Targeting DNA Vaccines to Myeloid Cells Using a Small Peptide

Abstract

Targeting DNA vaccines to dendritic cells (DCs) greatly enhances immunity. Although several approaches have been used to target protein antigens to DCs, currently there is no method that targets DNA vaccines directly to DCs. Here, we show that a small peptide derived from the rabies virus glycoprotein, fused to protamine residues (RVG-P) can target DNA to myeloid cells, including DCs, that results in enhanced humoral and T-cell responses. DCs targeted with a DNA vaccine encoding the immunodominant vaccinia B8R gene via RVG-P were able to restimulate vaccinia-specific memory T cells in vitro. Importantly, a single i.v. injection of B8R gene bound to RVG-P was able prime a vaccinia-specific T-cell response that was able to rapidly clear a subsequent vaccinia challenge in mice. Moreover, delivery of DNA in DCs was enough to induce DC maturation and efficient antigen presentation without the need for adjuvants. Finally, immunization of mice with a DNA-vaccine encoding West Nile virus (WNV) prM and E proteins via RVG-P elicited high titers of WN neutralizing antibodies that protected mice from lethal WNV challenge. Thus, RVG-P provides a reagent to target DNA vaccines to myeloid cells and elicit robust T-cell and humoral immune responses.

Introduction

DNA vaccines provide an attractive method for vaccination against intracellular pathogens and cancer. DNA vaccines are ideally suited for global mass vaccination because in contrast to other subunit vaccines based on vector systems such as modified vaccinia Ankara (MVA) and recombinant adenoviruses, they are easy to manufacture, cheap, stable and without problems of pre-existing immunity against the vector (reviewed in (1). However, although some DNA vaccines have been licensed for use in animals (e.g. WNV DNA vaccine for horses (2) and melanoma DNA vaccine for dogs (3)), and many vaccines are in human clinical trial (e.g. Vaccines for WNV, HIV, malaria, HPV and many cancers (reviewed in (4, 5)), no DNA vaccine has yet been approved for use in humans. This is mainly because of their generally low immunogenicity profile in humans. Thus, If immunogenicity can be enhanced it would greatly aid the utility of DNA vaccines in humans.

Dendritic cells (DCs) are critical antigen presenting cells in the initiation of an adaptive immune response. Targeting protein antigens to DCs can enhance an immune response by over 100-fold. Thus, in recent years there has been a tremendous interest in targeting antigens to DCs to enhance immunogenicity (6, 7). This approach is rapidly being applied to several diseases including malaria, tuberculosis, Leishmania, cancer, and AIDS (8). In fact, the first DC-targeted vaccine in clinical trial, DEC-205/HIV gag vaccine has been recently reported to effectively induce a robust antibody, as well as T-cell response in human volunteers (8). The commonly used approach to target antigens to DCs is to use antigenic proteins conjugated to monoclonal antibodies to DC antigen-uptake receptors such as DEC205, Dectin-1, Langerin or Clec9A (9-13). This approach results in the generation of strong antibody, as well as T-cell responses. However, strong adjuvants such as poly IC and/or TLR agonists are needed to elicit immunity and in their absence, this approach induces tolerance rather than immunity (12, 14, 15). Other approaches to target DCs have used DNA or viral (adenovirus, Newcastle disease virus, etc) vectors engineered to express single chain antibody to DC receptors fused to the antigen of interest (16-19). However, this approach relies on infection of other cell types followed by secretion of the DC targeting Ab/antigen conjugate and thus, does not allow selective antigen expression in DCs. Moreover this approach cannot be used to deliver DNA vaccines directly to DCs.

Several reagents including d-glucan-encapsulated siRNA particles, ionizable cationic lipids called DLinKC2DMA and a lipidoids nanoparticle called C12-200 have been developed to deliver small oligonucleotides (principally siRNA) to murine macrophages and dendritic cells (20-24). However, these reagents are unlikely to be able to deliver large molecules such as DNA vaccines to DCs. Thus, currently there is no method to target DNA vaccines directly to DCs.

