Strong educating impact leads to more efficient NK cells within the responsive pool. (A) Degranulation and IFN-γ production in NK cells from naive MHC^−/−, D^b, K^bD^b, and K^bD^bD^d mice in response to anti-NK1.1 stimulation (stim.). The large gate defines frequencies of total CD107a⁺ and the small gate CD107a⁺ and IFN-γ⁺ double-positive NK cells. One representative experiment of 7. (B) Summary of all 7 experiments performed, as in panel A, with the percentage of CD107a⁺ IFN-γ⁺ NK cells expressed as the mean with SEM. (C) Mean GeoMFI for IFN-γ⁺ with SEM for all 7 experiments performed. (D) GeoMFI of IFN-γ in Ly49A⁺C⁻I⁻Ly49G2⁻ IL-15/18–activated NK cells from MHC^−/−, D^b, and D^d mice after stimulation with antibodies against Ly49D, NK1.1, and NKG2D. Numbers indicate the percentage of IFN-γ⁺ cells and GeoMFI for IFN-γ. One representative experiment of 3. (E) Similar experiment as in panel D. Anti-Ly49D stimulation of poly IC–activated NK cells, numbers indicate the percentage of IFN-γ⁺ cells and GeoMFI for IFN-γ⁺. (F) The percentage of IFN-γ⁺ (top graph) and GeoMFI (bottom graph) in Ly49A⁺C⁻I⁻Ly49G2⁻ IL-15/18–activated NK cells from D^b and D^d mice after stimulation with either YAC-1 or its β2m-deficient variant AH-2⁻. YAC-1 and A.H-2⁻ both display very low levels of MHC class I at the cell surface and, as they gave identical results, all experiments were combined in the same figures. In total, 7 experiments were performed. Data are expressed as mean relative values compared with MHC^−/− NK cells from the same experiment. *P < .05; **P < .01; ***P < .001.