Electroporation as a “Prime/Boost” Strategy for Naked DNA Vaccination against a Tumor Antigen¹ Free

Deoxyribonucleic acid vaccination is emerging as an effective and safe strategy for inducing protective immunity in preclinical models of infectious disease (1, 2, 3, 4), cancer (5, 6, 7, 8), and autoimmunity (9, 10, 11). It is evident that DNA vaccines have the ability to stimulate a broad spectrum of immunological activities (12). In mice, they are capable of inducing potent cell-mediated and humoral immunity, although they are typically weaker at promoting Ab responses than protein-based vaccines (3, 13). However, transferring this technology into large animals or human subjects has generally produced only modest results (14, 15).

Nevertheless, DNA vaccines are effective at priming immune responses in humans and large animals, a quality that can be exploited using the heterologous prime/boost approach, whereby the initial immune response to naked DNA vaccination is boosted by delivery of the same Ag in a different vaccine vehicle (e.g., via viral or bacterial vectors, or as protein) (16, 17, 18). However, despite a large body of evidence in animal models and in the clinic (19) demonstrating the efficacy of heterologous prime/boost procedures, this approach has its limitations (20). Additional vectors or proteins raise regulatory and manufacturing issues. Importantly, pre-existing or induced blocking immunity against the viral or bacterial vector is a major concern, especially for patients with cancer, in which it is likely that the vaccination program will be prolonged.

We have developed DNA fusion vaccines, encoding tumor Ags linked to pathogen-derived sequences aimed to provide CD4⁺ T cell help critical for the induction and maintenance of antitumor immunity (20, 21). By using the fragment C sequence of tetanus toxin, we can activate robust tumor-specific Ab, CD4⁺, and CD8⁺ T cell responses and protect mice from tumor (13, 22, 23, 24). Because vaccine dose and volume, known to be critical for responses in mice (25, 26, 27), are difficult to scale up for human subjects, other delivery strategies are required. Important factors are the level of Ag expression and activation of innate immunity (12, 28). Numerous techniques are being developed to increase efficiency (28), with electroporation being particularly attractive, as it has been shown to increase DNA uptake and protein expression in various tissues in vivo (29, 30, 31). Improvement in vaccine potency has been observed in small and large animal models of infectious disease (26, 27, 32, 33).

We report in this work the application of electroporation to two models, the CT26 carcinoma and the BCL₁ lymphoma, susceptible to attack via either CD8⁺ T cells or Ab, respectively (23, 34, 35). We demonstrate an increase in priming by electroporation in both. Importantly, we show that a prime/boost approach with naked DNA, followed by DNA plus electroporation, amplifies both effector functions. Thus, the induction of effective tumor-specific immunity may now be feasible in cancer patients using only a single naked DNA vaccine format.

The murine CT26 colon carcinoma cell line and a cell line derived from the B cell lymphoma BCL₁ (36) were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated FCS (Invitrogen Life Technologies), 1 mM sodium pyruvate, 2 mM l-glutamine, nonessential amino acids (1% of 100× stock), 25 mM HEPES buffer, and 50 μM 2-ME (hereafter referred to as complete medium). CT26 cells were harvested by incubation with Ca/Mg-free medium, as previously described (23).

The H-2L^d-restricted gp70 epitope (AH1) has been described previously (35). The peptide (SPSYVYHQF) was synthesized commercially and supplied at >95% purity (Peptide Protein Research). Peptide stocks (2 mM) were dissolved in PBS, filter sterilized, and stored at −20°C.

Construction of the DNA fusion vaccine p.DOM-AH1 has been described (23). It encodes the first domain of fragment C (FrC)³ from tetanus toxin (DOM; TT_865–1120) with sequence encoding the AH1 CTL epitope fused to the 3′ terminus. The p.DOM control vaccine encodes the first domain of FrC alone.

p.BCL₁, encoding the idiotypic V_L and V_H regions (single chain Fv (scFv)) derived from the murine B cell lymphoma, BCL₁, fused to human C_H3 from IgG1, has been described previously (37). This was used as a template to construct p.BCL₁-FrC (kindly supplied by D. Zhu University of Southampton, Southhampton, U.K.), which encodes BCL₁ scFv upstream of sequence-encoding FrC. Briefly, C_H3-encoding sequence was cut from p.BCL₁ using BspEI and NotI. FrC sequence (8) was amplified using the primers 5′-TATTCCGGAGGACCCGGACCTATGAAA-3′ (forward) and 5′-TAATGCGGCCGCTTAGTCGTTGGTCCAACCTTC-3′ (reverse), each of which introduced either a BspEI site or NotI site, respectively, to the FrC termini. The resulting PCR product was gel purified, digested, and cloned into p.BCL₁ in the place of C_H3, creating p.BCL₁-FrC, which encodes signal peptide-V_L-linker peptide-V_H-linker peptide-FrC.