In previous studies, we have shown that a small peptide derived from rabies virus glycoprotein (RVG) can specifically recognize both murine and human DCs and macrophages by binding to acetylcholine receptor expressed by these cells (25, 26). We have also shown that a chimeric RVG protein containing oligoarginine (RVG-9R) residues can bind siRNA by charge interaction and deliver it to DCs in vitro, as well as in vivo, resulting in specific gene silencing (25, 26). However, RVG-9R is unable to bind DNA. Since protamine has been extensively used to condense DNA (27, 28), in this study we tested RVG-protamine fusion peptide (RVG-P) for its ability to target DNA vaccines to myeloid cells. We find that targeting DNA vaccines to DCs via RVG-P resulted in enhanced immune responses against Vaccinia and WNV infection.

Results and Discussion

RVG-protamine can deliver DNA to DCs

Since protamine has been traditionally used to condense DNA, we reasoned that adding it to our RVG peptide might eanble DNA delivery to DCs. Thus, we synthesized RVG peptide with minimal protamine residues at the carboxy terminus. RVG-P was tested for DNA binding in a gel shift assay as dsecribed earlier (29). RVG-P bound DNA in a dose dependent manner as evidenced by decreased mobility and quenching of flourescense with complete binding at 1:10 wt/wt ratio of DNA:peptide (Fig. 1a). Binding was also confirmed by testing for restoration of mobility after treatment of DNA/RVG-P complex with proteinase K. We used 1:10 weight/weight ratio of DNA:peptide in further studies.

Fig. 1. RVG-P binds and delivers DNA to MDDCs resulting in DC maturation and effective antigen presentation.

(A) Vaccinia B8R gene-encoding plasmid, mixed with RVG-P at indicated concentrations was tested for DNA binding by a gel shift assay on 1% agarose gel. The quenching and mobility shift seen in lanes 2, 3 and 4 could be reversed by proteinase K digestion of the complex (lanes 6, 7, 8). Lane 1 shows 1kb DNA ladder. (B) RAW cells or (C) MDDCs were incubated with RVG-P/B8R DNA or the control RV-MAT/B8R DNA for 4 h, and cultured for 48 h before testing for myc tag expression by flow cytometry. Overlay histograms of untreated cells with isotype control antibody and RV-MAT-P and RVG-P treated cells with myc tag antibody is shown as indicated. Gating strategy and a representative dot plot analysis is shown in Supporting Information Figure 1. (D) MDDCs were untreated or treated with B8R DNA bound to RVG-P or RV-MAT-P or activated with LPS (10 μg/ml) and tested for indicated DC maturation marker expression 24 h later, gating on live cells as shown in Supporting Information Figure 1. (E-G) The concentrations of (E) TNF-α, (F) IL-6, and (G) IFN-γ in culture supernatants from (D) were measured by cytometric bead array (CBA). (H) MDDCs were treated with mock, B8R DNA/RVG-P and B8R DNA/RV-MAT-P or pulsed with cognate peptide, and were then used to stimulate vaccinia-specific memory T cells and antigen-specific T-cell expansion followed over time by B8R pentamer staining, gating on CD8+ T cells. (A-H) Data are shown as mean ± SEM (n=3 samples) and are from one experiment representative of at least two independent experiments all performed in triplicates. * p < 0.05;Anova.

Next we tested if we can target a vaccine DNA to DCs using RVG-P. During vaccinia infection in C57BL/6 mice, the immunodominat CD8⁺ T-cell response is focussed on a single peptide epitope called the B8R epitope (30). In fact, B8R peptide-MHC pentamer⁺ cells constitute ~10% of CD8⁺ T cells at the peak of response on day 7 postinfection (31). Vaccinia virus B8R gene encodes a 278 aa long secreted protein (in which the epitope is located) and we chose to use this gene as a DNA vaccine to test DC targeting with RVG-P. We synthesized this gene with a myc tag (to identify protein expression) and cloned it in front of a CMV promoter in a eukaryotic expression vector. To test gene delivery, RAW cells and primary bone marrow-derived dendritic cells (MDDC) were treated with DNA bound to RVG-P or a control peptide, RV-MAT-P and gene expression tested by staining the cells with anti-myc tag antibody 2 days later. As shown in Figure 1b-c, myc tag was expressed in RVG-P/DNA, but not in the control peptide, RV-MAT-P/DNA treated cells (both RAW and MDDCs).

