The HIV-1 envelope glycoprotein (Env) functional spike has evolved multiple immune evasion strategies, and only a few broadly neutralizing determinants on the assembled spike are accessible to Abs. Serological studies, based upon Ab binding and neutralization activity in vitro, suggest that vaccination with current Env-based immunogens predominantly elicits Abs that bind nonneutralizing or strain-restricted neutralizing epitopes. However, the fractional specificities of the polyclonal mixture of Abs present in serum, especially those directed to conformational Env epitopes, are often difficult to determine. Furthermore, serological analyses do not provide information regarding how repeated Ag inoculation impacts the expansion and maintenance of Env-specific B cell subpopulations. Therefore, we developed a highly sensitive Env-specific B cell ELISPOT system, which allows the enumeration of Ab-secreting cells (ASC) from diverse anatomical compartments directed against different structural determinants of Env. In this study, we use this system to examine the evolution of B cell responses in mice immunized with engineered Env trimers in adjuvant. We demonstrate that the relative proportion of ASC specific for defined structural elements of Env is altered significantly by homologous booster immunizations. This results in the selective expansion of ASC directed against the variable regions of Env. We suggest that the B cell specificity and compartment analysis described in this study are important complements to serological mapping studies for the examination of B cell responses against subspecificities of a variety of immunogens.
A challenge to the development of prophylactic vaccines against highly variable pathogens, such as HIV-1, is to achieve broad rather than strain-restricted protective responses. Pathogens that establish an active, chronic infection are particularly difficult vaccine targets because they evolve efficient mechanisms to shield conserved antigenic neutralization determinants from B cell recognition to persist long-term in the host (1). During each infection cycle in the new host, Abs are abundantly elicited against both surface-exposed and internal viral elements. Those elicited against internal elements are not neutralizing, whereas most Abs elicited against the surface envelope glycoprotein (Env)3 are either nonneutralizing or neutralize only a limited subset of circulating HIV-1 variants. Broadly neutralizing Abs appear only in some individuals after several years of chronic infection (2, 3, 4, 5, 6, 7). Thus, the virus appears to persist by selection of resistant variants that outgrow and replace the originally dominant virus. Immune pressure at both the T and B cell level and selection of escape variants most likely explain most of the variability observed in the pool of HIV-1 circulating in the world today (8, 9, 10, 11, 12).
The HIV-1 Env consists of the exterior subunit, gp120, and the transmembrane glycoprotein, gp41, which are noncovalently associated to form a trimer of gp120/gp41 heterodimers. This complex comprises the functional viral spike, which mediates entry into CD4 and CCR5 receptor-positive target cells. Early analysis of the sequence of gp120 from different HIV-1 isolates revealed five V regions (V1–V5) and five conserved regions (C1–C5) (13). Strain-specific neutralization, the earliest autologous neutralizing response detected after seroconversion, is usually mediated by Abs that recognize the V regions of gp120 (14, 15, 16, 17, 18, 19). The V3 region of gp120 was referred to as the principal neutralizing determinant of HIV-1 (20, 21), and V3-directed Abs are largely responsible for neutralization of homologous laboratory-adapted HIV-1 isolates (3, 22). However, the V3 region is poorly exposed on most HIV-1 primary isolates, rendering it a less desirable vaccine target (23). The V regions of Env most likely evolved as immunodominant decoys for the host Ab response to subvert responses away from subdominant conserved epitopes (24, 25). A similar hierarchy of immunodominance between B cell epitopes may exist after vaccination with different Env immunogens, rendering the elicitation of broadly protective Ab responses difficult.
Due to the failure of monomeric gp120 to elicit protective responses in humans (26) and in preclinical animal models, efforts to design immunogens that more closely mimic the viral spike were undertaken. Different versions of soluble trimeric molecules containing full-length gp120 covalently linked to the gp41 ectodomain were designed (27, 28, 29, 30, 31, 32). In one approach, cleavage-defective Env trimers derived from the primary R5 isolate YU2 possessing a heterologous trimerization motif, derived from T4 bacteriophage (fibritin; gp140-F), were developed. Small animal studies (33, 34, 35, 36) using well standardized binding and neutralization assays (37, 38) demonstrated that the engineered soluble trimeric Env spike mimetics elicited superior neutralizing Abs compared with monomeric gp120, but that the neutralization breadth and potency were still relatively limited (34, 35).
