Toxoplasma persists in the face of a functional immune system. This success critically depends on the ability of parasites to activate a strong adaptive immune response during acute infection with tachyzoites that eliminates most of the parasites and to undergo stage conversion to bradyzoites that encyst and persist predominantly in the brain. A dramatic change in antigenic composition occurs during stage conversion, such that tachyzoites and bradyzoites express closely related but antigenically distinct sets of surface Ags belonging to the surface Ag 1 (SAG1)-related sequence (SRS) family. To test the contribution of this antigenic switch to parasite persistence, we engineered parasites to constitutively express the normally bradyzoite-specific SRS9 (SRS9c) mutants and tachyzoite-specific SAG1 (SAG1c) mutants. SRS9c but not wild-type parasites elicited a SRS9-specific immune response marked by IFN-γ production, suggesting that stage-specificity of SRS Ags determines their immunogenicity in infection. The induction of a SRS9-specific immune response correlated with a continual decrease in the number of SRS9c cysts persisting in the brain. In contrast, SAG1c mutants produced reduced brain cyst loads early in chronic infection, but these substantially increased over time accompanying a hyperproduction of IFN-γ, TNF-α, and IL-10, and severe encephalitis. We conclude that stage-specific expression of SRS Ags is among the key mechanisms by which optimal parasite persistency is established and maintained.

Toxoplasma gondii is an obligate intracellular, protozoan parasite highly prevalent in warm-blooded vertebrates (1). One route of infection is consumption of meat harboring cysts that contain bradyzoites (BZs).3 Once in the gut, BZs rapidly invade and convert to tachyzoites (TZs) that proliferate and disseminate throughout the body (2). This acute infection with TZs activates strong, long-lasting Ab and T cell responses that eliminate most of the parasites (3). Some parasites that survive undergo stage conversion to BZs, a process that may be induced by the immune response itself (4). BZs then encyst, establish a chronic infection primarily in the brain and harmlessly persist for the life of an immunocompetent host.

A key feature of Toxoplasma persistence is the requirement of a functional immune system to control the acute infection with TZs. In an immunodeficient host, uncontrolled TZ growth causes tissue destruction, and persistent infection is not established (5). When a chronically infected host later becomes immunodeficient, BZs reactivate to TZs causing severe neurological diseases (6). Another characteristic of Toxoplasma persistence is that the parasite persists in the face of a long-lasting anti-Toxoplasma immune response. Toxoplasma succeeds in this daunting task by using a variety of strategies applied by microbial agents, which allow them to persist until transmission to a new host (by carnivorism for Toxoplasma) can be accomplished. Such strategies include: modification of the intracellular environment within host cells to favor parasite survival (7); manipulation of the host immune response by producing immunomodulatory molecules (8, 9); establishment of chronic infection predominantly in the brain, where immune surveillance is controlled differently than in other peripheral tissues (10); and, the fact that BZs are metabolically quiescent and relatively nonproliferative during chronic infection (11).

Another immune-evasion mechanism commonly used in microbial persistence is the variation of antigenic composition that eliminates epitopes that would otherwise be targeted by protective immune responses (12, 13, 14). Toxoplasma stage conversion involves a differential expression of numerous genes in a stage-specific manner (4, 15), among which are the members of the surface Ag (SAG)1-related sequence (SRS) superfamily that encode GPI-anchored surface proteins (>160 putative genes) (16). The prototypic SAG1 is the most abundant TZ SRS Ag (17). SRS Ags are structurally related (18), yet antigenically distinct, sharing 25–35% overall amino acid sequence identity and up to 50–90% identity within a given subfamily (16). The structure of SRS Ags suggests their possible role as cell adhesion molecules (18). In addition, extensive polymorphisms found among the three canonical lines of Toxoplasma imply that SRS Ags are under selective pressure from the immune response (19). A distinctive feature of the SRS antigenic switch is that TZs and BZs express largely nonoverlapping sets of SRS Ags, but a single set of multiple SRS Ags is simultaneously expressed by most if not all of the parasite population in a given developmental stage (16).

Because stage conversion and a dramatic change of the parasite surface occur with the emergence of the adaptive immune response, it has been speculated that stage-specific expression of SRS Ags has evolved as an immune-evasion mechanism that allows Toxoplasma to persist. Surface proteins are excellent targets of neutralizing Abs and cell-mediated immune responses. In fact, SAG1 and SAG2A dominate the humoral response in the acute infection (20, 21), and SAG1 also activates T cells producing IFN-γ (22, 23), the major cytokine mediating resistance against Toxoplasma (24). However, an immune response to TZ-specific Ags such as SAG1 and SAG2A is expected to be ineffective against BZs that do not express these molecules. Although parasite Ags shared by both TZs and BZs can also activate an immune response (22, 25), this is clearly insufficient to eliminate BZs. Notably, immune responses to BZ-specific Ags, such as SAG2C/D (a member of the SRS family) and SAG4, LDH2, ENO1, and p-ATPase were undetectable in chronically infected humans (26). In this study (26), BAG1 and MAG1 were thought to be the only BZ-specific molecules immunogenic in infection. Recent data, however, demonstrated that MAG1 mRNA (15) and protein (27) are expressed in both the TZ and BZ stages, making it likely that a MAG1-specific response may have been elicited during acute infection with TZs. Because BAG1 shares a significant homology with other small heat shock proteins across species (28), cross-reactivity of an immune response to other microorganisms remains a distinct possibility. Thus, overall, BZ-specific Ags appear to be poorly or not at all immunogenic in infection and this condition may be one of the mechanisms by which BZs escape immune surveillance.

To explore the role of stage-specific expression of SRS Ags in Toxoplasma persistence, we chose to focus on SAG1 and SRS9, two of the most abundant SRS Ags specific to TZs and BZs (15), respectively. A key aspect of our work is the use of mutant parasites expressing SAG1 and SRS9 in a constitutive manner in both the TZ and BZ stages, which allowed us to study the interaction between the host immune response and SRS Ags in relation to their stage-specificity. We show that an immune response to SRS9, as well as to other BZ-specific SRS Ags, is lacking in natural infection and that stage-specific expression of SRS Ags has important consequences regarding their immunogenicity in infection and the ability of parasites to achieve optimal persistence in infected hosts.

Prugniaud (Pru) strain parasites were maintained as TZs by passage in human foreskin fibroblast (HFF) monolayers cultured in DMEM supplemented with 10% FCS, penicillin (100 U/ml), streptomycin (100 μg/ml), and l-glutamine (2 mM) (Invitrogen Life Technologies) in a humidified, 5% CO2 incubator. Switching to BZs was induced by culturing infected HFFs in HEPES-buffered RPMI 1640 (pH 8.1) supplemented with 2% FCS in an air incubator.

rSRS9 was produced in insect cells as a secreted protein using the pAcGP67.A baculovirus expression vector, as described for recombinant (rSAG1) production (18). SRS9 coding region lacking the N-terminal signal sequence and C-terminal GPI anchor signal was PCR cloned from Pru strain genomic DNA using the following primers: 5′-ATCGGATCCTCTGCATGAAGGACTTCAGAG-3′ and 5′-CTATCTAGACTAGTGATGGTGATGGTGATGCGCGTACGAAGCAGAACTG-3′ (contains a 6-His tag). Highly pure (>98%) rSRS9 and rSAG1 were obtained by using nickel agarose beads. Mouse polyclonal anti-SRS9 antisera were raised in BALB/c mice (The Jackson Laboratory) by i.p. immunization with 20 μg of rSRS9 per mouse in 200 μl of Ribi adjuvant composed of purified monophosphoryl lipid A and synthetic trehalose dicorynomycolate (Corixa). Mice were given four boosts spaced 3 wk apart with the same Ag preparations. Tail vein blood was allowed to coagulate overnight at 4°C and spun for 5 min at 14,000 rpm in a microcentrifuge to collect serum. Rabbit polyclonal anti-SRS9 antisera were raised by intradermal immunization (100 μg of rSRS9 in CFA) followed by four s.c. boosts spaced 3 wk apart (50 μg of rSRS9 in IFA; Covance Research Products).

TZs lysed out of HFFs were washed in PBS by centrifugation at 250 × g for 10 min, resuspended in 0.4% SDS sample buffer containing 100 mM DTT, boiled for 5 min and subjected to a 12% SDS-PAGE. Proteins were transferred to a nitrocellulose membrane, which was blocked in PBS/5% nonfat dry milk and incubated with rabbit anti-SAG1 or -SRS9 antisera followed by peroxidase-conjugated goat anti-rabbit IgG (Kirkegaard & Perry Laboratories). Detection was by ECL chemiluminescence reagents (Amersham Biosciences).