Next we tested the ability of DCs treated with RVG-P/B8R DNA to induce DC maturation and to present antigen to memory T cells. MDDCs were treated with DNA bound to RVG-P and 1 day later examined for DC maturation marker expression. RVG-P/DNA-treated cells upregulated MHC Class II, CD40, CD80 and CD86 to similar extent as LPS-stimulated DCs (Fig. 1d). Moreover, the DCs also secreted several pro-inflammatory cytokines (Fig. 1e-g). Thus, introduction of DNA into DCs is enough to cause DC-maturation. To test antigen presentation, the vaccine DNA/RVG-P treated DCs were used to stimulate splenocytes from mice infected with vaccinia virus 2 months earlier. Cognate peptide-pulsed DCs were used as positive control. Memory T cells as determined by MHC-pentamer staining constituted ~1% of CD8⁺ T cells before stimulation. Pentamer⁺ T cells steadily increased following stimulation with DCs treated with DNA/RVG-P or pulsed with cognate peptide, but not with DCs treated with RV-MAT-P bound DNA (Fig. 1h), showing that RVG-P efficiently targets DNA to DCs to stimulate antigen-specific T cells.

Immunization with RVG-P-bound vaccinia DNA vaccine primes a vaccinia-specific immune response

Next we tested if RVG-protamine can target DNA to myeloid cells in vivo and whether it primes a vaccinia-specific immune response. Mice were immunized with 50 μg of B8R DNA bound to RVG-P or RV-MAT-P, and 5 days later infected with vaccinia virus. Mice were followed for the generation of an immune response and clearance of virus on different days after infection. As ascertained by B8R pentamer staining and IFN-γ production in response to peptide stimulation, an accelerated and heightened T-cell response was seen for up to 7 days after infection in mice vaccinated with DNA bound to RVG-P compared to the endogenous response seen in mice immunized with DNA bound to RV-MAT-P (Fig. 2a-b). Pentamer⁺ cells were detectable (constituting ~5% of CD8⁺ T cells) as early as day 3 after infection in immunized mice and progressively increased with time. Correspondingly there was a four-fold reduction in peak viral titers on day 5 (Fig. 2c). Co-administration of the adjuvant, poly IC, did not further enhance the response, confirming that expression of DNA within DC is enough to activate them (not shown). Vaccinia is a self-limiting disease and immunocompetent mice clear the infection by day 7-10 and around that time, the acute phase T-cell response is decreased. Consistent with this, there were no differences between the groups by day 9. These results show that immunization with RVG-P/DNA could prime a vaccinia-specific T-cell response.

Fig. 2. Immunization with B8R DNA bound to RVG-P primes vaccinia-specific T-cell response.

Mice were injected with PBS (Mock), B8R DNA bound to RVG-P (vaccinated mice) or RV-MAT-P (control mice), challenged with vaccinia virus intraperitoneally 5 days later, and the peritoneal exudate cells followed for generation of, (A) antigen-specific T cells by serial B8R pentamer staining on indicated days after infection, (B) IFN-γ production following ex vivo peptide stimulation, determined by intracellular staining and flow cytometry (gating strategy and a representative dot plot analysis is shown in Supporting Information Figure 1), and (C) viral titers in the ovaries determined by plaque assay. (D_F) Mice were immunized as in (A) and challenged with vaccinia virus after 6 weeks. Subsequently the peritoneal exudate cells were tested for (D) B8R pentamer⁺ cells, (E) IFN-γ production and (F) virus titers in the ovaries. (A-F) Data are shown as mean ± SD (n=5 mice/group/each time point) and are pooled from 2 independent experiments. * p < 0.05; Anova.

We also tested if DNA immunization can induce long-term memory T cells. Mice were vaccinated with B8R DNA bound to RVG-P or control RV-MAT-P and rested for 6 weeks before challenging with vaccinia virus. Immune response and virus clearance was tested as above. Again, a heightened antigen-specific T-cell response (Fig. 2d-e) accompanied with lessened viral burden (Fig. 2f) was seen in RVG-P/DNA immunized mice compared to control mice. Collectively, our results show that DC-targeted DNA vaccination results in an enhanced vaccinia-specific T-cell response.

