Novel drugs that turn HIV against itself might clear infected cells from the body

Scientists from the pharmaceutical company Merck have published the first detailed data about a subset of the familiar non-nucleoside reverse transcriptase inhibitor (NNRTI) drugs that, in addition to inhibiting HIV’s ability to copy itself, may be able to kill off HIV-infected cells by causing the virus to ‘poison’ the cells it depends on for replication.

At a later stage in the viral life cycle than the initial inhibition of reverse transcriptase, these drugs cause the proteins that will comprise the new generation of viruses to be synthesised prematurely. One of these proteins is HIV’s protease enzyme, which is normally employed at the final stage of HIV replication to snip individual viral components out of a ‘polyprotein’ that the virally infected cell has made, like a tailor cutting a pattern out of whole cloth.

The premature synthesis of protease, however, causes the enzyme instead to snip out cellular proteins that form a so-called inflammasome – a self-destruct bomb that almost literally blows up the cell.

The compounds that can do this have been dubbed Targeted Activators of Cell Kill (TACK) by Merck.

The potential importance of the TACK drugs lies in the fact that for the first time, they suggest a cure therapy might be possible that does not rely on host immune responses to make it happen and that, by definition, only destroys HIV infected cells.

HIV is lifelong because some of the instructions that it inserts into our DNA for making new viruses are hidden inside a population of long-lived ‘central memory’ cells. Their job is to do remember previous infections so that the immune system can mount a swift and specific response should they ever appear again, sometimes decades later. Till then, they wait. The immune system’s ability to remember is also the basis of vaccines. The population of HIV-infected memory cells has been called the ‘HIV reservoir’.

The ‘kick and kill’ strategy for curing HIV is to wake up the reservoir cells using drugs that stimulate their activity, and thus make them visible to the immune system.

At first it was hoped the immune system would then be able to pick them off by itself, but as it did not, they are now targeted with therapeutic vaccines, antibodies or modified T-cells to kill them off. The trouble is that while there has been modest success in ‘kicking’ the reservoir cells into action, there has been little success in achieving ‘kill’ – the shrinkage of the reservoir. This is partly because not all reservoir cells can be activated at once but mainly because immune therapies cannot be targeted solely at HIV-infected cells. Anything powerful enough to kill all HIV-infected cells would kill a lot of other cells – a familiar problem in cancer chemotherapy.

The TACK drugs solve this problem by only using HIV’s own proteins to effect a kill – so only target HIV-infected cells. The researchers believe cells would probably still have to be activated but the other advantage of TACK drugs is that they belong to a drug class, the NNRTIs, that we already know a lot about and can probably be used just like regular ART as well.

The TACK properties were in fact first noticed in some of the NNRTI drugs we already use, including efavirenz and rilpivirine. They seemed to increase the death rate of activated HIV-infected cells. The first NNRTI, nevirapine, on the other hand, completely lacks this property. The authors do not comment on the other two NNRTIs, etravirine and doravirine, though they note their leading candidate is chemically similar to doravirine.

This leading candidate currently has the investigatory name of Pyr01 (for pyrimidone, the chemical class it belongs to). It, and to a lesser extent these other NNRTIs, work at two stages of the HIV lifestyle.

"The protease helps to assemble an inflammasome – a kind of self-destruct module that sends out toxins that destroy the cell."

The first is the familiar one common to all Reverse Transcriptase Inhibitors. They interfere with an early, ‘upstream’ stage in HIV’s life cycle. After the virus has entered a cell and shed its outer layers, its first job is to turn its genetic material – its instructions for making new viruses – into instructions human cells can read. The viral genes are encoded on to RNA – a rather fragile molecule. The Reverse Transcriptase enzyme ready-packaged inside HIV transcribes this into the more robust form of DNA. Once this is done, in the central event in HIV infection, the DNA is integrated into the human DNA of the host cell. The subverted cell then starts making new viruses instead of its proper job, guarding against them.

The TACK activity happens in the latter, ‘downstream’ half of the life cycle, when the infected cell turns out HIV proteins. Among these proteins are the components of the new reverse transcriptase for new viruses.

Reverse transcriptase is a complex molecule with several tasks, but to do its job properly, it must ‘dimerise’, meaning that two proteins pair up and stick together as a dimer – a molecule made of two separate units. The TACK drugs make this happen too quickly – in a matter of a few hours rather than a day or two. This initiates a chain of events that involves HIV protease being made prematurely too. The protease then, instead of assembling new viruses, helps to assemble a protein complex called an inflammasome – a kind of self-destruct module that sends out toxins that destroy the cell.

