Effective Induction of Simian Immunodeficiency Virus-Specific Systemic and Mucosal Immune Responses in Primates by Vaccination with Proviral DNA Producing Intact but Noninfectious Virions

Abstract

We report a pilot evaluation of a DNA vaccine producing genetically inactivated simian immunodeficiency virus (SIV) particles in primates, with a focus on eliciting mucosal immunity. Our results demonstrate that DNA vaccines can be used to stimulate strong virus-specific mucosal immune responses in primates. The levels of immunoglobulin A (IgA) detected in rectal secretions of macaques that received the DNA vaccine intradermally and at the rectal mucosa were the most striking of all measured immune responses and were higher than usually achieved through natural infection. However, cytotoxic T lymphocyte responses were generally low and sporadically present in different animals. Upon rectal challenge with cloned SIVmac239, resistance to infection was observed, but some animals with high SIV-specific IgA levels in rectal secretions became infected. Our results suggest that high levels of IgA alone are not sufficient to prevent the establishment of chronic infection, although mucosal IgA responses may have a role in reducing the infectivity of the initial viral inoculum.

Globally, more than 80% of the transmission of human immunodeficiency virus (HIV) infection is via mucosal routes. The ability of vaccines to induce mucosal immunity may be required for protection against HIV infection or the immunodeficiency syndrome that emerges after infection.

Stimulation of simian immunodeficiency virus (SIV)-specific mucosal responses has been achieved with particulate antigens or with microencapsulated killed virus (20, 21, 27, 29, 31–33, 37). When the site of immunization targeted the iliac lymph node (TILN), total protection from rectal challenge was achieved, while protection from vaginal challenge was less consistent (33, 35). These results suggest that mucosal responses might be a desirable feature of an HIV vaccine (39). Since TILN vaccination or mucosal administration of infected and fixed cells is unlikely to be employed for large-scale immunization of humans, it is important to identify an alternative vaccination strategy that engenders a similar protective response but is more easily administered. Mucosal administration of DNA vaccines may provide an alternative and safe strategy. DNA vaccines have been successful in inducing antigen-specific mucosal responses in mice, but little is known about the ability of DNA vaccines to stimulate mucosal responses in primates. Furthermore, DNA vaccines have been surprisingly less successful in stimulating antigen-specific systemic immunoglobulin G (IgG) responses in primates than in mice, so it is particularly important to determine whether antigen-specific IgA production can be elicited in primates through DNA vaccination.

DNA vaccines expressing HIV genes have been investigated in humans to determine their safety and their ability to induce or boost virus-specific immune responses (6). Several DNA vaccine constructs and vaccination protocols have been evaluated alone or combined with other approaches for their ability to induce protection against challenge with retroviruses (13–15, 19, 25, 34, 45). When challenged intravenously (i.v.), the animals sometimes resisted the establishment of chronic infection and more frequently achieved decreased viral load and had a more prolonged asymptomatic state (13, 17, 45). These studies represent an interesting first step in the study of SIV DNA vaccines, but they are limited by several factors. In particular, these studies were not designed to evaluate mucosal immunity, and the challenges did not involve mucosal exposure.

Viral genomes that produce noninfectious virus-like particles have several features that make them attractive candidates for an AIDS vaccine. They may be capable of engendering immunity similar to that obtained with attenuated viral vaccines but do not establish the persistent infection associated with attenuated viruses (8). Noninfectious virus-like particles are produced in host cells in the same manner as a normal replicating virus and have protein components whose conformational integrity is maintained. Ideally, an altered viral genome would express proteins that assemble into noninfectious particles which contain all of the immunogenic components of the virus but which are unable to productively infect new cells.

We constructed a DNA vaccine with mutations in multiple structural genes that produces SIV particles that are noninfectious and yet are similar to normal SIV particles in protein content. We find that this DNA vaccine candidate is immunogenic in rhesus macaques and can stimulate significant levels of IgA antibodies in secretions when administered at the rectal mucosa. Detailed analyses of the immunological responses engendered by this vaccine and the ability of vaccinated primates to resist a challenge with live SIV are presented.

MATERIALS AND METHODS

Vector construction.

All mutants of SIVmac239 were constructed using the infectious clone pMA239 (14,110 bp) (47), which carries a full copy of the molecular clone of SIV mac239. Mutations were introduced in the SIV genome using oligonucleotide-mediated site directed mutagenesis by overlapping extension PCR (18) (Table 1). The individual changes introduced in three SIV proteins are listed in Table 1. In addition, the SIV 5′ long terminal repeat (LTR) was replaced by the cytomegalovirus (CMV) promoter or eukaryotic polypeptide chain elongation factor 1a (EF1a) promoter, and the SIV 3′ LTR was substituted with the polyadenylation site, poly(A), from pSG5 (Stratagene). The poly(A) fragment from the pSG5 vector replace sequences 9505 to 10709 of SIVmac239 (44). A fragment containing the CMV promoter, derived by PCR from the pRL CMV vector, replaced the 5′ SIV sequences up to the NarI site of pMA239. The modified plasmid containing the CMV promoter, multiple mutations in the HIV structural genes, and a replacement of the 3′ LTR with a poly(A) site was designated pVacc1. A similar strategy was used to replace the 5′ LTR with the EF1a promoter. The EF1a promoter DNA fragment, obtained by PCR from pEBB (50), replaced the CMV promoter of pVacc1 to obtain the plasmid designated pVacc2. All mutated viral sequences were confirmed by dideoxy sequencing. All DNA manipulations were carried out according to previously published procedures (3).

