We still have no HIV cure. But we have a better strategy for finding one

There are many pre-conference meetings that lead up to each International AIDS Society conference. One of the most eagerly awaited, by researchers and treatment activists alike, is run by the Towards an HIV Cure research consortium.

The 2022 meeting, held before the 24th International AIDS Conference (AIDS 2022), was titled 'Pathways to an HIV cure: Research and advocacy priorities' and this reflected a feeling in the HIV cure research community that, as we said in a report on a previous Towards an HIV Cure meeting, “targets were proliferating” in HIV cure research, and they needed a structure.

There was a feeling that researchers were pursuing a number of different and even contradictory hypotheses and needed to reach a consensus on the priority of discoveries that needed to be made, and the order in which we needed to make them, to bring coherence to the field.

One of the problems was that though the first cure achieved – that of Timothy Ray Brown, reported in 2009 – was a proof of concept, and immediately re-established curing HIV as a serious target of research, it was done with a bone-marrow transplant he barely survived. It could never be applied to people who didn’t need such a procedure to cure cancer.

In other words the field has had to tackle an “I wouldn’t start from here” problem that begs the question of where you would start.

The programme this year reflected and was structured around a paper designed to tackle that question. Research priorities for an HIV cure: International AIDS Society Global Scientific Strategy 2021 was published in Nature Medicine last December and set out a series of questions that needed to be answered in order to draw nearer to an HIV cure that will be safe, affordable and applicable at scale to everyone in the world with HIV. It is not the first such paper on cure research priorities but by far the most informed.

The meeting started with two presentations by two co-authors of the paper: Professor Sharon Lewin, the incoming IAS President, and Simon Collins of HIV i-Base, the paper’s primary community advisor.

Lewin said that the research priorities paper itself grew out of the discussions at a meeting at the Sunnylands Center in California, a retreat and discussion centre more used by top politicians. Here the concept of a ‘Target Product Profile’ (TPP) for an HIV cure had been discussed, resulting in a paper in the The Lancet HIV in 2020.

TPP is an acronym we are likely to encounter more often in the future, and means a product that sufficiently combines efficacy, safety, affordability, accessibility and acceptability to be worth taking forward – scientifically, practically, financially, culturally and ethically.

Simon Collins said that repeated surveys of opinion among people with HIV indicated that a cure therapy needed to be safe; affordable in all settings; equally effective, including to people with non-B HIV subtypes; accessible and easy to apply; and permanent. This last criterion implies that people could not be re-infected with HIV, which implies either T-cell immunity or a cure that has vaccine-like properties.

The surveys don’t indicate much concern about it needing to be a once-only therapy. With COVID vaccines, people have got more used to periodic boosters and, as Collins remarked, “With long-lasting antiretroviral therapy and PrEP we may approach something similar to a cure anyway.” When asked, people thought a shot that needed to be given once every three years was about the minimum desirable dosing frequency.

But The Lancet HIV paper acknowledged that the specifications for a TPP might change over time, as products edge closer to the ideal attributes of a TPP. The first generation might not, instance, adhere to the rigorous safety demanded of long-term antiretroviral therapy (ART) use.

The research priorities lays out a set of questions to be answered – and for the rest of this piece we’ll use it as an introduction to topics discussed at the cure meeting, or during the main conference.

The HIV reservoir: how can we map and measure it?

This is a central topic in HIV cure research. An original hope, that HIV would eventually fade away under the pressure of long-term therapy, had been dashed in 1999 when it was established that a small percentage of HIV hides away as a 'reservoir' in long-lived memory cells within the immune system, establishing lifelong infection. 

Although we know a lot more about the reservoir than we did and have used advanced genetic techniques to estimate the number of cells containing HIV in individuals, we still don’t know which cells typically form part of the reservoir, including cells that are hard to access such as those in the nervous system. It may even look different in different parts of the world, depending on the predominant HIV strain and different host characteristics.

We don’t have an easy way to count the number of cells containing HIV DNA in individuals. We do know it keeps itself going by cell division, so that new clones of the same HIV DNA persist when old cells die off; but we don’t know what prompts cells to divide and whether division can be turned off. We do know that a lot of HIV is defective and can’t give rise to new viruses, but we don’t know whether pure chance or specific factors push HIV into becoming defective or ending up in locations where it can’t activate. We don’t know whether and to what extent defective HIV DNA still gives rise to viral proteins that could cause inflammatory states or, conversely, sustain a useful immune reaction against HIV.

