The modern era of HIV-1 vaccine development

See allHide authors and affiliations

Science  10 Jul 2015:
Vol. 349, Issue 6244, pp. 139-140
DOI: 10.1126/science.aac7800

More than 30 years after the discovery of HIV-1, where are we in our quest to develop a vaccine that could prevent human infections and help stem the global HIV-1 pandemic? The truth is, unfortunately, that we are not yet close. Even at the laboratory bench, we do not have a vaccine that can induce cross-reactive neutralizing antibodies—the type of response likely needed to provide high-level protective immunity (1, 2). This goal of inducing neutralizing antibodies has proved especially intractable and remains a focus of HIV-1 vaccine researchers. A major culprit in this story is the structurally complex surface envelope glycoprotein (Env) that mediates entry of HIV-1 into host cells (3). On pages 10.1126/science.aac4223, 191, and 156 of this issue, Sanders et al. (4), Chen et al. (5), and Jardine et al. (6), respectively, and another study (7), provide new insights into the nature of the HIV-1 Env trimer, its potential use as a vaccine, and how the humoral immune system generates antibodies with the necessary characteristics to neutralize HIV-1.

Env is composed of three identical precursor glycoprotein molecules (gp160) that are cleaved into a surface component (gp120) and a membrane component (gp41), which noncovalently associate to form the entry-competent Env trimer—also called the viral spike. HIV-1 Env resists neutralization by a combination of remarkable antigenic diversity and concealing critical epitopes via its quaternary structure and heavy glycosylation. Moreover, antibodies must undergo extensive evolution (somatic hypermutation) to gain the ability to recognize the native trimer and block HIV-1 infection (1, 8) (see the figure).

Shortly after the discovery of HIV-1, the viral genes for gp160 and gp120 were cloned and encoded into viral vaccine vectors or expressed as recombinant protein vaccines. However, these envelope vaccines did not elicit broadly reactive neutralizing antibodies, and gp120 protein vaccines failed to protect against HIV-1 infection in human efficacy trials (2). The RV144 vaccine trial, which used a pox virus prime and gp120 boost, showed a modest 31% protective efficacy (2). There is now a sense of why these vaccines were suboptimal, and how to address the problem. A key concept is that much of the antigenic surface of a monomeric gp120 molecule is buried within trimeric Env, and thus, most antibodies elicited by gp120 do not bind the functional viral spike (3). The solution to this problem is conceptually simple—express the native Env trimer as a vaccine immunogen. Technically, this is a challenge, partly because the noncovalent interactions between gp120 and gp41 do not allow for trimer stability outside the context of the viral membrane.

Sanders et al. solved this problem by engineering a protein with specific disulfide bonds to hold gp120 and gp41 together called BG505 SOSIP; BG505 is the strain of virus used to make the protein, SOS refers to the disulfide bond, and IP is a specific mutation of gp41 required for SOSIP trimer stability (9). The authors have now used BG505 SOSIP to immunize rabbits and monkeys. To a casual reader, their results would seem disappointing: The vaccine elicited neutralizing antibodies to the autologous virus (matching the vaccine), and not to heterologous viruses. Because HIV-1 is genetically diverse, this trimer has certainly not solved the HIV-1 vaccine problem. The key question is whether these results are a first step on a pathway to generate more cross-reactive responses. Some of these next steps are already underway, including development of alternative strains and platforms to make additional trimers (10) and the ability to further stabilize the trimer to potentially improve its immunogenicity (11). In this regard, Chen et al. report the ability to produce near-homogeneous native Env trimers on the cell surface of stably transfected cell lines. Interestingly, they demonstrate that alterations made to the intracytoplasmic region of gp41, something that is routinely done when soluble gp140 proteins are expressed, can adversely affect trimer conformation and antigenicity.

