Introduction

Mature cells of the non-lymphoid haemopoietic lineages are relatively short-lived, and maintenance of constant cell numbers requires a continuous process of new cell formation. These cells are generated from a small common set of multipotential stem cells mainly located in the bone marrow. Haemopoiesis therefore involves, as a continuing process throughout normal adult life, massive cell production, irreversible differentiation commitment of the progeny to one or other lineage, and progressive maturation of the cells within each lineage then release of the cells to the circulation. This is often followed by seeding of the mature cells to particular tissues, with major changes in their morphology and function in these locations. Finally, the mature cells may, or may not, need to become functionally activated to perform required physiological functions.

These complex events are executed with precision under basal conditions but the haemopoietic system also needs the capacity to respond quickly to emergency situations by major changes in cell production, relocation of mature cells and their functional activation.

It is not surprising, therefore, that the regulatory control of haemopoiesis is highly complex. Unlike cells in most tissues, haemopoietic cells do not form intercellular bridges, so all extrinsic regulation of haemopoietic cell behaviour requires activation of specific membrane-displayed receptors by secreted or membrane-displayed regulatory molecules.

The initiation of haemopoiesis

To initiate haemopoiesis as seen in adult life, three essential elements need to be developed: (i) a population of haemopoietic stem cells and their progeny; (ii) cell populations producing regulatory molecules; and (iii) cell populations (micro-environmental cells) providing the specialized tissue niches that are mandatory to support and perhaps regulate blood cell formation (Figure 1).

Figure 1.

Three special cell populations need to be generated to establish haemopoiesis: 1, a population of haemopoietic stem cells; 2, a set of cells of various types able to produce haemopoietic regulators; and 3, specialized micro-environmental cells able to sustain proliferating and maturing haemopoietic cells.

Full figure and legend (16K)

Haemopoietic stem cells are formed from non-haemopoietic precursors in two episodes of finite duration, the first in the yolk sac¹ and the second in the foetal para-aortic (AGM) region.² Nuclear transcription factors have been identified whose activation in these precursor cells is required to convert them to haemopoietic stem cells and their immediate more specialized progeny.³ In line with general embryological dogma, it can be presumed that selective transcription is likely to need to be activated by a combination of cell-position effects and gradients of inducing signals. Both represent forms of cell–cell signalling, but knowledge of these signals is rudimentary. While cell- position effects seem less likely to be of relevance for adult haemopoiesis, blood cell formation in adult life does require continuous commitment decisions, probably involving the same and additional nuclear transcription factors, and more needs to be understood about these genes and their regulation.

After the initial haemopoietic stem cells have been formed, they and their immediate progeny migrate to the developing foetal liver and there generate an extensive population of haemopoietic cells. Then follows a second migration of cells to populate the developing bone marrow and spleen. During late foetal and early postnatal life, liver haemopoiesis ceases and splenic haemopoiesis declines, leaving the bone marrow as the dominant site of blood cell formation.

Uncertainty exists regarding the manner in which cells acquire the ability to produce haemopoietic regulators. Many different adult cell types have the capacity to produce one or more of these haemopoietic regulators and the inductive signals able to elicit regulator production by such adult cells are known in many instances.⁴ During development, it is possible that some pre-programming by cell–cell signalling is required to allow such cells to be inducible, for example by low oxygen tension or microbial products. However, because most cells appear able to produce at least some haemopoietic regulators after induction, this may be a basic functional capacity possessed by most cells and thus may not require any specific gene pre-programming. This may not be the case for some regulators, such as erythropoietin, that are products of only a few cell types and, here, pre-programming would appear to be necessary.

Finally, the stromal cells that provide the special permissive environment, in which haemopoietic cell proliferation and maturation can occur, presumably achieve this capacity by activation of yet other genes. What controls this activation is possibly the most obscure problem of all because haemopoietic organ anlagen often develop in sites remote from the initial location of the first haemopoietic cells. For example, what signals could conceivably lead to the formation of two, and only two, lymph node anlagen in the axilla of a mouse, ready for the future arrival of the first T or B lymphocytes and where are these signals coming from?

