Myelogenous leukemic hematopoiesis is monoclonal, but otherwise simulates normal hematopoiesis in that it is generated by a stem cell, and differentiation to progenitor cells and maturation to mature cells occur. Thus, the hierarchy of cells from stem to multipotential and monopotential progenitors and their maturation to precursors and mature cells mimics normal hematopoiesis. The execution of these genetic programs, however, ranges from closely simulating normal to bizarrely and unpredictably anarchic and dysfunctional. Moreover disturbances in anchoring relationships of cells to stroma and insufficient or exaggerated expression of cell death programs adds to the variety of malignant phenotypes. In this report, the evidence for the nature of leukemic hematopoiesis and the critical interplay between leukemic and normal hematopoiesis is reviewed from a historical perspective. A future focus on (1) the chemistry underlying the mechanisms of leukemic inhibition of normal hematopoiesis; (2) potential mechanisms to impair leukemic stem cell funtion; and (3) enhancement of normal stem cell function, so as to tip the balance in favor of the latter may provide needed new approaches to therapy.
Tissues that have a high rate of mature cell turnover must be organized in a hierarchy of cells with tight control over the processes of, first, stem cell self-renewal and differentiation and second, proliferation and maturation of progenitor cells, in order to meet the demands of programmed cell death or pathological consumption of mature cells. Hematopoiesis is particularly complex since ten unique basic lineages are derived from marrow parenchyma.
Early twentieth century morphological observations
At the beginning of the twentieth century, several schools of thought evolved regarding the formation of blood cells, including the monophyleticists, the dualists, the triadists, and the polyphyleticists. In 1909, Maximow in a lecture to the Hematologic Society in Berlin argued, based on his histologic studies, that the different blood cell lineages were derived from a common cell, which he described as a lymphocyte, implying that a morphologically bland primitive cell with a high nuclear to cytoplasmic ratio was capable of differentiating into multiple distinct blood cell lineages.1 Some have inferred that by calling the stem cell a lymphocyte, Maximow implied that the cell circulates, a concept established experimentally over 50 years later.2
In 1916, Vera Danchakoff, a Russian emigré working at the Rockefeller Institute, delivered an extraordinary lecture at the College of Physicians and Surgeons of Columbia University. She described morphological and embryological studies in several species that led her to verify that a common morphologic unit, the stem cell, gives rise to blood cells. She stated in her 1916 lecture,3 ‘the erythrocytes, the small lymphocytes, the different leucocytes, the wandering cells of the connective tissue, the mast cells, and the plasma cells – all these cells are different cell units, morphologically as well as physiologically but in the early embryonic stages they all had a common mother cell, and this mother cell is preserved in the adult organism and becomes the source of differentiation and regeneration and most probably also the source of pathological proliferation’. Danchakoff's last statement that the stem cell might undergo pathological change represented one of the earliest suggestions that there were hematopoietic stem cell diseases. The rest of the century has been spent filling in the details of the monophyleticists's experimental insights! Medicine was slow to accept the lymphohematopoietic stem cell concept and at the mid-twentieth century textbooks of the day still promulgated separate cells of origin for lymphocytes from lymphatic tissue, for monocytes from the so-called reticuloendothelial system, and for granulocytes, erythrocytes and platelets from the marrow.4
Mid-twentieth century cell physiological observations
Notwithstanding the genius of the best embryologists of the first half of the twentieth century, few additional insights into the process of hematopoiesis were discovered until the 1950’s when, motivated by the Manhattan District Project, first secret and then unclassified studies of radioprotection, under the aegis of the Atomic Energy Commission, established that cell transfer from a normal mouse, rat or dog could re-establish lymphohematopoiesis in a lethally irradiated animal.5 However, the cell or cells responsible for this effect were unknown.
