P47phox−/− Mice Are Compromised in Expansion and Activation of CD8+ T Cells and Susceptible to Trypanosoma cruzi Infection

Macrophage activation of NAD(P)H oxidase (NOX2) and reactive oxygen species (ROS) is suggested to kill Trypanosoma cruzi that causes Chagas disease. However, the role of NOX2 in generation of protective immunity and whether these mechanisms are deregulated in the event of NOX2 deficiency are not known, and examined in this study. Our data showed that C57BL/6 p47phox−/− mice (lack NOX2 activity), as compared to wild-type (WT) mice, succumbed within 30 days post-infection (pi) to low doses of T. cruzi and exhibited inability to control tissue parasites. P47phox−/− bone-marrow and splenic monocytes were not compromised in maturation, phagocytosis and parasite uptake capacity. The deficiency of NOX2 mediated ROS was compensated by higher level of inducible nitric oxide synthase (iNOS) expression, and nitric oxide and inflammatory cytokine (TNF-α, IFN-γ, IL-1β) release by p47phox−/− macrophages as compared to that noted in WT controls infected by T. cruzi. Splenic activation of Th1 CD4+T cells and tissue infiltration of immune cells in T. cruzi infected p47phox−/− mice were comparable to that noted in infected control mice. However, generation and activation of type 1 CD8+T cells was severely compromised in p47phox−/− mice. In comparison, WT mice exhibited a robust T. cruzi-specific CD8+T cell response with type 1 (IFN-γ+TNF-α>IL-4+IL-10), cytolytic effector (CD8+CD107a+IFN-γ+) phenotype. We conclude that NOX2/ROS activity in macrophages signals the development of antigen-specific CD8+T cell response. In the event of NOX2 deficiency, a compromised CD8+T cell response is generated, leading to increased parasite burden, tissue pathogenesis and mortality in chagasic mice.


Introduction
Chagas disease is caused by the protozoan Trypanosoma cruzi [1,2]. During acute phase of infection, parasites can be found in the circulating blood, and host may develop fever or swelling around the site of inoculation, and rarely, severe inflammation in heart muscle or brain. Several years after exposure to T. cruzi, ,30% of the infected individuals develop clinical symptoms of chronic cardiomyopathy associated with progressive cardiomegaly, arrhythmia, thromboembolic events, and heart failure [3,4].
Both innate and acquired immune responses are required for control of T. cruzi and critical for host survival (reviewed in [5,6]). Upon infection, macrophages serve as first responders by activation of phagocytic NADPH oxidase, referred as NOX2. NADPH oxidase is a multi-subunit complex and utilizes NADPH as an electron donor to reduce O 2 to superoxide (O 2 ?2 ), that is then dismutated into other oxidants (e.g. H 2 O 2 ) [7]. The plasma membrane-associated proteins gp91 phox and p22 phox compose the flavocytochrome-b558 complex that is the major component responsible for enzyme stability and activity. Phosphorylation of cytosolic factors (p47 phox , p67 phox , and p40 phox ), and small Rho GTPases in response to exogenous or endogenous stimuli initiates their translocation to the cell membrane, and NADPH oxidase activation [7][8][9]. Activated phagocytes exert cytotoxic effects via NOX2-dependent reactive oxygen species (ROS) production that mediates pathogen killing by oxidative damage of DNA, proteins and lipids, and suggested to play an important role in control of T. cruzi [10][11][12][13][14].
Besides innate immune mechanisms, a body of literature demonstrates that adaptive immune responses are required for parasite control. CD4 + T cells assist in the control of T. cruzi through secretion of Th1 cytokines, amplification of the phagocytic activity of macrophages, stimulation of B cell proliferation and antibody production, and enhancement of the CD8 + T cells response (reviewed in [6,15]. CD8 + T cells recognize processed parasite antigens presented in association with MHC class I molecules on the surface of infected host cells and contribute to the control of T. cruzi, either by cytolysis of parasite-infected cells or by the secretion of cytokines that may induce trypanocidal activity (reviewed in [6,16]). Current literature suggests that NADPH oxidase activity may modulate adaptive immune responses via ROS signaling of cytokine gene expression and regulation of the efficient antigen presentation for T cell activation and proliferation [17,18], though the cell type involved in NADPH oxidasemediated regulation of adaptive immunity are not fully detailed.
In this study, we have assessed the host response to T. cruzi infection in the event of phagocytic NADPH oxidase deficiency. We first monitored the susceptibility of wild-type (WT) versus p47 phox2/2 mice to T. cruzi infection, and then proceeded with a step-wise approach to identify the immune mechanisms that may be altered and contributed to susceptibility of p47 phox2/2 mice to T. cruzi. Our data show that p47 phox2/2 macrophages were not compromised in phagocytic activity, and mounted enhanced levels of inducible nitric oxide synthase (iNOS), nitric oxide (NO), and cytokines in response to T. cruzi infection. In vivo activation of CD4 + T cell subset and inflammatory cytokine response was also similar to or more pronounced in p47 phox2/2 mice when compared to that observed in WT controls in response to T. cruzi infection. However, in the event of NOX2 deficiency, generation and activation of CD8 + T cell response was severely compromised leading to increased parasite burden, tissue pathogenesis and mortality. We discuss the involvement of distinct innate receptor signaling pathways governing the activation and proliferation of T cell subsets and the various mechanisms contributing to increased susceptibility of p47 phox2/2 mice to T. cruzi infection.

