Cape Town Connections

By Michael Dumiak

From her base in Johannesburg, South Africa, Glenda Gray is getting ready for a new round of clinical trials. Kicking off a month or so after the new year, one study is testing the only vaccine regimen to date that is effective in preventing HIV infection. The tests build on promising results she delivered in Cape Town at the end of October during the debut of HIV R4P—the R4P means Research for Prevention—which is billed as the first-ever conference dedicated to every aspect of biomedical HIV prevention research. Gray is not alone in looking for the new year to nurture green shoots brought to bear in Cape Town: as HIV R4P fades in the distance, its themes hint at what’s expected to happen across a number of HIV research fields in coming months.

As executive director of the Perinatal HIV Research Unit in Soweto, Gray has a broad perspective of what’s happening in HIV science and in society. Organizers planned HIV R4P to make a similarly broad statement: while researchers tend to work in narrowly circumscribed fields because of how science is funded, the need for specialized expertise, and the difficulty of the work, emerging results are blurring the divisions between HIV prevention, treatment, and cure efforts.

HIV R4P reflected this trend. Data was presented by scientists experimenting with new microbicides that would employ the same broadly neutralizing antibodies (bNAbs) that serve as a crux of vaccine research and studying the idea of combining partially effective topical microbicides with partially effective vaccines. There was also a strong showing from molecular biologists, who are developing new images of atomic-scale protein crystal structures—and perhaps new ways of using them to develop immunogens—and ongoing discussions about developing improved animal models with which to conduct research and how to apply those results to humans. Carl Dieffenbach, director of the Division of AIDS at the US National Institute of Allergy and Infectious Diseases (NIAID), sees opportunity in HIV R4P’s inclusive spirit. “All the fields now are mixing and mingling based on scientific opportunity. That’s what you would hope would bring the innovation that we all need.”

Activists were also among the 1,300 who attended HIV R4P from Oct. 28-31, and they, along with behavioral and social science researchers, brought their own concerns about how people might—or, if careful attention isn’t paid, might not—in the end use the products resulting from these decades of research.

Cape Town, long a melting pot and trade center with aspects of endurance and elegance, bound to freedom and brutal history alike, seemed like the right place to begin with HIV R4P. In South Africa the HIV epidemic is very real. Two million people globally are infected with HIV every year; the virus still kills a million a year. Two thirds of those deaths are in sub-Saharan Africa. South Africa alone is home to the largest population of people in the world living with HIV. Amid strong calls for a boost in research efforts from African teams, a third of presenters selected by HIV R4P organizers came from the continent, with organizers granting 300 full and partial scholarships to researchers and advocates who otherwise wouldn’t have attended.

Bigger might be better

One of the top priorities in HIV vaccine research remains figuring out and following up on the partial efficacy provided by the vaccine regimen tested in the now landmark RV144 trial in Thailand conducted by the US Army and Thai Health Ministry. This trial showed that a prime-boost combination of two experimental vaccine candidates reduced the risk of HIV infection from clade B and recombinant E/A virus (the types of HIV most commonly circulating in southeast Asia) by a modest 31% among 16,000 volunteers. The vaccine candidates were ALVAC HIV, a live, recombinant, non-replicating canarypox vector encoding clade B gag/pro and clade Eenv, and AIDSVAX B/E, a genetically modified version of HIV gp120 from clade B and E.

In Cape Town Gray reported on her team’s local follow-up study, named HVTN 87, which tested the same regimen as RV144 even though the most common circulating virus in South Africa is clade C. Gray and colleagues tested the vaccine candidates at research centers in Cape Town, Klerksdorp, and Johannesburg. Preliminary results show that the immune responses induced by the vaccine regimen among the South African volunteers are equally expansive to those induced in Thai volunteers—if not more so—even given that the vaccine regimen was not designed using clade C HIV. Researchers worried because prior studies with DNA and replication-defective pox and adenoviral vectors showed larger people—specifically larger women—had weaker immune responses to the vaccine candidate and obesity rates are on the rise in South Africa. The population is also distinctly different, genetically speaking, from Thais. Researchers enrolled 100 South African volunteers: 51 men and 49 women, with 28 of the women and six men either overweight or obese.

