glucocorticoid-induced TNFR family-related protein
moth cytochrome c
parasitized red blood cells
non-lethal strain of Plasmodium yoelii
lethal strain of Plasmodium yoelii
regulatory T cells
CD4+ T cells are the major effector T cells against blood-stage Plasmodium yoelii infection. On the other hand, the lethal strain of P. yoelii (PyL) has acquired an escape mechanism from host T cell immunity by activating CD4+CD25+ regulatory T cells (Treg). Although the activation of Treg during PyL infection precludes the clearance of PyL from mice, it remains unclear whether activation of Treg is attributable to a specific response against PyL infection. Thus, we examined here whether Treg proliferate in an antigen-dependent manner during PyL infection. We also investigated the effector and regulatory mechanisms of Treg. Infection with PyL increased the number of CD4+CD25+ T cells, in which expression of Foxp3 mRNA is up-regulated. The Treg that were transferred into mice infected with PyL, but not with a non-lethal strain of P. yoelii (PyNL), proliferated during the initial 5 days following infection. The Treg from PyL-infected mice showed strong suppression compared with those from naive or PyNL-infected mice, and could suppress T cell activation by recognizing PyL- but not PyNL-derived antigens. Furthermore, the suppressive function of Treg activated in PyL-infected but not in naive mice could not be inhibited by treatment with an anti-glucocorticoid-induced TNFR family-related protein (GITR) mAb. These findings indicate that PyL infection specifically activates Treg that are specific for PyL-derived antigens. The infection also induces resistance for Treg to GITR signaling, and this eventually contributes to the escape of parasites from host T cell immunity.
The immune system exploits clonal diversity in order to combat pathogens that produce a variety of antigens. In contrast, pathogens such as parasites and viruses have evolved mechanisms to escape from host immune systems in order to survive within the host. Many such evasion strategies have been demonstrated 1, 2. For instance, certain viruses encode viral cytokines or chemokines that inhibit the host immune system, which is then unable to clear the viruses 3. Also, adenoviruses suppress the MHC class I antigen presentation pathway, and this eventually impairs maturation of lytic CD8+ T cells to kill viral infected cells 4. Optimization of therapeutic strategies and vaccine against pathogens will benefit from detailed knowledge of these escape mechanisms.
Malaria parasites cause approximately 2 million deaths per year 5. Antibodies and T cells play crucial roles in protective immunity against malaria parasites 6. Antibodies to merozoites and parasitized red blood cells (pRBC) can reduce disease severity by inhibiting new RBC infections by merozoites, or by inhibiting the cytoadherence of pRBC to the capillary endothelium 7. CD8+ T cells exhibit cytotoxicity to hepatocytes, where sporozoites mature 8. CD4+ T cells are indispensable for protection against blood-stage parasites 9. However, it is difficult to acquire long-lasting protective immunity, despite frequent exposure to the parasite in endemic areas. There are several reasons to explain this failure. For example, antigenic diversity results in altered T cell epitopes that prevent a protective T cell response 10. Clonal antigenic variation of molecules expressed on merozoites or pRBC hide the parasite from antibody recognition 11–14. One recent report demonstrated that pRBC impair the maturation of dendritic cells that initiate T cell immune responses 15. Malaria patients also show some level of global immunosuppression 16–18.
We have recently revealed that T cells expressing both CD4 and CD25 (IL-2R α chain) (regulatory T cells, Treg), which are known to suppress autoreactive T cells 19–22, inhibit the activation of effector T cells against antigens from the blood stage of the lethal strain of Plasmodium yoelii (PyL) 23. This inhibition of effector T cell activation finally induces host death, while the depletion of Treg enables mice to survive 23. These studies suggest that activation of Treg is one of the escape mechanisms for malaria parasites from host T cell immunity 23. Although those studies have revealed the clear contribution of Treg to the inhibition of effector T cell activation, it remains unclear whether activation of Treg is attributable to the specific recognition of PyL-derived antigens.
We here demonstrate that infection with PyL induces proliferation of Treg during a very early stage of infection, which is, at least partly, dependent on the recognition of PyL-derived antigens. Furthermore, the suppressive function of Treg during PyL infection could not be inhibited by anti-glucocorticoid-induced TNFR family-related protein (GITR) mAb, in contrast to complete relief of suppressive activity of Treg from naive mice. These studies have revealed that PyL infection induces rapid proliferation of Treg and enables Treg to acquire GITR-resistant mechanisms, which are responsible for the strong function of Treg during PyL infection.
