Annual AIDS vaccine meeting highlights recent data from clinical trials and lessons on recruitment and retention of volunteers
By Kristen Jill Kresge
Vaccine researchers and immunologists can be forgiven if, at times, they seem overwhelmed by the challenges that lie ahead. Working through the scientific difficulties inherent in developing a vaccine, the disappointing results from clinical trials, and the inconsistent data from animal models could make anyone a little skeptical, if not downright depressed. Seemingly, this despair has not gone unnoticed. "This field could use a little Elavil or Prozac," said Larry Corey of the HIV Vaccine Trials Network (HVTN) to scientists gathered at the AIDS Vaccine 2005 conference in Montreal in September. In conversation, several other researchers shared Corey's opinion about the level of frustration in the field.
But over the following days the hundreds of researchers who gathered were handed a remedy for their depression, albeit in small doses. Results from two Phase I studies with DNA-based vaccines showed promise and sparked a renewed interest in this strategy, illustrating that not all DNA candidates are created equal. Meanwhile Merck's lead candidate, MRKAd5, yielded promising results on many fronts, including its immunogenicity in people with pre-existing immunity to the adenovirus serotype 5 (Ad5) vector followed by encouraging news from the field's only ongoing Phase IIb trial. More preliminary research into new vaccine vectors also attracted attention at the meeting. Wyeth's manipulations of vesicular stomatitis virus (VSV) have rendered the viral vector nearly ready for testing in human volunteers and data from several animal models demonstrating the improved safety of this approach dominated the presentations.
There was also a wealth of information on recruiting and retaining volunteers from the clinical trials and vaccine preparatory studies that continue across sites spanning several continents. Researchers brought up some of the remaining ethical dilemmas involved with the design of clinical trials, including the need to evaluate AIDS vaccines in adolescent volunteers. Dealing with these issues will become essential as candidates advance and the field readies for large-scale efficacy trials.
When Corey addressed the delegates again at the closing ceremony, he encouraged the field to move forward with guarded optimism despite remaining obstacles. He predicted a fruitful year of research that will provide substantial data on vaccines that induce cell-mediated immunity. And if that translates into progress, then it could be even better than Prozac.
DNA down but not out
The Vaccine Research Center (VRC) at the US National Institutes of Health has tested a series of DNA vaccines in several Phase I clinical trials. In one of these trials, VRC 004, volunteers received a 4 plasmid DNA vaccination including env genes from subtypes A, B, C and a fused gag/pol/nef construct from subtype B. These volunteers were subsequently rolled over into the VRC 009 trial where they received a booster inoculation with a combination of 4 Ad5 recombinants that were previously tested alone in the VRC 006 trial (see Research at the extremes, IAVI Report, 9(2), 2005).
Barney Graham of the VRC presented the impressive results of this rollover study in Montreal. All volunteers received a 1010 dose of Ad5 after receiving either 4 or 8 mg of DNA in the previous trial. Researchers observed an 11 to 21-fold rise in immune responses following the adenovirus boost, including a substantial increase in antibody response in volunteers with low Ad5 titers at the start of the VRC 009 trial. Graham also reported that the CD8+ T-cell responses after boosting were comparable to those seen in long term nonprogressors. In volunteers with higher Ad5 antibody titers, the responses were dampened similarly to those reported from other studies.
The results of this study differ from those Merck reported with its DNA/Ad5 vaccine combination, which was abandoned because it offered no advantage over the MRKAd5 candidate alone (now in Phase IIb trial). Graham speculated that there are several possible reasons for this difference, including the addition of the env gene that was not included in Merck's DNA candidate as well as the design of the VRC 009 study, which as a rollover study had at least a two-year gap between the DNA prime vaccination and the Ad5 boost. "The long boosting interval probably does play a role in the improved immune response," Graham said at the meeting. He acknowledged that this represents an impractical vaccination schedule, but argued that these results warrant further research. "DNA is immunogenic and this data shows that it clearly primes."
The VRC is now running a Phase II clinical trial of this DNA/Ad5 combination vaccine to assess its immunogenicity with a more standard prime/boost schedule. The randomized, placebo-controlled study, HVTN 204, will enroll 480 volunteers at HVTN sites in North and South America, Africa, and the Caribbean. Volunteers will receive three injections with the DNA prime and a single Ad5 boost over a period of six months. This candidate is also being tested in a series of Phase I and II trials at sites in East Africa including Kigali, Rwanda, in cooperation with IAVI, at other African sites with the US Military Research Program, and sites in Haiti, Puerto Rico, Jamaica, Brazil, and the US with the HVTN (see Vaccine Briefs).
Results from another Phase I trial with a DNA candidate developed by Advanced BioScience Laboratories and tested at the University of Massachusetts Medical School were also presented at the meeting. The candidate includes a DNA prime containing a subtype C gag gene and env from five HIV subtypes followed by a homologous gp120 protein booster vaccination with an adjuvant called QS-21. The trial included two groups of volunteers that received a 1.2 mg dose of DNA administered either intradermally or intramuscularly, and a third group that received a 7.2 mg dose of DNA intramuscularly, followed in all arms by a 0.375 mg protein boost.
