INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

DNA vaccines have been shown to induce immune responses and protective immunity in many animal models of infectious disease (for a review, see reference 11). In mice, such responses can often be achieved with low doses (<1 µg) of naked DNA. However, the immunogenicity of DNA vaccines in larger animals (e.g., guinea pigs, rabbits, and nonhuman primates) has been much lower than that observed in mice, even at higher doses of DNA. In human clinical trials, certain DNA vaccines have been shown to induce immune responses (5, 19, 30), but multiple immunizations of high doses of DNA were required. Therefore, in order to provide protective efficacy in humans, the potency of DNA vaccines needs to be increased. So far, it appears that DNA vaccines are more effective at priming T-cell responses than antibodies, as exemplified by induction of cytotoxic T lymphocytes (CTL) but no antibodies against malaria circumsporozoite protein in humans (30). Similarly, we show here that in mice, human immunodeficiency virus (HIV) gag DNA primed CD8⁺ T-cell responses at doses of DNA 10- to 100-fold lower than that required for priming of antibody responses. Therefore, technologies aimed at increasing the potency of DNA vaccines need to be especially effective at boosting humoral responses.

The technology described herein, formulation of DNA onto cationic poly(lactide-coglycolide) (PLG) microparticles, has been developed as a means to better target DNA to antigen-presenting cells (APCs). PLG microparticles are an attractive approach for vaccine delivery, since the polymer is biodegradable and biocompatible and has been used to develop several drug delivery systems (21). In addition, PLG microparticles have also been used for a number of years as delivery systems for entrapped vaccine antigens (24). More recently, PLG microparticles have been described as a delivery system for entrapped DNA vaccines (15, 18). Nevertheless, recent observations have shown that DNA is damaged during microencapsulation, leading to a significant reduction in supercoiled DNA (2, 29). Moreover, the encapsulation efficiency is often low. Therefore, we developed a novel approach of adsorbing DNA onto the surface of PLG microparticles to avoid the problems associated with microencapsulation of DNA. This approach, involving adsorption of DNA to cationic PLG, was previously demonstrated to markedly increase the potency of DNA vaccines in mice (25) and was shown here to increase potency in larger animal species, i.e., guinea pigs and rhesus macaques, as well. Particularly striking was the increase in antibody titers: the levels induced by PLG/DNA were equal to or better than with recombinant protein given together with a potent adjuvant. Therefore, this approach facilitates both cellular and humoral immune responses and holds promise as an enabling technology to allow the successful use of DNA vaccines in humans.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

DNA plasmids. The plasmid encoding HIV gag driven by the cytomegalovirus (CMV) promoter (pCMV HIV gag) was grown in Escherichia coli strain HB101, purified using a Qiagen Endofree Plasmid Giga kit (Qiagen, Inc.), and resuspended in 0.9% sodium chloride (Abbott Laboratories, North Chicago, Ill.). The pCMV vector used contains the immediate-early enhancer/promoter of CMV and a bovine growth hormone terminator and is described in detail elsewhere (6). The HIV gag DNA vaccine (pCMV HIV gag) contains a synthetically constructed p55^gag gene, with codons reflecting mammalian usage, derived from the HIV-1 SF2 strain as previously described (35). The HIV env DNA vaccines (pCMV HIV gp120 and gp140) consist of a human tissue plasminogen activator (tPA) signal sequence and the following gene inserts: gp120 from HIV-1 SF2 strain and gp140 from HIV-1 US4 strain, codon optimized for expression in mammalian cells (33). Expression of HIV Gag and Env proteins was determined as described previously (33).

Preparation of PLG microparticles. The PLG polymer (RG505) was obtained from Boehringer Ingelheim. Cationic microparticles were prepared using a modified solvent evaporation process. Briefly, the microparticles were prepared by emulsifying 10 ml of a 5% (wt/vol) polymer solution in methylene chloride with 1 ml of phosphate-buffered saline (PBS) at high speed using an IKA homogenizer. The primary emulsion was then added to 50 ml of distilled water containing cetyltrimethylammonium bromide (CTAB) (0.5% wt/vol), resulting in the formation of a water-in-oil-in-water emulsion, which was stirred at 6,000 rpm for 12 h at room temperature, allowing the methylene chloride to evaporate. The resulting microparticles were washed twice in distilled water by centrifugation at 10,000 × g and freeze-dried. DNA was adsorbed onto the microparticles by incubating 100 mg of cationic microparticles in a 1-mg/ml solution of DNA at 4°C for 6 h. The microparticles were then separated by centrifugation, the pellet was washed with TE (Tris-EDTA) buffer, and the microparticles were freeze-dried. Physical characteristics were monitored as previously described (25).

