HIV BROADLY NEUTRALIZING ANTIBODIES: TAKING GOOD CARE OF THE 98%

SUMMARY BLURB

In this issue of Immunity, Huang et al. describe an exceptionally broad and potent neutralizing antibody to HIV. This antibody, N6, is capable of neutralizing up to 98% of global isolates with a potent median IC₅₀ of 0.04 μg/ml, making it the current “best-in-class” for bnAbs targeting the CD4 binding site.

MAIN TEXT

Developing an HIV vaccine is among the most difficult challenges in modern medicine. The challenge is due, in part, to the virus being a molecular escape artist, capable of continually generating mutants that are resistant to immune responses. In addition to this high molecular diversity, the HIV surface spikes, which are the sole target of protective neutralizing antibodies, are coated in a high density of self-sugars that derive from the host glycosylation machinery and are therefore poorly recognized by the immune system. In this tug of war between the human immune system and the virus, antibodies must not only pursue escape viruses, but must also puncture through the glycan shield to reach the protein surface of the spikes (Burton and Hangartner, 2016).

Despite the formidable defenses, prolonged exposure to HIV in some infected individuals results in the generation of antibodies that meet both of the challenges of epitope variability and glycan shielding. These are the so-called broadly neutralizing antibodies (bnAbs) that are able to recognize a diversity of HIV isolates. Biochemical, structural and virological studies have revealed that bnAbs recognize relatively conserved regions on HIV Env, avoid variable regions, and accommodate, avoid and/or directly bind sugars that surround the critical epitopes.

In the study presented by Huang et al., the authors describe an antibody called N6, which is a new member of the “VRC01-class” of bnAbs that target the CD4 binding site of gp120 subunits of the HIV surface spike. VRC01-class antibodies characteristically derive from the V_H1-2*02 heavy chain gene and harbor a rare L-CDR3 that is 5 amino acids in length (Zhou et al., 2015). The V_H1-2*02 heavy chain mimics CD4 in how it binds the receptor binding site, which explains the generally high neutralization breadth of this class of antibodies, hovering between 80 and 90% cross-clade neutralization. Although extraordinary in neutralization breadth, this class of antibodies typically exhibits modest potency compared to bnAbs targeting other epitopes on the surface spike. N6, however, is exceptional: this monoclonal antibody neutralizes up to 98% of a large panel of global isolates of different subtypes or clades at a median IC₅₀ of 0.04 μg/ml (Figure 1). Such high potency rivals bnAbs such as PGT121 and PG9, which target the V3-glycan and V2-apex epitopes at the “top” of the surface spike and are known for being exceptionally potent (Walker et al., 2011). Importantly, N6 also shows minimal autoreactivity by a number of measurements, favoring this antibody for use in therapeutic or prophylactic modalities. Based on this profile, N6 represents the current “best-in-class” for bnAbs targeting the CD4 binding site.

Figure 1. Breadth and potency of N6 relative to a number of other prototype bnAbs.

Broadly neutralizing antibodies targeting different sites of vulnerability on HIV Env are plotted based on their percent coverage and potency. Coverage is defined as the percent of cross-clade viruses neutralized in a large virus panel and potency is defined as the median IC₅₀ of neutralized viruses in μg/mL. The main bnAb epitopes are indicated by color as shown in the figure key. Note that different panels of viruses have been used for many of the antibodies, so comparisons are approximate. For example, PGDM1400 was reported to have 83% breadth and a median IC₅₀ of 0.003 μg/ml {Sok:2014dl}, but measured on the panel in Huang et al., the breadth was reported to be 78% with a median IC₅₀ of 0.008 μg/ml. See Burton and Hangartner (2016) for further details of this type of plot.

The extraordinary neutralization breadth and potency of N6 can be attributed to the ability of the humoral immune response to hone antibodies to precisely target a relatively conserved epitope region. Honing of N6 to the CD4 binding site involves two main elements. First, unlike other members of the VRC01-class of bnAbs, N6 is less dependent on variable loop 5 of gp120 and instead has shifted its binding surface towards the more conserved loop D. Second, the antibody features a GlyGlyGly motif in the H-CDR2 that, in combination with binding at an angle that brings the antibody closer to loop D, allows N6 to accommodate potential elongation and/or the addition of glycan sites on variable loop 5. Thus, N6 has enhanced binding to the most conserved features on HIV Env, while also being able to accommodate surrounding variability and addition of glycan sites that would otherwise obstruct binding for other VRC01-class antibodies.

