Harnessing human plasmacytoid dendritic cells as professional APCs

Most DC-based immunotherapy has been performed with monocyte-derived DCs that were generated ex vivo [11‐13]. Although clinical outcomes were observed in a fraction of patients treated with DCs [12‐17], it has been postulated that moDCs might be less effective than naturally occurring DC subsets. pDCs comprise one of these natural DC subsets circulating in the blood and are potential candidates for DC-based antitumor immunotherapy as detailed below.

The observation that human pDCs were able to induce T-cell responses in vitro confirmed results observed in mice where CpG or influenza virus matured pDCs-induced ovalbumin (OVA)-specific CD4⁺ and CD8⁺ T-cell responses [18‐22]. While CpG-matured pDCs induced CD8⁺ T-cell responses against endogenous antigens, no T-cell responses were observed against exogenous OVA antigen [23]. This discrepancy could in part be attributed to the fact that in some studies, pDCs were pulsed with the entire OVA protein, while others used OVA fragments. In other studies with exogenous antigens like influenza A or HSV, pDCs also induced CD4⁺ and CD8⁺ T-cell responses, again indicating their capacity to present antigens and stimulate T cells [24‐26]. Notwithstanding these results, pDCs were less effective in CD4⁺ priming when compared to mDCs when both are stimulated with TLR agonists, like LPS [19]. This might be explained by the fact that mDCs express high levels of TLR4 where pDCs do not. Salio et al. [23] proposed that high numbers of peptide-pulsed pDCs are equally efficient as peptide-pulsed mDCs in priming of CD8⁺ T cells. This notion is supported by the finding that murine pDCs exposed to influenza virus induced the generation of memory CD8⁺ T cells as well as effector CD8⁺ T cells upon rechallenge with the virus, leading to protective immunity [20]. Further evidence that pDCs can induce protective immunity comes from studies where mice were vaccinated with CpG-matured tumor peptide-pulsed pDCs or pulsed by a Leishmania major (L. major) lysate. Mice vaccinated with tumor peptide-loaded pDCs gained antitumor immunity and were protected upon tumor challenge, whereas nonvaccinated animals showed unhampered tumor growth [23]. This indicates that mature pDCs can effectively generate protective immunity [27]. Together, these data emphasize the potential of pDCs as type I IFN-secreting professional APCs.

A potent immune response requires both CD4⁺ T-cell and effector CD8⁺ responses. To this end, the capacity of DCs to cross-present antigen is absolutely essential. In contrast to the well-defined capacity of human CD141⁺ mDCs and mouse CD8α⁺ mDCs to cross-present antigens [28‐31], controversy exists about ability of pDCs to cross-present. Although in some studies murine pDCs fail to induce functional CD8⁺ T cells [18, 32], human pDCs appear to be competent in processing exogenous antigens to be presented in MHC class I molecules [33‐35] (Fig. 1). Exogenous antigens captured by pDCs might end up in MHC class I through distinct routes; via an endosome-to-cytosol pathway or via a cytosol-independent pathway [36‐38]. Di Pucchio et al. [35] showed that a functional proteasome is not necessary for pDCs to present captured viral antigens in MHC class I molecules, indicating that pDCs have the capacity to process and load captured antigens directly onto MHC class I molecules in endosomal compartments. This is in contrast with findings showing that exogenous antigens need to be transported from endosomal vesicles into the cytosol, become processed by the proteasome and subsequently loaded on MHC class I molecules, similar to endogenous antigens [39, 40]. Other studies showed that pDCs can cross-present antigens directly after stimulation with viral pathogens with comparable efficacy as mDC subsets. It has been reported that pDCs have larger amounts of MHC class I molecules stored in endosomes when compared to other DC subsets and can therefore rapidly transport these MHC class I molecules to the plasma membrane [35]. Taken together, studies performed with human pDCs suggest that an effective cross-presentation machinery is present. Human pDCs can be regarded as professional APCs and should therefore be considered for immunotherapy.

Two main strategies are exploited in DC-based cancer immunotherapy: (1) isolation and ex vivo stimulation of autologous DCs and (2) via direct targeting of DCs in vivo. Until now, clinical trials have primarily focused on the generation, stimulation and manipulation of DCs ex vivo. To compose a potent vaccine, the antigenic cargo needs to be efficiently and specifically delivered, avoiding inappropriate release of vaccine content. Furthermore, to explore the most potent mechanism to stimulate immature pDCs with antigens in vivo, it is important to dissect the antigen-uptake and stimulation mechanisms of pDCs. Immature DCs are able to take up various types of exogenous antigens through macropinocytosis, phagocytosis or receptor-mediated endocytosis [41]. Macropinocytosis refers to a process that mediates nonspecific capture of soluble antigens, without the involvement of receptors. While phagocytosis comprises receptor-mediated engulfment of larger antigenic particles or pathogens, receptor-mediated endocytosis is considered the most specific and efficient mechanism to capture antigens [41].

