Immunogenicity of a Vaccine Regimen Composed of Simian Immunodeficiency Virus DNA, rMVA, and Viral Particles Administered to Female Rhesus Macaques via Four Different Mucosal Routes

Abstract

A comparative evaluation of the immunity stimulated with a vaccine regimen that includes simian immunodeficiency virus (SIV), interleukin 2 (IL-2), and IL-15 DNAs, recombinant modified vaccinia virus Ankara (rMVA), and inactivated SIVmac239 particles administered into the oral and nasal cavities, small intestine, and vagina was carried out in female rhesus macaques to determine the best route to induce diverse anti-SIV immunity that may be critical to protection from SIV infection and disease. All four immunizations generated mucosal SIV-specific IgA. Oral immunization was as effective as vaginal immunization in inducing SIV-specific IgA in vaginal secretions and generated greater IgA responses in rectal secretions and saliva samples compared to the other immunization routes. All four immunizations stimulated systemic T-cell responses against Gag and Env, albeit to a different extent, with oral immunization providing greater magnitude and nasal immunization providing wider functional heterogeneity. SIV-specific T cells producing gamma interferon (IFN-γ) dominated these responses. Limited levels of SIV-specific IgG antibodies were detected in plasma samples, and no SIV-specific IgG antibodies were detected in secretions. Vaccination also induced CD4⁺ and CD8⁺ T-cell responses in the rectal and vaginal mucosa with greater functional heterogeneity than in blood samples. Rectal T-cell responses were significantly greater in the orally vaccinated animals than in the other animals. The most balanced, diverse, and higher-magnitude vaginal T-cell responses were observed after intestinal vaccination. Significantly higher CD8⁺ granzyme B-positive T-cell responses were observed systemically after intestinal vaccination and in rectal cells after oral immunization. The majority of SIV-specific T cells that produced granzyme B did not produce cytokines. Of the immunization routes tested, oral vaccination provided the most diverse and significant response to the vaccine.

INTRODUCTION

Natural transmission of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) occurs predominantly via mucosal surfaces. Systemic dissemination usually occurs within a few days, and at that point, the intestinal mucosa is also a site of major virus replication and CD4⁺ T-cell depletion in addition to lymphoid organs (1–6). In order to control both entry and systemic dissemination, an effective HIV may need to stimulate both arms of the adaptive immune system, eliciting cellular and humoral immunity systemically as well as at mucosal surfaces. In humans, only a few vaccines are administered via the oral and intranasal route (7). One of the most successful mucosal vaccines has been the polio vaccine, and the live attenuated oral polio vaccine (OPV) is more effective than the inactivated polio vaccine (IPV), which is given intramuscularly (i.m.). The extremely low prevalence of polio in the United States and some risk associated with the use of OPV led to discontinuing it, and since 2000, the IPV has been used in the United States. The OPV is still used in countries with a high prevalence of polio (8, 9). Other examples of vaccines currently in use that are given via the mucosal route are the live-attenuated mucosal vaccines against influenza virus (FluMist), rotavirus, and Salmonella enterica and nonliving whole-cell oral vaccines against Vibrio cholerae and Shigella flexneri (10–16). Different routes for the delivery of mucosal vaccines are being explored; these routes include nasal aerosol, intravaginal, rectal, and sublingual routes (17). In the case of the HIV vaccine, most of the research emphasis is devoted to exploring the intramuscular route of immunization. Thus far, only one vaccine tested in clinical trails and administered intramuscularly has achieved partial protection (31.2% efficacy), the RV-144 ALVAC-HIV (v CP1521) plus AIDSVAX (18), supporting the feasibility of achieving protection but also requiring further improvement.

We have shown that rectal immunizations in rhesus macaques (RM) with SIV DNA/recombinant modified vaccinia virus Ankara (rMVA) vaccine were effective in eliciting virus-specific cellular immune responses systemically and mucosally and also anti-SIV IgA antibodies in rectal secretions, but these humoral responses were sporadic and declined quickly over time. However, protection from progression to AIDS was achieved (19, 20). The same vaccine administered intranasally was more efficient in eliciting cellular and humoral virus-specific responses at mucosal sites than the same regimen administered systemically (i.m.) and provided better protection from disease progression (21). Intranasal immunization with the same vaccine was able to protect from disease progression in female RM following vaginal challenge with SIVmac251 (22). SIV-specific CD4⁺ and CD8⁺ gamma interferon (IFN-γ)-producing T cells present at the time of challenge correlated with the subsequent control of the viremia and longer survival of these animals. This nasal DNA/rMVA vaccine did not stimulate significant humoral responses systemically or in vaginal mucosa (22). Modifications to stimulate greater mucosal antibody responses may further enhance the efficacy of this vaccine, as others have correlated vaccine-mediated induction of SIV-specific mucosal IgA at sites of viral challenge in nonhuman primates (NHP) with sterile protection, delayed acquisition of infection, or reduced viral loads after infection (23, 24). In another study, RM were intranasally vaccinated with a virosome-coupled trimeric gp41 protein that failed to generate systemic antibodies but elicited IgA responses in the genital tract and prevented vaginal simian-human immunodeficiency virus (SHIV) transmission. Phase I clinical trials are under way with intranasal delivery of this HIV-1 candidate vaccine (17). The importance of mucosal IgA humoral responses is highlighted in humans by the detection of HIV-specific IgA in semen or vaginal wash samples collected from some cohorts of sex workers exposed to HIV type 1 (HIV-1) but seronegative, which has been interpreted as an indication that local IgA, induced by viral exposure, can protect during subsequent exposures by interacting with HIV-1 in mucosal secretions (24–26).