Infection with PyL increases CD4+CD25+ cells and their Foxp3 expression
We have previously shown that C57BL/6 mice can clear the infections with a non-lethal strain of P. yoelii (PyNL) by 20 days after infection, while mice die around 10 days after infection with PyL 23. Furthermore, the injection of anti-CD25 mAb 3 and 1 day before, and 5 days after PyL infection, enabled mice to become resistant to PyL infection, and these mice finally cleared the parasites 23.
In order to address the question as to whether CD25+ Treg proliferated in response to PyL-derived antigens, we first examined the kinetics of relative CD4+CD25+ T cell number during the course of infection. We found that the ratio of CD25+ cells in total spleen cells or in the splenic CD4+ T cell fraction was slightly higher in both PyL-infected and PyNL-infected mice compared with naive mice as early as 5 days after infection (Fig. 1a, b). Although the increase was transient in mice infected with PyNL, that in mice infected with PyL sustained until 7 days after infection (Fig. 1a, b).
Since CD25 is not a specific marker for Treg, and activated T cells also express CD25 19, we evaluated the expression of Foxp3. The Foxp3, a forkhead transcription factor, has been identified as a gene selectively expressed in CD4+CD25+ Treg. The expression of Foxp3 is associated with the development and function of CD4+CD25+ Treg 24–26. As previously demonstrated, among CD4+ T cells purified from naive mice, CD25+ cells specifically expressed Foxp3 and CD25– cells showed only a trace expression (Fig. 2a). We next analyzed CD4+CD25+ cells purified from mice after infection with the malaria parasites. CD4+CD25+ cells from mice infected with PyNL expressed slightly more Foxp3 transcripts 5 days after infection when transient increase of CD4+CD25+ cells was observed (Fig. 2b). In contrast, those from mice infected with PyL showed remarkable increase of Foxp3 5 days after infection and returned to normal level 7 days after infection (Fig. 2b).
Treg proliferated in PyL-infected mice
In order to examine whether Treg proliferate in response to PyL but not PyNL infection, CD4+CD25+ or CD4+CD25– T cells labeled with CFSE were transferred into C57BL/6 mice, and they were subsequently infected with each parasite strain. About 17% of CD4+CD25+ T cells isolated from mice 5 days after infection with PyL had divided, while PyNL infection induced few cell divisions of CD4+CD25+ T cells (Fig. 3a, b). The CD4+CD25– T cells isolated from mice 5 days after PyL or PyNL infection did not divide significantly (Fig. 3a, b). These findings indicate that PyL infection allows CD4+CD25+ T cells to proliferate at an early phase of infection before proliferation of CD4+CD25– T cells.
Treg induced by PyL infection show strong suppressive function
We next evaluated the functional properties of CD25+ T cells obtained from PyL- or PyNL-infected mice. The CD4+CD25+ T cells from uninfected, PyL- or PyNL-infected mice were cultured with naive CD4+CD25– T cells in the presence of anti-CD3 mAb. CD4+CD25+ T cells from naive and PyNL-infected mice similarly suppressed the anti-CD3 mAb-induced proliferation of naive CD4+CD25– T cells in a dose-dependent manner (Fig. 4). The CD4+CD25+ T cells from PyL-infected mice had much stronger suppressive ability than those from PyNL-infected or naive mice (Fig. 4).
Treg induced by PyL infection show suppression in an antigen-specific manner
As Treg requires T cell receptor (TCR) engagement for exertion of their suppressive function 27, we next examined whether this suppression occurs in an antigen-specific manner in vitro. Splenic total CD4+ T cells purified from mice 7 days after infection with PyL proliferated less than those isolated from PyNL-infected mice when stimulated with PyL- or PyNL-pRBC (Fig. 5a, left and right panels), as we previously described 23. These results indicate that CD4+ T cells from PyL-infected mice contain cell population able to respond to PyNL-derived antigen. When CD25+ cells were removed from CD4+ T cells in spleen cells of PyL-infected mice, the resultant cells could proliferate in response to PyL-pRBC comparably to PyNL-pRBC-derived antigens (Fig. 5b). Furthermore, the addition of CD25+ cells back to CD4+CD25– cells from PyL-infected mice remarkably inhibited the response against PyL-pRBC compared with that against PyNL-pRBC-derived antigens (Fig. 5b).
In order to further confirm the antigen specificity of the Treg, we performed the experiments using 2B4 TCR-transgenic mice that have only T cells bearing TCR specific for moth cytochrome c (MCC) peptide presented by I-Ek MHC class II molecules 28. The co-culture of Treg from PyL-infected mice and 2B4 T cells reduced the proliferative responses of 2B4 T cells against MCC peptide in the presence of PyL-pRBC (Fig. 5c). Taken together, these results indicate that Treg from PyL-infected mice are able to exert suppressive activity by recognizing antigens preferentially expressed in PyL-pRBC.