Each DNA vaccine encoded the signal sequence derived from the V_H of the IgM of the BCL₁ tumor, and was incorporated into the pcDNA3 vector backbone (Invitrogen Life Technologies). Vaccine integrity was confirmed by DNA sequencing, while expression and product size were checked in vitro using the TNT T7 Coupled Reticulocyte Lysate System (Promega).

BALB/c (H-2^d) mice were vaccinated at 6–12 wk of age by injection of DNA, in 0.9% saline (w/v), into the quadriceps muscle of each hind limb. Injection volume per leg and total DNA dose are indicated, but ranged from 10 to 50 μl/leg and 5 to 100 μg of DNA/leg (10–200 μg dose). A Hamilton Microliter syringe (Scientific Laboratory Supplies) was used to administer injection volumes smaller than 50 μl. All injections were administered using a 26G needle. Animal welfare and experimentation were conducted in accordance with local Ethical Committee and United Kingdom Coordinating Committee for Cancer Research guidelines, under Home Office license.

Mice were anesthetized before electroporation using 1 part midazolam (5 mg/ml), 1 part hypnorm (fentanyl citrate (0.315 mg/ml) and fuanisone (10 mg/ml)), and 2 parts water. The mice received 7 μl/g body weight by i.p. injection. The skin overlying the quadriceps muscle was shaved, and DNA vaccine was administered using the indicated dose and volume. Following the application of a conductance gel, silver electrodes were placed on the skin on either side of the injection site and a local electrical field was immediately applied using a custom-made pulse generator, Elgen (Inovio), as previously described (31). The electrical field comprised 10 trains of 1000 square wave pulses delivered at a frequency of 1000 Hz, with each pulse lasting a total of 400 μs (200 μs positive and 200 μs negative). The electrical field strength varied with the resistance in the tissue of each animal and was ∼50 V over 3–4 mm. Each train was delivered at 1-s intervals; the electrical pulse was kept constant at ± 50 mAmp (31).

To assess priming of CD8⁺ T cells, mice were culled at day 14 following DNA vaccination (using the dose and volume of vaccine, as indicated), and spleens were harvested and processed for detection of intracellular IFN-γ. To monitor the potential to boost existing CD8⁺ T cell responses, mice were vaccinated at day 0 (25 μg of DNA in 50 μl of saline per rear limb) and given booster injections of vaccine at day 28, either with or without electroporation at each time point; spleens were harvested at day 36 to monitor CD8⁺ T cell responses. Viable, pooled splenocytes were selected by density centrifugation, and cells were incubated for 4 h at 37°C in 96-well plates, at 1 × 10⁶ cells/well, in complete medium together with 10 U/well human rIL-2 (PerkinElmer), 1 μM AH1 peptide, and 1 μl/well Golgi Plug. Samples were then processed to label intracellular IFN-γ, as previously described (23), before analysis by FACSCalibur using CellQuest software (BD Biosciences). Analyses were performed on lymphocyte populations with MHC class II-positive cells gated out.

Mice were culled at day 14 postvaccination, spleens were pooled, and splenocyte suspensions (3 × 10⁶ cells/ml) were prepared in complete medium, together with IL-2 (20 U/ml) and AH1 peptide (1 μM). Bulk splenocyte cultures were incubated at 37°C, 5% CO₂, for 6 days before assessing cytolytic activity in a standard 4- to 6-h ⁵¹Cr release assay, as previously described (23). Targets included BCL₁ cells, either alone or labeled with AH1 peptide. Specific lysis was calculated by the standard formula ((release by CTL − spontaneous release)/(total release − spontaneous release) × 100%). Spontaneous release was always <30%.

Mice were vaccinated with a total dose of 50 μg of DNA (25 μg per rear leg), using the indicated injection volumes administered with or without electroporation. During tumor challenge, mice were injected s.c. with 1 × 10⁵ CT26 tumor cells into the rear flank. For prophylactic immunization, mice were challenged with tumor cells 14 days after DNA vaccination, while for therapeutic immunization tumor cells were injected 1 day before DNA vaccination. All mice were monitored twice daily for tumor development and were culled when mean tumor diameter reached 15 mm, in accordance with humane end point guidelines (United Kingdom Coordinating Committee for Cancer Research).

