In order to avoid imbalance in covariates likely to influence outcomes that may result from an unfortunate instance of simple random sampling, the method of rerandomization (20) was employed to randomly assign RMs to four groups: three experimental groups of 12 animals (Groups K, L and M) as well as a titration group (n=6) to confirm the infectious challenge virus dose (a second titration group (n=6) was enrolled later). Covariates to be balanced across groups were age, weight, number of males with which each animal was previously co-housed, and number of offspring (Table 1). For each of 500,000 completely random allocations to study group, the Mahalanobis distance (21) was computed as the scaled difference between vectors formed in such a way that all pairwise differences in covariates between groups were computed for a given randomization instance, with the scaling matrix block-diagonal, with each block obtained as the generalized inverse of the covariance matrix of the variables of interest from all animals available for randomization. A cutoff value was chosen such that ~5% of the best (most balanced) potential randomizations were eligible to be selected for use. The subset was further limited by requiring no cells in the contingency table of FcR3A by group to be 0, no group would have fewer than 4 of the individuals previously co-housed with no males, and no group would have fewer than two Mamu A*01-positive animals, and two animals which had already been designated for the titration group were always assigned to this group. In order to enable randomization hypothesis tests across the three study groups, the final set of eligible randomizations to be used were then obtained via implementation of the same method but excluding those randomized to the titration group, by retaining those with the dissimilarity metric less than the median of those used in the first round (Table 1).

HIV-derived antigens inserted into the virosome membrane were previously described (14, 15, 22, 23). The synthetic P1 lipopeptide (amino acid sequence 649-683 of gp41) was produced by Bachem AG (Bubendorf, Switzerland). The recombinant gp41-derived antigen (amino acid 540-664 with a deletion of 25 amino acids from 593-617, plus a C-terminal 4xHis-tag for purification followed by a free cysteine) was expressed in E. coli and purified as trimers under non-denaturing conditions by PX’Therapeutics (Grenoble, France). Lipidation of the C-terminal cysteine of the rgp41 to 1,2-Dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-[4-(p-maleinimidomethyl)cyclohexane-carboxamide]) (N-MCC-DPPE, Corden Pharma, Liestal, Switzerland) allowed antigen anchorage into the virosome lipid membrane produced under liquid form, as described (14). The antigens were dissolved in 100 mM octaethylene glycol monododecyl ether (OEG, Sigma, Buchs, Switzerland) prepared in HN buffer (50 mM HEPES pH 7.4, 142 mM NaCl) and added to the virosome excipients during manufacturing. The final HIV-1 liquid virosomal vaccine MYM-V201 for the i.m. route (0.5 ml) contained 85 µg/ml hemagglutinin (HA), 90 µg/ml P1, 130 µg/ml rgp41, and was supplied in HN buffer pH 7.4. For the i.n. formulation, a similar dose was delivered but in 0.2 ml volume (0.1 ml per nostril) using the BD Accuspray. Nanoparticle tracking analysis (NTA) for the virosome particle size was performed on a Malvern NS300 instrument. Dynamic light scattering (DLS) to determine virosome population homogeneity based on the polydispersity index (PDI) was performed on a Malvern Zetasizer Nano. Microbiological quality was determined according to E.P. section 5.1.4. Absence of specific microorganisms was demonstrated according to E.P. section 2.6.13 - Pseudomonas aeruginosa and Staphylococcus aureus.

Four weeks before the first virosome administration, Groups K, L, and M were given 10 μg of HA from inactivated influenza virus (strain A/Brisbane/59/2007 H1N1) to mimic the pre-existing natural or vaccine-induced anti-influenza immunity in humans. This influenza virus was propagated in the allantoic cavity of embryonated eggs and purified as described (24). The vaccine doses were 45 µg of P1 and 65 µg of rgp41 for each of the i.m. and i.n. immunizations. The immunization schedule for the RMs in Groups K, L, and M is depicted in Figure 3; all investigators and veterinary staff were blinded during the vaccinations.

