Stage-Dependent Role of Nitric Oxide in Control ofTrypanosoma cruzi Infection

ABSTRACT Trypanosoma cruzi, the causative agent of Chagas' disease, is known to be susceptible to nitric oxide (NO)-dependent killing by gamma interferon-activated macrophages. Mice deficient for inducible nitric oxide synthase (iNOS) are highly susceptible toT. cruzi, and inhibition of iNOS from the beginning of infection was reported to lead to an increase in trypomastigotes in the blood and to high mortality. In the present study, we investigated whether NO production is essential for the control of T. cruzi in all phases of the infection. BALB/c mice were treated at different time intervals after T. cruzi infection with an iNOS inhibitor, aminoguanidine orl-N6-(1-iminoethyl)-lysine (L-NIL). Treatment initiated with the beginning of the infection resulted in 100% mortality by day 16 postinfection (p.i.). If treatment was started later during the acute phase at the peak of parasitemia (day 20 p.i.), all the mice survived. Parasitemia was cleared and tissue amastigotes became undetectable in these mice even in the presence of the iNOS inhibitor L-NIL. Inhibition of iNOS in the chronic phase of the infection, i.e., from day 60 to day 120 p.i., with L-NIL did not result in a reappearance of parasitemia. These data suggest that while NO is essential for T. cruzi control in the early phase of acute infection, it is dispensable in the late acute and chronic phase, revealing a fundamental difference in control mechanisms compared to those in infections by other members of the orderKinetoplastida, e.g., Leishmania major.

Trypanosoma cruzi is an obligate intracellular protozoan parasite of mammals and the etiologic agent of Chagas' disease. The parasite infects a variety of host cell types, including macrophages; intracellular replication as amastigotes is followed by the release of trypomastigotes that can reach the bloodstream before infecting other host cells. The acute phase is characterized by a large increase in parasite replication, and trypomastigotes are observed in the blood of infected mice. After control of the acute phase in immunocompetent mice, the infection turns into a chronic phase (starting around day 21 postinfection [p.i.]) where parasites are no longer detectable by light microscopy in the bloodstream but form inflammatory nests in various tissues, a process associated with chagasic pathology, in which antiparasite cytotoxic T lymphocytes or autoimmune mechanisms may play a role (14,21,29).
Several cell subsets of both the innate and the specific immune system were reported to be required for survival during the acute phase of the infection in murine T. cruzi infection, such as NK cells (4), CD4 ϩ (20,27) and CD8 ϩ T cells (20,26,28), and B cells (11). Two essential mediators of resistance to T. cruzi have been found to be gamma interferon (IFN-␥) (1,10,12,18,23,31) and nitric oxide (NO) (10,17,33), which has direct strong cytotoxic effects on T. cruzi (6,33). IFN-␥ is thought to be the most important inducer of the inducible NO synthase (iNOS) for increased NO production by macrophages (2,8,9) and thus essential for mediation of NO-dependent parasite control during acute infection.
There have been several reports that effector cell pathways essential for T. cruzi control during acute infection are dispensable during chronic infection (4,16,26). In this study, we found that this also applies for NO production during the chronic and also during the late acute phase; both phases are characterized by control through the adaptive immune system and not through NK cells (4

MATERIALS AND METHODS
Mice and parasites. Six-to eight-week-old IFN-␥ knockout (KO) BALB/c mice and wild-type BALB/c littermates as well as C57BL/6 mice maintained under specific-pathogen-free conditions were used for the experiments.
Tulahuen strain T. cruzi blood trypomastigotes were routinely maintained by weekly intraperitoneal infection of BALB/c mice (7,13). Blood was collected from mice by orbital puncture under anesthesia in tubes containing EDTA. Blood (10 l) was diluted in 40 l of Tris-ammonium chloride to lyse erythrocytes. Trypomastigotes were microscopically counted in a hemocytometer. For intraperitoneal infection of mice, the number of trypomastigotes was adjusted by dilution in phosphate-buffered saline.
