Special Report: Thai Trial Results
Results from RV144 Send Scientists in Search of Clues
Whether or not this trial is the key that helps researchers unlock the mysteries of immunological protection against HIV, it provides an important clue—actually 8,000 of them
By Kristen Jill Kresge
Results from an AIDS vaccine trial known as RV144 involving more than 16,000 participants from Thailand show that a prime-boost regimen of two genetically engineered vaccine candidates lowered the rate of HIV infection by about 31%, but had no effect on set point viral load in those who became HIV infected despite vaccination. This is the first time an AIDS vaccine candidate has shown any efficacy in preventing transmission of the virus—a finding that has boosted the field and opened many doors of scientific investigation.
“To be clear, the level of efficacy demonstrated by this vaccine is a modest one. However, it is the first time that we have ever seen a positive signal of efficacy in a human trial of any HIV vaccine—a welcome and exciting result in a field that has been characterized by many disappointments for more than two decades,” says Anthony Fauci, director of the US National Institute of Allergy and Infectious Diseases (NIAID), the agency that funded the bulk of the trial’s $105 million cost (see RV144 in Detail).
So far, limited data has been shared publicly, but trial investigators say some data analyses from RV144 are being readied for publication in a major medical journal and will also be presented later this month at the AIDS Vaccine Conference in Paris (October 19-22). Meanwhile, committees of experts have already been established to start poring over the data. These committees will make recommendations about how best to use the specimens collected during the six-year study, with a focus on trying to tease out any possible correlates of protection. “If we can identify a correlate we can then accelerate development enormously,” says Peggy Johnston, director of the vaccine research program at the division of AIDS at NIAID.
Another major focus will be using this first positive signal in humans to fine-tune the current assays used to evaluate the immunogenicity of vaccine candidates, as well as to potentially improve upon or even validate a nonhuman primate model. Progress in either of these areas could advance the field substantially and lead to the development of improved candidates.
“RV144 is the largest HIV vaccine trial to date [but] it’s just the first step in a longer journey to a globally effective HIV vaccine,” says Jerome Kim, deputy director of science at the US Military HIV Research Program (MHRP).
The trial’s significance
RV144 tested Sanofi Pasteur’s canarypox vector-based candidate ALVAC-HIV (vCP1521) in a prime-boost combination with AIDSVAX B/E, a genetically engineered version of HIV’s gp120 surface protein. Among the 8,198 volunteers who received placebo, there were 74 HIV infections through three years of follow up, compared to 51 infections among the other half of the volunteers who received six shots of the two vaccine candidates (four of ALVAC-HIV and two of AIDSVAX B/E) over a six-month period (see RV144 in Detail for dosing schedule). The difference of 23 infections between the two groups was statistically significant, with a two-tailed p-value of 0.039 and a 95% confidence interval that ranged from 1.1% to 52.1%, according to trial investigators. This means that if the true vaccine efficacy is zero, then there is a 3.9% probability that the split in the number of HIV infections between the vaccine and placebo groups would have been at least as wide as what was seen in the trial, according to Peter Gilbert, co-director of the statistical center for the HIV Vaccine Trials Network (HVTN). He explains the confidence interval like this: “There is 95% probability that the true vaccine efficacy lies within that interval between 1% and 51%.”
These results, referred to as “modest” by the investigators, led some researchers to speculate about the veracity of the finding. They noted that if just a couple more infections occurred among vaccine recipients, the p-value would have risen above 0.05, at which point it would no longer be statistically significant. But Johnston says this is all speculation. “The data is the data,” she says.
According to Gilbert it is difficult to draw conclusions about how real the observed efficacy is without having seen the follow-up analysis from the trial. “The vaccine efficacy might have changed with time. When those analyses are done it will really help gauge whether this is real or not,” he says. “It would probably be a mistake to discount the result and also a mistake not to acknowledge the possibility that this could be a fluke. They are both possibilities. Our best guess from the data is that the true vaccine efficacy is 31%.”
Donald Francis, executive director of Global Solutions for Infectious Diseases, who is considered by many to be the man behind AIDSVAX, reacted strongly to speculation over the significance of the results. “Those are the numbers,” he said. “Just take the numbers and try to make a vaccine out of it.”
