Technologies for enhanced efficacy of DNA vaccines

Mode of action of DNA vaccines

Following intramuscular injection, the plasmid DNA is taken up by muscle cells and monocytes that then express the antigen. Antigen peptides are then loaded into MHC-I molecules on muscle cells, or both MHC-I and MHC-II molecules on APCs [3]. APCs may also capture protein secreted by transfected cells or contained in transfected apoptotic muscle cells [3,4]. Antigen-loaded APCs traffic to the draining lymph node where they activate B cells and CD4⁺ and CD8⁺ T cells (Figure 1).

Since the amount of antigen expressed is small, in the order of nanograms to picograms, immunogenicity is dependent upon help from CpG motifs in the plasmid backbone that play a role as ‘built-in’ adjuvants, promoting the induction of T-cell responses [5,6]. The core CpG motif consists of an unmethylated CpG flanked by two 5´ purines and two 3´ pyrimidines. CpG sites are relatively rare (~1%) in vertebrate in comparison to bacterial genomes or viral DNA. Toll-like receptor (TLR)9 is expressed in the endosome of APCs including DCs and B cells. Upon recognition of CpG, TLR9 triggers a MyD88-dependent signaling cascade resulting in a proinflammatory response. HMGB1 binds nucleic acids and is required for CpG-activated innate immune responses [7]. CpG motifs thereby increase the immunogenicity of DNA vaccines [8], although recently TLR9 has been shown not to be the sole mediator of adjuvant effects, with TLR9 knockout mice still responding to DNA vaccines. Plasmid DNA has a double-stranded structure that interacts with cytoplasmic DNA sensors such as TBK-1 and STING, thereby activating TLR-independent pathways and inducing type 1 interferon [9–12]. Unlike the activation of TLR9 by CpG motifs, the methylation of dsDNA does not alter its activity. Rather, poly(dA-dT)• poly(dT-dA) induced higher levels of IFN-1 compared with poly(dG-dC)•poly(dC-dG), suggesting that the right-handed helical structure of B-form DNA is essential for cellular activation of IFN-1 production [13]. IFN-1 induced via the STING/TBK1 pathway was found to be crucial for both direct and indirect antigen presentation via DCs and muscle cells, respectively [14]. More recently, the helicase DDX41 was identified as a new intracellular DNA sensor in myeloid DC. DDX41 bound both DNA and STING and localized together in the cytosol. The knockdown of DDX41 expression blocked the activation of TBK1, NF-κB and IRF3 by B-form DNA [15]. In addition, RIG-I has been described to sense cytosolic B-form DNA via RNA polymerase III to trigger responses by transcription factor IRF3-dependent IFN-1[16]. Furthermore, AIM2 senses cytosolic B-form DNA and activates inflammasome responses via the adaptor apoptosis-associated speck-like protein [13,16].

Recent research has revealed a plethora of endosomal and cytoplasmic DNA sensors that may recognize plasmid DNA and hence contribute to the immunogenicity of DNA vaccines. This new knowledge should thereby create opportunities to exploit these mechanisms to further enhance DNA vaccine immunogenicity.

DNA vaccine safety

Preclinical vaccine safety evaluation includes assessment of local reactogenicity and systemic toxicity plus histopathology in appropriate animal models. One of the initial safety concerns for DNA vaccines was the risk of integration of partial or complete plasmid sequences into the host genome by insertion mutagenesis, thereby risking inactivation of tumor suppressor genes or activation of oncogenes or causing chromosomal instability (breaks and rearrangements). Fortunately, these concerns have not been realized with experimental data showing that the rate of plasmid integration is negligible and lower than the spontaneous rate of mutation in mammalian genomes [17]. DNA injection rarely results in the long-term persistence of plasmid in tissues distal from the site of vaccine administration [1]. However, plasmid with modified backbones to enhance gene expression, or modified by a new delivery method, may increase the risk of integration and needs to be evaluated for persistence of plasmid DNA before clinical use. Other safety issues include the possibility that DNA vaccines could stimulate the production of anti-DNA antibodies, associated with autoimmune disorders, such as systemic lupus erythematosus. Data from animal studies showed that DNA vaccines could increase the production of anti-DNA autoantibodies but did not increase disease severity in lupus-prone animals nor induce autoimmunity in healthy animals [18,19]. Injection of mice with a plasmid encoding hepatitis B surface antigen (HBsAg) induced liver and kidney damage that was explained by the prolonged expression of HBsAg resulting in the formation of immune complexes [20]. Parker et al. failed to find evidence of pathological changes after repeated injections of DNA in mice or rabbits [21]. The 2007 US FDA guidance on DNA vaccines concluded that no preclinical studies are required from the sponsors for assessment of the effect on autoimmunity. Given their immature immune system, newborns exposed to foreign antigens could develop tolerance rather than protection [22]. However, immunity rather than tolerance occurred when DNA vaccines were administered to children [23,24]. Thus, despite initial concerns, because of the novelty of this approach, extensive recent evidence including multiple clinical trials supports the safety of DNA vaccines for routine human prophylactic and therapeutic use.

