Role of CD4⁺ Foxp3⁺ Regulatory T Cells in Protection Induced by a Live Attenuated, Replicating Type I Vaccine Strain of Toxoplasma gondii

Abstract

Vaccination with the live attenuated Toxoplasma gondii Mic1.3KO strain induced long-lasting immunity against challenge with Toxoplasma gondii type I and type II strains. The involvement of regulatory T cells (Tregs) in the protection mechanism was investigated. Intraperitoneal injection of Mic1.3KO induced a weak and transient influx of CD4⁺ Foxp3⁺ T regulatory cells followed by recruitment/expansion of CD4⁺ Foxp3⁻ CD25⁺ effector cells and control of the parasite at the site of infection. The local and systemic cytokine responses associated with this recruitment of Tregs were of the TH1/Treg-like type. In contrast, injection of RH, the wild-type strain from which the vaccinal strain is derived, induced a low CD4⁺ Foxp3⁺ cell influx and uncontrolled multiplication of the parasites at this local site, followed by death of the mice. The associated local and systemic cytokine responses were of the TH1/TH17-like type. In addition, in vivo Treg induction in RH-infected mice with interleukin-2 (IL-2)/anti-IL-2 complexes induced control of the parasite and a TH1/Treg cytokine response similar to the response after Mic1.3KO vaccination. These results suggest that Tregs may contribute to the protective response after vaccination with Mic1.3KO.

INTRODUCTION

Toxoplasmosis is caused by a protozoan parasite that infects humans and other warm-blooded animals. Infections in humans are generally asymptomatic, although immunosuppressed patients may exhibit severe symptoms. Similarly, primary infection during pregnancy can lead to miscarriage and neonatal malformations. Toxoplasmosis can be transmitted to humans via ingestion of oocysts or via the consumption of meat products contaminated with tissue cysts (1). Effective vaccination of domestic livestock can therefore prevent human infection with Toxoplasma gondii. It is possible to induce strong protection by immunization with a live attenuated strain (2). Live attenuated vaccine strain models are also useful for advancing the understanding of the protective host immune response (3–8). However, the information available about the mechanism of protection involved after vaccination with a type I attenuated strain was obtained with nonreplicating, nonpersistent strains, such as strains cps-1 (3–7) and ts4 (8). A strain known as Mic1.3KO was obtained in our laboratory by deleting the MIC1 and MIC3 genes (9). This strain was derived from the highly virulent type I RH strain, and its reduced invasion capacity in vitro was correlated with decreased virulence when injected into outbred Swiss OF1 mice (9). Type I strains are characterized by the rapid dissemination of the parasite and by a high parasite burden that results in death soon after infection by a single viable parasite in mice (10). High levels of gamma interferon (IFN-γ) were produced following infection with a type I parasite, and mice succumbed to uncontrolled parasite growth and associated inflammation (11, 12). The Mic1.3KO strain with the MIC1 and MIC3 gene deletions showed reduced virulence, lower levels of dissemination throughout tissues, and lower levels of IFN-γ production than the parental RH strain after injection into mice (13).

In a model of lethal toxoplasmosis induced after oral administration of infection with a type II strain causing infection in C57BL/6 mice, overproduction of IFN-γ was also responsible for mortality and was correlated with a sharp decline in the percentage of regulatory T cells (Tregs) just before death, supporting the hypothesis of defective immunoregulation (14). Tregs are a subpopulation of CD4⁺ T cells, and their main function is to maintain immune homeostasis and tolerance (15). They constitutively express the interleukin-2 (IL-2) receptor alpha chain (IL2Rα), a surface receptor also known as CD25, and the intracellular fork head box-p3 transcription factor (Foxp-3) marker (16). The role of Tregs after infection with type II strains has been fully described (14, 17–21), and Tregs have been clearly implicated in the mortality of C57BL/6 mice after oral infection in the lethal ileitis model (14). The collapse of Tregs is correlated with pathogenicity and occurs only under highly pathogenic conditions since oral infection of BALB/c mice with a type II strain did not induce a reduction in the levels of Tregs (14). However, depletion of Tregs in these mice resulted in morbidity associated with a high parasite burden and increased ileal pathology compared to that in control BALB/c mice (19), suggesting a role of Tregs in protection during acute infection.

In the present study, we compared the involvement of Tregs after infection with the vaccinal Mic1.3KO strain with that after infection with the parental lethal RH strain in an attempt to identify their involvement in the protection induced by vaccination. We showed a small increase in the absolute CD4⁺ Foxp3⁺ Treg count at the site of infection with Mic1.3KO followed by an increase in CD4⁺ CD25⁺ Foxp3⁻ effector cells and control of the parasite. In contrast, the increase in the Treg count in RH-infected mice was lower and the parasites were not controlled locally. Our studies show that CD4⁺ Foxp3⁺ Tregs are involved in protection, since specific expansion of these cells using IL-2/anti-IL-2 complexes in mice infected with RH induced a reduction in the parasite burden and a decrease in proinflammatory cytokine levels. These features were similar to those of mice infected with Mic1.3KO, supporting the role of CD4⁺ Foxp3⁺ Tregs in the protection induced by vaccination.

MATERIALS AND METHODS

Animals and parasites.

