Protein Progress

Like the development of the antibodies themselves, understanding and optimizing immunogens designed to elicit a broadly neutralizing response against HIV takes time.

By Morgane Rolland and Yegor Voronin

In science, results often come slowly. Vaccine research is a painstaking process and sometimes it can be hard to tell whether progress is being Rolland Morgane 8x10Fmade at all. But the results presented at the Keystone Symposium on HIV Vaccines, held in Steamboat Springs, CO, March 26-30, are unambiguous—researchers are making strides in their attempts to develop antibody-based vaccine candidates. Thanks to continuing efforts to learn how broadly neutralizing antibodies (bNAbs) are generated during natural infection and application of this knowledge to design and test novel immunogens, there is a wealth of promising results and discoveries to report.

Responses induced by trimeric proteins

One major effort to develop bNAb-based vaccine candidates relies on presenting the immune system with an HIV protein that mimics the natural structure of the virus’s trimeric Envelope protein(1). These native-like Env proteins (including the SOSIP family of native-like trimers and related approaches; see Many Keys to Protection but Many Locks Remain, IAVI Report, Vol. 20, Issue 4, 2016) have not yet been tested in humans, but several groups are studying them in various animal models(2,3).

At Keystone, Andrew Ward, associate professor at The Scripps Research Institute (TSRI) in La Jolla, CA, synthesized the lessons learned from several independent studies of antibodies isolated from trimer-vaccinated rabbits and macaques. Antibodies to autologous viruses (those that carry the exact versions of Envs that were used for immunization) were observed at different levels across studies. The antibodies were primarily able to block the easier to neutralize Tier-1 viruses, with a few native-like trimeric proteins eliciting antibodies to the harder to neutralize Tier-2 panel of viruses.

Some of the trimers evaluated to date have chinks in their glycan armor, characterized by glycans missing in key spots on the surface. Yegor AIDS Conference 9 12 12 143These so-called “glycan holes” were targeted by antibodies elicited in response to these trimeric immunogens in multiple animals(3,4). In some cases these responses led to effective neutralization of autologous viruses. But it remains unclear whether a vaccine strategy based on glycan hole-targeting antibodies would lead to the development of a broad antibody response targeting multiple viral variants, which is what would be required to fend off the broad spectrum of viruses in circulation, or whether it will lead to a potent, but narrowly focused response. Data from an in silico analysis presented by Kshitij Wagh, a staff scientist at the Los Alamos National Laboratory, indicated that glycan holes have a negative effect on the development of neutralization breadth and that focusing immune responses on glycan holes may favor strain-specific antibodies at the expense of a broader antibody response. Ward pointed out that the first of the constructed native trimers, BG505 SOSIP.664, is missing two glycans at positions 241 and 289. Presence of the glycan 289 would completely block binding of antibodies targeting this “hole” and this glycan is found on a substantial proportion of viruses around the world, thereby making them resistant to the glycan hole-focused antibodies induced by this trimeric protein.

Other SOSIP-elicited antibodies discussed at Keystone targeted the gp120/gp41 interface, fusion peptide, V3, and V1/N332 supersite, often in a similar manner in multiple animals (see Figure 1, below). These are promising results because they are reminiscent of responses found in natural HIV infection, including those that lead to the development of bNAbs in a subset of chronically HIV-infected individuals. However, some antibody responses were not representative of what has been seen in people, such as those targeting the base of the soluble trimer that normally faces the lipid membrane of the virion and is not easily accessible by antibodies(5). The consequences of these off-target responses need to be further evaluated to understand whether they can be ignored or will need to be silenced so as not to distract the immune response from the more critical spots of weakness on the viral Envelope.

Figure 1. HIV Env trimer  

Fig bnAb Epitopes.r2

Fully glycosylated crystal structure of BG505 SOSIP.664 Env trimer (side view on left and top view on right) with Env-gp120 colored in white and Env-gp41 in gray. Key Env sites mentioned in the article are shown here with the epitope sites corresponding to representative antibodies: V2-Glycan (PG9), V3-Glycan (PGT122), and fusion peptide (PGT151). The glycan hole created by the absence of glycosylation at amino acid sites 241 and 289 is represented in pink. Image courtesy of Hongjun Bai, US Military HIV Research Program, Walter Reed Army Institute of Research.