The first physiological experimental evidence that blood cells are derived from a common cell was published in 1961 by Till and McCulloch who developed an ingenious assay of the radiosensitivity of marrow cells.6 They used lethally irradiated mice as their ‘culture’ vehicle, measuring the number and size of spleen colonies resulting from the infusion of donor mouse marrow cells. The donor cells could be treated with varying doses of radiation before or after their harvest from donor animals. The infused donor cells seeded the spleen and, depending on their proliferative potential, formed colonies of varying size in that site. The spleen nodules and smaller colonies could be studied histologically as to their cell composition. The objective of the study was to develop a means of studying the radiosensitivity of marrow cells. Two observations in this model led Till and McCulloch from radiobiology to hematology and profoundly energized the study of hematopoiesis: the nodules or colonies were shown to be derived from a single implanted donor primitive cell, that is, they were a clone and, second, the presence in a single colony of mixtures of granulocytic, erythroid, megakaryocytic and sometimes lymphoid cells indicated that some spleen colony-forming cells could differentiate into all blood cell lineages including lymphocytes.
These seminal studies provided the first physiological evidence in support of the concept of a primitive undifferentiated cell as the origin of lymphoid and hematopoietic cells. A series of clinical observations suggested that the somatic mutations that leads to AML is in such an early multipotential hematopoietic cell, the cell proposed by the monophyleticists and assayed by Till and McCulloch. First, the predominant morphological appearance of AML, although usually granulocytic could be monocytic, erythrocytic, megakaryocytic, macrophagic, basophilic, mastocytic, or eosinophilic and, importantly, often a mosaic of two, three, or more of these phenotypes.7
The most parsimonious interpretation of these observations is that the several phenotypes of AML are a reflection of the neoplastic multipotential hemopoietic cells capability to differentiate into committed progenitor cells and that these cells retain the capability, albeit imperfectly, to mature into identifiable cells of the erythroid, granulocytic, monocytic, or megakaryocytic lineage. The variable degree of differentiation and lineage maturation of the clone accounts for the numerous phenotypes of myelogenous leukemias. Of course, genetic lesions ultimately account for the phenotype but the correlation of genotype as determined at the level of current cytogenetic techniques with phenotype is not precise. The ability of the leukemic stem cell to differentiate became even more apparent after the introduction of immunophenotyping to determine the lineage of amorphous leukemic cells. At about the time of Till and McCulloch's initial studies, the introduction of cancer cytogenetic techniques provided the opportunity to show that marrow erythroblasts and multinucleated giant cells, presumptive megakaryocytes, contained the same clonal cytogenetic marker as did leukemic myeloblasts, providing further circumstantial evidence that AML arose in a cell capable of at least trilineage differentiation.
In the mid-1960s, the introduction of a method to determine the clonal origin of neoplastic tissue by examining the expression of an X chromosome-linked polymorphism in the tumors of informative women permitted experimental verification of the cell lineages that were part of a clone. This was an extremely important application of the x-inactivation hypothesis by Linder and Gartler, a pathologist and geneticist at the University of Washington.8 They analyzed glucose-6-phosphate dehydrogenase isotypes extracted from samples of uterine tissue and established the utility of the method. Leiomyomata in a woman who was heterozygous for isoenzymes A and B expressed one glucose-6-phosphate isotype, either type B or type A. Intervening normal myometrium expressed both enzymes. They concluded that among the several possible explanations for this result, the most attractive was that the leiomyomas were clonal outgrowths of a somatic mutation in one uterine muscle cell.
This method and its more sensitive successors provided the first universal definition of a neoplasm, that is a tissue alteration that arises from a somatic mutation in a single tissue cell. The technique was applied next to blood cells in leukemia. The confirmation by this method and by cytogenetic techniques of the clonal derivation of at least eight blood cell lineages in chronic myelogenous leukemia and its status as an invariable forerunner of acute leukemia, was powerful support for a stem cell origin of the myelogenous leukemias.9 Thus, the three techniques shown in Figure 1, the spleen colony assay, visualization of human chromosomes in hematopoietic cells, and X chromosome-linked allelic polymorphisms resulted in a concept of leukemic hematopoiesis that mimicked, albeit imperfectly, normal processes.