The granulocyte-macrophage regulatory system

Rather than attempt a global summary of what is known of the regulation in adult life of the nine major lineages of blood cells, it is more useful to consider the situation with the dual granulocyte-macrophage lineage. This lineage shows the cardinal features of regulatory control, and the principles emerging are applicable to the situation with the other lineages.

Following the development of high-efficiency solid-state culture systems in the 1960s that were able to support the growth of clonal populations of granulocytes and/or macrophages, it became possible to explore the manner in which proliferation and maturation were regulated in developing colonies.

Such colonies were shown to be formed by a distinct class of committed progenitor cells, a transit population lying intermediate between haemopoietic stem cells and the morphologically identifiable cells in each lineage, such as myeloblasts, promyelocytes and myelocytes.⁴ These progenitor cells have little capacity for self-generation and are irreversibly committed to generating cells in the granulocyte-macrophage lineage.

The generation of maturing progeny cells in vitro by granulocyte-macrophage (GM) progenitor cells requires continuous stimulation by factors able to be detected in tissues and also in the serum and urine. These were obvious candidates for lineage-specific regulatory factors whose normal function was to control these populations in vivo. They were given the operational name colony stimulating factors (CSF), and purification and characterization of the four recognized CSF proceeded slowly during the next 15 years in step with necessary advances in the technology of protein purification.

By 1983, purification had been achieved of minute amounts of four CSF: GM-CSF, M-CSF, G-CSF and multi-CSF (more usually now termed interleukin-3, IL-3). The CSF are glycoproteins in the molecular weight range 18–35 000, and crystallographic studies have shown that their polypeptide chains have a 4- helical configuration held in shape by mandatory disulphide bridges.⁴ Regulator molecules of this type have two working faces that are able to enter into specific interactions with regions on two receptor chains. The logistical problems posed by the low abundance of the CSF and their complex structure were resolved by cloning cDNA for all four murine and corresponding human CSF. Virtually all subsequent studies and clinical applications have used recombinant CSF.

Colony stimulating factors have been shown to be able to be produced by a wide variety of cell types, such as fibroblasts, endothelial cells, stromal cells, lymphocytes and even mature macrophages.⁴ It is likely that all cell types in the body can produce one or other CSF if the cells are stimulated by an appropriate inducing agent. Highly effective inducing agents for many cell types are bacterial products such as endotoxin, foreign antigens and signalling molecules, such as tumour necrosis factor or interleukin-1, themselves induced by bacterial products. These responses occur with great rapidity and can elevate CSF levels more than 100-fold within 1–3 h.⁴

The body wide distribution of cells able to produce CSF, the ready accessibility of these cells to micro-organisms that have entered the body and the strong inducing action of such micro-organisms, strongly suggested that the CSF-producing system was one designed to respond promptly to the entry of micro-organisms or to be triggered by the products of effete or damaged cells. This was in line with the very low levels of CSF detectable in normal health, but the rapid and often extreme elevation of CSF in acute infections and induced inflammatory states.⁴

The latter observations led some to the view that, in fact, the CSF might merely be designed to regulate cellular responses to emergency situations, such as infections, and that they might not be important regulators of basal haemopoiesis.

Elimination of this extreme view has required the generation of mice with inactivation of the genes encoding the CSF or their receptors. Such animals have clearly documented the important role of the CSF in basal haemopoiesis as well as during emergency responses, although the abnormalities encountered in such mice were not always what might have been expected. Three examples can be used to illustrate this. Mice with inactivation of the G-CSF gene have a major (70%) reduction in blood neutrophil levels and a 50% reduction in marrow neutrophil and progenitor cell levels.⁵ They are also subnormal in their ability to mount neutrophil responses after challenge infections. These data clearly establish G-CSF as probably the most important single regulator of neutrophil production. Similarly, mice with the op/op (osteopetrosis) mutation, in which a nucleotide insertion in the M-CSF gene prevents production of M-CSF, exhibit marked deficiencies in macrophages in various tissues.^{6, 7} Because osteoclasts are derivatives of macrophages, such mice are also deficient in bone-resorbing osteoclasts, leading to the disease, osteopetrosis. The op/op mice are a little odd in being able to cure these deficiencies spontaneously after some months,⁸ presumably based on the development of alternative methods for stimulating macrophage formation. This sequence is surprising merely because of the length of time required to establish this alternative regulatory system.