The anti-GITR mAb could block the suppressive function of Treg from naive mice but not from PyL-infected mice
Several reports have shown that the suppressive function of CD25+ Treg is mediated by cytokines as well as by direct cell contact 19. We examined whether CD25+ Treg from PyL-infected mice have similar suppressive mediators to those from naive mice. The CD4+CD25– T cells from naive mice and CD25+ T cells from PyL-infected or naive mice were stimulated with anti-CD3 mAb in the presence or absence of anti-IL-10 and anti-TGF-β mAb, and total T cell proliferation was measured. The anti-IL-10 and TGF-β mAb could not inhibit the suppressive ability of Treg from both naive and PyL-infected mice against anti-CD3 mAb-treated T cell proliferation (Fig. 6a). The addition of anti-IL-10 and anti-TGF-β mAb could not also affect Treg activity against PyL-derived antigen-specific T cell proliferation (Fig. 6b).
Recent studies have shown that GITR signaling in Treg inhibits the suppressive function of Treg 29, 30. Thus, we next examined whether GITR signaling inhibits the suppressive function of Treg from PyL-infected mice. Treating Treg from naive mice with anti-GITR mAb completely inhibited the suppressive function of Treg, while the same treatment could not inhibit the suppressive function of Treg from mice infected with PyL (Fig. 6a). Anti-GITR mAb also inhibited suppression of CD25+ cells from PyNL-infected mice (data not shown). Similarly, the anti-GITR mAb could not inhibit the suppressive function of Treg that were activated by PyL-pRBC (Fig. 6b). To confirm the resistance of Treg from PyL-infected mice to GITR stimulation, we also examined the dose response to anti-GITR mAb. The anti-GITR mAb dramatically increased the T cell proliferation when Treg from naive mice were used, while only slight T cell proliferation was observed when Treg from PyL-infected mice were used with all mAb concentrations (Fig. 6c). We finally checked the expression level or pattern of GITR on CD4+ T cells from naive or PyL-infected mice. Splenic CD4+CD25– cells expressed significant level of GITR and expressed CD25+ fraction at much higher level. There was no difference in expression profiles between naive and PyL-infected mice (Fig. 6d).
GITR is expressed not only on Treg but also on non-Treg T cells and several reports demonstrated that anti-GITR antibodies act on non-Treg cells by providing costimulation 31, 32. Thus, the outcome of treatment with anti-GITR in mixed culture with Treg and effector cells depends on the potency of Treg and on the strength of the effector T cell proliferation responses. Although Treg from PyL-infected mice are more efficient in suppression than naive Treg (Fig. 6a), it is possible that resistance to reversal of suppression induced by anti-GITR is due to the quantity of Treg potential. To test this possibility, we titrated effector cell-to-Treg ratios in the presence or absence of anti-GITR antibodies.
The addition of anti-GITR antibodies to cultures mixed with CD4+CD25– T cells and CD25+ cells from naive mice increased the baseline proliferation and Treg could not inhibit the proliferation as much as 1:6 CD4+CD25– T cell-to-CD25+ T cell ratio. However, the suppression began to be observed in cultures containing half number of naive Treg-to-effector cells (Fig. 6e, left panel). In contrast, when CD25+ cells from PyL-infected mice were used, the addition of anti-GITR antibodies did not reverse the suppression at the all ratios of CD25+ cells to effector cells (Fig. 6e, right panel). It should be noted that 1:54 ratio of CD25+ cells from PyL-infected mice suppressed the proliferation equivalently to that used in 1:6 ratio of CD25+ cells from naive mice in the absence of anti-GITR antibodies, and that remarkable suppression was observed in the former but not in the latter cultures in the presence of anti-GITR antibodies (Fig. 6e). Taken together, these findings suggest that Treg from PyL-infected mice have different regulatory mechanisms from those by Treg from naive mice.
Treg play a vital role in the induction and maintenance of peripheral self tolerance 19, 21, 22. These professional regulatory cells prevent the activation and proliferation of potentially autoreactive T cells that have escaped thymic deletion or that recognize extrathymic antigens 19. Several recent studies have demonstrated that Treg regulate effector T cell responses against infectious organisms 33–37. We also have revealed that activation of Treg is one of the escape mechanisms of malaria parasites from host immunity 23. Here, we provide evidence that PyL infection resulted in Treg that proliferated in a PyL-specific manner at a very early stage of infection, although Treg are known to be anergic in vitro38. Furthermore, the activation of Treg during PyL infection enabled the Treg to become resistant to GITR signaling that generally makes Treg losing their suppressive function 29, 30.
