Medical Hypotheses MedialHypaem w-92) 35-, llo-ll8 0 LongmanGroupUK Ltd1992
The Origin and Spread of Human Leukemia A. ISLAM Division of Hematologic Oncology, Department of Medicine, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, New York 14263, USA
Abstract - The human leukemiasare a group of hematologicneoplasmscharacterizedby uncontrolled proliferationof cells concerned with blood cell production. The cause(s) of human leukemia remains unknown. Bone marrow (BM) is believedto be the site of origin of human leukemias, although the specific locus(i) and/or cell(s) from which it arises have not been definitively identified. Generally, human leukemiasand related proliferativediseases are thought to be clonal in nature; affecting a single hematopoieticstem cell, which then proliferatesand replaces the marrowof normal hematopoieticstem cell systems. The condition is believedto be malignant in nature. Results of our current morphologicstudies on well-fixed,ideally-stainedthin sections of plastic-embeddedbone marrowbiopsies (BMB)from a large numberof acute (AML, ALL) and chronic (CGL, CLL) leukemia patients suggestthat human leukemiasmay not be clonal diseases. Instead, a large population of other resident cells - ‘endosteal cells’ - appearsto become involved in the process and it is possiblethat all membersof this group enter the activity simultaneously. This change (transformation)in the endosteal cell population might be due to an abnormality (qualitativeor quantitative) of diffusable, humoralfactors (yet to be identified)that are responsible for the growth and proliferation of these hematopoieticprecursorcells. In this context, the human leukemias may be considered not as malignant,but rather the result of an aberration of factor(s) that control hematopoiesis. In this respect,the human leukemia%particularlyAML, ALL and CML, might be analogousto pernicious anemia (megaloblasticanemia) as it was understood 40-50 years ago. The major deficiency at present is our lack of knowledgeabout the humoralfactors and the activity of a different group of cells (endosteocytes),their relationshipto hematopoiesisand their role in causing human leukemias.
Introduction In recent years, there has been considerable progress in our knowledge and understanding of human leukemia. However, the morphologic identification of a cell(s) from which it arises remains indefinite. Date received 9 September 199 1 Date accepted 30 September 1991
Leukemias represent a segment of the hematologic neoplastic spectrum generally believed to affect hematopoietic stem cells (HSC). Although progress has been made in the differentiation of various hematopoietic progenitor cells (l-4), and functional assays have significantly promoted the study of 110
111
ORIGlNAM)SFREADOFHUMAN-
hematopoietic precursor cells (5,6,7), morphological investigationshave not been rewarding. Even after the application of enrichment techniques, the morphologic characterization of hematopoietic stem cells and colony forming cells has been restricted to hypothetical ‘candidate stem cells’ (8). The reason is that their precursor cell qualities cannot be deduced from their morphology, but can only be evaluated retrospectively and statistically from their function, i.e. in vitro colony formation or in vivo hematopoietic reconstitution from enriched cell suspensions (9). Our studies, which am based on morphologic observations in healthy human fetuses, normal adults and leukemia patients, on cellular changes observed in the BM of healthy human donors in which hematopoiesis had been mechanically perturbed (10, 1I), and on histologic study of cultured bony fragments in threedimensional collagen gels (12). have led us to hypothesize that the endosteal cells (EC), i.e. the cells normally lining the bony trabeculae, represent part of the undifferentiated mesenchyme and, when necessary, are able to ‘revert’ to their fetal phenotype, modulate and transform into primitive multipotential hematopoietic stem cells or their committed (e.g. granulocytic) precursors. Current evidence suggests that human leukemias are clonal in origin, beginning as single altered cells that proliferate and replace normal hematopoiesis. Contrary to this conventional belief, our morphologic evidence (the apparent origin, modulation and transformation of EC into leukemic blast cells) strongly suggests that the origin of leukemia is a multicellular event involving the entire EC population as ancestors of hematopoietic stem cells. Hematopoieticstem cells
Hematopoietic stem cells (HSC) are considered to be a self-renewing population of cells capable of giving rise to several progenitor cells, each serving as the origin of a different cell line (13). Methods used to study HSC have improved greatly and characterization of HSC has become more precise, particularly in the mouse model. However, much less is known about human HSC, although the availabledata strongly suggest a close parallelism with the murine system It is generally believed that a totipotential HSC capable of giving rise to all blood cells exists, and that certain progeny of this cell are also stem cells, but with increasing degrees of restriction of multi- potentiality. Thus, a totipotential HSC (primitive stem cell (CFU-L-M)) (14, 15.16) may give rise to two other
HSC, one of which is a lymphoid HSC and the other a myeloid HSC. The latter, in turn, may give rise to progenitor cells that am increasingly restricted in their potential to follow the different routes of differentiation in the myeloid system. This is evidenced by the existence of unipotent HSC that can form colonies of either erythroid (BFU-E, CFU-E) (17. 18), granulocytic (CFU-C) (19.20). or mega- karyocytic cells (CFU-M) (21.22). These are the most ‘mature’ or committed stem cells. Eosinophils and basophils may be derived from such unipotent stem cells, while neutrophils and monocytes are apparently derived from a bipotential progenitor. Clonalityof hwnan leukemia
