|Home | About | Journals | Submit | Contact Us | Français|
Broadly neutralizing antibodies (bnAbs) against highly variable viral pathogens are much sought-after to treat or protect against global circulating viruses. We have probed the neutralizing antibody repertoires of four HIV-infected donors with remarkably broad and potent neutralizing responses and rescued 17 new monoclonal antibodies (MAbs) that neutralize broadly across clades. Many of the new MAbs are almost 10-fold more potent than the recently described PG9, PG16, and VRC01 bnMAbs and 100-fold more potent than the original prototype HIV bnMAbs1–3. The MAbs largely recapitulate the neutralization breadth found in the corresponding donor serum and many recognize novel epitopes on envelope (Env) glycoprotein gp120, illuminating new targets for vaccine design. Analysis of neutralization by the full complement of anti-HIV bnMAbs now available reveals that certain combinations of antibodies provide significantly more favorable coverage of the enormous diversity of global circulating viruses than others and these combinations might be sought in active or passive immunization regimes. Overall, the isolation of multiple HIV bnMAbs, from several donors, that, in aggregate, provide broad coverage at low concentrations is a highly positive indicator for the eventual design of an effective antibody-based HIV vaccine.
Most successful anti-viral vaccines elicit neutralizing antibodies as a correlate of protection4,5. For highly variable viruses, such as HIV, HCV and, to a lesser extent, influenza, vaccine design efforts have been hampered by the difficulties associated with eliciting neutralizing antibodies that are effective against the enormous diversity of global circulating isolates (i.e. bnAbs)6,7. However, for HIV for example, 10–30% of infected individuals do, in fact, develop broadly neutralizing sera, and protective bnMAbs have been isolated from infected donors1,2,8–12. It has been suggested that, given the appropriate immunogen, it should be possible to elicit these types of responses by vaccination13 and understanding the properties of bnMAbs has become a major thrust in research on highly variable viruses.
We have previously screened sera from approximately 1,800 HIV infected donors for neutralization breadth and potency, designating the top 1% as “elite neutralizers”, based on a score incorporating both breadth and potency14. In this study, we set out to isolate bnMAbs from the top four elite neutralizers (Supplementary Table 1) by screening antibody-containing memory B cell supernatants for broad neutralizing activity using a recently described high-throughput functional approach2. Antibody variable genes were rescued from B cell cultures that displayed cross-clade neutralizing activity and expressed as full-length IgGs. Analysis of the sequences revealed that all of the MAbs isolated from each individual donor belong to a distant, but clonally related cluster of antibodies (Supplementary Table 2). Since it has been proposed that antibodies from HIV infected patients are often polyreactive15,16, we tested the new MAbs for binding to a panel of antigens and showed that they were not polyreactive (Supplementary Fig. 2).
The potency and breadth of the MAbs were next assessed on a 162-pseudovirus panel representing all major circulating HIV subtypes (Fig. 1, Supplementary Tables 3 and 4)2. All of the MAbs exhibited cross-clade neutralizing activity, but more strikingly, several displayed exceptional potency. The median IC50s and IC90s of PGT MAbs 121-123 and 125-128 were almost 10-fold lower (i.e. more potent) than the recently described PG9, PG16, VRC01, and PGV04 bnMAbs1,2 (Falkowska et al., manuscript in preparation), and approximately 100-fold lower than other bnMAbs described earlier (Fig. 1). At concentrations less than 0.1 μg/ml, these MAbs still neutralized 27% to 50% of viruses in the panel (Fig. 1 and Fig. 4a). Although PGT MAbs 135, 136, and 137 displayed lesser neutralization breadth than the other MAbs, they all still potently neutralized over 30% of the clade C viruses on the panel (Supplementary Fig. 2 and Supplementary Table 3b). This result is significant considering that HIV clade C predominates in sub-Saharan Africa and accounts for more than 50% of all HIV infections worldwide.
Interestingly, many of the clonally related MAbs exhibited differing degrees of overall neutralization potency. For example, the median IC50s of PGT MAbs 131, 136, 137, and 144 were approximately 10- to 50-fold higher than those of their somatically related sister clones (Fig. 1). Also, in some cases, the somatically related MAbs exhibited similar neutralization potency, but differing degrees of neutralization breadth, against the panel of viruses tested (Fig. 1 and Supplementary Tables 3 and 4). For example, PGT 128 neutralized with comparable overall potency but significantly greater neutralization breadth than the clonally related PGT 125, 126, and 127 MAbs (Fig. 1 and Supplementary Tables 3 and 4). Overall, these observations suggest that serum neutralization breadth may develop from the successive selection of somatic variants that bind to a modified epitope or a slightly different envelope (Env) conformation expressed on virus escape variants. Comparison of the neutralization profiles of the MAbs isolated from a given donor with that from the corresponding serum revealed that the isolated MAbs could largely recapitulate the serum neutralization breadth and potency (Fig. 2 and Supplementary Fig. 3).