Treatment with L-NIL and AG and determination of nitrite. L-N 6 -(1-iminoethyl)-lysine (L-NIL) (Alexis, Grünwald, Federal Republic of Germany) and aminoguanidine (AG) (Sigma, Munich, Federal Republic of Germany) were dissolved at 3 mM and 90 mM, respectively, in drinking water, which was the only source for fluid intake of mice during the duration of blockade experiments. That the mice actually continued drinking normally was verified by recording of their weight twice a week. Production of NO in serum was assessed by determination of NO 2 Ϫ and NO 3 Ϫ in mouse sera (Griess reaction) as described elsewhere (19). Determination of IFN-␥ in serum. IFN-␥ concentrations in sera of mice were determined by specific two-site enzyme-linked immunosorbent assays using standard protocols. The antibody pairs for capture and detection (biotinylated) were purchased from Pharmingen (Hamburg, Federal Republic of Germany) in the combination recommended. Recombinant IFN-␥ (Pharmingen) was used as a standard. Enzyme-linked immunosorbent assays were developed after incubation with streptavidin-peroxidase complex (1:10,000; Boehringer, Mannheim, Federal Republic of Germany), using 3,5,3Ј,5Ј-tetramethylbenzidine as substrate (dis-solved [6 mg/ml] in dimethyl sulfoxide; Roth, Karlsruhe, Federal Republic of Germany). Sensitivity was 5 pg/ml. PCR. DNA was purified from blood samples using the QIAamp blood mini DNA kit from Qiagen (Hilden, Federal Republic of Germany). Two oligonucleotides (T1, 5Ј GAC GGC AAG AAC GCC AAG GCA 3Ј; T2, 5Ј TCA CGC GCT CTC CGG CAC GTT GTC 3Ј) derived from T. cruzi cDNA sequence coding for a 24-kDa protein were used as described elsewhere (15).
Histology. Previously described protocols for histology were followed (5). Complete hearts and complete femoral muscles from each animal were fixed in 70% ethanol and embedded in paraffin using standard methods. Sections (4 m thick) were prepared from each tissue. Tissue sections were first stained in hematoxylin (Merck, Darmstadt, Germany) for 20 min and rinsed in tap water for 30 min, and this was followed by counterstaining with 1% (wt/vol) eosin (Merck) for 3 min. The tissue sections were analyzed with a light microscope at a magnification of ϫ1,000. For quantification of amastigote pseudocysts, two sections from each tissue of each animal were examined by an investigator who was not aware if the mice had received NO-blocking treatment before.

NO is produced early in infection during trypomastigote parasitemia.
When infected with sublethal doses of 10 or 100 trypomastigotes, BALB/c mice developed peak parasitemia within 15 to 20 days p.i. (Fig. 1), consistent with the acute phase of T. cruzi infection. A high amount of IFN-␥ was observed in the serum from 10 days p.i., and it was followed by the appearance of NO (Fig. 1).
After 30 days p.i. mice developed the chronic phase of infection with no trypomastigote forms found in the blood (Fig.  1). Consistent with this, NO production decreased to levels undetectable in serum (Fig. 1). However, T. cruzi DNA was detected by PCR in the blood of BALB/c mice during both the acute and the chronic phases of T. cruzi infection (Fig. 2). T. cruzi DNA detectable in blood from the chronic phase was correlated with viable trypomastigotes since transfusion of those blood aliquots led to fulminant infection, with high parasitemia and death in IFN-␥ KO mice (not shown). Interestingly, IFN-␥ in the serum of wild-type BALB/c mice remained elevated in the chronic phase of infection ( Fig. 1).
NO is essential for control of parasitemia and for survival in the early acute phase of infection. Administration of AG or L-NIL starting 1 day before the infection led to a high parasitemia ( Fig. 3) and to 100% mortality of mice by day 17 p.i.  4). The efficacies of AG and L-NIL were equivalent in this trial. When NO production was blocked by AG from day 7 p.i., the development of high parasitemia was delayed by 8 days (i.e., for the time span treatment was delayed; i.e., day Ϫ1 versus day 7 [ Fig. 3A]), but mortality was still 100% (Fig. 4A). NO synthase inhibition from day 13 p.i. by AG resulted in only 50% mortality (Fig. 4A); in this group, those animals that died developed high parasitemia before death while those animals that survived eventually controlled their parasitemia (Fig. 3A). NO is dispensable for survival as well as blood and tissue parasite burden in the late acute phase of infection. The above results suggested that NO is indispensable at a time during early T. cruzi infection (day 1 to 20) when the adaptive immune system is not yet effective but that it may become dispensable To gain further evidence for the existence of such a time window where NO is essential, NO synthase activity was blocked during the time when the adaptive immune system in this model is known to be involved in the control of the parasites (26) but the innate, NK-derived IFN-␥-dependent defense mechanisms are no longer essential (4), i.e., from day 20 p.i. on. Note that at this time, parasitemia is not yet under control since blood trypomastigote levels do not drop before 22 days p.i. (Fig. 1). NO in mouse serum has similar kinetics (Fig. 5). NO synthase inhibition was carried out (i) with AG from day 20 p.i. (Fig. 5) to the time when there were no more trypomastigotes detectable in blood smears (day 30 p.i.) or (ii) with L-NIL from day 20 p.i. until day 42 p.i. This treatment led to an immediate drop in serum NO levels (Fig. 5). However, the parasitemia, which in this experiment was at its peak at day 20 p.i. (Fig. 3), declined to levels below the detection limit in blood smears in spite of either inhibitor, AG or L-NIL (Fig.  3B).