TABLE 1.

SIV proviral DNA constructs

Construct	5′ End	No. of mutations			3′ End
		Gag NCa	RTb	INc
pMA239	LTR	wtg	wt	wt	nef-stop, 3′ LTR
pMA22polyA	LTR	12	3	7	Δnef, poly(A)d
pVacc1	CMVe	12	3	7	Δnef, poly(A)
pVacc2	EF1af	12	3	7	Δnef, poly(A)

^aThe NC mutations are Q3→G, C3→W, five R→A, and five K→A. 

^bThe reverse transcriptase (RT) mutations are W71→A, R72→A, and R78→A. 

^cThe integrase (IN) mutations are alanine substitutions of the histidine and cysteine in the HHCC domain and of the aspartic acid and glutamic acid residues in the DD35E domain. 

^dThe simian virus 40 polyadenylation site replaces the 3′ end of the SIV genome from the termination codon of the env gene to the end of the 3′ LTR. 

^eThe CMV replaces the SIV 5′ LTR from the beginning of the SIV sequence in pMA239 to the SIV NarI site in pVacc1. 

^fThe eukaryotic polypeptide chain elongation factor 1a promoter (EF1a) replaces the SIV 5′ LTR from the beginning of the SIV sequence in pMA239 to the SIV NarI site in pVacc2. 

^gwt, wild type (no mutations). 

In vitro analysis of the SIV DNA vaccine candidates.

Transfection in 293T cells was carried out by the calcium phosphate method. Quantitative reverse transcription-PCR (RT-PCR) assay was carried out on cellular RNA extracted from 293T cells and DNase treated to eliminate contaminating DNA according to a previously described procedure (42). The protein and RNA contents of the viral particles produced from the vectors were characterized biochemically by Western blot and quantitative RT-PCR as previously described (43). Electron microscopy was carried out on sections of embedded particles according to standard procedures. Viral supernatants derived from two independent transfections per construct were tested in infectivity assays on CEMx174 cells as described elsewhere (43). Cultures were maintained for 30 days after infection. Cleared supernatants were tested for virus content by SIV p27 enzyme-linked immunosorbent assay (ELISA). Nested PCR on cellular DNA and RT-PCR on RNA from pelleted supernatants were also carried out on cultures that scored negative in SIV p27 ELISA.

Vaccine formulation.

Plasmid DNA was purified by a CsCl gradient, followed by passage through an endotoxin-free column (Qiagen). For intradermal (i.d.) and intramuscular (i.m.) administration of the DNA vaccine, saline solution (Sigma) was used to resuspend the DNA, and the concentration was adjusted to 1 mg/ml. For mucosal administration, the vaccine DNA was formulated in 20 mM DOTAP-cholesterol liposomes (51) at a concentration of 0.5 mg/ml. For i.d. gene gun administration, the plasmid DNA was precipitated onto 1.6-μm gold particles according to the manufacturer protocol (Bio-Rad).

Macaque vaccination.

Nine rhesus macaques were vaccinated at time zero and at 9 and 25 weeks with the SIV construct i.d. (group 1), i.d. and at the rectal mucosa (i.d., R) (group 2), and i.d., at the rectal mucosa, and i.m. (i.d., R, i.m.) (group 3). The animals in groups 1, 2, and 3 received 0.5 mg of pVacc1 DNA i.d. (0.4 mg of DNA in saline by needle injection and 0.1 mg DNA by gene gun) in the skin that covers the gluteal area. Additionally, at all three vaccination time points, 1 mg of pVacc1 DNA, mixed with liposomes according to the protocol described above, was administered to the rectal mucosa of animals in groups 2 and 3. Group 3 animals also received 1 mg of vaccine pVacc2 DNA administered i.m. to the gluteal muscle. Control animals were vaccinated like the animals in group 3, except that the pUC19 plasmid replaced the SIV-related plasmids.

SIV-specific IgA analysis in rectal secretions.

Rectal secretions were collected before and at intervals after immunization with absorbent Weck-Cel sponges (Windsor BioMedical, Newton, N.H.) using a modified wicking method that has been described in detail elsewhere (27a). The volume of secretion eluted from each sponge and dilution factors introduced by the premoistening saline and elution buffer were calculated based on weights of fluid centrifuged into 2-ml microcentrifuge lower-chamber tubes (Kozlowski et al., submitted). Blood contamination in secretions (assessed by measuring hemoglobin using Boehringer-Mannheim ChemStrips 4) was found to be negligible, representing only 0.01% of that in blood on average.

