Dendritic cells (DCs) are antigen-presenting cells (APC) that play a central role in both innate and adaptive immune responses. To initiate T cell-dependent immune responses to microbial infections, DCs phagocytose antigens in peripheral tissues and migrate to the draining lymph nodes, where they interact with antigen-specific T cells. Maturation of DCs, involving up-regulation of the major histocompatibility complex (MHC) and peptide complexes and the costimulatory molecules at the surface, is required to efficiently prime naïve T cells 1. Upon maturation, DCs reorganize their actin cytoskeleton, projecting long and motile membrane extensions, called dendrites. The initial encounters between antigen-presenting DCs and specific naïve T cells are characterized by the directional projection of abundant membrane extensions from the DC toward the naïve T cell, followed by entrapping of the T cell within a complex net of membrane extensions 2.

The activation of T cells by DCs during Plasmodium infection has been previously studied. Although different effects have been described depending on the parasite strain used, time after infection or subpopulation of DC analysed, a number of reports found defective activation of T cells 3. These findings may be related with the low parasite-specific T cell responses induced by human malaria infections 45.

This report shows that DCs from Plasmodium yoelii-infected mice are able to present antigens associated with MHC-II, but do not establish strong interactions with naïve CD4⁺ T cells. Accordingly, it was also found that activation of naïve CD4⁺ T cells is inhibited during late malaria infections.

DCs present exogenous antigens in the context of MHC class II molecules for the activation of CD4⁺ T cells. DCs are also able to cross-present exogenous antigens in the context of MHC-I molecules to activate CD8⁺ T cells. It was determined whether MHC-II and MHC-I antigen presentation was affected during infection with P. yoelii. For this purpose, specific antibodies were used, that recognize defined peptide epitopes bound to particular MHC molecules. These antibodies do not recognize the antigen or the MHC molecules alone and can be used to determine the presence of MHC-epitope bound complexes in the surface of antigen presenting cells. To determine MHC-I and MHC-II antigen presentation anti-OVA epitope 257–264/H-2^b (25-D1.16) and anti-LACK epitope 156–173/I-A^d (2C44) antibodies were used, respectively. Bone-marrow derived DCs were incubated either alone, with uninfected or P. yoelii-infected erythrocytes for 24 h before addition of purified OVA or LACK proteins for 5 h, followed by stimulation with LPS to induce antigen presentation. It was observed that incubation with P. yoelii-infected erythrocytes inhibits presentation of exogenous antigens on MHC-I (Figures 1A and 1B), but not on MHC-II molecules (Figures 1C and 1D).

The activation of T cell hybridomas that recognize specific OVA peptides in the context of MHC-I and II molecules was used as an alternative method to determine antigen presentation. The T cell hybridomas B3Z 9 and DO.11.10 10 recognize specific OVA peptides on MHC class I (H-2^b) and II (I-A^d), respectively. T hybridoma cells are less dependent than primary naive T cells on costimulatory molecules and they can be activated by fixed antigen presenting cells. Recognition of peptide-MHC complexes by T hybridoma cells results in increased IL-2 secretion. DCs were incubated with uninfected or P. yoelii-infected erythrocytes for 24 h before addition of purified OVA proteins for 5 h. DCs were then fixed to prevent secretion of cytokines and incubated with the different hybridomas for 24 h. Activation of hybridomas was determined by detection of IL-2 in the incubation medium. It was found that DCs pre-incubated with P. yoelii-infected erythrocytes activate the T cell hybridoma recognizing OVA-MHC-II, but the activation of the hybridoma recognizing OVA-MHC-I was significantly reduced. The activation of the hybridomas was dose dependent (Figures 2A and 2B).

Maturation of DCs increases their capacity for antigen presentation, since it increases the expression of MHC molecules on the cell surface 1. As expected, it was found that the addition of LPS to induce maturation of DCs 5 h after incubation with OVA, resulted in increased activation of both hybridomas recognizing OVA-MHC-I and II (Figures 2C and 2D). The effect of pre-incubating DCs with P. yoelii-infected or uninfected erythrocytes in antigen presentation of OVA by MHC-I and MHC-II after inducing maturation with LPS was analysed. It was again observed that P. yoelii-infected erythrocytes inhibit the activation of the hybridoma recognizing OVA-MHC-I, but not OVA-MHC-II (Figures 2E and 2F).

