HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lennartsson, J. Articles by Shivakrupa, R. Search for Related Content PubMed PubMed Citation Articles by Lennartsson, J. Articles by Shivakrupa, R.

Normal and Oncogenic Forms of the Receptor Tyrosine Kinase Kit

Basic Research Laboratory, Center for Cancer Research, National Cancer Institute–Frederick, Frederick, Maryland, USA

Key Words. Kit • Signal transduction • Hematopoiesis • Oncogene • Cancer

Correspondence: Johan Lennartsson, Ph.D., Ludwig Institute for Cancer Research, Box 595, Biomedical Center, SE-751 24 Uppsala, Sweden. Telephone: 46-18-160406; Fax: 46-18-160420; e-mail: Johan.Lennartsson{at}LICR.uu.se

	ABSTRACT

Top Abstract Introduction Hematopoiesis Signal Transduction Kit and Human Disease Oncogenic Kit Signaling Conclusion References

 
Kit is a receptor tyrosine kinase (RTK) that binds stem cell factor. This receptor ligand combination is important for normal hematopoiesis, as well as pigmentation, gut function, and reproduction. Structurally, Kit has both an extracellular and intracellular region. Theintra-cellular region is comprised of a juxtamembrane domain (JMD), a kinase domain, a kinase insert, and a carboxyl tail. Inappropriate expression or activation of Kit is associated with a variety of diseases in humans. Activating mutations in Kit have been identified primarily in the JMD and the second part of the kinase domain and have been associated with gastrointestinal stromal cell tumors and mastocytosis, respectively. There are also reports of activating mutations in some forms of germ cell tumors and core binding factor leukemias. Since the cloning of the Kit ligand in the early 1990s, there has been an explosion of information relating to the mechanism of action of normal forms of Kit as well as activated mutants. This is important because understanding this RTK at the biochemical level could assist in the development of therapeutics to treat primary and secondary defects in the tissues that require Kit. Furthermore, understanding the mechanisms mediating transformation of cells by activated Kit mutants will help in the design of interventions for human disease associated with these mutations. The objective of this review is to summarize what is known about normal and oncogenic forms of Kit. We will place particular emphasis on recent developments in understanding the mechanisms of action of normal and activated forms of this RTK and its association with human disease, particularly in hematopoietic cells.

	INTRODUCTION

Top Abstract Introduction Hematopoiesis Signal Transduction Kit and Human Disease Oncogenic Kit Signaling Conclusion References

 
Kit and Stem Cell Factor
The v-kit oncogene was identified in 1986 as the transforming gene in the Hardy-Zuckerman 4 feline sarcoma virus [1]. One year later, the cellular counterpart, c-kit, was identified based on sequence similarities [2]. The protein product of the c-kit gene will be referred to as Kit in this article. Kit is a receptor tyrosine kinase (RTK) belonging to subclass III, which also includes platelet-derived growth factor receptor (PDGFR), PDGFRß, Flt3, and the CSF1R (c-Fms and M-CSFR). This subclass is characterized by an extracellular region consisting of five immunoglobulin (Ig)-like domains, a single transmembrane domain, and an intracellular tyrosine kinase domain split in two by an insert region (Fig. 1). Studies on the extracellular domain have shown that the first three N-terminal Ig-like domains of Kit are involved in ligand binding [3, 4]. Alternative splicing of Kit results in the presence or absence of four amino acids (GNNK) just outside the plasma membrane, and these isoforms exist in both mice and humans [5]. In humans, additional splice forms of Kit exist, which are characterized by the presence or absence of a serine residue in the kinase insert [6].

View larger version (25K): [in this window] [in a new window]   Figure 1. Schematic illustration of the Kit receptor. The receptor is drawn to indicate the domain structure of Kit. Phosphorylated tyrosine residues (in black) in the human Kit and their number are indicated. The corresponding numbers in the murine receptor are designated in parenthesis. Activating mutations are shown in red. Abbreviation: Ig, immunoglobulin.

The White Spotting and Steel Loci
In 1988 it was established that c-kit is allelic with the White spotting (W) locus [8, 9]. Several naturally occurring loss-of-function mutations have been found in the c-kit gene. In mice with alleles that result in complete loss of Kit expression, the homozygotes die in utero or perinatally, likely because of anemia. In the heterozygous state, phenotypes include defects in pigmentation and reduced fertility. Mutations that affect kinase activity often result in death in utero or shortly after birth in the homozygous state, whereas heterozygous mice are anemic and have reduced fertility. In general, the severity of the phenotype correlates with the magnitude of the decrease in kinase activity of Kit [10].

In 1990, the product of the Steel (Sl) locus was demonstrated to be identical to the ligand for Kit [11, 12]. Naturally occurring mutations in the Sl locus lead to complete or partial deletion of the gene as well as reduced SCF expression. The phenotypes of the Sl mutants are in most cases similar to those of W mutant mice; complete loss of SCF results in death in utero or perinatally, whereas heterozygous mice display less-severe phenotypes, with defects in fur color and mild anemia [10].

The phenotypes of W and Sl mice suggest that Kit signaling is important for normal hematopoiesis, gametogenesis, and melanogenesis, and studies of Kit function have indeed indicated roles for the receptor in these processes.

	HEMATOPOIESIS

Top Abstract Introduction Hematopoiesis Signal Transduction Kit and Human Disease Oncogenic Kit Signaling Conclusion References

 
During mammalian embryonic development, primitive hematopoiesis occurs in the yolk sac and in the aorta-gonad-mesonephros region [13]. Subsequently, hematopoietic stem cells (HSCs) migrate from these regions to the embryonic liver, thymus, and spleen. Kit is expressed by hematopoietic cells in the embryonic liver throughout its development. In contrast, expression of SCF is more restricted during this time [14]. Around the time of birth, the major site for hematopoiesis becomes the bone marrow, and this is maintained during adult life.

Most mature blood cells are short lived and must be continuously replenished through out life. HSCs are rare cells with the capacity to self-renew and give rise to all blood lineages, including lymphocytes, megakaryocytes, granulocytes, monocytes, and erythroid cells. CD34 is a transmembrane glycoprotein of unknown function that is expressed on hematopoietic stem/progenitor cells [15]. Most of the CD34⁺ cells in the bone marrow also express Kit, suggesting an important role for Kit in primitive hematopoietic cells [16]. Most HSCs remain in a quiescent state, but they can be mobilized in response to stress on the hematopoietic system. Specific cytokines in combination induce HSC proliferation and differentiation. In vitro SCF in combination with Tpo has a central role in HSC proliferation [17]. In addition, HSC mobilization involves interactions between SCF and Kit [18]. Furthermore, G-CSF in combination with SCF is a particularly potent means to mobilize stem/progenitor cells in both humans and mice [19]. Physiological regulation of HSC growth, differentiation, and mobilization is a complicated process that is likely mediated by combinations of factors, including several cytokines and components of the extracellular matrix [20].

HSCs have the capacity to differentiate into multipotential progenitor cells, thus initiating the process of lineage commitment. These latter cells have a more restricted ability to differentiate than stem cells and are functionally defined by their ability to form colonies in vitro. In the absence of Kit, other pathways can compensate and produce small numbers of HSCs as well as more differentiated hematopoietic cells. For example, HSCs can be maintained on stromal cells in the presence of an antibody that blocks SCF binding to Kit. This suggests that stromal cells can support HSCs in the absence of functional Kit signaling [21]. Indeed, Epo receptor (EpoR) signaling partially supported colony-forming unit–erythroid formation in W/W mice lacking functional Kit [22]. In addition, other growth factors may compensate for loss of Kit. Hepatocyte growth factor (HGF) promotes proliferation and differentiation of the HSCs and progenitor cells from W/W mice [23]. HGF in combination with GM-CSF, G-CSF, or M-CSF induces colony formation of CD34⁺ cells from human cord blood [24]. Furthermore, insulin-like growth factor-I (IGF-I) in combination with G-CSF can maintain proliferation of CD34⁺ cells isolated from human cord blood. The number of differentiated myelopoietic cells, as measured by increased expression of the granulocytic marker CD15 and the myelomonocytic marker CD64, increased in the presence of IGF-I in combination with G-CSF [25].

