BALB/c-background SCID, which lack T cells, and RAG1-knockout (RAG1^−/−) mice, which lack both T and B cells, and wild-type (WT) BALB/c mice were from the Jackson Laboratory (Bar Harbor, ME). BALB/c-background perforin-deficient (Prf1^−/−) mice (14) were originally provided by John T. Harty (University of Iowa) and bred in our animal facility. Female mice were used for all studies. SCID or RAG1^−/− mice were used as recipients of CD8⁺ T cells from WT or Prf1^−/− BALB/c mice. There were three to five mice in each experimental group. Mouse care and experimental procedures were performed under specific pathogen-free conditions in accordance with established institutional guidance and approved protocols from the Institutional Animal Care and Use Committee.

The ME49 strain of T. gondii was maintained by infecting Swiss-Webster mice intraperitoneally with 10 cysts obtained from the brains of chronically infected mice of this strain (3–5). The WT and GRA3-deficient (ΔGRA3) Prugniaud (Pru) strains of the parasite were maintained as tachyzoites in their cultures with monolayers of human foreskin fibroblasts in DMEM medium (Gibco/Thermo Fisher Scientific, Waltham, MA) containing 10% fetal bovine serum (Gibco). Both the ME49 and Pru strains belong to the genotype II of T. gondii.

Mice were infected with 10 cysts (for T-cell donors) or 20 cysts (for T-cell recipients) of the ME 49 strain of T. gondii orally by gavage. SCID and RAG1^−/− mice were treated with sulfadiazine (Sigma-Aldrich, St. Louis, MO) in the drinking water (400 mg/L) beginning at 9 days after infection for the entire periods of the experiments to control proliferation of tachyzoites and establish a chronic infection (5, 15). The Prf1^−/− mice received sulfadiazine in the same manner beginning at 26 days after infection. In a part of the experiments, SCID mice were infected intraperitoneally with 2 × 10³ tachyzoite of the WT or ΔGRA3 Pru strains (16) and treated with sulfadiazine in the drinking water beginning at 7 days after infection to establish chronic infection, and sulfadiazine dose was increased to 1 g/L in drinking water from 13 days after infection to maintain the chronic stage of the infection.

Spleen cells were obtained from WT and Prf1^−/− mice chronically infected with the ME49 strain and suspended in Hanks’ balanced salt solution (Hyclone, Logan, UT) with 2% fetal bovine serum (Sigma). CD8⁺ immune T cells were purified from the spleen cells using MACS with microbeads-conjugated anti-mouse CD8 (53-6.7) monoclonal antibody (Miltenyi Biotec, Auburn, CA) (5, 17). As a control, CD8⁺ normal T cells were purified from the spleens of uninfected WT mice in the same manner. Infected SCID and RAG1^−/− mice received the purified CD8⁺ T cells (2.1–3.5 × 10⁶ cells) intravenously from a tail vein at 3 weeks after infection (5).

At 1 day before (Day −1) or 7 days after (Day 7) the transfer of CD8⁺ normal or immune T cells, half of each brain of infected SCID and RAG1^−/− mice was triturated with a mortar and pestle in 0.5 mL of PBS (3, 5), and the number of cysts in at least three aliquots (20 μl each) of each brain suspension was counted microscopically.

RNA was purified from half of each brain of infected SCID and RAG1^−/− mice obtained at Day −1 or Day 7 of the CD8⁺ T-cell transfer, and the amount of mRNA for 13 secretory molecules (GRA1–8, ROP5, ROP16–18, and ROP35), bradyzoite antigen 1 (BAG1), and bradyzoite-specific surface antigen 2C (SAG2C) was measured by reverse transcription real-time PCR (RT-PCR) using the StepOne Plus real-time PCR system with TaqMan reagents (Allied Biosystems, Norwalk, CT) (5, 15). BAG1 was used as a molecule constitutively expressed only in the bradyzoites stage of T. gondii, and its mRNA levels were used to indicate the cyst burden in the brains of infected mice. SAG2C is a bradyzoite-specific molecule expressed on the surface of the bradyzoites and used as a control molecule for comparison with the secretory GRA and ROP molecules. In infection with the WT Pru and ΔGRA3 strains of T. gondii, mRNA levels for GRA3 were also measured to confirm the absence of GRA3 mRNA in the brains of mice infected with the ΔGRA3 strain. Sequences of the primers and probe for each of these T. gondii molecules are described in Table 1 . Relative expression levels of GRA1–8, ROP5, ROP16–18, ROP35, and SAG2C in the cysts were calculated as ratios of their mRNA levels to BAG1 mRNA levels. The mRNA levels for mouse CD8β and Prf1 were also measured using ready-made primer and probes from Applied Biosystems.

