The utility of plasmid DNA as an immunogen has been limited by its weak immunogenicity. In the present study, we evaluated the ability of a family of linear polyethylenimine (PEI) polymers, complexed to plasmid DNA, to augment DNA expression in vivo and to enhance antigen-specific adaptive immune responses. We showed that four of five structurally different PEIs that we evaluated increased in vivo DNA expression 20- to 400-fold, and enhanced DNA-induced epitope-specific CD8+ T-cell responses 10- to 25-fold in BALB/c and C57BL/6J mice respectively, when delivered intravenously. Functional studies of the PEI-DNA-induced CD8+ T-cell responses demonstrated that formulation of DNA with PEI was associated with increased numbers of cells secreting type I cytokines. In addition, PEI-DNA complexes improved antigen-specific TH1-helper cell and humoral responses. Most importantly, the PEI-DNA complexes elicited memory cellular responses, capable of rapid expansion and accelerated clearance of a lethal dose of recombinant Listeria monocytogenes. Lastly, we identified physical properties of PEI-DNA complexes that are associated with enhanced DNA-elicited immunogenicity. These findings demonstrate that PEI polymers can play an important role in the development of DNA-based vaccines in the setting of infectious disease prevention and cancer therapy.
DNA vaccine constructs have been evaluated and are being tested in a number of human clinical trials for generating immunity to many types of diseases, including infectious, cancer, allergic, and autoimmune diseases [1, 2]. In addition, the use of DNA vaccines has proven effective in some current veterinary products such as a vaccine against infectious hematopoietic necrosis virus for salmon , melanoma immunotherapeutic vaccine for dogs  and a vaccine against the West Nile Virus for horses . Unfortunately, the immunogenicity of plasmid DNA in humans has proven to be modest in comparison with the immunogenicity observed in other species of microbial expression vectors. Consequently, efforts are being made to increase the immunogenicity of DNA vaccine constructs. Strategies have been developed to increase plasmid DNA expression via codon modification , the use of enhanced promoters  and the use of microbial transcriptional control elements . Since plasmid DNA has poor in vivo transfection efficiency, a number of formulations of plasmid DNA have been evaluated to protect DNA from degradation and to enhance transfection efficiency [9-11]. The “gene gun” technology  was developed to enhance DNA delivery and immunogenicity. Furthermore, immunomodulation strategies have been evaluated for their ability to enhance DNA immunogenicity [13-15].
The synthetic polycationic polymer polyethylenimine (PEI), consisting of chains of ethylenimine units—CH2CH2NH-, has become the gold standard for enhancing the in vivo expression of administered genes [16, 17]. PEI is superior to other nonmicrobial transfection agents because of its ability to protect DNA from degradation , to escape intracellular endosomal lysis  and to efficiently deliver DNA into the nucleus of cells . Formulation with PEI polymers has been employed in DNA- and siRNA-based immunotherapy against cancer [21, 22] and in vaccination studies targeting different infectious agents [10, 23, 24].
In the present study, we have evaluated the physicochemical characteristics of a series of PEI polymers and have assessed associations between particular characteristics of these polymers and their ability to facilitate in vivo DNA expression and enhanced adaptive immune responses. Moreover, we have examined plasmid DNA constructs formulated with these various PEI polymers for their ability to generate immune responses under conditions that maximize the immunogenicity of these PEI-DNA complexes. We demonstrate that selected PEI-DNA complexes can have dramatic immunogenicity and provide protection against challenge with pathogenic microbes.
The majority of PEI-DNA complexes are nanometer-sized and positively charged
The physical properties of DNA complexed to various forms of the PEI polymers were investigated to confirm the formation of PEI-DNA complexes. Since the size and charge of polymer-DNA complexes have been shown to be important factors that affect the kinetics of cellular uptake and the mode of immune stimulation , the size and charge of the PEI-DNA complexes were measured. Five PEI polymers were complexed to plasmid DNA at an N/P ratio of 7.5. Four of the five polymers, namely the max linear 25 kDA PEI (M25), linear 22 kDa PEI (L22), linear 87 kDa PEI (L87), and JetPEI (JET) polymers, formed nanometer-sized, positively charged complexes with similar physical properties (Fig. 1). Their size ranged between 110 nm and 140 nm (Fig. 1A), and their charge ranged between +35 mV and +42 mV (Fig. 1B). In contrast, the complexes of the linear 25 kDa PEI (L25) polymer were larger and more negatively charged than the other PEI-DNA complexes, with a size and a charge of 200 nM and −18 mV, respectively. These data demonstrate that all five PEI polymers form complexes with DNA. In addition, the PEI-DNA complexes created have unique size and charge characteristics, which are dependent on the type of PEI polymer used.
