Heterologous prime-boost vaccination has been shown to be an efficient way of inducing T cell responses in animals and in humans. We have used three vaccine vectors, naked DNA, modified vaccinia virus Ankara (MVA), and attenuated fowlpox strain, FP9, for prime-boost vaccination approaches against Plasmodium falciparum malaria in humans. In this study, we characterize, using two types of ELISPOT assays and FACS analysis, cell-mediated immune responses induced by different prime-boost combinations where all vectors encode a multiepitope string fused to the pre-erythrocytic Ag thrombospondin-related adhesion protein. We show that these different vectors need to be used in a specific order for an optimal ex vivo IFN-γ response. From the different combinations, DNA priming followed by MVA boosting and FP9 priming followed by MVA boosting were most immunogenic and in both cases the IFN-γ response was of broad specificity and cross-reactive against two P. falciparum strains (3D7 and T9/96). Immunization with all three vectors showed no improvement over optimal two vector regimes. Strong ex vivo IFN-γ responses peaked 1 wk after the booster dose, but cultured ELISPOT assays revealed longer-lasting T cell memory responses for at least 6 mo. In the DNA-primed vaccinees the IFN-γ response was mainly due to CD4+ T cells, whereas in the FP9-primed vaccinees it was mainly due to CD4-dependent CD8+ T cells. This difference may be of importance for the protective efficacy of these vaccination approaches against various diseases.

The main bottleneck in developing vaccines for intracellular infections such as HIV, tuberculosis, and malaria and chronic diseases such as cancer is the ability to induce strong and long-lasting cell-mediated immunity (CMI).4 Stimulation of a functional CD8 response is often crucial in addition to a Th1-type CD4 T cell response. DNA immunization was originally shown to induce strong CD8+ T cell responses in murine models ( 1) but it is now clear that induction of CMI by DNA vectors is not as potent in humans and thus new adjuvants and Ag delivery systems are being developed for improved immunogenicity ( 2). The use of recombinant viral vectors is an increasingly popular alternative to achieve intracellular Ag expression that can result in Ag presentation on MHC class I molecules that are recognized by specific CD8+ T cells ( 3). In 1993, Li et al. (4) showed that sequential immunization with two different replicating recombinant viral vectors could induce a strong CD8+ T cell response against malaria. Further studies in mice have shown that this immunization strategy, called heterologous prime-boost strategy, that combines two different vectors encoding the same Ag is more efficient in inducing CMI response than the use of a single vector and is surprisingly immunogenic with nonreplicating vectors ( 5, 6). Studies in nonhuman primates have revealed that the prime-boost approach induces strong CMI and can lead to protection against diseases such as malaria or simian-HIV also in primates ( 7, 8, 9, 10).

Our research group has been focusing on the immunogenicity and protective efficacy of prime-boost immunization strategies against malaria in humans. We have used three different vaccine vectors: plasmid DNA, modified vaccinia virus Ankara (MVA), and an attenuated strain of fowlpox, FP9, in various combinations. Both of the poxviral vectors can infect and express Ags in mammalian cells but are replication-deficient and thus the infection is self-limited ( 11). We have administered DNA and MVA constructs alone or sequentially to over 150 subjects and shown them to be well-tolerated, causing no serious or severe adverse events ( 12). FP9 has now been administered to over 40 volunteers in phase I/IIa malaria vaccine trials and again no serious or severe adverse events have occurred.5 In these phase I and IIa vaccine trials, we have shown that prime-boost strategy, using DNA priming and MVA boosting, induces strong IFN-γ responses in humans and is partially protective against an experimental Plasmodium falciparum challenge ( 13). Furthermore, FP9 prime and MVA boost vaccine regimens are also capable of inducing Th1-type CMI response and partial or even complete protection against P. falciparum challenge.6

Despite these encouraging results, few data have been reported on the detailed characteristics of the immune response induced by prime-boost vaccines in humans. In the present report, we have combined data from several prime-boost vaccine trials using different vector combinations, to compare the type of immune response they induced. In addition to detecting T effector cell responses, we included a cultured ELISPOT assay for the detection of less activated memory T cell responses ( 14) and measured intracellular cytokines by FACS analysis. As expected the heterologous prime-boost vaccine regimens were more immunogenic than homologous boosting but there were also differences in the priming and boosting capabilities for specific T cells of the different vaccine vectors. These differences may have an important impact for choosing an optimal prime-boost vaccination strategy for a given pathogen.

