In the present study, we investigated the cellular and molecular mechanisms that may account for the suboptimal immunity of specific CD8+ T cells generated during infection with T. cruzi in mice. Accordingly, our strategy was to compare the suboptimal immune response elicited by parasite infection with the response elicited by a highly effective adenoviral vaccine expressing the parasite's immunodominant epitopes. Our main hypothesis was that CD8+ T cells induced by T. cruzi infection had one or more properties not present in cells expanded by adenoviral vaccines, and this caused a suboptimal immune response.
A detailed comparison of the specific CD8+
T cells elicited by infection or vaccination revealed that, in general, the cells overlapped in their functional or phenotypic characteristics. Epitope-specific CD8+
T cells were highly cytotoxic in vivo
and secreted IFN-γ, a cytokine that is critical for control of T. cruzi
. Further, these cells mobilize CD107a to their surface after in vitro
stimulation with the cognate peptide, indicating that a large proportion of these cells are prepared for granule exocytosis. This observation is supported by cytotoxicity experiments showing that the specific T cells are highly cytotoxic in vivo
. Analysis of the expression of IFN-γ and TNF-α by ICS also revealed that after experimental infection or immunization with AdASP-2 vaccine, the majority of specific T cells were a population of multifunctional T cells that mobilized surface CD107a and expressed both cytokines. Other cytokines evaluated, such as IL-2 or IL-10, were not detected by ICS (data not shown, 
Similarly to the functional aspects described above, the surface phenotype of the specific CD8+ T cells was characteristic of Teffector cells. The fact that effector functions of CD8+ T cells elicited by T. cruzi infection overlap with those elicited by the AdASP-2 vaccine emphasizes that these cell functions may be important for control of the pathogen multiplication and, in the case of the most mouse strains, the acute-phase pathology.
A clear difference we observed between the immune responses elicited by T. cruzi
infection or AdASP-2 vaccine was in the kinetics of the immune response. The generation of cytotoxicity and IFN-γ-producing cells occurred faster in mice immunized with the AdASP-2 vaccine than in those infected with T. cruzi
a fact previously described by us and others 
. The precise reasons for the delay in the immune response after T. cruzi
challenge are debatable and certainly may be important for the parasite to establish a successful infection. In earlier studies, the timing for the development of the T-cell immune response was critically controlled by the parasite load 
. Because parasites contain antigen and, possibly, yet to be identified adjuvant molecules, the effect of the parasite load could be associated with either one individually, or both. Padilla et al. 
proposed that the delay was not due to the absence of parasite antigen but to the absence of sufficient parasite TLR agonists considered as T. cruzi
adjuvant molecules. We question this interpretation because, recently, we described that the development of the CD8+
T cell immune response during experimental T. cruzi
infection progressed well in genetically deficient mice that do not express TLR-2, TLR-4, TLR-9, or even the adaptor molecule MyD88 
. Further, even when the host immune response was accelerated by the addition of TLR agonists, no changes in the acute-phase pathology were described 
. Therefore, other intrinsic characteristics of the CD8+
T cells may explain the difference in the timing of specific CD8+
In accord with this possibility, an important observation from our study was that epitope-specific CD8+
T cells expanded during infection expressed higher levels of CD95 than naive CD8+
cells or epitope-specific cells from mice immunized with adenoviral vaccines. This phenomenon occurred in different mouse strains (A/Sn, C57BL/6 and BALB/c) infected with 2 distinct parasite isolates (Y and Brazil). This aspect is important because in many cases there may be differences when using different parasites isolates in terms of mechanism of immunity activated during infection 
. This aberrant pattern of upregulated expression was the only one consistently observed when compared to any other surface adhesion/homing/activation receptor that we analyzed. In addition to the specific CD8+
T cells expressing higher levels of CD95, CD95hi
cells stained positively for annexin V, a marker for apoptotic cells. Compatible with an increased apoptosis rate, we found that specific CD8+
T cells accumulated less in vivo
. These smaller numbers cannot be explained by lower proliferation rates. In vivo
incorporation of BrdU indicated similar proliferation rates of cells from infected and infected AdASP-2-vaccinated mice.