Figure 2 depicts a schematic of the behavior of the presumptive myelogenous leukemia stem cells.10 The leukemic stem cell must self-renew and can differentiate. The latter step is variable and unpredictable but occurs, no matter how undifferentiated the blast cells appear. Cell multiplication or proliferation is not regulated normally. Blast cells may be amitotic, in G zero (potentially mitotic), or mitotic. The latter blasts may mature to recognizable cells. Maturation is variable and unpredictable. Because of the loss of anchoring relationships to the stroma, any array of these cells may be released into the blood. The process most closely simulates normal in the clonal cytopenias (inappropriately referred to as myelodysplasia),11 the clonal cytoses, such as polycythemia vera, and deviates most from normal in acute myeloblastic leukemia with every conceivable pattern in between. Dysregulation of apoptosis of blast cells and later progenitors also affects leukemic cell accumulation in marrow and blood and thus the phenotype of the leukemia.
No sooner had cytogenetic studies and the glucose-6-phosphate dehydrogenase isoenzymes assay establish the clonal origin of all blood cell lineages in acute and chronic myelogenous leukemia, some exceptions were found. In some cases of AML, the blood cells at presentation were a mosaic of clonally-derived leukemic blast cells and polyclonally-derived red cells and platelets, providing evidence that in some cases of AML the somatic mutation occurred in a more differentiated progenitor cell, closer to a cell committed to the granulocytic lineage. There is precedence for this phenomenon since virtually all the lymphoid malignancies ranging from acute to chronic lymphocytic leukemia to myeloma originate as progenitor cell tumors. Young persons with AML, some cases of translocation 8; 21, and some cases of monocytic leukemia were found to have red cells and platelets that were not part of the leukemic clone suggesting that the neoplasm was a ‘progenitor cell leukemia’.10 Moreover, in this circumstance basal levels of normal stem cell function must continue.
Late-twentieth century cell phenotyping, genotyping, and xenografting observations
Further study of these propositions was made possible by three key advances in analytical techniques shown in Figure 3. The development of high-speed fluorescence activated cell identification and sorting instruments, the identification of the antigenic phenotype profile at each level of the hierarchy of hematopoietic cell differentiation, and the development of profoundly immunodeficient mice that permits the growth and differentiation of transplanted primitive human normal or leukemic hematopoietic cells, were each major steps in understanding the pathogenesis of AML. An example of the impact of these methods is the evidence garnered that AML developing from a translocation between chromosome 15 and 17 originates in a more differentiated progenitor than the stem cell. Figure 4 depicts the results from a patient with promyelocytic leukemia. CD34-positive/CD38-negative, primitive marrow progenitor cells, putative stem cells, and BFU-e did not contain the translocation and were polyclonal as judged by PCR analysis of an X chromosome gene polymorphism, whereas CD34-positive, CD38-positive and CFU-GM, more differentiated progenitors, contained the fusion oncogene and were monoclonal.12 In other studies this observation was supported by the introduction of a t(15;17) human transgene in mice which produced a disease mimicking human promyelocytic leukemia. The transgene was shown to have its effects principally in the neutrophilic lineage. The majority of cases of AML originate in a multipotential hematopoietic cell and some cases develop in a pluripotential stem cell.13,14,15 Based on such studies Figure 5 depicts the levels of hematopoiesis from which myelogenous leukemia may originate: the pluripotential stem cell, the multipotential progenitor cell, and the granulocytic progenitor cell compartments.
The putative progenitor cell leukemias have on average a higher response to therapy and a better prognosis than those cases that originate in a pluripotential stem cell or primitive multipotential hematopoietic cell. The reasons why the stem cell leukemias are generally more resistant to therapy than are progenitor cell leukemias have not been deciphered, although a higher expression of genes that mediate drug resistance in more primitive cells is one factor. Although as therapists, we view drug resistance as perverse, the evolution of such genes may reflect the protection they afford critical, small, self-renewing cell populations from injury by naturally occurring toxins.