Although native trimers provide a major new tool for HIV-1 vaccine research and development, there is still the problem that high levels of somatic mutation, and in some cases preferential immunoglobulin variable gene usage, are required for effective neutralization. Antibodies to the CD4 binding site on gp120 are of particular interest because they bind to a functionally conserved region on the virus. A major category of such antibodies, called the VRC01 class, can neutralize HIV-1 by mimicking the interaction of the cellular receptor CD4 with gp120. Importantly, VRC01-class antibodies are found in multiple HIV-1–infected donors and, hence, appear to represent a reproducible solution to targeting a vulnerable site on the HIV-1 trimer (12). All VRC01-class antibodies have two key immunological characteristics: They derive from the same heavy-chain variable gene allele, VH1-2*02, and they have a short (five amino acid) CDR3 region of the light chain—needed to avoid clashes with parts of the Env trimer.

Stimulating antibody lineages.

The ability to express the native HIV-1 Env trimer as a recombinant protein is a major step forward for vaccine development. Through somatic mutation, an antibody lineage evolves from its unmutated common ancestor, through a series of intermediates, to a mature form that can neutralize the virus. Current vaccine designs, such as the BG505 SOSIP and engineered gp120 variants, appear to stimulate intermediate antibodies.


To study how the immune system pairs this unusual short light chain with the VH1-2*02–derived heavy chain, Jardine et al. and Dosenovic et al. (7) generated transgenic mice that express the full human heavy chain of a VRC01-class antibody. Antibody-producing B cells from these mice preferentially express the human VH1-2*02–derived heavy chain, but importantly, the murine light chain repertoire is intact. Jardine et al. generated mice that expressed the near-germline (i.e., unmutated) heavy chain of the VRC01 antibody and then immunized the mice with a modified gp120 protein specifically engineered to engage this antibody. Notably, less than 0.1% of mouse light chains have the required characteristics to produce an appropriate VRC01-class neutralizing antibody. Yet, over 90% of vaccine elicited B cells targeting the CD4 binding site region that contained the appropriate five–amino acid light chain. The results are illuminating—they suggest that with appropriate antigenic stimulation, the rare immunoglobulin gene rearrangements required to produce neutralizing antibodies are efficiently produced. Importantly, the antibodies generated were unable to neutralize HIV-1—they were early VRC01 precursor antibodies that would likely require further somatic mutation to display effective virus neutralization.

Dosenovic et al. engineered mice to express the germline-reverted VRC01-class antibody called 3BNC60, and immunized these mice with engineered variants of gp120. They demonstrate the generation of some 3BNC60 precursor antibodies—but again without neutralizing activity. However, in transgenic mice expressing the mature 3BNC60 heavy chain, immunization with the BG505 SOSIP trimer produced measurable amounts of serum neutralization by VRC01-class antibodies. Antigen-specific B cells were cloned, and the antibodies they produced contained the expected mature 3BNC60 heavy chain and the necessary five–amino acid CDR3 light-chain motif. Overall, the results of Jardine et al. and Dosenovic et al. show the proficiency of the humoral immune system in pairing rare light chains with the appropriate heavy chain to produce effective antibodies.

How do these four studies advance the potential for an effective HIV-1 vaccine? As a first step, they show that we can now immunize with authentic native trimer proteins—and thus, there is a greater potential to generate antibodies that will recognize the functional viral spike and neutralize the virus. Single trimer proteins are just a first step, but they should provide the necessary framework from which to design more effective vaccine immunogens. The studies also demonstrate proof of concept that HIV-1 Env immunogens can be modified, based on our knowledge of structure and immune recognition, and that such immunogens can efficiently engage the appropriate naïve B cells.

Based on our understanding of the circumstances that lead to the generation of broadly neutralizing antibodies in HIV-1–infected individuals (13, 14), success achieved through immunization will likely require various sequential immunogens and iterative human trials. Whether specifically engineered smaller portions of Env, or the full trimer—or perhaps both—will be needed for driving difficult-to-induce B cell lineages should be determined in these immunization trials. We now have novel vaccine immunogens along with an understanding of immune pathways leading to effective antibodies, and the technologic tools to study the B cell response in detail. It may be reasonable to conclude that we have finally entered the modern era of HIV-1 vaccinology.

References and Notes

  1. Acknowledgments: I thank J. Stuckey for help with the figure sketch.
View Abstract

Stay Connected to Science

Navigate This Article