Protection of macaques against vaginal SHIV challenge by systemic or mucosal and systemic vaccinations with HIV-envelope

Background: Worldwide, the majority of human immunodeficiency virus (HIV) infections occur by heterosexual transmission. Thus, the development of a vaccine that can prevent intravaginal HIV infection is an important goal of AIDS vaccine research.

Objectives: To determine which single or combination of systemic and mucosal routes of immunizations of female rhesus macaques with an HIV-1_SF162 envelope protein vaccine induced protection against intravaginal challenge with SHIV.

Design: Female rhesus macaques were immunized with an HIV-1_SF162 envelope protein vaccine administered systemically (intramuscularly), or mucosally (intranasally), or as a sequential combination of both routes. The macaques were then challenged intravaginally with SHIV_SF162P4, expressing an envelope that is closely matched (homologous) to the vaccine.

Results: Macaques receiving intramuscular immunizations, alone or in combination with intranasal immunizations, were protected from infection, with no detectable plasma viral RNA, provirus, or seroconversion to nonvaccine viral proteins, and better preservation of intestinal CD4+ T cells. Serum neutralizing antibodies against the challenge virus appeared to correlate with protection.

Conclusions: The results of this study demonstrate that, in the nonhuman primate model, it is possible for vaccine-elicited immune responses to prevent infection after intravaginal administration of virus.

Introduction

Heterosexual transmission of the human immunodeficiency virus (HIV) leading to the acquired immunodeficiency syndrome (AIDS) continues to fuel a major global pandemic ^[1]. The identification of an immunization strategy that can protect against mucosal HIV transmission via the female genital tract thus should provide an essential mechanism for halting the spread of HIV/AIDS. In the HIV research community, there is general consensus that both antibody and T-cell-mediated immune responses could contribute to an effective prophylactic vaccine ^[2,3].

An important role for serum neutralizing antibodies against viral surface protein antigens for protection has been demonstrated repeatedly for numerous other viral pathogens for which there are effective vaccines, including influenza, polio, hepatitis B virus, MUMPS, and measles ^[4-8]. Thus, it is likely that a successful HIV vaccine candidate should also generate protective neutralizing and/or other functional antibody responses directed against its viral surface envelope glycoprotein. This contention is well supported by passive immunizations studies performed that have amply demonstrated that antibodies alone can protect against challenges with HIV or SIV/HIV chimeric viruses (SHIV) ^[9-13]. Moreover, active immunization strategies using either Env protein-based vaccines or various prime-boost strategies employing combinations of recombinant DNA, viral vectored, and Env protein have also demonstrated varying degrees of protection against homologous or closely-related challenge viruses administered by either the intravenous or intrarectal routes ^[14-19]. Even so, there have been no reports to date of vaccine-induced protection against vaginal challenge with homologous or heterologous viruses using an Env-protein-based approach.

The goal of the present study was to explore differing routes of vaccination to evaluate their relative abilities to induce systemic and mucosal immune responses and protection against intravaginal (IVAG) challenge of macaques with a CCR5-tropic SHIV, i.e. the most relevant nonhuman primate model available for this purpose ^[20]. The vaccines examined in these studies employed exclusively the trimeric V2-deleted Env (SF162 o-gp140ΔV2) protein derived from the HIV-1 subtype B SF162 strain ^[21,22] The rationale for the combined mucosal and systemic routes of vaccinations was based on previous studies demonstrating that a combined regimen using intranasal priming and intramuscular boosting (IN/IM) with protein antigens and adjuvant induced enhanced mucosal and systemic immune responses in mice and rhesus macaques in comparison with regimens that used intranasal/intramuscular or intramuscular/intranasal (IM/IN) immunizations ^[23-25]. LTK63, a mutant of the E. coli heat labile toxin, was used in these studies and was also shown to be effective as an intranasal adjuvant for influenza and was found to be well tolerated by humans ^[26]. For systemic immunizations, a licensed adjuvant, oil in water emulsion, MF59, was used; shown to increase immune responses to co-administered protein-based antigens ^[27]. In the present study, the protective efficacy of the subtype B SF162 o-gp140ΔV2 protein plus adjuvants, was tested in female rhesus macaques through systemic vs. mucosal vs. combinations of mucosal and systemic immunizations, followed by IVAG challenge with the homologous and mildly pathogenic CCR5-tropic SHIV strain, SHIV_SF162P4 ^[28,29].