PEI polymers increase DNA expression following intravenous but not intramuscular inoculation
To determine the kinetics of DNA expression in vivo, the different PEI polymers were complexed to plasmid DNA encoding the firefly luciferase protein and were inoculated into mice. Light emission due to firefly luciferase protein activity was analyzed quantitatively using the Xenogen In Vivo Imaging System (IVIS) technology (Fig. 2). The ability of the PEI polymers to increase DNA expression in vivo was first evaluated by monitoring the magnitude of luciferase expression upon intramuscular inoculation. On day 1 postinoculation, high light emission was observed in the quadriceps of mice receiving DNA alone, and low to undetectable light emission was observed in the quadriceps of mice receiving the PEI-DNA complexes (Fig. 2A). At 7 days postinoculation, the magnitude of luciferase expression remained at least tenfold greater in mice receiving DNA alone compared with that in mice receiving PEI-DNA complexes (Fig. 2B).
It is well documented that PEI polymers enhance DNA expression most significantly when the PEI-DNA complexes are delivered intravenously in mice [26, 27]. Therefore, we assessed the magnitude and duration of plasmid DNA expression in vivo in mice receiving DNA complexed to the various forms of PEI by intravenous delivery. At 6-h postinoculation, light emission was detected in the cervical, thoracic, and abdominal regions of mice inoculated with DNA complexed to the M25, L22, L87, and JET polymers (Fig. 2C). The luciferase expression intensities associated with DNA complexed to the L87 and JET polymers were greater than those associated with DNA complexed to the M25 and L22 polymers. No light emission was visualized in mice inoculated with DNA alone or with L25-DNA complexes.
To identify the location of DNA expression, tissues and organs were harvested from mice at 6-h postinoculation and were imaged (Fig. 2D). In mice receiving the DNA alone or DNA complexed to the L25 polymer, DNA expression was barely detectable in the spleen and the cervical and inguinal lymph nodes (cLN and iLN) but was quantifiable in the liver, heart and lungs. In mice receiving the DNA complexed to the M25, L22, L87, and JET polymers, high DNA expression was detected in all previously mentioned organs. Interestingly, the intensity measured in mice receiving the DNA complexed to the M25, L22, L87, and JET polymers was more than 1000-fold greater in the spleen, cervical, and inguinal lymph nodes, than in the mice receiving DNA alone or L25-DNA complexes.
The quantitative data collected at 6-h postinoculation demonstrated that four of the five polymers enhanced DNA expression in vivo when delivered intravenously (Fig. 2E). The M25 and L22 polymers increased expression by 20- to 30-fold, and the L87 and JET polymers increased expression by 300- to 400-fold relative to expression in mice receiving the DNA alone. The expression kinetics in all mice receiving PEI-DNA complexes indicated that DNA expression was highest at 6-h postinoculation and declined linearly thereafter (Fig. 2F). Together, these data revealed the relative abilities of the different PEI polymers to enhance DNA expression in vivo. Additionally, the data provide evidence for the ability of these polymers to induce DNA expression in secondary lymphoid organs.
PEI polymers enhance immunogenicity of DNA and increase antigen-specific CD8+ T-cell responses
Given the differential ability of the PEI polymers to enhance DNA expression in vivo, we assessed the ability of the different PEI polymers to augment DNA-elicited epitope-specific CD8+ T-cell responses. We performed these studies using plasmid DNA encoding the HIV-1 gp120 protein as the immunogen. In BALB/c mice, CD8+ T cells recognize the immunodominant 10-amino acid p18 sequence of the HXB2 gp120 protein . We first explored the potential adjuvant effects of PEI-DNA complexes delivered by the intramuscular route. p18-specific CD8+ T-cell responses of comparable magnitudes, 0.2–0.5%, were induced in mice receiving DNA alone or DNA complexed to the L22 or L87 polymers (Fig. 3A). To determine if intramuscular delivery of PEI-DNA complexes primed more efficiently than DNA alone, we boosted the immunized mice with 107 viral particles of recombinant adenovirus expressing the HIV gp140 protein (rAd-gp140). DNA-primed mice showed peak p18-specific CD8+ T-cell responses of 5% (Fig. 3B), while mice that were previously primed with L22- and L87-DNA complexes showed peak p18-specific CD8+ T-cell responses of 10% on day14 postboosting. These data indicate that, despite the absence of an observed difference in the magnitudes of CD8+ T-cell responses in mice primed with DNA alone or with PEI-DNA complexes, the addition of the PEI polymers enhanced the priming of the DNA, resulting in enhanced secondary CD8+ T-cell responses after boosting.