To forward our understanding of why these immunogens are ineffective at eliciting broadly neutralizing Abs, analyses of the specificities of the elicited humoral immune responses are required. Such responses remain poorly characterized at the B cell level, and our understanding of how different immunization regimens affect the development of Env-specific memory B cell and plasma cell populations is incomplete. A successful prophylactic vaccine against HIV-1 may require the establishment of both circulating Abs and memory B cells capable of rapid expansion and differentiation into Ab-secreting cells (ASC) upon restimulation with Ag. Hence, it is important to investigate how selected B cell compartments evolve after immunization with candidate Env immunogens and to develop systems for the analysis and characterization of these compartments.
In this study, we systematically analyze Env-specific ASC following repeated immunization with soluble HIV-1 Env trimers using a differential B cell ELISPOT approach. We demonstrate that the relative proportion of ASC specific for different structural determinants of Env is altered significantly by repeated gp140-F immunizations, leading to a selective and undesired expansion of ASC directed against the gp120 V regions. These results have direct relevance for the design of improved immunization regimens aimed at eliciting more broadly reactive Env-specific B cell responses. The approach described in this study provides an important complement to the analysis of circulating serum Abs and represents a powerful system to study basic questions related to the evolution of B cell responses after immunization with complex protein Ags.
Materials and Methods
Constructs expressing HIV-1 Env glycoproteins
Soluble stable YU2-based gp140-F was used for the immunizations. YU2 gp140-F was formerly referred to as YU2gp140−/FT (32). To generate probes for the ELISPOT assay, the gp140-F and gp120-F sequences were modified by addition of an Avitag sequence for biotinylation (GLNDIFEAQKIEWHE) (39, 40) downstream of the fibritin trimerization motif (F) and the histidine tag (His)6, resulting in constructs encoding gp140-F-Avitag and gp120-F-Avitag, respectively. The gp120-F was derived from gp140-F by deletion of the gp41 sequence spanning residues 512–683 (41). A construct encoding a V3-deleted trimeric gp120 probe (gp120-F-ΔV3-Avitag) was derived by deleting residues 302–324 from gp120-F-Avitag. This construct was further modified by deleting aa 126–197 to encode a V1/V2/V3-deleted trimeric gp120 probe (gp120-F-ΔV1/2/3-Avitag) (42). Finally, a construct encoding a V1/V2-deleted trimeric gp120 probe (gp120-F-ΔV 1/2-Avitag) was generated by deleting aa 126–197 from gp120-F-Avitag, and a construct encoding a monomeric gp120 probe with a C-terminal Avitag (gp120-Avitag) was made by removing the fibritin trimerization domain from gp120-F-Avitag. The construct expressing soluble gp140 trimers stabilized with the GCN4 trimerization domain used to produce protein for generation of mouse hybridomas was previously described (31).
Production and purification of Env glycoproteins
All Env glycoproteins were produced and purified, as previously described (43
Enzymatic biotinylation of probes
Biotin ligase, Bir A (Avidity), was used for site-specific biotinylation of the Avitag sequence of all Env probes, per manufacturer’s instructions. This allows for covalent linkage of biotin to the lysine residue in the motif (one per monomer), distal to the Env antigenic surface. Biotinylation was confirmed by capturing Env on an ELISA plate (Maxisorp high binding plate; Nunc) using the mAb b12 and subsequent detection using HRP-conjugated streptavidin (Sigma-Aldrich).
Biochemical characterization of YU2 HIV-1 Env protein probes
Generation of hybridoma cell lines
To produce mouse hybridoma cell lines, BALB/c mice were immunized with YU2 gp140-GCN4 trimers (31). Inoculations were performed by intradermal injections of 50 μg of gp140-GCN4 in IFA three times with 1-wk intervals, with a final boost 3 days before the hybridoma fusion was performed. Splenocytes were harvested and fused with SP2 myeloma cells, as described (46). Hybridomas were screened by ELISA for gp120 binding and cloned by limiting dilution. Env-specific mouse hybridoma cells 5D4-F7 and 2B5-H8 were grown in RPMI 1640 containing 10% FCS, l-glutamine, and antibiotics.
