Kinetics and Phenotype of Vaccine-Induced CD8⁺ T-Cell Responses to Toxoplasma gondii^▿

Mice and infections.

C57BL/6, C57BL/6-Prf1^tm1Sdz/J (22), and B6.MRL-Fas^lpr/J mice (1) were purchased from Jackson Laboratory (Bar Harbor, ME). B6.SJL-Ptprc^a/BoyAiTac (CD45.1) mice were purchased from Taconic. For depletion of CD4⁺ or CD8⁺ T cells, mice were injected twice during the week prior to immunization with 0.5 mg anti-CD4 (GK1.5), anti-CD8 (2.43), or control antibody (rat immunoglobulin G2b). Mice were maintained under specific-pathogen-free conditions, and all animal work was done in accordance with the Institutional Animal Care and Use Guidelines of the University of Pennsylvania. For all experiments using CPS-OVA and Pru-OVA, mice were injected intraperitoneally with 10⁵ parasites. For challenge experiments, mice were given 10³ RH parental or RH-OVA (RH expressing secreted OVA) tachyzoites (29). Tachyzoites were grown in human foreskin fibroblast monolayers in Dulbecco modified Eagle medium containing 1% (CPS and RH parasites) or 10% (Pru parasites) fetal calf serum and 1% penicillin-streptomycin. CPS parasites were cultured in medium containing 0.2 mM uracil, and OVA-transgenic parasites were maintained in medium containing 20 μM chloramphenicol.

In vitro T-cell responses.

Spleen cells, lymph node cells, and peritoneal exudate cells (PECs) were harvested and dissociated into single-cell suspension in complete RPMI 1640 (Gibco/Invitrogen, Carlsbad, CA) containing 10% heat-inactivated fetal calf serum, 10 U/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, 25 mM HEPES, 0.1 mM nonessential amino acids, and 50 μM 2-mercaptoethanol. Spleen and lymph node cells (5 × 10⁵/well) and PECs (10⁵/well) were plated out in 96-well round-bottom plates (Costar, Carlsbad, CA) and cultured to assess cytokine production. Cells were restimulated with anti-CD3/CD28 (1 μg/ml each), soluble Toxoplasma antigen (25 μg/ml), 500 μg/ml OVA (Worthington Biochemical Corporation, Lakewood, NJ), or 1 μg/ml SIINFEKL peptide (CHI Scientific, Maynard, MA). Supernatants were removed after 40 h and assayed for the production of IFN-γ by enzyme-linked immunosorbent assay.

Flow cytometry.

For identification of OVA-specific T-cell responses, single-cell suspensions were washed in fluorescence-activated cell sorter buffer (phosphate-buffered saline, 2 mM EDTA, and 2% bovine serum albumin) and incubated for 15 to 20 min with Fc block (fluorescence-activated cell sorter buffer containing 1 μg/ml 2.4G2 from BD Pharmingen and 1 μg/ml rat and mouse immunoglobulin G from Caltag/Invitrogen). Cells were stained with OTI-SIINFEKL tetramer conjugated to phycoerythrin (PE) or allophycocyanin (APC) (from Beckman Coulter or a generous gift from the Wherry lab at Wistar Institute) for 25 min at room temperature, washed once, and then stained for other surface markers for 15 min at 4°C. The following monoclonal antibodies were used: CD8 (conjugated to fluorescein isothiocyanate, peridinin chlorophyll protein, or APC), CD45.1-PE, CD127-biotin, CD122-biotin, and CD62L-APC (BD Biosciences); tumor necrosis factor alpha-APC, IFN-γ-APC, IFN-γ-PE, KLRG1-APC, and streptavidin-APC (eBioscience, San Diego, CA); and anti-human granzyme B-APC (Caltag, Carlsbad, CA).

For intracellular cytokine analysis, splenocytes or PECs were incubated (with cell and antigen concentrations as noted above) for 5.5 h total, with the addition of 10 μg/ml brefeldin A (Sigma) for the final 4 h. Cells were first stained for surface markers, followed by fixation overnight with 2% paraformaldehyde (Electron Microscopy Sciences). Cells were permeabilized with 0.1% saponin and then stained for intracellular cytokines for 1 h at 4°C. Flow cytometry samples were collected on a FACSCalibur or FACSCanto machine (BD) and analyzed with FlowJo software (Tree Star Inc. Ashland, OR).

