Retroviral-mediated gene transfer into hematopoietic precursors often results in only short-term gene transduction in vivo. Loss of the transduced genetic material over time may be caused by the limited ability of retroviral infection to transduce genes into early, pluripotent hematopoietic stem cells. Because fetal liver contains actively proliferating multipotential stem cells that should be more susceptible to retroviral-mediated gene transfer than quiescent cells derived from adult bone marrow, these cells may be an ideal target for gene transduction. Furthermore, physiologic expansion of these cells during development obviates the need for marrow ablation during gene therapy in vivo. We performed in utero gene transfer by injecting high titer replication-defective retrovirus in vivo into the livers of 11, 14, 16, and 18 day gestation rats. After birth, the rats were analyzed for the presence of proviral integration and gene expression. The efficiency of gene transfer into bone marrow cells was greatest in rats infected at day 14 to 16 of gestation. In rats killed at 1 to 26 weeks of age, gene transfer was detected by Southern analysis in 48% and by polymerase chain reaction in 86% of bone marrow samples. The provirus was also detected in white blood cells, the granulocyte-macrophage colony-forming unit, thymus, spleen, liver, and lung. The presence of the transgene in bone marrow and other hematopoietic tissues at 26 weeks of age suggests that early hematopoietic precursors present in the fetal liver are susceptible targets for gene transfer and that these cells become resident in the bone marrow of the adult animal. This model is a new technique for gene transduction into proliferating hematopoietic cells in vivo that avoids bone marrow transplantation and has potential application in the correction of genetic defects in utero.