Down, But Not Out

At the first Keystone meeting on HIV vaccines following the STEP trial, researchers discuss recalibrating research and development priorities

By Kristen Jill Kresge and Simon Noble

Just a few days after many of the leading researchers in the AIDS vaccine field gathered for an HIV Vaccine Summit sponsored by the National Institute of Allergy and Infectious Diseases (NIAID) to discuss US government spending and the research priorities for the field (see Balancing AIDS vaccine research), they reconvened in vastly different environs for the annual Keystone Symposium on HIV Vaccines: Progress and Prospects. This year’s meeting took place from March 26 to April 1 in Banff, Canada, and like the field itself, was much more focused on fundamental immunology and discovery research than the clinical pipeline of vaccine candidates.

Many speakers remarked in some way on the results of the STEP trial and its repercussions. Updates on the ongoing analysis were also scattered throughout the conference. Larry Corey of the University of Washington said during his opening keynote presentation that the STEP trial has “recalibrated” the HIV vaccine field, but he dismissed the notion that nothing positive has come out of it. “We should still have optimism,” he said, making it clear that in his estimation, “T-cell vaccines are not dead,” because there is still no clear explanation for the failure of Merck’s adenovirus serotype 5 (Ad5)-based candidate, known as MRKAd5. “The uncertainty in the mechanism suggests we should not give up on T cell-based vaccines,” added Corey. “The ability to make such vaccines may be more approachable than getting effective neutralizing antibody vaccines.”

Alan Bernstein, executive director of the Global HIV Vaccine Enterprise, said there is currently a lot of “navel gazing” going on in the wake of the STEP trial as researchers try to figure out where to go next. The way forward, for now, seems to include figuring out some basic science questions. Efforts to more fully understand mucosal immunology, the types of T-cell responses a vaccine should induce, the still mostly uncharacterized role of innate immunity in HIV infection, the mechanism of protection for live-attenuated vaccines in nonhuman primates (NHPs), and the mysteries of long-term nonprogressors figured prominently at this meeting and remain clear priorities for the field. Researchers also presented novel strategies for combining microbicides and partially-effective vaccines to try to prevent HIV infection. “There isn’t one way forward or one simple way forward,” said Bernstein. “If anyone says there is they’ve got a crystal ball that I don’t have.”

Side-stepping

Corey presented what he referred to as a “glimmer of data” from the STEP trial. In a very small number of vaccinated individuals who had no pre-existing Ad5 immunity and who mounted high levels of immune responses to HIV Gag, there was evidence of lower viral loads. But he also acknowledged some caveats, including that this was only discovered in a post-hoc exploratory analysis and that the sample size is small. Corey said an understanding of the way immune responses against HIV are related to disease progression “is just beginning to emerge.”

Human leukocyte antigen (HLA) typing for individuals in the STEP trial and epitope mapping of the infecting viruses is still ongoing. Results from these analyses might shed light on the reason the candidate failed to provide any protection, but Corey suggested the “character of the T-cell response with three sequential doses of Ad5 vector may not be optimal.”

Corey also addressed possible explanations for why the HIV incidence increased with a higher frequency of anti-Ad5 antibodies. Susan Buchbinder of the San Francisco Department of Public Health, and a principle investigator on the STEP trial, said there was a two- to three-and-a-half-fold increase in risk of infection in the vaccine group as the Ad5 antibody titers increased. The majority of volunteers in the trial were men who have sex with men and so one possible mechanism is that more Ad5-specific CD4+ T cells were present in the rectal mucosa, creating more targets for HIV, according to Corey. He also said an indirect biological mechanism could be at play and that perhaps MRKAd5 interfered with innate immune responses against HIV.

A Phase I trial conducted by the Collaboration for AIDS Vaccine Discovery (CAVD), an initiative funded by the Bill & Melinda Gates Foundation, was analyzing the effect of MRKAd5 on innate immune responses until it was halted following the results of the STEP trial. Julie McElrath of the Fred Hutchinson Cancer Research Center said that in this study the innate immune responses appeared different in individuals with high pre-existing immunity, but she didn’t specify how. She did suggest results from this prematurely-halted trial might help elucidate the role of Ad5 immunity in increasing susceptibility to HIV.