For detection of SIV and gp130-specific IgA and IgG in rectal secretions, plates were coated with 250 ng of SIV viral lysate or purified native gp130 (both from Advanced Biotechnologies, Rockville, Md.) per well. Antibodies measured in these SIV ELISAs likely do not include those to gp130 since this envelope protein could not be detected on plates coated with viral lysate using 5 μg of anti-gp130 antibody (Advanced Biotechnologies) per ml. Pooled serum from SIV-infected monkeys was used to generate standard curves in these assays for interpolation of antibody concentrations in samples. A highly specific anti-monkey IgA mouse IgG monoclonal antibody (52) was used as a secondary antibody, followed by biotinylated goat anti-mouse IgG antibody (Southern Biotechnology Associates, Birmingham, Ala.) from which antibodies cross-reactive with monkey IgG had been removed by passage through a column of CNBr-activated Sepharose (Pharmacia) conjugated to monkey IgG. In assays for SIV-specific IgG in rectal secretions, plates were developed with a biotinylated (Pierce Sulfo-NHS-LC-Biotin EZ-link kit) goat anti-monkey IgG antibody (Accurate, Westbury, N.Y.). To determine the endpoint titers of antibody in secretions, the last sample dilution producing an absorbance value of greater than or equal to the mean absorbance ± 3 standard deviations in eight control wells was multiplied by the dilution factor introduced into the secretion during elution from sponges.

To determine with accuracy whether rectal secretions contained significant levels of SIV-specific antibodies and to facilitate comparisons among animals in which total immunoglobulin concentrations in secretions were highly variable, measured antibody concentrations were divided by the total IgA or the total IgG concentration in each sample. Total IgA and IgG concentrations were quantitated by ELISA using plates coated with goat anti-monkey IgA or IgG (Accurate), a calibrated monkey serum standard provided by M. W. Russell (University of Alabama at Birmingham), and the above-described secondary reagents.

SIV-specific humoral responses in serum.

SIV-specific serum IgG were measured by ELISA using plates coated with whole virus or purified virus-derived gp130. Incubations were performed in duplicate, and multiple double dilutions of each sample were evaluated. Bound SIV-specific IgG were detected by incubation with an affinity-purified donkey anti-human IgG-alkaline phosphatase conjugate (Jackson Laboratories). Antibody-mediated neutralization of SIV was measured in a CEMx174 cell-killing assay as described previously (30, 40). Neutralization was measured with two stocks of SIV: (i) a laboratory-adapted stock of SIVmac251 produced in H9 cells and (ii) molecularly cloned SIVmac239/nef-open produced in rhesus peripheral blood mononuclear cells (PBMC) by using a vial of the original animal challenge virus as seed stock. The former virus is highly sensitive to neutralization, whereas the latter virus is extremely difficult to neutralize in vitro (30, 40).

Cell-mediated immune responses.

Cytotoxic-T-lymphocyte (CTL) assays were carried out according to previously described procedures (23). Autologous herpesvirus papio-transformed B lymphoblastoid cell lines (B-LCL) infected with a recombinant vaccinia virus vectors expressing the SIV Gag, Pol, and Env were used as stimulators and targets. Cytolytic activity was determined in a standard ⁵¹Cr release assay. Lysis was generally examined at effector/target (E/T) ratios of 40:1, 20:1, and 10:1. Based on examination of SIV-specific CTL activity in over 20 negative controls studied to date, SIV-specific CTL activity of ≥5% at two different E/T ratios was considered significant (23). In the subset of vaccinated animals that express the Mamu-A*01 allele, the frequency of CD3⁺ CD8⁺ cells specific for the SIV Gag 11C-M epitope was determined using major histocompatibility complex (MHC) tetramers.

RT-PCR and PBMC limiting-dilution viral loads.

Plasma SIV RNA levels were measured by a real-time RT-PCR assay, as described elsewhere (49). The assay has a threshold sensitivity of 300 copy Eq/ml. Interassay variation is <25% (coefficient of variation). Cell-associated virus loads were measured by limiting-dilution culture of PBMC every month during the postchallenge time course as previously described (53).

Flow cytometry.

Whole blood collected in EDTA was analyzed for lymphocyte subset CD4 (OKT4a, Ortho, and/or Anti-Leu3a; Becton Dickinson), CD8 (Anti-Leu2a; Becton Dickinson), and CDw29 (4B4; Coulter Immunology) by a whole-blood lysis technique described previously (53).

RESULTS

DNA constructs expressing genetically inactivated SIV particles.