A more unexpected outcome was observed in mice in which the genes encoding GM-CSF or the -chain of the GM-CSF receptor were inactivated.^{9, 10, 11} These mice show no numerical deficiencies in granulocytic or macrophage populations or their precursors. However, they exhibit two defects based on inadequate functional activity of their macrophages and those dendritic cells that are a derivative population of macrophage precursors. Thus, these mice invariably develop the lung disease, alveolar proteinosis, in which surfactant turnover is defective with accumulation of surfactant and protein in the alveoli together with the development of prominent para-bronchial accumulations of T and B lymphocytes. The mice also exhibit inadequate immune responses to foreign antigens based on defective function of the dendritic cells needed to present these foreign antigens to T lymphocytes.¹²

The natural history of both G-CSF- and GM-CSF-deficient mice is premature death in middle age with a miscellany of indolent infections; a situation which is compounded in mice interbred to have inactivation of both genes.¹³ This outcome highlights the essential role of both granulocytes and macrophages in adequate responses to micro-organisms, without assigning to either white cell type a dominant role in such resistance.

The converse responses of activation and expansion of granulocyte-macrophage populations are seen when CSF are injected into animals or humans exhibiting reduced responses to infections. The injection of either G-CSF or GM-CSF can increase granulocyte and/or macrophage production and can enhance resistance to such infections, forming the basis for their current extensive clinical use, most commonly in cancer patients in whom prior administration of chemotherapeutic agents has damaged bone marrow populations.⁴

The cell–cell signalling system involving the CSF has clear elements of a demand-driven regulatory arrangement. Entry of micro-organisms triggers increased production of CSF, first locally, then systemically. These agents either activate local granulocytes and macrophages to resolve the infection or if this fails, regulate the production of increased numbers of granulocytes and macrophages to cope with the situation (Figure 2). Underpinning these emergency regulatory responses is the continuous low-level production of CSF, able to regulate the basal production of granulocytes and macrophages appropriate for the demand for such cells in tissue remodelling and/or removal of effete cells occurring throughout normal life. The short half-life of induced transcription responses and of the CSF themselves means that, following removal of the inducing signal (usually microbial products), the signalling system rapidly returns to basal levels (Figure 3). There is little evidence to support the possibility that the control of CSF production is based on sensor systems monitoring, or dependent upon, the absolute numbers of mature cells, even though receptor-mediated clearance of such regulatory factors clearly is an important aspect of their biology and turnover.^{14, 15}

Figure 2.

Entry of microbial organisms triggers local cells to respond by producing a variety of regulator molecules. Agents such as the chemokine IL-8 attract pre-formed granulocytes and monocytes to the local site. Agents such as IL-1 can increase the local production of colony stimulating factor (CSF) and other molecules that can increase the functional activity of these cells. If these responses fail to eliminate the micro-organisms, regulator production becomes elevated and systemic elevations of CSF levels amplify cell production in the marrow, allowing larger numbers of mature granulocytes and macrophages to be recruited to the infection site.

Full figure and legend (25K)

Figure 3.

Colony stimulating factor (CSF) production is regulated by a demand-generated system. Entry of micro-organisms triggers enhanced production of CSF to activate or generate the needed granulocytes and macrophages. With elimination of the micro-organisms the initiating signal is removed, and the short half-life of CSF mRNA and protein rapidly restores the system to basal levels. The total number of granulocytes or macrophages appears not to be a significant factor in regulating CSF production.

Full figure and legend (38K)

The regulatory system involving four CSF appears unnecessarily complex, the more so because most granulocyte-macrophage progenitor cells and their progeny appear to co-express receptors for all four CSF.⁴ In vitro, all four have some ability to stimulate the production of both granulocytic and macrophage populations, although the action of G-CSF is strongly skewed in favour of granulocyte production and M-CSF in favour of macrophage stimulation.⁴

A feature of regulation by these types of molecule is the common occurrence of superadditive synergistic responses when two or more are combined. For the physiological control of responding populations in vivo, the ability to use combinations of factors not only introduces subtlety into the responses that are achievable, but also considerable efficiency.¹⁶ With combinations of regulators, required levels of cell production can be achieved by the production and use of lower concentrations of individual regulators. However, the apparently over-elaborate control system has raised in some the view that the control system exhibits evidence of redundancy, perhaps the consequence of evolutionary changes. The growing body of data from gene inactivation studies is clearly discounting the redundancy hypothesis with respect to haemopoietic regulators. There is no example of induced inactivation of a regulator gene which has no phenotypic consequences. All regulators play an irreplaceable role in one or other aspect of the biology of haemopoietic cells and, although many of their actions may overlap, none of the regulators is genuinely redundant.

