A quantitative analysis of expression levels of GM-CSF receptors was performed by flow cytometry in different disease categories, ie AML (n = 72), ALL (n = 18), and MDS (n = 12), as well as 12 healthy volunteers, using three different unconjugated GM-CSF/R monoclonal antibodies (McAbs) (HGM-CSFR (CD116), M5D12, 4B5F5), and appropriate standards. By using the reference HGM-CSFR McAb, in healthy subjects we found detectable levels of GM-CSF/R on blood monocytes (mean MESF (molecules of equivalent soluble fluorochrome)/cell: 36.1 x 10[3]), neutrophils (mean MESF/cell: 7.4 x 10[3]), bone marrow (BM) myelo-monocytic precursors (MESF range for the myeloid component, ie promyelocytes, myelocytes, metamyelocytes: 11.7-40.5 x 10[3], and for the monocytic lineage: 25.7-69.2 x 10[3]), and in two distinct subsets of BM CD34+ progenitor cells (GM-CSF/R dim: 2.5 x 10[3] MESF/cell, GM-CSF/R bright (10% of the total number of CD34 cells: 22.0 x 10[3] MESF/cell). In these subjects, there was no correlation between the expression levels of GM-CSF/R and CFU (CFU-GM, CFU-GEMM, BFU-E) colony production. Among the AML samples, M5D12 McAb was positive in 33%, 4B5F5 McAb in 90%, and HGM-CSF/R McAb in 78% of the cases examined (range of MESF/cell for the HGM-CSFR McAb: 0.9 x 10[3]-106.7 x 10[3]). The highest MESF values were seen in the M5 FAB subvariety (mean: 39.4 x 10[3]), where all the patients tested (n = 20) showed a strong positivity for the HGM-CSFR McAb. On the contrary, all ALL samples were GM-CSF/R negative except in two patients, who displayed a dim GM-CSF/R positivity (My+ALL: 1.3 x 10[3] MESF/cell; pro-B ALL: 1.0 x 10[3] MESF/cell). In most (>70%) M1 FAB subtypes, GM-CSF/R+ blasts co-expressed CD34low, HLA-DRhigh, CD33, CD38 antigens, and had little or no capacity to form CFU-GM colonies. GM-CSF/R+ blasts from the M5 FAB category were also positive for CD14, CD11c, CD33 and CD87. Furthermore, the number of GM-CSF/R expressed by leukemic cells from five out of 72 (7%) AML patients was above the highest values seen in normal samples (>69.2 x 10[3] MESF/cell), allowing the possibility of using this marker for the monitoring of the minimal residual disease (MRD) in a subset of AML. Cell culture studies aimed at evaluating GM-CSF receptor modulation following AML blast exposure to rhGM-CSF showed two distinct patterns of response; in the first group (6/10 cases) rhGM-CSF down-modulated GM-CSF receptors, whereas in the second group (4/10 cases), rhGM-CSF treatment was associated with either an increase or no change in the number of GM-CSF/R. In conclusion, cellular GM-CSF/R expression was variable and ranged from undetectable (ALL and a minority of AML) to very high intensities in M5 AML, and were also documented in some M0 AML, thus suggesting the concept that GM-CSF/R detection may be of help in lineage assignment of undifferentiated forms. Since the number of GM-CSF/R on AML blasts may be modulated after GM-CSF treatment, it can be postulated that the clinical use of GM-CSF in this disease may be optimized by a dynamic analysis of the number and the affinity status of GM-CSF-R in blasts and normal hemopoietic cells.