AMP-ticipation: Context and concepts for understanding the AMP Trials

Bill Snow
November 2020

Infusions from the Antibody Mediated Prevention Trials (AMP) have concluded, and publication of the primary results on prevention efficacy are expected soon. The AMP trials are groundbreaking, logistically complex and well-run clinical trials. The pre-specified primary endpoint at week 80 has passed for all participants, enabling the all-important, primary prevention efficacy results to be analyzed. Once the analysis is completed, it’s expected the findings will be shared publically to inform further bNAb product development. Meanwhile, the studies continue with ongoing visits as COVID allows, to complete pre-specified safety follow-up.

The trials were conducted to test the concept that a broadly neutralizing antibody could prevent infection, and, if so, to estimate what blood level of that antibody would be protective. Subsequent bNAb approaches could then simplify the method of delivery and add breadth with multiple antibodies to stop virtually all infecting viruses. (For background, see The Rise of Broadly Neutralizing Antibodies published by AVAC in 2018.)

Basics of the AMP Studies

VRC01 is a single, potent, broadly neutralizing, monoclonal antibody. Collectively known as the AMP studies, two parallel prevention efficacy trials of VRC01 were launched in 2016, enrolling a total of 4,600 people globally.

Each trial tested a higher and lower dose of VRC01, administered intravenously every eight weeks to evaluate it as a long-acting intervention against HIV versus a placebo. The AMP Studies were conducted jointly by the HIV Prevention Trials Network (HPTN) and HIV Vaccine Trials Network (HVTN) and funded by the US National Institutes of Health.

This document discusses the multiple factors that could contribute to the primary endpoint of efficacy, which looks at the rate, number, and timing of infections, and other factors related to virology and immunology. Those include differences in populations, PrEP use, side effects and safety to date.

The table below reflects the language of the protocol and summarizes the study objectives, sample size, populations, and countries participating in each trial.

Populations and Susceptibility

While the trial designs are essentially identical, the two trials differ greatly by location and population – in terms of gender, geography, virus subtypes, transmission routes, and access to PrEP – all factors that could have influenced the outcome of each trial. Consequently, findings from each trial should be evaluated for both differences and similarities. These two separate but parallel trials may reinforce one another or differ, as we have seen in other paired HIV prevention efficacy trials. VRC01 may have a different benefit in different parts of the world with different circulating strains and in different contexts of risk.

Participants’ Reaction and Response During Infusion Period

While human antibodies such as VRC01 generally have good safety profiles, and adverse events are evaluated continuously in real time, the full safety data will be reported after the follow-up period. In the meantime, early reporting on participant adherence to trial visits, how well they tolerated the infusions, and how many dropped out and why will offer valuable insights. Future antibody products will most likely be delivered by injection under the skin (subcutaneous or “sub-Q”) as a more practical and broadly accessible alternative to infusions, but the ability to concentrate the antibodies enough for small, subcutaneous injection has not been fully resolved. Participant attitudes toward passive infusion can and should guide future development.

The Top Line, Number of Infections in Each Arm and Antibody Levels

These trials were designed to demonstrate a statistical difference between antibody and placebo at each dose level. If the differences from the placebo are significant, that would clearly be a positive outcome and a good justification for moving forward with broader combinations and more potent antibodies. Each trial was designed with 90 percent power to detect efficacy above 60 percent with the lower bound of confidence above zero. If that endpoint has been met, the effort to improve on this result with current and combined antibodies will continue. If there’s less difference between active and placebo arms of these trials, the field needs to look beyond the dose levels of VRC01 to understand if and how neutralizing antibodies can induce meaningful protection. Relevant factors may be frequency of dosing and type of breakthrough infections, among others.

The trials were developed to determine a baseline level of antibody for protection. These data will also validate how durable the antibodies are (how they persist and remain viable). Earlier phase studies supported the 8-week dosing schedule, and participants returned to the trial site every 4 weeks for blood draws, so there should be sufficient data to evaluate antibody levels at different durations from the time of infusion, and threshold concentrations that may correlate with protection.

It’s possible that the efficacy won’t be high enough to reach the 60 percent level overall or in any subgroups. In either case, additional, multiple analyses will be needed to break down the components of any observed efficacy. We can expect that the trial team will complete all the analysis according to the objectives and statistical plan and give appropriate explanations and limitations on any unplanned findings.

PrEP Use

Oral PrEP was offered to all participants in both trials as part of the standard of prevention, but availability and use in the two populations was likely different. For the MSM/TG trial it was provided by Gilead through a mail order pharmacy. For the women’s trial, the trial connected participants to PrEP through country-specific approved mechanisms. In both trials, participants were asked about their PrEP use and tested for it periodically. How much PrEP use actually took place in each trial will be an important factor in understanding and interpreting the results. It is predictable that there will be varying levels of PrEP uptake between the two trials, as PrEP implementation is, broadly-speaking, much further along in the countries hosting HVTN 704/HPTN 085 compared to HVTN 703/HPTN 081. It will be important to understand if and how PrEP affected the results.

