Individuals living in malaria-endemic areas show generally low T cell responses to malaria Ags. In this study, we show murine dendritic cell (DC) interaction with parasitized erythrocytes (pRBC) arrested their maturation, resulting in impaired ability to stimulate naive, but not recall T cell responses in vitro and in vivo. Moreover, within the naive T cell population, pRBC-treated DC were selectively deficient in priming CD8+ but not CD4+ T cells. Indeed, DC that had taken up pRBC were shown for the first time to efficiently prime CD4+ T cell responses to a known protective merozoite Ag, MSP4/5. In contrast, impaired priming resulted in decreases in both proliferation and cytokine production by CD8+ T cells. Deficient priming was observed to both a model and a Plasmodium berghei-specific CD8+ T cell epitope. The mechanisms underlying the inability of parasite-treated DC to prime CD8+ T cells were explored. pRBC treatment of DC from wild-type C57BL/6, but not from IL-10 knockout animals, suppressed DC-mediated T cell priming across a Transwell, suggesting active IL-10-dependent suppression. CD8+ T cells were arrested at the G0 stage of the cell cycle after two cell divisions post-Ag stimulation. The proliferation arrest was partially reversible by the addition of IL-2 or IL-7 to responder cultures. These results suggest that in malaria-endemic areas, priming of CD8+ T cell responses may be more difficult to induce via vaccination than the priming of CD4+ T cells. Moreover, pathogens may selectively target the CD8+ T cell arm of protective immunity for immune evasion.
Malaria remains the most important tropical disease afflicting humans, with recent mortality and incidence estimates of 2–3 million deaths and 300–500 million clinical cases annually (1). A better understanding of the immune response to the malaria parasite and the nature and regulation of protective mechanisms would facilitate the development of a much needed vaccine. A number of mechanisms have been proposed to explain how the asexual erythrocytic parasites survive in the face of the host’s immune response, including antigenic diversity, clonal Ag variation, and altered peptide ligand antagonism (reviewed in Ref.2).
Low numbers of Plasmodium falciparum-specific memory T cells, particularly CD8+ T cells, have been observed in individuals living in malaria-endemic areas (3, 4, 5, 6, 7). Impaired APC function as a result of exposure to parasitized erythrocytes (pRBC)5 may contribute to this observation. Indeed, Urban et al. reported that exposure of human dendritic cells (DC) to pRBC promoted a maturation defect, and although the consequences of this interaction for immunity were not assessed in vivo, a subsequent impairment in T cell priming was apparent in vitro (8).
As well as recent reports of pRBC-induced impairment of DC maturation (9), there are reports suggesting the opposite: pRBC-induced DC activation and up-regulation of costimulatory molecules (10). Results from the present study supported and extended the former finding, showing a reproducible DC-maturation defect induced by both Plasmodium yoelii nigeriensis lethal (P. yoelii) or Plasmodium chabaudi chabaudi (P. chabaudi) pRBC. Further investigating this observation, the ability of pRBC-exposed DC to stimulate CD4+ as compared with CD8+ T cells was determined, as was their ability to prime recall compared with nonrecall T cell responses.
Our results suggest that although maturation-arrested DC have a specific defect in priming of naive CD8+ T cells, other important functions can remain largely unaffected. The mechanism of this impaired immunity was found to be novel, with active IL-10-mediated suppression of proliferation after two CD8+ T cell divisions resulting in G0-arrested T cells. The arrest was partly reversible by the addition of IL-2 or IL-7.
Materials and Methods
C57BL/6 (H-2Kb), BALB/c (H-2Kd), and IL-10BL/6KO (H-2Kb) female mice, 6–8 wk old were used in the experiments. Mice were bred at the Austin Research Institute Biomedical Animal Research Laboratory. The Institutional Animal Care and Use Committee approved all animal procedures.
P. yoelii and P. chabaudi (AS)-infected erythrocytes were generated in mice from frozen stock (kindly provided by Ross Coppel, Monash University, Melbourne, Australia) and freshly harvested for i.p. injection into naive mice. Mice were monitored for “percent parasitaemia” daily, and a collection of moderately (5–10%) and highly (40–50%) parasitized blood was obtained by eye bleed. The blood was washed twice in sterile PBS, and packed RBC were used at 10 RBC per one cultured DC. Other pRBC:DC ratios, 1:1 and 1:10, were used in preliminary experiments. After this preliminary titration, we used a ratio of 10:1 throughout the studies shown in this study.