Levels of difference between experimental groups were determined by one-way ANOVA with Tukey’s multiple comparison test or Mann–Whitney test (GraphPad Prism software, version 9.3.1). Levels of significance in correlations between the degree of increased expression levels of selected GRA and ROP molecules and the successful persistence of T. gondii cysts through the condition in which greater portions of cysts from their total populations are eliminated were determined by Pearson test (GraphPad Prism software). The correlations that provided p < 0.05 were considered significant.

SCID mice lacking T cells were infected with T. gondii and treated with sulfadiazine to establish and maintain the tissue cysts of the parasite in their brains (3–5). At 3 weeks after infection, one mouse group received a systemic transfer of CD8⁺ immune T cells (3.5 × 10⁶ cells) purified from the spleens of WT mice chronically infected with the parasite. Another group of mice received CD8⁺ normal T cells from uninfected WT mice as a control in the same manner. Two additional groups of infected SCID mice did not receive any T cells as another control, and cyst numbers and mRNA levels for bradyzoite (cyst)-specific BAG1 in their brains were measured at 1 day before (Day −1) and 7 days after (Day 7) the T-cell transfer. The cyst burden and BAG1 mRNA levels in the brains of the recipients of the CD8⁺ T cells were measured at Day 7. We also performed an independent experiment using RAG1^−/− mice that lack both T and B cells as the recipients of the CD8⁺ T cells (3.2 × 10⁶ cells) from infected and uninfected control WT mice in the same manner.

Cyst numbers in the control group without any T-cell transfer did not differ between Day −1 and Day 7, indicating that their cyst numbers were stable during this time period in the absence of T cells ( Figure 1A , the data from both SCID and RAG1^−/− mice are combined). In contrast, the number of cysts in the recipients of the CD8⁺ immune T cells was 21 times less than that in the control mice with no T-cell transfer at Day 7 (p < 0.001) ( Figure 1A ). Consistently, cerebral mRNA levels for BAG1 in the CD8⁺ immune T-cell recipients were 27 times less than the control mice with no T-cell transfer at Day 7 (p < 0.001) ( Figure 1B , the data from both SCID and RAG1^−/− mice are combined). In contrast, both cyst numbers and BAG1 mRNA levels in the brains of the mice that had received the normal T cells did not differ from those of the control mice without any T-cell transfer ( Figures 1A, B ). Thus, CD8⁺ immune T cells eliminated a majority (95%) of T. gondii cysts from the brains of the recipients within 7 days after the transfer of T cells. In other words, a small population (5%) of cysts had successfully persisted despite the presence of the anti-cyst T-cell immunity that had eliminated a majority (95%) of the cysts.

We also examined whether cerebral BAG1 mRNA levels directly correlate with T. gondii cysts numbers in the brains of these mice. In this analysis, we applied the BAG1 mRNA level and cyst number of each mouse in all of the experimental groups at Day 7 described above. The BAG1 mRNA levels strongly correlated with the cyst numbers (p = 0.0018, Figure 1C ), indicating that cerebral BAG1 mRNA levels are an effective indicator for cyst burdens in the brains of these mice with and without the transfer of CD8⁺ normal and immune T cells.

We next address a possibility that the small portion (5%) of the cysts, which successfully persisted in the presence of CD8⁺ immune T cells, have a unique molecular expression profile when compared to the total cyst population that persisted in the absence of those T cells in the control group. T. gondii has two major groups of secretory proteins, ROPs and GRAs. The ROPs are secreted during an invasion of the parasite into host cells (8), whereas GRAs are secreted after the parasite invaded into host cells to assist the formation of the PV and the cyst wall (9). We examined whether relative mRNA expression levels for eight GRAs (GRA1–8) and five ROPs (ROP5, ROP16–18, and ROP35) differ between the total population of cysts, which existed in the absence of T cells in the control mice, and the small portion of the cysts that successfully persisted for 7 days in the presence of the CD8⁺ immune T cells. Among GRA proteins, GRA1–8 were chosen since they are not only secretory proteins but also known to be present in the cysts and/or the cyst wall detected in the brains of infected mice (18) and their absence often disrupts the cyst wall integrity and/or number of cysts in infected mice (16, 19, 20). Regarding ROP proteins, ROP5, ROP17, ROP18, and ROP35 were chosen, since the absence of any of these molecules results in marked decreases in the number of cysts in infected mice (21). ROP16 was used as a control for the other four ROPs, since a deletion of this molecule rather increases the number of cysts in the brains of infected mice (21). SAG2C was used as a control molecule that is not a secretory protein and expressed on the surface of the bradyzoites present within the cysts (22).