Susceptibility of p47 phox2/2 mice to T. cruzi infection
We used well-established experimental models [19,20] to assess the role of NAD(P)H oxidase (NOX2) in immunity to T. cruzi infection. C57BL/6 (WT and p47 phox2/2 ) mice were assessed at day 7 post-infection (pi) for the expression level of p47 phox as an indicator of NOX2 activation in innate immune cells. The low level of baseline expression of p47 phox was increased by 2-fold in splenic (Fig. 1A) and bone-marrow monocytes/macrophages of WT mice. The splenic and BM monocytes of p47 phox2/2 mice exhibited no expression of p47 before or after T. cruzi infection. These data confirmed that p47 phox2/2 mice lacked the ability to induce NOX2 activity in phagocytes in response to T. cruzi infection.
Challenge infection with 10,000 T. cruzi per mouse proved to be lethal for p47 phox2/2 mice as all mice succumbed within 28 days pi (Fig. 1B.a). When inoculum was reduced to 2000 parasites, 70% of p47 phox2/2 mice still succumbed by 30 days pi. In comparison, 100% of WT mice challenged with 2000 or 10000 parasites survived (Fig. 1B.a).
The increased mortality of p47 phox2/2 mice was associated with increased tissue parasites ( Fig. 1B.b, Fig. 1.C-E). Histological analysis of skeletal muscle and heart tissue sections (three sections/ tissue .10-microscopic fields (mf) per slide, n = 8 mice/group) was conducted to obtain a score of parasite foci in tissues (Table 1). An average of skeletal tissue parasite foci in WT and p47 phox2/2 mice infected with 2000 or 10000 parasites is presented in Fig. 1B.b. The p47 phox2/2 mice infected with 2000 or 10000 parasites exhibited an early increase in tissue parasitemia by day 7 pi that further increased in a linear manner at days 14 and 21 pi (Fig. 1B.b). In comparison, WT mice exhibited a delayed, 2-5-fold lower level of parasite foci in skeletal muscle tissue during 7-21 days pi ( Fig. 1B.b). At 30 days pi, parasite foci in WT mice infected with 2000 or 10000 parasites were controlled, while p47 phox2/2 continued to exhibit an increase in tissue parasite foci ( Fig. 1B.b). A similar pattern of increase in parasite foci was observed in heart tissue of p47 phox2/2 mice during the 7-30 days pi (Table 1). We noted .5 parasite foci/mf in heart tissue sections of p47 phox2/2 mice infected with 2000 or 10000 parasites and harvested at day 30 pi (Fig. 1C.b). In comparison, contained (0-2 pseudocysts/mf) were noted in heart tissue of WT mice infected with 2000 or 10000 parasites ( Fig. 1C.a, Table 1) A semiquantitative PCR showed the Tc18SrDNA signal was significantly higher in the myocardium of infected/p47 phox2/2 mice at day 30 pi than was observed in the myocardium of WT mice infected with the same dose or 5-fold higher dose of parasites (Fig. 1D). Quantitative real-time PCR validated the findings of semiquantitative PCR and showed 2-5-fold increase in myocardial (Fig. 1E, ## p,0.01), and skeletal muscle and circulatory parasite burden in infected/p47 phox2/2 mice as compared to that detected in infected/WT mice. These data suggested the p47 phox2/2 mice failed to control tissue parasites and succumbed to T. cruzi infection.
Phagocytic activity of P47 phox2/2 macrophages is not compromised One plausible explanation for increased susceptibility of p47 phox2/2 mice to T. cruzi could be that p47 phox2/2 macrophages were compromised in phagocytic activity, and, therefore, failed to control parasites' dissemination. To test this, we isolated BM and splenic monocytes from WT and p47 phox2/2 mice, in vitro differentiated to macrophages with interferon gamma (IFNc), and incubated in presence of T. cruzi for 0, 6, 12, 24 h. The data presented in Fig. S1 are from splenic monocytes and representative of the results from triplicate experiments with splenic and BM monocytes. Giemsa staining showed the monocytes of p47 phox2/2 mice had a similar or better capacity than the WT monocytes to differentiate to macrophages by 6 h pi ( Fig. S1.b&g). Likewise, p47 phox2/2 phagocytes' capacity to uptake parasites (i.e. phagocytic efficiency) was not significantly compromised. Counting of .200 cells/slide showed that by 6 h, 25% and 15% of WT and p47 phox2/2 macrophages were infected (average 6-15 parasites/cell), and at 12 h, .50% of WT and p47 phox2/2 macrophages were full of replicative, amastigote form of parasites ( Fig. S1.c,d,h,i). At 24 h pi, some of the WT macrophages were ruptured releasing parasites while

Author Summary
Macrophage activation of NADPH oxidase NOX2) and reactive oxygen species (ROS) is suggested to mediate control of Trypanosoma cruzi infection that is the causative agent of Chagas disease. However, how NOX2/ROS deficiency affects parasite persistence and chronic disease is not known. In this study, we present the first evidence that NOX2 and ROS shape the T cell-mediated adaptive immunity, and its deficiency result in compromised splenic activation of type 1 cytotoxic CD8 + T cell response to T. cruzi infection. Subsequently, p47 phox2/2 mice that lack NOX2 activity were more unable to control parasite replication and dissemination and succumbed to susceptible to T. cruzi infection. Our study highlights how redox state of innate immune cells alters the adaptive immunity to intracellular pathogens; and suggests that understanding the molecular and cellular mechanisms affected by redox state of immune cells at basal level could be exploited in designing future therapeutic and vaccination strategies against T. cruzi infection and Chagas disease. p47 phox2/2 macrophages continued to exhibit parasites contained within phagosome (Fig. S1.e&j, p,0.01).
