A Change of Tune

Following the first trial showing efficacy and continuing progress in other areas of research, a new chord of optimism was struck at AIDS Vaccine 2010

By Kristen Jill Kresge and Regina McEnery

A few years ago it was not uncommon to hear Anthony Fauci, the veteran director of the National Institute of Allergy and Infectious Diseases (NIAID) at the US National Institutes of Health (NIH), publicly question whether it would be possible to develop an AIDS vaccine. And although the prime-boost vaccine regimen tested in the controversial RV144 trial in Thailand provided only modest efficacy (31.2%), it was enough to make Fauci a convert.

“It is feasible to block acquisition of HIV infection. We know from the Thai trial that it can be done,” Fauci said at AIDS Vaccine 2010, held in Atlanta, Georgia, from Sep. 28 to Oct. 1. “Before, I was not so sure it was feasible. The proof of concept here is huge.”

During his signature overview talk at the annual conference, Fauci highlighted recent progress in the isolation of HIV-specific broadly neutralizing antibodies and their role in structure-based vaccine design, as well as the insights from and plans to build upon the results of RV144. Advances in these areas, as well as many others, including new plans for designing clinical trials and additional data from viral vectors, were presented in the jam-packed sessions that occurred throughout the four-day meeting.

“I believe we are seeing a real reason for optimism,” said Alan Bernstein, executive director of the Global HIV Vaccine Enterprise, which co-hosted AIDS Vaccine 2010 with Emory University’s Center for AIDS Research. Fauci was also enthusiastic about the progress being made. “Our task now is to use the science to get us closer to a much more effective vaccine,” said Fauci. “I don’t think there’s any question we’re going to get there. We have a light at the end of the tunnel. All we need to do is follow the light, follow the science.”

RV144: The search continues

Without a doubt, the results of the RV144 trial helped galvanize AIDS vaccine research. Now, researchers are mining the samples from the trial in the search for possible immune correlates of protection that could potentially allow researchers to rationally improve upon what Nelson Michael, director of the US Military HIV Research Program (MHRP), called the “early but nondurable efficacy” seen in RV144 with the ALVAC canarypox-based prime and AIDSVAX gp120 boost. One year into the RV144 trial the efficacy was as high as 60% (based on the modified intent-to-treat analysis), but it slowly waned over the course of the three-and-a-half-year trial.

Michael reported that MHRP and the 35 investigators at 20 different institutions who are collaborating on the RV144 correlates analysis are still in the initial phase of the process that involves evaluating a broad range of assays and selecting those that will be used, come January 2011, to evaluate the precious RV144 samples in a case-control design. Although the case-controlled studies have not yet begun, researchers at MHRP have already made some intriguing observations, which Michael summarized.

In an exploratory analysis, researchers found that a small group of 60 vaccinated volunteers from RV144 who remained HIV uninfected and who had a positive response by interferon (IFN)-γ ELISPOT (greater than or equal to 20 million spot-forming cells per million cells and at least four times background) had a high frequency of CD4+ T-cell responses that targeted two distinct epitopes (peptides 44 and 49) in the V2 loop of gp120. T-cell responses directed toward these epitopes were not seen in 68 volunteers with breakthrough infections (who are being followed in a separate trial, RV152), and are very rare in HIV-infected Thais (only one person from natural history studies has been found to have these responses).

“That may be on the pathway to a correlate,” said Michael. But he cautioned that this finding was far from conclusive because these immune responses may have occurred in all vaccinated volunteers but then been eradicated when an individual became HIV infected.

What is unique about peptide 44 on the V2 loop is that it includes the binding motif for the integrin α4β7, which was shown by James Arthos and colleagues in Fauci’s lab to be an additional receptor on CD4+ T cells that HIV gp120 binds to and signals through (Nat. Immunol. 9, 301, 2008). “This receptor defines a subset of T cells in the cervix and rectum that is highly susceptible to HIV infection,” said Fauci. The α4β7 integrin mediates the migration of activated T cells to gut-associated lymphoid tissue, where HIV wreaks the most substantial, and often irreversible, damage to the immune system very soon after infection. Arthos and colleagues found that the efficiency of gp120 binding to α4β7 varied considerably. According to Fauci, Env from newly transmitted HIV has been shown to bind better to α4β7 than chronically replicating virus that is more heavily glycosylated. He concluded that an Env confirmation that is defined by easy binding to α4β7 “should seriously be considered as a vaccine target.”

Linear epitope mapping of the B-cell responses in RV144 volunteers conducted by the Vaccine Research Center (VRC) at NIAID also indicate that antibody responses to the V2 loop are more frequent in RV144 volunteers than in volunteers in the HIV Vaccine Trials Network (HVTN) 204 study, which tested the DNA/adenovirus (Ad) serotype 5 prime-boost regimen developed at the VRC that is now being tested in the large Phase II trial, HVTN 505.

Taken together, these observations have piqued the interests of investigators and Michael said that assays looking at vaccine-induced immune responses to the V2 loop, and to the peptides that contain the α4β7 binding site specifically, “will make the cut for case-control studies.”

Meanwhile, there are also plans in place for additional clinical trials to build on the results of RV144. One of these follow-up studies is focused on determining the correlates of protection and is being planned with optimal specimen collection. Another study will test the effect of administering an additional gp120 protein boost to already vaccinated individuals from RV144.