Adenovirus (Ad)-based vaccine vectors are one of the most popular platforms for AIDS vaccine development. It has been shown that replicating Ad-HIV/SIV recombinant prime/envelope protein boost is able to elicit broad cellular immunity and functional humoral responses in serum and mucosal sites. Different routes of mucosal immunization with Ad have been explored; these routes include sublingual, intranasal/intratracheal, vaginal, and rectal routes (24, 27–29). The administration of adenovirus type 5 (Ad5)-SIV recombinants by any of these mucosal routes resulted in highly effective priming of systemic immunity as well as mucosal cellular immune responses and mucosal antibody responses after envelope subunit boosts at multiple mucosal sites. These responses correlated with delayed acquisition of SIV infection in the rhesus macaque model.

Few comparative mucosal immunization studies have been performed in humans in order to determine which route of immunization induces the greatest levels of immune responses in the rectum and genital tract, two mucosal sites of HIV entrance and viral replication. In comparative studies that utilized protein immunogens that are efficiently internalized at all mucosal surfaces, it has been established that local immunization produces the greatest rectal or vaginal IgA responses in humans and NHP, but these IgA responses are often limited to the vaccination site (reviewed in reference 5). Nasal immunization typically induces more widely disseminated mucosal IgA responses, and it generates greater systemic IgG responses than does peroral, rectal, or vaginal immunization (5). These differential antibody responses may be less apparent with replicating vaccines. For example, a replication-competent adenovirus type 5 vector with SIV genes was recently reported to induce SIV-specific mucosal IgA responses in multiple mucosal tissues regardless of whether it was administered to female macaques by the sublingual, rectal, vaginal, or nasal plus intratracheal mucosal routes (24). However, in this study, the vaginal immunization route was found to generate the greatest SIV-specific vaginal T-cell responses (24), suggesting that strong stimulation of cellular responses in the genital tract may require local vaccination.

We previously explored the rectal and intranasal routes of immunization, but we never directly compared multiple mucosal vaccination routes in the same study. This study aims to compare the immune responses elicited systemically and at mucosal sites of HIV entry by a vaccine composed of SIV, interleukin 2 (IL-2), and IL-15 plasmid DNAs followed by boosts with SIV Gag/Pol/Env MVA and inactivated SIVmac239 particles administered at four different mucosal surfaces. The simultaneous comparison of oral, intestinal, intranasal, and vaginal mucosal routes of immunization could be helpful to determine which immunization route is the most effective at eliciting diverse humoral and cell-mediated virus specific responses that could be needed in a preventive HIV vaccine.

MATERIALS AND METHODS

Vaccine components.

The DNA plasmid pVacc6 used in the vaccination is a derivative of pVacc1 (20). pVacc6 includes a full SIVmac239 genome with multiple mutations in the NC basic domain and in the functional domains of RT, INT, and PR and a stop codon at the beginning of the vpr gene. Gene expression is under the control of the cytomegalovirus (CMV) promoter, with deletions of both long terminal repeats (LTRs). The DNA sequence was confirmed by sequencing, and the profile of viral particle produced was evaluated by 293T transfection and subsequent Western blotting using macaque SIV-positive sera. The IL-2/Ig and IL-15 plasmids were previously described (30, 31). IL-2 and IL-15 production was tested in supernatants of transfected 293T cells by an enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN). rMVA expressing SIV Gag-Pol and Env proteins was prepared as previously described (32, 33). 2,2″-Dithiodipyridine (aldrithol-2 [AT2])-inactivated SIVmac239 particles were provided by Jeff Lifson and were previously described (34–36).

Experimental groups and vaccination schedule.

Female rhesus macaque monkeys were cared for at the New England Regional Primate Research Center using approved protocol under the guidelines established by the Animal Welfare Act (37) and the Guide for the Care and Use of Laboratory Animals (38). Twenty-eight animals, divided into four groups, received the vaccine as follows: group 1 (n = 7), oral administration; group 2 (n = 7), intestinal delivery; group 3 (n = 7), nasal administration; and group 4 (n = 7), vaginal administration. Each animal received a total of three DNA doses on day 1, week 8, and week 24 that consisted of 1 mg pVacc6, 0.5 mg IL-2/Ig, and 0.5 mg IL-15 DNA, formulated in 1 ml of 20 mM DOTAP (1,2-dioleoyl-3-trimethylammonium-propane) (Encapsula Nanosciences) and cholesterol (1:1). The DNA was administered as follows: for oral immunization, the vaccine solution was applied in the oral cavity to the cheek mucosa, while the anesthetized animal laid on one side; for intestinal immunization, a gastric endoscope was used to deliver the vaccine to the duodenum; for nasal immunization, a total of 500 μl of liposomes containing the same amount of DNA as for the other routes was applied to each of the two nostrils in a 250-μl volume (the volume was changed to avoid inhalation of excessive fluids, as the nasal cavity is small); for vaginal immunization, the DNA was applied directly in the vaginal lumen. A total of 10⁹ PFU of rMVA, expressing SIV gag, pol, and env genes, in phosphate-buffered saline (PBS) were mucosally delivered on week 32. On weeks 48 and 50, each group received a dose of 2,2″-dithiodipyridine (aldrithol-2 [AT2])-inactivated SIVmac239 particles equal to 300 μg total proteins, and corresponding to 50 μg of p28^CA, with 100 μg of heat-labile toxin (LT) of Escherichia coli (39) in PBS in a final volume of 350 μl. For vaginal immunization with the inactivated SIV particles, 350 μg of Eldexomer in PBS were mixed with particles and LT in a final 700-μl dose (Perstop Specialty Chemical AB) (40). Each group received the inactivated particles on the same mucosal surface where they had been previously immunized. One animal per group was Mamu-A*01 positive (Mamu-A*01⁺) (animals 248, 258, 145, 265, and 301), animal 249 was Mamu-B*08⁺, and animal 265 was also Mamu-B*17⁺, as determined by PCR (41–43).

Analysis of SIV-specific IgA in rectal and vaginal secretions and IgG in plasma samples.

