In keeping with the in vitro mitogenic properties of anti-CD3 MoAbs, the first injections of anti-CD3 are invariably responsible for an in vivo cellular activation. This activation induces a massive cytokine release in the circulation (TNF, IFN gamma, IL-2, IL-6, and IL-3). Paralleling this release, a severe clinical reaction occurs in OKT3-treated patients and in 145 2C11-treated mice. Corticosteroids both in vitro and in vivo inhibit the production of several cytokines involved in the anti-CD3 reaction. A single 1 mg hydrocortisone dose was administered to 145 2C11-treated mice according to different kinetics schedules. When given 1 hr prior to the anti-CD3 MoAb, hydrocortisone exerted a beneficial effect on the mouse physical reaction. Hypothermia was totally abrogated at the 4-hr time point. Diarrhea decreased by 50%. Hypomotility improved although not significantly. This improvement correlated with a major modification in the anti-CD3 pattern of cytokine release. At the 90-min blood withdrawal time point cytokine serum levels showed a 100% decrease for IFN gamma, an 88% decrease for IL-6, and 85% decrease for IL-2, and a 75% decrease for TNF. At 4 hr IL-2 serum levels were diminished by 65%; IL-6, IL-3, and IFN gamma serum levels were comparable to controls; and, interestingly, TNF was still detected, whereas it has already disappeared when 145 2C11 was administered alone. Importantly, when given more than 1 hr prior to anti-CD3 injection, corticosteroids were ineffective. To conclude, high doses of corticosteroids must be given with a precise kinetics--i.e. 1 hr prior to anti-CD3 MoAb--to achieve their maximal beneficial effect in the prevention of the anti-CD3 reaction.