Infected HFFs grown on cover slips were fixed in 3.7% formaldehyde and permeabilized in 0.2% Triton X-100. Brain homogenates containing cysts were fixed in methanol. Samples were incubated in PBS/3% BSA with indicated anti-parasite Ag Abs followed by appropriate goat IgG coupled with Alexa Fluor 488, Alexa Fluor 594, or Cascade Blue (Molecular Probes). Cover slips were mounted on a glass slide with Vectashield (Vector Laboratories), and photographs were taken using the Image Pro Plus software and a 35 mm digital camera (model C4742-95; Hamamatsu) connected to an upright (model BX60; magnification, ×1000; Olympus) or an inverted fluorescence microscope (model TE300; magnification, ×200; Nikon).

TZs lysed from HFFs were incubated in PBS/2% FCS (106 TZ in 100 μl volume) with indicated anti-parasite Ag Abs followed by appropriate goat IgG coupled with FITC or PE (BD Biosciences). Data acquisition and analysis were by FACScan and CellQuest software (BD Biosciences).

SRS9c and SAG1c were derived from PruΔhpt (hypoxanthine-xanthine-guanine phosphoribosyltransferase-deficient Pru strain), a gift from D. Soldati (University of Geneva, Geneva, Switzerland) (29). Upstream (promoter) sequences from the TZ-specific GRA1 (pGRA1) (30) and from the BZ-specific SRS9 (pSRS9) were used to express transgenic SRS9 and SAG1 in the TZ and BZ stage, respectively. The SRS9 promoter region (∼1.5 kb) was PCR cloned from the Pru strain genomic DNA using the following primers: 5′-GGGGAAGCTTTGTCACCGGTTCGGTGCACT-3′ and 5′-GCCCATGCATTGTGTCGACCCGTGTGCACG-3′. Full-length SRS9 and SAG1 coding regions were cloned with the following primers: 5′-GGGCATGCATGAAAGGACAGGCAATATGCAG-3′ and 5′-GCCCTTAATTAATTACAATGAAGCAACAACGAACC-3′ for SRS9; 5′-GGGCATGCATTCGGTTTCGCTGCACCACTT-3′ and 5′-GCCCTTAATTAATCACGCGACACAAGCTGCGA-3′ for SAG1. PCR products were digested with NsiI and PacI, and cloned between the indicated promoter and 3′ GRA2 downstream sequence in a plasmid vector containing a copy of HPT gene driven by the dihydrofolate reductase promoter (pDHFR) (31) (see Fig. 3 A). The plasmids were linearized with NotI and electroporated into PruΔhpt as described (32). Clones were derived from over two independent populations after selecting for HPT activity using mycophenolic acid and xanthine (50 μg/ml each).

FIGURE 3.

Generation of SRS9c and SAG1c mutants. A, Plasmid constructs used to create SRS9c, SAG1c, and WT parasites from PruΔhpt by electroporation. B, Western blot with WT and SRS9c TZ lysates probed with mouse anti-SRS9 antisera and later reprobed with rabbit anti-SAG1 antisera. A total of 5 × 106 TZs per lane. SRS9c mutants S7 and S8 were cloned from a population distinct from the one K2 and K10 were cloned from. C, Flow cytometry with live WT and SRS9c TZs using SAG1-specific mAb DG52 and rabbit anti-SRS9 antisera. PE-coupled goat anti-mouse IgG and FITC-coupled goat anti-rabbit IgG were used for detection. Dead host cell debris were gated out based on forward light scatter and negative SAG1 staining. Contour plots of SAG1 vs SRS9 expression were drawn to show that transgenic SRS9 expression did not alter SAG1 expression levels at the SRS9c parasite surface. D, Brain cysts from WT- or SAG1c-infected mice (3 wk postinfection) fixed in methanol and stained with rabbit anti-SAG1 antisera followed by Alexa Fluor 488-coupled goat anti-rabbit IgG. SAG1c-C5 and -D7 were cloned from two independent populations.

FIGURE 3.

Generation of SRS9c and SAG1c mutants. A, Plasmid constructs used to create SRS9c, SAG1c, and WT parasites from PruΔhpt by electroporation. B, Western blot with WT and SRS9c TZ lysates probed with mouse anti-SRS9 antisera and later reprobed with rabbit anti-SAG1 antisera. A total of 5 × 106 TZs per lane. SRS9c mutants S7 and S8 were cloned from a population distinct from the one K2 and K10 were cloned from. C, Flow cytometry with live WT and SRS9c TZs using SAG1-specific mAb DG52 and rabbit anti-SRS9 antisera. PE-coupled goat anti-mouse IgG and FITC-coupled goat anti-rabbit IgG were used for detection. Dead host cell debris were gated out based on forward light scatter and negative SAG1 staining. Contour plots of SAG1 vs SRS9 expression were drawn to show that transgenic SRS9 expression did not alter SAG1 expression levels at the SRS9c parasite surface. D, Brain cysts from WT- or SAG1c-infected mice (3 wk postinfection) fixed in methanol and stained with rabbit anti-SAG1 antisera followed by Alexa Fluor 488-coupled goat anti-rabbit IgG. SAG1c-C5 and -D7 were cloned from two independent populations.

Close modal

Eight-week-old CBA/J female mice (The Jackson Laboratory) were infected i.p. with TZs. Infected HFFs were syringed-lysed using a 27-gauge needle to release TZs. TZs were washed in PBS by centrifugation at 250 × g for 10 min and counted with a hemacytometer. Mice were injected i.p. with 400 or 4000 TZs in 200 μl of PBS. We considered >3 wks postinfection as the chronic phase of infection. All animal studies have been reviewed and approved by Stanford University Administrative Panel for Laboratory Animal Care.

CBA/J females (8 wk of age) were immunized s.c. with rSRS9 (20 μg/mouse in a 100 μl volume) prepared in Ribi adjuvant or adjuvant alone into one site at the back of the neck. Mice were boosted s.c. twice, 3 wk apart, with the same Ag preparations and challenged i.p. with 4000 wild-type (WT) TZs 7 days after the second boost.

Tissues were homogenized over 100-μm cell strainers. Aliquots of tissue homogenates were syringe-lysed using 27-gauge needles to prevent intact mouse cells from serving as host cells and thus interfering with plaque assays. Two-fold serial dilutions starting from 1% of the whole tissue were added to HFF monolayers in 12-well plates and cultured for 24 h in DMEM/3% FCS. Medium was replaced with DMEM/10% FCS the next day. HFFs were fixed in methanol at 4 days postinfection (dpi) and stained with rabbit anti-SAG1 antisera followed by Alexa Fluor 488. Plaques were counted using an inverted fluorescence microscope (magnification, ×200). Parasite loads are expressed as PFU per entire organ.

Brains were homogenized over 100-μm cell strainers. Cells were washed in 25 ml of PBS by centrifugation at 150 × g for 10 min and resuspended in 20 ml of PBS/2% FCS. The 5% of the entire brain sample was stained with fluorescein-conjugated Dolichos biflorus agglutinin (Vector Laboratories) to stain the cyst wall (33). After washing, samples were resuspended in 500 μl of PBS/2% FCS, and 50 μl of aliquots were seeded in a flat-bottom 96-well plate. Cysts were counted in all 10 wells using an inverted fluorescence microscope (magnification, ×200). To count T cells, 1% of the total brain homogenates preincubated with Fc block were stained with FITC-coupled anti-CD4 or anti-CD8 Abs (BD Biosciences). Serial dilutions of each sample were seeded in a flat-bottom 96-well plate and FITC+ cells were counted using an inverted fluorescence microscope (magnification, ×200).

Brains were fixed for >24 h in 3.7% formaldehyde solution. Sections (10-μm thick) were cut 100 μm apart and stained with H&E (Histotec). Photographs of H&E-stained sections were taken with a camera attached to a microscope (magnification, ×200) and used to measure the diameter of each cyst as an arbitrary scale.

To quantitate Ag-specific IgG, serial dilutions of serum prepared from tail vein blood were added to ELISA plates coated with rSRS9 or rSAG1 (50 ng/well). After incubation with peroxidase-coupled goat anti-rabbit IgG and its substrate (Kirkegaard & Perry Laboratories), plates were read colorimetrically at 450 nm. To detect cytokines secreted by cultured splenocytes, spleens were homogenized over 100-μm cell strainers. After lysing RBC (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM Na2EDTA), splenocytes were stimulated with indicated Ags in a 12-well plate (8 × 106 cells/well in 2 ml of medium) using DMEM supplemented with 10% FCS, penicillin (100 U/ml), and streptomycin (100 μg/ml), l-glutamine (2 mM), and 2-ME (55 μM). At desired time points, 250 μl of supernatant were collected and analyzed for IFN-γ, TNF-α, and IL-10 levels using ELISA kits according to the manufacturer’s instructions (Pierce). Following incubation with peroxidase-coupled detection Ab and its substrate, plates were read colorimetrically at 450 nm. Serum cytokine levels were determined using the same ELISA kits.