To monitor priming of humoral immunity, mice were vaccinated i.m. with 50 μg of p.BCL₁-FrC (25 μg in 50 μl of saline per rear limb) on day 0, either with or without electroporation. Serum samples were collected on days 28 and 42 and analyzed by ELISA for the presence of IgG specific for BCL₁ Id IgM or FrC, as described previously (8, 38, 39). For the prime/boost setting, we vaccinated mice at day 0 and gave booster vaccinations at day 21, with or without electroporation at each time point, and collected serum samples at day 41 to monitor Ab responses. The injection schedule (day 0, day 21) follows the previously published protocol for Ab induction, established using Id IgM protein vaccination (34). For protein vaccinations, Id IgM from BCL₁ was coupled to FrC protein using a one-step glutaraldehyde method, as used for coupling to keyhole limpet hemocyanin (38). Mice were injected with IgM, or IgM coupled to FrC, in CFA before serum analysis, as described previously (34, 38). ELISA plates were analyzed using a Dynex MRX plate reader at 450 nM wavelength.

CTL responses were analyzed using the Mann-Whitney U test. Serum IgG titers were compared using a two-tailed t test on log normalized data. Survival curves were compared using the χ² log rank test. Experimental groups were considered significantly different from control groups if p < 0.05.

We tested the ability of the p.DOM-AH1 vaccine to induce AH1-specific CD8⁺ T cell responses when the vaccine dose was varied (Fig. 1, a and b). Using a constant injection volume of 2 × 50 μl, a dose of 10 μg was inadequate to induce a detectable CD8⁺ T cell response. Increasing the dose to 30 μg induced significant IFN-γ-producing CD8⁺ T cell responses, and this was not amplified markedly by using 50–200 μg doses (Fig. 1,a, Expt. 1). A second experiment confirmed this trend (Fig. 1,a, Expt. 2); an AH1-specific CD8⁺ T cell response was only detected when a dose of >5 μg was used, with the proportion of responding cells reaching a plateau at a dose of 30 μg of p.DOM-AH1. Responses to the control vaccine (p.DOM) were insignificant. The functional efficacy of the IFN-γ-producing CD8⁺ T cells was confirmed in a cytotoxicity assay following a 6-day expansion in vitro with AH1 peptide (Fig. 1,b). The CTL specifically lysed BCL₁ cells when pulsed with peptide (Fig. 1 b), as well as tumor cells expressing endogenous AH1 (23) (data not shown).

Maintaining a constant DNA vaccine dose of 50 μg per mouse, we then assessed the impact of injection volume on AH1-specific CD8⁺ T cell induction (Fig. 1, c and d). It was necessary to use 30 μl per limb to generate a significant response, and 40 μl for maximum response, not increased further using 50 μl (Fig. 1,c, Expt. 1). The effect of injection volume on immune outcome was confirmed in a second experiment (Fig. 1,c, Expt. 2); an injection volume of 25 μl was required to prime a significant AH1-specific CD8⁺ T cell response, with 50 μl producing an increased response. Responses to the control vaccine (p.DOM) were insignificant. Again, functional activity was confirmed by cytotoxicity assay following a 6-day expansion in vitro (Fig. 1 d).

We next assessed the effects of combining DNA vaccine delivery with electroporation. Mice were vaccinated with 50 μg of p.DOM-AH1 using injection volumes that were either optimal (2 × 50 μl) or suboptimal (2 × 25 μl and 2 × 10 μl) for priming of AH1-specific CD8⁺ T cells (Fig. 1,c), with electroporation of the injection sites. Results (Fig. 2) indicate that an injection volume of 50 μl led to effective priming of AH1-specific CD8⁺ T cells, with decreasing responses as volumes were reduced (Fig. 2). Electroporation did not influence the performance of the 50 μl delivery (Fig. 2,a). Three experiments were conducted to confirm this point, because experiment 1 showed a decrease in response when using electroporation. However, the additional two experiments showed no change, and data compiled from the three identical experiments indicated no effect of electroporation using this optimized volume (Fig. 2,a). In contrast, electroporation did significantly improve CD8⁺ T cell responses when using suboptimal injection volumes of 25 and 10 μl per limb, p = 0.021 and p = 0.01, respectively, and the trend was evident in each of the two experiments (Fig. 2, b and c).

We have previously demonstrated that following vaccination with p.DOM-AH1, the induced CD8⁺ CTL of single epitope specificity can protect against tumor (23). Results (Fig. 3) confirm this using our standard injection volume of 50 μl/leg, with tumor challenge 14 days later. Delivery in a suboptimal volume (2 × 10 μl) did not mediate protection. However, protective efficacy was completely restored when suboptimal volume was combined with electroporation (p < 0.003), demonstrating a clear correlation between the ability to induce cytolytic T cells by vaccination and protection from CT26 tumor in vivo.