Since PEI polymers are most potent as transfection agents when delivered intravenously, we next explored the adjuvant effects of the PEI polymers when the PEI-DNA complexes were delivered intravenously. Four of the five polymers, M25, L22, L87, and JET, augmented the DNA-induced p18-specific CD8+ T-cell responses compared with the L25 polymer and DNA alone (Fig. 3C). The highest responses were induced by DNA complexed to the L87 and JET polymers: 18 and 17% respectively, at peak (day 7) and 4% at plateau (day 70). The M25 and L22 polymers improved responses that were of similar magnitudes, with peak values of 8 and 12% respectively, and plateau values of 2 and 3% respectively. DNA alone and DNA complexed to the L25 polymer also elicited responses that were similar in magnitude, with peak values of 0.5% and plateau values of 0.2%. These data demonstrate the potent adjuvant effects of PEI polymers when delivered with DNA. Additionally, they highlight the importance of the route of delivery on the observed adjuvant properties of PEI polymers.
PEI-DNA-elicited CD8+ T cells are functional
Given the large magnitudes of PEI-DNA-elicited CD8+ T-cell responses when PEI-DNA complexes were delivered intravenously, we explored the functionality of these PEI-adjuvanted DNA-induced CD8+ T-cell populations. The effect of PEI polymers on the functionality of this cellular immune response was investigated in C57BL/6J mice using chicken ovalbumin (OVA) as the antigen. The production of IL-2, IFN-γ, and TNF-α by the SIINFEKL-specific CD8+ T cells was initially examined. For DNA alone and the different PEI-DNA formulations, the greatest portion of the cytokine-secreting cells was in the one-cytokine group, followed, in descending order, by the two-cytokine and three-cytokine groups (Fig. 4A). In the mice receiving DNA complexed to the M25, L22, L87, and JET polymers, approximately half of the cytokine-secreting SIINFEKL-specific CD8+ T cells produced only one cytokine; about one-third of these cells secreted two cytokines and the remaining one-sixth of the cells secreted all three cytokines. In the mice receiving DNA alone, approximately two-thirds of the cytokine-producing SIINFEKL-specific CD8+T cells produced one cytokine; a quarter of the cells secreted two cytokines and the remaining one-eighth of the cells produced all three cytokines. The PEI-adjuvanted DNA elicited a greater percentage of cytokine-secreting SIINFEKL-specific CD8+ T cells than DNA alone (Fig. 4B). The majority of the PEI-DNA-induced SIINFEKL-specific CD8+ T cells secreted IFN-γ only, ranging from 15 to 27%; concurrent secretion of IFN-γ and TNF-α only ranged from 10 to 16%, and concurrent secretion of all three cytokines ranged from 5 to 10% (Fig. 4B).
The cytotoxic activity of the epitope-specific CD8+ T cells was also assessed (Fig. 4C). Isolated CD8+ T cells from mice receiving PEI-DNA complexes were incubated with EL4 cells pulsed with the SIINFEKL peptide. The numbers of effectors noted in the effector-to-target ratios were the numbers of SIINFEKL-specific CD8+ T cells. This allowed us to detect any differences at the cellular level in the cytotoxic abilities of CD8+ T cells elicited by the different PEI-adjuvanted DNA. At all effector-to-target ratios investigated, the specific lysis mediated by SIINFEKL-specific CD8+T cells induced by the different PEI-DNA complexes was comparable. No cytotoxicity data were generated for the mice inoculated with the DNA alone because of the small number of SIINFEKL-specific CD8+ T cells that were elicited. The data demonstrate that the strong CD8+ T-cell responses elicited by PEI-adjuvanted DNA are functional, with polyfunctional cytokine-producing and cytotoxic capabilities. Moreover, the functionality of the responses was more or less comparable using the different PEI-DNA complexes.
