Enhancement of the priming efficacy of DNA vaccines encoding dendritic cell-targeted antigens by synergistic toll-like receptor ligands

Depending on their state of maturation or activation dendritic cells (DC) can prime antigen-specific T cells with substantially different functional activities. While DC can mediate peripheral T cell tolerance in the steady-state, activation and maturation turns them into potent inducers of CD4⁺ and CD8⁺ T cell immunity (reviewed in [1]). Thus, targeting antigens to DCs is a promising strategy to improve prevention and treatment of autoimmunity and to raise the efficacy of vaccines against tumors and infectious diseases. For targeting of DC in vivo, most of the studies use recombinant protein antigens coupled or fused to an antibody specific for a DC surface marker. In addition to targeting the DCs their differentiation and activation status needs to be controlled in order to obtain the desired T cell response. A striking example of how DC activation and differentiation signals can modulate T cell responses has been obtained in experiments using an antibody to the DEC205 receptor (α-DEC) to target proteins coupled to the antibody to DCs [2]. Targeting of antigens to dendritic cells via the DEC205 receptor enhances presentation of antigen-derived peptides on MHC-I and MHC-II molecules [2, 3]. Efficient loading of MHC-II molecules with DEC205-targeted antigenic peptides might occur during the recycling of the DEC205 receptor with its ligands through late, MHC-II-rich endosomal compartments [4]. MHC-I-restricted presentation of exogenous DEC-205-targeted antigens by dendritic cells in vivo was shown to be dependent on the transporter of antigenic peptides [3]. In the absence of costimulatory signals DEC205-targeted antibody fused to antigens induced initial T cell proliferation followed by peripheral deletion and unresponsiveness of CD4⁺ and CD8⁺ T cells [2, 3]. Induction of regulatory T cells [2, 3, 5, 6] and a novel form of peripheral tolerance [7] could also be observed after immunization with different DEC-205-targeted antigens in the absence of co-stimulation. Although the precise mechanisms leading to these different forms of peripheral tolerance remain to be defined, simultaneous co-stimulation via anti-CD40 antibodies and/or toll-like receptor (TLR) ligands induces antigen-specific CD4⁺ and CD8⁺ T cell immunity rather than tolerance [8‐10].

Since gene-based vaccines are usually more potent inducers of cytotoxic T-cell responses than protein vaccines, we previously investigated whether the immunogenicity of DNA vaccines could be enhanced by targeting the encoded protein to DEC205 expressing dendritic cells [11]. In addition to the differences in immune responses induced by DNA and protein vaccines, a practical advantage of DNA vaccines encoding DC-targeted antigens is that the DNA vaccine can be produced in large scale under standardized conditions more or less independent of the vaccine antigen, while production and purification of recombinant protein vaccines needs to be adjusted for each antigen. Immunizing mice by in vivo electroporation with DNA vaccines encoding HIV Gag fused to single-chain antibodies to DEC205 allowed to reduce the DNA dose approximately 100-fold without impairment of T cell immunity and protection from viral challenge even in the absence of exogenous co-stimulation. Although this might allow overcoming the requirement for large doses, a key obstacle to the application of DNA vaccines in humans, the DC-targeted DNA vaccines did not enhance immune responses above the levels obtained by DNA immunization with high doses of non-targeted DNA vaccines.

Since the immunogenicity of viral vector vaccines can be improved by priming with DNA vaccines in various animal models including non-human primates (reviewed in [12]) and humans [13‐15], we now explored whether targeting of DNA vaccine encoded antigens to DCs can increase the priming efficacy of the DNA vaccines in DNA prime viral vector boost regimens. In addition, the requirement for additional stimuli during priming with the DEC205-targeted DNA vaccines was analysed.

In several experimental systems, DNA prime viral vector boost immunization regimens are the most potent inducers of CD8⁺ T-cell responses (reviewed in [21]). Although the DNA prime itself only induces weak immune responses, CD8⁺ T-cell responses are substantially higher after the viral boost compared to DNA prime DNA boost immunization or viral vector immunizations in the absence of DNA priming. We therefore modified the DNA priming conditions and analysed the effect of DNA priming on CD8⁺ T-cell responses after a constant viral vector boost to directly assess the potency of the combined DNA prime viral vector boost regimen.

