|Home | About | Journals | Submit | Contact Us | Français|
Gene delivery holds great potential for the treatment of many different diseases. Vaccination with DNA holds particular promise, and may provide a solution to many technical challenges that hinder traditional vaccine systems including rapid development and production and induction of robust cell-mediated immune responses. However, few candidate DNA vaccines have progressed past preclinical development and none have been approved for human use. This Review focuses on the recent progress and challenges facing materials design for nonviral DNA vaccine drug delivery systems. In particular, we highlight work on new polymeric materials and their effects on protective immune activation, gene delivery, and current efforts to optimize polymeric delivery systems for DNA vaccination.
The principles of vaccination pioneered by Jenner and others in the early 19th century remain relatively unchanged in clinical practice. Despite almost two centuries of use and development, only 27 human diseases are recognized by the CDC as preventable by vaccination. All of these diseases are infectious in nature but do not include rapidly emerging or mutating entities such as cancers, malaria, and HIV, for which classical vaccines are typically ineffective. In general, traditional vaccination is largely protein-based and entails direct administration of dead or attenuated pathogens, recombinant proteins, or virus-like particles. For many targets, protein-based vaccines generate incomplete immune responses and fail to induce protective or therapeutic affects.[2,3] Vaccination with traditional protein-based vaccines typically generates only antibody-mediated (“humoral”) immune responses and often requires periodic booster injections.[2–4] However, cell-mediated responses are required for clearance of intracellular pathogens and generation of cytotoxic T-lymphocyte (CTL) cells that kill infected cells and suppress cancers. Currently, only attenuated live organism vaccines generate significant cell-mediated immune responses, but these are associated with certain safety concerns and can be difficult to manufacture.
Gene-based vaccination offers an attractive alternative to traditional vaccine strategies. The protein targets of immune responses, termed antigens, are encoded in DNA and produced within the body’s own cells, which can mimic live infection more closely than injection of traditional nonreplicating vaccines. Intracellular production of antigens from DNA can result in coordinated activation of both humoral and cell-mediated responses. DNA-based vaccines offer practical advantages as well. DNA vaccine targets can be simply altered by changing the sequence of the DNA. Additionally, DNA vaccines potentially allow for both prophylactic and therapeutic vaccination strategies. [3,7–9] Finally, the ability to cheaply and rapidly produce plasmid DNA from bacteria makes DNA vaccination an attractive prospect economically for future medicine and global health.[11,12] The largest drawback facing gene-based vaccination remains the difficulty in intracellular delivery of DNA.
The concept of DNA vaccination was developed almost two decades ago when it was recognized that intramuscular (IM) injection of naked plasmid DNA in mice yielded production of encoded proteins. Early demonstrations of DNA vaccines in mice used DNA-coated gold microparticles administered via ballistic injection (gene gun technology) and naked DNA injection IM. Within a decade, a myriad of potential applications of DNA vaccines had been developed targeting infectious agents, various cancers, allergy, and immune dysfunction,[8,17] and by 1998 early clinical trials reported induction of immune responses against HIV and malaria in humans. Building upon experiences with other applications of gene therapy, numerous clinical trials have been conducted for DNA vaccines yet no system has received FDA approval. The only licensed and approved DNA-based vaccines are for animal use. One targets flavivirus (West Nile virus) infection in horses and has also been used to protect wild Californian condors; the other has been used to protect commercial salmon against infectious hematopoietic necrosis virus. Though not for human use, these vaccines are significant for establishing the technical potential and commercial value of DNA vaccines.
One of the most commonly cited problems with DNA vaccines in particular, and gene therapy in general, is low levels of gene expression (“transfection”).[23–25] Weak transfection limits immune responses. As in all gene therapies, the large plasmid DNA molecule must be transported across multiple organ, tissue, and cellular barriers into the nucleus for expression of antigen to occur. Gene delivery events must also occur in the proper setting and to the proper cell-type in order to coordinate innate and adaptive immune responses and generate strong immunity.
Herein, we review recent advances in polymeric materials for DNA transfection and their application to DNA vaccines. We begin by outlining the steps involved in immune responses to a DNA vaccine and the role played by material interactions with the immune system. Barriers to antigen expression and generation of immune responses are discussed in the context of novel biomaterials-based approaches to gene delivery systems. High lighting recent advances in materials for gene delivery, we focus on polymeric systems applied to DNA vaccination and the lessons learned that provide insights toward design criteria and challenges for development of novel materials for DNA vaccines.
There are three major branches of the immune system classically described as the cell-mediated, humoral, and innate immune responses. The cell-mediated and humoral branches together generate the adaptive immune responses responsible for neutralizing pathogens, clearing infection, and creating very specific memory immune responses such as CTLs and high affinity antibodies. These responses are mediated by antigen presenting cells (APCs), and B and T lymphocytes. The great diversity of the adaptive immune response is generated through controlled and random genetic rearrangements occurring in the precursors to B and T cells. By contrast, the innate immune system is responsible for the passive immune response and relies upon genetically pre-determined pattern recognition receptors (PRR) to detect foreign substances or tissue damage. These three subsystems have specific components but do not necessarily function independently. The goal of vaccination is to activate the adaptive branches to induce protective memory responses, and generating a robust protective immune response typically requires all three systems.
Specialized immune cells such as the dendritic cell (DC) and macrophage (MP) play a central role in integrating innate immune signals and coordinating adaptive immune responses.[4,27,28] These professional APCs capture antigens and process them into small peptide fragments for presentation to T-cells loaded on cell-surface major histocompatibility complex (MHC) molecules. Antigen signal strength, sustained antigen presentation from DCs, and DC activation drive T-cells to proliferate, differentiate into effector cells, and eventually generate memory T-cells. T-cell mediated signals in turn drive the proliferation and maturation of B-cells. Thus, the first steps in DNA vaccination are significant production of antigen and APC capture of antigen (Fig. 1).
Antigens of intracellular origin, produced within the APC itself, are presented on MHC class I for recognition by CD8+ T cells. When activated, CD8+ T-cells differentiate into CTLs. CTLs recognize and kill cells infected with intracellular pathogens such as viruses or cells presenting abnormally expressed proteins such as tumors. Antigens produced by transfection of a bystander cell can be captured by the APC; antigens of extracellular origin are presented on MHC class II for recognition by CD4+ cells (Fig. 1). CD4+ T-cells mature into different types of helper T-lymphocytes (TH1, TH2, TH17, or Treg) and secrete soluble signals that accelerate or suppress different aspects of cell-mediated and humoral responses.[4,31] TH1 signals such as interferon gamma (IFN-γ) typically combat intracellular pathogens and are desired vaccination outcomes for preventing viral infection and controlling cancers. TH2 signals, such as interleukin-4 (IL-4), are associated with IgE production and are typically necessary for controlling parasite infection but are also involved in asthma and allergy.
While highly compartmentalized, the MHCI and MHCII presentation pathways are not entirely distinct, and through a process termed “cross-presentation” antigens from one pathway may be presented through the other (Fig. 1). Hence, CTL activation to DNA vaccines does not require direct transfection of APCs, as evidenced by effective cell-mediated responses following transfection of muscle cells. Antigens produced and released from a non-APC can be taken up by an APC, processed, and cross-presented on MHCI. Cross-presentation is very important in generation of complete immune responses to vaccines, and activation of cross-presentation can be particularly important in DNA vaccination.[34,35]
Extracellular antigens are also specifically recognized by B-cells, which reside in lymph nodes and produce antibodies to neutralize extracellular pathogens. Thus, humoral activation requires sufficient extracellular antigen accumulation in lymph nodes following DNA vaccination. When B-cells are activated by TH cells they produce high affinity antibody molecules, differentiate into plasma cells to secrete massive amounts of antibodies, and also become memory B-cells (Fig. 1). Preformed antibodies secreted at the mucosal barriers and circulating in the blood become the first line of passive immune defense to invading pathogens and are important for preventing future infections.