PEI polymers improved CD4+ T-cell and antibody responses
If DNA is to be a useful immunogen, it must elicit potent antigen-specific CD8+ T cell, CD4+ T-cell, and antibody responses. Therefore, we assessed the adjuvant effect of PEI-DNA complexes on the cytokine-producing ability of CD4+ T cells that recognize the immunodominant ISQAVHAAHAEINEAGR peptide sequence of chicken ovalbumin. The production of IL-2, IFN-γ, and TNF-α by the CD4+ T cells upon ex vivo peptide stimulation of splenocytes isolated on day 7 following intravenous inoculation of mice with DNA alone or PEI-DNA complexes was analyzed. In mice receiving DNA complexed to the PEI polymers, the percentages of responding CD4+ T cells producing IFN-γ and TNF-α were approximately an order of magnitude greater than the percentages of responding CD4+ T cells in mice receiving DNA alone (Fig. 5A). Moreover, there were no significant differences in the magnitudes of IFN-γ- and TNF-α-producing CD4+ T cells in mice receiving the different PEI-DNA complexes, with one exception. In mice receiving DNA complexed to the M25 polymer, the magnitude of IFN-γ-producing CD4+ T cells was much greater, with a value of 0.5%. Interestingly, the percentages of responding CD4+ T cells producing IL-2 varied in the mice receiving different PEI adjuvants. In mice receiving DNA complexed to the M25 and JET polymers, the percentages of responding CD4+ T cells were greater, with a magnitude of 0.5% each, than the percentages of responding CD4+ T cells in mice receiving DNA alone or DNA complexed to the L22 and L87 polymers, where the responses were 0.1%. Responding CD4+ T cells that produced IL-4 or IL-17 were not detected.
The ability of PEI polymers to enhance DNA vaccine-induced humoral immunity was also investigated. Low titer anti-ovalbumin IgG responses were detected on days 21 and 42 postinoculation in mice receiving DNA alone or DNA complexed to the PEI polymers intravenously (data not shown). However, high titer anti-ovalbumin IgG responses were measured in the serum of mice receiving DNA alone or DNA complexed to PEI polymers by day 70 postinoculation (Fig. 5B). In mice receiving DNA alone or DNA complexed to the M25 polymer, IgG levels of 0.7 μg/mL were measured. However, in mice receiving DNA complexed to the L22, L87, and JET polymers, the IgG responses were substantially greater with values of 4 μg/mL, 4 μg/mL, and 7 μg/mL respectively. These data show that the adjuvant effects of PEI polymers are not limited to CD8+ T-cell responses. The data demonstrate that PEI polymers can also enhance both DNA-elicited CD4+ T-cell and IgG responses.
Memory PEI-DNA-elicited CD8+ T-cell responses are protective against a microbial challenge
The protection against a pathogen challenge was assessed in the mice inoculated with plasmid DNA complexed to each of the different PEI polymers. The response against a microbial pathogen exposure was examined using OVA-primed C57BL/6J mice challenged with recombinant Listeria monocytogenes (rLM) expressing OVA (rLM-OVA) as a model challenge pathogen. Clearance of this pathogen is known to be highly CD8+ T cell-dependent . An expansion of SIINFEKL-specific CD8+ T-cell populations in mice previously inoculated with PEI-DNA complexes or DNA alone occurred when the mice were exposed to a sublethal challenge of rLM-OVA (Fig. 6A). The magnitudes of the responding cell populations were greater in mice inoculated with PEI-DNA complexes than in mice inoculated with DNA only. The magnitudes of the SIINFEKL-specific CD8+ T-cell populations in mice receiving L87- and JET-DNA complexes increased by 20-fold; in the mice inoculated with M25- and L22-DNA complexes, the SIINFEKL-specific CD8+ T-cell responses increased by 35-fold. In contrast, these responses in mice inoculated with DNA alone increased approximately tenfold.
Upon challenge with a lethal dose of rLM-OVA, the PEI-adjuvanted DNA induced CD8+ T-cell responses that accelerated the clearance of the pathogen (Fig. 6B). By day 3 postchallenge, no colony-forming units of rLM-OVA were detected in the spleens of mice receiving DNA complexed to the L87 and JET polymers; colony-forming units (CFUs) of 102–103 per spleen were detected in the mice that received DNA complexed to the M25 and L22 polymers. In contrast, CFUs of greater than 106 per spleen were detected in the mice receiving DNA alone and in naive mice. Therefore, the DNA complexed to the optimal PEI polymers afforded significant protection against a challenge with a model microbe. Together, these data provide evidence that the strong PEI-DNA-elicited CD8+ T-cell responses are comprised of functional memory cell populations that are capable of responding to and protecting against microbes.
Previous studies have shown a moderate enhancement of epitope-specific CD8+ T-cell responses elicited by plasmid DNA when it was formulated with polymers [10, 30]. Nevertheless, the responses were less robust than those induced by a number of microbial expression vectors [31, 32]. The family of PEI polymers used in this study was chosen due to their ability to significantly enhance DNA expression in vivo with low cellular toxicity, compared with other forms of PEI . We have identified four PEI polymers, M25, L22, L87, and JET that enhanced epitope-specific CD8+ T-cell responses induced by a single inoculation of PEI-DNA complexes, using less than half the usual quantity of DNA for immunization. More importantly, the magnitudes of the responses elicited were comparable with those induced by a number of microbial expression vectors [31, 32].
