Intramuscular immunization with a naked DNA plasmid expressing the Plasmodium yoelii circumsporozoite protein (pPyCSP) protects mice against challenge with P. yoelii sporozoites. This protection can be improved either by coadministration of a plasmid expressing murine GM-CSF (pGMCSF) or by boosting with recombinant poxvirus expressing the PyCSP. We now report that combining these two strategies, by first mixing the priming dose of pPyCSP with pGMCSF and then boosting with recombinant virus, can substantially increase vaccine effectiveness. Not only were immune responses and protection improved but the pPyCSP dose could be lowered from 100 μg to 1 μg with little loss of immunogenicity after boost with recombinant poxvirus. Comparing mice primed by the 1-μg doses of pPyCSP plus 1 μg pGMCSF with mice primed by 1-μg doses of pPyCSP alone, the former were better protected (60% vs 0) and had higher concentrations of Abs (titers of 163, 840 vs 5, 120 by indirect fluorescent Ab test against sporozoites), more ex vivo CTL activity (25% vs 7% specific lysis), and more IFN-γ-secreting cells by enzyme-linked immunospot assay (1460 vs 280 IFN-γ spot-forming cells/106 cells). Priming with plasmid vaccine plus pGMCSF and boosting with recombinant poxviruses strongly improves the immunogenicity and protective efficacy of DNA vaccination and allows for significant reduction of dose.
Intramuscular injection of BALB/c mice with plasmid DNA encoding the Plasmodium yoelii circumsporozoite protein (PyCSP)4 protects against challenge with P. yoelii sporozoites (1), and this protection is dependent on CD8+ T cells, IFN-γ, and nitric oxide and is directed against the parasite-infected hepatocyte (2). Protection ranges from 25 to 75% depending on the dosage regimen of vaccine, infectivity, and dose of sporozoites used for challenge (1, 2, 3). However, this level of protection is significantly less than the >95% protection in BALB/c mice after immunization with radiation-attenuated sporozoites (4, 5), and we have continued to search for ways to improve the PyCSP plasmid vaccine. We have recently reported that coadministration of PyCSP DNA with a DNA plasmid that expresses murine GM-CSF (pGMCSF) (3) improves vaccine efficacy. In a separate paper, we described that boosting of pPyCSP DNA-primed animals with recombinant poxvirus expressing the PyCSP (rvPyCSP) (6) increases protective efficacy. Both these strategies, boosting with recombinant poxvirus (7, 8, 9, 10, 11, 12, 13) and the addition of pGMCSF (14, 15, 16, 17), have been shown to enhance immunogenicity and protection of several other DNA plasmid vaccines.
In an effort to further improve the protective efficacy of pPyCSP DNA vaccine and to reduce the dose of DNA required for immunization, we have combined these two approaches, first priming with pPyCSP mixed with pGMCSF and then boosting with rvPyCSP. The results demonstrate that this approach leads to greater Ag-specific Ab, IFN-γ, and cytotoxic T cell activity and a significant increase in protective efficacy. Furthermore, priming with a low DNA dose of 1 μg pPyCSP and 1 μg pGMCSF gives excellent protection, only slightly less than that achieved after priming with high doses of 100 μg pPyCSP and 30 μg pGMCSF.
Materials and Methods
Female BALB/cByJ (H-2d), A/J(H-2a), and C57BL/6 (H-2b)mice, 6–8 wk old, purchased from The Jackson Laboratory, Bar Harbor, ME, were used in all experiments.
P. yoelii (17XNL) clone 1.1 parasites were used. Sporozoites for challenges were obtained from infected mosquito glands in M199 medium containing 5% normal mouse serum. All challenges were accomplished by injecting 50–100 sporozoites into the tail vein.
The P. yoelii plasmid DNA construct, 1020PyCSP, designated here as pPyCSP and the murine GM-CSF plasmid, designated here as pGMCSF, used for all the immunizations have been described (3, 6). All DNA for injection was purified using cesium chloride-ethidium bromide density gradient centrifugation as previously described (6). Plasmids stored at −20°C were diluted in PBS (pH 7.2) to the appropriate concentration for injections. Priming with plasmids was done either with pPyCSP alone or mixed with pGMCSF. Plasmid DNA lacking the PyCSP gene was used as control plasmid and was added to other groups to equalize the total amount of injected plasmid. A 0.3-ml insulin syringe with a 29-gauge 0.5-inch-long needle was used for the i.m. injections in the tibialis anterior muscle of the two hind limbs.
