Induction of Antigen-Specific Cytotoxic T Lymphocytes in Humans by a Malaria DNA Vaccine

See allHide authors and affiliations

Science  16 Oct 1998:
Vol. 282, Issue 5388, pp. 476-480
DOI: 10.1126/science.282.5388.476


CD8+ cytotoxic T lymphocytes (CTLs) are critical for protection against intracellular pathogens but often have been difficult to induce by subunit vaccines in animals. DNA vaccines elicit protective CD8+ T cell responses. Malaria-naı̈ve volunteers who were vaccinated with plasmid DNA encoding a malaria protein developed antigen-specific, genetically restricted, CD8+ T cell–dependent CTLs. Responses were directed against all 10 peptides tested and were restricted by six human lymphocyte antigen (HLA) class I alleles. This first demonstration in healthy naı̈ve humans of the induction of CD8+ CTLs by DNA vaccines, including CTLs that were restricted by multiple HLA alleles in the same individual, provides a foundation for further human testing of this potentially revolutionary vaccine technology.

During 1990–1994, the administration of “naked” plasmid DNA encoding a specific protein antigen was shown to induce expression of the protein in mouse myocytes (1), to elicit antibodies against the protein (2), and to manifest protection against influenza (3) and malaria (4) that was dependent on CD8+ T cell responses against the expressed protein. Hundreds of publications have now reported the efficacy of DNA vaccines in small and large animal models of infectious diseases, cancer, and autoimmune diseases (5).

DNA vaccines elicit antibodies and CD4+ T cell responses in animals, but their major advantage at the immunological level has been their capacity to induce antigen-specific CD8+ T cell responses, including CTLs, which is a major mechanism of protection against intracellular pathogens. Important to our method of developing a malaria vaccine is the induction of CD8+ T cell responses against Plasmodium falciparum –infected hepatocytes (6). The lysis of cells in a standard chromium release assay was used as a surrogate for antihepatocyte responses, because it has been established that CD8+ CTLs, which recognize peptide-pulsed target cells, also recognize and eliminate parasite-infected hepatocytes (6). On the basis of our work with rodents (4,7) and our work and that of others with rhesus monkeys (8, 9), we have developed a plan for manufacturing and testing the efficacy of a multigene P. falciparumliver-stage DNA vaccine in humans (10). This has been contingent on establishing that DNA vaccination of humans is safe and induces antigen-specific, genetically restricted, CD8+ T cell–dependent CTLs. Recently, the presence of CTL responses in human immunodeficiency virus (HIV)–infected individuals after vaccination with plasmid DNA encoding the nef, rev, ortat genes or the env and rev genes of HIV was reported (11). Interpreting these results is difficult because of the concurrent HIV infection, which has been demonstrated to prime individuals for a CTL response that is independent of immunization.

Accordingly, 20 healthy, malaria-naı̈ve adults were recruited and randomized into four dosage groups of five individuals. Three injections of 20, 100, 500, or 2500 μg of plasmid DNA encoding the P. falciparum circumsporozoite protein (PfCSP) (12) were administered at 4-week intervals in alternate deltoids (13). The details of recruitment, safety, and tolerability were reported elsewhere (14). To assess CTL responses, we collected peripheral blood mononuclear cells (PBMCs) from each volunteer before vaccination, 2 weeks after the second immunization, and 2 and 6 weeks after the third immunization. These cells were either assayed while fresh for recall antigen-specific CTL responses (15) or were frozen (16) for subsequent study. In parallel, CTL assays were carried out with PBMCs from nonimmunized control volunteers. Cytolytic activity was assessed after both primary and secondary in vitro restimulation against HLA-matched and HLA-mismatched PfCSP-specific and control targets. The percent lysis and the percent specific lysis were determined as described (15). The most sensitive and specific method (17) for demonstrating the presence of CTLs was with effector cells that were expanded in vitro by exposure to cells infected with canary pox (ALVAC) expressing the PfCSP (18) and with target cells that were sensitized with PfCSP-derived synthetic peptides (19). There was no apparent difference between the primary and secondary assays (20) or between the fresh and frozen specimens (21).

For logistical reasons, fresh PBMCs were studied only before vaccination and after the second immunization in the 20- and 100-μg-dosage groups but were studied before vaccination and after all immunizations in the 500- and 2500-μg-dosage groups, with the exception of one individual (13). For 14 individuals, adequate amounts of frozen PBMCs were available for further analysis. A typical pattern of CTL responses is presented in Fig. 1A. These responses were peptide-specific and genetically restricted because there was little or no recognition of autologous targets that were incubated with the control peptide or of HLA class I–mismatched targets that were incubated with the specific peptide. This activity was shown to be CD8+T cell–dependent by restimulating the effector cells and repeating the assay after the depletion of T cell subsets (Fig. 1B) (22). The simultaneous assessment of coded frozen PBMCs that were collected before and after vaccination (Fig. 1C) confirmed that these CTLs were induced after vaccination with a plasmid DNA and after subsequent translation of the encoded PfCSP.