We here found that Treg proliferated significantly after infection with PyL, but not PyNL, suggesting that PyL infection specifically activated Treg in vivo. There are at least three possible explanations for these findings. One explanation for the increased proliferation of Treg in PyL-infected mice in vivo is that PyL secretes soluble factors that cause proliferation of Treg. Our preliminary experiments, however, indicate that this explanation is unlikely, because the addition of parasite lysate or pRBC did not affect the proliferative response of Treg in vitro (data not shown). This also could not explain the difference between PyL and PyNL, because PyL, a variant derived from PyNL 39, is supposed to be identical in its genetic and antigenic background to PyNL.
The second possibility is that the qualitative or quantitative properties of antigen displayed by MHC class II are different between PyL- and PyNL-infected mice. Such differences might contribute to the preferred activation of Treg in PyL-infected mice. Treg are thought to have TCR with a relatively high affinity against self antigens compared with conventional CD4+ T cells 40, 41. Thus, the TCR repertoire might differ between the two populations. PyL might enable the escape mechanism by expressing peptides on MHC class II that have high affinity against TCR on Treg rather than on conventional T cells. The final possibility is that the parasite or pRBC directly interact with Treg or non-T cells, thus augmenting Treg activation. Indeed, recent findings have shown that Toll-like receptors expressed on Treg regulate Treg function even in the absence of TCR signaling 42. In any case, the mechanisms that activate Treg in PyL infection would be the next logical step to study.
We here observed that CD4+CD25+ T cells from PyL-infected mice have stronger suppressive activity than those from naive or PyNL-infected mice. Furthermore, the expression of Foxp3 was higher in CD4+CD25+ T cells from PyL- than PyNL-infected mice. However, it remains unclear whether the relative percentage of Treg is increased in CD4+CD25+ T cells, or suppressive effector ability in individual Treg is augmented in PyL-infected mice. These issues may be solved by comparing Foxp3 expression in CD4+CD25+ T cells from PyL- and PyNL-infected mice by flow cytometry.
The effectors involved in the suppressive function of Treg and inducers of the suppressive mechanism of Treg remain controversial. As for effectors of Treg, Nakamura et al.43 reported that TGF-β plays a role in the suppressive function of Treg, but other groups do not support this concept 44. Also, the blockade of IL-10 in vitro does not affect Treg function 43 while IL-10 has been shown to block Treg function in some in vivo studies 45. We also examined whether TGF-β or IL-10 is responsible for the suppressive function of Treg from naive or PyL-infected mice. Using blocking antibodies, we found that neither cytokine could suppress Treg function.
As a regulator of Treg function, the signaling through GITR, which is highly expressed on Treg, is known to inhibit the suppressive function of Treg 29, 30. The stimulation of GITR by an antibody could inhibit the suppressive function of Treg from naive mice, as reported 29, 30. In contrast, the same stimulation did not affect the function of Treg from PyL-infected mice, indicating that Treg acquire GITR-resistant mechanisms during PyL infection. Since we have not observed such a loss of GITR effect in Treg that were activated during tumor invasion (H.H. et al., unpublished observation), the GITR resistance in Treg is somehow specific for PyL infection. In addition the GITR signaling enhances the proliferative responses of effector T cells 31, 46. Given these findings it will be interesting to examine whether the observed resistance against GITR signaling reflects defective GITR signaling in Treg, or the acquisition of strong suppressive function of Treg overriding the GITR-mediated costimulation to effector T cells. Either way, GITR resistance in Treg function is likely to play a role in the escape of parasites from T cell immunity. It would be interesting to examine whether TCR signaling through the recognition of PyL-derived antigens, or a TCR independent effect on Treg, contribute to GITR resistance.
Collectively, our data provide a new perspective to evaluate mechanisms by which Treg are specifically activated during PyL infection by acquiring GITR resistance. Further studies to address the activator of Treg and inhibitors of GITR signaling will be important for understanding molecular mechanisms of parasite escape mechanisms from host immunity, as well as regulatory pathways of naturally occurring Treg.
