In a fully MHC plus multiple minor antigen-mismatched murine bone marrow transplantation (BMT) model, we have demonstrated that a short course of high dose IL-2, begun on the day of BMT, protects against graft-versus-host disease (GVHD). This inhibitory effect is directed against donor CD4+ cells. To determine whether the mechanism of IL-2-induced GVHD protection involves clonal deletion or anergy of host-reactive donor T helper cells (Th), we performed limiting dilution analyses to measure the frequency of activated Th that reacted to donor, host, and thirdparty antigens in GVHD control and IL-2-protected mice. Marked and specific expansion of host-reactive Th was observed to a similar extent in GVHD control and IL-2-protected mice by day 5 after BMT, and the number of these cells in the spleen increased by several orders of magnitude between days 3 and 5 after BMT, which suggests that recirculation from other tissues occurred in this period. A high proportion (approximately 80%) of donor T cells expressed CD25 in both GVHD control and IL-2-protected mice on day 4 after BMT, which suggests a high level of bystander T cell activation. Since marked quantitative differences in the GVH response were not observed between GVHD control and IL-2-protected mice, we assessed both groups for qualitative differences in the Th response. Spleen cells isolated in the first 8 days after BMT were cultured with host-type, donor-type, or third-party stimulators or without stimulators, and cytokines were measured in supernatants harvested at 24 hr. GVHD was associated with marked increases in supernatant IFN-levels from day 3 to day 6 after BMT, and with increases in IL-2 levels compared with naive A/J controls or syngeneic BMT controls stimulated with host antigens. Production of these cytokines was specifically induced by host-type antigens. Supernatants from spleens of IL-2-treated mice showed delayed kinetics of IFN-y production, and tended to contain higher levels of IL-4 in response to host antigen compared with GVHD controls on days 2 and 4 after BMT. Both IL-4 and IFN-were produced almost exclusively by CD4+ cells in spleens of GVHD control and IL-2-protected mice on day 4. However, no consistent difference was observed between the groups in supernatant IL-2 or IL-10 levels, ruling out a simple Th1 to Th2 switch. Neutralizing antibody to IL-4 did not inhibit the protective effect of IL-2 against GVHD, and previous studies have indicated that changes in IFN-kinetics do not explain IL-2-induced GVHD protection. Thus, in vivo high dose exogenous IL-2 treatment does not inhibit the marked Th expansion and T cell activation that occurs in the first week of GVHD, but significantly perturbs the pattern of GVHD-associated CD4 cytokine production.