Generation and characterization of murine bone marrow-derived DC
Bone marrow cells from C57BL/6 female mice were cultured at 106 cells/ml. Petri dishes contained conditioned RPMI 1640 medium (CSL) supplemented with 1000 U/ml GM-CSF, 10 ng/ml IL-4, 10% (v/v) heat-inactivated FCS, 4 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin sulfate, and 100 μM 2-ME. At day 6 the cells were washed, resuspended in the same cultured media at 106 cells/ml, and incubated overnight with RBC or P. yoelii or P. chabaudi pRBC. Maturation of in vitro grown DC was achieved by adding 1 μg/ml LPS, 100 ng/ml TNF-α, and further 4 h incubation. DC for flow cytometry were washed and incubated with PBS and 5% (v/v) normal human serum for 30 min on ice with mAbs to CD11c and costimulatory molecules, CD40, CD80, and CD86, labeled with PE or FITC (BD Pharmingen). Cells were washed and analyzed using a BD Biosciences FACSCalibur machine. PE- or FITC-labeled mAb isotype controls were used as negative controls (BD Pharmingen). Data were analyzed using CellQuest software (BD Biosciences).
DC cultured for 6 days in GM-CSF and IL-4-supplemented media were left untreated or pulsed with RBC or Plasmodium pRBC overnight. Pulsed DC were then washed thoroughly and resuspended at 107 cells/ml in PBS, and 100 μl aliquots (106 cells) were injected without further treatment intradermally into the footpads of experimental mice or further pulsed in vitro for 3 h with peptides and washed with RPMI 1640 (three times) before injection. Peptides corresponding to the CD8+ T cell epitope SIINFEKL and the CD4 helper epitope OVA323–339 from OVA protein were pulsed at 5 mg/ml and 25 μg/ml. Responses to these peptides were assessed in H-2kb animals (C57BL/6) using C57BL/6-derived DC. The CD8+ T cell epitope from the circumsporozoite protein of Plasmodium berghei, pb9, was also pulsed at 5 μg/ml and 25 μg/ml. In this case, H-2d BALB/c mice were used both as responders and as the source of DC.
In previous experiments, testing splenocyte reactivity to OVA CD8+ and CD4+ T cell epitopes, comparable and optimal CD8+, and CD4+ T cell IFN-γ ELISPOT responses were found 10–16 days after the final immunization. Spleens were therefore isolated for analysis 10–14 days after immunization in this study. A total of 106 splenocytes was incubated in conditioned media (control), or with experimental Ags for 18 h on prepared ELISPOT-mixed acetate wells (MAHA; Millipore) precoated with rat anti-mouse IFN-γ (clone R4 from ETCC) as described previously (9). Triplicate wells were set up for each condition. To test for CD8+ T cell responses, the peptide corresponding to the SIINFEKL epitope was used at 0.5 μg/ml and 5 μg/ml or the P. berghei circumsporozoite protein-derived pb9 peptide used at 0.5 μg/ml and 5 μg/ml. To test for CD4+ T cell effector responses, the helper epitope OVA323–339 was added at 2.5 μg/ml and 5 μg/ml or the recombinant Plasmodium yoelii merozoite surface protein 4/5 (MSP4/5) (11) at 25 μg/ml. Cells were discarded after 18 h, and plates were incubated for 2 h with biotinylated anti-IFN-γ mAb (XMG.21-biotin) (BD Pharmingen) followed by extravidin-alkaline phosphatase (AP) at 0.1 μg/ml (Sigma-Aldrich). Spots of activity were detected using a colorimetric AP kit (Bio-Rad), counted using an ELISPOT reader (Autoimmun Diagnostika), and scored using the AID ELISPOT software version 2.0 (Autoimmun, Diagnostika). Data are presented as mean spot-forming units (SFU) per million cells ±SD of the mean.
CD8+ and CD4+ T cell purification
BALB/c mice were immunized with DC pulsed with either nonparasitized or pRBC. Fourteen days later, splenocytes were isolated and counted. T cells were purified by Ab-bead depletion using Ab against the following surface molecules: CD11b, GR-1, B220, TER119, and MHC class II. To purify CD4+ or CD8+ T cells, Ab against CD8α or CD4 were additionally used. Two rounds of magnetic bead depletion were performed using goat anti-rat Ig-coated beads (Paesel & Lorei), and purity was confirmed to be >95% as described previously (10).
