Importance of CD 8 T cell-mediated immune response during intracellular parasitic infections and its implications for the development of effective vaccines

Obligatory intracellular parasites such as Plasmodium sp, Trypanosoma cruzi, Toxoplasma gondii and Leishmania sp are responsible for the infection of hundreds of millions of individuals every year. These parasites can deliver antigens to the host cell cytoplasm that are presented through MHC class I molecules to protective CD8 T cells. The in vivo priming conditions of specific CD8 T cells during natural infection are largely unknown and remain as an area that has been poorly explored. The antiparasitic mechanisms mediated by CD8 T cells include both interferon-γ -dependent and -independent pathways. The fact that CD8 T cells are potent inhibitors of parasitic development prompted many investigators to explore whether induction of these T cells can be a feasible strategy for the development of effective subunit vaccines against these parasitic diseases. Studies performed on experimental models supported the hypothesis that CD8 T cells induced by recombinant viral vectors or DNA vaccines could serve as the basis for human vaccination. Regimens of immunization consisting of two different vectors (heterologous prime-boost) are much more efficient in terms of expansion of protective CD8 T lymphocytes than immunization with a single vector. The results obtained using experimental models have led to clinical vaccination trials that are currently underway.

cases of the disease per year that result in the death of more than 1 million people, mostly African children under the age of five.Altogether, the other most important intracellular parasites (Plasmodium vivax, Trypanosoma cruzi, Toxoplasma gondii and Leishmania sp) are not responsible for as many infections or kill as many individuals as P. falciparum.Nevertheless, they are responsible for infecting millions of individuals yearly causing enormous social distress and economic losses.
Although the pattern of disease transmission, infection and mortality caused by these obligatory intracellular parasites has changed in the past 50 years, worldwide they had never been a greater problem than they are today.In developing countries, these circumstances are very unlikely to improve in the next ten years.In view of these problems, the development of new strategies for prevention and control of these diseases are in great need.
Mass vaccination against these parasitic infections is not available.Research in this field can be considered important because in the future vaccines may complement other strategies such as chemotherapy, vector control etc., for the prevention and control of these parasitic diseases.Based on this conviction, studies on the mechanisms of host protective immunity and its target antigens have been pursued by immuno-parasitologists for the last 50 years.

HISTORICAL BACKGROUND
Pioneering studies performed in the sixties by Dr. Ruth S. Nussenzweig and collaborators at New York University firmly established that immunization with radiation-attenuated sporozoites could confer an exceptional degree of protection against experimental infection with viable sporozoites of Plasmodium berghei, a species that causes malaria in rodents.In the following decades, this work was extended to several other host species including humans.Individuals immunized with radiationattenuated sporozoites of P. falciparum develop a protective immunity that is sterile, long lived and prevents the development of any symptom of the disease (Hoffman et al. 2002).Because of these exceptional attributes, this type of immunity would be ideal for an effective vaccine against malaria.
In 1987 and 1988, two groups independently reported that the immunity induced in mice by radiation-attenuated sporozoites was dependent on CD8 T lymphocytes (Schofield et al. 1987, Weiss et al. 1988).It was the first evidence that this subpopulation of lymphocytes participates in the immunity against any form of a parasite.In the subsequent years, it was described that CD8 T cells also participate in immunity to infections caused by other intracellular parasites such as Toxoplasma gondii and Trypanosoma cruzi.In the case of Leishmania sp, the participation of CD8 T cells is more controversial; however, very recent studies have implicated CD8 T cells in the pathogenesis and immunity of certain types of cutaneous infection (Belkaid et al. 2002).
The definition of the importance of CD8 T lymphocytes in the immunity against these intracellular protozoan parasites has opened a new direction in the studies of the immunologic mechanisms of resistance to parasitic diseases.From a practical point of view, it may have important implications in the development of vaccines against these parasitic diseases.

Malaria Liver Stages
The importance of CD8 T lymphocytes in the resistance to infection against a challenge with malaria sporozoites was demonstrated in several studies performed during the last 15 years.In initial studies, BALB/c mice immunized with radiation-attenuated sporozoites were depleted of CD8 T cells before a challenge with viable sporozoites of P. yoelii.CD8depleted immunized mice became completely susceptible to infection (Weiss et al. 1988).This study demonstrated that in the absence of CD8 T lymphocytes, other immunological mechanisms induced by radiation-attenuated sporozoites were not capable to promote resistance against the infection.
To further study the importance of CD8 T cells in the immunity to experimental malaria, T cell clones were obtained from spleen cells of BALB/c mice immunized with radiation-attenuated P. yoelii sporozoites.CD8 T cell clones were generated and characterized in vitro in terms of phenotype and specificity.These clones were Thy1+, αβ T-Cell Receptor+, CD4-and CD8+ and they recognized with high affinity a synthetic peptide represented by amino acids SYVPSAEQI present in the C-terminal portion of P. yoelii CS protein (Rodrigues et al. 1991).
In vitro these CD8 T cells presented intense lytic activity and capacity of inducing DNA degradation of H-2kd target cells in the presence of the malaria peptide.In addition, these cells secrete large amounts of IFN-γ , TNF-α and the enzyme BLTesterase when stimulated by the antigen in vitro (Rodrigues et al. 1991(Rodrigues et al. , 1992)).Most important, when adoptively transferred to naïve BALB/c mice, some of these clones were capable to provide specific immunity against the challenge with viable P. yoelii sporozoites (Rodrigues et al. 1991(Rodrigues et al. , 1992)).
To demonstrate that liver stages of the malaria parasite were in fact targets for the immunity mediated by these CD8 T lymphocytes, an assay was developed to quantify the P. yoelii rRNA in RNA isolated from livers of mice infected with P. yoelii.Using this assay it was possible to demonstrate that there was a reduced amount of parasite rRNA in the liver of animals that received cells of a CD8 T cell clone 20 hours after the challenge with sporozoites.In contrast, control animals that did not receive T cells presented large amounts of parasite rRNA in their livers (Rodrigues et al. 1991).
CD8 T cells specific for epitopes present within the P. falciparum liver stage antigens have been commonly observed in humans immunized with radiation-attenuated sporozoites of P. falciparum (Doolan et al. 1997).Whether these cells are as important for human protective immunity as seen in the mouse model of infection cannot be directly addressed experimentally.Nevertheless, it is very likely that they actively participate in the protective immunity observed in human vaccinees.
During natural infection in humans, CD8 T cells specific for epitopes present within different plasmodial antigens have been detected in many individuals with previous contact with P. falciparum (Hill et al. 1992, Aidoo et al. 1995, Plebanski et al. 1997, Gonzalez et al. 2000).Because these individuals do not develop sterile immunity against re-infection, these CD8 T cells do not seem to provide a degree of protective immunity similar to the one observed in individuals vaccinated with radiationattenuated sporozoites of P. falciparum.Whether this is a matter of specificity or quantity of specific T cells it is unknown.Alternatively, mechanisms of parasite evasion of the immune response that are absent after immunization with radiation-attenuated sporozoites of P. falciparum may operate during natural human infection.