Francis, who was involved in the two earlier Phase III trials of AIDSVAX while at VaxGen and whose organization now holds the intellectual property rights to the product, is no stranger to criticism. According to him, the difference between the vaccine and placebo groups was more dramatic earlier in the trial. “There was a 50% to 60% efficacy in the first year,” he says. “Then it started waning.” He speculates that if more booster shots of AIDSVAX were given the results may have been more dramatic. Nelson Michael, director of MHRP, also says that it looks like there was more of an effect in the first year of the study, and so investigators will be looking in particular at the durability of the responses.
Sponsors of the trial stressed that the study was not intended as the final hurdle before seeking licensure for the vaccine candidates. Michael says six shots over four visits spanning six months is “not exactly a deployable regimen.” While the official title of RV144 was a Phase III efficacy trial, Michael says that the trial was actually a Phase IIb test-of-concept trial, based on the number of HIV infection endpoints investigators anticipated when the trial commenced. In the early 1990s, Thailand waged a multi-faceted HIV prevention campaign that required commercial sex workers to use condoms during every sexual act. The country’s HIV prevalence dropped from 2.4% in 1993 to 1.9% in 2003 when RV144 started. The most recent data available suggest HIV prevalence is about 1.5%, according to the Joint United Nations Programme on HIV/AIDS.
Also, the potential efficacy of this regimen outside of HIV clades B and E, which are the predominantly circulating subtypes in Thailand, is unknown. “The vaccines were designed for use in Thailand,” says Kim.
Digging for correlates of protection
Generally, researchers were surprised by the result from RV144, even the trial investigators (see Surprise and a Burst of Scientific Energy, below). This surprise was in part due to the observation that the prime-boost regimen prevented acquisition, without having any affect on viral load. “I expected that if we saw anything it would be a difference in viral load, not a prevention of [HIV] acquisition,” says Johnston. According to Michael, the trial was powered to detect a small difference in viral load between the two groups, yet the viral load set points were “spot on the same.” The 51 individuals who became infected despite vaccination are being followed in a companion study, RV152, which will continue collecting information from these volunteers at six-month intervals. This information will be critical in the hunt for possible correlates of protection.
Several researchers speculate that the level of protection afforded by this prime-boost regimen was likely mediated by antibodies and not cellular mediated immunity (CMI), which is thought to be associated with control of viral load. “The absence of an effect on viral load suggests this [protective effect] wasn’t due to CMI,” says Stanley Plotkin, a veteran vaccinologist and consultant with Sanofi Pasteur, the company that developed ALVAC. “Most people are pinning this efficacy to antibodies,” adds Michael. The ALVAC vaccine candidate had been genetically engineered to express HIV subtype B gag and pro and subtype E env. “Based upon prior work done on these similar vaccines, my bet is that it’s an antibody effect from the env component,” says Louis Picker, associate director of the Vaccine and Gene Therapy Institute at Oregon Health & Science University.
In multiple Phase I and II clinical trials, ALVAC induced only a moderate level of cellular immune responses. And neither ALVAC nor AIDSVAX has previously induced broadly neutralizing antibodies against HIV, which are how most, if not all, licensed vaccines are thought to protect. Data collected from previous trials of both ALVAC and AIDSVAX led 22 prominent AIDS vaccine researchers in early 2004, shortly after the launch of RV144, to publish a policy forum in Science magazine questioning the scientific rationale for pursuing a large-scale trial of this prime-boost regimen. They argued that other candidates in early phase clinical trials at that time had a greater hope of success. “I find it amusing that although many people looked down their noses at the ALVAC vector it is giving us the first immunological evidence of protection against HIV,” says Plotkin.
In 2003, prior to the publication of the policy forum, the HVTN scrapped plans to conduct a Phase IIb test-of-concept trial of similar design to RV144 in the US because the CD8+ T-cell responses induced by the prime-boost regimen did not meet a pre-specified frequency of at least 30%, as measured by interferon (IFN)-γ ELISPOT, in a Phase II trial. “We had set up certain criteria for a level of immunogenicity using ELISPOT data and it didn’t meet those criteria. We thought that was going to be the important immunologic correlate,” says Susan Buchbinder, a principal investigator of the HVTN. However, when the results of the Phase II ALVAC/AIDSVAX trial were published, the study’s authors concluded that it was still possible that the canarypox candidate could demonstrate some efficacy that was not correlated with the ELISPOT assay (J. Acquir. Immune Defic. Syndr. 44, 203, 2007).
