Research Article

Control of a Mucosal Challenge and Prevention of AIDS by a Multiprotein DNA/MVA Vaccine

See allHide authors and affiliations

Science  06 Apr 2001:
Vol. 292, Issue 5514, pp. 69-74
DOI: 10.1126/science.292.5514.69


Heterologous prime/boost regimens have the potential for raising high levels of immune responses. Here we report that DNA priming followed by a recombinant modified vaccinia Ankara (rMVA) booster controlled a highly pathogenic immunodeficiency virus challenge in a rhesus macaque model. Both the DNA and rMVA components of the vaccine expressed multiple immunodeficiency virus proteins. Two DNA inoculations at 0 and 8 weeks and a single rMVA booster at 24 weeks effectively controlled an intrarectal challenge administered 7 months after the booster. These findings provide hope that a relatively simple multiprotein DNA/MVA vaccine can help to control the acquired immune deficiency syndrome epidemic.

Cellular immunity plays an important role in the control of immunodeficiency virus infections (1). Recently, a DNA vaccine designed to enhance cellular immunity by cytokine augmentation successfully contained a highly virulent immunodeficiency virus challenge (2). Another promising approach to raising cellular immunity is DNA priming followed by recombinant poxvirus boosters (3). This heterologous prime/boost regimen induces 10- to 100-fold higher frequencies of T cells than priming and boosting with DNA or recombinant poxvirus vaccines alone. Previously, we showed that boosting a DNA-primed response with a poxvirus was superior to boosting with DNA or protein for the control of a nonpathogenic immunodeficiency virus (4). Here we test DNA priming and poxvirus boosting for the ability to protect against a highly pathogenic mucosal challenge. The 89.6 chimera of simian and human immunodeficiency viruses (SHIV-89.6) was used for the construction of immunogens and its highly pathogenic derivative, SHIV-89.6P, for challenge (5). SHIV-89.6 and SHIV-89.6P do not generate cross-neutralizing antibody (6) and allowed us to address the ability of vaccine-raised T cells and nonneutralizing antibodies to control an immunodeficiency virus challenge. Modified vaccinia Ankara (MVA) was used for the construction of the recombinant poxvirus. MVA has been highly effective at boosting DNA-primed CD8 T cells and enjoys the safety feature of not replicating in human or monkey cells (3).

To ensure a broad immune response, both the DNA and recombinant MVA (rMVA) components of the vaccine expressed multiple immunodeficiency virus proteins. The DNA prime (DNA/89.6) expressed simian immunodeficiency virus (SIV) Gag, Pol, Vif, Vpx, and Vpr and human immunodeficiency virus–1 (HIV-1) Env, Tat, and Rev from a single transcript (7–9). The rMVA booster (MVA/89.6) expressed SIV Gag, Pol, and HIV-1 Env under the control of vaccinia virus early/late promoters (10). Vaccination was accomplished by priming with DNA at 0 and 8 weeks and boosting with rMVA at 24 weeks (Fig. 1A) (11).Four groups of six rhesus macaques each were primed with either 2.5 mg (high-dose) or 250 μg (low-dose) of DNA by intradermal (i.d.) or intramuscular (i.m.) routes with a needleless jet injection device (Bioject, Portland, Oregon) (12, 13). Gene gun deliveries of DNA were not used because these had primed nonprotective immune responses in our 1996–98 trial (4). The MVA/89.6 booster immunization [2 × 108 plaque-forming units (pfu)] was injected with a needle both i.d. and i.m. A control group included two mock immunized animals and two naïve animals. The challenge was given at 7 months after the rMVA booster to test for the generation of long-term immunity. Because most HIV-1 infections are transmitted across mucosal surfaces, an intrarectal challenge was administered.

Figure 1

Temporal frequencies of Gag-specific T cells. (A) Gag-specific CD8 T cell responses raised by DNA priming and rMVA booster immunizations. The schematic presents mean Gag-CM9–tetramer data generated in the high-dose i.d. DNA-immunized animals. (B) Gag-specific IFN-γ ELISPOTs inA*01 (open bars) and non-A*01 (filled bars) macaques at various times before challenge and at 2 weeks after challenge. Three pools of 10 to 13 Gag peptides (22-mers overlapping by 12) were used for the analyses. The numbers above data bars represent the arithmetic mean ± the SD for the ELISPOTs within each group. The numbers at the top of the graphs designate individual animals. *, data not available; #, <20 ELISPOTs per 1 × 106PBMC. Temporal data for Gag-CM9–Mamu-A*01 tetramer-specific T cells can be found in the supplementary data (17).