During seven months of trial follow-up, immune responses to the vaccine candidates among the South African volunteers were even better than their Thai counterparts—for example, 69.2% of the South Africans had a peak CD4+ T-cell response to a specific HIV protein, 92TH023-Env, versus 50.3% of the Thai volunteers. The non-neutralizing antibody concentrations (a specific type of which is associated with protection against infection in the RV144 trial) following vaccination are also similar to that seen in the RV144 participants, Gray said. So far there are no significant differences in responses between the two studies given body mass, gender, or age, but Gray and colleagues are still conducting a formal statistical comparison of the antibody responses among the volunteers.

It remains to be seen, however, if the good cross-clade immunogenicity observed in HVTN 87 implies an equivalent efficacy. Researchers have long wondered about cross-clade immune responses and whether vaccine candidates need to be strain-specific. More trials are needed for that, and more are coming. By February 2015 Gray expects to begin a Phase I trial in South Africa testing a version of the vaccine candidates tested in the RV144 trial that are based particularly on clade C HIV.

The Phase I HVTN 100 study will evaluate a clade C version of the Thai vaccine, vCP2438, delivered along with a new adjuvant—the Novartis-made squalene adjuvant, MF59, a proprietary compound originally developed for and still used to boost immune responses to the company’s flu vaccine. Researchers hope this adjuvant will boost the potency of the vaccine candidates and the durability of the immune responses they induce. Researchers plan to eventually conduct efficacy trials of this regimen involving as many as 7,000 volunteers.

Just how much better the vaccine candidate will need to be to potentially gain licensure was a recurring question at HIV R4P. The experimental vaccine regimen tested in RV144 was only 31% effective at reducing HIV infection risk. Researchers are hopeful that modifications to the regimen, such as a better adjuvant or additional boosts, might improve the efficacy to the point that it might be considered for licensure. Asked about it several times at HIV R4P, Gray pointed out that government licensure of a vaccine might be within reach at a level as low as 40% to 50% efficacy. “A partially efficacious intervention to prevent HIV acquisition would have public health benefit,” she said during a press conference. Such a vaccine could be combined with other partially effective prevention strategies—such as a microbicide, use of female condoms, or male circumcision—one of the main messages of HIV R4P. Not only does this reflect the urgency of the situation—the kitchen-sink approach over the one-shot solution—it pointed to the field’s mantra that the epidemic will not come to an end with a single ‘home run.’

To that end, combining vaccines with antiretroviral-based microbicides is a new effort that received attention in Cape Town. Robin Shattock, a virologist at Imperial College London, collaborated with French immunovirologist Roger LeGrand and colleagues to test a vaccine candidate combined with a 1% tenofovir gel—an antiretroviral-based microbicide—in three groups of rhesus macaques, all compared to a group of untreated control monkeys.

Pharmaceutical company Novartis provided a nasally-delivered vaccine candidate derived from two HIV proteins—gp140 TV1 (clade C) and SF162 (clade B) that researchers administered to the monkeys along with an adjuvant, R848 (a toll-like receptor 7/8 agonist). This was followed by two booster injections of MF59. Although the vaccine on its own failed to provide protection, when used together with the microbicide the combination provided a higher level of protection than the microbicide alone. “Can we get more out of putting vaccines and microbicides together?” Shattock asked, a question at the heart of HIV R4P.

Shattock’s findings will gain a boost if the tenofovir microbicide gel gains regulatory approval following an ongoing Phase III trial in South Africa expected to produce results next year.

