What an HIV vaccine would have to do

A vaccine is essentially a ‘fake infection’. It is a way of priming the body by getting it to mount an immune response to essentially harmless microbes – or to parts of microbes called antigens – so that these immune responses work faster and with more potency against subsequent infection with a similar but disease-causing microbe. The principle has not essentially changed since Robert Jenner observed that dairymaids exposed to the relatively harmless cowpox virus (though he did not know it was a virus then) were later immune to the ravages of the smallpox virus.

Vaccination happens all the time naturally, in the spirit of Nietzsche’s saying “That which does not kill us, makes us stronger”. Malaria, for instance, is a particularly tricky infection because, like HIV, it constantly changes its shape in order to fool the immune system. However children in Africa who do not die of repeated malaria infections within their first three years will eventually develop a broad-enough immune response to malaria to either repel further infections or develop only mild symptoms.

One theory as to why allergies like asthma are so much more common in the modern world is the so-called hygiene hypothesis. This states that children these days are not exposed to enough allergens and germs when they are young. As a result, their immune system does not ‘learn’ to respond appropriately to certain foreign substances and mounts a disproportionate response when it finally encounters them.

Vaccines set in motion an immune response the body would mount against the dangerous pathogen (disease-causing organism) anyway.

Many diseases kill not because the body mounts no fight against them but because there is always a timelag between an invasion by a previously unknown infection and the immune system learning how to fight it.

In a few cases the invader will win and kill or cripple – either by directly causing damage before the immune system can stop it, or by generating an immune response so extreme that it starts to damage the body’s own cells (this is what is thought to happen in illnesses like SARS and bird flu, and it is also the cause of the liver damage in chronic hepatitis B infection).

In most cases the immune system will eventually win and the invader will be driven out. What the vaccines do to most diseases is prime the immune system to an invader so that when it eventually arrives, it is already ‘known’ to the immune system and there is a much shorter timelag between infection and the generation of an effective immune response.

Vaccines do this in the same way that infections do. An antigen is any foreign protein or protein component that causes an immune reaction in the body. Once an antigen of any sort – a bacterium, a virus, a parasite, even certain chemicals, drugs and dust – enters the body for the first time, the immune system sets about devising an immune response that will, in future, neutralise this invader. A vaccine is an antigen designed to create a very specific response. Antigens set off three different types of immune response:

1. Innate immunity. This, evolutionarily the most primitive part of the immune system, is a set of chemicals that recognise and neutralise common foreign protein sequences non-specifically. However it does not ‘remember’ infections and so cannot be used as the basis for a conventional vaccine.

2. Adaptive immunity, which is subdivided into:

2a. Humoral immunity. This comprises a set of free-floating proteins called antibodies that link to proteins from specific invaders, and when they reappear either chemically neutralise them or tag them for destruction. Antibodies are generated by B-cells and are immensely variable molecules that have the capacity to ‘remember’ infections.

2b. Cellular immunity. This is a set of roving cells that destroy infected cells by recognising bits of foreign antigen, called epitopes, that cells display on their surfaces. Much of the time these cells exist in embryo form but at times when the body is under attack by one or more pathogens they differentiate – first into T-cells, then into CD4 cells (which orchestrate the immune response) and CD8 cells (which do the actual cell-killing) and finally into memory cells.

In the case of both humoral and cellular responses, the initial attack leaves behind a few memory cells. These are cells that have ‘learned’ the signature of the invader so that when the same one (or apparently the same one) turns up again, the immune system can spring into action much faster and contain an infection before it has time to do damage.

It is this memory effect that vaccines exploit, and the goal of an HIV vaccine would be to produce enough broadly effective memory B-cells (which make antibodies) and/or T-cells (which direct and operate the cell-killing mechanism) to recognise any strain of HIV when it arrives and quickly neutralise it.

HIV has one unique property that has made cellular vaccines especially difficult to develop. The central-memory T-cells, which mainly live in the lymph nodes and other internal immune system sites, are precisely the ones that HIV infects and the ones within which it establishes a ‘reservoir’ of integrated viral DNA. Vaccines that stimulate this fraction of the immune system to proliferate on encountering HIV might therefore be acting too late, or could even only be creating a better environment for the virus. This is why mucosal vaccines or ones that stimulate the effector-memory T-cells (which mainly circulate outside the immune sanctuaries and patrol mucous surfaces) might work better.

The body does mount an immune response to HIV – indeed, without one the virus would destroy the average person’s immune system within weeks rather than years.

However in the case of HIV infection the immune response is sufficient neither to prevent infection in the first place nor to prevent the virus circumventing the body’s immune defences in the long run.

An HIV vaccine, therefore, would have to do ‘better than nature’ – and that is why it has proven so difficult to develop.

An HIV vaccine would have to do one of three things, which we will explore in more detail.

Humoral immunity is the type that involves antibodies. These are extremely variable Y-shaped protein molecules that are produced in huge quantities by the B-cells of the immune system. They either destroy invading microbes themselves or tag them for destruction by other components of the immune system.

