Balancing Act

Regulatory T cells suppress immune responses and researchers are now working to determine precisely how, but their role in HIV pathogenesis is still unclear

By Andreas von Bubnoff

What goes up must come down. That's not only true for everyday life but also for complex biological systems like the immune system: Where there is activation, there must also be suppression. Such homeostasis is a tenet of biology, but finding the mechanism that prevents the immune system from spiraling out of control was not easy. After following many blind alleys, it would take until 1995 to identify T cells that suppress the immune system.

Today, it is clear that these regulatory T cells—or Tregs—are indeed crucial to keeping the immune system in check, although the exact mechanism as to how they suppress their target cells has not yet been resolved. Tregs are important in suppressing autoimmune disease in that they suppress T cells directed against self antigens. They also appear to play an important role in HIV pathology, although it's an open question whether they make things better or worse. Already, some researchers are thinking about modifying Tregs to help boost the response to AIDS vaccine candidates, but potential side effects such as autoimmune disease could make this a difficult balancing act.

Road to discovery

Immunologists thought they had discovered T cells that suppress the immune response over 30 years ago. In the early 1970s, Richard Gershon of Yale University coined the term "suppressor T cell" after experiments suggested that a certain type of T cell suppressed immune responses through a soluble factor and not through direct interaction. In the following years, many prominent groups tried to characterize these suppressing factors. With 1300 journal citations during 1981, "this was a tremendously active field," says Ethan Shevach of the US National Institutes of Health (NIH).

But the genes for the postulated suppressing factors couldn't be found, and the field shrank dramatically. In 1987 there were only 150 papers published on suppressor T cells. "Nobody could find how they were working," says Claire Chougnet of the Children's Hospital Medical Center in Cincinnati, Ohio. "So they became this thing that nobody wanted to touch. Some people started to doubt that they were real."

But not everyone was on the wrong track. In 1969 a Japanese group reported that removing the thymus during a certain period of mouse development resulted in autoimmune disease (Science 166, 753, 1969). This could be prevented, however, by transferring a normal thymus back into the mice. It appeared, then, that the mouse thymus produces T cells that suppress the action of T cells recognizing self-antigens, and thereby prevent autoimmune disease. For the most part, though, immunologists ignored these early studies. "They were more interested in these soluble suppressor factors that the boys at Harvard or Yale were working on," Shevach says.

The renaissance of the field came in 1995, when Shimon Sakaguchi, now at Kyoto University, showed that removing a subset of CD4+ T cells that express a marker called CD25 resulted in severe autoimmune disease in mice (J. Immunol. 155, 1151, 1995). The condition resembled the one described in the 1969 study. This time, Sakaguchi called the cells regulatory T cells. Suppressor T cells had become the "s-word," Shevach says, "based on 10-15 years of terrible studies which are basically regarded as fundamental artifacts."

Still, the earlier name probably describes their function better: These Tregs suppress immune responses, including the activation of CD4+ and CD8+ T cells and also dendritic cells. They are important because, for example, not all T cells against self-antigens are deleted in the thymus, and Tregs are thought to suppress the activity of those that escape. They can also suppress T-cell activation in response to pathogens.

Defining Tregs

The 1995 study made it possible for the first time to identify Tregs at a molecular level. Then, in 2001, several groups identified a similar population of Tregs in cultured human cells. One problem was that CD25 is expressed not only by Tregs but also by CD4+ and CD8+ T cells once they are activated. But David Hafler's group at Harvard Medical School showed that Tregs could be better defined as the ones that express the highest levels of CD25 (J. Immunol. 167, 1245, 2001). Still, what's high is in the eye of the beholder. "How you define high CD25 is highly subjective," Chougnet says.

Then several papers added one more marker to the molecular definition of Tregs: The transcription factor Foxp3. A mutant mouse strain called scurfy with a severe autoimmune condition proved to have a mutation in Foxp3(Nature Genetics 27, 68, 2001). Similarly, humans with an autoimmune condition called IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X-linked) syndrome also had mutations in Foxp3 (Nature Genetics 27, 20, 2001). What's more, Sakaguchi's group showed that Foxp3 could convert naive T cells into Tregs (Science 299, 1057, 2003). "[Foxp3] really established the Tregs as something real," says Derya Unutmaz of New York University, "because now you have a molecular program that you can turn on and show that you develop suppression or autoimmunity." But like CD25, human Foxp3 expression is not specific to Tregs: Recently-activated CD4+ T cells express it as well. "It's not clear-cut in humans as to how specific [Foxp3] is," Chougnet says.

Figure 3. Generation and Function of Regulatory T Cells.


View larger image

Checks and balances

In the past few years researchers have also started to look into the role of Tregs in infectious diseases. One early study found that inhibiting Tregs leads to clearance of infection with the parasite Leishmania major in mice (Nature420, 502, 2002). Intuitively that's a good thing, says Lishan Su of the University of North Carolina at Chapel Hill. "But the bad thing is that you also lose immunity against secondary infection with Leishmania," Su says. This implicates Tregs in keeping immunity to secondary infection intact by preventing complete clearance. "It's a perfect balance," he says.