Student’s t test was performed to compare parasite loads in different experimental groups.

SRS9 was first identified as an abundant, BZ-specific mRNA (15). To confirm that SRS9 is expressed as a BZ-specific surface protein, we performed immunofluorescence assays using polyclonal antisera raised against rSRS9 produced by the baculovirus system. In vitro BZs were obtained by culturing in a high pH medium (34) HFFs infected with the PruΔhpt parasite engineered to express GFP from the TZ-specific GRA1 promoter (pGRA1), a gift from G. Arrizabalaga (University of Idaho, Moscow, ID). We show that SRS9 is expressed by GFPdim BZs but not by GFPbright TZs (Fig. 1,A). SAG1, as expected, is expressed by TZs but not by BZs (Fig. 1,B). BZs within brain cysts isolated from a PruΔhpt-infected mouse express SRS9 but not SAG1 at the parasite surface (Fig. 1 C).

FIGURE 1.

Stage-specific expression of SAG1 and SRS9. HFFs infected with PruΔhpt/pGRA1-GFP were grown on cover slips in a pH 8.1 medium for 4 days to induce stage conversion. Parasite surface was stained with mouse anti-SRS9 antisera (A) or SAG1-specific mAb DG52 (B) (58 ), followed by Cascade Blue-coupled goat anti-mouse IgG. C, Brain cysts isolated from a PruΔhpt-infected CBA/J mouse (3 wk postinfection) were costained with fluorescein-coupled Dolichos that binds to the cyst wall and rabbit anti-SRS9 or anti-SAG1 followed by Alexa Fluor 594-coupled goat anti-rabbit IgG. Ten-micron scale bars are shown.

FIGURE 1.

Stage-specific expression of SAG1 and SRS9. HFFs infected with PruΔhpt/pGRA1-GFP were grown on cover slips in a pH 8.1 medium for 4 days to induce stage conversion. Parasite surface was stained with mouse anti-SRS9 antisera (A) or SAG1-specific mAb DG52 (B) (58 ), followed by Cascade Blue-coupled goat anti-mouse IgG. C, Brain cysts isolated from a PruΔhpt-infected CBA/J mouse (3 wk postinfection) were costained with fluorescein-coupled Dolichos that binds to the cyst wall and rabbit anti-SRS9 or anti-SAG1 followed by Alexa Fluor 594-coupled goat anti-rabbit IgG. Ten-micron scale bars are shown.

Close modal

To compare the nature of immune responses to SAG1 and SRS9 in infection, CBA/J mice were infected i.p. with PruΔhpt TZs. As expected, a robust SAG1-specific serum IgG was detectable in chronic infection sera (6 wk postinfection) (Fig. 2,A), and in vitro stimulation with rSAG1 of splenocytes from the same mice led to the production of IFN-γ (Fig. 2,B). In addition to IFN-γ, we found that the SAG1-specific T cells produced IL-10 (Fig. 2 B), a potent antagonist of IFN-γ.

FIGURE 2.

Lack of a SRS9-specific immune response in natural infection. A, Serum IgG titers specific for SAG1 or SRS9 determined by ELISA in PruΔhpt-infected CBA/J mice (6 wk postinfection; 400 TZs i.p.) and shown as absorbance readings at 450 nm. Each line represents a single mouse. B, IFN-γ and IL-10 production by splenocytes from the same mice. Cells were stimulated with rSAG1 or rSRS9 (20 μg/ml) for 48 h, and cytokines secreted into culture supernatants were measured by ELISA (lower limit of detection ∼30 pg/ml). The mean of duplicate determinations is shown.

FIGURE 2.

Lack of a SRS9-specific immune response in natural infection. A, Serum IgG titers specific for SAG1 or SRS9 determined by ELISA in PruΔhpt-infected CBA/J mice (6 wk postinfection; 400 TZs i.p.) and shown as absorbance readings at 450 nm. Each line represents a single mouse. B, IFN-γ and IL-10 production by splenocytes from the same mice. Cells were stimulated with rSAG1 or rSRS9 (20 μg/ml) for 48 h, and cytokines secreted into culture supernatants were measured by ELISA (lower limit of detection ∼30 pg/ml). The mean of duplicate determinations is shown.

Close modal

In contrast, SRS9-specific serum IgG was conspicuously absent in the same infected mice (Fig. 2,A), as well as in 12 different chronically infected people who have most likely acquired infection by ingesting BZ cysts (data not shown). The lack of detectable IgG in infection was not unique to SRS9 but was also true for other BZ-specific SRS Ags, such as SAG2C, SAG2X, and SAG2Y, despite the fact that all these molecules were highly immunogenic in animals when given along with an appropriate adjuvant (data not shown). A T cell response to SRS9 was also undetectable: splenocytes from the infected mice did not produce IFN-γ or IL-10 when stimulated with rSRS9 in vitro (Fig. 2 B).

The stage-specific expression of SRS Ags may be one of the reasons why BZs are not cleared by the immune system but persist. Because rSRS9 given with an adjuvant is highly immunogenic in animals, we predicted that it should be immunogenic during an infection if it could be expressed in the TZ stage and that the resulting immune response could eliminate BZs that subsequently appeared. Likewise, SAG1, if still expressed after TZ-to-BZ conversion, might continue to attract the robust SAG1-specific response leading to the elimination of BZs. To test these hypotheses, we engineered parasites to express SAG1 or SRS9 in a developmentally inappropriate, “constitutive” manner. SRS9c and SAG1c mutants were derived from PruΔhpt by using the TZ-specific pGRA1 and BZ-specific pSRS9 promoters to drive SRS9 and SAG1 expression, respectively (Fig. 3,A). To control for the presence of the HPT gene in the expression vectors, WT parasites were engineered by transforming PruΔhpt with the same vectors from which the SAG1 or SRS9 coding region was removed (Fig. 3 A). Clones were obtained from at least two independent populations per construct after selecting for HPT activity.

Because WT TZs lack SRS9 expression, SRS9c mutants were readily identified by Western blotting with TZ lysates (Fig. 3,B), as well as by flow cytometry with live TZs (Fig. 3,C). Transgenic SRS9 reaches the parasite surface (Fig. 3,C) and can be cleaved with phosphatidylinositol phospholipase C indicating that it is GPI-anchored as strongly predicted from its sequence (data not shown). The relative abundance of SAG1 vs SRS9 in SRS9c mutants was estimated to be ∼5-fold (∼178,000 SAG1 and ∼36,000 SRS9 molecules per parasite) by comparing band intensities in a Western blot performed with rabbit anti-SAG1, a gift from M. Grigg (University of British Columbia, Vancouver, BC, Canada) or anti-SRS9 antisera and serial dilutions of denatured recombinant proteins and TZ lysates (data not shown). Because of the TZ-specific nature of pGRA1 (Fig. 1), it is highly unlikely that SRS9 is over-expressed in the BZ stage of SRS9c mutants compared with WT parasites.

SAG1c mutants were identified after infecting mice with individual clones and examining SAG1 expression in brain cysts by immunofluorescence microscopy. Different mutant clones showed similar SAG1 expression levels in BZs (Fig. 3,D), which appeared lower than in TZs (data not shown). Due to the difficulty in obtaining a sufficient number of pure cysts from infected mice, the exact levels of SAG1 expression by SAG1c BZs vs TZs could not be determined. In the TZ stage, SAG1 expression levels between WT and SAG1c mutants were indistinguishable (data not shown), as expected from the endogenous pSRS9 being active exclusively in the BZ stage (Fig. 1). Also, even in the context of the plasmid construct we used, pSRS9 is BZ-specific because the same construct containing the GFP gene instead of the SAG1 gene results in GFP expression only in the BZ stage (data not shown).

To avoid possible positional effects associated with random integration of the expression vectors into the genome that could obscure the interpretation of in vivo phenotypes, mutant clones from at least two independent populations were used in all studies. WT parasites were maintained as either nonclonal or clonal populations, both of which showed similar growth and dissemination rates in acute infection (data not shown), as well as similar brain cyst loads in chronic infection (Fig. 4, B and D). WT and SRS9c mutants showed similar growth rates during coinfection of HFFs in vitro (data not shown).

FIGURE 4.

