Cancer and Meta.sttuis Reviews 8: 253-262, 1989.

(~) 1989 Kluwer Academic Publishers. Printed in the Netherlattds.

The role of growth factors in haemopoietic development: Clinical and biological implications C. Paul Daniel and T. Michael Dexter

Department of Experimental Haematology, Paterson bzstitutefor Cancer Research, Christie Hospital, Manchester, U.K.

Key words: growth factors, haemopoiesis, leukaemia, stem cells, differentiation Abstract

Mature blood cells of all lineages are derived from a single class of cell, the haemopoietic stem cell. Stem cells are pluripotent and capable of almost limitless self-renewal. In the bone marrow they form part of a hierarchy that includes progenitor cells, which are more restricted in the lineages their progeny can adopt, and precursor cells, which are committed to differentiation. The mechanisms that regulate progression through this hierarchy are not fully understood, but evidence suggests that both bone marrow stromal cells and soluble growth factors have a role in controlling haemopoiesis. Four growth factors act on progenitor cells to promote their survival, proliferation, differentiation, and maturation: interleukin-3 (IL-3), granulocyte/macrophage-colony stimulating factor (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF). They can also activate the function of mature cells. Considerable overlap is found in the target cells for these four growth factors. We have found that growth factors acting in syncrgy can recruit more primitive cells than had previously been appreciated. These factors can also determine the lineage that the progeny of multipotential progenitors will adopt. Thus, colonystimulating factors (CSFs) have the potential to regulate the development of primitive haemopoietic cells in vivo. The properties of CSFs have made them useful in treating malignant disease: G-CSF, in particular, has been used to reduce the period of neutropaenia that follows cytotoxic therapy for various malignancies. The success of these early trials gives ground for cautious optimism about the clinical use of these compounds.

Introduction

Normal haemopoiesis fulfills two important functions: It maintains a dynamic equilibrium, constantly replacing mature blood cells lost through apoptosis or damage, and it responds to acute challenges such as infection and haemorrhage. In recent years, there has been considerable research into the mechanisms that regulate these processes. Understanding these mechanisms is important not only because their malfunction is implicated in the aetiology of diseases such as leukaemia and myelodysplastic syndrome, but also for the potential

therapeutic benefits that being able to manipulate the production of mature blood cells would bring. Although complete understanding of the haemopoietic system remains elusive, useful information has emerged from these investigations. Recently, a number of soluble factors have been characterised that have potent effects on blood cells. In the course of this review we shall argue that, while their function in the regulation of steady-state haemopoiesis in the bone marrow is unclear, the observed effects of these growth factors on the growth, development, and activation of haemopoietic cells has important conceptual and therapeutic implications.

254

The haemopoietic system The phenoptypes of mature blood cells are diverse, ranging from enucleate erythrocytes to multinucleate megakaryocytes. Despite this heterogeneity, experimental evidence suggests that not only myeloid cells, which include granulocytes (neutrophils, eosinphils, and basophils), monocytes, platelets, and erythrocytes, but also T and B lymphocytes derive from a single class of cell called the haemopoietic stem cell. In adults the stem cell resides in the bone marrow. This is known because the entire haemopoietic systerfi of a mouse in which endogenous haemopoiesis has been ablated may be reconstituted by transplanting marrow from a normal animal [1]. In experiments in which random chromosomal markers were generated in the donor marrow either by radiation-induced mutation or retroviral insertion, it was shown that reconstitution could take place from a single cell [2-4]. In fact, bone marrow can be serially transplanted through several generations of mice, demonstrating the extensive self-renewal capability of the haemopoietic stem cell [5]. Stem cells, therefore, have two important properties: Their progeny can differentiate into "all the different types of mature haemopoietic cells, but the stem cell population itself is maintained (or expanded) by self-renewal. Transplanting bone marrow cells from normal to potentially lethally irradiated mice also gives rise to colonies of haemopoietic cells in the spleen of the recipients [6]. These colonies are derived from a single cell designated the colony forming unitspleen (CFU-S). They contain not only cells of several haemopoietic lineages, but also CFU-S, which are capable of giving rise to more spleen colonies. Recent work has shown that CFU-S are heterogenous and contain a continuum of cells with different self-renewal capacities [7]. They are not, therefore, identical with stem cells, although they provide a convenient assay for primitive haemopoietic populations. Also present in the bone marrow are cells which, although capable of extensive proliferation, have a much reduced capacity for self-renewal and are no longer pluripotent [8]. These progenitor cells are characterised by their ability to form colonies of

morphologically recognisable mature cells in semisolid medium. Such cells are capable of rapid expansion, giving rise to colonies of several thousand cells in a few days. These colonies, unlike spleen colonies, however, do not usually contain more colony-forming cells. Thus progenitor cells are committed to differentiation rather than self-renewal. The lineages along which progenitor cells can differentiate are restricted. CFC-mix can give rise to colonies containing granulocytic cells, monocytes, megakaryocytes, and erythrocytes but not, apparently, T or B cells. In this respect, therefore, CFC-mix are similar to some CFU-S. More restricted still are GM-CFC, which produce colonies consisting of both neutrophils and monocytes, while colonies derived from G-CFC and M-CFC contain only neutrophils or monocytes respectively. Other lineage-restricted progenitors have been described that give rise to eosinophii (eos-CFC), erythroid (BFU-e), or megakaryocyte (CFC-meg) colonies. The haemopoietic system, therefore, consists of a hierarchy of cells, from the pluripotent, self-renewing stem cell through committed progenitor cells that are increasingly lineage restricted, to the mature, terminally differentiated end cell. This hierarchy accomplishes both the massive expansion of cell numbers required to replace senescent cells as well as their simultaneous differentiation. The mechanisms that control haemopoiesis, therefore, must control the progress of cells through this hierarchy, determining whether they undergo selfrenewal or differentiation, and which lineage they adopt. In vitro, a central role in this process is played by a class of factors called colony stimulating factors (CSFs).