To obtain a quantitative measure of parasite uptake, primary BM and splenic cells were incubated for 24 h with CFSE-labeled T. cruzi, and then labeled with fluorescence-conjugated antibodies to examine the frequency of CFSE + macrophages (APC-CD11b + ) and neutrophils (PE-Ly6B + ) by flow cytometry. Representative flow cytometry data from BM-macrophages and BM-neutrophils incubated with CFSE-labeled parasites are presented in Fig. 2A, and percentage of CFSE + CD11b + and CFSE + Ly6B + macrophages and neutrophils, respectively, from BM and splenic cells are presented in Fig. 2B. We noted a higher extent of infection of BM cells derived from p47 phox2/2 mice as compared to that noted in BM cells from WT mice (CD11b + macrophages: 71% versus 36%; Ly6 + neutrophils: 23.5% versus 13.5%; p47 phox2/2 versus WT, respectively, Fig. 2A & Fig. 2B.a). The splenic cells from p47 2/2 and WT mice exhibited comparable rate of infection efficiency at 24 h post-incubation that were not statistically different (CD11b + macrophages: 10.65% versus 16.8%; Ly6+ neutrophils: 3.28% versus 9.45%, p47 phox2/2 versus WT, respectively, Fig. 2B.b). The hemacytometer counting of parasites in supernatants showed comparable number of parasites were released from p47 phox2/2 and WT macrophages at 24 h and 48 h pi (Fig. 2C). Together, the data presented in Fig. S1 and Fig. 2 suggested that BM and splenic monocytes from WT and p47 phox2/2 mice were equally competent in differentiating to macrophages and parasite uptake,  Table 1. Tissue burden of parasite foci and inflammation in wild type and p47 phox2/2 mice infected by T. cruzi. C57BL/6 mice (wild type and p47phox2/2) were infected with T. cruzi, as detailed in Materials and Methods. Tissue sections (skeletal muscle, heart) were stained with H&E. In general, we analyzed each tissue-section for .10microscopic fields (1006 magnification), and examined three different skeletal muscle or left ventricular (LV) tissue sections/mouse (4 mice/group) to obtain a semi-quantitative score of parasitic foci (cells filled with parasites).
Myocarditis (presence of inflammatory cells) was scored as 0 (absent), 1 (focal or mild with #1 foci), 2 (moderate with $2 inflammatory foci), 3 (extensive with generalized coalescing of inflammatory foci or disseminated inflammation), and 4 (diffused inflammation with severe tissue necrosis, interstitial edema, and loss of integrity and NOX2 deficiency did not result in increased parasite release from infected macrophages.
In vitro functional activation of p47 phox2/2 macrophages in response to T. cruzi Macrophages control the invading pathogen through production of ROS, NO, and inflammatory cytokines. Isolated BM and splenic monocytes from WT and p47 phox2/2 mice were incubated for 0, 6, 12 and 24 h with T. cruzi (6 recombinant IFN-c). We measured ROS levels using H 2 DCFDA that is cell permeable and when oxidized by ROS, releases fluorescent DCF. A gradual increase in DCF fluorescence beginning at 6 h that reached the maximal level at 24 h pi was observed in WT macrophages. Macrophages from WT mice responded to T. cruzi infection (6 rIFN-c) by 5-fold increase in DCF fluorescence ( Fig. 3A.a). The p47 phox2/2 splenic macrophages (6rIFN-c) exhibited 2.5-fold lesser DCF fluorescence in response to T. cruzi when compared to that noted in infected/WT cells ( Fig. 3A.a, ## p,0.01). Likewise, the p47 phox2/2 BM macrophages (6 rIFN-c) exhibited a 2-fold decline in ROS levels as compared to that noted in WT BM macrophages upon T. cruzi infection. The superoxide-dependent formation of formazan blue crystals was noted to be significantly increased in WT and slightly increased in p47 phox2/2 splenic macrophages incubated for 24 h with T. cruzi ( Fig. 3A.b). T. cruzi-induced increase in DCF fluorescence and nitroblue tetrazolium (NBT) reduction was quenched by .90% when cells were incubated in presence of 0.5 mM apocynin (inhibits NADPH oxidase activity) or 10 mM N-acetyl cysteine (ROS scavenger), suggesting the observed increase in ROS is primarily due to NOXdependent ROS from infected macrophages. The iNOS mRNA level, determined by qRT-PCR, was increased by 4-fold in p47 phox2/2 splenocytes as compared to that noted in WT controls, infected in vitro for 24 h (Fig. 3B). The release of cytokines in supernatants of primary BM and splenic cells incubated with T. cruzi for 24 h was measured by an ELISA. The p47 phox2/2 BM and splenic cells responded to T. cruzi by .10-fold increase in IFN-c and TNF-a release that was significantly higher than that observed in infected/WT cells (Fig. 3C). Together the data presented in Fig. 3 suggested that p47 phox2/2 macrophages lacked the ability to mount a strong NOX2-dependent ROS; however, exhibited a higher extent of iNOS and proinflammatory cytokine expression in response to T. cruzi infection. The p47 phox2/2 monocytes/macrophages also responded to heatinactivated T. cruzi with a strong proinflammatory cytokines production.

Tissue inflammatory infiltrate in p47 phox2/2 mice infected by T. cruzi
To further gain an indication of the effects of phagocytes' NOX2 deficiency on host immunity to T. cruzi; we looked at the tissue infiltration of immune cells in T. cruzi infected WT and p47 phox2/2 mice by histological studies (Fig. 4, Fig. S2, and Table 1). The p47 phox2/2 mice injected with 2000 parasites exhibited infiltration of inflammatory infiltrate in skeletal muscle and heart tissue as early as day 7 post-infection (pi, Fig. 4.e, Table 1). The inflammatory foci were observed in all tissue sections by 14 days pi (score: 2), and extensive inflammation with large inflammatory foci or diffused inflammation throughout the tissue section (score: 2-4) was observed at 21-30 days pi in skeletal muscle (Fig. 4.f-h) and heart tissue ( Fig. S1.m-o) of p47 phox2/2 mice. In WT mice, infection with 2000 parasites resulted in minimal inflammation of the skeletal muscle (Fig. 4.a&b) and heart tissue (Table 1) at 7-14 days pi; and inflammatory infiltrate was moderately increased (score: 1-2) at 21-30 days pi ( Fig. 4.c&d, Table 1). These data suggested that p47 phox2/2 mice responded to T. cruzi infection (2000 parasites/mouse) with an increase in tissue infiltration of inflammatory infiltrate that was higher than that observed in WT mice given the same dose of parasites. Infection with a 5-fold higher dose of parasites was required to elicit the extent of increase in inflammatory infiltrate in skeletal muscle (score: 2-4, Fig. S1.d&e) and heart tissue (score: 1-2, Fig. S1.i&j) of WT mice as was noted in skeletal muscle and heart tissue of p47 phox2/2 mice infected with 2,000 parasites.