In vivo CTL assay.

Cytolytic activity was assessed using the in vivo cytotoxic T-lymphocyte (CTL) assay, modified slightly from previous protocols (3). Briefly, spleen and lymph node cells from CD45.1 mice were pooled and pulsed with 1 μg/ml OTI peptide (CHI Scientific) for 1 h at 37°C. Cells were washed extensively in phosphate-buffered saline, labeled with 5(6)-carboxyfluorescein diacetate N-succinimidyl ester (Molecular Probes/Invitrogen) at a concentration of 5 μM (pulsed with OTI peptide) or 0.1 μM (unpulsed), and then counted and resuspended at a 1:1 ratio. A total of 6 × 10⁶ cells were then transferred intravenously to anesthetized recipient mice. Mice were sacrificed 4 h later, and spleens were analyzed for specific lysis of the peptide-pulsed population by gating on CD45.1⁺ donor cells. Specific lysis was calculated as described previously (3).

Detection of parasite DNA.

Tissue samples were taken from spleen, liver, and peritoneal exudate at various time points following immunization, and DNA was extracted using the High Pure PCR template preparation kit (Roche). For measurement of parasite burden, the 35-fold repetitive T. gondii B1 gene was amplified by real-time PCR with SYBR green PCR master mix (Applied Biosystems, Foster City, CA) in an AB7500 fast real-time PCR machine (Applied Biosystems) using published conditions (36).

Statistics.

Absolute numbers and frequencies of tetramer-positive cells were statistically analyzed using a Student t test. Where appropriate, data are shown as means ± standard deviations.

T cells are required for CPS-OVA-induced protective immunity.

Earlier work demonstrated that cps1-1 parasites could provide protective immunity to challenge with the virulent RH strain of T. gondii (13). However, these studies did not provide an analysis of the adaptive immune response to immunization or indicate which cell types were important for protection. In order to define which T-cell subsets were required for CPS-induced protective immunity, mice were injected intraperitoneally with a single dose of 10⁵ CPS parasites that expressed the model antigen OVA and then challenged 30 days later. Similar to what was seen previously with the parental (non-OVA-expressing) strain, unimmunized mice succumbed to RH-OVA infection by day 12, while mice immunized with CPS-OVA survived this challenge (Fig. 1). One week prior to RH-OVA challenge, immunized mice were depleted of either CD4⁺ or CD8⁺ T cells. As shown in Fig. 1, depletion of CD8⁺ T cells resulted in increased susceptibility to RH-OVA challenge. In contrast, while there was some variation between experiments, CD4⁺ depletion had only a modest effect on the survival of challenged mice. Thus, in this model of immunization with a single dose of 10⁵ CPS-OVA parasites, CD8⁺ T cells appeared to be the most important subset required for protection following a lethal RH-OVA challenge. It is also worth noting that CPS-OVA immunization was also able to provide resistance to challenge with the parental RH strain (Fig. 1), indicating that CPS-OVA can induce a protective polyclonal response to T. gondii.

Immunization with a replication-deficient parasite induces a robust endogenous CD8⁺ T-cell response and upregulation of activation markers.

Because the parasite used in these experiments expressed the model antigen OVA, a robust endogenous CD8⁺ T-cell response that could be tracked using an OVA-specific class I tetramer was induced (Fig. 2A [day 7 data are shown]). These experiments revealed that OVA-specific cells could be readily detected in the spleen as early as 5 days following immunization (Fig. 2B). Analysis of the kinetics indicated that the peak of this response occurred at 10 days postinfection, after which the response declined, though low frequencies of OVA-specific cells were maintained (Fig. 2B, top panel). A similar pattern of expansion and contraction of OVA-specific CD8⁺ T cells was also seen in the PECs, the site of infection (Fig. 2B, bottom panel). In both the spleen cells and PECs, the total number of OVA-specific CD8⁺ T cells followed a similar pattern of expansion and contraction (data not shown). Splenocytes and PECs were stimulated directly ex vivo with OTI peptide for 5 h, revealing antigen-specific IFN-γ production following intracellular cytokine staining (Fig. 2C, day 7). Further, antigen-specific IFN-γ production was detected in the supernatant following a 40-hour restimulation of splenocytes with OTI peptide, whole OVA protein, or soluble Toxoplasma antigens (Fig. 2D).