T cell proliferation assay
C57BL/6 mice were immunized with in vitro grown DC pulsed with either the CD8 T cell epitope SIINFEKL alone (5 μg/ml) or SIINFEKL with pRBC. Ten to 14 days later, splenocytes were isolated from each mouse and stimulated in culture with in vitro grown DC pulsed with nothing (control), SIINFEKL (0.5 μg/ml) alone (KL), SIINFEKL and RBC (KL-RBC), or SIINFEKL and pRBC (KL-pRBC) for 24 h in a 96-well plate (2 × 105cell/well). Overnight culture was chosen to correspond with the overnight incubation for the IFN-γ ELISPOT assay. Triplicates from each well were pulsed with 1 μCi/ml tritiated thymidine overnight, and incorporation into DNA was assessed using a Top Count machine (Packard Instrument).
DC were left untreated or pulsed with RBC (control), P. yoelii, or P. chabaudi pRBC overnight, and matured by stimulation with LPS and TNF-α for 24 h. A total of 105 pulsed DC was incubated with 106 responder splenocytes from BALB/c for 72 h, then triplicate samples were pulsed overnight with 1 μCi/ml tritiated thymidine, and incorporation was assessed the next day. Transwell MLR assays were conducted similarly, 105 pulsed DC were incubated with 106 responder splenocytes from BALB/c mice in a 24-well plate. In these experiments, 105 GM-CSF/IL-4 in vitro grown DC left untreated (control) or pulsed with P. yoelii or P. chabaudi pRBC were further added to the MLR cultures separated by a 0.4-μm Transwell (BD Biosciences).
Cytokine MLR assay
After 72-h exposure to allogeneic cells, lymphocytes from MLR assays were aliquoted at 106 cells per well onto mixed acetate plates (MAHA; Millipore) coated with rat anti-murine IFN-γ (clone R4 from ETCC) or rat anti-mouse IL-10 (BD Pharmingen) as described previously (9) and incubated for 18 h without in vitro stimulation. Triplicate wells were set up for each condition. Cells were discarded, and plates were incubated for 2 h with anti-IFN-γ mAb biotin (BD Pharmingen) or anti-IL-10 mAb biotin (BD Pharmingen) followed by extravidin-AP at 0.1 μg/ml (Sigma-Aldrich). Spots of activity were detected using a colorimetric AP kit (Bio-Rad) and counted as described above. Data are presented as mean SFU per million cells ±SD of the mean.
Cell division analysis
Cells from MLR assays described above were pulsed with CFSE (Molecular Probes). A total of 5 μg/ml CSFE diluted in potassium/phosphate buffer was used to stain 107 cells for 15 min at 37°C. Cells were then washed twice in RPMI 1640 media (CSL), incubated for a further 72 h at 37°C, then washed in PBS, and analyzed by flow cytometry. rIL-2 (BD Pharmingen) was added to cultures at 10 U/ml and IL-7 (kindly donated by Dr. Vaios Karanikas, Austin Research Institute, Melbourne, Australia) at 10 ng/ml on day 0 of MLR culture. As described in Ref.12 , quantitative analysis of division of CD8 and CD4 T cells in a MLR response was performed after stimulation with normal and pRBC. Histograms of cells were obtained 72 h after CFSE staining, and cells were gated based on expression of CD4 or CD8 markers, identified using PE-labeled MAbs and appropriate isotype controls (BD Pharmingen). Each division was calculated on the basis that upon each division, the fluorescence of CFSE becomes half of the previous load. This was calculated for each histogram plot, and markers were set for each division to determine the proportion of cells present. The area under each histogram (proportion of cells) was calculated.
DNA staining using propidium iodide (PI)
For experiments that involved PI DNA staining, 106 cells from MLR cultures were washed in PBS twice and permeabilized in PBS containing 0.3% saponin for 20 min. PI (1 μg/ml; ICN Chemicals) was added to 106 cells, and the preparation was fixed as described previously (13, 14). These were costained with FITC-labeled MAbs to T cell markers CD4 or CD8 and analyzed using flow cytometry. Analysis was performed according to procedures described in Ref.12 . Briefly, cells gated on the CD4+ or CD8+ FITC-positive population on the FL1 channel were plotted on linear histogram plots on the FL3 channel, which detects PI staining. The regions of the plot corresponding to G0, S, or G2 were calculated based on the fact that cells at G0 have half the DNA content of those at G2, and S represents the transition stage during which variable quantities of DNA are present (12).
