To facilitate the investigation of the direct interaction between hematopoietic progenitors and colony-stimulating factors, we have developed a method to purify human marrow progenitor cells. Using density centrifugation, negative panning with concanavalin A coated plates, positive selection of CD34-positive cells with immunomagnetic microspheres, overnight adherence to a plastic dish, negative selection with a panel of monoclonal antibodies, and density centrifugation, human marrow progenitor cells were purified from 1.5% to 53.2%, a 42- fold purification, with a 4.8% yield. The purified cells consisted of 38% erythroid, 9% colony forming unit-granulocyte (CFU-G), 29% CFU- macrophage (CFU-M), 12% CFU-eosinophil/basophil (CFU-Eo/Ba), and 4% CFU- mix. The purified cells cultured in serum-free fibrin clots with recombinant human macrophage colony-stimulating factor (rM-CSF) for 14 days developed a pure population of CFU-M colonies. An appearance of CFU-M colonies was present after the addition of 1 U/mL of rM-CSF and the maximum stimulation was found at 100 U/mL. When the purified cells were cultured in serum-free medium with rM-CSF in a limiting dilution assay and the percentage of nonresponder wells for CFU-M colonies was plotted against cell concentration, serum-free cultures yielded a straight line through the origin, indicating that CFU-M development did not depend on accessory cells and that rM-CSF acted directly on the CFU- M.

To facilitate the investigation of the direct interaction between hematopoietic progenitors and colony-stimulating factors, we have developed a method to purify human marrow progenitor cells. Using density centrifugation, negative panning with concanavalin A coated plates, positive selection of CD34-positive cells with immunomagnetic microspheres, overnight adherence to a plastic dish, negative selection with a panel of monoclonal antibodies, and density centrifugation, human marrow progenitor cells were purified from 1.5% to 53.2%, a 42- fold purification, with a 4.8% yield. The purified cells consisted of 38% erythroid, 9% colony forming unit-granulocyte (CFU-G), 29% CFU- macrophage (CFU-M), 12% CFU-eosinophil/basophil (CFU-Eo/Ba), and 4% CFU- mix. The purified cells cultured in serum-free fibrin clots with recombinant human macrophage colony-stimulating factor (rM-CSF) for 14 days developed a pure population of CFU-M colonies. An appearance of CFU-M colonies was present after the addition of 1 U/mL of rM-CSF and the maximum stimulation was found at 100 U/mL. When the purified cells were cultured in serum-free medium with rM-CSF in a limiting dilution assay and the percentage of nonresponder wells for CFU-M colonies was plotted against cell concentration, serum-free cultures yielded a straight line through the origin, indicating that CFU-M development did not depend on accessory cells and that rM-CSF acted directly on the CFU- M.