Properties of colony stimulating factors Although many substances influence haemopoietic cells, CSFs are recognised by their ability to promote the survival, proliferation, differentiation, and activation of function of these cells. Four major CSFs have been purified, molecularly cloned, and sequenced: interleukin-3 (1L-3), granulocyte macrophage-CSF (GM-CSF), granulocyte-CSF (G-CSF), and macrophage-CSF (M-CSF) [9-12].

255 All are glycoproteins, although glycosylation does not appear to be essential to their function. M-CSF is a homodimer; the others are monomers with disulphide bridges [13]. The factors are not related (there is little homolo~, between their sequences), but the genes for three of them (IL-3, GM-CSF and M-CSF) are located in the q23--q33 region of chromosome 5 [14-16] in humans, while c-fms, thc gene that codes for the M-CSF receptor [15], is also located on the long arm of this chromosome. The functional significance of this clustering is unclear, since coordinate expression does not apparently occur in T cell clones capable of .secreting several factors [17]. However, 5q- is a chromosome abnormality as~)ciated with haemopoietic disorders [18]. Although there is overlap in their functions, the four CSFs fit into the hierarchy of progenitor cells already described. Thus, IL-3 stimulates the growth of CFC-mix in vitro [19], but can also increase the survival of CFU-S [20], and is essential for the growth of some haemopoietic cell lines that resemble stem cells [21]. IL-3, therefore, appears to act on the primitive, pluripotent cells in the hacmopoietic hierarchy, but does not determine the lineage they adopt. GM-CSF acts on cells that are more restricted in their potential lineages, promoting the growth of GM-CFC as well as granulocyte and macrophage progenitors [22], while GCSF and M-CSF act on these latter cells respectively [23]. While colony stimulating acti~4ty provides a convenient way of characterising these factors, it does not completely describe the spectrum of their activity. IL-3, for example, can act on a wide range of cell types, stimulating the growth of GM-CFC and primitive crythroid cells as well as eosinophil, basophil, and megakaryocyte progenitors. IL-3 can also promote haemoglobinisation of a subset of erythroid cells, a property previously thought to be unique to erythropoietin [23]. Cells that respond to GM-CSF include multipotent blast cells and early erythroid progenitors and, at high concentrations, the factor will support the growth of megakaryocyte colonies [24-26]. The activity of G-CSF overlaps with that of GM-CSF to the extent that some GM-CFC will respond to high concentrations of the factor [23].

These examples demonstrate that, while the activities of the CSFs overlap, they may still be ranked according to the range of target cells that respond to them. The greatest range of cell types respond to IL-3, which also stimulates the most primitive cells. GM-CSF and M-CSF stimulate progressively smaller groups of cells with greater lineage rcstriction.

Synergistic effects of colony-stimulating factors While it is useful to investigate the effects of individual CSFs acting alone on haemopoietic cells, in reality these cells exist in a far more complex environment. Haemopoietic cells in vivo are more likely to be exposed to combinations of CSFs and other factors. In consequence, it is necessary to consider the possible synergy between factors. Since normal bone marrow contains accessory cells which are themselves the source of various cytokincs, it is preferable, when studying the effects of CSFs on progenitor cells, to purify the target cell population in order to eliminate these variables. We have used fluorescence-activated cell sorting (FACS) to highly enrich for a population of multipotent colony-forming cells [27]. This population has proved invaluable in elucidating the synergistic effects of CSFs and other haemopoietic factors [28]. Two factors in particular were found to synergise with the others to promote colony growth from FACS-enriched cells. These were G-CSF and interleukin-1 (IL-1). IL-1, which is produced by monocytes, macrophages, and endothelial cells [29], does not, by itself, stimulate colony growth. In combination with GM-CSF or M-CSF, however, it increases the number of colonies formed (by more than 100%) without influencing the morphology of the mature cells that are produced. The lineage of the recruited differentiating cells appears, therefore, to be determined by the CSF with which IL-1 is combined. No additive or synergistic effects were secn when IL-I was combined with G-CSF. The combination of G-CSF with either GM-CSF or M-CSF, however, incrcased the number of colonies formed to

256 an extent similar to that seen with IL-1. As with IL-I, the outcome of the response of the cells depended on whether the factor was combined with GM-CSF or M-CSF. Thus IL-1 and G-CSF have similar synergistic actions on haemopoietic cells. Neither IL-1 nor G-CSF could synergise with IL-3, suggesting that the same cells are recruited by IL-3 and by combinations such as G-CSF and MCSF. This conclusion is supported by experiments in which combinations of three factors were used. IL-3 plus M-CSF and IL-1 did not stimulate more colonies than IL-3 alone. In other words, the effect of IL-3 was not additive with that of IL-I and M-CSF together, implying that the factors act on the same population of cells. Similar results were obtained for the combination of IL-3, M-CSF, and G-CSF. At least some multipotent cells, therefore, can respond to IL-3, M-CSF, G-CSF, and IL-1, suggesting that they have receptors for all these factors. This observation has important implications for understanding the mechanism that regulates haemopoietic development.

