HIV's Hiding Places
Tracking latent infection, scientists confront new fears about the future.
by Anne-christine d'Adesky

Over the past two years, the world's attention on HIV has shifted to a new obstacle in our battle: Latent infection. While we've developed potent therapies to treat active HIV infection and curb the disease, many reports have documented that even when no virus can be found in the blood, a tiny amount remains locked inside long-living immune T-cells in lymphoid tissue compartments and so-called "sanctuary" sites like the brain. In essence, HIV causes two parallel infections-active and silent-and how they are connected forms a new area of research.

Rather than one big fight, researchers describe many small or local battles being waged on different fronts. The infection may smolder in one site, then rekindle itself and cause a brushfire elsewhere. There is also evidence that in people on HIV therapy, different strains of virus with different mutations can be found in protected compartments like the testes and female reproductive tract, a discovery that has implications for sexual transmission of the disease. Looking at the big picture, these findings move the goal of treatment away from complete eradication of the virus to the more viable goal of remission, using new vaccines and other strategies to harness the power of the immune system to control any outbreaks of the virus. If this approach works, the future might look very different for people now struggling with difficult drug regimens that require taking a dozen or more pills a day on a strict schedule. The emergence of easier maintenance regimens would also address mounting concerns over side effects that are the downside of HIV therapy (see "The Good, the Bad, and the Ugly").

A Stable Pool
That's the rosy picture of the future. The present reality is rather different and a bit darker. Over at the Johns Hopkins School of Medicine, viral hunter Bob Siliciano has revised his earlier estimates of the pool size and life expectancy of latently infected "memory" T-cells. These cells are found in blood and lymphoid tissue and are thought to be the main reservoir for HIV, though other reservoirs exist. Memory T-cells are a subset of once-active immune cells that have returned to a dormant state but carry a "memory," or ability to recognize and fight infections. According to Siliciano, only a tiny fraction of the body's memory cell pool is HIV-infected in people with chronic HIV infection; the number is smaller in recently exposed individuals.

Searching through millions of resting T-cells in blood and lymph node samples of people on HIV therapy, he occasionally comes across a cell with a virus particle (called provirus) lodged inside or just outside its nucleus. Some of these rare cells carry virus particles that can reproduce and are infectious; in other cases, the virus particles appear defective or can't reproduce. Hundreds of copies of HIV can be released from a single cell when it reproduces itself (see "Copycat: How HIV Enters Cells").

The bad news is that this latent reservoir appears stable, even in people on therapy who have no signs of virus in their blood. That means the cells that harbor HIV aren't gradually dying or being affected by current therapies. Normally, when an infected cell dies naturally, so does the virus it carries. "We're now looking at estimates of at least 60 years for this pool to die out," says Siliciano, wishing it wasn't so. While that doesn't mean eradication is a bust, right now the virus holds a trump card.

Sixty years is a lifetime, and for those now coping with harsh drug regimens and side effects, it represents an eternity. "What's clear is that we're gonna need something else to tackle this pool," says Siliciano, adding, "It may not be feasible for some people to stay on these regimens for their entire lives. We've also seen that some people can do just fine."

Frontline scientists are now looking to other latent and chronic diseases for clues about what a lifetime of HIV might look like. Some sexually transmitted herpesviruses, for example, permanently lodge themselves in nerve cells and can stay dormant for years, unless a bout of the flu or lowered immune defenses spark an active outbreak. For his part, Siliciano points to diabetes as a good model for how we might approach long-term management of HIV. "Many physicians don't tell patients that this is like having diabetes," he says. "Insulin-dependent diabetes will certainly kill you if you are not treated, but with treatment, you can live a normal life."

Fears of Transmission
The public health implications of latent infection are quite serious, raising questions about how well we are tracking those who may be infected with HIV and who could pass on the virus unwittingly. In a small study recently presented at the Sixth Conference on Retroviruses and Opportunistic Infections in Chicago, a group of 37 "exposed seronegative" people who tested negative on HIV-antibody tests were actually found to be latently infected when more sensitive tests were done. Some of them harbored an infectious virus. Latent virus has also been found in long-term survivors with HIV who remain healthy. In this group, HIV-specific immune responses appear essential for controlling the virus. The immune responses are usually lost in people with chronic HIV infection. Studies of individuals treated with HIV drugs shortly after exposure also show that drugs can preserve these protective immune responses and stop HIV from taking off.

There are also renewed concerns about sexual transmission of undetected latent infection and the role that genital secretions may play. In both men and women on HAART, latently infected cells have been found in genital secretions when there is no detectable trace of virus in the blood. In other cases, virus levels in semen and vaginal secretions are simply different from those found in the blood. This may be because different compartments in the body are surrounded by different barriers. "Sanctuary sites," like the testes, brain, and retina, are protected by a blood-tissue barrier that allows the virus to hide within them. The female genital tract is not considered a real sanctuary site due to its lack of a true blood-tissue barrier. The difference between a reservoir and a "sanctuary site" is basically that viruses can travel more easily into and out of reservoirs.

