Several previous reports have suggested a generalized suppression of immune responses during infection with Plasmodium [2–10, 12–21, 24–34, 59]. One possible mechanism could be impairment of the function of DCs, a cell type that is essential in the generation of the primary immune responses . But the significance of this observation for in vivo studies, the implications for downstream immunological function, and what (if any) parasite component mediated this effect remained unclear. Here, we have shown that DCs are modulated by the malaria parasite and are suppressed by infection with P. chabaudi through the malarial pigment HZ. Importantly, Plasmodium infection also causes a significant defect in the induction of immune responses in vivo: expansion and migration of CD4+ T cells is greatly reduced, resulting in a consequent reduction in the interaction between these T cells and B cells and in the help they can provide to the B cells. Despite severely impaired T-cell migration and effector function, early stimulation of antigen-specific CD4+ T cells is not affected by malaria infection, as T cells stimulated in vitro and in vivo upregulate CD69, suggesting that, despite suppression of DC function, there is sufficient antigen presentation to induce initial T-cell activation.
Although the observed defect in CD4+ T-cell function seems to be directly related to inhibition of DCs, it has also been suggested that optimal T-cell expansion and differentiation requires the interaction of T and B cells . Thus it may be that the failure of DCs to activate properly and subsequently induce T-cell migration into B-cell follicles also results in the reduced expansion of CD4+ T cells observed here. Nevertheless, the failure of optimal CD4+ T-cell expansion and migration during infection clearly results in the abrogation of B-cell expansion and a subsequent absence of antibody production at specific time-points during infection. Thus, protective, systemic immunity against non-parasite antigens, and presumably against parasite antigens, fails to develop effectively.
Although alterations in the architecture of the spleen during infection with P. chabaudi have been described , these changes were less evident in our model, with intact B220+ B-cell follicles still clearly discernible by day 17 of infection (see Figure 9b). Interestingly, infected mice immunized on day 4 of infection (when lymphoid structure is already disrupted ) were able to generate antibody responses, suggesting that despite the breakdown in architecture, immune priming is still normal in these mice.
Our important finding that activated T cells fail to migrate into follicles may in part be due to splenic alterations not visualized in our immunohistochemical staining for B220, or to alteration of chemokine gradients important for T-cell migration in infected mice. To address these issues, we transferred T cells that had been primed in vitro in the presence of malaria parasites into uninfected recipient mice with normal lymphoid architecture. In addition, we immunized naive, uninfected mice with OVA-loaded DCs which had been cultured in vitro with infected erythrocytes. In all of these situations we could separate the effect of parasites upon DCs or T cells from the described disruption of lymphoid architecture (see Figure 9). In each case, T cells primed in the presence of malaria parasites or T cells primed by DCs exposed to infection failed to differentiate fully following transfer into uninfected recipients. Thus, alteration of splenic architecture alone cannot account for our observations, suggesting that the failure in T-cell differentiation is due to DC modulation.
Another possible explanation for the observed defect in the induction of immunity could be through competition for access to antigen between adoptively transferred OVA-specific T cells and endogenous malarial antigen-specific T cells. Thus, a large dose of blood-borne parasite antigens could impede the induction of other ('bystander') immune responses. But studies examining bystander responses in other parasitic diseases (for example, [62–65]), as well as the potent stimulatory capacity of mycobacteria-containing complete Freunds' adjuvant , suggest that such inhibition of immunity does not occur, despite potentially large amounts of competing antigens. In our experiments, naive OVA-specific CD4+ T cells activated in vitro and in vivo in the presence of infected erythrocytes upregulated CD69 to the same extent as T cells stimulated in uninfected controls, suggesting that sufficient access to antigen was available to initiate T-cell signaling cascades (see Figures 7 and 8). Furthermore, transfer of T cells activated in the presence of infected erythrocytes, or transfer of purified DCs from infected mice into uninfected recipients, transferred the immunosuppressive phenotype, suggesting that the effect cannot be ascribed to out-competition of the OVA-specific T cells by malaria-specific cells (see Figures 8 and 9).
In search of a mechanistic explanation for these observations, we initially focused on analyzing the effect that parasite proteins expressed on the erythrocyte membrane might have on DC function. It is known that the development of the parasites within erythrocytes is coupled with changes in the host cells, including the host-cell plasma membranes , and it is well established that parasites express 'neo-proteins' on the host-cell surface, some of which are reported to induce protective immunity [68–70]. Erythrocytes infected with the rodent-specific strain P. chabaudi adhere to specific cell types by interacting with molecules such as CD36 ; this is known as cytoadherence. Importantly, the interaction of P. falciparum-infected RBCs with CD36 has been shown to mediate suppression of DC function . Examination of the P. chabaudi genome has not, however, revealed any homologs of the P. falciparum protein PfEMP1 , which is important in sequestration . Rather, a separate multigene family expressing surface antigens was identified in P. chabaudi , which may have a role in cytoadherence .