The capacity of CD11c⁺ DCs isolated from P. yoelii-infected mice to present antigens by MHC-II was analysed. It was found that that MHC-II antigen presentation is maintained during the course of infection, but MHC-I cross-presentation of exogenous antigens is inhibited throughout the course of the disease (Figures 2G and 2H). These results confirm previous observations 11 that MHC-I antigen presentation is inhibited during Plasmodium infections and indicate that DCs are able to process and present exogenous antigens by MHC-II during P. yoelii infections.

The initial encounters between antigen-presenting DCs and specific naïve T cells include directional projection of abundant membrane extensions from the DC toward the naïve T cell and prolonged interactions between both 2. To analyse whether Plasmodium interferes with this process, the interactions between bone marrow derived DCs pre-incubated with infected erythrocytes and naïve CD4⁺ T cells were first studied using time-lapse video microscopy. DCs were pre-incubated with uninfected or P. yoelii-infected erythrocytes and incubated with naïve anti-OVA CD4⁺ T cells isolated from transgenic mice (DO11.10) recognizing a specific OVA epitope. DCs were loaded with the same peptide epitope (OVA 323–339) before addition of T cells. DCs presenting antigen normally interact with antigen-specific naïve T cells for long periods of time (more than 5 min) with abundant membrane extensions projected in the direction of the T cell 8. Pre-incubation of DCs with uninfected erythrocytes did not affect DC-T cell interactions, as prolonged interactions with membrane extensions were frequently found in the co-cultures (Figure 3A and Additional file 1). In contrast, pre-incubation with P. yoelii-infected erythrocytes inhibits the capacity of DCs to maintain prolonged interactions with membrane extensions with naïve T cells (Figure 3B and Additional file 2). Only short interactions without projection of membrane extensions were observed.

To perform a quantitave analysis of this phenomenon, the formation of stable conjugates between antigen-presenting DCs and specific T cells was observed. Each cell type population was labeled with a different fluorescent dye to allow the determination of dual positives for both labels that correspond to stable conjugates 8. After a 30 min co-incubation of equal numbers of DCs presenting the ovalbumin epitope and the specific naïve anti-OVA CD4⁺ T cells formed stable conjugates where DCs 'engulf' T cells (Figure 4A, upper panels), which are different from loose contacts between these cells (Figure 4A, lower panel). See Additional file 3.

Quantification of the number of dual positives by fluorescence microscopy revealed that DC-T cell conjugates are formed when control DCs loaded with the specific peptide epitope are incubated with specific T cells and the number of conjugates is increased when DCs had been activated by addition of LPS. Pre-incubation of DCs with P. yoelii-infected erythrocytes significantly inhibited the formation of stable DC-T cell conjugates (Figure 4B). Analysis of double positives for each fluorescent label by FACs provided similar results (Figure 4C). Addition of LPS as a maturation signal increased the number of DC-T cell conjugates in the control DCs pre-incubated with uninfected erythrocytes, but did not improve DC-T cell interactions impaired by P. yoelii-infected erythrocytes (Figures 4B and 4C).

When a similar analysis was performed using CD11c⁺ DCs isolated from P. yoelii-infected animals at different times after infection, we also found a significant decrease in the numbers of DC-T cell stable conjugates compared to uninfected mice (day 0). Even if lower, detectable levels of stable conjugates formed by DCs from infected mice and specific T cells were found by microscopy and FACs (Figures 4D and 4E), suggesting that DC-T cell interactions are impaired but not completely inhibited by P. yoelii infection. These levels were minimally increased by addition of LPS to DCs (Figures 4D and 4E).

To determine the level of T cell activation in the co-cultures of DC and T cells, DCs were activated by addition of LPS and the surface expression of the early activation marker CD69 in T cells were analysed. In the co-cultures of DCs incubated with P. yoelii-infected erythrocytes or isolated from infected mice, the level of T cells expressing CD69 was greatly decreased compared to co-cultures with control DCs (Figures 4F and 4G).