In the late 1980s, several groups reported early success transplanting human hematopoietic cells into severe combined immunodeficient mice. These studies required massive infusions of cells to achieve engraftment in part because SCID mice retain natural killer cell activity and macrophage functions that limit transplantation of small numbers of human donor cells. The cross-breeding of non-obese diabetic with severe combined immunodeficient mouse strains resulted in sufficiently severe immunodeficiency that small numbers of human CD34-positive cells could engraft, reconstitute multilineage human hematopoiesis and B lymphopoiesis, and retain a high responsivity to human cytokines. The model was used to better understand the nature of leukemic hematopoiesis.16,17 It was found that various subtypes of human leukemia could be introduced into immunodeficient mice, especially with cytokine stimulation. Mimicry of the clinical features of the human leukemia was observed such as the highly infiltrative character of monocytic leukemia.
For at least 30 years, the hierarchical organization of AML had been presumed based on strong circumstantial evidence. This concept implied that the great bulk of leukemic blast cells had no or limited proliferative potential. Rather the disease is sustained by leukemic stem cells that feed cells into a compartment of proliferating and effete cells.17 To establish experimental evidence for this concept, an assay for a leukemic stem cell was needed that paralleled the assay for normal stem cells, that is one had to reconstitute the disease in vivo with one or a very few cells. Limiting dilution assays of the infused donor human leukemia cells found that there was a linear relationship between the number of cells infused into recipient mice and the probability of transplanting the human leukemia. The frequency of what has been dubbed the leukemia–initiating cell was 1/10 000 cells in human donor marrow and 1/250 000 blood cells. Moreover, the subfraction containing the leukemia-initiating cell is enriched with highly proliferating blast colony-forming cells that can differentiate into myeloid and lymphoid cells in vitro. Furthermore, CD34-positive/CD38-negative cells, putative stem cells, but not CD34-positive/CD38-positive cells could be used to establish human leukemia in mice. The frequency of leukemia-initiating cells in mice can be extrapolated to a large number of leukemic stem cells, perhaps about one million, in the marrow in human leukemia.
Normal and leukemic hematopoietic competition
The ramifications of a leukemic stem cell pool at the base of a hierarchy simulating normal hematopoiesis are several. A classic but presumably erroneous study by Furth and Kahn in 1937 reported that murine leukemia could be transplanted by inoculation of mice with a single leukemic cell, selected randomly.18 Thus, the requirement to kill the last leukemic cell was set as the goal of early therapists. With the appreciation of the total body burden of cells in acute leukemia and the kinetics of cytotoxic therapy, this task was daunting, perhaps impossible, because of drug-resistant subclones and host intolerance to continued intensive therapy. When chemotherapy was introduced after World War II, it was not known that normal stem cells were retained in the marrow in leukemia, although their presence is a prerequisite for the strategy of intensive cytotoxic therapy. In AML with cytogenetic abnormalities, the normal karyotype of dividing marrow cells during remission strongly suggested that normal hematopoiesis was restored, although the possibility that a karyotypically normal ‘leukemia minor’ clone was responsible for remission could not be excluded. The concept was that chemotherapy, having decreased the more disturbed subclones derived from the leukemic stem cell, may have permitted expression of the seminal leukemic clone without cytogenetic changes and with a greater degree of differentiation and maturation. The argument, with uncommon exceptions, was put to rest when the polyclonal nature of blood cell production of women, and presumably men, in remission was established. The presence of normal stem cells in the marrow in myelogenous leukemia is now established to wit their use for autologous stem cell rescue. Figure 6 depicts the circumstance of leukemia in relapse.19 Leukemic hematopoiesis dominates. It has been estimated that about one trillion cells are present in the patient with AML. Their presence suppresses normal hematopoiesis. During intensive cytotoxic therapy at least a three-log reduction in the cell population to one billion cells is sufficient to make leukemic cells inapparent in marrow and blood. In this setting in many patients such a decrease in leukemic cells or a larger one, permits the resumption of polyclonal hematopoiesis and a morphological remission. Thus, in the aftermath of a significant leukemic cell kill, normal stem cells gain hegemony for a time.
