Virus-Like Particles Displaying Trimeric Simian Immunodeficiency Virus (SIV) Envelope gp160 Enhance the Breadth of DNA/Modified Vaccinia Virus Ankara SIV Vaccine-Induced Antibody Responses in Rhesus Macaques

All animal protocols were approved by Emory University Institutional Animal Care and Use Committee (IACUC) protocol YER-2002343. All animal experiments were conducted in strict accordance with USDA regulations and recommendations for conducting experiments in accord with the highest scientific, humane, and ethical principles, as stated in the Guide for the Care and Use of Laboratory Animals (13). Animals were housed in pairs in standard nonhuman primate cages. Animals received standard primate feed (Purina monkey chow) as well as fresh fruit and enrichment daily and had free access to water. Upon infection, animals were housed singly. Immunizations, infections, blood draws, and biopsy procedures were performed under anesthesia by trained research staff.

Twenty-four male Indian rhesus macaques were included in this study, 14 of which were immunized as described below. Animals ranged in age from 2 to 4 years and ranged in weight from 3 to 6 kg at the start of the study. Animals were negative for simian T cell lymphotropic virus, herpes B virus, simian retrovirus type D (SRV-D), and SIV at study commencement. None of the animals expressed Mamu class 1 alleles B08 and B17. The vaccine and control groups had 3 Mamu A*01 animals each. Animals from experimental groups were randomized into four sampling groups for immunization, challenges, and sampling of blood and tissue biopsy specimens. All animals were housed at the Yerkes National Primate Research Center (YNPRC) in Atlanta, GA, and treated in accordance with YNPRC IACUC regulations.

The vaccine study consisted of two experimental groups; all 14 vaccinated animals received two CD40L-adjuvanted DNA primes (D) (0 and 8 weeks) followed by two MVA boosts (M) (16 and 32 weeks) (DDMM regimen). To determine whether an extended protein boost after the second MVA administration augmented antibody responses, 7 of the 14 animals received a VLP protein boost at week 76. The DNA immunogen expressed SIV239 Gag-Pol, Env, Tat, and Rev and rhesus macaque membrane-anchored CD40L (14) and was delivered at 3 mg/dose. The MVA immunogen expressed SIV239 Gag, Pol, and Env and was delivered at a 10⁸ PFU/dose (15, 16). Protein vaccination consisted of purified VLPs (containing 50 μg gp160 Env and 1.5 mg Gag) premixed with poly(d,l-lactic-coglycolic acid) (PLGA)-based nanoparticle-encapsulated monophosphoryl lipid A (MPL) (TLR4 agonist) (50 μg) and R848 (TLR7/8 agonist) (750 μg) adjuvants prior to vaccination. Briefly, 200 μl of PLGA nanoparticles containing the respective adjuvants was admixed with 1 ml of VLPs by mild vortexing, followed by gently pipetting the mixture up and down prior to vaccination. The prepared inoculum was stored on ice until the time of inoculation. DNA and MVA immunizations were delivered in phosphate-buffered saline (PBS) intramuscularly in a single shot in the outer thigh. VLPs plus TLR4/7/8 adjuvants were administered subcutaneously behind the knee.

Animals were challenged with SIVmac251 intrarectally on a weekly basis for a maximum of 5 weeks or until detection of plasma viremia at levels above 250 copies/ml for two consecutive weeks. Infection with SIVmac251 (day 8) (7.9.10 virus stock from Nancy Miller at the NIH) was performed by using a 1-ml slip tip syringe containing 1 ml of SIVmac251 at 200 50% tissue culture infective doses (TCID₅₀). The syringe was inserted gently ∼4 cm into the rectum, the plunger was depressed to instill the virus, and the animal was returned to the cage in a prone position.

Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood collected into sodium citrate tubes and isolated by density gradient centrifugation according to standard procedures, as described previously (15). PBMCs were isolated, counted, and utilized for various assays within 6 h of blood collection. After determination of cell counts, cells were used immediately or cryopreserved by using standard techniques. Cell viability in PBMC suspensions was determined to be above 90%. Rectal secretions were collected with and eluted from premoistened Weck-Cel sponges as previously described (17).

The linear epitope specificity of the antibodies was determined by profiling serum IgG responses against linear epitopes representing full-length SIV239 Env in a peptide microarray, as described previously (18). Sera from baseline, 2 weeks after the second MVA immunization, and 2 weeks after VLP immunization were examined. The peptide library consisted of 15-mer peptides overlapping by 12 covering SIVmac239 gp160. Serum was tested at a 1:50 dilution against slides (slide lot JPT 2413) coated with a library of peptide probes. A secondary goat anti-human fluorescent IgG antibody was added to tag peptide-bound serum antibodies, and slides were scanned (photomultiplier tube [PMT] setting 660, with 100% power) to yield fluorescence intensity values for each probe (19). Final intensity values were computed after baseline subtraction, and the maximum binding for defined epitopes was determined based on the highest level of binding to a single peptide within the epitope region.

Concentrations of serum IgG binding antibodies to recombinant SIVmac239 gp140 (Immune Technology) or trimeric gp120 were measured by using gp140-coated plates or SIV239 VLPs captured on concanavalin A (ConA)-coated plates, respectively, as described previously (20, 21). Concentrations of antibodies to SIVmac251 recombinant gp140 (rgp140) (Immune Technology) and total IgA and IgG in rectal secretions were measured by an enzyme-linked immunosorbent assay (ELISA) as previously described (22). All samples intended for IgA antibody analysis were first depleted of IgG by using protein G-agarose (GenScript) as described previously (17). Concentrations of gp140-specific IgA or IgG in secretions were divided by the total concentration of IgA or IgG to obtain the specific activity. Secretions were considered antibody positive if the specific activity was greater than or equal to the mean specific activity plus 3 standard deviations (SD) measured in secretions collected at baseline.

A customized binding antibody multiplex assay (BAMA) (23) was used in the Kozlowski laboratory to measure antibodies to recombinant SIVmac251 gp120 (Immune Technology), gp70-V1V2mac239 (24) (from D.C.M.), SIV p55gag (a gift from Francois Villinger, New Iberia Research Center), and HIV-2 gp36 (Prospec), which has 85% homology to the SIV gp41 ectodomain (25). Each viral protein was conjugated to carboxylated Luminex beads (Bio-Rad) by placing 1 × 10⁷ beads into the upper chamber of an Ultrafree 0.1-μm centrifugal filter unit (Millipore), washing the beads with PBS, and then adding 5 mg/ml N-hydroxysulfosuccinimide (Pierce) and 5 mg/ml ethyl-3-[3-diethylaminopropyl]carbodiimide hydrochloride (Sigma) in 0.1 M Na phosphate buffer (pH 6.2). After 20 min, the beads were washed and mixed at 900 rpm in an Eppendorf MixMate vortexer with 100 μg of protein that had been dialyzed in PBS. After 2 h, the beads were washed, resuspended in PBS containing 1% bovine serum albumin (BSA) and 0.05% azide, and counted in a hemocytometer. Assays were performed by using 2 panels of beads: gp120 and gp36 (panel 1) and V1V2 and p55 (panel 2). For each assay, 2,500 beads expressing each protein were added in a final volume of 25 μl to wells of a blocked, low-protein-binding, 96-well white plate (Costar), which contained 25 μl of serum and a standard that had been diluted in PBS containing 1% BSA, 0.05% Tween 20, and 0.05% azide (assay buffer). For IgG assays, the standard was IgG purified from pooled serum of SIV-infected macaques. For IgA assays, the standard was IgG-depleted serum from SIV-vaccinated/infected macaques (22). All standards had been previously calibrated by an ELISA as described previously (26) and were adjusted in the bead assay mixture to achieve test sample concentrations similar to those obtained in ELISAs. After overnight mixing at 4°C, the plate contents were transferred onto a Durapore 96-well filtration plate (Millipore) and washed 4 times with PBS containing 0.05% Tween 20 in a BioTek ELx50 washer. The beads were mixed with 100 μl of 2 μg/ml biotinylated goat anti-monkey IgA or IgG (Rockland) for 30 min at room temperature, washed, and then treated with 100 μl of a 1/200 dilution of Neutralite-phycoerythrin (SouthernBiotech) in assay buffer. After 30 min at 900 rpm, the beads were washed, and the plastic tray on the bottom of the plate was removed. The filtration plate was then placed on top of the white plate. Beads were transferred back to the white plate by mixing with 100 μl of assay buffer, puncturing the filters with pipette tips, and then centrifuging the plates at 2,300 × g for 5 min. The beads were briefly mixed at 700 rpm on a rotating platform and then analyzed for fluorescence with a Bioplex 200 system (Bio-Rad). Concentrations of antibody were interpolated from standard curves constructed by using Bioplex Manager software.