This efficacy was seen in some studies with nonhuman primates (NHPs). Data from studies in neonate rhesus macaques hinted that ALVAC encoding Env could induce humoral immunity—a version of ALVAC expressing Gag, Pol, and Env from simian immunodeficiency virus (SIV) demonstrated protection against a low-dose oral inoculation of SIVmac251 in 10 out of 16 neonate rhesus macaques (J. Acquir. Immune Defic. Syndr. 38, 124, 2005). This study and route of challenge was designed to mimic HIV transmission in infants during breast feeding. A similar result was observed in neonate macaques with modified vaccinia Ankara (MVA), another poxvirus vector. Genoveffa Franchini, chief of the animal models and retroviral vaccine sections at the National Cancer Institute, has been involved in numerous studies of ALVAC in nonhuman primates. She says that the partial protection seen in neonate macaques was encouraging, although there are obvious differences between the infant model and adults. In older macaques, her group first observed some protection against non-pathogenic virus (HIV-2) infection with ALVAC, but against a more pathogenic challenge virus (SIVmac251), there was only a transient reduction in viral load.
However, now Franchini says that based on recently collected and still unpublished data from her lab, there is some evidence that protection against infection with the prime-boost regimen of ALVAC and gp120 can also be demonstrated by altering the SIVmac251 challenge dose in nonhuman primates. “If we decrease the dose of virus challenge, we start to see protection from acquisition and from [high] virus load,” she says, adding that this is only true in the vaccinated animals, because all control animals get infected. “In humans so far they are not seeing the protection from virus load but I think that the analysis is probably still going so we will probably know more in a couple of years.”
According to a review article on poxvirus-based vaccine candidates that Franchini co-authored, studies with both ALVAC and MVA provided “the first evidence that when low doses of virus are used in a rigorous macaque model, protection from overt infection can be observed even with vaccines that do not elicit neutralizing antibodies to the viral challenge” (Expert. Rev. Vaccines 3, Supplement, 2004).
But overall, the levels of antibodies induced by poxvirus-based vaccine candidates were lower than what was seen in response to those induced by recombinant protein vaccine candidates, including AIDSVAX, which is what led to the investigation of the prime-boost strategy tested in RV144. “The antibody responses from the glycoprotein have been good but have not given evidence of prevention of acquisition before,” says Plotkin. In 1994, NIAID refused to fund a large-scale trial of an earlier version of AIDSVAX after experiments showed that the antibodies triggered by the vaccine, while effective in neutralizing cell-line adapted HIV strains in the laboratory, could not neutralize strains in circulation at that time. AIDSVAX B/E was then tested alone in two Phase III clinical trials, and although it induced antibodies against HIV, the vaccine candidate had no effect on HIV infection rates. These two studies involved primarily men who have sex with men (MSM) or injection-drug users (IDUs).
So could the prime-boost combination of ALVAC and AIDSVAX be the trick? Some believe so. “The speculation I would make is that it was somehow the combination. It suggests the antibody response was improved by canarypox,” says Plotkin. Others decline to speculate until more data is shared. “It’s very early days,” says Dennis Burton, a professor of immunology and molecular biology at The Scripps Research Institute. “From the antibody side we just don’t even know enough to comment.”
As samples from RV144 are analyzed and the data is reported, it should shed more light on the role antibodies played in the 31% protection afforded by the prime-boost regimen. In addition to screening for neutralizing antibodies, researchers will also likely consider the role of non-neutralizing antibodies and antibody-dependent cellular cytotoxicity (ADCC). Barton Haynes, director of the Center for HIV/AIDS Vaccine Immunology, will be leading the committee that will analyze B-cell immunity, while Don Forthal, a viral immunologist at the University of California in Irvine, will be running the ADCC assays on the RV144 samples. Although the role of ADCC in either prevention of HIV infection or control of disease progression in humans is unclear, there is evidence to suggest it may be important. Through ADCC, antibodies can link HIV-infected cells with effector cells, including natural killer (NK) cells, which can then kill the HIV-infected cell and help stop the spread of the virus. “We were the first to show that ADCC could affect protection in the macaque model,” says Burton.