The vaccine regimes are described in detail.6 Briefly, the grouped data are from six different vaccine regimens where the Ag is the P. falciparum multiepitope (ME) string including mostly P. falciparum T cell and B cell epitopes fused to the whole thrombospondin-related adhesion protein (TRAP) ( 13). The different regimes used were as follows: DNA-DNA-MVA-FP9 (n = 4); DNA-DNA-FP9-MVA (n = 3); FP9-FP9-MVA (n = 12); FP9-MVA (n = 4); MVA-FP9 (n = 5); and MVA-MVA-MVA (n = 5). Some of the data from individual vaccinees are also from an earlier trial with three DNA primes followed by one MVA boost ( 13). Time intervals between all the immunizations were 3–4 wk. All vaccinees were challenged with P. falciparum strain 3D7 2–7 wk after the last vaccine dose as described in detail in Ref. 13 and D. P. Webster et al.6 The vaccine regimens studied were chosen on the bases of results from rodent studies that indicate DNA vaccines to be good in priming and pox viruses to be good in boosting. Thus, in any combination including both DNA vaccines and pox virus vectors, the pox viruses were always given last.

PBMC were isolated from heparinized blood samples using Lymphoprep (Nycomed Pharma) gradient centrifugation. Mononuclear cells were counted with a automated cell counter (CasyCounter TT; Schärfe System) and resuspended in RPMI 1640 (Sigma-Aldrich) containing 1000 U/ml penicillin-streptomycin (Invitrogen Life Technologies), 20 mM l-glutamine (Sigma-Aldrich), and 10% heat inactivated normal human AB sera (Blood Bank Service, National Health Service). Some of the PBMC were frozen in 10% DMSO-FCS (both from Sigma-Aldrich) and stored in liquid nitrogen. In the thawing procedure, 25 U/ml Benzonase nuclease (Novagen) was added to avoid cell clumping. For cell depletions, human anti-CD4 and -CD8 Dynabeads (Dynal) were used with a ratio of seven beads per cell according to the manufacturer’s instructions. The depletion efficacy was assessed with flow cytometric analysis and was typically 90 and 94% for the CD4+ and CD8+ cells, respectively.

For ex vivo ELISPOT, 0.4 × 106 freshly isolated PBMC/well were plated on MultiScreen Immobilon-P 96-well filtration plates (MAIP S45; Millipore) coated with 10 μg/ml anti-IFN-γ (1-D1K; MabTech) and blocked with 10% inactivated FCS. Cells were stimulated (as duplicates) with 25 μg/ml/peptide pools of TRAP derived 20-mer peptides overlapping by 10 amino acids. After 18–20 h incubation at +37°C 5% CO2 atmosphere, the cells were washed away and 1 μg/ml biotin labeled anti-IFN-γ (7-B6-1; MabTech) was added. A color reaction (spots) was obtained by adding alkaline phosphatase (MabTech) followed by a substrate (Bio-Rad). The number of spots was counted with an ELISPOT reader (AutoImmune Diagnostika). The results are expressed as number of spots per 106 cells in stimulated wells minus the number in unstimulated wells.

For cultured ELISPOT, 1 × 106 cryopreserved PBMC were stimulated with 10 μg/ml/peptide TRAP peptide pools and plated on a 24-well plate. After a 3-day incubation period at +37°C 5% CO2 atmosphere half of the cell culture supernatant (500 μl) was removed and replaced with 10 IU/ml Lymphocult (an IL-2 containing growth factor supplement obtained from Biotest). Fresh Lymphocult was added again on day 7. On day 9, the cells were washed three times and left to rest overnight in +37°C 5% CO2 atmosphere. On day 10, the cells were again washed and an amount corresponding to 2.5 × 104 originally plated cells was plated for a standard ELISPOT assay similarly as described above for the ex vivo ELISPOT.

A total of 0.5–1 × 106 cryopreserved PBMC were stimulated with 12.5 mg/ml TRAP peptide pools and 0.1 μg/ml anti-CD28 and anti-CD49d Abs (BD Pharmingen). After 2 h of stimulation, GolgiPlug (BD Pharmingen) was added and the cells were incubated for another 18 h at +37°C 5% CO2 atmosphere. After this the cells were first stained with surface marker-specific Abs (anti-CD4 PE and anti-CD8 PerCP), then fixed and permeabilized with Cytofix/Cytoperm solution and stained with cytokine-specific Abs (anti-IFN-γ FITC and anti-IL-2 allophycocyanin (all purchased from BD Pharmingen)). Stimulation with a streptococcal superantigen staphylococcal enterotoxin B (Sigma-Aldrich) was used as a positive control for all samples. The proportions of fluorochrome-labeled cells were assessed with FACSCalibur (BD Pharmingen) and analyzed by CellQuest (BD Pharmingen) to gate the lymphocytes according to size and morphology. Data was typically collected from 70,000 lymphocytes.

Statistical significance of differences between two groups was assessed with the nonparametric Mann-Whitney U test using SPSS software.

