Differentiation of natural killer (NK) cells from human primitive marrow progenitors in a stroma-based long-term culture system: identification of a CD34+7+ NK progenitor

We have recently described a marrow stroma-dependent long-term culture system that supports differentiation of CD34+ human marrow primitive progenitors into natural killer (NK) cells. We postulate that CD7 expression may be an early event in commitment of hematopoietic progenitors to the NK lineage. Here we compare the characteristics of CD34+7- and CD34+7+ marrow cells cultivated in the stroma-based NK culture system. These CD34+ populations were further compared with a marrow derived, more committed, CD34-7+ progenitor to emphasize the continuum of NK development and to highlight differences between progenitors in our assays. No progenitor proliferated when plated in media without stroma, underscoring the importance of stroma in NK differentiation. Plating progenitor populations in interleukin-2 containing media directly on preestablished, allogeneic, irradiated marrow stroma for 5 weeks resulted in CD56+CD3- NK cells; however, characteristics of the cultured populations differed. Fold expansion and cloning efficiency of the CD34+7+ population, determined by a functional limiting dilution assay was significantly higher than of the CD34+7- or CD34+7+ populations. This suggests that the CD34+7+ population is highly enriched for an NK progenitor and a possible intermediate in NK lineage differentiation. Further dividing the CD34+7+ population by the relative fluorescence of CD7 into CD34+7+dim and CD34+7+bright populations showed that the CD34+7+bright population exhibited a significantly higher cloning frequency than parallel experiments with CD34+7+dim cells (11.8% +/- 2.4% v 2.4% +/- 0.7%, n = 6; P = .005). Plating of the more primitive CD34+7- population in a transwell system (which separates progenitors from stroma by a microporous membrane) prevents differentiation into NK cells. In contrast, plating of CD34+7+ progenitors in transwells resulted in generation of NK cells. These data suggest that primitive, but not more mature NK progenitors may require direct contact with stroma for the initial differentiation steps. Finally, differentiation of the NK progenitors in this stroma-dependent model results in expression of CD2 not present on any of the starting populations. This observation suggests that marrow stroma can stimulate CD2 expression on NK progenitors in a previously undescribed fashion that may be analogous to the thymic effect on CD2 expression in immature T lymphocytes. These observations identify early steps in the commitment of primitive marrow CD34+ hematopoietic progenitors to a lymphoid lineage and underscore the importance of coexpression of CD7 with CD34 as an early lymphoid commitment characteristic and direct progenitor-stroma interactions in this process.