Myeloid cell-targeting of WNV DNA elicits neutralizing antibodies and protects from viral challenge

A WNV DNA vaccine encoding prM and E protein has been approved for use in horses (2). It has also been used in human clinical trial and was found to elicit modest levels of neutralizing antibodies in healthy volunteers (32). We therefore tested if DC targeting of this vaccine could enhance immunogenicity. We obtained the same plasmid that was used in clinical trial from VRC, National Institutes of Health. Mice were immunized twice, two weeks apart by injection of DNA bound to RVG-MAT-P (control) or RVG-P (75 μg DNA bound to 800 μg of peptide). Mice were bled 2, 4 and 6 weeks later and on week 6 after boost injection, challenged i.p. with 4X10⁵ PFU of WNV. DNA/RVG-P immunized mice developed robust levels of WNV antibodies, including neutralizing antibodies (Fig. 3a-b) and were completely protected (Fig. 3c), whereas DNA/RV-MAT-P-immunized mice showed no antibodies and 80% of animals succumbed by day 10 after challenge. In summary, we have shown that RVG-P provides a reagent to target DNA vaccines to DCs in vitro and in vivo and that such targeted delivery can enhance both humoral (WNV) and T cell (vaccinia) immune responses. In contrast to the other approaches used for DC targeting discussed in the introduction, our method of delivering DNA directly to DCs using RVG-P offers several advantages. Compared to the complexity of generating the available DC-targeting reagents, RVG-P is a small peptide that can be readily synthesized. Moreover, it allows direct antigen expression in DCs thereby enabling classical (not cross presentation) antigen presentation via both MHC Class I and Class II pathways. It also obviates the need for strong and potentially harmful adjuvants such as poly IC or TLR agonists and avoids vector-directed immune response that could render the targeting ineffective. In addition, RVG is nontoxic, non immunogenic and peptide binding protects nucleic acids from serum nuclease cleavage, making in vivo use possible (29). Since AchR is highly conserved in mouse and humans, another key advantage of RVG-P is that the same reagent can be used to target mouse and human DC and thus, allows preclinical mouse studies to be immediately translated to human cells. Moreover, since DNA binding to RVG-protamine occurs by charge interaction (we just need to mix DNA with RVG-P for 5-10 min before injecting), virtually any DNA vaccine or even a combination of DNA vaccines can be rapidly targeted to DCs without a need for generating specialized targeting constructs for individual vaccines. Although RVG-9R can deliver small amounts of siRNA to brain cells (29), RVG-P is not expected to deliver DNA molecules across the blood-brain barrier because of large size differences (~20 kD for RVG-9R/siRNA complex versus >2000 kD for RVG-P/DNA complex), making neuronal expression of the vaccine gene with potential toxicity unlikely. Similarly since RVG essentially binds to the α7 subunit of nicotinic acetylcholine receptor, delivery at muscarinic AchR expressing neuromuscular junction is unlikely.

Fig. 3. Immunization with WNV DNA vaccine bound to RVG-P protects against WNV infection.

Mice were immunized with a plasmid encoding WNV PrM/E proteins bound to RVG-P (vaccinated mice) or RV-MAT-P (control mice) on days 0 (prime) and 14 (boost). (A) Sera collected 2, 4 and 6 weeks after boost were tested at 1:50 dilution for E protein binding by ELISA. (B) Sera collected 6 weeks after boosting were tested for WNV neutralization by plaque reduction neutralization test (PRNT). PI= pre-immune sera. (C) Six weeks after booster immunization, the mice were challenged with WNV and monitored for survival. Kaplan-Meier survival curve is shown. (A-C) Data are shown as mean ± SEM (n=16 mice/group) and are pooled from two independent experiments. . * p < 0.05; Anova for (A) and (B), Log-rank test for (C).

Materials and methods

Mice

C57BL/6 mice were purchased from Jackson Laboratory. All mice were maintained in specific pathogen free conditions and used when they were 6–10 weeks of age. All animal experiments had been approved by the Institutional Animal Care and Use Committee (IACUC) at TTUHSC.

Peptides

RVG-protamine (NH2-YTIWMPENPRPGTPCDIFTNSRGKRASNGGGGRSQSRSRYYR QRQRSRRRRRRS-COOH) and the control RV-MAT-protamine (NH2-MNLLRKIVKNRRDEDTQKSS PASAPLDDGGGGRSQSRSRYYRQRQRSRRRRRRS-COOH) peptides were synthesized and purified by high performance liquid chromatography (HPLC) at Peptide 2.0 (Chantilly, VA). RVG peptide (derived from the Rabies virus glycoprotein) binds to acetylcholine receptor whereas RV-MAT peptide (derived from the Rabies virus matrix protein) does not, and thus serves as negative control (29).