IFN-γ responses against peptides from the ME string and TRAP were measured from freshly isolated PBMC using an ex vivo ELISPOT assay. The arithmetic mean summed IFN-γ responses to ME or TRAP detected for different vaccine regimens are shown in Fig. 1. None of the three different vaccine vectors, naked DNA, MVA, nor FP9, induced statistically significantly increased IFN-γ responses as compared with the baseline responses, when given alone. However, when the vaccination was conducted as priming with one delivery system and boosting with another, a clear induction of IFN-γ response was detected. These responses were highest at 1 wk after the booster dose and were lower at 3–4 wk postvaccination.

FIGURE 1.

Ex vivo IFN-γ ELISPOT responses (mean ± SD of each group) assayed at baseline and different time points during vaccination with six different prime-boost vaccine regimens DNA-DNA-MVA-FP9 (A), DNA-DNA-FP9-MVA (B), FP9-FP9-MVA (C), FP9-MVA (D), MVA-FP9 (E), MVA-MVA (F).

FIGURE 1.

Ex vivo IFN-γ ELISPOT responses (mean ± SD of each group) assayed at baseline and different time points during vaccination with six different prime-boost vaccine regimens DNA-DNA-MVA-FP9 (A), DNA-DNA-FP9-MVA (B), FP9-FP9-MVA (C), FP9-MVA (D), MVA-FP9 (E), MVA-MVA (F).

Close modal

The IFN-γ response reached highest values after priming with DNA and boosting with MVA (Fig. 1,A). This response was of similar magnitude to that we have previously observed for three doses of DNA priming followed by MVA boosting using similar time intervals ( 13). A more moderate but clearly increased IFN-γ response was detected after two doses of FP9 as a priming agent and MVA boosting (Fig. 1,C). An almost similar level of IFN-γ production was detected after a single dose of FP9 as a priming agent followed by MVA boosting (Fig. 1,D) and when MVA was used as a second booster after initial DNA priming followed by FP9 boosting (Fig. 1,B). FP9 resulted in moderate boosting of a DNA-primed immune response (Fig. 1,B) but the level was ∼6-fold lower than that detected after DNA priming followed by MVA boosting (Fig. 1,A). Furthermore, FP9 did not provide any booster effect after MVA priming (Fig. 1,E) or as a second booster after DNA priming and MVA boosting (Fig. 1 A). Despite the differences in the magnitudes of the IFN-γ responses to TRAP obtained by different vaccine regimens, all the responses were broadly cross-reactive between the vaccine strain (T9/96) and another P. falciparum strain 3D7 that differ by 6% in their amino acid sequence.

The longevity of the IFN-γ response was assessed at 6–7 mo after final vaccine using ex vivo and cultured ELISPOT methods. The ex vivo ELISPOT, where cells are stimulated only for 18 h, detects cytokine production of effector T cells whereas the cultured ELISPOT where cells are stimulated for 10 days is likely to detect cytokine production of CCR7+ central memory cells ( 14, 15).

All volunteers received a P. falciparum sporozoite challenge after the final vaccination of the various vaccine regimens. When detecting IFN-γ responses with ex vivo ELISPOT assay at the 6–7 mo time point, the mean responses of all vaccine groups were very low (Fig. 2,A). However, because most of the vaccine groups had higher IFN-γ responses than the unvaccinated but P. falciparum-challenged controls, these remaining albeit low IFN-γ responses seemed to be vaccine-related in contrast to being challenge-related. When detecting IFN-γ responses with a cultured ELISPOT assay, it was shown that some vaccine regimens had induced memory T cells that could be expanded to IFN-γ-producing T cell populations in vitro (Fig. 2 B). Unvaccinated but challenged controls as well as unvaccinated and unchallenged controls gave very weak or absent cultured IFN-γ responses indicating that the observed memory T cell response was vaccine-specific and not due to the P. falciparum challenge given after vaccination. Two FP9 primes followed by an MVA boost as well as DNA priming followed by MVA and FP9 boosting were effective in inducing a long-lasting T cell response detected as increased IFN-γ release as compared with the unvaccinated controls. DNA priming followed by a single MVA boost was also capable of inducing a similar response (data not shown). Furthermore, three consecutive MVA vaccines were able to induce a long-term T cell memory response. In contrast, MVA priming followed by FP9 boosting and two DNA primes followed by FP9 and MVA boosting did not induce an elevated memory T cell response as compared with nonvaccinated controls. In contrast to the effector responses detected with the ex vivo ELISPOT method, elevated cultured ELISPOT responses were not detected at 7 days after final vaccination in any of the vaccine regimens (all responses below 500 spots per 106 plated cells), contrasting with the strong cultured ELISPOT responses detected 6 mo after the final vaccination.

FIGURE 2.

