by Dr. Geoffrey W. Hoffmann, PhD
In January 1974 I had the pleasure of hearing a lecture about the immune system by Niels Jerne at Manfred Eigen’s Winter Seminar in the Swiss Alps. The members of the audience were mostly not immunologists, and Jerne told the story of network interactions in the immune system in a way that made it clear that here was an exciting problem for a theoretical biologist to work on. After the talk I said to Peter Richter, a theoretical physicist in Eigen’s group “It’s going to be a race, isn’t it?” Peter grinned.
We returned to Göttingen and Peter and I both did some work on the problem. Peter was able to tackle it analytically, while I tried out various ideas using a computer. The problem was to find a set of postulates such that the system could have multiple stable states for any of many antigens, and to determine how the system could switch between different stable states in response to antigenic stimulation. The simplest form of switching was low dose tolerance, which is a paradox in the context of clonal selection without taking account of network interactions, so showing how it could be explained using network interactions was a first challenge. My initial efforts did not take me far. No matter how I tweaked the postulates of my models by trial and error, there seemed to be only error, with the system stubbornly exhibiting oscillations, rather than multiple stable states.
Peter on the other hand did not initially go near the computer. He formulated an idea in his head for how not only low dose tolerance might occur, but also high dose tolerance. He determined that for his idea to work, he needed to include not only stimulatory and suppressive interactions as postulated by Jerne, but also inhibitory interactions. This was a significant modification of Jerne’s thinking, who had emphasized “dualisms”. With these three types of interactions, Peter did the sums and was able to find values of the parameters that indeed resulted in three types of switching between stable steady states – low dose tolerance, the immune response, and high dose tolerance. It was inspiring to see that this could all be done analytically, without the aid of a computer.
Then one day an advertisement for positions at the Basel Institute for Immunology, where Jerne was the Director, appeared on the notice board of Max Planck Institute for Biophysical Chemistry, and I needed no second bidding. This was the way to learn immunology. I applied, and was accepted on the basis of a theoretical paper I had written on the origin of life problem.
When I arrived in Basel I spent a couple sessions with Niels, in which he would describe a phenomenon, and then say “It doesn’t make any sense, does it?” One of them was the fact that an antiidiotypic antibody could stimulate T cells, seemingly taking the place of the carrier in the stimulation of the system by hapten-carrier conjugates. Another was a report of antigen-specific T cell factors. These had a molecular weight of about 50,000, and from my exposure to Peter Richter’s work it immediately occurred to me that these could be responsible for inhibition. I knew very little about haptens and carriers, but the temptation was to lump them together as the antigen, and to assume that antiidiotypic antibodies could be stimulatory. Jerne was skeptical, since he had emphasized the distinction between the recognizing part of an antibody (“paratope”) and the recognized parts (“idiotopes”). But I made a shape-averaging argument, in which the average shape of antibodies that bind to an antigen has a shape that is complementary to the antigen, and the average shape of the antiidiotype is similar to that of the antigen. This is now of course familiar and even obvious, but at the time Jerne looked at me quizzically, and said “Maybe you will solve immunology before you leave here, yes?” Maybe you will solve…? That was enough to set the ego spinning out of control.
While in Göttingen I had already worked on a model that incorporated symmetrical killing, and now it looked as though stimulation could also be symmetrical. With the possibility that specific T cell factors are inhibitory it was a small step to assume that inhibitory interactions are also symmetrical. I now had enough ideas to be able to go back to the computer, and see what would work. In Richter’s theory low dose tolerance occurred when the level of antiidiotypic cells surpasses that of antigen-specific cells and the antigen-specific clones are suppressed. It was natural to look for something similar. But the computer told me something else. With symmetrical stimulation, low doses of antigen took the system to elevated levels of both antigen-specific and antiidiotypic cells, such that the induced suppressed state was symmetrical. I immersed myself in the literature, and soon became convinced that everything was falling into place nicely. The theory provided a way for understanding the helper and suppressor roles of T cells in their interactions with B cells. I wrote it up and submitted it for publication in the same journal where Richter had published his.
When the manuscript was ready, I asked Niels what he thought of it, and he said “I don’t believe in those little fellows”, meaning specific T cell factors. After that there did not seem to be much point in spending time talking to him about the theory, and we went our separate ways. He evidently did not think I had solved immunology, and I was not invited to participate in the small meetings of scientists working on immune network science that he organized. This exclusion was strange. I was working full time on the theory that was dear to his heart, in his own Institute. But the Institute nevertheless provided me with an excellent environment for continuing to develop the theory until 1979, when I moved to the University of British Columbia. In retrospect, I could and should have been more of a promoter for the theory, including to Jerne; at the time I felt that the logic of the theory spoke for itself, and if the generic you did not like it, that was your choice.
I have put together a list of publications on the symmetrical immune network theory from 1975 to 2012. This list documents in detail the development of the theory since 1975.
 N. K. Jerne (1974) Towards a Network Theory of the Immune System. Ann. Immunol. (Inst. Pasteur), 125C, 373-389.
 P. H. Richter (1975) A network theory of the immune response. Eur. J. Immunol., 5, 350-354.
 G. W. Hoffmann (1975) “A network theory of the immune system.” Eur. J. Immunol., 5, 638-647.
 The Symmetrical Immune Network Theory – Publications