Granulocyte colony-stimulating factor enhances bone marrow stem cell damage caused by repeated administration of cytotoxic agents.

Despite the increasing use of cytokines to circumvent the acute dose-limiting myelotoxicity of cancer treatment, little is known about the combined effects of cytotoxic agents and cytokines on the primitive stem cells responsible for long-term hematopoiesis. In an experimental model, we administered cytotoxic agents that have variable effects on primitive stem cells in C57BL/6 (B6)-mice. Mice received six every-other-week doses of cyclophosphamide (CY, 84 mg/kg), VP-16 (24 mg/kg) + cisplatinum (2.4 mg/kg), carboplatinum (50 mg/kg), chlorambucil (12 mg/kg), BCNU (13.2 mg/kg), or TBI (80 cGy). Granulocyte colony-stimulating factor (G-CSF; 250 microg/kg/day) was administered subcutaneously twice daily on days 3 to 6 after each dose of the cytotoxic agent. Comparison with animals receiving the cytotoxic agent alone was made to investigate the effects of G-CSF on long-term hematopoiesis. Hematopoiesis was measured 20 weeks after the last dose of the cytotoxic agent by assessment of peripheral blood counts, marrow cellularity, progenitor cell content (colony-forming units-spleen; CFU-S), and primitive stem cell number (long-term repopulating ability and day 28 and day 35 cobblestone area-forming cell [CAFC] frequencies). Exposure to cytotoxic agents alone resulted in a significant decrease in primitive stem cells (as measured by repopulating units [RU] and day 28 and day 35 CAFC content) in animals given carboplatinum, chlorambucil, BCNU, and TBI, but not in animals treated with cyclophosphamide or VP-16 and cisplatinum. The addition of G-CSF resulted in a significant decrease in stem cell content when compared with no G-CSF administration in animals treated with chlorambucil, BCNU, or TBI. Thus, G-CSF administered after repeated exposure to cytotoxic agents, appeared to damage the primitive stem cell compartment when used in combination with agents known to damage primitive stem cells. These results, although obtained in an experimental model, should raise concerns for the indiscriminate use of G-CSF in the clinic.