Currently, no single marker specific for HSCs has been found. Consequently, these cells are often purified using combinations of markers. A common strategy for enrichment of murine HSCs has been to use lineage-negative cells that express Sca-1, Kit, and CD34 [26]. Progenitor cells differentiate into precursor cells, which are committed to a specific lineage and have no capacity for self-renewal. However, precursor cells have a high proliferative potential, leading to extensive amplification of cell numbers. The precursor cells differentiate further into fully mature blood cells. Although the mechanisms regulating these events are not fully understood, a variety of cytokines and growth factors play critical roles. Moreover, the hematopoietic microenvironment also plays an important role through structural support and adhesive interactions required for colocalization of primitive cells and regulatory factors. In fact, differentiation and proliferation of hematopoietic progenitor cells can be controlled by communication with stromal cells in the microenvironment [27]. Furthermore, it has been demonstrated that autocrine/paracrine loops, including Kit and SCF, are important for growth of primitive hematopoietic cells [28, 29]. Because these cells secrete a multitude of cytokines and growth factors, it is likely that synergistic interactions between these factors are important for survival and proliferation. For further information on autocrine/paracrine loops in normal and oncogenic hematopoiesis, please refer to the study by Janowska-Wieczorek et al. [30].

Kit is also expressed on more committed progenitors, including myeloid, erythroid, megakaryocytic, natural killer, and dendritic progenitor cells, as well as pro-B and -T cells and mature mast cells [31, 32]. The expression of Kit is low on the earliest HSC and peaks in more mature progenitor cells. Interestingly, SCF is a potent chemoattractant for this population of cells [33]. As a single factor, SCF acts primarily to promote survival of hematopoietic progenitor and stem cells [34, 35]. However, it is insufficient to maintain self-renewal in vitro [36]. A recent study demonstrated that activation of JAK2 in combination with signals emanating from Kit or Flt3 promoted self-renewal of primary multipotential hematopoietic cells [37]. Interestingly, progenitor cells with different characteristics were obtained when Kit was activated instead of Flt3. Differentiation of HSCs to progenitor cell populations is, at least in some cases, linked to increased mitogenic activity of SCF [36].

Although Kit expression is generally lost during hematopoietic cell differentiation, some differentiated cell types, such as mast cells, retain this RTK (Fig. 2). SCF induces mast cell differentiation from primitive fetal liver hematopoietic cells [38]. Furthermore, SCF supports proliferation, survival, and final maturation of mast cells in vitro, as well as induces chemotaxis, adhesion, and survival of mature mast cells [39]. One striking phenotype of viable mice with W and Sl alleles is the deficiency in tissue mast cells [40]. Kit is not expressed on differentiated lymphocytes. Hence, in the lymphoid lineage, SCF acts primarily on lymphoid progenitors. Because mice with severe mutations in Kit or SCF die early in development, little is known about the effects these mutations have in adult animals. However, a recent study showed an age-dependent decline in the number of pro-T and pro-B cells, as well as reductions in common lymphoid progenitors in the bone marrow of the viable W/W mouse mutant Vickid [41]. The identity of the compensatory mutation that allows the Vickid mice to survive in the absence of Kit and the molecular mechanism by which this occurs are not presently known.

View larger version (23K): [in this window] [in a new window]   Figure 2. The hematopoietic tree. This figure illustrates the differentiation of hematopoietic cells and expression of Kit during this process. Abbreviations: HSC, hematopoietic stem cell; CMP, common myeloid progenitor; CLP, common lymphoid progenitor; CFU, colony-forming unit; Kit+, Kit is expressed; Kit–, Kit is not expressed; Kit+/–, Kit is expressed on certain subpopulations.

	SIGNAL TRANSDUCTION

Top Abstract Introduction Hematopoiesis Signal Transduction Kit and Human Disease Oncogenic Kit Signaling Conclusion References

 
The focus of this section is on biochemical pathways activated by wild-type Kit, with special consideration given to mechanisms in hematopoietic cells. Table 1 and Figure 3 are summaries of signaling proteins that interact with Kit.

View larger version (27K): [in this window] [in a new window]   Figure 3. Schematic illustration of Kit and interacting proteins. This figure summarizes signaling proteins activated by Kit and interaction sites on the receptor. Abbreviations: SCF, stem cell factor; SFK, Src family kinases.

The soluble and membrane-associated forms of SCF have different effects on Kit autophosphorylation. Stimulation with the soluble form results in rapid but transient auto-phosphorylation of Kit, whereas stimulation with the membrane-bound form results in more sustained phosphorylation [57]. Interestingly, soluble and membrane-associated SCFs have overlapping but distinct functions in vivo. This was clearly demonstrated in a mutant mouse strain that expresses only the soluble form of SCF (Sl^d) [58]. The homozygous Sl^d mice are viable but have severe anemia, are sterile, and lack pigmentation, and their skin is deficient of mast cells. In mice with a disrupted SCF gene (Sl), the homozygotes die in utero or perinatally [10]. Moreover, in an erythroid progenitor cell line, stimulation with membrane-associated SCF results in higher proliferation compared with soluble SCF [59]. Biochemical analysis showed that this correlated with more persistent phosphorylation of Erk1/2 and p38 mitogen-activated protein kinase (MAPK) after stimulation with membrane-associated SCF compared with soluble ligand [60].

In summary, ligand-induced dimerization of Kit monomers increases the local concentration of kinase domains, thus facilitating autophosphorylation. Most of the tyrosine phosphorylation sites are located outside the kinase domain, and these function as docking sites for intracellular signal transduction proteins.

The Phosphoinositide 3'-Kinase Pathway
Phosphoinositide 3'-kinase (PI3K) phosphorylates the 3'-hydroxyl group in the inositol ring of phosphati-dylinositol (PtdIns), phosphatidylinositol-4-phosphate [PtdIns(4)P], or phosphatidylinositol-4,5-bisphosphate [PtdIns(4,5)P₂] in vitro. However, in vivo, the preferred substrate is PtdIns(4,5)P₂, and this subsequently generates PtdIns(3,4,5)P₃ [61]. PI3K is a heterodimeric complex with a regulatory (p85) and a catalytic (p110) subunit [62]. PI3K has been implicated in many aspects of cellular signaling, such as DNA synthesis, cell survival, membrane ruffling, and chemotaxis, as well as receptor and vesicular trafficking. Activation of PI3K involves association of the SH2 domain of the regulatory subunit, with phosphorylated receptors leading to translocation of PI3K to the plasma membrane, where its sub-strates are located. Downstream targets for 3'-phospholipids generated by PI3K include the serine/threonine kinases Akt and protein kinase C (PKC). Akt is important for antiapoptotic signaling. The PH (Pleckstrin homology) domain of Akt interacts with phospholipids generated by PI3K, resulting in its localization to the plasma membrane, where it is activated by serine and threonine phosphorylation.

SCF induces association of the SH2 domain of the 85-kDa regulatory subunit of PI3K with tyrosine residue 721 in the kinase insert of human Kit (Y719 of murine Kit). The role of this interaction in SCF signaling in hematopoietic cells has been examined by several groups, and many of the earlier studies were reviewed previously [63]. In brief, infection of murine mast cells lacking endogenous Kit with Kit^Y719F demonstrated that interaction with PI3K was required for optimal membrane ruffling, adhesion, actin assembly, and proliferation in response to SCF [64]. In contrast, a recent study showed that restoration of the PI3K binding site in a Kit receptor where all autophosphorylation sites were mutated to phenylalanine led to only minor increases in cell survival and migration and had no effect on proliferation [65]. A human Kit^Y721F mutant was unable to protect against Bad-induced apoptosis in U2-OS cells [66]. Similar results were obtained with bone marrow mast cells (BMMCs) expressing the Kit^Y719F mutant. On the cellular level, there were reductions in SCF-induced proliferation and protection from apoptosis. Biochemically, this resulted in reductions in SCF-induced activation of Akt, as well as c-Jun NH2-terminal kinase (JNK) in BMMCs [67]. These data, in conjunction with earlier work, demonstrate that direct association of PI3K and Kit contribute to, but are not required for, most SCF-mediated responses. Studies from Besmer and colleagues [68] suggest that the Kit binding sites for both PI3K and Src family kinases (SFKs) are required for most SCF-mediated responses. Consistent with this notion, Ueda et al. [69] demonstrated that PI3K and SFK binding sites were the only tyrosines in the Kit cytoplasmic domain sufficient for SCF-induced chemotaxis and calcium mobilization.

