Alpha-particle exposures were used to determine whether cells of the human prostate carcinoma cell line DU-145 can produce and respond to a bystander effect signal. An apparatus for alpha-particle irradiation of cells growing as a monolayer on a 1.4-microm-thick Mylar membrane directly above an 241Am alpha-particle source was constructed and calibrated. At the cell irradiation position, the alpha-particle fluence was 998 counts/mm2 s(-1), the average alpha-particle energy was 3.14 MeV, and the average linear energy transfer was 128 keV/microm. The average dose rate to the cells growing on the Mylar surface was 1.2 Gy/min. A co-culture system was used to examine bystander effects transmitted through the medium from the directly targeted cells to tumor cells growing on an insert well beyond the range of the alpha particles. Alpha-particle doses from 0.1 to 6.0 Gy to the targeted cells on the Mylar membrane, followed by a 2-h co-incubation of the cells on the insert in the irradiated medium above the irradiated cells, all caused an approximately 50% increase in micronucleus formation in the nontargeted co-cultured cells. Addition of the radical scavenger DMSO to the medium during the irradiation and the 2-h postirradiation incubation period completely blocked the bystander effect, whereas addition of a nitric oxide scavenger had no effect. Irradiation of medium containing serum, followed by a 2-h incubation, caused no bystander effect in the co-cultured cells. When the co-cultured cells on the insert were placed into the irradiated medium above the directly targeted cells immediately (approximately 1 min) after the irradiation and co-incubated for 2 h, there was no bystander effect. These data indicate that the observed bystander effect requires that the co-cultured cells be present in the medium during the irradiation of the directly targeted cells and suggest the involvement of a short-lived radical species.