EDITORIAL Pre-Progenitor Cells: a Proposed New Category of Hematopoietic Precursor Cells D Metcalf

EDITORIAL Pre-progenitor cells: a proposed new category of hematopoietic precursor cells D Metcalf

The Walter and Eliza Hall Institute of Medical Research, PO Royal Melbourne Hospital, 3050 Victoria, Australia

Hematopoietic precursor cells capable of forming blast colon- Certain surrogate assays for stem cells have been used exten- ies in vitro appear to be a distinct transit population between sively – long-term culture initiating cells (LTCIC),7 cobble- stem cells and lineage-committed progenitor cells because the stone-forming cells with a varying capacity for long-term pro- cells in such blast colonies are mainly lineage-committed pro- 8 9,10 genitor cells. It is proposed that blast colony-forming cells be liferation and spleen colony-forming cells (CFU-S). In termed pre-progenitor cells (PPC) to encourage efforts to each case, there is evidence that the cells detected have some establish their biology and potential for initiating some types capacity for self-generation as well as a capacity for multilin- of leukemic populations more clearly. eage differentiation. It is also apparent however that most cells Keywords: stem cells; progenitor cells; blast colony-forming cells; detected in these latter assays are slightly more mature than hematopoiesis the cells able to sustain long-term hematopoiesis in vivo. What has become increasingly unsatisfactory is how to regard cells capable of forming colonies of blast cells in With increasing knowledge of the hematopoietic cells that are vitro.11–13 Most cells, sorted as stem cells, can probably form ancestral to the morphologically identifiable cells of the vari- blast colonies when cultured in vitro, unless the analysis of ous lineages, it has so far proved useful to divide these cells such cultures is delayed to the point (at 2–3 weeks) when mat- into two categories – lineage-committed progenitor cells and uring cells have begun to appear in the colonies. This latter stem cells. practice of delayed culture readout has almost certainly Lineage-committed progenitor cells are a transit population, resulted in blast colonies also being labeled CFU-A14 or high generated by stem cells and expending themselves by the for- proliferative potential colonies.15,16 The problem with blast mation of maturing cells in various lineages. For most hemato- colonies is that cells with the surface markers of stem cells poietic lineages, these cells are identifiable with high may not be the only cells capable of blast colony formation efficiency because of their ability to generate colonies of mat- in vitro. 1 uring progeny in semisolid cultures. The most numerous Two apparently distinct populations of blast colony-forming committed progenitor cells form colonies composed of cells cells have now been identified in murine bone marrow. One of a single lineage, eg erythroid or granulocytic, but some pro- set is responsive to stimulation by stem cell factor, with or genitor cells are bipotential or multipotential. In murine popu- without potentiation by other growth factors such as the col- lations, common bipotential progenitor cells are granulocyte– ony-stimulating factors or IL-6.13,17 The second set is respon- macrophage or erythroid–megakaryocytic and there is a sive to stimulation by Flk-ligand with potentiation by gp130- numerically smaller subset that exhibit multilineage potential signaling growth factors – LIF, oncostatin M, IL-11 and IL-6.18 if appropriate combinations of stimulating factors are used. It Blast colony-forming cells until now have been regarded is reasonable to assume that multilineage progenitors are with some reservation as stem cells. However, as currently slightly ancestral to either bipotential or unipotential progeni- grown in semisolid cultures, blast colony-forming cells do not tor cells. However, the relative frequencies and varying pro- fit satisfactorily either into the category of stem cells or into liferative potential of these different progenitor cells suggests that of lineage-committed progenitor cells for a number of that not all single lineage progenitors are necessarily the lineal reasons: descendants of bipotential or multipotential progenitor cells. In contrast to the reasonably clear situation with committed • Stem cells capable of repopulation are usually estimated to progenitor cells, there are major problems in determining the have a frequency of only about 1–2 per 105 murine bone interrelationships between the heterogeneous collection of marrow cells but (as shown in Figure 1) the frequency of more ancestral cells currently classified as hematopoietic stem blast colony-forming cells in adult bone marrow approxi- cells. Cells in this class have a uniform morphology of small mates 40 per 105 cells. In contrast, total lineage-committed mononuclear cells, lack basophilic cytoplasm, are mainly not progenitor cells approximate 600 per 105 cells. in cell cycle and can be segregated from other hematopoietic • Blast colony-forming cells, stimulated to proliferate either cells by the presence or absence of various surface