Assays were set up in triplicate. Mean values were compared using the two-tailed unpaired t test. Two p-value thresholds are noted in the text for all immunogenicity assays: ∗∗, p < 0.001, to indicate a highly significant difference; and ∗, p < 0.05, to indicate a significant difference.
Murine DC maturation defect upon interaction with pRBC results in impaired priming of CD8+ but not CD4+ T cell responses in vivo
We assessed the effect of preincubation with Plasmodium yoelii pRBC on the expression of costimulatory molecules CD40, CD80, and CD86 by DC in vitro. Direct DC treatment did not change CD40 or CD80 costimulatory molecule expression from basal levels; however, the number of DC expressing CD86 was significantly decreased (immature DC; Fig. 1,A). Addition of LPS and TNF-α to DC normally results in up-regulation of costimulatory molecules and a fully mature DC phenotype (8, 15, 16). Consistent with some previous studies, pRBC pretreatment of DC significantly decreased the number of DC that up-regulated CD40, CD80, and CD86 in response to these maturation stimuli (mature DC; Fig. 1,B) (8, 15, 16). MHC class I and II expression remained unchanged by pRBC treatment of either immature or mature DC (data not shown). Individual flow cytometry plots are shown to illustrate that mean fluorescence intensity (MFI) as well as percentage costimulatory marker-positive cells were both decreased, and that similar results were observed using P. yoelii and P. chaubadi pRBC (Fig. 1,C). The pRBC-pretreated DC were less efficient in stimulating in vitro MLR responses than normally matured DC (Fig. 1 D; p < 0.001).
To compare the in vivo effect of pRBC interaction with APC on their ability to prime CD8+ and CD4+ T cell responses, we further pulsed RBC- or pRBC-treated DC in vitro with peptides representing specific OVA protein CD8+ (SIINFEKL) or CD4+ (OVA323) T cell epitopes. Low (Fig. 2, A and C) and high (Fig. 2, B and D) doses of Ag were pulsed, and mice were immunized with these DC. Fig. 2, A and B, show that induction of IFN-γ-producing CD8+ T cells was significantly impaired in mice primed with DC treated with P. yoelii pRBC (∗∗, p < 0.001). In contrast, the precursor frequency of IFN-γ-producing CD4+ T cells in response to OVA323–339 was not affected (Fig. 2, C and D). Similar impairment of CD8+ but not CD4+ T cell stimulation was observed using either optimal (25 μg/ml; Fig. 2,C) or suboptimal (5 μg/ml; Fig. 2 D) doses of peptide to pulse the DC for injection, or to restimulate the IFN-γ response in vitro.
Studies with malaria parasite Ags
Further studies using the protective CD8+ T cell epitope from the CS protein of P. berghei, pb9, as the Ag were performed. Mice immunized with DC pulsed with pb9 at 25 μg/ml and 5 μg/ml exhibited high IFN-γ responses as measured by ELISPOT assay (Fig. 2, E and F, respectively). When DC were pulsed with pb9 together with P. yoelii pRBC, IFN-γ responses were significantly decreased at all peptide concentrations tested (Fig. 2, E and F). These results confirm our original data using the model CD8+ T cell epitope from OVA, SIINFEKL, demonstrating that CD8+ T cell responses may be generally inhibited, regardless of the CD8+ T cell epitope. In addition, experiments with SYIPSAEKI were performed in BALB/c mice and those with SIINFEKL in C57BL/6, thus this CD8 T cell priming deficiency was observed across mouse strains, further expanding the generality of this finding.
We then investigated the induction of CD4+ IFN-γ-producing T cells to a natural protein expressed by blood-stage merozoites, merozoite surface protein 4/5 (MSP4/5), after immunization with pRBC-pulsed DC. Fig. 2 G shows that 14 days after immunization with pRBC-pulsed DC, purified CD4+, but not CD8+ T cells, responded specifically in vitro by IFN-γ production to MSP4/5. These results support the above observations, confirming the ability of pRBC-pulsed DC to induce substantial CD4+ T cell responses, not only to a model CD4+ helper epitope from OVA, but also to an Ag expressed naturally by the parasite. Moreover, our recent studies also show that mice immunized with pRBC-pulsed DC (whether left untreated or further matured with LPS and TNF-α) exhibit strong IFN-γ responses to pRBC lysate and protect against a blood-stage P. yoelii challenge (15).
pRBC treatment of DC impairs their ability to prime CD8+ T cells capable of cytokine production and proliferation
Recent studies suggest that T cell cytokine production and proliferation may be elicited from different T cell subsets (17, 18). The observed parasite-induced CD8+ T cell priming defect was therefore further tested in the context of the induction of SIINFEKL-specific T cells capable of responding by proliferation to SIINFEKL. Fig. 3 shows that as for IFN-γ production, proliferative responses were severely impaired by pRBC-pulsed DC treatment.