Antigen expression alone is not enough, however, to induce efficient immune responses. The body does not normally activate immune responses against “self antigens and generally requires a “danger signal” to discriminate between “foreign” and “self antigens. Thus, supplying a second danger signal is equally as important as achieving efficient antigen expression. With traditional protein-based vaccines the incorporation of adjuvants is often required to couple a strong early innate immune response to adaptive immune activation. Adjuvants are broadly defined as substances that enhance vaccine responses but are not immunogenic themselves, and in the case of genetic vaccines the materials used for DNA delivery can also serve as adjuvants.
APCs express PRRs that recognize structurally and evolutionarily conserved motifs on bacteria or viruses (see also Section 3.3). During infection, these PRRs provide the secondary danger signal early (minutes to hours) prior to the generation of an adaptive immune response.[7,27] Danger signals can also include chemicals released by damaged cells, stimuli from other immune cells,[27,28] or interaction with foreign materials. When immature APCs are activated by danger signals they mature to efficiently present antigen on MHCI and MHCII to T-cells alongside costimulatory molecules, which communicate the danger signal to T-cells. The costimulatory signals on the APC surface include CD40, CD80 (B7-1), and CD86 (B7-2), and soluble cytokines may also be released from APCs (Fig. 1). Quantifying changes in expression levels of these costimulatory molecules is often used as a surrogate measure of DC maturation in response to DNA vaccine delivery platforms.
Antigen presentation without costimulation results in immune tolerance and negative deletion of auto-reactive T-cells.[4,17] Additionally, innate immune activation affects DC levels of MHCI and MHCII, cross-presentation, and DC-mediated signals that polarize the helper CD4+ T-cell response (TH1 vs. TH2, etc.).[28,32] As DCs mature, they reduce their potential for cross-presentation while increasing peptide-loaded MHCI and MHCII at the surface together with costimulatory molecules. Therefore, innate immune recognition, APC maturation, and presentation of costimulatory signals is as important for generation of immune responses to DNA vaccines as is transfection and presentation of the antigen itself.
The goal of DNA vaccination is transfection of an APC or a bystander cell to produce antigens in an immunostimulatory setting. However, the field of genetic vaccines has so far been limited by a lack of safe and effective gene delivery systems. The main recombinant viral vectors used for gene delivery are adenovirus, adeno-associated virus (AAVs), retrovirus, and lentivirus. The advantages of adenovirus are infection of a wide range of human cell types, ability to infect nondividing cells, and lower risk of insertional mutagenesis. However, adenovirus expression is short lived and adenoviruses can cause a severe, even lethal, inflammatory response due to prior immune exposure. AAV, which depends on adenovirus or another virus for replication, has also been used for gene delivery with the advantages of predictable chromosomal insertion and no known pathological consequence of infection. The main advantage of retroviruses, their ability to integrate into the host genome for long-term expression, is also their main disadvantage as this integration can cause mutagenesis and potentially cancer. Retroviruses are also further limited by their inability to infect nondividing cells. Lentiviruses, which can transfect a broad spectrum of cell types, are the most efficient method to transfect DCs in vitro and in vivo. Yang et al. recently reported very high levels of immune activation and therapeutic tumor rejection following immunization with a lentiviral vector engineered to target DCs by the cell surface receptor DC-SIGN.
In particular, Merck & Co. has advanced the use of viral DNA vaccines for HIV vaccination. While there have been some successes in using viral gene therapy and many clinical trials are currently ongoing there are currently no approved protocols. Problems with viral delivery systems include immunologic priming to the vector itself, oncogenicity due to insertional mutagenesis, difficult manufacturing, and limited DNA cargo capacity. [48,49] Clinical trials have tragically highlighted some of these safety risks as viral gene delivery has resulted in both cancer and deaths.[42,50,51] Recently Merck & Co. halted its Phase III HIV adenovirus vaccine prompting renewed questions about the utility of viral vectors. The safety challenges and limitations of viral vectors have resulted in increased interest in nonviral approaches to gene delivery using nonviral materials.
In general, the nonviral methods of DNA vaccination utilized in clinical trials, recently reviewed by Lu et al., rely on physical methods. Injection of naked DNA plasmids has found limited success in humans particularly when injected intramuscularly, even though in smaller animal models naked DNA vaccination produces robust humoral and cell-mediated responses. However, the rapid degradation/clearance [half-life of under 5 min if injected intravenously (IV)] of unprotected nucleic acids, poor induction of humoral immune responses in DNA vaccination in larger animals, and requirement for large doses[54,55] has hindered progress into clinical trials.[53,56] Clinically relevant physical methods that have been employed include electroporation, ballistics (gene gun), ultrasound, and magnetofection. Encapsulation or complexation of DNA with a biomaterial can significantly enhance DNA stability, cellular uptake of DNA, and final protein expression. Materials shown to possess potential for the delivery of genes include inorganic nanoparticles and surfaces that bind to or encapsulate DNA.[61–64] Cationic biomolecules including lipids,[65–68] polysaccharides,[69,[70 polymers,[24,71,72] and dendrimers[73,74] can also electrostatically complex anionic DNA to facilitate transfection. Unless specifically designed to do so, DNA delivered nonvirally has low potential for genomic integration. Nonviral delivery systems for gene therapy are generally cheaper to manufacture, easily scalable from laboratory to GMP-scale production, and are typically more robust for long-term storage compared to their viral counterparts.
Despite achieving greater efficacy than naked DNA administration (IM, IV, or otherwise), physical methods for gene delivery are often limited due to local tissue damage and insufficient gene expression. Research into nonviral gene delivery has been ongoing since the 1970s, and as understanding of the mechanisms of gene delivery has grown, the design of synthetic biomaterials has become more advanced. However, while there have been advances, nonviral methods of gene delivery generally still have lower efficacy than viruses.[24,26,77]
There are many potential bottlenecks that must be overcome for successful DNA delivery (see Table 1),[24,26,77,78]] a process broadly defined as transfection. Plasmid DNA must first be packaged into particles. Requirements for gene delivery include protection of plasmid DNA from degradation, localization to the tissue and cell types of interest avoiding off-target distribution, minimal inactivation by interaction with serum proteins, low clearance from the blood or interstitial space, and efficient transport through the extracellular matrix to the surface of target cells. Next, the DNA-containing particles must associate with cells and become internalized into them by cellular uptake processes (Fig. 2 and Section 3.3). Following uptake, DNA-containing particles must escape the endosomal/phagosomal compartment into the cytoplasm and release their DNA cargo. DNA must finally translocate into the nucleus to be transcribed into mRNA and subsequently translated into protein antigen. Viruses have evolved to accomplish these steps and provide a framework for the design of synthetic delivery particles. Despite efficient uptake of particles of a variety of sizes, in vitro and in vivo transfection of DCs is still notoriously difficult to achieve. APCs are specialized not only for uptake of antigen but also rapid and efficient antigen processing. As a key role of APCs is to internalize and process pathogens for immune activation, APCs may have greater protection against foreign (viral) DNA entry into the nucleus, which may be a barrier to DNA vaccination.
