| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Concise Reviews |
a Division of Human Immunology, The Hanson Centre for Cancer Research, The Institute of Medical and Veterinary Science, Adelaide, SA, Australia;
b The Baker Research Institute, Melbourne, Victoria, Australia
Key Words. Colony-stimulating factor • Hematopoiesis • Receptor • Signal transduction
Dr. Angel F. Lopez, Division of Human Immunology, The Hanson Centre for Cancer Research, The Institute of Medical and Veterinary Science, Adelaide, 5000 SA, Australia.
 |
Abstract |
|---|
 Top Abstract IntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
Central to an organism's need to meet hemopoietic demands and to mount an immune/inflammatory response is its ability to stimulate specific blood cell populations. GM-CSF, interleukin 3 (IL-3), and IL-5 are cytokines principally produced by activated T cells which exhibit pleiotropic activities, stimulating proliferation, survival, and differentiation of myeloid hemopoietic cells and the effector function of the terminally differentiated myeloid cells [1, 2]. GM-CSF receptors have been identified on most types of myeloid progenitors and on mature monocytes, neutrophils, eosinophils, basophils, and dendritic cells [2]. IL-3 receptors are present on early hemopoietic progenitor cells, on certain committed myeloid progenitors, eosinophils, and basophils, while IL-5 receptors have been shown to be expressed mainly on eosinophils [2]. Consistent with the pattern of receptor expression, GM-CSF and IL-3 can stimulate immature myelomonocytic cells in the hemopoietic system and cause differentiation of the granulocyte and macrophage populations [2]. In addition, GM-CSF and IL-3 are important for inducing the effector functions of these cells and hence contribute to the body's defense against microbial pathogens [1, 2]. IL-5 can chiefly stimulate production, survival, and granule release of eosinophils and participate in allergic reactions and anti-parasite responses [1, 2]. Experiments aimed at investigating the in vivo function of GM-CSF, IL-3, and IL-5 using a gene knockout approach would suggest that these cytokines are not essential in maintaining "steady-state" populations of granulocytes, macrophages, and eosinophils, but instead are central for the accelerated production of these cells in inflammation and infection. As such, they may be considered more as "reactive" cytokines required in emergency situations.
Apart from its proposed role in stimulated hemopoiesis and inflammatory responses, GM-CSF has also been implicated in cell transformation. The GM-CSF receptor has been identified in different types of cancer cells, including those of acute myeloid leukemia, chronic myeloid leukemia, juvenile myelomonocytic leukemia, melanoma, certain breast cancer cell lines, small-cell lung carcinoma, and the prostate [3, 4]. In all of these cells (with the apparent exception of some melanoma cell lines), both subunits of the GM-CSF receptor have been detected, suggesting that the receptor is capable of fully transducing GM-CSF signals and hence contributing to proliferative and/or survival functions.
The receptors for GM-CSF, IL-3, and IL-5 consist of a ligand-specific 
subunit (GMR, IL-3R, and IL-5R respectively) [5-7] and a common ß subunit (ßC) [8]. The subunit is the major ligand-binding subunit and on its own does not seem to transduce any of the biological activities ascribed to GM-CSF, IL-3 and IL-5 in hemopoietic cells [9, 10]. The ßC subunit, on the other hand, converts the ligand-bound subunit to a high affinity state and is important for most, if not all, of the signaling [7, 8, 11]. Curiously, there are two ß subunits in mice; a ßC which can be activated by GM-CSF/IL-3/IL-5, and the IL-3-specific ß subunit, ßIL-3, which binds IL-3 with low affinity and only forms a high-affinity receptor with IL-3R [12, 13]. In terms of cross-reactivity, human GM-CSF does not bind to mouse GM-CSF receptors detectably; however, crosstalk between human and mouse receptors has been observed. FDCP1 cells that express endogenous mouse GM-CSF receptors and which are transfected with the human GMR exhibit crosstalk between the receptor species [14]. This trans-species crosstalk phenomenon should be borne in mind for interpreting results, as proliferative and survival signals apparently mediated by human receptors may be confounded by the presence of endogenous mouse receptors.
