Feature: Three-day event

T-cell behaviour in type 1 diabetes

Type 1 diabetes, which affects one in 500 children, arises when the body's own immune cells attack insulin-producing beta cells in the pancreas. If a way could be found to stop this immunological 'friendly fire', those affected would be free of the need to inject insulin to control blood sugar levels. In Cambridge, Allison Green believes just such a strategy is possible. The reason? A particular type of cell that has revolutionised immunology.

It is one of immunology's grand challenges: to work out why an immune system that is supposed to defend us from dangerous pathogens and cancer cells can also turn against us.

Dr Allison Green, a Wellcome Trust Senior Research Fellow at the Cambridge Institute of Medical Research, crossed the Atlantic in 1995 to join the search for the mechanisms of cellular destruction that end in type 1 diabetes. She joined a leading immunology lab at Yale University to zero in on the cellular triggers of this autoimmune disease, and to work out the immune mechanisms controlling the assault. But Dr Green had not bargained for what she found.

For three years, as is so often the case in science, she met with little but frustration. Her findings were puzzling, even disappointing. "The results were not very good. The animals always got sick. So I finally lost interest in this system," she admits. As it turned out, though, those years were far from wasted.

Magical mice

The mice Dr Green was studying were predestined to become diabetic. She had genetically altered them to churn out tumour necrosis factor (TNF), a naturally occurring compound, in islet cells. TNF derives its name from its ability to kill cancer cells, but in excess it also leads to inflammation and autoimmune disease.

The mice were doomed because the TNF gene, unlike the normal one, was constantly active. Crucially, however, Dr Green’s system was constructed so that TNF production could be turned on and off at will. As long as the mice’s food was spiked with the antibiotic doxycycline, the gene remained silent. But as soon as Dr Green withheld the doxycycline, the gene was switched back on.

This system enabled her to see whether the timing of TNF activity was important. At first, she gleaned nothing helpful from these mice. If she kept TNF switched on from birth, the mice soon developed diabetes. If TNF was turned on when the mice were adults, they then also became ill.

Eventually, serendipity lent an obliging hand. By chance, Dr Green took a look at the insulin-producing islets before the destruction set in, and was surprised to find them teeming with lymphocytes. As she pointed out to Richard Flavell, Chairman of Immunobiology at Yale: "Something really weird is going on with these animals."

Timing is everything

It was only when she collected the data from three years of experiments that she realised something fascinating was happening in these mice. The pancreatic islets were being invaded in a tightly choreographed fashion.

If TNF was on from birth, at 21 days there were only a few infiltrating CD4 T lymphocytes (see box, below). By day 25, the numbers of CD4 cells had increased dramatically, but there were still no signs of aggression. Finally, by day 28, the assault by destructive CD8 T lymphocytes had begun.

"We found exactly when the regulation stage was happening," says Dr Green. Between day 21 and day 25, the CD4 regulatory T cells accumulate, trying to keep a tight grip on the potentially destructive CD8 cells. Then, between days 25 and 28, TNF prevents the regulatory cells expanding; as the CD8 cells are unleashed from regulation, the mouse is pushed towards type 1 diabetes.

What makes Dr Green's system so powerful is that TNF production can be switched on and off at different times in the same animal model – there are no confounding factors. "It hit us [that] we have a beautiful model for looking at regulation, tolerance and breakdown in tolerance," she says.

The key period for regulation was clearly between day 21 and day 25. But where exactly did the regulatory cells reside? Dr Green discovered that they congregated exclusively in the pancreatic islets and lymph nodes. Had she looked in the usual places – the spleen or lymph nodes farther around the body – she would have missed them. "If you remove all CD4 cells [at this point], you lose regulation of islet-specific CD8 cells in these animals." And once regulation fails, diabetes is inevitable.

The ultimate test for whether these regulatory cells played a protective role was to transfer them into mice that were already getting sick. Could they delay disease? "To prove how good these regulatory T cells are, we had to do the transfer experiment," Dr Green explains. But this was a tough call, because there are few cells in the pancreatic lymph nodes, and they are hard to isolate.

In desperation, she took as many regulatory cells as she could gather – just 2000. "People would have just said: 'Don't be ridiculous'," she admits. But to her amazement, it worked. These cells were so potent that even in small numbers they stopped mice from becoming diabetic.

Expanding numbers

Nevertheless, in therapeutic terms, delaying or halting diabetes will probably require a hefty dose of regulatory T cells. But what are the signals that prompt regulatory T cells to multiply? A good candidate is a protein called transforming growth factor b (TGF-b). TGF-b binds to immune cells and tells them to stop attacking. In the inflamed tissue, TGF-b levels rocket during the protective, regulatory phase. But as soon as the mice slip into the aggressive phase, levels plummet.

