American Scientist

Tending Adam's Garden: Evolving the Cognitive Immune Self. Irun R. Cohen. 288 pp. Academic Press, 2000. $49.95 (hardcover), $24.95 (paper).

There is much wisdom in this book. It makes the immune system, which serves to protect our body from attack, very accessible to those willing to cut through the Aristotelian verbiage that makes up much of the early parts of the book. Unfortunately, most people, amateur or professional scientists alike, are likely to be scared off before they get to the crux of Cohen's argument. Not until page 209 (?152) does he discuss his central concept, the immunological homunculus.

Those readers who have studied neuroscience or medicine will be familiar with the neurological homunculus. It is a funny-looking little being with large lips and a large thumb, published by the Canadian neurosurgeon Wilder Penfield and based on sensations reported to him by conscious patients when they were stimulated along the parietal fissure (experiments that would be difficult to get by an institutional review committee today!). Cohen contends that the immunological homunculus exists through co-respondence, a term he made up to describe cellular interactions that mediate selection of the immune repertoire and its action in producing the adaptive immune response, which involves signals from phagocytic cells, T lymphocytes and B lymphocytes.

Although I am a believer in the immunological homunculus, I was disappointed by Cohen's disquisition on two fronts. Not only did I find the introductory chapters very heavy going, but to read Cohen one would think that idiotypic networks had been proved through immunization with T cells and B cells, and that was that. Nothing could be further from the truth. What such experiments do demonstrate is that all of the elements of an idiotypic network exist, as should have been obvious to immunologists (and was obvious to many of them). What has not been demonstrated is what such immune networks are for. Cohen accepts them as proven, but they are far from that.

Finally, I find Cohen's discussions of phagocytes particularly objectionable. He assumes that they respond to a whole host of microbial signals, and I agree. But he is not straightforward about the most important consequence of this interaction, which is that it stimulates the migration to the cell surface of two classes of molecules: foreign peptides—bound to self-MHC (major histocompatibility complex) molecules—and other costimulatory molecules. Both are needed to produce a response from naive T cells; only when both are expressed on the same cell can the immune system be activated and differentiate to perform its effector function.

On a different topic, it now appears that regulatory T cells, which can control autoimmune responses, recognize antigen just as all other T cells do, as fragments of foreign (or self) proteins bound to conventional MHC molecules. Regulatory T cells specific for, say, insulin can prevent type I diabetes, at least in mice, and in human beings as well, I hope. Surprisingly, at least to me, autoimmunity appears to be a loss of regulatory T cells, as my former boss Richard K. Gershon always maintained, rather than, or in addition to, the activation of T cells that kill the insulin-secreting pancreatic β cells. Therein lie the lessons of modern immunology, and they are perfectly able to accommodate immunological homunculi.

To me, Cohen spends far too much time on Aristotle and other irrelevancies such as "basins of attraction," and far too little on the central message, which is that the adaptive immune system is referential to self. However, others may love the book's flavor and logic. Putting my reservations aside, I can recommend Tending Adam's Garden highly to laypeople and scientists alike, whether they are knowledgeable about immunology or not, as one way into the immune system.—Charles A. Janeway, Jr., Immunobiology, Yale University School of Medicine