Slow progress on HIV vaccines reported at CROI

This article is more than 17 years old. Click here for more recent articles on this topic

A symposium at the Conference on Retroviruses and Opportunistic Infections on HIV vaccine development served as a reminder of many unanswered questions about the prospects for effective vaccines. Leading US vaccine researchers focussed in turn on neutralising antibodies, viral vectors, cytotoxic T-lymphocytes, and the relevance of animal studies to results seen in human clinical trials.

In search of neutralising antibodies

Peter Kwong of the National Institutes of Health Vaccine Research Center reported on studies of neutralising antibodies from volunteers in clinical trials of a prime-boost vaccine system. This used a DNA primer followed by a recombinant Adenovirus type 5 (Ad5) booster, both containing the same HIV genes. Specifically, envelope sequences from viruses of subtypes (or clades) A, B and C, together with gag, pol and nef genes from subtype B. Most of the neutralising antibodies raised are directed against the V3 loop, on the outer envelope protein, gp120, where the V stands for ‘variable’ and hints at the problem, that these antibodies are normally strain specific and the virus can easily escape their effects.

Dr Kwong highlighted one of the more broadly active neutralising antibodies, called 2F5, directed against part of the V3 loop closest to the viral membrane on the HIV envelope protein. This antibody includes a region that binds to the viral membrane and he suggested this was important in enabling the antibody to access a relatively protected and constant part of the otherwise very variable loop structure. This was also why early attempts to raise such antibodies using vaccines based solely on fragments of the V3 loop itself failed, and could point the way to better vaccine design for broadly active antibodies.

What he didn’t say, but which needs to be kept in mind, is that the broadly active neutralising antibodies identified to date have not been particularly potent. It is therefore always necessary to ask about the quantity of an antibody that may be needed, and whether this can be achieved with a vaccine, as well as how active it is against different HIV strains.

Setting one virus against another


deoxyribonucleic acid (DNA)

The material in the nucleus of a cell where genetic information is stored.

neutralising antibody

An antibody that neutralises (renders harmless) an infectious microorganism.


A molecule on the surface of some white blood cells. Some of these cells can kill other cells that are infected with foreign organisms.

simian immunodeficiency virus (SIV)

An HIV-like virus that can infect monkeys and apes and can cause a disease similar to AIDS. Because HIV and simian immunodeficiency virus (SIV) are closely related viruses, researchers study SIV as a way to learn more about HIV. However, SIV cannot infect humans, and HIV cannot infect monkeys. 


One of the three proteins encoded within the retroviral genome.

Philip Johnson of the Children’s Hospital Research Foundation, Columbus, Ohio, reviewed the use of viral vectors as HIV vaccines. The idea is to take a vaccine against one virus (like Ad5) and include HIV genes, to get immune responses against HIV.

He observed that there is a "working" vaccine, in the form of live attenuated SIV (a monkey equivalent of HIV) which can sometimes protect against infection with SIV, although we still don’t understand how this happens. The rationale for viral vectors is that they're much safer than live SIV/HIV and can express high levels of multiple HIV genes. Vaccines based on Venezuelan Equine Encephalitis (VEE) can express higher levels of proteins than are seen in many natural infections. Modified Vaccinia Ankara (MVA, a smallpox vaccine) can be engineered to contain six different viral gene sequences. Recombinant Adeno-Associated Virus (rAAV) can elicit strong and enduring immune responses.

A number of viral vectors are now in clinical or pre-clinical development. There is a Phase III (full-scale efficacy) trial under way in Thailand, using Aventis Pasteur’s ALVAC canarypox vaccine. Phase I or Phase II trials are testing vaccines based on Ad5, MVA, VEE, fowlpox, NYVAC (a smallpox vaccine), and AAV. Other vectors are being studied in animals, including Semliki Forest Virus, Vesicular Stomatitis Virus, Sendai, Measles Virus, and Herpes Simplex Virus (HSV).

Pre-existing immunity to particular vectors has a substantial impact on Ad5 based vaccines, since Ad5 is a widespread human virus. Immunity to vectors also means recombinant vaccines cannot usually be given more than twice.