The trial was stopped prematurely upon advisement from a Data Safety Monitoring Board due to a case of cutaneous small-vesicle vasculitis that developed in a single volunteer in the high dose arm. Other adverse events that were reported included grade 2 flu-like symptoms. But Jeff Kennedy, assistant professor of medicine at the University of Massachusetts, believes that further study is still merited. "We're not sure how concerned we are about this. In vaccine development sometimes reactogenicity isn't a bad thing," he says. Kennedy cited smallpox and yellow fever vaccines as clear examples of approved products that can have side-effects but were developed because of the human toll these diseases wrought.
The candidate induced robust immune responses as determined by interferon-g (IFN- g) ELISPOT even prior to the Ad5 boost, and high antibody titers were detected by ELISA at all doses after boosting, with 60-70% of volunteers mounting a substantial antibody response to HIV subtype B. Although promising, Kennedy invokes a baseball analogy in describing the results from this Phase I trial. "It's more like hitting a double with no outs than a walk-off home run. Now we have to figure out how to get these responses even higher." He predicts the future of this candidate is to test it in combination with other promising approaches, including an adenovirus vector.
Modifications in test of concept study
Robin Isaacs, executive director of HIV Vaccine Clinical Research at Merck, presented the rationale behind the company's decision to modify the Phase IIb or test of concept trial—which is testing the trivalent Ad5 vector expressing subtype B Gag, Pol, and Nef proteins—to include volunteers with high vector-specific antibody titers. The trial was originally designed to exclude any volunteers with Ad5 specific titers greater than 1:200, but after analysis of now completed studies with the trivalent vaccine researchers concluded that at higher doses there was less inhibition of immune responses in people with pre-existing immunity than expected.
At the 3x1010 particle dose of the trivalent vaccine, 69% of the 118 volunteers produced an immune response to at least two proteins in IFN-g ELISPOT. When this data was broken down by the level of pre-existing immunity, 77% of the volunteers with titers less than or equal to 1:200 were classified as having the same response. This finding was consistent with other studies that illustrated a broader immune response than previously seen with the monovalent Gag Ad5 candidate. The trivalent vaccine even seems to be immunogenic in people with very high antibody titers, reaching beyond 1:1000.
"We have found that the trivalent vaccine consistently produces a detectable immune response in 60-70% of people and that the impact of Ad5 was less when we used it in people who had pre-existing immunity," says Isaacs. These results led the company to include an equal number of volunteers with high (>1:200) and low levels (≤1:200) of pre-existing antibody immunity to the viral vector, in essence doubling the enrollment in the Phase IIb trial by including an additional 1500 volunteers in the new cohort. The so-called Step study will expand enrollment across all currently active Merck and HVTN sites and may also begin recruiting volunteers at new sites in North America, South America, the Caribbean, and Australia. The primary and secondary endpoints of this ongoing trial are the prevention of infection and the ability to slow disease progression in volunteers that become infected during the course of the trial. Isaacs reported that approximately 1200 volunteers have received the Ad5 candidate so far in all clinical trials and that the only side-effects observed are injection site reactions and fever after initial dosing, especially in volunteers with low adenovirus-specific antibody titers. Results of this intermediate study are not expected until 2008.
"If we continue to find that people with high Ad5 antibodies have a good response to the vaccine then it may well make the vaccine useful for a larger number of people," says Isaacs. In the meantime Merck continues to look at other serotypes, including Ad6, in order to broaden the immunogenicity of the vaccine and find what he refers to as the "ultimate" adenovirus vector.
VSV vector nears testing in human trials
John Rose of Yale University has been developing VSV as a vaccine vector for many years and this negative-stranded RNA virus, which is a natural pathogen of cattle, has proven highly effective. Just a single intranasal inoculation with a VSV vector expressing hemagglutinin from influenza virus elicits impressive neutralizing antibody titers in mice. Proteins from measles virus or the Severe Acute Respiratory Syndrome (SARS) coronavirus expressed in VSV have also provided impressive results in animal models.
In 2001 Rose reported that an AIDS vaccine candidate based on attenuated VSV vectors was able to protect seven rhesus macaques from disease up to 14 months after a SHIV89.6P virus challenge (Cell 106, 539, 2001). All but one of the control animals in this experiment progressed to AIDS in an average of 148 days. These results caught the attention of researchers at Wyeth who began working on developing this approach into a candidate that could be evaluated in human clinical trials.
Although VSV does not cause serious disease in humans it is a replication competent vector and so determining its neurovirulence profile was a priority for Wyeth. Researchers injected wild-type and recombinant forms of the virus directly into the brains of macaques. All animals that received the wild-type VSV developed necrotizing meningoencephalitis, a slowly progressing viral infection of the brain, while some also showed hemorrhaging and inflammation around the blood vessels. The recombinant forms induced less severe necrosis, but it was clear that further attenuation of the vector was necessary.
Wyeth researchers tried various genetic manipulations, including shuffling the nucleocapsid gene further downstream in the viral sequence in order to decrease expression, truncating or deleting the transmembrane glycoprotein ectodomain that is responsible for viral attachment, and introducing mutations into the matrix protein to reduce cytopathic effects. These strategies were tested alone and in combination in several animal models, including mice and ferrets.