Ex vivo IFN-γ ELISPOT (A) and cultured IFN-γ ELISPOT (B) responses assayed at 6–7 mo after the final vaccine dose and a P. falciparum challenge in different vaccine groups. The controls in A contain nonvaccinated but P. falciparum-challenged volunteers and in B nonvaccinated but challenged volunteers as well as nonvaccinated nonchallenged volunteers. The numbers studied in A and B, respectively, are 2/1 for DDMF, 3/2 for DDFM, 10/4 for FFM, 2/0 for FM, 3/4 for MF, 3/2 for MMM, and 2/2 for control groups.

FIGURE 2.

Ex vivo IFN-γ ELISPOT (A) and cultured IFN-γ ELISPOT (B) responses assayed at 6–7 mo after the final vaccine dose and a P. falciparum challenge in different vaccine groups. The controls in A contain nonvaccinated but P. falciparum-challenged volunteers and in B nonvaccinated but challenged volunteers as well as nonvaccinated nonchallenged volunteers. The numbers studied in A and B, respectively, are 2/1 for DDMF, 3/2 for DDFM, 10/4 for FFM, 2/0 for FM, 3/4 for MF, 3/2 for MMM, and 2/2 for control groups.

Close modal

The IFN-γ-producing cells induced by the vaccinations were mainly specific for the TRAP part of the ME-TRAP vaccine construct. In the ME region only the tetanus toxoid and bacillus Calmette-Guérin-derived epitopes gave some IFN-γ response, but the P. falciparum epitope-specific IFN-γ responses were mostly negligible as observed previously by McConkey et al. in Ref. 13 (data not shown). The immune response against TRAP T9/96 (the vaccine strain) was measured using four pools of 20-mer peptides overlapping by 10 amino acids whereas that against 3D7 strain (the challenge strain) was measured using six pools of 20-mer peptides. The IFN-γ responses to different regions of the TRAP amino acid sequence are shown in Fig. 3 for all the vaccine regimens that had somewhat elevated responses at 7 days after final vaccine (DDMF, DDFM, FFM, and FM). The highest responses were obtained against peptides near the N terminus and moderate responses were obtained against C-terminal peptides, while peptides in the middle parts of TRAP were constantly weakly recognized by cells from any of the vaccinated volunteers. Interestingly, the data indicate that two FP9 priming doses (instead of one) were needed for a broad immune response similar to that seen in DDMF and DDFM vaccine regimens.

FIGURE 3.

The breadth of the observed ex vivo IFN-γ ELISPOT response detected 7 days after the final vaccine dose in the four vaccine groups that expressed elevated IFN-γ responses. Proportion (%) of the total IFN-γ response toward individual peptide pools consisting of 7–30 20-mer peptides overlapping by 10 amino acids.

FIGURE 3.

The breadth of the observed ex vivo IFN-γ ELISPOT response detected 7 days after the final vaccine dose in the four vaccine groups that expressed elevated IFN-γ responses. Proportion (%) of the total IFN-γ response toward individual peptide pools consisting of 7–30 20-mer peptides overlapping by 10 amino acids.

Close modal

The dependence of IFN-γ production on CD4+ or CD8+ cells was assessed with in vitro depletion of specific cell subsets using cell surface marker-specific magnetic beads before the ex vivo ELISPOT assay. This was performed using cryopreserved cells from the samples that had been found IFN-γ-positive in the initial ex vivo ELISPOT analysis, and it typically resulted in 90 and 94% depletion of CD4+ and CD8+ cells, respectively. The results in Table I show that in all four vaccinees where priming was performed using naked DNA as a delivery system, the IFN-γ responses were mainly CD4-dependent. In contrast, in five of the eight vaccinees with FP9 priming the response was clearly dependent on both CD4- and CD8-positive cells, whereas in the three other vaccinees it was mainly CD4-dependent.

Table I.

Percentage (%) of IFN-γ response in CD8- and CD4-depleted samples as compared to the response in the total PBMC cell population

Vaccine Regimena% in CD4 Depletedb (SFU in all PBMC)% in CD8 Depletedb (SFU in all PBMC)Type of Response
DDM 3.6 (278) 79.6 (278) CD4 
DDM 18.3 (2947) 110 (2947) CD4 
DDM 0 (213) 134 (213) CD4 
DDFM 28 (42) 144 (42) CD4 
FM 47.8 (38) 95.7 (38) CD4 
FM 18.0 (490) 8.2 (490) CD4 and CD8 
FM 28.0 (477) 17.1 (477) CD4 and CD8 
FFM 34.6 (43) 61.5 (43) CD4 (and CD8) 
FFM 6.3 (662) 7.1 (662) CD4 and CD8 
FFM 33.5 (542) 102 (542) CD4 
FFM 19.1 (183) 46.4 (183) CD4 and CD8 
FFM 0 (328) 13.7 (328) CD4 and CD8 
Vaccine Regimena% in CD4 Depletedb (SFU in all PBMC)% in CD8 Depletedb (SFU in all PBMC)Type of Response
DDM 3.6 (278) 79.6 (278) CD4 
DDM 18.3 (2947) 110 (2947) CD4 
DDM 0 (213) 134 (213) CD4 
DDFM 28 (42) 144 (42) CD4 
FM 47.8 (38) 95.7 (38) CD4 
FM 18.0 (490) 8.2 (490) CD4 and CD8 
FM 28.0 (477) 17.1 (477) CD4 and CD8 
FFM 34.6 (43) 61.5 (43) CD4 (and CD8) 
FFM 6.3 (662) 7.1 (662) CD4 and CD8 
FFM 33.5 (542) 102 (542) CD4 
FFM 19.1 (183) 46.4 (183) CD4 and CD8 
FFM 0 (328) 13.7 (328) CD4 and CD8 
a