Subsequent to the initial work with the CSF, additional regulators have been identified with proliferative actions on granulocytes. These include stem cell factor (SCF), with strong proliferative effects on granulocytic populations,¹⁷ interleukin-6 (IL-6) with an in vitro action on granulocytic populations similar to that of G-CSF,¹⁸ and flk3-ligand (FL) with weaker effects on granulocyte proliferation.¹⁹

There is a lack of quantitative concordance between the relative magnitude of responses to individual regulators observable in vitro and those seen in vivo which remains puzzling. For example, G-CSF and IL-6 appear almost identical in their actions in marrow cultures yet, in vivo, G-CSF is a powerful stimulus for granulocyte production while IL-6 has very little measurable action. It is recognized that the strong in vivo action of G-CSF is likely to depend on enhancement by SCF,²⁰ yet SCF has an equally strong enhancing action in vitro on IL-6,²¹ so a discrepancy remains.

One recurring theme emerging from individual gene inactivation studies on haemopoietic regulators is that there must exist additional regulators, yet to be identified, for all haemopoietic lineages. This current deficiency is most apparent for macrophage regulators.

A continuing issue raised by the gene inactivation studies is whether it will be possible to establish some sort of hierarchy of quantitative importance for the six to eight regulators with actions on each of the haemopoietic lineages. Granted that all play a key role in one or other aspect of the biology of the lineages, can some be identified as being of exceptional importance? This is a necessary question to pose when considering the practical and cost implications of transposing this knowledge to clinical medicine.

Seven regulators have so far been identified with proliferative actions on granulocytic and macrophage populations with the likelihood of additional active factors. Furthermore, complex interactions are possible between these factors and with a variety of modulating or inhibitory factors such as tumour growth factor- (TGF-). In this situation, pessimists have argued that the resulting complex networks would prevent any useful responses being achievable by the injection of any one of the factors to an animal or patient. Experience with the CSF has shown this view to be incorrect because highly reproducible responses are achievable by the administration of single agents; a source of satisfaction to clinicians.^{22, 23} However, these are still early days in attempts to translate experimental biology to clinical medicine. To use a single agent is not truly mimicking the presumably highly efficient combinatorial system used by the body, and clinicians ultimately will need to use combinations of injected factors to achieve superior cellular responses in their patients.

Receptors and receptor signalling

All actions of haemopoietic regulators are mediated by interactions with specific receptors on responding cells. This interaction results in cross-linking of two or more receptor chains, then their activation by phosphorylation to initiate signalling cascades in the cell.

The distribution, structure and function of these receptors introduces elements that are at the same time untidy, elegant, sophisticated and puzzling in their complexity. Display of a particular receptor dictates which cells in the body can respond to the corresponding regulator. The apparent untidiness in receptor biology stems from the unexpectedly broad range of cell types displaying particular haemopoietic receptors. The simplest design system would have one regulator restricted in its action to cells of a single haemopoietic lineage because only that lineage expresses the appropriate receptors. This is possibly most closely approximated for the erythroid regulator, erythropoietin. In fact, however, most haemopoietic receptors are expressed on more than one lineage of haemopoietic cells and thus no haemopoietic regulator exhibits strict lineage specificity. Not only can multiple haemopoietic lineages respond to a particular regulator but it is common enough for some non-haemopoietic cells also to express receptors and to respond. Why such an arrangement makes sense is not presently understood but the problem reaches bizarre proportions with regulators such as leukaemia inhibitory factor (LIF) with receptor display and the consequent actions of LIF on haemopoietic cells, embryonic stem cells, adipocytes, liver cells, uterine cells, nerve cells, pituitary cells and osteoblasts to mention only some of the responding cell types.²⁴