Recombinant vaccinia virus immunization
The construction of P. yoelii recombinant vaccinia virus, designated (rvPyCSP) used for boosting has already been described (6). Mice were boosted i.p. with rvPyCSP at 3, 6, 9, or 12 wk after priming with DNA. The dose of virus was delivered in a total volume of 0.2 ml in PBS, pH 7.2.
Indirect fluorescent Ab test (IFAT).
The IFAT technique (1) was used to detect anti-P. yoelii sporozoite Abs in pooled sera from blood taken just before challenge.
Serum Ab levels against recombinant protein or synthetic peptide were determined by standard ELISA protocols as previously described (18), using a full length recombinant PyCSP protein produced in yeast (rPyCSP) or linear synthetic peptide containing four copies of the major repeat of PyCSP protein, (QGPGAP)4, as solid phase Ags. Briefly, 50 μl of 0.1 μg/ml rPyCSP or 0.1 μg/ml (QGPGAP)4 in PBS were added into wells of Immunolon II ELISA plates (Dynatech Laboratory, Chantilly, VA) and incubated for 6 h at room temperature. The wells were washed three times with PBS containing 0.05% Tween 20 (washing buffer) and incubated overnight at 4°C with 100 μl 5% nonfat dry milk in PBS (blocking buffer). After three washes with washing buffer, the wells were incubated for 2 h with 50 μl of different dilutions of sera diluted in PBS containing 3% nonfat dry milk (diluting buffer). The wells were washed three times, incubated for 1 h with peroxidase-labeled goat anti-mouse IgG (Kirkegaard and Perry, Gaithersburg, MD) diluted 1:2000 in diluting buffer, then washed again three times. The wells were incubated for 20 min with 100 μl of a solution containing ABTS substrate (2,2′-azinodi(3-ethylbenzthiazoline sulfonate) (Kirkegaard and Perry)) and H2O2. Color reaction was measured in a microELISA automated reader (Dynatech, MR5000) at OD410 nm. All reaction steps except blocking were performed at room temperature. Means ± SD of the OD readings of quadruplicate assays were recorded. The results were reported as OD 0.5 units (the reciprocal of the serum dilution at which the mean OD reading was 0.5).
Restimulation CTL assay.
51Cr release assays using bulk cultures was performed as previously described (1). In this method, spleen cells (5 × 106) were stimulated in vitro for 7 days with a 16-amino acid peptide, PyCSP 280–295 (SYVPSAEQILEFVKQI), which contains the H-2Kd-restricted epitope, PyCSP 280–288 (SYVPSAEQI) in 24-well plates at peptide concentration of 2.5 μM. Culture medium used was RPMI supplemented with 10% heat-inactivated FCS, l-glutamine, 50 U/ml each of penicillin and streptomycin, and 2-ME at 5 × 10−5 M. P815 (H-2d) cells (American Type Culture Collection, Manassas, VA) labeled with 0.1 mCi 51Cr (DuPont-New England Nuclear, Boston, MA) and pulsed with the CTL peptide (PyCSP 280–288 (SYVPSAEQI)) at 0.025 μM were used as positive target cells. Control targets consisted of unpulsed 51Cr-labeled P815 cells. Varying numbers of effector cells were added to both positive and negative targets at 5000 target cells per well of 96-well U-bottom plates, and 5-h standard chromium release methods were followed. Percent lysis was calculated for both positive and negative targets by the formula: [(experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm)] × 100%. Spontaneous release was defined as background lysis of target cells in presence of medium alone (generally <10% of maximum lysis), and maximum lysis was defined as lysis in the presence of 5% Triton X-100. Net percent specific lysis was calculated as: % lysis of positive targets − % lysis of negative control targets.
Ex vivo CTL assay.