Order of vaccinations for a given volunteer (D = DNA, M = MVA, F = FP9).

b

Percentage of IFN-γ spots detected in a CD4- or CD8-depleted sample of a given volunteer at 6–7 days after last vaccine as compared to the number of spots in the undepleted sample. Number of spots detected in the undepleted sample is shown in parentheses.

Next, the phenotype of IFN-γ-producing cells was assessed using intracellular cytokine staining combined with surface marker staining and followed by flow cytometric analysis. In addition to IFN-γ staining we combined IL-2 staining in this protocol and assessed the production of these cytokines by different cell phenotypes. Gating was first performed on lymphocytes and then on the basis of the surface marker expression. Typically, the PBMC samples consisted of ∼52% CD4+ cells, 20% CD8+ cells, and 26% CD4CD8 cells. Approximately 85% of all CD8+ cells were CD8bright which were all CD3+, but the remaining CD8dim population consisted of both CD3+ and CD3 cells (data not shown). For this reason, only the CD8bright cell population was included in the final analysis.

As can be seen in Fig. 4,A, in the DNA-primed vaccinees, a higher proportion of CD4+ cells were secreting IFN-γ (either alone or together with IL-2) than in the FP9-primed vaccinees (p = 0.01). In contrast, in the case of CD8+ cells, the FP9-primed vaccinees had higher proportions of activated CD8+ cells (producing IL-2 and/or IFN-γ) as compared with the DNA-primed vaccinees (Fig. 4 B). This difference reached statistical significance for both IL-2 alone and total IL-2/IFN-γ production in CD8bright cells (p = 0.02 and p = 0.04, respectively).

FIGURE 4.

Proportion of cytokine-secreting cells of all CD4+ (A) or CD8+ (B) cells in eight FP9-primed vaccinees and six DNA-primed vaccinees detected at 7–28 days after the final vaccine dose. Black parts of the bars represent cells producing only IFN-γ, white parts represent cells producing both IFN-γ and IL-2, and gray parts of the bars represent cells producing only IL-2.

FIGURE 4.

Proportion of cytokine-secreting cells of all CD4+ (A) or CD8+ (B) cells in eight FP9-primed vaccinees and six DNA-primed vaccinees detected at 7–28 days after the final vaccine dose. Black parts of the bars represent cells producing only IFN-γ, white parts represent cells producing both IFN-γ and IL-2, and gray parts of the bars represent cells producing only IL-2.

Close modal

In the present study we have conducted a detailed characterization of the cellular immune responses induced in humans by different heterologous prime-boost immunization strategies using various combinations of three different vaccine vectors (naked DNA, and MVA or FP9 viral vectors) each encoding the same Ag P. falciparum ME-TRAP ( 13).

As judged by ex vivo ELISPOT assays that detect the presence of Ag-specific effector T cells, single or consecutive administration with one delivery system did not induce IFN-γ responses that were significantly greater than the baseline values in these small group sizes. However, when priming was conducted with DNA or the FP9 vector and boosting with MVA, a clear induction of IFN-γ response was typically detected 7 days after the booster dose. These responses waned rapidly during the next couple of weeks and at 6–7 mo after vaccination all the ex vivo IFN-γ responses were rather low, although still elevated as compared with unvaccinated controls in some vaccine regimens. However, using a cultured ELISPOT assay where cells are allowed to go through an Ag-specific proliferation process before the detection of an Ag-specific effector function (IFN-γ production), the presence of T cell memory was detected. DNA- or FP9-primed and MVA-boosted vaccine regimens were clearly able to induce long-term memory detected at 6–7 mo after vaccination. In addition to these heterologous prime-boost vaccine regimens that also induced the strongest effector cell responses detected with ex vivo assay, a boosting immunization with MVA was able to induce long-term T cell memory.

These data point to strikingly different kinetic profiles of the vaccine-induced T cell responses as measured by the ex vivo and culture ELISPOT assays. In the former, the peak response is at ∼7 days following the final vaccination with perhaps 10% of the peak response detectable 6 mo later. In contrast, the cultured ELISPOT response induced by vaccination is markedly higher at 6 mo than at 7 days post-final vaccination, presumably reflecting differential kinetics of the differentiation process and effector functions of the evolving immune response ( 16). Strong cultured ELISPOT responses were not detected in unvaccinated volunteers 6 mo after challenge, or before vaccination indicating that the observed memory T cell responses were vaccine-induced.