The challenge virus, SHIV_SF162P3 (a tier 2, R5 clade B strain (18, 19), was kindly provided by Dr. Nancy Miller, NIAID (NIH-AIDS Research and Reference Reagent Program (NIH-ARRRP)); the genome of this SHIV contains the env, tat, rev, and vpu genes of HIV-1 SF162 inserted into the backbone of the pathogenic SIVmac239. The stock was grown in RM PBMC; it had a p27 concentration of 183 ng/ml and 3.5 x 10⁵ 50% tissue culture infectious doses (TCID₅₀)/ml as measured in TZM-bl cells. Using ELISA, the total gp41 concentration present in the virus stock was found to be 11 ng/ml. Therefore, the gp41 concentration was 0.13 ng/ml for the first seven challenges (virus stock diluted 1/86 to obtain 15 TCID₅₀), and 0.19 ng/ml for the subsequent challenges (virus stock diluted 1/57 to obtain 22 TCID₅₀ as measured in RM PBMC); these gp41 concentrations were used to estimate the approximate molar ratio of gp41 antigen molecules to specific Ab molecules in vaginal fluid.

The SHIV_SF162P3 stock used in the Chinese RM study (14) was no longer available for our study in Indian RMs. Hence, the new virus stock was tested in RMs by intravaginal titration experiments, as in vivo infectivity for a given challenge route is the gold standard. Our most important criterion was to achieve persistent viral infection using approximately the same number of intravaginal challenges, at the same time intervals, as in the original study (14). We enrolled a total of 12 Indian RMs in the titration to extrapolate the new SHIV_SF162P3 stock’s intravaginal challenge dose. The latter was also expressed as TCID₅₀ (measured in Indian RM PBMC) to show that its infectivity was in the same order of magnitude as that of the stock used previously in Chinese RMs. Importantly, TCID₅₀ measurements are only operational terms but not absolute units of measure. TCID₅₀ values for any given stock vary from lab to lab for the following reasons: i) PBMC collected from RM donors are virus target cells that cannot be standardized given outbred nature of RMs; ii) TCID₅₀ data also depend on the cell culture conditions and individual laboratory procedures, and iii) assay readout sensitivity can vary greatly. Consequently, TCID₅₀ values measured in primary RM PBMC give only order-of-magnitude information rather than absolute values.

Following similar timeline and virus TCID₅₀ doses as described in the Chinese RM study (14), five weeks after the second i.n. boost (week 29), all animals were inoculated intravaginally with low-dose virus once or twice per week, with 15 TCID₅₀ of the heterologous SHIV_SF162P3 for the first 7 challenges and 22 TCID₅₀ for challenges 8 to 22. The veterinary procedures outlined in Chenine et al. (25) were used; RMs undergoing vaginal SHIV challenges were not treated with hormones.

Plasma vRNA was isolated by QiaAmp Viral RNA Mini-Kits (Qiagen, Germantown, MD, USA); vRNA levels were measured by quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) for SIV gag sequences (26, 27). Assay sensitivity was 50 copies/ml.

Time-to-event methodology was used to evaluate the occurrence of primary and secondary endpoints. As the intravaginal challenges were not equally spaced, the time variable is presented and analyzed in units of study days, rather than number of challenges. The endpoints analyzed included: time to first viremia, defined as the first observation of viral load >50 copies/ml; time to peak viremia, using the study day of observed peak viral load; and time to persistent systemic infection (PSI), in which the endpoint is reached when viral RNA is ≥10,000 copies/ml. As the challenge dose was increased from the 8^th challenge onward (Challenge Phase II), results are presented overall (Challenge Phases I plus II) and also with endpoints censored on day 32, the date of the 8^th challenge (Challenge Phase I).

Randomization tests were used to compare endpoints across group. The survival (28) and interval (29) packages for R statistical software were used to compute the corresponding log-rank test statistic for each eligible randomization instance, with p-values obtained as the number of test statistics exceeding the value of the statistic for the actual randomization. As a sensitivity evaluation of the randomization test method, the standard log-rank p-value obtained from the null distribution, which assumes random allocation, rather than the restricted allocation described above, was also computed. Vaccine efficacy is computed as one minus the ratio of infection rates (vaccinated vs control group), with supportive confidence intervals obtained by the exact conditional binomial method, assuming Poisson distribution of endpoints within each group (30). No adjustment for multiplicity was done.