The SIV-specific neutralization antibody response was measured as a function of the reduction in luciferase reporter gene expression levels after a single round of infection in TZM-bl cells, as described previously (27). Neutralization against homologous SIVmac251.6 (tier 1) and SIVmac251.30 (tier 2) and heterologous SIVsmE660/CP3C-P-48 (tier 1) and SIVsmE660/CR54-PK-2A5 (tier 2) was measured at peak and memory time points following the second MVA immunization and VLP boost. Values were considered positive for neutralization based on the criterion of >3× background signal with the negative-control virus simian virus amphotropic murine leukemia virus (SVA-MLV).

Phagocytosis assays were performed as described previously (28), with the modifications noted below. Briefly, 1 × 10⁹ 1-μm Neutravidin Fluorospheres (Invitrogen) were labeled with 7 μg biotinylated recombinant SIV gp120mac251 (Immune Technology) or 5 μg biotinylated concanavalin A (vector) followed by 10 ml of VLP-containing medium produced by transfection of 293T cells with SIV DNA, as described above. After washing, 1 × 10⁷ beads per well were added to V-bottom plates containing triplicate serial dilutions of heat-inactivated serum that had been absorbed against 293T cells. THP-1 cells (2 × 10⁴ per well) were then added. After 6 h at 37°C in 5% CO₂, the cells were washed with Ca²⁺/Mg²⁺-free Dulbecco's PBS (DPBS) and incubated for 10 min with 50 μl of 0.05% trypsin-EDTA (Gibco). The cells were washed in DPBS, resuspended in 1% paraformaldehyde, and analyzed for fluorescence. The score was calculated as described previously (28), by multiplying the number of bead-positive cells by their median fluorescence intensity. To obtain the score ratio (29), the average score for the test samples was divided by the average score for naive monkey serum at the same dilution.

The SIV copy number was determined by using quantitative real-time PCR as previously described (14). All PCRs were performed in duplicates, with a limit of detection of 60 copies per reaction.

Prior to intracellular staining for cytokines, fresh PBMCs were stimulated with peptide pools of Gag and Env for 5 h in the presence of brefeldin A (Golgi Plug; BD Biosciences, San Jose, CA). Unstimulated cells from each animal served as a negative control, and PBMCs stimulated with phorbol myristate acetate-ionomycin served as positive controls. Cells suspensions were first stained with the Live/Dead Near-IR dead-cell stain from Molecular Probes, Invitrogen (Grand Island, NY), in PBS containing 2% fetal bovine serum (FBS) (fluorescence-activated cell sorter [FACS] buffer) for 30 min at 4°C. Intracellular staining was performed after cells were fixed with Cytofix/Cytoperm (BD Biosciences, San Jose, CA), followed by permeabilization with BD Perm/Wash buffer according to the manufacturer's instructions. Cells were stained in Perm/Wash buffer with CD3 (SP34-2), CD4 (OKT4), CD8 (SK1), gamma interferon (IFN-γ) (B27), tumor necrosis factor alpha (TNF-α) (MAb11), interleukin-2 (IL-2) (MQ1-17H12), and IL-21 (3A3-N2.1), from BD Pharmingen, at 4°C for 45 min. After two washes, samples were acquired on an LSR Fortessa instrument (BD Biosciences), and 500,000 total events were collected for each sample. Data were analyzed by using FlowJo software vX.0.7 (Tree Star, Inc., Ashland, OR). Dead cells were excluded from the analysis.