Another interesting finding was that although one of the viral vectors, MVA, was a potent booster of cellular response, the other tested viral vector FP9 was not able to boost IFN-γ response after MVA priming or DNA and MVA priming. However, FP9 was able to provide moderate boosting of DNA-primed IFN-γ response when no MVA was involved. This finding might be explained with the type of anti-vector immunity induced by these viral vectors. MVA has a genome size of 186 kb with 193 open reading frames although only 152 intact genes ( 17), while FP9 has a genome size of 266 kb and 244 open reading frames ( 18). A plausible mechanism could be that because MVA has a smaller genome size with regions highly conserved among poxviruses, the anti-vector immune response developed after priming might be directed against common Ags and inhibit subsequent infection with FP9. In contrast, FP9 might induce anti-vector immunity to Ags that are not expressed in MVA and thus did not inhibit the subsequent infection with MVA. If this hypothesis, that needs to be confirmed by assessing the anti-vector immunity induced by the two viral vectors, were true then it would contrast with recent findings in mice. In a malaria mouse model, MVA and FP9 viral vectors were shown not to induce detectable cross-reactive anti-vector immunity, and there was no impairment of Ag-specific boosting by cross-immunity between these vectors. However, only a CD8 epitope-specific booster effect was measured in this mouse model and in a malaria challenge experiment, FP9 priming followed by MVA boosting was indeed superior in the protective capacity as compared with MVA priming and FP9 boosting ( 19).

The strongest booster effect (judged by the ex vivo IFN-γ ELISPOT responses) was obtained when DNA or FP9 was used for priming and MVA for boosting. This conclusion is based on the results of this study and of an earlier study by our research group ( 13). All these vaccine regimens (DDDMM, DDMF, and FFM) have also been shown to confer some level of protection against an experimental challenge with P. falciparum strain 3D7, detected as a statistically significant delay in the time to detectable parasitemia as compared with unvaccinated volunteers (Ref. 13 and D. P. Webster et al.).6 However, the ∼3-fold higher ex vivo IFN-γ responses in the DNA-MVA vaccine regimens as compared with the FP9-MVA vaccine regimen were not reflected in the magnitude of protection. On the contrary, the only two volunteers that have thus far shown complete protection against P. falciparum challenge after ME-TRAP vaccination both had been administered two FP9 priming immunizations and a single MVA booster.6 For this reason, we continued to characterize the cellular immune responses induced by DNA-primed and FP9-primed vaccines in more detail.

Both DNA- and FP9-primed vaccine regimes induced a cellular response that was specific not only to P. falciparum T9/96 strain (isolated from Thailand) but also to the 3D7 strain (thought to originate from Africa) used in the experimental challenge protocol. Cross-strain recognition is important in protection against malaria where the disease can be spread by a large number of different P. falciparum strains with much antigenic polymorphism. Furthermore, the cellular immune response was broad with responses against several of the T9/96 and 3D7 TRAP peptide pools. This is an important feature of CMI-inducing vaccines because vaccines with very narrow CMI responses directed against one or a couple of epitopes can, over time, lose protective efficacy due to escape mutants of the infectious agent, as has been shown in the case of simian-HIV infection ( 20). In accordance with the known locations of epitopes for CD4+ T cells, CD8+ T cells, and B cells within TRAP ( 21) the main part of the T cell response was directed toward the N- or C-terminal parts of TRAP while the middle part that contains mainly B cell epitopes had lower reactivity.

When the T cell phenotypes responsible for the IFN-γ production were assessed from cryopreserved cells some potentially important differences were observed between the DNA-primed and the FP9-primed vaccine regimens. Whether assessed either with in vitro depletion or intracellular FACS staining assays, the results indicated that DNA priming induced mainly a CD4+ T cell-specific response while FP9 priming induced a mixed response dependent on both CD4+ and CD8+ T cells. CD4+ T cells that comprise ∼50% of all lymphocytes (whereas CD8+ comprise ∼20%), contained higher numbers of IFN-γ-secreting cells in vaccinees given DNA priming as compared with vaccinees that were given FP9 priming. This probably explains the very high IFN-γ responses in the DNA-primed vaccinees detected in the ex vivo ELISPOT assay. In contrast, the number of CD8+ T cells recognizing TRAP peptides (assessed by IFN-γ and/or IL-2 secretion) was statistically significantly higher in the volunteers that had received a vaccine regimen with FP9 priming. This CD8 response was clearly CD4-dependent, however. CD4-dependent CD8 T cells have previously been reported in a study of vaccination with a different plasmid DNA malaria vaccine ( 22). This finding may reflect the importance of CD4 help mediated by CD40-CD40L interaction in memory CD8 T cell responses ( 23). Thus, a live viral vector (FP9) appears more efficient in priming for a CD8 response detectable after MVA boosting than naked DNA is as a priming agent, while both delivery systems are capable of priming for a CD4+ T cell response post-MVA boosting. This might relate to more efficient MHC class I presentation following viral infection and cytoplasmic Ag expression as compared with transfection of cells with naked DNA.