Freshly isolated immune spleen cells were used without the 7-day in vitro restimulation period. In this method, varying numbers of freshly isolated spleen cells were added to 5000 peptide-pulsed, 51Cr-labeled P815 target cells with medium containing 2% rat T-cell stim (Collaborative Biomedical Products, Bedford, MA) in 96-well U-bottom plates. Both positive targets, 51Cr-labeled P815 cells pulsed with the CTL peptide (PyCSP 280–288 (SYVPSAEQI)) at 0.025 μM, and control targets, consisting of unpulsed 51Cr-labeled P815, cells were tested. A 12-h chromium release method was followed. Net percent specific lysis was calculated as: % lysis of positive target cells − % lysis of negative control target cells.
ELISPOT assay for detection of Ag-specific IFN-γ-secreting CD8+ T cells
The numbers of PyCTL epitope-specific IFN-γ-producing CD8+ T cells were determined using freshly isolated spleen cells from mice that had received their second immunization 14 days earlier. These measurements were made after 24–28 h incubation of spleen cells with irradiated P815 target cells pulsed with a 1 μM concentration of the H-2d-restricted PyCSP peptide SYVPSAEQI as previously described (6, 19). Nitrocellulose plates (96-well, Millipore, Bedford, MA) were coated with 75 μl PBS containing 1 μg/ml purified rat anti-mouse IFN-γ mAb (PharMingen, San Diego, CA). After overnight incubation at room temperature, the wells were washed 6 times with culture medium and incubated for 1 h with 100 μl culture medium containing 10% FCS. Irradiated P815 target cells (100 μl of 2 × 105 cells/ml) were placed in the Ab-coated wells. The starting concentration for the freshly isolated unrestimulated effector cells was 1–4 × 106/ml, and 2-fold dilutions in duplicate were assayed. One set of duplicate cells was cocultured with irradiated P815 cells, which had been incubated with 1 μM SYVPSAEQI. The other set of cells was cocultured with irradiated P815 cells that had not been incubated with peptide. After incubation at 37°C and 5% CO2 for 24–28 h, the plates were washed six times with PBS containing 0.05% Tween 20 (PBS/T). The wells were then incubated with 100 μl of a solution of 1 μg/ml biotinylated anti-mouse IFN-γ-mAb (PharMingen) in PBS/T. After overnight incubation at 4°C, wells were washed with PBS/T and 100 μl peroxidase-labeled streptavidin (Kirkegaard and Perry) at a dilution of 1/1000 in PBS/T, was added to each well. After 1 h incubation at room temperature, wells were washed twice with PBS/T and twice with PBS. The spots were developed by following the instructions provided with the DAB Reagent set (Kirkegaard and Perry Laboratories). After 10–15 min, the number of spots corresponding to IFN-γ-producing cells in wells containing the different spleen cell dilutions was determined using a stereomicroscope. The results were expressed as the number of IFN-γ-secreting cells per 106 spleen cells. Net spots per 106 spleen cells were calculated as: number of spots with peptide pulsed targets − number of spots with unpulsed targets.
Inhibition of liver stage development assay (ILSDA)
Mouse hepatocytes were obtained by in situ collagenase perfusion of mouse liver as previously described (20). Briefly, livers were perfused in situ sequentially with HBSS and a collagenase solution. The cell suspension generated was then centrifuged over a Percoll gradient to remove dead cells. The hepatocytes were then seeded onto 8-well Lab-Tek chamber slides in complete medium (EMEM with Earle’s balanced salts supplemented with 0.2% BSA (fraction 5), 10% FCS, 2% penicillin-streptomycin solution, insulin, 1% l-glutamine solution (100×), and 1% nonessential amino acids solution (100×)) at a concentration of 1 × 105 cells/well. The slides were incubated overnight at 37°C in a 5% CO2, 95% air environment. The medium was changed the following day and fresh media containing dexamethasone (7 × 10−4) was added to the cultures.
ILSDA: restimulated spleen cells.
Hepatocyte cultures, which had been seeded onto 8-well chamber slides 24 h previously, were incubated with 7.5 × 104 P. yoelii sporozoites for 3 h. The cultures were then washed with medium and incubated for 24 h. Immune spleen cells, which had been restimulated with peptide PyCSP 280–295, were tested for their ability to inhibit the development of liver stage parasites. Spleen cells from naive mice were stimulated and tested in the same manner as the immunized mouse spleens and used as a control. Stimulated spleen cells (5 × 105) were added to the liver cultures and allowed to remain in culture with the infected hepatocytes for an additional 24 h. Assays were run in triplicate. The chamber slides were then fixed for 10 min in ice-cold absolute methanol and stained with NYLS3 mAb (21) and a FITC-labeled goat anti-mouse IgG. The stained parasites on the slides were then viewed and counted using an epifluorescence microscope. The slide reader was blinded and did not know which group received which vaccine combination. Percent inhibition was then calculated and expressed as a percentage according to the formula: % inhibition = [1 − (mean number of parasites in experimental wells/mean number of parasites in control wells)] × 100.