The capacity of pRBC to impair the ability of DC to restimulate CD8+ T cell responses was then investigated. C57BL/6 mice were primed with DC pulsed with SIINFEKL to generate recall T cells. These were then used as responder cells in in vitro assays. There was no significant difference between the ability of untreated DC-SIINFEKL and those further preincubated with P. yoelii pRBC to restimulate SIINFEKL-specific CD8+ T cells to produce IFN-γ (Fig. 3). In this experimental series, the same pRBC-treated DC preparations were severely impaired in their ability to prime SIINFEKL-specific T cells in vivo (Fig. 3; ∗∗, p < 0.001), confirming previous results. These results indicate that initial priming, rather than the restimulation of CD8+-specific T cells, is being affected.
Mechanism: proliferation arrest after two division cycles of CD8+ but not CD4+ T cells in the same MLR reaction
CD8+ T cell priming can be enhanced under some conditions by CD4+ T cell help. To confirm a selective defect in priming of CD8+ but not CD4+ T cell responses, CD8+ compared with CD4+ T cell division in an MLR assay was quantitated in parallel (Fig. 4). Both cell types were activated and cultured together and, therefore, in the same microenvironment throughout the assay. BALB/c splenocytes were labeled with CFSE, and stimulated in an MLR response with matured allogeneic (C57BL/6) DC prepulsed with RBC (DC:RBC) or P. yoelii pRBC (DC:PRBC). Cell divisions were assessed after 72 h of culture by CSFE dilution flow cytometry analysis with selective gating on CD4+ or CD8+ T cells. CD8+ cells stimulated with DC:pRBC had a different pattern of cell division to control cultures stimulated by DC:RBC. Most of the DC:pRBC-stimulated CD8+ T cells completed two rounds of division, but none progressed to three or four, whereas in control DC:RBC-stimulated cultures, >30% of cells progressed to three and >5% to four cell divisions (Fig. 4,A). This result suggests that cell cycle arrest prevents further CD8+ cell expansion after two divisions in DC:pRBC-stimulated cultures. In contrast, cell division of CD4+ cells from the same cultures was similar in groups that were stimulated with either DC:RBC or DC:pRBC (Fig. 4 B).
Mechanism: active parasite-induced primary CD8+ T cell proliferation arrest is partially reversed by IL-2 or IL-7
The arrest in CD8+ T cell proliferation during priming was further investigated by quantifying the progression of T cells throughout the cell cycle in an MLR response. BALB/c splenocytes were stimulated for 72 h with C57BL/6-derived DC prepulsed with RBC or P. yoelii pRBC. Fig. 5 shows histogram plots of the DNA content of T cells (detected by intracellular staining with PI) positively gated on CD8+ or CD4+ cells as identified by costaining with FITC-labeled Abs. As illustrated in Fig. 5,A, G0 (resting T cells), S (blast/dividing cells), and G2 (duplicated DNA) phases of the cell cycle can be separately examined by differential cellular DNA content using linear analysis on the FL3 channel by FACScan (13, 14). Fig. 5,A shows that in MLR cultures, CD8+ T cells stimulated with pRBC-pulsed DC (black line) were primarily at the G0 phase of the cell cycle (G0, 63%; S, 36%; and G2, 1%), whereas those stimulated with RBC-pulsed DC (gray line) were primarily at S phase (G0, 11%; S, 64%; and G2, 20%). CD4+ T cells in the same cultures showed no difference in cell cycle distribution between DC:pRBC (G0, 10%; GS, 59%; and G2, 25%) or DC:RBC (G0, 8%; S, 58%; and G2, 25%) treatment, with most cells at the S phase of the cell cycle (Fig. 5 A). These results were consistent with all of the previous results showing a selective defect in CD8+ but not CD4+ T cell priming, and further implicate proliferation arrest at G0 as the mechanistic basis of this defect.