Our vaccine design strategy was to introduce a large number of inactivating mutations into the viral genome, thereby minimizing the probability of genetic reversion to an infectious virus while retaining the ability to produce particles with protein content and immunogenic properties similar to wild-type virions. We have identified a combination of 22 mutations in three independent SIV genes and substitution of the SIV LTRs that minimize the potential for genetic reversion while retaining virus particle production, and we have introduced these into a single construct. Knowledge of the elements required for RNA packaging has led us to construct various SIV and HIV-1 genomes that are capable of expression of particles which lack genomic RNA (2, 43). To further improve the safety of these particles, we selected residues that are crucial to the function of reverse transcriptase and integrase as additional targets for viral inactivation (9, 26). The collection of SIV mutated viruses that have been produced is summarized in Table 1. All constructs contain multiple CpG motifs, which are associated with improved immunostimulatory properties of DNA vaccines (reference 28 and references therein).

Various promoters have been engineered to drive expression of these SIV genes, as different promoters might produce different levels of antigen in the various cells targeted by DNA vaccination. The different promoter efficiencies were measured by evaluating genomic viral RNA accumulation by RT-PCR in 293T transfected cells 48 h after transfection (Fig. 1A). The construct pVacc2, containing the EF1a promoter, produced higher levels of RNA than construct pVacc1, containing the cytomegalovirus (CMV) promoter, or pMA22polyA, containing the SIV LTR, and the increased RNA accumulation correlated with the increase in particle production from the transfected cells.

FIG. 1.

Analysis of viral particle production by SIV mutated constructs upon transfection in 293T. (A) RT-PCR analysis of total viral RNA accumulated intracellularly after transfection with SIV constructs. Nucleic acid amplification was carried out with SIV Tat-related primers on total cellular RNA extracted from transfected 293T cells. Lanes show RT-PCR (a) and PCR (b) reactions carried out on the total cellular RNA samples. The PCR control was a PCR reaction carried out on pMA239 DNA (a) and in the absence of template DNA (b). The mock control was a reaction carried out on RNA from supernatant from mock-transfected 293T. Values reported in the panel reflect the evaluation of the bands visualized in the figure. A value of 100 was assigned to the band from pMA239. The intensity of the other bands was calculated by dividing the intensity of each band by the intensity of the pMA239 band. The average values from three independent experiments performed in duplicate ± their standard errors were as follows: pMA239, 100; pMA22polyA, 35.2 ± 1.73; pVacc1, 175.5 ± 22.4; pVacc2, 292.2 ± 34.1. (B and C) Western blot analysis of pelleted particles. In panel A the blot was probed with a macaque SIV polyclonal serum that reacts predominantly with the SIV Env products. In panel B the blot was probed with a macaque SIV polyclonal serum that reacts predominantly with the SIV Gag products. The molecular weights of SIV Env and Gag products are indicated in kilodaltons (kD). (D) RT-PCR analysis of the genomic RNA content of SIV mutated particles. RNA amplification was carried out with SIV Gag-related primers on viral RNA extracted from pelleted virions. Lanes a and b show, respectively, RT-PCR and PCR reactions carried out on the RNA samples. Standards are derived from twofold dilutions of an RNA sample obtained from virions produced by pMA239, which produces wild-type SIVmac239. (E) Electron micrograph of noninfectious particles produced by pVacc1 after transfection into 293T. Images were obtained at a magnification of ×90,000.

The SIV vectors were tested for particle production and infectivity in a tissue culture system. These vectors, containing a total of 22 mutations affecting the function of three essential genes of SIV, efficiently produce particles containing all major SIV proteins and no detectable viral RNA (Fig. 1B and C). The morphology of these mutant particles was consistent with that of immature particles lacking RNA (43). These particles were determined to be noninfectious when tested by a number of different assays on CEMx174 (data not shown).

Evaluation of immune responses to SIV antigens in macaques vaccinated with a genetically inactivated SIV genome.

To evaluate the induction of SIV-specific mucosal and systemic immunity in primates, nine rhesus macaques were inoculated with SIV DNA (groups 1 to 3) and three rhesus macaques were inoculated with the control plasmid pUC19 (group 4) according to the schedule and doses indicated in Materials and Methods. Three different vaccination regimens were used in order to investigate the ability of the DNA vaccine to prime different immunological compartments. The rationale was to compare a relatively simple regimen of immunization to more complex regimens. Because hepatitis B virus is the only chronic virus for which a protective vaccine is available and only one schedule could be investigated with the limited number of animals available, we reasoned that it might be appropriate to follow the hepatitis B virus vaccination schedule.

Various samples were harvested during the course of the immunizations, and the following immunological assays were performed: SIV-specific IgA and IgG in the rectal secretions, SIV-specific IgG, IgA and neutralizing activity in the serum, CTL activity in PBMC, and tetramer staining in Mamu A*01-positive macaques (two of the nine that received the vaccine). The results of the immune response assays prior to live virus challenge are briefly summarized in Table 2. The most striking of all measured immune responses were the levels of virus-specific IgA detected in rectal secretions of animals in group 2, which received the i.d., R regimen.

