The identification of broadly neutralizing monoclonal antibodies against HIV-1 may aid efforts to design a vaccine.
Despite a massive research effort stretching back more than 20 years, development of a prophylactic HIV-1 vaccine has been hindered by the ability of the virus to evade host immune defenses through rapid antigenic variation and epitope masking. Neutralizing antibodies generated in infected individuals often provide a useful starting point for vaccine design, but antibody responses against HIV-1 are in most cases highly specific for the original viral strain and do not keep pace with newly evolving quasi-species. Two new studies in Science by Wu et al.1 and Zhou et al.2 present an approach that may lead to an HIV-1 vaccine that can induce a broad and potent neutralizing antibody response. They describe the identification1 and structural characterization2 of human broadly neutralizing monoclonal antibodies directed toward the conserved CD4 receptor binding site on the viral envelope glycoprotein gp120. Importantly, the antibodies were selected directly from infected donor sera using a recently developed technique of direct clonal B-cell sorting3, and the selection probes were rationally designed to identify antibodies specific for the targeted receptor binding site. The hope is that these antibodies will facilitate the engineering of vaccine candidates capable of focusing the immune response on highly conserved protective epitopes and inducing a broadly neutralizing antibody repertoire in immunized individuals.
The selection strategy employed by Wu et al.1 is a critical aspect of the work as it offers a way of identifying broadly neutralizing antibodies in the small percentage of infected individuals able to produce a protective antibody repertoire (called 'nonprogressers' or 'elite controllers'). Over the last decade, several such antibodies have been found4. Their immunologic targets include structurally conserved or functionally important epitopes, such as the CD4 binding site, chemokine co-receptor binding sites, the high-mannan glycan shield, the membrane proximal region of the viral envelope protein gp41 and the gp41 pre-hairpin intermediate. However, the coverage breadth of the antibodies to these targets is generally limited to ∼40–50% of viral strains across all clades, and their potency varies widely.
Vaccine researchers have used several experimental techniques for generating neutralizing monoclonal antibodies, including human hybridoma generation, immortalization of B cells with Epstein Barr virus and combinatorial display5. However, the quality of the antibodies discovered with such methods is directly related to the quality of the antigens used for panning and selection. When antigen selection is uninformed by structural knowledge of immunologically relevant conformations—as has often been the case in HIV-1 vaccine research—the resulting antibodies are likely to be suboptimal.
The new work1,2 uses two strategies that offer a significant advantage over previous efforts (Fig. 1). First, the antigen probes were designed using a technique called 'resurfacing' in which a relevant neutralizing epitope—in this case the HIV-1 CD4 binding site—is presented in the context of an immunologically irrelevant scaffold—here, a simian immunodeficiency virus (SIV) gp120 framework. The investigators used knowledge of immunologically relevant CD4 binding site conformations garnered from previous studies with neutralizing antibodies along with computational modeling to precisely define the desired epitope.
Second, monoclonal antibodies were directly isolated from individual B cells of HIV-1–infected individuals by antigen-specific, memory-B-cell sorting3. In this approach, donors with a strong positive serum reactivity to the probes were selected and their memory-B-cell repertoires were screened to identify clones that bind the antigen with high affinity. Nonimmortalized B-cell culture has been used previously to identify the HIV-1 broadly neutralizing monoclonal antibodies PG9 and PG16, with the primary screening assay being virus neutralization6. Interestingly, both antibodies recognize a cryptic epitope on the native HIV-1 envelope trimer but do not bind gp120 or gp41 in enzyme-linked immunosorbent assays.
The source of antibodies is an important distinguishing feature of the current work1,2 because phage display and other cloned antibody libraries may show selection biases, often yielding monoclonal antibodies of relatively low affinity and moderate specificity. The most potent HIV-1 broadly neutralizing monoclonal antibodies have all been isolated directly from human B cells and share the common characteristic of extensive maturation relative to germline sequence, reflecting the immune response of the host to the evolving HIV-1 infection. One of the newly identified broadly neutralizing antibodies, VRC01, exhibited ∼30% and ∼20% divergence for VH and VL chains, respectively1, and the antibody contained an additional disulfide bond as well as residue deletions within its Lκ-chain2. Similarly, PG9 and PG16 vary by ∼25% from their germline parent, with no single mutation accounting for their broad cross-reactivity7.