IL-2 can reverse cell cycle arrest in tolerized T cells by promoting induction of the S phase (19, 20, 21). Addition of exogenous IL-2 at 10 U/ml largely reversed the G0 phase cell cycle arrest in CD8+ T cells stimulated with DC:pRBC (black line), resulting in clear progression into S phase (G0, 22%; S, 64%; and G2, 12%) (Fig. 5,B), and a more comparable cell cycle distribution to cultures stimulated with DC:RBC and IL-2 alone (gray line) (G0, 6%; S, 55%; and G2, 30%); however, reversal was not complete (Fig. 5 B).
Cytokines such as IL-7 share many pro-proliferative functions with IL-2, and have also been used to reverse anergic or tolerized cells (22). As observed with IL-2 in this study, addition of 10 ng/ml IL-7 (Fig. 5,C) was able to partially reverse the DC:pRBC (black line)-induced CD8+ T cell cycle arrest (G0, 23%; S, 65%; and G2, 11%), promoting a cell cycle distribution more comparable to cultures stimulated with DC:RBC and IL-7 alone (gray line) (G0, 4%; S, 57%; and G2, 30%). In the same cultures, exogenous cytokine addition had no effect on cell cycle distribution in CD4+ T cells stimulated with either DC:RBC or DC:pRBC (Fig. 5, B and C). Therefore, treatment of DC with pRBC resulted in CD8+ (but not CD4+) T cell arrest at the G0 phase of the cell cycle, with a characteristic IL-2 (and IL-7) partial reversibility profile shared with classical anergic T cells (22).
Mechanism: parasites promote IL-10 production by DC, which actively suppresses T cell proliferation
We sought to further distinguish between an inability of the DC:pRBC to stimulate T cells and active suppression of an MLR response. Matured DC from C57BL/6 mice were treated with either pRBC or RBC, washed and added, separated by a Transwell, to a BALB/c splenocyte MLR response to irradiated C57BL/6 DC. Proliferation was significantly reduced in cultures exposed to DC:pRBC compared with DC:RBC (Fig. 6,A; p < 0.001), suggesting active suppression by a soluble mediator. To investigate possible candidates, production of cytokines by the suppressed MLR cultures was assessed using ELISPOT assays. IFN-γ production was significantly decreased in cultures containing P. yoelii (∗∗, p < 0.001) and P. chabaudi (∗, p < 0.05) pRBC-pulsed DC (Fig. 6,B). As well as decreased IFN-γ production, these cultures exhibited increased IL-10 production following exposure to P. yoelii (∗∗, p < 0.001) and P. chabaudi pRBC (∗, p < 0.05) -pulsed DC, as compared with cultures exposed to DC:RBC (Fig. 6 C). There was no detectable IL-10 in supernatants of DC pulsed with pRBC using either the cytokine ELISA or ELISPOT assay (data not shown). Thus, IL-10 was only detectable after the addition of splenocytes, suggesting further secretion by T cells.
IL-10 is a cytokine capable of suppressing proliferative responses and whose production can be elicited from DC under specific circumstances (23, 24). We tested DC from IL-10 knockout (KO) mice on C57BL/6 background, pulsed with P. yoelii pRBC, for their ability to stimulate an MLR response using splenocytes from BALB/c mice as responder cells. In contrast to our results comparing RBC- and pRBC-treated DC from normal C57BL/6 mice (Figs. 3–5 and 6,A), IL-10 KO pRBC-treated DC did not have impaired ability to stimulate MLR responses, compared with untreated KO DC (Fig. 6 D). These results further supported the finding that IL-10 is a critical suppressive factor associated with parasite-induced DC-mediated immunomodulation.
There has been considerable interest in immune evasion strategies used by the malaria parasite to evade their host. In this study we show that two Plasmodia strains are able to prevent the maturation of murine DC, leading to decreased expression of costimulatory molecules upon further stimulation with “danger signals” and to a failure to stimulate the priming of Ag-specific CD8+ T cells. In our study, the induction of both IFN-γ-producing and -proliferating CD8+ T cells was impaired. Interestingly, defects in DC maturation did not affect their ability to induce CD4+ T cells in vivo or restimulate recall CD8+ T cell responses. The mechanism of impaired CD8+ T cell priming was found to be based on a CD8+ T cell cycle arrest at the G0 stage after two rounds of cell division. The arrest was partially reversible in vitro by cytokines able to reactivate anergic T cells: IL-2 and IL-7. We speculate that malaria-exposed individuals could harbor large numbers of such cell cycle-arrested CD8+ but not CD4+ T cells. Our results predict that malaria-specific CD4 T cell responses, and consequently Ab responses associated with protection against blood-stage malaria, would be effectively induced by DC that had taken up pRBC.