In this context the announcement in the main conference of a new assay that, for the first time ever, can finally distinguish between the ordinary T-memory cells in the immune system and the ones that contain concealed HIV DNA, and what their genetic state is, marks a major advance.

How do some people manage to control their HIV? Exceptional elite controllers

A fifth person cured of HIV in California was announced at the conference, adding to Timothy Ray Brown, Adam Castillejo, and people in New York and Düsseldorf in Germany, the latter rumoured to be thinking of going public. But these were all stem cell transplants in which an immune system receptive to HIV was replaced by one resistant to it.

Arguably more interesting, because it's more feasible at this stage for more people, are techniques that concentrate on long-term remission rather than complete eradication.

Firstly, there are the exceptional elite controllers like Loreen Willenberg and the Esperanza patient, who are able to completely control their HIV infection from the start, without needing antiretroviral therapy.

At the cure meeting, Professor Javier Martinez-Picado of the irsiCaixa AIDS research institute in Barcelona, which last year reported promising results for a therapeutic vaccine that helped to control HIV viral levels, said that although these exceptional cases were rare, there were more reported in the literature than these two.

“We are beginning to see a clear phenotype: a weak antibody reaction to HIV but a strong cellular reaction, and a defective virus.”

Four people were reported in 2012 – this included the first report of Loreen Willenberg’s case. They had lived with HIV from 1 to 14 years off therapy without a detectable viral load. They were distinguished from other elite controllers, who have CD8 cells with strong reactivity to HIV, by immune systems with somewhat weaker but very broad reactions – in other words their CD8 cells were sensitive to a wide variety of HIV variants that might arise. They avoided the inflammation that sometimes leads to other illnesses in ‘ordinary’ elite controllers.

Another three cases were reported from Spain in 2020. All three had maintained undetectable viral loads for an exceptionally long time – 25, 28 and 29 years – without any therapy. They all had protective variants in their HLA genes that, in non-controllers, confer slower progression to AIDS, and had strong and broad CD8 T-cell responses to HIV. HIV DNA could be found in their cells, but it was all defective; in most cases the pol and env genes that govern viral entry and gene-copying were absent.

The researchers speculated that CD8 control of their HIV might have pushed it towards replicative incompetence, but also found evidence suggesting that the HIV they were infected with in the first place was poor at attaching itself to CD4 T-cells.

In 2020 a second paper looked at the remaining viral DNA in Loreen Willenberg’s cells but also in one other exceptional controller. This reinforced other findings that these people had experienced a form of immune surveillance that had culled cells containing replication-competent HIV, leaving only cells in which the HIV DNA was exiled to so-called ‘gene deserts’: parts of the genome never likely to be activated.

With the Esperanza patient, we have found nine exceptional controllers who appear to be in long-term and probably life-long remission from active HIV infections, though two from the 2012 paper have been lost to follow-up. Significantly, at least four of them are women (in one patient, their sex was not specified). We know that even with normal courses of HIV infection, women tend to develop lower viral loads, possibly because of hormonal influences.

With these exceptional people, we are beginning to see a clear phenotype: a weak antibody reaction to HIV but a strong cellular reaction, and a defective virus that may have been a poor replicator from the start but is now fatally defective.

Martinez-Picado also described a separate group of ‘viraemic non-progressors’. These maintained significant viral loads in the absence of ART, but also maintained relatively normal CD4 counts for prolonged periods of time.

A study comparing 19 of these people with 24 normal progressors appeared as a poster at AIDS 2022 (see Bayón-Gil in references). They did progress, but very slowly; the average time to developing a CD4 count below 500 was three years in the normal progressors but 17 years in viraemic non-progressors, so they might better be described as ‘slow progressors’.

They were characterised by normal immune reactivity to HIV in their CD4 T-cells cells but, unlike the exceptional elite controllers, low CD8 T-cell activity. Their CD4 cells were deficient in the CCR5 co-receptor cell-surface molecule, suggesting that their state was a halfway house towards the complete lack of CCR5 seen in some people with natural immunity to HIV, and which was engineered in the transplant cases like Brown and Castillejo.

How do some people manage to control their HIV? Post-treatment controllers

The third group of controllers may be the most interesting, because they could be the most numerous, though we will not know how many there are unless they stop their HIV treatment. These are the post-treatment controllers, who have a relatively normal initial course of HIV infection and start ART, but whose virus does not reappear when they stop ART.