Eight-week-old female Swiss OF1 and C57BL/6 mice were obtained from Janvier (France). All experiments using animals were approved by the local ethics committee (CEEA VdL) and registered under reference number 2011-06-6.

Two weeks before injection, RH tachyzoites harvested from the peritoneal cavity of the mice were cultured on human foreskin fibroblast (HFF) monolayers (ATCC CRL-1634; American Type Culture Collection), as previously described (13). Mic1.3KO parasites were obtained by targeted disruption of the MIC1 and MIC3 genes in the ΔHX RH strain of T. gondii, as previously described (9). Mic1.3KO tachyzoites were also propagated by serial passage on HFF monolayers. Tachyzoites freshly harvested from a cell culture were counted using a Malassez counting chamber.

Mice were infected by intraperitoneal (i.p.) inoculation of 100 freshly harvested tachyzoites diluted in 0.2 ml Dulbecco modified Eagle saline.

Cell culture conditions and cytokine quantification.

For cytokine detection, splenocytes were recovered and purified as described previously (13, 21) and stimulated for 72 h with 10 μg/ml Toxoplasma extract (TE) or with purified anti-CD3 (clone 145-2C11) (eBioscience) at 1 μg/ml. The cells (5 × 10⁵) were seeded into 24-well plates in 1 ml RPMI 1640 containing 5% fetal calf serum (FCS), and supernatants were collected at 24, 48, and 72 h after activation.

Peritoneal exudate cells (PECs) were obtained from uninfected mice (day 0 controls) and from infected mice at 4, 7, and 11 days postinfection (dpi) by peritoneal lavage with 5 ml of ice-cold phosphate-buffered saline (PBS), as already described (13), and centrifuged to collect the supernatants. The supernatants were kept frozen at −20°C until assayed for cytokines. The cytokines (IFN-γ, IL-2, IL-10, IL12p40, IL12p70, and IL-23) and chemokines (CCL2, CCL3, and CCL20) in the serum, cell culture supernatants, and peritoneal washes were quantified by enzyme-linked immunosorbent assay (ELISA) using a Ready-Set-Go kit (eBioscience) and R/D Duoset kits.

Cell surface staining and intracellular staining.

PECs and splenocytes were obtained from mice on day 0 (noninfected controls) and on 4, 7, and 11 dpi. Cells were washed once in PBS and counted to determine total viable cell numbers by trypan blue exclusion. Standard procedures were used to stain 2 × 10⁵ to 5 × 10⁵ PECs and 10⁶ splenocytes in 5% FCS in PBS and mouse Fc block (BD Biosciences) as previously described (21). Antibodies for the detection of CD4 (GK1.5), CD8 (eBioH35-17.2), CD25 (PC61), CD69 (H1.2F3), Ki-67 (SolA15), and Foxp3 (FJK-16S) were purchased from eBioscience.

For IL-10 detection, splenocytes were seeded into 6-well plates at 10⁶/ml in a final volume of 5 ml RPMI 1640 containing 5% FCS and stimulated for 18 h with 10 μg/ml TE. The culture medium was then removed and replaced by fresh culture medium containing phorbol myristate acetate (50 ng/ml), ionomycin (1 μg/ml), and brefeldin A (5 μg/ml) for a further 4-h period. Antibodies for the detection of IL-10 were purchased from BD Biosciences.

Cell acquisition was undertaken with a BD FACSCalibur cytometer, and cells were analyzed using CellQuest software (BD Bioscience).

In vivo Treg expansion.

IL-2/anti-IL-2 complexes were prepared as described in the literature (14, 20) with minor modifications. Recombinant IL-2 and anti-IL-2 monoclonal antibody (MAb; clone JES6-1A12) were obtained from eBioscience. IL-2 (1 μg) was mixed with the anti-IL-2 MAb (10 μg), and the mixture was incubated for 15 min at room temperature prior to i.p. injection at 0, 3, and 5 dpi.

Statistical analyses.

Differences between the groups were compared using the Mann-Whitney or Kruskal-Wallis nonparametric test followed by Dunn's posttest using GraphPad Prism software. All statistical tests were two-sided, and a P value of <0.05 was considered statistically significant in all tests. Some analyses were performed with StatXact software (Cytel Studio) using nonparametric exact tests with strata to take into account variability between repetitions of the same experiment.

All data in the graphs are expressed as the median plus the range (unless otherwise specified).

RESULTS

Lack of control of the parasite was not associated with decreased IFN-γ production in mice infected with RH.

The parasite burden in the peritoneum was monitored by direct counting on days 4, 7, and 11 after inoculation of 100 tachyzoites of both strains (Fig. 1A). On day 4, few parasites (≤0.02 × 10⁶) were found in either group, and the difference between the two groups was not significant. However, on day 7 parasite counts were significantly increased in the group infected with strain RH (referred to here as the RH group; 16 × 10⁶) but not in the group infected with strain Mic1.3KO (referred to here as the Mic1.3KO group; 0.002 × 10⁶). In the Mic1.3KO group, the parasite counts remained very low even at 11 dpi, and the number of parasites decreased between days 4 and 11. Taken together, these results suggest that after day 4 the parasite was being controlled at the local level in the Mic1.3KO-infected mice but not in the RH-infected mice, where the parasite count increased exponentially until the animals died. This high parasite count at 7 dpi was correlated with a high cell mortality rate (between 25 and 90%, as determined by trypan blue exclusion) for PECs recovered from the RH mice. Similar results were obtained when using inbred C57BL/6 mice, in which lower parasite counts were found in the Mic1.3KO group than in the RH group (see Table S1 in the supplemental material), suggesting that these mice were as resistant to the infection with the vaccinal strain as outbred mice.