The pRBC-treated DC had decreased levels of CD86 and were incapable of up-regulating CD40, CD80, or CD86 upon further stimulation with LPS and TNF-α. There was no significant difference in surface expression or MFI of MHC class I and II, suggesting that limited Ag presentation on MHC class I did not underlie their inability to prime CD8+ T cell responses. It is known that DC stimulation via the Fc receptor (25), signaling via the CD40 molecule (usually by interaction with CD40L-expressing CD4+ T cells) (26), or by addition of danger signals that may interact with Toll-like receptors (27, 28), results in CD86 up-regulation and renders DC further capable of effectively priming CD8+ T cells: the “DC licensing” phenomenon. DC may also be susceptible to modulation of costimulatory molecule expression following pRBC interaction in vivo (16, 29). The low CD86+ levels on both “immature” (in this context referred to as cells that have not received exogenous maturation signals) and mature (after receiving the danger signals) DC upon pRBC treatment are likely to limit their CD8+ T cell-stimulating capacity. In contrast, the levels of other costimulatory molecules, notably the unchanged CD40 levels on immature DC, may remain sufficient to promote adequate CD4+ T cell priming. In a recent study this prediction was validated, with both immature and LPS/TNF matured pRBC-pulsed DC able to prime protection (15). In this study, we further confirmed DC that had taken up pRBC can induce strong CD4+ T cell responses to a specific protein expressed by the parasite, MSP4/5. It may have been further interesting to use Abs against these costimulatory molecules to block their function; however, such Abs can themselves exert a regulatory function that would complicate the interpretation of the findings, for example, an anti-CD80 mAb can induce IL-10 secretion from murine bone marrow DC in vitro (26).
The functions of DC are also influenced by cytokines in the local microenvironment. IL-10 is present during acute malaria infection and may contribute to the suppression of T cell effector responses such as proliferation and Th1 cytokine production (30). IL-10 suppresses the expression of costimulatory molecules including CD80 and CD86 in monocyte (24) and bone marrow-derived DC (31). Production of important Th1 type cytokines, such as IFN-γ, TNF-α, and IL-12, is also suppressed, which, in turn, may hamper the priming of naive T cells (32). IL-10-induced suppression may be particularly relevant for individuals living in malaria-endemic areas, who evidently produce significant levels of this cytokine (33). This suggested to us that IL-10 might be associated with a maturational defect of DC, leading to impaired priming of CD8+ T cell responses. Indeed, a soluble suppressor mediator was induced during DC interaction with pRBC, because suppression of proliferative responses was observed across a Transwell. When DC from C57BL/6 IL-10 KO were used in similar experiments, no suppression was observed, suggesting IL-10 as an important molecular mediator in this process. The central role for IL-10 was further supported by the reversal of suppression in MLR cultures affected by pRBC-treated DC from C57BL/6 mice by using a neutralizing anti-IL-10 mAb (data not shown). However, the simple interaction of the pRBC with DC was insufficient to generate detectable levels of IL-10 by ELISA or ELISPOT, suggesting either that a minimal amount of IL-10 is required to generate a further positive feedback loop for IL-10 production by both DC and T cells (34), or that pRBC-treated DC may secrete some as yet uncharacterized soluble factor, which, in turn, promotes the IL-10 production required to suppress the MLR. In this context, it is interesting to note that IL-10 can promote the induction of anergic and suppressor cells other than the well-known IL-10-producing Tr1 phenotype, for example, a subset of CD4+ CD25+ regulators (27, 34).
IFN-γ production is associated with malaria-protective immunity (5, 9, 33), and recent studies indicate that DC can induce protective CD8+ T cell immune responses to liver stage infection (28). There is also evidence showing that sporozoites can be phagocytosed by DC that then induce specific CD8+ T cells (45). Therefore, the ability of the malaria parasites to impair DC maturation and selectively suppress CD8+ T cell proliferation is likely to compromise the generation of protective malaria immunity in endemic populations. Urban et al. (8) have shown that erythrocytes infected with some strains of P. falciparum can impair human DC maturation in vitro, suggesting that this phenomenon is not restricted to murine malaria parasites. Indeed, the selective defect in CD8+ T cell priming is consistent with the low precursor frequency of CD8+ T cells specific for P. falciparum Ags evident in humans living in malaria-endemic areas (3, 4, 6, 7, 35, 36).
