Revolutionary HIV vaccine that ‘trains’ the immune system passes its first hurdle

Gustavo Fring/Pexels

A completely new kind of vaccine that aims to induce a response to HIV that is ‘better than nature’ has completed its first immunogenicity trial in humans, with promising results.

The vaccine, currently dubbed eOD-GT8, is being developed by Professor William Schief and colleagues at the Scripps Research Institute in California, in collaboration with the Fred Hutchinson Cancer Center, the International AIDS Vaccine Initiative, the US Institute of Allergies and Infectious Diseases, and a number of other research institutes. The pharmaceutical company Moderna are involved in the next stage of trials, as the same mRNA vaccine technology that the firm used in its SARS-CoV-2 vaccine will be used.

The vaccine aims to bring into being immune cells (B-lymphocytes) that will flood the body with broadly neutralising antibodies (bnAbs) of the type called VRC01 in the event of an exposure to HIV. This particular bnAb has the ability to neutralise (block) a very specific part of the HIV envelope protein that is necessary for the virus to enter CD4 cells.


broadly neutralising antibodies (bNAbs)

A neutralising antibody (NAb) is an antibody that fully defends its target cell from an antigen. A broadly neutralising antibody (bNAb) is a neutralising antibody that has this effect against a wide range of antigens. A number of broadly neutralising antibodies have been isolated from persons living with HIV. Some of them are being studied and, in some cases, used in clinical trials, to defend humans against HIV infection, treat HIV infection, and kill HIV-infected CD4+ T cells in latent reservoirs.


A substance which forms the structure of most cells and enzymes.


Genes are instruction manuals for our bodies. They determine characteristics like our eye and hair colour. Every human has a set of around 20,000 genes. We get one copy of each gene from each of our parents. Genes can also play a part in our health and may affect our risk of developing some health condition.


The outer surface of a virus, also called the coat. Not all viruses have an envelope. In the case of HIV, the envelope contains two viral proteins (gp120 and gp41), which are initially produced as a single, larger protein (gp160) that is then cleaved in two. 

immune system

The body's mechanisms for fighting infections and eradicating dysfunctional cells.

The problem is that bnAbs have, so far, never been induced by a vaccine in humans, only by chronic HIV infection in a minority of people. This is because very few B-lymphocytes have the inbuilt capacity to ever make these uncommon, ‘hypermutated’ bNAbs from scratch.

The Scripps vaccine uses a technique called ‘germline targeting’ to greatly expand the pool of B-lymphocytes that have the necessary genes to be able to make bNAbs, and the current trial, G001, has managed to do this – a remarkable feat that essentially redesigns part of recipients’ immune systems.

The eOD-GT8 vaccine was designed to elicit this genetic response, not to actually induce bNAbs, and did not contain all the components of the HIV envelope protein that usually give rise to an antibody response. In future trials, it is hoped to turn the potential capability to produce VRC01 bNAbs into a reality by using vaccines ever more closely aligned to the subsections of the HIV envelope protein that give rise to the strongest bNAb responses. These are called epitopes.

Findings from the G001 study, the first of its kind, are published in Science. In the study, the higher of two doses of the vaccine induced the production of 550 times as many cells capable of producing the VRC01 bnAb as naturally occur, and in responders, the percentage of B-cells in lymph nodes that had the genetic mutations (called VRC01-2*02 or VRC01-2*04) that were capable of making bNAbs increased to 6.7%, compared with an average of 0.00023% (just one cell in 429,000) in unvaccinated people.

The affinity of the receptor molecules produced on the surface of these cells – their sensitivity to HIV epitopes or to antibodies to those epitopes – increased 840-fold in comparison to cells that did not have the VRC01-2 mutations.

More on the background

For more details on why HIV is such difficult virus to develop a vaccine against HIV, see this, but  perhaps the core reason is HIV’s ability to rapidly mutate away from immune control. In the vast majority of cases of infection the virus has already won the race against the immune system before the latter is scarcely off the starting blocks.

Some people eventually develop broadly neutralising antibodies, but do so too late, when the virus has already evolved the ability to outwit them. But if we could infuse the capability into enough B-lymphocytes to produce bnAbs against HIV the moment they saw it – then the immune system would win the race.

The approach used by Schief’s team is not the only one that tries to do this. In the approach called epitope-focused design, researchers try to develop a multivalent vaccine that recognises as many HIV epitopes as possible and produces bNAbs against them.

Lineage-focused design reverse-engineers this. Scientists look at the bNAbs in people who develop them in chronic infection, work out what HIV epitopes they developed in reaction to, and then design vaccines to stimulate vaccines to those epitopes.

The problem with both these approaches is that there are so few B-lymphocytes that have the genes to make bNAbs that effective amounts are not generated.

What germline targeting does is to take the reverse-engineering one stage further back and ask: can we design an evolutionary immune process that starts with naïve B-lymphocytes that don’t specifically recognise any bug, and ends up with memory B-cells that can produce floods of varied bnAbs in response to HIV?