FIG 1.

A lack of parasite control in the RH group is not correlated with low levels of IFN-γ secretion. (A) Parasites were counted directly after peritoneal lavage at the indicated times postinfection with 100 RH or Mic1.3KO tachyzoites. Results are expressed as the median plus range. Cumulative data from four different experiments are shown (4 to 6 mice per group per time point per experiment). (B) Sera were recovered at the indicated times postinfection, and the IFN-γ levels were quantified. Results are representative of those from three independent experiments with 5 or 6 mice per group per time point in which mice were injected with 100 tachyzoites of RH or Mic1.3KO. **, P < 0.01 using the Kruskal-Wallis test followed by Dunn's posttest; ***, P < 0.001 using the Kruskal-Wallis test followed by Dunn's posttest.

We also measured the serum levels of systemic IFN-γ at 4, 7, and 11 dpi (Fig. 1B). In control naive mice, the level of IFN-γ was below the detection threshold (data not shown). The serum levels of IFN-γ rose significantly following infection. In the RH group, the level of IFN-γ increased significantly between days 4 and 7 and paralleled the level of parasite multiplication. In the Mic1.3KO group, levels also increased between days 4 and 7 and subsequently remained at similar levels until day 11. Taken together, these results suggest that mortality in the RH group may not have originated from a lack of IFN-γ production but, as for type II-infected C57BL/6 mice, may have originated from uncontrolled IFN-γ production (14) or a lack of parasite control, or both.

The total numbers of PECs and splenocytes were checked at 4, 7, and 11 dpi and compared to those of naive mice in response to Mic1.3KO and to RH (data not shown). Both infected groups displayed a slightly higher total PEC count after infection than the control naive group. This difference was significant only at day 11 for the Mic1.3KO-injected group. At 4 dpi, the number of spleen cells increased in both infected groups compared to the number in the naive group, although the difference was significant only for the RH group. In the Mic1.3KO group, this increase was significant at 7 and 11 dpi. At 7 dpi, there were significantly higher spleen cell counts in the Mic1.3KO group than in the control and RH groups. In the RH group, the cell counts were also significantly lower than those at 4 dpi. This suggests that a significant immune response was induced on day 4 after infection in both infected groups.

Kinetics of CD4⁺ Foxp3⁺ Tregs after infection.

After infection with a lethal dose of a type II strain of T. gondii, the numbers and frequencies of CD4⁺ Foxp3⁺ Tregs have been reported to be reduced at the site of infection and systemically in C57BL/6 mice, and this decrease is correlated with high levels of IFN-γ production (14). We therefore followed the CD4⁺ Foxp3⁺ Tregs after infection at local (peritoneum) and systemic (spleen) sites. Three populations (CD25⁺ Foxp3⁺, CD25⁻ Foxp3⁺, and CD25⁺ Foxp3⁻) were observed within the CD4⁺ gated population in both groups. The percentage of total Foxp3⁺ cells in the peritoneum at 4 dpi was slightly but nonsignificantly higher in the RH group than in the naive mice (Fig. 2A). Moreover, there was no significant difference in the numbers (data not shown) or percentages of Foxp3⁺ cells (Fig. 2A) between the two infected groups at this time. The total percentage of Foxp3⁺ cells at 4 dpi was significantly greater in the Mic1.3KO group than in the naive group (Fig. 2A). The kinetics of both effector (CD4⁺ Foxp3⁻ CD25⁺) and regulator (CD4⁺ Foxp3⁺) cells were followed until 11 dpi in the Mic1.3KO group (Fig. 2B and C). This analysis could not be performed for the RH group because there were high levels of cell mortality at 7 dpi, and the permeabilization step further increased cell mortality to 100% (data not shown). The relative number of regulatory Foxp3⁺ cells at day 4 was significantly higher in the Mic1.3KO group than in the naive group (Fig. 2B) (25.9 × 10⁴ versus 1.8 × 10⁴ for the naive group). This increase was transient, since there was no significant difference between the groups at 7 and 11 dpi.

FIG 2.