Spotting an effective HIV vaccine may be difficult. It's not at all clear whether current tests are measuring the right dimension of immune responses. Viral vectors offer no advantage for the induction of broadly reactive neutralising antibodies. It is the antigen – especially the HIV-related components - which will determine that.

There are also many manufacturing and regulatory issues about genetic instability, low yields, the acceptability of cell substrates, questions about replication competent vectors, and possible virulence of some vaccine strains, which have been problems for vector-based systems.

VEE, Sindbis, VSV, fowlpox and others can avoid the problem of pre-existing immunity because these viruses do not circulate among people. Ad35 is much less prevalent than Ad5, although we need to know it gets similar immune responses before the adenovirus-based vaccine effort goes in that direction. It may be possible to blunt the immune response to a vector by modifying its outer proteins, although this may also risk losing the properties that made the vector useful in the first place.

DNA plus vector worked effectively in monkeys as a way to overcome pre-existing immunity, but there seems to have been no such effect from using a DNA prime in human studies with Ad5. DNA priming simply does not work as well in humans as in monkeys, let alone in mice.

Two different vectors could be used, as in the Merck - Aventis Phase I trial of Ad5 and canarypox. This worked well in monkeys but, in human trials, a canarypox booster was no better than a second dose of Ad5.

Another possibility would be to use a viral vector to transfer genes for a neutralising antibody, so that muscle cells produce neutralising antibodies against HIV. Long-lasting production of human antibodies against HIV has been achieved in animals and should be investigated further.

Stimulating T Cells

Juliana McElrath of the Fred Hutchinson Cancer Research Center and University of Washington, Seattle, spoke about HIV vaccines that aim for cellular immune responses to HIV-infected cells. The hope is that Cytotoxic T-Lymphocytes (CTLs), also known as CD8+ or killer T-cells, and CD4+ T-cells (also known as helper T-cells) can either abort or control infection to prevent disease.

Responses to HIV by CD8+ and CD4+ T-cells depend on recognising peptides – small fragments of HIV proteins - called T-cell epitopes, displayed on the surface of infected cells. A CD8+ cell which recognises the particular combination of epitope and human protein it responds to may then kill that cell. A CD4+ cell which recognises its unique combination of epitope and human protein will send signals using substances called cytokines to alert other cells to the need to respond, and switch on the ability of CTLs to kill infected cells.

Measurement of CTL responses typically look for cytokine responses by cells to pools of epitopes spanning the viral proteins that researchers are interested in.

A number of approaches to stimulating CD8+ responses have now been tested in clinical trials, among which she focussed on two, namely the trials sponsored by the International AIDS Vaccine Initiative (IAVI) in London and Oxford, as reported at the AIDS Vaccine 2004 meeting in Lausanne, and a study sponsored by the US government-supported HIV Vaccine Trials Network (HVTN) in partnership with Merck.

The results of the London and Oxford trials of DNA followed by recombinant MVA showed approximately 20% of volunteers had at least one response to one pool of peptides but observed responses were not sustained.

The HVTN/Merck study which is being conducted at four sites in Thailand, Caribbean, South America and the USA, involves three immunisations on day 1, week 4 and week 26. Some strong responses have already been seen in particular volunteers, in terms of HIV-specific CD4 and CD8 cells which seem to be functional and to be persisting for at least one year.

Initial studies of a trivalent Ad5 gag/pol/nef vaccine are finding that vaccine is immunogenic in around 70% of volunteers at the highest dose given, when there is no pre-existing immunity to Ad5. The majority of volunteers are showing responses to multiple epitopes from HIV.

There are now 13 trials within the HVTN alone and ten more due to start in the next year involving a variety of vaccines designed to stimulate CTLs.

There are still many unanswered questions about the properties of CTLs needed for a protective effect. For example, should the vaccines target or avoid viral epitopes that are likely to escape? High avidity (tight-binding) CTLs may put more pressure on the virus to escape, but escape can occur with low avidity CTLs too. Escape from CTL control may sometimes reduce viral fitness, and these epitopes may be those most worth targeting.