We calculated the relative expression levels of the mRNA levels for each of the eight GRAs, five ROPs, and the control molecule, SAG2C, in the ratios to BAG1 mRNA levels. Since BAG1 mRNA levels indicate cerebral cyst burden in infected mice as shown in Figures 1A–C , the relative expression levels of mRNA for the GRAs and ROPs in ratios to BAG1 mRNA levels indicate relative expression levels of these secretory molecules in the cysts present in the brains of these mice.

The relative expression levels of the control molecule, SAG2C, to BAG1 mRNA levels did not differ between the total population of the cysts present in the absence of T cells and the small portion of the cysts that successfully persisted in the presence of the CD8⁺ immune T cells in both SCID and RAG1^−/− mice ( Figure 1D ). In contrast, relative expression levels of mRNA for four GRAs (GRA1, GRA2, GRA3, and GRA7) and one ROP (ROP35) consistently showed significantly greater mRNA levels in the cysts that successfully persisted in the presence of the CD8⁺ immune T cells than the cysts that persisted in the absence of any T cells in both SCID and RAG1^−/− mice (p < 0.05, p < 0.01, p < 0.001, or p < 0.0001, Figures 1E–I ). In contrast, the relative mRNA expression levels for these five secretory molecules in the cysts that persisted in the recipients of CD8⁺ normal T cells from uninfected mice did not differ from the cysts present in the control mice that did not receive T cells in both SCID and RAG1^−/− mice ( Figures 1E–I ). The transfer of the normal T cells did not induce significant reduction of cyst numbers and BAG1 mRNA levels as described earlier ( Figures 1A, B ). These results indicate that the elimination of a majority of T. gondii cysts by anti-cyst activity of CD8⁺ immune T cells resulted in a selection of the cysts that express significantly increased mRNA levels for at least the five selected secretory molecules, GRA1, GRA2, GRA3, GRA7, and ROP35, suggesting that these GRA and ROP proteins could mediate in the capability of T. gondii cysts to persist in the presence of anti-cyst CD8⁺ T-cell immunity.

We previously demonstrated that perforin is required for the activity of CD8⁺ immune T cells to remove T. gondii cysts from the brains of infected mice (5, 23). Therefore, we examined whether the persistence of the cysts with upregulated expressions of mRNA for GRA1, GRA2, GRA3, GRA7, and ROP35 occur only when the immune T cells express perforin. Infected and sulfadiazine-treated SCID mice received CD8⁺ immune T cells from chronically infected WT and Prf1^−/− mice. As a control, an additional group of infected SCID mice did not receive any T cells.

Seven days after the T-cell transfer, BAG1 mRNA levels in the brains of the recipients of WT CD8⁺ T cells were significantly less than those of the control mice that did not receive T cells (p < 0.01, Figure 2A ), whereas BAG1 mRNA levels in the brains of the mice that received Prf1^−/− CD8⁺ T cells did not differ from those of the control mice with no T-cell transfer ( Figure 2A ). Their cerebral BAG1 mRNA levels significantly correlated with cyst numbers in their brains ( Figure 2B ), which is consistent with the results shown in Figure 1C , confirming that cerebral BAG1 mRNA levels are an effective indicator of the cyst burden in the brain of these mice as well.