CD4⁺ Foxp3⁺ Tregs were recruited/expanded before CD4⁺ Foxp3⁻ CD25⁺ T effector cells at the local site after infection with Mic1.3KO but not at the systemic level. PECs recovered at the indicated times postinfection and splenocytes recovered at 7 dpi were stained with anti-CD4 anti-CD25 and anti-Foxp3 MAbs and analyzed by flow cytometry. (A) Percentage of the total Foxp3⁺ population among CD4⁺ cells recovered at 4 dpi. *, P < 0.05 using the Kruskal-Wallis test. (B and C) For calculation of the count, the absolute number of lymphocytes was evaluated by multiplying the total PEC count by the percentage of the lymphocyte gate compared to the whole population. Tregs were defined as Foxp3⁺ CD25⁺ and Fopx3⁺ CD25⁻ cells within the CD4⁺ population (B), and T effector cells (T eff) were defined as Foxp3⁻ CD25⁺ cells within the CD4⁺ population (C). The data presented are representative of those from two independent experiments with 5 mice per group per injection time point. *, P < 0.05 using the Kruskal-Wallis test followed by Dunn's posttest; **, P < 0.01 using the Kruskal-Wallis test followed by Dunn's posttest. (D) Percentage of the total Foxp3⁺ and CD25⁺ Foxp3⁺population within CD4⁺ splenocytes at 7 dpi with RH or Mic1.3KO tachyzoites. Results are representative of those from four independent experiments with 4 to 6 mice per group. *, P < 0.05 using the Kruskall-Wallis test followed by Dunn's posttest. (E) Ratio of the percentage of Foxp3⁺ cells expressing CD25 compared to the total percentage of Foxp3⁺ cells. Results are cumulative data from 4 different experiments. *, P < 0.05 using the Kruskall-Wallis test followed by Dunn's posttest.

Similar results were obtained using C57BL/6 mice (see Fig. S1 in the supplemental material). C57BL/6 mice infected with the RH strain showed reduced percentages of Foxp3⁺ cells at the local level compared to those in the mice infected with strain Mic1.3KO. The percentages of Foxp3⁺ cells in the spleens of the two groups were similar.

The numbers and percentages of effector cells at day 4 were slightly and nonsignificantly higher in the Mic1.3KO group than in the naive mice (Fig. 2C). However, the numbers of CD4⁺ Foxp3⁻ CD25⁺ cells increased exponentially between days 0 and 11 (Fig. 2C) and at 7 and 11 dpi were significantly different from those at day 0. This high CD4⁺ Foxp3⁻ CD25⁺ cell count at the local level was correlated with low parasite counts in Mic1.3KO-infected mice.

The total percentage (Fig. 2D) and absolute number of the Foxp3⁺ population in the spleen (data not shown) did not differ significantly between the two groups at 7 dpi. However, significantly higher percentages of CD25⁺ Foxp3⁺ cells were found in the RH group. The ratio of the percentage of Foxp3⁺-expressing CD25 cells to the total percentage of Foxp3⁺ cells was similar between the naive and the Mic1.3KO groups (0.61 and 0.60, respectively) but was significantly higher for the RH group (about 0.71) than for both the naive and the Mic1.3KO groups (Fig. 2E).

We investigated the expression of phenotypic markers associated with activation (CD69) and proliferation (Ki-67). Cells were analyzed directly ex vivo from the peritoneal cavity and spleen by flow cytometry on day 7 (Tables 1 and 2).

TABLE 1.

PECs from the RH-infected group of mice presenting an activated phenotype after infection^b

Group	% of cells that were:
	CD4+	CD8+	CD69+	Ki-67+
Naive	8.1 ± 4.1	0.82 ± 0.74	5.8 ± 4.2	16.4 ± 3.4
RH	15.4 ± 2.4	2.8 ± 1.1	19.9 ± 5.1a	62.8 ± 3.5a
Mic1.3KO	16.1 ± 3.1	3.3 ± 2.5	11.0 ± 3.3	52.7 ± 7a

^aThe difference between the infected group and the control group using the Kruskal-Wallis test followed by Dunn's posttest was significant (P < 0.05).

^bPECs obtained at 7 dpi were stained with different MAbs, as indicated, and analyzed by cytometry. Results are presented as means ± standard deviations and are representative of those from two independent experiments with 4 to 6 mice per group.

TABLE 2.

Splenocytes of the RH-infected group of mice presenting an activated phenotype after infection^c

Group	% of cells that were:
	CD69+	CD25+	CD4+ CD25+	CD8+ CD25+	Ki-67+
Naive	3.7 ± 2.7	3.8 ± 0.6	2.8 ± 0.6	0.9 ± 0.6	14.0 ± 5.5
RH	11.8 ± 5a,b	6.5 ± 0.8b	3.4 ± 1.1b	1.5 ± 0.6	24.3 ± 3.4
Mic1.3KO	1.9 ± 1.2	2.0 ± 0.7	1.2 ± 0.7	0.8 ± 0.6	37.6 ± 4.9a

^aThe difference between the infected groups and the control group using the Kruskal-Wallis test was significant (P < 0.05).

^bThe difference between the two infected groups using the Kruskal-Wallis test followed by Dunn's posttest was significant (P < 0.05).

^cSplenocytes obtained at 7 dpi were stained with different MAbs, as indicated, and analyzed by cytometry. Results are presented as means ± standard deviations and are representative of those from three independent experiments with 4 to 6 mice per group.

There was no significant difference in the percentages of CD4⁺ and CD8⁺ cells in PECs between the control and the infected groups (Table 1). The percentage of CD69⁺ cells in PECs was significantly higher in the RH group than in the naive group (19.9 versus 5.8%) at 7 dpi (Table 1). The percentage of CD69⁺ cells in splenocytes was also significantly higher in the RH group (11.8%) than in both the naive (3.7%) and the Mic1.3KO (1.9%) groups. The kinetics of the expression of CD69 on splenocytes were monitored in both groups up to 11 dpi (for the Mic1.3KO group alone), and this higher level was only seen at 7 dpi in the RH group (data not shown). The level had not increased at 11 dpi in the Mic1.3KO group, suggesting that CD69 expression may be correlated with the mortality observed in the RH group.