Is has been theorised that as many as 14% of people currently on ART, or as few as 2%, might be post-treatment controllers. One was announced last year (the 'Buenos Aires patient') and AIDS 2022 heard about a woman from Barcelona who has maintained viral control for 15 years since stopping ART.

She might be described as having had ‘ART++’ because she received four different immune-modulating drugs in addition to her normal ART, as part of a clinical trial (see our report for a full description). She was the only person out of 20 participants in the trial to maintain long-term viral control off ART, so it is difficult to know whether to ascribe her control to the extra treatment or not.

However, her phenotype was interesting, and very different to that seen in the exceptional elite controllers. Like the Buenos Aires patient, she had had typical or even severe initial HIV infection. Her CD4 T-cells were receptive to HIV and her viral DNA turned out to produce replication-competent virus.

But the CD8 T-cells of her cellular immune system and the natural-killer (NK) cells of her innate immune system both proved to have particularly strong activity against HIV. She had particularly high levels of two cellular subtypes called gamma-delta CD8 T-cells and G2C+ memory-like NK cells. Even if her control was achieved only with extra therapy, these controller phenotypes are interesting because they point the way towards how viral control might be induced in other people.

Why are HIV cure studies with children so important?

We were re-introduced to another post-treatment controller at the cure meeting – a South African boy who aidsmap last reported on when he had maintained viral control off ART for 8.5 years. He is now 13 and still has an undetectable viral load without ART. He was in a study that gave ART to babies very soon after birth and stayed undetectable when he was taken off them at a year of age.

Since last reported, he has had an additional HIV western blot test for antibodies, which found very weak but detectable antibody response to two HIV proteins. His HIV genome and the activity of some of his host genes have been sequenced. No replication-competent HIV has been found. His immune response to HIV is weak, but a gene coding for PD-1, an ‘immune checkpoint’ cell-surface protein that forces immune cells into latency, is exceptionally active.

Interestingly, PD-1 inhibitors, which force immune cells containing HIV DNA to come out of hiding, have been investigated as possible immunotherapies for HIV, but this finding suggests that maintaining PD-1 directed latency might be better than waking it up.

Professor Deborah Persaud was one of the first researchers to present a case of prolonged post-treatment viral control in a child, the 'Mississippi baby', who was able to stay off ART for two years. She emphasised that cure research was particularly important in paediatric HIV.

Firstly, children were losing out when it came to access to standard ART: the 2022 UNAIDS report shows that while 70% of adults with HIV globally are on ART and virally suppressed, only 54% of children are. The COVID epidemic worsened this situation by restricting access to maternity and paediatric facilities.

There are positive reasons to study children as subjects in cure studies too. Infants’ date of infection and ART start date are precisely known, and they can be given ART very soon after infection. Children have far more undifferentiated ‘naïve’ T-cells than effector T-cells, which means that, as cells differentiate and then die over the course of childhood and adolescence, that the HIV latent reservoir often shrinks along with them. This implies that it might be easier to induce deep and irreversible latency in children.

One long-term remission study in children with HIV is IMPAACT p1115, which started in 2015 but whose phases will not all be completed until 2028. It aims to give a large cohort of infants infected before birth several different immediate regimens of ART, including in one arm the broadly neutralising antibody (bNAb) VRC01. Children who reach an undetectable viral load within six months and maintain it for another 14 months are then eligible to be taken off ART, at any time before the age of five, to see if they stay undetectable.

Aidsmap reported on a much smaller study using bNAbs in children last year. The Tatelo study enabled 44% of a group of 28 children to stay virally undetectable on a regimen of two bNAbs alone, without ART, for at least six months. While Tatelo was not designed to demonstrate remission, the use of bNAbs in children, as well as improving adherence and viral suppression, may make long-term remission more likely too.

Targeting the viral reservoir: should latency be reversed or perpetuated?

The idea of an HIV cure is to stop the virus from ever replicating and causing immune damage again, without the need for constant therapy. But there are still not one but several possible pathways towards this end.

Researchers are working on gene therapies that can either precisely target and kill HIV-infected reservoir cells and no other cells, or on gene therapies that work inside cells and can snip out viral code. But these are probably many years away.

In the meantime, the two other principal strategies have been labelled 'kick and kill' and 'block and lock'. The first aims to reverse the latency of reservoir cells and therefore their invisibility to the immune system. The second aims to do the opposite; to make reservoir cells permanently latent and incapable of ever producing virus.