Potential mechanisms for the regulation of haemopoiesis The mechanisms that regulate haemopoiesis must ensure the production of mature cells in the appropriate numbers and proportions from the stem cell pool. It has been suggested that this development may occur stochastically, with there being a fixed probability of a stem cell undergoing self-renewal or commitment and subsequently differentiating along a particular lineage [30]. In this model CSFs do no more than permit the development of cells along predetermined lineages. The formation of mixed colonies in the presence of IL-3 may be interpreted in this way. IL-3 permits the cells to express their full developmental potential but does not specify the lineage the progeny will adopt, since mature cells of different types are present in a single colony. The stochastic model also predicts that primitive cells have to undergo a degree of commitment before they are able to respond to more lineagespecific factors such as G-CSF and M-CSF. Thus,

although it might be sufficient for steady-state haemopoiesis, the mechanism seems too inflexible to be able to respond to acute challenges. The synergistic actions of combinations of growth factors make it more likely that these factors play a role in determining the direction of differentiation as well as its probability. Growth factors in combination can clearly act on a more primitive population of cells than had previously been believed. In addition, unlike IL-3, they determine lineage, so that the phenotype of the mature cell produced depends on the combination of factors to which its multipotent progenitor is exposed. Growth factors clearly have the potential to be sensitive, flexible, and potent regulators of haemopoiesis. What is the evidence to suggest that they fulfil this function in vivo? Although they are essential to colony growth in vitro, growth factors do not need to be added to haemopoietic cultures if the haemopoietic cells are in contact with an adherent stromal layer. Long term bone marrow cultures (LTBMC) encourage the development of a layer of adherent cells including endothelial cells, adipocytes, and macrophages [31]. Physical contact between the developing haemopoietic cells and the stroma in these cultures seems to be essential for haemopoiesis [32]. In this respect, therefore, LTBMC resemble the situation in vivo in which it is believed that haemopoiesis occurs in specific microenvironmental niches in the marrow stroma [33]. LTBMC can remain viable and active for several months, producing both progenitors and mature cells in large numbers [34]. The long-term persistence of these cultures suggests that a degree of stem cell self-renewal occurs supported by the stromal micro-environment. The addition of IL-3 or GM-CSF to LTBMC increases the production of mature cells and, to a lesser extent, of GM-CFC, but does not affect the longevity of the cultures (L. Coutinho, personal communication). This, therefore, is indirect evidence that the factors do not influence stem cell self-renewal in this system. Stromal cells can, however, produce some haemopoietic factors. M-CSF is constitutively produced [351, and GM-CSF and IL-1 (but not IL-3) have been detected at low levels [36, 37]. In addition, experiments have shown that

257 stromal cells can bind exogenously added GM-CSF and IL-3 [38, 39] and present these factors in a conformation that supports haemopoietic development. These factors may, therefore, be present but masked in LTBMC. The evidence from LTBMC demonstrates the fundamental importance of stromal cells to the maintenance of haemopoiesis. In addition, the ability of these cells to bind and present factors suggests a putative mechanism for the regulation of haemopoiesis by the combined action of the cells and the factors working in synergy. Whether this is in fact the mechanism that regulates haemopoiesis, however, awaits the application of more discriminating analytical techniques to LTBMC.

Negative regulation of haemopoiesis

The mechanisms discussed so far have emphasised stimulators of haemopoiesis. In theory, however, there could also be factors that limit the growth of stem cells and their progeny. Some workers have suggested that the two responses of proliferation and differentiation are in fact separately controlled. According to this hypothesis, CSFs stimulate both the proliferation of progenitor cells and the production of a different class of factor responsible for inducing differcntiation [40]. Thus proliferation and differentiation are normally coupled and, since cells become capable of fewcr divisions as they differentiate, the process is self-limiting. Breakdown of the system caused by failure to produce or respond to the differentiation-inducing factor would lead to uncontrolled proliferation, as in some leukaemias. The attraction of this model is that it couples the stimulation of proliferation with the induction of differentiation. There remain, however, some unanswered questions. Are there, for instance, different differentiation-inducing factors for different haemopoietic lineages, and why are four CSFs required if their function is the relatively simple one of stimulating proliferation? Whether or not differentiation is induced by a separate class of factor, it is clear that CSFs, directly or indirectly, must coordinate the expression of genes responsible for mod-

ulating the phenotype of haemopoietic cells. Little is known yet about this area, although an intracellular trans-acting factor capable of inducing differentiation has been reported [41], and several workers have noted the differential regulation of oncogene expression during induction of differentiation in haemopoietic cell lines [42-44]. We would hope that future molecular studies might morc precisely elucidate the relationship between proliferation and differentiation. A more direct mcchanism for the negative regulation of hacmopoiesis would be the inhibition of colony stimulating activity. There is evidence that extracts of mature granulocytes [45] and erythrocytes [46] contain substances that specifically inhibit the proliferation of granulocyte and erythrocyte precursors respectively, while macrophages from normal bone marrow produce a specific inhibitor of CFU-S [47]. Furthermore, in regenerating marrow taken from animals previously treated with cytotoxic drugs (or radiation), the inhibitory activity cannot be found, suggesting a feedback mechanism which suppresses its production in times of haemopoietic stress. Transforming growth factor beta (TGF-beta) is a well-characterised factor with haemopoietic activity. TGF-beta has both stimulatory and inhibitory effects on a wide range of cells [48]. We have found that it inhibits the response of multipotent cells to IL-3. It also inhibits GM-CFC colony formation stimulated by IL-3 or M-CSF, but not those stimulated by GM-CSF [49]. Its activity, therefore, appears to be both differentiation-specific and growth factor related. Since, among other sources, TGFbeta is produced by activated lymphocytes [50] and has been detected in bone marrow [51], it may have a physiological role in haemopoiesis.