Ki-67 expression was used to follow the proliferation of PECs and splenocytes after infection with both strains. Ki-67 expression was significantly higher in the PECs of both infected groups than in those of the naive control group (16.4% in the control group versus 62.8 and 52.7% in the RH and Mic1.3KO groups, respectively). The difference between the two infected groups was not significant. Similar results were obtained for splenocytes, with higher percentages of Ki-67-expressing (Ki-67⁺) cells being detected in both infected groups (14.0% in the control group versus 24.3 and 37.6% in the RH and Mic1.3KO groups, respectively), but the difference between the control and RH groups was not significant.

The expression of CD25 was monitored in the CD4⁺ and CD8⁺ splenocytes and compared to that in the naive group (Table 2). Both the CD4⁺ and CD8⁺ populations expressed CD25 at 7 dpi. Moreover, the percentage of CD25⁺ cells was not significantly different between the naive and infected groups. However, the difference between the two infected groups was significant for the CD4⁺ population and was correlated with the higher level of expression of CD25 on Foxp3⁺ cells observed in the RH group.

Although the proliferation rates were similar in both groups, the RH group showed features of uncontrolled inflammation, with expression of high levels of CD69 molecules.

Induction of proinflammatory chemokines and cytokines after infection.

To determine whether protection may be due to equilibrium between effector and regulator cells, chemokines and cytokines associated with TH1/Tregs were quantified at the local level (peritoneum) and in the sera of both infected groups at 7 dpi.

Chemokine levels in the peritoneal washes and in the sera were quantified at 7 dpi. In the sera, only CCL2 was above the detection threshold and the level was higher in the RH group (Fig. 3A). The levels of CCL2, CCL3, and CCL20 in the peritoneal washes were significantly higher in the RH group than in the Mic1.3KO group (Fig. 3B to D).

FIG 3.

Higher chemokine levels in the RH group. Chemokines (CCL2, CCL3, CCL20) in sera (A) and peritoneal washes (B to D) were quantified at 7 dpi with RH or Mic1.3KO tachyzoites. The levels of all chemokines were below the detection threshold in control naive mice. The data presented are representative of those from five independent experiments with 5 to 6 mice per group. **, P < 0.01 using the Wilcoxon-Mann-Whitney exact test with strata; ***, P < 0.001 using the Wilcoxon-Mann-Whitney exact test with strata.

Systemic (sera) and local (peritoneum) IL-12 and IL-23 levels were also quantified at 7 dpi (Fig. 4). The levels of the IL12p40 subunit in the peritoneal washes (Fig. 4A) and in the serum (Fig. 4B) were significantly higher for the RH group. IL12p70 was undetectable in the peritoneal washes, and the levels of the biological form of this cytokine were not significantly different between the two groups (Fig. 4C). IL-23 was detected at low levels in the sera of RH-infected mice but not in the sera of Mic1.3KO-infected mice (Fig. 4D). The IL-23 level in the peritoneal washes was below the threshold of detection.

FIG 4.

Higher proinflammatory cytokine levels in the RH group. (A to D) IL12p40, IL12p70, and IL-23 were quantified by ELISA in peritoneal washes (IL12p40 only) (A) and in sera (B to D) at 7 dpi. (E to H) IFN-γ and IL-6 levels in peritoneal washes (E, G) and sera (F, H) at 7 dpi were quantified by ELISA. The levels of all cytokines were below the detection threshold in control naive mice. The data presented are representative of those from four independent experiments with 4 to 6 mice per group. **, P < 0.01 using the Wilcoxon-Mann-Whitney exact test with strata; ***, P < 0.001 using the Wilcoxon-Mann-Whitney exact test with strata.

As IL12p70 directs the differentiation of naive T cells in TH1 cells and IL-23 is linked to TH17 differentiation and pathogeny, cytokines related to TH1 and TH17 cells were then quantified.

Serum IFN-γ levels were significantly higher in the RH group than in the Mic1.3KO group (Fig. 4E). Slightly higher levels of IFN-γ in the peritoneal washes were observed for the RH group, but the difference from the Mic1.3KO group was not statistically significant (Fig. 4F). IL-6 levels were also significantly higher in the sera and peritoneal washes of the RH group than in those of the Mic1.3KO group (Fig. 4G and H). IL-17A levels were at the detection limit for the RH group and below the detection limit for the Mic1.3KO group (data not shown).

Taken together, these results suggest that the responses of the RH group were more inflammatory, and the RH group revealed a TH1/TH17-like response, whereas the Mic1.3KO group showed a more regulated TH1/Treg-like response.

To evaluate the TH1/Treg response further, antigen (Ag)-specific IFN-γ production in spleen cell supernatant was measured following 48 h of restimulation with soluble Toxoplasma antigens (TE) (Fig. 5A). The IFN-γ levels were below the detection threshold on day 4 in both groups (data not shown). At day 7, the levels were very high in the RH group and significantly different between the two groups. Cells from the RH group produced higher levels of IFN-γ than those from the Mic1.3KO group with or without TE stimulation.

FIG 5.