The biological activity of the Abs produced by the various regimens was assessed in an ILSDA. Hepatocyte cultures, seeded onto 8-well chamber slides and allowed to attach overnight, were used. After 24 h of incubation at 37°C in an atmosphere of 5% CO2 in air, medium was removed, and 7.5 × 104 salivary gland dissected sporozoites suspended in 25 μl of medium, and 25 μl of varying dilutions of sera from immunized mice or medium were added. After 3 h incubation, the cultures were washed to remove sporozoites that did not invade hepatocytes, and fresh medium was added. Assays were run in triplicate. After a further 24 h, the medium was changed, and at 48 h the cultures were fixed and incubated with a mAb directed against liver stage parasites of P. yoelii (NYLS3) and a FITC-labeled goat anti-mouse IgG. The slide reader was blinded and did not know which group received which vaccine combination. The stained parasites on the slides were then viewed and counted as already described: % inhibition = [1 − (mean number of parasites in cultures with immune serum/mean number of parasites in cultures with medium control)] × 100.
Protection against challenge
Immunized mice were challenged i.v. with 50 to 100 infective P. yoelii sporozoites 2 wk after the second immunization, and protection was defined as absence of patent parasitemia in Giemsa-stained blood smears during the 14-day follow-up period.
Simple linear regression was performed using SPSS for Windows version 8.0 (SPSS, Chicago, IL). Comparisons between groups were also performed using the same SPSS for Windows version 8.0. χ2 was used except when any cell contained fewer than five observations; then Fisher’s exact test (two-tailed) was substituted.
Effect of dose of rvPyCSP used for boosting after pPyCSP priming
In our previous publication using DNA priming followed by recombinant vaccinia boosting in the Py malaria model, we used 1 × 107 rvPyCSP PFU as the viral dose and were able to achieve 69% protection. We wished to know whether increasing the viral dose would increase the immune response. The results in Table I⇓ show that larger doses of virus provided better protection. Nine weeks after DNA priming, boosting with 1 × 108 PFU rvPyCSP resulted in a 100% protection, whereas a boost with 1 or 2 × 107 rvPyCSP PFU resulted in 40 and 80% protection. As in our previous studies (6), boosting with recombinant virus did not protect unless there had been a previous immunization with pPyCSP.
Priming with pPyCSP and pGMCSF and boosting with rvPyCSP induces increased protection against malaria
Because pGMCSF can enhance DNA immune responses to pPyCSP (3), we wished to know whether a mixture of pPyCSP and pGMCSF would improve priming for rvPyCSP. Our strategy was to use a suboptimal immunization schedule giving lower levels of protection to easily detect immune enhancement. We primed with a single dose of DNA followed 3 wk later by a boost with 2 × 107 PFU of rvPyCSP, a less than optimal amount of virus. Table II⇓ shows the results of three separate protection experiments admixing pGMCSF in the prime-boost regimen. In Experiment 1, using high doses of plasmid (100 μg pPyCSP and 30 μg pGMCSF), priming with admixtures of pGMCSF protected better than did priming with pPyCSP alone. In Experiment 2, we compared smaller and larger doses of DNA for priming. When a mix of the two plasmids was used, protection was maintained although plasmid doses were reduced 30- to 100-fold (1 μg pPyCSP plus 1 μg pGMCSF). In Experiment 3, we systematically compared low and high plasmid doses and extended the plasmid titration down to 0.001 μg of each DNA. Experiment 3 confirmed the results of Experiments 1 and 2 and showed that some protection was achieved with as little as 0.1 μg priming doses of pPyCSP and pGMCSF. It thus appears that pGMCSF is a powerful enhancer of pPyCSP at the priming step. Summing the results of the three experiments, we found that priming with high dose PyCSP plus GM-CSF DNA (DG-V) protected 23 of 28 mice (82%), whereas priming with low dose PyCSP plus GM-CSF (dg-V) protected 13 of 18 mice (72%) (comparison of DG-V vs dg-V (p = 0.430, χ2 test)). This degree of protection with dg-V, 13 of 18 mice (72%) was higher than that recorded for recombinant virus boosting with high dose PyCSP plasmid priming D-V, 9 of 20 mice (45%) (comparison of dg-V vs D-V, p = 0.089, χ2 test). Furthermore, the degree of protection by dg-V, 13 of 18 mice (72%), was significantly better than that recorded for recombinant virus boosting with low dose PyCSP plasmid priming alone, d-V, 0 of 10 mice (0%) (comparison of dg-V vs d-V, p = 0.00023, Fisher’s exact test, two-tailed), or two doses of high dose PyCSP plus GMCSF plasmid DG-DG, 3 of 20 mice (15%) (comparison of dg-V vs DG-DG, p = 0.0004, χ2). This is an important improvement in antimalarial immunity. Not only does the addition of pGMCSF to the priming dose increase protection but it also allows a large decrease in the amount of DNA required for immunization.