Methods

Animals, immunizations and SHIV challenge

A total of 21 adult female rhesus macaques (Macaca mulatta) of Indian origin were used for this study, four groups of four macaques and 5 naive controls. The subtype B HIV-1 Env antigen used in this study was derived from the R5 primary isolate SF162 oligomeric gp140 with a deletion in the V2 loop (o-gp140SF160ΔV2) (o-gp140) ^[30]. All immunizations and sample collections were performed on anesthetized macaques. The time intervals between immunizations, challenge and sample collections are shown in Fig. 1. Intramuscular immunizations were with o-gp140 in MF59 (Novartis Vaccines and Diagnostics, Inc., Emeryville, California, USA; 50 μg per dose in the left arm and 50 μg per dose in the right arm, 0.5 ml each arm). Intranasal immunizations were performed with o-gp140 protein (300 μg per dose) LTK63 (Novartis) (100 μg per dose). Intranasal immunizations were performed in a total volume of 400 μl given as 200 μl per nostril. The first group (M352, M356, M359, M368) was immunized intramuscularly twice followed by three intranasal immunizations. The second group (M349, M361, M363, M365) was immunized three times intranasally followed by twice intramuscularly. The third group (M350, M353, M357, M364) was immunized five times intranasally. The fourth group was immunized five times intramuscularly (M351, M358, M362, M366). The SHIV_162P4 challenge stock was prepared by culturing phytohemoagglutinin (PHA)-activated CD4+ T-cell-enriched peripheral blood mononuclear cells (PBMCs) from a SHIV_162P4-infected Indian rhesus macaque with PHA-activated PBMC from naive macaques. Virus was titered in vivo via the vaginal route in seven rhesus macaques (data not shown). As animals inoculated with 1 ml of undiluted virus [approximately 3000 median tissue culture infective dose (TCID₅₀) in rhesus PBMC] became readily infected, as evidenced by high plasma viremia, this dose of virus was selected for challenge. All immunized and a group of unvaccinated (M075, M670, L869, L939, M367) animals were anesthetized and inoculated once intravaginally with 1 ml containing approximately 3000 TCID₅₀ of an undiluted SHIV_SF162P4 stock. Of note, unlike other groups, only three animals were challenged in the five times intramuscular immunization group as one animal from that group was found dead before challenge. Sera were collected from 4 ml of whole blood and stored at -20°C. At specific time points, 1 ml each of vaginal washes and nasal washes, rectal lavage and saliva were collected from each macaque under anesthesia, frozen immediately on dry ice and stored at -80°C. Seven days prior to vaginal challenge, all animals were treated with antibiotics to ensure lack of bacterial infections and inflammation in the vaginal/uterine mucosa. All animals were housed at the Advanced BioScience Laboratories, Inc. (Kensington, Maryland, USA) in accordance with standards of the Association for Assessment and Accreditation of Laboratory Animal Care.

Fig. 1

Measurement of virus infection in plasma and tissues

Viral RNA in plasma was quantitated by using the real-time nucleic acid sequence-based amplification (NASBA) method as described ^[31]. Peripheral blood mononuclear cells (PBMC) collected post-challenge at 2 weeks following challenge were also assayed for proviral DNA copies by real-time PCR (data not shown). The limit of detection of proviral DNA was 10 copies per million PMBC.

