Identification of functionally distinct domains of human granulocyte-macrophage colony-stimulating factor using monoclonal antibodies.

ABSTRACT

Granulocyte-macrophage colony-stimulating factor (GM-CSF) is a glycoprotein that is required for the survival, growth, and differentiation of hematopoietic progenitor cells. Although the primary structure of GM-CSF is known from cDNA cloning, the relationship between structure and function of GM-CSF is not fully understood. Fifteen different monoclonal antibodies (MoAbs) to human GM-CSF were generated to map immunologically distinct areas of the molecule. Each of the MoAbs was biotinylated and shown by enzyme-linked immunosorbent assay to bind to recombinant GM-CSF that had been affixed to a solid phase. Each of the 15 unconjugated MoAbs was then used to compete with each biotinylated MoAb for binding to GM-CSF. These cross-blocking studies identified eight distinct epitopes of native GM-CSF. Seven of these epitopes were also present in denatured GM-CSF by Western blotting, and four of the epitopes were at least partially conserved on GM-CSF that was reduced in beta-mercaptoethanol. MoAbs to four of eight epitopes neutralized both recombinant (glycosylated and nonglycosylated) and natural human GM-CSF in a GM colony-forming unit (CFU-GM) assay and blocked GM-CSF-induced activation of neutrophils. For most of the antibodies there was a good correlation between neutralizing activity and the capacity to block binding of 125I-GM-CSF to neutrophils or blasts. Non-neutralizing antibodies to one epitope partially blocked binding of 125I-GM-CSF to neutrophils. None of the MoAbs neutralized interleukin-3, G-CSF, or M-CSF. The locations of seven of the epitopes could be partially mapped with regard to the amino acid structure by determining reactivity to GM-CSF synthetic peptides or to human-mouse chimeric GM-CSFs. The neutralizing antibodies were found to map to amino acids 40-77, 78-94, or 110-127. Thus, these MoAbs are useful to identify functional domains of GM-CSF and in identifying regions that are likely to be involved in receptor interaction.