Continual decrease in brain cyst loads in SRS9c infection. A, Survival curves of CBA/J mice infected i.p. with 4000 TZs of indicated parasite strains (n = 12 per group). WT used is not a clone but a stable population. B, Brain cyst loads in representative mice from A determined at 3 and 6 wk postinfection (wpi). C, Survival curves of mice infected i.p. with 400 TZs of indicated parasite strains (n = 13 per group). WT c3 and c28 are clonal populations. D, Brain cyst loads in representative mice from C determined at 4 and 9 wk postinfection (wpi). Each diamond represents a single mouse. Bars represent average cyst counts. ∗, Significant difference (p < 0.05) in cyst loads between WT and SRS9c at the given week postinfection; #, significant difference (p < 0.05) in SRS9c cyst loads between the two time points.

FIGURE 4.

Continual decrease in brain cyst loads in SRS9c infection. A, Survival curves of CBA/J mice infected i.p. with 4000 TZs of indicated parasite strains (n = 12 per group). WT used is not a clone but a stable population. B, Brain cyst loads in representative mice from A determined at 3 and 6 wk postinfection (wpi). C, Survival curves of mice infected i.p. with 400 TZs of indicated parasite strains (n = 13 per group). WT c3 and c28 are clonal populations. D, Brain cyst loads in representative mice from C determined at 4 and 9 wk postinfection (wpi). Each diamond represents a single mouse. Bars represent average cyst counts. ∗, Significant difference (p < 0.05) in cyst loads between WT and SRS9c at the given week postinfection; #, significant difference (p < 0.05) in SRS9c cyst loads between the two time points.

Close modal

In CBA/J mice infected i.p. with 4000 TZs of WT and SRS9c mutants, mortality usually began to occur 12–14 dpi, following the peak of acute infection (Fig. 4,A) and around the time most parasites are found in the brain (35). With all infecting strains, mortality often occurred throughout the chronic phase (>3 wk postinfection) with SRS9c-infected mice showing similar or slightly better (but not statistically significant) survival than WT-infected mice (Fig. 4,A). The number of cysts persisting in the brains of WT-infected mice remained relatively constant during the chronic infection (Fig. 4,B). Brain cyst loads in SRS9c-infected mice were similar to or sometimes lower than those in WT infection in the early phase of chronic infection (3 wk postinfection), but they significantly declined over time and were 17–50% of those in WT-infected brains by the late phase of chronic infection (6 wk postinfection) (Fig. 4 B).

When the inoculum size was reduced by 10-fold to 400 TZ per mouse, the onset of death was generally delayed by more than a week and the occurrence of mortality was still observed throughout the chronic phase (Fig. 4,C). SRS9c-infected mice showed a similar or marginally better survival during chronic infection than WT-infected mice, which was not statistically significant but consistently observed in all five independent experiments performed (data not shown). SRS9c-infected mice showed a better overall health status than WT-infected mice as they weighed more (27.6 ± 3.4 g; n = 7) than WT-infected mice (20.1 ± 2.5 g; n = 6) at 9 wk postinfection (age-matched, naive mice weighed 30.5 ± 0.8 g; n = 3) (data not shown). As with the higher inoculum and after initial parity, the brain cyst loads continually decreased in SRS9c-infected mice and were 20–30% of those in WT infection by 9 wk postinfection (Fig. 4,D). Notably, the 10-fold difference in inoculum size, which causes a clear difference in acute phase parasite loads in the whole body (data not shown), yielded similar brain cyst loads in chronic infection (Fig. 4, B and D). This implies that the number of persisting parasites is not simply determined by the inoculum size but may be controlled by complex factors including, e.g., the host response.

Because brain cyst loads in the early phase of chronic infection were similar between WT and SRS9c mutants, the subsequent reduction in SRS9c cyst loads is unlikely due to a growth defect. Indeed, up to 15 dpi, both WT and SRS9c mutants showed similar parasite loads in the spleen (the major site of parasite replication after i.p. infection) (Fig. 5,A) and in the brain (Fig. 5 B), indicating that dissemination from the site of infection and subsequent growth in target tissues during the acute phase is comparable. Brain cyst counts at 15 dpi were also similar between WT (323 ± 140; n = 3) and SRS9c-K2 (300 ± 46; n = 3) (data not shown). Moreover, there was no measurable difference in brain cyst size between WT and SRS9c mutants measured in the H&E-stained brain sections (data not shown), suggesting that a defect in BZ growth is not responsible for the decrease in SRS9c cyst loads.

FIGURE 5.

Comparable growth of WT and SRS9c mutants in vivo. Parasite loads determined by plaque assays at indicated day postinfection in the spleen (A) and brain (B) of CBA/J mice infected with 400 TZs of WT or SRS9c mutants. Shown are mean PFU ± SD (n = 3 per group per time point).

FIGURE 5.

Comparable growth of WT and SRS9c mutants in vivo. Parasite loads determined by plaque assays at indicated day postinfection in the spleen (A) and brain (B) of CBA/J mice infected with 400 TZs of WT or SRS9c mutants. Shown are mean PFU ± SD (n = 3 per group per time point).

Close modal

Another possible reason for observing decreased brain cyst loads in SRS9c infection would be if the transgenic SRS9 expression in TZs altered tissue tropism of the parasite such that the parasite preferentially located to and encysted in other tissues than in the brain. This possibility is unlikely because parasite loads were similar between WT- and SRS9c-infected brains up to 4 wk postinfection (Figs. 5,B, and 4, B and D). Also, regardless of the infecting strain, cysts detectable after dolichos-fluorescein staining were present only in the brains and were absent in the lungs and livers that are targets of the acute infection (35) (zero cysts in 5% of the total tissue homogenates, both at 4 and 9 wk postinfection; data not shown).

A SRS9-specific immune response that was absent in WT infection was readily detectable in SRS9c infection. Sera from SRS9c-infected mice (6 wk postinfection) had a high titer SRS9-specific IgG unlike WT-infected mice (Fig. 6,A). As expected, SAG1-specific IgG titers were high in all infected mice (data not shown). SRS9c infection also activated a SRS9-specific T cell response marked by IFN-γ but not IL-10 production by cultured splenocytes (Fig. 6,B). Although SAG1 clearly dominates the humoral response in Toxoplasma infection, it is not known whether it also dominates the T cell response. In SRS9c infection, the amount of rSAG1-stimulated IFN-γ production was substantially higher than that induced by rSRS9. Because SAG1 is the most abundant SRS Ag in TZs and expressed in ∼5-fold excess over SRS9 in SRS9c TZs, the size of the T cell response to SAG1 is expected to be greater than that to SRS9 (and most other TZ-specific Ags for that matter). Neither T cell population produced detectable levels of TNF-α regardless of the infecting strain (data not shown). Generally, cytokine secretion in the absence of an antigenic stimulus (Fig. 6,B) or in the presence of PMA and ionomycin (Fig. 6 C) was higher with splenocytes from infected mice than with cells from naive mice, indicating an infection-driven accumulation of immune cells still detectable in the spleen at 6 wk postinfection.

FIGURE 6.

Induction of a SRS9-specific immune response in SRS9c infection. A, Serum IgG titers specific for SRS9 determined by ELISA in CBA/J mice infected (6 wk postinfection) with 400 TZs of WT (n = 4) or SRS9c mutants (n = 2 per strain) and shown as absorbance readings at 450 nm. Each line represents a single mouse. B, Cytokines produced by splenocytes from naive and infected (6 wk postinfection) mice cultured with no Ag, rSAG1, or rSRS9 (20 μg/ml) for 6, 24, 48, or 72 h before supernatants were analyzed in ELISA. For WT, mean ± SD of two duplicate determinations from two mice is shown. For others, mean of duplicate determinations is shown. C, Same mean as in B except PMA (25 ng/ml) and ionomycin (1 μg/ml) were used for stimulation. Lower limit of detection was ∼30 pg/ml for all cytokines tested.

FIGURE 6.

Induction of a SRS9-specific immune response in SRS9c infection. A, Serum IgG titers specific for SRS9 determined by ELISA in CBA/J mice infected (6 wk postinfection) with 400 TZs of WT (n = 4) or SRS9c mutants (n = 2 per strain) and shown as absorbance readings at 450 nm. Each line represents a single mouse. B, Cytokines produced by splenocytes from naive and infected (6 wk postinfection) mice cultured with no Ag, rSAG1, or rSRS9 (20 μg/ml) for 6, 24, 48, or 72 h before supernatants were analyzed in ELISA. For WT, mean ± SD of two duplicate determinations from two mice is shown. For others, mean of duplicate determinations is shown. C, Same mean as in B except PMA (25 ng/ml) and ionomycin (1 μg/ml) were used for stimulation. Lower limit of detection was ∼30 pg/ml for all cytokines tested.