Figure 1

(A) Peptide-specific recall CTL response to vaccination with plasmid DNA. Fresh PBMCs from volunteer 37 (2500-μg-dosage group) [which were collected before vaccination (pre-bleed), 2 weeks after the second immunization (2 wk p2), and 2 weeks after the third immunization (2 wk p3)] were stimulated in vitro with ALVAC expressing the PfCSP for 7 days and then were assayed against HLA class I–matched or HLA class I–mismatched PfCSP peptide–pulsed or control peptide 242–pulsed targets in a conventional chromium release assay (15). MHC, major histocompatibility complex. (B) DNA-induced CTL responses are CD8+ T cell–dependent. ALVAC PfCSP–stimulated effector cell populations from volunteer 37 were depleted in vitro of CD8+ or CD4+ T cells (22) immediately before the chromium release assay. Results of a 13-day assay at an E:T ratio of 20:1 are presented. (C) DNA-induced CTL responses with frozen PBMCs (16). Coded frozen PBMCs from volunteer 37 were assayed for peptide-specific CTLs. The % specific lysis in a 7-day assay at an E:T ratio of 20:1 is presented.

The CTL responses with fresh PBMCs from 9 of 20 volunteers and with frozen PBMCs from 6 of 14 volunteers met our criteria for positivity. Eleven of 20 volunteers were shown to have antigen-specific, genetically restricted CTL activity. The effect of T cell subset depletion was studied in fresh cells that were acquired from volunteers in the 500- and 2500-μg-dosage groups after the third immunization. In the five responders tested, CD8+ T cell depletion eliminated the CTL activity (Fig. 1B), and CD8+ T cell dependence was demonstrated for all 10 peptides, except peptide B35353. Volunteer 33 was the only volunteer who expressed HLA-B35. Accordingly, peptide B35353 was tested only once [at 2 weeks after immunization (13)]—at which time peptide-specific, genetically restricted CTLs were detected, but CD8+ dependence was not assessed. The presence of CD4+ CTLs could not be completely excluded, because (in some cases) there was a reduction in cytolytic activity upon the depletion of CD4+ T cells. However, this reduction was minor in relation to the effect of CD8+ T cell depletion (Fig. 1B).

In nine volunteers, CTLs could not be detected in the three assays that were conducted after immunization. In three of these volunteers, the lack of response could not be attributed to a failure to respond to the vaccine, because these individuals did not express any of the HLA alleles restricting the peptides under study (volunteers 8 and 19, 20 μg; volunteer 20, 500 μg). Four of the other six nonresponders were in the two lower dosage groups.

The CTL responses of all volunteers to all peptides after immunization are summarized in Table 1. The representative data for each responder to each peptide are presented in Fig. 2. The responses after immunization (80 of 341 assays, 23.4%) were significantly greater than those before immunization (6 of 139 assays, 4.3%) (P = 0.0000007, chi-squared test). Two of the 82 assays (2.4%) that were conducted with PBMCs from control (nonimmunized) volunteers were positive, which is significantly less than those assays that were conducted with PBMCs after immunization (P = 0.000015).

Figure 2

Representative data of positive (difference between the percent lysis of target cells pulsed with experimental or control peptides ≥10%) CTL responses for each volunteer for each peptide. Fresh or frozen PBMCs, taken at the same or different time points, were assayed for peptide-specific, genetically restricted CTLs as described in the caption of Fig. 1A. Shown is the percent lysis (mean ± SEM) for each peptide with its simultaneously assessed controls at a single E:T ratio. Error bars indicate SEM.

Table 1

Summary of overall CTL responses as assessed by vaccinia-stimulated effectors and peptide-sensitized targets with fresh and frozen PBMCs. Numbers separated by slashes without parentheses indicate the number of positive peptides out of the number of tested peptides; numbers separated by slashes in parentheses indicate the number of positive assays out of the number of assays. A positive assay result is defined as a percent specific lysis after vaccination that is >10% for at least two E:T ratios in the same assay, with a percent specific lysis before vaccination of <10%. In volunteers 7, 5, 13, and 37, CTLs against some peptides were detected before vaccination; these peptides were excluded from analysis for the respective individuals. nt, not tested.