Functional response of splenocytes and BM cells in p47 phox2/2 mice infected with T. cruzi
To examine the quality of inflammatory response in vivo, WT and p47 phox2/2 mice were harvested at day 7, 14, 21, and 30 postinfection. BM and splenic cells from infected mice were either directly analyzed or in vitro stimulated in presence of T. cruzi trypomastigote lysate (TcL) and utilized for functional assessment. Shown in Fig. 5A are intracellular ROS levels in splenic cells of infected mice at day 30 pi, determined by dihydroethidium (DHE) fluorescence. DHE is cell permeable, and when oxidized to ethidium, accumulates in nuclei and fluoresces bright red. We noted a significant increase in ethidium fluorescence in splenocytes of infected/WT, but not of infected/p47 phox2/2 mice, at all timepoints pi (Fig. 5A.a&b). Likewise, BM cells isolated at day 7, 14, 21, and 30 from infected/WT mice, but not from infected/ p47 phox2/2 mice, exhibited a significant increase in DHE fluorescence. DHE fluorescence was quenched when cells were incubated in presence of 0.5 mM apocynin (NOX2 inhibitor). Note that 49-6-diamidino-2-phenylindole-dihydrochloride (DAPI, binds nuclear DNA) staining of the splenocytes (Fig. 5A.c&d) of infected/WT and infected/p47 phox2/2 mice was comparable.
The intracellular nitric oxide levels in BM and splenic cells harvested at day 7, 14, 21 and 30 pi was first determined by DAF-FM-based fluorimetry. DAF-FM is cell permeable, and forms fluorescent benzotriazole upon reaction with nitric oxide. Shown in Fig. 5B.a are arbitrary units of DAF-FM fluorescence in splenic cells of infected mice harvested at day 30 pi. Our data showed a 4fold increase in DAF-FM fluorescence in splenocytes of infected/ p47phox2/2 mice that was further increased upon in vitro stimulation with TcL. In comparison, splenic cells of infected/WT mice exhibited a significant increase in intracellular nitric oxide (DAF-FM fluorescence) only after secondary in vitro stimulation with TcL, and this response was ,3-fold lesser than that observed with splenocytes of infected/p47phox2/2 mice (Fig. 5B.a). Because DAF-FM may exhibit non-specific signal by reacting with N compounds others than nitric oxide, we also performed a Griess reagent assay to evaluate the nitric oxide production rate, reflected by nitrite release. Splenocytes of infected/p47 phox2/2 mice, in vitro stimulated with TcL, exhibited a robust increase in nitrite release that was .5.8-fold higher than that noted with splenic cells from infected/WT mice ( Fig. 5B.b). Likewise, BM monocytes of infected/p47 phox2/2 mice responded to in vitro antigenic stimulus (TcL) by a robust 7-fold and 8-fold increase in DAF-FM fluorescence and nitrite release, respectively. The extent of TcL-stimulated nitrite release was 4-fold (32.864.7 versus 8.0760.6 pg nitrite/ml) higher in BM cells of infected/ p47 phox2/2 mice than that noted in BM cells of infected/WT mice ( ## p,0.001). In all experiments, incubation of splenic or BM monocytes from infected mice with 5 mmol/ml L-NAME (inhibits iNOS activity/nitric oxide) abolished the DAF-FM fluorescence and nitrite release. To examine the in vivo cytokine profile in response to infection, BM and splenic cells from WT and p47 phox2/2 mice were harvested at 7, 14, 21, and 30 days pi, in vitro incubated with or without second antigenic stimulus for 48 h, and supernatants were submitted to an ELISA. Overall, splenocytes of WT and p47 phox2/2 mice (6 T. cruzi lysate) were activated early upon infection, as is evidenced by a significant increase in TNF-a, IFN-c and IL-10 levels at day 7 pi (Fig. 5C). No IL-4 release was observed. The splenocytes of infected/WT mice exhibited a predominance of TNF-a (TNF-a.IL-10.IFN-c) release throughout the course of infection (Fig. 5C.a,c,e). In comparison, splenocytes of infected/p47 phox2/2 mice (6 TcL) exhibited a mixed response with a predominance of IL-10 (IL-10.TNF-a. IFN-c) at 7, 14, 21 and 30 days post-infection ( Fig. 5C.b,d,f). The BM cells of WT and p47 phox2/2 mice infected with T. cruzi (6TcL) exhibited a similar pattern of cytokine response as was noted in splenocytes. Together, the data presented in Fig. 5 suggested that compromised ROS production capacity due to NOX2 deficiency was compensated by an increased iNOS and nitric oxide levels in p47 phox2/2 mice infected with T. cruzi. However, p47 phox2/2 mice exhibited a subdued proinflammatory cytokine response (IL-10.TNF-a) during T. cruzi infection. Characterization of T cell response in p47 phox2/2 mice infected by T. cruzi CD4 + and CD8 + T cells are important constituents of the adaptive immunity against T. cruzi. To gain an appreciation for the role of NOX2 in determining T cell functional profile, we evaluated the in vivo quality and potency of the cellular immune responses elicited in WT versus p47 phox2/2 mice. Splenocytes, harvested at 30 days post-infection, were incubated in presence and absence of TcL antigenic stimulus, and T cell proliferation determined by an MTT assay (Fig. 6A). The CD4 + and CD8 + T cells were examined for proliferative capacity (Ki67 + ), intracellular cytokine profile (IFN-c, TNF-a) and marker of lytic capacity (CD107a) by flow cytometry. The mean fluorescence intensity (6SD) indicative of T cell profile (Fig. 6B, n = 6/group) were derived from representative quadrant images of flow cytometry results presented in Fig. 6C. Splenic lymphocytes of WT mice exhibited 3.6-4.2-fold increase in proliferation in response to T. cruzi infection (6 in vitro stimulation with TcL, Fig. 6A). In comparison, p47 phox2/2 mice exhibited a ,40% lower rate of splenic lymphocyte proliferation in response to T. cruzi infection, and no effect of in vitro stimulation with TcL was noted (Fig. 6A, ## p,0.01).
