We have investigated the leukocyte-repopulating-predictive value of granulocyte-macrophage colony-forming unit (CFU-GM) analyses in ex vivo-expanded versus fresh murine bone marrow (BM) grafts. After the transplantation of graded numbers of normal BM cells (from 15 to 5 x 10(3) CFU-GMs/mice), a dose-dependent increase in the recipient leukocytes was observed between the first and third weeks posttransplantation. During these stages, increases in the graft size of 100-fold improved the leukocyte counts up to 30-fold and shortened the leukopenia period by 5 to 11 days, depending on the leukocyte threshold considered. To investigate whether similar correlations could be established using ex vivo-expanded samples, the size of the CFU-GM population was maximized by means of the preactivation of the BM with 5-fluorouracil (9-day 5FU-BM), followed by 3 days of incubation with interleukin-1 plus stem cell factor. Under these conditions, the CFU-GM content of the ex vivo-expanded grafts was 73-fold higher than that observed in equivalent femoral fractions of normal fresh BM. When equivalent fractions of both graft types were transplanted, an improved leukocyte recovery was observed in mice transfused with the expanded grafts. However, the leukocyte values obtained after the transplantation of the ex vivo-expanded samples were not as high as expected, based on the number of transplanted CFU-GMs. Analyses performed during the second week posttransplantation showed that, in comparison with normal fresh BM, ex vivo-expanded grafts containing 6 to 50 times more CFU-GMs were required to generate a similar number of leukocytes. These results were confirmed in both the peripheral blood leukocytes and the myeloid Gr1+ cells, when similar numbers of CFU-GMs were transfused in the fresh and the ex vivo-expanded BM. The possibility that the preactivation of the ex vivo-expanded grafts with 5FU had a role in this effect was ruled out, because the leukocyte repopulation capacity of fresh 5FU-treated BM was as high as that observed in normal fresh BM which contained a similar number of CFU-GMs. Neither by extending the ex vivo incubation period nor by using other hematopoietic growth factor combinations was the functional capacity of the expanded grafts improved. The results presented in this study are consistent with the belief that ex vivo expansion procedures will be a useful tool for improving the hematologic recovery of patients who receive hematopoietic transplants. However, our data indicate that predicting the leukocyte repopulating capacity of ex vivo-expanded grafts according to correlations established with numbers of fresh CFU-GMs can lead to overestimations of their function, and therefore to unexpected and delayed hematopoietic engraftments.