Enzyme-linked immunosorbent assay, ELISPOT and serum neutralization assays

HIV-1 o-gp140 env-specific serum IgG titers were quantified by a standard colorimetric enzyme-linked immunosorbent assay (ELISA) exactly as described ^[32]. Vaginal and rectal lavage, nasal wash, and saliva samples were assayed for anti-o-gp140 and total IgG and IgA using a Europium-based ELISA assay. Ninety-six-well plates (Nunc, Denmark) were coated with o-gp140 or goat anti-monkey IgA (KPL) or goat anti-monkey Ig (H+L) (Southern Biotechnology Associates, Birmingham, Alabama, USA). The samples were then added in DELFIA assay buffer (PerkinElmer, Boston, Massachusetts, USA). The plates were developed following incubations first with biotinylated goat antimonkey IgA (Fitzgerald (RDI), Concord, Massachusetts, USA) or biotinylated goat anti-monkey IgG (Nordic Labs, Capistrano Beach, California, USA) and then with ^[31] Streptavidin-Europium (PerkinElmer). The plates were read on a Wallac Victor 21429 Multilabel fluorescence reader at 616 nm. Neutralization assays were performed using a standardized TZM-bl assay and pseudovirus stocks as described elsewhere ^[33]. Neutralizing antibody titers were expressed as the reciprocal of the dilution at which relative luminescence levels were reduced to 50% as compared to control wells (no test serum). IFNγ secretion by PBMC was measured by ELISPOT as described ^[2] Peptide pools from HIV-1 gp120 peptide pool (NIH AIDS reagent program) were added at 1 μg/ml. As positive control, the cells were stimulated with PHA at 10 μg/ml (Sigma/Aldrich, St Louis, Missouri, USA), and as negative control the cells were stimulated with media alone. Results are presented as mean of triplicate wells of IFNγ secreting cells per million mononuclear cells (MNC) from individual macaques with values from two relevant un-stimulated wells subtracted.

Cell preparations and flow cytometric cell analysis

The macaques were given antibiotics a week before the biopsy collections, were given analgesics at the time of collection and observed closely following collection for any signs of pain or distress, vaginal bleeding, or vaginal infection, none of which were observed in any of the animals. Mononuclear cells from peripheral blood, jejunal and vaginal pinch biopsies were isolated, stained and analyzed as described ^[34].

Statistical analysis

The low sample sizes did not allow the assumption of a normal distribution of the data. Therefore, the nonparametric, one-way analysis of variance (ANOVA), Tukey's analysis of variance was performed for multiple comparisons of differences between the groups at the 95% confidence interval, and the p values were reported, using the GraphPad Prism statistical analysis software (Graphpad Software Inc., San Diego, California, USA).

Results

Serum binding and neutralizing antibody responses in vaccinated macaques

Env-specific binding antibody titers (Fig. 1a) and neutralizing antibody responses (Fig. 1b) in sera were measured during the immunization phase and post-virus challenge. Two or three intramuscular immunizations (2×IM or 3×IM) induced relatively strong serum anti-gp140 IgG responses in all macaques (Fig. 1a). In contrast, three intranasal immunizations (3×IN) generally induced low or undetectable serum antibody responses; however, in the group that was immunized three times intranasally, serum antibody responses were raised substantially following one and two intramuscular booster immunizations. Indeed, the level of serum anti-gp140 IgG in the 3×IN/2×IM group (range: 18692-117047; average: 63480) was significantly higher than that of the five intramuscular immunizations (5×IM) group (range: 19613-38619; average: 29231) on the day of challenge (P < 0.03). Beginning at 2 weeks after the fifth and final immunization, the binding antibody responses in the 2×IM/3×IN group, were significantly lower than the 3×IN/2×IM or 5×IM groups (Fig. 1a). Intranasal immunizations alone did not induce appreciable serum antibody responses until after intramuscular boosting, when they increased to the level seen with four intramuscular (4×IM) immunizations. The second (or fifth dose) intramuscular dose raised the antibodies to even higher levels. An increase in serum antibodies seen in the 5×IN group after challenge seemed to be due to infection; however the antibodies in the 5×IN group after challenge showed an anamnestic response, increasing sooner than seen in the unvaccinated/challenged animals.

