A quantitative model suggests immune memory involves the colocalization of B and Th cells.

A prominent and essential feature of the humoral immune response of vertebrates is immunologic memory: the ability to recall previous exposure to antigen. We present a mathematical model of the growth and interactions of the major cell populations involved in the humoral immune response. Our analysis of this model predicts that the formation of a dynamic association between small numbers of antigen-specific B and Th cells, "colocalization", is sufficient to account for memory and the kinetics of the secondary response--neither specifically differentiated Th or B memory cells nor networks of antigen and anti-idiotypes are required. The colocalization hypothesis explains a number of existing experimental observations and can be tested by straightforward experiments which we describe.