-γR^−/− and iNOS^−/− mice was responsible for their inability to kill intracellular T. cruzi, which in turn led to severe acute experimental Chagas’ disease and eventually death.

TABLE 1.

Nitric oxide production in vivo and in vitro

Source	NO secretion (mM)c
	IFN-γR		iNOS
	+/+	−/−	+/+	−/−
Plasmaa
Uninfected	<1.5	<1.5	<1.5	<1.5
Infected	41.0 ± 3.0	<1.5	53.8 ± 1.3	<1.5
Peritoneal cellsb
Medium	4.3 ± 0.6	3.3 ± 0.6	<1.5	<1.5
iTC	46.3 ± 1.5	3.0 ± 1.0	31.7 ± 1.7	<1.5
LPS	64.3 ± 3.2	5.3 ± 3.1	56.3 ± 1.7	<1.5
Spleen cellsb
Medium	12.3 ± 1.3	<1.5	11.0 ± 0.8	<1.5
iTC	33.7 ± 2.1	1.7 ± 1.3	13.3 ± 2.9	1.7 ± 0.9
LPS	51.0 ± 5.9	6.3 ± 0.9	29.7 ± 1.3	2.0 ± 0.0

^aGroups of five mice were each infected with 1,000 (IFN-γR^−/−) or 500 (iNOS^−/−) blood trypomastigotes, and blood was collected for NO determination at day 10 postinfection. 

^bCells of five mice per group were pooled at day 10 postinfection; triplicates of single-cell suspensions were restimulated with either medium, iTC, or LPS for 48 h; and the NO content of the supernatants was determined. 

^cResults are expressed as the means ± standard deviations. The results shown are from one representative of three experiments. 

FIG. 7.

Trypanocidal activity and NO production of IFN-γR^−/− (a) and iNOS^−/− (b) mouse-derived BMMφ after in vitro infection with T. cruzi. BMMφ were isolated from femora of naive wild-type (hatched bars) or naive mutant (open bars) mice, infected with T. cruzi trypomastigotes, and incubated with medium alone, IFN-γ (100 U/ml), IFN-γ (100 U/ml) plus l-NMMA (500 μM), or IFN-γ (100 U/ml) plus anti-IFN-γ (500 ng/ml) as described in Materials and Methods. Two days later, the NO content of the supernatants and the percentage of T. cruzi-infected cells (250 scored cells/well) were determined in relation to those for infected BMMφ incubated with medium alone. Comparable results were obtained in another independent experiment. Error bars, standard deviations.

DISCUSSION

Both IFN-γR^−/− and iNOS^−/− mice showed related susceptibilities to T. cruzi and succumbed during the acute phase of infection, with similar mortality kinetics and severe acute Chagas’ disease, even after infection with very low doses of trypomastigotes. We were able to demonstrate a common defect, responsible for the severe outcome of the disease. Both mutant strains were unable to produce NO in response to infection with T. cruzi in vivo and in vitro. Moreover, macrophages from both mutant mouse strains were defective in trypanocidal activities and parasites developed rapidly in these cells. These results strongly suggest the crucial importance of iNOS-mediated NO production as the major defense mechanism for surviving in vivo T. cruzi infection. A similarly crucial role of NO effector functions has been observed in experimental infections with Leishmania major (42, 48) and Mycobacterium bovis (3, 49). However, the importance of NO-mediated defense functions against intracellular pathogens cannot be generalized. For example, in toxoplasmosis or listeriosis the role of this effector mechanism is rather limited (14, 23, 28, 40) and other mechanisms seem to play a protective role.

On the other hand, IFN-γ is a crucial cytokine in all the above-mentioned diseases, as IFN-γ^−/− or IFN-γR^−/− mice are unable to survive these infections (6, 12, 24, 39, 46). After activation of the IFN-γR on macrophages, various tyrosine kinases are induced and proteins that are signal transducers and activators of transcription are phosphorylated, both of which regulate the induction and activation of transcription factors of the interferon regulatory factor family. Among these, interferon regulatory factor 1 directly mediates the expression of iNOS and subsequently the production of NO (22). We have shown that this IFN-γ-dependent iNOS expression (18, 19, 47) and the consequent NO production seem to be mandatory for surviving T. cruzi infection. However, recent studies have shown evidence that T. cruzi-infected IFN-γR^−/− mice are able to express low but detectable levels of iNOS mRNA, suggesting a dual pathway of iNOS induction and subsequent NO production (37). We also observed some residual NO in T. cruzi-infected IFN-γR^−/− peritoneal exudate cells (Table 1). Interestingly, IFN-γ-independent iNOS protein has been located in a different subcellular compartment than IFN-γ-dependent iNOS and could be localized on the membranes of amastigotes within parasite nests (37). Rottenberg et al. hypothesized that IFN-γ-independent iNOS may be involved in the enhancement of parasite proliferation (37). In Listeria monocytogenes-infected IFN-γR^−/− mice, a similar IFN-γ-independent pathway of iNOS expression was observed (12). In this case, neutrophils were major contributors to the overall iNOS transcript levels (unpublished data). Nevertheless, in experimental listeriosis NO effector functions have a limited protective capacity. Different IFN-γ-dependent effector mechanisms, such as the subsequent fusion of the phagosome with the lysosomal compartment, may be involved in efficient pathogen elimination (12).