|
Activation of the GM-CSF Family of Receptors: The Extracellular Story |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
subunits for GM-CSF, IL-3 and IL-5 as well as ßC are members of the type I cytokine receptor superfamily and contain conserved extracellulardomains termed cytokine receptor modules (CRMs) [15]. Each CRM consists of two repeats of a fibronectin type III-like domain. These repeats carry two sets of conserved motifs typical of this family of receptors. The first repeat contains four cysteines with conserved spacing, while the second repeat contains a WSXWS motif. The homology between ßC and the growth hormone (GH) receptor has allowed the extracellular structure of ßC to be modeled from the crystal structure of the GH receptor [16, 17]. This model predicts that the four spatially conserved cysteines within the CRM of ßC will form disulfide bonds and are important for maintaining the structural integrity of ßC. In terms of the WSXWS motif, mutagenesis of this sequence in a number of cytokine receptors has been shown to disrupt cytokine binding and receptor activation. However, from the crystal structure of GH bound to the GH receptor, the WSXWS-equivalent motif does not appear to be directly involved in ligand binding [17]. Mutagenesis studies in both the GH receptor and the erythropoietin (EPO) receptor suggest that the WSXWS motif is important for correct folding of the extracellular domain of cytokine receptors [18, 19].
High-Affinity Binding
Reconstitution studies where both the GMR or ßC are expressed at the cell surface indicate that GM-CSF first binds to GMR with low affinity (Kd ~5 x 10–9M) and that this complex is converted to high affinity following recruitment of ßC (Kd ~10–11M) [7, 8, 11]. Such an observation would imply that GM-CSF would have at least two binding interfaces, one that first binds GMR and one that subsequently recruits and binds ßC to form a 1:1:1 complex of GM-CSF:GMR:ßC. A model for such an interaction in illustrated in Figure 1. Since GM-CSF cannot detectably bind ßC directly in the absence of GMR, it would also seem likely that there are domains within GMR that cooperate with GM-CSF in binding ßC ( Fig. 1). Mutagenesis of GM-CSF, GMR, and ßC indicate that, in fact, there are likely to be two binding interfaces on GM-CSF important for the formation of a 1:1:1 complex. For low-affinity binding, residues centered on the fourth helix of GM-CSF involving Asp112 are likely to be important for establishing an ionic interaction, possibly through Arg280 of GMR [20, 21]. On the second binding interface, a single conserved glutamate residue located in the first helix of GM-CSF, IL-3 and IL-5 is essential for high-affinity binding [22-24]. Mutants of this residue in GM-CSF (E21R) [23], IL-3 (E22R) [22], or IL-5 (E13Q) [24] abolish high-affinity binding and result in either functional antagonism (GM-CSF:E21R and IL-5:E13Q) or weak agonism (IL-3:E22R). On the other hand, substitution with a neutral amino acid at this position such as alanine (GM-CSF-E21A) resulted in weak agonism, indicating that a GM-CSF/GMR complex may recruit ßC through a charge interaction involving Glu21 of GM-CSF and a positively charged amino acid on ßC [23]. The close association of ligand and ßC is also supported by the finding that GM-CSF can be cross-linked to ßC following receptor activation [8].
|
Receptor Dimerization and Activation
While type I transmembrane receptors vary considerably in terms of their structure, subunit composition, and the nature of ligand binding, it has been repeatedly shown that receptor dimerization is an obligatory event in receptor activation [28]. The GM-CSF/IL-3/IL-5 receptors are no exception to this general rule. However, while the formation of a 1:1:1 complex of GM-CSF:GMR:ßC may represent a high-affinity binding complex, it is much less clear what the stoichiometry of the active receptor complex is. Studies using dominant-negative, chimeric, and mutant receptors as well as modeling studies would indicate that at least two ßC subunits are required for receptor activation and signaling and that the active GM-CSF receptor is composed of a 2:2:2 complex of GM-CSF:GMR:ßC. A chimeric receptor consisting of an EPO receptor extracellular domain, which is known to dimerize in response to EPO, and a ßC cytoplasmic domain were examined for their ability to confer growth signals in response to EPO. Expression of these receptors in Ba/F3 cells resulted in EPO-dependent growth, indicating that ßC cytoplasmic-domain dimerization is sufficient for signaling [29]. In addition, GM-CSF was able to support the growth of cells coexpressing wild-type ßC and a chimeric receptor consisting of the GMR extracellular domain and the ßC intracellular domain, further supporting the idea that dimerization of ßC intracellular domains is sufficient for receptor activation [10]. In a reverse experiment, GM-CSF was not able to support cell growth in cells coexpressing wild-type GMR and a chimeric receptor consisting of the extracellular domain of ßC and the cytoplasmic domain of GMR suggesting that dimerization of ßC but not chains is sufficient for receptor activation [10]. Two chains nevertheless appear to be necessary for the formation of an active complex. Coexpression of a cytoplasmically truncated form of GMR with wild-type GMR and ßC in NIH 3T3 cells impairs cell proliferation and focus formation in response to GM-CSF [30].