Neutralizing antibody titers against the challenge virus measured at 2 weeks postfinal immunization, also were highest in the 3×IN/2×IM (range: 574-1104) or 5×IM (range: 394-2423) groups in comparison with the 2×IM/3×IN (range: < 20-161) and 5×IN (all < 20) groups; this trend was similar on the day of challenge, which was at 6 weeks after the final immunization (Fig. 1b).

In addition, at 2 weeks after the fifth immunization, sera were tested for the ability to neutralize the heterologous, neutralization sensitive isolates HIV-1_MN (subtype B) and HIV-1_MW965 (subtype C). Sera from the intranasal/intramuscular immunized (average titers 114, 186, respectively) and intramuscular alone (average 300, 485) immunized animals, but not the intranmuscular/intranasal (IN/IM) (average 27, 27) and intranasal alone immunized (average 27, 27), animals were able to neutralize these isolates. However, none of the animals showed neutralization of other heterologous virus strains, such as SHIV-SF162P3, HIV-1Ba-1, SHIV89.6, HIV-1B×08.16, and HIV-1 SS1196.1. Overall, IN/IM immunizations induced equal or better serum IgG responses, as well as serum neutralizing antibody responses, in comparison with intramuscular-alone immunizations.

Mucosal antibody responses in vaccinated macaques

Mucosal antibody responses were measured in vaginal, nasal, and rectal lavages as well as saliva at 2 weeks after the final immunization. The IN/IM immunizations induced significantly higher IgG responses in vaginal washes at this time-point (Fig. 2a), and also in nasal washes (data not shown) and saliva (data not shown), in comparison with all other immunization groups (Table 1). Interestingly, IgA responses in vaginal washes were significantly higher in the macaques immunized IN/IM compared with the macaques immunized IM/IN or intranasally, but not in the macaques immunized intramuscularly (Fig. 2b and Table 1). Serum IgA responses were significantly higher in the macaques immunized IN/IM in comparison with all other groups (Fig. 2c and Table 1). Env-specific IgG and IgA responses were not detectable in rectal lavages from any animals.

Fig. 2

Table 1

After the vaginal challenge with SHIV, the IgG and IgA responses in vaginal washes were generally maintained or slightly declined in the macaques immunized by the IN/IM, IM/IN or intramuscular routes indicative of the absence of infection (Fig. 2a and c). By 4 weeks after challenge, the IgG and IgA vaginal wash responses were highest in the infected intranasally vaccinated macaques, suggesting that the intranasal priming immunizations rapidly gave rise to relatively high vaginal antibodies following vaginal challenge as may be expected by an anamnestic response (Fig. 2a and b). Overall, mucosal followed by systemic immunizations induced the highest IgG responses in vaginal and nasal washes and saliva as well as the highest serum IgA responses in comparison with all other immunization routes prior to vaginal challenge with SHIV.

Cellular immune responses in vaccinated macaques

Cellular immune responses in all animals were measured by the ELISPOT assay throughout the immunization period. These showed that, overall, intramuscular immunizations induced greater IFNγ responses than intranasal immunizations (Fig. 3); however, 2×IM booster immunizations of macaques that had previously received 3×IN immunizations, induced the same or higher level of IFNγ responses compared to 5×IM immunizations. Intranasal immunizations alone, even after five doses, induced very low or negative IFNγ responses (Fig. 3). These data demonstrate that mucosal followed by systemic immunizations induced similar or better peripheral IFNγ responses in comparison with systemic alone immunizations.