Strong systemic, nonspecific IFN-γ production and low levels of IL-10 production induced by infection with RH. Splenocytes were recovered at 7 dpi with RH or Mic1.3KO tachyzoites and stimulated with medium or with parasite extract (TE). Culture supernatants were collected after a 48-h stimulation period. IFN-γ (A) and IL-10 (B) levels were determined by ELISA. The data presented are representative of those from four independent experiments with 4 to 6 mice per group.*, P < 0.05 using the Kruskall-Wallis test followed by Dunn's posttest; **, P < 0.01 using the Kruskall-Wallis test followed by Dunn's posttest.

IL-10 plays an important role in the suppressive function of antigen-specific Tregs, and IL-10 production by Tregs has been described to be a prominent Treg suppressor mechanism (22). IL-10 levels in the splenocyte supernatants were significantly different in both groups (Fig. 5B). The splenocytes of animals from the RH group produced less IL-10 than those from the Mic1.3KO group. However, for both groups IL-10 production was dependent on Ag restimulation. To investigate the cellular source of IL-10, spleen cells were cultured with TE in order to track the IL-10-secreting cells by flow cytometry. There was no significant difference in the percentage of IL-10-secreting cells between the two groups (median values, 10.5 for the RH group and 7 for the Mic1.3KO group). We were unable to detect IL-10 in the Foxp3⁺ population. Moreover, the main sources of IL-10 in both groups were the CD4⁻ cells, since the percentages of CD4⁺ cells among IL-10-producing cells in two different experiments were 28% and 30% for the RH and the Mic1.3KO groups, respectively.

IL-2 is a cytokine that is essential for Treg development and survival (23) and is also involved in the immunosuppression induced during the acute phase of toxoplasmosis (24, 25). IL-2 levels were quantified in spleen culture supernatants after activation with anti-CD3 (data not shown). IL-2 levels were significantly lower in the RH group than in the naive mice at 7 dpi. The high levels of IFN-γ in the RH group supernatants indicated that the cellular cytokine secretion capacity was not impaired.

Taken together, these results suggest that, despite a strong cellular response and high levels of production of IFN-γ, the parasite burden was not controlled in the RH group; in contrast, in the Mic1.3KO group the cellular response and the parasite burden were controlled.

Induction of Tregs in vivo in the RH group induced better control of the parasites and reduced the level of inflammatory cytokines.

Administration of a complex consisting of a low dose of IL-2 and anti-IL-2 antibody JES6-1 specifically amplifies Tregs in vivo (26). To evaluate further the role of CD4⁺ Foxp3⁺ cells in protection, RH-infected mice were treated with IL-2/anti-IL-2 complexes, and the levels of cytokine secretion and the parasite burden were evaluated at 7 dpi. Treatment with IL-2/anti-IL-2 complexes significantly increased the frequency of Foxp3⁺ cells in the spleen. The percentages of CD4⁺ Foxp3⁺ splenocytes in control and IL-2/anti-IL-2-treated mice were 1.6% ± 0.5% and 3.7% ± 1.6%, respectively, in the first experiment and 1.3% ± 0.4% and 3.5% ± 0.3%, respectively, in the second experiment. Mice treated with IL-2/anti-IL-2 complexes showed a diminished number of parasites in the peritoneal cavity compared to control mice injected with PBS, with 0.5 × 10⁶ parasites being found in the IL-2-treated group and 5.5 × 10⁶ parasites being found in the group injected with PBS (corresponding to median values from two independent experiments with 5 to 6 mice per group per experiment), and these diminished parasite burdens were correlated with increased numbers of cells (data not shown).

CCL2 and CCL3 levels were higher in the control group than in the IL-2-treated group (Fig. 6 A and B) at the local level (peritoneum). IFN-γ and IL-6 levels in the serum and peritoneum were also significantly higher in the control group than in the IL-2-treated group (Fig. 6C to F). IL-17A was detectable at low levels at the local level only in the control group (data not shown).

FIG 6.

Higher chemokine and proinflammatory cytokine levels in the RH-infected control group than the RH-infected and IL-2-treated group. (A and B) The chemokine CCL2 (A) and CCL3 (B) levels in peritoneal washes were quantified at 7 dpi. (C to F) IFN-γ (C, D) and IL-6 (E, F) levels in peritoneal washes (C, E) and sera (D, F) were quantified at 7 dpi by ELISA. The control and IL-2 groups were injected by the i.p. route with PBS and IL-2/anti-IL-2 antibody, respectively. Results are expressed as the median plus range. The data presented are representative of those from two independent experiments with 5 to 6 mice per group. *, P < 0.05; **, P < 0.01; ***, P < 0.001 using Wilcoxon-Mann Whitney exact test with strata on the cumulative data of the two experiments.

Specific IFN-γ and IL-10 secretion by splenocytes after restimulation with the antigen was also significantly higher in the IL-2-treated group than in the control group (Fig. 7A and B).

FIG 7.

