Soluble HIV-1 Env trimers in adjuvant elicit potent and diverse functional B cell responses in primates

Memory B cells do not produce Abs unless they are reactivated by antigen or polyclonal stimulation to differentiate into ASC. Enumeration of total and antigen-specific memory B cells is greatly facilitated by methods to differentiate memory B cells into ASC to allow their detection by B cell ELISpot analysis. Several cytokines, mitogens, and toll-like receptor ligands were described as important factors for the proliferation and differentiation of human memory B cells into ASC (Arpin et al., 1995; Krieg et al., 1995; Crotty et al., 2004; Amanna and Slifka, 2006; Huggins et al., 2007). To establish protocols for efficient macaque memory B cell differentiation, we used rhesus and cynomolgus macaque PBMCs collected 14 d after Env immunization. At this time, circulating PCs are not detected by direct B cell ELISpot analysis, whereas memory B cells require a period of in vitro expansion and differentiation into ASC before they are detected (Fig. S1). The ELISpot assay used in this study is based upon a recently described protocol where total IgG is captured on the plate and antigen-specific IgG is detected with biotinylated Env probes, generating high definition spots with low background (Dosenovic et al., 2009).

We evaluated a panel of stimuli for in vitro differentiation of macaque memory B cells. The efficiency of the stimulations varied considerably, with a combination consisting of PWM, SAC, and CpG (PWM + SAC + CpG) being the most potent under the conditions tested in this study (Fig. 1 B). There was a strong correlation between the number of Env-specific and total IgG⁺ memory B cells per million cultured PBMCs, indicating nonbiased polyclonal stimulation (r² = 0.9209; Fig. 1 B). A strong correlation between the number of Env-specific memory B cells detected by ELISpot assay and secreted Ab measured by ELISA was also observed (r² = 0.9273; Fig. 1 C). To compare the quality of the response stimulated by PWM + SAC + CpG or IL-2 + IL-10 + CpG + CD40L, we also measured the kinetics of the response. For both stimulation conditions, the highest number of ASC (total and Env-specific) was present in the culture after 4 d and declined thereafter (Fig. S2 A), whereas maximal accumulation of Abs in the supernatant required an additional 4 d of culture (Fig. S2 B). The frequencies of Env-specific ASC and secreted IgG in the supernatant were similar at all time points measured for both stimulation conditions (Fig. S2, A and B).

The phenotype of the B cells cultured in the presence of PWM + SAC + CpG or IL-2 + IL-10 + CpG + CD40L was also investigated. As shown in Fig. 2 A, the largest fraction of IgG-expressing cells (20%) of the CD4/CD8-negative population was found in the PWM + SAC + CpG–stimulated cultures. The corresponding population in the IL-2 + IL-10 + CpG + CD40L–stimulated culture was 7.8%. As expected, only cells displaying low CD20 expression were IgG producers and no CD20 down-regulation was observed in nonstimulated control cultures (Fig. 2 A). To assess the proliferative capacity of lymphocytes in response to the different stimuli, PBMCs were CFSE stained and incubated with PWM + SAC + CpG, IL-2 + IL-10 + CpG + CD40L, or CpG alone. The proliferation of T cells (CD4⁺, CD8⁺, and CD20⁻) and B cells (CD4⁻, CD8⁻, and CD20^+/low) was evaluated after 6 d of culture. Consistent with the B cell ELISpot results, the PWM + SAC + CpG cocktail stimulated B cell expansion and differentiation most efficiently. This cocktail also stimulated T cell proliferation (Fig. 2 B). In contrast, IL-2 + IL-10 + CpG + CD40L and, to a lesser extent CpG alone, stimulated B cell proliferation and down-regulation of CD20 but not T cell proliferation. To further elucidate the role of T cells for the responsiveness of human and rhesus PBMCs to stimulation by PWM, PWM + SAC, PWM + SAC + CpG, IL-2 + IL-10 + CpG + CD40L, or CpG alone, B cells were sorted and cultured in the presence or absence of autologous T cells and the different stimuli. The results demonstrate that the stimulatory effect of PWM on both human and rhesus memory B cells was dependent on the presence of T cells. Only the IL-2 + IL-10 + CpG + CD40L cocktail induced differentiation of human and rhesus memory B cells in the absence of T cells (Fig. 2 C). The data also show that human memory B cells were more responsive to CpG than rhesus memory B cells under the assay conditions used in this study (Fig. 2 C).

To examine the kinetics by which antigen-specific Abs, memory B cells, and PCs develop in response to Env immunizations, rhesus macaques were inoculated with HIV-1 Env trimers in adjuvant (immunized) or adjuvant alone (controls) via the i.m. route. We used cleavage-defective soluble gp140 trimers (gp140-F) derived from the primary clade B isolate YU2. The trimers possess a heterologous trimerization motif from fibritin (F), as previously described (Yang et al., 2002). The schedule for inoculations and sample collections is shown in Fig. 3 A. Plasma and PBMCs were collected before the first protein inoculation and 1 and 2 wk after each inoculation. Bone marrow samples were collected 3 wk after the second and fourth inoculations. 1 mo after the last inoculation, all animals were challenged with SHIV-SF162P4 to assess the protective capacity of the immune response.