Another combination approach is using the broadly neutralizing monoclonal antibody VRC01, isolated by researchers at the Vaccine Research Center (VRC) at NIAID, in a vaginal microbicide film or ring. Deborah Anderson, a Boston University obstetrics professor and microbiologist, is exploring this concept in collaboration with Kevin Whaley of Mapp Biopharmaceuticals and others. Together they are working with Kentucky Bioprocessing, a company that uses tobacco plants to make genetically-modified lots of human monoclonal antibodies, or “plantibodies.” They expect to start human safety trials this spring. “It’s the first plantibody in a human study in North America,” Anderson says. Mapp is growing pilot lots of VRC01 in tobacco plants, and will test it in combination with an antibody that prevents herpes simplex virus (HSV) infection. “If we prevent HSV, we might prevent HIV. That’s the idea behind the cocktail. You’ll have a lot of different antibodies against different mechanisms that might work better together,” Anderson says. Another goal is to add sperm-directed antibodies to the microbicide cocktail for contraceptive use.

Anderson’s group will test the antibody cocktail in a film substrate for topical use. The idea is to deliver the antibodies using a device that is a combination diaphragm and microbicide provider. The diaphragm contains a ring with holes in it; these holes can contain pods carrying film or other material that would release the antibody-containing microbicide.

Antibody infusions

Researchers also continue to study whether directly injecting bNAbs into people—passive immunization, as it’s called—will be an effective means of HIV prevention or as a treatment for those already infected. Barney Graham, deputy director of the VRC, presented initial results in Cape Town of passive immunization safety trials with VRC01.

Two Phase I studies of passive immunization are currently ongoing, one involving a group of 25 HIV-infected volunteers (VRC 601) and the other a group of 24 uninfected volunteers (VRC 602). Researchers are administering VRC01 at different dosage levels, both intravenously and subcutaneously, ranging from one milligram per kilo to 40 milligrams per kilo in different subgroups. So far there are no serious adverse events after more than 80 doses, Graham says. Early data for five-milligram doses show intravenous delivery produces peak concentrations in the HIV-uninfected group of up to 100 micrograms of the antibody per milliliter blood within a few hours following administration, with similar kinetics at four weeks. The 20-milligram doses produce much higher antibody concentrations, of up to close to 1,000 micrograms per milliliter. At the higher dose, antibody concentrations remain in the body at what Graham calls a “meaningful” level—40 micrograms per milliliter—for a month.

Meanwhile, the team is planning tests administering VRC01 shortly after birth to babies born to HIV-infected mothers to prevent HIV transmission to the child. Monthly antibody injections would continue until the end of the breastfeeding period to prevent subsequent transmission through breast milk. That would be in addition to standard antiretroviral therapy, which is already proven to be up to 95% effective in preventing mother-to-child HIV transmission. The idea is that a long-acting antibody injection could cover the gaps in adherence to antiretroviral therapy. Whatever the method, the goal remains the same: to create long-lasting, low-maintenance, effective ways to stop HIV. “Can an antibody with a particular level of neutralizing activity prevent HIV infection, either in the setting of mother-to-child transmission or in the setting of high-risk adult exposures?” Graham asks.

Graham and his colleagues are also developing other variations of the VRC01 antibody by mutating the antibody’s amino acid structure to make it more potent and longer lasting. One variation, an antibody billed VRC07-523LS, was made by inserting four amino acids in VRC01’s CDR3 loop and deleting a few from the end of its light chain. A CDR loop, or complementary determining region loop, is a structure that a bNAb uses to bind to its target epitopes. “With those minor modifications,” Graham says, “it has quite a bit more potency.” 

There are several bNAbs that are also candidates for passive immunization, all with different viral targets: the bNAbs PG9 and PG16 target the V1/V2 glycan; PGT121 and PGT128 bind to the N332 glycan supersite; 8ANC195, PGT151, and 35O22 all bind to the gp120/gp41 trimer; and 2F5 and 4E10 that target the gp41 membrane-proximal external region. “We have targets in at least six different areas of the glycoprotein,” Graham says.