The strongest evidence for the clonal origin of human leukemia is derived from G6PDisoenzyme studies and from cytogenetic data (23-26). Fialkow and coworkers (23,24) have studied four common types of human leukemia (ALL, AML, CLL, CML) in black female patients, and have provided evidence for a cfonal origin in each disease. CLL clonality is also suggested by studies of the surface immunoglobulin (Ig) of the cells involved in this condition. CLL is usually a neoplasm of B lymphocytes, a cell which normally expresses Ig on its surface. Normally, a B-cell population contains ahnost innumerable types of Ig but, in CLL, the B-cell populations usually contain an idiotypic immunoglobulin. This would be anticipated if the disease began in a specific precursor for the B cell, i.e. one of which was already committed to differentiationto yield progeny that produce a single common immune surface marker. Still other manifestations, such as the pattern of chromosomal abnormalities demonstrable in most leukemias, are also compatible with clonal disease (25.26). The trisomy 8 in CML, with its t&22) tramlocation. is not only the hallmark of this disease but also a prime indicator of its clonality. Site and cellular origin of human leukemia
Although it is known that BM is the site of origin of human leukemia, the specific locus within this BM from which it arises and the morphologic cellular initiators remain to be determined. It is generally believed that, in the leukemic process, an HSC (primitive or its descendant) is affected. In myeloid leukemias (AML, CML), there is evidence for the involvement of a myeloid HSC, while in lymphoid
112 leukemias (ALL, CLL), it is the lymphoid counterpart which is the aberrant cell. In cases when AML and ALL co-exist, it is most likely that a hierarchy of stem cell(s) capable of giving rise to both myeloid and lymphoid progenitors is affected. Most of this information is derived from G6PD and chromosomal studies. Current data show no evidence of theparticipation of myeloid cell lines in lymphoid leukemias (ALL, CLL), while in AML and CML, the myeloid leukemias, there is evidence for involvement of all of the progeny of a pluripotent myeloid HSC. In CML; G6PD analysis of isolated lymphocyte subpopulations indicates that some may be part of the clone, while others are not (14). In CIVIL,the thesis that it originates from the totipotential stem cell is further strengthened by the observation that either myeloid or lymphoid blast crises can arise, that ‘mixed’ blast crises can be observed, and that a given patient can have a lymphoid crisis, enter remission, and subsequently develop a myeloid crisis and vice versa (27).
A new concept of the site and cellular origin of human leukemia
MEDICAL HYPo.I-HE.sEs
that human leukemia appears in a region of the endosteum and originates from the endostcal cells. Supporting observations Acute MyeloblasticLeukemia (AML), Acute LymphoblasticLeukemia (ALL)
BM biopsy sections from newly-diagnosed AML and ALL patients revealed highly cellular marrows with dense infiltrations of similar appearing blast cells (Fig. la). The blast cells were concentrated and packed near the bony lrabeculae (BT). In some areas of the sections, waves of blast cells were seen protruding and budding out of the endosteal surface of the trabeculae in a sequential progression; suggesting an origin from endosteal progenitors (Fig. lb). In some sites, paratrabecular blast cells were also seen to migrate into intertrabecularmarrow space, where they entered into the circulation by way of the marrow sinusoids (Fig. lc). Chronic MyeloidLeukemia (CML)
Human leukemias m, usuallydiagnosed by examining dry film samples of peripheral blood (PB) and aspir- BM biopsy sections fromchronic phase CML patients ated bone marrow (BM). BM biopsies (histologic showed characteristic cellular proliferations adjacent sections) are not routinely used for this purpose. BM to bone spicules. Early granulopoieticcells (blastsand biopsies give an in situ view of a segment of hemato- promyelocytes) were the more numerous near the poietic tissue which, if appropriately fixed, processed trabeculae (Fig. 2a), while further advanced forms (plastic embedding) and properly stained (Roman- (myelocytes, metamyelocytes and segmented owsky type stain, e.g. May-Gmnwald-Giemsa stain), polymorphonuclear cells) were more numerous in the provide a wealth of information on the topography and central intertrabecular spaces (Fig. 2b). The overall cytoarchitectural details that cannot be obtained from histomorphologic appearance suggested that early a PB sample or BM aspirate, Yet, films, not sections, hematopoietic cells were originating (developing and remain the most universally accepted and widely used proliferating) in the paratrabecular region and then method of diagnosing and examining marrows in migrating with concmnitant maturation toward the human leukemia. The conventional rationalehas been central intertrabecular marrow space. BM biopsy that BM aspiration is easier to perform and Roman- specimens from CML patients in blastic transowsky stained smears of BM do provide excellent formation (BT) demonstrated cellular proliferations morphological preservation of hematopoietic cells very similar to those observed in AML and ALL. In CML-BT, as in the acute @ML, ALL) leukcmias, with which most hematologists are familiar. We have investigated the site and cellular origin of waves of blast cells were seen emerging from the human leukemia in a large number of acute @ML, endosteal BT surface in some areas of the sections ALL) and chronic (CML, CLL) leukemia patients (Fig. 2c). using an improved method of obtaining BM biopsies and embedding the specimens in plastic. These patients were studied at various stages of the disease Chronic LymphocyticLeukemia (CLL) - at diagnosis, following chemotherapy, at complete remission, during hemopoietic regenerationfollowing BM biopsy sections t?om CLL patients demonstrated chemotherapy, and at relapse. Our findings suggest 4 types of marrow infiltration (nodular, interstitial,
113
ORIGIN AND SPREAD OF HUhfAN LEUKEMIA
