Quantitative aspects of an in vivo diffusion chamber assay for normal human hemopoietic colony forming units

The purpose of the present study was to definite some of the quantitative characteristics of the colonies formed by normal human hemopoietic cells in diffusion chambers in vivo. The cells were cultured in fibrin clots in diffusion chambers, which were implanted i.p. into 450 R irradiated mice, reimplanted into newly irradiated mice after 7 days and harvested by a previously described chamber centrifugation technique after a total of 14 days. At low cell doses per chamber, no colonies were observed; however when more than 3-8 X 10(4) cells were inoculated per chamber, the number of granulocytic colonies was proportional to the cell dose. Similar dose-reponse curves were obtained when the cells were harvested in suspensions from the chambers. The granulocytic colonies do not conform to a Poisson distribution when identically prepared chamberbeare compared, but show a more heterogeneous distribution. Evidence is presented that this heterogeneity is not due to technical errors in the process of chamber filling and harvesting. A host factor was found to influence the mean number of cells per neutrophilic colony, but the heterogeneity of the total number of neutrophilic or eosinophilic colonies of more than five cells could not be ascribed to variable host factors. There was a mean of 3.6 granulocytic colonies (of more than 30 cells) per 10(5) cells implanted in a series of 15 normal bone marrows cultured. The mean size of neutrophilic colonies was 23 cells. Determinations in duplicate suggest that the method may serve as an assay for certain classes of colony forming cells, provided the number of chambers in each assay is sufficiently high.