TABLE 2.

Summary of immune responses to SIV DNA vaccines

Group (n, construct)	Route(s)	No. of animals with positive antibody responses		No. of animals with positive CTLsc	Tetramer stainingd
		SIV IgGa	SIV IgAb
I (3, pVacc1)	i.d.	3	1	3
II (3, pVacc1)	i.d., R	2	3	1	2
III (3, pVacc1/2)	i.d., R, i.m.	3	1	2
IV (3, pUC19)	i.d., R, i.m.	0	0	0

^aSerum samples, 1:100 to 1:2,560 dilution. 

^bRectal secretion samples, 1:23 to 1:2,179 dilution. 

^cSIV-specific CTL activity was scored as positive when ≥5% at any time after vaccination. 

^dTetramer staining of PBMC from two Mamu-A*01-positive animals with 11/C-M Gag peptide was considered positive when it was >0.05%. 

The data shown in Table 3 demonstrate that the administration of a DNA vaccine at the rectal mucosa can stimulate significant SIV-specific IgA responses in primate rectal secretions. Samples from five of nine vaccinated animals were positive at a secretion dilution of 1:23 to 1:2,179. The absence of detectable SIV-specific serum IgA (data not shown) indicated that the IgA was locally produced. No virus-specific IgA was detected in serum samples collected at the same time. The secretions from two animals were also SIV-IgG positive. Analysis of SIV-specific IgA content in secretions collected after the first and second vaccinations indicated that three rectal mucosal doses are necessary to induce significant and consistent SIV-specific IgA levels (data not shown). The magnitude of the increase in SIV-specific IgA content in most of the positive rectal samples was substantially higher than that seen thus far in any other sample analyzed in SIV-vaccinated animals or in animals infected with SIV (29, 33, 37, 38; Kozlowski et al., submitted).

TABLE 3.

SIV-specific IgA antibodies in rectal secretions on day of challenge

Group and animal no.	Route(s)	Fold increasea in:
		SIV IgA	gp130 IgA	SIV IgG
I
19775	i.d.	0.7	6.0*	0b
19796	i.d.	2.2	0	0
19831	i.d.	1.3	2.7	5.2*
II
19777	i.d., R	23.9*	0	0
19786	i.d., R	54.1*	87.5*	17.55*
19821	i.d., R	39.5*	26.2*	0
III
18781	i.d., R, i.m.	0	0	0
18784	i.d., R, i.m.	3.0	1.7	1.3
19856	i.d., R, i.m.	6.6*	0	0
IV (controls)
19783	i.d., R, i.m.	3.2	1.0	1.6
19816	i.d., R, i.m.	0	0	0
19845	i.d., R, i.m.	3.0	1.3	0

^aAn asterisk indicates samples with significant levels of virus-specific IgA antibodies, where significance refers to values of specific activity above the mean of all preimmune samples ± 3 standard deviations (99% confidence interval). The fold increase is defined as the ratio between the specific activity in the postimmunization sample and the specific activity in the preimmune sample. The total IgA concentrations in secretions from significant responders were 1,290 (animal 19775), 23.5 (animal 19777), 34.8 (animal 19786), 3.8 (animal 19821), and 32.9 (animal 19856) μg/ml. 

^b0, no detectable SIV-specific antibodies. 

The i.m. administration of DNA together with rectal and i.d. inoculations appeared to negatively affect the mucosal responses (compare fold increase in animals of groups 2 and 3 in Table 3). This result provides preliminary evidence that simultaneous mucosal and i.m. DNA vaccination cannot stimulate both the systemic and the mucosal arms of the immune system. However, the outcome might be different if simultaneous mucosal and systemic antigenic stimulation is provided by vaccines that are not DNA based or are administered via different routes.

Humoral systemic virus-specific immunity was investigated by measuring SIV-specific serum IgG in an ELISA assay. SIV-specific IgG responses were weak, ranging from 1:100 to 1:2,560 on the day of challenge (Table 4). Neutralization assays carried out with the same samples were negative when SIVmac251 or the challenge virus SIVmac239 was used in the assay. Systemic cell-mediated immunity was investigated by measuring virus-specific CTL activity in PBMC. CTL responses were sporadically present at different levels in different animals (Table 5). Animals with different genetic backgrounds respond differently to a vaccine. One animal (animal 19775) showed a very high level of CTL activity against Env (25%) and Pol (50%) when assayed 2 weeks after the third vaccination, indicating that this vaccine has the potential to stimulate significant cellular responses.

TABLE 4.