We next sought to gain information on the epitopes recognized by the newly isolated bnMAbs. ELISA binding assays indicated that PGT MAbs 121-123, 125-128, 130, 131, and 135-137, bound to monomeric gp120 (Supplementary Table 5). In contrast, the PGT 141-145 bnMAbs exhibited a strong preference for membrane-bound, trimeric HIV Env (Supplementary Fig. 4). Based on this result, we postulated that these bnMAbs bound to quaternary epitopes similar to those of the recently described PG9 and PG16 bnMAbs2. Indeed, this hypothesis was confirmed by competition studies, N160K sensitivity, and, for PGT MAbs 141-144, an inability to neutralize JR-CSF pseudoviruses expressing homogenous Man9GlcNAc2 glycans17 (Supplementary Fig. 5).
To define the epitopes recognized by the remaining PGT antibodies, competition ELISA assays were carried out with a panel of well-characterized neutralizing and non-neutralizing antibodies (Fig. 3a). Unexpectedly, all of the remaining antibodies (PGT MAbs 121-123, 125-128, 130, 131, and 135-137) competed with the glycan-specific bnMAb 2G12. This result was surprising given that 2G12 had previously formed its own unique competition group. All of the MAbs, except for PGT MAbs 135, 136 and 137, also competed with a V3 loop-specific mAb and failed to bind to gp120 ΔV3, suggesting their epitopes were in proximity to or contiguous with V3 (Fig. 3a and Supplementary Table 5). Deglycosylation of gp120 with Endo H abolished binding by all the MAbs, indicating that certain oligomannose glycans were important for epitope recognition (Supplementary Table 5). Competition of these MAbs with 2G12 and lack of binding to deglycosylated gp120 prompted us to investigate whether these antibodies contacted glycans directly. Glycan array analysis revealed that PGT MAbs 125-128 and 130 bound specifically to both Man8GlcNAc2 and Man9GlcNAc2, whereas the remaining antibodies showed no detectable binding to high-mannose glycans (Fig. 3b). Interestingly, binding of PGT MAbs 125-128 and 130 to gp120 was competed by Man9, but, unlike 2G12, was not competed by monomeric mannose or Man4 (D1 arm of Man9GlcNAc2) (Fig. 3c,d), suggesting a different mode of glycan recognition. Furthermore, in contrast to 2G12, no evidence was found for domain exchange and monomeric Fab fragments still exhibited potent neutralizing activity (Supplementary Fig. 7 and data not shown).
To further define the epitopes recognized by the MAbs, neutralizing activity against a large panel of HIV-1JR-CSF variants incorporating single alanine substitutions was assessed using a single round of replication pseudovirus assay (Supplementary Table 6). In the panel of mutants, the N-linked glycans at positions 332 and/or 301 were important for neutralization by PGT MAbs 125-128, 130, and 131 suggesting their direct involvement in epitope formation. The apparent dependency on so few glycans suggests that, although these PGT MAbs contact Man8-9GlcNAc2 glycans directly, their arrangement in the context of gp120 is critical for high affinity glycan recognition and neutralization potency. This is further highlighted by the inability of the PGT MAbs to neutralize SIVmac239, HIV-2 or HCV, which display a high level of glycosylation (data not shown). Interestingly, although PGT MAbs 121-123 failed to exhibit detectable binding to high-mannose glycans and be competed by mannose sugars (Supplementary Fig. 6), the only substitutions that completely abolished neutralization by these MAbs were those that resulted in removal of the glycan at position 332. Although structural studies will be required to fully define the epitopes recognized by these antibodies, the above results suggest either that the PGT MAbs 121-123 bind to a protein epitope along the gp120 polypeptide backbone that is conformationally dependent on the N332 glycan or that the glycan contributes more strongly to binding in the context of the intact protein.
