THE NEW BODY OF EVIDENCE

Rather Like Archimedes, Polly Matzinger has done some of her best work in the bath. It was while she was having a soak about three years ago that she suddenly had her own Eureka moment: she saw an entirely new way to understand the immune system.

Like all the best ideas, it is elegant and simple, yet explains a number of phenomena that we can't presently explain with the existing ideas - rather like the proposal that the earth orbited the sun, rather than vice versa, put forward in the Middle Ages by the astronomer Copernicus.

And, rather as Copernicus's heliocentric view took the earth out of the middle of the picture and relegated it to an observer's position in favour of the sun, Matzinger's view removes our "self" from the centre of the action, in favour of a disinterested, yet vigilant, body system.

Maybe for that reason, her suggestions have upset many immunologists, who see her new way of thinking as threatening 50 years' work. Most people reckon they have a fairly good idea of what the immune system does and how it works. When viruses and bacteria invade, it co-ordinates the reaction which fights them off and eventually kills them - just like the white blood cells trying to devour Raquel Welch's swimsuit in the Sixties film Fantastic Voyage. The reddening of the skin around an infected wound or the high temperature of somebody fighting 'flu are both indicative of the immune system at work. We can be immunised against diseases by being given tiny doses of them, because that shows the white blood cells what that invader will look like if it ever tries to attack.

But how do they know how to do that? Most people - and immunologists, too - would say that it's straightforward. The constituents of the immune system can recognise what is "ourself" and what is "foreign". Our bodies attack what they recognise as foreign. This is the "self/ not-self" model to which most mainstream immunologists adhere.

But Matzinger disagrees. She says that the answer is simpler, and that the components of our immune system listen out for body cells suffering "damage, distress or (bad) death". (The "bad" qualification is explained further on.) That then triggers the responses with which we are all so familiar.

It may not seem like such a big shift, but the elegance of the idea is that it actually requires no distinction between self and not-self, simply a constant vigilance for bad news. But, at the same time, it is a notable conceptual shift in which the person disappears. The "ego" of an immune system which knows who it is part of disappears, replaced instead by the "id" of an indifferent monitor.

Matzinger backs up her idea with an example. "Your 'self' changes all the time. The immune system can't discriminate for that. Think of puberty: your body undergoes huge changes. But you don't get an immune reaction." But to immunologists, who have based years of work, if not entire careers, around the "self/ not-self" paradigm, Matzinger seems nothing short of a loose cannon on the decks of science.

Meeting this happy woman, who looks like a cross between the actresses Sally Field and Helen Mirren, one would not think so. She gives every sign of being at ease with herself and her work. At the prestigious American Association for the Advancement of Science (AAAS) conference this February, which brings together scientists from across the US, the audience for her talk on this subject included her dog, a bright-eyed collie. Afterwards, while talking to journalists, she kept the dog happy by throwing a rubber frisbee around the room for it to retrieve (which, happily, it did).

Also like Copernicus, Matzinger has got us- ed to not having her ideas readily accepted. Last June she was visited by "two old fogey reviewers", as she calls them, at her offices at the National Institute of Allergic and Infectious Diseases. Concluding that she was carrying out "naive experiments", they earmarked her for a 15 per cent budget cut, denied her the funding to pay for the extra person needed to carry out the next stage of experimentation, and denied her promotion.

Was she disappointed? Of course. "But if you're swimming upstream," she says, "then you're going to get hit by dead wood from time to time." There's a hard edge to the smile that accompanies the remark.

In fact, in March the pivotal journal Science published an important piece of work by Matzinger and some colleagues which backed up her theory. They had repeated the experiments which first led to the "self/non-self" paradigm, which were carried out by Peter Medewar in the Fifties, and subsequently earned him the Nobel Prize. Medewar had injected newborn female mice with tissues from newborn male mice. The male tissues were not rejected. Medewar concluded that this was because the newborn immune system was still learning what was "self", and so had no reason to mount a reaction to the intruder cells.

Matzinger repeated that work - perhaps the cause of the reviewers' "naive experiments" jibe - and got the same results. But then she took another set of newborn mice and, before injecting the tissues, added a dose of sentry-type cells to the male cells being transplanted. The sentry cells, known as dendritic cells, signal danger to cells known as T-lymphocytes, which are key soldiers in the immune system. This time, the newborn females immediately rejected the transplants. No learning was required: the immune system was fully functional at birth.

The "self/not-self" theory struggles to explain this, although Arthur Silverstein, a critic of Matzinger based at the Institute of the History of Medicine, Johns Hopkins University, contends that "there is nothing mystical" about it. "The mouse has long been known to slowly expand its immunological repertoire during the neonatal period," he responded when the work was published.

It's true that the existing theory copes with the results; but it has to be tied into complex immunological knots to do so. Matzinger contends that the "danger" theory copes with it easily. And it helps to explain a host of other observations, including some which presently tie scientists in knots, just as Copernicus's re-siting of the earth helped to explain the peculiar movements of the planets in the sky that the geocentric model of the universe could not.