Close modal

To test whether the SRS9-specific immune response contributed to the decrease in brain cyst loads in SRS9c infection, CBA/J mice were immunized s.c. with rSRS9 prepared in Ribi adjuvant or with adjuvant alone and then challenged i.p. with 4000 WT TZs. Following infection, rSRS9-immune mice showed better survival (Fig. 7,A) and lower brain cyst loads (Fig. 7 B) than the control group. This result suggests that the induction of a SRS9-specific immune response in SRS9c infection may have been a key mechanism that led to the clearance of BZ cysts during chronic infection.

FIGURE 7.

Improved survival and reduced brain cyst loads in mice immunized with rSRS9 following challenge infection with WT parasite. A, Survival curves of CBA/J mice immunized s.c. with Ribi adjuvant with or without rSRS9 (n = 10 per group) and then challenged i.p. with 4000 WT TZs. B, Brain cyst loads at 6 wk postinfection. The difference between the two groups is significant (∗, p < 0.05).

FIGURE 7.

Improved survival and reduced brain cyst loads in mice immunized with rSRS9 following challenge infection with WT parasite. A, Survival curves of CBA/J mice immunized s.c. with Ribi adjuvant with or without rSRS9 (n = 10 per group) and then challenged i.p. with 4000 WT TZs. B, Brain cyst loads at 6 wk postinfection. The difference between the two groups is significant (∗, p < 0.05).

Close modal

Because SAG1 dominates the immune response during the acute infection with TZs, one would anticipate that BZs still expressing SAG1 will be effectively cleared by the immune response. Indeed, the average brain cyst count in SAG1c infection was often lower than that in WT infection at 3–4 wk postinfection (Fig. 8, A and B). However, in contrast to SRS9c infection, the cyst count substantially increased over time in SAG1c-infected brains and, by 6–7 wk postinfection, far exceeded that in WT-infected brains by up to 3.5-fold (Fig. 8, A and B). This increase in SAG1c cyst loads was associated with an increase in mortality during the late phase of chronic infection (Fig. 8 C), which was not statistically significant but consistently observed in all five independent experiments we performed (data not shown). At 6 wk postinfection, SAG1c-infected mice weighed 17.5 ± 1.5 g (n = 3), less than WT-infected (20.5 ± 1.8 g; n = 3) or SRS9c-K2-infected mice (21.2 ± 0.7 g; n = 3) (age-matched, naive mice weighed 25.2 ± 0.8 g; n = 3), reflecting the poor overall health status (data not shown).

FIGURE 8.

Continual increase in brain cyst loads in SAG1c infection. A, Brain cyst loads at 4 and 6 wk postinfection in CBA/J mice infected i.p. with 400 TZs of WT, SRS9c-K2, SAG1c-A4, and SAG1c-D7 (n = 10 per group). Each diamond represents a single mouse. Bars represent average cyst counts. B, Brain cyst loads at 3.5 and 7 wk postinfection (wpi) in representative CBA/J mice from C that were infected i.p. with 400 TZs of WT and SAG1c-C5 and SAG1c-D7. C, Survival curves of CBA/J mice following i.p. infection with 400 TZs of indicated strains (n = 12 per group). ∗, Significant (p < 0.05) difference in cyst loads between WT and mutants at the given week postinfection; #, Significant difference (p < 0.05) in SRS9c and SAG1c cyst loads between the two time points.

FIGURE 8.

Continual increase in brain cyst loads in SAG1c infection. A, Brain cyst loads at 4 and 6 wk postinfection in CBA/J mice infected i.p. with 400 TZs of WT, SRS9c-K2, SAG1c-A4, and SAG1c-D7 (n = 10 per group). Each diamond represents a single mouse. Bars represent average cyst counts. B, Brain cyst loads at 3.5 and 7 wk postinfection (wpi) in representative CBA/J mice from C that were infected i.p. with 400 TZs of WT and SAG1c-C5 and SAG1c-D7. C, Survival curves of CBA/J mice following i.p. infection with 400 TZs of indicated strains (n = 12 per group). ∗, Significant (p < 0.05) difference in cyst loads between WT and mutants at the given week postinfection; #, Significant difference (p < 0.05) in SRS9c and SAG1c cyst loads between the two time points.

Close modal

The increase in brain cyst loads in SAG1c infection is unlikely due to an enhanced growth of SAG1c mutants. Following i.p. infection, parasite loads expanded almost equally from 3 to 7 dpi in the spleens of all infected mice (Fig. 9,A). At 11 dpi, although statistically not significant, the mean parasite loads in the spleen tend to be somewhat lower in SAG1c infection than in WT infection (Fig. 9,A), possibly indicating an improved parasite clearance due to SAG1 expression by BZs. Parasite loads in the brain at 11 dpi were similar in WT and SAG1c infection (Fig. 9 A). The brain cyst size measured in the H&E-stained brain sections was also similar between WT and SAG1c mutants (data not shown), indicating no significant difference in the rate of BZ growth.

FIGURE 9.

Cytokine hyperproduction in SAG1c infection despite comparable in vivo growth of WT and SAG1c mutants. A, Parasite loads in spleen and brain determined by plaque assays at indicated day postinfection in CBA/J mice infected i.p. with 400 TZs of indicated parasite strains. Shown are mean PFU ± SD (n = 3 per group per time point). B, Cytokine production by splenocytes from mice uninfected or infected with indicated parasite strains (6 wk postinfection) detected by ELISA after stimulating cells in vitro with rSAG1 or rSRS9 (20 μg/ml) for up to 72 h. For WT, mean ± SD of two duplicate determinations from two mice are shown. For others, means of duplicate determinations are shown. Lower limit of detection was ∼30 pg/ml for all cytokines tested.

FIGURE 9.

Cytokine hyperproduction in SAG1c infection despite comparable in vivo growth of WT and SAG1c mutants. A, Parasite loads in spleen and brain determined by plaque assays at indicated day postinfection in CBA/J mice infected i.p. with 400 TZs of indicated parasite strains. Shown are mean PFU ± SD (n = 3 per group per time point). B, Cytokine production by splenocytes from mice uninfected or infected with indicated parasite strains (6 wk postinfection) detected by ELISA after stimulating cells in vitro with rSAG1 or rSRS9 (20 μg/ml) for up to 72 h. For WT, mean ± SD of two duplicate determinations from two mice are shown. For others, means of duplicate determinations are shown. Lower limit of detection was ∼30 pg/ml for all cytokines tested.

Close modal

As with SRS9c mutants, cysts formed by SAG1c mutants were only detectable in the brain and not in the lung and liver (data not shown), indicating that constitutive SAG1 (SAG1c) expression probably did not result in altered tissue tropism.

In immunocompetent mice infected with WT parasites, brain cysts occasionally rupture during chronic infection but this does not lead to active parasite replication because of an immune response resulting in the killing of the parasites (36). The increase in brain cyst loads in SAG1c infection suggests that the immune response may not be properly regulated in infected brains and, therefore, it may not effectively control parasite replication or it could even promote differentiation to BZs.

To test this possibility, we compared immune responses elicited by WT, SRS9c and SAG1c mutants and found that splenocytes from SAG1c-infected mice (6 wk postinfection) produced much higher levels of IFN-γ when stimulated with rSAG1 in vitro (>50 ng/ml at 48 h culture) (Fig. 9,B) than did the cells from WT- and SRS9c-infected mice (<20 ng/ml at 48 h culture) (Fig. 6,B). Moreover, unlike in WT and SRS9c infection, significant levels of TNF-α were produced by splenocytes from SAG1c-infected mice when stimulated with rSAG1 (Fig. 9,B). SAG1-specific IL-10 production was also enhanced in SAG1c infection (Fig. 9,B) compared with WT and SRS9c infection (Fig. 6 B).

This result suggests that SAG1c infection causes cytokine dysregulation due to an excessive expansion and/or activation of SAG1-specific T cells in the spleen despite the comparable in vivo growth rates of WT and SAG1c parasites. However, serum IFN-γ levels were similar among mice infected with WT, SRS9c, or SAG1c mutants: ∼2000 and ∼700 pg/ml at 4 and 6 wk postinfection, respectively, compared with ∼300 pg/ml in naive mice (data not shown). Serum IL-10 and TNF-α were undetectable in all mice (data not shown). This result is perhaps not surprising given that the major site of parasite replication during chronic infection is the brain. Cytokines locally produced in the brain may be more relevant to the contrasting phenotypes of SRS9c and SAG1c mutants.

In histologic examination of H&E-stained brain sections, meningitis and encephalitis were apparent in all infected mice (6 wk postinfection) regardless of the infecting strain (Fig. 10, A–F). Mononuclear infiltrates were mostly found in the cerebral cortex where the great majority of cysts were localized. Compared with WT (Fig. 10,B) and SRS9c infection (Fig. 10, C and D), SAG1c-infected brains showed much severe inflammation with prominent areas of necrosis (Fig. 10, E and F).

FIGURE 10.