When we performed the specific T cell population analysis, we found that in vivo population of CD4 + T cells was comparable in naïve WT and p47 phox2/2 mice (range: 12-14%, Fig. 6B.a), and exhibited no significant change in proliferation (Ki67 + ) or IFN-c + phenotype in response to T. cruzi infection or subsequent incubation with TcL antigenic lysate. Instead, in vivo percentage of CD8 + T cells in p47 phox2/2 naïve mice (4.5%) was ,2-fold lower than that noted in WT normal controls (7.9%) (Fig. 6B.a). In response to T. cruzi infection, CD8 + T cells expanded by 3.4-fold in WT mice (27% of total) while these cells expanded at a very low frequency in p47 phox2/2 mice (Fig. 6B.a). Functional characterization of CD8 + T cells showed that a majority of the CD8 + T cells were proliferative (Ki67 + ) in infected/WT mice (Fig. 6B.b) and a significant proportion of the CD8 + Ki67 + cells (up to 13%) produced IFN-c in an antigen-specific manner (Fig. 6B.c). In p47 phox2/2 mice, CD8 + T cells exhibited no significant proliferating phenotype in response to T. cruzi infection (Fig. 6B.b). Further, up to 16% of the CD8 + T cells exhibited IFN-c + phenotype in naïve p47 phox2/2 mice and these didn't increase in response to T. cruzi infection or second antigenic stimulation with TcL ( Fig. 6B.b).
Transport of CD107a and CD107b to the plasma membrane of effector T cells is required for a) the cytolytic activity mediated by perforin and granzymes and b) the release of IFN-c which exerts pleiotropic effects to suppress intracellular pathogens. Our data showed T. cruzi infection induced CD107a + CD8 + T cells (2-4%) in infected/p47 phox2/2 mice (Fig. 6B.d), and a majority of these exhibited dual-positive (CD107a + IFN-c + ) cytolytic phenotype, comparable to that noted in infected/WT mice. Together, the data presented in Fig. 6 suggested that splenic CD8 + T cells in p47 phox2/2 mice were low in number and failed to expand in response to T. cruzi infection, resulting in a substantial decline in proliferating, IFN-c-producing cytolytic CD8 + T cell response. In comparison, an expansive CD8 + T cells proliferation that were predominantly IFN-c + with cytolytic capacity, and, thus, had a potential to act as effector T cells was induced in WT mice infected by T. cruzi.

Discussion
The present study shows that in the absence of NOX2 activity, a defective activation of CD8 + T cell occurs, and contributes to the inability of mice to successfully control T. cruzi infection. Our data suggested that the NOX2 deficiency was compensated by enhanced levels of iNOS, nitric oxide, and inflammatory cytokines in macrophages; however, p47 phox2/2 mice were highly susceptible to T. cruzi because of the inability to activate a type 1 CD8 + T cell response that is known to be essential for intracellular parasite control. Our study highlights how redox state of innate immune cells alters the adaptive immunity to intracellular pathogens, and understanding the molecular and cellular mechanisms affected by redox state of immune cells at basal level could be exploited in designing future vaccination strategies against T. cruzi infection and Chagas disease.
Current literature demonstrates that macrophage-derived free radicals (O 2 N 2 , nitric oxide) generated by the NOX2 complex and iNOS participate in cytotoxic mechanisms against microorganisms (reviewed in [21] [22]). In the context of T. cruzi, it is suggested that nitric oxide plays a central role through its action on macrophage-derived peroxynitrite formation, a strong cytotoxic oxidant that is formed by the reaction of nitric oxide with O 2 N 2 [14,23,24]. Our in vitro studies showed that p47 phox2/2 monocytes were better than (or equal to) the WT controls in their ability to differentiate into macrophages and phagocytize parasites (Fig. S1). Though it appeared that a higher number of intracellular parasites were present in infected p47 phox2/2 macrophages at 24 h pi ( Fig. S1 & Fig. 2); however, the extent of parasite release from p47 phox2/2 cells at 24 and 48 h pi was comparable to that noted in WT controls (Fig. 2C), thus suggesting that NOX2 deficiency did not result in increased parasite survival in infected macrophages. Our observations are supported by others demonstrating the enhanced replication of bacteria (e.g. Coxiella burnetii) in p47 phox2/2 macrophages, that was followed by a slightly delayed control of infection at a rate similar to the WT macrophages [25]. We propose that p47 phox2/2 macrophages, despite a lack of NOX2/ROS, were equipped to phagocytize and control the parasites through compensatory mechanisms. One, a low but detectable level of O 2 N 2 production in p47 phox2/2 macrophages (Fig. 3A&B) was sufficient to support the nitric oxide mediated cytotoxic peroxynitrite formation for parasite killing. Indeed, O 2 N 2 and nitric oxide can rapidly diffuse (diffusion control rates: k,1010 M21 s21) and react to form peroxynitrite that is significantly more potent cytotoxin against trypomastigotes than H 2 O 2 only [13,14]. Secondly, a significant up regulation of the iNOS, nitric oxide and inflammatory cytokines (IFN-c/TNF-a) in p47 phox2/2 macrophages in response to T. cruzi infection (Fig. 3) could have controlled the infectious pathogen. Others have also demonstrated increased iNOS and nitric oxide levels in gp91 phox2/2 mice infected by T. cruzi [26]. We surmise that in the event of defects in mounting NOX2/ROS, macrophages are capable of using alternative, compensatory mechanisms for pathogen control. Further studies will be required to conclusively establish if the peroxynitrite formation rate is indeed enhanced and identify the signaling mechanisms that were up regulated resulting in enhanced iNOS and inflammatory cytokines' expression in p47 phox2/2 macrophages in response to T. cruzi infection.