In nonhematopoietic tissue, studies with tyrosine 719 knock-in mice demonstrated that direct interaction of PI3K and Kit is vital for mouse fertility [67, 70]. Sterility in males was attributable to a block in spermatogenesis with decreased proliferation and extensive apoptosis. Reductions in fertility in females were caused by impaired follicle development. A recent study supports an important role for Akt in survival and growth of primordial germ cells [71].

Data with p85^PI3K-deficient mice support an important role for PI3K in SCF-mediated responses beyond that mediated by direct interaction of Kit with p85. SCF-induced proliferation of BMMCs from p85-deficient mice was dramatically reduced compared with wild-type cells. This decrease in proliferation correlated with a defect in the SCF-induced activation of JNK in the p85-deficient BMMCs. This is consistent with previous studies demonstrating that activation of the JNK pathway in BMMCs is required for SCF-induced proliferation [68, 72]. Surprisingly, SCF-induced activation of Akt was only partially affected in p85-deficient BMMCs, and no defects in SCF-mediated survival were observed in this study. Similar results were obtained in fetal liver mast cells from p85-deficient mice [73]. Thus, other pathways must be contributing to survival mediated through Kit. Kapur and colleagues [74] demonstrated using mast cells from p85 or Rac2-deficient mice that SCF-induced chemotaxis is dependent on signal transduction involving these proteins. Furthermore, they showed that the 4 integrin and Kit cooperate in promoting directed migration of mast cells.

The differences in results using p85-deficient cells and cells expressing the Kit^Y719F receptor mutant are of note. In the knock-in studies, other proteins that potentially bind to tyrosine 719 are unable to interact. In addition, p85 can associate indirectly with other signaling components, independent of interaction with the PI3K docking site on Kit. Thus, further study using both these important models will be required to define biochemical mechanisms underlying the contribution of PI3K to SCF-mediated responses.

The JAK-STAT Pathway
The Janus kinases (JAKs) are cytoplasmic tyrosine kinases that are rapidly activated by ligand binding to cytokine receptors or RTKs. Important targets for activated JAKs are the signal transducers and activators of transcription (STAT) [75]. STATs are a group of transcription factors that have a central DNA binding domain followed by a SH2 domain and a C-terminal transactivation domain. Activated JAKs or RTKs phosphorylate the cytoplasmic STATs, which then undergo a SH2 domain–dependent dimerization and subsequent translocation to the nucleus. Nuclear STATs regulate expression of responsive genes by binding to their promotors. STATs (at least 1, 3, and 5) are also serine phosphorylated at the C-terminus, which regulates their transcriptional activity.

SCF stimulation results in activation of the JAK-STAT pathway [76, 77]. JAK2 constitutively associates with Kit and becomes transiently activated in response to SCF [78]. Studies using antisense oligonucleotides specific for the JAK2 sequence suggest that it is important for maximal SCF-induced proliferation [77]. Recently, we showed that partially purified populations of fetal liver hemato-poietic progenitor cells are dependent on JAK2 for proliferation and differentiation in response to SCF [79]. After SCF stimulation, STAT1, STAT5A, and STAT5B associate with Kit and become tyrosine phosphorylated [80]. SCF-induced STAT tyrosine phosphorylation has been associated with increased DNA binding [81, 82]. In addition to tyrosine phosphorylation, SCF also induces serine phosphorylation of STAT3 [83]. Interestingly, different regions of Kit are involved in the activation of STAT1 and STAT5A/B [80]. Complete activation of STAT5, but not STAT1, requires the C-terminal region of Kit.

In contrast to the above findings, several studies have not observed activation of JAK-STAT pathway by SCF [84–86]. One potential explanation for these differences is lineage-specific differences in JAK-STAT activation. Consistent with this possibility, we recently showed that JAK2 is important for SCF-induced growth of Kit-positive progenitor cells but not critical for mast cell response to SCF [79]. A second potential explanation is differences in the experimental procedures. For example, many studies looked at only one time point after stimulation with SCF. In contrast to interleukin 3 (IL-3), SCF-induced activation of this kinase is extremely rapid and transient [77, 78]. In addition, JAK2 is rapidly dephosphorylated in the absence of high levels of phosphatase inhibitors [78]. Thus, differences in the components of the lysis buffer can have a dramatic effect on the ability to detect phosphorylation of JAK2.

The Ras-Erk Pathway
Activation of the Ras-Erk pathway is of critical importance for several cellular responses such as cell division and survival [75]. Ras has two conformations, an active guanine triphosphate (GTP)-associated form and an inactive guanosine diphosphate (GDP)-associated form. RTKs activate Ras by associating with Sos, a guanine nucleotide exchange factor that favors the guanosine diphosphate–bound state. Sos associates constitutively with the adaptor protein Grb2, which in turn associates with activated RTKs via its SH2 domain either directly or indirectly via other molecules such as SHP2 and ShcA. The active receptor recruits the Grb2/Sos complex, bringing it close to the plasma membrane where Ras is located. Activated Ras can interact with and initiate the activation of the serine/threonine kinase Raf-1. Active Raf-1 can phosphorylate and activate the dual-specificity kinases Mek1 and 2. However, Raf-1–deficient cells are still able to activate Erk1/2, suggesting a redundancy with other kinases such as A-Raf and B-Raf. The mitogen-activated protein kinases Erk1 and Erk2 are serine/threonine kinases that are activated through phosphorylation by Mek1/2. Targets for Erks in the cytoplasm include the serine/threonine kinases RSK and Mnk1/2, as well as phospholipase A₂. A fraction of the activated Erk and Ribosomal S6 kinase(RSK) proteins translocate to the nucleus, where they phosphorylate various transcription factors, thereby influencing gene transcription. A recent study showed that disruption of the Erk2 gene is embryonic lethal despite the presence of Erk1 [87]. Thus, Erk1 and Erk2 have unique nonredundant functions during mouse development.

It has been well established that SCF activates the Ras-Erk pathway through multiple signaling components. The adaptor protein Grb2 can interact directly with tyrosine residues 703 and 936 of human Kit or indirectly via proteins that associate with Kit such as ShcA or SHP2 [88–90]. Some studies, but not all, have demonstrated that SCF-induced activation of SFKs is important for activation of the Ras-Erk cascade [68, 69, 91, 92]. In a recent study, Kapur and colleagues [65] showed that activation of the Ras-Erk cascade requires only an intact SFK binding site in Kit. Consistent with these findings, knock-in mice expressing a Kit receptor lacking the SFK interaction site are unable to activate the Erk cascade [93]. Variability relating to the role of SFKs in activation of the Ras-Erk pathway may relate to differences in experimental procedures, lineage specificity, and which GNNK splice form of Kit that is predominantly expressed. In addition, Ras can be activated by the guanine nucleotide exchange factor Vav, which is phosphorylated in response to SCF in hematopoietic cells [94, 95]. The adaptor protein Gab2 can link to the Ras-Erk pathway through association with SHP2 [96]. Indeed, Gab2 was demonstrated to be involved in the activation of the Ras-Erk pathway after SCF stimulation of mast cells [97].