VRC01 was one of the first broadly neutralising antibodies to be discovered, in the blood of a person with long-term HIV infection. Its structure and potency were determined in 2010. In studies using infusions of the antibody to treat HIV or to prevent HIV infection, it had some efficacy but was in itself not sufficient to suppress HIV replication and its efficacy in preventing HIV was only modest. However having an immune system that is primed to produce large amounts of VRC01 as soon as it detects HIV may work better.

The reason the Scripps researchers picked VRC01 as the bNAb they wished to generate is not because it has already been used, but because the cells with the capability to produce it have a unique genetic signature. All antibodies are Y-shaped molecules. The genes that it needs to produce the particular series of amino acids (protein components) that becomes the VRC01 antibody are either one or both of the VRC01-2*02 or *04 genes, which produce the ‘heavy chain’ of the antibody (one arm and the stem of the Y) and any five, but not any other number, of amino acids that produce the ‘light chain’, the other arm of the Y.

Only one B-cell in over 400,000 naturally has this genetic conformation, and HIV envelope proteins encountered upon infection do not seem to stimulate these ‘precursor’ B-cells to specifically proliferate, nor to generate bNAbs.

What Schief’s team did was to design, on the evidence from previous clinical trials in mice, a vaccine that would preferentially cause cells with this genetic conformation to divide and proliferate. It was not designed to, and did not, produce the VRC01 bNAb itself, but to produce more of the cells that make it.

This made the study a lot more complicated to do: it is a relatively simple process to measure the amounts of an antibody in blood, but it’s much harder to look at the genes in the cells that make them, and the researchers had to use several ways to do this, including sorting cells with flow cytometry and using RNA and DNA PCR assays.

The vaccine and the study participants

The eOD-GT8 vaccine is a virus-like particle – a hollow shell of a non-human protein. Virus-like particle vaccines are not new – the HPV vaccine that prevents genital warts and cervical and anal cancers is one. In this case, the non-human protein is lumazine synthase, an enzyme found in organisms that can make their own riboflavin (vitamin B2).  Sixty enzyme molecules spontaneously assemble into a hollow shell as part of its synthesis function. To this, the researchers attached 60 HIV envelope protein molecules, one per enzyme unit, attuned to elicit proliferation in cells with the VRC01 genetic profile.

This vaccine, or a placebo, was given to 48 volunteers in two doses eight weeks apart; 12 received placebo, 18 received a vaccine dose of 20 micrograms and 18 a dose of 100 micrograms. To both vaccine doses were added 50 micrograms of an adjuvant, a compound that amplifies an immune response to a vaccine. In this case the adjuvant was ASO1B, a mix of two naturally-occurring compounds that assist the vaccine to attach itself to B-lymphocyte receptors (surface proteins that sense foreign substances).

The volunteers were quite balanced in terms of gender (18 women, 24 men) and sex at birth (four of the men were trans; none of the women were, but two volunteers described themselves as genderqueer or non-conforming).

This US study wasn’t so balanced in term of race; 33 (69%) were classed as White of whom six were Latino; four were Black, five Asian, three of mixed race and one unknown. Their mean age was 30.

The vaccine did have some side effects: more people receiving the vaccine than people receiving placebo complained of pain at the injection site, mild headache, chills, joint pain, mild fever and malaise or fatigue. However the rate of severe symptoms (grade 3) was not higher than in placebo recipients.

The vaccine’s effects

The vaccine did produce a general reaction in non-bNAb-producing naïve B-cells and T-cells, showing it did have general immunogenicity (as also shown by the aside effects). B-cells circulating in the blood increased about a thousandfold. B-cells in the lymph nodes, where they are taken to be repeatedly exposed to foreign antigens and to turn into memory cells that recognise specific antigens, before being returned to the bloodstream, increased by a factor on 100,000 or more.

But they did not develop any HIV-specific responses. This was expected, as this priming vaccine was not designed to produce a reaction directly to HIV and did not contain HIV surface proteins known to be directly immunogenic in this way.

A strong generalised immune reaction to HIV env proteins would in fact have been counterproductive, as it would divert B-cells into producing non-neutralising antibodies against HIV that don’t work. The point was to magnify the population of precursor cells whose genetic signature showed that, if a later vaccine containing a specific immunogenic HIV epitope was used, a useful proportion of the antibodies produced would be bnAbs.

The main goal of the vaccine, to magnify the population of B-cells capable of making bnAbs, was achieved. Without going into exactly how here, the researchers looked at three populations of B-lymphocytes and counted the proportion that had the VRC01 genetic profile. Two, ones circulating in the blood and ones maturing in the lymph nodes, make the most common type of antibody, immunoglobulin G (IgG), which makes up 75% of circulating antibodies. They also measured the amounts of cells in the blood making a rare type of antibody, immunoglobulin D, which only forms 1% of antibodies but appears to have a role directing the antibody production of other B-cells.

In pre-vaccination samples, the researchers found six participants where precursor cells with the VCR01 profile could be identified, but literally in only one or two individual cells, indicating that even in these participants, only one in 429,000 B-cells had this signature. Post-vaccination, they also found this signature in two placebo recipients.