Trypanosoma cruzi
The importance of CD8 T cells in the naturally acquired immunity to T. cruzi infection was determined by the use of genetically modified mice that do not express either MHC class I or the CD8 molecule.These two genetically deficient mouse strains are highly susceptible to infection, being unable to control acute parasitemia (Tarleton et al. 1992, Rottenberg et al. 1993).
Subsequent studies generated CD8 T-cell lines specific for an epitope of trypomastigote surface antigen-1 (TSA).These T-cell lines were Thy1+, αβ T-cell receptor+, CD4-and CD8+.In vitro these CD8 T cells presented intense lytic activity against T. cruzi-infected target cells.These cells also secreted large amounts of IFN-γ and TNF-α when stimulated with the antigen in vitro (Wizel et al. 1997).When adoptively transferred to naïve mice, they provided specific immunity against the challenge with infective forms of T. cruzi (Wizel et al. 1997).
T. cruzi is a parasite that persists for long periods of time causing a chronic inflammatory disease.Under certain circumstances T cells exacerbate the immune response causing an increase in tissue pathology or autoimmunity (Soares et al. 2001a).However, this autoimmune response has been associated with CD4 Th1 cells (dos Santos et al. 1992, Soares et al. 2001b).
As mentioned above, in the murine model, the adoptive transfer of large amounts of IFN-γproducing CD8 T cells specific for a parasite epitope significantly reduced infection and promoted survival of mice against a lethal infection with T. cruzi (Wizel et al. 1997).These results indicate that IFN-γ -producing CD8 T cells did not aggravate inflammatory reactions or autoimmunity, but rather promoted elimination of the parasites.
During natural infection in humans, CD8 T cells specific for T. cruzi have been detected in most individuals in the chronic phase of infection (Wizel et al. 1998b).Also, there is an association of an increase in CD8 T cells with the presence of T. cruzi antigens in chronic human chagasic myocarditis (Higuchi et al. 1997).
Although it is difficult to prove definitively that these cells are important in host resistance, it is plausible to believe that they exert an antiparasitic role similar to that seen in the mouse model of infection.

Toxoplasma gondii
Early studies by the group of Kasper and collaborators demonstrated that mice and humans immune to T. gondii develop CD8 T cells that recognize parasite antigen presented by antigen-pulsed or infected target cells.These CD8 lymphocytes proliferate, secrete IFN-γ , lyse the target cells and block the multiplication of parasites in vitro (Khan et al. 1988(Khan et al. , 1990)).
In subsequent years, the concept that CD8 T cells are a very important effector mechanism against this parasitic infection became well established (reviewed by Denkers and Gazzinelli 1998).In mice vaccinated with an attenuated strain of T. gondii (ts-4), the depletion of CD8 T cells was shown to partially abrogate resistance to infection with a highly virulent strain of the parasite.Among the αβ T lymphocytes, CD8 T cells seem to be the major effector mechanism because the depletion of CD4 T cells fails to produce any detectable loss of immunity.Nevertheless, both subpopulations of T lymphocytes cooperate in the immune response because the depletion of CD4 and CD8 T cells renders mice immunized with attenuated T. gondii ts-4 completely susceptible to infection (Gazzinelli et al. 1991).
The importance of CD8 T cells in the resistance to murine toxoplasmosis was confirmed in parallel studies in which CD4-depleted T lymphocytes were adoptively transferred to naïve athymic recipients.These cells (mainly CD8 cells) led to a significant increase in mouse survival after infection and a significant reduction in the formation of brain cysts (Parker et al. 1991).Finally, the importance of CD8 T cells was demonstrated by the adoptive transfer of T cell clones specific for a T. gondii antigen.αβ Tcell receptor+, CD4-and CD8+ T cells specific for the P30 (Sag1) antigen conferred a high degree of protective immunity against a lethal challenge with T. gondii (Khan et al. 1994).
Several lines of evidence implicate T lymphocytes in pathologic changes associated with acute infection in mice.Most evidence pointed at lymphocytes of the CD4 subset and granulocytes during this detrimental inflammatory response that can be the cause of death.Therefore, so far, CD8 T cells seem to be associated with a particularly beneficial immune response by the host, being unable to cause significant pathology (reviewed by Denkers and Gazzinelli 1998).

