Extended Data Fig. 1: The frequency issue in T cell–B cell interactions for GC positive selection and the model of chemoattraction-driven remote sensing of affinity information.

a–c, Conceptualization of the frequency dilemma. Red arrow, a hypothetical unit of T cell help delivered during a cognate T cell–B cell contact; blue shade, a hypothetical range of potential that does not ensure survival; red shade, a hypothetical range of potential that ensure positive selection. a, Receipt of T cell help is necessary to ensure survival (black curve), as GC cells die in its absence (blue curve). b, Frequent T cell contacts raise the positive selection potential (black curve), which decays when T cell contacts and help signals are not as frequent (grey curve). c, Sufficiently frequent T cell contacts help B cells to cross the threshold for positive selection (red curve), leading to cyclic re-entry, memory B cell formation or plasma cell formation as possible outcomes (red arrows). d, The working model of CCL22-mediated dynamic highlighting and remote sensing of B cell affinity information. Cognate T cell–B cell contact and help-signal strength are ultimately determined by the B cell affinity; help signals received during brief cognate contacts with T cells (green) drive CCL22 (cyan) upregulation in the B cell (red); higher-affinity cells produce more CCL22, attracting more T_FH cells from a distance; more attempts to contact by T_FH cells are translated into more contacts and better help signals experienced by a given B cell, which in turn can produce more CCL22 to attract more help; without experiencing sufficient T cell help, either because of low affinities or by chance, B cells also rapidly lose CCL22 production (not depicted).