In conclusion, we show that DNA priming and MVA boosting as well as FP9 priming and MVA boosting are efficient heterologous vaccination strategies for inducing strong and long-lasting Th1 type CMI response in humans. However, the choice of the delivery system used for priming affects the outcome of the T cell phenotype (CD4 vs CD8) responsible for the detected effector function. Thus, the optimal priming system should be chosen on the basis of the desired immune response and what is known of protective immunity against a given pathogen.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by the Wellcome Trust. J.M.V. was in part supported by the Academy of Finland.

4

Abbreviations used in this paper: CMI, cell-mediated immunity; MVA, modified vaccinia virus Ankara; FP9, attenuated fowlpox strain 9; ME, multiepitope; TRAP, thrombospondin-related adhesion protein.

5

D. P. Webster, S. Dunachie, S. McConkey, I. Poulton, A. C. Moore, M. Walther, S. M. Laidlaw, T. Peto, M. A. Skinner, S. C. Gilbert, and A. V. S. Hill. Safety of recombinant fowlpox strain 9 and modified vaccinia virus Ankara vaccines against liver-stage P. falciparum malaria non-immune volunteers. Submitted for publication.

6

D. P. Webster, S. Dunachie, J. M. Vuola, T. Berthoud, S. Keating, S. Laidlaw, S. J. McConkey, I. Poulton, L. Andrews, R. F. Andersen, P. Bejon, G. Butcher, R. Sinden, M. Skinner, S. C. Gilbert, and A. V. S. Hill. Prime-boost vaccination using the recombinant poxviruses FP9 and MVA: enhanced T cell-mediated protection against malaria challenge in humans. Submitted for publication.