Ng la Bone marrow biopsy section from a newly-diagnosed patient with AML, showing dense infiltration appearing blast cells. Methylmethacrylste (MMA), May-Grunwald-Giemsa (MGG) stain
of morphologically
similar
Ng lb Bone maxrow biopsy section from a newly-diagnosed patient with AML showing part of a bone trabecula (BT) and demonstrating theoriginofblastcellsfromtheendosteum(arrows). Notethesequentialoutgmwthofendostealcells whichare baxxningpartofthe spectrum of leukemic blast cells. MMA. MGG stain
114
MEDICALHYPOTHESEs
Fig lc Bone marrow biopsy section from a newly-diagnosed patient with AML showing a marrow sinusoid surrounded by blast cells. A few blast cells that have already entered the sinusoidal lumen can also be seen. h4h4A, MGG stain
Fig. 2a Bone marmw biopsy section from a patient with Ch5L showing localization of early gramdopoietic cells (blasts and pmmyelocytcs) near a bone trabecula (BT). MMA, MGG stain
ORIGIN AND SPREAD OF HUMAN LEmMIA
115
J%. 2b Bone marmw biopsy section from the same case as in Figure 2a showing mostly mature granulocytic cells (myelocytes, mets .myelocytes and segmented polymorphonuclear cells) in the central intertrabecular marmw space. MMA, MGG stain
Ng 2e Bone marrow biopsy section from a Ch% patient in blast cell transformation showing the apparent origin of blast cells from the of endosteal cells from the endosteal surface of BT. MMA. MGG stain endosteum. Note the budding (outgrowth)
116
3 Bone marrow biopsy section from a patient with CLL showing apparent origin of the lymphoid cells from the endoste Glycol-methaccylate, MGG stain
mixed nodular and interstitial, and diffuse) by predominantly small, mature lymphocytes. In the first 3 types, the endosteal BM appeared benign and nonreactive, unless a lymphoid infiltrate was in intimate contact with the BT. In these cases, the endosteum sometimes appeared to be the site of origin of the lymphoid precursors. However, in the fourth type (diffuse), there were always some areas where morphologically recognizable lymphoid precursor cells could be seen ‘sprouting out’ from the endosteal surface of a BT, once again suggesting an origin from endosteal progenitors (Fig. 3).
al Pw3 !niton.
thyme. In the normal. steady-state, the great majority of them remain quiescent and, at a given time, only a representative few move away from their fixed endosteal position to enter the marrow cavity (intertrabecular marrow space). We proposed that they remain in this new locus in an uncommitted form as pluripotential stem cells (mobilized stem cells) or, if stimulated, differentiate into one of the hematopoieticpathways, depending on the demand and nature of the stimulus (signals) that they receive. One important feature of the resident endosteal cells is that, as long as they remain in their original locus, they probably maintain their fetal equivalent pluripotentiality.
Proposed cellular genesis of leukemia Conclusion Figure 4 illustrates the new hypothesis as to the site of origin of human leukemia and the cells involved in its genesis. It suggests that human leukemia originates at the endosteum and the cells that are primarily involved in the leukemic process are the endosteal cells. Trabecular bone surfaces are covered by a single layer of thin, inconspicuous, flattened, endothelium-like c&Is - bone lining cells or endosteun In fact, there are many millions of these cells and we believe that they are fixed hemoconnective tissue totipotential stem cells (10) descendant from embryonal mesen-
Human leukemia is a hematologic neoplasm that originates in the bone marrow (BM). No one has yet definitively demonstrated the locus in the BM where leukemia originates and what cell types are involved in the process. In this communication, we suggest that human leukemia arises from the endosteal cells near the trabecular surfaces of BM. Although this hypothesis is based solely on morphologic observations made on histologic sections, a medium in which it is difficult to distinguish between cell-cell interaction or
ORIGIN
AND SPREAD OF HUhIAN
117
LEUKEMIA
Origin of Human Leukemia from the Endosteum (BT=Bone Trabecula)
Migration and Spread to lntertrabecular Marrow Space and Replacement of All Normal Hematopoietk Cells
Systemic Spread by entering through the Marrow Sinusolds (S=Sinusoid)
Fig. 4 A schematic presentation of the proposed concept of human leukemia. According to this endosteal cells (equivalent of embxyonal stage undifferentiated mesenchyme) are the fmt to be affected in the genesis of the leukemic process. These cells modulate and transform into blast cells, migrate into the intertrabecular marrow space and gradually replace the normal hematopoietic elements. They then spill over from the marrow and enter the general circulation via the marrow
transformation of one cell type to another, the observed manner of cellular origin leads us to believe that endosteum is the site of genesis of human leukemia. Based primarily on G6PD andchromosomal studies, human leukernias have been considered to be of clonal origin. However, from our morphologic observations of the simultaneous outgrowth of a large number of hematopoietic precursor cells from the endosteum in the leukemic process, it seems apparent that not one, but a large number of cells are involved in the initiation of the disease. In light of these observations, we believe that the present concept of clonality in leukemia should be modified/re-evaluated, and that more emphasis should be given to finding factors that are responsible for the initiation (causation) of these diseases. While current medical treatment of leukemia as a malignant disease is appropriate for the present state of knowledge, one should also consider that leukemia, in general, may be very similar to megaloblastic anemia. At one time considered a fatal disease, it is now recognized as a deficiency of one ormore factors, the absence of which generates an imbalanced, unbridled hemoproliferative phenomenon.
Acknowledgments The author is greatly in debt to Dr E S Henderson and Dr C Glomski for critical review of the manuscript and to Mr Kevin Craig for his editorial assistance.
