x x






Dr Richard Hunt
Department of Pathology, Microbiology and Immunology
University of South Carolina School of Medicine

En Espaņol
Let us know what you think





AIDS Vaccines

AIDS Vaccines (from NIAID)


Since the development of an effective vaccine against smallpox, many diseases caused by viruses are no longer a major public health problem or their incidence has been greatly reduced. Initially, when the causative agent for AIDS was discovered to be a virus, it was hoped that a protective vaccine would soon be available. But the AIDS epidemic has been with us for more than twenty five years and we still have no effective vaccine, despite many clinical trials. During this time highly active anti-retroviral therapy (HAART) has made AIDS a tractable disease in many but not all patients, but as we have seen there is little hope that a drug regimen can cure the patient of AIDS. A sterilizing (neutralizing) vaccine, as in other viral diseases, is the best prospect to stop HIV infection


Lessons from the past from vaccines against other viruses

  • Many vaccines stimulate immunologic memory but do not block infection

  • Most successful vaccines induce potent antibody production

  • Most vaccines were developed empirically

  • The best vaccines involve live attenuated or whole killed virus

Goals for an anti-HIV vaccine

1. It must provide sterilizing immunity. It will need to act against free virus and HIV-infected cells. Humoral responses to viral infection are usually narrower than cytotoxic T lymphocyte responses

2. If this cannot be achieved, it must depress initial rates of HIV replication and lower the set point of infection maintaining a low viral load and therefore a state of long term non-progression. Patients who naturally have a low viral load either do not progress to disease or progress more slowly. By decreasing the viremia, there should be a lower virus load in blood and semen and therefore a reduced chance of transmission.

Therapeutic vaccines

These are designed to boost the immune system of an already-infected person.

In 1986, Zagury and colleagues used fixed B lymphocytes expressing HIV proteins but there was no evidence of improved cellular immunity. They then tried a vaccine consisting of the fixed B cells, inactive interferon-alpha, fixed PHA-activated CD8 T cells, formalin-stabilized RNA-less HIV and a mix of HIV peptides. It appeared safe. All 6 patients had higher T4 cell count. Other therapeutic vaccines have followed but no sustained clinical benefit has been demonstrated in any of them.

Protective vaccines

A major problem in development of a protective vaccine is that we know too little about the immune responses that correlate with protection. As noted elsewhere, it has been observed that a small number of people remain uninfected despite the fact that they have extensive exposure to the virus. This means that immunological protection is possible.

Initially, after exposure, there is a high level of viremia with infection of T4 helper cells and dissemination to the lymphoid system. Clearly, this initial viremia is effectively controlled and HIV-specific CD8+ cytotoxic T lymphocytes are produced. There are also HIV-binding antibodies early in infection although neutralizing antibodies only appear after viremia has been controlled. The asymptomatic phase then ensues for 8 to 10 years


The problems involved in the production of an anti-HIV vaccine

  • HIV is a retrovirus...Although HIV does not carry an oncogene, a vaccine strain will have LTRs that could lead to oncogenesis by promoter insertion and may still be oncogenic.

  • HIV is a retrovirus...The use of reverse transcriptase (genomic RNA to DNA) and RNA polymerase II (DNA to genomic RNA), which have no proof reading, leads to rapid population polymorphism. Since the error rate for RNA polymerase II is about one in ten thousand and the size of the HIV genome is about ten thousand nucleotides, this results in all daughter virions having a different genome leading to rapid population polymorphism. Therefore a vaccine against a laboratory strain will be unlikely to protect against the strains in the population.The virus is constantly evolving within the host and worldwide. In addition to variation that results from mutation and natural selection with an individual, there are several groups (M, N and O) of HIV found around the world. These groups contain sub-groups (or clades) and any effective vaccine will have to give immunity against all of them. The envelope protein (gp120) can vary as much as 20 % within a clade and 35% in sequence between various HIV groups. In addition, recombinants between different groups are being identified.