There is some evidence from a Phase I/II trial in Thailand that suggests the ALVAC/AIDSVAX prime-boost regimen can induce ADCC, while ALVAC alone does not (Vaccine 23, 2522, 2005). In this study, there was a significant difference between the magnitude of ADCC responses between vaccinees who received the prime-boost regimen and placebo recipients. ADCC activity in this study was measured through chromium release from gp120 subtype B and subtype A/E recombinant-coated target cells. The study’s authors concluded that the prime-boost regimen elicits HIV-specific ADCC and that “this enhancement may give rise to protection from and/or prevention of HIV infection.”
While there is plenty of plasma and serum from volunteers in RV144 for antibody analyses, cell samples from the volunteers are much more limited and no mucosal samples were collected. This may make unraveling any cellular responses that may correlate with protection more complicated. “There wasn’t a lot of material that was collected to do ancillary studies,” says Picker. Michael says that the volume of cells collected during the trial—there are about 20 million cells for analysis—was limited because investigators amended the trial protocol to collect fewer cell samples from volunteers. “This goes to the lack of enthusiasm that surrounded this trial at the start,” he says.
Because so few cells are available, Michael says no analysis of these samples will be done until the committee on T cells, which is being headed by Julie McElrath, director of the HIV vaccine evaluation unit at the Fred Hutchinson Cancer Research Center, agrees as to which assays should be prioritized. Michael says they are considering viral inhibition assays and polychromatic flow cytometry, as well as others. “To me the critical thing is to make sure those precious samples are being used to answer the very best questions by the very best groups,” says Alan Bernstein, executive director of the Global HIV Vaccine Enterprise.
Some researchers question whether it will be possible to identify correlates of protection from this trial with only 51 HIV-infected vaccinated volunteers to compare to approximately 8,000 uninfected vaccinated volunteers. Moreover, it may be that the correlate is something current assays are not even set up to measure. “We do not know whether our current measures of the human immune response are even relevant to the protection that we see in this trial,” warned Fauci. But others are remaining optimistic, or if not optimistic at least not pessimistic. Plotkin says that although the trial was not set up ideally to determine the correlates, they still could be found.
Most everyone agrees that if researchers could identify immunological correlates from this study, it would help propel HIV vaccine research and lead to the development of improved candidates. “If those insights emerge, this will be a tremendous advance,” says Mark Feinberg, vice president of policy, public health and medical affairs at Merck. “That would put the field in a totally different place than it has ever been.” Feinberg says that if a correlate of protection is found, it would be cause to start a major vaccine development program. And at that point, he says the field would have to consider which model would be best suited to development of an HIV vaccine. “If you do have a direction to advance, maybe the right model is a more broadly collaborative approach than one entity like a given company,” he says.
Informing animal models and assays
Another avenue of exploration will be trying to use studies in nonhuman primates to learn more about the protection afforded by the ALVAC/AIDSVAX prime-boost regimen. “This trial will give us a chance to validate animal models and look at assays,” says Seth Berkley, president and chief executive officer of IAVI. “That is really a big deal.”
Franchini has been asked to coordinate the group that will be evaluating the prime-boost regimen in animal models, and she thinks studies in NHPs may also help in identifying correlates of protection. “I think we can dissect this in macaques if we can figure out the [right] dose to challenge with,” Franchini says, which is an ongoing debate in the field (see Looking for the Perfect Challenge, IAVI Report, July-Aug. 2009). “We want to dilute [the challenge virus] down to the point where we can reproduce the results of the Thai trial with the same vaccine.” Then, according to Franchini, this model could be used to make improvements to the ALVAC/AIDSVAX regimen, as well as to compare other existing vaccines to help select which ones to test in clinical trials. “I think this trial is going to help us to retool the macaque model,” she says.