References Metcalf D, Johnson G R, Nicola N A. Separation and commitment of hemopoietic stem and progenitor cells. In Killmann S A, Cronkite E P, Muller-Berat C N, eds. Alfred Benson Symposium 18. Haemopoietic stem cells, characterizations, proliferation, regulation. Copenhagen: Munksgaard, 29-39, 1982. Morstyn G, Nicola N A, MetcalfD. Purification of hemopoietic progenitor cells from human marrow using a fucose-binding lectin and cell sorting. Blood 6: 798-805, 1980. Nicola N A, Metcalf D, Von Melchner H, Burgess A W. Isolation of murine fetal hemopoietic progenitor cells and selective fractionation of various erythmid precursors. Blood 58: 376-386.1981. Watt S M, Waldmann
D, Gilmore D J, Stenning G M, Clark M R, H. Selective isolation of murine erythropoietic-re-
Metcalf
118 sponsive progenitor cells (CFU-E) with monoclonal antibodies. Mol Biol Med 1: 95-115.1983. 5. van den Engh G, Bauman J. Muldcr D, Viser J. Measurement of antigen expression of hemopoietic stem cells and progenitor cells by fluorescence activated cell sorting. In: Kilbna~ S A, Cronkite E P, Muller-Berat C N, eds. Alfred Benson Symposium 18. Haemopoietic stem cells, charactcrixations, proliferation, regulation. Copenhagen: Munksgaard, 59-72,1982. 6. Goldschneider I, Metcalf D.Battye F, Mandel T. Analysis of rat hemopoietic cells on the fluorescence-activated cell sorter. I. Isolation of pluripotent hemopoietic stem cells and granulocyte-macrophage progenitor cells. J Exp Med 152: 419-437, 1980. 7. Metcalf D. Detection and analysis of human granulocyte-monocyte precursors using semi-solid cuhums. Clin Haemat 8: 263-285,1979. 8. Van Beldcum D W, van den Engh G, Wagemaker G, Bol S J L. Viser J W M. St~ctutal identity of the phuipotcnt hemopoietic stem cell. Blood Cells 5: 143-159,1979. 9. Thierfelder S, Hoffman-Fezer G. Immunohistochemical staining of hemopoietic prccunor cells. In: Killmann S A, Cronkite E P, Muller-Bemt C N, eds. Alfmd Benson Symposium 18, Haemopoietic stem cells, characterizations, proliferation, regulation. Copenhagen: Munksgaard. 75-84,1982. 10. Islam A. Haemopoietic stem cell: A new concept. Leuk Res 9: 1415-1432,198s. 11. Islam A, Glomski C, Henderson E S. Bone lining (endostcal) cells and hcmatopoiesis: A light microscopic study of normal and pathologic human bone marrow in plastic-embedded sections. Anat Res 227: 300-306.1990. 12. Islam A, Steiner R. Primary culture of marrow core in collagen g&z Modulation and transformation of endosteal cells. I. Morphologic observations. J Med 2Q 241-250.1989. 13. LaithaL G. Stemcellconcepts. Nouv RevFrHemat21: 59-65, 1979. 14. Boggs D R. Clonal origin of leukemia: site of origin in the stem cell hierarchy and the significance of chromosomal changes. Blood Cells 7: 205-215.1981. 15. Curry J L, Ttentin J J. Hemopoietic spleen colony studies. I.
h4ED1cAL BYFQTHFSES
Growth and dilferentiation. Dev Bioll5: 395-413,1967. 16. Curry J L. Trentin J J, Chcng U. Hemopoietic colony studies. III. Hemopoietic nature of spleen colonies induced by lymph node or thymus cells, with or without phytohemagghuinin. J Immunol99: 907-916,1967. 17. Curry J L, Trentin J J. Wolf N. Hemopoietic spleen colony studies. II. Erythropoiesis. J Exp Med 125: 703-719,1967. 18. Feldman M, Bleiberg I, Liron M. Regulation of intmsplcenic formation of erythtoid colonies. AM NY Acad Sci 129: 864-871,1966. 19. Ischove N N, Till J E. MaCulloch E A. The proliferative states of mouse granulopoietic progenitor cells. Pmc Sot Exp Biol Med 134: 33-36,197O. 20. Moom M A S, Williams N. Functional. morphologic, and kinetic analysis of the granulocyte-macrophage progenitor cell, hemopoiesis in culture. In: Robinson W A, ed. Second International Workshop (DHEW Publication No NIH 74-205). Government Printing Offhe. Washington, DC 17: 1974. 21. MetcalfD, MacDonald H R, Odartchenko N. Sordat B. Growth of mouse megakatyocyte colonies in vitro. Proc Nat1Acad Sci USA. 72: 1744-1748.1975. 22. Nakeff A, Dicke K A. van Noon-l M J. Megakaryocytes in agar culture of mouse bone marrow. Scr Haemat 8: 4-21,197s. 23. Fialkow P J, Singer J W. Adamson J W, et al. Acute non-lymphocytic leukemia: heterogeneity of stem cell origin. Blood 57: 1068-1073.1981. 24. Fialkow P J, Singer J W, Raskind W H. et al. Clonal develop met& stem cell differentiation, and clinical remission in acute nonlymphocytic leukemia. N Eng J Mcd 317: 468-473,1987. 25. The Fourth International Workshop on Chromosomes in Lcukemia: A prospective study of acute nonlymphocytic leukemia. Chicago, Illinois, USA. September 2-7,1982. Cancer Genet Cytogenet. 11: 249-360.1984. 26. Keinanen M, GrifEm J D, Bloomtield C D, et al. Clonal chromosomal abnormalities showing multiple-cell-lineage involvement in acute myeloid leukemia. N Eng J Med 3 18: 1153-I 158,1988. 27. Wintmbe M M, Lee G R, Boggs D R, et al. Clinical hematology. Philadelphia: Lea and Febiger, 1981.