    Thus, any successful vaccine must recognize invariant regions of gp120 (such as the CD4 antigen binding site). HIV gp120 is, nevertheless, highly immunogenic and the highly variable regions promote the formation of type-specific antibodies; however, these may be responsible for directing the immune system away from the conserved regions of gp120.

  • Three dimensional epitopes in conserved regions of gp120 may only form transiently. For example, the binding site for the chemokine co-receptor is only formed after gp120 binds to CD4 antigen on the T4 cell. The CD4 antigen binding site itself is recessed into a pocket in gp120 making antibody access difficult. Gp120, in its natural state in the viral membrane, is a trimer and intermolecular associations may lead to epitopes that do not exist in purified gp120 used in subunit vaccines; thus, it may be necessary to design gp120 molecules so that they are stabilized in such a way as they present epitopes that occur, perhaps only transiently, in the native virus. For example, it may be possible to synthesize regions of gp120 attached to foreign proteins that are stabilized.

  • The major antigen is covered by immuno-silent sugars. The crystal structure of gp120 shows the problem. The conserved epitopes that bind to the CD4 antigen are, for most of the time, hidden deep in the molecule. The variable loops and the sugar chains protect these epitopes. The N-glycans are also subject to variation as the virus evolves

  • Antibodies against HIV elicited by a vaccine may increase uptake by macrophages

  • Antibodies against HIV elicited by vaccine may give rise to autoimmunity

  • We still do not know completely how HIV leads to CD4+ T4 cell depletion

  • We have no good animal models. Chimpanzees can be infected but do not develop AIDS. So human trials are necessary early in vaccine development. Alternatively, for development of a vaccine, we can use SIV and monkeys but this is not the same virus.

  • Clinical latency means we need other markers than onset of AIDS. Viral load has been shown to be a useful correlate that seems to predict clinical outcome.

  • How do we design trials? If volunteers are counseled, could this skew the results? If a false sense of security is given to vaccinees could this skew the results in the opposite direction?

  • Cell to cell transmission may make standard humoral antibodies elicited by a vaccine unimportant.

  • HIV infects specifically CD4+ cells, that is mainly the T4 lymphocytes. Thus, this virus infects and destroys the very cells that are needed to fight off an infection.

Nevertheless, the initial course of the disease shows that a good cytotoxic T-cell mediated immunological response against HIV can be mounted and the virus can be cleared from the circulation, albeit not entirely. Moreover, there are rare individuals who have been exposed to HIV-1 on multiple occasions who completely clear the virus showing that sterilizing immunity is possible. In these patients, antibodies seem to be directed (as might be expected) against the conserved parts of gp120, particularly the CD4 binding site and they mount a potent cellular immune response.

For a vaccine, the following questions are important:

  • What are correlates of protection?

  • How should antigens be presented?

  • What key antigens confer protection?

  • Should we even think of using a killed or attenuated whole virus preparation? There is also the problem of failure to inactivate formalin-killed virus completely. High-tech approaches, such as subunit vaccines, seemed more promising (but they have failed to keep their promise)

  • What about mucosal immunity? Since most HIV entry is rectal or vaginal, this is essential since the virus soon migrates to T cells in which it can set up a long term infection with no viral protein synthesis and is therefore immunologically silent.

  • Since most people are not at risk, should we aim at a therapeutic vaccine despite past lack of success?

  • Are animal models relevant?

  • What should a vaccine elicit? We need to kill infected cells since virus may be spread from cell to cell. So a cellular response is very important.




International AIDS Vaccine Initiative



Subunit vaccines

Vaccines using purified proteins do not induce a cell-mediated response; rather, they stimulate neutralizing antibodies. Until recently, most anti-HIV vaccines fell into this category. Subunit vaccines consist of recombinant proteins (e.g. gp160, gp41 or gp120) presented in a soluble form. Because of the need for a cell-mediated response, genes for individual proteins made in situ have been expressed from recombinant vectors (see below).

Original attempts at making a vaccine in chimpanzees used soluble gp120 antigen and it was found that immunized chimps can resist subsequent intra-venous challenge by virus of the same type from which the antigen was made. However, as has been noted, HIV is constantly changing and challenge by other types showed that there was no protection against these.