Researchers are also exploring ways to augment the responses induced by trimeric protein immunogens. Data presented by Colin Havenar-Daughton, a scientific associate in Shane Crotty’s laboratory at the La Jolla Institute for Allergy and Immunology, showed that rhesus macaques generated strong immune responses to native-like trimers administered along with the adjuvant iscomatrix. He highlighted results from Dennis Burton’s laboratory at TSRI showing that despite the presence of glycan holes on the BG505 SOSIPs, the majority of responses in macaques targeted other parts of the trimers and were variable in specificity and magnitude.

Havenar-Daughton also presented an innovative way to follow the development of antibody responses that may help explain this variability and possibly suggest ways to make the induced immune response more reproducible. The approach involves inserting a very fine needle into a lymph node in a vaccinated animal and then aspirating a small sample of germinal center cells(6). Curiously, Havenar-Daughton found a higher number of B cells as well as higher ratio of B cells to helper T cells in the germinal centers from aspirates collected early in the immunization schedule strongly correlated with neutralizing titers observed after the second or third dose of vaccine. This technique should enable investigation of antibody responses almost in real-time, and should therefore help to uncover the underlying mechanisms responsible for the development of antibody responses, their maturation, and the eventual development of neutralization breadth that researchers are hoping to induce through vaccination.

A combination approach

Despite the current focus on native trimers, many researchers feel they represent only one piece of the puzzle that will have to be put together to create a vaccine that can induce bNAbs. To this end, Peter Kwong, senior investigator at the Vaccine Research Center (VRC) at the US National Institute of Allergy and Infectious Diseases, presented results from studies testing a combination of SOSIP trimers with a scaffolded immunogen in a prime-boost regimen. The scaffold is designed to present an eight-residue peptide that corresponds to the N-terminus of the fusion peptide (FP), a conserved region of gp41 targeted by some bNAbs. Kwong and colleagues tested a wide variety of immunization strategies in mice and identified an approach that resulted in neutralizing responses, with some antibodies neutralizing select Tier-2 strains of HIV from subtypes/CRF A, AE, B, BC, and C. Optimal immunization required priming with a trimer, boosts with FP-carrying scaffolds, and an additional boost with a trimer. To improve on this result researchers are optimizing the immunization strategy on two fronts. First, they are testing various versions and combinations of the FP peptides to improve the breadth of the immune responses. Second, they are comparing different scaffolds, some of which appear to be far more immunogenic in mice than the scaffold that was used in this study. The long-term objective of this work is to reproducibly elicit antibodies that neutralize at least 30 percent of the Tier-2 strains of HIV.

Ryan Meyerhoff, an MD/PhD student at Duke University, presented results from Bart Haynes’ group that also pointed to the need to combine trimers with immunogens that focus responses on a particular part of the Env protein. Their work makes use of a synthetic glycopeptide mimicking glycans on the V3 region(7,8) of HIV Env, targeted by such bNAbs as 2G12 and PGT125. Immunizations with a polymeric version of this peptide resulted in neutralizing responses in all four of the immunized macaques. Isolation and analysis of antibodies induced in these animals showed that they targeted the V3 region in a similar manner to V3-targeting bNAbs, by recognizing both the N301 glycan as well as the GDIR amino acid motif, which corresponds to the tip of the V3 loop (amino acids 312-315). When boosted with a SOSIP trimer, these antibody lineages greatly expanded and showed evidence of evolution, presumably to acquire enhanced binding to the native envelope. Further analysis of these responses is underway, as well as efforts to design better boosting immunogens that would select for greater breadth of antibody responses.

Presenting proteins

Meanwhile researchers are also exploring novel ways of presenting these immunogens to enhance the immune response. Paola Martinez-Murillo, a PhD student at Karolinska Institutet, presented results from Gunilla Karlsson Hedestam’s group showing that arranging native-like clade C trimers on liposomes improved the quality of the immune response as observed by the formation of larger germinal centers that contained more CD4+ T follicular helper (Tfh) cells than observed after immunization with soluble trimers. The liposome-arrayed trimers also elicited stronger and more consistent autologous Tier-2 neutralizing antibody responses. By sorting Env-specific memory B cells and isolating a set of monoclonal antibodies, the investigators showed that the autologous Tier-2 neutralization was mediated by antibodies that target the V2 cap region of the trimeric Env spike(9).