Statistical analysis was performed by using GraphPad Prism v5.0. A two-tailed nonparametric t test was used for all comparisons unless otherwise specified. Spearman correlation was used to determine associations between variables. Statistical significance was set at a P value of <0.05.

The decline in Env antibody titers postvaccination and the associated increase in HIV acquisition risk in RV144 vaccinees spurred interest in extended protein boosting regimens as a strategy to augment waning Env titers and enhance vaccine efficacy (30). The objective of the present study was to determine whether an extended protein boost contributes significantly to enhancing antibody titers induced by a DNA prime-MVA boost immunization. We evaluated this question by employing DNA and MVA constructs designed to express trimeric Env on VLPs in vivo. In an effort to augment conformation-dependent Env antibody responses, we employed a VLP protein boost. The VLP protein immunogen was generated by using an inducible 293F SIV Gag-Env stable production cell line. Codon-optimized SIVmac239 Gag and Env genes were synthesized (GenScript) and cloned into the pcDNA4/TO and pcDNA5/TO vectors (Thermo Fisher Scientific), respectively. 293F cells were transfected with both vectors and selected by culture in zeocin and puromycin, followed by single-cell cloning and expansion of high-producer clones. VLP production was induced by the addition of doxycycline for 48 h, after which VLPs were pelleted through a 20% sucrose cushion. Purified VLPs were resuspended in PBS and were quantified by p27 and gp120 ELISAs for Gag and Env contents prior to immunization. VLP production lots were characterized on a 20 to 60% sucrose gradient, followed by Western blotting (Fig. 1A). As expected, fractions 5 to 7 (density from 1.14 to 1.16 g/cm³) were enriched in VLPs, as measured by the enrichment of Env and Gag in these fractions. Electron micrograph analysis of transfected cells confirmed the presence of budding VLPs on the cell membrane (Fig. 1B).

Fourteen rhesus macaques were primed intramuscularly with 3 mg of CD40L-adjuvanted plasmid DNA expressing SIVmac239 Gag, protease, reverse transcriptase (RT), Env gp160, Tat, and Rev and coexpressing macaque CD40L at 0 and 8 weeks (14) (Fig. 2A). Booster immunizations comprised recombinant MVA expressing SIVmac239 Gag, protease, RT, and Env gp150 administered intramuscularly at 10⁸ PFU on weeks 16 and 32. After a 24-week rest period, seven animals were challenged intrarectally at weekly intervals with 200 TCID₅₀ (1.16 × 10⁷ copies of viral RNA) of SIVmac251 for 5 weeks. Animals that tested positive for SIV Gag at >250 copies/ml plasma for two consecutive weeks were not challenged further. Unvaccinated control animals were similarly challenged. The remaining seven animals received a subcutaneous booster immunization with SIVmac239 VLPs containing Env gp160 (50 μg) and Gag (1.5 mg), together with PLGA-based nanoparticle (NP)-encapsulated MPL (TLR4 agonist) (50 μg) and R848 (TLR7/8 agonist) (750 μg) adjuvants at week 76 (31). Following a 12- or 20-week rest period, animals in the VLP group were challenged with the same SIVmac251 viral stock by using identical infection protocols. All animals were monitored for viral control and anti-SIV immune responses for 24 weeks following infection. Animals were male and negative for Mamu B*08 and B*17 alleles.