Animals. Female CB6 F₁ mice (Jackson Labs) and female guinea pigs (Elm Hill Laboratories) were housed at Chiron in an American Association for Accreditation of Laboratory Animal Care-accredited facility. Male and female rhesus macaques were housed at Southern Research Institute (Frederick, Md.).

Measurement of antibody responses. At various times following immunization, blood was collected from anesthetized animals and serum was recovered by centrifugation. Anti-HIV Gag antibodies were measured by enzyme-linked immunosorbent assay (ELISA) as follows. Wells of Immulon 2 HB U-bottomed microtiter plates (Dynex Technologies, Chantilly, Va.) were coated with HIV p55 protein at 5 µg/ml in PBS, 50 µl per well, and incubated at 4°C overnight. The plates were washed six times with wash buffer (PBS, 0.1% Tween 20 [Sigma, St. Louis, Mo.]) and blocked at 37°C for 1 h with 150 µl of blocking buffer (PBS, 0.1% Tween 20 [Sigma], 1% goat serum)/well. Test sera were diluted 1:25 and then serially diluted threefold in blocking buffer. The block solution was aspirated, and then the plates were incubated at 37°C for 2 h with 50 µl of each serum dilution/well. After being washed six times, the plates were incubated for 1 h at 37°C with appropriate horseradish peroxidase-conjugated antibodies (1:20,000 dilution) in blocking buffer for 30 min. Following a final six washes, the plates were developed with OPD for 30 min. The OPD developer consists of one tablet (10 mg) of o-phenylenediamine, 12 ml of buffer (0.1 M citric acid, 0.1 M dibasic sodium phosphate), and 5 µl of 30% H₂O₂. The reaction was stopped with 50 µl of 4 N H₂SO₄ per well, and optical density (OD) was measured at dual wavelengths, 492 and 690 nm. The reported titers correspond to the reciprocal of the serum dilution producing an absorbance value of 1.0.

Measurement of T-cell responses in mice. A recombinant vaccinia virus encoding the HIV-1_SF2 gag-pol genes (rVVgag-pol) has been described previously (10). Four or more weeks following gag DNA immunization, mice were challenged with an intraperitoneal injection of 10⁷ PFU of rVVgag-pol. Five days later, spleens were harvested and stimulated with the H-2K^d-restricted p7g Gag peptide (10) and then stained for intracellular gamma interferon (IFN-), as follows. Erythrocyte-depleted single-cell suspensions were prepared by treatment with Tris-buffered NH₄Cl (Sigma), and 1 × 10⁶ to 2 × 10⁶ nucleated spleen cells were cultured in duplicate at 37°C in the presence or absence of 10-µg/ml p7g peptide. Monensin (Pharmingen, San Diego, Calif.) was added to block cytokine secretion. After 3 to 5 h, cells were washed, incubated with anti-CD16/32 (Pharmingen) to block Fc receptors, stained with fluorescein isothiocyanate (FITC)-conjugated CD8 monoclonal antibody (Pharmingen), and fixed overnight at 4°C in 2% (wt/vol) paraformaldehyde. The following day, cells were treated with 0.5% (wt/vol) saponin (Sigma) and then incubated with phycoerythrin (PE)-conjugated mouse IFN- monoclonal antibody (Pharmingen) in the presence of 0.1% (wt/vol) saponin, washed, and analyzed using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, Calif.).