Use of traditional adjuvants with DNA vaccines

Traditional adjuvants act as immune stimulators or antigen delivery systems, or both. These include killed bacteria, bacterial components, aluminum salts, oil emulsions, polysaccharide particles and biopolymers, which when coadministered with a protein antigen, enhance its immunogenicity. As detailed below, most if not all of these traditional adjuvant compounds have been tested for their ability to enhance the immunogenicity of DNA vaccines, with mixed results (Table 1).

Alum has been widely used as a vaccine adjuvant since 1926. Recent studies indicate that alum may activate caspase 1 through inflammasomes containing the cytosolic receptor NLRP3 [25]. Flach et al. used the sophisticated experimental approach of atomic force microscopy to indicate that alum interacts directly with membrane lipids on the surface of DCs [26]. A recent study suggested that the adjuvant effect of alum is mediated by cell death and the subsequent release of host cell DNA, which acts as a potent endogenous immunostimulatory signal [27]. Addition of alum adjuvant to DNA-encoding HBsAg enhanced antibody responses in mice, guinea pigs and nonhuman primates [28]. Boosting HBsAg protein-primed mice with alum-adjuvanted DNA vaccine increased IgG2a production reflecting a Th1 shift [29]. Subsequent studies did not find a major adjuvant effect of alum when combined with antigens such as CMV glycoprotein B (gB) [30], botulinum neurotoxin [31] or Leishmania Mexicana GP63 antigen [32]. Hence, the use of alum as a DNA vaccine adjuvant has been supplanted by more promising strategies.

Polysaccharides are sugar molecules on the surface of many plants and microorganisms (e.g., fungi and bacteria). Mammalian cells have evolved receptors to recognize such sugars on pathogens and thereby trigger innate immune activation. Advax™ (Vaxine Pty Ltd, Australia), a polysaccharide adjuvant based on deltainulin particles, has recently emerged as a strong adjuvant candidate, combining high potency with safety and low reactogenicity [33–35]. Advax adjuvant enhanced both humoral and cellular immune responses and enabled antigen sparing when combined with a range of vaccine antigens, including Japanese encephalitis antigen [36]. Advax adjuvant has a high potency to reactogenicity ratio as compared with other adjuvants [37] and when combined with gp120 protein boost significantly enhanced the humoral and cellular immune response following initial priming of mice with an env-encoding DNA vaccine [38]. The combination of DNA priming with sequential adjuvanted protein boost via both intramuscular and intranasal routes provided high systemic and mucosal immune responses against gp120 able to neutralize homologous HIV strains. The immune responses generated by priming with DNA vaccine followed by nasal and intramuscular protein boosts with Advax adjuvant were long-lived and showed minimal attenuation 26 weeks postimmunization. Another polysaccharide, zymosan, a yeast cell wall consisting of protein/carbohydrate complexes with the major constituent being β1–3-glucan, enhanced the humoral and cellular response to a DNA vaccine in mice, an effect blocked by neutralizing antibody to complement factor 3 [39]. The adjuvant effect of zymosan was absent in complement factor 5-deficient DDD and AKR mice, consistent with the adjuvant effect being mediated by complement [39]. A synthetic glucohexaose (β-glu6), an analogue of lentinan, enhanced hepatitis B core antigen (HBcAg)-specific cytotoxic T lymphocytes (CTLs) and IgG and IgG2a against HBcAg, consistent with an adjuvant effect [40]. In addition to immune activation, sugar structures may also help plasmid uptake into cells. Formulation of DNA vaccine with mannosylated cationic liposomes enhanced plasmid transfection of macrophages [41]. The coating of cationic liposomes with mannan also enhanced the ability of a DNA vaccine to induce HIV-specific cellular immunity [42]. Nasal administration of plasmid DNA-loaded chitosan nanoparticles enhanced seroprotection and mucosal IgA against HBsAg, an effect not seen in groups immunized with naked plasmids or plasmid DNA adsorbed to alum [43]. Their excellent human safety record, low reactogenicity, convenient manufacture and ability to be used as immune targeting systems thereby make polysaccharide compounds strong contenders for human DNA vaccine adjuvants.