HIV-1 gp41-derived antigens for ELISA plates. The P1 peptide powder stock was dissolved first in hexafluoroisopropanol prior adding protease-free sterile water for obtaining a solution stock at 0.7 mg/ml in 10% hexafluoroisopropanol. P1 peptide was then diluted at 2 µg/ml. The frozen rgp41 stock was thawed and diluted with cold phosphate-buffered saline (PBS) pH 7.4 to obtain a solution at 2 µg/ml.

To coat 96-well MaxiSorp flat bottom plates (Nunc), 0.1 ml of freshly prepared antigen solution was added to wells (0.2 µg/well). Plates were wrapped into aluminum foil and incubated 16 h at 4°C. The next day, plates were washed four times with 300 µl/well with PBS pH 7.4 with Tween 20 at 0.05% (PBST) prior adding 300 µl/well of casein as blocking reagent (Abcam, ab 171535), followed by a 2 h incubation at 37°C. Plates were then washed four times with 300 µl/well with PBST and 100 µl of diluted sera, positive RM serum controls, and human monoclonal antibodies (mAbs) 2F5 (31) and 98.6 (32), both from NIH-ARRRP were added to the appropriate wells, plates were wrapped in aluminum foil, and incubated for 2 h at 37°C. Plates were then washed 8x with 300 µl/well with PBST and 100 µl of diluted rabbit anti-monkey IgG horse radish peroxidase (HRP) into PBST (Sigma, A2054, dilution according to the supplier’s recommendation) were added to wells and plates wrapped in aluminum foil prior placing them into the incubator at 37°C for 1 h. Wells were then washed 8x with 300 µl/well, followed by the addition of 100 µl of substrate (TMB single solution, Life Technologies) and plates incubated in the dark (wrapped in aluminum foil) at room temperature for 10-15 min or until the color signal was considered strong enough, then the reaction was stopped by the addition of 100 µl/well of 1 M H₂SO₄ and the optical density of the reaction was read at 450 nm.

NAbs were measured as a function of reduction in luciferase (Luc) reporter gene expression after a single round of infection in either TZM-bl cells (33, 34), TZM-bl/FcγRI cells (35) or A3R5 cells (36). TZM-bl cells (also called JC57BL-13) were obtained from the NIH-ARRRP, as contributed by John Kappes and Xiaoyun Wu. This is a HeLa cell clone engineered to express CD4 and CCR5 (37) and to contain integrated reporter genes for firefly luciferase and E. coli beta-galactosidase under control of an HIV-1 long terminal repeat (LTR) (38). The cells were engineered by stable transduction to express human FcγRI cells, which made them ultra-sensitive for detecting gp41 MPER-specific nAbs (35).

For assays in TZM-bl and TZM-bl/FcγRI cells, a pre-titrated dose of virus was incubated with serial 3-fold dilutions of heat-inactivated (56°C, 30 min) serum samples in duplicate in a total volume of 150 μl for 1 h at 37°C in 96-well flat-bottom culture plates. Freshly trypsinized cells (10,000 cells in 100 μl of growth medium containing 75 μg/ml DEAE dextran) were added to each well. One set of control wells received cells + virus (virus control) and another set received cells only (background control). After 48 h of incubation, 100 μl of cells was transferred to a 96-well black solid plate (Costar) for measurements of luminescence using the Britelite Luminescence Reporter Gene Assay System (PerkinElmer Life Sciences). Neutralization titers are the dilution (serum/plasma samples) or concentration (mAbs) at which relative luminescence units (RLU) were reduced by 50% or 80% compared to virus control wells after subtraction of background RLUs. Assays in TZM-bl and TZM-bl/FcγR1 cells used a rhesus PBMC-grown stock of SHIV_SF162P3.R157.