markers. by stem cell factor or Flk-ligand have no apparent capacity Thus murine stem cells are kit+ lineage− Ly6A/E+ Rh123lo 2,3 + − + for self-generation because no blast colony-forming cells are cells and corresponding human cells are CD34 CD33 kit 13,18 − lo − 4,5 detectable in such colonies. lineage Rh123 HLA DR cells. In principle, stem cells • After 7 days of incubation, blast colonies are composed pre- should be capable of at least some level of self-generation dominantly of lineage-committed progenitor cells. This is and have an extended capacity for generating cells in multiple in sharp contrast to colonies formed by lineage-committed lineages. The gold standard for a stem cell is that it should be progenitor cells which either contain no progenitor cells or 6 capable of long-term repopulation of an irradiated recipient. only a small number of such cells with a very limited proliferative potential. • Stem cell factor-stimulated blast colonies contain progenitor Correspondence: D Metcalf; Fax: 61 3 9347 0852 cells of multiple lineages, eg granulocytic-granulocyte– Received 17 August 1997; accepted 22 September 1997 macrophage, macrophage or eosinophil and for human col- Editorial D Metcalf 2 proliferative capacity of the progeny progenitor cells. (3) It is also possible that under conditions of basal hematopoiesis, differentiation commitment occurs earlier than in semisolid cultures and restricts the capacity of individual blast colony- forming cells to continue to generate progeny with the proper- ties of progenitor cells. If these alternative behavior patterns can occur in vivo, then the pre-progenitor cell population becomes an important subset of precursor cells, capable of major expansion of progeni- Figure 1 Diagram, not drawn to scale, of the likely relationship tor cell production when required by emergency situations. between stem cells, blast colony-forming cells and committed progenitor cells. Numbers in the boxes indicate the frequency of cells in This would permit a major expansion of hematopoiesis with- each category per 105 murine bone marrow cells. Two distinct popu- out placing corresponding demands on cells in the stem cell lations of blast colony-forming cells exist and these probably include compartment. Furthermore, in some multilineage leukemias, some cells currently labeled stem cells and some regarded as early the initiating clonogenic cell for the leukemic population progenitor cells. It is proposed that blast colony-forming cells be could be a pre-progenitor cell. If so, this would provide an regarded as a distinct subset of ancestral hematopoietic cells and be alternative explanation for the continuing presence of normal renamed pre-progenitor cells (PPC). stem cells able to generate normal hematopoietic populations during remission. The ability to grow identifiable blast colon- onies also megakaryocytic and erythroid progenitors.17 ies in vitro should permit a variety of transformation studies Murine blast colonies initiated by Flk-ligand plus potentiat- to be performed on such cells to establish their potential to ing factors are committed exclusively to the formation of generate leukemic populations. macrophages with some capacity to generate dendritic Whether there is any particular value in designating blast cells.18 colony-forming cells as a special stage in hematopoiesis depends on the results of future observations. How often are These considerations seem to require that blast colony-for- lineage-committed progenitor cells generated by stem cells ming cells be placed in a distinct category and given another without passing through a pre-progenitor stage? The two name. The terminology applied to ancestral cells of B-lympho- known types of pre-progenitor cells appear to be unrelated: cytes in the marrow has proved satisfactory for immunologists does this mean that they are generated by stem cells that are and it may therefore be useful to group the subsets of blast already segregated into different subtypes? Are there pertur- colony-forming cells into a new category of precursor cell, bations in hematopoiesis that selectively involve, or are based termed pre-progenitor cells (PPC). Pre-progenitor cells would on, fluctuations in the size and proliferative behavior of pre- then represent a distinct class of transit cells that are the pro- progenitor cell populations rather than genuine repopulating geny of stem cells and the ancestors of many lineage-commit- stem cells? How often is leukemic transformation able to be ted progenitor cells. It may well be that not all lineage-com- traced to pre-progenitor cells rather than to stem cells? mitted progenitor cells are necessarily the progeny of pre- It seems useful for the present to provisionally regard pre- progenitor cells and some may well be formed directly by progenitor cells as a distinct subclass of hematopoietic precur- stem cells. sors. Because efficient clonal culture methods exist for their The defining characteristic of pre-progenitor cells would be enumeration, the way is open to rapidly collect information their ability to form blast colonies composed of progenitor on their biology, surface