These targets also avail themselves to other uses. Structural biologists are using atomic-level analysis to build new models to aid in vaccine development. “We’re excited about all these human monoclonal antibodies that help us to define the important structural features of the glycoprotein, and how that might also then lead to active vaccination,” Graham says. The ultimate goal, what Graham refers to as active vaccination, is getting the body to produce these antibodies on its own rather than having to deliver regular infusions. The way to do that is to design a vaccine immunogen that can provoke the immune system to generate such powerful antibodies against HIV.

From structure to immunogen

At HIV R4P Peter Kwong, chief of the VRC’s Structural Biology Section, presented his team’s recently published 3.5-angstrom resolution structural model of the HIV envelope protein (Nature 514, 455, 2014). The published structure, which specifically shows the long-sought pre-fusion closed form of HIV, was derived using a procedure he thinks could emerge as a template for the design of effective vaccines.

The group is applying structural techniques used to produce vaccine candidates against the pediatric respiratory syncytial virus (RSV), which causes severe respiratory tract infections in infants. Kwong and colleagues are also building on the work that Rogier Sanders and John Moore, both of Weill Medical College of Cornell; and Ian Wilson and Andrew Ward, both of The Scripps Research Institute (TSRI), did to produce a soluble trimeric complex called BG505-SOSIP.664, the first immunogenic mimic of the native HIV trimer (see CROI: Progress on Prevention and Cure), a key discovery for vaccine researchers. As TSRI immunologist and the IAVI’s Neutralizing Antibody Consortium Director Dennis Burton said in Barcelona last year, having a stable mimic of the HIV trimer was a holy grail of HIV vaccine research—for more years than anyone likes to remember.

What Kwong and colleagues at the VRC did with RSV was to examine in detail how a particular RSV glycoprotein, RSV-F (for fusion), undergoes a conformational change from its pre-fusion to post-fusion states. The pre-fusion RSV-F trimer changes quite a bit during this process, Kwong says. The VRC team used its understanding of these changes to engineer double cysteine mutations that form disulfides which keep the RSV-F trimer in its vulnerable pre-fusion state. For reasons that are not yet completely clear, humans make impressive neutralizing antibody responses to pre-fusion RSV-F. Kwong and his team were able to inject pre-fusion-stabilized RSV-F trimers into rhesus macaques and elicit very high titers of effective neutralizing antibodies, illustrating how atomic-level structural information can potentially lead to improved vaccine candidates. Human studies of this immunogen are slated to start in the next 10 months or so.

Kwong’s group used neutralizing antibodies to help with stabilizing its RSV-F structural model and to guide their understanding of what happens during the fusion process. The Dutch pharmaceutical company AIMM Therapeutics identified two such neutralizing antibodies, D25 and AM22, and Graham at the VRC isolated another, 5C4. These antibodies target the pre-fusion RSV-F, not the post-fusion version. By placing the D25 antibody in complex with RSV-F, the team knew they had a pre-fusion trimer structure on their hands. They were then able to experiment with nearly 100 mutations, Kwong estimates, which allowed them to fix it in this state before settling on cavity-filling alterations and a disulfide that seems to be most effective in keeping RSV in its pre-fusion state. Ergo the new antigen.

HIV, like RSV, is a fusion engine, changing its properties while binding to its target cell. But whether an approach similar to that used to identify immunogens against RSV will work against a more complex and cagey virus like HIV remains to be seen. “There’s so much evasion that occurs with HIV, so much glycosylation,” Kwong says. But he’s hopeful, and the structural biology team is already racing ahead with a new mutated form of the HIV Env trimer structure. Kwong said he derived a structural working method from the RSV experience. The first step is to characterize—from natural infection—the most frequently elicited, effective neutralizing antibody responses. “If you want to make effective HIV-1 neutralizing antibodies,” he says, “figure out how this happens naturally. I’m not saying natural infection is a precise model to follow, but rather follow the development of the antibodies.”