Whereas in vitro investigation of transfection efficiency in cell culture can be used to identify promising materials for transfection, there exist multiple extracellular barriers to effective DNA vaccination in vivo. DCs reside in the blood, in the skin (Langerhans cells), other mucosal barriers, and in lymph nodes. MPs also exist in lymph nodes, as circulating precursors in the blood (monocytes) that differentiate as they enter inflamed tissue cites, and as specialized MPs lining the spleen and liver (Kupfer cells) forming the phagocytic part of the reticuloendothelial system (RES). Access to these APCs is therefore determined not only by route of injection, but also by ability of a particle to drain into lymphatic systems or activate inflammatory signals to recruit APCs (Table 1). For example, Reddy et al. have illustrated size-based targeting of lymph-node resident DCs by accessing lymphatic vessels with 25 nm particles; lymphatic drainage and DC uptake was significantly reduced with 100 nm particles.
Many cationic delivery materials, both polymeric and lipid, form vector-nucleic acid particle (VNP) complexes by electrostatic interactions with the negative charges on the phosphate groups of the DNA backbone (Fig. 3). A net positive surface charge can facilitate transfection by interacting with the negatively charged glycoproteins at the cell membrane.[80,81] However, electrostatic interactions can also rapidly lead to aggregation of these VNPs with serum proteins; as VNP-protein aggregate size increases they can be eliminated from circulation by the phagocytic MPs of the RES, deposit nonspecifically in microvascular beds, or crash out of solution, which may cause acute toxicity. [82,83] In addition to containing high concentrations of negatively charged proteins, plasma also has a significant ionic strength. Interactions between serum proteins, blood and interstitial fluid solutes, and polycationic carrier materials can lead to competitive binding, destabilization of the VNP, and subsequent premature release of the nucleic acid payload. Interaction with complement proteins, C3 and C4 in particular, can activate the innate immune responses resulting in acute inflammation and lead to severe acute toxicity or death. Mucosal surfaces and serum also contain DNase and RNase enzymes that specifically degrade nucleic acids. Condensed VNPs prevent the degradation of the nucleic acid payload by steric inhibition of these D/RNases. The addition of poly(ethylene glycol) (PEG) and other hydrophilic polymers can be used to prevent aggregation with serum proteins and subsequent rapid clearance. This simple functionality can sharply increase the serum half-life of a particle and prevent acute toxic events due to nonspecific interactions, but also results in lower transfection efficiency and reduced cellular targeting. [24,54]
Toxicity at the cellular level and/or due to interactions with the immune system, liver, kidneys, or other complex organ systems can be a concern with nonviral gene delivery. For example, polyethylenimine (PEI) has been shown to be an effective transfection agent but has also been reported to be toxic in animal models.[85,86] Polycations such as PEI and cationic lipids such as 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) also tend to activate complement and the RES, aggregate with serum proteins, and can aggregate with red blood cells as well. Some toxicity issues both in vitro and in vivo can be addressed by chemical modification of PEI. For increased safety, biodegradable gene delivery systems have also been developed[24,88] including degradable cross-linked PEI, poly(ortho esters) (POE), and poly(β-amino esters) (PBAE), and are further discussed in Section 4.2.
Particle uptake by phagocytosis (particles > 500 nm), macropinocytosis, and receptor-mediated endocytosis are particularly important routes of entry into APCs.[4,78,92] Delivery systems can be designed to exploit these avenues (Fig. 2). Cell-specific targeting can significantly enhance transfection efficiency and the desired therapeutic outcome.[24,54,83] The direct conjugation of targeting moieties such as receptor ligands, peptides, sugars, aptamers, and antibodies can increase cell and tissue specificity and transfection efficiency. Additionally, targeting can be based upon size-specific signals to avoid off-target affects.
A variety of strategies exist for targeting APCs. First, APCs express Fc-receptors, which bind to the constant region of antibodies to facilitate uptake of antibody-coated foreign bodies. APCs also express complement receptors that help clear complement-opsonized particles. Lectin-binding receptors, such as the mannose receptor, and scavenger receptors that recognize apoptotic bodies, certain bacterial components, and other nonself motifs are PRRs commonly found on APCs that can enhance particle uptake and may trigger innate immune activation (Sections 2.4 and 3.4). Second, unlike most other cell types, immature DCs constitutively sample their extracellular fluid environment nonspecifically through macropinocytosis to maintain immune surveillance and vigilance for foreign particles, as well as to present endogenous proteins for maintenance of self-tolerance.[4,98,99] This constant sampling may explain DC-targeting by particle size. Particle size may also influence the specific route of entry, as reviewed recently by Xiang et al. Particle surface characteristics also play a role in uptake, as cationic particles more readily associate with the negatively charged glycoproteins on the cell surface and promote nonspecific uptake, [78,92] and other surface characteristics can activate opsonization.
Champion et al. have illuminated the role of particle shape in influencing phagocytosis by MPs, as well as provided simple methods for creating materials with complex shapes and sizes to take advantage of particle physical properties. While spherical micro- and nanoparticles are efficiently phagocytosed by lung alveolar MPs, phagocytosis can be inhibited by contact with odd geometries due to an inability to form the necessary actin structures. In general, size plays a more significant role with particle association with the cell surface than with internalization. These studies suggest a role for particle surface nano-and micro structure in the design of APC-targeted DNA vaccine delivery systems.
As a key integrator between innate and adaptive immune responses (Section 2.4), the APC is a significant target for DNA vaccination. For example, DCs can control the type of helper T-cells (TH1 vs. TH2 vs. Treg) generated in response to vaccination, and innate immune signaling is a key factor in delineating this level of immune control. Biomaterials play a significant role as vaccine adjuvants in controlling the activation of APCs in the setting of DNA vaccine delivery. A wide variety of polymers are known to activate APCs, although the mechanisms of these interactions are not completely understood.[37,78] Many of the PRRs involved with binding and internalization of particles also serve the dual role of activating APCs to provide secondary danger signals.[78,104] Poly(lactic-co-glycolic)acid (PLGA) is one of the best studied DNA delivery vehicles in terms of innate immune activation, recently reviewed in detail by Babensee and Jilek et al. Fischer et al. reported that PLGA does not induce high levels of APC activation, even with a variety of surface coatings, but others have reported that PLGA and PLGA-based microparticles can indeed induce primary DC maturation.[106–109] Other materials can intrinsically generate higher levels of activation, such as POE and those that efficiently bind and activate complement with surface hydroxyl groups.[79,111]
DNA may itself be immunostimulatory as a danger signal, and localized delivery may determine the adjuvant properties of DNA. Toll-like receptors (TLRs) are a class of evolutionarily conserved PRRs heavily expressed on APCs that may be targets for adjuvanting DNA vaccine delivery[27,113] by developing materials that mimic the natural ligands of TLRs. TLR9 in particular recognizes CpG motifs often found in most plasmid DNA and has been suggested as a target for DNA vaccine adjuvants,[114,115] which would require DNA delivery to endosomes of the plasmacytoid DC that expresses TLR9 in high copy. Recently, Ishii et al. reported that TBK-1, an intracellular cytosolic PRR of DNA, is required for DNA vaccine immune responses in mice after IM injection coupled with electroporation. TBK-1 therefore represents another target for delivery of DNA to adjuvant DNA vaccines.
A myriad of materials have been developed for gene delivery. Early chemical methods of increasing the efficacy of gene delivery focused on co-precipitation of the DNA with salts such as calcium phosphate.[61,62,117] More recently, inorganic materials have also been combined with polymers to form hybrid gene delivery nanoparticles. For example, textured surfaces and silica nanoparticles have been shown to be effective for gene transfer in vitro, and organically modified silica nanoparticles have been shown to deliver genes in vivo. Gold nanoparticles have also been combined with PEI for hybrid gene delivery systems.[119,120] A sample of gene delivery materials applied for DNA vaccines can be found in Table 2.