Some of these studies will get underway as early as next year, but it will take until 2013 or 2014 for the next efficacy trial to begin with the same regimen that was tested in RV144. “The field wasn’t in the position to capitalize on success,” said Michael, who reviewed plans for two RV144 follow-up efficacy trials. The first is a Phase IIb trial in Thailand that Michael called a “top priority” because it has the potential to lead to licensure of the vaccine candidate in this region. This trial, which will be funded by the US Army, the Thai Government, the NIH, and Sanofi Pasteur, will test the same ALVAC/AIDSVAX regimen used in RV144 with an additional gp120 protein boost administered six months after the fourth vaccination (12 months after the first vaccination). This trial will enroll men who have sex with men (MSM) at high risk of HIV infection, a much different population than was enrolled in RV144. Immediately after the RV144 results were reported, some researchers speculated that the almost exclusively heterosexual and low- to moderate-risk trial population may have been an important factor in the vaccine’s modest success. Several studies have shown that despite substantial viral diversity in the infected partner, the majority of heterosexual infections can be traced back to a single transmitted founder virus—only about 20% of cases are the result of more than one founder virus. However, more than one transmitted founder virus is evident in about 40% of the infections studied among MSM, suggesting this mode of transmission may widen the range of viruses a vaccine would have to block to be protective.

Another Phase IIb efficacy trial, which would also start in 2014, is being planned in southern Africa. This trial will involve high-risk heterosexual volunteers and is being funded by the Bill & Melinda Gates Foundation, the NIH, the HVTN, Sanofi Pasteur, and Novartis RSA, among others, according to Michael. He said the objective of this trial is to see if the efficacy seen in RV144 can be extended to other geographic regions where there is greater viral diversity. Investigators collaborating on this trial are still deciding which gp120 protein will be used as the boost in this trial.

At the NIH’s AIDS Vaccine Research Subcommittee meeting, which took place the week before the Atlanta conference in Rockville, Maryland, there was extensive discussion about whether to use two clade C HIV gp120s or one clade A and one clade C. Although clade C is predominant in southern Africa, there is evidence from the Phambili trial, the Phase IIb trial of Merck’s adenovirus-based candidate that was halted early, that clade A virus may be emerging in South Africa. Additionally, researchers think that including two gp120 boosts of different clades may increase the breadth of the immune response induced by the prime-boost regimen.


First Trial in Infants 

Results from the first AIDS vaccine trial in Africa in breast-fed infants born to HIV-infected mothers were presented at AIDS Vaccine 2010 in Atlanta. The Phase I, double-blind, randomized, placebo-controlled HPTN 027 trial evaluated the ALVAC-HIV vCP1521 canarypox vector-based vaccine candidate—the same one used in the RV144 trial as part of a prime-boost regimen—in a cohort of 60 infants.

The four-year study found the four-dose regimen, given at birth, and then at one, two, and three months of age, was safe, but not very immunogenic. HIV-specific T-cell responses were not observed by interferon (IFN)-γ ELISPOT assay, but were observed in vaccine recipients using two other assays, cytokine staining and carboxyfluorescein succinimidyl ester CFSE, which measures T-cell proliferation using flow cytometry. However, the difference in immunogenicity between the vaccine and placebo groups was not statistically different during any time point in the trial. Results of antibody assays should be presented early next year, said Huyen Cao, director of cellular immunology at the California Department of Public Health.

Cao, who conducted the T-cell analysis, said the trial was significant in that it showed that “neonates were able to mount an immune response that is comparable to adult populations.” Until now, ALVAC cp1521 had only been tested in clinical trials in combination with other vaccines, but other ALVAC formulations have been tested alone and were found to be poorly immunogenic. —RM 

Longer-term follow up of STEP

The RV144 trial may have opened the long-awaited door to AIDS vaccine efficacy, but it is not the only late-stage clinical trial informing researchers these days. Clues continue to emerge from the Phase IIb test-of-concept trial known as STEP, which was stopped in 2007 after an interim analysis revealed that Merck’s Ad5 vector-based vaccine candidate MRKAd5 appeared to be ineffective in both preventing acquisition of HIV and in reducing levels of the virus in individuals who became infected despite vaccination. And going even further back, scientists have also turned their attention to a 12-year-old efficacy trial, Vax004, which tested AIDSVAX, the same gp120 protein candidate tested in RV144, by itself.

Ann Duerr, HVTN associate director for Scientific Affairs and an investigator in the STEP trial that enrolled 3,000 volunteers at high risk of HIV infection in North and South America, the Caribbean, and Australia, reported on a long-term follow-up study (HVTN 504) of STEP participants. The most complete analysis conducted to date showed that among 1,836 male participants there were 99 HIV infections reported among vaccinated men and 73 infections among unvaccinated men. This analysis excluded women because at the time vaccinations were halted in the STEP trial in 2007 only one HIV infection had been reported among 1,134 female volunteers. An analysis of the combined STEP and HVTN 504 data showed that all vaccinated men were at an increased risk of HIV acquisition compared to placebo recipients, based on the modified intent to treat analysis. However, the increased risk of infection in men who were uncircumcised and/or who had preexisting Ad5 immunity waned over the course of the four-year study. Interpreting the results of the post-hoc analyses has been tricky, however, because retention rates for vaccine recipients were slightly higher than they were for placebo recipients, particularly when the trial was halted and unblinding occurred. Differences in behaviors between men who received vaccine and placebo could also have impacted the relative risk, according to Duerr.

