Kinetics of colony formation by BFU-E grown under different culture conditions in vitro.

Colony formation by erythroid burst-forming units (BFU-E) involves a variable number of cell divisions before individual 'subcolonies' begin to appear. Consequently the numbers of subcolonies vary amongst individual bursts. If this observation is interpreted as a reflection of a stochastic process, the number of subcolonies in each individual burst represents the number of divisions by the BFU-E prior to commitment to terminal differentiation. This provides a means for quantitating the probability of erythroid differentiation (pD) and the probability of renewal (1 - pD). In order to determine whether these kinetics of burst formation can be influenced by exogenous factors we used three commercially available media designed for the growth of BFU-E. We found that subcolony numbers per burst ranged from one to 64 and that the cumulative distributions of subcolonies per burst followed a logarithmic curve (r > 0.90). Differences were observed in the distribution of subcolonies per burst when BFU-E were grown in different media (P=0.03; Kruskall-Wallis test). The probability of immediate terminal differentiation (i.e. committment to form a subcolony) was 0.25 for two of the media and 0.7 for the third. The corresponding renewal probabilities were O.75 and O.3. These data indicate that the proliferation kinetics of BFU-E are susceptible to regulation by exogenous factors.