Rectal and vaginal secretions were collected using Weck-Cel sponges premoistened with 50 μl Dulbecco's phosphate-buffered saline (DPBS) as described previously (44). Saliva samples were collected by placing 2 dry sponges between the cheek and gum, near the back of the mouth, for a minimum of 5 min. To collect nasal secretions, 100 μl of DPBS was first dropped into one of the animal's nostrils using a pipetman. The nostril was immediately pressed shut and massaged gently. A premoistened sponge was then inserted, allowed to absorb fluid for 5 min, and then removed. The procedure was then repeated for the second nostril. Sponges with secretions were stored at −70°C. Secretions were eluted from the sponges by centrifugation as described previously (44). Antiviral and total IgA and IgG in secretions and IgG in serum were measured in an ELISA as described previously (19, 21). Briefly, for the SIV ELISAs, microtiter plates were coated overnight with 100 ng/well SIV recombinant gp140mac239 (rgp140mac239) (Immune Technology, New York, NY) or 100 μl per well of 1/400 SIVmac251 viral lysate (Advanced Biotechnologies Inc., Columbia, MD) that lacked detectable envelope protein at this dilution. Reactivity detected against the SIV lysate is referred to as being against SIV Gag/Pol proteins, as no reactivity against Env is detected at the lysate dilution used to coat the plate. For SIV assays, the standards were pooled serum samples from SIV-infected macaques. All SIV standards were calibrated relative to the total IgA or IgG serum standard by coating portions of the same plate with the SIV antigen, goat anti-monkey IgA (Rockland Immunochemicals) or goat anti-monkey IgG (MP BioMedicals). The plates were developed using biotinylated goat anti-monkey IgA (Rockland) or human IgG (Southern Biotech), neutravidin-labeled peroxidase (Pierce), and 3,3′,5,5′-tetramethylbenzidine (TMB) substrate and stop solution (Southern Biotech). The concentration of anti-SIV IgA or IgG in each secretion was subsequently divided by the concentration of total IgA or IgG to obtain the specific activity (nanogram of anti-SIV antibody per microgram of immunoglobulin). The secretion was considered positive for SIV antibody if it had a specific activity that was greater than the mean specific activity plus 3 standard deviations (SD) in secretions from naïve animals. The concentrations of SIV-specific IgG in plasma were considered significant if they were 3.4-fold greater than that measured in the animal's preimmune plasma sample. If a preimmunization secretion had no detectable IgA antibody (Ab), it was assigned the mean specific activity value of naïve macaques.

Isolation of intestinal lymphocytes.

Isolation of mononuclear cells (MNC) from colon-rectal mucosa was carried out according to previously published procedures (45). Briefly, after Telazol anesthesia, seven or eight biopsy specimens/animal/time point were obtained from the rectum and cervicovaginal tissue using sterile forceps and a small pinch biopsy device (Olympus endoscopic biopsy forceps). MNC from colon-rectal or vaginal biopsy specimens, and peripheral blood mononuclear cells (PBMC) were isolated according to previously published procedures (45).

Intracellular staining (ICS) and antibodies.

MNC (10⁵ cells) and PBMC (10⁶) were incubated for 14 h with medium alone (unstimulated) or 1 μg/ml pool of 15-mer SIV Gag or SIV Env peptides. For a positive control, the cells were incubated with 10 ng/ml PMA (4-α-phorbol 12-myristate 13-acetate; Sigma) and 1 μg/ml ionomycin (Sigma); unstimulated cells provided background controls. All cultures contained brefeldin A (BD GolgiPlug [catalog number 555029; BD Biosciences]) and 1 mg/ml anti-CD49d and anti-CD28. The PBMC and MNC were stained and evaluated for expression of cytokines according to previously described procedures and reagents (19). For MNC, 200 μl of a viability dye (VIVID, LIVE/DEAD kit, Invitrogen) was added to the antibody cocktail to exclude dead cell background. The CD3⁺ cells were used as the gate for CD4⁺ or CD8⁺ cells. Data for peptide-stimulated populations are reported as a percentage, determined after subtracting the percentage of positive cells detected in unstimulated cells for each sample.

The Abs used in this study were Mab11 (antibody against tumor necrosis factor alpha [anti-TNF-α]), B27 (antibody against gamma interferon [anti-IFN-γ]), MQ1-17H12 (anti-interleukin 2 [anti-IL-2]), L200 (anti-CD4), SP34-2 (anti-CD3), GB11 (anti-granzyme B) (BD Biosciences, Pharmingen, San Diego, CA), and SK1 (anti-CD8) (Caltag, Invitrogen Corporation, Carlsbad, CA). The acquisition of the flow cytometry data was performed using MoFlo (Dako North America, Inc., Carpinteria, CA). Flow cytometry data were analyzed using FlowJo v9.1 (TreeStar, Ashland, OR).

Quantitative analyses of blood-derived immune cells.

Characterization of CD4⁺ and CD8⁺ T cells in PBMC was conducted according to previously published procedures (46).

Statistical analysis.

Calculations and statistical analyses were performed using GraphPad Prism version 3 software. The two-tailed Fisher's exact test was used to compare the frequency of IgA responses between groups. Comparisons between groups were carried out by two-tailed t test or Mann-Whitney test; for comparisons among four groups, one-way analysis of variance (ANOVA) and Bonferroni posttest was used. The results of statistical analyses were considered significant if they produced P values of ≤0.05. Display of multicomponent distributions was performed with SPICE v5.2 (freely available from http://exon.niaid.nih.gov/spice/) (47).

RESULTS

Study design.