Other studies have used chimeric receptors consisting of either the Fos or Jun leucine zippers in the extracellular domains and either the GMR or ßC intracellular domains to investigate the stoichiometry of the active receptor. Heterodimerization of GMR and ßC intracellular domains or homodimerization of ßC intracellular domains mediated by leucine zipper dimerization both promoted cytokine-independent growth in Ba/F3 cells [31]. While the ability of leucine zipper-mediated ßC intracellular homodimers to promote cell proliferation is in line with the above experiments examining the role of dimerization in receptor activation, the finding that leucine zipper-mediated heterodimerization of GMR and ßC intracellular domains supported cell growth would imply that a 1:1:1 complex of GM-CSF:GMR:ßC would also be capable of signaling. However, it is difficult to exclude the possibility in these experiments that, in addition to heterodimerization of GMR and ßC intracellular domains, homodimers of ßC intracellular domains were also formed or that the endogenous mouse ß is recruited into these complexes.
Overall, these data together with modeling studies, suggest a model in which the active receptor complex is formed following the dimerization of two 1:1:1 high-affinity complexes of GM-CSF:GMR:ßC to form an activated receptor composed of a 2:2:2 complex ( Fig. 1). In its simplest form, this model proposes that the active receptor complex would contain two ßC subunits; however, higher-order oligomeric complexes may also form functionally activated receptors.
A critical event in the production of an active receptor complex is not only its initial formation but also its subsequent stability. In the case of the GM-CSF/IL-3/IL-5 receptors, stability is likely to be promoted by a number of factors. First, there are likely to be domains within the 1:1:1 complex of GM-CSF:GMR:ßC that are responsible for the recruitment and binding of a second 1:1:1 complex. In a model similar to that proposed by Stahl and Yancopolous for IL-6 binding, GM-CSF may have three binding interfaces: one that interacts with GMR, one that interacts with ßC (as discussed above and shown in Fig. 1), and one that interacts with and recruits a second high-affinity GM-CSF:GMR:ßC complex [32]. While this possibility remains to be tested, molecular modeling of the GM-CSF:GMR:ßC complex using the crystal structure of GH bound to the GH receptor would suggest that an alternative mechanism for receptor dimerization may be at work [16, 17]. In terms of size, GM-CSF, IL-3, and IL-5 are smaller than the IL-6 family of cytokines. It would seem unlikely from these models that the smaller size of GM-CSF, IL-3, and IL-5 would permit a third binding interface for the recruitment of a second ligand:R:ßC complex as is proposed to occur for IL-6 [32]. Alternatively, we propose that high-affinity binding of GM-CSF, IL-3, or IL-5 and the formation of a ligand:R:ßC complex may create specific motifs or expose specific residues that mediate dimerization and the formation of a 2:2:2 complex of ligand:R:ßC. This is based on our observation that ßC contains, apart from the spatially conserved cysteines of the CRM, two additional cysteine residues, C86 and C91, that are conserved between mouse and human ßC and which are involved in the formation of a disulfide linkage with cysteines in the subunit [33]. IL-3 stimulation of COS cells expressing IL-3R and either ßCC86A or ßCC91A resulted in high-affinity binding but no ßC tyrosine phosphorylation. While IL-3R/ßC disulfide-linked dimers are normally observed following ligand stimulation, these dimers were not detected following stimulation of cells expressing either ßCC86A or ßCC91A [33]. As proposed by Stomski et al. [33] and illustrated in Figure 1, these disulfide linkages most likely occur between two different ligand:R:ßC complexes and thus may be responsible for the dimerization and stability of an active 2:2:2 complex.