DNA priming followed by rMVA boosting generated high frequencies of virus-specific T cells that peaked at 1 week after the rMVA booster (Fig. 1). The frequencies of T cells recognizing the Gag-CM9 epitope were assessed by means of Mamu-A*01 tetramers (14), and the frequencies of T cells recognizing epitopes throughout Gag were assessed with pools of overlapping peptides and an enzyme-linked immunospot (ELISPOT) assay (15, 16). Gag-CM9 tetramer analyses were restricted to macaques that expressed theMamu-A*01 histocompatibility type, whereas ELISPOT responses did not depend on a specific histocompatibility type. As expected, the DNA immunizations raised low levels of memory cells that expanded to high frequencies within 1 week of the rMVA booster (Fig. 1) (17). In Mamu-A*01 macaques, CD8 cells specific to the Gag-CM9 epitope expanded to frequencies as high as 19% of total CD8 T cells (17). This peak of specific cells underwent a 10- to 100-fold contraction into the DNA/MVA memory pool (Fig. 1A) (17). ELISPOTs for three pools of Gag peptides also underwent a major expansion [frequencies up to 4000 spots for 1 × 106 peripheral blood mononuclear cells (PBMC)] before contracting from 5- to 20-fold into the DNA/MVA memory response (Fig. 1B). The frequencies of ELISPOTs were the same in macaques with and without the A*01 histocompatibility type (P > 0.2) (18). At both peak and memory phases of the vaccine response, the rank order for the height of the ELISPOTs in the vaccine groups was 2.5 mg i.d. > 2.5 mg i.m. > 250 μg i.d. > 250 μg i.m. (Fig. 1B). The interferon-γ (IFN-γ) ELISPOTs included both CD4 and CD8 cells (19). Gag-CM9–specific CD8 cells had good lytic activity after restimulation with peptide (19).

The highly pathogenic SHIV-89.6P challenge was administered intrarectally at 7 months after the rMVA booster (20), when vaccine-raised T cells were in memory (Fig. 1). The challenge infected all of the vaccinated and control animals (Fig. 2). However, by 2 weeks after challenge, titers of plasma viral RNA were at least 10-fold lower in the vaccine groups (geometric means of 1 × 107 to 5 × 107) than in the control animals (geometric mean of 4 × 108) (Fig. 2A) (21–23). By 8 weeks after challenge, both high-dose DNA-primed groups and the low-dose i.d. DNA-primed group had reduced their geometric mean loads to about 1000 copies of viral RNA per milliliter. At this time, the low-dose i.m. DNA-primed group had a geometric mean of 6 × 103 copies of viral RNA and the nonvaccinated controls had a geometric mean of 2 × 106. By 20 weeks after challenge, even the low-dose i.m. group had reduced its geometric mean copies of viral RNA to 1000. Among the 24 vaccinated animals, only one animal, animal number 22 in the low-dose i.m. group, had intermittent viral loads above 1 × 104 copies per milliliter (Fig. 2D).

Figure 2

Temporal viral loads, CD4 counts, and survival after challenge of vaccinated and control animals. (A) Geometric mean viral loads and (B) geometric mean CD4 counts. (C) Survival curve for vaccinated and control animals. The dotted line represents all 24 vaccinated animals. (D) Viral loads and (E) CD4 counts for individual animals in the vaccine and control groups. The key to animal numbers is presented in (E). Assays for the first 12 weeks after challenge had a detection level of 1000 copies of RNA per milliliter of plasma. Animals with loads below 1000 were scored with a load of 500. For weeks 16 and 20, the detection level was 300 copies of RNA per milliliter. Animals with levels of virus below 300 were scored at 300.