Immunizations and preparation of cells for B cell ELISPOT measurements
Adult male BALB/c mice, obtained from Taconic Farms, were immunized s.c. two or three times, 2 wk apart, with 10 μg of HIV-1 YU2 gp140-F, or control β-galactosidase (β-gal) protein (Sigma-Aldrich) with 10 μg of AbISCO-100 adjuvant (Isconova). Single-cell suspensions were prepared from spleen and bone marrow (BM) by passing the tissue through a nylon mesh. RBC were lysed with hypotonic ammonium chloride solution. After washing, the cells were resuspended in complete RPMI 1640 medium containing 5% FCS, 50 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μM streptomycin at a final concentration of 1 × 107 cells/ml, and then diluted in 3-fold dilution series to 1.2 × 105 cells/ml. Cells were then added to the ELISPOT assay immediately or stimulated in vitro at a concentration of 1 × 106 cells/ml in complete RPMI 1640 medium containing 2 μg/ml LPS for 7 days. All animal experiments were approved by the Committee for Animal Ethics (Stockholm, Sweden), and performed according to given guidelines.
B cell ELISPOT assays
Ninety-six-well MultiScreen-IP filter plates (Millipore) were pretreated with 70% ethanol and washed three times in sterile PBS, before being coated with either 0.5 μg/well (5 μg/ml) Env protein (insect cell S2-produced gp120 monomers or gp140-F) for the conventional ELISPOT, or 1 μg/well (10 μg/ml) polyclonal goat anti-mouse IgG Ab (Mabtech) for the optimized and the total IgG ELISPOT. The plates were incubated overnight (ON) at 4°C. Before addition of the cells, the plates were washed five times in sterile PBS and then blocked with complete RPMI 1640 medium at 37°C for 2 h. Cells were added in duplicates or triplicates to the wells in 3-fold serial dilutions, starting at 5000 cells/well for hybridoma cells and 1 × 106 cells/well for splenocytes and BM cells. Plates were wrapped in plastic and incubated for 12 h at 37°C. For detection of spots, the cells were removed by washing the plates six times in PBS containing 0.05% Tween 20. For the conventional ELISPOT and for detection of total IgG-secreting cells, 0.1 μg/well (1 μg/ml) biotinylated polyclonal goat anti-mouse IgG (Mabtech) was added in blocking buffer (PBS containing 1% FCS and 0.05% Tween 20). For the optimized ELISPOT, 200 ng/well (2 μg/ml) biotinylated protein (variants of Env or β-gal) was added diluted in blocking buffer. Biotinylated Ab and protein were incubated in the plates for 2 h at room temperature (RT). Plates were washed six times in PBS before addition of 100 μl of ALP-conjugated streptavidin (Mabtech) diluted 1/1000 in PBS. Plates were incubated for 45 min at RT and then washed six times in water. A total of 100 μl of 5-bromo-4-chloro-3-indolyl phosphate/NBT-plus substrate (Mabtech) was added and incubated for 10 min at RT. Plates were then washed extensively with water and air dried. Spots were counted in an ImmunoSpot analyzer (Cellular Technology).
ELISA for serum Ab responses
Determination of Env-specific Ab levels in mouse serum was done by ELISA. Briefly, ELISA plates (Nunc) were precoated with 100 ng/well (1 μg/ml) of Galanthus nivalis lectin (Sigma-Aldrich) diluted in PBS and incubated ON at 4°C. Excess lectin was removed, and plates were washed in PBS before the addition of 100 ng/well (1 μg/ml) Env protein diluted in PBS. Plates were incubated for 2 h in RT before the excess protein was removed, and the plates were washed in wash buffer (PBS containing 0.05% Tween 20) and then incubated in blocking buffer (PBS containing 2% nonfat dry milk) at RT for 1 h. Serum from immunized mice was added in serial dilutions and incubated at RT for 2 h before washing. Plates were then incubated for 1 h at RT with a goat anti-mouse IgG-HRP Ab (Southern Biotechnology Associates) at a dilution of 1/5000. After washing the plates in water, the assay was developed using the OPD Fast kit (Sigma-Aldrich), and the OD was read at 450 nm (or at 492 nm if reaction was stopped with 50 μl/well H2SO4) using a Multiscan EX (MTX Lab Systems). In Fig. 5B, a different ELISA format was used. Briefly, ELISA plates were coated with 100 ng/well (1 μg/ml) goat anti-mouse IgG (Mabtech) diluted in PBS and incubated ON at 4°C. The solution was removed and plates were washed before the addition of blocking buffer. After 1 h of incubation at RT, plates were washed in wash buffer and serum from immunized mice was added in serial dilutions and incubated at RT for 2 h. Plates were washed in wash buffer, and 200 ng (2 μg/ml) of biotinylated Env protein diluted in PBS was added. Plates were incubated for 2 h at RT before excess protein was removed and the plates were washed. A total of 100 μl of streptavidin-HRP (Mabtech) was then added at a dilution of 1/1000 in PBS and incubated for 1 h at RT. After washing the plates in water, the assay was developed using the OPD Fast kit (Sigma-Aldrich) and the OD was read as above.