Isolation of N6 was critically dependent on two factors: 1) the identification of a donor who demonstrated exceptional serum neutralization breadth and potency, and 2) the use of a technology to isolate rare antibodies from an overwhelming preponderance of non-neutralizing and isolate-specific antibodies. There has been a marked evolution in both of these factors over the years. Phage display yielded the first broadly neutralizing antibody to the CD4bs called b12 (Burton et al., 1994), which exhibits 34% neutralization breadth with a median IC₅₀ of ~ 2.8 μg/ml. Subsequent efforts to isolate new bnAbs to the CD4bs involved single memory B cell sorting using epitope specific baits. In the case of VRC01, the first VRC01-class bnAb isolated, memory B cells were positively selected using a resurfaced stabilized gp120 core called RSC3 coupled to a negative selection step using a CD4 binding site knock-out gp120 core that had an isoleucine at position 371 removed (ΔRSC3) (Wu et al., 2010). The advantage of using such antigen baits for sorting memory B cells is that a specific phenotype can be selected for (in this case, isolating antibodies that target the CD4 binding site) and the recovery is limited largely by reverse transcription and PCR efficiencies. Indeed one of the most potent bnAbs isolated to date called PGDM1400 (median IC₅₀ of 0.003 μg/ml neutralizing 83% of a large panel of global isolates) (Sok et al., 2014) was generated by this strategy. The PGDM1400 antibody is highly quaternary-structure dependent and was discovered using a recombinant surface spike mimic (BG505 SOSIP.664) as a positive and monomeric gp120 as a negative selecting agent. For the N6 donor, however, the best CD4 binding site bnAb isolated (VRC27) using selection baits was not outstanding and the authors turned to an alternative approach in which they performed sorting and activation of memory B cells using cytokine-secreting feeder cells and subsequent high-throughput screening of supernatants for neutralization. In this approach, memory B cells are not selected for binding to specific epitopes, but are instead screened directly for neutralization activity. This approach has successfully led to the isolation of a number of highly broad and potent antibodies (Walker et al., 2011), but the functional screen is dependent on the effective concentration of secreted antibody in the supernatant and therefore typically selects for only the most potent neutralizing antibodies. The approach is also limited by the in vitro activation efficiency of memory B cells. Once N6 was discovered by this approach, it was apparent that the RSC/ΔRSC3 strategy would not be effective for isolating N6 since its binding is reasonably insensitive to single point mutations, notably the substitution at position 371. In summary, the use of baits for isolating bnAbs is highly efficient, but only if the baits accurately mimic the native antigen and there is an effective negative selection bait. Otherwise, activation and screening of memory B cells can be used to select for Abs based on neutralization without selecting for epitope specificities, although this approach is relatively more resource intensive in terms of cost and labor.

N6 is an exceptionally broad and potent neutralizing antibody to the CD4 binding site of HIV Env. The continued isolation of bnAbs like N6 will contribute not only to vaccine efforts, but antibodies with such levels of breadth and potency and minimal autoreactivity are also being championed for potential use as prophylactic and therapeutic agents. With regard to the former, the discovery of additional bnAbs will help to define common features that are shared between bnAbs, such as those that define the VRC01-class of bnAbs (Zhou et al., 2015), and will also help to define shared features between overlapping epitopes, which would feed directly into immunogen design to faithfully recapitulate the critical features of the epitope. For prophylaxis and treatment, bnAbs have potential advantage over small molecules in terms of half-life that may improve adherence by requiring infrequent administration compared to the current daily intake of small molecule antiretrovirals (Caskey et al., 2015; Ledgerwood et al., 2015). Additionally, bnAbs typically retain important Fc effector functions, which might play a key role in activity against infected cells (Boesch et al., 2015), which is of potential import in cure strategies. Finally, antibody delivery for prophylaxis or treatment holds particular promise if, like N6, the antibody is highly potent, thereby reducing required dose and consequently unit costs.