Effects of CSFs on mature haemopoietic cells

Myeloid cells have a number of functions in immune and inflammatory responses, and there is evidence that CSFs play a part in modulating these processes by activating the mature cells after release from the bone marrow. IL-3 can enhance the ability of human monocytes to kill tumour cells; an

258 effect which appears to be modulated by the production of tumour necrosis factor-alpha (TNF-alpha) by the haemopoietic cells [52]. IL-3 also stimulates histamine release by bone marrow, probably by activating basophils [53]. Both TNF-alpha and histamine are important mediators of inflammatory responses. Release of 1L-3 may, therefore, be a key event in inflammation, particularly as it has been shown that the factor can be produced by keratinocytes [54]. GM-CSF also activates inflammatory cells by increasing the phagocytic ability of neutrophils [55] as well as their production of reactive oxygen intermediates in response to agonists such as phorbol esters [56]. Monocytes and eosinophils also respond to GM-CSF with increased cytocidal activity [57, 58] while histamine release is stimulated in basophils [59]. Like IL-3, GM-CSF stimulates cytokine release from haemopoietic cells, including TNF-alpha from macrophages [52]. The multiplicity of these reactions suggests that GM-CSF is a major mediator of the inflammatory response. The effects of G-CSF and M-CSF are more restricted. G-CSF acts mainly on granulocytes, promoting the antibody-dependent killing of malignant cells by neutrophils [60] as well as increasing superoxide production in response to the chemotactic peptide FMLP [611. M-CSF enhances the survival, proliferation [62], and cytocidal capacity of monocytes [63]. and the production of prostaglandin E, plasminogen activator, and TNF-alpha [63, 64]. The evidence for CSF involvement in inflammation is, perhaps, more compelling than it is for its role in haemopoiesis - a concept worth remembering when considering the effects of these factors in vivo.

Clinical use of growth factors

In view of the complex synergies between growth factors in vitro, the uncertainty about their role in normal haemopoiesis, and their ability to activate mature cells, it might be expected that their effects in vivo would be difficult to predict. Studies on the effects of CSFs in vivo have only become practicable with the availability of the recombinant prod-

ucts. A number of such studies have now been done, however, using IL-3, GM-CSF, and G-CSF. In mice infused with IL-3, the major effect was an enlargement of the spleen associated with an increase in the number of committed progenitor cells it contained [65]. There was, however, no overall increase in the number of progenitors in the spleen and bone marrow combined and only a slight rise in the number of peripheral neutrophils and monocytes. In primates, IL-3 has been reported to cause a modest increase in levels of reticulocytes, eosinophiis, and platelets [66]. Injection of GM-CSF into mice had little effect on leukocyte numbers in the peripheral blood. In primates, however, GM-CSF caused rapid leukoeytosis, increasing the numbers of peripheral neutrophils, eosinophils, monocytes, and lymphocytes [67]. These effects, however, were rapidly reversed upon withdrawal of the factor. The effect of G-CSF was more specific; infusion into either mice or primates affected only neutrophils, but their numbers were increased more than tenfold [68]. Again, the blood count fell rapidly to normal levels upon cessation of the treatment. In view of their ability to increase numbers of functional leukocytes, therapeutic applications of GM-CSF and G-CSF might be in the alleviation of conditions of haemopoietic insufficiency. These would include defective haemopoiesis as in myelodysplastic syndromes (MDS) and AIDS, foUowing chemotherapy, radiotherapy, bone marrow transplantation, and, possibly, in infection. Trials in some of these areas have already taken place. GM-CSF has been used in phase I/I1 trials in patients with AIDS [69] and MDS [70], leading in both cases to increases in the numbers of peripheral neutrophils, eosinophils, and monocytes. At doses greater than 30 ~g/kg/day side effects were severe and included bone pain, myalgia, fever and, most important, capillary leak syndrome leading to oedaema, renal dysfunction, pleural effusions, and erythroderma. While these reactions limit the upper dose of the factor, effects on haemopoiesis may be achieved at lower levels, which are less toxic. G-CSF, on the other hand, has been better tolerated, and no dose-limiting toxicity has been reported. In patients with malignant diseases G-CSF can