Measurement of T-cell responses in macaques. A set of 51 Gag peptides 20 residues long, overlapping by 10 amino acids (aa) and spanning residues 1 to 496 of HIV-1_SF2 p55^gag, was synthesized (Chiron Mimotopes, Clayton, Australia). Peptides were used as a pool.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Priming of antibody and T-cell responses in mice by formulated DNA vaccines. To assess the relative ability of a promising DNA vaccine technology to prime antibody and T-cell responses, mice were immunized with HIV gag DNA with and without formulation onto cationic PLG microparticles. To quantify T-cell responses, mice were challenged at 28 days after a single immunization with a recombinant vaccinia virus expressing Gag. Then 5 days later, Gag-specific CD4⁺ and CD8⁺ T cells were measured in vitro by determining IFN- production in spleen cells in response to brief restimulation with an H-2^d-restricted CTL epitope (10), as measured by flow cytometry. In an extensive dose-response titration, naked DNA primed Gag-specific CD8⁺ T cells after a single dose of DNA as low as 100 ng (Table 1). In contrast, PLG/DNA was effective at 1 ng, indicating a ~100-fold increase in DNA vaccine potency, as judged by reduction of DNA vaccine dose. Anti-Gag antibodies were measured in serum before and after recombinant vaccinia virus challenge. In mice immunized with naked DNA, measurable anti-Gag antibodies were observed in unchallenged mice only at the highest DNA dose of 10 µg, whereas PLG/DNA induced antibodies at a dose as low as 100 ng. To determine the level of antibody priming by Gag DNA, anti-Gag antibodies were also measured at 5 days after recombinant vaccinia virus challenge. In naive challenged mice, no measurable antibody (or T-cell) responses were seen at this time. In challenged mice previously primed with naked DNA, substantial anti-Gag antibodies were seen only in the high-dose group, with only one of five mice showing a measurable titer at the 1-µg DNA dose. However, significant antibody priming was observed in mice immunized with as little as 1 ng of PLG/DNA (lowest dose tested). Therefore, the PLG/DNA formulation increased antibody priming by 1,000- to 10,000-fold. These data also illustrate that naked DNA primed CD8⁺ T-cell responses much more effectively than antibody responses (~100-fold, based on the lowest dose of DNA required to prime a measurable response). However, formulation of DNA onto PLG microparticles substantially enhanced immune responses, with a particularly marked effect on humoral responses, such that T-cell and antibody priming appeared equally potent. In a separate study, PLG/DNA was similarly potent, priming CD8⁺ and CD4⁺ T-cell responses at 1- and 10-ng doses of gag DNA given a single time (Table 2).

Potency of DNA vaccines in guinea pigs and rhesus macaques. To test the efficacy of PLG formulation of DNA in a larger animal species, HIV DNA vaccines were evaluated in guinea pigs and rhesus macaques. First, the potency of DNA encoding HIV gp120 derived from the HIV-1 SF2 strain was assessed at 25- and 250-µg DNA doses in guinea pigs and compared to that induced by a very immunogenic recombinant gp120 protein (SF2) administered with the powerful adjuvant MF-59. Two immunizations of naked DNA at the 25-µg dose induced very modest levels of antibody responses, as measured by ELISA (geometric mean titer [GMT] = ~100) (Fig. 1). In contrast, substantially higher levels of response were observed in animals that received PLG/DNA (GMT = ~1,000) (P < 0.0001). Further enhancement was achieved by formulation of PLG/DNA with aluminum phosphate (AlPO₄) (GMT = ~4,000) (P < 0.001). This combination of PLG/DNA/AlPO₄ induced levels of anti-Env antibodies equivalent to that induced by recombinant protein with MF-59 (P = 0.881). The levels of neutralizing antibody titers correlated well with ELISA, as very similar results were seen with naked DNA < PLG/DNA < PLG/DNA/AlPO₄  rgp120/MF-59 (P = 0.003, 0.09, and 0.266, respectively) (Table 3). In general, higher levels of antibody responses (both ELISA and neutralizing titers) were induced by the 250-µg dose of DNA than by 25 µg. However, formulation of 25 µg of DNA (with either PLG or PLG/AlPO₄) increased potency to levels equivalent to that induced by 250 µg of naked DNA. This was true for both ELISA and neutralizing antibody titers. Therefore, by two separate criteria, increase in titer and decrease in DNA dose, PLG formulation of DNA increased HIV gp120 DNA vaccine potency in guinea pigs by ~10-fold.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Groups of 10 guinea pigs were immunized with HIV gp120 DNA in saline, PLG, or PLG plus AlPO4 at doses of 25 or 250 µg of DNA. As positive controls, animals were immunized with 50 µg of recombinant gp120 protein with or without MF59 adjuvant. All animals were immunized at 0 and 4 weeks, and sera were collected at 2 weeks post-second immunization. Data are presented as geometric mean ELISA titer ± standard error of the mean (SEM) for n = 10.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Groups of 10 guinea pigs were immunized with HIV gp140 DNA in saline or PLG at a dose of 50 µg of DNA. As positive controls, animals were immunized with 25 µg of recombinant gp120 protein with MF59 adjuvant. All animals were immunized at 0, 4, and 8 weeks, and sera were collected at 2 weeks post-second (A) and -third (B) immunization. Some groups were immunized with DNA only (D/D/D) or protein plus MF-59 only (P/P/P) or primed with DNA and boosted with protein plus MF-59 (D/P or D/D/P). Data are presented as geometric mean ELISA titer ± SEM for n = 10.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Groups of five guinea pigs were immunized with a combination of HIV gp140 DNA and HIV gag DNA in saline or PLG at doses of 1 and 0.5 mg of DNA, respectively. Animals were immunized at 0, 4, and 8 weeks, and sera were collected at 3, 7, 11, and 32 weeks. Data are presented as geometric mean ELISA titer ± SEM (n = 5) for anti-Gag (A) and anti-Env (B) antibodies.