To evaluate humoral immune responses, serum samples were collected at peak and memory time points after each of the booster immunizations. Binding IgG antibody responses to homologous SIVmac239 Env gp140 were detected by an ELISA in all animals after the first MVA boost, with median peak titers of 10 μg/ml, and responses significantly increased after the second MVA immunization, with median peak titers of 175 μg/ml (Fig. 2B). These responses declined to a median of 35 μg/ml by 10 weeks and were maintained at these levels for the next 34 weeks. A single immunization with NP-adjuvanted VLP protein robustly boosted gp140 titers by an average of 20-fold (range, 12- to 28-fold), resulting in median peak gp140 titers of 468 μg/ml, which significantly exceeded peak responses after the second MVA boost (P < 0.01) (Fig. 2C). Similar increases in binding antibody titers were also observed against Env gp160 derived from Triton X-100-dissociated VLPs and captured on ConA-coated plates (Fig. 2B and C), SIVmac251 gp120, and HIV-2 gp36, which is highly homologous to SIVmac251 gp41 (25) (Fig. 2D). Due to the high concentrations of Gag contained in the VLP immunogen, we measured antibody titers against the recombinant SIV239 Gag p55 protein (Fig. 2E) and observed a 17-fold increase in peak titers following the VLP boost (median of 70.3 μg/ml, compared to a median of 4.25 μg/ml after the second MVA boost; P < 0.05). Assessment of titers against gp140 and ConA-captured gp160 10 to 12 weeks following the second MVA and VLP boosts (memory of VLP) showed that the VLP boost significantly increased memory titers against ConA-captured gp160 but not against gp140 (Fig. 2F). These results demonstrated that VLPs displaying trimeric SIV Env induce a strong recall of SIV gp120-, gp41-, and Gag-specific antibody responses in serum.

Next, we determined whether antibody-neutralizing activity was boosted after VLP immunization. We measured neutralization responses against homologous (SIVmac251) and heterologous (SIVsmE660) SIV isolates using a TZM-bl cell-based assay (27). While a response against harder-to-neutralize tier 2 homologous and heterologous viruses was not detected, the VLP boost augmented neutralization activity against tier 1 viruses relative to the second MVA vaccination (Fig. 3A and B). This was true at peak (week 2) as well as memory (weeks 8 to 10) time points (Fig. 3A and B). We also sought to evaluate associations between binding antibody titers and neutralizing antibody responses after MVA and VLP boosts; the data showed that neutralizing responses were highly correlated with the magnitude of gp140 binding titers at peak and memory time points after MVA immunization but not at the peak time point post-VLP boost (Fig. 3C). Indeed, neutralization responses were constant over a range of antibody titers in VLP-boosted animals, suggesting that neutralization activity did not increase in proportion to the increase in binding antibody titers following the VLP boost. Together, the data demonstrate that the VLP protein boost effectively elicits a robust recall of DNA/MVA-induced neutralizing antibody responses.

We also evaluated whether nonneutralizing functional antibody responses were correspondingly boosted at the peak time point following VLP immunization. We measured antibody-dependent phagocytosis (ADP) against gp120 and VLP-coated beads in serum at the peak time points after the second MVA immunization and post-VLP immunization. We computed the phagocytic score by computing the ratio of ADP responses in vaccine sera to the response from naive monkey sera. Assessment of ADP responses using THP-1 monocyte-mediated capture of SIVmac251 gp120 or ConA-VLP-coated fluorescent beads (28) showed a significant increase at the peak time point post-VLP immunization relative to that at the peak time point after the second MVA immunization (Fig. 3D and E). The magnitude of ADP responses against gp120 and VLP-coated beads directly correlated with anti-gp120 binding antibody titers. Thus, our data indicate that a VLP booster following DNA/MVA immunization induced antibodies that direct tier 1 neutralizing responses and ADP activity.