Danny Casimiro of Merck said that his company has committed support for conducting in-depth analyses of MRKAd5 in NHPs to see if they can duplicate the results of the STEP trial and answer some of these lingering questions. But Corey said conducting a prospective clinical trial, in which biopsies are taken from trial volunteers, is the only way to know if and how this vaccine enhanced susceptibility to HIV infection, and he said he hopes that the US Food and Drug Administration will approve such a trial.

Buchbinder said analyses of other potential confounding factors that may also help explain this observation are still ongoing. These factors include host genetics, sexual networks, clusters of HIV infections at certain sites, and sexual risk behavior; there is some evidence that risk behaviors increased among vaccine recipients during the course of the study, indicating that perhaps these volunteers suspected they had received the vaccine candidate and not placebo, according to Buchbinder.

McElrath also presented data from the ongoing analysis of the immune responses induced by MRKAd5. Only 15% of vaccine recipients who became HIV infected did not develop CD4+ or CD8+ T-cell responses against HIV. While the magnitude of CD4+ T-cell responses was similar in vaccinated individuals regardless of their Ad5 serostatus, the magnitude of CD8+ T-cell responses was lower in vaccinated individuals with pre-existing Ad5 immunity. However, even in these individuals there was no correlation between the magnitude of response and whether they became HIV infected. Most of the CD8+ T cells induced by the vaccine secreted IFN-γ and TNF-α, and McElrath and colleagues are now analyzing the proliferative capacity of these T cells.

In the meantime, a Phase I clinical trial with another adenovirus vector, based on adenovirus serotype 26 (Ad26), was recently initiated by Dan Barouch and colleagues at the Beth Israel Deaconess Medical Center. This is the first time an Ad26-based vaccine candidate is being analyzed in human volunteers and Barouch said it shows that regulatory authorities are comfortable allowing trials with other Ad vectors to proceed.

In NHP studies Barouch found that immunizing macaques with an Ad26 vector-based candidate followed by an Ad5 candidate afforded a 1.4 log decrease in peak viral loads and a statistically significant improved survival ratio after simian immunodeficiency virus (SIV)mac251 challenge. This Ad26 vector “outperforms Ad5 vectors in rhesus macaques,” said Barouch.

Pushing forward

One of the burning questions in the field since the MRKAd5 candidate showed no efficacy in the STEP trial has been how to determine which vaccine candidates should be advanced to clinical, and particularly efficacy, trials. Gary Nabel, director of the Vaccine Research Center (VRC) at NIAID gave a presentation on the ‘Criteria for advancement of novel vaccine candidates.’ He said the highest priority now is to develop vaccines that induce neutralizing antibodies (Nab) against HIV, not only for direct neutralization but also antibody-dependent cell-mediated cytotoxicity (ADCC), induction of complement, or other innate immune responses to virus-infected cells. For these vaccines, Nabel said, the advancement criteria are more straightforward.

But what comes next for an optimal T-cell vaccine is trickier. Nabel said the field needs to identify the characteristics of a protective T-cell response, increase the breadth of coverage against natural HIV isolates, enhance the magnitude of the relevant responses, and address the potential mechanisms of enhancement of susceptibility to HIV infection seen in the STEP trial.

To the first point, Nabel said defining the immune correlates of protection in human and NHP studies was a priority, and that this was the best rationale for going ahead with clinical testing of the VRC’s DNA and Ad5 candidates in the planned PAVE 100 trial since this heterologous prime-boost approach induces immune responses qualitatively different to those induced by MRKAd5.