Examples of inhibition of DC maturation with particular reference to expression of costimulatory molecules include cancer (multiple myeloma), cytomegalovirus, HIV, measles, herpes simplex virus, Trypanosoma cruzi, and Mycobacterium tuberculosis (35, 36, 37, 38, 39, 40, 41, 42, 43). The possibility of a selective defect in CD8+ T cell priming has not been studied with regard to these diseases.
Impaired ability of DC to prime CD8+ T cells in malaria-infected individuals may impair more than malaria-specific immunity. Indeed, in malaria-endemic areas other infectious diseases have been shown to elicit low CD8+ T cell responses (3, 4, 5, 6, 7). In practical terms, the present study suggests that blood-stage malaria infection may interfere with efficient CD8+ T cell priming in vivo by natural exposure or by vaccination. In this study, we show a cell cycle arrest of CD8+ T cells during priming. Other studies have shown that factors produced by DC such as IL-10 can induce anergic T cells and result in cell cycle arrest (27). As well as cytokine-induced cell cycle arrest, T. cruzi, HIV, and bacteria such as Escherichia coli can reportedly induce cell cycle arrest at S phase in anergic T cells in vitro via alteration of IL-2 secretion (37, 38, 39, 40). Kierszenbaum et al. (38) showed that adding exogenous IL-2 to in vitro grown human T cells infected with T. cruzi partially reversed cell cycle arrest. We observed similar results, with exogenous IL-2 or IL-7 re-establishing cell cycling in suppressed CD8+ T cells that had been stimulated with pRBC-treated DC. By using MLR to study CSFE cell division arrest, we ensured that both CD8+ and CD4+ T cells were stimulated to proliferate in the same well, and particular factors such as cytokines could potentially affect both CD8+ and CD4+ T cells in the same culture, providing a rigorous test for being differentially affected, despite being in the same microenvironment.
Malaria-associated immunosuppression has been associated with the presence of high numbers of circulating anergic T cells, particularly in individuals with chronic infection living in malaria-endemic areas (41, 42, 43). We hypothesize that this type of malaria-associated immunosuppression occurs as a consequence of cell cycle arrest of CD8+ T cells due to parasite interaction with DC. In our study, DC:pRBC pulsed with pb9 completely failed to induce a significant CD8+ T cells response, indicating that none of the potential DC subsets present in these cultures was capable of priming such cells. However, given the plethora of DC subsets with differential T cell stimulatory function in vitro and in vivo, including, for example, DC that may be specifically deficient in activating IL-2 production from CD8+ T cells (44), we have performed studies to address whether culturing DC under different conditions could generate DC subsets resistant to the pRBC maturation impairment described in this study. Our studies suggest that certain less mature DC subsets may be, surprisingly, less susceptible (D. S. Pouniotis et al., manuscript in preparation). Also, on a positive note, the lack of effect on the induction of CD4+ T cells indicates that pRBC-exposed DC may still be able to fully induce protective blood-stage immunity. The ability of DC, even after treatment with pRBC, to restimulate recall CD8+ T cell responses also supports the development of liver stage DC-targeted vaccines designed specifically to restimulate pre-existing CD8+ T cell immunity.
We thank Dr. Annemiek van Spriel for assistance with CFSE study analysis; Dr. Vaios Karankas for supplying rIL-7; and Professor Mauro Sandrin, Professor Ian Mckenzie, and Dr. Denise Jackson for useful comments on the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
D.S.P. was supported by National Health and Medical Research Council Grant 223302. O.P. was supported by a LaTrobe University Doctor of Philosophy Scholarship. M.P. and R.L.C., Howard Hughes Medical Institute International Scholars, are supported by National Health and Medical Research Council Program Grant 334012. M.P. is a National Health and Medical Research Council Senior Fellow.
Abbreviations used in this paper: pRBC, parasitized erythrocyte; DC, dendritic cell; P. yoelii, Plasmodium yoelii nigeriensis lethal; P. chabaudi, Plasmodium chabaudi chabaudi; AP, alkaline phosphatase; SFU, spot-forming units; PI, propidium iodide; MFI, mean fluorescence intensity; KO, knockout.