David Clarke of Wyeth reported in Montreal that the newly engineered VSV vectors were all "extremely attenuated". In ferrets these attenuated vectors induced only minor lesions upon direct injection into the brain. Intranasal administration led to a further reduction of lesions and caused only minor swelling, but as Clarke notes there is bound to be some inflammation when any agent is delivered directly into the brain.
The researchers then identified two of these attenuated vectors as lead candidates—referred to as N4CT9 and N4CT1—and tested them in both mice and young macaques. With the wild-type VSV they observed mortality in mice with only a 102 particle dose; however with the N4CT1 it took a 107 particle dose to induce mortality. Kevin Wright, also of Wyeth, presented immunogenicity data of these rationally designed variants at the conference. Mice that received a 107 particle intramuscular injection both as a prime and a boost generated cellular responses after the boost equivalent to a single inoculation with wild-type. The anti-Gag responses and the antibody responses as measured by ELISA were even greater with the attenuated viruses than wild type VSV.
The immunogenicity data in monkeys parallels that in mice, according to Stephen Udem, vice president of viral vaccine research at Wyeth. Udem also presented on these vectors at the meeting and waxed poetic over their potential, even going as far as calling it the "best vector on Earth," at least in animals. He remarked that although he is continually surprised by the mutation capacity of HIV, he is also repeatedly impressed by the potency of VSV as a vector. And the genetically modified vector, N4CT9, is now being manufactured in preparation for Phase I clinical trials.
"We've done almost everything we can do in animals," says Wright. "Now we have to see how it works in humans. Let's hope it bears fruit."
Wyeth has had several meetings with the US Food and Drug Administration (FDA) who will be responsible for approving the Investigational New Drug application that the company is now filing. The approach is "well regarded" by the FDA according to Udem, who predicts that the vaccine will be in human trials within a year. "The N4CT9 vector produced less injury than vaccines we've been using for hundreds of years, like smallpox," says Udem. "It's a remarkably effective agent." The trial will evaluate an intramuscular rather than intranasal administration because it induces more robust immune responses.
Learning from experience
As more vaccine candidates enter or progress in clinical trials it will be necessary to recruit more volunteers, especially women, and optimizing the design of vaccine trials was another important focus of the conference. Several presentations in Montreal centered on some of the difficulties experienced or progress made during enrollment for ongoing trials or cohort studies.
Stephen Mawa of Makerere University and the Walter Reed Project in Kampala, Uganda gave a presentation on recruiting for a Phase I trial (VRC 009) in 2004 with the VRC's DNA candidate. The trial site staff conducted more than 20 informative seminars for both community leaders and the general public—each attended by as many as 100 people—and used newspaper and radio advertisements to recruit potential volunteers. In total the site interacted with more than 4000 people and although several hundred women were screened, only a few were enrolled. Despite these extensive outreach efforts the majority of volunteers reported hearing about the study from friends and Mawa emphasized the need to engage in continuous advocacy and education since "word of mouth" is always likely to be an important recruitment tool.
Sanjay Mehendale of the National AIDS Research Institute in Pune, India, where the country's first AIDS vaccine trial began earlier this year, presented more encouraging news on recruiting women for this Phase I vaccine trial. Preparations included sensitizing the trial staff about the need to recruit women, conducting meetings with local women's groups, as well as training the staff on gender-related issues that could affect participation. The IAVI-sponsored Phase I study enrolled 11 male and 9 female volunteers to receive either the low or medium-level doses of the adeno-associated virus based vaccine called tgAAC09.
But to get just nine female volunteers enrolled the site staff had to screen five times that number, indicating that significant screening resources will be necessary if investigators are to recruit a balanced number of women in larger Phase II or III clinical trials in this country.
In preparation for these trials Project San Francisco (PSF), a project founded in Kigali, Rwanda in 1986, has been working extensively with discordant male-female African couples where one partner is HIV infected and the other is not. The PSF sites have been successful at bringing women into clinics for HIV testing and possible enrollment in a vaccine trial and are the world's largest cohorts of HIV serodiscordant couples.
More than 20,000 couples have been screened at one of PSF's sites in Kigali over the past three years, 950 of whom have been identified as serodiscordant, according to a presentation by Erin Shutes who works at this site. The retention rates for these cohorts hover around 90% through one year, offering a unique opportunity for investigators to reach potential female volunteers for vaccine trials. The Kigali cohort is now enrolling volunteers for a Phase I vaccine trial.
Beyond the strategies for recruiting female volunteers, researchers within the AIDS vaccine community are also starting to grapple with the possibility of evaluating promising candidates in adolescent volunteers. The inclusion of adolescents in AIDS vaccine trials bring up completely new ethical questions and will be an important issue for debate and discussion in the coming years. The Montreal meeting provided merely the background for what will be a thorny issue in vaccine research, but as Glenda Gray of the University of Witswatersand reminded researchers, "HIV prevalence among adolescents in South Africa is horrific and excluding them from efficacy trials is a big mistake."