The Origin and Spread of Human Leukemia A. ISLAM Division of Hematologic Oncology, Department of Medicine, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, New York 14263, USA
Abstract - The human leukemiasare a group of hematologicneoplasmscharacterizedby uncontrolled proliferationof cells concerned with blood cell production. The cause(s) of human leukemia remains unknown. Bone marrow (BM) is believedto be the site of origin of human leukemias, although the specific locus(i) and/or cell(s) from which it arises have not been definitively identified. Generally, human leukemiasand related proliferativediseases are thought to be clonal in nature; affecting a single hematopoieticstem cell, which then proliferatesand replaces the marrowof normal hematopoieticstem cell systems. The condition is believedto be malignant in nature. Results of our current morphologicstudies on well-fixed,ideally-stainedthin sections of plastic-embeddedbone marrowbiopsies (BMB)from a large numberof acute (AML, ALL) and chronic (CGL, CLL) leukemia patients suggestthat human leukemiasmay not be clonal diseases. Instead, a large population of other resident cells - ‘endosteal cells’ - appearsto become involved in the process and it is possiblethat all membersof this group enter the activity simultaneously. This change (transformation)in the endosteal cell population might be due to an abnormality (qualitativeor quantitative) of diffusable, humoralfactors (yet to be identified)that are responsible for the growth and proliferation of these hematopoieticprecursorcells. In this context, the human leukemias may be considered not as malignant,but rather the result of an aberration of factor(s) that control hematopoiesis. In this respect,the human leukemia%particularlyAML, ALL and CML, might be analogousto pernicious anemia (megaloblasticanemia) as it was understood 40-50 years ago. The major deficiency at present is our lack of knowledgeabout the humoralfactors and the activity of a different group of cells (endosteocytes),their relationshipto hematopoiesisand their role in causing human leukemias.
Introduction In recent years, there has been considerable progress in our knowledge and understanding of human leukemia. However, the morphologic identification of a cell(s) from which it arises remains indefinite. Date received 9 September 199 1 Date accepted 30 September 1991
Leukemias represent a segment of the hematologic neoplastic spectrum generally believed to affect hematopoietic stem cells (HSC). Although progress has been made in the differentiation of various hematopoietic progenitor cells (l-4), and functional assays have significantly promoted the study of 110
111
ORIGlNAM)SFREADOFHUMAN-
hematopoietic precursor cells (5,6,7), morphological investigationshave not been rewarding. Even after the application of enrichment techniques, the morphologic characterization of hematopoietic stem cells and colony forming cells has been restricted to hypothetical ‘candidate stem cells’ (8). The reason is that their precursor cell qualities cannot be deduced from their morphology, but can only be evaluated retrospectively and statistically from their function, i.e. in vitro colony formation or in vivo hematopoietic reconstitution from enriched cell suspensions (9). Our studies, which am based on morphologic observations in healthy human fetuses, normal adults and leukemia patients, on cellular changes observed in the BM of healthy human donors in which hematopoiesis had been mechanically perturbed (10, 1I), and on histologic study of cultured bony fragments in threedimensional collagen gels (12). have led us to hypothesize that the endosteal cells (EC), i.e. the cells normally lining the bony trabeculae, represent part of the undifferentiated mesenchyme and, when necessary, are able to ‘revert’ to their fetal phenotype, modulate and transform into primitive multipotential hematopoietic stem cells or their committed (e.g. granulocytic) precursors. Current evidence suggests that human leukemias are clonal in origin, beginning as single altered cells that proliferate and replace normal hematopoiesis. Contrary to this conventional belief, our morphologic evidence (the apparent origin, modulation and transformation of EC into leukemic blast cells) strongly suggests that the origin of leukemia is a multicellular event involving the entire EC population as ancestors of hematopoietic stem cells. Hematopoieticstem cells
Hematopoietic stem cells (HSC) are considered to be a self-renewing population of cells capable of giving rise to several progenitor cells, each serving as the origin of a different cell line (13). Methods used to study HSC have improved greatly and characterization of HSC has become more precise, particularly in the mouse model. However, much less is known about human HSC, although the availabledata strongly suggest a close parallelism with the murine system It is generally believed that a totipotential HSC capable of giving rise to all blood cells exists, and that certain progeny of this cell are also stem cells, but with increasing degrees of restriction of multi- potentiality. Thus, a totipotential HSC (primitive stem cell (CFU-L-M)) (14, 15.16) may give rise to two other
HSC, one of which is a lymphoid HSC and the other a myeloid HSC. The latter, in turn, may give rise to progenitor cells that am increasingly restricted in their potential to follow the different routes of differentiation in the myeloid system. This is evidenced by the existence of unipotent HSC that can form colonies of either erythroid (BFU-E, CFU-E) (17. 18), granulocytic (CFU-C) (19.20). or mega- karyocytic cells (CFU-M) (21.22). These are the most ‘mature’ or committed stem cells. Eosinophils and basophils may be derived from such unipotent stem cells, while neutrophils and monocytes are apparently derived from a bipotential progenitor. Clonalityof hwnan leukemia