We next sought to ascertain whether VLP immunization also boosted Env-specific antibody titers in the rectal mucosa. For this purpose, we assayed colorectal secretions nontraumatically using premoistened Weck-Cel sponges according to previously described protocols (17, 22). Total and Env-specific IgG and IgA responses were assessed at baseline and longitudinally after the second MVA and VLP boosts (Fig. 4). Responses were considered positive if the absorbance was greater than the mean plus 3 standard deviations of the absorbance of baseline samples. Based on this criterion, the cutoffs for absorbance were 0.03 for IgG and 0.23 for IgA responses in mucosa.

Similarly to serum IgG, the VLP boost induced a robust increase in mucosal gp140-specific IgG titers (Fig. 4A). The magnitude of the IgG response 2 weeks after the VLP boost (median of 29.4 μg/ml) was 3.8-fold higher (P < 0.01) than the magnitude 2 weeks after the second MVA boost (median of 7.8). This increase in gp140-specific IgG titers was due to an increase in titers against both gp120 and gp41 (Fig. 4B). As expected, Env-specific mucosal IgG responses strongly correlated with systemic IgG responses (P < 0.01) (data not shown) (32). Similarly to the responses in sera, we found a significant increase in anti-Gag IgG titers in the rectal mucosa.

Peak SIV Env-specific IgA responses after the second MVA immunization were low and detectable in only 5/14 animals (median, 0.21 μg/ng total IgA). However, booster immunization with VLPs significantly enhanced mucosal IgA responses (median, 1.5 ng/μg total IgA) (Fig. 4C), resulting in peak titers that were, on average, 6-fold higher (P < 0.01) than those for the second MVA immunization. Consistently, IgA titers against gp120 and gp41 were significantly higher after VLP immunization than after the second MVA immunization (Fig. 4D). Measurement of gp140-specific IgA responses in sera showed that VLP immunization induced strong gp140-specific responses in the systemic compartment, which did not correlate with mucosal IgA responses (data not shown). However, unlike the significant increase in Gag-specific IgG titers, we did not observe an increase in Gag-specific IgA titers in the mucosa. In summary, the data show robust induction of mucosal Env-specific IgG and IgA responses following VLP booster immunization.

Having established the robust induction of binding antibody responses after VLP immunization, we next determined whether the VLP boost changed the epitope specificity. For this purpose, we assessed Env-specific IgG responses to overlapping linear peptides (15-mers overlapping by 12) spanning the entire SIV239 gp160 protein. The intensity of binding to each peptide was determined at week 2 following the second MVA and VLP immunizations after baseline subtraction. The maximum binding intensity for each epitope was computed from the highest intensity of binding to a single peptide within the region. As described previously, the binding antibodies elicited after the second MVA immunization showed response rates (i.e., signal intensity value above 200) for linear epitopes comprising the V1b (85% responders), V2 (100%), C5 (71%), gp41 (100%), and Kennedy epitope (KE) (71%) regions of SIVmac239 gp160 (18). The breadth of the response to various regions of gp140 was highly correlated with the magnitude of gp140 binding titers: the animal with the lowest gp140 titers (68 μg/ml) showed responses to 3 regions, comprising the V1b, V2, and gp41-immunodominant (ID) regions, whereas animals with binding titers of >400 μg/ml showed responses to 8 to 10 regions. Figure 5A compares the antibody profiles expressed as median binding intensities at the peak time point after the second MVA immunization and after VLP immunization. Figure 5B shows the heat map aggregate of baseline-subtracted binding intensity values against the gp160 peptide for seven vaccinated animals at peak MVA and VLP time points. VLP immunization significantly increased the breadth of the response, with most new specificities being induced for the C1a and C1b regions (Fig. 5B). In addition, the magnitude of binding to the C5, V2, gp41-ID, and KE regions was significantly enhanced after the VLP boost (Fig. 5C).

