Umbilical cord blood is capable of hematopoietic stem cell reconstitution in children. However, the major limitation of cord blood is a relatively low content of pluripotent progenitor cells. Thus, safe engraftment for adolescents and for adults is still not predictable and a technology for ex vivo expansion of umbilical cord blood cells is desirable. In a first step, four different methods of red cell depletion followed by magnetic cell sorting of CD34+ cells were evaluated in this study in order to assess the efficacy and safety of optimal stem cell recovery. A modified two-step Ficoll gradient separation and a hydroxyethyl starch separation tended to produce a better WBC/MNC recovery (median 94.2+/-2.44% vs. 90.2+/-5. 8%) as compared with standard Ficoll gradient separation and a gelatin-based procedure (median 78.35+/-7.1% vs. 67.2+/-5.5%). However, the recovery of CD34+ cells after magnetic cell sorting did not reach a statistically significant difference after the four different methods of red cell depletion, indicating that the recovery of WBC/MNC is not predictably correlated with the recovery of stem cells within these fractions. In a second step, we established three different cytokine combinations by adding the megakaryocyte growth and development factor +/- erythropoietin and granulocyte colony-stimulating factor to a fetal calf serum containing medium with Flt 3, stem cell factor, and interleukin-3. Net expansion of total colony-forming cells 20- to 50-fold and expansion of colony-forming cells after 5 weeks of culture 1.5- to 3-fold were obtained over a period of 7-14 days. These results demonstrate that cord blood stem cells can be expanded substantially in this short-term culture system.