259 shorten, but not prevent, the period of neutropaenia that follows chemotherapy [71-73]. The factor appears to affect both the development of neutrophils and their release from the bone marrow. Infusion of the factor beginning after chemotherapy results in an early but brief drop in the peripheral neutrophil count followed by a release of mature cells into the circulation. Proliferation and differentiation of G-CFC in the bone marrow is also increased by about 20%. This expansion probably contributes to the rapid increase in peripheral neutrophils that is characteristic of G-CSF infusion. As in the animal studies, however, other lineages are unaffected. The advantage of G-CSF therapy is, therefore, that it reduces the time during which the patient is vulnerable to infection after chemotherapy, thus improving the quality of life and reducing the period of hospitalization and the need for intravenous antibiotics. Another potential use of CSFs is in intensive cytotoxic therapy. As one of the limitations on chemotherapy is marrow toxicity, growth factors may allow the use of higher doses of cytotoxic drugs than could otherwise be tolerated. The success of this approach depends on the susceptibility of the tumour to more vigorous attempts to eradicate it. A recent trial has, however, demonstrated the efficacy of dose intensification [74]. These preliminary results indicate the potential usefulness of haemopoietic growth factors in the treatment of neoplasia. A possible limitation, however, is the effect they may have on the proliferation of tumour cells. Factors such as EGF and PDGF are implicated in the growth of tumours [75], so it is a natural concern that CSFs might promote the growth of neoplastic as well as haemopoietic cells. Some non-haemopoietic tissues have receptors for CSFs and can proliferate in response to them. G-CSF, for example, stimulates the ~owth of endothelial cells [76]. It has alto been shown that small cell lung carcinoma, breast carcinoma, and osteosarcoma cell lines all respond to GM-CSF [77]. These findings, however, relate to established cell lines adapted to growth in vitro, and it may be premature to extrapolate them to the response of tumour cells in vivo. Controlled clinical trials are the only meaningful way to answer these questions.

Similar concerns have so far prevented the use of CSFs in the treatment of leukaemia. The response of primary leukaemic cells to CSFs in vitro is heterogenous, but in most cases a stimulation of proliferation has been recorded [78]. G-CSF has, however, been found to induce differentiation of leukaemic cells in mice [79]. Induction of differentiation of leukaemic cells leads to their clonal extinction [80], and it has been suggested that this may be exploited therapeutically [81]. Clearly, however, the extent to which growth factors alter the balance between differentiation and proliferation in leukaemic cells will determine their effect on the progression of the disease, and this is likely to vary according to the cell phenotype. /

Conclusions

Experiments using combinations of growth factors have clearly demonstrated the ability of these compounds to work in synergy when stimulating colony formation by primitive haemopoietic cells. Not only can combinations of factors act on more primitive cells than can single factors, but they also apparently determine the lineage that the progeny of these ceils will adopt. This strengthens the argument that the direction of differentiation of haemopoietic cells is influenced by the environment to which they are exposed, rather than being a preprogrammed characteristic. In the bone marrow haemopoietic cells develop in close association with stromal cells. We have presented evidence that growth factors are capable of interacting with these cells to create microenvironments which support haemopoiesis. Synergy between factors, as well as their effects on cytokine production by mature cells, make it difficult to predict their effects when administered to patients. Early results using GM-CSF or G-CSF to reduce neutropaenia after chemotherapy have, however, been encouraging. One of the problems of this type of therapy - the fact that G-CSF only increases peripheral neutrophils - might in the future be overcome by the use of combinations of factors. Since animal models arc not necessarily predictive of the response of patients, however, progress in this, as

260

in other areas of growth factor therapy, will probably only be made by clinical trials.

15.

References 16. 1. Ford CE, HamertonJL, Barnes DWH, Loutit JF: Cytological identification of radiation chimeras. Nature 177: 452454, 1956 2. Abramson S, Miller RG, Phillips RA: The identification in the adult bone marrow of pluripotent and restricted stem cells of the myeloid and lymphoid systems. J Exp Med 145: 1567-1579, 1979 3. Lemischka IR, Raulet DH, Mulligan RC: Developmental potential and dynamic behaviour of hematopoietic stem cells. Cell 45: 917-927, 1986 4. Snodgrass R, Keller G: Clonal fluctuations within the haematopoietic system of mice reconstituted with retrovirus infected cells. EMBO J 6: 3955-3960, 1987 5. Harrison DE: Normalproductionoferythrocytesbymouse marrow continuous for 73 months. Proe Nat Acad Sci 70: 3184-3188, 1973 6. Till JE, McCull0ch EA: A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res 14: 215-222, 1961 7. Dexter TM, Spooncer E, Schofield R, Lord BI, Simmons P: Haemopoietic stem cells and the problem of self-renewal. Blood Cells 10: 315-339, 1984 8. Metcalf D: Haemopoietic colonies. Berlin: Springer-Verlag, 1977 9. Fung MC, Hapel AJ, Ymer S, Cohen DR, Johnson RM, Campbell HD, Young IG: Molecular cloning of cDNA for murine interleukin-3. Nature 307: 233-237, 1984 10. Gough NM, Gough J, Metcalf D, Kelso A, Grail D, Nicola NA, Burgess AW, Dunn AR: Molecular cloning nf cDNA encoding a murine haematopoictic growth regulator, granulocyte-macrophage colony stimulating factor. Nature 309: 763-767, 1984 11. Tsuchiya M, Asano S, Kaziro Y, Nagata S: Isolation and characterization of the cDNA for routine granulocyte colony-stimulating factor. Proc Nat Acad Sci 83: 763_3-7637, 1986 12. Kawasaki ES, Lander MB, Wang AM, Van Ardsell J, Warren MK, Coync MY, Schweickart VL, Lcc MT, Wilson KJ, Boosman A. Stanley ER, Ralph P, Marks DF: Molecular cloning of a complementary DNA encoding human macrophage-specifie colony stimulating factor (CSF-I). Science 230: 291-296, 1985 13. Nicola NA: Hemopoietic growth factors and their interactions with specific receptors. J Cell Physiol (Suppl) 5: 9-14, 1987 14. Lc Beau M, Epstein ND, O'Brien SJ, Nienhuis AW, Yang YC, Clark SC, Rowley JD: The intcrleukin-3 gene is located on human chromosome 5 and is deleted in myeloid

17.