By 5 weeks after challenge, all of the nonvaccinated controls had undergone a profound depletion of CD4 cells (Fig. 2B). All of the vaccinated animals maintained their CD4 cells, with the exception of animal 22 in the low-dose i.m. group (see above), which underwent a slow CD4 decline (Fig. 2E). By 23 weeks after challenge, three of the four control animals had succumbed to AIDS (Fig. 2C). These animals had variable degrees of enterocolitis with diarrhea, cryptosporidiosis, colicystitis, enteric campylobacter infection, splenomegaly, lymphadenopathy, and SIV-associated giant cell pneumonia. In contrast, all 24 vaccinated animals maintained their health.

Containment of the viral challenge was associated with a burst of antiviral T cells (Figs. 1 and 3A). At 1 week after challenge, the frequency of tetramer+ cells in the peripheral blood had decreased, potentially reflecting the recruitment of specific T cells to the site of infection (Fig. 3A). However, by 2 weeks after challenge, tetramer+ cells in the peripheral blood had expanded to frequencies as high as, or higher than, after the rMVA booster (Figs. 1 and 3A). The majority of the tetramer+ cells produced IFN-γ in response to a 6-hour peptide stimulation (Fig. 3B) (24, 25) and did not have the “stunned” IFN-γ negative phenotype sometimes observed in viral infections (26). The postchallenge burst of T cells contracted concomitant with the decline of the viral load. By 12 weeks after challenge, virus-specific T cells were present at about one-tenth of their peak height (Figs. 1A and 3A) (19). In contrast to the vigorous secondary response in the vaccinated animals, the naïve animals mounted a modest primary response (Figs. 1B and 3A). Tetramer+ cells peaked at less than 1% of total CD8 cells (Fig. 3A), and IFN-γ–producing ELISPOTs were present at a mean frequency of about 300 as opposed to the much higher frequencies of 1000 to 6000 in the vaccine groups (Fig. 1B) (P < 0.05) (18). The tetramer+ cells in the control group, like those in the vaccine group, produced IFN-γ after peptide stimulation (Fig. 3B). By 12 weeks after challenge, three of the four controls had undetectable levels of IFN-γ–producing ELISPOTs (19). This rapid loss of antiviral T cells in the presence of high viral loads may reflect the lack of CD4 help.

Figure 3

Postchallenge T cell responses in vaccine and control groups. (A) Temporal tetramer+ cells (dashed blue line) and viral loads (solid pink line). (B) Intracellular cytokine assays for IFN-γ production in response to stimulation with the Gag-CM9 peptide at 2 weeks after challenge. This ex vivo assay allows evaluation of the functional status of the peak postchallenge tetramer+ cells displayed in Fig. 1A. (C) Proliferation assay at 12 weeks after challenge. Gag-Pol-Env (open bars) and Gag-Pol (hatched bars) produced by transient transfections were used for stimulation. Supernatants from mock-transfected cultures served as control antigen. Stimulation indices are the growth of cultures in the presence of viral antigens divided by the growth of cultures in the presence of mock antigen.

T cell proliferative responses demonstrated that virus-specific CD4 cells had survived the challenge and were available to support the antiviral immune response (Fig. 3C) (27). At 12 weeks after challenge, mean stimulation indices for Gag-Pol-Env or Gag-Pol proteins ranged from 35 to 14 in the vaccine groups but were undetectable in the control group. Consistent with the proliferation assays, intracellular cytokine assays demonstrated the presence of virus-specific CD4 cells in vaccinated but not control animals (19). The overall rank order of the vaccine groups for the magnitude of the proliferative response was 2.5 mg i.d. > 2.5 mg i.m. > 250 μg i.d. > 250 μg i.m.

At 12 weeks after challenge, lymph nodes from the vaccinated animals were morphologically intact and responding to the infection, whereas those from the infected controls had been functionally destroyed (Fig. 4). Nodes from vaccinated animals contained large numbers of reactive secondary follicles with expanded germinal centers and discrete dark and light zones (Fig. 4A). By contrast, lymph nodes from the nonvaccinated control animals showed follicular and paracortical depletion (Fig. 4B), whereas those from unvaccinated and unchallenged animals displayed normal numbers of minimally reactive germinal centers (Fig. 4C). Germinal centers occupied <0.05% of total lymph node area in the infected controls, 2% of the lymph node area in the uninfected controls, and up to 18% of the lymph node area in the vaccinated groups (Fig. 4D). More vigorous immune reactivity in the low-dose than the high-dose DNA-primed animals was suggested by more extensive germinal centers in the low-dose group (Fig. 4D). At 12 weeks after challenge, in situ hybridization for viral RNA revealed rare virus-expressing cells in lymph nodes from 3 of the 24 vaccinated macaques, whereas virus-expressing cells were readily detected in lymph nodes from each of the infected control animals (19). In the controls, which had undergone a profound depletion in CD4 T cells, the cytomorphology of infected lymph node cells was consistent with a macrophage phenotype (19).