View this table:

An apparent positive response was noted with one peptide for volunteers 5 (1 of 12 assays), 7 (1 of 10 assays), and 13 (1 of 8 assays) and with three peptides for volunteer 37 (3 of 11 assays) in secondary but not primary assays of fresh PBMCs that were collected before immunization. However, CTL responses after vaccination were significantly enhanced in relation to the levels before vaccination. Furthermore, no activity was detected when the assay was repeated with frozen PBMCs (21). Nevertheless, in accordance with our conservative definition of positivity, all peptides with ≥10% specific lysis before vaccination were excluded from subsequent analysis for the respective individual.

Peptide-specific, genetically restricted, and CD8+ T cell–dependent CTL responses were induced by as little as two 20-μg doses of DNA (Table 1). The induction of CTLs after a single immunization was not tested. CTL responses were detected in two of five volunteers immunized with 20 μg of DNA or 100 μg of DNA, in three of five volunteers immunized with 500 μg of DNA, and in four of five volunteers immunized with 2500 μg of DNA. Data indicate that immunization with either 500 or 2500 μg of DNA induced a significantly better CTL response in comparison with either 20 or 100 μg of DNA (P ≤ 0.0003). There was no significant difference between 500- and 2500-μg dosages overall or after the third immunization (P ≥ 0.53), but a significantly higher frequency of response was induced with 2500 μg of DNA in comparison to 500 μg of DNA after the second immunization (P = 0.000001). There was no significant difference between 20- and 100-μg dosages at any time (P ≥ 0.43).

With regard to the immunization schedule, overall, the rate of positive assays 2 weeks after the third immunization (31 of 106) was significantly greater than the rate after the second immunization (19 of 114) (P = 0.026); there was no significant difference between the 2- and 6-week time points after the third immunization (31 of 106 versus 30 of 121) (P = 0.45) (Table 1).

The frequency and magnitude of the CTL responses to specific peptides are summarized in Table 2. Overall, 5 of the 11 responders recognized 100% of the peptides studied, 3 responders recognized 60 to 70% of the peptides, and 3 others recognized 43 to 50% of the peptides.

Table 2

HLA restriction and magnitude and frequency of CTL responses for each of the 10 peptides studied.

View this table:

The DNA-induced CTLs were genetically restricted by multiple HLA alleles (Table 2). Representative data are presented in Fig. 3. There was no apparent hierarchy in terms of allele-specific recognition. The magnitude of the induced CTL responses to defined epitopes varied between volunteers. Overall, the best response was detected for the HLA-A2 restricted epitope, A2386 (Table 2). This response was not substantially different than the responses that were noted for the peptides A1310, A27, A3/11336, B7285, B35353, and B35+353. Induction by subunit vaccines of CD8+T cell–dependent immune responses of multiple HLA restrictions in the same individual has, to our knowledge, not previously been reported for any infectious agent in humans. This has been a major obstacle to vaccine development and will be critical to the success of a malaria vaccine, because T cell responses to individual epitopes are genetically restricted and there is a substantial allelic variation of CTL epitopes among P. falciparum isolates in nature (23). Indeed, it has been demonstrated that the irradiatedP. falciparum sporozoite vaccine, which confers potent protective immunity in humans (6), induces CTL responses that are restricted by multiple HLA alleles in genetically diverse individuals (24).

Figure 3

DNA-induced CTL responses are restricted by multiple HLA alleles. Fresh PBMCs from volunteer 36 (500-μg-dosage group), who expressed the alleles HLA-A2, A3, and B7, were assayed for antigen-specific, genetically restricted CTLs (15). The assay was repeated with coded frozen PBMCs (16) that were collected before and after vaccination; the results confirmed that the peptide-specific (the same five peptides), genetically restricted CTLs were induced by vaccination with plasmid DNA (21).

In mice, immunization with a P. yoeliicircumsporozoite protein DNA vaccine elicits a substantially greater CTL response than does immunization with irradiated sporozoites (4). We did not simultaneously compare CTL responses in our vaccine recipients with CTL responses in individuals who were immunized with irradiated sporozoites or naturally exposed to malaria. However, as in the rodent model (4), the magnitude of the CTL response that was seen in some of the volunteers (Fig. 1) was considerably higher than the response that is generally seen in humans exposed to irradiated sporozoites or to natural infection (24–27).

  • * These authors contributed equally to this work.

  • Present address: Clinical Research, Pasteur-Merieux Connaught-USA, Swiftwater, PA 18370, USA.

  • To whom correspondence should be addressed. E-mail: hoffmans{at}


View Abstract

Navigate This Article