Plasma from Env trimer-inoculated animals were tested for binding Abs to YU2 Env by ELISA (gp120; Fig. 3 B). The data are shown as half-maximum binding titers of Env-specific IgG and IgA (Log10 optical density [OD] 50 titers). Env-specific IgG responses were detected 2 wk after the first inoculation and they were boosted by the second inoculation, resulting in OD50 titers of ∼10⁵, which were maintained after the subsequent inoculations. Env-specific IgA responses were ∼1.5 log lower but followed the same kinetics. The response was comparable between individual animals for both IgG and IgA. When the plasmas were examined for increases in functional affinity, often referred to as avidity, using a more stringent binding assay involving incubation in 1.5 M NaSCN (see Materials and methods), we recorded an enhancement in avidity with repeated boosting (Fig. 3 C). These data demonstrate that i.m. immunizations with Env trimers administered in Abisco-100 and CpG ODN 2395 elicit potent plasma Ab responses, which evolve with repeated antigen exposure. To investigate if this increase was selective for subspecificities of Env, we performed the avidity assay with three trimeric YU2-based Env probes: gp140-F, gp120-F, and gp120ΔV1/2/3-F (Dosenovic et al., 2009). We observed a significant increase in the avidity index between the samples after immunization two and after immunization five for all three probes (Fig. 3 D). Furthermore, there was no significant difference between the avidity indexes for the three probes at either time point (Fig. 3 E). These results indicate that the increase in avidity between the second and fifth immunization was not selective for variable region 1–3 determinants (which are absent from gp120ΔV1/2/3-F) or for gp41 determinants (which are absent from gp120-F and gp120ΔV1/2/3-F) but probably reflect overall avidity increases for several determinants displayed by these complex glycoprotein immunogens.

To examine the dynamics of Env-specific memory B cells and PCs in the Env trimer-inoculated animals over the prime/boost regimen, we collected PBMCs 1 and 2 wk after each inoculation. The cells were either plated directly in ELISpot plates for detection of PCs or stimulated in vitro for 4 d with PWM, SAC, and CpG for detection of memory B cell–derived ASC. Circulating PCs are presented as ASC per 10⁶ PBMCs (Fig. 4 A, top) and as the percent Env-specific of total IgG-producing cells (Fig. 4 A, bottom). Circulating memory B cells are presented as ASC per 10⁶ cultured PBMCs (Fig. 4 B, top) and as the percent Env-specific of total IgG-producing cells (Fig. 4 B, bottom). The results demonstrate that the frequencies of total IgG⁺ PCs and memory B cells were stable during the course of the immunization schedule in both control and Env trimer-inoculated animals. In contrast, Env-specific PCs peaked 1 wk after each Env inoculation and then declined to near baseline levels. Env-specific memory B cells were detectable 2 wk after the first inoculation and achieved peak levels 1 wk after the second inoculation. Thereafter, the frequency of Env-specific memory B cells was maintained at stable levels at the time points measured. Though not shown in these animals, a decline in specific memory B cells at later time points after protein inoculation is expected based on data from another set of animals where we analyzed additional time points after immunization (Fig. S3). No Env-specific PC or memory B cells were detected in the adjuvant-only inoculated control animals.

We also examined the frequency of Env-specific PCs in the bone marrow after both the second and the fourth trimer inoculations. Approximately 3% of total IgG⁺ PCs were Env specific at both time points (Fig. 4 C). We did not detect any memory B cells in the bone marrow using either PWM + SAC + CpG stimulations or IL-2 + IL-10 + CpG + CD40L, which stimulates memory B cells in the absence of T cells (Fig. 4 D). These data suggest that memory B cells are rare in the bone marrow of rhesus macaques.

An important issue to address is if Env selectively boosts B cells that recognize specific regions of the trimers during the immunization regimen. To determine the subspecificities of circulating Abs and memory B cells, and to ask if these specificities shift during the course of the immunization schedule, we used the same series of trimeric Env probes as described in the previous section (gp140-F, gp120-F, and gp120ΔV1/2/3-F). We determined the frequency of memory B cells reactive to the three probes using subtractive analysis to calculate the percentage of gp41-reactive, V1/2/3-reactive, and non-gp41/non-V1/2/3 (termed “other”) fractions. By plotting the data as the percent region specific of total Env-specific memory B cells, the results reveal that a considerable fraction (∼30% of memory B cells) was directed against gp41 after two immunizations, but this fraction diminished during the course of the vaccination schedule to ∼10% of the total Env-specific response (Fig. 4 E). In contrast, the fraction directed against the gp120 variable regions expanded over time and especially after the third immunization. The results were different when the supernatants from PWM + SAC + CpG–stimulated PBMCs were analyzed in ELISA coated with the different Env probes. After converting the data to the percent region specific of total Env-specific reactivity, there was nearly no detectable gp41-specific response. Also, there was no difference in the relative reactivity against variable region versus other during the course of the immunizations. A similar pattern was observed when the plasma was analyzed by differential ELISA (Fig. 4 E). These results suggest that analyses based on ELISA, which measures secreted Ab in solution, are not sufficiently sensitive to measure subtle shifts in Ab subspecificities, likely as a result of competition between Abs of different specificities in this assay format.

To investigate the NAb activity elicited by the Env trimers, we examined the capacity of the plasma collected after two and five immunizations to neutralize a panel of HIV-1 isolates. The selected isolates are known to display varying degrees of neutralization sensitivity (Li et al., 2005). The panel included clade B isolates representing neutralization-sensitive viruses (tier 1) and isolates representing more neutralization-resistant primary circulating variants (tier 2). Recombinant viruses were pseudotyped with the following clade B Envs: MN, HxBc2, SF162, and BaL.01; and the following tier 2 clade B Envs: YU2, 89.6, 6535, and ADA. In addition, viruses were pseudotyped with one clade A (DJ263) and one clade C (MW965) Envs. The data are presented as 50% plasma inhibitory dilution (ID₅₀) values. In agreement with previous studies in rhesus macaques (Mörner et al., 2009; Douagi et al., 2010), we observed potent ID₅₀ neutralizing titers against several Tier 1 viruses and detectable, but less potent, neutralization against some Tier 2 viruses (Fig. 5 A). Mean ID₅₀ neutralization titers against MN were ∼1,000 after two immunization and these responses were significantly increased after five immunizations. High ID₅₀ titers against HxBc2 and SF162 were also measured after two immunization and these responses were boosted by the subsequent immunizations. Neutralization of viruses pseudotyped with the neutralization-resistant homologous YU2 Env was observed in three animals after two immunizations and in five animals after five immunizations. A similar pattern was observed for ADA. There was no or only sporadic neutralization of 89.6, 6535, TRO.11, JRCSF, or REJO (Fig. 5 A and not depicted). In contrast, neutralizing activity against Tier 1 viruses DJ263 and MW965 was obtained, suggesting some breadth of neutralization across HIV-1 clades. Furthermore, we also observed neutralization of viruses pseudotyped with the HIV-2 Env in the presence of subinhibitory concentrations of soluble CD4. This assay detects the presence of coreceptor binding site–directed Abs (Decker et al., 2005), which are elicited in NHPs inoculated with soluble HIV-1 Env immunogens (Forsell et al., 2008; Douagi et al., 2010).