Statistical differences were determined by unpaired, two-tailed, Student’s t tests using GraphPad Prism software version 5 and considered significant at ∗ for p ≤ 0.05, ∗∗ for p ≤ 0.01, and ∗∗∗ for p ≤ 0.001.
Analysis of B cell ELISPOT assay formats to enumerate Env-specific ASC
To establish a robust system to analyze vaccine-induced B cell responses against HIV-1 Env, we sought to increase the sensitivity of the conventional B cell ELISPOT assay and to extend the range of immune parameters that can be quantified by such analysis. Ag-specific B cell ELISPOT assays are usually performed by coating Ag directly onto the membrane of ELISPOT plates and, following incubation with cells, probing with biotinylated anti-IgG (39, 47, 48, 49, 50, 51). We asked whether the specificity and the sensitivity of the assay could be improved by capturing all IgG generated by ASC during the cell incubation using anti-Fc Ab-coated membranes and, following removal of the cells, probing with biotinylated Ag in solution. To test this approach, we generated a probe based on the YU2 gp140-F trimers used as an immunogen in our studies (31, 32, 33, 34, 35, 43, 52). The gp140-F trimers were modified to contain a site-specific biotinylation motif (Avitag) at the C terminus of the protein, distant from all relevant B cell epitopes (39, 40) (Fig. 1⇓A). Matching the probe and the immunogen, and exposing all antigenic sites on the probe, would potentially allow detection of all vaccine-induced Env-directed ASC. We confirmed that the antigenicity of gp140-F probe was not altered by the enzymatic biotinylation (supplemental Fig. S1A), whereas, in contrast, chemical biotinylation of Env by lysine-based coupling disrupted Env recognition by several mAbs (supplemental Fig. S1B). These results demonstrated that enzymatically biotinylated Env was a preferable probe for the B cell ELISPOT assay.
The general procedure for a B cell ELISPOT assay is shown in Fig. 1⇑B. To directly compare the two Ag-specific B cell ELISPOT formats, splenocytes were harvested from mice immunized twice with gp140-F in adjuvant and Ag-specific ASC were measured using the conventional (Fig. 1⇑C, lower left panel) and the alternative (Fig. 1⇑C, lower right panel) assay formats. Total IgG-secreting cells were quantified by incubating cells on plates coated with anti-IgG, followed by detection with biotinylated anti-IgG (Fig. 1⇑C, upper left panel). As shown by representative images, the conventional assay for Ag-specific ASC resulted in poorly defined spots and high background compared with the alternative assay format, referred to in this study as the optimized B cell ELISPOT assay (Fig. 1⇑C).
To determine whether the optimized B cell ELISPOT assay was Ag sparing, mice were immunized twice with purified gp140-F in adjuvant or with β-gal protein in adjuvant, and the amount of Ag required for optimal sensitivity in the two assay formats was evaluated. Results from two representative animals are shown as ELISPOT images from wells with 3.3 × 105 splenocytes/well (Fig. 1⇑D). In the conventional assay, higher concentrations of Env for coating resulted in an increased signal, but also increased levels of background staining (Fig. 1⇑D, left panel). Furthermore, nonspecific spots were measured from Env-immunized mice probed with the β-gal protein and from β-gal-immunized mice probed with Env. In contrast, the optimized B cell ELISPOT format resulted in distinct spots that were more uniform in size, with substantially less background. The total number of ASC was considerably higher, and nonspecific spots were completely eliminated (Fig. 1⇑D, right panel). The amount of Ag used for probing in the optimized assay format was reduced from 10 to 0.5 μg/ml without compromising the signal. In addition, enumeration of total Env-specific ASC per 106 splenocytes showed that the optimized B cell ELISPOT was more sensitive than the conventional B cell ELISPOT assay (Fig. 1⇑E). A similar improvement in sensitivity and reduced background were observed when the two assay formats were compared using other Ags, including IL-2, IFN-γ, and IgE (supplemental Fig. S2). When the sensitivity of the optimized assay was evaluated using an Env-specific mouse hybridoma cell line (Fig. 1⇑F), there was a close concordance between the number of ASC detected with the total IgG ELISPOT and the Ag-specific ELISPOT assay, demonstrating that the two assays were equally sensitive.
