The obstacle to curing HIV is not the virus circulating in blood — antiretroviral therapy handles that. It is the latent reservoir: infected cells that persist silently, invisible to the immune system and untouched by drugs, ready to restart the infection the moment treatment stops. An approved cancer drug has now shrunk that reservoir in monkeys.

Researchers at Emory University, led by Mirko Paiardini, PhD of the Emory National Biomedical Research Center, published the work in Nature Microbiology on September 3, 2026.

Why the reservoir is so hard to remove

When HIV infects a CD4+ T cell, it integrates its genetic material into the cell’s own DNA. Most infected cells die or are cleared. A small number enter a resting state in which the virus is not actively produced — and a cell making no viral protein displays nothing for the immune system to recognise, while antiretrovirals, which block steps in active replication, have nothing to act on.

These cells are long-lived and can self-renew. Stopping therapy allows some to reactivate, and viral load rebounds within weeks. It is why HIV requires lifelong treatment rather than a course.

The BCL-2 idea

The strategy tested here targets not the virus but how the infected cells stay alive.

BCL-2 is a protein that blocks apoptosis — programmed cell death. Latently infected CD4+ T cells appear to overexpress BCL-2, which gives them a survival advantage and helps explain why they persist for years. Venetoclax is a BCL-2 inhibitor already approved for certain blood cancers, where it works by removing exactly that protection and letting malignant cells die.

Applying it to HIV means using an existing, well-characterised drug to strip the survival advantage from reservoir cells — an approach with a substantially shorter path to human testing than a novel molecule would have.

What the study did

24 rhesus macaques infected with SIV, the simian counterpart of HIV, were divided into three arms: antiretroviral therapy alone, ART plus venetoclax, and ART plus venetoclax plus CD8α depletion. Venetoclax was given as 10 daily doses starting when ART began, 14 days after infection. Animals were followed for nearly 10 months.

The timing is notable — a very short course of drug administered at the moment the reservoir is being established, rather than prolonged treatment after it has settled.

The results

Venetoclax-treated animals showed a rapid decline in intact SIV DNA within blood CD4+ T cells relative to controls. Intact viral DNA is the meaningful measure: most integrated HIV or SIV DNA is defective and incapable of producing functional virus, so counting total viral DNA overstates the true reservoir.

The reduction was sustained through day 294 after infection, reductions were also seen in lymph nodes — a major reservoir site less accessible than blood — and absolute numbers of circulating SIV-infected cells fell. Most striking, the effect persisted for months after venetoclax was stopped, while ART continued.

“Eliminating the viral reservoir is a priority in pursuing an HIV cure,” the team noted. “Our study offers hope for accelerating the timeline to cure using approved medication.”

The limitations the authors flag

The most important gap is the absence of an analytical treatment interruption — the experiment in which therapy is stopped to see whether and when the virus rebounds. Without it, the study shows the reservoir got smaller but not that the animals are any closer to controlling the virus without drugs. Rebound can occur from a very small number of surviving cells, so reservoir size and time to rebound are related but far from equivalent.

The researchers also identified compensatory survival pathways — elevated BCL-xL and reduced PUMA expression — indicating that cells adapt to BCL-2 inhibition and that clearance was incomplete. They recommend longer dosing studies and further mechanistic work before human trials.

Why the 14-day timing is both a strength and a limit

Venetoclax was given as antiretroviral therapy began, 14 days after infection — and that choice cuts both ways.

It is well established that starting antiretroviral therapy early in infection results in a substantially smaller latent reservoir than starting late, because the reservoir is seeded progressively during uncontrolled replication. Intervening at the moment ART begins therefore attacks the reservoir while it is still small and while the cells forming it are newly established rather than settled into long-term persistence — plausibly the most favourable moment such a drug will ever have.

Which is precisely the limitation. Most people living with HIV are diagnosed and treated well after infection, often years later, with a reservoir that is larger, more genetically diverse and maintained partly through the slow self-renewal of long-lived cells. Whether a ten-day course of a BCL-2 inhibitor can make a comparable dent in that far more entrenched population is the central unanswered question, and it is not one this study was designed to address.

What it would take to reach patients

Venetoclax being approved shortens the path but does not clear it. In oncology the drug is used in patients with life-threatening malignancy, where meaningful toxicity — including neutropenia and tumour lysis syndrome — is an accepted trade-off. People living with HIV on effective antiretroviral therapy have a near-normal life expectancy, which sets a far higher safety bar for any addition to their regimen.

The dosing schedule here is encouraging on that point: ten days is a short exposure, and the durability after stopping suggests a brief intervention may be enough. Whether a comparable effect can be achieved when treatment begins years into established infection, rather than 14 days in, is the question this study cannot answer. Preclinical research; not medical advice.