Reprinted from JOURNAL OF CELLULAR AND COMPARATIVE PHYSIOLOGY

Many of the models of hemopoiesis that have been proposed (see, for example, Cronkite et al., '59) are based on the assumption that the continued production of blood cells requires the presence of progenitor cells with the capacity for continued proliferation. From this point of view, hemopoietic tissue may be considered to consist of two compartments; the first, or stem cell compartment, consists of progenitor cells with the capacity to give rise to progeny consisting of both differentiated cells and new stem cells; the second, or differentiated cell compartment, contains cells with limited capacity for cell division, giving rise only to fully differentiated cells. It follows that studies on the processes involved in hemopoiesis require methods for determining the composition of the two compartments. Members of the differentiated cell compartment can frequently be recognized by clear functional markers, for example, the ability to incorporate radioiron (Alpen and Cranmore, '59). In contrast, recognition and assay of stem cells must involve a procedure in which the descendants of the stem cell are examined. Ideally, such a procedure should test the stem cell not only for its ability to give rise to differentiated descendants (which is the basis for the assay for stem cells described by Gurney and co-workers (Gurney et al., '62)), but should also test for other key properties of stem cells. These include the capacities for self-renewal and extensive proliferation, both of which are required for the maintenance of the stem cell compartment. Recently, a method has been developed which may fulfill these requirements for studies of stem cells. The method depends on the observation that mouse hemopoietic tissue contains a population of cells that have the capacity to give rise to macroscopic colonies in the spleens of irradiated mice (Till and McCulloch, '61; McCulloch and Till, '62). It has been demonstrated by direct cytological means that these colonies originate from single cells (Becker et al., '63), showing that their cells of origin (colony-forming cells) possess sufficient proliferative capacity to give rise to a macroscopic colony containing of the order of a million cells. Further, histological examination of individual colonies reveals that they contain differentiated cell precursors; indeed, colonies frequently are found to contain more than one kind of differentiated cell, and often erythrocytic, granulocytic and megakaryocytic precursors may be observed within a single colony (McCulloch, '63). Thus the spleen colony technique appears to satisfy some of the requirements of an assay for stem cells, in that the cell type it detects has the capacity for extensive proliferation and is capable of giving rise to progeny containing differentiated cells. However, it remained to be demonstrated that the colony-forming cells possess the capacity for self-renewal. Information bearing on this problem was obtained in the work described here, by testing single colonies for their content of colonyforming cells. The results obtained indicate that colony-forming cells are indeed capable of self-renewal, in that many colonies were found to contain new colony-forming cells. Moreover, the distribution of colony-forming cells per colony was found to be extremely heterogeneous, with some colonies containing many new colony-forming cells and others containing very few. Experiments designed to investigate this heterogeneity have provided