Examination of the proportions of responses to different regions in the total gp160 array demonstrated that VLP immunization altered the specificity of the response from a V2-dominant response after the second MVA immunization to a KE-dominant response (Fig. 5D). While the mechanisms driving immunodominance to the KE region, situated in the cytoplasmic tail of gp41, after VLP immunization is not presently clear, it could be raised by looping out of the KE from a transfected cell in vivo or by the presence of disrupted VLPs in the vaccine preparation, as we have observed for our HIV VLP preparations in vitro (Fig. 5G).

We next determined whether antibody responses to gp70-scaffolded V1V2 antigens (Ags) were induced after VLP immunization. The data showed that the anti-gp70 V1V2 IgG antibody titer was increased in the sera of 4/7 animals after VLP immunization. The magnitude of V1V2 responses at the peak time point post-VLP immunization correlated with the magnitude of binding antibody titers to gp120 and gp140 (data not shown). Furthermore, reactivity to gp70-scaffolded V1V2 also directly correlated with the intensity of binding to V1b linear peptides (Fig. 5E). In rectal mucosal secretions, we observed a significant enhancement in IgG-specific activity against V1V2 antigens, which was highly correlated with serum V1V2 responses (Fig. 5F). In summary, this comprehensive analysis demonstrated that a single booster immunization with NP-adjuvanted VLP protein resulted in an increased breadth of antibody responses, a shift in epitope dominance, and an increase in rectal IgG binding responses to the gp70-scaffolded V1V2 protein.

After establishing the ability of a late protein boost to efficiently induce humoral immune responses, we next compared MVA and protein immunizations for the magnitude and quality of cellular responses. We assessed antigen-specific CD4 T cells by stimulating PBMCs at peak and memory time points following each boost with 15-mer peptide pools spanning SIVmac239 Gag and Env. Each of the three booster immunizations induced a strong recall of Gag- and Env-specific CD4 T cells, with responses after the first MVA immunization being significantly higher than those after the second MVA and VLP boosts (Fig. 6A and B). Stratifying Ag-specific responses by Gag and Env revealed an equivalent recall of Gag- and Env-specific CD4 T cell responses after VLP immunization relative to that induced after the second MVA immunization (Fig. 6B).

Next, we evaluated how a VLP protein boost altered the cytokine profile of antigen-specific CD4 T cells. CD4 T cells were examined for the production of IFN-γ, TNF-α, IL-2, and IL-21 after peptide stimulation. Boolean analysis revealed that ∼20% of SIV-specific CD4 T cells produced all four cytokines after the second MVA boost, and this polyfunctional response was substantially reduced after VLP immunization. Indeed, the frequency of cells producing 3 or more cytokines was dramatically decreased after VLP immunization, and antigen-specific CD4 T cells were preponderantly IFN-γ producers, with a large fraction coproducing either TNF-α or IL-21. Whether this marked shift from a polyfunctional to a bifunctional cytokine profile was shaped by the adjuvant or is a feature of recall responses to multiple boosts is not clear. Temporal kinetics of vaccine-induced CD8 T cells showed that VLP immunization did not boost DNA/MVA-induced memory CD8 T cells (Fig. 6D); the lack of anamnestic CD8 T cell responses after VLP immunization is consistent with the inefficiency of protein antigens in engaging the class I cellular machinery needed to elicit effective CD8 responses.