The conserved motif represented by Cys 86 and Cys 91 in human and mouse ßC is not found in related cytokine receptors and suggests an evolutionary conserved structural and functional basis for activation of the GM-CSF, IL-3, and IL-5 receptors. A second unique feature of this receptor subfamily is the demonstration that, in the case of the GM-CSF receptor, a proportion of these receptors are preassembled (before the addition of GM-CSF). We and others have shown that mutations of GMR that abolish GM-CSF binding are compensated for when coexpressed with ßC [21, 34]. As ßC is unable to detectably bind GM-CSF in the absence of GMR, this suggests that ßC is pre-associated with GMR prior to the addition of GM-CSF. We have now established that a proportion of GM-CSF receptors are preassociated by showing that GMR and ßC can be coimmunoprecipitated in the absence of GM-CSF from primary leukemic cells and myeloid cell lines, and in GM-CSF receptor transfected cells [35]. This is in direct contrast to the IL-3 and IL-5 receptors in which the subunit and ßC associate only in the presence of IL-3 or IL-5, respectively [35]. The preformed GMR:ßC complex may be responsible for the rapid association of GM-CSF to monocytes and eosinophils relative to IL-3 and IL-5. The kinetics of association for GM-CSF binding to eosinophils and monocytes corresponds to two classes of receptors, a relatively slow one with association characteristics identical to those of IL-3 and IL-5, presumably representing ligand-induced receptors, and a second class which exhibits virtually instantaneous GM-CSF binding and possibly represents preformed receptors [35].
Interestingly, the preformed GMR/ßC complex may be activated not only by GM-CSF but also by IL-3 or IL-5 [35]. The significance of this is unclear; however, it may provide a mechanism for regulating or increasing the diversity of cytokine signaling. As has been proposed in other receptor systems, heterodimerization of different members of a cytokine receptor family may allow the recruitment of different signaling molecules. For example, several members of the epidermal growth factor (EGF) receptor family, such as Neu, ErbB3 and the EGF receptor itself, have shown the potential to heterodimerize in response to EGF or a related ligand, heregulin [36]. In fact, the ability of EGF to generate diversity in signaling through heterodimerization of receptors has been demonstrated with respect to phosphatidylinositol 3-kinase (PI 3-kinase). Although EGF is known to activate PI 3-kinase activity in some cells, the EGF receptor does not appear to contain a motif for the recruitment of PI 3-kinase. However, ErbB3 contains a PI 3-kinase binding motif. EGF binding to the EGF receptor results in the recruitment and activation of ErbB3 which in turn allows PI 3-kinase activation [37]. Clearly, crosstalk between different receptors has the potential to increase the signaling repertoire for a receptor family and may account for some of the diverse biological functions attributed to GM-CSF, IL-3, and IL-5.
A preformed GM-CSF receptor complex and the indirect recruitment of ßC may also have implications pertinent to the cytokine receptor family at large, since they may explain crosstalk by other cytokine receptors and the universal expression of the GM-CSF receptor in the hemopoietic system. In murine cells, tyrosine phosphorylation of mouse ßC chains has been demonstrated in response to G-CSF [38] and EPO. Similarly, stem cell factor stimulation has been shown to result in cross-phosphorylation of mouse ß chains on serine or threonine [39]. In the human system, EPO stimulation has been shown to unidirectionally cross-phosphorylate ßC in UT7 cells [40]. More recently, thrombopoietin has also been demonstrated to unidirectionally cross-phosphorylate ßC in TF-1 cells. The significance of these crosstalk phenomena is not understood, but the unidirectionality suggests that ßC phosphorylation may contribute to the signaling of cytokines other than GM-CSF, IL-3 and IL-5. Together with the possible involvement of the preformed GM-CSF receptor complex, these observations of receptor crosstalk suggest that ßC may have a role in cytokine receptor signaling, perhaps facilitating certain universal and essential functions such as cellular survival.
|
Activation of the GM-CSF Family of Receptors: The Intracellular Story |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
|
Following receptor activation and phosphorylation by either JAK2 or other tyrosine kinases, phosphotyrosine residues within the cytoplasmic domain of ßC could function as high-affinity binding sites for SH2 (src-homology) or PTB (phosphotyrosine binding) domain proteins [55]. Recruitment of SH2 or PTB domain proteins to the ßC would then serve to couple the activated receptor to downstream signaling pathways [55] ( Fig. 2). One such example is the signal transducer and activator of transcription molecule, STAT5. The STAT proteins were originally identified as latent cytoplasmic transcription factors that were activated by JAK proteins in interferon signaling [56]. Although not formally demonstrated for ßC, STAT proteins have been shown to be recruited to tyrosine phosphorylated cytokine receptors via their SH2 domains. Phosphorylation of STAT5 by JAK2 would then result in STAT5 activation, dimerization, and translocation to the nucleus, where it is directly involved in regulating gene transcription [50, 52]. There appears to be a high level of redundancy in terms of which ßC tyrosines are important for STAT5 activation, and all six conserved cytoplasmic tyrosines can mediate STAT5 activation, although to varying extents [43] ( Fig. 2). Using dominant-negative STAT5 to abolish STAT5 activity, the JAK/STAT pathway has been shown to be important in the regulation of several GM-CSF-inducible genes, including pim-1, oncostatin M, and Id-1 [57].