MDDC and antigen presentation

Mouse bone marrow cells were cultured in RPMI 1640 medium (Gibco BRL) supplemented with 10% heat-inactivated fetal bovine serum (FBS) in the presence of GM-CSF (R&D Systems; 1,000 U/ml) and IL-4 (500 U/ml) for 6 days at 37°C in a CO2 incubator. To deliver DNA to DCs, B8R DNA was mixed with RVG-P or the control RV-MAT-P peptides and incubated at room temperature for 10 min before incubating with DCs for 4 h. The medium was then replaced and DCs cultured for 48 h before testing for myc expression by intracellular staining. The cells were also tested for DC maturation markers by flow cytometry. Cytokines secreted in the supernatant were tested using the BD Cytometric Bead Array (CBA) as per the manufacturer’s instructions. In some experiments, the RVG-P/DNA treated cells were used to stimulate splenocytes isolated from mice infected with vaccinia virus 6 weeks earlier, as previously described (31).

Gel shift assay

Gel shift assay was done as described earlier (29). Briefly 1 μg of DNA (~4kb) was mixed with different concentrations of RVG-P, incubated at room temperature for 10 min and analyzed for binding by 1% agarose gel electrophoresis. Mobility shift and quenching of fluorescence indicated binding. Binding was also confirmed by testing for restoration of mobility after treatment of DNA/RVG-P complex with proteinase K.

DNA vaccines and immunization

pCMV-Myc-N plasmid was purchased from Clontech Laboratories (Mountain View, CA); vaccinia virus B8R gene (GenBank accession: JN654982) synthesized at IDT was cloned into pCMV-Myc-N. West Nile DNA vaccine was a kind gift of VRC, NIH. For immunization, DNA vaccines (extracted using endotoxin free plasmid extraction kit from Qiagen) were mixed with RVG-P or RV-MAT-P peptides (75 μg DNA with 750 μg peptide) and incubated at room temperature for 10 min before i.v. injection.

Viral infection

The WR strain vaccinia virus and B956 strain of WNV were obtained from ATCC (Manassas, Virginia, United States). Mice were infected i.p. with 10⁶ PFU/mouse of vaccinia virus or West Nile virus (5LD₅₀) in 100 μl PBS.

ELISA for WNV antibodies

WNV E protein (Aviva Systems Biology) was coated on to 96-well plates overnight at 4°C. The wells were then washed, blocked and treated with 1:50 dilution of sera followed by anti-mouse IgG-Alk Phosphatase (Bio-Rad Laboratories). Color was developed with the substrate, p-nitrophenyl phosphate (Sigma) and the plates read at 405 nm using FLUOstar Omega (BMG Labtech).

Plaque and PRNT Assay

To determine vaccinia viral load, serial dilutions of homogenates of ovaries harvested from infected mice were inoculated on CV1 cells and after 2 days, stained with cristal/formalin and the plaques counted.

To test for WNV neutralizing antibodies, serial dilutions of heat-inactivated sera were incubated with 70 PFU of WNV for 1 h at 37°C and were added to BHK21 cells. After 2 h incubation, media was replaced with 1% agarose DMEM. After 72 h , the plates were stained with cristal/formalin and the plaques counted.

Flow Cytometry

Cell surface antigen staining was performed by 20 min incubation on ice with pertinent antibodies. The following monoclonal antibodies were used: anti-mouse CD40, CD80, CD86, MHC II, CD4, CD8, IFN-γ and isotype control Abs (all purchased from BD Pharmingen). B8R-Pentamer was phased from ProImmune (Sarasota, FL). Intracellular staining was performed with Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Bioscience) according to the manufacturer’s protocol. Data were acquired by BD FACS Canto II and analyzed on BD FACS Diva software. To test for IFN-γ production, peritoneal exudate lymphocytes were stimulated with B8R peptide as described earlier (31).

Statistical analysis

Differences in values between experimental groups were examined for significance with Graph Pad Prism software using one-way analysis of variance (Anova). We considered probability (ρ) values < 0.05 as significant. Values are presented as means ± SEM.

Abbreviations

RVG

Rabies virus glycoprotein

WNV

West Nile virus

DC

dendritic cell