For agents such as the CSF, their relatively restricted receptor distribution permits them still to be regarded as essentially haemopoietic regulators. Nonetheless, an apparent design untidiness remains. In addition to its actions on granulocytic and macrophage cells, GM-CSF has clear actions on eosinophil precursors and weaker actions on megakaryocytic and erythroid cells. Interleukin-3 has even broader actions, including actions on mast cells and B lymphocytes.⁴

Reasons can be advanced why, in some situations, it is appropriate that cells of more than one haemopoietic lineage may need to respond. However, a major improvement is urgently needed in our understanding of coordinated cell and organ behaviour to appreciate why it is efficient and safe for an agent such as LIF to possess such pleiotropic actions.²⁵

A feature of the CSF that became apparent early during their development was that they were not simply mitotic stimuli. Colony stimulating factors have clear actions on differentiation commitment, maturation induction, cell survival and even on the functional activity of mature end cells.⁴ At the time, these were regarded by others as slightly bizarre findings, but this pattern of responses is now recognized to be a characteristic of other regulators of cell proliferation. There is probably no such thing as a simple mitogenic factor; all such factors have other important actions on responding cells.

How a variety of cellular responses is achievable using only a single type of receptor has become apparent from a dissection of the structure of membrane receptors, an analysis that has also revealed the economy of design and common ancestry of many receptors.

There are three basic types of receptor for haemopoietic regulators.²⁶ A few are homodimeric receptors with classical tyrosine kinase regions in their cytoplasmic domains, as typified by the receptors for M-CSF, SCF and FL. Binding of the regulator results in cross-linking of two receptor chains, then transphosphorylation of the receptors occurs to achieve an activated signalling state. The majority of specific receptor chains for haemopoietic regulators fall into a large class characterizable by a common WSXWS motif close to the membrane and one or more haemopoietin domains containing spaced cysteine residues. These shared features strongly suggest a common evolutionary origin for this receptor class, which also includes receptors for some non-haemopoietic regulators such as prolactin. A small subset of these receptors again forms homodimers, including the receptors for G-CSF, erythropoietin and thrombopoietin. Receptors in this class do not contain kinase domains in their cytoplasmic regions but are now known to be activated by phosphorylation through the action of one or other of the Janus kinases (JAK-kinases). The third class of receptors are heterodimers comprising a specific non-signalling -chain that can become coupled to a signalling -chain, again able to be activated by JAK-kinases. Binding of regulators to the -chain is of low affinity due to fast off-kinetics, but when this complex associates with a -chain, the binding interactions become of very high affinity, due to slow off-kinetics.²⁷ Activation of the -chain is dependent on association with the -chain–ligand complex. A curious feature of this last class of receptors is their promiscuity. -Chains tend to be shared by two or more different specific -chains. For example, the -chain of the GM-CSF receptor is shared also with the -chains for IL-3 and IL-5,²⁸ and the equivalent gp130-chain of the LIF receptor is shared also by oncostatin M, interleukin-6, interleukin-11 and ciliary neurotrophic growth factor.²⁴ In heterodimeric form, all are highly specific receptors for their appropriate ligands but the cells have economized on the production of necessary signalling chains by communal use.

It is important to emphasize the high efficiency of the haemopoietic regulator–receptor signalling systems. Haemopoietic regulators are typically active in concentrations of 10–1000 pg/mL and receptor numbers on responding cells are quite low, typically a few hundred per cell. Signalling is achievable with a receptor occupancy of as little as 30%. Where -chains share a common -receptor chain, competition between -chains is demonstrable.²⁹ The high speed of generating high-affinity regulator–receptor complexes that require the conjunction of two receptor chains, and the competition evident between -chains, despite their low numbers on the cell membrane, strongly imply that receptors occur in clusters on the cell membrane.

