If one antibody won’t work, how about three? Or three in one?

Dr Boris Juelg at CROI 2024. Photo by Krishen Samuel.
Dr Boris Juelg at CROI 2024. Photo by Krishen Samuel.

This week’s Conference on Retroviruses and Opportunistic Infections (CROI 2024) heard the latest data from research involving giving people a cocktail of three broadly neutralising antibodies (bnAbs), which aidsmap last reported on two years ago.

In the latest study, Dr Boris Juelg and colleagues from the Ragon Institute in Massachusetts worked with 12 people with HIV, who stopped their antiretrovirals (ARVs) and instead received the triple bnAb combination. Of these, four and possibly five people (one was lost to follow-up) maintained undetectable viral loads off any HIV therapy to the end of the study period, more than six months after their last bnAb injection. Another five study participants did experience viral rebound, but not till 4 to 6 months after their last bnAb injection. Only two people had early viral failure.


BnAbs have been an intense focus of HIV treatment and prevention research for around 15 years.

These highly evolved protein molecules are capable of killing HIV-viruses in their own right – neutralising them. This is unlike the normal antibodies produced against HIV (the ones diagnostic tests look for), which can only signal to other parts of the immune system to do that job. In both types of antibodies, however, they arise too late; HIV has long outpaced them, because it can mutate around these immune defences and develop resistance to them.


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.


The amount of time it takes for a concentration in blood to be reduced by 50%. After one half-life, the concentration of a drug in the body amounts to half the starting dose of any drug to be eliminated from the body.


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


In pharmacology, a medication which maintains its effects over a long period of time, such as an injection or implant.


A variant characterised by a specific genotype.


The focus of bnAb research has therefore been to find a way to introduce them earlier, while they still work against HIV. In prevention science, the goal is to induce the immune system to make bnAbs itself, in advance of infection. This is an intensely difficult challenge, as it involves trying to get the body to react to HIV with a stronger immune response than occurs naturally (partly because HIV destroys the very immune cells that should protect us). But scientists remain hopeful it can be done.

As treatments, bnAbs or derivatives of them can be used as drugs; in oncology, their ability to identify and attack cancer cells has revolutionised the field. Hopes have been high that they could do the same in HIV treatment, partly because they have long half-lives in the body so could be used as long-acting therapies. Also, they might have additional vaccine-like effects that could continue to suppress HIV long after the end of therapy.

But bnAbs, and antibodies in general, have limitations as drugs too. As natural products, they have evolved in individuals’ bodies to be potent against a limited number of HIV strains. Researchers sometimes talk about ‘tier 1’ and ‘tier 2’ varieties of virus, which indicates how easily they are neutralised by bnAbs (tier 2 are the tougher ones). This is a bit like having an antiretroviral that only works against, say, subtype B viruses. In addition, individual viral strains may have specific spot mutations that confer something much more akin to conventional drug resistance to antibodies, and antibody treatment that isn’t completely suppressive may ‘select for’ resistance, i.e. cause it to arise.

This is why, by analogy with ARVs, when single bnAbs or dual combinations seemed too vulnerable to resistance, it was theorised that a triple combination might work.

This study

Juelg and colleagues therefore used three antibodies that bind to different parts of the HIV gp120 envelope molecule, named PGT121, PGDM1400 and VRC07-523. Pharmacokinetic studies showed that in people with no resistance to any of the three antibodies, a concentration of one microgram of each antibody per millilitre was its IC90 (sufficient to neutralise 90% of HIV viruses). In someone resistant to one of the antibodies, this concentration of the other two instead became sufficient to neutralise 80% of viruses.

The 12 participants (demographic details were not given) were all living with HIV and had fully suppressed virus on antiretroviral therapy (ART). They stopped ARVs at day one of the study and instead were given the three antibodies intravenously every four weeks until week 20. They were then followed-up for at least another 20 weeks and, if their viral load rebounded to at least 1000, restarted ART.

The half-lives of the bnAbs in blood indicated that this four-weekly dosing regimen was more than sufficient to maintain drug levels over the IC90 in people without resistance. The time taken for the bnAb levels to fall by half was 20 days in PGT121 and 24 days in PGDM1400. In the case of VRC07-523 it had a modification that conferred a longer half-life of 45 days; this implies that, in itself, it could be dosed every three months.

Two trial participants had viral loads that rebounded quickly, before the planned end of bnAb dosing, and both to high levels of over 100,000. For one participant this happened just after the second dose (week 4) and for the other just after the third dose (week 8).

It turned out that both participants had pre-existing resistance to both PGDM1400 and PGT121. While in one case this did not worsen because the person already had high-level resistance, in the second person, who initially had intermediate resistance to PGT121, this had become high-level by the third dose, with additional resistance mutations selected. Both participants started with some intermediate-level resistance to VRC07-523. In the person whose virus rebounded after the second dose, this remained intermediate because their virus did not need to change any further to evade the bnAbs, but the second person did acquire high-level resistance to that bnAb too.

