An HIV vaccine candidate has proved it can guide B-cells, the part of the immune system that makes antibodies, into being able to produce broadly neutralising antibodies (bnAbs) with the ability to block the entry into T-cells of many strains of HIV. Dr Wilton Williams of Duke University, US, described the vaccine candidate in a plenary talk at July’s 12th International AIDS Society Conference on HIV Science (IAS 2023).
The particular part of HIV’s gp41 fusion ‘spike’ that the antibodies target and neutralise is only exposed for a few seconds during the process of virus/cell fusion. These bnAbs were therefore also designed to attach to the viral membrane adjacent to the spikes, whose splitting into gp120 and gp41 proteins initiates the fusion sequence. At the crucial moment, gp41’s ‘membrane proximal external region’ (MPER) is uncovered, and the antibody that has been waiting, attached to the viral membrane, can ‘pounce’ and swiftly block its action. (The more ‘closed’ the shape gp41 is, and the less time its MPER is exposed, is what makes the difference between so-called ‘tier 1’ viruses that are easy to neutralise and ‘tier 2’ ones that are hard to.)
This vaccine is another example of so-called germline targeting, an approach that ‘trains’ B-cells to become more and more sensitised to HIV and produce antibodies with progressively more potency and breadth of efficacy for the part of the virus they are sensitised to.
The reason to do germline targeting is because a normal vaccine would not produce antibodies capable of neutralising HIV.
Many viruses, such as the one that causes COVID, induce strong immune responses. These may be directed against a very specific part of the virus – in COVID’s case, its spike protein. In comparison, HIV produces an ineffective immune response when it first enters the body. This is because it is naturally less immunogenic than other viruses; because it subverts the immune response by preferentially infecting the very cells that normally direct antiviral immunity; and because it is tremendously variable, so the antibodies that are produced only work against (‘neutralise’) very specific viral strains, and HIV can easily mutate away from them.
BnAbs are broadly-neutralising, ‘hypermutated’ antibodies that have the ability to target and penetrate the most ‘conserved’ parts of HIV, the components it can least afford to change genetically. Because of this they can neutralise a wider variety of subtypes of this notoriously variable virus.
BnAbs eventually form in the bodies of about 15% of people with chronic HIV infection, by which time they are too late to be effective. We also know that there exists, somewhere in the immune system, a few B-cells that are capable of making bnAbs – or of changing into cells that can (B-cells can mutate to produce neutralising antibodies). But these are tremendously rare and their ability to make bnAbs may initially exist only as potential. The aim of germline targeting is to ‘train’ these precursor cells to become combat-hardened virus fighters.
The bnAbs that the researchers were hoping to generate have already been isolated from the blood of people who have them, and their broadly neutralising ability has been quantified. The two found to have the greatest affinity for the viral membrane and the strongest MPER response are called 2F5 and m66.6. But the point of a bnAb-based vaccine is to induce these antibodies to form in HIV-negative people so that any attempt at infection by the virus is met by a blockade of bnAbs that already recognise it.
Dr Wilton Williams told the conference that the vaccine that induced the formation of these bnAbs – in a study known as HVTN 133 – was a relatively simple construct. It is composed of a lipid nanoparticle – a fat droplet – that contains three relatively short ‘epitopes’. These are lengths of viral protein – often only a few links long – that excite a strong and specific antibody response. One of them was part of the MPER sequence in gp41.
The vaccine was originally intended to be given in four inoculations at months zero, three, six and twelve of the study. The increasing binding ability (affinity) and specificity of the antibodies produced would be demonstrated with a lab-dish assay that uses a type of modified cancer cell (called TZM-bl) that HIV infects more easily than a CD4 cell. The more infection of these cells was blocked, the stronger and broader the antibody’s neutralising abilities.
Twenty participants were to be given the vaccine and four a placebo. Halfway through the trial, however, it had to be stopped. One of the participants developed a severe allergic reaction (anaphylaxis) to one of the ‘buffer’ ingredients in the vaccine – the normally inert chemical polyethylene glycol or PEG, which has often been used to improve the stability and half-life of injected medicines. Although allergy to PEG is very rare, the researchers will now repeat the study using a PEG-free vaccine formulation. It it not expected to affect its immunogenicity.
