Previous results from this laboratory demonstrated that treatment of mice with the adenosine analog tubercidin (Tub) reduced natural killer (NK) cell activity while stimulating antibody production whereas the deoxyadenosine analog, 2-fluoroadenine arabinoside-5'-monophosphate (FaraAMP), produced opposite effects; i.e., it stimulated NK cell activity at doses that inhibited antibody formation (Cancer Res. 48, 4799, 1988). Since NK cells have been reported to play a suppressor role in immunoglobulin induction, it was hypothesized that the actions of Tub and FaraAMP on antibody production occurred secondary to their opposing effects on NK cells. To test this hypothesis, abilities of these nucleoside analogs to modulate primary antibody response to sheep red blood cells were evaluated in a C57BL/6 mutant mouse lacking NK cell activity (the beige mutation. C57BL/6-bg/bg). As previously found with C3H/He mice. NK cell activity was inhibited (Tub, doses 2-6 mg/kg/day for 3 days) or stimulated (FaraAMP, doses 75-250 mg/kg/day for 3 days) in heterozygous mice C57BL/6-bg/+. In support of the hypothesis, these nucleosides had no effect on primary antibody formation in the homozygous mutant mice at doses that clearly stimulated (Tub) or inhibited (FaraAMP) this immune response in heterozygous C57BL/6-bg/+ animals. This results was corroborated in C57BL/6 wild-type mice by abrogation of NK cell activity using a monoclonal antibody to the NK cell surface glycophisingolipid, ganglio-n-tetraosylceramide. We conclude that under the conditions of drug administration, modulation of primary antibody formation by Tub and FaraAMP in mice occurs indirectly via NK cells. Similar experiments using the potent ADA inhibitor, deoxycoformycin, indicated that its enhancement of primary antibody formation is independent of NK cell activity.