Hindsight versus Foresight
By Richard Jefferys
While still analyzing data from the third ever AIDS vaccine efficacy trial, the field is serenaded by a chorus of opinions.
On March 25, the National Institute of Allergy and Infectious Diseases (NIAID) convened a special one-day summit in Bethesda, Maryland to solicit input on three major issues in AIDS vaccine research:
- vaccine-related basic research, discovery, and development
- animal model development and utilization
- clinical research and trials
On the first topic, there was widespread agreement that the generation of effective antibodies remains a key goal, and a better understanding of whether nonneutralizing antibodies can aid in protection is urgently needed. (Some scientists have suggested that just binding to HIV can be a useful property for antibodies, as it may facilitate elimination via a mechanism called antibodydependent cellular cytotoxicity.)
Unfortunately, however, some panelists could not resist resorting to hyperbole. Cellular immunologist Rafi Ahmed from Emory University claimed that he “cringed” when hearing that Merck was developing a T cell–based vaccine, even though he is on the immunology advisory committee of the HIV Vaccine Trials Network and does not appear to have voiced this feeling at any time during the trials development. Ahmed has also previously expressed optimism that T cells could mediate control of viral replication in the absence of antibody, because he has documented this in his own studies of the murine virus LCMV. In arguing at the summit that antibodies—even nonneutralizing antibodies—are likely to be critical, Ahmed may have been alluding to the DNA/MVA vaccine developed by his colleague Harriet Robinson at Emory; Robinson reports that her vaccine candidate induces binding antibodies, and claims (based on evidence from SHIV89.6P challenge studies) that these antibodies will increase the likelihood of efficacy.
Discussion of T-cell responses revolved around the likely inadequacies of current assays, which may not measure the most important functions (e.g., for CD8 T cells, the ability to recognize and kill HIV-infected cells). Lab assays typically test responses to small protein slices from HIV called peptides, and in the lab dish the high concentrations of these peptides may be far easier for T cells to recognize than they are when they’re being presented (by infected cells or antigen-presenting cells) in the body. The ability of a T cell to proliferate was frequently cited as critical. Notably to date, there has been just one poster presentation regarding the proliferative capacity of the HIV-specific T cells induced by the Merck vaccine (the results were not impressive). Several participants raised the issue of breadth; the researcher David Watkins pointed out that macaque monkeys typically respond to multiple epitopes in the Gag protein while recipients of the Merck vaccine displayed an average response to just one epitope from each of the three proteins included in the vaccine (Gag, Pol and Nef). Responses to multiple epitopes in Gag were also mentioned as a correlate of lower viral load infected people.
The section on animal models and utilization lamented the lack of resources and also reviewed the shortcomings of the SHIV89.6P challenge virus. SHIV89.6P is a lab-created hybrid comprised of the internal components of SIV, HIV’s simian relative, and HIV’s outer envelope. It was originally created to facilitate studies of antibody-based vaccines. However, because it was shown to consistently cause a very rapid and immediate crash in macaque CD4 T cell counts (unlike SIV, which typically has a variable and inconsistent impact on CD4 T-cell counts), it was also used in tests of T-cell vaccines, including Merck’s. The reason was that the consistency of CD4 T cell loss meant that, if a vaccine preserved CD4 T-cell counts in immunized animals, statistically significant differences between vaccine and placebo groups could be captured in studies involving only small groups of animals. With macaques in short supply for research, this was not a trivial issue. But it eventually became clear that SHIV89.6P was not useful for evaluating T cell–based vaccines: a number of vaccines known to offer no protection against other viruses showed some efficacy against SHIV89.6P, and long-term control—a key goal for T cell approaches—was shown to also be associated with the presence neutralizing antibody responses. Merck conducted a follow-up study using an SIV challenge of macaques, and in this case adenovirus alone had no effect while a DNA vaccine followed by adenovirus immunization had a significant effect, but only in macaques with particular immune response genes known to be beneficial in controlling SIV.
The general view among the summit panelists was that the stringent SIV/ macaque model remains the best method for preclinical evaluation of HIV vaccine candidates, as there is no doubt that it predicted the STEP trial outcome better than any other animal study. Some participants noted that SHIV viruses utilizing HIV envelopes that target the CCR5 coreceptor appear to more closely mimic HIV’s impact on the immune system than the X4-using SHIV89.P, and suggested that these viruses may deserve additional evaluation. Other issues raised included the importance of analyzing the ability of vaccine candidates to protect against SIV containing proteins that are different from those used in the vaccine (this has rarely been done to date) and also refining models that attempt to mimic the acquisition of HIV infection (e.g., by using repeated, low-dose mucosal challenges with SIV instead of a single high-dose challenge).