To determine the effect of targeting the antigen expressed by the DNA vaccine to DCs on the immunogenicity of a DNA prime viral vector boost regimen, C57/Bl6 mice were primed by subcutaneous immunization with a DNA vaccine encoding a DEC-205-single chain antibody fused to ovalbumin (DEC-OVA) and boosted with an adenoviral vector expressing ovalbumin (Ad-Ova). The Ad-Ova vector was used at a suboptimal dose (data not shown) of 5 × 10⁸ particles corresponding to approximately 2 × 10⁷ transducing units in order to better detect the influence of the DNA prime. Priming with the DEC205-targeted DNA vaccine (pDEC-Ova) reduced the percentage of Ova tetramer-positive CD8⁺ T-cells after the Ad-Ova boost in comparison to a non-targeted DNA vaccine (pCon-Ova) encoding ovalbumin fused to a non-reactive single-chain antibody (Figure 1A). The functional activity of OVA-specific CD8⁺ T cells was also assessed by stimulation with the SIINFEKL peptide followed by staining for interferon-γ and the degranulation marker CD107a. Consistent with the tetramer staining, the percentage of Ova-specific CD107a- and IFNγ-positive CD8⁺ T cells was lower after priming with the DEC205-targeted DNA vaccine (Figure 1B). The CD8⁺ T cell response in mice primed with the DEC205-targeted DNA Ova vaccine was even lower than in mice that received the Ad-Ova vaccine in the absence of a DNA Ova prime, indicating that DEC-205 targeting during DNA priming suppressed CD8⁺ T cell responses induced by the Ad-Ova vector. This suppressive effect was dependent on the dose of the DEC205-targeted DNA vaccine (Figure 1C) and was also observed for the subpopulation of CD8+ T cells co-expressing IFN-γ and interleukin-2 (Figure 1D). Reduction of this subpopulation is particularly problematic, since secretion of IL-2 can promote the expansion of T cells and has been shown to enhance CD8+ T-cell memory function [22].

Since co-stimulation changes the outcome of DEC205-targeted protein immunizations from peripheral tolerance to immunity, the DEC205-targeted DNA vaccines were co-injected with expression plasmids for trimeric soluble CD40L, a dimeric form of this trimer, or a tetrameric form of the trimeric CD40L [19]. Although there might be a trend to enhancement of the priming efficacy of the DEC205-targeted DNA vaccine by co-expression of the tetrameric CD40L, this did not reach statistical significance (Figure 2A, B). We therefore also explored the effect of two toll-like receptor (TLR) ligands: CpG-C and Poly I:C. CpG-C binds to TLR 9 leading to signalling via the myeloid differentiation primary-response protein 88 (MyD88), while Poly I:C triggers TLR 3 resulting in the activation of TRIF (TIR domain-containing adaptor protein inducing interferon-β) (reviewed in [23]). In addition, Poly I:C also activates the melanoma differentiation-associated gene-5 (MDA-5) pattern recognition receptor [24], that is also expressed by many cell types and involved in anti-viral immune responses (reviewed in [25]).

Addition of CpG or Poly I:C during priming with the DEC205-targeted DNA led to a small but significant enhancement of CD8⁺ T cell responses after the Ad-Ova boost. However, a much stronger stimulation of the CD8⁺ T cell responses was observed with the combination of CpG and Poly I:C, designated CpI:C. Importantly, the DEC205-targeted DNA vaccine in the presence of CpI:C primed significantly stronger CD8⁺ T cell responses than the adjuvanted non-targeted DNA vaccine (Figure 2C, D). This could be confirmed for a wide dose range of the DEC-205-targeted DNA vaccine (Figure 2E). Priming with 2 μg of the DEC205-targeted DNA vaccine primed for the viral vector boost as efficiently as 50 μg of non-targeted DNA vaccine. Co-stimulation with different combinations of the TLR ligands and the CD40L did not improve CD8⁺ T cell responses of the DNA prime adenoviral vector boost immunization above the levels obtained by co-stimulation with CpI:C (Figure 2F).