Facilitating the targeted delivery of particulate vaccines to DCs in vivo via receptor-mediated endocytosis seems an interesting opportunity, since this mechanism allows the specific and simultaneous delivery of maturation factors and antigens. But which receptor should be selected to obtain a potent immune response when considering antigen-presenting pDCs?

Several studies identified receptors for in vivo targeting of pDCs. In mice, Siglec-H and bone marrow stromal cell antigen 2 (BST2) were identified as pDC-specific receptors that opt for interesting targets [68]. Siglec-H is an endocytic receptor and member of the Siglec receptor family. Although most Siglec family members contain an ITIM sequence, Siglec-H lacks this domain [69, 70]. Siglec-H associates with DAP-12, which in spite of having this ITAM motif, result in impaired secretion of type I IFN [71, 72]. In mice, Siglec-H is involved in antigen cross-presentation, as demonstrated by the priming of antigen-specific CD8⁺ T cells upon antigen uptake via Siglec-H [69]. Recently, Loschko et al. [73] underscored the potency of targeting pDCs in vivo via BST2. In their study, they reported that targeted pDCs were efficient inducers of the expansion of both antigen-specific CD4⁺ and CD8⁺ T-cell responses. Interestingly, they observed protective antitumor responses when targeting pDCs with simultaneous administration of a TLR agonist [73]. This makes Siglec-H and BST2 potent receptors for targeting murine pDCs for the induction of CD8⁺ T-cell responses and protective tumor immunity. Although specific for murine pDCs, human pDCs do not express Siglec-H [74]. Freshly isolated as well as activated human pDCs do express BST2 [75]. However, in man expression of BST2 is not restricted to pDCs as it is also expressed by B cells. Moreover, it is upregulated on various cells upon IFNα treatment. Therefore, other receptors expressed by pDCs might be better suited [75]. Regarding the characteristics of the different surface receptors expressed by pDCs, specifically targeting antigen to the ITAM-containing FcγRIIa seems to be a promising strategy to induce immunity, since FcγRII is involved in T-cell priming, and moreover triggering FcγRIIa does not negatively affect type I IFN production. One disadvantage of FcγRIIa is that it is not uniquely expressed by pDCs, meaning that pDCs cannot be specifically targeted by triggering FcγRIIa [52]. Similarly, although DEC-205 expression is largely restricted to DCs in mice, it is broadly expressed on different immune cells in man (Table 1) [76]. In this respect, DCIR is expressed by less diverse immune cell types, making this receptor more potent for specific targeting of pDCs. Nevertheless, targeting DCIR might also activate other DC types like mDCs. Combining the stimulation of both DC types unlocks an interesting approach to potentially establish a more potent vaccine, since interaction between pDCs and mDCs has been demonstrated to increase antigen-specific immune responses. Activating pDCs along with mDCs leads also to induction of innate immune responses, likely resulting in an intensified adaptive immune response. pDCs are found to stimulate and enhance the cytokine secretion and cross-presentation of antigens leading to CD8⁺ T-cell priming by mDCs and induction of an antiviral immune response by moDCs [77‐79]. In turn, pDCs cocultured with mDCs are capable of inducing an immune response against bacteria where they fail to respond on their own [45]. In mice, an enhanced antitumor response was found when mDCs and pDCs were cocultured during pulsing with tumor antigens [80]. Moreover, pDCs were found to cross-talk indirectly with mDCs, via activation of specific lymphocyte subsets that can interact with, and might thereby stimulate, mDCs [81]. Together, these observations strongly suggest that combining pDC activation with the activation of other DC subsets might be advantageous and result in a more powerful immune response. Triggering DCIR could potentially establish such a synergetic immune response, while triggering pDC-specific receptors, like BDCA-2, initiate a more restricted induced immune response. Moreover, triggering DCIR does not completely inhibit TLR-induced type I IFN secretion by pDC as would be caused by BDCA-2 ligation. We hypothesize that the locally secreted type I IFN is important to establish an effective immune response, since type I IFNs links innate and adaptive immune responses by cross-talk with mDCs, natural killer T cells, natural killer cells and B cells. Alternatively, targeting CD40 could be considered, since it induces TLR-independent pDC maturation without negatively affecting type I IFN secretion. However, like FcγRIIa, BST2 and DEC-205, CD40 is expressed on many different cell types other then DCs and might therefore be less powerful for targeting pDCs.