News

The Lancet last Friday published the 52-week results of APPROACH, an HIV vaccine whose 28-week results, reported by aidsmap.com last year, were sufficiently impressive for it to be taken forward into a human efficacy trial among 2600 young women in southern Africa. 

These latest results provided more details of the immune responses to the vaccine, which is already being taken forward into the phase 2b efficacy trial, the Imbokodo or HVTN705 trial (this is the second efficacy trial now to be happening: the HVTN 702 trial started at the end of 2016).

What has excited special interest with APPROACH is that a parallel trial conducted in rhesus monkeys, with a vaccine designed to closely mimic the human one, and which produced immune responses similar to those in humans, protected two-thirds (67%) of monkeys challenged with six rectal doses of a highly pathogenic Simian Immunodeficiency Virus analogous to HIV. The efficacy against each single challenge was 94%.

It’s important to emphasise that the immune responses produced by both the human and the monkey vaccines tailed off over time (though somewhat more slowly in humans) and that we do not know how long the immunity produced by this vaccine, which requires four doses spaced over a year, may last. Previous vaccine concepts have produced promising results in animals but have failed in humans. But the APPROACH study at least suggests that we are getting nearer to having a vaccine with useful efficacy.

APPROACH was conducted by professors Dan Barouch, Frank Tomaka and Frank Wegmann and colleagues at Harvard Medical School and at Janssen, the research branch of the drug company Johnson & Johnson, with the collaboration of a number of other institutes. It involved three differently formulated vaccines that were given in seven different permutations to groups of 50 volunteers each. An eighth group received a placebo-only regimen. 

The vaccine consisted of three components of HIV. Two were lengths of RNA from the env and gag/pol genes in RNA’s genome, its ‘instructions’ that induce human cells to make more copies of itself. The third was a part of the complete viral particle, gp140, the mushroom-shaped protein that dots the viral surface and acts as HIV’s tether and molecular 'key' that gets it into human cells.

HIV’s genome has huge variability. This has been an Achilles heel of previous vaccines, as they have only worked against a narrow range of viruses and HIV finds it easy to escape the immune response generated by the vaccine. The RNA in the APPROACH vaccine consisted of a mosaic of many different genomic sequences from different types of HIV, designed to stimulate immune responses against a broad range of viruses.

The env and gag/pol gene sequences were enclosed inside ‘vectors’ – the outer shells of two quite different viruses called Ad26 (adenovirus 26) and MVA (modified vaccinia ankara). These vectors can get inside human cells and mimic an infection, which includes cells flagging up that they are infected and so stimulating an immune response. Though these vectors cannot replicate and cause ongoing infection, in the event of a real HIV infection, the immune 'memories' they produce should recognise the distress signals given off by HIV-infected cells and kill them.

The third vaccine component was pure gp140 with the mineral aluminium phosphate added to it as an adjuvant (immune stimulant). Protein vaccines like this are designed to further stimulate the humoral or antibody branch of the immune system. Antibodies are protein molecules generated by the immune system that either destroy viruses directly (‘neutralise’ them) or stick to the surface of virus-infected cells and flag them up for destruction by immune system cells, in a process called antibody-dependent cellular phagocytosis or ADCP.

The vaccine as given as four shots at zero, 12, 24 and 48 weeks after the start of the study. The vector vaccines were given at weeks zero and 12 as ‘primes’, and the gp140 at weeks 24 and 48 as 'boosts', in either high or low doses. This prime-boost design aims at generating a stronger and more varied suite of responses from different branches of the immune system.

There were eight different vaccine schedules. As well as a placebo-only arm, there were two arms without boost doses of gp140 and one arm that only had one vaccine prime dose (of ad26).

There were 47-50 people in each arm. By week 52, after various dropouts, there were 41-47 people in each arm; 394 people entered the study between February and October 2015, and there were 349 by week 52 a year later, a 10% discontinuation rate.

APPROACH was an international study, taking place in the US, Rwanda, Uganda, South Africa and Thailand in people defined as at low risk of HIV. Thirty-eight per cent of participants were from the US, 33% from Rwanda or Uganda, 15% from Thailand and 14% from South Africa. Fifty-four per cent were men and the average age was 29. Despite participants being defined as low-risk, three new HIV infections occurred, all at a single site in South Africa, the highest-prevalence country.

There were some side-effects reported. For instance, 69 to 88% of vaccine recipients (depending on what group they were in) reported pain at the injection site within the first week after injection, compared with 49% in the placebo-only group. In other words, 1.4 to 1.8 times as many people receiving vaccine reported pain than those receiving placebo.