To assess therapeutic efficacy, we investigated the effects of DNA vaccination 1 day after tumor injection. Again, mice received a total dose of 50 μg of DNA, but the vaccine injection volume was varied. Results (Fig. 4) indicate that vaccination with p.DOM-AH1 using our standard injection volume (2 × 50 μl) activates protective immunity in that setting compared with nonvaccinated mice or those given the control vaccine (p.DOM). Delivery in a suboptimal injection volume (2 × 10 μl) was ineffective, but protection could be fully restored by combination with electroporation (p = 0.03).

The effects of electroporation on priming were clear only when using suboptimal vaccination conditions. We then investigated whether electroporation could improve performance of optimal delivery when combined with boosting. Electroporation was given either at priming alone, at boosting alone, or at both time points. Boosts were given at day 28, and the levels of AH1-specific IFN-γ-producing CD8⁺ T cells were measured ex vivo 8 days later (day 36) (Fig. 5), because previous data using this vaccine design indicated that peak CD8⁺ T cell responses occur 7–10 days postboost (40).

At this later time point after only the first injection, the proportions of detectable AH1-specific CD8⁺ T cells observed in mice primed with p.DOM-AH1 at day 0 only (without electroporation) were low (mean 0.76%), even undetectable in five mice, probably due to the natural kinetics of the CD8⁺ T cell response. Booster injections at day 28 (without electroporation) generated a significant increase in the proportion of AH1-specific IFN-γ-positive CD8⁺ T cells, enabling them to be detected in all mice (mean 1.5%, p = 0.0066). However, the application of electroporation at the time of boosting amplified this response, generating high levels of IFN-γ-positive CD8⁺ T cells (mean 3.8%, p = 0.014) (Fig. 5). Electroporation at both the priming and boosting stages was also effective in amplifying the response to naked DNA (mean 2.7%, p = 0.0017). There was a trend for double electroporation to be less effective than priming with DNA alone plus boosting with electroporation, but the difference was not statistically significant (p = 0.24). Interestingly, reversing the prime/boost strategy (priming with DNA plus electroporation and boosting with DNA alone) did not enhance the number of responding T cells compared with DNA injection alone (data not shown), indicating that the correct sequence of vaccination and electroporation is critical for boosting the CD8⁺ T cell response.

To measure the effect of electroporation on induction of Ab, we used the DNA fusion vaccine containing the V regions of the BCL₁ lymphoma linked as scFv to full-length FrC (p.scFv-FrC) (8, 13). This vaccine is known to induce significant levels of Ab against both tumor-derived idiotypic Ig and FrC components of the fusion gene, and this is confirmed in Fig. 6. A single injection of the p.scFv-FrC vaccine induced detectable anti-Id Ab at day 28, which had increased only slightly by day 42. Electroporation at priming increased anti-Id Ab at both time points. Similar effects were evident in the anti-FrC Ab responses (Fig. 6 b).

We then tested the effects of electroporation used at the stage of priming and/or boosting (day 21) on Ab responses to both Id and FrC measured at day 41.

Priming and boosting with DNA alone induced significant levels of anti-Id IgG, detectable in all vaccinated mice (Fig. 7,a). However, priming with DNA alone and boosting with DNA plus electroporation led to a striking amplification of the anti-Id response (p < 0.0001). Electroporation at both time points is superior to no electroporation (p = 0.0013), but electroporation at the time of boosting only is clearly the most effective combination (p = 0.0025) (Fig. 7,a). Interestingly, reversal of this prime/boost strategy (electroporation at the time of priming only) did not improve the anti-Id response by day 41, compared with DNA alone (p = 0.40). Again, similar data were obtained for the anti-FrC response (Fig. 7,b). The anti-Id levels achieved by adding electroporation to boosting are ∼7-fold higher than those achieved using DNA alone (Fig. 7 a). These levels are similar to those induced by the previous gold standard of idiotypic protein plus CFA (34) and are comparable with the ∼11-fold increase over DNA scFv-FrC observed by using Id-FrC fusion protein in CFA (data not shown).

There are two major problems in developing DNA vaccination as a treatment for cancer. The first is the poor immunogenicity of most candidate tumor Ags. There are many strategies aimed to increase this (20), and we have chosen to use fusion genes that encode tumor Ags in combination with immunogenic pathogen-derived sequences, mainly derived from tetanus toxin (20). Different designs have been optimized to induce effector pathways for precision attack on tumor targets (21). Currently, we are testing these in clinical trials, with early evidence for immune responses.