We also sought to determine whether the assay could be applied to analysis of B cell responses in other animal model systems. Rabbits are commonly used to evaluate immunogens for induction of Ab responses. Accordingly, we adapted the optimized B cell ELISPOT assay to the detection of rabbit IgG and performed the assay on splenocytes from two Env-hyperimmune rabbits. The results showed that the detection of Env-specific ASC was sensitive and specific also in this vaccine-relevant animal model (Fig. 1⇑G). When the optimized B cell ELISPOT assay was adapted to the analysis of peripheral blood cells from Env-immunized nonhuman primates, ASC-producing Env-specific Ab of several different isotypes were detected (data not shown).
Development of Env probes for use in a differential B cell ELISPOT analysis
Analysis of Env-elicited B cell responses is often limited to serological binding analysis of serum Abs to native Env or to Env-derived peptides, followed by evaluation of the in vitro neutralizing capacity of the serum Abs. Studies describing the induction and evolution of B cell responses at the cellular level are generally lacking. Having established the optimized B cell ELISPOT assay format, we reasoned that it should be possible to determine the frequency of ASCs that are specific for different antigenic elements of Env, using probes that possess or lack specific structural determinants for subtractive analysis. Such a differential B cell ELISPOT assay would provide an improved means to evaluate vaccine-induced B cell responses as a powerful complement to serological analysis. Because most current Env-based vaccines do not elicit broad neutralization, a distinct possibility is that responses against the immunogenic V regions dominate over those directed against conserved elements. A sensitive, specific, and convenient system that defines such subspecificities at the B cell level would be of substantial value to identify vaccine candidates capable of directing B cell responses toward conserved and away from variable determinants.
To perform differential B cell ELISPOT assays, we next generated a panel of Env probes by systematically removing selected structural elements of the gp140-F probe shown in Fig. 1⇑A. The selective deletions, guided by previous structural and antigenic analysis (42, 53, 54, 55), were performed in a manner to ensure the conformational integrity of the resulting deleted glycoproteins, while maintaining the site-specific biotinylation motif at the C terminus of each probe. A trimeric and a monomeric form of gp120 (gp120-F and gp120, respectively), as well as selectively truncated versions of gp120-F (gp120-F-ΔV1/2, gp120-F-ΔV3, gp120-F-ΔV1/2/3), were generated (Fig. 2⇓A), as described in Materials and Methods. Following expression and purification from the culture supernatants, the Env probes were analyzed by gel electrophoresis. All proteins migrated with the expected apparent m.w. by SDS-PAGE (Fig. 2⇓B) and by blue native gel (supplemental Fig. S3A). We also confirmed that each probe possessed the expected antigenic profile by examining their capacity to be recognized by several well-characterized mAbs (supplemental Fig. S3B).
To ascertain that the biotinylated Env probes were functional and specific when used in the optimized ELISPOT assay, each probe was tested against two Env-specific mouse hybridoma cell lines, 5D4-F7 (specific for V2) and 2B5-H8 (specific for V3). The total number of IgG-producing cells was similar to Env-specific ASC for all probes, except for the gp120-F-ΔV3 and gp120-F-ΔV1/2/3 probes on 2B5-H8 cells and the gp120-F-ΔV1/2 and gp120-F-ΔV1/2/3 probes on 5D4-F7 cells. In those cases, no spots were detected. Also, no spots were detected when the cells were probed with the biotinylated β-gal-negative control probe. Representative ELISPOT images (Fig. 2⇑C) and percentage of detected cells of the total number of cells added per well (Fig. 2⇑D) are shown. In sum, these results validate the purity, antigenic integrity, and specificity of the HIV-1 Env probes.
