More Surprises Stem from RV144
At AIDS Vaccine 2011 investigators presented the much anticipated results of a two-year effort to elucidate how the RV144 vaccine regimen protected against HIV
By Kristen Jill Kresge
Two years ago, the prevailing sentiment in the HIV vaccine field was surprise. The source of the surprise was the results of the RV144 efficacy trial that provided the first evidence of vaccine-induced protection against HIV. From the start, many leading researchers opposed the trial, involving more than 16,000 volunteers in Thailand, because they thought the candidates being tested were unlikely to work. Many were also skeptical when the results were released because the trial demonstrated only a 31.2% reduction in HIV infection risk.
Now, after a two-year hunt to try to determine just how the prime-boost vaccine regimen tested in RV144 provided this modest level of protection against HIV infection, the trial has once again yielded surprising findings (see VAX Nov. 2009 Primer on Understanding the Hunt for Immune Correlates of Protection from RV144). At the AIDS Vaccine 2011 conference that, symbolically, took place in Bangkok, Thailand, from Sep. 12-15, investigators presented the much anticipated results of their massive effort to identify the specific types of immune responses that were induced by the vaccine candidates that led to the observed protection. The analysis showed that two different types of antibodies—infection-fighting proteins that are generated in response to specific pathogens—were correlated with the risk of HIV infection among volunteers. The first surprise was that they identified any correlates of risk. Given the lack of public support for RV144, the trial was a scaled-back version of what was originally planned. This meant researchers were left with few samples to analyze in the correlates analysis, making it very much a needle-in-a-haystack search.
The second surprise was that while one specific type of antibody response was correlated with a reduced risk of HIV infection, the other was correlated with an increased risk of HIV infection. These intriguing findings provide valuable clues about how this vaccine regimen might have worked and help bolster the credibility of the RV144 results. “The findings lend credence to the vaccine efficacy seen in the RV144 trial,” said Barton Haynes, who led the RV144 correlates of protection analysis team comprised of four teams of researchers, adding that these results are “intriguing clues.”
Haynes was careful to warn that the antibody responses they identified as correlates may or may not actually be related to the HIV infection risk among volunteers in RV144, and are merely hypothesis generating, not conclusive. “Without this we had uninformed hypotheses,” said Haynes. “Now we have informed hypotheses and directions that come from a trial.” Jerome Kim, deputy director of science at the US Military HIV Research Program, which was a key collaborator on RV144, also urged caution in trying to extend these findings to other vaccine candidates. “Any results may be unique to this vaccine. We have to bear that in mind as we look to the next step in HIV vaccine development,” said Kim.
However, the findings presented in Bangkok give investigators the opportunity to study these specific antibody responses more closely in both human and non-human primate studies, as well as in previous trials to determine if there is a causal relationship between these immune responses and HIV infection risk. Researchers will ultimately use the information generated by these studies and analyses to try to improve upon the 31% protection seen in the trial. “The RV144 correlates work is clearly going to guide us on the future of HIV vaccine development,” said Giuseppe Pantaleo, chief of the division of immunology and allergy at the Centre Hospitalier Universitaire Vaudois in Lausanne, Switzerland.
On the hunt
The goal of the correlates analysis team for RV144 was to determine what immune responses, if any, predicted the HIV infection risk of RV144 volunteers over a three-year period. To do this, the team of investigators carefully analyzed the immune responses present in blood samples collected from 41 volunteers in the vaccine group that eventually became HIV infected through natural exposure, as compared to the immune responses in blood samples from 205 volunteers in the vaccine group that remained HIV uninfected and 40 volunteers who received placebo.
After a series of test runs, researchers settled on six primary assays, or tests, and approximately 30 secondary assays to measure immune responses in this collection of samples. The assays were carefully selected based on several selection criteria. The assays were all run in July, and statisticians at the Fred Hutchinson Cancer Research Center in Seattle then analyzed the results. The final results were shared with the investigators in late August and presented publicly for the first time at the conference in Bangkok.