The strongest evidence for the clonal origin of human leukemia is derived from G6PDisoenzyme studies and from cytogenetic data (23-26). Fialkow and coworkers (23,24) have studied four common types of human leukemia (ALL, AML, CLL, CML) in black female patients, and have provided evidence for a cfonal origin in each disease. CLL clonality is also suggested by studies of the surface immunoglobulin (Ig) of the cells involved in this condition. CLL is usually a neoplasm of B lymphocytes, a cell which normally expresses Ig on its surface. Normally, a B-cell population contains ahnost innumerable types of Ig but, in CLL, the B-cell populations usually contain an idiotypic immunoglobulin. This would be anticipated if the disease began in a specific precursor for the B cell, i.e. one of which was already committed to differentiationto yield progeny that produce a single common immune surface marker. Still other manifestations, such as the pattern of chromosomal abnormalities demonstrable in most leukemias, are also compatible with clonal disease (25.26). The trisomy 8 in CML, with its t&22) tramlocation. is not only the hallmark of this disease but also a prime indicator of its clonality. Site and cellular origin of human leukemia
Although it is known that BM is the site of origin of human leukemia, the specific locus within this BM from which it arises and the morphologic cellular initiators remain to be determined. It is generally believed that, in the leukemic process, an HSC (primitive or its descendant) is affected. In myeloid leukemias (AML, CML), there is evidence for the involvement of a myeloid HSC, while in lymphoid
112 leukemias (ALL, CLL), it is the lymphoid counterpart which is the aberrant cell. In cases when AML and ALL co-exist, it is most likely that a hierarchy of stem cell(s) capable of giving rise to both myeloid and lymphoid progenitors is affected. Most of this information is derived from G6PD and chromosomal studies. Current data show no evidence of theparticipation of myeloid cell lines in lymphoid leukemias (ALL, CLL), while in AML and CML, the myeloid leukemias, there is evidence for involvement of all of the progeny of a pluripotent myeloid HSC. In CML; G6PD analysis of isolated lymphocyte subpopulations indicates that some may be part of the clone, while others are not (14). In CIVIL,the thesis that it originates from the totipotential stem cell is further strengthened by the observation that either myeloid or lymphoid blast crises can arise, that ‘mixed’ blast crises can be observed, and that a given patient can have a lymphoid crisis, enter remission, and subsequently develop a myeloid crisis and vice versa (27).
A new concept of the site and cellular origin of human leukemia
MEDICAL HYPo.I-HE.sEs
that human leukemia appears in a region of the endosteum and originates from the endostcal cells. Supporting observations Acute MyeloblasticLeukemia (AML), Acute LymphoblasticLeukemia (ALL)
BM biopsy sections from newly-diagnosed AML and ALL patients revealed highly cellular marrows with dense infiltrations of similar appearing blast cells (Fig. la). The blast cells were concentrated and packed near the bony lrabeculae (BT). In some areas of the sections, waves of blast cells were seen protruding and budding out of the endosteal surface of the trabeculae in a sequential progression; suggesting an origin from endosteal progenitors (Fig. lb). In some sites, paratrabecular blast cells were also seen to migrate into intertrabecularmarrow space, where they entered into the circulation by way of the marrow sinusoids (Fig. lc). Chronic MyeloidLeukemia (CML)
Human leukemias m, usuallydiagnosed by examining dry film samples of peripheral blood (PB) and aspir- BM biopsy sections fromchronic phase CML patients ated bone marrow (BM). BM biopsies (histologic showed characteristic cellular proliferations adjacent sections) are not routinely used for this purpose. BM to bone spicules. Early granulopoieticcells (blastsand biopsies give an in situ view of a segment of hemato- promyelocytes) were the more numerous near the poietic tissue which, if appropriately fixed, processed trabeculae (Fig. 2a), while further advanced forms (plastic embedding) and properly stained (Roman- (myelocytes, metamyelocytes and segmented owsky type stain, e.g. May-Gmnwald-Giemsa stain), polymorphonuclear cells) were more numerous in the provide a wealth of information on the topography and central intertrabecular spaces (Fig. 2b). The overall cytoarchitectural details that cannot be obtained from histomorphologic appearance suggested that early a PB sample or BM aspirate, Yet, films, not sections, hematopoietic cells were originating (developing and remain the most universally accepted and widely used proliferating) in the paratrabecular region and then method of diagnosing and examining marrows in migrating with concmnitant maturation toward the human leukemia. The conventional rationalehas been central intertrabecular marrow space. BM biopsy that BM aspiration is easier to perform and Roman- specimens from CML patients in blastic transowsky stained smears of BM do provide excellent formation (BT) demonstrated cellular proliferations morphological preservation of hematopoietic cells very similar to those observed in AML and ALL. In CML-BT, as in the acute @ML, ALL) leukcmias, with which most hematologists are familiar. We have investigated the site and cellular origin of waves of blast cells were seen emerging from the human leukemia in a large number of acute @ML, endosteal BT surface in some areas of the sections ALL) and chronic (CML, CLL) leukemia patients (Fig. 2c). using an improved method of obtaining BM biopsies and embedding the specimens in plastic. These patients were studied at various stages of the disease Chronic LymphocyticLeukemia (CLL) - at diagnosis, following chemotherapy, at complete remission, during hemopoietic regenerationfollowing BM biopsy sections t?om CLL patients demonstrated chemotherapy, and at relapse. Our findings suggest 4 types of marrow infiltration (nodular, interstitial,
113
ORIGIN AND SPREAD OF HUhfAN LEUKEMIA
Ng la Bone marrow biopsy section from a newly-diagnosed patient with AML, showing dense infiltration appearing blast cells. Methylmethacrylste (MMA), May-Grunwald-Giemsa (MGG) stain