The size and charge of polymer-DNA complexes are important factors that affect the kinetics of cellular uptake of polymer-DNA particulates and their ability to stimulate immune responses . Studies have shown that polymer-DNA particles between 20 and 200 nm in size are taken up by cells via receptor-mediated endocytosis ; in contrast, particles greater than 0.5 μm are taken up by cells via phagocytosis or micro-pinocytosis . As a consequence of differences in intracellular processing, nanometer-sized particles have been shown to trigger a TH1-biased cellular immune response  whereas larger microsized particles induce humoral responses . However, reports on the effect of charge on cellular uptake and immune stimulation of polymer-DNA particles are conflicting. Studies have shown that both negatively charged and positively charged particulates have high transfection capabilities and are able to induce immune responses [34, 36], suggesting that the charge may have no significant effect on the immunostimulatory properties of polymer-DNA complexes.
The immunostimulatory PEI polymers assessed in the present study, namely M25, L22, L87, and JET, shared physical properties that were distinct from the L25 PEI polymer that did not enhance immune responses. The complexes of DNA with the immunostimulatory polymers were smaller (110–150 nm) than those made with the nonstimulatory polymer (200 nm). Moreover, the immunostimulatory polymers formed positively charged complexes with DNA while the nonstimulatory polymer formed more negatively charged complexes. The size and charge measurements and the CD8+ T-cell responses elicited postintravenous delivery of PEI-DNA complexes suggest a negative correlation between size of PEI-DNA complexes and the CD8+ T-cell responses, and a positive correlation between the charge of the PEI-DNA complexes and the CD8+ T-cell responses. These findings suggest that there is an association between the physical properties of the PEI-DNA complexes and their immunostimulatory effects.
The ability of PEI polymers to potentiate DNA expression in the lungs and liver is well documented [26, 27]. It was therefore not surprising to detect high luciferase expression in the lungs and livers of inoculated mice. However, we also observed substantial differences in expression of DNA in secondary lymphoid organs (SLOs), namely the spleen and cervical and inguinal lymph nodes, between the mice receiving immunostimulatory PEI polymers and those receiving DNA alone. While studies have documented the ability of PEI polymers to induce DNA expression in SLOs , the present observations suggest that increased DNA expression in SLOs is critical for the enhancement of epitope-specific CD8+ T-cell responses induced by PEI-DNA complexes when delivered intravenously. These findings suggest that increased antigen expression in SLOs may be responsible for the robust adaptive immune responses generated in the PEI-DNA-vaccinated mice.
The adjuvant effect of PEI polymers was lost when PEI-DNA complexes were delivered intramuscularly. Our findings showed that intramuscular delivery of PEI-DNA complexes decreased transgene expression compared with expression of DNA alone. Previous studies have shown that intramuscular delivery of PEI-DNA complexes reduced DNA expression from that observed following intramuscular delivery of DNA alone [24, 38]. This decrease in DNA expression may explain the reduced immunogenicity of PEI-DNA complexes when delivered by the intramuscular route. In addition, a different route of administration may also change the mechanism by which the immune response is induced. Studies have shown that the intramuscular delivery of DNA alone results in the transfection of muscle cells . Cross-priming of APCs is believed to be the usual mechanism by which plasmid DNA elicits immune responses. The present study demonstrated high levels of DNA expression in SLOs following intravenous delivery of PEI-DNA complexes. Therefore, direct transfection of APCs following intravenous delivery of PEI-DNA complexes is a plausible mechanism to explain the immune responses induced in the current series of experiments. Interestingly, upon boosting with rAd-gp140, a dramatic increase in CD8+ T-cell responses was observed in mice primed by the intramuscular route with PEI-DNA complexes compared with the responses seen in mice primed by this route with DNA alone. These data demonstrate that, despite the apparent absence of enhanced priming by PEI-formulated DNA delivered by the intramuscular route, the complexation of DNA with PEI polymers did, in fact, enhance the priming by plasmid DNA. This finding is not surprising. Studies have shown that immunogens that more effectively elicit potent CD4+ T-cell responses, which play an important role in functional primary and secondary CD8+ T-cell immunity, are generally more immunogenic than immunogens that elicit weak CD4+ T-cell responses [40, 41]. Given the enhanced CD4+ T-cell responses observed when PEI-DNA complexes are delivered intravenously compared with responses elicited by DNA alone, it is reasonable to propose that intramuscular delivery of PEI-DNA complexes may elicit better CD4+ T-cell help than plasmid DNA alone.