Since quantification of residual leukemic cells is not routinely applied in myelogenous leukemia, it had been difficult to be certain that the amount of reduction in leukemic cells is correlated with length of remission, although such a presumption is logical and generally accepted. Indeed, further cytotoxic therapy after remission is predicated on this assumption. In patients with AML in whom application of polymerase chain reaction or multiparametric flow cytometry immunophenotyping is used, evidence that extent of residual leukemia is correlated with the probability of relapse has become compelling.20,21,22
The twenty-first century: new opportunities for treatment
Two of the most important pathogenetic phenomena in myelogenous leukemia are the growth advantage of the somatically mutated leukemic multipotential cell and the concurrent suppression of the numerous normal stem cells in human marrow. Can the growth advantage of the leukemic stem cell be decreased by medical intervention so as to favor the normal stem cells? This of course happens to a modest proportion of patients treated with cytotoxic drugs. Remission ensues because polyclonal hematopoiesis is restored. In most cases this restoration is short-lived; the leukemic clone replaces normal hematopoiesis. Even in the less common circumstance in which normal hematopoiesis is sustained for years, it is very unlikely that the estimated one trillion leukemic cells and perhaps one million stem cells present have been annihilated. Moreover, we have little idea what produces quiescence in the leukemic stem cells. It is probable that in most cases of indefinite remission, a clinical not a biological cure has occurred. A new steady state develops as a result of significant reduction of the leukemic cell population such that leukemic cell reaccumulation does not occur. It seems improbable that solely endogenous immune cells mediate this symbiosis. This suggests a type of chemotherapeutic shock therapy.
Further exploration of the growth-sustaining cells in the human acute myelogenous leukemia hierarchy should result in targets for genetic or immune therapy so as to decrease the growth advantage of leukemic stem cells (Figure 7). Chemically enhancing the differentiation potential of normal stem cells is also theoretically possible, if there is some dissociation in the receptor array of normal as compared to leukemic stem cells. That is to say that receptor-ligand stimulators of normal stem cells might be used to foster their activation and alter the competition between normal and leukemic stem cells in favor of the former. A combination of these two approaches may be beneficial. Inducing normal differentiation of leukemic stem cells is problematic both conceptually and in practice.24
The factors released or expressed by leukemic cells that inhibit normal stem cell development should be identified with an eye to reversing their suppressive effects. The application of agents that may act to decrease leukemic progenitor cells by targeting their signal transduction pathways may contribute to release of the suppression of normal hematopoiesis. These are not easy research objectives but in the future fundamental leukemia research will focus on a better assessment of the behavior and possible vulnerability of both the earliest leukemic cells in the hierarchy and their progenitors. Such a multitargeted approach may be required to suppress the disease sufficiently to achieve clinical cures in most patients.
Maximow A . Der lymphozyt als gemeinsame Stammzelle der verschieden Blutelemente in der embryonalen Entwicklung und postfetalen Leben der Säugetiere Folia Haematologica (Leipzig) 1909 8: 125–141
Goodman JW, Hodgson GS . Evidence for stem cells in the peripheral blood of mice Blood 1962 19: 702–714
Danchakoff V . Origin of blood cells. Development of the haematopoitic organs and regeneration of the blood cells from the standpoint of the monophyletic school Anatomical Record 1916 10: 397–414
Kracke RR . Diseases of the Blood and Atlas of Hematology 2nd edn J.B. Lippincott Company: Philadelphia 1941