The next step is to determine (through analyzing atomic-level and crystal structure information gained from assays, existing data, techniques such as measuring single-molecule fluorescence resonance energy transfer responses, and X-ray crystallization) the atomic-level characteristics of the prospective antigen. The third step is to create a matrix of physical properties, design, and structure in order to improve immunogenicity, and, finally, to use that information, according to Kwong, to recreate very specific B-cell ontogenies through vaccination.

Working with Ivelin Georgiev, a US National Institutes of Health (NIH) computational biologist, and John Mascola, the VRC’s director, Kwong analyzed sera from HIV-infected cohorts, characterizing the ability of these sera to neutralize a panel of diverse HIV-1 isolates. This produced what he called ‘neutralization fingerprints’. Kwong and colleagues then mapped these responses to the SOSIP HIV Env trimer in a mature closed state. The group’s recently published trimer structure, crystallized and solved by Marie Pancera, a research fellow at the NIH, is BG505 SOSIP bound by two antibodies, PGT122 and 35O22 (see image, below). The antibodies were used to hold the HIV envelope in its closed shape, and Kwong’s team thinks this shape could be vulnerable. “We could show the most prominent responses,” he says. “These are the ones you want to go after. Those are the ones you find from natural infection.”

Structure of the HIV-1 Env trimer

Ribbon representation of the HIV-1 Env trimer in the pre-fusion closed conformation with gp120 (receptor-binding subunit) shown in orange, gp41 ectodomain (fusion subunit) shown in red, and with dashed lines indicating disordered regions. In this orientation the viral membrane would be located towards the bottom of the page. Image courtesy of Peter D. Kwong, Marie Pancera, and Jonathan Stuckey, National Institutes of Health/National Institute of Allergy and Infectious Diseases (NIAID)/Vaccine Research Center.

 

Because HIV glycoproteins are unusually conformationally flexible, in Kwong’s words, the immunogen is actually the structure of an unliganded, which is to say non-binding, mature envelope trimer. “That’s the eliciting immunogen that you use when you immunize,” Kwong says. “When you immunize, you don’t immunize with the whole antibody-bound complex. You have to immunize with the naked molecule.”

On its own, the immunogenicity of the SOSIP trimers looks good in rabbits but perhaps not so in non-human primates or humans, who have lots of CD4 which could potentially prompt a SOSIP-based immunogen to open from its closed state, making it ineffective. If a SOSIP antigen could be modified to stay in its pre-fusion closed form, however, perhaps that’s a different story.

This is what the VRC team is now focused on. It’s already used the Pancera structure to guide its effort to grow two fully unliganded SOSIP crystals. “With that template we had a way to start analyzing,” Kwong says. “We had the shape of the molecule that’s seen by broadly neutralizing antibodies, and we could then say, aha—that’s the exact atomic-level structure that we now need to fix.” By using the SOSIP trimer containing Moore’s and Wilson’s stabilizing mutations and the new structural models, Kwong is experimenting further with disulfide fixers. Early results for one new mutation—a stabilizing disulfide, resulting in modified SOSIP trimer now called DS 201-433—seem promising.

Updating animal models

Wayne Koff, IAVI’s chief scientific officer, queried Kwong at HIV R4P on how immunogenicity studies might move the VRC’s structural efforts ahead, specifically which animal models would be employed.

This is a subject that has riddled the HIV field for decades. The varying utility of different animal models is an increasingly important topic for researchers both because of recent advances in altering animal genetics to more closely resemble the distinctly human biological environment in which HIV operates—and the fact that it’s difficult to ethically conduct experiments in humans.

Kwong weighed his options. Testing more broadly neutralizing antibodies requires the proteins involved to have long RNA loops, and mice don’t have these. Guinea pigs might be a better possibility.