Assays that are currently used to measure immunogenicity, including ELISPOT, may also undergo additional scrutiny following the RV144 results. After results from the STEP trial showed that Merck’s adenovirus vector-based candidate did not impact HIV acquisition or viral load, researchers began calling into question the value of the ELISPOT assay in determining immunogenicity. “Adenovirus had very good immunogenicity, very nice ELISPOT, and we have no protection, so what does immunogenicity tell us? We don’t know what is protective,” says Franchini. “I really support nonhuman primate studies but we also have to be humble and say, at the end of the day what we really care about is what happens in humans.”
Others echoed Franchini’s sentiments about the value of conducting clinical trials. “You can only show things that are important in humans, in humans,” Plotkin says.
The study population
One factor that may have contributed to the protection observed in RV144 is the population in which the vaccine candidates were evaluated. Rather than enrolling volunteers at high risk of HIV infection, volunteers in RV144 were recruited from the general population. “This includes people at higher risk and people at very low or no risk,” says Kim.
This could impact the route of exposure and the amount of virus people were exposed to. Feinberg says the low-risk study population could have played a key role in RV144. The STEP trial was conducted in a much different population. “Many of the volunteers were high risk and the significant majority of infections in STEP were in MSM,” says Feinberg. While some MSM and IDUs were enrolled in RV144, the majority of volunteers were heterosexual men and women, which according to Fauci allowed investigators to study a low-dose mucosal exposure to HIV. “One of the reasons people have postulated that this trial may have succeeded is that the intensity of [HIV] exposure in this trial might have been lower,” added Kim. “We know from work on other vaccines that even if you have a highly effective vaccine, you can overcome that vaccine immune response with a sufficiently high-dose challenge.” This is also what Franchini has observed in NHPs—only at a lower challenge dose is there protection against acquisition of SIV. Plotkin says dose and repeat exposure are important considerations for any vaccine.
In addition to studying the samples from RV144 and analyzing a similar regimen in nonhuman primates, there is also consideration being given to conducting additional clinical trials with this prime-boost regimen, or with ALVAC alone, which has never been tested by itself in a Phase IIb or Phase III efficacy trial.
Samples still exist from the Phase III trials of AIDSVAX. “We have about 20 freezers full of specimens from Thailand and the US,” says Francis. This frozen material could be useful in looking at antibody difference in individuals who were vaccinated with AIDSVAX alone versus the prime-boost, he adds.
There may also be the need to manufacture additional doses of AIDSVAX, according to Francis. Although some residual vaccine still exists, Francis says he has started building a team to look into manufacturing more doses. “It’s going to be expensive but the cost is probably not going to be an issue at this point,” he says. “I think there will be great support to do this.”
NIAID has already provided approximately $20 million in 2009 to the US Army Medical Research and Material Command for follow-up studies associated with RV144. How this money is spent will be determined after consultations with the expert committees. NIAID says it will consider the provision of additional resources depending on the results of those consultations and its budget limitations. “We need to bring the best minds together and map the way forward,” says Fauci.
And given the first evidence of efficacy from any HIV vaccine candidate, the way forward seems a bit brighter to many in the field. “Today I have a renewed sense of cautious optimism that the possibility of improving on these encouraging results and ultimately developing a highly effective vaccine to protect against HIV infection is within our reach,” Fauci says. Plotkin agrees. “This should make everyone more hopeful.”
Regina McEnery and Andreas von Bubnoff contributed to this article.
Prominent players in the field respond to RV144
By Regina McEnery
On his wall at the University of Washington, Larry Corey, principal investigator of the HIV Vaccine Clinical Trials Network (HVTN), has a plaque that reads, “If at first you succeed, try hard not to look astonished.”
Suffice it to say, there weren’t many researchers who met that challenge after hearing the results of RV144. The first evidence of efficacy in an AIDS vaccine clinical trial even caught some of the study’s investigators by surprise. “I was stunned,” said Nelson Michael, director of the US Military HIV Research Program (MHRP), which helped execute the trial. Results of the trial, which was conducted by the Thailand Ministry of Public Health, were announced September 24 in Thailand and videocast to the Rayong and Chon Buri provinces where the clinical research centers were located.