An antibody response is, as far as we know, the only specific response that can generate so-called sterilising immunity - the complete prevention of an infection. Other responses (see cellular immunity) may moderate, rather than prevent one.

If the invader is one the body already recognises, an antibody response can be generated so fast that an infection never becomes established. If it is not recognised, it may take some time for enough antibodies that ‘fit’ the invader to be generated.

Some vaccinations, so-called passive ones, actually consist of antibodies rather than antigens that generate an antibody response. Passive inoculation with anti-hepatitis B antibodies, for instance, is used to strengthen the immune response and augment the regular hepatitis B vaccine, especially in cases where exposure may have already happened, as in a needlestick injury. However passive inoculation is similar to using a drug – the antibodies quickly disappear from the body and no permanent immunity is generated.

Antibodies generally only recognise the surface molecules of bacteria, viruses, parasites etc. The first generation of candidate HIV vaccines, therefore, used this principle. They consisted of parts of HIV’s envelope – the outer viral covering. In particular, they used the gp120 protein that forms the ‘knobs’ on the surface of HIV that are the virus’s mechanism for entering cells.

Initial hopes that a simple envelope vaccine might work were dented when the AIDSVAX vaccine trial proved ineffective in February 2003¹ and dashed when the second AIDSVAX trial in Thailand proved equally ineffective two years later.² However they received an unexpected revival when the RV144 trial announced its results in 2008. The RV144 trial involved using two vaccines – an initial shot of an adenovirus vaccine similar to the one used in the STEP trial, and a booster of the same gp120 vaccine used in the AIDSVAX trials. Although it is still uncertain why this vaccine exerted some protective effect when others have not, the immune response create by the vaccine appeared to be humoral rather than a cellular one. Simple antibody-stimulating vaccines may still, then, have a place in HIV vaccine research, at least as a component of ‘combination’ vaccines.

Why did the AIDSVAX vaccines not work? The answer lies in the hyper-variability of the HIV envelope.

The gp120 protein, and in particular the part of the molecule called the V3 loop that actually makes contact with cellular receptors, is the most variable part of HIV. Not only is the amino acid sequence that makes up the core chain of the protein more variable than any other part of HIV, it is also heavily glycosylated. This means that HIV, as it evolves, coats its envelope protein with a sticky ‘fuzz’ of sugar molecules that frustrate the attempts of antibodies to latch on to it.

What this means, essentially, is that an HIV envelope vaccine would produce an antibody response – but only one that worked against the exact strain of virus that the vaccine was developed from, or imitated. The first generation of vaccines did not work because they, and the antibodies they elicited, were uselessly specific.

HIV does have highly ‘conserved’ regions. These are areas of the viral genome and of viral proteins that are forced to stay evolutionarily stable because they are essential to its core tasks. These regions vary little from one virus to the next. However in the case of HIV they are also areas that evolution has taken great care to guard. An example is the fusion peptide of the gp41 viral protein, the part of the HIV ‘spike’ that changes shape so that it can inject HIV’s genetic material into the cell. This part of the virus is only exposed for a fraction of a second in the sequence of events that comprise infection so antibodies have to be extremely potent and rapid in their action to neutralise these conserved regions.

Potent broadly neutralising antibodies exist, discovered by painstaking searches through samples donated by long-term non-progressors and other people with HIV worldwide. A central part of HIV vaccine development, including the Antibody Project sponsored by the International AIDS Vaccine Initiative (IAVI) now concerns discovering vaccines that can stimulate the production of broadly neutralising antibodies consistently, and at levels capable of generating sterilising immunity, in the majority of vaccinated individuals.

Broadly neutralising antibodies are rare; the first was not isolated from a person’s blood and described till 2001.³

However, only a handful have been found to neutralise a large majority of strains of HIV. They are also not typical of most antibodies. Molecular analysis and crystallography has found that they tend to have unusual structures featuring molecular ‘spikes’ that can penetrate into the normally well-concealed conserved areas of HIV’s envelope. So far only a few have been isolated from the blood of exposed seronegative individuals. One study in 2004⁴ found that just one antibody, 4E10, neutralised every one of a panel of 90 HIV strains with moderate potency. One called 2F5 neutralized 67% of isolates, but none from clade C of HIV, the most common type in Africa. An antibody called b12 neutralised 50% of strains, including some from almost every clade, while one called 2G12 neutralised 41% of the strains, but none from clades C or E. A second study from Africa in 2006⁵ found two more broadly neutralising antibodies.

Experiments with these antibodies have so far mainly involved using them as passive inoculations and studying how they are eliminated in the body. Here they act more like potential long-lasting anti-HIV drugs, as they are eliminated from the body over a timescale of one to three weeks. Some artificially created antibodies such as the experimental drug ibaluzimab (TNX-355) use the same principle.