To determine the most efficient mucosal route of immunization to stimulate the most diverse vaccine-induced immunity at different mucosal and systemic sites with the potential to protect rhesus macaques from SIV infection or to delay disease progression, four groups, each with seven female rhesus macaques, were vaccinated with 3 doses of SIV/cytokine DNAs in liposomes on day 1, week 8, and week 24, followed by one dose of SIV Gag/Pol/Env rMVA on week 32, previously tested in male and female animals (21, 22). On week 48 and 50, all animals were boosted with 2,2′-dithiodipyridine (aldrithol-2 [AT-2])-inactivated SIVmac239 particles equivalent to 50 μg of p28^CA (35). The IL-2/IL-15 combination is known to broadly stimulate cell-mediated immune responses, affect their longevity in mice, and stimulate immune responses in general (48, 49). The two boosts of inactivated SIVmac239 virions were added to improve humoral responses, since protein immunization has been shown to better stimulate antibody responses. Immunizations with AT-2-inactivated SIV virions were effective in inducing both binding and neutralizing antibody responses along with cellular immune responses (34–36). Group 1 received the entire regimen orally, and more specifically in the oral cavity, while group 2 received it intestinally (in the duodenum), group 3 received it nasally, and group 4 received it intravaginally.

Virus-specific humoral responses during immunization.

Systemic and mucosal antibody responses were evaluated using plasma samples and secretions that were collected 2 to 5 weeks after each vaccination. The mucosal vaccinations did elicit mucosal IgA responses to SIV antigens in many animals (Fig. 1), albeit of low magnitude. Regardless of the immunization group, peak mucosal IgA responses to Env were more often observed after the last DNA immunization or after the MVA boost than after the SIV particle boost (Fig. 1A to C). However, in 5/7 nasally immunized macaques, there were greater levels of Env-specific IgA in week 54 rectal secretions than in week 52 secretions (Fig. 1A). Thus, the SIV particle boosts may have stimulated or increased anti-Env IgA responses in some animals. This result is surprising, given that anti-Env IgA in nasal secretions from nasally vaccinated animals were not equally boosted by the particle vaccination (not shown). Nasal immunization induced moderate levels of anti-Env IgA responses in nasal secretions (not shown).

Fig 1.

Anti-SIV humoral (IgA) responses in secretions during immunization. (A to C) IgA antibodies to SIV gp140 Env and SIV lysate Gag and Pol proteins were measured by an ELISA performed on rectal secretions (A), vaginal secretions (B), and salivary secretions (C) collected at intervals after immunization by each mucosal route. To compare the magnitude of the mucosal IgA responses generated in animals, the SIV antibody concentration in each secretion was normalized against the total IgA concentration and is expressed as specific activity (nanogram of SIV-specific IgA per microgram of total IgA). Graphs on the left depict the specific activity in individual animals at time points corresponding to 2 and 4 weeks after the third DNA immunization (weeks 26 and 28), 2 weeks after the MVA immunization (week 34), and 2 and 5 weeks after the second viral particle immunization (weeks 52 and 54). The specific activity had to fall outside the shaded portion in the graphs to be considered statistically significant. Symbols in the shaded region may be overlapping, but all symbols above this have been separated to permit visualization of positive responders. Peak IgA responses measured during this phase of the study are illustrated in the Tukey graphs on the right. P values were obtained using the Mann-Whitney rank sum test. Int, intestinal; Nas, nasal; Vag, vaginal. (D) The percentage of animals in each group that were found to have significant IgA responses to both Env and Gag/Pol at the mucosal site is shown. (E) The number of animals that demonstrated a significant Env or Gag/Pol IgA response at all three or fewer mucosal sites is presented for each immunization group. Animal identification (ID) numbers for the four immunization groups were as follows: oral, animals 153, 247, 248, 249, 253, 255, and 256; intestinal, animals 146, 155, 246, 257, 258, 259, and 260; nasal, animals 145, 147, 148, 149, 150, 156, and 250; and vaginal, animals 151, 152, 154, 157, 158, 254, and 265.

In these small study groups, significant differences in the magnitude and frequency of SIV-specific mucosal IgA responses between immunization groups often could not be demonstrated. However, it is noteworthy that animals found to have the greatest rectal IgA responses were in the oral immunization group (Fig. 1A), and oral immunization proved just as effective as local vaginal immunization for induction of SIV-specific IgA in vaginal secretions (Fig. 1B). Animals immunized in the oral cavity also tended to have broader mucosal IgA responses to SIV, scoring positive in both the SIV Env and Gag/Pol assays (Fig. 1D), and oral immunization induced IgA responses to the SIV antigens at multiple mucosal sites (Fig. 1E).

In our previous nasal immunization studies with pVacc6 DNA and SIV MVA, systemic IgG responses were not induced (21, 22). In this study, plasma anti-SIV IgG antibody concentration increased after particle immunization against Gag/Pol antigens in one animal vaccinated intestinally, four animals vaccinated nasally, and two animals vaccinated vaginally (Fig. 2A). Antibodies against SIV Env were observed in one animal vaccinated intestinally and two vaccinated nasally (Fig. 2B). In some animals, the increase at week 52 or 57 was significant compared to the value for the preimmune plasma samples, but these values remained within the shaded area because the average plus 3 SD used to set the cutoff was high because of high background levels in some animals (Env concentrations in animals 249, 253, and 153 and Gag/Pol concentrations in animals 146 and 147). Anti-SIV IgG antibodies were not detected in secretions (not shown).

Fig 2.

Anti-SIV plasma IgG concentrations during the immunization regimen. (A and B) Plasma levels of IgG antibodies against SIV Gag/Pol (A) and gp140 (B) were measured by ELISA on the day of the first DNA immunization (day 1), 2 weeks after the second particle immunization (week 52), and at the end of the immunization regimen (week 57). The concentration, measured in nanogram/milliliter for individual animals at the three time points, is reported. The cutoff was calculated as the preimmune sample mean plus 3 SD, and it is 241 ng/ml for Gag/Pol and 111 ng/ml for Env. The concentration had to fall outside the shaded region and be higher than the value for the preimmune sample to be considered significant. Animal identification (ID) numbers for the four immunization groups were as follows: oral, animals 153, 247, 248, 249, 253, 255, and 256; intestinal, animals 146, 155, 246, 257, 258, 259, and 260; nasal, animals 145, 147, 148, 149, 150, 156, and 250; and vaginal, animals 151, 152, 154, 157, 158, 254, and 265.