Leishmania sp.
Years of studies have been unable to firmly establish that CD8 T cells are required for naturally acquired immunity against Leishmania sp.This is not unexpected because it is well established that MHC class II restricted CD4 T cells are extremely dominant during the development of immunity against Leishmania sp.The importance of CD4 T cells was illustrated in studies using mice deficient for the expression of CD8 or class II MHC molecules.After primary subcutaneous infection with L. major, CD8deficient mice controlled the infection for over 1 year (Huber et al. 1998).On the other hand, mice genetically deficient for MHC class II are highly susceptible to infection (Locksley et al. 1993).These results support the notion that in the absence of CD8 T cells, other cell types (mainly class II restricted helper T cells) could provide an efficient control of a primary infection with L. major.
In spite of these earlier observations, very recent studies using low doses of L. major promastigotes inoculated into a dermal site (ear dermis) de-monstrated that CD8 T-cell-deficient mice failed to control parasite growth.Also, these animals had a minor and delayed dermal pathology when compared to wild-type animals.In this model of infection, reconstitution of resistance was obtained when both CD4 and CD8 T cells were adoptively transferred.This study strongly suggests that CD8 T lymphocytes participate in both the pathology and immunity to primary infection with L. major in the skin (Belkaid et al. 2002).
These results are in agreement with a number of previous observations indicating that specific CD8 T cells are primed during natural infection or vaccination in humans.When re-stimulated in vitro with Leishmania antigens, human CD8 T cells secrete IFN-γ , an important mediator of resistance against certain types of infection with Leishmania sp (De Luca et al. 1999, Bottrel et al. 2001, Pompeu et al. 2001).Also, in individuals that develop the mucocutaneous form of the disease, CD8 T lymphocytes are highly cytolytic in vitro against Leishmania-infected macrophages (Brodskyn et al. 1997).
Together with the evidence obtained recently with the murine model of infection, it seems that CD8 T cells participate in both the pathogenesis and immunity to cutaneous and mucocutaneous Leishmaniasis.Nevertheless, the importance of these cells in relation to CD4 T cells is far from clear and remains to be further elucidated.

Malaria Liver Stages
Initial studies using the mouse model of malaria searched for CD8 epitopes in the only protein that had its primary structure known at that time, the CS protein.As a source of CD8 lymphocytes, spleen cells of mice vaccinated with radiation-attenuated sporozoites were used because these animals develop strong protective immunity against the infection.So far, to our knowledge, in mice vaccinated with radiation-attenuated sporozoites, CD8 lymphocytes were specific for a single epitope found in the C-terminal region of the P. berghei or P. yoelii CS protein (Romero et al. 1989, Rodrigues et al. 1991, 1992).This epitope is recognized by CD8 T cells restricted by H-2 Kd molecules.
Figure 1 shows an outline of the P. yoelii CS protein with its different immunodominant epitopes recognized by antibodies (lymphocytes B) or CD4 or CD8 T lymphocytes.The epitopes of P. yoelii CS protein are important because the infectivity of P. yoelii sporozoites is comparable to that of P. falciparum, the most virulent human malaria parasite.Due to these characteristics, P. yoelii infection in mice is considered a valuable model for studies of preclinical development of subunit vaccines (see below).
Similar to the mouse studies, epitopes recognized by human CD8 T cells from individuals vaccinated with radiation-attenuated sporozoites of P. falciparum were initially identified within the CS protein of this human malaria parasite (Malik et al. 1991).In subsequent studies, CD8 epitopes were also found in several other distinct antigens expressed in liver stages of P. falciparum (Doolan et al. 1997, Wizel et al. 1995a,b, Kumar et al. 2001).Currently a total of 35 epitopes have been described in different liver stage proteins of P. falciparum recognized by CD8 T cells from individuals of diverse HLA haplotypes vaccinated with radiation-attenuated sporozoites (Table I).
In addition to CD8 epitopes recognized by T cells from individuals vaccinated with radiationattenuated sporozoites, other epitopes were determined using cells collected from individuals naturally exposed to malaria (Hill et al. 1992, Aidoo et al. 1995, Plebanski et al. 1997).A list of the CD8 epitopes described in the distinct liver stage antigens of P. falciparum is depicted in Table II.As can be noticed, some of these epitopes are common to the ones mapped in individuals vaccinated with radiation-attenuated sporozoites.New epitopes recognized by individuals of the HLA-A2.1 haplotype have been recently described suggesting that the number of CD8 epitopes might be underestimated (Gonzalez et al. 2000).The identification of these CD8 epitopes is the basis for the development of multi-antigen subunit vaccines aimed at the induction of CD8 T cells in humans (see below).

Trypanosoma cruzi
Spleen cells from mice chronically infected with T. cruzi were used to identify a number of epitopes recognized by mouse H-2Kb restricted CD8 T cells.These epitopes were located within the trypomastig- Bras Cienc (2003) 75 (4) ote surface antigen-1 (TSA-1) and amastigote surface protein -1 or 2 (ASP-1 or ASP2, respectively, Wizel et al. 1997, Low et al. 1998).In addition, a CD8 T cell epitope was located in the trans-sialidase of T. cruzi by using spleen cells from mice immunized with plasmid DNA (Rodrigues et al. 1999).Table III contains a list of the murine CD8 epitopes that have been described so far.

Toxoplasma gondii and Leishmania sp.
Few publications have studied the antigens that are target of CD8 T cells specific for T. gondii.Khan et al. (1988), using spleen cells from infected mice, demonstrated the presence of CD8 T cells specific for the purified membrane protein denominated P30 or Surface Antigen 1 (SAG1).This result was confirmed by the immunization with native protein that elicited protective immunity dependent on the activation of this subpopulation of lymphocytes (Khan et al. 1991).CD8 T cell clones specific for the purified P30 antigen (SAG1) were generated and, when adoptively transferred to naive mice, protected them against acute infection with T. gondii (Khan et al. 1994).
In recent studies a putative CD8 T cell epitope recognized by spleen cells of mice immunized with a plasmid DNA containing the SAG1gene was identified.This epitope is represented by amino acids AESKSVII and it is possibly recognized by CD8 T cells restricted by H-2Kk (Nielsen et al. 1999).
In humans with the HLA-A*0201 haplotype, one epitope recognized by CD8 T cells was also identified within the SAG1 antigen.A peptide representing amino acids 289-297 (SPEKHHCTV) was recognized by HLA-A*0201 restricted human CD8 T cells that lysed T. gondii-infected target cells (Yano et al. 1997).
CD8 T cells specific for a single antigen (GP46/M-2) expressed in Leishmania amazonensis were shown to recognize Leishmania-infected macrophages (Kima et al. 1997).However, no subsequent epitope mapping was performed.