Researchers still do not know why uncircumcised male vaccine recipients with high Ad5 antibody titers at baseline had a higher rate of HIV acquisition. Some researchers speculated it might have been because the vaccine induced a long-lasting increase in HIV target cells that homed to the gut mucosa. But, Nicole Frahm, associate laboratory director of the HVTN, presented data from a study of blood and mucosal tissue from the lower GI tract of 23 HIV uninfected STEP trial participants—half of whom received MRKAd5—which showed that 30 months after the final injection there were no significant differences in expression of CCR5, CCR9 (a chemokine receptor that is usually present at higher levels on activated Ad5-specific CD4+ T cells), and CD103 (an integrin expressed on activated T cells) in either CD4+ or CD8+ T cells in the gut mucosa of vaccine or placebo recipients, even when stratified by baseline Ad5 titer, therefore failing to explain the increased risk of HIV infection among vaccinated volunteers who were uncircumcised and had higher pre-existing Ad5 immunity. “Even though we don’t have a mechanism, we’re pretty sure this is a real biologic effect,” said Larry Corey, principal investigator of the HVTN.

Another look at AIDSVAX

It’s certainly not news that AIDSVAX, which was tested in the first two Phase III AIDS vaccine trials, failed to protect against HIV infection. Still, the 5,400-person Vax004 trial that enrolled mostly MSM at high risk of HIV infection in North America and the Netherlands continues to engage and inform researchers, most recently in Atlanta, where it was highlighted by theoretical biologist Bette Korber, who heads the HIV Database and Analysis Project at Los Alamos National Laboratory (see Tracking HIV Evolution, IAVI Report, May-June 2010).

The renewed interest in Vax004 began earlier this year when a team of researchers released results of a study that analyzed the neutralizing antibody (NAb) responses against highly sensitive tier 1 and moderately sensitive tier 2 strains of HIV among vaccinated volunteers (J. Infect. Dis. 202, 595, 2010). A subset of plasma samples from over 100 randomly selected vaccine and placebo recipients were tested for their ability to neutralize HIV subtype B strain MN, which was part of the vaccine; a heterologous but readily neutralized tier 1 strain, SF162; as well as a panel of 12 tier 2 viruses that represent a spectrum of clade B viruses, including transmitted founder viruses. A comparison of responses among vaccinated and placebo recipients showed that the vaccine elicited high titers of NAbs against HIV MN and SF162 and weak overall neutralizing antibody responses against tier 2 viruses. The study’s authors concluded that the lack of neutralization of tier 2 viruses was consistent with the lack of protection seen in the trial and suggested that, “one way to improve the efficacy of current HIV-1 vaccines may be to elicit stronger NAb responses against tier 2 strains of the virus.”

Following this study, Korber, who was not involved in the original Vax004 studies, did an additional comparison of NAb responses in vaccinated individuals to see if there was any association between NAb responses to tier 2 viruses and being uninfected. In Atlanta, she reported results from what she called a “preliminary re-examination of the Vax004 data.” Korber found that essentially none of the 21 vaccinated males that were subsequently HIV infected had a detectable tier 2 neutralizing antibody response; their tier 2 responses were comparable to placebo recipients. In contrast, only about 20% of the uninfected vaccinees had a low-level tier 2 antibody response. Vaccinated individuals with detectable tier 2 responses remained uninfected. The difference in the tier 2 responses between infected and uninfected vaccinees was “highly significant,” said Korber. This observation was confirmed by Peter Gilbert, research professor in biostatistics at the Fred Hutchinson Cancer Research Center, and David Montefiori, director of the laboratory for AIDS vaccine research and development in the department of surgery at Duke University Medical Center.

A larger study including new samples from male and female subjects from the Vax004 study is being planned to see if the results from Korber’s and Gilbert’s analyses can be confirmed, but Korber said these findings are intriguing, while also urging caution about over-interpretation of the data. “It suggests that maybe the level of tier 2 responses observed, or some surrogate they are a marker for, did provide some protection, but individuals who made these responses were too infrequent to give an overall vaccine effect.” Montefiori said that based on Korber’s “interesting and provocative” data there is now “greater confidence that there is a significant association between the level of neutralizing antibody response and the risk of HIV acquisition.” Though, he added, “what that means, we still are not certain of.”

Korber said these observations suggest that it is worth looking at the tier 2 antibody responses among uninfected and infected vaccinees in RV144 to see if there is a similar correlation. However, a poster presented in Atlanta by Montefiori’s group suggests that the level of NAbs to tier 2 viruses may not explain the modest protection seen in RV144. An analysis of RV144 samples by Montefiori’s laboratory found that the NAb responses to tier 1 and 2 viruses are actually lower than what was seen in the Vax004 trial. Montefiori cautioned that the populations enrolled in RV144 and Vax004 were considerably different—one involved low-risk heterosexual men and women from Thailand and the other high-risk MSM from the US—yet he was still taken aback by the results. “I was hoping it would be the opposite,” said Montefiori. “The fact that the response in RV144 is weaker suggests it might not be neutralizing antibodies that are providing efficacy.”