Cationic lipids have been the nonviral gene delivery vectors of choice for clinical application since Felgner first introduced their use in 1987. The cationic lipid molecule consists of a hydrophilic positively charged head group, a linker that may impart some functionality such as pH sensitivity, and a hydrophobic long chain tail. A prototypical cationic lipid is DOTAP; it is the most widely used lipid for gene delivery. For in vivo delivery, nucleic acids are usually encapsulated into liposomes: vesicles with lipid bilayer membranes that exist as large unilamellar vesicles (LUVs) or multi-lamellar vesicles (MLVs). Liposomes generally consist of a single cationic lipid or a mixture of cationic lipids that facilitate nucleic acid binding and transfection, cholesterol or diolelphophatidylethanolamine (DOPE) to impart some rigidity or stability to the complex, and PEG to shield particles from aggregation, serum components, or other nonspecific interactions. Lipids have been used extensively in gene therapy and are the main nonviral delivery vectors used in clinical trials. Unfortunately some lipoplexes are toxic, interact nonspecifically with serum proteins and cells, aggregate quickly, activate the complement system, or have low in vivo efficacy.[66,67] One promising approach that may address these problems is to increase the chemical diversity of lipid-like materials through combinatorial synthesis approaches.[125,126]
Early cationic polymers for gene delivery include poly(l-lysine) (PLL) and PEI and are shown in Figure 4. These polymers can electrostatically bind and condense plasmid DNA into nanometer sized complexes. Although PLL can be used to form small VNPs, PLL is not an effective polymeric vector by itself presumably because it does not have an efficient mechanism for endosomal escape. Consequently, in order to have DNA released to the cytoplasm when using PLL, an additional endosomolytic agent is also required such usually chloroquine, but this is not amenable to in vivo use.
PEI can condense DNA as this polymer has high cationic charge potential, with one out of every three atoms of the polymer backbone being nitrogen (Fig. 4). PEIs high density of near-neutral pKa groups buffer the acidic environment of the endosome, and it has been hypothesized that this provides for endosomal escape through the “proton sponge” mechanism.[127,128] In this model, endosomes acidify through an ATP-mediated, pH-dependant hydrogen-ion pump and passive channels allow internalization of chloride ions to maintain electroneutrality. The primary, secondary, and tertiary amines (in the branched form) of PEI have pKas spanning the physiologic pH. The amines absorb incoming hydrogen ions like a “proton-sponge,” and this has been hypothesized to induce the continued influx of chloride ions leading to an osmotic pressure increase that eventually destabilizes the endosome membrane. PEI VNP complexes then escape the late endosome into the cytoplasm before the vesicles encounter the lysosome thereby avoiding degradation of the nucleic acid payload by lysosomal nucleases. However, PEI is not an ideal gene delivery agent by itself as it can also causes cell necrosis and apoptosis.
One promising approach to overcome some of these safety concerns is chemically modifying PEI. Thomas and Klibanov recently found that N-acylation of branched 25 kDa PEI can double its gene delivery efficacy while reducing its cytotoxicity, and dodecylation of 2 kDa PEI can create a nontoxic polymer five-fold better than branched 25 kDa PEI. Subsequent work showed that linear PEIs could be synthesized by the acid-catalyzed hydrolysis of poly(2-ethyl-2-oxazoline) to form PEI87 and PEI217, polymers that are two orders of magnitude more effective than “off-the-shelf” linear PEI (PEI25), although these polymers still show cytotoxicity. PEI87 also delivered DNA in vivo as efficiently as deacylated PEI25, which was 10000-fold higher than “off-the-shelf” linear PEI25, and with 200-fold increase in specificity to the lungs. Deacylation of commercial PEIs removes residual N-propionyl groups, which increases the DNA binding affinity as well as the pH buffering capability of the PEI. These results demonstrate how important polymer structure is to gene delivery function.
Chitosan is a linear cationic amine-containing polysaccharide that is made from the deacetylation of chitin, a naturally occurring polymer and the major component of crustacean exoskeletons (Fig. 4). Chitosan can be used to condense DNA into 100–200 nm nanoparticles, although they can have variable stability in the presence of serum proteins. Transfection using lactosylated chitosan was found to be comparable to PEI for in vitro delivery to HeLa cells, and chitosan/DNA particles have also been shown to work in vivo for intratracheal delivery to epithelial cells in the central airways of mice.
Another promising approach for polymeric gene delivery is the use of dendrimers, recently reviewed by Lee et al. Dendrimers are repeatedly branched cascade polymers that are synthesized from a central core and branch outward (Fig. 12). With each increasing synthesis generation, dendrimers become more branched, larger in size, and have a multiplicative increase in the number of end groups. Like PEI, several dendrimers including poly(amidoamine) (PAMAM), make effective gene delivery polymers due to their large number of secondary and tertiary amines that can buffer the endosome. Particular advantages of dendrimers include a more monodisperse population than other polymers, which are typically polydisperse, and fine control over dendrimer generation number that can be tuned for a desired application thereby controlling dendrimer size, number of functional groups, etc. A disadvantage of dendrimers is that they require a multi-step synthesis that can be more expensive and time-consuming than the synthesis of linear polymers. Dendrimers do not need to be in their perfect spherical shape to function. In fact, efficacy can be improved by effectively pruning the branches of the dendrimer or by using the branches themselves. Partially degraded or “fractured” PAMAM dendrimers have been shown to have dramatic > 50-fold enhancement for gene delivery as compared to complete PAMAM. Alternatively, the core of the dendrimer can be removed, leaving identical branch fragments known as dendrons.
PBAEs are cationic, biodegradable polymers of the general structure shown in Figure 4 that are useful for gene delivery. Their facile synthesis allows for the creation of large combinatorial libraries of materials.[72,134] They can be synthesized in parallel by the conjugate addition of amines to diacrylates, a one-step reaction that can be performed without solvent and that does not produce any byproducts. PBAEs can be synthesized to have a range of molecular structures (over 2 000 unique structures synthesized), molecular weights (2–50 kDa) and degradation half-lives (hours-to-days).[72,134,135]
Lead PBAEs bind to and condense DNA to form cationic nanoparticles that have high cellular uptake, and like PEI, facilitate endosomal escape.[72,136,137] Semi-automated approaches have been developed to create large libraries of PBAEs and evaluate their gene delivery efficacy. These studies have also been helpful to determine how certain cationic polymer structures affect polymer/DNA nanoparticle biophysical properties and final gene expression.[138,139] Lead PBAE VNP were positively charged in HEPES, PBS, or sodium acetate buffers, but neutrally charged in the presence of serum containing media. Top-performing PBAEs converged in structure to polymers composed of a linear diacrylate-based backbone with amino alcohol side-chains that differed by only single carbons. Lead PBAEs had 100-fold lower cytotoxicity than PEI and higher gene delivery efficacy.
Next generation PBAEs have been synthesized by the combinatorial modification of the ends of lead polymers (Fig. 5).[140–142] Interestingly, significant enhancement in polymer performance could be generated by simple modification of the polymer ends. Polymers synthesized with end groups that contained terminal primary amines were found to bind DNA more tightly, form smaller polymer/DNA nanoparticles, facilitate higher cellular DNA uptake, and mediate increased protein expression. Leading end-modified PBAEs enabled transfection 100-fold higher than PEI for in vitro delivery to human primary cells and were also promising for in vivo delivery (Fig. 6a). In comparison to lentivirus and adenovirus, these end-modified PBAEs achieve comparable gene delivery in vitro (Fig. 6b). This polymer library approach demonstrates that even subtle single-carbon changes to the structure of the polymer backbone, side-chain, or end-groups can have dramatic effects on gene delivery efficacy.
