Standardization of the CFU-GM assay: Advantages of plating a fixed number of CD34+ cells in collagen gels

We investigated whether plating a stable amount of CD34(+) cells improves the CFU-GM assay. Data of CFU-GM assays performed with leukaphereses products in two transplant centers using a commercial collagen-based medium and unified CFU-GM scoring criteria were pooled and analyzed according to the numbers of CD34(+) cells plated. A first series of 113 CFU-GM assays was performed with a fixed number of mononuclear cells (i.e., a variable number of CD34(+) cells). In these cultures the CFU-GM/CD34 ratio varied according to the number of CD34(+) cells plated: median CFUGM/CD34 ratios were 1/6.2 to 1/6.6 for grafts containing <2% CD34(+) cells, vs. 1/10.2 for grafts containing > or =2% CD34(+) cells. The median CFU-GM/CD34 ratio also varied depending on pathology: 1/9.3 for multiple myeloma (MM), 1/6.8 for Hodgkin's disease (HD), 1/6.5 for non-Hodgkin lymphoma (NHL), and 1/4.5 for solid tumors (ST). A second series of 95 CFU-GM assays was performed with a fixed number of CD34(+) cells (220/ml). The range of median CFU-GM/CD34 ratios was narrowed to 1/7.0 to 1/5.2, and coefficients of variation for CFU-GM counts decreased by half to 38.1% (NHL), 36.1% (MM), 49.9% (HD), and 22.4% (ST). In addition, CFU-GM scoring was facilitated as the percentages of cultures with >50 CFU/GM/ml decreased from 6.7% to 43.8% when a variable number of CD34(+) cells was plated, to 4.5% to 16.7% when 220 CD34(+) cells/ml were plated. Hence, plating a fixed number of CD34(+) cells in collagen gels improves the CFU-GM assay by eliminating cell number-related variability and reducing pathology-related variability in colony growth.