The statisticians found that the correlation between two of the immune responses and HIV infection risk was statistically significant, a measurement scientists use to give credibility to a finding (see VAX Oct. 2009 Primer onUnderstanding the Statistical Analysis of Clinical Trial Results). The first immune response that was significantly correlated with HIV infection risk was antibodies known as immunoglobulin G or IgG that bind to a specific portion of HIV’s outer coat, a protein structure known as HIV Envelope, or Env. Vaccinated volunteers in RV144 who had this antibody response at week 26 of the trial (which corresponds with two weeks following all six vaccinations that were administered over a six-month period, and was the time when the immune responses peaked) were 43% less likely to acquire HIV than volunteers who did not generate this specific antibody against HIV.
When researchers compared vaccinated volunteers with low versus high levels of this type of IgG antibody they found that those with a higher level of IgG were 71% less likely to become HIV infected than vaccinated volunteers with either low or mid levels of IgG antibodies.
The second immune response that was significantly correlated with HIV infection risk was a different class of antibody that binds to HIV Env. This type of antibody, known as IgA, is most commonly generated in mucosal secretions, such as genital secretions, but can also be found in serum, a component of blood. The responses evaluated for RV144 were all from serum samples because no mucosal secretions were collected in the trial in an effort to cut costs. Vaccinated volunteers who developed IgA antibodies that targeted HIV Env at week 26 of the study were actually 54% more likely to subsequently become HIV infected. While these antibodies were significantly associated with an increased risk of infection among vaccinated volunteers, Haynes reported that vaccinated volunteers with high levels of this type of IgA antibody became HIV infected at the same rate as placebo recipients, suggesting that presence of this type of an antibody response did not enhance an individual’s risk of acquiring HIV. Rather, these antibodies reduced the protective effect of the vaccine candidates, which meant the vaccine efficacy was actually higher among volunteers with low levels of this type of IgA antibody.
Although it is unclear precisely how this IgA antibody could increase the risk of HIV infection among vaccinated volunteers, Haynes presented one hypothesis to explain this finding. Antibodies can act against viruses, including HIV, in many ways. One of the ways is through something called antibody-dependent cellular cytotoxicity or ADCC (see VAX Jan. 2010 Primer on Understanding Antibody Functions: Beyond Neutralization). In ADCC, antibodies bind directly to cells infected with HIV and facilitate the killing of these cells by other cells of the immune system. There is evidence from other diseases, such as cancer, that IgA can block ADCC immune responses against tumors. Haynes said this type of blocking activity would be explored in the RV144 follow-up trials as well as animal studies.
Now that these two correlates of risk have been identified, researchers are planning several analyses and studies to determine if they are causally related to protection against HIV. Or, in the case of the IgA antibody, causally related to a lack of protection. This of course will be a major focus of the post-RV144 trials with the same or a similar vaccine regimen.
Other clinical trial data will come from analyzing previous trials in which one of the same vaccine candidates as was used in RV144 was administered. This includes the two Phase III trials known as VAX003 and VAX004 that tested AIDSVAX gp120 alone in either men who have sex with men in the US or injection drug users in Thailand. There was no protection against HIV infection observed in either of these trials; however, the populations differed greatly from the RV144 volunteers—both the mode of transmission and the level of risk were different in the VAX003 and 004 trials. It is possible that the same responses were induced in volunteers in these trials but the level of exposure to HIV and the diversity of the viruses that infected these volunteers overcame the vaccine’s effect. A similar observation has been made in non-human primates. Genoveffa Franchini, who led the animal models scientific working group as part of the RV144 correlates team, has been able to reproduce the RV144 results in non-human primate studies, but found the protection is dependent on the amount of virus the animals are exposed to. “If you’re using too much virus, you can’t see the vaccine efficacy,” she said.
Researchers are now analyzing samples from the VAX003 and 004 trials, which are available in a repository, to see if this vaccine candidate alone induced similar immune responses to the prime-boost regimen tested in RV144.
Researchers are also planning studies in non-human primates to see if administration of antibodies of the type that were elicited by the RV144 vaccine regimen are capable of protecting monkeys against infection. All of these studies will inform the design of future vaccine candidates that might be able to improve on the efficacy seen in RV144.