In vitro immune responses to PyCSP are increased after immunization with pPyCSP and pGMCSF and boosting with rvPyCSP
Having determined that pGMCSF improved priming, we wished to know whether any in vitro measures of immunity correlated with increased protection. We measured T cell and Ab responses in mice immunized with or without pGMCSF at the priming dose.
T cell responses to the PyCSP 280–288 Kd-restricted epitope.
The results of an experiment comparing T cell responses to the Kd-restricted epitope PyCSP 280–288 (SYVPSAEQI) among mice receiving the various immunizations are shown in Fig. 1⇓. Fig. 1⇓A shows CTL activity against targets cultured with peptide PyCSP 280–288 after 7 days of in vitro restimulation of splenocytes with PyCSP 280–295 (SYVPSAEQILEFVKQI) peptide. Fig. 1⇓B shows the number of cells producing IFN-γ in the same spleens. In the ELISPOT assay, spleen cells were incubated for 24–28 h with P815 cells pulsed with the PyCSP 280–288 peptide, and IFN-γ spot-forming cells were counted. Spleens from two mice per group were pooled for these assays. CTL activity and IFN-γ-producing cells was highest in the DG-V group with dg-V and D-V next highest in that order.
ILSDA is an in vitro assay system that quantifies the reduction of malaria infected hepatocytes after the addition of cells or serum. In mice from Experiment 2 in Table II⇑, splenocytes from two mice per group were pooled and restimulated for 7 days with the PyCSP 280–295 peptide. These cultures were then divided; one-half were used for a CTL assay, and one-half were used for ILSDA. Table III⇓ compares results for CTL, ILSDA, and protection in these groups. All three measures of immunity showed a consistent ranking of response with the highest levels of in vitro activity seen in the mice primed with high doses of pPyCSP plus pGMCSF plasmids.
The ability of Abs produced by the different immunization regimen to inhibit the entry and development of sporozoites in hepatocytes was also investigated. Pooled sera from the best protected groups regimens DG-V and dg-V gave the highest inhibitory effect, up to 82% (Table IV⇓). Our experience with passive transfer of protective mAbs (22) and immunization with synthetic peptides (23) has indicated that only levels of inhibition >90–95% have a high predictive value for protection.
Correlation of Ab and T cell responses with protection
Table V⇓ is a summary of our analysis of mice from Table II⇑, Experiment 3. Abs to PyCSP were measured three ways: by IFAT against P. yoelii sporozoites (column 3), and by ELISA using as capture Ag either a full length recombinant PyCSP protein produced in yeast (column 4), or a synthetic 18-mer peptide representing three copies of the 6-aa PyCSP repeat motif (column 5). T cell responses to the PyCSP 280–288 epitope were measured using three methods: ELISPOT measuring IFN-γ producing cells (column 6), and two CTL assays measuring lysis from splenocytes, either directly from spleens without restimulation (column 7), or after 7 days of culture with PyCSP 288–295 peptide (column 8). High dose DNA priming with pPyCSP + pGMCSF plasmids followed by boosting with recombinant virus gave the highest protection, as well as the highest levels of T cell and Ab responses by every measure. Other immunization regimens, using less DNA with or without virus boosting, gave less protection and lower Ab and T cell responses to the Kd-restricted epitope, PyCSP 280–288 (SYVPSAEQI) in vitro. There is a strong association between protection, and all measured immune responses as measured by simple linear regression. From this analysis, we conclude that overall immunity is increased by pPyCSP + pGMCSF DNA priming, but we cannot determine which immune responses are responsible for protection.