Groundbreaking studies by Roger Pomerantz of the Center for Human Virology at Thomas Jefferson University suggest that these virus particles may be more infectious in the semen of the patients he has studied than in their blood, based on laboratory tests. Pomerantz has also proposed that viruses may develop differently depending on the tissue reservoir where they are produced. To further complicate matters, Joe Eron, a researcher at the University of North Carolina at Chapel Hill, has detected different patterns of drug resistance in viruses found in the semen and blood of some men who experienced drug failure on HAART. Based on this data, he has warned of the risk of sexual transmission of multiple-drug-resistant viruses, something that was reported in a first case last fall. Eron's work also supports the notion of local or site-specific battles occurring in different places.

These findings raise the additional concern that in people on therapy with no detectable virus in their blood, latently infected cells will escape from protected sites like the testes and trigger a new infection, something called reinfection. Recently, University of Alabama at Birmingham vaccine researcher Patricia Fultz proved she could establish reinfection in nine of nine chimpanzees, so we know it's possible.

But is it happening? That's a different question. A number of studies show that in people on HAART therapy who temporarily quit their drugs and have seen HIV "rebound," the virus that comes back is "wild type," meaning that it doesn't contain any mutations and is sensitive to drugs. These are the viruses that first infect the body and form the initial latent pool, thinks Siliciano. Until we know more, he urges us to be careful. "I think we have to be on our guard with respect to reinfection," says Siliciano. "We don't know enough about what's happening to people yet." But what could also be happening is that viruses in sanctuary sites are being contained by a HAART-bolstered immune system. New studies of primates with SIV show that viruses in such protected sites as the brain don't traffic out when the immune system is strong but do so regularly when it is weakened (see accompanying story). By suppressing HIV, antiviral drugs strengthen the immune system.

Tracking The Virus
Ashley Haase is a slow virologist, which doesn't mean he's a slow scientist, only that he's interested in viruses that don't reveal themselves or their actions right away. Boyishly handsome and friendly, with a flat Minnesota twang and '50s-style dark-framed glasses, Haase has led the field of HIV with pioneering studies of where the virus heads and how to find it when it first enters the body. About 30 years ago, as a junior faculty member at the University of California at San Francisco, he began poking around sheep viruses called visna that are members of the lentivirus family (lenti, as in slow). "When I was going through residence, a fellow resident talked about a mysterious group of unconventional agents that caused a disease called Kuru 25 to 30 years after exposure to the agent," recalls Haase. (Kuru has since captured headlines due to its link to "mad cow disease," an infection that appears to have crossed over from sheep and cows to people.) "It was fascinating to me that you could have an infectious agent that could be transmitted and not be manifested for so long," he adds, recalling that his teachers weren't as impressed. "They'd say, 'What are you gonna be, a sheep pediatrician?'" he says with a laugh.

In the early '80s, when HIV suddenly appeared, Haase quickly observed that this myserious new virus shared the classic features of a lentivirus. With HIV there is an acute, transient infection followed first by an inactive state lasting months or years and then by the advent of symptoms leading to the progressive immune dysfunction known as AIDS. "Visna is a distant prototype for HIV," explains Haase. "The original hypothesis I had in visna was that these viruses persist because they go underground. They're introduced into a cell and complete part of the virus' life cycle and then stop. And because the immune system can't see the cell that's infected, [the virus] persists for the lifetime of that cell, whatever it turns out to be."

That also makes it harder to study viruses in tissue. Meanwhile, getting a blood test is easy, but giving over a lymph node sample requires invasive surgery, a procedure for which only a few brave individuals have volunteered. More often, animal models are helping fill in the picture of latency. Primates are natural hosts for SIV, a virus virtually identical to HIV; but unlike humans, primates with the virus don't usually get sick. About 10 years ago, Haase began testing his theories, looking at lymphoid-tissue samples taken from humans and primates to see whether the virus was stowed away in dormant tissue cells. Why there? he's asked. "It's sort of why Willy Sutton said he robbed banks"-Haase shoots back quickly-"because it's where the money is." Since HIV causes an immune disease, it made sense to look at the tissue reservoirs of the immune system. The work, he adds, was like "looking for a grain of sand in a pile of sand."

But it paid off: Several years ago he issued a somber report showing that lymphoid-tissue reservoirs harbored many infected cells and that copies of the virus could be seen inside and on the surface of immune cells called follicular dendritic cells (see "Who's Who Among Cells"). The infected cell population-counting cells both latent and those actively producing virus-included T-cells, another type of dendritic cell, and macrophage cells. The latter two cells represent the first line of defense against foreign invaders in tissue.