The sharing of common signalling chains by different specific -chains predicts that signals coming from an activated receptor chain are not likely to be able to convey detailed specific instructions to a cell to execute complex responses such as maturation. Clearly, if GM-CSF can stimulate both neutrophil and eosinophil formation using a common -signalling chain, this chain cannot issue detailed instructions for modifying a granulocyte precursor to adopt the highly specialized morphology of a neutrophil or an eosinophil precursor to develop the equally specialized, but completely different, morphology of an eosinophil. In fact, studies in which ectopic receptors have been inserted into cells of an incorrect lineage have confirmed this expectation. Insertion of erythropoietin receptors into macrophage precursors allows these cells to be stimulated to proliferate by erythropoietin but the progeny are macrophages, not erythroid cells.³⁰ Conversely, insertion of the M-CSF receptor into erythroid precursors allows M-CSF to stimulate colony formation but the colony cells are erythroid, not macrophages.³¹

The conclusion from a wide variety of such studies is that adequate mitotic signalling in factor-dependent cell lines is achievable using virtually any regulator after insertion of the appropriate specific receptor chains. In such cell lines, it is also feasible to induce some types of differentiation commitment and to initiate maturation in responding cells again using a wide variety of inserted receptors and relevant regulators. However, the exact type of response elicited, whether mitotic, maturation induction or merely functional activation, is dictated by the responding cell, not the type of receptor used.

Cell–cell signalling via this type of regulator–receptor system is thus very much a passive control system. Receptor distribution dictates what cell types respond and, in turn, the type of response elicited is dictated by the responding cell. At best, the regulator can achieve an on–off response. This is in contrast to the type of cell–cell inductive signalling referred to earlier in the brief review of embryonic and haemopoietic cell development. In these latter situations, cell–cell signalling seems to actively direct the future fate and function of the responding cells. To achieve these qualitative changes, a quite distinct signalling system seems to be required to regulate the necessary nuclear transcription factors.

Mutagenesis studies on a variety of receptors have identified distinct regions in their cytoplasmic domains that permit receptors to initiate multiple types of response in a cell.^{4, 28} Receptors have a region (box 1, box 2) close to the membrane that is responsible for initiating mitotic responses. More distal to the membrane is a region from which signals originate that initiate responses such as differentiation commitment and maturation initiation. At the extreme C terminus of many receptors is often a suppressor region that is able to dampen down proliferative responses to receptor activation. Deletions of these various regions have been noted as the likely basis for certain abnormal states. For example, deletion of the maturation-initiating region of the G-CSF receptor is seen in some patients with maturation arrest of granulocytic cells and consequent congenital neutropenia.³² Similarly, deletion of the suppressor region of the erythropoietin receptor can result in hyperresponsiveness to erythropoietin and mild polycythaemia.³³

Conclusions

To date more than 20 proliferation-inducing regulators for haemopoietic cells have been identified and produced in recombinant form. While this represents impressive progress in understanding the regulatory control of haemopoiesis, there are reasons for believing that an equal number of regulators have yet to be identified. Balancing these positive regulators, presumably, are numerous negative regulators. While some of these, such as TGF-, have been studied extensively, our knowledge of these negative regulators is very incomplete.

For the best understood granulocyte-macrophage, erythroid and megakaryocytic populations, a cohesive scheme can be built up for the likely manner in which their biology is regulated. This has culminated in the routine clinical use of regulators such as the CSF and erythropoietin. For some, this watershed transition from laboratory to the clinic, has indicated that the regulation of haemopoiesis has now been solved and that the question can now be closed.

The true situation is very different. We really understand very little about the biology of regulator production and degradation. The rationale for the pleiotropic actions of most regulators on cells of multiple types is quite unknown and indicates a profound lack of understanding of cell–cell interactions in the body. The most evident of our defects in understanding is our ignorance regarding the molecular basis for the choices that a responding cell is able to make in determining what type of response will actually occur. While this latter question may lie outside the realm of cell–cell signalling, it remains critical to our understanding of how haemopoiesis actually occurs and, until solved, impairs our capacity to intervene efficiently to manipulate haemopoiesis in the clinic.

