Donor-to-donor variability in the expansion potential of human bone marrow cells is reduced by accessory cells but not by soluble growth factors.

Clinical trials assessing the utility of cultured hematopoietic cells for the support of patients receiving high-dose chemotherapy are beginning. Although many reports have described these cultures, little is known about the donor-to-donor variability that might be expected to occur in widespread use. Therefore, this study was undertaken to assess variables which might predict and reduce the donor-to-donor variability in cell expansion potential. CD34-enriched cell cultures, plated to contain 3000 CD34+lin- cells per well, exhibited a wide range of cell output (0.02 to 5.07 x 10(6)) with a high coefficient of variation (CV = 0.69, n = 52). The range in CFU-GM output was even greater (12 to 9455, CV = 0.90). Addition of preformed stroma had a significant positive effect, and resulted in narrower ranges of cell (0.19 to 8.27 x 10(6), CV = 0.41) and CFU-GM (218 to 17586, CV = 0.54) output. A wide range of stromal-dependency was exhibited by CD34-enriched cells from different donors, with stroma augmenting cell output by 1.2- to 14-fold (mean 3.5), and CFU-GM output by 1.7- to 24-fold (mean 6.5). In contrast, changes in the soluble growth factor combination affected cells from different donors in a similar fashion, thereby altering the mean level of performance without reducing donor-to-donor variability. Experiments were next performed to assess the relative contribution of CD34+lin- cells and stromal cells to culture variability by culturing CD34+lin- cells from three donors on preformed stroma from three donors in parallel. Variability in culture output was attributed to the CD34+lin- cell donor, whereas stroma from different autologous or allogeneic donors gave similar performance. Therefore, both expansion potential and stromal-dependency were inherent characteristics of CD34+lin- cells from different donors. Donor characteristics (i.e., sex, age, weight, and height) and flow cytometric assays (i.e., CD34+lin- cell purity, and CD38-, Thy-1+, and c-kit+ subsets thereof) were not well correlated with expansion potential. In contrast, many of the different biological characteristics (i.e., inoculum CFU-GM, cell and CFU-GM output, and stromal-dependency) were strongly correlated with each other. Mononuclear cell (MNC) cultures, which provide an accessory cell environment (including endogenous stroma) in which CD34+lin- cells grow, were compared with CD34-enriched cell cultures. MNC cultures (containing 3000 CD34+lin- cells) were found to give the greatest and most consistent cell (2.51 to 5.20 x 10(6), CV = 0.17) and CFU-GM (2618 to 14,745, CV = 0.46) output. These results have significant implications for the design of clinical trials of cultured hematopoietic cells, as well as for the understanding of diversity in human stem cell behavior. Furthermore, the results demonstrate the importance of a large sample size in scientific studies of primary human hematopoietic cell behavior.