B cell responses against different Env determinants following immunization
We next investigated the evolution of the Env-directed B cell response in adult BALB/c mice following two and three immunizations with gp140-F trimers (groups 1 and 3) or control protein (groups 2 and 4) in adjuvant. Spleens and sera were harvested 3 days after the second immunization (groups 1 and 2, post 2 immunizations) or 3 days after the third immunization (groups 3 and 4, post 3 immunizations). See supplemental Fig. S4A for experimental design. The serological response after two immunizations was confirmed to be similar in groups 1 and 3, as shown by the Env-specific IgG titers in sera 3 days after the second immunizations (Fig. 3⇓A), allowing a direct comparison between the two groups in the subsequent analysis. As expected, the circulating Env-specific IgG titer in group 3 was increased after the third immunization (Fig. 3⇓A, right panel).
The frequencies of ASC reactive against selected structural Env elements were next determined using the differential B cell ELISPOT assay. We observed ∼25% gp140-F-specific ASC of total IgG-secreting cells after two immunizations and close to 30% after three immunizations (Fig. 3⇑, B and C). A significant fraction of the response post 2 immunizations was gp41 specific, as detected by comparing the percentage of gp120-F-specific spots with the percentage of gp140-F-specific spots (p < 0.0001). Significant differences were also detected with the gp120-F-ΔV1/2/3 and gp120-F-ΔV3 probes compared with the gp120-F probe (p = 0.0096 and p = 0.0351, respectively) (Fig. 3⇑B, upper panel). When the percentage of gp140-F-specific ASC was normalized to 100%, 59% of the response was detected with the gp120-F probe, suggesting that the remaining 41% was gp41 specific (Fig. 3⇑B, lower left panel). Similarly, when the gp120-F-specific response was normalized to 100%, the response detected with the gp120-F-ΔV1/2/3 probe amounted to 71%, suggesting that 29% of the gp120-specific response was directed against V regions 1–3. The majority of these were directed against V3, as shown by using the gp120-F-ΔV3 probe (Fig. 3⇑B, lower right panel). As expected, the β-gal-immunized mice showed a response with the β-gal probe, but not with the Env probes.
When the post 3 immunization samples were analyzed, we observed a different pattern of responses. The dominant response after three immunizations was directed against the V regions of gp120 (Fig. 3⇑C). The percentages of ASC detected with the gp120-F-ΔV1/2/3, gp120-F-ΔV3, and gp120-F-ΔV1/2 probes were all significantly lower than those detected with gp120-F (p < 0.0001, p = 0.0002, and p = 0.0434, respectively) In contrast, the fraction of ASC directed against gp41 was no longer significant (Fig. 3⇑C, upper panel). When the gp120-F-specific response was normalized to 100%, ∼56% of the total gp120-specific response was directed against V regions 1–3, and the majority of this response was directed against V3 (Fig. 3⇑C, lower right panel). There was no significant difference between the percentage of ASC detected with the gp120-F trimeric and the gp120 monomeric probes, suggesting that the stable gp140-F trimers used for immunization did not stimulate a detectable number of trimer-specific B cells. Taken together, these results demonstrated that the gp41-specific response, which dominated after two immunizations, was not boosted by the third immunization. Instead, the third immunization resulted in an expansion of B cells specific for the V regions of gp120, consistent with the reported V3- and V1-directed neutralizing responses elicited by trimers and monomers, respectively (34).
Evolution of Env-specific B cell responses over time in spleen and BM
Having characterized the population of Env-specific ASC present in spleen 3 days after gp140-F trimer immunizations, we asked how the response would evolve over time in the spleen compared with the BM. We immunized mice twice or three times with gp140-F in adjuvant. The spleen and BM were harvested 3 days after the second (group 1) and third (group 3) immunizations, and 21 days after the second (group 2) and third (group 4) immunizations (supplemental Fig. S4B) for analysis using the differential B cell ELISPOT assay. The results in Fig. 4⇓ are shown in two ways, as follows: bar diagrams on the left represent percentage of Env-specific ASC of total IgG-secreting cells as detected with the different probes, and the pie charts on the right show percentage of reactivity against gp41 (calculated as the differential between the gp140-F and the gp120-F probe) and percentage of reactivity against V regions 1–3 (calculated as the differential between the gp140-F and the gp120-F-ΔV1/2/3 probe after the gp41 reactivity was subtracted). The analyses of splenocytes harvested 3 days after immunization were consistent with the results shown in Fig. 3⇑ (Fig. 4⇓). In group 1, there was a significant difference in the percentage of ASC detected with the gp140-F probe compared with the gp120-F probe (p = 0.0046), whereas no significant difference between these probes was detected after the third immunization. As previously observed, the V3-directed responses dominated after three immunizations (group 3) in a statistically significant manner (p < 0.0001 for the V3 region differential), whereas the differential between the gp120-F and gp120-F-ΔV1/2 probes was not significant at either time point.