First data on a new TACK drug

How effective is the lead candidate Pyr01? In lab dish experiments, its ART activity - the upstream blockage of reverse transcriptase – was similar to that of efavirenz. Its IC₅₀ – the concentration needed to reduce viral replication by 50% – was 39.7 nanomols (nM), which is close to that of efavirenz (34.1 nM – the lower the figure, the more potent the drug). Its ability to kill HIV-infected cells, however, was 100 times greater than that of efavirenz (38.4 nM versus 4006 nM in the case of its ability to kill HIV-infected CD4 cells). It was not toxic to non-infected cells.

Its ability to dimerise the HIV RT molecule – to force its premature maturity – was ten times greater than that of efavirenz, at 24 nM versus 210 nM, and its ability to do this was not affected by drug resistance. Whereas the most common NNRTI drug resistance mutation, K103N, completely annulled the ability of efavirenz to dimerise reverse transcriptase, it did not affect Pyr01 at all. Few of the common NNRTI resistance mutations or combinations of mutations affected Pyr01’s ability to dimerise, with no more than a threefold increase in IC₅₀ for single mutations and a tenfold decrease – still probably clinically manageable – for a threefold combination of resistance mutations, K101E, G190A and Y181C.

The Merck scientists also investigated a compound very similar to Pyr01 they called Pyr02. Even though it only different to Pyr01 in the placement of two atoms in its molecule, it had a rather weaker ART ‘upstream’ effect (its IC₅₀ was 131nM, though that is still stronger than the effective drug nevirapine, at 214nM).

But it had hardly any ability to force ‘downstream’ reverse transcriptase to adopt its dimer conformation prematurely. There is a natural mutation in HIV’s reverse transcriptase molecule called W401A that means it cannot dimerise and so the virus cannot replicate: but Pyr01 could overcome this and still force dimerisation, leading to the TACK activity, whereas Pyr02 could not. Because Pyr01 and Pyr02 are so similar chemically yet so different in their effects, this has enabled the scientists to ‘zoom in’ on the specific molecular bonds that make Pyr01 so effective.

The fact that it is indeed the premature production of HIV protease that kills the infected cells was proven by adding the first-generation protease inhibitor drug indinavir to cell cultures containing infected lymphocytes and Pyr-1: the addition of indinavir negated the action of Pyr01 and the cells weren’t killed. This also shows that if they get to the clinic, TACK drugs and protease inhibitors will be incompatible.

"Are efavirenz’s neuropsychiatric side effects related to its cell-killing TACK activity?"

The experiments described above were on cells infected with a defective form of HIV in the lab dish that could only undergo one cycle of replication (which showed that the effect of Pyr01 was swift, specific and highly efficacious).

But the researchers also took CD4 cells from people with HIV to directly measure their HIV replicative abilities. They took CD4 central-memory cells (not just cells containing HIV) from eight volunteers on ART, activated them to allow some HIV replication within the HIV-containing cells, dosed them with Pyr01, Pyr02, efavirenz and nevirapine and measured their TACK activity. While this confirmed that efavirenz had significant TACK activity, with an 85% fall in HIV gag protein (a viral component) in the sample fluid, Pyr 01 produced a 94% fall. As these samples were from subjects already taking ART, the further suppression of HIV gag production was attributed to cell-killing activity.

The cell activity in these subjects was not enough to produce countable numbers of HIV-infected CD4 cells so the researchers also took central memory CD4 cells from one person who was not virally suppressed and treated them with Pyr 01 and efavirenz, producing similar drops in HIV proteins but also a drop in HIV-infected cells that was directly countable via flow cytometry. In contrast, the number of HIV-uninfected memory cells was unchanged, showing that the drugs were not killing these cells.

Unanswered questions

This is a very early, pre-clinical investigation of these interesting drugs and their bimodal, ‘upstream’ and ‘downstream’ effects on HIV’s reverse transcriptase enzyme. There are lots of unanswered questions.

Although efavirenz shows TACK activity too, it is ten times less potent and Pyr01 is much less susceptible to NNRTI resistance mutations. But the other reason efavirenz has largely been discontinued and would not be taken further as a TACK drug  is because of its neuropsychiatric side effects – sleeplessness, nightmares dizziness and depression. Are these related to its cell-killing TACK activity? We need safety studies to show if Pyr01 and related pyrimidone compounds (Merck has found several with similar properties) have similar side effects or if they’re unique to efavirenz.

The other unanswered question is whether TACK drugs will be able to ‘reach in’ and by their mere presence eliminate cells in the quiescent HIV reservoir. It seems unlikely in the case of completely latent cells. But the good news is that Pyr01 will probably work as a typical ART drug, as well as a TACK drug. We know that ART still allows a degree of low-level HIV replication to carry on (which is why a typical viral load on ART is in the region of 2 to 4 copies/ml, not zero). But this implies that the TACK drugs may be able to mount a slow campaign of assassination against any reservoir drugs that dare to activate and start making viral proteins, strengthening the process whereby HIV-infected reservoir cells become more deeply latent over time and possibly producing many more post-treatment controllers.