References

1. Moore MAS, Metcalf D. Ontogeny of the haemopoietic system: Yolk sac origin of in vivo and in vitro colony forming cells in the developing mouse embryo. Br. J. Haematol. 1970; 18: 279–96. | PubMed | ISI | ChemPort |
2. Dieterlen-Lievre F, Martin C. Diffuse intraembryonic hemopoiesis in normal and chimeric avian development. Dev. Biol. 1981; 88: 180–91. | PubMed | ISI | ChemPort |
3. Zon LI. Developmental biology of hematopoiesis. Blood 1995; 86: 2876–91. | PubMed | ISI | ChemPort |
4. Metcalf D, Nicola NA. The Hemopoietic Colony-Stimulating Factors. From Biology to Clinical Applications. Cambridge: Cambridge University Press, 1994.
5. Lieschke GJ, Grail D, Hodgson G et al. Mice lacking granulocyte colony-stimulating factor have chronic neutropenia, granulocyte and macrophage progenitor cell deficiency, and impaired neutrophil mobilization. Blood 1994; 84: 1737–46. | PubMed | ISI | ChemPort |
6. Yoshida H, Hayashi S, Kunisada T et al. The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene. Nature 1990; 345: 442–4. | Article | PubMed | ISI | ChemPort |
7. Wiktor-Jedrzejczak W, Ratajczak MZ, Ptasznik A, Sell KW, Ahmed-Ansari A, Ostertag W. CSF-1 deficiency in the op/op mouse has differential effects on macrophage populations and differentiation stages. Exp. Hematol. 1992; 20: 1004–10. | PubMed | ChemPort |
8. Begg SK, Radley JM, Pollard JW, Chisholm OT, Stanley ER, Bertoncello I. Delayed hematopoietic development in osteopetrotic (op/op) mice. J. Exp. Med. 1993; 177: 237–42. | Article | PubMed | ISI | ChemPort |
9. Stanley E, Lieschke GJ, Grail D et al. Granulocyte/macrophage colony-stimulating factor-deficient mice show no major perturbation of hematopoiesis but develop a characteristic pulmonary pathology. Proc. Natl Acad. Sci. USA 1994; 91: 5592–6. | Article | PubMed | ChemPort |
10. Dranoff G, Crawford AD, Sadelain M et al. Involvement of granulocyte-macrophage colony stimulating factor in pulmonary homeostasis. Science 1994; 264: 713–16. | Article | PubMed | ISI | ChemPort |
11. Robb L, Drinkwater CC, Metcalf D et al. Hematopoietic and lung abnormalities in mice with a null mutation of the common subunit of the receptors for granulocyte-macrophage colony-stimulating factor and interleukins 3 and 5. Proc. Natl Acad. Sci. USA 1995; 92: 9565–9. | Article | PubMed | ChemPort |
12. Wada H, Noguchi Y, Marino MW, Dunn AR, Old LJ. T cell functions in granulocyte/macrophage colony-stimulating factor deficient mice. Proc. Natl Acad. Sci. USA 1997; 94: 12557–61. | Article | PubMed | ChemPort |
13. Seymour JF, Lieschke GJ, Grail D, Quilici C, Hodgson G, Dunn AR. Mice lacking both granulocyte colony-stimulating factor (CSF) and granulocyte-macrophage CSF have impaired reproductive capacity, perturbed neonatal granulopoiesis, lung disease, amyloidosis, and reduced long-term survival. Blood 1997; 90: 3037–49. | PubMed | ChemPort |
14. Kuter DJ, Rosenberg RD. The reciprocal relationship of thrombopoietin (c-mpl ligand) to changes in the platelet mass during busulfan-induced thrombocytopenia in rabbits. Blood 1995; 85: 2720–30. | PubMed | ChemPort |
15. Metcalf D, Mifsud S, Di Rago L, Robb L, Nicola NA, Alexander W. The biological consequences of excess GM-CSF levels in transgenic mice also lacking high-affinity receptors for GM-CSF. Leukemia 1998; 12: 353–62. | Article | PubMed | ISI | ChemPort |
16. Metcalf D. Hemopoietic regulators: Redundancy or subtlety? Blood 1993; 82: 3515–23. | PubMed | ISI | ChemPort |
17. Metcalf D, Nicola NA. Direct proliferative actions of stem cell factor on murine bone marrow cells in vitro: Effects of combination with colony-stimulating factors. Proc. Natl Acad. Sci. USA 1991; 88: 6239–43. | Article | PubMed | ChemPort |