When the ASC were enumerated 21 days after the second immunization (group 2), a significant fraction was directed against the V regions (p = 0.0038), suggesting selective expansion of these cells over time. Furthermore, a significant fraction (p = 0.0122) of V region-directed ASC was also measured in the BM 21 days after the second immunization. When responses in BM were measured 3 days (group 3) and 21 days (group 4) after the third immunization, a significant fraction of gp41-directed ASC was observed at day 3, but not at day 21. This difference may be explained by the time required for ASC expansion and redistribution to the BM following the third immunization. The relative frequencies of Env-specific ASC in the BM measured 21 days after both the second and the third immunization were qualitatively similar to those measured in the spleen at the same time point, except that the fraction of non-gp41, non-V region-directed ASC was higher in the BM at both time points, as illustrated by the pie charts. The V region-directed ASC in the pie charts were primarily represented by V3-directed responses, as shown by the bar diagrams.
In many instances, it is of interest to characterize the memory B cell response induced by candidate vaccines. Unlike plasma cells, memory B cells do not spontaneously secrete Abs, but can be stimulated to do so in vivo by re-exposure to Ag, or, in vitro, if cultured in the presence of antigenic or polyclonal stimulators that promote their expansion and differentiation into ASC (50). An advantage of the B cell ELISPOT analysis over serological analysis is that B cell ELISPOT approach allows an examination of the memory B cell compartment also. We applied the differential B cell ELISPOT assay to measure Env-specific memory B cells by culturing splenocytes collected 21 days after two or three immunizations with gp140-F trimers in LPS to expand and differentiate memory B cells to ASC. This analysis demonstrates that the specificities present in the total ASC population were reflected in the memory B cell pool (supplemental Fig. S5).
Finally, we asked whether the probes developed for the differential B cell ELISPOT assay described in this work could be used in ELISA for differential detection of Abs in sera. We selected the gp120-F and gp120-F-ΔV1/2/3 probes and performed ELISA on sera taken 21 days after three immunizations with gp140-F. The analysis was performed using two different formats, as follows: a standard ELISA in which the proteins were used for coating (Fig. 5⇓A), and an ELISA that mimics the format of the optimized B cell ELISPOT assay, i.e., by coating the ELISA wells with anti-IgG and probing with biotinylated proteins (Fig. 5⇓B). In the former ELISA format, we detected a difference between the two Ags, indicating a concordance between B cell ELISPOT responses and the serological responses.
In this study, we describe a comprehensive analysis of the B cell response elicited by soluble gp140-F Env trimers. Using a new and innovative differential B cell ELISPOT system, we examined B cell responses directed against distinct native structural elements of HIV-1 Env. This approach allowed us to examine, and for the first time quantify, the evolution of the B cell response directed against conformational determinants of a viral glycoprotein elicited by recombinant protein immunizations.
We demonstrate that ASC directed against gp41 dominated in the spleen after two immunizations, but their relative proportion diminished significantly during the course of the vaccination schedule. Instead, ASC directed against the gp120 V regions, in particular V3, expanded over time and especially following the third immunization. These analyses suggest that the soluble gp140-F trimers, although designed to be mimics of the viral functional spike, most likely expose and present the V3 region more than it is exposed on primary HIV-1 isolate spikes, many of which are resistant to V3-directed Abs. The response detected in the spleen 21 days after immunization was qualitatively similar to the response in the BM at the same time point, except that the fraction of non-gp41, non-V region-directed ASC was higher in the BM. Although not defined in the current study, such responses may represent ASC directed against more conserved epitopes such as those in the CD4 binding region, consistent with their appearance after multiple Env inoculations (45, 56). Probes containing point mutations designed to disrupt the CD4 binding site could be added in future analyses to address this question in a quantitative manner.