A3R5 (A3.01/CCR5), a derivative of the A3.01 human lymphoblastoid cell line that naturally expresses CD4 and CXCR4, was engineered to express CCR5 (36). The A3R5 assay was performed with Env.IMC.LucR viruses as described (36). Briefly, serum and plasma samples were assayed at 3-fold dilutions ranging from 1:20 to 1:43,740. Neutralization titers are the sample dilution at which RLU were reduced by 50% as compared to RLU in virus-control wells after subtraction of background RLU in cell-control wells. Assays in A3R5 cells used SHIV-SF162P3.LucR infectious molecular clone produced by transfection in 293T cells.

The ADCP assay was adapted from (39). Briefly, gp140 SHIV_SF162p3 and the recombinant Mymetics gp41 antigens were biotinylated using sulfo-NHS LC-LC biotin, coupled to yellow-green, fluorescent Neutravidin 1 μm beads (Invitrogen, F8776) for 2 h at 37°C and washed two times in 0.1% bovine serum albumin (BSA) in PBS. Ten μl/well of coupled beads were added to 96-well plates with 10 μl/well of diluted sample for 2 h at 37°C to form immune complexes. After incubation, the immune complexes were spun down, and supernatants were removed. THP-1 cells were added at a concentration of 2.5 x 10⁴ cells/well and incubated for 18 h at 37°C. After incubation, the plates were spun down, the supernatant was removed, and cells were fixed with 4% paraformaldehyde (PFA) for 10 min. Fluorescence was acquired with a Stratedigm 1300EXi cytometer. Phagocytic score was calculated using the following formula: (percentage of FITC+ cells) * (the geometric mean fluorescent intensity (gMFI) of the FITC+ cells)/10,000. A polyclonal HIVIg pool available from the NIH AIDS Reagent Program was used as a positive control. Serum from a human HIV-seronegative donor was used as negative control.

The ADNP assay was adapted from Karsten et al. (40); gp140 SHIV_SF162p3 and the recombinant Mymetics rgp41 antigens were coupled to beads and immune complexes were formed as described for ADCP. Neutrophils were isolated from fresh whole ACD-anticoagulated blood using EasySep Direct Human Neutrophil Isolation kit (Stem Cell, 19666), resuspended in R10 medium, and added to plates at a concentration of 5x10⁴ cells/well. The plates were incubated for 30 min at 37°C. The neutrophil marker CD66b (Pacific Blue conjugated anti-CD66b; BioLegend, 305112) was used to stain cells. Cells were fixed for 10 min in 4% PFA. Fluorescence was acquired with a Stratedigm 1300EXi cytometer and phagocytic score was calculated as described for ADCP.

HIV-1-specific antibodies were measured by HIV-1 Binding Antibody Multiplex Assay (BAMA) for IgG and IgA as described (41–44). For IgA assays, the samples were depleted by Protein G for sensitivity of IgA detection. The following antigens were used for both IgG and IgA assays: Bio-MPR.03 (MPER) (NEQELLELDKWASLWNWFDITNWLWYIR), MYM gp41 (vaccine gp41 from Mymetics), MYM-P1-PE (P1 from Mymetics), SP62 (MPER-2F5 epitope, QQEKNEQELLELDKWASLWN), gp41 (recombinant MN, Immunodiagnostics) and for IgA, two antigens that corresponded with decreased HIV-1 risk in RV144 were included: 00MSA 4076 gp140 (clade A gp140), A1.con.env03 140 (consensus clade A gp140) (45, 46). Assays were run under GCLP compliant conditions, including tracking of positive controls by Levy-Jennings charts using 21CFR Part 11 compliant software. Positive controls included a HIVIG and SHIVIG (DBM5, purified IgG from SHIV-infected RMs kindly provided by Dr. Mario Roederer, Vaccine Research Center). Additional positive controls included 2F5 IgG, 7B2 IgG, 4E10 IgG and purified RM IgA and negative controls included in every assay were blank well control and uncoupled beads. Antibody measurements were acquired on a Bio-Plex instrument (Bio-Rad, Hercules, CA, United States) using 21CFR Part 11 compliant software and the readout is in Mean Fluorescence Intensity (MFI). The preset assay criteria for sample reporting are: coefficient of variation (CV) per duplicate values for each sample were ≤15% and >100 beads counted per sample. The preset positivity criteria were: 1. MFI-blank well-blank bead ≥antigen specific MFI cutoff (95^th percentile of W0 or 100 MFI) and 2. MFI-blank well-blank bead > 3X W0 MFI-blank well-blank bead and 3. MFI-blank well > 3X W0 MFI-blank well.