In a recent update, Haase showed that a subset of T-cells are actually the initial target for HIV and make up 90 percent of the latently infected cell pool; macrophages and dendritic cells make up the other 10 percent. What happens after that, where and how the virus spreads, is a matter of good guesswork.

"Not nearly as much has been done with measuring the level of infection in macrophages in vivo [in the body]," admits Siliciano. "Part of the problem is that macrophages don't circulate around; they pretty much stay where they are and it's harder to get at them and study them." It's easy to study T-cells, he explains, because they're always on the move. "You can just take some blood and catch a sample of them."

Early Events
Shortly after Thanksgiving of last year, Haase was a guest speaker at a small meeting of vaccine experts in Virginia who spent three days politely, and sometimes pointedly, arguing where the virus goes during the first minutes and hours after infection and where it later turns up. Most of the work is being done with SIV and macaque monkeys.

This emerging picture of sexual transmission suggests that HIV first enters the genital tract inside semen, either as free-floating virus or inside seminal cells. It then heads for the mucosal lining of the female reproductive tract (or if transmitted anally, the rectum). The very next step is not known, says Haase. But as early as day three, dendritic and macrophage cells lying close to the membrane grab this viral invader and carry the virus over to T-cells in the lymph nodes or to those found deeper inside the mucosal tissue of the lamina propria, just below the cervix and vagina. The lamina propria is an elastic membrane that lines the reproductive and digestive tract. Another possibility is that the T-cells directly engulf the virus and carry it to the lymph nodes.

As the immune system becomes alerted to the presence of the invader, other immune cells are activated and reproduce, spreading the infection to other cells in the tissue, as well as to follicular dendritic cells and macrophages. In a study of acute SIV transmission, Haase found that by day seven there was a sevenfold increase in the number of infected cells in the lamina propria, and by day 12, an explosive 800-fold increase in this area.

In a further step, the infected T-cells travel from the lymph nodes to the primary lymphoid organs, then to secondary lymphoid organs. At the Chicago meeting, Andrew Lackner of the New England Regional Primate Research Center reported that in his studies of acute infection in macaque monkeys, most of the SIV-infected cells are found in the T-cell zones of the lymph nodes, spleen, and thymus. Researchers have also named the retina, tonsils, central nervous system, brain, and possibly the kidney and lungs as other viral hiding places.

As some of these activated yet infected T-cells revert to the memory state, they form the initial latent cell reservoir, suggest researchers. Other types of tissue cells like macrophages form a secondary infected pool. But exactly how does this happen? At the Chicago meeting, a Rockefeller University researcher named Angela Granelli-Piperno presented evidence that HIV is captured by dendritic cells and carried over to T-cells in the lymph nodes. In laboratory tests, reverse transcription of the virus in dendritic cells only took place when T-cells were present. This might help explain how the virus infects different cell types in tissue reservoirs.

Others have noted that a great number of immune T-cells die during this early period, a phenomenon called cell suicide, or apoptosis, that is poorly understood. In his SIV studies, Lackner found that this widespread T-cell death occurs just as large numbers of tissue macrophage cells in the intestine are getting infected.

Dormant or Secretly Spreading?
Up to this point in the debate, there's been a general agreement among experts about some aspects of acute infection, but not all. For example, Haase says he has come under fire for suggesting that unactivated T-cells may be the real players behind the latent story rather than activated ones and that macrophages play only a bit role. This radical theory is based on his belief that HIV is somehow spread by local or neighborly contact without requiring cells to undergo the normal process of activation and reproduction. As he explains, HIV can enter dormant T-cells without waking them up. Other steps are required to activate the cell allowing HIV to reproduce.

"There's gonna be a lot of contention and discussion about this," admits Haase. "We're not saying these cells are resting cells (memory T-cells); we actually don't know what they are. All we know is that those cells don't have activation markers. What it does is lower the bar for a cell that is producing virus to pass it on to its neighbor: It doesn't have to be activated."

Siliciano isn't convinced: "It is very clear from a lot of work that in a truly resting cell, the virus does not replicate. So what he [Haase] may be seeing are cells that are in the process of being activated or had been previously activated and are going back into a resting state." He offers an alternative theory: "You could have what I call sort of a smoldering infection with ongoing replication that doesn't actually involve a formal state of latency and that is occurring even though the patient is on therapy." All you need to sustain a latent infection, he suggests, is for one infected cell capable of releasing many virus particles to infect another cell before it dies.