of morphologically
similar
Ng lb Bone maxrow biopsy section from a newly-diagnosed patient with AML showing part of a bone trabecula (BT) and demonstrating theoriginofblastcellsfromtheendosteum(arrows). Notethesequentialoutgmwthofendostealcells whichare baxxningpartofthe spectrum of leukemic blast cells. MMA. MGG stain
114
MEDICALHYPOTHESEs
Fig lc Bone marrow biopsy section from a newly-diagnosed patient with AML showing a marrow sinusoid surrounded by blast cells. A few blast cells that have already entered the sinusoidal lumen can also be seen. h4h4A, MGG stain
Fig. 2a Bone marmw biopsy section from a patient with Ch5L showing localization of early gramdopoietic cells (blasts and pmmyelocytcs) near a bone trabecula (BT). MMA, MGG stain
ORIGIN AND SPREAD OF HUMAN LEmMIA
115
J%. 2b Bone marmw biopsy section from the same case as in Figure 2a showing mostly mature granulocytic cells (myelocytes, mets .myelocytes and segmented polymorphonuclear cells) in the central intertrabecular marmw space. MMA, MGG stain
Ng 2e Bone marrow biopsy section from a Ch% patient in blast cell transformation showing the apparent origin of blast cells from the of endosteal cells from the endosteal surface of BT. MMA. MGG stain endosteum. Note the budding (outgrowth)
116
3 Bone marrow biopsy section from a patient with CLL showing apparent origin of the lymphoid cells from the endoste Glycol-methaccylate, MGG stain
mixed nodular and interstitial, and diffuse) by predominantly small, mature lymphocytes. In the first 3 types, the endosteal BM appeared benign and nonreactive, unless a lymphoid infiltrate was in intimate contact with the BT. In these cases, the endosteum sometimes appeared to be the site of origin of the lymphoid precursors. However, in the fourth type (diffuse), there were always some areas where morphologically recognizable lymphoid precursor cells could be seen ‘sprouting out’ from the endosteal surface of a BT, once again suggesting an origin from endosteal progenitors (Fig. 3).
al Pw3 !niton.
thyme. In the normal. steady-state, the great majority of them remain quiescent and, at a given time, only a representative few move away from their fixed endosteal position to enter the marrow cavity (intertrabecular marrow space). We proposed that they remain in this new locus in an uncommitted form as pluripotential stem cells (mobilized stem cells) or, if stimulated, differentiate into one of the hematopoieticpathways, depending on the demand and nature of the stimulus (signals) that they receive. One important feature of the resident endosteal cells is that, as long as they remain in their original locus, they probably maintain their fetal equivalent pluripotentiality.
Proposed cellular genesis of leukemia Conclusion Figure 4 illustrates the new hypothesis as to the site of origin of human leukemia and the cells involved in its genesis. It suggests that human leukemia originates at the endosteum and the cells that are primarily involved in the leukemic process are the endosteal cells. Trabecular bone surfaces are covered by a single layer of thin, inconspicuous, flattened, endothelium-like c&Is - bone lining cells or endosteun In fact, there are many millions of these cells and we believe that they are fixed hemoconnective tissue totipotential stem cells (10) descendant from embryonal mesen-
Human leukemia is a hematologic neoplasm that originates in the bone marrow (BM). No one has yet definitively demonstrated the locus in the BM where leukemia originates and what cell types are involved in the process. In this communication, we suggest that human leukemia arises from the endosteal cells near the trabecular surfaces of BM. Although this hypothesis is based solely on morphologic observations made on histologic sections, a medium in which it is difficult to distinguish between cell-cell interaction or
ORIGIN
AND SPREAD OF HUhIAN
117
LEUKEMIA
Origin of Human Leukemia from the Endosteum (BT=Bone Trabecula)
Migration and Spread to lntertrabecular Marrow Space and Replacement of All Normal Hematopoietk Cells
Systemic Spread by entering through the Marrow Sinusolds (S=Sinusoid)
Fig. 4 A schematic presentation of the proposed concept of human leukemia. According to this endosteal cells (equivalent of embxyonal stage undifferentiated mesenchyme) are the fmt to be affected in the genesis of the leukemic process. These cells modulate and transform into blast cells, migrate into the intertrabecular marrow space and gradually replace the normal hematopoietic elements. They then spill over from the marrow and enter the general circulation via the marrow
transformation of one cell type to another, the observed manner of cellular origin leads us to believe that endosteum is the site of genesis of human leukemia. Based primarily on G6PD andchromosomal studies, human leukernias have been considered to be of clonal origin. However, from our morphologic observations of the simultaneous outgrowth of a large number of hematopoietic precursor cells from the endosteum in the leukemic process, it seems apparent that not one, but a large number of cells are involved in the initiation of the disease. In light of these observations, we believe that the present concept of clonality in leukemia should be modified/re-evaluated, and that more emphasis should be given to finding factors that are responsible for the initiation (causation) of these diseases. While current medical treatment of leukemia as a malignant disease is appropriate for the present state of knowledge, one should also consider that leukemia, in general, may be very similar to megaloblastic anemia. At one time considered a fatal disease, it is now recognized as a deficiency of one ormore factors, the absence of which generates an imbalanced, unbridled hemoproliferative phenomenon.
Acknowledgments The author is greatly in debt to Dr E S Henderson and Dr C Glomski for critical review of the manuscript and to Mr Kevin Craig for his editorial assistance.