Among vaccinated participants, VRC01 cells proliferated in all but one participant. This participant turned out to have an uncommon genetic VRC01 signature, VCR01-2*05/*06, showing that maybe 2% (of predominantly White US participants) might be non-responders to this vaccine. Whether people from other parts of the world would have different genetic profiles is a question yet to be answered.

Four weeks after the first vaccination, one in 10,000 circulating blood cells in the lower-dose recipients and one in 4500 in the higher-dose recipients had VRC01-producing potentiality, representing respectively 43 and 94 times the pre-vaccination frequency.

By week 10, two weeks after the second vaccination, one in 1139 B-memory cells in low-dose recipients and one in 777 in high-dose ones were of the VRC01 precursor type, representing 377 and 552 times the pre-vaccination frequency. By eight weeks after the second dose this had gone down to one in 3764 and one in 2019 cells respectively.

It's important to emphasise that even at peak, although all but one participant now had B-cells with receptors that indicated they could produce bnAbs when encountering HIV epitopes, this still only represented 0.088% of all memory B-cells. The hope of the researchers is that, now that these precursor cells have been elicited, further immunisations with vaccines that do contain actual HIV epitopes (unlike this one) will now produce effective amounts of the VRC01 bNAb.

In the B-cells maturing in lymph nodes, only about 60% of participants had an expansion in the number of VRC01-type B-cells maturing in their lymph nodes. Amongst those that did, however, the number of cells with the VRC01 gene signature was about 100,000 times greater than before, with up to 6.7% of all IgG-producing B-cells having the VRC01 gene signature.

A promising sign was that within the population of cells with the VRC01 genes, there was wide variation in the rest of the genes that instruct the cell what antibodies to make. This ‘polyclonality’ indicates that they should be able to respond to a wide variety of HIV variants.

Sensitive gene sequencing also found that, even while the absolute number of VRC01 cells was already declining within eight weeks of the second vaccine, the number of mutations in cells with the VRC01 signature continued to accumulate, indicating increased potentiality to make a variety of VRC01-type bnAbs.

These mutations also confer increased sensitivity of VRC01 bnAb-making cells to typical HIV env protein epitopes, indicating an increase not only in the amount of bnAb cells that might be produced in response to an HIV exposure, but also in their sensitivity. Three weeks after the first vaccine, the affinity of B-cell receptors to HIV env epitopes in VCR01-type cells was 840 times higher (meaning 840 times less env protein would produce a reaction than what would produce a reaction in non-VRC01 cells) and three weeks after the second, 32,400 times higher. 

This is important because it suggests that, even though initially VRC01-producing B-cells might be a minority population, their bNAb production would not be outcompeted by the majority of cells that produce non-effective, non-neutralising antibodies which HIV could easily mutate away from.

Can this be turned into an actual vaccine?

In a commentary on the research paper, Professor Penny Moore of the South African National Institute for Communicable Diseases, herself a prominent HIV vaccine researcher, hails this as the first proof of a very complex concept.

She notes that the research needs to be repeated in participants from other regions and ethnicities, and also that studies need to be conducted with vaccines that expand cell populations capable of producing other bnAbs with greater potency than VRC01, and which target parts of the HIV envelope protein other than the CD4 binding site. This may be more difficult, as other bnAbs derive from cells with less specific genetic signatures.

The main challenge, however, it that this is only the first step in this vaccine concept. The next step will be to sequentially inoculate people with a series of vaccines containing epitopes based ever more closely on actual HIV virus env proteins.

It’s problematic as to how this can be done. A vaccine containing numerous HIV epitopes might have to have high volume, and also runs the risk that it could overwhelm the immune system and produce a massive but non-specific reaction that might not only be unsafe but would also not work against HIV. The alternative is to give people a long sequence of vaccines, aiming to ‘shepherd’ the immune system into producing bnAbs, but this would be challenging logistically and clinically.

One possible answer is that the next-stage germline targeting vaccines, in trials G002 and so on, will use mRNA-based vaccines, not virus-like particles. These will tell cells to make their own particles, not inject then direct. This implies that a number of different instructions could be packaged into a smaller dose of vaccine and also that the manufacturing of varied vaccine jabs should be easier. The G003 trial is taking place in Rwanda and South Africa, which may answer the question of whether a vaccine designed to elicit VRC01-works as well in African populations.

Essentially what the Scripps researchers have done is to create a significant population of cells, at least potentially capable of making one particular type of bnAb, that was insignificant before. The idea, to use a metaphor, in the hare-versus-tortoise race of HIV versus the immune system, is to equip the tortoise with a jet pack so the hare finds it waiting at the finish line.

Schief and colleagues comment that they have achieve “an unprecedented level of control over the specificity of immune responses” that may “herald new era of precision vaccine design for HIV and other pathogens”. In other words, while it may still be challenging to turn this priming vaccine into an entire, effective prime-and-boost course of vaccines, the scientific approach it uses has huge implications for more sophisticated vaccine design generally.


Leggat DJ, Cohen KW, Schief WR et al. Vaccination induces HIV broadly neutralising antibody precursors in humans. Science, 378(6623): pre-print publication, 2 December 2022.

Moore PL. Triggering rare HIV antibodies by vaccination. Science 378 (6623): 949-950, 2 December 2022.