At this point all 20 vaccine recipients had had two inoculations, there were five participants who had also had their third inoculation, and no one had received all four. It was found that after the second inoculation, all but one (95%) of vaccine recipients had an immune response to the MPER region of gp41; 65% had the specific response to the same peptide (sequence of MPER components) that the already-studied bnAb 2E5 had; and 35% of vaccine recipients had already developed detectable numbers of the kind of B-cells that were capable of producing such antibodies.
Only 0.03% or less than one in three thousand of the B-cells that had any reaction to MPER had developed this capability, but the point is that in situations of real infection, comparatively rare B-cells can hugely proliferate and turn into antibody factories.
The researchers then turned their attention to the five people who had received three vaccinations.
By using a technique called flow cytometry, they isolated individual B-cells that were reactive to the HIV MPER sequence. They only isolated a tiny number of these – a total of 80 cells from the five participants – and out of these picked 38 individual cells that they then cloned, turning the initial 38 into a sufficient number of cells to block HIV from infecting the TZM-bl cells in the lab-dish assay.
In the two best vaccine responders, it was calculated that the proportion of all of their B-cells that were capable of producing 2E5-type bnAbs had expanded from less than one per million before their first inoculation to one per 15,000 after their third.
Of the 36 species of cloned cells the researchers produced, 14 neutralised tier 1 strains of HIV but Dr Williams said that “strikingly”, two (5%) neutralised tier 2 strains of HIV – the mark of a true bNab.
The genes of these two types of B-cell were sequenced. It was found that they were enriched with certain gene sequences that were uncommon at baseline, proving that the B-cells had responded to the vaccine by generating a “polyclonal heterologous bNab response”, in other words one of sufficient breadth to cope with multiple strains of HIV.
The most common kind of antibody component produced by the cloned cells was a length of antibody protein found in seven of them, but it varied considerably in its composition, being from 14 to 23 amino acids long – again, evidence of the breadth of its response. This was called the DH1317 clonal lineage, meaning its development could be traced using a phylogenetic tree to show cells that must have developed from precursors.
They found that 11 different versions of the cloned B-cells had developed the capability of producing antibodies that could neutralise tier 2 HIV viruses.
A word of caution, however. Only 15% of HIV tier 2 viruses selected from a panel of different varieties could be neutralised with antibodies expressed by the DH1317 cells, though a higher proportion (35%) of HIV subtype B viruses could be (subtype B is the one that predominantly affects the gay population in higher-income countries). However while the DH1317 cells expressed no antibodies that could broadly neutralise the predominant African subtypes (A and C) they did find evidence that sequences capable of neutralising them were developing and that antibodies from these B-cells were capable of neutralising some tier 1 viruses.
More evidence that the vaccine was pushing B-cells into creating the right kind of bnAbs was provided by the fact that antibodies derived from the two most effective members of the DH1317 clonal lineage were also the two with the highest binding affinity to the virus’s lipid membrane, as well as to its MPER protein sequence, showing that its hypothesised mechanism of action appeared to be borne out in reality.
This may sound like a long way from proving that such a vaccine could produce effective anti-HIV antibodies in a large efficacy trial. But it is more evidence that a concept that appeared highly theoretical and elusive – that of pushing the immune system into mounting a response to HIV that it does not start out with naturally – can be made to work.
The researchers now aim to repeat HVTN 133 with a new trial using the PEG-free formulation and at least four vaccinations, as originally intended.
Wilton Williams hopes for improved results in the next study. “Lack of additional breadth is due to insufficient bNAb affinity maturation, and insufficient recognition of HIV sequence diversity,” he said.
Williams W. Engineered HIV-1 immunogens to induce broadly neutralising antibody (bnAb) responses. 12th IAS Conference on HIV Science, Brisbane, plenary PL01, 2023.
Update: This article was amended on 14 September 2023 to correct the name of the study (HVTN 133, not HPTN 133).