Better Together? 

Things are looking up. In the past year, the HIV prevention field has been buoyed by results from trials that show a prime-boost vaccine regimen and an antiretroviral-based microbicide gel are both partially effective at blocking HIV infection. And, as early as February of next year, the results from the first efficacy trials of oral pre-exposure prophylaxis (PrEP) as a means of HIV prevention will be released.

While trials to both confirm and improve upon these results are being planned, researchers are also considering ways to evaluate the possible synergy of partially effective HIV prevention strategies in clinical trials. At the AIDS Vaccine 2010 conference, Steven Self, co-director of the Vaccine and Infectious Disease Institute at the Fred Hutchinson Cancer Research Center, noted that there is a limited window of opportunity for such trials, including one that would evaluate the combined efficacy of the partially effective vaccine regimen tested in the RV144 trial in Thailand and the 1% tenofovir microbicide gel that was recently shown to block acquisition of HIV among women in the recent CAPRISA 004 trial in South Africa (see Microbicides Finally Gel, Securing Spotlight at the International AIDS Conference, IAVI Report, July-Aug. 2010). Self said that within two or three years the opportunity to conduct such trials could pass as efficacy data on both oral PrEP and ARV-based microbicides becomes robust enough to perhaps warrant licensure of these modalities. “You really want to do them sooner rather than later,” he said.

Mathematical models suggest varying degrees of efficacy when two partially effective interventions—such as a vaccine or a PrEP regimen—are offered in combination, according to Self. One model predicted a combined efficacy as high as 70% when a vaccine with 30% efficacy and a PrEP regimen with 40% efficacy were given in combination. —RM and KJK

Adapting to more flexible clinical trials

Given the valuable data being extracted from these and other trials, many researchers are calling for more clinical trials and more efficient ways of conducting these trials. “We start an efficacy trial every five years,” said Corey, who suggested AIDS vaccine researchers need to get past their twin afflictions: a fear of failure after a few candidates were unsuccessful, and a fear of success that has led to long development timelines for each product. “I don’t know, maybe what the field needs is a psychiatrist,” he joked.

Corey lamented that the first follow-up efficacy trials to RV144 will not begin before 2013. “The reality is we’ve been pleasantly surprised by RV144 but it’s taken us a lot of time to really line up what’s next,” he said. In contrast, Corey pointed out that the microbicide field is preparing to begin a confirmatory trial of the 1% tenofovir vaginal microbicide gel just four months after the first trial results were released in July, showing it was 39% effective in protecting women from HIV infection (see Microbicides Finally Gel, Securing Spotlight at the International AIDS Conference, IAVI Report, July-Aug. 2010; Vaccine Briefs, page 18). Corey acknowledged that with the antiretroviral-based microbicide researchers have the advantage of knowing the mechanism of protection. Still, he said, “we’re falling behind other prevention modalities. The number and the pace [of trials] needs to be increased.”

The exploration of new trial design strategies was one of the objectives outlined by the Global HIV Vaccine Enterprise in its 2010 Scientific Strategic Plan, published last month and officially launched at the conference (Nat. Med. 16, 981, 2010). At a special session focused on the plan, Bernstein said that efficacy trials should not be viewed as the culmination of a “series of basic science experiments,” but as part of the entire vaccine discovery process. The present system of moving candidates through efficacy trials is “not the most productive way of going forward,” he added.

During his plenary address, Corey argued that the best way to increase the pace and number of clinical trials is to employ adaptive trial designs that can test multiple vaccine candidates concurrently or sequentially in randomized, blinded, placebo-controlled trials that are conducted as standard Phase IIb trials with prevention of infection as an endpoint. Each candidate would be compared to placebo, and although investigators could not directly compare one candidate to another, Corey said investigators would have some ability to rank the efficacy of the different candidates.

Adaptive trials would be designed with more frequent analyses, or looks at the data, by the data safety monitoring board, allowing them to adapt the trial while it’s underway based on how the vaccine candidates are performing. This approach would permit investigators to identify candidates with low or high efficacy more quickly and give researchers the flexibility to add another arm or even trial if a positive signal is detected. In a trial population with a 4% HIV incidence rate and 2,000 volunteers per group, Corey said it would be possible to reach a decision point on whether a vaccine candidate is working in approximately 20 months, as long as volunteers are rapidly enrolled into the trial. If this type of adaptive trial design was employed in past efficacy trials, Corey said the four-and-a-half-year Vax004 trial could have been stopped two to two-and-a-half years earlier, the STEP trial could have been stopped nine months earlier, and RV144 could have been stopped two-and-a-half years earlier. For candidates with an efficacy in the range of 40%-50%, Corey said the trial would have to run the full duration. One important distinction of these adaptive trial designs is that they are not considered licensure trials because the more frequent data analyses while the trial is underway reduce the overall power of the study, making it more of a research tool that allows investigators to rapidly prioritize candidates for further study.