Lichtman MA (ed.) . Commentary on stem cell transplantation In Hematology: Landmark Papers of the Twentieth Century Academic Press: San Diego 2000 pp 669–670 and 715–717
Till JE, McCulloch EA . A direct measurement of the radiosensitivity of normal mouse bone marrow cells Radiat Res 1961 14: 213–222
Lichtman MA, Liesveld JN . Acute myelogenous leukemia In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U (eds) Williams Hematology 6th edn McGraw-Hill Book Company: New York 2001 pp 1047–1083
Linder D, Gartler SM . Glucose-6-phospate dehydrogenase mosaicism: utilization as a cell marker in studies of leiomyomas Science 1965 150: 67–69
Lichtman MA, Liesveld JL . Chronic myelogenous leukemia and related disorders (section on Pathogenesis; origin from a stem cell clone, p 1086) In Williams Hematology 6th edn McGraw-Hill Book Company: New York 2001 pp 1085–1123
Lichtman MA . Classification and clinical manifestations of the clonal myeloid disorders In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, Seligsohn U (eds) Williams Hematology 6th edn McGraw-Hill Book Company: New York 2001 pp 1019–1027
Lichtman, MA . Myelodysplasia or myeloneoplasia: thoughts on the nosology of clonal myeloid diseases Blood Cells, Mol Dis 2000 26: 572–581
Turhan AG, Lemoine FM, Debert C, Bonnet ML, Baillou C, Picard F, Macintyre EA, Varet B . Highly purified primitive hematopoietic cells are PML-RARα negative and generate non-clonal progenitors in acute promyelocytic leukemia Blood 1995 85: 2154–2161
Brendel C, Neubauer A . Characteristics and analysis of normal and leukemic stem cells.: current concepts and future directions Leukemia 2000 14: 1711–1717
Feuring-Buske M, Hogge D . Hoechst 33342 efflux identifies a subpopulation of CD34+CD38− cells from patients with acute myelogenous leukemia (AML) which lack chromosomal abnormalities and are enriched for colony forming cells (CFC) and long-term culture initiating cells (LTC-IC) Blood 1999 94: (Suppl. 1) 656a (Abstr.)
Wulf GG, Wang R-Y, Kuehnle I, Weidner D, Marini F, Brenner MK, Andreef M, Goodell MA . Leukemic involvement of the CD34 negative bone marrow side population (SP) in patients with AML Blood 1999 94: (Suppl. 1) (Abstr.)
Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, Caceres-Cortes J, Minden M, Paterson B, Caligiuri MA, Dick JE . A cell initiating human acute myeloid leukaemia after transplantation into SCID mice Nature 1994 367: 645–648
Bonnet D, Dick JE . Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell Nature Med 1997 3: 730–737
Furth J, Kahn MC . The transmission of leukemia in mice with a single cell Am J Cancer 1937 31: 276–282
Lichtman MA . Interrupting the inhibition of normal hematopoiesis in myelogenous leukemia: a hypothetical approach to therapy Stem Cells 2000 18: 304–306
Kim SC, Yoo NC, Hahn JS, Lee S, Chong SY, Min YH, Ko YM . Monitoring of the WT-1 gene in peripheral blood of patients with acute leukemia by semiquantitative RT-PCR; possible marker for detection of minimal residual leukemia Yonsei Med 1997 38: 212–219
Venditti A, Buccisano F, Del Poeta G, Maurillo L, Tamburini A, Cox, C, Battaglia A, Catalano G, Del Moro B, Cudillo L, Postorino M, Masi M, Amadori S . Level of minimal residual disease after consolidation therapy predicts outcome in acute myeloid leukemia Blood 2000 96: 3948–3952
Orfao A, Vidriales B, López-Berges, MC, San Miguel JF . Immunophenotypic investigation of MRD in AML: technical approaches and clinical impact Proc 2nd Intl Conf Minimal Residual Dis 2001 p 37, (Abstr.)
Russell NH . Biology of acute leukaemia Lancet 1997 349: 118–122
Lichtman MA . Differentiation versus maturation of neoplastic hematopoietic cells: an important distinction Blood Cells Mol Dis 2001 27: 659–662
This paper is part of a series of keynote addresses to be published in Leukemia. They were presented at the Acute Leukemia Forum, San Francisco, 20 April 2001. Supported by an unrestricted educational grant from Immunex.
About this article
Cite this article
Lichtman, M. The stem cell in the pathogenesis and treatment of myelogenous leukemia: a perspective. Leukemia 15, 1489–1494 (2001). https://doi.org/10.1038/sj.leu.2402247
- stem cells
- myelogenous leukemia
- clonal myeloid diseases