Regarding the breadth of the immune response, Nabel discussed at length the informatics approach to genetic diversity that Bette Korber of Los Alamos National Laboratory and colleagues are pursuing. Her group is using computational optimization methods to design polyvalent vaccine antigens from sets of mosaic proteins that maximize the representation of potential T-cell epitopes for a viral population (Nature Medicine 13, 100, 2007). The optimization offers the advantage that, at least in computational models, mosaic coverage exceeds any equivalent-sized natural sequence cocktail, and the immune evaluation of the mosaic concept in mice has indicated that vaccine-induced CD8+ T-cell responses to different antigen designs is much broader using mosaic antigens.

A leaky gut

One of the distinguishing features of HIV infection is the chronic immune activation that occurs, leading to immunodeficiency and eventually progression to AIDS. This is in marked contrast to the nonpathogenic immunodeficiency virus infection in some NHP species, notably sooty mangabeys and African green monkeys, which have negligible immune activation and few detectable consequences of SIV infection.

In contrast, immune activation in HIV infection has many deleterious consequences; it results in a higher frequency of activated T cells that are the preferred targets for HIV, leading to the selective destruction of CD4+ T cells and the maintenance of virus replication. Daniel Douek of the VRC and his group have been looking into why HIV disease is progressive and what drives the immune activation in chronic HIV infection that is the strongest predictor of disease progression. “The extent of CD4+ T-cell loss in the acute phase is not the sole determinant of progression rate in the chronic phase,” said Douek.

Previous studies from as early as the mid-1980s documented enteropathy in HIV-infected individuals and observed that gut permeability can be increased up to 10-fold. These and other findings led Douek to postulate that microbial translocation—the transfer of microbes and/or microbial products across the gastrointestinal barrier and into systemic circulation without overt bacteremia—across the compromised gut mucosal surface contributes to immune activation. “You have more bacterial cells in your gut than you have human cells in your body,” said Douek.

As a marker of microbial translocation, Douek’s group measured the amount of lipopolysaccharide (LPS)—an immunostimulatory signature component of Gram negative bacterial cell walls—in the plasma of HIV-infected and uninfected individuals, and found levels of plasma LPS correlate with HIV disease progression (Nature Medicine12, 1365, 2006). As a more direct measure of microbial translocation, they have also measured the amount of 16S ribosomal RNA (seen only in bacteria) in plasma and again found that significantly higher levels are seen in HIV-infected individuals as compared to uninfected controls.

CD14+ monocytes and macrophages are known to secrete soluble CD14 (sCD14) and proinflammatory cytokines after LPS stimulation. Douek’s group has demonstrated that the elevated levels of LPS in HIV-infected individuals correlates with significantly higher levels of plasma sCD14, suggesting chronic stimulation of monocytes in vivo, as well as a number of other measures of innate and adaptive immune activation. They have also documented differences in microbial translocation between progressive and non-progressive HIV infection in humans, seeing less elevated levels of plasma LPS and sCD14 in long-term nonprogressors or elite controllers, as compared to chronic progressors. Moreover, plasma LPS levels decreased after initiation of highly-active antiretroviral therapy (HAART) in HIV-infected individuals, suggesting that reduction in viral load might allow immunological and structural reconstitution of the gastrointestinal barrier. The reduced LPS levels in these individuals after starting HAART also inversely correlated with the number of blood CD4+ T cells.

Douek and his group have also compared natural SIV infection in sooty mangabeys—which is nonpathogenic, with low immune activation despite a high viral load—to pathogenic SIV infection in rhesus macaques. They saw no evidence for microbial translocation in the sooty mangabeys, whereas SIV-infected rhesus macaques had elevated plasma LPS levels compared to uninfected macaques.

Taken together, this evidence indicates that the immune activation seen in HIV disease is at least partly due to insult at the gut mucosa, which enables microbial flora to cross this barrier, and further reinforces the idea that HIV is primarily a disease of the gut mucosa. “We’ve got a leaky gut,” said Douek.

