Huge diversity in current HIV vaccine research, Research for Prevention conference hears

The HIV vaccine research field is currently going through probably its most fertile and diverse period yet, the second HIV Research for Prevention conference (HIVR4P) in Chicago heard last week. A high proportion of presentations there were devoted to the multiplicity of different approaches scientists are taking towards making an effective a vaccine.

At the opening plenary, Georgia Tomaras of Duke University in North Carolina, USA gave an overview of the field. It has been a long journey towards developing vaccines with even partial efficacy: the first trial of any kind was in 1987 and the first large efficacy trial – which failed – was in 2003. But the RV144 vaccine, which in 2009 showed limited efficacy, reducing infections in recipients by 31%, injected new energy into the field. This was not least because its effect seemed due to an unexpected kind of anti-HIV response.

The International AIDS Conference (AIDS 2016) held in Durban in July heard that a pilot study, HVTN100, of an RV144-type vaccine adapted to the strain predominant in South Africa had shown evidence that it produced a stronger response than the RV144 vaccine. This meant it had passed the criteria for being advanced to a large efficacy trial, HVTN 702. This will start next month – the first efficacy trial for seven years, since HVTN 505 started in July 2009.

How immune responses always surprise us

As if to illustrate the surprises the immune system springs on vaccine researchers, however, it was found that the immune response seen in HVTN100 included a type not seen in RV144. In that trial, antibodies of the type called immunoglobulin G (IgG) seemed protective, working by stimulating antibodies that did not themselves neutralise HIV but induced the immune system to mount a generalised attack on it. A second type of antibody, of the immunoglobulin A (IgA) type, was positively unhelpful, correlating with increased infection.

Now researchers have found a subtype of IgG called IgG3 that enhances immunity even more, reacting as it does to two specific parts of the HIV envelope protein’s ‘spikes’. This type was not seen in RV144 but was in HVTN 100 – a very positive sign. However, when RV144 recipients were revaccinated, given a shot of the vaccine up to six years after they initially received it, IgG3 reactions started to appear. With this happening here and with the cytomegalovirus (CMV) vaccine (see below), it implies that an effective HIV vaccine could involve eliciting HIV immunity initially in children, which could be enhanced with another shot in adolescence.

Tomaras emphasised the multi-dimensional complexity of the kind of immune response that researchers want to induce. There are three branches of the immune system, the antibody response, the cellular response and the innate response, and in each case these responses may vary by type, by strength, and by how broad they are (a vaccine would need to work against most different strains of HIV).

Vectors, primes and boosts

There are a large number of choices facing vaccine manufacturers. RV144 and subsequent studies show that an HIV vaccine may work best if given as two types of jab, a ‘prime’, usually of HIV genes encased in the empty shell of another virus, followed by a ‘boost’, usually of naked HIV proteins. So far these have been given in schedules of three to four doses over a year, though clearly a vaccine that worked with fewer doses would be better.

The pharmaceutical company Janssen, a subsidiary of Johnson & Johnson, has an ambitious research programme to develop a prime-boost vaccine that would act against a wide variety of HIV subtypes. The company’s Hanneke Schuitermaker told the conference that Janssen gives two or more prime doses of a ‘mosaic antigen’ made up of gene sequences from a wide variety of different HIV viruses in a viral shell (usually an adenovirus similar to the common cold) and follows it with two boost doses of HIV envelope protein. So far their best vaccine regimen has produced a 94% reduction in infections from one viral challenge in monkeys, and 66% from six viral challenges.

Janssen has a trial called APPROACH, of seven different vaccine regimens, underway in 400 volunteers in the US, Rwanda, Uganda, South Africa and Thailand. There are also two other trials called TRAVERSE and ASCENT with 348 volunteers between them which compare two different HIV mosaics. These are pilot trials and, as with HVTN100, if immune responses look good by the end of this year from APPROACH, and by mid-2017 from APPROACH and TRAVERSE, decisions will be taken as to whether to advance a vaccine into a larger efficacy trial, and if so which one.

Making the body make antibodies

The vector-prime-and-boost strategy is by no mean the universal one. Some vaccines prime with naked HIV DNA, which would be easier and cheaper to make than a vector; the challenge here is to get it into cells. Other vaccines dispense with the prime/boost strategy altogether.

A major problem with antibodies is that although the body does produce, in many cases, highly adapted and unusual antibodies that are ‘broadly neutralising’ and do eliminate the viruses that infect people (and can therefore be used as a kind of high-tech, long-lasting pre-exposure prophylaxis [PrEP] if given by infusion), once in the body, HIV develops immunity to them within weeks, staying ahead in an arms race which it usually wins.

Dennis Burton of the Scripps institute in La Jolla, California and Stephen Kent of Melbourne University both talked in separate presentations about developing more immunogenic versions of the HIV envelop proteins that more closely mimic what the immune system actually ‘sees’ in an infection. This has only been possible in recent years by using micro-manipulation techniques to get the gp120 surface protein to actually fold up into pairs and trios of protein called dimers and trimers, which is what the ‘knobs’ on the surface of HIV actually look like, instead of using unfolded HIV envelope proteins.

