Lack of Efficacy and Safety Concerns Blight Adenovirus Vaccine Vectors

Vaccine vectors based on adenoviruses—which are common in nature and cause severe colds—were once viewed as promising for HIV and other intractable diseases due to their ability to induce antigen-specific CD8 T cell responses in the majority of recipients. The subsequent Icarus-like trajectory of the approach in HIV has been covered in some detail on this blog, starting with the results of the first efficacy trial of Merck’s now-discontinued adenovirus serotype 5 (Ad5) candidate (the Step study), which showed that the vaccine significantly increased the risk of HIV acquisition and did not significantly lower viral load or preserve CD4 T cell counts in participants who became infected.

The HIV vaccine section of TAG’s new 2013 Pipeline Report covers the latest round of grim news, which emerged earlier this year with the announcement that immunizations in the lone ongoing HIV vaccine efficacy trial, HVTN 505, were being stopped at the recommendation of the Data Safety Monitoring Board (DSMB). An interim analyses conducted by the DSMB on April 22, 2013 found that the vaccine regimen—comprised of three priming immunizations with a DNA vaccine and a booster shot with an Ad5 vector (both designed by the NIH’s Vaccine Research Center)—did not show efficacy either in terms of preventing HIV infection or lowering post-infection viral load set points, and there was no possibility of efficacy being demonstrated if the trial ran to completion. The total number of infections was greater in the vaccine arm compared to placebo (41 vs. 30), but many of these infections occurred prior to the administration of the Ad5 boost and the difference was not statistically significant. Post-boost, there were 27 infections in the vaccine arm and 21 in the placebo arm.

On the heels of this announcement came word that extended follow up from the second efficacy trial of Merck’s Ad5-based HIV vaccine, the Phambili trial in South Africa, had documented significantly more infections among vaccine recipients compared to placebo recipients. The data were described by principal investigator Glenda Gray at the HIV Vaccine Trials Network meeting in Washington DC on May 7th, and reported by Jon Cohen in Science on May 10th. Although it went largely unnoticed at the time, on May 14th the National Institute of Allergy and Infectious Diseases (NIAID) issued a bulletin reporting that regulatory authorities had been notified of the Phambili findings. The bulletin notes that while the initial published analysis of the trial results did not reveal significant differences between vaccine and placebo arms, “in the current analysis of the full 3.5 year follow-up period, the researchers found that 100 study participants became HIV-infected: 63 participants who received the investigational vaccine and 37 participants who received a placebo injection. The increased number of HIV infections among the vaccinated recipients was greatest among men and more pronounced roughly 30 months after initial vaccination.” There are some potential caveats: the additional infections occurred after the study was unblinded, and there was a substantial drop-out rate that was slightly greater in the placebo arm. Also, in contrast to the Step trial where the enhanced HIV risk has been reported to have waned over time, the increased rate of infection occurred late (>2 years post vaccination). 

The accumulation of negative data has left a cloud of suspicion hanging over all investigational adenovirus-based vaccines, including those based on alternate human serotypes to Ad5 and chimpanzee-derived adenovirus serotypes. The mechanism of the apparent enhancement of HIV acquisition risk in the Step and Phambili trials remains unclear. Importantly, it is not yet known if the adenovirus vectors were somehow solely responsible or if the effect was interlinked with the immune responses to HIV that were induced by the constructs (as one small but as yet unconfirmed macaque study has suggested). Complicating matters further, evidence from the Step study indicated that the degree of pre-existing immunity to natural Ad5 (as measured by antibody titers) was correlated with the magnitude of the vaccine-associated enhancement of HIV risk. In the absence of clarity regarding these issues, there remains the worrying possibility that experimental adenovirus-based vaccines for other diseases (such as TB, malaria, and hepatitis C) could place recipients at increased risk of becoming HIV infected if they are exposed to the virus.

Currently the various stakeholders in the development of adenovirus-based vaccine candidates, including major sponsors such as NIAID and the International AIDS Vaccine Initiative (IAVI), are conducting meetings to discuss the new findings and their implications for the future. It is not yet known if planned or ongoing trials will be affected. In a related development—as reported by Mike Frick in the TB vaccine section of TAG’s Pipeline Report—plans for a large efficacy trial of an Ad35-based TB vaccine candidate have been significantly scaled back due to the observation that the approach is less immunogenic than preliminary results had suggested.

Comments

[Regina McEnery]

Stay tuned, I guess. NIAID bulletin also noted that the findings in the 503 trial were inconclusive, apparently because the HIV status of 189 participants is unknown at this time. http://www.vaxreport.org/Back-Issues/Pages/old-vaccine-trial-draws-new-scrutiny.aspx. I think NIAID is hosting a meeting in September on ad vectors. Not sure of the date though.

[Richard Jefferys]

I've heard a date of Sept 19, immediately after the next AVRS. Will add a sentence re: the Phambili caveats (I also forgot to specifically note that it's unblinded follow-up). There's yet another wrinkle in that the enhanced risk in Step is reported to have waned over time, whereas in Phambili it emerged late. The small, non-significant difference post-boost in HVTN 505 also was due to several infections in the vaccine arm that occurred late in the course of follow-up. But despite the confusing aspects of the data I think the overarching concern about three trials, all with findings going in the wrong direction, remains.