In 1996, scientists thought they were on the verge of curing HIV. Then they discovered the virus’s secret: it can lie dormant, virtually hidden inside human cells. Frustrated, some of the best minds in the world started using words such as “evil” and “diabolical” to describe a virus. Within a year, the search for a cure had petered out.  

Here’s what happened. When HIV was discovered in the early 1980s, scientists quickly figured out how it leads to AIDS. The virus attacks T cells—the soldiers of our immune systems—and meshes with human DNA. Then HIV uses enzymes to replicate so it can attack more T cells. Eventually, the immune system becomes compromised. Patients develop AIDS and die, usually of opportunistic infections or cancer.

National Institutes of Health

But in the mid-1990s, scientists created a second class of drugs—there are now six—to disrupt viral replication. When infected cells can’t replicate, they eventually die. Doctors began giving patients drug cocktails to beat back their viral loads. In most patients, so many infected T cells would die over the course of several months of treatment that standard blood tests couldn’t even detect the virus. Meanwhile, the body would naturally produce healthy T cells, the immune system would strengthen, and patients wouldn’t develop AIDS.

Some scientists thought they could completely wipe out HIV. But the virus had a trick up its sleeve. HIV can infect a cell and then go quiet—scientists use the word “latent.” The virus just sits inside the T cell as genetic information. It doesn’t replicate. It doesn’t express enzymes involved in replication. Medications can’t recognize the infected cell as anything but a typical T cell. It’s as if the virus just falls asleep.

In an average person with HIV, only one in a million T cells is latently infected. Sometimes it’s as many as ten in a million or as few as one in a billion. But all patients have latently infected cells. Standard blood tests still can’t detect such a small viral load; researchers have to use a laborious procedure to find latently infected cells.

UNC researchers Joe Eron and David Margolis found that latently infected cells are present in a patient within three weeks of the initial HIV transmission. And Eron suspects that latency doesn’t happen by accident. It’s as if the latent virus was always meant to go to sleep. Then, when patients stop taking medication, it’s as if the latent virus realizes it has a job to do. “It might take two weeks or several months,” Eron says. “But the virus will wake up.” It will produce typically infected cells and latently infected cells. And patients have to resume medication to avoid developing AIDS.

As soon as scientists discovered HIV latency, some of the most prominent HIV researchers thought there could never be a cure because the latent virus so insidiously entwined itself with human DNA. Finding the couple-hundred base pairs in the entire HIV genome that cause the virus to go dormant seemed next to impossible. At conferences, some of the most respected HIV researchers discouraged young investigators from pursuing certain avenues of research. They even discouraged colleagues from using the word “cure” because finding one seemed so far-fetched that scientists didn’t want to give HIV patients false hope.

Research dollars went instead toward vaccines, prevention, and better treatments. But Margolis and Eron didn’t buy it. They saw latency as a problem to be solved.

Six years ago, when Margolis began looking for what he calls curative therapies, he compared solving HIV latency to curing cancer. No one discourages cancer researchers from using the c-word; no one tells them to stop searching for cures.

“With chemotherapy, you’re trying to kill a cancer cell that’s very much like a normal cell,” Margolis said in 2007. “But it has some genetic differences.” A latently infected T cell is very much like a normal T cell with some genetic differences. “We want to find out what the differences are,” he said, “and how to exploit them.”

Donn Young

Back in 2004 Margolis started investigating a class of drugs that he thought might cause latently infected cells to express their HIV genes and start replicating. He thought standard HIV therapy would recognize the enzymes used in replication and essentially kill the virus.

The first drug Margolis’s team used was valproic acid, a treatment for epilepsy and bipolar disorder. The drug disrupts many different enzymes, and Margolis was confident that a few of those enzymes were the same ones HIV used to stay dormant. He gave valproic acid to four patients, three of whom showed a significant depletion of latent infection at the end of three months. But in a later study, only four of eleven patients showed a significant decline.

Margolis says there were two major problems with the clinical trial. First, waiting three months wasn’t the best way to figure out what valproic acid was doing to latently infected cells. Second, valproic acid disrupts a slew of enzymes involved in gene expression, but HIV hijacks specific enzymes to stay hidden. A compound that targeted those specific enzymes would be more potent than valproic acid.