Although the functional significance of the induction of these genes is unclear, the JAK/STAT pathway is also important for the induction of cytokine-inducible SH2-containing protein (CIS) [57-59] and the related suppressor of cytokine signaling proteins (SOCS) [60]. These proteins have recently been identified as being involved in a negative feedback pathway which downregulates JAK2 activity following receptor activation [59, 60]. While JAK2 activation has also been found to be essential for c-myc induction, dominant-negative STAT5 did not inhibit this induction, indicating that JAK2 may activate additional, as yet uncharacterized, pathways ( Fig. 2).
The ras/MAP Kinase Pathway
Tyrosine mutants of ßC have implicated Tyr577, Tyr612, and Tyr695 in the activation of the ras/mitogen-activated protein (MAP) kinase pathway [48] ( Fig. 2). Tyr577 has been shown to interact directly with the PTB domain of the adaptor protein, SHC [43, 61, 62], while tyrosine phosphorylation of Tyr577 and Tyr612 have been shown to be important for the activation of another adaptor protein, grb2 [43]. Other studies have shown that ßC, SHC, and grb2 form a ternary complex [63] which could, in turn, recruit the guanine nucleotide exchange factor, sos, enabling activation of ras and downstream partners of the MAP kinase pathway ( Fig. 2). Furthermore, the activation of several components of the ras/MAP kinase pathway, including SHC, ras, Raf, and MAP kinase, can be inhibited by deletion of a domain-encompassing amino acid (a.a.) 626-763, indicating that the tyrosines within this domain (Tyr695 and Tyr750) may also be important in regulating this pathway [48].
The PI 3-Kinase Pathway
PI 3-kinase consists of an 85kDa regulatory subunit containing both SH2 and SH3 domains (p85) and a 110kDa catalytic subunit (p110); it is activated in response to GM-CSF and IL-3 stimulation [48, 64, 65]. Phosphorylation of phosphoinositols by PI 3-kinase generates a class of second messenger molecules such as phosphatidylinositol-3,4-bisphosphate, that in turn regulate the activity of a number of kinases and cytoskeletal proteins [66]. There is no apparent consensus for the direct binding of PI 3-kinase to the ßC; however, recruitment of PI 3-kinase to the GM-CSF and IL-3 receptors may utilize certain adaptor molecules [64] ( Fig. 2).
|
Which Signaling Pathways Emanating from ßC Are Important for Cell Proliferation, Survival, and Differentiation? |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
|
|
These results would suggest that: A) there are GM-CSF-mediated survival signals that can, at least in part, be substituted by the presence of serum (ßC626-763), and B) that Ba/F3 cells require a specific GM-CSF-mediated survival signal that cannot be substituted by serum that is mediated by a.a. 544-626 of ßC ( Fig. 3, survival domain 1). Indeed, this domain contains a region encompassing a.a. 570 to 626 that we term "box 3," which exhibits unusually high sequence identity between mouse and human ßC (72% identity) when compared with the overall identity of the cytoplasmic domain of ßC (55% identity). The requirement of box 1 and JAK2 activation for the survival response mediated by box 3 is unknown. Although JAK2 activation is required for all known signaling originating from ßC and an essential role seems highly likely, its importance in survival remains to be tested [61, 68].
Identification of a putative survival domain (box 3) raises the question: What motifs within this conserved domain are responsible for signaling, and how do they signal? Although this question addresses one of the most fundamental aspects of GM-CSF biology, the answer has remained elusive. What is apparent, however, is that the survival signals involved are likely to utilize novel pathways and signaling mechanisms. When the possible role of tyrosine phosphorylation in mediating survival signals was examined, some interesting and somewhat surprising results were obtained. Substitution of each of two tyrosines in the box 3 ßC survival domain to phenylalanine (ßCY577F and ßCY612F) had no effect on survival of Ba/F3 cells, although SHC and grb2 activation was reduced for the ßCY577F mutant [72]. Substitution of a tyrosine outside box 3 (ßCY750F) resulted in a modest decrease in GM-CSF-mediated survival in the presence of serum [47].