Liposomes are vesicles composed of phospholipids and cholesterol, suitable for antigen or plasmid delivery. Liposomes entrap or bind plasmid DNA and facilitate DNA entry into cells by penetrating the lipid bilayer of the cell membrane [44]. Liposomes can protect DNA from degradation by serum proteins and after the release of plasmid DNA following fusion with endosomes [45]. Formulation of DNA vaccine into liposomes enhanced cellular and humoral immunity [46,47] with the efficacy further enhanced by exploiting various scavenger and other receptors to target liposomes to APCs [48,49]. A DNA vaccine against measles virus hemagglutinin (HA) and fusion glycoproteins formulated with cationic liposome in Rhesus monkeys enhanced neutralizing titers, H- and F-specific IFN-γ production and protection against infection [50]. A cationic lipid-based adjuvant (Vaxfectin^®; Vical, USA) formulated with a DNA vaccine targeting influenza virus NP and M2 proteins enhanced protection of mice against lethal viral challenge [51]. In a Phase I clinical trial, Vaxfectin-formulated H5N1 influenza DNA vaccine induced seroprotective hemagglutination inhibition titers in 50–67% and T-cell responses in 75–100% of healthy subjects [52]. Liposome adjuvants may be effective by other routes. An oral liposome-formulated vaccine based on plasmid encoding Mycobacterium tuberculosis antigen (Ag85A) induced antigen-specific cellular and humoral mucosal immune responses in mice [53]. Topical administration of plasmid DNA with cationic liposomes enhanced immunogenicity of an HBsAg-encoding vaccine [54]. Significant hurdles that need to be overcome before routine adoption of liposome formulations include managing a high rate of injection-site reactogenicity and potential long-term stability issues.

Nanoparticles based on biodegradable polymers have been developed as DNA vaccine delivery systems. Synthetic polymers such as poly(vinylpyridine), polylactide-co-glycolides (PLG) and polylactide-co-glycolide acid (PLGA) have a long history of biocompatibility and safety in humans [55,56]. Encapsulation of antigens into PLG microparticles was first demonstrated in the early 1990s [57]. Encapsulation of DNA helps protect the plasmid from nuclease degradation and provides prolonged release. HBsAg-encoding plasmid DNA formulated with PLGA and cetyltrimethylammonium bromide increased HBsAg-expressing APCs in the draining lymph nodes and was associated with increased antibody titers and T-cell immunity that translated into enhanced protection against challenge with transplanted HBsAg-expressing tumor cells [58]. A DNA vaccine expressing foot and mouth disease antigens (P1-2A3C3D) and GM-CSF formulated with PLG enhanced T-cell responses and neutralizing antibody in sheep and enhanced protection against clinical symptoms, viremia and carrier status [59]. Administration of cationic PLG with HIV protein (p55Gag)-encoding plasmid DNA enhanced vaccine responses and protection in vaccinated animals [60–62]. Injection of PLG-encapsulated DNA microparticles followed by electroporation increased cytokine expression, recruitment of APCs and plasmid antigen expression, as compared with immunization with PLG-DNA alone [63]. Tumor antigen (ZYC300 and ZYC101)-encoding plasmid DNA encapsulated in biodegradable polymer microparticles was evaluated in a Phase clinical trial in cancer patients and was shown to induce detectable immune responses and clinical improvement [64,65]. Polymer nanoparticles show promise, therefore, as DNA vaccine adjuvants.