CONDITIONS OF IN VIVO PRIMING AND EXPANSION OF PROTECTIVE CD8 T CELLS
The immune response is initiated when antigenic fragments (epitopes) of pathogen-derived proteins are presented by the molecules of the major histocompatibility complex (MHC).While MHC class I products are responsible for the presentation to classic CD8 cytotoxic T cells, MHC class II products will prime classic CD4 helper T cells.
Entry into the class I-restricted pathway of antigen presentation usually occurs through the presence of the antigenic proteins in the cytosol of antigen-presenting cells.In addition, dendritic cells can ingest proteins and have them processed and presented by the MHC class I pathway.The MHC class II pathway is mostly centered on peptides generated in the endocytic pathway.A detailed description of the MHC class I or class II pathways of antigen presentation can be found elsewhere (reviewed by Ploegh 1998).
The conditions of priming and expansion of CD8 T cells specific for any of these parasitic antigens during infection are completely obscure.It is unknown whether priming occurs through infection or antigen-pulsing of antigen-presenting cells, or by cross priming.Cross priming occurs when dendritic cells or macrophages take up cell-associated antigens and present them in the context of MHC class I molecules to classic CD8 T cells.This pathway of antigen presentation is particularly important when the antigen is not expressed in dendritic cells or macrophages.It will be very important in the future to characterize in detail the priming conditions of CD8 T cells specific for the different parasitic antigens.
expansion of specific CD8 T cells was addressed in few studies.CD4 T cells were important for the development of CD8 T cells specific for both T. gondii and malaria parasites (Gazzinelli et al. 1991, Parker et al. 1991).In the case of malaria (P.yoelii), it was possible to elegantly demonstrate that IL-4 secreted by CD4 T cells was crucial for the maturation, but not for the initial proliferation of specific CD8 T cells (Carvalho et al. 2002).

Malaria Liver Stages
To elucidate the role of IFN-γ as an immunological mechanism used by CD8 T cells to eliminate the liver stages of malaria parasites in vivo, studies were performed in mice genetically deficient for the IFN-γ receptor.After a single dose of radiationattenuated sporozoites, these mice do not develop protective immunity against malaria liver stages.In contrast, wild type mice display remarkable immunity against these forms of malaria (Tsuji et al. 1995).This experimental evidence strongly argues for the fact that after a single immunization with radiation-attenuated sporozoites, CD8 T cells use IFN-γ -dependent mechanism(s) to elimi-nate malaria liver stages.Because IFN-γ is a potent inducer of nitric oxide (NO) in mice, a similar approach was used to determine whether NO could mediate elimination of malaria liver stages.Unexpectedly, NO synthase 2 (NOS2)-deficient mice develop protective immunity in a manner similar to wild type animals (M.Tsuji, personal communication).
In spite of the IFN-γ dependence detected after a single immunization with radiation-attenuated sporozoites, after several doses of this immunogen, strong protective immunity can develop in mice genetically deficient for the IFN-γ receptor (Tsuji et al. 1995).The presence of IFN-γ independent mechanisms was confirmed by the observations that CD8 T cell clones, as well as CD8 T cells induced by immunization with a recombinant adenovirus, did not require a functional IFN-γ receptor to eliminate liver stages of P. yoelii (Rodrigues et al. 2000a).The downstream molecules activated by IFN-γ , as well as the nature of the IFN-γ -independent mechanisms remains completely unknown at present.
cise mechanism(s) used by these cells to eliminate T. cruzi has(ve) not been defined.The fact that CD8 T cells secrete IFN-γ may suggest that this mechanism may be relevant to the elimination of intracellular forms of the parasite.IFN-γ is an important mediator of naturally acquired immunity against the infection.Genetically deficient mice that do not express the IFN-γ receptor or Stat4 are unable to control T. cruzi infection (Holscher et al. 1998, Tarleton et al. 2000).However, a direct link between IFN-γ secretion by CD8 T cells and the in vivo antiparasitic activity of these cells has not been provided so far.
The IFN-γ -dependent mechanisms are at least in part mediated by the production of NO.In the absence of IFN-γ receptors, NO is not produced by macrophages, suggesting that IFN-γ is the main inducer of NOS2 activation.The evidence that NO is critical for the elimination of T. cruzi during acute experimental infection was obtained with genetically deficient mice that do not express NOS2.These animals are highly susceptible to infection with different parasite strains (Holscher et al. 1998, Rodrigues et al. 2000b).
In addition to producing IFN-γ , CD8 T cells can lyse host cells infected with T. cruzi, or secrete other potentially active mediators such as TNFα, granulisin or a number of different chemokines (Munoz-Fernandez et al. 1992, Aliberti et al. 1999, Stenger et al. 1999).In fact, CD8 T cells specific for amastigote or trypomastigote antigens are capable of lysing non-phagocytic cells infected with T. cruzi in vitro (Wizel et al. 1997, Low et al. 1998).However, it is unclear whether lysis and DNA degradation of infected target cells by CD8 T cells can be an effective mechanism to restrain T. cruzi infection in vivo.Genetically deficient mice that do not express perforin or granzyme B are not more susceptible to infection than wild type animals (Kumar and Tarleton 1998).These observations argue for the fact that in the absence of perforin-or granzyme B-mediated lysis other mechanism(s) can provide resistance against T. cruzi infection in mice.The elucidation of the anti-parasitic mechanisms medi-ated by T. cruzi-specific CD8 T cells will require further investigation using more complex experimental models.