The production of cytokines (IL-12, TNF-a) by innate immune cells (macrophages, dendritic cells (DCs) shapes the adaptive immunity via activation of T cells. CD4 + and CD8 + T cells producing type 1 cytokines and CD8 + T cell mediated cytolytic activity are required for control of T. cruzi infection (reviewed in [6,16,27]. Our observation of increased release of IFN-c/TNF-a by p47 phox2/2 macrophages in vitro infected with T. cruzi (Fig. 3) suggest that NOX2/ROS might control the cytokinopathy via regulating the cytokine gene expression; however, NOX2 deficiency did not inhibit the phagocytes ability to provide inflammatory cytokine milieu for the recruitment and activation of T cells. Indeed flow cytometry analysis showed the CD4 + T cells in p47 phox2/2 mice responded to T. cruzi infection and/or in vitro antigenic stimulus by activation and proliferation to a similar extent as was noted in WT mice (Fig. 6). Others have shown that IFN-c/LPS-treated p47 phox2/2 mice secrete more IL-12 from DCs than similarly treated WT mice, and IFNc/LPS matured p47 phox2/2 DCs biased more ovalbumin-specific CD4 + T cells toward a Th1 phenotype than the WT controls in a ROSdependent manner [28]. It is also suggested that CD4 + T cells from p47 phox deficient mice exhibit augmented IFN-c and diminished IL-4 production and an increased ratio of expression of T-bet (Th1-specific transcription factor) versus GATA-3 (Th2specfic transcription factor), consistent with a Th1 skewing of naïve T cells [29]. Selective inhibition of TCR-induced STAT5 phosphorylation was identified as a potential mechanism for skewed helper CD4 + T cell differentiation in p47 phox2/2 mice [29]. We surmise that p47 phox -dependent NOX2 deficiency enhanced the macrophage maturation and inflammatory cytokine response; and provided help for CD4 + T cell activation in the context to T. cruzi infection in p47 phox2/2 mice. Yet, early splenic response to T. cruzi infection (7 days pi) in p47 phox2/2 mice was dominated by type 2 cytokines evidenced by a .2-fold decline in splenic TNF-a production and ,2-fold increase in IL-10 release when compared to that noted in infected/WT controls, and likely responsible for susceptibility to T. cruzi infection (Fig. 5).
The phenotypic and functional characterization of CD8 + T cells in p47 phox2/2 mice provides clues to the cellular mechanisms contributing to increased susceptibility to T. cruzi infection. It was intriguing to find that splenocytes from p47 phox2/2 mice, as compared to WT controls, contained ,40% lower number of naïve CD8 + T cells. Further, CD8 + T cells in p47 phox2/2 mice exhibited no proliferation and activation evidenced by none-tominimal increase in cell frequency overall or the frequency of IFNc + , CD107 + or IFNc + CD107 + CD8 + T cells in response to T. cruzi infection and secondary in vitro stimulation with antigenic lysate (Fig. 6). Others have shown that T: B cell ratio is lower in p47 phox2/2 mice as compared to the WT mice [30] and the CD8 + T cells from p47 phox2/2 mice express higher levels of proapoptotic Bim and Puma proteins that promoted their removal by apoptosis [31]. Since FOXO3 dephosphorylation (activation) by protein phosphatase 2A (PP2A) is known to contribute to transcriptional control of various apoptosis factors including proapoptotic Bim, blocking the PP2A activity attenuated the FOXO3 activation and Bim transcription and prolonged the survival of CD8 + T lymphocytes in p47 phox2/2 mice [31]. These studies suggest that p47phox deficiency adversely affects the development and survival of naive CD8 + T cells. Additionally, treatment with apocynin that suppresses ROS production by NOX2 directly inhibited the production of proinflammatory cytokines (e.g. TNFa, IFN-c, and IL-2) in anti-CD3/anti-CD28-stimulated CD8 + T cells. It is proposed that apocynin effects were mediated via attenuation of anti-CD3/anti-CD28-induced NF-kB activation in CD8 + T cells [32]. The compromised CD8 + T cell activation was not likely due to inefficient antigen presentation as p47 phox2/2 dendritic cells are shown to be highly efficient in presentation of antigen to B cells in the context of antibody response to Streptococcus and Listeria infection [33]. Others have shown p47 phox2/2 DCs elicit enhanced ovalbumin-specific CD4 + T lymphocytes [28]. Further studies will be required to delineate the complex role of p47 phox in antigen presentation by DCs, CD8 + T lymphocytes survival and ROS-dependent mechanisms involved in NF-kB activation in cell-dependent manner. However, the literature discussed above and our findings allow us to surmise that compromised development of splenic CD8 + T cells and their inability to respond to antigenic stimulus by generation of IFN-c and cytolytic activity contributed to high tissue parasite burden in p47 phox2/2 mice.