The addition of the polymers was associated with significant increases in the proportion of plasmid DNA-induced epitope-specific CD8+ T cells that secreted IL-2, IFN-γ, and TNF-α, which are critical for mediating robust adaptive immune responses. The PEI-adjuvanted plasmid DNA-induced responses also included large numbers of cells that simultaneously secreted all three of these cytokines. This finding demonstrates the ability of PEI polymers to enhance the polyfunctionality of epitope-specific CD8+ T-cell responses, a property associated with effective adaptive immune responses. Furthermore, the results of the cell-mediated cytotoxicity assay revealed that the cytotoxicity of epitope-specific CD8+ T cells induced by the various PEI-DNA formulations was comparable, despite differences in the overall magnitudes of these responses. This finding suggests that at a cellular level, the PEI-adjuvanted DNA-induced CD8+ effector cells have comparable lytic potential.
Plasmid DNA has been shown to elicit TH1-biased helper T cells [42, 43]. This observation was confirmed in the present experiments by the detection of IFN-γ-producing CD4+ T cells in mice inoculated with plasmid DNA. The present data suggest that the formulation of plasmid DNA with PEI enhanced TH1-biased helper T-cell responses, with higher magnitudes of IFN-γ- and TNF-α-producing CD4+ T cells, than the responses induced by DNA alone. The absence of IL-4- and IL-17-producing CD4+ T cells induced by this vaccination regimen further corroborates this conclusion.
Humoral immunity is a major defense against extracellular microbes and toxins. The formulation of plasmid DNA with three of the four evaluated PEI polymers, namely L22, L87, and JET, enhanced antigen-specific IgG responses compared with those elicited by DNA alone. The fourth polymer, M25, induced IgG responses comparable to those elicited by DNA alone. Isotype switching from IgM to IgG requires B-cell activation by TH2 helper CD4+ T cells. The current study suggests that the addition of the L22, L87, or JET polymers to DNA enhanced TH2-helper CD4+ T-cell responses, although these responses were undetectable in CD4+ T-cell functional assays. The inability of the M25-DNA complexes to augment DNA-induced IgG responses suggests that M25-DNA complexes did not enhance TH2-helper CD4+ T-cell responses.
CD8+ T cells can recognize and kill cells infected with intracellular microbes, and in so doing, eliminate or control the spread of these pathogens. The PEI-adjuvanted plasmid DNA-induced CD8+ T cells elicited a population of cells that mediated this function. When exposed to rLM-OVA, the SIINFEKL-specific CD8+ T cells induced by PEI-DNA complexes expanded more than 20-fold, showing that PEI-DNA complexes elicited a memory cellular immune response. Within 5 days, greater than a third of circulating CD8+ T cells were SIINFEKL specific. As a result, when exposed to a lethal dose of rLM-OVA, the PEI-adjuvanted DNA-induced CD8+ T cells effectively eliminated infected cells. Within 3 days, no pathogen was detected in the spleens of mice inoculated with DNA complexed to the L87 and JET polymers, and the microbial load was decreased by greater than 100-fold in mice inoculated with DNA complexed to the M25 and L22 polymers compared with that in mice inoculated with DNA alone. These studies confirm that the PEI-DNA-induced antigen-specific CD8+ T cells have the functional ability to secrete cytokines and lyse target cells.
We have assumed thus far that the increased potency of PEI-DNA complexes is due only to adjuvanticity. However, the difference in DNA expression observed when PEI-DNA complexes are delivered intravenously, compared with intramuscular delivery, indicate a potentially significant role for increased transfection of antigen-presenting cells. Accordingly, we acknowledge that there may be other factors, including transfection efficiency that may explain the profound immunogenicity of PEI-DNA complexes. Indeed, the transfection efficiency of PEI made it a promising gene transfection agent for gene therapy over a decade ago. Unfortunately, due to the high toxicity of early forms of PEI , the use of PEI has had a limited role in previous clinical trials. Linear PEI polymers have been shown to have substantially reduced toxicity  and have been tested in humans [46, 47]. While some investigators have attempted to synthesize potent transfection agents such as cationic liposomes and novel polycationic polymers with transfection efficiencies comparable to PEI, PEI is still the gold standard for nonviral transfection agents.
Our studies demonstrate the adjuvant properties of PEI polymers when complexed to plasmid DNA immunogens and administered intramuscularly or intravenously. Intravenous delivery of PEI-DNA complexes elicited potent cellular and humoral immune responses. Currently, there are many FDA-approved systemic cancer therapies in clinical use [48, 49]. In addition, the JET polymer has been used intravenously in two human studies [46, 47]. Hence, PEI-DNA complexes can play a vital role in the development of potent cancer therapies. However, intravenous delivery is not an ideal route for vaccinations against infectious diseases. Our data showed that intramuscular delivery of PEI-DNA complexes did not elicit higher magnitudes of epitope-specific CD8+ T-cell responses compared with those elicited by DNA alone. Still, complexation of PEI to DNA did enhance the priming of plasmid DNA, as demonstrated by improved boost responses. Recent studies have also suggested that the adjuvant effect of PEI can be observed with pulmonary delivery . Other novel delivery routes, such as intradermal or oral delivery, should be explored to determine the best approach to harness the profound effect of polymer adjuvants on the immunogenicity of plasmid DNA vaccine constructs.