Mechanisms of protection after immunizing with plasmids and boosting with rvPyCSP
To understand the immune mechanisms responsible for protection in the DNA prime-viral boost immunizations, we tried to abrogate protection by injecting immunized mice with mAbs to IFN-γ and CD4 or CD8. We first tested mice immunized without the addition of pGMCSF (Table VI⇓). Mice were primed with 100 μg pPyCSP and boosted with 2 × 107 PFU rvPyCSP 9 wk later. Administration of an anti-CD8 mAb, which removed >99% of all CD8+ T cells, eliminated protection (p = 0.0001, Fisher’s exact test, two-tailed). (Table VI⇓, Experiment A). A single i.v. injection of 1 mg anti-IFN-γ mAb at the time of challenge reduced protection from 80% to 30% (Table VI⇓, Experiment B) (p = 0.069, Fisher’s exact test, two-tailed). These data indicated that the mechanism of protection with this prime-boost regimen was similar to that induced by pPyCSP immunization alone and probably involved CD8+ T cells that recognize infected hepatocytes and produced IFN-γ. Our previous data (2, 24, 25) indicate that the IFN-γ either induces the infected hepatocyte to produce nitric oxide, which eliminates the infected hepatocyte, or induces mononuclear cells to produce IL-12, which leads to production of IFN-γ by other T cells or NK cells (26).
We then attempted to repeat these in vivo depletion experiments in mice immunized with with pPyCSP + pGMCSF and rvPyCSP (Table VI⇑, Experiment C). Results revealed that after treatment with the anti-CD4 mAb, CD4+ T cells were depleted at least 98.3, 97.9, and 98.8% in the D-V, DG-V, and dg-V groups, respectively. CD8+ T cells were depleted at least 99.3, 99.5, and 99.4% in the D-V, DG-V, and dg-V groups, respectively, after treatment with the anti-CD8 mAb. In the low dose plasmid group (dg-V), treatment with anti-CD8 mAb completely eliminated protection (p = 0.0009, Fisher’s exact test, two-tailed). However, in the high dose plasmid group (DG-V), this same Ab reduced immunity, but 45% of the mice remained protected (p = 0.414, Fisher’s exact test, two-tailed), even though >99% of CD8+ T cells were removed by the Ab treatment. Treatment with anti-IFN-γ mAb only reduced protection to 83% in the DG-V group (p = 0.640, Fisher’s exact test, two-tailed), and to 58% in the dg-V group (p = 0.155 Fisher’s exact test, two-tailed), as compared with 30% in the previous D-V experiment. These data do not allow an unequivocal assignment of immune effector function in protection with the dg-V and DG-V prime-boost regimens. It appears that immune effector function with low doses of pGMCSF DNA priming is dependent on CD8+ T cells. However, with large plasmid doses of pGMCSF DNA, the fact that we cannot completely abrogate protection means either that the mAb treatment cannot neutralize higher levels of CD8+ T cells and IFN-γ, or that additional immune effector functions have come into play.