To benchmark the soluble YU2 trimers used in the immunization regimen in this study against the recombinant gp120 protein regimen used in the VAX04 trial, we obtained 20 clinical serum samples from VAX04 collected 2 wk after four inoculations with MN and GNE8-derived gp120 protein in Alum. The samples were analyzed for neutralizing activity using the same panel of viruses and the same assay conditions as used for the NHP samples described in the previous paragraph. To strengthen the comparison, we included data from 10 animals described in our previous studies: five rhesus macaques inoculated with YU2 trimers in Abisco-100 + CpG (Douagi et al., 2010) and five cynomolgus macaques inoculated with YU2 trimers in GSK Adjuvant system AS01B (Mörner et al., 2009). The results demonstrate that the neutralizing activity elicited by the monomers (VAX04) and the trimers (NHP) was comparable against MN, whereas the trimers elicited superior neutralizing activity against all other viruses tested except ADA (Fig. 6). With the caveat that somewhat higher anti-Env binding Ab titers were achieved in the NHP study than in the VAX04 study (unpublished data), perhaps as a result of the use of different adjuvants, these data demonstrate that improved NAb responses can be generated by optimized subunit Env protein immunization regimens.

Although the major and primary goal of this study was to define the kinetics and magnitude of the B cell response elicited by the soluble HIV-1 Env trimers in healthy primates, the potent response observed in this study led us to ask if this could mediate protection against heterologous SHIV challenge. We used the SHIV-SF162P4 challenge virus, which induces high but transient viremia in rhesus macaques (Harouse et al., 1999; Barnett et al., 2008; Polacino et al., 2008). To model one route of human mucosal transmission via sexual exposure, we subjected the animals to repeated rectal exposures, two per week, using a moderate dose of virus (50 TCID₅₀). This dose corresponds to ∼3 × 10⁶ virus RNA copies, which is within the physiological range of one transmission event (Chakraborty et al., 2001; Pilcher et al., 2004). From previous in vivo titration results, we predicted that this dose would infect the Env-naive control animals in a few exposures, thus representing a higher dose than that estimated in humans under average transmission conditions but a lower dose than in some other SHIV studies (Barnett et al., 2008, 2010; Bogers et al., 2008). Blood was collected at each challenge occasion for detection of plasma viremia.

The Kaplan-Meier graph in Fig. 7 A (left) shows the percentage of uninfected animals after each challenge. After one challenge, one macaque in the control group became infected but no animal in the vaccine group. After the second challenge, four control animals were infected compared with two in the Env-immunized group. The remaining animals in the control group were infected after the third exposure, whereas six exposures were required to infect all animals in the Env-immunized group. This result suggests that there is a trend toward a modest degree of protection in the vaccine group; however, it is not statistically significant. Further studies involving larger groups of animals are needed to confirm this trend. We also asked if the high plasma Ab titers measured in the immunized animals (Fig. 3 B) provided a benefit in terms of control of viremia once the animals were infected. We measured this by cumulative viral loads during the viremic period (area under the curve calculated from viral load curves shown in Fig. S6; Fig. 7 A, middle) and by peak viral load (Fig. 7 A, right panel). By both measures, there was a trend for the immunized animals to better control the infection, but the difference between the groups was not statistically significant.

The modest, but suggestive, effect observed in the challenge experiment led us to analyze additional parameters of the prechallenge immune response. We first asked if the prechallenge plasma contained Abs capable of neutralizing either the SHIV-SF162P4 stock or recombinant virus pseudotyped with an Env clone derived from the SHIV-SF162P4 stock, clone 41.1. Moderate ID₅₀ neutralization titers ranging between 200 and 300 were measured against both the stock and the cloned virus (Fig. 7 B). Some background activity against the stock was observed in the plasma from the control immunized animals, but not against the pseudotyped cloned virus, suggesting that the latter assay is less prone to nonspecific serum effects.

To further address why more efficient protection against SHIV-SF162P4 challenge was not achieved despite detectable plasma neutralizing activity against the challenge virus in the immunized animals, we measured Env-specific and total IgG and IgA titers in mucosal lavages collected 2 wk after each immunization (Fig. 7 C, top). Env-specific IgG was detected in both rectal lavages and vaginal lavages and these responses correlated weakly (rectal lavages) or strongly (vaginal lavages) with the plasma Ab titers measured in the corresponding animal (Fig. 7 C, bottom). Env-specific IgA in vaginal lavages and rectal lavages were below detection (unpublished data). By comparing all data points from the vaginal and rectal samples collected between the second and the fifth immunization with the corresponding IgG concentrations measured in plasma, we conclude that the IgG concentrations in vaginal lavages are ∼1,000-fold lower than in plasma and the IgG concentrations in the rectal lavages are ∼10,000-fold lower than in plasma (Fig. 7 D). These lower levels of antibody just outside the mucosa, presumably as a result of inefficient exudation, may not be sufficient to neutralize all incoming virus before interaction with CD4⁺/CCR5⁺ target cells in this challenge model.