SIV-specific serum IgG titers during DNA immunizations and postchallenge

Group and animal no.	Route(s)	Serum IgG titer (titer of neutralizing antibodies)a at:
		Wk 13 (4 wk post-v2)	Wk 25 (16 wk post-v2)	Wk 27 (2 wk post v3)	Wk 29 (2 wk p.c.)	Wk 30 (3 wk p.c.)	Wk 39 (12 wk p.c.)
I
19775	i.d.	N	N	100 (N)	2,560	2,560 (443)	10,240
19796	i.d.	N	N	1,280 (N)	320	100 (N)	20
19831	i.d.	N	N	160 (N)	640	2,560 (3,827)	5,120
II
19777	i.d., R	100	N	640 (N)	5,120	2,560 (313)	10,240
19786	i.d., R	N	N	N (N)	N	100 (58)	5,120
19821b	i.d., R	N	N	100 (N)	20	20 (N)	20
III
19781	i.d., R, i.m.	20	N	320 (N)	2,560	2,560 (310)	100
19784	i.d., R, i.m.	160	20	1,280 (N)	5,120 (1,034)	20,480 (1,589)	20,480
19856	i.d., R, i.m.	80	N	2,560 (N)	5,120 (141)	20,480 (415)	5,120
IV (controls)
19783	i.d., R, i.m.	N	N	N	N	40 (107)	10,240
19816	i.d., R, i.m.	N	N	N	N	40 (334)	5,120
19845	i.d., R, i.m.	N	N	N	N	100 (252)	5,120

^a4 wk post-v2, 4 weeks after the second dose of vaccine. Samples collected 4 weeks after the first and 2 weeks after the second dose of vaccine were SIV antidody negative. Numbers in parentheses indicate the titer of neutralizing antibodies. All samples were tested in neutralization assays with SIVmac239 and SIVmac251. N, negative with both SIVmac239 and SIVmac251; p.c., postchallenge. 

^bAnimal did not become infected after challenge. 

TABLE 5.

SIV-specific cell-mediated immunity during DNA immunizations

Group and animal no.b	DNA route(s)	% CTL activity for Gag/Pol/Env (% tetramer staining)a at:
		Wk 4 (4 wk post-v1)	Wk 8 (8 wk post-v1)	Wk 15 (6 wk post-v2)	Wk 17 (8 wk post-v2)	Wk 25 (16 wk post-v2)	Wk 27 (2 wk post-v2)	Wk 29 (2 wk p.c.)	Wk 31 (4 wk p.c.)	Wk 36 (9 wk p.c.)
I
19775	i.d.	N	N	N	N	5/4/9	–/50/25		40/53/44	NA
19796*	i.d.	N	6/–/–	N	N	N	N		9/–/–	10/–/7
19831	i.d.	10/8/–	N	N	N	20/18/15	7/–/12		20/21/45	21/29/23
II
19777	i.d., R	5/–/–	–/–/12	N	10/7/6	11/9/7	N (0.05)	(10.49)	NA (1.21)	9/9/13 (0.54)
19786	i.d., R	NA	N	N	N	N	N		N	N
19821*	i.d., R	N	N	N	N	N	N (0.12)	(0.1)	N (0.05)	N (0.06)
III
19781	i.d., R, i.m.	12/–/11	–/6/13	11/17/32	–/5/8	–/5/15	N		33/–/–	N
19856	i.d., R, i.m.	–/6/–	N	N	N	N	N		–/7/30	N
19784	i.d., R, i.m.	N	N	N	N	N	N		N	7/–/41
IV (controls)
19783	i.d., R, i.m.	ND	ND	ND	ND	ND	N		–/–/10	NA
19816	i.d., R, i.m.	ND	ND	ND	ND	ND	NA		5/–/46	8/–/28
19845	i.d., R, i.m.	ND	ND	ND	ND	ND	N		9/–/9	9/9/13

^aThe three results reported at each time point for each animal are the percentages of CTL activity for the following different targets: Gag, Pol, and Env. The number indicates the percentage of lysis at highest E/T ratio after background subtraction. The numbers in parentheses indicate the percentages of tetramer staining. Tetramer assays were carried out on PBMC from the two Mamu-A*01-positive animals (animals 19777 and 19821) and from one Mamu-A*01-negative animal. N, negative CTL activity against all three targets; ND, not done; NA, not available because of technical reasons; 4 wk post-v1, 4 weeks after the first vaccination; p.c., postchallenge; –, negative CTL activity against specific target. 

^bAnimals that did not become infected p.c. are indicated by an asterisk. 

Evaluation of protection from the establishment of a persistent SIV infection and/or simian AIDS after mucosal SIV challenge.

The significant levels of IgA antibodies in rectal secretions elicited in all three animals vaccinated i.d., R provided an opportunity for a preliminary evaluation of the role of virus-specific IgA in prevention of infection. Animals in all groups were challenged with live virus 2 weeks after the third immunization with 5,000 50% tissue culture infectious doses of cloned SIVmac239, administered to the rectal mucosa. This amount of virus was equivalent to 9 ng of p27 and is estimated to be approximately 10 rectal macaque infectious doses (AID₅₀) (R. Desrosiers, personal communication). This challenge dose corresponds to 10⁵ AID₅₀ by i.v. titration. Because the small size of the animal groups prevents a meaningful statistical analysis of the challenge results, investigation of larger groups of animals immunized via the mucosal route will be necessary to fully elucidate the role of virus-specific mucosal immunity in the prevention of infection.