Distinct immunological events occurring in the brains in WT, SRS9c, and SAG1c infection. H&E stained brain sections from CBA/J mice uninfected (A) and infected (B–F) (6 wk postinfection) with indicated parasite strains. Arrowheads indicate cysts. G, Number of T cells per brain at indicated week postinfection (wpi). Shown are mean ± SD (n = 3 per group). When examined, all CD4+ and CD8+ T cells were costained with anti-CD3 Ab (data not shown). H, Proportion of brain cysts associated with immune response (6 wk postinfection) was calculated based on the numbers of lone cysts (a), cysts surrounded by infiltrates (b), and inflammation foci apparently lacking cysts (c). Infiltrates around blood vessels or those broadly distributed over a large area of the cortex were excluded from the analysis.

FIGURE 10.

Distinct immunological events occurring in the brains in WT, SRS9c, and SAG1c infection. H&E stained brain sections from CBA/J mice uninfected (A) and infected (B–F) (6 wk postinfection) with indicated parasite strains. Arrowheads indicate cysts. G, Number of T cells per brain at indicated week postinfection (wpi). Shown are mean ± SD (n = 3 per group). When examined, all CD4+ and CD8+ T cells were costained with anti-CD3 Ab (data not shown). H, Proportion of brain cysts associated with immune response (6 wk postinfection) was calculated based on the numbers of lone cysts (a), cysts surrounded by infiltrates (b), and inflammation foci apparently lacking cysts (c). Infiltrates around blood vessels or those broadly distributed over a large area of the cortex were excluded from the analysis.

Close modal

Both CD4+ and CD8+ T cells are known to significantly contribute to Toxoplasma encephalitis (37), and were recruited to and accumulated in the brains in WT, SRS9c, and SAG1c infection and gradually decreased in numbers during the chronic infection (Fig. 10,G). Given that T cells, especially IFN-γ-secreting CD8+ T cells, play a critical role in resistance to Toxoplasma (38, 39), it is notable that more CD8+ T cells were recruited to the brain in SAG1c infection than in WT or SRS9c infection (Fig. 10 G).

In chronically infected brains, the majority of cysts were not associated with mononuclear infiltrates (Fig. 10,Ha). However, many sites of focal inflammation were found throughout the cerebral cortex, some of which were seen to be intimately surrounding cysts (Fig. 10,Hb) whereas others were not (Fig. 10,Hc). The latter (Fig. 10,Hc) might have been seen to associate with a cyst if sections had been cut at different levels or may represent the situation after a cyst had been cleared by the infiltrates. Counting these sites of focal inflammation revealed that a significantly higher proportion of SAG1c cysts were closely associated with infiltrates than WT cysts (Fig. 10,H). This is not simply due to the increased numbers of cysts and infiltrates present in SAG1c infection because SRS9c infection that caused moderate inflammation and produced fewer cysts than WT also had a higher proportion of cysts associated with infiltrates (Fig. 10 H).

Toxoplasma surface proteins belonging to the SRS superfamily are stage-specifically expressed during TZ-to-BZ conversion concomitant with the emergence of an adaptive immune response in acute infection. The stage conversion is critical in the establishment and maintenance of a chronic infection, but it has not been previously tested whether SRS antigenic switch indeed serves as an immune-evasion mechanism that allows parasites to persist. Using genetically manipulated parasites expressing the normally stage-specific SAG1 and SRS9 Ags in a constitutive manner, we have demonstrated that altering parasite stage-specificity substantially affects the immune response they elicit in infection and interferes with the ability of parasites to achieve optimal persistency.

BZ cysts persist despite the long-lasting immune response to TZ-specific SRS Ags, such as immunodominant (at least in terms of humoral response) SAG1 and SAG2A, as well as the response to Ags shared by both TZs and BZs, suggesting that BZ-specific immune responses are required to clear BZs. Somehow, Toxoplasma avoids eliciting such a response because BZ-specific SRS Ags, such as SRS9, SAG2C, SAG2X, and SAG2Y, are not immunogenic in infection. This result is true not only in mice infected i.p. with TZs but also in humans who have most likely acquired infection through oral ingestion of BZ cysts. Therefore, the lack of immunogenicity of BZ SRS Ags appears to hold regardless of the infection route (perhaps because BZs convert to TZs quickly upon reaching the gut before BZ Ags can be seen by the immune system). However, we found that constitutive SRS9 expression elicited a readily detectable SRS9-specific immune response and this response correlated with a significant decrease in cyst loads persisting in the brain. This finding supports the hypothesis that the BZ-specific nature of SRS9 expression is one of the key factors contributing to parasite persistence. The timing of SRS expression likely determines parasite immunogenicity in infection by regulating the way they interact with the immune system: the amount of SRS Ags available for presentation to the immune system is far greater during acute infection with rapidly multiplying TZs than in chronic infection with metabolically quiescent BZs. Also, systemically disseminating TZs may more readily encounter and activate immune cells than do BZs that encyst and predominantly reside in the brain.

The contrasting phenotypes of SRS9c and SAG1c mutants may explain why evolution has led to the immunodominant SAG1 molecule being expressed only in the TZ stage. Although SAG1c expression effectively targets BZ cysts for clearance in the early phase of chronic infection, an over-stimulation of SAG1-specific T cells seems to occur leading to severe encephalitis, and somehow, an increase in brain cyst loads over time. A similar phenomenon has been seen in an attempt to obtain a protective immunity by immunizing mice with SAG1: an i.p. immunization with purified SAG1 protein along with CFA resulted in an increase in mortality and brain cyst burdens in mice subsequently challenged with avirulent Toxoplasma strain (40). This observation and the phenotype of SAG1c in infected mice indicate that an excessive immune response to SAG1 can somehow lead to increased parasite growth and exacerbated disease in the chronic phase of infection. The studies with SRS9c and SAG1c mutants have demonstrated one of many purposes that may be served by stage-specific SRS expression: BZ-specific expression avoids the induction of a potentially protective immune response, whereas TZ-specific expression avoids BZ clearance by an immune response and minimizes immunopathology to the host.

The brain has been viewed as an immunoprivileged site due to such characteristics as the blood-brain barrier, lack of conventional lymphatics, better allograft acceptance, and low T cell trafficking (10). These features will undoubtedly contribute to the lack of immunogenicity of BZ Ags in infection and facilitate BZ persistence in the brain. Recently, however, many studies have shown that resident cells of the brain do have the ability to serve as potent APCs that are quite capable of supporting the induction of immune responses (41). Moreover, Ags exclusively localized in the brain can be transported to the cervical lymph nodes and activate T cells, which are then recruited to the brain (42, 43, 44), and selectively act on target cells (45). Thus, it seems reasonable to speculate that BZ Ags should be able to activate an immune response during chronic infection of the brain and that the lack of immunogenicity of SRS Ags may be the result of complex mechanisms including their localization in the brain.

The type of the immune response induced by TZ-specific SRS Ags may be one of the factors contributing to parasite persistence by down-modulating the induction of an immune response to BZ Ags. T cells are critical for controlling Toxoplasma replication and their protective effect is largely mediated by cytokines, especially IFN-γ (24). We found that, in addition to IFN-γ as previously shown (22, 23), SAG1-specific T cells produce IL-10. As a potent antagonist of IFN-γ, IL-10 is a key regulator in preventing excessive inflammation in the brain and other tissues infected with Toxoplasma (46, 47, 48). T cells simultaneously producing IFN-γ and IL-10 have been frequently observed in persistent infections, including tuberculosis (49), malaria (50), leishmaniasis (51), and Borrelia burgdorferi infection (52). IL-10 produced by these T cells limits the IFN-γ-driven inflammation, and in doing so, hampers complete clearance of infectious agents. Previously, an elevated level of IL-10 mediating immunosuppression was noted in the spleen of a mouse acutely infected with Toxoplasma (53). Conceivably, a T cell response to the dominant SAG1 Ag may greatly contribute to parasite clearance via IFN-γ production and also be the major source of IL-10 that serves as a self-regulatory mechanism to turn off the IFN-γ-driven Th1 type response. While preventing immunopathology in favor of host survival, IL-10 may also create an immunosuppressive environment that is not ideal for eliciting primary responses to newly emerging BZ-specific Ags.