Materials and methods
Plasmid DNA vectors
The firefly luciferase gene, the codon-optimized HIV-1 HXB2 env gene, and the chicken ovalbumin gene were each cloned into the Vaccine Research Center (VRC) vector, kindly provided by G. Nabel (Vaccine Research Center, National Institutes of Health), to form pVRC-DNALUC, pVRC-DNAgp120, and pVRC-DNAOVA respectively.
Commercially available linear 25kDa PEI (L25) and Max Linear 25kDa PEI (M25) were purchased from Polysciences. The L25 polymer, (C2H5N)x, with a molecular weight of 25 kDa, contains all secondary amines and up to 7–8% poly(2-ethyl-2-oxazoline) side-chains. The M25 polymer is a fully hydrolyzed form of the L25 polymer, and lacks the poly(2-ethyl-2-oxazoline) side-chains. Elemental analyses of these polymers indicate the absence of impurities. JETPEI (JET), purchased from Polyplus Transfection, is a patented linear polyethyleimine derivate that has been produced to GMP standards for use in clinical trials [46, 47]. Linear 22 kDa PEI (L22) and linear 87 kDa PEI (L87) polymers were synthesized to high purity by the Klibanov lab (MIT) as previously described .
Six- to 8-week-old female BALB/c and C57BL/6J mice were purchased from Charles River Laboratories and The Jackson Laboratory, respectively, and maintained under specific-pathogenic-free conditions. Research on the mice was approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee.
Plasmid DNA was complexed to the PEI polymers as previously described [26, 50]. Briefly, 150 mM stock solutions of PEI polymer were diluted in sterile 5% glucose and added to an equal volume of 0.7 μg/μL plasmid DNA solution in sterile 5% glucose, to achieve a nitrogen-to-phosphate (N/P) ratio of 7.5. The PEI-DNA mixture was incubated at room temperature for 15 min for complex formation. Complexes were further diluted in sterile 5% glucose before use.
Size and charge measurements
PEI polymers were complexed to 100 μg of pVRC-DNALUC as previously described. The size and charge measurements of the PEI-DNA complexes were made using ZetaPALS (Brookhaven) based on dynamic light scattering and zeta potential analysis.
In vivo DNA expression
BALB/c mice were injected intravenously with 30 μg of pVRC-DNALUC alone, PEI-DNA complexes, or 5% glucose only. At 6 h postinoculation, mice were anesthetized with ketamine and xylazine (Sigma Aldrich) and injected intraperitoneally with luciferin (Xenogen). Photonic emissions were measured using In Vivo Imaging System Lumina II (Xenogen). Luminescence was quantified using Living Image (Xenogen). To determine DNA expression in tissues and organs, mice were injected intraperitoneally 6-h postinoculation with luciferin. Murine cervical and inguinal lymph nodes, lungs, heart, liver, spleen, small, and large bowel were harvested and placed in luciferin. The organs were imaged and the luminescence was quantified.
Epitope-specific CD8+ T-cell analyses
BALB/c mice were injected intravenously with 20 μg of pVRC-DNAgp120 alone, DNA complexed to each PEI polymer at an N/P ratio of 7.5, or 5% glucose only. On days of interest postinoculation, mice were anesthetized with isofluorane and retro-obitally bled. Lymphocytes were isolated by density separation using Lympholyte- M (Cedarlane). Isolated lymphocytes were stained with H-2Dd tetramer complexes folded with the HIV gp120 p18-epitope peptide (RGPGRAFVTI). Cells were stained with anti-murine CD8α monoclonal antibodies (BD), fixed and analyzed using a LSRII flow cytometer (BD Biosciences) and FlowJo software (Tree Star). To quantify SIINFEKL-specific CD8+ T cells, C57BL/6J mice were injected intravenously with 20 μg of pVRC-DNAOVA alone, DNA complexed to each PEI polymer at an N/P ratio of 7.5, or 5% glucose only. The mice and blood samples were treated as abovementioned except for the use of H-Kb tetramer complexes folded with the chicken ovalbumin SIINFEKL-epitope peptide for tetramer staining. For boost studies, mice were inoculated intramuscularly with 107 viral particles of recombinant Adenovirus serotype 5 expressing the HIV gp140 protein (rAd5-gp140).