View larger version (16K): [in this window] [in a new window]   FIG. 4.   Groups of five rhesus macaques were immunized with HIV gag DNA in saline (solid triangles) or PLG (open circles) at a DNA dose of 0.5 mg. Animals were immunized at 0 and 4 weeks, and sera were collected at 2, 6, and 11 weeks. Data are presented as geometric mean ELISA titer ± SEM for n = 5. For comparison, anti-Gag antibody titers are shown for unimmunized animals (solid squares) (n = 4).

View larger version (11K): [in this window] [in a new window]   FIG. 5.   Gag-specific cytolytic activity of PBMC from individual rhesus macaques given two doses of PLG/CTAB-formulated pCMV-gag DNA (A) or unformulated pCMV-gag DNA (B). Two weeks after the second DNA immunization, PBMC were stimulated with a pool of overlapping Gag peptides and cultured in the presence of IL-2 for 8 days. Serial dilutions of cultures (1:15, 1:45, 1:135, and 1:405) were added to 51Cr-labeled rVVgagpolSF2-infected autologous B-LCL () or rVVgp160envSF162-infected autologous B-LCL (). 51Cr release was determined 4 h after addition of cultured PBMC to B-LCL.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

DNA vaccines were initially investigated because of their potential to induce CTL, as a consequence of antigen expression within cells of the vaccinated animals. In fact, this technology has been demonstrated to effectively induce CTL in mice (28), monkeys (31), and humans (30). Antibody and, to a lesser extent, helper T-cell responses, however, have proven more difficult to elicit by DNA vaccines. This relative deficiency is particularly true for larger animals, such as primates, but is also apparent in mice. For example, Table 1 demonstrates that CTL are primed by naked DNA at much lower doses (10 to 100 ng) than are antibodies (1 to 10 µg). The reasons for low immunogenicity in larger animals are unclear but may be related to relative body mass (32), poor delivery of DNA to APCs, and/or less efficient distribution of DNA within the tissue. In support of the latter hypothesis, we observed that administration to mice of potent DNA vaccines in low volume (i.e., 5 µl) limited the distribution of DNA and uptake by cells within injected tissue, resulting in very poor immunogenicity (12). Moreover, increased distribution and delivery of DNA within the tissue through the use of electroporation (12, 33) overcame this limitation in potency of DNA given in low volume. Therefore, the low volume of inoculum relative to muscle mass, which is particularly true in larger animals, may be an important factor in limiting DNA delivery and immunogenicity.