Other commonly reported side-effects included mild-to-moderate headache, fatigue and muscle aches – not uncommon in other vaccines too. Five participants reported severe side-effects including abdominal pain and diarrhoea in one, raised liver enzymes in another, and in the other three dizziness, back pain and malaise. One person reported a possible allergic reaction and was removed from the study .

In terms of the immune response, to cut to the ‘reveal’: the ad26/ad26/high-dose gp140/high-dose gp140 regimen produced the strongest immune reactions and will be the one used in the Imbokodo trial. The high-boost regime with a gp26 and then an MVA prime generated immune reactions that were just as strong in humans but were slightly less strongly associated with efficacy in the monkey study; this, plus the ease of manufacturing one primer rather than two, ruled it out.

This vaccine produced an antibody response to HIV in 100% of recipients by week 12, and an ADCP response in 80% of recipients by week 52. It also produced CD4/CD8 responses in 83% of recipients; so both branches of the immune system were stimulated in the majority of recipients. The antibody responses showed activity to a panel of HIV viruses taken from different people at different stages of infection, showing that the responses generated by the mosaic design were broad. 

The researchers noted that the vaccine did not generate a strong neutralising-antibody response. This was not unexpected; it’s known that the so-called broadly neutralising antibodies (bNAbs) that can by themselves annihilate a wide range of HIV are rare and exotic molecules (and the subject of study as drugs to cure or prevent HIV in their own right). 

Instead, the antibody responses seen recruit other branches of the immune system to kill off HIV-infected cells, as predicted – ADCP, in other words.

There have been other studies that generated what appear to be strong immune responses in animals but which haven’t worked in humans. So it’s important that a vaccine designed to be as close as possible to the human one was also given to 72 rhesus monkeys. The same two primes were used, though those who also received boosts all received the same dose, so the vaccine regimen was not split into high- or low-dose boosts. There were six different vaccine groups of 12 monkeys each.

The monkeys developed immune responses of a similar breadth and magnitude to the human volunteers, as measured at weeks 54-56, which was when they received their last booster dose. Twenty-two weeks after their last dose, when the monkeys’ immune responses had declined somewhat, they were challenged with six successive weekly doses of a highly infectious laboratory-developed SIV/HIV virus, developed to be able to infect monkeys while being as close as possible to human strains. 

Two-thirds of the monkeys given placebo were infected after just one viral challenge and all by five challenges. In contrast, while one monkey receiving the most efficacious vaccine – also the ad26/ad26/gp140/gp140 regime – was infected after one challenge, 83% remained uninfected after four challenges and two-thirds after all six challenges, representing an overall efficacy of 67% of infections prevented over the challenge period compared with placebo, and a 94% reduction in the risk of infection per exposure.

The researchers looked at the immune profiles of the monkeys and developed a model of the immune responses that were the best correlates of protection, in other words that were most strongly associated with not being infected. The antibody and CD4/CD8 responses at week 28 correlated most strongly with protection. 

The strength and rapidity of immune responses to specific regimens in humans and monkeys was compared – with good news for humans. The antibody responses in both humans and monkeys correlated with the type of vaccine regimen, but declined more slowly in humans than in monkeys. The CD4/CD8 responses were not quite so well correlated; for instance, the ad26/MVA prime regimens produced the best response in monkeys than in humans; but the humans’ immune systems seemed to respond more quickly, with significant responses after the third dose, but only after the fourth dose in monkeys. Longer-term data on the immune response and how fast it declines will be presented at the forthcoming International AIDS Conference in Amsterdam in two weeks’ time.

In an accompanying editorial, George Pavlakis and Barbara Felber of the US National Cancer Institute commented that the approach of vaccinating humans and monkeys in parallel and then challenging the monkeys was an interesting one.

Even if the efficacy seen in monkeys is not matched in humans (or declines more quickly), comparing what was protective in monkeys with humans would help scientists develop good correlates of infection and predictors of efficacy in vaccine candidates, a field which for too long had had to rely on best guesses about what should work. Pavlakis and Felber also praised the strategy of testing multiple candidates in parallel in different studies, and in different arms of the same study.

Barouch DH et al. Evaluation of a mosaic HIV-1 vaccine in a multicentre, randomised, double-blind, placebo-controlled, phase 1/2a clinical trial (APPROACH) and in rhesus monkeys (NHP 13-19).The Lancet, early online publication, 6 July 2018. http://dx.doi.org/10.1016/S0140-6736(18)31364-3.

Pavlakis GN and Felber BK. A new step towards an HIV/AIDS vaccine.The Lancet, early online publication, 6 July 2018. http://dx.doi.org/10.1016/S0140-6736(18)31548-4.