If HIV vaccine investigators have been acting like serial monogamists, Corey is advocating that they become speed daters, viewing trials more as experiments that allow them to quickly assess which candidates to court further. “We need to initiate one to two such trials each year for the next four years,” said Corey.

But exactly how these adaptive trials would be structured won’t be known until regulators and investigators are able to sort through logistical and ethical questions. The Joint United Nations Programme on HIV/AIDS (UNAIDS), the World Health Organization (WHO), and IAVI will be hosting a meeting in February 2011 to discuss the challenges and advantages of adaptive clinical trial designs. Catherine Hankins, chief scientific officer of UNAIDS, said the biggest challenge will be communication. “People will need to understand why one candidate gets dropped and another does not and that when it is dropped it doesn’t mean that it is a bad product, just that it is not doing as well as others.”

How full is the pipeline?

When asked if there were enough candidates in the pipeline to support adaptive clinical trials that call for testing multiple candidates simultaneously, Corey answered, without hesitation, “yes.” Some of these candidates will be different combinations of the now familiar cast of viral vectors. Giuseppe Pantaleo, principal investigator of the poxvirus vaccine development consortium of the Collaboration for AIDS Vaccine Discovery, provided an overview of viral vectors now in development.

The RV144 follow-up trials will involve canarypox vectors, either ALVAC, the vector developed by Sanofi Pasteur that was tested in RV144, or NYVAC, another canarypox vector that has been tested as part of a prime-boost regimen in Phase II trials by EuroVacc. Michael said ALVAC has a commanding position among viral vectors because it is the only one to have shown any efficacy and joked that Stanley Plotkin, a veteran vaccinologist and advisor to Sanofi Pasteur, calls ALVAC the “Rodney Dangerfield of vectors because it doesn’t get much respect.”

Other replication-defective vectors in development include alternate serotype Ad vectors, including an Ad35 candidate developed by IAVI, and an Ad26 candidate developed by Dan Barouch, an associate professor of medicine at the Beth Israel Deaconness Medical Center (BIDMC) and Harvard Medical School. Both of these candidates have been tested alone in Phase I safety and immunogenicity studies and preliminary results suggest they were both safe and immunogenic, and a Phase I trial of an Ad35/26 prime-boost regimen recently started in Boston (see Vaccine Briefs).

There are also plans to begin a Phase I trial next year to test Barouch’s Ad26 candidate in a prime-boost combination with a modified vaccinia Ankara (MVA) vector-based candidate developed by MHRP. Michael presented results in Atlanta from a nonhuman primate (NHP) study of Ad26/MVA conducted by MHRP and the Integrated Preclinical/Clinical AIDS Vaccine Development program, funded by a NIAID grant and led by Barouch. In this study, five groups of eight macaques were immunized with placebo, MVA/MVA, DNA/MVA, Ad26/MVA, or MVA/Ad26, encoding Gag, Pol, and Env from simian immunodeficiency virus (SIV)smE660. Michael showed that the Ad26/MVA prime-boost regimen was a more potent inducer of Gag- and Env-specific antibody responses by ELISA, and Gag-, Pol-, and Env-specific CD8+ and CD4+ T cell responses as measured by IFN-γ ELISPOT than DNA/MVA, MVA/MVA, or MVA/Ad26 prime-boost regimens.

To see how well these regimens protected, all animals received six low-dose, intra-rectal, heterologous SIVmac251 challenges. After only one challenge, 50% of the macaques in both the placebo and MVA/MVA groups were infected. Two challenges were required to infect 50% of the animals in the DNA/MVA group, and three challenges were required to infect an equivalent percentage of monkeys in both the MVA/Ad26 and Ad26/MVA groups. Michael noted that the DNA/Ad5 prime-boost regimen developed at the VRC, which is currently being tested in a Phase II trial, did not show any protection against heterologous SIVmac251 challenge in NHP studies. Four of the eight monkeys in the Ad26/MVA group that were infected had an approximately 1 log lower set-point viral load than infected animals in the other groups.

Michael said additional NHP studies would be conducted to see if the level of protection in this study can be augmented by adding an Env protein boost to the Ad26/MVA regimen and to evaluate the immunogenicity of the vectors formulated with mosaic inserts. There are also plans to test the Ad26/MVA prime-boost regimen with mosaic inserts in a Phase I clinical trial starting next year that is a collaboration of MHRP, BIDMC, NIAID, and Crucell. Mosaic antigens are computationally designed to achieve optimal coverage of the many different versions of HIV circulating globally. NHP studies have shown that mosaic antigens induce T-cell responses with greater breadth (number of epitopes recognized) and depth (number of viral variants recognized) than either natural proteins or consensus antigens (Nat. Med. 16, 324, 2010).

The enhanced breadth of T-cell responses that is achieved with mosaic antigens directly translates into the ability of the T cells to recognize more epitopes in circulating strains, according to Korber, who designed the mosaic antigens. “Breadth correlates with viral control in monkey studies,” she said. “Although we don’t have a single clade vaccine yet, a global vaccine is a worthy target.”

