Naive CD8+ T cells can be activated via dendritic cell 'cross-priming' of antigens obtained exogenously. Dendritic cells cannot cross-prime, however, after systemic activation in vivo, potentially contributing to immunosuppression associated with severe infections.
Dendritic cells (DCs) can potentially initiate antiviral CD8+ T cell responses through direct priming or cross-priming, which refer to the generation of peptide–major histocompatibility complex (MHC) ligands from endogenously synthesized viral proteins or from exogenous viral proteins acquired from infected cells, respectively1. During viral infection, DC are 'attentive' to environmental cues, with myriad stimuli inducing their maturation, including the binding of viral components to Toll-like receptors (TLRs)2. In vitro studies have indicated that DC maturation leads to enhanced expression of the costimulatory molecules required for activation of naive CD8+ T cells. At the same time, matured DCs show a considerably reduced capacity to phagocytose exogenous antigens and a concomitant reduction in cross-presentation of exogenous antigens. Until now, the effect of DC maturation on antiviral immunity in vivo has not been clearly established. In this issue of Nature Immunology, Wilson et al. show that activated DCs fail to cross-prime CD8+ T cell responses to subsequently encountered antigens but maintain full capacity for direct priming 3.
Although an efficient method for activating CD8+ T cells, direct presentation of antigens by DCs relies on the ability of viruses to infect them (and possibly macrophages or other 'professional' antigen-presenting cells capable of providing costimulatory signals4). However, many viruses show strict tissue tropism and do not detectably infect professional antigen-presenting cells. Furthermore, viruses that do infect professional antigen-presenting cells can express gene products that interfere with antigen presentation5. These factors may have contributed to the evolution of cross-priming, which enables CD8+ T cell priming based on the acquisition of viral antigens from infected cells of any 'stripe'.
Cross-priming of CD8+ T cells requires three specialized features of DCs: expression of costimulatory molecules; antigen acquisition through phagocytosis-macropinocytosis; and transport of antigens from endosomal compartments to the cytosol, where antigen processing can be initiated by proteasomal degradation to generate peptides that can be further processed for binding to MHC class I molecules. But what happens when DC maturation shuts down antigen acquisition? Wilson et al. tackle this question through well timed studies of DC activation and antigen delivery in mice.
Wilson et al. first formally demonstrate that in vitro activation of DCs results in inhibition of cross-priming but not direct priming. They mature purified splenic DCs with TLR ligands, expose them to soluble ovalbumin (OVA) or infect them with herpes simplex virus (HSV) and then test their ability to induce proliferation of naive OVA- or HSV-specific transgenic CD8+ T cells. Matured DCs lose the ability to cross-present soluble OVA but maintain the capacity to directly present an HSV-encoded antigen.
Extending these findings in vivo, they inject mice with a TLR ligand to systemically activate DC and then 9–12 h later with cell-associated OVA to assess cross-priming. Confirming the in vitro findings, DC activation is associated with a large reduction in the phagocytic activity of DCs and concomitantly with the activation of either OVA-specific TCR-transgenic or endogenous CD8+ T cells. Control experiments neatly establish that this effect cannot be attributed to general suppression of CD8+ T cell activation by the TLR ligands.
Armed with this tool to 'dissect' the contribution of cross-priming to CD8+ T cell activation, the authors wade into the raging controversy over the function of cross-priming versus direct priming in the activation of antiviral CD8+ T cells6. They find that systemic DC activation results in decreased CD8+ T cell division in response to infection with either HSV or influenza A virus. In an elegant demonstration of the specificity of immunosuppression, the authors immunize mice with OVA and with a systemic TLR ligand simultaneously, then later infect the mice with HSV. In these circumstances, OVA-specific CD8+ T cells are activated normally, whereas activation of HSV-specific-CD8+ T cells (either transgenic or endogenous) is inhibited. This effect on anti-HSV CD8+ T cells cannot be trivially attributed to reduced HSV replication resulting from activation of innate immunity by TLR ligands. Thus, CD8+ T cells can be activated by cross-presented antigens as long as cross-presentation occurs in the temporal window before TLR ligand–induced DC maturation (Fig. 1). These findings strongly support the claim that CD8+ T cell responses to the HSV and influenza A virus gene products is by cross-priming.
Notably, the HSV-encoded and influenza A virus–encoded antigens examined by Wilson et al. are viral glycoproteins, whose properties may favor presentation via cross-priming. It will be useful to extend this approach to evaluate the contribution of cross-priming to CD8+ T cells specific for other types of viral proteins (especially metabolically unstable proteins, which seem to be presented mainly by direct priming6) and for viruses adept at infecting DCs, such as vaccinia virus6. It is also important to integrate the lack of effect of DC maturation on direct priming with in vitro evidence for delayed endogenous antigen processing associated with the formation of dendritic cell aggresome-like induced structures7,8.
