The efficacy of immunosuppressive therapy in aplastic anemia (AA) provides the strongest argument to support its immune-mediated pathophysiology. While several immunosuppressive effects of antithymocyte globulin (ATG) can be demonstrated in vitro, some reports have implied that the activity of ATG in AA may be rather due to a variety of positive hematopoietic effects. We studied the effects of horse (h) and rabbit (r) ATG on marrow progenitors in vitro. Both types of ATG bound to CD34 cells and, in colony assays performed with total marrow cells, hATG had a dose-dependent, triphasic effect, with maximal increase in colony formation between 1 and 10 microg/ml and inhibition between 100 and 1000 microg/ml. As determined using CD34 cells, these effects did not require accessory cells. rATG showed similar activity, but was about 10-fold more potent than hATG. In the presence of complement, no increased cytotoxicity was observed. At concentrations equivalent to those measured in patients immediately after infusion, ATG showed moderate suppression of colony formation, while the stimulatory concentrations in vitro correspond to those seen in vivo within the first weeks after ATG administration. In control experiments, the patterns of the biologic effects of preimmune rIgG or hIgG preparations were similar to those of rATG and hATG, indicating a nonspecific nature of the effects of ATG on progenitor cells. Biological activity in methylcellulose cultures was observed with the F(ab)(2) fragments but was not found in purified Fc IgG. In summary, the spectrum of effects of ATG on hematopoietic progenitors is dependent upon the concentrations of ATG and may not be related to its antigenic specificity.