The second problem, relevant for all DNA vaccines, relates to the translation of promising data in animal models to human subjects. Although safety does not appear to be an issue, the efficacy in humans has been disappointing (41, 42, 43), partly due to difficulties in scaling up DNA vaccine dose and injection volume for human application (21). Cellular uptake of DNA appears to be a significant limiting factor on transfection in vivo, and low vaccine dose results in poor Ag expression and reduced immunogenicity (27). Similarly, injection volume can influence Ag expression and immunogenicity in vivo (26). Hydrostatic pressure created by a relatively large injection volume into a small muscle may distend the extracellular space between muscle cells and facilitate the transfer of macromolecules across the plasma membrane (26). This effect will be reduced in large animals and humans, because the ratio of injection volume to muscle mass is far lower (26). In vivo electroporation can increase DNA uptake by muscle cells and mononuclear cells at the site of injection (27, 44), leading to increased Ag expression (29, 30, 31). Dendritic cells at the draining lymph nodes have been shown to contain DNA originating from the injection site (26), and electroporation might also contribute an undefined adjuvant effect, possibly mediated through local tissue damage and release of inflammatory factors (44, 45, 46).

Our murine data confirm that induction of antitumor CTL by DNA fusion vaccines is dependent on dose and volume of injection (26, 27). It is clearly possible to achieve an optimal dose/volume in mice, and electroporation then has no additive effect on priming. However, induction of Ab appears far from optimal under the same conditions and electroporation amplifies priming significantly. This could reflect a need for higher levels of Ag for priming of Ab responses (47, 48). Electroporation therefore offers a strategy to amplify priming, which could be useful in the clinic.

However, a more striking effect of electroporation was evident in a prime/boost setting, with naked DNA at both time points. The amplification is reminiscent of that achieved by boosting with Ags delivered via viral vectors (49). These vectors are presumed both to increase protein expression and to stimulate an inflammatory response (50, 51). Their disadvantages, particularly for cancer patients, are that pre-existing or developing immunity can neutralize the delivery agent and negate continued use (52, 53, 54). A more general disadvantage is that highly immunogenic viral or bacterial vectors may introduce potentially immunodominant T cell epitopes, possibly outcompeting weakly immunogenic tumor Ags in the ensuing immune response (55, 56, 57). Efforts are being made to overcome these problems by removing viral genes (58), but success there may deplete efficacy, and two vaccine vehicles mean more safety/regulatory issues.

The mechanism by which electroporation amplifies CTL or Ab responses when administered at the stage of boosting is unclear. Increased Ag expression is likely to be important for boosting CTL, possibly by increasing the numbers of Ag-loaded APC. Our prime/boost strategy will drive increased Ag expression at the crucial stage of boosting, leading to more effective activation of vaccine-specific CD8⁺ T cells. Electroporation at both priming and boosting also enhanced CD8⁺ T cell induction, but was no more effective than using electroporation only at the boosting stage, confirming that the availability of Ag at boosting, rather than priming, is critical for CD8⁺ T cell induction. For Ab induction, in addition to a more effective induction of T cell help, more available Ag would be provided on boosting for uptake by B cells (47, 48, 59). This may explain why priming with DNA plus electroporation and boosting with DNA alone was no more effective at raising specific Ab levels than injecting DNA alone at both time points. Electroporation also leads to an inflammatory response, which is likely to recruit specific T and B cells to the injection site (44, 45, 46).

With this in mind, we delineated a homologous prime/boost strategy in which mice received the same naked DNA fusion vaccine, with electroporation only at the critical time of boosting. This turned suboptimal delivery for CTL induction into effective vaccination and should be translatable to human subjects. Electroporation devices are now acceptable for human subjects (60) and have already been tested in volunteers (61). We are beginning a clinical trial in patients shortly using the same device (61). Protocols for electrical stimulation have to balance immune outcome with patient acceptability, and further trials in large animals and patients will assist optimization. The apparently suboptimal performance of DNA vaccines in inducing Ab responses can be improved by the same prime/boost strategy. The priming qualities of DNA vaccines, together with the improved Ag expression offered by electroporation, can now be combined in a homologous prime/boost approach to generate superior immune responses. This simple modification should facilitate application to the clinic.

We thank Drs. Federica Benvenuti and Oscar Burrone (International Centre for Genetic Engineering and Biotechnology, Trieste, Italy) for the kind gift of the p.BCL₁ plasmid.

E. Grønevik and I. Mathiesen are shareholders in, and I. Mathiesen is also CEO of, Inovio, which is a wholly owned subsidiary of Genetronics. J. Rice and F. K. Stevenson have patent applications relating to the DNA Fusion Vaccine designs, which were discussed in this manuscript.