Since 1986, there have been many other anti-HIV vaccines  based on gp160/gp120/gp41 using various adjuvants. These can protect chimps against HIV under very controlled conditions but fail to protect against clinical isolates. In the SIV/Macaque model, vaccines based on envelope proteins do not seem to protect against subsequent challenge.

The immune response generated by vaccines that stimulate neutralizing antibodies is usually short-lived (often a matter of weeks) and of low titer.  Among the reasons for the poor response is the necessity for the gp120 to remain in its natural conformation, particularly its trimeric state, to elicit neutralizing antibodies but this also means that the large number of oligosaccharide chains and the folding of the protein itself shields potential epitopes. One site that presumably cannot change too much as a result of mutation is the receptor binding site of gp120 (i.e. either the CD4 antigen binding site or the co-receptor binding site) since mutation would preclude virus-cell interaction. But these sites remain buried in the gp120 molecule. One interesting approach has been to design vaccines against epitopes that are exposed only after the conformation change that occurs when CD4 antigen and gp120 interact.

So, in the case of glycoprotein-based subunit vaccines, we can say: All are safe! None are likely to work! Nevertheless, it is apparent that neutralizing antibodies against HIVgp120 or gp41 can be raised.


Cell-mediated immunity

Since there has been little success with soluble proteins and eliciting protective neutralizing antibodies, vaccine researchers have turned to the cell-meditated response. In fact, it is clear that the major drop in HIV virions that occurs in the patient soon after infection is the result of a cytotoxic T cell response rather than a humoral response. Moreover, depletion of CD8 lymphocytes in rhesus monkeys stops the immune control of SIV underlining the importance of the cell-mediated immunity. Thus, new vaccines attempt to display HIV epitopes on cells in association with HLA molecules and destroy infected cells rather than free virions.



Whole virus (killed) vaccines in animals

In early studies with the SIV/Macaque model, whole virus vaccines were used. These seemed to give protection although later studies showed that the observed protection was due to responses to cellular proteins in the vaccine preparation and in the challenge virus rather than a specific response to the virus itself.

Initial chimpanzee studies also used whole virus preparations and showed some degree of protection.  But these experiments were designed for success. The chimps were vaccinated with a killed laboratory strain of the virus and challenged intravenously with small doses of the same strain at the peak of antibody production. No rectal or vaginal immunity was detected and protection only extended to the same strain of HIV that was used as the immunogen. Clearly, a successful vaccine will need to act against the myriad of of sub-strains of HIV that arise during a normal infection or will need to neutralize the initial infecting virions. Using a killed virus preparation, a very complex vaccine that would elicit antibodies against numerous strains of virus would be required. In addition, since no mucosal immunity was elicited, it is doubtful if this type of vaccine could ever be prophylactic.

The early whole virus vaccine studies were done with syncytium-inducing virus (which binds to X4 co-receptor and is found in the patient late in infection) but we now know non-syncytium-inducing, macrophage-tropic, viruses are the infectious form. Clearly, what is required is a vaccine that acts against the initial infection by non-syncytium inducing virus (that binds to the R5 co-receptor).  Attempts have been made to do this but primary HIV isolates are a problem because they seem to be far better at concealing their neutralization sites from antibodies by burying vulnerable epitopes within the protein. Nevertheless, these critical conserved  epitopes must be exposed at some time in the virus life cycle, for example when the virus binds its receptors at which time a fusion structure is exposed.  In one investigation, cells were fixed with formaldehyde at the point of initiation of virus-cell fusion to capture this fusion structure. To do this, monkey fibroblasts were constructed that express the surface glycoprotein of a primary HIV isolate.  Human cell lines were made that express both CD4 antigen and CCR5 co-receptor. After mixing the cells so that the two types bound to each other, they were fixed to capture the fusion epitope and used to make antibodies in mice. The antibodies that the mice produced neutralized 23 of 24 primary isolates, many of which were non-syncytium-inducing.