In contrast to iNOS-deficient mice, which showed a normal proinflammatory cytokine response to T. cruzi infection, infected IFN-γR^−/− mice showed impaired TNF-α and IL-1α production in vivo and after in vitro restimulation with LPS. These data suggest that IFN-γR^−/− macrophages are defective in mediating these cytokine responses, demonstrating that TNF-α and IL-1α production is dependent on IFN-γ activation. This impaired cytokine production may contribute to the observed susceptibility of IFN-γR^−/− mice, since a protective role of TNF-α and IL-1α in experimental Chagas’ disease has been demonstrated (1, 26, 34). In contrast, in vitro IL-12 and IL-6 responses after antigen and LPS restimulation were comparable with those of the controls, suggesting that T. cruzi-induced expression of these cytokines by macrophages is independent of IFN-γR signalling. An IFN-γ-independent IL-12 response is in agreement with recent data which show that T. cruzi is able to induce IL-12 directly in macrophages without priming by IFN-γ (2, 16), as has been found in other infection models, e.g., L. monocytogenes (20), Staphylococcus aureus (5), and Toxoplasma gondii (17). IL-12 induces IFN-γ production by NK cells, which in turn leads to IFN-γ-dependent macrophage stimulation and subsequent activation of an effector function, which promotes an efficient early innate immune response. Normal IL-12 production after antigen restimulation of IFN-γR^−/− spleen or peritoneal exudate cells indicates that further stimulation by IFN-γ in a positive feedback loop is not mandatory for efficient production of both cytokines. This was recently demonstrated also in Listeria-infected IFN-γR^−/− mice (12). IL-12 is important in promoting Th1 development, which bridges innate and adoptive immunity (6) and leads to a Th1-type response which is usually protective in intracellular infections. As infected IFN-γR^−/− mice showed normal T-cell polarization, we suggest that sufficient IL-12 was present in these mice. The role of IFN-γ in directing T-cell differentiation is controversial, as in vitro studies and L. major infection studies using IFN-γ- and IFN-γR-deficient mice showed contradictary results (6). However, the normal Th1/Th2 and Tc1/Tc2 responses during T. cruzi infection in IFN-γR^−/− mice strongly suggest that IFN-γ is not crucial in the T-cell polarization. As IFN-γR^−/−–IL-4^−/− double-deficient mice had a prolonged survival time (unpublished data), the absence of a Th2 response due to IL-4 deficiency (IL-4^−/−) (25) may partially compensate for the impaired macrophage response. Nevertheless, normal mice seem to develop an optimal Th1 response, as no significant change in parasitemia and survival rate during the acute phase of infection was found in infected IL-4^−/− and wild-type mice (unpublished results).

NO is known to be involved in immune response regulation in that it inhibits the expansion of cloned Th1 but not Th2 cells (44). Consistently, restimulated spleen cells of L. major-infected iNOS^−/− mice produce more IFN-γ but less IL-4 than do wild-type mice (48). Moreover, an enhanced T-cell proliferation was found after inactivation of NO production of spleen cells isolated from T. cruzi-infected wild-type mice (1). Despite these observed down-regulatory properties of NO on T cells, the absence of inflammatory NO in T. cruzi-infected IFN-γR^−/− mice seemed not to have any effects on T-cell differentiation. We have not directly addressed T-cell differentiation in iNOS^−/− mice. However, restimulation of splenic T cells from T. cruzi-infected iNOS^−/− mice showed no increased IFN-γ levels compared to controls, indicating that in the absence of iNOS-mediated NO production Th1 responses were not increased.

The normal T-cell polarization in IFN-γR^−/− mice may also explain the normal humoral responses found with respect to T. cruzi-specific antibody production observed. Antigen-specific IgG2a response was also normal in IFN-γR^−/− mice, even though IFN-γ is known to regulate IgG2a isotype switches (41), shown by reduced 2,4-dinitrophenol–ovalbumin-specific IgG2a responses in IFN-R^−/− mice after immunization (21). However, our results suggest that T. cruzi-specific IgG2a responses are independent of IFN-γ signalling and that other factors may be involved in this isotype regulation.

In summary, these data demonstrate the limited effect of IFN-γR signalling on lymphocyte-specific immune responses in experimental Chagas’ disease. IFN-γ-independent but parasite-induced IL-12 seems to initiate and mediate the normal T-cell response observed in susceptible IFN-γR^−/− mice. This response is protective in immunocompetent mice. Hence, IFN-γ-dependent and iNOS-mediated NO production is crucial for macrophage trypanocidal activity and survival by mice of the acute phase of T. cruzi infection.