Fc array analysis was performed blinded to group as described (47, 48) to evaluate polyclonal Ab responses. Briefly, plasma was centrifuged for two min at 14.8 x g, then aliquoted and analyzed for binding to a panel of antigens listed in Supplementary Table 1. Plasma was diluted 1:1,000 for detection reagents (Supplementary Table 2) of tetramerized Fcγ receptors (Source: Duke Protein Production Facility) and anti-rhesus IgG (Source: Southern Biotech). For increased sensitivity, detection reagents C1q (Source: Sigma) and anti-human IgA (Source: Southern Biotech) were collected with a plasma dilution of 1:250, along with a replicate of the anti-rhesus IgG. Data collection was performed using Luminex Exponent version 4.2 software.

We utilized the flow-based GranToxiLux (GTL) assay to test the samples. Recombinant HIV-1 Con-S gp140 was used to coat the cells in the GTL assay (49, 50). The cut-off for positivity in the GTL assay was >8% of Granzyme B activity. The recombinant gp140 was chosen because it is the best immunogen to capture anti-Env binding Abs; gp140 was selected to detect the presence of ADCC responses directed against gp120 and gp41 epitopes. We also tested the samples using the Luciferase-based (Luc) ADCC assay against the SHIV_SF162P3.5 that expresses the Luciferase reporter genes (51). The analysis of the results was conducted after subtracting the background detected with the pre-immunization samples. After background subtraction, results would be considered positive if the percent specific killing is >15%. The ADCC laboratory tested samples collected from the pre-immune and immunized animals at weeks 25, 26, 29, and 33.

Imperacer^® combines the ELISA-based method with the qPCR technique to amplify the artificial DNA, conjugated to the detecting molecule (52–55). As previously described (23), DNA-labeled P1 and DNA-labeled rgp41 (Chimera proprietary expertise) were used in bridging assays to quantify specific IgG and IgA antibodies and DNA-labeled IgG anti-IgG or anti-IgA to quantify total IgG and IgA antibodies in a sandwich Imperacer assay. This method is very sensitive (lower detection limit for the current bridge assay is ≥ 0.001 ng/ml), specific, and species independent, as it can detect a broad range of antibody concentrations of any isotype and from any animal origin. Real-time PCR signals were converted to approximate antibody concentrations (ng/ml) by analysis against a reference antibody curve. These antibody concentrations were provided only as indicative values.

The prime goal of the current study was to confirm the strong protection, which had been reported for the efficacy study performed earlier in Chinese RMs (14), in the more widely used Indian RM subspecies. Thus, detailed immunogenicity analyses were contingent on first demonstrating the reproducibility of vaccine protection, and a “go/no-go” decision to initiate intravaginal SHIV challenges involved only a minimal confirmation of vaccine immunogenicity. Since both experimental Groups K and L (Figure 3) had received virosomes displaying the P1 peptide, we measured anti-P1 Ab responses using ELISA. For each animal, plasma samples collected before the first vaccination and at specific time points during and post vaccination were tested. Animals from Groups K and L had measurable anti-P1 Abs compared to autologous pre-immune samples (data not shown); the anti-P1 responses in two of the three groups led us to start the SHIV challenges.

In Group M (given placebo virosomes; Figure 3), no anti-P1 Abs were detectable. Only two scientists leading the NHP study were aware that Group M RMs were controls; veterinary staff directly involved in the animal experiments did not know the status of any of the three groups during the ensuing SHIV challenges. Regular viremia levels were assessed immediately to decide whether SHIV challenges should continue for any given RM; the experimental design called for stopping challenges once an RM reached vRNA levels ≥10⁴ copies/ml.