CD95 functions during the immune response are multiple and still the subject of intense studies 
. The presence of low levels of CD95L on the surface of antigen-presenting cells (APCs) may serve as a co-stimulatory signal for naive CD95-expressing T cells 
. On the other hand, higher levels of the ligand provide a negative signal for these same cells by blocking their activation through the CD3 complex 
. The presence of CD95L on the surface of pathogen-infected APCs has been suggested as a possible mechanism to downmodulate the immune response 
. We could not find evidence of this theory as we could not detect CD95L on the surface of dendritic cells infected with T. cruzi
(J. Ersching and MMR, unpublished results).
Because CD95 is an important initiator of the intrinsic pathway of apoptosis in T lymphocytes, later it cooperates with Bim in retraction of the T-cell immune response 
. The observation that specific CD8+
T cells undergo modulation of CD95 expression after infection was described in specific CD8+
T cells following mouse infection with LCMV or in individuals infected with HIV. In the case of LCMV infection, lower levels of CD95 expression on specific CD8+
T cells located in the peripheral organs is associated with an increased resistance to activation-induced cell death in vitro
. In the case of human HIV infection, higher expression of CD95 leads to increased susceptibility to CD95/CD95L-mediated apoptosis and may compromise the immune response of specific CD8+
T cells 
Upregulation of CD95 expression was also described in splenic T cells during acute infection with certain T. cruzi
. Most relevant biologically was that in this model of infection, in vivo
injection of antibody to CD95L, which blocks the interaction with CD95, reduced apoptosis, improved type-1 immune responses, and reduced the infection severity as estimated by parasitemia 
. Similar treatment with anti-CD95L or anti-PD1 during infection of the highly susceptible A/Sn mice with parasites of the Y strain of T. cruzi
did not reduce the parasitemia or the mortality, indicating that these treatments alone cannot substitute for immunization with AdASP-2 vaccine (AFA and MMR, unpublished results). The reasons for the failure observed with the anti-CD95L treatment are not clear. However the fact that CD95 is expressed in a number of different cell types which play a diverse role during the immune response may account for the difficulty to block solely the interaction of CD95 expressed on the surface of specific-CD8+
The precise reason(s) why CD95 is differentially modulated after infection with T. cruzi or adenoviral vaccination is unknown at present. Herein, we tested whether 2 important mediators of inflammation during T. cruzi experimental infection (MyD88, IL-12 and IFN type I) could account for this pattern of CD95 expression on specific CD8+ T cells. Our results demonstrated that during infection the CD95 upregulation process was independent of the expression of each one of these mediators individually. Because these molecules are involved directly or indirectly with most cytokines that are largely produced during T. cruzi infection, these results may indicate that the signal for CD95 up-regulation may be given by direct interaction between APC, CD8 and CD4.
To date and to our knowledge, generally the cellular and molecular mechanisms modulating CD95 expression on lymphocytes have been poorly explored. Cytokines IL-6 and type I IFN were described as factors that either downregulate or upregulate CD95 expression, respectively 
. The environment of T. cruzi
infection is relatively poor in IL-6 but rich in type 1 IFN 
. It is possible that this imbalance favors the expression of CD95. On the other hand, the in vivo
environment following immunization with human adenovirus (including recombinant type 5) is rich in type 1 IFN and IL-6. This balance may block the upregulation of CD95 
A very recent observation also correlated a defective CD8+
T cell-mediated anti-tumor immune response with the aberrant expression of CD95 and PD1 by these cells. This pattern of CD95 expression by CD8+
T cells was reproduced by stimulating transgenic CD8+
T cells in vitro
with immature dendritic cells pulsed with the cognate peptide 
. Whether immature or T. cruzi
-infected dendritic cells also have the ability to upregulate the CD95 expression of CD8+
T cells in vivo
is an interesting possibility that should be evaluated. Finally, cyclon, a newly identified cytokine-inducible protein produced by T cells upon TCR activation, has been shown to exert a key role in the process of CD95 expression 
. Whether cyclon expression following TCR activation plays a role in our system also remains to be determined.
