The Education of T Cells

Science, April 13, 2007

Almost 3 decades ago, a team of immunologists made an intriguing observation. They collected white blood cells called lymphocytes from lymphatic fluid (lymph) that drained the skin or the gut of a healthy sheep, labeled those lymphocytes, and injected them back into the same sheep's bloodstream. To their surprise, the injected cells didn't patrol the whole body: Cells from the skin region returned mostly to the skin, whereas those from the intestine homed mostly back to the gut.

T cells, the infection-fighting immune cells born in the thymus, were thought to cruise the entire body via the bloodstream and the lymphatic circulation, stopping where they spotted signs of trouble. So how did those sheep T cells know to navigate to and patrol a particular tissue? The question matters because immunologists hope to battle tumors or autoimmune diseases by controlling the cellular immune response in one organ, while leaving the immune system alone elsewhere.

The first clues to an answer came from Eugene Butcher and Irving Weissman of the Stanford University School of Medicine in Palo Alto, California. In the 1980s, they showed that certain squads of T cells can distinguish between tiny blood vessels near the skin or near the intestine. Then Butcher's team and others identified dozens of cell-surface receptors and soluble signaling chemicals called chemokines that helped those T cells penetrate and patrol particular tissues. In the 1990s, Butcher and other biologists uncovered a molecular code--the unique combination of receptors and chemokines--that directed T cells to, say, the skin or the gut. But one crucial mystery remained: How does a newborn T cell, fresh from the thymus, become programmed, or educated, to express the combination of receptors that will let them home to a particular tissue? "It's a fundamentally important problem in cellular immunology," says Jeffrey Frelinger of the University of North Carolina, Chapel Hill.

Over the past 5 years, researchers have begun to crack that mystery. The most recent work, which shows how immune sentinels called dendritic cells instruct T cells where to go, is revealing a layer of intelligence in the body's immune surveillance mechanisms that had gone undetected, say Frelinger and other immunologists. Ultimately, physicians hope to use the emerging understanding of T cell targeting to develop immune-modulating compounds more specific than today's drugs, which for the most part are blunt instruments that can cause serious side effects. Drugs that direct T cells to specific sites could battle tumors, improve vaccines, or ease autoimmune diseases. "One can conceivably generate drugs that interfere with organ-specific [T cell] recruitment without paralyzing immune defenses everywhere else," says immunologist Ulrich von Andrian of Harvard Medical School (HMS) in Boston.

T cells on patrol

When tissue is infected by a foreign agent, its first line of defense is inflammation, the nonspecific response involving pain, redness, heat, and swelling. Then, over several days, the immune system activates squads of T cell clones, lines of cells each of which can latch onto a single bit of pathogen on an infected cell. T cells then neutralize the threat, call for backup from other immune cells, or both.

T cell activation begins when dendritic cells, octopuslike cells that roam the body's tissues, spot infection and chew up infected cells to obtain antigen--a small piece of a pathogen or tumor that can trigger an immune response. Dendritic cells then travel through the lymphatic ducts to the nearest lymph node, spongelike sacs that serve as regional field stations for the immune system. There the dendritic cells encounter many so-called naïve T cells but only activate for battle the ones bearing receptors that recognize the antigen they carry. The newly vigilant T cells multiply into an army of clones known as effector T cells that can fight infected or rogue cells.

The effector T cells then move from the lymph nodes through lymphatic vessels to the bloodstream, where they circulate throughout the entire body. But to fight pathogens, they need to find the site of the infection. Immunologists believe that some effector T cells stop in any tissue or organ where there are signs of trouble, or inflammation. But Butcher and others have long concentrated on the more specialized T cells that can home back from the bloodstream to a particular tissue, such as skin or gut.

By the early 2000s, Butcher and others had uncovered a clever addressing system that targets those tissue-specific T cells to the correct home. These T cells use a fourstep process to exit the bloodstream across the walls of tiny veins called high endothelial venules. Each of the four steps requires either matching pairs of Velcro-like receptors on T cells and the venule walls, or matching pairs of other T cell receptors and chemoattractants, small molecules that make up a tissue's unique chemical scent. If the four correct pairs of receptors and chemoattractants are present in the right combination, the T cell recognizes that it's in the correct tissue, then squeezes through the venule wall to the tissue beyond. Today, Butcher says, the field is starting to ask how a naïve T cell learns to express the correct combination of homing receptors for the gut, skin, or other tissues--a process called T cell education, or imprinting.