The encapsulation and release of protein antigen or plasmid DNA from microparticles has proven to be a very effective strategy for passively targeting vaccines to pro-APCs by size-exclusion.[25,38,92,144] Microparticles in the 1–10 µm diameter range are too large for endocytosis and therefore avoid general cellular uptake, but are small enough to be phagocytosed by MPs and DCs (Fig. 2). Encapsulation of plasmid DNA also protects from nuclease degradation. Because of the bulk size of microparticles, a depot is formed at the injection site allowing for sustained exposure to DNA over time as microparticles are slowly cleared by phagocytosing cells. Randolph et al. showed that monocytes are recruited into the injection site within hours, where they phagocytose microparticles and differentiate into DCs by the time they have migrated to draining lymph nodes. Microencapsulation of DNA also provides a method of controlling release rates of DNA, which may be important for timing immune responses by coordinating DC migration to lymph nodes, maturation, and presentation of costimulatory molecules, peak gene expression, and antigen presentation.
The prototypical polymer used for microsphere encapsulation is PLGA (Fig. 4). PLGA degrades into the natural products lactic acid and glycolic acid that can be completely metabolized, and PLGA is FDA approved making it an attractive material for drug delivery systems. PLGA microspheres were used by Hedley and co-workers[149–151] over a decade ago to encapsulate plasmid DNA and elicit CTL-mediated immune responses in vivo by passive targeting to APCs. PLGA–DNA microspheres were also shown to generate mucosal and systemic antibody responses via oral delivery. PLGA–DNA microspheres were subsequently shown to generate therapeutic anti-tumor immunity after parenteral (IM or IV) administration, as well as protect from rotavirus challenge following oral administration.[153,154] Although early studies with PLGA microspheres were promising, PLGA is not an optimal gene delivery material. The encapsulation of DNA in PLGA can lead to DNA instability, poor loading efficiency, slow release, and subsequently low transfection efficiency.[144,155–157] Significant efforts have been made to improve PLGA microsphere function for DNA vaccination by changing polymer physical chemistry (MW, hydrophilicity), methods of microsphere formulation, and addition of secondary materials such as cationic transfection agents (Fig. 3).[38,156–160]
The method of microsphere preparation greatly influences plasmid stability and particle architecture, as well as DNA release kinetics and transfection efficiency.[147,161–164] The processing of PLGA and similar hydrophobic polymers into microspheres is a multi-step process, during which DNA can potentially be damaged.[156,157,164] Formulation typically relies on generating a primary emulsion of DNA in an aqueous or solid phase surrounded by polymer dissolved in an organic solvent phase.
With double emulsion techniques, the first emulsion is then subjected to shear stresses to form small particle sizes within a secondary aqueous emulsion, and the resulting particles are lyophilized after evaporation of solvent. Ando et al. developed a method of formulating microspheres by freezing the aqueous phase of the primary emulsion to reduce shear stresses and incorporating a cryoprotectant (lactose). These improvements help preserve the more functional supercoiled form of DNA during freezing and lyophilization. Alternatively, spray drying of primary emulsions reduces shear forces on plasmid DNA by avoiding the need to generate a secondary emulsion. Additionally, spray drying avoids lyophilization steps that can also damage DNA. When directly compared to double-emulsion, Atuah et al. reported better preservation of functional plasmid DNA in PLGA microspheres by spray drying. Methods of encapsulating DNA into PLGA are still heavily investigated, particularly with other functional materials such as PEI, and have been recently been reviewed by Abbas et al.
PLGA undergoes bulk erosion by hydrolysis. Release of DNA and other large molecules from PLGA can be tuned from days to months (reviewed in detail by Anderson and Shive). DNA release from microspheres typically exhibits an initial rapid burst-release phase over a matter of hours, followed by a second phase with kinetics depending on material type, particle size, and processing parameter (Fig. 7a and and9).9). Long term release of PLGA-encapsulated DNA injected IM resulted in transfection and continual gene expression lasting almost 200 days in mice.
Although the fine structure of the microsphere is typically porous enough for diffusion, polymer hydrolysis can result in acidic products and lower the pH of the microenvironment to between 1.5 and 3.5 for larger particles. Smaller particles may not experience such low extremes of pH but are acidic enough to catalyze hydrolysis of the encapsulated DNA [155,156,167] particularly during the slow, continuous release phase. Both Walter et al. and Ando et al. included buffering salts to reduce pH-mediated DNA degradation.
It is unlikely that continuous release of low amounts of DNA for weeks is beneficial for DNA vaccines considering that the typical phagocytosing cell has a lifespan of days. Faster DNA release may result in greater antigen dose. Tinsley-Bown et al. investigated different molecular weights and blends of PLGA to quicken DNA release rates. While most formulation techniques result in a porous microsphere matrix, they reported a modified double-emulsion technique to form hollow microspheres that allowed for high encapsulation of DNA, up to 11 µg mg−1. This interesting morphology completely released DNA within one week using lower molecular weight PLGA. Diez and Tros de Ilarduya more recently confirmed the effects of lower molecular weights and hydrophilic PLGA resulting in faster release rates from microspheres.
Microparticles generated by double emulsion often show low burst release followed by slow continuous DNA release. By comparison, a greater burst release of DNA was observed by Atuah et al. with spray-dried PLGA microparticles with little or no continuous release. Oster and Kissel directly compared spray drying and double emulsion methods for making PLGA-based microparticles encapsulating pre-formed PEI-DNA complexes. They observed similar trends of high burst release/low continuous release associated with spray drying and low burst release/slow continuous release of DNA from double emulsion PLGA microspheres. While both methods can be adjusted to result in similar-sized particles, it appears that spray drying makes DNA more labile either by increasing porosity or changing the distribution of DNA at the surface.
Adsorption of DNA onto the microsphere surface is one strategy to increase DNA loading and avoid plasmid degradation during formulation.[38,144] Surface-loaded DNA also avoids the difficulties associated with slower erosion-dependent release kinetics, and may increase transfection efficiency by making DNA immediately available at the surface. Multiple different plasmids expressing different genes can be easily absorbed onto the same particles. An additional benefit of a high surface DNA concentration may be increased interaction with PRRs of the innate immune that recognize DNA such as the endosomal TLR9 (Section 2.4). Surface loading can be facilitated by supplementation with the cationic surfactant cetyltrimethylammonium bromide (CTAB) as well as by the incorporation of PEI either into the matrix of the microsphere or at the surface. Many other materials have been investigated including DOTAP, DEAM,, and PLL, have been used to absorb DNA to PLGA microparticles.
Absorbing plasmid DNA onto PLGA/CTAB microspheres has been researched by the Chiron corporation (now Novartis Vaccines) for IM vaccination as recently reviewed by O’Hagan et al. and Singh et al. Surface absorption of plasmid DNA results in high DNA loading (effectively 100% absorption of DNA), transfection of DCs, and antigen presentation in vitro.[170,173] In vivo the PLGA/CTAB system resulted in increased transfection of muscle tissue following IM injection, increased production of antibody titers, and greater activation of CTL responses to plasmid-encoded HIV antigens when compared to PLGA microparticles or naked DNA (IM) alone, Prophylactic IM immunization with a plasmid encoding human carcinoembryonic antigen (CEA), which is associated with colon cancer, protected against tumor growth in a transgenic mouse model when absorbed onto PLGA/CTAB microspheres. In mice, compared to immunization with naked DNA (IM), CD8+ T-cell activity was 100-fold higher and antibody titers were 1 000-fold higher with PLGA/CTAB particles, which was similar to traditional protein vaccination with a strong adjuvant.