Figure 4

Lymph node histomorphology at 12 weeks after challenge. (A) Typical lymph node from a vaccinated macaque showing evidence of follicular hyperplasia characterized by the presence of numerous secondary follicles with expanded germinal centers and discrete dark and light zones. (B) Typical lymph node from an infected control animal showing follicular depletion and paracortical lymphocellular atrophy. (C) A representative lymph node from an age-matched, uninfected macaque displaying nonreactive germinal centers. (D) The percentage of the total lymph node area occupied by germinal centers was measured to give a nonspecific indicator of follicular hyperplasia. Data for uninfected controls are for four age-matched rhesus macaques.

The prime/boost strategy raised low levels of antibody to Gag and undetectable levels of antibody to Env (Fig. 5). Postchallenge, antibodies to both Env and Gag underwent anamnestic responses with total Gag antibody reaching heights approaching 1 mg/ml and total Env antibody reaching heights of up to 100 μg/ml (28). By 2 weeks after challenge, neutralizing antibodies for the 89.6 immunogen, but not the SHIV-89.6P challenge, were present in the high-dose DNA-primed groups (geometric mean titers of 352 in the i.d. and 303 in the i.m. groups) (Fig. 5C) (29). By 5 weeks after challenge, neutralizing antibody to 89.6P had been generated (geometric mean titers of 200 in the high-dose i.d. and 126 in the high-dose i.m. group) (Fig. 5D) and neutralizing antibody to 89.6 had started to decline. By 16 to 20 weeks after challenge, antibodies to Gag and Env had fallen in most animals.

Figure 5

Temporal antibody responses. Micrograms of total Gag (A) or Env (B) antibody were determined with ELISAs. The titers of neutralizing antibody for SHIV-89.6 (C) and SHIV-89.6P (D) were determined with MT-2 cell killing and neutral red staining (29). Titers are the reciprocal of the serum dilution giving 50% neutralization of the indicated viruses grown in human PBMC. Symbols for animals are the same as in Fig. 2.

Discussion. Our results demonstrate that a multiprotein DNA/MVA vaccine can raise a memory immune response capable of controlling a highly virulent mucosal immunodeficiency virus challenge. Our levels of viral control were more favorable than have been achieved with only DNA (30) or rMVA vaccines (31) and were comparable to those obtained for DNA immunizations adjuvanted with interleukin-2 (2). All of these previous studies have used more than three vaccine inoculations, none have used mucosal challenges, and most have challenged at peak effector responses and not allowed a prolonged postvaccination period to test for “long-term” efficacy.

The dose of DNA had statistically significant effects on both cellular and humoral responses (P < 0.05), whereas the route of DNA administration affected only humoral responses (18). Intradermal DNA delivery was about 10 times more effective than i.m. inoculations for generating antibody to Gag (P = 0.02) (18). Neither route nor dose of DNA appeared to have a significant effect on protection. At 20 weeks after challenge, the high-dose DNA-primed animals had slightly lower geometric mean levels of viral RNA (7 × 102 and 5 × 102) than the low-dose DNA-primed animals (9 × 102 and 1 × 103).

The DNA/MVA vaccine controlled the infection, rapidly reducing viral loads to near or below 1000 copies of viral RNA per milliliter of blood. Containment, rather than prevention of infection, affords the opportunity to establish a chronic infection (4). By rapidly reducing viral loads, a multiprotein DNA/MVA vaccine will extend the prospect for long-term nonprogression and limit HIV transmission (32, 33).

  • * Present address: Department of Microbiology, University of Toronto, Toronto, Ontario, Canada M5G 1X5.

  • Present address: MIT Center for Cancer Research, Cambridge, MA 02139, USA.

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


View Abstract

Navigate This Article