Restoring TGF-b may therefore be a good way to boost the numbers of regulatory cells. But as well as a regulatory signal being present, the autoreactive CD8 cells must be responsive to it. For example, without a molecule called CD40 ligand on their surface, CD8 autoaggressive cells pay no heed to regulatory cells, and the disease accelerates.

Dr Green ultimately envisages combination therapy as a two-pronged approach to diabetes. "First, we need to know how to expand regulatory cells," she suggests. "Second, we need to know what signals will keep an autoaggressive cell sensitised to regulatory cues."

Excitement is running high as these potent regulatory cells live up to their reputation in clinical trials. At the Necker Hospital in Paris, Lucienne Chatenoud has followed newly diagnosed diabetic patients given short-term antibody treatments to activate regulatory cells, and found that it is possible to stop disease progression.1 And in the UK, Professor David Wraith from the University of Bristol will soon be starting clinical trials of a peptide vaccine for patients with early-onset multiple sclerosis. The hope is that this strategy will induce regulatory T cells and dampen the disease.

But Dr Green cautions: "I have a strong belief that unless we do investigations into what particular signalling molecules are important in different states – inflammation, responses to outside antigens or host antigens – there is a potential for danger."

Gaining that additional insight is critical, and, now back in Cambridge, Dr Green is scrutinising those three days when regulation is at its peak. By finding out which genes are activated, and which proteins change, it may be possible to design a specific therapy. "That's the beauty of the small animal models. You can go straight into the islets to find out what is going on: the type of cells, cytokines, and get a full picture of what is crucial for developing regulatory T cells."

And once the best strategy for manipulating regulatory cells is confirmed, it may offer therapies not only for type 1 diabetes, but for other autoimmune diseases too.

Judging a book by its cover

Many types of immune cell look the same but can be distinguished on the basis of marker proteins found on their surfaces. Unique markers (or combinations of markers) can therefore be used to identify cells with particular roles in the immune system.

Lymphocytes are typically characterised according to so-called ‘CD’ antigens (rather unhelpfully, CD stands for ‘cluster of differentiation’). Over many years, numerous antigens were discovered on the surfaces of lymphocytes, and because no one had any idea what the proteins did, they were just given numbers: CD1, CD2, etc.

As well as distinguishing different types of cell, CD antigens (and other surface proteins) also allow types of cell to be separated, so they can be studied on their own.

It soon became apparent that some CD antigens on T lymphocytes were diagnostic of the cells’ functions. So T lymphocytes with CD4 (but not CD8) were ‘helper’ cells, boosting immune responses, while cells with CD8 (but not CD4) were ‘killer’ cells, able to destroy cells they recognised.

Different forms of CD4 cell are now known to exist, with slightly different roles in the immune response. And one of the most exciting types of CD4 cell is the regulatory T cell.

Back with a vengeance

Regulatory T cells have had a chequered past but promise a bright future.

They are the flavour of the moment in immunological circles. Manipulate them astutely, say the scientists, and these special types of T cell could finally crack autoimmune disease, and ensure transplantation tolerance with few or no drugs. ‘Regulatory’ or ‘suppressor’ cells are now firmly in the vanguard of research, but, bizarrely, only a few years ago, they were the pariahs of the immunological world – no one dared talk about them.

Why spurn such remarkable cells? For more than 30 years, researchers suspected that certain cells could snuff out an immune response, but made little progress because they could not tell for sure which of the myriad cells in the immune system performed this function. They could not identify the proteins on the cells’ surfaces that were diagnostic of suppressor function (see box, above).

The experiments were also hard to reproduce, so although numerous labs reported ‘suppression’ – the dampening down of the immune system – many immunologists remained sceptical, and suppressor cells were dismissed as artifacts.

Until 1996, that is, when Shimon Sakaguchi from Kyoto University, Japan, showed that the cell the immunologists had been looking for was a particular type of T cell. Crucially, he also located the marker on the surface, a molecule known as CD25.

After 30 years of ignominy, the study of suppressor cells has undergone a renaissance. As promising results from animal experiments pour out of leading research labs, the way is open to treat autoimmune disorders such as diabetes by artificially boosting the number of regulatory T cells, or to fight tumours and infections by turning them off.

The vital question is whether successes in the laboratory will translate into the clinic. Clinical trials are, so far, yielding promising results. The hope is that once the details are clinched, regulatory cells will offer therapeutic solutions where none existed before.

References