Importantly, NO depletion between day 20 and day 42 p.i. by L-NIL did not also affect the parasite burden in the hearts and skeletal muscles: the eight mice which had been treated with L-NIL from days 20 to 42 p.i. had been infected together with 24 control BALB/c mice which remained untreated. At weekly intervals, four mice of the untreated group were sacrificed, and amastigote pseudocysts in hearts and skeletal muscles were counted by histology. The number of pseudocysts peaked between days 14 and 20 and was reduced to one or zero per eight tissue sections (two sections from each animal per time point [ Fig. 6]). The subgroup of eight mice which were treated with L-NIL from days 20 to 42 p.i. was sacrificed at day 42 p.i.; two sections were counted from each tissue per animal. The results revealed that the amastigote pseudocyst number had also declined to levels equivalent to those of untreated animals (Fig.  6).
NO inhibition does not affect the chronic phase of T. cruzi infection. Inhibition of NO production by either AG or L-NIL during the chronic phase (days 60 to 120 p.i.) did not lead to mortality or reappearance of parasitemia (Table 1; Fig. 3B).

DISCUSSION
The results generated in this study allow a more precise view of the trypanocidal role of NO in the control of murine T. cruzi infection. The essentiality of iNOS induction for parasite control has been demonstrated using enzyme blockers (17) or iNOS KO mice (10); the latter mice had much higher parasite loads in blood and tissue, which were associated with extreme susceptibility (100% mortality after application of only 15 blood forms/mouse [10]). NO activity was linked to trypanocidal action of macrophages given that IFN-␥ rendered infected macrophages capable of inhibiting intracellular replication of T. cruzi in vitro by a mechanism that involves the production of NO (8). Recently, more molecular pathways of NO action on the parasites have been elucidated (30,32). By depletion studies as well as gene KO (10,17), IFN-␥ as well as a variety of cells that produce this cytokine (NK [4], ␥␦ [22], CD4 ϩ [20,27] and CD8 ϩ T cells [26,28]) was shown to be essential for protection against T. cruzi. However, the consensus of these studies is that depletion of any of these cell subsets affects only the early, acute phase of infection while it does not lead to a reversion of parasitemia if depletion is initiated after blood stage trypomastigotes have disappeared from the blood (3,12,26). FIG. 5. NO in the serum of BALB/c mice infected with 100 blood-derived trypomastigotes of the Tulahuen strain. The treatment with AG or L-NIL was started as indicated in the figure. For each time point, the arithmetic mean Ϯ standard deviation (error bar) for four, six, or eight animals is given. The data are representative of three independent experiments (except for treatment from days 20 to 30 p.i. which was done two times, and with L-NIL, which was done once).
We show here that this also applies to NO-mediated parasite control: while there was a strong exacerbation of parasitemia, with 100% mortality in animals treated from the beginning of infection with NO-synthase inhibitor L-NIL or AG (Table 1; Fig. 3), it was impossible to render animals parasitemic by long-term (60 days) AG or L-NIL application if started at day 60 p.i. (Table 1; Fig. 3). In addition, it was impossible to increase parasite loads both in blood and in tissue by NO blockade after day 20 p.i.; in contrast, tissue and blood parasite levels declined in the presence of either NOS inhibitor (Fig. 3). This is in contrast to murine infection with another kinetoplastid, Leishmania major. Here it was shown that in chronic infection (123 days p.i.), L-NIL application led to a reactivation of cutaneous lesions in C57BL/6 mice which became significant after only 17 days of treatment (24). In our own laboratory, we extended these findings when treating C57BL/6 mice with L-NIL towards the end of the acute phase of L. major infection, i.e., at a time when the footpad swelling is already declining: we observed a fast reactivation of cutaneous lesions in these mice which became significant as early as 10 days following the initiation of L-NIL administration (data not shown). This is in strong contrast to the further decline of T. cruzi parasite loads in blood and tissue at the end of the acute phase under L-NIL treatment (days 20 to 42 p.i.).