Timing: Correlating infections with antibody levels

A great deal of early analytical work was done to give these trials statistical power to estimate a threshold level of antibody to achieve protection, if protection is demonstated in the trial. A paper published in early 2017 – Basis and Statistical Design of the Passive HIV-1 Antibody Mediated Prevention (AMP) Test of Concept Efficacy Trials – lays out the approach and rationale in great detail (Gilbert et al, HHS Public Access Stat Commun Infect Dis, 2017). The timing and nature of infections that did take place in the trial will be critical to determine what level of bNAb actually reduces the risk of infection.

Each trial is designed to identify a threshold level of VRC01 antibody that reliably confers protection. This will be calculated by measuring antibody level and non-neutralizing viral activity in individuals who seroconvert. Those data come from the two dose levels given every eight weeks, HIV testing monthly, and approximating decay of antibody between visits. That will provide a measure of what wasn’t protective, and conversely, what was.

Comparing Infecting Viruses: Genetics, sensitivity sieving

In addition to antibody levels, any significant differences between viruses that infected antibody recipients and those that infected the placebo recipients has the potential to identify more precise aspects of the antibody protection. This is the secondary objective, to determine if some viruses were neutralized by VRC01, but others escaped.

This comparison of viruses is known as sieving. Investigators compare viruses from the active and placebo arm of the trial, looking at both their genetic sequences (genotypes) and their neutralization susceptibility (phenotypes). Genotyping is based on analyzing the genetic sequence from each infected participant, whereas phenotyping measures how easily those viruses are neutralized by VRC01. “Virus panels” have been developed that represent a range of HIV types, with any virus categorized as easier or harder to neutralize.

Another critical aspect will be whether differences in the circulating viruses in each trial are differently susceptible. VRC01 has been described as effective against 90 percent of globally circulating viral isolates; this was based on laboratory testing against one global panel prior to the trials. The AMP trials allow a real-time comparison of circulating infecting virus in the active and placebo arms, and will update knowledge of currenly circulating viruses in the various trial populations.

Mapping Antibody Effector Functions

Much of the talk about neutralizing antibodies focuses on the traditional neutralizing antibody function. Neutralizing antibodies depend on the adaptive immune system, which uses the variable regions (at the forked end of the antibody) to make precise contact to the anitigen binding site, as shown in the graphic.

The RV144 Thai HIV vaccine trial and other more recent studies have identified killing mechanisms aside from or in addition to classic neutralization that may also play a role in clearing HIV. Building on that body of research, AMP is also looking at antibody effector functions that are driven by the third, heavy chain-only arm of the antibody. These functions which are non-neutralizing, can be an important supplement or alternative to neutralization, and perhaps play a role in passive protection with bNAbs. Therefore, sophisticated and complex analysis of multiple effector functions will also be conducted on the AMP samples as part of the secondary objective. Those results, which may not be available until after the initial findings are published, will not be as straightforward to analyze as traditional neutralization – but they may be critical for defining mechanisms of protection.

Validating Methods and Assays

The AMP studies are also the first opportunity to confirm scientific hypotheses that HIV monoclonal antibodies can protect humans from HIV infection. Until now, animal challenge studies have been an imperfect substitute; defining a threshold of antibody that results in protection in AMP would be both novel and of great use in developing other antibody prevention regimens and for vaccines. Lab assays, determining the amount of bNAb associated with protection, can therefore be compared to real world human infection in the presence of a bNAb. AMP will show how predictive such measurements actually are.

Ramifications for Passive Protection

Since monoclonal antibodies are a powerful medical tool and a proven therapeutic and preventive measure against other infections, cancer and other diseases, a basic premise of the AMP trials was to gauge the potential of broadly neutralizing monoclonal antibodies as a method for HIV prevention. Monoclonal antibody technology is advancing quickly, becoming more cost effective and practical, and COVID-19 has been transforming vaccine and monoclonal antibody development. Rapid development of SARS CoV-2 prevention and treatment with neutralizing antibodies is receiving unprecedented attention, which may also lead to advances in monoclonal antibody engineering that can inform future work on bNAbs for HIV prevention and treatment.

Externally, and of great importance, is the concurrent development of long-acting PrEP using antivirals. Multiple long-acting, drug-based prevention options are moving forward at this time, most notably the injectable cabotegravir that has recently shown efficacy among men who have sex with men and transgender women, while the trial among cisgender women is ongoing. These small molecule drugs are, in general, less expensive to manufacture. Antibody mediated prevention as an option must show comparative benefits to continue to advance, such as fewer side effects, less potential for resistance, or greater acceptance from potential users.

Implications for Vaccine Development

At present, a large portion of HIV vaccine research and development is focused solely on generating broadly neutralizing antibodies, but we don’t know if a bNAb response alone will be sufficient or at what levels. The AMP results could help set direction and limit guesswork in developing a truly effective preventive HIV vaccine.

No matter what the AMP results are, critical thinking and dialogue will be essential among all stakeholder in order to understand new findings and move ever forward.

Resources

• The Rise of Broadly Neutralizing Antibodies, AVAC blog
• AMP Study FAQ, HVTN & HPTN
• What is HVTN 703/HPTN 081?
• What is HVTN 704/HPTN 085?
• Broadly Neutralizing Antibody Combinations, AVAC infographic
• HIV-Specific Neutralizing Antibodies - Targets and research status, AVAC infographic