Dehydroepiandrosterone (DHEA) and its sulfate conjugate are the major circulating steroids in human plasma. Low levels of these adrenal steroids are associated with a number of human diseases including certain cancers. In animal studies, DHEA is chemopreventive toward both spontaneous and chemically induced cancers. A potential concern for long-term usage of DHEA in humans is the finding that DHEA is hepatocarcinogenic in rats. The human health risk has been thought to be minimal, however, as the mechanism of DHEA hepatocarcinogenesis is assumed to be due to its properties as a peroxisome proliferator, a class of compounds to which humans are relatively insensitive. Recently, we have found DHEA to be a potent promoter of aflatoxin B1-initiation as well as a complete hepatocarcinogen in the rainbow trout, a species which is also insensitive to peroxisome proliferators. In order to determine the initiator- and tissue-specificity of DHEA promotion, we examined the effects of DHEA on N-methyl-N'-nitro-nitrosoguanidine (MNNG)-initiated carcinogenesis. Trout fry were initiated by a bath exposure (30 min at 35 ppm) to MNNG and then fed DHEA at levels of 0, 55, 111, 222, 444, or 888 ppm for 7 months. DHEA increased liver tumor incidence, multiplicity, and size in a dose-dependent manner. The liver tumor incidence ranged from 0 in the MNNG-initiated controls to 99% in initiated trout fed 888 ppm DHEA. The latter represents a potential synergistic interaction in liver between MNNG and DHEA, as tumor incidence in sham-initiated trout fed this level of DHEA was 41%. The kidney tumor incidence was also enhanced two- and threefold over initiated controls by 111 and 888 ppm DHEA, respectively. In contrast, the total number of stomach and swim bladder tumors was reduced by DHEA treatment. This study demonstrates differential effects of DHEA on MNNG-initiated carcinogenesis in liver, kidney, stomach, and swim bladder.