In the five people who had viral rebound much later, levels of PGDM1400 and PGT121 had fallen to low levels by the time of their treatment failure. One of three participants tested for resistance became completely resistant to these two bnAbs but the other two only developed some intermediate resistance. Because of the longer half-life of VRC07-523, its blood levels were still quite high at the time of virological failure and no or only intermediate resistance developed.

In the four (or five) participants who maintained virological suppression, it was difficult to get resistance data as by definition there wasn’t much virus to be tested for it. In the case of one individual, they already had some mutations against each of the three antibodies but further mutations were not selected for over time, even though PGDM1400 and PGT121 levels were near-zero by the end of the study. This might be because the low but still detectable level of VRC07-523 was enough to protect them or, more interestingly, that some other immune effect of the bnAbs was limiting viral replication.

That this might be a possibility was suggested by the fact that HIV-specific immune responses to the Gag (structural) protein of HIV were considerably enhanced in people whose HIV reappeared late, and even more so in those with no rebound. They also had some response to the HIV Env surface protein compared to none in the early rebounders. The type of T-cells that developed HIV responses was different too: the early rebounders mainly developed responses in their shorter-lived effector-memory cells, but the late rebounders and non-rebounders had more response in their longer-lived central-memory cells, indicating some vaccine-like effect. However there were too few study participants for these differences to be statistically significant.

Boris Juelg said that considerably larger and longer studies of this kind of triple-bnAb combination were needed before deciding whether it is a viable treatment option.

Two more studies

Two other studies of bnAbs independent of other ARVs were presented in the same session. Dr Athe Tsibris of the Brigham and Women’s Hospital in Boston presented a first in-human study, not of three bnAbs, but of a novel ‘trispecific’ antibody. This combines sequences responsive to three different parts of the HIV envelope – the V2 loop, the CD4 binding site, and the Membrane Proximal External Region (MPER) or gp41 protein.

This had a surprisingly poor effect, reducing viral load by only 0.38 logs (a 2.4-fold drop) in people who started the bnAb therapy before they started ART. It is unclear whether this was due to pre-existing resistance, or due to something about the assembly of something as complex as a trispecific antibody, that made it less potent than expected against HIV.

Secondly, the BANNER study, presented by Dr Jan Losos of ViiV Healthcare, gave just one dose of one bnAb, N6LS, to people with HIV who had not started ART. This produced a 1.5 log (32-fold) drop in viral load by day nine after dosing, but viral load started increasing again after that. ViiV intends to take N6LS into a larger study, EMBRACE, which will combine the antibody with hyaluronidase, a chemical that should enable subcutaneous rather than intravenous dosing, with or without long-acting cabotegravir.

Comments and controversy

These bnAb studies attracted some criticism from the audience. The UK’s Dr Laura Waters questioned the ethics of not testing participants in advance for pre-existing resistance. The reason it was not done seems to be in part that testing for bnAb sensitivity in people’s HIV is a long and expensive process, though Jan Losos indicated it would be done, at least in some participants, in EMBRACE.

Ugandan activist Moses Supercharger questioned the feasibility of a class of antivirals that would require refrigeration and expensive efficacy assays that are not available in many parts of Africa. And Dr Ricky Hsu of the AIDS Healthcare Foundation in New York also questioned the resources generally being poured into bnAbs, given the need for expensive assays – not just for resistance, but because of the fact that anti-antibodies can develop against these large molecules, which is another cause of loss of efficacy.

Presenters commented that it was more the case that their focus was on a need, rather than a class of drug. Many people with HIV had made it clear that they would value an injectable therapy that they only had to take every few months, and bnAbs fit that bill. Long-acting small molecule drugs were also in development, and data were also presented in this session on a combination of a small-molecule drug, lenacapavir, plus two bnAbs. Even though bnAbs have been in development for more than a decade, it was still “early days” in understanding their potential, and their pitfalls, panellists said.


Juelg B et al. Therapeutic efficacy of a triple combination of HIV-1 broadly neutralizing antibodies. Conference on Retroviruses and Opportunistic Infections, Denver, abstract 121, 2024.

View the abstract on the conference website.

Tsibris A et al. A first in-human study of the trispecific HIV-1 broadly neutralising antibody, SAR441236. Conference on Retroviruses and Opportunistic Infections, Denver, abstract 118, 2024.

View the abstract on the conference website.

Leone P et al. (Presenter Losos J) VH3810109 (N6LS) in adults with HIV-1 who are ART-naïve: phase IIa BANNER efficacy data. Conference on Retroviruses and Opportunistic Infections, Denver, abstract 117, 2024.

View the abstract on the conference website.