Vaccines against pathogens with low antigenic diversity, such as hepatitis B virus or measles, commonly achieve 90–95 % efficacy18. Similarly, the influenza vaccine achieves 85–90% efficacy in years when the vaccine and circulating seasonal strain are well-matched19,20. However, efficacy drops severely in years when there is a mismatch between the vaccine and circulating strain. In the case of HIV, the global diversity of circulating viruses is such that the match between the prophylactic antibodies and the circulating viruses, i.e. the antibody viral coverage, will be crucial for the degree of efficacy of active or passive prophylaxis approaches. To date, although the recent RV144 trial has led to speculation that some degree of protection against HIV may be achieved through extra-neutralizing activities of antibodies, such as antibody-dependent cell-mediated cytotoxicity or phagocytosis, the strongest evidence for protection is for neutralizing antibodies in non-human primate models using simian-human immunodeficiency virus (SHIV) challenge21–25. Passive administration of neutralizing antibodies in these models suggests that a serum antibody concentration of approximately or greater than 100 times the in vitro pseudovirus assay IC50 is required to achieve a meaningful level of protection21–25. Therefore, if a vaccine elicits a serum bNAb concentration on the order of 10 μg/ml26 and if an IC50: protective serum concentration ratio of 1:100 is assumed, then protection would be only achieved against viruses for which the bNAb IC50 is lower than 0.1 μg/ml. As a second more conservative scenario, for an IC50: protective serum concentration ratio of 1:500, protection would be achieved against viruses for which the bNAb IC50 is lower than 0.02 μg/ml. As shown in Figure 4, although various bnMAbs display breadth at high concentrations, viral coverage often drops sharply at lower concentrations. Therefore, if elicited or delivered singly, only the most potent Abs, such as 121 and 128, would be able to achieve a meaningful level of viral coverage, in particular at concentrations corresponding to the more conservative scenario given above. As bnMAbs display different and in some cases complementary breadth, we further looked at the coverage achieved by antibody combinations. For the two IC50: protective serum concentration ratios above, a combination of PGV04 and VRC01, the two most potent CD4bs bnMAbs, would provide protection against 50% and 3% of viruses, respectively (Fig. 4b). In contrast, for a vaccine eliciting antibodies with high potency and favorable non-overlapping breadth, such as 128 and 145, coverage would be achieved against 70% and 40% of viruses for the two scenarios (Fig 4c). Several combinations of two bnMAbs, including those directed to overlapping epitopes, can yield this degree of coverage (Supplementary Fig. 8). In addition, a combination of all of the bnMAbs would cover 89% and 62% of viruses, correspondingly. Coverage against such a large proportion of viruses would likely have an important impact on the pandemic.
In summary, an effective vaccine against HIV will likely require the elicitation of a combination of complementary potent neutralizing antibodies. The demonstration that large numbers of potent and diverse bnMAbs can be isolated from several different individuals provides grounds for renewed optimism that an antibody-based vaccine may be achievable.
Activated memory B cell supernatants were screened in a high throughput format for neutralization activity using a micro-neutralization assay, as described2. Heavy and light chain variable regions were isolated from B cell lysates of selected neutralizing hits by reverse transcription from RNA followed by multiplex PCR amplification using family-specific V-gene primer sets. For some antibodies, traditional cloning methods were used for antibody isolation, as described2. For other antibodies, amplicons from each lysate were uniquely tagged with multiplex identifier (MID) sequences and 454 sequencing regions (Roche). Single round of replication pseudovirus neutralization assays and cell surface binding assays were performed as described previously2,27,28. Glycan reactivities were profiled on a printed glycan microarray (version 5.0 from the Consortium for Functional Glycomics (CFG)) as described previously29.
The following antibodies and reagents were procured by the IAVI Neutralizing Antibody Consortium: antibody 2G12 (Polymun Scientific, Vienna, Austria), antibody F425/b4E8 (provided by Lisa Cavacini, Beth Israel Deaconess Medical Center, Boston, MA), soluble CD4 (Progenics, Tarrytown, NY), HxB2 gp120, SF162 gp120, BaL gp120, JR-FL gp120, JR-CSF gp120 and YU2 gp120 (provided by Guillaume Stewart-Jones, Oxford University). Purified ADA gp120 was produced in the laboratory of Robert Doms, University of Pennsylvania. Fab X5 was expressed in E.coli and purified using an anti-human Fab specific affinity column. Deglycosylated gp120 JRFL was expressed in HEK 293S GnTI−/− cells and treated with Endo H (Roche).
The donors identified for this study were selected from the IAVI sponsored study, Protocol G14. Eligibility for enrolment into Protocol G was defined as: male or female at least 18 years of age with documented HIV infection for a least three years, clinically asymptomatic at the time of enrolment, and not currently receiving antiretroviral therapy. Selection of individuals for monoclonal antibody generation was based on a rank-order high throughput screening and analytical algorithm14. Volunteers were identified as elite neutralizers based on broad and potent neutralizing activity against a cross-clade pseudovirus panel14.