Activation of Erk1/2 is important for hematopoietic differentiation [98]. This was demonstrated in the UT-7/GM cell line, which can differentiate into erythroid or mega-karyocytic lineages. Transient activation of Erk1/2 by Epo results in erythroid differentiation. In contrast, Erk1/2 activation by Tpo is sustained and induces megakaryocytic differentiation. Furthermore, pharmacological inhibition of the Ras-Erk pathway with PD98059 promotes Epo-induced erythroid differentiation and suppresses Tpo-induced mega-karyocytic differentiation in UT-7/GM cells. Hence, this study shows that the duration of Erk1/2 activity is important for the biological outcome. Furthermore, inhibition of the Ras-Erk pathway with PD98059 suppresses megakaryocytic differentiation from CD34⁺ hematopoietic progenitor cells from human cord blood [99].

The SFK Pathway
SFKs are involved in a wide range of cellular functions, including cell cycle progression, chemotaxis, adhesion, survival, and protein trafficking [100]. Gene disruption studies have demonstrated that SFKs are essential for normal development and that there is a great deal of functional redundancy between the different family members.

Activation of Kit results in a rapid increase in SFK activity [101, 102]. Furthermore, SFKs associate with tyrosine 568 and to some extent with tyrosine 570 in the juxta-membrane domain (JMD) of human Kit (Y567 and Y569 in murine Kit) through their SH2 domain [68, 91, 103].

SFKs are important for ligand-induced internalization of Kit [54, 104]. This was demonstrated by the use of point mutants of the receptor and pharmacological inhibitors of SFKs and in Lyn-deficient cells. Moreover, overexpression of c-Src results in an increased rate of epidermal growth factor receptor (EGFR) internalization, suggesting a conserved role for SFKs in internalization of members of the RTK super family [105].

The involvement of SFK in Kit-induced mitogenicity is not clear. In some studies, SFKs have no or only a partial effect on proliferation, whereas other studies have shown them to be essential [68, 91, 101, 106, 107]. Tan et al. [107] used chimeric Kit receptor mutants to show that inactivation of the SFK pathway had the most severe effect on proliferation compared with other signaling pathways. The same study also showed that a Kit mutant that only had the SFK binding site intact restored proliferation to 50% of that induced by wild-type Kit. SFKs have been shown to phosphorylate Kit on tyrosine residue 900 within the second part of the kinase domain [108]. Mutating tyrosine 900 to phenylalanine significantly inhibited SCF-induced DNA synthesis compared with wild-type Kit in transfected NIH3T3 cells. Tan et al. [107] could also demonstrate a reduction in c-Myc expression using a chimeric Kit receptor in which the SFK binding sites had been mutated compared with the chimeric receptor with the wild-type intracellular domain of Kit. In contrast, no defect could be seen in the mutated receptor’s ability to induce STAT5 phosphorylation or Bcl-x_L expression. In addition, this study also demonstrated that activation of Src family kinases plays a crucial role in the proliferative cooperation between Kit and the EpoR.

The SFK member Lyn has an important role in SCF-induced chemotaxis in primary hematopoietic cells [106]. Furthermore, a recent study in which all individual tyrosine residues in the intracellular part of murine Kit were mutated to phenylalanine showed that tyrosine residues 567 and 719 were important for SCF-induced chemotaxis and Ca²⁺ mobilization [69]. Restoring individual tyrosine residues in a receptor in which all tyrosines had been mutated to phenylalanine demonstrated that the SFK binding site in Kit is critical for survival and migration [65]. However, this receptor mutant only partially restored proliferation compared with wild-type Kit.

Besmer et al. [109] recently generated Kit^Y567F knock-in mice. Homozygous Kit^Y567F mice displayed normal hema-topoiesis but defective lymphocyte development in older animals with a block at the pro-B and pro-T stage. These results are in agreement with studies on Vickid mice, in which loss of Kit signaling led to an age-dependent reduction in pro-B and pro-T cells [41].

Two splice forms of Kit differ only by the presence or absence of four amino acids in the extracellular juxtamem-brane region. However, these splice forms display major differences in kinetics of autophosphorylation and signaling capabilities [110]. Recently, the shorter splice form was reported to bind and activate SFKs more efficiently [111]. This led to stronger phosphorylation of ShcA, Erk1/2, and Cbl compared with the longer splice form. The difference in kinetics of receptor autophosphorylation was also dependent on SFK activation.

Knock-in of a Kit receptor lacking the SFK binding site in mice resulted in decrease in receptor phosphorylation and a selective inhibition of the Erk phosphorylation [93]. Further analysis demonstrated these mice have defects in the development of mast cells and in the melanogenesis but normal development of erythroid blood cells, germ cells, and interstitial cells of Cajal (ICC).

Recent studies with knock-in mice have demonstrated the importance of cellular context on signaling pathways activated downstream of Kit [93, 109]. For example, loss of SFK binding(Kit^Y567F)leads to defects primarily in lymphopoiesis, whereas mice harboring Kit^Y567F/Y569Falso had major defects in pigmentation, mast cell development, and splenomegaly [93, 109]. Defects in pigmentation and reduced numbers of mast cells were also observed in mice with single tyrosine mutations but were more severe in the double tyrosine mutant [93]. This indicates that these two tyrosines in the JMD of Kit cooperate to activate certain signaling pathways. It is interesting to note that in the hematopoietic system, Kimura et al. [93] report age-dependent defects of B-cell development and megakaryopoiesis, whereas the study by Agosti et al. [109] found defects in B and T cells. This could be due to differences in the genetic background, because Kimura et al. studied the phenotype on mixed 129/SvCP and C57 BL/6 background, whereas the work from Agosti et al. used C57 BL/6J. Another potential explanation is difference in the age of mice when they were analyzed. In contrast, loss of PI3K binding (Kit^Y719F) resulted in defects in gametogenesis but not lymphopoiesis, as seen for loss of SFK binding. Thus, signaling pathways have distinct roles in different cell lineages.

Phospholipase C
Two isoforms of phospholipase C (PLC) have been identified: PLC1, which is expressed ubiquitously, and PLC2, which is mainly expressed in hematopoietic cells [112]. PLC hydrolyzes phosphatidylinositol-4,5-bisphosphate (PIP2), generating diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP3). DAG binds and activates conventional and novel PKCs, whereas IP3 induces release of Ca²⁺ from intracellular stores. A variety of cellular functions are regulated by Ca²⁺, including membrane localization of conventional PKCs. The PLC association with Kit is dependent on tyrosine residues 730 and 936 of the human receptor [113, 114]. Studies using soluble SCF have in most cases failed to detect an increase in activity or phosphorylation of PLC [115, 116]. These studies have instead suggested activation of phospholipase D after SCF stimulation. In contrast, studies using the membrane-bound form of SCF have indicated a role for PLC in SCF-induced proliferation and survival [114, 117]. This suggests a differential effect on signal transduction by soluble and membrane-bound SCF. How this occurs is not clear, but one possibility is that the membrane-bound SCF results in slower internalization of active Kit, whereas soluble SCF induces rapid internalization [57, 118, 119]. The difference in localization of the active receptor could explain the differences in coupling to PLC and potentially to other molecules. It is well established that SCF protects Kit-expressing cells against apoptosis [66, 120–122]. Studies have demonstrated that PLC1 has a central role in SCF-induced protection against apoptosis induced by irradiation as well as the cytotoxic agent daunorubicin [123, 124].

Adaptor Proteins
Adaptor proteins can be defined as proteins lacking enzymatic activity but with the ability to simultaneously interact with at least two proteins. The capability of linking proteins together via specific and in many cases inducible protein–protein interactions offers multiple flexible mechanisms to regulate signal transduction.

Grb2 is a widely expressed adaptor protein that couples RTKs to the Ras-Erk pathway. This occurs through association with the Ras guanine nucleotide exchange factor Sos. Activated Kit has been shown to bind Grb2 [88]. The interaction sites of the Grb2 SH2 domain have been mapped in vitro to tyrosine residues 703 and 936 in human Kit [90]. However, the contributions to Ras-Erk pathway activation in vivo by direct binding of Grb2 to Kit compared with indirect contributions through other adaptors, such as ShcA or SHP2, remains to be defined. Grap is an adaptor protein related to Grb2. Studies on Grap have suggested a role analogous to Grb2, i.e., coupling Sos to activated receptors, including Kit [125].