Vaccine-induced neutralizing antibody and memory B cell recall responses play critical roles in preventing infection and limiting virus dissemination. Although there is a pressing need to understand relevant B cell responses an effective HIV-1 vaccine should elicit, a detailed analysis of the magnitude and kinetics of Env-specific memory B cells and PCs generated in response to Env protein immunizations in humans or NHPs is currently lacking. Accordingly, we analyzed the generation of Env-specific B cell populations in rhesus macaques inoculated with recombinant Env trimers in adjuvant. We assessed the functionality of the response at several levels, including requirements for activation of resting macaque memory B cells, elicitation of Env-specific binding Abs and increases in avidity after prolonged antigen exposure, neutralization capacity of Env-elicited antibodies in vitro, and protection against in vivo SHIV challenge.

We demonstrate that Env-specific IgG-producing cells peaked 7 d after each boost and then declined sharply. This is similar to the kinetics reported for other vaccine antigens where boosting results in a transient burst of PCs (Blink et al., 2005; Wrammert et al., 2008; Blanchard-Rohner et al., 2009). The rapidity of the recall response suggests that it originates from circulating memory B cells that differentiate into PCs in the periphery or that it arises from circulating memory B cells that enter already existing germinal center reactions, as shown in a recent study (Schwickert et al., 2009). We also demonstrate that Env-specific IgG⁺ memory B cells peaked 2 wk after the second immunization and remained at stable levels between immunizations, comprising 10–20% of total IgG⁺ memory B cells. This pool is likely maintained by the frequent boosting regimen and would be expected to wane with time in the absence of antigen. We did not investigate the sustainability of the Env-specific memory B cell and PC pool over time once inoculation of antigen was terminated, but this is an important issue to address in future studies. Previous studies showed that the Ab response to gp120 immunization was poorly sustained after clearance of antigen (Anderson et al., 1989; Graham et al., 1996), suggesting a lack of long-lived PCs. Evidence for short half-lives of Env-specific Abs was also presented in a recent study based on gp120-vaccinated and HIV-1–infected individuals (Bonsignori et al., 2009).

Although the frequency of Env-specific memory B cells in PBMCs was high at the time points measured in our study, we were unable to detect total IgG⁺ or Env-specific memory B cells in the bone marrow. This is consistent with a recent study reporting a lack of detectable memory B cells in the human bone marrow. Instead, the spleen was shown to be the major reservoir for human memory B cells (Mamani-Matsuda et al., 2008). In this regard, we note that many basic aspects of rhesus macaque B cell biology remain uncharacterized despite their frequent use as a nonhuman model. We suggest that the systems of analysis described in the current study, including the protocols for in vitro expansion and differentiation of memory B cells, have broad applications for studies aimed at understanding the distribution and frequency of functionally distinct B cell subsets in NHPs and to assess them relative to human biology and vaccine-elicited immune responses.

High titers of circulating Abs capable of neutralizing a panel of sensitive viruses spanning clades A, B, and C were shown here, but improvements in Env immunogen design and/or mode of administration are needed to achieve neutralization of primary viruses represented in the Tier 2 panel used here. The relatively short immunization schedule used here and in many other preclinical immunogenicity studies may not allow optimal maturation of the B cell response. With the intervals of antigen exposure used in this study, we observed no improvement in the breadth of neutralization with boosting despite an increase in neutralizing activity against sensitive Tier 1 viruses. Early HIV-1 Env studies in baboons suggest that longer intervals between immunizations may provide some benefit (Anderson et al., 1989). Further analyses are needed to address this question in a comprehensive and systematic manner.

In light of the renewed interest in the responses elicited by the monomeric gp120 proteins used in the VAX04 and RV144 phase III clinical trials, we sought to directly compare the neutralizing Ab activity elicited by the monomeric gp120 used in VAX04 with that elicited by the trimers in NHPs. Previous evaluation of the antigenic properties of the YU2 trimers show that they possess several of the features desired of an immunogen aimed to elicit bNAbs, such as occlusion of nonneutralizing epitopes in the C- and N-terminal domains of gp120 and recognition by a set of broadly neutralizing mAbs (Yang et al., 2000). Consistent with this, the YU2 trimers are superior to monomeric YU2 gp120 for elicitation of neutralizing Abs in small animals (Li et al., 2006). However, a direct comparison between the VAX04 monomeric gp120 immunogen and the YU2 trimeric immunogens was not previously performed. In our in vitro neutralization comparison, we observed a broader response in the trimer-inoculated macaques than in the monomer-inoculated humans. Although there are several caveats to this comparison, this analysis provides initial benchmark information about the gp140 trimer-elicited response in NHPs relative to the response elicited by monomeric gp120 in humans.

The recent results from RV144 indicate that boosting with purified Env protein in adjuvant after priming with a recombinant canary pox vector provided a modest but transient protective effect in low risk populations (Rerks-Ngarm et al., 2009). Several preclinical studies also demonstrate that boosting with recombinant Env protein elicits high Ab responses after priming with recombinant viral vectors or DNA (Montefiori et al., 1992; Lubeck et al., 1997; Forsell et al., 2005; Shu et al., 2007; Mörner et al., 2009). Both VAX04 and RV144 used gp120 administered in Alum, an adjuvant known to stimulate Th2-biased immune responses (Marrack et al., 2009). However, because RV144 used a viral vector to prime the response, this might lead to a different spectrum of Ab isotypes than those elicited in VAX04, which used protein in Alum adjuvant alone. A difference in Ab isotypes could affect Ab-mediated activities such as ADCC. Consistent with shifts in Ab isotypes, we demonstrated previously that priming with an Env-expressing recombinant Semliki Forest virus vector resulted in an Ab response that remained Th1 biased even after boosting with Env protein in a Th2-biasing adjuvant (Forsell et al., 2005). In the current study, we used an adjuvant consisting of Abisco-100 and CpG, which stimulates Th1-biased responses including generation of IFN-γ–producing CD4⁺ and CD8⁺ T cells (Douagi et al., 2010). A better understanding of the role for Ab isotypes relative to antibody neutralization and protection against infection is an important issue yet to be clarified (Liu et al., 2003; Huber and Trkola, 2007; Nimmerjahn and Ravetch, 2008).