Next, we evaluated vaccine efficacy against a repeat, moderate-dose (200 TCID₅₀), intrarectal challenge with highly stringent SIVmac251 and observed that vaccination did not significantly delay acquisition (Fig. 7A). However, we noted some observations of interest. Of the unvaccinated controls, 40% were productively infected after the first challenge, the majority were infected after the second challenge, and all unvaccinated controls were infected after the fourth challenge. In the DM group, 30% of animals became infected after the first challenge, 80% became infected after the second challenge, and it took 5 challenges to infect the remaining animal. Inclusion of the VLP boost delayed acquisition, with 57% of animals in the DM plus VLP (DM+VLP) group becoming infected after the second challenge and 28% resisting infection (2/7 protected) (Fig. 7A). Of the two protected animals in this group, animal RWp13 showed a blip, with 90 SIV RNA copies/ml after the fourth challenge and remained below 60 copies after the fifth challenge. Animal RUe13 showed apparent sterilizing protection, with viral copy numbers being below 60 copies during and subsequent to 5 challenges. Viral copy numbers in both animals remained at <60 copies of plasma SIV RNA for 24 weeks postinfection.

We next asked whether vaccine-induced humoral responses correlated with the acquisition of infection in either the DM or the DM+VLP group. Neither binding titers nor neutralization and nonneutralizing responses were significantly associated with acquisition. However, the magnitude of the Env IgG response to gp70-scaffolded V1V2 proteins at the peak time point postboost in sera was positively associated with a delay in acquisition in vaccinated animals (Fig. 7B).

Although neither vaccine regimen had a significant impact on acquisition, vaccination significantly aided in acute viral control (Fig. 7C). The analysis of viral loads was performed only on infected animals, and protected animals were not included. At 3 weeks postinfection, unvaccinated animals showed a robust peak in viremia ranging from 5.3 to 8 log RNA copies/ml, while the viral load in vaccinated animals ranged between 2 and 6 log RNA copies/ml plasma. However, we noted that the peak viral load in VLP-boosted groups was higher, with significantly higher viral loads at week 1 (Fig. 7D), while at week 2 and 3, a trend for higher viral loads was observed (Fig. 7D). We did not observe an association with Mamu A*01 status (Fig. 7D, green circles) and peak viremia. By week 7, VLP-boosted animal achieved viral control similarly to the DM groups, and at week 24, animals in the DM+VLP group showed significantly lower viral loads than did non-Mamu A*01 animals in the DM group (Fig. 7D). Unexpectedly, we observed viral control in some of the unvaccinated animals by 24 weeks postinfection (Fig. 7E).

To gain an understanding of how vaccine-induced immune responses influenced early viral control, we examined SIV-specific T cell and antibody responses following infection. Vaccinated animals showed an overall higher frequency of Gag- and Env-specific CD8 T cells at weeks 3 and 7 postinfection than did unvaccinated controls (Fig. 8A). However, at week 3 postinfection, VLP-boosted animals showed on average lower frequencies of SIV-specific CD8 T cells than did the DM group (Fig. 8A). Because VLP immunization did not boost DNA/MVA-induced memory CD8 T cells, it is possible that the longer time interval between the final MVA boost and challenge in VLP-boosted animals contributed to the slower recall of Gag-specific CD8 T cells following infection. Indeed, the magnitude of the Gag-specific CD8 T cell response at week 3 postinfection was significantly associated with acute viral control in both vaccinated and control animals.

Similarly, the magnitudes of SIV-specific CD4 T cell responses were higher in vaccinated animals and were significantly higher in the DM group than in the DM+VLP group at the peak time point (Fig. 8B). The magnitude of the Gag-specific CD4 T cell response was inversely associated with viral load in control animals, consistent with virus-induced CD4 depletion. Assessment of gp140-specific IgG responses showed robust recall in vaccinated groups at significantly higher titers than those in controls. Unlike the T cell responses, the DM+VLP group mounted gp140 responses that were equivalent to those mounted by the DM group at weeks 3 and 7 postinfection (Fig. 8C). Titers peaked at week 7 in unvaccinated controls, and the magnitude of titers was inversely associated with viral load. In summary, the data show that DM+VLP immunization resulted in a trend for better protection and evidence for better viral control at week 24 than for the DM group.