Toxoplasma gondii
The participation of IFN-γ as an important immunological mechanism used by immune cells to eliminate tachyzoites of T. gondii was determined initially by the treatment of mice with specific neutralizing monoclonal antibodies.Mice injected with a monoclonal antibody to IFN-γ succumb due to acute toxoplasmosis.In contrast, animals that received control antibodies survived infection and developed chronic infection (Suzuki et al. 1988).In subsequent studies, it was shown that treatment with neutralizing monoclonal antibodies had the ability to dramatically change the resistant phenotype of mice vaccinated with an attenuated strain of T. gondii (Gazzinelli et al. 1991).Using genetically modified mice that do not express IFN-γ , it became clear that this cytokine is crucial for the resistance against acute toxoplamosis in mice (Denkers et al. 1997).Because the resistance elicited by vaccination with attenuated parasites had been previously thought to be partially dependent on CD8 T cells, it was plausible to assume that the IFN-γ was secreted in part by these cells.
However, the most compelling evidence showing the importance of IFN-γ produced by CD8 T cells in mediating resistance to murine acute toxoplasmosis was obtained by the adoptive transfer of protective immunity by CD8 T cell clones specific for T. gondii.The treatment with specific neutralizing monoclonal antibodies to mouse IFN-γ completely reversed the immunity mediated by these T cells in naïve mice (Khan et al. 1994).
The IFN-γ secreted by CD4, CD8 and NK cells has the ability to activate in vivo nonhematopoietic or hematopoietic cells to eliminate intracellular T. gondii parasites.Using mouse bone marrow chimeras between wild-type and IFN-γ receptordeficient mice, Yap and Sher (1999) elegantly demonstrated that both cell types have to express the IFN-γ receptor to provide efficient immunity to acute and persistent infection.
Because IFN-γ is a major inducer of NO synthesis by phagocytic cells, the participation of NOS2 was evaluated with the aid of genetically modified mice deficient for the expression of this enzyme.NOS2-deficient mice infected with an avirulent strain of T. gondii die at approximately 21 to 30 days postinfection.Death was associated with large numbers of cysts and tachyzoites suggesting the importance of NO in the control of chronic infection.A similar observation was made in mice genetically deficient for IFN-γ Regulatory Factor-1, the main regulator of IFN-γ -induced NO (reviewed by Denkers and Gazzinelli 1998).
Using mouse bone marrow chimeras between wild type and NOS2-deficient mice, it was possible to determine that NO synthesis was important for the control of tachyzoite replication in infected hematopoietic but not in non-hematopoietic cells (Yap and Sher 1999).
In spite of the importance of reactive nitrogen intermediates in the control of the chronic stage of infection, several lines of evidence strongly indicate that IFN-γ induces other antiparasitic mechanisms independent of NO during murine toxoplasmosis.First, IFN-γ -deficient mice die much earlier (less than 10 days post-infection) than NOS2deficient animals.Also, NOS2-deficient mice vaccinated with the ts-4 strain were protected against a challenge with a virulent T. gondii strain (Khan et al. 1998).This protective immunity had been previously shown to be entirely dependent on IFNγ (Denkers and Gazzinelli 1998).Finally, nonhematopoietic cells deficient in the expression of NOS2 can efficiently control parasite multiplication in vivo (Yap and Sher 1999).
Other possible mechanisms operating in vivo in the control of acute toxoplasmosis can be the release of reactive oxygen intermediates or tryptophan degradation (reviewed by Denkers and Gazzinelli 1998).There is also evidence of a third still unknown mechanism of intracellular T. gondii elimination.Murine astrocytes can be activated by IFNγ to eliminate intracellular forms of T. gondii.The elimination of these parasites does not depend on reactive oxygen intermediates, NO or tryptophan degradation (Halonen and Weiss 2000).Whether or not these mechanisms operate in vivo and which one is the most relevant is still unknown.
As mentioned above, in addition to secreting IFN-γ , T. gondii-specific CD8 effector lymphocytes from immune mice or humans are capable of efficiently lysing parasite-infected target cells.The requirement for IFN-γ in immunity to T. gondii has made it difficult to evaluate the contribution of the lytic activity of CD8 T cells to the resistance to infection.
To determine the importance of perforin as a mediator of the CD8 T cell-mediated immune response, perforin-deficient mice were infected with T. gondii.After vaccination with the ts-4 strain, resistance to acute infection was as good in perforindeficient mice as in wild type animals.In the chronic phase of infection there was a 3-to 4-fold increase in the number of cysts in mice infected with a nonlethal strain of T. gondii.This result indicates that a perforin-mediated mechanism may operate in the elimination of the brain cysts observed during the chronic stages of mouse infection (Denkers et al. 1997).