Material and Methods
Mice and parasites
Female C57BL/6 mice (6–8 wk old) purchased from SLC (Hamamatsu, Japan) and RAG2-deficient mice from CLEA (Kawasaki, Japan) were utilized. 2B4 TCR-transgenic mice 28 were a generous gift from Dr. Fukui (Kyushu University, Japan) under permission of Dr. Davis (Stanford University, CA). Blood-stage parasites of PyL or PyNL (generous gifts from Dr. Torii, Ehime University, Japan) were obtained after fresh passage through a donor mouse 3–7 days after inoculation with frozen stock. Mice were infected with 10 000 to 15 000 pRBC intraperitoneally. To purify pRBC for use as a stimulant, blood collected by heart puncture from infected mice was passed through a CF11 column to eliminate host white cells, followed by discontinuous gradient centrifugation on 55% Percoll (Pharmacia, Uppsala, Sweden). The interphase fraction was collected as pRBC.
Purified anti-CD25 (PC61.5) was purchased from eBioscience (San Diego, CA). Purified anti-CD16/32, anti-CD11b, anti-MHC class II, anti-DX5, PE-anti-CD25 (PC61.5), allophycocyanin-anti-CD4 (RM4-4) and anti-IL-10 (JES5–16E3) mAb were purchased from BD PharMingen (San Diego, CA). Anti-rat IgG microbeads were purchased from Miltenyi (Auburn, CA). Anti-rat IgG Dynabeads were purchased from Dynal AS (Oslo, Norway). The goat anti-GITR polyclonal antibody was purchased from R&D Systems (Minneapolis, MN). The anti-CD3 (2C11), anti-CD4 (GK1.5), anti-CD8 (53.6.7), anti-CD25 (7D4), anti-B220 (RA3–6B2), anti-GITR (YGITR 765.4.16) 31 and anti-TGF-β (1D11) mAb were purified from ascites of mice injected with hybridomas.
Flow cytometric analysis
Cells in single suspension were stained with allophycocyanin-anti-CD4, PE-anti-CD25 in combination with various antibodies. Cells were analyzed by FACS Calibur (Becton Dickinson, Mountain View, CA) and data were analyzed using CellQuest software (Becton Dickinson).
Cell purification and culture
Spleens of mice were prepared as single suspensions. To remove non-T cells, the suspensions were incubated with anti-B220, anti-CD11b, anti-CD16/32, anti-MHC class II and anti-DX5 mAb, followed by incubation with anti-rat IgG Dynabeads. To purify CD4+CD25+ T cells, enriched T cells were incubated with anti-CD25 (PC61.5) followed by incubation with anti-rat IgG microbeads. Positive selection was performed according to the manufacturer's protocol. CD4+CD25– T cells were purified by depletion of CD8+ and CD25+ cells from the enriched T cell fraction. The purity of each cell subset usually exceeded 85%. Purified cells were stimulated with soluble anti-CD3 antibody at a concentration of 2.5 μg/mL or with pRBC (up to 1×106) in the presence of APC in 0.2 mL of media for 72 h and incubated with 1 μCi/well of [3H]thymidine for the final 6 h. Radioactivity was measured by a liquid scintillation counter. In some experiments, anti-IL-10 (10 ng/mL), anti-TGF-β (20 ng/mL) or anti-GITR (20 ng/mL) mAb was added to the culture.
Total RNA was extracted from purified CD4+CD25+ T cells and then was reverse-transcribed to cDNA. cDNA was analyzed for the expression of foxp3 mRNA by SYBR Premix Ex Taq (Takara, Tokyo, Japan) using a Perkin-Elmer ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). PCR was performed according to manufacturer's instruction. Quantity of foxp3 mRNA was expressed as ratio to expression of 18S rRNA. The sequences of PCR primers for foxp3 were 5′–CCCAGGAAAGACAGCAACCTT-3′ and 5′–TTCTCACAACCAGGCCACTTG-3′, and those for 18S rRNA were 5′–GTAACCCGTTGAACCCCATT-3′ and 5′–CCATCCAATCGGTAGTAGCG-3′.
Cell division analysis
Purified cells were incubated with 5 μM of CFSE (Molecular Probes, Eugene, OR) at a concentration of 1×107/mL for 10 min at 37°C. After three washes, labeled cells were transfused intravenously into recipient mice 1 day prior to infection. Transferred cells were recovered from mice from 5 to 8 days after infection with P. yoelii and were stained with anti-CD4 and anti-CD25 mAb.
We thank Dr. Torii for providing malaria parasites. This work was supported by Grants-in-Aid for Young Scientists (A) from the Ministry of Education, Science, Technology, Sports and Culture of Japan and a grant from Japan Research Foundation for Clinical Pharmacology to K.Y.