The dominating gp41-directed B cell response that we detected at early time points after the second gp140-F trimer inoculations (Fig. 3⇑) is intriguing given the early appearance of gp41-directed Abs in HIV-1-infected individuals (57, 58). HIV-1 gp41 Ab specificities are, except in rare cases, nonneutralizing, and broadly neutralizing gp41-directed Abs have not been elicited by any immunization regimen reported to date. Our data suggest that the early gp41-specific B cells stimulated by gp140-F trimer immunization are eventually outcompeted by B cells directed against the V regions, in particular V3. This may be because the V region-directed B cells, or a subset of these cells, progress through more productive affinity maturation processes than the gp41-directed B cells, which remain of lower affinity and less capable of expanding upon a subsequent boost.
In addition to providing information about region-specific responses to Env of relevance to HIV-1 vaccine development, the studies presented in this work show that structurally and antigenically well-defined Env immunogens provide unique model Ags for basic B cell questions aimed at dissecting the evolution of the B cell response during and following immunization with complex protein Ags. Despite our current understanding of different classes of B cell epitopes on HIV-1 Env, methods to analyze Env-specific B cell responses at the cellular level are limited. Attempts to define the neutralization specificities present in polyclonal sera were recently described (7, 59, 60, 61, 62). However, less is known about the cells that produce these Abs, including how they are affected by sequential immunization and which specificities are archived in the memory B cell pool. The generation and analysis of mAbs using single cell RT-PCR (40, 64) or other means (63), RT-PCR, followed by sequencing of Ab genes represent approaches to sample the B cell repertoire. The differential B cell ELISPOT system provides a complementary high throughput approach to evaluate the B cell compartment and to characterize different specificities of ASC induced by vaccination or natural infection.
Neutralizing Ab responses elicited by the gp140-F trimers were not analyzed in the current study, as we have reported such analyses previously (34, 52). Instead, we focused on measurements of total Env-specific ASC, a sum of both short- and long-lived plasma cells. More detailed analyses of the Env-specific memory B cell and plasma cell pools generated by immunization are critical for the optimal design of vaccine regimens aimed to focus the Ab response on broad neutralization determinants. Furthermore, the relative roles of memory B cells and long-lived plasma cells for maintaining serological responses after natural infection and vaccination are still under debate and may depend on the nature of the Ag (65, 66, 67, 68, 69, 70, 71). The use of the differential B cell ELISPOT combined with serological studies could help to address this question as it relates to HIV-1 Env.
In conclusion, the studies described in this work show that repeated homologous immunizations with the YU2 gp140-F trimers resulted in selective expansion of ASC directed against the V regions of Env. Approaches to shift the response away from these immunogenic nondesired specificities are needed. A better understanding of the hierarchy of different B cell epitopes and how to selectively stimulate B cells capable of producing high-affinity Abs that bind subdominant conserved and broadly neutralizing epitopes of the HIV-1 spike complex may be an important step toward the development of an effective HIV-1 vaccine.
We thank the personnel at the animal facility of the Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, for excellent technical assistance; Gerry McInerney and Christopher Sundling for valuable comments on the study; and Mythreyi Shastri for help with editing the manuscript. We also thank Linda Stertman and Karin Lövgren Bengtsson at Isconova for generous use of their adjuvant, and Brenda Hartman and Morteza Loghmani for help with the graphics.
P.D., S.P., R.T.W., and G.K.H. are listed as co-inventors on a patent application related to the optimized and differential B cell ELISpot. S.P. is an employee of Mabtech AB. The rest of the authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by grants from Karolinska Institutet, Swedish Research Council, Sida/Swedish Agency of Research Cooperation with Developing Countries, International AIDS Vaccine Initiative, Bill and Melinda Gates Foundation, National Institute of Allergy and Infectious Diseases, and National Institutes of Health intramural research program.
↵2 Address correspondence and reprint requests to Dr. Gunilla B. Karlsson Hedestam, Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Box 280, S-171 77 Stockholm, Sweden. E-mail address:
↵3 Abbreviations used in this paper: Env, envelope glycoprotein; ASC, Ab-secreting cell; β-gal, β-galactosidase; BM, bone marrow; ON, overnight; RT, room temperature.
↵4 The online version of this article contains supplemental material.
- Received February 10, 2009.
- Accepted July 8, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.