Malaria Liver Stages
The fact that CD8 T cells are potent inhibitors of malaria liver stages and the characterization of epitopes recognized by these CD8 T lymphocytes prompted many investigators to explore whether induction of these T cells by subunit vaccines could be a feasible strategy for vaccination against malaria.
A number of delivery systems have been used thus far to induce mouse CD8 T lymphocytes specific for the malarial epitope.They include recombinant viruses, bacteria, plasmid DNA, virus-like particles and synthetic peptides.
Recombinant viruses are the single class of im-munogens that have been most systematically studied.Several recombinant viruses were used as vectors to elicit specific CD8 T cells and protective immunity in BALB/c mice against a challenge with sporozoites of P. yoelii.From these studies it became clear that, when used individually, distinct viral vectors expressing a malarial CD8 epitope provided variable degrees of protective immunity against liver stages of P. yoelii (reviewed by Soares and Rodrigues 1998).
When used alone in a single immunization dose, the replication-defective recombinant adenovirus containing the entire P. yoelii CS protein gene was shown to be the most efficient vector capable of reducing the development of liver stages of P. yoelii by ∼ 94%.Immunity was largely mediated by CD8 T lymphocytes, but CD4 T cells also participate in this process.Nevertheless, complete protection against malaria could only be observed in 40% of the mice challenged with P. yoelii sporozoites (Rodrigues et al. 1997).In very recent studies, the immunogenicity and protective efficacy of the recombinant adenoviral vector was improved by a simultaneous injection of the Natural Killer T cell ligand α-galactosylceramide.Under these circumstances, complete protection against infection with P. yoelii sporozoites was achieved in 80% of BALB/c mice (Gonzalez-Aseguinolaza et al. 2002).
Immunization with other recombinant viruses elicited CD8 T cells specific for the malarial epitope; however, the degree of elimination of malaria liver stage was lower than that observed with the adenoviral vector.Table V summarizes the results obtained after a single immunization with a variety of recombinant viral vectors.
Due to the limited degree of protective immunity observed after immunization with a single recombinant virus, studies were performed to improve protective immunity using sequential immunization with two different viral vectors expressing the same CD8 T-cell epitope for priming and boosting.This strategy has been denominated heterologous primeboost immunization regimen.
Earlier studies performed with recombi-  nant viruses expressing the P. yoelii malarial epitope showed that heterologous prime-boost immunization with recombinant influenza virus followed by recombinant vaccinia virus elicited a high degree of protective immunity against malaria infection (Li et al. 1993, Rodrigues et al. 1994).A similar protective immunity could not be achieved with two doses of only recombinant influenza or vaccinia viruses.Table VI summarizes the results obtained using different heterologous prime-boost immunization regimes to elicit immunological elimination of liver stages of P. yoelii and protective immunity against malaria.This unexpected observation was explained by the fact that mice immunized with the combination of these two different carrier viruses had a frequency of epitope-specific CD8 T-cells more than 20 times higher than animals immunized twice with one of these recombinant viruses (Murata et al. 1996).
In figure 2 there is an example of the synergistic effect observed after the heterologous prime-boost immunization with recombinant influenza virus followed by recombinant vaccinia virus expressing the malarial epitope.As it can be noticed there is an increase of more than 40 folds in the number of malaria peptide specific CD8 T cells.
The expansion of epitope-specific CD8 T-cells depends on the sequence of immunization with these two viruses.In mice primed with recombinant vaccinia virus followed by a booster injection with recombinant influenza virus, the frequency of CD8 T-cells was 10 times lower than after priming with influenza virus followed by a vaccinia virus booster (Figure 2 and Murata et al. 1996).No protective immunity could be observed in this case (Li et al. 1993).These findings were recently extended to other malaria CD8 epitopes expressed in the CS protein of P. falciparum, a human malaria parasite (Miyahira et al. 1998).
Based on these observations, several authors have explored the heterologous prime-boost immunization to increase the CD8 immune response.Independently performed studies using two distinct rodent malaria parasites, P. yoelii and P. berghei, observed that the heterologous prime-boost immunization regimes with a plasmid containing the CS gene followed by injection with recombinant vaccinia virus greatly enhanced the specific CD8 T cell response.This protocol of immunization did not only lead to a significant increase in the CD8 immune response, but also provided a significant degree of protection against a challenge with P. berghei or P. yoelii sporozoites (Table VI andSchneider et al. 1998, Sedegah et al. 1998).The increase in frequency of specific CD8 cells was dependent on the order of immunization.Priming with recombinant vaccinia virus followed by a booster injection of DNA did not evoke a significant increase in the frequency of specific CD8 T-cells or protection against malaria infection (Table VI and Schneider et al. 1998, Sedegah et al. 1998).
The immunogenicity of DNA vaccines can be further enhanced by co-immunization with plasmids containing genes of cytokines or co-stimulatory molecules of the immune system.Using P. yoelii infection, it was possible to demonstrate that coadministration of a plasmid containing the gene of murine GM-CSF enhanced both the CD8 response and the protective immunity to malaria in mice immunized with the CS gene and subsequently boosted with a recombinant vaccinia virus (Table VI and Sedegah et al. 2000).
Also using murine malaria as a vaccination model, it has been shown that priming with recom-binant virus-like particles derived from a yeast retrotransposon (TyVLPs) containing the P. berghei or P. yoelii CS CD8 epitope could only induce strong CD8 T cell responses and protective immunity to malaria when followed by a booster injection with the recombinant vaccinia virus expressing the P. berghei or P. yoelii CS protein (Table VI andPlebanski et al. 1998, Oliveira-Ferreira et al. 2000).
Very recently, total inhibition of liver stage development and 100% protection against malaria was obtained in mice primed with the recombinant adenovirus expressing the P. yoelii CS protein, followed by a booster injection with a recombinant vaccinia virus expressing this same antigen.This protocol of immunization led to increased numbers specific CD8 and CD4 T cells, high anti-sporozoite antibody titers, and 100% protection of BALB/c mice when the time between priming and boosting with these two vectors was extended from 2 to 8 weeks.Most importantly, this immunization proto- col elicited long-lasting protective immunity against experimental infection (TableVI and Bruna- Romero et al. 2001).
In addition to malaria, similar observations have been made in several experimental models, thereby confirming the potential of the heterologous prime-boost immunization regimen to enhance the specific immune response against multiple pathogens (reviewed by Zavala et al. 2001).It is also important to point out that this increase in the frequency of specific CD8 T cells described initially in mice was recently confirmed in several types of non-human primates (Kent et al. 1998, Robinson et al. 1999, Hanke et al. 1999, Allen et al. 2000).

Trypanosoma cruzi
The recent identification of epitopes recognized by T. cruzi-specific CD8 T lymphocytes has led to studies aiming at the development of subunit vaccines against experimental Chagas´disease.So far, to our knowledge, the only delivery system used to induce T. cruzi-specific CD8 T lymphocytes was plasmid DNA.
DNA vaccination studies were performed with 3 distinct genes of T. cruzi encoding for major surface antigens of trypomastigote forms of this parasite.In studies conducted in our laboratory we used a gene encoding for the catalytic domain of the enzyme trans-sialidase (TS).Immunization of BALB/c mice with the TS gene elicited specific antibodies and promoted T-cell activation (Costa et al. 1998).The T-cell immune response was later characterized as being mediated mainly by CD4 Th1 and CD8 TC1 cells, which secreted a large amount of IFN-γ , but not IL-4 or IL-10 (Rodrigues et al. 1999).Upon challenge with infective T. cruzi trypomastigotes, immunized mice showed a significant reduction in parasitemia and survived acute lethal infection (Costa et al. 1998).
Parallel studies were performed on BALB/c and C57BL/6 mice immunized with plasmids containing the TSA-1 gene.DNA vaccination was shown to generate antibodies, as well as CD8 T cells specific for parasitic epitopes.After challenge with T. cruzi trypomastigotes, a significant reduction in parasitemia was observed in C57Bl/6 mice, but not in BALB/c mice.Nevertheless, in both cases the DNA-vaccinated mice were more likely to survive acute infection than mice injected with the plasmid vector alone (Wizel et al. 1998a).
DNA-based immunization was also performed with a plasmid containing the gene of T. cruzi complement regulatory protein.This study demonstrated that both recombinant protein and DNA immunization could elicit activation of specific CD4 T cells.However, only DNA immunization could confer protective immunity on BALB/c mice.Based on that comparison, the authors proposed that CD8 T cell activation occurring during DNA immunization could be important for the protective immunity observed (Sepulveda et al. 2000).
Most recently, studies were carried out with genes encoding antigens expressed by the intracellular replicative forms of T. cruzi.They demonstrated that the genes encoding ASP-1, ASP-2 and LYT1 could provide a high degree of protective immunity against experimental infection (Garg andTarleton 2002 andFralish andTarleton 2003).
From mice immunized with the TS gene, we isolated CD4 Th1 and CD8 Tc1 clones which displayed remarkable antiparasitic activities in vitro (Rodrigues et al. 1999(Rodrigues et al. , 2000b)).Because DNA immunization with the TS gene could elicit distinct immunological mechanisms, we considered that a detailed comparison of the immunogenicity of plasmids containing either the entire TS gene or DNA sequences encoding its immunogenic portions would be important.For that purpose, the levels of T cell response and protective immunity against experimental infection were evaluated in mice immunized with plasmids containing DNA sequences encoding either the TS CD4 or CD8 or both T-cell epitopes.
Distinct plasmids containing DNA sequences encoding both TS CD4 and CD8 T-cell epitopes could provide a degree of immunity sufficient to reduce the parasitemia and mortality of DNAimmunized animals caused by a lethal challenge with T. cruzi.On the other hand, plasmids expressing either CD4 or CD8 T-cell epitopes of TS were unable to provide a similar degree of protective immunity against infection (Table VII).These observations suggest that activation of CD4 Th1 and CD8Tc1 T cells are important for host immunoprotection induced by immunization with a DNA subunit vaccine (Fujimura et al. 2001).