Anamnestic IgG responses were observed in all of the animals that were previously SIV-IgG positive and became infected. Seroconversion could be documented in control animals 3 weeks after challenge (Table 4). Clear evidence of an anamnestic neutralizing antibody response was detected in serum from some animals 2 to 3 weeks after challenge, suggesting that priming for neutralization epitopes was induced by the DNA vaccine (Table 4). This anamnestic response was detected with SIVmac251 that is highly sensitive to neutralization but not with SIVmac239. This result was not surprising as SIVmac239 is extremely difficult to neutralize (30, 40).

RT-PCR was carried out to detect RNA viral loads in serum samples from the day of challenge to 25 weeks after challenge (Fig. 2). Cell-associated virus loads were also measured in a limiting-dilution cocultivation assay on a monthly basis (data not shown). In the infected animals, viral loads peaked 2 weeks postchallenge and subsequently decreased. Average viral loads were lower for the group vaccinated intradermally than for the control group, with differences of approximately 10-fold as measured by RT-PCR (Fig. 2), possibly because on the day of challenge CTL responses, which have been associated with viremia control (22, 46), were more consistent and significant in the two i.d. vaccinated animals that subsequently became infected. Two of the nine vaccinated animals (animal 19796 of the i.d. group and animal 19821 of the i.d., R group) remained RT-PCR and PBMC cocultivation negative postchallenge (the last measurement was week 63 postchallenge). PCR analysis of PBMC DNA obtained from samples collected 2 weeks after challenge, when viremia peaked in all infected animals, was also negative in the two animals that resisted challenge (data not shown). PBMC fluorescence-activated cell sorter (FACS) analysis was carried out for the T-cell immunological markers CDw29, CD4, and CD8 (Table 6). CDw29 measures a subpopulation of CD4 cells (memory CD4 cells), and its decline has been observed as an early indicator of the immunological decline that is correlated with subsequent disease progression (16). Two consecutive measurements of this marker that are <10% are considered an indication of incipient immunological decline. SIV-infected animals vaccinated i.d. maintained values of CDw29, CD4, and CD8 within the normal range for a longer period of time than the other infected animals, while a decline affecting in particular the CDw29 marker was evident in most of the other infected animals. Animals 19781 and 19784 in group 3 and animal 19816 in group 4 were diagnosed with an AIDS-related illness and euthanized 41 to 49 weeks after challenge. These data suggest that disease progression might be delayed in the i.d. vaccinated group compared to the other groups.

FIG. 2.

SIV viral loads in macaques challenged rectally by SIVmac239 (serum RT-PCR). Each time point represents the average ± the standard error of the values of viral loads detected in the three animals of one regimen group.

TABLE 6.

PBMC FACS analysis

Group and animal no.a	DNA route(s)	% PBMCb staining with antibodies to:
		CDw29 at wk:				CD4 at wk:				CD8 at wk:
		0	35	47	63	0	35	47	63	0	35	47	63
I
19775	i.d.	16	9	10	9	27	22	23	17	15	24	27	21
19796*	i.d.	17	18	13	18	25	29	30	37	30	39	33	38
19831	i.d.	21	12	8	12	38	35	29	35	30	42	43	43
II
19777	i.d., R	12	5	2	3	20	13	8	8	27	47	35	27
19786	i.d., R	12	5	3	3	20	14	11	8	16	39	36	28
19821*	i.d., R	14	14	11	15	22	30	31	39	33	34	31	29
III
18781†	i.d., R, i.m.	9	8	4	3	13	22	17	16	21	42	40	52
19784	i.d., R, i.m.	20	9	1	2	35	22	5	7	31	37	40	21
19856	i.d., R, i.m.	17	9	6	8	33	28	22	25	28	44	23	41
IV (controls)
19783	i.d., R, i.m.	8	12	4	4	10	35	15	20	13	29	24	22
19816†	i.d., R, i.m.	17	8	4	5	31	51	42	47	15	28	32	42
19845	i.d., R, i.m.	15	7	9	10	29	20	27	35	19	24	26	30

^a∗, Animals that did not become infected after challenge; †, animals that were diagnosed with AIDS and euthanized 9.5 months after challenge. 

^bNumbers indicate the percentages of PBMC that stain with antibodies against CDw29, CD4, and CD8. Two consecutive measurements of the CDw29 marker that were <10% were considered an indication of immunological decline. Week numbers indicate weeks from first immunization. Weeks 35, 47, and 63 thus correspond, respectively, to weeks 8, 20, and 36 after challenge. 