The dispensability of ßC tyrosine phosphorylation for cell survival is also illustrated by a ßC mutant in which all eight cytoplasmic tyrosines are mutated to phenylalanine (ßCFALL), which is able to support survival in Ba/F3 cells in response to GM-CSF [43, 50]. Clearly, the results obtained with the ßCY577F, ßCY612F, and ßCFALL mutants would suggest that ßC activation invokes novel signals that are important for inducing cell survival that cannot be solely accounted for by receptor tyrosine phosphorylation.
While the presumptive motifs within box 3 of ßC and the mechanism by which it promotes cell survival are unclear, the recent discovery of a novel PI 3-kinase pathway has provided new insights as to how cytokines such as IL-3 promote survival by the prevention of apoptosis. Although PI 3-kinase activity has been well characterized as activating a survival program in response to certain cytokines and growth factors, it was only recently that the downstream signaling partners were identified [66]. IL-3 stimulation of Ba/F3 cells results in the activation of PI 3-kinase and the production of phosphatidylinositols [48, 64]. The binding of phosphatidylinositols to protein kinase B (PKB or Akt) results in activation of its kinase activity [73, 74]. PKB in turn phosphorylates the pro-apoptotic bcl-2 homolog, BAD at Ser136 and Ser112, both of which lie within a suitable context for binding to the 14-3-3 family of adaptor proteins [75, 76].
The binding of 14-3-3 to BAD is thought to sequester the complex to the cytoplasm, thus preventing BAD from exerting its proapoptotic activity (presumably by preventing Bcl-XL binding on mitochondria) and allowing IL-3-mediated survival [77]. However, in the absence of IL-3 stimulation, the PI 3-kinase/PKB pathway is not activated, and Ser136 of BAD remains unphosphorylated. It is then proposed that nonphosphorylated BAD is no longer sequestered in the cytoplasm by 14-3-3 and translocates to the mitochondria where, in a complex with Bcl-XL, the events leading to apoptosis are triggered [77]. PI 3-kinase is thought to be recruited to phosphotyrosine residues on activated receptors via its SH2 domain [78]. An interesting question that arises from these studies into IL-3-mediated survival is: How is PI 3-kinase activated by IL-3? Although ßC does not have a tyrosine motif for the recruitment of PI 3-kinase, alternative motifs may exist. Of relevance to this, we have recently obtained evidence that Ser585 in ßC mediates the association of ßC to 14-3-3 upon phosphorylation (unpublished observations). This is potentially an important novel finding because: A) serine phosphorylation of ßC offers an alternative to tyrosine phosphorylation for coupling to specific signaling molecules; B) it has the potential to link ßC via 14-3-3 to signaling molecules that contain a 14-3-3 binding motif, and C) the location of Ser585 is precisely in the ßC region shown to be involved in cell survival and differentiation (box 3, Fig. 3). Biological experiments are currently under way, with ßCS585A mutants to define the functional role of this interaction and the regulation of Ser585 phosphorylation by GM-CSF and IL-3.
Differentiation
The ability of GM-CSF to stimulate differentiation of an M1 murine leukemic cell line expressing a series of deletion ßC mutants has also been examined [70]. M1 cells expressing human GM-CSF receptors differentiate in response to GM-CSF and exhibit many of the morphological and biochemical hallmarks of macrophages. These studies defined a domain encompassing a.a. 525-610 ( Fig. 3) that is essential for mediating the ability of GM-CSF to stimulate M1 cell differentiation [70]. It is clear from Figure 3 that the domains responsible for survival and differentiation substantially overlap and also encompass the highly conserved box 3 domain. Thus, one prediction from these findings would be that box 3 contains signaling motifs that are important for mediating survival and differentiation, with the conserved Ser585 as a possible candidate.