Toxoplasma gondii and Leishmania sp.
The fact that no clear epitope recognized by CD8 T cells has been characterized in T. gondii or Leishmania sp.precludes testing of subunit vaccines in experimental models.Several reports using plasmid immunization have been able to demonstrate heterogeneous degrees of protective immunity against T. gondii infection in immunized mice.These studies used genes encoding the antigens SAG1, GRA1, GRA7 and ROP2.
Although DNA immunization can elicit some degree of protective immunity, the activation of spe-cific CD8 T cells has been poorly or not at all characterized (Nielsen et al. 1999, Angus et al. 2000, Vercammen et al. 2000, Desolme et al. 2000).Whether subunit vaccines may induce protective CD8 T cells specific for these antigen remains to be further characterized in the future.
In Leishmania major, s.c.immunization with a plasmid containing the gene coding L. major LACK elicited protective immunity against infection in a highly susceptible mouse strain (BALB/c).The protective immunity detected was mostly mediated by specific CD4 Th1 cells, although protective immunity was also partly dependent on CD8 T lymphocytes (Gurunathan et al. 1997).A further evaluation of the role of CD8 T cells found that they may function as helpers for the development of effector CD4 T cells.No evidence so far has indicated that the participation of CD8 T cells in protective immunity against L. major is related to their effector function (Gurunathan et al. 2000).Also, a CD8 T-cell epitope is yet to be identified within the LACK antigen.

HUMAN CLINICAL TRIALS USING SUBUNIT VACCINES
The successful results of preclinical vaccination studies performed on rodent malarias have encouraged human phase I/IIa vaccination trials aimed at inducing CD8 T cells specific for malaria liver stage antigens.Phase I/IIa trials tested the safety, tolerability and immunogenicity of plasmid DNA, recombinant viruses and synthetic peptides.Table VIII presents a summary of the different human trials.
The human vaccination trials performed so have provided evidence that DNA vaccines can be successfully used to elicit a specific CD8 T cells response (Doolan and Hoffman 2001).In the first study, groups of 5 healthy individuals were immunized with different doses of a plasmid containing the P. falciparum CS protein gene produced under Good Manufacturing Practice conditions.Individuals received 3 i.m. doses of plasmid, each consisting of 20 to 2, 500µg of DNA.Specific cytotoxic T lymphocytes (CTL) responses were detected in 2 of 5 individuals immunized with 3 doses of ei-  ther 20 or 100µg of DNA.The higher immunizing doses (500 and 2, 500µg of DNA) generated a CTL response in 3 of 5 and 4 of 5 individuals, respectively.MHC-restricted CD8 T-cells mediated CTL activity.CTL responses were significantly higher in individuals who received either 500 or 2, 500µg of DNA.However, no difference was observed in the magnitude of the CTL response after 3 doses in individuals immunized with these two doses of DNA.In general, the CTL responses were stronger in DNAimmunized vaccinees compared to individuals naturally exposed to P. falciparum infection (Wang et al. 1998).
A second human trial was designed to test immune responses induced by different routes of immunization.Individuals received three doses of 2.5 mg of DNA i.m. by needle injection or i.m. by needle-less BiojectorTM 2000, or i.m. andi.d. by BiojectorTM 2000 (70% of the total dose i.m. and30% i.d.).Thirteen of the 14 volunteers had antigen-specific CD8 T cells that produced IFN-γ responses against 20-100% of peptides containing defined class I binding epitopes.The data also suggested that Biojector injection i.m. was the most effective route of immunization (Wang et al. 2001).
Current studies are being performed to improve cell-mediated immune responses by the coadministration of a plasmid containing the human GM-CSF gene.Also, individuals are being immunized with several plasmids containing different malarial genes.Results of these new trials are expected to be published soon (reviewed by Doolan and Hoffman 2001).
Phase I/IIa trials have also been performed with NYVAC-Pf7, a highly attenuated vaccinia virus containing 7 P. falciparum genes inserted into its genome.The vaccine was considered to be safe and well tolerated.Seven of the 15 individuals that received three doses of 10E8 virus developed a detectable CTL response to at least one of the three liver stage antigens (CS or LSA-1 or TRAP antigen, Ockenhouse et al. 1998).Nevertheless, it is important to notice that the individuals immunized were not HLA matched.Therefore, it is difficult to determine whether the absence of the immune response reflected variable immunogenicity of the vector or lack of CD8 T-cell epitopes.
In a third Phase I/IIa vaccination trial, individuals were immunized with a synthetic peptide representing the C-terminal region of the P. falciparum CS protein (amino acids 282-383).Individuals were immunized i.m. with the peptide in the presence of the adjuvants Montanide 720 or Alum.The CD8 T cell immune response was evaluated in HLA-A*0201 individuals.Upon vaccination, 6 of 8 individuals developed a CD8 T lymphocyte response, as measured by IFN-γ secretion.CD8 T cells were mainly directed toward epitopes 327-335 (5 of 8 volunteers) and 299-308 (2 of 8 volunteers, Lopez et al. 2001).
In a very recent publication, a heterologous prime-boost vaccination regime of DNA either in-tramuscularly or epidermally, followed by intradermal recombinant modified vaccinia virus Ankara (MVA), induced high frequencies of interferon IFNγ -secreting, antigen-specific T-cell responses in humans to a pre-erythrocytic malaria antigen, thrombospondin-related adhesion protein (TRAP, McConkey et al. 2003).T-cell responses induced by the DNA vaccine priming-recombinant MVA boosting produced partial protection manifested as delayed parasitemia after sporozoite challenge with Plasmodium falciparum.
Taken together, these human vaccination trials open new avenues for new vaccination strategies aimed at inducing malaria-specific CD8 T cells.