By absorbing DNA to the surface as opposed to encapsulating DNA within the microsphere, DNA release is less dependent upon material degradation. Increasing the amounts of CTAB led to increased DNA loading and characteristically delayed DNA release, but immune responses were not significantly improved by increasing the DNA load per microparticle[169,176] indicating that DNA dose was not a limiting factor in this format. Additionally, larger PLGA/CTAB particles (30 µm) were not as effective as smaller particles (1 µm) at recruiting an inflammatory cellular infiltrate, trafficking to lymph nodes, activating infiltrating APCs, and inducing immune responses. The mechanism of PLGA/CTAB-mediated DNA vaccination following IM injection therefore seems to be related to increased availability of plasmid DNA to phagocytosing APCs early in the inflammatory response, leading to early transfection and antigen presentation in the draining lymph node.[106,144] Larger animals such as guinea pigs and nonhuman primates also responded to vaccination with PLGA/CTAB microparticles with increased humoral and cell-mediated activity compared to naked DNA vaccination.[175,176] Most recently, PLGA/CTAB particles were evaluated for long term efficacy in generation of a protective immune response against measles virus challenge 1 year following immunization of rhesus macaques. Immunization by IM and subcutaneous (SC) routes induced humoral and cellular immune responses greater than naked DNA IM vaccination, but these responses were not sustained. Vaccination ID was inferior to IM vaccination with respect to both antibody production, CTL generation, and protection from sequelae of challenge with measles virus. Because naked DNA vaccination by the SC route has previously been shown to be slightly more efficient at generation of both humoral and CTL-mediated immune responses, the results with PLGA/CTAB microparticles indicates the need to further understand the differences between microparticle-based and naked plasmid DNA vaccines in terms of mechanism and route of immunization. PLGA/CTAB microparticles were recently advanced into stage I clinical trials by Novartis for HIV-1 DNA vaccination.
Incorporating PEI into PLGA microspheres has also been developed as a method for avoiding the problems associated with internal encapsulation of plasmid DNA.[38,171] PEI imparts a positive charge to the PLGA microsphere. Unlike CTAB, which by itself is not a transfection agent, PEI has intrinsic ability to form nanoparticles with DNA and increase transfection efficiency. Kasturi et al. have reported a method for covalently attaching PEI to the PLGA microparticle surface that minimizes the amount of labile PEI, thereby incorporating the endosomal buffering capacity of PEI without high toxicity. PLGA–PEI microspheres prepared by this method improve in vitro transfection and cause upregulation of costimulatory signals on APCs.[107,108] Increased survival against a lethal dose lymphoma tumor challenge was observed following ID vaccination with PLGA microspheres prepared with branched PEI on the surface. While both IM and ID PLGA–PEI vaccination routes provided protection against lymphoma tumor challenge, IM vaccination was superior in a follow-up direct comparison of vaccination routes.
PLGA can also be used to encapsulate and release pre-formed PEI–DNA nanoparticles. PEI protects DNA during encapsulation and release. Microspheres release DNA in PEI–DNA nanoparticle with kinetics similar to that of simple PLGA microparticles and also efficiently transfect nonphagocytic cells as well as APCs.[168,171,180] Howard et al. reported increased splenic transfection following oral administration in mice of PLGA containing PEI–DNA particles. Recently, Zhou et al. demonstrated that PLGA–PEI–DNA microspheres increase humoral responses compared to naked DNA following IM vaccination and can induce efficient CTL responses at doses lower than that obtained with naked DNA vaccination IM.
PBAE materials for gene delivery have also been used for improving the characteristics of PLGA microspheres. Little et al. [109,182] described the modification of PLGA microspheres with up to 50% w/w PBAE polymer (Fig. 7). The addition of PBAE required modifications of the double emulsion method including incorporating high salt buffers to stabilize the secondary emulsion. The addition of PBAE to the PLGA matrix increased DNA loading and protected plasmid DNA in the supercoiled state, possibly by forming PBAE–DNA complexes during the encapsulation process. The added PBAE assists in buffering the pH through tertiary amine groups resulting in a more pH-neutral microenvironment within the degrading microsphere. Both the total release and release rate of plasmid DNA from the PBAE/PLGA microspheres over 1 week is reduced by incorporation of increasing amounts of cationic PBAE polymer. This phenomenon reflects a balance between burst release of surface-labile DNA and ionic retention of DNA (Fig. 7).
PBAE-modified microspheres exhibited drastically increased transfection efficiencies. Transfection of the phagocytosing P388D1 murine MP cell line in vitro with 25% PBAE microspheres resulted in up to five orders of magnitude greater transfection efficiency than simple PLGA and approached levels observed by transfecting with lipofectamine, a commercially available cationic lipid.[109,182] Transfection of P388D1 cells with PBAE/PLGA microspheres was shown to depend entirely upon phagocytic processes as the addition of cytochalasin-D, which inhibits actin-mediated phagocytosis, completely abolished transfection with microspheres.
Little et al. further demonstrated that PBAE/PLGA microspheres are promising candidates for DNA vaccine delivery. Incorporation of PBAE into PLGA activated primary murine DCs in vitro at levels similar to or exceeding that achieved with soluble lipopolysaccharide (LPS), which is a strong adjuvant. In vivo investigations revealed that immunization with either 15 or 25% PBAE/PLGA microspheres caused an efficient CTL response that was robust and target-specific resulting in tumor regression when mice were vaccinated against the model SIY antigen and then challenged with a SIY-bearing syngenic tumor (Fig. 8).
POE are a class of pH-sensitive polymers originally developed in the 1970s to be biodegradable and nontoxic alternatives to PLGA. The fourth generation POEs, synthesized from the reaction of a diketene acetal and a diol, contain the pH-sensitive orthoester linkage as well as glycolic acid or lactic acid monomers in the polymer backbone, which upon initial hydrolysis act as latent acids to catalyze further acid-mediated degradation of the orthoester linkages (Fig. 4). As a result, POE IV polymers have reproducibly controllable erosion rates depending on the monomer compositions, can exhibit surface erosion, are pH sensitive, and are biodegradable. These properties make POE IV an ideal candidate for controlled drug delivery.
Wang et al. reported the development of two versions of POE IV specifically designed for DNA delivery with pH-sensitive rapid release of DNA in the acidifying phagosome following microparticle uptake by APCs. In vitro studies demonstrated low cytotoxicity in a MP cell line (P388D1) and preservation of supercoiled DNA structure during encapsulation and release. The addition of a tertiary amine in POE2, which would be protonated and charged at the acidic pH of a late endosome, slowed the escape of DNA from the degrading polymer matrix as it eroded. POE1 completely released the encapsulated DNA payload within 2 days whereas the amine groups in POE2 retarded DNA release by 3–4 days (Fig. 9).
In vivo these polymers can function as DNA vaccine carriers, and this immune function was correlated to the timing of DNA release.[147,161] Both POE polymers elicited stronger humoral and cell-mediated responses than PLGA. Further, activation of CTLs was gene-specific, induced memory responses, and resulted in tumor regression when mice were vaccinated against the model SIY antigen and then challenged with a SIY-bearing syngenic tumor (Fig. 10). However, the POE1 polymer, which released DNA quickly, did not function as well as the slow releasing amine-containing POE2 polymer following ID vaccination.