Before immunologists could find out how T cells undergo such imprinting, they had to make sure it really happened in living animals and that the cells were not born "precommitted to homing to gut or skin or joints," Butcher says. Butcher and Daniel Campbell, now at the University of Washington, Seattle, did that in 2002. They injected mice with millions of identical, fluorescently labeled mouse T cells, all of which had been genetically engineered to recognize an egg-white protein. They immunized the mice with that egg-white protein, then 2 days later, surgically removed lymph nodes and other lymphoid tissue from the gut and the skin. Inside all the lymphoid tissue they examined, the quiescent T cells were being activated into effector T cells that were ready to battle the foreign protein. But T cells found in the gut lymph nodes produced receptors that would help them find their way to the gut itself once they had reentered the bloodstream from the nodes, whereas otherwise identical T cells from the skin lymph nodes produced receptors that would direct them to skin, the researchers reported in the Journal of Experimental Medicine. "Where you get stimulated determines which homing receptors are expressed," Butcher explains.

What happens within a tissue's lymph node to program a T cell to migrate from the bloodstream to that tissue? Von Andrian suspected that dendritic cells teach T cells to home to the tissue where those foreign bits are found. That's because dendritic cells are on the scene in lymph nodes, embracing and helping activate the T cells.

Von Andrian's team purified dendritic cells from lymphoid tissue (lymph nodes or other specialized immune tissue) from three parts of the body: spleen (a central lymphoid organ), skin, and intestine. They incubated each tissue-specific type of dendritic cell in separate petri dishes with naïve T cells. After 5 days, T cells were ready to do battle with pathogens. But in a test-tube experiment, only T cells exposed to dendritic cells from the Peyer's patch, lymphoid tissue in the intestinal wall, migrated toward a gut chemokine.

Then, to see whether the same thing happened in living animals, the researchers injected mice with fluorescent T cells that had been stimulated by one of the three types of dendritic cells. T cells ended up mostly in the gut when they'd been activated by dendritic cells from gut lymphoid tissue, but not when they'd been activated by dendritic cells from skin lymph nodes, the researchers reported in 2003 in Nature. The same year, immunologist William Agace's team at Lund University in Sweden reported that dendritic cells from mesenteric lymph nodes, another immune site in the gut, also educate T cells they touch to home in on the intestines. Together, the results mean that "antigen-presenting cells from different lymphoid tissues are not equal in terms of the story they're telling," von Andrian says.

Since then, immunologists have worked out some of the chapters of that story. In a pivotal 2004 paper in Immunity, Makoto Iwata of the Mitsubishi Kagaku Institute of Life Sciences in Tokyo discovered that vitamin A (retinol), which is abundant in the intestine but scarce in other tissues, plays a key instructional role in T cell homing. In test-tube experiments, they found that dendritic cells from the intestinal lymph nodes convert retinol to retinoic acid, which induces T cells to make gut-homing receptors but not skin-homing receptors. Subsequent animal experiments confirmed the importance of this conversion to T cell homing: Mice starved for vitamin A had far fewer intestinal T cells than mice that consumed enough of the vitamin.

Recently, Butcher and research scientists Hekla Sigmundsdottir and Junliang Pan and their colleagues probed for a comparable molecular mechanism in the skin. "We wondered if a similar vitamin or metabolite that might be restricted to the skin might imprint skin homing," Butcher says. Vitamin D, which is mass-produced by skin cells in response to sunlight, "was the obvious candidate," he adds.

Butcher's team isolated lymphatic fluid from the skin of sheep, purified dendritic cells from that fluid, and found that the immune cells convert vitamin D3, the sun-induced variant of vitamin D, into its active form. In other test-tube experiments, this activated vitamin D3 induced T cells to make a receptor that helps them follow their nose to a chemoattractant in the epidermis, the skin's outer layer, the team reported in the February issue of Nature Immunology. An evolutionarily related chemoattractant in the gut lures T cells using a different receptor to that tissue, Butcher points out. These studies indicate that dendritic cells can exploit a tissue's unique biochemical fingerprint--its unique mix of metabolites--to educate T cells to patrol that tissue, Butcher says.

T cells specialized for one tissue can also be retrained to patrol another area, von Andrian, HMS immunologist Rodrigo Mora, and their colleagues reported in 2005 in the Journal of Experimental Medicine. They cocultured T cells for 5 days with dendritic cells from the gut, spleen, or skin, which imprinted T cells for those tissues. They then washed each group of T cells and cultured them with dendritic cells from a different tissue. After 5 more days with their new instructors, "the T cell phenotype would always match the flavor of the dendritic cells they had seen last," von Andrian says. That ability to reassign T cells to new tissues may give the immune system an important degree of flexibility in fighting infection. If the pathogen stays put, the immune response is concentrated in that tissue, von Andrian says. "But if the pathogen spreads, you have not put all your eggs in one basket."