At the front lines

HIV is not unique in causing the most substantial damage within the first days and weeks of infection; all lentivirus infections begin with a fast and furious phase of infection, followed by a slowly progressive stage, said Ashley Haase of the University of Minnesota. It is the very earliest stages of HIV infection, within 10 days of transmission, which is the subject of his research. His laboratory is studying early events in transmission of SIV because acute HIV infection is discovered too late to study the earliest interactions between the virus and the immune system. Studying SIV transmission and acute infection offers many clues about critical immunologic time points at which a vaccine could act.

In the female genital tract, which Haase called the “active front line of the mucosal immune system,” the naturally-occurring cellular immune responses against HIV are actually quite robust (see The great barrier). Four days after rhesus macaques are challenged mucosally with a high dose of SIV, only concentrated foci of virus-infected cells can be found after exhaustive examination of several layers of cervical tissues, illustrating the success of the mucosal barrier. Haase said that a vaccine capable of focusing and augmenting these natural cellular immune responses could prevent establishment of an HIV infection. It could also provide an opportunity for microbicides or early antiretroviral treatment to limit this founder population of HIV-infected cells even further, said Haase.

Part of the reason for the small initial founder population is that there are limited target cells for HIV in vaginal mucosal tissues. Most of the target cells for the virus are on the other side of the mucosal barrier, said Tom Hope of Northwestern University. To overcome this, Haase speculated that the virus induces an influx of plasmacytoid dendritic cells (pDCs) through a mucosal outside-in signaling mechanism. An influx of inflammatory cells creates a favorable environment for expansion of the infection. The pDCs secrete chemokines that stimulate the expression of the CCR5 receptor on mucosal cells, providing “fuel” for HIV, Haase said. This allows the limited founder population of HIV-infected cells to expand massively and, once this occurs, the virus is then “broadcast” distally to other sites. HIV then quickly establishes an infection in lymphoid tissues and in the epithelial cells of the intestine, where it wreaks havoc on the critical cells of the immune system.

But before this happens microbicides and/or vaccines working in concert could subjugate the virus. Haase said a microbicide could potentially interfere with this proposed signaling mechanism, and a vaccine that induces a potent T-cell response at mucosal sites could help snuff out the virus before a systemic infection occurs.

Recent studies with live-attenuated SIV vaccines, cited by Haase, show a T-cell response could be enough to stop infection, especially if these responses are present at mucosal sites. Vaccinated rhesus macaques that are exposed to a high-dose mucosal SIV challenge develop a “very robust” CD8+ T-cell response in cervical vaginal tissues within three weeks after challenge, said Haase. There is also a correlation between this CD8+ T-cell response and the reduction in the size of the founder population of virus-infected cells. If a microbicide could help reduce the founder population even further, it would increase the likelihood of success for a vaccine.

A study by Meritxell Genesca of the University of California in Davis corroborated the idea that mucosally-available CD8+ T cells may play a role in protection. Genesca presented results from a study in which 12 rhesus macaques immunized with a hybrid simian-human immunodeficiency virus (SHIV)89.6 live-attenuated vaccine were challenged intravaginally with SIVmac239. The majority (60%) of vaccinated macaques controlled viral replication. However when the CD8+ T cells were completely depleted, the partial protection afforded by the live-attenuated vaccine was eliminated, leading Genesca to conclude that this protection is at least in part mediated by CD8+ T-cell responses. Gag-specific T-cell responses were present on the day of challenge at the site of inoculation and, after challenge, vaccinated macaques have more SIV-specific T cells secreting multiple cytokines and chemokines in their vaginal tissues.

In an impromptu talk, John Moore of Cornell University also presented a rationale for combining vaccines and microbicides in an effort to bolster immune defenses against HIV. Moore suggested that a vaginal microbicide that could deliver monoclonal antibodies against HIV might be a substitute for the mucosally available broadly-neutralizing antibodies that current AIDS vaccine candidates can not, as of yet, induce. “Science suggests these fields could actually work together,” said Moore.