There are several other theories bopping around to explain what researchers think may be the mysterious local spread of HIV in tissue reservoirs. For now, we can see parts of this puzzle but not how they fit together. There's even sharp disagreement over whether replenishment of the latent reservoir occurs. While Siliciano says that in people on HAART the latently infected T-cells aren't dying or naturally going away, others say this pool declines. Leading AIDS researchers David Ho of the Aaron Diamond AIDS Research Center and Tae-Wook Chun, who works with Anthony Fauci at the National Institute of Allergy and Infectious Diseases, have argued that in people on HAART there is a very slow decrease of the amount of virus in lymphoid tissue compared with the decrease in blood. Now that long-term studies have been done, Fauci says, "I don't think the pool is declining to the degree we first estimated. The evidence suggests that it's a pretty stable pool."

"The major revision in my thinking recently is that I think there's a pattern of local and long-term transmission of propagation that is really critical to pathogenesis of this disease," states Haase. While HIV is a systemic infection, we need to focus more on local events, he suggests.

Here Siliciano agrees: "I think that you will have some CD4 T-cells that are infected everywhere. The greater the degree of tissue barrier, the more of a chance of perhaps a localized spread of infection and slightly different [viral] sequences here and there-that all makes sense."

Weighing in, Fauci admits he's less optimistic about eradication: "I think what you'll see in people who are on therapy is, if they go off the drugs, the vast majority are going to see the virus bounce back out from the latent reservoirs. Even if you don't detect it [the virus], it's somewhere." For now, he says, the "C" word is control.

New Treatment Approaches
Let's suppose Ashley Haase is right about the local model for establishing latent HIV infection. What might this mean in regard to treating these various reservoirs?

Many studies have highlighted the protective role played by immune cells called cytotoxic T-lymphocytes (CTLs) in controlling HIV in long-term survivors. The CTLs are part of the first arm of the immune response. Other cells called CD4 helper T-cells are also involved in mounting a vigorous systemic (all-body) immune defense to the virus. Recent studies of a new immune booster called Remune suggest that it may work as an adjunct therapy to boost HIV-specific responses in people on HAART. We also have some clues that a strengthened immune system helps contain viral escape in sanctuary sites. Could we generate local or mucosal CTL responses to control local infections in tissue reservoirs?

Siliciano has his doubts: "The problem with CTLs is that they don't spread around as easily as the virus does. So they can come and destroy a collection of infected cells, but if any of those cells have moved in the meantime, then you've got to mobilize those CTL cells to another site. Can you always get those cells localized quickly enough to the site of infection? That seems a tall order."

Maybe so, but others are more encouraged. Looking at the mucosa of the gut, an early site of infection, University of Pittsburgh researcher Micky Murphey-Corb recently found that she could induce CTL responses that remained localized in the gut mucosa of primates and protect them against challenge viruses. Using rectal inoculations, she's also seen early local, SIV-specific CTL responses stronger than those seen in the gut. "This finding suggests that the rectal mucosa may provide an optimum site for the induction of mucosal and systemic CTL," she concludes.

What about flushing the virus out of dormant cells, as another strategy? Here, early experiments are positive. IL-2 is an experimental immune booster that can stimulate resting T-cells to become active. At Chicago, Tae-Wook Chun reported the exciting news that when IL-2 was added to the HAART regimens of 14 people with traces of latent infection in their blood and lymph nodes, six of them had no trace of infectious virus after therapy. By comparison, all of the people receiving only HAART were latently infected with infectious virus. That doesn't rule out other compartments that may contain HIV, as Chun stresses, but it's encouraged many researchers nonetheless.

There's also the interesting case of hydroxyurea, an anticancer drug used in combination with HAART regimens that can work against resting T-cells; and at Chicago, the latest reports showed that an "unprecedented low rate" of HIV was found in latently infected people taking this drug. Nor was there any viral rebound in 12 people with chronic infection treated for 122 weeks with hydroxyurea and ddI. That's good news.

Slow Progress
Where does all this leave us? Siliciano's own view of where we stand today is sanguine and positive, despite the fact that eradication appears unlikely, barring new developments. As he likes to stress, the drugs we're using today have flaws, but they are getting better and keeping people alive. And while latent infection has thrown us for a loop, it's also revealed new treatment approaches. "The problem is that we don't know what the natural history is going to be for people who have been on effective therapy," he says. "It's virgin territory, whether some low-level replication will be the reason we see resistance, whether there will be some horrendous long-term toxicity, or whether people will just swallow a couple of pills a day and live to be 80 years old."

What says Ashley Haase? "The big thing is always to think, 'Where could we have the greatest impact?'" he states. That's why he's refocused his efforts on developing an HIV vaccine. "Are we on the right track?" he asks rhetorically. "Yes," he answers firmly. Is he encouraged by IL-2 and the prospect of immune control? Definitely. "It may be incremental to those afflicted-agonizingly slow-but it is progress," he says, picking his words carefully. "We didn't win the war on cancer in one quick battle, but we've certainly come a long way. I hope and expect to see similar advances in AIDS and HIV over the long haul."

In the meantime, keep your seat belts fastened.

illustrations by Kathleen Brady