Attenuated Virus Vaccines

More than a decade ago, a live, attenuated vaccine was found to protect Macaque monkeys against SIV. All vaccinated animals were protected, all controls died. This confirms that cell-mediated immunity is important. In the human population, there has always been the phenomenon of the long-term HIV-infected survivors. Why some people survive a long time is unclear but some people who have clearly had many contacts with the virus (e.g. prostitutes or other promiscuous persons) show no antibody response but some of these people show evidence of having mounted a very strong anti-HIV cell-mediated response.

What about an attenuated vaccine for human use? As noted above, there are many problems that have caused researchers to shy away from anti-HIV live attenuated vaccines but there does seem to be a natural attenuated form of HIV that might be a good candidate vaccine strain. Some people, who are actively infected and show no signs of AIDS, harbor HIV NEF deletion mutants. For example, there is a cohort of Australians who have an HIV strain with multiple deletions in the NEF/LTR region of the genome. All of these people got the virus from transfusions and showed no symptoms after more than 15 years.

NEF is not required for HIV replication in vitro and, indeed, the high mutation rate leads to the closure of the reading frame in vitro. However, NEF is important for a productive infection in vivo since, when a virus with a closed NEF reading frame is used to infect a chimp, the reading frame opens. This led to the idea that an attenuated HIV with a NEF deletion would be a good potential vaccine. A deletion mutation rather than a point mutation is needed since the former is less likely to be suppressed and therefore reversion of the attenuated virus to the virulent wild type form should be less of a problem.

It was found that NEF deletion mutants of SIV can very effectively protect monkeys against simian AIDS without causing disease. However, a natural NEF/LTR deletion mutant was found to be pathogenic for humans who were exposed to it via a blood transfusion. It resulted in immunosuppression although onset was later than with the wild type virus.

It appears also that there is still the problem of reversion, even if it is small, and unfortunately this occurred. A NEF deletion of SIV reverted to a virulent strain by a process of gene duplication of an adjacent sequence. Moreover,  juvenile monkeys developed AIDS when given a high dose of the vaccine. So there might be a good chance of an infection among vaccine recipients especially those who are immune compromised (cancer patients or the elderly). It is thought that the vaccine strain may continue to replicate at low levels in some compartment of the immune system.

Recombinant vaccines

In order to present antigen in the context of HLA molecules and raise a cell-mediated response, recombinant vaccines that contain a gene (or genes) from HIV in a non-pathogenic vector have been developed. One vector that has been used is vaccinia, the live virus that is used as a smallpox vaccine. A vaccinia strain that has been used for vaccine production is the further attenuated Ankara strain. It is a very effective against smallpox which means that thousands of people have received it with a high level of safety. The anti-HIV vaccine based on the Ankara vaccinia containing HIV genes elicits both humoral and cell-mediated immunity and seems effective in animals. Unfortunately, in humans the results have given less cause for optimism. There was only limited anti-HIV activity, perhaps because most of us have immunity to vaccinia a result of vaccination. As with the attenuated vaccine strains, there is also the problem that live recombinant virus vaccines may spread through populations of immuno-compromised individuals.

Another vector, canarypox, has been used as the basis of an HIV recombinant vaccine.  Again, those tested have shown limited development of a long-lived cell-mediated response. In one trial (2007), a canarypox vaccine was tested along with administration of interleukin-2 to boost the patients' immune response. The subjects were HIV positive and being treated with HAART. The investigators wished to determine if the vaccine, ALVAC vCP1452, would maintain low levels of HIV in the blood of people after stopping antiretroviral therapy. The vaccine contains several HIV-1 genes: gp120, expressed by a part of the env gene of the MN HIV-1 strain; the anchoring transmembrane domain of gp41 of the LAI HIV-1 strain; the p55 polyprotein expressed by the gag gene of the LAI HIV-1 strain; part of the pol gene expressing protease activity from the LAI HIV-1 strain in order to process the p55gag polyprotein; in addition there are genes expressing peptides from pol and nef that are HLA-A2-restricted cytotoxic T-cell lymphocyte epitopes. Two vaccinia  coding sequences are also present in the vector to improve RNA translation and the expression of HIV-1 proteins. Unfortunately, there was no difference between the treatment and the placebo group.