Another important difference between P. chabaudi (the rodent-specific strain used here) and the human-specific strain P. falciparum is the presence of surface complexes, known as knobs, in P. falciparum-infected erythrocytes, which strengthen interactions between pRBCs and receptors expressed on other cells . P. chabaudi-infected pRBCs have no evident knobs . Thus, although P. falciparum and P. chabaudi show specific differences in their mechanisms of cytoadherence, pRBC-mediated DC suppression might occur through interactions with membrane molecules, such as CD36. We therefore exposed DCs to erythrocyte ghosts from infected or uninfected erythrocytes, before the LPS challenge. Ghosts isolated from pRBCs did not alter the ability of DCs to respond to LPS treatment in vitro, suggesting that parasite antigens expressed on the erythrocyte plasma membranes do not induce the suppression previously described following contact with DCs in vitro. In support of this finding is the observation that immunization of mice with pRBC ghosts can induce protection from parasite challenge , suggesting that protein structure is maintained on erythrocyte ghosts and that ghosts, unlike intact parasites, are not inherently immunosuppressive.
As fixed pRBCs suppressed the LPS-induced maturation of DCs whereas pRBC membranes did not, HZ seemed to be a good candidate to investigate when looking at the mechanism of parasite-induced modulation of DC function. Previous reports have suggested that HZ impairs the differentiation and functional capacity of human monocytes and murine macrophages through the production of IL-10 and/or the induction of peroxisome proliferator-activated receptor-γ (PPAR-γ) [78–82]. The extrapolation of these results to DCs remains somewhat controversial, however, with other workers suggesting a proinflammatory role for HZ, possibly via the Toll-like receptor-9 [83–86]. Our results suggest that intracellular parasite components, including HZ, do indeed suppress DC maturation and function in vitro. The exact mechanism involved in this suppression remains undefined, although it is interesting that phagocytosis of HZ has been found to increase degradation of protein kinase C . Thus, degradation of key intracellular signaling molecules may be one mechanism by which Plasmodium parasites suppress DC function. Together, these results suggest that in contrast to P. falciparum, intracellular HZ rather than P. chabaudi-derived membrane-expressed proteins is responsible for the suppression of APC function.
The results presented here clearly demonstrate that DC function is dynamically modulated in vitro and in vivo by asexual blood-stage malaria parasites. These findings support previous studies with P. falciparum-infected erythrocytes and human monocyte-derived DCs  as well as studies in vitro and in vivo of Plasmodium yoelii with murine DCs . Other studies suggest, however, that P. chabaudi schizonts activate DCs in vitro . Similarly, in previous studies [45, 89], DCs isolated from P. yoelii-infected mice during peak parasitemia were found to be activated and to efficiently process and present antigen to naive T cells. In the present study, DCs exposed to trophozoite-infected erythrocytes show impaired maturation in response to stimulation, indicating that it is not only different parasites (P. yoelii versus P. chabaudi) but also different stages (trophozoites versus schizonts) that have different effects on DC function. Interestingly, the observed differences in the ability of pRBCs to stimulate DC maturation may arise through contamination of parasite material with mycoplasmas, which are known to contain potent Toll-like receptor (TLR) ligands that efficiently activate DCs [90–92]).
It has recently been suggested that, during malaria infection in vivo, DCs are activated during early infection and then show TLR tolerance later in infection, becoming unresponsive to LPS stimulation . We believe this not to be the case with P. chabaudi, however, because we see no evidence of direct maturation of the DCs by the parasite (see Figure 2). In addition, activated DCs show an increased ability to stimulate T-cell proliferation and cytokine production , neither of which were observed in the T-cell assays in the current study. Rather, we suggest that the transient increased expression of activation markers on DCs ex vivo reflects the high concentrations of pro-inflammatory cytokines caused by the early stage of infection . In support of this, a recent report described that DCs activated through inflammatory cytokines without pathogenic stimulation upregulated markers of activation but were unable to drive CD4+ T-cell differentiation .
The ability of DCs to interact with CD40L on T cells in vivo has also been used to explain the differences between in vivo and in vitro studies . Interestingly, in our study we could not rescue DC maturation when pRBC-treated DCs were stimulated with CD40L-transfected fibroblasts, also suggesting that TLR tolerance (the refractory state of DCs to a second stimulation with a TLR ligand) is not involved in the failure of DCs to respond to stimulation. This suggests that the effects that P. chabaudi-infected erythrocytes exert on DC function in vitro might be more profound than those induced by P. yoelii infection. Whether these changes to the CD11c+ population as a whole reflect changes in individual subsets of DCs awaits further investigation. Importantly, we have shown that the immunosuppression seen is due to this inhibition of DC function rather than to suppression of T cells or breakdown of splenic architecture, as transfer of T cells activated in the context of parasites in vitro or of DCs from infected mice was sufficient to prevent subsequent T-cell differentiation in uninfected recipients (see Figure 9).
The results presented here demonstrate, for the first time, that suppression of immunity associated with P. chabaudi affects multiple populations of cells essential for development of immunity. DC function is impaired during parasite infection, as a result of ingestion of HZ, and although CD4+ T cells specific for a non-parasite antigen become activated following immunization, they fail to expand clonally as efficiently as in uninfected controls. Crucially, these T cells subsequently show a defect in their ability to migrate into B-cell areas and, consequently, fail to provide effective help for B-cell expansion and antibody production. These results demonstrate an overall defect in priming of heterologous immune responses during Plasmodium infection and provide an explanation for increased secondary infections and the reduced efficacy of vaccines in areas where malaria is endemic.