The pattern of changes in brain cyst loads and cellular infiltration during chronic infections with SRS9c and SAG1c mutants is clearly different from that seen in WT infection. This finding indicates that mutant BZs are attracting an active immune response to the brain that is greater than what is seen in WT infection resulting in BZ clearance in SRS9c infection and the emergence of new cysts in SAG1c infection. The fact that more brain cysts are intimately surrounded by immune infiltrates in the SAG1c and even SRS9c infection than in WT infection argues for the possibility that some of these infiltrates may be T cells recruited to the brain due to their specificity for SAG1 and SRS9. Alternatively, or in addition, the T cell response may be targeting an extracerebral parasite reservoir, thus changing the number of parasites colonizing the brain during chronic infection. Understanding how SAG1- and SRS9-specific T cells are activated and maintained in the periphery, how they are recruited to the brain and what their exact effector functions are in dealing with persisting parasites will require the ability to track Ag-specific T cells and parasites in vivo in a quantitative manner. Data available at the present time indicate that the cytokine response induced by SRS9c and SAG1c mutants in the periphery is qualitatively and quantitatively different: while SRS9-specific T cells produce moderate levels of IFN-γ in SRS9c infection, IFN-γ, TNF-α, and IL-10 are hyperproduced by SAG1-specific T cells in SAG1c infection. Dysregulated cytokine production has been linked to many immunologic diseases and shown to be a critical event in some cases (54, 55). Although it is difficult to fully appreciate exact mechanisms, a similar phenomenon may be responsible for regulating parasite growth/differentiation and encephalitis seen in this study in the brains of mice infected with SRS9c and SAG1c mutants.

Another challenging issue in understanding the biology of the SRS superfamily and its role in parasite persistence is the possible immune-interference by presentation of altered peptide ligands (56, 57). Some SRS Ags may contain Ab and T cell epitopes that can interfere with immune responses to closely related SRS Ags. Such interference unlikely exists between SRS9 and SAG1, as immune responses to both Ags were readily detectable in SRS9c infection. However, SRS9 and SAG1 belong to different subgroups within the SRS superfamily sharing only ∼25% sequence identity. An altered peptide ligand phenomenon is more likely to be operative among the SRS Ags of the same subgroup (in which sequence identity can be up to 90%) and could be one of the mechanisms contributing to the lack of immunogenicity of some of the SRS Ags in infection.

Overall, the results presented broaden our understanding of the immunologic roles played by the developmentally regulated SRS Ags of Toxoplasma and clearly argue that the SRS antigenic switch accompanying TZ-to-BZ stage conversion has evolved as one of the critical mechanisms that allow optimal parasite persistence in its intermediate host until such time as transmission to a new host becomes feasible.

We thank Drs. D. Soldati, G. Arrizabalaga, and M. Grigg for kindly providing the PruΔhpt parasite, PruΔhpt/pGRA1-GFP parasite, and rabbit anti-SAG1 antisera, respectively.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This research was supported by National Institutes of Health Grants AI41014 and AI21423.

3

Abbreviations used in this paper: BZ, bradyzoite; TZ, tachyzoite; SRS, surface Ag 1-related sequence; Pru, Prugniaud; HFF, human foreskin fibroblast; HPT, hypoxanthine-xanthine-guanine phosphoribosyltransferase; dpi, days postinfection; WT, wild type; SAG1, surface Ag 1.