ShcA is an adaptor protein that interacts with Grb2 after phosphorylation. Thus, ShcA can mediate recruitment of Grb2 to receptors that lack direct Grb2 binding sites. Activation of Kit results in increased tyrosine phosphorylation of ShcA, which correlates with the activation of the Ras-Erk pathway [91, 126]. However, ShcA likely serves additional functions in signal transduction through association with proteins such as Gab1 and the hematopoietic protein Gads [127, 128]. In vitro studies have suggested that ShcA interacts with the JMD of Kit [103].

Grb10 association with activated Kit requires SH2 domains [129]. Furthermore, this adaptor protein forms a constitutive complex with Akt and is a cofactor in SCF-induced activation. Grb10 belongs to the Grb7 family of adaptor proteins. In the case of Grb7, the interaction site with Kit has been mapped in vitro to tyrosine residue 936 in the C-terminal tail of the receptor [90].

Adaptor protein containing PH and SH2 domains (APS) is an adaptor protein cloned in a yeast two-hybrid screen using the activated mutant Kit^D816V as bait [130]. APS becomes phosphorylated in response to SCF but also by stimulation with growth factor and cytokines such as PDGF-BB, IGF-I, IGF-II, and GM-CSF [131]. The SH2 domain of APS interacts with tyrosine residues 568 and 936 in human Kit [132]. Phosphorylation of APS leads to its association with the ubiquitin ligase Cbl. Recruitment of Cbl through APS to the PDGFRß leads to inhibition of PDGF-BB–induced proliferation [131]. Lnk is a protein structurally related to APS but expressed mainly in hematopoietic cells. Lnk^–/– mice have increased numbers of hematopoietic progenitor cells in part because of increased signaling from Kit [133].

The adaptor protein Dok-1 was originally discovered as a tyrosine-phosphorylated protein associated with RasGAP [134]. Several growth factors, including SCF, induce Dok-1 tyrosine phosphorylation. Recently, it was demonstrated that SCF-induced Dok-1 phosphorylation was dependent on PI3K as well as the SFK Lyn [135, 136]. Liang et al. [136] showed that Dok-1 associated readily with the JMD as well as the C-terminal tail of Kit. Phosphorylated Dok-1 forms a complex with the kinases Lyn and Tec as well as other as-yet unidentified molecules and hence functions as an adaptor protein [135]. Studies on Dok-1^–/– mice have suggested that Dok-1 is a negative regulator of proliferation [137, 138].

SCF stimulation of MO7e cells leads to tyrosine phosphorylation of CrkL [139]. Sattler et al. [139] showed CrkL associates with Kit indirectly through the p85 subunit of PI3K. Furthermore, CrkL recruits Cbl to the Kit complex, likely downregulating its activity by increasing ubiquitination and degradation of this RTK. In NIH3T3 cells transfected with Kit, SCF induces tyrosine phosphorylation of Crk-II, a protein related to CrkL [108]. Moreover, SCF-induced Crk-II phosphorylation was dependent on tyrosine residue 900 in Kit.

The adaptor protein Gab2 becomes tyrosine phosphorylated in response to SCF [140]. Nishida et al. [140] showed that Gab2 forms a complex with SHP2 and PI3K and that in vitro Gab2 is a substrate for SHP2 [140]. A study using BMMCs derived from Gab2^–/– mice demonstrated that Gab2 is important for SCF-induced activation of Erk1/2 and Akt [97]. BMMCs from Gab2^–/– mice grow poorly in response to SCF, indicating an important role for Gab2 in mast cells proliferation.

These data highlight the importance of adaptor proteins in amplification and integration of signals downstream of Kit. They contribute to activation of proteins that cannot directly associate with the receptor as well as provide multiple means to activate a particular pathway. For example, the Ras-Erk pathway can be activated by association of Grb2 directly with the receptor or through proteins such as ShcA and SHP2.

Synergistic Signaling
An important feature of SCF is its strong capability to synergize with ligands for the cytokine receptor super family members [141]. This includes GM-CSF, IL-3, and Epo. Although the subject of rather intense study, the molecular mechanisms mediating synergy of SCF in combination with these cytokines remain to be determined. Below is a summary of some of the salient features of published work on this intriguing phenomenon.

Most studies indicate it is unlikely that synergistic growth of SCF in combination with other growth factors is mediated by alterations in expression or affinity of either population of receptors. Several studies suggest that SCF does not alter the number or affinity of GM-CSF or IL-3 receptors in either human or murine cell lines [142–144]. Furthermore, neither IL-3 nor GM-CSF influences Kit mRNA production [145]. However, GM-CSF has been reported to downregulate Kit expression in FDC-P1 cells [146]. Although receptor expression is not altered by SCF, Kit does interact with some cytokine receptors. This includes both the IL-7 receptor as well as the Epo receptor. Furthermore, our preliminary findings suggest that Kit also interacts with the ß-chain subunit of the GM-CSF receptor (unpublished data).

Several groups have examined tyrosine phosphorylation of receptors after costimulation with two growth factors. In general, GM-CSF and SCF costimulation does not induce synergistic increases in Kit tyrosine phosphorylation, although one group did report an increase in tyrosine phosphorylation of Kit in HML-2 cells stimulated with both SCF and GM-CSF compared with SCF alone [143, 147, 148]. Although one report indicates that SCF induces serine and threonine phosphorylation of the ß-chain common for GM-CSF, IL-3, and IL-5 receptor complexes in a PKC-dependent manner, the significance of this finding is not understood [144].

Downstream of Kit and the GM-CSF receptor, most early signaling pathways are not synergistically activated by SCF and GM-CSF. This includes JAK family members, PLC, Akt, Jnks, and p38 MAPK [84, 143, 149, 150]. However, several studies have implicated activation of Erk1/2 in synergistic signaling induced by SCF in combination with either GM-CSF or IL-3 [84, 85, 142, 143, 151, 152]. Broxmeyer et al. [149] showed that SCF and GM-CSF synergistically activate Erk but not the stress activated protein kinase (SAPK)/JNK and p38 MAPK pathways in MO7e cells. Similar findings have been reported in normal erythroid progenitors stimulated with SCF in combination with Epo [152].

Downstream of early signaling pathways, which are activated within minutes of costimulation, are proteins involved in cell cycle progression. Concurrent stimulation with SCF and GM-CSF synergistically induces phosphorylation of the Rb protein as well as increasing the cyclin-dependent kinase inhibitor Cip-1 and decreasing the Kip-1 expression [153]. SCF in combination with IL-3 induces the expression of the 68-kDa calmodulin-binding protein (CaM-BP68) [154]. Both Rb phosphorylation and expression of CaM-BP68 are required for growth factor–induced proliferation.

The studies above have focused primarily on SCF in combination with GM-CSF or IL-3 with regards to mechanisms mediating synergy. However, Epo in combination with SCF has also been examined in some detail. As mentioned above, Kit physically interacts with the EpoR. Kit also induces phosphorylation of the EpoR and vice versa [155]. It was recently shown that a Kit mutant unable to activate SFKs could not efficiently mediate phosphorylation of the EpoR [107]. Studies on the EpoR have demonstrated that a truncated recept or that only retains the STAT5 binding site (tyrosine residue 343) can support Epo-induced proliferation and erythroid development as efficiently as the wild-type EpoR [156]. In addition, the truncated receptor still synergizes with Kit. Studies on the molecular mechanism mediating synergy in FDCER cells have identified both EpoR tyrosine phosphorylation–dependent and –independent contributions [157]. A mutated EpoR with no intracellular tyrosine phosphorylation sites could still synergize with Kit in proliferation. In contrast, the same study also showed that SCF-induced Bcl-x_L gene expression was dependent on tyrosine residue 343 in the EpoR. Furthermore, it was also demonstrated that activation of the EpoR initiates a negative feedback loop involving SOCS-3 (suppressor of cytokine signaling-3) that inhibits both Kit and EpoR-induced proliferation [157]. Studies using G1E-ER2 cells revealed that activation of Kit is critical for maintenance of EpoR and STAT5 protein expression [158]. Subsequently, the EpoR activates STAT5 that in turn induces Bcl-x_L expression, ultimately leading to cell survival. A recent study in AS-E2 cells demonstrated that SCF can increase Epo-induced STAT5 transactivation through a mechanism that does not involve serine or tyrosine phosphorylation of STAT5 [159]. However, the increase in STAT5 transactivation was dependent on SCF-induced protein kinase A activity and CREB phosphorylation. Thus, the synergistic interaction between Kit and the EpoR occurs at the level of transcription, as well as through alterations in protein interaction and phosphorylation.