References Metcalf D, Johnson G R, Nicola N A. Separation and commitment of hemopoietic stem and progenitor cells. In Killmann S A, Cronkite E P, Muller-Berat C N, eds. Alfred Benson Symposium 18. Haemopoietic stem cells, characterizations, proliferation, regulation. Copenhagen: Munksgaard, 29-39, 1982. Morstyn G, Nicola N A, MetcalfD. Purification of hemopoietic progenitor cells from human marrow using a fucose-binding lectin and cell sorting. Blood 6: 798-805, 1980. Nicola N A, Metcalf D, Von Melchner H, Burgess A W. Isolation of murine fetal hemopoietic progenitor cells and selective fractionation of various erythmid precursors. Blood 58: 376-386.1981. Watt S M, Waldmann
D, Gilmore D J, Stenning G M, Clark M R, H. Selective isolation of murine erythropoietic-re-
Metcalf
118 sponsive progenitor cells (CFU-E) with monoclonal antibodies. Mol Biol Med 1: 95-115.1983. 5. van den Engh G, Bauman J. Muldcr D, Viser J. Measurement of antigen expression of hemopoietic stem cells and progenitor cells by fluorescence activated cell sorting. In: Kilbna~ S A, Cronkite E P, Muller-Berat C N, eds. Alfred Benson Symposium 18. Haemopoietic stem cells, charactcrixations, proliferation, regulation. Copenhagen: Munksgaard, 59-72,1982. 6. Goldschneider I, Metcalf D.Battye F, Mandel T. Analysis of rat hemopoietic cells on the fluorescence-activated cell sorter. I. Isolation of pluripotent hemopoietic stem cells and granulocyte-macrophage progenitor cells. J Exp Med 152: 419-437, 1980. 7. Metcalf D. Detection and analysis of human granulocyte-monocyte precursors using semi-solid cuhums. Clin Haemat 8: 263-285,1979. 8. Van Beldcum D W, van den Engh G, Wagemaker G, Bol S J L. Viser J W M. St~ctutal identity of the phuipotcnt hemopoietic stem cell. Blood Cells 5: 143-159,1979. 9. Thierfelder S, Hoffman-Fezer G. Immunohistochemical staining of hemopoietic prccunor cells. In: Killmann S A, Cronkite E P, Muller-Bemt C N, eds. Alfmd Benson Symposium 18, Haemopoietic stem cells, characterizations, proliferation, regulation. Copenhagen: Munksgaard. 75-84,1982. 10. Islam A. Haemopoietic stem cell: A new concept. Leuk Res 9: 1415-1432,198s. 11. Islam A, Glomski C, Henderson E S. Bone lining (endostcal) cells and hcmatopoiesis: A light microscopic study of normal and pathologic human bone marrow in plastic-embedded sections. Anat Res 227: 300-306.1990. 12. Islam A, Steiner R. Primary culture of marrow core in collagen g&z Modulation and transformation of endosteal cells. I. Morphologic observations. J Med 2Q 241-250.1989. 13. LaithaL G. Stemcellconcepts. Nouv RevFrHemat21: 59-65, 1979. 14. Boggs D R. Clonal origin of leukemia: site of origin in the stem cell hierarchy and the significance of chromosomal changes. Blood Cells 7: 205-215.1981. 15. Curry J L, Ttentin J J. Hemopoietic spleen colony studies. I.
h4ED1cAL BYFQTHFSES
Growth and dilferentiation. Dev Bioll5: 395-413,1967. 16. Curry J L. Trentin J J, Chcng U. Hemopoietic colony studies. III. Hemopoietic nature of spleen colonies induced by lymph node or thymus cells, with or without phytohemagghuinin. J Immunol99: 907-916,1967. 17. Curry J L, Trentin J J. Wolf N. Hemopoietic spleen colony studies. II. Erythropoiesis. J Exp Med 125: 703-719,1967. 18. Feldman M, Bleiberg I, Liron M. Regulation of intmsplcenic formation of erythtoid colonies. AM NY Acad Sci 129: 864-871,1966. 19. Ischove N N, Till J E. MaCulloch E A. The proliferative states of mouse granulopoietic progenitor cells. Pmc Sot Exp Biol Med 134: 33-36,197O. 20. Moom M A S, Williams N. Functional. morphologic, and kinetic analysis of the granulocyte-macrophage progenitor cell, hemopoiesis in culture. In: Robinson W A, ed. Second International Workshop (DHEW Publication No NIH 74-205). Government Printing Offhe. Washington, DC 17: 1974. 21. MetcalfD, MacDonald H R, Odartchenko N. Sordat B. Growth of mouse megakatyocyte colonies in vitro. Proc Nat1Acad Sci USA. 72: 1744-1748.1975. 22. Nakeff A, Dicke K A. van Noon-l M J. Megakaryocytes in agar culture of mouse bone marrow. Scr Haemat 8: 4-21,197s. 23. Fialkow P J, Singer J W. Adamson J W, et al. Acute non-lymphocytic leukemia: heterogeneity of stem cell origin. Blood 57: 1068-1073.1981. 24. Fialkow P J, Singer J W, Raskind W H. et al. Clonal develop met& stem cell differentiation, and clinical remission in acute nonlymphocytic leukemia. N Eng J Mcd 317: 468-473,1987. 25. The Fourth International Workshop on Chromosomes in Lcukemia: A prospective study of acute nonlymphocytic leukemia. Chicago, Illinois, USA. September 2-7,1982. Cancer Genet Cytogenet. 11: 249-360.1984. 26. Keinanen M, GrifEm J D, Bloomtield C D, et al. Clonal chromosomal abnormalities showing multiple-cell-lineage involvement in acute myeloid leukemia. N Eng J Med 3 18: 1153-I 158,1988. 27. Wintmbe M M, Lee G R, Boggs D R, et al. Clinical hematology. Philadelphia: Lea and Febiger, 1981.