1
Tenter, A. M., A. R. Heckeroth, L. M. Weiss.
2000
. Toxoplasma gondii: from animals to humans.
Int. J. Parasitol.
30
:
1217
-1258.
2
Dubey, J. P..
1997
. Bradyzoite-induced murine toxoplasmosis: stage conversion, pathogenesis, and tissue cyst formation in mice fed bradyzoites of different strains of Toxoplasma gondii.
J. Euk. Microbiol.
44
:
592
-602.
3
McLeod, R., R. G. Estes, D. G. Mack, H. Cohen.
1984
. Immune response of mice to ingested Toxoplasma gondii: a model of toxoplasma infection acquired by ingestion.
J. Infect. Dis.
149
:
234
-244.
4
Lyons, R. E., R. McLeod, C. W. Roberts.
2002
. Toxoplasma gondii tachyzoite-bradyzoite interconversion.
Trends Parasitol.
18
:
198
-201.
5
Beaman, M. H., F. G. Araujo, J. S. Remington.
1994
. Protective reconstitution of the SCID mouse against reactivation of toxoplasmic encephalitis.
J. Infect. Dis.
169
:
375
-383.
6
Luft, B. J., J. S. Remington.
1992
. Toxoplasmic encephalitis in AIDS.
Clin. Infect. Dis.
15
:
211
-222.
7
Sinai, A. P., K. A. Joiner.
1997
. Safe haven: the cell biology of nonfusogenic pathogen vacuoles.
Annu. Rev. Microbiol.
51
:
415
-462.
8
Denkers, E. Y., B. A. Butcher.
2005
. Sabotage and exploitation in macrophages parasitized by intracellular protozoans.
Trends Parasitol.
21
:
35
-41.
9
Aliberti, J., D. Jankovic, A. Sher.
2004
. Turning it on and off: regulation of dendritic cell function in Toxoplasma gondii infection.
Immunol. Rev.
201
:
26
-34.
10
Fabry, Z., C. S. Raine, M. N. Hart.
1994
. Nervous tissue as an immune compartment: the dialect of the immune response in the CNS.
Immunol. Today
15
:
218
-224.
11
Radke, J. R., M. N. Guerini, M. Jerome, M. W. White.
2003
. A change in the premitotic period of the cell cycle is associated with bradyzoite differentiation in Toxoplasma gondii.
Mol. Biochem. Parasitol.
131
:
119
-127.
12
van der Woude, M. W., A. J. Baumler.
2004
. Phase and antigenic variation in bacteria.
Clin. Microbiol. Rev.
17
:
581
-611.
13
Tortorella, D., B. E. Gewurz, M. H. Furman, D. J. Schust, H. L. Ploegh.
2000
. Viral subversion of the immune system.
Annu. Rev. Immunol.
18
:
861
-926.
14
Kyes, S., P. Horrocks, C. Newbold.
2001
. Antigenic variation at the infected red cell surface in malaria.
Annu. Rev. Microbiol.
55
:
673
-707.
15
Cleary, M. D., U. Singh, I. J. Blader, J. L. Brewer, J. C. Boothroyd.
2002
. Toxoplasma gondii asexual development: identification of developmentally regulated genes and distinct patterns of gene expression.
Eukaryot. Cell
1
:
329
-340.
16
Jung, C., C. Y.-F. Lee, M. E. Grigg.
2004
. The SRS superfamily of Toxoplasma surface proteins.
Int. J. Parasitol.
34
:
285
-296.
17
Tomavo, S..
1996
. The major surface proteins of Toxoplasma gondii: structures and functions.
Curr. Top. Microbiol. Immunol.
219
:
45
-54.
18
He, X.-L., M. E. Grigg, J. C. Boothroyd, K. C. Garcia.
2002
. Structure of the immunodominant surface antigen from the Toxoplasma gondii SRS superfamily.
Nat. Struct. Biol.
9
:
606
-611.
19
Grigg, M. E., S. Bonnefoy, A. B. Hehl, Y. Suzuki, J. C. Boothroyd.
2001
. Success and virulence in Toxoplasma as the result of sexual recombination between two distinct ancestries.
Science
294
:
161
-165.
20
Bessieres, M. H., S. Le Breton, J. P. Seguela.
1992
. Analysis by immunoblotting of Toxoplasma gondii exo-antigens and comparison with somatic antigens.
Parasitol. Res.
78
:
222
-228.
21
Partanen, P., H. J. Turunen, R. T. Paasivuo, P. O. Leinikki.
1984
. Immunoblot analysis of Toxoplasma gondii antigens by human immunoglobulins G, M, and A antibodies at different stages of infection.
J. Clin. Microbiol.
20
:
133
-135.
22
Prigione, I., P. Facchetti, L. Lecordier, D. Deslée, S. Chiesa, M.-F. Cesbron-Delauw, V. Pistoia.
2000
. T cell clones raised from chronically infected healthy humans by stimulation with Toxoplasma gondii excretory-secretory antigens cross-react with live tachyzoites: characterization of the fine antigenic specificity of the clones and implications for vaccine development.
J. Immunol.
164
:
3741
-3748.
23
Khan, I. A., M. E. Eckel, E. R. Pfefferkorn, L. H. Kasper.
1988
. Production of γ interferon by cultured human lymphocytes stimulated with a purified membrane protein (P30) from Toxoplasma gondii.
J. Infect. Dis.
157
:
979
-984.
24
Suzuki, Y., M. A. Orellana, R. D. Schreiber, J. S. Remington.
1988
. Interferon-γ: the major mediator of resistance against Toxoplasma gondii.
Science
240
:
516
-518.
25
Fatoohi, A. F., G. J. Cozon, T. Greenland, J. Ferrandiz, J. Bienvenu, S. Picot, F. Peyron.
2002
. Cellular immune responses to recombinant antigens in pregnant women chronically infected with Toxoplasma gondii.
Clin. Diagn. Lab. Immunol.
9
:
704
-707.
26
Di Cristina, M., P. Del Porto, W. Buffolano, E. Beghetto, A. Spadoni, S. Guglietta, E. Piccolella, F. Felici, N. Gargano.
2004
. The Toxoplasma gondii bradyzoite antigens BAG1 and MAG1 induce early humoral and cell-mediated immune responses upon human infection.
Microbes Infect.
6
:
164
-171.
27
Ferguson, D. J., S. F. Parmley.
2002
. Toxoplasma gondii MAG1 protein expression.
Trends Parasitol.
18
:
482
.
28
Bohne, W., U. Gross, D. J. Ferguson, J. Heesemann.
1995
. Cloning and characterization of a bradyzoite-specifically expressed gene (hsp30/bag1) of Toxoplasma gondii, related to genes encoding small heat-shock proteins of plants.
Mol. Microbiol.
16
:
1221
-1230.
29
Donald, R. G., D. S. Roos.
1998
. Gene knock-outs and allelic replacements in Toxoplasma gondii: HXGPRT as a selectable marker for hit-and-run mutagenesis.
Mol. Biochem. Parasitol.
91
:
295
-305.
30
Kim, K., M. S. Eaton, W. Schubert, S. Wu, J. Tang.
2001
. Optimized expression of green fluorescent protein in Toxoplasma gondii using thermostable green fluorescent protein mutants.
Mol. Biochem. Parasitol.
113
:
309
-313.
31
Black, M. W., J. C. Boothroyd.
1998
. Development of a stable episomal shuttle vector for Toxoplasma gondii.
J. Biol. Chem.
273
:
3972
-3979.
32
Soldati, D., J. C. Boothroyd.
1993
. Transient transfection and expression in the obligate intracellular parasite Toxoplasma gondii.
Science
260
:
349
-352.
33
Zhang, Y. W., S. K. Halonen, Y. F. Ma, M. Wittner, L. M. Weiss.
2001
. Initial characterization of CST1, a Toxoplasma gondii cyst wall glycoprotein.
Infect. Immun.
69
:
501
-507.
34
Weiss, L. M., D. Laplace, P. M. Takvorian, H. B. Tanowitz, A. Cali, M. Wittner.
1995
. A cell culture system for study of the development of Toxoplasma gondii bradyzoites.
J. Eukaryot. Microbiol.
42
:
150
-157.
35
Zenner, L., A. Foulet, Y. Caudrelier, F. Darcy, B. Gosselin, A. Capron, M. F. Cesbron-Delauw.
1999
. Infection with Toxoplasma gondii RH and Prugniaud strains in mice, rats and nude rats: kinetics of infection in blood and tissues related to pathology in acute and chronic infection.
Pathol. Res. Pract.
195
:
475
-485.
36
Ferguson, D. J., W. M. Hutchison, E. Pettersen.
1989
. Tissue cyst rupture in mice chronically infected with Toxoplasma gondii: an immunocytochemical and ultrastructural study.
Parasitol. Res.
75
:
599
-603.
37
Suzuki, Y..
2002
. Immunopathogenesis of cerebral toxoplasmosis.
J. Infect. Dis.
186
:(Suppl 2):
S234
-S240.
38
Shirahata, T., T. Yamashita, C. Ohta, H. Goto, A. Nakane.
1994
. CD8+ T lymphocytes are the major cell population involved in the early γ interferon response and resistance to acute primary Toxoplasma gondii infection in mice.
Microbiol. Immunol.
38
:
789
-796.
39
Khan, I. A., W. R. Green, L. H. Kasper, K. A. Green, J. D. Schwartzman.
1999
. Immune CD8+ T cells prevent reactivation of Toxoplasma gondii infection in the immunocompromised host.
Infect. Immun.
67
:
5869
-5876.
40
Kasper, L. H., K. M. Currie, M. S. Bradley.
1985
. An unexpected response to vaccination with a purified major membrane tachyzoite antigen (P30) of Toxoplasma gondii.
J. Immunol.
134
:
3426
-3431.
41
Olson, J. K., S. D. Miller.
2004
. Microglia initiate central nervous system innate and adaptive immune responses through multiple TLRs.
J. Immunol.
173
:
3916
-3924.
42
Karman, J., C. Ling, M. Sandor, Z. Fabry.
2004
. Initiation of immune responses in brain is promoted by local dendritic cells.
J. Immunol.
173
:
2353
-2361.
43
van der Most, R. G., K. Murali-Krishna, R. Ahmed.
2003
. Prolonged presence of effector-memory CD8 T cells in the central nervous system after dengue virus encephalitis.
Int. Immunol.
15
:
119
-125.
44
Marten, N. W., S. A. Stohlman, J. Zhou, C. C. Bergmann.
2003
. Kinetics of virus-specific CD8+-T-cell expansion and trafficking following central nervous system infection.
J. Virol.
77
:
2775
-2778.
45
Cabarrocas, J., J. Bauer, E. Piaggio, R. Liblau, H. Lassmann.
2003
. Effective and selective immune surveillance of the brain by MHC class I-restricted cytotoxic T lymphocytes.
Eur. J. Immunol.
33
:
1174
-1182.
46
Lu, F., S. Huang, L. H. Kasper.
2003
. Interleukin-10 and pathogenesis of murine ocular toxoplasmosis.
Infect. Immun.
71
:
7159
-7163.
47
Deckert-Schluter, M., C. Buck, D. Weiner, N. Kaefer, A. Rang, H. Hof, O. D. Wiestler, D. Schluter.
1997
. Interleukin-10 downregulates the intracerebral immune response in chronic Toxoplasma encephalitis.
J. Neuroimmunol.
76
:
167
-176.
48
Gazzinelli, R. T., M. Wysocka, S. Hieny, T. Scharton-Kersten, A. Cheever, R. Kuhn, W. Muller, G. Trinchieri, A. Sher.
1996
. In the absence of endogenous IL-10, mice acutely infected with Toxoplasma gondii succumb to a lethal immune response dependent on CD4+ T cells and accompanied by overproduction of IL-12, IFN-γ and TNF-α.
J. Immunol.
157
:
798
-805.
49
Gerosa, F., C. Nisii, S. Righetti, R. Micciolo, M. Marchesini, A. Cazzadori, G. Trinchieri.
1999
. CD4+ T cell clones producing both interferon-γ and interleukin-10 predominate in bronchoalveolar lavages of active pulmonary tuberculosis patients.
Clin. Immunol.
92
:
224
-234.
50
Plebanski, M., K. L. Flanagan, E. A. Lee, W. H. H. Reece, K. Hart, C. Gelder, G. Gillespie, M. Pinder, A. V. Hill.
1999
. Interleukin 10-mediated immunosuppression by a variant CD4 T cell epitope of Plasmodium falciparum.
Immunity
10
:
651
-660.
51
Belkaid, Y., K. F. Hoffmann, S. Mendez, S. Kamhawi, M. C. Udey, T. A. Wynn, D. L. Sacks.
2001
. The role of interleukin (IL)-10 in the persistence of Leishmania major in the skin after healing and the therapeutic potential of anti-IL-10 receptor antibody for sterile cure.
J. Exp. Med.
194
:
1497
-1506.
52
Pohl-Koppe, A., K. E. Balashov, A. C. Steere, E. L. Logigian, D. A. Hafler.
1998
. Identification of a T cell subset capable of both IFN-γ and IL-10 secretion in patients with chronic Borrelia burgdorferi infection.
J. Immunol.
160
:
1804
-1810.
53
Khan, I. A., T. Matsuura, L. H. Kasper.
1995
. IL-10 mediates immunosuppression following primary infection with Toxoplasma gondii in mice.
Parasite Immunol.
17
:
185
-195.
54
O’Garra, A..
1998
. Cytokines induce the development of functionally heterogeneous T helper cell subsets.
Immunity
8
:
275
-283.
55
Abbas, A. K., K. M. Murphy, A. Sher.
1996
. Functional diversity of helper T lymphocytes.
Nature
383
:
787
-793.
56
Plebanski, M., E. A. Lee, C. M. Hannan, K. L. Flanagan, S. C. Gilbert, M. B. Gravenor, A. V. Hill.
1999
. Altered peptide ligands narrow the repertoire of cellular immune responses by interfering with T-cell priming.
Nat. Med.
5
:
565
-571.
57
Sloan-Lancaster, J., P. M. Allen.
1996
. Altered peptide ligand-induced partial T cell activation: molecular mechanisms and role in T cell biology.
Annu. Rev. Immunol.
14
:
1
-27.
58
Burg, J. L., D. Perelman, L. H. Kasper, P. L. Ware, J. C. Boothroyd.
1988
. Molecular analysis of the gene encoding the major surface antigen of Toxoplasma gondii.
J. Immunol.
141
:
3584
-3591.