It is important to note that the components of NADPH oxidase have diverse effects in heart failure. For example, the survival rate of p47 phox2/2 mice 4-weeks after myocardial infarction (MI) was significantly higher than that of WT mice (72% versus 48%) and the survival benefits were associated with a decline in LV dilatation and dysfunction, cardiomyocyte hypertrophy, apoptosis, and interstitial fibrosis in p47 phox2/2 mice [34]. Others have suggested the loss of p47 phox enhanced the susceptibility to heart failure. Patel et al [35] showed that the expression of N-cadherin and b-catenin was up regulated in p47 phox2/2 mice subjected to biomechanical stress; however, actin filament cytoskeleton was disrupted because these mice lacked the ability to induce p47 phox dependent cortactin-N-cadherin interaction required for adaptive cytoskeletal remodeling. In comparison, gp91 phox2/2 mice exhibited no increase in susceptibility to pressure overload and were equally capable of adaptive cytoskeletal modeling as was noted in controls. In the context of T. cruzi infection, Santiago et al [26] showed that gp91 phox2/2 mice develop increased circulatory collapse and succumbed to infection. Authors proposed that while a lack of superoxide from phagocytes was not detrimental in hosts' ability to control parasites, superoxide regulates nitric oxide concentrations, and enhanced nitric oxide levels in these mice resulted in a critical drop in blood pressure. These studies suggest that targeting NADPH oxidase system as a potential novel therapeutic target to prevent cardiac failure should be considered with caution.
In summary, we present the first evidence that NOX2/ROS of macrophage origin shapes the T cell-mediated adaptive immunity, and its deficiency results in compromised CD8 + T cell response to T. cruzi infection. Our data show that macrophages from p47 phox2/2 mice were not compromised in the phagocytic activity and showed an enhanced iNOS/nitric oxide and pro-inflammatory cytokine levels in response to T. cruzi infection. However, in the event of NOX2 deficiency, generation and activation of CD8 + T cell response was compromised, leading to increased parasite burden, tissue pathogenesis and mortality. We propose that future studies focused on understanding how NOX2/ROS induced innate receptor signaling pathways govern the activation and proliferation of T cell subsets will have the potential to identify specific targets for modulating the adaptive immunity and prevent T. cruzi infection and persistence in Chagas disease.

Ethics statement
All animal experiments were conducted following NIH guidelines for housing and care of laboratory animals and in accordance with The University of Texas Medical Branch at Galveston in accordance with protocols approved by the institution's Institutional Animal Care and Use Committee (protocol number 08-05-029).

Tissue parasite burden
Skeletal muscle and heart tissues (50 mg) were subjected to Proteinase K lysis, and total DNA purified by phenol/chloroform extraction and ethanol precipitation method. Total DNA (100 ng) was used as a template in a PCR reaction for 28 cycles with oligonucleotides specific for T. cruzi 18S rDNA sequence (Forward: 59-TAGTCATATGCTTGTTTC-39, Reverse: 59-GCAACAGCATTAATATACGC-39) [19]. Quantitative estimate of parasite burden was obtained by real-time PCR on an iCycler thermal cycler with SYBR Green Supermix and Tc18S-sepecific oligonucleotides. Fold change was calculated as 2 2DCt , where D Ct represents the Ct (infected sample) -Ct (control) [20]. All data were normalized with host-specific GAPDH.

RT-PCR
Total RNA was isolated by using the RNeasy plus Kit (Qiagen), and analyzed for quality and quantity on a SpectraMax UV microplate reader. After reverse transcription of 2 mg RNA with poly(dT)18, first-strand cDNA was used as a template in a realtime PCR on an iCycler Thermal Cycler with SYBR-Green Supermix (Bio-Rad) and specific oligonucleotides for iNOS (59-GTTTCTGGCAGCAGCGGCTC-39 and 59-GCTCCfTCGCT-CAAGTTCAGC-39) and GAPDH (59-TGG CAA AGT GGA GAT TGT TG-39 and 59-TTC AGC TCT GGG ATG ACC TT-39). The PCR Base Line Subtracted Curve Fit mode was applied for Threshold Cycle (Ct) and mRNA level measured by iCycler iQ Real-Time Detection Software (Bio-Rad). The threshold cycle (Ct) values for target mRNA were normalized to GAPDH mRNA, and the relative expression level of iNOS was calculated with the formula n-fold change = 22DCt, where DCt represents Ct (iNOS) -Ct (GAPDH) [37].

Histology
Tissue sections were fixed in 10% buffered formalin for 24 h, dehydrated in absolute ethanol, cleared in xylene, and embedded in paraffin. Five-micron tissue-sections were stained with hematoxylin and eosin, and evaluated by light microscopy using an Olympus BX-15 microscope equipped with a digital camera. In general, we analyzed each tissue-section for .10microscopic fields (1006 magnification), and examined three different tissue sections/mouse (4 mice/group) to obtain a semiquantitative score of parasitic pseudocysts (foci). Myocarditis (presence of inflammatory cells) was scored as 0 (absent), 1 (focal or mild with #1 foci), 2 (moderate with $2 inflammatory foci), 3 (extensive with generalized coalescing of inflammatory foci or disseminated inflammation), and 4 (diffused inflammation with severe tissue necrosis, interstitial edema, and loss of integrity) [38]. Inflammatory infiltrates was characterized as diffused or focal depending upon how closely the inflammatory cells were associated [39].