These are interesting times in the field of AIDS vaccine research. Two years and three days after AIDS researchers began sifting through the disappointing results of the 3,000-person Phase IIb test-of-concept STEP trial with Merck's vaccine candidate, the larger RV144 trial of 16,000 participants gave researchers a much more encouraging sign. “I think it is a big step forward,” says Alan Bernstein, director of the Global HIV Vaccine Enterprise. “It has energized the field.”
The canarypox vector-based ALVAC-HIV in combination with a genetically engineered version of gp120 known as AIDSVAX showed a modest 31.2% efficacy in preventing acquisition, but had no impact on viral load among participants who became infected despite vaccination. This illustrates one of the many puzzles that scientists will now endeavor to solve through careful analysis of specimens collected in the trial, animal models, and perhaps future clinical trials (see Results from RV144 Send Scientists in Search of Clues).
RV144—the largest AIDS vaccine trial conducted so far—offers scientists a platform to try and locate the immune correlates of protection. “Mechanistically, it’s going to be a lot of work, but it will be very important,” says Seth Berkley, president and chief executive officer of IAVI. Identifying the correlates may still be like finding the proverbial needle in a haystack, but at least RV144 may provide a haystack.
“This kind of signal in prevention of infection is something we have all been seeking,” says Corey, who studied a previous version of ALVAC-HIV, but was not directly involved in RV144. Corey says the field now has a framework that suggests it is possible to make an AIDS vaccine. “We have certainly broken that barrier,” he says. “That’s a milestone.”
Peggy Johnston, director of the vaccine research program at the division of AIDS at the US National Institute of Allergy and Infectious Diseases (NIAID) and NIAID’s representative on the RV144 protocol team, says she was elated and surprised by the findings. When reflecting on the controversy that surrounded the trial’s launch, and its impact on the study’s design, Johnston was somewhat philosophical. “Controversy and disagreement [are] part of the scientific process,” she says. “It’s one of those things that even though it might be painful when it happens, it leads to better experiments. I am very happy with the way that all turned out.”
Though the sponsors of RV144 haven’t released data yet beyond the results—more findings are expected to be presented later this month at the AIDS Vaccine Conference—the study gave researchers and advocates plenty to ponder. “This is a vaccine that nobody, including myself, had much hope for so I think it’s positive,” says Louis Picker, associate director of the Vaccine and Gene Therapy Institute at Oregon Health & Science University. Yet he urges caution when interpreting the results. “I think there are going to be calls [saying] we should take this and put it in people for real. I think that would be a huge mistake.”
Indeed, Supatra Nacapew, an activist with the Foundation for AIDS Rights in Bangkok, has many questions about the RV144 trial, not just about the impact of the prime-boost regimen on acquisition of HIV and viral load, or how the trial population will be monitored now that the study is completed, but also on how to temper expectations among Thai citizens who may call for the vaccine to be available. “Some may wrongly claim that there is already an effective vaccine, and generate public demand,” says Nacapew. “What can we do to prevent this?”
Mitchell Warren, executive director of the AIDS Vaccine Advocacy Coalition, agreed that the field has a long journey ahead. “This was a Herculean task of a trial. It lasted six years,” says Warren, who describes RV144 as a huge shot in the arm for the field. “How do you communicate this result, not just what comes next for this vaccine but what comes next for vaccine research generally?”
Both Sanofi Pasteur, maker of ALVAC-HIV, and Global Solutions for Infectious Diseases (GSID), which holds the intellectual property rights to AIDSVAX, say the immediate goal should be to learn more about the vaccines and what immune responses they elicited. “There is going to be serious discussion in Thailand of all different issues, and we have to be prepared to respond,” says Donald Francis, who heads up GSID and has been involved with the AIDSVAX vaccine since it was developed by biotechnology giant Genentech.
For now, though, it appears as though RV144 will function as a kind of virtual laboratory, sprouting questions that are already engaging some of the best minds in AIDS vaccine research. Time will tell what evidence gets uncovered along the way. “Let’s hope that we can learn something useful from this,” says Dennis Burton, professor of immunology and molecular biology at The Scripps Research Institute. “The field reacted too quickly to the Merck trial and subsequent investigations indicated that the early responses were not correct. There is a lesson in that for all of us.”
Andreas von Bubnoff and Kristen Jill Kresge contributed to this article.