Genetic control of protection from prime/boost immunization
A/J and C57BL/6 mice were immunized with DNA i.m., boosted 12 wk later with DNA i.m. or recombinant vaccinia virus i.p. Mice were bled for serum at week 14 and then challenged with P. yoelii sporozoites (100 sporozoites for A/J mice and 50 sporozoites for C57BL/6 mice) to measure protection. Table VII⇓ shows protection data comparing the prime-boost immunization to other regimens of immunization. In the past, A/J mice have been difficult to protect with vaccination using pPyCSP alone (2). It appears the DG-V regimen is slightly more effective than other immunization regimens in A/J mice (Table VII⇓). Only the DG-V regimen was significantly better than its control, CG-V (p = 0.0171, Fisher’s exact test, two-tailed) The DG-V regimen elicited the highest levels of Abs against sporozoites (Table VII⇓), and sera from mice immunized with this regimen had the highest inhibitory activity against sporozoite invasion and development within hepatocytes in vitro (Table VIII⇓). This supported the interpretation that Abs might play a major role in the protection seen. However, in efforts to determine a protective role for CD8+ T cells in A/J mice, the nine mice that had been protected by the DG-V regimen in Table VII⇓ were treated with a control Ab or an anti-CD8 mAb. In these mice, Ab injections for depletions were started at 68 days after boost with recombinant virus and these depleted mice were rechallenged at 74 days after boost, and 57 days subsequent to first challenge. Three of the five mice (60%) treated with the control Ab were still protected, but none of the four mice treated with the anti-CD8 mAb were protected. The numbers of animals are too few to draw firm conclusions. However, the data support the interpretation that CD8+ T cells were required for the protection in the four previously protected A/J mice despite the high levels of inhibitory Abs. C57BL/6 mice were not significantly protected by any prime-boost regimen and may require immunization with additional Ags to achieve T cell-mediated protection or further optimization of the regimen.
Since our first descriptions of a protective DNA malaria vaccine in mice (1) and the safety and immunogenicity of a DNA malaria vaccine in humans (27), our goals have been to improve DNA vaccines by: 1) increasing protective efficacy, 2) increasing Ab response, and 3) decreasing on a dose. Although the protective efficacy of our plasmid DNA malaria vaccines has been consistent, they do not completely protect mice challenged with highly infective P. yoelii sporozoites. In most of our experiments, immunization with two or three doses of 100 μg PyCSP plasmid DNA alone has protected only 25–50% of BALB/c mice against challenge with 50 or 100 P. yoelii sporozoites (3, 6). CD8+ T cells are the most important effector cells in this model system, but protection is also dependent on IFN-γ induction of nitric oxide that apparently mediates killing of developing liver stage parasites (2, 26). One method for increasing the protection from a DNA vaccine is to include more than one malarial Ag. With mixtures of plasmids, immune responses are made independently to each plasmid Ag, and protection is apparently additive (2). However, even with multiple Ags, we have not consistently protected all animals and all mouse strains against malaria. Therefore, a major goal has been to improve protection from our P. yoelii DNA vaccines.
Our second goal has been to increase the humoral response to DNA vaccines. Results of the first human trial of a DNA malaria vaccine showed no Ab responses in any volunteer despite the induction of CTL responses in the majority of vaccinees (27). In mice, Ab responses after i.m. plasmid immunization vary widely, from moderately strong to undetectable, depending on the encoded Ag (2). Abs to malaria Ags are an important source of protection, particularly against the blood stages of malaria, and we have tried to devise ways to boost the Ab response to DNA vaccines.
Our third goal has been to decrease the amount of plasmid needed for immunization. In our first clinical trial (27), volunteers received three doses of 20, 100, 500, or 2500 μg of Plasmodium falciparum CSP plasmid DNA. CTL assays were positive more often in the 2500- and 500-μg groups than in the 20- and 100-μg groups. Obviously, reducing the dose would reduce the cost of production of a vaccine that is intended in large part for population groups with few financial resources. Epidermal administration of DNA on gold beads by a gene gun has been shown to require less DNA per dose, often <4 μg (28) but requires a complex technology that may be impractical for a vaccine which will be used in the developing world. Furthermore, in a recent study of a simian HIV DNA vaccine, rhesus monkeys primed with DNA administered by the intradermal route and boosted with recombinant fowlpox were protected. However, monkeys primed with the same DNA delivered on gold particles by gene gun were not protected (29).
Recent publications from our laboratory have identified two methods for increasing both CD8 T+ cell responses and Abs to pPyCSP. When plasmid encoding murine GMCSF is mixed with PyCSP plasmid, protection is doubled and Abs to PyCSP increase 8-fold, and CD4+ and CD8+ T cell responses increase 4-fold (3). It appears that the GMCSF causes an influx of APC into the injected muscle, leading to stronger Ag-specific immune responses.5 The second novel approach we have described is combining plasmid vaccination as a first dose followed by a vaccination with a recombinant poxvirus expressing the same Ag as the second dose (6). This “DNA prime-viral boost” technique increases protective efficacy, Abs, and T cell responses. We do not understand why the combination of DNA followed by poxvirus is superior to two doses of DNA or two doses of virus, but recombinant poxviruses also boost immune responses primed by recombinant adenoviruses (30). Similar data with the DNA prime-viral boost have been generated by other laboratories studying malaria in mice (7, 8) and simian HIV in rhesus monkeys (29).