Both strategies are less simple to achieve than they sound. With 'kick and kill', latency-reversing agents exist and have been used in cure research for a decade. However, the ones that are tolerable, such as the TLR-7 agonist vesatolimod and the HDAC inhibitors, though they wake HIV nuclear DNA to the extent it starts to be transcribed into measurable amounts of RNA (the active template for proteins), the amount of viral proteins expressed is generally too small to be ‘seen’ by the immune system and leaves researchers having to invent immune-system boosters that can detect it.

Conversely, stronger latency-reversing agents such as mitogens (cell-division promoters) wake up other immune cells as well as HIV-infected ones, potentially induce dangerous immune over-reactions, can sow more HIV DNA into reservoir cells rather than weed it out, and can even cause cancer.

Kick and kill is still very much a strategy being investigated, but the search is now on for the 'kill' – for immune enhancers such as CAR T-cells, genetically modified T-cells which are more sensitive to HIV-infected cells than natural T-cells, and therapeutic vaccines that can mop up the remainder of HIV-infected cells both in post-treatment controllers and in people with reservoir cells that have been awakened.

With 'block and lock' the problem is that though there are plenty of drugs that can repress gene expression – including HIV reverse transcriptase inhibitors – it was initially difficult to imagine a drug that could induce a permanent lock on HIV genes, long after it is withdrawn. Two more recent developments have stimulated higher hopes for this strategy.

Firstly, it has been found that selective ‘pruning’ of HIV-infected cells by the immune system naturally happens during HIV infection – possibly during times when reservoir cells divide. This leads to the remaining HIV DNA in the cells that survive being locked away in ‘gene deserts’ – parts of the human genome that aren’t genes and never activate.

So block and lock happens naturally in some people, implying that the right therapeutic vaccine could make it happen in more of them.

Secondly, a drug has been found – DCA (didehydro-cortistatin A). This is an inhibitor of tat, HIV’s ‘kickstart protein’, which is the first one made in the viral lifecycle, starting and amplifying viral replication.

DCA may exert lasting effects on the ability of reservoir cells to make HIV. Exactly how it does this is still unclear, but it looks as if inhibiting tat ‘knots up’ the DNA in a process called methylation at a position just in advance of the site where the transcription of RNA should normally start.

At the cure meeting, these two apparently contrary strategies were illustrated by two presentations, one of a tat inhibitor, and one of a tat promoter.

Professor Lishomwa Ndhlovu of Weill Cornell Medical School in New York is a Principal Investigator in a global research consortium called HOPE which stands for HIV Obstruction by Programmed Epigenetics. The type of obstruction that tat inhibition may induce is an example of epigenetics.

HOPE has already discovered at least one new small-molecule tat inhibitor, currently codenamed 44856, and is looking for others.

They are also investigating the genetic landscape surrounding other endogenous retroviruses. Because retroviruses integrate into our DNA, they become part of our genetic inheritance. Most that have been passed on to subsequent generations cause no effects because they have been silenced. HOPE is doing ‘genetic archaeology’ looking for clues in the DNA as to how this happened, to see if it can be made to happen with HIV.

Beyond this, they are exploring the world of overt gene therapy to use a vaccine that could insert a STOP sequence into reservoir-cell HIV DNA. This may sound like blue-sky science but they have already taken the first steps to show that, in mice at least, their experimental delivery vaccine targets exactly the right T-cells in the reservoir.

Meanwhile, Dr Marion Pardons of Ghent University in Belgium outlined early work being done by her consortium, supported by Johnson and Johnson, which is using an analogue of the tat protein, with or without the HDAC inhibitor panobinostat, to shock cells into coming out of latency.

They succeeded in inducing CD4 cells taken from four human volunteers to start producing significant amounts of the p24 capsid protein – the one detected in fourth-generation HIV tests – without causing them to differentiate into the other, short-lived varieties of T-cell that are productively infected by HIV and produce large amounts of new virus off ART.

"'Kick and kill’ and ‘block and lock’ are less simple to achieve than they sound."

So far the p24-positive cells produced this way are being studied to look at the genetic and epigenetic changes accompanying their activation. Using a single-cell assay, they are looking at whether the proviruses in each cell are defective or capable of producing replication-competent whole viruses, and where their sites of integration into human DNA are.