Lipopolysaccharide (LPS), which binds and activates TLR4 inducing proinflammatory cytokines, is a potent activator of innate immunity. The price for enhanced immunogenicity with LPS is increased reactogenicity. LPS coadministered with DNA vaccine decreased IgG responses in a dose-dependent manner upon intradermal but not intramuscular injection, with no effect on CTL responses [66]. With plasmids injected by gene gun there was no relationship between DNA vaccine LPS content and CTL responses [67]. High-dose LPS with a DNA vaccine encoding HER2/neu (HER2) increased the Th1 antitumor response, whereas low-dose LPS reduced the efficacy of the vaccine and skewed the response in a Th2 direction [68]. Coadministration of high-dose LPS (≥500 IU LPS/mg plasmid DNA) with an influenza A-encoding intradermal DNA vaccine enhanced the CD8⁺ response [69]. The high reactogenicity and toxicity of LPS, combined with variable results when it has been included in DNA vaccines, makes it an unlikely candidate for human DNA vaccines. Although monophosphoryl lipid A, a less toxic derivative of LPS that retains some adjuvant activity, has been approved in combination with alum in several protein-based vaccines, little data is available on its utility as a DNA vaccine adjuvant [42,70]. CpG motifs are known to increase DNA vaccine immunogenicity [8]. Vaccine immunogenicity was reduced by methylation of the CpG motifs but was restored by coadministration of exogenous CpG or by addition of CpG sequences to the plasmid backbone [8,71], with CpG motifs helping vaccines induce a Th1 response [72]. The addition of CpG sequences in plasmid DNA targeting HPV E7 enhanced the IFN-γ, granzyme B and antitumor response, an effect further enhanced by electroporation [73]. Plasmid CpG may also influence transgene expression in vitro as depletion of CpG was associated with decreased protein expression [74]. On a note of caution, high-dose CpG reduced the immunogenicity of a DNA vaccine [75] and introducing 16 CpG motifs into a DNA vaccine enhanced, while 50 CpG motifs reduced, vaccine immunogenicity [76,77]. Thus, whilst CpG motifs are easy to design into vaccine plasmids, their unpredictable dose–response behavior makes them less attractive candidates for human DNA vaccine adjuvants.

Overall, combination of plasmid DNA with traditional adjuvants has proved to be only modestly beneficial, at best, for enhancement of DNA vaccine immunogenicity. Consequently, adjuvant approaches using alum or LPS have been largely superseded in DNA vaccines by use of plasmid-encoded molecular adjuvants, as discussed in subsequent sections. However, there is still promise for some traditional vaccine adjuvant approaches, in particular polysaccharide-based nanoparticles, to be used in DNA vaccines to modulate the immune response, induce chemotaxis to the immunization site and act as delivery vehicles to enhance plasmid stability and uptake into cells. The aforementioned and other [53,78–81] clinical trials of conventional vaccine adjuvants with DNA vaccines are detailed in Table 1.

Plasmid-encoded adjuvants

DNA vaccine delivery

Heterologous immunization

Vaccination strategies based on prime–boost regimens can enhance responses against difficult pathogens such as HIV or malaria. These strategies involve the priming of the immune responses by using an antigen expressed first by a vector, typically a plasmid DNA, followed by a boost with administration of the same antigen in a different vector, typically a recombinant virus. Common vectors include adenovirus, fowlpox and vaccinia virus. Heterologous immunization often shows a strong synergistic effect compared with homologous immunization [195]. In clinical trials, DNA-prime adenovirus-5 boost induced both CD4⁺ and CD8⁺ T cells against HIV [196]. DNA-prime, MVA-boost strategies similarly enhanced immunity against HIV [197]. Recently, a DNA prime–poxvirus boost (NIVAC) approach targeting HIV-1 was evaluated in a Phase I trial [198]. Two DNA followed by two poxvirus-boost injections induced polyfunctional, durable T-cell responses to Env. Responses were higher with heterologous DNA–NIVAC compared with homologous NIVAC–NIVAC regimens [198]. Extensive efforts have been made to develop an optimized prime–boost malaria vaccine, based on priming with DNA or viral vector followed by a boost with a range of recombinant viral vectors including MVA, fowlpox or adenovirus [199]. A prime–boost strategy of DNA boosted with ME-TRAP (MVA multiple epitope-thrombospondin-related adhesion protein) reduced malaria parasite burden in the liver, as compared with homologous immunization, delayed parasitemia and reduced mortality following sporozoite challenge [195,200,201]. Substitution of plasmid DNA for priming with fowlpox virus elicited sterile protection in two of five volunteers, with persistent memory T cells observed for 20 months [202]. A DNA-prime protein-boost strategy targeting malaria CSP induced antibody and T-cell responses in the clinic [203]. In a recent study, DNA priming followed by an adenovirus- 5 boost targeting HIV CAM-1 Gag successfully induced Gag-specific T-cell responses [204]. DNA-prime protein-boost strategies have also been tested, using inactivated or recombinant protein antigen [205,206]. DNA priming with Env followed by intramuscular gp120 protein boost formulated in Advax polysaccharide adjuvant significantly enhanced the humoral and cellular immune response to HIV Env protein [38].