Developing a vaccine which induces the body to generate them will be much more difficult. It may be that some people’s B-cells are simply incapable of manufacturing such proteins. Because these antibodies act against parts of the viral infection mechanism that are only exposed for a fraction of a second during the intricate unfolding and insertion process that happens during the infection of a cell, it is challenging to establish what epitopes could elicit such antibodies in the average person, and so far we are nowhere near achieving this feat. Another problem is that while the antibodies do neutralise HIV at the moment of fusing - when the gp160 molecule on the surface of HIV splits apart to reveal its inner workings, including the gp41 fusion protein - in intact viral particles, the gp160 molecule exists as a heavily glycosylated trimer (triplet) of molecules to which the antibodies have no access.

An alternative approach, since finding epitopes that elicit broadly neutralising antibodies for everyone may be impossible, is to construct genetically engineered bacteria that express these antibody molecules.

The other thing a vaccine could do is delay or halt the damage that an established infection can do. It would do this by stimulating the other branch of the immune system – the cellular immunity.

The prime movers in the cellular immune system are the cytotoxic T-lymphocytes (CTLs), otherwise known as the CD8 cells. This branch of the immune system developed during evolution to deal with the problem that once a virus is inside a cell, it is essentially invisible to the humoral immune system.

However cells have a mechanism whereby they ‘advertise’ their contents by displaying tiny fragments of their internal constituents, called epitopes, on their surface. This is the way the body distinguishes between self and not-self, as well as between healthy cells and ones subverted into virus-making factories.

When the immune system senses the presence of foreign epitopes, a cascade of immune activation is generated which ends with the CTL cells destroying the infected cell.

The advantage of this kind of immunity is that the cell displays protein fragments from all parts of the invading virus and not just its envelope. In the case of HIV, this means that an immune response can be generated against deeper, more conserved parts of HIV.

The disadvantage of the cellular immune response is that it does not prevent an infection, but acts against already infected cells.

In most illnesses, this does not matter; the cellular immune response wipes the body clean of infected cells and the disease is gone. Serious damage only occurs if so many cells are infected that the immune response itself becomes harmful.

However in the case of retroviruses like HIV and HTLV, the virus becomes incorporated into the cell’s genetic code itself – as proviral DNA.

By the time this has happened, the virus has essentially lost its identity as an independent entity and become so much part of the cell that it is not recognised as foreign. It is only when the cell is activated and starts producing new viruses that the immune system can recognise it as infected.

For this reason, a vaccine that generated cellular immunity could have immensely variable effects depending on whether it acted in time to prevent the incorporation of HIV’s genes into the human cells’ DNA.

At best, it might be able to turn people into exposed seronegatives. In the natural history of HIV, these are people who remain HIV antibody-negative but where extremely sensitive tests detect signs of a historical infection by HIV. This infection remains so well-contained that not enough virus is ever present to trip the humoral immune response and induce antibodies to HIV. There is evidence that exposed seronegativity is quite common, but the majority of exposed seronegatives remain little-studied and we do not know how many there are and why they do not develop HIV infection. Even though exposed seronegatives do not have antibodies to HIV, immune experiments showed that their T-cells ‘recognise’ HIV in the test tube – so they must have seen it before.⁶

What appears to be the case with most of them, however, is that by good luck, good genes or good timing, their immune system developed a CD8 response against actively infected cells so efficient that it nipped any productive viral infection in the bud – exactly what we hope a truly efficient CD8 vaccine could do.

Some examples of this phenomenon were presented in a study at the Bangkok International AIDS Conference in 2004,⁷ which investigated the long-term exposed seronegative partners of HIV-positive gay men. The group consisted of the HIV-negative partners of HIV-positive men who had been diagnosed between 1994 and 1998.

Out of 94 HIV-negative regular partners of positive men, two were found who appeared to have in fact caught HIV from their partners early on in their relationship, but to have mounted a successful immune response to it. They had no antibodies to HIV and therefore did not test HIV positive. The fact that they had HIV at all could only be detected by hypersensitive viral-load testing, which picked up HIV in their blood at a count of 0.05 copies – one thousandth of the amount described as "undetectable" by standard tests.

However no CD8 vaccine has come anywhere close to producing an immune response to HIV as effective as this, and the exposed seronegatives remain with their immunity secrets tantalisingly elusive (it appears now that they may also generate broadly neutralising antibodies). Some CD8 vaccines have, at least in animal studies, blunted subsequent HIV infection by reducing the viral load in those infected, but this has not so far been observed in human studies.

There is a third kind of immunity a vaccine might be able to generate, but it is not one that previous vaccines have attempted to stimulate. This refers to humoral or cellular immune responses that are concentrated at the mucosal surfaces where most HIV transmission takes place, such as the vagina and rectum. Vaccines may be able to induce immune responses acting only at these surfaces, to prevent HIV transmission through sex. They would not work against infection through injection, but since the majority of HIV in the world is spread through sex, they would potentially contain the epidemic.

What would a mucosal vaccine look like? It might look a lot more like a microbicide than a vaccine, though it would be one that generated an immune response. It does not take a big leap of science to move from the idea of a microbicide that would work by getting genetically altered versions of natural gut and genital bacteria to produce microbicidal substances like cyanovirin-N (see Microbicides) to getting genetically altered bacteria to produce HIV antigens that would then generate an immune response.