Although the results described above are impressive, they still inhabit the realm of the artificial; the immune system was never intended to deal with injection of artificial TLR ligands. Establishing a vital link to 'real-world' conditions, Wilson et al. show that mice infected with the malaria agent Plasmodium berghei demonstrate systemic DC activation, resulting in decreased DC phagocytosis and a failure to cross-present OVA to transgenic CD8+ T cells. Although these data still must be extended to HSV and additional superinfection models, they provide important evidence that in certain circumstances, such as bacterial sepsis or parasitic infection, viral superinfection (or tumors) may escape CD8+ T cell surveillance because of suppression of DC-mediated cross-priming. Moreover, the ability of plasmodia to block cross-priming demonstrates the possibility for viruses to achieve the same feat. Though this is not known to occur, given the protean capacity of viruses to modify host immunity, it would be surprising if at least some viruses had not managed the feat of DC suppression, perhaps during viremia.
The data by Wilson et al. raise the issue of why DC activation is associated with diminished cross-presentation. The authors suggest that diminished cross-presentation allows DCs to focus on presenting antigens from the pathogens that triggered their activation. But pathogen-derived peptides would still presumably represent only a small fraction of the self peptides presented by DCs, so continued presentation itself would not be expected to be detrimental to activating antipathogen CD8+ T cells. The answer may instead lie in the need to preserve pathogen peptide–MHC class I complexes, as it is possible that abrogating phagocytosis-macropinocytosis prolongs the life of pre-existing peptide–MHC class I complexes at the cell surface. Another possibility is that at some point in an immune response, continued cross-presentation of self antigens becomes harmful, perhaps because of the danger of breaking tolerance associated with increased DC expression of costimulatory signals.
Vaccinologists need to take heed of this work for two reasons. First, protein-based vaccination strategies aimed at eliciting CD8+ T cell responses through cross-priming should ensure antigen delivery to immature DCs. The optimal use of TLR ligands as vaccine adjuvants may require more than simply adding them to vaccine preparations; instead, TLR ligands should be introduced only after a suitable interval following antigen exposure. Second, as the authors suggest, live attenuated vaccines whose CD8+ T cell immunogenicity is based on cross-priming may be harmful to people with globally activated DCs, which may be relevant in areas of endemic parasitic infection. In these circumstances, however, vaccines based on direct priming should maintain immunogenicity. Finally, it is important to note that though much remains to be learned, it is apparent that advances in the basic knowledge of the cell biology and cellular immunology of antigen processing and presentation is leading inexorably to the rational design of vaccines for eliciting CD8+ T cell responses.
Heath, W.R. et al. Immunol. Rev. 199, 9–26 (2004).
Guermonprez, P., Valladeau, J., Zitvogel, L., Thery, C. & Amigorena, S. Annu. Rev. Immunol. 20, 621–667 (2002).
Wilson, N.S. et al. Nat. Immunol. 7, 165–172 (2006).
Pozzi, L.A., Maciaszek, J.W. & Rock, K.L. J. Immunol. 175, 2071–2081 (2005).
Yewdell, J.W. & Hill, A.B. Nat. Immunol. 3, 1019–1025 (2002).
Yewdell, J.W. & Haeryfar, S.M. Annu. Rev. Immunol. 23, 651–682 (2005).
Pierre, P. Immunol. Rev. 207, 184–190 (2005).
Herter, S. et al. J. Immunol. 175, 891–898 (2005).
About this article
Cite this article
Hickman-Miller, H., Yewdell, J. Youth has its privileges: maturation inhibits DC cross-priming. Nat Immunol 7, 125–126 (2006). https://doi.org/10.1038/ni0206-125
Cytokine network analysis of immune responses before and after autologous dendritic cell and tumor cell vaccine immunotherapies in a randomized trial
Journal of Translational Medicine (2020)
Delivering all in one: Antigen-nanocapsule loaded with dual adjuvant yields superadditive effects by DC-directed T cell stimulation
Journal of Controlled Release (2018)
Inhibitor of apoptosis protein (IAP) antagonists demonstrate divergent immunomodulatory properties in human immune subsets with implications for combination therapy
Cancer Immunology, Immunotherapy (2013)
Critical Role of TLR7 Signaling in the Priming of Cross-Protective Cytotoxic T Lymphocyte Responses by a Whole Inactivated Influenza Virus Vaccine
PLoS ONE (2013)
The Journal of Immunology (2010)