We then examined mRNA expression levels for the four selected GRAs (GRA1, GRA2, GRA3, and GRA7), one ROP (ROP35), and one control molecule (SAG2C) of T. gondii molecules between the cysts that persisted in the presence and the absence of the perforin-mediated anti-cyst activity in CD8⁺ immune T cells. The relative expression levels of mRNA for the control molecule, SAG2C, in ratios to BAG1 mRNA levels did not differ between the recipients of WT CD8⁺ immune T cells and the control mice that did not receive T cells ( Figure 2C ). Similarly, the ratios of SAG2C mRNA levels to BAG1 mRNA levels did not differ between the recipients of the Prf1^−/− T cells and the control mice with no T-cell transfer ( Figure 2C ). In contrast, the relative expression levels of mRNA levels for GRA1, GRA2, GRA3, GRA7, and ROP35 in ratios to BAG1 mRNA levels were markedly and significantly greater in the brains of the recipients of WT CD8⁺ immune T cells than the control mice with no T-cell transfer (p < 0.01 for each of these GRA and ROP proteins, Figure 2D ). Notably, the relative expression levels of mRNA levels for these selected GRA and ROP proteins in ratios to BAG1 mRNA levels did not differ between the recipients of Prf1^−/− CD8⁺ T cells and the control mice with no T-cell transfer ( Figure 2D ). These results together indicate that the selection of the cysts with increased mRNA expression for GRA1, GRA2, GRA3, GRA7, and ROP35 occurred when the perforin-mediated anti-cyst activity in CD8⁺ immune T cells is present. Therefore, it would be possible that increased expression of these selected secretory molecules plays important roles in immune evasion of T. gondii cysts from the perforin-mediated protective activity of CD8⁺ T cells for the successful survival of those selected population of T. gondii cysts.

To further address the possibility that increased expressions of GRA1, GRA2, GRA3, GRA7, and ROP35 are involved in the immune evasion of T. gondii cysts to successfully persist in the presence of the perforin-mediated anti-cyst CD8⁺ T-cell immunity, we examined whether the degree of increase in the relative mRNA expression levels for GRA1, GRA2, GRA3, GRA7, and ROP35 in the cysts directly correlates with their capability to successfully persist through the immune environment in which greater portions of cysts are eliminated. We applied the data from SCID and RAG1^−/− mice that received CD8⁺ T cells from uninfected and infected WT mice and infected Prf1^−/− mice shown in Figures 1 and 2 in this analysis. SAG2C was included as a non-secretory control molecule whose expression level does not correlate with the capability of T. gondii cysts to persist in the presence of anti-cyst CD8⁺ immune T cells.

In this analysis, the mean value of the relative mRNA levels of GRA1, GRA2, GRA3, GRA7, and ROP35 in ratios to BAG1 mRNA levels in the control mice that did not receive any T cells was used as the base value, and the fold increases in the relative expression levels of these selected secretory molecules in the cysts that had persisted in the presence of CD8⁺ normal T cells and CD8⁺ immune T cells from infected WT and Prf1^−/− mice were calculated as their ratios to their base values in the control mice with no T-cell transfer. The degree of the decrease in the cyst burden was calculated as the fold decreases in cerebral BAG1 levels in the recipients of the normal T cells and immune T cells from infected WT and Prf1^−/− mice in comparison with the mean value of BAG1 mRNA levels in the control mice with no T-cell transfer.

Greater expression levels of only GRA1, GRA3, and ROP35 in ratios to BAG1 mRNA levels significantly correlated with their persistence through the condition in which greater degrees of reduction of BAG1 mRNA levels occurred in the recipients of the CD8⁺ T cells (p < 0.05 for GRA1 and GRA3, and p < 0.001 for ROP35, Figure 3 ). GRA7 also shows a tendency of the correlation ( Figure 3 ), but the degree of the correlation did not reach significance (p = 0.1074, Figure 3 ). GRA2 did not show any trend of the correlation (p = 0.9426, Figure 3 ). The degree of increased expression of SAG2C mRNA levels did not correlate with the efficiency of cysts to persist in the presence of the T cells as expected (p = 0.5864, Figure 3 ). Thus, GRA1, GRA3, and ROP35 appear to be the most likely candidates for the T. gondii molecules that play important roles in the immune evasion of the cyst stage of the parasite to persist in the presence of anti-cyst CD8⁺ immunity in infected hosts.

