The report by Ehrlich et al in this issue of Blood reveals that assigning lineage potential to hematopoietic progenitors can be a challenging task, with interpretations differing widely depending on which functional assays are used.1 The findings add a new angle to the controversy of how the lymphoid and myeloid lineages are related during normal differentiation of hematopoietic stem cells.
The most familiar model of hematopoiesis describes the first branch point in lineage commitment as a decision between lymphoid (B, T, and NK) and nonlymphoid (myeloid, erythroid, and megakaryocytic) lineages. Incorporated into this model is the concept that T cells are generated in the thymus from progenitors that first undergo lymphoid commitment in the bone marrow. In support of this model, early studies identified “common lymphoid progenitors” (CLPs) in human2 and murine3 bone marrow that, after transplantation, could produce only lymphoid lineages in the recipient. This lymphomyeloid dichotomy has been more recently challenged by 2 findings: (1) the identification of lymphomyeloid progenitors in murine bone marrow that lack erythroid or megakaryocytic potential,4 and (2) the finding of progenitors in the human5,6 and murine7,8 thymus with both T and myeloid potential.
In the current study, Ehrlich et al compared how in vitro and in vivo assays perform in the assessment of lineage potential of previously described progenitor populations in murine bone marrow and thymus. Early thymic progenitors produced T cells but almost no myeloid cells after transplantation into either sublethally or lethally irradiated congenic hosts. However, when the same thymic progenitors were cultured in commonly used stromal-based assays (on the murine stromal line OP9 in supraphysiologic concentrations of growth factors), they readily generated both myeloid and T cells. Even CLPs, a population with lymphoid restriction as its sine qua non based on in vivo assays, were able to generate myeloid cells in the OP9 stromal cultures. Interestingly, a different type of in vitro assay that does not rely on stromal cocultivation, the colony-forming unit–cell (CFU-C) assay, did not reveal clonogenic myeloid potential in either the thymic progenitor or bone marrow CLP population.
The question then is which assay(s) should be used as the “gold standard” for measuring lineage potential? Ehrlich et al argue that only in vivo readouts should be used to assign lineage potential because the stromal cocultivation system produces artificial signals that induce abnormal differentiation patterns. Based on this standard, the thymic progenitors and CLP populations would be defined as lymphoid-restricted. It could be argued that the CFU assay is a more stringent in vitro assay than the stromal cocultures, as it detects only clonogenic myeloid progenitors and correlated well with in vivo readouts.1
It is interesting to note that the myeloid cells produced by these “lymphoid” populations are predominantly monocyte/macrophages, a cell type that can differentiate into dendritic cells. As dendritic cells can be readily generated from both myeloid and lymphoid progenitors,9 the common presence of macrophages and dendritic cells in cultures might be seen as a common default pathway of all progenitors, at least in vitro. To put it another way, are macrophages really myeloid-restricted?
Although no in vitro system fully recapitulates the physiologic signals of the hematopoietic niche, transplantation assays may also have limitations in certain settings. Murine transplantation assays have long reigned as the best proof of the hematopoietic stem cell, as they measure the defining characteristics of stem cells: self-renewal and multilineage repopulation potential. The transplantation assay works for hematopoietic stem cells because these cells naturally function to repeatedly enter and exit the circulation during normal homeostasis. However, some types of progenitors which mark important stages of lineage commitment during hematopoiesis may have little or no ability to home to and proliferate in the bone marrow after transplantation, and will thus be difficult or impossible to detect in transplantation assays. This is particularly likely for progenitors isolated from nonmarrow sites such as the thymus, where very different environmental signals exist and homing to marrow does not occur. In addition, the environment of the irradiated host is different to that of the resting marrow space and may favor engraftment of certain progenitor types over others. Studies of human hematopoietic progenitor pathways confront even greater challenges, as in vivo measurements rely on xenogeneic transplants in which poor homing efficiency and lineage skewing are seen in even the most recently developed immune-deficient mouse models.10
The question of how to interpret these widely used in vitro and in vivo assays is not as esoteric as it may seem. The strengths and weaknesses of these assays have produced the blueprints hematologists have used for many decades to draw, and more recently to redraw, the complex pathways of hematopoiesis.
Conflict-of-interest disclosure: The author declares no competing financial interests. ■
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal