A two-signal mechanism for the induction of cytotoxic T lymphocytes.

Depletion of adherent cells from stimulator and responder lymphocytes by a single filtration through nylon wool columns led to complete abrogation of the cytotoxic response to the stimulating alloantigen. Cytotoxic responses were restored by adding anti-Thy-1 + complement-greated normal peritoneal exudate cells (PEC) syngeneic or allogeneic to the responding population. Alternatively, the response could be reconstituted with costimulator, a lymphokine obtained by stimulation of spleen cells with concanavalin A. Costimulator was not itself cytotoxic and induced few or no cytotoxic T lymphocytes (CL) in the absence of stimulator cells. Costimulator was also more efficient than allogeneic PEC, which in turn were more efficient than syngeneic PEC, in reconstituting the cytotoxic response. The number of CL produced to the activating alloantigen was shown to increase with increasing concentration of costimulator. More interestingly, in the presence of a relatively high concentration of costimulator, CL were also activated to target cells that differ in H-2 haplotype from the stimulating alloantigen. Lysis of the third-party target cells could not be inhibited by cold targets syngeneic to the activating alloantigen. A clonal assay for cytotoxic precursors was used to confirm that CL for the activating alloantigen and CL for the third-party H-2 antigens were derived from different progenitors. Only about 37% of the cytotoxic clones produced were specific for the activating alloantigen. These observations are explained in terms of a two-signal model of CL activation.