To study the activation of CD4⁺ T-cells during P. yoelii infection in vivo, naïve CD4⁺ T cells were transferred from DO11.10 transgenic mice that are specific for the OVA epitope 323–339 into P. yoelii-infected (day 10 p.i.) or uninfected mice. Mice were immunized with OVA 24 h after transfer and CD4⁺ T-cell activation was measured 72 h after OVA injection. Naïve CD4⁺ T-cells were fluorescently labeled before transfer to allow identification. The number of transferred CD4⁺ T-cells in the spleen after 72 h was similar in uninfected and infected animals (Figure 5A), but after immunization with antigen, infected mice primed transferred naïve CD4⁺ T-cells with lower efficiency than uninfected mice. T-cell activation was lower in infected mice as determined by proliferation (decrease in CFSE labeling, Figure 5B), up-regulation of CD11a (Figure 5C), down-regulation of CD62L (Figure 5D) and increased intracellular IL-2 (Figure 5E). The effect of transferring naïve CD4⁺ T cells from DO11.10 transgenic mice into P. yoelii-infected mice earlier in infection (day 5), was also analysed. Under these conditions, only proliferation (decrease in CFSE labeling) and expression of CD11a were significantly decreased in infected animals. No significant changes were found in IL-2 secretion and expression of CD62L (see Additional file 4).

The capacity of DCs to provide adequate antigen presentation to T cells in the context of P. yoelii infection was analysed in detail. The results suggest that during P. yoelii infections DCs maintain the ability to process and present exogenous antigens by forming the MHC-II-epitope complex in the surface of DC, however, as shown before for P. berghei 11, impaired cross-presentation of exogenous antigens by MHC-I was found.

Efficient antigen presentation is not sufficient for the activation of naïve CD4⁺ and CD8⁺ T cells, which also requires co-stimulatory signals from DCs and specific cytokines 12. The interactions of naïve T cells with mature priming-inducing DCs are more stable than the contacts with resting tolerance-inducing DCs. It has been proposed that stable DC-T cell interactions participate in the induction of antigen-specific T cell activation through the delivery of an activatory signal by the DCs. In contrast, unstable contacts with resting DCs might induce short-term activation and proliferation signals in T cells 2, which may explain the lack of T cell activation that we observed during late malaria infection. In the absence of mature DCs, serial unstable contacts between T cells and resting DCs would result in T cell clonal deletion 2, a phenomenon that has also been observed during malaria infections, where there is specific deletion of T cells recognizing Plasmodium antigens 13.

Maturation of DCs increases the duration of DC-T cell interactions and allows the formation of a complex net of membrane extensions in which DCs entrap T cells 8. The capacity of DCs to establish these effective interactions with naïve CD4⁺ T cells was found to be inhibited during infection. As actin-mediated reorganization of DC morphology is required to form the strong, long-lasting interactions with T cells 8, it is possible that the parasite may interfere with the cytoskeleton of the DCs. In fact, numerous genes in the DC cytoskeleton are modulated during P. yoelii infection 14.

Previous studies have shown that DCs can process and present antigens associated with MHC class II during blood stage infections, suggesting that efficient activation of CD4⁺ T cells could take place during murine infection 15, however, defective activation and specific depletion of CD4⁺ T cells recognizing Plasmodium antigens is observed during murine malaria infections 1316. Suppression of OVA-specific CD4⁺ T cell proliferation has also been observed in P. chaubaudi-infected mice, but in this model inhibition was evident since early infection 17. Furthermore, these authors also found that DC and CD4⁺ T cell interactions are inhibited by P. chaubaudi infection in vitro and in vivo, while antigen presentation by MHC-II is not affected 18.

Recently, two subpopulations of CD11c⁺ DCs with differential abilities to activate antigen-specific T cells have been identified in P. chaubaudi infected mice 19, suggesting that there may be a balance of opposing forces on the host response. Since the interactions between DCs and T cells are decreased, but are still detectable, it is possible that we have observed the sum of different effects contributed by different subpopulations of DCs. It is currently believed that tolerance of DCs, induced by exposure to TLR ligands is induced during malaria infection 20. In this context, it is likely that tolerized DCs that are found increasingly during late infection would have impaired interactions with T cells leading to decreased activation.

In this work, it was observed that DCs isolated from infected mice at different times after infection are not able to establish strong interactions and prime naïve CD4⁺ T cells. Protection generated by Plasmodium infections against blood-stage infection is mediated by helper and effector functions of CD4⁺ T cells 21. However, this T cell response does not induce complete protection or long-term immunity, suggesting that T cell activation or maintenance is impaired 22. These results indicate that activation of naïve CD4⁺ T cells by DCs is impaired during late malaria blood-stage infection in mice.

DCs: Dendritic cells; OVA: ovalbumin; CFSE: 5-(and-6)-carboxyfluorescein diacetate, succinimidyl ester.

The authors declare that they have no competing interests.

CO, KAW and AR conceived the study, participated in its design and coordination and wrote the manuscript. All authors read and approved the final manuscript.