Immunologists have begun investigating whether the T cell's instructors--the dendritic cells--themselves specialize to function in a particular tissue, or whether they simply sense their environment and respond. A definitive answer is not yet in. Butcher's team found data suggesting that dendritic cells have two vitamin D-activating enzymes no matter what tissue they're from, but only in the skin do they have access to the sunlight-produced vitamin. Agace's team, in contrast, has found evidence that at least some dendritic cells are more specialized. In a 2005 study in the Journal of Experimental Medicine, his Swedish team reported evidence of two types of gut dendritic cells: one that has visited the intestinal wall and can train T cells to migrate to the gut, and another, of unknown origins, that can't.

Steering cells right

The new work on tissue homing is raising immunologists' hopes of specifically boosting or suppressing immunity in selected tissues. Most autoimmune diseases involve an overactive, self-destructive immune response toward a particular tissue: the pancreas in type 1 diabetes, the central nervous system in multiple sclerosis (MS), the joint in rheumatoid arthritis. Typically, treatments for such diseases dampen the entire immune system and increase the risk of infection. Similarly, stimulating the immune system nonspecifically to fight a tissue-specific tumor can increase the risk for autoimmune side effects.

That's where the new knowledge of T cell homing can help, Butcher says. Drugs that alter homing are not themselves new; in 1997, Butcher and HMS biochemist Timothy Springer co-founded a biotech company called LeukoSite, which was later bought by Millennium Pharmaceuticals, to develop drugs that block the Velcro-like interactions and molecular sniffing that help T cells find their way into tissues. Many drug and biotech companies are still pursuing that approach, which has produced a U.S. Food and Drug Administration-approved drug for MS and drugs for ulcerative colitis and Crohn's disease that are currently in clinical trials. But blocking a single receptor often fails to prevent T cell entry into tissues because the receptors involved in homing can often fill in for one another.

Drugs that alter T cell imprinting "might be a way around the problem of redundancy," Butcher says. Both gut-homing and skin-homing T cells interpret their respective signals, retinoic acid and activated vitamin D, using members of a large family of receptors that sense hormones and metabolites and directly control gene expression. Drugs that stimulate or alter these nuclear-hormone receptors already exist, and some are being tested for autoimmune diseases such as rheumatoid arthritis or psoriasis. That gives researchers a head start, as those drugs might alter the instructions that tell T cells where to migrate, explains Butcher. "The exciting thing about imprinting is that we're just learning about its potential," he says.

The recent advances in T cell imprinting also create several possible new ways to fight disease, Agace says. Most pathogens enter the body through the surface, or mucosa, of a particular tissue, which means that a drug that directs T cells to the mucosa could enhance the cellular immune response, making vaccines more effective in warding off intruders. Other compounds could help battle localized tumors. For example, coinjecting lab-grown dendritic cells, which are already used as an antitumor therapy, with compounds modeled on retinoic acid could potentially program T cells to migrate to a gut tumor and boost the treatment's effectiveness, Agace says.

Retraining T cells could backfire by working too well, caution some immunologists. In a recent clinical trial, the MS drug Tysabri stopped abnormal T cell homing to the brain and eased MS symptoms. But it also suppressed the brain's immune surveillance system so much that a normally benign virus began reproducing in three patients, ultimately killing them.

What's more, T cells may not take instruction in all tissues, says pulmonary physician Jeffrey Curtis of the University of Michigan, Ann Arbor. Immunologists still debate whether specific squads of T cells are assigned to patrol tissues other than the skin and gut. Researchers have been unable to find a combination of adhesion molecules or chemoattractants that lures specific T cells into lungs, he notes. But physiologist Klaus Ley of the University of Virginia, Charlottesville, who studies T cell migration in lung and blood vessel disease, disagrees: "If I project into the future, we will see more homing specificity--for gut and lung and I hope for [atherosclerotic] blood vessels."

The research on T cell homing has also now begun to merge with another hot topic in immunology: regulatory T cells, a much-touted cell type that naturally suppresses autoimmune reactions. Several years ago, Alf Hamann of Charité University of Medicine in Berlin and his colleagues reported that regulatory T cells isolated from different tissues have homing receptors like those that effector T cells sport. Now, in a March online paper in the European Journal of Immunology, they report that these cells, like effector T cells, can be programmed by dendritic cells, an interleukin, and retinoic acid to home to skin or gut. In theory, sub-populations of regulatory cells could therefore be prepared to target a tissue and suppress an autoimmune response. "If you could make a regulatory T cell in vitro and make it go where you want it to go, that's a cool thing," Butcher says.