In the last section on clinical research, there was general agreement that human trials remain important but should be viewed as part of the “discovery” process. A number of participants argued for small, preliminary screening trials that might provide information as to whether a candidate had the potential for efficacy. The International AIDS Vaccine Initiative describes this idea as “screening test of concept,” or STOC. These trials would be conducted in very high-risk populations. One potential downside to this approach is that in such studies there is a much greater likelihood of participants being exposed to HIV while still receiving immunizations; in lowerrisk settings, exposure would be more likely to occur after a vaccine-induced immune response has had time to develop and mature, and this difference could conceivably be important.
South African researcher Glenda Gray made an eloquent plea for continued clinical research, stressing that the incremental progress that results from analyzing data on unsuccessful candidates is critical for moving the field forward. Gray also argued that in the high-incidence area in which she works, research participation is not so much based on expecting instant success but on the hope of contributing information that will ultimately lead to success. Gray’s comments in this regard were echoed from the floor by summit attendees from Uganda.
One of the other members of the final panel, immunologist Mark Connors from NIAID, offered some intriguing comments regarding ongoing analyses of the STEP data. Connors has studied HIV-specific T-cell responses and authored many skeptical articles questioning some of the correlations with control of HIV replication that have been reported in the literature. In 2002, however, he identified HIV-specific CD8 T-cell proliferation as a potentially important correlate, a finding several other research groups have now confirmed. In his comments at the summit, he cited additional assays his lab is working on, and expressed confidence that robust and broadly applicable correlates of immunological control are within striking distance. Furthermore, he cited the fact that several Merck vaccine recipients who became infected in the STEP trial and carry the favorable immune response gene HLA B*57 are controlling their viral loads at undetectable levels. The implication is that while the Merck vaccine was far from optimal, it may have been able to enhance the HLA B*57 effect.
Connors’s comments, like several made from the floor by STEP investigators such as Juliana McElrath, highlighted the slightly bizarre nature of listening to a day’s worth of opinions on the future of HIV vaccine research when—for only the second time in the history of the epidemic—there is a trove of actual data from an HIV vaccine efficacy trial that may shed more substantive light on the issues at hand. The HIV vaccine field has long been beleaguered by an excess of opinions and a dearth of data, and it is unfortunate that the summit—while clearly well-intentioned—ended up continuing this trend.
The specter that stalked the entire event was diminishing NIH funding for investigator-initiated grants. This severe problem is the result of the flat funding of NIH by the Bush administration and Congress over the past several years, leading to a net decrease because of inflation. It was clear that many people at the summit were hoping that more money would be shifted into the investigator-initiated grant pool, either from clinical vaccine research or somewhere else (such as the federally funded AIDS Clinical Trials Group). At least one person questioned the wisdom of the multimillion-dollar, multiyear NIAID grant to the Center for HIV/AIDS Vaccine Immunology (CHAVI), posing the question, “Would four CHAVIs be better than one?” The implication being that a large mechanism under a single investigator (in this case, Barton Haynes from Duke University) may not be the best way to foster innovation in AIDS vaccine research. This issue of where additional funding might come from—certainly the biggest concern for many summit attendees—was rarely commented on directly (Martin Delaney from Project Inform was one of the few who mentioned it explicitly) and certainly wasn’t resolved by the end of the event.
In terms of buzzwords of the day, innovation was certainly one of them, although how exactly to define or encourage it was not resolved. The importance of enticing young investigators into the field was also emphasized, although no one seemed to contemplate what a young investigator might make of the summit itself, and what it said about the HIV vaccine field’s ability to react rationally to the failure of a single AIDS vaccine candidate. Finally, the most repeated quotation was surely “We’re all in this together,” a friendly platitude but perhaps inadequate in terms of balancing the worries of some scientists about basic research grant funding with those of participants focused more on HIV decimating their immediate communities.
The immediate results of the summit are two requests for advice from NIAID on the creation of a new grant funding pool aimed to foster innovation in biomedical prevention science, including vaccine research, and “to solicit input and ideas on priorities in basic vaccine discovery research.” Responses to these announcements will undoubtedly shape the direction HIV vaccine research takes from here.
There was much more discussion at the summit than any one article can capture, and the entire day can be viewed via webcast online: www.macrovolt.com/ live/dgi_032508/.
Also hanging in the balance as a result of the STEP trial is another phase IIB efficacy trial, dubbed PAVE100. This trial also involves an adenovirus vector, given once after a series of three DNA vaccine shots. Originally designed as an 8,500-person study involving multiple populations internationally, there is now a proposal for a smaller trial (PAVE100A) enrolling 2,400 men who have sex with men. This trial will be discussed at a meeting of NIAID’s AIDS Vaccine Research Subcommittee. TAG has serious concerns about whether this trial should be conducted and will release a detailed position statement prior to the meeting.