Dubbed "the person least likely to succeed in life" by her schoolmates, Dr Matzinger is now challenging core beliefs about how our immune system works.
Since the Second World War and the work of Peter Medawar, who subsequently won the Nobel Prize, received wisdom has been that the immune system attacks anything that it recognises as foreign - the so-called "self/non-self" divide. The flaw is that there are increasing numbers of problems with this rule - for example, when does an embryo or a newborn learn what is self and what is not? Dr Matzinger's theory - the "Danger Model" - proposes that the immune system only attacks when it receives signals of danger or damage to cells. If she is right, her model has major implications for the way we look at transplant surgery, immunology and cancer treatment.
Her work has been published in the most respected science journals, such as Nature and Science, but she has proved controversial ever since her scientific career began - with more conventional scientists finding this kooky and direct woman difficult to come to terms with: her first published paper gave her dog as co-author, a fact that annoyed the journal publishing it considerably.
The chances of the young Polly Matzinger astonishing the scientific community seemed dim. Not regarded as university material, she left school to study music but dropped out because "the music I compose is dead boring and I realised 'I'm going to end up teaching high school choirs somewhere'."
Ten years of drifting followed, during which she worked as a jazz musician, trained problem dogs, ironed shirts, cleaned bricks and worked as a cocktail waitress: "Then I worked as a pool bunny - a Playboy bunny playing billiards. It paid 40 cents an hour and I got to listen to a lot of great conversations. It was a great job."
She decided to remain a cocktail waitress so that she could make a lot of money in the evening and do what she wanted during the day. "And then I got picked up in a bar," she says. "But not in the normal way that one gets picked up in a bar. I was serving these two people who used to come in a lot. These two professors would come in and talk science and I would listen.
"One day they were talking about mimicry - how one butterfly will mimic another butterfly and how a good tasting butterfly will mimic a bad-tasting butterfly. So I asked them a question that I'd wondered about for years and years and never known the answer to - why has no animal ever mimicked a skunk? A raccoon with a white stripe down its back would have a selective advantage. Their mouths fell open - a cocktail waitress asking this sort of question, and they didn't know how to answer it."
One of the scientists was so impressed that he started a nine-month campaign to persuade Polly Matzinger to take up science, bringing her articles which they would argue over in all-night diners after she had finished her shift, until "he finally convinced me that science might be a job that doesn't get boring".
At university Dr Matzinger specialised in immunology and contributed to several advances in the field. Then she moved to work at the prestigious National Institutes of Health. She became increasingly convinced, as did some of her colleagues, that the "self/non-self" model was not adequate to explain how the immune system works. A system that attacked everything foreign would lead to the system attacking the food we eat; a mother's body would reject the foetus it carried.
Dr Matzinger and her colleague Ephraim Fuchs, assistant professor of oncology at Johns Hopkins School of Medicine, became convinced that a really effective immune system would only attack things it perceived to be dangerous. But discovering how the system defines danger proved more difficult. It took months to work it out.
"There are a lot of qualities it takes to be a scientist and one of them is that you have to be a masochist," says Dr Matzinger. "You have to be willing to live a life that is 90 per cent depression and 10 per cent elation - and have the elation keep you going."
Her elation came in an Archimedean moment: "I was in the bath and I had had by that time months of trying to work out what was dangerous, and then all of a sudden - there it was! It was good there was nobody in the house. I jumped out of the bath - I was totally naked, dripping water all over the house, going from one room to another unable to sit still."
Her realisation? Cells die in two ways: programmed cell death, where the cell shrivels up, never releasing its contents, and a cell being killed, when its contents are let out, sending a signal there is damage. "Things that are dangerous are things that cause damage. If you don't cause damage you're not dangerous. It was so simple and it had taken so long."
The final piece in the puzzle came when she was sitting on a hill watching her sheepdog (she is a sheepdog expert as well). "I suddenly realised that there was a cell in the body which behaves like a good sheepdog - the dendritic cell. The dendritic cell would be activated by a cell dying in its midst and kickstart the immune response. And that puts the model together".
If correct, her theory has important implications for the treatment of cancers and transplants. At present those who undergo transplant surgery are forced to take huge doses of immuno-suppressant drugs for the rest of their lives to stop the body rejecting the new organ. Dr Matzinger's theory is that rejection is caused by the trauma of surgery which sets off the dendritic cells. But by blocking these danger signals at the time of the transplant with a short course of drugs, a patient could live drug- free for the rest of their life.
At the Emory University School of Medicine in Atlanta, scientists seem to be proving her idea right. They have successfully transplanted a rat's heart into a mouse, giving it a brief course of immuno-suppressant drugs at the time of the transplant and no more. The mouse is thriving.
In the field of cancer, Professor Gus Dalgleish at St George's Hospital NHS Trust is putting Matzinger's theory into practice on one of his patients with advanced skin cancer, applying the Danger Model in an attempt to alert the immune system to the damage caused by the cancer. Injecting the patient with a bacterium that expresses heat shock protein - a chemical signal to other cells which is produced when they are dead or dying - alerts other cells and the immune system to the intruder tumour.
Despite favourable reviews when her research was published in Science a year ago, Matzinger still has her critics who have refused to take her model seriously. Dr Matzinger is not cowed: "It is said the scientist who is willing to stick his neck out and be clear will contribute to the field whether he or she is wrong or not, because if they are wrong someone will do the experiments to prove they're wrong and in the process will learn something about nature. So whether I'm right or wrong doesn't matter.
What matters is that it is clear and I'm going out and saying it"n
Polly Matzinger's story is told in 'Horizon - Turned On By Danger' tonight on BBC2 at 9.30pm