Another important ligand that synergizes with SCF is IL-7. The IL-7 receptor (IL-7R) consists of an and a subunit. Comparison of the IL-7^–/– and IL-7R ^–/– mice has shown that the phenotype is more severe in IL-7R ^–/– mice. This suggests the existence of an IL-7–independent way to activate the IL-7R. Interestingly, Kit interacts with and phosphorylates the IL-7R in 293T cells, and this led to activation of the JAK-STAT pathway in the absence of IL-7 [160].

G-CSF is another important hematopoietic growth factor that synergizes with SCF. Although G-CSF induces tyrosine and partial serine phosphorylation of STAT1 and STAT3, SCF only induces serine phosphorylation of STAT3 [161]. Stimulation with the combination of SCF and G-CSF results in complete serine phosphorylation of STAT3, which is necessary for maximal DNA binding. Furthermore, SCF and G-CSF induce STAT3 serine phosphorylation through different pathways, which to some extent are dependent on PI3K and Erk activity.

In total, these data support the possibility that synergistic growth induced by SCF in combination with other growth factors is mediated by multiple mechanisms that could include receptor interaction, amplification of activation of the Ras-Erk cascade, and activation of STAT family members, as well as synergistic expression of components of the cell cycle. Much remains to be learned about this interesting and important biological phenomenon.

Negative Regulators of Kit Signaling
Because Kit activates multiple signaling pathways, it is not surprising that negative regulation of this potent receptor also occurs through a variety of mechanisms.

PKC is a large family of structurally related serine/threonine kinases. PKCs are important in downregulating several RTKs, including the EGFR and Kit [162, 163]. In cells transfected with Kit, activation of DAG-responsive PKCs by soluble SCF occurs through a mechanism involving a PI3K-dependent activation of phospholipase D [116]. The down-regulation of Kit by PKC occurs in two ways: Activation of PKC induces proteolytic release of the extracellular ligand binding domain of Kit [119, 164], and PKC phosphorylates serine residues 741 and 746 in the kinase insert, which inhibits Kit kinase activity by an unknown mechanism [165].

The tyrosine phosphatase SHP1 (PTP1C, HCP, SH-PTP1) interacts with Kit in the JMD [88, 166]. One report suggestst hat SHP1 binds to tyrosine 569 of murine Kit [166]. Expression of Kit^Y569F in BaF3 cells results in increased SCF-induced proliferation compared with wild-type Kit. Furthermore, genetic experiments support a negative role for SHP1 in Kit signaling. Motheaten (me) mice have a loss-of-function mutation in SHP1, and these mice have hyper-proliferative hematopoietic progenitor cells [167]. Crossing me mice with mice expressing a mutated Kit receptor (W ^v) to produce mice harboring both mutations showed that this combination had a milder phenotype than me or W ^v mice [168, 169]. This suggests a functional relationship in which SHP1 negatively regulates Kit. Interestingly, loss of SHP1 had no effect on SCF-induced bone marrow–derived mast cell proliferation, suggesting a cell type–specific role of SHP1 in Kit signaling [169].

SOCS-1 protein levels increase after SCF stimulation of mast cells, and this leads to inhibition of proliferation [170]. Although SOCS-1 inhibits JAK and Tec by blocking kinase activity [171–173], it is unlikely that this mechanism inhibits Kit signaling. De Sepulveda et al. [170] demonstrated that SOCS-1 binds to Kit as well as several downstream signal transduction proteins, including Grb2 and the guanine nucleotide exchange factor Vav. This suggests that the inhibitory effect of SOCS-1 on Kit could be attributable to interference with coupling and function of downstream molecules. Interestingly, a recent study using BMMCs from SOCS-1–deficient mice demonstrated that these cells had a reduced proliferative response to both IL-3 and SCF compared with wild-type cells [174]. This unexpected finding was explained by accumulation of proteases in the absence of SOCS-1, resulting in degradation of essential signal transduction molecules. Hence, it appears that besides its function in negative feedback loops, SOCS-1 is also important for maintenance of certain protein levels.

SHIP is a 5'-inositol phosphatase hydrolyzing phosphatidylinositol-3,4,5-trisphosphate (PIP3) to PIP2. SCF stimulation of mast cells leads to SHIP phosphorylation [175]. Disruption of the SHIP gene has revealed that it is a negative regulator of SCF-induced mast cell degranulation [176]. Furthermore, SHIP ^–/–mice have granulocytes and macrophages with increased responsiveness to IL-3, GM-CSF, and SCF compared with wild-type cells, supporting a role for SHIP as a negative regulator of cytokine signaling [177].

The GTPase activating protein neurofibromin-1 (NF-1) has been suggested to be a negative regulator of Kit function [178]. In NF-1^–/– progenitor cells, SCF induces a stronger Erk activation and increases colony formation compared with normal cells. The NF-1 protein is a tumor suppressor commonly lost in juvenile myelomonocytic leukemia (JMML) [179]. It is not clear if Kit has a role in the development of JMML. Interestingly, it has been suggested that JMML originates from a pluripotent hematopoietic stem cell, indicating a potential role for increased Kit signaling in tumorigenesis [180].

Ubiquitin molecules are covalently attached to lysine residues on target proteins, and this labels them for degradation via the 26S proteosome complex. In RTK signaling, the ubiquitin ligase Cbl has an important role, because it can bind to activated receptors, or other tyrosine phosphorylated proteins, with its SH2 domain [181]. Activation of Kit leads to its poly-ubiquitination [118]. Cbl can associate with Kit through CrkL or APS and becomes phosphorylated in response to SCF [132, 139, 182]. Hence, it is likely that Cbl has an important role in Kit ubiquitination and subsequent downregulation.

Signaling and Cellular Context
When reading the literature on Kit signaling, it is important to note that studies have been done with a variety of experimental designs. This includes the use of primary cells from different species, primary cells from genetically engineered mice, cell lines expressing endogenous Kit, and cell lines that do not express endogenous Kit but are transfected with wild-type or mutant Kit. Although each of these approaches generates valuable information, the data must be interpreted in the context of the experimental approach. For example, in primary cells, there are differences in pathways activated by SCF in different lineages [67, 93, 109]. There are also differences in the outcome of disrupting the same pathway in different lineages [93, 109]. Thus, stimulus-response coupling mechanisms of Kit vary in mast cells, hematopoietic progenitors, lymphoid cells, melanocytes, and germ cells. This is also a consideration in work done with cell lines, in which lineage as well as extent of differentiation is a factor. In addition, as mentioned above, cell lines expressing endogenous Kit, as well as cell lines transfected with wild-type Kit or mutants, have been used extensively. An advantage of transfected cell lines is comparison of wild-type Kit and Kit mutants in a specific cell context. However, ectopic expression of Kit may generate an inappropriate biochemical or biological readout due to absence of certain signaling proteins, an inability to couple with Kit, or activation of pathways not physiologically relevant. In summary, consideration of the strengths and weaknesses of each of these experimental approaches is critical in integrating the vast amount of data available in this field.

	KIT AND HUMAN DISEASE

Top Abstract Introduction Hematopoiesis Signal Transduction Kit and Human Disease Oncogenic Kit Signaling Conclusion References

 
Abnormalities in the expression and function of Kit are associated with several human diseases. Although the primary focus of this article is Kit and hematopoiesis, recent work indicates that targeting Kit holds promise in the treatment of diseases in nonhematopoietic tissues also. Consequently, studies of both hematopoieitc and nonhematopoietic tissues will be included in this section.