Previous studies established that protection against SHIV challenge via the mucosal route was achievable by either active immunization with a homologous Env-based immunogen (Barnett et al., 2008, 2010; Bogers et al., 2008) or by passive infusion of Env-directed neutralizing Abs (Mascola et al., 2000; Hessell et al., 2009a,b). However, heterologous challenge to Env vaccine-elicited B cell responses is rarely attempted even though it is likely a better model to mimic real-world human-to-human transmission events. Although the circulating Abs elicited by the regimen described in this paper neutralized SHIV-SF162P4 in vitro with reasonable potency, only a modest vaccine effect was observed upon in vivo rectal challenge. Upon closer examination, this outcome is perhaps not surprising. In the Barnett studies, an effective impact on acquisition was observed with homologous vaginal (Barnett et al., 2008) or rectal (Barnett et al., 2010) SHIV-SF162P4 challenge. The plasma neutralizing titers observed in the Barnett studies were generally higher against the homologous SHIV-SF162P4 than those elicited by the heterologous YU2 trimers here. To mediate protection, some circulating IgG induced by systemic immunizations would be expected to enter the mucosa by transcytosis (Rojas and Apodaca, 2002), which is consistent with the detection of mAbs in mucosal washes after intravenous infusion of mAbs (Mascola et al., 2000; Parren et al., 2001; Hessell et al., 2009b). In our measurements, the Ab titers in the rectal mucosa were approximately four orders of magnitude lower than those in plasma, likely bringing the neutralization titers below effectiveness with regard to viral challenge via this route of exposure. Further improvements in the specificity and affinity of vaccine-elicited neutralizing Abs may be needed for this process to be efficient enough to achieve protection against heterologous rectal challenge. For an antibody-dependent HIV-1 vaccine, the rectal route of challenge may be more difficult to protect against because of lower levels of exudates and more readily available CD4⁺/CCR5⁺ gut-associated target cells. Regimens designed to stimulate mucosal immune responses, including the production of antigen-specific secretory IgA, may provide benefit (Kozlowski and Neutra, 2003; Sundling et al., 2008), even though mucosal immunization was not required for protection against homologous infection via the vaginal route in a previous study (Barnett et al., 2008).

In conclusion, the extreme genetic diversity of circulating HIV-1 variants worldwide remains the main challenge for the development of an effective HIV-1 vaccine. Because of the difficulties in eliciting bNAbs through vaccination, the field has been spurred to dissect B cell responses in chronically infected individuals who display such activities. In parallel with these efforts, we contend that it is important to characterize B cell responses elicited by Env vaccination in healthy primates where potential pathogenic effects caused by chronic HIV-1 infection are not operational. In this regard, macaques offer several advantages as a result of their close genetic relationship with humans, their susceptibility to infection with SIV/SHIV, and because a more limited number of vaccine candidates can be tested in human clinical trials compared with in NHPs. The data presented here provide substantial insight into many issues relative to Env-specific primate B cell responses. Additional well designed HIV-1 vaccine and SHIV challenge studies, in coordination with selected human immunogenicity studies, are likely required to fully dissect the protective requirements at relevant portals of viral entry.

Soluble gp140 trimers (gp140-F) were produced by transient transfection of Freestyle 293F suspension cells (Invitrogen) as previously described (Forsell et al., 2008). In brief, cells were transfected at a density of 1 × 10⁶/ml in GIBCO Freestyle293 expression media using 293Fectin, according to manufacturer’s instructions (Invitrogen). Supernatants were collected 4 d after transfection. After collection, all supernatants were centrifuged at 3,500 g to remove cells or cell debris, filtered through a 0.22-µm filter, supplemented with Complete EDTA-free protease inhibitor cocktail (Roche) and penicillin-streptomycin (Invitrogen), and stored at 4°C until further purification. First, the proteins were captured via glycans by lentil-lectin affinity chromatography (GE Healthcare). After extensive washing with PBS, supplemented with 0.5 M NaCl, the proteins were eluted with 1 M methyl-α-d-mannopyranoside and captured in the second step via the His-tag by nickel-chelation chromatography (GE Healthcare). After a wash with 40 mM imidazole (IM) and 0.5M NaCl in PBS, proteins were eluted with 300mM IM in PBS. Trimers were then separated from lower molecular weight forms by gel filtration chromatography using a superdex200 26/60 prep grade column by the ÄKTA Fast protein liquid chromatography system (GE Healthcare). Biotinylated Env probes used in the B cell ELISPOT assay were purified by lentil-lectin and nickel-chelation chromatography but not subjected to subsequent gel filtration chromatography. The biotinylation was performed using site-specific biotinylation of an Avitag sequence inserted at the C terminus of each Env probe, allowing for covalent linkage of biotin to the lysine residue in the motif distal to the Env antigenic surface.

12 rhesus macaques (Macaca Mulatta) of Chinese origin, ∼3 yr old, and four cynomolgus macaques of Chinese origin, ∼5 yr old, were housed at the animal facility Astrid Fagraeus Laboratory at the Swedish Institute for Infectious Disease Control. Housing and care procedures were in compliance with the provisions and general guidelines of the Swedish Board of Agriculture, and the facility has been assigned an Animal Welfare Assurance number by the Office of Laboratory Animal Welfare (OLAW) at National Institutes of Health. All procedures were approved by the Local Ethical Committee on Animal Experiments (Stockholms Norra Djurförsöksetiska Nämnd). The animals were housed in pairs in 4-m³ cages, enriched to give them possibility to express their physiological and behavioral needs. They were habituated to the housing conditions for >6 wk before the start of the experiment and subjected to positive reinforcement training to reduce the stress associated with experimental procedures. All immunizations and blood samplings were performed under sedation with 10 mg/kg ketamine i.m. (100 mg/ml Ketaminol; Intervet), and when sampled for rectal or vaginal lavages or bone marrow the macaques were given an additional 0.5 mg/kg i.m. Xylazine (Rompun; Bayer) to induce total muscle relaxation and analgesia. The macaques were weighed at each immunization or sampling occasion. Before entering the study, all animals were confirmed negative for SIV, simian T cell lymphotropic virus, and simian retrovirus type D.