Thus, comparison of these two infections strongly suggests that fundamental differences in the control of chronic T. cruzi and L. major infections exist. A potential reason for this discrepancy is that L. major resides in macrophages which control infection by NO whereas T. cruzi also invades other types of cells which do not express iNOS.
A main novel finding of our study is that NO is not even essential for the whole period of acute infection: blood parasite levels, in our experimental system, did not decline before day 22 p.i. In untreated mice, the levels of NO reaction products in the serum fell concomitantly with blood parasite loads (Fig. 1), and NO synthase inhibition during that period did not block clearance of parasites from blood (Fig. 3). Consistently, NO synthase inhibition, if initiated from day 20 p.i., did not affect survival, in contrast to earlier inhibition, i.e., from day Ϫ1 or day 7 p.i. (Fig. 4). When treatment was started at day 13 p.i., mortality was only 50% (Fig. 4). This reflects a decreasing need for NO in parasitemia control from day 7 p.i. (essential) to day 20 p.i. (dispensable). Importantly, tissue amastigotes, after day 20 p.i., are also controlled by mechanisms other than NO, since FIG. 6. Quantification of amastigote pseudocysts from hearts and skeletal muscles of BALB/c mice infected with 100 blood-derived trypomastigotes. Tissues were subjected to histological analysis as described in Materials and Methods. Amastigote pseudocysts were counted with a light microscope at a magnification of ϫ1,000. Two sections of different parts of each tissue from each animal were completely analyzed. Thirty-two animals were used for this experiment. At each time point from day 7 until day 42 p.i., four animals were sacrificed and tissues were analyzed. The arithmetic means ϩ standard deviations (error bars) are given. ‫,ء‬ eight animals which had been infected in the same experimental series were given L-NIL from days 20 to 42 p.i. and were sacrificed at day 42 p.i. L-NIL treatment did not prevent the reduction of pseudocysts in hearts and skeletal muscles, so that equivalent low levels of amastigotes were found both in the presence and in the absence of L-NIL at day 42 p.i. (Fig. 6). The question of whether AG is as potent an NO synthesis blocker as is L-NIL may be raised. AG was shown to be indeed inferior to L-NIL on a molar ratio in vitro (25). However, these authors did not compare the two inhibitors in vivo, in contrast to us, and we adapted the molarity of the inhibitors for our in vivo study to equipotent levels. We always got consistent results with AG and L-NIL (Table 1; Fig. 3 and 4).
Our data fit very well to earlier observations that IFN-␥ itself is not needed during the whole acute phase of T. cruzi infection: IFN-␥ neutralization did not alter the parasite load of mice if applied later than day 11 p.i. (4). It was also shown that this early IFN-␥ production is dependent on the presence of NK cells in T. cruzi infection (4). In contrast, if T. cruzi-infected mice are depleted of CD8 ϩ T cells, parasitemia is unaltered compared to that in control infected mice until day 21 p.i., but thereafter, blood parasite loads rise quickly and mice die from infection (26). Mice lacking CD8 ϩ T cells through gene KO also die around day 28 (i.e., later than in our study when NO production was blocked) despite elevated IFN-␥ (28).
Taking all these results into account, it appears that NO is essential for the parasite control during the first 2 weeks of infection, i.e., a time when NK cells are known to control the parasite load through IFN-␥ production (4). It is, however, dispensable thereafter in late acute infection, i.e., during the period when blood parasite levels decrease under the control of CD8 ϩ T cells (26). Thus, the synopsis of the results suggests a cascade of events where NK cells, by their production of IFN-␥, activate effector cells, most likely macrophages, to induce NOS2 and to exert control of early parasite replication. After 2 weeks p.i., this mechanism is gradually replaced by the action of the adaptive immune system in which CD8 ϩ T cells (20,26,28), and apparently CD4 ϩ T cells (20,27) and B cells (11) producing neutralizing antibodies, act in concert to further limit parasite replication in an NO-independent way.