Systemic, antigen-specific IFN-γ and IL-10 production after treatment with IL-2/anti-IL-2 in RH-infected mice. Splenocytes were recovered at 7 dpi with RH tachyzoites and stimulated for 72 h with TE. Culture supernatants were collected after a 48-h stimulation period. The data presented are representative of those from two independent experiments with 5 to 6 mice per group. The control and IL-2 groups were injected by the i.p. route with PBS and IL-2/anti-IL-2 antibody, respectively. Values were corrected by subtraction of the level in control wells without antigen. IFN-γ (A) and IL-10 (B) levels were measured by ELISA.*, P < 0.05 using the Wilcoxon-Mann-Whitney exact test with strata on the cumulative data from the two experiments.

These findings suggest that the cytokine response after IL-2 treatment may be switched from an inflammatory response to a more controlled response.

In addition, mice vaccinated with Mic1.3KO and treated with the anti-IL-2 receptor antibody PC61, which is known to deplete Tregs preferentially, revealed features similar to those of RH-infected mice, including a higher parasite count, higher levels of morbidity, and higher inflammatory cytokine levels (data not shown).

Taken together, these results indicate that mice infected with RH and treated with IL-2/anti-IL-2 complexes showed features similar to those of mice vaccinated with Mic1.3KO (a reduced parasite count, a more regulated TH1 response), suggesting a role for Tregs in the protection developed with this vaccine strain.

DISCUSSION

The aim of this study was to investigate the role of CD4⁺ Foxp3⁺ T regulatory cells (Tregs) in the protection induced by an attenuated replicating type I Mic1.3KO strain. Type I strains are uniformly lethal except after attenuation, which leads to lower levels of replication and dissemination in the host and the development of long-term immunity, as for type II strains. Several studies have been performed to examine the role of Tregs using a susceptible C57BL/6 mouse strain (14, 17, 18, 20) and a type II Toxoplasma strain. Decreased Foxp3⁺ Treg counts at local and systemic sites were associated with such mortality in C57BL/6-susceptible mice after lethal oral infection with the type II strain (14). This decrease occurred very shortly before the animals died and was correlated with virulence. When the dose and the route of infection were changed and the mice recovered (14, 18), the decrease was only transient. In the present study, we investigated the Foxp3⁺ Tregs after infection with the attenuated Mic1.3KO parasite and compared them with the Tregs from the parental RH type I strain, from which Mic1.3KO is derived. In contrast to the results obtained after infection with the type II strain (14), the mortality induced by infection with the parental RH strain was not associated with a decrease in CD4⁺ Foxp3⁺ Treg numbers or frequencies. Neither the number nor the frequency differed from that in the naive control group. A lack of Treg reduction seemed to contradict the uncontrolled immune response observed in the RH group and previous studies (14, 18), showing that a decrease in Tregs was associated with mortality. However, we cannot exclude the possibility that this decrease appeared just before death or that it was significant only at the local site. Oldenhove et al. (14) showed a highly significant reduction in Foxp3⁺ Treg frequency in both the intestine (local) and the spleen at 10 dpi. However, the reduction in Treg splenocytes was significant only at 12 dpi, which was the day that the mice died.

In our experiment, a weak increase in Tregs was observed at the local site after infection with the vaccinal strain. This increase was small and transient and was followed by increases in CD4⁺ CD25⁺ T effector cells, which were correlated with parasite control. This increase in Treg number was limited to local sites, since both the numbers and percentages of CD4⁺ Foxp3⁺ T cells remained unchanged in spleens until 7 dpi in both infected groups compared to those in naive control mice. However, the percentage of CD4⁺ Foxp3⁺ splenocytes expressing CD25 was slightly higher in the RH group than in the Mic1.3KO group on day 7. Tenorio et al. (24) reported that, after infection with a type II strain of T. gondii, Tregs were activated and showed increased expression of CD25; in addition, there were increased percentages of Tregs expressing CD25. Tregs expanded in vivo following IL-2 injection also expressed higher levels of CD25 (27).

The lower increase in the numbers of Tregs in the RH group at the local site of infection may explain the difference between the Mic1.3KO and RH groups. In resistant BALB/c mice, Tregs are thought to reduce immunopathology since depletion of these cells induces morbidity due to the increased production of proinflammatory cytokines and higher parasite burden after oral infection (19). In this study, mice of both the resistant BALB/c and susceptible C57BL/6 strains had higher CD4⁺ Foxp3⁺ counts at the local level at 6 days postinfection than naive mice, and the absolute count in the resistant strain was higher than that in the susceptible strain. This suggests that differences in Treg counts may account for the differences in mortality. This may also have been true in our experiment, since the Treg count was lower for the RH group than for the Mic1.3KO group.

Treg expansion/recruitment and accumulation at local sites of infection have been extensively described in various parasite infections, especially in chronic infections (28).

After infection, the microenvironment may also influence the ability of Tregs to determine the outcome of an immune response toward tolerance or immunity. TH17 and Treg differentiation is closely related, and in the presence of proinflammatory cytokines, transforming growth factor β-induced Foxp3 expression is reduced, favoring the differentiation of TH17 (29).

Although IL-17A was undetectable in our study, several cytokines or chemokines induced in TH17 differentiation or induced by IL-17 were preferentially expressed in the RH group. CCL2, CCL3, and CCL20 levels were higher in RH-infected mice than in Mic1.3KO-infected mice both at the local level and at the systemic level. CCL20 is produced by TH17 cells (30), and IL-17 stimulates the production of CCL2 and IL-6 (29). CCL2 is produced early after infection and plays an essential role in resistance to acute toxoplasmosis by recruitment of Gr1-positive monocytes to the site of injection (31). Dal Secco et al. (32) also showed that Tregs inhibit CCL2 and CCL3 production by dendritic cells in vivo and limit the recruitment of inflammatory cells.