In this paper, we describe the combination of these two approaches: priming with pPyCSP + pGMCSF plasmids followed by boosting with recombinant poxvirus. The results are remarkable. Using this combined method, we have been able simultaneously to achieve all three of our goals for plasmid DNA vaccination: protective efficacy and T cell responses were increased; Abs were enhanced; and we were able to reduce the dose of DNA required for priming by 1–2 logs. If this same approach can be made to work in humans, it will be an important advance toward an effective malaria vaccine.
Despite these impressive results, there are important shortcomings of the prime-boost vaccine, in terms of both our understanding and application. As mentioned above, we are still unsure how GMCSF enhances priming and why poxvirus boosting is so effective. Also, we have not completely defined the immune effector mechanisms that are responsible for immunity in the prime-boost model of PyCSP in BALB/c mice. In all our previous experience, in vivo treatments of immunized mice with mAbs to CD8+ T cells, or to IFN-γ always abrogated immunity. This has allowed us to conclude unambiguously that these two aspects of the immune response were most significant in killing malaria parasites. However, these in vivo depletions have not completely neutralized immunity in the prime-boost experiments. This may indicate that additional arms of the protective immune response such as specific Abs or CD4+ T cells are playing an important role in protecting these mice. Alternatively, our inability to abrogate the protective immune responses with anti-CD8 and anti-IFN-γ mAbs may simply reflect the very high levels of immune responses, which cannot be completely neutralized by the in vivo Ab treatments we used. Preliminary analysis of our data by multiple regression techniques also seems to indicate that the high levels of specific Ab in our mice may be contributing to the high levels of protection.
Although BALB/c mice have superior protection from the prime-boost PyCSP vaccine, protection in other strains of mice is much less complete. A/J mice have a modest level of protection (45%) from this same vaccine, whereas C57BL/6 mice show little if any protection. Although both the A/J and C57BL/6 of these strains show important increases in specific serum Abs from the prime-boost vaccine, the levels in the C57BL/6 mice are not as high. In the small number of animals tested, depletion of CD8+ T cells eliminated protection in the AJ mice previously protected, whereas 60% of mice receiving control Abs were still protected (>10 wk since the last immunization). Seeing protection in a previously unprotected strain (A/J) just by improving the immunization regimen suggests that further optimization of immunization regimens may increase this protection further. We are optimistic that the inclusion of several parasite Ags in the prime-boost format will allow an expanded vaccine to protect a wider range of mouse strains. An analogous multigene vaccine would hopefully be similarly effective in a wide range of human HLA types.
We thank Arnell Belmonte, Romeo Wallace, Stephen Abot, and Steve Matheny for excellent technical assistance and for providing P. yoelii sporozoites. We also thank T. R. Jones for statistical analysis.
↵1 This work was supported by the Naval Medical Research Center Work Unit STOF 6.2.622787A.0101.870.EFX.
↵2 The assertions herein are the private ones of the authors and are not to be construed as official or as reflecting the views of the U.S. Navy or the Naval service at large. The experiments reported herein were conducted according to the principles set forth in the Guide for the Care and Use of Laboratory Animals (31).
↵3 Address correspondence and reprint requests to Dr. S. L. Hoffman, Malaria Program, Naval Medical Research Center, 503 Robert Grant Avenue, Silver Spring, MD 20910-7500. E-mail address:
↵4 Abbreviations used in this paper: pPyCSP, circumsporozoite protein plasmid DNA; PyCSP(rvPyCSP), recombinant vaccinia virus expressing; pGMCSF, murine GM-CSF plasmid DNA; ELISPOT, enzyme-linked immunospot; IFAT, indirect fluorescent Ab test; ILSDA, inhibition of liver stage development assay.
↵5 D. Haddad. Granulocyte-macrophage colony-stimulating factor expressing plasmid vaccine elicits highly localized infiltrates of immature dendetric cells in injected muscles. Submitted for publication.
- Received December 3, 1999.
- Accepted March 21, 2000.
- Copyright © 2000 by The American Association of Immunologists