These results indicate that a mucosal particle immunization added to the DNA/MVA regimen can stimulate or expand humoral responses both mucosally and systemically. A higher dose than that used, which was relatively low, may result in more consistent and higher responses among animals.

Systemic virus-specific cell-mediated responses.

Cell-mediated responses against SIV are important to control viremia and provide protection from disease progression (19, 21, 22, 50, 51). Thus, one of the aims of this study was to characterize the magnitude and function of the vaccine-induced cell-mediated immunity against SIV. Virus-specific T-cell responses were measured at 2- or 4-week intervals in the peripheral blood mononuclear cells (PBMC). Antigen-specific CD4⁺ or CD8⁺ T-cell responses were measured as a percentage of T cells producing IL-2, IFN-γ, and TNF-α in PBMC after stimulation with SIV Gag or Env peptide pools by intracellular staining and flow cytometry. SIV-specific IFN-γ⁺, IL-2⁺, TNF-α⁺, or triple-positive CD4⁺ and CD8⁺ T-cell populations are reported as a percentage of the total CD4⁺ or CD8⁺ T-cell populations. The numbers obtained by adding single and multiple cytokine-positive, CD4⁺ or CD8⁺ T-cell responses against SIV antigens (Fig. 3A and D columns [Gag plus Env], Fig. 3B and E columns [Gag], Fig. 3C and F columns [Env]), and averaging the group values were plotted for each group (top row of graphs in Fig. 3) during the time course of the immunization, as well as for each individual animal in each group (Fig. 3, second to fifth rows of graphs). Immunization given via all four routes was able to elicit systemic T-cell responses against Gag or Env (Fig. 3), albeit to a different extent in each group. Oral and intestinal delivery of the vaccine appeared to be more effective than nasal or vaginal vaccination in eliciting virus-specific CD4⁺ and CD8⁺ T-cell responses. Positive responses were detected after each immunization and at a much higher level after rMVA and inactivated SIV particle boostings (Fig. 3). However, these differences did not reach statistical significance, probably because of the small size of the groups and the variability in the magnitude of the responses among animals within the same groups. Significant, but more moderate compared to the other groups, CD4⁺ and CD8⁺ T-cell responses against Gag or Env were detected only after SIV particle boosting in the animals immunized nasally or vaginally, suggesting that this vaccine given via the nasal or vaginal mucosa is less effective in stimulating systemic immune responses than when given orally or in the duodenum. Antigen uptake in the nasal and vaginal mucosa may be less efficient than when given orally or in the small intestine, or alternatively, there might be more limited recirculation of T cells between these mucosal sites and the systemic compartment. The fact that higher level IgA responses were detected after the nasal immunization in the rectal secretions appear to make the first interpretation less likely and favor the second. Similar levels of anti-Gag and anti-Env CD4⁺ and CD8⁺ T-cell responses were elicited by the vaccine regardless of the immunization route, excluding the possibility that one antigen might be immunodominant over the other. Boosting with SIV particles further expanded preexisting CD4⁺ and CD8⁺ T-cell responses in blood, supporting diverse boosting as a strategy to continue to expand the preexisting response. Responses to these two immunizations varied among the animals. In some animals, the values declined after the peak induced by the first immunization; in some animals, the values increased further after the second particle immunization and declined at subsequent time points; and in some animals, these responses had individual peaks after each particle immunization. Given that these immunizations were all mucosal, it is possible that recirculation of antigen-positive cells from the site of immunization to other mucosal compartments has different kinetics after each kind of mucosal vaccination, and what is detected in blood represents the extent of this recirculation. When two immunizations are administered so close to each other, it is also possible that the effector cells expanded by the first immunization quickly eliminate the cells presenting the antigen given with the second immunization, rather than stimulating a further expansion, and that better results would be observed by waiting longer before the second particle immunization.

Fig 3.

Circulating cell-mediated total SIV and Gag- and Env-specific T-cell responses during immunization phase. The geometric mean of each vaccinated group is shown during the immunization phase as a total percentage of CD4⁺ T-cell (A to C) and CD8⁺ T-cell (D to F) responses measured in PBMC as production of IFN-γ, TNF-α, and IL-2 and their combination by ICS and flow cytometry analysis upon stimulation with Gag or Env peptide pools. The graphs in columns A and D show the total SIV-specific T-cell responses (Gag plus Env) for vaccinated groups and for each individual animal of the vaccinated group, while the graphs in columns B and E illustrate the T-cell responses elicited against Gag and the graphs in columns C and F illustrate the T-cell responses elicited against Env during immunization until the end of the immunization phase. The immunization route of the groups are shown as follows: orally vaccinated animals are represented in blue, intestinally vaccinated in red, nasally in orange, and vaginally in light blue. Statistical significance between groups was tested by the one-way ANOVA assay (P < 0.05). (G) Qualitative analysis of systemic SIV Gag- and Env-specific cell-mediated responses for each vaccinated group 2 weeks after the third DNA immunization (week 26), 2 weeks after the boost with rMVA (week 34), 2 weeks after the second boost with inactivated SIVmac239 particles (week 52), and 7 weeks after the last immunization (week 57). Pie graphs representing the diversity of CD4⁺ (left) and CD8⁺ (right) T-cell responses in production of IFN-γ, TNF-α, and IL-2 and the simultaneous production of 2 or more cytokines combined are shown for each of the four time points. The total mean percentage SIV-specific response for each group is shown on the upper left side of each pie. Animal identification (ID) numbers for the four immunization groups were as follows: oral, animals 153, 247, 248, 249, 253, 255, and 256; intestinal, animals 146, 155, 246, 257, 258, 259, and 260; nasal, animals 145, 147, 148, 149, 150, 156, and 250; and vaginal, animals 151, 152, 154, 157, 158, 254, and 265.