It was hypothesized that the release kinetics of DNA from POE2 microspheres are more closely matched to the natural timing of DC transfection and migration to lymph nodes.[147,161] In follow-up studies, Nguyen et al. investigated in vitro the properties of these POE microspheres as well as the effects of blending PEI into the POE microsphere matrix to enhance transfection efficiency. Interestingly, POE1 microspheres displayed a six-fold greater transfection efficiency than POE2 microspheres. The timing of maximal transfection correlated well with the timing of DNA release, with maximal transfection after 2–3 days with POE1 microspheres and after 5–6 days with POE2 microspheres ( and unpublished results). Despite greater transfection by POE1 microspheres in vitro, the in vivo performance of POE microspheres seems dominated by the slower release kinetics of POE2. By comparison, PLGA microspheres had very low transfection even with incorporation of PEI to enhance transfection efficiency. Blending POE with PEI greatly enhanced transfection of all sphere types, slowed DNA release, and induced high levels of DC maturation.
Without PEI or PBAEs, the mechanism of phagosome escape for microspheres is unclear, though mechanical rupture or increased osmotic pressure from degradation products are possible. In the POE system, it appears that release kinetics correlate with transfection in vitro, as the addition of PEI was not able to increase POE2 transfection efficiency to that achieved with more rapid release from POE1 microspheres. PLGA and POE alone do not facilitate high levels of transfection in vitro even though they generate measurable immune responses in vivo. In the POE system, though POE1 has greater transfection, the slower releasing POE2 polymer was more immunogenic. As the microsphere degrades inside a DC, the DC must migrate to the lymph node, upregulate danger signaling, and present antigen. It is possible that the release kinetics of DNA from the POE2 polymer, PLGA-PBAE, and cationic PLGA with surface-absorbed DNA are more appropriate for this natural timing.
The majority of polymer-mediated DNA vaccine delivery technology has focused on microparticle formulations due to the convenience of passive size-based targeting to APCs. However, polymeric nanoparticles offer attractive size-based properties such as ability to traffic directly to lymph nodes, multiple routes of uptake, and potential to influence types of immune outcomes. Additionally, given the difficulty in formulating microspheres, nanoparticles may offer an industrial advantage as they may be cheaper to produce and are readily filtered for sterilization. Generally speaking, nanoparticle formulations achieve greater transfection efficiency than microparticle formulations. Nanoparticles have been shown to increase uptake across skin and mucosal surfaces to target mucosal associated lymphatic tissues (MALT). Nanoparticles of the proper size can traffic to lymph nodes directly, [25,79] and particles size influences the TH1 versus TH2 bias of an immune response to protein antigens.
Chitosan in particular has been investigated as a nanoparticle carrier for DNA vaccines due its mucoadhesive and drug delivery properties across mucosal surfaces, both as a stand alone material and in conjunction with transfection agents such as PEI. As the structurally related chitin is recognized by MPs in the induction of allergy, it is possible that chitosan triggers innate immune PRRs on APCs. Chitosan–DNA nanoparticles induce stronger activation of DCs than naked DNA alone. Chitosan may be potentially useful for DNA vaccines to induce immune tolerance by shifting the TH1 versus TH2 response, and oral immunization with chitosan nanoparticles modulates anaphylactic allergic immune responses. Chitosan derivatives have been shown to increase epithelial permeability, perhaps by affecting tight junctions. Oral delivery of chitosan–DNA nanoparticles approximately 150 nm in size induced gene expression in the intestinal epithelium. Intranasal (IN) administration of similar chitosan–DNA nanoparticles protected against challenge with respiratory syncytial virus (RSV) due to generation of efficient mucosal antibody responses and cell-mediated responses. Larger (~300nm) chitosan–DNA particles were also shown to induce both humoral and cellular immune responses that are protective against RSV challenge following IN and ID administration.
PLL, which can condense DNA due to positive charges, has been examined as a material for DNA nanoparticle vaccines. Putnam et al. developed a class of transfection agents based upon PLL grafted with imidazole with an optimized balance of high transfection activity and low cytotoxicity. PLL–imidazole nanoparticles formed ~150 nm particles with plasmid DNA, and when administered ID in mice high antibody titers were developed against HIV envelope antigens. Minigo et al. have developed formulations made of PLL-coated polystyrene nanoparticles that condense plasmid DNA. In vitro, transfection was more efficient with 50 nm particles than 1 µm particles. Transfection of less mature DCs was greater than that observed with matured DCs, likely due to the reduced uptake in matured DCs. Size dependence of immune response was also observed in vivo following ID immunization, as only 50 nm PLL-PS particles induced cellular and humoral immune responses while neither 20 nm nor 1 µm particles were immunogenic. The preferential immune activation to 50 nm particles was independent of CD4+ TH activation and occurred at low DNA doses. This work implies a role for in vivo size-based targeting to Langerhans cells or other immature DCs, and confirms size-dependent uptake studies by the same group and others.[92,198]
Greenland et al. investigated the use of some first generation PBAE polymers for adjuvanting IM immunization as nanopartide complexes with plasmid DNA. They discovered multiple polymers that increased cell-mediated immune responses, and observed that polymers with hydrophobic backbones and alcohol end caps were best able to enhance muscle cell transfection in vivo. It was also observed that increased transgene expression in vivo correlated with increased CTL activation. However, seeming dose–response effects followed a log–linear relationship, implying that a practical threshold level of transfection where incremental improvements in gene transfection may not result in increased immune responses. Vaccine responses were better than both naked DNA administration (IM) and administration with the poloxamer CRL1005 (see Section 5.8).
We are currently developing novel PBAE-based materials specifically for DNA vaccination through reselection of core PBAE polymers, optimization of side-chain termini, as well as next-generation PBAE iterations. The availability of such a large, structurally diverse, and modifiable polymer library for genetic vaccination may allow further improvement upon immune responses, and will hopefully generate molecular-level mechanisms of APC transfection. End-modified PBAE C32 polymers (Fig. 4) form nanoparticles with DNA that are small (~100 nm) in size (Fig. 11a). When these nanoparticles are added to murine DC2.4 cell line DCs growing in cell culture, they show effective gene delivery as indicated by high expression of green fluorescent protein in positively transfected cells (Fig. 11b).
An interesting recent approach to build on dendrimer technology (see Section 4.2) is the development of hybrid linear dendrons for targeted gene delivery. Wood et al. have developed a self-assembling PAMAM dendrimer PEG linear block copolymer for targeting APCs. The linear-dendrimer block copolymer can be functionalized for APC-specific targeting by the addition of a mannose receptor ligand at the end of the PEG chain (Fig. 12a,b). This multi-functional platform system was rationally designed to efficiently deliver DNA plasmids for DNA vaccination and exhibits specific structure-based functionalities for bypassing many of the barriers of intracellular and extracellular nucleic acid delivery. The PAMAM block of the copolymer enhances transfection while the PEG linker prevents nonspecific adherence to cell membranes. Mannose targeting increases APC transfection, and engagement of these receptors may also stimulate APC activation. The PAMAM–PEG–mannose block copolymer demonstrated effective condensation of DNA into organized, uniform nanoparticles approximately 150 nm in size (Fig. 12c). In vitro these VNPs exhibited low cytotoxicity and high ligand-specific transfection efficiencies when compared to commercially available PEI, particularly the fifth and sixth generation (MG5 and MG6) dendrimers. These in vitro results are promising and we are currently investigating the use of this system in vivo as a DNA vaccine delivery agent.
