The interview

Shortly after the announcement of the 1996 Nobel prize winners, the Australian Academy of Science (AAS) spoke to Peter Doherty about his early life, the process of scientific research and the nature of the prize-wining research. The full text of this interview can be found at www.science.org.au/scientists/doherty.htm. This extract covers the prize-winning research.

AAS: So what exactly was the discovery that you and Rolf Zinkernagel made back in 1973–4 and how did you go about it?

PD: Well, of course, you don't set out to make a discovery. What you are doing is checking a hypothesis – strictly speaking, you try to falsify a hypothesis. In our case, there was a certain amount of serendipity involved. We started out looking at ways of assaying [measuring the activity of] killer T-cells. We wanted to know how T-cells recognise and respond to viruses in mice. Bob Blanden at the School was then working on the cytotoxic T-cell response to the ectromelia virus in mice. Rolf Zinkernagel worked with him, learning several techniques, including T-cell assays. As Bob's lab got a bit crowded, Rolf moved into my lab. Rolf loved to sing grand opera, and I was the only guy with sufficient musical taste to have him in the lab!

AAS: And then?

PD: Well, sure enough, we found killer T-cells in mice infected with a lymphocytic choriomeningitis virus (LCMV). (See Diagram 1.)

And, in the lab, we could get these T-cells to kill the virus-infected cells. But then we connected this work up with the system of transplantation antigens – something which hadn't really been done before.

AAS: In what way?

PD: Other people had suggested that there was some kind of relationship between the immune response genes, which map in the region of the major histocompatibility [transplantation] antigens and the susceptibility of mice to LCMV. Certain mouse strains, carrying particular histocompatibility antigens, were more susceptible to the virus, others less so. We did experiments to see whether the activity of the killer T-cells correlated in any way with the type of major histocompatibility antigens (MHC) in the mouse from which the T-cells came.

AAS: And did it?

PD: Well, yes, but not as we had supposed. It turned out, quite unexpectedly, that killer T-cells from one mouse were not active against virus-infected cells from another mouse – which of course, would have a different class of MHC antigens. In other words, the MHC antigens were certainly having an effect. In fact, the killer T-cells were not doing any killing of virus-infected cells unless the infected cells were showing the 'right' MHC antigens that the killer T-cells expected. (See Diagram 2.)

AAS: So the thing that decides whether or not a virus-infected cell is eliminated by these T-cells is not only the fact of having virus antigens on the outside, but also the possession of the 'correct' variant of the MHC antigens?

PD: Exactly. The MHC antigens have to be of the same sort as the individual that the T-cells come from. This means that what the killer T-cells are recognising is two things – viral antigens and MHC antigens. Viral antigen by itself is 'invisible' unless it is there with the MHC antigens; and it can't be with any old MHC antigens, it has to be with the MHC type of that individual (or inbred strain of mouse in the case of our work). (See Diagram 3.)

And incidentally all this work was only possible because the inbred mouse strains with different but consistent MHC types already existed.

The significance of the discovery

PD: I'll try....The fact that cytotoxic T-lymphocytes – the killer T-cells – cannot usually recognise foreign antigens unless these antigens are paired with MHC antigens is pretty fundamental.

AAS: That's what you called MHC restriction, isn't it? It's in every textbook now...

PD: It shows us that the immune system can recognise a third state – altered self – as well as self and non-self. When a virus has infected a cell and the cell is displaying viral antigens in addition to its own, it has become altered self. That's what's recognised and dealt with, rather than the viral antigens per se. The point is that the body treats altered self in much the same way as non-self. A virally modified cell is destroyed in the same way that a transplanted cell from another individual would be. (See Diagram 4.)

AAS: And what does that tell us?

PD: It gives us a biological role for the MHC system. People were wondering why the body should have a system for combating transplanted tissue when this state clearly never arises in nature. We suggested that the recognition of alloantigens – MHC antigens differing from your own – was there not to frustrate transplant surgeons but to help the body 'see' altered self.

AAS: But wouldn't it be easier just to see the viral antigen on an infected cell, rather than recognise the viral antigen in combination with the MHC antigens as an altered self?

PD: But altered self recognition allows the body to conduct surveillance on its own cells. A cell's antigens can be changed not just in virus infections but in certain cancers, for instance.

AAS: And such cells could be destroyed before they spread and threaten the whole organism?

PD: If all is working well, yes.

AAS: So you found a biological role for the MHC antigens....

In fact, Doherty and Zinkernagel had opened up a whole new highway that led to major advances of great significance to clinical medicine. Many common and severe diseases depend on the function of the cellular immune system and thus on the mechanisms for specific recognition.

Viral disease is only one part of the story; various long-term inflammatory illnesses such as rheumatism, rheumatoid arthritis, multiple sclerosis and even diabetes involve damage caused by the cellular immune system, probably as a result of altered self.

Certain forms of cancer, where the body's cells escape the controls on their multiplication, are also a form of altered self.

And we are now in a position to explain better the associations between tissue types (ie, HLA) and susceptibility to various diseases. Knowing a person's HLA-type it is possible to give a statistical likelihood of their developing certain diseases, based on the observed correlations between HLA-type and disease. Why particular tissue types are associated with a greater propensity towards certain conditions is under investigation.

Click pictures to enlarge
(will open in a new window)

Diagram 1

Diagram 2

Diagram 3

Diagram 4