The 12 rhesus macaques (two groups of six animals) were inoculated five times with the immunogens described in the Expression and purification of Env immunogens section. Immunizations were performed at weeks 0, 4, 8, 12, and 18 by the i.m. route of injection. All protein immunizations were administered in 75 µg ABISCO-100 (Isconova AB) and 500 µg CpG ODN2395 (Coley Pharmaceutical Group) as adjuvant (Abisco-100 + CpG). Protein doses were 200 µg per animal for the first inoculation and 100 µg for the subsequent injections. The vaccine was given in a total volume of 1 ml, divided equally between the left and right hind leg. Blood and rectal lavage samples were taken before and 1 and 2 wk after each immunization. The rectal mucosa was sampled through insertion of a syringe and rinsing with 4 ml PBS. The lavage was then added to a concentrated cocktail of protease inhibitors, spun, and the liquid phase frozen at −80°C. Vaginal lavage was taken 2 wk after each immunization through insertion of a pediatric feeding tube and rinsing with 2 ml PBS. The vaginal lavage was then treated as for rectal samples. 3 wk after immunizations two and four, the macaques were sampled for bone marrow through strictly aseptic insertion of a pediatric spinal needle between the greater and lesser tuberosity of humerus. A maximum of 5 ml was aspirated from each animal into a heparinized syringe and transferred to an EDTA tube.

Bleeds were taken in EDTA-coated tubes and spun at 1,200 g for 15 min for separation of plasma. The mononuclear cells were purified through density gradient centrifugation with Ficoll-Paque (GE Healthcare) according to the manufacturer’s instructions and then counted and frozen in 90% heat-inactivated FBS and 10% DMSO (Sigma-Aldrich). Experiments were performed on both frozen and fresh cells. PBMCs were thawed in a 37°C water bath followed by two washes in RPMI 1640 media supplemented with 10% FBS, 2 mM L-Glutamine, 100 U/ml penicillin, 100 µM streptomycin, 2% Hepes (all from Sigma-Aldrich), and 50 mM 2-ME (Invitrogen) before resting at 37°C supplemented with 5% CO₂ overnight (ON). The PBMCs were cultured at 2 × 10⁶ cells/ml in a total of 500 µl per well in 48-well flat bottom plates (Thermo Fisher Scientific). The stimulation mix was composed of different combinations of 5 µg/ml CpG ODN 10103 (TCGTCGTTTCGTCGTTTTGTCGTT; Coley Pharmaceuticals), PWM (Sigma-Aldrich), SAC strain 1/10,000 (Sigma-Aldrich), 100 ng/ml interleukin 2, 100 ng/ml interleukin 10, and 1 µg/ml of soluble CD40 ligand (Peprotech).

To evaluate the stimulatory effect on B cell and T cell proliferation through FACS analysis, PBMCs were first labeled with 0.5 µM CFSE for 7 min at 37°C and then washed with complete media. The cells were then incubated with the indicated stimulation for 6 d, after which the cells were stained with Abs for CD20 (clone 2H7), CD4 (clone L200), CD8 (clone SK1), and CD27 (clone M-T271). All Abs (BD) were previously titrated for optimal staining using rhesus macaque PBMCs. Samples were collected on a FACSCalibur (BD) and analyzed using FlowJo software (Tree Star, Inc.). For experiments using sorted macaque cells, PBMCs were incubated with CD20- or CD4-conjugated magnetic beads (Miltenyi Biotec) and positive fractions were isolated using autoMACS (Miltenyi Biotec) separation according to the manufacturer’s instructions. The purity of sorted cells was evaluated through FACS analysis and was >90% for the CD20⁺ fraction and between 50 and 90% for the CD4⁺ fraction. Contaminating cells in the CD4⁺ fraction were mainly CD20⁺ B cells. Human B cells were sorted by either CD20 or CD19 positive selection with magnetic beads as described according to the manufacturer’s instructions and were >90% CD20⁺ B cells. Human CD4⁺ T cells were sorted from buffy coats by negative selection using CD4⁺ T cell enrichment cocktail (STEMCELL Technologies Inc.) and were >95% CD3⁺ and CD4⁺, as determined by flow cytometry.

HIV-1 gp120-specific plasma, rectal or vaginal IgG, and IgA was measured by ELISA as previously described (Mörner et al., 2009) with modifications. In brief, insect cell–produced YU2 gp120 protein was coated onto MaxiSorp (Thermo Fisher Scientific) microtiter plates at 100 ng/well in 50 mM carbonate buffer, pH 9.6, ON at 4°C. After blocking in PBS containing 2% nonfat dry milk, samples were added and incubated for 1.5 h at 37°C. The gp120-specific IgG or IgA was detected by adding secondary HRP-conjugated anti–monkey IgG (Nordic Immunology) or anti–monkey IgA (KPL). The signal was developed by addition of O-phenylenediamine (OPD; Sigma-Aldrich) or TMB⁺ (Invitrogen). Reactions were terminated by adding 1 M HCl and the OD was read at 492 (for OPD) or 450 (for TMB⁺) and 620 nm. Between each incubation step, the plates were washed six times with PBS 0.05% Tween 20. The half-max binding titers (OD50) for each sample was calculated by interpolation from mean OD50 values calculated from an immunized control plasma using the formula ([(ODmax-ODmin)/2] + ODmin). As a result of dilution of the vaginal and rectal lavages upon sample collection, data were set in relation to the total amount of IgG or IgA in the sample. Total IgG and IgA was measured by coating plates with 200 ng/well of goat anti–human IgG (Jackson ImmunoResearch Laboratories) or 100 ng/well of goat anti–monkey IgA (KPL).