Poloxamers, known by the trade name Pluronics, are a well-studied group of copolymers used as surfactants in a variety of pharmaceutical applications including vaccine delivery and as adjuvants for DNA vaccines.[201,202] Poloxamers are thought to act as adjuvants by recruiting and activating APCs. Poloxamers consist of blocks of poly(ethylene oxide) (PEO) flanking a central poly(propylene oxide) (POP) core, and CRL1005 is a triblock copolymer that has a POP core of 12kDa flanked with 350 Da PEO. Evans et al. reported that CRL1005 forms micro-particles spontaneously above a phase transition temperature (~6–8°C), though formulation with the cationic surfactant benzalkonium chloride (BAK) reduces particle sizes into the nanometer range (200–300 nm). Without BAK, DNA does not associate with CRL1005. However adding plasmid DNA to BAK–CRL1005 particles increases the size to 300+ nm, indicating CRL1005–BAK–DNA particle formation, and these ternary nanoparticles were shown to increase cell-mediated immune responses in nonhuman primates. While the mechanism of CRL1005 without BAK as a DNA vaccine adjuvant is poorly understood, there is some evidence that CRL1005 enhances delivery of DNA in vivo. Greenland et al. demonstrated that increased expression of antigen in vivo following IM injection of CRL1005–DNA was correlated to increased immune responses in mice. Hartikka et al. have demonstrated increased antigen expression leading to greater immune responses in a DNA dose-dependent manner following injection of CRL1005–BAK–DNA particles. CRL1005 has also been shown to be safe and practical. Vilalta et al. demonstrated long plasmid life in vivo following IM injection compared to naked DNA, and plasmid persistence was independent of integration into the host genome. Additionally, CRL1005-DNA particles can be sterile-filtered prior to forming particles above the phase change temperature as well as frozen for later use.[202,203] In pre-clinical trials, CRL1005–BAK–DNA nanoparticles increased cell-mediated and humoral responses to cytomegalovirus (CMV) antigens. Wloch et al. reported that CRL1005–BAK–DNA induced cell-mediated and humoral CMV responses in humans in phase I clinical trials, and this formulation is now in phase II clinical trials. Poloxamer was also blended with PLGA to form nanoparticles for nasal delivery of DNA to elicit a strong humoral response.
PLGA microspheres containing DNA encoding for the E6 and E7 genes of human papillomavirus virus (HPV) 16 and 18 have been advanced into clinical trials by Zycos, Inc. (now MGI Pharma, Inc.) to treat advanced pre-cancerous cervical intraepithelial neoplasia (CIN) by inducing CTL-mediated responses to HPV-infected pre-cancerous epithelial cells.[208,209] A phase I dose-escalation trial of PLGA microparticles encapsulating plasmid DNA encoding only HPV-16 E7 antigen (ZYC101) established T-cell immunologic responses in 11/15 study participants and complete clinical response in 5/15 patients following three IM vaccinations. In the most recently published results of a phase II study, patients with CIN grade 2/3 were injected IM with 100 or 200 µg DNA encoding for both the E6 and E7 antigens in PLGA microspheres (ZYC101a/Amolimogene) or saline placebo. In patients under 25 years of age, 70% experienced clinical resolution after treatment compared to 23% with placebo (p = 0.007). Direct measurements of immunological responses were not reported. However, these two clinical trials demonstrate the clinical potential of DNA vaccines delivered by PLGA microspheres. Perhaps more significantly, the ability to change the plasmid DNA and retain clinical responses is a perfect demonstration of the versatility of DNA vaccine delivery technology and validates the concept of sequence independence. This result suggests that new vaccines could be created simply by modifying existing platforms to encode for different antigens.
PEI is also under clinical investigation for DNA vaccine delivery. Mannose–PEI was originally used for the ex vivo transfection of DCs, which successfully generated effector and memory CTL responses following SC injection in nonhuman primates mediated by transfection of Langerhans cells in vivo. Interestingly, these robust cell-mediated responses were not accompanied by antibody production. DermaVir is a topical administration of linear PEI conjugated to mannose administered through an intradermal patch for the purpose of generating HIV immunity, and is currently in phase I/II studies.
DNA vaccination holds great potential for combating a variety of diseases. Initial results are promising and some technologies have advanced to clinical trials. Yet challenges remain, and despite decades of research, safe and efficient delivery of plasmid DNA to initiate immune responses remains a major bottleneck in bringing DNA vaccination into human medicine. The use of polymeric materials to elicit DNA vaccine responses holds promise. The tailoring of readily available polymers such as PLGA, chitosan, and PEI has shown much promise in pre-clinical and clinical studies. Polymers tailored for gene delivery including POEs, PAMAMs, and PBAEs have only recently emerged as promising strategies for DNA vaccine delivery.
Development of novel nonviral delivery strategies for DNA vaccines must continue to serve as both methods of biological insight and clinically relevant outcomes. Specific concerns include the observed difficulty in transfecting DCs, methods to target APC uptake and lymph node trafficking, and providing strong danger signals without sacrificing biocompatibility. We are only beginning to understand the effects of particle size, surface characteristics, and material interactions with the innate immune system. Investigating of the underlying biological mechanisms of DNA vaccination requires strategies that can isolate one polymer function from another, such as DNA release kinetics and transfection efficiency. Future development of polymeric and other synthetic materials must focus on these considerations unique to drug delivery for DNA vaccination.
The authors thank Yong Zhang for assistance with TEM and the CMSE Experimental Facilities supported in part by the MRSEC Program of the National Science Foundation under Award Number DMR 02-13282. This work was supported by NIH Grant EB000244. J. M. Chan acknowledges the financial support from the Agency for Science, Technology and Research, Singapore.
David N. Nguyen received his Ph.D. in Materials Science and Medical Engineering from the MIT and Harvard Medical School joint Division of Health Sciences and Technology in 2008. Dr. Nguyen is currently a postdoctoral fellow in Institute Professor Robert Langer’s lab after which he plans to finish medical training pursuing the field of infectious diseases. His research interests include DNA vaccines, small interfering and immunostimulatory RNA drug delivery, and novel anti-viral and anti-cancer therapies.
Jordan J. Green is an assistant professor of Biomedical Engineering at the Johns Hopkins University. He received his Ph.D. in biological engineering from the Massachusetts Institute of Technology in 2007 and was a postdoctoral associate in Institute Professor Robert Langer’s lab from 2007 to 2008. His work has resulted in over a dozen papers and patents in the areas of chemistry, biomaterials, and drug delivery.
Daniel C. Anderson is appointed at the D. H. Koch Institute for Integrative Cancer Research at the Massachusetts Institute of Technology. He received his Ph.D. in molecular genetics from the University of California at Davis. At MIT, he pioneered the use of robotic methods for the development of smart biomaterials for drug delivery and tissue engineering. In particular, the advanced drug delivery systems he has developed provide new methods for nanoparticulate and microparticulate drug delivery, nonviral gene therapy, siRNA delivery, and vaccines.
Dr. David N. Nguyen, Massachusetts Institute of Technology, 77 Massachusetts Ave, E25 Room 342, Cambridge, MA 02139 (USA)
Jordan J. Green, Massachusetts Institute of Technology, 77 Massachusetts Ave, E25 Room 342, Cambridge, MA 02139 (USA)
Juliana M. Chan, Massachusetts Institute of Technology, 77 Massachusetts Ave, E25 Room 342, Cambridge, MA 02139 (USA)
Robert Longer, Massachusetts Institute of Technology, 77 Massachusetts Ave, E25 Room 342, Cambridge, MA 02139 (USA)
Dr. Daniel G. Anderson, David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, 77 Massachusetts Ave, E25 Room 342, Cambridge, MA 02139 (USA)