The method for isolating human MAbs from memory B cells in circulation has previously been described2. Surface IgG+ B cells seeded at near clonal density in 384-well microplates were activated in short-term culture. Supernatants were screened for neutralization activity against 2-4 pseudotyped viruses for which neutralization activity was detected at high titers in the donor serum. Heavy and light chain variable regions were isolated from B cell lysates of selected neutralizing hits by reverse transcription from RNA followed by multiplex PCR amplification using family-specific V-gene primer sets. Amplicons from each lysate were uniquely tagged with multiplex identifier (MID) sequences and 454 sequencing regions (Roche, Indianapolis, IN). A normalized pooling of gamma, kappa and lambda chains was performed based on agarose gel image quantitation and the pool was analysed by 454 TitaniumR sequencing. Consensus sequences of the VH and VL chains were generated using the Amplicon Variant Analyzer (Roche) and assigned to specific B cell culture wells by decoding the MID tags. Selected VH and VL chains were synthesized and cloned in expression vectors with the appropriate IgG1, Igκ or Igλ constant domain. Monoclonal antibodies were reconstituted by transient transfection in HEK293 cells followed by purification from serum-free culture supernatants.
Antibody genes were cloned into an expression vector and transiently expressed with the FreeStyle 293 Expression System (Invitrogen, Carlsbad, CA). Antibodies were purified using affinity chromatography (Protein A Sepharose Fast Flow, GE Healthcare, UK) and purity and integrity checked with SDS-PAGE.
Neutralization by monoclonal antibodies and donor sera was performed by Monogram Biosciences using a single round of replication pseudovirus assay as previously described31. Briefly, pseudoviruses capable of a single round of infection were produced by co-transfection of HEK293 cells with a subgenomic plasmid, pHIV-1lucu3, that incorporates a firefly luciferase indicator gene and a second plasmid, pCXAS that expressed HIV-1 Env libraries or clones. Following transfection, pseudoviruses were harvested and used to infect U87 cell lines expressing co-receptors CCR5 or CXCR4. Pseudovirus neutralization assays using HIV-1JR-CSF alanine mutants are fully described elsewhere2. Neutralization activity of MAbs against HIV-1JR-CSF alanine mutants was measured using a TZM-BL assay, as described2. Kifunensine-treated pseudoviruses were produced by treating 293T cells with 25 μM kifunensine on the day of transfection. Memory B cell supernatants were screened in a micro-neutralization assay against a cross-clade panel of HIV-1 isolates and SIVmac239 (negative control). This assay was based on the 96- well pseudotyped HIV-1 neutralization assay (Monogram Biosciences) and was modified for screening 15 μl of B cell culture supernatants in a 384-well format.
Titrating amounts of antibodies were added to HIV-1 Env transfected 293T cells, incubated for 1 hr at 37°C, washed with FACS buffer, and stained with goat anti- human IgG F(ab′)2 conjugated to phycoerythin (Jackson ImmunoResearch, West Grove, PA). Binding was analyzed using flow cytometry, and binding curves were generated by plotting the mean fluorescence intensity of antigen binding as a function of antibody concentration. For competition assays, titrating amounts of competitor antibodies were added to the cells 30 min prior to adding biotinylated PGT MAbs at a concentration required to give EC50.
For antigen-binding ELISAs, serial dilutions of MAbs were added to antigen-coated wells and binding was probed with alkaline phosphatase-conjugated goat anti-human immunoglobulin G (IgG) F(ab′)2 Ab (Pierce, Rockford, IL). For competition ELISAs, titrating amounts of competitor MAbs were added to gp120-coated ELISA wells and incubated for 30 min prior to adding biotinylated PGT MAbs at a concentration required to give IC70. Biotinylated PGT MAbs were detected using alkaline phosphatase conjugated streptavidin (Pierce) and visualized using p-nitrophenol phosphate substrate (Sigma, St. Louis, MO).
Monoclonal antibodies were screened on a printed glycan microarray version 5.0 from the Consortium for Functional Glycomics (CFG) as described previously29. Antibodies were used at a concentration of 30 μg/ml and were precomplexed with 15 μg/ml secondary antibody (goat-anti-human-Fc-rPE, Jackson Immunoresearch) before addition to the slide. Complete glycan array data sets for all antibodies may be found at www.functionalglycomics.org in the CFG data archive under “cfg_rRequest_2250”.