As the primary goal of this current vaccine efficacy study in Indian RMs was to confirm the efficacy observed in Chinese RMs (14), SHIV challenges were performed as in the original study in two phases, starting with 15 TCID₅₀ (as assessed in RM PBMC) for the first seven challenges (Challenge Phase I, Figures 4A–C), followed by a 50% increase of the inoculum from challenge 8 onwards (Challenge Phase II, Figures 4D–F). The results of the plasma vRNA loads for Challenge Phase I up to day 32 are shown in Figures 4A–C. During this time span, Groups L and M showed significant differences by three parameters (Table 2). Using time-to-first viremia, efficacy was 78.4% (p=0.0456); using time-to-peak viremia, efficacy was 85.3% (p=0.0359), and time-to-persistent systemic infection (PSI, defined as vRNA loads ≥10⁴ copies/ml), efficacy was 87% (p=0.0319) (Figure 5). In contrast, the analysis of vRNA loads of Group K versus M revealed no protection. We conclude that compared with control Group M, Group L animals had significant vaccine protection during Challenge Phase I; these animals had been vaccinated with the combination of virosomes displaying either P1 or rgp41 (Figure 3).

Following the blueprint of the study in Chinese RMs (14), the SHIV inoculum had to be increased by 50% from intravaginal challenge #8 onwards. One week later, five previously aviremic RMs in Group L became infected and progressed to PSI (Figure 4E; red dots in green box, Figure 6). The protection seen during Challenge Phase I was no longer seen, and Group K versus Group M again showed no evidence of protection. An in-depth statistical analysis of both Challenges Phase I and II can be found in Table 2.

The abrupt loss of protection in Group L immediately following exposure to the higher challenge virus dose implies a threshold effect, according to which vaccine-induced host immune defenses were able to hold the incoming virus at bay as long as the inoculum was not overpowering. Essentially, the situation from challenge #8 onwards is akin to a flash flood overcoming protection provided by a dam (Figure 6; please see Discussion). Nevertheless, our analysis is focused on Challenge Phase I in an attempt to identify and describe any protective immune responses observed to be associated with the protection that was evident only in Phase I; importantly, our primary analysis was not limited to this part of the overall study.

Given the significant protection seen in Group L during Challenge Phase I, we decided to proceed with analyzing the vaccine-induced antibody responses over time, including pre-immunization, during the i.m. and i.n. vaccine administrations, and at week 29, the day of the first virus exposure (Figure 7). Individual data points reflect values of each experimental RM. Animals represented by red symbols had breakthrough infections during Challenge Phase I; their anti-rgp41 IgG responses revealed no obvious link with the SHIV challenge outcome.

Next, we sought to determine whether the sudden loss of protection seen in Group L vaccinees might be due to decreases in vaccine-induced, antigen-specific Ab levels. On day 38, the first plasma samples collected after the 50% challenge SHIV dose escalation, revealed that five of the Group L animals had sudden breakthrough infection (large red dots in green boxes, Figures 6, 8, animals L1, L5, L7, L9, and L10). Figure 8 shows the anti-rgp41 (blue lines) and anti-P1 IgG responses (green lines) during the intravaginal SHIV challenges. There were no drops in the antigen-specific IgG responses just preceding the breakthrough infections; however, shortly after viremia became apparent, there was a marked boost in the vaccine-induced Ab responses, especially in animals L1, L11, and L12. Clearly, waning of vaccine-induced Ab responses did not account for the breakthrough infections seen after the SHIV dose escalation for challenge #8.

Next, we assessed whether Group L vaccinees had developed nAbs against the challenge virus, the tier 2 SHIV_SF162P3. None were detected at any time points in the plasma/serum samples examined (data not shown). This finding prompted us to examine the epitope specificity of the vaccine-induced IgGs by Binding Antibody Multiplex Assay (BAMA); the data are shown in Figure 9. Consistent with the lack of nAb responses in the functional assays, no reactivity was found against the 2F5 epitope (56) by IgG BAMA in Group L animals (Figure 9, top panels). This epitope, recognized by the broadly neutralizing mAb 2F5, is located in the MPER, and reactivity to this epitope would have been expected in animals vaccinated with virosomes displaying the P1 peptide. However, the experimental RMs had high background reactivity, including controls in Group M (Figure 9), possibly due to antigen mimicry between HIV gp41 regions and microbiome antigens (57, 58). For Group L, the anti-gp41 antibody responses were generally much higher during the vaccination phase compared to anti-P1 antibody responses (panels MYM gp41 versus MYM-P1-PE; Figure 9). In parallel, plasma IgA responses were also tested by the BAMA methods; no responses were found (data not shown).