While the magnitude of the vaccine-induced responses has been shown to be important for a vaccine to be efficient, the functional quality of these responses may also be relevant to confer protection from infection or disease progression (52–54). Analysis of cytokine production in systemic SIV-specific CD4⁺ and CD8⁺ T-cell responses is shown in Fig. 3G. Production of IL-2, TNF-α, and IFN-γ was evaluated and reported as the average percentage of single, double, and triple cytokine-positive cells detected after SIV Gag or Env peptide stimulation in each group. These responses were analyzed 2 weeks after the third DNA dose, the MVA boost, and after the second boosting with inactivated SIV particles, and 7 weeks after the last immunization, on week 57. This analysis revealed that these responses were mainly dominated by SIV-specific cells producing IFN-γ at most of the investigated time points (Fig. 3G). Three doses of DNA immunization induced predominantly antigen-specific CD4⁺ and CD8⁺ T cells that produced IFN-γ in all groups (Fig. 3G, week 26 graphs), with the oral vaccination providing the most heterogeneous response in both populations. Interestingly, mucosal boosting with rMVA appears to expand both the magnitude of the responses and the heterogeneity of the cytokines produced by them (Fig. 3G, week 34 graphs). Boosting with inactivated particles, also given via each of the mucosal routes, significantly expanded the pool size of antigen-specific cells, in particular CD8⁺ T cells, and the multifunctional phenotype of these cells (Fig. 3G, week 52 graphs). This was particularly true for the vaginal immunization, despite the more limited size of the SIV-specific population in these animals. Seven weeks after the last immunization, when the vaccine-induced responses had contracted and returned to memory, production of IFN-γ was mostly detected in these cells, with the exception of SIV-specific CD4⁺ T cells in nasally and vaginally vaccinated animals (Fig. 3G, week 57 graphs). The size of the SIV-specific memory pool was approximately twice as big in the orally vaccinated animals compared to the other groups. We conclude that all the four mucosal routes of immunization we explored manage to stimulate significant SIV-specific systemic T-cell-mediated responses, with oral immunization providing the best results in terms of the magnitude of the response and nasal immunization providing the best results in terms of functional heterogeneity of the antigen-specific cells.

Vaccine-induced CD4⁺ and CD8⁺ T-cell responses in rectal and vaginal mucosa.

We evaluated the ability of the vaccine candidate to induce SIV-specific T-cell responses in the rectal and vaginal mucosa during the time course of the immunization regimen by investigating the number of SIV-specific mucosal mononuclear cells (MNC) obtained from colorectal and cervicovaginal biopsy specimens collected during immunization. Duodenum biopsy specimens were also obtained for the oral and intestinal immunization groups. Samples were collected at three time points during the immunization, when the responses were likely to approximate their peak after a specific immunization: 2 weeks after the third DNA vaccination, 2 weeks after MVA boosting, and 4 weeks after the second immunization with SIV particles. We found that the mucosal administration of the vaccine successfully stimulated CD4⁺ and CD8⁺ T-cell responses at both sites (Fig. 4 and 5).

Fig 4.

Rectal SIV-specific cell-mediated T-cell responses during immunization. (A and B) SIV-specific CD4⁺ (Gag [A] and Env [left graph in panel B]) and CD8⁺ T-cell responses (Gag [A] and Env [right graph in panel B]) were detected in MNC isolated from rectal biopsy specimens 2 weeks after the third DNA dose (week 26), 2 weeks after the boost with rMVA (week 34), and 4 weeks after the boost with inactivated SIVmac239 particles (week 54) and are represented as a bar graph for each animal in all four vaccinated groups. Each bar represents the sum of anti-Gag or anti-Env-specific T cells producing IFN-γ, TNF-α, and IL-2 alone or as a double- or triple-cytokine combination, after stimulation with Gag or Env peptide pools. (C to E) Diversity of rectal SIV-specific cell-mediated T-cell responses during immunization. Pie graphs illustrate the diversity of Gag-specific (C) and Env-specific (D) T cells producing IFN-γ, TNF-α, and IL-2 for each group detected at 3 time points, after the third DNA dose (week 26), 2 weeks after the MVA boost (week 34), and 4 weeks after the boost with inactivated SIVmac239 particles (week 54). The SIV-specific T-cell mean percentage of the total CD4⁺ or CD8⁺ T cells is shown for each group beneath each pie. (E) Total mean percentage of SIV-specific Gag⁺ plus Env⁺ CD4⁺ and CD8⁺ T-cell response detected 6 weeks after the last immunization is shown for each vaccinated group. Statistical significance was evaluated by one-way ANOVA and Bonferroni posttest. A P of <0.05 was considered statistically significant.

Fig 5.