18.

19.

20.

21.

22.

23. 24. 25.

26.

27. 28.

leukemias with a dcletion of 5q. Proc Nat Acad Sci 84: 5913-5917, 1987 Le Beau M, Westbrook CA, Diaz NO, Larson RA, Rowley JD, Gasson JC, Goldc DW, Shcrr CJ: Evidence for involvement of GM-CSF and FMS in the deletion of 5q in myeloid disorder. Science 231: 984-987, 1986 Pettenati MJ, Le Bcau MM, Lemons RS, Shima EA, Kawasaki ES, Larsou RA, Shcrr CJ, Diaz MO, Rowley JD: Assignment of CSF-1 to 5q 33.1: Evidence for clustering of genes regulating haematopoiesis and for their involvement in the deletion of the long arm of chromosome 5 in mycloid disorders. Proc Nat Acad Sci 84: 2970-2974, 1987 Kelso A, Gough NM: Co-expression ofgranulocyte-macrophage colony-stimulating factor, gamma interferon and interleukin 3 and 4 is random in murinc alloreactivc T-lymphocyte clones. Proc Nat Acad Sci 85: 9189-9193, 1988 Paietta E, Papcnhausen P, Gucalp R, Reed J, Wieink PH: The 5q- deletion: correlation of breakpoints with the immuno-phenotype of leukemic blast cells. Leukemia 1" 822824, 1988 Bazill GW, llaynes M, Garland J, Dexter TM: Characterization and partial purification of a haemopoietic cell growth factor in WEHI-3B cell conditioned medium. Biochem J 210: 747-759, 1983 Spivak JL, Smith RR, lhle IN: Interleukin 3 promotes the in vivo proliferation of murine pluripotent hematopoietic stem cells. J Clin Invest 76: 1613-1621, 1985 Spooncer E, lleyworth CM, Dtmn A, Dexter TM: Selfrenewal and differentiation of interleukin-3-dependent multipotent stem cells are modulated by stromal cells and serum factors. Differentiation 31: 111-118, 1986 Metcalf D, Burgess AW, Johnson GR, Nicola NA, Nice EC, DeLamartcr J, Thatcher DR, Mermod J J: In vivo actions on hemopoietic cells of recombinant murine GMCSF purified after production in Escherichia coil: Comparison with purified native GM-CSF. J Cell Physiol 128: 421431, 1986 MetcalfD: Ilemopoietic Colony Stimulating Factors. Elsevier, Amsterdam, 1984 Metcalf D: The granulocyte-macrophage colony stimulating factors. Science 229: 16-22, 1985 Koikc K, Ogawa M, lhle JN, Miyake T, Shimizu T, Miyajimu A, ¥okota T, Arai K: Recombinant murine granulocyte-macrophage (GM) colony-stimulating factor supports formation of GM and multipotential blast cell colonies in culture: comparison with the effects of Interleukin-3. J Cell Physiol 131: 458-464, 1987 Robinson BE, McGrath HE, Quesenberry PJ: Recombinant murine granulocyte macrophage colony stimulating factor has megakaryocyte colony stimulating activity and augments megakaryocyte colony stimulation by Interleukin 3. J Clin Invest 79: 1048-1052, 1987 Lord BI, Spooncer E: Isolation of haemopoietic spleen colony forming cells. Lymphokine Res 5: 59-72, 1986 Heyworth CM, Ponting ILO, Dexter TM: The response of haemopoietic cells to growth factors: developmental impli-

261

29. 30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

4(I. 41.

42.

43.

cations of synergistic interactions. J Cell Science 91: Z39247, 1988 Dinarello CA: Interlcukin 1. Rev Infect Diseases 6: 51-86, 1984 Hayushi M, Gotoh O, Okabe-Kado J, Ilozumi M: Mechanisms controlling the kinetics in proliferation and differentiation of populations of mouse myeloid leukemic cells in vitro. J Cell Physiol 108: 123-128, 1981 Dexter TM, Allen TD, Lajtha LG: Conditions controlling the proliferation of hacmopoietic stem cells in vitro. J Cell Physiol 91: 335-344, 1977 Bentley SA: Close range cell:cell interaction required for stem cell maintenance in long term I'xme marrow cultures. Exp Hematol 9:308-311 Schofield R: The stem cell system. Biomedicine and Pharmacotherapy 37: 375, 1983 Dexter TM, Spoonccr E, Simmons P, Allcn TD: Long-term marrow culture: An overview of techniques and experience. In: Wright DG, Greenbergcr J'S (eds.) Long-term Bone Marrow Culture. Alan R Liss Inc., New York, 1984, 57-96 Shadduck RK, Wahecd A, Grecnberger JS, Dexter TM: Production of colony stimulating factor in long term bone marrow cultures. J Cell Physiol 114: 88-92, 1983 Oblon D J, Castromaspani H, Broxmeyer II: The production of colony stimulating activity by monocyte-enriched fractions from routine continuous bone marrow culture adherent layer. Br J Haematol 54: 291-294, 1983 Fibbe WE, van Damme J, Billiau A, Goselink HM, Voogt PJ, van Eeden G, Ralph P, Altrock BW, Falkenberg JItF: Interleukin-1 induces human marrow stromal cells in long term culture to produce granulocyte colony stimulating factor and macrophagc colony stimulating f~,ctor. Blood 71: 430-437, 1988 Gordon MY, Riley GP, Watt SM, Greaves MF: Compartmentalization of a hacmatopoietic growth factor (GMCSF) by glycosaminoglycans in the bone marrow microenvironment. Nature 326: 403--405, 1987 Roberts R, Gallagher J, Spooncer E, Allen TD, Bloomfield F, Dexter TM: Heparan sulphate bound growth factors: a mechanism for stromal cell mediated haemopoiesis. Nature 332: 376-378, 1988 Sachs L: The molccular control of blood development. Science 238: 1374-1379, 1987 Okazaki T, Kato Y, Tashima/Vl, Sawada 1t, Uchino H: Evidence of intraccllular and trans-acting differentiationinducing activity in human promyelocytic leukemia HL-60 cells: Its possible involvement in process of cell differentiation from a commitmcnt step to a phenotype expression step. J Cell Physiol 134: 261-268, 1988 Todokaro K, lkawa Y: Sequential expression of protooncogcne during mouse erythroleukcmia cell differentiation. Biochem Biophys Res Comm 135: 1112-1118, 1986 Einat M, Resnitzky D, Kimch A: Close link between reduction of c-myc expression by interferon and G0,'Gj arrest. Nature 313:597-600