Loss-of-Function Mutations
Autosomal-dominant piebaldism is associated with loss-of-function mutations in Kit [183]. This syndrome results in deafness, megacolon, and abnormalities in pigmentation of skin and hair. The pigmentation defect in persons with piebaldism is remarkably similar to that seen in heterozygous W mice. Study of megacolon in patients with piebaldism helped define the critical role for Kit in development of the ICC. Kit mutations found in patients with piebaldism include point mutations and deletions.

Inappropriate Expression of Kit or SCF
One mechanism leading to inappropriate activation of Kit is autocrine secretion of SCF. Tumors reported to have Kit/SCF autocrine loops include small cell lung carcinomas (SCLC), colorectal carcinoma, breast carcinoma, gynecological tumors, and neuroblastomas [184]. These data suggest that these diseases could respond to drugs that inhibit wild-type Kit. Indeed, in vitro, the tyrosine kinase inhibitor STI571 (imatinib mesylate/gleevec/glivec; Novartis Pharmaceuticals, St. Louis) does inhibit proliferation of SCLC cell lines [185, 186]. Furthermore, Krystal et al. [187] have identified two other Kit inhibitors, SU5416 and SU6597, that can inhibit proliferation of SCLC cell lines. A similar inhibitory effect of STI571 on proliferation of neuroblastoma cells was also recently demonstrated [188]. Thus, inhibition of Kit kinase activity may be one approach for treatment of diseases associated with autocrine activation of Kit.

Mast cells and hematopoietic progenitor cells express Kit on their surface, whereas most differentiated hematopoietic cells lack Kit expression. However, certain human leukemia cells, especially blasts from patients with acute myelogenous leukemia (AML) [43, 189], and certain non-hematopoietic tumors, such as gastrointestinal stromal cell tumors (GIST) [190] and seminoma germ cell tumors [191], also express surface Kit.

Kit is also expressed in cells such as cutaneous and choroidal melanocytes. Melanomas are variable with regard to chromosomal abnormalities, but loss of Kit expression seems to be associated with progression of certain forms of melanoma. Approximately 70% of metastatic lesions and human melanoma cell lines do not express detectable levels of Kit receptor [192]. Interestingly, transient expression of Kit in melanoma cell lines significantly inhibited tumor growth and metastasis [193]. Exposure of Kit-positive mela-noma cells to SCF, in vitro and in vivo, triggered apoptosis of these cells but not of normal melanocytes. These results suggested that loss of Kit may allow malignant melanoma cells to escape Kit-mediated apoptosis, thus allowing tumor growth and eventually metastasis. Further support for this possibility is a study demonstrating that the absence of Kit expression directly correlated with the metastatic potential of human melanoma cells in nude mice [194]. Although the exact mechanism for the loss of Kit gene expression during melanoma progression is unknown, Bar-Eli and colleagues [194] recently demonstrated that the transcription factor AP-2 is downregulated in melanomas. The Kit promoter has three AP-2 binding sites, suggesting loss of function of AP-2 plays a role in regulation of Kit expression and subsequently melanoma progression [195]. In contrast to these findings, Kit is expressed on choroidal melanoma tumors [196].

Gain-of-Function Mutations
In 1993, Furistu et al. [197] identified two gain-of-function mutations in Kit (V560G and D816V) in the human mast cell line HMC-1. These mutations map to the JMD and the second part of the kinase domain of Kit, respectively. A variety of activating mutations in the JMD have been identified, including deletions, point mutations, tandem duplications, and combinations of deletions and point mutations. These mutations are frequently associated with GISTs [198]. Similar JMD mutations have also been identified in RTKs closely related to Kit such as Flt3 (internal tandem repeats in the JMD of AML patients) and Met (JMD missense mutations or insertional mutations in SCLC) [199, 200].

Activating mutations in the second part of the kinase domain have been identified as point mutations at codon 816 of human Kit, which corresponds to codon 814 in murine Kit. Aspartic acid 816/814 is located in the activation loop and can be mutated to tyrosine, histidine, asparagine, or valine with similar consequences [201]. Activating mutations in the Kit kinase domain are associated with human malignancies such as mast cell leukemia, masto-cytosis [202, 203], acute myeloid leukemia (D816Y) [204, 205], and germ cell tumors (D816H) [206].

In a recent review, Longley et al. [207] classified activating mutations of Kit as either regulatory or enzymatic. The JMD mutations are considered regulatory in nature. These likely change the conformation of Kit and lead to constitutive dimerization and activation. This may also impact the binding of Kit to secondary signaling molecules. Mutations of the kinase domain are considered the enzymatic type. Interestingly, these two classes of mutations differ in their sensitivity to inhibitors of wild-type Kit, including STI571 and SU9529. JMD mutants are more sensitive to these inhibitors than the kinase domain mutants [208, 209]. In vitro studies have identified four small molecule inhibitors to be effective against the activated form of Kit mutant found in many patients with mastocytosis. These drugs, SU6577, SU11652, SU11654, and SU11655, all indolinone derivatives, are active against Kit JMD as well as kinase domain mutants [210, 211]. As of now there are no drugs identified that can specifically target the kinase domain mutants. Our laboratory is currently involved in screening for novel drugs to inhibit the kinase domain mutant of Kit.

In addition, a considerable number of patients with sinonasal natural killer/T-cell lymphomas have mutations in codon 825 of Kit [212]. Last, small subsets of patients with myeloproliferative disease contain point mutations in the extracellular domain of Kit, resulting in a D52N substitution [213, 214]. It is as yet unclear if gain of function of Kit is involved in disease progression in either of these cases.

The following sections will highlight the salient features of three of the most prevalent diseases associated with gain-of-function mutations in Kit, as well as summarize available therapeutic approaches.

Gastrointestinal Stromal Cell Tumors
Although GISTs constitute less than 1% of all tumors of the gastrointestinal (GI) tract, they are the most common form of mesenchymal tumors of the digestive tract [198]. GISTs occur in the GI tract from the esophagus to the rectum (5%) but are most commonly found in the gastric (6%–70%) and small intestine (25%–30%) regions. These tumors are believed to arise from the ICC or their mesenchymal stem cell precursors. On average, 10 to 20 GIST cases per million people are recorded annually. Of these, 20%–30% are malignant. Expression of Kit, readily detectable with immunohistochemistry, is often associated with this disease. Ninety percent of GIST patients express Kit, and many of them contain activating mutations in the JMD. Specifically, 57%–71% of GIST patients have mutations in exon 11, and 4%–17% of patients have mutations in exons 9 or 13. In certain rare cases, mutation in the kinase domain of Kit is also observed in GIST patients. The presence of these mutations is associated with poor prognosis. The percentage of patients surviving 5 years with GIST without mutations in Kit is 68%, compared with 49% for GIST patients with mutation in Kit.

As mentioned earlier, mutations in the JMD are considered regulatory and are believed to work through disruption of an inhibitory alpha helical structure [215]. This alpha helix is implicated in suppression of kinase activity and phosphorylation of Kit in the absence of ligand. The presence of mutations in this region is believed to relieve this inhibitory configuration, rendering Kit constitutively active.

Although most GISTs have activating mutations in Kit, a certain population of GIST patients lacks Kit mutations [216, 217]. These findings suggest that there are other mechanisms that could potentially result in GIST. A recent study by Heinrich et al. [218] has demonstrated that 35% of GIST patients (14 of 40) did not carry a Kit mutation but instead had mutations in PDGFR, a related RTK [218]. They also demonstrated that the tumors with either Kit or PDGFR mutations did not differ in their potential to activate downstream signaling molecules such as Akt, STAT1, STAT3, and Erk1/2. These tumors, irrespective of a mutation in Kit or PDGFR, have deletions in chromosome 1p and monosomies of chromosomes 14 and 22. These results imply that Kit and PDGFR could be alternative and mutually excusive oncogenic mechanisms leading to GISTs.