For instance, says Matzinger, with the "self/ not-self" model, why doesn't a mother reject the foetus in the womb? After all, the baby is not the mother's "self": the genes in its cells are different from its mother's, which means the proteins on its cell surface are, too. They are linked by the umbilical cord which should promote a huge immune response. Conceptually, a foetus is the same as an implanted, donated organ.

"The reason is, babies don't look dangerous to the mother's immune system," says Mat-zinger. "The foetus grows, and has a lot of cell death, true - but it's programmed cell death."

Such cell death, also known as "apoptosis", is the "good" way for body cells to die, she argues. It happens because there are internal mechanisms in cells which check that replication is carrying on safely. If they detect that the DNA is damaged, for example, or that the cell function has gone wrong, that starts a process by which the cell shuts itself down and quietly dies. No other cells are involved. "Programmed cell death is silent," says Matzinger. And that, she argues, means that the immune system ignores it.

By contrast, in "bad" cell death - such as that caused by a virus which invades the cell and takes over its DNA to start making endless copies of itself - the result, when the virus has finished its job of usurping the cellular factory, is that the cell walls break apart as the cell is killed. Those signals of bad cell death are what first alerts the immune system, she contends, and not the recognition of the virus as foreign.

"It's interesting," she adds, "that if the foetus gets infected, then you often get spontaneous abortion." The explanation is that the mother's immune system picks up the untimely cell death in the womb, and treats the whole foetus as potentially dangerous. The abortion is an immune reaction.

Matzinger offers more questions, such as the evolutionary one. Which is: given that evolution generally picks the most efficient design for a system, why have something that generates an immune response to absolutely everything the body encounters - including the air we breathe and food we eat? Surely it's simpler only to react if something bad happens.

For instance, why should the body worry if a virus invades, takes over a cell to make a few copies of itself, and then departs without causing damage? Such viruses might have useful genes in them, she argues; there would be little point in fighting them off. Evolution offers witnesses for her case: every genome contains historical fragments of "retroviruses" - viruses which incorporate themselves into the cell DNA. And the mitochondria, which are the power stations of normal cells, are thought to be evolved from bacteria which our cells captured billions of years ago for their usefulness.

Matzinger credits her shift in thinking to Ephraim Fuchs, a graduate student who join- ed her at the National Institutes of Health. Questions such as the non-rejection of the foetus, and the supposed enormous activity of the immune system to every bite of food, "were questions that I had asked when I was a graduate student and then forgotten about. I was reminded of them by Ephraim.

"It took him two years of arguing to re- convince me that the immune response is not governed by the foreignness or selfness of the antigen." Fuchs's reasoning stemmed from a complex series of experiments carried out on elements of the immune system, known as T and B cell lymphocytes. The conclusion: it didn't seem really to matter whether the antigen is foreign or self. It can have similar effects - either making the immune system tolerant of a "foreign" protein (and letting it provoke no response) or immunising it, in which case it will watch for and attack the antigen.

This helps to explain why cancers do not cause immune reactions in their early stages, since there is no cell death of any sort going on in cancerous cells, just division out of control. Yet the cell proteins are different, because of the mutation causing the cancer. Isn't that therefore non-self? By the time that the cancer is large enough to damage normal cells, and trigger an immune response, the disease has often advanced a long way.

Matzinger's idea may give clues to understanding diseases that are presently categorised as "autoimmune" - the body attacking itself. Multiple sclerosis, she suspects, could be caused by a viral invasion that has not yet been identified. If that sounds absurd, then it's worth remembering that millions of pounds were spent investigating stomach ulcers - but the Australian scientist who, in 1983, suggested that it could be caused by a bacterium was laughed at. He proved it by making a drink laced with the bacterium (called Helicobacter pylori) and drinking it. He rapidly developed gastritis - a symptom related to stomach ulcers. But it was not until 1994 that the National Institutes of Health agreed that "Helicobacter pylori plays a significant role in the development of ulcers". Now, researchers are positing bacterial causes for all sorts of illnesses, including heart disease.

The big problem with Matzinger's idea, though, is that it is very hard to prove wrong. And that, to scientists, makes it almost a useless theory; rather like the person who suggests that all our actions are controlled by aliens who always cover their tracks. "I wish there was someone out there trying to prove me wrong," she says. The trouble is that you can't isolate immune cells to do the test; as soon as you cut into the body, you're creating that initial damage signal. Yet her idea does explain so much. It might also offer new methods to treat cancer precisely (by injecting immune- system triggers that would attach to the cancer cells). And it would give the whole mode of thinking a degree of elegance that it really lacks.

It could take sometime. In a response to one critic of her Science paper, Matzinger wrote, "Roughly two millennia before Copernicus, Aristarchus proposed a heliocentric model to explain the motion of the planets." But that was ignored for almost 20 centuries because people felt happy with the old paradigm, of an earth-centred universe. "By the time Copernicus suggested his version, the motion of planets could no longer be easily explained by the view that the earth was central in the solar system, and the intellectual community was ready for a 'new' idea," she added.

Well, maybe it's time for a new idea. Certainly, one gets the feeling that Polly Matzinger would be happy if she doesn't have to wait 20 centuries to be proved correct. !