When ELISA was used to measure Env region-specific Abs, the plates were coated with the same molar ratios of gp140-F, gp120-F, or gp120ΔV1/2/3-F as in the previous paragraph. Percentages of gp41 and variable region 1/2/3–specific Abs were calculated using the formulas (gp140-gp120)/gp140 × 100 and (gp120-gp120ΔV1/2/3)/gp140 × 100. To calculate absolute concentrations of gp120-specific Abs in samples, a gp140-F–immunized rhesus control sera was quantified for IgG gp120-specific Abs via a rhesus IgG standard (SouthernBiotech) and subsequently used as standard for unknowns.

Avidity index was evaluated as described previously (Vermont et al., 2002; Lai et al., 2007) with some modifications. ELISA plates were coated overnight at 4°C with 100 ng/well Galanthus nivalis lectin (Sigma-Aldrich) in PBS. S2-produced gp120 or 293F-produced gp140-F, gp120-F, or gp120Δ1/2/3-F was added to the wells and incubated for 2 h at 37°C. Plasma was added in dilution series and incubated for 1.5 h before exposure to 1.5 M NaSCN or PBS for 10 min. The bound Abs were detected as in the ELISA described in the previous paragraph. Titers for each sample were calculated from the OD50 value of an in-house standard plasma sample. The avidity index was calculated by dividing the titer for NaSCN-treated and PBS-treated samples and multiplying by 100. To determine if treatment with 1.5 M NaSCN disrupted the integrity of Env, we performed a dose titration experiment with increasing concentrations of NaSCN and assessed binding by a panel of mAbs directed against distinct gp120 epitopes (Fig. S4). The results demonstrate that recognition by several conformation-sensitive Abs (17b, b6, and F91) but not all (E51, 2.1C, b12, and F105) was retained after the 1.5-M NaSCN treatment, suggesting that Abs recognizing both linear and conformational epitopes are detected under the assay conditions used in the present study.

The frequency of antigen-specific memory B cells was determined as previously described (Douagi et al., 2010). In brief, total PBMCs were cultured at 2 × 10⁶ cells/ml and stimulated with 5 µg/ml CpG-B (ODN-10103; Coley), 10 µg/ml PWM (Sigma-Aldrich), and SAC (1:10,000; Sigma-Aldrich) for 4 d in 48-well plates. To detect Ab-secreting cells, MAIPSWU10 96-well plates (Millipore) were coated with 10 µg/ml of anti–human IgG (Fcγ; Jackson ImmunoResearch Laboratories) and PBMCs were transferred to the plates in dilution series and incubated ON at 37°C in 5% CO₂. The plates were then washed with PBS containing 0.05% Tween and incubated with biotinylated gp140-F, gp120-F, or gp120ΔV1/2/3-F (Dosenovic et al., 2009) followed by washing and incubation with streptavidin-AP (Mabtech). The reactions were developed using BCIP/NBT substrate (Sigma-Aldrich) and stopped by washing in water. Spots corresponding to ASC were counted using an Immunospot analyzer (Cellular Technology Ltd.). The results were converted to Ab-secreting cells per million cultured PBMCs (ASCs/10⁶ PBMCs cells). Percentages of gp41 and variable region 1/2/3-specific memory B cells were calculated using the formulas (gp140-gp120)/gp140 × 100 and (gp120-gp120ΔV1/2/3)/gp140 × 100, respectively.

Plasmas from immunized animals were tested for virus neutralization capacity against a panel of diverse HIV-1 isolates. Neutralization assays were performed using a single round of infection HIV-1 Env pseudovirus assay and TZM-bl target cells as previously described (Li et al., 2005; Shu et al., 2007). Env pseudoviruses were prepared by cotransfecting 293T cells with an Env expression plasmid containing a full gp160 env gene and an env-deficient HIV-1 backbone vector (pSG3ΔEnv). To determine the serum dilution that resulted in a 50% reduction in RLU, serial dilutions of sera were performed and the neutralization dose–response curves were fit by nonlinear regression using a four-parameter hill slope equation programmed into JMP statistical software (JMP 5.1; SAS Institute Inc.). The results are reported as the serum neutralization ID₅₀, which is the reciprocal of the serum dilution producing 50% virus neutralization. Diverse HIV-1 virus isolates, including viruses from clades A, B, and C, were used in the neutralization assays. Clade B viruses included a panel of Env pseudoviruses that were recently characterized and recommended for use in assessing neutralization by HIV-1 immune sera (Li et al., 2005). Several investigators also provided replication-competent viruses or functional Env plasmids for pseudoviruses. D. Gabuzda (Dana Farber Cancer Institute, Boston, MA) provided the Env plasmids for YU2 and MuLV. Env plasmids for SF162 and JRFL were provided by L. Stamatatos (Seattle Biomedical Research Institute, Seattle, WA) and J. Binley (Torrey Pines Institute, La Jolla, CA), respectively. The clade A DJ263.8 sequence was cloned from the original PBMC-derived virus (provided by F. McCutchan and V. Polonis, U.S. Military HIV Research Program, Rockville, MD) and the clade C MW965 Env plasmid was obtained from the AIDS Research and Reagent Repository. Isolation of the Env plasmids BaL.01 were recently described by our laboratory (Shu et al., 2007) and the SS1196.1 Env was previously described (Li et al., 2005). Neutralization was also measured against the SHI-SF162P4 stock and against a cloned Env from this stock, Env clone 41.1.