A second possible limitation in larger animals is poor DNA uptake by and transfection of APCs. In small animals, such as mice, intramuscular injection of DNA results primarily in the transfection of myocytes (34). However, APCs may also be transfected, albeit at a very low frequency (1, 4, 7). It is well known that expression of antigens in APCs by DNA vaccines is a potent means of priming immune responses (20, 26), including those in nonhuman primates (3). Hence, targeting DNA to APCs is likely to increase DNA vaccine potency. To this end, we have developed a microparticle-based PLG formulation of DNA (25), based on the hypothesis that particles ~1 µm in diameter would be readily internalized by phagocytic cells, such as immature dendritic cells (16), thereby facilitating nonspecific targeting. Indeed, this formulation is very potent in mice (25) and appears to function, at least in part, by facilitating DNA uptake by APCs (9). PLG has been used previously for delivery of small-molecule drugs (17), proteins (21, 22), and DNA (8, 15, 18, 27). However, in these cases the delivered cargo was encapsulated inside the particles and, hence, may have functioned in a slow-release depot manner. In contrast, the present formulation, consisting of surface-adsorbed DNA, has a twofold rationale: (i) adsorption onto preformed PLG particles avoids the harsh emulsion conditions, which are known to affect the integrity of DNA (2, 27, 29); and (ii) it provides a means of delivery to and rapid release of DNA in APCs.

Previously, we have shown a marked enhancement of DNA vaccine potency using PLG in mice, based on antibody and CTL responses (25). We have now extended those observations in mice to include the enhancement of CD8⁺ and CD4⁺ T-cell responses, based on a quantitative assay measuring intracellular IFN- production, across a wide dose range of DNA (1 ng to 10 µg). In terms of dose reduction, PLG/DNA was 100- to 1,000-fold more potent than naked DNA for CD8⁺ T-cell and antibody responses, respectively. Similarly, in much more limited dose-range studies, PLG/DNA was also able to reduce the dose of DNA in guinea pigs (at least 10-fold). More importantly, though, PLG/DNA was able to elicit antibody responses to levels equivalent to those induced by recombinant HIV gp120 administered in a strong adjuvant, which is currently the best means short of live virus infection of inducing anti-Env antibody responses. This was true for both ELISA and neutralization assays. The ability of a DNA vaccine technology to reduce the dose of DNA is likely to be an important parameter. First, at high doses of DNA in mice (~10 µg), an apparent ceiling of T-cell responsiveness was achieved, with ~20% of CD8⁺ T cells responding to Gag. At these doses, naked and formulated DNA showed equivalent responses, and further enhancement is unlikely. Therefore, enhancement at low doses may be a more relevant indicator of potency. Second, to date CTL responses in humans have only been reliably seen with high doses of DNA (>1 mg), and antibody responses are not induced (30). Hence, for reasons of potency, ability to manufacture, cost of goods, and safety, it is imperative that the DNA dose be reduced.

During the course of these studies, an apparent interference between the two components of a combination vaccine consisting of unformulated HIV gag and env DNA was observed. Specifically, anti-Env antibody responses were considerably lower in guinea pigs given the combination vaccine containing 1 mg of naked env DNA than in those animals given as little as 50 µg of naked env DNA alone (compare Fig. 2B and 3B). This hyporesponsiveness to the env component may be related to antigenic competition between gag and env or to differential uptake of the DNA plasmids by cells in vivo. Whatever the cause, this interference was not observed in other studies with mice or rabbits (C. Goldbeck, M. Selby, and J. Ulmer, unpublished observations) and, hence, appears to be specific to this animal species. Moreover, formulation of the DNA components with PLG overcame this hyporesponsiveness to the env component. Therefore, the use of PLG microparticles for DNA delivery may facilitate the development of combination vaccines.

The ability of vaccines to induce potent immune responses in small animals does not ensure that the vaccines will be effective in primates. This has been shown to be true for several naked DNA vaccines. We show here that naked HIV gag DNA did not induce significant anti-Gag antibodies in rhesus macaques unless it was formulated on PLG microparticles. In summary, we have demonstrated markedly increased potency of HIV gag and/or env DNA vaccines in mice, guinea pigs, and rhesus macaques by formulation with PLG. Based on dose reduction and the ability to enhance both cellular and humoral immune responses (particularly the latter), this technology holds promise for use in humans.