To examine the expression of phenotypic markers associated with activation and effector capacity, cells were stained directly ex vivo from the spleen and peritoneal exudate for tetramer binding as well as phenotypic markers, including CD62L, KLRG1, CD122, and CD127. These particular markers were chosen because they have been used to discriminate between naïve, effector, and memory subpopulations in a variety of viral and bacterial systems (21). However, antigen-specific responses to T. gondii in particular have not been extensively characterized. We found that at day 7, the majority of tetramer-positive cells in the spleen (Fig. 3) as well as the PECs (data not shown) had upregulated KLRG1 and had also downregulated CD62L, thus exhibiting an effector phenotype similar to what has been noted in other infection models. Here, OVA-specific cells are compared to the tetramer-negative CD8⁺ T-cell population, which contains activated cells but at a very low frequency compared with the immunodominant tetramer-positive population. Antigen-specific cells also showed higher expression of the IL-2/IL-15 receptor β (IL-2/IL-15Rβ) (CD122) than tetramer-negative cells (Fig. 3). Consistent with recent publications (21), at 7 days following immunization expression of IL-7Rα was decreased on tetramer-positive cells compared to CD3⁺ CD8⁺ tetramer-negative cells, a population that would include primarily naïve cells but also a small population of activated cells (Fig. 3).

To assess the phenotype of the memory cell population, later time points were also examined. Parasite DNA was not detected in the PECs or spleen at 3 weeks following CPS infection, and it was difficult to consistently detect parasite DNA even at 10 days following immunization (Fig. 3B and data not shown); thus, we decided to investigate the phenotype of antigen-specific cells at 30 days after CPS-OVA immunization. At this time point, and as late as 75 days following immunization, the tetramer-positive cells maintained their KLRG1^hi CD62L^lo IL-7Rα^hi effector-like phenotype (data not shown). Overall, these experiments demonstrated that antigen-specific cells at the site of infection as well as in secondary lymphoid tissues resembled an effector population up to 75 days following immunization.

As data emerged from these studies, it became apparent that the kinetics of the CD8⁺ T-cell response to CPS-OVA differ from what has been reported for replication-sufficient strains of T. gondii. While the antigen-specific response induced by CPS-OVA peaks at approximately 10 days following immunization in both the spleen and PECs (Fig. 2), work from other groups has demonstrated that CD8⁺ T-cell responses induced by replicating parasites are not detectable until 2 weeks following peroral infection (24, 26). To directly compare the development of antigen-specific responses, mice were given the same dose of 10⁵ CPS-OVA or Pru-OVA tachyzoites intraperitoneally, and tetramer responses were assessed at two time points. In this comparison, CPS-OVA induced significantly higher CD8⁺ T-cell responses in the spleen as well as the PECs at 8 days following infection (Fig. 4; Table 1). However, by day 14 following infection, the pattern had reversed such that Pru-OVA responses were now significantly higher in the spleen with a trend toward higher responses in the PECs, while CPS-OVA CD8⁺ T-cell responses had started to contract (Table 1). CPS-OVA and Pru-OVA are derived from different strains of T. gondii and have been previously shown to induce disparate responses (15). However, work from our laboratories (E. D. Tait et al., unpublished data) has compared replicating and nonreplicating type 1 strains and demonstrated that CPS-OVA did induce higher frequencies of antigen-specific CD8⁺ T cells than its replication-sufficient parental strain, RH-OVA. These data must be interpreted with care, but the same construct to drive production of OVA was used in all strains, and these parasites appear to express similar levels of OVA. Both CPS-OVA and Pru-OVA generated antigen-specific splenocytes with a similar effector-like phenotype at these time points, expressing low levels of CD62L and CD127 and high levels of CD44, KLRG1, and CD122 (data not shown).

OVA-specific antigen-specific CD8⁺ T cells are cytotoxic.