Darrell Irvine, a professor at the Massachusetts Institute of Technology, is taking a very different approach by putting SOSIPs into microneedle patcdissolving array fingerhes—arrays of tiny cone-like needles that pierce the outermost layer of the skin and deliver their content intradermally by dissolving at a defined rate (see image, right)(10). His lab has previously shown that a sustained, or even escalating, release of antigen is better at stimulating antibody responses than a single shot. The same was true in a head-to-head comparison of SOSIPs delivered by either microneedles or traditional soluble protein in mice. With the use of microneedle injection there was a quantitatively stronger response, with an increased frequency of Tfh cells in germinal centers and, most importantly, approximately 100-fold higher titer of Env-binding antibodies after immunization. An additional benefit of microneedles is that they are made from non-immunogenic silk protein used in dissolvable sutures. The protein protects SOSIP structures during lyophilization, removing the need for cold chain storage.

These and other studies presented at the meeting illustrate the extensive analyses that native-like trimers are undergoing in animal models, both in search of a better understanding of the specific responses being activated by these immunogens and also for empirical investigation of candidate Env vaccines and immunization regimens that result in higher neutralization titers and wider breadth of response in animals and ultimately in humans.

bNAb development in natural infection

Studies of antibody development in HIV-infected adults have revealed a number of important concepts that inform vaccine development. An important observation is that bNAbs bind to different epitopes on HIV Env proteins, suggesting there are multiple targets for vaccine researchers to pursue. Another is that development of bNAbs usually takes years. And, as a result of continuous exposure to perpetually mutating viruses, bNAbs often have very high somatic hypermutation rates, suggesting that co-evolution of antibody lineages and viruses is dynamic over a period of several years.

To create vaccines that would be widely applicable, it is therefore important to better understand bNAb development in different contexts. Only a few studies have thoroughly characterized the development of bNAb responses over time in infected individuals, so additional studies are needed to find how generalizable these findings are. Identifying shared patterns of bNAb development involving different antibody epitopes, viral subtypes, and ethnic groups would pave the way toward developing a broadly applicable vaccine. As such, the Keystone symposium highlighted a variety of approaches—from the large-scale neutralization breadth analysis of more than 4,500 individuals in the Swiss cohort to the minute analysis of bNAb development in select individuals, as well as the comparison of bNAb development in infants versus adults.

Alexandra Trkola, a professor at the Institute of Medical Virology at the University of Zurich used data from the Swiss 4.5K Screen, a large cross-sectional analysis of HIV neutralization activity in two longitudinal Swiss cohorts that included close to 4,500 individuals, to further evaluate the association between bNAbs and host, virus, and disease characteristics. In a recent study the group identified four factors that contribute to the development of broad responses(11). One host characteristic was identified: individuals with black ethnicity showed greater neutralization breadth. And three factors linked to the virus—higher viral load, longer infection length, and a higher degree of viral diversity—were associated with greater neutralization breadth. They then looked at the binding antibodies using 13 HIV antigens and obtained immune signatures linked with the same four determinants that were found for development of neutralization breadth. Emerging data gathered from studying more than 300 HIV transmission pairs showed that viral characteristics explain up to 15 percent of the variation in development of neutralization breadth between individuals, providing the first delineation of the contribution of host versus viral factors in the development of a broadly neutralizing antibody response.

Additional studies of antibody development in specific individuals can also complement large-scale studies such as that being pursued by Trkola. One example of studying a single HIV-infected volunteer came from Elise Landais, a senior scientist at the International AIDS Vaccine Initiative (IAVI). She presented data from an individual in an IAVI cohort in whom viral diversity, as suggested by the Swiss cohort data, appears to be implicated in the development of neutralization breadth(12). Understanding the ontogeny of bNAb responses in a given subject provides new ideas for vaccine design.