Higher levels of IL-6 and IL-23 were found in the RH group than in the Mic1.3KO group. Both cytokines are involved in TH17 differentiation. Moreover, TH17 cells that differentiate in the presence of IL-23 are pathogenic (33). The pathological role of IL-17 has been shown in ocular toxoplasmosis. Sauer et al. (34) demonstrated that the balance between TH17 and TH1 responses was crucial for the outcome of infection. They also showed a deleterious TH17 response at the local level after primary infection which was reduced after reinfection. TH2/Treg responses were enhanced under the reinfection condition (35).

Contrary to the TH1/TH17-type signature after RH infection, the cytokine signature after Mic1.3KO infection in our study was of the TH1/Treg type.

We observed a higher level of IL-10 in the spleen cell supernatant of the Mic1.3KO group than in that of the RH group. IL-10 is an immunosuppressive cytokine, and the common feature of all effector Tregs is the expression of IL-10 (22). Most of the IL-10-secreting splenocytes of both infected groups (75%) were CD4⁻, and within the CD4 population, the secreting cells were Foxp3⁻. This result is consistent with previous results reported by Jankovic et al. (36), who showed that the main producer of IL-10 during T. gondii infection within a CD4 population was the Foxp3⁻ IFN-γ-positive cell population.

IL-2 is essential for Treg development and survival (23) and is also known to inhibit TH17 and to favor Treg differentiation. After anti-CD3 activation in vitro, a decrease in the IL-2 level was observed in the RH group compared to that in the control naive group and the Mic1.3KO group. The low level of IL-2 in the RH group may have resulted from the greater consumption of IL-2 rather than from lower levels of IL-2 release, since cells were not impaired in IFN-γ production. This low IL-2 level may therefore be correlated with the higher level of expression of CD25 by Foxp3⁺ cells. This reduced IL-2 level in cells from animals infected with T. gondii has been described by others (14, 19, 24). A reduced number of Tregs during the acute phase of T. gondii infection is a consequence of a reduced IL-2 availability (14, 19). IL-2 is also involved in the transient immunosuppression observed during acute toxoplasmosis (24, 25).

IFN-γ levels were higher in the serum, peritoneal cavity, and spleen cell supernatant of the RH group. However, the production of IFN-γ by splenocytes of the RH group was not antigen dependent and was not increased by antigen restimulation. These findings suggest that splenocytes from the RH group were in an activated state, and the expression of CD69 by these cells further supports this hypothesis. Expression was not observed until day 7, 1 to 3 days before the mice died. CD69 expression was not observed in the Mic1.3KO group even at day 11. Expression of CD69⁺ T cells is known to be localized at sites of chronic inflammation and at sites of an active immune response in vivo (37). The role of CD69 has not yet been elucidated, and previous findings have suggested that CD69 is an activating molecule, although recent studies have also reported that it has regulatory functions. CD69 regulates immune and inflammatory responses by acting as a brake on the differentiation of TH17 effector cells (38).

In summary, the protective response after infection with the attenuated type I-derived Mic1.3KO strain may originate from a TH1 response regulated by an appropriate Treg response. The mortality induced by infection with the parental RH strain may have been due to a TH1/TH17-type response.

When the balance was in favor of Tregs by using IL-2/anti-IL-2 complexes, we demonstrated that treated mice infected with RH showed features (a lower parasite burden and TH1/Treg cytokines and chemokines) similar to those of mice vaccinated with Mic1.3KO. These complexes work by inducing and expanding peripheral Tregs (27). Treatment of C57BL/6 mice with IL-2/anti-IL-2 during T. gondii infection with a lethal dose of type II cysts was shown to prevent the loss of Tregs and reduce the level of morbidity, but it induced a higher parasite burden in the brain (14). The lower parasite burden that we observed at the local site in the group in which Tregs were induced is surprising and seems contradictory to the earlier findings (14). However, a higher parasite burden in the lamina propria tissue of Treg-depleted BALB/c mice after oral infection has also been reported (19). The authors argued that this high parasite burden may be an indirect effect of the high levels of proinflammatory cytokines at the local site causing tissue destruction. This may also have been true in our experiments, since RH infection induced high levels of IFN-γ at the infection site. Alternatively, very high levels of IFN-γ and the constitutive production of IFN-γ may lead to the exhaustion of the response as a result of negative feedback or by saturation of the IFN receptor, leading to impairment of the IFN-γ-induced response. This could also be related to a failure to induce Ag-specific IFN-γ production, which may be ineffective in parasite control.

In conclusion, this study provides evidence that Tregs may have contributed to protecting mice after infection with an attenuated, replicating type I strain of T. gondii. It is important to determine whether Tregs have a beneficial or detrimental role in protection against toxoplasmosis in order to design an efficient vaccine. Although a complex mechanism may underlie the protection induced by an attenuated type I strain, we showed that a change in the absolute number of Tregs may be crucial in determining whether animals are protected or whether they will succumb.