Oregon Health & Science University immunologist Louis Picker—for some years now—has generated interesting hypotheses using non-human primate (NHP) models. In the last 18 months the slow fine-tuning of creating antiretroviral therapies for monkeys that closely recapitulate the therapy effects in humans has helped his cause. In a parallel but separate effort, Picker’s lab published research last year (Nature 502, 100, 2013) showing that the use of a cytomegalovirus (CMV) vector-based vaccine candidate appeared to cure nine rhesus monkeys of simian immunodeficiency virus (SIV), without antiretroviral treatment, in the context of a prophylactic vaccine given before SIV infection. This finding excited many researchers and is fueling efforts to develop this vector as both a preventive and therapeutic vaccine candidate, and there are many more than nine monkeys now. There may be only one Berlin patient, but there are 40 to 50 Portland monkeys who are now cured, Picker said at HIV R4P, and though it’s a laugh line, Picker is serious about the monkeys.

“I think we should have 500 monkeys on antiretroviral therapy now that we can do experiments on,” Picker said, which could fuel research on understanding how the viral reservoir is established—specifically where the virus goes to hide in sanctuaries—perhaps in the follicles of memory T cells—to escape destruction and enter a latent phase only to reappear again later. Picker’s team showed in NHPs that even after reducing the viral reservoir from three to four logs—a factor ranging from 1,000 to 10,000—interruption of therapy allows the virus to come back. He says this means the reservoir needs to be reduced by five or six logs.

But from some corners the monkey models draw criticism because while SIV is similar to HIV, it is not the same; the same is true with monkeys and humans themselves. “I agree monkeys aren’t humans. The clinical situation is different, but the biology is really similar,” Picker says. “To ask intelligent questions and make sure our resources for clinical trials are used appropriately, it behooves us to invest in making monkey models as biologically relevant as we can.” Picker argues that using monkeys speeds fundamental research. “You need to do things quickly and do experiments that make things worse as well as better,” he says. “You can’t do that in humans.”

Anderson and LeGrand, immunovirology division coordinator at the French Commissariat à l’Energie Atomique et aux Energies Alternatives (CEA), say that improved animal models, especially NHP models, are needed to understand unresolved questions about cell-associated transmission of HIV. “Despite evidence for cell-associated transmission, most infection models used for screening vaccines and microbicides use cell-free HIV viral challenge,” Anderson said in Cape Town. “The failure of HIV vaccines and microbicides to date could be in part due to the failure to address cell-associated HIV transmission.”

Anderson and LeGrand are working with a small group of researchers to address this. In mid-December, theJournal of Infectious Diseases is publishing an 80-page supplement summarizing the group’s work from a 2013 workshop; another supplement will come out next spring as a result of a workshop held in Cape Town at HIV R4P. In the Infectious Diseases supplement Le Grand calls for increased efforts to develop models recapitulating the complexity of natural sexual transmission, including mucosa and mucosal secretions, in order to improve the relevance of animal models for HIV prevention research. “The community needs to make efforts to try to improve the relevance of the models we are using,” he said in Cape Town.

Shattock and Anderson say matching models to questions is the important—and sometimes maddening—thing to consider. “Monkeys work for me for the right questions, but I see them not as a gate-keeper,” Shattock says. “You can do a level of depth in monkeys that you can’t do in humans, but you need to know what you’re modeling is what you’d see in humans.” Anderson says it’s easy for research teams to develop vested interests in existing models, because producing ongoing results is important for funding. “The models you choose to work on kind of frames the answers you’ll get,” she says. “You have to be careful and know the shortcomings of the models you’re using.”

