The aim of this study was to evaluate the possible carcinogenic potential of residual DNA derived from immortalized and possibly tumorigenic cell lines due to activated oncogenic sequences (oncogenes). These cell lines have been used for the production of biologicals, i.e. monoclonal antibodies, lymphokines and vaccines. The authors used hybridoma DNA as a first model. For this reason experiments in two species were performed, namely in 3-4 week-old female Balb/c mice and newborn Riv:TOX rats. Doses of 250 micrograms DNA, derived from Balb/c hybridoma cells, were injected subcutaneously (s.c.) in 200 mice. These mice also received a s.c. injection of the solvent only (TE buffer) at another site of the back skin (negative control for local tumour development). An additional group of 50 mice was treated intraperitoneally (i.p.) with the solvent only to serve as a negative control group for possible systemic tumorigenic effects. Doses of 5 micrograms plasmid pPy1 DNA, containing the entire Polyoma virus genome, served as positive control and were injected s.c. and i.p. in 20 and 50 mice, respectively. Doses of 50 micrograms hybridoma DNA or 5 micrograms pPy1 DNA were injected s.c. in rats too, using nine animals per group. During the experiment, animals were observed weekly, especially for the occurrence of subcutaneous tumours at the injection sites. The mouse study was terminated after more than 2 years, the rat study after 1 year. Gross necropsy was performed on all animals and histopathological examination of grossly suspected neoplastic lesions was performed. In the mouse experiment, tumour development at the s.c. injection site of the DNA was observed in one out of 20 animals in the pPy1-treated positive control group (neurofibrosarcoma) and one out of 200 animals in the hybridoma DNA-treated group (haemangioma-like lesion). Tumour development at or near the s.c. injection site of the solvent only was observed in two out of 200 animals. In the rat study none out of nine hybridoma DNA-treated rats developed tumours at the injection site, while three out of nine rats of the positive control group, injected with the pPy1 DNA, showed local tumour development (benign and malignant soft tissue tumours.) It is concluded that, at the high dose and numbers of animals tested, parenteral administration of hybridoma DNA does not induce local tumour development. Furthermore, no indications were found for systemic carcinogenic potential of the hybridoma DNA used. Based on a worst case approach of our data, the oncogenic risk of 100 pg residual DNA was estimated to be 2 x 10(-9), a value intermediate of the estimations of the WHO (1987) and the Dutch Health Council (1988) 5 x 10(-11) and 2 x 10(-7), respectively. Therefore, it is unlikely that the risk of 100 pg of DNA derived from other immortalized cell lines will exceed the level of generally accepted cancer risk of 10(-6).