In HIV infection it's still unclear if Tregs are beneficial or detrimental. The area is still only a few years and about two dozen papers old, Unutmaz says. The first studies suggested that Tregs suppress the immune response to HIV. In 2004 several groups removed CD25+ Tregs with magnetic beads from blood-derived, ex vivo cells of HIV-infected individuals and showed that depletion of Tregs increased the T-cell response to HIV antigens in vitro. "That suggested that Tregs play a role by decreasing the response to HIV," Chougnet says, indicating perhaps that Tregs are a "bad" thing because they suppress cellular immune responses to HIV.

But other studies have suggested that Tregs could be a "good" thing because they also suppress chronic immune activation, which often correlates with progression of HIV to AIDS. For example, SIV-infected African Sooty Mangabeys don't develop disease or chronic activation of the immune system. And a recent study showed that they also maintain more Tregs than SIV-infected rhesus macaques, which do develop disease and show chronic immune activation (J. Virol. 81, 4445, 2007).

It's possible that Tregs are "bad" early in infection but "good" later on, Unutmaz says. Early in infection, or when giving a vaccine, a potent HIV-specific immune response is a good thing, so suppressive Tregs are "bad". Later on, Tregs are "good" as they suppress the chronic activation of the immune response. "They are a double-edged sword," Unutmaz says. "You don't want to mess with it too much, it could work both ways."

Getting the story straight

In HIV infection, Tregs appear to behave differently in different sites in the body. They disappear from the blood, some studies have found, but accumulate in the lymphoid tissues where most of the HIV infection occurs.

But just how Tregs disappear from the blood after HIV infection is unclear. One possibility is that HIV infects and kills the circulating Tregs, and Unutmaz has found that this can happen with cultured human Tregs. If HIV does deplete Tregs in vivo, this could contribute to the chronic activation of the immune system that's observed in HIV-infected individuals. "The more immune activation, the quicker they will develop AIDS and disease," Unutmaz says.

Intuitively, that doesn't fit with Chougnet's observation that there are many Tregs and also many HIV particles in the lymphoid tissues. The Tregs may migrate from the blood to the tissues where most HIV replication occurs, and HIV may actually promote accumulation and perhaps even survival of Tregs, which then further suppress the immune response. "Tregs may be one way that HIV uses to limit the capacity of the immune system," Chougnet says.

In a field this young, it's perhaps not all that unexpected that there are contradictory hypotheses. "The field has turned into quite a mess," Unutmaz says. "It is a bit like the story of blind scientists trying to figure out the elephant by touching different body parts."

Manipulating Tregs

Since it's unclear whether Tregs in HIV-infected patients are a good or a bad thing, Chougnet says, "it's difficult to predict how we can manipulate them in a clinical setting." Genoveffa Franchini's lab at the National Cancer Institute, with Chougnet, Israel Lowy of Medarex and others, has been trying to inhibit Tregs in monkeys. They used an antibody to block CTLA-4 (cytotoxic T lymphocyte antigen 4), a receptor that, among other things, inhibits activation and proliferation of T cells. Franchini says the antibody modestly boosted the immune response in SIV-infected macaques that were treated with antiretroviral therapy after they were infected with SIV (Blood 108, 3834, 2006). Franchini's lab has also combined the CTLA-4 antibody with a therapeutic SIV vaccine in macaques, but did not observe an increase in T-cell response.

Medarex has also used the CTLA-4 antibody in a Phase I clinical trial in HIV-infected patients and shown that it was safe and well tolerated, says Lowy, the lead physician for that trial. But Lowy and Shevach say there is little evidence that the CTLA-4 antibody actually affects Tregs and not effector T cells, which also express CTLA-4.

With experimental tumor vaccines, studies in humans have shown that a drug called ONTAK, which binds CD25 and kills Tregs, can enhance the response to the vaccines. And CTLA-4 antibodies can lead to remission but also autoimmune side effects in some patients. However, Shevach says, the precise specificity of ONTAK is unclear, so they may not only affect Tregs but other T cells that also express CD25.

For now, it's still up in the air what the implications of the continuing research on Tregs will be for AIDS vaccines. Su cautions that preventive vaccination together with an anti-Treg treatment could exacerbate the chronic immune activation in HIV-infected people. "We have to be very careful with modulating the immune system," Su says. "You may accelerate the disease."

Hindsight is 20/20

But when it comes to the way Tregs were discovered, one thing is clear: "When one looks back, one can find what's a right and what's a wrong experiment 20 years later," Shevach says. "Some [of the experiments] people paid little attention to in the end proved to be correct."

In a paper describing the CD25 antibody in 1983, Shevach had also found that even in uninfected mice, 8% of the normal T cells expressed CD25—a similar fraction to the Tregs described by Sakaguchi. But at the time, he was not thinking about Tregs, he says. CD25 was thought to be only expressed by activated T cells. So he thought he was looking at regular T cells that expressed CD25 because they had become activated in the animal facility which was quite dirty. "[I] didn't pay any attention to that," he says.

So how did Sakaguchi have the idea to look at CD25 as a marker? "I asked him that," Shevach says, laughing. "He wouldn't tell me."