DISCUSSION

We have carried out a small pilot study to investigate whether SIV-specific mucosal immunity can be induced by a DNA vaccine candidate that produces noninfectious virus and whether virus-specific IgA antibodies are a desirable component of a vaccine aimed toward the prevention of mucosally transmitted AIDS. From this study, we can infer the following. (i) A DNA construct that expresses all the SIV proteins except Nef and produces a noninfectious virus due to mutations in three structural genes is a safe and immunogenic reagent in macaques. (ii) This DNA vaccine administered to macaques in a liposome formulation at the rectal mucosa can stimulate significant levels of antigen-specific IgA in mucosal secretions. These levels are higher than those achieved through natural infection. (iii) Virus-specific IgA antibodies present in rectal secretions may have a role in decreasing the infectivity of the initial viral inoculum but alone are unlikely to be sufficient to prevent infection. (iv) Simultaneous DNA vaccination via multiple routes as described here did not result in efficient priming of various immunological compartments.

There are a number of potential advantages to DNA vaccines that produce virus particles compared to vaccines that produce individual antigens. These noninfectious viral particles incorporate the Env proteins and bind the viral receptor on target cells as well as the wild-type virus. Binding of the receptor is thought to trigger conformational changes in the Env protein that may reveal critical neutralizing epitopes on both gp120 and gp41. Exposure of these epitopes may not occur with Env-based vaccine preparations where the Env protein is not part of a virus particle. It is also possible that DNA vaccines expressing viral particles may be more efficient in priming CTLs. Although DNA vaccines that produce virus particles and DNA vaccines that produce individual antigens may prime equally well via the endogenous antigen presentation pathway when targeted to hematopoietic cells, CTL priming might occur more efficiently via the exogenous pathway with particles, once they are taken up by antigen-presenting cells (APCs). This may be particularly important if the cells targeted by the DNA vaccination are nonhematopoietic cells. When the DNA vaccine-derived antigen or the virus-associated antigen is expressed in a nonhematopoietic cell, uptake of antigens by bone-marrow-derived APCs is required to initiate antiviral CTL responses (10, 48). The efficiency of this process might increase with particle-associated antigens.

The potential role of virus-specific mucosal immunity was not clarified in this study. Antibodies in mucosal secretions alone seem unlikely to protect from the establishment of chronic infection, since animal 19786, who had the highest levels of IgA in secretions but no detectable serum IgG, IgA, or CTLs on the day of challenge, became infected. It has been suggested that secreted antibodies might provide the first line of defense against the virus inoculum transmitted at the time of exposure and that local interstitial antibodies and CTLs could act as the second line of defense against virus that enters the mucosa (4, 11–12, 14, 36). Indeed this combination has been suggested as a key to resistance to HIV infection in studies of exposed sex workers in Nairobi and Northern Thailand and discordant heterosexual couples (5, 7, 41). Immunologically mediated containment of the infection during its initial local phase might be possible, and the presence of mucosal immunity could be critical to achieve this goal. A future goal should be to devise practical immunization regimens that are capable of stimulating consistent high levels of both mucosal and systemic immunity.

Two animals in the experiment described here may have been protected from infection by the immune responses induced through vaccination. It is unlikely that the lack of chronic infection in these two SIV-negative vaccinated animals resulted from poor infectivity of the viral stock or natural resistance. The SIVmac239 utilized for the rectal challenge in our experiments is a highly pathogenic virus that quickly establishes high viremia and could easily infect PBMC from these two animals in culture. Protection against this virus has been difficult to achieve, and most successes have been in cases where animals were vaccinated with attenuated viruses (8, 24). The SIVmac239 virus stock used for the rectal challenge has been administered rectally at the same concentration used in our experiment to a total of 15 naive animals so far (Desrosiers, personal communication). All of these animals have become persistently infected. Despite multiple attempts, we were unable to demonstrate any direct or indirect evidence of infection in the two animals that remained virus negative after challenge. In addition, the postchallenge decline of anti-SIV antibody titers in these animals implies a lack of further antigenic exposure and is consistent with the animals being protected from infection. No immunological correlates of protection could be established in these two animals. Nevertheless, because some immunological parameters potentially relevant to protection such as levels of inhibitory chemokines (1) and SIV-specific mucosal cellular responses were not evaluated in this study, it is possible that immune responses induced by the vaccination may have played a role in protecting these animals.

The small size of the groups does not lend statistical power to any conclusions about protection. Nevertheless, it is intriguing that the observed protection of one in three animals in two of the vaccine groups is comparable to the rate of success obtained in other SIV DNA vaccine studies (13, 45). A more thorough evaluation of the role of mucosal immunity in protection from infection will require a more extensive investigation with a larger number of animals. Nevertheless, our ability to easily induce a high level of SIV-specific IgA in rectal secretions by local DNA vaccination provides a simple immunization strategy that could be easily transferred to the clinical setting, if stimulation of mucosal immunity is a desirable feature in an SIV and/or HIV vaccine.