Two other Phase I clinical trials involving mosaic antigens are also planned to start in the next few years. One, which is being conducted by Barton Haynes, director of the Center for HIV/AIDS Vaccine Immunology (CHAVI), will compare a wild type clade B HIV Env, a trivalent mosaic Env, and a group M consensus Env in a DNA /NYVAC prime-boost regimen. This trial is expected to begin in 2012 and is being sponsored by the Bill & Melinda Gates Foundation and NIAID. Dan Barouch is also collaborating with IAVI and Crucell to manufacture Ad35/26 vectors encoding mosaic antigens in preparation for clinical trials.

Korber reported that her team is also working on optimizing B cell mosaic antigens. “We’ve done it for the entire Envelope but you could also do it for a specific region such as the CD4 binding site,” said Korber. This, she suggests, may be one way that researchers would be able to guide antibodies along the affinity maturation process that creates mutations in the variable regions of an antibody. This may be important given the observation that many of the HIV-specific broadly neutralizing antibodies have an unusually high degree of affinity maturation compared with other antibodies (see Vaccines to Antibodies: Grow Up!, IAVI Report, July-Aug. 2010). Korber said that the hypothesis is that most good neutralizing antibodies develop in chronic infection because of continuous exposure to virus variation. A vaccine could mimic that diversity in target epitopes using B cell mosaics either all at once or as a series of immunizations, she posited.

Other viral vectors in various stages of preclinical development include a Lymphocytic Choriomeningitis Virus vector, which is being developed by the VRC, and several replication-competent viral vectors. Pantaleo said a replication-competent form of NYVAC is in preclinical testing and should enter clinical trials by 2012.

One replication-competent virus vector that has shown promise is a rhesus cytomegalovirus (RhCMV) vector developed by Louis Picker, associate director of the Vaccine & Gene Therapy Institute at Oregon Health & Science University. At last year’s meeting Picker showed that 54% (13 of 24) of macaques vaccinated with the replicating RhCMV vector exhibited immediate viral control upon infection with SIVmac239 (see Raft of Results Energizes Researchers, IAVI Report, Sep.-Oct. 2009). In fact, 12 of the 13 animals maintained undetectable plasma viral loads for a year with occasional blips of detectable virus decreasing in frequency. Picker reported results in Atlanta from a study designed to further analyze what was happening in these 12 animals. He showed that depleting either the CD8+ or CD4+ T cells in these animals had no effect on viral replication, in contrast to control animals.

Picker and colleagues then sacrificed four of the RhCMV vector vaccinated macaques to conduct extensive tissue analysis of the animals using ultrasensitive polymerase chain reaction to try to detect SIV DNA or RNA in the spleen, liver, tonsils, jejunum, ileum, colon, lymph nodes, bone marrow, or thymus of the animals. They found no inducible, replication-competent SIV in any of these tissues, indicating to Picker that the RhCMV vector-based vaccine is controlling SIV infection in these animals over the long term, and may possibly even lead to eventual clearing of the virus altogether.

Picker contends that this observation provides strong support for the hypothesis that vaccine-elicited cellular immunity, if present early at the sites of SIV infection and replication, could prevent or abort infection, or provide complete control of the infection before massive systemic viral replication occurs. He compared the control afforded by the RhCMV vector to that seen with live-attenuated SIV or SHIV, an SIV/HIV hybrid virus, which when administered either intravenously or mucosally can protect against highly pathogenic SIV challenge. Picker noted that the protection seen with live-attenuated SIV and SHIV occurs in the absence of any SIV-specific antibodies, and that successful live-attenuated vaccines induce persistent replication and a high frequency of SIV-specific T-cell responses in tissues.

The bigger question, it seems, is whether CMV vectors can be safely tested in clinical trials, or as Picker asked, “is this protection without practicality,” much like the live-attenuated vaccine concept. To this end, Picker is developing CMV vector-based candidates that retain the immunogenicity of wild-type CMV, yet are attenuated enough so that they are safe enough to be tested in CMV seronegative volunteers in clinical trials. He called this work “quite promising.”


A Cheaper Option  

Amid the talk of biomedical HIV prevention strategies, there was one voice at the AIDS Vaccine 2010 conference in Atlanta, calling for a much more low-tech strategy—behavior change. Susan Allen, who directs the Rwanda Zambia HIV Research Group, presented a mathematical modeling study that shows behavior change is not only a more cost-effective approach, it could avert more new HIV infections than test and treat, a strategy that calls for universal testing and immediate treatment for HIV-infected individuals as a way of reducing HIV transmission.

Allen used mathematical models to compare the impact of couples voluntary counseling and testing (CVCT)—the type of counseling she has pioneered in her well-established cohorts of HIV serodiscordant couples—with test and treat for HIV-infected partners of serodiscordant couples in Mozambique.

The model took into account that 20% of the HIV-infected partners among serodiscordant couples already meet the criteria for treatment and that an additional 5% of HIV-infected partners per year would require treatment.

The models predict that a five-year, nationwide rollout of CVCT for 3.3 million couples in Mozambique would cost about US$115 million and avert between 180,000 and 580,000 new HIV infections. Whereas, the cost of a five-year rollout of test and treat for 551,525 couples would avert only 41,364 infections and cost $372 million. The model only included the cost of treatment for those HIV-infected individuals who did not yet meet the criteria for starting treatment. “Why test and treat continues to be promoted in this economic climate is beyond me,” she said. —RM


Antibody frenzy continues

Over the past year there has been a flurry of new HIV-specific broadly neutralizing antibodies (bNAbs) isolated from chronically HIV-infected individuals (see Research Briefs, IAVI Report, Jan.-Feb. 2010). And it seems that now that researchers have developed the technology to pluck them out quite efficiently, there will be many more additions to the antibody armamentarium.