The oligomannose dendrons (Man4D and Man9D) were synthesized by Cu(I) catalyzed alkyne-azide cycloaddition between azido oligomannose and the second generation of AB3 type alkynyl dendron. Detailed procedures and characterization were previously reported30.
NHS-activated glass slides (Nexterion slide H, Schott North American) were printed with robotic pin (Arrayit 946) to deposit gp120 JRFL at concentrations of 750 or 250 μg/ml in printing buffer (120 mM phosphate, pH 8.5; containing 5% glycerol and 0.01 % Tween 20). 12 replicates were used for each concentration. The printed slides were incubated in relative humidity 75% chamber overnight and treated with blocking solution (superblock blocking buffer in PBS, Thermo) at room temperature for 1 h. The slides were then rinsed with PBS-T (0.05% Tween 20) and PBS buffer, and centrifuged at 200 g to remove residual solution from slide surface.
Serial diluted oligomannose dendrons were mixed with MAb (40 μg/ml) in PBS-BT buffer (1% BSA and 0.05 % Tween 20 in PBS). The mixtures were applied directly to each sub-array on slide. After incubation in a humidified chamber for 1 h at RT, the slides were rinsed sequentially with PBS-T and PBS buffer, and then centrifuged at 200 g. Each sub-array was then stained with Cy3 labeled goat anti-human Fc IgG (7.5 μg/ml in PBSBT) for 1 h in a humidified chamber. The slides were then rinsed sequentially with PBS-T and demonized water and centrifuged at 200 g. The fluorescence of the final arrays was imaged at 10 μm resolution (Ex: 540 nm; Em: 595 nm) with an ArrayWorx microarray reader (Applied Precision).
Germ line genes were predicted using the immunoglobulin sequence alignment tools IMGT/V-QUEST32 and SoDA233. Clonally related sequences were identified by common germ line V-genes and long stretches of identical N-nucleotides.
Statistical analyses were done with Prism 5.0 for Mac (GraphPad, La Jolla, CA). Viruses that are not neutralized at an IC50 or IC90 < 50 μg/ml were given a value of 50 μg/ml for median calculations. For combinations of antibodies, a virus was counted as covered if at least one of the MAbs was neutralized depending on individual concentrations (IC50). This approach does not take additivity into account and therefore underestimates the neutralization potency of antibody combinations.
We thank N. Schultz, K. Le, and V. Thaney for technical assistance at TSRI. We also thank C. Corbaci, S. M. Eagol and C. Williams for assistance with Figure 3. We would also like to thank the research staff members of the Theraclone I-STAR Discovery and Antibody Purification Team, and K. Limoli, S. Jauregui, K. Tran, P. Sitou, J. Perea, B. Huang, and P. Guinto at Monogram BioSciences. We thank R. McBride for assistance with the glycan array and are grateful to W. Olson for providing purified KNH1144 SOSIP trimer, R. Wyatt for providing the YU2 Foldon trimer, and J. Mascola for providing MAb VRC01. In addition, we would like to thank all of the IAVI Protocol G project, clinical and site team members and all of the Protocol G clinical investigators, specifically, G. Miiro, A. Pozniak, D. McPhee, O. Manigart, E. Karita, A. Inwoley, W. Jaoko, J. DeHovitz, L.-G. Bekker, P. Pitisuttithum, R. Paris, J. Serwanga, and S. Allen. This work was supported by the International AIDS Vaccine Initiative (IAVI), NIAID, the Ragon Institute, and the United States for International Development (USAID). The contents are the responsibility of the authors and do not necessarily reflect the views of USAID or the United States Government.
Author Contributions Project planning was performed by L.M.W., M.H., K.J.D., E.F., P.P., P-Y.C-H., R.P., J-P. J., T.W., S.K.P, W.K., I.A.W. S.F., and D.R.B.; experimental work by L.M.W., A.R., M.H, K.J.D., E.F., R.P., J-P. J., P-Y.C-H., S-K.W., P.W.H., O.A.O., and T.W.; data analysis by M.H, L.M.W., K.J.D., E.F., P.P., T.W., P-Y.C-H., and D.R.B.; and manuscript composition and revision by L.M.W, M.H., K.J.D., R.P., J-P. J., E.F., W.K., I.A.W, D.R.B, and P.P.
Author Information Reprints and permissions information is available at www.nature.com/reprints. Gene sequences of the reported antibodies and the primers used for antibody isolation have been deposited under GenBank accession numbers JN201894-JN201927. The authors declare competing financial interests: details accompany the paper on www.nature.com/nature. Readers are welcome to comment on the online version of this article at www.nature.com/nature.