Researchers have induced broadly neutralising antibody responses by using these techniques, but the antibodies show a frustrating tendency to react to the ‘wrong’ bits of the envelope protein, keying into ‘holes’ in the shield of sugar molecules that surround the protein rather than directly evolving to attach to the protein itself. But getting cells to produce broadly neutralising antibodies at all in response to a vaccine is a major advance.

One way round this is to alert the third, most non-specific but fastest branch of the immune system, the innate immune response. This does not specifically recognise HIV. But antibodies that are induced by vaccines like RV144 in turn alert cells of the innate immune system. These so-called ‘natural killer' (NK) cells recognise general characteristics of infected cells and send cell-killing chemicals around to neutralise them. Researchers found that NK cells responded to a higher proportion of people immunised with the HVTN100 rather than the RV144 vaccine – though when, six years after their first vaccine, RV144 recipients received a renewed dose, their systems reacted this way too.

Another strategy is to get ahead of the arms race and make a vaccine that ‘looks’, to the immune system, like one of the adapted viruses seen in chronic infection, thus inducing the body to make broadly neutralising antibodies without HIV infection.

Christopher Parks of the International AIDS Vaccine Initiative (IAVI) introduced programmes to develop virus-like particles that are ‘live vectors’ that spread from cell to cell and deliver what looks, to the immune system, like an ongoing HIV infection. So far, they have used a benign virus called VSV (vesicular stomatitis virus) as the shell of their live vector. In monkey studies it has so far only retarded HIV infection slightly, though there is a correlation between weak antibody response and faster infection. But they are trying to develop virus-like articles with more ‘punch’ on them, including one coated with the HIV envelope protein itself and one even coated with envelop protein from the Ebola virus. As these are not actually viruses, they should be harmless, though clearly, safety studies will have to be done.

The vaccine no-one expected keeps springing surprises

Although since RV144 most researchers have been concentrating on raising antibodies, which should prevent infection in the first place, at least one unconventional vaccine based on a CMV virus vector aims to produce a strong response by immune cells themselves, which would not necessarily stop HIV infection but instead contain it or even put it into permanent remission.

This is the CMV vaccine produced by Louis Picker in Oregon, USA, which caused great excitement three years ago when it appeared to completely cure 55% of a group of monkeys given it. CMV is another ‘live vector’ that mimics the spread of a real HIV infection. The monkeys developed infections when challenged with SIV, the monkey equivalent of HIV, but in the responding monkeys their peak viral loads were very low and eventually their virus disappeared altogether. Picker said they have now reproduced the same results in many other monkeys. In addition, the effect seems quite persistent; when rechallenged with new SIV 2 to 6 years after their initial infection and 3 to 7 years after vaccination, 80% of originally-responding monkeys were still protected from SIV via vaginal challenge and 56% via rectal.

The CMV vaccine does not work by producing antibodies: instead it stimulates cells that kill infected cells, which is why it snuffs out infections rather than stopping them in the first place. When Picker’s team looked at how it worked, however, they found a completely new kind of immune response. Infected cells normally display bits of the viral proteins they contain on their surface, like distress flags. The ‘flagpoles’ consist of the tremendously variable stranger-recognition proteins called MHC 1a or MHC 2 (MHC stands for Major Histocompatibility Class). These, classically, alert CD4 cells to send out cell-killing and system-alert proteins and attract CD8 cells to kill the virus-infected cell directly.

The CMV vaccine, on the other hand, while producing MHC class 2 responses, also aroused a completely different kind of immune response using MHC class E molecules. These alert the innate immune system and natural killer cells. This response had never been seen before – either in a vaccine or, strangely, in natural CMV infection.

What the researchers found was that by happy accident the strain of CMV used to make the vector had two crucial genes missing. It was only when both these genes were deleted that CMV stimulated an MHC-E response. Furthermore, though natural MHC responses vary from animal to animal, there were a few responses to specific SIV genes that were the same in all animals – what one wants from a broad-brush vaccine.

Picker’s team are now developing vaccines using the human variety of CMV. Human CMV has genes that are analogues of the two missing genes in the rhesus monkey CMV, and the big question is whether the human vector will produce the same uniquely effective immune response signature. A dose escalation study is planned with a CMV vector that will produce a more conventional MHC-1a and 2 response, and if this is successful they will move on to stimulating an MHC-E response.

One big problem remains: the vaccine has consistently only worked in about 55% of monkeys and so far there does not seem to be a way of telling which monkeys will respond and which will not. Louis Picker told aidsmap.com that the difference is probably caused by how speedy the CMV vaccine response is and whether it is able to stop SIV/HIV infection before it becomes systemic and spreads throughout the body. What determines the speed is as yet unknown, but it may mean the CMV vaccine would have to be combined with an antibody vaccine for better efficacy. As with other vaccines, the difference between success and failure is all about whether the immune response can stay one step ahead of HIV in the arms race between virus and body.