EVASION OF CD8 T-CELL IMMUNE RESPONSE
Viruses are the best known class of pathogens that use a variety of mechanisms to evade the immune recognition by MHC class I restricted CD8 T cells.Some of these strategies may function at the level of the antigen presentation by infected cells such as: i) down regulation of MHC class I expression, ii) interference with cytosolic proteolysis, iii) inhibition of peptide transport to the endoplasmic reticulum (ER), iv) retention and destruction of MHC class I molecules in the ER, v) presence of mutant epitopes that do not bind to MHC I class molecules, etc.Under these circumstances, infected cells have a limited amount of peptide/MHC-class I complexes on their surface that makes them extremely difficult to be recognized by specific CD8 T cells.A detailed review and examples of these evasion mechanisms can be found in Ploegh 1998.
Alternatively, escape mechanisms can interfere with the recognition by MHC I class restricted CD8 T cells.For example, mutant viruses can be selected with amino acid variations that retain binding to MHC class I molecules but result in reduced recognition by T cells or generate antagonistic peptides that inhibit activation of specific T cells by the MHC-peptide complex (reviewed by Ploegh 1998).
Most recently, viral proteins have been described that interfere with the effector mechanisms  mediated by T cells.For example, poxviruses encode several soluble cytokine receptors that bind to and block the activity of IFN-α, β and γ TNF, IL-1 etc. Also, these viruses express a protein termed crmA that functionally inhibits Asp-specific cysteine proteases (caspases), blocking apoptosis induced by CTL, TNF, or Fas (reviewed by Ploegh 1998).
It is unknown whether protozoan parasites have evolved in such a way as to use all these mechanisms of immune evasion.So far very few examples of mechanisms of evasion have been confirmed in the case of the CD8 T cell-mediated immune response to parasites.This lack of information most likely reflects the fact that the CD8 T cell-mediated immune response to parasites only recently has attracted the attention of most investigators.
Amino acid variations in CD8 epitopes within the CS protein of P. falciparum malaria parasites have been described.These amino acid substitutions affected binding to HLA I class molecules, generated peptides that failed to stimulate specific CD8 T cells or behaved as antagonistic peptides (Udhayakumar et al. 1997, Gilbert et al. 1998, Bonelo et al. 2000).
The best described example that a variation in a malaria parasite epitope may interfere with binding to HLA I class molecules or fail to stimulate specific naturally induced human CD8 T cells was described by Bonelo et al. (2000).As shown in Table IX, nine variant peptides based on the sequences of different strains of P. falciparum were tested for the ability to bind to HLA-A*0201 and to be recognized by a human CD8 T cell clone obtained from an individual naturally exposed to the parasite.Variant peptides from strains 7G8 and LE5 of P. falciparum displayed much lower binding affinities for HLA-A*0201 and were not recognized by the human CD8 T cell clone.Four other variant peptides still retain the ability to bind to HLA-A*0201.However, because of their amino acid substitutions, they also failed to be recognized by the human CD8 T cell clone.Finally, it is noteworthy that this T cell clone could still recognize fairly well two other variant epitopes represented in several strains of P. falciparum.A total of 7 of the 18 different strains could be recognized by a single T cell clone.
This experiment illustrates that there might be naturally occurring variant CD8 epitopes that interfere with recognition of MHC class I restricted T cells.The identification of these variant epitopes may increase our understanding of the importance of certain effector immune mechanisms, and ways in which vaccines can be best designed.

PERSPECTIVES
The knowledge that CD8 T cells are essential for efficient immunity to intracellular parasites has raised important questions with basic and applied implications.Several gaps in basic knowledge are expected to be filled in the years to come.Some of the most relevant questions are: i) How are specific CD8 T cells induced and maintained in vivo?ii) What are the main effector mechanisms operating in vivo against the different parasites?Are there specific or general mechanisms of immune evasion used by each of these parasites?These questions in humans are even more relevant and still more complex to study due to the obvious limitations.
From the applied point of view, it is now clear that the heterologous prime-boost regimes of immunization can be highly efficient in terms of the expansion of protective CD8 T lymphocytes.Nevertheless, the reason why this type of immunization is more efficient than the others is still largely unknown and should be addressed in the future.To our knowledge, few hypotheses have been formulated thus far to explain this event and none was tested experimentally with success.
Finally, human vaccination trials aimed at inducing CD8 T cell-mediated immunity have been initiated and should be further intensified to provide important information about the feasibility of this strategy to develop a new generation of vaccines against parasitic infections.

TABLE I Epitopes on P. falciparum liver stage proteins recognized by CD8 T cells from volunteers vaccinated with radiation-attenuated sporozoites of
P. falciparum.

TABLE V Summary of experiments using a single recombinant virus to elicit immunological elimination of liver stages of Plasmodium
yoelii.

TABLE VII Requirement of DNA coding sequences for CD4 Th1 and CD8 Tc1 epitopes during DNA immunization against experimental
T. cruzi infection.

TABLE IX Polymorphism of a human CD8 T cell epitope restricted by HLA-A*0201.
Bonelo et al. 2000 et al. 2000.a = Relative activity in relation to the binding of MA influenza virus peptide.b = Percent maximal lysis induced by a human CD8 T cell clone specific for peptide YLNKIQNSL.