This is a long way from the initial days of 'kick and kill' where it was hoped that waking up the reservoir would lead naturally to a self-cure. And it is interesting to see apparently opposed strategies that aim – in what is probably a still quite distant future – to converge on the same technique, namely being able to either excise or incapacitate the specific lengths of HIV DNA in the reservoir that give rise to whole viruses.

Is there a more gentle way to modify T-cells?

As mentioned earlier, transplants of the entire immune system are not a cure technique ever likely to be performed on more than a few people. But there may be gentler ways of modifying people’s immune systems so their T-cells are less responsive to HIV and ongoing replication ceases, thereby depleting the reservoir. Modifying people’s T-cells outside the body and then reintroducing them is not a new technique, first tried in 2011. But it has proved difficult to maintain the population of modified T-cells.

In 2019 Dr Pablo Tebas of the University of Pennsylvania was able to delay viral ‘rebound’ (the reappearance of detectable HIV viral load) in some of a group of trial participants given T-cells genetically altered outside the body to lack the CCR5 co-receptor. He added in cyclophosphamide, an immune modulator that reduced the number of ‘native’ unmodified T-cells in recipients so the modified ones had more chance.

Tebas has since conducted a third round of studies, dropping the cyclophosphamide but adding in CAR T-cells sensitised to HIV.

CAR T-cells were mentioned in the previous section. All cells naturally have receptor molecules on their cell surface called human leucocyte antigen (HLA) proteins and T-cells have another specialised type called HLA class 2. The job of HLA is to help the rest of the immune system distinguish between ‘self’ and ‘non-self’, in other words between cells that are OK and ones that are not OK and need destroying. Sometimes these are virally infected, but sometimes they have other things wrong with them like being cancer cells. The cells display fragments of the proteins they have inside them on the HLA molecules, like flags on flagpoles.

The problem with the HLA system is that it is not foolproof. Sometimes it starts looking like ‘nonself’ to the immune system, causing autoimmune disease. Sometimes viruses mutate in such a way that they no longer get ‘flagged’ – HIV can do this. Anything inside the nucleus – such as proviral DNA – is not displayed. And sometimes HLA ‘downregulates’, meaning that if the cells sense nothing wrong, they may cease displaying HLA molecules.

Chimeric Antigen Receptors (CAR) T-cells are, so to speak, genetically modified ‘superhero’ T-cells with special powers. They have had their HLA molecules replaced, in a process like vaccination, with receptors that are specifically and exquisitely attuned to HIV proteins (or cancer-cell proteins or anything else recognisable as foreign), do not downregulate, can’t be easily mutated away from, and don’t depend on the baffling variety of HLA genes to do the job of sensing ‘enemy’ cells.

In cancer therapy they are already being used to kill cancer cells. In HIV cure research, they can’t directly attack and kill reservoir cells. But in a study which hopes to be able to take people off ART, they can help to maintain viral control by more quickly sensing and destroying, or flagging for destruction, cells that are beginning to produce HIV proteins.

In the case of Tebas’ new research, his team gave six participants two kinds of modified T-cells: firstly, ones lacking the CCR5 co-receptor molecule, which had also been used in the previous study, and secondly, CAR T-cells. They gave the participants infusions of these two kinds of cell plus ART for eight weeks then took them off ART, only restarting it if their viral load rebounded to over 1000.

Two of the patients maintained low viral loads during their treatment interruption and the other four rebounded to lower levels than a comparison group of six participants who only received the two infusions of modified T-cells once. One maintained a viral load below 50 for 16 weeks and is still under study.

Another rebounded, but only to about 500 copies, has only had one viral load over 1000 in 48 weeks, and has not restarted ART. Essentially the modified T-cells appear to have turned this individual into a viral controller, though not an elite one because their viral load is detectable.

It may be a bit early to start calling this person the 'Philadelphia patient', as Tebas did, but it does begin to show that in the future we might be able to control HIV with methods comparable to those used on Brown, Castillejo and their cured confrères, but far gentler on the system.

A last word

This report on the cure workshop and related presentations at AIDS 2022 has had to miss out several fascinating presentations – one being two studies of people with HIV who, with terminal non-HIV-related illnesses, consented to have their tissues autopsied after death to measure the amount of HIV DNA in different organs. We’ve also excluded the panel discussions.

A cure for HIV that could be a universally available pill or injection is still many years away. But we have a much clearer idea of what it will have to involve instead of, as before, just trying different therapies and seeing if they will work.

As Simon Collins said, “A cure is always at least ten years away. But somehow this ten years feels closer than the previous ones.”