Fig. 3

Protection against intravaginal challenge with CCR5-utilizing SHIV

Six weeks following the final vaccination, all macaques were challenged vaginally with a high dose (approximately 3000 TDIC₅₀ in rhesus PBMC) of SHIV_SF162P4. Plasma viral loads were measured and monitored beginning at 3 days after challenge (DPC). No plasma viral RNA was detected (< 50 copies/ml) in any of the animals immunized IN/IM, IM/IN or intramuscularly (Fig. 4a, b and d). In contrast, the unvaccinated naive control animals and macaques receiving only intranasal immunizations displayed substantial viral plasma RNA beginning at 7 DPC and peaking at 21 DPC. Importantly, whereas plasma viral RNA in animals only immunized intranasally alone and unvaccinated had the same kinetics until peak at 21 DPC, the intranasally vaccinated animals displayed a substantial decrease in plasma viral load from peak and had undetectable viral RNA by about 3 months after challenge, whereas the unvaccinated animals still had significant levels of viral RNA at this time point. Thus, immunizations through the combinations of mucosal and systemic or systemic alone routes appeared to result in complete protection of the 12 monkeys in these three groups as determined by undetectable plasma viral RNA following vaginal challenge. These results were confirmed by the absence of detectable proviral DNA in the PBMC of these protected animals on 42 DPC (data not shown), the absence of SIV-Gag specific antibodies post-challenge (data not shown), and the absence of an anamnestic Env-Ab response post-challenge (Fig. 1a). Moreover, although intranasal immunizations alone did not similarly protect against infection, beginning at 28 DPC, there appeared to be a vaccine effect resulting in lower plasma viral RNA as compared to controls, albeit not at statistically significant levels.

Fig. 4

Maintenance of jejunal and peripheral CD4+ T cells in vaccine-protected macaques

From 1 day pre-challenge until 14 weeks post-challenge (WPC), the average values for peripheral blood CD3+CD4+ T cells were generally lower in each of the unprotected groups, in comparison with the protected groups, but these differences did not reach statistical significance (data not shown). the CD3+CD4+ T cells values, however, in all of the protected macaques were significantly higher in comparison with the unprotected macaques, by 10 WPC (P = 0.04) and 14 WPC (P = 0.03) (Fig. 5a).

Fig. 5

Lymphocytes (mononuclear cells) were isolated from jejunal pinch biopsies of the monkeys at 4 weeks after vaginal challenge, and the percentages of CD3+CD4+ T cells were determined. Although the uninfected animals showed a varying range of CD4+ T-cell percentages, and given the fact that this was a single time point with no comparison pre-challenge, there was clearly a statistical difference between the jejunal CD4+ T-cell levels in the protected monkeys in comparison with the infected naive control monkeys (IM/IN P = 0.01; IN/IM P = 0.02; intramuscular P = 0.05), indicating that the control monkeys were already showing jejunal (but not peripheral) CD4+ T-cell loss from infection with SHIV_SF162P4. The intranasally immunized animals, which were infected and showed plasma virus loads comparable to the naive controls, did not show a statistically significant preservation of jejunal CD4+ T cells in comparison with the naive control animals (P = 0.13).

Relationship of neutralizing antibodies to protection

At 2 weeks after the final vaccination, macaques immunized IN/IM, IM/IN or intramuscularly showed serum neutralization titers against the homologous SHIV_SF162P4, to varying degrees but were all above a titer of approximately 100. Macaques that were only immunized intranasally and naive macaques showed very little or no serum neutralizing activity (Fig. 1b) pre-challenge. Beginning at 14 DPC animals only immunized intranasally increased their neutralization titers rapidly and by 28 DPC sera from animals only immunized intranasally exhibited substantial neutralization titers. This rise in serum neutralization titers (Fig. 1b), which was significantly higher than all other groups (neutralization titers of intranasal immunization versus IM/IN P = 0.037; versus IN/IM P = 0.039; versus intramuscular immunization P = 0.04; versus naive P = 0.035) coincided with a significant decrease in viral loads from day 21 (1 week before rise in neutralizing antibody on week 4) to day 35 (1 week after rise in neutralizing antibody on week 4) (viral loads P = 0.05). The development of neutralizing antibodies in the unvaccinated controls took much longer in comparison (not shown). These data may suggest a relationship between the presence of serum neutralization titers and the observed protection (of the IN/IM, IM/IN and intramuscular immunization groups) or reduction in plasma viral loads (for the intranasal immunization group).