To determine whether the improvement of CD8⁺ T cell responses as determined by in vitro assays would be of functional relevance in vivo, mice were primed with the DEC-205-targeted DNA vaccine or the non-targeted DNA vaccine in the presence of CpI:C and boosted with Ad-Ova. Subsequently, immunized mice were co-injected with differentially CFSE-labelled syngeneic spleen cells either loaded with an immunodominant Ova peptide or not loaded with exogenous peptide. One day after injection, the percentage of Ova-peptide-loaded cells among the transferred cells declined to hardly detectable levels in mice primed with the DEC205-targeted DNA vaccine in the presence of CpI:C (Figure 3A). The in vivo cytotoxic activity for Ova-peptide loaded cells was significantly lower in mice immunized with the non-targeted DNA prime Ad-Ova boost. As expected, the percentage of Ova-peptide loaded cells was inversely correlated to the Ova tetramer positive CD8⁺ T cells (Figure 3B) measured in parallel for the same mice.

The HIVp41 Gag antigen was used as an independent antigen to confirm the key findings we had obtained with the ovalbumin model antigen. Since the immunodominant Gag epitope, AMQMLKETI, is H2-Kb restricted, Balb/C mice were primed with a DNA vaccine encoding DEC205-targeted HIVp41 Gag in the presence or absence of CpI:C prior to a boost with an adenoviral vector encoding HIV GagPol (Ad-Hgp^syn). Addition of CpI:C during priming with the DEC205-targeted HIV-p41 Gag DNA enhanced Gag-specific IFN-γ producing CD8⁺ T cells approximately 3-fold (Figure 4A). T-cells producing multiple cytokines appear to be particularly indicative of protective immunity (reviewed in [26]). We therefore also analysed the influence of CpI:C on the capacity of CD8⁺ T cells to simultaneously produce IFN-γ and IL2. Adding CpI:C during priming with the DEC205-targeted HIV-p41 Gag DNA enhanced the percentage of CD8⁺ T cells double positive for IFN-γ and IL2 approximately 4-fold (Figure 4B). The CD8⁺ T cell response in mice primed with the DEC205-targeted HIVp41 DNA in the presence of CpI:C was also higher than in mice primed with adjuvanted non-targeted DNA vaccines confirming the results obtained with the ovalbumin model antigen. However, the suppression of CD8⁺ T cell responses by targeting the encoded ovalbumin to DEC205 in the absence of adjuvant could not be observed for the DEC205-targeted HIVp41 DNA vaccine.

The CD8⁺ T cell responses were further analysed by in vivo CTL assays. Priming with the DEC205-targeted HIVp41 DNA in the absence of CpI:C did not enhance in vivo cell killing after the adenoviral vector immunization, while a priming effect of the non-targeted DNA vaccine could be observed (Fig 5). Adding CpI:C to the non-targeted DNA vaccine slightly enhanced in vivo cell killing, but this difference did not reach statistical significance. The strongest in vivo CTL activity was obtained by addition of CpI:C during DEC205-targeted HIVp41 DNA priming confirming the results of the in vivo CTL assays with the Ova model antigen.

Adjuvanting DEC205-targeted DNA vaccines with TLR-ligands substantially enhanced CD8⁺ T-cell responses of the DNA prime viral vector boost regimen. The DEC205-targeted DNA prime in the presence of CpG and Poly I:C adjuvants followed by an adenoviral vector boost induced stronger CD8⁺ T cell responses than non-targeted DNA prime adenoviral vector boost immunizations. Since the latter regimen is considered to be one of the most efficient ways to induce cytotoxic T cell responses in rodents, non-human primates and humans this could be an important achievement. Although pre-existing immunity of human adenoviral vectors could reduce the immune responses induced, several strategies to overcome this limitation, such as the use of rare adenoviral serotypes (reviewed in [27]) and non-human adenoviruses [28], are presently persued.