44. Mitchell RL, Zokas L, Schreiber RI), Verma IM: Rapid increase in expression of the proto-oncogene .COs during human monocytc differentiation. Cell 4(1: 2(19-217. 1983 45. Rytomaa T, Kivmicmi K: Control of granulocyte productkm. l. Chalonc and antichalone, two specific humoral regulators. Cell and Tissue Kinetics 1: 329-340, 1068 46. Kivilaakso E, Rytomaa T: Erythrocyte chalone, a tissue specific inhibitor of cell proliferation in the erylhron. Cell and Tissue Kinetics 4: 1-9, 197l 47. Simmons PJ, Lord BI: Enrichment of CFU-s proliferation inhibitor producing cells based on their identification by the monoclonal antibody F4/80. J Cell Science 78: 117-131, 1985 48. Massague J: The TGF-~ family of growth and differentiation factors. Cell 49: 437-458, 1987 49. Hampson J, Ponting ILO, Cook N, Vodinclich L, Redmond S, Roberts AB, Dexter TM: The effects of TGF-[3 on haemopoietic cells. Growth Factors (in press), 1989 5(1. Kerhl JH, Wakefield L, Roberts AB, Jakowlew SB, Aivarez-Mon M, Derginck R, Sporn MB, Fauci AS: Production of transforming growth factor beta by human T lymphocytes and its potential role in the regulation of T cell growth. J Exp Ivied 163: 1037-1050, 1986 51. Ellingsworth LR, Brcnnan JE, Fok K, Rosen DM, Bentz H, Piez RA, Seyedin SM: Antibodies to the N-terminal portion of cartilage-inducing factor A and transforming growth factor. J Biol Chem 261: 12362-1~67, 1986 52. Cannistra SA, Vellegna E, Groshek P, Rambaldi A, Griffin JD: Human granulocyte-monocyte colony-stimulating factor and interleukin-3 stimulate monocyte cytotoxicity through a tumour necrosis factor-dependent mechanism. Blood 71: 672-(')76, 1988 53. Schneider E, Pollard H, Lepault E, Guygrand D, Minkowski M, Dy M: Interleukin-3 and granulocyte macrophage colony stimulating factor induce de novo synthesis of histadine-decarboxylase in haemopoietic progenitor cells. J Immunol 139: 3710-3717, 1987 54. Lugcr TA, Wirth U, Kock A: Epidermal cells synthesize a cytokine with interleukin-3-1ike properties. J lmmuno1134: 915-919, 1985 55. Flcischmann J, Golde DW, Weisbart RH, Gasson JC: Granulocyte macrophage colony-stimulating factor enhances phagocytosis of bacteria by human neutrophils. Blood 68: 708-711, 1986 56. Lupez AF, Williamson D J, Gumble JR, Begley CG, Harlan JIvl, Klebanoff SJ, Walterdorph PS, Wong G, Clark SC. Vadas IvIA: Recombinant human granulocyte macrophage colony stimulating factor stimulates in vgro mature human neutrophil and eosinophil function, surface receptor expression and survival. J Clin Invest 78: 1220-1228, 1986 57. Grabstein KH, Urdal DE, Tushinski RJ, Mochizuki DY, Price VL, Cantrell MA, Gillis S, Conlon PJ: Induction of macrophage tumoricidal activity by granulocyte-macrophage colony stimulating factor. Science 232: 506-508, 1986 58. Dessein AJ, Vadas MA, Nicola NA, Metealf D, David JR: Enhancement of human blood eosinophil cytotoxicity by