Margolis thought of vorinostat, a drug that was developed to treat cutaneous T-cell lymphoma and, compared to valproic acid, is better at inhibiting only the enzymes HIV uses to stay quiet inside T cells. If vorinostat could fit properly into the spaces where those enzymes go, then the latent HIV would be forced to replicate. And if the virus replicated, then standard therapy could disrupt the replication and leave the virus for dead.

Margolis created a basic study of vorinostat to find out exactly what the drug did to latently infected cells at a moment in time.

First, his team took infected cells from patients and added vorinostat to them. On a lab plate, the latent virus started expressing its genes. That was a good sign. Then Margolis’s team gave a dose of the drug to patients, including two of Eron’s. The researchers measured how quickly the drug entered the patients’ blood, how much of the drug was in the blood at specific times, and how long it lasted there. This allowed Margolis to pinpoint when the lone dose of vorinostat would be most potent. “Then we gave each patient another dose of the drug,” he says, “and measured the HIV expression in the precise window of time when we thought we’d see a measurable effect on latent cells.”

In the first four patients, Margolis found a significant increase of RNA inside latently infected cells. RNA expression was a sign that the virus was expressing its genes and trying to replicate. “And that’s evidence that the virus was being forced out of its hiding place,” Margolis says. Three other patients were given a single dose of vorinostat, and Margolis saw a significant RNA response in each of them, as well. That’s seven for seven.

“Now we know vorinostat stirs up the reservoirs of latent cells,” he says. “But we’ve only looked at a snapshot in time. The virus could come out of latency and go back in when there’s no drug present.”

He says that if the latent cells are exposed to the drug over time, all of the virus that can come out will come out. “The question,” he says, “is going to be: once that process is done, will we have really gotten rid of the virus completely?”

HIV primarily infects CD4 T cells in lymph tissue throughout the body. But macrophages—a different kind of immune cell that’s among the first infected during HIV transmission—might also contribute to the long-term persistence of HIV, Eron says. And some scientists think macrophages carry HIV into the brain, though it’s not clear how long HIV-infected cells in the brain can persist in a patient who is on antiviral therapy. “We’d love to answer that question,” Eron says. “And we’re going to have to.”

Donn Young

Eron says scientists in the 1960s and 70s thought they had cured kids with leukemia. “But then patients relapsed,” he says. “Chemotherapy didn’t get into their brains, but a few cancer cells must have.” This observation forced scientists to modify treatments, and now children can be cured. The same situation could unfold with HIV.

Also, there might be other reservoirs throughout the body that current therapies don’t reach. “We don’t think this is going on,” Eron says, “because when HIV replicates, it makes errors. So usually we can see when the virus is replicating, and we don’t see any evidence of that in patients on medication.”

Those are among the problems yet to be solved, but the biggest, scientists agree, is latency. It’s the major roadblock to finding a cure—something that patients in clinical trials know well. “We don’t want to give patients false hope,” Eron says. “And I don’t want to oversell what Dave has shown.”

So Eron uses a NASA analogy when talking to patients. In the 1940s, no one thought it was possible to get out of Earth’s atmosphere. Then the Mercury astronauts did it. “But they were nowhere near landing on the moon,” Eron says. “To cure HIV, you have to wake up latent cells, but we’re still a long way from curing patients.”

Margolis, though, has come further than anyone thought possible just ten years ago.

“There were plenty of people who said we wouldn’t be able get cells out of latency, that we were wasting our time,” Eron says. “But Dave’s shown, especially over the past six months, that we can find these cells and we can wake up the virus in these cells. Now the next step is to figure out how to kill the cells.”

Margolis is seeking FDA approval to give patients more doses of vorinostat over time to see if that—along with standard therapy—can deplete the reservoirs of latent virus.

“I’m optimistic because we have lots of ways to enhance the immune response and kill infected cells,” Margolis says. “It’s just that the virus always escapes the immune response. But under the cover of antiviral care, we hope the virus won’t have anywhere to hide.”



David Margolis is a professor of medicine and microbiology and immunology in the School of Medicine, professor of epidemiology in the UNC Gillings School of Global Public Health, and director of the Program in Translational Clinical Research. Joe Eron is a professor of medicine in the School of Medicine and director of the clinical core for the UNC Center for AIDS Research. Their study on latency was funded by the National Institutes of Health, the pharmaceutical company Merck, and the James B. Pendleton Charitable Trust.