Discussion

The findings described in this study, provide proof-of-concept for Env protein-based vaccines delivered in potent adjuvant(s) as a means to protect against heterosexual transmission of HIV via a key mucosal portal of entry, the female genital tract. Safe and effective HIV vaccine candidates that can completely protect, as defined by lack of plasma virus RNA and preservation of peripheral and/or mucosal CD4+ T cells, against vaginal challenge with even homologous SHIV or SIV strains, have not been reported ^[35-39]. Furthermore, this is the first study demonstrating protection through vaccination with a clinically relevant vaccine candidate as the intranasal prime, intramuscular boost. The vaccination approach described here is now being tested in an ongoing Phase 1 clinical trial in the United Kingdom.

Although intramuscular immunizations alone were found to be protective against the homologous virus challenge in the present study, it also remains to be determined whether this route of delivery will be sufficient for more rigorous challenges with heterologous viruses. Thus, it is of interest, that at specific time points after immunizations or challenge, mucosal followed by systemic immunizations induced the highest mucosal and systemic responses as measured by serum IgA, vaginal, nasal and saliva IgG as well as peripheral IFNγ ELISPOT responses, in comparison with any other routes of immunization. This is consistent with previously published data in mice using Helicobacter pylori- or influenza-derived proteins or in macaques using HIV-1 envelope proteins ^[23-25]. The mechanism of the observed increases in these immune responses is unknown. This phenomenon is, however, supported by studies in which systemic immunizations with polio ^[40] or influenza ^[41] vaccines enhanced immunity only in individuals pre-exposed mucosally through natural or active immunization. Of note, for intranasal immunizations, although three-fold higher doses of o-gp140 was used, this route of immunization with this particular vaccine did not protect against vaginal challenge. Although it is widely accepted that mucosal immunizations require higher doses of antigen in comparison with systemic immunizations, the type of adjuvant or delivery system would directly affect the efficacy of the vaccine ^[42].

In cytotoxic T lymphocyte-targeted vaccine studies, correlations between cellular immune responses pre-challenge and control of virus replication after infection have been suggested, although infection was not prevented ^[43-46], albeit rarely to undetectable levels. In the current study, peripheral IFNγ ELISPOT responses following re-stimulation with HIV-1 envelope derived peptide pools were detected prechallenge. They correlated with peripheral and vaginal antibody responses for the IN/IM, intramuscular alone and intranasal alone groups, but not the IM/IN group. Therefore, a correlation between peripheral IFNγ ELISPOT responses and protection appears unlikely.

Several lines of evidence suggest that IgG responses may play a major role in the vaginal mucosa of primates ^[35,47,48]. In this study, o-gp140-specific IgG in nasal and vaginal washes and saliva were found to be the highest in IN/IM immunized animals. Immunity in vaginal and saliva secretions has direct implications for various sexual practices and the mucosal HIV transmission.

Despite the promising proof-of-concept described here, the challenge remains to demonstrate that similar vaccine strategies can be used to protect against heterologous mucosal challenges using diverse CCR5-tropic SHIV viruses whose virus env genes do not closely match the vaccine antigens ^[20]. Moreover, because the design of this study was such that the animals survived to measure the kinetics of the immune response and peripheral blood viral loads after challenge, we can not be certain that the protection was complete since we could not detect presence of virus in various tissues. Even so, the results described here indicate that further improvements in Env immunogen design, adjuvants, and/or vaccine delivery may induce protection against heterologous challenge.

Acknowledgements

We thank J. Treece and S. Orndorff for their assistance with the study. We are grateful for important contributions from Dr Nancy Miller, NIH, for the design, analysis and reporting of this study.

Sponsorship: the study was supported by NIAID contract N01-AI60005 for the Advanced BioScience Laboratories, Inc. Simian Vaccine Evaluation Unit, and by NIAID contracts N01-AI30033 and N01-AI30034 for cellular and humoral immune assays.