Vaginal SIV-specific cell-mediated T-cell responses during immunization. (A to D) SIV-specific CD4⁺ (Gag [A] and Env [B]) and CD8⁺ T-cell responses (Gag [C] and Env [D]) were detected in MNC isolated from cervicovaginal biopsy specimens 2 weeks after the third DNA dose (week 26), 2 weeks after the boost with rMVA (week 34), and 4 weeks after the last boost with AT-2-inactivated SIVmac239 particles (week 54). The T-cell responses are represented as a bar graph for each animal in all four vaccinated groups. Each bar represents the sum of anti-Gag or anti-Env-specific T cells producing IFN-γ, TNF-α, and IL-2 alone or as double- or triple-cytokine combination, after stimulation with Gag or Env peptide pools. (C to E) Diversity of vaginal SIV-specific cell-mediated T-cell responses during immunization. Pie graphs illustrate the diversity of SIV Gag-specific (C) and Env-specific (D) T cells producing IFN-γ, TNF-α, and IL-2 for each group at 3 time points, 2 weeks after the third DNA dose (week 26), 2 weeks after the MVA boost (week 34), and 4 weeks after the boost with inactivated SIVmac239 particles (week 54). (E) Mean percentage of SIV-specific Gag⁺ Env⁺ CD4⁺ and CD8⁺ T-cell response detected 6 weeks after the last immunization is shown for each vaccinated group. Statistical significance was evaluated by one-way ANOVA and Bonferroni posttest. A P value of <0.05 was considered statistically significant. Animal ID numbers for the four immunization groups were as follows: oral, animals 153, 247, 248, 249, 253, 255, and 256; intestinal, animals 146, 155, 246, 257, 258, 259, and 260; nasal, animals 145, 147, 148, 149, 150, 156, and 250; and vaginal, animals 151, 152, 154, 157, 158, 254, and 265.

When SIV-specific responses were evaluated in rectal MNC, the oral administration of rMVA and inactivated SIV particles was more effective at eliciting higher levels of Gag- and Env-specific CD4⁺ and CD8⁺ T-cell responses than the same immunizations via the other routes (Fig. 4E, P = 0.01 and P = 0.007, respectively, by ANOVA). Vaccination with SIV particles did not expand the magnitude of the responses observed after rMVA vaccination in the majority of the animals, and this was particularly true for the CD8⁺ T-cell responses (Fig. 4B, week 54). The assessment of the functional heterogeneity of the responses showed that monofunctional T cells producing IL-2, TNF-α, or IFN-γ were also predominant in SIV-specific rectal T cells, but they were more balanced than in the systemic compartment, with less polarization toward production of IFN-γ (Fig. 4C and D). Although SIV particle immunization did not further expand the levels of responses observed after rMVA immunization, it did modify the pattern of produced cytokines. After the last immunization, the overall magnitude of the SIV-specific rectal responses was significantly greater in the orally vaccinated animals than in all the other animals (Fig. 4E), and among Gag-specific CD8⁺ T cells, dually functional cells were well represented (Fig. 4D).

A similar evaluation was performed at the vaginal mucosa, another site of virus exposure in humans (Fig. 5). Overall, T-cell responses at this site were higher than in the rectal mucosa. The peak of anti-Gag and anti-Env CD8⁺ T-cell responses was observed after the three doses of DNA in the animals vaccinated nasally and vaginally, with no efficacy of the rMVA immunization and limited efficacy of the SIV particle immunization via these routes (Fig. 5A and B). It is possible that peak responses after boosts administered at these sites occurred later and were missed. In the animals vaccinated orally or intestinally, the responses were of greater magnitude at the time point analyzed after the last immunization, with the exception of the CD8⁺ response in the orally vaccinated animals which progressively decreased during the time course of the regimen and was at its lowest levels after the last immunization. As noticed for the rectal responses, a greater heterogeneity of function was observed in the vaginal SIV-specific T-cell responses, with a larger fraction of cells producing IL-2 and TNF-α than in blood (Fig. 3G). By the time the regimen was completed, the most balanced, diverse, and higher-magnitude vaginal T-cell response was observed in the animals vaccinated intestinally (Fig. 5E, bottom graph), with a statistically significant number of SIV-specific CD8⁺ T cells compared to those of the other groups (P = 0.03 by ANOVA). The CD4⁺ T-cell response was of significant size in the orally vaccinated animals, but antigen-specific CD8⁺ T cells were barely detected (Fig. 5E, top graph). These results suggest that the CD8⁺ antigen-specific T cells, detected after the third DNA oral vaccination, were predominantly effector T cells with a small component of memory cells. A direct evaluation of number and function of effector and memory cells was not carried out because of limitations in cell numbers obtained from the biopsy specimens. When the same analysis was carried out in MNC isolated from duodenal biopsy specimens, responses of similar magnitude and function were detected in both orally and intestinally vaccinated animals, and the magnitude of the responses was intermediate of those observed for the same groups vaccinated in the rectum (not shown).

Cytotoxic CD4⁺ and CD8⁺ T-cell responses after the immunization regimen.

The cytotoxic capability of the anti-SIV responses induced by the vaccine was investigated by measuring the production of granzyme B by CD4⁺ and CD8⁺ T cells after antigen-specific stimulation of PMBC and rectal and vaginal MNC during the immunization time course. Production of three cytokines, IFN-γ, IL-2, and TNF-α, was also evaluated in CD4⁺ or CD8⁺ T cells that were positive for granzyme B. The results observed in PBMC samples on week 57 or in rectal and vaginal MNC on week 54, 4 weeks after the last immunization, are reported as a percentage of CD4⁺ or CD8⁺ T cells producing granzyme B alone or in combination with one of the cytokines (Fig. 6). The number of cells positive for granzyme B and two or three cytokines was barely above detection, and it is not reported. In general, the majority of the antigen-specific CD4⁺ or CD8⁺ T cells that produce granzyme B did not produce other cytokines, with only a small fraction of these cells simultaneously producing both. In PBMC, significantly higher percentages of CD8⁺ granzyme B⁺ T cells were observed in intestinally vaccinated animals than in the other vaccinated groups (Fig. 6A, P = 0.03 for intestinal versus nasal values), although the opposite was true for CD4⁺ granzyme B⁺ cells (P = 0.02 for nasal versus intestinal values). Oral immunization was the most efficient at stimulating SIV-specific granzyme B⁺ cells in rectal MNC, since significantly higher numbers of CD4⁺ and CD8⁺ T cells producing granzyme B were detected in these animals compared to the other vaccinated groups (Fig. 6B). No significant differences were observed among groups when the same evaluation was carried out in cells from vaginal biopsy specimens (Fig. 6C).