It may be that the really important property of cellular immune responses is the functionality of the CTLs – their ability to multiply when they encounter their target and to destroy infected cells.

The bottom line is that the mere presence of T cells does not ensure protection; we need to know what they can and cannot do.

A question from the audience was whether proliferating CD4+ cells are truly useful or just generate more targets for the virus to infect. Dr McElrath thought that while there might be a risk of providing the virus with more vulnerable cells, it was probably more important to have an HIV-specific CD4+ response, to maximise the effectiveness of the CD8+ response.

Primate studies

Norman Letvin of Harvard University asked how predictive nonhuman primate studies are likely to be.

Vaccine immunogenicity was discussed with reference to NIH Vaccine Research Center vaccine candidate which includes subtype (or clade) A, B, C envelope genes and subtype B gag/pol/nef genes in a DNA/recombinant Ad5 prime-boost system.

In rhesus monkeys, there were broad and strong CTL responses to the DNA vaccines, both for the envelope and for the gag/pol/nef components. However, in human trials, there were excellent responses to the env antigens but not to the gag/pol/nef antigen. Studies in cynomolgus monkeys found similar results to the human clinical trials. The obvious conclusion is that all monkeys are not the same.

Using a new promoter sequence from cytomegalovirus to change the way the viral genes were expressed, it was possible to get much greater immune responses in the cynomolgus monkeys. Modified DNA vaccines of this kind are now being taken forwards into human trials.

Dr Letvin went on to discuss the Oxford and London studies referred to by Dr McElrath.

Initial animal studies with a candidate DNA/MVA vaccine found strong CD8+ responses in monkeys to epitopes from SIV.

However, in clinical trials - as previously mentioned - the immune responses to HIV were disappointing.

Subsequent studies of the HIV immunogens in monkeys found they were almost as poor in the monkeys.

His conclusion was that the HIV immunogens that were marginally immunogenic in humans were also marginally immunogenic in monkeys, so it could actually be worth evaluating prototype HIV immunogens in monkeys before going to human trials.

What kind of protection against viral challenge can we expect from vaccines that produce CTL responses?

Dr Letvin reported a study of DNA vaccines including elements of SIVmac239 gag, HIV-1 89.6P env, immunised at weeks 0, 4, 8, 40, followed by an intravenous challenge using the highly pathogenic SHIV-89.6P at week 46.

Viral RNA was monitored for 800 days after challenge. There was a striking difference between the survival of vaccinated and unvaccinated monkeys. In the monkeys which were vaccinated and appeared to be protected, the viral RNA became undetectable soon after challenge and remained so.

The biology of the pathogenic SHIV/macaque model is very different from HIV and also SIV, in that it it infects T-cells but not macrophages, and infection leads to rapid, near-complete loss of circulating CD4+ T cells.

Despite this difference, studies of T cell based vaccines in SIV/macaque model were able to find similar levels of protection to those observed in the SHIV/macaque model.

Two weeks after challenge vaccinated animals had statistically significant reduction in viral load compared to recipients of sham vaccine. Vaccinated animals survived while those which received a sham vaccine became ill and died.

Dr Letvin’s conclusion was therefore upbeat, taking the view that there is indeed a reasonable chance that a vaccine inducing CTLs against HIV can be clinically effective.


Recent Advances in HIV Vaccine Development Symposium Twelfth Conference on Retroviruses and Opportunistic Infections, Boston, 2005.

Kwong PD. Hide and Seek: Evasion and Exposure of the HIV Envelope. Twelfth Conference on Retroviruses and Opportunistic Infections, Boston, 2005.

Johnson PR. Viral Vectors as HIV Vaccines: Lessons Learned and Future Prospects. Twelfth Conference on Retroviruses and Opportunistic Infections, Boston, 2005.

McElrath MJ. CTLs: All T-cells Are Not Created Equal. Twelfth Conference on Retroviruses and Opportunistic Infections, Boston, 2005.

Letvin N. Nonhuman. Primate HIV Vaccine Studies: Will They Be Predictive? Twelfth Conference on Retroviruses and Opportunistic Infections, Boston, 2005.

Viewed as a webcast at (and follow links for Thursday, 4pm)