In Atlanta, researchers from the VRC reported for the first time the isolation of two bNAbs, referred to as VRCPG04 and VRCPG05 because they were isolated from samples collected from IAVI’s protocol G cohort of chronically HIV-infected individuals that also led to the isolation of PG9 and PG16 (Science 326, 285, 2009). John Mascola, deputy director of the VRC, said both of these antibodies target the CD4 binding site and were identified using the same technology that allowed them to isolate VRC01-03 (Science 329, 811, 2010). The identification of VRCPG04-05 shows that VRC01-like antibodies are not limited to a single donor.

Pascal Poignard, principal scientist at IAVI’s Neutralizing Antibody Center at The Scripps Research Institute (TSRI), reported the isolation of 13 new monoclonal antibodies from four so-called elite neutralizers—individuals whose sera can neutralize a large number of HIV isolates—identified from IAVI’s protocol G cohort. Three of these antibodies target a collection of overlapping, highly conserved epitopes on the viral spike not targeted by any of the other bNAbs described so far.

And researchers from CHAVI also reported isolation of new neutralizing antibodies from their cohorts of both acutely and chronically HIV-infected individuals. Georgia Tomaras, associate director of research at the Duke Human Vaccine Institute, reported that bNAbs targeting the membrane-proximal external region (MPER), which is also the target site for the bNAb 4E10, have been detected in three individuals from the CHAVI cohorts. The epitope targeted by these antibodies “binds to a similar but not identical epitope as 4E10,” said Tomaras. She and Mattia Bonsignori, also with CHAVI, also reported that five new monoclonal antibodies dubbed CH01-CH05 that target quaternary epitopes on the V2 and V3 loops of HIV Env were isolated from a single individual in the CHAVI cohorts.

Despite this windfall, the generation of bNAbs to HIV is still uncommon. “Individuals can make these antibodies, but they are a relatively rare occurrence,” said Haynes, who noted that while approximately 20% of chronically infected individuals have some degree of antibody neutralization breadth in plasma, only 1% -2% of individuals make very broadly neutralizing antibodies. “Multiple factors may predispose people to being able to make bNAbs,” said Haynes. One of these factors might have to do with lax tolerance controls that otherwise would delete such antibodies, he suggested. Haynes showed that in knock-in studies in mice in which the 2F5 heavy chain is expressed, the B cells that express the 2F5 antibody either get eliminated or are controlled by tolerance mechanisms.

The evolution of antibodies

As researchers home in on the structures of the new crop of antibodies, they are developing a clearer picture of some of their unique attributes (see Research Briefs, IAVI Report, May-June 2010; Science 329, 856, 2010). Studies have shown that these bNAbs are poly-reactive, have long CDR H3 regions (particularly PG9 and PG16, see image at right), and are extremely somatically hypermutated, according to Haynes.

The high level of affinity maturation is one attribute of these HIV-specific bNAbs that has been receiving more attention (see Vaccines to Antibodies: Grow Up!, IAVI Report, July-Aug. 2010). “The persistent nature of HIV may lead to affinity maturation,” said Peter Kwong, chief of the structural biology section at the VRC. Even the most recently identified antibodies such as VRCPG04 have accumulated a large number of mutations in their variable regions. For VRCPG04, which can neutralize approximately 90% of circulating isolates, 29% of its sequence differs from the germline version of the antibody. “This antibody again is highly affinity matured,” said Mascola. The sequence of VRCPG04 is only 51% homologous to VRC01 in its heavy chain, but “despite substantial sequence diversity, they have very similar modes of recognition by their structures,” said Mascola. This suggests to him that there may be more than one path to achieve this high level of affinity maturation. “Antibodies all develop in different ways,” added Kwong.

To better understand how germline antibodies evolve to attain the affinity maturation needed for potent neutralization, researchers at the VRC are beginning to track the precursors of the bNAb VRC01 by applying genomic technologies to the analysis of antibody development. Mascola, Kwong, and colleagues have used 454 Sequencing to identify more than 250,000 antibody heavy chain sequences from peripheral blood mononuclear cells of the same donor from whom the VRC01-03 antibodies were isolated. From these sequences, they have categorized a family of 20,000 based on how much they diverge from the germline precursor and how their sequence is similar to VRC01. One antibody they have studied is 59% homologous to VRC01 and only 9% divergent from the germline sequence. “This antibody is nowhere near as affinity matured as VRC01 and neutralizes quite potently,” said Mascola, adding, however, that it does not neutralize as potently or with the same breadth as VRC01. He said this method can be used to potentially identify the genetic precursors of VRC01. This information could reveal how the immune system got there, said Mascola, which “may provide insights on how to design immunogens and which strategies to use for vaccination.”

Haynes and colleagues are also exploring computationally derived, reverted versions of bNAbs to understand how affinity maturation and/or tolerance controls impact the ability of naive B cells to make bNAbs. Haynes referred to the germline ancestors of 2F5 as putative antibodies because there are several unmutated germline versions that can be inferred. “These are only approximations,” he said. “We don’t have the actual naive B cell in our hands.”