markers, cycling status and regulat- cells, allowing them to be accurately enumerated in clonal ory control in various developmental or disease situations. cultures and to be readily distinguished from other hematopoi- With this information, the position can be reassessed as to etic precursors. The assignment of the special name to such whether there is continuing value in regarding them as a cells may eliminate some of the confusion arising from the distinct stage in hematopoiesis. heterogeneity of stem cell populations and some of the reluc- tance to work on cells in the difficult grey area between obvious stem cells and obvious progenitor cells. It should also cat- Acknowledgements alyze experiments to determine the role of this subset of precursor cells in normal development, basal adult This work was supported by the Carden Fellowship Fund of hematopoiesis and in various disease states. the Anti-Cancer Council of Victoria, the National Health and One feature already evident regarding pre-progenitor cells Medical Research Council, Canberra and the National Insti- is the numerical discrepancy between the frequency of pre- tutes of Health, Bethesda, grant No. CA-22556. progenitor and lineage-committed progenitor cells. With a frequency ratio of 1:15, there appears to be an excess number of pre-progenitor cells because it is usual when recloning blast References colonies to detect much higher numbers of lineage-committed progenitors per colony – typically several hundred to more 1 Metcalf D. The Hemopoietic Colony-Stimulating Factors. Elsevier: than 1000. This discrepancy can be interpreted in several Amsterdam, 1984. ways: (1) some pre-progenitor cells may die by apoptosis 2 Spangrude GJ, Heimsfeld C, Weisman IL. Purification and charac- rather than generate progeny; (2) more likely, pre-progenitor terization of mouse hemopoietic stem cells. Science 1988; 241: cells may possess more proliferative potential than is usually 58–62. expressed with the low concentrations of regulatory factors 3 Li CL, Johnson GR. Long-term hemopoietic repopulation by Thy- lo − + impinging on the cells in vivo under basal hematopoietic con- l , Lin , Ly6A/E cells. Exp Hematol 1992; 20: 1309–1315. 4 Sutherland HJ, Lansdorp PM, Henkelman DH, Eaves AC, Eaves ditions. In this regard, the size of blast colonies can readily CJ. Functional characterization of individual human hematopoi- be increased in vitro by increasing the concentration either of etic stem cells cultured at limiting dilution on supportive marrow stem cell factor or Flk-ligand without affecting the subsequent stromal layers. Proc Natl Acad Sci USA 1990; 87: 3584–3588. Editorial D Metcalf 3 5 Udomsakdi C, Lansdorp PM, Hogge DE, Reid DS, Eaves AC, Eaves renew and generate multipotential colonies. Proc Natl Acad Sci CJ. Characterization of primitive hematopoietic cells in normal USA 1982; 79: 3843–3847. human peripheral blood. Blood 1992; 80: 2513–2521. 13 Metcalf D. Lineage commitment of hemopoietic progenitor cells 6 Li CL, Johnson GR. Murine hematopoietic stem and progenitor in developing blast colonies: influence of colony-stimulating fac- cells: I. Enrichment and biological characterization. Blood 1995; tors. Proc Natl Acad Sci USA 1991; 88: 11310–11314. 85: 1472–1479. 14 Holyoake TL, Freshney MG, Konwalinka G. Haun M, Petzer A, 7 Hogge DE, Sutherland HJ, Cashman JD, Lansdorp PM, Humphries Fitzsimons E, Lucie NP, Wright EG, Pragnell IB. Mixed colony RK, Eaves CJ. Cytokines acting early in human haematopoiesis. formation in vitro by the heterogeneous compartment of multipo- Baillie`re’s Clin Haematol 1994; 7: 49–63. tential progenitors in human bone marrow. Leukemia 1993; 7: 8 Breems DA, Blokland EAW, Neben S, Ploemacher RE. Frequency 207–213. analysis of human primitive haematopoietic stem cell subsets 15 Bradley TR, Hodgson GS. Detection of primitive macrophage pro- using a cobblestone area forming cell assay. Leukemia 1994; 8: genitor cells in mouse bone marrow. Blood 1979; 54: 1446–1450. 1095–1104. 16 McNiece IK, Bertoncello I, Kriegler AB, Quesenberry PJ, Zsebo 9 Till JE, McCulloch EA. A direct measurement of the radiation sensi- KM. Colony-forming cells with high proliferative potential. In: tivity of normal mouse bone marrow cells. Rad Res 1961;14: Murphy MJ (ed). Concise Reviews in Clinical and Experimental 213–222. Hematology. AlphaMed Press: Dayton, OH, 1992, pp 267–277. 10 Curry JL, Trentin JJ. Hemopoietic spleen colony studies. I. Growth 17 Metcalf D. The cellular basis for enhancement interactions and differentiation. Devel Biol 1967; 15: 395–413. between stem cell factor and the colony-stimulating factors. Stem 11 Keller GM, Phillips RA. Detection in vitro of a unique multipotent Cells 1993; 11: (Suppl. 2): 1–11. hemopoietic progenitor. J Cell Physiol 1982; 111 (Suppl. 1): 31– 18 Metcalf D. Murine hematopoietic stem cells committed to 36. macrophage/dendritic cell formation: stimulation by Flk2-ligand 12 Nakahata T, Ogawa M. Identification in culture of a class of hemo- with enhancement by regulators using the gp130 receptor chain. poietic colony-forming units with extensive capability to self- Proc Natl Acad Sci USA 1997; 94: 11552–11556.