Among the three molecules, GRA1, GRA3, and ROP35, that showed significant correlations between the degrees of their increased expressions and increased survival rates of T. gondii cysts in the presence of CD8⁺ T cells in Figure 3 , GRA3 has been shown to be present in both cyst wall and the space within the cysts detected in the brains of infected mice (18, 24), and was suggested to play a structural or organizational role during cyst development or in cyst maintenance (16). Thus, we examined whether a deficiency of GRA3 ablates the capability of T. gondii cysts to persist in the presence of anti-cyst activity of CD8⁺ immune T cells. Two groups of SCID mice were infected with 2 × 10³ tachyzoites of the ΔGRA3 mutant strain, and two additional groups of mice were infected with the WT Pru strain in the same manner. The ΔGRA3 strain was generated from the WT Pru strain (16), which belongs to the genotype II of T. gondii as does the ME49 strain that was used in the studies shown in Figures 1 – 3 . The mice received sulfadiazine treatment beginning at 7 days after infection for the entire period of the study. At 3 weeks after infection, one of the two groups of each of the WT Pru- and ΔGRA3-infected SCID mice received CD8⁺ immune T cells (3 × 10⁶ cells) from WT mice infected with the ME49 strain, and another group of each of the WT Pru- and ΔGRA3-infected SCID mice did not receive any T cells as a control.

The brains of the control group with no T-cell transfer in both the WT Pru- and ΔGRA3-infected SCID mice were obtained at 1 day before the T-cell transfer to indicate the cyst burden at the time of the T-cell transfer. The brains of the mice that received CD8⁺ immune T cells were obtained at 1 week after the T-cell transfer to determine the degrees of reduction of cerebral cyst burden when compared to the time of the T-cell transfer in the control group. Equivalent BAG1 mRNA levels were detected in the brains of the control group without the T-cell transfer in the WT Pru- and ΔGRA3-infected mice ( Figure 4A ). One week after the CD8⁺ immune T-cell transfer, the cerebral BAG1 mRNA levels in the ΔGRA3-infected mice were five times less than those in the control mice at the time of the T-cell transfer (p < 0.05, Figure 4A ). In contrast, the cerebral BAG1 mRNA levels of the WT Pru-infected mice at 7 days after the CD8⁺ T-cell transfer were slightly greater than half of those of the control mice at the time of the T-cell transfer, and the difference did not reach statistical significance (p = 0.208, Figure 4A ). These results indicate that ΔGRA3 T. gondii cysts have a significantly increased susceptibility to anti-cyst activity of CD8⁺ immune T cells for their elimination in the brain of infected mice.

To confirm the deficiency of GRA3 expression in ΔGRA3 T. gondii strain, we measured mRNA levels for GRA3 in the brains of SCID mice infected with the WT Pru and ΔGRA3 strains of the parasite. Large amounts of GRA3 mRNA were detected in the brains of the WT Pru-infected mice in parallel with those of BAG1 mRNA ( Figure 4B ). In the brains of the ΔGRA3-infected mice, GRA3 mRNA was somehow not completely absent but at the levels close to a detection limit (more than 3,500 times less than the levels detected in the WT Pru-infected mice) in the RT-PCR ( Figure 4B ). Even though trace amounts of GRA3 mRNA were detected in ΔGRA3 cysts, the marked reduction of the cyst burden in the ΔGRA3-infected SCID mice after receiving CD8⁺ immune T cells indicates that GRA3 plays an important role in the capability of T. gondii cysts to persist in the presence of anti-cyst CD8⁺ immune T cells.

In contrast to the notable differences in the reduction of BAG1 mRNA levels between the WT Pru- and ΔGRA3-infected SCID mice after receiving CD8⁺ immune T cells, the mRNA levels for CD8β did not differ between these two groups of mice ( Figure 4C ), suggesting that the transferred CD8⁺ immune T cells efficiently migrated into the brains of both WT Pru- and ΔGRA3-infected recipient mice in a similar manner. These results further support that GRA3 plays a critical role in the capability of T. gondii cysts to evade or resist anti-cyst CD8⁺ T-cell immunity and persist in the brains of infected hosts in the presence of those T cells.

The reason for the inefficient removal of the cysts of the WT Pru strain by CD8⁺ immune T cells in this study is unclear. One possible reason is that the CD8⁺ immune T cells were purified from mice infected with the ME49 strain, whereas the recipient SCID mice were infected with the Pru strain. Although both the Pru and ME49 strains belong to the genotype II of T. gondii as mentioned earlier, there could be some variations in the amino acid sequences and structures of a part of the parasite molecules that anti-cyst CD8⁺ T cells recognize, between these two strains of the parasite. These differences may have reduced the efficiencies of the CD8⁺ immune T cells obtained from mice infected with the ME49 strain to eliminate the cysts of the Pru strain.