This phase I clinical trial was designed to determine the feasibility of using rBCG as a live bacterial vaccine vector for the outer surface protein A (OspA) of Borrelia burgdorferi and as model for other vaccines based on a rBCG vector. To construct the vaccine, a signal peptide derived from a mycobacterial lipoprotein was used to direct the export, and membrane-associated surface expression, of OspA in a standard strain of BCG (Connaught). The rBCG OspA vaccine was safe and immunogenic in several animal species, and protective in a mouse model of Lyme borreliosis. An intradermal injection (0.1 ml) of rBCG OspA was administered to 24 healthy adult volunteers sequentially at one of four dose levels, ranging from 2.0 x 10(4) CFU to 2 x 10(7) CFU, using a dose-escalation design. All volunteers were initially PPD-skin test and OspA antibody negative, and they were monitored for 2 years after immunization. Three volunteers had mild flu-like reactions 1-2 days after vaccination. Local ulceration and drainage at the site of injection, which occurred in 50% and 83% of volunteers in the two highest dose groups, persisted for 1-70 days before the ulcers healed. Most of the drainage samples yielded rBCG colonies that contained the OspA plasmid. Thirteen of 24 vaccinees, principally in the two highest dose groups, converted their PPD skin tests from negative to positive. None of the 24 volunteers developed OspA antibody. In conclusion, the current rBCG vaccine construct, the first such construct tested in humans, had a safety profile comparable to that of licensed BCG, but it did not elicit primary humoral responses to the vectored antigen.