A microbicide that does not block infection would be considered a failure, but if it could lower the levels of viral inoculum in the genital mucosa and therefore increase the likelihood that a vaccine would be effective, it would be a great success, added Moore. To test this hypothesis he is preparing to conduct a combination microbicide/vaccine study in 20 rhesus macaques in collaboration with Barouch. This study will evaluate the efficacy of the vaginal microbicide T-1249—a gel formulation of an entry inhibitor developed by Roche Pharmaceuticals but never licensed for the treatment of HIV infection—in combination with a prime-boost vaccine regimen using an Ad26 vector-based vaccine followed by a chimeric Ad5/Ad48 vector-based candidate, both developed by Barouch, against a vaginal SIVmac251 challenge.

Quantity vs. quality

Rafi Ahmed’s laboratory at Emory University is looking at various vaccine vectors to see if they induce T-cell responses of differing qualities. His studies suggest that the most important marker of T-cell function is their proliferative capacity. In all other disease models, “expansion potential in vivo is the function that gives the best protection,” he said. In a study in mice, Ahmed compared an Ad5 vector expressing the full glycoprotein of lymphocytic choriomeningitis virus (LCMV) with a Listeria monocytogenes (LM) vector expressing the gp33 epitope of the LCMV glycoprotein. The control group included mice infected with LCMV, which is rapidly cleared and induces a long-lasting memory T-cell response in the animals.

T-cell responses were analyzed in the three groups of mice at various time points over a period of 60 days. In the mice challenged with the LM vector, the CD8+ T-cell responses peaked after eight days at about 8-9% of cells measured in peripheral blood, and leveled off at a markedly lower 2%. Comparatively, the Ad5 vector was highly immunogenic. In the Ad5 mice the levels of CD8+ T cells hovered around 8%. “Things look extremely good in terms of numbers,” said Ahmed. “Levels were as high as we’ve seen with LCMV and we did not see much of a contraction.”

When memory responses were analyzed in tissues, the quantity of Ad5-induced cells was also much higher—6.3% of cells in the spleen recognized gp33 in the mice given Ad5, compared to 2.5% in the control group. The absolute number of CD8+ T cells was also higher in the mice that received Ad5 both in the lungs and liver. “It’s hard to trump LCMV in numbers,” said Ahmed, yet Ad5 did.

But when the Ad5-induced memory cells in peripheral blood were analyzed for CD127 and interleukin (IL)-7 expression—markers of healthy memory cells—the levels of CD127 declined and did not recover, as they did in the other two groups. And when researchers honed in on the quality of these responses at other sites, they found that fewer of the Ad5-induced CD8+ T-cell responses were secreting interferon (IFN)-γ or IL-2.

To evaluate the recall potential—the proliferative capacity of memory T cells upon re-exposure to antigen—of the Ad5-induced CD8+ T cells, memory cells from the spleen or liver of all three groups of mice were transferred into naïve mice, which were subsequently challenged with vaccinia virus expressing gp33. This produced the most striking data, according to Ahmed. In response to viral challenge there was a dramatically lower expansion of the Ad5 memory T cells in the spleen, as compared to both the LM or control mice. When liver cells were transferred from the Ad5 group of mice, they did not proliferate at all. “Even though there was a very high frequency in the liver, when you look at their capacity to proliferate, they are very much compromised,” said Ahmed. Even at varying doses of Ad5, the cytokine expression profile and proliferative capacity, in both lymphoid and non-lymphoid compartments, of the CD8+ T cells were impaired, as compared to LCMV and LM. This led Ahmed to conclude that, at least in mouse experiments, the T-cell phenotype induced by the Ad5 vector “is not matching what a good memory T-cell phenotype should look like.” This raises questions about the value of the T-cell responses induced by the Ad5 vector in humans. As to how these findings relate to the human system, Ahmed said jokingly, “I’ll leave that to the experts to figure out.”