While there is not currently a study comparing the development of bNAbs in different individuals against a common Env epitope, this knowledge gap will likely narrow in the coming years as multiple studies in different cohorts are underway. Nicole Doria-Rose, a staff scientist at the VRC, traced virus-antibody co-evolution in an HIV-infected individual from the US Military’s HIV Research Program’s (MHRP) RV217 cohort and thereby discovered for the first time a pathway toward a membrane proximal external region (MPER)-directed bNAb. Meanwhile, Penny Moore, an associate professor at the National Institute for Communicable Diseases in Johannesburg, South Africa, presented follow-up analyses on the CAP256.VRC26 antibody lineage, which targets the V2 Env region, and is possibly the most extensively studied case of virus/antibody co-evolution(13). Moore and colleagues aimed to understand why some neutralizing antibodies within a lineage become bNAbs, while others retain narrow specificity despite equally high levels of somatic hypermutation. The CAP256.20 antibody has narrow specificity, while the CAP256.27 antibody is broadly neutralizing. Importantly, introducing three mutations found in CAP256.27 into CAP256.20 restored breadth for CAP256.20. The original amino acids at these sites in CAP256.20 allowed the antibody to bind to a virus variant that was common in the CAP256 donor, but was globally rare. This explains the lack of breadth of the CAP256.20 antibody and shows that antibody maturation during infection does not necessarily result in neutralization breadth.

While all these findings come from studies of adult samples, Julie Overbaugh, a member at the Fred Hutchinson Cancer Research Center, summarized findings on infant antibody responses. Previously, her lab has shown that approximately two thirds of HIV-infected infants develop bNAbs very rapidly, within 11 to 24 months after infection(14). At Keystone, Overbaugh and her lab presented detailed analyses they conducted of antibodies from two infants. These studies confirmed previous findings and also led to some unexpected and even puzzling observations. All of the neutralizing antibodies, including one bNAb, isolated from one of the infants were able to bind the founder virus that established the infection in that child but did not neutralize it. The antibodies did, however, neutralize viruses that appeared three months after infection. Similar results were obtained in a second infant, in whom antibodies that neutralized heterologous viruses did not neutralize the transmitted virus of that infant. This suggests that the antibodies that recognize the transmitted virus seem to be distinct from those that are responsible for breadth. This dichotomy between binding and neutralization, as well as viral escape from binding antibodies, has not been well documented in adults, although there are examples of it, such as the one from Penny Moore’s lab in which CAP256 lineage neutralizing antibodies retained their binding to escape viruses that were no longer neutralized by them.

Unlike in adults, in whom neutralization often depends on a single dominant antibody lineage, plasma mapping studies and detailed study of antibodies from one infant suggest that infant responses appeared to be polyclonal and simultaneously target multiple sites of vulnerability on the envelope, thus attaining breadth of coverage. Moreover, studies of the bNAb isolated from an infant showed that in contrast to bNAbs from adults, which usually show an unusually high level (15 percent to 30 percent) of somatic hypermutation, an infant bNAb was only 7 percent mutated, indicating a more rapid path to neutralization breadth(15). More studies are needed but the polyclonal nature of the response and the easier path to neutralization suggest that infants may have a particularly favorable immune environment for bNAb-generating vaccines.

All these studies provide a window to the ontogeny of breadth for humoral responses. With a detailed understanding of the viral and host parameters responsible for a strong, potent, and broad antibody response, researchers can design and develop improved vaccine strategies. In parallel, the field is focusing on optimizing vaccine constituents and regimens to find the best approach to elicit an effective vaccine-induced immune response.


1. Julien J-P, Cupo A, Sok D, Stanfield RL, Lyumkis D, Deller MC, Klasse P-J, Burton DR, Sanders RW, Moore JP, et al. (2013) Crystal structure of a soluble cleaved HIV-1 envelope trimer. Science 342:1477–1483.

2. Feng Y, Tran K, Bale S, Kumar S, Guenaga J, Wilson R, de Val N, Arendt H, DeStefano J, Ward AB, et al. (2016) Thermostability of Well-Ordered HIV Spikes Correlates with the Elicitation of Autologous Tier 2 Neutralizing Antibodies. PLoS Pathog. 12:e1005767.