Another possible vector is replication-defective adenovirus and these seem to raise better responses. These viruses can infect cells and are expressed but progeny virions are not formed. Some vaccines contain multiple viruses, each expressing different HIV proteins such as nef, gag and pol. These offer protection in animal models but there is always the question of how relevant such models are to human AIDS; indeed, past experience has shown that animal models may offer a lot of hope that is subsequently dashed in human trials.

STEP Study

A phase 2b "proof of concept" trial sponsored by Merke and NIH of a novel adenovirus-based vaccine was the STEP study (HVTN 502 or Merck V520-023 study). This involved over 3000 people in several locations around the world receiving a vaccine based on mixture of adenovirus 5 (rAd5) vectors which express nef, pol and gag from HIV1 clade B. Phase 1 trials showed that the virus was well tolerated and an immune response was elicited. The response was not as good in people who had pre-existing anti-Ad5 neutralizing antibodies. This might limit the effectiveness of the vaccine since many people in America and Europe have such antibodies, as do the large majority of people in sub-Saharan Africa. This could be overcome by using a different adenovirus vector. This trial was prematurely terminated in late 2007, however, because it was not only clear that the vaccine would not meet the end points set for efficacy but also people with pre-existing anti-Ad5 neutralizing antibodies showed more infections by HIV than the control participants who received a placebo. The reason why people who have anti-Ad5 antibodies and received the vaccine were more susceptible is unclear, although it has been suggested that they may have memory Ad5-specific CD4 T cells that could be better targets for infection by HIV or that the anti-Ad5 antibodies bound to the vector particles (opsonization) and in some way changed their tropism.

Another adenovirus-based vaccine is based on a dual approach in which a primary vaccination with DNA plasmids is followed by a rAD5 vaccine boost. Among the DNA vaccines tested are a six-plasmid combination expressing sequences encoding the HIV-1 subtype B gag, pol, and nef on separate plasmids with protein expression being driven by a novel cytomegalovirus promoter. In addition, a phase 1 trial with a four plasmid mixture encoding subtype B gag-pol-nef fusion protein and modified gp120 constructs from subtypes A, B, and C. However, a similar protocol in monkeys showed long lasting change in the set point viral loads after administration of the vaccine.

A Possible Success

In September, 2009, the first clinical trial that showed any efficacy of an AIDS vaccine was reported. The trial was conducted in Thailand using 16,402 volunteers who were followed over six years. Half received placebos and half received the vaccine. The volunteers were from the population at large rather than high risk men and women such as sex workers and intravenous drug users. All participants were given advice on avoiding infection and the use of condoms and were tested for three years for HIV infection. Seventy four of the vacinnees who received the placebo became infected while 51 who got the vaccine became infected. This small difference of 23 is statistically significant and suggests that the vaccine is 31% effective. This, clearly, is not great and normally an effective vaccine should give greater than 80% protection.

The vaccine (RV144) combined two vaccines that in previous trials had shown no efficacy. Alvac-HIV (Sanofi-Aventis) is a canarypox virus with three HIV genes cloned into it. The other vaccine, Aidsvax (Genentech), is a subunit vaccine containing purified surface protein of HIV, gp120, prepared in Chinese Hamster Ovary cells.

Why the combined vaccine seems to work, at least partially, when its two constituents failed to confer protection in previous trials is a mystery. Another mystery is why the vaccine did not lower viral load in vaccinees who became infected which would normally be the case.

A further analysis of these results has cast doubt on whether RV144 worked at all. This analysis which eliminated some participants showed that the effectiveness of the vaccine may have been as low as 26 percent with a 16 percent possibility the results were due to chance. A 16 percent possibility of chance is considered high in vaccine trials.





  Return to the HIV chapters

This page last changed on Sunday, August 28, 2016
Page maintained by
Richard Hunt