1
Ulmer, J. B., J. J. Donnelly, S. E. Parker, G. H. Rhodes, P. L. Felgner, V. J. Dwarki, S. H. Gromkowski, R. R. Deck, C. M. DeWitt, A. Friedman, et al
1993
. Heterologous protection against influenza by injection of DNA encoding a viral protein.
Science
19
:
1745
.
2
Donnelly, J., K. Berry, J. B. Ulmer.
2003
. Technical and regulatory hurdles for DNA vaccines.
Int. J. Parasitol.
33
:
457
.
3
Bonnet, M. C., J. Tartaglia, F. Verdier, P. Kourilsky, A. Lindberg, M. Klein, P. Moingeon.
2000
. Recombinant viruses as a tool for therapeutic vaccination against human cancers.
Immunol. Lett.
74
:
11
.
4
Li, S., M. Rodrigues, D. Rodriguez, J. R. Rodriguez, M. Esteban, P. Palese, R. S. Nussenzweig, F. Zavala.
1993
. Priming with recombinant influenza virus followed by administration of recombinant vaccinia virus induces CD8+ T-cell-mediated protective immunity against malaria.
Proc. Natl. Acad. Sci. USA
90
:
5214
.
5
Schneider, J., S. C. Gilbert, T. J. Blanchard, T. Hanke, K. J. Robson, C. M. Hannan, M. Becker, R. Sinden, G. L. Smith, A. V. S. Hill.
1998
. Enhanced immunogenicity for CD8+ T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara.
Nat. Med.
4
:
397
.
6
Sedegah, M., T. R. Jones, M. Kaur, R. Hedstrom, P. Hobart, J. A. Tine, S. L. Hoffman.
1998
. Boosting with recombinant vaccinia increases immunogenicity and protective efficacy of malaria DNA vaccine.
Proc. Natl. Acad. Sci. USA
95
:
7648
.
7
Schneider, J., J. A. M. Langermans, S. C. Gilbert, T. J. Blanchard, S. Twigg, S. Naitza, C. M. Hannan, M. Aidoo, A. Crisanti, K. J. Robson, et al
2001
. A prime-boost immunisation regimen using DNA followed by recombinant modified vaccinia virus Ankara induces strong cellular immune responses against the Plasmodium falciparum TRAP antigen in chimpanzees.
Vaccine
19
:
4595
.
8
Rogers, W. O., J. K. Baird, A. Kumar, J. A. Tine, W. Weiss, J. C. Aguiar, K. Gowda, R. Gwadz, S. Kumar, M. Gold, S. L. Hoffman.
2001
. Multistage multiantigen heterologous prime boost vaccine for Plasmodium knowlesi malaria provides partial protection in rhesus macaques.
Infect. Immun.
69
:
5565
.
9
Kent, S. J., A. Zhao, S. J. Best, J. D. Chandler, D. B. Boyle, I. A. Ramshaw.
1998
. Enhanced T-cell immunogenicity and protective efficacy of a human immunodeficiency virus type 1 vaccine regimen consisting of consecutive priming with DNA and boosting with recombinant fowlpox virus.
J. Virol.
72
:
10180
.
10
Amara, R. R., F. Villinger, J. D. Altman, S. L. Lydy, S. P. O’Neil, S. I. Staprans, D. C. Montefiori, Y. Xu, J. G. Herndon, L. S. Wyatt, et al
2001
. Control of a mucosal challenge and prevention of AIDS by a multiprotein DNA/MVA vaccine.
Science
292
:
69
.
11
Pastoret, P.-P., A. Vanderplasschen.
2003
. Poxviruses as vaccine vectors.
Comp. Immunol. Microbiol. Infect. Dis.
26
:
343
.
12
Moorthy, V. S., S. McConkey, M. Roberts, P. Gothard, N. Arulanantham, P. Degano, J. Schneider, C. Hannan, M. Roy, S. C. Gilbert, et al
2003
. Safety of DNA and modified vaccinia virus Ankara vaccines against liver-stage P. falciparum malaria in non-immune volunteers.
Vaccine
21
:
1995
.
13
McConkey, S. J., W. H. Reece, V. S. Moorthy, D. Webster, S. Dunachie, G. Butcher, J. M. Vuola, T. J. Blanchard, P. Gothard, K. Watkins, et al
2003
. Enhanced T-cell immunogenicity of plasmid DNA vaccines boosted by recombinant modified vaccinia virus Ankara in humans.
Nat. Med.
9
:
729
.
14
Godkin, A. J., H. C. Thomas, P. J. Openshaw.
2002
. Evolution of epitope-specific memory CD4+ T cells after clearance of hepatitis C virus.
J. Immunol.
15
:
2210
.
15
Flanagan, K. L., E. A. Lee, M. B. Gravenor, W. H. Reece, B. C. Urban, T. Doherty, K. A. Bojang, M. Pinder, A. V. Hill, M. Plebanski.
2001
. Unique T cell effector functions elicited by Plasmodium falciparum epitopes in malaria-exposed Africans tested by three T cell assays.
J. Immunol.
15
:
4729
.
16
Seder, R. A., R. Ahmed.
2003
. Similarities and differences in CD4+ and CD8+ effector and memory T cell generation.
Nat. Immunol.
4
:
835
.
17
Antoine, G., F. Scheiflinger, F. Dorner, F. G. Falkner.
1998
. The complete genomic sequence of the modified vaccinia Ankara strain: comparison with other orthopoxviruses.
Virology
10
:
365
.
18
Laidlaw, S. M., M. A. Skinner.
2004
. Comparison of the genome sequence of FP9, an attenuated, tissue culture-adapted European fowlpox virus, with those of virulent American and European viruses.
J. Gen. Virol.
85
:
305
.
19
Anderson, R. J., C. M. Hannan, S. C. Gilbert, S. M. Laidlaw, E. G. Sheu, S. Korten, R. Sinden, G. A. Butcher, M. A. Skinner, A. V. S. Hill.
2004
. Enhanced CD8+ T-cell immune responses and protection elicited against Plasmodium berghei malaria by prime boost immunisation regimens using a novel attenuated fowlpox virus.
J. Immunol.
172
:
3094
.
20
Barouch, D. H., J. Kunstman, M. J. Kuroda, J. E. Schmitz, S. Santra, F. W. Peyeri, G. R. Krivulka, K. Beaudry, M. A. Lifton, D. A. Gorgone, et al
2002
. Eventual AIDS vaccine failure in a rhesus monkey by viral escape from cytotoxic T lymphocytes.
Nature
415
:
335
.
21
Sinnis, P., E. Nardin.
2002
. Sporozoite antigens: biology and immunology of the circumsporozoite protein and thrombospondin-related anonymous protein. P. Perlmann, and M. Troye-Blomberg, eds.
Malaria Immunology
p. 70
. Basel,
22
Wang, R., J. Epstein, F. M. Baraceros, E. J. Gorak, Y. Charoenvit, D. J. Carucci, R. C. Hedstrom, N. Rahardjo, T. Gay, P. Hobart, et al
2001
. Induction of CD4+ T cell-dependent CD8+ type 1 responses in humans by a malaria DNA vaccine.
Proc. Natl. Acad. Sci. USA
11
:
10817
.
23
Bourgeois, C., B. Rocha, C. Tanchot.
2002
. A role for CD40 expression on CD8+ T cells in the generation of CD8+ T cell memory.
Science
297
:
2060
.