Learning from baby

Deborah Anderson, a Boston University obstetrics professor and microbiologist, might’ve said it best: HIV R4P wasn’t a cure meeting, but cure is the ultimate prevention. Cure research, a booming area of study, was the focus at a satellite session in Cape Town. Much of the talk was on the complex and shadowy subject of the HIV reservoir—the pool of latently HIV-infected cells or virus hideouts in the body that allow viral replication to continue in full force if effective antiretroviral therapy is interrupted. Carl Dieffenbach, director of the Division of AIDS at the US National Institute of Allergy and Infectious Diseases, focused specifically on the case of a Mississippi infant who received antiretroviral therapy beginning 30 hours after birth, starting with treatment even before medical staff had confirmed the baby’s HIV status. After a month, the Mississippi baby had no detectable virus. After two years, the child remained HIV-free, firing hopes that a cure was achieved. Unfortunately, the child’s virus eventually rebounded after discontinuing antiretroviral therapy. For Dieffenbach, the case of the Mississippi baby exemplifies the significance of understanding and quantifying the latent reservoir. “We are still in the stone age when it comes to assays on the reservoir,” she says. “There’s a range of tools. They’re all challenged.” To address this, NIAID recently approved a set of grants to seven labs searching for new and better assays to detect the latent reservoir. More broadly speaking, early treatment for infants is still a tantalizing prospect; researchers want to see if antiretroviral treatment given in the first two days to HIV-infected babies at birth can lead to viral remission, allowing the children to eventually stop treatment for an extended period. Pediatrics professor Yvonne Bryson at the University of California, Los Angeles, will lead such a study, called IMPAACT P1115, which will enroll nearly 500 volunteers. “We’re trying to develop a new cohort of infants treated in a similar way to the Mississippi baby,” she says. It’s a unique set of circumstances however, as the child’s mother was not on antiretrovirals during her pregnancy. Dieffenbach says they’re screening mothers coming into delivery rooms around the world, looking for volunteers who fit the profile. —MD

 

Realpolitik

For microbicides and pre-exposure prophylaxis (the use of antiretrovirals to prevent HIV infection), research is coming closer and closer to providing real-world interventions. At HIV R4P there was discussion about what to do with new products and approaches and how to get them to those who need them most. Helen Rees, virologist and director of the Wits Reproductive Health and HIV Institute, joined South African science minister Naledi Pandor in arguing for social science research to help with this. Gray says if there is a breakthrough, there needs to be follow-through. That’s why another focus of HIV R4P was the social science needed to turn study results into meaningful interventions for those at risk of HIV infection.

Thinking ahead to a time when an HIV vaccine may come, Rees pointed to the country’s vaccine rollout against the sexually transmitted human papillomavirus, which causes genital warts and can lead to cervical cancer, as a potential model for delivery. Even with something as intractable as HIV vaccine research, Rees says, advance planning and an understanding of the environments where it might be used are vital to making a vaccine effective. “Can we introduce vaccine service delivery in schools?” Rees says. “How do we reach nine- to 13-year-olds not targeted for immunization?”

Attitudes toward these products can greatly influence their real-world effectiveness, as Makarere University’s Teopista Nakyanzi pointed out in showing why Ugandan women didn’t join an otherwise promising study on topical HIV prophylaxis. Her studies suggest one main reason women didn’t enroll was the fear of knowing their HIV status; another is that many did not have any financial income and feared losing support from their partners if they were involved in the trial. Ariane van der Straten, an expert in female-initiated HIV prevention at RTI International in San Francisco, says more broadly that there’s potential stigma involved because there is a view that women who use HIV prevention products are promiscuous—which then implies that pre-exposure prophylaxis, for instance, is only appropriate for promiscuous women.

Dieffenbach sees an effective vaccine that could act both prophylactically and therapeutically as the ultimate way to deal with stigma. “If you had a safe and affordable and durable HIV vaccine which also worked therapeutically, you would not have to test people for HIV at the time you gave it. You could vaccinate your whole population,” he says. “Talk about a way of de-stigmatizing.”

Rees cites the World Bank President Jim Kim in trying to bring attention to these concerns. “I am just asking that we bring the same kind of rigorous approach and scientific thinking,” she quotes the former physician and anthropologist, “to actually delivering these tools for health that we bring to creating them."

Michael Dumiak reports on global science, technology, and public health and is based in Berlin.