262

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

semipurificd eosinophil colony stimulating factor(s). J Exp Med 156: 90-103, 1982 Haak-Frendsho M, Arai N, Arai KI, Baeza ML, Finn A, Kaplan AP: Human recombinant granulocyte macrophage colony stimulating factor and interleukin-3 cause basophil histamine release. J Clin Invest 82: 17-20, 1988 Begley CG, Lopez AF, Nicola NA, Warren D J, Vadas MA, Sanderson C J, Metcalf D: Purified colony stimulating factors enhance the survival of human neutrophils and eosinophils in vitro: a rapid and sensitive microassay for colony stimulating factors. Blood 68: 102-106, 1986 Kitagawa S, Yuo A, Souza LM, Saito M, Miura Y, Takaku F: Recombinant human granulocyte colony stimulating factor enhances superoxide release in human granulocytes stimulated by the chemotactic peptide. Biochem Biophys Res Commun 144: 1143-1146, 1987 Stanley ER, Guilbert LJ, Tushinski R.J, Bartlemez SH: CSF-I; a mononuclear phagocyte lineage-specific hemopoietic growth factor. J Cell Biochem 21: 151-159, 1983 Kurland JI, Bockman RS, Broxmeyer HE, Moore MAS: Limitation of excessive myelopoiesis by the intrinsic modulation of macrophage-derived prostaglandin E. Science 199: 552-555, 1978 Warren MK, Ralph P: Macrophage growth factor CSF-I stimulates human monocyte production of interferon, tumor necrosis factor and colony stimulating activity. J lmmunol 137:2281-2285 Lord BI, Molineux G, Testa NG, Kelly M, Spooncer E, Dexter TM: The in vivo effects of recombinant IL-3. Lymphokine Res 5: 97-104, 1986 Donahue RE, Seehra J, Metzger M, Lefebvre D, Rock B, Carbone S, Nathan DG, Garnick M, Sehgal PK, Laston D, Lavallie E, McCoy J, Schendel PF, Norton C, Turner K, Yang Y-C, Clark SC: Human 1L-3 and GM-CSF act synergistically in stimulating haematopoicsis in primates. Science 241: 1820-1823, 1988 Donahue RE, Wang EA, Stone DK, Kamcn R, Wong GG, Sehgal PK, Nathan DG, Clark SC: Stimulation of haematopoiesis in primates by continuous infusion of recombinant human GM-CSF. Nature 321: 872-875, 1987 Tamura M, Hattori K, Nomura H, Oheda M, Kubota N, lmazeki I, Ono M, Veyama Y, Nagata S, Shira Fuji N, Asano S: Induction of neutrophilic granulocytosis in mice by administration of purified human native granulocyte colony-stimulating factor (G-CSF). Biochcm Biophys Res Commun 142: 454-460, 1987 Groopman JE, Mitsuyasu RT, DeLeo MJ, Oette DH, Golde DW: Effect of recombinant human granulocytemacrophage colony stimulating factor on myelopoiesis in the acquired immunodeficiency syndrome. N Engl J Med 317: 593-598, 1987 Vadhan-Raj S, Keating M, LeMaistre A, Hittelman WN, McCredie K, Trujillo JM, Broxmeyer HE, Henney C, Gutterman JU: Effects of recombinant human granulocyte-

71.

72.

73_

74.

75. 76.

77.

78.

79.

80.

81.

macrophage colony-stimulating factor in patients with the myelodysplastic syndromes. N Engl J Mcd 317: 1545-1552, 1987 Morstyn G, Campbell L, Souza LM, Alton NK, Keech J, Green M, Sheridan W, Mctcalf D, Fox R: Effect of granuIocyte colony stimulating factor on neutropenia induced by cytotoxic chemotherapy. Lancet 1: 667~'572, 1988 Brouchud MH, Potter MR, Morgenstcrn G, Blaseo MJ, Scarffe JH, Thatcher N, Crowther D, Souza LM, Alton NK, Testa NG: In vitro and in vivo analysis of the effects of recombinant human granulocyte colony-stimulating factor in patients. Br J Cancer 58: 64-69, 1988 Gabrilove JL, Jakubowski A, Scher H, Sternberg C, Wong G, Grous J, Yagoda A, Fain K, Moore MA, Clarkson B: Effect of granulocyte colony-stimulating factor on neutropenia and associated morbidity due to chemotherapy for transitional carcinoma of the urothelium. N Engl J Med 318: 1414-1422, 1988 Bronchud MH, Scarffe Jlt, Thatcher N, Crowther D, Souza LM, Alton NR, Testa NG, Dexter TM: Phase I/II study of recombinant human granulocyte colony-stimulating factor in patients receiving intensive chemotherapy for small cell lung carcinoma. Br J Cancer 56: 809-813, 1987 Heldin C-H, Westermark B: Growth factors: mechanism of action and relation to oncogenes. Cell 37: 9-20, 1984 Bussolino F, Wang JM, DeFillipi P, Turrini F, Sanavio F, Edgell C-JS, Aglietta M, Arese P, Mantovani A: Granulocyte- and granulocyte-macrophage-colony stimulating factors induce human endothelial cells to migrate and proliferate. Nature 337: 471--473, 1989 Dcdhar S, Galloway P, Eaves C: Human granulocyte-macrophagc colony stimulating factor is a growth factor for a variety of cell types of non-hemopoietic origin. Proc Am Soc Clin Oncol 29: 51, 1988 Testa NG, DexterTM: Haemopoietic growth factors: their role in acute myeloblastic leukaemia. British Medical Society Bulletin 45: 582-599, 1989 Metcalf D: Regulator-induced suppression of myeiomonocytic leukemic cells: clonal analysis of early cellular events. lnt J Cancer 30: 203-210, 1982 Daniel CP, Parreira A, Goldman JM, McCarthy DM: The effect of 1,25-dihydroxyvitamin D~ on the relationship between growth and differentiation in HL-60 cells. Lcuk Res 11: 191-196, 1987 Koeffler HP: Induction of differentiation of human acute myelogenous leukaemia cells: therapeutic implications. Blood 62: 709-721, 1983

Address for offprint: T.M. Dexter, Department of Experimental Haematology, Paterson Institute for Cancer Research, Christie Hospital, Wilmslow Road, Withington, Manchester M20 9BX, United Kingdom