18. McKinstry W, Metcalf D, Li CL. Distribution of IL-6 receptors on murine hemopoietic and lymphoid cells and the anomalous action of IL-6 in stimulating granulocytic proliferation. Leukemia 1994; 8: 1726–33. | PubMed | ChemPort |
19. Rasko JEJ, Metcalf D, Rossner MT, Begley CG, Nicola NA. The flt3/flk-2 ligand: Receptor distribution and action on murine haemopoietic cell survival and proliferation. Leukemia 1995; 9: 2058–66. | PubMed | ISI | ChemPort |
20. Cynshi O, Satoh K, Shimonaka Y et al. Reduced response to granulocyte colony-stimulating factor in W/W^v and Sl/Sl^d mice. Leukemia 1991; 5: 75–7. | PubMed | ChemPort |
21. Metcalf D. The cellular basis for enhancement interactions between stem cell factor and the colony stimulating factors. Stem Cells 1993; 11: (Suppl. 2): 2–11.
22. Welte K, Gabrilove J, Bronchud MH, Platzer E, Morstyn G. Filgrastim (r-metHuG-CSF): The first 10 years. Blood 1996; 98: 1907–29.
23. Scarffe JH. Emerging clinical uses for GM-CSF. Eur. J. Cancer 1991; 27: 1493–504. | PubMed | ChemPort |
24. Nicola NA, Hilton DJ. Leukemia inhibitory factor and its receptor. In: Leroith, D, Bandy C (eds) Growth Factors and Cytokines in Health and Disease. JAI Press Inc, Greenwich, 1997; 613–68.
25. Metcalf D. Implications of the polyfunctionality of hemopoietic regulators: The Metcalf Forum. Stem Cells 1994; 12 (Suppl. 1): 259–75. | PubMed | ISI |
26. Hilton DJ. An introduction to cytokine receptors. In: Nicola NA (ed.) Guidebook to Cytokines and Their Receptors. Oxford: Oxford University Press, 1994; 8–16.
27. Nicola NA. Cytokine pleiotropy and redundancy: A view from the receptor. The Metcalf Forum. Stem Cells 1994; 12 (Suppl. 1): 3–14. | PubMed |
28. Nicola NA, Smith A, Robb L, Metcalf D, Begley CG. The structural basis of the biological actions of the GM-CSF receptor. In: Goode J (ed.) The Molecular Basis of Cellular Defence Mechanisms. (Ciba Foundation Symposium no. 204). Chichester: John Wiley and Sons, 1997; 19–32.
29. Nicola NA, Robb L, Metcalf D, Cary D, Drinkwater CC, Begley CG. Functional inactivation in mice of the gene for the interleukin-3 (IL-3) -specific receptor -chain: Implications for IL-3 function and the mechanism of receptor transmodulation in hematopoietic cells. Blood 1996; 87: 2665–74. | PubMed | ChemPort |
30. McArthur GA, Longmore GL, Klingler K, Johnson GR. Lineage-restricted recruitment of immature hematopoietic cells in response to erythropoietin after normal hematopoietic cell transfection with erythropoietin receptor. Exp. Hematol. 1995; 23: 645–54. | PubMed | ChemPort |
31. McArthur GA, Rohrschneider LR, Johnson GR. Induced expression of c-fms in normal hematopoietic cells shows evidence for both conservation and lineage restriction of signal transduction in response to macrophage colony-stimulating factor. Blood 1994; 83: 972–81. | PubMed | ChemPort |
32. Dong F, Dale DC, Bonilla MA et al. Mutations in the granulocyte colony-stimulating factor receptor gene in patients with severe congenital neutropenia. Leukemia 1997; 11: 120–5. | Article | PubMed | ChemPort |
33. Yi T, Zhang J, Miura O, Ihle J. Hemopoietic cell phosphatase associates with erythropoietin (Epo) receptor after Epo- induced receptor tyrosine phosphorylation: Identification of potential binding sites. Blood 1995; 85: 87–95. | PubMed | ISI | ChemPort |

Acknowledgements

The work from the author's laboratory was supported by the Carden Fellowship Fund of the Anti-Cancer Council of Victoria, the National Health and Medical Research Council, Canberra, the AMRAD Corporation, Melbourne, and the National Institutes of Health, Bethesda, Grant No. CA22556.