GISTs represent a wide spectrum of disease, ranging from benign to highly malignant forms. This makes prediction of clinical responses difficult. The malignant potential of GIST relies on tumor size, mitotic activity, and anatomical origin. Certain conventional treatment strategies are available to treat GISTs. Surgical resection is currently the only option for treatment of primary GIST. Unfortunately, these tumors are fragile and often require secondary removal. Adjuvant therapy is useful after surgical resection in patients with rectal GISTs with close or positive microscopic margins [219]. Metastatic GISTs are treated with conventional chemotherapeutic agents such as doxorubicin and ifosfamide together with surgical resection. However, conventional chemotherapy has limited efficacy. For patients with liver metastasis, hepatic artery embolization (i.e., cutting off the primary blood supply to liver) is an alternative that may be used in combination with traditional chemotherapy [220]. In certain rare cases, radiotherapy is also used mostly for palliation or to stop bleeding from peritoneal recurrence [221].

More recent, the tyrosine kinase inhibitor STI571, which is active against BCR-ABL fusion protein in chronic myelogenous leukemia (CML) as well as PDGFRs and Kit, has been shown to be effective in the treatment of GIST [222]. STI571 has recently been cocrystallized with Kit [223]. STI571 binds near the active site and stabilizes a conformation that is incompatible with kinase activity. Because of its ability to directly target the JMD mutant of Kit and its low toxicity in CML patients, use of STI571 as a treatment for GIST is increasing. In 2001, Joensuu et al. [224] reported promising results of treating a patient with metastatic GIST with STI571. Data obtained in follow-upstudies indicated STI571 had minimal toxicity and substantial activity against GIST. There was a partial response rate of 59%, with 23% showing disease stabilization [225]. Only approximately 13% of patients experienced progression of disease. STI571 is now considered a prime adjuvant therapy for primary GISTs after resection. Most patients are at high risk of recurrence after resection and do not respond to conventional therapies. Thus, the adjuvant therapy is expected at least to delay recurrence and hence prolong survival. However, because STI571 does not elicit a complete response, it may be best used in combination with conventional therapies. Despite the effectiveness of STI571, there are concerns regarding the development of resistance. In certain patients with CML, gene amplification and point mutations in the kinase domain in response to STI571 have already been demonstrated [226–228].

Mastocytosis
Mastocytosis is characterized by accumulation of mast cells in various organs and release of mast cell mediators. Mastocytosis is reported in 1,000 to 8,000 patients in the U.S. every year belonging to all races and both sexes [229]. Peak incidence is during infancy and early childhood, with a second peak occurring in middle age. It usually occurs as a transient and limited disease in children, whereas it is persistent and progressive in adults. Because of its heterogeneous clinical presentation, mastocytosis is classified into four categories. Indolent mastocytosis is the most common form and involves the skin, bone marrow, and GI tract. Clinical features range from a single cutaneous nodule to multiple pigmented macules resulting from increased epidermal melanin and papules (urticaria pigmentosa) or diffuse cutaneous involvement; bullae, vesicles, and abnormal telangiectasia may be seen. GI involvement leads to symptoms such as nausea, vomiting, and abdominal pain. Mastocytosis with an associated hematologic disorder refers to cases in which urticaria pigmentosa symptoms are accompanied by a variety of hematological findings due to mast cell infiltration of the bone marrow, spleen, liver, and lymph nodes. Mast cell leukemia is characterized by proliferation and infiltration of immature mast cells in bone marrow, peripheral blood, and various extramedullary tissues. Aggressive mastocytosis is characterized by severe involvement of bone marrow, liver, spleen, and lymphatic system.

The prognosis of mastocytosis is highly dependent on the form. Although indolent mastocytosis, the most common form, is not generally fatal, the other three forms can be.

Interaction of SCF and Kit is critical for mast cell development, survival, and function. Although abnormalities in this receptor ligand combination have been implicated in mastocytosis, mechanisms mediating the diverse clinical patterns have not been elucidated. As previously discussed, activating mutations in Kit have been reported in a variety of patients with mastocytosis. Most of these mutations map to the kinase domain of Kit. In contrast, familial mastocytosis may occur in the absence of Kit mutations [230]. These mutations are summarized in Table 2. Direct and indirect evidence supports the contention that these mutations promote ligand-independent autophosphorylation of the mutant receptor. This will be discussed in greater detail in the following section dealing with oncogenic Kit signaling.

Altered expression of SCF may also be involved in mastocytosis. The soluble form of SCF is prevalent in the dermis and extracellular spaces between kertinocytes in skin from mastocytosis patients [231]. Altered distribution of SCF in the skin of patients with cutaneous mastocytosis, a disease characterized by dermal accumulations of mast cells and increased epidermal melanin, is consistent with abnormal production of the soluble form of SCF. This likely results from an increase in cleavage of the membrane-bound form of SCF [232]. This suggests cutaneous mastocytosis could be the result of a hyperplastic response to a local increase in soluble SCF in the tissue.

To date, there is no cure for mastocytosis. Patients are asked to avoid classic mast cell triggers, as well as extreme temperatures, alcohol, and certain drugs. Pharmacological treatment is aimed at reducing discomfort associated with symptoms and includes stabilizing mast cell membranes to decrease the extent of degranulation in conjunction with blocking the action of inflammatory mediators. Surgical removal of isolated mastocytomas, topical corti-costeroid therapy, and psoralen plus ultraviolet A therapy to fade multiple pigmentation lesions are other available approaches for treatment.

Acute Myelogenous Leukemia
Wild-type Kit is expressed in most cases (80%–90%) of human AML and in some cases during CML blast crisis [184]. The presence of Kit has been attributed to the immature stage of differentiation of AML cells [233]. In contrast, Kit is rarely expressed in leukemia cells of lymphoid origin. In addition, Kit-positive AML specimens show constitutive phosphorylation of Kit [189, 234, 235]. Kit is also activated in some AML blasts in the absence of SCF. Furthermore, SCF can stimulate proliferation of certain AML blasts in vitro [189, 234, 235]. Autocrine and paracrine stimulation of Kit by SCF may thus play a role in the pathobiology of AML. Indeed, it has been shown that approximately 30% of AML blasts coexpress Kit and SCF [236]. In addition, Gewirtz et al. [44] reported that 25% of AML blasts showed decreased growth in the presence of Kit antisense oligonucleotides. The same study also found this to be the case for 33% of ALL and 50% of CML blasts. For a detailed review on autocrine/paracrine loops in normal and oncogenic hematopoiesis, please refer to the study by Janowska-Wieczorek et al. [30].

One third of the cases of AML with inv (16) karyotype also have deletion and insertion in exon 8 that code for the Kit extracellular domain [237]. Mutations in codon 416 are involved in all of these cases, and this is presumed to be a gain-of-function mutation. However, this mutation has not been linked with increased activation of Kit [237]. The other site for Kit mutations in patients with AML is codon 816 (D816Y or D816V). This mutation was observed in 7 of 15 patients with core binding factor leukemia [t(8;21) or inv(16)] [238, 239]. The heterodimeric core binding transcription complex is crucial for normal hematopoiesis [240]. Interestingly, mutations and rearrangements of the core binding complex have often been found in leukemia. However, the functional significance of the correlation between the translocations of core binding factors and Kit mutations is as-yet unclear.

Although a significant number of Kit-positive AMLs have been identified, there is still insufficient evidence to use Kit as a prognostic marker. Overall, Kit expression does not correlate well with cytogenetic analysis. Despite accepted prognostic indicators, approximately 40% of patients with newly diagnosed AML do not achieve complete remission after conventional therapy. The response rate can be improved to 50%–75% with the use of sibling bone marrow transplant therapy. Recent studies have shown that the small molecule inhibitors SU5416 and SU6668 effectively block kinase activity of wild-type Kit in cultured AML blasts from patients [241, 242]. Complex biological defects in hematopoietic progenitor cell differentiation, proliferation, and survival are critical in the pathogenesis of AML and pose significant challenges in development of treatment strategies.

AML is also associated with mutations in Flt3, a RTK closely related to Kit. These occur as either internal tandem duplications of the JMD (25%–30%) or point mutations in the kinase domain region (7%–8%) [243].