In animals, Ad-mediated gene transfer initiates anti-Ad host immune responses that vary, depending on vector design, dose, host, and transgene. To begin to understand whether the anti-Ad vector responses in humans simulate those in animals, Ad(GV)CD.10, an E1-E3- Ad5 vector encoding the E. coli cytosine deaminase gene, was administered by the intradermal route to six normal individuals (8 x 10(7) to 8 x 10(9) particle units, each dose administered to two sites; n = 2 per group). No adverse events were observed. Polymerase chain reaction/Southern analysis demonstrated vector genome in the skin through 28 days in all individuals except one of two at the lowest dose. Local induration, independent of vector dose and baseline systemic anti-Ad5 neutralizing antibodies, developed in all subjects (6 to 17 mm, peak by day 3). Biopsies revealed a mild to moderate T cell (CD3+, CD4+, CD8+), B cell, and macrophage infiltrate at day 3, all decreased by day 28. Langerhans cells accumulated primarily in the papillary dermis. The day 3 cellular response was dose independent. On day 28, CD4+ and CD8+ T lymphocytes and macrophages showed dose dependency. There was minimal systemic Ad5-specific lymphocyte proliferation induced by Ad vector administration in three individuals studied, and no Ad5-specific cytotoxic T lymphocytes (evaluated in two subjects) could be detected. Thus, intradermal administration of an E1-E3- Ad vector to normal subjects induces mild/moderate local cellular responses, even in Ad-immunized individuals. These observations provide a baseline to determine if these human anti-Ad vector host responses can be circumvented by using "stealth" vectors and/or immunosuppression.