3. Klasse PJ, LaBranche CC, Ketas TJ, Ozorowski G, Cupo A, Pugach P, Ringe RP, Golabek M, van Gils MJ, Guttman M, et al. (2016) Sequential and Simultaneous Immunization of Rabbits with HIV-1 Envelope Glycoprotein SOSIP.664 Trimers from Clades A, B and C. PLoS Pathog. 12:e1005864.

4. McCoy LE, van Gils MJ, Ozorowski G, Messmer T, Briney B, Voss JE, Kulp DW, Macauley MS, Sok D, Pauthner M, et al. (2016) Holes in the Glycan Shield of the Native HIV Envelope Are a Target of Trimer-Elicited Neutralizing Antibodies. Cell Rep. 16:2327–2338.

5. Derking R, Ozorowski G, Sliepen K, Yasmeen A, Cupo A, Torres JL, Julien J-P, Lee JH, van Montfort T, de Taeye SW, et al. (2015) Comprehensive antigenic map of a cleaved soluble HIV-1 envelope trimer. PLoS Pathog. 11:e1004767.

6. Havenar-Daughton C, Carnathan DG, Torrents de la Peña A, Pauthner M, Briney B, Reiss SM, Wood JS, Kaushik K, van Gils MJ, Rosales SL, et al. (2016) Direct Probing of Germinal Center Responses Reveals Immunological Features and Bottlenecks for Neutralizing Antibody Responses to HIV Env Trimer. Cell Rep. 17:2195–2209.

7. Alam SM, Aussedat B, Vohra Y, Ryan Meyerhoff R, Cale EM, Walkowicz WE, Radakovich NA, Anasti K, Armand L, Parks R, et al. (2017) Mimicry of an HIV broadly neutralizing antibody epitope with a synthetic glycopeptide. Sci. Transl. Med. 9.

8. Bonsignori M, Kreider EF, Fera D, Meyerhoff RR, Bradley T, Wiehe K, Alam SM, Aussedat B, Walkowicz WE, Hwang K-K, et al. (2017) Staged induction of HIV-1 glycan-dependent broadly neutralizing antibodies. Sci. Transl. Med. 9.

9. Martinez-Murillo P, Tran K, Guenaga J, Lindgren G, Àdori M, Feng Y, Phad GE, Bernat NV, Bale S, Ingale J, et al. (2017) Particulate Array of Well-Ordered HIV Clade C Env Trimers Elicits Neutralizing Antibodies that Display a Unique V2 Cap Approach. Immunity 46:804–817.e7.

10. DeMuth PC, Li AV, Abbink P, Liu J, Li H, Stanley KA, Smith KM, Lavine CL, Seaman MS, Kramer JA, et al. (2013) Vaccine delivery with microneedle skin patches in nonhuman primates. Nat. Biotechnol. 31:1082–1085.

11. Rusert P, Kouyos RD, Kadelka C, Ebner H, Schanz M, Huber M, Braun DL, Hozé N, Scherrer A, Magnus C, et al. (2016) Determinants of HIV-1 broadly neutralizing antibody induction. Nat. Med. 22:1260–1267.

12. Landais E, Huang X, Havenar-Daughton C, Murrell B, Price MA, Wickramasinghe L, Ramos A, Bian CB, Simek M, Allen S, et al. (2016) Broadly Neutralizing Antibody Responses in a Large Longitudinal Sub-Saharan HIV Primary Infection Cohort. PLoS Pathog. 12:e1005369.

13. Moore PL, Gray ES, Sheward D, Madiga M, Ranchobe N, Lai Z, Honnen WJ, Nonyane M, Tumba N, Hermanus T, et al. (2011) Potent and broad neutralization of HIV-1 subtype C by plasma antibodies targeting a quaternary epitope including residues in the V2 loop. J. Virol. 85:3128–3141.

14. Goo L, Chohan V, Nduati R, Overbaugh J (2014) Early development of broadly neutralizing antibodies in HIV-1-infected infants. Nat. Med. 20:655–658.

15. Simonich CA, Williams KL, Verkerke HP, Williams JA, Nduati R, Lee KK, Overbaugh J (2016) HIV-1 Neutralizing Antibodies with Limited Hypermutation from an Infant. Cell 166:77–87.