The use of DNA vaccines for priming followed by a proteinor vector-based boost strategy is highly promising as it provides the best of both worlds in terms of priming a strong T-cell response while at the same time maximizing neutralizing antibody responses. Furthermore, when combined with an adjuvanted protein boost strategy, this approach faces minimal regulatory hurdles as the individual DNA and protein components are both known to be safe and well tolerated, such that no problems would be anticipated with the combined approach.

Plasmid design

Essential plasmid features for a DNA vaccine include a strong promoter for optimal expression in mammalian cells, a selection marker and polyadenylation (polyA) sequences to stabilize the transcripts. In addition, since optimal expression of some genes depends on the splicing of their transcripts, the inclusion of an intron in the promoter–enhancer complex may improve expression. Approaches to increase the transcription and translation may thereby improve DNA vaccine immunogenicity in humans. This can be achieved by an optimization of transcription elements in the plasmid backbone. The promoter is an important component of the plasmid that drives the expression of the gene of interest. Classic promoters for DNA vaccines included the human CMV/immediate early or CMV-chicken-β actin (CAGG) promoters. CMV promoters are used for most DNA vaccines since they drive high constitutive expression levels in a wide range of mammalian tissues [207] and do not suppress downstream read-through. Improvement of expression and immunogenicity have been observed by modifying CMV promoters (i.e., incorporation of HTLV-1R-U5 downstream of the CMV promoter) or by using chimeric SV40-CMV promoter [208]. Alternatives to CMV promoters include host tissue-specific promoters, which avoid constitutive expression of antigens in inappropriate tissues, but in general these have resulted in lower immunogenicity [209]. The presence of an intron in the vector backbone downstream of the promoter can enhance the stability of mRNA and increase gene expression. A kozak sequence immediately prior to the ATG start codon may further enhance antigen expression [210]. The use of species-specific codons increases antigen expression and may thereby enhance DNA vaccine immunogenicity [211]. Gene expression can be manipulated by altering the polyA sequence, which is required for proper termination of transcription and export of mRNA from the nucleus. Many current DNA vaccines use the bovine hormone terminator sequence [212]. Alteration of the polyA sequence may enhance gene expression of DNA vaccines [213] and thereby the immune response [214].

Hence, optimal plasmid design for DNA vaccines should combine bacterial and eukaryotic elements into a vector designed to allow a high copy number during production and high mammalian expression, thereby ensuring maximal immunogenicity.

Expert commentary & five-year view

DNA vaccines are a promising alternative approach to traditional protein vaccines for the prevention and treatment of infectious diseases, cancer and allergy. However, while DNA vaccines have proved effective in animal models, they have continued to suffer because of suboptimal human immunogenicity. Hence, a major frustration in the DNA vaccine field is the inability to predict human vaccine responses based on mouse immunogenicity data. This means that quantitative differences observed between different DNA vaccine approaches in mice rarely translate to humans. For example, electroporation or cytokine adjuvants that may deliver 100–1000-fold improvements in antibody titers in mice may deliver at best a two-to three-fold improvement in titers in humans. Thus, small animal models that are typically used in adjuvant comparison studies are more likely to be misleading than informative, and instead the field would be advised to only use large animal models or best of all humans to undertake adjuvant comparison studies. Hence, while major advances to DNA vaccine design have moved them closer to their true potential for human use, head-to-head human comparative studies of the different DNA vaccine technologies are required to enable selection of the most effective technology for a given disease. Improvements have been made in plasmid construction, such as codon optimization and removal of cryptic splice sites, the use of molecular adjuvants, electroporation and intradermal needle-free injection, and heterologous prime–boost strategies. Among recent promising advances is the use of cytokine, chemokine or costimulatory molecules as molecular adjuvants in DNA vaccines. When combined with strategies to enhance antigen expression, such as electroporation or intradermal needleless injection, molecular adjuvant approaches have the potential to significantly enhance DNA vaccine immunogenicity. The large number of ongoing human clinical trials of DNA vaccines is a testament to the belief that DNA vaccines will ultimately deliver on their potential [2]. Given the regulatory challenges it is most likely that the first DNA vaccine breakthroughs will be in the area of therapeutic vaccines for cancer, with infectious disease applications following much later. This is reflected in the fact that cancer indications currently represent the largest group of human DNA vaccine trials. To realize the full potential of DNA vaccines there is an ongoing need for research into their mechanisms of action and, in particular, into strategies to further enhance their immunogenicity. Such research should allow one or more human therapeutic cancer DNA vaccines to achieve licensure within the next 5–7 years.