Together, the results obtained from studies examining proliferation, survival, and differentiation prompt the question: What is the function of ßC tyrosine phosphorylation in GM-CSF/IL-3/IL-5 receptor activation? One possibility is that within an in vivo context, tyrosine phosphorylation of ßC is important for amplifying, or in some way regulating, GM-CSF/IL-3/IL-5 signals and that such amplification or regulation is difficult to measure in cultured cell lines. Regardless, a number of laboratories have now observed a phosphotyrosine-independent mode of signaling for ßC that allows for proliferation, survival, and differentiation in certain myeloid cells.
Other cytokine receptors have been observed to promote proliferation and differentiation in the absence of tyrosine phosphorylation. Activation of the EPO receptor is essential for the proliferation, survival, and differentiation of erythroid progenitors. Mutant EPO receptors in which all eight cytoplasmic tyrosines were substituted by phenylalanine afforded reduced mitogenesis and ß-globin mRNA accumulation compared with wild-type controls [79]. Nevertheless they were able to sustain the growth and survival of the Ba/F3 cells in an EPO-dependent manner.
|
Biological Roles of GM-CSF/IL-3/IL-5 |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
Consistent with the phenotype observed in mice, humans with low levels of ßC expression on peripheral blood cells exhibited PAP with no other apparent hemopoietic disorder [83]. This pathology is characterized by an excessive accumulation of surfactant lipids and proteins in the alveolar spaces. In mice, this accumulation appears to result from a defect in clearance rather than one of overproduction [84]. While numbers of macrophages in GM-CSF -/- lungs appear normal, they are filled with abnormally large amounts of surfactant proteins and lipids [81, 82]. These observations would imply that GM-CSF is important for macrophage function in the lung. The lungs from GM-CSF -/- mice also have increased bacterial and fungal infections with occasional inflammation [82]. Local production of GM-CSF in the lungs of GM-CSF-/- mice using a transgene under the control of a surfactant promoter restores normal lung function, implying that GM-CSF is likely to function as a classical autocrine/paracrine factor within the lung, a role not previously suspected from in vitro studies [84]. The lungs represent a unique dilemma in biology: How can an organ that allows gas exchange and is in intimate contact with the environment also prevent invasion of a wide range of airborne bacteria, fungus, and other pathogens? While the molecular basis of PAP is unclear, one possibility is that GM-CSF is important for the ability of macrophages in the mammalian lung to phagocytose and degrade surfactant proteins, lipids, and other pathogens in lung alveoli.
Consistent with the proposed role of GM-CSF in inflammation, Listeria monocytogenes infection of GM-CSF -/- mice resulted in 50-fold more organisms in the spleen and liver than in wild-type mice [85]. In general, these mice demonstrated reduced long-term survival compared with wild-type controls, with a higher incidence of lung and soft-tissue infections, indicating that GM-CSF expression provides a selective advantage for the survival of the organism [86].
|
Future Directions/Conclusions |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
The in vitro activity of GM-CSF and IL-3 led to the expectation that these cytokines would be central factors for the production of granulocytes and macrophages in vivo. Recent studies using a gene knockout approach suggest that GM-CSF and IL-3 are dispensable for the steady-state production of these cells in vivo. Although much has been learned using a variety of in vitro systems, the findings in knockout mice highlight the importance of considering function within an in vivo context. Future in vivo approaches for examining some of the unique features of the GM-CSF, IL-3, and IL-5 receptors may rely on the introduction of discrete receptor mutants (knock-ins) on a null receptor background in mice.
The molecular mechanism by which GM-CSF, IL-3, or IL-5 stimulates survival/differentiation remains to be determined, but the possibility that these signals arise from box 3 independently of both proliferation and ßC tyrosine phosphorylation may allow intervention of these pathways with more specific alternatives than phosphotyrosine-based approaches. The ability to increase survival and consequently enhance an inflammatory or immune response to GM-CSF, IL-3, and IL-5 or to decrease GM-CSF autocrine loops associated with leukemia would both be of clinical benefit. Pivotal to such an approach will be the identification of the nature of the survival and differentiative signals.
|
References |
|---|
|
TopAbstractIntroductionActivation of the GM-CSF...Activation of the GM-CSF...Which Signaling Pathways...Biological Roles of GM-CSF/IL...Future Directions/ConclusionsReferences |
|---|
chain and a ß chain shared with the receptor for GM-CSF. Cell 1991;66:1175-1184.[Medline]- and ß- chains of the receptor. Blood 1994;83:3500-3508.- and ß-chains of its receptor and participate in function. J Biol Chem 1994;269:8488-8492.