The quadriceps muscles of BALB/c mice were each injected with 10 μg of pVRC-DNAgp120 alone, DNA complexed to each PEI polymer at an N/P ratio of 7.5, or 5% glucose only. Quantification of p18-specific CD8+ T cells followed the protocol previously described.
Cytokine production by CD8+ T cells
On day 7 postinoculation of C57BL/6J mice with PEI-DNA complexes or DNA alone as abovementioned, splenocytes were isolated and stained with H-Kb/SIINFEKL tetramer. For CD8+ T-cell stimulation, SIINFEKL peptide, anti-CD28, and anti-CD49d (BD) were added. Poststimulation, cells were stained with H-Kb/SIINFEKL tetramer and anti-mouse CD8α (Abcam). Cells were fixed, permeabilized, stained with anti-mouse -IFN-γ, -TNF-α, and -IL-2 antibodies (BD) and analyzed.
Splenic lymphocytes were harvested from C57BL/6J mice on day 7 postinoculation with PEI-DNA complexes. Cells were incubated with SIINFEKL peptide for 7 days. On day 2 of incubation, rat IL-2 (BD) was added to cells. SIINFEKL-specific CD8+ T lymphocytes were isolated using a murine CD8+ T-cell selection kit (Miltenyi Biotec), quantified using cell staining, Guava cell counter (Millipore) and flow cytometry, and incubated with SIINFEKL-pulsed EL4 cells (ATCC) at different SIINFEKL-specific effector-to-target cell ratios. Cell lysis was quantified using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega).
Cytokine production by CD4+ T cells
Cytokine production by CD4+ T cells was determined in a similar fashion as the determination of cytokine production by CD8+ T cells except: (i) splenocytes were not stained with H-Kb/SIINFEKL tetramer before and after stimulation; (ii) for stimulation of epitope-specific CD4+ T cells, ISQAVHAAHAEINEAGR peptide was used; (iii) poststimulation, cells were stained with anti-mouse CD4 (eBioscience); and lastly, (iv) after fixing and permeabilizing, cells were stained with anti-mouse–IFN-γ, -TNF–α, -IL-2, -IL-4, and -IL-17 antibodies (BD and eBiosciences).
Enzyme-linked immunosorbent assay (ELISA)
C57BL/6J mice were injected intravenously with 20 μg of pVRC-DNAOVA alone, PEI polymers only, DNA complexed to each PEI polymer at an N/P ratio of 7.5 or 5% glucose only. Mice were retro-orbitally bled and the serum was isolated. ELISA plates (NUNC) were coated with ovalbumin (Sigma Aldrich) and blocked with 10% FBS/PBS. Anti-ovalbumin IgG antibody (GenScript) and serum samples were incubated in plates. Postincubation, horseradish peroxidase (HRP)-conjugated anti-mouse IgG secondary antibody solution (Southern Biotech) was added. Next, plates were washed after incubation and SureBlue Reserve TMB peroxidase substrate solution (KPL) was placed in the wells, and plates were incubated for 15 min are room temperature. TMB stop solution (KPL) was added, and the absorbance at 450 nm was measured.
rLM-OVA infection and CFU determination
Recombinant Listeria monocytogenes that secretes chicken ovalbumin and contains an erythromycin-resistance marker (rLM-OVA) was kindly provided by H. Shen (University of Pennsylvania School of Medicine). Frozen stocks of rLM-OVA were grown in brain-heart infusion (BHI, BD) supplemented with erythromycin (Sigma Aldrich), to mid-log phase as measured by optical density (A600). Mice were inoculated with approximately 105 CFUs for CD8+ T-cell memory expansion and with approximately 106 CFUs for lethal challenge by lateral tail-vein injection. To determine bacterial burden in spleens of infected mice, whole spleens were homogenized in 0.1% Tergitol (Sigma Aldrich)-PBS. Tenfold serial dilutions of spleen suspensions were spread on BHI plates. Plates were incubated and total CFUs per spleen were determined.
Data were expressed as the mean with SE. Statistical analyses were performed using an unpaired two-tailed exact Mann–Whitney test. A P value less than 0.05 was considered significant.
We would like to thank Thomas Rogers, Maytal Bivas-Benita, Ralf Geiben-Lynn, Michelle Lifton, and Christa Osuna-Gutierrez for their scientific contribution to this study. We are also grateful to Dr. Shen for kindly providing us with the recombinant Listeria monocytogenes. This work was supported by the NIAID Center for HIV/AIDS Vaccine Immunology grant AI-067854.
Conflict of interest
The authors declare no financial or commercial conflict of interest.