The fact that AdASP-2 vaccine fails to upregulate CD95 and changes the program of CD8+ T cells expanded after an infectious challenge is unexpected but may be desirable for an efficient immune response. The lower levels of CD95 expressed would render these cells resistant to parasite-induced apoptosis. In fact, we observed that adenoviral-induced CD8+ T cells of multiple specificities were not affected by ongoing T. cruzi infection (JRV and MMR, unpublished observations). These observations disfavor the hypothesis that cytokines generated during infection can nonspecifically trigger upregulation of the CD95 molecule on activated CD8+ T lymphocytes.
Based on our observations, we propose a model depicted in . The initial contact of specific CD8+ T cells with APC loaded with T. cruzi antigen will lead to a molecular program characterized by a higher expression of CD95 and a lower viability (Gr. 2, T. cruzi infected mice). The low viability of specific CD8+ T cells will preclude optimal immune response. Host pathology is developed that may lead to death or chronic infection.
Proposed model for the differences in activation of specific CD8+ T cells.
As oppose, initial contact with APC loaded with AdASP-2 antigens will lead to different molecular program of priming and expansion of specific CD8+ T cells (Gr. 3, AdASP-2 immunized mice). Upon a second contact, primed CD95Low specific CD8+ T cells will proliferate intensively and display a significantly higher viability (Gr. 4, AdASP-2 immunized and infected mice). This accelerated proliferation coupled to a high viability would lean the equilibrium towards an efficient host immune response and parasite elimination.
The absence of an aberrant CD95 expression not only facilitates the development of a protective immune response by slowing the rate of apoptosis but it also reduces the formation of apoptotic bodies. Apoptotic bodies are described as powerful immune modulators during T. cruzi
infection. Binding of apoptotic lymphocytes to αVβ3 expressed by macrophages leads to PGE2 and TGF-β production by macrophages, followed by the induction of ornithine decarboxylase and the synthesis of putrescine, which function as growth factors for intracellular forms of T. cruzi
The reduction of CD95 expression on the surface of specific CD8+
T cells expanded following pathogen challenge may be important not only in our infection model but also in vaccination against other infectious diseases. As mentioned above, during HIV/SIV infection, CD8+
T cells upregulate CD95 expression 
. Blocking the interaction of CD95/CD95L in vivo
by treatment with anti-CD95L antibody preserved memory lymphocytes and cell-mediated immunity in SIV-infected primates 
. Further, by down-modulating CD95 expression, adenoviral vaccines may reduce the formation of apoptotic bodies. As in the case of T. cruzi
infection, these apoptotic bodies may fuel viral growth 
The possibility that vaccination may selectively alter specific CD8+
T cell phenotype and change the expression of surface receptors is only now beginning to be explored. Recently, it was described in humans that during immune responses to melanoma tumors, antigen-specific CD8+
T cells maintained high expression levels of the inhibitory receptor B and T lymphocyte attenuator (BTLA). Multiple vaccination doses with peptides representing CD8 epitopes in the presence, but not in the absence, of the TLR-9 agonist CpG ODN led to progressive downregulation of BTLA in vivo
, increased resistance to BTLA-mediated inhibition, and improved T cell function 
. This study provides important evidence that vaccination can modify antigen-specific CD8+
T cells, thereby qualitatively improving their function. Further, by using this same adjuvant (a TLR-9 agonist), Muraoka et al.
modified a defective CD8+
T cell-mediated anti-tumor immune response to a productive one by counteracting the aberrant expression of CD95 and PD1 on CD8+
We consider that these results may facilitate understanding of the host–parasite relationship and present some new interpretations of the functions of T-cell vaccines. Recombinant replication-defective human and simian adenoviral vectors have proven successful against a variety of experimental infections, such as SIV, TB, malaria, toxoplasmosis, Marburg virus infection, and Ebola 
. Nonetheless, the precise reason for their success is still unknown. By understanding the potential of these vectors to elicit an immune response and simultaneously to modulate acquired immune responses in the host, it might be possible to design simpler and more effective vaccine formulations.