Both retro- and adenovirus-mediated gene therapy have been suggested as a novel approach to the treatment of malignant brain tumors. However, little information is available about the gene transfer efficiency in human malignant glioma in vivo. We compared the feasibility and safety of retrovirus- and adenovirus-mediated beta-galactosidase gene transfer in human malignant glioma. Beta-galactosidase gene was transferred to 10 patients with malignant glioma via a catheter inserted into the tumor. The catheter was left in place until the tumor resection. To maximize gene transfer efficiency, gene transfer vectors (BAG retroviruses, titer, 6 x 10(5) CFU; and adenoviruses, titer from 3 x 10(8) to 3 x 10(10) PFU) were injected into the tumor via the catheter once a day for three consecutive days, followed by tumor resection 1-2 days later. Tumor was resected in such a way that the catheter was still in place inside the tumor, which permitted accurate histological analysis of the transduced tumors. X-Gal staining for beta-galactosidase activity was used to study gene transfer efficiency and distribution of the marker gene. Beta-galactosidase gene transfer was well tolerated with both vectors. Except for two patients with clear increases in serum adenovirus antibody titers, no adverse tissue responses or systemic complications were noticed in any of the patients. Gene transfer was successful in all patients. Gene transfer efficiency varied between <0.01 and 4% with retroviruses and between <0.01 and 11% with adenoviruses. However, the transgene activity was not evenly distributed in the tumors. Both glioma cells and endothelium in the tumor blood vessels were transduced with retro- and adenovirus vectors. In conclusion, the safety and feasibility of in vivo gene transfer to human malignant glioma was established with retro- and adenovirus vectors. Adenoviruses were more efficient than retroviruses in achieving in vivo gene transfer. Transduction of endothelial cells may have important consequences for the proposed treatment strategies and selection of treatment genes. The results justify clinical gene therapy trials for malignant glioma.