The DEC205-targeted DNA prime adenoviral boost regimen also seems to induce substantially stronger CD8⁺ T cell responses in comparison to previous DEC205-targeted protein vaccines and DEC205-targeted DNA immunization experiments [8, 11]. In these previous experiments with DEC205-targeted protein vaccines co-stimulation with anti-CD40 antibody induced optimal CD8⁺ and CD4⁺ T-cell responses [8, 10]. We therefore also explored whether co-expression of different forms of CD40L could increase the priming efficacy of DEC205-targeted DNA vaccines. These CD40L expression plasmids enhance the immunogenicity of DNA vaccines after repeated intramuscular co-delivery with the DNA vaccine [19]. We also observed a trend towards enhancement of CD8⁺ T-cell responses by co-expression of CD40L during DNA priming via a single subcutaneous DNA immunization. However, due to a larger variation in immune responses using the subcutaneous vaccination route, the stimulatory effect by co-expression of CD40L did not reach statistical significance. Our results also indicate that DEC205-targeting of antigens encoded by DNA vaccines could be a double-edged sword. In the absence of additional stimuli, suppression of CD8⁺ T cell responses was observed for the ovalbumin model antigen, but not for HIVp41 Gag. What governs this different outcome is not understood, but could be related to the affinity of the T-cell receptor for the antigenic peptide/MHC complex or antigen-mediated modulation of the state of differentiation or activation of the DCs. Similar mechanisms were proposed to underlay the different forms of peripheral tolerance induced by DEC205-targeted peptide vaccines [6].

At a first glance, the suppressive effects we observed in our non-adjuvanted ovalbumin DNA prime adenoviral boost regimen also seems to contradict the enhancement of DNA vaccination efficacy by DEC205-targeted HIVp41 Gag DNA in the absence of co-stimulation [11]. However, in the latter study, the DEC205-targeted DNA vaccine was administered by in vivo electroporation. Since in vivo electroporation induces a strong inflammatory response [29], the delivery mode might have overcome the requirement for an additional stimulus, which we observed in the present study after subcutaneous immunization with the DNA vaccines. The inflammatory response induced by electroporation might explain why the DEC205-targeted HIVp41 DNA vaccine enhanced Gag-specific immune responses in the absence of adjuvant after in vivo electroporation, but not after subcutaneous immunization.

The type of immune response induced by DEC205-targeted protein or DNA vaccines critically depends on the activation and maturation status of the targeted DCs. This may vary within one individual depending on the site of vaccine injection and/or local infections and between individuals due to differences in the overall activation status of the immune system. It might therefore be important to override the endogenous differentiation status of the DCs at the injection site of DEC205-targeted vaccines. For DEC205-targeted DNA vaccines, the innate stimulatory properties of the injected DNA itself does not seem to be sufficient for all antigens as indicated by the suppressive effects induced by the DEC205-targeted DNA expressing ovalbumin. Although the detection of antigen-specific, IFN-γ and interleukin 2 secreting CD8+ T cells after priming with DEC205-targeted DNA vaccines in the presence of co-stimuli is suggestive for induction of memory immune responses, it will also be important to evaluate memory immune responses directly, since they are necessary for long-term protective efficacy after vaccination.

The present study demonstrates that antigen specific CD8⁺ T cell responses induced by DNA prime adenoviral vector boost regimens can be consistently enhanced by priming with DEC205-targeted DNA vaccines in the presence of TLR 3 and 9 ligands. Simultaneous stimulation of the TRIF and MyD88 pathways via TLR3 and TLR9 has been shown previously to lead to DCs with enhanced T helper type 1 polarizing capacity [30‐32]. Thus, combining adjuvants acting synergistically on DCs with DC-targeted vaccines seems to be an particularly attractive strategy for development of prophylactic or therapeutic vaccines against chronic viral infections such as HIV and HCV, in which strong cytotoxic T cell responses are considered to be of benefit.