The discovery of broadly neutralizing monoclonal antibodies such as VRC01 raises expectations that an effective prophylactic vaccine can be generated by reverse engineering of appropriate immunogens. In principle, the antibody can be used to design candidate immunogens that focus the immune response on the desired protective epitope. For example, the crystal structure of VRC01 bound to the resurfaced gp120 probe used in its identification2 could inform modifications of the epitope surface to increase binding affinity or mask nonproductive irrelevant antibody responses. Such modifications might include single-residue substitutions, addition of glycan shielding sites or introduction of conformational constraints. Designed immunogens could then be tested in animal models for functional responses such as competitive binding or neutralization.
Thus far, attempts to reverse engineer immunogens from neutralizing monoclonal antibodies have not met with success, although promising results were recently reported for both HIV-1 (ref. 8) and influenza9. Challenges include correct structural presentation of complex, discontinuous epitopes and focusing of the immune response on desired regions of the molecule. The resurfaced gp120 probe used to identify VRC01 presents the HIV-1 CD4 binding site in the context of an SIV framework1, and immunization with this protein would be expected to produce antibodies against both target and framework. It is an open question whether the framework-specific response would be immunodominant and whether the proportion of antibodies directed to the HIV-1 CD4 binding site would be high enough to effect protection. In addition, the antibody response directed to sterically restricted or transient conformational intermediates, such as those presented on gp41 and CD4-inducible epitopes on gp120, may be thermodynamically or kinetically limited in potency.
Finally, many HIV-1–specific broadly neutralizing monoclonal antibodies have a distinctive architecture that may itself pose a challenge to vaccine design. Such antibodies are characterized by extended complementarity-determining regions, lipid-binding capability and use of domain swapping. If these structural features are genetically restricted in the general population and are critical to neutralization potency, there is at present no way to bias the immune response towards production of such antibodies. Furthermore, the extended complementarity-determining regions and hydrophobic combining sites of HIV-1 broadly neutralizing monoclonal antibodies that bind near the membrane surface may mediate polyreactivity with human proteins, such as cardiolipin and various nuclear antigens. One model postulates that polyreactive antibodies are actively eliminated from the repertoire during B-cell maturation owing to their anti-self activity, and that this accounts for the poor ability to elicit them with designed immunogens.
Despite the remaining obstacles, there is considerable cause for optimism as greater understanding of the immune system opens the door to rational manipulation. Challenges, such as the need to induce highly matured antibodies to specific epitopes, may be addressable through optimized immunization regimens and novel adjuvants. For example, a heterologous prime-boost regimen that involved priming with a canarypox vector followed by boosting with recombinant gp120 antigens has demonstrated a modicum of efficacy in a large phase 2 clinical trial10. Taken together, recent developments in HIV-1 research raise the prospect of an effective vaccine in the not-too-distant future.
Wu, X. et al. Science 329, 856–861 (2010).
Zhou, T. et al. Science 329, 811–817 (2010).
Scheid, J.F. et al. J. Immunol. Methods 343, 65–67 (2009).
Kwong, P.D. & Wilson, I.A. Nat. Immunol. 10, 573–578 (2009).
Hammond, P.W. mAbs 2, 157–164 (2010).
Walker, L.M. et al. Science 326, 285–289 (2009).
Pancera, M. et al. J. Virol. 84, 8098–8110 (2010).
Bianchi, E. et al. Proc. Natl. Acad. Sci. USA 107, 10655–10660 (2010).
Bommakanti, G. et al. Proc. Natl. Acad. Sci. USA. 107, 13701–13706 (2010).
Rerks-Ngarm, S. et al. N. Engl. J. Med. 361, 2209–2220 (2009).
J.G.J. and J.t.M. are employees of Merck & Co.
About this article
Cite this article
Joyce, J., ter Meulen, J. Pushing the envelope on HIV-1 neutralization. Nat Biotechnol 28, 929–931 (2010). https://doi.org/10.1038/nbt0910-929