Parasite uptake, intracellular replication, and release by macrophages Splenic and BM monocytes from WT and p47 phox2/2 mice were isolated as described [39]. Monocytes were distributed in 24-well plates (10 5 /well) incubated with T. cruzi trypomastigotes (live or heat-inactivated; cell: parasite ratio, 1:3) for 0, 6, 12, and 24 h at 37uC, 5% CO 2 . In some experiments, monocytes were incubated with 5-mg/ml IFN-c for 4 h before exposure to T. cruzi. Cells were submitted to Giemsa staining (Sigma-Aldrich, St. Louis, MO), and parasite uptake and intracellular replication monitored in 200 randomly selected cells by light microscopy.
Trypomastigotes release in supernatants of T. cruzi-infected macrophages after 24 h and 48 h incubation was counted under a light microscope by using a hemacytometer.

ROS production
Isolated primary monocytes from WT and p47 phox2/2 mice were in vitro exposed to T. cruzi for 0-24 h as above. Cells were incubated for 30 min with 5-mM CM-H 2 DCF-DA (detects intracellular ROS, Ex 498 nm /Em 598 nm ) in Hank's Balanced Salt Solution (HBSS), and signal was monitored on a SpectraMax M5 microplate reader (Molecular Devices, Sunnyvale, CA). In some experiments, isolated primary monocytes from WT and p47 phox2/2 mice were in vitro exposed to T. cruzi for 24 h, and then incubated for 30 min with 0.1% nitroblue tetrazolium (NBT). NBT is a yellow water-soluble nitro-substituted aromatic tetrazolium compound that reacts with cellular superoxide ions to form water insoluble blue formazan crystals. Cells were counter-stained with safranin, and the percentage of NBT + cells monitored by monitoring .200 randomly selected cells by light microscopy.
Mice (WT and p47 phox2/2 ) were harvested at day 7-30 pi, and single cell suspension of splenic and BM cells were depleted of red blood cells by hypotonic lysis. Cells were cyto-spinned on glass slides (10 4 cells/slide), equilibrated in Kreb's buffer, and incubated with 5-mM dihydroethidium (DHE, detects intracellular ROS, Ex 518 nm /Em 605 nm ) and images captured by fluorescence microscopy [40]. Cells stained with DAPI (stains all nuclei, blue) were used as controls. Splenocytes (10 6 -cells/well/50 ml) were also incubated with APC-conjugated anti-CD11b antibody (e-Biosciences) and DHE, and macrophage-specific ROS production monitored by flow cytometry. All assays for monitoring the DCF or DHE fluorescence, NBT-based formazan crystal formation were performed in the presence and absence of 0.5 mM apocynin (NADPH oxidase inhibitor) to confirm the source of ROS.
Nitrite level in supernatants of splenocytes, in vitro stimulated in presence or absence of TcL, was measured by Griess reagent assay. Briefly, supernatants were reduced with 0.01 unit/100 ml of nitrate reductase, and incubated for 10 min with 100 ml of 1% sulfanilamide made in 5% phosphoric acid/0.1% N-(1-napthyl) ethylenediamine dihydrochloride (1:1,v/v). Formation of diazonium salt was monitored at 545 nm (standard curve: 2-50 mM sodium nitrite). DAF-FM fluorescence and Griess reagent assays were performed in presence and absence of 5 mmol/ml N(G)nitro-L-arginine methyl ester (L-NAME) that is an inhibitor of nitric oxide synthase [42].

Cytokine release
Isolated primary splenocytes or BM monocytes were in vitro incubated with T. cruzi (live or heat-inactivated) for 48 h. The release of cytokines (IFN-c, TNF-a, IL-4, IL-10) in cell free supernatants was determined by using optEIA TM ELISA kits, according to the manufacturer's specifications (BD Biosciences (San Jose, CA).
For estimating splenic production of cytokines, infected mice were harvested at day 7, 14, 21 and 30 pi, and single cell suspension of splenocytes (10 6 -cells/well/100 ml) incubated with media for 48 h (6TcL). Cytokine release was measured by an ELISA, as above.
Single-cell splenocytes from WT and p47 phox2/2 mice harvested at day 30 pi were suspended in RPMI-5% FBS and distributed in 24-well plates (10 6 cells/well/200 ml). Cells were incubated in presence of Con A (5 mg/ml), or T cruzi trypomastigote lysate (TcL, 25-mg/ml) at 37uC, 5% CO 2 for 48 h. The cell suspensions were utilized to measure the T cell proliferation by MTT assay [43].
To identify the T cell subsets in infected mice, splenocytes were incubated with or without (TcL), and then labeled for 30 min on ice with PE-Cy7-anti-CD3 (binds all T cells), FITC-anti-CD8 and PE-anti-CD4 antibodies (0.5-1 mg/100 ml, e-Biosciences). Following incubation, cells were fixed, washed and re-suspended in 100 ml PBS/2% BSA, and analyzed by flow cytometry [44].
To monitor the intracellular cytokine response, splenocytes were in vitro stimulated as above except that brefeldin A (10-mg/ml; Sigma) was added in the final 6 h to prevent protein secretion. Cells were labeled with PE-anti-CD4 and FITC-anti-CD8 antibodies, fixed, suspended in 100-ml permeabilization buffer (0.1% saponin/1% FBS in PBS) and then utilized for intracellular staining with e-Fluor-anti-IFN-c, Cy5-anti-TNF-a and PerCP-PAanti-Ki67 antibodies (0.5-2-mg/100-ml, e-Biosciences). Splenocytes were also incubated with Alexa-Fluor488-anti-CD107 antibody to determine the cytolytic activity of the activated/proliferating T cell subpopulations. Cells stained with isotype-matched IgGs were used as controls. Samples were visualized on a LSRII Fortessa Cell Analyzer by six-color flow cytometry, acquiring 30-50,000 events in a live lymphocyte gate, and further analysis performed using FlowJo software (ver.10.0.6, Tree-Star, San Carlo, CA) [44].