On December 12th, NIAID's advisory body on AIDS vaccine research - the AIDS Vaccine Research Subcommittee (AVRS), formerly known as the AIDS Vaccine Research Working Group - met to discuss the results from the Merck HIV vaccine trial and mull the implications for NIAID's next planned HIV vaccine efficacy trial, dubbed PAVE 100. The two issues are inextricably intertwined because PAVE 100 also involves the use of an adenovirus serotype 5 (Ad5) vector, in this case designed by researchers at the NIH's Vaccine Research Center in collaboration with a company called GenVec.
Only two media outlets appear to have covered the AVRS meeting so far; Bob Roehr wrote an excellent, detailed article for Medscape (free registration required) and a much shorter piece appeared online in The Scientist (free registration also required). Based on this coverage, it appears that the VRC is arguing that their approach is sufficiently different from Merck's that PAVE 100 should go ahead.
The VRC's vaccine regimen starts with a series of three DNA vaccine immunizations (a month apart) and involves a single shot of the Ad5 vector at the six month timepoint after study entry (the Merck trial used three Ad5 shots at months 0, 1 & 6). The VRC's vaccines also include additional HIV antigens (Env proteins from HIV-1 clade A, B & C) and the Ad5 vector has more Ad5 genes deleted from it than Merck's (this is intended to reduce the expression of Ad5 proteins and thus focus the immune response more against the HIV proteins than the vector itself). While these differences may well be important, they do not rule out the possibility that the VRC's Ad5 vector could have a deleterious effect on susceptibility to HIV infection similar to that seen with Merck's construct.
One potentially critical point raised at the AVRS meeting came from antibody expert John Moore. As reported by Bob Roehr, "Dr. Moore suggested that the naturally occurring Ad5 antibody titers might be a surrogate 'that is tracking correlates of susceptibility; the higher the antibody titer to Ad5, the less likely you are to be infected at common risk.' In the AIDSVAX trials, individuals with the highest titers were less likely to be infected."
Moore was referring to data from phase III trials of an HIV vaccine candidate called AIDSVAX; this vaccine was comprised of a monomeric gp120 protein and it failed to offer any protection against HIV infection in two large efficacy studies. However, when researchers looked at the magnitude of the antibody titers that study participants generated against the vaccine, a very interesting finding emerged: people who generated high antibody titers against the vaccine were less likely to become HIV infected than people who received placebo, while the AIDSVAX recipients who generated low titers of antibodies against the vaccine were more likely to become HIV infected than placebo recipients. These findings were published in the Journal of Infectious Diseases and the full article can now be accessed free of charge. To quote from the abstract:
"Despite inducing a complex, robust immune response, the vaccine was unable to reduce the incidence of HIV‐1. Two interpretations of the correlative results are that the levels of antibodies (i) caused both an increased (low responders) and decreased (high responders) risk of HIV‐1 acquisition or (ii) represented a correlate of susceptibility to HIV‐1 but had no causal effect on susceptibility. Although the data cannot definitively discriminate between these 2 explanations, (ii) appears to be more likely."
In other words, the magnitude of the antibody responses to the vaccine likely correlated with the ability of an individual's immune system to mount a robust immune response, and the ability to make a robust immune response is linked to reduced susceptibility to HIV infection. These data strongly suggest that, in the Merck trial, the baseline antibody levels to Ad5 did not just indicate whether participants had previously been exposed to Ad5, they also - like the levels of antibodies against AIDSVAX - correlated with the ability to mount a strong immune response against other pathogens, including HIV. This would offer a compelling explanation of the reduction in HIV incidence associated with increasing anti-Ad5 antibody titers in the Merck trial.
|Anti-Ad5 antibody titer||HIV incidence (vaccine)||HIV incidence (placebo)|
|less than 1:18||4%||4%|
But it would also mean that Merck's Ad5 vector had the most detrimental effect in individuals who would otherwise have been less likely to acquire HIV infection than the study population overall, and such individuals should theoretically be most likely to benefit from an effective HIV vaccine. This could be consistent with these individuals mounting a potent immune response to the Ad5 vector and thus facing increased risk because of the magnitude and duration of CD4 T cell activation subsequent to receiving the immunizations, but until additional results from ongoing studies of both Ad5- and HIV-specific T cell responses among Merck trial participants are available this remains completely speculative. One thing seems certain, however: more data is needed before the VRC's (or any other) Ad5 vector can be declared safe. The impact of the additional gene deletions in the VRC's Ad5 vector on Ad5-specific T cell activation is one of the issues Juliania McElrath specifically mentioned addressing in her presentation regarding follow-up studies at the November 7th HVTN meeting.
Another important issue raised by the correlations between antibody titers and susceptibility to HIV infection relates to the possibility - as contemplated by the AVRS Chair Eric Hunter in the article from The Scientist - that the PAVE 100 trial will be restricted to individuals with low or zero anti-Ad5 titers. The AIDSVAX data and the data on anti-Ad5 titers from the Merck HIV vaccine trial strongly suggest that these individuals face an increased risk of acquiring HIV infection compared to the overall population, and this is even more likely to be the case in places where exposure to adenoviruses is very common (such as the African sites where PAVE 100 is slated to be conducted). So restricting entry to PAVE 100 based on baseline antibody titers against Ad5 would likely mean that the population enrolled in the study would be those least likely to benefit from the vaccine. Conversely, if the trial is opened to anyone without regard to anti-Ad5 titer, the Ad5 vector could potentially have the same detrimental effect on susceptibility to HIV infection seen in the Merck trial.
Until further data emerges from the ongoing analyses of the Merck trial results, a decision on whether to go ahead with PAVE 100 cannot and should not be made. It is disturbing to read in Bob Roehr's article that "there was a general consensus among the committee that the PAVE 100 trial is sufficiently different to allow it to move forward." Given that Merck's Ad5 vaccine showed not only lack of benefit but compelling evidence of harm among people with pre-existing immunity to Ad5, a consensus that another trial involving Ad5 should go ahead seems bizarrely, wildly premature. Although the article in The Scientist cites NIAID director Tony Fauci as saying that a decision on PAVE 100 will be made at an AVRS meeting next month, it appears extremely unlikely that sufficient additional data from the Merck trial will be available to facilitate a well-informed decision by then, and the last thing the HIV vaccine field needs at this juncture is a poorly informed decision.
One somewhat separate issue discussed at the December 12th AVRS was the differences between the Merck and VRC vaccines in terms of the HIV-specific T cell responses they induce; AVRS member Jerry Sadoff complained that comparable immunogenicity data has not been presented, which is painfully ironic given that both the Global HIV/AIDS Vaccine Enterprise and the NIAID-sponsored Partnership for AIDS Vaccine Evaluation (PAVE) have convened many, many meetings about standardizing immunogenicity assays over the past few years. I'll cover this subject in more detail in a follow-up post.