In addition to production of cytokines, another important function of antigen-specific CD8⁺ T cells is the lysis of infected target cells. Indeed, at 1 week following immunization, many tetramer-positive splenocytes had upregulated expression of granzyme B, consistent with possible cytolytic activity (Fig. 5A, day 7). While a proportion of the total CD8⁺ tetramer-negative population had also upregulated granzyme B (5.5% ± 1.6%), this frequency was much lower than that for the tetramer-positive population (89.8% ± 0.7%). Two weeks following immunization, a lower proportion (40.9% ± 3.7%) of antigen-specific cells showed increased expression of this cytolytic effector molecule directly ex vivo (Fig. 5A, day 15). To directly assess the cytolytic ability of these cells, an in vivo CTL assay was modified based on earlier work (3). Briefly, unpulsed or OTI peptide-pulsed cells from congenic CD45.1 mice were CFSE labeled, mixed at a 1:1 ratio, and injected into immunized (CD45.2) mice. When immunized recipient mice were analyzed, loss of the OTI-pulsed CFSE^hi population indicated that CPS-OVA induced robust cytolytic activity in CD8⁺ T cells (Fig. 5B). While cytolytic activity could be detected as soon as 4 h after target cell transfer in the recipient spleens, maximum lysis was achieved after a 16-hour incubation of the target cells in the recipient mice (Fig. 5C). When multiple experiments were combined from various time points from 7 to 30 days following immunization, there was a correlation between the level of cytotoxic activity and the frequency of antigen-specific cells present at the time of the CTL assay (Fig. 5D) (all mice were analyzed at 4 h posttransfer of target cells).

To better understand the requirements for cytotoxic activity, WT and perforin-deficient mice were immunized with a single dose of 10⁵ CPS-OVA. As seen in other infection models (2), perforin-deficient mice generated increased frequencies of antigen-specific CD8⁺ T cells (Fig. 6A). Despite this phenotype, the perforin-deficient mice demonstrated significantly lower levels of cytolytic activity against peptide-pulsed target cells when analyzed at 4 or 16 h following target cell transfer (Fig. 6B). In addition to granule-mediated cytotoxicity, CD8⁺ T cells can also mediate cytolysis via Fas-FasL interactions (4). To examine perforin-independent pathways of cytotoxicity in this model, Fas knockout (KO) peptide-pulsed target cells were mixed with WT target cells and unpulsed cells and transferred into immunized mice. WT and Fas KO target cells were killed equally well, indicating that the cytotoxic activity generated in response to immunization is not mediated via the Fas-FasL pathway (data not shown). Together, these findings indicate that immunization induces a perforin-dependent pathway but that a perforin-independent pathway of cytolysis also exists.

Depletion of CD4⁺ T cells abrogates effector CD8⁺ T-cell response to CPS-OVA.

Different models of infection have shown diverse requirements for CD4 help in the generation of CD8⁺ T cells. Following infection with a replicating strain of T. gondii, CD4⁺ T-cell help is not required for the generation of CD8⁺ effector responses in the spleen (26). To address whether CD4 help was required in the CPS-OVA experimental system, mice were depleted of CD4⁺ T cells 1 week prior to immunization with two doses of 0.5 mg anti-CD4 antibody. At 7 days following immunization, CD4-depleted mice demonstrated a decrease in the frequency (Fig. 7A) as well as total numbers (Fig. 7B) of OVA-specific CD8⁺ T cells that were generated. However, the tetramer-positive T cells generated in the absence of CD4⁺ help were still capable of making IFN-γ in response to OTI peptide stimulation (data not shown). Expression of KLRG1 on antigen-specific CD8⁺ T cells generated in the absence of CD4⁺ T-cell help was decreased (Fig. 7C) (76.3% ± 3.2%, versus 21.77% ± 4% for CD4-depleted mice; P = 0.0004). Additionally, tetramer-positive cells generated without CD4⁺ help were capable of killing peptide-pulsed target cells as demonstrated by an in vivo CTL assay, though the rates of cytolysis were decreased (Fig. 7D). Based on the results shown in Fig. 4, it seems likely that the decreased levels of cytotoxicity seen in the absence of CD4 help are due to lower overall OVA-specific cell numbers, since these cells were capable of expressing granzyme B (data not shown).