PG16 Fab structure with variable (V) and constant (C) parts of the heavy chain in orange and the variable and constant parts of the light chain in yellow. The CDR H3 can be seen on top.

Courtesy of Robert Pejchal at The Scripps Research Institute; also published in Proc. Natl. Acad. Sci. 107, 11483, 2010.


Effector functions of antibodies

In addition to the focus on bNAbs, there is also a more concerted effort to understand the role of Fc-receptor mediated antibody function, including antibody-dependent cellular viral inhibition (ADCVI), or what Don Forthal, associate professor of medicine at the University of California Irvine School of Medicine, referred to as “an immune function in search of a better acronym.” ADCVI is a measure of the overall antiviral effects due to the Fc-receptor effector functions of antibodies, including antibody-dependent cellular cytotoxicity (ADCC), which occurs when antibodies coat an HIV-infected cell and recruit innate immune cells to kill the infected cell. “The Fc region is really the link between the innate and the adaptive immune system,” said Galit Alter, principal investigator at the Ragon Institute.

Interest in ADCVI has been fueled by both animal studies and results from clinical trials. There is direct evidence from studies by Ann Hessell, a staff scientist at TSRI, and colleagues that shows Fc-receptor mediated antibody functions play a role in protection against SHIV infection in rhesus macaques. In another NHP study, Genoveffa Franchini, chief of the animal models and retroviral vaccine section at the National Cancer Institute, has been able to mimic the RV144 results in rhesus macaques immunized with an ALVAC prime/AIDSVAX boost and challenged rectally with repeated low-doses of SIVmac251. In this study, highlighted by Forthal in Atlanta, three of the 12 animals were completely protected. Two of the protected animals had good ADCVI responses, while the other had no ADCVI activity. Forthal said some of the variability in ADCVI responses is due to levels of immunoglobulin G2, which correlate inversely with ADCVI activity. And in Vax004, the Phase III clinical trial of AIDSVAX, there was higher ADCVI activity among uninfected subjects (J. Immunol. 178, 596, 2007). “ADCVI seems to play an important role in this study,” said Forthal.

All of this suggests to Forthal that improved Fc-mediated antibody function should improve vaccine efficacy. “This might be achieved by having the right adjuvant, which shouldn’t be that hard, and the right immunogens, which might be harder,” said Forthal.

Guido Ferrari, a CHAVI investigator, and colleagues are studying the development of ADCC and ADCVI responses in natural HIV infection. In two individuals from CHAVI’s cohort of acutely infected individuals, Ferrari reported that both ADCC and ADCVI mediating antibodies develop rather quickly against the autologous transmitted virus, and at least three months earlier than neutralizing antibody responses against heterologous virus.

Meanwhile, Justin Pollara, from the Duke University Medical Center, and colleagues are studying the role of ADCC-mediating antibody responses in nine elite neutralizers selected from a CHAVI cohort of 308 chronically HIV-infected individuals. Pollara found that these elite neutralizers had a significantly higher magnitude of ADCC mediating antibodies than neutralizing antibodies. Pollara said this suggests that antibodies capable of mediating ADCC are more commonly produced in natural infection and therefore may be more readily induced by vaccination. Future studies will focus on what makes a good ADCC response in natural HIV infection and on mapping the epitopes recognized by ADCC mediating antibodies to identify the HIV Env epitopes that may be important to include in vaccine candidates.

Structure-based design

One obstacle to developing vaccine immunogens that could induce bNAbs against HIV is the unstable trimeric structure of HIV’s Envelope spikes, what Rich Wyatt, director of viral immunology at IAVI’s Neutralizing Antibody Center at TSRI, called “a well-shielded, shape-shifting target.” In the absence of stable, engineered trimers that mimic the structure of HIV’s functional spikes, researchers have taken a different tack to developing immunogens. This approach, referred to as epitope scaffolding, marries the precision of structural biology to elucidate the epitopes on HIV that bNAbs bind to, with computational biology, which allows researchers to manipulate and design protein structures.

Just before the conference in Atlanta, progress was reported in the use of epitope scaffolds to elicit antibodies (Proc. Natl. Acad. Sci. doi:10.1073/pnas.1004728107). Bill Schief, a research assistant professor at the University of Washington, along with Kwong, Wyatt, and others, grafted the epitope of the HIV gp41 transmembrane glycoprotein that is the target for the bNAb 2F5 onto a heterologous protein scaffold using techniques of computational protein design. Immunization of guinea pigs with either a single epitope scaffold or a prime-boost combination of two different epitope-scaffold platforms elicited structure-specific antibodies that mimic 2F5. Crystal structures of the antibodies elicited by the 2F5 epitope-scaffolds in complex with gp41 showed that these antibodies were able to induce the gp41 conformation that allowed the antibodies to bind to the 2F5 epitope. In his opening remarks, Fauci trumpeted this approach to immunogen design, calling these “very exciting results.”

Schief and colleagues have also developed CD4 binding-site scaffolds that he says are now being tested as immunogens. “Epitope scaffolds have utility as immunogens,” said Schief. “We do have a chance to elicit neutralizing antibodies.”