CD4 memory has a hierarchical structure created by requirements for infection-derived signals at an effector checkpoint.

Our recent studies reveal that the persistence, location, and amount of both antigen and signals that induce pathogen recognition responses determine the number of CD4 memory cells, the subsets that develop, their location, and hence their protective efficacy. Non-replicating vaccines provide antigen that is short-lived and generate low levels of only some memory subsets that are mostly restricted to secondary lymphoid tissue. In contrast, exposure to long-lived replicating viruses and bacteria provides high levels of diverse antigens in sites of infection and induces strong pathogen recognition signals for extended periods of time, resulting in much higher levels of memory cells of diverse subsets in both lymphoid and nonlymphoid sites. These include memory subsets with highly potent functions such as T follicular helpers and cytotoxic CD4 effectors at sites of infection, where they can most effectively combat the pathogen early after re-infection. These effectors also do not develop without antigen and pathogen recognition signals at the effector stage, and both subsets must receive these signals in the tissue sites where they will become resident. We postulate that this leads to a hierarchical structure of memory, with the strongest memory induced only by replicating pathogens. This paradigm suggests a likely roadmap for markedly improving vaccine design.