20 randomly chosen serum samples were obtained via a Materials Transfer Agreement with the Global Solutions for Infectious Diseases. These sera were derived from volunteers from the VaxGen Inc phase III clinical trial. At the time of sampling (month 12.5), the participants had received four injections (months 0, 1, 6, and 12) of the AIDSVAX B/B vaccine containing 300 µg each of recombinant HIV-1MN– and HIV-1GNE8–derived gp120 in Alum adjuvant (Pitisuttithum et al., 2006).

Plasma from an acutely SHIV-SF162P4 (Kraft et al., 2007)–infected rhesus macaque was originally provided by L. Stamatatos and was first inoculated i.v. into a naive rhesus macaque. During peak viremia, blood and bone marrow from this animal was injected into a second macaque (32M). Virus was reisolated from 32M at peak viremia by co-culture with naive rhesus PBMC, and a virus stock was prepared. This virus was used to prepare a new virus stock by i.v. inoculation into a naive rhesus macaque of Indian origin and subsequent reisolation as described in Virus neutralization assays. The resulting virus was further expanded in vitro on C8166-CCR5 cells to produce the challenge stock used in the present study. The stock was titrated on rhesus PBMC cells in vitro and in vivo in six monkeys via the rectal route. 5 wk after the last immunization, 3 ml, corresponding to 50 TCID₅₀ (tissue culture infectious dose, 50), was inoculated rectally in 12 macaques. The macaques were then challenged with this dose twice a week until confirmed positive for viral load as was monitored by quantitative (Q) RT-PCR with a cutoff of 214 copies per ml. Anamnestic Ab responses were evaluated up to 7 wk after confirmed infection by ELISA.

A DNA template was created by inserting a 458-bp segment of p27 from SHIV-SF162P4, using the forward primer SIVMac239Gag5 (5′-GTGCATTCACGCAGAAGAGA-3′) and reverse primer SIVMac239Gag.r.out (5′-TTCCTGCAATATCTGATCCTGA-3′), into the pJET1.2-blunt vector (Fermentas; designated pJET1.2-gag458). The insert was verified through sequencing (GATC Biotech AG) and alignment with SIVMac239. RNA standard was produced by in vitro transcription of NcoI-linearized pJET1.2-gag458 with T7 polymerase followed by DNase I digestion (Fermentas). The RNA was then purified using the RNeasy mini kit (QIAGEN) and frozen in aliquots, corresponding to 10-fold dilution series ranging from 10⁶ to 10¹ RNA copies, supplemented with 30 µg/ml S. cerevisiae transfer RNA (Sigma-Aldrich).

Viral RNA was extracted from 280 µl EDTA-plasma using the QIAamp Viral RNA mini kit (QIAGEN) and eluted in 60 µl RNase-free water containing 0.04% sodium azide according to the manufacturer’s instructions. The eluted RNA was assayed immediately after recovery using a One-Step Q-RT-PCR protocol. The primers and probe chosen for amplification of the SIVMac239 p27 were published previously (Leutenegger et al., 2001) and worked with high efficiency for detection of SHIV-SF162P4 (Fig. S4). In brief, the forward primer SIV.510f (5′-GCCAGGATTTCAGGCACTGT-3′) and reverse primer SIV.591r (5′-GCTTGATGGTCTCCCACACA-3′) amplified an 81-bp segment, where the probe SIV.535p (5′-AAGGTTGCAC­CCCCTATGACA­TTAATCAGATG­TTA-3′) bound. The one-step Q-RT-PCR was run using the qScript One-Step Fast qRT-PCR Low ROX kit (Quanta BioSciences Inc.) according to the manufacturer’s instructions under the following conditions: 20 µl total reaction containing 80-nM probes, 400 nM of each primer, and 10 µl RNA. Cycling conditions were 5 min at 50°C for initial cDNA synthesis and 95°C at 30 s, followed by 40 cycles of 95°C for 3 s and 60°C for 30 s The reactions were performed on an ABI PRISM 7500 Fast cycler (Applied Biosystems). A standard curve and variability coefficients are shown in Fig. S4.

In all group comparisons including three or more than three groups, statistical significances were determined by one-way ANOVA with Bonferroni’s posttest for individual comparisons on log-transformed data using Prism software (version 5; GraphPad Software, Inc.) and considered significant at P ≤ 0.05, P ≤ 0.01, and P ≤ 0.001. When comparing three or less than three groups the Mann-Whitney test was used.

Fig. S1 shows a schematic representation of the experimental procedure for quantifying PCs and memory B cells. Fig. S2 demonstrates the kinetics of memory B cell differentiation into ASC under different in vitro stimulations conditions. Fig. S3 describes the frequency of PC and memory B cell populations at multiple time points after Env trimer inoculation. Fig. S4 shows a standard curve and the variability of the Q-RT-PCR assay used to evaluate virus load after in vivo SHIV challenge. Fig. S5 shows how the binding of anti-Env mAbs is affected by SCN treatment. Fig. S6 shows the plasma viral loads in individual animals challenged with SHIV-SF162P4.

We thank Mats Spångberg, Helene Fredlund, and the personnel at Astrid Fagraeus laboratory at the Swedish Institute for Infectious Disease Control for expert assistance with rhesus macaques. We are grateful to Pia Dosenovic for helpful discussions and for critically reading the manuscript, Martina Soldemo, Andreas Mörner, William Adams, and Cornelia Gujer for technical contributions, Linda Stertman and Karin Lövgren Bengtsson at Isconova AB for sharing Abisco-100, and Dr. Frank Sinangil and the Global Solutions for Infectious Diseases scientists for providing the VaxGen Inc phase III clinical serum samples.

This study was supported by grants from the Swedish Research Council, Sida/Swedish Agency for Research Cooperation, the International AIDS Vaccine Initiative, the Bill and Melinda Gates Foundation, the intramural program of the Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health intramural research program, and a Karolinska Institutet Doctoral grant from Karolinska Institutet.

The authors have no conflicting financial interests.