But in NHP studies conducted by Louis Picker of the Oregon Health & Science University, optimal protection from a live attenuated SIV vaccine does not seem to depend on the recall potential of SIV-specific CD4+ or CD8+ T cells. He suggests other possibilities for protection, including the qualitative aspects of the T-cell response, still unidentified non-classically neutralizing antibodies, or innate immune responses such as natural killer cells. Picker is currently utilizing a genomic approach to try to elucidate the mechanism for protection from live-attenuated SIV vaccines. He emphasized that intrinsic “innate” mechanisms may play a key role and he is hopeful that a large study involving 120 rhesus macaques—a collaborative experiment with IAVI and the CAVD—will be able to determine just what is responsible for the partial protection mediated by live-attenuated SIV vaccines.

Andrew McMichael is looking at human T-cell immune escape in very early acute HIV infection using samples from plasma donors in the US, some of whom donate as much as once or twice a week. If one of these donors becomes infected with HIV, researchers have a series of samples that can be analyzed for extremely early time points in their infection. McMichael reported that in several individuals there is evidence of HIV escaping T-cell immune responses within 16 days. At the same time there is a rapid loss in the quantity of CD4+ T cells, suggesting that early cytotoxic T lymphocyte (CTL) responses quickly decay once virus mutation occurs. This led McMichael to conclude that many of the CTL responses may be of limited utility. “It’s a lot like the neutralizing antibody problem,” he said. “A lot of the T-cell responses are probably useless.”

Taking control

Bruce Walker of Massachusetts General Hospital is also studying the role of T-cell responses, among other factors, in HIV-infected individuals who are dubbed elite controllers because they effectively control viral replication without the aid of antiretroviral therapy. Most elite controllers have some measurable level of ongoing viral replication by ultra-sensitive assays. In some, the viral load is as low as 4 copies/ml of blood. “Do we have a model for a successful vaccination in people who control the virus at those levels?” asked Walker.

He is seeking an explanation for this control by studying the T-cell and neutralizing antibody responses in these individuals, as well as the characteristics of the infecting virus and the genetic makeup of the individuals. So far, neither T cells nor neutralizing antibodies seem to be the key to control. In the elite controllers Walker is studying, in cooperation with Steve Deeks at the University of California in San Francisco, the magnitude of T-cell immune responses are actually lower than those seen in chronic progressors. These individuals can control virus replication without any remarkable T-cell responses, at least by the way they are currently being measured, said Walker. The same is true for neutralizing antibodies. “Neutralizing antibody responses are much, much weaker in these individuals.” He said some of the elite controllers would not even test positive for HIV infection by Western blot methods because their HIV-specific antibodies target so few epitopes. Uniquely, the T-cell responses in elite and viremic controllers (individuals who control the virus but at higher levels of viral replication than elite controllers) primarily target HIV’s Gag protein.

So far there does not appear to be an explanation in studying the properties of the infecting virus either—no genetic defects in the viruses of elite controllers have been identified that contribute to decreased replication capacity or reduced viral fitness. Only three out of a group of 63 elite controllers have any deletions in HIV’s nefgene, which are known to be associated with weakened viral replication capacity. Some viral polymorphisms have been identified in the HIV Gag protein that are strongly associated with the elite controller phenotype, but when a virus is constructed in the laboratory with these same mutations, researchers observe no effect on viral replication capacity. However, viruses isolated from some elite controllers do seem to have a lower replicative capacity, suggesting something else may be contributing to their reduced fitness.

Genome-wide association scans, using a single nucleotide polymorphism (SNP) platform, are ongoing in Walker’s elite controller cohort (see HIV Controllers: Can the Human Genome Project advance AIDS vaccine development?, IAVI Report, May-June 2007). More than 600 elite controllers have been recruited and so far three SNP associations have been identified in these individuals, said Walker. For now, “the basis of elite control remains undefined from a genomic standpoint,” he added.