Fig 6.

Analysis of the vaccine-induced SIV-specific cytotoxic T-cell responses in circulating PBMC and rectal and vaginal MNC at the end of the immunization regimen. (A to C) Granzyme B (GB) production in combination with IFN-γ, TNF-α, and IL-2 was measured in circulating PBMC (A) and rectal (B) and vaginal (C) MNC at the end of the immunization regimen. The variety of total SIV-specific responses in CD4⁺ and CD8⁺ T-cell population is represented as pie graphs for each vaccinated group, and the total mean percentage is shown for each group beneath each pie graph. The total SIV-specific CD4⁺ (left graph) or CD8⁺ (right graph) mean production of granzyme B and its combination with the three cytokines analyzed is represented as bar plots for each vaccinated group. Two-way ANOVA and Bonferroni tests were performed to analyze statistical significance. A P value of < 0.05 was considered statistically significant.

DISCUSSION

The main function of the mucosal immune system is to recognize and protect against the majority of microbial pathogens, forming the primary barrier against pathogens. This barrier is comprised of specialized mucosal immune cells that accumulate or are in transit between various mucosa-associated lymphoid tissues (MALT) (2, 55, 56). The MALT is a highly compartmentalized immunological system that functions independently of the systemic immune apparatus. It is comprised of anatomically defined sites that serve as mucosal inductive sites for initiation of immune responses. Because these sites are anatomically separated but functionally connected, the induction of an immune response at one site can lead to an effector response at a different mucosal site mediated by homing receptors on induced memory T and B cells (7, 57). This feature is important for the development of vaccine against a mucosal pathogen such as HIV, since effective immune responses able to protect from infection and control viral replication may be needed at different anatomical locations as well as at systemic sites of virus spreading and replication. In this sense, the different mucosal routes of immunization, such as the oral, intranasal, rectal, and vaginal routes, may play an important role, since they can induce generalized mucosal immune responses not only at the site where they are administered but at distal mucosal effector sites as well. It is well-known that intranasal immunization can elicit immune responses in respiratory tract and genital mucosal tissues in mice, monkeys, and humans (58, 59, 60).

To elucidate which immunization route is the most effective to stimulate antigen-specific humoral and cell-mediated virus-specific responses at HIV-1 entry and replication sites, we compared four mucosal routes of vaccination in female rhesus macaques, oral, intestinal, nasal, and vaginal vaccination, using as vaccine a previously tested SIV DNA plasmid in conjunction with rMVA and inactivated SIV particles as boosts. Regardless of the route of immunization, mucosal vaccination was able to induce SIV-specific humoral responses in rectal and vaginal secretions and T-cell-mediated virus-specific responses in blood and mucosal tissues. It is interesting that compared to other routes, the oral route tended to generate more widely disseminated SIV-specific mucosal IgA responses of broader specificity. Immunization in the oral cavity similarly stimulated the most significant and diverse T-cell responses at multiple sites, with the only exception of the anti-SIV CD8⁺ T-cell response detected in vaginal MNC that was more significant after intestinal immunization. In the case of oral immunization, these responses progressively declined after DNA immunization, while they kept expanding during the multiple intestinal immunizations. Compartmentalization within the mucosal system may provide an explanation for these results, suggesting different recirculation pathways between different parts of the intestinal tract and genital system. Env-specific IgA binding antibodies were detected in vaginal secretions from animals in all groups. It is possible that these antibodies could mediate antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent virus inhibition (ADVI). Unfortunately, these analyses were not planned for this trial.

In previous studies, we found that antigen-specific IFN-γ⁺, CD4⁺, and CD8⁺ T-cell levels correlated with delayed progression to AIDS (22). These responses were also detected in this study and were more significant in the blood and rectal and vaginal mucosa of orally vaccinated animals, being detected at lower levels in the others. On the basis of these results, we expect to achieve control of virus replication and delayed disease progression in many of the vaccinated animals.

Vaccine-induced immune responses elicited with this SIV DNA/MVA vaccine given nasally significantly delayed disease progression after challenge with pathogenic SIVmac251 but did not protect from vaginal infection (22). Stimulation of SIV-specific antibodies in vaginal secretions was sporadic and short-lived. Vaccines made of proteins have been shown to stimulate antibody responses very efficiently (17, 61, 62). Toward this goal, two additional immunizations consisting of inactivated SIVmac239 virions were included in this study. While these immunizations provided excellent boosting of preexisting cell-mediated SIV responses, only a limited improvement of mucosal IgA responses was achieved, and no induction of anti-SIV systemic responses was evident. The dose chosen to immunize the animals was likely insufficient to successfully stimulate strong IgA and IgG responses in mucosal and systemic compartments. Alternatively, a different adjuvant or systemic administration may better achieve the desired results.

Vaginal vaccination provided responses similar to those observed after nasal vaccination in the different compartments that were evaluated. The expectation was that vaginal immunization would provide the best responses in vaginal secretions and associated mucosal tissue. However, this was not the case. This result may simply be related to reduced uptake of vaccine components at this site, which can also be affected by the menstrual cycle phase (48), which was not controlled for in this study. The use of different adjuvants and formulations may also be able to provide a different outcome. However, given that vaccination via this route is less practical than others, the interest in this route for further vaccine development is substantially diminished.

Overall, regardless of the route of vaccination, mucosal vaccination was able to elicit vaccine-induced immune responses locally as well as in distal mucosal and systemic compartments. The excellent results obtained with the oral cavity vaccination are very encouraging for a number of reasons. This modality of vaccination has been used in millions of people for the polio vaccine, and it is a safe route. The fact that no needle injection is required is particularly appealing in developing countries, as side effects at the site of immunization could be avoided and the health care cost is reduced. It will be interesting to evaluate what level of protection this immunization route provides and how it compares with the other immunization routes.