IgG responses specific for the rgp41 antigen were tested in RMs of Groups L and M (Figure 9, 3^rd row of panels from top). No background reactivity was detected in any preimmune samples. Among Group L vaccinees, 7 out of 12 RMs mounted such IgG reactivity during the vaccination phase (blue symbols, Figure 9; 3^rd left panel). After the SHIV challenges resulted in infection, these specific anti-rgp41 IgG responses formed de novo in 3 out of the 11 control Group M animals.

Next, we performed Fc array analysis to examine the IgG features between Groups L vs M and K vs M. Briefly, plasma from the experimental animals at the time points indicated were incubated with beads conjugated to target antigens of interest followed by detection with FcγR, lectins, as well as IgG and IgA detection reagents. The data are presented as volcano plots that depict the fold change and significance of differences between the experimental groups versus control Group M (Figure 10). Between one week and two weeks after the second i.n. boost, Group L showed clear, specific reactivity compared to Group M; significant differences were also present on the day of the first SHIV challenge (Figure 10, top row, 4^th panel). The difference was no longer noticed on the day of SHIV challenge #7. In contrast, none of the time points showed any significant differences in volcano plots comparing Group K vs Group M.

We ascribe the differential outcome of the Fcγ arrays, and the functional assays for Fcγ-related activities, to differences in assay sensitivity, with the Fcγ array being significantly more sensitive. Overall, the volcano plots revealed significant differences between vaccinees in Group L and controls in Group M with regards to antigen-specific FcγR2A, FcγR2B, FcγR3A, and FcγR3B. Figure 11 highlights specific reactivity to P1 peptide (top panels) and rgp41 (bottom panels). The three experimental groups are color coded as in other figures (Black, Group M; green, Group K; blue, Group L). After the second i.n. boost and on the day of challenge #1, the levels of antigen-specific IgG binding to FcγR2A-4 are increased in Group L vaccinees, compared to the levels in the other two groups.

Prompted by the finding of increased antigen-specific IgG to FcγR2A, we sought to correlate increased binding with functional activities. FcγR2A is preferentially expressed on phagocytes, including macrophages, monocytes, neutrophils, and dendritic cells, and plays a role in phagocytosis. Thus, we examined plasma samples for ADCP and ADNP.

Overall, antigen-specific functional activities of the samples were low, with only a few data points above the negative control. On the day of challenge, neither ADCP nor ADNP were above background (Figure 12). Importantly, the samples with increased ADCP at week 33 (four weeks after the initiation of SHIV challenges) also demonstrated increased ADNP activity, suggesting the observed responses were vaccine-induced, as ADNP and ADCP activity often track together. Although Ab titers were not evaluated as part of the ADCP/ADNP assays, the weak functional responses coupled with the lack of detectable antigen-specific glycan suggests that relatively low-titer responses were elicited in the vaccine trial.

Finally, we sought to determine the concentrations of antigen-specific IgG + IgA in vaginal washes using the ultrasensitive Imperacer^® assay (Methods) that takes advantage of PCR amplification in the last step of an ELISA. This assay allows measurements of Ab concentrations with a sub-ng/ml sensitivity. Vaginal washes were collected at weeks 15 and 25, which corresponds to one week after the first or second i.n. boosts, respectively. Only the sum of the IgG + IgA responses could be determined (Figure 13). Week 25 corresponds to four weeks prior to the first intravaginal SHIV challenge; no vaginal wash samples were collected after this time point to avoid introducing microabrasions that could compromise mucosal barrier integrity. Of note, absolute antigen-specific Ab responses were higher in Group L compared to Group K – possibly explaining the initial vaccine protection seen in Group L vaccinees but not in Group K animals.