Studies by April Ferre of the University of California in Davis, suggest that strong, polyfunctional CD8+ T-cell responses at mucosal sites might contribute to the control of viral replication in elite controllers. In a cohort of 28 elite controllers, she reported that these individuals have stronger, more polyfunctional (as defined by secretion of INF-γ, TNF-α, IL-2, and MIP-1β) mucosal immune responses, as compared to HIV-infected individuals on HAART. Elite controllers also have more CD4+ T cells and more polyfunctional CD8+ T cells in the rectal mucosal tissues than either chronic progressors or individuals on HAART.

Regulating responses

Regulatory T cells (Tregs), which were given a bad name in the past, have had their immunological status restored in recent years, including within HIV pathogenesis studies (see Balancing Act, IAVI Report, July-August 2007). Tregs are a specialized subset of CD4+ T cells that are still somewhat uncharacterized, but their signature markers include constitutive expression of CD25, low levels of CD127, and, most specifically, Foxp3.

Claire Chougnet and colleagues at Cincinnati Children’s Hospital Medical Center are investigating the role of Tregs in HIV infection. Chronic progressive HIV infection is usually characterized by weak HIV-specific T-cell responses and a multitude of studies published in the past 2-3 years have indicated that Tregs could play some part in this ineffective immune control. For instance, depletion of CD4+CD25+ T cells in vitro during chronic HIV infection has been shown to increase HIV-specific T-cell responses, chronic progressive HIV disease correlates with increased numbers of Tregs in lymphoid tissues compared to non-progressors or uninfected individuals, and Foxp3 levels correlate closely with HIV and SIV viral loads during infection of humans and macaques, respectively.

However, Chougnet’s group now has in vitro data indicating that Tregs can also provide some benefit by inhibiting HIV replication in effector T cells in a dose-dependent manner. Whether these data have in vivo relevance must now be determined but Chougnet thinks Tregs may be “double-edged swords” in HIV infection. She also noted that almost all of the published studies describe chronic infection and there is clear variation in the function and dynamics of Tregs depending on the model studied; therefore further studies are required to elucidate the precise role of Tregs during HIV infection.

Enterprising strategy

Bernstein gave a special lecture entitled ‘HIV Vaccines: Progress and Prospects’ which he said represented “preliminary, and therefore mutable, thoughts” about the state of the field. He began by listing what he called structural challenges, including insufficient scale of some projects to solve the major scientific problems, the need for more coordination and information sharing, challenges in manufacturing and clinical trials capacity, and the need for new approaches and innovation. He emphasized that the Enterprise is an alliance of organizations that would serve as a “convener and honest broker.”

Bernstein gave a brief overview of how the Enterprise is, and will be, structured. The secretariat that he heads is now located in New York City, and that body will work with three other arms: the Enterprise Council and an associated board of directors, a scientific stewardship committee, and stakeholder assemblies. In addition, working groups will be convened to share ideas and advise; Bernstein has already asked Bob Seder of NIAID and Rafick Sekaly of the University of Montreal to organize the first of these regarding the immune correlates of protection, a key question that he was surprised the field doesn’t have a better handle on.

The original scientific strategic plan that the Enterprise coordinated for the HIV vaccine field was published in 2005 and one of Bernstein’s first priorities will be to ask whether it should be amended, or if a new plan is required in light of progress in the intervening years. He also said that “we need to reorient our thinking from product development to interrogations of the human immune response to HIV,” and that the field was pinning its hopes on a home run, which doesn’t often happen. He also asked if there was a need to establish transparent criteria to decide on which candidates should progress to efficacy trials.

Bernstein finished with some parallels from his previous career in cancer research. Much like cancer therapy, he said, there is no clear path to an HIV vaccine. Childhood leukemia used to be fatal in the majority of cases but is now treatable about 80% of the time, and yet we still don’t have a clear idea of precisely how the therapy works. Improvements in this field have come from incremental, empirical advances over the past 50 years or so. In contrast, some of the new anti-cancer drugs like Gleevec and Herceptin, which are also extremely effective, were developed only after the culmination of 40 years of fundamental research into the molecular mechanisms of cancer and combinatorial chemistry, as well as successful clinical trials. The lesson, Bernstein said, is that a multitude of approaches are needed.